Langenbecks Arch Surg (2008) 393:759–815 DOI 10.1007/s00423-008-0409-0

ABSTRACTS

12th Annual Meeting on Surgical Research/12. Chirurgische Forschungstage: Chirurgische Forschung — Brückenschlag zwischen Klinik, Forschung und Industrie, 25.–27. September 2008, Freiburg i. Breisgau, Germany © Springer-Verlag 2008

President: Prof. Dr. Dr. h. c. Ulrich T. Hopt, Abteilung Allgemein- und Viszeralchirurgie, Chirurgische Universitätsklinik Freiburg, Hugstetter Str. 55, 79106 Freiburg, Germany Local Organisers: PD Dr. med. Tobias Keck PD Dr. med. Ernst von Dobschütz

1 IN VIVO ENGINEERING OF A HUMAN VASCULATURE FOR BONE TISSUE ENGINEERING APPLICATIONS Günter Finkenzeller, Lilian Steffens, Sven Graner, G. Björn Stark Department of Plastic and Hand Surgery, University of Freiburg Medical Center, Freiburg, Germany Keywords: endothelial cell, angiogenesis, vascularization, bone tissue engineering, SCID mouse Background: Neovascularization represents a major challenge in tissue engineering applications since implantation of voluminous grafts without sufficient vascularity results in hypoxic cell death of implanted cells. An attractive therapeutic approach to overcome the problem of insufficient vascularization of grafts is based on the coimplantation of endothelial cells to create vascular networks. In this context, we have developed a three-dimensional spheroidal coculture system consisting of human primary endothelial cells (EC) and human primary osteoblasts (hOB) to improve angiogenesis in bone tissue engineering applications. In this study, we investigated the survival and vascularization of the engineered implants in vivo. Methods: The in vivo vasculogenic potential of human endothelial cells was evaluated in a xenograft model where spheroidal endothelial aggregates were co-implanted with human primary osteoblasts subcutaneously into immunodeficient mice. After various time points, the implants were retrieved and analyzed for the formation of a human EC-derived neovasculature using antibodies to human CD31 and CD34 to visualize transplanted EC and anti -SMA to detect recruited murine mural cells. Survival of transplanted human osteoblasts and formation of a calcified extracellular matrix was evaluated by anti-osteonectin and alizarin red stainings. Results: Subcutaneous co-implantation of human spheroidal endothelial aggregates and human osteoblasts into SCID mice resulted in the

formation of a complex three-dimensional network of human blood vessels. The human vasculature anastomosed with the mouse vasculature at the scaffold boundary, matured by recruiting mouse mural cells, and was connected to the mouse circulation. Anti-osteonectin and alizarin red stainings revealed that the co-implanted human osteoblasts survived and formed calcified bone nodules. Conclusions: We conclude that this EC-spheroid system can be used to create a network of functional perfused blood vessels in vivo. The finding that this process takes place with high efficacy in the presence of co-implanted primary osteoblasts suggests that this system may be suitable for improving vascularization in bone tissue engineering.

2 RAPAMYCIN, BUT NOT CYCLOSPORINE A, INHIBITS VASCULARIZATION AND INCORPORATION OF IMPLANTED SURGICAL MESHES Jonas Roller, Matthias W. Laschke, Jörg M. Häufel, Michael D. Menger Institute for Clinical & Experimental Surgery, University of Saarland, Homburg/Saar, Germany Keywords:rapamycin, cyclosporine, surgical mesh, hernia repair, angiogenesis, vascularization Background: Incisional hernias are a frequent complication of upper abdominal wall interventions, especially in patients undergoing liver transplantation with subsequent immunosuppressive therapy. This may be due to the fact that immunosuppressive therapy impairs the formation of granulation tissue. Therefore we analyzed in the present study, how the incorporation of a surgical mesh for hernia repair is affected by the immunosuppressant drugs rapamycin and cyclosporine A (CsA). Methods: Ultrapro meshes were implanted into the dorsal skinfold chambers of Syrian golden hamsters, which were treated daily with rapamycin (1.5 mg/kg i.p.; n=5) or CsA (20 mg/kg i.p.; n=5). Untreated animals served as controls (n=7). Subsequently, we analyzed the angiogenic and inflammatory host tissue response to the mesh implants over a 14-day period by means of intravital fluorescence microscopy. Moreover, mesh incorporation was determined histologically by sirius red staining to assess the collagen content of the newly formed granulation tissue surrounding the implants and by measurement of the explantation strength. Results: Rapamycin inhibited vascularization of implanted meshes, as indicated by a significantly reduced number of angiogenesis positive regions of interest (48±14%; p<0.05) and microvessel density (5.5± 1.6 cm/cm2; p<0.05) at day 14, when compared to CsA-treated hamsters (100±0% and 129.9±27.6 cm/cm2) and untreated controls (100±0% and 110.2±15.7 cm/cm2). In addition, the granulation tissue

760 surrounding the meshes of rapamcyin-treated animals exhibited only a low collagen content, resulting in an impaired mesh incorporation with a significantly reduced explantation strength (0.6±0.2cN/mm2 vs. 6.7± 1.3cN/mm2 and 7.4±0.8cN/mm2; p<0.05). Leukocyte-endothelial cell interaction, indicating inflammatory host tissue response, did not markedly differ between the three observation groups. Conclusions: Our novel findings indicate that immunosuppressive therapy with rapamycin impairs incorporation of surgical meshes by inhibition of angiogenesis and collagen synthesis within the newly developing granulation tissue around the implants. Thus, immunosuppressed patients should not be treated with rapamycin in case of incisional hernia repair in order to guarantee adequate mesh incorporation.

3 ANGIOGENESIS INHIBITION BY ENDOSTATIN AND TUMSTATIN VARIES DEPENDING ON THE HOST ORGAN ENVIRONMENT Moritz Koch, Nuh Rahbari, Namali T. Fernando, Sam S. Yoon, Jürgen Weitz Chirurgische Universitätsklinik Heidelberg, Division of Surgical Oncology, Massachusetts General Hospital, Harvard Medical School, Boston, USA Keywords: endogenous angiogenesis inhibitors, host organ environment, tumstatin, endostatin Background: Anti-angiogenic agents are being increasingly utilized for the treatment of metastatic cancers. Endostatin and tumstatin are matrixderived endogenous angiogenesis inhibitors, but it is unclear if they have equal efficacy in suppressing angiogenesis in different organ sites. Methods: Cell lines stably overexpressing endostatin and tumstatin were generated using murine renal carcinoma cells (RenCa) and colon carcinoma cells (CT26). These cell lines were analyzed in vitro for effects on endothelial cell proliferation and in vivo for effects on metastasis formation. Results: Endostatin and tumstatin cell lines showed equal growth rates compared to control cell lines in vitro. Conditioned media from endostatin- or tumstatin-secreting cells inhibited endothelial cell proliferation up to 53%. Flank tumor growth was greater than 1000 mm3 for RenCa and CT26 control cells compared to less than 200 mm3 for endostatin or tumstatin cell lines two weeks after tumor inoculation with a concomitant decrease in microvessel density. Endostatin overexpression in RenCa cells markedly reduced the formation of experimental lung metastases but had no effect on the formation of liver metastases, while tumstatin overexpression in CT26 cells inhibited liver metastases but had no effect on lung metastases. Western blot analysis of the endostatin receptor (alpha5beta1 integrin) and tumstatin receptor (alphavbeta3 integrin) revealed varying protein and isoform levels in liver and lung tissue and metastases. Conclusions: The endogenous angiogenesis inhibitors endostatin and tumstatin have widely variable efficacy in inhibiting metastases depending on the host organ environment. Clinical trials of these agents should take into consideration the sites of metastatic disease.

4 DARBEPOETIN PROMOTES ANGIOGENESIS AND GROWTH OF INTRAHEPATIC COLORECTAL METASTASES AFTER LIVER RESECTION Kathrin Rupertus, Markus Corsten, Martin K. Schilling, Michael D. Menger, Otto Kollmar Department of General, Visceral, Vascular and Pediatric Surgery and Institute for Clinical and Experimental Surgery, University of Saarland, 66421 Homburg/Saar, Germany

Langenbecks Arch Surg (2008) 393:759–815 Keywords: colorectal cancer, liver metastasis, DPO, EPO, angiogenesis Background: Mortality related to colorectal cancer results from uncontrolled metastatic disease. More than 50% of patients present with synchronous or metachronous liver metastases. Radical surgical resection of liver metastases is the only curative treatment and can be performed with low mortality in specialized centers. After major hepatectomy, postoperative hepatic failure still remains an unsolved problem. In experimental models, erythropoietin (EPO) and its analogue darbepoetin (DPO) have been shown to improve liver function after toxic liver failure. In liver malignancies treatment with EPO is still discussed controversially, suggesting that EPO could stimulate growth of tumor cells or micrometastases which may lead to early tumor relapse. We therefore studied the influence of DPO-treatment on tumor growth in a model of experimental colorectal liver metastasis. Methods: 29 female BALB/c mice were injected with 1× 105 syngeneic CT26.WT colorectal cancer cells into the left liver lobe and were randomized into 4 groups: one group received sham treatment and served as control (n=8; control). Animals of the second group received 10 µg/kg BW DPO injected into the portal vein (n=7; DPO). The third group underwent a 50% hepatectomy before tumor cell implantation (n=7; Phx). The fourth group additionally received DPO injection after 50% hepatectomy and tumor cell implantation (n=7; Phx+DPO). After 7 days, animals were relaparotomized to study tumor volume, growth characteristics, angiogenesis and leukocyte adhesion using intravital fluorescence microscopy. Results: Application of DPO after tumor cell implantation lead to a slight, but not significant increase of tumor volume as compared to controls (0.11 vs. 0.15 mm3), whereas Phx significantly increased tumor volume (0.20 mm3). Additional application of DPO after Phx caused a further stimulation of tumor growth with significantly increased tumor volume compared to Phx only (0.27 mm3). DPO also induced increased angiogenic activity in treated tumors. The angiogenic front surrounding the hepatic tumors was significantly larger in the DPO group than in controls (607 vs. 408 µm) as well as the number of draining tumor venules (15 vs. 8). After Phx, DPO injection also increased the number of draining tumor venules (17 vs. 10) and caused an enlargement of the angiogenic front. The capillary density of tumor vessels was significantly higher after treatment with DPO compared to sham treatment (24.92 vs. 70.94 cm/cm2). DPO after Phx caused an increase of capillary density as compared to Phx alone (70.87 vs. 45.31 cm/cm2). Capillary diameters were not significantly affected. After Phx, the normal liver showed a significant increase in leukocyte adherence as compared to non-resected controls (8.95 vs. 2.98/mm2). DPO reduced the number of adherent leukocytes compared to control (0.68/mm2) and Phx (2.39/mm2). This effect was also visible within the tumors where the number of the adherent leukocytes was significantly reduced after DPO treatment. Conclusions: Our study shows for the first time that DPO stimulates tumor growth especially after liver resection due to an increased angiogenesis of the colorectal metastases within the regenerating liver. These results indicate that the use of DPO as a stimulator of liver regeneration after liver resection in the context of liver malignancies should therefore be used cautiously to avoid early relapse after surgical treatment of liver tumors.

5 IMPROVED REGENERATION OF AUTOLOGOUS NERVE TRANSPLANTS BY MEANS OF VEGF-GENE THERAPY Thomas Holzbach (1), Rupprecht Milojcic (1), Martina Anton (2), Thomas Brill (2), Moritz A. Konerding (3), Bernd Gänsbacher (2), Hans-Günther Machens (1), Riccardo E. Giunta (1) 1: Department of Plastic Surgery and Hand Surgery (Head: Univ.Prof. H.-G. Machens), 2: Institute for Experimental Oncology (Head: Univ.-Prof. B. Gänsbacher)/Klinikum rechts der Isar, Technische Universität München, Munich, Germany; 3: Department of Anatomy, Johannes Gutenberg Universität, Mainz, Germany

Langenbecks Arch Surg (2008) 393:759–815 Keywords: nerve regeneration, VEGF, Gene Therapy Background: The impact of the Vascular Endothelial Growth Factor (VEGF) on the angiogenic cascade is proven. Recently its neuroprotective effect after peripheral nerve injuries on α-motoneurons in the spinal cord was shown. The biological effect is mainly mediated by the binding of two tyrosine kinase receptors (VEGFR1&VEGFR2), but in addition an effect via Neuropilin (NP)-1 and NP-2, receptors essential for the development of the nervous system, was reported. Thus a sprouting of axons could be monitored as well as an improved survival of neurons and glial cells. Experiments on α-motoneurons demonstrated a decreased sensitivity to ischemia under VEGF-therapy. Aim of the study was to elucidate the effect of a localized VEGF-gene-therapy using an adenoviral vector construct in the model of a peripheral nerve defect in the rat treated with an autologous nerve transplant. Methods: A 2 cm segment was resected in the course of the right sciatic nerve of the rat (n=24) and reversely coapted under the microscope in terms of an autologous interposition. Thereafter we injected a total volume of 200μl Ad.CMV.VEGF165 (108pfU) in 4 fractions into the surrounding muscle and sheet. During the trial period of 18 weeks we conducted weekly walking-track and static foot-print-analyses. After 18 weeks the gastrocnemius muscle was evaluated electrophysiologically and weighed. The axons in the sciatic nerve and the α-motoneurons of the corresponding neuronpool in the spinal cord were counted. Additionally we evaluated the coaptation sites histologically. Results: In the experimental groups we detected signs of reinnervation earlier than in controls, the innervationindex at the end was significantly increased in VEGF-experimental groups (66% of opposite side) when compared to controls (48%; p<0.05). The muscle weight was significantly increased, as well (57% vs. 48% of opposite side; p<0.05). The electrophysiological assessment yielded significantly higher amplitude between tarsus and trochanter (p<0.05) with comparable conduction time. Conclusions: Local VEGF-Genetherapy produced a quicker nerval regeneration with an improved functional outcome in this setting. These results cannot only be explained by a quicker incorporation of the transplant by dint of induced angiogenesis but rather by a direct impact on neurons resp. axons. Notwithstanding the impact of these findings a transfer into clinic appears limited due to the used vector construct.

6 INHIBITION OF THE ENZYME DIHYDROOROTATE DEHYDROGENASE BY LEFLUNOMID AND THE ACTIVE METABOLITE A 771726 INHIBITS IN VITRO TUMOUR CELL PROLIFERATION AND IN VIVO TUMOUR GROWTH Carsten Dietz, Sebastian Hennig, Julia Sara Becker, Verena Welter, Elias Karakas, Detlef Bartsch, Ilhan Celik Department of Visceral, Thoracic and Vascular Surgery, Institute of Theoretical Surgery, Philipps University Marburg, Germany Keywords: Leflunomid, tumour growth inhibition, BxPC-3, SCID Background: Leflunomid is in clinical use for the treatment of rheumatoid arthritis. Inhibition of the enzyme dihydroorotate dehydrogenase (DHODH) leads to antiproliferative effects of proliferating cells by the reduction of the pyrimidine nucleotide pool. Aim of the study was the investigation of the inhibitory effect of the active metabolite A 771726 in vitro with the human pancreatic carcinoma cell line BxPC-3 and in vivo a comparison between the pro-drug leflunomid and the active metabolite A 771726 according to tumour growth inhibition. Methods: Male immunodeficient mice (SCID), 6–8 weeks old were used. BxPC-3 pancreatic cancer cells (5×106) in 0.2 ml RPMI 1640 medium were implanted subcutaneously (s.c.) in the dorsa of the mice (n=5/group). Tumour volumes were measured every 3–5 days with

761 the digital calliper. Mice were randomised in therapy and control groups when tumour size reached 100+/−20 mm3. Animals in the 3 therapy groups were treated with 0,5 mg, 0,25 mg leflunomid and 0,25 mg A 771726 by daily i.p. application. The control group received placebo. Therapy was continued for a period of 21 days. Proliferation assay was performed for 96 hours in 24 well plates. The final cell amount with or without drug incubation was analysed using the coulter counter model Z1. Results: During 21 days therapy (i.p.) with leflunomid (0,25 and 0,5 mg/mouse) tumour growth was inhibited by 46% and 79%. The inhibition rate with A 771726 was 77% versus placebo. During the whole experiment there was a weight loss observed in the animals. In vitro investigation revealed an IC50 value of 37 µM for A 771726. Conclusions: The in vitro data has shown an IC50 value which is after searching the literature data bases easily reproducible in vivo. We found a dose dependent inhibition of tumour growth in this experiment. Further studies should be done to analyse also the mechanism of the therapy. CD 31, TUNEL and Ki-67 staining are necessary for the tumour sections for a better interpretation of the results. The side effects in our experiment should be reduced if the drugs are not diluted in DMSO, so that optimal dosing and application will also be investigated in further experiments.

7 NEOANGIOGENESIS IMPAIRS BIOMECHANICAL PROPERTIES DURING TENDON HEALING Hacer Sahin, Nancy Tholema, Britta Wieskötter, Steffen Schanz, Michael J. Raschke, Wolf Petersen, Richard Stange Klinik für Unfall-, Hand- und Wiederherstellungschirurgie, Universitätsklinikum Münster, Germany Keywords: Angiogenesis, degenerative tendon disease Background: Recent studies were able to demonstrate an important role of neoangiogenesis in degenerative tendon diseases. Vascular endothelial cell growth factor (VEGF) is known to be one of the key factors in angiogenesis. VEGF expression is regulated by the transcription factor hypoxia inducible factor 1 (HIF-1), which consists of α- (120 kDa) and β-subunit (about 90 kDa). Several factors are involved in the regulation of VEGF expression, like hypoxia, inflammatory cytokines and mechanical load. The way how VEGF influences biomechanical properties of the tendons is not well understood. We hypothesised that impairment of the tendon leads to vascular disruption and to local hypoxia due to micro traumata and micro ruptures. Hypoxia along with inflammatory cytokines stimulates the VEGF production which leads to matrix metalloproteinases (MMPs) induction and proteolysis of the extracellular matrix. The hypothesis was investigated in vivo in a rat tendon healing model using biomechanical and histological methods. Methods: An in-situ freezing model modified to Yamamoto [Yamamoto et al., 2000] was established in 5 months old wistar rats. Using liquide nitrogen the patellar tendon was repetitively frozen (3×10 sec) to create a tendon disease. 0, 7, 14 and 28 days after surgery, animals were sacrificed and the patellar tendons dissected. For immunohistochemical analysis (n=8) sections of the patellar tendons were immunostained with anti-HIF-1α, anti-MMP3 and anti-VEGF and quantified by digital image analysis. The microvessel density was determined by smooth-muscle-actin staining and also quantified. For biomechanical testing (n=8) patellar tendons were harvested along with proximal tibia, and patella. Subsequently, a load to failure test was performed with a speed of 0.1 mm/s to messure max. load, tension and E modulus. Statistical differences are analyzed by using Mann-Whitney-U-Test and P<0.05 was considered to be significant. Results: Immunohistochemical staining showed a peak in HIF1expression immediately after surgery. Both, VEGF

762 and MMP-3 were significantly increased 7 days after surgery and afterwards returned back to baseline. Along with the increase of VEGF and MMP3 expression, vascularisation was also significantly elevated after 7 days. In contrast, the biomechanical stability of the tendon was significantly decreased after 7 days and raised after 14 and 28 days. Conclusions: We were able to demonstrate that increased VEGF levels lead to a decrease in biomechanical stability of the tendon in vivo due to MMP-3 expression and its proteolytic activity. These findings support our hypothesis of VEGF-induced neoangiogenesis that leads to impaired mechanical tendon stability. Thus VEGF seems to play a decisive role in the development of degenerative tendon diseases. The modulation of neoangiogenesis might be a promising target for new therapeutic approaches in degenerative tendon diseases.

8 IMPAIRED IN VIVO VASCULOGENIC POTENTIAL OF ENDOTHELIAL PROGENITOR CELLS IN COMPARISON TO HUMAN UMBILICAL VEIN ENDOTHELIAL CELLS IN A SPHEROID-BASED IMPLANTATION MODEL Sven Graner, G. Björn Stark, Günter Finkenzeller Department of Plastic and Hand Surgery, University of Freiburg Medical Center, 79106 Freiburg, Germany Keywords: angiogenesis, vasculogenesis, endothelial cell, endothelial progenitor cell, SCID mouse Background: The lack of a sufficient vascularization is a limiting factor in tissue engineering applications. Without connection to a vascular system, survival of voluminous biological grafts is not possible. Therapeutic approaches to overcome the problem of insufficient vascularization comprise the use of angiogenic growth factors embedded into appropriate scaffolds to promote ingrowth of microvessels and cell-based concepts to enhance vasculogenesis. In this study, we investigated the potential of human endothelial progenitor cells (EPCs) to form functional blood vessels in vivo in direct comparison to mature endothelial cells, represented by human umbilical vein endothelial cells (HUVECs). Methods: EPCs were isolated from human peripheral blood, expanded in vitro and analyzed in vitro for phenotypical and functional parameters. The in vivo vasculogenic potential of EPCs and HUVECs was evaluated in a xenograft model where spheroidal endothelial aggregates were implanted subcutaneously into immunodeficient mice. Results: EPCs were indistinguishable form HUVECs in terms of the expression of the classical endothelial markers CD31, vWF, VECadherin and VEGF-R2 and in their ability to endocytose acetylated low-density lipoprotein. Moreover, EPCs and HUVECs displayed an almost identical angiogenic potential in vitro, as assessed by an in vitro matrigel sprouting assay. However, in vivo, a striking and unexpected difference between EPCs and HUVECs could be detected. Whereas implanted HUVEC spheroids gave rise to the formation of a stable network of perfused microvessels, implanted EPC spheroids showed a significantly impaired ability to form vascular structures under identical experimental conditions. Conclusions: Our results indicate that mature endothelial cells, such as HUVECs, are superior to EPCs in terms of promoting in vivo vascularization of engineered tissues.

9 THE INFLUENCE OF NANOSILVER ON HUMAN MESENCHYMAL STEM CELLS Christina Greulich, Stefanie Kittler, Matthias Epple, Manfred Köller Institute of Surgical Research, BG Kliniken Bergmannsheil, University Hospital of Bochum, Germany

Langenbecks Arch Surg (2008) 393:759–815 Keywords: silver-nanoparticles; human mesenchymal stem cells; proliferation; cytokine Background: Nanosilver (silver-nanoparticles, silver NPs) has wellknown antimicrobial activities and is increasingly used in different areas, such as, clothes, catheters, electric home appliances, food, and electronic industry. Apart from the antimicrobial properties there is only limited information on the biological effects of nanosilver exposure on living cells. Methods: It was the purpose of this study to analyze the effects of silver NPs on the cell proliferation and the release of cytokines of human mesenchymal stem cells (hMSCs). The silver NPs were prepared by the reduction of silver salts with solvent ethylene glycol and then stabilized by the polymer polyvinylpyrrolidone (PVP). HMSCs were incubated with different concentrations of silver NPs up to 7 days under cell culture conditions. In addition, control experiments with dissolved silver ions (CH3COOAg) were performed to separate particle and ion effects. The proliferation of hMSCs was measured using the alamarBlue assay and cytokine release was analyzed using enzyme-linked immunosorbent assay (ELISA). Results: Our results demonstrate a concentration-dependent toxicity. At levels of over 5 µg/ml silver NPs were found to have a significant cytotoxic effect on hMSCs. The cytokine production was significantly inhibited (IL-6) or activated (IL-8) by the silver NPs. Conclusions: Although silver NPs had a cytotoxic effect on hMSCs at a concentration of over 5 µg/ml, a concentration-dependent increase in cell activation indicated by synthesis of IL-8 was observed. Decreased cytokine generation rates at high particle concentrations were related to cytotoxic effects. These data clearly show that silver NPs are able to activate mesenchymal stem cells at sub-lethal concentrations.

10 A NEW EXPERIMENTAL APPROACH TO INVESTIGATE THE BIOCOMPATIBILITY OF NICKEL-TITANIUM UNDER DYNAMIC MECHANICAL LOADING Tim Habijan, Thomas Glogowski, Sebastian Kühn, Michael Pohl, Gert Muhr, Manfred Köller Bergmannsheil University-Hospital/Surgical Research, Ruhr-University Bochum, Germany/ Faculty of Mechanical Engineering/Material Testing, Ruhr-University Bochum, Germany Keywords: nickel-titanium, human mesenchymal stem cells, biomaterial, tensile testing machine Background: Nickel-Titanium shape memory alloys (NiTi) are of biomedical interest due to an unusual range of pure elastic deformability (pseudo-elasticity) with an elastic modulus (28 GPa) closer to that of bone (0,3–20 GPa) than any other metallic or ceramic material. However, the high nickel content of NiTi-SMA may result in adverse tissue reactions especially with long-term implants under mechanical strain. In NiTi-SMA, nickel and titanium are distributed in a regular crystal lattice order, exhibiting high atomic bonding forces with mixed covalent and metallic character, thus it is difficult for nickel to leave the bulk material. The surface of NiTi-SMA is well passivated because titanium is more readily oxidised than nickel. Therefore, NiTi-SMA devices exhibit a Ti-based oxide layer which is responsible for the corrosion resistance of this material, and acts as an effective barrier to nickel ion release. From a biological point of view, the integrity of the outermost surface layer appears to be of crucial importance for the biocompatibility of NiTi-SMA. However, mainly orthopaedic implants in the body are exposed to loading/unloading conditions, being deformed in an elastic manner, and biocompatibility studies are often performed under static conditions. Stress may aggravate metallic ion release resulting in a degradation of both the body tissue and

Langenbecks Arch Surg (2008) 393:759–815 implant. Systematic studies on the corrosion behaviour of NiTi-SMA under dynamic conditions are missing. Methods: In order to analyze biocompatibility of mechanically strained NiTi-SMA, tensile specimens were preloaded with human mesenchymal stem cells (hMSCs). The specimens were transferred to a sterile poly-tetrafluoroethylene (PTFE) cell culture tube equipped with a cell culture perfusion system and fixed to the pull rods of a tensile testing machine. The cell culture tube was located into a conventional cell culture incubator located within the tensile testing machine. 86.400 strain cycles were performed for a period of 24 h and 7d. Subsequently, the cell culture tube was disconnected from the tensile testing machine and transferred to a cell culture hood. The cell culture medium was aspirated and stored at −80°C. Interleukin-6 and nickel ion release were determined. The cells on the tensile specimen were stained by calceinAM and propidium iodine to analyze cell viability. Results: The dynamic loading under the used conditions did not influence the biocompatibility of NiTi-SMA. The release of IL-6 from hMSCs cultured under dynamic conditions was significantly higher after mechanical load compared to static conditions. Dynamic cycles of loading and unloading increased the nickel ion release from the tensile specimen. Conclusions: The presented experimental approach will provide information on the biocompatibility and fatigue behavior of metallic specimens using sample size and dynamic strain relevant for orthopedic implants. We successfully established a cell culture perfusion system which provides nearly in vivo conditions for the culture and testing of living cells onto tensile specimens, bridging the gap between static in vitro cell culture experiments and in vivo animal experiments.

11 SURFACE MODIFICATION OF TRACHEAL STENTS BY SILVER NANO-PARTICLES TO PREVENT BACTERIAL COLONIZATION Thorsten Walles, Heike Mertsching, Iris Trick, Michael Müller, Christian Oehr Dept. of Thoracic Surgery, Klinik Schillerhöhe, Gerlingen, and Depts. of Cell and Tissue Engineering and Interfacial Engineering and Material Science, Fraunhofer IGB, Stuttgart, Germany Keywords: tracheal stents; surface modification; biocompatibility Background: Airway stenting is increasingly used in the management of obstructive lesions of the central airway. Microbiological colonization of the airway stents causes therapy resistant halitosis and severe airway infections. We coated clinically applied tracheal stents with a thin layer of silver nano-particles to prevent bacterial colonization in vitro. Methods: Tecoflex® Polyurethane sheets (n=5) were coated with silver nano-particles by immersion in a colloidal silver solution, generated from silver nitrate by reduction with sodium citrate in distilled water, for 2 hours and exposed to a hydrogen plasma glow discharge for silver colloid fixation. To test for bacterial colonization, stent materials were incubated for 1–3 days in vitro with i) Micrococcus luteus (ML), ii) E. coli (EC) and iii) Pseudomonas aeruginosa (PA). Antimicrobial activity was evaluated by bacterial growth inhibition. Stent surface colonization was investigated by raster electron microscopy. Results: The covering procedure resulted in an 80 nm silver nanocoating. It shows antibacterial activity against gram-positive (ML) and gram-negative (EC) bacterial strains, especially PA, by inhibiting bacterial stent surface seeding. Additionally, the coating exerts its antibacterial activity on the stent environment, suppressing colony formation. Conclusions: Nano-silver coating using plasma technology may increase the biocompatibility of tracheal stents in clinical applications, thereby eliminating the cause for halitosis and airway infections.

763

12 IN VIVO GAIT ANALYSIS IN MURINE FEMORAL FRACTURE MODEL Tina Histing, Alexander Kristen, Christian Roth, Jörg Holstein, Patric Garcia, Michael Menger, Tim Pohlemann Department of Trauma, Hand and Reconstructive Surgery, Institute for Clinical & Experimental Surgery, University of Saarland, Homburg/ Saar, Germany, Collaborative Research Center, AO Foundation, Switzerland, Departemt of Diagnostic and Interventional Neuroradiology, University of Saarland, Homburg/ Saar, Germany Keywords: gait analysis, mouse, femur fracture, fracture stabilization Background: Gait analysis is used as a powerful technique for evaluating locomotion such as motion and load patterns in humans and large laboratory animals. In small animal models the gait analysis is only limited useful. A murine fracture model has not been described yet. Therefore the aim of the study was to define and describe parameters of gait analysis in murine fracture model Methods: Ten CD-1 mice (32+/−3 g) were divided into two groups. All mice underwent prior training to become accustomed to walking in the running wheel. In 5 mice, a standardized closed midshaft fracture was produced using a 3-point bending device. For fracture stabilization a common pin (10 mm) was used. 5 mice served as control group, one femoral pin (10 mm) was inserted without producing a fracture as femoral marker. Both groups were additionally marked by one tibial pin (5 mm). An x-ray permeable running wheel was used for gait analysis. Multiple running cycles were digitized with 30 images/ s by a digital angiography x-ray system to identify the markers. Fourteen days after surgery the following gait parameters were determined: the minimum and maximum tibio-femoral angle, the stride frequency, the stride time, the stride length and the stride velocity. Eighteen representative strides per mouse were analyzed. All measurements were done using the Java based open source program ImageJ. Results: The control group showed a significantly higher maximum (96.6+/−1.2° vs. 74.7+/−1.2°) and minimum (41.6+/−1.4° vs. 51.4+/−1.2°) tibio-femoral angle compared to the fracture group. Measuring the stride frequency and stride time both groups showed comparable results without significant differences (stride frequency: fracture group: 4.8+/−0.3 vs. control group: 4.6+/−0.3 strides/ s; stride time: control group: 0.3+/−0.1 vs. fracture group: 0.2+/−0.01 s). In the control group stride length was significantly larger (9.2+/−0.2 cm) than in the fracture group (5 +/− 0.1 cm). By consequence, a significantly higher stride velocity was observed in the control group (42.7+/−5.3 cm/ s vs. 23.6+/−0.4 cm/ s). Conclusions: In the present study we describe a novel approach to quantify the fixation techniques and the fracture healing process using gait analysis. This study was conducted to measure different characteristic parameters to investigate the stability served by different methods of fracture stabilization. Further the herein presented results will help to design future studies with standardized mechanical conditions for analyzing mechanisms of fracture healing.

13 BIOMECHANICAL PROPERTIES OF FASCIA LATA GRAFTS Lars Hinrich Evers, Dhaval Bhavsar, Richard Bodor, Peter Mailänder Dept. of Plastic Surgery, University of Luebeck, Germany, Division of Plastic Surgery, University of California, San Diego, USA Keywords: Grafts, Flaps, Biomechanical Stability Background: Fascia lata grafts are commonly employed as support structures in plastic and reconstructive surgery. Despite this widespread application, little objective testing has been published

764 regarding the biomechanical features of this graft particularly as related to the clinical relevance of directional orientation. Methods: Cadaveric study was approved by institutional review board. 16 sheets of fascia lata (2.5 cm×0.8 cm) were surgically and meticulously dissected, either in horizontal or vertical orientation, from fresh adult post-mortem specimens. The end of each sheet was affixed to the ZP-44 force gauge (IMADA Northbrook, IL) for measurement of bursting strength. Group A (n=8) were tested with fascia oriented horizontally, Group B (n=8) were tested with fascia oriented vertically. Force gauge was set for peak values and bursting strength as well as course of the biomechanical curve were recorded. Difference between the groups was analyzed with ANOVA test. Results: The mean peak bursting strength of the fascia lata graft oriented in the horizontal direction was 11.5 N (SD 5.8) and in vertical direction 76.8 N (SD 17.1). The difference between bursting strengths for horizontal and vertical orientations is statistically significant (p< 0.01). Conclusions: Fascial fibre orientation is responsible for significant differences in strength and should be considered when employing these grafts for biomechanical and structural support applications. Fascial fibre orientation is likely derived from functional demands and graft harvest should reflect a balance of recipient site needs and donor site morbidity. Fascia lata in horizontal orientation does not provide adequate strength (11.5 N) when compared to native rectus fascia (peak bursting strength 59 N, Choe et al). Our data could form a base for further refinement in orienting the clinical applications of fascia lata for abdominal wall and other reconstructions.

14 EFFICACY OF VARIOUS DURATIONS OF IN-VITRO PREDEGENERATION ON THE CELL COUNT AND PURITY OF RAT SCHWANN CELL CULTURES Armin Kraus (1), Joachim Täger (2), Thomas Skutella (2), Sabine Conrad (2), Konrad Kohler (3), Hans-Eberhard Schaller (1), Nektarios Sinis (1) 1: BG-Unfallklinik, Klinik für Plastische, Hand-, Rekonstruktive und Verbrennungschirurgie an der Eberhard-Karls-Universität Tübingen, 2: Institut für Experimentelle Embryologie, Tissue Engineering und klinische Anatomie, Anatomisches Institut der Universität Tübingen; 3: Zentrum für Regenerationsbiologie und Regenerative Medzin, Universitätsklinikum Tübingen, Germany Keywords: Schwann cells, methods, nerve regeneration,peripheral nerve reconstruction, nerve injury Background: For reconstruction of peripheral nerve defects, tubulization with artificial grafts coated with cultivated Schwann cells is a promising approach. By in vitro-predegeneration of the harvested nerve, efficacy of cell cultivation can be enhanced. It is aim of this study to improve Schwann cell culture efficacy by comparison of three different durations of predegeneration. Methods: The ischiadic nerves of 6–8 week old Lewis rats were harvested and subjected to either 2-day, 7-day or 14-day predgeneration in DMEM, 10% FCS and 1%Pen/Strep. Afterwards, tissue was dissociated in 0,125% collagenase IV and 0,25% trypsine for 24 hours and given into modified melanocyte-growth medium. Fibroblast growth factor was added at 10 ng/mL, in addition to 2 µM forskoline and 5 µg/ml bovine pituitary extract. Cell count was determined immediately after dissociation, cell purity was determined 16 hours later after cell attachment to the culture plate by means of optical microscopy. Special value was set on Schwann cell to fibroblast relation hereby. Results: Cumulative cell count in culture was 5.8×10^4 for 2 days, 1.12×10^5 for 7 days and 1.48×10^5 for 14 days of predegeneration. Culture purity was approximately equal for 2 day and 7 days of

Langenbecks Arch Surg (2008) 393:759–815 predegeneration (88% Schwann cells, 12% fibroblasts after 2 days, 85% Schwann cells, 15% fibroblasts after 7 days). After 14 days of predegeneration however, cell cultures were significantly higher debased by fibroblast proliferation (57% Schwann cells, 43% fibroblasts) Conclusions: The yield of cultivated rat Schwann-cells is duplicated by 7-day in vitro predgeneration in comparison to 2-day predegeneration. After 14-day predegeneration, culture debasement with fibroblasts is significantly higher. Therefore, 7-day in vitro predgeneration of the harvested nerve tissue is an advisable procedural method in the cultivation of rat Schwann cells.

15 3-D SURFACE IMAGING FOR OBJECTIVE QUANTIFICATION AND PREDICTION OF BREAST REDUCTION MAMMAPLASTY? Maximilian Eder, Nikolaos A. Papadopulos, Hans-Günther Machens, Edgar Biemer, Laszlo Kovacs Department of Plastic Surgery and Hand Surgery, Klinikum rechts der Isar, Technische Universität München, Germany Keywords: breast, breast volume, breast reduction, 3-D scan, 3-D surface imaging Background: Especially in reduction mammaplasty the prediction of resection weight could be very helpful to obtain symmetrical breasts, but presently volume estimations are largely based on the surgeons experience and the currently used methods are unreliable and cumbersome. Three-dimensional (3-D) surface imaging proved to be highly accurate and precise for the quantification of the female breast region. In this study we analysed the postoperative changes regarding volume, surface, shape and symmetry in three-dimensions and tried to develop a more clinically reliable method of predicting resection weight and breast shape changes. Methods: 15 patients (n=30) obtaining a superior pedicled breast reduction (Höhler technique) and 15 patients (n=30) obtaining a vertical scar reduction (Lejour technique) were 3-D assessed regarding volume [cc], surface [cm2], linear distance measurements [cm] preand postoperatively and the results were analysed regarding existing breast asymmetry and compared with the breast resection weight [g]. The changes between the pre- and postoperative measurements are analysed by correlation and regression equations to gain calculation formulas to predict volume, surface and linear distance changes preoperatively for each technique. Results: There were no significant differences between the left and the right breast in both reduction groups ragarding volume, surface and linear distances pre- and postoperatively, expressing existing breast symmetry. In both groups the calculated 3-D volume (surface) difference highly correlated with the resection weight (Höhler: r= 0.993 (r=0.963), Lejour: r=0.978 (r=0.887), for all p<0.001) and expresses high accuracy of the volume (surface) measurement. The most important regression equations revealed the following relationship for the preoperative prediction: pre OP Volume = sternal notch to nipple distance − 26.626/0.003 (Höhler) and sternal notch to nipple distance − 19.020/0.01 (Lejour). Resection volume ¼ 156:931þ 0:325
pre OP volumeðHö hlerÞ a n d 204 þ 0:506
pre OP volume ðLejourÞ. Surfacechange ¼ 112:679 þ 0:81
resection
volumeðHöhlerÞ and 16:283 þ 0:719
resection volume Lejour . Conclusions: We presented a new approach to predict breast shape changes and resections weights for different techniques in reduction mammaplasty more objectively using 3-D surface imaging. Due to the small sample size we consider our findings as preliminary requiring larger studies to validate the method before introducing it as a routine tool in clinical evaluation,

Langenbecks Arch Surg (2008) 393:759–815 although we belive in future promising applications of the 3-D technology in Plastic Surgery.

16 POSTOPERATIVE EVALUATION OF SOFT-TISSUE SWELLING FOLLOWING BREAST SURGERY IN THREE DIMENSIONS Maximilian Eder, Nikolaos A. Papadopulos, Hans-Günther Machens, Edgar Biemer, Laszlo Kovacs Department of Plastic Surgery and Hand Surgery, Klinikum rechts der Isar, Technische Universität München, Germany Keywords: breast, breast volume, swelling, 3-D scan, 3-D surface imaging, objective quantification Background: Three-Dimensional (3-D) surface imaging offers the ability to quantitatively evaluate the breast region regarding shape, surface and volume after surgery sufficiently precise and accurate. Soft-tissue swelling is usually observed after surgery and current information has been gained from two-dimensional data studies only and no study described chronologically the breast changes following surgery. Clinicians may not have been able to estimate the posoperatively soft-tissue changes accurately and to satisfactorily answer the anxious patients request regarding the treatment effects. This study aims to evaluate for the first time in a prospective clinical trial the change of soft-tissue swelling in three-dimensions following breast surgery over time. Methods: 10 breast augmentation and 10 breast reduction patients were recruited for the study and 3-D surface imaging of the breast region was accomplished using a linear laser scanner. Laser scans were obtained using a standardized three-dimensional scanning protocol over six time periods (pre OP and posoperatively: post OP 1–2–3 days, post OP 2–1 week, post OP 3–1 month, post OP 4– 3 month and post OP 5–6 months). The 3-D surface scans were analysed at each time period and for each patient regarding braest volume changes [cc] and breast surface changes [cm2]. The changes of these parameters are expressed as soft-tissue swelling in percentage. Results: The breast volume and braest surface changes highly correlated (r=0.963, p<0.001). Improvements in soft-tissue swelling were recorded for breast augmentation (breast reduction) as 18.38% (15.22%) after 1 week, 52.75% (55.48%) after 1 month, 83.31% (85.61%) after 3 months and 89.20% (91.62%) after 6 months compared to post OP 1. Conclusions: For the first time a new method for quantitave evaluation of postoperative soft-tissue swelling after breast surgery was described. A significant reduction in soft-tissue swelling is reached after one month and furthermore the results suggest that after 6 months the soft-tissue swelling decreased to approximately 90%. 3D surface imaging offers new clinical applications regarding postoperative evaluation, qualtiy assurance and further clical trials should be designed to clarify existing postoperative process controls and judgement of the order of precedence of competitive operational methods.

17 INNOVATION IN URODYNAMICS CYSTOMETRY WITH GAS PERFUSED CATHETERS WITH HELIUM R. Mund, R. Finke Department of Pediatric Surgery, Martin-Luther-University HalleWittenberg, Germany Keywords: urodynamics Background: Proof of the fitness of the method Gaserfusion with helium for urodynamical studies on the basis of simultaneous comparative cystometries between Waterperfusion and Gasperfusion.

765 Introduce the gasperfusion to the cystometry with the child. Development of a combination measuring instrument for urodynamic and manometry of anorectum and esophagus. Methods: The pressure changing was measured with Gasperfusion in an artificial bladder in different positions and under movements. During 62 urodynamic investigations of 33 young patients between 5 month and 25 years 562 active (coughing) or passive (manual pressure at the bladder) provocation tests were done. With the simultaneous measurement of the amplitudes of pressure changing in time in the bladder with two intraurethraly placed double channel catheters (6French), we get the velocity of pressure increase and could compare the common Waterperfusion with the new Gasperfusion using helium perfused catheters. Results: In the artificial bladder the pressure was not changing neither due to movements nor the different positions of the bladder. The Investigation of the children shows the statistic significant higher velocities of pressure changes of the Gasperfusion than those measured with Waterperfusion. Conclusions: The Gasperfusion has a lot of methodical advantages for urodynamic investigations especially in the use in children. The child can be measured in all possible body positions and during changing the body position. Helium is not dangerous for the human organism. The study with children could show the better measuring results of the Gasperfusion epsacially in fast pressure changes like provocation and stress. We can conclude that the Gasperfusion with helium is an excellent alternative to the common techniques using Waterperfusion and recommend the use of Gasperfusion in urodynamic studies espacially for children. Following these results it could be developed a new combination measuring instrument for urodynamics an manometry of the gastrointestinal tract with the same technical equipment.

18 SACRAL NERVE STIMULATION (SNS): CT GUIDED ELECTRODE PLACEMENT FOR PERCUTANEOUS NERVE EVALUATION (PNE) AND TEMPORARY SACRAL NERVE STIMULATION (TSNS) IN FECAL INCONTINENCE (FI) TREATMENT Matthias Goos, Günther Ruf Abteilung Allg./Viszeralchirurgie Universität Freiburg; Chirurgische Klinik Keywords: Fecal incontinence; sacral nerve stimulation Background: Sacral Nerve Stimulation (SNS) is used in the treatment of fecal incontinence caused by non-morphological sphincter deficiency provided that an intact “wiring” between sacral plexus and pelvic floor is given (Leroi 2005). Methods: PNE has been performed on 14 patients under local anaesthesia in an outpatient setting. The Assessment prior to the PNE procedure comprised a complete clinical history (including Cleveland Clinical Incontinence Score (CCIS)), physical examination, anorectal physiological evaluation (including anorectal manometry, state of the reflex arcs feeding the pelvic floor). Sphincteric lesions were excluded with anal ultrasound. The procedure was performed in the controlling area of a 64 slice CT-scanner (Somatom Sensation 64 ®; Siemens, Germany) under sterile conditions. Repeated CT-scans have guided the needle and electrode placement procedure (see pict. 1–2). Each time CT-scans were performed prior to PNE to evaluate the feasibility and at the end of the PNE procedure to exclude perirectal haematoma. Results: PNE was performed 8 times on the left and 6 times on the right site. The testing electrode was placed 10 times in the sacral foramen 3 (6×3R/4×3 L) and 3 times in the sacral foramen 4 (3×4 L). On average the procedure lasts 37 minutes (range: 21–52 min.). In one case CT-scan showing a an anatomical variant with angled course and

766 small calibre of the sacral foramen, in this case a PNE was not possible (see picture 4). In 2 cases the PNE wasn’t successful; a testing electrode was not implanted. In 3 cases the motor response to stimulation during PNE was weak, but yet a testing electrode was implanted: in 2 cases the patients experienced a significant improvement at the end of tSNS; hence an indication für pSNS was given. In none case there was a major surgical complication. Conclusions: CT-guided PNE under local anaesthesia in an outpatient setting enables the physician to perform a quick, reliable and safe PNE and electrode placement for temporary sacral nerve stimulation (tSNS) and also supports a more reliable decision whether or not a permanent sacral nerve stimulation (pSNS) should be performed. The “guidance” of a CT scanner delivers valuable information on the high, width and course of the sacral foramina as well as the peripheral course of the sacral nerve roots in the sacral plexus.

19 OPTIMISING POST-CONDITIONING TIME OF MARGINAL LIVERS Steffen Manekeller, Alexandra Seinsche, Judith Stegemann, Andreas Hirner Departement of surgery, Surgical research division, Rheinische Friedrich-Wilhelms-Universität, Bonn, Germany Keywords: Cold storage, Machine perfusion, Post-conditioning Background: Due to the discrepancy between organ donors and receptors the use of marginal livers (steatosis livers or organs from non-heart-beating donors) for transplantation purpose increased. The potential of a short-term aerobic machine perfusion for”less then optimal” grafts after cold storage was recently demonstrated. In this context it was shown that this post-conditioning mainly depends on the provision of oxygen, but was independent on the allowance of nutrients. In our study the optimal time course of post-conditioning is to be evaluated. Methods: Livers from male Wistar rats were withdrawn 30 min after cardiac arrest and flushed with 60 ml of HTK preservation solution (via portal vein). The organs were then stored at 4°C for 18 h under ischemic conditions (CS). After 16 h a part of the livers were then transferred to an aerobic machine perfusion circuit for 0,5 h, 1 h, 2 h or 3 h. Afterwards the viability of the organs was estimated by an acellular, aerobic normothermic reperfusion (2 h) in vitro. The vascular resistance (Pa/s/ml), the enzyme release into the perfusat (U/l), bile production (µl/g/h), the O2- consumption (µl/g/min), the ammonium-clearance (µmol urea/gxh), the ATP content (%/CS) and expression of apoptotic factors in the tissues (TUNEL) were evaluated. Results: After 1 h of post-conditioning a significant increase in bile production and a decrease in enzyme release could be detected in comparison to CS. Also for the vascular resistance, the oxygen consumption and the urea clearance a positive tendency was noted starting with 1 h of PK. The ATP content of the PK livers after 1 h of treatment was 60% higher than in CS organs. No markers for apoptosis could be detected after 1 h of PK. Conclusions: It can be concluded that a post-conditioning time of 1 h after cold storage can ameliorate the organviability of marginal livers. The extension or abbreviation of PK time seems to have no further beneficial effects under these experimental conditions. After 1 h of PK the ATP content reaches his maximum, afterwards the values decrease. Also the apoptotic induction is triggered on PK times over 1 hour. It seems possible that early alterations of the sinusoidal endothelial cell and hepatic sinusoid associated with liver injury are mediated by free radical species, such as superoxide anion and nitric oxide. Further experiments trying to scavenge of free radicals (SOD or NAcethylcystein) will be done to investigate this hypothesis.

Langenbecks Arch Surg (2008) 393:759–815

20 HEMODYNAMIC AND METABOLIC EFFECTS OF H2S DURING PORCINE ISCHEMIA/REPERFUSION INJURY Florian Simon 1,2, R. Giudici 2, C.D. Nguyen 2, S. Öter 2, M. Gröger 2, U. Wachter 2, J. Vogt 2, G. Speit 3, C. Szabó 3, P. Radermacher 2, E. Calzia 2, H. Schelzig 1 Thorax- und Gefäßchirurgie, Anästhesiologische Pathophysiologie und Verfahrensentwicklung, Humangenetik, Universitätsklinikum, 89070 Ulm, Germany; Ikaria, Seattle WA 98102, USA Keywords: H2S, suspended animation, hypoxia, aortic occlusion, metabolism Background: In awake mice inhaled H2S produced a “suspended animation-like” metabolic status with hypothermia and reduced O2 demand, thus protecting from lethal hypoxia [1]. Murine models may be questioned, however, since due to their large surface area/mass ratio rodents can rapidly drop their core temperature. Therefore, we investigated whether intravenous H2S (sulphide) would induce a comparable metabolic response in anesthetised and mechanically ventilated pigs. Since H2S was also reported to improve post-ischemic heart function, we also investigated whether sulfide would influence the norepinephrine (NE) responsiveness during reperfusion after aortic occlusion. Methods: After 2 hours of i.v. sulphide or vehicle, animals underwent 30 minutes of aortic occlusion with nitroglycerine, esmolol and ATP adjusted to maintain mean arterial pressure (MAP) at 80–120% of baseline. During reperfusion NE was titrated to keep MAP≥80% of this level. Results: sulfide reduced heart rate and cardiac output without affecting stroke volume, markedly the NE needs required to maintain hemodynamic targets, and caused a drop in core temperature concomitant with lower VO2 and VCO2. Sulfide attenuated the reperfusion-related hyperlactatemia, while glycemia was higher at the end of the experiment. Conclusions: Intravenous sulfide allowed reducing energy expenditure in an anesthetised large animal model, improved the noradrenaline responsiveness during reperfusion after aortic occlusion, and thus may be organ-protective after ischemia/reperfusion injury. 1. Szabó C: Hydrogen sulfide and its therapeutic potential. Nat Rev Drug Disc 2007;6:917–35

21 INTRACAPILARY LEUKOCYTE ACCUMULATION AS A NOVEL ANTIHEMORRHAGIC MECHANISM IN ACUTE PANCREATITIS IN MICE Eduard Ryschich, Vachtang Kerkadze, Olegas Deduchovas, Audrius Parseliunas, Angela Märten, Werner Hartwig, Markus Sperandio, Jan Schmidt Chirurgische Klinik Universität Heidelberg; Walter-Brendel-Center für Experimentelle Medizin, Ludwig-Maximilians-Universität, München, Germany Keywords: leukocytes, acute pancreatitis, hemorrhage, transgenic mice, intravital microscopy Background: Pancreatic infiltration by leukocytes represents a hallmarks in acute pancreatitis. Although leukocytes play an active role in the pathophysiology of this disease, the relation between leukocyte activation, microvascular injury and hemorrhage has not been adequately addressed. Methods: We investigated intrapancreatic leukocyte migration, leukocyte extravasation and pancreatic microperfusion during acute pancreatitis in lys-EGFP-ki mice using fluorescent imaging and time-lapse intravital microscopy. Results: In contrast to the current paradigm of leukocyte recruitment, the initial event of leukocyte activation in acute pancreatitis was

Langenbecks Arch Surg (2008) 393:759–815 represented through a dose- and time-dependent occlusion of pancreatic capillaries by intraluminally migrating leukocytes. Intracapillary leukocyte accumulation (ILA) resulted in dense filling of almost all capillaries close to the area of inflammation and preceded transvenular leukocyte extravasation. ILA was also initiated by isolated exposure of the pancreas to IL-8, PMA or fMLP demonstrating the causal role of chemotactic stimuli in the induction of ILA. The onset of intracapillary leukocyte accumulation was strongly inhibited in LFA-1−/− and ICAM-1−/− mice, but not in Mac-1−/− mice. Moreover, prevention of intracapillary leukocyte accumulation led to the development of massive capillary hemorrhages and transformed mild pancreatitis into lethal hemorrhagic disease. Conclusions: ILA represents a novel protective and potentially lifesaving mechanism of hemostasis in acute pancreatitis. This process depends on expression of LFA-1 and ICAM-1 and precedes the classical steps of the leukocyte recruitment cascade.

22 GENETIC DEFICIENCY OF THE HIF PROLYL HYDROXYLASE 1 (PHD1) PROTECTS MURINE LIVERS AGAINST HYPOXIC DAMAGE Martin Schneider, Katie van Geyte, Judit Kiss, Martin Mollenhauer, Massimiliano Mazzone, Julian Aragones, Peter Fraisl, Patrick Maxwell, Jürgen Weitz, Peter Ratcliffe, Peter Carmeliet (1) Department of General, Visceral and Transplantation Surgery, University of Heidelberg, Heidelberg, Germany; (2) Vesalius Research Center (VRC), VIB, Leuven, B-3000, Belgium; (3) Renal Section, Hammersmith Campus, Imperial College London, London, UK; (4) The Henry Wellcome Building of Molecular Physiology, Oxford, UK. Keywords: hypoxia, liver ischemia, HIF-prolyl hydroxylase Background: Liver ischemia/reperfusion (I/R) injury is a frequent cause of organ dysfunction associated with liver surgery or transplantation. We recently documented that loss the HIF prolyl hydroxylase 1 (PHD1) caused hypoxia tolerance of skeletal muscle via reprogramming of cellular energy metabolism. Here, we aimed to assess whether loss or short-term silencing of PHD1 likewise induces hypoxia tolerance of mouse hepatocytes in vivo, providing protection against hepatic I/R damage. Methods: Hepatocyte damage following ischemia or I/R was investigated in PHD1-deficient (PHD1−/−) and wild type (WT) mice, or following short hairpin interference RNA (shRNA)-mediated short-term inhibition of PHD1 in vivo. Metabolic studies involved expression analysis of pyruvate dehydrogenase kinase isoenzymes (PDK1-4) via real-time RTPCR and protein blotting, and determination of hepatic pyruvate dehydrogenase complex (PDC) activity. Relevant in vivo findings were confirmed by in vitro assays applying siRNA-mediated knock down of PHD1 in murine AML12 liver cells. Oxygen consumption rates in cultured AML12 cells were assessed by electron paramagnetic resonance- (EPR-) oxymetry. Results: Livers from PHD1-/- mice displayed reduced hepatocyte swelling and intracellular hypoxia during acute ischemia. Necrosis, invasion by inflammatory cells and generation of toxic reactive oxygen species (ROS) were strongly attenuated in PHD1-/- livers after reperfusion of the ischemic organ. Importantly, similar protective effects could be induced by shRNA-mediated short-term inhibition of PHD1 in vivo. In accordance with these findings, inhibition of PHD1 likewise induced hypoxia tolerance of cultured hepatocytes in vitro. Mechanistically, the hypoxia tolerance of PHD1-deficient hepatocytes was attributed to an adaptation of hepatocellular energy metabolism, involving an increased expression of PDK1, that, in turn, lowered the activity of PDC and, hence, oxygen consumption. Conclusions: Loss of PHD1 increases the tolerance of hepatocytes against acute hypoxia, and protects against hepatic I/R-damage. Short-

767 term inhibition of PHD1 may offer a novel treatment perspective to minimize I/R-induced liver injury.

23 SIMVASTATIN DEPENDENT EXPRESSION OF HEME OXYGENASE-1 PREVENTS LIVER INJURY AFTER HEMORRHAGE/RESUSCITATION IN RATS Borna Relja, Mark Lehnert, Korbinian Seyboth, Fabian Bormann, Christoph Czerny, Dirk Henrich, Ingo Marzi Department of Trauma-, Hand- and Reconstructive Surgery, Hospital of the Johann Wolfgang Goethe-University, 60590 Frankfurt/Main, Germany Keywords: Simvastatin, HO-1, hemorrhage, liver Background: Hemorrhagic shock and resuscitation (H/R) is associated with free radical production and proinflammatory changes leading to hepatocellular cell death. Simvastatin inhibits cholesterol biosynthesis by blocking HMG-CoA reductase. Its pleiotropic effects include modulation of endothelial NO synthase activity as well as changes in inflammatory responses. Statins are known to inhibit Rho-kinase (ROCK). ROCK activity increases under hypoxic conditions and is normalized with Simvastatin treatment. Previous studies pointed at a potential effect on HO-1 expression after ischemia-reperfusion, however the effect of Simvastatin after H/R is still unknown. Hence, we evaluated the potential effect of Simvastatin treatment and its mechanisms on hepatocellular damage induced by H/R. Methods: Before H/R, female Lewis rats were pretreated for 6 days with Simvastatin (5 mg/kg per day, intraperitoneally, i. p.) or vehicle. Then, rats were hemorrhaged to a mean arterial blood pressure of 30±2 mmHg for 60 min and resuscitated. The control group underwent surgical procedures without the onset of H/R. Two hours later, hepatic cell death, oxidative (4-hydroxynonenal, 4-HNE) and nitrosative (3-nitrotyrosine, 3-NT) stress, markers of inflammation (serum IL-6 levels and hepatic infiltration with polymorphonuclear leukocytes (PMNL)) as well as actin cytoskeleton rearrangements were determined. Hepatic protein levels of HO-1, eNOS and ROCK-I were evaluated by determination of integrated density of individual bands in Western blot; p<0.05 was considered significant (ANOVA, Fisher’s exact test, mean±S.E.M). Results: Serum activity of alanineaminotransferase (ALT) following H/R increased 2.2-fold in vehicle treated rats as compared to Simvastatin treated rats. Hepatic necrosis as well as apoptosis were significantly inhibited in Simvastatin treated rats (necrosis: 7±1, and apoptosis 5±1 cells per high power field, respectively) after H/R as compared to shock ctrl (necrosis: 69±8, and apoptosis: 26±3 cells per high power field, respectively; p<0.05). Simvastatin blunted neutrophil infiltration after H/ R by 50%. After H/R, Simvastatin treatment reduced serum IL-6 levels as compared to vehicle treated rats (217±73 vs. 1596±815 pg/ml; p< 0.05). Simvastatin treatment lowered hepatic 4-HNE staining after H/R by 25% and hepatic 3-NT staining by 27% as compared to vehicle treated rats (p<0.05). Actin cytoskeleton rearrangements after H/R were reduced by Simvastatin. Elevated hepatic ROCK-I protein levels were largely blunted after Simvastatin pre-treatment. By contrast, Simvastatin treatment induced an enhanced hepatic eNOS and HO-1 protein expression and augmented eNOS phosphorylation. The mortality rate after 72 h after H/R was attenuated by Simvastatin pre-treatment from 80% in vehicle treated rats to 20%. Conclusions: Simvastatin application before H/R blunts hepatic damage, hepatic oxidative and nitrosative stress, systemic and hepatic inflammation, cytoskeleton rearrangements as well as mortality rate. These changes are mediated, at least in part, by Simvastatin-induced HO1 and eNOS overexpression as well as ROCK-I expression modification. These results highlight specific mechanisms by which Simvastatin might be beneficial in patients undergoing H/R. Supported in part by DFG MA 1119/3-3

768

24 THE LEUKOCYTE-INHIBITION MODULE (LIM) A BENEFICIAL TOOL TO PREVENT NEUTROPHIL MEDIATED SIRS AFTER HEMORRHAGIC SHOCK? Tim Tobias Lögters, Stefan Margraf, Jens Altrichter, Martin Scholz, Joachim Windolf Klinik für Unfall- und Handchirurgie, Universitätsklinikum Düsseldorf, Germany Keywords: inflammation, neutrophils, organ failure, sepsis, SIRS, trauma Background: Neutrophil-mediated SIRS after hemorrhagic shock is a well known phenomenon being associated with mortality and multiple organ failure. The leukocyte-inhibition module (LIM) rapidly inactivates neutrophils through CD95 stimulation. In a study the efficacy of the LIM in preventing neutrophil-mediated pathophysiologic mechanisms was evaluated in an experimental porcine hemorrhagic shock model. Methods: Munich Mini Pigs (n=24; 30.3±3.3 kg) were rapidly hemorrhaged to reach a mean arterial pressure (MAP) of 35±5 mmHg, maintained hypotensive for 45 minutes, and then resuscitated with Ringer’s solution to baseline MAP. With beginning of resuscitation 12 pigs (LIM-group) underwent LIM therapy with extracorporeal circulation for 3 hours. Hemogram, hemodynamic variables, blood gas analyses, lactate levels, and organ specific enzymes were followed throughout the experiment. The animals were allowed to survive. They underwent necropsy 48 hours (n=6 per group) and 72 hours (n= 6 per group) after hemorrhagic shock. Results: Neutrophil counts were lower in the LIM group throughout the LIM therapy with statistical significant differences at 90 minutes after resuscitation (shock with LIM: 5.1±0.9×103/mm3; shock without LIM: 7.4±0.6×103/mm3). For lymphocyte counts no intergroup differences were found (shock with LIM: 0.9±0.07×103/mm3; shock without LIM: 1.0±0.08×103/mm3). Histological analysis showed reduced neutrophil infiltration for the LIM group in lung, liver, and heart. Furthermore, MAP (shock without LIM: 44.0±6.5 mmHg; shock with LIM: 51.0 ± 5.7 mmHg), cardiac output (shock without LIM: 2.1±0.2 l/min; shock with LIM: 3.1±0.6 l/min) and central venous oxygen saturation (shock without LIM: 48.2±9.1%; shock with LIM: 61.3±11.7%) were significantly higher in the LIM group. Conclusions: The leukocyte-inhibition module (LIM) specifically downregulates neutrophil counts in the bloodstream and reduces shockmediated neutrophil tissue infiltration. In addition LIM therapy is associated with improved hemodynamics after hemorrhagic shock. Thus, LIM may represent a beneficial tool for the prevention of immunepathogenic sequelae after hemorrhagic shock or other settings of unappreciated neutrophil activation. Further studies transferring the LIM therapy to humans are being planned.

25 PRESERVING SKIN MICROCIRCULATION WITH DOLPHINBED TECHNOLOGY Lars Hinrich Evers, Richard Bodor, Peter Mailänder Dept. of Plastic Surgery, University of Luebeck, Germany, Div. of Plastic Surgery, University of California, San Diego, USA Keywords: Microcirculation, Pressure Ulcer Background: Pressure ulcers remain a major concern for healthcare facilities, from both economical and quality-of-care points of view, but with proper patient care their occurrence can be reduced. Studies examining factors most likely associated with intra-operatively acquired pressure ulcers have not resulted in the development of reliable risk reduction tools adequately intervening in the surgical patient population. Current technology in pressure sores prevention consist in air flow mattresses. Regardless of previous improvements in these beds, the

Langenbecks Arch Surg (2008) 393:759–815 deflation during the surgery is still a significant drawback. Recently, a new technology has been introduced which uses three dimensional volumetric pressure redistribution. Our study aims to investigate the impact of this naval engineered bedding material on the effect of pressure on vulnerable cutaneous microcirculation. Methods: A total number of 10 healty volunteers were tested with the Microsolo 900 T surgical table surface (Biologics, Inc., Clearwater, Flo) and compared to a regular operating gurney. Objective measurement of the perfusion at pressure points of shoulder prominence were achieved with Laser Doppler Flowmetry (Periflux System 5000, Perimed, Sweden). With this technique the dynamic changes in the microcirculation were followed closely in real time. The percentage of the perfusion changes was calculated. Statistical analysis was performed using Student’s t-Test. Results: The mean reduction of perfusion from a warmed subject (maximum heat), which represents maximum vasodilation, to regular gurney was −90.52%, to the Dolphin bed only −22.31%.The mean change of perfusion from Dolphin bed to regular gurney was −88.71%. The differences are statistically significant (p<0.01). Conclusions: Our results indicate a significant improvement in microcirculation using the Dolphin Bed technology in comparison to regular gurneys. This results may lead to a prevention of pressure ulcer development. Further studies especially in high risk populations are warranted to investigate this new technology.

26 POXVIRUS DERIVED CYTOKINE RESPONSE MODIFIER A (CRMA) DOES NOT PROTECT AGAINST FOCAL CEREBRAL ISCHEMIA Mia Kim, Ertugrul Kilic, Herman van der Putten, Christian Wiessner, Bernd W. Böttiger, Dirk M. Hermann Department of General, Visceral and Thoracic Surgery, Klinikum Nürnberg, Department of Neurology, University Hospital Zürich, Nervous System Research, Novartis Pharma AG, Department of Anaesthesiology, University Hospital Cologne, Department of Neurology, University Hospital Essen Keywords: ischemic stroke, apoptotic cell death, caspase inhibitor Background: Pox-derived cytokine response modifier A (CrmA), a viral broad-spectrum caspase-inhibitor, is a neuroprotectivum inhibiting apoptosis on different levels. It has already shown its antiapoptotic potential in vivo and in vitro. Aim of this study was to show the neuroprotective potential of CrmA, neuronally expressed in a transgene mice model, measured by neuronal deficit-score, infarct volume, brain swelling, number of TUNEL-positive, surviving cells and of activatedcaspase-3-positive neurons after transient focal cerebral ischemia. Methods: Middle cerebral artery (MCA) occlusion was carried out according to National Institute of Health guidelines for the care and use of the laboratory animals with governmental approval. Expression of transgene CrmA in the mouse brain was analyzed with FISH and Re-FISH, PCR and western blot. CrmA-tg C57bl6 animals and their wildtype littermates were anesthetized with 1% halothan (30% O2, 70% N20). Cerebral blood flow was measured by laser doppler flowmetry and occlusion of the MCA was performed by intraluminal occlusion model. Therefore a silicone-coated filament was placed on the outlet of the right MCA 90 minutes or 30 minutes (n=6–7 animals/group). After 24 hours of reperfusion, neurological deficit scores was evaluated using a 5-point deficit score ranging from 0= normal function to 4= absence of spontaneous motor activity. Animals were re-anesthetized and decapitated. Brains were removed and cut on a cryostat in 18 mm coronal sections for TUNEL-, cresyl violet staining and immunohistochemistry for activated caspase-3. Neurological deficit score, infarct volume and brain swelling were analyzed on the basis of those animals, undergoing 90 minutes of

Langenbecks Arch Surg (2008) 393:759–815 ischemia. Surviving neurons, TUNEL-staining and activated caspase3 were analyzed on the basis of those animals, undergoing 30 minutes ischemia. Coronal brain sections from five equidistant brain levels, 2 mm apart, were used. All animals were randomized and investigator blinded. Differences between groups were compared by two-tailed-ttest. P<0.05 were considered to indicate statistical significance. Results: FISH, Re-FISH, PCR and western blot confirmed the chromosomal integration and neuronal expression of CrmA. Laser doppler flowmetry showed the significantly reduced perfusion on the MCA territory. Neuronal deficit-score, infarct volume and brain swelling confirmed successful infarction of all mouses. But there were no differences between transgene and wildtype concerning neurological deficit score, infarct volume and brain swelling after 90 minutes of ischemia and there were no differences between the groups concerning TUNEL-positive activated caspase-3 and surviving neurons after 30 minutes of occlusion. Conclusions: These data show for the first time the influence of transgene, antiapoptotic, neuronal expression of CrmA after transient, focal cerebral ischemia. Inhibition by CrmA alone in this setting doesn’t seem to be sufficient to reduce the detected neuronal damage. A possible reason for these results can be the by-passing of caspases by activating alternative, caspase-independent pathways. Those pathways could be mediated by different mitochondrial proteins or the Calpain-Cathepsine cascade. The examination of those pathways and their influence on apoptotic celldeath are interesting for future investigations.

27 ERYTHROPOIETIN INHIBITS POSTISCHEMIC LEUKOCYTE ADHESION IN ALLOGENEICALLY TRANSPLANTED MOUSE HEARTS WITHOUT AFFECTING CORONARY MICROCIRCULATORY DYSFUNCTION Schramm R., Kirsch S., Schäfers H-J., Harder Y., Menger M. Dept. of Thoracic and Cardiovascular SurgeryUniversity Hospital Homburg/SaarUniversity of Saarland, Germany Keywords: Erythropoietin, microcirculatory dysfunction Background: This study was meant to analyze the effect of erythropoietin (Epo) on microcirculatory dysfunction and inflammation in murine cardiac allografts. Methods: Balb C mouse hearts were transplanted into C57BL/6 mice after 3-hour cold ischemia. Epo was given i.p. in recipients at 2 hours before reperfusion (n=6), while controls received saline only (n=6). The subepicardial microcirculation was assessed by intravital fluorescence microscopy (IVM) at 1, 3 and 6 hours of reperfusion. Results: In controls, subepicardial capillary blood flow velocities and functional capillary densities (FCD) decreased during reperfusion from 0.34±0.04 mm/s and 351±73 cm/cm2 to 0.30±0.01 mm/s and 239±41 cm/cm2, however not significantly. Capillary diameters and venular blood flow characteristics showed no significant changes over time, ranging between 4.5 and 5.5 µm as well as 0.76 and 0.96 mm/s. Epo-treatment had no effect on coronary microhemodynamics. Postischemic inflammation was characterized by augmented microvascular leakage ranging between 71 and 99% throughout the entire observation period. This was comparable between controls and Epo-treated mice. During reperfusion, control allografts showed decreasing numbers of rolling leukocytes and increasing numbers of firmly attached leukocytes from 64±16 cells/min and 238±84 cells/mm2 to 19±16 cells/min and 479±154 cells/mm2 (P>0.05). Capillary leukocyte plugging remained stationary over time in controls with 5.7±0.4 cells/HPF at 1 h and 5.0±0.5 cells/HPF at 6 h of reperfusion. Epo-treatment did not alter leukocyte rolling interactions. In contrast, firm leukocyte arrest in postcapillary venules was inhibited by Epo-treatment, resulting in 84±34 cells/mm2 at 6 h of reperfusion (P<0.05). Epo-treatment also reduced capillary leukocyte plugging to 3.6±0.3, 2.6±0.3 and 3.0±1.3 cells/HPF at 1, 3 and 6 h of reperfusion (P<0.05).

769 Conclusions: These are the first data on microcirculatory dysfunction and inflammation in murine cardiac allografts assessed by IVM. We demonstrate that non-hematopoietic treatment with Epo exerts antiinflammatory effects, reducing leukocyte-coronary endothelium adhesive interactions, without affecting microhemodynamics.

28 AMIODARONE PRETREATMENT OF ORGAN DONORS INDUCES LIVER PRESERVATION AND REPERFUSION INJURY Mohammed R. Moussavian, Otto Kollmar, Michael Schmidt, Jan E. Slotta, Michael D. Menger, Martin K. Schilling Department of General, Visceral, Vascular and Pediatric Surgery Keywords: amiodarone, liver transplantation Background: Based on the continuous organ shortage and increasing critical patients for liver transplantation, always more marginal donors and organs are allocated. Often donors manifest cardiopulmonary diseases with primary or secondary cardiac arrhythmia. Furthermore several study groups already reported about cardiac arrhythmia within reperfusion after liver transplantation. Usually these were treated with anti arrhythmic drugs. One of the most frequent used anti arrhythmic drug is amiodarone (AM), which find especially use in the emergency situation based on its wide indication position for atrial- and ventriculary cardiac arrhythmia. Moreover Lapenna D et al. demonstrated an antioxidative effect of AM. Considering the double function of AM as an antioxidant and as an anti arrhythmic drug, was of interest how far that compound could affect the organs as pretreatment of livers in an isolated reperfusion rat liver model. Methods: Livers were reperfused for 60 min through the portal vein after 2- or 24-hours storage, in a nonrecirculating model with freshly prepared Krebs-Henseleit bicarbonate buffer saturated with carbogen at a flow rate of 2 mL/min*g liver tissue using a pulsatile perfusion pump. Liver tissue was kept for histology and Western blot analysis of caspase-3 and heme-oxygenase 1 (HO-1). Livers were subjected to three groups (n=8 livers each): AM, I/R or Sham-treatment. AM was injected in a concentration of 5 mg/kg/body weight 3 minutes before liver flushing with HTK solution. Ischemia and reperfusion control (I/ R) livers were only perfused with HTK solution. Cold preservation of livers was performed for 24 hours in the AM and I/R group. Livers that were stored as described above but immediately reperfused with an ischemic period of less than 2 hour served as sham-operated controls. Results: The I/R group and the addition of AM demonstrated significantly higher K+ in first effluat compared to the sham group (11.62±0.81 and 12.87±1.41 vs. 8.02±0.47, p<0.05). The mitochondrial damage, determined by the GLDH activity, was significantly increased in first effluat within the AM compared to the I/R and Sham group (34.29±13.35U/l vs. 23.36±4.02U/l und 5.73±2.03U/l, p< 0.05). After 30 and 60 min GLDH were substantially lower in all groups whereby AM still induced significantly higher activities than the Sham group (2.24±0.47 vs. 0.16±0.05U/l, p<0.05). MDA concentration was significantly decreased by addition of AM measured directly after organ storage and 30 min compared to the I/R group (0.35±0.12 and 0.22±0.07 µmol/l vs. 0.81±0.2 and 0.75±0.08 µmol/l, p<0.05). Again the I/R group presented a significantly higher MDA concentration in the first effluat and after 30 min. Hepatocellular excretory function was found in situ and ex situ after 30 and 60 min maintained by supplementation of AM as indicated by a significantly reduced bile secretion compared to I/R and Sham group. Histological analyzes showed no significant differences between the groups. AM was effective to significantly inhibit cold ischemia-induced caspase activation as demonstrated by 68% reduction of cleaved caspase-3 in 24-hour stored organs. The expression of HO-1 was significantly

770 reduced after warm hepatic reperfusion by addition of AM with approximately 57%. Conclusions: Our data showed that the in vivo application of the potassium antagonist AM significantly decreased the in-situ and exsitu hepatic excretory capacity including mitochondrial damage. Its qualities as an antioxidants measured by reduced MDA during reperfusion could be approved. Due to domination of mitochondrial toxicity and reduction of the excretory liver function by AM, for this reason, pretreatment of marginal donors with AM has to be discussed controversially.

29 A META-ANALYSIS ON THE CLINICAL EFFECTIVENESS OF SELECTIVE GRANULOCYTE AND MONOCYTE ADSORPTIVE (GMA) APHERESIS WITH THE ADACOLUMN® device in ulcerative colitis Stefan Sauerland, Brigitte Habermalz, Edmund A.M. Neugebauer Institute for Research in Operative Medicine (IFOM), University of Witten/Herdecke Keywords: apheresis, ulcerative colitis, metaanalysis Background: Therapeutic apheresis is an extracorporeal blood purification method, by which undesirable cells or proteins are removed from the streaming blood. In several studies, selective granulocyte and monocyte adsorptive (GMA) apheresis was found effective in treating diseases which are associated with an excess of activated immune cells. The aim of this meta-analysis was to determine whether selective adsorptive granulocyte and monocyte apheresis (GMA apheresis) using the Adacolumn® device can effectively reduce clinical symptoms and endoscopic signs of inflammation in patients with ulcerative colitis. Methods: A comprehensive search for randomized controlled trials (RCTs) published up to May 2008 was performed. Studies were excluded if they used a filter for removal of granulocytes and lymphocytes (e.g. Cellsorba®). Each study’s quality was evaluated and results data were abstracted. One Japanese article was translated. Pooled relative risk (RR) estimates and 95% confidence intervals (CI) were calculated using the fixed-effects model. We also calculated the number-needed-to-treat (NNT). Heterogeneity was quantified statistically using I2 and explained by variation in patient sample, treatments, and trial design. Funnel plots were constructed to check for the presence of publication bias. Results: Seven RCTs including 594 patients were found, and six RCTs on active ulcerative colitis contributed to the main analyses. In half of the trials, GMA apheresis was compared to steroids. Only one trial was fully blinded. A response or remission after 6 weeks was achieved more often in patients treated with GMA apheresis (RR 1.42; CI 1.15 −1.75). Also, after 12 weeks, GMA apheresis produced significantly higher remission rates (RR 1.41; CI 1.08–1.83), but long-term data were sparse. From the risk difference (20.1%), the NNT was calculated as 5. In the trials that compared GMA apheresis versus steroids (n=220 patients), side effects were much less frequent in the GMA apheresis groups (RR 0.19; CI 0.11–0.34). No evidence for publication bias was found. Conclusions: Homogeneous evidence from several RCTs shows that GMA apheresis induces a clinical remission in a higher proportion of UC patients as compared to conventional medical therapy.

Langenbecks Arch Surg (2008) 393:759–815

30 MULTILAYER HYBRID CUFF ELECTRODE FOR REAL-TIME MONITORING IN THYROID SURGERY Christoph Ulmer*, Klaus Peter Koch***, Andreas Seimer**, KlausPeter Thon*, Wolfram Lamadé* * Abteilung für Allgemein-, Viszeral- und Unfallchirurgie, RobertBosch-Krankenhaus,** Abteilung für Phoneatrie und Pädaudiologie, Marienhospital, Stuttgart, Germany Keywords: thyroid surgery, cuff electrode Background: Continuous vagus nerve stimulation (VNS) via fully implantable stimulators is well-established in therapy of depression and seizure disorder. Aim of this study was to develop a real-time intraoperative nerve monitoring system for the recurrent laryngeal nerve (RLN) via vagus nerve. We present technical and clinical results of a new developed micro technological cuff-electrode for a real-time nerve monitoring system in thyroid surgery. Methods: On the basis of VNS probes, a new stimulator of the vagus nerve for a real-time nerve monitoring of the recurrent laryngeal nerve should be developed in cooperation with the Fraunhofer Institute, St. Ingbert. Up to now, all known VNS probes were tested only for afferent nerve fibers; we want to stimulate the efferent fibres. Implantation should be done through the OP-field during thyroidectomy, therefore the geometry of the stimulator should be as big as an ergonomic implantation is possible and no interference of the operation occurs. As sensing electrode to monitor compound muscle action potentials of the internal laryngeal muscles, we used a bipolar needle electrode, placed into the ipsilateral vocalis muscle via the cricrothyroid ligament. The new developed vagal nerve stimulator should be tested in a clinical trail for following properties: 1. Possibility of to stimulate efferent vagal nerve fibres and the RLN, 2. feasibility of e real-time Monitoring of the RLN, 3. Ergonomics of the stimulator and 4. safety of the stimulator. All patients received a laryngeal examination of the vocal cord before and after surgery. Results: A tripolar hybrid cuff electrode was developed. Three <1 µm thin golden film electrodes were fixed on a polyimide cuff and linked by platinum films. It resembles an open tube that firmly surrounds the nerve to avoid stray current, it is fully implantable during the operation and is highly flexible not to damage the nerve by extraction or incidental dislocation. As extension of the electrode three wires were connected to an EMG via Craggs connector. Stimulation current was 0.5–1.5 mA, bi-phasic, for 200 µsec with a frequency of 1 Hz by a multichannel EMG-system. 29 patients scheduled for a thyroid resection were enrolled into a non-randomized observational trail. In all patients a homogenous and reproducible signal could be sensed (mean 197 µV, range 45–401 µV). Mean supramaximal current was 1.0 mA. Maximal current necessary for supramaximal stimulation was 1.5 mA, minimal 0.5 mA. Duration of cuff electrode imlantation was medium 4,5 min, range 2–10 min. Three incidental dislocations without any harm to the nerve occurred. Duration of re-implantation was mean 5 min, range 2–8 min. In three cases we saw a reduction of the sensing signal of >60%, while the thyroid gland was pulled towards the trachea to expose the lateral part of the gland. Change in operative strategy led to immediate and spontaneous reconstitution of normal signal amplitude. In this series no permanent RLN-paralysis occurred. One patient with a postoperative hemorrhage from a strap muscle vein was revised without nerve monitoring. This patient experienced a temporarily partial impairment of the left vocal cord without complete loss of mobility. No patient suffered from prolonged hypoparathyroidism.

Langenbecks Arch Surg (2008) 393:759–815 Conclusions: With the presented multilayer hybrid cuff electrode a real time laryngeal nerve monitoring by stimulation of the vagal nerve is feasible. The electrode is completely atraumatic, fully implantable and easy to handle. Evoked potentials of the larynx are highly reproducible throughout the operation and can be handled for real time signal analysis with audio feed back. To avoid dislocation different sizes of the cuff could be developed.

31 A MDR1 (ABCB1) GENE SINGLE NUCLEOTIDE POLYMORPHISM PREDICTS OUTCOME OF TEMOZOLOMIDE TREATMENT IN GLIOBLASTOMA PATIENTS Markus Schaich, Lydia Kestel, Markus Pfirrmann, Katja Robel, Thomas Illmer, Michael Kramer, Claudia Dill, Gerhard Ehninger, Gabriele Schackert, Dietmar Krex Department of Medicine I, University Hospital C. G. Carus, Dresden, Germany. Department of Medical Informatics, Biometry and Epidemiology (IBE), Ludwig Maximilians University Munich, Germany. Department of Neurosurgery, University Hospital C. G. Carus, Dresden, Germany. Keywords: glioblastoma, MDR1, polymorphism, temozolomide Background: Some patients with glioblastoma multiforme do not respond to temozolomide even though they have aberrant promoter methylation of the DNA repair enzyme MGMT. This suggests that additional factors hamper temozolomide cytotoxicity. We aimed to confirm first that temozolomide is a target for the multidrug-resistance transporter MDR1/ABCB1, and second to investigate whether genetic variants of the MDR1 gene are associated with the survival of glioblastoma patients treated with temozolomide. Methods: Temozolomide mediated cytotoxicity was determined by the colorimetric methyl-thiazol-tetrazolium (MTT) assay in MDR-expressing and-non-expressing cell lines. Genotypes of three single nucleotide polymorphisms of the MDR1 gene (C1236T, G2677T, and C3435T), MDR1 mRNA expression levels and the MGMT promoter methylation status were analyzed in 112 glioblastoma patients who had been treated either by surgery plus radiotherapy alone or by additional temozolomide chemotherapy. Results: In vitro analysis revealed that temozolomide mediated cytotoxicity is dependent on MDR1 expression. Multivariate analysis of MDR1 genotypes showed that the C/C variant of the exon12 C1236T SNP is predictive for survival of patients treated with temozolomide. This effect was independent of the MGMT methylation status. Patients with the C/C genotype had a 2-year overall survival of 37% compared to 8% and 10% for patients with C/T and T/T genotypes, respectively (p=0.02). No influence was seen in the group of patients with radiotherapy only. Conclusions: The genotype of the MDR1 exon12 C1236T SNP is a novel independent predictive factor for outcome of temozolomide treatment in glioblastoma patients.

32 DIFFERENTIAL GENE-EXPRESSION IN RECURRENT GLIOBLASTOMA MULTIFORME Dietmar Krex, Katja Robel, Christian Pilarsky, Gabriele Schackert Klinik für Neurochirurgie, Klinik für Viszeral, Thorax- und Gefäßchirurgie, Universitätsklinik Carl Gustav Carus, Technische Universität Dresden, Germany

771 Keywords: glioblastoma, recurrencies, gene expression, GAS1 Background: Glioblastoma multiforme (GBM) is a morphological and molecular extremely heterogeneous tumor. Recent results indicate that aimed individualized therapeutically approaches are most promising to improve the still miserable prognosis of those patients. Therefore, identification of GBM subgroups is a prerequisite. Consequently, in our study we looked at patients with early versus late tumor recurrences in order to identify molecular markers, which might be associated with those different growth patterns. Methods: Study population comprised fifteen pairs each of primary and recurrent tumors occurring within 180 days after initial diagnosis (short-term recurrences, STR) and later than 300 days after initial diagnosis (long-term recurrences, LTR), respectively. The TP53 and PTEN genes were investigated by sequence analysis. EGFR amplification was determined. Loss of heterozygosity (LOH) analysis of markers known to be involved in glioma pathogenesis was performed. A chip-based expression analysis was performed in 6 samples (3 STR, and 3 LTR) using Affymetrix, GeneChip® Human genome U133 Plus 2.0. Expression of two candidate genes (GAS-1, SNTG) was confirmed by real-time PCR. Results: TP53 mutations are less common in STR than in LTR, while genetic variants of the PTEN gene are more frequent in STR. In STR, TP53 and PTEN variants occur in combination, while LTR harbour either TP53 or PTEN variants. Clustering after Chip analysis revealed clear differentiation between STR and LTR. Construction of a 6-gene comprising predictor is possible. Differential expression of GAS-1 is recorded between STR and LTR. Conclusions: STR and LTR can be distinguished by TP53 and PTEN mutation status and by gene expression profile. Out of a set of 53 differentially expressed genes, GAS-1 was confirmed to be differential expressed in STR and LTR, and might be used as predictive factor, taken that our results are confirmed in further series.

33 WORLDWIDE COLLECTION OF GENOTYPE AND PHENOTYPE DATA FOR HNPCC COLON CANCER AND A MODEL FOR THE HUMAN VARIOME PROJECT (HVP) Gabriela Möslein, Richard G.H. Cotton, Finlay A. Macrae St. Josefs Hospital Bochum-Linden, Genomic Disorders Research Center, Melbourne, Australia, Dept. of Colorectal Medicine and Genetics, The Royal Melbourne Hospital, Parkville, Australia Keywords: HNPCC, Human Variome Project, genotype phenotype correlation Background: InSiGHT (The International Society for Gastrointestinal Hereditary Tumours) has embraced ther HVP’s vision, and directed its attention to coordinationg the submission of variant information derived from a wide range of diagnostic research laboratories in Europe, Africa, USA, Canada, Japan and Australia, to its Locus Specific Database(LOVD) (www.insight-group.org). In Parallel, integration of other mismatch repair databases serving useful functions of collating the literature and other helpful annotations, is being unified onto a common LOVD platform to allow ready cross database interrogation. This is being supported by academic interest in defining search strategies used in clinical/Laboratory services, to locate published literature relating to the variants in question, and subsequently developing automated text searching strategies to populate the InSiGHT databases

772 Methods: Complementing this, dedicated InSiGHT committees are formulating databases to capture phenotype data across clinical, functional and histopathological domains, again linked to variant findings. The aim is to develop a common portal to allow searching of all available data relating to particular variants, informed by generic databases, e.g. ENSEMBL, to allow the best informed interpretation of unclassifies variant pathogenicity. To this end, an international, regionally represented, expert committee engaging a range of disciplines, has been assembled to InSiGHT to provide the best available interpretation of the variant data, with a process of dynamic updating. Results: Interaction with EBI, NCBI and GEN2PHEN received a strong foundation through a meeting of the HVP in May. Mutual benefits to each have been explored, and a RoadMap for this and the entire process described. Conclusions: This work involving a team of committed bioinformatics teams and information system expertise, interacting with clinicians, genetecists and privacy experts, is proving a model for extrapolation to other genes and LSDBs worldwide

34 EPIGENETIC INFLUENCE OF CD40/CD40L ENCODING MICRORNAS FOR INVASION AND METASTASIS OF THE DUCTAL ADENOCARCINOMA OF THE PANCREAS Soeren Torge Mees, Wolf Mardin, Christina Schleicher, Mario Colombo-Benkmann, Norbert Senninger, Joerg Haier Dept. of General Surgery, University Hospital of Muenster, Muenster, Germany Keywords: pancreatic carcinoma, miRNA, expression profiling Background: Pancreatic ductal adenocarcinoma (PDAC) is known for its very poor overall prognosis. Accurate early diagnosis and new therapeutic modalities are therefore urgently needed. While global microRNA (miRNA) expression patterns of many embryologic, physiologic, and oncogenic processes have been described, description of the role of miRNAs in ductal adenocarcinoma of the pancreas is lacking. We investigated the global microRNA expression patterns in pancreatic carcinomas to evaluate their involvement in invasion and metastasis. Methods: 16 cell lines of well- to undifferentiated pancreatic cancers from different tumor sites (primary or metastatic lesions) were used in a murine orthotopic model for pancreatic carcinomas. Using a standardized dissemination score, tumor biology (primary tumor volume, local infiltration, systemic metastasis) of these cell lines was assessed. We studied specific mi-croRNA expression regarding invasion and metastasis in all 16 cell lines by Taqman low density assays (quantitative real-time PCR) and microRNA-microarrays. To verify the accuracy of our data, we performed additional quantitative RT-PCR experiments for 15 relevant miRNAs and for CD40 mRNA. A computational analysis was performed to identify miRNA encoded genes and potential pathways. Results: According to their dissemination scores cell lines were classified into 3 groups with a hierarchical order of malignancy. Highly-metastatic cell lines presented a significant overexpression of CD40 encoding miR-224 (28.9fold) (p<0.0001) compared to non- or low-metastatic cell lines. Regarding invasion a significant induction (p<0.0001) of CD40 encoding miR-224 (11.4fold) and miR-486 (40.8fold) was detected for highly-invasive cell lines compared to non- or low-invasive cell lines. In contrast, CD40L (CD154) encoding miRNAs miR-19b (1.8fold), miR-532 (3.0fold) and miR-210 (5.9fold) were significantly decreased in highly-invasive cell lines (p<0.0001). Appositely to the miRNA results, a significantly decreased expression of CD40 mRNA was seen for highly-metastatic cell lines compared to non- or low-metastatic cell lines.

Langenbecks Arch Surg (2008) 393:759–815 Conclusions: In this study we have investigated the expression of miRNAs in PDAC cell lines and have observed an upregulation of CD40 encoding and a downregulation of CD40L (CD154) encoding miRNAs. Impaired anti-tumour immune responses have been described as associated with reduced expression of CD40L on T cells. Most recently, CD40L (CD154) related treatment has been shown to induce strong antitumoral effects toward different malignances. Therefore, epigenetic alterations of CD40/CD40L encoding miRNAs might be promising diagnostic tools or novel therapeutic targets for PDAC. (Project funded by Peter Geiger e.V., Beilstein, Germany)

35 MICRORNA METHYLATION OF MIR-107 LEADS TO INCREASED CELL PROLIFERATION IN HUMAN PANCREATIC CANCER CELLS Nils Habbe (1,2), Kwan Hyuck Lee (1), Oliver Kent (3), Joshua T. Mendell (3), Anirban Maitra (1) 1: The Sol Goldman Pancreatic Cancer Research Center, Department of Pathology, Johns Hopkins University School of Medicine, 2: Department of Surgery, Philipps-University Marburg, Marburg, Germany, 3: The McKusick-Nathans Institute of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore Keywords: miRNA, pancreatic cancer, methylation, epigenetics Background: Epigenetic changes are a common event in the development of pancreatic cancer (PC), and much effort has been taken to elucidate the mechanisms behind methylation dependent silencing of tumor suppressor genes, in order to identify novel markers for early detection or targets for treatment of this devastating disease. Recently, changes in the expression of microRNAs (miRNAs) have been described in pancreatic cancer, and these changes are subject to several studies investigating the influence of miRNAs in pancreatic cancer. Interestingly, several studies were able to prove that miRNA expression is by itself regulated by promotor methylation. Therefore, our aim was to identify miRNAs which are regulated by promotor methylation in human pancreatic cancer cells and to elucidate the functions of these miRNAs. Methods: MiRNA expression in long-established PC cell lines as well as in patient derived low passage cell lines was examined by realtime-PCR and Northern Blot before and after treatment with the demethylating agent 5-Aza-2′-deoxycytidine (5-Aza) and the histone deacetylase inhibitor trichostatin A (TSA). After identification of the miRNA promotor sequences and CpG islands within these promotors, we performed methylation specific PCR (MSP) using bisulfite modified genomic DNA of those cell lines. To identify the function of the reexpressed miRNAs we used lentiviral vectors carrying the miRNA sequence of interest to create stably transfected cell lines and examined the cell growth of these cell lines. Results: In the cell lines used in our study, miR-107 and miR-103 revealed methylation dependent regulation and showed up to 3.2 fold upregulation after treatment with 5-Aza or in combination with TSA, whereas TSA alone was not sufficient to reexpress those miRNAs. CpG island methylation of the distinct promotors could be identified by MSP. After reexpression of miR-107 by stable transfection using a lentiviral vector, the cell growth was significantly decreased when compared to mock transfected lines. Conclusions: Our study revealed for the first time that miRNAs are subject to methylation dependent regulation in human PC cell lines. These findings foster the possible use of demethylating agents in the treatment of PC. Furthermore, miRNA methylation patterns can be used as early markers in the detection of PC.

Langenbecks Arch Surg (2008) 393:759–815

36 EXPRESSION OF THE EMT REGULATOR SNAIL CORRELATES WITH METASTATIC POTENTIAL IN EXPERIMENTAL HUMAN PANCREATIC CANCER Hubert G. Hotz, Birgit Hotz, Elisabeth Schellhaas, Heinz J. Buhr Chirurgische Klinik I, Charité Univerversitätsmedizin Berlin, Campus Benjamin Franklin Keywords: EMT, Snail, transcription factor, pancreatic cancer Background: Epithelial to mesenchymal transitions (EMT) are vital for tumor progression and metastasis. Several inducers of EMT are transcription factors that repress E-Cadherin expression such as Snail, Slug and Twist. In this study we aimed to examine the expression and role of these transcription factors in pancreatic cancer. Methods: The expression of Snail, Slug, Twist, and E-Cadherin was detected by immunohistochemistry in tissue samples from 36 patients with pancreatic ductal adenocarcinoma. Four human pancreatic cancer cell lines (Capan-1, HPAF-2, MIAPaCa-2, and Panc-1) and human endothelial cells (HUVEC; control) were analyzed by RT-PCR, real-time PCR and western blotting. An orthotopic nude mouse model was applied for in vivo experiments: tumors were derived from the four human pancreatic cancer cell lines and animals (12 per group) observed for 14 weeks. Expression of the transcription factors and E-Cadherin was correlated with tumor spread and metastasis (dissemination score) in the mice. Results: 78% of human pancreatic cancer tissues showed an expression of Snail, and 50% of the patients displayed positive expression of Slug. Twist showed no or only weak expression. A strong Snail expression in undifferentiated cancer cell lines (MIAPaCa-2 and Panc-1) was associated with low E-Cadherin and extensive metastasis in the mouse model. In contrast, low Snail and strong E-Cadherin expression in more differentiated cells lines (Capan-1, HPAF-2) corresponded with a significantly reduced tumor spread in the animals (table). Snail mRNA expression correlated positively with metastatic potential of the cancer cells (r=0.8), whereas a negative correlation was found between ECadherin and metastasis (r=−0.9). Cell line Relative mRNA concentration Metastasis (Dissemination-Score) Snail E-Cadherin Capan-1 0,39 7,64 5,9±0,6* HPAF-2 1,03 5,58 6,9±0,5* MIAPaCa-2 7,17 0,10 14,3± 0,9 Panc-1 11,42 0,69 10,8±0,7 HUVEC (cont.) 1 1 p<0.05: * vs. MIAPaCa-2/Panc-1 Conclusions: The transcription factors and EMT-regulators Snail and Slug are expressed in pancreatic cancer but not in normal tissue. The upregulation of Snail is associated with low expression of the adhesion molecule E-Cadherin, suggesting a role for Snail in the progression and metastasis of human pancreatic carcinomas.

37 MEMBRANE-PROTEIN EXCTRACTIONS FOR SUBSEQUENT PROTEIN PROFILING OF CLINICAL TISSUE S. Bünger, B. Fritzsche, T. Gemoll, H-P. Bruch, U.J. Roblick, H. Jörnvall, G. Auer, J.K. Habermann Laboratory for Surgical Research and Department of Surgery, University of Schleswig-Holstein, Lübeck, Germany; Department of Medical Biochemistry and Biophysics, and Biomics Center, Karolinska Institutet, Stockholm, Sweden Keywords: membrane proteins, isolation, extraction, methodical approach Background: Membrane proteins are critical for normal cellular differentiation and diversified functions. Alterations in these proteins often lead to cell dysfunction and diseases. Membrane proteins are also the target for 70–80% of all drugs and the identification of new, differentially expressed membrane proteins reflecting certain disease properties has the potential of providing novel biomarkers and

773 therapeutic targets (Carter, P et al., 2004). However, the study of membrane-bound proteins can be difficult due to their relative low abundance in total cell lysates, their frequently large size and their hydrophobic properties. We thus aim to define effective protocols and systems that allow the enrichment of membrane-bound proteins for subsequent protein profiling. Methods: Traditional methods for isolation of membrane proteins are interminable, time consuming, and require gradient separation with expensive ultracentrifugation equipment. By far the most commonly used method is a multistep detergent extraction combined with lengthy ultracentrifugation (Ohlendieck K, 1996, Helenius A. et al., 1975, Tanford C et al., 1976). The faster, easier and less expensive way to isolate membrane proteins from mammalian cells or tissue is to use one of the commercial membrane protein extraction kits. Results: Our review presents a summary and comparison of available membrane extraction kits and related literature. We compare five commercial extraction kits with conventional ultracentrifugation protocols regarding (technical) requirements, time, costs, sample volume, and protein yield. Conclusions: Ideally, the isolation process should be mild yet rapid und inexpensive. The ideal extraction method will also depend on subsequent protein profiling methodologies to use such as 2-DE, SELDI, and/or mass spectrometry.

38 EXPRESSION OF THE ZINC FINGER TRANSCRIPTION FACTOR SNAIL IN ADRENOCORTICAL CARCINOMA IS ASSOCIATED WITH DECREASED SURVIVAL Jens Waldmann, Georg Feldmann M.D., Emily P.Slater Ph.D., Peter Langer M.D., Malte Buchholz Ph.D., Annette Ramaswamy M.D., Wolfgang Saeger M.D., Matthias Rothmund M.D., Detlef K. Bartsch M.D., Volker Fendrich M.D. Department of Surgery, Department of Internal Medicine, Division of Gastroenterology and Department of Pathology, Philipps-Universität Marburg, Germany Keywords: Epithelial-mesenchymal transition, Snail, Adrenocortical carcinoma Background: In this study we evaluate whether Snail is expressed in adrenocortical cancer (ACC) and if its expression is related to patient outcome. One of the best known functions of the zinc-finger transcription factor Snail is to induce epithelial-to-mesenchymal transition (EMT). Increasing evidence suggests that EMT plays a pivotal role in tumor progression and metastatic spread. Methods: Snail and E-cadherin expression were assessed by immunohistochemistry in 26 resected ACCs and real-time quantitative RTPCR expression analysis was performed. Data were correlated with clinical outcome and in particular with overall patient survival. Results: Seventeen of 26 (65%) ACC tumor samples expressed Snail when assessed by immunohistochemistry. Snail expression was neither detected in normal adrenocortical tissue, nor in benign adrenocortical adenomas. Expression levels were confirmed on the mRNA level by Real-Time PCR. Survival rates were significantly decreased in Snail positive tumors compared to Snail negative tumors: 10/16 vs. 1/ 8 patients succumbed to disease after a median follow up of 14.5 and 28.5 months, respectively (p=0.03); 2 died of unrelated causes. Patients with Snail expressing ACCs presented in advanced disease (11/12 vs. 6/ 14, p=0.01) and tend to develop distant metastases more frequently than patients with negative staining (7/12 vs. 2/8, p=0.19), 6 patients presented with distant metastases at initial diagnosis. Conclusions: In conclusion we describe for the first time that Snail is expressed in a large subset of ACCs. Furthermore Snail expression is associated with decreased survival, advanced disease and higher risk of developing distant metastases.

774

39 THE ROLE OF KALLIKREINS 6 AND 10 IN PANCREATIC CANCER Felix Rückert, Mario Hennig (1), Constantina D. Petraki (2), Diana Wehrum (1), Marius Distler (1), Axel Denz (1), Michael Schröder (3), Gihan Dawelbait (3), Holger Kalthoff (4), Hans-Detlev Saeger (1), Eleftherios P. Diamandis (5), Christian Pilarsky (1)*, Robert Grützmann (1)* 1: Visceral, Thoracic and Vascular Surgery, University Hospital Carl Gustav Carus, Technical University of Dresden, 2: Department of Nephropathology, Evangelismos Hospital, Athens, 3: Bioinformatics Group, Biotechnological Centre, Technical University Dresden, 4: Division of Molecular Oncology, Clinic for General Surgery and Thoracic Surgery, Schleswig-Holstein University Hospitals, 5: Department of Pathology and Laboratory Medicine, Mount Sinai Hospital, Toronto Keywords: Kallikrein, Pancreatic cancer, Microenvironment Background: Kallikreins (KLK/hK) play an important role in tumor microenvironment and as cancer biomarkers in different cancer entities. Previous studies suggested an upregulation of KLK10 and KLK6 in pancreatic ductal adenocarcinoma (PDAC). Therefore we evaluated the clinicopathological role of these kallikreins and their value as biomarkers in PDAC. Methods: Differential expression was validated by DNA-microarrays and immunohistochemistry in normal and malignant pancreatic tissues. Sera concentrations of both kallikreins were evaluated using ELISA. In silico analysis of possible protein interactions and gene silencing of KLK10 in vitro using siRNAs gave further insights in the pathomechanisms. Results: Gene expression analysis and immunohistochemistry demonstrated a strong expression for KLK10 and KLK6 in PDAC. Statistical analysis showed that co-expression of these kallikreins correlated with a R1-resection status (p=0.027) and worse outcome for overall survival (p=0.028). Multivariate analysis proofed that co-expression is an independent prognostic factor for survival (p=0.017). Importantly, KLK10 knock-down in AsPC-1 cells significantly reduced cell migration, while computational analysis suggested interaction of KLK6 with angiogenetic factors as important mechanism. Conclusions: Co-expression of KLK10 and KLK6 plays an unfavourable role in PDAC. Our results suggest that this effect is likely mediated by interaction with factors of the extracellular matrix and enhancement of cancer cell motility.

40 LASER CAPTURE MICRO-DISSECTION- AN OPTIMIZED PROTOCOL FOR CYTOGENETIC CHARACTERIZATION OF COLON CANCER SPECIMEN Britta Fritzsche, Uwe J. Roblick (1), Stefanie Bünger (1), Marco Gerling (1), Timo Gemoll (1), Elke Gheribi (1), Kathrin Kalies (2), Hans-Peter Bruch (1), Jens K. Habermann (1) 1: Laboratory for Surgical Research, Department of Surgery, and 2: Department of Anatomy, University Hospital Schleswig-Holstein, Campus Lübeck, Ratzeburger Allee 160, 23538 Lübeck, Germany Keywords: colon cancer, laser capture microdissection, nucleic acids, method establishment Background: A late diagnosis of colorectal cancer leads to a significant reduction of average survival time. Colon cancer specimens consist of heterogeneous tissue including different cell types such as tumor-, inflammatory- and stroma-cells. For a better understanding of cancer tumorgenesis and for the tagging of potential biomarkers of high sensitivity and specificity, it is thus crucial to dissect uncontaminated cell types for further analyses. The method of Laser Capture Microdissection (LCM) allows the selection and

Langenbecks Arch Surg (2008) 393:759–815 dissection of e.g., tumor cells independent and free of surrounding cells. The major disadvantages of LCM however are the timeexcessive workload, the possible interference of section staining with subsequent techniques and the low yield of DNA-, RNA- and proteinextraction of the microdissected cells. Methods: In order to address these limitations of LCM we aimed at identifying an optimized workflow to gain a high DNA- and RNA-yield of high quality with decent workload and time investment. Thus, non aberrant, healthy colon samples were prepared for kryosectioning. Under RNase- and Dnase-free conditions sections at different thickness (e.g. 12, 25 and 30 micrometer) were placed on special slides for microdissection. After fixation and nuclear-staining sections were microdissected. The dissected cells were examined by different extraction methods for RNA and DNA (classical method adapted to Chomczynski and Sacchi vs column-assisted-method). The quantity and quality of the extracted nucleic acids has been estimated spectro-photometrically and electrophoretically. Consequently the necessary minimum amount (area) of cells for e.g. arrayCGH, microarrays, without further amplification of the nucleic acids was evaluated. Results: Our results showed 25 micrometers to be the optimum thickness of cryosections for further processing. Thinner and thicker sections do not provide a sufficient amount of nucleic acid resp. have not proved satisfactory handling during sectioning, staining and LCM process. The classical phenol-isothiocyanate extraction method often seems to lead to unspecific results in subsequent cytogenetic experiments (e.g. arrayCGH). Consequently we tested newly available commercial column-assisted extraction “kits” that also allow for simultaneous extraction of nucleic acids and proteins of the same sample. After protocol modifications, we defined the microdissect minimum cell area to be 35 million per square micrometer in order to gain high quality DNA- and RNA-yields of an amount that does not require amplification for subsequent analyses. Conclusions: The isolated nucleic acids are highly representative for the transcriptome and genome of the examined tumor. With this method we are able to detect characteristic chromosomal abbreviations, diseasespecific mutations in the examined genome and specific expression pattern of certain cell types within a given sample type. Such workflow will thus allow further elucidating tumorigenesis and development of innovative diagnostics and therapeutics at early disease stages.

41 ACTIVATING TRANSCRIPTION FACTOR-3 (ATF3) FUNCTIONS AS A TUMOR SUPPRESSOR IN COLON CANCER AND IS UP-REGULATED UPON HEAT SHOCK PROTEIN 90 (HSP90) INHIBITION Christina Hackl, Christian Moser, Sven A. Lang, Kathrin Stengel, Hans J. Schlitt, Edward K. Geissler, Oliver Stöltzing Department of Surgery and Surgical Oncology, University of Regensburg Medical Center, Regensburg, Germany Keywords: ATF3, Hsp90, CREB, colon cancer, tumor suppressor Background: Activating transcription factor-3 (ATF3) is a member of the ATF/cAMP-responsive element binding protein (CREB) transcription factor family and is involved in the cellular response to a large variety of stresses including DNA damage. Importantly, ATF3 elicits tumor suppressive functions and its dysfunction has been associated with an impaired p53-mediated cellular response to DNA damage, thus allowing cells to be transformed by oncogenes. In our study we sought to investigate the regulation and role of ATF3 in colon cancer cells, as this has not been well defined to date. In particular, we hypothesized that blocking Hsp90, which represents a novel target for molecular cancer therapies, would alter ATF3 expression and regulation through multiple pathway inhibition.

Langenbecks Arch Surg (2008) 393:759–815 Methods: Human colon cancer cell lines (HCT116, SW620, HT29) and the Hsp90 inhibitor 17-DMAG were used for experiments. HCT116 colon cancer cells were transfected with either a ATF3shRNA, or Luciferase-shRNA expression plasmid, and stable clones selected and screened. Changes in signaling pathway activation and ATF3 expression were determined by Western blotting and real-time PCR. The effects of ATF3 inhibition on tumor growth in vivo were determined in a subcutaneous tumor model in athymic nude mice (BALB/cnu/nu). The effects of ATF3 suppression on peritoneal carcinomatosis, and on the anti-neoplastic efficacy of the Hsp90 inhibitor 17-DMAG are currently being determined in vivo. Results: ATF3 expression was slightly detectable in all colon cancer cells. Surprisingly we found that blocking Hsp90 substantially upregulated ATF3 expression in colon cancer cells on both protein and mRNA levels, in addition to a 17-DMAG-associated inhibition of multiple oncogenic signaling pathways, suggesting that anti-metastatic effects of Hsp90 inhibitors may be due to an ATF3 upregulation. Using selective pathway inhibitors, we identified that blocking MAPK/Erk up-regulated ATF3, whereas inhibition of PI3K/Akt, SAPK, and p38 did not. Stable knock-down of ATF3 in HCT116 not only decreased ATF3 mRNA, but also markedly diminished the Hsp90 inhibitor mediated induction of ATF3 protein. In vivo, suppression of ATF3 by RNAi (2 clones tested) significantly accelerated subcutaneous tumor growth of HCT116 cells, as compared to Luciferase-shRNA transfected controls (P<0.05 for both), suggesting that ATF3 indeed functions as a tumor suppressor in colon cancer. Results from above mentioned additional in vivo experiments for determining effects on peritoneal carcinomatosis and response to Hsp90 inhibition are pending and will be presented at the meeting. Conclusions: The transcription factor ATF3 functions as a tumor suppressor in colon cancer and is inducible by blocking Hsp90. Hence, Hsp90 inhibitors appear to elicit their anti-tumor efficacy at least in part through up-regulating ATF3, suggesting that ATF3 could be an important modulator of the efficacy of Hsp90 inhibitors. This latter aspect is currently being investigated.

42 ELUCIDATING ANEUPLOIDY AS PROGNOSTIC MARKER FOR ULCERATIVE COLITIS-ASSOCIATED COLORECTAL CARCINOGENESIS Marco Gerling (1), Karl-Frederick Meyer (1), Sandra Freitag-Wolf (3), Sampsa Hautaniemi (4), Kari Nousiainen (4), Timo Gemoll (1), Britta Fritzsche (1), Stefanie Bünger S (1), Hendrik Schimmelpenning (2), Thomas Ried (5), Gert Auer (6), Hans-Peter Bruch (1,2), Uwe J. Roblick (1,2), Jens K. Habermann (1,2) 1: Laboratory for Surgical Research, 2: Department of Surgery, and 3: Institute of Pathology, University Clinic Schleswig-Holstein, Campus Lübeck, 23538 Lübeck, Germany, 3: Institute for Medical Informatics und Statistics, University Hospital Schleswig-Holstein, Campus Kiel, Germany, 4: Computational Systems Biology Laboratory, Biomedicum Helsinki, Finland, 6: Unit of Cancer Proteomics, Biomics Center Karolinska, Karolinska Institutet, Stockholm, Sweden, 5: Genetics Branch, National Cancer Institute, NIH, Bethesda, MD, USA Keywords: ulcerative colitis, colorectal cancer, carcinogenesis, DNA ploidy, gene expression, microarrays, diagnostics Background: Ulcerative colitis patients have a well known risk for colorectal cancer development. Since epithelial dysplasias have been defined as precursor lesions, intensive and expensive surveillance programs have been introduced. The diagnosis of dysplasias can be hampered by sampling and interpretation errors. Therefore, these programs still show uncertainty in the individual risk assessment. Additional cellular features with possible predictive value for

775 malignant transformation in ulcerative colitis are thus mandatory. Against this background, DNA ploidy was investigated for its diagnostic and prognostic potential and microarray analysis was applied to unravel target genes of distinct ploidy types. Methods: First, 683 mucosal biopsies of ulcerative-colitis patients were analyzed retrospectively for gross ploidy. These biopsies belonged to 16 colitis patients without and 8 colitis patients with subsequent carcinoma during follow-up period. The degree of dysplasia was graded according to Ridell (intraobserver reproducibility: K=0.78). Furthermore, 135 mucosal biopsies of additional 32 colitis patients and normal non-colitis mucosa samples of 7 patients were collected prospectively for microarray evaluation and analyzed for gross ploidy. DNA assessments were performed by means of image cytometry on tissue sections. The DNA histograms were classified into four different types according to Auer. Expression analysis was performed on 33K oligonucleotide microarrays using aRNA after Trizol extraction, Quiagen RNeasy purification and amplification (Ambion). Results: In the first sample cohort, no significant difference was found in the inflammatory activity or the quantity and degree of dysplasia between the patients that did or did not progress to cancer. However, nuclear DNA assessments revealed aneuploid epithelial cell populations more frequently in patients with subsequent carcinoma (p= 0.006). These aneuploid lesions were distributed over the whole colon and rectum and not related to dysplasia. Aneuploidy could be observed at average 7.8 years prior to the final carcinoma diagnosis. In contrast, the colon epithelium of the patients without a subsequent carcinoma mainly exhibited proliferative-diploid DNA distribution patterns. Gene expression analysis revealed 371 genes differentially expressed between normal non-colitis affected colon mucosa and ulcerative colitis-associated biopsies (fdr <0.05). We then compared diploid and aneuploid colitis-associated tissue and detected 28 genes that were 2–10.4-fold differentially expressed. These differentially expressed genes were analyzed using Ingenuity Pathway Analysis (IPA), which revealed that these genes are mainly involved in pathways of cellular movement, cell death, and cancer (IPA network score 18). Conclusions: Genomic instability, represented by DNA aneuploidy, might initiate the process of malignant transformation in colitis as an early event. Thus, ploidy assessment of mucosal biopsies may identify premalignant lesions with invasive capacity. Differentially expressed genes that are characteristic for distinct ploidy types might further ameliorate our understanding of carcinogenesis and provide useful targets for improved early diagnostics and innovative treatment interventions.

43 BLOCKADE OF ENDOTHELIAL PRECURSOR CELLS ENHANCES TUMOR GROWTH OF COLORECTAL METASTASIS DUE TO STROMAL CELL-DERIVED FACTOR (SDF)-1 RELATED STIMULATION OF ANGIOGENESIS AND TUMOR CELL PROLIFERATION Kathrin Rupertus, Gudrun Y Haberl, Martin K. Schilling, Michael D. Menger, Otto Kollmar Abt. für Allgemein-, Viszeral-, Gefäß- und Kinderchirurgie, Uniklinikum des Saarlandes Keywords: Colorectal cancer, angiogenesis, EPC, c-kit, SDF-1, chemokine Background: Colorectal cancer is one of the leading causes of cancerrelated mortality. Tumor progression and uncontrolled metastatic spread are closely related to tumor angiogenesis. Mobilization into the peripheral blood of c-kit+ bone marrow-derived endothelial precursor cells (EPCs) contributes to the formation of new blood vessels. The c-kit receptor is essential for survival and function of EPCs. Migration and ‘homing’ of EPCs to sites of tumor growth is

776 directed by chemotactic signaling in which binding of stromal cellderived factor (SDF)-1 to its receptor CXCR4 plays an important role. Targeting migration and proliferation of EPCs provides a promising new strategy of anti-angiogenic tumor therapy. Therefore, we studied the influence of the chemokine SDF-1 on angiogenesis and tumor growth related to c-kit+ EPCs. Methods: Angiogenesis and tumor growth in vivo was analyzed using GFP-transfected CT26.WT colorectal cancer cells which were implanted into the dorsal skinfold chambers of syngeneic BALB/cmice (n=24). The animals were randomized into 3 groups: 16 animals were pretreated with a monoclonal anti-c-kit-antibody starting 4 days before tumor cell implantation. Eight of these animals received additional injections with a monoclonal anti-SDF-1 antibody. Animals treated with a control IgG antibody (n =8) served as controls. Angiogenesis, tumor growth, tumor cell proliferation and apoptosis were studied using intravital fluorescence microscopy, histology and immunohistochemistry during an observation period of 14 days. Results: Blockade of c-kit stimulated significantly tumor growth with an increase of tumor area (3.02+/−0.26 vs. 4.25+/−0.58 mm2) compared to controls. Interestingly, angiogenesis and tumor cell migration were not markedly affected. This enhancement of tumor growth was associated with a significantly increased tumor cell proliferation (44 +/− 5 vs. 58 +/− 2%) and a decreased rate of apoptotic cell death (0.28+/−0.05 vs. 0.06+/−0.04%) especially next to the tumor border. Neutralization of SDF-1 significantly decreased the enhanced tumor growth by anti-c-kit treatment (2.61 +/− 0.21 mm2) comparable with values of the control group. These anti-c-kit/anti-SDF-1 treated tumors showed significantly decreased angiogenesis and capillary densities as well as a significant reduction of tumor cell infiltration, proliferation and apoptosis compared to the other groups. Conclusions: The results of our study indicate that targeting EPCs by blockade of the c-kit receptor enhances local tumor growth of extrahepatic colorectal metastasis. Interestingly, this stimulation of tumor growth by anti-c-kit treatment can be antagonized by additional blockade of SDF-1. We therefore conclude that the stimulation of tumor growth after blockade of the c-kit receptor is mediated by SDF-1.

44 POOR OUTCOME IN PRIMARY NON-SMALL CELL LUNG CANCERS IS PREDICTED BY TRANSKETOLASE-1 (TKTL1) EXPRESSION G. Kayser, W. Sienel, B. Kubitz, D. Mattern, J. Coy, B. Passlick, M. Werner, A. Zur Hausen Abteilung Thoraxchirurgie, Chirurgische Universitätsklinik, Universitätsklinikum Freiburg; Institut für Pathologie, Universitätsklinikum Freiburg; R-Biopharm AG, Darmstadt Keywords: Lung cancer; Transketolase; Warburg effect Background: Malignant tumors are able to ferment glucose to lactate even in the presence of oxygen (Warburg effect). Within the nonoxidative pentose phosphate pathway (PPP) controlled by transketolase, glucose is degraded into ribose for nucleic acid synthesis and into lactate. TKTL1 has previously been shown to encode a transketolase-like enzyme which is overexpressed in colon, urothelial and breast cancer. Here we investigated the prognostic impact of TKTL1 expression in nonsmall cell lung cancer (NSCLC). Methods: Primary tumors of 201 consecutive patients with completely resected NSCLC (pT1-4, pN0-2, cM0, R0) were stained by immunohistochemistry immunohistochemistry using a new monoclonal TKTL1 antibody. The prognostic relevance of TKTL1 expression was evaluated by Kaplan-Meier and multivariate Cox regression analysis.

Langenbecks Arch Surg (2008) 393:759–815 Results: 89 tumors (44.7%) showed no or mild TKTL1 expression, whereas in 110 tumors (55.3%) it was overexpressed. TKTL1 overexpression correlated with tumor-type (adenocarcinoma: 65%, squamous cell carcinoma: 82%, large cell carcinoma: 93%; p=0.0007) and histological grading (G2: 61% vs. G3: 76%; p=0.033). Overexpression of TKTL1 was significantly associated with poor patient survival (p= 0.008). In addition, TKTL1 overexpression identified patients with poor clinical outcome among lymph node negative (p=0.039) and well to moderately differentiated (p=0.005) NSCLCs. The prognostic effect of TKTL1 expression in NSCLC was independent in multivariate analysis (p<0.0223). Conclusions: Our data suggest that TKTL1 overexpression is a new and independent predictor of patient survival in primary NSCLC. Since inhibition of transketolase enzyme reactions has recently been shown to effectively suppress tumor growth, TKTL1 might represent a novel pharmacodiagnostic marker.

45 ELEVATED EXPRESSION OF VASCULAR ENDOTHELIAL GROWTH FACTOR RECEPTOR 3 IS CORRELATED WITH INCREASED RISK OF OCCURRENCE OF OCCULT LYMPH NODE METASTASIS AND IS ASSOCIATED WITH POOR CLINICAL OUTCOME IN NON-SMALL CELL LUNG CANCER (NSCLC) Sebastian Dango, Moritz Schreiber, Wulf Sienel, Gian Kayser, Bernward Passlick und Christian Stremmel Abt. Thoraxchirurgie, Chirurgische Universitätsklinik Freiburg und Abt. Pathologie, Universitätsklinik Freiburg Keywords: occult lymph node metastasis, vascular endothelial growth factor receptor (VEGFR)-3, lymphangiogenesis, prognosis, immune escape Background: Cancer cells spread out of the original tumor through the lymphatic capillaries and, therefore, promote tumor progression. Detection of cancer cells in lymphatic vessels and regional lymph nodes is the strongest prognostic factor and is associated with poor survival. Although the clinical relevance of lymph node involvement is well known, the molecular mechanisms by which tumor cells spread via lymphatic vessels are less understood. Enormous progress has occurred in this field during the past decade through the identification of specific lymphatic endothelial cell markers such as vascular endothelial growth factor receptor (VEGFR)-3 and its ligands VEGF-C and VEGF-D. The intention of this study was to investigate a possible impact of these endothelial markers on occult lymph node metastases and lymphangiogenesis in patients with operable non-small cell lung cancers (NSCLC). Methods: In 76 patients with in toto resected NSCLC the expression of VEGFR-3 and its ligand VEGF-C was investigated immunhistochemically. Further more, lymphadenectomy specimen were incubated with a monoclonal antibody Ber-EP4 against 17-1A antigene (EpCAM) to detect possible occult lymph node metastases. Presence of lymphatic vessels and microvessel density was investigated using an anti-podoplanin antibody and an anti-CD-31 antibody, respectively. Results were compared with clinico-pathological parameters and cancer-related survival was calculated with the univariate and multivariate analysis. Results: In 16 out of 22 patients with occult lymph node metastases a significant increased VEGFR-3 expression (p=0.03, Chi2-Test) was seen. An occult lymph node metastases was associated with a poor prognosis (p=0.008, rel. risk 2,41, Cox-Regeression-Analysis). Correlation was found between presence of occult lymph node metastases and pathological lymph node status (p=0.01). A high VEGFR-3 expression showed a tendency toward a decreased cancer-related

Langenbecks Arch Surg (2008) 393:759–815

777

survival (p=0.07). Presence of occult lymph node metastases was associated with a high microvessel density (p=0.01), no correlation was seen with measured lymphatic vascularisation. Conclusions: Herein, we show for the first time a significant increased VEGFR-3 expression is associated with a higher incidence of occult lymph node metastases and may promote lymphatic spread out in patients with NSCLC. A possible explanation might be that lymphangiogenesis is promoted due to an immune escape of the tumor by a decreased microvascularisation and, therefore, with a reduced cellular immune response. A possible benefit of a neoadjuvant antibody therapy against VEGFR-3 to decrease micrometastases and lymphatic spread out must be investigated in further studies.

significant independent factor was HPSE for predicting the grade of dedifferentiation. Conclusions: This study suggests that HPSE plays a crucial role for the aggressiveness of pancreatic cancer. Though these results were obtained on a relatively small number of patients, larger studies including patients treated with actual chemotherapeutics seem to be warranted.

46 PROGNOSTIC IMPACT OF HEPARANASE GENE EXPRESSION IN PANCREATIC CANCER AND ITS CORRELATION TO MARKERS ENCLOSED IN THE HEPARAN SULPHATE GLYCOSAMINOGLYCANS

Alexis Ulrich, Mohamed El-Ghamari; Frank Bergmann; Jürgen Weitz; Markus Büchler Department of Surgery, University of Heidelberg; Department of Pathology (F Bergmann)

Daniel Vallböhmer (1), Andreas C. Hoffmann (1,2), Jan Brabender (1), Uta Drebber (3), Stephan Baldus (4), Ralf Metzger (1), Arnulf H. Hölscher (1), Kathleen D. Danenberg (2), Klaus L. Prenzel (1), Peter V. Danenberg (2) 1: Klinik und Poliklinik für Visceral- und Gefäßchirurgie, Universität zu Köln, 2: Department of Biochemistry and Molecular Biology and Norris Comprehensive Cancer Center, University of Southern California; Response Genetics Inc.; Los Angeles, CA, USA, 3: Institut für Pathologie, Universität zu Köln, 4: Institut für Pathologie, Universität Düsseldorf Keywords: pancreatic cancer; prognostic factors; heparanase Background: Pancreatic cancer still has the worst prognosis of all solid tumors with a 5-year survival rate of 5%. Due to late symptoms and therefore often delayed diagnosis, only 10–20% of patients with pancreatic cancer are eligible for complete resection with curative intention. Consequently, markers or gene-sets are required which would further classify patients into different risk categories and thus allow more individually adapted multimodality treatment regimens. Especially heparanase (HPSE) has recently been discussed as a key molecular factor in pancreatic carcinogenesis. The aim of this study was to analyze the correlation of intratumoral HPSE gene expression with clinical and histopathological variables and its correlation to markers enclosed in the heparan sulphate glycosaminoglycans, such as basic fibroblast growth factor (bFGF), heparin-binding EGF-like growth factor (HB-EGF), hypoxia inducible factor-1 alpha (HIF1a) and platelet-derived growth factor alpha (PDGFA). Methods: Paraffin-embedded surgical specimens were obtained from 41 patients [23 male, 18 female; median age: 65 years (range: 34–85 years)] with pancreatic cancer who underwent primary surgical resection. After laser capture microdissection, direct quantitative real-time reverse transcriptase PCR (RT-PCR, TaqMan™) assays were performed to determine HPSE, bFGF, HB-EGF, HIF1a and PDGF-A gene expression levels. Gene expression was normalized with beta-actin. Decision tree analysis and the maximal chi-square method were adapted to determine which gene expression value best segregated patients into lymph node negative and positive, and high and low dedifferentiation subgroups. Results: HPSE was significantly correlated to PDGFA (p=0.04) and to HIF1a (p=0.04). The correlation of HIF1a to bFGF and HB-EGF expression was significant (p=0.04, p=0.02). We put all clinical, histopathological parameters and the used genes as independent variables in a stepwise multiple linear regression model with lymph node metastasis as the dependent variable. The overall model fit had a significance level of p=0.029 with HPSE as the only significant predictor of lymph node metastasis. Using a stepwise multiple regression analysis to evaluate the most influential of the accessible factors on the dedifferentiation of the tumor the overall model fit was significant at a level of p=0.003. The most

47 LANGERHANS ISLETS PLAY A ROLE IN CARCINOGENESIS OF PANCREATIC DUCTAL ADENOCARCINOMAS IN AN ANIMAL MODEL

Keywords: pancreatic cancer; islet cells; animal model Background: In the Syrian Golden Hamstermodel pancreatic ductal adenocarcinoma developed after treatment with N-nitrosobis(2-oxopropyl)amin (BOP). Further investigations indicated a central role of the Langerhans islets in the process of carcinogenesis. In contrast, mice were not affected and showed no tumour formation after treatment with BOP. It was the goal to investigate if pancreatic tumours develop after orthotopic implantation of hamster islets into the pancreas of scid mice and treatment with BOP. This would indicate that pancreatrophic carcinogens are metabolized primarily by the target cells and that islet cells are involved in pancreatic carcinogenesis. Methods: scid mice were seperated into two groups of 12 animals, each. 500 hamster islets were implanted in the splenic lobe of the mice pancreas in the treatment group by mini laparotomy, the animals of the control group received a sham operation with injection of 150 µl NaCl. All animals were treated with BOP over 5 weeks with aweekly application of 1 mg/kg body weight. One year later the animals were killed and investigated for the formation of tumours. Tissue from pancreas, liver, spleen, kidney and lung was fixated in paraffin for immunohistochemical investigations and cryoasservated for molecular tests. Results: Pancreatic adenocarcinoma developed in three animals of the treatment group compared to no tumour in the control group. The average size of the tumours was 1.5 cm. The tumours showed histopathologic signs of ductal adenocarcinomas, partly with mucinous dilated, dysplastic epithelia comparable to intraepithelial neoplasias. One animal had metastases in the liver and the lung. Immunhistochemical investigations have shown the expression of Muc4 in the ductal adenocarcinomas whereas in non-tumerous pancreatic tissue only the islet cells were stained. A similar distribution was detected for drug metabolizing enzymes. Conclusions: In this animal model the carcinogen BOP could only be metabolized or activated within the implanted hamster cells, causing the formation of ductal adenocarcinomas. Therefore, islet cells seem to play a role in pancreatic carcinogenesis.

48 ROLE OF METASTASIS SUPPRESSOR GENES AND THEIR EPIGENETIC REGULATION BY PROMOTER METHYLATION IN PANCREATIC CARCINOMA Wolf Arif Mardin, Sören Torge Mees, Kostadin Petrov, Andreas Enns, Jörg Haier Klinik für Allgemeine Chirurgie, Universitätsklinikum Münster, Münster, Germany

778 Keywords: Pancreatic Carcinoma, Metastasis Suppressor Gene, Epigenetic Regulation, Promoter Methylation Background: The adenocarcinoma of the exocrine pancreas belongs to the most aggressive malignancies with an overall 5-year survival rate of less than 5%. A characteristic of this particular tumor entity is the early development of distant metastases, the mechanisms of which remain poorly understood. In a series of newly established orthotopic models for pancreatic carcinoma tumor growth, metastatic and infiltrative potential were investigated and compared to data generated from genetic and epigenetic analyses. Methods: Using 16 pancreatic cancer cell lines of varying metastatic potential, we investigated their metastatic and infiltrative potential in a series of newly established mouse models. For all cell lines Methylation Specific PCR (MSP) and Bisulfite Sequencing PCR (BSP) were performed in order to determine the promoter methylation status of the majority(12) of the known Metastasis-Suppressor-Genes. The respective mRNA expression levels were determined by TaqMan Assay based Real Time Reverse Transcriptase PCR for each gene and cell line. Infiltration and metastasis scores of the mouse models were compared to the data obtained from the epigenetic and mRNA analyses. Results: 5 of 12 investigated genes (AKAP12, CDH1, GPR54, MASPIN, TIMP3) showed significant promoter methylation while the remaining genes did not show methylation. This methylation pattern correlated with loss of expression in 2 of 5 cases (MASPIN, CDH1) and did not show significant correlation in AKAP12, TIMP3 and GPR54. Using bisulfite direct sequencing, we were able to show inhomogeneous methylation at certain CpG sites. Expression of Metastasis-Suppressor-Genes only differentially corresponded with a less aggressive phenotype. For AKAP12 and NDRG1, higher mRNA expression corresponded to lower metastatic and infiltrative potential. Increased levels of CRSP3 mRNA corresponded to reduced infiltrational potential. For BRMS1 and MASPIN, higher levels of mRNA expression correlated with elevated tumor metastatic potential. Statistically significant correlations were not found for the other genes. Conclusions: The Metastasis Suppressor MASPIN is mainly, if not exclusively, regulated by promoter methylation. In normal Human Pancreatic tissue, the MASPIN promoter is methylated which means that certain cancer cell lines undergo loss of methylation in order to reexpress MASPIN. CDH1 is also regulated by promoter methylation, but other mechanisms must also be relevant. BRMS1 and MASPIN mRNA expression correlate with increased metastatic potential, indicating that Metastasis Suppressor genes should be evaluated for single tumor entities and generalization of results from different tumor entities is not always possible.

49 RNAI-BASED IDENTIFICATION OF POTENTIAL TARGETS IN COLORECTAL CANCERS Marian Grade, Amanda B. Hummon, Patrick Hörmann, Georg Emons, Jordi Camps, Jochen Gaedcke, Michael J. Difilippantonio, B. Michael Ghadimi, Natasha J. Caplen, and Thomas Ried Department of General and Visceral Surgery, University Medicine, Göttingen; Genetics Branch, National Cancer Institute, NIH, Bethesda, Maryland Keywords: Colorectal cancer, expression profiling, differentially expressed genes, RNA interference, therapeutic targets Background: Despite the implementation of sophisticated therapeutic strategies into clinical practice, colorectal cancer still represents a major cause of cancer death in the Western world. Thus, understanding cancer progression and establishing novel therapeutic options remains of considerable clinical interest. Methods: We have recently profiled a series of 90 primary colorectal cancers and 50 matched normal mucosa biopsies using gene

Langenbecks Arch Surg (2008) 393:759–815 expression microarrays and identified a set of differentially expressed genes, among them up-regulated mRNA levels of MYC and HMGA1. Towards functional validation, mRNA expression levels of 28 differentially expressed genes were established in 25 colorectal cancer cell lines by semi-quantitative real-time PCR. Using RNAi analysis, we then systematically silenced a subset of these genes in SW480 cells, and screened for siRNA duplexes that reduced cellular viability. The siRNA-mediated reduction in mRNA levels was validated 48 hours after transfection using a branched-DNA/RNA assay. Results: Screening our panel of 25 colorectal cell lines, we first confirmed that the majority of genes were similarly deregulated comparing colorectal cancers and matched mucosa samples and comparing the cell lines and a mucosa pool. Using RNAi analysis, we then silenced 14 highly up-regulated genes in SW480 cells using two siRNAs per gene, and could show that knockdown of a subset of these genes resulted in reduced cellular viability. This effect was independently confirmed for up to four different siRNA duplexes, and for HT-29 as well as DLD-1 cells. Conclusions: Our experimental strategy led to the identification of novel genes critical for colorectal tumorigenesis, and we have now started to analyze the global transcriptomic changes that occurred as a consequence of gene silencing using gene expression microarrays. We surmise that some of these genes represent potential targets for therapeutic intervention.

50 EFEMP PROTEIN OVEREXPRESSED IN HIGHLY METASTATIC PANCREATIC CARCINOMA CELLS BINDS TO EGF RECEPTOR AND ACTIVATES MAPKAND AKT- PATHWAYS Peter Camaj, Ivan Ischenko, Hendrik Seeliger, Stefan Krebs, KarlWalter Jauch, Christiane J. Bruns Experimentelle Forschung: Chirurgie, Klinikum Großhadern, Marchioninistr. 15, 81377 München, Laboratory for Functional Genome Analysis, Feodor-Lynen-Str. 25, 81377 München Keywords: EFEMP1, EGFR, Akt, Erk, Phosphorylation Background: Pancreas carcinoma belongs in very aggressive tumor disease with very bad prognosis. Therefore it is very important to understand the processes leading to change of low aggressive non metastatic tumor to the highly metastatic and proangiogenic one. We have available very elegant model consisting of two pancreas carcinoma cell lines: parental FG cell line and L3.6PL generated from FG via three cycle of in vivo selection, when liver metastases were isolated from orthotopically injected nude mice, cultivated and further intra-pancreatic injected. Comparison of the transcriptomes of highly metastatic and proanagiogenic cell line pancreas carcinoma L3.6pl versus closely-related low metastatic and low angiogenic cells FG has revealed upregulation of the gene EFEMP1 in aggressive cell line. Animal experiment has confirmed protumorigene role of this gene in vivo. Multiple sequence alignment has shown strong similarity of EFEMP1 protein sequence with EGFR-ligand heparinbinding EGF-like protein especially in the putative receptor-binding region conserved in all EGFR-ligands. Therefore we have investigated possible interaction of EFEMP1 protein with EGFR. Methods: Cell culture Immunoprecipitaiton: anti-EGFR antibody Immunoblotting: anti-EFEMP1, anti-phospho-EGFR, anti-phosphoAkt, anti-phospho-Erk1/2, anti total EGFR, Akt, Erk, anti-b-actin treatment with PD15035 Results: Using anti-EGFR antibody we are able to coprecipitate a protein of a identical molecular mass 54 kDa as EFEMP1 and being recognized by anti-EFEMP1 antibody. Amount of coimmunoprecipitated EFEMP1 was competitive reduced under treatment with EGF. Therefore we assume that EFEMP1 is able to bind EGFR on the same

Langenbecks Arch Surg (2008) 393:759–815 site as EGF or in its neighbourhood. Furthermore, we have investigated a signal transduction pathway activated upon treatment of L3.6pl cells with the purified EFEMP1 protein. Our results have shown that EGFR sites Tyr-992 and Tyr-1068 are phosphorylated under treatment of the starved L3.6pl cells with purified EFEMP1. Since this phosphorylation is the signal activating downstream MAPK- or Akt-signalling, we have investigated phosphorylation of these target molecules under treatment with EFEMP1. Phosphorylation of Erk p44 and p42 at the positions Tyr-202 and Tyr204 or prosphorylation of Akt at the position Thr-308 respectively, have been detected via immunoblotting using appropriate antiphospho- antibody. All these phosphorylation events are initiated by EGFR-signaling induced by binding of EFEMP1 protein o:n the receptor. EFEMP1-induced activation can be inhibited by treatment with PD15035: an inhibitor of EGFR-Tyrosine kinase activity. Conclusions: EFEMP1 is novel ligand of EGFR. Its binding to this receptor activates downstream pathways such as MAPK or Akt signalings respectively. This activation could substantively contribute to increased aggressiveness of L3.6pl cell line

51 EXTRACELLULAR MATRIX COMPOSITION INFLUENCES CHEMOKIN-INDUCED ADHESIVE AND MIGRATORY PROPERTIES OF BREAST CANCER CELLS Claudia Wendel, Andre Hemping-Bovenkerk, Julia Krasnyanska, Sören Torge Mees, Jörg Haier Dept. of General Surgery, University Hospital Münster, Germany Keywords: breast cancer, metastasis, liver, chemokines, integrins, extracellular matrix, cell migration Background: Adhesion and migration to extracellular matrix are important steps in metastasis. In breast cancer, CXCR4 is often overexpressed and the tumor cells are directed to their destinations where metastases are developed. This process differs between various organs that have different chemokine expression but also different compositions of extracellular matrix (ECM) proteins. The aim of this study was to investigate the effect of chemokine stimulation on breast cancer cells which adhere and migrate to different matrix proteins. Methods: Adhesion and migration of different tumor cell lines (MDAMB231 and 468) to the liver were investigated by intravital microscopy in Sprague-Dawley rats. Static adhesion assays and transwell assays were used to evaluate adhesion to and migration at ECM (Collagen, Fibronectin, Laminin) under chemotactic stimulation with SDF1α. The expression of chemokine receptors was shown using flow cytometry. Downstream signaling targeted at actin cytoskeleton was tested by Rho, Rac and Cdc42 pull down assays. Immuncytochemistry was performed to investigate CXCR4 localisation at cell surfaces. The migratory distance of tumor cells was observed using time lapse microscopy. Results: In vivo migration of MDA-MB231 cells, but not adhesion was impaired by CXCR4 inhibition through siRNA. Stimulation with SDF1α induced a concentration and matrix dependent increase in cell migration. Strong adhesion to type I collagen and fibronectin were found whereas migration was stimulated at laminin. Small GTPases Rho and Rac/Cdc42 were activated through SDF1α. Conclusions: Breast cancer cells adhere to and migrate at ECM components to arrest within their metastatic target organ. Their ultrastructure including ECM composition appears to determine the preference of the tumor cells for certain metastatic targets. The chemokine SDF1α serves as a chemoattractant to guide tumor cells to their target. Actin regulators such as Rho and Rac are activated by SDF1α to modulate these adhesive and migratory properties.

779

52 IDENTIFICATION AND FUNCTIONAL VALIDATION OF CANDIDATE ONCOGENES IN COLORECTAL CANCER Georg Emons, Jordi Camps, Marian Grade, Amanda B. Hummon, Quang T. Nguyen, Jochen Gaedcke, Heinz Becker, B. Michael Ghadimi, Thomas Ried Department of General and Visceral Surgery, University Medical Center, Georg-August-University Göttingen, Germany; Genetics Branch, National Cancer Institute, National Institutes of Health, Bethesda, Maryland, USA Keywords: Colorectal cancer, expression profiling, oncogenes, siRNA screen Background: Colorectal carcinomas are characterized by a specific pattern of genomic imbalances, including, very prominently, copy number increases of chromosome arm 13q. The genes targeted by this recurrent aberration remain elusive Methods: To detect minimal regions of amplification on chromosome 13q, we first performed high-resolution arrayCGH of 31 primary colon carcinomas. To identify deregulated genes mapping to these regions, we then profiled a subset of these tumors (n=25) using whole-genome expression arrays. Next, we validated the expression levels of the over-expressed genes in 25 colorectal cancer cell lines using real-time PCR. Subsequently, we conducted a lipid-based RNAi screen to assess the resulting phenotypic consequences of gene knockdown in aneuploid SW480 cells, measured 72 and 96 hours after transfection using a CellTiter-Blue assay. mRNA silencing was confirmed 48 hours after transfection by real-time PCR. Results: Minimal regions of recurrent amplification were detected to be at chromosome 13q12.13-q12.3 and 13q34. We then identified 69 genes within these amplicons that were over-expressed in the tumors compared to normal samples. Using real-time PCR, 44 of these genes were also shown to be highly up-regulated in 25 colorectal cancer cell lines. Next, we silenced these 44 genes in SW480 cells using two siRNAs per gene, and observed a viability reduction of at least 20% for 10 genes (up to 70% reduction for some targets). This was confirmed in a secondary screen using two additional siRNAs per gene, and screening diploid DLD-1 cells. Conclusions: We identified genes that were amplified and/or overexpressed in primary colorectal cancers. For a subset of those, transfection with synthetic siRNAs dramatically impaired cellular viability. These genes might therefore represent potential oncogenes residing in chromosome 13q. While we were reassured to identify genes encoding for transcription factors, ATP and zinc ion binding proteins, receptor activity proteins, nuclear pore complex proteins, and protein/DNA binding gene products, the function of some of these genes has yet not been discovered.

53 PANCREATIC-DUODENAL HOMEOBOX 1 (PDX1) EXPRESSION DISTINGUISH BETWEEN DUODENAL AND PANCREATIC GASTRINOMAS Volker Fendrich, Ricarda Ramerth, Jens Waldmann, Peter Langer, Detlef K. Bartsch, Emily P. Slater, Matthias Rothmund Klinik für Viszeral-, Thorax- und Gefäßchirurgie, Uniklinik Marburg, Germany Keywords: PDX1, gastrinoma Background: Gastrinomas are responsible for Zollinger-Ellison syndrome (ZES). 80–90 per cent of all gastrinomas are located in the so-called gastrinoma triangle, which includes the duodenum, the pancreatic head and the hepatoduodenal ligament. The natural history of the tumors depends on their origin. Duodenal gastrinomas are much less aggressive than pancreatic primaries and infrequently develop liver metastases, the reason here fore is unclear. Pancreatic-duodenal

780 homeobox 1b (PDX1) is a transcription factor of homeobox genes family important in differentiation and development of the pancreas, duodenum and antrum. PDX1 is first present at 8.5 days of the embryo and selectively expressed in adult endocrine glands such as pancreatic beta-cells, or Brunner’s glands of the duodenum. This study aims to clarify the expression pattern of PDX1 in duodenal and pancreatic gastrinomas thereby evaluating a different origin of these tumors. Methods: Tumor tissue from 28 patients that underwent surgery between 1987 and 2007 at our institution because of a ZES were evaluated for Pdx1 expression by Immunohistochemistry (IHC) and real-time PCR. The diagnosis of ZES was based on clinical symptoms, biochemical tests and histopathology. Results: Tumor tissue from 15 patients with duodenal gastrinomas and from 11 patients with pancreatic gastrinomas were analyzed. Furthermore tissue from lymph node metastases from two patients with a so far undetected primary was analyzed. IHC and real-time PCR revealed strong PDX1 expression in pancreatic gastrinomas. In contrast, there was no PDX1 Expression detectable in all duodenal gastrinomas. This pattern was also true for their associated lymph node or liver metastases. PDX1 expression was absent in lymph node metastases from the two patients with an unknown location of the primary, suggesting a so far undetected duodenal gastrinoma. Conclusions: We show for the first time that clinical differences of duodenal and pancreatic gastrinomas can be explained by their different organic origin. Pancreatic gastrinomas express PDX1 which proves their pancreatic endocrine background, whereas duodenal gastrinoma develop from a PDX1 negative cell cluster. Furthermore, the expression pattern of PDX1 in a resected metastases can help to locate the primary gastrinoma.

54 ASSOCIATION OF ACTIVATING RET-PROTO-ONCOGENE MUTATIONS WITH SPORADIC NEOPLASM OF THE GASTRO-ENTERO-PANCREATIC AXIS Diana Wehrum, Felix Rückert, Heike Görgens, Hans-Konrad Schackert, Guido Fitze, Stephan Kersting, Hans-Detlev Saeger, Christian Pilarsky, Robert Grützmann Department of Visceral, Thoracic and Vascular Surgery, University Hospital Carl Gustav Carus, Technical University of Dresden, Fetscherstr. 74, 01307 Dresden; Department of Surgical Research, University Hospital Carl Gustav Carus, Technical University of Dresden, Fetscherstr. 74, 01307 Dresden; Department of Paediatric Surgery, University Hospital Carl Gustav Carus, Technical University of Dresden, Fetscherstr. 74, 01307 Dresden Keywords: RET, sporadic neoplasm, pancreatic cancer, gastric cancer Background: RET is a member of the cadherin superfamily and encodes a receptor tyrosine kinase. Gain-of-function mutations are causative for malignancies originating from the neural crest, like familial medullary thyroid carcinoma (FMTC). Activating mutations of exon 13 show a high incidence in sporadic MTC. This activation is influenced by other mutations, e.g. the two SNPs −5G>A and −1C>A and the polymorphism c.135G>A of exon 2, which decreases promoter activity. Latest reports suggest an involvement of RET in different sporadic tumours of the gastro-entero-pancreatic axis. Although mutations and polymorphisms of this gene are well documented, there are until now no systematic studies analysing if germline and somatic mutations and/or polymorphisms may be involved in the carcinogenesis of these sporadic tumours. Because of the high actual interest in the RET-proto-oncogene and the good knowledge about existing mutations we wanted to analyse if the existence of mutations of this gene might be underlying the malignant transformation of those tumour entities. Methods: We included 65 patients with gastric (GCA) and 54 patients with pancreatic carcinoma (PCA). As control group served 131

Langenbecks Arch Surg (2008) 393:759–815 anonymous blood donors. We isolated genomic DNA from peripheral blood and analysed mutations of RET-proto-oncogene by sequencing of the promoter, exon 2 and 13 including intron/extron boundaries. To assess the Hardy-Weinberg-Equilibrium (HWE) the frequencies of the identified variants of the RET-Proto-oncogene in each population were compared and the significance was estimated by the Pearson test with a significance-level of 5%. For each mutation the odds ratio for the different populations was assessed by the Cochrane-Armitage trend test. The frequencies of the haplotypes from the −5G>A, −1C>A, c.135G>A and c.2307T>G polymorphisms were estimated by the Likelihood-Ratio-Test. Results: In all groups we found the genetic variations c.2307T>G in codon 769 (L769L) and c.2372A>T in codon 791 (Y791F) of exon 13. The latter is a known activating missense mutation, which could not be found in any individual of our control population. Although not statistically significant, we could detect this mutation in 2 of 54 PDAC patients and 2 of 65 patients with gastric carcinoma (all p>0.05). Like previously reported we further found a distinct linkage disequilibrium between the variants −5G>A in the promoter and c.135G>A in exon 2 of the RET-proto-oncogene (Fitze et al., 2003). However, the estimation of the haplotype frequency of the variants −5G>A, −1C>A and c.135G>A showed no differences to the control population. Conclusions: We could not find new mutations of the RET protooncogene. However, we found the known activating mutation c.2372A>T (Y791F) which plays an unfavourable role in sporadic MTC in 2 of 54 cases of the PDAC-population and 2 of 65 in the GCpopulation. Plaza Menacho et al. showed that RET, which is activated normally by dimerisation of two receptor proteins, can also be activated in the presence of the Y791F mutation as monomeric protein. Above mentioned mutations cause a modified substrate specificity and ATP-binding capacity of the receptor, resulting in cellular transformation. Although statistically not significant, the tendency of our p-value with 0,0841 in the PDCA-population and 0,1088 in the GC-population suggests an general association with sporadic tumour, not only originating from the neural crest, but also from other tissues.

55 THE ROLE OF ORNITHINE DECARBOXYLASE (ODC) MRNA EXPRESSION IN THE PATHOGENESIS AND PROGNOSIS OF NON-SMALL LUNG CELL CANCER (NSCLC) Jan Brabender, Daniel Vallböhmer, Peter Grimminger, Georg Lurje, Frederike Ling, Paul M. Schneider, Elfriede Bollschweiler, Arnulf H. Hölscher, Ralf Metzger. Klinik und Poliklinik für Allgemein, Visceral- und Tumorchirurgie der Universität zu Köln; Klinik für Visceral und Transplantationschirurgie, Universität Zürich. Keywords: Lung Cancer, molecular prognostic marker, Ornithine Decarboxylase (ODC), Background: Ornithine Decarboxylase (ODC) is the first and rate limiting enzyme in the biosynthesis of polyamines. The role of ODC mRNA expression in the pathogenesis and prognosis of NSCLC remains unclear. Aim of this study was to investigate the mRNA expression of ODC in tumor and normal tissues of patients with NSCLC and to determine the potential of this gene as a molecular biomarker in this disease. Methods: ODC mRNA expression was analyzed in tumor and normal tissue of 91 patients with NSCLC by quantitative real-time RT-PCR (Taqman) with beta-actin as the internal control. Tumor stages were as follows: Stage I: 45 (49%) patients, Stage II: 19 (21%), Stage IIIA 27 (30%). 43 (47%) patients had squamous cell carcinoma and 33 (36%) had adenocarcinoma and 15 (17%) had large cell carcinoma. All tumors were R0 resected.

Langenbecks Arch Surg (2008) 393:759–815 Results: ODC mRNA expression was detectable in 100% of tumor and in 99% (90 of 91) of matching normal tissues. The median ODC mRNA expression in tumor tissue was 9.11 (minimum: 0.9; maximum 155.3) and in matching normal lung tissue 7.88 (min.: 0; max. 45.8; p<0.001; Wilcoxon test). There were no significant associations between histology, tumor stage, grading, age or gender and levels of ODC mRNA expression. The follow up was 85.9 months for the study population with a median survival of 59.7 months. A cut-off for an ODC expression of 10.0 best segregated patients into good and bad prognostic subgroups in terms of likelihood of surviving. Patients with a high ODC expression showed a significant better survival than patients with a low ODC expression in their tumor tissue (p=0.035, Log rank test). Multivariate analysis revealed ODC expression (p=0.022) and tumor stage (p<0.001) as independent unfavorable prognostic factors. Conclusions: A high ODC mRNA expression is associated with the pathogenesis of NSCLC. ODC mRNA expression is an important biomarker for a biologically aggressive disease in NSCLC. Quantitation of the mRNA expression of these genes might be helpful in identifying patients who would benefit from additional therapies for controlling their disease.

56 THE T393C POLYMORPHISM IN THE GENE GNAS1 AS A PROGNOSTIC AND PREDICTIVE MOLECULAR MARKER FOR RECTAL CANCER Hakan Alakus, Hakan Alakus (1), Daniel Vallböhmer (1), Ute Warnecke-Eberz (1), Elfriede Bollschweiler (1), Jan Brabender (1), Uta Drebber (2), Stefan P. Mönig (1), Arnulf H. Hölscher (1), Ralf Metzger (1) 1: Department of Visceral- and Vascular Surgery, 2: Institute of Pathology, University of Cologne, Cologne, Germany Keywords: SNP, G Protein, rectal cancer Background: The gene GNAS1 encodes the ubiquitously expressed Gαs subunit of heterotrimeric G proteins. Several recent studies have demonstrated that the T393C polymorphism of the gene GNAS1 is an independent prognostic factor for different tumor types. In these studies patients with TT genotypes showed a prolonged overall survival in bladder cancer, clear cell renal cell carcinoma as well in UICC I–II stage of sporadic colorectal cancer, whereas overall survival was higher for CC genotypes in invasive breast cancer and intrahepatic cholangiocarcinoma. Last year our institution could show that the T393C polymorphism is a predictive molecular marker for tumor response to Cisplatin/5-FU-based radiochemotherapy in esophageal cancer. Aim of the present study was to analyze the value of the T393C polymorphism as a predictive and prognostic molecular marker for 5-FU-based treatment radiochemotherapy in rectal cancer Methods: 73 patients (median age 59 years, range 31–86 years) from a prospective observation trial on neoadjuvant radiochemotherapy for rectal cancer were retrospectively genotyped to examine a potential association between T393C genotypes and histomorphologic tumor regression after 5-FU-based radiochemotherapy. Histomorphologic tumor regression was defined as major histopathologic response (MaHR) when tumor specimens contained less than 10% of residual vital tumor cells. Associations were evaluated using Pearson`s 2. Kaplan-Meier plots and the log-rank test for trend were used to evaluate the relationship between T393C genotypes and clinical outcome. Differences were regarded significant at a p-value <0.05. All statistical analysis was performed using SPSS 14.0 (SPSS, Chicago, USA). Results: The genotype distribution in the patient group was: 20 (28%) CC-genotypes, 22 (40%) TT-genotypes and 29 (41%) CTgenotypes. The C-allele frequency in the patient group was 0.48, which is not different from healthy control groups. 47 (64%) patients

781 had a minor response and 26 (36%) patients had a major response. Pearson’s 2-test showed no significant correlation between tumor regression grades with T393C genotypes. Survival was also not significantly associated with T393C genotypes. Conclusions: In contrast to our results in esophageal cancer, the T393C polymorphism of the gene GNAS1 seems not to be useful as a predictive molecular marker for tumor response to 5-FU-based radiochemotherapy in rectal cancer.

57 GENOME-WIDE EXPRESSION PATTERNS OF INVASION FRONT, CENTRAL TUMOR MASS AND CORRESPONDING NORMAL EPITHELIUM OF COLORECTAL CANCER PATIENTS Jörn Gröne, Heinz J Buhr, Eike Staub Chirurgische Klinik I, CBF, Charité — Universitätsmedizin Berlin, Germany; present address: Merck Serono, Bio- & Chemoinformatics, Darmstadt, Germany Keywords: colorectal cancer; gene expression; invasion front Background: Colorectal tumors have characteristic genome-wide expression patterns that allow their distinction from normal colon epithelia and facilitate clinical prognosis. The expression heterogeneity within a primary colorectal tumor has not been studied on a genome scale yet. Methods: Here we investigated three compartments of eight colorectal cancer patients (cryo tissues), the invasion front, the inner tumor mass, and corresponding normal epithelial tissue by laser microdissection and large scale microarray-based expression profiling. Results: In both tumor compartments many genes were differentially expressed when compared to normal epithelium. The sets of significantly deregulated genes in both compartments overlapped to a large extent and revealed various interesting known and novel pathways that could have contributed to tumorigenesis. Cells from the invasion front and inner tumor mass, however, did not show significant differences in their expression profile, neither on the single gene level nor on the pathway level. Instead, gene expression differences between individuals are more pronounced as all patientmatched tumor samples clustered in close proximity to each other. Conclusions: With respect to invasion front and inner tumor mass we conclude that the specific tumor cell micro-environment does not have a strong influence on expression patterns: largely similar genome-wide expression programs operate in the invasion front and interior compartment of a colorectal tumor.

58 ABERRANT EXPRESSION OF CHECKPOINT-KINASES IN PANCREATIC CARCINOMAS Frank Traub, Caroline Zug, Christine Kienzle, Susan Kupka, Alfred Königsrainer Department of General, Visceral and Transplant Surgery Tübingen, Germany Keywords: CHK1, ATM, ATR, pancreatic carcinoma, Background: Over the past two decades, research aimed at understanding and fighting the complex mechanisms of cellular transformation which lead to tumor development and helped to improve the detection and treatment of many cancers. However, some tumors have not benefited from these advances. Among them, tumors from the pancreas, which are rapidly invasive, metastatic, and resistant to standard therapies, remain the type of cancer with the worst clinical outcome—having a median survival of 6 months.

782 Methods: Pancreas tissue from 72 patients with pancreatic cancer and from 26 patients with pancreas adenoma or server pancreatitis was analysed for the expression of CHK1, ATM and ATR (mRNA). The snap-frozen pancreas tissue was micro-dissected to gain a higher proportion of carcinoma, adenoma or inflamation. The gene promoter methylation status was examined, to exclude a epigenetic phenomenon. Results: In 68% of pancreas cancer a suppression of CHK1 was detected, while an overexpression was found in all adenoma/pancreatitis tissues. Compared to histological normal tissue no hypermethylation could be detected. For ATM and ATR a non-uniformly expression pattern was observed. ATR was significant overexpressed in about ∼46% of carcinoma tissue. 9 patients had a significant suppression of ATR expression. An overexpression of ATR was detected in two-thirds of the adenoma/pancreatitis tissues. Regarding ATM again a mixed expression was found. 61% of the pancreas cancer samples showed an overexpression. In the adenoma and pancreatitis samples a non significant elevated expression was measured. Conclusions: The expression of the tumor suppressor gene CHK1 was significantly reduced in carcinoma tissue of the pancreas, while the regulating genes ATM and ATR showed a significant overexpression. An epigenetic cause for this result could be excluded. In further investigations the function of ATM and ATR as well as for CHK1 will be analysed using SNP screening.

59 TUMOR INHERENT DEGREE OF GENOMIC INSTABILITY LINKS BREAST CANCER SPECIFIC GENE EXPRESSION SIGNATURES WITH CLINICAL OUTCOME Jens K. Habermann (1,2,5), Jana Doering (2), Sampsa Hautaniemi (3), Uwe J. Roblick (2,5), Nana K. Bündgen (2), Ulrike Kronenwett (5), Gert Auer (5), Hans-Peter Bruch (2), Thomas Ried (1) 1: Genetics Branch, National Cancer Institute, NIH, Bethesda, MD, USA, 2: Department of Surgery, University Hospital SchleswigHolstein, Campus Lübeck, Germany, 3: Computational Systems Biology Laboratory, Genome-Scale Biology Research Program, Biomedicum Helsinki, and Institute of Biomedicine, University of Helsinki, Finland, 4: Institute of Signal Processing, Tampere University of Technology, Tampere, Finland, 5: Karolinska Biomic Center (KBC), Karolinska Institutet, Stockholm, Sweden Keywords: genomic instability, prognosis, gene expression, CGH, breast cancer Background: Breast cancer patients whose tumor cells show aneuploidy, i.e., a heterogeneous distribution of the nuclear DNA content, have an average disease free survival time of five years, which increases to ten years in patients whose tumors are genetically stable. Methods: In order to identify the gene expression signature of aneuploidy we analyzed 48 primary breast cancer specimens using global gene expression profiling on cDNA arrays. These analyses were complemented by mapping of genomic imbalances using comparative genomic hybridization. Results: Seventeen of these tumors were classified as diploid, 15 as aneuploid, yet genomically stable with a defined stemline, and 16 as aneuploid and unstable. These tumors displayed significant differences in the number and distribution of genomic imbalances: diploid tumors mainly revealed gains of chromosome arm 1q. The average number of gains and losses increased in the aneuploid, yet stable tumors, and the highest degree of genomic imbalances, including localized amplifications, were detected in the aneuploid unstable tumors. These distinct differences in the genomic aberration profiles were accompanied by specific transcriptional alterations, which showed that 81 genes were differentially expressed between the aneuploid unstable and diploid tumors and 79 genes between the aneuploid unstable and aneuploid stable tumors, whereas only 19 genes were different when comparing the diploid and aneuploid stable tumors.

Langenbecks Arch Surg (2008) 393:759–815 Conclusions: We conclude from these data that the major parameter affecting global gene expression profiles is not the actual ploidy status, but the inherent genomic instability; this genetic blueprint is also reflected in the poor clinical prognosis of the unstable tumors.

60 E-CADHERIN AND ITS TRANSCRIPTION REGULATORS TWIST AND SNAIL IN HUMAN COLORECTAL ADENOMAS F. Aydin, Z. Totikov, S.E. Baldus, G. Flügen, S. Seidschner, A. Rehders, C.F. Eisenberger, W.T. Knoefel, N.H. Stoecklein Klinik für Allgemein-, Viszeral- und Kinderchirurgie, Uniklinik Düsseldorf, Germany Keywords: EMT, colorectal adenom, snail, twist, E-Cadherin Background: Epithelial-mesenchymal transition (EMT) seems to be a critical event for invasion and metastasis of carcinoma cells. Downregulation of E-cadherin is one of the important steps in this mechanism. E-cadherin contributes to the maintenance of the adhesive and polarized phenotype of epithelial cells where it is primarily expressed. The aim of the present work was to study the expression profiles of E- cadherin transcription regulators Snail and Twist in human colorectal adenomas as a precursor lesion of colorectal cancer and to correlate the expression patterns with E-cadherin expression. Methods: A continuous series of formalin-fixed and paraffin embedded colorectal adenomas from our institution (n=42) was analysed. We also examined normal colorectal tissue (n=10). After RNA extraction the expression of E-Cadherin and its transcription regulators were examined by RT-PCR analysis. E-cadherin and Snail expression were validated by immunohistochemical analysis. The results of the RT-PCR analysis were statistically evaluated by Mann-Whitney-U-test. Results: In comparison to normal colon tissue we found a significantly down-regulated E-cadherin expression in colorectal adenoma (n=42). In 7% (n=3/42) of the adenomas we could not observe E-cadherin expression. While expression of the E-cadherin transcription regulator Snail and Twist were not detectable in normal colon tissue, we found Snail expression in 76% (n=32/42), Twist expression in 40% (n=17/ 42) of the adenomas. A correlation between Snail and Twist expression and a significant E-cadherin down-regulation was observed. The co-expression of Snail and Twist caused the most effective E-Cadherin down-regulation. Conclusions: Snail and Twist expression already exist in colorectal adenomas and a negative effect of the transcription regulators Snail and Twist on E-cadherin expression was shown. Important EMT regulators are already activated in precursor lesions of colorectal cancer and have a potential influence on malignant progression. They may reflect important markers for prognostic assessment in colorectal adenomas.

61 PREVALENCE OF BRCA2 AND CDKN2A MUTATIONS IN FAMILIAL PANCREATIC CANCER FAMILIES Emily P. Slater, Mercedes Sina, Peter Langer, Volker Fendrich, Nils Habbe, Brunhilde Chaloupka, Elvira Matthäi, Stephan A. Hahn*, Matthias Rothmund, Detlef K. Bartsch Department of Surgery, Philipps-University, Marburg, Germany; *Department of Internal Medicine, Universitätsklinik Knappschaftskrankenhaus, Bochum, Germany Keywords: pancreatic cancer, BRCA2, CDKN2a Background: Previous small scale studies reported BRCA2 germline mutations in about 15% of familial pancreatic cancer families. CDKN2a mutations were reported in up to 50% of FAMMM-PC families. As the genetic basis of familial pancreatic cancer is still

Langenbecks Arch Surg (2008) 393:759–815 widely unknown, we evaluated the prevalence of BRCA2 and CDKN2a germline mutations in a large cohort of familial pancreatic cancer (FPC) families. Methods: 70 families with at least two first-degree relatives with confirmed pancreatic cancer of the German National Case Collection for Familial Pancreatic Cancer of the Deutsche Krebshilfe (FaPaCa), that did not fulfill the criteria of other tumor predisposition syndromes, were analyzed for BRCA2 and CDKN2a germline mutations by DHPLC and/or direct sequencing. Results: Mutation analysis identified relevant BRCA2 mutations in 3 families (4%) and unclassified variants in 6 families (8%), of which one (p.R2034) potentially cosegregated with the disease. In addition, 67 families (96%) revealed BRCA2 polymorphisms. No relevant CDKN2a mutations were detected in the 70 FPC families. Conclusions: The prevalence of BRCA2 mutations in FPC families is much lower than previously described (4% vs. 15%). CDKN2a mutations do not predispose to FPC. Thus, CDKN2a mutation testing is not indicated in FPC families and BRCA2 mutation analysis should only be performed in the setting of controlled study protocols.

62 CLINICAL AND GENETIC ANALYSIS OF PANCREATIC CARCINOMA-MELANOMA PRONE FAMILIES Emily P. Slater, Mercedes Sina, Peter Langer, Nils Habbe, Elvira Matthäi, Brunhilde Chaloupka, Detlef K. Bartsch Department of Surgery, Philipps-University, Marburg, Germany Keywords: pancreatic cancer, melanoma, FAMMM, CDKN2a, BRCA2 Background: The familial atypical multiple mole melanoma (FAMMM) syndrome may predispose affected families to nonmelanoma carcinomas, predominantly to adenocarcinoma of the pancreas. About 25% of these families with an occurrence of pancreatic carcinoma (FAMMM-PC) harbor germline mutations in the CDKN2a gene. However, the phenotypic and genotypic expression of families with an association of pancreatic cancer and melanoma has not been fully characterized. Methods: The National Case Collection of Familial Pancreatic Cancer of the Deutsche Krebshilfe (FaPaCa) comprises 110 pancreatic cancer families of which 15 (14%) show an association of pancreatic carcinoma and melanoma. After genetic counseling, these 15 families were thoroughly analysed regarding their phenotype and the prevalence of germline mutations in the candidate genes CDKN2a, BRCA2, NOD2, ARLTS1 and Palladin, respectively. Results: Only 5 of the 15 families revealed the FAMMM phenotype (>50 atypical cutaneous nevi). PC was the predominant tumor type in 6 families, whereas melanoma was in none. Early age onset of either pancreatic carcinoma (<55 years) or melanoma (<35 years) occurred in 3 and 5 families, respectively. There were no incidences of melanoma and pancreatic cancer as double primaries in the same patient. The most frequent associated other tumor type was breast cancer in 5 families (33%). CDKN2a mutations were identified in 2 families (13%) and a BRCA2 mutation in 1 family (7%), all not exhibiting the FAMMM phenotype. No relevant germline mutations were detected in NOD2, Palladin, ARLTS1 and CHEK2 genes. Conclusions: Families with an accumulation of PC and melanoma show a large variety of phenotypic expression, which is often not consistent with the proposed FAMMM-PC phenotype. Only a subset of these families carry predisposing CDKN2a mutations, whereas other families might be linked to BRCA2 or other yet unidentified genes. This proposes the possibility of a distinct hereditary cancer syndrome, which might be classified as the “Pancreatic CarcinomaMelanoma Syndrome (PCMS)”.

783

63 OVEREXPRESSION OF THE GENE IFIT3 ENHANCE TUMOR GROWTH, ANGIOGENESIS, METASTASING AND CHEMORESISTANCE OF THE PANCREAS CARCINOMA CELLS Peter Camaj, Ivan A. Ischenko, Hendrik Seeliger, Georg Arnold, KarlWalter Jauch, Christiane J. Bruns Experimentelle Forschung: Chirurgie, Klinikum Großhadern, Marchioninistr. 15, 81377 München, Laboratory for Functional Genome Analysis, Feodor-Lynen-Str. 25, 81377 München Keywords: IFIT3, angiogenesis, chemoresistance, apoptosis, NFkB, IFNAR Background: Pancreas carcinoma belongs to one of the most dangerous tumor disease. Therefore it is very important to understand mechanism leading to change of the low angiogenic and low metastatic tumors to its aggressive counterparts. To address this issue, the experimental model consisting of the low metastatic pancreas carcinoma cell line FG and its closely related highly metastatic and proangiogenic cell line L3.6pl has been used. Transcriptome analysis has revealed that expression of gene IFIT3 as well as Interferon /β receptor has been upregulated in aggressive cell line L3.6pl. Here is the first time described upregulation of this gene in the tumor. Methods: Expression study: real time RT-PCR or western blotting (antiIFIT3 Ab) using material from FG and L3.6pl cells. Cloning of IFIT3 gene: expression vector pcDNA3.2 V5, Animal experiment: Orthotopic injection of nude mice with FG vector control and with FG cells stable transfected with IFIT3 plasmid. IHC: staining with anti-CD31 antibody. Effect on apoptosis and chemoresistance: Apoptosis assay via propidium iodide staining of both transfected cell lines untreated or treated with different concentration of the antitumor chemotherapeutics. VEGF production of transfectants: ELISA assay of conditioned media. Regulation of IFIT3 transcription: IFIT3-real time RT-PCR upon treatment with NFκB-inhibitor BAY 11-7082. Results: We have shown that highly metastatic and proangiogenic cell line L3.6pl overexpress IFIT3 on the level of RT-PCR, western blotting and IHC as well. Overexpression of IFIT3 in FG cells leads to increased production of VEGF in vitro determined by ELISA. Overexpresssion of this gene leads to inhibition of apoptosis by untreated cells and to increase of chemoresistance by the cells treated with gemcitabine, 5FU and irrinotecan as well. Tumors initiated by implantation of the FG transfected with IFIT3-plasmid grow significantly faster then tumors obtained by implantation of FG empty vector control cells. Cells transfected with IFIT3 exhibited increased metastatic potential. CD31 staining showed increased microvessels density in tumors derived from FG IFIT3 cells. Treatment with NFκB-inhibitor BAY 11-7082 diminish amount of IFIT3-transcript. Conclusions: Overexpression of IFn regulated gene IFIT3 leads to increased tumor growth, metastasing, angiogenesis and chemoresistance. This gene is probably regulated via NFκB signalling. Presented results could contribute to understanding of the mechanism of the crosstalk between inflammation and pancreas carcinogenesis.

64 IDENTIFICATION OF A MSI-H TUMOR SPECIFIC CYTOTOXIC T CELL EPITOPE GENERATED BY THE (1) FRAME OF U79260(FT0) Michael Linnebacher, Anne Wienck, Mathias Krohn, Inga Boeck, Ernst Klar, Sven Eisold Sektion Molekulare Onkologie und Immunotherapie, Abteilung für Allgemeinchirurgie, Universität Rostock, Schillingallee 35, 18057 Rostock, Germany

784 Keywords: MSI-H, tumor antigen, immunotherapy Background: Colorectal cancer is caused by genetic instability of somatic cells. Two different molecular pathways generate either chromosomal or high level microsatellite instability (MSI-H). The latter is induced by defects of the DNA mismatch repair system and results in insertion or deletion of single nucleotides at short repetitive DNA sequences. About 15% of sporadic and more than 90% of hereditary non-polyposis colorectal cancers display MSI-H. Instability of microsatellites in coding regions results in frameshift mutations preceding expression of C-terminally truncated proteins with attached neo-peptide tails (frameshift peptides; FSPs). Several genes known to play important roles in colorectal carcinogenesis as for example transforming growth factor beta receptor type II (TGFbetaRII), Bax, AIM2 and TCF-4 are targeted by this mutational mechanism, leading to the hypothesis, that these truncated gene products have an impaired function and might thereby contribute to MSI-driven carcinogenesis. On the other hand, it is likely, that immune reactions against FSPs are involved in the immune surveillance of MSI-H cancers. Methods: To analyze the immunogenic potential of the HLAA0201-restricted peptide FSP11 derived from a (−1) mutation of a T(15) tract within the U79260(FTO) gene, we generated a T cell bulk culture from a HLA-0201(pos) donor. FSP11 was able to stabilize HLA-A0201 molecules in a T2 stabilization assay, thus confirming HLA-A0201 binding properties. As antigen presenting cells, we used CD40-activated B cells (CD40 Bs) from the autologous blood donor taking advantage of the CD40L culture system. These CD40 Bs were loaded with FSP11 and used for repetitive stimulation of autologous T cells. After 28 days in culture the number of T cells had increased 5.6fold. Stimulation was continued for more than 100 days and resulted in a T cell expansion exceeding 1000fold. Results: In the present study we provide conclusive data that the (−1) frame of the MSI-H target gene U79260(FTO) encodes an HLAA0201-restricted cytotoxic T cell epitope (FSP11; TLSPGWSAV). T cells specific for FSP11 efficiently recognized HLA-A0201+ tumor cells harboring the mutated reading frame. Conclusions: These FSPs are highly specific for MSI-H tumors and consequently may be useful in specific antitumoral immunotherapy. However, for creation of successful vaccines for potential treatment of MSI-H tumors it seems reasonable to combine several epitopes derived from different antigens. Considering the exceptionally high mutation rate of U79260(FTO) in MSI-H colorectal carcinoma (81.8%) these data strongly recommend FSP11 as a component for vaccines against MSI-H cancers.

65 EXPRESSION OF INDOLEAMINE 2,3-DIOXYGENASE IN TUMOR TISSUE OF GASTRIC CANCER PATIENTS IS ASSOCIATED WITH LONG-TERM SURVIVAL Natasja K. van den Engel, Dominik Rüttinger, Nina Schupp, Georgios Meimarakis, Claudia Fernsebner, Christoph Fischer, Karl-Walter Jauch, Rudolf A. Hatz, Hauke Winter Department of Surgery, Grosshadern Medical Center, University of Munich, Germany Keywords: IDO, Helicobacter pylori, gastric cancer Background: The inflammatory enzyme indoleamine 2,3-dioxygenase (IDO) participates in immune tolerance and tumor immune escape processes by degradation of the essential amino acid tryptophan and formation of toxic catabolites. Helicobacter pylori (H. pylori) infection was shown to be an independent positive prognostic factor for patients’ survival. Here, we analyzed the role of IDO in tumor growth and survival of H. pylori + and H. pylori — patients with gastric cancer.

Langenbecks Arch Surg (2008) 393:759–815 Methods: Expression of IDO mRNA was analyzed by quantitative reverse transcription-PCR in 41 gastric cancer patients (18 H. pylori -, 23 H. pylori +). Immunohistochemistry analyses were used to semiquantitatively determine IDO in a subset of tumor samples, and gastric cancer cell lines. The preoperative H. pylori status of the patients was determined by bacterial culture, histology (H&E and Warthin-Starry stain) and serology. Results: IDO expression was significantly higher in the tumor than in the tumor-free mucosa as determined by quantitative PCR. Patients expressing high amounts of IDO in the tumor tissue had a better survival than patients expressing low amounts. Furthermore H. pylori + patients had a significantly better survival than H. pylori — gastric cancer patients. IDO was constitutively expressed by 4 out of 6 tumor cell-lines generated from the patients’ tumors. IFN-gamma further induced the expression of IDO in all cell lines in vitro, whereas a coculture of the cell lines with H. pylori had no influence on the expression of IDO. Conclusions: In contrast to other reports that support a correlation between IDO overexpression and decreased survival, we found that higher expression levels of IDO were associated with better survival. We hypothesize that IDO is induced in the tumor of H. pylori + patients by IFN-gamma secreting T-cells. This might induce the generation of tumor-toxic metabolites, thus restricting tumor growth and contributing to survival.

66 POTENTIAL BIOMARKERS FOR COLORECTAL LIVER METASTASIS Vilma Oliveira Frick, Claudia Rubie, Mathias Wagner, Martin K. Schilling Department of General, Visceral, Vascular and Paediatric Surgery, University of the Saarland, 66421 Homburg/Saar, Germany, Institute of Pathology, University of the Saarland, 66421 Homburg/Saar, Germany Keywords: chemokines, hepatocellular carcinoma, colorectal liver metastasis, gene expression Background: Like hepatocellular carcinoma (HCC), which is a highly malignant tumor with a poor prognosis due to its rapidly progressing and infiltrating growth, also colorectal cancer (CRC) still remains one of the leading causes of cancer-related death worldwide. The mortality of CRC is principally attributable to the development of metastases which primarily infest the liver. Colorectal liver metastases (CRLM) are present in up to 95% of patients in the advanced disease stage. In the last few years, chemokines have been shown to participate in tumor growth and the lymphatic and even distant spread of malignant tumors. Therefore, chemokines represent important targets in the battle against hepatic metastases formation. The aim of the present study was to evaluate and compare the expression profiles of CXCL12 (SDF-1), CCL19 (MIP-3ß), CCL20 (MIP-3α) and CCL21 (6Ckine) and their receptors on RNA and protein level in HCC versus CRLM and to elucidate their impact on the carcinogenesis and progression of malignant liver diseases. Methods: Chemokine expression was analyzed by RT-PCR and ELISA in 11 cases of HCC specimens and in 23 cases of CRLM and corresponding adjacent non-tumorous liver tissues, respectively. Expressions of their receptors CXCR4, CCR6 and CCR7 were analyzed by RTPCR and Western blot analysis in the same cases of HCC and CRLM. Results: Significant up-regulation for CCL20/CCR6 was detected in both cancer types. Moreover, CCL20 demonstrated significant overexpression in CRLM in relation to the HCC tissues. Being significantly up-regulated only in CRLM, CXCR4 displayed an aberrant expression pattern with respect to the HCC tissues. Conclusions: Correlation of CXCR4 expression with CRLM suggests CXCR4 as a potential predictive factor for CRLM. High level

Langenbecks Arch Surg (2008) 393:759–815 expression of CCL20 and its receptor CCR6 in HCC and CRLM with marked up-regulation of CCL20 in CRLM in relation to HCC tissues indicates involvement of the CCL20/CCR6 ligand-receptor pair in the carcinogenesis of hepatic malignancies and particularly in the progression of colorectal metastases of the liver.

67 INVOLVEMENT OF CHEMOKINE RECEPTOR CCR6 IN COLORECTAL CANCER METASTASIS Claudia Rubie, Vilma Oliveira Frick, Mathias Wagner, Martin K. Schilling Dept. of General -, Vascular - and Paediatric Surgery University of the Saarland, 66421 Homburg/Saar, Germany, Institute of Pathology, University of the Saarland, 66421 Homburg/Saar, Germany Keywords: chemokine receptor CCR6, colorectal cancer, gene expression Background: Various chemokine receptors namely CXCR4, CCR6 and CCR7 have recently been shown to be involved in the regulation of metastasis in malignant tumors. However, little is known about the role of these receptors in promoting tumor metastasation in colorectal cancer to the primary site of colorectal cancer metastasis in the liver. To investigate this issue, we analysed the expression of the chemokine receptors CXCR4, CCR6 and CCR7 and their respective ligands CXCL12/ SDF1α, CCL20/MIP3α, CCL19/MIP3ß and CCL21/6Ckine in colorectal tumors and colorectal liver metastases. Methods: In the present study, 30 human cancer samples from colorectal tissue, 30 human samples from colorectal liver metastases and the corresponding normal tissues were screened using quantitative real-time PCR (Q-RT-PCR), Western blot analysis, immunohistochemistry (IHC) and the enzyme-linked immunosorbent assay (ELISA). Results: Among the chemokine receptors CXCR4, CCR6 and CCR7 only CCR6 was significantly overexpressed in both malignant colorectal tumors and colorectal liver metastases. Consequently, we investigated the expression of the corresponding CCR6 receptor ligand CCL20/MIP3α in different organs such as stomach, oesophagus, pancreas, colon and rectum in comparison to its expression in the liver. Compared with these organs we found that CCL20 exhibits peak levels of expression in the liver thus indicating that an increased production of CCL20 may contribute to the selective recruitment of CCR6-expressing cells in colorectal cancer. Furthermore, we demonstrated that colorectal cancer patients who developed liver metastases express significantly more CCL20 in the liver in comparison to an unaffected control group. Conclusions: Therefore, our findings strongly indicate an association between CCL20/CCR6 expression in human colorectal cancer and the promotion of colorectal liver metastasis.

68 IL-8 EXPRESSION CORRELATES WITH INDUCTION, PROGRESSION AND METASTATIC POTENTIAL OF COLORECTAL CANCER Claudia Rubie, Vilma Oliveira Frick, Mathias Wagner, Martin K. Schilling Dept. of General -, Visceral-, Vascular - and Paediatric Surgery, University of the Saarland, 66421 Homburg/Saar, Germany, Institute of Pathology, University of the Saarland, 66421 Homburg/Saar, Germany Keywords: colorectal benign and malign diseases, IL8 expression Background: In the present study we investigated the expression profile of Interleukin-8 (IL-8) in inflammatory and malignant

785 colorectal diseases to evaluate its potential role in the regulation of colorectal cancer (CRC) and the development of colorectal liver metastases (CRLM). Methods: IL-8 expression was assessed by quantitative real-time PCR (Q-RT-PCR) and the enzyme-linked immunosorbent assay (ELISA) in resection specimens from patients with ulcerative colitis (UC, n=6) colorectal adenomas (CRA, n=8), different stages of colorectal cancer (n=48) as well as synchronous and metachronous CRLM along with their corresponding primary colorectal tumors (n=16). Results: IL-8 mRNA and protein expression was significantly upregulated in all pathological colorectal entities investigated compared with the corresponding neighbor tissues. However, in the CRC specimens IL-8 revealed a significantly more pronounced overexpression in relation to the CRA and UC tissues with an average 30-fold IL8 protein up-regulation in the CRC specimens in comparison to the CRA tissues. Moreover, IL-8 expression revealed a close corelation with tumor grading. Most interestingly, IL-8 up-regulation was most enhanced in synchronous and metachronous CRLM, if compared with the corresponding primary CRC tissues. Herein, an up to 80-fold IL8 overexpression in individual metachronous metastases compared to normal tumor neighbor tissues was found. Conclusions: Our results strongly suggest an association between IL8 expression, induction and progression of colorectal carcinoma and the development of colorectal liver metastases. Thus, IL-8 may represent a potential target to prevent hepatic recurrence of CRC.

69 CHARACTERIZATION OF THE HUMAN PROGRAMMED CELL DEATH PROTEIN 4 (PDCD4) PROMOTER Jörg Hendrik Leupold, Irfan Asangani, Stefan Post and Heike Allgayer Dept. of Experimental Surgery and Molecular Oncology of Solid Tumors, Medical Faculty Mannheim, University Heidelberg and DKFZ (German Cancer Research Center) Heidelberg, Germany Keywords: apoptosis Background: The pdcd4-gene has originally been identified as a suppressor of malignant transformation and translation, getting upregulated during apoptosis. Recent data suggest Pdcd4 to be an important novel tumor suppressor, inhibiting carcinogenesis, tumor progression, invasion, and intravasation, and in a first study we implicated the loss of Pdcd4 as an independent prognostic marker for colorectal cancer. Pdcd4 expression is downregulated in several cancers, however, very little is known yet about essential molecular regulators leading to a loss of this tumor suppressor in tumor cells. For chicken Pdcd4, it has been shown that it is targeted by c-Myb and its oncogenic homolog v-Myb, and that the expression of Pdcd4 is regulated by topoisomerase inhibitors, COX-2 inhibitors, interleukins and retinoic acid. Moreover, we recently showed that Pdcd4 is negatively regulated at the post-transcriptional level by miR-21. However, the transcriptional regulation and the promoter of this important tumor suppressor still has not been characterized, and little is known about potential cis-regulatory elements and transcription factors targeting the human Pdcd4 promoter. Methods: Rapid amplification of cDNA-ends (RACE) was used to determine the transcriptional start site within the promoter. Referring to the start sites identified, a 3891 kb sequence upstream of the first exon was cloned into a luciferase reporter plasmid, and sequential 5′-deletions were generated and transfected into mammalian cell lines. To evaluate an appropriate induction for Pdcd4 expression, different cell lines (breast and colorectal cancer) were stimulated for different time points and with different amounts of PMA, EGF or ATRA. Next, in order to elucidate potential transcription factors bound to the human Pdcd4 promoter,

786 Electromobility Shift Assays (EMSA) and Supershifts were performed. Finally, site-directed mutagenesis was performed to distinguish the influence of individual cis-acting elements within the promoter. Results: RACE identified different transcriptional start sites located within a region between +56 bp and +150 bp in the first exon (the first basepair of this exon defined as +1). Transfection and reportergene analysis of serially deleted luciferase constructs showed a 2456 bp fragment having the highest activity, as well as an important basal promoter region spanning 548 base pairs upstream of the first exon, which lost activity completely when further deleted. More detailed investigations on these fragments revealed several putative binding sites for transcription factors including GATA, RAR/RXR and several Sp binding sites. EMSAs identified four Sp1/Sp3/Sp4 binding elements within the Pdcd4 basal promoter and a potential Sp1/2/4 and GATA site in the 2456 bp fragment showing maximum promoter activity. Supershift assays using Sp- and GATA-antibodies confirmed the binding of these transcription factors to their respective sites. Site-directed mutagenesis of the individual Sp-sites led to a reduced or enhanced activity, suggesting that all Sp-sites are functionally relevant for promoter activity. Additionally, stimulation with ATRA leads to an enhanced binding of a transcription factor to a potential RAR/RXR site in front of the first exon. These preliminary results on specific sites and transcription factors are currently completed by further and more detailed analyses. Conclusions: Taken together, this is the first study characterizing the human pdcd4-gene promoter, and implicating first promoter motifs important for regulation. Especially, Sp binding sites appear as regulators of constitutive pdcd4-gene expression in diverse colorectal and breast cancer cell lines. Furthermore, this study gives a first implication on a potential RAR/RXR binding site leading to enhanced expression of Pdcd4.

70 ZEB1 PROMOTES EMT AND INVASION OF CANCER CELLS BY INDUCING A MICRORNA MEDIATED POSITIVE FEEDBACK LOOP Ulrike Burk, Ulrich Wellner, Jörg Schubert, Otto Schmalhofer, Simone Spaderna, Thomas Brabletz Institution Dept. of Visceral Surgery, University of Freiburg, Germany Keywords: microRNA, ZEB1, EMT, invasion, feedback loop Background: The embryonic program `epithelial-mesenchymal transition’ (EMT) is considered to promote malignant tumor progression. The transcriptional repressor ZEB1 is a crucial inducer of EMT in different human tumors and was recently shown to promote invasion and metastasis of tumor cells. We here describe that ZEB1 directly suppresses transcription of two microRNAs, which strongly preserve epithelial differentiation in pancreatic, colorectal and breast cancer cells. Off note, ZEB1 itself and TGF2 are prime targets downregulated by these miRNAs. The results indicate that ZEB1 triggers a microRNA-mediated positive feedback-loop, which stabilizes EMT and promotes invasion of cancer cells. Methods: Total RNA was isolated from different cell lines and a miRNA expression screen was performed. Real time RT-PCR was used for quantification of mRNA and miRNA expression after transient transfection of cell lines with miRNAs or after transient knock down of ZEB1. Results: ZEB1 directly suppresses transcription of two microRNAs, which strongly preserve epithelial differentiation in pancreatic, colorectal and breast cancer cells. ZEB1 itself is a prime target downregulated by these miRNAs. The results indicate that ZEB1 triggers a microRNA-mediated positive feedback-loop, which stabilizes EMT and promotes invasion of cancer cells Conclusions: In conclusion we suggest that ZEB1 is a crucial promoter of tumor progression by both reducing transcription of mRNAs and miRNAs. Thereby ZEB1 is a central molecular regulator of a miRNA-mediated positive feedback loop, which stabilizes EMT and thus promotes malignant tumor progression

Langenbecks Arch Surg (2008) 393:759–815

71 THE GROWTH PATTERN OF RECURRENT GLIOBLASTOMA AFTER GLIADEL WAFER IMPLANTATION IN FIRST RECURRENCIES Patrick Weigel, Gabriele Schackert, Dietmar Krex Klinik und Poliklinik für Neurochirurgie, Universitätsklinikum Carl Gustav Carus, Dresden, Germany Keywords: recurrent glioma, carmustin wafer Background: Although in recent years, several approaches for local therapies in the treatment of glioblastoma multiforme have been tested in clinical trials, carmustin polymeres (Gliadel™ wafer) is the only evaluated local therapy to date. Because of the predominance of temozolomide, Gliadel™ is frequently used in recurrent gliomas only. However, data about effectivity and the pattern of re-recurrences are rare. Therefore, we initiated the present MRI-based retrospective study. Methods: 20 patients had surgery for first recurrence of glioblastoma, where Gliadel™ wafer were implanted. Early post-op MRI was performed documenting the extent of resection, tumor remnants and wafer placements. Follow-up MRI was performed every two months looking particularly for tumor growth in relation to the wafer placements. Progression-free and overall survivals were recorded. Results: 18 patients were available for evaluation, while 2 patients had incomplete data. In 16 (89%) patients an early tumor growth was recorded in the first MRI follow-up, 2 months post-op. If the wafers had been placed in areas with suspicious tumor according the early post-op MRI, tumor progression was recorded in the follow-up in 12 (75%) patients. If the wafers were placed in areas without suspicious tumor remnants, regrowth in those areas was recorded only in 9 (56%) patients. However, tumor pseudo-progression has to be taken into account for all cases. Survival data will be determined. Conclusions: The use of Gliadel™ wafer in recurrent glioblastoma is most effective in areas with no tumor remnants, underlining the meaning of surgical resection also of recurrent tumors.

72 PROTEIN-PROFILING OF DIPLOID AND ANEUPLOID COLORECTAL CANCER CELL LINES Timo Gemoll, Uwe Roblick, Susanne Becker, Britta Fritzsche, Franz Bader, Stefanie Bünger, Hans-Peter Bruch, Ulf Hellmann, Hans Jörnvall, Gert Auer, Jens Habermann Department of Surgery, University Hospital Schleswig-Holstein, Campus Lübeck, Germany; Karolinska Biomic Center, Karolinska Institutet, Stockholm, Sweden; Ludwig Institute of Cancer Research, Uppsala, Sweden; Department of Medical Biochemistry and Biophysics, Karolinska Institutet, Stockholm, Sweden Keywords: Protein-Profiling, genomic instability, aneuploidy, 2-de, image cytometrie, colorectal cancer Background: DNA aneuploidy has been identified as a prognostic factor in the majority of epithelial malignancies such as breast- and endometrial cancer. Neither cause nor consequences of DNA aneuploidy have been sufficiently understood yet. However, our understanding of the translation of aneuploidy into protein expression will impact on our ability to define successful therapies or even prevention. We therefore aimed at identifying differentially expressed genes that characterize aneuploid from diploid colorectal cell lines, in order to elucidate mechanisms of polyploidization and to screen for innovative biomarkers for early detection, prognosis and therapeutic intervention. Methods: We compared two-dimensional gel electrophoresis (2-DE)based protein expression pattern between diploid and aneuploid colorectal

Langenbecks Arch Surg (2008) 393:759–815 carcinoma cell lines. DNA ploidy was assessed by image cytometry. 2-DE results with differences in expression levels between diploid and aneuploid cell lines were compared and statistically evaluated. Mass spectromy served for identification of these differentially expressed protein spots. Results: Three diploid and four aneuploid colorectal tumor cell lines were defined by image cytometry and analyzed by two-dimensional gel electrophoresis. Two independent bioinformatic analysis revealed 38 significant (p-value<0.05) protein spots that were differentially expressed in between the diploid and aneuploid groups. These protein spots are now subsequently identified by mass spectrometry. Conclusions: This study showed significant differences in protein expression between diploid and aneuploid colorectal carcinoma cell lines. These targets are identified by mass spectrometry and will be further validated for their biomarker potential on clinical samples.

73 PROGNOSTIC IMPLICATIONS OF TUMOR-ASSOCIATED FACTORS IN RESECTABLE PANCREATIC ADENOCARCINOMA Martin Gasser, Martin Grimm, Meike Munt, Arnulf Thiede, Ana Maria Waaga-Gasser Department of Surgery I and Molecular Oncology and Immunology, University of Wuerzburg, Germany Keywords: pancreatic adenocarcinoma, tumor-associated factors Background: Reproducible prognostic markers could potentially allow patients with pancreatic cancer to be stratified into treatment groups. In this study, we asked whether EGF, EGFR, erb B2, TGF-α and -, p53, and PCNA expression might be useful in detecting different aggressive pancreatic cancers and if survival would be worse in high marker expressors. Methods: In this retrospective study a total of 282 patients with histologically verified pancreatic ductal adenocarcinoma were treated at our Department from 1993 to 2003. In addition, 93 patients were selected based on the availability of tissue specimens from the primary tumor and R0-resection (75% head, 13% corpus, 12% caudal). We used Elisa, immunohistochemistry and real time PCR to investigate expression of the different tumor associated factors and compared with standard marker CEA and CA19-9 expression. The values of morphological and molecular parameters were correlated with clinicopathological characteristics for their predictive value of tumor recurrence using chi-square test, Kaplan-Meier method, multivariate analysis by Cox regression, and Mann-Whitney-U-Test (follow-up period: 32±6.2 months). Results: A total of 34% of the patients underwent R0-resection, which significantly influenced their survival (78% stage UICC IIb, p< 0.0001). Mean survival and time to progress were also influenced by N-stage (mean survival N1: 21.5 Mo vs. N0: 49.1 Mo) and tumor differentiation (G1: 63.7 Mo; G2: 26.3 Mo; G3: 15.2 Mo; p=0.009). The prognosis of patients undergoing curative R0-resection seems to be determined by perineural invasion (p=0.019) with a mean survival of 12 months (no invasion: 29 months). With exception of MIC-A all of the analyzed parameters showed an increased protein expression during tumor progression. A prognostic significance was observed for increased CD4+ and CD8+ T cell infiltration, and poor EGF, erb B2, and PCNA expression in the tumor (p<0.001). Conclusions: Our data suggest that patients, even if R0-resected and with no lymph node metastasis, may benefit from analysis of the prognostic markers EGF, erb B2, PCNA, and MIC-A in their tumors. Those nodal negative patients with increased marker expression profiles would potentially profit from chemotherapy combined with small molecules targeting intracellular signal transduction pathways.

787

74 CLINICAL SIGNIFICANCE AND THERAPEUTIC POTENTIAL OF PROGRAMMED DEATH 1 LIGAND-1 AND PROGRAMMED DEATH-1LIGAND-2 EXPRESSION IN HUMAN COLORECTAL CANCER Martin Gasser, Martin Grimm, Matthias Königshausen, Arnulf Thiede, Ana Maria Waaga-Gasser Department of Surgery I and Molecular Oncology and Immunology, University of Wuerzburg, Germany Keywords: Clinical Significance and Therapeutic Potential of Programmed Death 1 Ligand-1 (PD-L1), PD-L2, colorectal cancer Background: The negative regulatory programmed death-1/programmed death-1ligand (PD-1/PD-L) pathway in T-cell activation has been suggested to play an important role in tumor evasion from host immunity. Levels of immune cells expressing PD-1 in clinical colorectal carcinoma (CRC) tumors have not been evaluated. Thus, we investigated whether PD-1 positive T cells were expressed within CRC tumors and the expression of PD-L1 and PD-L2 in human CRC to define their clinical significance in patients’ prognosis after surgery. Methods: Tissue samples from 116 patients operated between 2001 and 2003 were collected in our institution and histologically confirmed CRC were evaluated retrospectively for this study. PD-L1 and PD-L2 gene expression was evaluated by real time quantitative PCR and the samples were immunostained and anti-PD-1 Outcome analyses were performed. Results: The protein and the mRNA levels of determination by immunohistochemistry and real time quantitative PCR were closely correlated. PD-L expression was inversely correlated with tumorinfiltrating T lymphocytes, particularly CD8+ T cells. T cell infiltration was observed in 105 (90.5%) specimens. 93 (80.2%) specimens were PD-1+ T cells. Intratumoral PD-1+ T cells were associated with advanced tumor stage (p=0.002). Patients with PD-1+ T cells had significantly more PD-L1 tumor cell expression. Multivariate analysis indicated that PD-L positive patients and those with PD-1+ T cells had a significantly poorer prognosis than the negative patients. This was more pronounced in the advanced stage of tumor than in the early stage. Conclusions: These data suggest that interactions of T cells expressing PD-1 and PD-L may promote cancer progression. PD-L1 and PD-L2 status may be a new predictor of prognosis for patients with CRC.

75 LEVO- BUT NOT DEXTRO-1-METHYL TRYPTOPHAN ABROGATES THE IDO ACTIVITY OF HUMAN TUMOR- AND DENDRITIC CELLS Stefan Löb, Peter Terness, Derek Zieker, Richard Schäfer, Björn Brücher, Hans-Georg Rammensee, Alfred Königsrainer Department of General, Visceral and Transplant Surgery, University Hospital of Tübingen, Hoppe-Seyler-Straße 3, 72076 Tübingen, Germany; Institute of Clinical and Experimental Transfusion Medicine, Eberhard Karls University of Tübingen, Otfried-Muller-Straße 4/ 1, 72076 Tübingen, Germany; Institute for Cell Biology, Department of Immunology, Eberhard Karls University of Tübingen, Auf der Morgenstelle 15, 72076 Tübingen, Germany; Institute of Immunology, Department of Transplantation Immunology, University of Heidelberg, INF-305, 69120 Heidelberg, Germany Keywords: IDO, 1-MT, human tumors, dendritic cells, cancer immunotherapy Background: Clinical phase-I-trials with cancer patients have been started with the aim of inducing tumor immunity by blocking the immunosuppressive action of indoleamine-2,3-dioxygenase (IDO) with the IDO2-inhibitor dextro-1-methyl-tryptophan (D-1MT). In this study we analyzed IDO2- and IDO1-expression in human dendritic cells (DCs)

788 and tumor cells and further investigated the impact of their respective inhibitors, D-1MT and L-1MT on the abrogation of IDO-activity. Methods: Freshly isolated human monocytes were differentiated into DCs. Established tumor cells lines, surgically removed human primary gastric-, colon- and renal cell carcinoma specimens and DCs were analyzed for IDO1- and IDO2-expression by RT-PCR. siRNA treatment specific for IDO1 was evaluated by qRT-PCR. IDO-activity was measured and quantified by RP-HPLC. Results: Human primary tumors investigated in this study constitutively expressed various amounts of IDO1 and IDO2, whereas tumor cell lines and DCs needed to be induced by Interfon-gamma (IFNγ). Transfection with IDO1-specific siRNA blocked IDO1transcription in cell lines and DCs. IDO2-expression remained unaffected. Tryptophan breakdown was completely restored upon IDO1-siRNA transfection of tumor cell lines and DCs. Addition of L1MT, but not D-1MT abrogated IDO-activity of tumor cell lines, DCs and freshly isolated colon carcinoma tumor. Conclusions: The present results lead us to conclude that although tumor cells and DCs express IDO2, IDO2 is produced in a functionally inactive form. Most importantly, D-1MT does not block the IDO-activity of tumor cells. Therefore, D-1MT-based clinical trials aiming at abrogating the immuno-suppressive effect of IDO are based on incorrect assumptions and L-1MT is the appropriate candidate for inhibiting IDO.

76 PREOPERATIVE FDG-PET-CT CORRELATES WITH INTRAOPERATIVE FINDINGS IN PATIENTS WITH PERITONEAL CARCINOMATOSIS Ingmar Königsrainer (1), Derek Zieker (1), Philip Aschoff (2), Stefan Beckert (1), Jörg Glatzle (1), Jörg T. Hartmann (3), Christina Pfannenberg (2), Falko Fend (4) Björn Brücher (1), Alfred Königsrainer (1). University of Tübingen; 1: General, Visceral- and Transplant Surgery, 2: Radiology, 3: Oncology, 4: Pathology Keywords: PET CT, peritoneal carcinomatosis Background: Peritoneal carcinomatosis until recently were considered incurable. Interdisciplinary multimodal therapy including extensive cytoreductive surgery followed by hyperthermic intraperitoneal chemotherapy (HIPEC) offers significant survival benefit. So far, no radiomorphological imaging can predict the intraoperative tumor load. Combined FDG-PET-CT-scan might offer an impact on patient selection, because its findings might influence making the decision for peritonectomy and HIPEC. We evaluated the results of preoperative FDG-PET-CTscan by correlating the findings of metabolic imaging with the intraoperative findings of patients who underwent peritonectomy and HIPEC. Methods: Eleven consecutive patients (mean age: 61,45years) underwent preoperative FDG-PET-scans, followed by cytoreductive surgery and HIPEC. The intraperitoneal tumor load was assessed preoperatively by FDG-PET-CT-scan according to the Peritoneal-Carcinomatosis-Index (PCI-PET-CT) according to the intraoperative PCI reported by Sugarbaker (PCI-Sugarbaker). Both indices (PCI-PETCT and PCI-Sugarbaker) underwent correlation analysis. Moreover we divided the abdominal quadrants into 4 regions (1: upper abdomen, 2: middle abdomen, 3: lower abdomen-pelvis, 4: intestine including jejunum and ileum) and correlated again the different regions. Results: The PCI-PET-CT (mean: 20,54+/−12,18) was statistically significant correlated with the PCI-Sugarbaker (mean: 21,09+/−12,8) (r2=0.929; p<0.0001). In addition, this correlation could be found in respect to the 4 PCI subgroups (upper abdomen: r2=0.765; p=0.006; middle abdomen: r2=0.779; p=0.005; lower abdomen-pelvis: r2= 0.746; p=0.008 and intestine: r2=0.862; p=0.001). Conclusions: The PCI-FDG-PET-CT correlates well with the intraoperative Peritoneal-Carcinomatosis-Index by Sugarbaker. Further

Langenbecks Arch Surg (2008) 393:759–815 studies are warranted investigating the potential impact of FDGPET-CT on preoperative patient selection for peritonectomy and HIPEC in patients suffering from peritoneal carcinomatosis.

77 LOW-DOSE SURAMIN IN COMBINATION WITH DOCETAXEL OR GEMCITABINE IN PANCREATIC CANCER Elisabeth Schellhaas, Birgit Hotz; Heinz J. Buhr; Hubert G. Hotz Chirurgische Klinik I; Charité Campus Benjamin Franklin; Hindenburgdamm 30, 12200 Berlin Keywords: pancreatic cancer; low-dose suramin; combination therapy; gemcitabine; docetaxel Background: Suramin, a naphthyl urea derivative, inhibits growth of several tumours, among them ductal pancreatic adenocarcinoma. Its use as a single therapeutic agent is problematic due to its toxicity. The addition of low dose suramin (in sub-toxic doses) to standard cytotoxic therapy was shown to enhance efficacy. Gemcitabine is considered to be the gold standard of therapy in pancreatic cancer. The aim of this study was to test combination therapies with low-dose suramin in pancreatic cancer. Methods: Investigations were performed using DSL6A, a rat ductal pancreatic adenocarcinoma cell line. Cells were incubated for 72 hours in the presence of increasing concentrations of suramin (0–200 µg/ mL) or docetaxel (0–3 µg/mL) or gemcitabine (0–100 µmol/L) or a combination of suramin with either docetaxel or gemcitabine. Cell proliferation and viability were assessed using a hemocytometer for direct cell counting and an MTT assay. Results: Docetaxel, gemcitabine, and suramin dose-dependently reduced cell proliferation and viability (p<0.01). The highest concentrations used as monotherapy inhibited cell growth by 74% (suramin) and 95% (docetaxel and gemcitabine), respectively. A combination of suramin with either cytotoxic agent was more effective than each of the single agents given alone. This was especially true for very low dose docetaxel and gemcitabine, whereas the synergetic effect was less pronounced with high dose cytostatics. A combination of suramin 10 µg/mL and docetaxel 0.003 µg/mL reduced cell proliferation by more than 50%. Conclusions: The addition of low-dose suramin enhances docetaxel as well as gemcitabine efficacy in vitro. Results by other study groups showed a higher IC50 for a combination of docetaxel and low-dose suramin in non-small-cell lung cancer cells than that found in our study in DSL6A pancreatic cancer cells. Therefore, a combination therapy of low-dose suramin and either docetaxel or gemcitabine seems to be able to improve efficacy as compared to standard cytotoxic therapy. Further studies using an orthotopic tumour model of rat ductal pancreatic adenocarcinoma are necessary to confirm whether these combinations are of relevance in a preclinical therapeutic setting. As suramin has already extensively been used in humans suffering from sleeping sickness, there are data on side effects of this drug. So it seems reasonable to assume that low-dose suramin in addition to standard cytotoxic tumour therapy will quickly become available for routine use once its efficacy has been proven.

78 CANCER STEM CELL TARGETED THERAPY WITH RAPAMYCIN AND CYCLOPAMINE REVERTS CHEMORESISTANCE TOWARDS 5-FLUOROURACIL IN HUMAN PANCREATIC CARCINOMA CELLS Ivan Ischenko, Hendrik Seeliger, Peter Camaj, Karl-Walter Jauch, Christiane J. Bruns Chirurgische Klinik und Poliklinik der Ludwig-Maximilians-Universität München Großhadern

Langenbecks Arch Surg (2008) 393:759–815 Keywords: human pancreatic cancer, cancer stem cells, chemotherapy resistance, cancer stem cell targeted therapy Background: Recently, cancer stem cells have been implicated in pancreatic tumor growth and metastatic activity, but the specific role of these cells in tumor chemoresistance mechanisms is still uncertain. In the present study, we aimed to identify and to eliminate putative cancer stem cells (CSC) within different human pancreatic cancer cell lines as well as within 5-Fluorouracil-resistant pancreatic cancer cell lines established in our laboratory. Methods: Pancreatic CSC were identified by flow cytometry, using the following antibodies: anti-CD133/1 (clone AC133), anti-CD133/2 (clone 293C3), anti-CD44, anti-CD34, all individually or in combination. The relative expression of stem cell markers was confirmed by Western blotting. The effects of cancer stem cell targeted therapeutics (Rapamycin, Cyclopamine) on tumor cell proliferation and apoptosis were assessed by MTT and propidium iodide staining, respectively. Results: All human pancreatic cancer cell lines as well as their 5-FUresistant sublines were different in the CSC content. Interestingly, flow cytometry and Western blotting analyses revealed a significantly high amount of CSC in all chemotherapy-resistant cell lines tested. The combined use of 5-FU with Rapamycin or Cyclopamine induced apoptosis and efficiently killed pancreatic cancer cells with the acquired 5-FU-resistance properties. Conclusions: Our results demonstrate that chemotherapy-resistant cancer cells contain the increased CSC population (as compared to their parental sensitive cells) that is highly resistant to standard chemotherapy but not towards CSC-targeted therapy. The further characterization of such cells might therefore lead to the development of new molecular and pharmaceutical therapeutics and better anti-cancer strategies.

79 INHIBITION OF THE TUMOR GROWTH OF COLON CARCINOMA BY GEFITINIB, A SELECTIVE TYROSINKINASE INHIBITOR OF THE EGFR IN VITRO AND IN A LAPAROSCOPIC ANIMAL MODEL Roger Kuhn (1), Daniel Schubert (1), Linda Flohr (1), Regine Schneider-Stock (2), Sabine. Krüger (2), Dörthe Küster (2),Walter Halangk (1), Hans Lippert (1), Matthias Pross (1) 1: Klinik für Allgemein-, Viszeral- und Gefäßchirurgie, Ottovon-Guericke-Universität Magdeburg, 2: Institut für Pathologie Ottovon-Guericke-Universität Magdeburg, Germany Keywords: : geftinib — iressa -ZD1839- colon-adenocarcinoma cell — tumor growth — laparoscopy — rats Background: EGFR is involved in epithelial cancers and is being considered as a potential target for antineoplastic therapy. We investigated cytotoxicity, anti-adhesive and anti-invasive effects of the EGFR tyrosine kinase inhibitor Gefitinib in-vitro using colon-adenocarcinoma cells CC531. Methods: In-vitro: Adhesion and cytotoxicity assay were performed in cell line CC531 with microtiter plates. Transwell dual chambers were used for the invasion assay. CEA and EGFR-expression was measured by fluocytometry. In-vivo: CC531 adenocarcinoma cells were intraperitoneally applied to WAG rats. 90 animals were randomized and received gefitinib in different dosages intraperitoneally at time of tumorimplantation or daily and in some gefitinib was administered orally daily in two different dosages. Controls were applicated with saline or polysorb. After 21 days, the animals were euthanized. Results: In-vitro: We found dose dependent effect of gefitinib in the cytotoxicity assay. A significant inhibition of tumor cell adhesion (p< 0.05) by all gefitinib concentrations but only with higher gefitinib concentrations inhibition of cell invasion was possible. In-vivo: Nearly complete tumor inhibition was achieved by daily intraperitoneal application of Gefitinib (p<0,0001). Geftinib only applied single shot at time of tumor cell implantation significantly decreased the intraperitoneal

789 tumor weight compared to controls (p<0.001). Oral administration of gefitinib reduce tumor growth dose dependent (p<0,01). Conclusions: Gefitinib given direct intraperitonally inhibits tumor growth effective. Regular intraperitoneal application diminishes intraperitoneal tumor growth better than single intraperitoneal or recurrent peroral in rats undergoing laparoscopy. Gefitinib was at this time not evaluated with daily intraperitoneal application. This may offer additional therapeutic options for colorectal cancer and new perspectives for Gefitinib therapies.

80 THE MESOPANCREAS IS THE PRIMARY SITE FOR POSITIVE RESECTION MARGINS AFTER PANCREATIC HEAD RESECTION Jochen Gaedcke, Bastian Gunawan, Marian Grade, Réka Szöke, Torsten Liersch, Heinz Becker, Michael Ghadimi Department of General- and Visceral Surgery, University Medical Center Göttingen, Germany Keywords: pancreatic cancer, mesopancreas, histopathological work up, pancreatic head resection Background: The prognosis of patients with pancreatic cancer remains poor, even after potentially curative R0 resection. This discrepancy may be due to the histopathological misclassification of R1 cases as curative resections (R0) in the past. Methods: To test this hypothesis, color coding of all resection margins and organ surfaces as part of a standardized histopathological work up was implemented. This modified system was prospectively introduced into the clinical routine. Results: 84 pancreatic head specimens were processed using this modified protocol. 19 patients were excluded from the analysis, owing to a nonmalignant or borderline tumor, lymphoma, or duodenal carcinoma. Applying the UICC criteria, 33 cancer resections were classified R0 (50.8%), while 32 cases turned out to be R1 resections (49.2%). Interestingly, the mesopancreas was infiltrated in 21 of the 32 R1 resection specimens (65.6%). It proved to be the only site of tumor infiltration in 19 specimens (59.4%). Applying the Royal College of Pathologists’ criteria, 42 resections were classified R1 (64.6%). As expected, the total number of infiltrated margins increased, with the mesopancreas again being the most frequent site for non-curative resection (n=26; 61.9%). Conclusions: We confirmed that the application of an intensified histopathological work up for pancreatic head cancer specimens results in an increased rate of R1 resections. Importantly, we demonstrated that the mesopancreas represents the primary site for positive resection margins. Such results have major implications for adjuvant and palliative chemotherapy trials.

81 A COMPARATIVE PROSPECTIVE STUDY OF THREE AVAILABLE COLD STORAGE SOLUTIONS IN HUMAN AND MURINE ADRENAL TUMOURS Alexandra Ozimek, Inga Johnsen, Urs Lichtenauer, Constanze Hantel, Felix Beuschlein, Klaus Hallfeldt, Thomas Mussack Department of Surgery, LMU University of Munich, 80 336 Munich, Germany; Department of Medicine/Endocrine Research,LMU University of Munich, 80 336 Munich, Germany Keywords: tumour storage, cold storage solutions, adrenocortical carcinoma, postoperativ tissue preservation Background: The poor overall survival time of patients with adrenocortical carcinoma is increasingly paving the way towards innovative in vitro and in vivo tumor models in order to provide individualized tumor therapy concepts in future. However, the quality of resected patient tissues can

790 dramatically influence the quality of results gained from molecular screening. Thus, maintaining tumour tissue viability during tissue preservation after surgical tumour resection is an important prerequisite for successful tumour research. We have performed a prospective study to compare the role of different cold-storage solutions on murine and human adrenal tumour tissue after surgical tumour resection. Methods: Murine adrenal tumours were induced by subcutaneous inoculation of 15×106 NCI-H295 cells. Tumours developing at the inoculation site were resected 6 weeks after tumour cell injection. Human tumour tissue was obtained from a patient undergoing adrenalectomy due to an adrenocortical carcinoma. All tumours were minced into small pieces and placed in three different cold storage solutions at 2–8°C for 72 h (GibcoTM CO2 Independent Medium-Invitrogen, Cool Star-PAA, University of Wisconsin Solution). In the two control groups, tissue pieces were stored in RPMI- Media 1640-Invitrogen at 2–8°C for 72 h or processed directly without cold storage. Primary tumour cell cultures were derived 72 h after tumour storage from all tissue pieces. Tumour cell suspensions were compared regarding obtained tumour cell number in relation to tumour volume, tumour cell viability (determined by FACS analysis) and tumour cell functionality (determined by analysis of unstimulated and stimulated cortisol production). Results: Regarding tumour cell number, tumour cell viability and tumour cell functionality there was no significant difference neither between the murine cold storage solution groups nor between these groups and the control groups. In the human cold storage groups tumour cell viability was significantly lower in the UWS group in comparison with the GIBCO group (46,49% vs. 81,9%) However, there were no significant differences regarding the tumour cell number in these groups. Conclusions: Cell viability and cell functionality after adrenocortical tumour tissue storage for 72 h are comparable to results obtained from directly processed tissue without cold storage. RPMI- Media 1640Invitrogen was shown to be as effective as selected cold storage solutions in adrenocortical tumour tissue preservation. Thus, cold storage of tumour tissue represents an efficacious option in tumour tissue research enabling an appropriate transport of resected human tumour tissue and enlarging the trading area for human tumour tissue.

82 INTRODUCTION OF HUMAN TUMOUR XENOGRAFTS IN IMMUNODEFICIENT MICE FOR HUMAN ENDOCRINE TUMOURS Alexandra Ozimek, Constanze Hantel, Felix Beuschlein, Klaus Hallfeldt, Thomas Mussack Department of Surgery, LMU University of Munich, 80 336 Munich, Germany; Department of Medicine/Endocrine Research,LMU University of Munich, 80 336 Munich, Germany Keywords: in vivo tumour models, xenograft tumour models, malignant endocrine tumours Background: The poor overall survival time of patients with malignant endocrine tumors of the adrenal gland and the gastroenteropancreatic system gives reason for an increasing need of individualized tumor therapy concepts. Heterotransplantation of human cancer cells or tumour tissue into immunodeficient xenograft model constitutes the major preclinical screen for the development of novel cancer therapeutics for different tumour entities. However, there is a lack of appropriate in vivo tumour models obtained from human adrenocortical carcinomas or malignant gastroenteropancreatic tumours. We introduce a subcutaneous tumour xenograft model in immunodeficient mice for adrenocortical carcinoma and malignant gastroenteropancreatic tumours. Methods: Up to now, human tumour tissue was obtained from patients undergoing adrenalectomy due to an adrenocortical carcinoma (n=2) or resection of a malignant gastroenteropancreatic tumour (n=1). All tumours were minced into uniform small pieces (5×5×5 mm). Two

Langenbecks Arch Surg (2008) 393:759–815 tissue pieces were transplanted into subcutaneous pockets of the neck and flank region of 6–8 weeks old immunodeficient mice at a time. When reaching a maximum tumour diameter of 10 mm, tumours were explanted and characterized using immunohistochemistry. To enable a characterization of the tissue implanted, primary human cell cultures were obtained from all tumours used for human xenograft. Results: In all human xenografts, a small bulge was visible at the implantation site, but disappeared 2–4 weeks after inoculation. At 8– 12 weeks following implantation, a tumour reappeared at the site of implantation in 2/5 mice with adrenocortical carcinoma and 2/2 mice with malignant gastroenteropancreatic tumour. For 1/3 malignant gastroenteropancreatic tumour immunhistochemical analysis was performed and confirmed histological analysis. Primary human tumour cells were obtained from all 3 tumour tissue samples and are successfully maintained in culture for experimental, diagnostic or therapeutic purposes. Actually, molecular biological analysis of all primary tumour cells is still performed in order to confirm tumour integrity with first preliminary results being expected in four weeks. Conclusions: We introduce a human xenograft tumour model for malignant endocrine tumour of the adrenal gland and the gastroenteropancreatic system. The surgical procedure is simple and provides results within 2–4 months after tumour implantation.

83 PREDICTION OF TUMOR INFILTRATION WITH MOLECULAR MARKERS AND IMPACT ON LIMITED SURGERY OF THE ESOPHAGUS Oulfa Dhaouadi (1), Elfriede Bollschweiler (1), Uta Drebber (2), Stephan E. Baldus (3), Stefan P. Mönig (1), Arnulf H. Hölscher (1), Ralf Metzger (1) 1: Department of Visceral and Vascular Surgery, University of Cologne, 2: Department of Pathology, University of Cologne; (3) Department of Pathology, University of Düsseldorf Keywords: tumor infiltration, early esophageal cancer, limited surgery Background: Early cancer of the esophagus is a potentially curable disease. Limited procedures such as endoscopic resection is indicated for dysplasia and mucosal esophageal adenocarcinomas (AC) and squamous cell carcinomas (SCC) under the prerequisite of complete removal. In all cases of submucosal infiltration en-bloc esophagectomy with standardized lymphadenectomy is the treatment of choice. The reason is that the depth of tumor infiltration into the mucosa (T1a carcinoma) and the submucosa (T1b carcinoma) with its three layers (sm1-sm3) is correlated with the rate of nodal metastases and therefore with long-term prognosis. The crucial point for limited therapy is therefore to define the right stage of T1 carcinoma. However, the diagnostic validity of pre-intervention lymph-node identification is very low. For this reason the impact of the molecular markers MMP2, Timp2, Pim-1 and Survivin was analyzed to predict depth of infiltration and consecutively the risk for lymph node metastasis in early esophageal cancer. Methods: 67 patients (43 men, 24 women) with early cancer of the esophagus (pT1) and a median age of 62 years (min:39 years, max: 80 years) were included in the study. There were 28 SCC and 39 AC. 22 patients were pT1a, 45 patients were pT1b. All patients underwent enbloc esophagectomy with standardized lymphadenectomy. From all patients surgical specimens were available for immunohistochemical detection of MMP2, Pim-1, Timp2 and Survivin protein expression. A high-sensitivity immunohistochemical staining was performed at 5 µm sections applying the DAKO EnVision System according to the manufacturer’s instructions. Nuclei were counterstained with hematoxylin. The microscopic evaluation was performed by two pathologists (U.D., S.E.B.) using a scoring system described as follows: score 0: none expression <5%, score 1: low expression of 5–35%, score 2: medium expression was 35–65%, and score 3: strong expression >65%

Langenbecks Arch Surg (2008) 393:759–815 positive cells. Correlations between the degree of staining and clinical parameters i.e. depth of tumor infiltration were calculated by Chi2-test. P-values <0.05 were considered significant. Results: The data showed that a submucosal tumor infiltration could be predicted in 14/45 (31%) patients with pT1b (p=0.047) if at least one of the protein expressions of MMP2, Timp2 or Survivin was scored 3 (>65% positive cells). There was a significant (p=0.021) correlation with the depth of tumor infiltration (sm1-sm3). In the group with pT1a tumors (n=22) 2 patients were overestimated due to a staining score 3. Pim-1 had no predictive value. Conclusions: High protein expression of MMP2, Timp2 or Survivin identifies submucosal tumorinfiltration in 31% of patients with early cancer of the esophagus. Thus, these patients must be excluded from limited procedures such as endoscopic resection because submucosal infiltration has a risk of lymph-node metastases up to 45% independent of histological suptype.

84 SUCCESSFUL ISOLATION OF ENDOTHELIAL CELLS AND FIBROBLASTS OF HUMAN COLORECTAL CANCER AND HEALTHY COLON Vera Schellerer, Roland S. Croner, Kristina Weinlaender, Werner Hohenberger, Michael Stürzl, Elisabeth Naschberger Department of Surgery, Friedrich-Alexander-University, Germany Keywords: Angiogenesis; colorectal cancer; endothelial cells; fibroblasts Background: Colorectal cancer (CRC) is one of the leading cancers in western countries with more than 500 000 deaths worldwide per year. Recently anti-angiogenic drugs have been used successfully for treatment. Our target was to establish a protocol for the successful isolation of human endothelial cells and fibroblasts of CRC to optimize treatment strategies in-vitro. Methods: Tissues of tumor and healthy colon were obtained from patients undergoing standard surgical procedure for primarily diagnosed CRC at the Department of Surgery of the Friedrich-AlexanderUniversity Erlangen. Endothelial cells and fibroblast of have been isolated and immunocytochemically stained. Results: Viable endothelial cells (EC) were isolated with a purity of >93% via magnetic cell sorting of tissue samples obtained from CRC and healthy colon of ten patients undergoing surgical therapy. TEC and NEC expressed CD31, CD105, VE-cadherin, V-CAM-1, I-CAM-1 and Eselectin, formed capillaries in matrigel and were able to take up acetylated LDL. In addition the isolated cells were negative for podoplanin, CD45, CD68 and CK-20 demonstrating blood vessel EC lineage of the cells. Viable fibroblasts from 8 patients were isolated. Immunocytochemical stainigs were positive for vimentin and negative for CK 20 and CD31. Conclusions: Endothelial cells and fibroblasts isolated from CRC and corresponding healthy colon may provide a useful in vitro model to elucidate epigenetic effects on angiogenesis in cancer and to optimize anti-angiogenic therapy.

85 THE IMPACT OF CHRONIC STRESS ON TUMOR GROWTH AND INVASION IN AN ORTHOTOPIC PANCREATIC CANCER MODEL IN IMMUNCOMPETENT C57/BL6-MICE Lars Ivo Partecke, A. Käding (1), M. Sendler (2), K. Cziupka (1), N. Siebert (1), S. Speerforck (1), F.U. Weiss (2), J. Mayerle (2), M.M. Lerch (2), A. Stier (1), C.D. Heidecke (1) und W. v. Bernstorff (1) 1: Klinik und Poliklinik für Chirurgie, Abteilung für Allgemeine Chirurgie-, Viszeral-, Thorax- und Gefäßchirurgie der Ernst-MoritzArndt Universität Greifswald, 2: Klinik für Innere Medizin A der Ernst-Moritz-Arndt Universität Greifswald

791 Keywords: chronic stress — immune dysfunction — pancreatic cancer — tumor growth Background: Chronic stress is a common problem in the perioperative management of cancer patients. Pancreatic cancer is associated with a poor prognosis anyway and might be influenced supplementary as chronic stress leads to immune dysfunction. In this study we investigated the impact of chronic stress on the immune system, the tumor growth and metastasis in an established model of pancreatic cancer in immuncompetend mice. Methods: The murine cancer cell line 6606 PDA was injected orthotopically into the pancreas of C57/Bl6 mice. Afterwards the mice were exposed to chronic combined acoustic and restraint stress in an well established stress modell. For control purposes non-stressed tumor bearing mice were used. The individual stress was scored with an adapted stress severity score. Tumor growth and metastasis were detected by small animal MRI and compared to those of non-stressed mice. After induction of stress the levels of IL-6, IL-10, IL-12, TNF and IFN- were determined in the supernatant of LPS stimulated splenocytes, whereas the concentrations of corticosterone and norepinephrine were determined in the urine and blood. The diameter of the adrenal gland was examined via small animal MRI and served also as an indicator for the stress severity. The survival was described using the Kaplan-Meier curve. Finally the tumors were evaluated using different histological parameters. Results: The stress severity score showed a highly significant difference between stressed and non-stressed mice (p<0,0001). With increasing time of observation stressed mice had considerable higher tumor volumes than non-stressed mice (median 992,8 mm3 vs 287,0 mm3 42 days after tumor implantation). Stressed mice exhibited significant (p<0,05) reduced levels of Cytokines in the supernatant of LPS stimulated splenocytes. The level of corticosterone was significantly (p<0,05) elevated and the diameter of the adrenal gland was significantly (p=0,0025) increased in stressed mice. Histological exploration of the tumors of stressed mice showed an increased level of infiltration and we found a trend to exceeding lymphatic spread. The Kaplan-Meier survival curve showed a significant (p<0,05) poorer survival of stressed mice. Conclusions: Chronic stress causes a considerable immunosuppression in a model of pancreatic cancer in immuncompetend mice and has an impact on tumor growth and survival. Therfore a perioperative reduction of chronic stress might improve the prognosis of patients with pancreatic cancer.

86 LYMPH NODE METASTASES IN SOFT TISSUE SARCOMAS A SINGLE CENTER ANALYSIS OF 1597 PATIENTS Marcus Lehnhardt, Cornelius Kuhnen, Ingo Stricker, Rose Moritz, Hans U Steinau, Adrien Daigeler Universitätsklinik für Plastische Chirurgie und Schwerbrandverletzte, BG-Kliniken Bergmannsheil, Ruhr-Uni-Bochum, Germany Keywords: sarcoma, lymph node, metastasis, prognosis, therapy Background: The aim of this study was to examine the clinical course of patients with the rare finding of regional lymph node metastasis (RLNM) from soft tissue sarcoma. Methods: Data from 28 out of 1597 consecutive soft tissue sarcoma patients with RLNM was from the patients’ charts and interviewing patients and general practitioners. Survival, including possible influencing factors, was statistically calculated. Results: RLNM were seen in 21.4% for epithelioid sarcoma and 17.6% for clear cell sarcoma. All other entities presented RLNM rates below 10%. At follow-up after an average of 9.6 years, only 3 patients were alive with no evidence of disease. Survival was independent from surgical resection status of the primary tumor and the RLNM, as well as

792 from adjuvant radiation and chemotherapy. Tumor entity, as well as the length of the time period from primary to RLNM, affect survival. Conclusions: Surgical treatment, as well as radiation and chemotherapy, may improve survival in selected cases but probably have their value much more in terms of local disease control and improvement life quality of patients who probably already suffer from an aggressive systemic disease at time of nodal involvement. TRAIL and Taurolidine induce apoptosis and decrease proliferation in human fibrosarcoma.

87 TRAIL AND TAUROLIDINE INDUCE APOPTOSIS AND DECREASE PROLIFERATION IN HUMAN FIBROSARCOMA Marcus Lehnhardt, Christina Brenzel, Hans-Ulrich Steinau, Ulrich Mittelkötter, Waldemar Uhl, Ansgar Michael Chromik, Adrien Daigeler Universitätsklinik für Plastische Chirurgie und Schwerbrandverletzte, BGKliniken Bergmannsheil, Ruhr-Uni-Bochum, Germany Keywords: taurolidine, tauroline, apoptosis, fibrosarcoma, HT1080, TRAIL, Background: Disseminated soft tissue sarcoma still represents a therapeutic dilemma because effective cytostatics are missing. Therefore we tested TRAIL and Tarolidine (TRD), two substances with apoptogenic properties on human fibrosarcoma (HT1080). Methods: Viability, apoptosis and necrosis were visualized by TUNELAssay and quantitated by FACS analysis (Propidiumiodide/AnnexinV staining). Gene expression was analysed by RNA-Microarray and the results validated for selected genes by rtPCR. Changes on protein level were documented by Western Blot analysis. NFKB activity was analysed by ELISA and proliferation assays (BrdU) were performed Results: The single substances TRAIL and TRD induced apoptotic cell death and decreased proliferation in HT1080 cells significantly. Gene expression of several genes related to apoptotic pathways (TRAIL: ARHGDIA, NFKBIA, TNFAIP3; TRD: HSPA1A/B, NFKBIA, GADD45A, SGK, JUN, MAP3K14) was changed. The combination of TRD and TRAIL significantly increased apoptotic cell death compared to the single substances and lead to expression changes in a variety of genes (HSPA1A/B, NFKBIA, PPP1R15A, GADD45A, AXL, SGK, DUSP1, JUN, IRF1, MYC, BAG5, BIRC3). NFKB activity assay revealed a reduction of the p50 subunit by TRD, TRAIL and the combination therapy. Conclusions: TRD and TRAIL are effective to induce apoptosis and decrease proliferation in human fibrosarcoma. A variety of genes seems to be involved pointing to the NFKB pathway as key regulator in TRD/TRAIL-mediated apoptosis.

88 EX VIVO TREATMENT OF MALIGNANT PLEURAL EFFUSIONS OF METASTATIC BREAST CANCER PATIENTS WITH THE BISPECIFIC ANTIBODY MT110 TARGETING EPCAM AND CD3 Juliane Witthauer (1), Bernd Schlereth (2), Klaus Brischwein (2), Ilona Funke (3), Karl-Walter Jauch (1), Patrick Baeuerle (2), Barbara Mayer (1) 1: Ludwig-Maximilians Universität, Klinische Forschung Chirurgie, Klinikum Großhadern, München, Deutschland, 2: Micromet AG, München, Deutschland; 3 Chirurgische Klinik Dr. Rinecker, München, Germany Keywords: malignant pleural effusion, breast cancer, EpCAM, CD3, MT110, ex-vivo therapy

Langenbecks Arch Surg (2008) 393:759–815 Background: Despite therapy, approximately half of patients with metastatic breast cancer develop a malignant pleural effusion (MPE) associated with short survival. Hence, new treatment strategies are urgently needed. In the present study, the efficacy of the bispecific single-chain antibody MT110 targeting both the epithelial antigen EpCAM (CD326) and the T cell antigen CD3 was investigated ex vivo with MPE samples of breast cancer patients. Methods: Target antigen expression of MPE cells was analyzed by immunohistochemistry (ABC-method on cytospins, FACS). Percentage of redirected target cell lysis by MT110 was determined by double staining of cells with 7-amino actinomycin (7-AAD) and an antiEpCAM antibody using FACS analysis. Activation of autologous CD4+ and CD8+ cells in response to MT110 was studied by the expression of CD25, granzyme B and different cytokines i. e. IL-2, IL4, IL-6, IL-10, TNF-α and IFN-γ (FACS). Results: EpCAM+ cells were found in 14 out of 18 (78%) MPE samples from metastatic breast cancer patients. The fraction of EpCAM+ pleural carcinoma cells varied between 30% and 100% (mean 78%). CD3positive cells were detected in all MPE samples ranging from 60% to 93% (mean 80%). Seven effusion samples were analyzed for MT110 treatment and revealed a dose-dependent and specific redirected lysis of EpCAM+ cells after 48 h and 72 h with 10 ng/ml (48 h, 72 h: p=0.03) and 1000 ng/ml (48 h, p=0.03; 72 h, p=0.016). After 72 h, 57%±29.5% (mean ± SD) of EpCAM+ cells were killed using 1000 ng/ml MT110. Antibody treatment (1000 ng/ml) increased the fraction of CD25+/CD4+ cells after 48 h (p=0.03) and 72 h (p=0.016). Similar, CD25 expression on CD8+ cells was stimulated after 48 h (p=0.03) and 72 h (p=0.016) using 1000 ng/ml MT110. Cytokine analysis revealed a strong TH1 immune response detecting an increased TNF-α (p=0.016) and INF-γ (p=0.03) secretion. Conclusions: Single-agent treatment with MT110 is capable of activating unstimulated autologous T cells for efficient and specific redirected lysis of EpCAM+ tumor cells in MPE from breast cancer patients. Future clinical studies are warranted to investigate the efficacy of bispecific antibody MT110 in breast cancer patients resistant to conventional treatment using the bispecific antibody MT110.

89 COMBINATION CHEMOTHERAPY PRIOR TO RECONSTITUTION WITH AUTOLOGOUS PBMC AND VACCINATION WITH AUTOLOGOUS TUMOR AND GM-CSF FAILS TO INDUCE LONG-TERM REDUCTION OF CIRCULATING REGULATORY T CELLS IN RESECTED NSCLC PATIENTS Natasja K. van den Engel (1), Hauke Winter (1), Heike Pohla (2), Dolores J. Schendel (2), Nina Schupp (1), Beate Wagner (4), Tarsem Moudgil (3), Karl-Walter Jauch (1), Bernard A. Fox (3), Rudolf A. Hatz (1), Dominik Rüttinger (1) 1: Department of Surgery, Grosshadern Medical Center, University of Munich, Germany, 2: Institute of Molecular Immunology, GSF National Research Center for Environment and Health, Munich, Germany, 3: Earle A. Chiles Research Institute, Providence Portland Medical Center, Portland, OR, 4: Department of Transfusion Medicine, Grosshadern Medical Center, University, Munich, Germany Keywords: NSCLC, lymphopenia, tumor cell vaccination, cyclophosphamide, fludarabine, regulatory T cells, GM-CSF Background: To date, standard adjuvant chemotherapy for non-small cell lung cancer (NSCLC) provides only modest survival benefit and is associated with substantial toxicity. Clearly, there is a need for new therapies in the adjuvant setting that demonstrate efficacy in NSCLC with less toxicity. Active specific immunotherapy applied during a phase of lymphopenia-driven T cell proliferation is an innovative strategy that has been shown to augment the anti-tumor immune

Langenbecks Arch Surg (2008) 393:759–815 response and result in therapeutic efficacy in preclinical models. This beneficial effect may be explained by an increased sensitivity of lymphocytes to react to antigens and through a reduction of regulatory T cells induced by the chemotherapy regimen. Methods: We performed a Pilot-Phase I clinical trial in patients with NSCLC (stages I–IIIA) following surgical resection using a strategy of lymphopenia induction, reconstitution and vaccination. Tumor cell vaccines were generated from enzyme-digested autologous lung tumor. Patients received preparative chemotherapy (cyclophosphamide 350 mg/m2 and fludarabine 20 mg/m2) on 3 consecutive days for induction of lymphopenia followed by reconstitution with an autologous apheresis product. Irradiated vaccines were administered intradermally on day 1 following reconstitution and every two weeks for a total of 5 vaccinations. Granulocyte-macrophage-colony-stimulating-factor (GM-CSF) was administered at the vaccine site via a minipump (50 µg/24 h) for six days after each vaccination. Results: Five patients were recruited on the trial and four patients successfully underwent the complete treatment procedure. T cell subsets (CD4, CD8, CD25) as well as B cells and NK cells were significantly reduced by the preparative chemotherapy in all patients. However, CD4+ CD25+ T cells followed the same kinetics of recovery as the other subsets. Flow cytometry analysis showed an increased frequency of FoxP3+ CD25+ cells in the peripheral blood of all patients following the 5th vaccination as compared to the peripheral blood before treatment. Isolated CD25+ cells inhibited proliferation of autologous CD25- PBMC in vitro, illustrating their regulatory properties. Conclusions: Our results do not support a persistent reduction of FoxP3+ regulatory T cells by the combination chemotherapy applied in our study. Subsequent studies may therefore include depletion or functional inactivation of CD25+ cells from the apheresis product prior to reconstitution.

90 PERITUMORAL BLOOD VESSELS ARE THE MAJOR SITE OF LEUKOCYTE RECRUITMENT IN EXPERIMENTAL PANCREATIC CANCER Olegas Deduchovas, Vachtang Kerkadze, Jan Schmidt, Eduard Ryschich Chirurgische Klinik, Universität Heidelberg, Germany Keywords: pancreatic cancer, leukocyte recruitment, time-lapse microscopy, transgenic mice Background: Pancreatic cancer induce cellular immune response which can destroy a part of tumor cells. The major problem of insufficient immune response is that leukocytes strongly infiltrate peritumoral tissue whereas intratumoral infiltration is low or absent. The mechanism of such differential infiltration was not known. In the present study, we investigated the site of recruitment of tumor-infiltrating leukocytes and the dynamic leukocyte- tumor endothelium interaction as well as leukocyte migration in experimental pancreatic cancer. Methods: Pancreatic cancer was induced using intraperitoneal inoculation of Panc02 tumor cells in lys-EGFP (fluorescent neutrophils/monocytes) and in CD2-EGFP+ (fluorescent T-cells/NK-cells). Leukocyte-tumor endothelium interaction, leukocyte extravasation and extravascular migration were analysed using digital time-lapse intravital microscopy. The density of peritumoral and intratumoral leukocyte infiltration was measured using laser scanning confocal microscopy. Expression of ICAM-1 and ICAM-2 on tumor endothelium was analysed immunohistochemically. Results: Immunohistochemistry demonstrated that the fraction of ICAM-1 and ICAM-2-expressing blood vessels in tumor (25.3+/− 6.1%; 9.0+/−2.9% respectively) was significantly lower than the fraction of ICAM-1 and ICAM-2-expressing venules in normal pancreas (100%). Basal and chemoattractant-stimulated leukocyteendothelium interaction in the intratumoral blood vessels was strongly reduced compared to the peritumoral and normal pancreatic venules.

793 Leukocytes emigrated mainly from the peritumoral venules, but not from intratumoral blood vessels. The velocity of extravasated lymphocytes was significantly higher in tumor than in peritumoral tissue, whereas neutrophils/ monocytes cells migrated significantly faster in peritumoral tissue. Although leukocytes showed high migratory activity in the peritumoral and in peripheral tumor tissue, they were not able to penetrate into the tumor depth. Conclusions: Intratumoral vascular system fails to recruit leukocytes. Peritumoral, but not intratumoral blood vessels represent the major site for the leukocyte extravasation and determine the high leukocyte infiltration in peritumoral tissue.

91 INHIBITION OF P38-MAP-KINASE PATHWAY ATTENUATES INFLAMMATORY REACTIONS IN THE TUNICA MUSCULARIS AND PREVENTS POSTOPERATIVE ILEUS Tim Vilz, Sven Wehner, Stefan Straesser, Andreas Hirner, Jörg C. Kalff Department of Surgery, University of Bonn, Germany Keywords: Postoperative Ileus, p38-MAP-Kinase-Inhibition, Muscularis-Macrophages Background: The mechanical trauma of the gut is an unavoidable consequence of abdominal surgery leading to massive inflammation of the entire bowel wall and subsequently in postoperative ileus. Further studies revealed the activation of the resident macrophages in the tunica muscularis as an initial step in the inflammatory cascade. The aim of this study was to investigate effectivity of a macrophage-specific p38 MAP-Kinase Inhibitor in preventing postoperative ileus. Methods: C57BL6/J mice were treated with placebo or a p38 MAPKinase inhibitor 90 minutes before intestinal manipulation (IM) or sham operation. Animals were sacrificed after 3, 6 and 24 h after IM and tunica muscularis was prepared for further investigations. The expression of the proinflammatory mediators IL-6, TNF-alpha, MIP-1 and MCP-1 were determined by quantitative PCR. NO-production as the major intestinal inhibitory neurontransmitter was measured by Griess-Reaction. The infiltration of neutrophils in the tunica muscularis was detected by MPO-staining and counted to quantify the inflammatory extend. Subsequently, contractility of muscularis strips was examined in an in vitro organ bath and gastrointestinal transit was measured to analyze the motility function in vivo. Results: RNA expression of inflammatory mediators and NO production were significantly reduced after preoperative p38 MAPKinase treatment. Neutrophile infiltration was nearly completely abregated by p38 MAP-Kinase Inhibition. Finally, jejunal contracility and gastrointestinal transit were normalized after treatment. Conclusions: Our results demonstrate that postoperative inflammation of the tunica muscularis can be prevented by blocking the p38 MAP-Kinase pathway in macrophages. This inhibition prevents development of postoperative ileus in rodents.

92 ANALYSIS OF RET, ZFHX1B, EDN3 AND GDNF GENOMIC REARRANGEMENTS IN 80 PATIENTS WITH HIRSCHSPRUNG DISEASE (USING MULTIPLEX LIGATION-DEPENDENT PROBE AMPLIFICATION) Alexandre Serra, H. Görgens (†), K. Alhadad (†), A. Ziegler (§), G. Fitze (*), H.K. Schackert (†) * †: Departments of Pediatric Surgery and Surgical Research, Technische Universität Dresden, §: Institute of Medical Biometry and Statistics, Medical University of Lübeck, Germany

794 Keywords: Hirschsprung disease, genomic rearrangements, RET proto-oncogene, multiplex ligation-dependent probe amplification Background: Hirschsprung disease (HSCR) is transmitted in a complex pattern of inheritance and is mostly associated with variants in the RET proto-oncogene. However, RET mutations are only identified in 15– 20% of sporadic HSCR cases and solely in 50% of the familial cases. Since genomic rearrangements in particularly sensitive areas of the RET proto-oncogene and/or associated genes may account for the HSCR phenotype in patients without other detectable RET variants, the aim of the present study was to identify rearrangements in the coding sequence of RET as well as in three HSCR-associated genes (ZFHX1B, EDN3 and GDNF) in HSCR patients by using Multiplex Ligation-dependent Probe Amplification (MLPA). Methods: DNA from blood leukocytes in 80 HSCR patients were screened for genomic rearrangements in RET, ZFHX1B, EDN3 and GDNF using MLPA, according to standard techniques. Results: We have not identified any rearrangements (deletions or amplifications) in these four genes in all patients studied. Conclusions: We conclude that genomic rearrangements in RET are rare and were not responsible for the HSCR phenotype in individuals without identifiable germline RET variants in our group of patients, yet this possibility cannot be excluded altogether because the confidence to identify variation in at least two percent of the individuals was only 95%.

93 CD4 CELLS, BUT NOT TREG OR TH17-CELLS CONTRIBUTE TO THE PROPAGATION OF THE GASTROINTESTINAL FIELD EFFECT Arne Koscielny, Daniel Engel (*), Christian Kurts (*), Jörg Kalff Klinik und Poliklinik für Allgemein-, Viszeral-, Thorax- und Gefäßchirurgie, *: Institut für Molekulare Medizin und Experimentelle Immunologie am Universitätsklinikum Bonn, Germany Keywords: field effect, CD4 cells, Th17 cells, Treg cells Background: Intestinal manipulation (IM) leads to local bowel wall inflammation subsequently spreading over the entire gastrointestinal tract. Previously, we have demonstrated this gastrointestinal field effect (FE) in rodent models. We showed IM-mediated recruitment of dendritic cells (DC) into the small bowel muscularis and mesenteric lymph nodes (MLN) combined with upregulation of costimulatory molecules on DCs. Furthermore, we could show a central involvement of DC, presumabely by CCR7-dependent mechanisms operating few hours after IM, of the GALT and the MLN in the FE. The aim of this study was to investigate the role of T-cells in the gastrointestinal field effect. Methods: Mice underwent standardized IM or sham operation and tissues (small bowel muscularis, colonic muscularis, mesenteric lymph nodes) were obtained at different times after IM. The FE was analysed by measuring the gastrointestinal transit time or muscularis contractility in wildtype mice (WT) versus T cell deficient mice. Furthermore, expression of inflammation markers (e.g. IL-6, TNF-α) and of adhesion molecules (e.g. ICAM-1) was determined by Taqman-PCR. CD4deficient- (CD4-/-), regulatory T cell-deficient- (DEREG) and TH17deficient mice (p19-/-) were used to determine the role of T cell subset in FE. The local inflammation was determined by preparing immunohistochemistry for leucocytes on jejunal and colonic specimen of those mice. Results: Lack of CD4 cells significantly reduced the FE after IM as compared to WT. Concomittantly, we observed a significant downregulation of inflammation markers (IL-6, TNF-α) in CD4 deficient intestinal and colonic muscularis and a lower amount of inflammatoric cells in these T cell-deficient mice. In contrast, lack of TH17 cells and depletion of regulatory T cells mice did not ameliorate the FE. Additionally, there was no upregulation of IL-17, IL-23, STAT3 or FOXp3 in WT mice.

Langenbecks Arch Surg (2008) 393:759–815 Conclusions: CD4 cells are centrally involved in the generation and propagation of the FE. Analysis of T cell subsets showed that neither Th17 nor regulatory T cells contribute to the FE.

94 DIFFERENTIAL TIMP-1 MRNA EXPRESSION IN MONOCYTES OF PATIENTS WITH MULTIPLE INJURIES V. Stoecklein (1), V. Bogner (1), L. Keil (1), T. Giese (2), W. Mutschler (1), P. Biberthaler (1) 1: Department of Traumatology and Orthopedic Surgery — Downtown Campus, Ludwig-Maximilians-University Munich, 2: Institute for Immunology und Molecular Diagnostics, Rupprecht-Karls-University, Heidelberg, Germany Keywords: multiple injuries, immune response, MMPs Background: Altered monocyte function plays a major role in the development of Systemic Inflammatory Response Syndrome (SIRS) and consecutive Multiple Organ Failure (MOF) in multiple trauma patients. Precedent genome wide microarray studies showed, that tissue inhibitor of metalloproteinases-1 (TIMP1), an inhibitor of matrix metalloproteinase 9 (MMP9), is upregulated and differentially expressed depending on clinical outcome. To further evaluate these findings, monocyte mRNA levels of TIMP-1 was investigated in a larger patient collective. Methods: 23 patients presenting with blunt multiple injuries (ISS>16 points) were included. Blood samples were drawn on admission (<90 minutes after trauma) and at 6, 12, 24, 48 and 72. Monocytes were isolated from whole blood. After mRNA extraction (RNeasy Midi Kit, Qiagen, Hilden), rt-PCR was performed. The data was statistically analyzed using ANOVA on ranks and t-test (Sigma STAT, SpSS Inc.). Results: 8 of 23 patients did not survive the traumatic event (90-day survival). After a peak at 6 h (P=<0,001, t-test), monocytes TIMP-1 expression continuously decreases in all patients until baseline levels at 48 h. Patients with adverse clinical outcome exhibit higher expression values at 6 h, 12 h, 24 h and 72 h. At 24 h, the differences are significant (p<0,032, t-test). Conclusions: Following our genome wide monocytes gene expression studies in multiple injured patients, we now present an rt-PCR investigation of TIMP-1 in an extended patient collective. TIMP-1 shows enhanced expression levels in the very early posttraumatic period in all patients. Furthermore, TIMP-1 is elevated in patients with adverse outcome. This suggests a potential role for TIMP1 and the inhibited MMPs in the pathogenesis of posttraumatic immunologic disorders. Further studies to more in detail clarify the pathophysiological meaning of these findings are currently under way.

95 BILIVERDIN ATTENUATES POSTOPERATIVE ILEUS AFTER INTESTINAL MANIPULATION Maria Sioutis, Johannes Chang, Stephan Bortscher, Andreas Hirner, Marcus Overhaus Klinik und Poliklinik für Allgemein-, Viszeral-, Thorax- und Gefäßchirurgie, Bonn Keywords: Biliverdin, Postoperative Ileus, IL-6, Transit, Neutrophils Background: Postoperative atony and ileus are major obstacles in postsurgical rehabilitation characterized by persistent activated enteric and systemic pro-inflammatory cytokines and their pathways. In heme degradation, the anti-inflammatory efficiency of carbon monoxide and hemeoxygenase-1 (HO-1) has already been demonstrated. This study investigates biliverdin’s protective effect on small bowel and colonic integrity after intestinal manipulation.

Langenbecks Arch Surg (2008) 393:759–815 Methods: Sprangue-Dawley rats were subjected to intestinal manipulation. Biliverdin-treated animals with intestinal manipulation (IM+B) were compared to intestinal-manipulated animals without biliverdin treatment (IM) and to sham operated animals (C). In vivo gastrointestinal transit and in vitro intestinal muscle contractility were measured. RT-PCR was performed 3, 6 and 24 h after intestinal manipulation on pro- inflammatory mRNA (IL-6) in the small bowel muscularis. Small bowel whole mounts were histochemically stained for myeloperoxidase leukocytes infiltration (N=6/group; p<0.05). Results: Biliverdin treatment reversed the surgical-induced significant transit delay (Geometric Center C: 9.18±0.67 vs. IM: 6.92±1.47 vs. IM +B: 8.36±1.23). 24 h after intestinal manipulation the small bowel muscle contractility was attenuated by 65% (jejunum: C: 3.24±0.55 vs. IM: 1.14± 0.25 g/mm2/sec at 100μM bethanechol). With a simultaneous biliverdin injection this attenuation was almost completely reversed (IM +B: 3.14±0.13 g/mm2/sec, p<0.05). Additionally, biliverdin significantly inhibited the neutrophilic infiltration into the intestinal muscularis after manipulation. The surgical-induced mRNA up-regulation at 3, 6 and 24 h of pro-inflammatory IL-6 in the small bowel muscularis compared to controls was significantly reduced with biliverdin treatment (IM+B at 3 h: 67%, 6 h: 78%, 24 h: 92% reduction vs. IM, p<0.05). Conclusions: During intestinal manipulation biliverdin treatment preserved the intestinal integrity. This beneficial effect was due to the inhibition of neutrophilic migration and suppression of the pro-inflammatoric (Il-6) cascade. Hence, biliverdin prevented the intestinal atony and inflammation and interrupted the intestinal momentum in postoperative ileus.

96 INHIBITION OF P38-MAP-KINASE PATHWAY ATTENUATES INFLAMMATORY REACTIONS IN THE TUNICA MUSCULARIS AND PREVENTS POSTOPERATIVE ILEUS Tim Vilz, Sven Wehner, Stefan Strässer, Andreas Hirner, Jörg Kalff Department of Surgery, University of Bonn, Germany Keywords: Postoperative Ileus, Macrophages, p38 MAP-KinaseInhibition Background: The mechanical trauma of the gut is an unavoidable consequence of abdominal surgery leading to massive inflammation of the entire bowel wall and subsequently in postoperative ileus. Further studies revealed the activation of the resident macrophages in the tunica muscularis as an initial step in the inflammatory cascade. The aim of this study was to investigate effectivity of a macrophagespecific p38 MAP-Kinase Inhibitor in preventing postoperative ileus. Methods: C57BL6/J mice were treated with placebo or a p38 MAPKinase inhibitor 90 minutes before intestinal manipulation (IM) or sham operation. Animals were sacrificed after 3, 6 and 24 h after IM and tunica muscularis was prepared for further investigations. The expression of the proinflammatory mediators IL-6, TNF-alpha, MIP-1 and MCP-1 were determined by quantitative PCR. NO-production as the major intestinal inhibitory neurontransmitter was measured by Griess-Reaction. The infiltration of neutrophils in the tunica muscularis was detected by MPO-staining and counted to quantify the inflammatory extend. Subsequently, contractility of muscularis strips was examined in an in vitro organ bath and gastrointestinal transit was measured to analyze the motility function in vivo. Results: RNA expression of inflammatory mediators and NO production were significantly reduced after preoperative blocking of the p38 MAP-Kinase pathway. Neutrophile infiltration was nearly completely abregated by p38 MAP-Kinase Inhibition. Finally, jejunal contracility and gastrointestinal transit were normalized after treatment. Conclusions: Our results demonstrate that postoperative inflammation of the tunica muscularis after intestinal manipulation can be prevented by blocking the p38 MAP-Kinase pathway in macrophages. This inhibition prevents the development of postoperative ileus in rodents.

795

97 LOSS OF CO2-REACTIVITY IN CEREBRAL RESISTANCE VESSELS AFTER SUBARACHNOID HEMORRHAGE IN MICE Benjamin Friedrich, Sergej Feiler, Sebastian Sonanini, Nikolaus Plesnila Laboratory of Experimental Neurosurgery, Department of Neurosurgery & Institute for Surgical Research, University of Munich Medical Center — Großhadern, Munich, Germany Keywords: SAH, CO2, MICE, REACTIVITY, CEREBRAL, VESSEL Background: Carbon dioxide (CO2) plays a major role in the regulation of the brains microcirculation. The aim of our study was to asses the CO2reactivity after experimental subarachnoid haemorrhage (SAH). We compared the reactivity to different CO2 partial pressures (pCO2) of control animals’ pial vessels with those 3 and 24 hours after SAH. Methods: SAH was induced in male C57/Bl6 mice by the previously described filament perforation model. 3 respectively 24 hours later pial microcirculation of the brain was visualized by a fluorescent intravital microscopy technique through an opened cranial window, using fluorescein isothiocyanate bound to dextran. Mean arterial blood pressure was kept constant above 60 mmHG. In one group (called “Ventilation”) the vessel images where recorded every 15 minutes over an hour to achieve a proper baseline. Afterwards the animals where ventilated with 10% CO2 for 15 minutes and the vessel images where recorded likewise. In the other group (called “Frequency”) the animals where successively normo-/ hyper-/normo-/hypoventilated. Every ventilation period lasted for 15 minutes and always two images per vessel where recorded. The diameter of the pial vessels was measured by a computer-assisted method. Results: The first measured diameter of the vessel in every group was defined as 100%. In group “Ventilation” the arterial diameter was constant (103±9,4%) before ventilation with CO2. After ventilation with 10% CO2 (pCO2=274±11,3 mmHg) the vessel diameter in control animals enlarged significant (164±16,4%; p<0,001). Neither the 3 nor the 24 hour post SAH animals showed any significant vessel reaction (99±10,1% respectively 107±3,5%; p>0,05) to ventilation with 10% CO2. In group “Frequency” the arterial diameter of control animals shrank (86±8,3%; p<0,001) after hyperventilation (pCO2= 22±2,3 mmHg), returned to normal (102±12,1%) when normoventilated again and dilatated (134±17,7%) after hypoventilation (pCO2= 49±1,9 mmHg). Mice which suffered a SAH did not show any significant reaction to different ventilation frequencies (101±7,78% after hyperventilation; 105±8,48% after hypoventilation). Conclusions: SAH is a major cause of mortality and morbidity in the western civilization. The neurological deficits after SAH, mainly attributed to vasospasm, are one of the most feared complications after SAH. Our data suggest, that quite rapidly after SAH there are dramatic changes in the regulation of the brains microcirculation, which may contribute to the dangerous situations occurring after SAH. The mechanisms of the abrogated CO2-reactivity after SAH are yet unclear and require further investigation.

98 HYDROGEN SULFIDE (H2S) MEDIATES ANTI-COAGULATORY EFFECTS BY A DOWNREGULATION OF P-SELECTIN EXPRESSION Heiko Sorg, Hanna Vörsmann, Brigitte Vollmar Institut für Experimentelle Chirurgie, Universität Rostock, Rostock, Germany Keywords: thrombosis, coagulation, inflammation Background: There is considerable evidence that the inflammation and the coagulation cascades are not independently acting pathways but are rather interconnected assuming that both, inflammatory mediators and thrombin

796 production are leading to endothelial and microvascular injury. Thus, it is suggested that the supplementation of endogenous substances might have beneficial effects on the course and outcome of inflammatory diseases. In this context, the gasotransmitters hydrogen sulfide (H2S) is increasingly recognized for its anti-inflammatory and anti-adhesive properties. In order to evaluate its anti-coagulatory effects, we studied the role of H2S on platelets in vitro and on microvascular vessel occlusion in vivo. Methods: In a light/dye-injury model of microvascular thrombosis in the ear of the hairless mouse, kinetics of thrombus formation in venules were quantitatively assessed by 50% vessel diameter reduction and complete vessel occlusion (CVO) using intravital fluorescence microscopy. Animals received either a continuous intravenous infusion of a H2S-donor (Na2S; 15 µmol/kgxh; n=7) or a single iv-bolus of a H2S-inhibitor (propargylglycin (PAG); 100 mg/ kg; n=7). Control animals received equivalent volumes of physiologic saline (n=10). Thrombus formation has further been characterized by laboratory analyses, histology and immunohistochemistry. Flow cytometry was used to evaluate the effect of H2S on TRAP (thrombin receptor activating peptide)-induced human platelet activation. Results: In Na2S-pretreated animals venular thrombus formation was significantly delayed (CVO: 726±66 s; p<0.001) as compared to saline-controls (552±24 s), while application of the H2S-inhibitor PAG significantly (p<0.001) accelerated thrombus kinetics (346±21 s) as compared to both other groups. Moreover, complete venular occlusion upon continuous epi-illumination was only found in 50% of all vessels studied in Na2S-treated animals, while saline- and PAG-administration could not prevent vessel occlusion at all (100% occlusion rate). Concomitantly, bleeding time was markedly prolonged in Na2S- versus control- and PAG-animals. Laboratory analyses revealed physiological Quick-values as well as partial thromboplastin time in all groups studied. Animals treated with Na2S presented with markedly reduced fibrinogen concentrations, while PAG led to significantly elevated fibrinogen concentrations compared to the control group, implying a higher risk for vascular events in case of reduced systemic H2S-levels. In immunohistochemical staining for P-selectin and PAI-1 we could observe a significantly decreased expression of both molecules on the endothelial surface of vessels in Na2S-treated animals compared to the PAG-group. TRAP-induced platelet P-selectin expression, as assessed by flow cytometry, was almost completely abolished by concomitant Na2S-exposure, whereas upon TRAP-exposure PAG slightly enhanced the fraction of P-selectin expressing platelets. Conclusions: Our results clearly demonstrate that H2S effectively prevents venular thrombus formation as given by significantly decelerated thrombus formation. Furthermore, this effect is associated with decreased endothelial cell and platelet P-selectin expression. In summary, H2S seems to play a distinct role in microvascular thrombus formation implying that the application of H2S-donors could represent a potential preventive strategy in diseases associated with thromboembolism or as successful thrombosis prophylaxis during inflammatory diseases.

99 LONG-TERM ERYTHROPOIETIN (EPO) TREATMENT IMPROVES FRACTURE HEALING IN A MODEL OF DELAYED UNION Victor Speidel, Patric Garcia, Tina Histing; Jörg, Holstein, Tim Pohlemann, Michael D. Menger Klinik für Unfall-, Hand- und Wiederherstellungschirurgie und Insititut für klinisch-experimentelle Chirurgie; Universitätsklinikum des Saarlandes, Homburg, Germany Keywords: fracture healing, bone, delayed union, erythropoietin Background: The glucopeptide EPO is widely known for the treatment of anemia in patients with renal failure. However, there is increasing

Langenbecks Arch Surg (2008) 393:759–815 evidence that EPO additionally has a variety of non-hematopoetic and tissue protective effects. Holstein et al. have previously shown that the EPO-receptor is expressed in hypertrophic chondrocytes in the periosteal callus. They further showed that short-term EPO treatment with 5000 U/ kg for 6 days improves endochondral ossification after 2 weeks in a closed femoral fracture model in mice. However, EPO treatment did not influence the outcome of fracture healing after 5 weeks. The aim of our study was to analyze, whether long-term EPO treatment is also able to stimulate the healing process in a model of delayed union. Methods: We used a mouse femoral fracture model which results in a delayed union. Segmental defects of 0.25 mm were created in the right femur using a gigly wire saw. Osteotomies were stabilized with a pinclip fixation technique, as described previously. 16 CD1 mice were treated with EPO (500 U/kg i.p.) and 16 animals received vehicle only (i.p.) Fracture healing was analyzed after 2 and 5 weeks by radiological, biomechanical and histological analysis. Results: X-ray analysis (callus density/spongiosa density of tibia) showed an increased callus density in EPO-treated animals after 2 weeks (161.1±2.9% vs. 122.4±5.4%, p<0.0001) and after 5 weeks (140.3±3.4% vs. 126.2±4.6%, p<0.05) when compared to controls. Biomechanical testing of the femora also revealed a higher bending stiffness at failure of the EPO-treated animals after 2 weeks (42.2± 6.7% vs. 3.4±0.8%, p<0.0001) and after 5 weeks (57.4±9.9% vs. 32.2±4.8%, p<0.05). Histomorphometric analysis after 2 weeks showed bone bridging of the fracture gap in 2/8 animals treated with EPO compared to 0/8 animals in the control group. After 5 weeks all EPO-treated animals showed complete bone bridging (8/ 8), compared to 4/8 animals in the control group (p<0.05). We observed no difference in the callus diameter after 2 weeks (3.8± 0.5 mm vs. 2.9±0.4 mm, p=0.2) and after 5 weeks (2.1±0.1 mm vs. 2.5±0.5 mm, p=0.4) between the two groups. Of interest, after 2 weeks EPO treated animals showed a significantly increased proportion of bone in the periosteal callus area (96.2±2.1% vs. 26.5±6.3%, p<0.0001), and a significantly decreased proportion of cartilage (2.8±2.8% vs. 46.7±13.9%, p<0.01) and fibrous tissue (1.0 ±0.6 vs. 26.8±11.2%, p<0.05). Conclusions: Long-term EPO treatment improves fracture healing in a model of delayed union with an improved union rate and an increased biomechanical stiffness. Our data indicates that the better mechanical stiffness of the healing bone is not due to an increased callus size, but rather due to a better quality of the callus with an increased proportion of bone. Currently, further studies are focusing on the molecular mechanism of EPO in the healing bone. Thus, EPO may represent a novel approach to improve fracture healing in patients with delayed union.

100 EFFECTS OF ANGIOTENSIN-1-RECEPTOR BLOCKADE AFTER RESECTION OF CIRRHOTIC RAT LIVERS Linus Kebschull, Ralf Bahde, Sandra Stöppeler, Jens Peter Hölzen, Evgeni Minin, Hans-Ullrich Spiegel, Daniel Palmes Abteilung Chirurgische Forschung, Klinik und Poliklinik für Allgemeinund Viszeralchirurgie, Universitätsklinikum Münster, Germany Keywords: liver regeneration, cirrhosis, hepatic stellate cells Background: Liver resection in cirrhosis has a poor outcome. In this context, hepatic stellate cells (HSC) play both a key role in development of hepatic cirrhosis and in liver regeneration. Recent studies have shown that HSCs are activated via an Angiotensin-1 (AT) receptor. In this study we investigated the effect of HSC modulation by selective AT-1-receptor blocking. Methods: In 136 male Lewis rats liver cirrhosis was induced by repetitive intraperitoneal CCl4 injection. In half of the animals a 2/3liver resection was performed. In a pilot study different dosages of

Langenbecks Arch Surg (2008) 393:759–815 daily administered Losartan (1, 5 and 10 mg/kg b.w. by gavage) were tested. In the main study the most effective Losartan dosage was investigated in full-size livers (controls) and after 2/3 liver resection. After 6 hours, 3, 7 and 14 days 7 rats per group were sacrificed. At each time point hemodynamic (blood pressure, portal vein flow), morphologic (liver volume, degree of cirrhosis) and serum parameters (liver enzymes) were measured (KruskalWallis-test, p<0.05). Results: In the pilot study a dosage of 5 mg/kg b.w. was most efficient. In the main study after induction of liver cirrhosis all animals survived until 14 days after 2/3 liver resection. Treatment with Losartan reduced liver cirrhosis (preoperatively 2.8±0.3 points to 2.2±0.4 points after 14 days) in contrast to control animals (3.1±0.5 points preoperatively vs. 2.8±0.4 points postoperatively). After liver resection, Losartan-treated rats revealed a significant lowered systemic blood pressure (day 3: 92±13.4 mmHg vs. 188±16.7 mmHg) and elevated portal vein flow (6 hours postoperatively: 11.1±2.6 ml/min vs. 6.1±1.6 ml/min) in comparison to control animals. Liver volume and serum parameters were not affected by Losartan treatment after liver resection. Conclusions: AT1-receptor blockade improves portal vein flow and reduces cirrhosis after liver resection but does not impair regeneration and function in cirrhotic livers.

101 P-SELECTIN GLYCOPROTEIN-LIGAND-1 REGULATES HEPATIC LEUKOCYTE RECRUITMENT DURING CHOLESTATSIS INDUCED LIVER INJURY IN A PLATELET-INDEPENDENT MANNER Stefan Dold (1,2), Matthias W. Laschke (1,3), Sven Richter (2), Martin K. Schilling (2), Michael D. Menger (3), Henrik Thorlacius (1) 1: Department of Surgery, Malmö University Hospital, Malmö, Lund University, Sweden, 2: Department of General-, Visceral-, Vascular and Pediatric Surgery and Institute for Clinical & Experimental Surgery, University of Saarland, Homburg/Saar, Germany Keywords: leukocyte recruitment; cholestasis; PSGL-1; platelet depletion Background: Hepatic leukocyte accumulation represents a key-step in the development of liver injury during the early period of extrahepatic cholestasis. However, the adhesive mechanisms regulating recruitment of leukocytes to the liver remain elusive. Herein, we investigated the role of P-selectin glycoprotein-ligand-1 (PSGL-1) in cholestasis-induced leukocyte recruitment and tissue injury in the liver. Methods: C57BL/6 mice (n = 23) underwent bile duct ligation (BDL) for 12 hours. Mice were pretreated with a control antibody (n = 7), a monoclonal antibody directed against PSGL-1 (n = 6), a platelet-depleting antibody against GP1ba (n = 5) or the combination of the latter two (n = 5). Hepatic leukocyte recruitment and sinusoidal perfusion failure were analyzed by means of intravital fluorescence microscopy. Hepatocellular damage was determined by serum levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST). Moreover, CXC chemokine concentration in the liver tissue was quantified by ELISA. In addition, plateletleukocyte aggregate formation was characterized by FACSanalysis. Results: BDL caused clear-cut liver injury, as indicated by massively increased levels of liver enzymes (ALT: 68,9±5,2 vs. 0,4±0,1 kat/l) and distinct sinusoidal perfusion failure (18±2 vs. 4±1%) when compared to non bile duct-ligated controls. This was associated with formation of leukocyte-platelet aggregates, hepatic leukocyte recruitment and increased levels of CXC chemokines (4659±783 vs. 708±104 pg/g) in the liver tissue. Immunoneutrali-

797 zation of PSGL-1 reduced BDL-induced leukocyte adhesion in postsinusoidal venules by 79% and hepatic MPO activity by 65%. Platelet-leukocyte aggregate formation was also diminished by antiPSGL-1 treatment. Notably, the inhibitory effect of the anti-PSGL-1 antibody on BDL-induced leukocyte infiltration could also be observed in platelet-depleted mice. Moreover, inhibition of PSGL-1 significantly improved the hepatic microcirculation (10±2 vs. 18± 2%) and reduced the serum levels of transaminases (ALT: 14,7±6,1 vs. 68,9±5,2 kat/l) as well as the hepatic accumulation of CXC chemokines (4659±783 vs 1581±319 pg/g). Conclusions: Our novel data demonstrate that PSGL-1 plays a key role in hepatic leukocyte recruitment during obstructive cholestasis. Moreover, our findings suggest that PSGL-1-dependent leukocyte recruitment is independent of circulating platelets. Our findings indicate that PSGL-1 may be a useful target in order to protect against cholestasis-induced liver injury.

102 DIFFERENTIAL ACTIVATION OF DENDRITIC CELL SUBTYPES AFTER MULTIPLE TRAUMA Dirk Henrich, Marcus Maier, Emanuel Geiger, Kerstin Wilhelm, Ingo Marzi Department of Trauma-, Hand- and Reconstructive Surgery, Hospital of the Johann Wolfgang Goethe-University, 60590 Frankfurt/Main, Germany Keywords: dendritic cells Background: Dendritic cells (DC) are potent initiators of inflammation. Two distinct subsets of DC could be identified which differ in many aspects: Myeloid DC (MDC) and plasmacytoid DC (PDC). Chemokines are important factors that regulate amongst others the adhesion and emigration of DC. Our previous serial microarray analysis of circulating DC revealed that the gene expression of the chemokines CCL5, CXCL5, their corresponding receptor CCR1 and the platelet derived factor-4 (PF4) were significantly upregulated after multiple trauma. The concomitant upregulation of CCR1, its ligands and PF4 may influence basic DC functions such as adhesion due to an autocrine activation. Hence, this study was conducted 1) to assess wether the results of the microarray analysis can be confirmed by alternative methods, 2) to analyse these changes on the level of the DC-subtypes, and 3) to analyse if these changes affect the DC’s adhesion to endothelium. Methods: Dendritic cells (DC) are potent initiators of inflammation. Two distinct subsets of DC could be identified which differ in many aspects: Myeloid DC (MDC) and plasmacytoid DC (PDC). Chemokines are important factors that regulate amongst others the adhesion and emigration of DC. Our previous serial microarray analysis of circulating DC revealed that the gene expression of the chemokines CCL5, CXCL5, their corresponding receptor CCR1 and the platelet derived factor-4 (PF4) were significantly upregulated after multiple trauma. The concomitant upregulation of CCR1, its ligands and PF4 may influence basic DC functions such as adhesion due to an autocrine activation. Hence, this study was conducted 1) to assess wether the results of the microarray analysis can be confirmed by alternative methods, 2) to analyse these changes on the level of the DC-subtypes, and 3) to analyse if these changes affect the DC’s adhesion to endothelium. Results: CCR-1 was raised on MDC on d1 after trauma in comparison to volunteers (D1: 31+/−5.1 vs 14+/−3.5, p<0.05). The expression of CD195 on MDC and PDC was comparable to healthy volunteers but declined significantly on d4 after admission (con, MDC: 22+/−4.5 vs. D4: 12+/−5.1; con PDC:22+/−1.2 vs. D4: 11+/−2.8). CCL5 and CXCL5 peptides were not detectable, whereas PF4 in MDC and PDC

798 was significantly raised on D1 and D4 after trauma in comparison to controls (MDC D1: 57+/−5.7; D4: 52+/−7.7 vs. con: 28+/−4.9; value of PDC were comparable). The surface expression of HLA-DR was significantly upregulated on MDC on d1 and d4 after trauma in comparison to healthy volunteers (D1: 227+/−41, D4: 205+/−48 vs con: 84+/−32) whereas the HLA-DR expression on PDC remained comparable to healthy volunteers. Neither the adhesion to endothelium of MDC nor the the adhesion of PDC was altered after multiple trauma. Conclusions: The actual findings confirm our former results of the microarray analysis with respect to CCR1 and PF4. No CCL5 and CXCL5 proteins were found suggesting that their expression is maybe restricted to the mRNA level. The subtype analysis revealed that MDC and PDC are differentially activated after multiple trauma and that the state of activation declines over the time. These changes are not associated with an alteration of the adhesive properties of both MDC and PDC. Taken together, our results indicate a complex regulation of DC subtypes during multiple trauma.

103 IN-VITRO STUDY OF SMALL INTESTINAL MOTILITY IN PATIENTS WITH CHRONIC INTESTINAL PSEUDOOBSTRUCTION Hanno v. Koschitzky, S. Schmidt, D. Svoboda, LM. Wessel, T. Wedel, M. Kohl Department of Pediatric Surgery, University Hospital Luebeck, Germany Keywords: intestinal motility Background: The diagnosis of chronic intestinal pseudoobstruction (CIPO) is based on clinical and radiographic findings. Routine histological stainings of intestinal biopsies are not always appropiate to reveal specific abnormalities. Immunohistochemical markers may help to differentiate between neuropathic and myopathic type of CIPO, but often yield inconsistent findings. In addition to histopathologic studies specimens of infantile CIPO were subjected to functional studies in order to characterize abnormal in-vitro motility patterns. Methods: Ileal full-thickness specimens from 4 girls with CIPO (age range: 1.2–9 y) were compared with corresponding specimens from 11 patients (7 male, 4 female, age range: 2–85 y) without intestinal motility disorders. In-vitro contractility of circular smooth muscle strips was assessed in an organ bath by recording isometric tension forces. Spontaneous activity and contractile responses to acetylcholine (ACh), electric field stimulation (EFS) of enteric nerves under non-cholinergic non-adrenergic (NANC) condition, blockade of NO-production by L-NNA and direct muscular stimulation by KCl were measured. In addition, specimens were processed for both conventional histology and immunohistochemistry using antibodies against alpha-smooth muscle actin (SMA), smooth muscle myosin heavy chain (SMMHC), smoothelin (SM), histone deacetylase 8 (HDAC8), protein gene product (PGP) 9.5, Hu C/D and c-kit. Results: In CIPO patients active muscle tone (2.3±1.6 vs. 4.3±2.6 mN/mm2, p<0.001), amplitude of oscillations (1.4±1.3 vs. 4.4±3.9 mN/mm2, p<0.005) and maximum response to ACh (164±114 vs. 444 ± 162% increase of basal tone, p < 0.001) were decreased. Relaxation after EFS under NANC-condition was less pronounced (43±86 vs. 54±35% of basal tone, p<0.001) and EFS induced relaxation after prior stimulation with KCl or L-NNA was not significantly altered. Contractile response to KCl was markedly decreased in 3 patients and increased in 1 patient. Histopathology of CIPO specimens showed either myopathic (fibrosis, thinned muscularis propria, thickening of muscularis mucosae) or neuropathic (myenteric hypo- or hyperganglionosis, nerve fiber hypertrophy)

Langenbecks Arch Surg (2008) 393:759–815 alterations or both (n=2). The child with an increased contractile response to KCl had marked hypoganglionosis but intact muscle layers. Conclusions: Routine histology in CIPO is often inconclusive, and immunohistochemistry does not always allow to identify the underlying etiology. In contrast, in-vitro motility patterns showed characteristical and uniform abnormalties in CIPO specimens. In particular, the neuropathic type of CIPO could be readily identified by organ-bath technique. Increased inhibitory nervous input as a cause of the motility disorder could be excluded. Once full-thickness biopsies are available, we therefore recommend to perform in-vitro motility recordings in addition to histopathologic studies in those cases of suspected CIPO in which a definitive diagnosis can not be made on clinical grounds.

104 ANALYSIS OF ERCC1 GENE POLYMORPHISM BUT NOT OF PROTEIN EXPRESSION PREDICTS RESPONSE TO NEOADJUVANT RADIOCHEMOTHERAPY IN ESOPHAGEAL CANCER Ute Warnecke-Eberz, Elfriede Bollschweiler, Fabian Kütting, Hakan Alakus, Daniel Vallböhmer, Uta Drebber, Jan Brabender, Stefan Mönig, Arnulf H Hölscher, Ralf Metzger Visceral and Vascular Surgery, and Institute for Pathology, University of Cologne, Germany Keywords: ERCC1, Single nucleotide polymorphism, multimodality treatment, esophageal cancer Background: Neoadjuvant treatment strategies have been developed to improve survival of patients with locally advanced esophageal cancer. Since only patients with major histopathological response benefit from this treatment, predictive markers indicating response or non-response are needed. A panel of 17 gene polymorphisms was analyzed to predict response to neoadjuvant radiochemotherapy (cis-platin,5FU,36 Gy) of esophageal cancer patients. Methods: Genomic DNA was extracted from paraffin-embedded tissues of 52 patients. Allelic discrimination was performed by quantitative real-time PCR using two allele-specific TaqMan probes in competition for amplification of each gene. Allelic genotyping was correlated with therapy response. ERCC1 Protein was immunohistochemically stained in pretherapeutic biopsies and resected tissues. Results: ERCC1 single nucleotide polymorphism (SNP) A/G (rs11615) was predictive for therapy response (p<0.003). Within the AA genotype group of 25 patients 20 (80%) did not respond to chemoradiation, whereas 14 patients (70%) of 20 patients with the heterogenous A/G genotype were major responders. The GG genotype comprising 7 patients was not of predictive importance. The ERCC1 polymorphism was significantly (p<0.02) associated with formation of lymph node metastases. The three genotypes did not show different levels of ERCC1 protein expression. Immunohistochemically detected expression of ERCC1 protein was not predictive for therapy response. Conclusions: Our data strongly support the role of ERCC1 as a predictive marker for therapy response. Single nucleotide polymorphisms of ERCC1 could be applied to further individualize treatment of patients with locally advanced esophageal cancer

105 EXPRESSION OF ESTROGEN RECEPTOR BETA (ER BETA) IN PANCREATIC ADENOCARCINOMA Hendrik Seeliger, Nina Seel, Ivan Ischenko, Gerald Assmann*, KarlWalter Jauch, Christiane J. Bruns Chirurgische Klinik und Poliklinik und Pathologisches Institut (*), Klinikum der Universität München, München, Germany

Langenbecks Arch Surg (2008) 393:759–815 Keywords: pancreatic cancer, estrogen receptor expression, multivariate survival analysis Background: The role of the expression of estrogen receptors in pancreatic adenocarcinoma is largely unknown. Historic trials aimed to establish selective estrogen receptor modulators (SERMs) in pancreatic cancer did not show any benefit of this therapeutic approach. Here, we analyze the impact of the recently characterized estrogen receptor isoform ER beta on overall survival and time to tumor progression in a cohort of patients with resected pancreatic cancer. Methods: Ninety-four patients having undergone pancreatic resection for pancreatic ductal adenocarcinoma between 2003 and 2006 were identified. Tissue microarrays were constructed of archival tumor specimens. The expression of ER beta was determined immunohistochemically and quantified using a system established for estrogen receptor expression in breast cancer. ER beta expression was then correlated with clinicopathological parameters, and univariate and multivariate survival analyses were performed. Results: Nuclear expression of ER beta was found in 33% of the tumors. No significant correlation was found between ER beta expression and TNM status, tumor grade, age or sex. Univariate analysis revealed nodal metastasis and the expression of ER beta as factors correlating with a shorter overall survival and disease free survival. When comparing ER beta expression in patients surviving more than 24 months with those who died from the tumor within 12 or 24 months, respectively, a significantly lower ER beta expression was found in the long term survivors. In multivariate analysis, nodal metastasis and ER beta expression were shown as independent predictors of poor prognosis. Conclusions: In resected ductal pancreatic adenocarcinoma, expression of ER beta seems to correlate with poor prognosis. These data may help in identifying patients who benefit from additional therapeutic regimens, including selective estrogen receptor modulators.

106 IN VIVO REFLECTANCE MODE CONFOCAL-LASER-SCANNING MICROSCOPY CONTRIBUTES TO A BETTER UNDERSTANDING OF PATHOPHYSIOLOGY IN BURN SHOCK ON MICROCIRCULATORY AND CELLULAR LEVEL Mehmet Ali Altintas, Altintas AA (**), Knobloch K (*), Busch K (*), Rennekampff O (*), Vogt PM (*) *: Department of Plastic, Hand and Reconstructive Surgery, Medical School Hannover, 30625 Hannover, Germany, **: Department of Plastic and Reconstructive Surgery, Cologne-Merheim University of Witten-Herdecke, Cologne, Germany Keywords: Confocal laser scanning microscopy; noninvasive imaging; optical biopsy; burn shock, histomorphology Background: Shock induced microcirculatory disturbances are well known, however, their direct influence on cutaneous micromorphology is widely unknown. In vivo confocal-laser scanning microscopy allows real-time non- invasive evaluation of cutaneous histomorphology with a high cellular resolution. We observed microcirculatory and cellular alterations in burn shock using confocal-laser-scanning microscopy. Methods: Confocal-laser-scanning microscopy was performed on an unburned skin site in 10 burn shock patients (4 female, 6 male; aged 40.6 ±11.4 years, burn extent >20% BSA) prior and after fluid replacement. Ten matched hemodynamic stable ICU patients served as controls. The following parameters were evaluated: quantitative blood cell flow, cell size of the granular-layer, thickness of the basal-layer, thickness of the epidermis. Results: Quantitative blood cell flow was 62.45±3.39/min in controls, decreased significantly in burn shock (37.27±3.64/min) and increased (65.18±3.76/min;P<0.05) after fluid replacement. The granular cell size was 793.61±41.58 µm2 (controls), pointed in burn shock at 644.27± 42.96 µm2, and increased after fluid replacement of up to 932.74±

799 38.83 µm2. The basal layer thickness was in controls at 14.84±0.59 µm, decreased (13.26±0.54 µm) prior- and increased after fluid substitution (17.50±0.46 µm). The epidermal thickness was in controls at 49.60± 2.36 µm, decreased prior fluid replacement to 37.83 ±2.47 µm and increased after fluid substitution (69.50±3.18 µm). The granular cell size, epidermal and basal layer thickness prior fluid replacement differed significantly to both, the controls and the values after resuscitation. Conclusions: Confocal-laser-scanning microscopy appears to be a useful non invasive tool for detecting microcirculatory alterations and their influence on the cutaneous morphology on cellular and subcellular level and may help to consider the efficiency of the specific treatment course in real time.

107 PROTECTIVE ROLE OF TLR7-AGONIST IN MURINE POLYMICROBIAL SEPSIS Pia Körner, Hendrik Mehmcke, Katharina Cziupka, Wolfram Keßler, Tobias Traeger, Alexandra Westerholt, Gunther Hartmann, ClausDieter Heidecke, Stefan Maier Klinik und Poliklinik für Chirurgie, Abteilung für Allgemeine Chirurgie, Viszeral-, Thorax- und Gefäßchirurgie, Ernst Moritz Arndt Universität Greifswald, Institut für Klinische Pharmakologie, Rheinische Friedrich-Wilhelms-Universität Bonn, Germany Keywords: toll like receptor, polymicrobial sepsis, Colon Ascendens Stent Peritonitis, mice Background: Toll-like receptor 7 (TLR7) is considered to serve as receptor for viral recognition patterns. We now found evidence that TLR7 seems to play an important role in postoperative polymicrobial infections. Therefore, TLR-agonists may provide a new possibility in prophylaxis of postoperative sepsis. Methods: We used Colon Ascendens Stent Peritonitis model (CASP) in mice pre-treated with TLR7-agonist (R-848, resiquimod). We determined survival rates over a period of 10 days post infection. Cytokine levels and bacterial load were detected in spleen and peritoneum 12 h after induction of CASP. Results: In polymicrobial sepsis model TLR7-agonist pre-treated mice showed increased probability of survival. Furthermore, bacterial clearance in polymicrobial sepsis was significantly increased in spleen and peritoneum of mice pre-treated with TLR7-agonist. In these mice TNF-α release was decreased in spleen whereas TNF-α levels were elevated in the peritoneal cavity when compared to nonpre-treated septic mice. Conclusions: In conclusion, TLR7 seems to be essential for pathogen defence not only in viral but also in bacterial infections. Pharmacological stimulation of this receptor supports the infected organism in coping with pathogens and could be a prophylactical or potentially therapeutic approach for severe infectious diseases.

108 CCL-3 CONCENTRATIONS IN MONOCYTE MRNA AND SERUM OF MULTIPLE INJURED PATIENTS: A COMPARATIVE ANALYSIS BETWEEN GENOME RESPONSE AND SERUM PROTEIN LEVELS Peter Richter, V. Bogner, W. Mutschler, P. Biberthaler Chirurgische Klinik und Poliklinik Innenstadt, Ludwig Maximilians University Munich, Germany Keywords: multiple injuries, immune response, CCL-3 Background: Alterations of the monocyte function seem to advantage the development of SIRS (Systemic Inflammatory Response Syndrome) and consecutive multiple organ failure in multiple trauma

800 patients. However, thereby underlying initial intracellular mechanisms are not well understood yet. In our precedent genome-wide microarray monocyte mRNA profiling showed a significant up-regulation of CCL3 (Mip-1-beta), a protein that is involved in the MAP-Kinase Pathway, especially in patients with adverse outcome. However, it remains unclear if this mRNA up-regulation also ends up in higher CCL-3 protein concentrations in the serum of multiple trauma patients. Methods: 41 patients with multiple injuries ISS >16 (Injury Severity Score) were included. Serum samples were taken on admission and 6 h, 12 h, 24 h, 48 h and 72 h after trauma. The patient collective was divided into two groups according to their clinical outcome after 90 days. Serum samples were tested for MIP-1 beta concentrations using a human cytokine panel (Microbionix, Germany). Statistics were performed by ANOVA on Ranks, followed by Dunn’s post-hoc test (Sigma Stat, SpSS). Results: 33 patients survived the trauma, 8 patients had an adverse outcome. CCl-3 serum concentrations were elevated in all patients exhibiting a peak at 12 h (ANOVA on Ranks, p<0.05; 12 h vs. 48 h/ 72 h) and a consecutive continuous decrease after 48 h (ANOVA on Ranks, p<0.05; 0 h vs. 48 h/72 h). Though there is a slight tendency of lower CCL-3 concentration in patients who deceased, no significant differences in CCL-3 levels depending on the patients’ clinical outcome can be observed. Conclusions: On the basis of our genome wide gene expression profiling study in monocytes of multiple injured patients we thereby present a follow up CCL-3 investigation study in serum of 41 multiple injured patients in the early posttraumatic period. In contrast to our precedent genome investigations that exhibited elevated CCL-3 expression in patients with adverse outcome, CCL-3 serum protein concentrations are not outcome-dependent. Further studies to clarify the apparent discrepancy between serum-protein concentrations and mRNA monocyte levels have currently been initiated.

109 KONTINUIERLICHE MESSUNG DER INTESTINALEN PERMEABILITa´T MIT FITC-INULIN ALS PARAMETER DES MOV IM EXPERIMENTELLEN MODELL Dirk Jargon, V. Friebe, O. Sommer, T. Keck, U.T. Hopt, E. von Dobschütz Abteilung für Allgemein- und Viszeralchirurgie, Chirurgische Universitätsklinik der Albert-Ludwigs-Universität, Freiburg, Germany Keywords: Sepsis, Intestinale Permeabilität, LPS, Ratte Background: Die Inzidenz des Multiorganversagens (MOV) bei grossen Bauch- und Gefässeingriffen beträgt 9% bis 25%. Nach der Darm-Sepsis-Hypothese wird ein Zusammenhang zwischen der Aktivierung des Immunsystems mit gesteigerter intestinaler Permeabilität (IP) und konsektutiver bakterieller Translokation vermutet. Zahlreiche tierexperimentelle Untersuchungen konnten zeigen, dass ein Multiorganversagen mit einer gesteigerten IP verbunden ist. Die Messung der IP in vivo erfolgte bislang zu einzelnen Messzeitpunkten mittels eines 5 h-Sammelurins. Wir haben im Gegensatz zu bislang verwendeten Versuchsmodellen ein Tiermodell entwickelt, mit dem eine kontinuierliche Messung der IP in vivo (parazelluläre Permeabilität von FITC-Inulin) 1 h vor Sepsisinduktion durch LPS, nach 1; 2; und 3 h sowie nach 24 h nach LPS-Gabe möglich ist. Methods: Es wurden drei Gruppen gebildet: eine Kontrollgruppe (n=6), eine zweite Gruppe (n=6), in der 1 h vor und 3-mal stündlich nach LPS-Gabe gemessen wurde. In einer weiteren Gruppe (n=6) erfolgte die Messung nach 24 h. Wir induzierten in männlichen Sprague-Dawley-Ratten (350–450 g) eine Endotoxinämie durch intraperitoneale Applikation von LPS (1 mg/kgKG). Über einen nach Oesophagotomie gelegten Duodenalkatheter wurde FITC-Inulin (250 mg /60 ml NaCl) enteral kontinuierlich appliziert (0,1 ml/min).

Langenbecks Arch Surg (2008) 393:759–815 Gemessen wurde FITC-Inulin im Urin mittels Fluoreszenzspektroskopie (Exzitation 490 nm, Emission 518 nm). Zusätzlich erfolgte die Bestimmung von Laktat und Glucose im Plasma sowie Urinvolumen und Kreatinin-Clearance. Als proinflammatorischer Parameter wurde IL6 im Plasma mittels ELISA gemessen. Die Stabilität der Versuchstiere wurde durch kontinuierliche Messungen von Blutdruck und stündliche BGA-Analysen überprüft. Results: Die FITC-Inulin-Konzentrationen im Urin 3 h und 24 h nach Sepsisinduktion stiegen auf 407±33 units und 508±108 units im Vergleich zum Wert vor Sepsisinduktion (125±25 units) an. In der Kontrollgruppe wurde nach 3 h ein Wert von 342±33 units gemessen. Die Kreatinin-Clearance der 24 h-LPS-Gruppe fiel im Vergleich zur Kontrollgruppe und der 3 h-LPS-Gruppe signifikant ab. Die Urinvolumina der 3 h-LPS-Gruppe waren im Vergleich zur Kontrollgruppe nicht unterschiedlich, wohingegen die Urinvolumina der 24 h-LPS-Gruppe überproportional anstiegen. IL-6 als proinflammatorischer Marker ist in der 24 h-LPS-Gruppe signifikant erhöht im Vergleich zur Kontrollgruppe und 3 h-LPS-Gruppe (1924±448; 324±55; 677±68 pg/ml). Conclusions: Mit diesem Versuchsmodell ist eine kontinuierliche Messung der IP in vivo vor und nach Sepsisinduktion möglich. Potentielle Interventionsmöglichkeiten zur akuten Stabilisierung der IP können somit überprüft werden.

110 DIFFERENTIAL ACTIVATION OF DENDRITIC CELL SUBTYPES AFTER MULTIPLE TRAUMA Dirk Henrich, Maier, Emanuel Geiger, Kerstin Wilhelm, Ingo Marzi Department of Trauma-, Hand- and Reconstructive Surgery, Hospital of the Johann Wolfgang Goethe-University, 60590 Frankfurt/Main, Germany Keywords: denfritic cell, chemokine, multiple trauma Background: Dendritic cells (DC) are potent initiators of inflammation. Two distinct subsets of DC could be identified which differ in many aspects: Myeloid DC (MDC) and plasmacytoid DC (PDC). Chemokines are important factors that regulate amongst others the adhesion and emigration of DC. Our previous serial microarray analysis of circulating DC revealed that the gene expression of the chemokines CCL5, CXCL5, their corresponding receptor CCR1 and the platelet derived factor-4 (PF4) were significantly upregulated after multiple trauma. The concomitant upregulation of CCR1, its ligands and PF4 may influence basic DC functions such as adhesion due to an autocrine activation. Hence, this study was conducted 1) to assess wether the results of the microarray analysis can be confirmed by alternative methods, 2) to analyse these changes on the level of the DC-subtypes, and 3) to analyse if these changes affect the DC’s adhesion to endothelium. Methods: Blood samples were obtained from 10 Multiple trauma patients (ISS 35.4+/−10.6 points) on day 1 and day 4 after admission. As control, 10 healthy volunteers were also included. Peripheral blood mononucleated cells (PBMC) were stained for CCR1, CD195 (MIP1alpha receptor) and HLA-DR. PBMC were incubated 4 h in presence of Brefeldin-A, supplemented with or without PMA [10^−6 M] for intracellulary detection of CLC5, CXCL5 and PF4. Myeloid DC and PDC were identified as lineage negative, CD11c+ or, respectively, as lineage negative, CD303 + PBMC. Mean fluorescence intensities (MFI) were presented. Adhesion assay: PBMC were incubated for 1 h on a HUVEC-monolayer. Adhering cells were counted and subsequentely the distribution of DC subtypes was assessed by flowcytometry. The study was approved by the local ethics committee. Kruskal-Wallis Test with Dunn’s post hoc analysis was applied for statistical analysis. Results are presented as mean and standard errror of mean. A p<0.05 was significant. Results: CCR-1 was raised on MDC on d1 after trauma in comparison to volunteers (D1: 31+/−5.1 vs 14+/−3.5, p<0.05). The expression of

Langenbecks Arch Surg (2008) 393:759–815 CD195 on MDC and PDC was comparable to healthy volunteers but declined significantly on d4 after admission (con, MDC: 22+/−4.5 vs. D4: 12+/−5.1; con PDC:22+/−1.2 vs. D4: 11+/−2.8). CCL5 and CXCL5 peptides were not detectable, whereas PF4 in MDC and PDC was significantly raised on D1 and D4 after trauma in comparison to controls (MDC D1: 57+/−5.7; D4: 52+/−7.7 vs. con: 28+/−4.9; value of PDC were comparable). The surface expression of HLA-DR was significantly upregulated on MDC on d1 and d4 after trauma in comparison to healthy volunteers (D1: 227+/−41, D4: 205+/−48 vs con: 84+/−32) whereas the HLA-DR expression on PDC remained comparable to healthy volunteers. Neither the adhesion to endothelium of MDC nor the the adhesion of PDC was altered after multiple trauma. Conclusions: The actual findings confirm our former results of the microarray analysis with respect to CCR1 and PF4. No CCL5 and CXCL5 proteins were found suggesting that their expression is maybe restricted to the mRNA level. The subtype analysis revealed that MDC and PDC are differentially activated after multiple trauma and that the state of activation declines over the time. These changes are not associated with an alteration of the adhesive properties of both MDC and PDC. Taken together, our results indicate a complex regulation of DC subtypes during multiple trauma.

111 IMPACT OF ENTEROCOCCUS SSP. ON MORTALITY AND MORBIDITY OF PATIENTS WITH PERFORATION OF THE SMALL AND LARGE BOWEL. RETROSPECTIVE ANALYSIS ON 473 PATIENTS Magnus Kaffarnik, Max Urban, Stefan Utzolino, Ulrich Theodor Hopt Department of General Surgery, Universityhospital Freiburg, Germany Keywords: enterococcus, abdominal perforation, peritonitis Background: Among other bacteria Enterococcus ssp. are frequently detected in abdominal fluid after bowel perforation, as well as after major abdominal surgergy. The commonly administered antibiotics are not or just partially effective against Enterococcus ssp. The need for a specific treatment of Enterococcus ssp. is discussed controversely. Agents that are effective ex-vivo penetrate tissue insufficiently or are very expensive. Based on this background and the growing number of vancomycine resistant Enterococci there is demand to identify patients after abdominal surgery who could benefit from therapy. Methods: We included 473 patients with intestinal perforation. The patients were divided into three different groups, 1.: Enterococcus ssp. not detected, no therapy, 2.: Enterococcus ssp. detected, no therapy, and 3.: Enterococcus ssp. detected, specific therapy. We compared the groups and analyzed how far age, sex, immunosuppression, and detection of Enterococcus ssp. influence the course of the disease. Primary endpoint of the study was lethality. Secondary endpoints were morbidity, LOS-ICU and LOS-Hospital. The severity of the disease was assessed using the SOFA-Score (Sepsis-related Organ Failure Assessment). Results: 90-day-mortality was significantly higher in group 3 compared with group 1 and 2 (p<0,001). There was no difference between group 1 and group 2. LOS-ICU, LOS- Hospital and SOFAscore were significantly higher in group 3 than in group 1 or 2. Even though we did not find a significant difference in mortality between enterococci negative and enterococci positive (not treated) patients, there is strong evidence that enterococci cause postoperative complications and prolong the ICU-LOS and hospital-LOS. Enterococcus ssp. were not an independent risk factor for mortality. Age, peritonitis, anastomosis leackage and immunosuppression were independent risk factors for mortality. Conclusions: Enterococcus ssp. detected in peritoneal fluid after major abdominal surgery are often due to colonization and do not warrant specific treatment. Nevertheless, Enterococci occurring in peritoneal

801 fluid become increasingly relevant in critically ill patients, as they can worsen the prognosis. Therefore, in critically ill and immunocompromised patients specific treatment of Enterococcus ssp. must be considered. Because of the higher complication rate in enterococci positive patients, physicians should be aware of the right time to start specific anti-enterococcal therapy.

112 TNF-RELATED APOPTOSIS-INDUCING LIGAND (TRAIL) ENHANCES BACTERIAL CLEARANCE AND SURVIVAL IN MURINE POLYMICROBIAL SEPSIS Katharina Cziupka, Alexandra Westerholt, Tobias Traeger, Lars Ivo Partecke, Christian Poetschke*, Pia Koerner, Wolfram Kessler, Stephan Diedrich, Wolfram v. Bernstorff, Barbara Broeker*, Stefan Maier, Claus-Dieter Heidecke Department of Surgery, * Institute of Immunology, University of Greifswald, Germany Keywords: TRAIL, sepsis, CASP Background: TNF-related apoptosis-inducing Ligand (TRAIL) is a member of the TNF superfamily and is currently being evaluated as one of the most promising natural immune modifier molecules proposed for trials in cancer patients. Apart from inducing apoptosis in transformed cells, there is growing evidence that endogenous TRAIL exhibits physiological functions in several other immunological reactions. The aim of this study was to investigate the role of TRAIL in the colon ascendens stent peritonitis (CASP) as an established model for postoperative polymicrobial sepsis. In this model a stent is inserted into the ascending colon, thereby generating a septic focus. The CASP model closely mimics the clinical course of human diffuse peritonitis with early and steadily increasing systemic infection and inflammation. Methods: For all experiments C57BL/6 mice were used. For induction of CASP a stent was surgically inserted into the ascending colon of mice. The resulting leakage of gut bacteria into the peritoneal cavity lead to peritonitis and consecutive sepsis. Recombinant TRAIL was intravenously injected at a concentration of 1g TRAIL per 1 g mice. It was given 3 hours, 24 hours and 48 hours after induction of CASP. Controls received equal volumes of sterile saline solution. We examined the survival rates of mice as well as bacterial colonisation of different organ systems. The local and systemic cytokine response was detected by cytometric bead arrays 20 hours after induction of CASP. Apoptosis in the thymus was detected using the propidium iodide cell cycle assay. The expression of TRAIL receptors 2, 3 and 4 was determined in the spleen and the blood of naïve, septic and TRAIL-treated septic mice using FACS and western blot analyses. Surface bound TRAIL was examined in naïve and septic mice by FACS analyses. 12 hours after CASP peritoneal lavage fluid was collected and cells were counted. FACS analyses of the cells from peritoneal lavage fluid were performed. Results: TRAIL treatment led to a significantly better survival of mice in polymicrobial sepsis. TRAIL treated animals showed a significantly enhanced bacterial clearance within the peritoneal cavity, the blood as well as all other investigated organs. At the same time, the fraction of apoptotic cells of septic mice within their thymus was significantly decreased. Interestingly, we could not detect any significant change in cytokine levels neither in serum nor in supernatants of all tested organs. In contrast, expression of the TRAIL-receptor 2 (DR5) in the spleen and the blood was significantly increased in septic animals when compared to naïve mice. However, TRAIL-receptors 3 and 4 were not detectable in any settings. The up-regulation of DR5 during sepsis could be reversed by treating mice with TRAIL during sepsis. Furthermore, TRAIL expression on the cell surface of immune cells was significantly up-regulated in the spleen and in the blood during sepsis. Surprisingly, cell counts were significantly elevated within the peritoneal lavage fluid of TRAIL-treated septic mice

802 when compared to placebo-treated septic mice. TRAIL treatment led to elevated counts of granulocytes, macrophages and dendritic cells within the peritoneal lavage fluid of septic mice. Furthermore, the injection of TRAIL into the peritoneal cavity of naïve mice led to a massive influx of immune cells, particularly macrophages and dendritic cells. Conclusions: Treatment of septic mice with TRAIL significantly improved survival of CASP. This is due to a strongly enhanced bacterial clearance. In TRAIL treated animals the fraction of apoptotic cells within the thymus is significantly decreased indicating a milder course of sepsis. Our results suggest that the enhanced bacterial clearance following TRAIL treatment is caused by an increased influx of effector cells and professional antigen presenting cells of the innate immune system to the site of infection. Our results demonstrate a new feature of TRAIL: TRAIL can also act as a regulator of systemic inflammation. Therefore, TRAIL may become a new therapeutic modality in the treatment of sepsis.

113 SEVERE NECROTIZING PANCREATITIS INDUCED BY TAUROCHOLATE IN MICE Uwe A Wittel, Ulrich T Hopt Abt. fuer Allgemeinchirurgie und Visceralchirurgie, Uniklinik Freiburg, Germany Keywords: pancreatitis, sirs Background: Lethality of acute pancreatitis depends on the severity of systemic complications. So far, the induction of severe necrotizing pancreatitis in mice, where both wild-type and genetically engineered animal models can be used is difficult. We developed and characterized a mouse model of severe necrotizing pancreatitis which is based on the method of taurocholate induced pancreatitis in rats. Methods: Pancreatitis was induced by retrograde infusion of sodium taurocholate into the common bile duct. In a first step the optimum volume and concentration of the injected sodium taurocholate solution was determined. In a second step, pancreatic damage and systemic inflammatory response in taurocholate pancreatitis were compared with cerulein-induced pancreatitis, the most frequently used model. Results: Morphologic pancreatic damage was higher in taurocholate pancreatitis than cerulein pancreatitis (24 hours: cerulein, 5.8+/−0.2 points; taurocholate, 14.8+/−0.8 points; P<0.001). Mortality reached up to 60% within the first 24 hours after taurocholate administration when 5% taurocholate were used. Pulmonary damage was quantified by an increase in albumin in bronchoalveolar lavage fluid and was only detected in taurocholate-induced pancreatitis (12 hours: cerulein, 97.1+/− 22.83 mg/g of protein; taurocholate, 234.0+/−32.7 mg/g of protein; P<0.001). In addition, plasma interleukin 6 levels were significantly elevated in mice with taurocholate-induced pancreatitis when compared to the other experimental groups (12 hours: cerulein, 2.6+/−6.1 pg/mL; taurocholate, 2168.8+/−941.7 microg/mL; P<0.001). Conclusions: Taurocholate pancreatitis in mice has proven to be a reliable model for severe necrotizing pancreatitis with significantly greater pancreatic systemic damage than cerulein-induced pancreatitis.

114 CHONDROGENESIS OF ADIPOSE-DERIVED ADULT STEM CELLS IN A POLY-LACTIDE-CO-GLYCOLIDE (PLGA) SCAFFOLD Alexander Mehlhorn, Jörn Zwingmann, Phillip Niemeyer, Norbert Südkamp, Hagen Schmal Department für Orthopädie und Traumatologie Freiburg, Germany Keywords: cartilage repair, mesenchymal stem cells, polymer scaffolds

Langenbecks Arch Surg (2008) 393:759–815 Background: Adipose-derived adult stem cells (ASCs) are considered as an alternative cell source for cell-based cartilage repair due to their multiple differentiation potentials. This article addresses the chondrogenic differentiation of ASCs seeded into poly-lactide-co-glycolide (PLGA) scaffolds following implantation in a subcutaneous pocket of nude mice. Methods: Human ASCs were seeded into PLGA (PLA:PGA=90:10) scaffolds and cultured in Transforming Growth Factor ß1 (TGFβ1)-containing medium for three weeks in vitro. Then, specimen were implanted into a subcutaneous pocket of SCID mice and harvested after 8 weeks. Chondrospecific mRNA expression was analyzed using RT-PCR. Corresponding extracellular matrix synthesis was demonstrated using immunohistochemical staining. The biomechanical properties of the explants were tested with a confined compression test. Results: Chondrospecific marker molecules like collagen type II, type X, Cartilage Oligomeric Matrix Protein (COMP) and aggrecan subsequently increased during the 3 weeks period in vitro. After further 8 weeks in vivo samples pre-treated with TGF-β1 continued expressing collagen type II and aggrecan mRNA and collagen type II was demonstrated within the extracellular matrix by immunohistochemistry. In control samples chondrospecific mRNA was not detected. Conclusions: ASCs-seeded PLGA scaffolds express a stable chondrogenic phenotype in a heterotopic model of cartilage transplantation and represent a suitable tool for tissue engineering of cartilage.

115 DE NOVO SIGNIFICANT ROLE OF PANCREAS DERIVED STEM CELLS IN THE IN VIVO EPIDERMALIZATION OF SKIN WOUNDS Haitham Salem (1,2), D.H. Rapoport (3), P. Ciba (3), J.T. Egana (1), M. Kadry (2), K. Reithamyer (4), C. Kruse (3), H.G. Machens (1) 1: Dept. of Plastic Surgery, TU Munich, Germany, 2: Dept. of Plastic Surgery, Kasr Alainy Faculty of Medicine (KAFM), Cairo University, Egypt, 3: Fraunhofer Institute for Marine Biotechnology (EMB), Campus Lübeck, Germany, 4: Dept. of Dermatology, Venerology and Allergology, University of Lübeck, Germany Keywords: Tissue Engineering, Skin Regeneration, Stem cells Background: In the rapidly growing field of stem cells, pancreas derived stem cells (PSCs) have been explored mainly with respect to their therapeutic potential in the treatment of diabetes. Different research groups have reported that PSCs differentiate not only in pancreatic cell types, but in a multitude of different cells, including ectodermal cell types. In this study, we present the first in vivo attempt to use PSCs in the field of skin regeneration. Methods: 5×105 PSCs from wistar rats were used to bioactivate Matriderm® sheets. Cell survival and proliferation were tested in vitro after an over night incubation with the scaffolds to ensure their survival and homogenous distribution. The bioactivated scaffolds were then used in bilateral full-thickness skin wounds made on the dorsum of nude mice. A control group of nude mice received the Matriderm® scaffolds without cells. Two weaks after transplantation the wound areas were analyzed with respect to epithelialization, vascularization and wound closure. Results: After two weeks; the healed area and regeneration rate were significantly increased in the group of PSC-covered scaffolds (factor of 2.1). Transillumination of the wound area and vessel quantification showed a significant increase in the vascularization percentage with the bioactivated scaffolds (factor of 1.5). Skin-like structures positive for epidermal markers (K14, K10 & Fillagrin) were evident in the healing wound bed via morphology and immunohistochemistry. PSCs were detected in the healing wound tissues using PCR. In vitro cytokine arrays of PSC-bioactivated scaffolds showed that the cells

Langenbecks Arch Surg (2008) 393:759–815 produce a significant amount of inflammatory cytokines responsible for immune defense (GM-CSF, IL-1 alpha and IL-6) together with the VEGF signaling for both vasculogenesis and angiogenesis. Conclusions: In our study we prove that the abilities of PSCs extend beyond the replacement of beta cells. They improve epidermalization percentage, vascularization and thus healing of full-thickness skin wounds.

116 USE OF PANCREAS-DERIVED STEM CELLS TO IMPROVE VASCULARIZATION IN SKIN TISSUE ENGINEERING Egaña J.T. (1,2,3), Danner S. (4), Kremer M. (1), Rapoport D.H. (4), Lavandero S. (3), Kruse C. (4), Machens H.G. (1,2). 1: Dept. Plastic and Hand Surgery, University of Lübeck, Germany, 2: Dept. Plastic and Hand Surgery, Technische Universität München, Munich, Germany, 3: FONDAP Center for Molecular Studies of the Cell. University of Chile, Santiago, Chile, 4: Fraunhofer-Institute of Biomedical Engineering, Group of Cell Differentiation and Cell Technology, Lübeck, Germany Keywords: Scaffolds, vascularization, stem cells, tissue engineering Background: Clinical success in skin tissue engineering requires improvements in the vascularization capacity of scaffolds for dermal regeneration (SDR). Several efforts have been done in this field including cellular and acellular technologies. Methods: In this work we combined the use of GFP+ Pancreasderived stem cells (PdSCs) with a commercially available SDR. 1× 106 cells were isolated, characterized, seeded and cultured in a 15 mm SDR. Cell viability, proliferation and cytokine release from scaffolds containing cells was assayed in vitro. In vivo, scaffolds containing cells were used to induce dermal regeneration in a 15 mm full skin defect model. After three weeks of in vivo regeneration tissues were harvested and vascularization analyses were performed. Results: Experimental data showed that PSCs can survive and proliferate in the scaffold, releasing significant amounts of cytokines. In vivo, the presence of PSCs significantly enhances the vascularization levels compared to empty scaffolds (p≤ 0,05). Presence of PSCs within the tissue in regeneration was confirmed by detection of GFP+ cells in the neodermis. In vitro PdSC were able to from capillary like structures and induce transmigration of bone marrow endothelial progenitor cells, suggesting and possible angiogenic and vasculogenic contribution of the cells in vivo. Conclusions: These results suggest that combined use of PSCs and SDR could be a rational way to improve vascularization in scaffolddependent dermal regeneration. However, further studies must be performed in order to better analyze the cellular and molecular mechanisms involved in the contribution of PSCs and its possible physiological role in wound healing and tissue regeneration.

117 IN VIVO ANALYSIS OF BIOCOMPATIBILITY AND VASCULARIZATION OF ELASTIC POLYURETHANE SCAFFOLDS FOR TISSUE ENGINEERING Anja Strohe, Matthias W. Laschke, David Eglin, Sophie Verrier, Mauro Alini, Michael D. Menger Institute for Clinical & Experimental Surgery, University of Saarland, Homburg/Saar, Germany; AO Research Institute, Clavadelerstrasse, Davos, Switzerland Keywords: tissue engineering, scaffolds, polyurethane, biocompatibility, angiogenesis, vascularization Background: In tissue engineering, scaffolds provide a matrix for cells to attach and proliferate that can be implanted into a tissue defect site.

803 For this purpose, the scaffolds should exhibit a good biocompatibility. Moreover, they should stimulate the fast ingrowth of new blood vessels after implantation in order to guarantee the survival and longterm function of the implanted cells. In the present study, we analyzed in vivo the biocompatibility and vascularization of a novel type of polyurethane scaffolds. These non-toxic scaffolds have been developed for tissue engineering by introducing labile units into the stable polyurethane chains, ensuring an almost frictionless integration into the host tissue due to the elastic material properties of these modified polyurethanes. Methods: A dorsal skinfold chamber was prepared in 28 balb/c mice for the implantation of three different polyurethane scaffolds (3×3× 1 mm), i.e. EGD (n=10), GRS-13 (n=9) and 070724 (n=9). Using the technique of intravital fluorescence microscopy we repetitively analyzed vascularization of the implants and venular leukocyteendothelial cell interaction in the surrounding host tissue over a time period of 14 days. Subsequently, incorporation of the scaffolds was analyzed by histology. Moreover, we performed a WST-1 assay to analyze in vitro the biocompatibility of the different scaffold types. Results: We could demonstrate by WST-1 assay that all three types of polyurethane scaffolds exhibited a good biocompatibility. Accordingly, also in vivo the numbers of rolling and adherent leukocytes in venules of the dorsal skinfold chamber were found in a physiological range (∼17–21 cells/min and ∼100–200 cells/mm2) and did not significantly differ between the observation groups. Interestingly, implantation of EGD, GRS-13 and 070724 induced a poor angiogenic response with a microvessel density of only ∼47–60 cm/cm2 and ∼3–10 cm/cm2 in the border and center zones of the scaffolds at day 14 after implantation. Histology confirmed our intravital microscopic findings. After 14 days, the scaffolds were incorporated in a granulation tissue, which showed only a few newly formed blood vessels. Conclusions: In the present study we could demonstrate that the novel elastic polyurethane scaffolds EGD, GRS-13 and 070724 exhibit a good biocompatibility and, thus, may be used to generate tissue constructs which do not induce a strong inflammatory reaction after implantation into patients. However, the scaffolds should be further modified in order to accelerate and improve the process of vascularization.

118 POSSIBLE ROLE OF INTRAARTICULAR CYTOKINES IN NATURAL AND SURGICAL CARTILAGE REPAIR Hagen Schmal, Alexander Mehlhorn, Fabian Stoffel, Philipp Niemeyer, Norbert P. Südkamp Department Orthopädie und Traumatologie, Uniklinik Freiburg, Germany Keywords: cytokine, cartilage repair, Background: Cartilage defects are considered as an initial event in the progress of osteoarthrosis. Although there is much known about cartilage metabolism including significant mediators reliable data about in vivo regulation of natural cartilage repair and consequences of surgical interventions are still missing. Methods: Lavage fluids of knee joints of 47 patients undergoing an arthroscopy were collected between August 2006 and September 2007. 5 patients had no cartilage lesion and served as a control group, in the case of the other 42 patients the cartilage defects were treated by microfracturing (19 patients) or by an Autologous condrocyte transplantation (23 patients). The average age of the patients with cartilage lesions was 42±10 years, the gender distribution was equal. The concentrations of bFGF, IL-1beta, BMP-2, Aggrecan, BMP-7 and IGF1 were determined by ELISAs. Results: High level expression in the controls was found for BMP-2 and Aggrecan, low level constitutive expression for bFGF and IGF-1, concentrations of IL-1beta and BMP-7 in the control group remained

804 below detection levels. The concentrations of IGF-1 in the lavage fluids of knees with cartilage lesions were significantly higher (p< 0.05) than in the control group. A similar association was found for bFGF, but differences did not reach statistical significance. bFGF concentrations seem strongly to depend on size of cartilage lesions, because levels in the knees of patients undergoing an Autologous condrocyte transplantation, that is used only in cases with defects greater than 3 cm^2, were significantly higher compared to the control group (p<0.05) or the group of patients undergoing microfracturing (p<0.001), that is used in case of smaller lesions. Levels of all other measured cytokines did not show any dependence on cartilage defects. Levels of Aggrecan and BMP-7 did not change after surgical cartilage repair, whereas concentrations of bFGF (p < 0.001), IL-1beta (p<0.01), BMP-2 (p<0.001) and IGF-1 (p< 0.001) significantly increased after the arthroscopic intervention. Intraarticular concentrations of IL-1beta significantly correlated with systemic concentrations of the C-reactive protein determined at day 4 (Pearson 0.74, p<0.0001), but not on day 2. Conclusions: bFGF and IGF-1 seem to play a pivotal role in natural and surgical cartilage repair. Arthroscopic intervention is additionally associated with inflammation-like reactions, because IL-1beta is elevated after the operation and correlates with a systemic marker of inflammation. Aggrecan, a known specific component of the extracellular cartilage matrix, is constitutively expressed in knee joints and not dependent on procedures of natural or surgical cartilage repair.

119 BERSTUNGSDRUCK, EXPRESSIONSVERHALTEN VON MMP13 UND VEGF VON GENa´HTEN UND GEKLEBTEN ANANSTOMOSEN IM RATTENMODELL Arne Weinhold, Hayo Rieger, Heinz-Johannes Buhr, Uwe Pohlen Chirurgische Klinik I, Charite, Campus Benjamin Franklin, Berlin, Berlin Keywords: MMP13, VEGF, geklebte Anastomosen Background: Hohe Insuffizienzraten bis 25% mit einer Mortalität bis zu 50% bei Insuffizienz kolorektaler Anastomosen zeigen deutlich die Problematik chirurgischer Interventionen auf. Die Versorgung einer solchen Insuffizienz ggf. die primäre Protektion der Anastomose durch Verklebung mit Kollagenfliess wäre eine Möglichkeit die Insuffizienzrate und den delitären Verlauf zu beeinflussen. Ziel der Studie war der Vergleich genähter mit geklebten kolorektalen Anastomosen im Rattenmodel. Zielparameter waren der Berstungsdruck der Anastomosen sowie das Expressionsverhalten von MMP13 als Haupteffektor des Kollagenremodeling und VEGF als einen Hauptgewebsfaktor im Zeitverlauf. Methods: 60 männliche Wistar-Ratten wurden in 2 Gruppen randomisiert. Gruppe I: Zirkulär genähte Anastomose im rektosigmoidalen Übergang. Gruppe II: semizirkulär genähte Anastomose und Klebung der Vorderwand mit Fibrin beschichteten Kollagenvlies (TachosilR). Je 10 Tiere wurden pro Zeitpunkt gemessen. Nach 12, 48 und 120 Stunden erfolgte die in situ Messung des Berstungsdrucks. Nach Euthanasie der Tiere wurde die MMP13- und VEGF-Expression in den jeweiligen Anastomosenregionen mittels Real-Time-PCR evaluiert. Results: Bestimmung des Berstungsdrucks, MMP 13 und VEGFExpression im Gruppenvergleich: Es zeigten sich ein signifkanter Stabilitätsvorteil der Kleb-Anastomose nach 120 h und eine signifikante Minderaktivität der MMP13-Expression und eine signifikant erhöhte Expression von VEGF der Klebanastomosen im Gruppenvergleich. Gruppe Naht Gruppe Klebung p Berstungsdruck in mmHg 12 h 61 21.5<0.01 48 h 33 28.8 n.s. 120 h 130 160<0.01 MMP13 Expression 12 h 6.12 2.26 n.s. 48 h 34.39 3.77<0.01 120 h 51.66 11.98<0.01 VEGF-Expression 12 h 1.14 1.65 n.s. 48 h 1.73 2.09 n.s. 120 h 0.97 0.67 n.s.

Langenbecks Arch Surg (2008) 393:759–815 Conclusions: Berstungsdruck: Nach 12 h zeigt sich ein signifikanter Stabilitätsvorteil für die genähte Anastomose. Nach Ausheilung der Anastomose (120 h) war die geklebte Anastomose signifikant stabiler als die genähte. MMP13 und VEGF- Expression: Der Stabilitätsvorteil der Kleb-Anastomose könnte durch eine geringere Remodellingaktivität (MMP13) erklärt sein. Die höhere VEGF-Expression in der KlebAnastomose spricht zudem für eine Beschleunigung der Stabilisierung innerhalb der geklebten Anastomose, die nach 120 h abgeschlossen scheint.

120 VASCULARIZED ENGINEERED TRANSPLANTS FOR RECONSTRUCTIVE SURGERY: IMPLANT GENERATION AND FIRST CLINICAL EXPERIENCES Thorsten Walles, Johanna Schanz, Markus Schandar, Volker Steger, Godehard Friedel, Heike Mertsching Dept. of Thoracic Surgery, Schillerhoehe Hospital, StuttgartGerlingen, Germany and Dept. of Cell and Tissue Engineering, Fraunhofer IGB, Stuttgart, Germany Keywords: tissue engineering; vascularization; transplantation; clinical application Background: The repair of hollow organ lesions (f.e. oesophagus, central airways) still remains a challenging task in reconstructive surgery. In the prevalent setting of chronic tissue inflammation and distinctive bacterial contamination no suitable (bio-) materials are available for defect closure so far. Tissue engineering applications represent promising solutions for surgical reconstruction, however, their success bases on sufficient implant vascularisation. Based on previous pre-clinical work we developed an engineered transplant (EnTrans) equipped with an innate vascular system that can be connected to the patient’s blood-circuit. Methods: Fibroblasts and myocytes were obtained by muscle biopsy from the patient and cells were isolated following standard protocols and cultured under GMP conditions for 6 weeks. The mononuclear cell fraction containing endothelial progenitor cells was separated from the peripheral blood (400 ml). Mononuclear cells were seeded on the preserved basement membrane within the vascular structures of a decellularized porcine jejunal segment by cumulative pulsatile perfusion. The differentiation process of endothelial precursor cells into endothelial cells was verified by the determination of CD34, CD31, and VE cadherin expression. Within 14 days an autologous vascular human network was formed. Subsequently, the circumjacent matrix was seeded with patient’s fibroblasts and muscle cells. For clinical application, tissue viability and thrombogenicity of the engineered vascular network was surveyed by connecting the EnTrans to the patients vasculature in the left upper arm for one week. Following explantation, permeability and composition of the vascular network, viability of the EnTrans, and immunological hostversus-graft reactions were analysed. Additional applications followed Results: The pre-clinical experiences regarding implant-generation could be transferred to the clinical setting. The EnTrans was endued with a closed circuit vascular system enabling the recirculation of cell culture medium in vitro. EnTrans connection to the patient’s blood-circuit could be realized applying standard surgical techniques. Aggressive clinical anticoagulation protocols were sufficient to maintain graft patency. At EnTrans explantation, a moderate venous backflow was detectable clinically. Histological work up identified a confluent endothelial lining of the artificial vascular network within the EnTrans and a strong tissue viability. There was no clinical or histological evidence of implant site inflammation or EnTrans rejection after 1 week of implantation. Conclusions: Pre-clinical techniques to engineer vascularized transplants for reconstructive surgery have been successfully applied in the clinical setting. Although anecdotal, these encouraging experiences may unlock the door for the clinical application of engineered repair tissues for reconstructive surgery in numerous paediatric and adult

Langenbecks Arch Surg (2008) 393:759–815 pathologies where therapeutic interventions are limited. Moreover, vascularized bioartificial tissues seeded with human islet cells could represent a new treatment modality for diabetic patients type I. However, long-term immunogenicity and vascular supply, especially the venous drainage, remain critical issues in transplant generation and function.

121 BIOARTIFCIAL HUMAN TISSUES AS TEST SYSTEMS FOR BASIC AND APPLIED RESEARCH: THE VASCULARISED LIVER MODULE Johanna Schanz, Kirstin Linke, Jan Hansmann, Heike Mertsching Fraunhofer Institute for Interfacial Engineering and Biotechnolgy, Stuttgart, Germany Keywords: tissue engineerning; artificial liver; test system Background: Currently applied in vitro tissue cultures are highly artificial and the findings from animal studies often can not be transferred to the clinic. The liver centers the human metabolic and detoxification pathways. Therefore human liver tissue that is functional and viable for several weeks in vitro represents an auspicious test-system for screening active agents in pharmacy, chemistry and cosmetics and for identifying toxic metabolites. Such tissues need to recapitulate both the 3D organization and multicellular complexity of the liver but at the same time accommodate systematic experimental intervention. Here, tissue engineering, the generation of human tissues and organs in vitro, provides new perspectives for basic and applied research by offering 3D tissue-cultures resolving fundamental obstacles encountered in currently applied 2D and 3D cell-culture systems. We developed a bioartificial liver tissue that is equipped with an artificial vascular network enabling the co-culture of hepatocytes (HC) and endothelial cells (EC). Methods: A porcine jejunal segment with its preserved was network served as biological vascularized scaffold (BioVaSc). The BioVaSc was chemically decellularised by sodium desoxycholate and washed to remove all cells. HC were isolated from porcine and human liver biopsies by the use of collagenase IV. For endothelial reseeding, porcine progenitors were isolated from bone marrow aspirate, and human microvascular EC from foreskin biopsies, respectively. Cocultures were implemented by reseeding the decellularized vascular network within the BioVaSc. After 7 to 14 days, HC were seeded in a collagen gel suspension on the matrix lumen. The BioVaSc was perfused with cell culture medium over the artery in a bioreactor system. The flow rate was controlled via a sensor and a PC to simulate in vivo blood flow. The cells were characterized for the expression of liver specific markers, vitality and for metabolic activities like urea and albumin synthesis, phase I and II metabolism and lactate formation. Results: EC were successfully seeded via the arterial inflow in the vascular bed of the matrix. After 4 weeks of culture Progenitor cells isolated from the bone marrow aspirate could be differentiated into EC and human EC keep their differentiation. They are vital and express endothelial cell specific markers. Vital HC could be effectively populated on the surface of the lumen and partially migrate through the deeper lying structures of the matrix. Thereby they form cell layers with typical hepatic morphology and bile canaliculi structures. The culture of HC on the matrix shows good results for vitality, expression of liver specific markers, and formation of cell-matrix and cell-cell contacts. Liver specific functions (e.g urea and albumin synthesis), phase I and phase II metabolism could be obtained over three weeks. Conclusions: First steps towards the development of a vascularised bioartificial liver module with physiological liver cell functions were realized. EC shall be co-cultivated with the HC in such a way, that

805 they can form out a filtration barrier. Thus it will be possible to simulate the way of a drug in vitro including the passage through the vascular system and the EC layer. For the first time a testing system should enable the arterial application of substances and the identification of metabolites in a venous system after a single application as well as after repeated applications.

122 XENOGENIC TRANSPLANTATION OF HUMAN MESENCHYMAL STEM CELLS FOR BONE REGENERATION IN A CRITICAL SIZE DEFECT MODEL OF THE SHEEP TIBIA Thomas S. Schönberger (1), Nina Siebert (1), Stefan Milz (2), Joachim Hahn (2), Simon Pearce (2), Norbert P. Südkamp (1), Philipp Niemeyer (1) 1: Department of Orthopedic Surgery and Traumatology, Freiburg University Hospital, Freiburg, Germany, 2: AO Research Institute, AO Foundation, Davos, Switzerland Keywords: xenogenic, transplantation, MSC, bone regeneration, In Situ Hybridisation, critical size defect, sheep Background: Mesenchymal stem cells (MSC) isolated from bone marrow represent an attractive cell source for bone tissue engineering. Due to the lack of immunological relevant surface antigen expression, this cell population might also be available for allogenic or xenogenic transplantation and regeneration of bone defects. Although the immunopriviliged character of MSC was demonstrated in vitro and the engraftment of allogenic and xenogenic MSC was demonstrated in immunocompetent mice and fetal sheep, its remains unclear if a HLA independent (allogenic or xenogenic) transplantation of MSC is able to regnerate bone defects in an HLA independent host. The aim of this study was to investigate, whether the xenogenic transplantation of MSC leads to increased bone formation in critical size defect compared to unloaded controls and autologous MSC transplantation. Methods: After isolation, in vitro expansion and cultivation on a mineralized collagen scaffold, xenogenic human and autologous ovine MSC were transplanted into a diaphyseal critical size defect of the tibia of swiss alpine sheep. Unloaded mineralized collagen scaffolds served as controls. The animals were sacrificed after 3 (2 animals per group) and 6 months (5 animals per group) Radiographic controls were performed every two weeks to evaluate newly formed bone in a quantitative analysis using the digital image software analysis program GIMP. After euthanasia histological analysis was performed descriptively and quantitatively using semi quantitative scores. For evaluating an engraftment of xenogenic human MSC in the immunocompetent model of the sheep, In Situ hybridization was performed. Results: Autologous MSC transplantation lead to significant increased bone formation in the radiological evaluation after 10 weeks (p<0,05) compared with unloaded controls. Histological evaluation also leads to significant increased bone formation with statistical significance (p<0,05). Compared to the autologous MSC group, the transplantation of xenogenic MSC leads to inferior bone formation (p < 0.05). Although bone formation in xenogenic group was significantly higher than in the unloaded control group using radiographic evaluation (p < 0,05), histological evaluation failed statistical significance. Nevertheless, engraftment of xenogenic human MSC could be detected 6 months post implantation using human specific In Situ hybridisation, while no severe systemic or local immune response could be detected. Conclusions: Although local engraftment of human MSC could be detected in a xenogenic immunocompetent animal model 6 months

806 after transplantation and no severe immuno reaction could be detected, xenogenic transplantation of MSC seems to lead to inferior bone formation compared to autologous transplantation. The reasons for this discrepancy, explainable perhaps with species specific growth factors, need to be investigated in further studies.

123 NITRIC OXIDE INDUCED DILATORY EFFECT AFTER ERYTHROPOIETIN PRETREATMENT PREVENTS SKIN FROM ISCHAEMIC NECROSIS Farid Rezaeian (1), Wettstein (1), Philipp Mörsdorf (2), Annick Bächle (2), Michael D. Menger (2), Yves Harder (1) 1: Division of Plastic, Reconstructive & Esthetic Surgery, Geneva University Hospitals, Switzerland, 2: Institute for Clinical & Experimental Surgery, University of Saarland, Homburg/Saar, Germany Keywords: Angiogenesis — apoptotic cell death — critical ischemia — erythropoietin — inflammation — intravital epi-fluorescence microscopy — microcirculation — preconditioning — flap surgery Background: Erythropoietin (EPO), the main regulator of erythropoiesis, is commonly used to treat certain types of anaemia. Irrespective of its erythropoietic properties, recent findings have attested EPO pleiotropy allowing for tissue-protection in a variety of organ systems suffering from ischaemic injury such as brain, heart and striated muscle. In flap surgery, persistent ischaemia might cause wound breakdown and necrosis. Herein we studied the putative effectiveness of EPO to reduce tissue necrosis aiming at elementary microcirculatory and cellular mechanisms. Methods: A randomly perfused musculocutaneous flap integrated in a dorsal skinfold chamber of C57BL/6-mice was used as a model for persistent ischaemia. 28 mice were assigned to four experimental groups of 7 animals each: 1. EPO (500IU/kg bodyweight (bw)); 2. LName (50 mg/kg bw; unspecific nitric oxide synthase (NOS) inhibitor); 3. EPO+ L-Name (500IU EPO/kg bw +50mgL-NAME/kg bw); 4. CON (NaCl). The substances were administered intraperitoneally 30 minutes before as well as 30 minutes and 24 hours after flap elevation. Arteriolar diameter, functional capillary density (FCD), angiogenesis (mean vessel density; MVD) and flap necrosis were assessed with repetitive epi-fluorescence microscopy over a 10-day period. Western blot experiments and immunohistochemistry were performed to analyze the expression of inducible NO-synthase (iNOS), endothelial NO-synthase (eNOS), vascular endothelial growth factor (VEGF) and EPO-receptor (EPO-R). Blood values including haematocrit were measured in separate animals (n=8 each). Results: Increased expression of iNOS (6-fold vs CON) and eNOS (2fold) in EPO-pretreated mice correlated with significant arteriolar dilation (CON: day1 after surgery: 49±3m, day10: 52± 5m; EPO: day1: 65±3m, day10: 77±5 m; p<0.05 vs CON) and maintained FCD at day 10 (CON: 60±2 cm/ cm2; EPO: 119±13 cm/cm2; p<0.05). Also, EPO induced a VEGFupregulation (3-fold) resulting in newly formed capillaries. Consequently, EPO pretreatment significantly reduced flap necrosis (CON: 48±2%; EPO: 20±3%; p<0.05). In contrast, L-Name abolished tissue protection by abrogating the dilatory effect (L-Name+ EPO: day 1: 46±3m, day10: 54±4m; L-Name: day 1: 38± 4m, day10: 46±4m) finally resulting in reduced FCD and tissue survival (L-Name+ EPO: 39±4%; L-Name: 59±7%), without counteracting angiogenesis. Haematocrit was not influenced by EPO, which was in line with similar EPO-R expression in both CON and EPO-groups. Conclusions: Improved tissue survival following non-haematocritrelevant EPO-administration results from a NO-dependent increase in microcirculatory perfusion and a VEGF-mediated angiogenic response. Accordingly, pretreatment with EPO seems a promising non-

Langenbecks Arch Surg (2008) 393:759–815 invasive approach to pharmacologically reduce ischaemia-related complications in elective and even emergency surgery at risk of wound breakdown and tissue necrosis.

124 THE INFLUENCE OF NONSTEROIDAL ANTIINFLAMMATORY DRUGS (NSAID) AND IL-1 BETA ON THE PROLIFERATION OF HUMAN ARTICULAR CHONDROCYTES Barbara Maria Dirhold, Alexander T. Mehlhorn, Norbert P. Südkamp, Hagen Schmal Department of Orthopaedic & Trauma Surgery, University Medical Center, Albert-Ludwigs University of Freiburg, Germany Keywords: Chondrocytes, Osteoarthritis, Cartilage, Inflammation, Nonsteroidal Antiinflammatory Drugs, IL-1 beta Background: Nonsteroidal Antiinflammatory Drugs (NSAID) are applied in orthopedics and trauma surgery as a standard analgetic medication in the therapy of osteoarthritis, posttraumatic- and postoperative pain. Animal studies have shown a delayed fracture union or nonunion and a reduced bone density following NSAID application during the healing process. The influence of NSAID on differentiation and metabolism of the osteoarthritic joint cartilage is not yet completely analyzed. During this process the interaction of cytokines, NSAID and the cellular components seem to be the key for understanding. Therefore, this study was performed to elucidate this issue. Methods: Human femoral heads were obtained during hip arthroplasty operations following femoral neck fractures. Afterwards, chondrocytes were enzymatically released from their natural matrix. Cells from patients with advanced arthrosis (≥ grade 3 Kellgren and Lawrence) were not used for the experiments. After isolation and amplification of the chondrocytes in monolayer, cells of 3 patients were pooled. Primary and chondrocytes of the second passage were cultured in alginate beads. The cells were stimulated with either IL-1 beta or IL1 beta in combination with the NSAIDs Ibuprofen, Paracetamol, or Diclofenac simulating inflammation and intervention. As controls, experiments were performed with cells cultured in monolayer’s and co-cultured with peripheral blood mononuclear cells (PBMCs) stimulated with lipopolysaccharide (LPS). Cell proliferation and activity were determined utilizing an MTT-assay and a pico-green assay. Furthermore, the bFGF content in the supernatants was measured by ELISA. Results: For both primary and passaged chondrocytes the MTT-assay following 14 days of incubation showed a significant increase in the metabolic activity of IL-1beta or IL-1beta and NSAID treated chondrocytes. Intervention groups were compared to chondrocytes without additions or cells exclusively treated with NSAIDs. The level of increase was similar for both cell types used, but even more distinct in the groups with added NSAIDs. In contrast, cell activity of chondrocytes co-cultured with LPS-stimulated PBMCs was significantly reduced by 23%, the activity of chondrocytes in monolayer’s was not significantly alternated by the addition of IL-1 beta. The measurement of DNA-content correlated with the results of the MTTAssay. bFGF, an important chondrogenic stimulator of proliferation, was measured in the supernatant of the investigated cell cultures. We found increased concentrations in all groups exhibiting higher proliferation rates. Conclusions: Isolated addition of IL-1 beta resulted in a stimulation of proliferation of primary and passaged human chondrocytes. NSAID did not influence proliferation of chondrocytes without stimulation, but enhanced IL-1beta induced proliferation boost. The effect of IL-1 beta might be caused by induction of apoptosis and a release of intracellular bFGF.

Langenbecks Arch Surg (2008) 393:759–815

125 CO-CULTIVATION OF DIFFERENT CELL TYPES ON SPIDER SILK FIBRES Anja Hillmer, Christina Allmeling, Kerstin Reimers, Jörn Kuhbier, Peter M. Vogt Department of Plastic, Hand, and Reconstructive Surgery, Medical School Hannover, Germany Keywords: spider silk, tissue engineering Background: After a long time of evolution spiders became interesting for tissue engineering. The optimal matrix for tissue engineering has no immunological effect after transplantation, has positive effects of cell growth and is biodegradable. Spider silk serves as a conventional matrix for tissue engineering. Methods: In the present study, spider silk from Nephila clavipes was used to weave a cross pattern (150–300 µm3) matrix on stainless steel frames. Different types of cells were seeded on the silk fibres to generate a three dimensional organotypic construct. Petri-dishes were coated with 0.2% Pluronic 127 to make the surface non-adhesive. The first fluorescent labelled cell type was seeded on the frame and incubated. The cells attached on the spider silk surface and proliferated. The seeding was repeated two times with different cell types after a 7 day incubation period. Cells were observed daily under a light microscope and analysed with a fluorescent microscope. Results: After 7 days a confluent monolayer was observed. The cells adhered preferentially in the corners of the spider silk and spread into the interspaces to form a monolayer. The next cell layer was able to arrange on the spider silk sustained cell layer. We were able to cultivate three different cell types, labelled with different fluorescent dyes CellTracker orange, green and Hoechst 33324. The spider silk cells formed a monolayer on the spider silk and generated a organotypic model. Conclusions: In the future spider silk could be a innovative matrix for cultivation different types of cells (e.g. skin) to generate artificial cell constructs, use as matrix for guided growth or cell sheets. The perfect tissue-construct would be spider silk cultivated with autologues cells.

126 LOCAL DFO-APPLICATION IMPROVES REGENERATION AFTER PRIMARY NERVE SUTURE Nektarios Sinis (1), Philipp Schoenle (1), Frank Werdin (1), Timm Danker (2), Elke Guenther (2), Guido Koopmanns (3) 1: Department of Hand-, Plastic-, Reconstructive Surgery, Burn-Unit, University of Tuebingen, Schnarrenbergstr. 95, Tuebingen, Germany, 2: Hand- and Plastic Surgery, Orthopaedische Klinik Markgroeningen, Markgroeningen, Germany, 3: NMI-Reutlingen, University of Tuebingen, Markwiesenstr. 55, Reutlingen, Germany, 4: Dipartimento di Scienze Cliniche e Biologiche, University de Torino, Ospedale San Luigi, Regione Gonzole 10, Orbassano, Italy, 5: Neuraxo Biopharmaceuticals GmbH, Erkrath, Germany Keywords: primary nerve suture, iron chelator, collagen scar Background: Despite the progress of microsurgical techniques the outcome of repaired nerves after primary nerve suture remains incomplete in a lot of cases. One reason that explains these results is the development of a fibrous collagen scar at the site of coaptation with expression of different growth inhibiting substances. This collagen scar prevents axons from passing the lesion and reaching the end organ. The aim of this study was to analyze the impact of a scar inhibiting substance, namely a potent iron chelator on peripheral nerves after transection and primary nerve suture. Methods: In a rat median nerve model four experimental groups were operated. Group I — transection of the median nerve and primary nerve suture (N=12). Group II — transection and venous ensheatment of the median nerve at the coaptation site (external jugular vein) (N= 12). Group III — transection of the nerve, venous ensheatment and

807 filling the vein with a lipid carrier (N=12). Group IV — transection, venous ensheatment, filling of the vein with the iron chelator combined with the lipid carrier (N=12). The observation period was 12 weeks. During this time the animals were all examined with the grasping test for functional assessment of regeneration. After 12 weeks the flexor digitorum sublimis muscle was harvested and weighed to estimate the grade of atrophy. Furthermore, electrophysiological studies were performed measuring the amplitude and latency of the signals. Afterwards, the animals were perfused and specimens were gathered for histological analysis (electron microscopy, histomorphometry). Results: Twelve weeks post operation all animals displayed comparable functional data regarding their final force. However, in group IV animals where the iron chelator was applied, functional regeneration and grasping started sooner compared to the other groups (2nd to 3rd week, other animals 4th to 7th week). Analysis of the flexor digitorum sublimis muscle demonstrated less atrophy in these animals treated with the scar inhibitor. The trigger point for muscle contraction was decreased in rats treated with the iron chelator compared to other animals (∼30%). Histomorphometric data revealed comparable results among all the experimental groups (myelin thickness, axon counts, etc.). Conclusions: Using the iron chelator for inhibition of the fibrous scar reduces the atrophy of flexor digitorum sublimis muscle. Further functional regeneration sets on early thus minimizing the potential loss of muscle.

127 TISSUE ENGINEERING APPLYING A 3D IN VITRO CULTURE SYSTEM USING COLLAGEN COATED NANO-FIBROUS SCAFFOLDS Jeanette Bierwolf, Jeanette Bierwolf, Steffen Deichmann, Johannes Erbes, Christina Stieglitz, Eva Toronyi, Björn Nashan, Kai Fung, Peter X. Ma, Marc Lütgehetmann, Jörg-Matthias Pollok Klinik für Hepatobiliäre Chirurgie und Transplantationschirurgie, Transplantations-Center, Universitätsklinikum Hamburg-Eppendorf, University of Michigan, Ann Arbor, Michigan, USA Keywords: Tissue engineering, hepatocytes, pharmacology, toxicology, polymer scaffolds, nano Background: In contrast to animal experiments and clinical trials, cell culture studies are more economical, more comparable to each other and easier to carry out. But a very low seeding efficiency and the loss of liver cell-specific functions are only some of the typical problems in currently used 2D culture systems. The use of polymeric scaffolds could overcome these problems and support the establishment of 3D culture systems in pharmaceutical research. Methods: Primary rat hepatocytes were seeded onto PLLA-scaffolds in two different manners: The first one was a static seeding method performed by pipetting a defined number of cells onto the scaffold. In the second seeding method centrifugal force was used. The scaffolds were incubated under static conditions in a humidified atmosphere of 5% CO2 and 95% air at 37°C. Cell loading efficiency and extent of hepatocyte leakage from scaffolds were quantified via DNA content measurement. Viability and liver cell specific functions, such as albumin and glycogen storage, were evaluated at different times of culture. The activity of liverspecific factors (Connexin32, ZO-1, HNF-4) were analyzed by immunofluorescent staining. Furthermore, cytochrome p450 enzyme activity was evaluated by RT-PCR and immunofluorescent methods. Results: Our results indicate that with centrifugal force, a significantly higher loading efficiency compared to monolayer culture is possible. Hepatocytes cultured on nano sugar scaffolds revealed high cell viability and well preserved glycogen storage during the entire culture period of 7 days, as demonstrated by HE staining and positive PAS reaction. HNF-4 is one of the major nuclear hepatocyte transcription factors. Within our 3D culture system, using nano sugar scaffolds, mainly those hepatocytes, which are organized in clusters, remain HNF-4 positive during the entire

808 culture period. Connexin 32 is a marker for highly differentiated hepatocyte interaction. Aggregated hepatocytes in 3D culture on nano-fibrous scaffolds remarkably re-establish positive signalling for this marker within 7 days of culture. Zo-1 is a marker for tight junctions, which in the liver indicates bile canaliculi formation between two adjacent hepatocytes. Interestingly, this marker also stains positive as an indication for re-formation of bile canaliculi and re-establishment of the apical and basolateral membrane and therefore, bipolar configuration of the cultured hepatocytes. Cytochrome p450 enzyme activity was positive in the cultured hepatocytes during the entire culture period as demonstrated by immunofluorescent staining. Furthermore, albumin secretion determined by ELISA was existent during the entire culture period in our 3D culture system. Conclusions: Our results indicate that nano sugar scaffolds with interconnected spherical macropores provide a good microenvironment for viability and cell attachment of primary rat hepatocytes. A high cell density is necessary to compare in vivo with in vitro detoxification activity, since in sparsely seeded scaffolds there is little probability of hepatocyte aggregation. Our centrifugation seeding technique enables efficient and high density 3D culture of primary rat hepatocytes using nano sugar polymer scaffolds.

128 HETEROTYPIC CELL CONTACT BETWEEN HUMAN ENDOTHELIAL CELLS AND HUMAN OSTEOPROGENITOR CELLS SUPPORTS OSTEOGENIC DIFFERENTIATION Sven Hager, G. Björn Stark, Günter Finkenzeller Department of Plastic and Hand Surgery, University of Freiburg Medical Center, 79106 Freiburg, Germany Keywords: osteogenic differentiation, alkaline phosphatase, osteoblast, endothelial cell, co-cultivation Background: Bone development and remodeling depend on complex interactions between bone-forming osteoblasts and other cell types present within the bone microenvironment, particularly vascular endothelial cells that may be pivotal members of a complex interactive communication network in bone. In this study, we investigated the interaction between human endothelial cells (ECs) and human primary osteoblasts (hOBs). Co-cultivation of endothelial cells supports osteogenic differentiation which is characterized by a strong upregulation in the expression of the osteogenic marker alkaline phosphatase (ALP). The effect on ALP expression is mediated by the formation of heterotypic cell contacts between endothelial cells and osteoblasts. Methods: The effect of endothelial cells on osteoblastic ALP expression was investigated on mRNA level by quantitative real-time RT-PCR and on protein level by measuring alkaline phosphatase activity. Promoter analysis of the ALP gene was performed by transient transfection of a ALP-promoter-luciferase reporter plasmid into primary osteoblasts and subsequent analysis of luciferase activity in cell extracts. Results: Co-cultivation of human ECs and human hOBs leads to a 14-fold upregulation of ALP expression on mRNA level and a 3.4-fold upregulation on protein level. This effect is dependent on the formation of heterotypic cell contacts between EC and hOB and is not mediated by paracrine-acting diffusible factors. Interstingly, upregulation of ALP expression occurs only in primary osteoblasts or primary mesenchymal stem cells, but not in immortalized permanent ostoblastic cell lines. Molecular analysis of the activation mechanism of ALP expression revealed that co-cultivation of ECs does not lead to a transcriptional activation of the ALP promoter suggesting that the effect of EC co-cultivation on ALP expression may be mediated by stabilization of ALP mRNA. Conclusions: Based on the finding that co-cultivation of ECs with hOBs leads to a strong upregulation in the expression of the most important osteogenic marker alkaline phosphatase, we conclude that endothelial cells have a positive impact on osteogenic differentiation of osteoprogenitor cells.

Langenbecks Arch Surg (2008) 393:759–815

129 MORPHOLOGICAL AND IMMUNOLOGICAL CHARACTERISTICS INDICATE THE YIELD OF EARLY PROGENITORS AND REPRESENT A QUALITY CONTROL METHOD FOR HUMAN MESENCHYMAL STEM CELL CULTURING Wolf Christian Prall (1), Florian Haasters (1), David Anz (2), Denitsa Docheva (1), Stefan Endres (2), Wolf Mutschler (1), Matthias Schieker (1) 1: Experimental Surgery and Regenerative Medicine, Department of Surgery, University of Munich (LMU), Nussbaumstr. 20, 80336 Munich, Germany, 2: Division of Clinical Pharmacology, Department of Internal Medicine, University of Munich (LMU), Ziemsenstr. 1, 80336 Munich, Germany Keywords: Mesenchymal stem cells (MSCs), characterization, subtypes, heterogeneity, morphology, RS cells Background: Human mesenchymal stem cells (hMSC) are a heterogeneous cell population, which is reflected in varying morphology and biological properties. Three subpopulations with intrinsic characteristics can be distinguished: small rapidly self-renewing (RS) cells, spindleshaped (SS) cells and large flat cells (FC). Unfortunately, it has not been possible to relate specific surface marker features to these subpopulations. Methods: Here, the hMSC subpopulations of three donors were morphologically detected by cell length and cell area and compared to an immortalized stem cell line. Afterwards these cells were stained for the putative hMSC surface markers CD105, CD90 as well as CD73 and evaluated by three colour flow cytometry and simultaneous multicolour immunocytochemistry. Results: Interestingly, the percentage of RS cells was positively correlated with the percentage of cells positive for all three hMSC surface markers. On the other hand, the percentage of FC was inversely correlated with percentage of cells positive for CD105, CD90 and CD73. Expression of CD73 showed the highest heterogeneity. Immunocytochemistry further confirmed that mainly FC lack CD73 expression whereas mostly RS cells were positive for all three hMSC markers. Conclusions: In literature, hMSC properties are especially conceded to RS cells whereas FC are discussed to represent early stages of differentiation. Taken together, analysis of the morphological shape of cell culture populations can be adducted to provide further reference for stem cell criteria. Among the recently suggested surface markers, CD73 seems to be the most sensitive since it may already be downregulated in early stages of differentiation.

130 INTRODUCING A SINGLE-CELL-DERIVED HUMAN MESENCHYMAL STEM CELL LINE EXPRESSING HTERT AFTER LENTIVIRAL GENE TRANSFER Wolfgang Böcker (1), Zhanhai Yin (2), Inga Drosse (1), Florian Haasters (1), Matthias Wierer (1), Cvetan Popov (1), Melanie Locher (3), Wolf Mutschler (1), Denitsa Docheva (1), Matthias Schieker (1) 1: Experimental Surgery and Regenerative Medicine, Department of Surgery, Ludwig- Maximilians-University (LMU), Nussbaumstraße 20, 80336 Munich, Germany, 2: Department of Orthopaedics, First Affiliated Hospital, School of Medicine, Xi’an Jiaotong University, Yanta West Road 277, 710061 Xi’an, Shaanxi Province, P. R. China, 3: Center for Human Genetics and Laboratory Medicine, Lochhammerstr. 29, 82152 Munich, Germany Keywords: Mesenchymal stem cells, Telomerase reverse transcriptase, Gene transfer techniques, Lentivirus Background: Human mesenchymal stem cells (hMSCs) can be readily isolated from bone marrow and differentiate into multiple tissues,

Langenbecks Arch Surg (2008) 393:759–815 making them a promising target for future cell and gene therapy applications. The low frequency of hMSCs in bone marrow necessitates large expansion in vitro prior to clinical use. But due to senescenceassociated growth arrest during culture, only limited cell numbers can be generated. Since malignant transformation was observed in hMSCs and the use of retroviral vectors caused insertional mutagenesis, we ectopically expressed hTERT using lentiviral gene transfer. Methods: Lentiviruses containing the gene of hTERT were generated from the lentiviral expression construct pLenti6/V5-hTERT. After lentiviral transduction of hMSCs four clones were isolated by single cell picking. HTERT expression was confirmed by PCR, immunocytochemistry and telomerase activity assay. We observed morphology, senescence and growth kinetics by monitoring the population doubling level (PDL) for more than 2 years, β-galactosidase activity and BrdU proliferation assay. Evaluation of potential neoplastic transformation was carried out by RT-PCR for tumor suppressor proteins Rb-1, p14, p16, p53, p21, a soft agar assay and upon in vivo transplantation. Chromosomal aberration was evaluated by fluorescence in situ hybridization (FISH). Stem cell plasticity was confirmed by In vitro differentiation capacity osteogenic, chondrogenic and adipogenic lineage. Results: Single cell derived hMSC clones expressing hTERT did not show malignant transformation in vitro and in vivo after extended culture periods. There were no changes observed in the expression of tumor suppressor genes and karyotype. Cultured hMSCs lack telomerase activity, but it was significantly increased by ectopic expression of hTERT. Furthermore hMSC senescence was prevented and the cells showed significantly higher and unlimited proliferation capacity. Even after an extended culture period (>PDL 320), hMSCs expressing hTERT preserved their stem cells character as shown by osteogenic, adipogenic and chondrogenic differentiation. Conclusions: In summary, we report for the first time a hMSC clone successfully immortalized by ectopic expression of hTERT using lentiviral gene transfer without malignant transformation. Extending the lifespan of human mesenchymal stem cells by ectopic expression of hTERT may be an attractive and safe way to generate appropriate cell numbers for applications in the field of Regenerative Medicine. The immortalized clonal hMSCs revealed no change of karyotype with no signs of malignant transformation in vitro and in vivo, making them an attractive and safe tool for ex vivo cell and gene therapy.

131 PROLIFERATION OF CD8-POSITIVE T CELLS IN THE LUMINA OF RAT RENAL ALLOGRAFT BLOOD VESSELS Veronika Grau, Gabriele Fuchs-Moll; Winfried Padberg Laboratory of Experimental Surgery, Department of General and Thoracic Surgery, Justus-Liebig-University Giessen, Rudolf-Buchheim-Str. 7, 35385 Giessen, Germany Keywords: T cells, proliferation, kidney transplantation, rat Background: Activation and subsequent proliferation of alloreactive T cells are essential steps in the rejection of vascularised organ transplants. The cellular mechanisms and the anatomical compartments where alloreactive T cell activation takes place are not fully understood. In this study, we test the hypothesis that T lymphocytes proliferate in the lumina of the blood vessels of renal allografts. Methods: Kidneys were transplanted in the fully allogeneic Dark Agouti to Lewis rat strain combination. Cells in the S-phase of mitosis were pulse-labelled with 5-bromo-2′-deoxyuridine (BrdU). T lymphocytes were detected and characterized inside the lumina graft blood vessels by immunohistochemical triple-staining with antibodies to endothelial cells, BrdU, and cell surface markers of T lymphocytes. Results: On day 3 post-transplantation, 22±6% (mean±standard deviation) of the intravascular allograft T cells were labelled with BrdU whereas

809 only 4±1% of isograft T cells incorporated BrdU (p=0.014). These data were further corroborated by immunohistochemical detection of the proliferating cell nuclear antigen (PCNA) in the nuclei of intravascular allograft T lymphocytes. Subset analysis revealed that 72±5% of the cells labelled with BrdU expressed CD8 and 10±2% of them CD4. Conclusions: Our data suggest that alloantigens are directly or semidirectly presented to host T lymphocytes within the blood vessels of the graft resulting in intravascular T cell proliferation which has never been directly evidenced before.

132 INTRAVITAL MICROSCOPIC ANALYSIS OF VASCULARIZATION OF TRANSPLANTED PANCREATIC ISLETS, PSEUDOISLETS AND MODIFIED PSEUDOISLETS Christine Wittig, Mathias W. Laschke, Michael D. Menger Institute for Clinical & Experimental Surgery, University of Saarland, Homburg/Saar, Germany Keywords: diabetes mellitus, islets, bone marrow, insulin, revascularization Background: During the last years, the transplantation of isolated pancreatic islets has been proposed as an ideal therapeutic strategy for the treatment of type 1 diabetes mellitus. However, the applicability of this therapy is limited, because it requires the transplantation of a large number of islets in order to induce normoglycemia. In order to overcome this problem, generation of modified islets by combining islet cells with bone marrow derived cells may represent a promising approach to increase the amount of transplantable tissue. Moreover, a major prerequisite for the success of this therapy is a rapid and sufficient vascularization of the grafts in order to guarantee their survival and longterm function. The aim of the present study was therefore to analyze whether the addition of bone marrow cells to islets cells accelerates the vascularization after transplantation. Methods: Langerhans islets of Syrian golden hamsters were isolated by collagenase digestion of the pancreas. The isolated islets were dispersed into single cells by trypsin treatment and then cultured for 3 days to generate pseudoislets by reaggregation of the cells. Furthermore, the bone marrow of the femur and tibia of donor hamsters was removed and co-cultured with monolayer cultures of islet cells to create modified pseudoislets. Subsequently, 7–8 pancreatic islets, pseudoislets (without bone marrow cells) and modified pseudolislets (with bone marrow cells) were transplanted into the dorsal skinfold chamber of recipient animals. Using the technique of intravital fluorescence microscopy we then analyzed angiogenesis, vascularization and microhemodynamics of the grafts over a 14-day observation period. Results: The islet grafts exhibited a size of ∼0.06–0.09 mm2 after transplantation into the dorsal skinfold chamber without any significant differences between the groups. First signs of angiogenesis, i.e. capillary sprouts, could be observed in the grafts of all three groups at day 3. Interestingly, during the further time course modified pseudoislets exhibited an improved vascularization, as indicated by a significantly increased functional capillary density of 365±18.4 cm/ cm2 (p<0.05)) at day 6 when compared to islets (223±20 cm/cm2) and pseudoislets (211±37 cm/cm2). Moreover, newly formed blood vessels in modified pseudoislets presented with increased diameters (day 6: 7.9±0.1 µm; p<0.05) and volumetric blood flow (day 6: 6.7± 0.4pl/s; p<0.05) when compared to those of islets (6.8±0.2 µm and 4.0±0.6 pl/s) and pseudoislets (7.0±0.1 µm and 4.5±0.8 pl/s). Conclusions: In the present study we could demonstrate that modified pseudoislets created by the combination of pancreatic islet cells and bone marrow derived cells exhibit an improved vascularization after transplantation when compared to normal islets. This may be most probably due to the angiogenic potential of stem cells included in the bone marrow. Thus, generation of modified pseudoislets may

810 represent a promising approach in order to improve the success rates of islet transplantation for the treatment of diabetes mellitus.

133 PROGNOSTIC IMPLICATIONS OF TUMOR-ASSOCIATED FACTORS IN RESECTABLE PANCREATIC ADENOCARCINOMA Martin Gasser, Martin Grimm, Meike Munt, Arnulf Thiede, Ana Maria Waaga-Gasser Department of Surgery I and Molecular Oncology and Immunology, University of Wuerzburg, Germany Keywords: pancreatic adenocarcinoma, tumor-associated factors Background: Reproducible prognostic markers could potentially allow patients with pancreatic cancer to be stratified into treatment groups. In this study, we asked whether EGF, EGFR, erb B2, TGF-α and -, p53, and PCNA expression might be useful in detecting different aggressive pancreatic cancers and if survival would be worse in high marker expressors. Methods: In this retrospective study a total of 282 patients with histologically verified pancreatic ductal adenocarcinoma were treated at our Department from 1993 to 2003. In addition, 93 patients were selected based on the availability of tissue specimens from the primary tumor and R0-resection (75% head, 13% corpus, 12% caudal). We used Elisa, immunohistochemistry and real time PCR to investigate expression of the different tumor associated factors and compared with standard marker CEA and CA19-9 expression. The values of morphological and molecular parameters were correlated with clinicopathological characteristics for their predictive value of tumor recurrence using chi-square test, Kaplan-Meier method, multivariate analysis by Cox regression, and Mann-Whitney-U-Test (follow-up period: 32±6.2 months). Results: A total of 34% of the patients underwent R0-resection, which significantly influenced their survival (78% stage UICC IIb, p< 0.0001). Mean survival and time to progress were also influenced by N-stage (mean survival N1: 21.5 Mo vs. N0: 49.1 Mo) and tumor differentiation (G1: 63.7 Mo; G2: 26.3 Mo; G3: 15.2 Mo; p=0.009). The prognosis of patients undergoing curative R0-resection seems to be determined by perineural invasion (p=0.019) with a mean survival of 12 months (no invasion: 29 months). With exception of MIC-A all of the analyzed parameters showed an increased protein expression during tumor progression. A prognostic significance was observed for increased CD4+ and CD8+ T cell infiltration, and poor EGF, erb B2, and PCNA expression in the tumor (p<0.001). Conclusions: Our data suggest that patients, even if R0-resected and with no lymph node metastasis, may benefit from analysis of the prognostic markers EGF, erb B2, PCNA, and MIC-A in their tumors. Those nodal negative patients with increased marker expression profiles would potentially profit from chemotherapy combined with small molecules targeting intracellular signal transduction pathways.

134 DIFFERENTIAL BETA-ARRESTIN 2 (ARRB2) EXPRESSION DURING REJECTION OF EXPERIMENTAL RENAL ALLOGRAFTS A. Zakrzewicz (1), G. Krasteva (2), J. Wilhelm (3), S. Wilker (1), H. Dietrich (4), P. Freitag (1), W. Padberg (1), V. Grau (1) 1: Laboratory of Experimental Surgery, Department of General and Thoracic Surgery, University of Giessen Lung Center, Justus-LiebigUniversity Giessen, Rudolf-Buchheim-Str. 7, 35385 Giessen, Germany, 2: Institute for Anatomy and Cell Biology, Justus-Liebig-University Giessen, Germany, 3: Department of Pathology, Justus-Liebig-University Giessen, Langhans Str. 10, 35385 Giessen, Germany, 4: Department of Internal Medicine, Justus-Liebig-University Giessen, Klinik Str. 36, 35385 Giessen, Germany

Langenbecks Arch Surg (2008) 393:759–815 Keywords: leucocytes, arrestin beta 2, NF kappa B Background: Numerous leukocytes accumulate in the blood vessels of experimental renal allografts. A majority of them are identified as activated, cytotoxic monocytes, which may cause graft rejection. To elucidate the mechanisms of monocyte activation during allograft rejection, we sought to identify novel regulating factors. Methods: Renal transplantation was performed in Lewis rats, and Dark Agouti rats were used as allogenic donors. Intravascular graft leukocytes were obtained from isografts and allografts by perfusion on day 4 after transplantation. RNA extracted from perfusate leukocytes was used for microarray analyses of genes regulated during acute rejection. Results: Microarray technology and quantitative RT-PCR revealed that mRNA expression of ARRB2 was reduced in leukocytes from allografts compared to isografts. Differential expression of ARRB2 protein was confirmed by Western blotting. In agreement with these findings, also the mRNA level of beta 2 adrenergic receptor which regulates the level of ARRB2 was markedly decreased in these cells. Confocal immunofluorescence microscopy revealed ARRB2 expression in graft monocytes and other leukocytes and the regulation of ARRB2 by monocytes was further confirmed by two-colour FACS analysis. Since ARRB2 can serve as an endogenous inhibitor that blocks signal-induced I kappaB degradation and subsequent activation of NF kappa B, we examined the expression of genes controlled by this transcription factor. We observed an up-regulation of TNF alpha, IL-1 beta and iNOS mRNA in cells from allografts. Concomitantly, I kappa B levels were significantly reduced during rejection. Conclusions: Our results suggest that the NF kappa B mediated activation of monocytes in the blood vessels of renal isografts is blocked by the presence of high levels of ARRB2. During acute rejection, however, ARRB2 levels are drastically reduced and classical monocyte activation via NF kappa B is enabled.

135 ALVEOLAR MACROPHAGES EXPRESS INCREASED LEVELS OF EPIDERMAL FATTY ACID-BINDING PROTEIN DURING ACUTE REJECTION OF RAT LUNG ALLOGRAFTS Julia Holler, Anna Zakrzewicz, Wolfgang Kummer, Holger Garn, Winfried Padberg, Veronika Grau Laboratory of Experimental Surgery, Department of General and Thoracic Surgery, University of Giessen Lung Center, Justus-LiebigUniversity Giessen; Institute for Anatomy and Cell Biology, JustusLiebig-University Giessen; Department of Clinical Chemistry and Molecular Diagnostics, Marburg, Germany Keywords: Lung Transplantation, E-FABP, Alveolar macrophages Background: In the lung, epidermal fatty acid-binding protein (EFABP) is expressed by alveolar macrophages (AM) and alveolar epithelial cells type II (AEII). E-FABP is involved in the metabolism of surfactant phospholipids and may regulate inflammation in macrophages. As macrophage activation and surfactant dysfunction are associated with acute rejection, we investigated E-FABP expression after experimental pulmonary transplantation. Methods: Orthotopic left lung transplantations were performed in the fully allogeneic Dark Agouti to Lewis and in the Lewis to Lewis rat strain combinations. E-FABP expression was analyzed in the total lung and in leukocytes obtained by bronchoalveolar lavage. In all experiments, the native right recipient lung was included as an internal control. Results: Immunohistochemistry revealed strong E-FABP expression by AEIIs and a moderate expression by AMs in isografts and right native lungs. In allografts undergoing acute rejection, the number of cells strongly expressing E-FABP was increased due to an accumulation of EFABP-positive AMs. To test for the specificity of the antibodies used in this study, immunoblots of lung homogenates were performed which

Langenbecks Arch Surg (2008) 393:759–815 resulted in a single band of about 14 kDa corresponding to the calculated molecular weight of rat E-FABP. The levels of E-FABP mRNA as revealed by real-time RT-PCR, were higher in allografts compared to the right native lungs and to isografts. Similarly, alveolar leukocytes from allografts displayed higher E-FABP mRNA expression levels than cell from the corresponding right lungs and isografts. Conclusions: After lung transplantation, E-FABP mRNA expression is strongly increased in allografts. We demonstrate for the first time, that E-FABP is up-regulated by AMs during severe inflammation. E-FABP expressed by allograft AMs may be involved in the degradation of damaged surfactant and may promote pro-inflammatory functions.

136 TISSUE PROTECTION IN LUNG TRANSPLANTATION: LOW PERFUSION PRESSURE COMBINED WITH HIGH POSITIVE END EXPIRATORY PRESSURE REDUCE EDEMA FORMATION IN AN EXPERIMENTAL MODEL Andreas Kirschbaum (1), Stefan Schumann (2), Stephan Schliessmann (2), Giskard Wagner (2), Ulrich Goebel (2), Bernward Passlick (1), Josef Guttmann (2) 1: Department of Thoracic Surgery and 2: Section for Experimental Anaesthesiology, University Medical Center, Freiburg, Germany Keywords: Lung transplantation — flush perfusion — edema formation Background: Flush perfusion with a cooled organ protection solution via pulmonary artery is a standard procedure before lung explantation in transplant surgery. Since flush perfusion is administered rapidly there is a risk for development of a pulmonary hyperperfusion edema with negative effects for lung function. We hypothesized that pulmonary edema can be reduced by positive and expiratory pressure (PEEP) in the mechanically ventilated and perfused lung. Methods: Anterograde perfusion, using cooled low-potassium-dextrane- solution was applied at either of two different pressure levels (low PAP: 27 mmHg, high PAP: 40 mmHg) to the pulmonary artery in ventilated, isolated procine lungs. Positive and expiratory presseure was set to either of two levels: low PEEP: 4 cm H2O, high PEEP: 8 cm H2O. The four combinations (lowPAP-lowPEEP; lowPAPhighPEEP; highPAP-lowPEEP; highPAP-highPEEP) have been investigated in 28 mechanically ventilated procine lungs. Before perfusion, subsequent to perfusion and after 4 hours of cold storage the lungs were weighted, the intratidal development of the dynamic compliance was calculated and histopathology was taken. Results: Relative weight increase of the isolated lungs (after perfusion/after storage) was smaller at low PAP (62+32%/42+26%) compared to high PAP (133+54%/87+30%). LowPAP combined with high PEEP resulted in smallest relative weight increase (44+9%/27+12%) compared to all other PAP-PEEP combinations (p<0,05).This group showed also a better compliance and no pathologic changes in the histologic examinations. Conclusions: We conclude that high PEEP alone does not prevent hyperperfusion edema in an isolated procine lung explantation model if PAP is high. Edema formation in isolated and ventilated lungs can be reduced, using small pulmonary artery pressures in combination with high PEEP.

137 IMPAIRED AUTOPHAGIC CLEARANCE AFTER COLD PRESERVATION OF FATTY LIVERS CORRELATES WITH TISSUE NECROSIS UPON REPERFUSION AND IS REVERSED BY ENDISCHEMIC HYPOTHERMIC OXYGENATION Judith Stegemann, Andreas Hirner, Thomas Minor Surgical Research Division, University Hospital Bonn, Bonn

811 Keywords: autophagy, ischemia, preservation, steatosis, oxygen, persufflation Background: Fatty livers are particularly susceptible to mitochondrial alterations after cold preservation including dysfunction and membrane permeabilisation. Depending on the amount of affected mitochondria, and the energetic status of the cell, this results in autophagy, apoptosis or necrosis. Autophagy occurs at low rates in cells to perform homeostatic functions (e.g. lysosomal degradation of defected proteins or injured organelles or depolarized mitochondria). The present study was aimed to evaluate the potential of postconditioning of liver grafts by short term gaseous oxygen insufflation prior to implantation of the organ, using a technique, which has already been shown to enhance energy charge during hypothermia. Methods: A histologically documented moderate macrovesicular steatosis was induced in rat livers by fasting for 2 days and subsequent feeding of a fat-free diat enriched in carbohydrates. Fatty livers were retrieved and flushed via the portal vein with 60 ml of HTK. Control livers (CON) were then stored ischemically at 4°C for 20 h. Postconditioning (PC) was performed in other livers, isolated and stored in the same manner, by insufflation of gaseous oxygen via the caval vein during the last 90 min of the preservation period, while the liver was still immersed in cold HTK. Viability of the livers was then assessed upon 120 min of isolated perfusion in vitro with oxygenated KrebsHenseleit buffer according to validated techniques. Results: PC resulted in a relevant (approx. 5-fold) and significant reduction of parenchymal (ALT, LDH) and mitochondrial (GLDH) enzyme release during reperfusion. Moreover, functional recovery (bile production, oxygen consumption and tissue levels of ATP) was significantly increased to more than twice the values of CON, respectively. On the molecular level we could document a significantly impaired autophagy index in CON (<50% of baseline) already at the end of ischemia and persistent after reperfusion, as evaluated by cleavage of LC3B and confirmed by protein expression of beclin-1, both of which were found restored to near normal values after PC. Moreover, PC induced a twofold increase of mitochondrial chaperones of the HSP70 family. Caspase 3 enzyme activity after 2 h of reperfusion was slightly above baseline values in CON and lightly (factor 1.5) but insifgnificantly enhanced after PC. By contrast, histological signs of tissue necrosis were abundant after reperfusion in CON but largely abrogated by PC, closely corresponding to the parenchymal enzyme leakage. Conclusions: It is concluded that endischemic oxygenation in the cold limits mitochondrial defects and restores basal rates of cellular autophagy, likey to represent a rescue mechanism to maintain cellular homeostasis and tissue survival. Aerobic post-conditioning may thus be regarded a useful tool for more successful preservation of ischemia sensitive steatotic livers.

138 PHARMACOLOGICAL INHIBITION OF P38MAPK IN MACROPHAGES AMELIORATES ISCHEMIA REPERFUSION INJURY AFTER SMALL BOWEL TRANSPLANTATION (SBTX) Michael Praktiknjo, Thomas Pech, Jun Fujishiro, Kareem AbuElmagd*, Jörg Kalff, Andreas Hirner, Nico Schäfer Department of Surgery, School of Medicine, University of Bonn; * Thomas E. Starzl Transplantation Institute, University of Pittsburgh, PA, USA 15213 Keywords: small bowel transplantation, p38MAPK, macrophages, ischemia reperfusion injury Background: Resident macrophages within the tunica muscularis are known to play a crucial role in initiating ischemia reperfusion injury (IRI)

812 after SBTx contributing to graft dysmotility and bacterial translocation. Therefore we investigated the effects of inhibition of macrophages cytokine releasing pathway on IRI in SBTx by pharmacological inhibition of p38MAPK a key player in cytokine synthesis. Methods: Orthotopic isogenic SBTx was performed in rats (LewisLewis). Recipient and donor animals were treated perioperatively with a macrophage specific p38MAPK inhibitor (1 mg/kg, i.v.). Nontransplanted native animals and vehicle treated animals served as control groups (n=8). Animals were sacrificed 3 h and 18 h after reperfusion. Park’s score was used for histological grading. Leukocyte infiltration was investigated by immunohistochemistry and histochemistry, apoptosis by TUNEL-staining and mediator expression by Real-Time-RT-PCR, ELISA and Griess reaction. Smooth muscle contractility was assessed in a standard organ bath under bethanechol stimulation. Statistics consisted of analysis of variance (ANOVA). Results: Inhibition of p38MAPK in macrophages results in significant less leukocyte (ED1, MPO) infiltration and amelioration of graft dysmotility 18 h after reperfusion compared to vehicle treated group. Proinflammatory cytokines and kinetic active mediators were significantly decreased 3 h after reperfusion whereas no significant differences were detectable after 18 h. Histologic evaluation revealed protective effects of p38MAPK inhibition at all timepoints. Conclusions: Early inflammatory processes in the tunica muscularis in SBTx due to ischemia reperfusion injury initiated by activated macrophages are leading to impaired graft motor function. Preoperative treatment with a macrophage specific p38MAPK inhibitor provides protection from IRI with reduced inflammation and graft dysmotility after isogenic SBTx.

139 NICOTINE ATTENUATES MACROPHAGE INFILTRATION IN EXPERIMENTAL RAT LUNG ALLOGRAFTS Markus Hirschburger, Anna Zakrzewicz, Winfried Padberg, Wolfgang Kummer, Veronika Grau Abteilung für Experimentelle Chirurgie, Abteilung für Allgemein-, Viszeral-, Thorax-, Transplantations- und Kinderchirurgie, Universität Giessen Lung Center, Justus-Liebig-Universität Giessen, RudolfBuchheim-Str. 7, 35385 Giessen, Germany Keywords: lung transplantation, acute rejection, acetylcholine Background: Monocytes and macrophages play an important role in acute pulmonary allograft rejection. Acetylcholine has been shown to exert anti-inflammatory effects on these cells via nicotinic acetylcholine receptors. The aim of this study was to test for the hypothesis that a global nicotinic stimulation of pulmonary allograft recipients attenuates acute rejection. Methods: Orthotopic left lung transplantation was performed in the Fisher 344 to Wistar Kyoto rat strain combination. Graft recipients treated with nicotine added to the drinking water were compared to untreated allograft recipients. Graft histopathology, leukocytic infiltration, expression of inducible NO synthase, and cytokine expression were analyzed during the process of acute rejection on day 7 post-transplantation. The right native lung of the experimental animals was included as an internal control. Results: Nicotine treatment resulted in a marked reduction in lung allograft infiltration by alveolar and tissue macrophages. Concomitantly, inducible NO synthase-expression, which was predominantly localized in alveolar macrophages of control allografts, decreased in response to nicotine. In line with these data, less interleukin-1beta was expressed in pulmonary allografts from nicotine-treated recipients compared to controls. Conclusions: Stimulation of nicotinic acetylcholine receptors results in a marked attenuation of important hallmarks of pulmonary allograft rejection. Further studies on the cholinergic system of pulmonary allografts will be needed to design new strategies involving more specific cholinergic therapeutics.

Langenbecks Arch Surg (2008) 393:759–815

140 PERIOPERATIVE USE OF THE CHIMERIC-MONOCLONAL-ANTI-TNF-α ANTIBODY INFLIXIMAB IN SMALL BOWEL TRANSPLANTATION (SBTX) REDUCES ACUTE REJECTION T. Pech, T. Finger, J. Fujishiro, D. Boeker, K. Abu-Elmagd*, J.C. Kalff, A. Hirner, N. Schaefer Department of Surgery, School of Medicine, University of Bonn * Thomas E. Starzl Transplantation Institute, University of Pittsburgh, PA, USA 15213 Keywords: Infliximab, small bowel transplantation, acute rejection Background: After small bowel transplantation TNF-α upregulation in resident macrophages in the tunica muscularis results in local amplification of inflammation and contributes decisively to graft dysmotility, as we have shown in the past. Therefore the aim of this study is to investigate the effectiveness of the chimeric-monoclonalanti-TNF-α antibody Infliximab as perioperative single shot treatment regarding ischemia reperfusion injury and acute rejection in rat small bowel transplantation. Methods: Orthotopic isogenic and allogenic SBTx was performed (Lewis — Lewis or Brown-Norway — Lewis). In the treatment groups Infliximab was injected i.v. directly after reperfusion (5 mg/ kg). Transplanted vehicle (isotonic saline) or sandoglobulin treated animals served as transplanted controls, non transplanted animals as native control. Animals were sacrificed after 3 h (only isogenic SBTx), 24 h and 7 days. Immunohistochemistry, histochemistry, TUNEL-staining, Real-Time-RT-PCR. Contractility was assessed in a standard organ bath. Statistics: (ANOVA). Results: Compared to allogenic controls, the Infliximab treated animals (allogenic SBTx) displayed significant less apoptosis in the tunica muscularis (24 h/7days). Infliximab resulted in significant less leukocyte infiltration (MPO positive cells, ED1 and ED2). Smooth muscle contractility improved significantly after 24 h compared to transplanted control groups with saline or sandoglobulin. Infliximab treatment revealed no significant improvements in isogenic SBTx. Conclusions: Acute rejection activates inflammatory processes in the tunica muscularis leading to acute graft dysfunction. The perioperative treatment with Infliximab during reperfusion improves motility and attenuates the inflammatory response in allogenic SBTx. Effects on ischemia reperfusion injury are not responsible for improved graft function and condition.

141 TGF-β SIGNALING IN BLOOD MONOCYTE DERIVED NEO-HEPATOCYTES FROM ALCOHOLIC PATIENTS AND CONTROLS Sabrina Ehnert, Antje Lehmann, Ulrich Böcker, Steven Dooley, Andreas Nussler Dept. of Traumatology, Technical University Munich, Germany; Dept. General Surgery, Universitätsmedizin Berlin, Campus Virchow, Berlin, Germany; Dept. of Medicine II, University Hospital Mannheim, University of Heidelberg, Germany Keywords: hepatocyte-like cells, TGF-beta, fibrosis, cell therapy Background: End stage liver disease induced by chronic viral infection or compound intoxication, e.g., alcohol and aflatoxin, represented as cirrhosis and hepatocellular carcinoma is continuously increasing in developed countries. Its only definite treatment is currently liver transplantation. However, donor organ scarcity raises a growing interest in new therapeutic options, e.g. cell therapy using hepatocyte-like (NeoHep) cells generated by trans-differentiation of peripheral blood monocytes.

Langenbecks Arch Surg (2008) 393:759–815 Fibrotic areas in the liver display active TGF-β signaling and the resulting signal transduction in liver cells is critically required for progression of chronic liver disease. Therefore, we were interested how NeoHep cells from alcoholic patients and controls respond to this cytokine. Methods: NeoHep cells were generated from monocytes of alcoholic patients and healthy controls and compared to human hepatocytes. We generated an expression profile of Smads and TGF-β receptors by RT-PCR. TGF-β signaling was analyzed by Westernblot, immunofluorescent staining, Smad1/4- and Smad3/4 luciferase reporter assays. Accumulation of fat droplets was visualized by Oil Red O staining. Results: Quality of NeoHep cells was determined by expression of hepatocyte marker genes, e.g. glucose-6-phosphate dehydrogenase, cytokeratin 18 and albumin or glucose and urea formation. NeoHep cells express all the Smads and TGF-β receptors necessary for TGF-β signaling. Except for Smad1 and Smad3 mRNA and protein levels are comparable between NeoHep cells and human hepatocytes. Smad1 and Smad3 is much lower in NeoHep cells (∼15–20%) compared to human hepatocytes. Upon stimulation with TGF-β the receptor (R)-Smads become phosphorylated and translocate to the nucleus. Comparable to the mRNA levels, Smad1/4- and Smad3/4 luciferase reporter assays confirmed a reduced TGF-β signaling via Smad1 and Smad3 in NeoHep cells. However, the signal is sufficient to trigger several cellular responses, e.g. reduction of E-cadherin and increase of β-cathenin. Furthermore, TGF-β induces accumulation of fat droplets in a dose dependent manner. With regard to TGF-β, there was no difference observed between NeoHep cells from alcoholic patients and controls. Conclusions: NeoHep cells have lower expression of Smad1 and Smad3 compared to human hepatocytes, this might give the NeoHep cells a proliferation advantage in fibrotic liver as Smad3 is reported to be involved in cellular damage by TGF-β. Quality control of the generated NeoHep cells showed comparable expression of hepatocyte marker genes as well as glucose and urea synthesis in cells generated from alcoholic patients and healthy controls. Furthermore, TGF-β signaling in NeoHep cells from alcoholic patients was comparable to control cells, suggesting that these cells are not more susceptible to TGF-β damage. This allows the generation of NeoHep cells from patients blood for autologous cell therapy.

142 MYOCARDIAL PROTECTION WITH THE USE OF DESFEROXAMINE AND L-ARGININE Achim Koch, Tamas Radovits, Sivakanan Loganathan, Falk-Udo Sack, Arthur Lichtenberg, Matthias Karck, Gabor B.Szabó Chirurgische Universitätsklinik Heidelberg, Abteilung Herzchirurgie, Im Neuenheimer Feld 110, 69120 Heidelberg, Germany Keywords: Heart transplantation, myocardial protection, desferoxamin Background: Effective myocardial preservation is an important condition for cardiac surgery especially in heart transplantation with long ischemic times. During ischemia and reperfusion myocardial function is altered by oxidative stress. One source of free oxygen radicals are ironiones. We investigated the efficacy of new modifications of the well established HTK solution (Custodiol) in a rat heart transplantation model Methods: Heterotopic transplantation was performed in Lewis rats (n= 30). After one hour of ischemic preservation and one hour reperfusion myocardial function and energy charge potential were assessed. The modifications of HTK solution included the addition of L-Arginine, partial replacement of histidine with acetyl-histidine and reduction of chloride concentration (HTK-1). In a second group, the iron-chelator

813 desferoxamine was added (HTK-2). A third group with Custodiol served as control. Results: After 1 h reperfusion left ventricular systolic pressure (106± 33 vs. 60±39, vs. 67±8 mmHg p<0.05) and dP/dt minimal (−1388± 627 vs −660±446, vs. 871±188 mmHg/s p<0,05) were significantly higher in the HTK-1 group in comparison to HTK-2 and controls. Energy charge potential did not differ significantly between the groups. Conclusions: This study showed that the novel modified HTK-1 solution improves myocardial contractility and relaxation after heart transplantation. Nevertheless addition of the iron-chelator desferoxamine diminished these beneficial effects.

143 MUTAGENIC AND CYTOTOXIC EFFECTS OF IMMUNOSUPPRESSIVE DRUGS ON HUMAN LYMPHOCYTE CULTURES Vilma Oliveira Frick, Claudia Rubie, Thomas Rath, Mathias Wagner and Martin K. Schilling Dept. of General -, Visceral-, Vascular - and Paediatric Surgery, University of the Saarland, 66421 Homburg/Saar, Germany, Department of Nephrology and Transplantation Medicine, Westpfalz-Klinikum, 67653 Kaiserslautern, Germany, Institute of Pathology, University of the Saarland, 66421 Homburg/Saar, Germany Keywords: Immunosuppressive drugs, micronucleus assay Background: Immunosuppressive drugs such as cyclosporine A, mycophenolate mofetil, tacrolimus, and the immunosuppressive agent sirolimus are used effectively to prevent immunologic rejection after solid organ transplantation. The most serious complication among patients undergoing immunosuppressive therapy is the risk of developing cancer. To date it is unknown whether the applied drugs mutagenic properties and thus potentially contribute to an increased cancer risk. Methods: We evaluated the mutagenic and cytotoxic effects of the above-mentioned drugs in human lymphocyte cultures with special consideration given to clinically relevant blood-drug concentrations. Mutagenicity was tested by analysis of micronuclei using the wellestablished cytokinesis-block micronucleus assay with cytochalasin B. To evaluate cytotoxicity, the cytokinesis block proliferation index was calculated. Concentrations used ranged from 0.1–2 µg/mL for cyclosporine A, 1–20 µg/mL for mycophenolate mofetil, 5–40 ng/ mL for tacrolimus, and 2.5–50 ng/mL for sirolimus. We also estimated mutagenicity and cytotoxicity in blood of kidney transplanted patients by using the above mentioned techniques. Results: Cultures supplemented with mycophenolate mofetil or tacrolimus showed higher amounts of micronuclei when compared with solvent controls at all concentrations tested. Cultures supplemented with cyclosporine A also led to a rise in micronuclei number at concentrations of 0.2 µg/mL and 0.4 µg/mL. In contrast with the other immunosuppressive drugs, sirolimus induced only weak mutagenic activity in the micronuclei test at its highest concentration (50 ng/mL). Cytotoxic effects were seen only in mycophenolate mofetil-supplemented cultures at all concentrations tested (P<0.01). Kidney transplanted patients under immunosuppression display a broad reduction of the CBPI (P<0,001) in comparison to healthy persons and a significant increase in MN frequency (P<0,001). Conclusions: All immuosuppressive drugs under investigation displayed mutagenic effects in vitro, indicating that mycophenolat mofetil and tacrolimus show more mutagenic effects in vitro than cyclosporine A or sirolimus. Moreover, all transplanted patients exhibit higher amounts of MN and a noteworthy reduction in the CBPI compared to healthy people.

814

144 THE TRANSCRIPTIONAL REPRESSOR ZEB1 PROMOTES METASTASIS AND LOSS OF CELL POLARITY IN CANCER Otto Schmalhofer (1), Simone Spaderna (1), Joerg Schubert (1), Ulrike Burk (1), Mandy Wahlbuhl (2), Dennis Strand (5), Andreas Eger (4), Jürgen Behrens (3) and Thomas Brabletz (1) 1: Dept. of Visceral Surgery, Univ. of Freiburg, Germany, 2: Dept. of Pathology and 3: Nikolaus-Fiebiger-Center, University of Erlangen, 4: Max F. Perutz Laboratories, Medical University Vienna, Austria, 5: First Dept. of Internal Medicine, Univ. of Mainz, Germany Keywords: ZEB1, LGL2, epithelial-mesenchymal transition, metastasis, cell polarity Background: Invasion and metastasis are the hallmarks of malignant tumor progression and the main cause of death in cancer. The embryonic program “epithelial-mesenchymal transition” (EMT) is thought to trigger invasion by allowing tumor cell dissemination. Disseminating tumor cells are characterized by acquisition of a mesenchymal phenotype, including loss of epithelial cell polarity. Methods: Immunoflourescence, immunohistochemistry, ChIP, nude mice xenografts, migration and invasion assays Results: Here, we describe that the EMT-inducing transcriptional repressor ZEB1 promotes cancer cell metastasis and loss of cellular polarity. Thereby, ZEB1 suppresses the expression of cell polarity factors, in particular of LGL2, which we found reduced in colorectal and breast cancers. We further show that retention of LGL2 expression is critical for the epithelial phenotype and that its loss might be involved in metastasis. Conclusions: By linking EMT, loss of polarity, and metastasis, ZEB1 is a crucial promoter of malignant tumor progression.

145 RECONSTRUCTION OF SEGMENTAL FEMORAL DEFECTS WITH LIVING BONE ALLOGRAFTS COMBINED WITH HOST-DERIVED NEOANGIOGENESIS: MECHANICAL, HISTOLOGIC AND RADIOGRAPHIC ANALYSIS Goetz A. Giessler, Patricia F. Friedrich, Allen T. Bishop Department of Orthopedic Surgery, Microvascular Research, Mayo Clinic, Rochester, MN Background: Living musculoskeletal allografts currently require longterm immunosuppression to maintain viability, impractical due to

Langenbecks Arch Surg (2008) 393:759–815 associated risks for non-life-critical tissue transplantation. We have previously demonstrated an alternative method using implanted hostderived vessels to replace the allogeneic nutrient circulation. These vessels maintain measurable blood flow, generate extensive neoangiogeneic capillaries and form new bone when combined with short-term immunosuppression. In this study, we have used this method to reconstruct large segmental femoral defects. Methods: A segmental femoral defect was created in Dutch-Belted rabbits. Reconstruction was performed using a free vascularized allogeneic femoral diaphyseal transplant from a New Zealand White rabbit. Rigid fixation allowing immediate full weight-bearing was performed. In addition to microvascular repair of the nutrient artery circulation, a pedicled inferior epigastric fascial flap was placed within the medullary canal. Survival time was 16 weeks. Five groups of 10 Dutch-Belted rabbits each included a pedicled autograft control group, and four allograft groups which varied in fascial flap patency (patent or ligated) and immunosuppression with 0.08 mg/kg Tacrolimus (+ or −). Healing was quantified by X-ray. Microangiography and Spalteholz bone clearing allowed quantification of neoangiogenesis. Mechanical properties were evaluated using 4-point bending. Quantitative histomorphometry assessed bone remodeling. Results: X-ray analysis using a grading schema demonstrated an equivalent healing response when autograft controls were compared to immunosuppressed allografts with patent fascial flaps. The roentgenograms of the latter group demonstrated faster healing as well as the lowest relative ultimate strength and elastic modulus values of all groups. This is an indication of biologic activity, including a greater blood supply and a higher rate of bone remodeling than other animals. It correlated with findings from microangiography (the highest amount of neoangiogenesis among all groups) and histomorphometric analysis of bone turnover. Not surprisingly, the lowest angiogenesis and bone remodeling values were found in the nonimmunosuppressed allograft femurs with a ligated intramedullary flap. Conclusions: Surgical angiogenesis from host-derived fascial flaps can provide greater blood flow and improved rates/grading of healing in immunosuppressed allogeneic bone transplants than other groups. Material properties of this group were less than the other groups. Thus, while the vascularized tissue allotransplants treated with immunosuppression and fascial flap implantation maintained flow and viability at levels higher than other groups, we found this to weaken the graft more as well. As the demonstrated active bone turnover ultimately replaces the graft with host-derived cells, this process in the long-term may result in a more stable graft with minor rejection.

Langenbecks Arch Surg (2008) 393:759–815

815

Author index

(The numbers are the abstract numbers) Alakus H 56 Alexis U 47 Aydin F 60 Bierwolf J 127 Borna RM 23 Brabender J 55 Burk U 70 Bünger S 37 Böcker W 130 Camaj P 50 Camaj P 63 Cziupka K 112 Dango S 45 Deduchovas O 90 Dhaouadi O 83 Dietz C 6 Dirhold BM 124 Dold S 101 Eder M 15 Eder M 16 Egaña JT 116 Ehnert S 141 Emons G 52 Evers LH 13 Evers LH 25 Fendrich V 53 Finkenzeller G 1 Frick VI 143 Frick VO 66 Friedrich B 97 Fritzsche B 40 Gaedcke J 80 Gasser M 73 Gasser M 74 Gasser M 133 Gemoll T 72 Gerling M 42

Goetz A 145 Goos M 18 Grade M 49 Graner S 8 Grau 131 Greulich C 9 Gröne J 57 Habbe N 35 Habermann JK 59 Habijan T 10 Hacer SB 7 Hackl C 41 Hager S 128 Henrich D 102 Henrich D 110 Hillmer A 125 Hirschburger M 139 Histing T 12 Holler J 135 Holzbach T 5 Hotz HG 36 Ischenko I 78 Jargon D 109 Kaffarnik M 111 Kayser G 44 Kebschull L 100 Kim M 26 Kirschbaum A 136 Koch A 142 Koch M 3 Koscielny A 93 Kraus A 14 Krex D 32 Kuhn R 79 Königsrainer I 76 Körner Ü 107 Lehnhardt M 86

Lehnhardt M 87 Leupold HG 69 Linnebacher M 64 Löb S 75 Lögters TT 24 Manekeller S 19 Mardin WA 48 Mees ST 34 Mehlhorn A 114 Mehmet AA 106 Moussavian MR 28 Mund R 17 Möslein G 33 Ozimek 82 Ozimek A 81 Partecke LI 85 Pech T 140 Praktiknjo M 138 Prall WC 129 Rezaeian F 123 Richter P 108 Roller J 2 Rubie C 67 Rubie C 68 Rupertus K 4 Rupertus K 43 Rückert F 39 Ryschich E 21 Salem H 115 Sauerland S 29 Schaich M 31 Schanz J 121 Schellerer V 84 Schellhaas E 77 Schmal H 118 Schmalhofer O 144 Schneider M 22

Schramm R 27 Schönberger TS 122 Seeliger H 105 Serra A 92 Simon F 20 Sinis N 126 Sioutis M 95 Slater EP 61 Slater EP 62 Sorg H 98 Speidel V 99 Stegemann J 137 Stoecklein V 94 Strohe A 117 Traub F 58 Ulmer C 30 van den Engel NK 65 van den Engel NK 89 Vilz T 91 Vilz T 96 von Koschitzky H 103 Waldmann J 38 Vallböhmer D 46 Walles T 11 Walles T 120 Warnecke-Eberz U 104 Wehrum D 54 Weigel P 71 Weinhold A 119 Wendel C 51 Wittel UA 113 Witthauer J 88 Wittig C 132 Zakrzewicz A 134