Cell Tissue Res (2007) 330:209–220 DOI 10.1007/s00441-007-0477-4

REGULAR ARTICLE

Gene regulation by homeobox transcription factor Prox1 in murine hepatoblasts Maria Papoutsi & Jozsef Dudas & Jürgen Becker & Marco Tripodi & Lennart Opitz & Giuliano Ramadori & Jörg Wilting

Received: 28 March 2007 / Accepted: 13 July 2007 / Published online: 9 September 2007 # Springer-Verlag 2007

Abstract The homeobox transcription factor Prox1 is expressed in embryonic hepatoblasts and remains expressed in adult hepatocytes. Prox1-null mice show severe deficiencies in liver development, although the underlying

This study was supported by the Deutsche Forschungsgemeinschaft (grant: SFB 402, D4). M. Papoutsi : J. Becker Children’s Hospital, Pediatrics I, Georg August University, Robert-Koch-Strasse 40, 37075 Göttingen, Germany J. Dudas : G. Ramadori Department of Internal Medicine, Section of Gastroenterology and Endocrinology, Georg August University, Robert-Koch-Strasse 40, 37075 Göttingen, Germany M. Tripodi Istituto Pasteur-Fondazione Cenci Bolognetti, Dipartimento di Biotecnologie Cellulari ed Ematologia, Università “La Sapienza”, Rome, Italy L. Opitz DNA Microarray Facility, Georg August University, Humboldtallee 23, 37073 Göttingen, Germany M. Papoutsi Novartis, Basel, Switzerland J. Wilting (*) Center of Anatomy, Department of Anatomy and Cell Biology, Georg August University, Kreuzbergring 36, 37075 Göttingen, Germany e-mail: [email protected]

mechanisms are unknown. We have studied the effects of Prox1 on the transcriptional profile of met-murine hepatocytes (MMH) obtained on embryonic day 14 (ED14). These immortalized murine hepatoblasts express numerous hepatoblast markers, but not Prox1. We have performed stable transfection with Prox1 cDNA, analyzed the transcriptome with Agilent mouse whole-genome microarrays, and validated genes by quantitative reverse transcription/ polymerase chain reaction. We have observed the upregulation of 22 genes and the down-regulation of 232 genes, by more than 12-fold. Many of these genes are involved in metabolic hepatocyte functions and may be regulated by Prox1 directly or indirectly, e.g., by the downregulation of hepatocyte nuclear factor 4α. Prox1 induces the down-regulation of transcription factors that are highly expressed in neighboring endodermal organs, suggesting a function during hepatoblast commitment. Prox1 does not influence the proliferative activity of MMH but regulates genes involved in liver morphogenesis. We have observed the up-regulation of both type-IVα3 procollagen and functionally active matrix metalloproteinase-2 (MMP-2), an observation that places Prox1 at the center of liver matrix turnover. This is consistent with MMP-2 expression in hepatoblasts during liver development and with the persistence of a basal lamina around the liver bud in Prox1deficient mice. Our studies suggest that Prox1 is a multifunctional regulator of liver morphogenesis and of hepatocyte function and commitment.

Keywords Liver . Prospero-related homeobox 1 . Hematopoietically expressed homeobox . Hepatocyte nuclear factor 4α . Matrix metalloproteinase-2 . Type-IV collagen . Met-murine hepatocytes

210

Introduction The embryonic liver develops by interactions between endodermal and mesodermal cells. The liver parenchyma (hepatocytes and cholangiocytes/biliary epithelial cells/ ductular cells) is derived from an endodermal bud located near the foregut/mid-gut boundary. The cells of the liver bud are generally regarded as bipotential hepatoblasts giving rise to hepatocytes and intra-hepatic ductular cells including the canals of Hering (Grompe and Finegold 2001). The stroma of the liver, blood vessels, lymph vessels, Kupffer cells, and peritoneal epithelium are of mesodermal origin, which also seems to hold true for the stellate/Ito cells (Cassiman et al. 2006). The development and differentiation of hepatocytes are regulated by a network of genetic and epigenetic mechanisms. The generation of the endodermal germ layer is an essential prerequisite for all endoderm-derived organs and is controlled by homeobox transcription factors such as Mixer and SRY-homeobox 17α/β (Henry and Melton 1998; Hudson et al. 1997). The transcription factor GATA4 regulates morphogenetic movements during the folding of the gut into a tube (Molkentin et al. 1997). Regionalization of the endoderm along the cranio-caudal axis is again regulated by homeobox transcription factors, e.g., in the anterior region, by orthodenticle homeobox 2, homeobox expressed in ES cells 1 (Hesx1), and hematopoietically expressed homeobox (Hex). Regional differentiation is thus progressively defined by transcription factors, which mark presumptive territories for esophagus and stomach, liver, duodenum and pancreas, and small and large intestine (for a review, see Grapin-Botton and Melton 2000). The liver primordium is initially located at the anterior intestinal portal and requires signals from the adjacent cardiac mesoderm for its development (Jung et al. 1999; Le Douarin 1970). In the mouse, the liver bud is visible at embryonic day (ED) 9.5, although the cells express albumin and α-fetoprotein (AFP) mRNA 24 h earlier (Gualdi et al. 1996). A function for fibroblast growth factors (FGF) 1, 2 and 8, and FGF receptors (FGFR) 1 and 4 during liver induction by cardiac mesoderm has been discussed (Jung et al. 1999). Expansion of the liver bud by the proliferation of hepatoblasts is controlled by growth and transcription factors, as evidenced by the severely reduced liver size in mutant mice. Hepatocyte growth factor (HGF), most likely secreted by blood vascular endothelial cells, acts via its transmembrane receptor met protooncogene (c-met), which is expressed by hepatoblasts, and activates the jun kinase JNK1 pathway (Schmidt et al. 1995). Similarly, mice deficient for the homeobox transcription factors Hlx (Hentsch et al. 1996), Hex (Martinez Barbera et al. 2000), and prospero-related homeobox 1 (Prox1; Sosa-Pineda

Cell Tissue Res (2007) 330:209–220

et al. 2000), all of which are expressed in hepatoblasts, show arrest of liver development at early stages. Prox1 is highly conserved between species. It is expressed in the liver of Xenopus tadpoles, zebrafish, chick, mouse, rat, and human (Dudas et al. 2004; Glasgow and Tomarev 1998; Ny et al. 2005; Oliver et al. 1993; Rodriguez-Niedenfuhr et al. 2001). Human and mouse Prox1 proteins are 98% identical. The homeodomains of chick and mouse Prox1 are identical at the amino acid level and are 65% and 67% similar to the prospero homeodomains of Drosophila and Caenorhabditis elegans, respectively (Tomarev et al. 1998; Zinovieva et al. 1996). Prox1 null mice die around ED14.5. They show a 70% reduction in liver size. The differentiation markers albumin and AFP are expressed, but high concentrations of Ecadherin and the continuous basal lamina surrounding the liver bud suggest that the Prox1-null hepatoblasts fail to migrate into the adjacent mesoderm (Sosa-Pineda et al. 2000). In the HepG2 hepatocarcinoma cell line, Prox1 binds the hepatocyte nuclear factor 4α (HNF4α) and thereby regulates bile acid synthesis and gluconeogenesis (Song et al. 2006). Furthermore, Prox1 represses the human liver receptor homolog-1, an essential regulator of cholesterol 7α hydroxylase (CYP7A1), which catalyzes a ratelimiting step in bile acid synthesis (Qin et al. 2004; Steffensen et al. 2004). Although some of the Prox1 functions in hepatocytes/ hepatocarcinoma cells have been described, the global functions of Prox1 in hepatoblasts have not been analyzed as yet. For this purpose, we have studied ED14 met-murine hepatocytes (MMH), which represent bipotential hepatoblasts, isolated from transgenic ED14 mouse embryos expressing a truncated, constitutively active form of the human HGF receptor, c-Met (Spagnoli et al. 1998). The cells represent non-transformed immortalized liver cells and possess several hepatocyte/hepatoblast markers in vitro. They support hematopoiesis and express aldolase B, albumin, β-fibrinogen, α1-antitrypsin, and hepatocyteenriched transcription factors such as HNF4α, HNF1α, HNF1β, and the CCAAT enhancer-binding protein (C/EBP; Aiuti et al. 1998; Amicone et al. 1997; Spagnoli et al. 1998). Our studies show that ED14-MMH are negative for Prox1 but positive for other hepatoblast markers, e.g., Hex. Transfection of the cells with human Prox1 cDNA reveals a large number of regulated genes involved in liver development and hepatocyte function. Specifically, Prox-1 regulates type-IV collagen turnover by inducing both active matrix metalloproteinase-2 (MMP-2) and type-IVα3 procollagen, which establishes a liver-specific extracellular matrix (ECM) and, during development, may facilitate migration of hepatoblasts into the septum transversum mesenchyme. Furthermore, Prox1 down-regulates transcription factors that are highly expressed in neighboring

Cell Tissue Res (2007) 330:209–220

endodermal organs, suggesting a function of Prox1 during hepatoblast commitment.

Materials and methods Cell lines, cell isolation, and cell culture Hepatocytes were isolated from adult rats by means of collagenase treatment in a recirculating in-situ perfusiontechnique and cultured as reported previously (Hartmann et al. 1990). Hepatic stellate cells (HSCs), Kupffer cells (KCs), and liver myofibroblasts (MFs) were isolated and cultured as described elsewhere (Knittel et al. 1999). HepG2 and Hep3B hepatoma cell lines were purchased from the American Type Culture Collection (ATCC, Manassas, Va., USA) and cultured according to the recommendations of the provider. Hepa 1–6 cells were purchased from DMSZ (Braunschweig, Germany). MMH were cultured as described (Spagnoli et al. 1998).

211

staining of nuclei was performed with 4,6-diamidino-2phenylindole (DAPI; Molecular Probes). RNA isolation, polymerase chain reaction, and Northern blot analysis Total RNA was isolated from cultured cells by the guanidinium isothiocyanate and CsCl ultracentrifugation method. A total of 5 μg RNA was separated by agarose gel electrophoresis, transferred onto nylon membrane, and finally hybridized with a specific 32P-dCTP-labeled human Prox1 cDNA probe. Probes were labeled by nick translation with a kit from Amersham (Buckinghamshire, UK). The human Prox1 cDNA probe was as published previously (Tomarev et al. 1998). In addition, a probe was generated from RNA of rat hepatocytes as described by Dudas et al. (2004). Reverse transcription (RT) was performed with a first-strand synthesis kit (Amersham). cDNA transcripts were used for the polymerase chain reaction (PCR) with the Faststart PCR kit (Roche, Mannheim, Germany). Transfection of ED14-MMH by electroporation

Immunohistology and immunocytology For immunofluorescence studies of histological sections, specimens were fixed in 4% paraformaldehyde for 10 min and rinsed in phosphate-buffered saline (PBS). Mouse embryos were embedded in Tissue Freeze Medium (Leica, Bensheim, Germany). Cryosections of 20 μm thickness were prepared. Non-specific binding of antibodies was blocked with 1% bovine serum albumin (BSA). The Prox1 antibody (Reliatech, Braunschweig, Germany) was diluted 1:1,000 and incubated with the sections overnight, as described previously (Papoutsi et al. 2000; RodriguezNiedenfuhr et al. 2001). Monoclonal anti-MMP2 antibody (Chemicon Int., Temecula, Calif.) was diluted 1:1,000, polyclonal anti-cMet antibody (Santa Cruz Biotechnology, Calif.) was diluted 1:80, and monoclonal anti-CK8 (cytokeratin 8) and polyclonal anti-AFP antibodies (DAKO, Glostrup, Denmark) were diluted 1:50. For immunofluorescence studies of cultured cells, we used unfixed fresh cultures of HepG2, MMH, and primary rat hepatocytes in small Petri dishes. The cells were permeabilized for 10 min with 0.1% Triton X-100 (Sigma, Steinheim, Germany) in blocking solution (1% BSA in PBS). Subsequently, cells were incubated with polyclonal anti-Prox1 antibody diluted 1:700 in PBS, for 1 h. AntiCK8 and anti-AFP were similarly applied (1:50). After the sections or cells had been rinsed, the secondary antibodies, viz., Alexa 488-conjugated goat-anti-mouse antibody (Molecular Probes) or Alexa 594-conjugated goat-anti-rabbit antibody were applied (diluted 1:200) for 1 h. In controls, primary antibody was omitted. Counter-

ED14-MMH cells were transfected with the pcDNA 3 plasmid (Invitrogen, Carlsbad, Calif.) containing the 2,500 bp of the human Prox1 coding region (Tomarev et al. 1998): 1×107 cells in 250 μl MMH-medium (Spagnoli et al. 1998) containing 25 mM HEPES were transfected with 10 μg pcDNA 3 or pcDNA 3-Prox1 in electroporation cuvettes (50×4 mm; Eurogentec, Belgium) with 250 V/1650 μF by using an Easy-jet electroporator (Equibio, UK). Microarray analysis Microarrays were performed at the core facility of the Medical Faculty of the Georg August University, Göttingen, Germany. We compared the expression profile of stably Prox1-transfected MMH with that of pcDNA 3-vector control transfectants. RNA was isolated from the cells with the RNeasy kit (Qiagen). The quality of the RNA was analyzed with Agilent Bioanalyzer 2100 Expert. Fluorescent-labeled probes for hybridization were produced with the SMART fluorescence probe amplification kit (BD Biosciences). In an amplification step, double-stranded cDNA was synthesized by PCR, the products being purified on PCR quick columns (Qiagen). Following cDNA synthesis, aminoallyl-modified dUTP was incorporated into the cDNA during several rounds of primer extension. In a coupling step, N-hydroxysuccinimide-activated Cy3 and Cy5 dyes (Amersham Bioscience) were used to produce labeled probes. For hybridization of the probes, a mixture of differently labeled (Cy5 and Cy3) probes was prepared; 0.7–1.5 μg Cy-labeled DNA was used per

212

Array (44 k mouse whole-genome oligo microarrays from Agilent). Hybridization was performed overnight at 63°C. According to the project design, various statistical models were used as designed by Prof. Dr. E. Brunner (Department of Medical Statistics, Georg-August University, Göttingen), in combination with the Resolver gene expression data analysis system (Rosetta Inpharmatics).

Cell Tissue Res (2007) 330:209–220 Table 1 Primers used (for an explanation of gene names, see Tables 2, 3; fwd forward, rev reverse) Gene

Sequence

β-Actin

fwd rev fwd rev fwd rev fwd

Amotl2 Cdh17

Real time RT-PCR Col4a3

Relative quantification of mRNA expression for Prox1 and other selected genes was carried out by real-time RT-PCR (qRT-PCR) with Sybr Green Dye on a thermal cycler ABI PRISM 7700 (Perkin-Elmer-Applied Biosynthesis, Foster City, Calif.). Total RNA was prepared from 106 cells with the RNeasy kit from Qiagen (Hilden, Germany) according to the manufacturer’s instructions. Only samples that gave an optical density (OD)260/OD280 ratio of 1.7–2.0 were used for further applications. RT-PCR was carried out according to the manufacturer’s instructions. cDNA was made from 1 μg total RNA, random primer, and Moloney murine leukemia virus reverse transcriptase. The RNA was randomly primed for 10 min at 60°C and subsequently subjected to RT for 1 h at 37°C followed by heating at 95°C for 5 min. PCR was performed with the cDNA aliquots listed in Table 1. The primers were purchased from Thermohybaid (Ulm, Germany). PCRs were carried out in triplicate with the PCR Mastermix for Sybr Green1-kit from Eurogentec (Seraing, Belgium) according to the manufacturer’s instructions. After an initial denaturation step of the cDNA (10 min at 95°C), a two-step PCR was performed (15 s at 95°C; 1 min at 60°C, 40 cycles). Template dilution experiments were performed to ensure that the efficiency of the PCRs for the target and the reference genes were identical. Standard curves were calculated by referring the threshold cycle (CT) to the logarithm of each cDNA dilution step. Detection of specific amplicons was assessed by the generation of melting curves and agarose-gel electrophoresis. The CT values obtained for Prox1 were normalized to corresponding CT values of β-actin. To generate the relative expression levels, each of the normalized values were divided by the normalized values of the calibrator. We designated the control-transfected cells as the calibrator.

Cyp2c40 Elf3 Fez2 Fgf1 Fgf15 Fgf2 Fgf21 Foxa3 Gjb1 Glipr1 Habp2 Hex Hnf4a Il24 Lbp Mmp2 Onecut2 Prox1 Tcea3

rev fwd rev fwd rev fwd rev fwd rev fwd rev fwd rev fwd rev fwd rev fwd rev fwd rev fwd rev fwd Rev fwd rev fwd rev fwd rev fwd rev fwd rev fwd rev fwd rev fwd rev

5′-GCATCCCCCAAAGTTCACAA-3′ 5′-AGGACTGGGCCATTCTCCTT-3′ 5′-TGGAGTAGGTTAGAGGTCACTTGAGA-3′ 5′-CTGAACCGGCACGCTTAGTC-3′ 5′-CGAGCTGATTGTGACGGATGT-3′ 5′-TCCTCGCTGACGTTTGCTT-3′ 5′-TCTAAACAGTACATAACATGGCA TTTACA-3′ 5′-ACATGGTTGCGAAGAATCTATAACAT-3′ 5′-CACCCTGTGATCCCCAATTC-3′ 5′-TGGAAAACAATGGAGCAGATGA-3′ 5′-AGTTTATCCGAGACATCCTAATCCA-3 5′-GCCGGTTCTCCCACTTCA-3′ 5′-GGGAGCCGCTTCCTGTTC-3′ 5′-CAGCATGTTCCAGAATAACACCAT-3′ 5′-GAAGCATGCGGAGAAGAACTG-3′ 5′-CGAGGACCGCGCTTACAG-3′ 5′-CAGGACCAGAAACCCTCAAACT-3′ 5′-GGTCCCCGGTTTCAAAGAAG-3′ 5′-TGGTATGTGGCACTGAAACGAA-3′ 5′-TCTGTCCAGGTCCCGTTTTG-3′ 5′-GTACCTCTACACAGATGACGACCAA-3′ 5′-TGCGCCTACCACTGTTCCA-3′ 5′-ACTACCGGGAGAACCAGCAA-3′ 5′-AAGCAGTCATTGAAGGACAGTGAA-3′ 5′-GCCTCCGGCATCTGCAT-3′ 5′-GCCCGGATGATGAGGTACAC-3′ 5′-CCACCAGCCCGGAATATG-3′ 5′-TCCATGCTTTTGCAATTTGG-3′ 5′-CGGCATGCAGGCCAAA-3′ 5′-AAACCTGGATCTCCGTCTGTGT-3′ 5′-CAGACGCACCACCATCAATTT-3′ 5′-TGTGAGGACAAGCCTCGAAAA-3′ 5′-TGTCACTAGTCAATGGGAACAAGAG-3′ 5′-CTATCCAGTCTCACAGCCCATTC-3′ 5′-GAGTGCACACCAGCGGTTT-3′ 5′-CAAAGCGACTTCTGTATCCAACTG-3′ 5′-GATGGAGATCGAAGGCTTTGTG-3′ 5′-CCACGCTAAGCCGGAAGA-3′ 5′-CCCGGTTTCCCTAAGCTCAT-3′ 5′-GTCCACGACGGCATCCA-3′ 5′-AGCGCATGTCTGCCTTACG-3′ 5′-CCTGTCTTTGTTTGGTTCTTGCT-3′ 5′-CATGCGCTCGGAGATCCT-3′ 5′-TCGCACATCTCATTATCTGACCTT-3′ 5′-GGCCTGAGGCGGAATGT-3′ 5′-CCATTTCCTCCGCTGTCATC-3′ 5′-GCAGGCAAAATCCGTCAGA-3′ 5′-CACCAAGTTACAGTGGAATGAAGAA-3′

Zymography

Tm4sf3

Supernatants of MMH-wt and the MMH-vec- and MMHProx1-transfected cells were examined by gelatin zymography in order to analyze their MMP-2 activity. Cells were grown in 10-cm dishes, washed twice with PBS, and cultured with serum-free RPMI medium for 24 h. Supernatants were collected and centrifuged at 4,000g for 30 min at 4°C

to remove cellular debris. Conditioned media were concentrated with the Centricon 10 system (Amicon, Millipore, Bedford, Mass.). SDS-polyacrylamide gel electrophoresis was performed with samples derived from 4×104 cells on a

Cell Tissue Res (2007) 330:209–220

10% polyacrylamide gel with 0.1% (w/v) gelatin under non-reducing conditions. Gels were washed twice with 2.5% (v/v) Triton X-100 to remove SDS and incubated with substrate buffer [50 mmol/l TRIS-HCl pH 7.6, with 5 mmol/l CaCl2 and 0.02% (v/v) NaN3] overnight at 37°C. Gels were stained with 0.5% (w/v) Coomassie Brilliant Blue R-250 in 30% (v/v) methanol and 10% (v/v) acetate. Cell proliferation assay Cells (104 per well) were seeded into 96-well culture plates in 200 μl culture medium. At various times, cells were fixed by adding 50 μl 25% glutaraldehyde. After incubation for 20 min, the medium was removed, and plates were washed twice with deionized water. The plates were allowed to dry, and 100 μl crystal violet solution (0.1% in deionized water) was added to each well. After 20 min, the dye was removed. The plates were washed three times with deionized water and dried again. The dye was resolved in 100 μl 10% acetic acid, and absorption at 570 nm was determined by using a micro-plate reader (Molecular Devices, Sunnyvale, Calif.).

213

Fig. 1 Prox1 Northern blot analysis of liver cells (28S 28S RNA used as loading control). Prox1 is expressed in hepatocarcinoma cell lines HepG2 (lane 1) and Hep3B (lane 11). Isolated primary hepatocytes (HC) are positive for Prox1 (lane 5), whereas ED14-MMH (lanes 2– 4), hepatic stellate cells (SC, lane 6), myofibroblasts (MF, lanes 7, 8), and Kupffer cells (KC, lanes 9, 10) are Prox1-negative

and rat (Dudas et al. 2006, 2004; Rodriguez-Niedenfuhr et al. 2001). Immunocytological studies showed a fine granular Prox1-staining pattern in the nuclei of primary hepatocytes (not shown) and HepG2 cells, whereas ED14MMH did not contain Prox1 (Fig. 2). CK8, a factor present in hepatocytes, cholangiocytes, and oval cells (the latter being regarded as adult bipotential hepatocyte progenitors;

Matrigel invasion assay Chambers with 8-μm-pore polycarbonate membranes were coated with Matrigel on the upper side (Becton-Dickinson, San Diego, Calif.).MMH-wt, MMH-vec and MMH-Prox1 cells (5×105) were seeded in fetal calf serum (FCS)-free medium in the upper compartment, whereas medium enriched with 10% FCS was added to the lower compartment. After 16 h and 24 h, cells attaching to the lower compartment were fixed in 4% paraformaldehyde and stained with hematoxylin and eosin. Representative areas were photographed at 20× magnification. The assay was performed in triplicate for each time point.

Results ED14-MMH express hepatoblast markers but do not express Prox1 We isolated various liver cell types from rat liver and studied their Prox1 mRNA expression in comparison with human hepatocarcinoma cell lines HepG2 and Hep3B and with MMH cells derived from ED14 mouse embryos. Northern blot analysis showed that primary hepatocytes (HC), HepG2, and Hep3B expressed Prox1, whereas ED14MMH, HSCs, MFs, and KCs were Prox1-negative (Fig. 1). This result was confirmed at the protein level with antiProx1 antibodies. We had previously shown that these antihuman-Prox1 antibodies cross-reacted with chick, mouse,

Fig. 2 Prox1 expression in cultured liver cells. Left DAPI staining of nuclei. Right Anti-Prox1 staining. a, b HepG2 cells. Negative control without primary antibody. c, d HepG2 cells. Prox1 protein is located in the nuclei. e, f ED14-MMH. Prox1 is not detectable. Bar30 μm

214

Grompe and Finegold 2001), was expressed in the cytoplasm of ED14-MMH (Fig. 3). Additionally, after treatment with transforming growth factor-β (TGFβ), the morphology of the cells changed as they took up a longitudinal “palmate” shape, as described previously (Spagnoli et al. 1998). This showed that the cells had preserved their characteristics over years. Furthermore, the cells exhibited AFP immunofluorescence, typical of bipotential liver cells (Fig. 4a). We also studied the expression of the homeobox transcription factor HEX in ED14-MMH by means of real-time RT-PCR. Other than the thymus, the hepatoblasts are the only endoderm-derived cells that express Hex (Martinez Barbera et al. 2000). Compared with the mouse hepatoma cell line Hepa 1–6 (Darlington et al. 1980), we found a four-fold higher expression in both vector controls and Prox1-transfected MMH (Fig. 4b). In summary, ED14-MMH possess characteristics of hepatoblasts. They support hematopoiesis and express numerous liver-specific genes. Their response to TGFβ is the same as has been described previously (Spagnoli et al. 1998). After isolation from ED14 mouse embryos, the cells have obviously lost their Prox1 expression and serve as an ideal cell line to study Prox1-regulated genes in hepatoblasts. Transfection of ED14-MMH with Prox-1 cDNA Transfection of ED14-MMH was performed with pcDNA 3 plasmid containing 2,500 bp of the human Prox1-coding region (Tomarev et al. 1998). Human and mouse Prox1 proteins are 98% identical. Transfection was performed by electroporation, and cell clones were selected with the

Fig. 3 Cytokeratin 8 (CK8) expression in ED14-MMH. a, c DAPI staining of nuclei. b, d CK8 immunofluorescence. CK8, a marker of hepatocytes, cholangiocytes, and oval cells, is present in the cytoplasm of ED14-MMH. The cells possess an epithelial morphology. c, d After treatment of ED14-MMH with TGFβ, the cells take up an elongated “palmate” morphology. Bar 20 μm

Cell Tissue Res (2007) 330:209–220

aminoglycoside antibiotic G418. Expression of Prox1 mRNA in stably transfected clones was quantified by realtime RT-PCR (Fig. 5). Expression of Prox1 in untreated MMH was set as 1. Expression in vector-transfected controls was less than 1. MMH transfected with the plasmid containing human Prox1 cDNA showed a 317- and 610fold increase of Prox1 in clones 3.0 and 3.2, respectively. Alexander cells (primary liver carcinoma) showed a 570fold expression. Prox1 protein in transfected cells was studied by immunofluorescence. The data showed that, in the cells of clone 3.2, Prox1 had a normal nuclear localization (Fig. 6). We chose clone 3.2 for all further analyses, because both the expression level and the localization of Prox1 were comparable to those of normal hepatoblasts. Prox1 mRNA expression in clone 3.2 was also verified with Agilent whole-mouse-genome microarrays (Table 2). Array data demonstrated a 32-fold (25) up-regulation as compared with the vector controls. In general, we observed the upregulation of 22 genes and the down-regulation of 232 genes by Prox1, by more than 12-fold. The complete list of regulated genes is published at NCBI/GEO (accession numbers: GSM190798, GSM190799, GSM190800, GSM190801, GSM190802, GSM190803; series number: GSE7867). A large number of transcription factors were regulated. Among these, the endodermally expressed transcription factors Foxa3 and HNF4α were downregulated (Table 2). Onecut2, which activates a number of liver genes such as HNF3β, was also down-regulated. Furthermore, we observed Prox1 regulation of transcription factors, which are highly expressed in endodermal cells

Cell Tissue Res (2007) 330:209–220

215

hepatoblasts expressed MMP-2, and those hepatoblasts located in the periphery of the liver and interacting with the mesenchyme had highest MMP-2 expression. Additionally, the transverse septum also displayed high levels of MMP-2 (Fig. 8a,b). Our data thus suggest that Prox1 induces MMP-2 expression in hepatoblasts; this may be of major importance for early liver morphogenesis. We therefore expected that Prox1-positive MMH would possess higher invasive capacities than would wild-type and vectorcontrol MMH. However, the Matrigel invasion assay showed that the opposite was true (Fig. 8c,d). The reason for this observation may be that Prox1 up-regulates not only MMP-2, but also type-IV α3 procollagen more than 12-fold (micro-array) and five-fold (qRT-PCR). This places Prox1 in the center of extracellular matrix turnover, making it important for both liver development and maintenance.

Discussion

Fig. 4 α-Fetoprotein (AFP) immunofluorescence and Hex expression in ED14-MMH. a AFP (green) is expressed in the cytoplasm of the cells. Nuclei are counterstained with DAPI. Bar20 μm. b Both vector control cells (MMH-vec) and Prox1-transfected ED14-MMH (MMHProx1) express four-fold more Hex than the mouse hepatoma cell line Hepa 1–6

other than hepatocytes. Elf3, which is highly expressed in gallbladder, was down-regulated. Ehf, which is high in gallbladder, lung, and intestine, was down-regulated. Rara, which is highly expressed in stomach, and Zfp99, which is high in intestine and pancreas, were also down-regulated. Additionally, Tcea3, which is highly expressed in stomach and pancreas, was down-regulated. Taken together, these results suggest that Prox1 down-regulates transcription factors that are highly expressed in neighboring endodermal organs and thereby that it specifies hepatoblast commitment. Proliferation of MMH was not altered by Prox1 (data not shown). Table 3 shows a selection of genes most highly regulated by Prox1 in ED14-MMH and validated by real-time RTPCR. Among these, we have concentrated on MMP-2, which is a collagenase specific for type-IV collagen of basal laminae. MMP-2 was up-regulated in Prox-1transfected MMH approximately 26-fold in the microarrays and 16-fold in qRT-PCR. Zymography showed that MMP-2 secreted by Prox1-transfected MMH was functionally active (Fig. 7). In mouse embryos, Prox1-positive

The master control gene for liver development has not as yet been found. Indeed, liver development seems to be regulated by a network of signaling molecules expressed either within endodermal compartments or secreted by neighboring tissues such as the heart and transverse septum (Grapin-Botton and Melton 2000). The homeobox transcription factor Prox1 is an essential member of this regulatory network. It is expressed in the embryonic midgut endoderm before the anlagen of liver and pancreas become morphologically visible (Burke and Oliver 2002). In the pancreas, Prox1 is constitutively expressed in

Fig. 5 Quantitative real-time RT-PCR of Prox1 mRNA expression in ED14-MMH. Expression of Prox1 in untreated met-murine hepatocytes (MMH-norm) was set as 1. Expression in vector-transfected controls (MMH-pcDNA) was less than 1. MMH transfected with the plasmid pcDNA3 containing human Prox1 cDNA showed a 300- and 600-fold increase of Prox1 in clone 3.0 and clone 3.2, respectively. Alexander cells (primary liver carcinoma; Alexander (PCL)) showed a 570-fold expression. We chose clone 3.2 for micro-array-gene analysis and functional analyses

216

Cell Tissue Res (2007) 330:209–220

Fig. 6 Prox1 protein expression in transfected ED14-MMH. a, c DAPI staining. b, d Anti-Prox1 immunofluorescence. a, b Control cells transfected with pcDNA 3.1 plasmid. Prox1 is not detectable. c, d Cells transfected with Prox1-pcDNA 3.1 plasmid. Prox1 is expressed in the nuclei. Bars 40 μm (a, b), 30 μm (c, d)

exocrine cells but is lost from endocrine islands. Highly differentiated pancreatic carcinoma cell lines display higher levels of Prox1 than poorly differentiated lines (Schneider et al. 2006). Similarly, during liver development, Prox1 remains stably expressed in the hepatocyte lineage but is lost from cholangiocytes (Dudas et al. 2004). Hepatocarcinoma cell lines HepG2 and Hep3B express Prox1. However, a comparison of their expression levels with normal primary hepatocytes is difficult, because the latter lose Prox1 expression after 72 h of culture (own unpublished data). The early expression of Prox1 in embryonic hepatoblasts and its stable expression in adult hepatocytes, even during liver regeneration (Dudas et al. 2006), suggest that it performs complex functions during both the development and differentiation of hepatocytes.

To study the complex functions of Prox1 in hepatoblasts, we have used the immortalized mouse embryo liver cell line MMH. Several MMH cell lines exist. The one that we have used was originally isolated from ED14 mouse embryos (Amicone et al. 1997; Spagnoli et al. 1998). The cells possess numerous hepatoblast/hepatocyte characteristics. As shown here, they express Hex, CK8, and AFP and, as shown previously, support hematopoiesis and additionally express aldolase B, albumin, β-fibrinogen, α1-antitrypsin, HNF4α, HNF1α, HNF1β, and C/EBP (Aiuti et al. 1998; Amicone et al. 1997; Spagnoli et al. 1998). The cells are negative for Prox1. Expression may have been lost during cultivation, as can be observed for primary hepatocytes. However, this makes the cells an ideal tool to study Prox1 effects in a hepatoblast lineage.

Table 2 Prox1-regulated transcription factors in ED14-MMH Gene

Symbol

Accession number

Agilent clone ID

Regulation status

P-value

qRT-PCR status

Prospero-related homeobox 1 Forkhead box A3 Hepatic nuclear factor 4, alpha

Prox1 Foxa3 Hnf4a

NM-008937 NM-008260 NM-008261

One cut domain, family member 2 E74-like factor 3 Ets homologous factor Retinoic acid receptor, alpha Zinc finger protein 99 Transcription elongation factor A 3

Onecut2 Elf3 Ehf Rara Zfp99 Tcea3

NM-194268 NM-007921 NM-007914 NM-009024 NM-023322 NM-011542

52006081 52039104 52013063 52040750 52024837 52002997 52041325 52007216 52036124 52006204 52038654

5.0↑ 2.76↓ 3.81↓ 3.27↓ 2.42↓ 2.35↓ 5.43↓ 2.12↓ 3.10↓ 1.36↓ 2.98↓

2.82e-05 0.0001 2.6e-06 1.26e-05 8.21e-05 8.31e-05 0.0004 1.9e-05 2.63e-05 0.001 0.0004

+ + + + + + + − − + +

Values of up-regulation (↑) and down-regulation (↓) are expressed as the log2 ratio compared with a pool-clone of control transfectants (+ genes verified by real-time RT-PCR, − genes not verified by real-time RT-PCR).

Cell Tissue Res (2007) 330:209–220

217

Table 3 Selection of highly Prox1-regulated genes in ED14-MMH Gene

Symbol, synonym

Accession number

Agilent clone ID

Regulation status

P-value

qRT-PCR status

Growth factors Fibroblast growth factor 1

Fgf1, Fgfa

NM-010197

Fibroblast growth factor Fibroblast growth factor Fibroblast growth factor Hyaluronic acid binding

Fgf2, bFGF Fgf15 Fgf21 Habp2

NM-008006 NM-008003 NM-020013 NM-146101

52011433 52027409 52009122 52040958 52040430 52007036 52024314

3.2↓ 4.8↓ 4.0↓ 3.7↑ 3.8↓ 5.4↓ 5.4↓

0.000587 3.95e-05 0.000298 0.00308 6.21e-05 2.77e-05 2.73e-05

+ + + + + + +

Mmp2

NM-008610

Col4a3

NM-007734

52035028 52031966 52020840

4.7↑ 4.9↑ 3.48↑

6.95e-05 5.68e-05 8.93e-05

+ + +

Lbp Tm4sf3

NM-008489 NM-146010

52027155 52008943

5.3↓ 4.40↓

6.8e-05 8.63e-08

+ +

Cdh17

NM-019753

52018540

5.81↓

2.39e-05

+

Cyp2c40

NM-010004

52002594 52040032

3.96↓ 7.88↓

7.4e-05 6.5e-05

+ +

Gjb1

NM-008124

52042940

6.55↓

4.23e-05

+

Il24

NM-053095

52037760 52016185

3.44↑ 2.2↑

5.16e-05 0.03159

+ +

Glipr1

NM-028608

52043209

3.92↑

3.95e-05

+

Amotl2 Fez2

NM-019764 NM-199448

52031486 52031321

3.32↑ 5.03↑

3.22e-05 6e-05

+ +

2 15 21 protein 2

ECM turnover Matrix metalloproteinase 2 Procollagene type-IV alpha 3 Transport proteins Lipopolysaccharide binding protein Tetraspanin 8 Cell adhesion Cadherin 17 Oxidoreductase activity Cytochrome P450, family 2, subfamily c, polypeptide 40 Cell cell signaling Gap junction membrane channel protein beta 1 Cytokines Interleukin 24 Cell membrane protein GLI pathogenesis-related 1 (glioma) Function unknown Angiomotin like 2 Fasciculation and elongation protein zeta 2 (zygin II)

Values of up-regulation (↑) and down-regulation (↓) are expressed as the log2 ratio compared with a pool-clone of control transfectants (+ genes verified by real-time RT-PCR).

Our studies have revealed a large number of Prox1-regulated genes in ED14-MMH, the majority being down-regulated (22 genes up-regulated and 232 genes down-regulated more than 12-fold). The complete data set is published at NCBI/GEO (accession numbers: GSM190798, GSM190799, GSM190800, GSM190801, GSM190802, GSM190803; series number: GSE7867). Notably, we have observed the down-regulation of

HNF4α, which has been confirmed by qRT-PCR. Additionally, a direct interaction of Prox1 and HNF4α proteins has been demonstrated recently (Song et al. 2006); this interaction takes place via the N-terminal LXXLL motif of Prox1 and the activation function 2 domain of HNF4α. Prox1 inhibits HNF4α-induced transactivation of both CYP7A1 and phosphoenolpyruvate carboxykinase genes, which are involved in bile acid synthesis and gluconeogenesis, respectively. Prox1 clearly down-regulates HNF4α functions at both gene and protein levels. Prox1 and hepatocyte commitment

Fig. 7 Zymographic analysis of supernatants from ED14-MMH after 24h of culture. Lane 1 Wild-type (wt) cells, lane 2 control transfectants (vec), lane 3 Prox1-transfected cells (Prox1). Note upregulation of active MMP-2 in lane 3

Our data also suggest that Prox1 is involved in hepatocyte commitment by down-regulating five transcription factors that are highly expressed in neighboring endodermal organs: Elf3, Ehf, Rara, Zfp99, and Tcea3. The microarray data have been confirmed by qRT-PCR in three (Elf3, Ehf,

218

Cell Tissue Res (2007) 330:209–220

Fig. 8 a, b Prox1 (red) and MMP-2 (green) immunofluorescence studies in ED11.5 mouse embryos. a Overview showing the liver (L) and transverse septum (asterisks). MMP2 expression occurs in both liver and transverse septum. b Higher magnification of a. Note the highest expression of MMP-2 lies in peripherally located hepatoblasts. Bars150 μm (a), 50 μm (b). c, d Matrigel invasion assay performed with ED14-MMH over a period of 24 h. No difference is seen between vector-transfected MMH (MMH-vec) and wildtype MMH (data not shown), but the Prox1-transfected MMH (d) migrate much less efficiently through Matrigel (MMH-Prox1). Representative pictures of triplicate experiments. Bar30 μm

Tcea3) out of the five cases, but not for Rara and Zfp99. Specification of the endodermal cell fate toward the hepatocyte lineage might take place by these means. Details of the transcription factors are as follows: (1) Elf3 is highly expressed in gallbladder and has two putative DNA binding domains, viz., an ETS domain and an A/T hook domain (Oettgen et al. 1997). (2) Ehf is highly expressed in gallbladder, lung, and intestine; it is an ets-homologous factor and transcriptional activator and is associated with both development and carcinogenesis (Kleinbaum et al. 1999). (3) Retinoic acid receptor α (Rara) is highly expressed in stomach; it is a nuclear receptor closely related to the thyroid hormone receptor α (Brand et al. 1988), also being involved in leukocyte development and acute promyelocytic leukaemia (Pandolfi 2001). (4) Zinc finger protein 99 (Zfp99) is highly expressed in intestine and pancreas. (5) Tcea3 is highly expressed in stomach and pancreas and acts as a transcription elongation factor (Labhart and Morgan 1998). The data suggest that the constitutive expression of Prox1 determines the hepatocyte lineage. Regulatory molecules, which are highly expressed in neighboring endodermal organs, are obviously down-regulated by Prox1 in hepatoblasts. Thereby, Prox1 may separate hepatoblasts from cholangiocytes, enterocytes, and other endodermal derivatives.

Prox1 and ECM turnover Our data provide evidence that Prox1 regulates early liver morphogenesis by inducing MMP-2 expression and secretion. Most MMPs are secreted as inactive proproteins, which are activated when cleaved by extra-cellular proteinases. MMP-2 is a collagenase (gelatinase), which primarily degrades type-IV collagen, the major structural component of basement membranes (Nagase et al. 1992). Degradation of a basal lamina is a characteristic feature of pathologic processes such as tumor metastasis (Rudek et al. 2002) and of morphogenetic mechanisms such as angiogenesis in which endothelial cells sprout into adjacent tissue (Brooks et al. 1998). Similarly, the early liver bud sprouts into the septum transversum mesenchyme. As a derivative of the endodermal epithelium, it is surrounded by a basal lamina. In Prox1-deficient mice, this basal lamina does not become dissolved, and the liver bud fails to migrate into the adjacent mesoderm (Sosa-Pineda et al. 2000). We have observed that MMP-2 expression in mouse embryos is highest in those hepatoblasts that are located at the periphery of the growing liver bud and in close contact with the mesoderm. Massive up-regulation of MMP-2 occurs in ED14-MMH after Prox1 transfection, and this MMP-2 is active, as has been shown by zymography. In addition, the high expression of MMP2 in the septum transversum mesoderm may facilitate invasive growth of the liver bud. However, the opposite up-regulation of the α3-chain of type-IV procollagen is also induced by Prox1.

Cell Tissue Res (2007) 330:209–220

The α3-chain is part of the basal lamina, in addition to the classical α1- and α2-chains (Saus et al. 1988). Although a continuous basal lamina is not characteristic of liver lobules, even in differentiated liver, type-IV collagen is part of the normal liver matrix, and Prox1 may have a permanent function in its turnover. When hepatocytes are grown in wells pre-coated with type-IV collagen, rapid spreading and prolonged survival can be observed, in contrast to hepatocytes seeded on type-I collagen (Pinkse et al. 2004). The differentiation of hepatocytes in vitro is retained for a prolonged period when the cells are cultured on liver-specific basal lamina (Zeisberg et al. 2006). Prox1 seems to have a central role in both the production and the degradation of type-IV collagen in the liver, and the balance may shift during development, differentiation, or the regression of fibrosis. Our in vitro invasion assay shows that high MMP2 does not necessarily correlate with high invasiveness, and invasive growth of the embryonic liver bud is most probably controlled by interactions with the MMP2-positive mesenchyme of the transverse septum. Not only cell-matrix interactions, but also cell-cell interactions are regulated by Prox1. Cadherin 1 (E-cadherin), a calcium ion-dependent epithelial adhesion molecule, remains highly expressed in the liver bud of Prox1 deficient mice (SosaPineda et al. 2000) and is significantly more highly expressed in Prox1-negative ED14-MMH (2.8-fold). Since these cells migrate through Matrigel more efficiently, MMH appear to migrate as cell cords rather than single cells. Acknowledgements We thank Mrs. M. Böning and Mr. M. Winkler for their excellent technical assistance.

References Aiuti A, Cicchini, C, Bernardini S, Fedele G, Amicone L, Fantoni A, Tripodi M (1998) Hematopoietic support and cytokine expression of murine-stable hepatocyte cell lines (MMH). Hepatology 28:1645–1654 Amicone L, Spagnoli FM, Spath G, Giordano S, Tommasini C, Bernardini S, De Luca V, Della Rocca C, Weiss MC, Comoglio PM, Tripodi M (1997) Transgenic expression in the liver of truncated Met blocks apoptosis and permits immortalization of hepatocytes. EMBO J 16:495–503 Brand N, Petkovich M, Krust A, Chambon P, The H de, Marchio A, Tiollais P, Dejean A (1988) Identification of a second human retinoic acid receptor. Nature 332:850–853 Brooks PC, Silletti S, Schalscha TL von, Friedlander M, Cheresh DA (1998) Disruption of angiogenesis by PEX, a noncatalytic metalloproteinase fragment with integrin binding activity. Cell 92:391–400 Burke Z, Oliver G (2002) Prox1 is an early specific marker for the developing liver and pancreas in the mammalian foregut endoderm. Mech Dev 118:147–155 Cassiman D, Barlow A, Vander Borght S, Libbrecht L, Pachnis V (2006) Hepatic stellate cells do not derive from the neural crest. J Hepatol 44:1098–1104

219 Darlington GJ, Bernhard HP, Miller RA, Ruddle FH (1980) Expression of liver phenotypes in cultured mouse hepatoma cells. J Natl Cancer Inst 64:809–819 Dudas J, Papoutsi M, Hecht M, Elmaouhoub A, Saile B, Christ B, Tomarev SI, Kaisenberg CS von, Schweigerer L, Ramadori G, Wilting J (2004) The homeobox transcription factor Prox1 is highly conserved in embryonic hepatoblasts and in adult and transformed hepatocytes, but is absent from bile duct epithelium. Anat Embryol (Berl) 208:359–366 Dudas J, Elmaouhoub A, Mansuroglu T, Batusic D, Tron K, Saile B, Papoutsi M, Pieler T, Wilting J, Ramadori G (2006) Prosperorelated homeobox 1 (Prox1) is a stable hepatocyte marker during liver development, injury and regeneration, and is absent from “oval cells”. Histochem Cell Biol 12:6549–6562 Glasgow E, Tomarev SI (1998) Restricted expression of the homeobox gene prox 1 in developing zebrafish. Mech Dev 76:175–178 Grapin-Botton A, Melton DA (2000) Endoderm development: from patterning to organogenesis. Trends Genet 16:124–130 Grompe M, Finegold MJ (2001) Liver stem cells, stem cell biology. Cold Spring Habor Laboratory Press, Cold Spring Habor, NY Gualdi R, Bossard P, Zheng M, Hamada Y, Coleman JR, Zaret KS (1996) Hepatic specification of the gut endoderm in vitro: cell signaling and transcriptional control. Genes Dev 10:1670–1682 Hartmann H, Schmitz F, Christ B, Jungermann K, Creutzfeldt W (1990) Metabolic actions of insulin-like growth factor-I in cultured hepatocytes from adult rats. Hepatology 12:1139–1143 Henry GL, Melton DA (1998) Mixer, a homeobox gene required for endoderm development. Science 281:91–96 Hentsch B, Lyons I, Li R, Hartley L, Lints TJ, Adams JM, Harvey RP (1996) Hlx homeo box gene is essential for an inductive tissue interaction that drives expansion of embryonic liver and gut. Genes Dev 10:70–79 Hudson C, Clements D, Friday RV, Stott D, Woodland HR (1997) Xsox17alpha and -beta mediate endoderm formation in Xenopus. Cell 91:397–405 Jung J, Zheng M, Goldfarb M, Zaret KS (1999) Initiation of mammalian liver development from endoderm by fibroblast growth factors. Science 284:1998–2003 Kleinbaum LA, Duggan C, Ferreira E, Coffey GP, Buttice G, Burton FH (1999) Human chromosomal localization, tissue/tumor expression, and regulatory function of the ets family gene EHF. Biochem Biophys Res Commun 264:119–126 Knittel T, Kobold D, Saile B, Grundmann A, Neubauer K, Piscaglia F, Ramadori G (1999) Rat liver myofibroblasts and hepatic stellate cells: different cell populations of the fibroblast lineage with fibrogenic potential. Gastroenterology 117:1205–1221 Labhart P, Morgan GT (1998) Identification of novel genes encoding transcription elongation factor TFIIS (TCEA) in vertebrates: conservation of three distinct TFIIS isoforms in frog, mouse, and human. Genomics 52:278–288 Le Douarin N (1970) Methods of culture and grafts in the experimental study of development of the liver in chick embryos. Ann Biol 9:335–348 Martinez Barbera JP, Clements M, Thomas P, Rodriguez T, Meloy D, Kioussis D, Beddington RS (2000) The homeobox gene Hex is required in definitive endodermal tissues for normal forebrain, liver and thyroid formation. Development 127:2433–2445 Molkentin JD, Lin Q, Duncan SA, Olson EN (1997) Requirement of the transcription factor GATA4 for heart tube formation and ventral morphogenesis. Genes Dev 11:1061–1072 Nagase H, Barrett AJ, Woessner JF Jr (1992) Nomenclature and glossary of the matrix metalloproteinases. Matrix Suppl 1:421–424 Ny A, Koch M, Schneider M, Neven E, Tong RT, Maity S, Fischer C, Plaisance S, Lambrechts D, Heligon C, Terclavers S, Ciesiolka M, Kalin R, Man WY, Senn I, Wyns S, Lupu F, Brandli A, Vleminckx K,

220 Collen D, Dewerchin M, Conway EM, Moons L, Jain RK, Carmeliet P (2005) A genetic Xenopus laevis tadpole model to study lymphangiogenesis. Nat Med 11:998–1004 Oettgen P, Alani RM, Barcinski MA, Brown L, Akbarali Y, Boltax J, Kunsch C, Munger K, Libermann TA (1997) Isolation and characterization of a novel epithelium-specific transcription factor, ESE-1, a member of the ets family. Mol Cell Biol 17: 4419–4433 Oliver G, Sosa-Pineda B, Geisendorf S, Spana EP, Doe CQ, Gruss P (1993) Prox 1, a prospero-related homeobox gene expressed during mouse development. Mech Dev 44:3–16 Pandolfi PP (2001) Oncogenes and tumor suppressors in the molecular pathogenesis of acute promyelocytic leukemia. Hum Mol Genet 10:769–775 Papoutsi M, Siemeister G, Weindel K, Tomarev SI, Kurz H, Schachtele C, Martiny-Baron G, Christ B, Marme D, Wilting J (2000) Active interaction of human A375 melanoma cells with the lymphatics in vivo. Histochem Cell Biol 114:373–385 Pinkse GG, Voorhoeve MP, Noteborn M, Terpstra OT, Bruijn JA, De Heer E (2004) Hepatocyte survival depends on beta1-integrinmediated attachment of hepatocytes to hepatic extracellular matrix. Liver Int 24:218–226 Qin J, Gao DM, Jiang QF, Zhou Q, Kong YY, Wang Y, Xie YH (2004) Prospero-related homeobox (Prox1) is a corepressor of human liver receptor homolog-1 and suppresses the transcription of the cholesterol 7-alpha-hydroxylase gene. Mol Endocrinol 18:2424–2439 Rodriguez-Niedenfuhr M, Papoutsi M, Christ B, Nicolaides KH, Kaisenberg CS von, Tomarev SI, Wilting J (2001) Prox1 is a marker of ectodermal placodes, endodermal compartments, lymphatic endothelium and lymphangioblasts. Anat Embryol (Berl) 204:399–406 Rudek MA, Venitz J, Figg WD (2002) Matrix metalloproteinase inhibitors: do they have a place in anticancer therapy? Pharmacotherapy 22:705–720

Cell Tissue Res (2007) 330:209–220 Saus J, Wieslander J, Langeveld JP, Quinones S, Hudson BG (1988) Identification of the Goodpasture antigen as the alpha 3(IV) chain of collagen IV. J Biol Chem 263:13374–13380 Schmidt C, Bladt F, Goedecke S, Brinkmann V, Zschiesche W, Sharpe M, Gherardi E, Birchmeier C (1995) Scatter factor/hepatocyte growth factor is essential for liver development. Nature 373: 699–702 Schneider M, Buchler P, Giese N, Giese T, Wilting J, Buchler MW, Friess H (2006) Role of lymphangiogenesis and lymphangiogenic factors during pancreatic cancer progression and lymphatic spread. Int J Oncol 28:883–890 Song KH, Li T, Chiang JY (2006) A prospero-related homeodomain protein is a novel co-regulator of hepatocyte nuclear factor 4alpha that regulates the cholesterol 7alpha-hydroxylase gene. J Biol Chem 281:10081–10088 Sosa-Pineda B, Wigle JT, Oliver G (2000) Hepatocyte migration during liver development requires Prox1. Nat Genet 25:254–255 Spagnoli FM, Amicone L, Tripodi M, Weiss MC (1998) Identification of a bipotential precursor cell in hepatic cell lines derived from transgenic mice expressing cyto-Met in the liver. J Cell Biol 143:1101–1112 Steffensen KR, Holter E, Bavner A, Nilsson M, Pelto-Huikko M, Tomarev S, Treuter E (2004) Functional conservation of interactions between a homeodomain cofactor and a mammalian FTZ-F1 homologue. EMBO Rep 5:613–619 Tomarev SI, Zinovieva RD, Chang B, Hawes NL (1998) Characterization of the mouse Prox1 gene. Biochem Biophys Res Commun 248:684–689 Zeisberg M, Kramer K, Sindhi N, Sarkar P, Upton M, Kalluri R (2006) De-differentiation of primary human hepatocytes depends on the composition of specialized liver basement membrane. Mol Cell Biochem 283:181–189 Zinovieva RD, Duncan MK, Johnson TR, Torres R, Polymeropoulos MH, Tomarev SI (1996) Structure and chromosomal localization of the human homeobox gene Prox 1. Genomics 35:517–522