Granulocyte-colony stimulating factor promotes invasion by human lung cancer cell lines in vitro Xin-Hai Pei, Yoichi Nakanishi, Koichi Takayama, Jun Yatsunami, Feng Bai, Masayuki Kawasaki, Kentaro Wakamatsu, Nobuko Tsuruta, Keiko Mizuno and Nobuyuki Hara Research Institute for Diseases of the Chest, Faculty of Medicine, Kyushu University, Fukuoka, Japan (Received 28 August 1995; accepted in revised form 12 December 1995)
The effects of exogenous and endogenous granulocyte colony-stimulating factor (G-CSF) on invasion by cancer cells were studied, using lung cancer cell lines that produce G-CSF (NCI-HI57) and lines that do not (PC-9 and NCI-H23). The invasive capacity of NCI-HI57 cells was 26- to 27-fold higher than that of PC-9 and NCI-H23 cells. The invasiveness of PC-9 cells was stimulated by exogenous G-CSF, while that of NCI-HI57 cells was not. Antibodies against G-CSF blocked the stimulation of PC-9 cell invasiveness by exogenous G-CSF. Anti G-CSF antibodies also inhibited invasion by NCI-H157 cells in the absence of exogenous G-CSF. These results suggest that endogenous and exogenous G-CSF both stimulate invasion by lung cancer cells. Keywords: granulocyte-colony stimulating factor, invasion, invasion assay, lung cancer
Introduction The ability of tumor cells to invade interstitial tissue is one of several steps necessary for the establishment of metastases . This process is regulated by interactions between tumor and host cells, which can regulate tumor growth and invasion ability directly or through various humoral factors [2-5]. It is known that such agents including hematopoietic growth factors, cytokines, enzymes and oxygen radicals mediate or regulate tumor growth and invasion capacity [6-9]. Granulocyte colony-stimulating factor (G-CSF), a human hematopoietic growth factor, is secreted by hematopoietic cells and some non-hematopoietic cells . It has classically been considered to regulate hematopoiesis. G-CSF can stimulate in vitro A d d r e s s correspondence to: Yoichi Nakanishi, Research Institute for Diseases of the Chest, Faculty of Medicine, K y u s h u University, 3-1-1 Maidashi, Higashiku, F u k u o k a 812-82, Japan. Tel: (+81) 92 641 1151 (ext. 2323); Fax: (+81) 92 633 4257.
colony formation not only of leukemia progenitor blasts  but also of non-hematopoietic cells, such as small cell lung cancer cells , endothelial cells , human colon adenocarcinoma cells, bladder cancer cells and human placenta trophoblastic cells [14,15]. However, no st~udies have been performed to determine whether G-CSF also increases the metastatic capacity of tumor cells. The present study was aimed at determining the effect of G-CSF on tumor cell invasiveness, using lung cancer cell lines which do or do not produce G-CSF.
Materials and methods Cells and chemicals PC-9 and NCI-H23 cells, derived from human pulmonary adenocarcinomas, and NCI-H157 cells, derived from a squamous cell carcinoma of the lung, were maintained in RPMI-1640 medium containing Clinical & Experimental Metastasis Vol 14 No 4 351
X.-H. Pei et al.
5% fetal bovine serum (FBS; CC Laboratories, Cleveland, OH, USA). Medium was changed every 3 days and cultures were routinely subcultured at a split of 1:2 once a week. Prior to and during experiments, cell viability was determined by trypan blue exclusion. Only cell suspensions exhibiting >95% viability were used. Nartograstim, human recombinant granulocyte colony-stimulating factor, was obtained from Kyowa H a k k o Kogyo Co. Ltd (Tokyo, Japan). Rabbit anti-human G-CSF polyclonal antibody was purchased from Genzyme Corp (Cambridge, MA, USA). Horseradish peroxidase ( H R P ) conjugated anti-G-CSF was a gift from Chugai Pharmaceutical Company (Tokyo, Japan). Cell culture inserts and 10% NuSerum were from Becton Dickinson (Bedford, MA, USA). Biocoat Matrigel invasion chambers (Becton Dickinson) consist of Falcon cell culture inserts containing 8 txm pore size polyethylene terephthalate (PET) membranes coated with Matrigel basement membrane matrix. Diff-Quick staining solution was from International Reagents Corp (Kobe, Japan). 3-(4,5Dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) was brought from Sigma Chemical Co (St. Louis, MO, USA). All reagents were reconstituted as indicated by the manufacturers. E n z y m e immunoassay
The concentration of intracellular G-CSF in tumor cells was measured by enzyme immunoassay (EIA) as described . In brief, cells were suspended with PBS at a concentration of 2 × 106/ml, and were then mechanically disrupted with a homogenizer. Lysate was centrifuged and 200 ml aliquots of supernatant or standards were added to polystyrene tubes precoated with rabbit anti-G-CSF IgG followed by addition of 500 txl of E I A buffer containing 2% polyethylene glycol. After 2 h incubation at room temperature, 100 txl of HRP-conjugated anti-G-CSF was added and further incubated for 2 h at room temperature. After washing three times with 20 mm T r i s - H C L (pH 8.0), 1 ml of reaction mixture (3 mg/ml O-phenylenediamino dihydrochloride 2.3% disodium hydrogen phosphate, 0.38% citric acid, 0.1% salicyclic acid, 0.015% H202) was added for the color reaction. After 1 h incubation at room temperature in the dark, the reaction was stopped by adding 1 ml of 4 N sulfuric acid and the resulting optical density was measured at 492 nm using a spectrophotometer. Invasion assay and chemoinvasion assay In vitro invasiveness as assayed by the methods of Filderman et al.  and Hendrix et al.  with modi352
Clinical & Experimental Metastasis Vo114 No 4
fications. Invasiveness was measured by use of the Biocoat invasion chamber. Prior to seeding, cells were starved in serum-free medium for 24 h and subsequently incubated with RPMI-1640 plus 10% NuSerum in the presence or absence of different concentrations of G-CSF or anti-G-CSF polyclonal antibody for another 24h. NuSerum was used instead of FBS for the invasion assay because it contains standardized amounts of defined growth factors (but no G-CSF) and lower concentrations of protease inhibitors. Cells were removed from the cell culture flask and suspended with NuSerum except when the effect of anti-G-CSF antibody on cancer cell invasiveness was evaluated in the absence of exogenous G-CSF. In this instance, cells were continuously exposed to antibody over the course of the experiment. Two ml of cell suspension (1 x 102 cells/ml) were placed in the upper well. Both the upper and lower wells were filled with RPMI-1640 supplemented with 10% NuSerum and the plates were incubated for 24 h. The cells that invaded the Matrigel-coated filters and floated in the medium of the lower well were collected. These samples were stored in culture tubes on ice and the lower well was filled with 2 mm E D T A . After 10 rain incubation at room temperature, cells passing through the filter and adhering to its bottom surface were collected. After the cells were trapped with the cell culture insert they were stained with Diff-Quick and counted. Invasiveness was evaluated as percentage invasion, which was calculated by the following formula : total no. of invading cells (lower well sample) × 100. total no. of seeded cells (upper well sample) The chemoinvasion assay differed from the invasion assay in the following respects. Cells were prepared by starving for only 24 h without pretreatment with growth factors or antibodies before seeding into Biocoat Matrigel invasion chambers. Growth factors, with or without antibodies, were added to the lower well as chemoattractants . M T T assay
M T T proliferation assays were carried out as described elsewhere [19, 20]. Briefly, cells were harvested and washed three times with RPMI-1640. Aliquots (100 ~1) of cell suspension were placed in 96-well tissue culture plates. G-CSF (110 jxl) diluted to different concentrations in a serum-free medium was added to each well. After 4 days in incubation, 0.1 mg (50 ixl of 2 mg/ml) M T T was added and the plates were incubated at 37°C for 4 h. The plates were then centrifuged at 800 g for 10 min and the
G-CSF promotes invasion by human lung cancer Table 1.
Production of G-CSF and tumor cell invasion PC-9
Endogenous < 30 < 30 G-CSF (pg/ml) a Invasiveness 0.054 + 0.003 0.055 + 0.004 (% invasion) b
Statistical analysis T h e d a t a w e r e a n a l y s e d for significance using S t u d e n t ' s t-test. A l l P v a l u e s cited w e r e t w o - s i d e d , a n d t h o s e less t h a n 0.05 w e r e j u d g e d to b e statistically significant.
NCI-H157 8940 1.45 _+0.04
Results aCells were suspended with PBS at a concentration of 2 x 106/ml and disrupted with a homogenizer. G-CSF concentrations in lysate supernatants were measured by EIA. bInvasiveness was expressed as a percentage of the n u m b e r of seeded cells.
G-CSF effect of invasion by PC-9 cells
Invasion assay a (% invasion)
Chemoinvasion assay (% invasion)
0 50 250 1000
0.054 0.059 0.104 0.125
0.054 0.110 0.140 0.244
_+0.003 + 0.003 +_0.014" +_0.021"
+ 0.003 _+0.017" _+0.018" + 0.008*
Neither PC-9 nor NCI-H23 cells p r o d u c e d d e t e c t a b l e levels o f G - C S F , while N C I - H 1 5 7 cells p r o d u c e d s u b s t a n t i a l a m o u n t s o f G - C S F ( T a b l e 1). I n v a s i o n b y t h e t w o a d e n o c a r c i n o m a cell lines (PC9 a n d N C I - H 2 3 ) was n e a r l y t h e same. M o r e o v e r , t h e i n v a s i v e n e s s o f N C I - H 1 5 7 cells was significantly g r e a t e r t h a n t h a t o f P C - 9 a n d N C I - H 2 3 cells (26- to 2V-fold).
Exogenous G-CSF effect on tumor cell invasion
alnvasiveness was expressed as a percentage of the n u m b e r of seeded cells. *P < 0.05 compared to control.
m e d i u m r e m o v e d . M T T f o r m a z a n crystals w e r e t h e n s o l u b i l i z e d b y a d d i n g 200 p.1 o f d i m e t h y l s u l f o x i d e a n d a b s o r b a n c e was m e a s u r e d using an a u t o m a t e d m i c r o p l a t e r e a d e r at a w a v e l e n g t h o f 540 n m ( E a s y R e a d e r E A R 340, S L T - L a b i n s t r u m e n t s , Austria).
Figure 1. Effect of exogenous G-CSF on invasion by tumor cells. Tumor cells were starved for 24 h, incubated with G-CSF for 24 h, then seeded into a Matrigel-precoated invasion chamber. After 24 h, invading cells were counted. The number of invading cells is expressed as the percentage of cells pretreated with the medium only (control). Values represent the mean + SE (n = 9 for PC-9, n = 6 for NCI-H23 and NCI-H157. *P = 0.0001 compared to control,
Production of G-CSF and invasiveness of tumor cells
T a b l e 2 shows t h e effect o f v a r y i n g d o s e s o f G - C S F o n P C - 9 cell invasiveness. I n t h e s t a n d a r d i n v a s i o n assay, e x o g e n o u s G - C S F p r o m o t e d i n v a s i o n in a d o s e - d e p e n d e n t m a n n e r . T h e m a x i m u m effect of G C S F (2.5-fold s t i m u l a t i o n ) was o b t a i n e d at a c o n c e n t r a t i o n o f 1000 ng/ml. A t a c o n c e n t r a t i o n of 250 ng/ml, a significant effect o n i n v a s i o n was still o b t a i n e d ( P <0.0001). Thus, w e s u b s e q u e n t l y u s e d this c o n c e n t r a t i o n as a s t a n d a r d G - C S F dose. In the c h e m o i n v a s i o n assay, G - C S F also s h o w e d significant s t i m u l a t i o n o f P C - 9 invasion. In c o n t r a s t , invasiveness o f t h e N C I - H 1 5 7 cell line, w h i c h p r o d u c e s high level o f G - C S F , was n o t i n c r e a s e d b y e x o g e n o u s G - C S F ( F i g u r e 1).
O n7~ oO 200
z O m X
50 250 1000 G-CSF (ng/m!) PC-9
50 250 1000 G-CSF (ng/ml) NCI-H157
Clinical & Experimental Metastasis Vol 14 No 4 353
X.-H. Pei et al. Figure 2. Effect of anti-G polyclonal antibody on invasion by exogenous G-CSF-treated tumor cells. Tumor cells were starved for 24h, incubated with G-CSF (250 ng/ml) in the absence or presence of anti-G-CSF antibody for 24 h, and then seeded into an invasion chamber. After 24 h, invading cells were counted. The number of invading cells is expressed as the percentage of cells pretreated with the medium only (control). Values represent the mean _+ SE ( n = 6 ) . P < 0 . 0 5 compared to the cells pretreated with G-CSF only.
O ~ O o z 0 ~ X _z
0 AB (gG/ML) G-CSF
0.25 1.25 2.5 +
Figure 3. Effect of anti-G-CSF polyclonal antibody on invasion by tumor cells. Cells were starved and incubated with anti-G-CSF antibody for 2 4 h and then seeded into an invasion chamber in the presence of the antibody. After 24 h, invading cells were counted. The number of invading cells is expressed as the percentage of cells pretreated with medium only (control). Values represent the mean _+SE (n =6). *P < 0.05 compared to the control.
[ - - I
zw 8 0 O o 60 z O ~ X _z
0.25 1.25 2.5
0: 0 1 5 ANTI-G-CSF(gg/m) PC-9
Figure 4. Effect of G-CSF on the growth of tumor cells. Cell growth was measured in the absence or presence of G-CSF as described in Materials and methods. The dose-response curve was plotted as a percentage of control (cells not exposed to G-CSF). Similar results were obtained in three experiments. @ = PC-9, • = N C I H157.
0 1 5 ANTI-G-CSF(gg/m) NCI-H 157
300 O nFZ 0 0
-11..0rr (.9 10
354 Clinical& Experimental Metastasis Vo114 No 4
G-CSF promotes invasion by human lung cancer Antibody effect on G-CSF-mediated tumor cell invasion As shown in Figure 2, the effect of e x o g e n o u s G-CSF on the invasiveness of PC-9 cells was blocked in a dose-dependent m a n n e r by a polyclonal antibody (anti-G-CSF). At a concentration of 2.5 txg/ml the anti-G-CSF antibody completely blocked the effect of G-CSF on PC-9 cell invasiveness in the invasion assay system. On the other hand, antiG-CSF antibody did not inhibit the invasiveness of NCI-H157 cells when 250ng/ml of exogenous G-CSF was added. However, when NCI-H157 cells were continuously exposed to anti-G-CSF polyclonal antibody in the absence of exogenous G-CSF, their invasive capacity was reduced in a dose-dependent manner. The invasiveness of PC-9 cells pre-incubated and cultured with anti-G-CSF antibody was not changed (Figure 3). G-CSF effect on tumor cell growth The M T T assay was used to determine if G-CSFinduced increases in tumor cell invasiveness were due to growth stimulation by G-CSF (Figure 4). GCSF had no significant effect on tumor cell growth, even at doses much higher than the optimum concentration.
Discussion T u m o r cell invasion through the basement m e m b r a n e and stroma consists of three steps: tumor cell attachment to the extracellular matrix, proteolytic dissolution of the matrix, and movement of the cells through the digested barrier . This process can occur repeatedly during the courses of intra-extravasation and can result in metastases at sites distant from the original tumor. However, the factors controlling these events are not well defined. G-CSF is one of a growing n u m b e r of cytokines recognized as being involved in the regulation of hematopoiesis. It is currently being used to alleviate chemotherapy-induced bone marrow toxicity . However, relatively few studies have been conducted to ascertain its effect on tumor cells, even though non-hematopoietic tumor cells have been shown to express functional receptors for G-CSF . It has been reported that G-CSF enhances growth of tumor cells in addition to its effect on the proliferation and differentiation of myeloid progenitors. Up to now, no study has examined whether it is also involved in human carcinoma cell invasion. Recently, Young reported that GM-CSF stimulates the metastatic properties of Lewis lung carcinoma cells [22, 23]. It also has been reported that CSF-1,
another colony-stimulating factor, stimulates the invasiveness of two lung carcinoma cell lines . In the present study, we found the cellular ]level of G-CSF correlated with the invasive potential of the cells studies. The highly invasive NCI-H157 cell line produced high cellular levels of G-CSF, while the less invasive cell lines, PC-9 and NCI-H23, did not produce detectable levels of G-CSF. To confirm the relation between endogenous G-CSF levels and invasion potential, we carried out experiments using neutralizing antibodies to G-CSF. When cells were pre-incubated and cultured continuously with antiG-CSF antibodies in the absence of exogenous GCSF, the invasiveness of NCI-H157 cells was significantly reduced, while that of PC-9 cells was not. However, the presence of exogenous G-CSF did not cause any significant difference in the invasiveness of NCI-H157 cells which produce endogenous G-CSF. In contrast, recombinant G-CSF stimulated in vitro invasion by PC-9 and NCI-N23 cells, which do not produce endogenous G-CSF. Additional experiments using antibody neutralization confirmed the specificity of invasion stimulation by exogenous G-CSF. T o confirm the effect of G-CSF and anti-G-CSF antibodies on the invasiveness and motility of lung cancer cells, we used both a standard invasion assay and a chemoinvasion assay. Similar results were observed in both assays. When the effects of two or more factors are to be studied, cells can be pretreated with factors individually or sequentially in the invasion assay, avoiding the possibility of interactions between factors in the chemoinvasion assay . Therefore, invasion assay would be preferable to chemoinvasion assay for the analysis of multiple factors on the invasive potential of tumor cells. Tumors secrete a variety of soluble products that can directly or indirectly stimulate the capacity of tumor cells to metastasize. GM-CSF and CSF-1, which have immunomodulatory activities that can facilitate establishment of metastasis are produced by hematopoietic and non-hematopoietic cells [7, 22]. G-CSF has also been reported to be produced by monocytes/macrophages, fibroblasts, endothelial cells and some tumor cells, such as bladder cancer cells, human squamous carcinoma cells and small cell lung cancer cells . It has been reported that some endogenous growth factors and peptides, such as urinary trypsin inhibitor, stem cell factor and interleukin 8 produced by tumor or normal cells, act via binding with their receptors on the host cells [24-26]. These receptors can be completely saturated by endogenous growth factors. In this study, invasion by NCI-H157 cells Clinical & Experimental Metastasis Voi 14 No 4
X . - H . Pei et al.
was inhibited in the presence of anti-G-CSF antibodies, which block the binding of endogenous G-CSF to its receptor. In addition, anti-G-CSF antibody did not inhibit the invasiveness of NCI-H157 cells in the presence of exogenous G-CSF. The lack of a significant response by NCI-H157 cells to exogenous G-CSF may have been due to near-complete saturation of receptors on the surface of these cells by endogenous G-CSF, giving these cells a high invasive potential. Our data demonstrate a dose-response effect of exogenous G-CSF on the invasiveness of PC-9 cells. After exposure to G-CSF, invasion by PC-9 cells was maintained, even after removal of G-CSF by washing. Therefore, G-CSF was most likely bound to cellular receptors after initial exposure although we did not analyse G-CSF receptors on the cell lines in this study. A M T T assay was used to determine whether G-CSF-mediated tumor cell invasion could be attributed to its effects on tumor cell growth. No significant stimulatory effect was found. This suggested that G-CSF-mediated PC-9 cell invasion in this system was not due to an increase in PC-9 cell proliferation. In summary, our results suggest that G-CSF may function directly on lung cancer cells to stimulate their invasiveness regardless of whether G-CSF is endogenously secreted or provided exogenously.
References 1. LiottaRao CN and Barsky SH, 1983, Tumor invasion and the extracellular matrix, Lab Invest, 49, 636-49. 2. Wei W and Heppner GH, 1988, Tumor infiltrating lymphocytes of spontaneous versus transplanted mouse mammary tumors. Cancer Immunol Immunother, 26, 257-62. 3. Rice GE, Gimbrone MA and Bevilacqua MP, 1988, Tumor cell-endothelial interaction. A m J Pathol, 133, 204-11. 4. Breillout F, Antoine E, Lascaus V, Rolland Y and Poupon MF, 1989, Promotion of micrometastasis proliferation in a rat rhabdomyosarcoma model by epidermal growth factor. J Natl Cancer Inst, 81, 702-5. 5. Li X, Nagayasu H, Hamada J, Hosokawa M and Takeichi N, 1993, Enhancement of tumorigenicity and invasion capacity of rat mammary adenocarcinoma Cell by epidermal growth factor and transforming growth factor-J3. Jpn J Cancer Res, 84, 1145-9. 6. Kacinski BM, Scata KA, Carter D, et al. 1991, FMS (CSF-1 receptor) and CSF-1 transcripts and protein are expressed by human breast carcinoma in vivo and in vitro. Oncogene, 6, 941-52. 7. Filderman AB, Bruckner A, Kacinski BM, Deng N and Remold HG, 1992, Macrophage colonystimulating factor (CSF-1) enhances invasiveness in 356
Clinical & Experimental Metastasis Vo114 No 4
CSF-1 receptor-positive carcinoma cell lines. Cancer Res, 32, 3661-6.
8. Dabbous MK, North SM, Haney L and Nicolson GL, 1988, Macrophage and lymphocyte potentiation of syngeneic tumor cell and host fibroblast collagenolytic activity in rats. Cancer Res, 48,6832-6. 9. Pedersen PH, Ness GO, Engebraaten O, Bjerkvig R, Liliehaug JR and Laerum OD, 1994, Heterogenous response to the growth factors [EGF, PDGF (bb), TGF-a, bFGF, IL-2] on glioma spheroid growth, migration and invasion. In J Cancer, 56, 255-61. 10. Harmenberg J, Hoglund M and Hellstrom-Lindberg E, 1994, G- and GM-CSF in oncology and oncological haematology. Eur J Haematol, 52, 1-28. 11. Motoji T, Watanabe M, Uzumaki H, et al. 1991, Granulocyte colony-stimulating factor (G-CSF) receptors on acute myeloblastic leukaemia cells and their relationship with the proliferative response to G-CSF in clonogenic assay. Br J Haematol, 77, 54-9. 12. Avoloa BR, Gasson JC, Hedvat C, et al. 1990, Human granulocyte colony-stimulating factor biologic activities and receptor characterization on hematopoietic cells and small cell lung cancer cell lines. Blood, 75, 851-7. 13. Bussolino F, Wang JM, Defilippi P, et al. 1989, Granulocyte and granulocyte-macrophage-colony stimulating factors induce human endothelial cells to migrate and proliferate. Nature, 337, 471-3. 14. Berdel WE, Danhouser-Riedl S, Steinhauser G and Winton EF, 1989, Various human hematopoietic growth factors (interleukin 3, GM-CSF, G-CSF) stimulate clonal growth of nonhematopoietic tumor cells. Blood, 73, 80-3. 15. Uzumaki H, Okabe T, Sasaki N, et al. 1989, Identification and characterization of receptors for granulocyte colony-stimulating factor on human placenta and trophoblastic cells. Proc Natl Acad Sci USA, 86, 9323-6. 16. Motojima H, Kobayashi T, Shimane M, Kamachi S and Fukushima M, 1989, Quantitative enzyme immunoassay for human granulocyte colony stimulating factor (G-CSF). J Immunol Method, IlK 187-92. 17 Hendrix MJC, Seftor EA, Seftor REB and Fidler IJ, 1987, A simple quantitative assay for studying the invasive potential of high and low human metastatic variants. Cancer Lett, 38, 137-47. 18. Albini A, Iwamoto Y, Kleinman HK, et al. 1987, A rapid in vitro assay for quantitating the invasive potential of tumor cells. Cancer Res, 47, 3239-45. 19. Mosmann T, 1983, Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assay. J Immunol Methods, 65, 55-63. 20. Nakanishi Y, Mulshine JL, Kasprzyk PG, et al, 1988, Insulin-like growth factor-1 can mediate autocrine proliferation of human small cell lung cancer cell lines in vitro. J Clin Invest, 82, 354-9. 21. Rapoport AP, Abboud CN and Dipersio JF, 1992, Granulocyte macrophage colony-stimulating factor (GM-CSF) and granulocyte colony-stimulating factor (G-CSF): receptor biology, signal transduction and neutrophil activation. Blood, 6, 42-57. 22. Young MRI, Lozano Y, Coogan M, Wright MA, Young ME and Bagash JM, 1992, Stimulation of the metastatic properties of Lewis-lung-carcinoma cells by
G - C S F promotes invasion by human lung cancer autologous granulocyte-macrophage colony stimulating-factor. Int J Cancer, 50, 628-34. 23. Young MRI, Lozano Y, Djordjevic A, et al. 1993, Granulocyte-macrophage colony-stimulating factor stimulates the metastatic properties of Lewis lung carcinoma cells through a protein kinase. A signaltransduction pathway. Int J Cancer, 53667-71. 24. Kobayashi H, Gotoh J, Fujie M and Terao T, 1994, Characterization of the cellular binding site for the urinary trypsin inhibitor. J Biol Chem, 269, 20642-7.
25. Pietsch T, Kyas U, Steffens U, et al. 1992, Effect of human stem cell factor (c-kit ligand) on proliferation of mycloid leukemia cells: heterogeneity in response and synergy with other hematopoietic growth factors. Blood, 80, 1199-206. 26. Singh RK, Gutman M, Radinsky R, Bucana CD and Fidler IJ, 1994, Expression of interleukin 8 correlates with the metastatic potential of human melanoma cells in nude mouse. Cancer Res, 54, 3242-47.