Enhanced VEGF production and decreased immunogenicity induced by TGF-β 1 promote liver metastasis of pancreatic cancer

TGF-βs are multifunctional polypeptides that regulate cell growth and differentiation, extracellular matrix deposition, cellular adhesion properties, angiogenesis and immune functions. In this study, we investigated the effect of TGF-β1 on liver metastasis and its mechanism by using human pancreatic cancer cell lines Panc-1, Capan-2, and SW1990. Capan-2 and SW1990 cells demonstrated enhanced liver metastatic potential by in vivo splenic injection with TGF-β1. Consequently, we examined the role of TGF-β1 on in vitro angiogenesis and received cytotoxicity by peripheral blood mononuclear leukocytes (PBMLs). While TGF-β1 slightly decreased cell proliferation, it also upregulated VEGF production in all cancer cells examined. The binding of PBMLs to cancer cells and cancer cell cytotoxicity during co-culture with PBMLs were remarkably decreased by treatment with TGF-β1. Panc-1 cells revealed no liver metastasis despite their high immunogenetic and angiogenetic abilities, which was attributed to a lack of expression of the cell surface carbohydrates that induce attachment to endothelial cells. We concluded that the presence of TGF-β1 in the microenvironment of tumour site might play an important role in enhancing liver metastasis of pancreatic cancer by modulating the capacity of angiogenesis and immunogenicity. © 2001 Cancer Research Campaign http://www.bjcancer.com

TGF-β 1 (20 ng/ml) for 48 h. The cells were then suspended to a final concentration of 5 ×10 5 /0.1 ml in phosphate-buffered saline (PBS) and injected into the spleen of 6-week-old female Balb/c nude mice under ether anaesthesia. After injection, the spleen was extracted. The mice were sacrificed after 5 weeks to measure the number of metastatic tumour nodules and the liver weight.

Measurement of vascular endothelial growth factor (VEGF) levels in culture supernatants
The cells were cultured in DMEM with 10% FCS, washed with DMEM, and then cultured in DMEM alone or with 0, 1, 10, or 20 ng/ml TGF-β 1 for 48 h. The culture supernatants were collected and their VEGF levels were measured using commercially available ELISA kits (Amersham International PLC, Buckinghamshire, England).

Isolation and purification of peripheral blood mononuclear leukocytes (PBMLs) from healthy human donors
Human peripheral blood mononuclear leukocytes (PBMLs) were prepared from healthy volunteers by Ficoll-Hypaque density gradient centrifugation. Mono-Poly Resolving Medium was added to plastic non-siliconized tubes and then fresh anti-coagulated blood were added onto the medium. The tubes were centrifuged at 1500 rpm for 30 min, and the supernatant was drawn off with a pasteur pipette. After two washes with PBS, PBMLs were collected and resuspended in DMEM with 10% FCS. The resuspended lymphocytes were seeded onto a plastic plate and incubated for 1 h to remove attached monocytes. PBMLs were resuspended in medium appropriate to the work being done.

Lymphocytes adhesion assay
The binding of PBMLs to cancer cells was also investigated. Panc-1, Capan-2, and SW1990 cells (2.5 × 10 5 /well) were incubated in DMEM with either 10% fetal calf serum alone or 10% fetal calf serum and TGF-β 1(20 ng/ml) for 48 h in 96-well microtitre plates. After washing these plates with PBS, PBMLs (2.5 × 10 6 /well) suspended in serum-free medium were allowed to attach to cancer cells on each well for 1 h at 37˚C. The binding of PBMLs was quantified by measuring the concentration of 3-(4,5dimethylthiazol-2yl)-2,5-dipheny l-2,4-tetrazolium bromide (MTT; Sigma Chemical Co., St Louis, MO) by colorimetric assay (Carmichael et al, 1987), using a MTP-120 Microplate reader at 550 nm. The percentage of total PBMLs that adhered to cancer cells (% of adhesion) was calculated as % of adhesion = (OD of experimental wells -OD of cancer cell wells) / OD of total PBML wells × 100.

Cancer cell cytotoxicity assay
The cell cytotoxicity assay was performed using a Cytotox 96 Non Radioactive Cytotoxicity Assay Kit (Promega Co., Madison, WI), that measured lactate dehydrogenase(LDH), a stable cytosolic enzyme released upon cell lysis during culturing of mixed leukocytes (effector cells) and tumour cells (target cells) culture (MLTC). Effector cells (PBMLs) were isolated from healthy donors by Ficoll-Hypaque density gradient centrifugation as described above. After target cells were pretreated with or without 20 ng/ml TGF-β 1 for 48 h prior to assay, approximately 1× 10 4 target cells were placed in each well of a 96-well plate, and then 1× 10 5 effector cells were added and incubated at 37˚C in a humidified atmosphere of 5% CO 2 for 18 h. After the addition of Stop Solution (1 M acetic acid), absorbance related to released LDH level in the medium was measured with an MTP-120 Microplate reader (Corana Electric Co., Ibaragi, Japan) at 492 nm. Then the percentage of total cancer cells that underwent lysis (% cytotoxicity) was calculated as % cytotoxicity = (Experimental -Effector spontaneous -Target Spontaneous) / (Target maximum -Target Spontaneous) × 100.

Enzyme-linked immunosorbent assay (ELISA)
To determine the effect of TGF-β 1 on cell surface antigen expression, cancer cells were cultured in 96-well microtitre plates (Coster, Cambridge, MA) for 48 h in the presence or absence of TGF-β 1(20 ng/ml). After the cells had been washed three times with PBS, they were fixed with 0.25% glutaraldehyde at room temperature for 1 h. After blocking the wells with 100 mM glycine/1%BSA in PBS, anti-sialyl Le a and Le x first antibody were reacted to cells and ELISAs were performed with horseradish peroxidase-conjugated second antibodies (Zymed Laboratories Inc., San Francisco, CA) as previously described (Sawada et al, 1994;Ho et al, 1993).

Cell proliferation assay
To examine the effect of TGF-β 1 on cell proliferation, 1 × 10 4 cancer cells with or without TGF-β 1 were inoculated onto a 24well culture plate (Falcon, Lincoln Park, NJ). After 24, 48, and 72 h of incubation, cells were harvested and cell numbers were counted using a Coulter Counter (Coulter Electronics, Luton, UK)

Statistical analysis
Results were expressed as the mean ± SD. Student's t-test was used for statistical analysis, and P values less than 0.05 were considered to indicate statistical significance.

Effect of TGF-β1 on adhesion of lymphocytes to cancer cells
The attachment of PBMLs to cancer cells was inhibited following the addition of TGF-β 1 at a concentration of 20 ng/ml, compared with untreated cancer cells. The percentages of PBMLs binding to Panc-1, Capan-2 and SW1990 cells following the addition of TGFβ 1 were 24.3 ± 4.7%, 36.7 ± 10.8% and 11.2 ± 6.0%, while those of the controls were 34.5 ± 10.0, 59.5 ± 8.7% and 19.4 ± 6.6% ( Figure 3).

Effect of TGF-β1 on cell surface antigen expression
Expression of the cell surface antigen of sialyl Lewis-A (sLe a ) was higher in SW1990 than Capan-2 cells. In contrast, sialyl Lewis-X (sLe x ) was only detected in SW1990 cells. Panc-1 cells did not express either sLe a or sLe x antigens. TGF-β 1 had no influence on either sLe a or sLe x expression ( Figure 5).

Effect of TGF-β 1 on cell proliferation
At a concentration of 20 ng/ml, the addition of TGF-β 1 slightly reduced the proliferation of all cells, but there were no significant difference in proliferation among the cell lines, even at 72 h of incubation (data not shown).

DISCUSSION
In the present study, we found that TGF-β1 enhanced the liver metastasis of pancreatic cancer by modulating the capacity for immunogenicity and angiogenesis. Haematogenous metastasis involves a multistep process that begins with detachment of tumour cells from the primary lesion and ends with their attachment to a different organ and formation of new tumour nodules (Fidler, 1995). In these steps, tumour growth depends on angiogenesis to a large degree, which means that the tumours are dependent on the ingrowth of a vascular supply from the surrounding tissue in order to proliferate and metastasize (Folkman, 1995;McCulloch et al, 1995). Any individual tumour may have dominant angiogenic factors that induce angiogenesis by favouring an imbalance between positive and negative regulators. These angiogenic regulators act either directly on endothelial cells or indirectly by inducing the production of other regulators. Among these factors, VEGF is particularly important, since it clearly acts on endothelial cells in a direct manner (Mattern et al, 1997). Recently, it has been reported that enhanced VEGF expression correlates with haematogenous metastasis and prognosis in human colon, gastric, breast and pancreatic cancers (Kang et al, 1997;Maeda et al, 1996;Toi et al, 1994;Ikeda et al, 1999). These findings support our results that the production of VEGF in the culture supernatant of SW1990 was higher than that in the other two cell lines, and that only SW1990 originally revealed liver metastatic potential by in vivo splenic injection.
The production of many angiogenic modulators, such as VEGF, is regulated by various factors. It has been demonstrated that TGF-β1 has an indirect effect on angiogenesis. TGF-β1 promotes blood vessel
formation by potentiating VEGF-and bFGF-dependent angiogenesis in vascular smooth muscle cells (Brogi et al, 1994). It also upregulates the production of numerous proangiogenic factors, including VEGF, bFGF, platelet-derived growth factor, tumor necrosis factor-α, and interleukin-1 (Pepper, 1997). Our results showed that TGF-β1 up regulates the production of VEGF in all three pancreatic cancer cell lines examined. These findings suggested that the presence of TGF-β1 in the microenvironment produced by tumours and their surrounding tissues may play an important role in enhancing the liver metastasis of pancreatic cancer by modulating the capacity of angiogenesis with up-regulation of VEGF production. Pertovaara et al (1994) reported that TGF-β treatment of human lung adenocarcinoma A549 cells results in the induction of VEGF mRNA. They also found that TGF-β induced the expression of c-jun, junB, and c-fos genes and the activation of AP-1 transcription factors in A549 cells (Pertovaara et al, 1989). On the other hand, it has been shown that the promoter region of the VEGF gene contains several potential binding sites for AP-1 (Tischer et al, 1991). Thus these transcription factors could mediate VEGF induction by TGF-β.
TGF-β1 has also been shown to promote angiogenesis directly. Some of the most compelling evidence comes from TGF-β1 null mice, in which homozygous deletion of TGF-β1 is lethal. These mice have significant defects in the yolk sac vasculature and hematopoietic system, including increased vessel wall fragility and frequent disruptions between endothelial cells. This demonstrates that a deficiency in the production of TGF-β1 markedly affects the establishment and maintenance of vessel wall integrity (Dickson et al, 1995).
Among its numerous functions, TGF-β1 has been shown to act as a potent immunosuppressive factor by inhibiting the proliferation and function of natural killer (NK) cells, lymphokineactivated killer (LAK) cells, cytotoxic T-cells and B-cells (Kehrl et al, 1986;Tada et al, 1991;Rook et al, 1986;Espevik et al, 1988;Wahl et al, 1989). Malignant tumour cells escape from these effector cells and then can grow and metastasize. Immunogenicity is thus also thought to be important in determining the capacity for tumour invasion and metastasis.
Cytotoxic lymphocytes adhere to cancer cells and lyse them by recognizing various molecules on their surface. Among the molecules, HLA class I and II antigens, ICAM-1, B 7, and LFA-3 have been reported to play important roles in mediating the cytotoxic effects of lymphocytes (Ishii et al, 1994;Damale, 1992). In the present study, we found that TGF-β1 inhibit the attachment of PBML to cancer cells and decreased cell cytotoxicity, both of which may increase the capacity for metastasis. These results suggested that TGF-β1 may influence the surface expression of these adhesion molecules. In the future, the detailed mechanisms of this phenomenon should be investigated.
There are many other adhesion molecules on the surface of cancer cells, such as the carbohydrate antigens, sLe a and sLe x , both of which are known to have a strong connection with hematogenous metastasis in pancreatic cancer (Sawada et al, 1994;Kishimoto et al, 1996). These antigens are important ligands for Eselectin, which mediates the attachment between cancer cells and endothelial cells in target organs in the cancer cell lines of several species (Takada et al, 1991;Iwai et al, 1993). Okazaki et al (1998) also reported that sLe a and sLe x appear to be involved in the increase of metastatic activity of colon cancer. The present finding that sLe a and sLe x were highly expressed in highly metastatic SW1990 cells was consistent with these previous reports. We then investigated the effect of TGF-β1 on the cell surface expression of sLe a and sLe x , but found that it had no influence. Our present results showed that the liver metastatic potential of Panc-1 cells was not enhanced by in vivo splenic injection after the treatment of TGF-β1, although the capacity for immunogenicity and angiogenesis were modulated by TGF-β1 as in the other two cancer cell lines. Panc-1 cells expressed neither sLe a nor sLe x , while Capan-2 cells expressed sLe a . Therefore Panc-1 cells must be lacking in adhering to endothelial cells, an important process in the development of hematogenous metastasis. This is supposed to be the reason why Panc-1 cells did not metastasize after treatment of TGF-β1, compared to Capan-2 cells.
TGF-β1 is known to be potent inhibitor of proliferation in most cells, including some cancer cells, and exerts its effects through an interaction with type I and type II membrane receptors. Recently, it has been demonstrated that a loss of responsiveness to TGF-βmediated growth inhibition was involved in alteration of these receptors and in tumour progression (Kim et al, 1998). In the present study, TGF-β1 only slightly affected the proliferation of cancer cells but significantly enhanced liver metastasis. These results might be related to the alteration of TGF-β receptors and / or to the alteration of postreceptors that are secondary to structural or functional abnormalities in the p53 or DPC4 tumour suppressor genes (Wyllie et al, 1991;Hahn et al, 1996).
In summary, we conclude that the presence of TGF-β1 in the microenvironment of the tumour site may play an important role in enhancing the liver metastasis of pancreatic cancer by promoting the escape of cancer cells from immunosurveillance, in addition to its effect on tumour angiogenesis. Further elucidation of this process might lead to development of new therapeutic strategies and, in turn, to a decrease in the high morbidity and mortality of pancreatic cancer.