Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 May 5.
Published in final edited form as: Cancer Res. 2009 Sep 8;69(18):7320–7328. doi: 10.1158/0008-5472.CAN-09-0315

Integrin α2 Mediates Selective Metastasis to the Liver

Kiyoshi Yoshimura 1,3, Kristen F Meckel 3, Lindsay S Laird 3, Christina Y Chia 3, Jang-June Park 3, Kelly L Olino 1, Ryouichi Tsunedomi 5, Toshio Harada 5, Norio Iizuka 6, Shoichi Hazama 5, Yukihiko Kato 3, Jesse W Keller 4, John M Thompson 3, Fumin Chang 3, Lewis H Romer 2, Ajay Jain 1,3, Christine Iacobuzio-Donahue 4, Masaaki Oka 5, Drew M Pardoll 3, Richard D Schulick 1,3
PMCID: PMC4857201  NIHMSID: NIHMS779910  PMID: 19738067

Abstract

Cancers display distinct patterns of organ-specific metastasis. Comparative analysis of a broad array of cell membrane molecules on a liver-metastasizing subline of B16 melanoma versus the parental B16-F0 revealed unique up-regulation of integrin α2. The direct role of integrin α2 in hepatic metastasis was shown by comparison of high versus low-expressing populations, antibody blockade, and ectopic expression. Integrin α2–mediated binding to collagen type IV (highly exposed in the liver sinusoids) and collagen type IV–dependent activation of focal adhesion kinase are both known to be important in the metastatic process. Analysis of primary colorectal cancers as well as coexisting liver and lung metastases from individual patients suggests that integrin α2 expression contributes to liver metastasis in human colorectal cancer. These findings define integrin α2 as a molecule conferring selective potential for formation of hepatic metastasis, as well as a possible target to prevent their formation.

Introduction

Most cancer patients ultimately die from the metastatic burden of their disease. Thus, understanding the molecular interactions involved in both general and organ-specific metastasis is critical. The metastatic process is dynamic and involves numerous complex interactions between tumor and host. Tumors will develop metastatic disease only if all steps in this cascade of sequential biological events occur. The metastatic process can be formalized into a series of discrete stages as illustrated by the model introduced by Fidler and colleagues (1). These stages include escape from the primary tumor, intravasation into the lymphatic or vascular systems, survival in the circulation, avoidance of host defense mechanisms, arrest at a new site, extravasation into the tissue, and finally growth at the new site (2).

Tumor progression and metastasis depend on factors intrinsic to the tumor cell such as growth factors and receptors, extracellular matrix proteins, proteases, chemokines, and cellular adhesion molecules, as well as host factors. Cellular adhesion molecules are important in various biological processes, including organogenesis, tissue homeostasis, wound healing, and inflammatory/immune responses (3). Adhesion molecules including cadherins, integrins, and immunoglobulin superfamily molecules have been implicated in many of the steps in the metastatic cascade (4). Adhesion molecules are expressed on the cell surface and mediate adhesion between cells of the same or different types.

Among the most important adhesion molecules implicated in metastasis are the integrins, heterodimeric transmembrane receptors found throughout metazoan development (5). In humans, there are 24 different functional integrins formed by the different combination of 18 α and 8 β subunits (6, 7). Many integrins bind to components of the extracellular matrix such as laminins, collagens, and fibronectin. Integrins are involved in various aspects of cellular behavior (8).

Integrin-mediated cell adhesions can cause cell signaling that trigger calcium fluxes, activate tyrosine and serine/threonine protein kinases, inositol lipid metabolism, and regulate the activity of GTPases that control the actin cytoskeleton (8). Besides mediating stable adhesion, integrins play a role in cellular motility (9). Cell migration is essential for embryonic development, immune responses, and tissue repair, whereas dysregulated migration characterizes metastatic cancers (10).

Integrin-mediated migration and interactions with other cells such as endothelial cells contributes to tumor cell invasion and metastasis (11). During tumor progression, tumor cells gain and lose expression of specific integrins depending on the environmental restraints encountered.

The liver is one of the most important targets for organ-specific metastasis in many cancer types and is commonly the sole site of metastasis for gastrointestinal cancers. It is likely that the portal drainage of the gastrointestinal tract is partially responsible for the high rate of liver metastasis in gastrointestinal tumors but additional molecular variables are undoubtedly critical in determining whether a given cancer will metastasize to the liver. To define cell membrane molecules mediating liver metastasis, we used the classic approach of repetitive in vivo metastasis selection to derive liver-metastasizing sublines of a tumor with normally low intrinsic potential for hepatic metastasis. Analysis of a broad array of cell membrane adhesion molecules and chemokine receptors revealed that the integrin α2 (CD49b) was uniquely up-regulated relative to the nonliver-metastasizing parental line. Antibody blockade and forced expression of integrin α2 showed a direct role in liver metastasis, collagen type IV binding, and collagen type I–dependent activation of focal adhesion kinase (FAK), a critical intracellular kinase involved in the metastatic process (12, 13). Analysis of metastatic colon tumors from patients provided evidence that integrin α2 plays a role in liver metastasis in human cancer.

Materials and Methods

Mice and tumor cell lines

Female C57BL/6J (B6) mice and BALB/c mice (ages 8–10 wk) were purchased from the National Cancer Institute. B16 is a melanoma cell line of B6 origin. Colon tumor 26 cells (CT26) are murine colon adenocarcinoma cells derived from BALB/c mice (14). 4T1 breast tumor cells are spontaneously developed in BALB/c mice. All tumor cell lines were purchased from American Type Culture Collection. These tumor cell lines were cultured as previously described (15).

Hepatic metastasis model

The hepatic metastasis model has been previously described (15–17). All experimental subjects were treated ethically in accordance with an approved animal use protocol.

B16 subpopulation with propensity to form hepatic metastases

Two to 3 wk after tumor injection, metastatic deposits within the liver were resected, mechanically dispersed into a single-cell suspension, and cultured in vitro. These tumor cells were then reinjected into naïve B6 mice again. After repeating this procedure eight times, a cell line with high propensity to form hepatic nodules was established, B16-KY8 (Fig. 1A). B16-KY4 (KY4) was an intermediate cell line derived from B16-F0 by four passages in the hepatic metastasis model.

Figure 1.

Figure 1

Open in a new tab

Establishment and characterization of B16-KY8. A, scheme for establishment of B16-KY8. B16-F0 was subjected to eight rounds of sequential splenic administration followed by extraction and culture of liver nodules. With each passage, tumors were selected from the liver until the eighth passage produced a cell line with high propensity to form liver metastases. B, patterns of hepatic and i.p. metastases at death or when moribund. C, appearance of liver disease in mice challenged with the various B16 cell lines and necropisied on day 17. D, weight of livers of mice from C.

Cell growth velocity

B16 melanoma subpopulations were seeded at an initial concentration of 1 × 104/10 mL into flasks and their contents counted after 3, 6, and 9 d.

Flow cytometric analysis of cell surface markers

The following staining and blocking antibodies were used. Cells were blocked with Fc receptor antibody and stained with: anti CD49b HMα2, CD49b DX5, CD29, CD49a, CD49d, CD49e, CD49f, CD51, CD11a, CD11b, CD11c, CD25, CD31, CD34, CD43, CD44, CD54, CD62L, CD106, CD137, CD154, CD184, CD195, ANNEX V, B7H1, ICOS, Gr-1, LPAM-1 (BD Biosciences), B7RP-1, NK1.1, OX40L (eBioscience), CCR10 (Capralogics, Inc.), c-Met (Upstate), and Integrin α7 (R&D Systems), and the cells were counted for specific cell surface marker expression on a FACScan flow cytometer (BD Biosciences). Matched isotype control antibodies were used. All staining experiments involving anti integrin α2 were done with antibody from clone HM α2.

Blocking of CD49b

Integrin α2 was blocked by purified anti CD49b (HM α2) antibody (BD Biosciences) at a dilution of 1:10. The efficacy of blocking was >99% and confirmed by flow cytometry (Data not shown).

Cell attachment assay to collagen type IV–coated plates

We assessed the ability of B16-F0 and B16-KY8 cells to attach to collagen type IV, collagen type I/III, fibronectin, and laminin-coated (Sigma-Aldrich) plates. Each reagent (Sigma-Aldrich) was diluted to 50 μg/mL with 0.02N acetic acid and added to plates. This was incubated at room temperature for 2 h, then rinsed with PBS. The cells were seeded at a density of 5 × 105 cells/plate and incubated for 2 h at 37 °C. After incubation, the plates were washed gently with PBS twice to remove the nonadherent B16 cells.

Western blotting

The cells were plated in Petri dish with or without collagen type IV–coated for 30, 60, 90, 120, and 240 min, then washed twice with PBS and resuspended and incubated for 30 min at 4°C with 1% Nonidet P-40, 50 mmol/L HEPES (pH 7.5), 5 mmol/L EDTA, 50 mmol/L NaCl, and 1× protease inhibitor mixture (Roche). The amount of proteins was measured by Micro BCA protein assay kit (Pierce). Proteins (100 Ag/lane) from the cell lysates were applied to 4% to 15% Tris-HCl gel (Bio-Rad) and blotted with primary antibodies as follows: anti-FAK polyclonal antisera 5158 (Romer lab), antiTyr 397 phosphorylated FAK (Santa Cruz), Akt (pan), and Phospho-Akt (Ser473; Cell Signaling).

Flow cytometric sorting of integrin α2 high and low cells

Integrin α2 high and Integrin α2 low expressing B16-KY8 cells were sorted using a FACSVantage SE cell sorter (BD Biosciences). Those cells in the highest and lowest 10 percentile of the total population were recovered and used.

Gene transfer of integrin α2 to B16-F0

A plasmid vector encoding murine Integrin α2 was kindly provided by Dr. Jeff Bergelson (University of Pennsylvania, Philadelphia, Pennsylvania). B16-F0 was transfected with murine integrin α2–expressing plasmid using Lipofectamine 2000 reagent (Invitrogen Life Technologies) according to the manufacturer’s instructions. The resultant cell lines, B16-SKY, expresses high levels of integrin α2 and B16-SKYP expresses similar levels of integrin α2 as KY8.

Human tissue samples and immunohistochemical labeling for CD49b

Deidentified clinicopathologic data and paraffin-embedded materials from subjects with metastatic colonic cancer were collected from the Surgical Pathology Files of the Johns Hopkins Hospital. The slides were cooled for 15 min then incubated with a 15 μg/mL dilution of monoclonal antibody to CD49b (Santa Cruz Biotechnology) overnight at room temperature. Immunohistochemical labeling of each tissue section was scored on an intensity scale of 0 to 3, with 0 corresponding to no positive cells were observed, 1 corresponding to weak positive labeling (labeling best seen at 10× power or greater), 2 corresponding to unequivocal positive labeling, and 3 to intense positive labeling. The percentage of cells with positive labeling was also scored from 1% to 100%. The H score is obtained by multiply the relative staining intensity by the percent of cells staining.

Statistical analysis

Statistical analysis was performed by a Mann-Whitney U test method and an analysis of χ2. A P value of <0.05 was considered statistically significant. Results are presented as mean ± SEM.

Results

In vivo selection of liver metastasizing subline of B16 melanoma

B16-F0 melanoma cells were subjected to eight rounds of in vivo selection for liver metastasis using an intrasplenic injection route (Fig. 1A). Intrasplenic injection of B16-F0 generates very few hepatic metastases; instead demonstrating massive peritoneal dissemination (Fig. 1B–D). A rare liver metastasis was explanted and expanded in vitro followed by intrasplenic readministration (Fig. 1A). After four (B16-KY4) and eight rounds of selection (B16-KY8), tumor lines were established. B6 mice were challenged with four different tumor lines (B16-F0, B16-KY4, B16-KY8, and B16-F10, a subline selected for lung metastatic potential) in the hepatic metastasis model. At the time of death, mice were necropsied and inspected for patterns of metastasis. Mice challenged with B16-F0 and B16-F10 had marked peritoneal dissemination of their disease and relatively small disease burden in the liver. In striking contrast, mice challenged with B16-KY8 had significant disease burden in the liver with little peritoneal dissemination. Mice challenged with B16-KY4 had an intermediate pattern of metastasis (Fig. 1B).

To better quantify hepatic metastases, B6 mice challenged with the B16-F0, B16-F10 B16-KY4, and B16-KY8 lines were necropsied on day 17 after intrasplenic injection, and livers were inspected and weighed (Fig. 1C and D). The B16-KY8 challenged mice had significantly higher liver weights and hepatic disease burden compared with the other groups (P < 0.02). In addition to having more disease in the liver, the peritoneal cavity of the mice challenged with B16-KY8 cells had much less extrahepatic dissemination. B16-F0, B16-F10 B16-KY4, and B16-KY8 cell lines were found to have similar growth rates as assessed by cell doubling time (Supplementary Fig. S1), and this cannot be the cause of increased tumor volume in the liver.

Assessment of cell surface molecule expression associated with liver metastasis

The four B16 cell lines were analyzed by flow cytometry to survey a broad panel of adhesion molecules. We concentrated on molecules related to cell attachment, and previously studied molecules described as important in the metastatic process. Of over 30 membrane molecules analyzed quantitatively for surface expression, including chemokine receptors previously implicated in hepatic homing, only integrin α2 (also known as CD49b) was differentially expressed on B16-KY8 cells relative to the other three B16 cell lines (Supplementary Fig. S2A–D). Because integrin α2 combines in a heterodimeric complex with integrin β1 (also known as CD29) to form VLA2, we also performed flow cytometric assays to test differences in expression of integrin β1. The four B16 cell lines expressed similar amounts of integrin β1 but only the B16-KY8 cell line expressed higher amounts of integrin α2 (Fig. 2A).

Figure 2.

Figure 2

Open in a new tab

Integrin α2 expression by B16-KY8 and effect of integrin α2 blockade on patterns of metastasis. A, FACS analysis of the two subunits of VLA2 on all B16 sublines. B, 10 of 10 mice (100%) challenged with integrin α2 high cells developed liver metastases, and 3 of 10 (30%) mice had limited peritoneal disease (P = 0.001). Eleven of 11 mice (100%) challenged with integrin α2 low cells developed peritoneal carcinomatosis, as well as hepatic metastases. Numbers of nodules in the liver and peritoneum for integrin α2 high versus integrin α2 low cell lines, and fewer peritoneal nodules compared with the integrin α2 low cell line.

B16-KY8 cells were sorted into two populations: integrin α2 high and low expressing cells. Mice were challenged with these cells using the intrasplenic injection model and then sacrificed on day 21. When the high integrin α2 cells were used, 30% of mice had limited peritoneal disease. In contrast, 100% of mice receiving intrasplenic injection of the low integrin α2 cells developed peritoneal dissemination. Although 100% of mice in both groups developed hepatic metastases, the number of nodules in the liver was significantly higher in the group with the high integrin α2 tumor cells challenge compared with the group challenged with the low integrin α2 B16-KY8 cells (Fig. 2B). At the same time, the number of peritoneal nodules in the high integrin α2–injected mice was much less than the number in the low integrin α2 (high, 0.7 ± 0.5; low, 31.1 ± 6.4; Fig. 2B). When the high integrin α2 cells were used to challenge mice via tail vein injection, a route that allows for initial passage of tumor cells through the lungs, 27% of the mice developed liver metastasis and all 100% developed lung metastases. When the low integrin α2 cells were used to challenge mice via tail vein injection, 100% developed lung metastases but none developed liver metastases (data not shown). The number of lung metastases did not significantly differ between the two groups (data not shown). These results provided evidence that integrin α2 expression might indeed enhance formation of liver metastasis.

Integrin α2 blockade inhibits hepatic metastasis, whereas α2 integrin overexpression promotes selective hepatic metastasis

Blocking B16-KY8 cells with anti–integrin α2 antibody in the hepatic metastasis model was studied. Although treatment of B16-KY8 cells by blocking with anti–integrin α2 antibody did not completely eliminate hepatic metastasis, it significantly reduced the number relative to treatment with the isotype control antibody (isotype, 27 ± 6; blocked, 18 ± 6; P = 0.013; data not shown). Importantly, anti–integrin α2 antibody treatment concomitantly increased peritoneal dissemination of B16-KY8 (isotype, 2 ± 3; blocked, 24 ± 17; Fig. 3A), demonstrating that integrin α2 blockade did not reduce generic metastatic potential but rather shifted metastasis away from the liver.

Figure 3.

Figure 3

Open in a new tab

Integrin α2 blockade inhibits formation of hepatic metastases. Forced expression of integrin α2 in a B16 cell line that normally has low expression alters the pattern of metastatic spread. A, number of liver nodules after challenge of B16-F0, B16-KY8, and B16-SKY via hemispleen was counted. Effect of anti–integrin α2 blockade on B16-SKY metastasis patterns to liver, lung, and peritoneum after hemispleen injection was shown as percent of mice with metastases in each of the organ sites (10 mice/group). Effect of anti–integrin α2 blockade on B16-SKY liver metastases was shown as number of nodules per liver (B16-SKY isotype antibody, 30 mice/group; B16-SKY anti–integrin α2 blockade, 19 mice/group). Effect of anti–integrin α2 blockade on B16-SKY peritoneal metastases was shown as number of nodules per peritoneum (10 mice/group). B, effect of anti–integrin α2 blockade on B16-SKY metastasis patterns to liver, lung, and peritoneum after i.v. injection was shown as percent of mice with metastases in each of the organ sites (10 mice/group). C, comparison of B16-SKY and B16-SKYP and effect of anti–integrin α2 blockade on B16-SKY and B16-SKYP liver metastases was shown as number of nodules per liver.

To further show that integrin α2 can promote selective hepatic metastasis, we transfected B16-F0 with the integrin α2 gene to generate the B16-SKY line expressing high levels of integrin α2 and the B16-SKYP line expressing similar levels of integrin α2 to B16-KY8 (Fig. 3A). We first used B16-SKY to challenge mice by both intrasplenic and i.v. routes. After intrasplenic injection, B16-SKY produced significantly more hepatic metastastic nodules than B16-F0 (although not as many as B16-KY8; B16-F0 versus B16-KY8, P = 0.0006; B16-F0 versus B16-SKY, P = 0.0023; Fig. 3A). We also studied the effect of blocking integrin α2 expression on B16-SKY on the pattern of metastases after intrasplenic and tail vein injection. The most notable effect of blocking integrin α2 expression on B16-SKY in the hepatic metastasis model was to increase the rate of peritoneal metastases (the number of nodules: B16-SKY with isotype antibody, 5.1 ± 9.8; B16-SKY blocked, 17.8 ± 6.1; P = 0.0011), although there was additionally a decrease in liver metastases (the number of nodules: B16-SKY with isotype antibody, 11.6 ± 1.1; B16-SKY blocked, 7.1 ± 1.3; P = 0.02; Fig. 3A). Blocking integrin α2 expression on B16-SKY decreased the rate of hepatic metastases and increased the rate of peritoneal metastases after i.v. injection (Fig. 3B). There was no difference in the number of pulmonary metastases, further supporting the notion that enhancement of metastatic potential by integrin α2 expression is selective for the liver. To determine whether physiologic levels of integrin α2 mediate selective liver metastasis, we injected B16-SKY and B16-SKYP with or without blockade of integrin α2 via hemispleen, we then counted the number of nodules in the liver on day 17 after injection. The number of nodules in B16-SKYP was similar to those in B16-SKY (B16-SKYP, 12.3 ± 1.8; B16-SKY, 11.6 ± 1.1; not statistically significant). This tropism was abrogated with blockade of integrin α2 expression (100% blockade of integrin α2 expression at the time of tumor injection; B16-SKYP blocked, 5.2 ± 1.6; B16-SKY blocked, 7.1 ± 1.3; B16-SKY versus B16-SKY blocked, P = 0.02; B16-SKYP versus B16-SKYP blocked, P = 0.008; Fig. 3C).

In previous studies to develop the intrasplenic injection model of hepatic metastases, we used a BALB/c-derived colon cancer, CT26, which, unlike B16-F0 or B16-F10, shows a high degree of liver metastasis and little peritoneal dissemination. CT26 was found to naturally express integrin α2, and we therefore studied whether antibody blockade of integrin α2 altered the metastatic pattern. Indeed, treatment of CT26 with anti–integrin α2 antibody before injection altered the metastatic pattern, diminishing liver metastasis (CT26 isotype control, 12.5 ± 2.4; CT26 blocked, 3.8 ± 1.5; P = 0.009) and increasing peritoneal dissemination (CT26 isotype control, 3.2 ± 1.9; CT26 blocked, 11.7 ± 2.4; P = 0.014; Fig. 4A and B). To evaluate additional tumors naturally expressing integrin α2, we tested an integrin α2 expression breast cancer, 4T-1 in the hemispleen model. 4T-1 efficiently generates liver metastases, and treatment with anti–integrin α2 antibody before injection showed the same findings as CT26 and B16-KY8 (4T-1 isotype control, 60.4 ± 11.8; 4T-1 blocked, 13.2 ± 9.4; P = 0.03; Fig. 4C). These results strengthen the notion that enhancement of hepatic metastasis by integrin α2 is not a phenomenon specific to the B16 tumor.

Figure 4.

Figure 4

Open in a new tab

Blocking integrin α2 expression on CT26 and 4T-1 alters their patterns of metastases. A and B, effect of blocking integrin α2 expression on CT26 on the pattern of metastases in mice challenged via splenic injection route. The number of nodules within the liver and peritoneum in the CT26 unblocked (isotype control) versus CT26 integrin α2 blocked is shown. C, effect of blocking integrin α2 expression on 4T-1 on the pattern of metastases in mice challenged via splenic injection route. The number of nodules within the liver in the 4T-1 unblocked versus 4T-1 blocked is shown. Ab, antibody.

Integrin α2 expression enhances binding to collagen type IV

Integrins α2 mediates collagen binding, and the liver is particularly rich in type IV collagen. We therefore asked whether integrin α2 up-regulation might enhance collagen type IV binding. An attachment assay to collagen type IV–coated plates was per formed using B16-F0 and B16-KY8. The counting was performed in a “blinded” fashion. The B16-F0 group displayed low numbers of attached tumor cells on the collagen type IV–coated plate (138 ± 17). In contrast, the B16-KY8 group displayed a >6-fold increase in attachment (881 ± 95, P < 0.001; 8 per group; Fig. 5A and B). When either B16-KY8 or B16-SKYP was first treated with anti– integrin α2 blocking antibody, attachment to collagen type IV–coated plates was significantly inhibited [B16-KY8, 786 ± 30; B16-KY8 blocked, 163 ± 42; P < 0.0001; B16-SKYP, 829 ± 28; B16-SKYP blocked, 142 ± 39; P < 0.0001; 8 high power field (HPF) per group; Fig. 5C]. Additionally, both B16-SKYP and B16-KY8 cells had high affinity for collagen type IV (B16-F0, 102 ± 12; B16-KY8, 367 ± 53; B16-SKYP, 502 ± 178; both B16-F0 versus B16-KY8 and B16-F0 versus B16-SKYP, P < 0.003; 4 HPF per group). These cells were also incubated on collagen type I/III or fibronectin-coated plates; there was very little attachment and no enhancement by integrin α2 expression. Integrin α2 expression correlated with a slight increase in attachment to laminin-coated plates. (uncoated: B16-F0, 7 ± 2; B16-KY8, 10 ± 2; B16-SKYP, 3 ± 1; Laminin: B16-F0, 47 ± 1; B16-KY8, 81 ± 7; B16-SKYP, 114 ± 8; both B16-F0 versus B16-KY8 and B16-F0 versus B16-SKYP, P < 0.003; collagen type I/III: B16-F0, 2 ± 1; B16-KY8, 4 ± 1; B16-SKYP, 3 ± 1; Fibronectin: B16-F0, 2 ± 1; B16-KY8, 19 ± 7; B16-SKYP, 6 ± 1; 4 HPF per group; Fig. 5C).

Figure 5.

Figure 5

Open in a new tab

Effect of collagen type IV, I/III, fibronectin, and lamin in attachment of the various B16 cell lines. A and B, to study the intensity of attachment of the B16 cell lines to collagen type IV–coated plates, an attachment assay was performed using B16-F0 and B16-KY8 (8 HPF) per group. C, when either B16-KY8 or B16-SKYP was blocked with anti–α2 integrin blocking antibody, both cell lines significantly lost their ability to attach to collagen type IV–coated plates. When the B16-SKYP and B16-KY8 cell lines were used in these assays comparing to B16-F0, they also attached strongly to the collagen IV–coated plates (4 HPF per group). When these cells were incubated on collagen type I/III or fibronectin-coated plates, there was very little attachment and no enhancement by integrin α2 expression. Integrin α2 expression marginally enhanced binding to laminin-coated plates.

Integrin α2 enhances collagen type IV–dependent FAK activation

FAK plays a major role in cell motility and has been shown to be activated by engagement of a number of integrins, including integrin α2, and is thought to be an important factor in the metastatic process. Indeed, attachment of B16-KY8 and B16-SKY to collagen type IV–coated plates resulted in significant FAK activation (as measured by tyrosine phosphorylation at Y397), peaking at 90 min, whereas B16-F0 cells plated on collagen IV failed to activate FAK (Fig. 6A). FAK activation was collagen type IV dependent (Fig. 6A and B) and specific, because another kinase involved in metabolic regulation of tumor cells, AKT, was equally activated in B16-F0, KY8, or SKY and was not affected by binding to collagen IV (data not shown). This study shows that induction of integrin α2 results in collagen type IV–dependent FAK activation and provides a potential mechanistic basis for its role in mediating selective hepatic metastasis.

Figure 6.

Figure 6

Open in a new tab

Integrin α2 binding to collagen type IV mediates FAK activation. A, Western blotting for total FAK and tyrosine-phosphorylated (Y397–) FAK in three cell lines. Plating B16-KY8 and B16-SKY but not B16-F0 on collagen type IV–coated plates increases activation of FAK (phospho-FAK). Total FAK remains unchanged. B, time course of total FAK expression and Y397 phosphorylation. Tyrosine-phosphorylation of FAK (Y397) with time following plating on collagen type IV until a peak of 90 min. Expression of integrin α2 and patterns of metastases of human colon cancer.

Correlation between integrin α2 expression and hepatic metastasis in human colorectal cancer

We also analyzed colorectal tumors from patients with simultaneous liver and lung metastases. Metastases are significantly less infiltrated with leukocytes than primary colorectal tumors, allowing direct analysis of integrin α2 on tumor cells by immunohistochemical staining. Sections of concurrent liver and lung metastases available from seven colorectal cancer patients were stained with anti–integrin α2 antibody and processes for immunohistochemistry as described. The staining patterns are summarized as in Supporting Table S1. In six of seven cases, the liver metastases had higher levels of staining than the pulmonary metastases (P = 0.0075). In roughly half the cases, hepatic metastases showed extremely strong staining for integrin α2, whereas the corresponding lung metastases had extremely low staining (Supplementary Fig. S3). These results in both primary and metastatic human colorectal cancers show a correlation between integrin α2 expression and propensity to metastasize to the liver.

Discussion

The formation of metastatic disease from a tumor primary requires multiple complex processes that have been associated with the up-regulation of various cell surface molecules, and in particular integrins (18, 19).

We hypothesized that differential expression of one or more cell surface molecules would be responsible for differences in metastatic pattern. Remarkably, analysis of a broad panel of integrins, chemokine receptors, and cellular adhesion molecules revealed a single molecule, integrin α2, which was differentially expressed between the nonliver metastasizing parental B16-F0 tumor and the liver metastasizing B16-KY8. We used multiple complimentary approaches to show that integrin α2 plays a causal role in liver metastasis. Additionally, two other cell lines that express integrin α2 showed liver-metastasizing capacity that was blocked by anti–integrin α2 antibodies. In vitro studies suggested that integrin α2 may facilitate hepatic metastasis via binding to collagen type IV and consequent activation of FAK.

We also studied integrin α2 expression in patient samples. We studied matched hepatic and pulmonary metastases from patients with colorectal cancer. In two cases, we also had the matched primary specimens. We found increased cell surface expression of the integrin α2 protein on tumor cells in hepatic metastases compared with pulmonary metastases. This finding is compatible with a report that uveal melanomas, which preferentially metastasize to the liver, show increases in expression of integrin α2 (20).

Integrin α2 β1 also known as very late antigen 2 (VLA2) is primarily found on activated T cells, platelets, and have been described on epithelial cells such as keratinocytes (21). Interestingly on keratinoyctes, VLA2 is considered a differentiation marker. Upon terminal differentiation, keratinocytes not only lose their contact to the basement membrane, but also decrease their expression of VLA2 drastically (22). The expression of the other two collagen-binding integrins, α10 β1 and α11 β1, is more restricted to chondrocytes and fibroblasts, respectively (23, 24).

Integrin α2 β1 can bind with collagen type I, II, and IV (25, 26). Collagen type I is found primarily in bone, dermis, tendon, ligaments, and cornea (27). Collagen type II is found primarily in cartilage, vitreous body, and nucleus pulposus (27). Collagen type IV is found primarily in basement membrane (27). In our murine models of hepatic, pulmonary, and peritoneal metastases, collagen type IV, which is highly enriched in the liver, is the most relevant collagen subtype to implicate in an integrin-collagen interaction.

Although a number of studies have implicated integrin expression in the metastatic process, including liver metastases, the findings presented here are the first to directly show a role for integrin α2 in selective liver metastasis. A few reports have implicated α6 integrin (associated with various β subunits), which binds laminin, in liver metastases (27–30). The most convincing report showed that both components of a complex of integrin α6 β4 and the D6.1A tetraspanin were necessary for liver metastasis in a rat pancreatic adenocarcinoma model (31). Other adhesion molecules, including integrin αv β3, osteopontin (a ligand for integrin αv β3), integrin αv β5, integrin α4 β1, and VCAM-1, have been associated with liver metastasis in various models (32–36), but their direct role has not been comprehensively confirmed by blockade and overexpression analyses, nor comparative expression analysis in human tumors as we have done for integrin α2 in the current report. In contrast to our findings, Haier and colleagues (37) reported similar levels of integrin α2 expression by poorly versus highly liver-metastasizing sublines of human colon tumor line, as well as lower collagen IV binding and higher collagen I binding by the highly liver-metastasizing line. The basis for these differences is unclear, although we note that no quantitative analysis [e.g., by fluorescence-activated cell sorting (FACS)] of integrin α2 cell surface expression was performed. It is possible that the xenogeneic nature of their system (human tumor cells injected into mice) obscured a role for species-specific interactions. Alternatively, it is possible that other molecular interactions besides integrin α2–collagen type IV mediate liver metastases in their xenogeneic system. Indeed, the fact that in vivo selected B16-KY8 shows somewhat greater liver-metastasizing efficiency than the integrin α2 transfectant, B16-SKYP, despite comparable levels of integrin α2 expression, indicates that multiple pathways contribute to this process.

Although this is the first report to implicate integrin α2 in hepatic metastasis, two earlier studies have shown that integrin α2 plays a role in the intrahepatic migration of tumor cells. One report showed that chemotactic and haptotactic migration of hepatocellular carcinoma cells in response to growth factors produced in a fibrotic microenvironment involves integrin α2 (38). A second report showed that although integrin α2 did not affect extravasation of rhabdomyosarcoma cells in the liver, it facilitated their intrahepatic migration to the subcapsular space once in the liver parenchyma (39). It remains to be determined whether α2 integrin–mediated adhesion to collagen type IV with subsequent FAK activation associated with hepatic metastasis in the current report directly enhances intrahepatic migration of tumor cells.

The link between integrins and intracellular signaling molecules involved in cell migration, such as FAK, is an emerging theme in metastasis research (40). Of note, FAK activation has been linked to multiple cell activities, such as tissue invasion, matrix assembly, cell mobility, and cell fate (41). Whether FAK signaling always enhances liver metastasis is unclear because it has been found to be up-regulated in metastases in other organs (12, 13). However, in our animal model and in human colorectal cancer, the first major capillary and tissue bed encountered by potentially metastatic cells from either splenic injection or from a colorectal primary is the liver. Additionally, it may be the combination of the extracellular adhesion interactions between integrin α2 β1 and collagen type IV in combination with the intracellular signaling through FAK that enhances liver metastases.

In summary, we have provided direct evidence in murine models and correlative evidence in human tumors that integrin α2 β1 is one of the determinants conferring potential for selective hepatic metastasis.

Supplementary Material

fig1
fig2

Acknowledgments

Grant support: Commonwealth Foundation, Charles Delmar Foundation, and gifts from Robert and Jacque Alvord, William and Betty Topercer, and Dorothy Needle.

We thank Jeffrey Bergelson M.D. (University of Pennsylvania) who provided the CD49b gene encoding plasmid vector.

Footnotes

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

References

  • 1.Fidler IJ. Critical factors in the biology of human cancer metastasis: twenty-eighth G.H.A. Clowes memorial award lecture. Cancer Res. 1990;50:6130–8. [PubMed] [Google Scholar]
  • 2.Chambers AF, Groom AC, MacDonald IC. Dissemination and growth of cancer cells in metastatic sites. Nat Rev Cancer. 2002;2:563–72. doi: 10.1038/nrc865. [DOI] [PubMed] [Google Scholar]
  • 3.Koukoulis GK, Patriarca C, Gould VE. Adhesion molecules and tumor metastasis. Hum Pathol. 1998;29:889–92. doi: 10.1016/s0046-8177(98)90191-5. [DOI] [PubMed] [Google Scholar]
  • 4.Chothia C, Jones EY. The molecular structure of cell adhesion molecules. Annu Rev Biochem. 1997;66:823–62. doi: 10.1146/annurev.biochem.66.1.823. [DOI] [PubMed] [Google Scholar]
  • 5.Hynes RO, Zhao Q. The evolution of cell adhesion. J Cell Biol. 2000;150:F89–96. doi: 10.1083/jcb.150.2.f89. [DOI] [PubMed] [Google Scholar]
  • 6.Hynes RO. Integrins: bidirectional, allosteric signaling machines. Cell. 2002;110:673–87. doi: 10.1016/s0092-8674(02)00971-6. [DOI] [PubMed] [Google Scholar]
  • 7.van der Flier A, Sonnenberg A. Function and interactions of integrins. Cell Tissue Res. 2001;305:285–98. doi: 10.1007/s004410100417. [DOI] [PubMed] [Google Scholar]
  • 8.Danen EH. Integrins: regulators of tissue function and cancer progression. Curr Pharm Des. 2005;11:881–91. doi: 10.2174/1381612053381756. [DOI] [PubMed] [Google Scholar]
  • 9.Ridley AJ, Schwartz MA, Burridge K, et al. Cell migration: integrating signals from front to back. Science. 2003;302:1704–9. doi: 10.1126/science.1092053. [DOI] [PubMed] [Google Scholar]
  • 10.Schmid BC, Rezniczek GA, Fabjani G, Yoneda T, Leodolter S, Zeillinger R. The neuronal guidance cue Slit2 induces targeted migration and may play a role in brain metastasis of breast cancer cells. Breast Cancer Res Treat. 2007;106:333–42. doi: 10.1007/s10549-007-9504-0. [DOI] [PubMed] [Google Scholar]
  • 11.Wu WS. The signaling mechanism of ROS in tumor progression. Cancer Metastasis Rev. 2006;25:695–705. doi: 10.1007/s10555-006-9037-8. [DOI] [PubMed] [Google Scholar]
  • 12.Rodrigo JP, Dominguez F, Suarez V, Canel M, Secades P, Chiara MD. Focal adhesion kinase and E-cadherin as markers for nodal metastasis in laryngeal cancer. Arch Otolaryngol Head Neck Surg. 2007;133:145–50. doi: 10.1001/archotol.133.2.145. [DOI] [PubMed] [Google Scholar]
  • 13.Weiner TM, Liu ET, Craven RJ, Cance WG. Expression of focal adhesion kinase gene and invasive cancer. Lancet. 1993;342:1024–5. doi: 10.1016/0140-6736(93)92881-s. [DOI] [PubMed] [Google Scholar]
  • 14.Corbett TH, Griswold DP, Jr, Roberts BJ, Peckham JC, Schabel FM., Jr Tumor induction relationships in development of transplantable cancers of the colon in mice for chemotherapy assays, with a note on carcinogen structure. Cancer Res. 1975;35:2434–9. [PubMed] [Google Scholar]
  • 15.Yoshimura K, Laird LS, Chia CY, et al. Live attenuated Listeria monocytogenes effectively treats hepatic colorectal cancer metastases and is strongly enhanced by depletion of regulatory T cells. Cancer Res. 2007;67:10058–66. doi: 10.1158/0008-5472.CAN-07-0573. [DOI] [PubMed] [Google Scholar]
  • 16.Jain A, Slansky JE, Matey LC, Allen HE, Pardoll DM, Schulick RD. Synergistic effect of a granulocyte-macrophage colony-stimulating factor-transduced tumor vaccine and systemic interleukin-2 in the treatment of murine colorectal cancer hepatic metastases. Ann Surg Oncol. 2003;10:810–20. doi: 10.1245/aso.2003.10.006. [DOI] [PubMed] [Google Scholar]
  • 17.Yoshimura K, Jain A, Allen HE, et al. Selective targeting of antitumor immune responses with engineered live-attenuated Listeria monocytogenes. Cancer Res. 2006;66:1096–104. doi: 10.1158/0008-5472.CAN-05-2307. [DOI] [PubMed] [Google Scholar]
  • 18.Winter JM, Ting AH, Vilardell F, et al. Absence of E-cadherin expression distinguishes noncohesive from cohesive pancreatic cancer. Clin Cancer Res. 2008;14:412–8. doi: 10.1158/1078-0432.CCR-07-0487. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Cairns RA, Khokha R, Hill RP. Molecular mechanisms of tumor invasion and metastasis: an integrated view. Curr Mol Med. 2003;3:659–71. doi: 10.2174/1566524033479447. [DOI] [PubMed] [Google Scholar]
  • 20.Baker JK, Elshaw SR, Mathewman GE, et al. Expression of integrins, degradative enzymes and their inhibitors in uveal melanoma: differences between in vitro and in vivo expression. Melanoma Res. 2001;11:265–73. doi: 10.1097/00008390-200106000-00008. [DOI] [PubMed] [Google Scholar]
  • 21.Santoro SA, Zutter MM. The α 2 β 1 integrin: a collagen receptor on platelets and other cells. Thromb Haemost. 1995;74:813–21. [PubMed] [Google Scholar]
  • 22.Watt FM. Role of integrins in regulating epidermal adhesion, growth and differentiation. EMBO J. 2002;21:3919–26. doi: 10.1093/emboj/cdf399. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Camper L, Hellman U, Lundgren-Akerlund E. Isolation, cloning, and sequence analysis of the integrin subunit α10, a β1-associated collagen binding integrin expressed on chondrocytes. J Biol Chem. 1998;273:20383–9. doi: 10.1074/jbc.273.32.20383. [DOI] [PubMed] [Google Scholar]
  • 24.Velling T, Kusche-Gullberg M, Sejersen T, Gullberg D. cDNA cloning and chromosomal localization of human α(11) integrin. A collagen-binding, I domain-containing, β(1)-associated integrin α-chain present in muscle tissues. J Biol Chem. 1999;274:25735–42. doi: 10.1074/jbc.274.36.25735. [DOI] [PubMed] [Google Scholar]
  • 25.Eble JA. Collagen-binding integrins as pharmaceutical targets. Curr Pharm Des. 2005;11:867–80. doi: 10.2174/1381612053381738. [DOI] [PubMed] [Google Scholar]
  • 26.White DJ, Puranen S, Johnson MS, Heino J. The collagen receptor subfamily of the integrins. Int J Biochem Cell Biol. 2004;36:1405–10. doi: 10.1016/j.biocel.2003.08.016. [DOI] [PubMed] [Google Scholar]
  • 27.Gelse K, Poschl E, Aigner T. Collagens-structure, function, and biosynthesis. Adv Drug Deliv Rev. 2003;55:1531–46. doi: 10.1016/j.addr.2003.08.002. [DOI] [PubMed] [Google Scholar]
  • 28.Sawai H, Funahashi H, Yamamoto M, et al. Interleukin-1α enhances integrin α(6)β(1) expression and metastatic capability of human pancreatic cancer. Oncology. 2003;65:167–73. doi: 10.1159/000072343. [DOI] [PubMed] [Google Scholar]
  • 29.Yamaguchi K, Ura H, Yasoshima T, Shishido T, Denno R, Hirata K. Liver metastatic model for human gastric cancer established by orthotopic tumor cell implantation. World J Surg. 2001;25:131–7. doi: 10.1007/s002680020047. [DOI] [PubMed] [Google Scholar]
  • 30.Kemperman H, Driessens MH, La Riviere G, Meijne AM, Roos E. Adhesion mechanisms in liver metastasis formation. Cancer Surv. 1995;24:67–79. [PubMed] [Google Scholar]
  • 31.Herlevsen M, Schmidt DS, Miyazaki K, Zoller M. The association of the tetraspanin D6.1A with the α6β4 integrin supports cell motility and liver metastasis formation. J Cell Sci. 2003;116:4373–90. doi: 10.1242/jcs.00760. [DOI] [PubMed] [Google Scholar]
  • 32.Kikkawa H, Kaihou M, Horaguchi N, et al. Role of integrin α(v)β3 in the early phase of liver metastasis: PET and IVM analyses. Clin Exp Metastasis. 2002;19:717–25. doi: 10.1023/a:1021356019563. [DOI] [PubMed] [Google Scholar]
  • 33.Nishimori H, Yasoshima T, Hata F, et al. A novel nude mouse model of liver metastasis and peritoneal dissemination from the same human pancreatic cancer line. Pancreas. 2002;24:242–50. doi: 10.1097/00006676-200204000-00006. [DOI] [PubMed] [Google Scholar]
  • 34.Kishimoto H, Sasahara K, Yamazaki K, et al. Obstructive jaundice facilitates hepatic metastasis of B16F1 mouse melanoma cells: participation of increased VCAM-1 expression in the liver. Oncol Rep. 2001;8:575–8. doi: 10.3892/or.8.3.575. [DOI] [PubMed] [Google Scholar]
  • 35.Papadimitriou MN, Menter DG, Konstantopoulos K, Nicolson GL, McIntire LV. Integrin α4β1/VCAM-1 pathway mediates primary adhesion of RAW117 lymphoma cells to hepatic sinusoidal endothelial cells under flow. Clin Exp Metastasis. 1999;17:669–76. doi: 10.1023/a:1006747106885. [DOI] [PubMed] [Google Scholar]
  • 36.Wai PY, Mi Z, Guo H, et al. Osteopontin silencing by small interfering RNA suppresses in vitro and in vivo CT26 murine colon adenocarcinoma metastasis. Carcinogenesis. 2005;26:741–51. doi: 10.1093/carcin/bgi027. [DOI] [PubMed] [Google Scholar]
  • 37.Haier J, Nasralla M, Nicolson GL. Different adhesion properties of highly and poorly metastatic HT-29 colon carcinoma cells with extracellular matrix components: role of integrin expression and cytoskeletal components. Br J Cancer. 1999;80:1867–74. doi: 10.1038/sj.bjc.6690614. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Yang C, Zeisberg M, Lively JC, Nyberg P, Afdhal N, Kalluri R. Integrin α1β1 and α2β1 are the key regulators of hepatocarcinoma cell invasion across the fibrotic matrix microenvironment. Cancer Res. 2003;63:8312–7. [PubMed] [Google Scholar]
  • 39.Ho WC, Heinemann C, Hangan D, Uniyal S, Morris VL, Chan BM. Modulation of in vivo migratory function of α2β1 integrin in mouse liver. Mol Biol Cell. 1997;8:1863–75. doi: 10.1091/mbc.8.10.1863. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Steeg PS. Metastasis suppressors alter the signal transduction of cancer cells. Nat Rev Cancer. 2003;3:55–63. doi: 10.1038/nrc967. [DOI] [PubMed] [Google Scholar]
  • 41.Romer LH, Birukov KG, Garcia JG. Focal adhesions: paradigm for a signaling nexus. Circ Res. 2006;98:606–16. doi: 10.1161/01.RES.0000207408.31270.db. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

fig1
NIHMS779910-supplement-fig1.pdf (10.1KB, pdf)
fig2
NIHMS779910-supplement-fig2.pdf (415.5KB, pdf)

RESOURCES