Therapeutic targeting of integrin αvβ6 in breast cancer

Abstract

Background: Integrin αvβ6 promotes migration, invasion, and survival of cancer cells; however, the relevance and role of αvβ6 has yet to be elucidated in breast cancer.

Methods: Protein expression of integrin subunit beta6 (β6) was measured in breast cancers by immunohistochemistry (n > 2000) and ITGB6 mRNA expression measured in the Molecular Taxonomy of Breast Cancer International Consortium dataset. Overall survival was assessed using Kaplan Meier curves, and bioinformatics statistical analyses were performed (Cox proportional hazards model, Wald test, and Chi-square test of association). Using antibody (264RAD) blockade and siRNA knockdown of β6 in breast cell lines, the role of αvβ6 in Human Epidermal Growth Factor Receptor 2 (HER2) biology (expression, proliferation, invasion, growth in vivo) was assessed by flow cytometry, MTT, Transwell invasion, proximity ligation assay, and xenografts (n ≥ 3), respectively. A student's t-test was used for two variables; three-plus variables used one-way analysis of variance with Bonferroni's Multiple Comparison Test. Xenograft growth was analyzed using linear mixed model analysis, followed by Wald testing and survival, analyzed using the Log-Rank test. All statistical tests were two sided.

Results: High expression of either the mRNA or protein for the integrin subunit β6 was associated with very poor survival (HR = 1.60, 95% CI = 1.19 to 2.15, P = .002) and increased metastases to distant sites. Co-expression of β6 and HER2 was associated with worse prognosis (HR = 1.97, 95% CI = 1.16 to 3.35, P = .01). Monotherapy with 264RAD or trastuzumab slowed growth of MCF-7/HER2-18 and BT-474 xenografts similarly (P < .001), but combining 264RAD with trastuzumab effectively stopped tumor growth, even in trastuzumab-resistant MCF-7/HER2-18 xenografts.

Conclusions: Targeting αvβ6 with 264RAD alone or in combination with trastuzumab may provide a novel therapy for treating high-risk and trastuzumab-resistant breast cancer patients.

Figure 1.

Coexpression of integrin αvβ6 and HER2 and overall survival in breast cancer patients. Kaplan–Meier curves by integrin αvβ6 expression status. Tick marks indicate patients who were still alive at the time of analyses or who were censored. All P values refer to Wald tests as determined by the Multivariable Cox model. All tests were two sided. A) Normal and B) cancerous breast cancer tissue sections immunohistochemically stained for integrin αvβ6 (brown staining) using 6.2G2 antibody (Biogen Idec). Magnification x10, scale bar = 100 μM. Overall survival in two cohorts of breast cancer patients from London (C) and Nottingham (D) by integrin αvβ6 status (high expression represented by a dashed line, low represented by a solid line). The P value for patients with high integrin αvβ6 vs low expression in tumors is <.001. E) Overall survival of HER2+ patients from the combined London and Nottingham patient cohorts by integrin αvβ6 status. The P value for patients with high integrin αvβ6 status versus low tumors is <.001. F) Overall survival of HER2+ (ERBB2) patients from the Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) cohort by integrin αvβ6 status. The survival of patients with high ITGB6-expressing tumors vs low-expressing tumors is significantly lower (P = .003). Please also see Supplementary Tables 1–3 and Supplementary Figure 1 (available online). OS = overall survival.

Figure 2.

Integrin αvβ6 and HER2-blockade and breast cancer cell invasion. A) Expression of integrin αvβ6 and HER2 in a breast cancer cell line panel assessed by flow cytometry. Isotype controls all showed lower expression nearer the y-axis side, while integrin αvβ6 and HER2 expression shift the curves to the right (see Supplementary Table 4, available online, for full panel of cell lines analyzed). B) Transwell invasion assay of breast cancer cell lines expressing varying levels of integrin αvβ6 and HER2. 5 x 10⁴ cells/well were seeded and the number of cells that invaded was counted after 72 hours. C and D) Breast cancer cell-line invasion is integrin αvβ6 dependent. Cells were subjected to either 30 minutes of incubation with IgG or αvβ6 blocking antibody (β6 Ab) (10 μg/mL) (C) or 72 hours of transfection with control or β6 siRNA (20μM) (D) and subjected to a Transwell invasion assay as before. E and F) Breast cancer cell-line invasion is HER2 dependent. Cells were pretreated for 30 minutes with IgG or Trastuzumab (TRA) (10 μg/mL) (E) or transfected for 72 hours with control or HER2 siRNA (20μM) (F) and subjected to a Transwell invasion assay. G) Cells were pretreated for 30 minutes with IgG, β6 Ab, TRA (all 10 μg/mL), or a combination of the blocking antibodies, and subjected to a Transwell invasion assay. All experiments were performed in triplicate, representative experiments shown (n = 6, error bars represent 95% confidence interval). *P = .05, **P = .01, ***P < .001, relative to IgG or control-treated cells. C-F; Student’s t-test, B and G; one-way analysis of variance with Bonferroni’s Multiple Comparison Test. All tests were two sided. Please also see Supplementary Figure 2, available online. HER2 = Human Epidermal Growth Factor Receptor 2; IgG = immunoglobulin; TRA = trastuzumab.

Figure 3.

Role of integrin αvβ6 in HER2-driven breast cancer cell-line invasion. Cells were pretreated for 30 minutes with IgG, HRGβ (1μM) in the presence and absence of αvβ6 blocking antibody (10 μg/mL) (A) or trastuzumab (TRA) (10 μg/mL) (B) and 5 x 10⁴ cells/well seeded into a Transwell invasion assay. The number of cells invaded was counted after 72 hours. All experiments were performed in triplicate, representative experiments shown (n = 6, error bars represent 95% confidence interval). *P = .05, **P = .01, ***P < .001 (relative to IgG, two-sided, one-way analysis of variance (ANOVA) with Bonferroni’s Multiple Comparison Test). C) Organotypic invasion of MCF10.CA1a (CA1a) cell line. Cells were pretreated for 30 minutes with IgG, αvβ6 blocking antibody or TRA (10 μg/mL) or transfected with siRNA to αvβ6 or HER2 for 72 hours (20μM) prior to seeding. Gels were fixed in formal saline after 5 to 6 days incubation, paraffin embedded, sectioned and sections stained with H&E. Magnification bar = 10 μM. Histograms quantify the invasion of each cell with the aforementioned treatments as invasion index (n = 3, error bars represent 95% confidence interval). Experiments were performed in triplicate (n = 2/experiment), representative experiments shown. *P = .05, **P = .01, ***P < .001 (relative to Control siRNA treated cells, two-sided, one-way ANOVA with Bonferroni’s Multiple Comparison Test). HER2 = Human Epidermal Growth Factor Receptor 2; IgG = immunoglobulin; TRA = trastuzumab.

Figure 4.

The effect of 264RAD in combination with trastuzumab on human breast cancer xenograft growth in SCID mice. A) Mice bearing human BT-474 tumors were treated with IgG (square, solid line), 264RAD (triangle, dashed line, in line with TRA treatment), trastuzumab (TRA) (square, dashed line), or 264RAD+TRA (triangle on lower dashed line) (10mg/kg; ip) twice weekly for two consecutive weeks (start of treatment indicated by arrow, day 0). Data are presented as mean tumor volume (error bars represent 95% confidence interval, n ≥ 4 mice/group). Treatment commenced when tumors reached 100mm³. B) Mice bearing human HER2-18 tumors were treated as in (A). (C) Photographic images of representative BT-474 and HER2-18 xenografts posttreatment outlined in (A). Magnification bar = 5mm. D) BT-474 xenograft protein expression. Xenografts were treated as in (A), harvested, protein extracted, and subjected to immunoblotting. Blots were probed for indicated proteins. E) Histograms of relative protein expression from blots shown in (D) determined by optical density (n = 3 individual tumors, error bars represent 95% confidence interval). *P = .05, **P = .01, ***P < .001 (relative to IgG or to treatment indicated by corresponding lines to the side of growth curves and above histograms, as determined by plotting individual growth curves and then applying a linear mixed model to test for differences between treatments [A and B], and two-sided, one-way analysis of variance with Bonferroni’s Multiple Comparison Test [E and G]). F and G) HER2-18 xenograft protein expression and quantification as outlined in (D and E). HER2 = Human Epidermal Growth Factor Receptor 2; IgG = immunoglobulin; TRA = trastuzumab.

Figure 5.

The effect of 264RAD in combination with trastuzumab on human xenograft BT-474 tumor growth and stroma. A) Micrographs of BT-474 tumor xenografts from mice treated with IgG or 264RAD+TRA (10mg/kg; ip) twice weekly for two consecutive weeks (tumors harvested after two weeks of treatment from Figure 4A). Tumors were harvested, fixed, parafin embedded, and sections subjected to immunohistochemical staining for the indicated molecules of interest, including pancytokeratin (CK, epithelial marker), Ki67 (proliferation), endomucin (vasculature), α-sma (myofibroblast), cleaved caspase 3 (apoptosis), as well as αvβ6 and HER2 expression. Representative images are shown of the three tumors harvested for IgG and the combination treatment, where the greatest effect was observed. Scale bar in whole tumor images = 1000 μM, magnified images (x60) are of indicated region of interest (CK+ cells). Scale bar in x60 magnification images = 20 μM. B) Bar graph of composition of xenografts (% CK+ cells = white bar, % necrotic area in black, and % stroma is in gray or blue) and histograms of specific marker expression of xenografts shown in (A). Assessed and scored by two individuals (n = 3 individual tumors, error bars represent 95% confidence interval). *P = .05, **P = .01, ***P < .001 (relative to IgG, as determined by two-sided, one-way analysis of variance with Bonferroni Multiple Comparison Test [B, upper panel] and two-sided Student’s t-test [B, lower panel]). CK = pancytokeratin; HER2 = Human Epidermal Growth Factor Receptor 2; IgG = immunoglobulin; TRA = trastuzumab.

Figure 6.

The effect of 264RAD in combination with trastuzumab on human xenograft MCF-7/HER2-18 tumor growth and stroma. A) Micrographs of MCF-7/HER2-18 tumor xenografts from mice treated with IgG, or 264RAD+TRA (10mg/kg; ip) twice weekly for two consecutive weeks (tumors harvested after two weeks treatment from Figure 4B). Tumors were harvested, fixed, parafin embedded, and sections subjected to immunohistochemical staining for the indicated molecules of interest including cytokeratin (CK, epithelial marker), Ki67 (proliferation), endomucin (vasculature), α-sma (myofibroblasts), as well as αvβ6 and HER2 expression. Representative images are shown of the three tumors harvested for IgG and the combination treatment, where the greatest effect was observed. Scale bar in whole tumor images = 2000 μM, magnified images (x60) are of indicated region of interest (CK+ cells). Scale bar in x60 magnification images = 20 μM. B) Bar graph of composition of xenografts (% CK+ cells = white bar, % necrotic area in black, and % stroma is in gray or blue) and histograms of specific marker expression of xenografts shown in (A). Assessed and scored by two individuals (n = 3 individual tumors, error bars represent 95% confidence interval). *P = .05, **P = .01, ***P < .001 (relative to IgG, as determined by two-sided, one-way analysis of variance with Bonferroni Multiple Comparison Test [B, upper panel] and two-sided Student’s t-test [B, lower panel]). CK = pancytokeratin; HER2 = Human Epidermal Growth Factor Receptor 2; IgG = immunoglobulin; TRA = trastuzumab.

Figure 7.

The effect of long-term (six-week) treatment of 264RAD in combination with trastuzumab on human xenograft MCF-7/HER2-18 cell growth in SCID mice. Mice bearing human MCF-7/HER2-18 tumors were treated with IgG (square, solid line), 264RAD (triangle, dashed line, in line with TRA), trastuzumab (TRA) (square, dashed line), or 264RAD+TRA (triangle on lower dashed line) (10mg/kg; ip) twice weekly for six consecutive weeks. Data are presented as mean tumor volume (error bars represent 95% confidence interval, n > 5 mice/group). Treatment commenced (indicated by arrows) when tumors were 4mm in any one dimension (A), and when tumors reached 200mm³ (n > 6 mice/group) (B). C) Kaplan–Meier survival plot shows survival of mice from study of larger tumors shown in (B). D) Tumors from treated mice in (A) were analyzed by immunoblotting for indicated targets (combination therapy treated xenografts were eradicated, hence were unavailable for analysis). Actin immunoblot shows equal protein input. E) Histograms quantifying changes in protein expression levels from (D) (β-actin corrected). *P = .05, **P = .01, ***P < .001 (relative to IgG, or to treatment indicated by corresponding lines to the side of growth curves and above histograms). For tumor xenograft models, individual growth curves were plotted and a linear mixed model was used to test for differences between treatments. It was fitted by maximum likelihood using the nlme package in the statistical software R (R Development Core Team, 2010) 2.11.1. P values are from Wald tests. Survival of mice was measured using the Log-Rank test. All tests were two sided. HER2 = Human Epidermal Growth Factor Receptor 2; IgG = immunoglobulin; TRA = trastuzumab.