CSPG4 protein as a new target for the antibody-based immunotherapy of triple-negative breast cancer

Abstract

Background: The cell surface proteoglycan, chondroitin sulfate proteoglycan 4 (CSPG4), is a potential target for monoclonal antibody (mAb)-based immunotherapy for many types of cancer. The lack of effective therapy for triple-negative breast cancer (TNBC) prompted us to examine whether CSPG4 is expressed in TNBC and can be targeted with CSPG4-specific mAb.

Methods: CSPG4 protein expression was assessed in 44 primary TNBC lesions, in TNBC cell lines HS578T, MDA-MB-231, MDA-MB-435, and SUM149, and in tumor cells in pleural effusions from 12 metastatic breast cancer patients. The effect of CSPG4-specific mAb 225.28 on growth, adhesion, and migration of TNBC cells was tested in vitro. The ability of mAb 225.28 to induce regression of tumor metastases (n = 7 mice) and to inhibit spontaneous metastasis and tumor recurrence (n = 12 mice per group) was tested in breast cancer models in mice. The mechanisms responsible for the antitumor effect of mAb 225.28 were also investigated in the cell lines and in the mouse models. All statistical tests were two-sided.

Results: CSPG4 protein was preferentially expressed in 32 of the 44 (72.7%) primary TNBC lesions tested, in TNBC cell lines, and in tumor cells in pleural effusions from 12 metastatic breast cancer patients. CSPG4-specific mAb 225.28 statistically significantly inhibited growth, adhesion, and migration of TNBC cells in vitro. mAb 225.28 induced 73.1% regression of tumor metastasis in a TNBC cell-derived experimental lung metastasis model (mAb 225.28 vs control, mean area of metastatic nodules = 44590.8 vs 165950.8 μm(2); difference of mean = 121360.0 μm(2), 95% confidence interval = 91010.7 to 151709.4 μm(2); P < .001). Additionally, mAb 225.28 statistically significantly reduced spontaneous lung metastases and tumor recurrences in an orthotopic xenograft mouse model. The mechanisms responsible for antitumor effect included increased apoptosis and reduced mitotic activity in tumor cells, decreased blood vessel density in the tumor microenvironment, and reduced activation of signaling pathways involved in cell survival, proliferation and metastasis.

Conclusions: This study identified CSPG4 as a new target for TNBC. The antitumor activity of CSPG4-specific mAb was mediated by multiple mechanisms, including the inhibition of signaling pathways crucial for TNBC cell survival, proliferation, and metastasis.

Figure 1

Analysis of chondroitin sulfate proteoglycan 4 (CSPG4) expression in breast cancer subtypes. A) Analysis of CSPG4 mRNA expression levels in published clinical microarray datasets (GSE5460). Closed circles represent individual patients, and horizontal bars denote the mean expression. CSPG4 mRNA expression level is statistically significantly higher in estrogen receptor–negative (ER^–)/HER2^– breast cancers than in ER^–/HER2⁺, ER⁺/HER2^–, and ER⁺/HER2⁺ breast cancers. The P value determined by the Mann–Whitney U test (ER^–/HER2^– vs ER⁺/HER2^–, P = .01) surpassed a Bonferroni threshold of .0167. B) Analysis of CSPG4 protein expression in triple-negative breast cancer (TNBC) lesions. Representative images of immunohistochemical (IHC) staining of TNBC lesions (arrow marked) stained with the CSPG4-specific monoclonal antibody (mAb) D2.8.5-C4B8 are shown. The staining was graded as − (negative), when no staining was detected; + (weak), when the staining was weak; ++ (moderate), when the staining was homogeneous; +++ (strong), when the staining was strong and homogeneous. The IHC staining was performed once. Magnification ×200. Scale bar = 100 μm. C) Analysis of CSPG4 protein expression and frequency of CD44^hi/CD24^lo/− cells in TNBC cell lines HS578T, MDA-MB-231, MDA-MB-435, and SUM149 and in luminal cell lines MCF7, SK-BR-3, and T47D. Cells were treated either with CSPG4-specific mAb 225.28 and rabbit anti-mouse IgG-PE antibodies or with mouse anti-human CD24-FITC mAb, mouse anti-human CD44-allophycocyanin mAb, and 7-aminoactinomycin D. Stained cells were subjected to flow cytometry analysis. The percentage of cells stained with the CSPG4-specific mAb 225.28 and the mean fluorescence intensity (MFI) are shown in each histogram (% positive/MFI). D) Enrichment of CSC subpopulation in CSPG4⁺ cell population in pleural effusions from patients with breast carcinoma. After exclusion of dead cells, CD45^–/CSPG4⁺ cells stained with CSPG4-specific mAb 225.28 (top panel) or mAb 763.74 (bottom panel) were analyzed for CD44 and CD24 expression. The percentages of CD44^hi/CD24^lo/− cells found in four patient samples (PC-P12, PC-P13, PC-P14, and PC-P15) are indicated in the bottom right quadrant of the dot plots.

Figure 2

Effect of chondroitin sulfate proteoglycan 4 (CSPG4)–specific monoclonal antibody (mAb) 225.28 on triple-negative breast cancer (TNBC) cell growth, adhesion, and transmigration in vitro. A) Image analysis of TNBC cell growth in the presence of mAb 225.28. MDA-MB-231 cells were incubated with mAb 225.28, the control mAb F3-C25, or phosphate-buffered saline (PBS), at 37°C for 6 days in a 96-well filled with Matrigel. Representative images show the clusters of cells. Magnification ×100. Scale bar =100 μm. B) Quantitative analysis of TNBC cell growth. The viable cells that were not stained by trypan blue were counted. The results are expressed as percent inhibition of cell growth using the values obtained in the PBS-treated cells as a 100% reference. The means and 95% confidence intervals from three independent experiments are shown. ***P < .001. C) Quantitative analysis of TNBC cell adhesion to fibronectin. MDA-MB-435 cells were incubated at 37°C for 40 minutes with mAb 225.28, mAb F3-C25, or PBS in an adhesion assay. Adherent cells were quantitated by measuring the optical density at 540 nm. The results are expressed as percent inhibition of adhesion using the values obtained from PBS-treated cells as a 100% reference. The means and 95% confidence intervals from three independent experiments are shown. ***P < .001. D) Quantitative analysis of TNBC cell migration. MDA-MB-231 cells were incubated for 48 hours at 37°C with mAb 225.28, mAb F3-C25, or PBS in a transmigration assay. Cells that migrated toward fibronectin were counted. The results are expressed as percent inhibition of migration using the values obtained from PBS-treated cells as a 100% reference. The means and 95% confidence intervals from three independent experiments are shown. ***P < .001. All P values were calculated using the two-sided Student t test.

Figure 3

Effect of chondroitin sulfate proteoglycan 4 (CSPG4)–specific monoclonal antibody (mAb) 225.28 on signaling pathways important for cell growth, adhesion, and migration in vitro. MDA-MB-231 and MDA-MB-435 cells were treated with CSPG4-specific mAb 225.28, the control mAb F3-C25, or phosphate-buffered saline (PBS) at 37°C for 48 hours. Representative immunoblot analysis of p-FAK (Tyr397), FAK, p-ERK1/2 (Thr202/Tyr204), ERK1/2, p-AKT (Ser473), AKT, and PKCα is shown. HLA class I heavy chain was used as the loading control. The experiment was performed three independent times. p = phosphorylated.

Figure 4

Effect of chondroitin sulfate proteoglycan 4 (CSPG4)–specific monoclonal antibody (mAb) 225.28 on the metastasis of triple-negative breast cancer cells in an experimental lung metastasis model. A) Analysis of MDA-MB-231 cell metastasis. Mice were injected intravenously with MDA-MB-231 cells (1 × 10⁶ cells in 200 μL phosphate-buffered saline [PBS] per mouse) and were randomized into two groups on day 3. One group (n = 8) was injected intravenously twice weekly with mAb 225.28, and the other group (n = 7) was injected with the control mAb F3-C25. Lungs were analyzed for metastatic nodules on day 79. Means and 95% confidence intervals from two independent experiments are shown. **P < .01. B) Analysis of MDA-MB-435 cell metastasis. On day 3, mice were injected intravenously with MDA-MB-435 cells (1 × 10⁶ in 200 μL PBS per mouse) and randomized into two groups. One group of mice (n = 5) was injected intravenously twice weekly with mAb 225.28, and the other group of mice (n = 5) was injected with the control mAb F3-C25. Lungs were analyzed for metastatic nodules on day 34. Means and 95% confidence intervals from two independent experiments are shown. **P < .01. C) Analysis of the effect of CSPG4-specific mAb 225.28 on established experimental lung metastases. Mice were injected intravenously with MDA-MB-231 cells (1 × 10⁶ in 200 μL PBS per mouse) and randomized into two groups. On day 20, one group of mice (n = 7) was injected intravenously with mAb 225.28 every 48 hours, three times, and the other group of mice (n = 7) was injected with mAb F3-C25 using the same schedule and dose as mAb 225.28. Lungs were collected on day 25 and formalin-fixed paraffin-embedded lung tissues were stained with hematoxylin and eosin. Representative images are shown. Magnification ×200. Scale bar = 100 μm. D) Quantitative analysis of tumor areas shown in (C). The area of metastatic nodules in each section was quantified from five randomly selected high-power (magnification ×200) fields. Means and 95% confidence intervals from two independent experiments are shown. ***P < .001. E) Analysis of apoptotic tumor cells in lung metastasis from mice treated with the CSPG4-specific mAb 225.28. Representative images are shown. Magnification ×200. Scale bar =100 μm. F) Quantitative analysis of the apoptotic cells shown in (E) from 10 randomly selected high-power (magnification ×200) fields per section. Means and 95% confidence intervals from two independent experiments are shown. **P < .01. G) Analysis of mitotic cells in metastatic tumors in lung tissue sections from mice treated with the CSPG4-specific mAb 225.28. Mitotic cells were detected by staining p-Histone H3 (Ser10) protein. Representative images are shown. Magnification ×200. Scale bar = 100 μm. H) Quantitative analysis of mitotic cells shown in (G) from 10 randomly selected high-power (magnification ×200) fields per section. Means and 95% confidence intervals from two independent experiments are shown. ** P < .01. All P values were calculated using the two-sided Student t test.

Figure 5

Effect of chondroitin sulfate proteoglycan 4 (CSPG4)–specific monoclonal antibody (mAb) 225.28 on spontaneous metastases of triple-negative breast cancer cells and tumor recurrence in the postsurgery recurrence and metastasis model. MDA-MB-435 cells (1.0 × 10⁶ in 50 μL phosphate-buffered saline per mouse) were injected into the mammary fat pad of female severe combined immunodeficient mice (n = 24). On day 7, mice were randomized into two groups using a stratified randomization strategy. Starting on day 7, one group (n = 12) was injected intraperitoneally with mAb 225.28 twice weekly for a total of 18 injections, and the other group (n = 12) was injected with the control mAb F3-C25. On day 79, all tumors were surgically removed. The treatment with mAb was continued with nine additional injections. A) On day 131, lungs were collected for quantitative analysis of metastatic nodules. Means and 95% confidence intervals from three independent experiments are shown. *P < .05. B) Mice were monitored twice weekly for local tumor recurrence. Results are expressed as the percent of mice treated with mAb 225.28 and control mAb F3-C25 that developed local tumor recurrence. **P = .01. C) Analysis of the effect of CSPG4-specific mAb 225.28 on the density of blood vessels in surgically removed primary tumors. Blood vessels in MDA-MB-435 cell–derived primary tumors were detected by staining with CD31-specific mAb. Representative images are shown. Magnification ×400. Scale bar = 100 μm. D) Quantitative analysis of the blood vessels shown in (C) by counting five randomly selected high-power (magnification ×400) fields per section. Means and 95% confidence intervals from three independent experiments are shown. *P < .05. All P values were calculated using the two-sided Student t test.

Figure 6

Effect of chondroitin sulfate proteoglycan 4 (CSPG4)–specific monoclonal antibody (mAb) 225.28 on signaling pathways important for cell growth, adhesion, and migration in vivo. MDA-MB-435 cells (1.0 × 10⁶ in 50 μL phosphate-buffered saline per mouse) were injected into the mammary fat pad of female SCID mice. A) On day 79, orthotopic primary tumors that originated from MDA-MB-435 cells were surgically removed from mice (n = 5 per group) treated with mAb 225.28 and control mAb F3-C25. Cell lysates were analyzed by immunoblot assay, and representative immunoblot of two independent experiments shows the levels of p-FAK (Tyr397), FAK, p-ERK1/2 (Thr202/Tyr204), ERK1/2, p-AKT (Ser473), AKT, and PKCα. HLA class I heavy chain was used as the loading control. p = phosphorylated. B) Quantitative densitometric analysis of the proteins shows the ratios of the density of each protein to the reference HLA class I heavy chain. Means and 95% confidence intervals from two independent experiments are shown. All P values were calculated using the two-sided Student t test.