Cilengitide sensitivity is predicted by overall integrin expression in breast cancer

Background: Treatment options for triple-negative breast cancer (TNBC) are limited and patients face a poor prognosis. Here, we sought to identify drugs that target TNBC vulnerabilities and understand the biology underlying these responses. We analyzed the Broad Institute DepMap to identify recurrent TNBC vulnerabilities and performed a 45-compound screen on vulnerability-related pathways on a set of up to 8 TNBC cell lines. We identified a subset of cell lines with an ITGAV vulnerability and a differential sensitivity to cilengitide, an integrin inhibitor targeting ITGAV:ITGB3 and ITGAV:ITGB5. Next, we sought to understand cilengitide resistance and response biomarkers. Clinical trials targeting integrins continue enrolling patients, necessitating an understanding of how these drugs affect tumors.

Methods: We combined in vitro assays with computational approaches to systematically explore the differential sensitivity to cilengitide and resistance mechanisms. We tested an additional pan-ITGAV inhibitor (GLPG0187) to determine how generalizable our findings on cilengitide sensitivity might be to integrin inhibition. ITGB4, ITGA3, and ITGA6 knockdown experiments assessed the importance of integrin monomers in cell attachment during cilengitide treatment. Additionally, we explored the role of extracellular matrix (ECM) proteins in cilengitide response by performing cell replating experiments and by culturing on collagen, fibronectin, or laminin coated plates.

Results: We discovered that cell-derived ECM modulates cilengitide sensitivity and exogenous fibronectin addition conferred resistance to all sensitive TNBC cell lines, though fibronectin expression did not correlate with sensitivity. Instead, elevated overall integrin protein levels, not specific integrins, in TNBC cells positively correlated with resistance. This suggested that high pan-integrin expression promotes cilengitide resistance. Thus, we tested cilengitide in six luminal breast cancer cell lines (which have low integrin levels); all were sensitive. Also, pan-ITGAV inhibitor, GLPG0187, showed the same sensitivity profile across our TNBC cell lines, suggesting our findings apply to other integrin inhibitors.

Conclusions: Integrin inhibitors are appealing candidates to pursue as anti-cancer drugs because they are generally well-tolerated, but their efficacy is mixed, possibly due to the absence of predictive markers. Cilengitide induces death in breast cancer cells with low integrin abundance, where complementary ECM promotes survival. Thus, integrin inhibition in breast cancer warrants further study.