Abstract P3-07-53: Chromosomal instability as a predictor of sensitivity to paclitaxel

Abstract

Abstract Background: Paclitaxel is one of the most effective therapies for breast cancer, although many patients do not benefit. Our goal is to identify those who will benefit, by understanding how this drug contributes to chromosomal instability (CIN). CIN is the gradual gain/loss of whole chromosomes that can occur with mitotic errors as tumors proliferate. Some breast cancers inherently have CIN whereas others lack CIN. Previous work suggests low rates of CIN can promote tumor growth by creating genetic diversity. By contrast, high rates of CIN are lethal, apparently due to a high incidence of deleterious karyotypes. We hypothesize that paclitaxel operates by increasing CIN, and that this has preferential anticancer effects in tumors with preexisting low CIN. Methods: To assess rates of underlying CIN in human breast cancer, we performed 6-chromosome FISH on 354 human breast cancers and correlated with outcomes on a cohort with median 8.4 year follow-up. We measured the physiologic levels of paclitaxel that occur in human breast tumors. To do this, we treated 5 women with neoadjuvant paclitaxel 175mg/m2, performed tumor biopsy at 20 hours, performed LC to quantify intratumoral levels and analyzed mitotic spindles by IHC. Additionally, we performed timelapse videomicroscopy to analyze mitosis in fluorescently-labeled breast cancer cells in the laboratory after exposure to these levels. To evaluate whether CIN controls paclitaxel sensitivity we artificially introduced low levels of CIN into breast cancer cell lines by doxycycline-inducible expression of GFP-Mad1, a protein involved in the mitotic checkpoint, and tested whether this enhanced sensitivity to physiologic doses of paclitaxel. Results: A total of 77% (270/349) of breast cancer have detectable underlying CIN, (average percentage of non-modal chromosomes averaged for 6 chromosomes) greater than the normals (n=11). CIN is higher in HER2+ and TNBC subtypes compared to HR+. CIN does not correlate with the proliferation marker, Ki67 (r2 = 0.04), which does not support the idea of a growth advantage. CIN greater than median levels correlated with worse breast cancer-specific survival (p=0.022 log rank), but no difference in OS or RFS. Paclitaxel in human breast cancer reaches a level mimicked by 5-50nM exposure in laboratory experiments. In the laboratory, breast cancer cells exposed to these levels exhibit multipolar divisions, and similar abnormal mitoses can be found in patient tumors. In breast cancer cells lacking CIN, chromosome analysis demonstrates that it can be artificially induced by conditionally expressing GFP-Mad1. Inducing GFP-Mad1 expression increases sensitivity to paclitaxel, demonstrating that CIN enhances taxane sensitivity. Conclusions: These data support the idea that excessively high levels of CIN can be lethal to cancer cells and that paclitaxel enhances CIN. We predict that the anticancer effects of paclitaxel are marked in tumors with intrinsic CIN, as the enhanced levels are lethal. Thus CIN may be an effective biomarker to predict which women will benefit from taxane therapy. Ultimately, this could be applied in the clinic to substantially improve patient care by decreasing primary resistance or by reducing side effects associated with paclitaxel use. Citation Format: Cavalcante LL, Denu R, Zasadil L, Weaver BA, Burkard M. Chromosomal instability as a predictor of sensitivity to paclitaxel. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P3-07-53.