Caitlin E. Mills

Abstract PD4-02: Advantageous polypharmacology of abemaciclib revealed by omics profiling of CDK4/6 inhibitors

Small molecule inhibitors of cyclin-dependent kinases (CDK) 4/6 have created new opportunities for the treatment of advanced hormone-receptor positive (HR+) breast cancer and show promise in other malignancies. Three CDK4/6 inhibitors, palbociclib (PD0332991; Ibrance), ribociclib (LEE011; Kisqali), and abemaciclib (LY2835219), which have been approved or are in late stage trials, are reported to be broadly similar although recent data suggest that abemaciclib has distinct single-agent activity in patients and a unique adverse effects profile. Differences in pharmacokinetics and relative potency for CDK4 versus CDK6 are postulated to account for these differences. In this paper, we use molecular and functional profiling by mRNA sequencing, mass spectrometry-based proteomics, and GR-based dose-response assays to obtain complementary views of the mechanisms of action of CDK4/6 inhibitors. We show that abemaciclib, but not ribociclib or palbociclib, is a potent inhibitor of kinases other than CDK4/6, including CDK1/Cyclin B, which appears to cause arrest in the G2 phase of the cell cycle, and CDK2/Cyclin E/A, which is implicated in resistance to palbociclib. Whereas ribociclib and palbociclib induce cytostasis, and cells adapt to these drugs within 2-3 days of exposure, abemaciclib induces cell death and durably blocks cell proliferation. Abemaciclib is active even in retinoblastoma protein (pRb)-deficient cells in which CDK4/6 inhibition by palbociclib or ribociclib is completely ineffective. The degree of polypharmacology of small molecule drugs is increasingly viewed as an important consideration in their design, with implications for efficacy, toxicity, and acquired resistance. In the case of CDK4/6 inhibitors, we propose that abemaciclib polypharmacology elicits unique molecular responses that are likely to be therapeutically advantageous. More generally, we propose that multi-omic approaches are required to fully elucidate the spectrum of targets relevant to drug action in tumor cells. We expect such understanding to assist in stratifying patient populations and ordering sequential therapies when resistance arises. Citation Format: Hafner M, Mills CE, Subramanian K, Chen C, Boswell SA, Everley RA, Juric D, Sorger PK. Advantageous polypharmacology of abemaciclib revealed by omics profiling of CDK4/6 inhibitors [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr PD4-02.

Abstract 203: Single cell imaging of kinase inhibitor-induced effects in breast cancer cell lines

There is need for improved targeted therapies for the treatment of breast cancer. Kinase inhibitors are candidates to be used in this context. With the goal of uncovering specific vulnerabilities in differing cell lines, we treated five breast cancer cell lines representing triple negative (BT20, MDAMB231, and Hs578T), hormone receptor positive (MCF7), and Her2 amplified (SKBR3) disease and the non-malignant MCF10A line with a panel of 105 kinase inhibitors covering a broad range of targets. Cells were treated with doses between 0.1 and 10 μM for 24 hours, at which time they were stained with DRAQ5 (DNA) and TMRE (mitochondrial membrane potential). Live-cell images were acquired using a high throughput, confocal Opera microscope. Cell segmentation, based on the DRAQ5 staining, and feature extraction (intensity, morphology, and texture) were performed with Acapella software. Over 300 features were extracted for ∼1.5 million cells. Analytical methods have been applied to identify those treatments that induced significant changes to the cells. Over half of the kinase inhibitors queried had a significant effect within 24 hours in all cell lines at 10 μM. Cell cycle inhibitors had the most common effects across cell lines. Cell line specific effects were also uncovered, for example, MDAMB231 cells were the most sensitive to MAPK inhibitors. These, early time point, results have been compared with the effects of the same inhibitors on cell growth to better understand the mechanisms leading to cell line and pathway specific effects. Citation Format: Caitlin Mills, David Andrews, Peter Sorger. Single cell imaging of kinase inhibitor-induced effects in breast cancer cell lines. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 203.

The relative biological effectiveness of low-dose mammography quality X rays in the human breast MCF-10A cell line.

Mammography is used to screen a large fraction of the population for breast cancer, and mammography quality X rays are speculated to be more damaging than the higher energy X rays used for other diagnostic procedures. The radiation dose delivered to breast cells as a result of these screening exposures may be a concern. The purpose of this current study was to determine the relative biological effectiveness (RBE) of low-energy mammography X rays for radiation-induced DNA double-strand breaks evaluated using a highly sensitive automated 53BP1 assay. Automation of the 53BP1 assay enabled the quantification and analysis of meaningful image-based features, including foci counting, within the cell nuclei. Nontumorigenic, human breast epithelial MCF-10A cells were irradiated in the low-dose range with approximately 3–30 mGy of 29 kVp mammography X rays or 137Cs (662 keV) gamma rays. The induction and resolution of the 53BP1 foci did not differ significantly between exposures to 137Cs gamma rays and 29 kVp X rays. The RBE was calculated to be 1.1 with a standard deviation of 0.2 for the initial number of radiation-induced double-strand breaks. The radiation dose from a single mammogram did not yield a significant change in the number of detectable foci. However, analysis of additional features revealed subtle differences in the distribution of 53BP1 throughout the nuclei after exposure to the different radiation qualities. A single mammogram was sufficient to alter the distribution of 53BP1 within the nuclear area, but not into discrete foci, while a dose-matched gamma exposure was not sufficient to alter the distribution of 53BP1. Our results indicate that exposure to clinically relevant doses of low-energy mammography quality X rays does not induce more DNA double-strand breaks than exposure to higher energy photons.

Therapeutically advantageous secondary targets of abemaciclib identified by multi-omics profiling of CDK4/6 inhibitors

The target profiles of many drugs are established early in their development and are not systematically revisited at the time of FDA approval. Thus, it is often unclear whether therapeutics with the same nominal targets but different chemical structures are functionally equivalent. In this paper we use five different phenotypic and biochemical assays to compare approved inhibitors of cyclin-dependent kinases 4/6 – collectively regarded as breakthroughs in the treatment of hormone receptor-positive breast cancer. We find that transcriptional, proteomic and phenotypic changes induced by palbociclib, ribociclib, and abemaciclib differ significantly; abemaciclib in particular has advantageous activities partially overlapping those of alvocidib, an older polyselective CDK inhibitor. In cells and mice, abemaciclib inhibits kinases other than CDK4/6 including CDK2/Cyclin A/E – implicated in resistance to CDK4/6 inhibition – and CDK1/Cyclin B. The multi-faceted experimental and computational approaches described here therefore uncover under-appreciated differences in CDK4/6 inhibitor activities with potential importance in treating human patients.Three inhibitors of the cyclin-dependent kinases CDK4/6, palbociclib, ribociclib, and abemaciclib, have emerged as highly promising therapies for the treatment of breast cancer and other solid tumors. These drugs are reported to have similar mechanisms of action although recent data suggest that abemaciclib exhibits distinct single-agent activity and toxicity. We compare their mechanisms of action using biochemical assays, mRNA profiling, mass spectrometry-based phospho-proteomics, and GR-based dose-response assays. We find that abemaciclib has activities not shared by palbociclib or ribociclib including: induction of cell death (even in pRb-deficient cells), arrest in the G2 phase of the cell cycle, and reduced drug adaptation. These activities appear to arise from inhibition of CDKs other than CDK4/6 including CDK2/Cyclin A/E and CDK1/Cyclin B. We propose that inhibition of these kinases by abemaciclib overcomes known mechanisms of resistance to CDK4/6 inhibition and may be therapeutically advantageous for patients whose tumors progress on palbociclib or ribociclib.

Abstract 5037: New concepts for quantifying the benefits of mono and combination therapy in an era of big data

I will describe new (unpublished) approaches to quantifying drug response at two points in the drug development pipeline: pre-clinical studies in cell lines and clinical trials of combination therapies in patient populations. Drug sensitivity and resistance in cell lines is conventionally quantified by IC50 or Emax values, but these measures suffer from a fundamental flaw when applied to growing cells: they are highly sensitive to cell division number, which varies with cell line, experimental condition, seeding density etc. The dependency of IC50 and Emax on division rate creates artefactual correlations between genotype and drug sensitivity while obscuring important biological insights and interfering with biomarker discovery. I will describe alternative growth rate inhibition (GR) metrics that are insensitive to division number and can directly measure both endpoint sensitivity and adaptive drug resistance. Theory and experiments show that GR50 and GRmax are superior to IC50 and Emax for assessing the effects of drugs in dividing cells. GR metrics promise to improve our ability to score drug sensitivity in specific-derived tumor cells, improves data reproducibility, and increase the translational potential of pharmacogenomics data. In patients, combination therapy improves tumor control compared to monotherapy and the development of combinations is motivated in most cases by pre-clinical data on synergism in cell lines. However I will describe a different way in which combinations can provide clinical benefit. Based on analysis of between-patient variability in existing trial data and a large set patient derived tumor xenograft (PDX) mice published by Gao H, Korn JM, Ferretti S, et al. (Nature medicine 2015;21:1318-25), I will argue for a simple principle: in nearly two-thirds of cases analyzed, efficacious combinations work simply by improving the likelihood that a tumor will experience an outlier response to a single drug. Thus, even in the absence of additive or synergistic tumor inhibition, combinations are generally superior to monotherapies. Superiority by “independent action” provides a principle on which to design new combinations and a scientific rationale for the use of combination therapies in poorly understood cancers whenever toxicity is acceptable.  These studies, at two different points of the drug discovery pipeline, illustrate the value of combining new first-principles theories about drug mechanism of action with “big data” that is increasingly available on pre-clinical and clinical drug response. Conversely, in the absence of conceptual innovation, we miss important mechanistic insights hidden in existing data. The purpose of such insights, when obtained, is drive new laboratory and real-world experiments. I will discuss how the process of coupling computation and experimentation works in practice. Citation Format: Peter K. Sorger, Marc Hafner, Mario Niepel, Caitlin Mills, Adam Palmer, Mohammed Fallahi Sichani. New concepts for quantifying the benefits of mono and combination therapy in an era of big data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 5037. doi:10.1158/1538-7445.AM2017-5037

Abstract P4-08-02: Integration of transcriptomic, proteomic and drug response data in triple negative breast cancer cell lines and PDX models

Drug response screens on panels of cell lines aimed at identifying markers of sensitivity or resistance have been limited in their successes. Unfortunately, the recent release of many such studies has been accompanied by concerns surrounding reproducibility. Since then, several publications have addressed these concerns by pointing out sources of variability and by suggesting better experimental methods as well as more robust analytical approaches. In the presented profiling effort, we integrated the latest advances in drug response measurement and focused on data diversity and quality rather than on breadth. We selected 32 breast cancer cell lines with a strong bias towards triple negative lines as well as 4 cell lines established from relevant patient-derived xenografts. We used high content microscopy to assay the phenotypic responses of the cell lines to a panel of 34 drugs made up largely of kinase inhibitors currently in the clinic along with some standard of care chemotherapeutics. The microscopy based drug response assay allowed us to measure drug potency, and to quantify the efficacy of the drugs in terms of growth inhibition and cell death. For the same cell lines, we used RNAseq to measure basal mRNA expression levels and shotgun mass spectrometry to measure endogenous protein levels. The completeness and controlled conditions under which these data sets were collected provide confidence in their integration. The complementarity of these multi-omics data has allowed us to address questions about the landscape of, particularly triple negative, breast cancer cell lines. Such questions include: where do the patient-derived lines lay among the established cell lines? and how different are the landscapes defined by drug response phenotypes, mRNA expression, and protein levels? We used network-based algorithms to identify eigenstates of signaling pathways related to genomic events, and further explored these states in the TCGA data. At the level of drug response, we have focused on important questions related to the clinical use of kinase inhibitors. In particular, we compared various CDK inhibitors in an effort to identify markers that are informative of response potency and efficacy. We have also looked at variability of the responses of the cell lines studied to multiple PI3K inhibitors that either target specific isoforms or all isoforms. Overall the data set that has been collected is a valuable resource for understanding drug response in triple negative breast cancer, and the transcriptomic and proteomic factors that influence it. Citation Format: Mills CE, Hafner MA, Sorger PK. Integration of transcriptomic, proteomic and drug response data in triple negative breast cancer cell lines and PDX models [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P4-08-02.