• breast cancer
• resistance to target therapy

Introduction: Loss of control of proliferation is a hallmark of cancer. In breast cancer, uncontrolled cell proliferation is made possible by different mechanisms, including deregulation of cyclin D1-cyclin dependent kinases 4 and 6 (CDK4/6)-retinoblastoma (Rb) pathway. Phosphorylation of the oncosuppressor Rb by active cyclin-CDK complexes leads to release of transcriptional factors of the E2F family and transcription of genes for cell cycle S-phase entry. Strategies to target the cyclin D1-CDK4/6-Rb pathway in patients with hormone receptor positive (HR+)/HER2 negative (HER2–) tumors have shown excellent results. Palbociclib is a highly selective CDK4/6 inhibitor which maintains Rb in a hypophosphorylated status leading to cell cycle blockade. Based on the results of the PALOMA trials, palbociclib has received accelerated approval by the Food and Drug Administration (FDA) for the treatment of HR+/HER2– advanced breast cancer in combination with endocrine therapy (ET). However, results of these clinical trials indicate that a proportion of patients do not receive clinical benefit and that acquired resistance invariably occurs. Clinically, the identification of biomarkers of sensitivity and resistance to palbociclib is of crucial importance. Moreover, clarifying the mechanisms of acquired resistance to palbociclib might help develop new therapeutic strategies in patients progressing after palbociclib.
Therefore our aims are:
1. To derive a gene expression signature of Rb loss-of-function (RBsig), to test its role in identifying breast cancer cell lines which are intrinsically resistant to palbociclib and to explore its prognostic role in well annotated datasets of patients with early breast cancer.
2. To investigate mechanisms of acquired resistance to palbociclib through the molecular analysis of a panel of estrogen receptor positive (ER+) breast cancer cell lines, including endocrine resistant cell lines, in which we induced acquired resistance to palbociclib.
Materials and Methods: We established a gene expression signature of Rb loss-of-function (RBsig) by identifying genes associated with E2F1 and E2F2 expression in breast cancers included in The Cancer Genome Atlas database. We assessed the prognostic role of the RBsig in the METABRIC dataset and in a comprehensive breast cancer gene expression meta-dataset and we analyzed whether RBsig would discriminate palbociclib sensitive and resistant breast cancer cell lines in a large RNA sequencing-based (RNAseq) dataset (EGAS00001000610).
To study the mechanisms of palbociclib acquired resistance we chronically exposed a panel of ER+ palbociclib-sensitive (PDS) breast cancer cell lines to increasing doses of palbociclib in order to develop palbociclib-resistant (PDR) derivatives. Cells were defined as PDR when growth rate was restored in the presence of the target concentration of 1 µM of palbociclib. PDR models were functionally characterized to confirm the development of palbociclib acquired resistance, and molecularly assessed by RT-PCR, western blot and gene expression profiles. In particular, to assess if the RBsig may have a role also in the acquired resistance to palbociclib, this was analyzed in all PDS and PDR gene expression data and confirmed by RT-PCR.
Results: The RBsig was associated with RB1 genetic status in the TCGA dataset in all tumors (p=<7e-32) and in luminal subtype (p<7e-11). The RBsig was prognostic in the METABRIC dataset, both in the discovery (HR=1.93, CI=1.5-2.4, p=1.4e-08) and in the validation set (HR=2.01, CI=1.6-2.5, p=1.3e-09). In the meta-dataset patients with ER+ early breast cancer expressing high RBsig had significantly worse recurrence free survival compared to those with low RBsig (HR = 2.37 [1.8 − 3.2] p= 1.87e−08 and HR = 2.62 [1.9 − 3.5] p= 8.6e−11, respectively). The RBsig was also able to identify palbociclib resistant and sensitive breast cancer cell lines (ROC AUC = 0,7778).
We were able to develop 7 ER+ PDR breast cancer models. PDR derivatives failed to undergo senescence during palbociclib treatment and were able to normally proliferate in the presence of palbociclib 1μM. PDR cells showed IC50 values significantly higher than their PDS counterparts (PDR IC50/ PDS IC50 ranged from 6 to 30; p<0.05).
By molecular analyses we observed that in PDS cells treated for 3 days with palbociclib the RBsig was generally downregulated, suggesting that in this models palbociclib is effectively blocking the cyclin D1-CDK4/6-Rb pathway. However, PDR models showed RBsig levels similar to those of the untreated parental cells, indicating that at the time of acquired resistance to palbociclib Rb pathway activity is restored despite the continued presence of the drug. Similar data were obtained by RT-PCR and western blot although heterogeneity was observed among different models.
Conclusions: Our data show that the RBsig is prognostic in breast cancer patients and is able to identify breast cancer cell lines intrinsically resistant to palbociclib. However, it is not able to identify cells with acquired resistance to palbociclib. Investigations of alternative pathways that might be implicated in palbociclib acquired resistance are ongoing.