Mutation Burden in CDKN2C, CDK1 and E2F2 Is Associated with Differential Response to Targeting CDK4/CDK6 in Combination with Bortezomib in Mantle Cell Lymphoma,

Abstract 3738 Dysregulation of the cell cycle is a hallmark of mantle cell lymphoma (MCL) in which cyclin D1 expression is constitutive due to the t (11:14) translocation and CDK4 levels are elevated. MCL remains incurable despite initial response to therapy. Our goal was to develop a mechanism- and genome-based therapy to both inhibit lymphoma cell proliferation and sensitize them for cytotoxic killing. We have recently developed such a regimen by inhibition of CDK4/CDK6 with PD 0332991 (PD), the only known selective inhibitor of CDK4 and CDK6 that is also potent, reversible and orally bioavailable, in combination with cytotoxic agents. We demonstrate, for the first time, that 1) inhibition of CDK4/CDK6 with PD leads to early G1 arrest; 2) upon release of the G1 block, synchronous cell cycle progression to S phase occurs, and 3) S phase synchronization following prolonged early G1 arrest (pG1-S) sensitizes MCL cells to killing by diverse clinically relevant cytotoxic agents at reduced doses, including proteasome inhibitors bortezomib and carfilzomib, and the nucleoside analog cytarabine, in vitro and in a mouse model of MCL (Huang et al, submitted). In a completed phase I clinical study in MCL, PD potently and preferentially inhibited CDK4/CDK6 in lymphoma cells despite extensive chromosomal abnormalities, with an excellent toxicity profile and promising clinical response (Leonard et al, submitted). To advance targeting CDK4/CDK6 in MCL, we have now combined PD with escalating dose of bortezomib in an ongoing phase I clinical study (PD-B) in MCL. In this proof-of-concept study, PD is administered on days 1–12 of a 21-day cycle; bortezomib is administered first in prolonged G1 arrest concurrent with PD on days 8 and 11, and again after PD withdrawal in pG1-S on days 15 and 18. CD19+ MCL tumor cells were isolated at baseline, on day 8 and day 21 for analysis. To elucidate the mechanisms that underlie the progression of MCL and the differential response to this novel, cell-sensitizing therapy, we preformed 50×50 paired-end RNA-Sequencing on a HiSeq2000, using one lane for each sample of clinically responding and non-responding patients enrolled in this clinical trial. We generated an average of 76 million reads for each sample, then used the Burrow-Wheeler Aligner (BWA) to align the reads to the genome (Build 37), and SAMtools and the Genome Analysis Toolkit (GATK) to call non-reference variants. We focused on examining genes in the cell cycle and apoptotic pathways, and our data show 400 mutations in 16 genes including CDKN2C (p18), CDK1, E2F2, BBC3 (PUMA), BCL2L11(BIM), JUN and TP53, which are specific to each patient and whose expression changes dynamically during treatment. Moreover, we observe that the overall mutation burden is higher in a non-responding patient relative to the responding patient, and that certain genes (CDKN2C, CDK1, E2F2) show a highly significant (p=2.2×10–16) enrichment of mutations at baseline in the non-responder. By inhibiting CDK4/CDK6, p18 (CDKN2C) is essential for homeostatic cell cycle control of B cell activation and plasma cell differentiation in immunity. Conversely, mutations and deletions of CDKN2C are frequent in MCL, suggesting that loss of CDKN2C contributes to cell cycle dysregulation in this disease. Our RNA-Seq data reveal specific mutations in CDKN2C that are associated with compromised clinical response to PD, in line with cooperative inhibition of CDK4/CDK6 by p18 and PD in BCR-activated B cells as we reported previously. Gene expression profiling and quantitative RNA and protein analyses further demonstrate that induction of prolonged G1 arrest by inhibition of CDK4/CDK6 with PD halts gene expression in early G1 and depletes the expression of those programmed for other phases of the cell cycle. This leads to a metabolic imbalance, which is not restored in pG1-S, thereby sensitizing MCL cells to cytotoxic killing. Mutations in E2F2, which promotes G1/S transition, and CDK1, which functions in G2/M, may therefore antagonize cell cycle sensitization to cytotoxic killing by CDK4/CDK6 inhibition. These data provide new mechanistic insight into therapeutic targeting of CDK4/CDK6 in MCL, and suggest novel molecular targets for personalizing and advancing cell cycle-based therapy in MCL. Disclosures: Martin: Millennium Pharmaceuticals, Inc.: Research Funding, Speakers Bureau. Leonard:Pfizer, Inc: Consultancy; Millenium: Consultancy; Johnson and Johnson: Consultancy; Onyx: Consultancy. Chen-Kiang:Pfizer, Inc.: Research Funding.