Mantle cell lymphoma (MCL) is a subtype of B-cell lymphomas characterized by deregulation of cell cycle progression at the G1-S phase transition as a consequence of cyclin D1 overexpression [1, 2]. Palbociclib, a highly selective cyclin-dependent kinase (CDK) 4/6 inhibitor demonstrated single-agent activity in patients with relapsed and/or refractory (R/R) MCL. Venetoclax, a BCL2 inhibitor also demonstrated clinical activity in MCL, but the remissions were short calling for rational drug combinations [3,4,5,6]. We analyzed efficacy, mode of action, and predictive markers of susceptibility to palbociclib, and provide sound preclinical rationale for its combination with venetoclax.

Anti-proliferative and cytotoxic effect of palbociclib was analyzed using a panel of nine MCL cell lines (Supplemental Table 1) with promising anti-proliferative effect (median IC50 13.6 nM), however the cytotoxic effect was detected only after extremely high concentrations of palbociclib with LD50 above 10µM in all cell lines.

In vitro, palbociclib increased number of cells arrested in the G1 phase, and downregulated critical cell cycle regulators including RB1 protein, while expression of CDKs and BCL2 family proteins remained without significant changes (Supplemental Figs. 1, 2 A). Despite this, in vitro screen demonstrated synergistic efficacy between palbociclib and venetoclax (Supplemental Table 2). The synergy was subsequently confirmed in vivo using a panel of 4 patient-derived xenograft (PDX) models derived from patients with R/R MCL including ibrutinib resistant cases (Fig. 1, Supplemental Table 3). Similar observations were recently reported by Yuxuan Che et al. [7]. Except for moderate downregulation of MCL1, ex vivo protein expression analysis did not significantly differ from in vitro results. Co-immunoprecipitation analyses after palbociclib revealed increased amounts of BCL2 and BCL-XL bound to BIM ex vivo and in vitro (but only after prolonged cultivation). (Supplemental Fig. 2B, 3).

Effect of palbociclib and venetoclax on PDX models of R/R MCL. Legend: Synergistic effect of CDK4/6 inhibition with palbociclib and BCL2 blockage with venetoclax on a panel of 4 PDX models of R/R MCL. X-axis shows days from initiation of therapy. Y-axis shows calculated tumor volume of PDX tumors (means ± standard deviations)

Because the mode of action behind the synergy could not be explained by significant changes in the levels of free or bound BCL2 proteins, we speculated that cell cycle arrest induced by palbociclib might lead to complex changes that prime MCL cells to venetoclax-triggered apoptosis. Indeed, reactive oxygen species (ROS) and JC-1 analyses revealed increased mitochondrial ROS production and increased mitochondrial depolarization of MCL cells after exposure to palbociclib (Supplemental Fig. 4A, B). Enhanced mitochondrial proapoptotic priming by complex mechanisms independent of BCL2 proteins was confirmed by intracellular BH3 profiling (Supplemental Fig. 4C). Although the SeaHorse analysis did not reveal any specific impact of palbociclib on oxidative phosphorylation, both basal and maximal glycolytic activity was decreased in 6 out of 7 analyzed MCL cell lines (Supplemental Figs. 5 and 6). Using a genetically encoded AKT activity biosensor we demonstrated that AKT activity corresponded with the change of glycolytic activity- it was decreased in MINO and UPF1H cells but increased in JEKO-1 (Supplemental Fig. 7).

These data point to the fact that palbociclib leads to complex deregulation of key energy metabolic pathways and priming to mitochondrial apoptosis. However, precise molecular mechanisms that mediate the observed palbociclib-mediated priming of MCL cells to venetoclax remain to be elucidated.

MCL is characterized by several other recurrent molecular / cytogenetic lesions that deregulate G1-S cell cycle transition, including deletions of CDKN2A or RB1, and amplification of CDK4 or MYC genes. To evaluate the impact of these aberrations on palbociclib (and venetoclax) sensitivity, we derived MCL clones with transgenic (over)expression of p16INK4A, MYC, and CDK4, as well as MCL clones with knockout of CDKN2A and RB1 genes. We demonstrated that RB1 gene knockout leads to acquired resistance to palbociclib, which is in line with previous reports in breast cancer [8, 9]. Other tested aberrations did not significantly change sensitivity to palbociclib. Of note, overexpression of MYC was associated with increased sensitivity to venetoclax (Fig. 2, Supplemental Fig. 8).

Effect of palbociclib on clones with overexpression of MYC, CDK4, CDKN2A and K/O of CDKN2A and RB. Legend: the figure shows relative comparison of half-maximal inhibitory concentration (IC50, in nM) in response to palbociclib between MCL clones (IC50clone) and corresponding controls (IC50CTRL). The bars were constructed according to the following formulas: 1 + IC50CTRL / IC50clone in the cases when IC50clone > IC50CTRL; 1 - IC50clone / IC50CTRL in the cases when IC50CTRL > IC50clone. Positive and negative bars represent more sensitive and resistant clones, respectively (compared to controls)

To sum up, the innovative, chemotherapy-free combination of an FDA-approved CDK4/6 inhibitor palbociclib, and a BCL2 inhibitor venetoclax is synthetically lethal in vitro and in vivo on a panel of chemotherapy and ibrutinib-refractory MCL. Molecular mechanisms behind the observed synergy include palbociclib-triggered downregulation of anti-apoptotic MCL1, increased levels of proapoptotic BIM bound on both BCL2, and BCL-XL and increased pro-apoptotic priming mediated by BCL2-independent mechanisms, predominantly palbociclib-induced metabolic and mitochondrial stress. Deletion of RB1 represents a predictive marker of palbociclib resistance. In translation, our data strongly support investigation of palbociclib in combination with venetoclax as an innovative treatment strategy for R/R MCL patients without RB1 deletion.