Splice of Life for Cancer: Missplicing of PPP2R5A by Mutant SF3B1 Leads to MYC Stabilization and Tumorigenesis

Mutations in RNA-splicing factors have been identified in multiple cancers and are most common in hematologic malignancies. These mutations most commonly occur in SF3B1; but are also found in other splicing factors including SRSF2 and U2AF1, and mutations found in these genes are typically mutually exclusive of each other (1). Mutations in SF3B1 have been shown to be early driver events, in both hematologic malignancies and solid tumors (1). Because these mutations result in numerous missplicing events, parsing out which splicing events drive tumorigenesis has remained challenging.

In this issue of Cancer Discovery, Liu and colleagues present a multidimensional approach to determine the functional consequences of SF3B1 mutations, with the goal of identifying therapeutically targetable dependencies in SF3B1 mutant–expressing cells (2). The authors first examine RNA-sequencing data from roughly 200 wild-type and SF3B1-mutant tumors and cell lines from multiple lineages. Analysis of the aberrant 3′ss sites reveals that there were differences between the missplicing driven by each distinct recurrent SF3B1 point mutation, and each SF3B1 mutant could have distinct downstream effects on splicing based on histologic context. In addition, the SF3B1-mutant 3′ss sites were distinct from the dysregulated alternative splicing resulting from SRSF2 and U2AF1 mutants, indicating that each of these splicing factors have nonoverlapping and distinguishable effects on the alternative splicing program. Further exploration into determining how different hotspot mutations in SF3B1 result in distinct RNA splicing and gene expression profiles may give insight into the tissue specificity of SF3B1 mutations in cancer and normal SF3B1 function.

Next, to determine the downstream signaling consequences, gene expression analysis was performed and revealed an enrichment of the MYC transcriptional program, which was specific to SF3B1-mutant chronic lymphoid leukemia. Liu and colleagues went on to show that SF3B1 mutations indeed enhance MYC-driven tumorigenesis in vivo. Interestingly, the enhanced MYC transcriptional program was not a result of changes in MYC mRNA expression or genetic amplification of the MYC locus in the SF3B1-mutant tumors, indicating either a post-transcriptional or post-translational mechanism for the observed changes in MYC signaling. Analysis of genes which have been shown to regulate MYC stabilization revealed that only PPP2R5A, the gene encoding the B56α regulatory subunit of the protein phosphatase 2A (PP2A), was aberrantly spliced in SF3B1-mutant cells. The authors went on to show that aberrant splicing of PPP2R5A resulted in the degradation of the PPP2R5A transcript by nonsense-mediated decay (NMD), resulting in decreased mRNA and protein expression (Fig. 1).

PP2A is a heterotrimeric enzyme family encoded by a scaffolding A subunit, catalytic C subunit, and one of many substrate-determining regulatory B subunits. When all subunit combinations are considered, the PP2A family is made up of more than 100 structurally distinct phosphatases which can be active within a cell at any given time. PP2A is thus not a single entity, but a complex system of more than a dozen subunits that assemble into specific heterotrimers in a highly ordered and regulated process involving specific chaperones and post-translational modifications that collectively regulate much of the serine–threonine phosphoproteome (3). Given its central role in regulating cellular signaling and homeostasis, PP2A inactivation, through a plethora of mechanisms, is a common event in human disease. The oncogene MYC is frequently amplified or overexpressed in many cancer types, independent of histologic subtype, and increased MYC expression is correlated with both more aggressive disease and resistance to standard-of-care treatments. Moreover, studies have demonstrated that MYC drives transformation, tumor growth, and metastasis in multiple cancer types. Interestingly, emerging data suggest that post-translational regulation of MYC stability is an important mechanism driving its aberrant activation in a broad range of cancers. Mechanistically, MYC stability is regulated by a series of highly ordered phosphorylation and dephosphorylation steps at specific residues, Thr-58 and Ser-62 (4). Phosphorylation at these sites is regulated by the balance of kinase activity (cyclin-dependent kinases and ERK) and phosphatase function (PP2A). Specifically, the PP2A B56α holoenzyme is responsible for the dephosphorylation of Ser-62, leaving the singly phosphorylated Thr-58 site, which marks MYC for ubiquitination and degradation (5). The authors show that SF3B1-mutant cells expressed higher amounts of phosphorylated MYC at Ser-62 than their wild-type counterparts, and that restoration of B56α into the mutant cells normalized the phosphorylated Ser-62 levels and resulted in MYC degradation.

Interestingly, PP2A B56α holoenzyme has also been shown to regulate MYC through phosphorylation of BCL2 at Ser-70, and this phosphorylation event of BCL2 is an important regulator of MYC-mediated apoptosis (6). Liu and colleagues show that the SF3B1-mutant cells had higher levels of phosphorylated BCL2 at Ser-70, and that these cells were more generally resistant to apoptosis than the SF3B1 wild-type cells. To mechanistically test the dependency of the effects of SF3B1 mutations on MYC and BCL2 phosphorylation, the authors use a series of elegant experiments utilizing phosphomimetic proteins and unphosphorylatable proteins.

Finally, the authors demonstrated that the aberrant splicing of PPP2R5A was directly linked to MYC activity by restoring B56α expression in their SF3B1- and double-mutant tumors, which resulted in decreased MYC protein levels and prolonged survival in disease-relevant mouse models. The mechanistic insight into the dependence of PPP2R5A and MYC activity could also be exploited therapeutically, and supportive of this the authors found that SF3B1-mutant cells are preferentially sensitive to PP2A activation. Multiple approaches to target PP2A therapeutically have been described, and the authors employ multiple strategies, including FTY-720, perphenazine, and a small-molecule PP2A modulator, DT-061, to preclinically translate the therapeutic tractability of this approach (7–9). In this series of experiments, the authors show that SF3B1-mutant cells and tumors are highly susceptible to PP2A activation therapies, resulting in decreased MYC expression and subsequent apoptosis. These therapeutic approaches were efficacious because there was not a complete loss of B56, just a reduction. These compounds allow for the remaining PP2A pool to be activated, thus activating the retained PP2A B56α heterotrimers. The clinical implications of this research are exciting and definitely warrant further study.

Beyond the novel insights demonstrated in this article, this research highlights a new mechanism of PP2A inactivation in cancer: aberrant splicing. Until this point, although alternatively spliced isoforms have been reported or described, the functional consequences of PP2A subunit splice variants have not been explored. Further studies analyzing whether this represents a more general nongenetic mechanism of PP2A inactivation in human cancer will clearly be an important area of research inquiry. Although numerous mechanisms of PP2A inactivation in human disease have been reported on, dysregulated alternative splicing of specific tumor-suppressive PP2A regulatory subunits has yet to be studied.

In summary, the elegant study by Liu and colleagues provides novel insight into the mechanism of how SF3B1-mutant tumors function as a biological driver of human cancers. Their combinatorial approach revealed that missplicing of PPP2R5A by mutant SF3B1 caused the stabilization and activation of oncogenic MYC, which could be reversed through genetic or pharmacologic restoration of PP2A activity (Fig. 1). More broadly, this work provides the mechanistic and biological foundation for the development of targeted therapeutics for the treatment of SF3B1-driven diseases.

C.M. O'Connor is a consultant at RAPPTA Therapeutics. G. Narla is chief scientific officer at RAPPTA Therapeutics, is an SAB member at Hera BioLabs, reports receiving commercial research support from RAPPTA Therapeutics, and has ownership interest (including patents) in RAPPTA Therapeutics. No other potential conflicts of interest were disclosed.