Dual Effects of BCOR-PRC1.1 Dependent Gene Regulation Mediate Cooperation of BCOR and TET2 Mutations in Myeloid Transformation

Myeloid transformation occurs via stepwise acquisition of driver mutations that have distinct positions in the clonal hierarchy. BCOR mutations are common in myelodysplastic syndromes (MDS), but rarely occur as initiating events, instead typically arising in the context of other pre-existing mutations, such as TET2. BCOR encodes a subunit of the Polycomb Repressive Complex 1.1 (PRC1.1) which mediates gene repression by regulating H2AK119 via RING1 ubiquitin ligase activity. We hypothesized that BCOR mutations exert context dependent effects on chromatin regulation, thereby mediating conditional clonal selection during disease progression.

We used CRISPR/Cas9 genome editing to establish an isogenic human leukemia cell line model harboring BCOR, TET2, or concurrent BCOR and TET2 mutations. Whereas TET2 deficiency alone caused no effect on cell growth properties, isolated loss of BCOR caused a significant decrease in cell proliferation. Concomitant inactivation of TET2 and BCOR reversed the negative impact of BCOR mutations and resulted in enhanced proliferation suggesting that TET2 and BCOR deficiencies cooperate to drive a distinct cell growth phenotype that is not evident when either gene is knocked out alone.

To define the impact of BCOR deficiency on PRC1.1 complex assembly we performed PCGF1 co-immunoprecipitation followed by mass spectrometry (IP/MS) in cells with and without BCOR deficiency. BCOR mutations caused a marked decrease in PRC1.1 complex abundance without disrupting complex integrity. Even in the absence of BCOR, PCGF1 interactions with other PRC1 components (RING1A/B, SKP1, KDM2B, USP7) were maintained. By contrast, reciprocal BCOR IP/MS in cells with and without PCGF1 deletion showed that PCGF1 is required for RING1A/B and RYBP binding, but does not regulate complex abundance. These data suggest that BCOR is critical for maintaining PRC1.1 stability while PCGF1 selectively recruits core enzymatic components to the complex.

To evaluate the role of PRC1.1 enzymatic function in regulating target loci, we performed RNA-seq and ChIP-seq analysis in PCGF1 deficient cells that lack PRC1.1-RING1 interactions. 931 genes were significantly upregulated (FC>2, padj≤0.05) in PCGF1 deficient cells compared to control. In wild type cells, these genes were bound by PRC1 (RING1B) and PRC2 (SUZ12) and marked by high levels of repressive histone marks (H3K27me3, H2AK119ub). In PCGF1 deficient cells, RING1B binding and repressive marks were lost at these loci, while BCOR binding was maintained and the active H3K27ac mark significantly increased. These data suggest that selective loss of PRC1.1 enzymatic activity causes derepression of PRC1.1 target loci.

To determine whether derepression of PRC1.1 targets underlies the cooperative phenotypic effect of TET2 and BCOR inactivation, we compared the transcriptional and epigenetic state of BCOR/TET2 deficient cells to that of control, TET2, and BCOR deficient cells. Indeed, a subset of genes that were derepressed in PCGF1 deficient cells also showed increased expression, loss of repressive and gain of activating marks in BCOR-TET2 deficient cells, but not in cells with sole BCOR or TET2 deficiency. Derepressed genes were specifically enriched for involvement in developmental processes and cell signaling, including the leukemia-associated HOXA gene cluster. In contrast to PCGF1 deficient cells, BCOR deficient cells also displayed significant transcriptional downregulation of genes involved in metabolism and cell cycle progression, consistent with the negative effect of BCOR mutations on cell growth properties. In wild type cells, these BCOR-dependent loci had high levels of PRC1.1 (BCOR) binding, active histone marks, and high chromatin accessibility. Remarkably, expression of nearly all (590/622) BCOR-dependent genes was restored after introduction of concurrent TET2 deletion.

Our data suggest that the BCOR-PRC1.1 complex fulfills a dual epigenetic role via RING1-dependent regulation of gene repression at some target loci and RING1-independent maintenance of active transcription at other loci. Accordingly, inactivation of BCOR alone is sufficient to derepress key leukemia-associated targets, but incurs a metabolic liability that limits transformation potential. Concurrent inactivation of TET2 may therefore enable transformation by attenuating negative impacts of BCOR mutations on active transcription.