In this study, we investigated the mechanisms by which PBRM1 functions as a tumor suppressor in clear cell renal cell carcinoma. PBRM1, also known as BAF180 or Polybromo, is a member of the PBAF SWI/SNF chromatin remodeling complex. Cancer sequencing studies have revealed that SWI/SNF components are widely mutated in cancer. PBRM1 is recurrently mutated in various human malignancies, but it has a particularly high mutation rate in clear cell renal cell carcinoma: ~40% of clear cell renal cell carcinomas have a PBRM1 mutation, making it the second most highly mutated gene in clear cell renal cell carcinoma behind VHL. Although many recent studies have looked at how other SWI/SNF components function in cancer control, relatively little is known about the tumor suppressive mechanisms of PBRM1 in clear cell renal cell carcinoma.
To investigate PBRM1 function, we manipulated its expression in clear cell renal cell carcinoma cell lines. In cell lines with intact PBRM1, we stably knocked down its expression using shRNA. In a cell line with mutant PBRM1, we stably restored expression of the wild-type protein. We found that PBRM1 deficiency significantly enhanced the growth properties of cells, but only when the cells were grown under stressful conditions, such as reduced serum or a 3-D culture environment. To investigate genes and pathways influenced by PBRM1 that may confer this growth advantage, we compared gene expression differences in the clear cell renal cell carcinoma cell lines and murine embryonic fibroblasts with or without PBRM1. We found that PBRM1 regulated numerous cancer-related genes and pathways.
One gene, ALDH1A1, was consistently upregulated with PBRM1 deficiency across our cell lines. Further expression analysis using two different clear cell renal cell carcinoma primary tumor datasets revealed that PBRM1 mutation in primary tumors was also associated with higher ALDH1A1 levels. ALDH1A1, or aldehyde dehydrogenase 1, is part of the retinoic acid metabolic pathway and irreversibly converts retinaldehyde to retinoic acid. It functions in hematopoietic stem cell development, white versus brown fat programming, and insulin signaling. Numerous studies have also identified ALDH1A1 as a marker of tumor-initiating cells, also known as cancer stem cells. Not much is known about the regulation of ALDH1A1 expression in cancer, and it has not previously been linked to PBRM1 or SWI/SNF. We confirmed that stable knockdown of PBRM1 in clear cell renal cell carcinoma cell lines resulted in higher ALDH1A1 mRNA and protein expression, and also higher ALDH1-class enzyme activity. Alternatively, re-expression of wild-type PBRM1, but not cancer-associated mutant PBRM1, lowered ALDH1A1 expression and activity in the PBRM1-mutant line. Additionally, inhibiting ALDH1A1 or knocking it down in the context of PBRM1 deficiency reduced anchorage-independent growth, while over-expressing ALDH1A1 in the PBRM1-normal setting increased tumorsphere-forming capacity. These results suggest that ALDH1A1 is not only a marker of tumor-initiating cells, but can also increase the tumorigenic potential of cells.
Based on our gene expression analysis, we additionally explored PBRM1 regulation of the EGFR and IFN pathways. PBRM1 decreased total EGFR protein levels and dampened downstream signaling. These changes had functional consequences, as PBRM1 deficiency led to faster growth in response to EGF stimulation. However, it did not create a setting of oncogenic addiction, as PBRM1 deficient cells were also more resistant to EGFR inhibition. Alternatively, PBRM1 deficiency reduced basal and IFNα-induced levels of IFI27, a pro-apoptotic interferon response gene, and made cells more resistant to growth inhibition by IFNα. PBRM1 mutations in cancer would thus be expected to have wide-ranging effects on a cell, and the targeting of any one specific downstream pathway might have limited efficacy.
Finally, we investigated the molecular mechanisms of how PBRM1 deficiency could alter transcription, keeping in mind that PBRM1 is one subunit of the larger PBAF complex. In our clear cell renal cell carcinoma cell lines, we found that mRNA and protein levels of another PBAF-specific subunit, ARID2, increased with PBRM1 deficiency. PBRM1 mutation in primary tumors was also associated with significantly higher ARID2 expression. Immunoprecipitation and glycerol gradient fractionation experiments suggested that more ARID2 may associate with the SWI/SNF components BRG1 and SNF5 after PBRM1 knockdown. ARID2 ChIP-seq analysis revealed that this remnant PBAF-like complex was bound to fewer locations in the genome, and its binding locations were broadly redistributed. Both gained and lost ARID2 binding were associated with differential gene expression, of both upregulated and downregulated genes, indicating that the genomic context influences whether PBAF-binding is activating or repressive. Interestingly, we also found that ARID2 was required for some of the pro-tumorigenic changes associated with PBRM1 deficiency, such as upregulation of ALDH1A1 and EGFR levels, but not others, such as decreased IFI27 levels, implying alternative modes of transcriptional regulation.
In total, this study implicates PBRM1 in the regulation of numerous cancer-related genes and pathways in clear cell renal cell carcinoma. PBRM1 mutation would alter the genomic binding of a residual PBAF-like complex containing ARID2, leading to transcriptional changes that promote tumor formation and growth. A better understanding of this oncogenic mechanism may reveal novel therapeutic opportunities.