Abstract
Recent studies have identified a previously uncharacterized protein C11orf53 (now named POU2AF2/OCA-T1), which functions as a robust co-activator of POU2F3, the master transcription factor which is critical for both normal and neoplastic tuft cell identity and viability. Here, we demonstrate that POU2AF2 dictates opposing transcriptional regulation at distal enhance elements. Loss of POU2AF2 leads to an inhibition of active enhancer nearby genes, such as tuft cell identity genes, and a derepression of Polycomb-dependent poised enhancer nearby genes, which are critical for cell viability and differentiation. Mechanistically, depletion of POU2AF2 results in a global redistribution of the chromatin occupancy of the SWI/SNF complex, leading to a significant 3D genome structure change and a subsequent transcriptional reprogramming. Our genome-wide CRISPR screen further demonstrates that POU2AF2 depletion or SWI/SNF inhibition leads to a PTEN-dependent cell growth defect, highlighting a potential role of POU2AF2-SWI/SNF axis in small cell lung cancer (SCLC) pathogenesis. Additionally, pharmacological inhibition of SWI/SNF phenocopies POU2AF2 depletion in terms of gene expression alteration and cell viability decrease in SCLC-P subtype cells. Therefore, impeding POU2AF2-mediated transcriptional regulation represents a potential therapeutic approach for human SCLC therapy.
Introduction
Lung cancer remains a prominent contributor to global cancer morbidity and mortality worldwide1, with small-cell lung carcinoma (SCLC) representing a particularly aggressive and lethal subtype2,3,4. SCLC is characterized by its development within lung tissues and is frequently diagnosed at advanced stages with metastatic spread to other organs, accounting for roughly 13% of all lung cancer cases5. Regrettably, the patient survival rate for SCLC is among the lowest of all cancer subtypes, and the prognosis for patients afflicted with this condition is often bleak.
Human SCLC has been identified as a highly malignant tumor of primitive neuroendocrine (NE) cells with inactive mutations in Rb1 and Trp53 tumor suppressors6,7,8. However, recent studies on molecular subtyping further classified human SCLC based on the relative expression of four key lineage-specific transcription and co-transcription regulators: achaete-scute homolog 1 (ASCL1; SCLC-A subtype), neurogenic differentiation factor 1 (NEUROD1; SCLC-N subtype), yes-associated protein 1 (YAP1; SCLC-Y subtype), and POU class 2 homeobox 3 (POU2F3; SCLC-P subtype)9. Notably, the SCLC-Y subtype cells appeared to be enriched for intact RB by immunohistochemistry. Therefore, whether high YAP1 represents a transcriptional driver of this subtype, or a subtype-specific correlation has yet to be determined10.
Distinguishing between subtypes of lung cancer is of great importance, as the effectiveness of treatment could vary substantially among these subtypes. Recently, we and other colleagues utilized a global genome-wide CRISPR screening database (DepMap) and have identified and characterized human SCLC subtype-specific essential factors and signaling pathways within all SCLC subtypes. Intriguingly, we have identified C11orf53 gene, which is the top essential factor for SCLC-P subtype cells11,12,13. We then worked with HUGO gene nomenclature committee to rename this gene POU2AF2 (POU Class 2 Homeobox Associating Factor 2), based on its robust function as a co-activator of the master transcriptional regulator POU2F314. By utilizing multiple unbiased genome-wide studies, we discovered that more than 90% of POU2AF2 occupies distal enhancer elements and regulates the tuft cell-specific gene signature at their nearby super-enhancers and maintains the cell identity11,12,13,14. However, why loss of POU2AF2 triggers cell growth inhibition and cell death remains to be discovered.
In our current studies, we uncover an unexpected role of POU2AF2 in maintaining cell viability by regulating PTEN expression, and further elucidate the mechanism of how POU2AF2 interacts with the SWI/SNF chromatin remodeling complex at H3K27me3 enriched distal enhancer elements and facilitates the transcriptional repression function of Polycomb repressive complex 2 (PRC2).
Results
POU2AF2 elicits opposing effects of gene expression at distinct distal enhancer elements
Previously, we have characterized the chromatin occupancy of POU2AF2, which co-localizes with POU2F3 at the chromatin (Fig. 1a, Supplementary Fig. 1a) and functions as its co-activator across the genome11. Despite being a nuclear protein lacking a known canonical chromatin binding domain, POU2AF2 is shown to be recruited by the POU domain of POU2F311. In our ChIP-seq analysis, we have identified 6,987 overlapping POU2F3 and POU2AF2 peaks in the NCI-H526 SCLC cell line. Notably, approximately 93% of these peaks occupy enhancers that are characterized by high levels of H3K4me1 (Fig. 1a, Cluster 2 and 3). Interestingly, we have observed the subdivision of these enhancers into two clusters through k-means clustering: Cluster 2 peaks were enriched with the active enhancer histone mark H3K27ac, while Cluster 3 peaks were more enriched with the poised enhancer histone mark H3K27me3. This phenomenon was also observed in a different SCLC cell line, NCI-H211 (Supplementary Fig. 1b). To determine how POU2AF2 and POU2F3 regulate gene expression at distinct enhancer clusters, we first compared the gene expression profile between wild-type and POU2F3 or POU2AF2-depleted cells. The scatter plots showed a significant correlation between POU2F3 and POU2AF2 target genes in different cell lines (Fig. 1b, Supplementary Fig. 1c, d). It has been shown that depletion of either POU2AF2 or POU2F3 leads to the repression of genes involved in cell identity such as tuft-cell lineage genes12,14. Therefore, we further conducted pathway analysis with the up-regulated genes in POU2F3/POU2AF2-depleted cells. Interestingly, we found that most of these genes were enriched in cell differentiation or development pathways (Fig. 1c), implying there is a distinct role of POU2F3/POU2AF2 in transcriptional activation or repression. Next, we integrated the RNA-seq results with the ChIP-seq data from Fig. 1a and analyzed the expression change of genes nearest to POU2F3 and POU2AF2 overlapped peaks at the genome-wide level. As expected, loss of either POU2AF2 or POU2F3 reduced the expression of these genes (e.g., tuft cell-specific genes) nearest to Cluster 2 peaks, which were enriched with active enhancer histone marks H3K4me1 and H3K27ac (Fig. 1d, left and middle panel). This phenomenon is consistent with the effects of JQ1 treatment (Fig. 1d, right panel, Fig. 1e).
Surprisingly, we found that the up-regulated genes in POU2AF2 (with median RPM 13.78 (sgNONT), 15.42 (sgPOU2AF2−1), and 16.03 (sgPOU2AF2−2) or POU2F3 depleted cells were dramatically enriched at Cluster 3 peaks that were marked by repressive histone marks H3K27me3 (Fig. 1d, Supplementary Fig. 1e). Intriguingly, these genes were not upregulated in JQ1 treated cells, indicating that the activation of Cluster 3 nearby genes may not be the secondary effect of the repression of Cluster 2 nearby genes (Fig. 1d). To study the function of POU2F3 and POU2AF2 at enhancer regions marked by H3K27me3 and determine whether there is a direct binding of POU2F3/POU2AF2 at H3K27me3-marked enhancer chromatin, we conducted motif analysis and identified 35,503 enhancer regions marked by H3K4me1 and H3K27ac (Fig. 1f, left panel), and 6,538 enhancers marked by H3K4me1 and H3K27me3 in NCI-H526 cells (Fig. 1f, right panel). Consistent with our previous studies, POU2F3 DNA binding motif (5’-TATGCAAATC-3’)11,15 is the most enriched motif in both regions (Fig. 1f). Similar results were observed at specific gene loci (Fig. 1g, Supplementary Fig. 1f) in different SCLC cell lines. These results revealed a potential bidirectional transcriptional regulatory function of POU2F3/POU2AF2 at different enhancers.
POU2AF2 is essential for PRC2 maintenance and repression of Polycomb target genes
Previous studies have demonstrated that a subclass of enhancers enriched for H3K4me1 (but lacking H3K27ac) is also marked by H3K27me3 and bound by PRC2 in human or mouse pluripotent cells. These elements, referred to as “poised enhancers”, are located near key developmental genes16,17,18,19,20. Given the presence of H3K27me3 at enhancer regions (Cluster 3), we postulated that POU2AF2 may be directly or indirectly involved in transcriptional repression at these sites through a potential collaboration with EZH2, the dominant enzymatic catalytic subunit of PRC2 responsible for catalyzing H3K27me3 histone modification.
To investigate this hypothesis, we treated cells with EZH2-specific inhibitor GSK126 and observed a similar effect on transcriptional activation that phenocopies POU2AF2 depletion (Fig. 2a), without affecting the protein levels of POU2AF2 (Supplementary Fig. 2a). In particular, these genes nearest to the Cluster 3 peaks were strongly upregulated, compared to genes nearest to Clusters 1 and 2 peaks (Fig. 2b, c). The co-upregulated genes by POU2AF2 depletion or GSK126 treatment were also enriched in differentiation and developmental pathways (Fig. 2d, Supplementary Fig. 2b), suggesting a potential co-function between POU2AF2 and the PRC2 complex in human small cell lung cancer cells.
We further evaluated the molecular mechanisms underlying the loss of POU2AF2 and its impact on Polycomb target genes. We found no obvious reduction in total EZH2 or SUZ12 protein levels in POU2AF2-depleted cells (Supplementary Fig. 2c). However, the ChIP-seq assay revealed a significant reduction in H3K27me3 levels at Cluster 3 regions, but not at Cluster 1 or 2 loci (Fig. 2e, Supplementary Fig. 2d), indicating that POU2AF2 depletion may attenuate the occupancy or catalytic activity of PRC2 at these sites.
Next, we conducted ChIP-seq to determine the chromatin occupancy of EZH2 and SUZ12 levels in cells with or without POU2AF2. The results indicated a dramatic reduction in chromatin occupancy of EZH2 and SUZ12 at Cluster 3 peaks in POU2AF2-depleted cells (Fig. 2f, g, Supplementary Fig. 2e–g), consistent with the reduction in H3K27me3 levels observed (Fig. 2e). Notably, the significantly reduced H3K27me3 (Fig. 2h) and EZH2 peaks (Fig. 2i) tend to be broader than the other peaks, indicating that POU2AF2 (narrow peaks) may not physically interact with the PRC2 complex or directly recruit EZH2 to the chromatin. Finally, we did not detect a significant increase of H3K27ac levels at the Cluster 3 peaks after POU2AF2 depletion (Supplementary Fig. 2h), despite elevated gene expression, which is consistent with recent studies demonstrating that enhancer H3K27ac levels may not be directly associated with the nearby gene expression21. The activation of the poised enhancer in cancer cells may be different from that in embryonic stem cells and may not require the deposition of H3K27ac18.
POU2AF2 interacts with the SWI/SNF complex and regulates chromatin accessibility
To further elucidate the mechanism by which POU2AF2 mediates the recruitment of EZH2 to chromatin and whether there is a direct interaction between POU2AF2 and PRC2, we employed GFP-tagged POU2AF2 purification followed by mass spectrometry analysis in both SCLC NCI-H526 and non-SCLC HEK293T cell lines. Our analysis yielded 44 enriched proteins common to both cell lines (Fig. 3a) but we did not detect any subunits of the Polycomb complexes. However, we found multiple subunits of the SWI/SNF complex to be significantly enriched in the POU2AF2 co-purified proteins from both cell lines (Supplementary Fig. 3a). Immunoprecipitation experiments with the POU2AF2-specific antibody confirmed a strong interaction between POU2AF2 and different subunits of the SWI/SNF complex in NCI-H526 cells (Fig. 3b). Consistent with the mass spectrometry results, we did not observe a detectable interaction between POU2AF2 and EZH2, despite the genetic interaction observed between both factors in Fig. 2. To further study the composition of endogenous POU2AF2 and the SWI/SNF complex, nuclear extracts from NCI-H526 cells were subjected to size exclusion (SE) chromatography, followed by western blot analysis of the elution profiles of POU2AF2 and SWI/SNF complex. As shown in Supplementary Fig. 3b, we observed that a large fraction of POU2AF2 was co-eluted with SWI/SNF components. These results led us to hypothesize that POU2AF2 may impact chromatin accessibility via its interaction with the SWI/SNF complex. Following confirmation that protein levels of major subunits within the SWI/SNF complex were not affected after POU2AF2 depletion (Supplementary Fig. 3c), we conducted ChIP-seq with a BRG1-specific antibody in POU2AF2 wild-type and depleted cells to investigate how the loss of POU2AF2 regulates the function of the SWI/SNF complex at the chromatin level. Our analysis revealed a drastic redirection of chromatin-bound BRG1 after POU2AF2 depletion, with a gain of 6,649 and a loss of 21,957 peaks of BRG1 across the whole genome in POU2AF2-depleted cells (Fig. 3c, Supplementary Fig. 3d). Interestingly, the motif analysis demonstrated that both the retained and lost peaks of BRG1 were significantly enriched with the POU2F3 motif, which was not observed in the gained BRG1 peaks (Fig. 3d)….
