PACT sustains proliferation in prostate cancer cells
To identify the functional effects of PACT expression on PCa cell growth we performed loss-of-function studies in LNCaP, C4-2B, 22Rv1, PC3, and DU145 human PCa cell lines. We initially used siRNA to mediate the transient knockdown of PACT in the cell lines. We firstly assessed three different PACT siRNAs (si-PACT) in LNCaP cells (Supplementary Fig. S1A) and selected #s16334 for use in subsequent experiments. We transiently transfected the PCa cells with si-PACT or a negative control siRNA (si-NC) and assessed cell proliferation using a Cell Titer end-point assay and the xCELLigence real time system. The cells with siRNA mediated knockdown of PACT exhibited a growth reduction as compared to the si-NC transfected cells using both methods of evaluation (Fig. 1A), indicating the requisite role for PACT in cell proliferation.
As an alternate approach to reducing PACT via siRNA, we used CRISPR-Cas9 to knock PACT out of the androgen-dependent LNCaP cells. PCR and sequencing confirmed correct targeting of CRISPR-Cas9 to the PACT locus (Supplementary Fig. S2A), and verification of PACT protein depletion was by western blot (Fig. 1B (i)). There was a significant reduction in the proliferation of the PACT knockout (KO) cells compared to the parental cells as assessed by Cell Titer assay (Fig. 1B (ii)), which was in concordance with the PACT siRNA knockdown data (Fig. 1A). Furthermore, colony formation assays showed that LNCaP parental cells formed more colonies than their PACT KO counterparts (Fig. 1B (iii)), indicating a requirement for PACT in cell proliferation, albeit dispensable for cell survival.
To further investigate if the reduced cell proliferation in PACT KO cells is directly due to the absence of PACT, we used lentiviral transduction of PACT cDNA to reconstitute PACT protein in the cells (confirmed by western blotting, Fig. 1C (i)). Re-expression of PACT in the PACT KO cells rescued the reduced cell proliferation phenotype back to the level equivalent to the LNCaP parental cells in Cell Titer and colony formation assays (Fig. 1C (ii-iii)). The loss-of-function studies using siRNA and CRISPR KO, together with the rescue experiment, suggest a direct role for PACT in sustaining the proliferation of PCa cells.
PACT expression positively correlates with genes involved in metabolic processes
Based on the loss-of-function studies using established PCa cell lines, PACT is required for sustaining the proliferation of PCa cells. To investigate if this association with proliferation exists in clinical samples, we interrogated the Prostate Adenocarcinoma samples within the TCGA PanCancer Atlas. Using the Database for Annotation, Visualization and Integrated Discovery (DAVID) we performed gene ontology (GO) pathway analyses on the top 200 genes which showed positive or negative correlation to PRKRA/PACT expression (see gene lists in Supplementary Tables 1 and 2). The biological processes of topmost co-expressed genes included those involved with mitochondrial function and associated metabolic pathways e.g., ATP5PB (Fig. 2A). In contrast, the genes that are inversely correlated to PACT are enriched with biological processes involving epigenetic regulation and chromatin organization e.g., ASXL1 (Fig. 2B). Based on the requirement for PACT in PCa cell proliferation, and the co-expressed gene pathways in patients, we hypothesize that functionally PACT enhances cell growth by regulating metabolic processes.
PACT regulates PCa cell cycle progression
To explore the alterations in gene expression upon PACT depletion in PCa cells, we performed RNA-Sequencing (RNA-seq) analysis of LNCaP parental versus PACT CRISPR KO cells. Using ≥1.5 absolute fold change (FC) (0.58 log2 FC) and a False Discovery Rate (FDR) of p<0.05, we identified 718 differentially expressed genes (DEGs) between the two groups; 343 DEGs were downregulated, and 375 DEGs were upregulated in the absence of PACT (Fig. 3A and Supplementary Tables 3 and 4). DAVID GO and Kyoto Encyclopedia of Genes and Genomes (KEGG)(35) pathway analyses of the downregulated DEGs identified mitotic nuclear division, cell division, cell cycle, and steroid hormone biosynthesis as the most enriched biological processes (Fig. 3B). The pathway analyses of the upregulated genes were mixed and less significant than the downregulated processes, without a clear trend (Fig. 3C).
The enrichment of pathways relating to cell division in the downregulated DEGs in the PACT KO cells corroborated with the reduced proliferation phenotypes observed in the PACT loss-of-function studies. To confirm the finding, we performed cell cycle analyses in LNCaP PCa cells with either transient (siRNA) or stable (CRISPR) mediated PACT depletion. Using propidium iodide (PI) staining and flow cytometry we detected an increase in cell cycle arrest at G0/G1 and a block in transition to S phase using both methods of PACT knockdown (Fig. 3D). Due to the cell cycle arrest at G0/G1, we also investigated the effect of PACT reduction on apoptosis in LNCaP cells using annexin V/PI staining and flow cytometry and observed an increase in apoptotic cells with PACT depletion (Fig. 3E). Taken together, these data support the notion that the growth reduction in PACT KO cells is, in part, via alteration of cell cycle progression and apoptosis, further validating the proproliferative function of PACT in PCa growth.
Targeted knockdown of genes downregulated in the absence of PACT affects PCa cell proliferation
We validated five of the most downregulated and three of the most upregulated genes in the PACT KO cells from the RNA-seq. H2AFJ (H2A histone family member J), PSMD5 (proteasome 26s non-ATPase subunit, 5), AQP3 (Aquaporin 3),TMEM45B (transmembrane protein 45B), and SLC22A3 (Solute carrier family 22 member 3) were respectively downregulated 10.76, 7.18, 5.88, 4.63, and 3.9 log2 fold change; and NOVA1 (neuro-oncological ventral antigen-1), PXDN (Peroxidasin), and RASSF2 (Ras association domain-containing protein 2),were respectively upregulated 2.77, 2.33, and 1.4 log2 fold change, in the PACT KO cells (Supplementary Tables 3 and 4). RT-qPCR quantification of each of these genes in LNCaP parental versus PACT KO cells verified the RNA-seq data (Fig. 4A).
We next used siRNA to transiently knockdown H2AFJ, PSMD5, AQP3, TMEM45B, and SLC22A3 gene expression in LNCaP PCa cells, to assess the functional effects of these molecules on PCa growth. We initially evaluated three different siRNAs for each gene (Supplementary Fig. S3A (i-v)) and selected optimal siRNAs for our experiments (#s31462 H2AFJ; #SI03167108 PSMD5; #s1521 AQP3; #s42359 TMEM45B; and #s13107 SLC22A3). We transiently transfected LNCaP cells with either gene-specific siRNA or si-NC and assessed cell proliferation using a Cell Titer assay at 2 to 6 days post-transfection. Gene-specific siRNA transfected cells exhibited a substantial growth reduction as compared to the si-NC transfected cells for each gene (Fig. 4B), which supports the notion that downregulating PACT, or the genes which are downregulated in the absence of PACT could also benefit PCa patient prognosis by reducing the tumor growth.
PACT modulates key androgen receptor signaling molecules
We next interrogated the entire RNA-seq expression data using Gene Set Enrichment Analysis (GSEA). The GO biological processes and Hallmark gene sets (FDR<0.25, p<0.01) both showed the depletion of gene sets relating to cell cycle and proliferation in the KO cells (e.g., regulation of cell cycle G1/S phase transition, and G2/M checkpoint) (Fig. 5A (i and iii) and 5B (i)). Moreover, biological processes relating to RNA transport and localization, chromosome organization and DNA repair were also depleted in the PACT KO cells (Fig. 5A (i-ii) and 5B (i)), congruent with the reported roles of PACT; acting via PKR and the RISC complex(18, 21).
Interestingly, some of the depleted Hallmark gene sets in the PACT KO cells were NR-function related, including fatty acid metabolism, cholesterol homeostasis, estrogen response late, and bile acid metabolism (Fig. 5B (i-v)), and as we previously described PACT as a NR-coregulator regulating AR activity and downstream gene expression in PCa, we performed GSEA against Hallmark for Androgen response. We observed the upregulation of Androgen response genes in the parental cells relative to the PACT KO cells, indicating a depletion of the gene set in the absence of PACT (Fig. 5C).
We further scrutinized the RNA-seq data to determine if known androgen-regulated genes were differentially expressed in the PACT KO cells, and notably the Kallikrein family members KLK3 (Kallikrein related peptidase 3, Prostate-specific antigen, PSA), KLK2 (Human kallikrein), and KLK4 (PSA-related serine protease) were all downregulated (2.49, 0.83 and 0.47 log2 fold change, respectively) in the PACT KO cells (Supplementary Table 3, and GSE253245). We treated LNCaP parental and PACT KO cells ± dihydrotestosterone (DHT) and validated the differential expression, and hormone responsiveness of these androgen-regulated genes (Fig. 5D). The PACT KO cells also exhibited a lower hormone mediated expression of each gene (Fig. 5D).
We subsequently used siRNA to transiently knockdown PSA in LNCaP PCa cells to assess the functional effects of PSA gene expression on PCa cell growth. We initially tested three different PSA targeting siRNAs (Supplementary Fig. S3C) and selected #SI03078299 for our studies. siRNA-mediated knockdown of PSA resulted in a reduction in cell proliferation at 3- and 6-days post-transfection via a Cell Titer assay, reduced colony formation in clonogenicity assays, and induced cell cycle arrest at G0/G1 (Fig. 5E).
This data supports the role of PACT in modulating key AR-signaling molecules, the targeting of which could abrogate the proproliferative function of PACT.
TRBP has a compensatory role for cell survival when PACT is depleted from LNCaP PCa cells
Our findings suggest that PACT sustains PCa cell proliferation, although it is dispensable for cell survival.Given that PACT and TRBP are both integral members of the RISC complex and co-regulate PKR activation, we next used si-TRBP to investigate the role of TRBP in LNCaP cells ± PACT. We firstly evaluated three different siRNAs targeting TRBP (Supplementary Fig. S3E) and selected #s13790 to use in subsequent experiments. We transiently transfected LNCaP parental cells with either si-PACT, si-TRBP or si-NC; and PACT KO (CRISPR) cells with either si-TRBP or si-NC, and assessed TRBP expression, cell proliferation using a Cell Titer assay, colony formation, and cell cycle. When PACT was depleted from LNCaP cells, there was an increase in the expression of TRBP (Fig. 6A). Parental LNCaP cells exhibited a substantial growth reduction with si-PACT or si-TRBP as compared to the si-NC transfected cells, with the levels comparable to PACT KO cells (Fig. 6B (i)). Additionally, in the PACT KO cells, when TRBP was also depleted the cells exhibited extremely poor viability and very few colonies formed in clonogenic assays (Fig. 6B). Cell cycle analysis of parental LNCaP cells transfected with each gene-specific siRNA or si-NC, or PACT KO cells transfected with si-TRBP or si-NC showed that PACT knockdown induced cell cycle arrest at G0/G1, and TRBP knockdown had no effect on cell cycle in parental cells, and a marginal, yet significant additive effect in the PACT KO cells (Fig. 6C).
Taken together, these data suggest that both PACT and TRBP are essential for LNCaP PCa cell growth, and TRBP expression compensates when PACT is either depleted or knocked out from these cells.