Thursday, September 21, 2023
BestWooCommerceThemeBuilttoBoostSales-728x90

Metformin escape in prostate cancer by activating the PTGR1 transcriptional program through a novel super-enhancer – Signal Transduction and Targeted Therapy


Development of metformin resistance in PCa cells after long-term treatment

Although metformin has been demonstrated to have a cancer-prevention effect, how PCa cells respond to long-term metformin exposure has rarely been explored. Here, we first tried to construct the metformin-resistant PCa cell model (MetR) by continuously treating DU145 and 22RV1 cells, which have distinct characteristics and genetic backgrounds, with the corresponding half-maximal inhibitory concentration (IC50) of the agent. CCK-8 assays, clone formation assays, and studies in subcutaneous xenograft tumor models showed that treated DU145 cells and 22RV1 cells were resistant to metformin after a long period of exposure (Fig. 1a–d). Metformin resistance was suggested to be associated with transcriptional programs that may induce reversible cell cycle arrest.19,21,28 Our clone formation assay showed that MetR cells restored metformin sensitivity after thirty days of drug withdrawal (Fig. 1e, f), for example, the proliferation rate of these MetR cells eventually became similar to that of wild-type (WT) cells. These results indicated that metformin resistance in PCa cells is likely to be a transient phenotype. This is in agreement with previous studies, which indicated that metformin resistance may be associated with reversible cell cycle arrest.19,21,28 Furthermore, MetR cells (DU145-MetR or 22RV1-MetR) and the corresponding control cells (DU145-WT or 22RV1-WT) were paired and subcutaneously injected into the flanks of male nude mice, which were treated with daily feeding of a diet without metformin or a diet intermittently containing metformin every three days. The results showed that the mice injected with the MetR cells had significantly smaller tumor sizes than those injected with the WT cells (Fig. 1g, h). Interestingly, we observed no significant difference in tumor size between the resistant cell injected groups and the control groups in the intermittent metformin feeding model (Fig. 1i, j). Taken together, our data indicated that the development of metformin resistance in PCa cells is due to continuous stimulation by the agent.

Fig. 1

Prostate cancer cells acquire resistance to metformin after long-term treatment. a Representative metformin half-maximal inhibitory concentration (IC50) in two wild-type (WT) and metformin-resistant (MetR) prostate cancer cell lines (DU145, 22RV1, n = 3) determined by the CCK-8 assay and calculated by fitting a nonlinear regression curve. b Colony formation assay of DU145-WT, DU145-MetR, 22RV1-WT and 22RV1-MetR cells treated with different concentrations of metformin. c, d Nude mice (DU145 n = 6, 22RV1 n = 4) received continuous metformin treatment. Tumor volume growth curves and representative images of DU145 (c) and 22RV1 (d) tumors are shown. Metformin was administered at a concentration of 250 mg/kg. e, f Colony formation assay comparing the proliferation of WT cells, MetR cells and MetR cells after 30 days of metformin withdrawal. The results of DU145 cells (e) and 22RV1 cells (f) are shown. g, h Nude mice (DU145 n = 9, 22RV1 n = 8) were fed a diet without metformin. Tumor volume growth curves and representative images of DU145 (g) and 22RV1 (h) tumors are shown. i, j Nude mice (DU145 n = 11, 22RV1 n = 8) received metformin every 3 days. Tumor volume growth curves and representative images of DU145 (i) and 22RV1 (j) tumors are shown. Metformin was administered at a concentration of 250 mg/kg. The tumor sizes were measured at 3-day intervals as soon as the tumors were palpable. *P < 0.05, **P < 0.01, the error bar indicates the standard deviation

Metformin resistance in PCa is acquired through cell cycle reactivation and metabolic reprogramming

As shown in Fig. 2a, b, metformin treatment induced cell cycle arrest in WT cells, as determined by the accumulation of G0/G1-phase cells, consistent with previous report.29 Furthermore, we found an increase in S-phase cells and a concomitant decrease in G0/G1-phase cells among the MetR cells, and such changes became more substantial after metformin treatment. However, there were no significant differences in cell invasion, migration and apoptosis between WT cells and MetR cells (Supplementary Fig. S1). On the other hand, we proved that metformin can effectively inhibit OXPHOS and the production of ATP in PCa in our recent study.30 Thus, we next sought to determine whether continued treatment with metformin would alter the primary metabolic pathways that maintain PCa cell growth. We quantified the level of OXPHOS and glycolytic activity by measuring the oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) respectively using a Seahorse assay, showing that DU145-MetR cells exhibited a higher OCR than DU145-WT cells when treated with metformin (Fig. 2c). The inhibition of basal respiration and ATP production by metformin was much weaker in DU145-MetR cells than those in DU145-WT cells (Fig. 2d). However, basal respiration and ATP production were significantly suppressed in both 22RV1-MetR and 22RV1-WT cells when metformin was applied (Fig. 2e, f). Interestingly, the ECAR in both MetR cell lines were significantly increased regardless of metformin treatment, suggesting that PCa cells have the ability to activate glycolysis to compensate for their energy supply needs (Fig. 2g, h). Taken together, these results indicated that PCa cells may acquire metformin resistance through cell cycle reactivation and metabolic reprogramming.

Fig. 2
figure 2

Reactivation of cell cycle progression and metabolic reprogramming contributes to metformin resistance in prostate cancer cells. a The percentages of G0/G1-, S-, and G2M-phase cells in each group of DU145 cells were determined by using flow cytometric analysis. Metformin (20 mM) was used in the treatment group. b The percentages of G0/G1-, S-, and G2M-phase cells in each group of 22RV1 cells were determined by using flow cytometric analysis. Metformin (10 mM) was used in the treatment group. c, d The oxygen consumption rate in DU145 cells treated with or without metformin (20 mM) was measured by Seahorse assay (c). Basal respiration, Spare respiratory capacity and ATP production were calculated (d). e, f The oxygen consumption rate in 22RV1 cells treated with or without metformin (10 mM) was measured by Seahorse assay (e). Basal respiration, Spare respiratory capacity and ATP production were calculated (f). g Extracellular acidification rate in each group of DU145 cells was measured by Seahorse assay. Metformin (20 mM) was used in the treatment group. h The extracellular acidification rate in each group of 22RV1 cells was measured by a Seahorse assay. Metformin (10 mM) was used in the treatment group. n = 3. Oligo oligomycin, FCCP carbonyl cyanide 4-trifluoromethoxy-phenylhydrazone, 2-DG 2-deoxy-d-glucose. *P < 0.05, **P < 0.01, the error bar indicates the standard deviation

Relevance of aberrant activation of SEs and their target genes in PCa cells with metformin resistance

To understand the underlying mechanisms, we first conducted RNA-seq on metformin-resistant cell lines (Fig. 3a, b). Differential expression analysis revealed that 602 genes were significantly upregulated and 687 genes were downregulated in DU145-MetR cells compared to DU145-WT cells (Supplementary Fig. S2a) and that 996 genes were significantly upregulated and 406 genes were downregulated in 22RV1-MetR cells compared to 22RV1-WT cells (Supplementary Fig. S2b). The differentially expressed genes (DEGs) between DU145-MetR and DU145-WT cells were enriched in cell growth and metabolic pathways (sterol metabolic process, fatty acid metabolic process, and prostanoid metabolic process) (Supplementary Fig. S2c), while the DEGs between 22RV1-MetR and 22RV1-WT cells were enriched in the cell cycle and DNA replication pathways (Supplementary Fig. S2d). The fold changes in the expression levels of metabolic DEGs and cell cycle-related DEGs shown in detail in Supplementary Fig. S2e,f, indicating that cell cycle-related genes and metabolic genes were differentially expressed in metformin-resistant cells. Studies have demonstrated that SEs and their associated networks of transcription factors play an important role in regulating drug resistance and progression in prostate cancer.25 SEs are characterized by higher levels of binding of chromatin factors associated with enhancer activity, such as cohesin, and histone modifications, including H3K27ac, dimethylation of histone H3 at lysine 4 (H3K4me2), and H3K4me1.22

Fig. 3
figure 3

Aberrant activation of super-enhancer and its associated gene PTGR1 in a preexist cluster of prostate cancer cells may be associated with metformin resistance. a Heatmap of the RNA-seq results in DU145-WT and DU145-MetR cells (n = 3). Each row represents the transformed FPKM Z-score of an individual gene. b Heatmap of the RNA-seq results in 22RV1-WT and 22RV1-MetR cells (n = 3). Each row represents the transformed FPKM Z-score of an individual gene. c Average intensity curves of the H3K27ac ChIP-seq signal at the super-enhancer regions and the ±3 kb flanking regions in DU145-WT and DU145-MetR cells. Boxplot of super-enhancer peak density between DU145-MetR and DU145-WT cells. d Average intensity curves of H3K27ac ChIP-seq signal at the super-enhancer regions and the ±3 kb flanking regions in 22RV1-WT and 22RV1-MetR cells. Boxplot of super-enhancer peak density between 22RV1-MetR and 22RV1-WT cells. e GSEA analysis based on the pre-ranking genes that ordered by the fold change (FC) from differentially expressed analysis in DU145-MetR cells versus. DU145-WT cells with the input annotation generated from DU145-MetR SE-associated genes. f GSEA analysis based on the pre-ranking genes that ordered by the fold change (FC) from differentially expressed analysis in 22RV1-MetR cells versus. 22RV1-WT cells with the input annotation generated from 22RV1-MetR SE-associated genes. g DU145-MetR enhancer ranking plot based on H3K27ac ChIP-Seq signals using the ROSE algorithm. h 22RV1-MetR enhancer ranking plot based on H3K27ac ChIP-Seq signals using the ROSE algorithm. i Heatmaps of H3K27ac ChIP-seq signals at super-enhancer (left, SE) or typical enhancer (right, TE) regions in DU145-MetR cells. j Heatmaps of H3K27ac ChIP-seq signals at super-enhancer (left, SE) or typical enhancer (right, TE) regions in 22RV1-MetR cells. k Log2(FPKM) of typical enhancer-associated genes and super-enhancers-associated genes in DU145-MetR cells (left) and 22RV1-MetR cells (right). l T-distributed stochastic neighbor embedding (t-SNE) plot for the sub-clusters identified by single-cell RNA-sequencing analysis in DU145 pre-MetR and DU145-WT, and stacked bar chart for the distribution of each subcluster in DU145 pre-MetR and DU145-WT. m The gene PTGR1 was identified by intersection analysis of H3K27ac ChIP-Seq results and marker genes of Cluster 0

In our study, to test the hypothesis that the genes contributing to metformin resistance are upregulated by transcriptional programs such as histone modification and cis-regulatory elements, we performed H3K27ac ChIP-Seq in both WT and MetR cells. ChIP-Seq data were processed through ROSE (Rank Ordering of super-enhancers) to identify super-enhancers and typical enhancers (TEs).31 In order to access the statistical difference of H3K27ac signal in DU145-MetR SE region between DU145-MetR and DU145-WT cells, we quantified and normalized the average read coverage as the peak density of each DU145-MetR SE by using Homer annotatePeaks.pl function32 and presented the difference by boxplot in Fig. 3c, d. The boxplot indicated significant difference between DU145-MetR and DU145-WT cells (Wilcoxon, P value = 0.017) and a significant difference between 22RV1-MetR and 22RV1-WT cells (Wilcoxon, P value < 2.2e-16). The results showed that the H3K27ac signal in the super-enhancer regions of MetR cells was significantly stronger than that in WT cells (Fig. 3c, d). We next conducted gene set enrichment analysis (GSEA) based on the fold changes in the expression of preranked genes differentially expressed between DU145-MetR and DU145-WT cells. The input annotations were generated from the SE-associated genes in DU145-MetR cells. Our results showed that SEs associated genes signature enriched in DU145-MetR cells versus DU145-WT cells (Fig. 3e, NES = 1.43, adjusted P = 0.01). However, the SEs associated genes signature between 22RV1-MetR cells and 22RV1-WT cells showed no significant difference (Fig. 3f, NES = 0.73, adjusted P = 1).

With the ROSE algorithm, 281 SEs and 7788 TEs were called in DU145-MetR cells, while 643 SEs and 10,610 TEs were called in 22RV1-MetR cells (Fig. 3g, h). Moreover, 269 and 603 genes were annotated as SE-associated genes in DU145-MetR cells and 22RV1-MetR cells, respectively. The composite heatmap showed that, in both MetR cell lines, the increases in H3K27ac in SEs were significantly greater than that in TEs (Fig. 3i, j). Although the number of SEs was much less than that of TEs in MetR cells, the expression levels of SE-associated genes were significantly higher than those of genes associated with TEs (Fig. 3k). Taken together, our data indicated that aberrant activation of SEs and their target genes may account for metformin resistance in PCa cells.

Upregulation of PTGR1 in PCa cells indicates metformin resistance

The androgen receptor (AR) has been demonstrated to have an impact on the effectiveness of metformin treatment, and typically, drug resistance develops in only a subset of cancer cells, not all cells.20,22 To explore whether a subgroup of DU145 cells with AR-negative expression is predisposed to acquiring metformin resistance, we performed single-cell RNA-seq analysis on a DU145 cell model undergoing the acquisition of metformin resistance (DU145 pre-MetR) and identified 6 clusters (0 through 5) using unsupervised clustering (Fig. 3l). Our comparison with DU145-WT cells revealed that the number of cells in Cluster 0 was increased by more than 30%, while the numbers of cells in the other clusters were decreased (Table 1). We designated the cells in Cluster 0 as DTP-like cells, which could potentially develop metformin resistance. Four marker genes were identified within Cluster 0. By intersecting these data with the H3K27ac ChIP-Seq data, we found a common gene, Prostaglandin Reductase 1 (PTGR1) between the marker genes of Cluster 0 and the SE-associated genes in DU145-MetR cells (Fig. 3m), in addition, mRNA level of PTGR1 was increased in Cluster 0 (Fig. 4a). Notably, among the four marker genes (PTGR1, DDIT4, CEBPD, and EEF1A1) associated with the metformin-resistant subcluster (Cluster 0), only PTGR1 was significantly positively correlated with the cell cycle in the TCGA-PRAD dataset (Supplement Fig. S2g).

Table 1 Cell numbers of each cluster in WT group and pre-MetR group of DU145 cells
Fig. 4
figure 4

Increased expression of the super-enhancer-associated gene PTGR1 is associated with acquired metformin resistance. a Relative mRNA levels of PTGR1 in the general population (left), Cluster 0 (resistant) and Cluster 2 (sensitive) of DU145 pre-MetR cells from single-cell RNA sequencing (DU145 pre-MetR, cells undergoing metformin resistance). b mRNA (i) and protein (ii) levels of PTGR1 in DU145 and 22RV1 cells with or without metformin resistance were measured by qRT‒PCR and western blot analysis (n = 3). c Representative immunohistochemistry (IHC) images showing the PTGR1 expression pattern in the subcutaneous xenograft tissue from mice in the DU145-WT+Met and DU145-MetR+Met groups. Scale bar = 50 μm. d Immunofluorescence (IF) analysis of PTGR1 expression in DU145-WT and DU145-MetR cells. Representative images are shown. Scale bar = 20 μm. e mRNA (i) and protein (ii) levels of PTGR1 in PC3 cells with or without metformin resistance. (iii) IF analysis of PTGR1 expression in PC3-WT and PC3-MetR cells are shown. Scale bar = 20 μm. f qRT‒PCR and western blot analysis were utilized to validate the overexpression of PTGR1 in DU145 and 22RV1 cells transfected by lentiviral plasmids (n = 3). g The effect of PTGR1 overexpression on metformin treatment in DU145 (20 mM) and 22RV1 (10 mM) cells was analyzed by a CCK-8 assay (n = 3). h, i PTGR1 expression was decreased in DU145-MetR cells and PC3-MetR cells using siRNA, and the effect was validated by qRT‒PCR and western blot analysis. j, k The effect of decreased expression of PTGR1 on metformin treatment in DU145-MetR (20 mM) and PC3-MetR (20 mM) cells was analyzed by a CCK-8 assay (n = 3). *P < 0.05, **P < 0.01, the error bar indicates the standard deviation

Next, by assessing its expression at both the mRNA and protein levels, we confirmed that PTGR1 was significantly upregulated in DU145-MetR cells compared with DU145-WT cells (Fig. 4b–d). However, we did not observe any difference in PTGR1 expression between 22RV1-WT cells and 22RV1-MetR cells (Fig. 4b), consistent with the assumption that the metformin efficacy is somehow compromised or confounded in AR-positive PCa cells. To gain further insights, we expanded our analysis by constructing another metformin-resistant cell model with PC3 cells (PC3-MetR), another AR-negative PCa cell line. The results of qRT‒PCR, western blot analysis, and immunofluorescence analysis showed that PTGR1 was overexpressed in PC3-MetR cells compared to PC3-WT cells (Fig. 4e). To assess the presumed relationship between PTGR1 expression and metformin efficacy, we applied lentiviral transduction to establish two PCa cell lines, i.e., DU145-PTGR1 and 22RV1-PTGR1, that can stably express PTGR1 (Fig. 4f). The high level of PTGR1 expression was found to significantly attenuate metformin efficacy in both DU145 and 22RV1 cells (Fig. 4g). Moreover, when we downregulated PTGR1 by transiently transfecting MetR cells with siRNA, the effect of metformin was enhanced as expected (Fig. 4h–k). In summary, our data and results implied that increased expression of PTGR1 in PCa indicates metformin resistance and that PTGR1 may serve as a biomarker for metformin treatment selection.

Upregulation of PTGR1 promotes cell cycle progression and is related to poor survival in PCa patients

To investigate how PTGR1 antagonizes metformin treatment, GSEA was employed to delineate the potential biological pathways involving PTGR1 in PCa. The results showed that cell cycle-related pathways, i.e., the MYC, G2M checkpoint, and E2F target pathways, were enriched in the PTGR1-upregulated group, indicating the relatedness between the antagonizing effect of PTGR1 on metformin treatment and the activation of cell cycle pathways (Fig. 5a, b). We then utilized flow cytometric analysis to demonstrate that upregulated PTGR1 can effectively abrogate metformin-induced G0/G1 arrest and promote S and G2/M-phase entry in cancer cells (Fig. 5c). In contrast, downregulated PTGR1 restored metformin sensitivity in MetR cells by blocking cells in G0/G1 phases and decreasing the population of cells in S phase or G2/M-phase (Fig. 5d). To investigate the impact of PTGR1 on the cell cycle, we performed western blot analysis. Our results showed that the expression of E2F Transcription Factor 3 (E2F3) was upregulated in DU145-MetR cells and in DU145 cells that stably overexpressed PTGR1. Furthermore, when PTGR1 expression was suppressed through siRNA transfection in DU145 cells, the expression of E2F3 was downregulated (Supplementary Fig. S3a). It is known that upregulation of E2F3 plays a crucial role in promoting the S-G2 transcriptional program.33 Therefore, we treated DU145-PTGR1 cells with CDK4/6 inhibitors, including Ribociclib, Palbociclib, and Abemaciclib. These inhibitors have been shown to induce G1 arrest by targeting the Retinoblastoma/E2F repressive complex.34 Our results showed that the group with high expression of PTGR1 exhibited a higher growth rate than the control group (Fig. 5e). These data and results suggested that PTGR1 likely exerts an antagonizing effect on metformin treatment by interfering with cell cycle arrest, which may be related to the role of E2F3.

Fig. 5
figure 5

Upregulation of PTGR1 promoted cell cycle progression and was associated with poor survival in prostate cancer patients. a Bubble plot shows the biological pathways activated/suppressed in the high PTGR1 expression group. b Gene set enrichment analysis (GSEA) plot of cell cycle-related pathways enriched in the high PTGR1 expression group. c The effect of PTGR1 overexpression on the cell cycle with or without metformin treatment (20 mM) was analyzed by flow cytometry analysis. The percentages of G0/G1-, S-, and G2M-phase cells were compared among the groups. (n = 3) d The effect of decreased expression of PTGR1 on the cell cycle with metformin treatment (20 mM) was analyzed by flow cytometry analysis. The percentages of G0/G1-, S-, and G2M-phase cells were compared among the groups. (n = 3) e The effect of PTGR1 overexpression on CDK4/6 inhibitors, including abemaciclib, palbociclib and ribociclib, was analyzed by a CCK-8 assay. The concentration of each CDK4/6 inhibitor was 10 nM. Optical density values were determined by CCK-8 assay and measured at 450 nm (n = 3). f The proportion of PCa patients with aberrant activation of PTGR1 was analyzed in public databases by cBioPortal (http://www.cbioportal.org/). g Biochemical recurrence (BCR)-free survival of patients with high PTGR1 expression was compared with that of patients with low PTGR1 expression in the TCGA and GEO databases. *P < 0.05, **P < 0.01, the error bar indicates the standard deviation

In the literature, high expression of PTGR1 has been reported to be associated with poor prognosis in several types of cancers.35 To assess the clinical significance of PTGR1 in PCa, we surveyed its expression in PCa patients across public databases. For each PCa cohort in Fig. 5f, aberrant activation of PTGR1 was frequently found in a subset of patients who may show no response to metformin treatment. Further exploration of The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases indicated that PCa patients with high PTGR1 expression had shorter biochemical recurrence (BCR)-free survival times than those with low PTGR1 expression (Fig. 5g).

PTGR1 expression is upregulated by super-enhancer bound by the master transcription factors SRF and RUNX3

To understand the relationship between PTGR1 expression and metformin resistance in DU145-MetR cells, we next investigated the underlying mechanism. super-enhancers (SEs) are critical regulatory elements and have been linked to the expression of genes associated with drug resistance.24,25,26 Our hypothesis was that the upregulation of PTGR1 in DU145-MetR cells might be associated with a specific SE. Thus, we conducted H3K27ac ChIP-Seq and applied the ROSE algorithm (Rank Ordering of super-enhancers), which is based on the active enhancer marker histone modification H3K27ac. Our analysis revealed a total of 281 SEs in DU145-MetR cells, one of which was located ~10 kb upstream of the transcription start site (TSS) of PTGR1. We selected the two constituent enhancers with the highest intensity signals and named them E1 and E2 (Fig. 6a). To examine the regulatory function of this SE on PTGR1 expression, we constructed two plasmids that contained minimal promoters of E1 and E2 to drive luciferase expression, and then transfected these plasmids into 293T cells. The results showed that the transcriptional activity was increased when either the E1 or E2 plasmid was successfully transfected and was significantly higher in cells with E2 transfection (Fig. 6b).

Fig. 6
figure 6

PTGR1 is upregulated by an upstream super-enhancer bound by the master transcription factors SRF and RUNX3. a Genome browser view of normalized H3K27ac ChIP-seq signals at the PTGR1 locus in DU145-WT and DU145-MetR cells. Two tracks were the average of two biological replicates. The super-enhancer (SE) region is marked by the blue line. The two constituent enhancers (Element 1, E1 and Element 2, E2) within the PTGR1-SE region are marked by the red lines. b Luciferase reporter assays were performed in 293T cells to validate the combination of SE and PTGR1. The Luciferase signal was normalized to the Renilla transfection control luciferase signal (n = 5). c qRT-PCR and western blot analysis were used to evaluate the expression level of PTGR1 in DU145-MetR cells with downregulation of BRD4 (n = 3). d qRT-PCR and western blot analysis were used to evaluate the expression level of PTGR1 in DU145-MetR cells treated with 1 μM JQ1 (n = 3). e Summary of TF motif occurrences within E1 and E2 elements. TFs expression in the metformin resistance cluster is shown in the dot plot for each motif. Statistically significant motif matches identified by FIMO were defined as P value (i.e., q value) <0.05. f Capture of the The Cistrome Data Browser showed the locations of the predicted TFs binding sites and the PTGR1-SE locus. g qRT-PCR and western blot analysis were performed to assess the mRNA and protein levels of PTGR1 after downregulating the predicted TFs (n = 3). The corresponding protein bands representing beta-actin are shown in Supplementary Fig. S5. h The binding sites indicated by the SRF ChIP-seq and RUNX3-ChIP-seq. i ChIP-qPCR analysis for enrichment of RUNX3 and SRF at the super-enhancer identified in Fig. 6a (n = 3). j Luciferase reporter assays were performed in 293T cells to validate the combination of SE with SRF and RUNX3, respectively. The Luciferase signal was normalized to the Renilla transfection control luciferase signal (n = 5). k The effect of decreased expression of SRF (left) or RUNX3 (right) on metformin treatment was analyzed by CCK-8 assay (n = 3). *P < 0.05, **P < 0.01, error bar indicates the standard deviation

To further confirm that the SE specifically regulates PTGR1, we also evaluated the expression of six neighboring genes located within a 20-kb flanking distance from the PTGR1-SE site: DNAJC2, GNG10, ECPAS, LRRC37A5P, SHOC1, and ZNF48. However, our analysis showed that the expression of these neighboring genes did not show statistically significant differences between DU145-MetR cells and DU145-WT cells (Supplementary Fig. S4a). It has been recognized that BRD4 is enriched in SEs, which is required for SE-regulated transcriptional activity.36 Therefore, we applied BRD4 siRNAs as well as JQ1, a BRD4 inhibitor, to impair the regulatory function of SEs in DU145-MetR cells. As expected, the expression of PTGR1 was significantly decreased at both the mRNA and protein levels after BRD4 downregulation with siRNAs, and the experiment with JQ1 produced consistent results (Fig. 6c, d). However, the expression levels of these six neighboring genes did not exhibit significant changes (Supplementary Fig. S4b, c).

Next, we sought to identify the transcription factors involved in the regulation of PTGR1 by the SE. To this end, we conducted motif analysis using Homer and scanned the motifs from the JASPAR 2022 database using FIMO with default parameters.37 We then explored the expression of the predicted TFs, including SRF, TFDP1, SOX12, ZSCAN29, and RUNX3, in the metformin-resistant cell cluster using single-cell RNA-Seq (Fig. 6e). We also obtained ChIP-Seq data of genomic regions enriched with these transcription factors and mapped their binding sites in the E1/E2 element using FIMO software, resulting in the identification of four transcription factors, SRF, RUNX3, TFDP1, and ZSCAN29 (Fig. 6f). Furthermore, qRT-PCR and western blot analyses revealed that PTGR1 expression was significantly reduced when SRF and RUNX3 were downregulated (Fig. 6g and Supplementary Fig. S5). Moreover, ChIP-seq analysis results showed that SRF and RUNX3 were able to bind to the SE regions of PTGR1 in DU145-MetR cells (Fig. 6h). To validate the ChIP-seq results, ChIP-qPCR was performed to quantify the occupancy of RUNX3 and SRF, and their enrichment was confirmed at the SE regions of PTGR1 (Fig. 6i). We then applied the luciferase reporter assay to evaluate the functionality of SRF and RUNX3 in the regulation of PTGR1 SE, and the results demonstrated that upregulation of SRF or RUNX3 can significantly increase luciferase expression (Fig. 6j). Notably, DU145-MetR cells became sensitive to metformin when SRF or RUNX3 was downregulated (Fig. 6k). Furthermore, we suppressed the expression of RUNX3 and SRF in DU145-MetR cells using transient transfection of siRNAs. The results showed that the expression levels of those six neighboring genes remained unchanged (as indicated in Supplementary Fig. S4d). Considering these results collectively, we concluded that SE activates PTGR1 expression by interacting with the key transcription factors SRF and RUNX3.



Source link

Related Articles

Leave a Reply

Stay Connected

9FansLike
4FollowersFollow
0SubscribersSubscribe
- Advertisement -spot_img

Latest Articles

%d bloggers like this: