Pan-cancer integrated bioinformatic analysis of RNA polymerase subunits reveal RNA Pol I member CD3EAP regulates cell growth by modulating autophagy

ABSTRACT

Transcription is a crucial stage in gene expression. An integrated study of 34 RNA polymerase subunits (RNAPS) in the six most frequent cancer types identified several genetic and epigenetic modification. We discovered nine mutant RNAPS with a mutation frequency of more than 1% in at least one tumor type. POLR2K and POLR2H were found to be amplified and overexpressed, whereas POLR3D was deleted and downregulated. Multiple RNAPS were also observed to be regulated by variations in promoter methylation. 5-Aza-2-deoxycytidine mediated re-expression in cell lines verified methylation-driven inhibition of POLR2F and POLR2L expression in BRCA and NSCLC, respectively. Next, we showed that CD3EAP, a Pol I subunit, was overexpressed in all cancer types and was associated with worst survival in breast, liver, lung, and prostate cancers. The knockdown studies showed that CD3EAP is required for cell proliferation and induces autophagy but not apoptosis. Furthermore, autophagy inhibition rescued the cell proliferation in CD3EAP knockdown cells. CD3EAP expression correlated with S and G2 phase cell cycle regulators, and CD3EAP knockdown inhibited the expression of S and G2 CDK/cyclins. We also identified POLR2D, an RNA pol II subunit, as a commonly overexpressed and prognostic gene in multiple cancers. POLR2D knockdown also decreased cell proliferation. POLR2D is related to the transcription of just a subset of RNA POL II transcribe genes, indicating a distinct role. Taken together, we have shown the genetic and epigenetic regulation of RNAPS genes in most common tumors. We have also demonstrated the cancer-specific function of CD3EAP and POLR2D genes.

Introduction

Cancer is the second most common cause of lifestyle-disease-related death worldwide after cardiovascular disease. The primary reason for the poor management of cancer is the lack of early diagnosis and resistance to the therapy [1]. This requires the continuous development of novel biomarkers and therapeutic targets. Most common driver genes are either already being utilized for therapy or classified as “undruggable”, creating a need for further identification of therapeutic targets [2]. Detection and analysis of such biomarkers for cancer can also be studied similarly through advanced biotechnological techniques, allowing for accurate and early diagnosis, personalized treatment approaches, and improved patient outcomes, ultimately contributing to the advancement of cancer research and clinical practice [3–7]. Transcription by RNA Polymerase I, II& III (RNAP) is essential for continuous protein synthesis and therefore is a fundamental process for cell survival. Among these, RNA Pol I is involved in rRNA, RNA pol II in mRNA, and RNA Pol III in tRNA synthesis. Cell cycle progression is not possible in the absence of cell growth which is dependent on adequate protein synthesis [8]. Even a slight decrease in protein synthesis causes cells to come out of the cell cycle in quiescence showing evidence of the dependence of proliferation on protein synthesis. As the number of mRNAs far exceeds the number of ribosomes, the availability of rRNA and tRNA becomes a rate-limiting step in protein synthesis [9]. For example, a 50% decrease in tRNA can increase cell doubling time by a factor of three in yeast [10]. Hence, RNAP I and RNAP III become an interesting target for cell proliferation inhibition.

Indeed, many studies have proposed targeting the RNA polymerase machinery to reduce cell growth in cancer. The rRNA transcription by RNA Pol I complex is a major step in ribosome biogenesis (RiBi). It has been shown that inhibition of the RiBi process leads to activation of pathways like p53 causing cell death, making the RiBi process a prime candidate for therapeutic target [11,12]. Given the fundamental role of rDNA transcription in ribosome biogenesis several studies feature, TP53INP2 as a key determinant in promoting ribosome biogenesis through regulation the assembly of POLR1 preinitiation complex during rDNA transcription in the nucleolus [13,14]. Multiple reports now suggest that RiBi may play a more extensive role in cancer development than previously thought [15]. Like RNA Pol I, many studies have shown increased RNAP III activity in cancer. The in vitro transformation assays using defined genetic steps have shown that level of RNAP III subunits is increased in transformed cells compared to untransformed cells [16]. Increased levels of RNA Pol III subunits like BRF1 and BRF2 have been reported in multiple tumor types. BRF1 level was increased after alcohol consumption in hepatocellular carcinoma patients, and knockdown of BRF1 in HCC cell lines reduced EGF-mediated transformation [17]. The role of RNAP II in cancer development and progression is not well established owing to the non-redundant function of RNA Pol II. The only known exception is the recurrent mutation of the TFIIH (ERCC2) subunit associated with nucleotide excision repair [18]. However, some RNAP II subunits, like POLR2A, POLR3C, etc., have been known to play a role in carcinogenesis. POLR2A expression is required for the cell growth, specifically in Tp53 deleted cells [19]. However, the expression of many RNA Pol II subunits has been shown to be upregulated in cancer. For example, TATA-binding protein expression is increased in cancer and is a contributing factor in oncogenesis [20]. Similarly, many transcription factors that regulate gene expression via RNAP II are deregulated in cancer [21–23]. These studies suggest that RNA polymerases play a central role in oncogenesis. However, there is no comprehensive study done to understand the genomic and expression landscape of RNAP subunits in cancer.

The RNA polymerase I, II, and III are consist of eight, fifteen, and eleven subunits correspondingly (Figure 1a) [24]. Subunits POLR1C and POLR1D are commonly shared by RNA pol I and III (Figure 1a). Additionally, POLR2E, POLR2F, POLR2H, POLR2K, and POLR2L are present in all the RNA polymerases (Figure 1a). Here in this work, we have used TCGA from the six most common cancer types (Breast, Colon, Liver, non-small cell lung, prostate, and stomach cancers). As lung cancer has two significant subtypes (LUAD and LUSC), we used both in this analysis. We have performed integrated analysis to understand the mutation and somatic copy number variation in RNAP subunits. We identified CD3EAP and POLR2D as the most commonly differentially expressed and prognostic RNAP subunits using expression analysis. We also showed that CD3EAP and POLR2D are required for cell growth. CD3EAP knockdown is associated with decreased RiBi and autophagy. We also found that inhibition of CD3EAP-mediated autophagy partially rescued cell proliferation.

Figure 1.

Mutation and CNV analysis of cancer: a) list of RNA polymerase subunits. The commonly shared subunits are colored, and keys are provided. b) mutation frequency of RNAPS in five common cancers studied. c) the mutation frequency of all the RNAPS was obtained from cBioportal, and genes with more than 1% mutation frequency are plotted. d) Somatic copy number variations of RNAPS were obtained from cBioportal, and genes with more than 5% CNV frequency were plotted. e) the CNV value was correlated with the expression of given RNAPS and plotted as a heatmap. Yellow represents high and blue represents low correlation. f, g, and h) expression of POLR2K, POLR2H and POL3D were divided into not changed, deleted, and amplified groups based on CNV and plotted. Each group was compared with the corresponding normal using t-test, and a p-value was obtained.

Material and methods

Datasets

In the current study, patients’ data associated with BRCA, COAD, LIHC, LUAD, LUSC, and PRAD (n ~ 3334) cancers and 734 normal samples from breast, colon, liver, lung, and prostate were used. Clinical data associated with BRCA, COAD, LIHC, LUAD, LUSC, and PRAD were downloaded from cBioPortal. The list of RNAPS was obtained from HUGO Gene Nomenclature Committee.

Genomic variation analysis

The mutation frequency and Copy number variation of RNAPS genes were examined using the cBioPortal database (https://www.cbioportal.org/). For mutation analysis, the mutation frequency of RNAPS was calculated in each cancer type separately. Genes with more than 1% mutation frequency in at least one tumor type were considered recurrent. Similarly, CNV data of RNAPS were calculated for each cancer type and plotted. The genes with more than 5% variation in at least one tumor type were considered recurrently amplified or deleted.

Expression analysis

Expression and methylation data were downloaded from the GDC data portal for The Cancer Genome Atlas (TCGA) program for all the tumor types. To identify the differentially expressed RNA polymerase genes, a t-test was carried out using the FPKM values from the tumor and the normal samples. Genes with log2FC ≥ 2 and FDR p-value ≤0.05 were considered to be differentially expressed. Along the same lines, methylation beta values were used to identify the differentially methylated CpG probes. CpG probes on X and Y chromosomes and those representing SNP were dropped using the rmSNPandCH function. Probes and samples with more than 10% missing data were dropped. And among the probes retained, the missing values were imputed by K-nearest neighbors (KNN) with K = 5. Further, the beta values were transformed into M-values, and differential methylation analysis was carried out using the limma package. CpG probes with log2FC > 2 and q-value ≤0.05 were considered to be differential methylated.

Survival analysis

To understand the association of RNAPS genes expression with survival, Cox regression analysis was performed, and Hazard ratio (HR) with < 0.05 p-value were considered significant. The Kaplan-Meier analysis was used for the comparison of the survival of two groups using the Mantle-Cox test. HR with a p-value <0.05 were considered significant.

Statistical analysis and network analysis

The two-sided t-test was used in GraphPad8.3 to compare two groups, and tests with p-values less than 0.05 were considered significant. FDR correction was used for multiple tests, and FDR 0.05 was judged significant. The Kruskal-Wallis test was used to compare three or more groups, and a p-value of 0.05 was considered significant. For Correlation analysis, Pearson correlation coefficients were calculated, and coefficients with < 0.05 p-value were considered significant. For network analysis, metascape databases was used. Gene lists were sued as input, and terms with less than 0.05 p-value were considered significant.

siRNA transfection

MDA-MB-231 and A549 cells were seeded at a density of 30,000 cells per well in a 96-well plate, 24 hours prior to the transfection. Next day RNAi duplex-Lipofectamine™ RNAiMAX (Invitrogen #13778–150) complexes were prepared for each 20 μM siRNA (siNT, siATG5 & siCD3EAP) and incubated at RT for 10 min followed by transfection in target cell lines using optiMEM(Gibco #31985062) without antibiotics. Post transfection, cells were incubated for 48 hours at 37°C in a CO2 incubator. Successive transfection was performed using siATG5 (Eurogentec #7801923) and siCD3EAP (Eurogentec #7801926), where on the first day the cells were transfected with siATG5 and the following day siCD3EAP transfection was performed on the same cells. Two days after transfection, the cells were trypsinized for RNA and protein isolation to check the knockdown.

RNA isolation and qPCR

RNA isolation was carried out using RNAiso Plus (TaKaRa, Cat.# 9108) as per the manufacturer’s instructions. Briefly, 0.1 million cells were collected in 1.5 mL microcentrifuge tubes and lysed using 500 uL of RNAiso Plus by pipetting in and out multiple times and incubating it for 10 minutes at room temperature. Later, 100 uL of chloroform (Finar, Cat.#10470PL025) was added, shaken the mixture for about 30 seconds and incubated at room temperature for about 5 minutes. The mixture was later centrifuged at 10,000 × g for 10 minutes at 4°C, which resulted in separation of organic and aqueous layers. Then, carefully transferred the upper aqueous layer into a new 1.5 mL microcentrifuge tube and added about 500 uL of isopropanol (SRL, Cat.# 62986) and mixed them gently by inverting the tubes 5–6 times. Later, the resultant mixture was centrifuged at 12,000 × g for about 15 minutes. The pellet was washed with 75% ethanol (MB228-500 ML, Himedia) and air-dried at room temperature for 5 minutes. The pellet was resuspended in molecular grade H₂O and 500 ng was used for cDNA synthesis using iScript (Bio Rad, Cat.# 1708841). For qPCR analysis, universal SYBR Green mix (Bio Rad: 1725271) was utilized, and real-time analysis was carried out in Bio Rad CFX96 using two-step amplification and melt curve protocol. The qPCR analysis was performed following the delta-delta Ct value and plotted.

Protein isolation and Western Blot

A total of 0.3 × 10 [6] to 0.6 × 10 [6] cells were counted and used to isolate protein. Briefly, the cells were detached from the culture dish using trypsin solution (Sigma, Cat.# T4049), and all the samples were collected in separate microcentrifuge tubes. They were washed with cold PBS (Himedia, Cat.# ML023-500 ML) and lysed in a lysis buffer (Pierce IP Lysis Buffer: 87787) supplemented with 1X protease inhibitor cocktail (Halt protease inhibitor cocktail: 87786). The cell lysates were centrifuged at 13,000×g for 15 minutes, and the supernatant was collected. The total protein concentration was estimated before loading onto polyacrylamide gel by BCA protein assay kit (Pierce: 23227), and a total of 40 µg of cell lysates from different experimental cells were loaded and resolved in 12% polyacrylamide gel (Biorad TGX FastCast Acrylamide Kit 12%: 161–0175)). The gel was blotted in a PVDF membrane (Immun-Blot PVDF Membranes for Protein Blotting, Biorad: 1620177), and the membrane was blocked using a 5% blocker solution (Blotting Grade Blocker, Biorad: 1706404). The membrane was cut as per molecular size of proteins and incubated with PARP1 (proteintech: 66520–1-Ig), Vimentin (Invitrogen: MA5–16409), SNAI1 (proteintech: 13099–1-AP), SNAI2 (proteintech: 12129–1-AP), POLR2D (ABclonal: A1859), CD3EAP (ABclonal: A16099), LC3 (proteintech: 14600–1-Ap), H2AX (proteintech: 10856–1-Ap), CDH1 (Invitrogen: 13–1700), CASP3 (proteintech: 19677–1-AP) and internal control Actin B (ABclonal: AC004), GAPDH (ABclonal: AC002) and/or Histone H3 (proteintech: 17168–1-Ap) antibodies. The membrane was later incubated with HRP-tagged goat anti-mouse (Biorad: 172–1011) or anti-rabbit (Invitrogen: 31460)) secondary antibodies. All the steps were followed by washing with TBS (Biorad: 1706435) consisting of 0.1% Tween 20 (Polysorbate 20, MP Biomedicals: 103168). The bound antibody complexes were detected with ECL Western Blotting substrate (Pierce: 32209) or femtoLUCENT PLUS HRP chemiluminescent reagents (G Biosciences: 786–003).

Cell proliferation assay

The cell proliferation assay was carried out using MTT [3-(4,5-dimethylthiazol-2-yl)- 2,5-diphenyl-tetrazolium bromide] (Sigma: M5655-1 G). Briefly, all the experimental cells were seeded in a 96-well plate at a density of 2000 cells/well in triplicate for 4 or 6 days. MTT was added to the well every 24 hours and incubated at 37°C and 5% CO₂ (95% atmospheric air) in a humidified incubator for 4 hours. The formazan complex was dissolved in DMSO (Himedia: MB058-500 ML) and recorded the absorbance at 570 nm. The data were normalized and plotted

Colony formation assay

Cells were seeded in a 12-well plate at a density of 300 cells/well in duplicate and allowed then to grow for about two weeks. Later, the media was removed, washed with PBS, and stained in 0.25% crystal violet (LOBA CHEMIE: 0305000025). The image was captured and shown.

Monitoring the autophagy flux using pcDNA3-GFP-LC3-RFP-LC3ΔG plasmid

MDA-MB-231 and A549 cell lines stably expressing shRNA against CD3EAP-mRNA were generated by lentiviral transduction. Following selection with puromycin (2 ug/ml), these cells were transfected with pcDNA-GFP-LC3-RFP-LC3ΔG construct (Addgene#168997) using Lipofectamine 3000 reagent (Invitrogen -2,173,180) according to manufacturer’s instructions. 24 hours post-transfection, cells were harvested and seeded in a fresh 96-well culture dish and subsequently selected with G418 at 300 ug/ml for 1 week. Following antibiotic selection, images were acquired using Leica Fluorescence Microscopy. For image processing and calculating the ratio, GFP/RFP so as to measure the autophagic flux, ImageJ software(NIH) was used.

Autophagy detection using mCherry-GFP-LC3B WT plasmid

MDA-MB-231 and A549 cell lines stably expressing shRNA against CD3EAP-mRNA were generated by lentiviral transduction. Following selection with puromycin (2 ug/ml), these cells were transfected with pDEST-CMV mCherry-GFP-LC3B WT construct (Addgene#123230) using Lipofectamine 3000 reagent (Invitrogen -2,173,180) according to manufacturer’s instructions. 24 hours post-transfection, cells were harvested and seeded in a fresh 96-well culture dish and subsequently selected with G418 at 300 ug/ml for 1 week. Following antibiotic selection, images of the punctas formed were acquired using Leica Fluorescence Microscopy.

SUnSET assay

To detect the relative changes in protein synthesis, SUrface SEnsing of Translation (SUnSET) assay was performed that measures the puromycin incorporated into nascent peptide chains using anti-puromycin antibody. After siCD3EAP knockdown, the cells were treated with 1 μM Puromycin (Sigma #P8833) and incubated at 37°C for 30 min. Post treatment, the cells were lysed using RIPA buffer (G Biosciences #786–490) and 50 μg protein was resolved using a 12% polyacrylamide gel. Further, proteins were transferred to a PVDF membrane (G Biosciences #GS-PVDF-302) and subsequently incubated in TBST buffer consisting of 5% Quick Blocker protein powder (G Biosciences #786–011) for two hours at room temperature. The PVDF membrane was cut and incubated overnight at 4°C with anti-Puromycin (DSHB #PMY-2A4), CD3EAP (Abclonal #A16099), GAPDH (ABclonal #AC002) antibodies. Following day, membranes were incubated for 1 Hour at room temperature in TBST buffer containing HRP tagged goat anti-mouse (Bio Rad #172–1011) or anti-rabbit (Invitrogen #31,460) secondary antibodies. All the membrane washing steps were performed four times at an interval of 10 min using TBS (HIMEDIA #ML029) along with 0.1% Polysorbate 20 (MP Biomedicals #103,168). Finally, to develop the image of the blot containing protein of interest, chemiluminescence detection method was performed using an ECL WesternBlotting substrate (Pierce #32,209) or femto LUCENT PLUSHRP (G Biosciences #786–003) reagents.

Results

Genomic aberration of RNA polymerase subunits (RNAPS) in six most common cancer types

To identify the mutation frequency of the RNAPS genes in human cancers, we used the TCGA mutation frequency data from cBioPortal. The analysis showed that BRCA and PRAD have low COAD, LUAD, and LUSC have high mutation frequencies of the RNAPS genes (Figure 1b). We searched for the genes with at least 5% mutation frequency and identified PLOR1A, POLR2A, POLR3A, POLR3B, and POLR1B as mutated genes in the majority of six tumor types (Figure 1c). Interestingly, although PRAD has a lower mutation frequency for RNAPS genes, PLOR1A, POLR2A, POLR3A, POLR3B, and POLR1B showed a higher mutation frequency in prostate cancer (Figure 1c). These genes were also frequently mutated in COAD, LIHC, LUAD, and STAD tumors (Figure 1c). However, the mutation frequency of these genes was less than 6% in any single tumor type, suggesting that RNAPS genes are required for transcription and rarely mutated.

Next, we tried to understand the copy number aberration in RNAPS genes. The analysis identified five genes, POLR2K, POLR2H, POLR3C, POLR3GL, and POLR3D, with a 5% or more CNV frequency (Figure 1d). POLR2K, POLR3C, and POLR3GL were most frequently amplified in BRCA, COAD, and LIHC (Figure 1d). In contrast, POLR2H was amplified in more than 45% of LUAD samples suggesting POLR2H is a crucial gene in LUAD tumors (Figure 1c). Next, to understand the effect of SCNV on the expression of these selected genes, we performed a correlation analysis between copy number and expression. The analysis showed that POLR2K, POLR2H, and POLR3D genes have a high correlation between SCNV and expression, suggesting that these genes are most frequently regulated by copy number variation (Figure 1d). Further detailed expression analysis showed that POLR2K is significantly overexpressed in amplified samples of BRCA, COAD, LIHC, LUAD, LUSC, and PRAD tumors compared to corresponding normal samples (Figure 1e). This analysis further confirms the CNV-based regulation of POLR2K gene in six tumor types studied here (Figure 1f) . Similarly, POLR2H and POLR3D genes also showed a significantly high correlation with CNV and correspondingly high expression in amplified tumor samples (Figure 1g,h). In contrast, POLR3GL and POLR3C did not show a strong correlation between CNV and expression, suggesting that these genes are not activated due to genomic aberrations (Figure 1d).

Many RNAPS are regulated by DNA methylation

DNA methylation plays a critical role in gene regulation. Generally, promoter hypermethylation is associated with the downregulation of gene expression. For DNA methylation analysis, we utilized TCGA illumina bead array data. For every RNAPS, multiple CpGs were probed in the bead array (Figure 2a). RNAPS were considered hypomethylated or hypermethylated if there was a two-fold difference between the tumor and normal in methylation value (beta value) as identified in Limma analysis at 5% FDR. The analysis identified the highest number of hypermethylated CpGs in BRCA and hypomethylated CpGs in LUSC associated with RNAPS genes (Figure 2b,c) . Next, to identify genes regulated by DNA methylation, we performed correlation analysis between the expression and methylation value of hypo- and hyper methylated CpGs and corresponding genes. We only considered genes regulated by methylation if at least three contiguous CpGs probes showed significant hypo or hypermethylation and <-0.15 correlation coefficient between methylation and expression. The analysis identified 11 CpGs probes associated with POLR2F in BRCA as hypermethylated with a strong negative correlation with expression (Figure 2d). Multiple hypomethylated CpGs associated with POLR2F also showed a strong negative correlation with expression in LUSC (Figure 2d). Further, we showed that expression of POLR2F is significantly less in hypermethylated BRCA samples compared to BRCA samples where the methylation level of these POLR2F CpG probes were similar to normal tissue (Figure 2e). Similarly, multiple probes associated with POLR2L showed hypermethylation and downregulation in LUAD and LUSC samples (Figure 2d,e). To confirm that POLR2F is regulated by DNA methylation in BRCA, we identified two BRCA cell line with low (MCF7) and high (T47D) expression of POLR2F (Figure 2f). The expression of POLR2F was then measured after 5-Aza-2-deoxycytidine (DAC) treatment. Confirming our initial finding, DAC treatment increased POLR2F expression in MCF7 (low expressing cells) but not in T47D (high expressing cells) (Figure 2f,g). Similarly, DNA methylation mediated expression inhibition of POLR2L was reversed after DAC treatment in H23 (low expressing cells) but not in A549 (high expressing cells) (Figure 2h,i). ZNRD1 and POLR1C were hypomethylated and overexpressed genes in LUAD and LIHC (Figure 2d,e). Interestingly, the POLR2F hypermethylation silenced gene in BRCA was overexpressed and hypomethylated in LUSC (Figure 2d,e) .

Figure 2.

DNA methylation analysis of RNAPS in cancer: a) the number of CpGs probed for each RNAPS was plotted and shown. b) the RNAPS CpGs (blue) with significantly increased methylation value in mentioned tumors compared to corresponding normal samples. The RNAPS genes (orange) with increased methylation compared to corresponding normal were also plotted. c) the RNAPS CpGs (blue) with significantly decreased methylation value in mentioned tumors compared to corresponding normal samples. The RNAPS genes (orange) with decreased methylation compared to corresponding normal were also plotted. d) each differentially methylated CpG was correlated with the expression of corresponding genes, and CpGs with negative correlation were selected and plotted. Genes with more than the three adjacent CpGs with negative correlation with expression were shown. e) the expression of genes with at least three contiguous hyper- or hypo-methylated CpGs were divided into two groups based on methylation compared to normal were plotted. The expression and methylation were compared using a t-test, and p-values were obtained. f) expression of POLR2F as measured by qRT-PCR. g) effect of DAC treatment on POLR2F in MCF7 and T47D. h) expression of POLR2L as measured by qRT-PCR. g) effect of DAC treatment on POLR2L in H23 and A549.

These observations suggest that RNAPS can be regulated differently in different tumor types and may have a contrasting function.

Some RNAPS are deregulated, and their expression is associated with survival in cancers

The deregulation of expression is a common phenomenon in cancer. To understand the expression changes in RNAPS in selected cancers, we compared the expression in tumors with corresponding normal using TCGA data. The LIHC showed the highest number of differentially expressed RNAPS genes. In contrast, there were no differentially expressed genes in PRAD (Figure 3a). Also, we noticed a higher number of overexpressed genes than under-expressed genes (Figure 3a). We also noticed that the majority of differentially expressed genes were deregulated in very few cancer types (Figure 3b,c). Only one RNAPS CD3EAP was overexpressed in all the cancers studied (Figure 3c). The POLR2F gene was downregulated in BRCA and COAD tumors (Figure 3c). The expression of all the deregulated genes in each cancer is shown in Figure 2c. Next, to identify the RNAPS whose expression has an association with patients’ survival, Cox regression analysis was performed. We identified 17 RNAPS as significant predictors of survival in LIHC (Figure 3d). In contrast, 3, 1, 3,0, and 1 RNAPS were significant predictors of survival in BRCA, COAD, LUAD, LUSC, and PRAD, respectively (Figure 3d). Interestingly, CD3EAP, the most commonly overexpressed gene, was also a prognostic gene in three tumor types LIHC, LUAD, and PRAD (Figure 3d).

Figure 3.

Expression analysis of RNAPS in cancer: a) the expression of RNAPS genes was compared in tumors with corresponding normal, and the number of upregulated and downregulated genes were plotted. b) the number of tumor types in which RNAPS genes were significantly differentially expressed was plotted. c) the expression of significantly differentially expressed RNAPS genes in each cancer types was plotted as a heatmap. d) each RNAPS gene expression was used for cox regression analysis, and the HR value with 95% CI of genes with significant correlation with survival was plotted as a forest plot.

CD3EAP is pan-cancer overexpressed gene and required for the cell growth

As we identified CD3EAP as one of the most overexpressed and prognostic genes in all the six cancer types, we studied the CD3EAP in function in further detail. CD3EAP is overexpressed in all six cancer types of studies in comparison to the corresponding normal (Figure 4a). We used ROC analysis and showed that CD3EAP RNA levels could be used as diagnostic markers for all six cancer types (Figure 4b). However, the diagnostic capabilities of CD3EAP were found to be highest in COAD and LUSC. Furthermore, we used Kaplan-Meier analysis to show that CD3EAP RNA expression is a strong prognostic predictor of overall survival in BRCA, LIHC, and LUAD and of progression-free survival in PRAD (Figure 4c). To understand the function of Cd3EAP in cancer, we checked the effect of CD3EAP on cell proliferation. We first studied the expression of CD3EAP in different cancer cell lines and identified MDAMB231 and A549 as cells with the highest expression of CD3EAP (Figure 4d). We used two shRNAs to knock down the CD3EAP expression in A549 and MDAMB231 cells and found that both shRNAs significantly decreased the expression of CD3EAP (Figure 4e). Next, the proliferation assay showed that cells with knockdown of CD3EAP significantly reduced cell proliferation rate in both MDAMB231 and A549 cells (Figure 4f). Colony suppression assay also showed that CD3EAP knockdown cells show reduced long-term cell growth (Figure 4g). Next, we used guilt-by-association analysis to understand the mechanism of CD3EAP-mediated growth regulation (Figure 4g). The genes correlated with CD3EAP expression were used for Metascape analysis. As expected, the most common enriched pathway in the analysis was the Metabolism of RNA (Figure 4h). However, interestingly, we also found significant enrichment of Cell cycle and Mitotic cell division as the top enriched pathway suggesting a role of CD3EAP in cell proliferation (Figure 4h). The analysis identified the enrichment of cell cycle genes as one of the significant pathways. Hence, we correlated the expression of CD3EAP with CDK and Cyclins. Interestingly, we identified a strong correlation between CD3EAP and S and G2/M phase CDK and cyclins (Figure 4i). We also checked the expression of these genes in CD3EAP knockdown samples and found a consistent decrease in CDK2, CCNE2, CDK1, and CCNB2 expression (Figure 4k). Interestingly, G1- CDK and cyclins did not show a consistent change in the expression (Figure 4k). These results suggest that CD3EAP knockdown is associated with decreased S and G2/M phase activity.

Figure 4.

CD3EAP is overexpressed and required for cell growth: a) expression of CD3EAP was compared for given tumor-normal pairs and plotted. P-values were obtained using a t-test. B) the expression of CD3EAP was used to construct a ROC plot for all the given types. The area under the curve is given against each tumor type. C) the patients of each cancer were divided into two groups based on the CD3EAP expression and plotted. The P-value was obtained using the Log-rank test. D) expression of CD3EAP was measured in given cell lines using qRT-PCR and plotted. e) the expression of CD3EAP was measured in a western blot after the knockdown of CD3EAP using two different shRnas. f) the proliferation of A549 and MDAMB231 cells was measured after CD3EAP knockdown and plotted. g) effect of CD3EAP knockdown on the long-term growth of cells was measured using colony suppression assay in MDAM231 and A549 cells – the representative images showing the colonies after two weeks. h) genes correlating with CD3EAP were used for metascape analysis, and the resulting network of GO terms was plotted. The groups of GO terms associated with a function are circled. i) the expression of CD3EAP was correlated with cell cycle genes in individual tumor types, and the correlation coefficient was plotted as a heatmap. J) the correlation plot of cell cycle genes with CD3EAP was constructed using all the tumor types studied. The correlation coefficient is also shown. k) the effect of CD3EAP knockdown was checked on the cell cycle gene with a high correlation with CD3EAP expression and plotted.

CD3EAP knockdown reduces ribosome biogenesis and induces autophagy

Further, to understand the reason for cell growth inhibition, firstly, we checked the localization of CD3EAP and found that it is majorly localized in the nucleus, confirming its role in transcription (Figure 5a). Next, we checked 18s and 28S rRNA levels in CD3EAP knockdown cells. Interestingly, we found that CD3EAP knockdown cells have consistently decreased in 18s and 28S rRNA levels, suggesting inhibition of RNA polymerase I mediated transcription (Figure 5b). The reduced rRNA levels are associated with reduced protein synthesis, hence, we checked the effect of CD3EAP knockdown on total protein level. Interestingly, we did not find any change in total protein level after CD3EAP knockdown (Figure 5c). However, as changes in total protein level were not the direct measurement of reduced protein synthesis, we performed surface sensing of translation (SUnSET) assay where puromycin incorporation is measured as indicator of active protein synthesis. In both the cell lines, we found the decreased puromycin incorporation after CD3EAP knockdown using siRNA (Figure 5d). The protein synthesis inhibition was more drastic in MDA MB 231 cells. This observation suggests that CD3EAP knockdown inhibits protein synthesis.

Figure 5.

CD3EAP knockdown induces autophagy: a) the A549 cells were fractionated, and level off CD3EAP and H3 histone and GAPDH was checked in whole cell lysate (WC), cytoplasmic (cyt.) and nuclear (Nucl.) fraction. b) qRT-PCR was performed to check the effect of CD3EAP knockdown on 18S and 28S RNA. c) effect of CD3EAP knockdown on the total protein was checked using Coomassie staining. d) SUnSET assay results are shown for A549 and MDAMB231 after CD3EAP knock down. e) the western blot was performed to check the effect of CD3EAP knockdown on given proteins in A549 cells F) and in MDAMB231 cells. g) the effect of CD3EAP on cell death was measured using PI staining and quantified. h) A549 and MDAMB231 cells were transfected with mentioned plasmids and CD3EAP knockdown was performed. Fluorescent imaging was done to show the increased LC3 puncta. i) A549 cells were transfected with mentioned plasmids and CD3EAP knockdown was performed. Fluorescent imaging was done to show the increased autophagy flux. The ratio was calculated and plotted. j) MDA MB 231 cells were transfected with mentioned plasmids and CD3EAP knockdown was performed. Fluorescent imaging was done to show the increased autophagy flux. The ratio was calculated and plotted. k) CD3EAP knockdown and control A549 cells were treated with 3-MA or vehicle control, and the effect on proliferation was measured. l) A549 cells were transfected with various combination of siRnas and expression of ATG5, CD3EAP and TPT1 was measured. The western blot was done to measure the effect on CD3EAP, LC3 and GAPDH. m) cell proliferation of cell was measured after knock down of ATG5, CD3EAP or both using incucyte.

Inhibition of RNA pol I mediated transcription and protein synthesis is associated with increased apoptosis and autophagy. We checked if the removal of CD3EAP induces apoptosis. Interestingly, we did not find appreciable change in Caspase 3 or cleaved PARP level, suggesting there is no apoptosis in CD3EAP knockdown A549 and MDA MB 231 cells (Figure 5e,f). However, PI staining showed significant increase in the cell death after CD3EAp knockdown (Figure 5g) . Next, we checked the autophagy status in CD3EAP knockdown A549 and MDAMB231 cells. CD3EAP knockdown A549 and MDAMB231 cells showed increased autophagy which was evident by increased LC3II/LC3I ratio (Figure 5e,f). The increased autophagy level after CD3EAP knockdown in A549 and MDAMB231 was also measured using LC3 puncta formation (Figure 5h). We also used GFP-LC3-RFP-LC3ΔG plasmid to show the increased autophagy flux in A549 and MDAMB231 cells after CD3EAP knockdown (Figure 5i,j).

Now to understand if the increased autophagy is cytoprotective or growth suppressing, we inhibited the autophagy in CD3EAP knockdown and control cells and checked the cell viability. The results showed that CD3EAP-mediated proliferation inhibition was partially rescued after 3-MA treatment (Figure 5k). As 3-MA is not a specific autophagy inhibitor, we also used ATG5 knockdown in presence and absence of CD3EAP, to show that cell proliferation changes induced after CD3EAP knockdown was completely rescued after ATG5 knockdown mediated autophagy inhibition (Figure 5l,m).

POLR2D is localized in the cytoplasm and required for cell growth and migration

Another RNAPS we found as prognostic in more than 50% of tumor types was POLR2D. Hence, we also studied POLR2D in further detail. Expression analysis showed that POLR2D is overexpressed in all the cancer types except prostate cancer when compared to the corresponding normal (Figure 6a). Using ROC analysis, we showed that POLR2D RNA level is a strong diagnostic marker for LUSC and COAD tumors (Figure 6b). Next, we divided the different cancer patients into the high and low-expression groups of POLR2D and analyzed the survival using Kaplan-Meier analysis. The analysis showed that POLR2D is associated with poor overall survival of STAD, BRCA, LIHC, and LUAD and poor progression-free survival of PRAD patients (Figure 6c). Next, we checked the effect of POLR2D knockdown to understand the role of POLR2D in cell growth. The qRT-PCR data identified A549 and MDAMB231 with the highest expression of the POLR2D gene (Figure 6d). Hence, we used two shRNAs to silence the expression of POLR2D in A549 and MDA MB 231 cells. Both the shRNAs significantly reduced the expression of the POLR2D gene (Figure 6e). Cell proliferation assay showed that the knockdown of POLR2D inhibited cell growth of both A459 and MDA MB 231 cells (Figure 6f). Colony suppression assay showed the POLR2D knockdown also suppresses the long-term growth of the cells (Figure 6g).

Figure 6.

POLR2D is required for the cell’s survival. a) expression of POLR2D was compared for given tumor-normal pairs and plotted. P-values were obtained using a t-test. b) the expression of POLR2D was used to construct a ROC plot for all the given types. The area under the curve is given against each tumor type. c) the patients of each cancer were divided into two groups based on the POLR2D expression and plotted. The P-value was obtained using the Log-rank test. d) expression of POLR2D was measured in given cell lines using qRT-PCR and plotted. e) the expression of POLR2D was measured in a western blot after the knockdown of POLR2D using two different shRnas. f) the proliferation of A549 and MDAMB231 cells was measured after POLR2D knockdown and plotted. g) effect of POLR2D knockdown on the long-term growth of cells was measured using colony suppression assay in MDA MB 231 and A549 cells. The representative images show the colonies after two weeks. h) the A549 cells were fractionated, and level off CD3EAP and H3 histone and GAPDH was checked in whole cell lysate (WC), cytoplasmic (cyt.), and nuclear (Nucl.) fraction. i) the expression of given genes was checked after POLR2D knockdown using semi-quantitative PCR.

Next, to understand the mechanism of POLR2D function, we first check the localization of POLR2D proteins. Interestingly, we found that POLR2D is localized both in the cytoplasm and nucleus; however, the localization was much higher in the cytoplasm (Figure 6h). As POLR2D is a subunit of RNA polymerase II, we checked the level of randomly selected ten genes in control and POLR2D knockdown cells. To our surprise, POLR2D is not associated with the transcription of all the genes; only a subset of genes is affected by POLR2D knockdown (Figure 6i). Suggesting POLR2D is involved in the transcription of a set of genes and may cause POLR2D-mediated cell growth inhibition.

Discussion

Many studies have associated the link between RNA polymerase and cancer development and progression. Here, we have performed the genomic and epigenomic analysis of the TCGA data to characterize the RNA polymerase subunits in cancer. Except for some RNAPS, there was no recurrent mutation found in this group of genes. It is an expected observation as transcription is the core process of cell physiology, and mutation in RNAPS genes can cause severe transcriptional defects. In contrast, many RNAPS genes showed significant variation in the copy number. Specifically, POLR2H was found to be amplified in more than 40% of LUAD samplers. POLR2H expression also showed a high correlation with CNV in all cancer studies; however, in LUSC tumors, the POLR2H was specifically overexpressed due to amplification. POLR2H expression has been associated with lung cancer survival [25]. We also identified POLR3D as the only deleted gene with low expression in LUSC.

DNA methylation changes are common mechanisms of expression regulation in cancer. We used Illumina methylation array data from TCGA to identify the RNAPS deregulated in cancer due to DNA methylation. As multiple CpGs are checked for each gene, we considered a gene regulated by methylation if at least three contiguous probes showed significant change in methylation and showed a correlation coefficient of <-0.15. We identified POLR2F as one of the hypermethylated and silenced RNAPS genes in BRCA. POLR2F was not repressed in all the BRCA samples; only samples with high methylation at promoter showed the silencing of this gene. This observation suggests that POLR2F silencing may be a critical step in some BRCA patients. Interestingly, POLR2F was hypomethylated and overexpressed in LUSC, which indicates a more complex function of POLR2F in cancer. POLR2F is a prognostic marker in COAD [26]. POLR2F is also downregulated in glioma and differentially expressed in Erlotinib and Vinorelbine sensitive NSCLC cells [27,28]. POLR2L was another RNAPS silenced due to methylation in NSCLC (LUAD and LUSC) samples. POLR2L has been shown to be associated with multiple cancer types [29,30]. However, this is the first report showing the methylation-mediated downregulation of an RNAPS. We have also identified ZNRD1 and POLR1C as hypomethylated and overexpressed in more than two tumor types. ZNRD1 has been shown to function as a growth promoter gene [31,32]. ZNRD1 is overexpressed in drug-resistant cancer cells, and its depletion could reduce the resistance [33]. The role of POLR1C mutation is well established in the Treacher-Collins syndrome [34]. POLR1C is also reported as a prognostic marker for triple-negative breast cancer [34]. ZNRD1 and POLR1C may prove to be crucial for carcinogenesis. In this work, we report CD3EAP, an RNA Pol I subunit, as one of the most commonly overexpressed RNAPS in cancer. Various studies have shown that CD3EAP polymorphism is associated with lung cancer risk in the Chinese population [35]. Another polymorphism associated with CD3EAP is associated with drug sensitivity [36]. Here, we found that CD3EAP is a prognostic marker in multiple cancer types. CD3EAP is also required for cell growth. The expression of the CD3EAP gene was associated with S and G2 phase cell cycle genes. Earlier studies have shown that RNA Pol I inhibition induces G2/M arrest [37,38]. Our observation validates that RiBi inhibition induces cell cycle arrest in the G2 phase. Interestingly, CD3EAP knockdown induced autophagy, but there was no effect on cleaved PARP and Caspase 3 levels. In contrast, inhibition of RNA pol I transcription increases caspase-dependent apoptosis [37]. We observed that CD3EAP knockdown induces autophagy-related cell growth inhibition. Autophagy induction is also a known phenomenon of RNA pol I transcription inhibition [39]. Multiple studies have demonstrated the crucial role of autophagy in regulating the synthesis of ribosomal RNA (rRNA), highlighting its significance in cellular processes [40,41]. These observations suggest that inhibition of RNA pol I machinery causes cell growth inhibition by activating the autophagy pathway. POLR2D, a subunit of RNA Pol II, is also found to be overexpressed and prognostic in multiple cancer types. The function of POLR2D is dispensable for yeast but not for Zebrafish [42]. We showed in this work that POLR2D is required for cell survival. Interestingly, we found that POLR2D is essential but required for the transcription of selected genes only. Taken together, we have performed a Pan-cancer analysis to identify RNAPS with genetic and epigenetic aberrations. We also identified POLR2F and POLR2L as DNA Methylation suppressed genes in BRCA and NSCLC. We showed that the RNA pol I subunit CD3EAP is upregulated in cancer, and its inhibition leads to a similar effect as pharmacological inhibition of RNA pol I. We found that POLR2D, the RNA Pol II subunit, is also overexpressed in cancer and required for cell survival.

Funding Statement

This work was funded by a research grant from the Indian Council of Medical Research, Govt. of India (NO. 2020-3806/CMB/ADHOC-BMS and 2021-9513/CMB/ADHOC-BMS).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Author Contribution

Nikita Bhandari: Designed and Performed Experiments, Wrote the manuscript

Disha Acharya: Designed and Performed Experiments

Annesha Chatterjee: Performed Experiments

Lakshana Mandve: Performed Experiments

Pranjal Kumar: Performed Experiments

Shreesh Pratap: Designed and Performed Experiments

Pushkar Malakar: Designed experiments

Sudhanshu K Shukla: Conceptualization, resources, supervision, writing (reviewing and editing),

Data availability statement

All the data used in this study are publicly available.