A pan-cancer analysis of the oncogenic role of ribonucleotide reductase subunit M2 in human tumors

doi:10.21203/rs.3.rs-2215057/v1

Download PDF

Article

A pan-cancer analysis of the oncogenic role of ribonucleotide reductase subunit M2 in human tumors

https://doi.org/10.21203/rs.3.rs-2215057/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Although there is evidence that ribonucleotide reductase subunit M2 (RRM2) is associated with many cancers, no pan-cancer analysis is available. This study aims to explore the potential oncogenic role of RRM2 in pan-cancer based on the datasets of TCGA (The cancer genome atlas). RRM2 was highly expressed in most cancers and obvious correlation existed between RRM2 expression and prognosis of tumor patients. RRM2 expression was associated with infiltration level of diverse immune cell as well as endothelial cells, immune checkpoint, tumor mutational burden (TMB) and microsatellite instability (MSI). Moreover, the cell cycle pathway was involved in the functional mechanisms of RRM2. Our pan-cancer study provides a fairly comprehensive understanding of the oncogenic roles of RRM2 across different tumors.

Biological sciences/Cancer

Biological sciences/Cell biology

Biological sciences/Immunology

Health sciences/Oncology

RRM2

Pan-cancer

TCGA

Prognosis

Tumor immunity

Given the complex causes of tumorigenesis, such as genetic loss or alteration(1) and immune infiltration(2), it is crucial to perform a pan-cancer analysis of genes that may have potential clinical value. Recently, with a deeper research of the initiation, development and treatment of tumors, many therapeutic concepts emerged, such as, immune checkpoint proteins(3) and immune infiltration(4). Inspiringly, the publicly funded TGCA project and many powerful bioinformatic tools allow us to perform pan-cancer analysis.

Ribonucleotide reductase subunit M2 (RRM2), a subunit of ribonucleotide reductase, supports DNA synthesis and repair by catalyzing the formation of deoxyribonucleotides(dNTP)(5). RRM2 governs tumor cell nucleotide metabolism(6) and is associated with multiple cancers(7, 8). Many studies have shown that RRM2 is a potential diagnostic or prognostic biomarker for Ewing sarcoma(9), liver cancer(10), glioma(11) and lung adenocarcinoma(12). Moreover, RRM2 is an important target of many factors or drugs for suppressing tumorigenicity in diverse cancers(13–16). For example, DHS (trans-4,4'-dihydroxystilbene) suppresses DNA replication and tumor growth by inhibiting RRM2 in pancreatic, ovarian, and colorectal cancer(17).

Although the oncogenic role of RRM2 has been uncovered in several tumors, there is still no pan-cancer evidence about the relationship between RRM2 and a variety of cancer types on the basis of big clinical data. Here, we carried out a pan-cancer analysis of RRM2. And we explored the molecular mechanism of RRM2 in different cancers by a series of factors, including gene expression, clinical correlation, survival analysis, genetic alteration, immune infiltration, immune checkpoint, enrichment analysis and potential drug prediction.

1. The expression of RRM2 in pan-cancer

To explore RRM2 expression in different cancers, the online Oncomine (https://www.oncomine.org/) was performed using the default parameters. Additionally, we downloaded gene expression RNAseq data of 33 cancers from the TCGA dataset in UCSC Xena(18). Perl language (https://www.perl.org/) was used for extracting RRM2 expression data in 33 cancers. The R language was performed for further analysis and visualization.

2. RRM2 expression and its clinical correlation in pan-cancer

To explore the correlation of RRM2 expression with tumor patients’ clinical characteristics (age, race, gender, tumor stage and status), we downloaded the clinical data of the pan-cancer from the TGCA database (https://www.cancer.gov/). R language with limma (version 3.38.3)(19) was used for visualization.

3. Survival prognosis analysis Cox regression models

The survival data of 33 cancers were downloaded from the TGCA database for survival analysis. The Kaplan-Meier (KM) and Cox regression models were applied to explore RRM2 potential prognostic value for pan-cancer. For estimating the prognostic value of RRM2 comprehensively, we performed Overall survival (years) (years) (OS), Disease-free interval (years) (DFI), Disease-specific survival (years) (DSS) and Progression-free interval (years) (PFI) by R language with limma, survival (version 3.1.12), forestplot (version 1.9) and survminer packages (version 0.4.7).

4. Genetic alteration analysis

We selected the “TCGA Pan Cancer Atlas Studies” in the “quick select” section and entered “RRM2” for queries of the genetic alteration characteristics of RRM2 in the cBioPortal web (https://www.cbioportal.org/). In the “Cancer Types Summary” module, we got the alteration frequency, Mutation data (the type of mutation tumors) and CNA data (Copy number alteration) of RRM2 in all TGCA tumors. In the “Mutations” module, the mutated sites information of RRM2 was visualized in the schematic diagram of the protein structure or the 3D structure. Meanwhile, we performed the R language with fmsb package to explore the correlation between RRM2 expression and tumor mutation in 33 TGCA cancers downloaded from the TGCA database.

Tumor mutational burden (TMB) is a biomarker which can present the total number of mutations in a tumor(20). Microsatellite instability (MSI) can reflect the loss or gain of nucleotides from repetitive DNA tracts, is a diagnostic phenotype of many cancers(21). We explored the correlation between RRM2 expression and TMB as well as MSI in pan-cancer using the Perl and R language with fmsb package (version 0.7.0) .

5. Immune infiltration analysis

we chose the “Immune” module of the TIMER2 webserver(22), then RRM2 and endothelial cell were selected for exploring the association between immune infiltrates of endothelial cell and RRM2 expression in all TCGA. The partial Spearman’s correlation was performed with “Purity Adjustment” option. We further investigated the correlation between RRM2 expression and the infiltration level of other 22 immune cells. The R language with ggplot2 (version 3.3.2), ggpubr and ggExtra (version 0.9) packages were performed to analyze and visualize.

6. Immune checkpoint analysis

Immune checkpoint molecules are co-stimulatory receptors on immune cells, and can regulate immune cells by immune response(23). Herein, we used R language with limma, reshape2 and RColorBrewer packages to analyze the relationship between RRM2 expression and the expression level of 47 common immune checkpoint genes in pan-cancer.

7. Enrichment analysis of SND1-related partners

In the STRING website (https://string-db.org/), we input protein name (“RRM2”), organism (“Homo sapiens”) in the “Protein by name” module. Then, we changed the basic settings: meaning of network edges: evidence, active interaction sources: Experiments, minimum required interaction score: low confidence (0.150) and max number of interactors to show: no more than 50 interactors in 1st shell. Eventually, the RRM2-binding proteins which had been authenticated by experiments were acquired. Cytoscape (3.6.1) was used for integrating and visualizing(24). In the “Similar Genes” section of GEPIA2 (http://gepia2.cancer-pku.cn/), we searched RRM2-correlated targeting genes by Gene (“RRM2”), Top # (“100”) similar Genes and all TGCA tumor and normal tissues. Meanwhile, we used the “correlation analysis” module to explore the Pearson correlation between RRM2 expression and the top 5 RRM2-correlated targeting genes in all TGCA tumor and normal tissues. The “Gene_Corr” module in TIMER2 was used to show the correlation between RRM2 with the top 5 RRM2-correlated targeting genes in various cancer types. Subsequently, we performed an intersection analysis to compare the genes that RRM2 binds and interacts with by Jvenn(25). Furthermore, we performed Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis for the two sets of data that RRM2-binding and correlated through R language with clusterProfiler, ggplot2 and enrichplot packages.

8. Potential drug prediction for pan-cancer

The KEGG gene sets as Gene Symbols were downloaded in GESE online dataset (https://www.gsea-msigdb.org/). Then, the R language with colorspace, stringi, ggplot2, limma and enrichplot packages was running to explore the RRM2 related KEGG enrichment analyses in pan-cencer. Finally, the enrichment pathways were counted for further drug prediction. In the Therapeutic Target Database (http://db.idrblab.net/), we chose the “Pathway Search” section of “Advanced Search” module(26). Then, step1: KEGG and step2: CELL CYCLE were chosen to obtain pathway associated drugs that are related to RRM2 expression in pan-cancer. And, only approved drugs were screened to show.

1. The expression of RRM2 in pan-cancer

In this study, to explore the oncogenic role of human RRM2, we first analyzed the expression pattern of RRM2 in different cancer types using the Oncomine database (Fig. 1A). The results showed that RRM2 was significantly increased in most cancer groups, including bladder, brain, breast, cervical, colorectal, esophageal, gastric, head and neck, kidney, liver, lung, lymphoma, melanoma, other, ovarian, pancreatic, prostate and sarcoma cancers. While the expression of RRM2 was decreased in leukemia. Further, we identified the expression levels of RRM2 in 33 different tumor and nontumor tissues based on the datasets of TCGA. As shown in Fig. 1B, the expression levels of RRM2 were extremely different between cancer and normal groups. The expression levels of RRM2 were significantly higher in 19 tumor tissues than the corresponding normal tissues, including BLCA (Bladder Urothelial Carcinoma), BRCA (Breast invasive carcinoma), CESC (Cervical squamous cell carcinoma and endocervical adenocarcinoma), CHOL (Cholangiocarcinoma), COAD (Colon adenocarcinoma), ESCA (Esophageal carcinoma), GBM (Glioblastoma multiforme), HNSC (Head and Neck squamous cell carcinoma), KIRC (Kidney renal clear cell carcinoma), KIRP (Kidney renal papillary cell carcinoma), LIHC (Liver hepatocellular carcinoma), LUAD (Lung adenocarcinoma), LUSC (Lung squamous cell carcinoma), PCPG (Pheochromocytoma and Paraganglioma), PRAD (Prostate adenocarcinoma), READ (Rectum adenocarcinoma ), STAD (Stomach adenocarcinoma), THCA (Thyroid carcinoma), and UCEC (Uterine Corpus Endometrial Carcinoma). Overall, these results indicated that RRM2 was highly expressed in most cancers and showed a potential oncogenic role.

2. RRM2 expression and clinical correlation in pan-cancer

We assessed the correlation between the expression of RRM2 and tumor patients’ clinical characteristics in pan-cancer, including age, race, tumor stage and status. For age, a significant correlation was observed in BRCA, ESCA KICH (Kidney Chromophobe), KIRC, KIRP, LAML (Acute Myeloid Leukemia), LGG (Brain Lower Grade Glioma), LIHC, LUAD, LUSC, PAAD, PCPG, READ, STAD, TGCT (Testicular Germ Cell Tumors) and THYM (Thymoma) (Fig. 2A-D and Fig. S1A-L). For race, RRM2 expression significantly correlated with BLCA, BRCA, KICH, KIRC, KIRP, LIHC and THYM (Fig. 2E-H and Fig. S1M-O). Regarding the tumor stage, RRM2 expression was significantly associated with ACC (Adrenocortical carcinoma), BRCA, COAD, KICH, KIRC, KIRP, LIHC, LUAD, LUSC, PAAD (Pancreatic adenocarcinoma), SKCM (Skin Cutaneous Melanoma), TGCT and THCA (Fig. 2I-L and Fig. S1P-X). In terms of tumor status, RRM2 expression levels significantly correlated with ACC, BLCA, COAD, KICH, KIRC, KIRP, LGG, LUAD, OV (Ovarian serous cystadenocarcinoma), PAAD, PCPG, PRAD and UVM (Uveal Melanoma) (Fig. 2M-P and Fig. S1Y-AG).

3. Survival analysis

For estimating the correlation between RRM2 expression with the prognosis of pan-cancer, we divided the tumor cases into two groups by the high- and low-expression levels of RRM2.

Kaplan-Meier method was applied to analyze Overall survival (years) (years) (OS), Disease-free interval (years) (DFI), Disease-specific survival (years) (DSS) and Progression-free interval (years) (PFI), respectively. As shown in Fig. 3A-E and Fig. S2A-I, OS analysis data showed high RRM2 expression was associated with poor prognosis for the TCGA cases of ACC(P < 0.001), KICH(P = 0.015), KIRC(P < 0.001), KIRP(P < 0.001), LGG(P < 0.001), LIHC(P = 0.005), LUAD(P < 0.001), MESO (Mesothelioma) (P < 0.001), PAAD(P = 0.006), PRAD(P = 0.004), SARC (Sarcoma)(P = 0.04), UCEC(P = 0.021) and UVM(P < 0.001). However, a low RRM2 expression was related to poor OS for THYM(P = 0.017). DSS analysis data indicated that a correlation between high RRM2 expression and poor prognosis for the TGCA cases of ACC(P < 0.001), KICH(P = 0.001), KIRC(P < 0.001), KIRP(P < 0.001), LGG(P < 0.001), LIHC(P = 0.028), LUAD(P = 0.004), MESO(P < 0.001), PAAD(P = 0.014), PRAD(P = 0.03) and UVM(P < 0.001) (Fig. 3F-J and Fig. S2J-O). For DFI (Fig. 3K-O and Fig. S2P-R), high RRM2 expression was linked to poor prognosis for cancers of KIRP(P = 0.007), LIHC(P = 0.038), LUAD(P = 0.005), PAAD(P = 0.022), SARC(P < 0.001), TGCT(P = 0.043) and THCA(P = 0.001). Additionally, a low RRM2 expression was related with poor DFI for OV(P = 0.022). For PFI analysis data (Fig. 3P-T and Fig. S2SA-B), the high RRM2 expression showed a correlation with poor prognosis for cancers of ACC(P = 0.002), KICH(P = 0.007), KIRC(P = 0.025), KIRP(P < 0.001), LGG(P = 0.001), LIHC(P = 0.003), LUAD(P = 0.003), MESO(P < 0.001), PAAD(P = 0.002), PRAD(P < 0.001), SARC(P = 0.001), THCA(P = 0.011) and UVM(P < 0.001). However, the low RRM2 expression was linked to poor PFI for cancers of COAD(P = 0.049) and STAD(P = 0.05).

Further, we calculated the survival data by the Cox regression models (Fig. 3U-X). The results of OS, DSS and PFI were similar to that we calculated by Kaplan-Meier method. The difference for OS was the high RRM2 expression did not show a significant correlation with poor prognosis for cancers of UCEC. Besides, high RRM2 expression was linked to poor DSS for SARC(P = 0.024). For PFI, the high RRM2 expression showed a significant correlation with poor prognosis for cancer of PCPG(P = 0.006) and did not show a significant correlation with poor prognosis for STAD. For DFI, the data indicated that the high RRM2 expression was associated with the poor prognosis for cancers of BRCA(P = 0.036), KIRP(P < 0.001), LIHC(P = 0.021), LUAD(P < 0.001), PAAD(P = 0.012), PRAD(P = 0.020), SARC(P = 0.002) and THCA(P < 0.001). These data suggested that RRM2 was differentially associated with the prognosis of cases with different cancers.

4. genetic alteration analysis data

Next, we explored the genetic alteration status of RRM2 in 10953 patients / 10967 samples in 32 studies of the TGCA cohorts. The analysis data showed that RRM2 has the highest alteration frequency (> 7%) in patients with uterine carcinosarcoma, among which "amplification" is the main type. Notably, the genomic alteration types of all DLBCL (Diffuse large B cell lymphomas) and KICH cases were deep deletion of RRM2(Fig. 4A). Then, we investigated the types, sites and case numbers of RRM2 mutation (Fig. 4B). We found that the primary genetic mutation types of RRM2 was missense mutation. R298Q/W mutation in the ribonucleotide reductase, which was detected in 3 cases of UCEC and 1 case of GBM, was able to product a missense mutation of the RRM2 gene, translation from R(Arginine) to W(Tryptophan) or Q(Glutamine) at the 298 site of RRM2 protein. We visualized the R298 site in the 3D structure of RRM2 protein (Fig. 4C).

Moreover, we explored the correlation between RRM2 expression and TMB (tumor mutational burden)/ MSI (microsatellite instability) in all cancers of TCGA. As shown in Fig. 4D, RRM2 expression had a positive correlation with TMB for ACC, BLCA, BRCA, CESC, CHOL, COAD, KICH, KIRC, LGG, LIHC, LUAD, LUSC, MESO, OV, PAAD, PRAD, READ, SARC, SKCM and STAD but a negative correlation for THYM. As shown in Fig. 4E, we found a positive correlation between RRM2 expression and MSI for cancers of COAD, LIHC, SARC, STAD, TGCT, UCEC, and UCS (Uterine Carcinosarcoma) but found a negative correlation for LAML and SKCM. The results deserved more in-depth analysis and induction.

5. immune infiltration

Immune cells, as primary components of tumor microenvironment (TME), were related with the progression and metastasis of tumors and therapy resistance(23, 27). Tumor-associated endothelial cells were reported to play a key role in sculpting immune responses, which was necessary for tumor out-growth and metastasis(28). Firstly, we explored the association between RRM2 gene expression and the infiltration level of immune cells in cancers of TCGA by the TIMER, CIBERSORT, CIBERSORT-ABS, QUANTISEQ, XCELL, MCPCOUNTER and EPIC algorithms. As shown in Fig. 5A, the RRM2 expression level in BRCA, KIRC, LUAD, LUSC, STAD, THCA and THYM was negatively correlated with the infiltration level of tumor-associated endothelial cells, but noted a positive correlation for KIRP and LGG. Then, we analyzed the correlation between RRM2 expression level and the immune infiltration of different tumor-associated cells (B cells, Plasma cells, T cells, NK cells, Monocytes, Macrophages, Dendritic cells, Mast cells, Eosinophils and Neutrophils) as well as their subtypes in the tumors of TGCA. The results showed that RRM2 expression level significantly correlated with the immune infiltration of various tumor-associated cells in 24 tumor types. For example, RRM2 expression level in ACC was negatively correlated with the infiltration level of Mast cells resting (Fig. 5B). In BLCA, we observed that RRM2 expression significantly correlated with the infiltration of 5 types of immune cells (Fig. 5C-G). In BRCA, RRM2 expression significantly correlated with the infiltration of 10 types of immune cells (Fig. 5H-Q). In other cancers, the correlations between RRM2 expression and immune cells’ infiltration are shown in Fig. S3A-CI.

6. Immune checkpoint analysis

We explored the correlation between RRM2 expression level and immune checkpoint genes in various cancers (Fig. 6). The results showed that the expression levels of more than 40 immune checkpoint genes were significantly associated with RRM2 expression level in TGCT and THCA. Moreover, the expression level of up to 30 immune checkpoint genes in BRCA, HNSC, KIRC, KIRP, LGG, LIHC, PRAD, THYM and UVM were related to RRM2 expression level. Additionally, there were more than 20 different immune checkpoint genes related with RRM2 expression level in 10 types of cancers.

7. Enrichment analysis of SND1-related partners

To deeply understand the molecular mechanism of the RRM2 gene in tumorigenesis, we tried to search the RRM2 expression-correlated targeting genes and the RRM2-binding proteins for functional enrichment analyses. We acquired 50 RRM2-binding proteins which had been authenticated by experiments using STRING. The protein-protein interaction network was shown in Fig. 7A. Then we explored the top 100 RRM2 expression-correlated genes based on datasets of TGCA by the GEPIA2. The results showed that the expression level of RRM2 was positively associated with that of MKI67 (Marker of proliferation Ki-67) (R = 0.79), ORC1 (Origin recognition complex subunit 1) (R = 0.78), CCNA2 (Cyclin A2) (R = 0.77), PLK1 (Polo like kinase 1) (R = 0.77) and KIF11 (Kinesin family member 11) (R = 0.76) (Fig. 7B). The heatmap indicated a positive correlation between RRM2 and the four genes in virtually all the cancer types (Fig. 7C). Intersecting the above two groups demonstrated 4 common genes: PLK1, CDK1 (Cyclin dependent kinase 1), DTL (Denticleless E3 ubiquitin protein ligase homolog) and ASF1B (anti-silencing function 1B histone chaperone) (Fig. 7D). The KEGG enrichment analyses were performed based the genes in the two groups. The result indicated that “cell cycle” might play a crucial role in the effect of RRM2 on tumor pathogenesis (Fig. 7E).

8. Potential drug prediction for pan-cancer

We performed GSEA to explore the pathways that RRM2 regulates in pan-cancer (Fig. S4). The results showed that the pathway of cell cycle was the most significantly affected by RRM2 expression in pan-cancer, which was the same to KEGG enrichment analyses above (Table S1). The drug prediction analysis showed that up to 14 types of approved drugs, such as Phenylbutyrate and Romidepsin, may have therapeutic effect on pan-cancer through targeting the cell cycle-associated pathways (Table S2).

Consistent with the extremely important biological process of RRM2 in DNA replication(29), RRM2 has been shown to exhibit a functional link with diverse tumors(7–9). Whether RRM2 can involve in the pathogenesis of multiple tumors through one common molecular mechanism remains to be explored. Through a bibliography retrieval, we failed to find any publication with a pan-cancer analysis of RRM2 based on the perspective of overall tumors. So, in this study, we thoroughly uncovered for the first time that the role of RRM2 gene in 33 different cancers. We showed that the expression of RRM2 in up to 19 tumor tissues was higher than that in the corresponding normal tissues by bioinformatics analysis of the Oncomine and TCGA public databases. These findings demonstrated that RRM2 plays an extremely important role in most tumors (Fig. 1). Subsequently, the significance of RRM2 in the clinical prognosis of pan-cancer was explored. Expression of RRM2 was significantly related with age, race, tumor stage and status in nearly one-third of cancers in TGCA (Fig. 2). Additionally, survival analyses of Kaplan-Meier and Cox regression models showed that most cancer patients with high RRM2 expression presented a noticeable worse survival rate than those with low expression (Fig. 3). In the light of the patient samples in the cBioPortal database, RRM2 gene mutations appeared in 23 kinds of tumor patients, the alteration frequency is 0.2–7.02%. Meanwhile, we unearthed the types that most of the alterations of RRM2 are gene mutation, amplifications and deep deletion (Fig. 4). We also exhibited evidence of the close association between RRM2 expression and TMB or MSI in all TCGA cancers. These results demonstrated that RRM2 may be an extremely powerful prognostic biomarker in pan-cancer and may expedite the development of precise targeted therapy for tumors.

RRM2 was highly expressed in most tumors. However, given that the limits of online databases and tools, the expression of RRM2 in some tumors were omitted. Through literature search, we supplemented these missing meaningful results. In Adrenocortical carcinoma (ACC), Tabbal et al. showed that high expression of RRM2 induced by EZH2 was associated with poor prognosis in ACC(30). Inhibition of RRM2 could block cell proliferation and cell migration, induce apoptosis, suggesting that RRM2 may be a target in ACC(30). In Acute Myeloid Leukemia (LAML), Aimiuwu et al. demonstrated that 5-Azacytidine (5-azaC) could cause a perturbation of deoxyribonucleotide pools and RRM2 may be a novel molecular target of 5-azaC in AML(31). In Mesothelioma, Wendy et al. showed that RRM2 was highly expressed in MESO by immunohistochemistry semi-quantitative analysis in 70 patients, but no significant correlation was found between RRM2 expression and clinical prognosis of OS and PFS(32). As a result, the current data showed us the strong relationship between RRM2 and up to 20 types of tumors, but further research is still needed to explore whether RRM2 is related to the rest of the tumors. Consistent with this result, our clinical data correlation analysis also showed that RRM2 has a great correlation with the age, race, tumor stage and status of cancer patients. These results give us great confidence to believe that RRM2 plays an important role in pan-cancer and has clinical value worthy of in-depth exploration.

In evaluating the survival analysis of RRM2 in tumors, we used Cox regression models and KM method. However, in some tumors, the results of the two algorithms were inconsistent. In this work, we used a Cox regression survival analysis of data in the TCGA- UCEC cohort and observed a statistical correlation between high expression of RRM2 and poor OS prognosis (P = 0.021). Nevertheless, no significant correlation was found between RRM2 expression and OS for UCEC by the Kaplan-Meier method(P = 0.133). Different data processing or updated survival data information may cause this result. Therefore, we used another web tool, GEPIA2(33), and failed to detect a correlation between RRM2 expression and Overall survival (years) prognosis of UCEC patients (P = 0.064). Consequently, the current evidence still cannot confirm the role of SND1 expression in the OS prognosis of UCEC. For other tumors with nonidentical results in DFI, DSS and PFI, larger sample sizes and clinical studies are needed to further confirm the relationship between RRM2 and survival prognosis of different tumors, which shows us the complexity of tumors and the diverse role of RRM2 in tumors.

For pancreatic adenocarcinoma (PAAD), we analyzed the dataset of the TCGA-PAAD (n = 183) projects and failed to observe a significant differential expression of RRM2 in PAAD compared with normal tissues. However, significant correlation was found between RRM2 expression and clinical prognosis of OS, PFS, RFS, and DSS in PAAD. A recent research by Anna et al. is line with the results of survival analysis(34). Perhaps larger sample sizes are warranted for confirming the role of SND1 expression in the PAAD.

In this study, we first described that RRM2 expression level in pan-cancer is correlated with the immune infiltration level of diverse tumor-associated cells (Fig. 5). In most cancers, the high expression of RRM2 will lead to a decrease in the level of immune infiltration, including ACC, KIRC, KIRP, LGG, LIHC, LUAD, PAAD, PRAD, SARC, TGCT, THCA, UCEC and UVM. Consistent with this, the high expression of RRM2 is associated with the poor prognosis of these tumors. These findings indicated that in at least 13 types of tumors, RRM2 can affect the prognosis of tumors by regulating the level of immune infiltration.

In the treatment of a series of cancer types, immune checkpoint blockade has shown remarkable advantages in numerous immunotherapy strategies(35). Immune checkpoint blockade increases antitumor immunity(35). Here, we showed the expression of RRM2 is closely related to immune checkpoints genes in various cancers (Fig. 6). In renal cell carcinoma, Xiong et al. found that RRM2 promotes sunitinib resistance and knockdown of RRM2 enhances the anti-tumor efficiency of PD-1 blockade in renal cancer(36). Our data analysis also revealed that the expression of RRM2 is related to 21, 38, and 30 checkpoint genes in KICH, KIRC and KIRP, respectively. Additionally, we also confirmed that the expression of RRM2 is positively associated with PDCD1 (PD-1) in KICH, KIRC and KIRP. These results suggest that RRM2 may be a promising therapeutic target for renal cell carcinoma. Nevertheless, the specific immunotherapy role of RRM2 in the immune checkpoint blockade still needs further study in pan-cancer.

We first integrated the data about RRM2-binding components and RRM2 expression- correlated targeting genes across all tumors (Fig. 7). The enrichment analyses of KEGG and KSEA indicated that “cell cycle”, “oocyte meiosis”, “p53 signaling pathway”, “cell senescence” and “DNA replication” etc. have potential roles in the effect of RRM2 on tumor pathogenesis.

In view of the importance of RRM2 in tumors, we further explored the drugs that may target RRM2. We identified drugs that target the cell cycle pathway, which is the most affected by RRM2 expression in pan-cancer. Nevertheless, whether these drugs have a role in the treatment of pan-cancer or certain tumors still needs to be further explored. It is worth noting that it was reported that RRM2 can convert the active metabolite of gemcitabine into an inactive form(37). And, alteration of RRM2 was related with resistance to gemcitabine-induced cell death in lung cancer and pancreatic ductal adenocarcinoma(37). So, we speculate that some of these predicted drugs can regulate efficacy of gemcitabine in some cancers. Of course, this needs more research to confirm.

In sum, our first pan-cancer analyses of RRM2 revealed statistical correlations of RRM2 expression with clinical prognosis, genetic alteration, tumor mutational burden or microsatellite instability, immune cell infiltration and immune checkpoint across multiple tumors. It helps understand the role of RRM2 in tumorigenesis from the perspective of clinical tumor samples.

Author Contributions

YW designed the experiments and carried out data analyses. JZ collated the data and performed the statistical analysis. RC produced the initial draft of the manuscript. YW contributed to drafting of the manuscript. PZ finalized the manuscript and had final responsibility for the decision to submit for publication. All authors contributed to manuscript revision, read, and approved the submitted version.

Data availability statement

The data used to support the findings of this study are available from the corresponding author upon request.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

LG M and TA C: Therapeutic targeting of tumor suppressor genes. Cancer 121: 1357–1368, 2015.
ZX L, ZQ Z, ZH W, et al.: Comprehensive characterization of the alternative splicing landscape in head and neck squamous cell carcinoma reveals novel events associated with tumorigenesis and the immune microenvironment. Theranostics 9: 7648–7665, 2019.
Weber J: Immune Checkpoint Proteins: A New Therapeutic Paradigm for Cancer—Preclinical Background: CTLA-4 and PD-1 Blockade. Semin Oncol 37: 430–439, 2010.
F P, J G, MC D-N, E T, C S-F and WH F: Immune infiltration in human tumors: a prognostic factor that should not be ignored. Oncogene 29: 1093–1102, 2010.
D K, AL A, J V, AK N, TA K and A C: Mechanisms of mutagenesis in vivo due to imbalanced dNTP pools. Nucleic Acids Res 39: 1360–1371, 2011.
Gandhi M, Groß M, Holler JM, et al.: The lncRNA lincNMR regulates nucleotide metabolism via a YBX1 - RRM2 axis in cancer. Nat Commun 11, 2020.
Mazzu YZ, Armenia J, Chakraborty G, et al.: A novel mechanism driving poor-prognosis prostate cancer: Overexpression of the DNA repair gene, ribonucleotide reductase small subunit M2 (RRM2). Clin Cancer Res 25: 4480–4492, 2019.
S Z, X W, YH W, et al.: siRNA Knockdown of RRM2 Effectively Suppressed Pancreatic Tumor Growth Alone or Synergistically with Doxorubicin. Mol Ther Nucleic Acids 12: 805–816, 2018.
S O, A M, MF O, et al.: Translational evidence for RRM2 as a prognostic biomarker and therapeutic target in Ewing sarcoma. Mol Cancer 20, 2021.
Yang Y, Lin J, Guo S, et al.: RRM2 protects against ferroptosis and is a tumor biomarker for liver cancer. Cancer Cell Int 20: 1–16, 2020.
H S, B Y, H Z, et al.: RRM2 is a potential prognostic biomarker with functional significance in glioma. Int J Biol Sci 15: 533–543, 2019.
CY J, L D, AH N, XL W, YW Y and R G: High expression of RRM2 as an independent predictive factor of poor prognosis in patients with lung adenocarcinoma. Aging (Albany NY) 13: 3518–3535, 2020.
N F, K S, C K, et al.: Targeting RRM2 and Mutant BRAF Is a Novel Combinatorial Strategy for Melanoma. Mol Cancer Res 14: 767–775, 2016.
N H, W G, K R, et al.: LncRNA AFAP1-AS1 Supresses miR-139-5p and Promotes Cell Proliferation and Chemotherapy Resistance of Non-small Cell Lung Cancer by Competitively Upregulating RRM2. Front Oncol 9, 2019.
Rasmussen RD, Gajjar MK, Tuckova L, et al.: BRCA1-regulated RRM2 expression protects glioblastoma cells from endogenous replication stress and promotes tumorigenicity. Nat Commun 7, 2016.
Goss KL, Koppenhafer SL, Waters T, et al.: The translational repressor 4E-BP1 regulates RRM2 levels and functions as a tumor suppressor in Ewing sarcoma tumors. Oncogene 40: 564–577, 2021.
CW C, Y L, S H, et al.: DHS (trans-4,4’-dihydroxystilbene) suppresses DNA replication and tumor growth by inhibiting RRM2 (ribonucleotide reductase regulatory subunit M2). Oncogene 38: 2364–2379, 2019.
Goldman MJ, Craft B, Hastie M, et al.: Visualizing and interpreting cancer genomics data via the Xena platform. Nat Biotechnol 2020 386 38: 675–678, 2020.
Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W and Smyth GK: Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43: e47, 2015.
Chalmers ZR, Connelly CF, Fabrizio D, et al.: Analysis of 100,000 human cancer genomes reveals the landscape of tumor mutational burden. Genome Med 9, 2017.
Hause RJ, Pritchard CC, Shendure J and Salipante SJ: Classification and characterization of microsatellite instability across 18 cancer types. Nat Med 2016 2211 22: 1342–1350, 2016.
T L, J F, Z Z, et al.: TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res 48: W509–W514, 2020.
Petitprez F, Meylan M, de Reyniès A, Sautès-Fridman C and Fridman WH: The Tumor Microenvironment in the Response to Immune Checkpoint Blockade Therapies. Front Immunol 11: 784, 2020.
Shannon P, Markiel A, Ozier O, et al.: Cytoscape: A software Environment for integrated models of biomolecular interaction networks. Genome Res 13: 2498–2504, 2003.
Bardou P, Mariette J, Escudié F, Djemiel C and Klopp C: jvenn: an interactive Venn diagram viewer. BMC Bioinforma 2014 151 15: 1–7, 2014.
Y W, S Z, F L, et al.: Therapeutic target database 2020: enriched resource for facilitating research and early development of targeted therapeutics. Nucleic Acids Res 48: D1031–D1041, 2020.
Osipov A, Saung MT, Zheng L and Murphy AG: Small molecule immunomodulation: The tumor microenvironment and overcoming immune escape. J Immunother Cancer 7: 224, 2019.
De Sanctis F, Ugel S, Facciponte J and Facciabene A: The dark side of tumor-associated endothelial cells., 2018.
Z S, Z L, H H, et al.: Cell-cycle-dependent phosphorylation of RRM1 ensures efficient DNA replication and regulates cancer vulnerability to ATR inhibition. Oncogene 39: 5721–5733, 2020.
Tabbal H, Septier A, Mathieu M, et al.: EZH2 cooperates with E2F1 to stimulate expression of genes involved in adrenocortical carcinoma aggressiveness. Br J Cancer 121: 384, 2019.
J A, H W, P C, et al.: RNA-dependent inhibition of ribonucleotide reductase is a major pathway for 5-azacytidine activity in acute myeloid leukemia. Blood 119: 5229–5238, 2012.
W M-M, S M-H, A A-S, et al.: RRM1 and ERCC1 as biomarkers in patients with locally advanced and metastatic malignant pleural mesothelioma treated with continuous infusion of low-dose gemcitabine plus cisplatin. BMC Cancer 21, 2021.
Tang Z, Kang B, Li C, Chen T and Zhang Z: GEPIA2: an enhanced web server for large-scale expression profiling and interactive analysis. Nucleic Acids Res 47: W556–W560, 2019.
A K-W, K B, I N-D, et al.: Expression of Genomic Instability-Related Molecules: Cyclin F, RRM2 and SPDL1 and Their Prognostic Significance in Pancreatic Adenocarcinoma. Cancers (Basel) 13: 1–19, 2021.
MA P, R S and MD H: Immune-Related Adverse Events Associated with Immune Checkpoint Blockade. N Engl J Med 378: 158–168, 2018.
W X, B Z, H Y, L Z, L Y and X J: RRM2 Regulates Sensitivity to Sunitinib and PD-1 Blockade in Renal Cancer by Stabilizing ANXA1 and Activating the AKT Pathway. Adv Sci (Weinheim, Baden-Wurttemberg, Ger, 2021.
R D, DJ M, S V and RJ B: Pegylated arginine deiminase synergistically increases the cytotoxicity of gemcitabine in human pancreatic cancer. J Exp Clin Cancer Res 33, 2014.

No competing interests reported.

SupplementaryFigure1.png
Figure S1. The expression level of RRM2 and its clinical correlation in pan-cancer. (A-L) The clinical correlation between RRM2 expression level and age in KIRP, LAML, LGG, LIHC, LUAD, LUSC, PAAD, PCPG, READ, STAD, TGCT and THYM, respectively. (M-O) The clinical correlation between RRM2 expression level and race in KIRP, LIHC and THYM, respectively. (P-X) The clinical correlation between RRM2 expression level and tumor stage of tumor patient in KIRC, KIRP, LIHC, LUAD, LUSC, PAAD, SKCM, TGCT, and YHCA, respectively. (Y-AG) The clinical correlation between RRM2 expression level and tumor status of tumor patient in KIRC, KIRP, LGG, LUAD, OV, PAAD, PCPG, PRAD and UVM, respectively. The number above the horizontal line represents the p-value between the two groups.
SupplementaryFigure2.png
Figure S2. Prognostic value of RRM2 in pan-cancer. (A-AB) Correlation between expression level of RRM2 and OS, DSS, DFI and PFI in pan-cancer, using the Kaplan-Meier method. OS, Overall survival (years); DSS, Disease-specific survival (years); DFI, Disease-free interval (years); PFI, Progression-free interval (years).
supplementaryFigure3.png
Figure S3. Correlation analysis between RRM2 expression and immune infiltration of cancer-associated cells. (A-CI) The correlation between RRM2 gene expression and the infiltration level of diverse immune cells in pan-cancer.
supplementaryFigure4.png
Figure S4. Gene set enrichment analysis (GSEA) of RRM2 in pan-cancer. (A-N) Gene set enrichment analysis (GSEA) of RRM2 in pan-cancer.
supplementarytable1.docx
Table S1. The KEGG pathway that RRM2 regulates in pan-cancer by using GSEA.
supplementarytable2.docx
Table S2. Potential drug prediction of RRM2 for pan-cancer.

Download PDF

Version 1

posted

You are reading this latest preprint version

A pan-cancer analysis of the oncogenic role of ribonucleotide reductase subunit M2 in human tumors

Status:

Version 1

Abstract

Figures

Introduction

Materials And Methods

1. The expression of RRM2 in pan-cancer

2. RRM2 expression and its clinical correlation in pan-cancer

3. Survival prognosis analysis Cox regression models

4. Genetic alteration analysis

5. Immune infiltration analysis

6. Immune checkpoint analysis

7. Enrichment analysis of SND1-related partners

8. Potential drug prediction for pan-cancer

Result

1. The expression of RRM2 in pan-cancer

2. RRM2 expression and clinical correlation in pan-cancer

3. Survival analysis

4. genetic alteration analysis data

5. immune infiltration

6. Immune checkpoint analysis

7. Enrichment analysis of SND1-related partners

8. Potential drug prediction for pan-cancer

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1