A Pan-Cancer Analysis of the Oncogenic Roles of RAD51 in Human Tumors

Lu, Han; Li, Zhenzhen; Liu, Liangeng; Tao, Yunjuan; Zhou, Yue; Mao, Xiaohong; Zhu, Aoxun; Wu, Honglin; Zheng, Xingzhong

doi:https://doi.org/10.1155/2022/1591377

Advanced Gut & Microbiome Research

On this page

Abstract Introduction Materials and Methods Results Discussion Conclusions Data Availability Conflicts of Interest Authors’ Contributions Acknowledgments Supplementary Materials References Copyright Related Articles

Research Article | Open Access

Volume 2022 | Article ID 1591377 | https://doi.org/10.1155/2022/1591377

A Pan-Cancer Analysis of the Oncogenic Roles of RAD51 in Human Tumors

Han Lu,¹Zhenzhen Li,²Liangeng Liu,¹Yunjuan Tao,¹Yue Zhou,¹Xiaohong Mao,¹Aoxun Zhu,¹Honglin Wu,³and Xingzhong Zheng¹

Academic Editor: Jiong Yu

Received27 May 2022

Revised01 Aug 2022

Accepted18 Aug 2022

Published17 Sept 2022

Abstract

Objective. RAD51 homolog 1 (RAD51) plays significant roles in DNA damage repair and apoptosis. These pathways are tightly associated with tumor initiation and progression. To unravel the roles of RAD51 in oncogenesis and progression of different cancers, herein, a comprehensive analysis of the RAD51 was carried out using multiomics datasets of 33 cancers. Methods. Raw data were obtained from The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) databases. We analyzed the correlation between RAD51 expression and drug response using datasets from CellMiner. Next, clinical characteristics and prognostic values of RAD51 were conducted based on TCGA data. The correlation between RAD51 expression and tumor immune infiltration was explored. This was followed by gene set enrichment analysis by Rsoftware. In addition, pan-cancer analysis was conducted to investigate genetic and epigenetic alterations, respectively. Results. RAD51 was upregulated in most tumors, and this was associated with poor overall survival (OS), progression-free survival (PFS), and disease-specific survival (DSS). The expression level of RAD51 is significantly associated with the IC50 of multiple antitumor drugs and the proportion of stromal and immune components in tumor microenvironment (TME). Moreover, RAD51 expression showed a positive relationship with multiple key immune checkpoint and immunosuppressive genes, including death-ligand 1 (PD-L1), cytotoxic T-lymphocyte associated protein 4 (CTLA-4), CD28, and several TNF-related immune genes. Gene set enrichment analysis uncovered that RAD51 correlated with cell cycle, cell division, and immune system pathways in diverse cancers. Our results revealed a strong link between RAD51 expression and microsatellite instability (MSI) or tumor mutation burden (TMB). Conclusions. Our pan-cancer analysis provides a comprehensive overview of the roles of RAD51 in multiple human cancers and infers that RAD51 has the potential as a biomarker for progression and immune infiltration of different tumor types.

1. Introduction

Genomic instability is a critical feature of tumorigenesis. The absence of DNA damage repair (DDR) pathway may lead to many catastrophic consequences, such as mutations or chromosomal rearrangements, promoting genomic instability and tumor progression [1]. RAD51, a DNA repair protein, functions in mediating the pairing of homologous DNA sequences and strand invasion [2]. RAD51 is central to homologous recombination repair (HRR), which is one of the major pathways for DNA double-strand break repair, directly taking charge of genomic stability [3]. There was a host of animal and cell-based data connecting RAD51 to several types of malignancies. RAD51 was shown to be elevated in brain cancer [4], head and neck cancer [5], thyroid cancer [6], nonsmall-cell lung cancer [7], and breast cancer [8]. Therefore, RAD51 has the potential to be a biomarker and a key mediator in a variety of human malignancies.

DNA damage response of tumor cells has a profound impact on the organization of tumor microenvironment (TME) that plays a critical role in the beginning and progression of human malignancies [9, 10]. TME can function in malignant transformation through inhibiting DDR pathways [11]. Hypoxia, one of the TME characteristics, was demonstrated to downregulate c-MYC, a proliferation promoting transcription factor that combines with promoters of Mlh1 and Mlh2, which are known to be involved in the mismatch repair (MMR) pathway [12]. And, it is worth noting that RAD51 and BRCA1 in the HRR pathway are also associated with hypoxia, so the TME can regulate genomic instability via key repair factors [13, 14]. And, some of the most common single nucleotide polymorphisms (SNPs) found in RAD51 may increase drug resistance of cancers [15]. SNPs in RAD51 were also manifested to increase the risk of glioblastoma and nonsmall cell lung cancer and influence overall survival (OS) [16, 17]. Despite these findings imply that TME and mutation are closely associated with tumorigenesis, additional research is required. As a result, it is paramount to elucidate the interactions between the biomarker gene and TME, mutation, or epigenetic modification.

Herein, a comprehensive analysis of the RAD51 was carried out using multiomics datasets of 33 cancers, including the response of antitumor drugs, expression profile comparison, survival status, immune cell infiltration, TMB, and microsatellite instability (MSI). In addition, genetic alteration, methylation, and relevant cellular pathways were investigated. In conclusion, the obtained results indicate the potential molecular mechanism of RAD51 in the clinical prognosis of multiple human cancers.

2. Materials and Methods

2.1. Data Collection

RNA sequence and related clinical data (10201 samples for 33 types of cancer) were retrieved from The Cancer Genome Atlas (TCGA) database [18]. Gene profile data from normal human tissues were retrieved from Genotype-Tissue Expression (GTEx) [19]. The R package “metafor” was used for data processing (filtering, deleting missing, and duplicated results) and transformed by log2 (R version: 4.1.2).

2.2. Gene Expression Analysis and Relationship with Pathological Stages

We used Tumor Immune Estimation Resource (TIMER) to investigate the expression profiling of RAD51 between tumors and adjacent normal tissues [20]. However, some samples lacked normal tissue. We combined the data from TCGA and GTEx to perform further analysis of the expression of RAD51 in tumors and adjacent normal tissues. The transformed expression data were applied for the box and radar plots. Gene Expression Profiling Interactive Analysis (GEPIA2) [21] is a web-based platform, and it was used for expression profiling based on pathological stages.

2.3. Survival Analysis and Cox Regression Analysis

We used UCSCXenaShiny [22] to perform survival analysis of RAD51 for pan-cancer based on TCGA. Cutoff-low (50%) and cutoff-high (50%) values were used as the thresholds of RAD51 expression to split the low-expression and high-expression cohorts. The log-rank test was utilized for differential analysis ( value < 0.05 as significant). With the aid of R package “survival,” univariate Cox regression analysis was used to explore the correlation of RAD51 expression with patients’ OS, progression-free survival (PFS), disease-specific survival (DSS), and disease-free survival (DFS) in each cancer type based on TCGA databases [23].

2.4. RAD51 Expression and Drug Response

CellMiner (https://discover.nci.nih.gov/cellminer/home.do) was utilized to explore the correlation between RAD51 expression and drug response. CellMiner database was designed to investigate transcript and drug patterns data in the NCI-60 cancerous cell line set, which is a dataset of 60 diverse human cancer cell lines used by the Developmental Therapeutics Program of the U.S. National Cancer Institute to screen over 100,000 chemical compounds [24]. RNA expression data (RNA: RNA-SEQ) and drug data (compound activity: DTPNCI-60) were downloaded. To ensure the reliability of the analysis results, clinical trials and FDA-approved drug results were selected for the analysis of drug sensitivity. The half-maximal inhibitory concentration (IC50) that represented drug response was estimated employing the pRRophetic package [25].

2.5. Immune Infiltration Analysis

TIMER was used to calculate immune cell infiltration scores for six main immune cell types (CD4⁺ T cells, CD 8⁺ T cells, B cells, dendritic cells, neutrophils, and macrophages) [26]. To increase the accuracy of the results, we also used xCell to calculate the correlation of RAD51 with various immune cells. ESTIMATE [27] (Estimation of STromal and Immune cells in MAlignant Tumor tissues using Expression data) is a tool for predicting tumor purity and the presence of infiltrating stromal/immune cells in tumor tissues using gene expression data [27]. We downloaded disease-centric results containing stromal, immune, and ESTIMATE scores of pan-cancer from the ESTIMATE database and combined them with the results of RAD51 expression in diverse tumors in TCGA, and the correlation of RAD51 expression with immune infiltration scores was performed ( value < 0.05 as significant).

The correlation of RAD51 expression with immune-related genes, including immunosuppressive genes, histocompatibility complex (MHC) genes, chemokine, and chemokine receptors downloaded from TISIDB [28], was calculated by Spearman correlation coefficients.

2.6. Gene Set Enrichment Analysis

Gene Set Enrichment Analysis (GSEA) was employed to explore the potential biological process of RAD51 in diverse tumors using clusterProfiler package [29]. Adjusted value < 0.05 was considered statistically significant. RAD51-related genes were selected by calculating the Spearman correlation coefficient between RAD51 and other genes in all samples of 33 tumors downloaded from TCGA ( value < 0.05 as significant).

2.7. Analysis of RAD51 Expression with TMB and MSI

We obtained the scores of tumor mutation burden (TMB) and microsatellite instability (MSI) from TCGA. Correlation analysis of the RAD51 expression with TMB and MSI was determined employing the Spearman correlation coefficient ( value < 0.05 as significant).

2.8. CNV and Methylation Profile of RAD51

We investigated the copy number variation (CNV) and methylation profile of RAD51 by using Gene Set Cancer Analysis (GSCA) [30], which is a web platform providing methylation, immune infiltration, and CNV based on the TCGA database. Also, the correlation of RAD51 expression with the level of CNV and methylation in pan-cancer was performed by the GSCA website ( value < 0.05 as significant).

2.9. Statistical Analysis

Correlation analysis between RAD51 expression and targets of research, such as drug sensitivity, immune cell infiltration scores, immune checkpoint, CNV, methylation, TMB, or MSI, was determined by using the Spearman Correlation coefficient. Paired -test or -test was performed for comparison of RAD51 expression levels in cancer tissues and normal tissues. For survival analysis, the hazard ratio (HR) and value were calculated using univariate Cox regression analysis. Kaplan-Meier analysis was utilized to examine the survival time of patients stratified according to the degree of RAD51 expression. value < 0.05 was considered the significant threshold for all statistical analyses. The R tools ggplot2, ggunchained, ggpubr, and forestplot were used for visualization.

3. Results

3.1. Pan-Cancer Expression Landscape of RAD51

TIMER database was utilized to study the differential expression of RAD51 between tumors and adjacent normal tissues presented in TCGA. As shown in Figure 1(a), the RAD51 expression was significantly higher in cancers versus adjacent normal tissues in bladder urothelial carcinoma (BLCA), breast invasive carcinoma (BRCA), cervical squamous cell carcinoma, endocervical adenocarcinoma (CESC), cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), esophageal carcinoma (ESCA), glioblastoma multiforme (GBM), head and neck squamous cell carcinoma (HNSC), kidney renal papillary cell carcinoma (KIRP), kidney renal clear cell carcinoma (KIRC), liver hepatocellular carcinoma (LIHC), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), prostate adenocarcinoma (PRAD), stomach adenocarcinoma (STAD), thyroid carcinoma (THCA), and uterine carcinosarcoma (UCEC). Some normal tissue data were not available in TCGA, and we further evaluated the expression differences of RAD51 between the tumor and normal tissues employing the GTEx dataset. According to the results from the GTEx database, the differences were highly significant for 27 of the 33 cancers (Figure 1(b)). Profiles of RAD51 expression across various types of cancers and corresponding tissues were presented and ranked from low to high (Figures 1(c) and 1(d)).

(a)

(b)

(c)

(d)

3.2. Correlation of RAD51 Expression with Pathological Stages of Cancer

GEPIA2 was utilized to examine the relationship between RAD51 expression and the pathological stages of cancers. We found that the expression profiles of RAD51 in kidney chromophobe (KICH), KIRP, LIHC, ACC, KIRC, LUAD, BRCA, COAD, THCA, testicular germ cell tumor (TGCT), and uterine carcinosarcoma (UCS) were all significantly associated with their pathological stages (Figure 2).

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

(k)

3.3. RAD51 and Drug Sensitivity

Targeted therapies have an important role in cancer management. Genetic alterations that contribute to cancer heterogeneity are associated with the action of targeted therapies. Reliable prediction of RAD51 for targeted therapies remains lacking in pan-cancer. Thus, we performed the correlation analysis of RAD51 expression and drug sensitivity in the NCI-60. RAD51 expression was positively connected with drug sensitivity in NCI-60 treated with methylprednisolone, nelarabine, vorinostat, ribavirin, 6-thioguanine, chelerythrine, cyclophosphamide, parthenolide, ZM-336372, 8-chloro-adenosine, and 6-thioguanine. Differently, there was a negative correlation between RAD51 expression and the anticancer drug BPTES, INK-128, LY-3023414, mithramycin, and GDC-0349 (Figure 3(a)). Further analysis was conducted to uncover whether RAD51 expression was predictive for drug responses in cancer patients. We noted that the high-expression group of RAD51 had significantly lower estimated IC50 of INK-128, LY-3023414, and GDC−0349, indicating that the subpopulations with high-expression RAD51 presented higher sensitivity to INK-128, LY-3023414, and GDC−0349. In the meantime, the low-expression group showed significantly lower estimated IC50 of vorinostat, indicating that low-expression subpopulations preferred to respond to vorinostat (Figure 3(b)). Therefore, RAD51 might be a favorable predictive biomarker for targeted therapies with the aforementioned drugs.

(a)

(b)

3.4. Effects of RAD51 on Pan-Cancer Prognosis

Cox proportional hazard model was utilized to examine the relationship between RAD51 expression and OS in various cancer types through TCGA. As shown in Figure 4(a), high RAD51 level was significantly linked to worse OS in patients with adrenocortical carcinoma (ACC), BRCA, KICH, KIRP, brain lower grade glioma (LGG), LIHC, LUAD, mesothelioma (MESO), pancreatic adenocarcinoma (PAAD), pheochromocytoma and paraganglioma (PCPG), PRAD, rectum adenocarcinoma (READ), and THYM. With the aid of survival analyses, Kaplan–Meier curves revealed that high expression of RAD51 was associated with shorter survival times in patients with ACC, BRCA, KICH, KIRP, LGG, LIHC, LUAD, MESO, PAAD, PCPG, and PRAD (Figures 4(b)–4(l)). Conversely, the high expressions of RAD51 in READ and THYM were associated with longer OS times (Figures 4(m) and 4(n)), and this was consistent with the result through Cox proportional hazards model.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

(k)

(l)

(m)

(n)

Next, we performed Cox regression analysis for PFS, DSS, and DFS to elucidate the relationship between RAD51 expression and various survival indicators. For PFS, the high expression of RAD51 was significantly associated with poor prognosis in cancer patients with ACC, BLCA, KIRC, KIRP, LGG, LIHC, LUAD, MESO, PAAD, PCPG, PRAD, SKCM, THCA, and UVM (Figure 5(a)). On the other side, it was demonstrated that high RAD51 level was correlated with significantly worse DSS in patients with ACC, BLCA, BRCA, KIRC, KIRP, LGG, LIHC, LUAD, MESO, SKCM, and STAD (Figure 5(b)). Similarly, a high level of RAD51 expression was also concerned to have a correlation with an unfavorable DFS in the TCGA cancer types, which are exhibited in Figure 5(c).

(a)

(b)

(c)

3.5. Correlation between RAD51 Expression and Immune Infiltrating Level in Cancers

As the central of HRR, RAD51 participates in the regulation of DDR and plays a key role in genomic stability. It is known that TME regulates genomic stability via DDR pathways [1], suggesting the close connection between RAD51 and TME in tumorigenesis. To determine whether RAD51 affects TME, we used TIMER to examine the correlation between RAD51 expression and the infiltrating degree of six immune cell types (CD8⁺ T cell, CD4⁺ T cell, neutrophil, dendritic cell, macrophage, and B cell). A clustering heat map revealed a significant correlation between RAD51 expression and six major immune cells: CD8⁺ T cells in 13 types of cancer, CD4⁺ T cells in 18 types of cancer, neutrophils in 13 types of cancer, dendritic cells in 18 types of cancer, macrophages in 19 types of cancer, and B cells in 10 types of cancer (Figure 6(a)). KIRC, LIHC, THCA, and THYM were the top four tumor groups that had a strong association with these six immune cells accessible in TIMER (Figure 6(a) and Figure S1). We further performed a pan-cancer analysis of the relationship between the RAD51 expression and the immune infiltration level based on xCell. As shown in Figure 6(b), RAD51 expression was significantly negatively associated with 38 subtypes of immune cells in most tumor types. Remarkably, Th2 CD4⁺ T cells exhibited a highly significant positive correlation with RAD51 expression in these cancers (Figure 6(b)).

(a)

(b)

We used the ESTIMATE algorithm to estimate the relationship between the level of RAD51 expression and the proportion of immune matrix components (immune cells, stromal cells, and the sum of the previous two) in the TME presented in the form of ImmuneScore, StromalScore, and ESTIMATEScore in cancers. The degree of RAD51 expression significantly correlated with the proportion of stromal and immune components of TME in UCEC, LUSC, THCA, LUAD, SKCM, CESC, GBM, STAD, KIRC, HNSC, ESCA, and PAAD (Figure 7).

3.6. Correlation of RAD51 Expression with Immune-Related Genes

Immunosurveillance has crucial influences on the therapy and prognosis of cancer patients. Immune checkpoints provide opportunities for tumors to evade immune responses. We determined gene coexpression analyses to illustrate the relationship between RAD51 expression and checkpoint-related genes that are tightly linked to and recognized as immune checkpoint components. Our results revealed that RAD51 expression was positively correlated with multiple checkpoint genes (Figure 8). The coexpression analysis of RAD51 and immunostimulatory gene indicated a high correlation with TNF-related immune genes, including TNFSF9, TNFSF4, TNFRSF18, TNFRSF9, TNFRSF8, TNFRSF4, TNFRSF25, TNFRSF18, TNFRSF14, and CD276 (Figure 8). PDCD1 (PD-1)/CD274 (PD-L1) and CD28/CTLA-4 axis are two fundamental pathways of immune checkpoint, which makes a huge effect on cancer therapy [31, 32]. Our results manifested that PDCD1 (PD-1), CD274 (PD-L1), CD28, and CTLA-4 significantly correlated with RAD51 expression in multiple tumor types (Figure 8). We also found a strong connection between RAD51 expression and immune-related genes, including immunosuppressive genes, MHC genes, and chemokine (Figure S2B-S2D). These results may indicate the potential mechanism of RAD51 influencing infiltration of TAMs.

3.7. Correlation between RAD51 Expression and TMB or MSI in Cancers

TMB and MSI are two emerging biomarkers associated with the immunotherapy response. We determined a pan-cancer analysis of the correlation of the RAD51 expression with TMB and MSI. In STAD, ACC, LUAD, LGG, PAAD, BRCA, and PRAD, RAD51 expression positively correlated with TMB (Figures 9(a) and 9(c)–9(h)). It is worth noting that THYM exhibited a remarkably negative correlation with TMB (Figure 9(a)). In STAD, CHOL, UCEC, MESO, COAD, ACC, UVM, LIHC, BLCA, and UCS, RAD51 expression positively correlated with MSI, whereas in TGCT and CESC, it was adversely connected with MSI (Figures 9(b) and 9(i)–9(n)).

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

(k)

(l)

(m)

(n)

3.8. GSEA Analysis of RAD51 in Pan-Cancer

To deeply examine potential biological pathways of RAD51 regulation in diverse tumors, we performed GSEA analysis using “clusterprofiler” in pan-cancer subjects. Then, we selected the 12 types of tumor with similar results. These results indicated that these pathways enriched by GESA focused on cell cycle, cell division, immune system, cell communication, DNA repair, DNA metabolic process, and DNA replication (Figure 10 and Figure S3). The etiologic fundamental of cancer is deregulation of cell cycle arrest, activation of genes, and gene products involved in DNA repair when DNA damage occurs [33]. Therefore, it can lead to the loss of heterozygosity and genetic instability, which is the kernel of carcinogenesis, if not regulated properly [33]. Besides, GSEA results also indicated that RAD51-related genes were enriched in immune-related pathways, which implied the correlation between RAD51 and tumor immunity.

3.9. CNV and DNA Methylation of RAD51 in Pan-Cancer

We investigated the genetic alterations of RAD51 using the cBioPortal. The results demonstrated that the highest frequency of alteration in RAD51 (>5%) occurred in patients with lung cancer, where “deep deletion” and “amplification” were the primary types. The “amplification” type of CNV was also the primary type in the soft tissue sarcoma and pancreatic cancer and occupied an important part in breast cancer (Figure 11(a)). Thereafter, we explored the relationship between CNV and mRNA expression. The results revealed a significantly positive correlation of RAD51 expression with CNV in LUSC, COAD, CESC, HNSC, TGCT, ESCA, SKCM, OV, STAD, READ, UCEC, UCS, SARC, THYM, and BLCA (Figure 11(b)).

(a)

(b)

(c)

DNA methylation directly affects tumor occurrence and progression [34]. We utilized Spearman’s correlation to explore the relationship between DNA methylation and the expression of RAD51. A significantly negative correlation of DNA methylation with RAD51 expression was observed in LIHC, OV, LGG, PRAD, SKCM, ACC, THYM, HNSC, SARC, and BRCA (Figure 11(c)).

4. Discussion

The maintenance of genome integrity is indispensable for any organism survival and the inheritance of traits. Deficiency in DDR results in prolonged DNA damage and can give rise to gene mutations, chromosome rearrangements, carcinogenesis, and cancer progression. These defects in DDR may have the potential to provide therapeutic implications for the clinical practice of malignancies. Previous studies have shown that the impact of a range of defects in the HRR is relevant to cancer inception, progression, and therapeutics [15]. RAD51, the vital molecular actor of HRR, directly takes charge of genomic stability. With the help of accessory proteins, RAD51 performs the tricky task of revolving around a single broken strand of DNA, then captures the backup DNA copy, matching the sequence of the broken stand to the homologous sequence. In the complex of RAD51, the single strand is swapped for another strand in the double-stranded DNA. Finally, a plethora of other proteins fill in the missing sections of DNA and restore two matching copies [3]. RAD51 was expressed at high levels in multiple types of cancers and shown to be elevated in several types of malignancies compared with corresponding normal tissues [35]. The reports abovementioned provide the evidence that RAD51 has the potential to be a biomarker in a variety of human malignancies. Thus, here, we set out to perform a pan-cancer analysis of RAD51.

In our result, the analyses of 33 cancer datasets from the TCGA were consistent with previous studies and indicated that the expression of RAD51 was significantly higher compared to adjacent normal tissues in BLCA, BRCA, CESC, CHOL, COAD, ESCA, GBM, HNSC, KIRP, KIRC, LIHC, LUAD, LUSC, PRAD, STAD, THCA, and UCEC (Figures 1(a) and 1(b)). Cox proportional hazard model analysis and Kaplan-Meier survival curves in various cancer types revealed that the high expression of RAD51 was associated with worse OS in patients with ACC, BRCA, KICH, KIRP, LGG, LIHC, LUAD, MESO, PAAD, PCPG, and PRAD (Figures 4(b)–4(l)). In contrast, the high expressions of RAD51 in READ and THYM were associated with longer OS times (Figures 4(m) and 4(n)). The upregulated expression of RAD51 correlated with worse DSS, PFS, and DFS in several cancers. It is worth noting that KIRP LIHC and LUAD simultaneously exhibited worse DSS, PFS, and DFS (Figures 5(a)–5(c)). Next, we will emphatically discuss the situation of RAD51 in LIHC and KIRC, which exhibited worse DSS and DFS. RAD51 expression closely correlated with its CNV and DNA methylation in diverse tumors, and a high RAD51 methylation level or CNV could serve as a biomarker of prognosis in patients with cancer. In particular, LIHC displayed the highest correlation with CNV. These results implied that RAD51 can be used as a significant prognostic marker in indicated tumor types.

It was reported that elevated levels of RAD51 contributed to drug resistance in BCR-ABL transformed cells, which is the etiological cause of chronic myeloid leukemia (CML) [36]. Thus, RAD51 could be assigned as a potential alternative approach for overcoming drug resistance in cancer therapeutics [37]. We performed the correlation analysis of RAD51 expression and drug sensitivity in the NCI-60. In our results, the drug sensitivity of methylprednisolone, nelarabine, vorinostat, ribavirin, 6-thioguanine, chelerythrine, cyclophosphamide, parthenolide, ZM-336372, 8-chloro-adenosine, and 6-thioguanin were positively associated with RAD51 expression (Figure 3(a)). Inversely, a negative correlation was exhibited between RAD51 expression and BPTES, INK-128, LY-3023414, mithramycin, and GDC-0349 (Figure 3(a)). Further, high-expression RAD51 presented higher sensitivity to INK-128, LY-3023414 and GDC−0349, and low-expression subpopulations preferred to respond to vorinostat (Figure 3(b)). These results not only implied the potential role of RAD51 in the mechanisms of tumor drug resistance but also provided new therapeutic regimens when RAD51 might be used for targeted therapies with the aforementioned drugs.

TME plays a critical role in tumorigenesis [38]. Alterations driven by oncogene may influence TME and inhibit immune responses. Tumor cells might escape immune surveillance by generating TAMs to restrict the cytotoxicity from immune cells [39]. In our results, TAM infiltration significantly correlated with RAD51 expression in multiple tumor types, especially M2-like macrophages, indicating that RAD51 may affect macrophage polarization (Figure S2A). CD8⁺ cytotoxic T lymphocytes (CTLs) are preferred immune cells for tumor immunotherapy [40]. We indeed found a significant association between CTLs and RAD51 expression in multiple malignancies (Figure 6). CD4⁺ and CD8⁺ T cell responses are part of the cancer immune cycle, and both populations significantly influence the clinical outcomes [41]. The balance between CD4⁺ and CD8⁺ T cells play important roles in tumor-suppressive mechanisms. In our results, Th2 and Th1 cells exhibited a highly significant positive correlation with RAD51 expression in these cancers (Figure 6). KIRC, LIHC, THCA, and THYM were the top four tumor groups that had a strong association with these six immune cells. (Figure 6(a)). The function of RAD51 in tumor-suppression may mainly depend on CD4⁺ T cells, which can regulate and cooperate with CTLs. Immune effector cells in the TME can suppress tumor growth [42].

Our results further demonstrated a strong link between RAD51 level and immune checkpoint genes, such as CD274 (PD-L1), CTLA-4, CD28, and TNF-related immune genes. PDCD1 (PD-1)/CD274 (PD-L1) axis is responsible for cancer immune escape and plays a crucial part in cancer therapy [31]. CTLA-4 and CD28 are members of immunoglobulin-related receptors that are responsible for T-cell immune regulation [32]. The correlation between RAD51 expression and immune checkpoint genes is very high in LIHC and KIRC, especially KIRC. These results suggested that RAD51 may influence the regulation of the tumor immune response via immune checkpoint activities. And, we estimated the proportion of immune matrix components in the TME for pan-cancer, suggesting that the degree of RAD51 expression significantly correlated with the proportion of stromal and immune components of TME in UCEC, LUSC, THCA, LUAD, SKCM, CESC, GBM, STAD, KIRC, HNSC, ESCA, and PAAD (Figure 8). Moreover, GSEA results indicated that RAD51-related genes were involved in the immune system pathway, which implies the relationship between RAD51 and the immune process (Figure 10). RAD51 is an essential protein related to HRR and directly functions in DNA damage repair. The above results prove that RAD51 also participates in immune infiltrating and has high correlation with immune checkpoint, implying RAD51 has crucial roles in DNA damage repair and immune infiltrating. These results provide evidence for RAD51 as a potential drug target for tumor immunotherapy. TMB and MSI are two emerging biomarkers associated with the immunotherapy response [43, 44]. Our results also revealed a strong link between RAD51 expression and MSI or TMB (Figure 9).

However, the limitations of our research should be pointed out. Our study was mainly based on bioinformatics. The patient data obtained from open databases have not been verified by experiments. Therefore, more studies to validate the expression and function of RAD51 in vivo and in vitro are needed in the future.

5. Conclusions

Overall, our study determined an integrative analysis to uncover the statistical association between the level of RAD51 expression and drug response, survival status, immune cell infiltration, tumor mutation burden, microsatellite instability, and genetic alteration in diverse tumors, contributing to elucidate the roles of RAD51 in tumorigenesis from multiple perspectives. Particularly, in LIHC and KIRC, RAD51 exhibited high correlation with prognosis, immune cell infiltration, and immune checkpoint.

Data Availability

The data that support the findings of this study are available from the corresponding authors upon reasonable request.

Conflicts of Interest

The authors declare no competing interests.

Authors’ Contributions

H.L., L.L., H.W., and X.Z. initiated the study. H.L. and Z.L. performed the analysis. Z.L., Y.T., A.Z., Y.Z., and X.M. collected the literature. H.L., Z.L., and H.W. wrote the manuscript. L.L., H.W., and X.Z. supervised sample collection and data processing.

Acknowledgments

We acknowledge the contributions from TCGA, GTEx, xCell, TIMER, ESTIMATE, CellMiner, and cBioPortal databases.

Supplementary Materials

Figure S1: relationship between RAD51 expression and the infiltration scores of six immune cell types (B cell, CD4⁺ T cell, CD8⁺ T cell, neutrophil, macrophage, and dendritic cell) in several tumors (KIRC, LGG, LIHC, THCA, and THYM) accessible in TIMER. Figure S2: correlation of RAD51 expression with (a) macrophages, (b) chemokine, (c) immunosuppressive genes, and (d) MHC genes. Red represents positive correlation, blue represents negative correlation, and the darker the color, the stronger the correlation. , , , and . Figure S3: GSEA of RAD51 in TCGA. The top 20 significant pathways of GSEA results across the indicated tumor types. Adjusted value < 0.05 was considered statistically significant. (Supplementary Materials)

References

M. Jiang, K. Jia, L. Wang et al., “Alterations of DNA damage response pathway: biomarker and therapeutic strategy for cancer immunotherapy,” Acta Pharmaceutica Sinica B, vol. 11, no. 10, pp. 2983–2994, 2021.
View at: Publisher Site | Google Scholar
B. Bonilla, S. R. Hengel, M. K. Grundy, and K. A. Bernstein, “RAD51Gene family structure and function,” Annual Review of Genetics, vol. 54, no. 1, pp. 25–46, 2020.
View at: Publisher Site | Google Scholar
M. A.-O. Gachechiladze, J. Škarda, A. Soltermann, and M. Joerger, “RAD51 as a potential surrogate marker for DNA repair capacity in solid malignancies,” International Journal of Cancer, vol. 141, no. 7, pp. 1286–1294, 2017.
View at: Publisher Site | Google Scholar
J. W. Welsh, R. K. Ellsworth, R. Kumar et al., “Rad51 protein expression and survival in patients with glioblastoma multiforme,” International Journal of Radiation Oncology • Biology • Physics, vol. 74, 2009.
View at: Google Scholar
P. P. Connell, K. Jayathilaka, D. Haraf, R. Weichselbaum, E. Vokes, and M. Lingen, “Pilot study examining tumor expression of RAD51 and clinical outcomes in human head cancers,” International Journal of Oncology, vol. 28, no. 5, pp. 1113–1119, 2006.
View at: Publisher Site | Google Scholar
D. Grimm, “Current knowledge in thyroid cancer-from bench to bedside,” International Journal of Molecular Sciences, vol. 18, no. 7, p. 1529, 2017.
View at: Publisher Site | Google Scholar
A. A.-O. Pataer, R. Shao, A. M. Correa et al., “Major pathologic response and RAD51 predict survival in lung cancer patients receiving neoadjuvant chemotherapy,” Cancer Medicine, vol. 7, no. 6, pp. 2405–2414, 2018.
View at: Publisher Site | Google Scholar
J. Hu, N. Wang, and Y. J. Wang, “XRCC3 and RAD51 expression are associated with clinical factors in breast cancer,” PLoS One, vol. 8, no. 8, 2013.
View at: Publisher Site | Google Scholar
F. R. Balkwill, M. Capasso, and T. Hagemann, “The tumor microenvironment at a glance,” Journal of Cell Science, vol. 125, no. 23, pp. 5591–5596, 2012.
View at: Publisher Site | Google Scholar
F. Klemm and J. A. Joyce, “Microenvironmental regulation of therapeutic response in cancer,” Trends in Cell Biology, vol. 25, no. 4, pp. 198–213, 2015.
View at: Publisher Site | Google Scholar
J. Yuan, L. Narayanan, S. Rockwell, and P. M. Glazer, “Diminished DNA repair and elevated mutagenesis in mammalian cells exposed to hypoxia and low pH,” Cancer Research, vol. 60, no. 16, pp. 4372–4376, 2000.
View at: Google Scholar
V. T. Mihaylova, R. S. Bindra, J. Yuan et al., “Decreased expression of the DNA mismatch repair gene Mlh1 under hypoxic stress in mammalian cells,” Molecular and Cellular Biology, vol. 23, no. 9, pp. 3265–3273, 2003.
View at: Publisher Site | Google Scholar
R. S. Bindra, S. L. Gibson, A. Meng et al., “Hypoxia-induced down-regulation of BRCA1 expression by E2Fs,” Cancer Research, vol. 65, no. 24, pp. 11597–11604, 2005.
View at: Publisher Site | Google Scholar
R. S. Bindra and P. M. Glazer, “Repression of _RAD51_ gene expression by E2F4/p130 complexes in hypoxia,” Oncogene, vol. 26, no. 14, pp. 2048–2057, 2007.
View at: Publisher Site | Google Scholar
E. Laurini, D. Marson, A. Fermeglia, S. Aulic, M. Fermeglia, and S. Pricl, “Role of Rad51 and DNA repair in cancer: a molecular perspective,” Pharmacology & Therapeutics, vol. 208, p. 107492, 2020.
View at: Publisher Site | Google Scholar
S. Franceschi, S. Tomei, C. M. Mazzanti et al., “Association between RAD 51 rs1801320 and susceptibility to glioblastoma,” Journal of Neuro-Oncology, vol. 126, no. 2, pp. 265–270, 2016.
View at: Publisher Site | Google Scholar
X. Yin, M. Liu, Y. Tian, J. Wang, and Y. Xu, “Cryo-EM structure of human DNA-PK holoenzyme,” Cell Research, vol. 27, no. 11, pp. 1341–1350, 2017.
View at: Publisher Site | Google Scholar
K. Tomczak, P. Czerwińska, and M. Wiznerowicz, “Review The Cancer Genome Atlas (TCGA): an immeasurable source of knowledge,” Contemporary Oncology, vol. 1A, pp. 68–77, 2015.
View at: Publisher Site | Google Scholar
GTEx Consortium, “The Genotype-Tissue Expression (GTEx) project,” Nature Genetics, vol. 45, no. 6, pp. 580–585, 2013.
View at: Publisher Site | Google Scholar
T. Li, J. Fan, B. Wang et al., “TIMER: a web server for comprehensive analysis of tumor-infiltrating immune cells,” Cancer Research, vol. 77, no. 21, pp. e108–e110, 2017.
View at: Publisher Site | Google Scholar
Z. Tang, B. Kang, C. Li, T. Chen, and Z. Zhang, “GEPIA2: an enhanced web server for large-scale expression profiling and interactive analysis,” Nucleic Acids Research, vol. 47, no. W1, pp. W556–W560, 2019.
View at: Publisher Site | Google Scholar
S. Wang, Y. Xiong, L. Zhao et al., “UCSCXenaShiny: an R/CRAN package for interactive analysis of UCSC Xena data,” Bioinformatics, vol. 38, no. 2, pp. 527–529, 2022.
View at: Publisher Site | Google Scholar
P. J. Heagerty, T. Lumley, and M. S. Pepe, “Time-dependent ROC curves for censored survival data and a diagnostic marker,” Biometrics, vol. 56, no. 2, pp. 337–344, 2000.
View at: Publisher Site | Google Scholar
W. C. Reinhold, M. Sunshine, H. Liu et al., “CellMiner: a web-based suite of genomic and pharmacologic tools to explore transcript and drug patterns in the NCI-60 cell line set,” Cancer Research, vol. 72, no. 14, pp. 3499–3511, 2012.
View at: Publisher Site | Google Scholar
P. Geeleher, N. Cox, and R. S. Huang, “pRRophetic: an R package for prediction of clinical chemotherapeutic response from tumor gene expression levels,” PLoS One, vol. 9, no. 9, p. e107468, 2014.
View at: Publisher Site | Google Scholar
B. Li, E. Severson, J. C. Pignon et al., “Comprehensive analyses of tumor immunity: implications for cancer immunotherapy,” Genome Biology, vol. 17, no. 1, p. 174, 2016.
View at: Publisher Site | Google Scholar
K. Yoshihara, M. Shahmoradgoli, E. Martínez et al., “Inferring tumour purity and stromal and immune cell admixture from expression data,” Nature Communications, vol. 4, no. 1, p. 2612, 2013.
View at: Publisher Site | Google Scholar
B. Ru, C. N. Wong, Y. Tong et al., “TISIDB: an integrated repository portal for tumor-immune system interactions,” Bioinformatics, vol. 35, no. 20, pp. 4200–4202, 2019.
View at: Publisher Site | Google Scholar
G. Yu, L. G. Wang, Y. Han, and Q. Y. He, “clusterProfiler: an R package for comparing biological themes among gene clusters,” OMICS, vol. 16, no. 5, pp. 284–287, 2012.
View at: Publisher Site | Google Scholar
C. J. Liu, F. F. Hu, M. X. Xia, L. Han, Q. Zhang, and A. Y. Guo, “GSCALite: a web server for gene set cancer analysis,” Bioinformatics, vol. 34, no. 21, pp. 3771-3772, 2018.
View at: Publisher Site | Google Scholar
Y. Han, D. Liu, and L. Li, “Potential predictive value of serum targeted metabolites and concurrently mutated genes for EGFR-TKI therapeutic efficacy in lung adenocarcinoma patients with <i>EGFR</i> sensitizing mutations,” American Journal of Cancer Research, vol. 10, no. 12, pp. 4266–4286, 2020.
View at: Publisher Site | Google Scholar
B. A.-O. Rowshanravan, N. A.-O. Halliday, and D. A.-O. Sansom, “CTLA-4: a moving target in immunotherapy,” Blood, The Journal of the American Society of Hematology, vol. 131, no. 1, pp. 58–67, 2018.
View at: Publisher Site | Google Scholar
W. Henning and H. W. Stürzbecher, “Homologous recombination and cell cycle checkpoints: Rad51 in tumour progression and therapy resistance,” Toxicology, vol. 193, no. 1-2, pp. 91–109, 2003.
View at: Publisher Site | Google Scholar
A. Mehdi and S. A.-O. Rabbani, “Role of methylation in pro- and anti-cancer immunity,” Cancers, vol. 13, no. 3, p. 545, 2021.
View at: Publisher Site | Google Scholar
E. Orhan, C. Velazquez, I. Tabet, C. Sardet, and C. Theillet, “Regulation of RAD51 at the transcriptional and functional levels: what prospects for cancer therapy?” Cancers, vol. 13, no. 12, p. 2930, 2021.
View at: Publisher Site | Google Scholar
A. Slupianek, G. Hoser, I. Majsterek et al., “Fusion tyrosine kinases induce drug resistance by stimulation of homology-dependent recombination repair, prolongation of G(2)/M phase, and protection from apoptosis,” Molecular and Cellular Biology, vol. 22, no. 12, pp. 4189–4201, 2002.
View at: Publisher Site | Google Scholar
S. J. Pearson, T. Roy Sarkar, C. M. McQueen et al., “ATM-dependent activation of SIM2s regulates homologous recombination and epithelial-mesenchymal transition,” Oncogene, vol. 38, no. 14, pp. 2611–2626, 2019.
View at: Publisher Site | Google Scholar
P. H. R. da Silva, B. C. Borges, I. A. Uehara, L. R. Soldi, R. A. de Araújo, and M. J. B. Silva, “Chemokines and the extracellular matrix: set of targets for tumor development and treatment,” Cytokine, vol. 144, p. 155548, 2021.
View at: Publisher Site | Google Scholar
A. Gao, X. Liu, W. Lin et al., “Tumor-derived ILT4 induces T cell senescence and suppresses tumor immunity,” Journal for Immunotherapy of Cancer, vol. 9, no. 3, p. e001536, 2021.
View at: Publisher Site | Google Scholar
B. Farhood, M. Najafi, and K. A.-O. Mortezaee, “CD8(+) cytotoxic T lymphocytes in cancer immunotherapy: a review,” Journal of Cellular Physiology, vol. 234, no. 6, pp. 8509–8521, 2019.
View at: Publisher Site | Google Scholar
D. Ostroumov, N. Fekete-Drimusz, M. Saborowski, F. Kühnel, and N. A.-O. Woller, “CD4 and CD8 T lymphocyte interplay in controlling tumor growth,” Cellular and Molecular Life Sciences, vol. 75, no. 4, pp. 689–713, 2018.
View at: Publisher Site | Google Scholar
H. A.-O. Rundqvist, P. Veliça, L. Barbieri et al., “Cytotoxic T-cells mediate exercise-induced reductions in tumor growth,” eLife, vol. 9, 2020.
View at: Publisher Site | Google Scholar
I. A.-O. Sahin, M. Akce, O. Alese et al., “Immune checkpoint inhibitors for the treatment of MSI-H/MMR-D colorectal cancer and a perspective on resistance mechanisms,” British Journal of Cancer, vol. 121, no. 10, pp. 809–818, 2019.
View at: Publisher Site | Google Scholar
A. Thomas, E. D. Routh, A. Pullikuth et al., “Tumor mutational burden is a determinant of immune-mediated survival in breast cancer,” Oncoimmunology, vol. 7, no. 10, p. e1490854, 2018.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2022 Han Lu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

463

Downloads

323

Citations