Insights into the Oncogenic, Prognostic, and Immunological Role of BRIP1 in Pan-Cancer: A Comprehensive Data-Mining-Based Study

Background BRCA1 interacting helicase 1 (BRIP1), an ATP-dependent DNA helicase which belongs to an Iron-Sulfur (Fe-S) helicase cluster family with a DEAH domain, plays a key role in DNA damage and repair, Fanconi anemia, and development of several cancers including breast and ovarian cancer. However, its role in pan-cancer remains largely unknown. Methods BRIP1 expression data of tumor and normal tissues were downloaded from the Cancer Genome Atlas, Genotype-Tissue Expression, and Human Protein Atlas databases. Correlation between BRIP1 and prognosis, genomic alterations, and copy number variation (CNV) as well as methylation in pan-cancer were further analyzed. Protein-protein interaction (PPI) and gene set enrichment and variation analysis (GSEA and GSVA) were performed to identify the potential pathways and functions of BRIP1. Besides, BRIP1 correlations with tumor microenvironment (TME), immune infiltration, immune-related genes, tumor mutation burden (TMB), microsatellite instability (MSI), and immunotherapy as well as antitumor drugs were explored in pan-cancer. Results Differential analyses showed an increased expression of BRIP1 in 28 cancer types and its aberrant expression could be an indicator for prognosis in most cancers. Among the various mutation types of BRIP1 in pan-cancer, amplification was the most common type. BRIP1 expression had a significant correlation with CNV and DNA methylation in 23 tumor types and 16 tumor types, respectively. PPI, GSEA, and GSVA results validated the association between BRIP1 and DNA damage and repair, cell cycle, and metabolism. In addition, the expression of BRIP1 and its correlation with TME, immune-infiltrating cells, immune-related genes, TMB, and MSI as well as a variety of antitumor drugs and immunotherapy were confirmed. Conclusions Our study indicates that BRIP1 plays an imperative role in the tumorigenesis and immunity of various tumors. It may not only serve as a diagnostic and prognostic biomarker but also can be a predictor for drug sensitivity and immunoreaction during antitumor treatment in pan-cancer.


Introduction
Cancer remains a thorny problem which brings immense sufering to individual health and fnancial burden. Despite the tremendous advances in the detection of novel biomarkers and development of targeted drugs as well as immunotherapies in recent decades, the high morbidity and mortality of cancer is still frustrating. According to the GLOBOCAN 2020 statistics, there were approximately 19.3 million new cases and 10.0 million deaths related to cancer worldwide in 2020, and the global cancer burden was expected to reach 28.4 million cases in 2040 with a rise of 47% from 2020 [1]. Terefore, persistent eforts are urgently needed to understand the complex mechanisms of tumorigenesis and identify novel biomarkers for early diagnosis, clinical prognosis, and therapy response. Tanks to various public databases, valuable data can be mined and pan-cancer analysis can be conducted for a comprehensive investigation of extracted genes. BRIP1 (BRCA1 interacting helicase 1), also known as FANCJ (as the gene mutated in the J complementation group of Fanconi anemia) or BACH1 (BRCA1-associated Cterminal helicase), was frst discovered in 2001 by its interaction with BRCA1 [2]. BRIP1 is a protein coding gene which encodes for homologous recombination repair (HRR)related protein and facilitates DNA single-strand break (SSB) and DNA double-strand break (DSB) repair during vital biological processes including DNA replication, transcriptional regulation, and overall metabolic health [3]. BRIP1, whose encoded protein belongs to an Iron-Sulfur (Fe-S) helicase cluster family with a DEAH domain, helps to preserve chromatin structure and function and may also maintain genomic and epigenetic stability. Besides its collaboration with numerous DNA metabolizing proteins implicated in the detection and repair of DNA damage, BRIP1 also participates in cell cycle checkpoint control [4]. Recent studies manifest that BRIP1 took part in miscellaneous tumorigeneses and pathological conditions. Te National Comprehensive Cancer Network (NCCN) guidelines identifed BRIP1 as a potential risk factor for breast cancer, especially for triple negative breast cancers [5]. In ovarian cancer, a deleterious mutation of BRIP1 was associated with low-grade histology and led to an increased risk of the disease [6]. In endometrial cancer, BRIP1 correlated to tumor recurrence and patients with mutations in BRIP1 might beneft from poly ADP-ribose polymerase (PARP) inhibitors [7]. Mikaeel et al. reported that BRIP1 might be a cancerpredisposing gene in young-onset colorectal cancer [8]. Mani et al. suggested that BRIP1 was of the imperative role in maintaining neuronal cell health and homeostasis by suppressing oxidative stress, excitotoxicity induced DNA damage, and protecting mitochondrial integrity [3]. However, there is a lack of a comprehensive pan-cancer analysis of BRIP1. Hence, we extracted diverse data from Te Cancer Genome Atlas (TCGA), Genotype-Tissue Expression (GTEx), Cancer Cell Line Encyclopedia (CCLE), Human Protein Atlas (HPA), cBioPortal and GeneMANIA databases and evaluated the expression, prognosis, and mutation as well as function of BRIP1 in various cancer types. We further carried out immune infltration analysis, and the relationships between BRIP1 and immune-related genes and tumor mutation burden (TMB)-microsatellite instability (MSI) as well as immunotherapy and targeted drug responses were subsequently analyzed. Tis in-depth data-mining based study helped us understand the role of BRIP1 in tumorigenesis, provided evidence for its diagnostic and prognostic evaluation in the clinic, and shed light on the novel targeted treatment as well as immunotherapy in pan-cancer.

Raw Data Collection and Diferential Expression Analysis.
Te mRNA expression profles and related clinical information of 33 human cancers and their corresponding normal samples were, respectively, downloaded from TCGA via the UCSC Xena platform (https://xena.ucsc.edu/) [9]. Additional gene expression data were also retrieved from GTEx (https://gtexportal.org/home/datasets) and CCLE (https://sites.broadinstitute.org/ccle). BRIP1 expression was transferred to transcripts per million (TPM) and then evaluated by log 2 transformation. T-test was carried out to identify its diferent expression between tumor and normal tissues as well as between diferent TNM stages. R software (Version 4.0.3, https://www.Rproject.org) and the "ggplot2" R package (Version 3.3.3) were applied to analyze the data and draw box diagrams. Te abbreviations and full names of the various cancer types were listed in Table 1. Besides, to evaluate the diferential expression of BRIP1 at the protein level, immunohistochemistry (IHC) images in multiple tumors and normal tissues were downloaded from HPA (https://www.proteinatlas.org/). Te antibody used for IHC was HPA005474.

Prognostic Value of BRIP1 in Pan-Cancer.
Overall survival (OS), disease-specifc survival (DSS), disease-free interval (DFI), and progression-free interval (PFI) were of vital importance in exploring the association between BRIP1 expression and prognosis. Related survival data were downloaded from the UCSC Xena platform. Te Kaplan-Meier (KM) method and log-rank test were utilized to carry out survival analyses in each cancer with the best cut-of value of BRIP1 expression by using R packages "survminer" and "survival." Univariate Cox regression and R package "forestplot" were also used to identify the relevancy between BRIP1 expression and survival in pan-cancer. Te hazard ratio (HR) and Cox's regression P values were shown in the plot.

BRIP1 Mutation and Its Correlation with Copy Number
Variation and DNA Methylation. To further investigate the modifcation of BRIP1 gene in pan-cancer, we used the cBioPortal database (https://cbioportal.org) to explore its mutation, structural variant, amplifcation, deep deletion, and multiple alterations [10]. As copy number variation (CNV) and copy number alteration (CNA) played a critical role in cancer initiation and progression, and promoter methylation was critical in gene silencing and inactivation, related data were downloaded from cBioPortal for further analyses. Association between the expression of BRIP1 and CNV as well as promoter methylation was further evaluated by carrying out Pearson correlation analysis. R software and the "ggplot2" R package were acquired to analyze the data and draw lollipop plots.

Gene Interaction of BRIP1 and Its Enrichment and
Variation Analysis. Te GeneMANIA database (https:// www.genemania.org) was applied to detect functionally similar genes to BRIP1 and construct the protein-protein interaction (PPI) network [11,12]. Subsequently, gene set enrichment analysis (GSEA) was performed in pan-cancer based on the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) database to explore the biological signalling pathway by using R package "cluster-Profler," and R package "ridgeplot" was used to draw the ridge plot [13,14]. We further downloaded the "gmt" fle of the 50 hallmark gene sets from the Molecular Signatures Database (MSigDB, via https://www.gsea-msigdb.org/gsea/ msigdb/index.jsp) [15,16] and performed gene set variation analysis (GSVA) using the "GSVA" R package to explore the correlation between BRIP1 expression and 50 hallmark pathways. Pearson correlation analysis was conducted, and the "pheatmap" R package was used to turn the results into heatmap.

BRIP1 Expression and Its Relationship with Immunity.
Tumor microenvironment (TME), a crucial element of tumor, has been reported to play a decisive role in cancer development and therapeutic responses. Hence, we carried out evaluation of the association between BRIP1 expression and the proportion of immune-stromal component in pancancer. Data were downloaded from TCGA via the UCSC Xena platform, while R package "ESTIMATE" was used to evaluate the immune score, stromal score, and tumor purity score. Subsequently, the specifc tumor-infltrating immune cells (TIICs) and its correlation to BRIP1 expression were assessed via Tumor Immune Estimation Resource (TIMER) database (https://timer.cistrome.org/) [17]. Te TIMER, EPIC, MCPCOUNTER, CIBERSORT, CIBERSORT_ABS, XCELL, and QUANTISEQ algorithms were utilized to estimate the immune infltration of the 21 TIICs. Relationship between BRIP1 expression and immune-related genes was also evaluated at the pan-cancer level. Te visualization of the results was implemented with R packages "ggplot2" and "pheatmap." Besides, TMB, which refects cancer mutation quantity, has been considered as a leading candidate biomarker for immune checkpoint blockade (ICB) [18]. Meanwhile, MSI, which facilitates mutation and acts as a biomarker of response to immune checkpoint inhibitors (ICPis), plays an important role in improving the possibility of a favorable response to immunotherapy [19]. We thus analyzed the TMB-MSI association with BRIP1 in pancancer by Pearson correlation using Sangerbox tools (https://vip.sangerbox.com/home.html), and the results were shown in radar maps.

BRIP1 Expression and Diferent Terapies.
To further validate the relationship between BRIP1 and ICB therapy response, data from the IMvigor210 cohort, which contains 298 metastatic urothelial cancer cases treated by atezolizumab (an antiprogrammed cell death ligand 1, anti-PD-L1 agent), were obtained and analyzed [20]. Patients were divided into two subgroups, one with a low level of BRIP1 and the other with a high level of BRIP1, according to the best cut-of value identifed by the "survminer" R package, and immunotherapy response of BRIP1 was then validated. A chi-square test was carried out to assess the proportion diferences of responses between subgroups. Furthermore, relationships between BRIP1 and IC50 of numerous antitumor drugs were explored via the Genomics of Drug Sensitivity in Cancer (GDSC) database (https://www.cancerrxgene.org). A Spearman correlation was used to evaluate the drug resistance.

Results and Discussion
3.1. BRIP1 Expression Profle. Te expression level of BRIP1 explored via the GTEx transcriptomics dataset indicated it was low in most normal tissues under physiological circumstances, whereas higher in bone marrow than other 30 tissues (Figure 1(a)). Results from CCLE revealed that its expression level was generally increased in various cancer cell lines as the highest expression was in NB, ALL, and SCLC ( Figure 1(b)). TCGA data showed a similar expression tendency to that of CCLE, and the highest expression level of BRIP1 was in LAML and genital cancers such as CESC and TGCT ( Figure 1 Figure S1). When further seeking for the association between BRIP1 expression and diferent tumor stages, we found that there was a signifcant diference between stage I, II and stage III, IV in ACC, KIRP, LUAD, and OV ( Figure S2). Subsequently, the protein level of BRIP1 was explored in multiple tumor and normal tissues. Representative IHC images showed that BRIP1 was mostly enriched in the nucleoplasm and nuclear membrane and had a low expression level in normal tissues than that of tumor tissues in breast, cerebellum, cervix, colon, endometrium, kidney, liver, lung, lymph node, ovary, pancreas, prostate, skin, stomach, thyroid gland, and urinary bladder, while high in normal testis tissues than tumor tissues ( Figure 2). When considering the relationship between BRIP1 expression and DFI, there was a signifcant association between them in KIRP (P < 0.001), THCA (P � 0.002), PAAD (P � 0.003), and LIHC (P � 0.026). Moreover, BRIP1 was a high-risk factor in all of these four cancers ( Figure 5(a)). In addition, poor DFI was perceived in BLCA, KIRP, LIHC, LUAD, LUSC, MESO, PAAD, SARC, and THCA as BRIP1 upregulated in these tumors, while in COAD, DLBC, KIRC, READ, STAD, UCEC, and UCS as BRIP1 downregulated (Figures 5(b)-5(q)).

Correlation between BRIP1 Expression and CNV as well as DNA Methylation in Pan-Cancer.
Te previous fndings indicated that BRIP1 might play a role in the carcinogeneses and it was widely accepted that the genomic mutation was associated with tumorigenesis. Terefore, a comparative analysis of genomic mutations of BRIP1 in pan-cancer was conducted. Results from the cBioPortal database consisting of 32 cancer types and 10953 tumor samples showed that the amplifcation of BRIP1 was one of the most vital single factors for alteration. It accounted for 6.92%, 4.6%, and 3.14% in BRCA, MESO, and sarcoma, respectively, as the largest proportion of all mutation types among these tumors (Figure 7

Interacting Genes of BRIP1 and Its Enrichment and
Variation Analysis. Te PPI network for BRIP1 and its coexpressed as well as colocalized genes were constructed by GeneMANIA. Te results showed the 20 most frequently altered proteins closely linked to BRIP1, in which BRCA1 had the most prominent correlation with BRIP1 as expected. Besides, the functional analysis indicated that BRIP1 and its similar genes had a signifcant association with DNA recombination, double-strand break repair, and recombinational repair (Figure 8). To uncover the function of BRIP1, we carried out GSEA in 33 cancer types, and the results suggested that the top 6 signalling pathways correlated with BRIP1 among all cancers based on KEGG were DNA replication, cell cycle, spliceosome, nucleocytoplasmic transport, homologous recombination, and Fanconi anemia pathway. Te specifc 20 signalling pathways associated with BRIP1 in each type of tumor are summarized in Figure S5. As for GSVA, the relationship between BRIP1 and various hallmark pathways in pan-cancer is shown in the heatmap (Figure 9). It was obvious that BRIP1 had the most signifcantly positive correlation with G2M checkpoint and E2F targets in ACC, BLCA, LGG, LUSC, PCPG, and THYM, with mitotic spindle and G2M checkpoint in BRCA, ESCA, MESO, OV, PAAD, and PRAD, with mitotic spindle in CESC, CHOL, DLBC, SARC, TGCT, and UCS, with G2M checkpoint in COAD, GBM, STAD, and THCA, with mitotic spindle, G2M checkpoint, and E2F targets in HNSC, KIRC, KIRP, LAML, LIHC, and LUAD, with G2M checkpoint and MYC targets V1 in KICH, with MYC targets V1 in READ and UCEC, with mitotic spindle and MYC targets V1 in SKCM, and with MYC targets V1 and protein secretion in UVM. Moreover, the most prominently negative correlation between BRIP1 and xenobiotic metabolism lay in ACC, BLCA, DLBC, READ, SARC, and UCEC, between BRIP1 and xenobiotic metabolism as well as myogenesis lay in BRCA, between BRIP1 and coagulation as well as KRAS signalling upregulation lay in CESC, between BRIP1 and xenobiotic metabolism and myogenesis as well as P53 pathway lay in COAD, between BRIP1 and xenobiotic metabolism and adipogenesis as well as complement lay in ESCA, between BRIP1 and xenobiotic metabolism as well as adipogenesis lay in GBM, between BRIP1 and KRAS signalling downregulation lay in HNSC, KICH, and UVM, between BRIP1 and KRAS signalling downregulation as well as oxidative phosphorylation lay in KIRC, between BRIP1 and xenobiotic metabolism as well as oxidative phosphorylation lay in KIRP, between BRIP1 and coagulation as well as P53 pathway lay in LAML, between BRIP1 and bile acid metabolism as well as heme metabolism lay in LGG, between BRIP1 and myogenesis lay in LIHC and STAD, between BRIP1 and bile acid metabolism as well as fatty acid metabolism lay in LUAD, between BRIP1 and  Journal of Oncology coagulation as well as adipogenesis lay in LUSC, between BRIP1 and xenobiotic metabolism and bile acid metabolism as well as fatty acid metabolism lay in MESO, between BRIP1 and bile acid metabolism lay in OV, between BRIP1 and pancreas beta cells lay in PAAD, between BRIP1 and apical surface lay in PCPG, between BRIP1 and xenobiotic metabolism as well as KRAS signalling downregulation lay in PRAD, between BRIP1 and xenobiotic metabolism, myogenesis as well as KRAS signalling downregulation lay in SKCM, between BRIP1 and P53 pathway lay in TGCT, between BRIP1 and fatty acid metabolism lay in THCA, and between BRIP1 and xenobiotic metabolism, myogenesis, and P53 pathway as well apical junction lay in THYM. In summary, the previous results elucidated the hallmark pathways and potential mechanisms of BRIP1 in pan-cancer. In essence, BRIP1 kept an intimate relationship with HRR, cell cycle, and varied metabolism in diferent cancers.

BRIP1 Expression and Its Correlation with TME and
Immune Infltration. Along with the above coexpressed genes and signalling pathways, TME and immune infltration also take part in the regulation of tumorigenesis. As part of the complex microenvironment, TIICs have a crucial role in cancer progression and therapeutic responses. Accordingly, we explored the correlation between BRIP1 expression and TME by ESTIMATE and evaluated the coefcient of BRIP1 expression and immune infltration level via TIMER. Te results revealed that the expression of BRIP1 had signifcant correlations with tumor purity and ESTIMATE-Score in 19 cancer types ( Figure 10). Te top three most signifcant cancers associated with BRIP1 expression were GBM, SARC, and LUSC based on ImmuneScore, Stromal-Score, and ESTIMATEScore. Te higher the level of BRIP1 as these three tumors expressed, the less stromal and immune cells as these tumors had. On the contrary, the higher level of BRIP1 as these tumors expressed, the higher purity as these tumors had. Results of ESTIMATEScore for all tumor types are listed in Figure S6. In addition, BRIP1 expression and its association with TIICs were signifcant in most cancer types ( Figure 11). Especially in THYM, BRIP1 had a positive correlation with B cells, memory and naive CD4+ T cells, CD8+ T cells, myeloid dendritic cells, neutrophils (by CIBERSORT, CIBERSORT_ABS, and TIMER algorithm), common lymphoid progenitor, granulocyte and monocyte progenitor, and a negative correlation with fbroblast, endothelial cell, eosinophil, macrophage, mast cell, monocyte, neutrophil (by XCELL and MCPCOUNTER algorithm), NK cell, and common myeloid progenitor. As for those digestive tumors, the most signifcant association between BRIP1 and TIICs was found in STAD. A signifcantly negative correlation was found between BRIP1 and fbroblast as well as hematopoietic stem cell in STAD. Moreover, the association between the expression of BRIP1 and immune-related genes including immuneactivating genes, immunosuppressive genes, mismatch repair (MMR) genes, and genes encoding the major histocompatibility complex (MHC), chemokine, and chemokine receptor proteins is evaluated (Figure 12). Results indicated that BRIP1 was positively correlated with the majority of immune-activating genes as well as immunosuppressive genes in UVM, KIRC, THCA, KICH, PAAD, HNSC, PRAD, and OV. In HNSC, OV, PRAD, UVM, LAML, TGCT, UCEC, READ, SKCM, LIHC, ESCA, DLBC, LUSC, BLCA, GBM, BRCA, CESC, KICH, STAD, SARC, LGG, PAAD, PCPG, and KIRC, the expression of BRIP1 was positively correlated with most of the MMR genes. As for the correlation between BRIP1 and the majority of genes encoding the chemokine and chemokine receptor proteins, a positive correlation was found in THCA for the former and in KIRC and PRAD for the latter. Besides, there was a positive correlation between BRIP1 and most of the MHC-related genes in KIRC, PAAD, UVM, and THCA, and a negative correlation between BRIP1 and most of the MHC-related genes in THYM, GBM, and LUSC.

Correlation between BRIP1, TMB/MSI, and Immunotherapy Response.
To discover the role of BRIP1 in predicting the response to ICPis, we assessed the correlation between BRIP1 expression and the two famous biomarkers, TMB and MSI. BRIP1 was positively associated with TMB in 2 digestive tumors (COAD and STAD) and other 7 tumors including KICH, LUAD, ACC, OV, PRAD, KIRC, and SKCM (Figures 13(a) and S7). As for MSI, it was negatively correlated with BRIP1 in DLBC and positively correlated with BRIP1 in 3 digestive tumors (COAD, READ, and STAD) and other 4 tumors including GBM, LUSC, KIRC, and LUAD (Figures 13(b) and S8). As indicated by previous studies that high TMB/MSI-H increased patients' response to ICPis and was correlated to better immunotherapy outcomes, we therefore, downloaded data of the IMvigor210 cohort to investigate the correlation between BRIP1 and treatment response. Results showed that in this urothelial cancer cohort, patients with a high level of BRIP1 had a better response to the treatment and a more favorable survival rate (Figures 13(c) and 13(d)). Moreover, the anti-PD-L1 response rate was 49% among patients with a high expression level of BRIP1, while there were only 19% of the low-BRIP1 patients responding to the treatment (Figure 13(e)). Tese results showed the potential of BRIP1 in predicting immunotherapy response and BRIP1 could be a promising candidate biomarker for immunotherapy of various cancers.

BRIP1 and Antitumor Drugs.
Other than immunotherapy, the relationships between BRIP1 and IC50 of numerous antitumor drugs are also evaluated (Table S1). Among the 192 antitumor drugs, 141 of them including Olaparib and Niraparib (two PARP inhibitors) were negatively correlated with BRIP1, which indicated a promising response in these treatments. Besides, 7 drugs including Trametinib, SCH772984, ERK_2440, ERK_6604, Selumetinib, Ulixertinib, and VX-11e were positively correlated with BRIP1 which indicated a potential resistance during treatment.

Discussion
Cancer is a complex polyfactorial disease with high morbidity and mortality, remaining as an unsolved threaten to human health. Tus, research of efective diagnostic biomarkers and therapeutic targets for tumors has always been a heated focus. With the availability of public databases, cancer-related data can be mined to explore novel biomarkers. Trough pan-cancer analysis, BRIP1 emerged from a bunch of candidate genes who were applicable for broadspectrum tumor diagnosis as it signifcantly upregulated in most tumors. Herein, we conducted a systematic and comprehensive analysis of BRIP1 in pan-cancer. We validated its diferential expression in various cancers between tumor and normal tissues at transcriptional and protein levels. Subsequently, we elucidated its role in prognosis, gene function, and regulatory pathways, and we discovered its  TGCT  THYM  SARC  ESCA  CHOL  LUSC  BRCA  UCEC  MESO  KICH  CESC  SKCM  BLCA  OV  COAD  LIHC  STAD  KIRP  HNSC  UVM  READ  LUAD  THCA  KIRC  PCPG  ACC  PAAD  PRAD  UCS LGG LAML GBM (c) association with TME, immune infltration, immune-related genes, and treatment responses. BRIP1, with a length of more than 180 kb, is located on chromosome 17q23.2 and encodes a protein of 1249 amino acids. Previous studies regarded BRIP1 as a tumor suppressor gene and revealed its diagnostic role in various types of cancer, such as breast cancer, ovarian cancer, cervical cancer, and colon cancer [21][22][23][24]. In our comprehensive data mining-based analysis, by analyzing data from the GTEx, CCLE, and TCGA databases, we revealed that BRIP1 expression was higher in 28 types of cancer tissues (including BRCA, CESC, COAD, and OV, in consistence with previous study results) and only lower in TGCT than in normal tissues. Furthermore, our results of diferential expression analysis of paired samples and the results of IHC analysis also confrmed the diagnostic role of BRIP1 in pan-cancer. Unfortunately, due to the lack of normal sample data, diferential expression analysis could not be conducted in MESO and UVM. Accumulating evidence will be needed for further exploration in these two tumors. Besides, we found a signifcant diferential expression between tumor stage I, II and stage III, IV in ACC, KIRP, LUAD, and OV, suggesting the predicting role of BRIP1 in early diagnosis of these cancers is worth looking forward to. Along with its predicting role in diagnosis, we also performed prognostic analyses in pan-cancer based on data from TCGA. Either from OS or DSS, as well as from DFI and PFI, we found a signifcant correlation between BRIP1 expression and survival probability in various cancers, among which, BRIP1 was basically a high-risk factor. Whether it be OS, DSS, DFI, or PFI, BRIP1 remained as a high-risk factor in KIRP and PAAD. Although our results from TCGA database did not  fnd a correlation between BRIP1 expression and prognosis of breast cancer patients, a study based specifcally on several breast cancer databases exhibited that higher BRIP1 expression was correlated with poor OS, DSS, DFI, and PFI [25]. Another study mining data of LUAD patients from the Genomic Data Commons (GDC) Data Portal indicated that BRIP1 might regulate fbroblast growth factor 22 and afect MAPK as well as Rap 1 signalling pathways in all tumor stages of LUAD, and a high level of BRIP1 showed boundary signifcance on OS [26], in consistence with our results. Synthesizing the previous results, we believed that the high expression of BRIP1 could hamper cancer patients' survival and it might be an independent prognostic factor for various tumors. Although BRIP1 seemed to be a novel biomarker of  MYC TARGETS V1   E2F TARGETS   G2M CHECKPOINT   MITOTIC SPINDLE   ACC  BLCA  BRCA  CESC  CHOL  COAD  DLBC  ESCA  GBM  HNSC  KICH  KIRC  KIRP  LAML  LGG  LIHC  LUAD  LUSC  MESO  OV  PAAD  PCPG  PRAD  READ  SARC  SKCM  STAD  TGCT  THCA  THYM  UCEC Figure 9: GSVA of BRIP1 in pan-cancer. * represents P < 0.05, * * represents P < 0.01, * * * represents P < 0.001, and * * * * represents P < 0.0001.
vital clinical utility in predicting diagnosis and prognosis in pan-cancer, the distinct efects of the diferential expression of BRIP1 on protein function in various cancer types remain largely unknown. Previously, a meta-analysis based on 29400 patients with 116000 controls from 63 studies found BRIP1 was associated with a high risk of ovarian cancer and the HRR pathway might be involved [27,28]. A cohort of more than 117000 patients elucidated the missense variant of BRIP1 conferred risk for ovarian and breast cancer. Researchers further studied the functional characterization of BRIP1, revealing an impaired interstrand crosslink (ICL) repair of DNA due to the missense variants of BRIP1 [29]. In an Asian esophageal squamous cell carcinoma cohort, researchers found that BRIP1 mutant was an adverse factor for OS and the cohort harboured TP53 signalling pathway alterations altered NOTCH, RTK-RAS, and cell cycle pathway, which might explain the phenomenon [30]. As reported by Singh, via quantitative real-time polymerase chain reaction (qRT-PCR) and Caspase-3 immunostaining, they found that the loss of DNA repair genes expression including BRIP1 in testis correlated with increased apoptosis [31]. To sum up, the distinct efects of the diferential expression of BRIP1 in various cancer types may rely on diferent signalling pathways. Furthermore, in vivo and in vitro experiments are needed to validate the above fndings and elucidate the specifc underlying mechanisms of BRIP1 in diferent types of cancer. Te genomic mutation analysis revealed that the amplifcation of BRIP1 was one of the most vital single factors for alteration. Interestingly, previous studies reported that amplifcation of the 17q23 region led to a gain of function in lung cancer, liver cancer, pancreatic cancer, bladder cancer, testis cancer, and ovarian cancer [32]. Since this is the region where BRIP1 locates and with our fnding of BRIP1 amplifcation and its role in pan-cancer, the phenomenon shall be explained to some extent. In addition, PPI analysis revealed that BRIP1 was mainly associated with DNA recombination, double-strand break repair, and recombinational repair. Enrichment analysis uncovered its correlation with homologous recombination, DNA replication, cell cycle, and Fanconi anemia. As indicated by previous studies, BRIP1 took part in HRR and helped in reducing the occurrence and persistence of DSB which was regarded as the last defense against feasibly mutagenic and carcinogenic injury [33]. Tese might explain the underlying mechanisms of BRIP1 in tumorigenesis and provide a theoretical foundation for the discovery and development of targeted drugs. For example, Hodgson et al. illuminated in their study that ovarian cancer patients with loss-of-function mutations in HRR genes, including BRIP1, would beneft from Olaparib treatment [34]. Our study has evaluated the association between BRIP1 and IC50 of various antitumor drugs via GDSC database and found the same promising response to Olaparib. Except a few of antitumor drugs, there were 141 drugs negatively correlated with BRIP1, which indicated a promising treatment response. Furthermore, clinical trials with diferent drugs in diverse cancers and research on their targeted signalling pathways are urgently needed to validate efective targeted-therapies.

Conclusions
Tis study highlights the potential role of BRIP1 in pancancer as a predictor for diagnosis, prognosis, and treatment response through in-depth analyses of diferential expression, relationships between BRIP1 and diferent prognostic parameters, gene functions, regulatory pathways, TME, TIICs, immune-related genes, and TMB-MSI as well as anticarcinogen. Furthermore, functional and mechanistic experiments are needed to elucidate the role of BRIP1 in specifc cancers.

Data Availability
Te original datasets analyzed in this study can be found in the corresponding websites as indicated in the article. Requests for further access to datasets can be directed to the corresponding author.

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Authors' Contributions
L-YR was involved in study design, data acquisition and analysis, result interpretation, and manuscript draft. WX was responsible for result discussion. Z-SY and WQ were (e) Response rate to PD-L1 therapy in diferent BRIP1 expression subgroups of IMvigor210 cohort. CR represents complete response, PR represents partial response, PD represents progressive disease, and SD represents stable disease. * represents P < 0.05, * * represents P < 0.01, * * * represents P < 0.001, and * * * * represents P < 0.0001.
involved in data acquisition. J-QW and F-YL were involved in data analysis. Y-AM was responsible for study concept, critical revision and supervision, and fnal approval of the submission. All authors have read and approved the fnal version of our study.