{"paper_id":"4555fed0-a0e9-4f73-a830-8c12996c01a6","body_text":"Analysis of BGN and pan-cancer correlations: based a Mendelian randomisation and bioinformatics study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Analysis of BGN and pan-cancer correlations: based a Mendelian randomisation and bioinformatics study GuangTao Min, Hao Tang, GuangNing Min, YuMin Li This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5873854/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background/Objectives: To preliminarily explore the significance of BGN in various cancers using pan-cancer analysis. Methods: Transcriptome data of 33 cancers were downloaded from TCGA database, and the expression levels of BGN in 33 cancers were extracted using Perl software. The limma package of R software was used to identify differential genes in some tumour types (paraneoplastic samples ≥5), and the clinical prognostic significance of BGN was analysed using Kaplan-Meier and Cox in conjunction with clinical information from TCGA. Mendelian randomisation was used to test for causal associations between BGN and multiple malignancies. The correlation between BGN and the tumour immune infiltration microenvironment was explored by ESTIMATE package, and the relationship between BGN and immune subtypes, clinical stage and tumour immune infiltration microenvironment were analysed in conjunction with TCGA-GTEx data. Analysis of clinical data of 180 patients with gastric cancer and immunohistochemical verification of the poor prognosis of BGN in gastric cancer. Results: BGN was significantly differentially expressed in most of the tumours, and MR analysis revealed potential causal associations with colorectal, lung and cervical cancers, etc. BGN showed prognostic correlations with a variety of cancers in survival analyses (P < 0.05). Single-tumour analyses showed correlations between BGN and TNM staging, immune subtypes, and the tumour microenvironment. BGN expression promotes immune cell infiltration and expression of immune checkpoint-associated genes in the tumour microenvironment, and the higher the level of expression, the greater the stromal component and the less the immune component.BGN may be expressed via ECM receptor interaction, BGN may be involved in the process of tumour immune and inflammatory responses through ECM receptor interaction, ascorbate and aldarate metabolism, and other signalling pathways. Conclusion: BGN plays an important role in tumour development and is expected to become a new prognostic marker and a potential target for immunotherapy in many types of cancers. Biglycan pan-cancer Mendelian randomization bioinformatics study Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 1. Introduction Cancer is a major public health and economic problem for a nation, a country, and even all of humanity. According to the latest statistics from the International Agency for Research on Cancer (IARC), there will be nearly 20 million new cancer cases and 9.7 million deaths due to cancer in the world in 2022(Bray et al., 2024 ). According to the latest statistics of the International Agency for Research on Cancer (IARC), there will be nearly 20 million new cancer cases and 9.7 million cancer-related deaths worldwide in 2022. Conventional cancer treatments (surgery, radiotherapy and chemotherapy) are often associated with serious complications and still have a poor prognosis and survival rate for patients with advanced disease(Mullard, 2020 ). The prognosis and survival rates for patients with advanced cancer are still unsatisfactory. Gene copy number changes, rearrangements, and mutations are key indicators of genomic instability, while changes in the genome, epigenome, and gene and protein expression are widely used in the diagnosis and treatment of cancer(Pilié, Tang, Mills, & Yap, 2019 ). The Cancer diagnostic biomarkers and therapies are ultimately combined in a \"therapeutic diagnostics\" platform, where personalized treatment plans enable oncologists to tailor treatments to the unique molecular profile of each cancer patient(Passaro et al., 2024 ). The development of cancer is a multifactorial process. Cancer development is a multifactorial process that involves changes in the cellular microenvironment and specific regulation of cellular functions. The identification of oncogenes can help us to better understand carcinogenesis and tumour progression, and broaden the potential therapeutic options for malignant tumours(Cripe, 2023 ). The identification of oncogenes can help us better understand carcinogenesis and tumour progression, and broaden potential treatment options for malignant tumours. Since 2012, the Cancer Genome Atlas has initiated a pan-cancer analysis project, and by 2020, the International Cancer Genome Consortium will have perfected genome-wide pan-cancer analysis data, providing a source of data for pan-cancer research(\"Pan-cancer analysis of whole genomes,\" 2020; Weinstein et al., 2013 ). This has laid a solid foundation for the data source of pan-cancer research. Biglycan (BGN) are hydrophilic proteoglycans (Proteoglycan, PG) present on the cell surface or in the extracellular matrix (ECM), which are released by cells from the endoplasmic reticulum and Golgi apparatus via biosignalling in response to tissue damage, thereby triggering a sustained inflammatory response(Schulz, Diehl, Trebicka, Wygrecka, & Schaefer, 2021 ). Meanwhile, BGN, decorin (DCN), the extracellular matrix2 (ECM2), and Asporin (ASPN) are collectively involved in constituting a family of class I small leucine-rich proteoglycans (SLRPs)(Gopinath, Natarajan, Sathyanarayanan, Veluswami, & Gopisetty, 2022 ). SLRP, as an extracellular compound, interacts with a variety of signalling receptors, such as toll-like receptors, transforming growth factor beta (TGF-β) and tumour necrosis factor-alpha, among others, and these interactions are involved in cell growth, proliferation, differentiation, survival, adhesion and migration under developmental, physiological and pathological conditions. migration under developmental and pathological conditions(Pietraszek-Gremplewicz et al., 2019 ). As there are no pan-cancer studies on the association between BGN and various cancers, and the relationship between BGN and the tumour immune microenvironment has not been reported in the literature. In this study, we comprehensively analysed the BGN expression level and its association with the prognosis of different types of malignant tumours through data mining analysis of various databases, explored the causal relationship between BGN and cancer based on Mendelian randomisation, and explored the potential associations between BGN expression and microsatellite instability (MSI), tumour mutational load (TMB), and the level of immune cell infiltration in cancer. 2. Materials and methods 2.1 Analysis of its subcellular localisation and BGN expression The HPA database ( https://www.proteinatlas.org/ ) is a public database designed to integrate various genomics technologies and explore the human proteome. It provides the tissue and cellular distribution and expression of more than 24,000 human proteins in a wide range of human normal tissues, tumour tissues, cell lines and blood cells, with the results represented by immunohistochemical staining maps. The immunofluorescence-based methodology used in subcellular sections allows for the simultaneous analysis of protein distribution in all organelles and subcellular structures, and we demonstrate the subcellular localisation of BGN proteins and their possible secretion modes through the \"SUBCELL\" module. Using the \"TISSUE\" module and \"PATHOLOGY\" module, we viewed the immunohistochemical staining of BGN in different tumour tissues and their adjacent normal tissues. Meanwhile, BGN mRNA expression data were obtained from deep sequencing of RNA (RNA-seq) from 40 different normal tissue types. Pan-cancer RNA sequencing expression (grade 3) profiles and corresponding clinical information were downloaded from the Cancer Genome Atlas ( https://portal.gdc.cancer.gov/ ). The harmonised TCGA-GTEx (PANCAN, N = 19131, G = 60499) dataset was downloaded from the TCGA database via UCSC Xena ( https://xena.ucsc.edu/ ) (Goldman et al., 2020 ). we further performed a joint analysis of the two data sets using the R language (version 4.4.1) to extract the expression data of the ENSG00000182492 (BGN) gene in individual samples and to analyse the expression of BGN in the above tumour tissues and normal tissues, \"ggpubr\" Box line plots of differential expression were drawn. The expression level of BGN was transformed by log2 (x + 0.001) and then statistically analysed, and Ρ < 0.05 indicated that the difference was statistically significant. 2.2 Analysis of the clinical relevance of BGN in pan-cancer From the TCGA Prognostic Study(J. Liu et al., 2018 ) A high-quality TCGA survival dataset was obtained, supplemented by follow-up data obtained from the UCSC Xena database and excluding samples with a follow-up time shorter than 30 days, and each expression value was log2(x + 0.001) transformed to exclude cancers with fewer than 10 samples in a single cancer type, and then the corresponding expression data and the patient Overall survival, Disease-specific survival, disease-free interval, progression-free interval data were obtained. survival, Disease-specific survival, disease-free interval, progression-free interval data. Univariate Cox proportional risk regression models were used to assess the correlation between BGN expression levels and survival in each cancer type. We used the coxph function of the R package \"survival\" to establish the relationship between BGN expression and the prognosis of the tumour, and performed statistical tests using the log-rank test. The \"Survival Analysis\" module of the GEPIA2 website was used to define the \"off-high (50%)\" and \"off-low (50%)\" values for splitting the high and low values. The \"off-high (50%)\" and \"off-low (50%)\" values were defined as the expression thresholds to separate the high and low expression groups, and a Kaplan-Meier plot with p-value of log-rank was generated to explore the expression patterns of BGN and the survival factors of tumour patients. The R package \"limma\" and \"ggpubr\" were used for clinical staging correlation analysis. 2.3 Mendelian randomisation analysis In order to determine the causal relationship between BGN and cancer, two-sample MR analyses were performed. A total of two common MR analysis methods were used, including the inverse variance weighting (IVW) method(Huang, Lin, He, Wang, & Zhan, 2022 ), MR- egger regression, and complementary weighted median, weighted modal and simple modal methods. According to previous studies, the IVW test has superior advantages over other methods. In most studies, it was used as the primary MR analysis method(Yang, Ma, Pang, & Jiang, 2023 ). Similarly, in our study, IVW was used as the primary test method and other methods were used as references. In addition, the MR-Egger regression test was used to verify the presence of horizontal pleiotropy, and a P < 0.05 was considered as horizontal pleiotropy. To ensure the validity of our findings, leave-one-out sensitivity analyses were performed to determine whether a single SNP was responsible for the results. All statistical tests were performed in the MR package. The test for heterogeneity was significant ( P < 0.05) and heterogeneity was considered to exist between instrumental variables (IVs). 2.4 Analysis of tumour mutational load and microsatellite instability cBioPortal ( http://www.cbioportal.org ) is a comprehensive and interactive web resource for analysing cancer data(Chandrashekar et al., 2022 ). We downloaded the Simple Nucleotide Variation dataset of all TCGA samples processed by MuTect2 software from GDC ( https://portal.gdc.cancer.gov/ ) and integrated the mutation data of the samples(Beroukhim et al., 2010 ). Integrated the mutation data of the samples, and obtained the structural domains of the proteins to evaluate the common mutation types of BGNs in various cancers by using the R language \"maftools\" package. In addition, we used the R language \"TCGAbiolinks\", \"ggplot2\", \"ggpubr\" and \"ggExtra\" packages to download and integrate the mutation annotation files in TCGA, and integrate the pan-cancer BGN expression data with microsatellite instability (MSI) and tumour mutation burden (TMB) scores of the samples. Spearman's correlation was calculated between BGN expression and these two scores in different types of cancers, and the R language \"radarchart\" package was used for image visualisation. 2.5 Association of BGN expression with tumour microenvironment and immune infiltration We also extracted immune checkpoint pathway genes (Inhibitory (24), Stimulatory (36)) and marker genes for immune pathways (chemokine (41), receptor (18), MHC (21), Immunoinhibitor (24)) from the TISIDB database, Immunostimulator(46))(Ru et al., 2019 ). The Spearman correlation between BGN expression and these immunomodulatory genes at the pan-cancer level was calculated using the Sangerbox online platform(Shen et al., 2022 ). The results were finally visualised using the R language packages \"limma\", \"reshape2\" and \"RColorBreyer\". Spearman correlation between BGN expression and immune infiltration in TCGA database tumours was mined using the \"Immune\" module of the Tumor Immune Estimation Resource 2.0 (TIMER 2.0) tool ( http://timer.cistrome.org/ ) (Li et al., 2020 ). Nineteen immune cells including \"Tregs\", \"Cancer associated fibroblast\", \"DC\", \"Macrophage\" and \"T cell CD8+\" Tregs were selected for the study, and evaluated for immune infiltration using different algorithms such as XCELL, QUANTISEQ, MCPCOUNTER, CIBERSORT, EPIC and TIMER. Subsequently, we further utilised the R software \"ESTIMATE\"(Yoshihara et al., 2013 ) and \"limma\" packages to assess the stromal score, immune score and ESTIMATE score of each patient in each tumour based on gene expression profiles and to calculate the Spearman's correlation between BGN expression and these three scores. 2.6 Gene set enrichment analysis (GSEA) GSEA gene sets (C2, and C5 collections) were downloaded from the Molecular Signatures Database by the R package 'clusterProfiler'. The R language \"limma\", \"org.Hs.eg.db\" and \"enrichplot\" packages were used for functional annotation and pathway enrichment analyses in gene ontology (GO) and the Kyoto encyclopedia of genes and genomes (KEGG). Fisher's test with p -value < 0.05 was considered significant. 2.7 Clinical data and immunohistochemical staining of gastric cancer One hundred and eighty cases of gastric cancer and its paracancerous normal tissue wax blocks were selected from the Second Hospital of Lanzhou University after surgical resection and postoperative pathological diagnosis, including 108 males and 72 females. The basic information and pathological data of patients were collected, including gender, age, tumour size, tumour stage and lymph node metastasis. Tumour staging was based on the eighth edition of the American Cancer Society.21 Postoperative recurrence time (time from surgical resection to disease recurrence or death) was collected from patients through outpatient clinic visit records and telephone follow-up, with a follow-up cut-off date of 31 June 2023.BGN antibody (Abcam, UK, cat no.: ab109369) was used for the immunohistochemistry stain. This study was approved by the Ethics Committee of the Second Hospital of Lanzhou University. 2.8 Statistical analysis All gene expression data were normalised by log2 transformation and statistically analysed using R software (version 4. 4. 3). All clinical correlations were analysed using IBM SPSS Statistics 29 software. Categorical data were analyzed using Pearson's chi-square test. Wilcoxon rank test was used to examine the expression levels of BGN mRNA in different tumour tissues and normal tissues, and the correlation between the expression of BGN and the survival rate of the patients was assessed by one-way survival analysis and Kaplan-Meier analysis, and the comparison between the groups with high and low expression of BGN was performed by Log-rank test. Spearman correlation analysis was used to examine the correlation between BGN expression and the level of immune cell infiltration and immunity. In Spearman's correlation analysis, a strong correlation was considered when the p -value was < 0.05, |R |> 0.8. 3. Results 3.1 BGN Expression Analysis Subcellular localisation analysis showed that BGN detected in Golgi apparatus and Endoplasmic reticulum; and is predicted to be secreted (Fig. 1 A). The TCGA-GTEx dataset showed that BGN was expressed at relatively low levels in most tissues, but at the highest expression in cardiomyocytes (Fig. 1 B). Meanwhile, we also analysed the physiological expression levels of BGN genes in different tissues using the FANTOM5, HPA, and GTEx datasets (Supplementary Fig. 1A-C), which confirmed this. Based on the TCGA data, we also compared BGN expression levels from 33 cancers and mutually matched normal samples (Fig. 1 C). Except for those cancers for which normal tissue data were not available, significant differences in BGN expression were detected in tumour and normal tissues from 14 cancers. Among them, BGN was highly expressed in BRCA, COAD, ESCA, GBM, HNSC, KIRC, LUAD, PRAD, READ, STAD and THCA tumour tissues ( P < 0.001), compared to KICH, LIHC and KIRP, where BGN levels were down-regulated in tumours relative to normal tissues. To assess BGN expression at the protein level, we analysed the IHC results provided by the HPA database and compared the results with the BGN gene expression data from TCGA-GTEx. As shown in Fig. 2 A-G, the results of data analysis from these two databases were consistent. Normal kidney, colon, rectum, brain, skeletal muscle and breast tissues were lowly stained by BGN IHC, and tumour tissues were strongly stained by BGN IHC. 3.2 MR analysis of the causal relationship between BGN and various cancers We investigated the causal association between BGN and a variety of cancers.MR analysis showed that BGN was associated with Malignant neoplasm of bladder (finn-b-C3_BLADDER), Cervical cancer (ebi-a-GCST90018597), Diffuse large B- cell lymphoma (finn-b-C3_DLBCL), Esophageal cancer (ebi-a-GCST90018841), Brain glioblastoma (finn-b-C3_GBM), Acute myeloid leukaemia (finn-b-C3_AML), Lung cancer (ebi-a-GCST004749, ieu-a-987), Ovarian cancer (ieu-b-4963), Pancreatic cancer (ebi-a-GCST90018673), Prostate cancer (ebi-a-GCST90018685, ebi-a-GCST90018905), and Colorectal cancer (ebi-a-GCST012878) were all significantly causally associated. the Cochrane's Q test did not provide evidence of heterogeneity between BGN and the appeal cancers ( P > 0.05). the MR-Egger test intercept did not detect SNP pleiotropy for BGN ( P > 0.05) (Table 1). The leave-one-out analysis also found little change in the overall error line after excluding each SNP (Fig. 3 B-M). Thus we found that BGN was a risk factor for Malignant neoplasm of bladder, Esophageal cancer, Brain glioblastoma, Acute myeloid leukaemia, Lung cancer, Ovarian cancer and Prostate cancer, whereas CCL15 was a protective factor for Cervical cancer, Diffuse large B-cell lymphoma, Pancreatic cancer and Colorectal cancer (Fig. 3 A). 3.3 Prognostic Value of BGN Across Cancers To investigate the relationship between BGN expression levels and prognosis, we performed survival association analysis for each cancer, including OS, DSS, DFI and PFI. in GBMLGG ( p = 1.4e-43, HR = 1.83(1.67,2.00)), LGG ( p = 3.0e-14, HR = 1.61(1.42,1.83)), STES ( p = 0.04, HR = 1.10(1.00,1.21)), KIRP ( p = 3.0e-3, HR = 1.29 (1.09,1.53)), KIPAN ( p = 4.0e-6,HR = 1.19(1.11, 1.29)), COAD ( p = 0.02,HR = 1.20(1.04,1.40)), COADREAD ( p = 0.01,HR = 1.20(1.04,1.38)), STAD ( p = 4.5e-3,HR = 1.19(1.06,1.35)), GBM ( p = 0.02,HR = 1.25( 1.03,1.52)), SKCM-P ( p = 0.02,HR = 1.49(1.05,2.10)), SKCM ( p = 0.02,HR = 1.12(1.02,1.24)), BLCA ( p = 8.2e-4,HR = 1.17(1.07,1.28)), SKCM-M ( p = 0.03,HR = 1.12(1.01,1.24)), MESO (, p = 0.04,HR = 1.18(1.01,1.38)), and ACC ( p = 9.8e-3,HR = 1.49(1.10,2.01)) in high expression with poor OS, and in TARGET-ALL-R ( p = 0.03,HR = 0.92(0.85,0.99) ) in low expression patients with worse OS (Fig. 4 A). In GBMLGG ( p = 4.9e-39,HR = 1.84(1.67,2.02)), LGG ( p = 2.6e-13,HR = 1.63(1.42,1.86)), STES ( p = 0.05,HR = 1.13(1.00,1.27)), KIRP ( p = 3.0e-3,HR = 1.38(1.11 1.71)), KIPAN ( p = 1.2e-4,HR = 1.20(1.09,1.32)), COAD ( p = 6.5e-3,HR = 1.34(1.08,1.65)), COADREAD ( p = 5.9e-3,HR = 1.33(1.09,1.63)), STAD ( p = 0.04,HR = 1.17 (1.01,1.37)), SKCM ( p = 0.02, HR = 1.13 (1.02,1.26)), BLCA ( p = 4.7e-3, HR = 1.17 (1.05,1.31)), SKCM-M ( p = 0.03, HR = 1.13 (1.01,1.26)), MESO ( p = 0.04, HR = 1.23(1.01,1.49)), ACC ( p = 7.7e-3,HR = 1.52(1.12,2.07)), and KICH ( p = 0.02,HR = 2.12(1.14,3.96)) with poorly expressed DSS (Fig. 4 B). In BRCA ( p = 0.02, HR = 1.28 (1.03,1.59)), CESC ( p = 0.01, HR = 1.38 (1.07,1.78)), STES ( p = 0.02, HR = 1.23 (1.03,1.48)), PAAD ( p = 0.05, HR = 1.51 (1.00,2.29)), ACC ( p = 8.9e-3,HR = 1.87(1.16,3.01)) with poor high expression DFI, and poor low expression DFI in LIHC ( p = 0.04,HR = 0.91(0.83,0.99)), and DLBC ( p = 0.03,HR = 0.43(0.15,1.21)) (Fig. 4 C). In GBMLGG ( p = 7.2e-32,HR = 1.58(1.46,1.71)), LGG ( p = 5.0e-10,HR = 1.40(1.26,1.56)), BRCA ( p = 0.05,HR = 1.18(1.00,1.39)), KIRP ( p = 8.6e-3,HR = 1.22(1.05 1.43)), KIPAN ( p = 9.1e-6,HR = 1.18(1.10,1.28)), COAD ( p = 0.02,HR = 1.18(1.03,1.35)), COADREAD ( p = 0.01,HR = 1.17(1.03,1.33)), PRAD ( p = 2.1e-3,HR = 1.39 (1.13,1.71)), BLCA ( p = 0.04,HR = 1.10 (1.01,1.21)), UVM ( p = 0.04,HR = 1.45 (1.01,2.09)), ACC ( p = 0.02,HR = 1.36 (1.05,1.77)), KICH ( p = 0.01,HR = 1.71 (1.11,2.63)) were poorly expressed in high PFI, and poorly expressed in low PFI in DLBC ( p = 9.7e-3,HR = 0.61(0.41,0.91)) (Fig. 4 D). We also analysed the relationship between BGN expression levels and patient survival curve correlation metrics by GEPIA2. KM analysis showed that among individuals with COAD, HNSC, KIRP, LGG and LUSC, patients with low BGN expression levels had longer OS times (Fig. 5 A). In patients with COAD, ESCA, GBM, LGG, KIRP, RADE, and UVM, DFS was worse in those with high BGN expression (Fig. 5 B). Pathological staging can evaluate the severity and extent of the tumour. Based on the combined data of BGN expression and tumour staging information in pan-cancer, we found a significant correlation between BGN expression and pathological staging of various cancers. We found that BGN expression was significantly correlated with tumour stage in nine cancers including BLCA, COAD, ESCA, HNSC, KIRP, READ, STAD, TGCT and THCA (Fig. 6 ). 3.4 Correlation of BGN expression levels with tumour mutation burden and tumour microsatellite instability Tumour development and progression are closely related to genomic mutations. We analysed the types and loci of BGN mutations in different cancers in the TCGA database. The frequency of BGN mutations (> 10%) was highest in lung cancer patients (Fig. 7 A). By integrating mutation data and protein structural domain information, we found that BGN had a high incidence of missense mutations in the LRR_8 structural domain in various cancers, especially in UCEC (Fig. 7 B). Subsequently, we investigated whether there was a correlation between BGN expression levels and TMB and MSI, both of which are importantly linked to sensitivity to immune checkpoint inhibitors. The results showed that BGN expression was strongly correlated with TMB in seven tumours, including STAD, SKCM, PRAD, LUSC, LIHC, LGG, and HNSC, with P < 0.001 (Fig. 7 C). In 10 tumours, including STAD, SKCM, PRAD, PAAD, LUSC, LAML, KIRP, HNSC, KIRC, and COAD, the expression of BGN was correlated with MSI P < 0.05 (Fig. 7 D). 3.5 BGN expression and level of immune cell infiltration in pan-cancer tissues The results of the TIMER database showed that, in most cancers, the expression of BGN was positively correlated with the expression of cancer- associated fibroblasts (CAFs), Endothelial cells and Hematopoietic stem cells in most cancers, while it was negatively correlated with many types of T cells and B cells, suggesting that BGN plays a role in the immune infiltration process of pan-cancer (Fig. 8 ). The tumour immune microenvironment plays a crucial role in tumour development. Therefore, it is important to further explore the pan-cancer relationship between TME and BGN expression. The ESTIMATE algorithm was used to assess the relationship between BGN expression and TME in pan-cancer. Our results showed that the BGN high expression group was more likely to have higher stromal score, immunity score and ESTIMATE score. The nine tumours with the highest correlation coefficients are shown in Fig. 9 (A-C). By analysing the co-expression analysis of BGN with 150 immune pathway marker-associated genes and 60 immune checkpoint genes, we found that there was a significant correlation between the expression of BGN and various immune regulators in pan-cancer (Fig. 9 D-E). 3.6 GSEA enrichment analysis In addition, our enrichment analyses suggest that BGN may be mediated through ECM receptor interaction, ascorbate and aldarate metabolism, WNT signaling pathway, focal adhesion, Notch signaling pathway, JAK/STAT signaling pathway, cell adhesion molecules cams, TGF-beta signaling pathway, chemokine signaling pathway, cytokine-cytokine receptor interaction, cytokine-cytokine receptor interaction(Fig. 10 ). 3.7 Clinicopathological features and immunohistochemical verification of gastric cancer Figure 11 A showed that BGN was highly expressed in gastric cancer tissues and lowly expressed in normal tissues adjacent to the cancer. Further comparison of the relationship between high and low expression of BGN and the clinicopathological characteristics of patients showed that the expression level of BGN was correlated with the TNM stage of the tumour (Table 1). Univariate and multivariate analyses of OS and PFS in patients with gastric cancer showed that high expression of BGN was an independent risk factor affecting the prognosis of patients with pancreatic cancer. high expression was an independent risk factor affecting the prognosis of patients with pancreatic cancer (Tables 2 and 3 ). One hundred and eighty patients were followed up for time to postoperative recurrence, excluding cases with incomplete follow-up data; the results showed a statistically significant difference in PFS between the high and low BGN expression groups (HR = 2.438, 95% Cl (1.513–3.442), Fig. 11 B). 4. Discussion Available studies have shown that the expression of BGN in gastric cancer (GC) tissues is higher than that in adjacent normal tissues, that BGN expression is significantly correlated with histological grading, histological staging, histological staging, T-staging, and Helicobacter pylori (HP) infection in patients with gastric cancer, and that high expression of BGN mRNA is significantly correlated with poor overall survival(Zhao, Yin, Zhao, Liu, & Wang, 2020 ). Another study in the direction of ubiquitination found that the SEMA3B-AS1/HMGB1/FBXW7 axis plays an inhibitory role in peritoneal metastasis of GC by regulating BGN protein ubiquitination(G. Huang et al., 2022 ). It was shown that BGN induced an increase in vascular endothelial growth factor expression by activating the ERK signalling pathway in CRC cells, which was significantly reversed by the ERK inhibitor PD98059(Xing, Gu, Ma, & Ye, 2015 ). Meanwhile, in colon cancer cells, overexpression of BGN promoted resistance to 5-FU chemotherapy by activating the NF-κB pathway(Liu, Xu, Xu, Cui, & Xing, 2018 ). A study of BGN-targeted therapy in colon cancer cells found that BGN induced G0/ G1 cell cycle block, decreased levels of cell cycle proteins A and D1, and increased levels of P21 and P27(Xing, Gu, & Ma, 2015 ). The LINC00460/miR-320a/BGN axis was involved in cell progression, migration, invasion, and EMT in Head and neck squamous cell carcinomas (HNSCC). Silencing of the LINC00460/miR-320a/BGN axis inhibited epithelial–mesenchymal transition (EMT) and was accompanied by a decrease in the expression of N-cadherin and Vimentin(Yang, Wang, Feng, Ma, & Fang, 2021 ). In addition, it was experimentally demonstrated that after BGN knockdown, breast cancer cells exhibited slow metabolism, reduced expression levels of NF-κB transcription factor and P65 and decreased cell metastasis and invasion ability(Manupati et al., 2022 ), BGN is expected to be a target for breast cancer metastasis treatment.TAp73, as a member of the P53 family, promotes pancreatic cancer EMT by affecting the TGF-β pathway and directly regulating BGN expression and Smads expression and activation(Thakur et al., 2016 ). Clinical studies also found that BGN was closely associated with androgen receptor levels, suggesting that BGN is regulated by hormones and that androgen levels are associated with prostate cancer, so it is hypothesised that up-regulation of BGN is a common feature of prostate cancer that parallels tumour progression(Jacobsen et al., 2017 ), suggesting that BGN can be used as a screening indicator for prostate cancer. Pan-cancer analysis is the analysis of genes in multiple cancers, comparing the differences and similarities in the expression of extracted genes from a genetic point of view to find their association with cancer. Understanding genomic changes in multiple cancers leads to further discovery of the causes of cancer(Srivastava & Hanash, 2020 ). Our study found that BGN is highly expressed in myocardial tissue under physiological conditions, which may be associated with the development of hypertrophic cardiomyopathy(Dong, Yin, Xiao, & Tang, 2024 ). Interestingly, BGN is highly expressed in most cancer cells but downregulated in KICH, LIHC and KIRP tumours, however there is no relevant literature report. Although BGN-mediated activation of TLR-2/4 is highly susceptible to triggering an inflammatory response in response to renal injury, particularly the recruitment of inflammatory cells (neutrophils, macrophages and T cells) and the release of cytokines and chemokines (TNF-α, CXCL1, CCL2 and CCL5)(Moreth et al., 2014 ). Meanwhile, BGN accumulates in the neointima of atherosclerotic vessels during renal fibrosis with aggregated deposits of type I collagen and Decorin(Stokes et al., 2000 ). However, its effect on renal malignancy still needs to be verified by more scientific experiments. Similarly, a significant increase in hepatic BGN expression positively regulates HSP47 to modulate ECM deposition and hepatic stellate cells activation to promote hepatic fibrosis(Yu et al., 2023 ). These contradictions may stem from the analytical errors of RNA-seq data, differences in sample sources, and limitations of experimental methods. In order to investigate the causal relationship between BGN and cancer, it was excluded that it might be influenced by factors such as the environment. We performed a Mendelian randomisation (MR) study between BGN and tumours.MR uses genetic variants as instrumental variables (IVs) to measure the potential causal relationship between exposure and outcome. Single nucleotide polymorphisms (SNPs) were obtained from genome-wide association studies, and SNPs with p-values less than 5e-8 were not detected for some BGNs, so a significance threshold of 1e-5 was set, excluding other external environmental and confounding factors. We finally obtained the association between BGN and the risk of 11 malignancies. One-way COX regression analysis revealed that OS, DSS, PFI, and DFI were closely associated with high BGN expression only in ACC patients. The OS, DSS, and PFI of GBMLGG, LGG, KIPR, KIPAN, COAD, and COADREAD patients with high expression of BGN had a poorer prognosis; the OS, DSS, and DFI of STES patients, and the OS, DSS, and DFI of STAD and SKCM patients had poorer OS, DSS prognosis.KM results also showed longer OS and DFS in COAD, KIRP and LGG patients with low BGN expression. We also found that the expression of BGN was closely related to the staging of various malignant tumours. These results suggest that BGN may serve as a potential prognostic marker for a variety of tumours. TMB is a promising pan-cancer cancer predictive biomarker that can guide immunotherapy in the era of precision medicine(Chan et al., 2019 ). Our study showed that BGN expression correlated with TMB in 7 cancer types and with MSI in 10 cancer types. This may indicate that the expression level of BGN affects the TMB and MSI of the tumour, thereby influencing the patient's response to immune checkpoint inhibition therapy. This would provide a new reference for the prognosis of immunotherapy. According to existing studies and our findings, tumours with high BGN expression, high TMB and high MSI may have a better prognosis after ICI treatment in cancers where BGN expression is positively correlated with TMB. Cancer progression is not only about changes in the cancer cells themselves, but changes in the tumour microenvironment have also been shown to play a key role in tumour development and progression(Hanahan & Weinberg, 2011 ). the TME profile can be used as a marker to assess the response of tumour cells to immunotherapy and to influence clinical outcomes, and tumour-infiltrating immune cells have a significant impact on tumour development, either antagonising or promoting tumour progression(Jin & Jin, 2020 ). According to the ESTIMATE score, BGN expression was positively correlated with stromal and immune cell content in the TME of most cancers. Our study further elucidates that BGN has broader tumour applicability and confirms that BGN expression is intimately involved in the biological processes of immune cells and immune-related molecules in most cancers. In addition, our study revealed that BGN was co-expressed with genes encoding MHC, immune activation, immune suppression, chemokine and chemokine receptor proteins. All these results suggest that BGN expression is closely related to the immune infiltration of tumour cells, which affects patient prognosis and provides a new target for the development of immunosuppressive agents. In addition, our enrichment analysis indicated that BGN may be expressed through ECM receptor interaction, JAK/STAT signaling pathway, TGF-beta signaling pathway, chemokine signaling pathway, cytokine- cytokine receptor interaction. The ECM is a dynamic structure composed of glycoproteins (mucins), proteoglycans, collagen, laminin, fibronectin, elastin fibres and reticulin. Among these components, proteoglycans, glycoproteins and collagen are considered to be the \"core matrix\" of the ECM(Zhang et al., 2022 ). Despite the abundance of BGN in the ECM, the mechanisms by which it affects the tumour microenvironment in other ways remain poorly understood, but it has been shown to be multifunctional in cancer progression, drug resistance, and the regulation of other features of cancer. Although the present study provides an integrated and comprehensive biochemical analysis of the role of BGN in pan-cancer, there are many shortcomings. This study used a large number of databases to validate the risk of BGN in a wide range of cancers, but it is only representative of some cancer patients. For this reason, we collected clinical characteristics of 180 gastric cancer patients to verify the correlation between BGN expression and the prognosis of gastric cancer, but experimental studies are still needed to verify its oncogenic function. Translated with DeepL.com (free version) 5. Conclusion In conclusion our results suggest that BGN can act as an independent prognostic factor in a wide range of tumours, and for different tumours, the level of its expression leads to different prognostic outcomes, which requires further investigation of the specific role of BGN in each tumour. In addition, BGN expression is associated with TMB, MSI and immune cell infiltration in various cancer types. Its effect on tumour immunity also varies by tumour type. These findings may help to elucidate the role of BGN in tumourigenesis and development, and may provide a reference for achieving more precise and personalised immunotherapy in the future. Declarations Conflict of interest The authors declare that there are no conficts of interest regarding the publication of this paper. Author Contribution All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by GuangNing Min and Hao Tang. The first draft of the manuscript was written by Guangtao Min and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Data availability Publicly available datasets were analysed in this study.data can be found here: