Abstract

Background

   Renal Cell Carcinoma (RCC) is a leading cause of cancer-related
   mortality worldwide, with Clear Cell Renal Cell Carcinoma (ccRCC)
   comprising 75% of cases. Surgical resection remains the cornerstone of
   treatment for localized RCC, but its asymptomatic progression and lack
   of reliable early biomarkers often result in advanced disease at
   diagnosis. Collagen VI alpha-2 chain (COL6A2), an extracellular matrix
   protein, has been implicated in tumor progression and metastasis.
   Despite its established roles in other malignancies, the specific
   contribution of COL6A2 to ccRCC pathogenesis is poorly understood.

Objective

   This study aims to systematically investigate COL6A2 expression in
   ccRCC, its prognostic value, and its potential impact on the tumor
   immune microenvironment, cancer stem cell characteristics, and drug
   response.

Methods

   The mRNA and protein expression data for ccRCC were sourced from TCGA,
   GEO, CPTAC, and ICPC. Single-cell and spatial transcriptomic data were
   processed using Seurat with quality control measures. Clinical
   correlations and survival analyses, including immune infiltration and
   COL6A2 expression, were assessed using Cox regression and Kaplan-Meier
   curves. Cancer stemness was evaluated using six stemness indices.
   Differential expression and pathway analyses (GO, KEGG, GSEA) were
   performed with DESeq2 and clusterProfiler. Drug sensitivity and
   immunotherapy response were predicted using GDSC, CTRP, and TIDE
   databases. Functional studies, including colony formation and invasion
   assays, as well as in vivo xenograft models, assessed the impact of
   COL6A2 on tumor progression and therapy response.

Results

   COL6A2 expression was significantly upregulated in ccRCC compared to
   normal tissues. High COL6A2 expression correlated with poorer overall
   survival (OS), progression-free interval (PFI), and progression-free
   survival (PFS), establishing it as an independent prognostic factor for
   ccRCC. Additionally, COL6A2 expression was positively associated with
   immune-suppressive cell infiltration, suggesting an immunosuppressive
   tumor microenvironment. COL6A2 was also linked to enhanced stem
   cell-like properties, invasiveness, and metastatic potential. Pathway
   enrichment analyses revealed that COL6A2 may influence tumor
   progression by regulating the epithelial-mesenchymal transition (EMT)
   process and activating the PI3K-Akt signaling pathway. Notably, high
   COL6A2 expression correlated with enhanced responsiveness to sunitinib
   but resistance to immunotherapy, highlighting its dual role in therapy
   selection.

Conclusion

   This study identifies COL6A2 as a powerful prognostic biomarker and a
   driver of ccRCC progression through EMT and immune suppression.
   Targeting COL6A2 holds promise for improving immunotherapy efficacy and
   personalizing treatment strategies, offering new hope for ccRCC
   patients facing limited options.

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12967-025-06793-9.

   Keywords: Renal cancer, COL6A2, Tumor microenvironment, Pharmacological
   prediction, Prognostic significance

Introduction

   Renal Cell Carcinoma (RCC), is the predominant form of malignant kidney
   tumors, with Clear Cell Renal Cell Carcinoma (ccRCC) accounting for 75%
   of cases [[30]1]. Globally, approximately 315,000 new cases of RCC are
   diagnosed annually, ranking it among the top ten most common cancers
   worldwide [[31]2]. Due to RCC’s low responsiveness to traditional
   radiotherapy and chemotherapy, surgical resection remains the preferred
   treatment for localized RCC. However, because of RCC’s asymptomatic
   growth and the lack of effective early diagnostic markers,
   approximately 30% of patients present with distant metastasis at
   diagnosis [[32]3]. Metastatic RCC treatment primarily relies on
   targeted therapies, such as Sorafenib and Sunitinib, but their high
   cost, resistance, toxicity, and side effects limit their widespread
   clinical application [[33]4]. Therefore, it is imperative to delve into
   the molecular mechanisms of RCC and identify new biomarkers and
   molecular targets.

   Collagen VI (COL6), a member of the collagen family, is ubiquitously
   expressed in various tissues [[34]5]. COL6-dependent assembly of
   vascular basement membranes is crucial for vascular function and tumor
   progression [[35]6]. The COL6A2 gene encodes the α2 chain of Collagen
   VI, which, along with the α1 (VI) and α3 (VI) chains encoded by COL6A1
   and COL6A3, forms the functional collagen VI structure [[36]7]. Studies
   have shown that COL6A2 is strongly associated with tumor proliferation,
   migration, and invasion. For instance, COL6A2 significantly inhibits
   the proliferation of human bladder cancer cells, induces G1-phase cell
   cycle arrest, and suppresses wound healing and invasion by inhibiting
   matrix metalloproteinases (MMP-2 and MMP-9) activity [[37]8].
   Additionally, COL6A2 is implicated as a key gene in the recurrence of
   Low-Grade Glioma (LGG), with the LncRNA HOXA-AS2/miR-184/COL6A2 axis
   serving as a potential regulatory mechanism for LGG recurrence [[38]9].
   Although previous studies have shown that COL6A2 expression is
   significantly elevated in ccRCC [[39]10], its functional role and
   clinical significance in ccRCC remain unclear. This study aims to
   systematically explore the expression pattern of COL6A2 in ccRCC, its
   prognostic value, and its potential impact on the tumor immune
   microenvironment, cancer stem cell characteristics, and drug
   responsiveness.

Materials and methods

Data sources and gene expression analysis

Bulk RNA-seq databases

   mRNA expression profiles and clinical data for various tumor types
   including bladder urothelial carcinoma (BLCA), breast invasive
   carcinoma (BRCA), cervical squamous cell carcinoma and endocervical
   adenocarcinoma (CESC), cholangiocarcinoma (CHOL), colon adenocarcinoma
   (COAD), esophageal carcinoma (ESCA), glioblastoma multiforme (GBM),
   head and neck cancer (HNSC), kidney chromophobe (KICH), kidney renal
   clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma
   (KIRP), liver hepatocellular carcinoma (LIHC), lung adenocarcinoma
   (LUAD), lung squamous cell carcinoma (LUSC), pancreatic adenocarcinoma
   (PAAD), pheochromocytoma and paraganglioma (PCPG), prostate
   adenocarcinoma (PRAD), rectum adenocarcinoma (READ), sarcoma (SARC),
   skin cutaneous melanoma (SKCM), stomach adenocarcinoma (STAD), thyroid
   carcinoma (THCA), thymoma (THYM), uterine corpus endometrial carcinoma
   (UCEC) were obtained from The Cancer Genome Atlas (TCGA) database.
   Public mRNA expression datasets for ccRCC were obtained from the Gene
   Expression Omnibus (GEO) database, including [40]GSE66272 (27 tumor and
   27 adjacent normal samples), [41]GSE53757 (76 normal samples, 68 tumor
   samples), [42]GSE36895 (23 normal samples, 29 tumor samples),
   [43]GSE46699 (63 normal samples, 69 tumor samples), [44]GSE167093 (155
   normal samples, 348 tumor samples), and [45]GSE126964 (11 normal
   samples, 55 tumor samples).Protein expression profiles for ccRCC and
   normal tissues were obtained from the Clinical Proteomic Tumor Analysis
   Consortium (CPTAC) and the International Cancer Proteogenome
   Consortium(ICPC) databases. Immunohistochemistry (IHC) staining images
   for COL6A2 were retrieved from the Human Protein Atlas (HPA) database.

   Single Cell and Spatial Transcriptomics RNA-seq Databases: Single cell
   and spatial transcriptomics RNA-seq datasets were obtained from the GEO
   database ([46]GSE210038, [47]GSE210041), including 7 ccRCC samples and
   2 adjacent normal samples. Quality control and downstream analysis were
   performed using the standard pipeline of the R package Seurat (version
   5.0.1) [[48]11], filtering out low-quality cells with fewer than 200
   expressed genes, more than 30% mitochondrial genes, and genes expressed
   in fewer than three cells. After excluding housekeeping and
   mitochondrial genes, a total of 22,935 genes and 61,343 cells were
   retained. Cluster annotation was conducted following standard data
   normalization, dimensionality reduction, and clustering, utilizing
   previously reported cellular markers. Pseudotime analysis was conducted
   using the R package Monocle [[49]12].

Human samples

   In accordance with ethical guidelines, informed consent was obtained
   from patients, and approval was granted by the institutional research
   ethics committee of Zhongnan Hospital, Wuhan University. Fresh renal
   cancer tissues and corresponding adjacent normal tissues were obtained
   from Zhongnan Hospital of Wuhan University. These tissues were stored
   at -80 °C.

Clinical correlation analysis and prognostic analysis

   The most recent survival data for the all cohorts of TCGA were
   retrieved from the study [[50]13].The relationship between COL6A2
   expression levels and clinical parameters such as TNM, Stage, and age
   in TCGA-KIRC cohort was assessed. Tumor samples from the TCGA-KIRC
   cohort were stratified into high-expression and low-expression groups
   using the median COL6A2 expression level as the cutoff. The association
   between COL6A2 expression and overall survival (OS), progression-free
   interval (PFI), progression-free survival (PFS) in ccRCC patients was
   analyzed using univariate Cox regression. The association between
   COL6A2 expression groups and OS, PFI, PFS was evaluated with
   Kaplan-Meier (KM) survival analysis. Multivariate Cox regression was
   employed to determine the independent impact of COL6A2 expression and
   clinical variables on OS, PFI, and PFS. To enhance prediction
   reliability of ccRCC patient outcomes, a prognostic nomogram
   integrating COL6A2 expression and clinical features was constructed
   based on the multivariate Cox regression results. Calibration curves
   were plotted to compare nomogram-predicted survival probabilities
   against observed outcomes. Predictive accuracy was assessed using
   receiver operating characteristic (ROC) curves. Statistical analyses
   were conducted with the R package “survival” and visualized using the
   “survminer” package. Nomogram construction was performed using the
   “rms” package.

Analysis of the correlation between COL6A2 expression and immune cell
infiltration

   We used the ESTIMATE algorithm to calculate the correlation between
   immune scores and COL6A2 expression levels, elucidating the
   relationship between COL6A2 and the immune contexture of ccRCC. We
   utilized several algorithms, including TIMER, Quantiseq, CIBERSORT,
   EPIC, and MCPcounter, to investigate the correlation between immune
   subset proportions and COL6A2 expression levels, assessing the impact
   of COL6A2 on immune cell infiltration in the ccRCC immune
   microenvironment.

Analysis of the correlation between COL6A2 and cancer stemness index

   In this study, six cancer stemness indices, derived from previous
   research, were calculated based on mRNA expression and DNA methylation
   signatures [[51]14]. These include the RNA Stemness Score (RNAss),
   based on RNA expression, the Epiregulin Expression Stemness Score
   (Epiregulin Expression Stemness Score), based on epigenetically
   regulated RNA expression involving 103 genes, the DNA Stemness Score
   (DNAss), based on DNA methylation, combining 219 stem cell signature
   probes, the Epigenetically Regulated DNA Methylation Stemness Score
   (EREG-METHss), based on 87 probes, the Differentially Methylated
   Positions Stemness Score (Differentially Methylated Positions Stemness
   Score), based on 62 differentially methylated probes, and the Enhancer
   Elements/DNA Methylation-based Stemness Score (ENHss), based on 82
   enhancer elements/DNA methylation probes. Additionally, this study
   integrates the stemness indices with COL6A2 gene expression data to
   explore the correlation between COL6A2 expression and tumor stemness.

Differential gene and pathway enrichment analysis

   To investigate the functions and pathways associated with COL6A2, we
   used the R package DESeq2 to identify differentially expressed genes
   (DEGs) between high and low COL6A2 expression groups. The selection
   criteria were based on an adjusted P-value < 0.05 and an absolute log
   fold change (|logFC|) > 1. Next, we used the R package
   ‘clusterProfiler’ to perform Gene Ontology (GO) analysis, Kyoto
   Encyclopedia of Genes and Genomes (KEGG) analysis, and Gene Set
   Enrichment Analysis (GSEA) on the identified DEGs [[52]15]. GSEA was
   performed using gene sets from the MSigDB database, including
   “c5.go.v2023.2”, “c2.cp.kegg_legacy.v2023.2”, and HALLMARK gene sets,
   to evaluate the enrichment of distinct pathways and molecular
   mechanisms between high and low COL6A2 expression groups. For GSEA, we
   set parameters with a minimum gene set size of 5 and a maximum of 5000,
   performing 1000 resampling iterations. Additionally, gene sets related
   to the Epithelial-Mesenchymal Transition (EMT) pathway were compiled
   from CancerSEA, and GSVA enrichment scoring was performed using the R
   package ‘GSVA’.

Drug sensitivity prediction

   Data from the GDSC (Genomics of Drug Sensitivity in Cancer) and CTRP
   (Cancer Therapeutics Response Portal) databases were used to construct
   our training set. We then employed the R package ‘oncoPredict’ to
   predict drug sensitivity scores for each patient with ccRCC, assessing
   their response to a variety of pharmaceuticals [[53]16]. Briefly, ridge
   regression models were trained using large-scale gene expression and
   drug sensitivity data (training dataset). These models were
   subsequently applied to independent gene expression profiles (testing
   dataset) to generate drug sensitivity predictions. Predicted drug
   sensitivity scores demonstrated close alignment with experimental IC50
   values, where lower scores corresponded to lower IC50 levels,
   indicating potentially enhanced patient therapeutic response.
   Additionally, the TIDE (Tumor-Immune Dysfunction and Exclusion) online
   tool was used to predict the response of patients with ccRCC to
   immunotherapy, based on TCGA-KIRC transcriptome data. The TIDE tool
   provides an in-depth analysis of immune cell infiltration and
   functional status within the tumor microenvironment, aiding in the
   evaluation of immune cell composition and their potential impact on
   tumor growth and therapeutic response [[54]17].

Cell culture and transfection

   Human RCC cell lines and RENCA cell were obtained from Land
   Biotechnology (Guangzhou, China). Cells were all cultured in RPMI 1640
   medium (Gibco, China), supplemented with 10% fetal bovine serum (FBS,
   Gibco, USA), and maintained in a humidified incubator at 37 °C with 5%
   CO^2. ShRNA plasmids were synthesized by Beijing Tsingke Biotech
   Co.,Ltd. Transfection of ACHN and OS-RC-2 cells was conducted using
   Lipofectamine 3000 (Invitrogen, USA). 48 h post-transfection, cells
   were collected for further experimental analyses.

RT-qPCR & Western blot

   RT-qPCR and Western blot were performed as previously described
   [[55]18]. Primer sequences for amplification were as follows: COL6A2
   forward primer, 5’-TCAAGGAGGCTGTCAAGA-3’ and reverse primer,
   5’-GGCGATGGAGTAGAGGTT-3’; GAPDH forward primer,
   5’-GAAGGTGAAGGTCGGAGTC-3’ and reverse primer,
   5’-GAAGATGGTGATGGGATTTC-3’. Primary antibodies, anti-COL6A2 (1:1000),
   anti-E-cadherin (1:1000), anti-N-cadherin (1:1000), anti-β-catenin
   (1:1000), anti-Vimentin (1:1000), anti-Snai1 (1:1000), anti-GAPDH(1:
   10000, proteintech), were used for probing, with GAPDH serving as a
   loading control.

Colony formation assay, transwell migration assay, scratch assay, cell
counting kit-8 and IC50 assay

   Colony formation, tumor invasion, and scratch assays were performed as
   previously described [[56]18]. Cell proliferation of ccRCC cell lines
   was assessed using the CCK-8 assay (MedChem Express). 2000 cells per
   well were seeded in 96-well plates with 100 µL of RPMI 1640 medium
   supplemented with 10% FBS for 96 h. After adding 10 µL of CCK-8
   solution, the cells were incubated for 2 h. Absorbance at 450 nm was
   measured daily using a DNM-9606 plate reader (Perlong, Beijing). These
   cells were exposed to varying concentrations of sunitinib (0, 0.01,
   0.05, 0.10, 0.50, 1.00, 2.50, 5.00, 7.50, 10.00 µmmol/L) for 24 h,
   followed by the addition of 10 µL of CCK-8 reagent and incubation at
   37 °C for 1 h. Each experiment was performed in triplicate. The IC[50]
   (half maximal inhibitory concentration) value for sunitinib was
   determined using best-fit dose-response inhibition curves in GraphPad
   Prism 10.0 software.

In vivo tumor xenograft model

   In accordance with standard statistical methods, we randomly selected 6
   mice from each group, following the approach commonly used in most
   studies. BALB/c-nu and BALB/c mice (4–5 weeks old, female) were
   purchased from Sibeifu Company (Beijing, China). The mice were randomly
   assigned to different experimental groups. The cells were mixed with 50
   µL of Matrigel (Corning) and 50 µL of PBS, then injected subcutaneously
   into the backs of the mice. A total of 6 × 10^6 OS-RC-2 cells were
   transfected with shControl and sh-COL6A2. Starting on the 5th day after
   subcutaneous tumor implantation, the mice were treated with or without
   sunitinib (25 mg/kg, administered orally once daily for 14 days). Tumor
   size was monitored and assessed every 2 days. Anti-mouse PD-1
   monoclonal antibody (clone RMP1-14), purchased from Bio X Cell, was
   used for anti-PD-1 treatment, while saline was used as the isotype
   control. Additionally, 1 × 10^7 Renca cells were separately infected
   with shControl and sh-COL6A2. From the 5th day after subcutaneous tumor
   implantation, the mice were treated with or without anti-mouse PD-1
   antibody (10 mg/kg, administered via intraperitoneal injection twice
   weekly). Tumor size was monitored and assessed every 3 days. Tumor
   volume was calculated using the formula: (L × W²)/2. Mice were
   euthanized, and tumors were excised and weighed. The experiment was
   conducted in accordance with the guidelines established by the Medical
   Ethics Committee of Zhongnan Hospital, Wuhan University (approval
   MRI2024-LAC209).

Statistical analysis

   All statistical analyses in this study were performed using R software
   (version 4.1.3) and GraphPad Prism 10 software (La Jolla, CA, USA).
   Intergroup differences were assessed for significance using Student’s
   t-test and the Wilcoxon rank-sum test, while correlation analyses were
   conducted using the Spearman correlation coefficient. Statistical
   significance was defined as * P < 0.05, ** P < 0.01, *** P < 0.001, and
   **** P < 0.0001.

Results

Pan-cancer landscape and prognostic significance of COL6A2 expression

   We first utilized TCGA database to analyze the expression patterns of
   the COL6A2 gene across various cancer types. The results showed that
   COL6A2 expression was significantly elevated in BRCA, CHOL, GBM, HNSC,
   KIRC, and significantly decreased in BLCA, CESC, COAD, KICH, PRAD, UCEC
   compared to normal tissues (P < 0.05) (Fig. [57]1A). For cancer types
   with significant expression differences, survival data, including OS,
   PFI, and PFS, were analyzed using univariate Cox regression models to
   assess the relationship between COL6A2 expression and prognosis. The
   analysis showed that COL6A2 acted as an independent risk factor for
   poor prognosis in most cancers with significant expression differences
   (Fig. [58]1B). Further analysis revealed that COL6A2 had statistical
   significance in BLCA, KICH, GBM, and KIRC (Fig. [59]1C). However, in
   BLCA and KICH, COL6A2 expression was inversely correlated with
   survival, requiring further investigation into the underlying
   mechanisms. Given the established role of COL6A2 in GBM, this study
   focused on its role in KIRC. COL6A2 expression in ccRCC was analyzed
   using TCGA, CPTAC, and ICPC databases, revealing significantly higher
   expression at both the transcriptional and protein levels compared to
   normal tissues (P < 0.05) (Fig. [60]1D, E). To validate this, COL6A2
   expression was analyzed using the [61]GSE66272, [62]GSE53757,
   [63]GSE36895, [64]GSE46699, [65]GSE167093, and [66]GSE126964 datasets
   from the GEO database, with results confirming higher expression in
   ccRCC compared to normal tissues (P < 0.05) (Fig. [67]1H).
   Additionally, IHC staining images from HPA confirmed the high
   expression of COL6A2 in renal cancer tissues (Fig. [68]1F). Finally,
   COL6A2 expression in clinical samples was validated at both the RNA and
   protein levels, confirming higher expression in cancer tissues compared
   to adjacent normal tissues (P < 0.05) (Fig. [69]1G, I).

Fig. 1.

   [70]Fig. 1
   [71]Open in a new tab

   COL6A2 Expression and the Correlation with Prognostic Indicators in
   Pan-Cancer, Particularly ccRCC. A. RNA expression of COL6A2 across
   various cancers and normal tissues from TCGA. B. Univariate COX
   regression forest plot for OS, PFI, and PFS of COL6A2. The horizontal
   line segments represent the 95% confidence intervals, with the two ends
   representing the lower limit (lower 95% HR) and upper limit (higher 95%
   HR) of the 95% confidence interval; the dots represent the HR values
   for each variable, where blue dots indicate HR < 1 and red dots
   indicate HR > 1; there is a vertical dashed line at HR = 1. If the line
   segment of a variable does not intersect with the dashed line, it means
   that the 95% confidence interval of the variable’s HR does not include
   1, the p-value is less than 0.05, and the variable has an impact on
   patient survival with statistical significance. C. Venn diagram
   displaying tumors with differential gene expression and significant
   univariate COX regression for OS, PFI, and PFS. D. Differential RNA
   expression of COL6A2 in paired tumor and normal samples from TCGA-KIRC.
   E. Protein expression of COL6A2 in ccRCC tumors and normal samples from
   the CPTAC. F. Immunohistochemical staining of COL6A2 in renal cancer
   and normal samples from HPA. G. RNA expression of COL6A2 in tumor and
   adjacent clinical samples. H. RNA expression of COL6A2 in cancer and
   adjacent samples from GEO datasets. I. Protein expression of COL6A2 in
   collected tumor and adjacent clinical samples

COL6A2 as independent prognostic factor in ccRCC

   Further analysis of the TCGA-KIRC dataset revealed correlations between
   COL6A2 expression levels and clinical characteristics of ccRCC
   patients. The analysis showed that patients over 65 had significantly
   lower COL6A2 expression than younger patients (P < 0.05) (Fig. [72]2A).
   Gender had no significant effect on COL6A2 expression (P > 0.05)
   (Fig. [73]2A). In TNM, patients in T3-4 had significantly higher COL6A2
   expression than those in T1-2 (P < 0.05). Similarly, patients in stage
   N1 had higher expression than those in stage N0 (P < 0.05), with no
   significant difference observed between M stages (P > 0.05)
   (Fig. [74]2A). In stage, patients in Stage3-4 had significantly higher
   COL6A2 expression than those in Stage1-2 (P < 0.05) (Fig. [75]2A).
   Kaplan-Meier survival curve analysis showed that patients with high
   COL6A2 expression had significantly shorter OS, PFI, and PFS
   (Fig. [76]2B). Similar trends were observed across subgroups defined by
   gender, age, TNM, and Stage classifications (P < 0.05) (Fig. [77]2C,
   SupFigure [78]1 A, F). Multivariate Cox regression analysis confirmed
   COL6A2 as an independent prognostic factor for OS in ccRCC patients
   (P < 0.05) (Fig. [79]2D). A nomogram clinical prediction model,
   incorporating age, N stage, and M stage, showed high predictive
   accuracy for 1-year, 3-year, and 5-year survival, as demonstrated by
   ROC curves and calibration plots (Fig. [80]2E-G). Analysis of PFI and
   PFS confirmed the trends observed for OS, with high COL6A2 expression
   correlating with poorer prognoses. Multivariate Cox regression analysis
   also confirmed COL6A2 as an independent prognostic factor for these
   survival indicators (P < 0.05) (SupFigure [81]1B, G). Similarly, a
   nomogram clinical prediction model, based on PFI and PFS multivariate
   Cox regression analysis results, demonstrated high predictive accuracy
   for 1-year, 3-year, and 5-year survival, as shown by ROC curves and
   calibration plots (SupFigure [82]1 C-E, H-J).

Fig. 2.

   [83]Fig. 2
   [84]Open in a new tab

   Association of COL6A2 with clinical characteristics and prognostic
   indicators in KIRC. A. Expression of COL6A2 across subgroups defined by
   age, gender, TNM, and stage in KIRC. Clinical characteristics (left to
   right): TNM stage, Clinical stage, Age, Gender. B. Kaplan-Meier
   survival curves for OS, PFI, and PFS in high vs. low COL6A2 expression
   groups. C. Kaplan-Meier survival curves for OS in high vs. low COL6A2
   expression groups across clinical subgroups. D. Multivariate COX
   regression forest plot for OS including COL6A2 and clinical indicators.
   E. Nomogram based on significant indicators from multivariate COX
   regression for predicting OS. F. Calibration curves at 1, 3, and 5
   years for the OS nomogram. G. Receiver operating characteristic (ROC)
   curves at 1, 3, and 5 years for the OS nomogram

Immune microenvironment remodeling by COL6A2

   We used the ESTIMATE algorithm to calculate the correlation between
   immune scores and COL6A2 expression, elucidating the relationship
   between COL6A2 and the immune microenvironment in ccRCC. The results
   showed a negative correlation between COL6A2 expression and tumor
   purity (P < 0.05), and a positive correlation with StromalScore,
   ImmuneScore, and ESTIMATEScore (P < 0.05) (Fig. [85]3A). Furthermore,
   we used various bioinformatics tools, including TIMER, Quantiseq,
   CIBERSORT, EPIC, and MCPcounter, to analyze immune cell infiltration
   within the tumor microenvironment. TIMER analysis showed no significant
   correlation between COL6A2 expression and the proportion of B cells
   (P > 0.05), while a positive correlation was observed with CD8 + T
   cells, CD4 + T cells, neutrophils, macrophages, and dendritic cells,
   which were more abundant in samples with high COL6A2 expression
   (P < 0.05) (P < 0.05) (Fig. [86]3B). EPIC analysis showed that samples
   with high COL6A2 expression had significantly higher proportions of B
   cells, cancer-associated fibroblasts, CD4 + T cells, and endothelial
   cells compared to low expression samples (P < 0.05), while CD8 + T
   cells, macrophages, and NK cells showed no significant differences
   (P > 0.05) (Fig. [87]3C). Quantiseq analysis showed that samples with
   high COL6A2 expression had higher proportions of both M1 and M2
   macrophages, with M2 macrophages being more abundant than M1, as well
   as higher proportions of NK cells, Tregs, and dendritic cells
   (P < 0.05), while proportions of B cells, monocytes, neutrophils,
   CD4 + T cells, and CD8 + T cells showed no significant differences
   (P > 0.05) (Fig. [88]3D). MCPcounter analysis showed that samples with
   high COL6A2 expression had significantly higher proportions of B cells,
   NK cells, monocytes, myeloid dendritic cells, neutrophils, endothelial
   cells, and fibroblasts (P < 0.05)(Fig. [89]3E). Although T cells and
   cytotoxic lymphocytes were more abundant in high expression samples
   (P < 0.05), the proportion of CD8 + T cells showed no significant
   difference (P > 0.05) (Fig. [90]3E). CIBERSORT analysis showed that Tfh
   cells, activated NK cells, and monocytes were more abundant in samples
   with low COL6A2 expression (P < 0.05), while Tregs were more abundant
   in samples with high COL6A2 expression (P < 0.05). No significant
   differences were observed in other cell types between high and low
   COL6A2 expression groups, including immature B cells, memory B cells,
   plasma cells, and CD8 + T cells (Fig. [91]3F). The comprehensive
   analysis suggests that samples with high COL6A2 expression have higher
   proportions of immune-suppressive cells, such as macrophages, Tregs,
   and dendritic cells, while proportions of CD8 + T cells and B cells
   show no significant differences. Additionally, the proportions of
   stromal cells, such as cancer-associated fibroblasts and endothelial
   cells, are significantly higher in samples with high COL6A2 expression
   compared to those with low expression, consistent with ESTIMATE
   analysis results, indicating a strong positive correlation between
   StromalScore and COL6A2 expression. These results suggest that,
   although samples with high COL6A2 expression have a higher immune
   score, the enrichment of tumor-associated fibroblasts and
   immune-suppressive cells may create an immunosuppressive
   microenvironment, thereby affecting tumor progression.

Fig. 3.

   [92]Fig. 3
   [93]Open in a new tab

   Exploration of the Relationship Between COL6A2 and Immune Infiltration
   in ccRCC Using Various Immune Infiltration Analyses. A. ESTIMATE. B.
   TIMER. C. EPIC. D. Quantiseq. E. MCPcounter. F. CIBERSORT. TCGA-KIRC
   tumor samples were dichotomized into high-expression (red) and
   low-expression (blue) groups based on the median COL6A2 expression
   level, as stratified in the box plot

COL6A2 promotes tumor stemness and EMT progression

   This study further analyzed DMPss, DNAss, ENHss, EREG − METHss,
   EREG.EXPss and RNAss to explore the association between COL6A2
   expression levels and tumor stemness. The findings revealed that in
   tumor samples with high COL6A2 expression, DMPss, DNAss, ENHss,
   EREG − METHss and EREG.EXPss scores were significantly higher than in
   samples with low expression (P < 0.05), while RNAss scores were
   relatively lower (P < 0.05) (Fig. [94]4A). These results suggest that
   tumor cells with high COL6A2 expression may possess stronger stem cell
   characteristics, invasiveness, and metastatic potential. Incorporating
   these stemness scores, we hypothesize that the expression level of
   COL6A2 may influence the sensitivity of tumor cells to treatment,
   thereby providing potential biomarkers for the development of
   personalized therapeutic strategies.

Fig. 4.

   [95]Fig. 4
   [96]Open in a new tab

   Relationship between COL6A2 and tumorigenesis in ccRCC: exploration of
   its function and molecular pathways. A. Distribution of tumorigenic
   scores in high vs. low COL6A2 expression groups. B. Volcano plot of
   differentially expressed genes between high and low COL6A2 expression
   groups. C. GSEA-KEGG enrichment bar plot for differentially expressed
   genes. D. GO Biological Process (BP) and KEGG enrichment bar plots for
   genes with high expression. E. GO-BP and KEGG enrichment bar plots for
   genes with low expression. F. Top 5 GSEA-GOBP enrichment plots for
   highly expressed genes. G. Uniform Manifold Approximation and
   Projection (UMAP) plot showing single-cell subtypes. H. UMAP plot
   showing COL6A2 expression across cell types. I. Volcano plot of
   differential analysis between tumor epithelial cells and normal renal
   tubular epithelial cells at the single-cell level. J. UMAP plot showing
   tumor cell subtypes. K. UMAP plot showing COL6A2 expression across
   tumor cell subtypes. L. UMAP plot showing EMT scores across tumor cell
   subtypes. M-N. Pseudotime trajectory plot for tumor cell subtypes. O.
   COL6A2 expression in the pseudotime trajectory plot

   Using the TCGA-KIRC dataset, we stratified tumor samples into high and
   low COL6A2 expression groups and performed differential gene expression
   analysis, identifying 996 differentially expressed genes. Specifically,
   compared to the low expression group, 585 genes were upregulated and
   411 downregulated in the high expression group (Fig. [97]4B). GSEA was
   performed on these differentially expressed genes, sorted by P-value
   from smallest to largest. The top 10 GSEA-KEGG enriched pathways
   revealed that Protein digestion and absorption, ECM − receptor
   interaction, PI3K − Akt signaling pathway, Proteoglycans in cancer,
   Focal adhesion, Relaxin signaling pathway, and other pathways were
   significantly enriched in the high COL6A2 expression group, while
   Metabolic pathways were enriched in the low expression group
   (Fig. [98]4C). The top 5 GSEA-GOBP (biological process) enriched
   pathways showed that biological processes significantly associated with
   the high COL6A2 expression group included extracellular matrix
   organization, anatomical structure morphogenesis, external
   encapsulating structure organization, extracellular structure
   organization, and system development (Fig. [99]4F). Further GO and KEGG
   enrichment analysis of the upregulated and downregulated genes
   indicated that the upregulated gene set was closely related to pathways
   such as PI3K − Akt signaling pathway, Focal adhesion, ECM − receptor
   interaction, TGF − beta signaling pathway, and primarily enriched in
   biological processes like extracellular matrix organization, external
   encapsulating structure organization, collagen fibril organization, and
   collagen metabolic process (Fig. [100]4D, SupFigure [101]2B). The
   downregulated gene set was related to metabolic pathways such as
   steroid hormone biosynthesis, retinol metabolism, pyruvate metabolism,
   and involved in ion transport processes including organic anion
   transport, carboxylic acid transport, organic acid transport,
   monocarboxylic acid transport, inorganic anion transport, inorganic
   anion transmembrane transport, nucleosome assembly, and sodium ion
   transport (Fig. [102]4E, SupFigure [103]2 C). The comprehensive
   analysis suggests that COL6A2 influences tumor invasiveness and
   metastatic capacity by regulating the EMT process and is associated
   with the activation of the PI3K-Akt signaling pathway. Additionally,
   the EMT-related gene set from the CancerSEA website confirmed the role
   of COL6A2 in the EMT process. Correlation analysis showed that high
   COL6A2 expression was significantly positively correlated with the
   expression of EMT-related genes (SupFigure [104]2 A), and EMT-related
   genes were significantly upregulated in the high COL6A2 expression
   group (P < 0.05) (SupFigure [105]2D), confirming that COL6A2 plays an
   important role in the progression of ccRCC by promoting the EMT
   process.

Single-cell and spatial profiling reveals mesenchymal-specific COL6A2
expression

   At the single-cell level, we used the Seurat software package to
   cluster cells, identifying major cell populations, including
   mesenchymal cells, B cells, MAST cells, immune cells, malignant cells,
   myeloid cells, NK cells, endothelial cells, mesangial cells, NKT cells,
   T cells, and normal tubule cells (Fig. [106]4G, SupFigure [107]3 A-C).
   By analyzing COL6A2 expression across different cell populations, we
   found that COL6A2 expression was significantly higher in tumor
   epithelial cells than in normal renal tubular cells (P < 0.05)
   (Fig. [108]4I). Additionally, COL6A2 expression was significantly
   elevated in tumor-associated fibroblasts and endothelial cells
   (Fig. [109]4H). Additionally, we performed a secondary clustering
   analysis on tumor epithelial cells, dividing them into four subtypes
   based on cellular markers reported in the literature: epithelial,
   inflammatory, intermediate, and mesenchymal (Fig. [110]4J). The
   analysis showed that COL6A2 was predominantly expressed in the
   mesenchymal subtype of tumor cells (Fig. [111]4K). Using the R package
   GSVA, EMT enrichment scores were analyzed for tumor epithelial cells
   and were found to be predominantly concentrated in the mesenchymal
   subtype (Fig. [112]4L). Pseudotime analysis showed epithelial tumor
   cells transitioning to mesenchymal tumor cells, marked by a gradual
   increase in COL6A2 expression, consistent with the trends of
   EMT-related genes (Fig. [113]4M-O, SupFigure [114]3D). Spatial
   transcriptomic analysis also showed a strong spatial correlation
   between COL6A2 expression and EMT scores (SupFigure [115]3E). These
   results further confirm the potential key role of COL6A2 in the EMT
   process of ccRCC.

COL6A2 drives tumor progression via PI3K-AKT/TGF-β and EMT

   We began by examining the RNA expression levels of COL6A2 at the
   cellular level. COL6A2 expression was significantly higher in the
   786-O, ACHN, and OS-RC-2 cell lines compared to normal HK2 cells
   (Fig. [116]5A). Among these, ACHN and OS-RC-2 cell lines, which
   exhibited the highest COL6A2 expression, were selected to create stable
   knockdown models for further functional analysis.

Fig. 5.

   [117]Fig. 5
   [118]Open in a new tab

   Functional exploration of COL6A2 in ccRCC at the cellular level. A. RNA
   expression of COL6A2 in HK2 and ccRCC cell lines. B. RNA expression of
   COL6A2 in OS-RC-2 and ACHN cell lines after knockdown. C. Cell
   proliferation curves of OS-RC-2 and ACHN cell lines following COL6A2
   knockdown, measured by CCK-8 assay. D-F. Transwell Migration Assays,
   Colony formation assays, Scratch assays in OS-RC-2 and ACHN cell lines
   following COL6A2 knockdown. G. Bar graphs of Transwell Migration Assay,
   Colony formation assay, Scratch Assay results. H. Protein expression of
   COL6A2 and EMT-related markers in OS-RC-2 and ACHN cell lines following
   knockdown. I. Expression changes of PI3K-AKT and TGF-β signaling
   pathway-related markers in OS-RC-2 and ACHN cell lines following COL6A2
   knockdown

   After successfully silencing COL6A2 (Fig. [119]5B), we used CCK-8
   assays and colony formation tests to assess its impact on ccRCC cell
   proliferation. Inhibition of COL6A2 significantly reduced the
   proliferation of both ACHN and OS-RC-2 cells (Fig. [120]5C, D, G).
   Transwell Migration and Scratch assays showed that silencing COL6A2
   significantly impaired the migratory potential of these cell lines
   (Fig. [121]5E, F, G). These findings underscore the critical role of
   COL6A2 in promoting ccRCC cell proliferation and migration.

   Bioinformatics analysis revealed a strong correlation between COL6A2
   and the EMT process. Based on this, we investigated changes in
   EMT-associated markers following COL6A2 knockdown. The results showed
   an increase in E-cadherin, β-catenin expression and a decrease in
   N-cadherin, Vimentin, and Snail1 levels (Fig. [122]5H). Additionally,
   the study identified a strong relationship between COL6A2 and the
   PI3K-AKT and TGF-β signaling pathways, suggesting that COL6A2 regulates
   EMT, as well as tumor cell growth and invasion, through these pathways.
   To further validate this hypothesis, we analyzed the expression of key
   markers in the PI3K-AKT and TGF-β pathways after COL6A2 knockdown.
   Although no significant changes were observed in the protein levels of
   PI3K and AKT, the phosphorylation levels of these proteins were notably
   reduced. Furthermore, the expression of TGF was markedly decreased
   following COL6A2 knockdown. These results provide support for the
   hypothesis that COL6A2 regulates EMT, tumor growth, and invasion
   through the PI3K-AKT and TGF-β signaling pathways (Fig. [123]5I).

COL6A2 modulates therapeutic response to targeted agents and immunotherapy

   Drug sensitivity analysis revealed that patients with high COL6A2
   expression exhibited an increased response to targeted therapies,
   including sunitinib, pazopanib, axitinib, cabozantinib, and sorafenib
   (Fig. [124]6A). To investigate the relationship between COL6A2 and
   immunotherapy, we used the TIDE tool to evaluate the response of tumor
   samples from the TCGA-KIRC dataset to immunotherapy. Tumor samples with
   high COL6A2 expression had significantly higher TIDE scores and lower
   immunotherapy response rates compared to the low-expression group
   (Fig. [125]6B). Further TIDE analysis revealed elevated scores for
   Dysfunction, Exclusion, Myeloid-derived suppressor cell (MDSC), and
   cancer-associated fibroblasts (CAF) in the high-expression group, while
   Microsatellite Instability (MSI) scores and CD274 levels were lower,
   despite higher Cytotoxic T lymphocyte (CTL) proportion (Fig. [126]6B).
   Collectively, these findings suggest that patients with high COL6A2
   expression exhibit an overall immunosuppressive profile and a poorer
   response to immunotherapy.

Fig. 6.

   [127]Fig. 6
   [128]Open in a new tab

   Key Role of COL6A2 in Regulating Sensitivity to Sunitinib and
   Immunotherapy in ccRCC. A. Distribution of drug sensitivity scores for
   Sunitinib, Pazopanib, Axitinib, Cabozantinib, and Sorafenib in high vs.
   low COL6A2 expression groups. B. Distribution of TIDE prediction
   results, TIDE scores, and other immune infiltration scores for immune
   therapy response in high vs. low COL6A2 expression groups. C. IC50
   curves of OS-RC-2 and ACHN cell lines in response to Sunitinib
   following COL6A2 knockdown. D-F. Tumor growth rate and body weight
   changes in BALB/c-nu mice following COL6A2 knockdown with/without
   Sunitinib. G-I. Tumor growth rate and body weight changes in BALB/c
   mice following COL6A2 knockdown with/without RMP1-14

   To validate the functional role and clinical relevance of COL6A2, we
   examined changes in tumorigenicity and sunitinib sensitivity following
   COL6A2 knockdown in both cellular and animal models. At the cellular
   level, IC50 experiments demonstrated that tumor cells with COL6A2
   knockdown exhibited significantly increased IC50 values compared to
   controls (Fig. [129]6C). In the animal model, a subcutaneous tumor
   implantation experiment was conducted in nude mice. Tumor growth was
   significantly suppressed in the COL6A2 knockdown group compared to
   controls (Fig. [130]6D-F). However, after sunitinib treatment, no
   significant difference in tumor size was observed between the knockdown
   and control groups (Fig. [131]6D-F). Additionally, to explore the
   relationship between COL6A2 and immunotherapy, we performed
   subcutaneous tumor implantation experiments on immunocompetent mice
   using normal and COL6A2-knockdown RENCA cells. These mice were also
   treated with anti-PD-1 antibodies. The results showed that the
   knockdown group exhibited significantly reduced tumor growth rates and
   body weight, and displayed higher sensitivity to anti-PD-1 treatment
   (Fig. [132]6G-I).

Discussion

   Renal cell carcinoma, particularly ccRCC, exhibits high incidence and
   mortality rates worldwide [[133]4]. COL6A2, a component of collagen VI,
   has demonstrated significant clinical relevance in various tumors
   [[134]8, [135]19, [136]20]. Research on the mechanisms of action and
   clinical value of the COL6A2 gene in ccRCC is still in its early
   stages. This study focuses on COL6A2 to investigate its potential
   clinical significance and biological functions in ccRCC.

   Through a comprehensive analysis of public databases and clinical
   samples, we found that COL6A2 expression in ccRCC is significantly
   elevated and closely associated with clinical features such as TNM
   staging and stage grouping. We confirmed using various statistical
   methods (including multivariate Cox regression, Kaplan-Meier survival
   curves, nomograms, and ROC curves) that COL6A2 is an independent
   prognostic factor for ccRCC, with high expression significantly
   associated with poor survival outcomes (OS, PFI, PFS). These findings
   suggest that COL6A2 plays a key role in prognosis assessment for ccRCC.

   This study explores the specific mechanisms of COL6A2 in ccRCC,
   focusing on differential gene analysis and pathway enrichment. The
   analysis reveals that ccRCC with varying COL6A2 expression levels
   exhibit distinct biological and functional characteristics. In ccRCC
   with high COL6A2 expression, pathway enrichment analysis shows that
   tumor biological processes are primarily related to extracellular
   structures, matrices, and collagen fibers, consistent with previous
   reports on COL6A2 function. Additionally, the high-expression group is
   associated with several key signaling pathways, including extracellular
   matrix-receptor interaction, focal adhesion, PI3K-Akt, and TGF-β
   pathways, which are known to play critical roles in tumor
   proliferation, invasion, EMT, and immune escape [[137]21–[138]25]. In
   contrast, ccRCC with low COL6A2 expression is primarily associated with
   metabolic processes, including organic/inorganic anion transport,
   retinol metabolism, pyruvate metabolism, and cytochrome P450-mediated
   drug metabolism. These findings suggest that varying COL6A2 expression
   levels in ccRCC may reflect tumor heterogeneity, with high expression
   promoting tumor development, while low expression impacts development
   through changes in metabolic pathways.

   EMT is a key process in tumor progression, influencing tumor
   invasiveness, metastasis, immune escape, and drug resistance, thus
   holding critical significance in cancer therapy [[139]26–[140]28].
   Previous studies have shown that COL6A2 promotes EMT in gliomas and
   colorectal tumors [[141]29, [142]30]. To explore the relationship
   between COL6A2 and EMT in ccRCC, we compiled an EMT-related gene set
   and calculated an EMT score using the GSVA algorithm. In the TCGA-KIRC
   database, COL6A2 showed a strong positive correlation with EMT-related
   genes, which were significantly upregulated in the high-expression
   group compared to the low-expression group. At the single-cell level,
   tumor cells were classified into four subtypes: epithelial,
   inflammatory, intermediate, and mesenchymal, with COL6A2 predominantly
   expressed in mesenchymal tumor cells, consistent with the distribution
   of EMT scores. Pseudotime analysis simulating the EMT transformation
   from epithelial to mesenchymal tumor cells showed that CDH1 and KRT18
   expression gradually decreased, while CDH2, VIM, SNAI2, TIMP1, and
   TGFBI were upregulated, confirming the reliability of the simulation.
   Meanwhile, COL6A2 expression gradually increased during this process.
   Spatial transcriptomic analysis showed that COL6A2 expression closely
   correlates with the spatial distribution of EMT scores. In summary,
   this study confirms that COL6A2 is closely associated with the EMT
   process in ccRCC.

   Through functional exploration and pathway enrichment analysis, this
   study identified three potential roles of COL6A2 in ccRCC. First,
   COL6A2 may promote proliferation, invasion, and metastasis in ccRCC by
   modulating the PI3K-AKT signaling pathway. Second, COL6A2 may promote
   EMT via the TGF-β signaling pathway, thereby influencing tumor
   development. Lastly, as a key extracellular matrix component, COL6A2
   promotes tumor development by regulating the tumor microenvironment. At
   the cellular level, COL6A2 knockdown reduced the expression of key
   molecules in the PI3K-Akt and TGF-β pathways, as well as EMT-related
   markers, with increased E-cadherin, β-catenin expression and decreased
   levels of N-cadherin, vimentin, Snail1. Furthermore, COL6A2 knockdown
   significantly reduced the proliferation, migration, and invasion of
   renal cancer cells, providing preliminary evidence of COL6A2’s role in
   ccRCC development.

   Regarding the tumor microenvironment, we conducted immune infiltration
   analysis in this study. ESTIMATE analysis showed that high COL6A2
   expression correlates with low tumor purity and high immune and stromal
   scores, reflecting an increased content of immune and stromal cells.
   Several immune infiltration methods revealed that COL6A2 is strongly
   correlated with immune-suppressive cells, such as M2 macrophages,
   myeloid-derived suppressor cells, Tregs, and dendritic cells, but not
   with anti-tumor immune cells like CD4 + T cells, CD8 + T cells, or NK
   cells. COL6A2 is strongly associated with CAFs and endothelial cells,
   with high expression in CAFs. CAFs promote tumor growth and invasion by
   producing extracellular matrix (ECM), cytokines, and growth factors,
   influencing the immune microenvironment [[143]31]. The ECM provides
   structural support for tumor cells and plays a critical role in tumor
   development by modulating cell signaling via interactions with surface
   receptors [[144]32]. ECM alterations regulate immune cell behavior,
   creating an immunosuppressive microenvironment and impairing
   immunotherapy efficacy [[145]33]. The comprehensive results suggest
   that COL6A2 promotes tumor immune escape by regulating the infiltration
   of immunosuppressive cells, thereby creating an immunosuppressive tumor
   microenvironment. Additionally, the interaction between COL6A2 and
   stromal cells enhances the tumor’s immunosuppressive environment,
   particularly in immune escape and EMT, creating a microenvironment that
   helps tumor cells evade immune surveillance and resist anti-tumor
   treatments. Therefore, COL6A2 plays a key role in renal cancer invasion
   and metastasis, and it also regulates immune escape by modulating the
   immune microenvironment.

   The Cancer Stemness Index (CSI) is a measure of the gene expression
   characteristics associated with cancer stem cells (CSCs) in tumor cells
   [[146]14]. CSCs are tumor cells with self-renewal and pluripotent
   differentiation capabilities, and they contribute to tumor initiation,
   progression, metastasis, and resistance to therapy [[147]34]. Cancer
   stemness analysis showed that the high COL6A2 expression group had
   higher tumor stemness scores, suggesting that COL6A2 is associated with
   the stemness characteristics of ccRCC. Tumor stem cells possess
   self-renewal and multidirectional differentiation potential, playing a
   critical role in tumor initiation, development, invasion, and
   metastasis. Tumor stem cells acquire invasiveness through the EMT
   process, which involves the loss of adhesion and enhanced migration,
   contributing to metastasis and drug resistance.

   This study also included pharmacological predictive analysis to assess
   the correlation between COL6A2 expression and renal cancer patients’
   sensitivity to various treatments. Higher COL6A2 expression was
   associated with increased sensitivity to targeted drugs like sunitinib,
   pazopanib, and cabozantinib. TIDE analysis revealed that patients with
   high COL6A2 expression had higher TIDE scores and a significantly lower
   proportion of responders to immunotherapy compared to the
   low-expression group, suggesting poorer responses to immune checkpoint
   inhibitors (e.g., PD-1/PD-L1 inhibitors). TIDE analysis further showed
   that the high COL6A2 expression group had lower MSI scores and CD274
   levels at immune checkpoints, while Dysfunction and Exclusion scores,
   as well as MDSC and CAF proportions, were significantly higher compared
   to the low-expression group. This suggests that COL6A2 plays a
   significant role in immune evasion in renal cancer, consistent with
   preliminary analysis results. Experimental validation demonstrated that
   COL6A2 knockdown significantly increased the IC50 values of tumor
   cells, while no significant change in tumor size was observed following
   intratumoral sunitinib injection. However, in the subcutaneous
   tumor-bearing model, tumor size was significantly reduced in the COL6A2
   knockdown group compared to the non-knockdown group. In the anti-PD-1
   treatment model, mice with COL6A2 knockdown exhibited significantly
   enhanced sensitivity to anti-PD-1 treatment compared to the normal
   group. These findings suggest that COL6A2 not only serves as a
   prognostic marker but may also be a potential therapeutic target for
   guiding clinical treatment in ccRCC patients.

   Although this study outlines the potential role of COL6A2 in ccRCC,
   further experimental validation is required. Limitations of this study
   include the unclear molecular mechanisms underlying COL6A2’s
   interaction with the PI3K-AKT and TGF-β pathways, as well as its
   potential synergy with other ECM proteins. Future research should focus
   on COL6A2’s role in the tumor immune microenvironment and its influence
   on immune evasion and EMT. The small sample sizes in existing studies
   limit generalizability, highlighting the need for large-scale,
   multicenter studies. While the association between COL6A2 and tumor
   prognosis is preliminarily established, its clinical application as a
   biomarker and its potential in targeted therapies for renal cancer
   require further validation through extensive clinical trials.

Conclusion

   Overall, this study demonstrates a strong association between high
   COL6A2 expression in ccRCC and tumor progression, the tumor immune
   microenvironment, and clinical prognosis. COL6A2 serves as a potential
   prognostic marker and offers new insights into the selection of
   targeted and immunotherapies for renal cancer. Future research should
   further investigate the biological role of COL6A2 in renal cancer and
   its clinical application, aiming to identify new targets for early
   diagnosis and personalized treatment.

Electronic supplementary material

   Below is the link to the electronic supplementary material.
   [148]Supplementary Material 1^ (2.9MB, docx)

Acknowledgements