Circulating bile acid profiles characteristics and the potential predictive role in clear cell renal cell carcinoma progression

Abstract

Background

The incidence of clear cell renal cell carcinoma (ccRCC) has steadily increased over the past decade, and recent studies have linked bile acid (BA) metabolism to its development. However, the metabolic profile of BAs and their potential as biomarkers in ccRCC pathogenesis remain poorly characterized, making their evaluation crucial for advancing disease understanding and management.

Methods

A total of 68 newly diagnosed ccRCC patients and 63 healthy controls were enrolled. Serum bile acid (BA) profiles were measured using Ultra Performance Liquid Chromatography-Tandem Mass Spectrometry (UPLC-MS/MS). The Orthogonal Projections to Latent Structures Discriminant Analysis (OPLS-DA) model analyzed differences in serum BA profiles between ccRCC patients and controls. Additionally, the relationship between BA profiles and tumor heterogeneity parameters was investigated. Receiver Operating Characteristic (ROC) analysis identified potential biomarkers for ccRCC pathogenesis.

Results

The BA profile was altered in ccRCC patients and was not influenced by sex or age. Specifically, primary and secondary unconjugated BA fractions were significantly higher in the ccRCC population. Five BA metabolite candidates exhibited the most significant differences between ccRCC patients and controls. Deoxycholic acid (DCA) was associated with pathological pTNM stage classification and grade. Chenodeoxycholic acid (CDCA) and lithocholic acid (LCA), combined with testosterone, showed potential as biomarkers for the pathogenesis of ccRCC.

Conclusion

Alterations in the serum BA profile are observed in ccRCC. Deoxycholic acid (DCA) correlates with pathological pTNM stage classification and tumor grade. Additionally, CDCA combined with LCA show potential as biomarkers for ccRCC pathogenesis.

Clinical trial number

Not applicable.

Introduction

Renal cell carcinoma (RCC) is characterized by high aggressiveness, significant histological subtypes, and mutational heterogeneity [1]. Among these, clear cell renal cell carcinoma (ccRCC) is the predominant pathological type, accounting for approximately 70–80% of all RCC cases [2]. The clinical presentation of ccRCC is often insidious, lacking effective tumor markers and showing resistance to radiotherapy and chemotherapy, making early diagnosis challenging [3]. Clinically, accurately diagnosing small ccRCC tumors and distinguishing RCC from benign renal lesions is difficult [4]. Serum is widely used in biomarker research for various diseases. However, metabolomics and lipidomics studies focusing on distinguishing ccRCC from benign tumors remain scarce, highlighting a significant area for potential diagnostic advancements in ccRCC.

The distinctive pathological feature of ccRCC is the accumulation of large lipid droplets within tumor cells, characterized by the presence of cholesterol esters, which is positively correlated with tumor development [5, 6]. The “clear” cytoplasm of ccRCC cells reflects the presence of these intracellular lipid droplets, which store neutral lipids including cholesterol, cholesterol esters, and triglycerides [7]. Studies indicate that ccRCC cells are heavily reliant on the formation of these lipid droplets for growth and survival, and inhibiting their formation results in cell death [8]. Although the exact mechanisms are not fully understood, cholesterol is the most abundant lipid in ccRCC lipid droplets, playing a crucial role in maintaining cellular homeostasis [9]. Dysregulated lipid metabolism in ccRCC has been implicated in disease progression.

Cholesterol plays multiple critical roles within cells, particularly in cancer cells, and serves as a precursor for bile acids (BAs) [10]. BAs are primarily synthesized in the liver and conjugated with taurine or glycine to form bile salts. These bile salts are released from the gallbladder into the duodenum, where they facilitate the emulsification and absorption of dietary fats, the excretion of excess cholesterol, and the maintenance of gut microbial homeostasis [11]. Previous research has predominantly focused on hepatic and gastrointestinal diseases, with limited studies on renal diseases. Recent studies have highlighted the importance of BA metabolism in ccRCC [7]. Elevated expression of genes in the BA biosynthetic pathway has been observed in ccRCC tumors compared to normal kidney tissue, suggesting a crucial role in ccRCC initiation and progression [11]. Additionally, BA receptors such as the farnesoid X receptor (FXR) and Takeda G protein-coupled receptor 5 (TGR5) are differentially expressed in renal tissues [12, 13]. Cell activity experiments have demonstrated that TGR5 enhances the activity of renal cell carcinoma [14], while FXR was involved in the growth of renal adenocarcinoma cells [13]. To date, there have been no studies on the changes in serum bile acid composition in ccRCC patients.

This study aims to characterize and compare the absolute concentration and composition of 15 serum BAs in ccRCC patients and control groups to investigate the feasibility of using serum BAs as screening or early warning indicators for ccRCC. Furthermore, we evaluate their correlation with cancer progression characteristics and identify potential biomarkers for the pathogenesis of ccRCC.

Materials and methods

Participants

This investigation encompassed 68 newly diagnosed, untreated ccRCC patients who underwent radical nephrectomy, with no prior targeted therapy or chemotherapy administered. Postoperative pathological examinations confirmed ccRCC in all cases. Serum samples from healthy individuals were included as controls for comparative analysis. All assessments and procedures within the study adhered strictly to the principles outlined in the Declaration of Helsinki regarding human experimentation. The study protocol received approval from the Ethical Committee of the Xuzhou Cancer Hospital (approval no. 2024-02-001-K01), and all participants provided informed written consent after being thoroughly briefed on the study details; consent for publication is not applicable.

Serum collection and biochemical index evaluation

Peripheral venous blood (4–5 ml) was drawn from each participant upon admission, centrifuged at 3500 rpm for 10 min at 4 °C to obtain serum, which was then stored at -80 °C for subsequent analysis. Serum biochemical parameters were analyzed using a Beckman AU5800 automated biochemical analyzer. Information collected included body mass index (BMI) [15], and various biochemical indicators such as aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyl transferase (GGT), alkaline phosphatase (ALP) creatinine, uric acid, and cystatin C. Participant characteristics and clinical parameters are detailed in Table 1.

Table 1 General characteristics of the CcRCC and control groups

Serum targeted bile acid profile assessment

BAs were quantified using an Ultrahigh-Performance Liquid Chromatography-Tandem Mass Spectrometry (UPLC-MS/MS) system (Waters Corporation, Milford, Massachusetts, USA) equipped with a triple quadrupole mass spectrometer and a Jet Stream electrospray ionization (ESI) source. Internal standards (ISs) and BAs were obtained from Sigma-Aldrich (Darmstadt, Germany), with abbreviations detailed in the supporting information (Table S1). ISs were prepared by mixing isotope-labeled standards in methanol and diluting with deionized water. For serum analysis, 50 µL of serum was mixed with 5 µL of IS and 150 µL of precooled methanol. After centrifugation at 12,000 g for 10 min at 4 °C, 50 µL of the supernatant was combined with 50 µL of deionized water and transferred to autosampler vials for analysis. Chromatographic separation was performed on an ACQUITY UPLC BEH C18 column (1.7 μm, 100 mm × 2.1 mm) at 45 °C with a gradient elution: 0–3 min, 35% B; 3–4.5 min, 35–60% B; 4.5–5 min, 65–100% B; 5–6 min, 100–35% B. Mobile phase A was 10 mmol/L ammonium acetate in water, and mobile phase B was 0.1% formic acid in acetonitrile, at a flow rate of 0.4 mL/min. Mass spectrometric detection was conducted in negative ion mode using Multiple Reaction Monitoring (MRM). Nitrogen was used as the nebulizer and cone gas, with source and desolvation temperatures set at 150 °C and 450 °C, respectively.

Data were analyzed using MassLynx software (Version 4.1), with calibration curves generated by linear regression. The linearity of all bile acids was validated within the range of 1.0–6000.0 nmol/L, with coefficients of determination (R²) exceeding 0.99 for all compounds. The calibration curves for all measured BAs demonstrated excellent linearity within the defined quantification range, with R² values exceeding 0.99 for all compounds. A total of 15 BAs were analyzed, including six primary BAs: cholic acid (CA, R² = 0.999), chenodeoxycholic acid (CDCA, R² = 0.999), glycocholic acid (GCA, R² = 0.999), taurocholic acid (TCA, R² = 0.999), glycochenodeoxycholic acid (GCDCA, R² = 0.998), and taurochenodeoxycholic acid (TCDCA, R² = 0.997); and nine secondary BAs: deoxycholic acid (DCA, R² = 0.996), ursodeoxycholic acid (UDCA, R² = 0.999), lithocholic acid (LCA, R² = 0.998), glycodeoxycholic acid (GDCA, R² = 0.999), taurodeoxycholic acid (TDCA, R² = 0.999), glycoursodeoxycholic acid (GUDCA, R² = 0.998), taurolithocholic acid (TLCA, R² = 0.999), glycolithocholic acid (GLCA, R² = 0.998), and tauroursodeoxycholic acid (TUDCA, R² = 0.997). These values underscore the accuracy and reliability of the quantification method employed in this study. The limits of quantification (LOQs) for the bile acids ranged from 1.0 to 10.0 nmol/L. Intra- and inter-day precision, expressed as relative standard deviations (RSDs), were below 15%. Additionally, the coefficients of variation (CoVs) for both high and low quality control (QC) samples of the 15 BAs were also below 15%. To ensure the accuracy of quantification and identification, chromatographic peaks were calibrated based on retention times and parameters related to the peak shape of each BA. This calibration enabled reliable comparison of BA concentration differences. Randomly inserted QC samples were included for every 10 serum samples to monitor the reproducibility of instrument performance. The overlap of the total ion chromatograms (TICs) for the QC samples was used as a measure of stability; higher overlap indicated better instrument stability. The final concentrations were calculated by integrating the peak areas and applying the calibration curve equation.

Statistical analysis

Statistical analyses were performed using JASP 19.0, with supplementary analysis in Origin. Bile acid distribution among groups was visualized using SIMCA-P software v14.1 (Umetrics AB, Umea, Sweden). Principal component analysis (PCA) and orthogonal partial least squares -discriminant analysis (OPLS-DA) models were constructed based on the metabolomics data. The variable importance in the projection (VIP) of the first principal component obtained from the OPLS-DA analysis was determined. In the univariate analysis, metabolites with a VIP > 1.0 and P-value < 0.05 were considered significantly different. Moreover, the quality of the OPLS-DA model was assessed using standard parameters (R2X and Q2) [16]. Categorical data were analyzed using the chi-square test. For continuous variables, data were expressed as means ± standard error of the mean (SEM) if normally distributed, or as medians and interquartile ranges if not. Independent sample t-tests were used for normally distributed data meeting parametric assumptions, while the Mann-Whitney U test was used for non-normally distributed data. Regression model to identify independent risk factors for ccRCC, employing forward stepwise regression. A P-value of < 0.05 was considered statistically significant.

Results

Characteristics and analysis of serum bile acid profile in CcRCC patients

Clinical data were collected for sex-matched patients with ccRCC and healthy controls, showing no significant differences in age, BMI, fasting blood glucose, blood pressure, liver and kidney function, total protein, albumin, or cystatin C (Table 1). Both primary and secondary unconjugated BAs were significantly higher in ccRCC patients (Fig. 1A). Ratio analysis of the 15 BA metabolites to total BAs indicated relatively higher percentages of LCA and DCA, and lower percentages of TLCA and GCDCA in ccRCC patients (Fig. 1B). Regarding the distribution of BA levels, we have provided bar plots of BA concentrations for control and ccRCC groups to facilitate visualization of the data distribution (Fig. 1C). These distributions clearly highlight variations in individual BA levels between the two groups, emphasizing the differences observed in our study. Serum BA level analysis revealed significantly elevated concentrations of free BAs (CA, CDCA, DCA, UDCA) and conjugated BAs (GDCA, GLCA) in the ccRCC group compared to controls (Table S2).

Fig. 1

Comparison of serum BA pools between the ccRCC and Control groups. (A) Serum BA composition of total BAs, including unconjugated and conjugated forms, in the ccRCC and Control groups. (B) Relative fractions of BAs as proportions in the ccRCC and Control groups (The proportions of BA species were calculated as concentrations of BA species / Total BA×100%). (C) Bar plots of BA concentrations for control and ccRCC groups. *P < 0.05, **P < 0.01, ***P < 0.001

Since strong significant differences were found in the BAs, to better quantify the difference between the ccRCC and control group, besides frequentist statistics, Bayesian T-test was used to compare the difference between the DCA and CA of ccRCC group and Control group. As shown in Fig. 2, A Bayes factor of 158 suggests that an alternative model is 158 times more favored than a null model, given the data.

Fig. 2

Bayesian Independent Samples T-Test for ccRCC Patients and Controls (A). DCA Bayesian Mann-Whitney. (B) CA Bayesian Mann-Whitney. The probability wheel on top visualizes the evidence that the data provide for the two rival hypotheses. The two gray dots indicate the prior and posterior density at the test value. The median and the 95% central credible interval of the posterior distribution are shown in the top right corner. BF₁₀ = 29300 means very strong evidence support the alternative hypothesis. Note: Following the proposals made by Wetzels, van Ravenzwaaij, and Wagenmakers (2015), based on Jeffreys (1961), the Bayesian findings were interpreted as follows. (I) Clear evidence for the alternate hypothesis (extremely strong evidence: BF₁₀ > 100; very strong evidence: 30–100; strong evidence: 10–30; moderate evidence: 3–10). II) Anecdotal evidence for the alternate hypothesis: BF10 = 1–3; (III) No evidence: BF₁₀ = 1

Identification of significantly different BAs associated with CcRCC

To comprehensively compare the BA metabolomic profiles between ccRCC patients and controls, we employed the PCA and OPLS-DA model to assess the degree of diversity. The PCA model demonstrated an inability to distinctly separate the two groups (R2X(cum) 0.454, Q2(cum) 0.095; Fig. 3A). The OPLS-DA model revealed a clear distinction between the ccRCC and Control groups, with serum BA profiles clustering into two distinct groups, indicating significant differences. The OPLS-DA model exhibited clearer discrimination, though not entirely conclusive (R2X(cum) = 0.728, Q2(cum) = 0.141; Fig. 3B). Considering the inherent variability in human biological samples [16, 17], the R2X value being above 0.5(with a relatively low Q2 value of at 0.141) was deemed acceptable. The analysis identified four BAs—DCA, CA and CDCA—with VIP values greater than 1, showing significant differences (P < 0.05) in the univariate analysis between the groups (Fig. 3C, Table S2). To further validate the prediction model, we conducted permutation tests. The Q2 intercept value falling below 0 indicated a reliable and non-overfitted model (Fig. 3D). Notably, DCA exhibited the most significant elevation in the ccRCC group and held substantial relevance in the BA composition compared to controls (VIP score = 3.24771, P < 0.001).

Fig. 3

PCA model and OPLS-DA model plots derived from BA metabolomic profiles comparing ccRCC and Control groups. (A) Serum BA profiles displayed in Principal component analysis (PCA) model scores plot for ccRCC group (green) and Control group (blue). (B) Serum BA profiles displayed in OPLS-DA scores plot for ccRCC group (green) and Control group (blue). (C) VIP scores from OPLS-DA highly ghting significant serum BA profile differences between the ccRCC and Control groups (VIP value > 1 indicates discrimination importance). (D) Permutation test of OPLS-DA Model of the BAs. The permutation plot strongly indicates that the model is valid. The criteria for validity are: (1) All blue Q2-value to the left are lower than the original points to the right; (2) The blue regression line of the Q2-points intersects the vertical axis (on the left) at, or below zero

The candidate value of BAs in serum for CcRCC

Our objective was to ascertain the predictive utility of BAs for ccRCC. We performed receiver operating characteristic (ROC) and logistic regression analyses on the differential serum BAs identified. ROC curves distinguished healthy controls from ccRCC patients, evaluating the area under the curve (AUC), and determining the sensitivity and specificity of cut-off values for each BA (Fig. 4A-B; Table 2). Serum DCA exhibited the highest correlation with ccRCC, yielding an AUC of 0.82 (P < 0.001). The sensitivity and specificity for predicting ccRCC were 61.8% and 81.0%, respectively. Subsequent logistic regression combined with ROC analysis, encompassing DCA, TCDCA, LCA, TLCA, further enhanced the predictive performance for ccRCC. The combined model yielded an AUC of 0.847, with a sensitivity of 67.6% and specificity of 84.1%, highlighting the potential of BA concentration measurement, particularly conjugated BAs, in ccRCC screening. Additionally, we performed a multivariate logistic regression analysis; the results indicated that while there was no significant effect of sex on the BA profiles, age may have an influence, as suggested by a near-significant difference(Table S4).

Fig. 4

ROC curve analysis of predictive value of DCA and CA indicators for ccRCC patients. (A) ROC curve for DCA in predicting ccRCC patients. (B) ROC curve for CA in predicting ccRCC patients. In this study, the ROC curves was analyzed by logistic regression part of JASP [JASP Team (2024). JASP (Version 0.19.2)]

Table 2 The model summary of the logistic regression

Discussion

ccRCC is the most prevalent and deadly subtype of renal cancer, accounting for 70-80% of cases [18]. Its high mortality rate is primarily due to its asymptomatic early stages, resulting in late diagnosis for about 30% of patients. Currently, invasive biopsy is the only definitive diagnostic method, as no specific clinical biomarkers are available for ccRCC [3]. Therefore, developing early diagnostic methods is crucial. Recent research into the metabolic pathways of ccRCC cells offers promising avenues for non-invasive and accurate screening technologies, which could significantly enhance diagnosis and treatment strategies [19, 20]. ccRCC is a metabolic disease marked by significant alterations in cellular metabolism, including changes in glycolysis, mitochondrial bioenergetics, oxidative phosphorylation (OxPhos), and lipid metabolism [21, 22]. Morphologically, ccRCC cells are characterized by an accumulation of lipids and glycogen, reflecting these metabolic reprogramming processes.

Recent studies have underscored the crucial role of BAs in regulating key metabolic pathways, including glycolysis, mitochondrial function, and lipid metabolism—processes central to RCC pathogenesis. Alterations in BA metabolism are linked to shifts in cellular energy production and lipid homeostasis, reinforcing the metabolic reprogramming characteristic of RCC [23]. These findings suggest that BAs may serve as both critical regulators of RCC metabolism and potential biomarkers for disease progression. Consequently, ccRCC exemplifies the role of metabolic reprogramming in cancer, highlighting the importance of metabolic shifts in its pathogenesis [24, 25]. Additionally, cholesterol, a key component in cancer cell physiology and a precursor for BAs, is vital for various cellular functions, further linking lipid metabolism with tumor progression. BAs are synthesized in the liver, conjugated with taurine or glycine to form bile salts, and stored in the gallbladder [26]. Upon release into the duodenum, these bile salts facilitate the emulsification and absorption of dietary fats, promote the excretion of excess cholesterol, and help maintain gut microbial homeostasis [27]. These processes are crucial for sustaining metabolic health and have significant implications in oncological contexts. BAs play roles in various cancers, synthesized in the liver but involving multiple organs for balance and excretion. Initially recognized as carcinogenic due to their signaling roles [28, 29], recent studies suggest anti-tumor properties in some cancers, such as breast cancer [30, 31]. The serum BA profile reflects the bile profile, making it feasible to use serum BAs as screening markers for ccRCC. However, few studies have focused on using serum BAs to distinguish ccRCC from benign tumors, which would significantly aid in differential diagnosis. Our study revealed significant alterations in the serum bile acid profile of ccRCC patients, identifying four free BAs (DCA, CDCA, UDCA) and two conjugated bile acids (GDCA, GLCA) with statistically significant differences. Among these, DCA was the most discriminative factor between groups. Additionally, DCA demonstrated the highest diagnostic efficacy as a single marker, with an AUC of 0.82.

DCA, primarily produced in the cecum and proximal colon by microbial activity, exhibits both antibacterial properties and toxicity [32]. DCA acts as a tumor-promoting agent by inducing apoptosis and promoting cancer cell proliferation, significantly contributing to various cancers. DCA induced oxidative stress, DNA damage, and inflammation, leading to esophageal adenocarcinoma (EAC), while UDCA mitigated DCA-induced injury [33, 34]. Acidic bile salts activated telomerase via a c-Myc-dependent pathway, and DCA induced a metaplastic phenotype in gastric cancer cells [35, 36]. Additionally, hydrophobic BAs like DCA were prominent promoters of liver cancer, contributing to hepatocellular carcinoma (HCC) [37, 38]. The Western diet’s tumor-promoting activity was linked to elevated colonic BA concentrations, mainly LCA and DCA, as observed in colorectal carcinoma (CRC) patients [39]. Notably, low concentrations of DCA (0.05–0.3 mM) inhibited colonic cell proliferation through cell cycle arrest and apoptosis pathways [40]. DCA also modulated the expression of breast cancer type 1 susceptibility protein (BRCA1) and the estrogen receptor, influencing the drug sensitivity of ovarian cancer cells [41].

BAs play a crucial role in kidney pathophysiology through the activation of the farnesoid X.

receptor (FXR) and G protein-coupled bile acid receptor 1 (GPBAR1/TGR5) [42]. These receptors have been implicated in RCC pathogenesis. TGR5 was reported to inhibit inflammation via the NF-κB pathway, mitigating diabetic nephropathy and potentially preventing RCC cell proliferation [43, 44]. FXR, essential for cholesterol/BA homeostasis, stimulated renal adenocarcinoma proliferation; its knockdown suppressed tumor growth without harming normal cells [13]. TCGA data revealed dysregulated TGR5 and FXR in RCC tissues, highlighting their significance [45]. DCA, a potent TGR5 and FXR agonist, inhibited acetylcholine-induced inositol phosphate formation, likely targeting inositol polyphosphate hydrolase, whose activity was reduced in renal tumors [46]. Elevated DCA conjugates in RCC patients inhibited this enzyme’s activity, contributing to cancer progression. Based on the analysis of BA profiles in (ccRCC, we identified potential BA biomarkers for ccRCC pathogenesis. Our findings indicated that the combined use of DCA, LCA, TCDCA and TLCA provided a more effective predictive model for ccRCC than DCA alone. These results enhance our understanding of BA-related signaling pathways, suggesting that future research should investigate downstream molecules within these pathways to further elucidate their roles in ccRCC.

Our study highlights the altered BA profile in the serum of ccRCC patients, revealing significant changes in the circulating levels of various bile acids. Interestingly, while our study focuses on the circulating bile acid profile in serum, Li et al. [47]. reported a decrease in bile acid levels within ccRCC tumor tissues, despite an increase in cholesterol. This apparent discrepancy may arise from differences in the biological compartments analyzed. In our study, the serum BA levels reflect systemic metabolic changes, influenced by factors such as liver function, bile acid synthesis, and intestinal absorption. In contrast, tumor tissues are known to undergo metabolic reprogramming, which may lead to localized alterations in bile acid metabolism, including reduced BA levels in the tumor microenvironment. These differences underscore the complexity of metabolic shifts in ccRCC and highlight the need for further studies that directly compare bile acid levels in both serum and tumor tissue to better understand their role in disease progression. This discrepancy between serum and tumor tissue may also be attributed to the unique properties of the tumor microenvironment. Tumors often exhibit altered lipid and energy metabolism, which may affect the local synthesis, transport, and catabolism of bile acids. Additionally, the interaction between tumor cells and the surrounding stromal cells can influence the uptake and secretion of BAs. These factors suggest that while serum BAs may serve as potential biomarkers of systemic disease, their role in the local tumor environment might differ, and further research is needed to explore these dynamics in greater detail.

Our study has several limitations that warrant a cautious interpretation of our findings. The single-center, retrospective, and cross-sectional design inherently limits the generalizability of our results and precludes establishing a causal relationship between altered bile acid (BA) profiles and the development of clear cell renal cell carcinoma (ccRCC). Although our findings indicate that deoxycholic acid (DCA) is correlated with pathological grade, stage, and related clinical indicators, the study design does not allow us to determine whether these alterations in DCA levels are a contributing factor to tumor progression or simply a reflection of underlying metabolic changes associated with ccRCC. Additionally, selection bias and the limited sample size, particularly the absence of late-stage ccRCC cases, may have affected the identification of independent risk factors and restricted the statistical power necessary for robust subgroup analyses. For instance, there is a near-significant age difference between ccRCC patients and controls (P = 0.06), and to address this, we performed a multivariate logistic regression analysis including both age and sex as covariates. The results indicated that while sex did not have a significant effect on BA profiles, age may have a near-significant effect, suggesting that the observed age influence might be attributable to the sample limitations inherent in our study design. Moreover, our study did not examine the interplay between BAs and gut microbiota—a relationship increasingly recognized as crucial in metabolic regulation, as the gut-liver axis plays a significant role in BA metabolism and alterations in the gut microbiome could impact BA synthesis, transformation, and systemic circulation. Future studies integrating microbiome analysis, as well as multicenter prospective randomized controlled trials with larger cohorts, will be critical to validate our findings and to develop potential BA-based diagnostic and therapeutic strategies for ccRCC.