Renal cell carcinoma (RCC) is the most common form of kidney cancer, claiming around 17,900 lives each year. Clear cell renal cell carcinoma (ccRCC) is the most common pathological subtype of RCC, and its diagnosis relies heavily on postoperative pathology. Although imaging features are helpful in certain respects, they are not absolutely convincing for diagnosing ccRCC. Today, liquid biopsy technologies are attracting considerable attention. The detection of exfoliated urothelial cells has become a routine diagnostic method for bladder cancer, valued for its simplicity and high sensitivity. For ccRCC, identifying molecular markers in serum and urine can support patient screening, diagnosis and follow-up. Most studies to date, however, have focused on serum or urine metabolomics, microRNA and ctDNA.

The Department of Urology at Changhai Hospital, Naval Medical University (Second Military Medical University), applied urinary proteomics to explore diagnostic biomarkers for ccRCC and construct a prognostic model, work that carries meaningful clinical value. PKHD1L1, ANGPTL6, FABP4 and C3 can aid in the diagnosis of ccRCC, and a prognostic model comprising VSIG4, HLA-DRA, SERPINF1 and IGLV2-23 can significantly predict prognosis in ccRCC patients. The findings were published in Cancer Molecular Targets and Therapeutics.

Urinary proteomic profiling of ccRCC patients and non-ccRCC patients

Urine samples from 12 ccRCC patients and 11 non-ccRCC patients underwent 4D label-free quantitative proteomic analysis, identifying 133 differentially expressed proteins, of which 85 were upregulated (fold change, FC > 1.5) and 48 downregulated (FC < 0.5). Figure 2A presents a heatmap of all differentially expressed genes, while the correlation analysis in Figure 2B shows heterogeneity across urine samples. Taken together, Figures 2A and 2B indicate that although urinary proteins are highly heterogeneous, much important information remains to be noted. The study therefore selected the top 10 proteins—S100A14, PKHD1L1, FABP4, ITIH2, C3, C8G, C2, ATF6, ANGPTL6 and F13B—as potential diagnostic biomarkers for ccRCC (Figure 2C).

Figure 2. Overview of the proteomic analysis of ccRCC patients and controls

(A) Heatmap of differentially expressed proteins (DEPs); (B) heatmap of Pearson correlation coefficients between samples; (C) volcano plot of all DEPs, with the top 10 upregulated proteins labeled.

PKHD1L1, C3, FABP4 and ANGPTL6 can be used to diagnose ccRCC

To verify whether the top 10 proteins can serve as diagnostic biomarkers for ccRCC, urine samples were collected from 46 ccRCC patients and 45 non-ccRCC patients for ELISA testing (Figure 3A). As expected, the urinary levels of four proteins—PKHD1L1, FABP4, C3 and ANGPTL6—were significantly higher in ccRCC patients than in non-ccRCC patients. Individually, however, PKHD1L1, FABP4, C3 and ANGPTL6 showed limited diagnostic efficiency (Figure 3B). The study therefore performed multivariate logistic regression analysis and constructed a nomogram (Figure 3C), which raised the diagnostic efficiency for ccRCC to 0.835.

Figure 3. Construction of the diagnostic model

(A) Scatter plot of ELISA results; (B) ROC curves of the individual diagnostic sensitivity of PKHD1L1, C3, FABP4 and ANGPTL6; (C) nomogram and ROC curve for the combined use of PKHD1L1, C3, FABP4 and ANGPTL6. *p < 0.05, **p < 0.01.

A prognostic model incorporating SERPINF1, HLA-DRA, VSIG4 and IGLV2-23

The researchers were keenly interested in whether urinary proteins could inform the prognosis of ccRCC patients. They therefore downloaded TCGA KIRC RNA-seq data and took the intersection of the two datasets (Figure 4A). Twenty-five genes appeared in the intersection; COX regression analysis showed that SERPINF1, HLA-DRA, VSIG4 and IGLV2-23 were statistically significant. These four molecules were combined with patient age, pathological grade and pathological stage to build a nomogram model for visualization (Figure 4B).

The study further explored the relationship between T stage and M stage and the expression of SERPINF1, HLA-DRA, VSIG4 and IGLV2-23. As shown in Figures 4B and 4C, HLA-DRA expression was unrelated to T and M stage; expression of IGLV2-23, SERPINF1 and VSIG4 was associated with T stage; and all except IGLV2-23 were also associated with M stage. For overall survival, high expression of HLA-DRA, IGLV2-23 and SERPINF1 predicted poorer outcomes, whereas no difference was observed between the high- and low-VSIG4 groups (Figure 4D). Although T and M stage showed no differences relative to HLA-DRA expression, the prognostic value of HLA-DRA remained statistically significant.

Figure 4. Construction of the prognostic model and prognosis of HLA-DRA, IGLV2-23, SERPINF1 and VSIG4

(A) Left: Venn diagram of TCGA RNA-seq data and urinary proteomics; right: prognostic model diagram for SERPINF1, HLA-DRA, VSIG4 and IGLV2-23; (B) scatter plot of the relationship between T stage and expression of HLA-DRA, IGLV2-23, SERPINF1 and VSIG4; (C) scatter plot of the relationship between M stage and expression of HLA-DRA, IGLV2-23, SERPINF1 and VSIG4; (D) Kaplan–Meier curves of overall survival (OS) for HLA-DRA, IGLV2-23, SERPINF1 and VSIG4. Ns: p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001.

Based on the expression levels of these five genes, a survival risk score was calculated for each case and patients were stratified into high-risk and low-risk groups (Figure 5A). The risk factor plot showed that high expression of SERPINF1, VSIG4 and IGLV2-23 was associated with higher risk scores, while higher HLA-DRA expression was associated with lower risk scores. As the study anticipated, the high-risk group had poorer prognosis, meaning that patients with high risk scores tend to have shorter survival (Figure 5B).

Figure 5. Predictive efficiency of the prognostic model

(A) Risk factor plot for the high-risk and low-risk groups; (B) K–M curves for the high-risk and low-risk groups.

For urological diseases, urine testing is a completely non-invasive and promising form of liquid biopsy. According to this study, urinary proteins differ significantly between ccRCC patients and controls, and diagnostic and prognostic models for ccRCC can be constructed from these differential proteins. Studies with larger sample sizes are needed to confirm these findings, so that relatively accurate preliminary screening for ccRCC can be achieved and the prognosis of ccRCC patients predicted through urine testing.

The ProteinT liquid biopsy discovery platform

The n-LAPE/MS™ novel liquid biopsy biomarker discovery platform, independently developed by ProteinT, draws on a library of dozens of self-synthesized, surface-modified nanomaterials. Based on high-resolution LC-MS/MS technology, it specifically and efficiently enriches low-abundance proteins in body fluids, delivering a revolutionary increase in the number of proteins detectable by mass spectrometry in body fluid samples.

In urinary proteomics specifically, using the Q Exactive series of mass spectrometry platforms, more than 6,000 proteins can be detected from a single sample within an effective half-hour gradient. This breaks through the technical bottleneck of urine-based liquid biopsy and can be widely applied to clinical research and exploration in urological diseases and cancers, enabling the discovery of protein biomarkers for early screening, early diagnosis and disease monitoring.