The tissue biopsies were analyzed in duplicate using liquid chromatography coupled to tandem mass spectrometry (nanoLC-ESI-HR-MS/MS). Raw spectral intensity values were converted into absolute concentrations using the TPA method [10]. A total of 1610 proteins were identified (1% FDR) across all samples, with 1379 (85.7%) of these proteins being quantified in at least seven out of the ten LC–MS/MS runs for one renal tissue type (chRCC, RO or NAT). Figure 1a displays the principal component analysis (PCA) of the quantified proteins, showing that each specimen class segregates into distinct clusters.

Classification of human renal tissue proteomes and protein expression profiles in tumors relative to NATs. Two instrumental replicates were run for each biopsy, resulting in 30 chromatograms that identified a total of 1610 proteins. Of these proteins, a total of 1379 were obtained from at least one tissue group (chRCC, RO, or NAT), with a reproducibility ranging from 70 to 100%. a Principal component analysis (PCA) of chRCC, RO, and NAT group samples. b Heatmap representation of Pearson correlation of biological (n = 5) and technical (n = 2, each sample) replicates (lower left) in combination with a scatterplot matrix protein of TPA-based quantification (upper right) demonstrating the reproducibility of the MS data. Volcano plots illustrate the significantly different (FDR = 0.01, S0 = 0.1) protein expression levels between (c) chRCC and NAT samples and (d) RO and NAT samples. Venn diagrams comparing proteomes of chRCC and RO in terms of the (e) downregulated proteins and (f) upregulated proteins

The reproducibility of the MS biological replicates was verified by comparing the Pearson correlation between all pairs of samples as well as visualizing pairwise scatter plots in a matrix (Fig. 1b). The calculated Pearson correlation coefficients spanned from 0.79 to 0.97 for chRCC, from 0.76 to 0.95 for RO, and from 0.83 to 0.95 for NAT. The Pearson coefficients were higher when tumors were compared to each other (0.68–0.86) than when compared to NAT (0.56–0.70). These results underline a marked deviation of both chRCC and RO from NAT. Furthermore, the data also demonstrate that there are discernible differences between chRCC and RO, highlighting the potential to differentiate these two neoplasms using this profiling approach.

A two-tailed Student’s t test (FDR = 0.01 and S0 = 0.1) was used to distinguish the abundance of proteins between each tumor and NAT. In chRCC, 532 proteins exhibited significant differences compared to NAT, with 260 upregulated and 272 downregulated proteins (Fig. 1c). Similarly, in the case of RO versus NAT, 578 proteins showed significant differences, with 226 upregulated and 352 downregulated proteins (Fig. 1d). Comparing the upregulated and downregulated proteins between each tumor and NAT, 241 downregulated proteins (63%) and 117 upregulated proteins (32%) were found to be common (Fig. 1e, f).

To elucidate the biochemical processes driving the phenotypes of chRCC and RO, the differentially expressed proteins for each tissue type were used to interrogate comprehensive, functional proteomic databases. We found a common signature of metabolic dysregulation for chRCC and RO compared to NAT (Fig. 2a), affecting carbohydrate metabolism, lipid metabolism, amino acid metabolism and oxidative phosphorylation. Among KEGG protein pathways linked to carbohydrate metabolism, the most dysregulated were glycolysis/gluconeogenesis, pyruvate metabolism and the citric acid cycle (FDR < 8.94 × 10–14). In our analysis, we identified up to 23 proteins implicated in these pathways that exhibited dysregulation across both tumor subtypes.

Metabolic pathway proteomes similarly affected in chRCC and RO tumors. a Network representation of a subset of pathways in four areas of metabolism in which proteins were differentially regulated in tumors. Significantly (FDR < 0.01) upregulated and downregulated proteins are shown in red and blue, respectively. Protein pathway analysis was performed by searching the KEGG database [19] using differentially expressed proteins between tumor and NAT biopsies. b Correlation between levels of dysregulation of individual proteins (log2 fold change (FC) in abundance in tumor vs NAT) in chRCC and RO for proteins in carbohydrate (n = 47), lipid (n = 27) and amino acid (n = 91) metabolism and oxidative phosphorylation (n = 47). The correlation coefficient r was calculated using the Pearson test. Shading areas represent the confidence of the interval, and p represents the p value of the test

The KEGG glycolysis/gluconeogenesis pathway includes processes involved in the degradation of glucose into pyruvate and the generation of glucose from noncarbohydrates. Based on the differential protein expression data, proteins involved in gluconeogenesis, such as fructose-1,6-bisphosphatase 1 (FBP1), pyruvate carboxylase (PC), and phosphoenolpyruvate carboxykinase 1 and 2 (PCK1 and PCK2), were significantly downregulated in tumors. In contrast, glycolytic proteins remained unchanged or upregulated, e.g., ATP-dependent 6-phosphofructokinase (PFKM). These data indicate that gluconeogenesis was downregulated in tumors compared to NATs. For instance, the results of FBP1 match well with its well-known inhibitory effects on glycolysis and tumor growth [22]. Furthermore, the downregulation of PC, which catalyzes the conversion of pyruvate to oxaloacetate in the first step of gluconeogenesis, suggests a shift in cellular metabolism toward the conversion of pyruvate into acetyl-CoA, as evidenced by the observed upregulation of proteins such as pyruvate dehydrogenase E1 component subunit alpha and beta (PDHA1 and PDHB) and the dihydrolipoyl lysine-residue acetyltransferase component of the pyruvate dehydrogenase complex (DLAT). Several TCA cycle proteins were upregulated, including citrate synthase (CS), aconitate hydratase (ACO2), isocitrate dehydrogenase [NADP] and [NAD] subunits alpha and beta (IDH2, IDH3A and IDH3B), dihydrolipoyl lysine-residue succinyltransferase component of the 2-oxoglutarate dehydrogenase complex (DLST), 2-oxoglutarate dehydrogenase (OGDH) and malate dehydrogenase (MDH2). The expression of proteins active in lipid metabolism was modified in both tumors, with most being downregulated relative to NAT controls. In the chRCC and RO samples, 91 downregulated proteins were linked to amino acid (AA) metabolism.

Pearson correlations were calculated for chRCC and RO based on the fold change in abundance of differentially expressed proteins in carbohydrate, lipid and AA metabolism pathways (Fig. 2b). Positive correlations were found for the four pathways compared, with the oxidative phosphorylation pathway showing the highest p and lowest Pearson correlation values.

The functional protein pathway analysis also revealed divergent features between chRCC and RO that may underlie differences in the tumor subtypes. Proteins involved in energy metabolism of chRCC were less dysregulated than RO. For instance, in chRCC, 28 proteins involved in the oxidative phosphorylation pathway were expressed at different levels compared to NAT, while in RO, 40 proteins belonging to this pathway were dysregulated (Fig. 2a). Comparing the variation between the proteins belonging to the oxidative phosphorylation pathway in both tumors versus NAT, a Pearson coefficient of 0.79 was obtained. The same comparison for carbohydrate, lipid, and AA metabolism resulted in Pearson coefficients of 0.94, 0.92 and 0.94, respectively (Fig. 2b).

Mitochondrial KEGG pathways were dysregulated more extensively in RO biopsies, with 256 differentially expressed proteins (Fig. 3a) compared to 175 in chRCC (Additional file 3: Table S2).

Differences in protein pathways underlying each tumor subtype. a Specific dysregulation of mitochondrial protein pathways in ROs, including those of respiratory chain complexes. Protein pathway analysis was performed by using the differentially expressed proteins in each tumor biopsy (relative to NAT) in searches against KEGG [19] and GO [21] databases. Significantly (FDR < 0.01) upregulated and downregulated proteins are shown in red and blue, respectively. b Comparison of the protein deregulation (log2 fold change (FC) in abundance in tumor relative to NAT) between chRCC and RO for the five respiratory chain complexes [Complex I, n = 44; Complex II, n = 4; Complex III, n = 10; Complex IV, n = 19; Complex V, n = 20]. Blue dots, downregulated proteins (RO vs NAT); red dots, upregulated proteins (RO vs NAT). The correlation coefficient r was calculated using the Pearson test. Shading areas represent the confidence of the interval, and p represents the p value of the test. c Absolute protein amount, expressed as pmol/mg of tissue, calculated through the TPA methodology

The common chRCC versus NAT and RO versus NAT dysregulations of proteins related to respiratory chain complexes were evaluated and compared. The results depicted in Fig. 3b present Pearson correlation coefficients of 0.81, 0.79, 0.93, 0.57, and 0.40 for complexes I, II, III, IV and V, respectively. The TPA-based concentration values for the respiratory chain complex proteins are presented in Fig. 3c, Additional file 4: Figs. S1, 2.