3.1 Construction and optimization of prognostic models

The univariate Cox regression analysis showed that 163 genes were strongly correlated with the prognosis of patients with renal clear cell carcinoma. In the LASSO analysis, the corresponding coefficients of the identified genes decreased to 0 as logλ changed (Fig. 1A). In the cross-validation, 28 genes reached the minimum value of partial fit deviation, and 28 genes showed some effect with HRs greater than 1, suggesting a positive effect on the development of bladder cancer (Fig. 1B). Multifactorial Cox regression analysis of the training set data yielded 21 genes with GRPs as independent predictors, namely CASR, ADGRV1, P2RY8, QRFPR, F2RL3, LGR4, scTR, GPRC5C, ADORA2B, OPN1sw, F2RL1, GABBR1, XCR1, ADGRG1 OR2B6, ADGRF5, PTGDR2, EDNRB, (Fig. 1C). The training group samples were scored, their GRPs correlation scores were calculated, and the median score was used as the boundary. Those with scores more outstanding than the median were the high GRPs group, and those with scores lower than the median were the low GRPs group, and the predictive analysis showed that the high GRPs group had a worse prognosis than the low GRPs group (Fig. 1D). Using the survfit function to assess those immune cell types associated with prognosis, the results showed that naïve B cells, monocytes, and activated dendritic cells were associated with prognosis. The CIBERSORT method to calculate immune infiltration in TCGA samples was bounded by the median score. Those with scores more remarkable than the medians were the TMEs high subgroup, and those below the score were the TMEs low subgroup, and the prognosis The results of the analysis showed that the TMEs high subgroup had a better prognosis than the TMEs low subgroup (Fig. 1E, F).

Fig. 1

LASSO regression model and results of multi-variate regression analysis. A LASSO coefficient distribution of 163 differentially expressed genes associated with prognosis of renal clear cell carcinoma, each curve represents a coefficient, when the tuning parameter changes, non-zero coefficients change with it into the LASSO regression model. B cross-validation of selecting the best λ, the red dashed line corresponds to the minimum of the multivariate Cox model when crossed with the best log λ. The two dashed lines indicate the distance minimum one standard deviation. C Results of multivariate Cox regression analysis based on the training set. D K–M curves based on GPR scores. E K–M curves based on TME scores. F Forest plots based on immune cells

3.2 GSEA enrichment analysis

We further showed the correlation between the 21 GRPs genes by Pearson correlation analysis and plotted the correlations (Fig. 2A). To test the reliability of GRPs and TMEs score in the prognosis of patients with renal clear cell carcinoma, we divided the patients into four gene groups according to the GRPs and TMEs gene expression levels in cancer patients, GRPs high + TMEs high, GRPs high + TMEs low, GRPs low + TMEs high, and GRPs low + TMEs low groups, and the predictive analysis showed that the GRPshigh−TMElow group had the worst prognosis and GRPslow−TMEhigh group had the best prognosis (Fig. 2B), and the AUCs based on the prediction model of GRPs and TMEs gene expression levels at 3, 5 and 7 years were 0.694, 0.761, and 0.724, respectively, all of which were greater than 0.6, indicating an excellent predictive efficacy (Fig. 2C). Differential analysis of GRPs scores and TME scores for high and low groups was performed using the limma package. Enrichment analysis of differential genes was performed. GSEA enrichment analysis showed that the significant differential downregulation pathways for high and low-risk groups were: Chemokine Signaling Pathway, Natural Killer Cell-Mediated Cytotoxicity, Tight Junction, Adherens Junction, Ppar Signaling Pathway, and Peroxisome. The upregulated pathways were the Renal Cell Carcinoma Erbb Signaling Pathway (Fig. 2D, E).

Fig. 2

Pathway enrichment analysis. A Correlation between 21 GRPs genes. B Prognostic analysis of different GRPs and TMEs gene groupings.C Prediction model based on GRPs and TMEs gene expression levels (D, E) regulation pathways in high- and low-risk groups

3.3 Differentially expressed genes related to overall survival status of patients with clear cell renal cell carcinoma

Gene mapping and clinical characterization data were extracted from 607 samples downloaded from XENA-TCGA-KIRC, including 535 patients and 72 control samples, with a soft threshold of 13. 12 color-coded modules were generated based on topological overlap matrix gene clustering with a dynamic tree cut of module size 30 (Fig. 3A). Overall survival status was correlated with all 7 modules, with turquoise and green modules showing the highest correlation with renal clear cell carcinoma (Fig. 3B). GO enrichment analysis revealed GPR_low + TME_High: enriched in metanephric epithelium development, the inhibitory postsynaptic potential, sulfur compound catabolic process, gland development, and actomyosin structure organization. Mixed: enriched in response to molecule of bacterial origin, histone ubiquitination, interferon alpha production, regulation of sysTMEic arterial blood pressure by hormone. gpr _high + TME_low: enriched in mesenchymal cell differentiation, response to BMP, endothelial cell migration, cardiac muscle cell action potential, and negative regulation of smooth muscle cell proliferation. corPvalueStudent function was used to determine the correlation type between the two phenotype scores and the modules. A correlation heat map was drawn to select the modules correlated with the two phenotype scores (Fig. 3C). The correlation heat map showed that CD8 T, NK, Th1 cell recruiting, and Macrophage recruiting expressions were relatively high in the GPR_high + TME_low group. In the GPR_low + TME_high group, The expression of CD4 T cell recruiting, Th1 cell recruiting, and Macrophage recruiting was relatively high in the Mixed group recruiting expression was relatively high (Fig. 3D).

Fig. 3

Prognosis-related differentially expressed genes in renal clear cell carcinoma. A Gene clustering based on topological overlap matrix, with relatively related genes located on the same or adjacent branches. B Correlation of module eigengenes with renal clear cell carcinoma. C GO enrichment analysis, D different Immune step correlated with subgroups

3.4 Prognosis of renal clear cell carcinoma

The prediction model results showed that the combination of GRPs low grouping and TME high grouping (as the best prognosis type) and the intermediate type defined as mixed. The prognosis of GRPs low grouping + TME high grouping was better, the prognosis of GRPs high grouping + TME low grouping was the worst, and the prognosis of the mixed group was intermediate (Fig. 4A). The results of forest plot analysis showed a correlation between GRPs and TME scores on the prognosis of renal clear cell carcinoma (Fig. 4B). We were repeated in the GSE167573 validation group, and the results were consistent with the verify group (Fig. 4C). Based on different clinical features, the KM curve showed significant difference in prognosis between high-risk group and low-risk group (Fig. 5).

Fig. 4

Correlation analysis of subgroup and prognostic outcome. A K–M prognostic model for different subgroups of renal clear cell carcinoma. B, C Regression analysis of age, sex, and disease of renal clear cell carcinoma based on multivariate Cox analysis

Fig. 5

Prognosis of patients with renal clear cell carcinoma based on age, sex, and disease staging classification was explored

3.5 Single-cell RNA sequencing analysis

ScRNA.seq was applied to analyze the cellular differences between cancer tissues and normal regional tissues in 2 patients with renal clear cell carcinoma. The single-cell sequencing data were subjected to data normalization, dimensionality reduction analysis, and cluster analysis, and then all cells were divided into 21 cell clusters (Fig. 6A). It included 7 immune cells clusters such as myeloid, endo, Masticells, Nkcells, CD8-T, Tregs, and Stromal (Fig. 6B). The GPRs scores of various immune cell clusters are also shown (Fig. 6C). Cell chat is a database containing receptor-ligand interactions with 2021 validated molecular interactions. cellChat can identify critical features of intercellular communication in a given scRNA-seq dataset and predict potential signaling pathways that are currently studied Differential analysis of intercellular communication indicates that myeloid, endo, Masticells, Nkcells, CD8-T, Tregs, Stromal, and other cell populations interact closely (Fig. 6D). Furthermore, PROGENy scores showed that Masticells, Nkcells, CD8-T, Tregs were positively correlated with activating “NFkB, TNFa, JAK-STAT, WNT, EGFR, MAPK, VEGF, PI3K, Estrogen Trail”. On the contrary, myeloids, endo, and Stromal cells were positively correlated with “TGFb, Hypoxia, Androgen, p53 activation” (Fig. 6E). And to assess the effect of immunotherapy in the training set (Fig. 6F).

Fig. 6

Analysis of single-cell RNA sequencing data for renal clear cell carcinoma. A, B UMAP was clustered into 21 clusters and divided into 7 cell subgroups. C GPR scores of different cell subgroups. D Number of intercellular interactions plotted. E PROGENy scores. F Evaluation of immunotherapy effect

3.6 Mutation analysis

Grouped according to GRPs and TME gene scores and presented vital differential genes between groups in box plots (Fig. 7A, B)—the differential expression of critical genes in tumor tissues and controls and predictive analysis. In addition, we constructed immune response True and False groups according to immune response gap grouping, in which the proportion of the True group was higher than the Mixed group in the GRPs low group + TME high group. The proportion of True in the Mixed group was higher than GRPs high group + TME low group (Fig. 7C). the GPRs score and TMEa score of the True group were higher than False group (Fig. 7D, E). The fragmentation plot shows the proportion of metabolic pathways among the four subgroups, where the functions and pathways of genes associated with high GRPs + high TMEs mainly included peptidases, NF-kappa B signaling pathway, Cytokine–cytokine receptor interaction, G protein-Functions and pathways of genes associated with GRPs high + TMEs low MAPK signaling pathway, Pl3K-Akt signaling pathway, Transcription., Functions and pathways of genes associated with GRPs low + TMEs high Ubiquitin labeling, MAPK signaling pathway, Calcium signaling pathway. Amino acid metabolism, PPAR signaling pathway, Carbohydrate digestion, absorption digestion, and absorption (Fig. 7F, G).

Fig. 7

Immunotherapy effect and pathway analysis. A, B Each immunoregulator was significantly different between disease and control groups C Proportion of immune response between different renal clear cell carcinoma subgroups. D, E GRPs/TME scores of high and low immune response disparity groups. F, G Fragmentation plots of genes associated with immune disparity and prognostic disparity among different renal clear cell carcinoma subgroups

3.6.1 XCR1 genes knockdown inhibits proliferation, migration, and EMT of renal cancer cells

siRNAs were transfected into kidney cancer 786-O and CAKI-1 cell lines to explore the detailed role of XCR1 genes in oncogenicity. The efficiency of the knockdown of the XCR1 genes was confirmed by RT-qPCR results shown, in which the mRNA expression level of the XCR1 genes was reduced after transfection of specific siRNA into kidney cancer 786-O and CAKI-1 cells (Fig. 8A). These data indicate the high specificity and transfection efficiency of siRNA. In addition, when the XCR1 genes were silenced intracellularly, EDU-positive cells were also reduced, which further confirmed that the knockdown of XCR1 genes inhibited cell proliferation (Fig. 8B, D). In addition, cell migration was detected by Transwell assay. The results showed that the deletion of the XCR1 genes hindered the migration ability of cancer cells, as the number of migrating cells was reduced in the transfected group (Fig. 8E, F). Meanwhile, a lack of XCR1 genes inhibited EMT in kidney cancer, as evidenced by decreased protein levels of N-calmodulin and Vimentin and increased E-calmodulin (Fig. 8G, I). Subsequently, clone formation experiments showed that the number of clones was significantly reduced after the XCR1 genes were knocked down in the cells (Fig. 8J, K). The results above suggest that XCR1 genes downregulation inhibited renal cancer cell proliferation, migration, and EMT.

Fig. 8

XCR1 genes knockdown inhibits proliferation, migration and EMT of kidney cancer cells. A Reduced mRNA expression levels of XCR1 genes after specific siRNA transfection into renal cancer 786-O and CAKI-1 cells. B–D Knockdown of the XCR1 genes inhibited cell proliferation. E, F The deletion of the XCR1 genes hindered the migratory ability of renal carcinoma cells, G–I decreased protein levels of N-calmodulin and Vimentin, increased protein levels of E-calmodulin. J, K The number of clones was significantly reduced after the knockdown of XCR1 genes in the cells