Cellular and secretome profiling uncover immunological biomarkers in the prognosis of renal cell carcinoma patients

Abstract

Renal cell carcinoma (RCC) is recognized as an immunogenic tumor, yet tumor-infiltrating lymphocytes often exhibit diminished effector function. However, the mechanisms underlying reduced T and NK cell activity in RCC remain unclear. Here, we examined the immune contexture in RCC patients undergoing nephrectomy to identify immune-related biomarkers associated with disease progression. Immune cell phenotypes and secretion profiles were assessed using flow cytometry and Luminex multiplex analysis. Supervised multivariate analysis revealed several changes of which frequencies of T and NK cells expressing CCR5, CXCR3, and PD-1 were elevated within tumors compared with peripheral blood. In addition, higher levels of regulatory T cells, PD-1+, and CXCR3+ T and NK cells were observed in patients with relapse following nephrectomy. With regards to soluble factors, tumor-derived CXCL8 was associated with higher Fuhrman grade and increased frequency of polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs). These biomarkers demonstrate potential relevance in the progression of RCC and merit further investigation in prospective studies.

Keywords: Biomarkers; CXCL8; NK cells; PD-1; T cells; prognosis; renal cell carcinoma.

Figure 1.

Immune cell phenotypes and secretion profiles in peripheral blood and tumors of RCC patients. (a) Experimental and analysis workflow of sample collection and analysis. (b) Semi-supervised hierarchical clustering of immune cell and soluble factor variables in tumor and pre-surgery blood. Compositions of three subgroups are summarized in pie charts. Three distinct subgroups were separated based on differences in the expression of chemokines and cytokines (red), chemokine receptors and Ki67 (blue), and DNAM1 and the chemokine receptors (green). Details of specific cell phenotypes and secretion profiles are specified in Table S4A-C. (c) OPLS-ep analysis of differences between paired tumor and blood variables for each patient at the time of surgery. Immune cell and soluble variables enriched or reduced in tumors compared with blood. Pie charts highlighting factors enriched and reduced in tumors compared with peripheral blood.

Figure 2.

Immune secretome is associated with tumor stage and Fuhrman grade in RCC patients. (a) Unsupervised hierarchical heatmap showing normalized z scores of all soluble analytes detected in pre-surgery blood, post-surgery blood, and tumor lysates in relation to clinical and patient parameters. Radar chart (snail plot) showing concentrations of chemokines (pg/mL) detected in tumor lysates and differentiated by tumor stage (b) and tumor grade (c).

Figure 3.

CXCL8 correlates with high tumor grade and increased PMN-MDSCs. (a) LEFT: OPSL analysis separating RCC patients according to Fuhrman grade (x-axis, predictive variation; ellipse, Hotelling’s T2 95% confidence region). Fuhrman grade for the patient with chromophobe histology is not available (NA). RIGHT: the 16 significant predictive loadings correlating with Fuhrman grade. (b) Frequency of CXCR3+CD56bright NK cells, and (c) CXCL8 concentration in patients with Fuhrman grade 2 vs. 3 + 4. Mean values ± SD are shown. (d) Correlation of CXCL8 concentration and MDSC in peripheral blood.

Figure 4.

CXCL8 and IFNβ levels correlate with disease relapse in RCC patients undergoing nephrectomy. (a) Significant changes (p < 0.05) of cellular factors displayed as fold change of mean values in pre-surgery blood samples between relapse and non-relapse patients. (b) CXCL8 and IFNβ levels in relapse and non-relapse patients. (c) CXCL8 expression across Fuhrman grade and primary tumor stage in TCGA-KIRC cohort. (d) Overall survival and disease specific survival in RCC patients stratified by high and low CXCL8 expression (set as 50%). p values were calculated by log-rank test. (e) CXCL8 expression in patients without (no) and with (yes) a new tumor event within the TCGA-KIRC cohort. (f) Ratio of CXCL8/IFNβ in relapse vs. non-relapse patients.

Figure 5.

Cellular and soluble factors in RCC patients. (a) Difference of immune phenotypes and secretome between blood and tumor. (b) Correlation of tumor and peripheral blood derived variables with tumor stage and Fuhrman grade. (c) Difference in pre- vs. post-surgery blood. (d) Factors correlating with RCC relapse.