Enrichment Analysis Of Genes In Patients With Kirc
We obtained 1465 DEGs, including 666 upregulated genes and 799 downregulated genes (Fig. 1a). We performed KEGG and GO analyses of DEGs. KEGG pathway analyses indicated that DEGs were mainly enriched in the B cell, T cell receptor signaling pathway, chemokine signaling pathway and Th1 cell differentiation, Th2 cell differentiation, Th17 cell differentiation, osteoclast differentiation and graft versus host disease, staphylococcus aureus infection, cytokine receptor interaction, natural killer cell mediated cytotoxicity, etc(Fig. 2a). GO analysis showed that DEGs were mainly enriched in the Leukocyte differentiation, peptidyl-tyrosine phosphorylation, endocytic vesicle membrane, cellular response to interferon-gamma, natural killer cell mediated cytotoxicity, positive regulation of innate immune response, regulation of calcium ion transport, etc(Fig. 2b).
Construction Of The Necroptosis-related Lncrna Predictive Signature
We analyzed the necroptosis-related lncRNA expression level in 541 KIRC samples and 72 normal samples from the TCGA. Then, 322 necroptosis-related lncRNAs were identified in the DE lncRNAs by the Pearson correlation algorithm. Univariate Cox regression analysis revealed that 28 lncRNAs were associated with the prognosis of KIRC patients(Fig. 3a). Then, the 28 lncRNAs were acquired by LASSO analysis(Fig. 3b, Fig. 3c). 18 of which were brought in the multi-Cox proportional hazards model(Fig. 3d). Multivariate Cox regression analysis revealed that 8 necroptosis-related lncRNAs (LINC00565, AL731567.1, PRKAR1B-AS1, PROX1-AS1, C3orf36, LINC02446, AL355377.4, LINC01738) were identified to construct a predictive signature. The risk score was calculated as follows: risk score=(0.10825×LINC00565 expression)+(0.08168×AL731567.1expression)+(0.29317×PRKAR1B-AS1expression)+(0.15311 ×PROX1-AS1expression)+(-0.16812×C3orf36expression)+(0.07065×LINC02446expression)+(0.13588 × AL355377.4 expression)+(-0.06933 × LINC01738 expression).
Correlation Between The Predictive Signature And The Prognosis Of Kirc Patients
The risk score of each patient was calculated according to the formula, and the patients were divided into high risk and low risk groups according to the median value of the risk score. To determine the value of the risk score in predicting the prognosis of KIRC patients, Kaplan-Meier analysis was used to analyze the OS time of the high and low risk groups. Compared with that of the low risk group, the OS time of the high risk group was significantly shorter (Fig. 4a, p = 2e-16). The 5-years survival rates of the high risk and low risk groups were 48.5 and 80.5%, respectively. The risk scores of the high and low risk groups are shown in Fig. 4B. With the increase of risk score, more and more patients died (Fig. 4b). The expression levels of 8 necroptosis-related lncRNAs of prediction model are shown in (Fig. 4c). To determine whether the predictive signature is an independent prognostic factor for KIRC patients, Cox regression analysis was performed. Univariate Cox regression analysis showed that age, T stage and risk score were significantly associated with the OS of KIRC patients (Fig. 4d). Multivariate Cox regression analysis showed that age, T stage and risk score were independent predictors of OS in KIRC patients (Fig. 4e). The AUC of the risk score was 0.725, which was better than that of clinicopathological variables in predicting the prognosis of KIRC patients (Fig. 4f). The AUCs of 1, 2, and 3-years survival were 0.7, 0.69 and 0.691, respectively, which indicated good predictive performance(Fig. 4g).
Construction Of Nomogram
Based on risk score, age, and clinicopathological factors, a nomogram was developed for predicting the 1, 2, 3- years OS incidences (Fig. 5a). The calibration plots were applied to testify that the good consistency with the actual observation(Fig. 5b-d).
Relationship Between The Predictive Signature And The Prognosis Of Kirc Patients In Different Clinicopathological Variables
To study the relationship between the predictive signature and the prognosis of KIRC patients sorted according to different clinicopathological variables, KIRC patients were separated into groups according to age, gender, stage, T stage and N stage. For each different classifications, the OS of patients in the high-risk group was significantly shorter than that of patients in the low-risk group (Fig. 6a-i). These results suggest that the predictive signature can predict the prognosis of KIRC patients without considering clinicopathological variables.
Internal Validation Of The Predictive Signature
To verify the applicability of the predictive signature for OS based on the entire TCGA dataset, we randomly divided the 613 KIRC patients into two cohorts. Consistent with the results observed in the entire dataset, the OS rate of patients in the high-risk group was lower than that of the low-risk group in the first internal cohort (Fig. 7a, p = 9.502e-06). In the second internal cohort, the prognosis of the high-risk group was worse than that of the low-risk group (Fig. 7c, p = 4.443e-06). The ROC curves of two cohorts showed good predictive performance. In the first internal cohort, the AUCs of 1, 2, and 3-years survival were 0.7, 0.65, and 0.669, respectively (Fig. 7b). In the second internal cohort, the AUCs of 1, 2, and 3-years survival were 0.8, 0.73, and 0.708, respectively (Fig. 7d)
Gene Enrichment Analysis
Because of the different prognoses of patients in the high and low risk groups, we conducted KEGG and GO analyses to study the possible differences between the high and low risk groups. KEGG analysis found that parathyroid hormones synthesis secretion and action, insulin secretion, synaptic vesicle cycle were significantly enriched in the high-risk group (Fig. 8a). GO analysis showed that high-risk group were mainly enriched in High-density lipoprotein particle, serotonin receptor signaling pathway, cyclic nucleotide metabolic process, negative regulation of reproductive process, etc(Fig. 8b).
Immune Cell Infiltration Analyses
To further explore the correlation between risk scores and immune cells, we predicted composition of immune cells in gene expression profiles (GEPs) using CIBERSORT. The results showed that B cells naive, plasma cells, CD8+T cells, regulatory T cells, NK cells resting, monocytes, M0 macrophages, M2 macrophages, Mast cells resting were significantly different in the high and low risk groups (Fig. 9a). In addition, the relationship between 22 types of common infiltrated immune cells and 8 prognostic lncRNAs was highlighted through a correlation heatmap. LINC02446 displayed a distinct negative correlation with CD8+T cells and M1 macrophages but positively correlated with monocytes and Dendritic cells activated, Mast cells resting(Fig. 9b).