Biological Characteristics and Predictive Model of Biopsy-Proven Acute Rejection (BPAR) After Kidney Transplantation: Evidences of Multi-Omics Analysis

Objectives: Early diagnosis and detection of acute rejection following kidney transplantation are of great significance for guiding the treatment and improving the prognosis of renal transplant recipients. In this study, we are aimed to explore the biological characteristics of biopsy-proven acute rejection (BPAR) and establish a predictive model. Methods: Gene expression matrix of the renal allograft samples in the GEO database were screened and included, using Limma R package to identify differentially expressed transcripts between BPAR and No-BPAR groups. Then a predictive model of BPAR was established based on logistic regression of which key transcripts involved in the predictive model were further explored using functional enrichment analyses including Gene Ontology analysis (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, and Gene Set Enrichment Analysis (GSEA). Results: A total of four studies (GSE129166, GSE48581, GSE36059, and GSE98320) were included for extensive analysis of differential expression. 32 differential expressed transcripts were observed to be significant between two groups after the pooled analysis. Afterward, a predictive model containing the five most significant transcripts (IDO1, CXCL10, IFNG, GBP1, PMAIP1) showed good predictive efficacy for BPAR after kidney transplantation (AUC = 0.919, 95%CI = 0.902–0.939). Results of functional enrichment analysis showed that The functions of differential genes are mainly manifested in chemokine receptor binding, chemokine activity, G protein-coupled receptor binding, etc. while the immune infiltration analysis indicated that immune cells mainly related to acute rejection include Macrophages. M1, T cells gamma delta, T cells CD4 memory activated, eosinophils, etc. Conclusion: We have identified a total of 32 differential expressed transcripts and based on that, a predictive model with five significant transcripts was established, which was suggested as a highly recommended tool for the prediction of BPAR after kidney transplantation. However, an extensive study should be performed for the evaluation of the predictive model and mechanism involved.