Predicting diagnostic gene expression profiles associated with immune infiltration in patients with lupus nephritis

Abstract

ObjectiveTo identify potential diagnostic markers of lupus nephritis (LN) based on bioinformatics and machine learning and to explore the significance of immune cell infiltration in this pathology.MethodsSeven LN gene expression datasets were downloaded from the GEO database, and the larger sample size was used as the training group to obtain differential genes (DEGs) between LN and healthy controls, and to perform gene function, disease ontology (DO), and gene set enrichment analyses (GSEA). Two machine learning algorithms, least absolute shrinkage and selection operator (LASSO) and support vector machine-recursive feature elimination (SVM-RFE), were applied to identify candidate biomarkers. The diagnostic value of LN diagnostic gene biomarkers was further evaluated in the area under the ROC curve observed in the validation dataset. CIBERSORT was used to analyze 22 immune cell fractions from LN patients and to analyze their correlation with diagnostic markers.ResultsThirty and twenty-one DEGs were screened in kidney tissue and peripheral blood, respectively. Both of which covered macrophages and interferons. The disease enrichment analysis of DEGs in kidney tissues showed that they were mainly involved in immune and renal diseases, and in peripheral blood it was mainly enriched in cardiovascular system, bone marrow, and oral cavity. The machine learning algorithm combined with external dataset validation revealed that C1QA(AUC = 0.741), C1QB(AUC = 0.758), MX1(AUC = 0.865), RORC(AUC = 0.911), CD177(AUC = 0.855), DEFA4(AUC= 0.843)and HERC5(AUC = 0.880) had high diagnostic value and could be used as diagnostic biomarkers of LN. Compared to controls, pathways such as cell adhesion molecule cam, and systemic lupus erythematosus were activated in kidney tissues; cell cycle, cytoplasmic DNA sensing pathways, NOD-like receptor signaling pathways, proteasome, and RIG-1-like receptors were activated in peripheral blood. Immune cell infiltration analysis showed that diagnostic markers in kidney tissue were associated with T cells CD8 and Dendritic cells resting, and in blood were associated with T cells CD4 memory resting, suggesting that CD4 T cells, CD8 T cells and dendritic cells are closely related to the development and progression of LN.ConclusionC1QA, C1QB, MX1, RORC, CD177, DEFA4 and HERC5 could be used as new candidate molecular markers for LN. It may provide new insights into the diagnosis and molecular treatment of LN in the future.

An interpretable machine learning pipeline based on transcriptomics predicts phenotypes of lupus patients

“…However, no other accuracy metrics were reported. 3 In contrast, our models developed using expression of genes grouped by disease-relevant cell types or functions achieved higher overall accuracies (with AUCs ranging from 0.842 to 0.989) and presented interpretable results that could be translated to improved understanding of the molecular systems involved in SLE pathology.…”

“… 1 , 2 Others have attempted to create machine learning (ML) models to predict phenotype by selecting individual genes to use as features, rather than considering how those genes fit into the broader context of disease pathogenesis. 3 , 4 , 5 In this article, we aim to establish an ML pipeline to identify and understand the specific molecular pathways implicated in phenotypic subsets of patients by using a systems biology lens.…”

“…However, many of the previous studies have not reported all the accuracy metrics necessary for validation or have employed small datasets from a single center and, thus, may be over-fitted. 3 , 4 , 18 As a result, no ML algorithms based on blood gene expression have been adapted in the clinic to aid in the diagnosis of SLE or the recognition of lupus phenotypes.…”

Systemic lupus in the era of machine learning medicine

“…Reports using genomic and genetic expression datasets identified several important biomarker for LN including C1QA, C1QB, MX1, RORC, CD177, DEFA4, and HERC5 for LN. 118 For non-renal SLE, FOXP3, 88 MX2, 106 HLA-DQA1, 90 HLA-DQB1, 90 HLA-DRB1, 90 neutrophil extracellular trap-related genes (HMGB1, ITGB2 and CREB5), 70 ABCB1, 120 IFI27 120 and PLSCR1 120 have been reported. Other types of biomarkers included proteomics (IFIT3, MX1, TOMM40, STAT1, STAT2 and OAS3), 101 metabolomics, 91 lipidomics 94 and microRNA profiles.…”

Decipher the Immunopathological Mechanisms and Set Up Potential Therapeutic Strategies for Patients with Lupus Nephritis

“…Wang et al [ 127 ] analyzed gene expression relevant to the macrophages and interferons in kidney tissues and the peripheral blood of LN patients. Their results showed that many pathways—including cell cycle, cytoplasmic DNA sensing, NOD-like receptor signaling, proteasome, and RIG-1 like receptors (RLRs)—were activated in peripheral blood.…”