OA25 Transcriptome analysis of peripheral blood and renal tissue identifies genes that predict treatment response in proliferative lupus nephritis: a prospective observational study

• Published in: Rheumatology (Oxford, England) Vol. 63; no. Supplement_1
• Main Authors: Bulusu, Sree Nethra, Bavikatte, Anitha, Shah, Sanket, Murthy, Sneha, Mariaselvam, Christina M, Kavadichanda, Chengappa, Vembar, Shruthi, Thabah, Molly, Negi, Vir S
• Format: Journal Article
• Language: English
• Published: 24-04-2024
• ISSN: 1462-0324; 1462-0332
• DOI: 10.1093/rheumatology/keae163.025

Abstract
Background/Aims
Mechanisms contributing to non-response in lupus nephritis (LN) are not clear. Peripheral blood mononuclear cells (PBMCs) serve as easily available less invasive alternative to kidney biopsy in monitoring treatment response. However, it is unclear to what extent peripheral blood mirrors the inflamed kidney. This study aims to identify changes in functional pathways from PBMCs and renal tissues after cyclophosphamide treatment and to identify baseline genes that predict treatment response by RNA-sequencing.
Methods
RNA from PBMCs and renal biopsies was extracted from 16 treatment-naïve proliferative LN at baseline, 6-months and at the time of flare within 2 years of follow-up. Patients were classified as clinical-responders (CR) (n = 4) or non-responders (NR) (n = 8) at 6-months using EULAR/EDTA criteria. CR who developed renal flare on follow-up were considered FL (n = 4). RNA was sequenced on Illumina Novaseq-6000 platform. Healthy control datasets were taken from public databases. Differentially expressed genes (DEGs) and enriched pathways were analysed using Reactome database-v74. Comparisons for each group were performed before vs after treatment. Weighted gene co-expression network analysis (WGCNA) was used to identify baseline hub genes that predict treatment-response and ROC analysis was used to test their prognostic utility.
Results
Functional enrichment analysis identified neutrophil degranulation, autophagy, regulation of NF-κβ signalling, FCGR-dependent phagocytosis and TNF signalling upregulated and IFN-γ signalling and ECM molecules downregulated in PBMCs of CR. ECM molecules and proteoglycans, neddylation and BCR signalling were enriched in PBMCs of NR. In NR renal tissue, apoptosis, NF-κβ signalling, IKK complex formation upon TLR signalling and MAPK signalling were enriched. In FL PBMCs, neutrophil degranulation, ROS and RNS production in phagocytes, antimicrobial peptide and immune system pathways were downregulated. In FL renal tissue, glycogen metabolism, viral mRNA translation, cell-ECM interactions, cell-cell communications, apoptosis and ISG15 antiviral mechanism were downregulated. Clinical traits for WGCNA were BCR, BNR, BFL and HC. The most correlated modules and hub genes from baseline PBMC were: green (BCR- IFI27, CD24, AZU1, MT2A, SNRNP70), brown (BNR- AC090809.1, TENM2, NLGN1, LSAMP, DLGAP2, AP005230.1, TMEM132D, AL445255.1, LINC00536, AC013287.1, BX470102.1, TM4SF20) and blue module (BFL- AC092436.3, AC046158.3, LINC01147). From baseline renal tissues, the most correlated modules and genes were: green (BCR- AP003068.1, BX470102.1, AC009088.3, AC136475.1, AC073611.1, AP000757.2) and turquoise module (BNR & BFL: TRDN-AS1, AC107973.1, AL672167.1, AC020637.1, AL137100.3, AC018437.3, LINC01982, TPH2, AC136759). From PBMCs, ROC analysis showed that combinational model of TENM2, NLGN1 and AP005230.1 predicted NR (AUC-0.94; 95% CI: 0.80-1; p = 0.02) and AC092436.3 predicted FL (AUC-0.81; 95% CI: 0.53-1; p = 0.036). AP005230.1 from renal tissue also predicted NR (AUC-0.94; 95% CI: 0.82-1; p = 0.01).
Conclusion
Type I IFN and fibrosis pathways were enriched in NR and downregulated in CR and FL. Baseline PBMC markers TENM2, NLGN1 & AP005230.1 predicted NR and FL. AP005230.1 from PBMCs & renal tissue predicted NR.
Disclosure
S. Bulusu: None. A. Bavikatte: None. S. Shah: None. S. Murthy: None. C.M. Mariaselvam: None. C. Kavadichanda: None. S. Vembar: None. M. Thabah: None. V.S. Negi: None.