Hsa_circ_0000479 as a Novel Diagnostic Biomarker of Systemic Lupus Erythematosus

Abstract

Background: Accumulating evidence suggests that differentially expressed non-coding circular RNAs (circRNAs) play critical roles in the progress of autoimmune diseases. However, the role of circRNAs in systemic lupus erythematosus (SLE) remains unclear. Methods: We initially used next-generation sequencing (NGS) to comprehensively analyze circRNA expression profiles in peripheral blood mononuclear cells (PBMCs) from 10 SLE patients, stratified by their disease activity characteristics (stable or active SLE), and 10 healthy controls (HCs). Candidate circRNAs identified were first validated by quantitative reverse-transcription (qRT)-PCR in PBMC samples from a training-phase cohort of five SLE patients and five HCs. The significantly dysregulated circRNAs were then confirmed by qRT-PCR in a validation cohort of 23 SLE patients and 21 HCs, and in an external validation cohort with 64 SLE patients, 58 HCs, and 50 patients with rheumatoid arthritis (RA). In addition, we conducted bioinformatics analysis and western blotting investigating the relationships between the candidate circRNAs and SLE progression. Results: Multilayer integrative analysis of circRNA regulation showed that 84 circRNAs were upregulated and 30 were downregulated in patients with SLE compared with HCs. We then analyzed the intersection of these differentially expressed circRNAs in an SLE-stable cohort, an SLE-active cohort, and HCs. This enabled us to narrow down dysregulated circRNAs to 15 upregulated circRNAs. Only hsa_circ_0000479 was significantly upregulated in PBMCs of patients with SLE compared with HCs (P < 0.05). Furthermore, the diagnostic potential of hsa_circ_0000479 expression to distinguish SLE patients from HCs and RA patients was also significantly increased in the validation-phase and external-validation-phase cohorts (P < 0.05). When distinguishing SLE patients from HCs, the diagnostic specificities of hsa_circ_0000479 were 0.619 and 1.0 in two validation cohorts, respectively (AUCs = 0.731 and 0.730, respectively). It was also significantly increased in either stable SLE patients or active SLE patients compared with HCs in these two cohorts (P < 0.05). We also used bioinformatics analysis to show that hsa_circ_0000479 regulates SLE progression by modulating metabolic pathways and the Wnt signaling pathway. Western blotting revealed that the expression of Wnt-16 protein was significantly decreased in SLE. Conclusion: Our results suggest that hsa_circ_0000479 has potential as a novel biomarker for the diagnosis of SLE.

Keywords: RNA-sequencing; biomarker; circular RNAs (circRNAs); hsa_circ_0000479; systemic lupus erythematosus.

Figure 1

Sequencing data on differential circRNA expression profiles in PBMCs from SLE patients and HCs in the discovery-phase cohort. (A) The percentage of significantly differentially expressed circRNAs arising from different genomic regions (exon, intron, and intergenic regions). (B) Volcano plot of differentially expressed circRNAs. The blue spots indicate significantly downregulated circRNAs, and the red spots indicate significantly upregulated circRNAs. (C,D) The overlapping significantly differentially expressed circRNAs in PBMCs of SLE patients vs. HCs. There were 84 significantly upregulated (C) and 30 downregulated (D) circRNAs in PBMCs of SLE patients vs. HCs (blue area). There were 30 significantly upregulated (C) and 19 downregulated (D) circRNAs in PBMCs of the SLE-stable group vs. HCs (red area). There were 32 significantly upregulated (C) and 15 downregulated (D) circRNAs in PBMCs of the SLE-active group vs. HCs (green area). Integrating these three comparisons, we found 15 overlapping significantly upregulated circRNAs in PBMCs of SLE patients vs. HCs. These 15 significantly altered circRNAs are detailed in Supplementary Table 4. (E) Hierarchical clustering of the differentially expressed circRNAs. Red represents relatively highly expressed circRNAs, and green represents relatively lowly expressed circRNAs.

Figure 2

The mapping network of circRNA-miRNA interactions in SLE. (A) The network map includes the 12 surviving candidates from the 15 significantly altered circRNAs (represented as red nodes) in the analysis for circRNA-miRNA network prediction. The other three differentially expressed circRNAs did not show reliable results in this interaction analysis. The blue nodes around the red node were the predicted miRNAs that interacted with the related circRNAs. (B) Gene ontology (GO) analysis for 12 circRNA-interacting miRNAs and their target genes showing significantly enriched pathways. Red indicates biological process (BP), green indicates molecular function (MF), and blue indicates cellular component (CC). (C) KEGG Pathway analysis for 12 circRNA-interacting miRNAs and their target genes showing significantly enriched signaling pathways. The Y-axis indicates pathway name, and the X-axis indicates the richness factor. The size of the spots represents the number of enriched differential target genes, and change of color from green to red represents the Q-value.

Figure 3

Expression of candidate circRNAs in PBMCs of SLE patients and healthy controls. The expression trends of seven circRNAs was consistent with the NGS profile results. qRT-PCR was conducted on RNA samples from five SLE patients and five HCs. Data are presented as 2^−ΔCt relative to GAPDH expression (mean ± standard deviation).

Figure 4

Double validation of hsa_circ_0000479 as an SLE diagnosis marker in two additional cohorts. (A) Expression of hsa_circ_0000479 in the validation-phase cohort with 23 SLE patients (including the active group and the stable group) and 21 HCs. (B) Expression of hsa_circ_0000479 in the external validation phase cohort with 64 SLE patients (including the active group and the stable group), 58 HCs, and 50 RA patients. Data are presented as a box plot. The “°” and “*,” respectively, indicate data that are more than 1.5-fold and 3-fold the quartile distance from the upper or lower bounds of the box. (C,D) Receiver operating characteristic (ROC) curves of hsa_circ_0000479 in the validation-phase and external validation-phase cohorts for SLE diagnosis.

Figure 5

Biological functions of hsa_circ_0000479 acting as a ceRNA. (A) Predicted circRNA-miRNA network. The SLE-related miRNA is annotated by blue nodes. (B) Biological processes associated with the target genes of hsa_circ_0000479. (C) Wnt-16 mRNA and protein expression in PBMCs from SLE patients and healthy controls. Wnt-16 mRNA expression was evaluated by NGS-Seq in the discovery phase. Wnt-16 protein expression levels were normalized to those of GAPDH.