Transcriptomic features of SLE patients in flare and changes during acute in-hospital treatment

Abstract

Systemic lupus erythematosus (SLE) is a heterogeneous autoimmune disease characterized by diverse clinical symptoms. Most patients suffer from varied disease relapse-remission cycles under current treatment regimens, which often leads to acute in-hospital treatment for disease flare. To better understand the dynamic transcriptomic features during SLE flare and in-hospital treatment, we investigated the transcriptome profiles of 65 SLE patients both before (SLE1) and after (SLE2) their acute in-hospital treatment, using RNA-sequencing profiling of their peripheral blood mononuclear cells (PBMCs). Together with 15 healthy donors as controls (HC), we studied the SLE-dysregulated transcriptomic features and the corresponding treatment effects. Among HC, SLE1, and SLE2 groups, more differentially expressed genes (DEGs) were identified in SLE1/SLE2 vs. HC than those in SLE2 vs. SLE1. And the DEGs of SLE1 vs. HC were largely overlapped with DEGs of SLE2 vs. HC, indicating the relatively small effect of treatment. Functional enrichment of DEGs confirmed that the overall gene expression profile did not change much after treatment, although the acute clinical manifestations of patients had been in great remission. Overlapping the DEGs identified three SLE disease signatures, the interferon (IFN), plasmablast, and neutrophil signatures, which were upregulated in disease flare (SLE1 vs. HC) and affected by the treatment (SLE2 vs. SLE1). Interestingly, the IFN and plasmablast signatures were largely downregulated upon treatment, while the neutrophil signature remained unchanged or even enhanced. Together with the well-known IFN score, a similar neutrophil signature score helped with the stratification of SLE samples. Furthermore, higher IFN signature scores were found well correlated with low cell counts on lymphocytes, neutrophils, and monocytes in SLE patients, while neutrophil scores were negatively correlated with lymphocyte percentage in white blood cells and positively correlated with neutrophile percentage but showed no correlation with monocyte percentage. These observations indicated the important roles of both IFN and neutrophil signatures and scores in reflecting the dynamic changes of molecular aberrations in SLE patients and the associated clinical development during disease flare and acute in-hospital treatment. Additionally, alternative splicing (AS) events were detected, and genes with differential alternative splicing (DSGs) were identified among HC, SLE1, and SLE2 groups. We found that most DSGs (>90%) were not DEGs and they tended not to be functionally enriched. More investigations on specific DSGs in the future, such as the NLRC4 gene, may result in novel knowledge on SLE pathogenesis. Finally, with a rationale that the transcriptome profile may bridge the gaps between genetic risks and disease phenotypes, the DEGs and DSGs were overlapped with the genes relevant to SLE risk loci reported by genome-wide association study (GWAS). The overlaps turned out to be few, indicating more complicated scenarios of interactions among genetic risks, transcriptomic aberrations, and disease clinical phenotypes. In summary, the dynamic transcriptomic features in SLE patients were observed during in-hospital treatment for disease flare, highlighting the complexity of neutrophil signatures and providing new insights about alternative splicing events in SLE pathogenesis. Further effort is required to clarify the underlying mechanisms of genetics and transcriptomics interactions.