Transcriptome signature in the blood of neuromyelitis optica spectrum disorder under steroid tapering

Abstract

Background: The advent of biologics has significantly transformed treatment strategies for neuromyelitis optica spectrum disorder (NMOSD). However, there are no biomarkers that predict relapses associated with steroid tapering; therefore, it is critical to identify potential indicators of disease activity. In this study, we collected peripheral blood mononuclear cells (PBMCs) from NMOSD patients during steroid tapering and performed bulk RNA sequencing to analyze changes in immune dynamics caused by steroid reduction.

Methods: PBMCs were collected at 3-5 timepoints from 10 NMOSD patients at our hospital (including one relapse case), and bulk RNA sequencing was performed. All patients were positive for anti-AQP4 antibodies and had no history of biologic use.

Results: In one relapsed patient, gene groups with decreased expression at relapse were observed predominantly in monocytes, with upregulation in anti-inflammatory pathways such as IL-10, while the upregulated genes were related to interferon signaling. Moreover, after steroid tapering, in non-relapsed patients, genes with increased expression were enriched in inflammatory pathways, represented by interferon signaling, while genes with decreased expression were enriched in pathways related to IL-10 and glucocorticoid receptors. Weighted gene co-expression network analysis identified modules that correlated with steroid dosage, and the modules inversely correlated with steroid dosage were enriched in monocytes, with marked immune signature of interferon pathway.

Conclusion: This study identified peripheral blood transcriptome signatures that could lead to the identification of clinically relevant NMOSD disease activity biomarkers, and further highlights the pivotal role of interferon and IL-10 signaling in NMOSD.

Keywords: IL-10; NMOSD; interferon; steroid; transcriptome signature.

Figure 1

Immune signature changes at the relapse of NMOSD. (A) PCA of samples derived from a patient with NMOSD in both remission and relapse phases. Relapse_1T: relapse timepoint; Relapse_2T: after acute-phase treatment; Remission_1T: remission timepoint; Remission_2T: remission timepoint. (B) K-means clustering of four samples derived from patients with NMOSD in both remission and relapse phases. Relapse_1T: relapse timepoint; Relapse_2T: after acute-phase treatment; Remission_1T: remission timepoint; Remission_2T: remission timepoint. (C) Single-cell annotation of PBMCs, and annotations colored according to cluster expressions determined by K-means analysis of bulk RNA-seq data (B). (D) Volcano plot showing DEGs between relapse and remission phases (Relapse v.s Remission).

Figure 2

qPCR results of DEGs. (A, B) qPCR results of five upregulated genes (A) and five downregulated genes (B) identified by DEG analysis, comparing the relapse and the remission phases. All samples were collected sequentially over time in the order of their sample numbers. Patient10_1T and Patient10_2T were collected during remission phase, Patient10_3T was collected at relapse timepoint, and Patient10_4T was collected after acute-phase treatment. Error bars represent standard errors (SE). *P < 0.05.

Figure 3

Biological pathways playing central roles in NMOSD relapse. (A–C) Pathway analysis results of clusters calculated by K-means in samples from the patient with NMOSD whose data are depicted in Figure 1 . Cluster pathways shown to be downregulated in the relapse phase according to either the Reactome2022 database (A) or the Wikipathway 2023 human database (B). (C) Cluster pathways showing enhanced expression in the relapse phase according to the Reactome2022 database. The bar charts show the top 10 most enriched terms, along with their corresponding p-values.

Figure 4

Blood transcriptome signature of patients with NMOSD under steroid tapering. (A) PCA of blood samples from nine patients with NMOSD. Sample names refer to the annotated patient number and the chronological timepoints at which the samples were collected, the latter described by “T” in numerical order. (B) For Patients 1, 2, and 5 (Group 1), and Patients 7 and 9 (Group 2), heatmaps under steroid tapering are shown, along with single-cell graphs of PBMCs colored according to the expression of genes in each cluster. For each patient, steroid dose decreases from left to right on the heatmaps.

Figure 5

qPCR results of representative genes identified by K-means method. The qPCR results of representative genes (CXCL8, EGR1, and SERPINB2) listed in Cluster 2 of Patient 5, identified using K-means method. Patient5_Monocytes, T cells, and B cells refer to three types of immune cells (monocytes, T cells, and B cells) that were isolated from the PBMCs of Patient 5 using flow cytometry. Error bars represent standard errors (SE). *P < 0.05.

Figure 6

WGCNA showing specific modules correlated with steroid tapering. (A) Gene cluster dendrogram. Each branch in the figure represents one gene, and each color represents the corresponding co-expression module. (B) Correlations between each module calculated in (A) and the steroid dosages of patients with NMOSD, excluding the relapsed patient. The numbers in each cell show the correlation (upper) and the p-value (lower).

Figure 7

Characteristics of steroid dosage correlated with weighted gene co-expression network analysis (WGCNA) module. (A) Heatmaps showing the correlation between the royalblue module, identified by WGCNA, and steroid tapering in patients with NMOSD. For Patients 5, 6, 7, and 9, the steroid dose decreases from left to right on the heatmaps. Patient 10 is the individual who relapsed, and the two left columns and two right columns represent samples obtained during the remission and relapse phases, respectively. More specifically, within the two right columns, the left column represents the sample obtained immediately after relapse, while the right column represents the sample immediately after acute-phase treatment. (B) Single-cell graphs of PBMCs colored according to the expression of the royalblue module identified by WGCNA as correlating with steroid tapering. (C) Pathways of modules identified by WGCNA as correlating with steroid tapering according to the Reactome2022 database. The bar chart shows the top 10 most enriched terms along with their corresponding p-values. .