Plasma microRNA Profiles as a Potential Biomarker in Differentiating Adult-Onset Still's Disease From Sepsis

Abstract

Adult-onset Still's disease (AOSD) is a systemic inflammatory disease characterized by cytokine storm. However, a diagnostic test for AOSD in clinical use is yet to be validated. The aim of our study was to identify non-invasive biomarkers with high specificity and sensitivity to diagnosis of AOSD. MicroRNA (miRNA) profiles in PBMC from new-onset AOSD patients without any treatment and healthy controls (HCs) were analyzed by miRNA deep sequencing. Plasma samples from 100 AOSD patients and 60 HCs were used to validated the expression levels of miRNA by qRT-PCR. The correlations between expression levels of miRNAs and clinical manifestations were analyzed using advanced statistical models. We found that plasma samples from AOSD patients showed a distinct miRNA expression profile. Five miRNAs (miR-142-5p, miR-101-3p, miR-29a-3p, miR-29c-3p, and miR-141-3p) were significantly upregulated in plasma of AOSD patients compared with HCs both in training and validation sets. We discovered a panel including 3 miRNAs (miR-142-5p, miR-101-3p, and miR-29a-3p) that can predict the probability of AOSD with an area under the receiver operating characteristic (ROC) curve of 0.8250 in training and validation sets. Moreover, the expression levels of 5 miRNAs were significantly higher in active AOSD patients compared with those in inactive patients. In addition, elevated level of miR-101-3p was found in AOSD patients with fever, sore throat and arthralgia symptoms; the miR-101-3p was also positively correlated with the levels of IL-6 and TNF-α in serum. Furthermore, five miRNAs (miR-142-5p, miR-101-3p, miR-29c-3p, miR-29a-3p, and miR-141-3p) expressed in plasma were significantly higher in AOSD patients than in sepsis patients (P < 0.05). The AUC value of 4-miRNA panel (miR-142-5p, miR-101-3p, miR-29c-3p, and miR-141-3p) for AOSD diagnosis from sepsis was 0.8448, revealing the potentially diagnostic value to distinguish AOSD patients from sepsis patients. Our results have identified a specific plasma miRNA signature that may serve as a potential non-invasive biomarker for diagnosis of AOSD and monitoring disease activity.

Keywords: AOSD; adult-onset still's disease; biomarker; inflammation; microRNAs; sepsis.

Figure 1

Overview of miRNAs Sequencing Data from AOSD patients. (A) Volcano plot of the 482 miRNAs of 6 samples (3 AOSD and 3HCs). The abscissa represents log2 (fold change) of the miRNA expression while the ordinate represents the logarithmic transformation of the P-value gained by t-test. (B) Heat map with two-dimensional clustering of the imputed and normalized miRNA expression in the 6 samples. Missing data are intercepted before this procedure.

Figure 2

Differentially expressed plasma miRNAs signature for AOSD diagnosis in a training set. Expression levels of (A) miR-142-5p, (B) miR-101-3p, (C) miR-29c-3p, (D) miR-29a-3p, (E) miR-141-3p were analyzed by qRT-PCR in plasma from AOSD (N = 68) and HCs (N = 41). The results of histograms represent the means ± SD, ^*P < 0.05, ^***P < 0.001, ^****P < 0.0001. (F-H) ROC analysis for individual miRNAs and combined miRNA panel was analyzed by KNN (F), SVM (G), and LR (H) analysis.

Figure 3

The plasma miRNAs panel for AOSD diagnosis in a validation set. Expression levels of (A) miR-142-5p, (B) miR-101-3p, (C) miR-29c-3p, (D) miR-29a-3p, (E) miR-141-3p were analyzed by qRT-PCR in plasma from AOSD (N = 32) and HCs (N = 19). The results of histograms represent the means ± SD, ^*P < 0.05, ^****P < 0.0001. (F-H) ROC analysis for individual miRNAs and combined miRNA panel was analyzed by KNN (F), SVM (G), and LR (H) analysis. Logistic regression demonstrated that a linear combination of values for miR-142-5p, miR-101-3p, and miR-29a-3p produced the best model for AOSD diagnosis.

Figure 4

Association of the miRNA signature with AOSD disease activity. Expression levels of (A) miR-142-5p, (B) miR-101-3p, (C) miR-29c-3p, (D) miR-29a-3p, (E) miR-141-3p in active AOSD (N = 74) and inactive AOSD (N = 26) patients, and plasma samples from 60 HCs were used as control. The results of histograms represent the means ± SD, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001.

Figure 5

The correlation of miR-101-3p with serum IL-6 and TNF-α. Correlation analysis of miR-101-3p with serum levels of (A) IL-6 and (B) TNF-α in AOSD patients.

Figure 6

Expression of AOSD miRNA signature in sepsis patients. Expression levels of (A) miR-142-5p, (B) miR-101-3p, (C) miR-29c-3p, (D) miR-29a-3p, (E) miR-141-3p were analyzed by analyzed by qRT-PCR in an independent set consisting of HC (N = 18), AOSD (N = 25) and sepsis (N = 21) patients. The results of histograms represent the means ± SD, ^*P < 0.01, ^***P < 0.001. (F-H) ROC analysis for individual miRNAs and combined miRNA panel was analyzed by KNN (F), SVM (G) and LR (H) analysis. Logistic regression demonstrated that a linear combination of values for miR-142-5p, miR-101-3p, miR-29c-3p, and miR-141-3p produced the best model to distinguish AOSD patients from sepsis patients.