Total leukocyte counts, haematocrit levels, platelet counts, serum levels of albumin, plasma leakage (pleural effusion and ascites), and duration of hospitalisation differed significantly different between patients with DHF and those without DHF (Table 1). Because SOCS1 is a crucial regulator of IL-12-mediated IFN-γ production,9 we determined the expression level of SOCS1 in PBMCs derived from OFI and patients with DHF or DF by using a real-time quantitative RT-PCR assay. Results showed that SOCS1 expression was elevated in DF patients, but not in those with DHF (Fig. 1(a)). In addition, we

assessed IFN-γ and IL-10 levels in blood to reflect the change of Th1/Th2 profiles related to severity. It was found that patients with DF elicited a higher IFN-γ production than those with DHF (P = 0.033, Fig. 1(b)). In contrast, patients with DHF had a higher SP600125 IL-10 level than those with 3 Methyladenine DF (P = 0.046, Fig. 1(b)). There were no difference in the IFN-α and IL-13 levels between patients with DF and DHF (data no showed). To further determine whether DENV-2 infection could induce SOCS-1 expression in vitro, we examined both of DENV-2 viral load and SOCS-1

expression in mRNA level of PBMCs isolated from healthy subjects at 12–48 h postinfection and at a multiplicity of infection (MOI) of 1, 5, and 10, respectively. We determine the viral titre both in at 12, 24 and 48 h postinfection and in MOI = 1, 5, and 10 by the TaqMan RT-PCR assay. The detected DENV-2 titres were related to the time course and dose dependent (data not shown). A significant increase in SOCS1 expression was induced in the primary DENV-2 infection in PBMCs from healthy individuals at 24 h postinfection (n = 6, P = 0.002,

Fig. 2(a)). To further determine what population of PBMCs induce the SOCS-1 expression, we found that CD14+ cells isolated by positive selection using CD14 microbeads were infected with DENV-2 at an MOI of 5, had significantly elevated SOCS1 expression, whereas CD14– cells did not (n = 6, P < 0.001, Fig. 2(b)). To determine which miRNAs were associated with DF or DHF, we initially screened 20 pairs of PBMC samples for 11 miRNAs that were potential regulators of SOCS1 based on bioinformatics-based analysis mafosfamide of the SOCS1 mRNA 3′ untranslated region (3′ UTR) (Fig. 3(a)). Using real-time RT-PCR, we analysed the expression levels of miR-150, 181a, 155, 221, and 572 in the PBMCs of DF and DHF patients (Fig. 3(b)). The expression levels of miR-221 (2.82-fold; P = 0.021) and miR-572 (4.60-fold; P = 0.012) in patients with DF were significantly higher than in DHF patients. Only miR-150 was expressed at elevated levels in DHF samples (7.16-fold; P = 0.008). We tested whether SOCS1 mRNA expression was inversely correlated to miR-150 expression in patients with DHF.