Demographic data, clinical and laboratory characteristics
A total of 122 children in the MPP group and 54 controls were enrolled in the present study. The demographic data, clinical and laboratory characteristics of children are shown in Table 1. In the MPP group, the mean age of the children was 6.86 ± 0.25 years, with 63/59 males/females accounting for 51.6%/48.4%, respectively. 11 patients (9.01%) had pleural effusion, and 29 (23.77%) had lobar pneumonia. Clinically, children with pneumonia mainly manifested with cough and fever, with an average fever duration of 7.6 ± 0.38 days and fever peak of 39 ± 0.34℃. The mean age of the control group was 3.37 ± 0.37 years, with 38 males (70.4%) and 16 females (29.6%).
TABLE 1 Demographic data and clinical characteristics of patients enrolled in this study
Altered CXCL10 protein levels in children with MPP
We first performed a protein chip analysis by using 3 blood samples from the groups of refractory Mycoplasma pneumoniae pneumonia (RMPP), non-refractory Mycoplasma pneumoniae pneumonia (NRMPP), and healthy patients to screen for differentially expressed proteins (DEPs) among these groups. Our data revealed that CXCL10 was the most significantly upregulated proteins in MPPs (P < 0.05; Figure 1A). Moreover, ontology analysis showed that most DEPs related to the cytokine-cytokine receptor interaction pathway, such as cell migration/chemotaxis, chemokine-mediated signaling pathway and CXCR3 chemokine receptor binding (Figures 1B). Then we used KEGG analysis to predict which the JAK/STAT signaling pathway would involve CXCL10 in the pathogenesis of NRMPP or RMPP (Figures 1C).
Clinical significance of IFN-γ/CXCL10 in the peripheral blood and BALF
In line with data obtained in protein chip analysis, the expression levels of inflammatory factors CXCL10 and IFN-γ were significantly elevated in the peripheral blood and BALF of MMP patients than those in controls (Figure 2A). Moreover, levels of CXCL10 were significantly higher in the BALF as compared to that in peripheral blood of MPP patients (Figure 2B). Notably, blood CXCL10 levels were positively correlated with days of hospitalization (r = 0.565, P < 0.05), fever duration (r = 0.341, P < 0.05) and serum IFN-g levels (r = 0.576, P < 0.0001), and BALF CXCL10 levels were positively correlated with total IgM in blood (r = 0.365, P < 0.05) in the MPP group (Figure 2C). In addition, the numbers of neutrophils and macrophages in the alveolar lavage fluid of MPP patients were remarkably increased as compared to those in the control group, and the majority (> 80%) of macrophages in BALF of MPP patients exhibited a M1-polarized phenotype (CD16+CD64+CD163-) (Figure 2D&E). In line with its biological function, CXCL10 levels showed a significant correlation with neutrophil/macrophages numbers in BALF (Figure 2F).
STAT1 down-regulation inhibits IFN-g-induced CXCL10 expression in M1-type monocytic macrophages
IFN-γ promoted the mRNA and protein levels of CXCL10 in THP-1-derived macrophages in a dose-dependent fashion in all time points (3, 6, and 9 h) analyzed (P < 0.05, Figure 3A&B), accompanied with the increased phosphorylation of STAT1 (Figure 3C), indicating that IFN-g induces CXCL10 via STAT1. Indeed, siRNA-mediated downregulation of STAT1 significantly diminished the ability of activated Th1 cells (Figure 3D) or IFN-g (Figure 3E) to elevate CXCL10 expressions in THP-1-derived macrophages as determined by western blot, qPCR and ELISA. Moreover, si-RNA mediated downregulation of STAT1 repressed the ability of THP-1-derived macrophages to activate Th1 cells, as evidenced by the significantly reduced IFN-g levels in culture supernatants (Figure 3F).
CXCL10 promotes the migration of Th1 cells
The results we showed above indicated that IFN-g and/or activated Th1 cells promotes CXCL10 productions in macrophages via STAT1. To determine the reciprocal effect of STAT1-induced CXCL10 in macrophages on Th cells, we next conducted transwell experiments. The addition of IFN-g in the lower chamber greatly promoted the migration of Th1 cells, which was significantly diminished by siRNA-mediated knockdown of STAT1 (Figure 4), indicating that macrophage-derived CXCL10, triggered by the IFN-g/STAT1 axis, induces the transmigration of Th1 cells.
Together, our data indicate that IFN-g promotes the production of CXCL10 in macrophages via STAT1. Subsequently, CXCL10 recruits more IFN-g-producing Th1 cells into the inflammatory sites, thereby constituting a vicious circle that exacerbates proinflammatory immune responses and tissue damages in patients with MPP infections