miR-122 promotes virus-induced lung disease by targeting SOCS1

Abstract

Virus-induced respiratory tract infections are a major health burden in childhood, and available treatments are supportive rather than disease modifying. Rhinoviruses (RVs), the cause of approximately 80% of common colds, are detected in nearly half of all infants with bronchiolitis and the majority of children with an asthma exacerbation. Bronchiolitis in early life is a strong risk factor for the development of asthma. Here, we found that RV infection induced the expression of miRNA 122 (miR-122) in mouse lungs and in human airway epithelial cells. In vivo inhibition specifically in the lung reduced neutrophilic inflammation and CXCL2 expression, boosted innate IFN responses, and ameliorated airway hyperreactivity in the absence and in the presence of allergic lung inflammation. Inhibition of miR-122 in the lung increased the levels of suppressor of cytokine signaling 1 (SOCS1), which is an in vitro-validated target of miR-122. Importantly, gene silencing of SOCS1 in vivo completely reversed the protective effects of miR-122 inhibition on RV-induced lung disease. Higher miR-122 expression in nasopharyngeal aspirates was associated with a longer time on oxygen therapy and a higher rate of treatment failure in 87 infants hospitalized with moderately severe bronchiolitis. These results suggest that miR-122 promotes RV-induced lung disease via suppression of its target SOCS1 in vivo. Higher miR-122 expression was associated with worse clinical outcomes, highlighting the potential use of anti-miR-122 oligonucleotides, successfully trialed for treatment of hepatitis C, as potential therapeutics for RV-induced bronchiolitis and asthma exacerbations.

Keywords: Asthma; Inflammation; Molecular biology; Virology.

Figure 1. Rhinovirus-induced lung disease and miR-122 expression in infant mice.

Rhinovirus (RV) replication could be detected in the lower airways at 24 hours after inoculation (A) (n = 3–6 mice per group). RV infection induced inflammation in bronchoalveolar lavage fluid (BALF) (B) (n = 3–6 mice per group) and the expression of the murine IL-8 analogs CXCL1 and CXCL2 (C and D) (n = 4–8 mice per group) as well as IFN-β and IFN-λ (E and F) (n = 4–8 mice per group). A TaqMan miRNA qPCR array, including 750 miRNAs from MIRBase v21, identified 4 miRNAs significantly upregulated in the airways of mice 24 hours after infection with RV when compared with UV controls normalized to the small nuclear–RNA (sn-RNA) control gene RNU6b (G) (n = 6 mice per group). miR-122 was induced by RV in immortalized human airways basal cells in submerged culture (H) (n = 3 independent cultures). *P < 0.05, calculated using 1-way ANOVA with multiple comparisons correcting the false discovery rate 2-stage step-up method of Benjamini, Krieger, and Yekutieli for all panels, except for G, where moderated 2-tailed t test and Westfall-Young correction for multiple testing was used. SAL, saline vehicle; UV-RV, UV-inactivated RV. Data are shown as the mean ± SEM.

Figure 2. Effect of miR-122 inhibition on rhinovirus-induced lung disease in adult mice.

Naive adult mice infected with RV had reduced miR-122 expression in their airways 24 hours after RV infection when treated with miR-122 antagomir (A.122/RV) as compared with mice treated with a scrambled control antagomir (SCR/RV) (A) (n = 4–5 mice per group). Mice treated with A.122/RV had reduced neutrophil influx into the BALF (B) (n = 6–8 mice per group)and fewer myeloperoxidase-positive cells (neutrophils) in the lung tissue (C) (n = 6–8 mice per group) when compared with mice treated with SCR/RV. Representative images of myeloperoxidase-positive neutrophils are shown (D) (n = 6–8 mice per group) (original magnification, ×400). Expression of mouse IL-8 analogs CXCL1 and CXCL2 was upregulated following RV infection, and CXCL2 was reduced by A.122 treatment (E and F) (n = 6–8 mice per group). Mice treated with miR-122 antagomir were protected from RV-induced AHR (G) (n = 6–8 mice per group). Expression of RV was increased in airways of infected mice at 24 hours and more pronounced with the inhibition of miR-122 but returned to baseline by 96 hours in all groups (H) (n = 6–8 mice per group). IFN-β and IFN-λ were increased in the airways of infected mice and more pronounced when miR-122 was inhibited (I and J). (n = 6–8 mice per group). *P < 0.05, calculated using 1-way ANOVA with multiple comparisons correcting the false discovery rate 2-stage step-up method of Benjamini, Krieger, and Yekutieli, except for A and G, where 2-tailed t test and 2-way ANOVA were used, respectively. UVRV, UV-inactivated RV. Data are shown as the mean ± SEM.

Figure 3. Effect of miR-122 inhibition on rhinovirus-induced exacerbation of allergic lung disease in adult mice.

House dust mite (HDM) sensitized and challenged mice infected with RV had increased expression of miR-122 in the airways (SCR/RV/HDM), which was ameliorated by antagomir treatment targeting miR-122 (A122/RV/HDM) (A) (n = 4–6 mice per group). Increased RV replication was observed in the airways in mice where miR-122 was inhibited (B) (n = 4–6 mice per group). Mice treated with miR-122 antagomir were protected from RV-induced AHR (C) (n = 4–6 mice per group) and had reduced neutrophil influx (D) into the BALF when compared with mice treated with a scrambled control antagomir (SCR/RV/HDM) (n = 3–4 mice per group). Expression of IFN-β and IFN-λ in the airways was further increased by inhibition of miR-122 (A122/RV/HDM) (E and F) (n = 4–6 mice per group). CXCL1 and CXCL2 were also increased by RV and reduced by inhibition of miR-122 (G and H) (n = 4–6 mice per group). *P < 0.05, calculated using 1-way ANOVA with multiple comparisons correcting the false discovery rate 2-stage step-up method of Benjamini, Krieger, and Yekutieli, except for A and C, where 2-tailed t test and 2-way ANOVA were used, respectively. UVRV, UV-inactivated RV. Data are shown as the mean ± SEM.

Figure 4. Modulation of rhinovirus-induced lung disease by SOCS1 and miR-122 inhibition in adult mice.

Naive BALB/c mice were treated with antagomirs targeting miR-122 (A.122) or scrambled control (SCR). RV infection reduced levels of SOCS1 in the lungs, which was reversed by miR-122 inhibition (A) (n = 3–5 mice per group). The NF-κB subunit p65 was upregulated by RV, and this upregulation was reversed by miR-122 inhibition (B) (n = 3–5 mice per group). In separate experiments, mice also received siRNA-targeting SOCS1 or a nonsense sequence siRNA as control (NON) along with A.122 or SCR before being infected with RV (C). The effects of miR-122 inhibition on neutrophil influx into the bronchoalveolar space (D), AHR (E), CXCL2 (F), and RV replication (G) were all neutralized by inhibition of SOCS1 (n = 6–8 mice per group). *P ≤ 0.05, calculated using 1-way ANOVA with multiple comparisons correcting the false discovery rate 2-stage step-up method of Benjamini, Krieger, and Yekutieli, except for C–G, where 2-tailed t test was used to determine the effects of SOCS1 inhibition in A.122-treated mice. UV-RV, UV-inactivated RV. Data are shown as the mean ± SEM.

Figure 5. miRNA 122 in nasopharyngeal aspirates from infants admitted to the hospital with moderately severe bronchiolitis.

Infants with high miR-122 expression spent more hours on oxygen during their admission (A and B) and failed treatment more often (C). Infants who failed treatment trended toward increased detectable levels of miR-122 expression (D). Infants with higher levels of control miRNAs miR-21 and miR-423 had no difference in time on oxygen (E, F, I, and J, respectively) and no differences in the rate of treatment failure (G and K). Those that failed treatment had no difference in the expression of either miR-21 (H) or miR-423 (L) (n = 87). *P ≤ 0.05, calculated using Mann-Whitney test, except for B, F, and J, where Gehan-Breslow-Wilcoxon test was used to compare survival curves.