Obesity dysregulates the pulmonary antiviral immune response

Abstract

Obesity is a well-recognized risk factor for severe influenza infections but the mechanisms underlying susceptibility are poorly understood. Here, we identify that obese individuals have deficient pulmonary antiviral immune responses in bronchoalveolar lavage cells but not in bronchial epithelial cells or peripheral blood dendritic cells. We show that the obese human airway metabolome is perturbed with associated increases in the airway concentrations of the adipokine leptin which correlated negatively with the magnitude of ex vivo antiviral responses. Exogenous pulmonary leptin administration in mice directly impaired antiviral type I interferon responses in vivo and ex vivo in cultured airway macrophages. Obese individuals hospitalised with influenza showed dysregulated upper airway immune responses. These studies provide insight into mechanisms driving propensity to severe influenza infections in obesity and raise the potential for development of leptin manipulation or interferon administration as novel strategies for conferring protection from severe infections in obese higher risk individuals.

Fig. 1. Schematic diagram showing recruitment into the study.

Subjects were recruited as part of a case control study. Total subject numbers screened in each group and subjects excluded are shown.

Fig. 2. Obesity does not alter bronchial epithelial cell antiviral or proinflammatory responses to influenza infection.

a Bronchial epithelial cells from 10 obese and 13 non-obese control subjects were cultured and infected ex vivo with H1N1/09 (H1N1), seasonal H3N2 and B/Florida (B/Flo) influenza viruses. Cell supernatants were collected at 24 h. b Interferon (IFN)-α, (c) IFN-β, (d) IFN-λ, (e) IL-6, (f) CXCL-8/IL-8 and (g) TNF protein concentrations were quantified by multiplex ELISA. Graphs show datapoints for individual subjects with lines indicating median (IQR). Data analysed by Kruskal Wallis with Dunn’s post-test or Mann–Whitney U test. Significance indicated above each group compares virus infected cells to medium-treated control cells. Brackets show comparisons between non-obese and obese subject groups. ns non-significant. Source data are provided as a Source Data file.

Fig. 3. Antiviral immune responses are impaired in bronchoalveolar lavage cells from obese subjects.

a Bronchoalveolar lavage (BAL) cells from 15 obese and 15 non-obese control subjects were cultured and infected ex vivo with H1N1/09 (H1N1), seasonal H3N2 and B/Florida (B/Flo) influenza viruses. Cell supernatants and lysates were collected at 24 and 48 h post-infection. b Interferon (IFN)-α, (c) IFN-β, and (d) IFN-λ. Influenza M gene mRNA expression was measured at 24 and 48 h following infection with (e), H1N1 (f) H3N2, and (g) B/Florida viruses. h IL-6, (i) CXCL-8/IL-8, and (j) TNF protein concentrations were quantified by multiplex ELISA. In b–d and h–j graphs show datapoints for individual subjects with lines indicating median (IQR). Significance indicated above each group compares virus infected cells to medium-treated control cells. In (e–g) data only measured in a subset due to sample dropout (RNA quality) and data expressed as median (IQR) and significance indicated above each time point compares non-obese with obese subjects. Data analysed by Kruskal–Wallis with Dunn’s post test. Brackets show comparisons between non-obese and obese individual groups. ns non-significant. Source data are provided as a Source Data file.

Fig. 4. Altered airway metabolomic milieu in obesity.

a Bronchoalveolar lavage (BAL) fluid from 15 obese and 15 non-obese control subjects was collected and processed for metabolomics analysis. b Orthogonal projections to latent structures-discriminant analysis (OPLS-DA) showing separation of obese (green) and non-obese (red) subjects. c Differential abundance analysis, performed via visualisation of the volcano plot. Data were generated using an unpaired t-test (2-sided) and p values refer to the unadjusted and -log₁₀ transformed value. d A heat map clustered for obese (green) and non-obese (red) subjects to show the top 50 metabolite correlations. An enlarged version of this plot to enable the reading of metabolite annotations can be found in Supplementary Fig. 4 (e) Enrichment analysis identifying the top 25 representative metabolic pathways significantly enriched in obese vs. non-obese BAL fluid. Data were computed using the R package globaltest which applies a generalised linear model to estimate a Q-statistic, with p values (2-sided) referring to the unadjusted and -log₁₀ transformed value. f Measurements of leptin in obese vs non-obese subjects in upper (nasosorption, left) and lower (bronchosorption, right, BAL fluid centre) airway samples shown as individual values and median (solid horizontal line) and IQR. g, h Correlations of bronchosorption leptin concentrations with (g) the magnitude of ex vivo BAL cell IFN-β responses to H1N1 (left), H3N2 (centre), and B/Flo (right) influenza strains and (h) BAL fluid AMP concentrations. Data in F analysed by Mann–Whitney U test. Data in G-H analysed by Spearman’s rank correlation test (two sided). Source data are provided as a Source Data file.

Fig. 5. Pulmonary leptin administration reduces airway immune responses to influenza infection in mice.

a BALB/c mice were treated with intranasal leptin protein (16μg) or vehicle control, 12 h prior to infection with Influenza X31 or PBS control. b Ifnβ, 2’-5’Oas, Viperin and PKR lung mRNA expression measured by qPCR at 6 h post-infection. c BAL neutrophils and activated neutrophil subsets (CD63+ and CD64+) enumerated by flow cytometry at 72 h post-infection. BAL concentrations of (d) IL-1β, (e) IL-6 and (f) TNF measured by ELISA at 72 h post-infection. g Lung Socs3 mRNA expression at 18 h following leptin administration. h BAL macrophages were harvested from naïve untreated mice and cultured ex vivo followed by treatment with leptin protein (32 and 320 µg/ml) for 12 h before infection with Influenza X31. i Ifnβ, 2′−5′Oas, and PKR mRNA expression in cell lysates at 6 h post-infection. j Socs3 mRNA expression at 18 h following leptin administration. Data are presented as mean (±SEM) for n = 5 mice per group in b–g and n = 5 for i, j. Statistical significance analysed using one-way ANOVA with Bonferroni post-test. *P < 0.05, **P < 0.01, ***P < 0.001. Source data are provided as a Source Data file.

Fig. 6. Altered upper airway immune responses in obese adult patients hospitalised with influenza infection.

a Schematic of the Mechanisms of Severe Acute Influenza Consortium (MOSAIC) study highlighting timepoints from hospitalisation of study samples. b Nasopharyngeal virus load between obese and non-obese patients. c Nasopharyngeal aspirate and (d) serum multiplex immune mediators comparing obese and non-obese patients. Data from 133 influenza positive adults (27 obese, 106 non-obese) compared using Mann–Whitney test (two-tailed). Box and whisker plots show median (line within box), interquartile range (box) and 1.5 x IQR (whiskers) *P < 0.05. Source data are provided as a Source Data file.