Chronic Household Air Pollution Exposure Is Associated with Impaired Alveolar Macrophage Function in Malawian Non-Smokers

Abstract

Background: Household air pollution in low income countries is an important cause of mortality from respiratory infection. We hypothesised that chronic smoke exposure is detrimental to alveolar macrophage function, causing failure of innate immunity. We report the relationship between macrophage function and prior smoke exposure in healthy Malawians.

Methods: Healthy subjects exposed daily to cooking smoke at home volunteered for bronchoalveolar lavage. Alveolar macrophage particulate content was measured as a known correlate of smoke exposure. Phagocytosis and intraphagosomal function (oxidative burst and proteolysis) were measured by a flow cytometric assay. Cytokine responses in macrophages were compared following re-exposure in vitro to wood smoke, before and after glutathione depletion.

Results: Volunteers had a range of alveolar macrophage particulate loading. The macrophage capacity for phagosomal oxidative burst was negatively associated with alveolar macrophage particulate content (n = 29, r2 = 0.16, p = 0.033), but phagocytosis per se and proteolytic function were unaffected. High particulate content was associated with lower baseline CXCL8 release (ratio 0.51, CI 0.29-0.89) and lower final concentrations on re-exposure to smoke in vitro (ratio 0.58, CI 0.34-0.97). Glutathione depletion augmented CXCL8 responses by 1.49x (CI 1.02-2.17) compared with wood smoke alone. This response was specific to smoke as macrophages response to LPS were not modulated by glutathione.

Conclusion: Chronic smoke exposure is associated with reduced human macrophage oxidative burst, and dampened inflammatory cytokine responses. These are critical processes in lung defence against infection and likely to underpin the relationship between air pollution and pneumonia.

Fig 1. Particulate burden is measured by the percentage of macrophage cytoplasm taken up with particulate matter as calculated by digital image analysis of light microscopy images.

ImageSXM software analyses cytospin images treated with Fields B stain (A), and identifies both cytoplasmic (B) and particulate areas (C). The proportion of particulate to cytoplasm in the output image (D) is used as a measure of recent particulate exposure. Fifty 40x fields are used, and the overall mean particulate burden used as the summary statistic. Panels B and C are generated internally within the software, and are shown here for illustrative purposes only.

Fig 2. Alveolar macrophages naturally exposed to higher levels of particulate have reduced capacity for intraphagosomal oxidative burst, but an unaffected capacity for phagocytosis and proteolysis.

Unselected participants naturally exposed to inhaled particulates underwent bronchoscopy. The burden particulate within the alveolar macrophage was used as a measure of overall exposure. The capacity of alveolar macrophages to phagocytose, produce oxidative burst and proteolytic responses was tested on the same day by co-incubation with fluorophore labelled silica beads. (A) phagocytosis as indicated by the percentage of macrophages associated with fluorophore labelled beads by flow cytometry. (B) Oxidative burst and (C) Proteolysis were measured using twin labelled beads with calibrator fluor and reporter fluor (DCFH-DA and DQ-bovine serum albumin conjugate respectively), reported as “activity index” (see Methods). Each dot represents an individual (n = 29). Solid and dotted lines show the linear regression model with 95% confidence intervals. The x-axis describes the particulate burden (% of cytoplasm taken up by particulate). This has been log transformed by x’ = log(x+0.05), as described in the methods in order to normalise the data.

Fig 3. CXCL8 release is decreased in particulate laden macrophages.

Prior exposure to particulates was measured by HAM particulate burden. Individuals were designated “low” and “high” exposures according to whether their HAM particulate burden was below or above the median value for this group (n = 24). Adherent HAM on day 1 after bronchoscopy were incubated with media only or with additional wood smoke suspension for 6 hours. CXCL8 concentrations in the cell culture supernatant following challenge are shown logarithmically transformed (log₁₀). For all panels, open circles represent individuals with “low” macrophage particulate burden. Filled circles are those with “high” baseline particulate burden. (A) shows concentrations of CXCL8 in media for control and wood smoke stimulated wells for the entire group. There is a significant increase in concentration in stimulated cells (paired t-test, p<0.0001). (B) shows CXCL8 concentrations from untreated (control) and wood smoke treated wells. Here, results are grouped according to baseline particulate burden (low and high as described above). Bars show the mean and 95% confidence intervals. There is a significantly higher CXCL8 concentration from cells with low baseline particulate burden in both control and wood smoke treated wells (two-sided t-test, p = 0.020 and p = 0.039 respectively). (C) represents the increase in CXCL8 (log₁₀ change in wood smoke treated wells compared with control). Bars show the mean and 95% confidence intervals. There is no difference in the magnitude of cytokine response between cells with low and high baseline particulate burden (two sided t-test, p = 0.93).

Fig 4. BSO treatment of HAM reduced total intracellular glutathione concentrations but did not alter overall proportion of glutathione in the oxidised state.

Ex vivo adherent macrophages (3x10⁶ per well) were cultured for 18 hours in medium only (control), or with addition of 0.2mM BSO. Concentrations of total and oxidised glutathione were measured using an enzymatic recycling assay (see Methods). (A) Total glutathione (a measure of buffering capacity against oxidative stress) is reduced by BSO treatment compared with control. (B) Oxidised glutathione, as conventionally expressed as a percentage of total (a measure of oxidative stress), was not significantly altered by BSO treatment. n = 4 in each group. Boxes indicate 25^th to 75^th centile with median, and whiskers at 95% CI. Two sided t-test used for comparison.

Fig 5. Cytokine responses to wood smoke particulates and LPS: the effect of intracellular redox balance.

Ex-vivo adherent alveolar macrophages were pre-incubated with media alone (control), 0.2mM BSO (glutathione depletion), or BSO plus 2mM NAC (antioxidant depletion-repletion) for 18 hours, followed by stimulation with either media alone (control), wood smoke (left panels, WS) or lipopolysaccharide (right panels, LPS). After WS stimulus, low level changes were seen in CXCL8 and CCL2 release, while other cytokines were no significantly different. Maximal responses were seen with LPS, with no consistent effect of redox manipulation. Boxes show mean cytokine concentration in cell culture supernatants with 25^th to 75^th centile, and whiskers at 95% CI. n = 8 for each group. p values represent repeated measures ANOVA on log-transformed data. * denotes difference from control. † denotes difference from stimulated condition without pre-treatment. Significance testing by Sidak’s multiple comparisons test (p<0.05 significance level). Horizontal broken line represents limit of detection for each cytokine.