Estrogen contributes to sex differences in M2a macrophages during multi-walled carbon nanotube-induced respiratory inflammation

Abstract

Lung diseases characterized by type 2 inflammation are reported to occur with a female bias in prevalence/severity in both humans and mice. This includes previous work examining multi-walled carbon nanotube (MWCNT)-induced eosinophilic inflammation, in which a more exaggerated M2a phenotype was observed in female alveolar macrophages (AMs) compared to males. The mechanisms responsible for this sex difference in AM phenotype are still unclear, but estrogen receptor (ER) signaling is a likely contributor. Accordingly, male AMs downregulated ERα expression after MWCNT exposure while female AMs did not. Thus, ER antagonist Fulvestrant was administered prior to MWCNT instillation. In females, Fulvestrant significantly attenuated MWCNT-induced M2a gene expression and eosinophilia without affecting IL-33. In males, Fulvestrant did not affect eosinophil recruitment but reduced IL-33 and M2a genes compared to controls. Regulation of cholesterol efflux and oxysterol synthesis is a potential mechanism through which estrogen promotes the M2a phenotype. Levels of oxysterols 25-OHC and 7α,25-OHC were higher in the airways of MWCNT-exposed males compared to MWCNT-females, which corresponds with the lower IL-1β production and greater macrophage recruitment previously observed in males. Sex-based changes in cholesterol efflux transporters Abca1 and Abcg1 were also observed after MWCNT exposure with or without Fulvestrant. In vitro culture with estrogen decreased cellular cholesterol and increased the M2a response in female AMs, but did not affect cholesterol content in male AMs and reduced M2a polarization. These results reveal the modulation of (oxy)sterols as a potential mechanism through which estrogen signaling may regulate AM phenotype resulting in sex differences in downstream respiratory inflammation.

Keywords: Fulvestrant; carbon nanotubes; cholesterol; estrogen; inflammation; lung; macrophage; oxysterol; respiratory; sex.

Figure 1.. Comparison of male and female AMs in vivo.

Male and female mice were exposed to MWCNT (2 mg/kg) via oropharyngeal aspiration and lung lavage cells collected 7-days post-exposure. AMs were isolated from lung lavage by adherence culture for A) gene expression analysis by qPCR, or D) quantitation of ERα and ERβ protein by western blot. All lung lavage cells were used for flow cytometry measurement of B) AM ST2 (IL-33Rα) expression, and C) AM SSC MFI. mean ± SEM, n = 3–9. Data presented in graphs A-C are from a single experiment for each respective endpoint; data presented in figure D are from 3 independent experiments with similar results. * p < .05, statistical significance was analyzed using a two-way ANOVA and Sidak’s post hoc testing.

Figure 2.. Fulvestrant treatment prior to MWCNT exposure in female mice.

As shown in (A), 7-days prior to dispersion media (DM) vehicle or MWCNTs (2 mg/kg BW) exposure, ER antagonist Fulvestrant (200 mg/kg), or vehicle (5% DMSO in corn oil), was administered via subcutaneous depot injection to female mice. Fulvestrant was administered again on the day of MWCNT instillation and mice euthanized 7-days post-MWCNT. Whole lung lavage was used to collect cells from the airways for (B) total airway cell counts (n = 4–8), and (C) determination of airway eosinophil content by flow cytometry; data expressed as percentage of eosinophils in vehicle treated groups for DM and MWCNT exposures, respectively (n = 8–13). (D) concentrated lung lavage fluid was used to assess airspace cytokine levels (n = 4–13). AMs were isolated from the lung lavage fluid and (E) protein collected for total STAT6 and pSTAT6 expression analysis by western blot (n = 4–8), or (F) RNA isolated for gene expression analysis by qPCR; data expressed as fold change (ΔΔCt) compared to vehicle or Fulvestrant-treated DM animals, respectively (n = 3–5). Data is from two independent experiments with comparable outcomes (Fig. S2 A,B); mean ± SEM, * p < .05, statistical significance was analyzed using two-way ANOVA and Sidak’s post hoc testing.

Figure 3.. Fulvestrant administration prior to MWCNT exposure in male mice.

As shown in Figure 2A, 7-days prior to dispersion media (DM) vehicle or MWCNTs (2 mg/kg BW) exposure, ER antagonist Fulvestrant (200 mg/kg), or vehicle (5% DMSO in corn oil), was administered via subcutaneous depot injection to male mice. Fulvestrant was administered again on the day of MWCNT instillation and mice euthanized 7-days post-MWCNT. Whole lung lavage was used to collect cells from the airways for (A) total airway cell counts and (B) determination of airway eosinophil content by flow cytometry; data expressed as percentage of eosinophils in vehicle treated groups for DM and MWCNT exposures, respectively. (C) concentrated lung lavage fluid was used to measure airspace cytokine levels. (D) AMs were isolated from the lung lavage fluid and RNA isolated for gene expression analysis by qPCR; data expressed as fold change (ΔΔCt) compared to vehicle or Fulvestrant-treated DM animals, respectively. mean ± SEM, n = 3–5 from a single experiment.* p < .05, Statistical significance was analyzed using two-way ANOVA and Sidak’s post hoc testing.

Figure 4.. AM cholesterol transport genes and oxysterol levels in the airways of MWCNT-exposed mice.

Male and female mice were exposed to MWCNT (2 mg/kg) via oropharyngeal aspiration and airways lavaged 7-days post-exposure. AMs were isolated from lung lavage by adherence culture for A) gene expression analysis by qPCR. B) oxysterol concentrations measured in concentrated lung lavage fluid via LC-MS/MS. C) Gene expression in AMs isolated from Fulvestrant-treated mice (Figure 2A). Gene expression data reported as fold change (ΔΔCt) compared to same-sex DM controls (if applicable: vehicle or Fulvestrant, respectively). Data points for Abca1 and Abcg1 expression are a combination of 3 independent experiments with similar results; all other graphs contain data points from a single experiment. mean ± SEM, n = 3–13. * p < .05, statistical significance was analyzed using two-way ANOVA and Sidak’s post hoc testing.

Figure 5.. Cholesterol transporters and BMDM phenotype in vitro.

A-C) male and female WT BMDM were polarized to M1 (1 ng/ml IFNγ + 10 pg/ml LPS) or M2a (1 ng/ml IL-4 or 1 ng/ml IL-13) phenotypes for 8 hrs before gene expression analysis by qPCR. All genes of interest were normalized to geomean of reference genes Hprt1 and Gapdh (ΔCt) and expressed as (A, B) fold change of reference genes (2^−ΔCt), or (C) fold change of animal-matched no-treatment controls (ΔΔCt). D-E) male WT and SR-BI KO BMDM were polarized to M1 (D) or M2a (E) phenotypes and associated genes measured. Genes of interest were normalized to reference gene Hprt1 (ΔCt) and expressed as fold change of WT no-treatment controls (ΔΔCt). Data are from 2 independent experiments; mean ± SEM, n = 3–5. p < .05 *compared to same sex control, ^&compared to the same cytokine treatment in the opposite sex, ^#compared to same genotype control, ^$compared to the same cytokine treatment in different genotype; statistical significance was analyzed using two-way ANOVA and Sidak’s post hoc testing.

Figure 6.. Effects of in vitro E₂ treatment on AM cholesterol content and M2a polarization.

Primary AMs from naïve male and female mice were cultured in hormone-free media and treated with vehicle (1% ethanol) or 100 nM E₂ in vitro for 48 hrs before A) measurement of cellular Free Cholesterol or B) treatment with IL-13 (0.5 ng/ml) for 0, 2, 4, or 8 hrs prior to protein collection for measurement of (p)STAT6 by western blot. All data presented as the change between vehicle and E₂ treatments of AMs from the same animal: (E₂-treated well) − (vehicle-treated well) = A) Δ μM free cholesterol per 100,000 cells, or, B) Δ pSTAT6:STAT6. Both assays include data from 2 separate experiments; mean ± SEM, n = 3–6. Statistical significance was determined using a one sample t test, * p < .05 compared to the hypothetical value of 0 (no change).