Figure 1. Airway and vascular cell infiltrates are increased in the airways when estrogen is given at steady-state levels.
Adult females are more likely to have increased allergic inflammation in comparison to men, yet steady-state levels of estrogen in the form of oral contraceptives improve asthma severity in adult women(14, 15, 33). With these data in mind, we hypothesized that estrogen may improve allergic inflammation in allergen-challenged animals. Ovarectomized animals (OVX) were implanted with subcutaneous estrogen (OVX-E2) or placebo pellets (OVX-Pl) and allergic inflammation was experimentally generated using chicken egg ovalbumin (OVA) sensitization and airway challenge protocols (Fig. 1A). High circulating levels of estrogen were appropriately detected in ovarectomized animals implanted with an estrogen pellet, while negligible levels of estrogen were detected when animals were implanted with a placebo pellet (Fig. 1B). Histological inflammatory scores were determined by a blinded pathologist (Fig. 1C). Cellular immune infiltration was determined by quantifying bronchial and vascular inflammation and cellular infiltrates (Fig. 1D-G). No inflammation was found in the F-Sham-saline treated animals even after methacholine challenge (Fig. 1C and 1D). F-Sham-OVA had more inflammation than Female-Sham-Saline mice indicating a significant induction of allergic inflammation with OVA sensitization and airway challenges (Fig. 1C,1D,1E; P < 0.05). OVX-E2-OVA had more inflammatory cell infiltrate than F-sham-OVA treated when comparing scoring results (Fig. 1C, 1D, 1E). Although not quantified basement membranes were thickened in all OVA treated animals and we noted airway contractility in the OVX-E2-OVA mice indicative of airway contractility (Indicated by black arrows in Fig. 1F, 1I and 1J). In summary these studies showed that estrogen significantly increased the allergic inflammatory response to airway challenge.
Airway resistance is increased in ovarectomized animals treated with estradiol in comparison to female-sham treated animals
In the next studies respiratory system mechanics were measured by determining resistance and elastance following methacholine challenge in each of the treatment groups: Female-Sham-Saline, Female-Sham-OVA, OVX-E2-OVA and OVX-Placebo-OVA. Airway resistance (Rrs) is an indicator of airway flow and airway obstruction, while elastance (Ers) determines the stiffness of the lung tissue. Rrs was higher in OVX-E2-OVA compared to Female-Sham-OVA (Fig. 2A; p = 0.0185). Differences in resistance between OVX-placebo-OVA compared to F-Sham-OVA did not reach statistical significance. Elastance (Ers) readings trended upwards in OVX-placebo-OVA animals compared to F-Sham-OVA (p = 0.0845), but only reached statistical significance in OVX-E2-OVA mice compared to F-Sham-OVA mice (Fig. 2B; P < 0.0001). Pressure volume (PV) loops were also determined over the course of pulmonary function testing (Fig. 2C). We observed changes, or flattening, of the pressure-volume loops which fits with the increased Ers in OVX-E2-OVA mice compared to F-Sham-OVA. No statistical differences were observed between OVX-placebo-OVA mice and F-Sham-OVA. Altogether the pulmonary function testing indicates that estrogen treatment stiffens the lung tissues in combination with increasing airway resistance in ovarectomized mice treated with ovalbumin.
Total immune cell infiltrate was reduced in BAL fluid from ovarectomized, estrogen-treated animals compared to female, sham-operated mice following allergen.
We compared the composition of immune airway infiltrates in each treatment group following OVA treatments and methacholine challenge (Fig. 3). OVA treated mice all had significant increases in neutrophils and eosinophils in comparison to F-Sham-Saline treated mice (Fig. 3A). OVX-E2-OVA treated animals had a higher percentage of macrophages represented in the BAL cells collected as compared to F-Sham-OVA, while OVX-placebo-OVA had a lower percentage of macrophages in comparison to F-Sham following OVA challenge. In addition, OVX-E2-OVA mice had a lower percentage of total lymphocytes as compared to F-Sham-OVA. Counts for each population were determined using the volume of BAL recovered multiplied by the percentage of cell type counted per field (Fig. 3B-3F). First, total BAL cells were reduced in OVX-E2-OVA treated animals compared to F-Sham-OVA; OVX-placebo-OVA treated mice maintained a comparable number of total cells to F-Sham-OVA in their BAL (Fig. 3B). The counts of macrophages were reduced in OVX-Placebo-OVA compared to F-Sham-OVA, but macrophages were comparable in OVX-E2-OVA and F-Sham-OVA (Fig. 3C). Most interestingly, OVX-E2-OVA had lower counts of neutrophils and eosinophils in comparison to F-Sham-OVA (Fig. 3D, 3E), while OVX-Placebo-OVA mice had increased numbers of neutrophils in comparison to F-Sham-OVA mice. Finally, less lymphocytes were recovered from the BAL fluid of OVX-E2-OVA treated animals compared to F-Sham-OVA controls (Fig. 2F). Together these data suggest that hormones or the lack of hormones significantly alter the immune cell compartment that contributes to pulmonary mechanics outcomes.
Allergic inflammatory cytokines are released at different rates in OVX-E2-OVA treated animals compared to intact female-sham mice treated with airway allergen. In the next studies we wanted to confirm that differences in BAL cell populations were not a byproduct of methacholine challenges, but present because of airway stress and inflammation induced by OVA treatment in conjunction with estrogen treatment. In the next studies we collected BAL following five consecutive daily intranasal OVA challenges. Sex differences are previously reported using the OVA model, and in our previous report, female BALB/c mice specifically had higher ILC2 responses to the allergy-associated type 2-inducing cytokine, IL-33, compared to males. In those studies, IL-5 and IL-13 responses were higher for females in comparison to males. In the remaining studies we include male animals to make an appropriate assessment on ‘sex differences’ in comparison to our ovariectomized-estrogen treated animals. IL-13 and CCL3 (MIP1a) were the cytokine and chemokine that were statistically higher in F-Sham-OVA treated animals compared to Male-Sham-OVA (Fig. 4B, 4G). IL-5 and IL-13 were decreased in OVX-E2-OVA animals compared to Female-Sham-OVA (Fig. 4A, 4B), and CCL11 trended downwards in these OVX-E2-OVA animals in comparison to F-Sham-OVA (Fig. 4F). BAL concentrations of CCL22, CC12, CCL11, CCL3 and IL-33 were equivalent in F-Sham-OVA and OVX-E2-OVA. Notably, CCL22, CCL12, and CCL11 were all increased in OVX-placebo-OVA treated animals compared to F-Sham-OVA, indicating that these allergic inflammatory proteins are regulated to some extent by ovary-derived hormones.
Circulating cytokines and chemokines are selectively altered with steady-state estrogen treatment. Serum was collected from OVA treated animals for detection of allergic-inflammatory cytokines and chemokines. Serum concentration of IL-13, CCL22, and CCL3 were increased in F-Sham-OVA mice compared to Male-Sham-OVA mice (Fig. 5B, 5C, 5D). In contrast to the findings in the BAL fluid, serum concentrations of CCL22 and CCL3 were increased in OVX-E2-OVA mice as compared with F-Sham-OVA mice (Fig. 5C, 5D). Although we saw increases in neutrophil numbers in OVX-Placebo-OVA mice compared to F-Sham-OVA in the previous studies (Fig. 3D), the neutrophil chemoattractant protein, KC, was not different between any of the groups.
Estrogen treatment reduces eosinophils in BAL of OVA sensitized and challenged ovarectomized mice. Eosinophils are regularly detected in circulation and in sputum of asthmatic patients (34, 35), as such eosinophils are thought of clinically as a determinant of asthma diagnosis and severity of disease (Fig. 6). In these studies, we prepared mice as in Fig. 1A, however these animals did not undergo methacholine challenges prior to characterizing eosinophil levels by flow cytometry. As expected, increased numbers of eosinophils were detected following OVA-allergen challenge in bronchoalveolar lavage fluids in comparison to saline treated (data not shown) animals. Eosinophils were determined at higher numbers in females compared to males in other studies (36, 37), however in the present studies statistical analysis showed no differences in eosinophil percentages or numbers in BAL between M-Sham-OVA and F-Sham OVA animals by flow cytometry. As with the previous data (Fig. 3), we again detected reduced numbers of eosinophils in OVX-E2-OVA treated animals in comparison to F-Sham-OVA. Together this indicates a suppressive effect on eosinophils, again, a well-excepted biomarker of asthma, with pharmacologically-delivered estrogen in the traditional OVA model.
Steady-state estrogen reduces airway T cells and B cells in BAL fluid collected from ovarectomized animals compared to intact female controls.
Total CD3 + T cells (7A, 7B) and CD19 + B cells (7C-7D) were assessed in the same treatment groups as before by flow cytometry. F-Sham-OVA did have higher numbers of CD3 + T cells detected in the BAL in comparison to M-Sham-OVA, however total CD3 + T cells were significantly reduced in OVA-E2-OVA mice compared to F-Sham-OVA mice. Again, this indicated that E2 is driving a suppressive program that is reducing the allergic inflammatory response to OVA. Similarly, CD19 + B cells (Fig. 7C, 7D) were reduced along with OVA-specific IgE (Fig. 7E) in the OVX-E2-OVA animals as compared with F-Sham-OVA mice. IgE levels are determinant of degree of allergic responses in the clinic and typically determined as another prototypical biomarker of asthmatic disease. These data revealed differential effects of female sex hormones whereby lung immune responses are decreased by estrogen.
Accumulation and IL-33-induced activation of lung ILC2 is reduced in OVX-E2-OVA mice compared to F-Sham-OVA mice.
ILC2 are important for allergic airway inflammation associated with asthma (38) and are increased in peripheral blood of asthmatic patients (39). ILC2 are predominantly responsible for IL-5 and IL-13 production following in vitro and in vivo stimulation of mouse tissues or cells with IL-33 (40). Importantly these cells have been shown to interact with type 2 helper T cells and to directly support eosinophil responses through their production of IL-5 (41–43). First, we identified ILC2 as LIN- CD127 + KLRG-1 + cells in the BAL and lungs following OVA challenge (Fig. 8A, 8B and 8C). Others have reported sex differences in the numbers of ILC2 in allergen challenge models, and we have intermittently determined ILC2 count and frequency differences in males versus females depending on the markers used for ILC2 specific detection. In these studies, we detected sex differences in counts and frequencies of ILC2, not only in BAL or airway-localized ILC2, but also in remaining lung tissue in M-Sham-OVA versus F-Sham-OVA animals (Fig. 8A, 8B). We assessed viability of the ILC2 during flow analysis using a fixable viability dye and confirmed that estrogen was not associated with lower viability in the BAL or lung ILC2 (Fig. 8D). Next lung ILC2 were sorted from each treatment group for ex vivo culture with and without IL-33. As in past studies we would expect higher allergic cytokine production in F-Sham-OVA mice as compared to M-Sham-OVA mice. As in our previous report, ILC2 were stimulated with IL-33 [10 ng/mL] for 3 days in culture to detect IL-5 and IL-13 production by ILC2. We add to that report by showing that ILC2 from F-Sham-OVA mice produced more CCL22 and CCL3 in response to IL-33 when compared to M-Sham-OVA. While the amount of IL-5, IL-13, CCL22, and CCL3 are higher in females compared to males, the excessive production of these cytokines in females was not due to estrogen stimulation, as we found reduced IL-5 and IL-13 production, and reductions in pro-inflammatory chemokines, CCL22 and CCL3, by lung ILC2 obtained 17β-E2 treated OVX mice (OVX-E2-OVA) compared to Female-Sham-OVA treated animals (Fig. 8E-8H). IL-5, IL-13, CCL22 and CCL3 were all comparable in OVX-Pl-OVA and F-Sham-OVA treated animals.