Bacterial regulation of macrophage bacterial recognition receptors in COPD are differentially modified by budesonide and fluticasone propionate

Abstract

Rationale: Patients with COPD have an increased risk for community-acquired pneumonia, which is further increased by inhaled corticosteroids.

Objective: To assess effects of the corticosteroids, budesonide and fluticasone propionate, on macrophage bacterial responses in COPD.

Methods: Monocyte-derived macrophages (MDMs) generated from blood monocytes from 10 non-smoker controls (NoS), 20 smokers without COPD (Sm), and 40 subjects with moderate to severe COPD (21 ex-smokers (COPD-ES) and 19 current smokers (COPD-S)) were pre-treated with budesonide or fluticasone (10 nM-1 μM) and challenged with live non-typeable Haemophilus influenzae (NTHI) or Streptococcus pneumoniae (SP). Cell surface bacterial recognition receptor expression (flow cytometry) and cytokine release (bead array) were analyzed.

Results: NTHI and SP reduced bacterial recognition receptor expression on MDMs from COPD and Sm, but not NoS (except TLR4). SR-AI and MARCO were reduced by both NTHI and SP, whereas other receptors by either NTHI or SP. Among COPD subjects, COPD-ES demonstrated a greater number of reductions as compared to COPD-S. NTHI reduced SR-AI, MARCO, CD11b, CD35 and CD206 in COPD-ES while only SR-AI and CD11b in COPD-S. SP reduced SRA-1, CD1d, TLR2 and TLR4 in both COPD-ES and COPD-S, and reduced MARCO and CD93 only in COPD-ES. All receptors reduced in COPD by NTHI and most by SP, were also reduced in Sm. Budesonide counteracted the receptor reductions induced by both NTHI (CD206 p = 0.03, MARCO p = 0.08) and SP (SR-AI p = 0.02) in COPD-ES. Fluticasone counteracted only SP-induced reductions in TLR2 (p = 0.008 COPD-ES and p = 0.04 COPD-S) and TLR4 (p = 0.02 COPD-ES). Cytokine release was equivalently reduced by both corticosteroids.

Conclusions: Reduction in macrophage bacterial recognition receptors during bacterial exposure could provide a mechanism for the increased pneumonia risk in COPD. Differential effects of budesonide and fluticasone propionate on macrophage bacterial recognition receptor expression may contribute to the higher pneumonia incidence reported with fluticasone propionate.

Fig 1. Cell surface bacterial recognition receptor expression at baseline was greater on MDMs from subjects with COPD and healthy smokers as compared to non-smokers.

Analysis by one-way ANOVA for receptor MFI (mean fluorescence intensity), representative of cell surface receptor density, presented as an aligned dot-plot with thin black bar at the mean, flanking short grey bars delineating the SD. Thick black horizontal bars above the entire graph mark the ANOVA results of statistical significance. Thick dark grey horizontal bars start and end at the comparisons of statistical significance identified in multiple comparison analysis (A). The bottom row (B) presents the receptors that did not have significant differences in cell surface expression between groups at baseline. Holm-Sidak correction was performed for multiple comparisons. COPD-ES = COPD ex-smoker, COPD-S = COPD current smoker, Sm = smoker without COPD, NoS = non-smoker healthy control, SR-AI = Scavenger receptor A-I, MARCO = Macrophage receptor with collagenous structure, TLR = Toll like receptor. S2 Fig demonstrates the Fig 1A differences in baseline receptor expression in heat map format.

Fig 2. NTHI and SP-mediated reductions of bacterial recognition receptor expression on MDMs from subjects with COPD.

Analysis by paired T-test of log-transformed receptor MFI (mean fluorescence intensity) representative of cell surface receptor density, as compared to baseline receptor expression. Boxes around SR-AI and MARCO denote the only receptors reduced by both NTHI and SP. Receptor expression is presented as an aligned dot-plot with thin black bar at the mean, flanking short grey bars delineating the SD. Effect size calculated by Cohen’s-d calculation. S3 Fig provides additional heat map format of statistically significant NTHI- and SP-induced reductions in bacterial recognition receptors.

Fig 3. Budesonide increased bacterial recognition receptor expression on MDMs from COPD ex-smokers for both NTHI and SP with greatest receptor upregulation at 10 nM, reflecting an inverse concentration-response relationship.

Analysis of log-transformed data on receptor MFI (mean fluorescence intensity). The p-values were calculated by paired T-test, for effect of budesonide (BUD) at 10 nM on SP- or NTHI mediated reduction of receptor expression on MDMs from COPD ex-smokers (COPD-ES) and COPD current smokers (COPD-S). Effect size for BUD effect on MDM receptor expression at 10 nM was calculated by Cohen’s-d calculation. Italicized p-value approached statistical significance, and is reported due to the large effect size. Light grey p-values indicate a statistically not-significant result (NS), and subsequent effect size and concentration-response values are shown as NS (A, Table). Receptor expression is presented as an aligned dot-plot with thin black bar at the mean, flanking short grey bars delineating the SD (B, Figures). Significant p-values for the BUD effect on receptor expression at 10 nM are noted above the thick, short black bar. Linear trend analysis (denoted by thick dark grey line) was done for concentration-response relationship between 10 nM, 100 nM and 1 μM of BUD calculated by repeated measures one-way ANOVA. Significant reductions in baseline receptor expression by NTHI or SP are noted by an asterisk (*), representing significant data presented in manuscript Fig 2, and S3 Fig. BUD increased expression of CD93 in NTHI-exposed MDMs although CD93 expression was not reduced by NTHI. NS = not statistically significant.

Fig 4. Fluticasone propionate increased bacterial recognition receptor expression on MDMs from subjects with COPD for SP only, with greatest receptor upregulation at 10 nM, reflecting an inverse concentration-response relationship.

Analysis of log-transformed data on receptor MFI (mean fluorescence intensity). The p-values were calculated by paired T-test, assessing the effect of fluticasone propionate (FP) at 10 nM on SP-induced reduction of receptor expression on MDMs from COPD ex-smokers (COPD-ES) and COPD current smokers (COPD-S). Effect size for FP effect on MDM receptor expression at 10 nM was calculated by Cohen’s-d calculation (A, Table). Only CD16 demonstrated a significant concentration–response relationship (inverse) to FP. Receptor expression is presented as an aligned dot-plot with thin black bar at the mean, flanking short grey bars delineating the SD (B, Figures). Significant p-values for fluticasone effect on receptor expression at 10 nM are noted above the thick short black bars. Linear trend analysis (denoted by thick dark grey line) was done for concentration-response relationship between 10 nM, 100 nM and 1 μM of FP calculated by repeated measures one-way ANOVA. Significant reductions in baseline receptor expression by SP are noted by an asterisk (*), representing significant data presented in manuscript Fig 2 and S3 Fig. Of note, FP increased expression of CD16 in SP-exposed MDMs although CD16 expression was not reduced by SP. NS = not statistically significant.

Fig 5. Budesonide and fluticasone propionate both effectively reduced cytokine release from COPD MDMs exposed to NTHI and SP.

Analysis of log-transformed data on cytokine concentrations (pg/mL) in MDM-bacteria-ICS co-culture supernatants. The p-values were calculated by paired T-test, assessing the effect of budesonide (BUD) or fluticasone propionate (FP) at 10 nM on NTHI- or SP-induced cytokine release by MDMs from COPD ex-smokers (COPD-ES) and COPD current smokers (COPD-S). Concentration of cytokine released at each concentration of ICS is presented as an aligned dot-plot with thin black bar at the mean, flanking short grey bars delineating the SD. Significant p-values for ICS effect on cytokine release (at 10 nM) and effect size are noted below the thick solid black bars. Only FP inhibition of NTHI-stimulated release of IL-1β from COPD-ES and COPD-S, and NTHI-stimulated release of IL-8 in COPD-ES demonstrated a statistically significant concentration-response relationship (inverse, greatest suppression at 10 nM) and is denoted by a dashed line over the concentrations with p-value above line. Extent of inhibition of cytokine release was not significantly different between BUD and FP and did not reach EC50 value for any cytokine. Effect size calculated at 10 nM by Cohen’s-d calculation.