Glucocorticoids relieve collectin-driven suppression of apoptotic cell uptake in murine alveolar macrophages through downregulation of SIRPα

Abstract

The lung environment actively inhibits apoptotic cell (AC) uptake by alveolar macrophages (AMø) via lung collectin signaling through signal regulatory protein α (SIRPα). Even brief glucocorticoid treatment during maturation of human blood monocyte-derived or murine bone marrow-derived macrophages (Mø) increases their AC uptake. Whether glucocorticoids similarly impact differentiated tissue Mø and the mechanisms for this rapid response are unknown and important to define, given the widespread therapeutic use of inhaled glucocorticoids. We found that the glucocorticoid fluticasone rapidly and dose-dependently increased AC uptake by murine AMø without a requirement for protein synthesis. Fluticasone rapidly suppressed AMø expression of SIRPα mRNA and surface protein, and also activated a more delayed, translation-dependent upregulation of AC recognition receptors that was not required for the early increase in AC uptake. Consistent with a role for SIRPα suppression in rapid glucocorticoid action, murine peritoneal macrophages (PMø) that had not been exposed to lung collectins showed delayed but not rapid increase in AC uptake. However, pretreatment of PMø with the lung collectin surfactant protein D inhibited AC uptake and fluticasone treatment rapidly reversed this inhibition. Thus, glucocorticoids act not only by upregulating AC recognition receptors during Mø maturation but also via a novel rapid down-regulation of SIRPα expression by differentiated tissue Mø. Release of AMø from inhibition of AC uptake by lung collectins may in part explain the beneficial role of inhaled glucocorticoids in inflammatory lung diseases, especially emphysema, in which there is both increased lung parenchymal cell apoptosis and defective AC uptake by AMø.

INTRODUCTION

Apoptotic cell (AC) uptake by phagocytes, also termed efferocytosis ⁽¹⁾, is an essential process that promotes the resolution of injury and inflammation, facilitating tissue repair in the lung and throughout the body ⁽²⁾. Impaired AC uptake has been found in phagocytes from human subjects with cystic fibrosis, asthma, and chronic obstructive pulmonary disease (COPD) ^(3-6). Because defective AC clearance clearly contributes to autoimmunity in murine models ^{(7, 8)}, and because there is growing evidence that human emphysema may have an autoimmune component ^{(9, 10)}, potential therapies designed to bolster AC clearance have been proposed ⁽¹¹⁾. This issue is of considerable importance, as COPD is now the third leading cause of death in the United States, and has been projected by the World Health Organization to become the leading worldwide cause of death by mid 21^st century ⁽¹²⁾.

In seeming contradiction to the importance of AC clearance, the resident lung phagocyte, alveolar macrophages (AMø), bind and engulf AC less avidly than do other professional phagocytes ^(13-15). Reduced efferocytosis by AMø results in part from very restricted adhesion pathway usage and markedly decreased expression of PKC βII ^{(15, 16)}. Importantly, however, the normal lung environment actively suppresses the ability of AMø to ingest AC, due to the inhibitory action of specific surfactant proteins (SP-), the lung collectins SP-A and SP-D, through their interaction with signal regulatory protein α (SIRPα) (CD172a) ⁽¹⁷⁾. This inhibitory effect persists in vitro for days after AMø removal from the lung environment. Whether increasing the ability of AMø to ingest AC would have beneficial heath affects is unproven, but better understanding of the unique mechanisms by AMø interact with AC uptake is essential to guide the development of any such future therapies.

Several pharmacological treatments can increase AC uptake in vitro. Glucocorticoids (GC) have been shown to increase in vitro AC uptake by human blood-derived monocytes, macrophage cell lines, and, in a single report, human AMø ^{(5, 18)}. In human blood-derived monocytes, this increase is dependent on Mertk, increased Rac phosphorylation and altered surface sialylation ^(19-21). It is unclear whether glucocorticoids act via these mechanisms in other cell types such as AMø. Defining whether and how GC and other agents increase AC uptake by murine AMø is an essential step to develop murine models to test whether manipulating AC clearance improves lung health.

In this study, we report that the potent GC fluticasone increased AC uptake by murine AMø in a rapid, dose-dependent fashion through downregulation of SIRPα. Our data show a novel facet of GC action: a rapid decrease in the sensitivity of murine AMø to the collectin-rich, inhibitory environment of the lung, thus lifting tonic inhibition and increasing AC uptake.

MATERIALS AND METHODS

Mice

We purchased C57BL/6 mice from Charles River Laboratories. Mice were housed under specific pathogen-free conditions and used for experiments between 8 and 16 weeks of age. Animal care and experimentation were conducted in accordance with U.S. Department of Health and Human Services Guide for the Care and Use of Laboratory Animals and were approved by the Animal Use Committee at VA Ann Arbor Healthsystem.

Cell isolation and culture conditions

We isolated alveolar cells by bronchoalveolar lavage using 10 mL PBS containing 0.5 mM EDTA ⁽¹⁴⁾. AMø were adhesion purified from this population; non-adherent cells were discarded after 1.5 h of culture. Unstimulated peritoneal cells were isolated by peritoneal lavage using 7-10 mL PBS containing 0.5 mM EDTA, administered in 1-2 mL aliquots. PMø were adhesion purified from this depleted population; non-adherent cells were discarded after 45 min of culture.

All culture was performed in a 5% CO₂ environment at 37°C. During adhesion purification, phagocytosis and adhesion assays, Mø were cultured in 10% FBS, 1 mM sodium pyruvate, 0.5 mM 2-Mercaptoethanol, 1 mM HEPES, 100 u/ml penicillin, 100 u/ml streptomycin, 0.292 mg/ml L-Glutamine in RPMI. During all other treatments, Mø were cultured in AIM-V (GIBCO) without serum.

Induction of thymocyte apoptosis and SRBC opsonization

To induce apoptosis, in most experiments we treated single cell suspensions of murine thymocytes with 10 μM dexamethasone (Sigma) for 4.5 h. These conditions consistently produced 50-60% Annexin+, PI- thymocytes, as we have previously shown ⁽²²⁾. In selected experiments, thymocyte suspensions were UV-irradiated using a gel box (FOTO/UV 15, Fotodyne, Hartland, WI) on high power for 15 min, then were incubated a further 4 h to allow apoptosis to progress. SRBC (Colorado Serum Company, Denver, CO) were opsonized with anti-SRBC (Sigma) for 1 h, as previously described ⁽²²⁾.

Phagocytosis and adhesion assays

We quantified AC phagocytosis and adhesion as previously described ⁽¹⁴⁾. Briefly, Mø were cultured in 8-well Permavox chamber slides (Nunc, Thermo Fisher Scientific) at 10⁵ cells per well. We added target cells to Mø at a ratio of 10:1 to measure phagocytosis after 2 h or at 100:1 to measure adhesion after 20 min. Cells were stained with H&E and at least 200 macrophages were counted at 100x magnification.

Fluticasone propionate, budesonide, azithromycin dihydrate, and simvastatin (Sigma) were all rehydrated according to the manufacturers’ instructions and used at the concentrations described. Simvastatin was activated before use by treatment with NaOH in ethanol as described ⁽²³⁾.

For certain adhesion assays, we treated AMø with mAb against anti-CD11c (HL3; Becton Dickenson Immunocytometry (BD), Mountain View, CA), anti-CD18 (GAME-46; BD), hamster IgG (eBioscience, San Diego, CA), or rat IgG (BD), all at 5 μg/mL final concentration, for 30 min before the addition of AC. For certain phagocytosis assays, AMø were treated with 2 μg/mL anti-CD36 (JC63.1; Cayman, Ann Arbor, MI) or mouse IgA (eBioscience) for 30 min before addition of AC. For other phagocytosis assays, we pre-treated AMø with 5 μM cycloheximide (Sigma) for 1 h, then washed prior to the addition of fluticasone, simvastatin or azithromycin. For some experiments, we treated AMø with indomethacin (Sigma) for 30 min, then washed before addition of fluticasone. For other experiments PMø were treated with 20 μM recombinant murine SP-D (R&D, Minneapolis, MN) for 4 h, then washed prior to the addition of fluticasone. Following all treatments, Mø were washed with warm media before the addition of AC.

RNA isolation and real-time RT-PCR

We isolated total RNA from murine AMø and PMø using the RiboPure Kit (Ambion) and removed DNA contamination using the TURBO DNA-free kit (Ambion). cDNA was prepared from total RNA using the RETROscript kit (Ambion). All reagents and kits were used according to the manufacturer's instructions. We performed real-time RT-PCR using TaqMan Gene Expression Master Mix with TaqMan primer-probe sets from Applied Biosystems for GAPD (4352932), Axl (Mm00437221_m1), Mertk (Mm00434920_m1), SIRPα (Mm00455928_m1), LRP (Mm00464608_m1) and PPARδ (Mm00803184_m1).

Flow Cytometry

AMø were cultured in 48-well tissue culture plates at 10⁵ cells per well with 2 μM fluticasone, 10 μM simvastatin, 500 ng/mL azithromycin, or control media for 6-24 h. Cells were released from culture plates using the dissociation enzyme TrypLE (Invitrogen) and stained after Fc-block with a panel of fluorochrome-conjugated Ab as previously described ⁽²⁴⁾. The following anti-murine Abs were used (clone; source): CD45 (30-F11; BD), TCRβ (57-597; BD), CD19 (MCA1439F; AB Serotech), CD11c (N418; eBioscience), CD11b (M1/70; eBioscience) and SIRPα (P84; BD).

Experiments were performed on an LSR II flow cytometer (BD Bioscience, San Jose, CA), equipped with the following lasers (#) and their associated filter sets (letters): (1) 488 nm blue (primary laser), (a) 550 nm long-pass (LP), 530/30 nm short band-pass (SBP), (b) 685 nm LP, 695/40 nm SBP; (2) 405 nm violet laser, (c) 505 nm LP, 530/30 BP, (d) 450/50 PB; (3) 633 nm red HeNe laser, (e) 735 nm LP, 780/60 nm SBP, (f) 685 nm LP, 710/50 nm SBP, (g) 660/20 nm BP; and (4) 561 nm yellow-green laser, (h) 735 nm LP, 780/60 SBP, (i) 685 nm LP 710/50 nm SBP, (j) 635 nm LP, 610/20 nm SBP, (k) 581/15 nm BP. In all experiments, we used isotype-matched controls, and collected a minimum of 10,000 CD45+ viable events per sample. Data were collected on an HP XW4300 Workstation (Hewlett-Packard, Palo Alto, CA) using FACSDiva software (version 6.1.3; BD Biosciences) with automatic compensation and were analyzed using FlowJo software (Tree Star, Ashland, OR) on an Intel iMac computer (Apple, Cupertino, CA).

Statistical analyses

We calculated significance using one-way ANOVA with Bonferroni post-hoc testing or using Student t-test where appropriate using GraphPad Prism4 (La Jolla, CA) on an Intel iMac computer. Results were considered significant at p<0.05.

RESULTS

Potent GC rapidly increases murine AMø uptake and binding of AC

To study the effect of GC used clinically as inhaled corticosteroids (ICS) on AC uptake by murine AMø, we first performed in vitro phagocytosis assays following treatment with the potent GC fluticasone (Fig. 1A; Supplemental Figs. S1A, S1B). Pre-treatment with fluticasone significantly increased the ability of murine AMø to ingest AC after only 3 h, with peak effect by 6 h (Figs. 1B, 1C). The magnitude of the effect was dose-responsive, increasing with higher doses of fluticasone; significance could be seen at 2 nM (Fig. 1D, 1E). Fluticasone treatment also increased AMø uptake of UV-killed thymocytes (Supplemental Figs. S1C, S1D), implying that the effect did not depend on the method used to induce apoptosis. This pro-efferocytic effect was not restricted to fluticasone, as increased AMø AC uptake could also be seen following treatment with budesonide, another potent GC used clinically (percent phagocytosis: untreated 16.5 ± 3.4% vs. budesonide 38.6 ± 4.0%; phagocytic index: untreated 0.19 ± 0.04 vs. budesonide 0.51 ± 0.09; mean ± SE of 3 mice assayed individually in two independent experiments; both measurements significant, p<0.05 by one-way ANOVA with Bonferroni post-hoc testing). In contrast, AC uptake by resident murine PMø did not increase on fluticasone treatment (Supplemental Figs. S2A, S4B), even on treatment up to 6 h (data not show). Additionally, fluticasone did not increase Fc-mediated clearance of IgG-opsonized SRBC (Supplemental Figs. S2C, S4D) or of 4 μm latex microspheres (data not shown) by murine AMø.

Figure 1.

Fluticasone rapidly and specifically increases uptake and binding of AC by murine AMø. (A-E) AC uptake. Adherence-purified AMø from normal C57 BL/6 mice were treated in chamber slides with fluticasone (2 nM unless indicated) for 0-6 h, then AC were added at a 10:1 ratio for 2 h. Slides were washed and stained using H&E, then ingested AC were counted at 100X magnification under oil. A. Graphic timeline of a phagocytosis assay. B, C. Kinetics of GC-augmented AC uptake. D, E. Dose-response of GC-augmented AC uptake. (F-J) AC binding. Adherence-purified AMø from normal C57 BL/6 mice were treated in chamber slides with fluticasone (2 nM unless indicated) for 0-6 h, then AC were added at a 100:1 ratio for 20 min. Slides were washed and stained using H&E, then surface bound AC were counted at 100X magnification under oil. F. Graphic timeline of a binding assay. G, H. Kinetics of GC-augmented AC binding. I, J. Dose-response of GC-augmented AC binding. Data are mean ± SE of 5-8 mice assayed individually in at least two independent experiments per condition. **, statistically significant for untreated, p<0.01 by one-way ANOVA with Bonferroni post-hoc testing.

To study the effect of GC on murine AMø binding of AC, we next performed adhesion assays (Fig. 1F). Similar to the effect on AC engulfment, 4 h treatment with fluticasone significantly increased the ability of murine AMø to bind AC, with the effect peaking by 6 h (Fig. 1G, 1H). The magnitude of the effect was also dose-responsive; significance could be seen at doses above 200 pM (Fig. 1I, 1J). To determine if fluticasone initiated novel adhesion pathways, we pre-treated AMø with mAbs to block CD11c and CD18, which we have previously shown mediate the majority of adhesion of AC to murine AMø ⁽¹⁵⁾. Blocking either integrin subunit reduced AMø binding to AC, regardless of treatment with fluticasone (Supplemental Fig. S3). In contrast, similar to the lack of effect on engulfment, fluticasone treatment did not increase PMø binding to AC regardless of fluticasone dose (2 pM-2 μM) or duration of treatment to 6 h (data not shown).

Thus, GC pretreatment is associated with rapidly increased AC binding and engulfment that is specific to AMø and not observed in a resting, fully-differentiated tissue Mø from another mucosal surface. Further, the ability to increase AC uptake appears to be a class effect of potent GC, which, however, does not alter phagocytosis by murine AMø of other types of particles.

Fluticasone initiates reprogramming towards a pro-clearance phenotype and increases AC uptake without a requirement for new protein synthesis

GC alter expression of large numbers of target genes, for the most part via the specific glucocorticoid receptor GRα, a member of the ligand-regulated family of nuclear receptors ⁽²⁵⁾, but also by incompletely understood translation-independent mechanisms ^{(26, 27)}. To begin to define how fluticasone upregulates murine AMø uptake of AC, we assessed the expression of several genes known to be involved in AC clearance, including Mertk and Axl, members of the TAM family of receptor tyrosine kinases ⁽²⁸⁾, CD91/LRP ⁽²⁹⁾ and the negative regulator SIRPα ⁽¹⁷⁾. We also examined mRNA expression of the nuclear receptor PPARδ, a positive regulator of the expression of opsonins involved in bridging AC and of Mø surface receptors including Mertk ⁽³⁰⁾. Within 3 h of fluticasone treatment, Mertk mRNA significantly increased, whereas SIRPα transcripts significantly decreased (Fig. 2A). These changes are consistent with an induction by GC of pro-clearance AMø phenotype, as previously described for human monocytes ⁽³¹⁾. Transcripts for Axl, LRP and PPARδ did not change during this period of fluticasone treatment.

Figure 2.

Fluticasone rapidly downregulates SIRPα and increases efferocytosis without a requirement for new protein synthesis. A, Murine AMø were treated with 2 nM fluticasone for 0, 1, 3 or 6 h. RNA was collected at each time point and analyzed by real-time RT-PCR with GAPDH as the housekeeping gene; results are displayed as fold increase from untreated. B, Murine AMø were pre-treated with 5 μM cycloheximide for 1 h followed by 2 μM fluticasone for 5 h, then AC were added at a 10:1 ratio for 2 h. Slides were washed and stained using H&E, then ingested AC were counted at 100X magnification under oil. C, D. Surface SIRPα protein. Murine AMø treated with 2μM fluticasone for 6 or 24 h, then analyzed by flow cytometry for surface expression of SIRPα. Cells shown are gated CD45⁺CD19^-TCRβ^-. C. Representative dot plot. D. Average percent of CD11c⁺SIRPα^- cells within gated CD11c⁺ population. Data are mean ± SE of 5-7 individual mice assayed individually in at least two independent experiments per condition. *, statistically significant from untreated, p<0.05 and **, statistically significant from untreated, p<0.01 by one-way ANOVA with Bonferroni post-hoc testing.

These mRNA changes not withstanding, the rapid kinetics of increased AC uptake in murine AMø led us to postulate that fluticasone may act on a short-lived inhibitor. To test that possibility, we blocked new protein synthesis using cycloheximide. Treatment of AMø with cycloheximide prior to an additional 5 h fluticasone treatment did not abrogate the increase in AC uptake (Fig. 2B). Thus, although Mertk and likely other AC recognition molecules were significantly increased by fluticasone treatment, translation-dependent increases in Mertk or any other protein are not required for the rapid (< 5 h) effect of fluticasone.

Fluticasone decreases protein expression of SIRPα

To test the significance of the observed fluticasone-induced gene repression of SIRPα (Fig. 2A), we examined protein expression of SIRPα. Using flow cytometry, we found that surface expression of SIRPα was decreased within 6 h of fluticasone treatment, with statistical significance reached by 24 h (Fig. 2C, 2D).

We also tested the involvement of several pathways that have been implicated in AC uptake by other types of tissue Mø, using pharmacological inhibitors or blocking mAbs. Neither fluticasone-treated AMø, nor as we have previously described, untreated murine AMø require CD36, alphaV integrin or autocrine prostanoid signaling for AC uptake (Supplemental Figs. S4A-S4F). These results complement those in which we blocked CD11c and CD18 (Supplemental Fig. S3) in indicating that GC-augmented AC uptake does not require engagement of new adhesion pathways but instead appears to result from increased efficiency of the same pathways used in the resting state.

Azithromycin but not simvastatin has additive effects on GC-augmented efferocytosis

In addition to GC, AC uptake is known to be increased by other commonly prescribed pharmaceuticals including statins and macrolides ^(32-34). To study interactions between these medications, we treated murine AMø with combinations of fluticasone, simvastatin and azithromycin, then assessed the effect on AC engulfment. Treatment with simvastatin or fluticasone alone each increased AC uptake, but the combination had no additive effect (Fig. 3A, 3B). By contrast, treatment of AMø with azithromycin and fluticasone was additive, resulting in near doubling of uptake capacity over either treatment alone (Fig. 3C, D).

Figure 3.

Azithromycin but not simvastatin has additive effects on efferocytosis by murine AMø. (A-D) Affect of multi-agent treatment on efferocytosis. Murine AMø were treated with Murine AMø were treated with 500 ng/mL azithromycin, 10 μM simvastatin or media alone. After 18 h, 2 μM fluticasone was added for a further 6 h, then AC were added at a 10:1 ratio for 2 h. Slides were washed and stained using H&E, then ingested AC were counted at 100X magnification under oil. A, B. Simvastatin and Fluticasone. C, D. Azithromycin and Fluticasone. Data are presented as the mean ± SE of seven mice assayed individually in three independent experiments. **, statistically significant than fluticasone alone, p<0.01 by one-way ANOVA with Bonferroni post-hoc testing.

Simvastatin affects AC uptake via the SIRPα pathway and mechanisms that require new protein translation

The lack of additive effect between simvastatin and fluticasone suggested that these agents likely affect AC uptake through the same molecular pathway. This possibility is supported by previous evidence that statin treatment decreases localization to the plasma membrane of RhoA, a downstream effector of SIRPα signaling; because RhoA antagonizes the essential action of Rac-1 on AC uptake, the net effect is increased efferocytosis ⁽²³⁾. We used flow cytometry to test whether either simvastatin or azithromycin also affected SIRPα surface expression. Azithromycin did not change SIRPα expression compared to untreated AMø, but simvastatin significantly decreased SIRPα surface expression after 24 h (Fig. 4A, 4B). However, in contrast to fluticasone, simvastatin did not change SIRPα mRNA levels (data not shown).

Figure 4.

Simvastatin downregulates SIRPα expression while azithromycin does not. A, B. Surface SIRPα protein. Murine AMø treated with 10 μM simvastatin or 500 ng/mL azithromycin for 24 h, then analyzed by flow cytometry for surface expression of SIRPα. Cells shown are gated CD45⁺CD19^-TCRβ^-. A. Representative dot plot. B. Average percent of CD11c⁺SIRPα^- cells within gated CD11c⁺ population. *, statistically significant from other conditions, p<0.05 by one-way ANOVA with Bonferroni post-hoc testing. C. Murine AMø were pre-treated with 5 μM cycloheximide (CHX) for 1 h followed by 10 μM simvastatin or 500 ng/mL azithromycin for 24 h, then AC were added at a 10:1 ratio for 2 h. Slides were washed and stained using H&E, then ingested AC were counted at 100X magnification under oil. Data in B, C are mean ± SE of 5-7 mice assayed individually in at least two independent experiments. **, statistically significant from no cyclohexamide, p<0.01 by one-way ANOVA with Bonferroni post-hoc testing.

To further differentiate possible mechanisms of action, we next blocked induction of new protein synthesis by these two agents. Treatment of murine AMø with cycloheximide prior to 24 h of treatment with simvastatin or azithromycin blocked the ability of either agent to increase AC uptake over that of untreated AMø (Fig. 4C). These results indicate that, unlike fluticasone, both simvastatin and azithromycin do require new protein synthesis to increase AC uptake in AMø.

SP-D treatment inhibits AC uptake by PMø which is reversed with fluticasone treatment

The inhibitory effect of SIRPα on AC uptake by murine AMø is tonically maintained by constant exposure in the alveolar space to high concentrations of the lung collectins SP-A and SP-D ⁽¹⁷⁾. By contrast, although PMø express surface SIRPα ⁽³⁵⁾, they receive limited exposure to lung collectins. These considerations led us to hypothesize that the absence of GC-augmented AC uptake by PMø (Supplemental Fig. S2A, S2B) might reflect limited activation of SIRPα in the peritoneal cavity, which unlike the alveolar spaces, do not contain significant concentrations of SP-A or SP-D. To test this possibility, we first used flow cytometry to test whether SIRPα expression on resident murine PMø was altered by fluticasone treatment in vitro. Similar to AMø, 24 h of fluticasone treatment significantly decreased PMø expression of SIRPα surface protein, whether expressed as percentage positive relative to isotype control or mean fluorescence index (MFI) (Fig. 5A-C). Next, by pre-incubating PMø with the SIRPα ligand SP-D, we investigated whether activation of SIRPα could repress AC uptake by murine PMø. SP-D significantly inhibited AC uptake by PMø within 4 h (Fig. 5D). Finally, we tested whether fluticasone treatment could rescue decreased PMø AC uptake following SP-D treatment. Although treatment with SP-D alone again significantly inhibited AC uptake, subsequent incubation with fluticasone for 5 h completely reversed this inhibition (Fig. 5D). These results provide a proof-of-concept that the rapid effect of GC on AC uptake by tissue Mø is mediated by release of collectin-induced repression acting via surface SIRPα expression (Fig. 6), and does not depend on GC-modification of other features of the AMø phenotype.

Figure 5.

SP-D activates SIRPα pathway in PMø and makes PMø sensitive to fluticasone-driven increase in AC clearance. (A-C) Surface SIRPα protein. Murine PMø were treated with 2 μM fluticasone for 6 or 24 h, then analyzed by flow cytometry for surface expression of SIRPα. Cells shown are gated CD45⁺CD19^-TCRβ^-. A. Representative dot plot. B. Average percent of CD11b⁺SIRPα^- cells within gated CD11b⁺ population. C, Average MFI of SIRPα on gated CD11b⁺ cells. D, Fluticasone rescues SP-D inhibition of AC uptake. Murine PMø were treated with 25 μg/mL SP-D for 4 h, followed by control media or 2 μM fluticasone for 5 h, then AC were added at a 10:1 ratio for 2 h. Slides were washed and stained using H&E, then ingested AC were counted at 100X magnification under oil. Data are mean ± SE of 5-8 mice assayed individually in at least two independent experiments per condition. **, statistically significant, p<0.01 by one-way ANOVA with Bonferroni post-hoc testing.

Figure 6.

Model of GC regulation of SIRPα-mediated control of murine AMø efferocytosis. A. In untreated AMø, which express high amounts of SIRPα, lung collectins SP-D and SP-A (not shown) signal constitutively through SIRPα, activating SHP-1 and leading to downstream activation of RhoA. By inhibiting Rac-dependent mobilization of actin, the lung collectins tonically impede efficient uptake of AC by AMø, even though SP-A and SP-D can also bind AC. B. Treatment with fluticasone (triangles) reduces SIRPα surface expression, in part via transrepression of SIRPα by ligand-occupied GRα homodimers (brackets). The consequent decreased activation of SHP-1 relieves inhibition of Rac, permitting efficient AC uptake. Based on data in the current study, plus previously published data ^{(17, 36, 37, 50)}.

DISCUSSION

The results of this study identify downregulation on AMø of the inhibitory receptor SIRPα, which releases them from tonic inhibition by lung collectins, as a novel mechanism by which clinically-relevant potent GC rapidly increase AMø uptake of AC. Using primary murine AMø, we found that treatment with fluticasone or budesonide increased both binding and uptake of AC within 2-3 h, without apparent induction of novel adhesive pathways. The effect did not require new protein synthesis, although its magnitude continued to increase through 5-6 h in association with significantly increased Mertk surface expression. Reduced SIRPα surface expression would be fully anticipated to prevent the previously described downstream activation of RhoA and hence Rho Kinase to inhibit Rac ⁽¹⁷⁾, on which AC ingestion depends crucially ^{(36, 37)}. Although fluticasone treatment of resting murine PMø did not show the same effect on AC uptake, brief treatment with SP-D induced a significant reduction in their AC uptake that was rapidly reversed by fluticasone. These findings emphasize the importance of the unique lung environment and thus, more globally, of studying primary phagocytes isolated from mucosal surfaces in attempting to understand host defense of specific organs.

The well-described distinctive characteristics of resident AMø include a low capacity for AC binding and uptake ^(13-15). This feature may be of evolutionary value by preventing AC-induced immunosuppression, thus maintaining AMø as sentinel immune responders. Uptake of AC activates multiple anti-inflammatory pathways within phagocytes, notably through upregulation of SOCS1 and SOCS3 and subsequent inhibition of Jak-STAT signaling ⁽³⁸⁾. The in vivo relevance of such AC-induced suppression on host defense has been shown in a murine model, in which intrapulmonary administration of AC reduced phagocytosis and killing of Streptococcus pneumoniae and impaired leukocyte recruitment through PGE₂-EP2-dependent signaling ⁽³⁹⁾. Conversely, the induction of lupus-like autoimmunity in mice by deletions of genes including C1q ⁽⁴⁰⁾, MFG-E8 ⁽⁴¹⁾, α_v integrins ^{(42, 43)} and the TAM receptors ^{(7, 44-46)} argue for concurrent evolutionary pressures to fine-tune AC clearance.

By defining a rapid, translation-independent effect on fully differentiated tissue Mø, these results extend previously described mechanisms of GC actions during Mø differentiation from precursors ^{(18-21, 47, 48)}. In contrast to the early SIRPα–dependent mechanism we show in mature AMø, results in those studies required new protein synthesis and more prolonged treatment, maximal when GC was added 3-5 days earlier. Thus, these studies were informative of the effects of systemic steroid treatments on Mø precursors, but not directly relevant to the question about how ICS might impact functions of resident AMø. Similarly, two groups have used microarray technology to define the effects of GC on gene regulation during in vitro differentiation of human monocytes ^{(31, 49)}. They found alterations in a range of molecules plausibly involved in AC clearance, including integrins, scavenger receptors, receptor tyrosine kinases, bridging molecules, molecules associated with engulfment, nuclear receptors, and members of the interferon regulatory family genes. Our finding of upregulation of Mertk transcripts is compatible with the initiation by GC of such a more prolonged multi-gene program in AMø, but the full range of such more delayed effects will require further study.

Our findings agree with and follow directly from recent publications that identified the importance of the alveolar environment to maintain a carefully regulated AMø phenotype ^{(17, 50)}, particularly in terms of AC uptake. We believe that this line of investigation highlights the ability for elegant control of AMø function by altered expression of key receptors rather than by disruption of this fragile environment. SP-A and SP-D serve at least three functions in the alveolar space: modulating basal AMø signaling in the absence of AC; binding directly to AC to increase their uptake; and as opsonins of multiple lung pathogens ⁽⁵¹⁾. Transgenic mice deficient in SP-A or SP-D have increased susceptibility to multiple viral, bacterial and fungal infections (reviewed in ⁽⁵²⁾). Deficiency of SP-D can also lead to chronic low-grade pulmonary inflammation and fibrosis ⁽⁵³⁾. We speculate that regulating SP-A and SP-D signaling by altering SIRPα expression on AMø, rather than directly by modulation of lung collectin levels, permits the continuation of other signaling and particularly opsonic functions of the lung collectins.

Increased AC uptake by inflammatory Mø in the alveolar spaces occurs in mice treated with intratracheal LPS ⁽¹⁷⁾ and has been shown in various phagocytes in vitro using a number of pharmacological agents including GC, statins and macrolides. To our knowledge, this is the first report describing how simultaneous treatment with these drugs, commonly prescribed to individuals with respiratory disease, affects AC uptake in any cell type. The lack of additive effect between simvastatin and fluticasone is congruent with a shared mechanism of action: inhibition of RhoA leading to increased Rac activity. Of more interest is the additive effect of azithromycin and fluticasone on AC uptake, especially given the recent demonstration that azithromycin reduces the frequency of acute exacerbations of COPD ⁽⁵⁴⁾. The mechanism for the positive effect of azithromycin on AC uptake remains undefined and will require considerable extra investigation; our results imply that azithromycin does not act on RhoA. Decreased AC uptake has been found in AMø from individuals with COPD ⁽⁶⁾ and asthma ⁽⁵⁾ when compared with healthy controls, which has prompted speculation that poor AC clearance may be contributing to various forms of inflammatory lung diseases. Our work does not address this hypothesis, but does identify a novel additive interaction between fluticasone and azithromycin that produces a robust increase in AC uptake and may be useful in future therapy.

The finding that SP-D can activate the pre-existing high levels of SIRPa on PMø merits discussion in relationship to acute lung injury, in which plasma concentrations of SP-A and SP-D increase significantly and correlate with clinical outcomes ^(55-57). Sepsis, the most common antecedent of acute lung injury, is associated both with massive apoptosis of circulating lymphocytes and with a delayed immunocompromised state. Results in murine models suggest that the first of these observations may explain the second, via the immunosuppressive effect of AC uptake on innate immunity ^{(58, 59)}. Although our results strongly imply that SIRPα signaling is not active in resident PMø harvested from untreated mice, they do suggest that increased circulating levels of lung collectins could contribute to reduced efferocytosis through the body during acute lung injury. Moreover, signaling via SIRPa also suppresses Mø phagocytosis mediated by FcγR and complement receptors ^{(60, 61)}. Thus, the possibility should be investigated that circulating SP-A and SP-D are not only biomarkers of severity during acute lung injury, but might also contribute to systemic immunosuppression that leads to the frequent superinfections that characterize this condition.

Defining how GC affect AMø is particularly important as a result of the widespread prescription of ICS for the treatment of lung disease. Multiple clinical trials have noted that receiving ICS is associated with increased hospitalization of COPD patients with pneumonia, compared to COPD patients receiving non-steroidal treatment, suggesting ICS treatment results in increased susceptibility to infection ⁽⁶²⁾. In contrast, mice pre-treated with fluticasone had significantly reduced lung bacterial burdens 24 and 48 h after Streptococcus pneumoniae infection, suggesting that fluticasone is protective and increases bacterial clearance ⁽⁶³⁾. Our findings in murine AMø and previous finding in human AMø strongly suggest that GC treatment, by increasing AC uptake, will enhance AC-mediated immunosuppression of AMø. It would be interesting to test whether increased immunosuppression from AC within the lung may explain these opposing results between COPD patients and model systems regarding ICS use and pneumonia infection, particularly for emphysema patients where lung destruction generates large numbers of AC. Our finding that murine AMø efferocytosis is increased following GC, azithromycin or simvastatin treatment demonstrates that mice provide an appropriate model system with which to predict consequences of pharmacologically-augmented AC clearance on human lung disease.

In summary, to our knowledge, our study demonstrates for the first time that GC increase AC uptake by murine AMø. We provide evidence that this rapid increase is caused by disruption of collectin-SIRPα signaling through downregulation of SIRPα transcript and surface protein, a novel GC mechanism. Finally, we demonstrate that regulation of AC uptake by SIRPα is not restricted to AMø and can be activated in PMø following exposure to SP-D.

Non-standard abbreviations

AMø

alveolar macrophage

PMø

peritoneal macrophage

GC

glucocorticoid(s)

SP-A

surfactant protein A

SP-D

surfactant protein D

COPD

chronic obstructive pulmonary disease

ICS

inhaled corticosteroids

MFI

mean fluorescence index

SIRPα

signal regulatory protein α