ABCG1, a member of the ATP-binding cassette transporter superfamily, is highly expressed in multiple cells of the lung. Loss of ABCG1 results in severe pulmonary lipidosis in mice, with massive deposition of cholesterol in both alveolar macrophages and type 2 cells and the accumulation of excessive surfactant phospholipids. These observations are consistent with ABCG1 controlling cellular sterol metabolism. Herein, we report on the progressive and chronic inflammatory process that accompanies the lipidosis in the lungs of Abcg1−/− mice. Compared with wild-type animals, the lungs of aged chow-fed mice deficient in ABCG1 show distinctive signs of inflammation that include macrophage accumulation, lymphocytic infiltration, hemorrhage, eosinophilic crystals, and elevated levels of numerous cytokines and cytokine receptors. Analysis of bronchoalveolar lavages obtained from Abcg1−/− mice revealed elevated numbers of foamy macrophages and leukocytes and the presence of multiple markers of inflammation including crystals of chitinase-3-like proteins. These data suggest that cholesterol and/or cholesterol metabolites that accumulate in Abcg1−/− lungs can trigger inflammatory signaling pathways. Consistent with this hypothesis, the expression of a number of cytokines was found to be significantly increased following increased cholesterol delivery to either primary peritoneal macrophages or Raw264.7 cells. Finally, cholesterol loading of primary mouse macrophages induced cytokine mRNAs to higher levels in Abcg1−/−, as compared with wild-type cells. These results demonstrate that ABCG1 plays critical roles in pulmonary homeostasis, balancing both lipid/cholesterol metabolism and inflammatory responses.
Inflammation is a complex biological response to tissue injury and/or infection that facilitates tissue repair and the removal of the causal stimuli. Whereas acute inflammation is generally considered a physiological process that resolves in a relative short period of time, chronic inflammation represents a prolonged pathological condition in which simultaneous destruction and healing of tissues eventually results in further tissue damage (1, 2). Chronic inflammation is also characterized by the presence of infiltrated mononuclear cells that include macrophages, lymphocytes, and plasma cells (1, 2).
The continuous exposure of pulmonary alveolar spaces to inhaled foreign Ags (bacteria, viruses, fungi, and particles of varying size) could potentially result in an excessive and violent inflammatory response. Thus, precise mechanisms that balance pro- and anti-inflammatory responses exist in the lung to ensure an appropriate response to environmental agents. Anti-inflammatory strategies that suppress the immune response in the lung include anergization of T cells by type 2 pneumocytes (3), disruption of LPS-stimulated cytokine secretion in alveolar macrophages by specific surfactant phospholipids or proteins (4, 5, 6), and neutralization of different pathogens by surfactant collectins SP-A and SP-D (7, 8, 9, 10). Recently, Lian et al. (11, 12) suggested that the neutral lipid content of the lungs can also modulate pulmonary inflammatory responses. These authors demonstrated that disruption of triglyceride and cholesteryl ester metabolism in alveolar macrophages, as a result of lysosomal acid lipase deficiency, results in respiratory inflammation, tissue remodeling, and emphysema (11, 12). Together, these results indicate that factors governing surfactant composition and neutral lipid metabolism in alveolar cells might be associated with chronic pulmonary inflammation resulting in asthma, allergy, chronic obstructive pulmonary disease (COPD),3 or lung cancer.
ABCG1 is a member of the ATP-binding cassette (ABC) family of transmembrane transporters (reviewed in Ref. 13). Earlier studies demonstrated that Abcg1 mRNA levels are highly induced following incubation of human and murine macrophages with either oxidized or acetylated low density lipoprotein (LDL) or with agonists that activate the nuclear liver X receptor (LXR) (14, 15). More recent studies have shown that ABCG1 protein facilitates cholesterol efflux from cells to a variety of exogenous lipid acceptors that include mature high density lipoprotein (HDL) particles, phospholipid vesicles, and phospholipid/apoprotein complexes but not lipid-poor apoA1 (15, 16, 17, 18, 19, 20, 21). Studies using both Abcg1−/− and ABCG1 transgenic mice demonstrated that ABCG1 plays a critical role in controlling pulmonary and hepatic lipid homeostasis in response to a high fat, high cholesterol diet challenge (20). Subsequently, we reported that Abcg1−/− mice exhibit an age-related, progressive pulmonary disease that has many of the properties associated with human respiratory distress syndromes (22). Thus, although the lungs of 3-mo-old Abcg1−/− mice appear normal by most criteria, by the age of 8 mo the lungs of chow-fed Abcg1−/− mice contain massive numbers of cholesterol- and cholesteryl ester-loaded macrophages and a 5-fold increase in type 2 cells that contain excessive numbers of lamellar bodies (22). Taken together, these studies identified pivotal roles for ABCG1 in controlling pulmonary homeostasis in vivo (20, 22). Interestingly, based on gene deletion in mice and/or natural mutations in humans, other members of the ABC family of transporters, ABCA1, ABCA3, and ABCC7, have also been shown to play important roles in pulmonary physiology (23, 24, 25, 26, 27).
In this report, we focus on the progressive, age-dependent, chronic inflammatory process that accompanies the lipidosis in the lungs of Abcg1−/− mice. The data are consistent with the proposal that the regulation of lipid/cholesterol metabolism has a critical role in balancing pulmonary pro- and anti-inflammatory responses and that ABCG1 plays an essential role in these processes.
Materials and Methods
Thioglycollate-elicited primary peritoneal macrophages were obtained and maintained as described (20). Raw264.7 macrophages (American Type Culture Collection) were maintained in DMEM plus 10% FBS. The day before the experiments, cells were plated in 6-well plates in DMEM plus 1% FBS. The next morning the media was switched to DMEM plus 1% FBS plus cyclodextrin ± cholesterol and the cells then incubated for 6 h. Where indicated, cells were incubated in media containing 50 ng/ml LPS. Cyclodextrin and cyclodextrin-cholesterol (50:1 molar ratio) media was prepared as described (28).
H&E staining of paraffin-embedded lung sections was performed as described (20).
Preparation of frozen lung sections were as described (20). For immunofluorescence studies, FITC-conjugated anti-CD45R/B220 (1/100 dilution, BD Biosciences no. 553087) and PE-conjugated anti-CD3ε (1/50 dilution, BD Biosciences no. 553063) were used following standard protocols.
RNA isolation and analysis
RNA was isolated and analyzed by real time quantitative PCR (RT-qPCR) as described (20). Each qPCR assay was performed in duplicate using cDNA samples isolated from individual mice (n = 4–6/genotype) or from duplicate dishes of cells. Primer sets are available upon request. Values were normalized to GAPDH and calculated using the comparative Ct method. Gene expression profile of inflammation markers was determined using the GEArray mouse inflammatory cytokines and receptors microarray system from Superarray Bioscience Corporation, following the manufacturer recommendations.
Protein isolation and analysis
Protein extracts were obtained from the lungs of wild-type and Abcg1−/− mice, as described (29). Fifty μg of protein were resolved in NuPAGE 12% Bis-Tris gels (Invitrogen Life Technologies) and transferred to polyvinylidene difluoride membranes. The expression of MCP-1, TNF-α, and β-actin was detected using specific Abs (PeproTech P113, Santa Cruz Technology sc-1350, and Abcam ab6276, respectively) diluted in TBS containing 0.1% Tween 20 and 4% nonfat dried milk. Immune complexes were detected with HRP-conjugated secondary Abs (Bio-Rad) diluted 1/5000.
Analysis of bronchoalveolar lavage (BAL)
BAL were performed as described (22). Recovered cells were pelleted by gentle centrifugation (500 × g, 5 min). Crystals present in the initial pellet were dissolved in PBS-Isolymph (Gallard-Schlesinger) and purified by successive centrifugations as described (30). Cell-free supernatants from BAL and purified crystals were resolved in SDS-PAGE gels under reducing conditions using NuPAGE 4–20% Bis-Tris gradient gels (Invitrogen Life Technologies). Commassie blue G-250 stained bands of interest were removed from SDS-PAGE and digested with trypsin as described previously (31). Recovered peptide mixtures were dried, resuspended in 30% CH3CN containing 1% trifluoroacetic acid, and stored at −20°C until use. LC-MS/MS experiments were performed on a Proteomex LTQ tandem mass spectrometry instrument (Thermo Electron) with a surveyor pump system using a reversed phase column (75 μm i.d. 10 cm, BioBasic C18 5 μm particle size, New Objective). The flow rate was 5 μl/min for sample loading and 250 nL/min for separation. Buffer A was 0.1% formic acid with 2% acetonitrile in water, and Buffer B was 0.1% formic acid with 20% water in acetonitrile. Analyses were performed using a shallow gradient from 5% B to 40% B over 70 min, then from 40% B to 100% B over 20 min, and finally using 100% B for 9 min. The ion transfer tube of the linear ion trap was held at 200°C; the normalized collision energy was 35% for MS/MS; and the spray voltage was set at 1.9 kV. Briefly, the mass spectrometer was operated in the data-dependent mode to switch automatically between MS and MS/MS acquisition. Survey full-scan MS spectra with 1 microscans (m/z 400-2000) were acquired, followed by five sequential scan events of MS/MS. Each subsequent MS/MS collision induced dissociation fragmentation (at a target value of 10,000 ions) was performed on a precursor ion, which was isolated using the data-dependent acquisition mode to automatically select ions with sequentially top five highest intensities from the survey scan, with a 3.0 m/z isolation width. In the acquisition mode, dynamic exclusion was used with two repeat counts within 10 s, and with an exclusion duration of 40 s. The spectra were searched with SEQUEST against the murine IPI database with cardamidomethyl and methionine oxidation as modifications and the following criteria used for peptide identification: Xcorr, ≥2.2 (+1), ≥2.5 (+2), ≥3.8 (+3); DeltaCN > 0.1; and a peptide probability of 0.001. All proteins were identified on the basis of ≥2 peptides.
Differences between samples were analyzed by Student’s t test.
Basal inflammation in the lungs of Abcg1−/− mice
We previously reported that chow-fed mice lacking ABCG1 exhibit a progressive and severe pulmonary lipidosis; at 3 mo of age the lungs appear normal, but by 6 mo they appear white as a result of massive lipid deposition (22). Interestingly, this phenotype was greatly accelerated when young Abcg1−/− mice were fed a diet supplemented with high levels of fat and cholesterol (20). Under these latter dietary conditions, the lungs of 3-mo-old Abcg1−/− mice were white, likely as a result of the excess lipids that included cholesterol esters, cholesterol crystals, and phospholipids. In addition, the lungs of these Abcg1−/− mice showed evidence of lymphocytic infiltration, excess foamy lipid-loaded macrophages and proliferating type 2 cells (22).
Herein, we report that the lungs of Abcg1−/− mice also show signs of severe basal inflammation. H&E-stained sections of the lungs of 8-mo-old chow-fed Abcg1−/− mice indicate they contain large numbers of subpleural macrophages and lymphocytes (Fig. 1, B and C), perivascular lymphocytes (Fig. 1,D), large orange-stained extracellular eosinophilic structures (Fig. 1,E), and extracellular clefts (Fig. 1, B and D). None of these features, with the exception of very sporadic, very small perivascular lymphocytic foci, were observed in lung sections from wild-type mice (Fig. 1,A; data not shown). Interestingly, giant cells present in the Abcg1−/− lungs showed a scattered pattern of nuclei distribution (Fig. 1, B and D, and data not shown) similar to foreign body granuloma cells, and in contrast to Langhans-type granuloma cells that typically contain a circular peripheral arrangement of the numerous nuclei (32). No signs of fibrosis, as measured by Masson’s trichrome staining, were noted in the lungs of Abcg1−/− or wild-type mice (data not shown). When a semiquantitative method, adapted from Card et al. (33), was used to compute different inflammatory parameters, Abcg1−/− mice showed an 8-fold increase in their basal pulmonary inflammation status, compared with their wild-type litter mates (data not shown).
Markers of inflammation are induced in BAL from Abcg1−/− mice
Analysis of the cells and proteins recovered after BAL revealed dramatic differences between 8-mo-old chow-fed wild-type and Abcg1−/− mice (Fig. 2). As compared with wild-type mice, the BAL recovered from Abcg1−/− mice contained increased numbers of total macrophages, lipid-loaded foamy macrophages, and leukocytes (22) (Fig. 2, A–C), in addition to crystals of unknown composition (Fig. 2,B). We also used SDS-PAGE to analyze proteins present in cell-free extracts recovered after BAL. Based on Coomassie blue staining of the gels, we identified four bands that are increased in the cell-free supernatants recovered from Abcg1−/− mice (Fig. 2,D, lanes 2 vs 1; bands a–d). Individual bands were subsequently excised from the acrylamide gel, trypsin-digested, and analyzed by LC-MS/MS (see Materials and Methods). Proteins shown to be highly enriched in, or present exclusively in, the lavages from Abcg1−/− mice included moesin, ezrin, radixin, l-plastin, cathepsin B precursor, and chitinase 3-like protein variants 3 and 4 (also known as Ym-1 and Ym-2) (Fig. 2,E). Interestingly, the presence of some of these proteins in sputum or bronchial lavages has been correlated with pulmonary inflammation (34, 35, 36, 37). Only trace amounts of chitinase 3-like variant 3 (and no variant 4) were detected in lavages obtained from wild-type mice (Fig. 2,E). Finally, based on the number of peptides identified using LC-MS/MS, we estimate that transferrin and albumin levels were increased ∼2-fold in lavages from Abcg1−/− mice (Fig. 2 E).
The crystals present in the BAL obtained from Abcg1−/− mice (Fig. 2,B) were purified by repeated centrifugation on PBS-Isolymph solutions and shown to migrate as a single band on SDS-PAGE (Fig. 2,D, lane 4). Analysis of this band by LC-MS/MS revealed that the crystals were a mixture of chitinase 3-like variants 3 and 4, and that together these two proteins comprised >90% of the protein recovered from the gel (Fig. 2,E). The data of Fig. 3,F show representative MS/MS spectra from two peptides that identify them as chitinase 3-like variants 3 and 4, respectively. Based on the MS/MS analysis of proteins present in the four major bands (Fig. 2, D and E), we subsequently determined the mRNA levels of l-plastin, cathepsin B, and chitinase 3-like protein. The data show that the expression of all three genes was increased 3- to 6-fold in the lungs of Abcg1−/− mice (Fig. 2,G), consistent with increased protein recovery from lavages (Fig. 2 E).
Lymphocytic infiltration in the lungs of Abcg1−/− mice
As shown above in Fig. 1,D, the lungs of chow-fed 8-mo-old Abcg1−/− mice contain large numbers of lymphocytic foci, especially in the subpleural areas, the internal regions of the pulmonary parenchyma, and in areas adjacent to bronchioles and larger blood vessels. Direct immunofluorescence, using primary Abs to CD45R or CD3ε, indicates that such lymphocytic infiltrates contain B cells, with few, if any, T cells (Fig. 3). Staining of spleen sections obtained from wild-type mice confirmed the specificity of the Abs to CD45R and CD3ε (data not shown). As expected, lymphocytic foci were absent in the lungs of either wild-type litter mates or young (3-mo-old) chow-fed Abcg1−/− mice (data not shown). Interestingly, the B cells that accumulate in the lungs of the 8-mo-old Abcg1−/− mice stain highly positive for β-galactosidase activity (22) (and data not shown), suggesting that ABCG1 is normally expressed at a relatively high level in these B cells. Whether deletion of ABCG1 affects B cell function remains to be established.
Increased expression of cytokines and cytokine receptors in the lungs of Abcg1−/− mice
The presence of macrophage giant cells, lymphocytic infiltration, pneumocyte type 2 proliferation, accumulation of chitinase 3-like protein, and induction of cathepsin B and l-plastin mRNAs in the lung are all signs of inflammation. Consequently, we studied the pattern of expression of a panel of different cytokines and their receptors in 8-mo-old chow-fed animals. The microarray data suggest that many genes associated with inflammation are highly induced in the lungs of these older chow-fed Abcg1−/− mice (Fig. 4, A–C; complete microarray data is accessible from the Gene Expression Omnibus (GEO) repository). RT-qPCR analysis of a selective subset of these genes confirmed that mRNA levels for Mcp-1, Tnf-α, Mip-1β, Il-1β, Ccr-5, and iNos are all significantly elevated (2.5- to 35-fold) in the lungs of Abcg1−/− mice (Fig. 4,D). The induction of inflammatory cytokines is specific because the levels of many other mRNAs (e.g., Vcam-1) are unchanged (Fig. 4,D). We next studied the expression of two representative cytokines, MCP-1 and TNF-α, in crude extracts of lung tissue. As shown in Fig. 4 E, protein levels of both cytokines were markedly increased in the lungs of Abcg1−/− mice, compared with wild-type lungs.
We next asked whether increased expression of inflammatory cytokines could be observed in the lungs of 3-mo-old chow-fed Abcg1−/− mice, at a time when there is no evidence of lipid accumulation. The data of Fig. 4,F show that of those cytokines tested, only Mip-1β mRNA levels were significantly increased in the lungs of the young chow-fed Abcg1−/− mice as compared with wild-type mice. However, administration of a high fat/high cholesterol (HF/HC) diet resulted in increased expression of a number of mRNAs, including those encoding for Mip-1β, Mcp-1, Ccr-5, and Il1-β (Fig. 4,F). Interestingly, the expression of these same mRNAs was significantly higher in the lungs of HF/HC-fed Abcg1−/− mice as compared with their wild-type litter mates (Fig. 4 F). These differences in cytokine expression between Abcg1−/− and wild-type mice occurred even though plasma lipid levels were not significantly different between the two genotypes (data not shown). These results strongly suggest that pulmonary lipid content plays a crucial role in eliciting inflammatory responses, and that small increases in cellular sterol levels might be sufficient to promote cytokine expression.
Altered expression of matrix metalloproteinases in the lungs of Abcg1−/− mice
Figs. 1 and 4 show that the lungs of Abcg1−/− mice undergo: i) a profound tissue remodeling, with massive accumulation of macrophages and lymphocytes, and ii) a chronic inflammatory process. Consequently, we next studied the pulmonary expression of several matrix metalloproteinases (MMPs) and their inhibitors (tissue inhibitors of matrix metalloproteinases; TIMPs). Both families of proteins are a diverse group of molecules that mediate extra-cellular matrix degradation in a variety of physiological processes such as development, growth, and wound repair (reviewed in Refs. 38, 39). Deregulation of the activity of MMPs and TIMPs in the lungs has been shown to result in a variety of pathological effects, including asthma, COPD, and respiratory distress syndrome (38).
Consistent with the altered histology and increased cytokine expression, we found elevated expression of Mmp-8 and Mmp-12 mRNAs in the lungs of Abcg1−/− mice, compared with wild-type controls (Fig. 5). The expression of Mmp-9 mRNA, however, was not significantly different between the two genotypes (Fig. 5). We also noted a small but significant increase in Timp-1, but not Timp-2 and Timp-3, mRNA levels in the lungs of Abcg1−/− mice (Fig. 5).
Taken together, the data from Figs. 1–5 suggest that the lungs of Abcg1−/− mice undergo a dramatic increase in the inflammatory process that is initiated at a time when there is no measurable change in tissue lipids. However, inflammation is progressive and parallels the subsequent accumulation of lipids that occurs in chow-fed Abcg1−/− mice. In additon, we also show that the inflammatory process in the lung is greatly accelerated following administration of a diet enriched in fat and cholesterol.
Cholesterol loading stimulates cytokine production in macrophages
The data of Fig. 4,F indicate that dietary fat and cholesterol is sufficient to increase pulmonary inflammation in both wild-type and Abcg1−/− mice. To test the hypothesis that increased intracellular cholesterol promotes the expression of inflammatory mediators, we analyzed the expression of three representative cytokines, Tnf-α, Il-1β and Mcp-1, in the murine macrophage cell line Raw264.7 after loading the cells with excess cholesterol. As shown in Fig. 6,A, compared with cells incubated for 6 h with cyclodextrin alone, cyclodextrin/cholesterol resulted in a significant increase in the expression of all three cytokines. As expected, incubation of the cells with cyclodextrin in the absence of cholesterol resulted in decreased Abcg1 expression (Fig. 6,A), presumably as a result of a decrease in LXR activation following the efflux of cellular sterols to exogenous cyclodextrin (that functions as a sterol sink). Interestingly, the basal mRNA expression of Tnf-α and Il-1β, but not Mcp-1, was elevated ∼4-fold in primary peritoneal macrophages derived from Abcg1−/− mice, as compared with wild-type controls (Fig. 6,B). Incubation of these primary macrophages with exogenous cholesterol/cyclodextrin resulted in a marked induction of all three cytokines (Fig. 6,B). However, it is notable that cytokine mRNA levels were always significantly higher in Abcg1−/−, as compared with wild-type macrophages, following treatment with either cholesterol/cyclodextrin or LPS (Fig. 6 B).
ABCG1 is broadly expressed in many tissues and cell types where it is thought to promote the efflux of cellular cholesterol to specific exogenous acceptors, such as HDL or phospholipid/apoA1 complexes (15, 16, 17, 18, 19, 20, 21). In the current report, we show that loss of ABCG1 results in an age-dependent chronic inflammation that is limited to the lungs. Thus, although the lungs of 3-mo-old chow-fed Abcg1−/− mice appear normal, by the age of 6–8 mo the lungs contain massive infiltrates of lymphocytes and macrophages, crystals of chitinase 3-like protein and cholesterol, and increased expression of multiple cytokines and cytokine receptors. The lungs of Abcg1−/− mice also show evidence of tissue hemorrhage. In contrast, no signs of inflammation or lipid deposition are observed in wild-type litter mates.
Histological studies show that numerous “giant cells” containing an enlarged cytoplasm, lipid droplets, and multiple centrally located nuclei are present in the subpleural region of Abcg1−/− lungs (Fig. 1). Such giant cells have been observed in several chronic inflammatory conditions and during osteoclastic remodeling (reviewed in Ref. 40). Although the formation of these cells is poorly understood, they are believed to be active in phagocytosis of large extracellular components including foreign bodies (32, 40, 41). Whether the giant cells accumulate in response to the extracellular crystals of cholesterol and chitinase 3-like in the Abcg1−/− lungs is unknown. Although the molecular and cellular mechanisms of macrophage fusion are still largely unknown, several cytokines and plasma membrane receptors/ligands have been proposed to be involved (42, 43, 44, 45, 46).
Surprisingly, the lymphocytic infiltrates observed throughout the lungs of Abcg1−/− mice are comprised of B220+ B cells with few, if any, T cells (Fig. 3). B cells can be subdivided into a number of subtypes including B-1 and B-2 (47). The activity of B-2 cells, that comprise the majority of B cell subtypes, is modulated/regulated by T cells in the so-called T cell-dependent B cell activation. However, B-1 cells show a T cell-independent activation and express a unique set of surface markers including CD11b and IgM (48, 49, 50). These cells are thought to be essential in innate immunity and in the generation of natural Abs (IgM) (51). The finding that the pleural cavity of Abcg1−/− mice shows a significant enrichment of B220+, CD11b+, and IgM+ cells (data not shown) is consistent with an enrichment of B-1 cells. Whether this enrichment of B-1 cells is a response to a specific lipid that accumulates in the lungs of Abcg1−/− mice will require extensive additional studies. However, previous reports demonstrating that a number of natural Abs (e.g., E06, E014, and T15) produced by B-1 cells bind to oxidized phosphatidylcholine present in oxidized LDL, to phosphatidylcholine from bacterial capsules, and to apoptotic cells (reviewed in Ref. 52) would be consistent with this proposal.
To identify changes in inflammatory mediators, we profiled the expression of inflammatory genes in the lungs of wild-type and Abcg1−/− mice. The results show that a large number of cytokines and cytokine receptors, including Tnfα, Il-1β, Mcp-1, Ccr-5, iNos, and Mip-1β, are induced in the lungs of Abcg1−/− mice (Fig. 4, A–E). We suggest that the massive lipid accumulation in the alveolar spaces of these animals acts as the triggering stimuli for the inflammatory response. Consistent with this hypothesis, we noted that pulmonary cytokine expression was induced in wild-type mice following administration of a diet enriched in fat and cholesterol for 9 wk (Fig. 4,F). This same diet resulted in even greater levels of cytokine expression in the lungs of Abcg1−/− mice as compared with wild-type litter mates (Fig. 4,F). Collectively, these data suggest that increased intracellular lipids/cholesterol induce the expression of certain inflammatory mediators by a process that is enhanced in cells lacking ABCG1. The finding that incubation of either Raw264.7 cells or primary macrophages with cyclodextrin-cholesterol resulted in the induction of Tnf-α, Il-1β, and Mcp-1 (Fig. 6) suggests that uptake of exogenous cholesterol is sufficient to induce the expression of certain cytokines.
Earlier studies reported that treatment of macrophages with oxidized LDL resulted in induction of Il-8 (53), and that this effect could be enhanced by cotreatment of the cells with an acyl-coenzymeA:cholesterol acyltransferase (ACAT) inhibitor (thus inhibiting cholesterol esterification) (54). A subsequent report showed that induction of Il-6 and Tnf-α in response to oxidized or acetylated LDL was dependent on the presence of the ACAT inhibitor (55). These authors proposed that the increased cholesterol content in the endoplasmic reticulum, following inhibition of ACAT, resulted in activation of NF-κB, multiple MAP kinase cascades, and the unfolded protein response pathway (55). Activation of this latter pathway in cultured cells, in response to elevated levels of intracellular unesterified cholesterol, can result in accelerated apoptosis (56, 57). Importantly, the lungs of Abcg1−/− mice contain high levels of unesterified cholesterol and cholesterol crystals (22), together with increased numbers of TUNEL-positive apoptotic cells (data not shown) in the absence of exogenous inhibitors of ACAT. Based on these data, we propose that ABCG1 plays a critical role in controlling intracellular cholesterol redistribution in macrophages and other cell types, and that loss of ABCG1 results in abnormal cholesterol metabolism. We hypothesize that, in the context of the lungs of Abcg1−/− mice, the chronic exposure of alveolar cells to surfactant-derived cholesterol results in a slow and progressive accumulation of intracellular sterols that, in turn, stimulates cytokine production and inflammation. According to this model, inflammation is a secondary process that develops in response to the lipid accumulation in the lungs of the Abcg1−/− mice. It is not clear, however, whether treating Abcg1−/− mice with anti-inflammatory drugs (such as glucocorticoids or nonsteroidal anti-inflammatory drugs) or decreasing the oxidative burden of the lungs (i.e., administering N-acetylcysteine) would ameliorate or exacerbate the pulmonary lipidosis. Additional experiments will be necessary to test this proposal.
Nevertheless, based on the profound changes that occur in the lungs of Abcg1−/− mice, it is tempting to speculate that loss of ABCG1 might affect pulmonary function in response to a variety of insults, such as bacterial, viral, or fungal infection. Interestingly, two recent papers describe a marked down-regulation of LXR target genes, including ABCG1, when peritoneal macrophages are incubated in vitro with influenza A virus or E. coli (58), or when J774 macrophages are cultured with LPS (59). These later studies suggest that certain pathogens might modulate cholesterol homeostasis affecting the transcriptional activity of LXR and, consequently, the expression of some of its targets, including ABCG1. Whether modulation of ABCG1 expression plays a critical role in pulmonary-pathogen infectivity remains to be established. Testing the hypothesis that loss of ABCG1 compromises the ability of the murine lungs to respond to environmental challenges, such as acute or chronic bacterial/viral/fungal infection, or allergens will require extensive additional experiments. Interestingly, elevated levels of chitinase-3-like proteins (Fig. 2) have been associated in humans and rodents with asthma and pulmonary allergic responses (34, 35, 60, 61).
Respiratory inflammation and increased expression of cytokines have been reported to increase in mice lacking functional lysosomal acid lipase, or following administration of a cholesterol-rich diet (11, 12, 62, 63). Conversely, recent studies suggest a protective role of the cholesterol-lowering drugs statins in smoke-driven respiratory inflammation, COPD, and asthma in both mice (64, 65, 66, 67) and humans (68, 69). To date, no functional mutations have been described in human ABCG1. However, Thomassen et al. (70) recently reported the intriguing finding that a subgroup of patients with pulmonary alveolar proteinosis show a marked decrease in ABCG1 mRNA and protein expression in alveolar macrophages recovered from BALs, compared with samples from healthy volunteers. Taken together, all these data suggest a critical role for ABCG1 in controlling pulmonary homeostasis and balancing both lipid/cholesterol metabolism and inflammatory responses.
We thank Drs. Robert Strieter and John Belperio from the Department of Pathology and Laboratory Medicine at University of California, Los Angeles for their help with the cytokine analysis. We thank Dr. Sam Hawgood from the Cardiovascular Research Institute and the Department of Pediatrics at University of California, San Francisco and Dr. Steve Bensinger (University of California, Los Angeles) for helpful discussions. We also thank the members of the Edwards lab for critical reading of the manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported in part by National Institutes of Health Grants 30568 and 68445 (to P.A.E.), a grant from the Laubisch Fund (to P.A.E.), and a grant from Pfizer, Inc. (to P.A.E.). Á.B. was partially supported by an American Heart Association (Western Affiliate) Postdoctoral Fellowship (0525010Y).
Abbreviations used in this paper: COPD, chronic obstructive pulmonary disease; ABC, ATP-binding cassette; LDL, low density lipoprotein; RT-qPCR, real time quantitative PCR; BAL, bronchoalveolar lavage; HF/HC, high fat/high cholesterol; MMP, matrix metalloproteinases; LXR, liver X receptor; HDL, high density lipoprotein; TIMP, tissue inhibitors of matrix metalloproteinases; ACAT, acyl-coenzymeA:cholesterol acyltransferase.