Regulation of Macrophage Foam Cell Formation During Nitrogen Mustard (NM)-Induced Pulmonary Fibrosis by Lung Lipids

Abstract

Nitrogen mustard (NM) is a vesicant known to target the lung, causing acute injury which progresses to fibrosis. Evidence suggests that activated macrophages contribute to the pathologic response to NM. In these studies, we analyzed the role of lung lipids generated following NM exposure on macrophage activation and phenotype. Treatment of rats with NM (0.125 mg/kg, i.t.) resulted in a time-related increase in enlarged vacuolated macrophages in the lung. At 28 days postexposure, macrophages stained positively for Oil Red O, a marker of neutral lipids. This was correlated with an accumulation of oxidized phospholipids in lung macrophages and epithelial cells and increases in bronchoalveolar lavage fluid (BAL) phospholipids and cholesterol. RNA-sequencing and immunohistochemical analysis revealed that lipid handling pathways under the control of the transcription factors liver-X receptor (LXR), farnesoid-X receptor (FXR), peroxisome proliferator-activated receptor (PPAR)-ɣ, and sterol regulatory element-binding protein (SREBP) were significantly altered following NM exposure. Whereas at 1–3 days post NM, FXR and the downstream oxidized low-density lipoprotein receptor, Cd36, were increased, Lxr and the lipid efflux transporters, Abca1 and Abcg1, were reduced. Treatment of naïve lung macrophages with phospholipid and cholesterol enriched large aggregate fractions of BAL prepared 3 days after NM exposure resulted in upregulation of Nos2 and Ptgs2, markers of proinflammatory activation, whereas large aggregate fractions prepared 28 days post NM upregulated expression of the anti-inflammatory markers, Il10, Cd163, and Cx₃cr1, and induced the formation of lipid-laden foamy macrophages. These data suggest that NM-induced alterations in lipid handling and metabolism drive macrophage foam cell formation, potentially contributing to the development of pulmonary fibrosis.

Nitrogen mustard (NM) is a bifunctional alkylating agent and cytotoxic vesicant known to cause acute damage to the respiratory tract which progresses to fibrosis. This is accompanied by the sequential accumulation of proinflammatory M1 macrophages and anti-inflammatory M2 macrophages in the lung, which have been implicated in these pathogenic responses (Venosa et al., 2016). Macrophage activation toward an M1 or M2 phenotype occurs in response to mediators the cells encounter in the tissue microenvironment. Thus, while IFN-γ, TNF-α, and TLR-4 agonists induce the development of M1 macrophages, M2 macrophages develop following stimulation with cytokines such as IL-4 and IL-13 (Martinez and Gordon, 2014). Macrophage phenotypic activation is also regulated by signaling pathways, transcription factors, cellular metabolism, and epigenetic modifiers (de Groot and Pienta, 2018; Li et al., 2018).

Evidence suggests that lipids contribute to the formation of foam cells, M2-like macrophages expressing high levels of scavenger receptors (eg, CD36) and lipid transporters (eg, ABCA1, ABCG1), as well as lipid chaperones (eg, ApoA, ApoE) (Boullier et al., 2005; Canton et al., 2013; Guo et al., 2006; Hazen and Chisolm, 2002; Rahaman et al., 2006). This is supported by studies demonstrating that M2 macrophages derived from peripheral blood monocytes readily take up oxidized lipids and become foam cells (van Tits et al., 2011). Similarly, in experimental models of fibrosis, alveolar proteinosis, and atherosclerosis, which are characterized by a lipid-rich microenvironment, macrophages exhibit a foamy appearance, stain positively for Oil Red O, upregulate scavenger receptors, and express markers of M2 activation (Basset-Leobon et al., 2010; Romero et al., 2015; Spann et al., 2012). Although lipids have been linked to the formation of foamy macrophages during the development of pulmonary fibrosis (Romero et al., 2015), mechanisms mediating this process and the role of these cells in disease pathogenesis remain largely unknown.

Lipid homeostasis is controlled by several transcription factors including liver-X receptor (LXR), farnesoid-X receptor (FXR), peroxisome proliferator-activated receptor (PPAR)-ɣ, and sterol regulatory element-binding proteins (SREBPs) (Kidani and Bensinger, 2012; Lambert et al., 2003; Madison, 2016). Pharmacologic and genetic modification of these transcription factors and/or their target genes has revealed a tight association between immune cell lipid metabolic programing and phenotypic activation (Reddy et al., 2012; Remmerie and Scott, 2018; Smoak et al., 2008; Sonett et al., 2018; Zhang et al., 2012). In this context, PPARɣ and LXR have been directly associated with M2 macrophage polarization by upregulating expression of scavenger receptors and transporters involved in lipid uptake and extrusion (Chawla et al., 2001; Gautier et al., 2012), and inhibiting proinflammatory transcription factors including STAT1 and NF-κB (Pascual-García et al., 2013).

In the present studies, we investigated whether lung lipids play a role in macrophage foam cell formation during the pathogenesis of NM-induced pulmonary fibrosis. We found that NM-induced fibrosis is associated with increases in lung lipids, upregulation of macrophage lipid transporters, and the appearance of enlarged foamy macrophages in the lung. Moreover, a large aggregate lipid-containing fraction of bronchoalveolar lavage fluid (BAL) prepared from fibrotic lung induced the formation of foamy macrophages and promoted M2 activation of these cells. Taken together, these data suggest that derangements in lung lipids after NM exposure contribute to M2 macrophage activation and fibrogenesis.

MATERIALS AND METHODS

Animals and Treatments

Male Wistar rats (225–250 g) were purchased from Harlan Laboratories (Indianapolis, Indiana) and maintained in an AALAC approved animal care facility. Animals were housed in filter top microisolation cages and provided food and water ad libitum. Animals received humane care in compliance with the Guide for the Care and Use of Laboratory Animals, published by the National Institutes of Health. Animals were anesthetized with 2.5% isoflurane, and then administered PBS or NM (0.125 mg/kg, mechlorethamine hydrochloride, Sigma-Aldrich, St Louis, Missouri) intratracheally as previously described (Sunil et al., 2011). All instillations were performed by one individual to reduce experimental variability. NM was prepared immediately before administration in a designated room under a chemical hood following Rutgers University Environmental Health and Safety guidelines.

Sample Collection and Preparation of BAL Large Aggregate Fractions

Animals were euthanized by i.p. injection of ketamine (80 mg/kg) and xylazine (10 mg/kg) 3, 7, or 28 days after administration of PBS or NM. BAL was collected by slowly instilling and withdrawing 10 ml of ice cold (4°C) PBS into the lung through a cannula inserted into the trachea. BAL was centrifuged (300 × g, 8 min) and cell-free supernatants aliquoted and stored at −80°C until preparation of large aggregate fractions (see below). The lung was then removed from the chest cavity and an additional 10 ml of ice cold PBS slowly instilled and withdrawn through the cannula while gently massaging the tissue; this procedure was repeated 4 times. Lavage fluid was centrifuged (300 × g, 8 min), cell pellets resuspended in 10 ml of PBS and combined with cells collected by BAL only. Cells were enumerated and viability assessed using a hemocytometer with trypan blue dye exclusion. Cytospins were prepared and stained with Giemsa; differential analysis of BAL plus massage cells showed >96% macrophage purity.

Large aggregate fractions of BAL, containing the majority of lipids in lung lining fluid, were prepared and analyzed as previously described (Bligh and Dyer, 1959). Briefly, frozen aliquots of BAL were thawed, centrifuged (20 000 × g, 1 h), and pellets containing large aggregate fractions resuspended in 0.9% NaCl saline. After evaporation on a heating plate, solutes were incubated for 30 min with chloroform/methanol (1:2), followed by dilution in 1 ml of water, and overnight incubation at 4°C. The lower phase of the suspension and freshly prepared standards, consisting of potassium phosphate monobasic (0–1000 mg), were evaporated using inert nitrogen gas, resuspended in perchloric acid and heated for 20 min. Subsequently, water (1.4 ml), ammonium molybdate (100 μl), and naphthalenesulfonic acid (50 μl) were added to the pellet, and the mixture heated in a water bath (95°C) for 7 min. Protein concentrations were measured using a BCA protein assay kit (Pierce Biotechnologies Inc, Rockford, Illinois) with bovine serum albumin as the standard. Phospholipid content was assayed spectrophotometrically by the method of Bligh and Dyer (1959); high-density lipoprotein (HDL) and very low density lipoprotein/low-density lipoprotein (LDL/VLDL) Cholesterol Assay Kit was used for analysis of cholesterol, according to the manufacturer’s instructions (Abcam, Cambridge, Massachusetts).

Lung Histology and Immunohistochemistry

Following BAL, the lung was removed, fixed by slow inflation at constant pressure with 10 ml of 2% paraformaldehyde using a syringe and paraffin embedded. As previously reported, this procedure did not cause alterations in lung morphology (Venosa et al., 2016). Sections (5 μm) were then prepared and stained with Masson’s trichrome. For immunohistochemistry, sections (5 μm) were deparaffinized with xylene followed by decreasing concentrations of ethanol (100%–50%) and then water. After antigen retrieval using citrate buffer (10.2 mM sodium citrate, 0.05% Tween 20, pH 6.0, 10 min) and quenching of endogenous peroxidase with 3% hydrogen peroxide in methanol (30 min), sections were incubated for 2 h at room temperature with 10% goat serum to block nonspecific binding. This was followed by overnight incubation at 4°C in a humidified chamber with rabbit polyclonal anti-CD36 (1:250, Santa Cruz Biotechnology, Dallas, Texas), anti-FXR (1:250, Santa Cruz), anti-ATP-binding cassette transporter-A1 (ABCA1, 1:200, Abcam), or anti-oxidized phospholipids (monoclonal IgM binding to the phosphocholine headgroup of oxidized phospholipids, 1:250, Avanti Polar Lipids Inc, Alabaster, Alabama) primary antibodies, or serum controls diluted in blocking buffer. Sections were then washed and incubated at room temperature for 30 min with biotinylated secondary antibody (Vectastain Elite ABC kit, Vector Labs, Burlingame, California). Binding was visualized using a Peroxidase Substrate Kit DAB (Vectastain). Random sections from at least three rats per treatment group were analyzed for each antibody. For neutral lipid staining, lung tissue was snap frozen in liquid nitrogen–cooled isopentane and embedded in OCT medium (Sakura Finetek, Torrance, California). Sections (6 µm) and cytospins were fixed in 10% formalin, rinsed with water and then with 60% isopropanol. This was followed by a 30-min incubation with Oil Red O, rinsing with 60% isopropanol, and then with water. Slides were then counterstained with hematoxylin (4 min) and cover slipped with glycerol vinyl alcohol aqueous mounting media (ThermoFisher Scientific, Grand Island, New York).

Macrophage activation studies

Cells, collected from naive animals by BAL plus lung massage as described above, were suspended in Dulbecco’s modified Eagle medium (DMEM) containing 10% fetal bovine serum (FBS) and inoculated into 6-well plates (4 × 10⁶ cells/well). After overnight incubation (37°C, 5% CO₂), macrophages were rinsed and cultured in DMEM containing 1% FBS ± 50 μg/ml of lipid enriched large aggregate BAL fractions prepared 3 or 28 days after administration of PBS (CTL) or NM to rats. Macrophages were harvested 24 h later using a rubber policeman for preparation of cytospins and analysis of gene expression.

mRNA isolation and real time (RT) qPCR

Total mRNA was extracted from lung macrophages using an RNeasy Mini kit (Qiagen, Valencia, California). RNA was reversed transcribed using a High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, California). Standard curves were generated using serial dilutions from pooled cDNA samples. RT-qPCR was performed using SYBR Green PCR Master Mix (Applied Biosystems) on an Applied Biosystems 7300HT RT-PCR system. Glyceraldehyde 3-phosphate dehydrogenase (Gapdh) was used to normalize the data. Full-length coding sequences were obtained from the NCBI Gene Bank. Primers were designed using Primer Express 3.0 software (Applied Biosystems). The following forward and reverse primers were used: Gapdh, CCTGGAGAAACCTGCCAAGTAT and CTCGGCCGCCTGCTT; Nos2, TGGTGAAAGCGGTGTTCTTTG and ACGCGGGAAGC CATGA; Prostaglandin-endoperoxide synthase 2 (Ptgs2), TGCTCACTTTGTTGAGTAGTCATTCAC and CATTCCTTCC CCCAGCAA; Il12b, CAGAAAGGTGCGTTCCTCGTA and GCCC CTTTGCATTGG; Cd163, CCTTTTCATCCTCATCTTCCTTCT and CAAAGGAGCACTGCGTGTTC; Il10, CCCAGAAATCAAGGA GCATTTG and CAGCTGTATCCAGAGGGTCTTCA; Cx₃cr1, GGAGCAGGCAGGACAGCAT and CCCTCTCCCTCGCTTGTGTA; Ccr2, TGACAGAGACTCTTGGAATGACACA and CTCACCAACAAA GGCATAAATGAT; apolipoprotein E (Apoe), TCCATTGCCTCCA CCACAGT and GGCGTAGGTGAGGGATGATC; transforming growth factor-β1 (Tgfb1); sterol regulatory element-binding protein 2 (Srebp2), GATGAGCTGACTCTCGGGGA and TGCTGCTG GATGGTAACTGG; 3-hydroxy-3-methylglutaryl-CoA synthase 1 (Hmgcs1), GGTCCGATCGCGTTTGGTG and TTGTCTCTGTCCCG ACTTCC; farnesyl diphosphate farnesyl transferase 1 (Fdft1), TTACCTGATCCTCCGAGCCA and TGCCCCTTCCGAATCTTCAC; isopentenyl-diphosphate delta isomerase 1 (Idi1), CCAACACCATCTCTTGGGCT and GCCAATTGCCAATCTAGCGT; farnesyl diphosphate synthase (Fdps), GTACAATCGGGGTCT GACGG and GGTTGAGGTAGTAGGGCTGC.

RNA-sequencing (RNA-seq) analysis

Total RNA was extracted from cells collected by BAL plus lung massage; RNA quality was analyzed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, California). RNA was then converted to cDNA and amplified using an Illumina TruSeq RNA Sample Preparation Kit following the manufacturer’s directions (Illumina, San Diego, California). RNA-seq was performed using an Illumina NextSeq 550 system. Ingenuity IPA Version 44691306 (QIAGEN Inc, https://www.qiagenbioinformatics.com/products/ingenuity-pathway-analysis/; last accessed August 6, 2019) was used to identify significantly enriched canonical pathways, upstream regulators, and diseases and functions (Kramer et al., 2014). Each analysis was performed on differentially expressed genes (experimental log ratio > ± 1.5 and experimental false discovery rate [q-value] < 0.05), and considered both direct and indirect relationships, and all node types, species, tissues, and cell lines, but was limited to experimentally observed interactions in the Ingenuity Knowledge Base. Data were deposited in Dryad (DOI: doi:10.5061/dryad.kr06n63) and NCBI’s Gene Expression Omnibus (Edgar et al., 2002) and are accessible through GEO Series accession number GSE125619 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE125619).

Statistical analysis

Data were analyzed using a one-way ANOVA with Tukey’s post hoc test; p ≤ .05 was considered statistically significant.

RESULTS

NM Exposure Results in an Accumulation of Lipid-Laden Macrophages in the Lung

Consistent with previous studies (Venosa et al., 2016), extensive perivascular edema and hemorrhage were observed in the lung within 3 days of NM exposure (Figure 1, left panels). By 7 days, thickening of the epithelial wall and bronchiolar remodeling, along with fibroplasia were noted, and by 28 days post NM, frank fibrosis was evident. These structural changes were associated with a persistent macrophage dominant inflammatory response (Venosa et al., 2016). Although at 3 days post NM, the macrophages were relatively small and round, when compared with macrophages from control animals, at 7 days they were enlarged and vacuolated. These foamy-looking macrophages persisted in the lung for at least 28 days after NM administration; at this time, they were mainly clustered within thickened alveoli surrounding areas of fibrosis. At 28 days post NM, we also noted a subpopulation of smaller macrophages in the lung (Figure 1, right panel).

Figure 1.

Effects of nitrogen mustard (NM) on lung histology and macrophage morphology. Tissue sections, prepared 3, 7, and 28 days after exposure of rats to NM or to PBS control (CTL) were stained with Masson’s Trichrome. Original magnification: ×100 (left panels) and ×400 (right panels). Cytospins of cells (10⁴ cells) collected by bronchoalveolar lavage fluid (BAL) plus massage were fixed in methanol and stained with Giemsa. Original magnification: ×600. Representative images from at least 3 rats/treatment group are shown.

Previous studies have demonstrated that foam cell formation is due to uptake of lipids (Kunjathoor et al., 2002). To determine if macrophages accumulating in the lung after NM were foam cells and contained lipids, histologic sections were stained with Oil Red O. In tissue sections prepared 28 days post NM exposure, Oil Red O staining was observed in macrophages (Figure 2A); this was more prevalent in macrophages localized in areas of thickened and fibrotic tissue. Macrophages isolated from the lung 28 days post NM exposure also stained positively for Oil Red O. In contrast, Oil Red O staining was not detected in macrophages from NM treated animals at 3 or 7 days postexposure, or from PBS controls (Figure 2A and not shown).

Figure 2.

Effects of NM on lung lipids. A, Frozen tissue sections and cytospins, prepared 28 days after exposure of rats to NM or to PBS control (CTL), were stained with Oil Red O. Arrows, cells in insets. Original magnification: ×100; inset magnification: ×600. Representative images from at least 3 rats/treatment group are shown. B, BAL was collected 3, 7, and 28 days after exposure of rats to NM or to PBS control (CTL). Large aggregate fractions were prepared and assayed for phospholipid content as described in Materials and Methods. Bars, mean ± SE (n = 3–5 rats). *Significantly different (p ≤ .05) from CTL. C, Tissue sections, prepared 3, 7, and 28 days after exposure of rats to NM or to PBS control (CTL), were immunostained with antibody to oxidized phospholipids. Binding was visualized using a Vectastain kit. Arrows, macrophages in insets. Original magnification, ×100; inset magnification, ×1000. Representative sections from 3 rats/treatment group are shown.

We next determined if the formation of foamy macrophages following NM administration was linked to alterations in lung lipids. Treatment of rats with NM resulted in increases in levels of phospholipids and cholesterol in BAL; although elevations in phospholipids were evident at all postexposure times (Figure 2B), cholesterol was only increased at 28 days post NM (∼1.8-fold, not shown). Oxidized phospholipids were also detected in macrophages in histologic sections of the lung at all times after NM, and in epithelial cells at 3 days postexposure (Figure 2C).

Effects of NM Exposure on Expression of Regulatory Molecules Involved in Macrophage Activation and Lipid Handling

To identify potential regulatory pathways involved in macrophage activation and foam cell formation, we performed RNA-seq analysis on lung macrophages isolated by BAL and lung massage 1 and 28 days post NM or PBS administration. Unbiased hierarchical gene clustering showed that the transcriptomes of macrophages collected 1 and 28 days after NM treatment of animals were different from the transcriptome of macrophages collected from PBS controls (Figure 3A). Moreover, principal component analysis demonstrated that the transcriptional profile of macrophages collected 1 day post NM was distinct from macrophages collected at 28 days (Figure 3B). To identify biological pathways that are altered by NM treatment, we used Ingenuity Pathway Analysis (IPA) software. Disease and function clustering analysis showed that pathways associated with leukocyte migration and inflammatory responses, and lipid accumulation and transport pathways were significantly altered in macrophages at 1 day post NM administration (Figure 3C). Although leukocyte migration and inflammatory response pathways were also upregulated at 28 days post NM, lipid accumulation, export, and deposition were largely unaltered. Canonical pathway analysis showed that at 1 day post NM, signaling associated with iNOS, NF-κB, interferon, TREM1, and production of reactive oxygen and nitrogen species were upregulated in macrophages (Figure 3D). These findings are consistent with our previous studies demonstrating proinflammatory macrophage activation in the lung early after NM administration (Malaviya et al., 2012, 2015; Venosa et al., 2016). We also found that genes involved in LXR signaling were reduced at both 1 and 28 days post NM. At 28 days post NM, PPAR signaling was decreased, along with pathways involved in cellular proliferation (mTOR) and inflammation (TNFR, chemokine signaling, JAK/STAT, ROS/RNS, HMGB1). Conversely, genes associated with cholesterol biosynthesis and PPARα pathways were increased (Figure 3D).

Figure 3.

RNA-seq and Ingenuity Pathway Analysis (IPA) of lung macrophages. Lung macrophages were isolated by BAL plus massage, 1 and 28 days after exposure of rats to NM or to PBS control (CTL) (n = 3 rats/treatment group). Total RNA was prepared and analyzed by RNA-seq. A, Heat-map of normalized expression of 1582 genes (rows) across treatment groups (columns); dendrogram (top) generated following hierarchical clustering of these genes. B, Two-dimension principle component analysis of macrophage genes identified as significantly different (fold change > 1.5 and false discovery rate [q-value] < 0.05) at 1 and 28 days post NM relative to PBS controls. C and D, Genes that were differentially expressed (fold change > 1.5 and false discovery rate [q-value] < 0.05) in macrophages 1 and 28 days after exposure to NM relative to PBS control (CTL) were analyzed for enriched diseases and functions (C) and canonical pathways (D). Bars in (C) show enrichment [−log(p value)] of diseases and functions. Heat-map in (D) shows activation z-scores of significantly enriched canonical pathways.

We next focused on lipid transport as this pathway exhibited the lowest activation z-score indicating predicted downregulation in macrophages 1 day post NM exposure. Diverse expression patterns were observed, with one subset of genes (fatty acid binding protein-4 [Fabp4], steroidogenic acute regulatory protein-related lipid transfer protein-3 [Stard3], annexin-6 [Anxa6], phospholipase-2 [Pld2], sterol O-acyltransferase-2 [Soat2]) exhibiting early (1 day post NM) downregulation followed by upregulation at 28 days, whereas a second group (Abcg1, Star, acyl-coenzyme A [CoA]: cholesterol acyltransferase-1 [Acat1], phosphatidylinositol transfer protein membrane associated-1 [Piptnm1]) was downregulated at both 1 and 28 days post NM administration (Figure 4A). Focused analysis of RNA-seq also showed upregulation of a cluster of genes canonically associated with inflammatory responses (Il1b, S100a9, S100a8, NF-κB inhibitor alpha [Nfkbia]) at both 1 and 28 days post NM exposure, suggesting a connection to lipid handling (Figures. 4A and 4C). In contrast, expression of the profibrotic growth factor, TGFβ was reduced, suggesting that in our model, macrophages are not a major source of this mediator. It may be that other profibrotic mediators are released from macrophages that are important in fibrogenesis. We also analyzed genes involved in cholesterol biosynthesis because this pathway was altered following NM administration. Significant increases in 3-hydroxy-3-methylglutaryl-CoA synthase 1 (Hmgcs1), cytochrome51A1 (Cyp51a1), 3-beta-hydroxysterol delta (14)-reductase (Tm7sf2), farnesyl diphosphate farnesyl transferase 1 (Fdft1), isopentenyl-diphosphate delta isomerase 1 (Idi1), and farnesyl diphosphate synthase (Fdps) expression were observed 28 days post NM (Figure 4B). These findings are consistent with increases in BAL cholesterol levels noted at this time.

Figure 4.

IPA of macrophage genes involved in lipid transport and cholesterol biosynthesis. Focused analysis was performed on the macrophage transcriptome 1 and 28 days after exposure of rats to NM or to PBS control (CTL) using IPA. A, Heat-map of differentially expressed genes involved in lipid transport (network with lowest activation z-score indicating predicted downregulation). Criteria for significance were set at fold change > 1.5 and false discovery rate (q-value) < 0.05. B, Heat-map of differentially expressed genes involved in cholesterol biosynthesis (network with lowest activation z-score indicating predicted downregulation). Criteria for significance were set at fold change > 1.5 and false discovery rate (q-value) < 0.05. C, Network of genes involved in lipid transport was grown to include 10 upstream molecules (cytokines or ligand-dependent nuclear receptors) that directly (solid lines) or indirectly (dashed lines) regulate pathway molecules. Blue ovals represent genes that exhibited the highest degree of specific connectivity. Green, downregulation; red, upregulation; white, not identified in analysis; gray, identified, but did not pass statistical threshold (fold change > 1.5, and/or q-value < 0.05).

To identify potential upstream regulators of these genes, we performed linkage analysis using a network grown to identify up to 10 regulatory molecules. IPA added upstream regulators that showed the greatest specific connectivity to molecules identified in our analysis, and then the overall network was limited to include only those genes with at least 2 connections (Figure 4C). The most connected regulators were PPARα, PPARγ, estrogen receptor-1 (ESR1), and TNF-α.

In further studies, we used RT-qPCR to analyze expression of lipid-associated regulators and targets which, based on our RNA-seq studies, were predicted to be altered in macrophages following NM administration. We found that the cholesterol efflux transporters, Abca1 and Abcg1, were constitutively expressed by lung macrophages (Figure 5). A transient decrease in mRNA expression of both transporters was observed at 1 and 3 days post NM exposure; Abcg1 was also downregulated at 7 days. Analysis of lung sections by immunohistochemistry showed that ABCA1 protein expression was also downregulated in macrophages 3 days after NM exposure; subsequently, levels of this transporter protein returned toward control. ABCA1 was also transiently expressed in lung epithelium 7 days post NM (Figure 6, left panels). This was correlated with reduced expression of Lxra, which is known to regulate expression of these genes (Figure 5) (van der Deen et al., 2005; Wojcik et al., 2008). In contrast, Srebp2, which regulates cholesterol metabolism in macrophages, was transiently upregulated 1 day post NM exposure. In accord with our RNA-seq data, expression of Pparg and its downstream targets cytochrome27A1 (Cyp27a1), Fdps, and Acat1, which are involved in cholesterol catabolism and esterification, respectively (Chang et al., 2009; Taylor et al., 2010), were also decreased in macrophages at 1 and 3 days following NM administration (Figure 5). Pparg was also decreased 7 days post NM. Conversely, we observed no significant changes in expression of the LDL receptor (Ldlr) (Figure 5).

Figure 5.

Effects of NM on macrophage expression of genes involved in lipid handling analyzed by RT-qPCR. Lung macrophages were collected 1, 3, 7, and 28 days after exposure of rats to NM or to PBS control (CTL). Gene expression was analyzed by RT-qPCR for transcription factors (Lxr, Pparg, and Srebp2), and target genes (Abca1, Abcg1, Cyp27a1, Acat1, Ldlr, Cd36, Apoa, Apoe, and Fdps) involved in lipid handling. Data are presented as fold change relative to CTL. Bars, mean ± SE (n = 3–5 replicates/group) from at least 2 independent experiments. *Significantly different (p ≤ .05) from CTL.

Figure 6.

Effects of NM on expression of ABCA1, FXR, and CD36. Lung sections, prepared 3, 7, and 28 days after exposure of rats to NM or to PBS control (CTL), were immunostained with antibody to ABCA1, FXR, and CD36. Binding was visualized using a Vectastain kit. Arrows, macrophages in insets. Original magnification, ×600; inset magnification, ×1000. Representative sections from 3 rats/treatment group are shown.

We next analyzed lung sections for expression of FXR, as it is also known to regulate levels of intracellular lipids. Macrophages from control rats constitutively expressed cytoplasmic FXR (Figure 6, center panels). Following NM administration, increased numbers of FXR positive macrophages were observed in the lung. Although at 3 days post NM, FXR was largely localized in the nucleus, at 7 and 28 days, it was mainly cytoplasmic. At 3 days, and to a lesser extent at 7 days post NM, expression of FXR was also observed in epithelial cells. Significant increases in macrophage mRNA expression of putative targets for FXR, including the scavenger receptor, Cd36, and the lipid chaperones, Apoa and Apoe, which are important in lipid uptake (Agardh et al., 2011; Castrillo and Tontonoz, 2004; Guo et al., 2006; Li et al., 2012; Shibata and Glass, 2009; Silverstein et al., 2010), were also observed in macrophages 3 and 7 days after NM exposure (Figure 5). Increases in macrophage CD36 protein expression were also noted in lung sections 3 days post NM (Figure 6, right panels).

Lung Lipids Generated Following NM Exposure Regulate Macrophage Phenotype

In our next series of studies, we assessed the effects of lipids released following NM exposure on lung macrophage phenotype and the formation of foam cells. In these studies, alveolar macrophages, collected from untreated rats by BAL plus lung massage, were incubated with lipid-containing large aggregate fractions of BAL prepared from control rats or from rats 3 or 28 days after NM administration. Treatment of macrophages with large aggregate fractions prepared 3 days post NM upregulated mRNA expression of M1 markers, including Nos2 and Ptgs2, along with Apoe, an acute phase protein involved in M1-to-M2 macrophage phenotypic switching (Baitsch et al., 2011). In contrast, large aggregate fractions prepared 28 days after NM upregulated the M2 macrophage markers, Il10, Cd163, and Cx₃cr1, as well as genes associated with cholesterol biosynthesis, including Idi1, Hmgcs1, Fdt1, and Fdps (Figure 7). In contrast, no effects were observed on TGFβ expression, indicating that its expression is not regulated by lung lipids in macrophages (not shown). Interestingly, Il12b and Ccr2 were also upregulated, consistent with notion that there is significant heterogeneity within the M2 macrophage subpopulation (Mosser and Edwards, 2008).

Figure 7.

Effects of lipid-containing large aggregate BAL fractions on macrophage expression of markers of activation and cholesterol metabolism. Naïve lung macrophages were incubated for 24 h with large aggregate BAL fractions (50 μg) prepared 3 and 28 days after exposure of rats to NM or to PBS control (CTL), or with Dulbecco’s modified Eagle medium (DMEM) + 1% fetal bovine serum (medium). RNA was then isolated and gene expression analyzed by RT-qPCR. Data are presented as fold change relative to macrophages incubated with medium. Bars, mean ± SE (n = 3–4 replicates/group) from at least 2 independent experiments. *Significantly different (p ≤ .05) from medium control. ^#Significantly different (p ≤ .05) from large aggregate BAL fractions from CTL animals.

Increases in M2 markers in macrophages treated with BAL fractions prepared 28 days post NM were correlated with rises in macrophage neutral lipid content, as measured by Oil Red O staining (Figure 8). Thus, the large aggregate BAL fraction prepared 28 days post NM caused an 8-fold increase in Oil Red O positive macrophages. Conversely, the large aggregate BAL fraction prepared from rats 3 days post administration of PBS control or NM had no major effects on macrophage staining for Oil Red O. The large aggregate BAL fraction from control rats also had no effect on macrophage gene expression.

Figure 8.

Effects of lipid-containing large aggregate BAL fractions of BAL on macrophage neutral lipid content. Naïve lung macrophages were incubated with large aggregate BAL fractions (50 μg) prepared 3 and 28 days after exposure of rats to NM, PBS control (CTL), or DMEM/1% fetal bovine serum (medium). Cytospins were prepared 24 h later and stained with Oil Red O. A, Cytospins of lung macrophages, original magnification, ×1000. B, Fold change in numbers of Oil Red O positive macrophages. A total of 300 macrophages/slide were evaluated microscopically; positive and negative Oil Red O staining was quantified. Data are presented as fold increase in number of Oil Red O positive cells over cells treated with medium. Bars, mean ± SE (n = 3–4 replicates/group) from at least 2 independent experiments. *Significantly different (p ≤ .05) from medium.

DISCUSSION

The appearance of lipid-laden foamy macrophages has been described in pathologies associated with chronic inflammation, including atherosclerosis, septic arthritis, persistent infection, and pulmonary fibrosis (Boven et al., 2006; Dushkin, 2012; Romero et al., 2015). Macrophage development into foam cells is thought to result from altered cellular metabolism and an imbalance between the influx and efflux of lipids, processes regulated by scavenger receptors and lipid transporters such as scavenger receptor A, CD36, ABCA1, and ABCG1 (Fernandez-Ruiz et al., 2016). Foamy macrophages have been shown to amplify fibrogenic signals, in particular in the context of atherosclerotic disease (Thomas et al., 2015). The present studies demonstrate that the formation of foamy macrophages in the lung is due, in part, to uptake of lipids generated following NM exposure. This is based on our findings that NM exposure results in derangements in lung lipids, increased lipid uptake by macrophages, and the appearance of enlarged lipid-laden macrophages in the lung. Moreover, treatment of naïve macrophages with lipid-containing large aggregate fractions of BAL from fibrotic lungs induced an M2 phenotype and the formation of foam cells. These findings provide novel insights into mechanisms mediating macrophage activation in the lung following NM exposure and the development of fibrosis.

In accord with our earlier studies (Malaviya et al., 2012; Venosa et al., 2016), following NM administration, we observed an influx of small round inflammatory macrophages into the lung within 1 day, which closely resembled blood-derived monocytes (Goasguen et al., 2009). With time, these cells became enlarged and vacuolated, features characteristic of macrophage maturation and activation (Daigneault et al., 2010). At 7 and 28 days post NM, macrophages were also highly vacuolated, indicative of phagocytosis of cellular debris, erythrocytes, and/or lipoproteinaceous material (Ishihara et al., 1987; Scott et al., 2003). Our findings that macrophages stained positively for Oil Red O at 28 days are consistent with the idea that these cells are accumulating intracellular lipids. A similar increase in lipid-laden macrophages has been described in the lung following exposure to other fibrogenic toxicants including amiodarone, bleomycin, silica, and thoracic radiation (Romero et al., 2015; Rossi et al., 2017). Our observation that foamy macrophages clustered within areas of thickened and fibrotic tissue provide support for a role of these cells in the fibrogenic process. Of note, only 10%–12% of macrophages recovered by BAL and lung massage 28 days post NM were enlarged, vacuolated and lipid-laden. This may be the result of clustering of the cells within fibrotic areas, making them more difficult to isolate.

Oxidative and nitrosative stress contribute to NM-induced pulmonary toxicity (Malaviya et al., 2012; Sunil et al., 2012, 2014). In addition to directly inducing tissue injury, reactive oxygen and nitrogen species (ROS/RNS) generated following NM exposure drive oxidation of fatty acids and lipoproteins (Zheng et al., 2013), resulting in the generation of products that modulate macrophage activation (Adamson and Leitinger, 2011; Kadl et al., 2010) and fibrogenesis (Allison, 2015; Romero et al., 2015; Ross et al., 2015). In line with this, we found that NM-induced oxidative stress is associated with an accumulation of oxidized phospholipids in lung macrophages and epithelial cells. Although this response was transient in epithelial cells, it persisted for up to 28 days in macrophages. These data suggest that lipid uptake and efflux are differentially regulated in macrophages and epithelial cells following NM-induced injury. Prolonged increases in oxidized lipids in macrophages may also be a consequence of persistent generation of reactive oxygen and nitrogen species by these cells following NM exposure (Fessler and Summer, 2016). This notion is supported by our RNA-sequencing findings showing prolonged upregulation of genes associated with NRF2-mediated oxidative stress and iNOS signaling in macrophages after NM administration.

Of note are our findings that increases in BAL phospholipids and oxidized lipids within macrophages were observed in the lung 3–28 days post NM exposure; these results indicate that NM causes prolonged aberrations in type 2 epithelial cells and oxidative stress, activities thought to be involved in triggering foam cell formation (Kaplan et al., 2012). In contrast, increases in cholesterol were only noted 28 days after NM administration, a time associated with increases in cholesterol biosynthesis and uptake pathways, and the formation of lipid-laden foam cells. Our findings that cholesterol efflux transporters and esterification enzymes are downregulated 1–7 days after NM administration, whereas cholesterol biosynthesis and uptake pathways are upregulated, are consistent with the notion that foam cell formation is driven by macrophage accumulation of intracellular cholesterol and dysregulation of cholesterol esterification and efflux (Maguire et al., 2019).

Transcriptomic analysis of macrophages confirmed our previous findings that proinflammatory signaling pathways related to nitric oxide, TNF-α and the transcription factor, NF-κB, are upregulated in lung macrophages within 1 day of NM exposure (Malaviya et al., 2012, 2015; Sunil et al., 2014). Conversely, at 28 days, coordinate with the appearance of lipid-laden foamy macrophages, alterations in pathways associated with lipid handling were observed; thus, LXR and PPARɣ were downregulated, whereas PPARα and lipid/cholesterol biosynthetic pathways were upregulated. These findings were corroborated by mRNA and/or protein analysis of LXR and its downstream targets, which showed transient decreases in expression of Lxr and lipid efflux genes, Abca1 and Abcg1, and increases in the PPARα targets, Apoa and Apoe (Kidani and Bensinger, 2012). In addition to their role in initiating foam cell formation, oxidized lipids promote inflammation by dampening LXR- and PPARγ-dependent anti-inflammatory activities in macrophages (Fessler, 2017; Ghisletti et al., 2007; Zhao et al., 2010; Zhu et al., 2008). Early downregulation of LXR and PPARγ in the lung after NM is in line with a role of oxidized lipids in promoting a proinflammatory environment and M1 macrophage activation.

Analysis of lung sections by immunohistochemistry also revealed that FXR was constitutively expressed in the cytoplasm of alveolar macrophages; moreover, following NM administration, numbers of FXR positive macrophages increased in the lung. At 3 days, but not 7 or 28 days post NM, FXR was localized in the nucleus, suggesting that it was activated. This may reflect an attempt of the cells to limit proinflammatory M1 macrophage activation via suppression of NF-κB signaling (Wang et al., 2008). Alveolar epithelial cells were also found to express FXR at 3 and 7 days post NM exposure. These findings are consistent with reports that FXR is involved in epithelial cell survival and surfactant lipid homeostasis (Yang et al., 2014). Increases in epithelial cell FXR may help these cells overcome NM-induced cytotoxicity and derangements in lung lipids. Recent reports suggest that FXR plays a role in pulmonary fibrosis; thus, in rat models, the FXR agonist, obeticholic acid, reduced the severity of established fibrosis and protected against the development of fibrosis induced by bleomycin (Comeglio et al., 2017, 2019). This appeared to be due to blunting the production of critical inflammatory mediators. Early upregulation of FXR in the lung after NM exposure may represent a compensatory attempt to limit the fibrogenic process. These findings suggest that pharmacological targeting of FXR may have therapeutic efficacy against NM-induced fibrosis. Pathways meditating increased FXR expression in the lung following NM exposure are unknown. IL-4 and IL-13 have been reported to upregulate FXR expression in human bronchial epithelial cells (Chen et al., 2016). We speculate that these cytokines or other M2 macrophage inducers similarly upregulate FXR in lung macrophages following NM administration; however, this remains to be determined.

Putative FXR targets including the class B oxidized phospholipid scavenger receptor, Cd36, and the lipid chaperone Apoe (Li et al., 2012; Mak et al., 2002; Silverstein et al., 2010), were also increased in macrophages at 3 and 7 days post NM. CD36 is known to bind native and oxidized LDL and other polyanionic lipids promoting their uptake into cells (Dahl et al., 2007; Ji et al., 2011). Our findings of increased CD36 protein expression in lung macrophages at 3 days, a time coordinate with increases in oxidized phospholipids in the lung, suggest that CD36 plays a role in lipid uptake during the acute inflammatory response to NM-induced injury. Evidence suggests that CD36 is also involved in promoting sterile inflammation and fibrogenesis (Pennathur et al., 2015) and it may play a similar role in NM-induced pulmonary toxicity. This is supported by earlier studies demonstrating proinflammatory functions of macrophage scavenger receptors during acute inflammatory responses in experimental models of stroke, atherosclerosis, and Alzheimer’s disease (Abumrad and Goldberg, 2016; Coraci et al., 2002; Kim et al., 2015; Oury, 2014).

Pulmonary lipid profiles have been reported to be distinct during acute inflammatory responses and fibrogenesis (Bernhard et al., 1997; Scaccabarozzi et al., 2015; Swendsen et al., 1996). In this regard, the present studies demonstrate that lipids generated in the lung at 3 days (acute injury) and at 28 days (chronic injury) following NM exposure exert diverse effects on macrophage cellular metabolism and phenotype. Thus, lipid-containing large aggregate fractions of BAL prepared 28 days, but not 3 days post NM, significantly upregulated macrophage expression of genes associated with cholesterol biosynthesis, including Hmgcs1, Fdft1, Idi1, and Fdps. Also observed were increases in Il10 and the scavenger receptor Cd163, markers of an anti-inflammatory/profibrotic M2 phenotype (Martinez and Gordon, 2014; Moestrup and Moller, 2004), along with the fractalkine receptor Cx₃cr1, which is involved in trafficking of M2 macrophages and their monocytic precursors to sites of injury (Ingersoll et al., 2011). This was correlated with increased numbers of Oil Red O positive lipid-laden M2 macrophages in the lung. Conversely, BAL large aggregate fractions prepared 3 days after NM administration caused lung macrophages to develop a proinflammatory M1 phenotype, characterized by high expression of Nos2 and Ptgs2. These findings are in accord with our earlier studies demonstrating that M1 macrophages accumulate in the lung early after NM-induced injury (within 3 days), whereas M2 macrophages accumulate at later times (14–28 days) in the pathogenic response (Malaviya et al., 2012; Venosa et al., 2016), and confirm that lipids generated following NM exposure participate in macrophage phenotypic activation. We also found that Apoe expression was upregulated in lung macrophages after exposure to lipid-rich BAL fractions collected 3 days post NM, consistent with the idea that in addition to functioning as an extracellular lipid binding protein (Steinberg et al., 1996), ApoE is involved in driving macrophage M1–M2 phenotypic switching (Baitsch et al., 2011). Interestingly, no changes in Ccr2 expression were observed in lung macrophages following treatment with lipid-containing BAL fractions prepared 3 days after NM exposure. In contrast, Ccr2, as well as Cx₃cr1, were markedly upregulated in macrophages in response to lung lipid fractions collected at 28 days. A number of reports have described a subset of anti-inflammatory M2 macrophages that coexpresses CCR2 and CX₃CR1 and exhibits enhanced mobilization from the periphery to inflammatory sites (Kohno et al., 2015; Schmall et al., 2015; Willenborg et al., 2012). It remains to be determined if a similar subset of M2 macrophages accumulates in the lung following NM exposure. We cannot exclude the possibility that proteins present in the large aggregate BAL fraction contribute to some of the changes observed in macrophages. Future studies characterizing the composition of the large aggregate BAL fraction at baseline and during NM-induced injury may help elucidate specific lipid moieties that drive macrophage phenotypic changes.

In summary, the present studies demonstrate that NM exposure is associated with derangements in lung lipid handling, a process that contributes to phenotypic activation of macrophages and foam cell formation. Thus, lipids from acutely injured lung induce a proinflammatory macrophage phenotype, whereas lipids generated during fibrogenesis induce an anti-inflammatory M2 phenotype and the formation of macrophage foam cells. The fact that enlarged lipid-laden macrophage foam cells were clustered in areas adjacent to fibrotic loci, suggest a potential contribution of these cells to the fibrogenic processes. Additional studies analyzing macrophages separated by size or density may provide further insights into the precise role of these cells in the pathogenic response to NM. One limitation of our studies is the fact that we did not distinguish between resident alveolar macrophages and monocyte-derived macrophages, or between different subsets of inflammatory macrophages in our analyses. Currently, these studies have limited feasibility, as commercial reagents specific for rat macrophages are not readily available. Nevertheless, our findings provide novel insights into mechanisms mediating acute lung injury and pulmonary fibrosis following mustard exposure and may lead to the development of novel therapeutic approaches including the use of hypolipidemic agents (eg, statins) to mitigate vesicant-induced toxicity.

DECLARATION OF CONFLICTING INTERESTS

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.