Differential Host Pro-Inflammatory Response to Mycobacterial Cell Wall Lipids Regulated by the Mce1 Operon

The cell wall of wild-type (WT) Mycobacterium tuberculosis (Mtb), an etiologic agent of tuberculosis (TB) and a Mtb strain disrupted in a 13-gene operon mce1 (Δmce1) varies by more than 400 lipid species. Here, we examined Mtb lipid-induced response in murine macrophage, as well as in human T-cell subpopulations in order to gain an insight into how changes in cell wall lipid composition may modulate host immune response. Relative to WT Mtb cell wall lipids, the non-polar lipid extracts from Δmce1 enhanced the mRNA expression of lipid-sense nuclear receptors TR4 and PPAR-γ and dampened the macrophage expression of genes encoding TNF-α, IL-6, and IL-1β. Relative to untreated control, WT lipid-pre-stimulated macrophages from healthy individuals induced a higher level of CD4−CD8− double negative T-cells (DN T-cells) producing TNF-α. Conversely, compared to WT, stimulation with Δmce1 lipids induced higher mean fluorescence intensity (MFI) in IL-10-producing DN T cells. Mononuclear cells from TB patients stimulated with WT Mtb lipids induced an increased production of TNF-α by CD8+ lymphocytes. Taken together, these observations suggest that changes in mce1 operon expression during a course of infection may serve as a strategy by Mtb to evade the host pro-inflammatory responses.


INTRODUCTION
Despite the availability of effective treatment regimens against Mycobacterium tuberculosis (Mtb), an etiologic agent of tuberculosis (TB), TB remains a major public health challenge, which has surpassed AIDS as the most common infectious disease cause of death in adults (1). A large proportion of the world's population latently infected with Mtb serves as the main reservoir of reactivation TB (2). The mechanisms by which Mtb remains persistent in a host and how a subset of the infected hosts reactivates to develop TB are still not fully understood.
Mtb contains four homologous copies of an operon called mce1-4, which resemble ATP-binding cassette transporters possibly involved in lipid important across its cell wall (3-5). Disruption of one of these operons (mce1) precludes Mtb to induce a strong Th1 type T-cell immune response and to form organized granuloma in mouse lungs (6). The expression of mce1 operon is repressed in the WT strain during the first 8 weeks of mouse infection (7) and its expression decreases 4 h post infection in macrophages (8). Thus, at some point during in vitro macrophage or mouse infection, the WT strain displays the same phenotype shown by the mce1 operon mutant ( mce1).
The cell wall lipid content of Mtb is highly affected by the repression of the mce1 operon. The mutant strain accumulates several-fold greater amount of free mycolic acids on its surface (9,10). As revealed by lipid metabolomics analysis, the mutant also contains diminished amounts of diacyltrehalose, diacylated sulfoglycolipid, phosphatidylethanolamine, phthienoic acid, and phthioceranic acid in its cell wall compared to its parent wild-type (WT) strain. In fact, a lipidomics analysis has shown differential expression of more than 400 lipid species between the mutant and WT Mtb strains (11).
The cell wall lipid rearrangement during infection could be a determining factor for bacterial persistence, as previously suggested (12). Individually, such lipids have been shown to induce granuloma formation (13), leukocyte migration (14) and inflammatory cytokines expression (15), or to inhibit macrophage responses (16)(17)(18). However, the evaluation of host cell behavior in response to Mtb cell wall lipid rearrangement has never been assessed. Here, we hypothesized that by comparing responses of host immune cells challenged in vitro with lipids extracted from the mce1 operon mutant or WT Mtb, we may be able to demonstrate how differences in cell wall lipid composition affect host cell responses that would be advantageous to Mtb.

Bacterial Strains, Media, and Growth Conditions
The following bacterial strains were used: mce1 operon mutant M. tuberculosis ( mce1) and its parent Erdman wild-type (WT) strain. The generation of mce1 was previously described by Shimono et al. (6). Both Mtb strains were initially cultured in Middlebrook 7H9 broth (Difco, MD) containing 10% ADC (Beckton-Dickinson, MD) and 0.2% glycerol (Fisher Scientific, NJ) to obtain similar numbers of bacterial cells for growth in detergent-free minimal media. Bacteria were incubated at 37 • C until stationary phase in Sauton's media (without Tween) as previously described (19). Briefly, bacteria were grown in 125 mL polycarbonate bottles containing 30 mL of Sauton's media with 300 µl of saturated planktonic culture (OD of each test strain was adjusted so that equal numbers of each bacteria strain were inoculated) and incubated without agitation at 37 • C for 19 days.

Lipid Extraction
The extraction of lipids from biofilm cultures was performed as previously described (19). Briefly, biofilm from 19-day M. tuberculosis cultures was harvested, and apolar lipids were extracted from using 5 mL of methanol: 0.3% NaCl (100:10) mixed with 2.5 mL of petroleum ether, incubated at room temperature for 30 min. The upper petroleum ether layer containing the apolar lipids was separated by centrifugation. After solvent evaporation, apolar lipids were weighed and resuspended at a concentration of 0.02 mg/mL in hexane/isopropanol. Lipid extracts (0.5 mL) were layered onto 24-well tissue culture plates (0.01 mg/well) and the solvent was allowed to evaporate. Control wells were layered only with hexane/isopropanol in the absence of lipid extracts.

RAW Macrophage Assay
RAW 264.7 Murine Macrophage Cultures RAW 264.7 murine macrophage-like cells (ATCC TIB-71) were cultured and maintained in Dulbecco's modified Eagles medium (DMEM; Gibco) supplemented with 10% FBS at 37 • C, 5% CO 2, in a humidified incubator. Macrophage cell number and viability were assessed by staining in a trypan blue (Gibco) exclusion assay. Prior to experimentation, macrophages were first seeded in 75-cm 2 flasks overnight to achieve 70% confluency, and then seeded onto Mtb lipid-coated 24-well tissue culture plates at a concentration of 3.7 x 10 5 cells/well, and incubated at 37 • C in a 5% CO 2 humidified atmosphere for 72 h. Control macrophages were cultured on layered wells with hexane/isopropanol, and incubated in the absence of lipid extracts.

RNA Extraction and Purification
RNA was extracted from RAW cells according to a standard Trizol RNA extraction protocol supplied by Invitrogen (Invitrogen, Life Technologies). Extracted RNA was treated with DNase (Qiagen) to ensure that no DNA was present in the samples. DNA-free RNA (500 ng) was mixed with 50 ng of random hexamers and 50 µM of oligo (dT) (Invitrogen) at a final volume of 10 µl, then reverse transcribed to cDNA with Superscript III reverse transcriptase (Invitrogen) following the manufacturer's recommendations.
qPCR Nine genes encoding lipid-sensing nuclear receptors (LSNR) were selected as targets, in addition to genes related to proinflammatory response (complete list of genes available in Supplementary Table 1). Primers were designed to produce a 100-195 bp amplicon for each gene. qPCR reactions were performed using 25 ng of previously generated cDNA and Maxima SYBR Green/ROX qPCR Master Mix (2X) (Fermentas) in accordance with the manufacturer's instructions. Relative changes in gene expression between lipid-induced and unstimulated RAW cells (controls) were analyzed by 2 − CT method according to a previously described method (20). The expression of all tested genes was normalized to both βactin and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene expression.

Study Participants
Seventeen participants were enrolled in this study, divided into healthy subjects (n = 11) and TB patients (n = 6). All healthy volunteers, recruited from the Gonçalo Moniz Institute (IGM-Fiocruz), were tested for latent TB infection (LTBI) by interferon γ release assay (QuatiFERON R TB Gold); two of them were tested positive for LTBI. Six TB patients were recruited at the Octavio Mangabeira Hospital (HEOM), a reference center for respiratory disease in Salvador, Bahia-Brazil. All patients were diagnosed with active TB based on GeneXpert MTB/RIF test results. The age of the study subjects ranged between 18 and 60 years and all participants were HIV negative.

PBMC Isolation From Study Subjects and Culture Conditions
Peripheral blood mononuclear cells (PBMC) were obtained by the Ficoll-Paque (GE Healthcare) density gradient method. PBMC concentrations were adjusted to 1 x 10 6 cells/mL in 1 mL of complete RPMI 1640 (100 µl/mL gentamicin, 2 mM Lglutamine, 30 mM HEPES) containing 10% inactivated bovine fetal serum (FBS) (Life technologies GIBCO BRL, Gaithersburg, MD). PBMCs from TB patients were cryopreserved (FBS containing 10% DMSO) in liquid nitrogen prior to performing culture assays. All PBMCs were dispensed into previously prepared 24-well plates and incubated at 37 • C under 5% CO 2 .

Co-culturing of Adherent and Non-adherent Cells
Adherent cells were isolated from six healthy subjects PBMC (2 × 10 7 ) on sterile 6-well tissue culture plates. After 2 h of adherence, cells in suspension (non-adherent cells) were harvested and cryopreserved (FBS with 10% DMSO) at −80 • C. Adherent cells were kept on 6-well plates for 5 days at 37 • C under 5% CO 2 . After differentiation, macrophages were collected by washing the wells with 2 mM EDTA saline at 4 • C, with the aid of a cell-scraper. Cells (1 × 10 6 ) were then cultured on 24-well tissue culture plates in RPMI medium supplemented with 10% FBS in the presence or absence of Mtb non-polar lipids for 72 h at 37 • C under 5% CO 2 . Non-adherent cryopreserved cells were then added to prestimulated macrophage cultures at a ratio of 10:1 (non-adherent cells:macrophage), incubated for 72 h at 37 • C under 5% CO 2 , and finally stained to perform flow cytometry.

PBMC Assay With TB Patients
Cryopreserved PBMCs from TB patients (1 × 10 6 ) were thawed and incubated for 72 h on 24-well tissue culture plates pre-treated with lipids at 37 • C under 5% CO 2 in the presence or absence of Mtb non-polar lipids (0.01 mg/well) extracted from WT and mce1 strains. After incubation, cells were stained for flow cytometry analysis.

Cytokine Analysis
Supernatants of PBMCs from healthy subjects and TB patients were collected and stored at −20 • C for cytokine analysis (TNFα, IL-6, IL-1β, IFN-γ, and IL-10) by either ELISA (R&D Systems, Minneapolis, MN) or Cytometric Bead Array (BD TM CBA Human TH1/TH2 cytokine kit) according to the manufacturer's instructions. Results are expressed as pg/mL.

Statistical Analysis and Data Representation
Statistical analyses were performed with GraphPad Prism v.7.0 (GraphPad Inc., San Diego, CA, USA). Statistical significance between variables was assessed by Mann-Whitney U test or Friedman test, followed by Dunn's post-test. Continuous variables with normal distribution were expressed as mean ± standard deviation, while those with non-normal distributions were expressed as median and interquartile interval. Significance was considered when p < 0.05.

Differential Induction of Lipid Sense Nuclear Receptors (LSNR) by Mtb Lipids
Mtb interacts with LSNR to modulate macrophage immune and metabolic functions (21,22). Since mycobacterial lipid extracts modulated pro-inflammatory response in macrophages, we attempted to determine whether LSNR activation was involved in this process. Here, we studied the expression by macrophages of LSNR PPAR-γ, TR4, RAR, and LXR-α genes. As depicted in Figure 1B, while the WT lipid extracts consistently repressed PPAR-γ , TR4, and RAR expression by 0.6-, 0.8-, and 0.7-fold, respectively, the lipid extracts of the mutant strain enhanced the expression of these same genes by 1.4-, 1.2-, and 1.2fold, respectively, relative to untreated controls. The resulting mce1/WT expression ratios were, respectively, 2.3, 1.5, and 1.7. Fold-differences between groups were statically significant for TR4 and RAR (p < 0.05). Both lipid extracts, WT and mce1, induced similar expression levels of another LSNR gene, LXR-α (1.3-and 1.5-fold, respectively). FIGURE 1 | qPCR analysis of selected cytokine genes (A) and lipid-sense nuclear receptor (LSNR) genes (B) in murine macrophages exposed to Mtb apolar lipid extracts. Data are presented as fold-difference in gene expression levels in cells stimulated with apolar lipids extracted from wild type or mce1 operon mutant M. tuberculosis, relative to untreated control, mean ± SD; gene expression was normalized to β-actin and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) genes; six replicates of each reaction were performed. Statistical significance was evaluated by Mann-Whitney U test. p < 0.05. *p < 0.05; **p <0.01. WT, Mycobacterium tuberculosis wild type strain; mce1, strain with disruption of mce1 operon.
Macrophages Pre-stimulated by Mycobacterial Lipids Activate Both CD4 + and Double-Negative (CD4 − CD8 − ) T Cells A co-culture assay was performed with both non-adherent and lipid-pre-treated adherent cells isolated from the peripheral blood of healthy subjects. Following the activation of IFNγ-, TNF-α-, and IL-10-producing CD4 + and CD8 + T cells by lipid-stimulated macrophages, the proportion of IL-10-producing CD4 + was found to be higher in mce1 lipid-induced cells than in those stimulated by WT (p < 0.05) (Supplementary Figure 1A). Also, compared to untreated control, stimulation with mce1 lipids induced lower mean fluorescent intensity (MFI) in TNF-α-producing CD4 + T cells (p < 0.05) (Supplementary Figure 1A). However, neither the frequency nor the MFI of CD8 + T cells changed (Supplementary Figure 1B).
The co-culture assay showed that pre-stimulation with both WT and mce1 lipids also increase the frequency and MFI of co-cultured DN T cells (Figure 2), which were defined as CD3 positive but CD4 and CD8 negative (CD3 + CD4 − CD8 − ) (Figure 2A). The proportion of TNF-α-producing DN T cells was only found to be higher in WT lipid-induced cells than in non-stimulated controls (p ≤ 0.01) ( Figure 2B). Interestingly, compared to WT, stimulation with mce1 lipids induced higher MFI in IL-10-producing DN T cells (p ≤ 0.01) (Figure 2C). Results are expressed as median ± IQR; n = 6/group. Statistical significance was determined by Friedman test followed by Dunn's post-test; significance was considered at **p < 0.01. WT, Mtb wild type strain; mce1, Mtb strain disrupted in mce1 operon; CTL, no stimulus.

Characterization of Lipid-Induced Response in PBMCs Isolated From Patients With Active TB
Next, to evaluate the effects of stimulation with WT and mce1 lipid extracts on the inflammatory response in PBMCs isolated from patients with active TB (Figure 4), the frequency of CD4 + , CD8 + , CD56 + , NKT, CD4 − /CD8 − DN, CD38 + , and CD14 + subpopulations were assessed, in addition to the production of IFN-γ, TNF-α, and IL-10 by these cells (see Supplementary Figure 2A). Compared to untreated control, stimulation with mce1 lipids induced lower MFI in IFN-γ-producing DN T cells (p ≤ 0.05) (Supplementary Figure 2B). In contrast to the response seen in DN T cells from healthy subjects stimulated with Mtb lipids, no significant differences were observed in the Mtb lipid-stimulated TNF-α-and IL-10-producing DN T cells from patients with active TB compared to untreated cells.

DISCUSSION
Mice infected with a strain of M. tuberculosis disrupted in the mce1 operon are unable to mount a strong pro-inflammatory response against WT Mtb, as evidenced by decreased levels of IL-6 and TNF-α; consequently, these animals do not form organized granulomas (6). Recent studies have demonstrated that free mycolic acid accumulates on the surface of Mtb mutant cells (9,11), and that the cell wall contains decreased levels of some saccharolipids and glycerophospholipids species. Only residual or no alteration were observed in the complemented strain, relative to WT (11). As the mce1 operon of WT Mtb becomes repressed during the course of infection in a mouse model of TB (7), we reasoned that the lipids in the cell wall of this mce1 mutant could be representative of the lipid composition of WT Mtb at some point during a course of in vivo infection. Thus, we hypothesized that the Mce1-mediated remodeling of lipids in the Mtb cell wall may contribute to the differential host proinflammatory response that could facilitate the persistence of Mtb in the infected host.
When stimulated with WT or mce1 lipid extracts, the RAW cells evaluated herein exhibited opposite expression responses in genes encoding TNF-α, IL-6 and IL-1β (Figure 1A), as lipids extracted from the mce1 operon mutant strain were found to virtually reverse the pro-inflammatory response triggered by stimulation with WT lipids. One mechanism for this modulation of inflammatory response may involve the mutant's lipids interacting with LSNRs PPAR-γ, TR4 and RAR (Figure 1B). Mtb lipids can serve as ligands of LSNRs, a large family of intracellular proteins responsible for regulating gene transcription (23). Previous studies have shown that Mtb infection is associated with increased expression of PPAR-γ and TR4 leading to Mtb survival through IL-10 production and prevention of phagolysosome maturation (21,22). Also, PPAR-γ ligands inhibit monocyte production of TNF-α, IL-6, and IL-1β (24) and the transcriptional activation of the pro-inflammatory transcription factor NF-κB in murine macrophage cells (25). Also, TR4 is sensed by mycolic acids (26), which is greatly increased in amount in the cell wall of the mutant strain (9,12). Results are expressed as median ± IQR; n = 6/group Statistical significance was determined by Friedman test followed by Dunn's post-test; significance was considered at *p < 0.05; **p < 0.01. WT, Mtb wild type strain; mce1, Mtb strain disrupted in mce1 operon; CTL, no stimulus.
FIGURE 5 | Schematic representation of Mycobacterium tuberculosis (Mtb) lipid-induced host inflammatory adaptation in macrophage. During infection and 4 h after phagocytosis by the macrophage, the mce1 operon of Mtb is naturally repressed (8) and the bacterial cell undergoes cell wall lipid rearrangement (12). At the early phase of infection, the WT lipids induce the macrophage to secrete TNF-α, IL-1β, and IL-6. Also, these lipids are presented by CD1 to TNF-α-producing CD4 − CD8 − double negative (DN) T cells. Together, these secreted cytokines should recruit immune cells to the site of infection. Conversely, during the later phase of infection, the mce1 lipids interact with the LSNR PPAR-γ and TR4, which in turn may negatively modulate the secretion of pro-inflammatory cytokines. These lipids are also presented by CD1 to IL-10-and IFN-γ-producing DN T cells. WT, Mtb wild type strain; mce1, Mtb strain disrupted in mce1 operon; TR4, testicular nuclear receptor 4; PPAR-γ, Peroxisome proliferator-activated receptor gamma.
On the other hand, we found that lipid extracts from both Mtb strains induced a similar expression level (1.3-and 1.5fold) of another LSNR LXR-α ( Figure 1B). This transcriptional factor triggers a protective immune response against Mtb (22,27) and it is activated by oxysterols, the oxygenated derivatives of cholesterol (28,29). As far as known, oxysterols are not included among those mycobacterial cell wall lipids that are altered by Mce1 repression (11), which may explain why we did not see a difference in the expression of LXR-α by cells exposed to lipids from the WT or mutant Mtb (30).
Macrophages can present mycobacterial lipids via the CD1 (CD1a-CD1e) family of MHC-class-I-like glycoproteins to T cells (27). Depending on their species, these lipids can be presented by CD1a, CD1b, or CD1c to either: (A) CD4 − CD8 − double negative or to CD4 + and CD8 + αβ TCR clonally diverse T-cells that mediate adaptive immunity, or (B) be presented by CD1d molecules to natural killer T (NKT) cells (15,31). Thus, since the lipid rearrangement on Mtb cell wall dampens the macrophage inflammatory response, we wondered if this modulation would determine the pattern of T cell activation. Here, compared to WT, mce1 lipid-pre-stimulated macrophages induced higher frequency and MFI level of, respectively, IL-10-producing CD4 + and DN T cells (Supplementary Figure 1 and Figure 2).
Conversely, relative to untreated control, WT lipids induced a higher proportion of TNF-α-producing DN T-cells and mce1 lipids decreased the MFI level in TNF-α-producing CD4 + T-cells. CD8 + T-cells, isolated from healthy subjects, were not activated by lipid-pre-stimulated macrophages (Supplementary Figure 1). Taken together, these results suggest that the Mtb's cell wall lipid-rearrangement modulates the macrophage's inflammatory response and, indirectly, determine the function of both CD4 + and DN T-cells. In Figure 5, we propose a model of mce1 operon-mediated differential Mtb lipid-induced host inflammatory control in macrophage.
We have not determined which DN T-cell population subsets respond to each lipid extract used here, but we can speculate that WT lipids would induce the production of TNF-α through αβ DN T-cells, while the mce1 lipids may activate the IL-10-producing γδ DN T-cells by lymphocyte-dependent immunoregulatory mechanisms, as previously suggested (32).
In contrast to the results with healthy donors, in PBMC collected from TB patients, the MFI of TNF-α-producing CD8 + T-cells induced by WT lipids was 3-fold higher than in non-stimulated or mce1 lipids-stimulated PBMC, suggesting distinct roles of Mtb lipids in activating T lymphocytes subpopulations in healthy and TB subjects. The TNF-α production by CD8 + T-cells is supported by the increased level of this cytokine in culture supernatants of WT lipid-induced PBMC relative to untreated control, suggesting that CD8 + T-cells stimulated by mycobacterial lipids maybe an important source of TNF-α during active disease.
In vivo studies showed that the expression of mce1 operon is repressed in the WT strain during the infection in both mouse (7) and macrophages (8) due the overexpression of mce1R gene. Interestingly, previous in vitro studies showed that under a hypoxic growth condition, the level of cell wall free mycolic acids in Mtb increases while the levels of trehalose mono (TMM) and dimycolate (TDM) decrease and re-aeration reverses such changes (33,34). TDM is widely known as cord factor and serve as potent triggers of host inflammation (35). Taken together, we can speculate that the hypoxic environment in tuberculosis granuloma may trigger a Mtb cell wall lipid remodeling via repression of mce1 operon. LSNR may serve as one component of this process to negatively dampen macrophage pro-inflammatory response, and perhaps indirectly influence T-cell-dependent proinflammatory responses. This lipid-mediated microbial defense strategy may contribute to Mtb's long-term survival in an infected host.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding author.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Gonçalo Moniz Institute (IGM-Fiocruz, Bahia) CAAE: 76009417.9.0000.0040. The patients/participants provided their written informed consent to participate in this study.