H37Rv-MTB was obtained from American Type Culture Collection (ATCC, Manassas, VA). HN878 W-Beijing strain of MTB (HN878-W-Beijing MTB) was originally a kind gift from Dr. B. Kreiswirth (Public Health Research Institute Center, Newark, NJ). The B6.PL(Thy1.1) and Foxp3⁺GFP⁺DTR⁺ (Thy1.2) mice were purchased from Jackson Laboratories, Bar Harbor, Maine. Cell Preparation Tubes (CPT) used to isolate peripheral blood mononuclear cells (PBMC) were purchased from BD Company (San Jose, CA). The Miltenyi Biotech Regulatory T Cell Isolation Kit II (human) (Auburn, CA) was used to isolate human Tregs from peripheral blood. 4′,6-diamidino-2-phenylindole (DAPI) and LysoTracker Red DND-99 were purchased from Invitrogen, Carlsbad, CA. Macrophage colony-stimulating factor (M-CSF, human) was purchased from Millipore Sigma, St. Louis, MO. Fetal bovine serum was acquired from Atlanta Biologicals (Norceross, GA) and inactivated at 56°C for one hour. The LC3-IIB and p62 primary rabbit antibodies were purchased from Cell Signaling Technology Inc. (Danvers, MA). Roswell Park Memorial Institute (RPMI) medium, CY3 goat anti-rabbit IgG (H+L) Cross-Adsorbed Secondary Antibody, and the four and eight chambered Nunc™ Lab-Tek™ II Chamber Slide™ System were purchased from ThermoFisher Scientific (Waltham, MA). PE-conjugated anti-CTLA-4 and APC-conjugated anti-CD279 (PD-1) were purchased from R&D System (Minneapolis, MN). Human cytokines from cell culture supernatants were analyzed using ELISA kits for TNF, IL-10, and TGFβ (R&D Systems, Minneapolis, MN and Invitrogen-ThermoFisher Scientific, Waltham, MA). Anti-CTLA-4 neutralizing antibody and non-immune human IgG antibody were purchased from BPS Bioscience Inc (San Diego, CA). Diphtheria toxin was purchased from List Biological Laboratories, Campbell, CA. Information on other reagents are discussed in the relevant methods sections below.

CS extract was prepared by mechanically “smoking” a single unfiltered 3R4F cigarette into 10 mL of pre-warmed RPMI medium via a vacuum pump, with the flow adjusted at 3 L/min using an Accucal flowmeter (Gilmont Instruments). The extract was then filtered through a 0.22 µm syringe filter and adjusted to a pH of 7.4. This solution is arbitrarily designated as 100% CS extract (Ouyang et al., 2000).

After informed consent (Colorado Multiple Institutional Board Review, protocol 16-1413), blood was drawn into six Cell Preparation Tubes^® from each healthy donor. Peripheral blood mononuclear cells (PBMC) were isolated by centrifugation. Following the determination of the total cell count, half of the PBMC aliquot was cryopreserved in freeze medium (85% FBS, 15% DMSO) for future Treg isolation. The other half was incubated with 20 ng/mL of monocyte-colony stimulating factor for one week to prepare monocyte-derived macrophages (MDM).

One day before completion of MDM differentiation, an autologous cryopreserved PBMC aliquot was thawed and then centrifuged to remove the DMSO. Tregs were isolated from PBMC using the Regulatory T Cell Isolation Kit II according to manufacturer’s instructions. In brief, non-CD4⁺ and CD127^high cells were magnetically labeled and then depleted. From the remaining cells, CD25⁺ cells were magnetically labeled and then isolated, yielding only cells with biomarkers CD4⁺CD25⁺CD127^dim/-, the cell surface signature of Tregs.

To confirm that Tregs were isolated, the cells were stained with CD127-BB515, CD25-PE, and CD4-APC-CY7, and analyzed with LSR Fortessa equipped with FACSDiva (BD Biosciences) and with the aid of FlowJo software. As shown in Supplementary Figure 1 , essentially all the isolated cells had markers entirely consistent with Tregs.

Differentiated MDM were infected with MTB H37Rv at a multiplicity-of-infection of 10 MTB:1 macrophage for one hour, and then washed gently to remove any free MTB. MDM were then incubated with: (i) RPMI medium + 10% FBS alone or with (ii) non-conditioned Tregs or (iii) Tregs previously conditioned in 5% CS extract for 18 to 24 hours at a ratio of 100 monocytes: 1 Treg. This ratio is the approximate ratio of lung macrophages:Tregs found in the bronchoalveolar lavage of healthy individuals (Heron et al., 2011). After an additional hour of incubation, Day 0 cells were washed, lysed, serially diluted, and plated on 7H10 agar to quantify MTB. Different aliquots of cells – incubated with MTB for one hour and then washed – were then incubated under the same three aforementioned conditions for 2 or 4 days before quantifying MTB.

PBMC were differentiated into MDM in four-chamber glass slides. After one week, the MDM were co-cultured with autologous Tregs (100 MDM:1 Treg ratio) that had been previously cultured in medium alone or with 5% CS extract for 18 to 24 hours. MDM alone and MDM co-cultured with air- or CSE-exposed Tregs were infected with GFP-MTB H37Rv at a MOI of 10 MTB:1 macrophage for six hours. At the 4-hour mark, 25 µL of LysoTracker Red (50 nM) was added to each chamber well containing 500 µL of medium to label the lysosomes. At the 6-hour time point, the medium was removed, the cells fixed with 4% paraformaldehyde, washed thrice with 1X PBS, stained with DAPI and stored overnight to dry in the dark at room temperature. The following day, the cells were viewed under 40X oil immersion lens with an inverted Zeiss 200M microscope and a 175-watt xenon lamp in a DG4 Sutter instrument lamp housing. The Cy3 filter was used to detect the LysoTracker-stained lysosomes while FITC was used to detect GFP-MTB. Co-localization of the GFP-MTB with lysosomes would appear yellow. Ten random pictures were taken per well with Live-Cell Marianis. Images were further analyzed using FIJI software. Phagosome-lysosome (P-L) fusion was calculated by dividing the number of cells with evidence of GFP-MTB and lysosome fusion by the total number of cells containing GFP-MTB, the latter cell type comprised of at least 100 cells per condition (Bai et al., 2019).

MDM were co-cultured in four-chamber glass slides with Tregs (100 MDM:1 Treg) that were previously exposed to medium alone or 5% CS extract for 18 to 24 hours. After one hour of incubation with Tregs, the cells were infected with GFP-MTB (MOI of 10 MTB:1 macrophage) and incubated for 18 hours at 37°C and 5% ambient CO₂. The cells were then washed with PBS, fixed with 4% paraformaldehyde for 30 minutes, washed, incubated with permeabilization buffer (0.5% Triton X in PBS) for 10 minutes, rinsed with PBS, and incubated with blocking buffer (5% BSA, 0.5% Tween 20 in PBS) for one hour. Anti-LC3 antibody at a dilution of 1.5:200 was incubated with the cells and left in a dark container at 4°C overnight. An anti-rabbit CY3 tagged antibody (1:1000) was added to the wells, left in the dark at room temperature for 45 minutes, and the cells were fixed with DAPI. The slides were dried overnight and viewed with the Live-Cell Marianis microscope to analyze and quantify autophagosome formation. Multiple images were taken from all conditions and five were picked at random. The number of LC3-II-positive puncta per cell was determined and averaged for 150 randomly chosen cells. The same process was replicated for all three conditions.

To determine autophagosome maturation, we quantified p62-positive puncta in a temporal fashion of MDM, MDM + unexposed Tregs, and MDM + CS extract-exposed Tregs that were infected with GFP-MTB for 3, 6, and 18 hours. At the end of each time point, the cells were washed with 1X PBS and fixed with 4% paraformaldehyde for 30 minutes. The cells were washed, permeabilized for 10 minutes, rinsed with 1X PBS, and incubated with a blocking buffer for one hour. Lastly, 25 µL of anti-p62-rabbit antibody [1:200] was added to each chamber well and incubated overnight at 4°C. Then CY3 anti-rabbit [1:1000] secondary antibody was added and incubated for 45 minutes in the dark. After incubation, slides were rinsed, DAPI added, and once dried, the cells were viewed under a 40X oil emersion lens. CY3 filter was used for detecting p62 and FITC for GFP-MTB. Further analyses of p62 in the presence of 200 nM rapamycin and 100 nM bafilomycin were conducted similarly.

The B6.PL(Thy1.1) mice were exposed to CS in a whole animal exposure chamber using smoke generated by the TE-10c cigarette smoking machine (Teague Enterprises) as previously described (Shang et al., 2011b). The relative proportion of smoke particulate, vapors, and gases generated approximate environmental CS exposure (89% side-stream smoke exposure and 11% mainstream CS exposure), maintaining a Total Suspended Particulate concentration of 85-120 mg/m³ for 5 hours a day and 5 days a week for four weeks.

Ten days before the adoptive transfer of Tregs, endogenous Tregs of the recipient Foxp3⁺GFP⁺DTR⁺ (Thy1.2) mice were depleted via administration of diphtheria toxin at 50 µg/kg x four doses intraperitoneally (Henao-Tamayo et al., 2016). Tregs of the air- and CS-exposed B6.PL(Thy1.1) mice were harvested by cell sorting. To sort CD4⁺CD25^hiCD127⁻ and CD4⁺CD25⁻ cells, single cell suspension was prepared from the spleens of donor B6.PL(Thy 1.1) mice. After incubating with 5 µg/mL Fc block (ebioscience) for 20 min at 4°C, the cells were attained with APC labeled anti-CD4 (clone gk1.1; ebioscience), Pe Cy7-labeled anti-CD25 (clone PC61.5; ebioscience), Alexa ef450 labeled CD127 (eBioRDR5; ebioscience) and FITC-labeled anti-CD3 (clone 17A2; ebioscience). Thereafter, the cells were stained with 0.5 µg/ml 7-AAD (eBioscience) for 5 min, sorted on a FACS Aria III (BD Biosciences) with a 70-μm nozzle, doublets gated out by FSC-A vs. FSC-H, and 7-AAD⁺ dead cells excluded from the singlet population using FL3. The live CD4⁺CD25^hiCD127⁻ or CD4⁺CD25⁻ were individually sorted. One million donor Tregs were transferred to each Treg-depleted Foxp3⁺GFP⁺DTR⁺ (Thy1.2) mouse via intravenous tail vein injection.

Once the Foxp3⁺GFP⁺DTR⁺ (Thy1.2) mice received the Tregs from the donor B6.PL(Thy1.1) mice, the recipient mice were infected with HN878 W-Beijing strain of MTB using the Glas-Col Aerosol Generator at an inoculum dose of 2x10⁶ CFU/mL. The mice were exposed to an aerosol infection in which approximately 100 bacteria were deposited in the lungs of each mouse.

One, 30 and 60 days after infection, the mice were sacrificed by CO₂ inhalation, and their lungs and spleen were prepared to quantify MTB (CFU) and immune cell phenotypes by flow cytometry and analyze the lungs by histopathology. The lungs and spleens were homogenized in saline and serial dilutions were plated on 7H11 agar plates supplemented with OADC (BD Biosciences, San Jose, CA). After 3-4 weeks of incubation at 37°C, CFUs were quantified and the data were expressed as the log₁₀ numbers per target organ.

For flow cytometric analysis, single-cell suspensions of lungs and spleen from each mouse were resuspended in PBS containing 0.1% sodium azide. Fc receptors were blocked with purified anti-mouse CD16/32. The cells were incubated in the dark for 25 minutes at 4°C with predetermined optimal titrations of specific antibodies. Cell surface expression was analyzed with antibodies for CD4 (clone GK1.5), CD11c (clone HL3), CD11b (clone M1/70), PD-1 (CD279) (clone J43), and CTLA-4 (clone 63828; R&D Systems). All antibodies and reagents were purchased from BD Pharmingen (San Diego, CA). The samples were probed using a Becton Dickinson FACSCanto instrument, and the data were analyzed using FlowJo software. Individual cell populations were identified according to the presence of specific fluorescence-labeled antibodies. A total of 300,000 gated events were collected for each sample and gated for each cell phenotype, differentiated by specific antibodies (Henao-Tamayo et al., 2010).

The cells were initially stimulated for 4 hours at 37°C with a 1X cell stimulation cocktail (eBioscience) diluted in complete DMEM. Thereafter, the cells were stained for cell surface markers, fixed, and permeabilized according to the manufacturer’s instructions for the Fix/Perm and Perm wash kit (eBioscience). The cells were then incubated for 30 minutes at 4°C with FcBlock plus anti–IL-10 (clone JES5-16E3; eBioscience), anti–IFNγ (clone XMG1.2; eBioscience), anti-Foxp3 (clones FJK-16s; eBioscience), anti-IL-12 (clone C15.6; BD), or anti-TNF (clone MP6-XT22; eBioscience), or with the respective isotype control. Data acquisition was performed on FACSCanto cytometer (BD) and analysis was performed as described (Bai et al., 2015).

Replicate experiments with duplicate wells for each condition per experiment are independent, and data are presented as means ± SEM or representative experiment. Group means were compared by repeated-measures ANOVA using Fisher’s least significant test or by two-way ANOVA with Bonferroni’s post-hoc test. Data were graphed in Prism 9^® and comparisons were considered significant when p<0.05.

To determine whether CS extract-exposed human Tregs impair macrophage control of MTB, MDM were first infected with MTB for 1 hour, washed to removed free MTB, and then co-cultured with autologous primary human Tregs (at a ratio of 100 monocytes: 1 Treg (Heron et al., 2011)) that were previously incubated in medium alone or with 5% CS extract for 18-24 hours. It is important to emphasize that the Tregs were washed to remove any medium containing CS extract before co-culture with the MDM to ensure that any effect of CS extract was caused by the CS extract-exposed Tregs and not due to any direct effects of CS extract on MDM. After the MTB-infected cell co-cultures were incubated for 1 hour, 2 and 4 days, cell-associated MTB were quantified as previously reported (Bai et al., 2017). As an additional control, MDM without Tregs were also infected with MTB. Compared to MDM alone, co-culture of MDM with control Tregs increased the bacterial burden, which further increased in co-culture with CS extract-exposed Tregs ( Figure 1A ).

An alternative approach to assess bacterial burden was employed in which MDM alone, MDM + unexposed Tregs, and MDM + CS extract-exposed Tregs were infected with GFP-MTB H37Rv for 1 hour, 2 and 4 days and the percentage of GFP-MTB-infected MDM determined by fluorescent microscopy. Compared to MDM alone, co-culture of Tregs with MDM increased the percentage of MDM containing GFP-labeled MTB, which further increased with the addition of Tregs that were pre-incubated with CS extract ( Figure 1B ). CS extract itself did not directly enhance the growth of MTB in the absence of macrophages ( Figure 1C ). We conclude that CS extract-exposed Tregs further enhance proliferation of MTB in macrophages.

An immune evasive mechanism of MTB is the inhibition of phagosome-lysosome (P-L) fusion. To determine whether CS extract-exposed Tregs affect P-L fusion in MTB-infected macrophages, MDM were cultured alone or co-cultured with autologous Tregs previously exposed to medium alone or with 5% CS extract for 18-24 hours; this was followed by infection with GFP-MTB H37Rv for 6 hours and quantitation of P-L fusion. Compared to GFP-MTB-infected MDM alone, co-incubation of MDM with Tregs reduced the co-localization of GFP-MTB with lysosomes, with even further decrease when MDM were co-cultured with CS extract-exposed Tregs ( Figures 1D, E ).

To quantify autophagosome formation, MDM, MDM + unexposed Tregs, and MDM + CS extract-exposed Tregs infected with MTB for 18 hours were immunostained with anti-LC3-II antibody and Cy3-tagged goat anti-rabbit antibody. MDM co-incubated with CS extract-exposed Tregs had significantly fewer LC3-II (+) puncta than MDM alone or MDM + unexposed Tregs ( Figures 2A, B ). These findings indicate that in the presence of CS extract-exposed Tregs, MTB-infected MDM had either decreased autophagosome formation, increased autophagosome-lysosome fusion with degradation of autophagosome content, or both.

To analyze which of these two autophagic processes is more dominant in MDM + CS extract-exposed Tregs, expression of the cargo protein sequestosome-1 (p62) – located within autophagosomes and degraded upon fusion with lysosomes – was temporally quantified. MDM, MDM + unexposed Tregs, and MDM + CS extract-exposed Tregs infected with MTB for 3, 6, and 18 hours were immunostained for p62. With MDM alone, there was a modest decrease in p62(+) autophagosomes with increasing time of infection, suggesting increased autophagosome-lysosome fusion ( Figures 2C, D , open bars). However, in the presence of either unexposed Tregs or CS extract-exposed Tregs, the number of p62(+) autophagosomes did not decrease but there was a modest increase, suggesting a component of blockade in autophagosome-lysosome fusion in the presence of Tregs. Examination of the 18-hour time point shows a modest stepwise increase in p62(+) autophagosomes with MDM, MDM + unexposed Tregs, and MDM + CS extract-exposed Tregs ( Figure 2D , last three bars), indicating decreased autophagosome-lysosome fusion in the presence of Tregs, especially those that were exposed to CS extract.

To further validate these findings, we quantified p62(+) puncta in the MDM, MDM + unexposed Tregs, and MDM + CS extract-exposed Tregs at 18 hours after infection with MTB in the presence of rapamycin, a known inducer of autophagy (Ko et al., 2017). As previously shown, there was a successive increase in p62 puncta with MDM alone, MDM + unexposed Tregs, and MDM + CS extract-exposed Tregs ( Figure 3B , open bars). In contrast, in MDM plus either unexposed or CS extract-exposed Tregs treated with rapamycin, there was no increase in p62(+) puncta compared to cells not treated with rapamycin ( Figure 3B , compare gray bars vs. respective open bars). Because rapamycin did not increase the number of autophagosomes in co-culture of MDM + CS extract-exposed Tregs, it suggests CS extract-exposed Tregs are inhibiting the signaling pathway that leads to autophagosome formation at a site “downstream” to where rapamycin is augmenting autophagosome formation ( Figure 3C ).

To determine whether CS extract-exposed Tregs are inducing autophagosome-lysosome fusion (= autophagosome maturation) to account for the decreased LC3-II(+) autophagosomes, we performed the aforementioned p62 assay in the absence or presence of bafilomycin, which inhibits both lysosome acidification and autophagosome-lysosome fusion (Mauvezin and Neufeld, 2015). We hypothesized that if CS extract-exposed Tregs were inducing autophagosome-lysosome fusion, then in the presence of bafilomycin, we would expect a relative increase in the number of p62(+) puncta because bafilomycin would be inhibiting the degradation of p62(+) autophagosomes. Unexpectedly, addition of bafilomycin decreased the number of p62(+) puncta, especially in MDM + 5% CS extract-exposed Tregs ( Figures 3A, B , last open vs. black bars). However, the size of the p62(+) puncta with bafilomycin treatment were noticeably larger than seen in MDM without bafilomycin treatment, suggesting autophagosome aggregation ( Figure 3A , Supplementary Figure 2A ). Using FIJI software, the p62(+) puncta were measured by area and fluorescence using a Corrected Total Cell Fluorescence equation provided by Luke Hammond from The University of Queensland, Australia (Hammond, 2014). These analyses showed that the total area fluorescence of the p62(+) puncta were not significantly different with or without bafilomycin ( Supplementary Figures 2B, C ). These findings support the paradigm that co-incubation of MDM with CS-exposed Tregs reduces autophagosome number by inhibiting their formation (as noted by decreased LC3-II autophagosomes) but also a component of decreased autophagosome maturation (as noted by increased p62(+) marker to the autophagosomes).

The supernatant of the MDM, MDM + unexposed Tregs, and MDM + CS extract-exposed Tregs used to obtain the aforementioned MTB CFU were saved and assayed for tumor necrosis factor (TNF), interleukin-10 (IL-10), and transforming growth factor-beta (TGFβ). Compared to MDM alone, co-culture with unexposed Tregs modestly decreased TNF and increased IL-10 and TGFβ at 2 and 4 days after culture ( Figures 4A-C ). Co-culture of MDM with Tregs exposed to CS extract further decreased TNF level and increased levels of IL-10 and TGFβ ( Figures 4A-C ).

In human lungs, CS exposure promotes Treg function (Ménard and Rola-Pleszczynski, 1987; Barceló et al., 2008; Wang et al., 2010). A plausible mechanism by which this occurs is the ability of CS to induce PD-L1/2 on antigen presenting cells, resulting in increased engagement to PD-1 on Tregs (Francisco et al., 2010; Yanagita et al., 2012); i.e., whereas engagement of PD-1 and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) on T cells to PD-L1/2 and B7 on macrophages, respectively, inhibits the activity of Foxp3-negative T effector cells, these same molecular interactions enhance the suppressive activity of Tregs (Saresella et al., 2008; Chan et al., 2014). Thus, to determine whether PD-1 and/or CTLA-4 plays a mechanistic role by which CS extract-exposed Tregs impair macrophage control of MTB infection, we first exposed Tregs to CS extract for 18 hours and quantified PD-1 and CTLA-4 expression. CS extract increased the expression of CTLA-4 but not PD-1 in primary human Tregs ( Figure 5A ).

To determine the role CS extract-induced CTLA-4 on Tregs may play in impairing MDM control of MTB, we incubated CS extract-exposed Tregs with 5 µg/mL non-immune IgG or with 5 µg/mL anti-CTLA-4 antibody before combining the Tregs with MTB-infected MDM (Machicote et al., 2018). As shown in Figure 5B , macrophages co-cultured with CS extract-exposed Tregs have greater MTB burden than MDM alone or MDM + unexposed Tregs. However, in the presence of an anti-CTLA-4 neutralizing antibody, there was a partial but significant abrogation of the increase in MTB burden in co-culture of MDM + CS extract-exposed Tregs; such abrogation was not seen with non-immune IgG. These findings indicate that CS extract induction of CTLA-4 on Tregs is a mechanism by which CS extract-exposed Tregs impair macrophage control of MTB infection.

The supernatant of the MDM, MDM + unexposed Tregs, MDM + CS extract-exposed Tregs, and MDM + CS extract-exposed Tregs incubated with IgG or anti-CTLA-4 (and all infected with MTB) were assayed for TNF, IL-10, and TGFβ. As previously seen, compared to MDM alone, co-culture with unexposed Tregs modestly decreased TNF and increased IL-10 and TGFβ at 2 and 4 days after culture ( Figures 5C-E ). Co-culture of MDM with Tregs exposed to CS extract further decreased TNF level and increased IL-10 and TGFβ. Compared to MDM + CS extract-exposed Tregs, neutralization of CTLA-4 significantly abrogated the reduction in TNF at Day 2 after MTB infection ( Figure 5C ) and abrogated the increase in IL-10 at Day 2 and TGFβ at Day 4 after MTB infection ( Figures 5D, E ).

Tregs in the Foxp3⁺GFP⁺DTR⁺ (Thy1.2) mice were depleted by administration of four doses of diphtheria toxin (DT) as we previously reported, wherein ~6-fold depletion of Tregs were achieved in the lungs and >10-fold depletion of Tregs in the spleen (Henao-Tamayo et al., 2016). In this study, we used higher amounts of DT for the last two doses. Following the adoptive transfer of Tregs from air-exposed and CS-exposed B6.PL(Thy1.1) donor mice into Treg depleted Foxp3⁺GFP⁺DTR⁺ (Thy1.2) mice, the latter recipient mice were infected with HN878-W-Beijing MTB for 1, 30, and 60 days. Because these mouse strains possess an immunocompetent phenotype, a hypervirulent strain of MTB was chosen to achieve a productive infection. Foxp3⁺GFP⁺DTR⁺ (Thy1.2) mice in which no Treg depletion or transfer was performed were also infected with HN878 MTB as an additional control. The lungs and spleen were harvested, processed, and the CFU of the organs were determined at the indicated times. Compared to control mice and mice that received Tregs from air-exposed mice, those that received Tregs from CS-exposed mice had significantly greater CFU in the lungs and spleens at Day 60 after infection ( Figures 6A, B ).

Histopathology of the lungs was analyzed at 1, 30 and 60 days after HN878 MTB infection. Compared to control mice and mice recipient of Tregs from air-exposed donors, mice that received Tregs from CS-exposed donors demonstrated increased size and number of granuloma-like lesions, increased granulocyte infiltration, and the presence of acid-fast bacilli, particularly in the lesions ( Figures 6C–E , arrows).

Lung macrophages and dendritic cells (DC) of uninfected mice, control MTB-infected mice, and Foxp3⁺GFP⁺DTR⁺ (Thy1.2) mice that received air- or CS-exposed Tregs followed by MTB infection were immunophenotyped for intracellular TNF, IL-12, and IL-10 at 1, 30 and 60 days after MTB infection. Compared to no infection, MTB infection increased the number of TNF⁺ lung macrophages and DC at 30 and 60 days after infection ( Figures 7A, B ). At 60 days after infection, there was decreased number of TNF⁺ macrophages in mice transferred Tregs from CS-exposed mice although this did not reach statistical significance ( Figure 7A ). At 60 days post infection, there was also a trend toward decreased TNF⁺ DC in mice transferred Tregs from either air- or CS-exposed mice ( Figure 7B ).

Compared to uninfected mice, MTB infection alone increased the number of IL-12⁺ lung macrophages, which further increased with the transfer of either air- or CS-exposed Tregs at 30 days post-infection ( Figure 7C ). Interestingly, at 60 days post-infection, IL-12⁺ lung macrophages decreased to levels even lower than at 1 day after infection. In contrast, IL-12⁺ DC progressively increased temporally with all three MTB-infected conditions ( Figure 7D ); in mice that received Tregs from CS-exposed mice, there was a significantly lower number of IL-12⁺ DC compared to uninfected mice, mice only infected with MTB, or mice that received Tregs from air-exposed mice and then infected with MTB ( Figure 7D ).

There was a temporal increase of IL-10⁺ lung macrophages and DC with MTB infection (except for DC at Day 30 after infection) ( Figures 7E, F ). Mice that received Tregs from either air- or CS-exposed mice have decreased IL-10⁺ macrophages at Day 60 compared to MTB-infected mice without Treg transfer ( Figure 7E ). In contrast, mice that received either air- or CS-exposed Tregs and MTB-infected have marked infiltration with IL-10⁺ DC at Day 30 compared to mice infected with MTB only, but this difference was lost at Day 60 ( Figure 7F ).

While CTLA-4 is canonically expressed on T cells, it is also present on human monocytes and mature monocyte-derived DC following stimulation with lipopolysaccharide, poly:IC, and various cytokines; CTLA-4 on DC also inhibits CD4⁺ T cell proliferation via CTLA-4-mediated production of IL-10 by the DC (Laurent et al., 2010). Thus, we quantified CTLA-4 expression in the mouse lungs and found that with MTB infection, there was a temporal increase in CTLA-4⁺ macrophages with MTB infection with or without Treg transfer ( Figure 8A ). However, in the mice that received air-exposed Tregs, there was a lower amount of CTLA-4⁺ macrophages at Day 60, but in the mice that received Tregs from CS-exposed mice, there was higher number of CTLA-4⁺ macrophages ( Figure 8A ). There was also a temporal increase of CTLA-4⁺ DC with all infections – except for Day 30 with the only MTB-infected mice ( Figure 8B ). In contrast to that seen with lung macrophages, there was a significant decrease of CTLA-4⁺ DC in the mice that received Tregs from CS-exposed mice at Day 60 compared to mice with MTB infection alone. We conclude that in mice that received Tregs from CS-exposed mice, there was greater decrease in the number of TNF⁺ macrophages, IL-12⁺ DC, and CTLA-4⁺ DC in the lungs compared to that seen in control mice that either received no Tregs or air-exposed Tregs.

The general gating strategy for T cells is shown in Figure 9A . Examining the phenotypic subsets of T cells revealed a temporal increase in CD4⁺IFNγ⁺ T cells in the lungs of all three mouse groups infected with MTB compared to uninfected mice ( Figure 9B ). But at Day 60, the number of CD4⁺IFNγ⁺ T cells in mice that received Tregs from CS-exposed mice and infected with MTB was less than all other mouse groups but these differences were not statistically significant ( Figure 9B , last bar vs. preceding two bars at Day 60).

Compared to uninfected mice, the number of CD4⁺IL-1⁺ T cells increased temporally in the other mouse groups infected with MTB ( Figure 9C ). While the number of CD4⁺IL-1⁺ T cells were similar at Days 1 and 30 between all three groups of MTB-infected mice, by Day 60 the mice that received air-exposed Tregs had the least number of IL-1⁺CD4⁺ T cells ( Figure 9C ). In examining an exhaustion marker of CD4⁺ T cells, mice that received Tregs from CS-exposed, MTB-infected mice had significantly more PD-1-positive T cells at Day 60 ( Figure 9D , last bar vs. preceding two bars at Day 60). However, there was no difference in the expression of either PD-1 or CTLA-4 donor Thy1.1⁺Tregs from unexposed or CS-exposed mice in the lungs of the recipient mice at 1, 30, or 60 days after infection (data not shown).

Quantifying the total number of CD4⁺ Tregs in the four groups of mice (uninfected, MTB-infected alone, and MTB-infected mice recipient of Tregs from air- or CS-exposed mice), we found that compared to uninfected mice, MTB-infected mice had significantly more total Tregs, which peaked at Day 30 after infection ( Supplementary Figure 3A ). While there was a lower number of total CD4⁺ Tregs at Day 30 after infection in the mice that received Tregs from CS-exposed mice, this was not statistically significant. Similarly, there was a trend of increased number of CD4⁺IL-10⁺ Tregs at Day 60 after MTB infection in the mice that received Tregs from either air- or CS-exposed mice ( Supplementary Figure 3B ). Since a small population of Tregs can produce IFNγ (Zhao et al., 2011; Koenecke et al., 2012; Sumida et al., 2018), we analyzed the population of transferred Thy1.1⁺CD4⁺IFNγ⁺ Tregs from the air- or CS-exposed Thy1.1 mice in the lungs of the recipient Thy1.2 mice following MTB infection. While there was no difference in the population of these Tregs in the lungs at Day 1 or 30, by Day 60 the recipient Thy1.2 mice had fewer donor Thy1.1⁺IFNγ⁺ Tregs from the CS-exposed mice compared to that from the air-exposed mice ( Figure 9E ). Furthermore, the temporal increase in the donor CD4⁺IFNγ⁺ Tregs indicates replication of these donor Tregs in the recipient mice over the course of the infection as previously reported (Henao-Tamayo et al., 2016).

Splenic macrophages and DC were also analyzed 2, 30, and 60 days after MTB infection in the three mouse groups. Uninfected mice also served as an additional control. Compared to uninfected mice, the three MTB-infected mouse groups demonstrated an increase in intracellular TNF⁺, IL-12⁺, and IL-10⁺ macrophages ( Supplementary Figures 4A-C ) and DC ( Supplementary Figure 4D-F ). With the time points examined, the number of such macrophage and DC phenotypes were highest at 60 days after MTB infection except for the IL-12⁺ macrophages, which peaked at Day 30 ( Supplementary Figure 4B ). Similarly, the three mouse groups infected with MTB had similar numbers of CTLA-4⁺ macrophages ( Supplementary Figure 4G ) and dendritic cells ( Supplementary Figure 4H ), with both cell type numbers highest at 60 days after infection.

Analysis of the total number of splenic CD4⁺ T cell subsets revealed that, compared to uninfected mice, the number of T cells significantly increased at 30 and 60 days in the three mouse groups that were infected with MTB ( Supplementary Figure 5 ). More specifically, the total number of IFNγ⁺ T cells increased progressively from Day 2 to Day 60 after infection in control mice that were only infected with MTB and in MTB-infected mice that received Tregs from air-exposed mice. In contrast, the temporal increase in the number of IFNγ⁺ T cells was attenuated in the animals that received Tregs from CS-exposed mice ( Supplementary Figure 5A , last bar vs. preceding two bars at Day 60). Similarly, compared to uninfected mice, there was an increase in the number of IL-1⁺ T cells in the three mouse groups infected with MTB ( Supplementary Figure 5B ). The number of such IL-1⁺ T cells peaked at Day 30 in the mice infected with MTB alone and mice that received Tregs from air-exposed mice whereas they peaked at Day 60 in the mice that received Tregs from CS-exposed mice and subsequently infected with MTB ( Supplementary Figure 5B ). In examining the CD4⁺ T cell subset expressing the PD-1 exhaustion marker, there was a progressive temporal increase in all three groups of mice infected with MTB ( Supplementary Figure 5C ). However, there was significantly greater number of PD-1⁺ splenic T cells in the MTB-infected mice that received Tregs from CS-exposed mice ( Supplementary Figure 5C ).

In analyzing the splenic Tregs, there was an increased in the total number at Days 30 and 60 in the three mouse groups infected with MTB compared to uninfected mice ( Supplementary Figure 5D ). Compared to MTB infection alone, those that received air-exposed mice had a robust rise in Tregs at Day 30, which decreased to near basal levels at Day 60. In contrast, there was a progressive increase in Tregs in mice that received Tregs from CS-exposed mice. The number of IL-10⁺ Tregs temporally quantified in the spleen followed a very similar pattern to that observed with the total number of Tregs ( Supplementary Figure 5E ).