Mucosal exposure to non-tuberculous mycobacteria elicits B-cell mediated immunity against pulmonary tuberculosis

Summary

Bacille Calmette Guerin (BCG) is the only licensed vaccine against Mycobacterium tuberculosis (Mtb), the causative agent of tuberculosis (TB) disease. However, BCG has limited efficacy, necessitating development of better vaccines. Non-tuberculous mycobacteria (NTM), ubiquitously present in the environment, can be opportunistic pathogens. Individuals in TB endemic countries experience higher exposure to NTM, but previous studies have not elucidated the relationship between NTM exposure and BCG efficacy against TB. Therefore, we developed a mouse model (BCG+NTM) to simulate human BCG immunization regime and continuous NTM exposure. BCG+NTM mice exhibit superior and prolonged protection against pulmonary TB, with increase in B-cell influx and anti-Mtb antibodies in serum and airways, compared to BCG alone. Notably, spatial transcriptomics and immunohistochemistry reveal that BCG+NTM mice form B-cell aggregates with features of germinal center development, which correlates with reduced Mtb burden. Our studies suggest a direct relationship between NTM exposure and TB protection, with B-cells playing a crucial role.

Introduction

Non-tuberculous mycobacteria (NTM) comprise a diverse group of mycobacterial species that do not cause tuberculosis (TB), but they can cause lung disease in individuals with local or systemic immune suppression^1,2. These opportunistic pathogens^3,4 are ubiquitous in the environment. Hence, humans are continuously exposed to NTM, most likely starting early in life, via mucosal routes such as the gastrointestinal tract (water/food), skin (wounds), or airways (inhalation). High exposure to NTM occurs in some regions of the world due to a combination of environmental, sanitary, and socio-economic reasons. Notably, TB endemicity overlaps areas of the world with high NTM exposure. NTM share a large number of antigens with Mtb and BCG; thus, it has been postulated that NTM modulate vaccine-mediated protection and immunity against TB ^5-7. However, the effect of pre-exposure to NTM on subsequent Mtb infection is still debatable, although highly studied.

TB is a severe public health problem, affecting a quarter of the world's population⁸. Approximately 1.5 million people die of TB annually, including HIV-associated deaths. Despite concerted efforts, it continues to spread worldwide ^9-11. TB is a treatable disease requiring six to nine months of multi-antibiotic treatment¹². However, long-term therapy may lead to drug-induced toxicity, human microbiota dysbiosis, and patient non-compliance, resulting in other secondary infections, a weakened immune system, and the emergence of antibiotic-resistant TB, respectively ^13-16. Current efforts to fight TB by improving diagnosis and developing better anti-TB vaccines and drugs are falling short ^17-19. Furthermore, the COVID-19 pandemic crippled the World Health Organization's TB guidelines, such as early patient diagnosis, treatment follow-up, and compliance ^20,21, erasing much of the last decade’s progress against TB and other infectious diseases. As a result, TB continues to cause immense human misery and considerable economic loss, and it is one of the leading causes of global inequity.

BCG is the only licensed vaccine against TB²². Unfortunately, its efficacy varies greatly depending on the country and demographics^23,24. BCG’s variable protection against TB has been attributed, amongst others, to genetic divergence of BCG strains used for vaccination^25,26, high virulence of clinical Mtb strains^19,27,28, host genetics^29,30, and interference by NTM^7,31,32. Data from animal models exploring the influence of NTM on Mtb pathogenesis and BCG vaccine responses vary between studies. While early research demonstrated that NTM infection (without BCG vaccination) protects animals against Mtb infection^33-35, others reported that mice pre-exposed to NTM with subsequent BCG vaccination appear to mask the overall protection provided by BCG vaccination^6,36. Such discrepancies could be attributed to a multitude of variables in experimental designs, including NTM concentration, duration, and route of NTM exposure, as well as NTM strain. Considering the preceding and to elucidate if and how NTM impacts BCG efficacy against TB, we developed a murine model closely simulating the natural history of human exposure to mycobacteria, including: 1) early BCG vaccination; 2) exposure to viable NTM via drinking water [1000 CFU/mL for high-income countries (HIC), 1.0x10⁵ CFU/mL for low to middle-income countries (LMIC)]; and 3) maintaining continuous NTM exposure even throughout Mtb infection.

Compared to BCG-only vaccinated mice, a high dose of NTM delivered via drinking water to BCG-vaccinated mice (BCG+NTM^high) provided more robust and long-lasting protection against Mtb infection 120 days post-infection (dpi), with lower Mtb burden and reduced pulmonary inflammation. Immunological studies showed significantly higher numbers of B220^hiMHC-II^hi B-cells and CX3CR1^hiMHC-II^hi macrophages in Peyer’s patches (PPs) of NTM exposed mice, regardless of BCG vaccination status before Mtb infection. In the lungs, BCG+NTM^high mice showed a significant increase in B-cells, CXCR5⁺ T-follicular helper cells (T_FH), and Perforin⁺Granzyme-B⁺ natural killer (NK) cells post-infection, strongly correlating with reduced Mtb burden. Additionally, mice exposed to high doses of NTM, irrespective of BCG vaccination status, had considerably higher titers of anti-mycobacterial IgG and IgA antibodies in serum and bronchoalveolar lavage fluid (BALF), both pre-and post-Mtb infection. Surprisingly, BCG+NTM^high mice also developed follicle-like structures in the lungs. Immunohistochemistry and spatial transcriptomics confirmed B-cell predominance in those structures, along with markers associated with B-cell activation, somatic hypermutation, and germinal center (GC) formation. These data suggest that NTM provide protective immunity against pulmonary TB and that B-cells and GC are crucial in this process.

Results

Oral NTM exposure boosts BCG-mediated protection against pulmonary tuberculosis

Initially, to establish a mouse model of NTM exposure (BCG+NTM mice) representative of the natural history of human exposure to mycobacteria, three-week-old C57BL/6, C3HeB/FeJ, and C3H/HeOuJ mice were vaccinated with BCG (1x10⁵ CFU/mouse). Fifteen days later and throughout the remainder of the experiment, mice were continuously exposed via the drinking water to 1000 CFU/ml Mycobacterium avium subsp. avium (MA), the most ubiquitous species of NTM³⁷]. Water containing NTM was replaced weekly, and bacterial concentration and viability in drinking water were monitored throughout the study (Figure S1). Ninety days after initiating NTM exposure (and 104 days post-BCG vaccination), mice were infected via the aerosol route with Mtb HN878 (50-60 CFU/mouse). Mtb HN878 is a hypervirulent W-Beijing clinical strain with high infectivity^38,39, reported to induce a granulomatous response in mice resembling the pathology observed in TB-infected human and non-human primates⁴⁰. Control mice were immunized with BCG or saline or fed NTM alone (Figure 1A).

Figure 1: Murine model of NTM exposure via drinking water and bacterial burden.

(A) Study-1: Experimental plan showing a timeline of BCG vaccination, NTM exposure, and Mtb infection using C3H/OuJ, C3HeB/FeJ, and C57BL/6 mice. Colony-forming units (CFUs) showing Mtb burden in the lungs of C3H/HeOuJ (B), C3HeB/FeJ (C), and C57BL/6 (D) mice at 30-, and 60 dpi). (E) Study-2: The experimental plan shows a timeline for BCG vaccination, NTM exposure to low (1000 CFU/mL) and high (1.0 x 10⁵ CFU/mL) doses, and Mtb infection in C3H/HeOuJ mice. (F) Plots showing Mtb CFUs in the lungs of C3H/HeOuJ mice at 30-, 60-, and 120 dpi. (G) Pathology scores represent the proportion of lesion area over total lung area per stained tissue section. (H) Lung sections were stained with hematoxylin and eosin at 120 dpi. Biological replicates, n= 5, technical replicates, n=2. Statistical significance was calculated using ANOVA with the Tukey-HSD test, and p<0.05 was considered significant. *p<0.05, **p<0.005.

To determine the effect of NTM on BCG's efficacy against Mtb, the bacterial burden in the lungs was enumerated at 30 and 60 dpi. As expected, at 30 dpi, BCG vaccination protected all strains of mice against Mtb infection, as evidenced by a one-log reduction in lung CFUs. Interestingly, NTM exposure did not abrogate BCG-mediated protection against Mtb. BCG's protective efficacy, however, began to wane at 60 dpi in vaccinated mice, as previously reported for BCG-vaccinated mice infected with W-Beijing strains^41,42. Strikingly, protection was preserved in BCG-vaccinated mice (C3H/HeOuJ and C57BL/6) continuously exposed to NTM via the oral route (Figure 1B-D).

To further investigate the long-term effect of NTM exposure and the impact of different NTM concentrations on BCG-mediated protection against TB (Figure 1E), C3H/HeOuJ mice (intermediate susceptibility to Mtb) were fed water containing 1000 (NTM^low) or 1x10⁵ CFU/mL (NTM^high) NTM, reported to be present in drinking water of HIC and LMIC countries^43-45, respectively. Consistent with the findings above, BCG+NTM^low mice had lower bacterial burden in the lungs at 30 and 60 dpi with Mtb HN878 (Figure 1F). Significantly, protection against TB at 120 dpi was only observed for BCG+NTM^high. Furthermore, a significant reduction in lung pathology was observed in BCG+NTM^high mice up to 120 dpi (Figures 1G and 1H). Improved histopathology confirmed that exposure to viable NTM via drinking water did not compromise the protective efficacy of BCG against Mtb infection but instead provided a boost effect.

NTM colonize Peyer’s patches and activate B-cells

Given that mice were administered NTM orally, we assessed their presence and effect on the cellular composition of PPs, the main lymphoid structure associated with the gut. Using an anti-lipoarabinomannan (LAM) antibody ⁴⁶, NTM were indeed shown to colonize PPs (Figure 2D), specifically the PPs interfollicular region (IFR) and follicles (high magnification insert, Figure 2D). The cellular composition of PPs was evaluated by flow cytometry (Table S1). One day before Mtb infection (−1 dpi), significantly increased numbers of B-cells (CD3⁻B220⁺MHC-II⁻ and CD3⁻B220⁺MHC-II⁺), phagocytic cells (CD3⁻CX3CR1⁺MHC-II⁺), and CD4⁺ T cells (CD3⁺CD4⁺) were observed in PPs of BCG+NTM^high mice (Figure 2A and 2B). Immune cell populations in PPs were also evaluated at 30, 60, and 120 dpi after Mtb infection (and continuously exposed to NTM) and correlated to the corresponding lung CFUs (Figure 2C). Mice exposed to high NTM concentrations, irrespective of their BCG vaccination status, showed increased B-cells in PPs, significantly (p<0.01) correlating to reduced Mtb burden in the lungs. Therefore, NTM colonization of PPs could prime the adaptive immune response against mycobacterial antigens.

Figure 2: Peyer's patches (PPs) immune profiling and immunohistochemistry.

(A) Flow cytometry dot plots showing increased B-cells in BCG+NTM^high mice. The bivariate plot here represents all mice concatenated together (n=5). (B) PPs cells were stained with B-cells, T-cells, macrophages, and dendritic cell markers and analyzed using flow cytometry. Data here represent total B-cell counts, CX3CR1⁺ macrophages, and T-cells on day-1 of Mtb infection. (C) Immune cell types present in PPs at days 30, 60, and 90 post-Mtb infection (heatmap) and the corresponding CFUs in the lung (barplot). (D) PPs sections were stained using the OPAL multiplex IHC (mIHC) staining technology and the Leica Bond III autostainer. The panel of antibodies used was anti-B220 (opal 570) antibody, anti-LAM antibody (opal 690), and GL-7 (opal 520) for B-cell, NTM, and GC visualization and location, respectively. Images were acquired using the multispectral camera of the Polaris^™ System (Akoya bioscience) and analyzed using the inForm tissue Finder and phenochart^™softwares. Biological replicates, n= 5, technical replicates, n=3. Statistical significance was calculated using ANOVA with the Tukey-HSD test, and p<0.05 was considered significant. *p<0.05, **p<0.005.

Oral NTM exposure augments pulmonary B-cell responses and production of cross-reactive anti-mycobacterial IgG and IgA antibodies before Mtb infection

Exposure of BCG-vaccinated mice to NTM via the oral route significantly protected against pulmonary TB (reduced lung CFUs and pathology); thus, we investigated whether NTM induced memory immune responses that cross-react and protect against Mtb. Flow cytometric analysis of the lungs was performed to evaluate the immune response elicited by NTM before Mtb infection (antibody panels, Tables S2 and S3). Our findings indicate that NTM^high and BCG+NTM^high mice had significantly higher numbers of both T-independent (B1) and T-dependent (B2) B-cells in their lungs (Figure 3A). Interestingly, mice exposed to NTM^high without BCG vaccination also had increased pulmonary influx of B1 memory (CD19⁺B220⁻CD44⁺CD62L⁻MHC-II⁺CD34⁺CD73⁺), B2 memory (CD19⁺B220⁺MHC-II⁺CD4⁺CD62L⁻CD73⁺), and progenitor B-cells (CD19⁺B220⁻CD44⁺CD62L⁻MHC-II⁺CD34⁺), corroborating that NTM generates memory B-cell responses. Additionally, we observed a significant increase in T_FH in the lungs of NTM^high and BCG+NTM^high mice (Figure 3B).

Figure 3: Immune response in lungs on day-1 before Mtb infection.

Graphs showing total cell numbers of each cell type present in lungs on day-1 before infection (A and B) Indirect ELISA targeting IgA and IgG antibodies against Mtb whole cell lysate (C). Biological replicates, n= 5, technical replicates, n=3. Statistical significance was calculated using ANOVA with the Tukey-HSD test, and p<0.05 was considered significant. *p<0.05, **p<0.005, ***p<0.0005.

Induction of cross-reactive anti-mycobacterial antibodies in mice orally exposed to NTM —yet not infected with Mtb— was evaluated in serum and BALF via an indirect enzyme-linked immunosorbent assay (ELISA) against Mtb HN878 whole cell lysate. A significantly higher titer of anti-mycobacterial IgA and IgG antibodies in BALF and serum was observed in mice exposed to a high NTM concentration regardless of BCG vaccination status than in those exposed to NTM^low or unexposed (Figure 3C). When combined with flow cytometry results presented above, these findings suggest that NTM exposure results in the generation of memory B-cells and cross-reactive antibodies against Mtb antigens.

Increased B-cells and anti-Mtb antibodies in lungs correlate with enhanced protection upon TB infection

To further investigate the possible immune mechanism(s) of protection, flow cytometry analysis was conducted on leukocytes obtained from lungs of Mtb infected mice at 30, 60, and 120 dpi (Figure 4A, Table S2, and S3). BCG-vaccinated mice not exposed to NTM had higher total T cell counts at 30 dpi, which declined significantly at 60 and 120 dpi, perhaps related to loss of protection at later time points. In contrast, NTM^high mice had significantly higher numbers of B cells (B220⁺/CD19⁺ cells), effector CD8 T cells (CD3⁺CD8⁺CD44⁺Granzyme-B⁺Perforin⁺), and NK cells (CD3⁻NKp46⁺Perforin⁺Granzyme-B⁺). A correlation analysis between lung CFUs and various immune cell populations was performed to evaluate the cell types associated with reduced bacterial burden in the lungs (Figure 4B and Figure S2). We found that at 30 dpi, the primary immune cells associated with protection (R-value > −0.5, and p-value < 0.05) were B1 and B2 B-cells, memory B-cells, effector NK cells, T_FH, and effector T cells. In contrast, at 120 dpi, NK cells and different B-cell subpopulations (B2, memory, and MHC-II⁺ B-cells) were significantly correlated with reduced bacterial burden (rather than B1 B-cells and effector T cells). Total B-cells also significantly correlated (p = 0.0037, R = −0.52) with lower lung CFUs at 30-, 60-, and 120-days post-infection (Figure 4C). Our results point to a crucial association between B-cells and NK cells and decreased Mtb burden in the lungs of BCG+NTM^high mice.

Figure 4: Comprehensive analysis of innate and adaptive immune cell populations at days 30-, 60-, and 120-days post Mtb infection.

(A) Immune cell populations in the lungs were stained using flow cytometry and analyzed using Cytotypr. Data here show the total cell numbers of each population converted to z-score and represented as a heatmap. (B) Spearman correlation analysis of various immune cell types with the bacterial burden (CFUs) in the lung at 30-, 60-, and 120-dpi. Significance was calculated using the Wilcoxon test with Benjamini-Hochberg correction. (C) Correlation analysis between total B-cells at all time points and lung CFUs using the linear regression model. R-value represents regression value. Statistical significance is calculated using a t-test. (D) Indirect ELISA showing the presence of anti-Mtb IgG and IgA antibodies in the bronchoalveolar lavage fluid (BALF) and serum of mice at 30, 60, and 120 days post-Mtb infection. Values represent the area under the curve (AUC) calculated using three different dilutions (1/50, 1/250, and 1/1250). Biological replicates: n=5, technical replicates for flow cytometry and CFU: n=2; technical replicates for ELISA: n=3. All graphs were created using ggplot2 in RStudio, and significance was calculated using R package stats. *p<0.05, **p<0.005, ***p<0.0005.

The presence of anti-Mtb antibodies in BALF and serum was evaluated by ELISA (Figure 4D), and results were expressed as the area under the curve (AUC). At day 120 post-Mtb infection, we found significantly higher IgA and IgG levels in the BALF of BCG+NTM^low and BCG+NTM^high mice than in all other groups. However, no significant differences were observed in serum IgA or IgG levels. These findings suggest that NTM confers enhanced protection via B-cell and antibody-mediated mechanisms and that B-cell activation and antibody generation occur within the lung.

Mice exposed to NTM develop B-cell aggregates in the lungs

Immunohistochemistry results demonstrated the presence of pulmonary B cell aggregates (B220⁺ cells forming clusters), which were very clear at 120 dpi, predominantly in BCG+NTM^high mice (Figure 5A). Notoriously, B cell aggregates (shown in yellow) localized around the lesion area and were surrounded by Mtb (shown in red). The number of B cell aggregates in the lungs of BCG+NTM^high mice (Figure 5B) inversely correlated with lower Mtb burden (R=−0.72, p <0.001, Figure 5C).

Figure 5: Immunohistochemistry showing the presence of B-cell aggregates in the lungs.

Five μm lung sections from saline, NTM^low, NTM^high, BCG, BCG+NTM^low, and BCG+NTM^high were stained using the OPAL multiplex IHC (mIHC) staining technology using the LabSat^® Research autostainer (Lunaphore platform; Epedria). Slides were scanned using the multispectral camera of the Polaris^™ System (Akoya bioscience) and analyzed with the inForm tissue Finder and phenoptr^™ softwares. The panel of antibodies used included anti-LAM antibody (opal 690), anti-mouse B220 antibody (opal 570), and anti-mouse CD4 antibody (opal 520) (A) B-cell aggregates were calculated in each mouse's lung lobe by drawing a boundary around each B-cell aggregate, teaching the software to count the number of follicles (B) Correlation analysis was performed between CFUs and B-cell aggregates in the lung using the Spearman rank correlation method (C). Biological replicates: n=5, technical replicates: n=3. Statistical significance was calculated using R package stats. *p<0.05, **p<0.005.

To further corroborate our histopathological findings and determine the cell types and gene signatures associated with these structures, we performed spatial transcriptomics using 10X Genomics Visium kit for FFPE tissues. Lung histology images and cell-type distribution of one mouse from each group are shown in Figure 6A (the remaining images are shown in Figure S3). Spatial transcriptomics analysis revealed that NTM^high and BCG+NTM^high mice had a significantly higher frequency of pulmonary B cells and dendritic cells than other groups (Figures 6B and C). In contrast, higher macrophage but lower B cell influx was observed in the lungs of saline mice and mice only vaccinated with BCG.

Figure 6: Spatial transcriptome profiling of lung sections at 120-day post-Mtb infection.

Spatial transcriptomics was performed using 10X visium kit for Formalin-fixed paraffin-embedded tissues on eight lung sections (two sections from each group from two separate mice). (A) H&E-stained lung sections from each group and corresponding cell types at each spot (second panel). (B) The cell type ratio in saline, NTM^high, BCG, and BCG+NTM^high mice. (C) Principal component analysis and representative UMAP embedding of single-cell data. (D) SpatialFeaturePlot showing marker expression of top variable features in different groups. Expression ranges from blue to red (lowest to highest expression, respectively).

Notably, mice exposed to NTM exhibited an increased stromal (CD45⁻Lin⁻Sca-1⁺CD29⁺CD105⁺) cell density in the lungs. Thus, we evaluated markers associated with follicle structures in the lungs of NTM-exposed mice and compared them to the saline and BCG groups (Figure 6D). The follicle structures in the lungs of mice exposed to NTM expressed Cd19 (a B-cell marker), Bcl-6 (a GC B-cell marker), Ltb (lymphotoxin-B, a marker associated with the formation of lymphoid tissue), Mki67 (a proliferation marker), and Cxcr5 (a marker of follicular B and T cells).

Gene ontology (GO) and gene set enrichment (GSE) analysis were performed with clusterProfiler to identify up or down-regulated pathways in each group. The top 25 biological pathways activated in saline, NTM^high, BCG, and BCG+NTM^high mice are depicted in Figure S4. The GSE analysis showed that a significant number of genes (p<0.05) activated in saline mice are involved in response to bacterial lipoproteins, removal of superoxide radicals, nitric oxide production, granulocyte migration, and chronic inflammatory response (Figure S4-A). Additionally, the top pathways activated in the BCG group revealed an inflammatory progression of infection, including interferon-alpha and beta responses and wound healing (Figure S4-B). However, in NTM mice, there was a significant upregulation (p<0.05) of pathways associated with humoral immune responses, B-cell activation, IgA, and complement activation (Figure S4-C). Interestingly, the top 25 pathways (p<0.05, gene involvement < 30) activated in BCG+NTM^high mice are involved in GC formation, B-cell activation and differentiation, somatic recombination, isotype-switching, immunoglobulin complex formation, IgA activation, and NK cell differentiation (Figure S4-D). Collectively, our findings suggest the formation of a lymphoid-like structure containing GC in the lungs of BCG+NTM^high mice, which may be involved in B-cell mediated immunity.

Transcriptional profiling revealed activation of memory B-cells and LT-LTBR signaling in BCG+NTM^high mice.

Using DESeq2, we investigated differentially expressed genes (DEGs) in each cluster and identified specific signatures for some groups. For instance, in TB-infected saline control mice, an inflammatory response characterized by a significant upregulation of Nos2, Thbs1, Cxcl10, and Tgfbi was observed (Figure 7A). Conversely, Cd38 activation (a marker for switched memory B-cells) was only observed in B-cells from BCG+NTM^high mice, a further indication of GC formation in the lungs of these mice (Figure 7B). We also found significantly (p<0.05) increased Ccr6 expression in macrophages, epithelial cells, and fibroblasts in BCG+NTM^high mice. It is worth noting that MKi67, a major cell proliferation marker, was highly upregulated in neutrophils and macrophages from the saline group. In contrast, its expression in BCG+NTM^high mice was significantly higher (p<0.01) in B-cells and dendritic cells.

Figure 7: Transcriptomic profile and cell-cell interaction among groups.

Differential gene expression of each cell type in each group was evaluated using the FindMarkers function in Seurat. (A) Selected markers highly differential between saline and NTM^high, BCG, and BCG+NTM^high groups are represented as Feature plots. Scale varies from grey to purple (least to high expression). (B) Violin plots showing marker expression of Cd38, Ccr6, Mki67, and Tnfrsf13c. Markers were selected based on significantly higher expression (p<0.01) between BCG+NTM^high groups compared to all. (C) Cell-cell interaction analysis using CellChat. Circle plots show the number of interactions between each cell type in an individual group. (D) Barplot indicates the number of total interactions between all cell types and their interaction strength. Values were calculated using the "compareInteractions" function in CellChat. (E) Heatmap showing signaling pathways upregulated in different groups. Red: activated; blue: not activated. (F) Dot plot shows the receptor-ligand interactions of the specific pathways activated in saline, NTM^high, BCG, and BCG+NTM^high and the cell types involved with those pathways. The circle's color represents communication probability (red: highest; blue: lowest), and the size of the circle represent the significance level. p<0.05 was considered significant.

Cellular interactions were interrogated using CellChat⁴⁷ (Figure 7). Figure 7C depicts significant networks of cellular interactions between all cell types within each group. The plots demonstrate that NTM^high and BCG+NTM^high mice had a significantly higher number of inferred interactions (1391 and 1805, respectively) between different cell types than saline (692) and BCG (731) mice (Figure 7D). The strength of these interactions also correlated with the number of interactions as follows: BCG+NTM^high (89.34 %) > NTM^high (62.45 %) > BCG (23.78 %) > saline (19.65 %). In BCG+NTM^high mice, B-cells strongly interacted with macrophages, dendritic and stromal cells (Figure 7C). We then determined which signaling pathways were up or downregulated in different groups. Seven signaling pathways were identified to be activated exclusively in BCG+NTM^high mice, including WNT (Wingless-related integration site), PROS (PIK3CA-Related Overgrowth Spectrum), TWEAK (TNF-related weak inducer of apoptosis), ALCAM (Activated Leukocyte Cell Adhesion Molecule), CD6, CSF (Colony Stimulation Factor), and TNF (Tumor Necrosis Factor). In saline mice, we discovered major interactions between Thbs1 (thrombospondin1) and Sdc1 (Syndecan-1), Sdc4 (Syndecan-4), and Cd36 (Figure 7E). Additionally, the saline group exhibited Notch1-Dll4 interaction in monocytes, indicating tuberculosis progression⁴⁸. BCG mice also showed similar cellular interactions to saline, including the Thbs1 interaction with Cd47 and Spp1, Icam, and Il16 gene interactions with their ligands. In NTM^high and BCG+NTM^high mice, Lta (lymphotoxin) —characteristic of secondary lymphoid organs and involved in lymphoid tissue development—contributed significant receptor-ligand interactions with Tnfrsf1b, Tnfrsf1a, Tnfrsf14, Lta-(Ltb Ltbr). Our results indicate that Lt signaling interaction was predominant between B-cells and dendritic and stromal cells (Figure 7F, panels 2 and 4). We also noticed that Lt and Ccl interaction probability with their respective receptors was more significant in BCG+NTM^high group than in NTM^high group. Interestingly, we also observed an interaction between Xcl1 (lymphotactin) and Xcr1(lymphotactin receptor) in DCs, but only in the NTM^high group. Cross-presenting DCs are known to express Xcr1, which binds to its ligand Xcl1 (present on effector T-cells/NK cells) and induces a cytotoxic immune response mediated by DCs^49,50. In comparison, the BCG+NTM^high group exhibited significant interactions (p<0.01) between Wnt4 and its receptors, and Tnf-Tnfrsf1a, Tnf-Tnfrsf1b, Pros1-Ax1, and Cd6-Alcam. These signaling pathways are required for immune cell maintenance, renewal, activation, proliferation, and effector/memory function regulation^51-53. These findings indicate that the BCG+NTM^high group significantly upregulates signaling pathways involved in B-cell activation and maintenance, B-cell-DC interaction, and Lt-Ltbr signaling, all of which may be involved in the lymphoid follicle and GC formation.

Discussion

Our findings suggest that exposure to viable NTM via drinking water boosts BCG protection against TB by activating B-cells and humoral immunity, a branch of the immune system not effectively induced by BCG. Using three different murine strains varying in susceptibility against TB (C3H/HeOuJ, C3HeB/FeJ, and C57BL/6), we observed substantially decreased Mtb burden and pathology in lungs of mice chronically exposed to NTM upon BCG vaccination. We developed a model to specifically address several shortcomings of previous publications, including: 1) administering heat-killed NTM intraperitoneally³¹, subcutaneously⁶, or via oral gavage^31,54; 2) ceasing animal exposure to NTM upon Mtb infection^6,36; 3) inconsistent time of NTM initiation, i.e., before or after BCG immunization^55-59; 4) animals exposed to an unrealistic concentration of NTM (sometimes reaching 1x10⁸ CFU/mL). These variables do not recapitulate the natural history of human-mycobacteria interaction, which we attempted to do by: 1) animals were BCG-vaccinated at an early age, before NTM exposure (humans are routinely BCG vaccinated within 48 hours of birth); 2) exposure to live, viable NTM via drinking water (without trauma), at concentrations previously reported for HIC (1000 CFU/mL) and LMIC (1.0x10⁵ CFU/mL)^44,45,60-62, and 3) sustaining continuous NTM exposure throughout Mtb infection. Our results showing the potent mucosal response induced by NTM could provide a roadmap for future TB vaccines.

Collectively, our results suggest the following scenario (Figure S5): 1) parenteral BCG elicits T-cell mediated immunity (Figure 4); 2) PPs in the gut are colonized by live NTM given orally (Figure 2); 3) within PPs, NTM potently induce memory mucosal immunity when uptaken and presented by CX3CR1⁺ APCs (Figure 2), activating T_FH (Figure 2) and stimulating B-cell proliferation and differentiation, antibody production and isotype switching (Figure 2 and 3); 4) mucosal immunity primed in the gut seeds other mucosae including the lung, leading to antibody (IgA and IgG) production locally in BALF preceding Mtb infection (Figure 3 and 4 ); 5) upon infection with Mtb, sharing similar antigens with NTM, ectopic pulmonary GC are generated to support the expansion of memory mucosal immunity, with lymphotoxin playing an important role (Figure 5, 6 and 7); and 6) effector mechanisms including NK antibody-dependent cell-mediated cytotoxicity (ADCC) (Figure S2 and S6), control Mtb replication and limit inflammation.

While B-cells are critical in the control of infectious diseases, research on TB has mainly focused on cell-mediated immune responses^39,69-79. A growing body of evidence now appreciates a protective role for B-cells and antibodies against Mtb infection^80-87. Data from human studies have demonstrated the presence of anti-Mtb antibodies in the serum of latently infected and active TB patients⁸⁸. Notably, antibodies and B-cells have also been observed in follicular-like formations in the lungs of Mtb-infected humans and animals^89-91, displaying characteristics of B-cell aggregates and GC^90,92,93, as we found in the lungs of BCG+NTM^high mice at 120 days post-Mtb infection (Figure 5 and 6). GCs are specialized structures in which B-cells undergo somatic hypermutation and clonal selection to generate long-lived and high-affinity antibody-secreting plasma cells, as well as memory B-cells^96,97. GCs have been reported in Mtb-infected mice, non-human primates, and humans over the past decade and are associated with improved Mtb containment in the lungs⁹⁸. However, the involvement of NTM exposure and cross-protective B-cells in the development of GCs is yet unknown.

In our study, we found that B-cells significantly correlated (R=−0.52, p=0.0037) with reduced bacterial burden in the lungs throughout the course of infection (Figure 4C). We also found the presence of B-cell aggregates in the lungs of BCG+NTM^high mice at 120 dpi, correlating with decreased bacterial load and pathology (Figure 5). Our DEGs and pathway analysis (spatial transcriptomics) suggest GC formation in the lungs is associated with highly upregulated genes, including Cd19, lymphotoxin alpha and beta (Lt⍺ and Ltβ), Mki67, Bcl6, Cxcr5, Cd83, Cd38, Igha, and Tnfrsf13 (a lymphotoxin beta receptor) (Figure 6 and 7). In secondary lymphoid organs, Lt⍺ and Ltβ produced by lymphoid tissue-inducer cells bind to the lymphotoxin receptor on lymphoid tissue-organizer cells, initiating a signaling cascade that generates GCs^99,100. The interaction between Ltα/Ltβ on B cells and Tnfrsf13 on DCs, endothelial cells, and stromal cells using CellChat analysis suggests a possible role of lymphotoxin-mediated signaling in the development of GCs in the lungs and sheds light on the cellular and molecular processes leading to their induction.

Other cell types (CD3⁺ T-cells and myeloid cells) increased at 60 and 120 dpi; however, they were not directly associated with protection (Figure S7-S9). In fact, increased numbers of myeloid cells correlated with a higher bacterial burden in the lungs. While total CD3⁺ T cells were not directly associated with bacterial burden and pathology, subpopulations of CD3⁺ T cells, including CD3⁺CD4⁺CXCR5⁺ follicular helper T cells and CD3⁺CD8⁺CD44⁺Perforin⁺ effector T cells, significantly correlated with the reduced bacterial burden. Follicular helper T cells play a critical role in protective immunity, helping B-cells produce antibodies against foreign pathogens, and effector T cells participate in ADCC. Correlation analysis also showed a significant association of anti-Mtb IgA/IgG antibodies with reduced bacterial burden in the lungs (Figure S10). These findings suggest a major contribution of B-cells, antibodies, effector T cells, and NK cells in reducing bacterial burden in the lungs. To further investigate the mechanism(s) of protection, we conducted a GSE analysis and pathway enrichment of lung transcriptome 120 dpi (Figure S6). Results indicated that a high fraction of genes are involved in the "Natural Killer Cell-Mediated Cytotoxicity" pathway, suggesting that ADCC may act as a protective mechanism.

In conclusion, our results demonstrate that exposure to NTM via drinking water (similar to human exposure to NTM) does not diminish the efficiency of the BCG vaccine against TB but rather confers additional protection to BCG-vaccinated mice. BCG has been known to induce TB-specific cell-mediated immune responses for a long time, and our data suggest that NTM elicits a complementary, potent humoral immune response. Consequently, an optimal balance of mucosal and systemic immune responses protects against TB and disease progression. It has been postulated that mucosal immunity plays a crucial role in respiratory infection, including Mtb^101-103. Mucosal antibodies, including IgA, have been extensively studied and shown to be important in protecting the host from Mtb infection^104,105. However, most conventional vaccinations are intended for systemic administration, which boosts systemic immunity but induces a weaker immune response at the mucosal site of pathogen entry. Interestingly, a 2018 study by Hoft et al. reported similar findings to the ones we described above. Specifically, oral vs. intradermal BCG vaccination in humans resulted in significant IgA-mediated mucosal immunity and Th1 immunity, respectively¹⁰⁶. Our joint findings underscore the need to develop TB vaccines that elicit systemic and mucosal protection. Based on our results, NTM could be engineered as a mucosal TB vaccine as they reach, colonize, and prime immunological responses in mucosal induction site (PPs). Further research is warranted.

Limitations of the study:

We were not able to completely rule out the possibility that NTM given orally reached the lungs, explaining the presence of memory B cells and anti-mycobacterial IgA and IgG in BALF of mice before Mtb infection. Our results using conventional microbiological techniques (plating on agar) and immunofluorescence with anti-LAM suggested otherwise (Figure S11). However, their sensitivity might not suffice to identify pauci-NTM colonization of the lungs or NTM in a non-culturable state. Our hypothesis that NTM colonization of PPs plays a key role is based on previous reports that PPs —but not spleen or lymph nodes— constitute an optimal niche conducive to the generation of higher affinity IgA-producing B-cells^63-65, which then migrate and populate other mucosal sites^66,67. To confirm this hypothesis, further experiments should have been performed where B cells from the PPs or lungs from BCG+NTM^high group should be adoptively transferred to saline (control) mice and evaluate their migration to the mucosal site (including lungs) and their protective efficacy. However, due to the limitation of the samples, we could not perform this experiment.

The association between B-cells, antibodies, and protection against TB is primarily correlative and represents the second limitation of our study. Transferring protection against TB to naïve animals with serum or sorted memory B-cells obtained from NTM-exposed mice would conclusively prove this.

Star methods

RESOURCE AVAILABILITY

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by Dr. Marcela Henao-Tamayo (Marcela.Henao_Tamayo@ColoState.EDU)

Materials availability

This study did not generate new unique reagents.

Data and code availability

Spatial transcriptomics data are available at the GEO database, and the accession number is GSE203037.

All original codes will be deposited to Github and will be publicly available as of the date of publication.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

All animal studies complied with NIH guidelines and were approved by the Animal Use and Care Committee of Colorado State University. Three-week-old female C3H/HeOuJ mice were purchased from the Jackson Laboratories (Bar Harbor, ME). Animals were maintained in a BSL-3 facility at Colorado State University and had ad libitum access to water and chow.

METHOD DETAILS

Bacterial strains and culture

Mycobacterium tuberculosis HN878 was grown at 37°C in Middlebrook 7H9 liquid medium (Sigma Aldrich, cat#M0178) supplemented with 10% oleic acid-albumin-dextrose-catalase (OADC), 0.5% glycerol, and 0.05% tween 80. Mycobacterium bovis Bacille Calmette Guerin (BCG) Pasteur was cultured in a Middle brook 7H9 liquid medium. Mycobacterium avium subsp. avium (TMC 724) was purchased from ATCC (cat# 25292) and was cultured according to the manufacturer's guidelines. Briefly, the lyophilized pellet was rehydrated in 0.5 mL of 7H9 broth, then transferred in a 5 mL tube of 7H9 broth aseptically. 400uL of the 7H9 containing lyophilized pellet was plated onto 7H11 plate and incubated at 37˚C for 21 days. Single colonies were inoculated in 10 mL of 7H9 media and incubated at 37°C with constant shaking. Optical density (OD) was measured at regular intervals. When reaching an OD of 0.8, cultures were aliquoted and kept at −80˚C until used

Experimental vaccination and infections

Study 1: Study 1 was performed with three mice strains, C3H/HeOuJ, C3HeB/FeJ, and C57BL/6, using 1000 CFU/mL NTM in drinking water. Study 2: Study 2 was performed with C3H/HeOuJ mice using 1000 CFU/mL and 1.0 x 10⁵ CFU/mL NTM in drinking water.

BCG, NTM, and Mtb regimen:

For BCG vaccination, four-week-old female C3H/HeOuJ, C3HeB/FeJ, and C57BL/6 mice were immunized with 1x10⁵ CFU of BCG Pasteur subcutaneously. For NTM exposure, we selected a Mycobacterium avium strain (MA) in our study because of its highest prevalence among all NTM ¹⁰⁷. Two separate water stocks containing MA at 1000 CFU/mL and 1x10⁵ CFU/mL were prepared and administered to the mice via drinking water. Water was changed every week to maintain consistency and plated on 7H11 plates to evaluate the correct concentration (Figure S1). For low-dose Mtb aerosol infections, one mL Mtb HN878 was diluted in 7.25 ml of sterile PBS to 2 × 10⁶ CFU/ml and placed in a nebulizer attached to an airborne infection system (Glass-Col, Terre Haute, IN, USA). Mice were exposed to an aerosol infection to deposit approximately 50-60 CFU in the lungs of each mouse. After infection, five mice were euthanized, and whole lungs were plated onto 7H11 plates to evaluate the CFUs in the lungs.

Enumeration of Mycobacterial burden in lungs.

The lungs of mice were homogenized using a Blender homogenizer (Next Advance, cat# BT24M) set with a speed of eight for four minutes, with 500 μL of PBS (Corning, cat# MT21040CV) containing four 3.2 mm stainless steel beads (Next Advance, cat# SSB32). To ensure complete cell lysis, initial organ homogenates were further homogenized by transferring 100 μL of the initial homogenate to new 2.0 mL conical screw cap (Cole Parmer, cat# EW-06245-51) tubes with 400 μL of 0.05% Tween80 (Sigma Aldrich, cat# P5188), PBS (Corning cat# MT21040CV) and with 2.0 mm zirconium beads (Next Advance, cat# ZROB20) using a homogenizer (Benchmark Scientific Inc., cat# D1030) for ten seconds at 2800 rpm. After complete homogenization, serial dilutions were performed in 96 well plates (Falcon, cat# 353227) with a dilution factor of 5, and 100 μL of each dilution was plated on 7H11 agar plates. CFUs were enumerated by counting the colonies after 28 days of incubation at 37°C.

Isolation of cells and preparation of single-cell suspension

Mice were euthanized by CO₂ asphyxiation, the thoracic cavity was opened, and whole blood was collected via cardiac puncture. An equal volume of PBS was added to the blood and centrifuged at 800× g for 10 min at 25˚C with the brake-off. The buffy coat was collected, washed, and erythrocytes were lysed using 1× Miltenyi RBC lysis buffer (Miltenyi, CA, USA). Cells were washed and resuspended in one mL of freezing media and saved at −80˚C. Bronchoalveolar lavage (BAL) was harvested and centrifuged at 1300 x g for 10 minutes, and supernatant and cell (pellets) were saved separately at −80˚C without and with freezing media, respectively. Subsequently, the lungs were isolated ¹⁰⁸ and treated with DNase IV (500 units/mL, cat# D5025) solution and Liberase (0.5 mg/mL, cat# 5401127001) for 30 min at 37˚C to dissociate and digest pulmonary collagen. Cells were homogenized, passed through a 70 μm filter, and erythrocytes were lysed using Gey's RBC lysis buffer (Sigma Aldrich, cat#R7757). Lung cells were resuspended in 1 mL of Dulbecco's modified Eagle's minimal essential medium (Corning, cat#15-017-CV). For Peyer's patches (PPs) isolation, the abdominal cavity was opened, and the whole small intestine and large intestine were removed. The intestine was placed in a 12mm round plate with complete media. The PPs were harvested from the intestines and placed into tubes for flow cytometry and histopathology/immunofluorescence staining. Each mouse generally has 5-6 PPs each. For flow cytometry, PPs cells were incubated with DNase IV (500 units/mL) and Liberase (0.5 mg/mL) for 30 min at 37˚C and homogenized using a syringe plunger and passed through a 70 μm filter to prepare a single-cell suspension.

Flow cytometry

Lungs and PPs cells were counted using the method described previously by ¹⁰⁸. Cells were seeded at 1 x 10⁶/well of a 96-well V-bottom plate (Greiner Bio-One, cat#651-180). For panels including surface and intracellular markers, cells were incubated with Brefeldin A (Biolegend, cat# 420601) (1x prepared in complete media) for 6 hours in a CO₂ incubator (VWR) at 37˚C. For surface markers staining, cells were incubated with 100 μl of Zombie NIR (Biolegend, cat# 423106) viability dye (1:2000 dilution prepared in PBS) for 15 minutes in the dark. Fc receptors were blocked using an anti-mouse CD16/32 antibody (Biolegend, cat# 156604) (1:100 dilution) and stained with predetermined optimal concentrations of specific antibodies (Tables S1 and S2) for 30 minutes in the dark at 4˚C. For intracellular cytokines and transcription factors staining, cells were further incubated with 1x Foxp3 Perm/Fix buffer (eBiosciences, cat# 00-5523-00) for 1 hour at 37°C, washed with 1x permeabilization buffer (eBiosciences, cat#00-5523-00), and incubated with intracellular antibody cocktail (prepared in 1x permeabilization buffer) overnight at 4˚C. The next day, cells were washed twice, resuspended in 300 μL of 1x permeabilization buffer, and acquired using a Cytek Aurora^™ spectral flow cytometer (Cytek Biosciences), where 50,000 events of PPs and 100,000 events of lungs were recorded. Data were analyzed using manual gating and cyto-feature engineering pipeline ¹⁰⁹.

Enzyme-linked Immunosorbent assay:

An enzyme-linked immunosorbent assay (ELISA) was performed to evaluate the presence of anti-Mtb antibodies in the serum and BAL. High binding 96-half-well microplates (Corning Life Sciences, cat#3690) were coated with 100 ng of Mtb HN878 lysate (BEI, cat# NR-14824) prepared in PBS and incubated overnight at 4˚C. The next day, plates were washed five times with 180 μL of wash buffer (PBS + 0.05% Tween-20), and non-specific interactions were blocked using 180 μL of blocking buffer (PBS + 0.05% Tween-20 + 2% BSA + 2% normal goat serum [Jackson ImmunoResearch Inc., cat#005-000-121 West Grove, PA, USA]). After 2 hours, plates were washed, and different dilutions of serum and BAL prepared in blocking buffer were added to the wells and incubated for 1 hour. Plates were then washed and incubated for 1 hour with horseradish peroxidase (HRP)-conjugated anti-mouse IgG (Jackson ImmunoResearch Inc, cat# 115-035-003) or anti-mouse IgA secondary antibodies (Southern Biotech, cat# 1040-05) prepared in blocking buffer. The colorimetric substrate was developed with the addition of 100 μl of TMB substrate (Thermo Fisher Scientific, Rockford, cat# ENN301), and the reaction was stopped by adding 50 μl of 1 M sulphuric acid. Absorbance was measured at 450 nm using a BioTek Synergy 2 plate reader (BioTek Instruments Inc., Winooski, VT, USA).

Histopathological evaluation and lesion scoring

The extent of lesion burden was analyzed by lung histopathology and expressed as the proportion of lesion area over total lung area per stained tissue section. Stained slides were blinded and analyzed by a veterinary pathologist (Dr. Brendan Podell, Colorado State University). H&E-stained sections were scanned at 20X magnification using an Olympus VS120 microscope, Hamamatsu ORCA-R2 camera, and Olympus VS-ASW 2.9 software. Visiopharm software was used for image analysis. For each tissue section, a region of interest (ROI) was generated at low magnification with a custom tissue detecting algorithm using decision forest training and classification to differentiate tissue versus background based on color and area. Lesions were identified within tissue ROIs at high magnification with an additional custom-made algorithm using decision forest training and classification based on staining intensity, color normalization and deconvolution, area, and morphological features. Percent lesion calculations were integrated into the same algorithm and calculated from tissue area and lesion area as designated by the ROI and lesions detected. Lesion identification and quantification were then reviewed and edited by a pathologist as needed.

Multiplex Immunohistochemistry

For multiplexed immunohistochemistry (mIHC), formalin-fixed and paraffin-embedded (FFPE) sections were cut at 5 μm and stained using the Opal Polaris 4 Color Kit (Akoya Biosciences Inc. cat# NEL830001KT) with 4-plex staining. Five μm lung sections were stained using the OPAL multiplex IHC (mIHC) staining technology using the LabSat^™ Research Automated staining instrument (Lunaphore Technologies SA, Epedria). Slides were scanned using the multispectral camera of the Polaris^™ System (Akoya bioscience) and analyzed using the inForm tissue Finder and Phenochart^™ softwares. The panel of antibodies used was anti-LAM antibody (opal 690), anti-mouse B220 antibody (opal 570), and anti-mouse CD4 antibody (opal 520). mIHC were finally stained using the OPAL multiplex staining technology and the Leica Bond III autostainer by Imaging Core at the University of Colorado, Anschutz Medical Campus, Denver. Images were acquired using the multispectral camera of the Polaris^™ System (Akoya Biosciences Inc.) and analyzed using the inForm tissue Finder and phenochart^™ softwares.

Sample preparation, cDNA amplification, and sequencing for spatial transcriptomics

Sectioning and tissue placement:

FFPE lung tissue blocks from saline, NTM^high, BCG, and BCG+NTM^high groups were soaked in ice-cold water for 2 hours, and 5 μm sections were cut using Leica microtome. Another 10 μm section was subjected to RNA quality assessment using Qiagen RNAeasy FFPE kit and evaluating the DV200 value. Sections were collected in the water bath and allowed to float on the water surface until they were flat and free from wrinkles. Sections were then placed within the frames of capture areas on Visium spatial slides. Slides were then dried at 42°C in an oven and left in a desiccator overnight to ensure complete dehydration.

Deparaffinization, histological staining, imaging, and decrosslinking:

Slides were retrieved from the desiccator and placed at 60°C in the oven for 2 hours, cooled, and deparaffinized using xylene. Slides were washed using 100%, 96%, and 70% alcohol, respectively, and stained with Hematoxylin (Leica, cat# 3801570) and Eosin (Leica, cat# 3801570). Imaging of the sections was performed using a KEYENCE BZ-X710 microscope (KEYENCE). For decrosslinking of the sections, slides were washed with 0.1N HCl, added 100uL of TE buffer (pH 9.0), and incubated at 70°C in a thermocycler for 1 hour.

Probe hybridization, probe ligation, and probe extension and release:

After decrosslinking, probes for the mouse whole transcriptome, consisting of a pair of specific probes for each targeted gene, were hybridized to the sections at 50°C overnight in a thermocycler. Slides were washed, and a ligation step was performed to seal the junction between the probe pairs hybridized to RNA by adding probe ligation enzyme and incubating at 37°C for 1 hour in a thermocycler. The tissue sections were then enzymatically permeabilized for 40 minutes at 37°C, and probes were extended by adding UMI, spatial barcode, and partial read 1.

Library preparation and sequencing:

Released ligated probe products were harvested from the slides and further processed to generate cDNA libraries for sequencing. To determine the sample index PCR cycle number, a qPCR was performed using Kappa Sybr master mix and primers (Roche, cat# 07959389001). The cDNA was then indexed and amplified. The indexed libraries were then cleaned, and size selected using SPRIselect (Beckman Coulter, cat# B23318). The average size of the indexed cDNA libraries was determined with an Agilent HS D1000 tape (Agilent, cat# 5067-5584) using Agilent tapestation (Agilent, cat# G2991BA). The molarity and concentrations of each library were measured using a qPCR, and libraries were pooled and diluted to 4 nM. Paired-end sequencing was performed on the Illumina NextSeq500 platform at 1.8pM loading concentration, with 29 bases from read 1 and 51 bases from read 2. A total of 80-100 million read pairs per sample were generated.

Spatial transcriptomics computational analysis

Processing raw reads and demultiplexing was performed using 10X Genomics’ Space Ranger (version 1.31.1) mkfastq. Demultiplexed samples were mapped to the annotated mouse whole transcriptome (mm10) targeted gene expression analysis using Space Ranger count pipeline. Filtered reads of individual samples were converted to Seurat object and normalized using SCTransform function ¹¹⁰. Each section was analyzed separately and as a merged object to evaluate the differences between the two. Variable features of integrated data were assessed, and principal component analyses were performed. Clusters were identified using FindNeighbors and FindClusters functions and represented as Uniform Manifold Approximation and Projection (UMAP). Differentially expressed genes (DEGs) in each cluster were identified using FindMarkers from the Seurat package ¹¹¹, and cell types were assigned via SingleR package ¹¹² using ImmGenData reference in the Celldex package. Cell-cell interaction for each group was performed using CellChat, and DEGs in each group were evaluated using DESeq2. Gene enrichment analysis of DEGs of each group was performed using clusterprofiler. We extracted the gene symbols of each group and subjected them to KEGG pathway analysis using mus musculus data. The top 25 pathways were visualized, and enrichment scores were represented as the positive and negative log10 values of the corrected p-value.

Quantification and Statistical Analyses

The experiment was performed on three separate occasions, the last repetition used C3H/HeOuJ only. The sample size was calculated by differentiation of proportions of 0.9 vs. 0.1 at a power of 80% (Fisher's exact test). Group means (n = 5) CFU, total cell counts of different populations, and ELISA were analyzed using Tukey-HSD two-way ANOVA analysis to analyze differences between groups. Data were considered significant if p < 0.05. Data analysis and plotting were performed using R studio (version 4.1.2). Mean (n = 5) subjective pathology scores were compared between groups using the Kruskal-Wallis test for non-parametric data with an alpha of 0.05.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-mouse CD11c Alexa-fluor 532	ThermoFisher	Cat# 58-0116-41; RRID: AB_11218876
Anti-mouse A4b7 integrin PE	Biolegend	Cat# 120606; RRID: AB_493267
Anti-mouse CD62L PE/Dazzle 594	Biolegend	Cat# 104448; RRID: AB_2566163
Anti-mouse CXCR5 PerCP-eFluor 710	ThermoFisher	Cat# 46-7185-82; RRID: AB_2573837
Anti-mouse B220 PE/Cy7	BD Biosciences	Cat# 552772; RRID: AB_394458
Anti-mouse CD73 PerCP-Cy5.5	Biolegend	Cat# 127214; RRID: AB_11219403
Anti-mouse CD34 BV421	Biolegend	Cat# 119321; RRID: AB_10900980
Anti-mouse CD11b Pacific Blue	Biolegend	Cat# 101224; RRID: AB_755986
Anti-mouse MHC-II BV480	BD Biosciences	Cat# 566088; RRID: AB_2869739
Anti-mouse CD3 BV510	Biolegend	Cat# 100234; RRID: AB_2562555
Anti-mouse CD45 BV570	Biolegend	Cat# 103136; RRID: AB_2562612
Anti-mouse CX3CR1 BV605	Biolegend	Cat# 149027; RRID: AB_2565937
Anti-mouse CD8a BV650	Biolegend	Cat# 100742; RRID: AB_2563056
Anti-mouse CD138 BV785	Biolegend	Cat# 142534; RRID: AB_2814047
Anti-mouse NKp46 BV711	Biolegend	Cat# 137621; RRID: AB_2563289
Anti-mouse NKM-16-2-4 APC	Miltenyi Biotec	Cat# 130-102-149; RRID: AB_2660295
Anti-mouse CCR9 AF-647	Biolegend	Cat# 129710; RRID: AB_2073247
Anti-mouse CD44 APC/Fire 750	Biolegend	Cat# 103062; RRID: AB_2616727
Anti-mouse CD4 AF-700	Biolegend	Cat# 100430; RRID: AB_4936999
Anti-mouse Ly6G PerCP	Biolegend	Cat# 127654; RRID: AB_2616999
Anti-mouse CD27 BV650	Biolegend	Cat# 124233; RRID: AB_2687192
Anti-mouse CCR2 BV711	Biolegend	Cat# 357232; RRID: AB_2800970
Anti-mouse CD19 BV605	Biolegend	Cat# 115540; RRID: AB_2563067
Anti-mouse Ly6C AF700	Biolegend	Cat# 128024; RRID: AB_10643270
Anti-mouse IL-10 FITC	Biolegend	Cat# 505006; RRID: AB_315360
Anti-mouse CD4 AF-532	ThermoFisher	Cat# 58-0042-82; RRID: AB_11218891
Anti-mouse Perforin PE	Biolegend	Cat# 154406; RRID: AB_2721641
Anti-mouse PD-1 PE/Cy7	Biolegend	Cat# 109110; RRID: AB_572017
Anti-mouse IL-12 PerCP-Cy5.5	Biolegend	Cat# 505212; RRID: AB_2566225
Anti-mouse IFNγ Pacific Blue	Biolegend	Cat# 505818; RRID: AB_893526
Anti-mouse CD8 BV570	Biolegend	Cat# 100740; RRID: AB_2563055
Anti-mouse IL-4 BV605	Biolegend	Cat# 504126; RRID: AB_2686971
Anti-mouse CD3 BV605	Biolegend	Cat# 100237; RRID: AB_2562039
Anti-mouse KLRG1 BV786	Biolegend	Cat# 138429; RRID: AB_2629749
Anti-mouse TGFbeta BV421	Biolegend	Cat# 141408; RRID: AB_2650898
Anti-mouse Granzyme-B AF-647	Biolegend	Cat# 515406; RRID: AB_2566333
Anti-mouse Bcl-6 APC	Biolegend	Cat# 358506; RRID: AB_2562472
Peroxidase AffiniPure Goat Anti-Mouse IgG (H+L)	Jackson Immunoresearch	Cat# 115-035-003
Goat anti-mouse IgA HRP	Southern Biotech	Cat# 1040-05
Anti-rat IgG, HRP-linked Antibody	CellSignaling/Fisher	Cat# 7077
Anti-mouse B220 (mIHC)	BDPharma	Cat# 550286; RRID: AB_393581
Anti-mouse CD4 (mIHC)	Invitrogen	Cat# 14-9766-82; RRID: AB_2573008
Anti-mouse GL7 (mIHC)	Invitrogen	Cat# 14-5902-82; RRID: AB_467715
Bacterial and virus strains
Mycobacterium tuberculosis HN878	In house	Legacy of Dr. Ian Orme’s lab
Mycobacterium bovis, BCG Pasteur TMC#1011	A kind gift from Dr. Angelo Izzo	Bickett et. al., 2020113
Mycobacterium avium (TMC 724)	ATCC	Cat# 25292
Mycobacterium tuberculosis HN878 whole cell lysate	NIH BEI	Cat# NR-14824
Chemicals, peptides, and recombinant proteins
TMB	Thermo Fisher Scientific	cat# ENN301
Hematoxylin	Leica Biosystems	Cat# 3801570
Eosin	Leica Biosystems	Cat# 3801615
Opal Polaris 4 Color Kit	Akoya Biosciences	Cat# NEL830001KT
Deposited data
Spatial Transcriptomics	10X Genomics	NCBI GEO database Accession number: GSE203037
Experimental models: Organisms/strains
C3H/HeOuJ	The Jackson Laboratory	JAX:000635
C57BL/6	The Jackson Laboratory	JAX: 000664
C3HeB/FeJ	The Jackson Laboratory	JAX: 000658
Software and algorithms
FlowJo	BD Biosciences	RRID: SCR_009520
Rstudio (Version 4.1.2)	Open Software	RRID: SCR_000432
Cytotypr	In-house	remotes::install_github("aef1004/cytotypr")
Spatial transcriptomics analyzed data	Seurat, SingleR, CellChat, and in house codes	https://github.com/tarudutt/NTM_manuscript_data_analysis DOI: 10.5281/zenodo.7242921
Phenochart™	Akoya Biosciences	Version 1.0.12 www.akoyabio.com
inForm®	Akoya Biosciences	Version 2.4.8 www.akoyabio.com