Animal experiments in this study were approved by the Animal Care and Use Committee of the Research Institute of Tuberculosis (RIT) (permit number: No. 2021-04). Animals were treated in accordance with the ethical guidelines of RIT. The endpoints were set to determine whether the mice were imminently dying of Mtb infection and/or required compassionate euthanasia: bodyweight loss >20% of the initial bodyweight at the time of infection.

Macrophage-specific Mafb conditional-knockout (Mafb^f/f::LysM-Cre^+/+ or Mafb^f/f::LysM-Cre^+/-, Mafb-cKO) and Mafb^f/f control mice were used (11). Mafb-cKO and control mice were maintained under pathogen-free conditions in a laminar-flow facility. Wild-type (WT) C57BL/6J mice were obtained from The Jackson Laboratory Japan, Inc. Specific pathogen-free status was verified by testing sentinel mice housed within the colony.

The Mtb strain Erdman was used and stored as previously described (12–14). For determining the bacterial burden in macrophages, the infected cells were lysed with PBS containing 0.1% SDS. Infected lungs or spleens were homogenized using a ShakeMaster Neo (Bio Medical Science). The resulting cell lysates or homogenates were serially diluted and plated in duplicate on 7H10 or 7H11 agar plates supplemented with 10% Middlebrook OADC (BD Bioscience) and 0.5% glycerol. Mtb colony-forming units (CFUs) were determined by calculating the mean CFU count from the two plates at each dilution.

BMMs were differentiated as previously described (15), with some modifications. Briefly, bone marrow was isolated from the hind legs of Mafb-cKO and control mice (6 weeks), washed, and suspended into a single cell by passing through a cell strainer. The bone marrow cells were then incubated at a concentration of 2 × 10⁶ cells/mL in DMEM (Sigma-Aldrich) supplemented with 10% inactivated-fetal bovine serum (FBS, JRH Biosciences Inc.) and 10% of L929-conditioned medium in a 12-well plate for 7 days. Differentiated macrophages in DMEM containing 10% FBS were infected with Mtb at a multiplicity of infection (MOI) of one. At one day postinfection (p.i.), BMMs were collected for mRNA sequencing (mRNA-seq). At 1, 3, and 7 days p.i., the number of the intracellular bacteria within BMMs was determined by CFU.

Cytotoxicity was evaluated colorimetrically by measuring lactate dehydrogenase (LDH) released into the culture supernatant using a Cytotoxicity LDH Assay Kit (Dojindo). Briefly, BMMs from control or Mafb-cKO mice were infected with Mtb at an MOI of 1, and the LDH assay was performed at 1, 3, and 7 days p.i. The optical density at 490 nm (OD) was measured using a Varioskan LUX multimode microplate reader (Thermo Scientific). For each condition, the mean OD of four replicate wells was calculated, the background (medium) value was subtracted, and cytotoxicity was expressed as a percentage of the maximal reaction obtained by complete cell lysis. Cytotoxicity (%) = (sample OD − medium OD)/(maximal reaction OD − medium OD) × 100.

The experimental mice (age: 6–10 weeks) were transferred to a biosafety level 3 animal facility at RIT. The mice were infected with Mtb via the aerosol route using an inhalation exposure system (Glas-Col). This method routinely gave Mtb infection at 100–200 CFU per lung one day p.i.

WT and Mafb-cKO mice infected with Mtb were monitored for 315 days. The mice that survived throughout the experiments or met the endpoint were euthanized by exsanguination under anesthesia with 0.75 mg/kg of medetomidine, 4.0 mg/kg of midazolam, and 5.0 mg/kg of butorphanol via the intraperitoneal route (16). Survival probabilities between the two groups were analyzed using Kaplan–Meier analysis and the log-rank test. The body weight of infected mice was monitored during the infection.

mRNA-seq of Mtb-infected BMMs or whole lung lobes of Mtb-infected mice was performed as previously described (14). Briefly, infected BMMs or whole lung lobes were homogenized with TRIzol Reagent (Invitrogen), followed by RNA purification using an RNeasy Mini kit (Qiagen). Total RNA qualified and quantified by a 4150 TapeStation system (Agilent) and a Qubit 4 Fluorometer (Invitrogen), respectively, was subjected to construct cDNA libraries using NEBNext^® Poly(A) mRNA Magnetic Isolation Module (New England Biolabs) and NEB Next Ultra™ II DNA Library Prep Kit for Illumina (New England Biolabs). All the cDNA libraries were examined for quality using a 4150 TapeStation system and quantified with a Qubit 4 Fluorometer. The libraries were sequenced with a NextSeq1000 system (Illumina).

For read-quality trimming, raw reads were processed with Trim Galore v0.6.7 (https://github.com/FelixKrueger/TrimGalore). The expressions of transcripts were estimated by Salmon v0.12.0 directly from the processed reads (17). Transcript-level expression data was then summarized into gene-level data by the R package tximport v1.30.0 (https://github.com/thelovelab/tximport) in R v4.3.3 (18). The analysis for differentially expressed genes (DEGs) was performed by edgeR v4.0.16 (19) using generalized linear models and quasi-likelihood tests (20). The DEGs were further utilized to perform Gene Ontology (GO) enrichment analysis to identify enriched BPs using the R package clusterProfiler v4.10 (21). To reduce redundancy among the identified GOBP categories, a simplification method in clusterProfiler was used. The gene concept network of the top 3 upregulated and downregulated GOBP categories was visualized by cnetplot in clusterProfiler. KEGG pathway enrichment analysis was conducted using ShinyGO v0.82 (22), an online gene set enrichment tool (http://bioinformatics.sdstate.edu/go/). Adapted KEGG pathway diagrams were visualized using Pathview v1.42.0 in R software (23). Pathway source: KEGG (https://www.kegg.jp). Gene set enrichment analysis (GSEA) was performed locally using the GSEA desktop application (Broad Institute, v4.2.3) with the WikiPathway gene sets (c2.cp.wikipathways.v2024.1.Hs.symbols.gmt) obtained from the Molecular Signature Database (MSigDB). MafB ChIP-seq peaks (GEO GSM1964739/SRA SRX1465586) (24) were downloaded via ChIP-Atlas (accessed 27 June 2025) (25) and compared with the DEGs identified in the present study. The ChIP-seq Atlas is accessible at https://chip-atlas.org/.

The sequencing data generated in this study were deposited in the DNA Data Bank of Japan under the BioProject accession number PRJDB20606.

Quantitative reverse transcription PCR (qRT-PCR) was performed as previously described (10) with minor modifications. Briefly, total RNA from BMMs or mouse lungs was reverse-transcribed into cDNA and subjected to qRT-PCR using a KAPA SYBR Fast qPCR kit (Roche) on a QuantStudio Pro 7 system (Invitrogen). The primers used in this study are listed in Supplementary Table 1. The threshold cycle (Ct) values of target genes were normalized to that of Rplp1 and compared with the control group.

Infected lung cells were obtained using the Lung Dissociation Kit (Miltenyi Biotec) according to the manufacturer’s instructions. Briefly, cell suspensions from the lungs were incubated with ACK buffer to lyse red blood cells. The cells were washed and diluted in MACS buffer (PBS supplemented with 2mM EDTA and 2% FBS) to achieve 1–3 × 10⁶ cells/mL. The cells were incubated with TruStain FcX™ PLUS (anti-mouse CD16/32) (Biolegend), followed by staining with antibodies against CD4, CD8, CD45R, CD3, SiglecF, CD64, CD11b, CD45, or Ly6G (BioLegend). The stained cells were then fixed with the fixation buffer (BioLegend) to inactivate infected Mtb for 24 h at 4°C. The cells were analyzed on a BD FACSLyric™ using analysis software BD FACSuite™ Application V1.4.0.7047 and FlowJo™ Software v10.10 (BD Biosciences).

Whole lung lobes from infected mice were fixed with 10% formalin in PBS for over 24 h at room temperature. Tissue sections were stained with hematoxylin and eosin (H&E). Immunohistochemistry (IHC) analysis was performed as previously described (26–28). Tissue sections were stained with anti-S100a9 (1:200, R&D Systems) and digitized using a NanoZoomer S60 slide scanner (Hamamatsu Photonics). The resulting IHC images were analyzed with QuPath (29) to perform cell detection followed by object-based classification within each annotated granuloma region to quantify S100a9⁺ cells.

BMMs from control and Mafb-cKO mice were grown on coverslips in 12-well plates and infected with DsRed-expressing Mtb. At 1, 3, and 7 days p.i., cells were fixed with 3% paraformaldehyde in PBS at 4°C for 24 h, washed three times with PBS, and mounted on microscope slides using Vectashield Antifade Mounting Medium with DAPI (Vector Laboratories). Fluorescence microscopy was performed using an Olympus IX81 microscope equipped with a DP74 camera (Olympus). DAPI and DsRed fluorescence images were merged, and the number of intracellular Mtb bacilli was quantified in ImageJ (version 1.54g) (30).

The concentrations of secreted MCP-1 and IP-10 from control and Mafb-cKO BMMs infected with Mtb were measured using the Mouse CCL2/JE/MCP-1 Quantikine SixPak 2nd Gen ELISA and the Mouse CXCL10/IP-10/CRG-2 DuoSet ELISA (R&D Systems), respectively. Culture supernatants were collected from infected macrophages at 1, 3, and 7 days p.i. and filtered through a 0.45-µm pore-size filter (Toyo Roshi Kaisha).

In our previous study, we demonstrated impaired inflammatory responses in PMA-stimulated MAFB-knockdown THP-1 cells (MAFB-KD macrophages). However, no significant difference in bacterial burden was observed between MAFB-KD macrophages and control macrophages at 24 h or 48 h p.i., suggesting that the knockdown effect and/or the duration of infection was insufficient to detect intracellular bacterial proliferation (10). In this study, we investigated the effect of Mafb deficiency on mycobacterial proliferation in BMMs (Figures 1A, 2). Using BMMs derived from Mafb-cKO or control mice, we compared Mtb proliferation within BMMs (Figure 2A) and Mtb infection-induced cytotoxicity (Figure 2B). We infected BMMs with Mtb at an MOI of one and monitored CFU and cytotoxicity at 1, 3, and 7 days p.i. We confirmed the depletion of Mafb expression in BMMs from Mafb-cKO mice by mRNA-seq (Supplementary Figure 1). At 3 and 7 days p.i., intracellular Mtb proliferation was significantly higher in Mafb-cKO BMMs, suggesting that the absence of Mafb transforms macrophages into a more permissive environment for Mtb proliferation. For Mtb-induced cytotoxicity, BMMs from Mafb-cKO also showed greater susceptibility at day 3 p.i., which aligns with the higher bacterial burden in Mafb-cKO BMMs. These results support the concept that host cell death accelerates intracellular Mtb growth (31).

When macrophages are exposed to Mtb, they internalize the bacteria, and Mtb begins adapting to the intracellular environment by 24 h p.i. During this period, macrophages undergo robust transcriptional changes, indicating active host–pathogen interactions (32). To investigate the transcriptional function of Mafb in Mtb-infected macrophages, we infected BMMs from Mafb-cKO and control with Mtb and conducted mRNA-seq at 24 h p.i. mRNA-seq comparing between Mtb-infected BMMs from Mafb-cKO and control mice identified 1223 DEGs (Figure 3A, Supplementary Table 2). GO analysis for BP (GOBP) identified 974 significantly enriched GOBP terms in 614 upregulated DEGs in Mtb-infected Mafb-cKO BMMs, including leukocyte cell–cell adhesion, reactive oxygen species (ROS) metabolic process, or nucleotide metabolic process (Figure 3B, Supplementary Table 3). In 609 downregulated DEGs, 493 significantly enriched GOBP terms were identified, including response to virus, defense response to symbiont, or regulation of innate immune response (Figure 3B). Some GOBP terms, such as leukocyte migration and response to virus, were shared between upregulated and downregulated DEGs.

To visualize the interactions of the genes annotated to each GO term, we constructed gene networks (Figure 3C). Upregulated genes annotated to leukocyte cell–cell adhesion are associated with adhesion molecules or integrins (e.g., Itgb2, Itgb7, Itgal), leukocyte surface receptors (e.g., Ccr2, Cx3cr1, Cd74), MHC molecules (e.g., H2-Ab1, H2-Aa), and immune modulation (e.g. Sirpb1b, Arg2, Thbs1), suggesting that the macrophages are in the state where they are actively participating in immune surveillance, cellular communication, and antigen presentation. Upregulated genes were also annotated to ROS metabolic process including ROS generation (e.g. Cybb, Cyba), ROS detoxification and antioxidant defense (e.g., Prdx1, Nnt), and oxidative stress modulation (e.g., Thbs1, Rhoa). The downregulated genes annotated to top significantly enriched GOBP terms were highly overlapped: RNA editing and modification (e.g., Apobec1, Adar, Ifi204), innate immune sensors and IFN-stimulated genes (ISGs) (e.g., Ifit1, Ifit2, Ifit3), transcription factors and signal transduction (e.g., Pou2f2, Il10rb, Il15), cell cycle, apoptosis, and DNA repair (e.g., Eif2ak2, Pml), and metabolism and miscellaneous functions (e.g., Lacc1, Apoe). These downregulated DEGs suggested weakened pathogen sensing and reduced IFN response or inflammatory signaling, indicating a potential shift to an anti-inflammatory phenotype in Mtb-infected BMMs of Mafb-cKO mice. KEGG pathway enrichment analysis revealed that oxidative phosphorylation and chemical carcinogenesis-ROS were enriched in the upregulated DEGs (Figure 4A), whereas ECM-receptor interaction, lysosome, and endocytosis were enriched in downregulated DEGs (Figure 4B). As depicted in the lysosome pathway diagram, the proton pump ATPeV, which plays a critical role in lysosomal acidification, is downregulated (Figure 4C) (33). We validated the expression of DEGs associated with selected GO terms in BMMS by qRT-PCR (Figure 5). As expected, Cd74, H2-Ab1, Mmp12, and Nnt were upregulated, whereas, Ccl2, Gas6, and Ifit3 were downregulated in BMMs from Mafb-cKO mice relative to controls.

By GSEA using all expressed genes, a pathway of immune response to TB (https://www.wikipathways.org/instance/WP4197) exhibited impairment in Mtb-infected Mafb-cKO BMMs (Supplementary Figure 2). These enriched GOBPs and pathways are consistent with the previous results obtained from mRNA-seq of Mtb-infected MAFB-KD macrophages (10). IFN-gamma inducible chemokines (Cxcl11, Ccl2, Ccl7, Cxcl9, Cxcl10) were downregulated in Mtb-infected MAFB-KD macrophages, as well as in Mtb-infected BMMs from Mafb-cKO mice (Supplementary Figure 3). Thus, the regulation of gene expression by Mafb in mouse BMMs resembles that in PMA-stimulated human THP-1 macrophages during Mtb infection (Supplementary Figure 4).

To examine whether Mafb deficiency in macrophages influences the outcome of Mtb infection in mice, we conducted aerosol infections in Mafb-cKO mice and WT mice and monitored them for 45 weeks (Figure 1B). Mafb-cKO mice began losing body weight and showed mortality starting at 20 weeks; by the end of the study, none remained alive (Figures 6A, B, Supplementary Figure 5). The survival probability was compared between groups of the same sex using Kaplan–Meier analysis and the log-rank test. The median survival of female (n = 6) and male (n = 3) Mafb-cKO mice was 212 and 208 days, respectively, which was significantly shorter than that of WT mice. Notably, male mice were more susceptible to Mtb infection than females, exhibiting greater body-weight loss and reduced survival. We next determined the bacterial burden in the murine organs after Mtb infection. At 10 and 20 weeks p.i., Mafb-cKO mice exhibited significantly higher burden in the lungs (Figure 6C). The spleens of Mafb-cKO mice also showed a higher burden at 10 and 20 weeks p.i., demonstrating the involvement of Mafb in the control of the bacterial burden in the lung and spleen. These results indicate that Mafb-cKO mice are more susceptible to Mtb infection than control mice, consistent with the phenotype observed in BMMs.

To investigate whether Mafb deficiency in macrophages alters BPs in the lungs during Mtb infection, we performed mRNA-seq on the whole lungs of Mtb-infected Mafb-cKO and control mice at 10 or 20 weeks p.i., respectively. At 10 weeks p.i., 89 genes were identified as DEGs in the lungs of Mtb-infected Mafb-cKO mice (Figure 7A, Supplementary Table 2). Among these 89 genes, 48 genes were upregulated and 41 genes were downregulated. GOBP of DEGs demonstrated that cell–cell adhesion, leukocyte proliferation, or the regulation of T-cell activation were activated, whereas complement activation, cellular response to type II IFN, synapse pruning, and response to protozoan were suppressed in the lungs of Mafb-cKO mice (Figure 7B, Supplementary Table 4). Concept gene network for GO categories visualized that Cd1d1, Cdkn2a, Tarm1, Havcr2, and Slfn1 were the key genes for T-cell regulation (Figure 7C). KEGG pathway enrichment analysis demonstrated the enrichment of osteoclast differentiation in upregulated DEGs (Supplementary Figure 6A). Previous research demonstrated that MafB negatively regulates RANKL-mediated osteoclast differentiation (34). Consistent with this finding, downregulation of Mafb in our study led to upregulation of osteoclast differentiation–related genes (e.g., Sirpb1c, Pira2, Sirpb1a, and Sirpb1b). Complement components such as C1qa, C1qb, or C1qc, identified in suppressed GO categories, played a central role in complement activation. The involvement of Mafb in regulating complement components was consistent with the previous report (11). In addition to complement components, downregulated DEGs included cytokine ligands such as Ccl8, which is also known as monocyte chemoattractant protein 2 (MCP2), Ccl12, known as monocyte chemoattractant protein 5 (MCP5), or Pf4, known as Cxcl4. KEGG pathway enrichment analysis demonstrated that complement and coagulation cascade, and chemokine signaling pathway were enriched in downregulated DEGs (Supplementary Figure 6B).

As Mafb-cKO mice began to succumb around 20 weeks p.i., we also performed mRNA-seq to examine transcriptional changes in lungs between Mafb-cKO and control mouse. Differential expression analysis identified 267 DEGs (Figure 8A), of which DEGs found at 10 weeks p.i. were included. Among the 267 genes, 110 genes were upregulated and 157 genes were downregulated. GOBP showed that myeloid leukocyte activation, myeloid leukocyte differentiation, the regulation of macrophage activation, the regulation of endocytosis, and the regulation of angiogenesis were upregulated, whereas leukocyte migration, leukocyte chemotaxis, leukocyte proliferation, and immune response cell-surface receptor-signaling pathway were downregulated in the lungs of Mtb-infected Mafb-cKO mice (Figure 8B, Supplementary Table 4). Concept gene network revealed Csfs (GM-CSF), a key regulator for macrophage and dendritic cell function, Mmp8, Cd177, genes associated with neutrophil activation and migration, or Sirpb1 family for phagocytosis and immune modulation in upregulated DEGs, highlights strong differentiation and the activation of myeloid-derived immune cells (Figure 8C). Down regulated DEGs included Ccl22, Ccl8, Ccl5, Cx3cr1, Pf4, and Ccr7, which are involved in chemokine signaling or leukocyte migration; P2rx7, Nfatc2, and Ptpn22, regulatory genes in T-cell activation and immune tolerance, Cd22 or Icosl, which are involved in B cell-mediated immune response, suggesting reduced adaptive immune activation and leukocyte or lymphocyte recruitment in the lungs of Mtb-infected Mafb-cKO mice compared to those of control mice (Figure 8C). We validated, by qRT-PCR, the expression of DEGs associated with selected GO terms in the lungs (Figure 9). While some genes (e.g. Cd1d1 and Tspan32) showed results inconsistent with the RNA-seq data, others were consistent.

Since MafB is a transcription factor that binds Maf recognition elements (MAREs) in gene promoters (6), we evaluated whether DEGs in Mtb-infected BMMs and lungs from Mafb-cKO mice were subject to direct or indirect regulation by MafB. We compared our DEG sets with published MafB ChIP-seq data (24). Among the 1,223 DEGs in Mtb-infected Mafb-cKO BMMs, 413 (33.8%) overlapped with MafB-bound targets, including 175 upregulated and 238 downregulated genes (Table 1). In lungs from Mtb-infected Mafb-cKO mice, 28 DEGs (31.5%) at 10 weeks p.i. and 55 DEGs (20.6%) at 20 weeks p.i. overlapped with MafB-bound targets (Table 1). The proportion of DEGs directly bound by MafB was similar in Mafb-cKO BMMs and in lungs at 10 weeks p.i.; however, this proportion decreased at 20 weeks p.i., despite a greater number of DEGs overall. These findings suggest that secondary effects of Mafb deficiency contribute to the increased mortality observed in Mafb-cKO mice during the later stage of infection.

Transcriptomics of the lungs of Mtb-infected Mafb-cKO mice suggested altered recruitment of immune cells during Mtb infection (Figures 7, 8). Therefore, we investigated the proportion of immune cells in the lungs of Mafb-cKO mice during Mtb infection by flow cytometry (Figure 10). The frequencies of both CD4⁺ and CD8⁺ T-cells were high at 10 weeks p.i., and then decreased at 20 weeks p.i. in control mice, whereas they were at the same levels in Mafb-cKO mice during infection, suggesting blocked early recruitment of CD4⁺ and CD8⁺ T-cells in the infected lungs of Mafb-cKO mice (Figure 7). The frequency of B-cells remained the same from 10 weeks to 20 weeks p.i. in the control mice; however, it decreased in Mafb-cKO mice at 20 weeks p.i., which supports the transcriptomics data. Despite the impaired chemokine signaling, the frequency of neutrophils was significantly higher in Mafb-cKO at 10 or 20 weeks p.i. compared to that in control mice. Although the frequency of interstitial macrophages was slightly lower in Mafb-cKO mice, the difference was not statistically significant, likely due to high inter-sample variability, which may explain the heterogeneity in disease development. Nonetheless, these results indicate that Mafb deficiency in macrophages affects the recruitment of various immune cells to the lungs of Mtb-infected mice.

We assessed the neutrophil recruitment by histpathological analysis (Supplementary Figure 7). On H&E staining, the whole lung architecture of control and Mafb-cKO mice at 10 weeks p.i. appeared similar. By immunohistochemistry for S100a9, a neutrophil marker, Mafb-cKO mice showed stronger S100a9 signals with more neutrophil infiltration into lymphocyte-rich granulomas than the control mice, consistent with the flow cytometric analysis (Figure 10).

Several limitations should be considered when interpreting our findings. Although Mafb-cKO BMMs exhibited significantly higher CFU at 7 days p.i., we did not observe a corresponding increase in cytotoxicity in the assay. This discrepancy may reflect differences in assay sensitivity, as the cytotoxicity assay predominantly measures cellular metabolic activity rather than direct cell death. Additionally, the medium change performed on day 3, required to maintain long-term cultures, may have influenced metabolic readouts and reduced the ability to detect subtle differences in viability at later time points. Using bacteria expressing a fluorescent protein, intracellular fluorescence signals did not reveal a clear difference between control and Mafb-cKO BMMs (Supplementary Figure 8), likely due to methodological limitations such as limited dynamic range, signal saturation, and inability to distinguish viable from nonviable bacteria. In contrast, CFU enumeration selectively quantifies viable replicating bacteria and is therefore more sensitive for detecting early differences in bacterial proliferation within macrophages. Although transcriptomic data indicated reduced expression of Ccl2 and Cxcl10 in Mafb-cKO BMMs during Mtb infection, the corresponding protein level measured by ELISA was not significantly different between groups (Supplementary Figure 9). Finally, we did not evaluate T cell activation in Mafb-cKO mouse lungs, and therefore our study cannot fully determine how Mafb deficiency influences the relationship between intracellular bacterial replication, macrophage death modalities, and downstream immune responses during Mtb infection.