Batf2/Irf1 Induces Inflammatory Responses in Classically Activated Macrophages, Lipopolysaccharides, and Mycobacterial Infection

Macrophages are widely distributed cells of the immune system that play an indispensable role in a vast range of biological processes by acquiring diverse states of activation in response to tissue signals and environmental stresses. The various immunological processes are carried out by phenotypically polarized macrophages called classically activated macrophages (M1) (1) and alternatively activated macrophages (M2) (2, 3). M2 is induced by IL-4 and IL-13 and involved in the regulation of inflammation in many biological processes, including type 2 infectious diseases (4), tissue repair, insulin resistance (5), and thermogenesis (6) among many others. M1 is induced by IFN-γ, mainly secreted from Th1 cells and NK cells, leading to the secretion of proinflammatory mediators, the release of reactive oxygen, and nitrogen intermediates (7–10), and is thus important for killing effector function, particularly for intracellular pathogens. M1 is usually associated with the control of acute infections.

Macrophages are the most important target cell population in which Mycobacterium tuberculosis resides and multiplies by challenging the host-protective immune response. Protective immunity against M. tuberculosis is dependent on an IL-12–driven dominant Th1 immune response and consequent microbicidal activities by M1 (11). This is supported by clinical data collected during the early phase of M. tuberculosis infection from patients, which demonstrated that macrophages were polarized toward an M1 profile (12). The major components of immune target genes for IFN-γ stimulation include MHC classes I and II, inflammatory and pyrogenic cytokines such as TNF-α, IL-6, IL-12, TGF-β, and IL-10, chemokines Ccl2, Ccl3, Ccl4, and RANTES (Ccl5), chemokine receptors Cxcl2, Cxcl3, Cxcl4, and Cxcl5 (13, 14), and antimicrobial proteins such as inducible NO synthase (Nos2) (7, 15). M1 also produces phagocyte oxidase and immune GTPases, which are major components of innate immunity (14). NO is usually associated with the control of acute infection (16), but it may be also responsible for necrosis and tissue damage, which leads to infectious pathogenesis when uncontrolled (17, 18). M1 also leads to the induction of immune regulatory transcription factors such as IFN regulatory factors (Irfs), Stats, Nfκb, and Ap1 (19, 20), which modulate inflammatory gene expression. Irf1, the first member of the IRF family to be identified, targets different sets of genes in various cell types in response to diverse cellular stimuli and evokes appropriate innate and adaptive immune responses (21). Irf8, also known as IFN consensus sequence-binding protein, seems to have broader effects on microbe-induced cytokine production than does Irf1. Irf8 can function both as a transcriptional repressor and as an activator, depending on the partners with which it interacts, and it plays crucial roles in macrophage activation (21). Previous studies indicated that Irf8 activates IL-12 p35 gene transcription in synergy with Irf1 (21, 22) and enhances Ccl5 gene transcription in cooperation with PU.1, Nfkb, and Irf1 (23).

Despite those extensive studies, there seems to be many unknown molecular mechanisms that govern this complicated transcriptional regulation in M1. In the present study, we report that a transcription factor Batf2 is involved in classical activation of macrophages. We found that Batf2 is drastically induced in IFN-γ–activated classical macrophages but not in IL-4–activated alternative macrophages. Batf2 knockdown experiments in IFN-γ– or LPS-stimulated macrophages revealed that Batf2 plays an important role in innate immune response by controlling expression of several important immune regulatory genes, such as Nos2, Tnf, Ccl5, Cxcl9, Cxcl11, Ccr5, Cxcr3, Il6, and Niarc1, possibly by the Batf2/Irf1 complex. M. tuberculosis infection amplified Batf2 expression in M1 via the TLR pathway for induction of inflammatory response in macrophages. Hence, the Batf2 pathway might be a potential target for drug and vaccine development against infectious disease.

All experiments were performed in accordance with the Animal Research Ethics Committee of the Riken Yokohama Institute (25-007) and the South African National Standard (SANS 10386:2008) and the University of Cape Town practice for laboratory animal procedures. The protocol (permit number: 012/036) was approved by the Animal Ethics Committee, Faculty of Health Sciences, University of Cape Town (Cape Town, South Africa).

Bone marrow–derived macrophages (BMDMs) were generated from 8- to 12-wk-old BALB/c male mice as described previously (24). Briefly, bone marrow was harvested from femurs and differentiated into macrophages in sterile tissue culture grade 140 × 20-mm petri dishes with Vent (Nunc, Roskilde, Denmark) containing Plutznik media (DMEM containing 10% FCS, 5% horse serum, 2 mM l-glutamine, 1 mM sodium pyruvate, 0.1 mM 2-ME, 30% L929 cell–conditioned medium, 100 U/ml penicillin G, and 100 μg/ml streptomycin). After 10 d of differentiation, macrophages were harvested and plated into six-well plates (Nunc) at a concentration of 2.5 × 10⁶ cells per well. Following overnight incubation, cells were either left untreated or stimulated with IFN-γ (100 U/ml, BD Biosciences, San Jose, CA), LPS from Escherichia coli 0111:B4 (100 ng/ml, Sigma Aldrich), heat-killed M. tuberculosis HN878 (HKTB; 5μg/ml), or with synthetic triacylated lipoprotein Pam3CSK4 (100 ng/ml, InvivoGen) and incubated at 37°C under 5% CO₂. At 2, 4, 6, and 24 h after stimulation for IFN-γ and 4 h for LPS, HKTB, or Pam3CSK4, supernatants and RNA were collected. Total RNA was extracted by an miRNAeasy kit (Qiagen, Valencia, CA).

For M. tuberculosis infection, the M. tuberculosis HN878 strain was cultivated in 7H9 broth and grown to exponential phase. BMDMs were cultured overnight into six-well plates (Nunc) at a concentration of 4.5 × 10⁶ cells per well and then left untreated or stimulated with IFN-γ or IL-4 (100 U/ml each, BD Biosciences, San Jose, CA) overnight. After 4 h of infection with M. tuberculosis (multiplicity of infection [MOI] of 5), BMDMs were washed once with prewarmed culture media to remove extracellular bacteria and incubated with media containing fresh IFN-γ or IL-4 activators and 10 μg/ml gentamicin. At 4, 12, 24, 48, and 96 h after infection, supernatants were collected and RNA was extracted.

Three percent Brewer-modified thioglycollate (Becton Dickinson) was prepared according to the manufacturer’s instructions. Sterile thioglycollate medium (1 ml) was injected into BALB/c mice i.p. Peritoneal exudate cells were harvested 7 d after the injection using 15 ml ice-cold sterile PBS with 5 mM EDTA. Cells were washed and RBCs were lysed. After washing twice with medium, peritoneal macrophages were seeded in six-well plates (2 × 10⁶ cells/well). After overnight incubation, cells were washed and used in stimulation experiments.

Total RNA (500 ng) was amplified using the Ambion total RNA amplification kit (Ambion, Carlsbad, CA). cRNA was hybridized to an Illumina mouse Sentrix BeadChip WG-6V2 array (Illumina, San Diego, CA) according to the manufacturer’s protocol. Scanning of the chip was performed using Illumina BeadScan and BeadStudio software packages, and data were generated using BeadStudio (version 3.1). Quantile normalization of the microarray data was executed using limma packages of Bioconductor in the R statistical language. Three biological replicates were analyzed for the microarray experiments. Microarray data were deposited in the Gene Expression Omnibus database, accession number GSE59210 (http://www.ncbi.nlm.nih.gov/geo).

Total RNA (500 ng) was reserve transcribed using the ThermoScript RT-PCR system (Life Technologies, Carlsbad, CA), and quantitative PCR was performed using SYBR Premix Ex Taq (TaKaRa Bio) and the 7900HT real-time PCR system (Applied Biosystems). Relative expression of mRNA was determined after normalization with Gapdh using the ΔΔCT method. A Student t test was used to evaluate statistical significance.

Protein was isolated using NE-PER nuclear and cytoplasmic extraction reagents (Thermo Fisher Scientific, Waltham, MA). Protein concentration was measured using a BCA kit (Thermo Fisher Scientific). Five micrograms nuclear proteins in each lane was subjected to 4–12% bis Tris Novex gel electrophoresis (Life Technologies) and transferred to nitrocellulose membrane. The membrane was probed with Batf2 Ab (L-24; Santa Cruz Biotechnology, Santa Cruz, CA), incubated with HRP-conjugated anti-rabbit IgG (Abcam, Cambridge, MA), and developed using an ECL plus kit (GE Healthcare, Buckinghamshire, United Kingdom). Chemoluminescence was recorded with a Fuji LAS-3000 luminescent analyzer. The membrane was stripped and reprobed with control Abs against β-actin (Sigma-Aldrich, St. Louis, MO).

BMDM cells were seeded in Nunc Lab-Tek Chamber Slide (Sigma-Aldrich) and stimulated with IFN-γ for 6 h. Cells were then fixed with methanol, washed, quenched with 0.3 M glycine, blocked with 0.3 M glycine in 0.1% Tween 20, followed by incubation overnight with Batf2 Ab (Santa Cruz Biotechnology). Cells were washed with PBS, incubated for 1 h with secondary Ab, Alexa Fluor 488 goat anti-rabbit IgG (Abcam) diluted in block solution. Cells were washed with PBS and incubated for 1 h with Alexa Fluor 594 WGA (Abcam) to stain the plasma membrane. Cells were washed and mounted in the ProLong Gold antifade reagent (Life Technologies) and viewed by confocal microscope (Leica TCS SPE system; Leica, Wetzlar, Germany) at ×63 magnification (HCX Plan Apochromat with a 63×/1.40-0.06 oil objective lens; Leica) using acquisition software (LAS AF/TCS SPE; Leica).

Control short hairpin RNA (shRNA), Batf2 shRNA, and Irf1 shRNA clones were purchased from Open Biosystems (Thermo Fisher Scientific). Plasmids were purified by endotoxin-free plasmid midiprep kit (Promega, Fitchburg, WI) according to the manufacturer’s instruction. Seventeen micrograms plasmid, 10 μg HIV-gp, and 10 μg vesicular stomatitis virus envelope genes were cotransfected into 4 × 10⁶ 293T cells using FuGENE HD (Roche Applied Science, Mannheim, Germany) in Opti-MEM (Life Technologies) medium containing 5% FBS at 37°C, 5% CO₂. After 2 d of incubation, the supernatant was collected and centrifuged at 19,400 rpm for 2 h at 20°C. Pellets were dissolved in 100 μl HBSS and stored at −80°C for future use.

For lentivirus infection, BMDM cells (1.5 × 10⁶ cell/well in a six-well plate) were incubated overnight followed by media change and replenishment with 8 μg/ml Polybrene (EMD Millipore, Billerica, MA) containing DMEM (with 10% FBS, antibiotics). BMDM cells were infected with lentivirus for negative control shRNA and Batf2 shRNA at an MOI of 1:10. After 24 h, the media were removed and 4 ml fresh DMEM with 10% FBS was added. The media were changed every third day. On day 6 of the transfection, medium was replenished with 3 μg/ml puromycin (Sigma-Aldrich) containing DMEM with 10% FCS. On day 10 of the transfection, cells were washed and used for the stimulation experiments.

For identification of Batf2-regulated genes, gene expression profiles were compared between IFN-γ–stimulated BMDMs at 4 h and unstimulated control, both negative shRNA treated, and between IFN-γ–stimulated Batf2 knockdown BMDMs and negative shRNA-treated IFN-γ–stimulated BMDMs. Genes with >1.6-fold change significantly (p < 0.001) were selected as up- and downregulated gene. Genes upregulated by IFN-γ stimulation and downregulated in IFN-γ–stimulated Batf2 knockdown were selected as positively regulated genes. Additionally, genes downregulated by IFN-γ stimulation and upregulated in IFN-γ–stimulated Batf2 knockdown were selected as negatively regulated genes.

Cell culture supernatants were analyzed for the production of NO₂⁻ using the Griess reagent assay as previously described (4, 25).

TNF-α and IL-12p40 levels in the culture supernatants were measured by sandwich ELISA using tetramethylbenzidine substrate solution (Roche Diagnostics, Mannheim, Germany) for development (4).

Coimmunoprecipitation experiments were done using the Dynabeads protein G kit (Life Technologies). Briefly, 10 μg Irf1 (Cell Signaling Technology) and Irf8 (Santa Cruz Biotechnology) Abs was diluted in 200 μl PBS containing Tween 20. Ab was incubated with the Dynabeads protein G at 4°C for 2 h. Nuclear protein was prepared from unstimulated and IFN-γ– or LPS-stimulated BMDM cells after 6 h. The Ab-conjugated Dynabeads were washed and incubated overnight with 5 μg nuclear protein extract. The beads were washed and protein was eluted using 20 μl elution buffer. Ten microliters premixed NuPAGE LDS (Life Technologies) sample buffer and NuPAGE sample reducing agent was added to the samples. The sample was heat denatured and subjected to Western blot experiment.

To identify novel regulation mechanisms in M1, mouse BMDMs were stimulated with IFN-γ (100 U/ml) in a time-dependent manner and total RNA was subjected to an Illumina beads microarray (Fig. 1A). Batf2 mRNA was highly induced in IFN-γ–stimulated BMDMs in a transient manner, together with known induced transcription factors Irf1, Irf8, Stat1, Stat3, Nfkb1, Jun, and Crem (Fig. 1B). The expression level of Batf2 was close to the detection limit at time 0, but drastically induced after 2 h of IFN-γ stimulation (28.0 ± 12.5-fold in comparison with the unstimulated control), keeping a high level until 6 h and downregulated at 24 h after stimulation. Because Batf2 is a member of the AP-1 basic leucine zipper transcription factor (activating transcription factor) family (26–29), we also explored expression profiles of other family members Batf and Batf3 (Fig. 1C). Batf and Batf3 showed low expression and were not induced by IFN-γ stimulation. Those intensities were quantitatively confirmed by the capped analysis of gene expression (CAGE) measurement (30), which shows that Batf2 was not induced by IL-4, IL-13, or IL-4/IL-13 (Supplemental Fig. 1A, 1B) in activated BMDMs, suggesting that Batf2 induction was specific for M1. In contrast, we found weak Batf and Batf3 induction by IL-4–stimulated M2 in the CAGE measurement (Supplemental Fig. 1C). The Batf2 induction was further confirmed using independent experiments by quantitative real-time RT-PCR (qRT-PCR) (Fig. 1D, Supplemental Fig. 1D) and thioglycollate-elicited peritoneal macrophages following IFN-γ stimulation (Supplemental Fig. 1E). Next we confirmed Batf2 induction at a protein level using Western blot analysis (Fig. 1E). As expected from the gene expression analysis, Batf2 was transiently upregulated at 4 h with a peak at 6 h and was downregulated to the unstimulated level at 24 h, which showed slightly delayed responses by protein induction in comparison with mRNA expression profiles. We further explored the localization of IFN-γ–induced Batf2 by immunohistochemical staining, which demonstrated a predominant localization in the nucleus (green fluorescence) at 6 h IFN-γ–stimulated BMDMs (Fig. 1F). Taken together, these results show that Batf2, but not Batf and Batf3, is transiently induced in BMDMs by IFN-γ stimulation.

To understand the functional importance of Batf2 in IFN-γ stimulation, we assessed the effect of Batf2 knockdown in IFN-γ–stimulated BMDMs. Because macrophages are one of the most difficult cells for lipofection-based gene knockdown by siRNA, we prepared lentivirus expressing Batf2 shRNA. BMDMs were infected by the lentivirus at an MOI of 1:10, and total RNA was extracted after 10 d of infection followed by qRT-PCR analysis. The mRNA expression indicated that knockdown efficiency of 75% on average was achieved in Batf2 shRNA-transfected cells as compared with the negative control shRNA-transfected cells (Fig. 2A). Additionally, significant knockdown (p < 0.01) of Batf2 mRNA level was also obtained after IFN-γ stimulation at 4 h (Fig. 2B), which was further confirmed at the protein level by Western blot at 6 h (Fig. 2C).

The consequences of Batf2 knockdown in IFN-γ–stimulated BMDMs was explored by qRT-PCR measurement of representative classical effector genes. IFN-γ–mediated induction of the major classical effector genes, NO synthase 2 (Nos2) and TNF-α (Tnf), were strikingly downregulated by Batf2 knockdown (Fig. 3A, 3C). The induction of the Nos2 enzyme leads to the production of NO, measured as nitrite in the culture medium by the Griess reagent assay. Nitrite was drastically reduced in Batf2 knockdown cells compared with the negative control cells (Fig. 3B), validating downregulation of IFN-γ–induced Nos2. We also confirmed by ELISA assay that IFN-γ–mediated induction of secreted TNF-α in the culture medium was significantly reduced by Batf2 knockdown (Fig. 3D). Of note, IFN-γ–mediated weak induction of the chemokine Rantes (Ccl5) also revealed downregulation by Batf2 knockdown (Fig. 3E). In contrast, IFN-γ–mediated induction of suppressor of cytokine signaling 1 gene (Socs1) was not perturbed by Batf2 knockdown (Fig. 3F). IL-10 (Il10) was induced by IFN-γ but not perturbed by Batf2 knockdown (Fig. 3G). Taken together, this demonstrates that important early immune response genes, including macrophage killing effector genes, are regulated by Batf2.

To globally understand Batf2-regulated genes in IFN-γ–stimulated BMDMs, the Batf2 knockdown samples were subjected to the Illumina beads microarray. Out of 670 IFN-γ–induced genes, 78 genes (11.6%) were found as positively regulated genes by Batf2 (Supplemental Fig. 2A), where 66 of 78 genes are early response genes with peak expression up to 6 h of IFN-γ stimulation. In addition to Tnf, Rantes (Ccl5) and Nos2 that were identified by qRT-PCR, well-characterized classical effector genes, such as Ccr5, Ccl11, Cxcl9, Cxcl11, Ifitm1, Niacr1, Igf2bp2, Il6ra, and Irg1, were regulated by Batf2. Because transcription factor genes Klf6, Crem, Bcl3, and Batf3 were also under the regulation of Batf2, it is possible that some of the 78 affected genes may be directly regulated by those transcription factors. Gene Ontology (GO) analysis for the Batf2 positively regulated genes revealed significant enrichment of the ontology term for immune response such as “immune response” (GO 0006955), “cytokine-mediated signaling pathway” (GO 0019221), “innate immune response” (GO 0045087), “response to wounding” (GO 0009611) “regulation of T cell activation” (GO 0050863), and “regulation of interleukin-6 production” (GO 0032675) (Fig. 4). Alternatively, of 458 IFN-γ–downregulated genes, 43 genes were selected as negatively regulated genes by Batf2 (p < 0.001) (Supplemental Fig. 2B). Although we found classical effector genes such as Cxcr3 and Slc17a9 among those genes, GO analysis of Batf2 negatively regulated genes did not show significant enrichment of ontology terms for the positively regulated genes, except for the term “fatty acid elongation” (GO 0019367). These results indicate that Batf2 plays an important role for the positive regulation of immune responses by altering expression of cytokines and chemokines and therefore possibly maintaining the balance in inflammatory processes.

To test the biological relevance of Batf2 regulation in an infectious disease model, BMDMs were either left unstimulated or stimulated with IFN-γ or IL-4 for 24 h (Fig. 5A). The treatment of BMDMs with IFN-γ resulted in rapid induction of Batf2 at 2 h after treatment, keeping the high level until 6 h and downregulation at 24 h (blue line in Fig. 5A). Subsequent infection with a clinical M. tuberculosis, Beijing strain HN878 (Fig. 5A time point 0, at 24 h), in a time course CAGE expression analysis demonstrated induction of Batf2 within 4 h after infection in unstimulated BMDMs (yellow line in Fig. 5A), indicating that M. tuberculosis itself is able to induce Batf2 independently of IFN-γ stimulation. However, Batf2 induction was strongest and prolonged in M. tuberculosis–infected IFN-γ–stimulated macrophages with up to 6-fold increased Batf2 gene expression at 4 and 12 h postinfection (green line in Fig. 5A), compared with unstimulated or IL-4–activated macrophages. These results indicated that IFN-γ and M. tuberculosis infection may act in concert to upregulate Batf2. Of note, infection also induced Batf in unstimulated macrophages and in IFN-γ–stimulated macrophages (Fig. 5B, 5C), whereas M. tuberculosis infection did not induce Batf3. A comparison from the Batf2 knockdown experiment in IFN-γ–stimulated macrophages (Supplemental Fig. 2A) demonstrated that many of the perturbed genes, including Tnf, Ccl5, and Irg1, but not all, were also induced by M. tuberculosis infection (Supplemental Fig. 3A). These results indicate that M. tuberculosis infection has an overlapping effect on the perturbation of Batf2 target genes with IFN-γ, but has also distinct differences. Biologically important for host defense, M. tuberculosis infection in IFN-γ–stimulated macrophages drastically induced known host-protective genes such as Tnf, Ccl5, and Cxcl11 (Supplemental Fig. 3A), including Nos2 (Fig. 5D, 5E, Supplemental Fig. 3B, 3C). Interestingly, M. tuberculosis–infected unstimulated and IL-4–stimulated macrophages also induced Nos2 to a certain level, which resulted in residual NO production in IL-4–stimulated macrophages but not in unstimulated macrophages (Fig. 5D, 5E, Supplemental Fig. 3B, 3C). This might indicate that additional factors are needed to produce sufficient NO, depending on the activation status of host macrophages. Taken together, these results demonstrate that M. tuberculosis infection induces Batf2, particular in IFN-γ classical macrophages, which upregulates important inflammatory and host-protective effector genes.

M. tuberculosis–induced gene expression change is known to be mediated by TLR-2 and TLR-4 (31), a different signal transduction pathway from IFN-γ stimulation by its receptors Ifngr1 and Ifngr2. To explore how M. tuberculosis activates Batf2 expression, BMDMs were stimulated with bacteria-derived LPS and HKTB, which induce macrophage activation through TLR-4 and TLR-2, respectively (31–33). Batf2 expression was induced in both LPS and HKTB (Fig. 5F) similarly to IFN-γ (Fig. 1D), although the induction was relatively weak in HKTB. TLR-2–mediated Batf2 induction in BMDM cells was also confirmed by stimulation with Pam3CSK4, a specific activator for TLR-2 (Supplemental Fig. 3D). Furthermore, we confirmed that costimulation with IFN-γ and LPS synergistically induced Batf2 mRNA in both BMDM cells (Fig. 5G) and thioglycollate-elicited peritoneal macrophages (Supplemental Fig. 3E).

Subsequently, we performed Batf2 knockdown experiments in LPS-stimulated BMDMs (Fig. 6A) and explored the effect to the above found set of representative classical effector genes (Fig. 6B–J). LPS stimulation induced Nos2, Tnf, and Ccl5 mRNA expression, which were significantly downregulated by Batf2 knockdown experiments (Fig. 6B–6F). Interestingly, LPS-mediated Socs1 mRNA induction was significantly downregulated by Batf2 knockdown (Fig. 6G), which was not observed in IFN-γ stimulation. Furthermore, we observed LPS-mediated induction of Il10, although it was not downregulated by Batf2 knockdown (Fig. 6H). Finally, Il12b was induced by LPS stimulation, which was significantly downregulated by Batf2 knockdown (Fig. 6I, 6J). The same experiment was also performed in HKTB-stimulated BMDMs (Fig. 6K–O). HKTB stimulation also induced Nos2, Tnf, Ccl5, and Il12b mRNA, which were significantly downregulated by Batf2 knockdown experiments (Fig. 6L–O). Taken together, these results demonstrate that Batf2 is induced by LPS and HKTB stimulation in macrophages via TLR-4/TLR-2, leading to Nos2, Tnf, and Ccl5 mRNA expression, also seen in M1 and M. tuberculosis infection.

Because basic leucine zipper transcription factors, including Batf2, lack the DNA binding domain, they regulate downstream genes by physical association with other transcription factors. Therefore, we performed the oPOSSUM program, an enrichment analysis of transcription factor binding motif (34). We identified RELA, SPIB, FOXA2, MZF1_1-4, IRF1, and NF-κB as significantly enriched motifs at the promoter sites of Batf2-regulated genes (Fig. 7A). Motif analysis exhibit that 27 of 79 Batf2-perturbed genes have Irf1 binding sites (Supplemental Table I).

So far it has been recently shown that Batf/Irf4 and Batf/Irf8 associations regulate compensatory dendritic cell development through the AICE motif (35–38), suggesting that Batf2 may also associate with Irf proteins. In fact, 17 of 78 Batf2 positively regulated genes possess the AICE motif significantly (p = 2.581 × 10⁻⁵). Thus, we explored physical association of Batf2 with IFN-γ–induced Irfs, Irf1, and Irf8 using immunoprecipitation. We found that Batf2 is coimmunoprecipitated by Irf1 but not by Irf8 (Fig. 7B). To further explore the role of Irf1 we performed Irf1 knockdown experiments in LPS-stimulated BMDMs (Fig. 7C) and explored the effect of the above found (Supplemental Table I) set of classical effector genes (Fig. 7D–G). Irf1 knockdown significantly downregulated IFN-γ– and LPS-activated classical effector genes Nos2, Tnf, Ccl5, and Il12b in LPS-stimulated BMDMs (Fig. 7D–G). These results indicate that Batf2 and Irf1 regulated the same classical genes and suggest that some of the Batf2-regulated genes may be regulated through Batf2/Irf1 interaction.

Batf2 belongs to the AP-1 basic leucine zipper transcription factor (activating transcription factor) family proteins, composed of Batf, Batf2, and Batf3 (26–29, 35), with Batf2 having very limited functional information available. Batf family proteins do not have a DNA binding domain and bind to DNA indirectly with support of other transcription factors. In this study, we demonstrate that Batf2 is involved in gene regulation of IFN-γ–activated classical macrophages as well as LPS/TLR-induced macrophage stimulation. Batf2, most probably in complex with the transcription factor Irf1, induces important classical effector genes in macrophages, among them Tnf, Nos2, Ccl5, and Il12b for inflammatory responses against type 1 infectious diseases, such as M. tuberculosis.

Gene expression studies in IFN-γ–activated M1 and in TLR-stimulated macrophages revealed a strong induction of the transcription factor Batf2, which was not observed in unstimulated or IL-4–activated M2, and hence Batf2 can be classified as a new marker for M1. Batf2 can compensate for Batf3 in CD8⁺ and CD103⁺ dendritic cell development during Toxoplasma gondii infection (36). In contrast, expression of Batf and Batf3 were not altered during macrophage stimulation; however, M. tuberculosis infection induced Batf in macrophage infection. A recent publication showed that Batf is involved in IL-12 production in CD103⁺ dendritic cells in response to mycobacteria and Listeria infection (36), associated with class switch recombination in B cells (39, 40), and is essential for the development of Th17 cells, T follicular helper cells, Th1 cells, and Foxp3⁺ T cells (41, 42), effector function of CD8⁺ T cells (43), as well as maintenance of hematopoietic stem cells (44). Moreover, Batf3 has been shown to be essential for the development of CD8α⁺ classical dendritic cells and related CD103⁺ dendritic cells that cross-present Ags to CD8⁺ T cells and produce IL-12 in response to pathogens (36). This is in agreement with our observation that induction of Batf3 was not found in IFN-γ, IL-4, or M. tuberculosis infection in macrophages.

Batf2 protein was predominantly found in the nucleus after 6 h of IFN-γ stimulation, indicating its involvement in signal transduction–mediated gene regulation. This was confirmed by Batf2 knockdown experiments, which perturbed a limited number of inflammatory effector genes, demonstrating that Batf2 is an essential transcription factor for gene regulation and effector functions in classical macrophage activation. Owing to the incapability of currently available chromatin immunoprecipitation-grade Ab against Batf2, we were limited by coimmunoprecipitation studies where we could demonstrated that Batf2 associates with Irf1 in IFN-γ– and LPS-stimulated BMDMs. Transcription factor binding motif analysis showed that 27 of 78 Batf2-perturbed gene such as Tnf, Ccl5, Ccr5, Irg1, Il6ra, and Cxcl9 has Irf1 binding sites (Supplemental Table I). Irf1 knockdown in LPS-stimulated BMDMs resulted in downregulation of classical effector genes. Thus, our results suggest that Batf2-mediated regulation of Irf1-regulated classical effector genes is carried out in a combinatorial way by association with Irf1. A recent publication revealed the importance of Batf/Irf complexes, where IRF4/BATF/JunB complexes bind to presumptive regulatory sequences in genes required by Th17 cells (37, 38, 45) and the Irf4, Irf8, and Batf complex is essential for the development of CD8⁺ classical dendritic cells (36, 37), supporting our findings. We observed that Socs1 expression is not affected by Batf2 knockdown in IFN-γ induction, but is affected in the LPS induction (46, 47). It is known that signal transduction of IFN-γ stimulation consists of Stat-dependent and -independent pathways where Nos2 and Tnf expression are regulated by the former pathway and Socs1 by the latter one (48), suggesting that Batf2 is not involved in the transcriptional regulation downstream of the Stat-independent pathway. However, our results show that Socs1 expression is alternatively transcriptionally regulated in a Batf2-dependent manner downstream of TLR-4–mediated signal transduction, where Il12b gene expression may be also under the same transcriptional regulation. Such complicated transcriptional regulation needs to be further explored to understand precise regulation mechanisms for classical effector genes.

M. tuberculosis infection induced Baf2 in macrophages, suggesting that IFN-γ stimulation and M. tuberculosis infection act in concert. This is supported by IFN-γ/LPS costimulation experiments in macrophages, which increased induction of Batf2 expression compared with single stimulation (Fig. 5G, Supplemental Fig. 3E) (36). Moreover, some of the downregulated genes found during Batf2 knockdown experiments in IFN-γ–stimulated macrophages were strongly upregulated during M. tuberculosis infection in macrophages, and particularly in infected IFN-γ–stimulated macrophages. One of the main functions of M1 is their inflammatory responses and antimicrobial functions, including the production of reactive nitrogen intermediates, which effectively kills M. tuberculosis and other intracellular pathogens in mouse macrophages. IFN-γ– and LPS-induced NO synthase is catalyzed by l-arginine to citrulline and NO, the later an effective antimicrobial radical and one of the main killers of Mycobacterium in mice, also demonstrated by Nos2-deficient mice, which are hypersusceptible to M. tuberculosis (16, 25). Global analysis of Batf2 knockdown genes show well-characterized classical effectors such as Ccl11, Cxcl9, Ccr5, Niacr1, Il6ra, and Irg1 were downregulated after Batf2 knockdown. Interestingly, M. tuberculosis infection experiments using knockout mice from our list of Batf2-regulated genes, such as Nos2, Ccr5, Niacr1, Cxcr3, Irf1, and Stat1, showed either susceptible or resistant phenotypes to tuberculosis infection (49–55). Therefore, Batf2 may play an important role in immune response by possibly maintaining the balance in the inflammatory process. Recently generated Batf2-deficient mice showed increased mortality to T. gondii infection compared with wild-type and Batf3-deficient mice. Parasite burden and cytokine levels were, however, similar compared with control and normal macrophage populations and were present in the peritoneum and liver but reduced in the lungs (36). However, a possible role of macrophages in toxoplasmosis was not explored. Host protection against T. gondii is also dependent on M1 effector functions, including NO. Together with our M. tuberculosis study, these results underline the role of Batf2/Irf1 in certain type 1 protective experimental human infectious diseases. Manipulating the Batf2 pathway may be therefore a valuable strategy to increase immunity to type 1 infectious diseases, particularly for tuberculosis.

The authors have no financial conflicts of interest.