Mycobacterium tuberculosis Mannose-Capped Lipoarabinomannan Induces IL-10-Producing B Cells and Hinders CD4+Th1 Immunity.

The importance of Th1/interferon (IFN)-γ-mediated responses in mycobacterial infection has been well established. However, little is known about B cell-mediated immunity during Mycobacterium tuberculosis (Mtb) infection. Interleukin (IL)-10-producing B cells (B10 cells), a subset of B regulatory cells (Bregs), are implicated in modulating the immune response. Herein, we found that B10 cells were significantly increased in patients with tuberculosis. Furthermore, mannose-capped lipoarabinomannan (ManLAM), a major surface lipoglycan component from Mtb, induced a significant increase in B10 cells, which enriched in CD5+ B1a B cells. ManLAM induced IL-10 production mainly by activating MyD88/PI3K/AKT/Ap-1 and K63-linked ubiquitination of NF-κB essential modulator/nuclear factor kappa-light-chain-enhancer of activated B cells signaling pathways in B cells via Toll-like receptor 2. IL-10 production by ManLAM-treated B cells further inhibited CD4+ Th1 polarization, leading to increased susceptibility to mycobacterial infection compared with ManLAM-treated IL-10-/- B group. Thus, we report a new immunoregulation mechanism in which Mtb ManLAM-induced B10 cells negatively regulate host anti-TB cellular immunity.

Introduction

Tuberculosis (TB) is the leading cause of death due to infectious diseases worldwide. In 2016, the World Health Organization reported approximately 10.4 million new TB cases worldwide (World Health Organization, 2017). The emergence of multidrug-resistant TB and co-infection with the human immunodeficiency virus (HIV) make TB control even more urgent. An attenuated strain vaccine of Mycobacterium bovis, termed bacillus Calmette–Guérin (BCG), the only available used TB vaccine, still has limited protective efficacy against TB. Therefore, it is important to understand the bacterial immunoregulation mechanism and host immune defense mechanisms against mycobacteria.

Characterization of the immune response to Mycobacterium tuberculosis (Mtb) has largely focused on Th1 cell-mediated immunity, whereas B cells are often overlooked in anti-Mtb immunity. Recently, emerging evidence suggests that B cells may orchestrate the immune response against Mtb by interacting with other immune cells such as T cells (Achkar et al., 2015, Hoff et al., 2015, Kozakiewicz et al., 2013, Maglione et al., 2007). Regulatory B cells (Bregs), which produce interleukin (IL)-10 or transforming growth factor β, participate in the immunomodulation of immune responses. A subset of Bregs, IL-10-producing B cells (B10 cells), has been shown to prevent excessive inflammatory responses in autoimmune diseases (Mauri and Bosma, 2012, Yang et al., 2013). B10 cells also appear to negatively regulate cellular immune responses in infectious diseases caused by intracellular pathogens, including hepatitis B virus (Das et al., 2012), HIV-1 (Liu et al., 2014a, Liu et al., 2014b), and Listeria (Horikawa et al., 2013). However, the roles of B10 cell in the immune response to Mtb remain elusive.

Mannose-capped lipoarabinomannan (ManLAM) is a major cell wall lipoglycan and an important immunomodulatory component of mycobacteria (Mishra et al., 2011). Bacterial ManLAM can also be secreted and recognized by macrophages and dendritic cells (DCs) via pattern recognition receptors, including mannose receptor (MR), Toll-like receptor 2 (TLR2), DC-specific intercellular adhesion molecule-3-grabbing non-integrin (DC-SIGN), CD1d, sphingosine-1-phosphate receptor 1 (S1P1), Dectin-2, and CD44, and triggers several cell signaling pathways (Pan et al., 2014, Osanya et al., 2011, Geijtenbeek et al., 2003, Sun et al., 2016, Richmond et al., 2012, Yonekawa et al., 2014, Zajonc et al., 2006). ManLAM inhibits phagosome maturation in macrophages, DC maturation, and CD4+ T cell activation (Osanya et al., 2011, Fratti et al., 2003, Mahon et al., 2012). Anti-ManLAM antibody treatment and anti-ManLAM aptamer treatment decrease bacterial loads and dissemination, prolong survival, and lead to better disease outcomes in an animal model of TB (Pan et al., 2014, Hamasur et al., 2004).

We were interested in determining the interaction between ManLAM and B cells. In the present study, we first reported that ManLAM induced IL-10 production by B cells (B10 cells) both in vitro and in vivo predominantly through TLR2. Molecular mechanism analysis revealed that the binding of ManLAM to TLR2 activated MyD88 and its downstream AP1 and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) signaling to promote IL-10 production by B cells. ManLAM-induced B10 cells hindered Th1 response compared with ManLAM-IL-10−/− B cells, facilitating mycobacterium survival. We report a new immunoregulation mechanism in which Mtb ManLAM-induced B10 cells negatively regulate host anti-TB cellular immunity. Our findings will help to understand the interaction between B cells and Mtb ManLAM and highlight the ManLAM-mediated B10 cells' immunomodulatory functions.

Results

Peripheral B10 Cells Are Elevated in Patients with TB

To assess the roles of human B10 cells in TB disease, we determined the serum concentration of IL-10 and the frequency of B10 cells in patients with active pulmonary TB. As shown in Figure 1A, the serum IL-10 concentrations in patients with active TB (ATB) were much higher than those in healthy donors (161.2 ± 21.34 pg/mL versus 40.9 ± 6.6 pg/mL). Consistent with the elevated serum IL-10 level, the percentages of IL-10+CD19+ B cells in peripheral blood mononuclear cells from patients with TB were significantly increased compared with those from healthy donors (4.0% ± 0.3% versus 1.0% ± 0.7%; Figures 1B and 1C). These results indicated that increased levels of IL-10 and B10 cells in patients with TB might be associated with TB disease.

Figure 1

Elevated Levels of B10 Cells in Peripheral Blood of Patients with TB

(A) Elevated serum IL-10 level in patients with ATB. IL-10 was detected by ELISA. Data are represented as mean ± SD. Two-tailed, unpaired t test; ***p < 0.001.

(B and C) (B) Human B10 cells were determined by flow cytometry analysis. (C) Representative dot plots. Data are represented as mean ± SD. ***p < 0.001.

(D) Serum ManLAM levels in patients with ATB and healthy donors. MR was coated on the microplates, and then the serum samples were added on the microplates. After washing, the biotin-labeled single-stranded DNA aptamer T9 (400 nM) was added to detect serum ManLAM and the horseradish peroxidase-streptavidin conjugate was used for color development. The absorbance at 450 nm was determined. Data are represented as mean ± SD. ***p < 0.001.

Because ManLAM is a virulent factor of Mtb, we determined the serum ManLAM concentrations in patients with ATB. We observed significantly higher serum levels of ManLAM in 30 patients with ATB (156.1 ng/mL, 95% confidence interval [CI] = 95.5–216.6) compared with those of 30 healthy donors (8.8 ng/mL, 95% CI = 6.8–10.8, Figure 1D). These results indicated that Mtb ManLAM might be related to the increase in serum IL-10 level in patients with ATB.

ManLAM Induces Murine B10 Cells In Vitro and In Vivo

To elucidate the roles of B10 cells in the immune response to Mtb, we first determined whether Mtb and ManLAM induced IL-10 production by murine B cells in vitro. Mtb H37Rv/BCG ManLAM was purified from the bacteria (Figures S1A and S1B). It has been reported that Mycobacteria do not have endotoxin (Jackson, 2014) (known as lipopolysaccharide [LPS]). The ManLAM used in our experiments was not contaminated by endotoxin because endotoxin scavenger polymyxin B-treated-ManLAM did not significantly alter IL-10 production and the cytotoxicity against B cells compared with ManLAM treatment (Figures S1C and S1D). According to the previous report (Kishimoto, 1985), we used IL-2 to promote B cell proliferation in our experiments. Both IL-2 and IL-4 had been used to promote B cell proliferation (Kishimoto, 1985, Karray et al., 1988, Wagner et al., 1998). However, IL-4 had the potential ability to negatively control B cell differentiation (Bod et al., 2018) and could inhibit IL-10 production by B cells upon stimulation with IL-5/mCD40L-expressing fibroblast in a STAT6-dependent manner (Taitano et al., 2018). Therefore, we used IL-2 to promote B cell proliferation in the current study. As shown in Figure S1C, the addition of IL-2 alone did not significantly affect the IL-10 production by B cells compared with the medium control group.

Flow cytometry (FCM) analysis showed that iH37Rv, iBCG (heat-inactivated Mtb H37Rv and BCG, in which heat-inactivated bacterial proteins were denatured [Ferrier, 2017], but bacterial glycolipids maintained activities [Worakitkanchanakul et al., 2008]), and ManLAM purified from both Mtb H37Rv and BCG significantly induced CD19+ B cells into B10 cells compared with the IL-2 control groups (Figures 2A, 2B, and S1E). As shown in Figures 2B and S1E, iBCG induced more B10 cells than iBCGΔ2196 (a ManLAM mutant strain of BCG that lacks the mannose cap of ManLAM) (Sun et al., 2016). These results demonstrated that Mtb ManLAM induced B10 cells in vitro.

Figure 2

BCG/iH37Rv/ManLAM Induces B10 Cells, whereas BCGΔ2196 Induces Less B10 Cells In Vitro and In Vivo

(A and B) B10 cells were detected by flow cytometry (FCM). B cells were stimulated with iH37Rv (MOI, 1:2) or Rv ManLAM (10 ng/mL), iBCG (MOI, 1:2), iBCGΔ2196 (MOI, 1:2), or BCG ManLAM (10 ng/mL) for 12 hr. Data are represented as mean ± SD. ***p < 0.001, **p < 0.01.

(C) IL-10 production in culture supernatant was detected by ELISA. B cells were stimulated with Rv ManLAM, LPS (10 μg/mL) or iH37Rv for 12 hr. Data are represented as mean ± SD. **p < 0.01.

(D) Time course of IL-10 production in ManLAM-treated B cells. B10 cells were detected by FCM. Data are represented as mean ± SD. ***p < 0.001, **p < 0.01.

(E) IL-10 production by B cells stimulated with various concentrations of ManLAM for 12 hr. Data are represented as mean ± SD. ***p < 0.001, *p < 0.05.

(F) B10 cells enriched in CD5+ B1a B cells (CD1dhi/loCD5+ B cells) upon Rv ManLAM stimulation for 12 hr.

(G and H) B cells produced IL-10 upon stimulation with iH37Rv/iBCG/iBCGΔ2196/ManLAM in vivo. WT (C57BL/6J) mice were injected (intraperitoneally [i.p.]) with iH37Rv/iBCG/iBCGΔ2196 (108 CFU/100 μL/mouse) or ManLAM (100 ng/100 μL/mouse), and each group had six mice; B10 cells were analyzed by FCM. Data are represented as mean ± SD. ***p < 0.001, **p < 0.01.

(I) WT mice were infected (i.p.) with BCG/BCGΔ2196 (6 mice/group). Left panel: schematic diagram. Right panel: splenocytes were collected at day 20 post-infection. B10 cells were determined by FCM. Data are represented as mean ± SD. *p < 0.05.

(A–E) B cells were cultured in medium containing IL-2.

See also Figure S1.

The IL-10 levels produced by the ManLAM-treated B cells were further measured. As shown in Figure 2C, both ManLAM and iH37Rv significantly induced IL-10 levels approximately 6- to 7-fold higher than that of the IL-2 control group and approximately 1- to 2-fold higher than that of the LPS-treated group. Both B10 cell numbers and IL-10 concentrations reached peak level approximately at 12 hr upon stimulation with ManLAM, and then slowly decreased, but were still higher than those in the control group (Figures 2D and S1F). The highest IL-10 and B10 cell levels were induced upon stimulation with 10 ng/mL ManLAM (Figures 2E and S1G). As shown in Figure 2F, the frequencies of B10 cells in CD1dhiCD5+ B and CD1dloCD5+ B subsets were 17.1% and 12.4%, respectively, in ManLAM treatment group, which were much higher than the B10 frequencies in CD1dhi/loCD5- B cells. Therefore, we identified ManLAM-induced B10 cells enriched in CD5+ B1a B cells (CD1dhi/loCD5+ B cells). CD1dhiCD5+ B cell population contains several B cell subsets, including marginal zone B cells and B1 B cells (Yanaba et al., 2008), which can be rapidly recruited into the early adaptive immune responses in a T cell-independent manner. The function of CD1dloCD5+ B cells are unknown, but Ding et al. reported that some CD1dloCD5+ B cells can secrete IL-10 as well (Ding et al., 2011). In the LPS group, LPS-induced B10 cells enriched in CD1dhiCD5+ B cells (22.9%), which was consistent with the previous report (Matsushita and Tedder. 2011).

ManLAM-induced B10 cells were then detected in vivo. Mice were intraperitoneally injected with purified ManLAM (from Mtb H37Rv/BCG), iH37Rv/iBCG/iBCGΔ2196, at days 1 and 7. After the second injection, we observed increased B10 cells in peripheral blood, spleens, and lymph nodes from the mice subjected to the ManLAM, iH37Rv, and iBCG treatment, respectively (Figures 2G, 2H, S1H, and S1I), whereas iBCGΔ2196 induced less B10 cells than BCG (Figure 2H). The B10 cell percentage was also significantly higher in the live BCG infection group compared with the PBS and live BCGΔ2196 infection groups (Figures 2I and S1J). These results indicated that B10 cells were induced by Mtb and ManLAM both in vitro and in vivo (Figures 2A–2I and S1C–S1J).

ManLAM Induces IL-10 Production by B Cells Mainly via TLR2

It has been reported that ManLAM can be recognized by TLR2 and MR (Pan et al., 2014, Osanya et al., 2011). Both TLR2 and MR signals are involved in IL-10 production by macrophages and monocytes during Mtb infection (Sun et al., 2016, Wang et al., 2012). Therefore, we investigated whether TLR2 and MR were involved in ManLAM-induced IL-10 production in B cells.

In the pull-down and immunoblotting assay, both TLR2 and MR were identified as ManLAM-associated proteins from wild-type (WT) B cells (Figure 3A). ManLAM lost the ability to bind to TLR2 in TLR2−/− B cells and showed decreased binding to MR in MR-short hairpin RNA (shRNA)-treated B cells (Figures 3A, S2A, and S2B), which further supported the binding of ManLAM to both TLR2 and MR. In the competition assay, the binding of TLR2 to ManLAM was not inhibited by a soluble MR competitor, and soluble TLR2 competitor did not inhibit MR binding to ManLAM (Figure S2C). This finding indicated that the binding sites of ManLAM for MR and TLR2 were different.

Figure 3

ManLAM Induces IL-10 Production by B Cells Mainly via TLR2

(A) Binding of ManLAM to TLR2 and MR proteins was detected by immunoblotting. Cell lysates from WT B cells, TLR2−/− B cells, and MR shRNA-transfected B cells were incubated with Rv ManLAM-coupled magnetic beads.

(B) IL-10 production by ManLAM-treated B cells was not mainly via MR. WT/TLR2−/− B cells were transfected with MR shRNA for 48 hr and stimulated with ManLAM (10 ng/mL) for 12 hr. IL-10 production in cell supernatant was detected by ELISA. Data are represented as mean ± SD. ***p < 0.001; NS, not significant.

(C–E) IL-10 production by ManLAM-treated B cells was dominantly via TLR2. Splenic B cells were stimulated with ManLAM for 12 hr. (C) IL-10 in culture supernatant was detected by ELISA. (D) B10 cells were analyzed by flow cytometry. (E) Representative dot plots. Data are represented as mean ± SD. ***p < 0.001; NS, not significant.

(B–E) B cells were cultured in medium containing IL-2.

See also Figures S2 and S3.

Next, we assessed whether the binding of ManLAM to MR and TLR2 participated in IL-10 production by B cells. As shown in Figure 3B, MR knockdown in both WT and TLR2−/− B cells had little effect on IL-10 production compared with the scramble group after ManLAM stimulation. IL-10 production was markedly reduced in TLR2−/− B cells, but not in TLR4−/− B cells and WT B cells after ManLAM treatment (Figure 3C). These findings suggest that ManLAM-MR binding has only a minor effect on IL-10 production by B cells but ManLAM-TLR2 binding is involved in IL-10 production in ManLAM-treated B cells.

We found that ManLAM induced IL-10 production by TLR2−/− B cells was significantly decreased by approximately 60%–80% compared with the WT B cell control or TLR4−/− B cells (Figures 3C–3E). Blocking ManLAM binding to TLR2 by anti-TLR2 antibody significantly decreased IL-10 production in WT B cells (Figure S2D). Moreover, the TLR2−/− B cells were actually capable of producing IL-10 at similar levels to WT B cells when the cells were stimulated with LPS and anti-IgM+ CD40 (Figure S2E). There was no significant difference in IL-10 production between TLR4−/− B and WT B cells upon ManLAM stimulation (Figures 3C–3E). These results strongly confirmed that ManLAM induced IL-10 production by B cells predominantly via TLR2.

To assess the effects of ManLAM as well as other TLR ligands on IL-10 production by B cells, ManLAM, TLR1/2 agonist Pam3CSK4, TLR6/2 agonist FSL-1 (synthetic diacylated lipoprotein), TLR4 agonist LPS, and TLR5 agonist FLA-ST (flagellin from Salmonella typhimurium) were used. As shown in Figure S2F, the B10 levels were increased in the ManLAM, TLR1/2 agonist, TLR6/2 agonist, and TLR4 agonist groups, whereas the B10 cell level did not change after treatment with TLR5 agonist. ManLAM induced IL-10 through TLR2, but this was not specific for ManLAM. The ligands for TLR2 could also induce IL-10 production by B cells.

TLR2 is working in heterodimers with TLR1 or TLR6. To further identify the roles of TLR2 heterodimers, we used TLR1 and TLR6 antagonists to assess the mechanism of IL-10 production by ManLAM-treated B cells. As shown in Figure S2G, either TLR1/2 or TLR2/6 heterodimers were involved in IL-10 production of ManLAM-treated B cells because ManLAM induced much lower levels of IL-10 in the presence of TLR1 or TLR6 antagonists. We used TLR1-shRNA and TLR6-shRNA to silence TLR1 and TLR6 expression on B cells (Figures S2H and S2I). Consistent with the above-mentioned results, the IL-10 production by TLR1 (or TLR6)-silenced B cells was greatly reduced upon stimulation with ManLAM (Figure S2J). These data suggest that ManLAM induces IL-10 production by B cells via TLR1/2 or TLR2/6 heterodimers.

It has been reported that ManLAM can be recognized by other receptors, including DC-SIGN, CD1d, S1P1, Dectin-2, and CD44, and triggers several cell signaling pathways (Geijtenbeek et al., 2003, Sun et al., 2016, Richmond et al., 2012, Yonekawa et al., 2014, Zajonc et al., 2006). However, DC-SIGN and Dectin-2 are undetectable in B cells (Ariizumi et al., 2000, Soilleux et al., 2002). The binding of ManLAM to CD44 is involved in the downregulation of IL-10 production (Sun et al., 2016). Therefore, we investigated whether CD1d and S1P1 were involved in ManLAM-induced IL-10 production by B cells. CD1d monoclonal antibody (mAb) blocked ManLAM-CD1d binding in a dose-dependent manner (Figures S3A). However, the addition of CD1d mAb did not reduce IL-10 production by ManLAM-treated B cells (Figures S3B, S3C, and S3E). The results indicated that ManLAM-CD1d binding was not involved in triggering IL-10 production by B cells. The S1P1 inhibitor W146 slightly reduced ManLAM-induced IL-10 production by B cells (approximately 20%, Figures S3D and S3E). Taken together, our findings demonstrated that ManLAM induced IL-10 production of B cells predominantly through TLR2.

ManLAM Induces IL-10 Production by B Cells via TLR2/MyD88/Ap-1 and K63-Linked Ubiquitination of NEMO/NF-κB Signaling Pathways

The signaling pathways in B cells that lead to IL-10 production have thus far remained elusive (Mauri and Bosma, 2012). We chose several key adapter proteins and transcription factors downstream of TLR2 to explore the signaling pathway that might contribute to IL-10 production in ManLAM-treated B cells (Du et al., 2016, Gaddis et al., 2013, Hu et al., 2006, Liu et al., 2014a, Liu et al., 2014b, Saraiva and O'Garra, 2010).

As shown in Figures 4A and S4A, IL-10 production was significantly reduced by approximately 60%–80% in MyD88−/− B cells, demonstrating that MyD88 was involved in IL-10 production in ManLAM-treated B cells. Specific inhibitors were applied to inhibit the activation of JNK (c-Jun N-terminal kinase), PI3K (phosphoinositide-3 kinase), MEK1 (mitogen-activated extracellular signal-regulated kinase 1), NF-κB, and AP1 (activator protein 1) in ManLAM-treated B cells. As shown in Figures 4B, S4B, and S4C, ManLAM-induced IL-10 production was significantly inhibited by inhibitors of the upstream signal transducers JNK, PI3K, and MEK1, as well as AP-1 and NF-κB. The two most potent inhibitors were PI3K (half maximal inhibitory concentration [IC50] = 6.029 μM) inhibitor and AP1 inhibitor (IC50 = 11.67 μM, Figures 4B, S4B, and S4C).

Figure 4

ManLAM-TLR2 Binding Stimulates IL-10 Production in B Cells by Activating MyD88/PI3K/AKT/Ap-1 and NF-κB Signaling Pathways

(A) B10 cells were analyzed by flow cytometry. WT and MyD88−/− B cells were stimulated with Rv ManLAM (10 ng/mL) for 12 hr. Data are represented as mean ± SD. ***p < 0.001.

(B) IL-10 production in ManLAM-treated B cells was inhibited by inhibitors of JNK, PI3K, MEK1, AP-1, and NF-κB. B cells were pretreated with the inhibitors for 1 hr. The cells were stimulated with ManLAM. IL-10 production in the supernatant was determined by ELISA.

(C) PI3K, AKT, AP1, and NF-κB activation in WT B cells upon ManLAM stimulation were analyzed at multiple time points by immunoblotting. Left panel: pooled data. Right panel: representative blots. Data are represented as mean ± SD versus 0 min at indicated time. ***p < 0.001, **p < 0.01,*p < 0.05.

(D) PI3K, AKT, NF-κB, and AP1 activation were determined in WT B cells and TLR2−/− B cells (or MyD88−/− B cells) upon stimulation with ManLAM for 1 hr by immunoblotting. Data are represented as mean ± SD. ***p < 0.001, **p < 0.01, *p < 0.05; NS, not significant.

(A–D) B cells were cultured in medium containing IL-2.

See also Figure S4.

Immunoblotting analysis revealed that PI3K and its downstream target AKT were highly phosphorylated at 30 min after ManLAM stimulation (Figure 4C). We found that cytoplasmic AP1, a downstream effector of PI3K/AKT, was decreased and nuclear AP1 was greatly increased 60 min after ManLAM stimulation (Figure 4C). The peak of NF-κB activation occurred between 60 and 120 min after ManLAM stimulation (Figure 4C). Both TLR2−/− B cells and MyD88−/− B cells showed a blunted increase in PI3K and AKT phosphorylation, nuclear AP1, and NF-κB activation upon ManLAM stimulation (Figures 4D and S4D).

As shown in Figures S4E and S4F, we did not observe ManLAM signaling and IL-10 production when (1) TLR2−/− B cells were treated with ManLAM/Pam3CSK4 and (2) WT B cells were stimulated with ManLAM/Pam3CSK4 in the presence of MyD88 inhibitor (ST2825). Therefore ManLAM signaling occurred predominantly through the TLR2-MyD88 pathway, and no compensatory pathway was detected. As shown in Figure S4G, knockdown of MR in TLR2−/− B cells had fairly limited effects on ManLAM signaling, further confirming that MR has only a minor effect on B10 cells. Taken together, these results suggested that ManLAM induced IL-10 production by B cells mainly via the TLR2/MyD88/PI3K/AKT/AP-1 and NF-κB signaling pathways.

It has been reported that the Lys63 (K63)-linked ubiquitination of NF-κB essential modulator ([NEMO] also known as IKK-γ) is involved in the IKK complex recruitment for NF-κB activation, whereas Lys48 (K48)-linked ubiquitination is associated with proteasomal degradation and inhibits NF-κB activation (Lopez-Castejon and Edelmann, 2016). As shown in Figure 5A, K63 polyubiquitylation, but not K48 polyubiquitylation, of NEMO was observed in ManLAM-treated B cells.

Figure 5

K63-Linked Ubiquitination of NEMO and USP40 Are Involved in TLR2-Dependent IL-10 Production in ManLAM-Treated B Cells

(A) ManLAM induces K63-linked ubiquitination of NEMO in B cells. B cells were transfected with the plasmids encoding NEMO, wild-type (WT) ubiquitin (H-Ub), ubiquitin mutant H-Ub (K48), and H-Ub (K63). After stimulation with Rv ManLAM, the supernatant from cell lysates was subjected to immunoprecipitation and immunoblotting (upper panel). The expression levels of NEMO were examined by immunoblotting (lower panel).

(B) The mRNA expression levels of NF-κB signal and ubiquitination-related proteins in ManLAM-treated CD1dhiCD5+ B cells were determined by transcriptome analysis. Left panel: differentially expressed mRNAs were presented as a heatmap. The threshold established for upregulated and downregulated genes was a log2 fold change ≥ 1.0. Red, upregulation; green, downregulation. Right panel: mRNA expression levels. FPKM (fragments per kilobase million) value indicated the relative expression of genes.

(C and D) ManLAM induces B10 cells via TLR2 and USP40. B cells were transfected with USP40-shRNA and stimulated with ManLAM. (C) B10 cells and (D) IL-10 production were detected by flow cytometry and ELISA. Data are represented as mean ± SD. ***p < 0.001, **p < 0.01, *p < 0.05; NS, not significant.

(A–D), B cells were cultured in medium containing IL-2.

See also Figure S5.

Because the highest frequency of ManLAM-induced B10 cells was found in CD1dhiCD5+ B subset, we sorted CD1dhiCD5+ B cells from the spleens of both WT and TLR2−/− mice and analyzed global gene expression changes by RNA sequencing after ManLAM treatment (Figures 5B, S5A, and S5B). The gene ontology analysis (Figure S5A) identified differentially expressed genes involved including 24 biological processes, 15 cellular components, and 10 molecular functions (false discovery rate, FDR ≤ 0.01) in the WT group compared with the TLR2−/− group. Also, biological processes include the response to stimulus and immune system process (Figure S5A). Chemokine signaling pathway and NF-κB signaling pathway were enriched in Kyoto Encyclopedia of Genes and Genomes pathway (FDR ≤ 0.01, Figure S5B). From the above-mentioned analysis, we found that ManLAM-TLR2 interaction markedly caused the changes of chemokine and NF-κB pathway in CD1dhiCD5+ B cells.

The results of hierarchical cluster analysis of gene expression patterns showed that that the mRNA expression of USP40 was upregulated in ManLAM-treated WT CD1dhiCD5+ B cells than the unstimulated control (Figure 5B). USP40 is a member of the ubiquitin proteasome system. The system is responsible for protein ubiquitination and proteasomal degradation. Our results from quantitative reverse-transcriptase PCR and immunoblotting analysis further confirmed that the mRNA and protein expression levels of USP40 were also greatly upregulated in the WT B cells compared with the TLR2−/− B cell groups when these cells were stimulated with ManLAM/Pam3CSK4 (Figure S5C). We constructed USP40-shRNA to silence the USP40 expression in B cells (Figures S5D and S5E). Knockdown of USP40 significantly decreased ManLAM/Pam3CSK4-induced IL-10 production by WT B cells, whereas IL-10 production by USP40-silenced TLR2−/− B cells was not reduced compared with scramble RNA-treated TLR2−/− B cells upon stimulation with ManLAM (Figures 5C, 5D, and S5F). Our data suggest that ManLAM induces IL-10 production in B cells via TLR2 and USP40. ManLAM-TLR2 induces USP40 increase, which positively regulated the IL-10 production in B cells (Figures 5B–5D and S5C–S5F).

ManLAM-Induced B10 Cells Inhibit Th1 Polarization, Promote Th2 Polarization, and Have No Effect on Th17 Cells In Vitro and In Vivo

Next, we investigated the effects of ManLAM-induced B10 cells on CD4+ T cell polarization. CD4+ T cells were co-cultured with ManLAM-treated WT/IL-10−/− B cells. The production of interferon (IFN)-γ, IL-4, and IL-17A of CD4+ T cells was then determined (Figure 6A). Compared with the ManLAM-treated IL-10−/− B cells group, IFN-γ-producing CD4+ T cells were significantly decreased when the cells were co-cultured with ManLAM-treated WT B cells, whereas IL-4-producing CD4+ T cells were markedly increased in the ManLAM-treated WT B cell group (Figures 6B, 6C, S6A, and S6B). There were no significant differences in IL-17A-producing CD4+ T cells among the groups (Figures 6D and S6C). Similar results were obtained in transwell assays (Figures 6B–6D and S6A–S6C). The upper chambers in transwell plates with a pore size of 0.4 μm prevented direct cell-cell contact but allowed the exchange of soluble factors (Noack et al., 2016). These results indicate that IL-10 secreted by ManLAM-treated B cells inhibits Th1 polarization, but promotes Th2 polarization, and has no effect on Th17 cells in vitro.

Figure 6

ManLAM-Induced B10 Cells Inhibit Th1 Polarization, Promote Th2 Polarization, and Have No Effect on Th17 In Vitro

Splenic B cells were isolated from WT/IL-10−/− mice and stimulated with Rv ManLAM (10 ng/mL) for 12 hr. Then ManLAM was removed by washing. CD4+ T cells were stimulated with plate-coated anti-CD3 antibody (5 μg/mL) and soluble anti-CD28 antibody (2 μg/mL) and then cultured with the ManLAM-treated B cells for 72 hr.

(A) Schematic diagram.

(B–D) (B) IFN-γ, (C) IL-4, and (D) IL-17A production by CD4+ T cells were analyzed by flow cytometry. (B–D) The cells were cultured in medium containing IL-2. Data are represented as mean ± SD. ***p < 0.001, *p < 0.05.

See also Figure S6.

To assess the regulation of CD4+ T cell polarization in vivo by ManLAM-induced B10 cells, we used the iH37Rv-primed mouse model. ManLAM-treated WT or IL-10−/− B cells were adoptively transferred into B cell-deficient μMT mice. The recipient mice were primed with iH37Rv. The polarization of splenic CD4+ T cells was evaluated after 7 days of adoptive transfer (Figure 7A). Consistent with the in vitro results, the transferred ManLAM-treated WT B cells induced a significant decrease in splenic IFN-γ-producing CD4+ T cells, but an increase in IL-4-producing CD4+ T cells compared with the ManLAM-treated IL-10−/− B cell group (Figures 7B and 7C). Adoptive transfer of ManLAM-treated WT B cells did not change the percentage of IL-17A-producing CD4+ T cells (Figures 7B and 7C). These findings strongly suggest that ManLAM-induced B10 cells inhibit Th1 polarization, promote Th2 polarization, and have no effect on Th17 cells in vitro and in vivo.

Figure 7

ManLAM-Induced B10 Cells Inhibit Th1 Polarization, Promote Th2 Polarization, and Have No Effect on Th17 In Vivo

(A–C) ManLAM-treated WT/IL-10−/− B cells were adoptively transferred into μMT mice. The mice were primed by iH37Rv. (A) Schematic diagram. (B) IFN-γ, IL-4, and IL-17A production by splenic CD4+ T cells were determined by flow cytometry. Data were shown as means ± SD (n = 6). ***p < 0.001, **p < 0.01; NS, not significant. (C) Representative dot plots.

ManLAM-Induced B10 Cells Increase Susceptibility to Mtb Infection in Mice

Next we used WT BCG infection model to explore the role of ManLAM-induced B10. Rag2−/− mice were adoptively transferred with ManLAM-treated WT/IL-10−/− B cells plus CD4+ T cells and infected (intravenously [i.v.]) with BCG (Figure 8A). Usually a low-dose aerosol of Mtb (approximately 50–100 colony-forming unit [CFU] in mice) causes chronic infection (Upton et al., 2015), whereas a high dose (>102 CFU, aerosol/i.v.) causes acute infection (Mayer-Barber et al., 2014, Aggarwal et al., 2017, Das et al., 2013). Because mice that inhaled 1 × 102–1×104 or i.v. administered 1 × 105–2 × 106 CFU of Mtb (Mtb H37Rv, H37Ra, or BCG) showed similar pathological changes in acute infection model (Mayer-Barber et al., 2014, Aggarwal et al., 2017, Das et al., 2013, Stüve et al., 2018), 1 × 105 CFU of Mtb/mouse (i.v. infection) was used in our mouse model. After 20 days of infection, transferred CD19+ B cells and CD3+CD4+ T cells were observed in the Rag2−/− mice (Figure S7A). Serum IL-10 and B10 cells were increased in the mice transferred with ManLAM-treated WT B cells compared with other groups (Figures 8B, 8C, and S7B). These results demonstrated that ManLAM-treated B cells secreted IL-10 and contributed to increased serum levels of IL-10.

Figure 8

ManLAM-Induced B10 Cells Increase Susceptibility to Mtb Infection in Mice

ManLAM-treated WT/IL-10−/− B cells were adoptively transferred into Rag2−/− mice. The mice were infected with BCG.

(A) Schematic diagram.

(B) After 20 days of infection, serum IL-10 level was determined by ELISA. Data are represented as mean ± SD. ***p < 0.001, **p < 0.01.

(C–G) Splenic B10 cells and cytokines produced by splenic CD4+ T cells were determined by flow cytometry. (C) B10 cells; (D) IFN-γ, (E) IL-4, (F) IL-17A, and (G) IL-10 production by CD4+ T cells. Data are represented as mean ± SD. ***p < 0.001, **p < 0.01.

(H) CFU assay in spleens. Data are represented as mean ± SD. ***p < 0.001, **p < 0.01.

See also Figures S7 and S8.

Consistent with the above-mentioned results, transfer of ManLAM-treated WT B cells reduced IFN-γ production, but promoted IL-4 production by splenic CD4+ T cells compared with the WT B cell group and ManLAM-treated IL-10−/− B cell group, whereas IL-17A production by CD4+ T cells did not significantly change among all groups during BCG infection (Figures 8D–8F and S7C). Moreover, the increased level of IL-10-producing CD4+ T cells was observed in ManLAM-treated WT B cell group (Figures 8G and S7C). BCG CFUs in spleens were significantly increased in the ManLAM-treated WT B group compared with WT B cell group and ManLAM-treated IL-10−/− B cell group (Figure 8H). These results clearly demonstrated that ManLAM-induced B10 cells increased susceptibility to BCG infection and facilitated the bacterial survival in mice.

Finally, we assessed the roles of the ManLAM-induced B10 cells during virulent Mtb H37Rv infection. Rag2−/− mice were adoptively transferred with ManLAM-treated WT/IL-10−/− B cells plus CD4+ T cells and infected (i.v.) with virulent Mtb H37Rv (Figure S8A). After 20 days of infection, the level of serum IL-10 was greatly elevated (Figure S8B), and Mtb H37Rv CFUs in both lung and spleen tissues were also higher in the ManLAM-treated WT B group compared with the ManLAM-treated IL10−/− B group (Figures S8C and S8D upper panel). In PBS-transferred group, Rag2−/− mice had no mature T and B cells, so the highest Mtb H37Rv CFUs were found in this PBS group (Figures S8C and S8D upper panel). Compared with the ManLAM-treated IL-10−/− B group, histopathological analysis of alveolar tissue from the ManLAM-treated WT B and PBS groups showed larger amount of lymphocyte infiltration (Figure S8D lower panel). Histological examination of splenic tissues demonstrated that white pulp and red pulp in spleen tissues from the ManLAM-treated WT B and PBS groups were more completely disorganized, with many basophilic granules and necrotic cellular debris, compared with the ManLAM-treated IL-10−/− B group (Figure S8D lower panel). These findings confirmed that ManLAM-induced B10 cells facilitated mycobacterial survival in the mice.

Discussion

Some evidence has shown that B cells favor anti-TB immunity and prolong mouse survival (Kozakiewicz et al., 2013, Maglione et al., 2007). Distinct B cell subsets might play different roles in anti-TB immunity. B10 cells, a subtype of Bregs, have been rarely reported in Mtb infection. It has been reported that TLR2/TLR4 ligands from Freund's adjuvant induce B10 cells (Lampropoulou et al., 2008). Freund's adjuvant from BCG components may contain ManLAM. We demonstrated that ManLAM and Mtb themselves induced B10 cells in a TLR2/MyD88-dependent manner, which was consistent with the previous study (Lampropoulou et al., 2008).

ManLAM has been reported to be released from the metabolically active or degrading bacterial cells and shed into body fluid (like sputum, blood, or urine) during TB infection, which appears to be present only in people with ATB disease (Paris et al., 2017). Based on this character, ManLAM has been used as a target for TB diagnosis (Pandie et al., 2016, Gupta-Wright et al., 2018). Therefore, we could measure serum ManLAM concentrations in patients with ATB (Figure 1D), although the enzyme-linked oligonucleotide assay used here could not distinguish between ManLAM and whole bacteria.

Several studies reported that IL-10 was also produced by human CD4+ T cells, DCs, and macrophages in patients with ATB (Kumar et al., 2015, Redford et al., 2011). In mouse TB models, IL-10 production was mainly from monocytes in the mouse lung after Mtb infection (Moreira-Teixeira et al., 2017). A small proportion cells (30%–50% in total IL-10+ cells), including DCs, macrophages, T cells, and B cells, produced IL-10 (Moreira-Teixeira et al., 2017). Here, we identified that increased B10 cells played an important role during Mtb infection by hindering Th1 immunity. A comparison of the functions among various types of IL-10+ cells in TB infection models may need to be further investigated.

The frequencies of various types of IL-10+ cells change dynamically during Mtb infection. Previous studies from our group and those of others have shown about 5%–40% IL-10+ DCs (in total DCs) and 10%–30% IL-10+ macrophages (in total macrophages) induced by Mtb in mice (Moreira-Teixeira et al., 2017, Sun et al., 2016). Here, we measured about 4.0% ± 0.3% B10 cells in patients with ATB and 6%–8% B10 in Mtb infected mice (Figures 1C and 2G–2I). It has been reported that the peak level of proportion of IL-10+ DCs/macrophages reached within 28 days of Mtb infection and then gradually decreased (Moreira-Teixeira et al., 2017). The proportion of B10 cells was about 10% at day 7 post-infection, sharply decreased at day 14, and then increased 2-fold during chronic Mtb infection (days 21–57) (Moreira-Teixeira et al., 2017). The B10 level (10%–12% in total IL-10+ cells) was similar to the level of IL-10+ DCs and macrophages at day 57 post-infection (Moreira-Teixeira et al., 2017). We speculate that levels of different IL-10-producing cells might be changed with the progress of infection. The effect of in vivo B10 cell depletion in infected WT mice will need to be detected in future studies to assess the dynamic effects of B10 and other IL-10+ cells during Mtb infection.

We found that IL-10 expression by B cells reached a peak at 12 hr and then decreased after ManLAM stimulation (Figures 2D and S1F). The reasons might be the rapid induction and degradation of IL-10 mRNA in the activation of TLR2 signaling triggered by microbial molecules (Gabrysova et al., 2014, Eixeira-Coelho et al., 2014). We also found that IL-10 production in B cells first increased (1–10 ng/mL of ManLAM) and then decreased as the concentrations further increased (>10 ng/mL of ManLAM) (Figures 2E and S1G). Unlike protein antigens, the structure of the aliphatic and carbohydrate chains are diverse and complex as the concentrations of these molecules increased (Holderness et al., 2011, Jagger et al., 2002). Therefore, the increased concentration of ManLAM would not always induce an enhancement of cytokine production by immune cells. Moreover, ManLAM can be recognized by CD44 on the macrophages, which may downregulate IL-10 production (Sun et al., 2016). Whether ManLAM-CD44 interaction in B cells and ManLAM-TLR2 interaction in macrophages have effects on IL-10 production needs to be further investigated.

As shown in the left panel of Figure 2E, at higher concentrations of ManLAM (>10 ng/mL), the IL-10 level tended to decrease but was still significantly higher than that in the control group. These results indicate that a wide range of ManLAM concentrations (from 0 to 10 μg/mL) induced B cells to produce IL-10, contributing to elevate total serum IL-10 level in patients with ATB. As shown in Figure 1D, the average concentration of ManLAM in 30 patients with ATB was 156.1 ng/mL (95% CI = 95.5–216.6 ng/mL). The above range of ManLAM concentrations in patients with ATB could induce B cells to produce IL-10, which might have biological function in patients with ATB.

As shown in Figures 2G and 2H, the percentage of splenic B10 cells from non-immunized mice (PBS group) was approximately 2%, which was consistent with a previous report (Matsushita and Tedder, 2011). Our present data showed that approximately 6%–8% of B10 cells were induced by immunization with ManLAM/iH37Rv/iBCG in vivo, in which the frequency of B10 cells was similar to that in LPS-stimulated mice (Maseda et al., 2012). To the best of our knowledge, our study is the first report to describe the numbers of B10 cells induced by ManLAM/iH37Rv/iBCG in mice (Figures 2G and 2H).

TLR ligands binding various TLRs can induce IL-10 production by human and murine B cells (Mauri and Bosma, 2012, Mizoguchi and Bhan, 2006). Here we reported that ManLAM induced B10 cells via TLR2-mediated signaling. TLR2 has been implicated as a major signaling receptor that binds to lipoarabinomannan (Osanya et al., 2011). MR knockdown could not cause a significant reduction of IL-10 in TLR2−/− B cells treated with ManLAM (Figure 3B), and knockdown of MR in TLR2−/− B cells had fairly limited effects on ManLAM signaling (Figure S4G). These results indicated that ManLAM-MR binding only had a minor effect on triggering IL-10 production by B cells. As shown in Figures 3C–3E, recognition by TLR2 was the predominant mechanism of IL-10 production by ManLAM-treated B cells.

K63-linked ubiquitination of NEMO was observed in the ManLAM-treated B cells (Figure 5A). NEMO ubiquitination-mediated NF-κB activation regulates many viral and bacterial infections (Ashida et al., 2010, Fang et al., 2017). It has been reported that Mtb PtpA, a secreted tyrosine phosphatase, directly blocks host TAB3 from binding K63-linked ubiquitin chains, suppresses NF-κB activation, and contributes to immune evasion (Wang et al., 2015). In the current study, we demonstrated that ManLAM-induced IL-10 production by B cells was associated with the promotion of NEMO K63 ubiquitination and NF-κB activation. Consistent with our study, other groups have also reported that ManLAM induces IL-10 and COX2 expression by activating TLR2/NF-κB signaling in macrophages and DCs (Kumar et al., 2013, Rajaram et al., 2010, Vergne et al., 2014).

B10 cells have been implicated in the Th1 to Th2 shift during Leishmania and Salmonella typhimurium infection in mice (Neves et al., 2010, Ronet et al., 2010). It is widely accepted that the dominant protective response in TB is a Th1-type response. Th1 cytokine IFN-γ is a cytokine that plays a critical role in resistance to Mtb infection (Fenton et al., 1997), whereas Th2 cytokine IL-4 facilitates Mtb escape (Rook et al., 2004). In our study, the results from the μMT mice primed by iH37Rv and Rag2−/− mice infected with BCG experiments showed that the IL-4 production in CD4+T cells were increased (Figures 7B, 7C, 8E, and S7C). Greatly increased Th1 development was demonstrated in co-culture and transwell experiments using IL-10−/− B cells in vitro (Figure 6B), and in vivo adoptive transfer experiments (Figures 7B, 7C, 8D, and S7C). These results indicated that ManLAM hindered Th1 development/function during Mtb infection via IL-10 from B10 and possibly from Th2 cells. Consistently, Schierloh et al. reported that human B cells isolated from the pleural fluid of patients with TB produced IL-10 upon stimulation with γ-irradiated Mt and that these B cells diminished in vitro Mtb-induced IFN-γ production by T cells and natural killer cells in an IL-10-dependent manner (Schierloh et al., 2014).

It has been reported that IL-10 from B cells downregulates T-bet activation and IFN-γ production in Th cells (Flores-Borja et al., 2013). We also confirmed that IL-10 produced by ManLAM-treated B cells inhibited the development of Th1 IFN-γ+ CD4+ T cells, but promoted the development Th2 IL-4+CD4+T cells (Figures 7 and 8). IFN-γ and IL-4 are signature cytokines for Th1 and Th2 cells, respectively, so we measured these cytokines. The effects of ManLAM-induced B10 cells on the expression of Th1/2-related transcription factors (such as T-bet and GATA3) and other Th1-type cytokines (such as IL-12 and tumor necrosis factor-α) need to be clarified in the future.

In the present study, we found ManLAM-TLR2 induced the IL-10 production in B cells due in part to intracellular increased USP40. We found that USP40, a member of the ubiquitin proteasome system, was involved in the upregulation of IL-10 production by ManLAM-treated B cells (Figures 5B–5D and S5C–S5F). However, molecular interactions with USP40, and the process of USP40-mediated upregulation of IL-10 production in ManLAM-treated B cells, require elucidation in further studies. Second, several studies have reported that NF-κB is an important mediator involved in the production of both anti- and pro-inflammatory cytokines in B cells (Lee et al., 2017, Walsh et al., 2015). Our study showed that ManLAM activated NF-κB pathway and induced anti-inflammatory cytokine IL-10 in B cells. These indicate that other inflammatory cytokines might be induced by ManLAM-treated B cells via NF-κB signaling.

Recently, Bénard et al. reported that type I IFN produced by Mtb-stimulated B cells favors macrophage polarization toward a regulatory/anti-inflammatory phenotype during Mtb infection (Bénard et al., 2018). In addition, they also found a modest increase of IL-10 mRNA in B cells isolated from the lungs of Mtb-infected WT mice, which is consistent with our results. In our current study, we demonstrate that ManLAM induces IL-10 production of B cells, which hinders the Th1 response. We hypothesize that Mtb may negatively regulate immune response via B10 cells. Whether the B10 cells induced by ManLAM are polyfunctional for the production of type I IFN should be investigated in the future.

B cells are the dominant antigen-presenting cells that activate naive CD4+ T cells upon immunization with a viral antigen (Hong et al., 2018). Besides antigen presentation, B cells also have multifunction, including antibody production and cytokine secretion, and have effects on the T cell activation during Mtb infection (Chan et al., 2014). Thus, when WT B cells were adoptively transferred into μMT mice, the transferred WT B cells promoted Th1 cell activation upon stimulation with iH37Rv (Figures 7B and 8D). More importantly, peak levels of IFNγ+CD4+ T cells were detected when ManLAM-treated IL-10−/− B cells were adoptively transferred (Figures 7B and 8D). The levels appear to be significantly higher than those observed when WT B cells are transferred. IL-10−/− B cells do not produce IL-10 for suppressing Th1 polarization, so these IL-10−/− B cells induce higher level of IFNγ+CD4+ T cells compared with the WT B cells group. These results clearly demonstrate that ManLAM hinders Th1 development/function during Mtb infection via IL-10 from B cells. Moreover, ManLAM-treated B cells might also be partially involved in anti-TB immune response like antigen presentation.

Collectively, we report a new immunoregulation mechanism in which Mtb ManLAM induced-B10 cells negatively regulate host anti-TB cellular immunity. Our data showed that Mtb ManLAM induced B10 cells with CD5+ B1a subset. Molecular mechanism analysis reveals that the interaction of ManLAM and TLR2 induces IL-10 production by B cells mainly via the TLR2/MyD88/PI3K/AKT/AP-1 and NEMO K63 ubiquitination NF-κB signaling pathways, and increased USP40. ManLAM-induced B10 cells hinder Th1 responses, promote Th2 polarization, and have no effect on Th17 cells. ManLAM-induced B10 cells increase susceptibility to Mtb infection in mice. Our findings will help to understand the interaction between B cells and Mtb ManLAM and highlight the ManLAM-mediated B cells' immunomodulatory functions.

Limitations of the Study

The overall cytokine expressions, including the production of type I and type II IFNs, as well as multiple functions (like antigen presentation) of ManLAM-treated B10 cells have not been detected and are worth investigating in the future.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.

Data and Software Availability

RNA-seq data reported in this study has been submitted to the NCBI SRA: PRJNA505400 (https://submit.ncbi.nlm.nih.gov/subs/sra/)