B‐cells with a FasL expressing regulatory phenotype are induced following successful anti‐tuberculosis treatment

Abstract Introduction Studies show that B‐cells, in addition to producing antibodies and antigen‐presentation, are able to produce cytokines as well. These include regulatory cytokines such as IL‐10 by regulatory B‐cells. Furthermore, a rare regulatory subset of B‐cells have the potential to express FasL, which is a death‐inducing ligand. This subset of B‐cells have a positive role during autoimmune disease, but has not yet been studied during tuberculosis. These FasL‐expressing B‐cells are induced by bacterial LPS and CpG, thus we hypothesized that this phenotype might be induced during tuberculosis as well. Methods B‐cells from participants with TB (at diagnosis and during treatment) and controls were collected, and analyzed by means of real‐time PCR and flow cytometry. In addition to this, BAL was collected from TB participants as well and analyzed by means of MAGPix (multi‐cytokine) technology. Results Gene expression analysis show that FASL transcript levels increase by the end of treatment. Similarly, phenotypic analysis show that there is a higher frequency of FasL‐expressing B‐cells by the end of treatment. Conclusion Collectively, these results indicate that these FasL‐expressing B‐cells are being induced during anti‐TB treatment, and thus may play a positive role. Further studies are required to elucidate this.


Introduction
B lymphocytes (B-cells) are predominantly recognized for their role during humoral immunity, whereby these cells produce antibodies to assist in fighting infections. However, B-cells are able to produce cytokine as well, and the type of cytokines produced is dependent on the microenvironment [1]. Effector type 1 B-cells (Be-1) produce Th1 type cytokines such as interferon gamma (IFN-g) and tumor necrosis factor alpha (TNF-a), Be-2 cells produce Th2 type cytokines while regulatory B-cells produce IL-10 [2]. Cytokine-producing B-cells play important roles in the regulation of immunity and have been extensively studied in the context of autoimmune disease [2,3]. B-cell culture studies have shown that there is a balance in the proinflammatory and regulatory responses of B-cells [4][5][6]. In the context of autoimmune disease, and more specifically multiple sclerosis (MS), this balance is skewed toward a more proinflammatory response due to a reduction in the production of IL-10 [3]. The balance is however restored upon treatment. An additional regulatory function utilized by B-cells is the expression of Fas-Ligand (FasL) upon activation [7,8]. This subset of B-cells is rare and has been studied in the context of autoimmune disease as well as helminth infections [9]. These regulatory ''killer'' B-cells are CD19 þ IgM þ and induce apoptosis of CD4 T-cells, and thus have a positive effect during autoimmune disease. The B-cells induce FasL-mediated apoptosis of autoreactive T-cells, thereby regulating/reducing the inflammation. Klinker and coworkers (2013) showed that this regulatory subset of B-cells are induced by interleukin(IL)-5, and that there is an increase in the expression of IL-5 receptor a (IL-5Ra) [9]. Lymphocytes may undergo exhaustion due to persistent infections and prolonged activation. This occurs as the cells fail to respond to antigenic challenge, and has increased expression of markers of exhaustion such as programmed cell death protein (PD-1). The role of these ''killer'' B-cells during tuberculosis has not yet been studied, but it was shown that this phenotype is induced in B-cells upon stimulation with TLR9 agonist CpG [10]. This suggests that the same regulatory phenotype could be induced in the context of TB, as B-cell expression of TLR9 is triggered by mycobacterial CpG motifs [11][12][13][14]. Thus, we hypothesized that FasL expression on B-cells is induced during Mycobacterium tuberculosis infection. We therefore aimed to evaluate gene expression patterns of FASL and IL5RA by B-cells using real-time PCR. Additionally, we aimed to evaluate the phenotype of B-cells from individuals with TB by means of flow cytometry, and how this changes throughout treatment. Finally, we evaluated cytotoxic host marker levels in BAL and plasma from individuals with TB.

Ethics statement
Ethical approval was obtained from the ethics committee of Stellenbosch University (N10/01/013 and N13/05/064) and the City of Cape Town City Health. The study was conducted according to the Helsinki Declaration and International Conference of Harmonization guidelines. Written informed consent was obtained from all study participants.

Participant recruitment
All participants were recruited in the Ravensmead/Uitsig Community of Cape Town, South Africa. Participants newly diagnosed and who had a first episode of TB were recruited for the study before initiation of treatment. The inclusion criteria is shown in Table 1. The TB cases were followed over the course of 6 month treatment. In addition, healthy controls (CTRL) from the same community were also recruited. The healthy controls were all QuantiFERON (QFN) positive, which is indicative of latent TB infection (LTBI). Participants with cancers were included in the study as controls for the bronchoscopies performed on TB cases. These participants displayed respiratory symptoms due to inflammatory or infective cause, but were M. tuberculosis sputum-culture and GeneXpert negative, which indicated that they did not have active TB disease. Individuals who had recently completed anti-TB treatment (within 3 months), were M. tuberculosis sputum-culture negative and who consented were selected for bronchoscopies. Bronchoscopies were performed by an experienced pulmonologist at Tygerberg Hospital, Cape Town, and transported to the laboratory under controlled conditions for processing.

B-cell phenotype analysis by flow cytometry
A 1 ml of sodium heparin blood was collected from 15 CTRLs (once-off) and 13 TB cases at diagnosis, week 2, months 1 and 6 of treatment (Supplementary Table S1). The blood was analysed using 1Â BD FACS Lysing solution (BD), and the leukocytes were cryopreserved in 10% DMSO, Saarchem), placed in a Mr. Frosty at À808C overnight and transferred to liquid nitrogen the following day.

B-cell gene expression analysis
Blood was collected into sodium heparin tubes from 19 TB cases at diagnosis, months 2 and 6 of treatment (Supplementary  Table S2), and peripheral blood mononuclear cells (PBMCs)

58
were isolated using the ficol-histopaque (Sigma-Aldrich, Missouri, USA) separation method. Subsequently, total B-cells were isolated from the PBMCs by positive selection using the B-cell MACS CD19 Microbeads, (Miltenyi) according to the manufacturer's instructions. Cell purities were checked by flow cytometry and were above 90%. RNA was isolated from these B-cells and used to synthesize cDNA. The procedure was carried out in a thermal cycler (Life Technologies, California, USA) using the First Strand Kit (Hilden, Germany) according to the manufacturer's instructions. Quantitative PCR was carried out using the ABI 7900HT platform (Applied Biosystems, California, USA). The RT 2 Profiler Custom Arrays (Qiagen) were utilized and manufacturer's instructions were followed. The arrays contained primers for the following genes: Fas-ligand (FASLG) (NM000639.2) and IL5RA receptor alpha (IL5RA) (NM_000564.4) as well as two housekeeping genes B2M (NM_004048) and GAPDH (NM_002046). FASLG and IL5RA is associated with regulatory B-cells.
The gene expression data obtained from the ABI 7900HT was represented as Ct values, which indicated the earliest visible cycle of amplification. These Ct values were converted to fold change values using the Qiagen online software (www.SABiosciences.com/pcrarraydataanalysis.php).

Host marker detection in BAL fluid and PLASMA using MAGPix platform
Bronchoscopies were performed on a group of TB patients at diagnosis and another group of TB patients at the end of treatment (Supplementary Table S3). During the procedure, broncho-alveolar lavage (BAL) was collected by washing the lung interior with saline. The BAL samples were spun down and the supernatants were used for protein analysis. In addition, blood was collected from these individuals into EDTA tubes and centrifuged in order to collect plasma. Bronchoscopies were also performed on 10 individuals with lung cancer, to serve as (other lung-disease) controls, for the collection of BAL samples.  (GM-CSF). The assay was carried out following the manufacturer's instructions and the concentration of the marker measured on the MAGPix platform (Bio-Rad Laboratories, California, USA). Two quality controls included in the kit were run in duplicate. Levels of all analytes in the quality controls were within the expected ranges (pg/ul). A standard curve ranging from 0.08 to 1.7 (pg/ul) was used for GM-CSF, 0.93-19 (pg/ul) for Granzyme A, 0.04-0.85 (pg/ul) for Granzyme B, 0.03-0.67 (pg/ul) for IL-5, 0.27-6.0 (pg/ul) for sFasL, and 0.24-5.6 (pg/ul) for Perforin. Bio-Plex Manager software (California, USA), version 6.1, was used to analyze the data.

Statistical analysis
Statistical analysis was carried out using Graph Pad Prism 5 Software (San Diego, CA). Where applicable, differences between groups were calculated using Student t-test, one-way ANOVA and Dunn's multiple comparison posthoc tests. Significant differences are indicated by P < 0.05.

Activated B-cells express a FasL þ IL5Raþ phenotype following anti-TB treatment
Studies have shown that there is a subset of regulatory B-cells that express FASL, and that these cells are induced by IL-5, with an increased expression of IL5RA, thus we were interested in the expression of FASL (CD178) and IL5RA (CD125) on B-cells. Additionally, we were interested in the expression of PD1 and CD40 as well, as increased expression of these receptors are linked to cell exhaustion. To evaluate the expression of the surface markers, the gating strategy shown in Figure 1 was followed.
When comparing the frequency of CD19 þ IgM þ CD38 þ cells expressing FASL (CD178) and IL5RA (CD125) between CTRLs and TB cases, there were no significant differences between the two groups (Fig. 2a). There was, however, a trend that the expression of FASL (CD178, P ¼ 0.06), was higher in CTRLs when compared to TB cases. When comparing the frequency CD19 þ IgM À CD38 þ cells   expressing FASL/CD178 and IL5RA/CD125, we found that there was a significant difference in the frequency of IL5RA between CTRL and TB groups (P ¼ 0.03; Fig. 2b). There were no significant differences in frequency of CD19 þ IgM þ CD38 þ and CD19 þ IgM À CD38 þ expressing CD40 between CTRLs and TB participants. PD1 (Fig. 2c and d).
There was, also, however a trend in that the expressing of CD40 by IgM À B-cells, where there was a higher expression in CTRLs (P ¼ 0.06). Overall, the expression of FASL/CD178 and IL5RA/CD125, as well as CD40 and PD1 was higher in CTRLs compared to TB participants (Fig. 2), even though these differences were not statistically significant.

Effects of anti-TB treatment on the expression of regulatory and exhaustion markers by B-cells
The frequency of CD19 þ IgM þ CD38 þ and CD19 þ IgM À CD38 þ cells expressing FASL/CD178 and IL5RA/CD125, as well as CD40 and PD1 were evaluated over the course of 6-month anti-TB treatment. Figure 3 shows the changes in the expression of FASL and IL5RA following TB treatment in order to assess the effects of treatment. The frequency of CD19 þ IgM þ CD38 þ of cells expressing FASL/CD178 (CD178 þ IL5RA À ) decreased during treatment, whereas the expression of IL5RA/CD125 (CD178 À IL5RA þ ) increased during treatment (Fig. 3a). Furthermore, the frequency of cells expressing both FASL/CD178 and IL5RA/CD125 (CD178 þ IL5RA þ ) was higher at week 2 and month 6 of treatment compared to diagnosis (Fig. 3a). These changes were, however, not statistically significant. The frequency of CD19 þ IgM À CD38 þ cells expressing FASL/CD178 decreased by month 6 of treatment, however expression was highest at month 1 of treatment (Fig. 3b). The expression of IL5RA/CD125 increased by the end of treatment, but was lowest at month 1 of treatment. (Fig. 3b). Additionally, the frequency of cells expressing FASL/CD178 and IL5RA/CD125 (CD178 þ IL5RA þ ) was higher at month 1 and month 6 of treatment compared to diagnosis. These changes, however, were not statistically significant either.
The expression of CD40 and PD1 by CD19 þ IgM þ CD38 þ and CD19 þ IgM À CD38 þ cells were evaluated as well (Fig. 4). The frequency of IgM þ and IgM À B-cells expressing CD40 (CD40 þ PD1 À ) increased during treatment, with the increase being more apparent in the IgM þ population, while the frequency of PD1 expressing (CD40 À PD1 þ ) B-cells decreased ( Fig. 4a and b). The frequency of IgM À B-cells expressing both CD40 and PD1 (CD40 þ PD1 þ ) was elevated at month 1; however, none of the latter changes were statistically significant.
Expression of genes linked to regulatory phenotype during treatment RT-qPCR was used to evaluate the expression of FASL and IL5RA transcripts in B-cells from TB participants during treatment (diagnosis, months 2 and 6). When the expression of genes were compared between the 3-time points, FASL expression was found to be significantly upregulated at month 6 compared to diagnosis (P ¼ 0.01; Fig. 5a). Similarly, the expression of IL5RA was significantly upregulated at month 2 (P ¼ 0.01) and month 6 (P ¼ 0.01; Fig. 5b) compared to diagnosis.
Changes in BAL-secreted cytokines following anti-TB treatment but not in the plasma of TB cases The levels of 6 analytes (Granzyme-A, Granzyme-B, Perforin, sFasL, IL5, and GM-CSF) were measured in the BAL fluid and plasma of TB patients at diagnosis and at the end of treatment (EOT), as well as controls using MAGPix technology. When comparing the analyte profiles measured at diagnosis and EOT in the BAL fluid of TB participants, we found that the profiles of the two different time points differ from each other (Fig. 6a). Furthermore, the analyte profiles of BAL fluid of TB participants at both time points differed from that of CA controls (Fig. 6a). When comparing the profiles in the plasma of these patients at diagnosis and EOT, no differences were seen (Fig. 6b). Statistical analysis on the analyte levels in the BAL fluid of the TB patients and CTRLs revealed, that Granzyme-B (P < 0.01) and Perforin (P < 0.05) levels are significantly lower at EOT when compared to Diagnosis (Dx) (Fig. 7b and e). EOT levels of Granzyme-B (P < 0.05), sFasL (P < 0.01), and Perforin (P < 0.05) were also significantly different from CTRLs (Fig. 7b, c, and e).

Discussion
In the current study, we aimed to investigate whether the so-called regulatory ''killer'' B-cells are present during tuberculosis and if there was an effect of treatment on the frequency of these cells. Previous studies suggest that these cells are present during autoimmune disease and may play an essential role in the regulation of inflammation [16,17]. Furthermore, there are characteristics about the regulatory ''killer'' cells that lead us to think that these cells (even though low in frequency) may play a role during tuberculosis as well, such as through the induction of The phenotype of B-cells was assessed by means of flow cytometry to identify cells expressing FASL and IL5RA, which are markers associated with and used to identify killer B-cells [9]. Even though there were no significant differences, apart from the difference in the frequency of IgM À cells expressing IL5RA, the data suggest that there is a trend to higher expression of FASL and IL5RA by cells from CTRLs compared to TB cases. Interestingly, there is an appearance of IgM À cells expressing both FASL and IL5Ra at month 1 of treatment, which increases by M6. The expression of FASL and IL5RA in IgM þ cells also increases during treatment, but appears earlier (week 2) during treatment. These results indicate the treatment may be inducing an increase in the expression of FASL by B-cells and that higher expression is associated with reversion back to healthy (or latently infected state) status, as the data show that B-cells from healthy controls has a higher expression of FASL compared to TB cases at diagnosis.
To assess the level of B-cell activation and immune exhaustion, we evaluated frequency of B-cells expressing CD40 and PD1. IgM þ B-cells expressing CD40 increased over time, whereas IgM À B-cells expressing CD40 only started increasing by month 1 of treatment, with lower frequencies compared to IgM þ B-cells. This suggests that there is an increase in the activation of B-cells during treatment, which implies that treatment may be inducing activation of B-cells.
Cliff and coworkers showed that B-cell associated activity increases as successful treatment progresses [18].
To assess the effects of anti-TB treatment on the expression of FASL by B-cells at the gene level, we evaluated the changes in transcript levels over time. The results indicated that there is an increase in the levels of both genes by month 6 of treatment. In a previous study, we found that the levels of FasL and Il5Ra of the untreated TB cases were lower when we compared it to that of community controls (data not shown) [19]. These results correlate with the phenotypic results which indicates that there is an increase in the surface expression of FASL and IL5RA during treatment, even though those changes are not statistically significant.
Analyte levels at the site of disease (BAL fluid) was evaluated by means of the multi-cytokine platform. sFasL secretion increased by the end of treatment, where perforin and granzyme-B levels were reduced at the end of treatment. The increase in sFasL in BAL fluid by the end of treatment also correlates with the phenotypic and gene expression results. FasL expression is mostly attributed to CD8þ T-cells and NK cells in response to viral infections [20,21]. Although B-cell frequencies in the lung is low, there is increased infiltration of B-cells into the lung during inflammation [22][23][24][25]. Thus, B-cells may also be contributing to the sFasL levels in the lung.
Collectively, these results show that TB treatment induces an increase in the expression of FASL by B-cells. These results are in agreement with studies which show that there is an increase in the expression of FASL by B-cells during parasitic infection and autoimmune disease [8,26]. Lundy and coworkers (2001) show that B-cells from mice infected with Schistosoma mansoni had a higher expression of FASL compared to healthy mice [8]. Similarly, another study shows an increase in B-cell FASL expression during airway inflammation and asthma, as well as systemic lupus erythematosus (SLE) [26][27][28]. Increased FASL expression by B-cells may have different effects, depending on the disease or infection. Culture experiments and killing assays carried out by Klinker and coworkers showed that B-cells stimulated with CD40L and IL-5-induced FasL-dependent apoptosis of CD4þ T-cells, and that these B-cells target T-cells based on the specificity of the antigen [9]. Thus, increased FASL expression may have a protective role in the case of autoimmune disease, where apoptosis of CD4 þ T-cells leads to decreased inflammation. In contrast, apoptosis of Th1 cells may cause a skewed balance between Th1 and Th2 cells, leading to airway inflammation and asthma due to cytokines produced by Th2 cells.
Here, the increase in the expression of FASL during treatment implies that these regulatory ''killer'' B-cells may play a protective role during TB. One hypothesis may be that these regulatory B-cells are somehow inducing activation of T-cells, which later undergoes apoptosis [29][30][31]. An alternative hypothesis may be that these regulatory B-cells are inducing apoptosis of T-cells via FASL, as a means to rid the host of the bacteria. This process is well documented for myeloid cells [32]. B-cells are infected by M. tuberculosis through a process of micropinocytosis [33,34] and express death receptors. The manner in which a general loss of T-cells occur during tuberculosis is not well defined [35]. Literature states that FasL plays a role during activation-induced cell death [29,[36][37][38]. When T-cells are activated, there is an increased expression of Fas by these cells and they become sensitized to FasL-induced apoptosis [37].
Even though the data in the relatively small study points in the same direction as the studies above, we acknowledge that the observations are preliminary and that no definite conclusions can be drawn from our findings. The studies need to be replicated and expanded by both our research group and that of others. This will then contribute further to our understanding of the phenotype of these B-cells induced upon completion of anti-TB treatment. However, functional studies are required to further elucidate the role of the ''killer'' B-cells during TB.

SUPPORTING INFORMATION
Additional supporting information may be found in the online version of this article at the publisher's web-site. Table S1. Cohort for B-cell phenotypic analysis. Table S2. Cohort for Gene expression analysis. Table S3. Cohort for Luminex.