Characterization of Regulatory B Cells in Graves’ Disease and Hashimoto’s Thyroiditis

A hallmark of regulatory B cells is IL-10 production, hence their designation as IL-10+ B cells. Little is known about the ability of self-antigens to induce IL-10+ B cells in Graves’ disease (GD), Hashimoto’s thyroiditis (HT), or other autoimmune disease. Here we pulsed purified B cells from 12 HT patients, 12 GD patients, and 12 healthy donors with the thyroid self-antigen, thyroglobulin (TG) and added the B cells back to the remaining peripheral blood mononuclear cells (PBMCs). This procedure induced IL-10+ B-cell differentiation in GD. A similar tendency was observed in healthy donors, but not in cells from patients with HT. In GD, B cells primed with TG induced IL-10-producing CD4+ T cells. To assess the maximal frequency of inducible IL-10+ B cells in the three donor groups PBMCs were stimulated with PMA/ionomycin. The resulting IL-10+ B-cell frequency was similar in the three groups and correlated with free T3 levels in GD patients. IL-10+ B cells from both patient groups displayed CD25 or TIM-1 more frequently than did those from healthy donors. B-cell expression of two surface marker combinations previously associated with regulatory B-cell functions, CD24hiCD38hi and CD27+CD43+, did not differ between patients and healthy donors. In conclusion, our findings indicate that autoimmune thyroiditis is not associated with reduced frequency of IL-10+ B cells. These results do not rule out regulatory B-cell dysfunction, however. The observed phenotypic differences between IL-10+ B cells from patients and healthy donors are discussed.


Introduction
Autoimmune thyroiditis (AITD) includes Graves' disease (GD) and Hashimoto's thyroiditis (HT), which are typically associated with hyper-and hypothyroidism, respectively. B cells are known to play an essential role in GD by virtue of their production of pathognomonic activating autoantibodies against the thyroid-stimulating hormone (TSH) receptor, leading to increased production and secretion of the thyroid hormones T 4 and T 3 and a compensatory decrease in TSH production by the anterior pituitary gland [1,2]. It is unclear whether B cells also play a pathogenic role in HT. Autoantibodies to the thyroid self-antigens thyroglobulin (TG) and thyroid peroxidase (TPO) are commonly found in both GD and HT, but T-cell mediated destruction of thyroid architecture plays a central role in HT [3,4]. This leads to low production of T 4 and T 3 , and a compensatory increase in TSH production [3,4]. The beneficial effect of the B cell-depleting antibody rituximab in a number of autoimmune diseases, including multiple sclerosis and type 1 diabetes mellitus, suggests a critical role for B cell endorsement in T-cell dominated diseases [5].
Little is known about IL-10 + B-cell frequency or the ability of B cells to induce IL-10 + T cells in AITD. Here we investigated the capacity of B cells from patients with GD, HT, and those from healthy donors to differentiate into IL-10 + B cells when challenged with TG or the mitogen PMA/ionomycin. Moreover, we assessed the capacity of B cells pulsed with TG to induce IL-10 production by CD4 + T cells and cytokine release from intact peripheral blood mononuclear cells (PBMCs). Finally, the expression by IL-10 + B cells of several surface markers that have previously been associated with regulatory functions was examined.

Subjects
Whole blood from 12 healthy donors (demographics: 9 females, 3 males; median age 44 yrs) with no history of autoimmune disease was provided by the Blood Bank at Copenhagen University Hospital. A total of 12 patients with HT and 12 patients with GD, attending the Endocrinology outpatient clinic at Odense University Hospital between November 2013 and March 2014 participated in the study. HT patients were characterized by elevated serum TSH levels, raised serum TPO Ab and/or TG Ab levels, and undetectable anti-TSHR Abs. Suppressed serum TSH levels, increased free T 4 (FT 4 ) and free T 3 (FT 3 ) levels, elevated serum anti-TSHR Ab levels, diffuse uptake on thyroid scintigraphy, and ultrasound demonstrating diffuse hypoechogenicity typified those with GD.
All patients were diagnosed within three years of study participation with the exception of one HT patient diagnosed in 2009 and two GD patients diagnosed in 2008 and 2009. At the time of blood collection, 9 out of 12 GD patients were receiving methimazole (median: 10 mg/day, IQR: 5 -15 mg/day) or levothyroxine (median: 75 μg/day, IQR: 50 -150 μg/day), while 5 out of 12 HT patients were receiving levothyroxine (median: 112.5 μg/day, IQR: 100 -125 μg/day). Duration of anti-thyroid treatment varied from 2 weeks to 5 years. Further clinical details of the study participants are outlined in Table 1. Written informed consent was obtained from all participating subjects prior to their participation. The study was approved by the Ethical Committee from the Region of Southern Denmark (project #28699) and followed the guidelines outlined in the Declaration of Helsinki.    Chromogenic LAL, Lonza) assay revealed endotoxin, which was removed using Triton X-114 as previously described [15].

Pulsing of B cells with antigen
Purified B cells were either preloaded with TG (30 μg/mL) or CpG ODN (10 μg/mL) for 1 h at 37°C or received no antigen and served as the negative control. Unless otherwise stated, 1.0x10 5 B cells were co-cultured with 2.0x10 5 residual PBMCs in RPMI 1640 media containing 30% (v/v) AB serum for 48 h. Supernatants were collected after 48 h of incubation and analyzed for IL-10 and IL-6 using Luminex (Austin, TX). Multiplex beads were supplied by BioRad (Hercules, CA).

Flow cytometry
The cells were acquired with a FACS Canto (BD Bioscience) flow cytometer with argon laser (488 nm) and Helium-Neon laser (633nm) excitation.
All analyses were carried out using FlowJo V10 (TreeStar, Ashland, OR). Dead cells were excluded based on Live/Dead Fixable Near InfraRed staining and B cells were identified as CD19 + events within a morphological lymphocyte gate. Individual IL-10 + B cells were identified using the gating strategy demonstrated in S1 Fig.

Statistics
Comparisons between each patient group and healthy donors were performed using the twotailed Mann-Whitney U-test. Differences between HT and GD patients were considered subordinate. In addition, comparisons within each group were performed using the Wilcoxon matched-pairs signed rank test. Correlations between thyroid hormones (FT 3 , FT 4 and TSH) and IL-10 + B cell frequency were evaluated using Spearman Rank correlation coefficient. All analyses were carried out using GraphPad Prism version 6 (GraphPad Software, La Jolla, CA). P values are presented in the figures. P-values < 0.05 were considered significant.

Induction of IL-10 + B cells by the thyroid self-antigen TG
In general, polyclonal stimulation with PMA/ionomycin or CpG has been used to study cytokine production by B cells. While this approach may show the potential of the entire B cell pool to differentiate into cytokine-producing cells, it does not reflect the more physiological situation where B cells may be stimulated clonally with self-antigens and receive help from antigen-specific T-helper cells. To mimic such conditions, B cells were purified and pulsed with the thyroid self-antigen TG before they were added back to the remaining PBMCs (Fig 1).
The hallmark of immunoregulatory function in B cells is the production of IL-10 [9]. As shown in Fig 1A, exposure to TG significantly increased the proportion of IL-10 + B cells in cultures from GD patients (p = 0.01). A similar tendency was observed among healthy donors (p = 0.10), but not among HT patients. After pulsing with TG, cultures derived from GD patients tended to contain more IL-10 + B cells than cultures derived from healthy donors (p = 0.059). As expected, CpG induced IL-10 + B cells substantially in all three donor groups.
We next examined whether uptake and presentation of TG by B cells promoted induction of IL-10 producing T cells in the co-cultures. As shown in Fig 1B, pulsing of B cells with TG resulted in an increase in the proportion of IL-10-producing CD4 + T cells in cultures derived from GD patients (p = 0.01), but not in cultures from HT patients or healthy donors.
As a supplement to the measurement of intracellular IL-10, levels of IL-10 released into the culture supernatants was determined. Unexpectedly, greater IL-10 production was observed in both patient groups compared with healthy donors in the presence of B cells not pulsed with TG (Fig 2A). The same applied to the production of the pro-inflammatory cytokine IL-6, the levels of which were one to two orders of magnitude higher than those of IL-10 ( Fig 2B). Pulsing of B cells with TG did not significantly alter these cytokine profiles. Therefore, baseline secretion of IL-10 and IL-6 by un-stimulated PBMCs from AITD patients is higher than in healthy donors independent of addition of exogenous self-antigen.

IL-10 + B-cell proportions in AITD patients and healthy donors
To compare the maximal achievable IL-10 + B-cell production in the two patient groups with those of healthy controls, polyclonal stimulation of PBMCs with PMA/ionomycin was used to  induce IL-10 expression (Fig 3A). Approximately, 1% IL-10 + B cells were identified following this stimulation in all three donor categories (Fig 3A).
A significant positive correlation was found between FT 3 levels and the frequency of IL-10 + B cells in GD patients (P = 0.0016, Fig 3B) while a borderline-significant correlation was identified between FT 4 levels and IL-10 + B-cell frequency (P = 0.059; data not shown). However, no correlation was found between IL-10 + B-cell abundance and serum TSH (data not shown).
The total B-cell count did not differ between healthy donors and patients with GD or HT (data not shown).
Representative dot plots of CD5, CD25 and TIM-1 surface expression on IL-10 + B cells from a healthy donor are shown in Fig 4A-4C. As expected from earlier studies, IL-10 + B cells were enriched with all three markers, compared to the rest of the B-cell pool, in all three donor groups (Fig 4D-4F). The median proportion of CD25 + IL-10 + B cells constituted 30% in both GD and HT, compared to only 17% in healthy donors (P = 0.039 and P = 0.0009, respectively; Fig 4E).
IL-10 + TIM-1 + B cells were more abundant in GD and HT than IL-10 + TIM-1 + B cells from healthy donors (P = 0.024 and P = 0.0009, respectively; Fig 4F). These data should be interpreted with caution; however, since the 4 h PMA/ionomycin stimulation used to induce differentiation of IL-10 + B cell more than doubled the frequency of B cells displaying TIM-1 (S1 Table).
The CD24 hi CD38 hi and CD27 + CD43 + phenotypes, previously associated with Breg function [22,23,25], were also investigated. Only a minority of IL-10 + B cells were CD24 hi CD38 hi ( Fig 5A), regardless of the donor group, and the proportion of these cells was similar in all three donor groups (data not shown). IL-10 + B cells were predominantly found within the CD24 hi CD38memory B-cell compartment in healthy donors (Fig 5B), while they were underrepresented in this compartment in HT patients (p = 0.012; Fig 5B). HT patients had a correspondingly higher proportion of mature CD24 int CD38 int IL-10 + B cells than healthy donors (37% vs. 21%, respectively, p = 0.0023; data not shown).
The median proportion of IL-10 + B cells expressing the CD27 + CD43 + phenotype comprised 19% in healthy donors and 29% in both patient groups (NS; data not shown).

Discussion
In this study, we examined the induction of IL-10 + B cells by the thyroid self-antigen TG, and by the polyclonal stimulators PMA/ionomycin and CpG. We consider TG to represent a more physiologically relevant stimulus than the mitogenic stimuli normally used to study regulatory B cells. In keeping with our previous findings in healthy donors [12], TG induced significant IL-10 + B-cell differentiation in GD, and a similar tendency was observed in healthy donors. These data suggest that IL-10 + B-cell differentiation in GD is not compromised. It should be noted, however, that we did not assess TSH receptor-reactive IL-10 + B cells, which are presumably more relevant to the pathogenesis of GD. Interestingly, B cells pulsed with TG were capable of inducing IL-10-producing CD4 + T cells after being added back to PBMCs cultures from GD patients. This is in keeping with our previous findings in cells from healthy donors [12].
No statistically significant changes in IL-10 or IL-6 levels could be detected in culture supernatants following exposure of B cells to TG, but trends confirmed our previous findings that B cells pulsed with TG can induce production of both cytokinesas well as tumor necrosis factor (TNF)-α and TGF-βin co-cultures with autologous T cells [12]. While the three donor groups did not differ with respect to the ability of TG to induce cytokine responses, the baseline production of both IL-10 and IL-6 was greater in the two patient groups. Considerably more IL-6 than IL-10 was produced, which might reflect a more pro-inflammatory environment in AITD patients compared to healthy donors. We have previously shown that addition of TG to PBMCs induced increased production of TNF-α, IL-2, interferon-γ and IL-10 in GD patients and HT patients, compared to healthy donors [26]; the data presented here suggest that B cells pulsed with TG do not provide a stimulus strong enough for similar changes in cytokine production.
We observed no difference between healthy donors and either patient group with respect to induction of IL-10 + B cells by the polyclonal stimuli PMA/ionomycin or CpG. In contrast to our results, Zha et al., also using CpG and PMA/ionomycin as stimuli, observed a significantly lower frequency of IL-10 producing B cells in new-onset GD patients than in healthy donors [27]. The reason for this discrepancy may be that Zha et al., investigated new-onset GD patients, while those with GD in our study cohort had disease of a longer duration, and 9 out of 12 of them had received methimazole. Significantly decreased IL-10 + B-cell frequency has also been described in patients with rheumatoid arthritis [28,29]. Importantly, our finding of similar frequencies of IL-10 + B cells in healthy donors and AITD patients does not rule out diseaseassociated B cell impairment with respect to IL-10-independent regulatory mechanisms affecting T-cell activity [27]. Alternative ways for B cells to regulate T-cell functions include mechanisms mediated by TGF-β [30], and expression of surface molecules such as programmed death ligand 2 [9,31] and Fas ligand [9,32,33].
Unexpectedly, the frequency of induced IL-10 + B cells correlated positively with FT 3 and FT 4 levels (the latter borderline significance), within the GD patients. The increase in IL-10 + B cell numbers may reflect an attempt of the B cell compartment to limit or eliminate disease activity. This has been demonstrated in experimental autoimmune encephalomyelitis where wild-type mice spontaneously remit or even recover within 30 days, but mice with a selective lack of IL-10 expression in B cells fail to do so [7]. Recently, we showed that the frequency of IL-10 + T cells is inversely correlated with TSHR-antibody levels, a marker of disease activity, in GD patients [34]. The increasing IL-10 + B-cell frequencies with increasing FT 3 levels reported in this study may reflect compensation for the relative IL-10 + T cell deficiency.
Within the IL-10 + B cell subset, a greater proportion of cells displaying CD25 or TIM-1 were found in patients with HT or GD than in healthy donors. CD25 is the IL-2 receptor α-chain, which promotes B-cell and T-cell proliferation [35]. Expression of CD25 may allow IL-10 + B cells to become activatedand regulate the immune responseunder circumstances with abundance of IL-2 in the environment, i.e. in presence of activated effector T cells. Our data confirm that IL-10 + B cells are enriched with CD25 + cells, as compared with the entire B cell population, but 50-75% of IL-10 + B cells did not express this marker. TIM-1 is a T-cell costimulatory molecule which regulates CD4 + T-effector cell differentiation and responses in autoimmune and alloimmune settings [21,36,37]. Ligation of TIM-1 induces IL-10 production by TIM-1 + B cells and, in so doing, may promote immune tolerance [21]. The finding of a significantly increased proportion of TIM-1 + IL-10 + B cells in HT and GD may thus reflect a compensatory increase in Bregs in this group. The findings concerning TIM-1 in this study should be interpreted with caution since TIM-1 expression on the B-cell surface was greatly enhanced by PMA/ionomycin stimulation. Even so, the majority of IL-10 + B cells in each group lacked this marker.
Immunoregulatory properties have been associated with CD24 hi CD38 hi B cells, a phenotype which defines transitional B cells [22,23,28,38], and which has been reported to be underrepresented in active RA [28]. In our study, the proportions of CD24 hi CD38 hi B cells were similar between healthy donors and patients with GD or HT; however, IL-10 + B cells were predominantly CD24 hi CD38 -(memory) cells in healthy donors, but almost evenly distributed among CD24 hi CD38 -(memory) cells and CD24 int CD38 int (mature) cells in both patient groups. Others have also found IL-10 production within the CD24 int CD38 int and CD24 hi CD38phenotypes, indicating that these subsets may also exert immunosuppressive activity [23]. We found no difference between the donor groups with respect to the frequencies of IL-10 + B cells with the CD27 + CD43 + phenotype, which has also been associated with IL-10 production and immunosuppression [24,25]. We did not examine CD24 in combination with CD27, but Zha et al., found a lower frequency and functional impairment of the CD24 hi CD27 + B cells in recent onset GD [27].
Among the subjects in this study, 9 out of 12 GD patients were undergoing anti-thyroid drug treatment, and 5 out of 12 HT patients were receiving levothyroxine treatment at the time of blood collection. It has been reported that anti-thyroid drugs may inhibit lymphocyte function and reduce production of pro-inflammatory cytokines, in part by inhibiting NF-κB [1,[39][40][41]. Due to the potentially normalizing effects of anti-thyroid treatment, differences between patients and healthy donors with respect to IL-10 production may have been underestimated in this study. It should also be noted that IL-10 production by circulating B cells might not mirror production in intra-thyroidal B cells.
In conclusion, differentiation of B cells into IL-10-producing cells was unimpaired in GD and HT after stimulation with TG as well as with PMA/ionomycin. As has been observed previously, IL-10 + B cells did not segregate into clearly defined phenotypic subsets, regardless of the donor group examined. Future treatment of AITD should be aimed at reinstituting immune tolerance in these patients with antigen-specific therapies. Our observation that pulsing of B cells with TG induces IL-10 + T cells as well as IL-10 + B cells in GD patients suggests that loading of B cells with appropriate self-antigens may be exploited in this respect.
Supporting Information S1 Table. Expression of surface markers by bulk B cells before and after PMA/ionomycin stimulation. PBMCs from healthy donors (N = 6) were stimulated with phorbol 12-myristate 13-acetate/ionomycin (PMA/ionomycin) or left unstimulated (baseline) for 4 hours. The proportion of B cells expressing each surface marker after stimulation was related to that of unstimulated B cells as a ratio. The median value of the ratios of 6 individual donors is shown.