Involvement of B Cell Subsets in Patients With Steven Johnson Syndrome/Toxic Epidermal Necrolysis ( SJS / TEN )

Steven-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are rare, acute inflammatory skin reactions with a high mortality rate of ~20%–50% [1]. It is mainly related to medications, which can trigger an acute reaction causing extensive skin necrosis and detachment [2, 3]. Presence of a drug-antigen-driven oligoclonal cytotoxic CD8+T-cell infiltrate in the affected skin is a hallmark of SJS/TEN pathogenesis [4]. In contrast, the involvement of B cells and their subsets in disease development has remained unexplored. To address this gap, we collected fresh blood samples from SJS/TEN patients (n = 16, naïve) and following triple therapy induction with cyclosporine, IVIG and a glucocorticosteroid at week 1 (n = 12, follow-up 1) and week 5 (n = 10, follow-up 2) [5]. Additionally, we obtained fresh blood from 13 healthy controls (HCs). Patient and HC demographic and clinical characteristics are summarized in Table 1. Using multiparameter spectral flow cytometry, we phenotypically characterize B cell subsets from fresh blood samples (Supporting Information). Initially, we analyzed the frequency of B cells in naïve, treated SJS/TEN patients and HC blood in a cross-sectional study. Our flow cytometry analysis showed no difference in the CD3−CD19+B cell population of naïve SJS/TEN and HCs (Figure 1A,B). However, within one week of treatment, naïve patients showed elevated B cell levels compared to HCs, which normalized by week-5 of treatment (Figure 1A,B). Interestingly, a similar pattern was observed in our longitudinal analysis (Figure 1C), suggesting a dynamic association of B cells and SJS/TEN treatment. We observed a profound increase in the CD27+CD38+ plasmablast subset in naïve patients compared to HCs. Interestingly, these increased plasmablast levels normalized by week 1 and 5 of treatment (Figure 1D,E). Since similar patterns were observed in our longitudinal cohorts (Figure 1F), this suggests that the CD27+CD38+ plasmablast subset is selectively expanded in SJS/TEN patients. Because B cell function is regulated by regulatory B (Breg) cells, we next investigated frequencies of Breg. Breg numbers were indeed elevated in naïve SJS/TEN patients compared to HCs. However, Breg frequency decreased by week 1 of treatment and returned to normal levels by week 5 (Figure 1G,H). This was also true when we analyzed. Breg frequency in longitudinal cohorts (Figure 1I). We next aimed to understand whether the frequency of plasmablast B cells is differentially regulated in our SJS/TEN subgroups. Based on the standard classification criteria of skin detachment (< 10% SJS, 10%–30% SJS/TEN-overlap, > 30% TEN) we identified 8 patients with SJS, 4 with SJS/TEN-overlap and four patients with TEN. Comparison was conducted between the grouped SJS and SJS/TEN-overlap cohorts, the TEN cohort and HC. Among these cohorts we found similar B cell frequencies (Figure 2A,B). However, the frequency of Breg cells was significantly higher in naïve TEN patients compared to naïve SJS and HC. Within 5 weeks of treatment, the frequency of Breg cells normalized to the level observed in HCs (Figure 2C,D). Comparing the frequency of plasmablast B cells revealed a significant expansion only in SJS patients, while the frequency of plasmablast cells in TEN patients remained on HC level. However, plasmablast B cells frequency in SJS patients achieved HC levels by week 5 of treatment (Figure 2E). Longitudinal follow-up in TEN patients confirmed unregulated low frequencies of plasmablast B cells in naïve and treated patients (Figure 2F). Our findings thus suggest that the expansion of plasmablast B cells and Bregs may help to differentiate SJS from TEN patients. Due to the limited availability of samples, we could not classify the type of plasmablasts or the secretion of different IgG subtypes from plasmablast B cells in our study. Additionally, we could not determine whether these plasmablast B cells resulted from T-cell-dependent activation. However, it is possible that activated CD8+T cells produce IFN-γ, which can enhance the differentiation of plasmablast B cells from memory B cells [6]. The presence of plasma cells in the bloodstream can serve as a biomarker for IgG4-related diseases, which are considered to belong to the class of autoimmune diseases. Here, plasma cells produce IL-10, which suppresses inflammation [7]. Consistent with this, IgG4 deficiency is associated with a worse disease outcome in 20% of inflammatory bowel disease patients, for instance [8]. In contrast, a loss of keratinocyte connectivity mediated by IgG4 has been reported in patients with pemphigus, in which IgG4 cross-reacts with the skin cell adhesion protein desmoglein [9, 10]. This mechanism could probably contribute to the skin necrolysis in SJS and TEN patients. However, our findings of elevated plasmablast B cells in SJS would suggest higher IgG4 levels in SJS patients whereas in fact TEN patients present with more skin necrolysis than SJS patients. Consequently, the loss of keratinocyte connectivity mediated by IgG4-secreting plasmablasts is a less plausible explanation for skin detachment in SJS/TEN patients. In addition, we could not confirm that plasmablast B cells produce IgG4 in SJS/TEN patients. Breg cells are immunosuppressive cells that promote immunological tolerance. They achieve this by producing IL-10, IL-35, and TGF-β, which all prevent the expansion of pathogenic T cells and other pro-inflammatory lymphocytes. Different inflammatory environments induce distinct Breg cell populations, which highlights the relevance of inflammatory signals in Breg cell differentiation [11]. Inflammation appears to trigger an increase in both the number and suppressive ability of Breg cells [12, 13]. In line with this, we noticed an enhanced frequency of Breg cells in SJS patients which was even more elevated in TEN patients (Figure 2C). However, despite TEN patients exhibiting a higher frequency of Breg cells, TEN patients also display extensive necrolysis of the skin suggesting that Breg cells are not primarily responsible for immunosuppression in these patients but rather form in consequence to a dynamic equilibrium between multiple B cell subsets and other immune cells. A key limitation of our study is its small sample size, particularly for TEN. Larger studies could help to determine whether the expansion of plasmablasts and Bregs is valid and linked to a specific mechanistic cascade within the dynamics of SJS/TEN disease processes and probably progression. In summary, our findings suggest that in addition to cytotoxic CD8+T cells, peripheral B cells and their subsets might play a role in SJS/TEN pathogenesis [6] as supported by intravenous immunoglobulin therapy for SJS/TEN [5]. We report on the importance of specific B cell subsets including plasmablasts and Bregs in distinguishing between SJS and TEN at the cellular level. Further in-depth phenotypic and functional characterization of these immune cell subsets could find additional evidence for their value as diagnostic, prognostic, and therapeutic markers in the pathogenesis of SJS/TEN. Fareed Ahmad, Martin Steinhoff and Jörg Buddenkotte conceptualized and wrote the manuscript. Jörg Buddenkotte and Martin Steinhoff had overall supervision and secured funding of the study. Majid Ali Alam analyzed data and wrote the manuscript. Jericha Miles P. Mateo, Gazala Anamangadan and Anh Jochebeth performed experiments and acquired data. Febu Elizabeth Joy, Seena Manjooran, Sara Al-Khawaga, Amina Al-Obaidli and Ayda AlHammadi recruited study individuals, obtained clinical data and collected patient samples. This work was supported by Medical Research Center routine grant (MRC-01-21-736 to J.B. and M.S.) of the MRC Fund, Hamad Medical Corporation, Qatar. M.S. has performed consultancy services for which he received compensation from Pfizer, Sanofi, Regeneron, Lilly, Novartis, Galderma, Leo, Merck, Avon, Pierre-Fabre, L'Oreal, BMS, Maruho, Toray, Mitsubishi, Maruho, Kiniksa, ZymoGenetics, Almirall; he served on advisory board for Pfizer, Novartis, Galderma, Leo, Avon, Pierre-Fabre, L'Oreal, BMS, Maruho, Toray, Mitsubishi, Maruho, ZymoGenetics, Almirall; Research was supported by Pfizer, Novartis, Galderma, Leo, Avon, Pierre-Fabre, L'Oreal, BMS, Maruho, Toray, Mitsubishi, Maruho, ZymoGenetics, Almirall. J.B. is an investigator in a Novartis-sponsored clinical study. Data available on request due to privacy/ethical restrictions. Figure S1: Manual gating strategy: The gating strategy used to identify the B cell subsets is presented. Arrows are used to visualize the relationships across plots, and numbers represent the frequency of each cell populations described here. Lymphocytes and monocytes were gated based on FSCA/SSC-A properties. Remaining RBCs were removed by plotting side-scatter from 488 nm laser (SSC-B-H) versus side-scatter from 405 nm laser (SSC-A). After dead cells and doublets exclusion, CD45+ leukocytes were identified. B cells were recognized as a CD19+CD3− gated on CD45+ leukocytes. Plasmablast cells were identified as a CD10−CD27+CD38+ and regulatory B cells (Bregs) as CD10−CD24+CD38+. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.