B cell phenotype and serum levels of interferons, BAFF, and APRIL in multisystem inflammatory syndrome in children associated with COVID-19 (MIS-C)

Multisystem inflammatory syndrome in children associated with COVID-19 (MIS-C) is now a well-specified entity described in a number of excellent publications that map in detail the immune/autoimmune/inflammatory responses accompanying this condition [1‐5]. Clinically, the hallmark symptoms include fever, rash, conjunctivitis, mucositis and serositis, lymphadenopathy, gastrointestinal, respiratory, cardiovascular, and neurocognitive symptoms, which develop on the background of markedly increased acute phase reactants and inflammatory markers in temporal association with SARS-CoV2 infection in genetically susceptible individuals. This pro-inflammatory state is accompanied by alterations in cellular populations of innate and acquired immunity, with prominent lymphopenia during the acute stage of the disease, which typically lags several weeks after acute SARS-CoV-2 infection, regardless of its severity. The lymphopenia is characterized by a decrease in T cells, but at the same time with their activation and clonal proliferation [1]. Significant alterations were shown in T cell subpopulations in MIS-C, and recent reports also show their clonality and exhaustion [2, 3]. Importantly, a number of publications document the presence of autoimmune phenomena and autoantibodies in MIS-C patients, targeting both systemic and tissue- or organ-specific antigens consistent with the systemic nature of the disease, yet also associated with its characteristic clinical presentation in specific organs, mainly the heart [4, 5]. These findings suggest a strong polyclonal antibody response driven by activated B cells. B cells, however, are less studied in the context of MIS-C. B cell counts were reported normal or decreased, in line with the general MIS-C-associated lymphopenia [6, 7]. Some studies have shown an increase in plasmablasts, as well as IgD⁻CD27⁻ double negative B cells [8, 9]. Similar IgD and CD27 double-negative activated B cells were previously documented in systemic lupus erythematosus (SLE) in association with disease activity and autoantibody secretion [4], pointing to a certain parallel in B cell activation and autoantibody production between systemic autoimmune diseases such as SLE and MIS-C. In SLE, this polyclonal autoantibody production is driven by the serum cytokine B cell activating factor (BAFF) and a proliferation-inducing ligand (APRIL) [5, 10]; however, this association has not yet been studied in MIS-C.

Similarly, while a dysregulated IFN-γ response is also a feature shared between MIS-C and SLE [11‐14], the activity of type I and type III interferons has so far only been shown in SLE and rare inborn autoinflammatory disorders [15, 16], but not in MIS-C, even though antibodies against type I interferons have been shown to contribute to COVID-19 mortality and severity [17].

To explore these immune factors contributing to the hyperinflammation in MIS-C, we set out to assess type I, II, and III interferons, serum BAFF, APRIL, and B cell phenotype and BAFFR expression in children with acute MIS-C and in healthy children after COVID-19.

To analyze the hyperinflammatory signature of MIS-C, we measured IFN-α, IFN-γ, and IFN-λ in the serum of 50 patients with MIS-C sampled shortly after admission to hospital, before administration of immunosuppressive therapy, and compared them to the sera of healthy children who underwent COVID-19 cca 6 weeks prior to sampling and had no signs of MIS-C. We saw no significant changes in IFN-α (t test with Welch’s correction p = 0.27) and IFN-λ levels (p = 0.33) (Fig. 1A, B), but IFN-γ was significantly elevated in MIS-C compared to healthy post-COVID children (p = 0.0004) (Fig. 1C).

Next, we measured the presence of autoantibodies and serum concentration of BAFF and APRIL, two cytokines supporting the development and survival of B cells, which are also implicated in other systemic autoimmune diseases reminiscent of MIS-C, such as SLE, systemic vasculitides, or Kawasaki disease [19‐21]. Of 35 patients in whom autoantibodies were tested, 4/35 (11%) had positive Ro60 antibodies, 1/35 (3%) had autoantibodies against the Sm antigen, and further, 4/35 (11%) had positive extractable nuclear antigen (ENA) screening. Serum levels of APRIL were largely below the assay detection limit, although in a subset of MIS-C patients we detected elevated APRIL levels, which were missing in post-COVID children, healthy children with no history of COVID-19 or MIS-C, and even in convalescent MIS-C patients sampled several months after full recovery (Fig. 1D). The clinical course of the disease was not distinctly different in these 4 children with high APRIL, and nevertheless, the majority of samples tested APRIL-negative, and as such, the differences were not significant.

Serum BAFF levels, on the other hand, varied significantly between acute MIS-C, convalescent MIS-C patients, post-COVID healthy children, and children with no history of COVID-19 or MIS-C (Brown-Forsythe ANOVA, p < 0.0001) (Fig. 1E). Acute MIS-C patients had the highest BAFF levels of all cohorts, which in particular were higher than those in post-COVID children (t test with Welch’s correction, p < 0.0001), but also than those in convalescent post-MIS-C patients (p < 0.0001). Interestingly, even healthy post-COVID children had elevated serum BAFF levels compared to healthy children without a history of COVID-19 (p = 0.0093). In both MIS-C patients and healthy children, the trend remained identical, with higher BAFF during/after disease, and lower after full recovery or in times of full health.

Finally, to assess the impact of upregulated BAFF signaling on B cell development, we performed B cell subpopulation phenotyping in patients and healthy donors. We noted a highly significant decrease of circulating plasmablasts (p < 0.0001), less significant decrease of transitional B cells (p = 0.029), and a slight but insignificant expansion of naïve B cells in MIS-C patients (Fig. 2A, B). Additionally, all MIS-C B cells had strikingly suppressed BAFF receptor (BAFFR) expression (Fig. 2C). This was true in all B cell subsets, including the naïve and transitional subsets.

In this study, we explored the role of type I and II interferons and the dysregulation of humoral immunity in MIS-C, drawing on parallels with other autoimmune diseases, such as SLE.

Our results point towards comparable type I interferon response between acute MIS-C and post-COVID-19 children without MIS-C, but demonstrate increased levels of type II interferon (IFN-γ) between these groups. Type II interferon has been described as a dominant feature of MIS-C previously [11, 12] and may reflect the concurrent T cell activation [1]. However, B cell-intrinsic IFNgR signaling has also been shown to evoke spontaneous generation of autoreactive germinal centers and its knock-out results in protection from systemic autoimmunity in animal models [19, 20]. Therefore, the increased IFN-γ levels may represent a contributing mechanism to the autoantibody induction in MIS-C. On the other hand, our data suggests that lingering type I and III interferon inflammation are not robust driving factors behind MIS-C pathogenesis, despite their role in the anti-SARS-CoV-2 immune response [17, 19].

The elevation of IFN-γ and the presence of autoimmune phenomena in MIS-C draws a parallel to SLE, which prompted the exploration of B cell immunity in our patients [13, 20]. Although we did not find any significant differences in the serum levels of the B cell survival-promoting cytokine APRIL between healthy donors, acute, convalescent MIS-C, and post-COVID children, we did observe a stark elevation of BAFF in acute MIS-C children and less so in healthy post-COVID children. Porritt et al. also found increased BAFF mRNA levels in MIS-C children [22]—this significant increase of BAFF levels in MIS-C is consistent with polyclonal B cell activation resulting in the spectrum of detected autoantibodies and clinical manifestations of MIS-C-associated autoimmune systemic and organ inflammation. In our cohort, of 35 patients in whom autoantibodies were tested, 4/35 (11%) had positive Ro60 antibodies and further 4/35 (11%) had positive extractable nuclear antigen (ENA) screening. Previous studies reported the presence of anti-La, a characteristic autoantigen of SLE and Sjogren’s disease, and anti-Jo-1, characteristic of idiopathic inflammatory myopathies, in MIS-C patients [21]. Although later works suggested that these autoantibodies may be derived from high-dose intravenous immunoglobulins [23], administered to the MIS-C patients, our samples were obtained prior to the treatment. The combination of increased IFN-γ and a consequent increase of BAFF has been previously described in SLE, again suggesting an analogy between MIS-C and SLE [24]. Interestingly, in the case of SLE, it has been shown that neutrophils can contribute to an increase in BAFF and a strengthening of the autoimmune process. In the case of MIS-C, hallmarked by marked neutrophilia and disturbed neutrophil functionalities [25, 26], a similar parallel might exist and contribute to the induction of autoimmunity [27].

Finally, we observed a shift in the maturation of B cells, skewing their subpopulations towards antibody-producing plasmablasts and away from the early transitional B cells in MIS-C children. Such reduction of plasmablasts is interesting given their lower reliance on BAFF for survival, which is also bolstered by APRIL [28]. Instead, naïve and transitional B cells, which are highly dependent on BAFF for survival [29], were both decreased in MIS-C children. However, at the same time, all MIS-C B cells had strikingly suppressed BAFF receptor (BAFFR) expression (Fig. 2C)—including naïve and transitional. These results are consistent with the similar situation observed in SLE, where high levels of BAFF are associated with decreased BAFFR expression on all B cell subtypes [30].

In summary, our brief report complements the current understanding of the dysregulation of humoral immunity and autoimmune phenomena seen in MIS-C and highlights the important role of the BAFF-BAFFR axis in their induction.