Enhanced Expression of Bruton’s Tyrosine Kinase in B Cells Drives Systemic Autoimmunity by Disrupting T Cell Homeostasis

A breakdown of B cell tolerance is thought to be a major pathogenic event in systemic autoimmune disorders, such as systemic lupus erythematosus (SLE) and Sjögren syndrome (1, 2). Autoantibodies directed against various nuclear self-antigens often appear in patient serum before the onset of clinical symptoms (3) and, particularly in SLE, form circulating immune complexes that elicit inflammation upon deposition in joints, eyes, skin, lungs, or kidneys (1, 4). Genetic studies indicate that B cell–intrinsic defects are central to autoimmune disease development (5–7). Susceptibility loci include genes implicated in BCR signaling (LYN, BLK, BANK1, PTPN22, and PXK) and NF-κB signaling. Furthermore, mouse models have shown that single-gene defects in BCR signaling molecules or inhibitory coreceptors can induce autoimmunity (reviewed in Ref. 8). Aberrant emergence, activation, and persistence of autoreactive B cells are thought to be dependent on other immune cells (9, 10). Whether and how autoreactive B cells propagate disease and to which extent they act through other cell types is largely unknown.

Conversely, follicular Th (Tfh) cells contribute to autoimmune disease development through activation of autoreactive B cells in germinal centers (GCs), resulting in the formation of plasma cells producing high-affinity autoantibodies (11–13). B cells are thought not to prime but to consolidate Tfh differentiation and support their function and expansion (14). For example, B cells guide Tfh responses through the production of IL-6 (15) or ICOS ligand (ICOSL) expression (16) and provide antigenic stimulation to Tfh cells when Ag availability in GCs is waning (17, 18). In contrast, Tfh cells are important producers of IL-21, a key cytokine involved in B cell activation. Dysregulated Tfh cell formation is found in various lupus-prone mouse models, including Sanroque mice harboring a mutation in the RNA-binding protein Roquin-1 (19) or BXSB-Yaa mice in which development of glomerulonephritis is critically dependent on IL-21R signals (20).

We recently observed that overexpression of the BCR signaling molecule Bruton’s tyrosine kinase (BTK) restricted to B cells in CD19-hBtk transgenic mice is sufficient to induce an SLE/Sjögren-like disease phenotype (21). Btk protein expression in mouse B cells is normally tightly controlled and upregulated upon BCR stimulation by complex regulatory mechanisms involving micro-RNAs and feed-forward NF-κB activation (22–24). B cells overexpressing BTK are selectively hyperresponsive to BCR stimulation, showing enhanced Ca²⁺ influx and NF-κB activation and resistance to Fas-mediated apoptosis in vitro, and are not effectively eliminated in vivo when autoreactive (21). A pathogenic role for BTK in rheumatic diseases is further emerging from mouse studies demonstrating that small-molecule BTK inhibitors prevent or ameliorate lupus nephritis and experimental arthritis (25–31). However, BTK inhibitors are likely to have effects beyond BCR signaling, because BTK functions downstream of many receptors, including chemokine and TLRs, and is also expressed in myeloid cells (32). Apart from direct effects on BTK-expressing cells, BTK inhibition was associated with diminished activation of both helper and cytotoxic T cells (25), raising the question how these T cells are indirectly activated in a BTK-dependent manner.

To elucidate how BTK signaling may engage pathogenic T cells in autoimmunity, here we investigated T cell activation, differentiation, and B–T cell interaction in aging CD19-hBtk mice. We found that increased BTK expression in mice supported spontaneous autoimmunity in vivo, both independently of T cells and by promoting the induction of Tfh cells. Tfh induction was also increased in Btk transgenic mice upon trinitrophenol keyhole limpet hemocyanin (TNP-KLH) immunization and collagen-induced arthritis (CIA) induction. However, affinity maturation was not affected. Interestingly, increased IL-6 production by B cells and IgG autoantibody formation was T cell dependent, as was increased CD86 expression by naive B cells. Taken together, these data demonstrate that enhanced BTK signaling in B cells can establish a T cell–driven proinflammatory loop resulting in autoimmune pathology, making BTK inhibition an attractive therapeutic strategy.

CD19-hBtk (33) and Cd40l^−/− (34) mice have been described. These mouse strains were backcrossed to the c57bl/6 genetic background for >10 generations. Mice were genotyped by PCR. Wild-type (WT) littermates were used as controls. Mice were kept at specified pathogen-free conditions at the Erasmus MC experimental animal facility, and experimental procedures were approved by the Erasmus MC committee for animal experiments.

Single-cell suspensions were prepared from mouse spleens, bone marrow, band lymph nodes using 100- μm cell strainers (BD Falcon) in MACS buffer (PBS/0.5% BSA/2 mmol EDTA). Fluorescent labeling of cell membrane molecules was performed as described previously (21). For isotype-specific staining of intracellular Ig in plasma cells, cells were fixed and permeabilized using Cytofix/Cytoperm Buffer and Perm/Wash Buffer, respectively (BD Biosciences, San Jose, CA), according to the manufacturer’s instructions. For the measurement of Foxp3 expression by CD4 T cells, cells were fixed and permeabilized with the eBioscience Foxp3 staining kit (eBioscience). For the measurement of intracellular cytokines, cells were fixed in PBS/2% paraformaldehyde, permeabilized, and stained in MACS buffer containing 0.5% saponin (Sigma-Aldrich). Gal-β (1–3)-GalNAc carbohydrates were stained with biotin-conjugated peanut agglutinin (Sigma-Aldrich), and indirect staining of biotin-conjugated Abs and reagents was performed with fluorochrome-coupled streptavidin (eBioscience). Details of the Abs used are available upon request.

Naive (CD3⁺CD4⁺CD25⁻CD62L⁺) T cells and total memory (CD3⁺CD4⁺CD25⁻CD62L⁻) T cells were sorted from spleens of 32-wk-old WT and CD19-hBtk mice, using an FACSAria equipped with BD FACS Diva software (BD Biosciences). Purity of the obtained populations was >96%. RNA was isolated from the cells using Qiagen RNeasy Micro Kit according to the manufacturer’s instructions. IL-21–specific primers were designed using the ProbeFinder software (Roche): 5′-CCATCAAACCCTGGAAACAA-3′ and 5′-TCACAGGAAGGGCATTTAGC-3′. Quantitative RT-PCR was performed using standard procedures and the ABI Prism 7300 setup (Applied Biosystems), and IL-21 expression values were normalized to GAPDH expression.

Mice were intradermally immunized with 100 μg chicken collagen type 2 (CII; Chondrex) emulsified in CFA. At day 21, a secondary intradermal immunization was performed by injecting 100 μg chicken CII emulsified in CFA. The evaluation of arthritis symptoms and severity was performed as described (35). Measurement of serum anti-chicken and anti-mouse CII IgG2c Abs by ELISA was performed as previously described (35).

Mouse serum and lymphoid organs were analyzed 7 d after i.p. injection of 0.1 mg TNP-KLH (Biosearch Technologies) and 0.8 mg Imject aluminum hydroxide adjuvant (Thermo Scientific) in PBS or 7 d after booster injection of 0.1 mg TNP-KLH in PBS given 6 wk after primary immunization. Serum IgG1-specific anti–TNP-KLH Ab levels were determined using a sandwich ELISA. To discriminate high- versus total affinity anti–TNP-KLH IgG1 Abs, plates were coated with TNP (5)-KLH and TNP (16)-KLH, respectively. After incubation of TNP-KLH–coated plates with diluted serum, bound TNP-KLH–specific IgG1 Abs were detected using goat anti-mouse IgG1 secondary Abs (Southern Biotechnology Associates).

Immunohistochemical stainings were performed as published (21). Sections were incubated with anti-CD3 (eBioscience) and anti-IgM Abs (BD Biosciences). Hep2 reactivity of IgM Abs was performed as described (21). For combination staining of IgM and IgG Abs, fluorescently labeled anti-IgM and anti-IgG F(ab)₂ fragments were used (Jackson ImmunoResearch Laboratories).

Plates were coated with DNA–cellulose, double-stranded from calf thymus DNA (Sigma-Aldrich). Serum samples were incubated in serial dilutions, and goat-anti-mouse biotinylated total IgG Ab (Southern Biotechnology Associates) was used to detect DNA-specific IgG Abs.

Significance of continuous data were calculated using the nonparametric Mann–Whitney U test. The p values <0.05 were considered significant.

We previously reported that 12-wk-old CD19-hBtk mice show spontaneous GC and increased plasma cell formation in the spleen and that aging CD19-hBtk mice develop anti-nuclear Abs and autoimmune pathology in kidney, lung, and salivary glands (21). To study the impact of BTK overexpression on the capacity of B cells to interact with T cells in aging mice, we first characterized the B cell compartment of CD19-hBtk mice and WT littermate controls at ∼30–33 wk of age. Total splenic B cell numbers were comparable between CD19-hBtk and WT mice (Fig. 1A). In line with our previous findings, proportions of CD95⁺IgD⁻ GC and programmed death-ligand 2 (PD-L2)⁺CD80⁺ memory B cells were significantly increased in the spleens of CD19-hBtk mice (Fig. 1B). Accordingly, CD138⁺ IgM⁺ and IgG⁺ plasma cells were also increased in spleen and bone marrow (Fig. 1C, 1D, respectively) (see gating strategy of B cell subsets and plasma cells in Supplemental Fig. 1A). In addition, the proportions of IL-6– and IL-10–producing B cells were significantly increased in spleens of CD19-hBtk mice (Fig. 1E). Interestingly, IL-10–producing B cells expressed higher levels of CD21 or CD5, indicating that many of these cells have a marginal zone (MZ) or B1 cell phenotype, whereas IL-6–producing B cells did not express these markers (Supplemental Fig. 1B). Similar to our previous findings in 12-wk-old mice (21), we noticed that surface expression of the costimulatory molecules CD80 and CD86 was increased on CD19-hBtk B cells compared with WT B cells (data not shown). Expression of ICOSL and the inhibitory molecule PD-L1 on GC B cells was comparable in CD19-hBtk and WT mice (Fig. 1F). Interestingly, ICOSL was significantly decreased on CD19-hBtk memory B cells (Fig. 1G), suggesting its downregulation as a result of recent T cell engagement, as described for memory B cells in patients with SLE (36).

Taken together, we observed: 1) increased formation of GC and memory B cells; 2) enhanced cytokine production by B cells; and 3) decreased ICOSL expression on memory B cells, all of which suggest that Btk overexpression stimulates B–T cell interaction.

Next, we studied the splenic T cell compartment in CD19-hBtk mice. Whereas at ∼8 wk of age, proportions of T cell subsets and cytokine production were comparable between CD19-hBtk and WT mice (data not shown), total numbers of CD4⁺ T cells were significantly increased in ∼30–33-wk-old CD19-hBtk mice (Fig. 2A). Moreover, CD19-hBtk mice showed significantly increased proportions of follicular T cells and Foxp3⁺ regulatory T cells (Tregs), both nonfollicular Tregs and follicular Tregs (Tfr; Fig. 2B). (Gating strategy shown in Supplemental Fig. 2.) Both programmed death-1 (PD-1)-high and PD-1-intermediate Tfh and Tfr cells were increased in CD19-hBtk mice. Proportions of residual CD4⁺ T cells, including Th1, Th2, and Th17 cells, were reduced. Expression of CTLA4, which transmits an inhibitory signal to T cells, was selectively increased on Tfh. Most interestingly, expression of ICOS, which is essential for Tfh survival and effective Th responses (37–40), was increased on all CD4⁺ T cell subsets in CD19-hBtk mice (Fig. 2C). Analysis of CD4⁺ T cells for intracellular cytokines after 4 h of PMA/ionomycin stimulation revealed that IFN-γ and IL-10 production were significantly increased in CD4⁺ T cells from CD19-hBtk mice (Fig. 2D). Also, CD8⁺ T cells manifested increased IFNγ production (data not shown). IL-21 mRNA expression, which was significantly higher in sorted memory CD4⁺CD25⁻CD62L⁻ T cells than in naive CD4⁺CD25⁻CD62L⁺ T cells, was increased in T cells from CD19-hBtk mice compared with nontransgenic littermates, reflecting the increased proportion of Tfh cells in these mice (Fig. 2E). Thus, there was no evidence that in CD19-hBTK mice the production of IL-21 per Tfh cell is increased.

Both excessive IFN-γ production and enhanced ICOS signaling have been shown to contribute to Tfh accumulation in lupus-prone Sanroque mice (19, 41). It is therefore conceivable that the capacity of Btk-overexpressing B cells to drive IFN-γ production by T cells and to promote expression of ICOS on the T cell surface supports their engagement in Tfh differentiation in vivo in aging CD19-hBtk mice.

Next, we aimed to directly investigate whether Btk-overexpressing B cells promote Tfh differentiation in young mice, both in the context of an in vivo immunization to a T cell–dependent Ag and in an autoimmune reaction based on molecular mimicry.

We studied Tfh cell induction and Ab affinity maturation in mice upon immunization with the T cell–dependent model Ag TNP-KLH. Mice were immunized with TNP-KLH in aluminum hydroxide and boosted 6 wk later with TNP-KLH in saline. At day 7 after the primary or secondary injection, we found higher numbers of GC B cells and Tfh cells in the spleens of CD19-hBtk mice, compared with WT controls (Fig. 3A, 3B for primary and secondary response, respectively). Interestingly, after the primary immunization, the sera contained increased levels of total anti-TNP IgG1 but comparable levels of high-affinity anti-TNP IgG1 Abs (Fig. 3C), indicating that low-affinity Abs were increased in CD19-hBtk mice. This would be consistent with the observed increased survival of Btk overexpressing B cells upon BCR stimulation in vitro (21). Despite this increase of low-affinity IgG1 Ab formation in the primary response in CD19-hBtk mice, we found that after the boost, serum levels of both total and high-affinity anti-TNP IgG1 Abs were comparable between the two groups of mice (Fig. 3D). In these experiments, we did not observe differences in TNP-specific IgM levels between WT and CD19-hBtk mice in the primary nor secondary response (Supplemental Fig. 3). Thus, an aberrant primary and secondary IgG1 response in CD19-hBtk mice, characterized by increased formation of GC B cells and Tfh cells, nevertheless allowed apparently normal affinity maturation and plasma cell differentiation after secondary immunization.

To investigate the effects of Btk overexpression on local Tfh differentiation in draining lymph nodes in an autoimmune model, we performed CIA experiments. Eight-week-old CD19-hBtk and WT mice were immunized with emulsified chicken CII in CFA and boosted at day 21. At day 38, the peak of disease severity, total splenic B cell numbers were comparable in CD19-hBtk and WT mice, but proportions of GC B cells and IgM⁺ or IgG⁺ plasma cells were increased in CD19-hBtk mice (Fig. 4A, 4B). Moreover, in CD19-hBTK mice, the proportions of Tfh cells were increased in spleens, although not significantly, and in popliteal lymph nodes (p < 0.05), compared with WT mice (Fig. 4C). Also, the expression of IFN-γ and IL-17A in CD4⁺ T cells was significantly increased in lymph nodes (Fig. 4D). Despite the enhanced T cell activation and Tfh differentiation, CD19-hBtk mice exhibited no detectable increase in disease severity and only a slight increase in disease incidence (Fig. 4E). Accordingly, serum levels of both anti-chicken CII IgG2c and autoreactive anti-mouse CII IgG2c Abs were comparable between CD19-hBtk and WT mice (Fig. 4F).

From these findings, we conclude that Btk overexpression in B cells induces enhanced Tfh differentiation in vivo, but does not substantially hamper affinity maturation nor augment autoantibody responses to collagen.

To determine the contribution of T cells to the autoimmune pathology in CD19-hBtk mice, we crossed these mice on a CD40L-deficient background. CD40L is expressed on activated T cells and engages CD40 on B cells, thereby facilitating T–B cell interaction, and is essential for the formation of GCs (42). Total B cell numbers in the spleens of ∼30–33-wk-old Cd40l^−/− CD19-hBtk mice were slightly decreased compared with WT, CD19-hBtk or Cd40l^−/− mice (Fig. 5A). Proportions of immature, follicular, and MZ B cells were comparable in the four mouse groups (data not shown; Supplemental Fig. 4A). Interestingly, IL-6–expressing and IFN-γ–expressing B cells, but not IL-10–expressing B cells, were restored to normal proportions in Cd40l^−/− CD19-hBtk mice (Fig. 5B). This would be consistent with the finding that many of the IL-10–expressing B cells have an MZ or B-1 phenotype (Supplemental Fig. 1B), which are B cells involved in T cell–independent responses (43). Also, CD86 expression on IgD⁺CD95⁻ naive B cells was increased in CD19-hBtk mice compared with WT, but not in Cd40l^−/− CD19-hBtk mice (Fig. 5C). This finding suggests an interaction of T cells with naive B cells. Indeed, enhanced cytokine production by T cells or differentiation into both PD-1-high and PD-1-intermediate follicular T cells was not observed in aging Cd40l^−/−CD19-hBtk mice (Fig. 5D, 5E).

As expected, GC B cells were lacking in Cd40l^−/− and Cd40l^−/− CD19-hBtk mice (Fig. 6A). The increase in the proportions of memory B cells and IgG⁺ plasma cells, as seen in CD19-hBtk mice, was absent in Cd40l^−/− CD19-hBtk mice (Fig. 6A, 6B). However, the numbers of IgM⁺ plasma cells in the spleens of Cd40l^−/− CD19-hBtk mice were comparable to those in CD40L-expressing CD19-hBtk mice and were significantly increased compared with WT or Cd40l^−/− control mice (Fig. 6B, Supplemental Fig. 4B). The autoimmune phenotype of CD19-hBtk mice featured IgM and IgG Ab deposition in the kidneys (21). In contrast, histological analysis of the kidneys from Cd40l^−/− CD19-hBtk mice revealed significant IgM, but no IgG deposition, nor thickening of glomerular membranes (Fig. 6C and data not shown). Accordingly, serum IgM from these mice clearly displayed more autoreactivity in a Hep2 assay than did IgM from WT or Cd40l^−/− mice, revealing a cytoplasmic staining pattern comparable to IgM from CD19-hBtk mice (Fig. 6E). In contrast to our previous findings in CD19-hBtk mice (21), no autoreactive IgG Abs were found in serum of Cd40l^−/− CD19-hBtk mice by Hep2 staining, and anti-DNA IgG Abs in serum were absent in Cd40l^−/− CD19-hBtk mice (Fig. 6F, 6G). Total serum IgM in CD19-hBtk and Cd40l^−/− CD19-hBtk mice was not increased compared with WT mice (data not shown). Infiltration of various organs, including lung, kidney, and salivary gland by immune cells, was found in CD19-hBtk (21) but not in Cd40l^−/− CD19-hBtk mice, indicating that the autoimmune pathology induced by BTK overexpression in B cells fully dependent on B–T cell interaction (shown for salivary glands in Fig. 6D).

Taken together, these data revealed that overexpression of Btk in B cells resulted in defects that were partly T cell independent, such as increased IL-10 expression by B cells and elevated levels of circulating autoreactive IgM Abs, and partly T cell dependent, including increased CD86 expression on naive B cells, increased GC and memory B cell formation, enhanced IL-6 production, autoreactive IgG Abs in serum, IgG Ab deposition in kidneys, and immune infiltration of target organs.

Various mouse models and patients with primary immunodeficiencies, including X-linked agammaglobulinemia resulting from Btk deficiency, show that B cells play a crucial role in the generation of Tfh cells (13, 44). Conversely, we now provide evidence that B cells with elevated Btk protein expression have the capacity to disrupt T cell homeostasis and induce excessive Tfh differentiation associated with spontaneous GC formation and autoimmunity. In aging mice, Btk-overexpressing B cells engage T cells in a positive-feedback loop that establishes T cell–propagated systemic autoimmune pathology. Btk overexpression induces IL-6 production in B cells in vivo (through a CD40L-dependent pathway) and promotes IFN-γ production and surface ICOS expression by T cells, as well as the formation of Tfh cells.

Mouse crosses on the Cd40l^−/− background showed that the anti-nuclear IgG associated autoimmune pathology in aging CD19h-hBtk mice required active involvement of T cells. Autoimmunity develops despite enhanced Treg and Tfr cell differentiation and increased IL-10 production by both B and T cells. Apparently, these Tfr cells, which are thought to be responsible for limiting the GC response (45, 46), fail to inhibit reciprocal activation of B and Tfh cells in aging CD19-hBtk mice. The inhibitory function of Treg and Tfr cells requires surface CTLA4 expression (47, 48), which in CD19h-hBtk mice was comparable to WT mice. Also, CTLA4 on the surface of Tfh cells has the capacity to inhibit B cell function (47, 48). Because CTLA4 expression was even increased in CD19-hBtk mice (Fig. 2C), we conclude that Btk overexpression in B cells reduces their susceptibility to inhibitory signals from T cells. This could be explained by high expression of CD80 and CD86 on B cells induced by Btk overexpression (21), because it has been shown that transfer of CD80- and CD86-expressing B cells alone was sufficient to induce Tfh formation (47).

Induction of Tfh by CD19-hBtk B cells in the context of spontaneously arising GCs, as seen in aging mice, could be enhanced in young animals, either by immunization with a model Ag or in a CIA autoimmune model. Primary immunization with the model Ag TNP-KLH resulted in higher serum levels of low-affinity anti-TNP Abs in CD19-hBtk mice than in WT mice. Because CD19-hBtk B cells are more resistant to FAS-mediated apoptosis (21), it is likely that B cells harboring low-affinity BCRs that would normally be counterselected in a GC response can survive. However, this did not lead to enhanced formation or maintenance of low-affinity memory B cells, as levels of both total and high-affinity anti-TNP IgG1 Abs were comparable after secondary immunization. Thus, repertoire selection after secondary immunization did not appear to be affected by the relative resistance to FAS-mediated apoptosis of CD19-hBtk transgenic GC B cells, showing that low-affinity CD19-hBtk transgenic B cells still had a selective disadvantage when competing with high-affinity Ag-binding CD19-hBtk transgenic B cells. Likewise, in the CIA experiments, the Tfh numbers in draining lymph nodes were increased in CD19-hBtk mice, but this was not associated with higher disease scores or enhanced anti-mouse CII autoantibody formation. From these findings, we conclude that affinity maturation in CD19-hBtk mice is unperturbed and that the induced Tfh cells are functional in BCR repertoire selection. We hypothesize that Tfh cells in CD19-hBtk mice may recognize peptides from nucleic acid–associated proteins, by which the formation of pathogenic anti-nuclear Abs depends on TLR signaling in B cells, as was previously shown for the extrafollicular plasmablast response in MRL.fs (lpr) mice (49). In addition, the increased GC B cell and Tfh development in CD19-hBtk mice depend on TLR signaling, as numbers were restored to WT levels after crosses with MyD88-deficient mice (O.B.J. Corneth and R.W. Hendriks, unpublished observations). An efficient GC-dependent autoreactive BCR repertoire selection mechanism in aging mice would convert specificities of circulating Abs from quite weak anti-cytoplasmic IgM (as also seen in Cd40l^−/− mice) into strong anti-nuclear IgG autoantibodies. The patterns of Hep2 reactivity that we found for IgM Abs suggest that they bind to ribosomes, similar to patterns found in patients with SLE.

Although CD40L is primarily expressed on activated CD4⁺ T cells and mostly studied because of its prominent role B–T cell interaction, CD40L is also present on several other cell types, including activated B cells, platelets, dendritic cells (DCs), and smooth muscle cells (50). It is therefore conceivable that, next to B–T cell interaction, additional immunological pathways involved in autoimmunity may be hampered in Cd40l^−/− mice. In this context, it is relevant that it has been shown that DCs can express both CD40 and CD40L and have the capacity to directly influence various B cell processes, including proliferation, differentiation, and Ig class-switch recombination (50–52). However, although MZ DCs were found to induce robust T cell–dependent Ab responses, they did not induce GCs or long-lived B cell responses but rather invoke T cell–dependent extrafollicular Ab responses after antigenic stimulation (52). Therefore, direct DC–B cell interactions do not likely explain the phenotype of CD19-hBTK Tg mice that is characterized by spontaneous GC formation.

Taken together, our findings support a model in which Btk overexpression induces spontaneous GCs and autoimmune pathology by upregulation of CD80 and CD86, which promotes Tfh cell formation and at the same time obstructs CTLA4-dependent regulation by Tregs and Tfr cells. As a result of ongoing B–Tfh cell interaction, B cells increase IL-6 production, and T cells exhibit excessive IFN-γ production and surface expression of ICOS, which lead to pathogenic accumulation of GCs and plasmablasts that produce anti-nuclear autoantibodies.

To date, several studies investigating the efficacy of BTK inhibitors in rodent models for autoimmune diseases, including SLE and rheumatoid arthritis, have shown a beneficial effect on disease progression (25–29). Although effects of BTK can partly be explained by direct effects on B cells and autoantibody production, BTK inhibition clearly affects other cell types, including monocytes, macrophages, and mast cells—for example, through disruption of FcR signaling in myeloid cells. However, it is conceivable that interference with B–T cell interaction may also significantly contribute to a therapeutic effect of BTK inhibition. The observed response in patients with SLE and progressive systemic sclerosis to BAFF/BLyS-neutralizing Ab belumimab further implies that targeting B cell survival and activation is an effective treatment strategy (53, 54). Interestingly, in human B cells, BTK protein levels increase upon activation in vitro, and BTK levels are increased in human patients with primary Sjögren syndrome in vivo (O.B.J. Corneth and R.W. Hendriks, unpublished observations). Because the BTK inhibitors ibrutinib and acalabrutinib, which show high response rates as a monotherapy in patients with various B cell malignancies, are generally well tolerated (55, 56), targeting BTK could be a promising new therapeutic strategy in the treatment of systemic autoimmune diseases.

In summary, we have established that increased expression of Btk is sufficient to induce excessive Tfh differentiation and spontaneous GC formation associated with the production of pathogenic anti-nuclear autoantibodies. We conclude that in autoimmune disease, BTK-mediated signaling in B cells establishes or maintains T cell–propagated pathology, making BTK an attractive therapeutic target.

We thank Ilke Ilgaz, Saravanan Yuvaraj, Lars van Greuningen, and Arndt Krause for excellent technical assistance.

The authors have no financial conflicts of interest.