Butyrophilin 2A2 promotes T cell immunoregulation via CD45 phosphatase activation and protects against murine autoimmune glomerulonephritis and pregnancy loss

Abstract

B7 costimulatory family member Butyrophilin 2A2 (BTN2A2) is predominantly expressed by antigen presenting cells and regulates T cell immunity, but molecular mechanisms are unclear. Using immunoblots analyzing TCR-initiated signaling intermediaries, co-immunoprecipitation studies, confocal microscopy, structural modeling-guided mutational analyses, and microscale thermophoresis, we demonstrate that BTN2A2 directly interacts with CD45RO, resulting in CD45 retention within the immune synapse during TCR activation. Recombinant BTN2A2 increases murine CD4+Foxp3+ regulatory T cells (Treg) and reduces T helper 17 (Th17) cells in vitro through mechanisms dependent on CD45 phosphatase activity. BTN2A2 therapy alleviates disease severity in murine nephrotoxic glomerulonephritis and autoimmune miscarriage while increasing Treg/Th17 ratios. Analyses of BTN2A2-deficient animals show exacerbation of disease associated with reduced Treg/Th17 ratios. BTN2A2 functions analogously on human T cells suppressing Th17, Th1 and Th2 responses while inducing Tregs. Together, our studies identify BTN2A2 as a modulator of CD45RO signaling in T cells, providing insight into how BTN2A2 regulates T cell-dependent immune responses including those mediating autoimmunity and transplant rejection.

Butyrophilin 2A2 is a member of the B7 costimulatory family that is expressed on antigen presenting cells and is linked to the regulation of T cells. Here the authors implicate butyrophilin 2A2 in enhancement of CD45 phosphatase activity within the immunological synapse during T cell activation, leading to expansion of regulatory T cells and reduction of proinflammatory Th17 CD4 T cells.

Introduction

T cells recognize antigens presented through peptide:MHC complexes using surface-expressed, heterodimeric αβT cell receptors (TCRs). As the TCR lacks intrinsic kinase activity, signals initiated by TCR ligation involve recruitment of the Src family kinase Lck, which then phosphorylates immunoreceptor tyrosine-kinase-based motifs (ITAMs) within the TCR-associated CD3 ζ-chain. Lck also phosphorylates subsequently recruited Zap70 kinase, thereby propagating requisite downstream signals required for full T cell activation. Studies performed over the past 30 years showed T cell activation is controlled in part by cell surface expressed CD45¹.

CD45 is a transmembrane glycoprotein that contains an intracellular tyrosine phosphatase domain capable of dephosphorylating multiple TCR immunoreceptor tyrosine activation (ITAM) motifs. Differential splicing results in the expression of multiple CD45 isoforms (i.e., RA, RB, RC, RO)². Following TCR stimulation, CD45 is initially recruited to the supramolecular activation cluster (SMAC) but is then expelled, segregating it from the TCR^3,4. Evidence suggests that this segregation of CD45’s phosphatase activity from the TCR is essential for Lck-initiated signal propagation that results in full T-cell activation. Conversely, retention of CD45 within the T cell immune synapse (IS) regulates the strength and duration of TCR activation^2,5, and perturbations of CD45 activity contribute to the development of autoimmune disease^6,7. Despite decades of work by multiple groups delineating these molecular mechanisms, it is not known how CD45 segregation versus retention during TCR activation may be regulated, or if ligands co-presented by antigen-presenting cells (APCs) play a role.

Butyrophilins (BTNs) are glycoproteins originally isolated from breast milk that have imprecisely understood immune-regulatory effects and are implicated in maintaining maternal-fetal tolerance^8–10. The mRNAs encoding for BTN and BTN-like molecules are widely expressed in lymphoid and non-lymphoid tissues¹¹. Butyrophilin immunoglobulin domains exhibit structural similarities to the B7 family of co-receptors, including B7-1/CD80, B7-2/CD86, ICOS-L, and PD-L1, and Butyrophilin 2A2 (BTN2A2) was previously shown to be expressed by professional APCs, including B cells, macrophages, and dendritic cells (DCs)¹². Results of in vitro studies suggest that BTN2A2 can modulate T cell receptor (TCR) signaling and promote de novo Foxp3 expression^12,13. Mice genetically deficient in BTN2A2 exhibit impaired CD4+ regulatory T cell function, potentiated anti-tumor immunity, and augmented clinical manifestations of experimental autoimmune encephalomyelitis, all of which were attributable to the deficiency of BTN2A2 in APCs¹⁴. While these cumulative findings implicate a key immunoregulatory function for BTN2A2, the exact molecular mechanisms underlying these effects remain unclear.

Herein, we demonstrate that BTN2A2 functions as a ligand for CD45, binding both to the TCR complex and CD45RO isoform on T cell surfaces, resulting in retention of CD45 phosphatase activity proximal to the TCR complex. Consequently, the enhanced phosphatase activity reduces downstream TCR signaling, which in turn promotes regulatory T cell (Treg) expansion, suppressing T effector cell differentiation. We further demonstrate the significant immunomodulatory effects of BTN2A2 in a mouse model of autoimmune glomerulonephritis, showing that BTN2A2 genetic deficiency exacerbated kidney injury, whereas administration of recombinant BTN2A2 limited disease severity. Furthermore, in a second mouse model of immune-mediated spontaneous abortion, treatment of pregnant dams with recombinant BTN2A2 helped animals tolerate their pregnancies, reducing fetal loss.

Results

BTN2A2 blocks CD3-dependent signaling in Jurkat Cells

To elucidate the mechanisms underlying BTN2A2’s ability to regulate T cell immunity^13,14, we generated a soluble recombinant human BTN2A2-Fc fusion protein using baculoviral expression systems. Consistent with prior reports performed in murine systems¹³, recombinant human BTN2A2-Fc, but not recombinant Fc control protein, blocked anti-CD3-induced IL-2 production by Jurkat cells (Fig. S1A–C).

When we next evaluated signaling components downstream of the TCR, we observed that BTN2A2-Fc reduced anti-CD3-induced phosphorylation of Zap70 and CD3ζ in Jurkat cells using western blot analysis (Fig. S1D) and reduced phosphorylation of Zap70, Akt, PI3kinase, and S6 ribosomal protein on phosphoflow analysis (Fig. S2). Addition of the phosphatase inhibitor pervanadate resulted in hyperphosphorylation of Zap70 and CD3ζ at baseline (unstimulated), and in contrast to the findings in the absence of pervanadate, addition of BTN2A2-Fc to the culture had no effect on the hyperphosphorylated proteins (Fig. S3). Together, these data suggested that BTN2A2-Fc activates a phosphatase that downregulates TCR signaling.

BTN2A2 blocks TCR signaling by binding to and enhancing CD45 phosphatase

Based on the known molecular mechanisms linking CD45’s phosphatase activity to TCR signaling^3,4, we tested the hypothesis BTN2A2’s inhibitory effect on TCR activation is mediated through interaction with CD45. Western blotting of CD45-associated proteins after co-immunoprecipitation (co-IP) of CD45 in unstimulated Jurkat cells demonstrated that in the absence of TCR activation, CD45 was normally associated with several components of the TCR complex (Zap70 and CD3ζ). However, activation of the TCR complex with anti-CD3 antibodies resulted in CD45 dissociating from Zap 70 and CD3ζ. In contrast, TCR activation in the presence of BTN2A2 caused CD45 to remain associated with the TCR complex (Fig. 1A). Similarly, when we performed co-IP experiments with Zap70 pulldown in activated Jurkat cells treated with BTN2A2, CD45 associated with Zap70 (Fig. S4A, B).

Fig. 1. BTN2A2-Fc enhances interaction and co-localization of CD45 with TCR signaling proteins.

A Jurkat cells were stimulated for 3 min with immobilized anti-CD3 antibody (10 μg/ml) in the presence or absence of BTN2A2-Fc (10 μg/ml). Cells were lysed in IP buffer, immunoprecipitated with anti-CD45 antibody, and immunoblotted for total Zap70 and CD3ζ. Input is 5% of the cell lysate. Right panel shows intensity plots depicted as Mean ± S.D from 5 independent experiments for Zap 70 and 4 independent experiments for CD3ζ; Kruskal–Wallis with Dunn’s multiple comparison test. B Immunostaining and confocal microscopy analysis shows segregation of CD45 from CD3ζ in Jurkat cells treated with or without immobilized BTN2A2-Fc protein (10 μg/ml) and activation by plate-bound anti-CD3/anti-CD28 antibody (10 μg/ml each) for 3 min. Right panel shows quantification of CD3ζ and CD45 colocalization in the presence or absence of BTN2A2-Fc protein from multiple fields (N = 8 untreated, N = 10 activated with anti-CD3/CD28, N = 9 for BTN2A2 treated group) acquired from two independent experiments. More than 500 cells from each group were included in the analysis shown in the right panel, and data is presented as Mean ± S.D; Kruskal–Wallis with Dunn’s multiple comparison test. Magnification 63x resolution with Zeiss “Immersol” immersion oil (refractive index of 1.518). C High-resolution immunostaining analysis showing single-cell orthogonal sections (z-stacks) in Jurkat cells with the same condition as in Fig. 1B. D Co-immunoprecipitation experiments were performed using anti-CD3ε antibody in Jurkat cells with the same condition as in (A), followed by immunoblot with anti-CD45 antibody. Input is 2% of the cell lysate. E CD45-specific phosphatase activity measured in immunoprecipitate from samples in D using FDP (Fluorescein Diphosphate, Tetraammonium Salt) substrate as described in the methods. Summary data depicted as mean ± S.D from five independent experiments; Kruskal–Wallis with Dunn’s multiple comparison test.

Cellular imaging with confocal microscopy confirmed colocalization of CD45 and CD3ζ in untreated cells at baseline, and segregation of CD45 following anti-CD3/anti-CD28-induced TCR activation. CD3ζ staining following TCR activation was largely perinuclear as previously noted¹⁵. Interestingly, we noted maintenance of CD45 and CD3ζ colocalization when cells were activated in the presence of BTN2A2, suggesting that BTN2A2 prevents the exclusion of CD45 from the immune synapse during the early phases of T cell activation. (Fig. 1B, C).

Co-IP experiments additionally demonstrated that compared to unstimulated cells, anti-CD3 activation reduced the association of CD3ε with CD45, whereas TCR activation in the presence of BTN2A2 augmented the interaction between CD45 and CD3ε (Fig. 1D). Functional experiments looking at TCR-associated CD45 phosphatase activity showed that when TCR complexes were isolated by co-IP with anti-CD3ε antibodies, CD45 phosphatase activity was detectable at baseline, decreased following TCR activation, and maintained when cells were activated in the presence of BTN2A2, suggesting segregation of CD45 from the TCR complex after activation and retention in the presence of BTN2A2^3,4 (Fig. 1E). Together with prior studies that activating CD45 phosphatase on lipid microdomains on T cell surfaces results in decreased sensitivity of TCR-mediated signaling⁵, our new data indicate BTN2A2 dampens TCR signaling by activating CD45 phosphatase activity in the TCR complex which in turn blocks critical phosphorylation events downstream of the TCR.

BTN2A2 interacts with CD45 phosphatase

We next employed a co-IP experimental strategy using Jurkat cells to test for direct interactions between CD45 and endogenously expressed BTN2A2 (Fig. 2A). These assays showed that CD45 co-immunoprecipitated with endogenous BTN2A2 in unstimulated Jurkat cells, and binding was significantly enhanced in activated Jurkat cells. Anti-CD3 activation of Jurkat cells did not change endogenous CD45 and BTN2A2 expression levels (Fig. S5A). Complementary experiments with exogenous BTN2A2-Fc or Fc protein control confirmed that exogenous BTN2A2-Fc but not Fc-tag protein binds to CD45 in activated T cells (Fig. 2B, C). This interaction between CD45 and BTN2A2 in activated T cells was specific as other abundantly expressed cell surface proteins, such as CD43, failed to interact with BTN2A2 under similar conditions as above (Fig. S5B).

Fig. 2. BTN2A2 interacts with CD45 phosphatase.

A Jurkat cells were immunoprecipitated with anti-BTN2A2 antibody or IgG and blotted with CD45 antibody at the unstimulated state or after activation using anti-CD3 and anti-CD28 antibodies (1 μg/ml each) for 48 h. Input is ~2% of cell lysate. Right panel show intensity plots depicted as Mean ± S.D from four independent experiments. B Jurkat cells activated with anti-CD3/anti-CD28 (1 μg/ml each) for 48 h were incubated with BTN2A2-Fc or Fc, and following Protein-G pulldown, eluted proteins were immunoblotted with anti-CD45 antibody. One representative experiment out of two experiments is depicted. C Jurkat cells activated with anti-CD3/anti-CD28 antibody (1 μg/ml each) were incubated with BTN2A2-Fc or Fc and crosslinked with BS3. Cell lysate was immunoprecipitated with protein G, and the eluted proteins were immunoblotted with anti-CD45 antibody. One representative blot out of two is depicted. D Ribbon diagram representation of protein-protein interaction between BTN2A2 (pink) and CD45 (PTPRC)(white), showing potential amino acids (blue sticks) determining their interaction. Two residues, N419 and N468, key glycosylation sites located in the fibronectin domain of CD45 that is critical for the interaction, is indicated in the spherical model. E Western blot from one experiment demonstrating expression of CD45 in wild type Jurkat cells (lane 1), Jurkat cells with CD45 knock-down (CD45-KD) (lane 2), CD45 expression in CD45-KD cells in which wild type CD45 (lane 3) or mutant CD45 (double mutant - N419A and N468A) (lane 4) is expressed. GAPDH shown as internal control. F Co-immunoprecipitation of CD45 with BTN2A2-Fc or Fc in activated Jurkat cells lacking CD45 in which wild type CD45 or mutant CD45 (double mutant - N419A and N468) were expressed as described in (E). Following Protein-G pulldown, eluted proteins were immunoblotted with anti-CD45 antibody or BTN2A2 antibody. Right panel shows summary data as Mean ± S.D from 5 independent experiments. (G)Co-immunoprecipitation of CD45 with BTN2A2-Fc or Fc in activated Jurkat cells (anti-CD3/anti-CD28) treated with or without PNGase. Following Protein-G pulldown, eluted proteins were immunoblotted with anti-CD45 antibody or BTN2A2 antibody. One representative experiments out of three is depicted. All statistical analysis done using Mann–Whitney U test (two-tailed).

We then performed molecular modeling to identify regions in the extracellular region of CD45 that might interact with BTN2A2. Protein homology modeling showed that BTN2A2 interacted with the extracellular fibronectin domain of CD45 protein and in particular amino acids—Asn-419 and Asn-468 (putative N-glycosylation sites based on their location within N-X-S/T consensus motifs) were critical for this interaction (Fig. 2D and Fig. S5C). It should be noted that the glycosylation of CD45 at N419 was previously reported by mass spectrometry¹⁶; interestingly, this glycosylation was detected in activated Jurkat cells, but not in non-activated cells. We then performed site-directed mutagenesis of asparagine to alanine and confirmed that mutant CD45 interaction with BTN2A2 was significantly attenuated (Fig. S5D, E and Fig. 2E, F). To confirm that N-glycosylation sites on CD45 were critical for the interaction, we treated Jurkat cells with endoglycosidic enzyme PNGase F (N-Glycosidase F) (Fig. S5F). As predicted, PNGase treatment significantly attenuated the interaction between CD45 and BTN2A2 in activated Jurkat cells (Fig. 2G).

Based on our observations that BTN2A2 showed enhanced interaction with CD45 in activated T cells, we suspected that BTN2A2 was preferentially binding with the CD45RO isoform, because CD45RO is predominantly expressed in activated T cells² and is upregulated in Jurkat cells after activation (Fig. S5G–I). We then repeated the co-immunoprecipitation assays and immunoblotted the eluted proteins using CD45 isoform-specific antibodies (Fig. 3A). These experiments confirmed that BTN2A2 binds CD45RO isoform (the predominant isoform in activated T cells), but did not bind CD45RA. Further, cell surface binding studies confirmed that CD45RO expression goes up during Jurkat cell activation (Fig. 3B). Treatment of activated Jurkat cells with anti-CD45RO antibody, but not anti-CD45RA antibody, inhibited cell surface binding of BTN2A2-Fc to activated Jurkat cells (Fig. 3C).

Fig. 3. BTN2A2 preferentially binds to cell surface CD45RO on T cells.

A Western blot analysis of CD45 isoforms (CD45RO and RA) following immunoprecipitation with protein G agarose beads of lysates obtained from activated Jurkat cells (1 μg/ml of anti-CD3 and anti-CD28 antibodies for 48 h) in the presence of recombinant BTN2A2-Fc or Fc. Input was used as 2% of total cell lysate. Right panel shows quantification of CD45RO and CD45RA co-immunoprecipitation depicted as Mean ± S.D from four independent experiments; Mann–Whitney U test (two-tailed); (B) CD45RO expression in resting Jurkat cells and activated Jurkat cells (1 μg/ml of anti-CD3 and anti-CD28 antibodies for 48 h). One of two experiments is depicted. C Binding of BTN2A2-Fc or Fc alone on activated Jurkat cells (1 μg/ml of anti-CD3 and anti-CD28 antibody for 48 h). Preincubation of activated Jurkat cells with anti-CD45RO antibody, but not anti-CD45RA antibody, inhibited the binding of BTN2A2-Fc on Jurkat cells. One out of two experiments is depicted. D MST analysis of the BTN2A2 binding interaction with CD45RO-Fc. The MST dose response data is obtained by titrating recombinant CD45RO-Fc protein (8 μM to 244 pM) against 20 nM fluorescent labeled recombinant BTN2A2 protein. The MST data analysis show that BTN2A2 interacts with CD45RO-Fc with a binding constant (K_D) of 63.5 nM. The data is represented as Mean ± S.D of triplicates of one independent experiment, and the experiment was repeated three times.

To demonstrate direct binding of BTN2A2 to CD45RO, we performed microscale thermopheresis (MST) using recombinant BTN2A2 and CD45RO, which demonstrated specific high-affinity binding of these two proteins with a Kd of 63.5 ± 14.7 nM (Fig. 3D).

These studies confirm that BTN2A2 specifically and directly binds to CD45RO on activated T cells and posit that activation of CD45RO mediates the functional effects reported for BTN2A2.

BTN2A2 enhances regulatory T cell expansion and suppresses Th17 cell populations in primary mouse T cells

Activated CD4 + T cells differentiate into several effector cell subsets based on activation events and the cytokine milieu present during activation. Since it was previously reported that BTN2A2 induced Foxp3 expression in CD4+ T cells¹³ in short-term culture studies, we tested whether T cells activated in the presence of BTN2A2 increased Treg populations in splenocytes from Foxp3-EGFP mice that co-express eGFP when Foxp3 is expressed. We cultured murine T cells (Foxp3+ reporter mice -BALB/cJ strain) in the presence of irradiated, allogeneic mouse splenocytes (CD1 mouse strain) with or without recombinant BTN2A2-Fc. These assays showed a 50% increase in CD4+CD25+Foxp3+ T cells, in the presence of BTN2A2-Fc compared to controls, a similar magnitude to that observed in positive control cultures using exogenous TGF-β (Fig. 4A and Fig. S6A, B). Consistent with published data¹³, we then confirmed that incubation of flow sorted, naïve CD4+Foxp3-GFP-CD62L + T cells with recombinant BTN2A2-Fc in anti-CD3 enhanced MLR reaction induced de novo Treg differentiation (Fig. S7A). Interesingly, recombinant BTN2A2-Fc did not enhance proliferation of preformed Tregs (Fig. S7B). Parallel assays showed that recombinant BTN2A2 blocked IL-6/TGF-β/IL-1β-induced production of RORγt, the signature transcription factor of Th17 cells and IL-17 (Fig. 4B and Supplementary Data Figs S8–9). Interestingly, short-term exposure of BTN2A2 to activated CD4⁺ T cells led to less proliferation and survival, phenotypes noted with a pro-differentiation state and similar to what has been reported with TGF-β¹⁷ (Fig. 4C-E and Fig. S10).

Fig. 4. BTN2A2-Fc enhances Tregs and suppress Th17 cell differentiation in mixed lymphocyte reaction (MLR).

A Flow cytometry analysis plot of primary CD4 + T cells (from spleen and lymph nodes of Foxp3-GFP transgenic mice) incubated with bound 1 μg/ml anti-CD3 and in the presence or absence of 10 μg/ml BTN2A2-Fc fusion protein in MLR for 7 days as described in methods and analyzed for CD4 + CD25 + GFP-Foxp3+ cells expression. As a positive control, effects of TGF-β (1 ng/ml) under the same conditions are also depicted. Right panel show summary plot depicted as Mean ± SD from 4 independent experiments; Kruskal–Wallis with Dunn’s multiple comparison test. B Flow cytometry analysis of primary CD4 + T cells (isolated from murine spleen and lymph nodes) incubated for 5 days with anti-CD3 (0.5 μg/ml) antibody alone or anti-CD3 + TGF-β (1.5 ng/ml) +IL6 (10 ng/ml) +IL-1β (10 ng/ml) and/or recombinant BTN2A2-Fc (10 μg/ml) in MLR and analyzed for CD4+RORγt+ cells. Right panel show summary plot depicted as Mean ± SD (N = 9 independent experiments); Kruskal–Wallis with Dunn’s multiple comparison test. C Flow cytometry plots of CellTrace violet stained purified mouse CD4 + T cells incubated with immobilized anti-CD3 and anti-CD28 antibodies (0.5 μg/ml) in the presence of immobilized Fc (10 μg/ml) or recombinant BTN2A2-Fc (10 μg/ml) for 3 days. Right panels show bar graph depicted as a percentage of non-proliferated cells (G0); Data is depicted as Mean ± SD (N = 4 independent experiments); Mann–Whitney U test (two-tailed). D Bar graph shows IL2 positive purified mouse CD4 + T cells incubated with immobilized anti-CD3 and anti-CD28 antibodies (0.5 μg/ml) in the presence of immobilized recombinant Fc (10 μg/ml) or BTN2A2-Fc (10 μg/ml) for 3 days. Data is depicted as Mean ± S.D (N = 4 independent experiments); Mann–Whitney U test (two-tailed). E Bar graph shows the percentage of live cells after incubation with immobilized anti-CD3 and anti-CD28 antibodies (0.5 μg/ml) in the presence of Fc (10 μg/ml) or recombinant BTN2A2-Fc (10 μg/ml) for 3 days. Data is depicted as Mean ± S.D (N = 4 independent experiments); Mann–Whitney U test (two-tailed). F Bar graph of GFP-Foxp3+ cell (%) population in total CD4 + T cells co-cultured with dendritic cells (DC) for 7 days. Purified T cells were incubated in RPMI ex vivo for day 7 with DC cells (1:10 ratio) isolated from wild-type or BTN2A2-/- mice. Data depicted as Mean ± S.D of 3 independent experiments; One-way ANOVA with Tukey’s multiple comparison test. G Bar graph of GFP-Foxp3+ cell (%) population in total CD4+ T cells co-cultured with B cells at day 7. Purified T cells were incubated ex vivo for 7 days with B cell (1:5 ratio) isolated from wild-type or BTN2A2-/- mice. Data depicted as Mean ± S.D of 3 independent experiments; one-way ANOVA with Tukey’s multiple comparison test.

To further explore the short-term effects of BTN2A2 on T cell subset expression patterns in mixed lymphocyte reaction (MLR) culture conditions, we assessed gene expression patterns for molecules related to Th1, Th2, Treg, and Th17 cells by RT-PCR. BTN2A2 upregulated Il2rb and Foxp3 and downregulated Il21, consistent with our in vitro findings of expansion of Tregs and suppression of Th17 cells (Fig. S11A–C). The assays further showed downregulation of transcription factors regulating Th1 (Tbx21) and Th2 (Gata3) differentiation pathway at day 5 (Fig. S11D–F).

To explore the role of endogenous BTN2A2, we genetically knocked out murine BTN2A2 using a CRISPR-Cas9 strategy (Fig. S12A, B) and confirmed that expression of BTN2A2 was absent in professional APCs (CD11c+ dendritic cells and CD19+ B cells) in knock-out mice compared to wild-type littermates (Fig. S12C, D). Interestingly, when isolated APCs from BTN2A2-/- animals were co-cultured with murine T cells from Foxp3+ reporter mice, the absence of BTN2A2 from both isolated dendritic cells and isolated B cells from BTN2A2-/- animals impaired mouse T cell differentiation to Foxp3+ Tregs when compared to dendritic cells and B cells from wildtype littermates (Fig. 4F, G and Fig. S13)

BTN2A2-induced enhancement of Treg/Th17 balance is dependent on CD45 phosphatase activity

To test whether CD45 phosphatase activity was essential for BTN2A2’s actions on TCR signaling and Treg expansion, we used a previously validated CD45-specific phosphatase inhibitor¹⁸. Addition of BTN2A2-Fc to activated Jurkat T cells decreased phosphorylation of ZAP70 kinase, while inhibition of CD45 phosphatase activity with this small molecule inhibitor¹⁸ abrogated the effect (Fig. S14A). When we inhibited CD45 phosphatase activity during BTN2A2-Fc treatment, we also observed complete abrogation of the effects of BTN2A2-Fc on increased Treg and decreased Th17 populations (Fig. S14B, C). Taken together with other experiments described above, these studies link BTN2A2 effects on CD45 phosphatase activity within the TCR complex and immune synapse and Treg and Th17 cellular balance.

BTN2A2-Fc therapy exhibits immunoregulatory function in vivo

To test whether and how the in vitro observed immunoregulatory effects of BTN2A2 correlate to findings in vivo, we employed two distinct murine immune-mediated model systems. Consistent with previous reports^19,20, injection of nephrotoxic serum (NTS) in wild-type B6 mice induced crescentic glomerulonephritis with significant proteinuria (Fig. 5A–D). Administration of BTN2A2-Fc during the days following NTS administration reduced proteinuria and glomerular crescent formation. Administration of BTN2A2-Fc also increased CD4+ Foxp3 gene expression (Fig. 5E), decreased CD4 + RORγt gene expression (Fig. 5F), and attenuated T cell activation marker CD5 levels (Fig. 5G) in CD4 + T cells purified from spleen and lymph nodes. Further, BTN2A2-Fc lowered IL-17A protein expression in the kidneys compared to controls (Fig. 5H, I).

Fig. 5. BTN2A2-Fc ameliorates crescentic glomerulonephritis in mice induced by nephrotoxic serum (NTS).

A Schematic protocol of NTS-induced nephrotoxic glomerulonephritis model and histological changes. GBM glomerular basement membrane. B Quantification of Mean ± SEM proteinuria (on Day 7) in mice with NTS induced glomerulonephritis treated with recombinant BTN2A2-Fc protein or vehicle control (control) (N = 12 mice each for both groups); Mann–Whitney U test (two-tailed). C, D Representative images of glomerular injury by PAS staining. Crescent formation in mice with NTS induced glomerulonephritis treated with recombinant BTN2A2-Fc protein or control (D). Scale bars are 25 μm. Summary (C) depicts quantification of % glomeruli with crescents (Mean ± S.E.M, N = 12 mice each for both groups); Mann–Whitney U test (two-tailed) (E, F) Relative Foxp3 (E) and RORγt (F) expression compared to GAPDH as internal control in CD4 + T cells purified from spleen and lymph nodes from mice with NTS induced glomerulonephritis treated with BTN2A2-Fc protein or control. Data depicted as mean ± S.E.M (N = 8 mice each for both groups); Mann–Whitney U test (two-tailed). G Quantitative PCR analysis show CD5 expression in CD4 + T cells from mice with NTS induced glomerulonephritis treated with BTN2A2-Fc protein or control. Data depicted as mean ± S.E.M (N = 8 mice each for both groups); Mann–Whitney U test (two-tailed) (H, I) Immunoblot analysis of IL-17A in kidney tissue lysate of mice with glomerulonephritis treated with recombinant BTN2A2-Fc or control (H). I shows quantification of western blot. Data represented as Mean ± S.E.M (N = 7 mice each for both groups); Mann–Whitney U test (two-tailed).

We next administered a lower dose of NTS (50% less than the amount used in experiments above) to BTN2A2-/- mice (Fig. S12) and wildtype littermate controls. These experiments demonstrated that lower doses of NTS induced relatively mild crescentic glomerulonephritis and proteinuria among wild-type control animals, however, in BTN2A2-/- animals, we saw an exacerbation of crescentic glomerulonephritis and severe proteinuria (Fig. 6A–C). We also observed reduced Foxp3 and enhanced RORγt gene expression in splenic/lymph node CD4 + T cells from the BTN2A2-/- mice vs. controls (Fig. 6D, E). CD5 was enhanced in CD4 + T cells in BTN2A2-/- mice compared to wild type mice (Fig. 6F). Analysis of kidney tissue 1 week after administration of nephrotoxic serum also showed increased IL-17A protein expression in the kidneys of BTN2A2-/- vs controls animals (Fig. 6G, H).

Fig. 6. BTN2A2 null mice show exacerbated crescentic glomerulonephritis.

A Scatter plot shows quantification of Mean ± S.E.M of proteinuria in wild-type and BTN2A2 (-/-) mice with nephrotoxic glomerulonephritis on Day 7 (N = 15 mice each for both groups); Mann-Whitney U test (two-tailed). B, C Representative images of glomerular injury by PAS staining. Wild-type mice show minimal glomerular injury with a low dose of nephrotoxic serum, while BTN2A2 (-/-) mice show severe glomerulonephritis (B). Scale bars are 25 μm. Summary data are depicted as mean ± S.E.M of %Glomerular crescents (C). N = 15 mice per group; Mann–Whitney U test (two-tailed). Relative Foxp3 mRNA (D) and RORγt mRNA (E) expression compared to GAPDH as internal control in CD4 + T cells purified from spleen and lymph nodes cells from wild-type and BTN2A2-knockout mice with nephrotoxic serum induced glomerulonephritis. Summary data are depicted as Mean ± S.E.M from N = 7 mice per group; Mann–Whitney U test (two-tailed). F Quantitative PCR analysis show CD5 expression in CD4 + T cells from wild-type and BTN2A2-knockout mice with nephrotoxic serum induced glomerulonephritis. Summary data are depicted as mean ± S.E.M from N = 7 mice per group; Mann-Whitney U test (two-tailed). Immunoblotting for IL-17A expression in kidney tissue lysate using anti-IL-17A antibody (G) and quantitation (H) in nephrotoxic serum induced glomerulonephritis from wild-type (N = 4) and BTN2A2 null mice (N = 5); Summary data are depicted as Mean ± S.E.M; Mann–Whitney U test (two-tailed).

To evaluate the generalizability of the role of BTN2A2 in immunoregulation, we studied the effects of recombinant BTN2A2-Fc in DBA/2 mice mated with CBA/J mice, a model of immunologically mediated abortions²¹. We confirmed that DBA/2 male x CBA/J female had reduced litter sizes and higher spontaneous adsorption rates compared to CBA/J male x DBA/2 females. Administration of BTN2A2-Fc improved litter size and rescued the excess abortion rates that were noted in the DBA/2 male x CBA/J female (Fig. 7A–C). The beneficial effects of BTN2A2-Fc were associated with increased frequencies of splenic/lymph node Foxp3+ Tregs, reduced frequencies of splenic/lymph node Th17 cells and attenuation of CD5, consistent with our findings in the autoimmune GN model (Fig. 7D–G and Fig. S15A, B). In addition, BTN2A2-Fc protein also reduced placental IL-17A protein expression that correlated with improved litter sizes (Fig. 7H, I).

Fig. 7. BTN2A2-Fc rescues fetal resorption and increases litter size in CBA/J x DBA/2 abortion model.

A Schematic of the Abortion prone model. Pregnant CBA/J x DBA/2 mice treated with BTN2A2 throughout pregnancy has improved litter size (B), reduced resorption (C) compared to untreated mice. As additional control, DBA/2 x CBA/J mice has been included. Data represented as mean ± S.E.M; N = 11 mice per group. Flow cytometry (D) Quantitative PCR (E) analysis show increase in Foxp3 expressing CD4+ T cells in pregnant CBA/J x DBA/2 J mice treated with recombinant BTN2A2-Fc as compared to control mice. Data represented as mean ± S.E.M; N = 8 mice per group. F Flow cytometry analysis (left panel) show reduced number of CD4+RORγt+ cells in pregnant CBA/J x DBA/2 J mice treated with recombinant BTN2A2-Fc as compared to control mice. Summary data depicted as Mean ± S.E.M from N = 8 mice per group. G Flow cytometry analysis show mean fluorescence intensity of CD4+ CD5+ cells in pregnant CBA/J x DBA/2 J mice treated with recombinant BTN2A2-Fc as compared to control mice DBA/2 J x CBA/J. Summary data depicted as Mean ± S.E.M from N = 6 mice per group. H, I Immunoblot analysis of IL-17A in placental tissue lysate of pregnant CBA/J x DBA/2 J mice treated with recombinant BTN2A2-Fc or control (left panel). Right panel shows quantification of summary data (N = 5 mice per group). All statistical analysis done using Kruskal–Wallis with Dunn’s multiple comparison test.

BTN2A2 enhances Treg cell expansion and suppresses Th17 cell populations in human PBMCs

To evaluate whether the effects of recombinant BTN2A2 apply to human T cells, we analyzed human PBMCs activated by mixed lymphocyte reactions (MLR, using allogeneic stimulator cells) with and without recombinant BTN2A2-Fc, and quantified CD4+CD25+ Foxp3+ Treg cell numbers 7 days later. These co-cultures showed that the BTN2A2-Fc induced 2-fold expansion of Tregs under MLR conditions compared with controls; however, BTN2A2-Fc was unable to induce Tregs in the presence of CD45 inhibitor (Fig. 8A, B, Figs. S6C and S16A). Parallel experiments revealed that recombinant BTN2A2-Fc robustly blocked the TGF-β-IL6-IL-1β induced increase in Th17 cells (Fig. 8C, D and Fig. S16B). Interestingly, BTN2A2-Fc was unable to block cytokines induced Th17 cells in the presence of CD45 inhibitor. Taken together, these observations suggest that CD45 phosphatase activity is necessary for BTN2A2’s actions in humans as well.

Fig. 8. BTN2A2-Fc enhances Tregs and suppresses Th17 cells differentiation in in vitro human PBMC mixed lymphocyte reactions.

B Flow cytometry plots in A depicts percentage of Treg cells (CD4+CD25+Foxp3+ cells) in human PBMC incubated with /without recombinant BTN2A2-Fc (10 μg/ml) in MLR for 7 days in the absence or presence of CD45 phosphatase inhibitor (125 nM). Summary data is B is depicted as Mean ± SD (N = 4 independent experiments) showing percentage of Treg cells (CD4+CD25+Foxp3+ cells) among total CD4+ cells; One-way ANOVA test with Tukey’s multiple comparison test. D Flow cytometry plots in C showing percentage of CD4+RORγt+ (Th17 cells) in human PBMC incubated with anti-CD3 antibody (0.5 μg/ml) only or anti-CD3 antibody +TGF-β (1.5 ng/ml) +IL-6 (10 ng/ml) +IL-1β (10 ng/ml) in absence or presence of recombinant BTN2A2-Fc (10 μg/ml) in MLR for 5 days in absence or presence of CD45 phosphatase inhibitor (125 nM). Summary data is depicted in (D) as Mean ± SD of 3 independent experiments; One-way ANOVA test with Tukey’s multiple comparison test.

BTN2A2 suppresses T cell proliferation, Th1 and Th2 cell response in human PBMCs

In human MLR reaction, we mixed CellTrace violot (CTV) labeled HLA-typed PBMCs with allogenic, HLA-typed primary human B cell lines for 4 days in the absence or presence of recombinant BTN2A2-Fc. Incubation of PBMCs with BTN2A2-Fc inhibited CD4+ and CD8+ T cell proliferation (Fig. 9A) as observed by reduced CTV dye dilution assay and reduced intracellular staining of IFN-γ production (Fig. 9B and Fig. S17A, B).

Fig. 9. BTN2A2-Fc inhibited Th1/Th2 cells activation in in vitro human MLR.

A, B Human PBMCs were co-cultured for 4 days with HLA-mismatched B cells in the absence and presence of BTN2A2-Fc. HLA type of PBMCs (P1) was defined as (HLA-A 03/30; HLA-B 14/15; HLA-C 03/08; HLA-DRB1 03/07; HLA DQB1 02/02). HLA type of PBMCs (P2) was defined as (HLA-A 11/11; HLA-B 35/35; HLA-C 04/04; HLA-DRB1 01/04; HLA DQB1 03/05); HLA type of B cells (B1) was defined as (HLA-A 02/03; HLA-B 07/44; HLA-C 05/07; HLA-DRB1 13/15; HLA DQB1 06/06); HLA type of B cells (B2) was defined as (HLA-A 01/02; HLA-B 08/62; HLA-C 07/10; HLA-DRB1 03/04; HLA DQB1 02/03). Total of six replicates from three independent experiments (P1 with B2; P2 with B1; and P2 with B2) represented by individual data points. Bar graphs depicted as Mean ± SD show the percentage of (A) CD4⁺ T and CD8⁺ T cells that proliferated; and B expressed IFN-γ as measured by flow cytometry; Mann–Whitney U test (two-tailed). C Human PBMCs from a tuberculin sensitive subject were cultured in fluorospot plates for 2 days in presence of tuberculin PPD and with or without BTN2A2-Fc. Left panels show representative image of IFN-γ secreting cells in fluorospot assay of three replicates. Bar graph shows quantification as Mean of number of spots ± SD. PHA-L was used as a positive control. N = 3 replicates per condition from one subject; One-way ANOVA with Tukey’s multiple comparison test. D Human PBMCs from a dust mite sensitive subject were cultured in fluorospot plates for 3 days in house dust mite and with and without BTN2A2-Fc. Left panels show representative image of IL-5 secreting cells in fluorospot assay of three replicates. Bar graph shows quantification as mean of number of spots ± S.D. PHA-L was used as a positive control. N = 3 replicates per condition from one subject; One-way ANOVA with Tukey’s multiple comparison test.

Separately, we tested the effects of recombinant BTN2A2-Fc protein on Th1 and Th2 cell cytokine production in PBMCs obtained from subjects sensitive to either Tuberculin purified protein derivative (PPD, induces IFNγ) or house dust mite (induces IL-5). FluoroSpot assays confirmed that recombinant BTN2A2-Fc markedly inhibited PPD-induced IFNγ production (57.67 ± 5.50 spots vs 151.3 ± 15.04 spots, p < 0.01, Fig. 9C) and similarly reduced house dust mite-induced IL-5 production (control, 133.7 ± 13.65 spots vs BTN2A2-Fc: 46.3 ± 11.93 spots, p < 0.01 Fig. 9D).

Discussion

Data presented here provide multiple lines of evidence supporting the conclusion that BTN2A2 binds to, and functions as a ligand for, CD45 (preferentially binding to the CD45RO isoform). Our data indicate that BTN2A2 prevents segregation of CD45 from the TCR/CD3 complex, resulting in sustained CD45 phosphatase activity within the immune synapse, which reduces the downstream TCR signaling cascade. Consequently, BTN2A2 reduces proliferation of effector Th17, Th1, and Th2 cells while promoting differentiation of Treg, and reduces the severity of NTS-induced glomerulonephritis and immune-mediated pregnancy loss in mouse models of immunologic diseases. Together with previous work showing that APC-derived BTN2A2 regulates T cells¹⁴, our findings suggest that during cognate T cell/APC interactions, APC-presented BTN2A2 ligates T cell-expressed CD45RO to the TCR and functions as a molecular brake at the time of activation by disrupting CD45 segregation from the TCR complex (an event previously implicated as necessary for T cell activation)⁴. Our data are also consistent with Payne et al., who reported that BTN3A1, another family member of BTN proteins, inhibits tumor-reactive T cell responses by preventing segregation of N-glycosylated CD45 from the immune synapse²². We confirm that BTN2A2 inhibits T cell activation by interacting with CD45RO, that specific N-glycosylated residues in CD45 (Asp-419 and Asn-468) are critical for this interaction, and confirm that point mutations of these residues or PNGase treatment significantly attenuated the CD45/BTN2A2 interaction. Our data suggest that in activated T cells, BTN2A2 binds to both CD45 and CD3ε, likely maintaining them in proximity, enabling CD45 phosphatase activity to persist, enabling dephosphorylation of CD3 ITAM domains and Zap70, and limiting downstream TCR signals required for full T cell activation (Fig. 10).

Fig. 10. Schematic of the model showing BTN2A2 as a co-inhibitor of TCR signaling via CD45 phosphatase.

(Created in BioRender. Ali, S. et al. (2025) https://BioRender.com/uitoe4e).

Our data indicate that BTN2A2 enhances CD45’s phosphatase activity within the TCR complex that reduces TCR signaling, which in turn suppresses Th17, Th1, and Th2 responses while inducing Tregs. We note that previous work by others showed that transient targeting of CD45 (independent of BTN2A2) similarly induced potent antigen-specific regulatory T cells and tolerance in a graft-vs-host disease model^23,24, although effects on TCR signaling were not fully described.

In conjunction with previous work by others demonstrating BTN2A2 is expressed on APCs¹², our data suggest that BTN2A2 expressed on APCs and DCs likely functions in trans as an endogenous ligand for CD45 on T cells. CD45 is expressed as multiple spliced products, and it is currently believed that CD45RA and CD45RB are replaced by CD45RO following T cell activation². Our structural and functional studies suggest that BTN2A2 predominantly binds to the CD45RO isoform that is expressed on activated T cells with a kD of 63.5 nM, which is in the range of binding affinities proposed for other B7 family members with receptors CD28 and CTLA-4²⁵. Our findings also support prior work where targeting naïve animals with anti-CD45RB antibody resulted in increased expression of CD45RO isoforms that correlated with upregulation of immunoregulatory T cell subset²⁶. While multiple human glycoproteins, including galectin-1 and CD22, were shown by others to bind CD45^27,28, there is no clear evidence that these putative ligands modulate CD45 phosphatase activity in T cells. Other identified ligands for CD45 include pUL11, a viral protein expressed by cytomegalovirus-infected cells, as well as E3/49 K protein expressed by adenovirus-infected cells; both of these viral proteins may play roles in inhibiting anti-viral immunity by inhibiting TCR signaling, although precise mechanisms remain unclear^29,30.

BTN2A2-Fc limits clinical manifestations of glomerular crescent formation and proteinuria in murine NTS glomerulonephritis and reduces miscarriages in a distinct orthogonal mouse model, both of which were associated with augmented Treg and reduced Th17 cell populations. Consistent with prior data, CD5, a T cell activation marker, was attenuated by BTN2A2-Fc³¹. Together with previous work in murine experimental autoimmune encephalomyeletis¹⁴, our data further indicate that BTN2A2-Fc is a potentially useful immune modulator with translational potential in a variety of immune-mediated diseases and transplantation. Our observations are an important step forward in understanding how naturally occurring proteins can modify inflammatory events by interacting with CD45 to sustain TCR complex phosphatase activity, increasing Treg and decreasing Th17 cell populations.

While we show that TCR signaling through ZAP70 is modulated by BTN2A2 in the presence of anti-CD3, we acknowledge that the specific molecular links between BTN2A2-induced alterations in TCR signaling and the induction of Foxp3+ Treg and the inhibition of Teff cytokine production remain to be fully elucidated. Prior studies from other groups have suggested that switching the antigen stimulation from an acute high level to chronic low-level stimulation may lead to more Treg cells and immune tolerance^32,33 but whether and if so how CD45 and BTN2A2 contribute to this mechanism is unknown. CD45 has been proposed as a signaling gatekeeper in T cells³⁴, however, structural and functional studies are now needed to precisely define how enabling graded signaling outputs in the presence or absence of BTN2A2 leads to the promotion of tolerogenic signals. Further understanding the changes in the CD45 isoform and glycosylation status following activation may offer new immune regulatory options³⁵. In experimental studies, CD28-B7 blockade with CTLA4-Ig can prevent the development of autoimmune glomerulonephritis³⁶. It would be interesting to evaluate whether BTN2A2 can synergize with co-stimulatory pathway molecules such as CTLA4-Ig that also promote tolerance by upregulating Tregs³⁷ particularly in the context of CD45RO+ memory T cells that are resistant to co-stimulatory blockade. Other butyrophilin family members, such as BTN2A1 and BTN3A1 are known to be important for phosphoantigen mediated γδ T cell activation³⁸. Additional studies will be needed to determine whether BTN2A2 may also regulate γδ T cell activation. Furthermore, since BTN2A2 is the only butyrophilin family member expressed by professional APCs, including antigen-specific B cells, further studies are needed to investigate whether BTN2A2 plays a role in communicating antigen-specific immune tolerance between autoantigen-specific APCs and CD45RO-expressing autoantigen-specific cognate memory T cells. Since CD45RO is also expressed on B cells, it is possible that some of the beneficial effects of BTN2A2 in our animal studies may be derived from its actions on B cells. In this regard, Szodoray et al. demonstrate that T helper signals upregulate CD45 phosphatase activity in B cells. High CD45 phosphatase activity in memory B cells controls their effective differentiation toward antibody-secreting cells in response to T helper signals³⁹. Additional work is needed to determine whether the immunomodulatory properties of BTN2A2 may act in cis⁴⁰, for example, in memory B cells that express high levels of ligand and receptor.

In conclusion, our findings suggest that BTN2A2 ligation of CD45 phosphatase to the TCR complex leads to dampened TCR signaling, resulting in expansion of Tregs and suppression of Th17 cells. While we demonstrate a beneficial effect of BTN2A2 in two orthogonal models of autoimmunity/immune tolerance, our data suggest that the BTN2A2/CD45 signaling pathway could be targeted in a wide variety of immune-mediated diseases such as inflammatory bowel disease, rheumatoid arthritis, multiple sclerosis, and transplant rejections. A better understanding of the precise targets of BTN2A2 based on its modes of action is critical for the development of new treatment strategies in autoimmune and chronic inflammatory diseases and transplant rejection.

Methods

Ethics statement

Human studies were approved by the Institutional Review Board (IRB) at the Cedars-Sinai Medical Center (Pro00017197). Since only anonymous discarded blood samples obtained from the blood bank at Cedars-Sinai was used for in vitro studies, the IRB deemed no informed consent was necessary and the protocol was approved as expediated research. All animal experiments were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and all the protocols used in this study were approved by the Cedars-Sinai Medical Center Institutional Animal Care and Use Committee.

BTN2A2-Fc construct generation and purification of recombinant protein

Soluble recombinant BTN2A2-Fc protein was generated using baculovirus infected insect cell system, a method that is both scalable and has been successfully used for generation of recombinant proteins in the immune system⁴¹. An upstream 711 bp region of human BTN2A2 gene (NP_008926.2) containing signal peptide and two extracellular domains (IgV and IgC2) was amplified using specific primers (Table S1) from cDNA from 293T cells. Amplified region was first cloned into pFUSE-hIgG1-Fc1 vector (InvivoGen Catalog # pfuse-hg1fc1) using Age1 and EcoRV restriction sites. Subsequently, BTN2A2 region along with Fc region was amplified using specific primers (Table S1) and cloned into pBacPak8 vector using Xba1 and Sac1 restriction sites. Paired-end sequencing of the clone was performed by sanger sequencing to confirm the sequence. Primers were synthesized from Integrated DNA Technologies, Inc (IDT) and Phusion® High-Fidelity PCR Master Mix with HF Buffer was used in all PCR amplification reaction (New England Biolabs). Baculovirus were generated by co-transfecting pBacPak8-BTN2A2-Fc clone with linearized baculovirus DNA (TakaraBio Catalog# 631401) into Sf9 insect cells according to manufacturer protocol. Preparation of baculovirus passage (P0, P1, P2) was done as described in manufactures protocol (TakaraBio, Cat # 631402). P3 baculovirus stock was used for large scale protein purification using Hi5 cells grown in serum free media (Express Five™ SFM, ThermoFisher, Cat# 10486025). Pierce™ Protein G Agarose columns (ThermoFisher Scientific, Cat# 20398) was used for purification of secreted soluble-BTN2A2-Fc protein from the supernatant. Purified protein was finally suspended in 1x phosphate buffer saline (1xPBS). Protein purity was checked by SDS-PAGE followed by Coomassie blue staining. Specificity of the purified protein was confirmed by western blot analysis using specific BTN2A2 antibody. Commercially available recombinant 293 cells derived human BTN2A2-Fc (Cat# 8918-BT; R&D systems) and mouse BTN2A2 (Cat# 8997-BT-050; R&D systems) was used for comparison studies with baculoviral-generated recombinant B2NT2A2-Fc. For some biochemical studies, we also generated BTN2A2-His tag protein (without Fc tag) in Hi5 cells and purified using Ni-NTA Purification System (Invitrogen cat# K95001).

T-cells activation and IL-2 secretion assay

To study T cell activation in vitro in Jurkat cells⁴², 96-well plates were coated with 1 μg/ml concentrations of anti-CD3 mAb (clone OKT3) in PBS at 4 °C overnight in absence or presence of recombinant BTN2A2-Fc (10 μg/ml) or Fc tag protein (Cat#10702-HNAH, Sino Biological Inc) or 293 cells derived human BTN2A2-Fc or mouse BTN2A2. A total of 2.5 × 10⁵ Jurkat cells/well were added to precoated flat-bottom 96-well plates. Cells were incubated for 24 hr in 37 °C incubator with 5% CO₂. Total 100 µl of media was used to measure IL-2 production using an IL-2 ELISA kit from Millipore-Sigma (Cat# RAB0286-1KT) according to the manufacturer’ protocol.

In vitro TCR stimulation and signal transduction analysis

For short-term activation of the Jurkat cells⁴³, 48-well tissue culture plate was coated with anti-CD3 mAb (OKT3, 10 μg/mL) and BTN2A2-Fc (10 μg/mL) or Fc-Tag (10 μg/mL). Plate was incubated overnight at 4 °C and each well was washed twice with PBS. A total of 2 × 10⁶ cells in 75 μl 1 x PBS were then added to each well for 3 min at 37 °C. The reaction was then stopped with ice cold PBS. Pervanadate was prepared by incubating vanadate (200 mM) and H₂O₂ (200 mM) in 1:2 ratio for 15 min at room temperature. Jurkat cells in 1 x PBS were treated with pervanadate at a final concentration of 0.1 mM for 5 min in an incubator (5% CO2 and 37 °C). The cells were then collected, lysed in lysis buffer (RIPA or IP lysis buffer), and subsequently subjected to immunoprecipitation reaction or immunoblotting. Phosphoflow analysis, Jurkat cells were either unstimulated or treated with plate bound 1 μg/ml anti-CD3 antibody in the presence or absence of 10 μg/ml BTN2A2-Fc fusion protein for 5 min and 30 min at 37 °C. Cells were fixed with BD Cytofix/Cytoperm-Fixation/Permeabilization buffer for flow cytometric analysis with p-ZAP70 (pY319)/Syk (Y352), p-Akt (pT308), pAkt (pS473), pERK1/2 (pT202/pY204), pS6 (pS235/pS236), and pPI3 Kinase (p85 alpha).

Immunoprecipitation and immunoblotting

Jurkat cells were activated with plate-bound anti-CD3/anti-CD28 with BTN2A2-Fc or Fc-tag protein or studied un-activated. Cells were collected, then washed in PBS and resuspended in IP lysis buffer (10 mM HEPES pH 7.5, 0.5 mM EDTA, 0.5% NP-40, 250 mM NaCl, 1x phosSTOP and protease inhibitors), incubated on ice for 30 min with vortexing, and cleared by centrifugation at 15,000 g for 10 min. An aliquot of protein lysate was used for western blotting, while the remainder were incubated at 4°C for 8–12 hours with specific antibodies or IgG as described in figure legends. Protein A/G Agarose (ThermoFisher Scientific) was added to lysate and incubated for additional 2 hours . Beads were washed once in IP-lysis buffer with 0.5% NP-40 and twice in IP-lysis buffer with 0.2% NP-40. Protein lysate prepared after Jurkat cell treatment and/or immunoprecipitation was separated in 4–12% Bis-Tris gel (ThermoFisher Scientific) and transferred to nitrocellulose membrane. Membrane was blocked for 1 hour in blocking solution (5% BSA in PBS with 0.1% Tween [PBST]). The membrane was incubated with primary antibodies overnight at 4°C, washed with PBST three times (10 min each), incubated with secondary antibodies for 1 hour at room temperature, and finally washed with PBST three times (10 min each). The blots were visualized using an ECL assay (ThermoFisher Scientific). Western blots were probed with specific primary antibodies as described in figure legends (Table S2). For measurement of phosphatase activity, immunoprecipitated material was directly mixed in Fluorescein Diphosphate, Tetraammonium Salt substrate (FDP, Catalog # F2999, Thermofisher Scientific), and activity was measured using Molecular Devices SpectraMax® M2 plate readers as recommended by the protocol included by the manufacturer.

For long-term activation studes, Jurkat cells in suspension were activated with anti-CD3/anti-CD28 antibody (1 μg/ml each) for 48–72 hours in RPMI medium supplemented with 10% FBS. Jurkat cells were collected, washed with PBS, and incubated on ice with 10 μg/ml BTNT2-Fc or Fc protein for 1–2  hours prior to immunoprecipitation studies. Extracellular crosslinking with BS3 (bis[sulfosuccinimidyl] suberate) was performed according to user manual (Thermo Scientific MAN0011240). BS3 (ThermoFisher) was added at a final concentration of 5 mM and incubated for 30 min at room temperature. The reaction was stopped using Tris-HCl, pH 7.5, at a final concentration of 20 mM for 15 min at room temperature. The cells were then washed extensively in PBS and resuspended in IP-lysis buffer and cleared by centrifugation at 14,000 g for 10 min. Protein G Agarose beads were used to pull down BTN2A2-Fc and Fc-tagged protein. Beads were washed and eluted with SDS sample buffer supplemented with dithiothreitol. Eluted samples were analyzed by western blot probed with specific antibodies (Table S2) as described in the figure legend.

Uncropped gels/scans for all western blots is included in Source data files.

Molecular modeling of BTN2A2 and PTPRC (CD45)

A molecular model of BTN2A2 was generated using homology modeling. A search for the homologous structure of the extracellular domain of BTN2A2 revealed that BTN3A2 shares about 47% of sequence homology with BTN2A2. Using the crystal structure of BTN3A2 as a template, the three-dimensional model of BTN2A2 was generated using SWISS-MODEL workspace⁴⁴. Subsequently, the molecular structure of BTN2A2 was subjected to a short 1.2 ns molecular dynamics simulation using Desmond (Schrodinger, Inc., San Diego, CA). The three-dimensional structure of PTPRC was retrieved from the protein data bank (PDB code: 5FN7)⁴⁵. To determine the interaction between BTN2A2 and PTPRC proteins, RosettaDock 4.0 was used^46,47. The top 10 predicted models were then subjected to 1.2 ns molecular simulation followed by minimization using Desmond. Then, the most energetically stable docking model was used to identify potential amino acids for mutation.

Microscale thermophoresis measurement of the BTN2A2 binding interaction with CD45RO-Fc

We measured the binding interaction between BTN2A2-His and CD45RO-Fc by Microscale Thermophoresis (MST)⁴⁸ using the Monolith NT.115 instrument (Nanotemper Technologies, München, Germany). Briefly, the BTN2A2-His protein was fluorescently labeled by covalent labeling method using the Monolith series protein labeling kit RED-NHS 2nd generation (Amine Reactive; Cat# MO-L011). The 1x PBS buffer supplemented with 0.05% Tween-20 was used as the MST assay buffer to perform the binding experiments. The non-labeled CD45RO-Fc protein (Catolog # 10642-CD, R&D systems, MN) was titrated in the concentration range of 8 μM–244 pM against the 20 nM of fluorescent labeled BTN2A2. The MST data were collected at settings including medium (40%) MST power, 20% LED/excitation power, Nano-RED excitation type and 25 °C thermostat setpoint. The BTN2A2-His binding affinity to CD45RO-Fc was determined by fitting the MST response data to the Kd model in the MO Affinity Analysis Software version v2.3 (Nanotemper Technologies, München, Germany).

CRISPR-Cas9 deletion of CD45 gene and site directed mutagenesis of CD45

CD45 knockdown cell lines were generated using the CRISPR-Cas9 system. Two different guide RNAs were designed complementary to upstream of CD45 exon-1 and downstream to exon-3 based on its protospacer adjacent motif (PAM) sequence (Table S1). These guide RNA sequences were cloned into lenti-Crispr v2 blast vector containing the Cas9 coding gene and a blasticidine resistant gene. Lentivirus were generated using envelop and packaging plasmid. Jurkat cells were then transduced with lentivirus expressing guideRNA and Cas9 gene and selected for blasticidine resistant cells. Single-cell cloning was performed to isolate independent clones. Deletion of Exon-1 to Exon-3 was confirmed by direct Sanger sequencing (Fig. S5D).

CD45 (PTPRC) (NM_002838) Human Tagged ORF Clone was purchased from origene technologies. Inc. Site-directed mutagenesis were performed using the QuickChange mutagenesis kit (Invitrogen) to introduce the N419A and N468A mutations. Introduction of the mutation was confirmed by direct sequencing (Fig. S5E). Lentivirus were generated using envelop and packaging plasmid. Jurkat cells deleted for CD45 gene were then transduced with lentivirus expressing mutant CD45 gene. For PNGase treatment, Jurkat cells were activated with anti-CD3/anti-CD28 antibody 1 μg/ml for at least 48 hours in suspension. Cells were collected and resuspended in 1x glycobuffer-2 along with PNGaseF (100U/10⁶ cells, NEB Cat# P0709S) for 4–8 hours at 37 °C in an incubator. Cells were lysed in IP lysis buffer for co-IP assay or used directly for Lectin binding assay using flow cytometry analysis (Fig. S5F). Ulex europaeus (gorse, furze, Sigma-Aldrich Cat#L5505) was purchased, conjugated with PE dye, and used 0.7 μg/ml for staining.

In vitro Treg and Th17 differentiation by mixed lymphocyte reaction (MLR)

For Treg differentiation, fresh spleen and lymph nodes (inguinal, brachial, and axillary) were harvested from mice and washed with RPMI media. Spleen was excised into small pieces, and single cells were prepared by mashing the tissue with the plunger end of syringe through a 70 μm strainer and cells were suspended in 5 ml of RPMI media. Cells were stimulated with plate bound anti-CD3 (1 μg/ml) with or without BTN2A2-Fc fusion protein (10 μg/ml) and co-cultured with CD1 mice irradiated splenocytes for 7 days (1:1 ratio). For denovo differentiation analysis, naïve (CD4+Foxp3-CD62L+ ) T cells sort purified from spleen and lymph nodes of Foxp3-GFP transgenic mice. Cells were labeled with immunofluorescent antibodies and analyzed for CD4+ CD25+GFP-Foxp3+ co-expression using flow cytometry. CD4+ cells incubated with anti-CD3 antibody in the absence or presence of recombinant BTN2A2-Fc for day-1 and day-5 were used to extract RNA for qPCR analysis. Purified APCs (DC and B cells) from WT and BTN2A2-/- mice were co-cultured in suspension for 7 days with isolated murine T cells from Foxp3+ reporter mice for measurement of Tregs. CD4 + T cells were isolated using EasySep™ Mouse CD4+ T Cell Isolation Kit (Stem Cell Tech); B cells using EasySep™ Mouse B Cell Isolation Kit (Stem Cell Tech); Pan-DC cells using EasySep™ Mouse Pan-DC Cell Enrichment Kit II (Stem Cell Tech); T cells using EasySep™ Mouse T Cell Isolation Kit (Stem Cell Tech) per protocols recommended by the manufacturer.

To explore the effects of BTN2A2 on Th17 differentiation, 1 × 10⁶ spleen cells were stimulated with soluble anti-CD3 antibody (0.5 μg/ml) and CD1 mice irradiated spleen cells (1:1 ratio) and Th17 differentiation cytokines cocktail (TGF-β1 1.5 ng/ml, IL-6 10 ng/ml, IL-1β 10 ng/ml) with or without recombinant BTN2A2-Fc (10 μg/ml)⁴⁹. Cells were co-cultured for 5 days and subjected to intracellular immunofluorescence staining of RORγt (Th17 cell marker). Cells were analyzed for CD4 + RORγt + co-expression using flow cytometry. CD45 phosphatase inhibitor (Compound 211, Catalog # 530197; Millipore Sigma) that has previously been characterized as irreversible and selective blocker of the allosteric pocket at the D1-D2 domains interface away from the substrate-binding/catalytic site¹⁸, was used at 125 nM concentration during the course of T cell differentiation studies.

Flow cytometry analysis

Immune cells were freshly harvested from spleen and lymph nodes (inguinal, brachial, and axillary) of mice. Cells were counted using a Hemavet 950FS hematology analyzer (Drew Scientific, Miami Lake, FL). Immunostaining and flow cytometry analysis of cells were performed⁵⁰ using cells isolated from mice or harvested from in vitro cultures and stained with Ghost Dye™ Red 780 Viability Dye (Cell Signaling). Cells were washed extensively with 0.5% BSA in 1xPBS and resuspended in staining buffer (FBS), and further stained with fluorochrome conjugated antibodies for cells surface marker anti-CD3, anti-CD4 and anti-CD25. Cells were washed with staining buffer to remove excess antibodies and then fixed/permeabilized using mouse Foxp3 Buffer Set (BD Biosciences, San Jose, CA) according to manufacturer protocol. Fixed/permeabilized cells were incubated in rat serum for 15 min and stained with intracellular markers (Foxp3 and/or RORγt). Cells harvested from Foxp3-GFP transgenic mice were stained with anti-CD3, anti-CD4 and anti-CD25 antibodies, washed and suspended buffer containing DAPI. Cells were directly analyzed for CD4+CD25+GFP-Foxp3+ or CD4+RORγt expression. The stained cells were analyzed using a BD LSR Fortessa (BD Biosciences, San Jose, CA). Details of the commercial antibodies used in this study are provided in Table S2.

For cell surface binding of BTN2A2-Fc, Jurkat cells were activated with anti-CD3/CD28 antibody (1 μg/ml) for at least 48 h. Recombinant BTN2A2-Fc was conjugated with biotin in 1:2 ratio and pre-complexed with Strep-PE-Cy7 for one hour on ice. Activated Jurkat cells (1 × 10⁶) were incubated with BTN2A2-Fc pre-complex (3 μg/1 × 10⁶ cells) directly or preceding with anti-CD45RO (GeneTex Cat# GTX00596)/CD45RA (invitrogen Cat#MA1-19113) antibody (3 μg/1 × 10⁶ cells) binding. Cells were washed with staining buffer and analysed in flow cytometer.

Cells survival, proliferation, and detection of intracellular IL2

CD4+ cells were isolated from the freshly harvested spleen of wild type B6 mice as described above. Cells were stained with CellTrace™ Violet dye at 5μM dye concentration in 10⁶ cells/ml dilution according to manufacturer protocol (Invitrogen). 2.5 × 10⁵ CellTrace™ Violet stained cells were incubated on bound anti-CD3 (0.5 μg/ml) plus anti-CD28 (0.5 μg/ml) with BTN2A2-Fc fusion protein (10 μg/ml) or Fc control (10 μg/ml). Cells were culture for 3 days at 37 °C and 5% CO2 in RPMI medium supplemented with 10% FBS. Cells survival was analysed based on cells negatively stained with Ghost Dye™ Red 780 Viability Dye (Cell Signalings Tech). Cell proliferation was analysed based on Dye dilution assay using flow cytometry analysis. For intracellular staining of IL2, cells were fixed, permeabilized, and stained with anti-IL-2 antibody (Clone, JES6-5H4, Biolegend) and analyzed using a BD LSR Fortessa (BD Biosciences, San Jose, CA).

Immunofluorescence studies and confocal microscopy

For immunofluorescence staining for CD45 and CD3ζ expression²², anti-CD3 antibody/anti-CD28 antibody (10 μg/ml each) and recombinant BTN2A2-Fc (10 μg/ml) were coated on glass cover slips in PBS overnight at 4 °C. 1 × 10⁶ Jurkat cells pre-incubated with mouse anti-human CD45 (Abcam cat# ab8216) in 75 μl of 1xPBS for 1 h were seeded on coated coverslips for 3 min at 37 °C and fixed using Cytofix/Cytoperm-Fixation/Permeabilization Kit (BD biosciences). Coverslips were then gently washed two times with 1x PBS. Cells were blocked with blocking buffer (5% goat/mouse serum in 0.3% triton-x100 in 1xPBS) and stained with rabbit anti-human CD3ζ (abcam cat# Ab226263). Coverslips were then gently washed 3 time with 1x PBS for 10 min each and stained with Donkey anti-Mouse IgG Antibody, Alexa Fluor™ 555 and Goat anti-Rabbit IgG Antibody, Alexa Fluor™ 488 (ThermoFisher Scientific). Coverslips were then gently washed 3 times with 1x PBS, incubated in DAPI for 5 min and mounted on slides using ProLong™ Diamond Antifade Mountant (ThermoFisher Scientific). Immunofluorescence of stained cells was acquired with a Zeiss LSM 780 confocal microscope through a 63x resolution with oil immersion (refractive index 1.518). Orthogonal sections (z-stacks) of single cell were acquired using z-stack tool of confocal microscope. The image was analysed as an 8-bit image and intensity of red and green channel was measured from 0 to 255 grayscale fluorescent units. Co-localization analysis was performed on Image-J plugin “Colocalization Finder”. At least nine fields were analysed per experimental condition, in which an average of 500 cells were analysed to yield the average amount of co-localization.

Quantitative polymerase chain reaction (QPCR)

For gene expression analysis, the total RNA was extracted using TRIzol™ Reagent (Thermo Fisher Scientific). The cDNA was synthesized from 500 ng of total RNA using SuperScript™ IV VILO™ Master Mix (Thermo Fisher Scientific). Quantitative real time PCR was performed with PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific) on a QuantStudio Real-Time PCR system (Thermo Fisher Scientific). GAPDH was used as internal controls. Gene expression was compared using the ΔΔCt method. Primers details are provided in (Table S1).

Animal models

All mice used in this study were housed in the Animal Research Facility at Cedars-Sinai Medical Center with a 12 hour light/12 hour dark cycle, ambient temperature of ~70 °F ( ~ 20 °C), 40–60% humidity, and with free access to water and chow throughout the experiment.

Foxp3EGFP (C.Cg-Foxp3^tm2Tch/J, Strain # 006769) mice

Foxp3EGFP mice that co-express EGFP and the regulatory T cell-specific transcription factor Foxp3 under the control of the endogenous promoter was acquired from the Jackson Laboratory.

BTN2A2 Knock-out mice (-/-)

BTN2A2(-/-) mice (strain: C57BL/6J-Btn2a2em1cyagen) were generated by deleting 10460 base pair region of btn2a2 gene comprising exon2–8 using CRISPR-Cas9 technique at commercial facility (Cyagen Biosciences) (Fig. S12A). Three PCR primers set (Table S1) were used for genotyping mice using polymerase chain reaction method. Wildtype mice shows single band of 500 bp, heterozygous shows two bands of 700 and 500 bp, while homozygous mice show 500 bp single band on agarose gel electrophoresis (Fig. S12B).

Glomerulonephritis model

We used nephrotoxic crescentic glomerulonephritis mouse model¹⁹ to evaluate the functional effects of BTN2A2 in vivo. Nephrotoxic sera were raised in rabbits by repeated immunization with the purified glomeruli in complete and incomplete Freund’s adjuvant. Female mice were preimmunized with normal rabbit IgG and complete Freund’s adjuvant 5 days prior to administration of nephrotoxic serum. Nephrotoxic serum nephritis was induced by the injection of 20 μl or 10 μl nephrotoxic serum intravenously at day 0. Animals were randomly assigned to receive either BTN2A2-Fc or vehicle. Four doses of 25 μg BTN2A2-Fc fusion protein or vehicle (control) were injected i.p. at day 0, 2, 4, 6 after nephrotoxic serum (NTS) injection. Mice were sacrificed at day 7 to collect tissues, urine, blood cells and plasma. Immune cells were harvested from fresh whole spleen and lymph-nodes (Inguinal, Brachial and Axillary) and CD4+ T cells were isolated using EasySep™ Mouse CD4+ T Cell Isolation Kit (STEM CELL TECH). Protein lysate of snap-freeze tissue in liquid nitrogen was prepared by homogenizing 5 mg tissue in 500 μl ice-cold RIPA lysis buffer. Lysate was agitated for 2 h at 4°C, centrifuged at 16,000 g for 20 min at 4°C and supernatant was collected. About 80 μg of tissue lysate was used to detect IL-17A  expression using a western blot.

To obtain the protein-to-creatinine ratio (mg/mg) in urine for evaluation of proteinuria, urine protein was measured with a protein assay dye (No. 500-0006, Bio-Rad, Hercules, CA), and urinary creatinine was measured using a Creatinine Assay Kit (No. DICT-500, BioAssay Systems, Hayward, CA). Fibrinoid necrosis (a precursor lesion for crescents, defined by finding GBM rupture, fibrin deposition, and karyorrhexis) and crescents were assessed in all glomeruli on one paraffin section for each mouse using periodic acid-methenamine silver and PAS stains, respectively. Representative images were captured using a light microscope (Nikon Eclipse 50i, Nikon, Tokyo, Japan). Histopathologic diagnosis of GN was evaluated by a board-certified renal pathologist (M.Y.)

Miscarriage model

DBA/2 and CBA/J mice were purchased from Jackson Laboratory. We crossed male DBA/2 with female CBA/J to document immunologically mediated pregnancy losses⁵¹. As controls, we crossed female DBA/2 with male CBA/J strain, and we evaluated litter size, pup weight, and litter resorption rates. To study the therapeutic effects of BTN2A2, four doses of 25 μg BTN2A2-Fc fusion protein were injected intraperitoneally at gestational days 6, 9, 12, and 15 in CBA/J female mice that were crossed with male DBA/2 mice. Animals were sacrificed at gestational day 18 and blood/tissues harvested for molecular analysis.

Human studies

Treg assays

Whole blood was drawn from healthy unrelated individuals to prepare peripheral blood mononuclear cells (PBMC) using the Ficoll–Hypaque gradient centrifugation method. After the stimulator PBMC (unrelated) was irradiated, it was exposed to the responder PBMC at 1:1 ratio (1 × 10⁶/ml of each PBMCs) in the absence or presence of BTN2A2-Fc at 0, 10 μg/ml, and then incubated for 7 days for measurement of Tregs. MLR mixture were first stained with antibodies to CD3, CD4, CD25. After permeabilization, the cells were stained with antibody to Foxp3. CD4+/CD25+/Foxp3+ cells were designated as Treg cells. Treg cell levels were expressed as Treg cell% of total CD4+ T cells.

Th17 assays

Human PBMCs (responder cells) isolated from Ficoll–Hypaque gradient centrifugation method were cultured in irradiated unrelated stimulator cells at 1:1 (1 × 10⁶/ml of each PBMCs) along with Th17 differentiation cytokines cocktail (anti-CD3 antibody 0.5 μg/ml, TGFβ 1.5 ng/ml, IL-6 10 ng/ml, IL-1β 10 ng/ml) with or without recombinant BTN2A2-Fc (10 μg/ml) for 5 days. Cells were stained with surface marker CD3 and CD4. After permeabilization, the cells were stained anti-RORγt antibody and analyzed for CD4+RORγt+ co-expression using flow cytometry. Th17 cells were expressed as RORγt+ cells % in CD4 T cells. CD45 phosphatase inhibitor was used at 125 nM concentration during the course of T cell differentiation studies.

Cell proliferation and detection of intracellular IFN-γ in Mixed Lymphocyte Reaction

PBMCs (Cell ID # CTLHHU20221018 and #CTLHHU20230601) purchased from commercial sources (CTL, Immunospot, Shaker Heights, OH) were labeled (200,000 cells per well) with 5 μM CTV and co-cultured with 50,000 HLA-mismatched primary B cell stimulators⁵² in 96-well round bottom plates in triplicates in RPMI-1640 media supplemented with 10% FBS at 37 °C and 5% CO2 for 4 days. HLA type of PBMCs (P1) was defined as (HLA-A 03/30; HLA-B 14/15; HLA-C 03/08; HLA-DRB1 03/07; HLA DQB1 02/02). HLA type of PBMCs (P2) was defined as (HLA-A 11/11; HLA-B 35/35; HLA-C 04/04; HLA-DRB1 01/04; HLA DQB1 03/05); HLA type of B cells (B1) was defined as (HLA-A 02/03; HLA-B 07/44; HLA-C 05/07; HLA-DRB1 13/15; HLA DQB1 06/06); HLA type of B cells (B2) was defined as (HLA-A 01/02; HLA-B 08/62; HLA-C 07/10; HLA-DRB1 03/04; HLA DQB1 02/03). Cultures were set up in the presence and absence of BTN2A2-Fc (10 µg/ml). After 4 days, cells were harvested, washed and stained with surface and intracellular cytokine antibodies. Flow cytometry was performed to assess T cell proliferation based on CTV dilution and IFN-γ expression. Cells were harvested from MLR and stained with Zombie NIR Viability Dye according to the manufacturer’s instructions (423106; BioLegend), then washed with staining buffer (1% FBS in PBS). Nonspecific antibody binding was blocked by 5% Fc Receptor Blocking Solution (422301; BioLegend) for 10 min on ice. Cells were stained with surface antibodies for 25 min at RT, washed and then fixed with Fixation Buffer (420801; Biolegend) for 20 min at RT and washed with 1X Intracellular Staining Permeabilization Wash Buffer (421002; Biolegend). Cells were subsequently stained with IFN-γ antibody for 20 min at RT and then washed with permeabilization wash buffer again. Cells were resuspended in staining buffer and run on a BD LSR Fortessa II.

Fluorospot assay

The IFN-γ and IL-5 Fluorospot was performed according to a modified version of Cellular Limited Technology’s (CTL) for Human IFN-γ/IL-17 and Human IL-4/IL-5 Double-Color FluoroSpot Assay protocol, respectively. Capture kit for Human IFN-γ (#hT2006F) and Human IL-5 (CTL; hT2023F) were purchased from CTL. High protein binding PVDF filter 96- well plates (Millipore Sigma MSIPS4510) were coated with capture antibodies. Human PBMC’s sentitive to Tuberculin PPD (Donor ID#626) and house dust mite (Donor ID# 595) were purchased from CTL and cultured (3.0 × 10⁵ cells) in the coated plates in CTL Serum-free medium (#CTLT-010) supplemented with 1% penicillin-streptomycin, 1% L-glutamine (Gibco 10378016). For IFN-γ assay, 50 µg/mL tuberculin PPD (Accurate Chemical SSIHF-100) for 2 days and for IL-5 assay 5 µg/mL house dust mite (Greer XPB81D3A2.5) for 72 h was added with and without 5 µg/mL BTN2A2-Fc at 37 °C/5% CO2. Phytohemagglutinin-L (PHA-L; Invitrogen 00-4977-03) was used as a positive control. Wells were then washed twice with PBS (Gibco 10010023) and twice with PBS-tween20 (PBS-T 0.05% v/v; Sigma P9416). Wells were incubated in detection solution and then washed three times with PBS-T. Wells were incubated with tertiary detection solution and washed three times with distilled water. Plates were allowed to dry overnight and were then imaged and counted on a ImmunoSpot® S6 ULTIMATE analyzer.

Statistical analysis

Data were graphed and statistics performed using GraphPad Prism version 9.2. Data are presented mean ± standard deviation (S.D) or standard error of mean (S.E.M) as indicated. Statistical significance was evaluated using the Mann–Whitney U test for comparison of 2 groups. For studies involving more than two groups, non-parametric Kruskal-Wallis test with Dunn’s test for multiple comparison or 1-way ANOVA with Tukey’s test for multiple comparison were used. Statistical analysis details with exact P values, and the number of experiments is indicated in the figures or figure legends.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

Conceptualization: S.A., A.H.B., S.J., and S.A.K.; Methodology: S.A., R.Z., P.H., and S.J.; Cell culture and signaling studies: S.A., R.Z., and B.S.; Animal models: M.Y., A.E.C., and V.D.; Structural modeling and MST: R.M., and M.K. Human studies: J.M., P.N.S., R.Z., B.S., and S.C.J., Writing, reviewing, and editing: S.A., A.H.B., M.Y., S.Y., M.P., R.T., P.H., S.J., and S.A.K.

Peer review

Peer review information

Nature Communications thanks Abdelhadi Saoudi, Jeong-Su do, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

Source data for all figures and uncropped gels/scans in the manuscript are provided with this paper. All reagents generated in this study including recombinant protein, baculoviral stocks, and knock-out mice will be made available to scientific community upon reasonable request. Source data are provided with this paper.

Competing interests

S.A., A.H.B., M.Y., J.M., R.Z., B.S., R.T., P.H., S.J., and S.A.K. are named as co-inventors on provisional patents (US PTO #63/688,964 and # 63/712,640) related to the subject matter of this manuscript that has been filed by Cedars-Sinai Medical Center. The remaining authors state no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-68077-6.