Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Precision targeting of autoantigen-specific B cells in muscle-specific tyrosine kinase myasthenia gravis with chimeric autoantibody receptor T cells

Precision targeting of autoantigen-specific B cells in muscle-specific tyrosine kinase myasthenia... nature biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Precision targeting of autoantigen-specific B cells in muscle-specific tyrosine kinase myasthenia gravis with chimeric autoantibody receptor T cells 1 1 1 2 Received: 22 March 2022 Sangwook Oh    , Xuming Mao    , Silvio Manfredo-Vieira    , Jinmin Lee , 2 1 2 2 Darshil Patel , Eun Jung Choi , Andrea Alvarado , Ebony Cottman-Thomas , Accepted: 8 December 2022 1 1 1 3 Damian Maseda    , Patricia Y. Tsao    , Christoph T. Ellebrecht , Sami L. Khella , 4 5 2 2 David P. Richman , Kevin C. O’Connor    , Uri Herzberg , Gwendolyn K. Binder , Published online: xx xx xxxx 6 2 1 Michael C. Milone , Samik Basu & Aimee S. Payne    Check for updates Muscle-specific tyrosine kinase myasthenia gravis (MuSK MG) is an autoimmune disease that causes life-threatening muscle weakness due to anti-MuSK autoantibodies that disrupt neuromuscular junction signaling. To avoid chronic immunosuppression from current therapies, we engineered T cells to express a MuSK chimeric autoantibody receptor with CD137-CD3ζ signaling domains (MuSK-CAART) for precision targeting of B cells expressing anti-MuSK autoantibodies. MuSK-CAART demonstrated similar efficacy as anti-CD19 chimeric antigen receptor T cells for depletion of anti-MuSK B cells and retained cytolytic activity in the presence of soluble anti-MuSK antibodies. In an experimental autoimmune MG mouse model, MuSK-CAART reduced anti-MuSK IgG without decreasing B cells or total IgG levels, reflecting MuSK-specific B cell depletion. Specific off-target interactions of MuSK-CAART were not identified in vivo, in primary human cell screens or by high-throughput human membrane proteome array. These data contributed to an investigational new drug application and phase 1 clinical study design for MuSK-CAART for the treatment of MuSK autoantibody-positive MG. Muscle-specific tyrosine kinase myasthenia gravis (MuSK MG) is a which leads to MuSK phosphorylation and formation of high-density chronic autoimmune disorder caused by MuSK autoantibodies that acetylcholine receptor (AChR) clusters that are essential for neuro- 1–3 6 result in potentially life-threatening muscle weakness . MuSK is a trans- muscular junction synaptic transmission . membrane receptor that interacts with lipoprotein receptor-related MuSK MG disease severity correlates with anti-MuSK antibody 4,5 7 protein 4 (LRP4) in complex with the neuronal proteoglycan agrin , titers , particularly IgG4 autoantibodies targeting the MuSK Ig1 1 2 Department of Dermatology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA. Cabaletta Bio, Philadelphia, PA, USA. 3 4 Department of Neurology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA. Department of Neurology, University of 5 6 California – Davis, Davis, CA, USA. Departments of Neurology and Immunobiology, Yale School of Medicine, New Haven, CT, USA. Department of Pathology and Laboratory Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA.  e-mail: [email protected]; [email protected] Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z 8–10 domain, which disrupt MuSK-LRP4 interactions . B cell depletion coincubation (Fig. 2a). MuSK-CAART and NTD-T produced similarly low with rituximab results in reduction of anti-MuSK IgG relative to total levels of IFNγ when cultured with Nalm-6 control and IgG from patients IgG , indicating that MuSK autoantibodies are primarily produced by with MuSK MG (Fig. 2b, left panel), whereas a significant increase in 12,13 short-lived plasma cells . Disease relapse after rituximab is attributed IFNγ production was observed in MuSK-CAART relative to NTD-T when to incomplete B cell depletion , requiring repeated rituximab infusions cocultured with anti-MuSK Nalm-6 cells (Fig. 2b, right panel, black for disease control, but chronic B cell depletion risks serious infections. versus red bars). IgG from patients with MG did not significantly affect Therapy should ideally eliminate only the pathogenic anti-MuSK B IFNγ levels, although a trend toward lower levels was observed (Fig. 2b, cells and spare healthy B cells to achieve durable remission of MuSK right panel, red bars). Coincubation of MuSK-CAART with monoclonal MG without generalized immunosuppression. Nalm-6 cells and matching soluble anti-MuSK mAb showed similar Here, we report the design and functional validation of a chimeric results, except 192-8/anti-Fz IgG4 potentiated MuSK-CAART cytotoxic- autoantibody receptor (CAAR) comprising the MuSK autoantigen, ity at higher concentrations (Extended Data Fig. 4a). tethered to tandem CD137-CD3ζ signaling domains. MuSK-CAAR To determine direct effects of anti-MuSK antibodies on expression in T cells directs cytotoxicity toward B cells expressing an MuSK-CAART, MuSK-CAART was incubated with anti-MuSK mAbs anti-MuSK surface autoantibody or B cell receptor (BCR). MuSK-CAAR (13-3B5(anti-Ig1)/3-28(anti-Ig2)/24C10(anti-Ig3)/192-8 (anti-Fz)). T cell (MuSK-CAART) technology is based on a clinically approved Anti-MuSK mAbs, mixed or individually, induced IFNγ production anti-CD19 chimeric antigen receptor T cell (CART-19) therapy that has and MuSK-CAART proliferation in a concentration-related manner 15,16 led to complete and durable remissions of B cell malignancies . We (Fig. 2c,d and Extended Data Fig. 4b,c). investigated MuSK-CAART efficacy and safety in preclinical models, To evaluate whether Fc-gamma receptors (FcγRs) or neonatal which support MuSK-CAART as a precision cellular immunotherapy Fc-receptor (FcRn) mediates indirect lysis by MuSK-CAART after with potential to induce complete and durable remission of MuSK MG. binding anti-MuSK antibodies, primary human monocytes (which express high-affinity FcγR/CD64, CD32 and FcRn) and natural killer Results (NK) cells (which express low-affinity FcγR/CD16) were cocultured MuSK-CAAR targets disease-relevant anti-MuSK B cell with MuSK-CAART and normal human IgG, mixed anti-MuSK IgG4 epitopes mAbs, purified plasma IgG from patients with MuSK MG or anti-CD3 MuSK is a transmembrane tyrosine kinase whose ectodomain comprises positive-control antibody (Fig. 2e,f ). After coincubation with three immunoglobulin-like (Ig1–Ig3) and frizzled-like (Fz) domains. An MuSK-CAART or NTD-T in the presence of anti-MuSK IgG4 mAbs or estimated 100, 58 and 23% of sera from patients with MuSK MG recog- purified plasma from IgG from patients with MG, caspase-3/7-positive nize Ig1, Ig2 and Ig3-Fz domains, respectively . To target anti-MuSK B monocytes or NK cells were similar to counts observed in the pres- cells, we designed a MuSK-CAAR comprising the complete MuSK ecto- ence of normal human IgG and less than those observed with anti-CD3 domain, linked to CD137-CD3ζ costimulatory and activation domains, positive-control antibody (UCHT1). and confirmed expression on primary human T cells (Fig. 1a,b). To validate MuSK-CAART cytotoxicity, we generated Nalm-6 B BCR and CD19-targeted lysis show similar in vivo efficacy cells expressing anti-MuSK domain-specific BCRs isolated from three We next evaluated whether anti-MuSK BCR-targeted cytolysis patients with MuSK MG or three MuSK-immunized mice. Epitope map- demonstrates comparable efficacy as CD19-targeted cytolysis in ping and previous literature confirmed MuSK domain specificity: eliminating anti-MuSK B cells in vivo, using a well-characterized nod 17 18 anti-MuSK Ig1 (13-3B5, ref. ), anti-MuSK Ig2 (189-1, ref. and 3-28, scid gamma (NSG) Nalm-6 xenograft model. NSG mice were engrafted 12,18 12,18 ref. ), anti-MuSK Ig3 (24C10) and anti-MuSK Fz (4A3, refs. and with mixed luciferase-expressing 13-3B5/3-28/24C10/192-8 or 4A3 192-8) (Extended Data Fig. 1a–c). Anti-MuSK BCR density on engineered Nalm-6 cells (anti-Ig1/Ig2/Ig3/Fz), followed by treatment with Nalm-6 cells was within twofold of IgG BCR density on primary human MuSK-CAART, NTD-T or CART-19. In parallel experiments to evaluate B cells (Extended Data Fig. 1d–f ). the potential neutralizing effect of anti-MuSK IgG on MuSK-CAART MuSK-CAART demonstrated specific lysis of Nalm-6 cells target- in vivo, NSG mice were engrafted with 13-3B5/anti-Ig1 or 13-3B5* ing each MuSK domain, but not control Nalm-6 cells (Fig. 1c–g and antibody-secreting Nalm-6 cells (described in Methods and Extended Extended Data Fig. 2), with increased specific cytotoxicity at 24 versus Data Fig. 5). 5 hours of coculture. Donor-matched nontransduced (NTD) T cells Bioluminescence imaging from all four experiments indicated (NTD-T) showed no cytotoxicity. IFNγ was detected in supernatants that CART-19 and MuSK-CAART significantly reduced Nalm-6 out- of MuSK-CAART cocultured with anti-MuSK Nalm-6 but not control growth relative to NTD-T (Fig. 3a,b), although recurrence of biolumines- Nalm-6 cells (Fig. 1h). cence flux signal was observed in a subset of MuSK-CAART-treated and CART-19-treated mice. Nalm-6 recurrence in CART-19-treated mice Anti-MuSK Abs have varying effects on MuSK-CAART activity engrafted with Nalm-6 13-3B5 cells was not due to loss of T cells, as MuSK autoantibodies might block MuSK-CAAR engagement with T cells were detectable in most CART-19-treated mice (Fig. 3c). Similarly, anti-MuSK BCRs, but they could also potentiate cytotoxicity by acti- Nalm-6 outgrowth in MuSK-CAART-treated mice engrafted with mixed vating MuSK-CAART. Anti-MuSK IgG concentrations in sera from Nalm-6 (192-8) cells was not due to failure of MuSK-CAART trafficking −1 patients with MuSK MG range from 0.5 to 49.5 nM (0.16–7.4 µg ml ), to cranial bone marrow (Extended Data Fig. 6c,d). In the subset of 11,19,20 assuming one antibody bound per MuSK molecule . To deter- mice (n = 3) with residual Nalm-6 target cells in cranial bone marrow, mine soluble anti-MuSK antibody effects on MuSK-CAART activity, the percentage of IgG BCR Nalm-6 cells was comparable between we performed cytotoxicity assays in the presence of a physiologic MuSK-CAART- and NTD-T-treated mice, but IgG BCR expression level −1 concentration (10 mg ml ) of polyclonal IgG from patients with MG, was significantly reduced in MuSK-CAART-treated mice (Extended or individual or mixed monoclonal antibody (mAb), at concentrations Data Fig. 6c–g). The low BCR density of Nalm-6 192-8, which is greater within or exceeding the expected range for autoantigen-specific IgG than two standard deviations lower than the mean density on pri- −1 + (0.2–25 µg ml ). Relative binding of MuSK IgG from patients MG3 mary human IgG B cells (Extended Data Fig. 1f ), may explain Nalm-6 and MG5 and anti-MuSK IgG4 mAbs appear in Extended Data Fig. 3. recurrence, since low target antigen density negatively affects CART MuSK-CAART cytotoxicity against mixed Nalm-6 target cells (13-3B5/3- cytotoxic efficacy . In MuSK-CAART-treated mice engrafted with 28/24C10/192-8 (anti-Ig1/Ig2/Ig3/Fz)) was partly inhibited by poly- mixed Nalm-6 (4A3) cells, all target cells were eliminated, comparable clonal IgG from patients with MuSK MG, although specific cytotoxicity to CART-19-treated mice, and no target cell recurrence was observed generally increased with higher effector to target ratios and longer (Fig. 3a, right). Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z a Gating strategy Extracellular domains TMD 0.023% **** Native MuSK Ig1 Ig2 Ig3 Fz Intracellular domain 72.3% MuSK CAAR CD137-CD3ζ Ig1 Ig2 Ig3 Fz FSC-A FSC-A Anti-MuSK-APC Linker CD8α TMD NTD-T MuSK-CAART c Nalm-6 control d Nalm-6 13-3B5 (anti-Ig1) e Nalm-6 189-1 (anti-Ig2) f Nalm-6 24C10 (anti-Ig3) 120 120 120 120 100 100 100 100 80 80 80 80 60 60 60 60 40 40 40 40 20 20 20 20 0 0 0 –20 –20 –20 –20 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 E:T ratio E:T ratio E:T ratio E:T ratio g h NTD-T Nalm-6 192-8 (anti-Fz) 5 hours MuSK-CAART CART-19 NTD-T 100 20 *** **** *** MuSK-CAART *** 80 **** **** **** CART-19 40 **** **** **** 24 hours –20 NTD-T 0 10 20 30 40 0 MuSK-CAART Nalm-6: Control 13-3B5 189-1 24C10 Control 192-8 E:T ratio CART-19 Fig. 1 | MuSK-CAAR expression on primary human T cells directs specific Nalm-6 cells (c) and Nalm-6 anti-MuSK target cells 13-3B5/anti-Ig1 (d), 189-1/anti- cytolysis of anti-MuSK B cells that target unique epitopes. a, Native MuSK Ig2 (e), 24C10/anti-Ig3 (f) and 192-8/anti-Fz (g) was measured using a luciferase- is a transmembrane tyrosine kinase whose ectodomain comprises three based cytotoxicity assay at 5 h (dashed line) and 24 h (solid line) after coculture immunoglobulin-like (Ig1–Ig3) and frizzled-like (Fz) domains. MuSK-CAAR with NTD-T (black), MuSK-CAART (red) and CART-19 (blue). The effector to target comprises the native MuSK ectodomain, followed by a glycine/serine-rich (E:T) ratio is based on total T cell number. h, Human IFNγ was measured in NTD-T, linker, CD8α transmembrane domain (TMD) and CD137-CD3ζ intracellular MuSK-CAART or CART-19 coculture supernatants (10:1 E:T, 24 h, experiments run costimulatory and activation domains. b, Primary human T cells were transduced using different T cell batches are shown in separate plots). One-way analysis of with MuSK-CAAR lentivirus or NTD-T, and MuSK-CAAR expression was detected variance (ANOVA) with the Holm–Sidak test for multiple comparisons. For c–h, using anti-MuSK 4A3 or 189-1 antibody. MuSK-CAAR transduction efficiency in error bars indicate mean ± s.d. of triplicate cocultures and are representative of six different donor T cell batches (NTD-T and MuSK-CAART). Error bars indicate 2–4 independent experiments. NS, P > 0.05; *P < 0.05; ***P < 0.001; ****P < 0.0001. mean ± s.d. Unpaired t-test (two-tailed). c–g, Cytolysis of wild-type (control) Higher MuSK-CAART percentages were observed in mixed MuSK-CAART efficacy in a syngeneic MuSK EAMG model Nalm-6 and 13-3B5*-engrafted mice (Fig. 3c), potentially due to solu- To determine whether MuSK-CAART demonstrates efficacy in an immu- ble mAb-induced expansion or proliferation of MuSK-CAART in vivo nocompetent mouse model that recapitulates features of autoim- after target cell encounter. MuSK-CAAR expression remained stable mune pathophysiology, including rare anti-MuSK B cells and polyclonal in T cells isolated from spleen (Fig. 3d) compared to MuSK-CAART autoantibodies, we evaluated MuSK-CAART efficacy in a syngeneic infusion product. However, anti-CD19 CAR expression was lower in MuSK experimental autoimmune myasthenia gravis (EAMG) model. spleen T cells (Fig. 3d) compared to CART-19 infusion product, which Classic MuSK EAMG models involve immunization of mice with rat 23,24 may explain Nalm-6 recurrence in CART-19-treated mice. MuSK ; use of human MuSK for EAMG induction is associated with In 13-3B5*-xenografted mice, anti-MuSK antibody titer increased variable symptom onset, although symptomatic mice demonstrate in mice treated with NTD-T from day 5 to 15, whereas titers in CART- reduction in miniature endplate potentials and postsynaptic AChR 25–27 19- and MuSK-CAART-treated mice were significantly reduced com- density, consistent with an MG phenotype . After day 0 immuniza- pared to NTD-T-treated mice by day 15, 11 days after T cell injection tion and day 26 boost of C57BL/6 (CD45.2 ) mice with human MuSK (Fig. 3e). Anti-MuSK IgG binding to diaphragm muscle cells in CART- ectodomain fragments, anti-MuSK IgG B cells comprised <0.2–1.5% of 19- and MuSK-CAART-treated mice was also reduced compared to total splenocyte IgG B cells gathered on day 34 (Extended Data Fig. 7a). NTD-T-treated mice (Fig. 3f ). Serum anti-MuSK antibodies were predominantly of the IgG1 and Taken together, these data indicate that MuSK-CAART can target IgG2c subclasses (Extended Data Fig. 7b). Anti-MuSK antibodies tar- MuSK Ig1/Ig2/Ig3/Fz domain-specific cells with comparable efficacy geted all four MuSK domains (Extended Data Fig. 7c). Immunized mice to CART-19. were treated on day 35 with murine CD45.1 T cells expressing human Nature Biotechnology Specific lysis (%) Specific lysis (%) MuSK-CAART NTD-T SSC-A –1 IFNγ (ng ml ) FSC-H FSC-H MuSK CAAR (%) MuSK mAb mixture MuSK mAb mixture Article https://doi.org/10.1038/s41587-022-01637-z 8 hours 24 hours NTD-T MuSK-CAART NTD-T MuSK-CAART Nalm-6 control 100 100 + medium + MG3 IgG 50 50 50 + MG5 IgG Nalm-6 mixed target 0 0 0 + medium + MG3 IgG + MG5 IgG –50 –50 –50 –50 (E:T ratio) 1:1 10:1 1:1 10:1 1:1 10:1 1:1 10:1 Nalm-6 control Nalm-6 mixed (anti-Ig1/Ig2/Ig3/Fz) (E:T ratio) 1:1 10:1 1:1 10:1 10 10 NS NTD-T *** NS NS 8 8 NS MuSK-CAART ** ** 6 6 ** ** ** 4 4 NS NS 2 2 0 0 MG3 IgG: – – + + – – – – + + – – – – + + – – – – + + – – MG5 IgG: – – – – + + – – – – + + – – – – + + – – – – + + c d NTD-T MuSK-CAART 3 Medium Unstained –1 0.2 μg ml Medium –1 –1 1 μg ml 0.2 μg ml –1 –1 1 μg ml 5 μg ml –1 5 μg ml –1 25 μg ml –1 25 μg ml CTV NTD-T MuSK-CAART e f Monocytes Monocytes NK cells NK cells 8 NTD-T 8 6 MuSK-CAART MuSK-CAART NTD-T 6 6 4 4 4 4 2 2 2 2 0 0 0 0 0 20 40 60 0 20 40 60 0 20 40 60 0 20 40 60 Time (h) Time (h) Time (h) Hours –1 –1 –1 –1 0.25 –g ml NH IgG 0.25 –g ml mixed mAbs 10.0 mg ml NH IgG 0.25 –g ml UCHT1 No antibody –1 –1 –1 –1 2.5 –g ml NH IgG 2.5 –g ml mixed mAbs 9.5 mg ml MG3 IgG 2.5 –g ml UCHT1 –1 –1 –1 –1 25 –g ml NH IgG 25 –g ml mixed mAbs 10.0 mg ml MG5 IgG 25 –g ml UCHT1 Fig. 2 | Evaluation of soluble anti-MuSK antibody effects on MuSK-CAART mAb concentration shown) and human IFNγ was quantitated by ELISA after cytotoxicity, IFNγ production and proliferation. a, NTD-T and MuSK- 24 h in duplicated samples (c), or proliferation of NTD-T and MuSK-CAART was CAART were coincubated with Nalm-6 control or mixed target cells (13-3B5/3- evaluated using CTV dye dilution by flow cytometry at 96 h (d). Representative 28/24C10/192-8 (anti-Ig1/Ig2/Ig3/Fz)) at 1:1 or 10:1 E:T ratios in the presence of flow plots from two individual experiments are shown. e,f, NTD-T or MuSK- −1 purified IgGs (10 mg ml ) from two patients with MuSK MG (MG3 and MG5; CAART were coincubated with monocytes (e) or NK cells (f) at a 5:1 E:T ratio in details in Methods) or medium alone. Cytotoxicity was evaluated at 8 and 24 h the presence of normal human IgG, mixed anti-MuSK mAbs (13-3B5/3-28/192-8 using a luciferase-based cytotoxicity assay. Error bars indicate mean ± s.d. of (anti-Ig1/Ig2/Fz), total mAb concentration shown), purified polyclonal IgG from triplicates. b, Human IFNγ was measured in NTD-T or MuSK-CAART coculture plasma from patients with MuSK MG (MG3 and MG5) or an anti-CD3 positive- supernatants (24 h) in two independent experiments. Two-way ANOVA with control mAb (clone UCHT1) for 48 h. Monocyte/NK cell death was detected by Tukey’s test for multiple comparisons: NS, P > 0.05; **P < 0.01; ***P < 0.001. incorporation of caspase-3/7 dye over time. Fold change of caspase cells relative c,d, NTD-T and MuSK-CAART were incubated with an equimolar mixture of to the 0 hour timepoint is plotted. Error bars indicate mean ± s.d. of triplicates. anti-MuSK IgG4 mAbs (13-3B5/3-28/24C10/192-8 (anti-Ig1/Ig2/Ig3/Fz), total MuSK-CAAR, a 1D3/anti-CD19 CAR with modified murine CD28-CD3ζ numbers of CAR/CAAR-transduced cells and matching numbers of signaling domains that confer greater in vivo efficacy as a positive con- NTD-T (Extended Data Fig. 7e). 1D3-CART and MuSK-CAART exhibited trol or NTD-T (Extended Data Fig. 7d). Mice were dosed with equivalent similar CD4:CD8 T cell ratios (Extended Data Fig. 7f ). Nature Biotechnology –1 –1 Caspase (fold change) Specific lysis (%) IFNγ (ng ml ) IFNγ (ng ml ) ND ND ND ND ND ND Caspase (fold change) Percentage of maximum **** **** ** **** NS **** * ** NS Article https://doi.org/10.1038/s41587-022-01637-z a b Nalm-6 mixed: anti-Ig1/Ig2/Ig3/Fz Nalm-6 mixed: anti-Ig1 13-3B5/3-28/24C10/4A3 13-3B5* (Ab-secreting) 13-3B5/3-28/24C10/192-8 13-3B5 11 11 10 10 10 10 10 10 9 9 10 10 8 8 10 10 7 7 10 10 6 6 10 10 0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25 Days after target cell injection Days after target cell injection Days after target cell injection Days after target cell injection c d IP Spleen IP Spleen IP Spleen IP Spleen 120 120 120 * * ** * NS NS NS NS 100 NS 100 100 NS NS **** NS ** NS NS *** *** 80 80 80 60 60 60 40 40 40 20 20 20 0 0 0 + 192-8 + 4A3 13-3B5 13-3B5* Nalm-6 mixed target Nalm-6 target Nalm-6 mixed target Nalm-6 target Nalm-6 mixed target Nalm-6 target e f **** NTD 10 **** NTD-T NS CART-19 ** CART-19 6 MuSK-CAART 0.2 MuSK-CAART 0.1 0 50 �m Neg control Day 5 Day 15 Fig. 3 | Targeting of anti-MuSK B cells through the BCR with MuSK-CAART cells relative to the infusion product (IP) (d). Error bars indicate mean ± s.e.m. demonstrates comparable efficacy as anti-CD19 CAR-mediated cytolysis, in One-sample t-test. Two MuSK-CAART-treated mice in Nalm-6 13-3B5/13-3B5* and −1 the presence or absence of soluble anti-MuSK antibody. a,b, Total flux (p s , two CART-19-treated mice in Nalm-6 13-3B5 experiments that were used for long- photons per second) after injection of 1:1:1:1 mixed 13-3B5/3-28/24C10/192-8 term follow-up were excluded from analysis. e, Anti-MuSK antibody titer in blood or 13-3B5/3-28/24C10/4A3 (anti-Ig1/Ig2/Ig3/Fz) (a), 13-3B5 (anti-Ig1) or 13-3B5* samples was measured on days 5 and 15 after target cell injection and quantitated (anti-Ig1, antibody-secreting) Nalm-6 cells (b), followed 4 days later by treatment relative to a 13-3B5 IgG4 mAb standard. Two-way ANOVA with Dunnett’s test for with 1 × 10 NTD-T (black, n = 5), CART-19 (blue, n = 5) or MuSK-CAART (red, n = 5). multiple comparisons. f, Direct immunofluorescence analysis of diaphragm Bioluminescence images appear in Extended Data Fig. 6a. One-way ANOVA with muscle harvested on day 25, stained with antihuman IgG to detect 13-3B5 IgG4 the Holm–Sidak test for multiple comparisons, day 23. c,d, Splenocytes were binding to MuSK on the muscle cell surface. Diaphragms from 13-3B5/NTD- analyzed on days 24 or 25 after target cell injection for CD3 T cell frequency T-treated mice served as a negative control. Scale bar, 50 µm. Representative (c) (median values are indicated; Kruskal–Wallis test with Dunnett’s test for images are shown from two independent staining experiments. NS, P > 0.05; + + multiple comparisons), and percentage of MuSK-CAAR and anti-CD19 CAR T *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. After treatment, few to no B cells were detected in the spleen in addition to the rarity of anti-MuSK B cells relative to CD19-expressing and lymph nodes of 1D3-CART-treated mice, whereas NTD-T- and B cells in immunized mice and/or the differing signaling and costimula- MuSK-CAART-treated mice showed comparable B cell frequency (Fig. 4b). tory domains of MuSK-CAAR and 1D3-CAR (Fig. 4f and Extended Data Anti-MuSK antibody titer was comparable across treatment groups Fig. 7a,d). before treatment (Fig. 4c) and progressively increased in NTD-T-treated These data indicate that in an EAMG model with rare anti-MuSK mice, whereas titers in MuSK-CAART- and 1D3-CART-treated mice sig- target B cells and circulating anti-MuSK antibodies, MuSK-CAART nificantly decreased, starting 1 week after injection (Fig. 4d). Unlike treatment results in antigen-specific IgG depletion without total 1D3-CART, which significantly decreased total serum IgG, MuSK-CAART B cell depletion and does not require previous lymphodepletion for did not reduce total serum IgG levels relative to NTD-T-treated mice therapeutic effect. (Fig. 4e). CD45.1 T cells were detected in spleen and lymph nodes with relatively lower percentages of engrafted T cells in MuSK-CAART-treated Specific off-target toxicity by MuSK-CAART was not observed mice, and higher T cell percentages in 1D3-CART-treated mice, the lat- To evaluate for potential off-target interactions of MuSK-CAART, ter in part due to the twofold higher number of T cells injected in set 1, we performed (1) comprehensive organ histopathology in an NSG Nature Biotechnology + 192-8 + 4A3 13-3B5 13-3B5* + 192-8 + 4A3 13-3B5 13-3B5* –1 –1 Total flux (p s ) T cell (%) Anti-MuSK Ab (�g ml ) MuSK CAAR (%) –1 Total flux (p s ) CD19 CAR (%) *** NS *** ** NS ** Article https://doi.org/10.1038/s41587-022-01637-z a b Gating strategy Spleen Lymph nodes NS 120 NS ** ** FSC-A FSC-A CD45.2-PE NTD-T 1D3-CART MuSK-CAART NTD-T 1D3 MuSK NTD-T 1D3 MuSK CART CAART CART CAART CD19-APC c NS d e Anti-MuSK Ab titer Total mouse IgG 2.5 NS NS Set 1 + Set 2 Set 1 + Set 2 2.0 2 8 1.5 1.0 –1 4 –2 0.5 –3 –4 2 0 1 2 3 4 5 0 1 2 3 4 5 Pretreatment Weeks after treatment Weeks after treatment Spleen Lymph nodes 2.5 Set 1 Set 2 Set 1 Set 2 2.0 Set 1 Set 2 NTD-T 1.5 1D3-CART 1.0 MuSK-CAART 0.5 −1 Fig. 4 | MuSK-CAART reduces anti-MuSK IgG but not total IgG or B cell counts (µg ml ). Kruskal–Wallis with Dunnett’s test for multiple comparisons. in a syngeneic MuSK EAMG model. CD45.2 C57BL/6J mice were immunized d,e, Anti-MuSK antibody titer and total mouse IgG were measured in mouse with MuSK Ig1-2 protein (30 µg in complete Freund’s adjuvant) on day 0 and blood samples drawn weekly after treatment. Graphs indicate fold change of anti- boosted with MuSK Ig1-Fz protein (30 µg in incomplete Freund’s adjuvant) on day MuSK antibody titer (d) or total mouse IgG (e) relative to week 1 after treatment 26. Results of two independent experiments are shown (total numbers NTD-T, (NTD-T, n = 8; 1D3-CART, n = 7 (d) and n = 6 (e); MuSK-CAART, n = 8 to include all n = 12, 1D3-CART, n = 8, MuSK-CAART, n = 12). Equivalent numbers of transduced mice with longitudinal samples through week 4 after treatment; one 1D3-CART + + CD45.1 T cells or a matching number of NTD CD45.1 cells were injected on day mouse was excluded in e due to low blood sample volume precluding analysis). + - + 35. a,b, Host B cells (CD45.2 CD3 CD19 ) were analyzed in spleen and lymph Error bars indicate mean ± s.e.m. Multiple linear regression-coefficient test for nodes at days 49–63 (2–4 weeks after treatment). Representative flow plot (a) difference between the slopes. f, Frequency of CD45.1 T cells were analyzed in + + and the frequency of CD45.2 CD19 B cells in the spleen and lymph nodes (b) are the spleen and lymph nodes on day 49–63. Statistical analysis was not performed shown. Kruskal–Wallis test with Dunnett’s test for multiple comparisons. c, Anti- since absolute number of CD45.1 T cells varied among treatment groups to MuSK antibody titer was measured in individual mouse blood samples drawn on achieve the same transduced cell dose in set 1. NS, P > 0.05; *P < 0.05; **P < 0.01; the day of treatment, normalized to 4A3 mouse antihuman MuSK mAb standard ***P < 0.001. 7 7 xenograft model, (2) screening of a high-throughput human mem- followed by treatment with vehicle, 1 × 10 NTD-T, 1 × 10 CART-19 or 6 7 brane proteome array (MPA), (3) primary human cell screens and (4) 3 × 10 –1 × 10 MuSK-CAART donor-matched cells. Representative targeted investigations based on potential MuSK-interacting proteins. bioluminescence images and graphs of total bioluminescence flux MuSK-CAART biodistribution was evaluated in 146 NSG mice allo- indicate control of Nalm-6 cell outgrowth in CART-19 and high-dose cated to 19 groups, injected with 1 × 10 3-28 Nalm-6 or no target cells, MuSK-CAART-treated mice, and delayed outgrowth in low-dose Nature Biotechnology CD45.1 (%) Normalized anti-MuSK Ab CD3-BV421 SSC-A FSC-H Fold change (relative to week 1) FSC-H + + CD45.2 CD19 (%) Fold change (relative to week 1) Article https://doi.org/10.1038/s41587-022-01637-z a MuSK-CAART b Vehicle Day 36, liver Day 36, lung Vehicle (n = 6) High-dose Low-dose CART-19 NTD-T 7 6 7 7 (10 cells) (3 × 10 cells) (10 cells) (10 cells) MuSK-CAART high-dose (n = 24) 3.0 MuSK-CAART low-dose (n = 24) CART-19 (n = 24) NTD-T (n = 8) ns 2.0 200 μm 200 μm 1.0 ** 7 10 ** (x10 ) 1 4 8 14 Days after target –1 2 –1 Radiance (p s cm sr ) cell injection 4 7 500 μm 1,000 μm Color scale: min = 4.17 × 10 , max = 3.04 × 10 c d Validation screen e Validation screen (positive controls removed) 9,000 8,000 Protein A MMP16 8,000 7,000 4A3 MMP16 7,000 MMP16 6,000 Vector Vector 6,000 5,000 5,000 4,000 4,000 3,000 3,000 2,000 2,000 1,000 1,000 20.0 5.0 1.3 0.3 20.0 5.0 1.3 0.3 20.0 5.0 1.3 0.3 –1 –1 –1 MuSK-Fc (μg ml ) MuSK-Fc (μg ml ) PD-1-Fc (μg ml ) Membrane proteome array f 800 g NTD-T Target only MuSK-CAART **** **** **** + Staurosporin + NTD-T + MuSK-CAART Nalm-6 control Nalm-6 3-28 (anti-Ig2) h i 160 200 Target only Target only ** ** *** ** 120 + Staurosporin 150 *** ** *** + Staurosporin or bortezomib + NTD-T 80 100 + MuSK-CAART + NTD-T 40 50 + MuSK-CAART HREC HUVEC NHEM Fig. 5 | Off-target cytotoxic interactions of MuSK-CAART were not identified curve overlaps with vector control). A representative graph from two validation in mouse tissue or using human MPAs. a, Representative bioluminescence screens is shown in e. e, Positive controls are removed in validation screens shown images from the MuSK-CAART biodistribution study in mice injected with 3-28/ in d and the y axis is rescaled. One of two validation screens confirmed MuSK-Fc anti-Ig2 Nalm-6 cells, then treated with vehicle only (n = 6), NTD-T (n = 8), CART- binding to MMP16, defined as MFI at least twofold higher than isotype (PD-1-Fc) 19 (n = 24) or MuSK-CAART (high- and low-dose, n = 24 per each dose). Graph control at two or more concentrations. f–i, Cytotoxicity of MuSK-CAART from indicates total bioluminescence flux for all mice in each treatment group. Error two donor T cell batches was measured after coincubation for 24 hours (5:1 E:T bars indicate mean ± s.e.m. One-way ANOVA with the Holm–Sidak test for multiple ratio) with Nalm-6 wild-type (negative control), Nalm-6 3-28/anti-Ig2 (positive comparisons, day 14. b, Example images from liver and lung on day 36 after control) or seven primary human cell types from each of two different donors. treatment (MuSK-CAART, n = 8; CART-19, n = 8). Liver and lung sections from high- Representative results are shown. IFNγ production was not detected in MuSK- dose MuSK-CAART-treated mice show lymphocytic infiltration without cytotoxic CAART cocultures with primary human cells (f). Viability was analyzed using effect (black arrows). CART-19-treated mice demonstrated focal hepatocellular high-content imaging analysis (HCA) for Nalm-6 control and anti-MuSK 3-28 necrosis and focal pulmonary thrombus (black arrows). c, Human MPA screened cells (g) and human-derived cells (h) or by flow cytometry (i) at 24 hours, using with MuSK-Fc protein identified a potential binding signal with MMP16. d, MuSK- staurosporin or bortezomib as toxicity controls. Error bars indicate mean ± s.d. of Fc binding to MMP16 demonstrated low mean fluorescence intensity (MFI) relative triplicates (f–i). Multiple t-test (two-tailed), Holm–Sidak correction for multiple to protein A and anti-MuSK 4A3 positive controls in validation screening (MMP16 comparisons. NS, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Nature Biotechnology T cell only Neurons Cardiomyocytes Hepatocytes HBEC HREC HUVEC NHEM Nalm-6 control Nalm-6 3-28 1,000 2,000 3,000 4,000 5,000 6,000 Neurons Cardiomyocytes Hepatocytes HBEC Live target cells Normalized target binding –1 Day 14 Day 8 Day 4 Day 1 IFNγ (pg ml ) (% of target only) MFI (×1,000) –1 Total fux (p s ) Live target cells Live target cells (% of target only) (% of target only) MFI (×1,000) CART-19 MuSK-CAART 7 7 (1 × 10 ) (1 × 10 ) Article https://doi.org/10.1038/s41587-022-01637-z a MuSK-CAART CART-19 –1 Agrin (ng ml ): 0 3 10 0 3 10 b c d Agrin –1 2.0 0 ng ml 5 × 10 120 –1 NTD-T 3 ng ml NS **** **** 1.8 –1 MuSK-CAART 10 ng ml **** 4 × 10 **** ***** **** Wise-CAART 1.6 **** **** 3 × 10 1.4 2 × 10 1.2 NS NS 1 × 10 1.0 0 0.8 0 MuSK-CAART CART-19 Medium Agrin Non Agrin −1 Fig. 6 | MuSK-CAART off-target effects on muscle are not observed. a, Effects agrin (5 ng ml ) or medium alone. Agrin-induced MuSK phosphorylation was of MuSK-CAART or CART-19 (E:T 10:1) on agrin-induced AChR clustering in C2C12 confirmed by phospho-MuSK ELISA. OD value relative to medium-alone 450/570 mouse myotubes were visualized by α-bungarotoxin staining (×25 magnification, control is shown. Mann–Whitney U-test (two-tailed). d, Human myotubes top row, plus bottom inset (red boxes); scale bar, 50 µm). Representative images were coincubated with NTD-T, MuSK-CAART or Wise-CAART in the presence are shown from two individual experiments. b, AChR clustering was quantitated or absence of agrin for 24 hours. IFNγ production was measured in cell-culture by fluorescence intensity, measured at six different sites in each image (error bars supernatants by ELISA. Error bars indicate mean ± s.d. of triplicates. One-way indicate mean ± s.e.m.). One-way ANOVA with the Holm–Sidak test for multiple ANOVA with the Holm–Sidak test for multiple comparisons. NS, P > 0.05; comparisons. c, Differentiated primary human myotubes were incubated with *P < 0.05; ****P < 0.0001. MuSK-CAART-treated mice relative to vehicle and NTD-T-treated mice second validation screen. MuSK-CAART cytotoxicity was subsequently (Fig. 5a). Comprehensive organ histopathologic analysis indicated lym- evaluated against U87-MG glioma cells, which express MMP16 as well phocytic infiltration in multiple organs, most notably at late timepoints as LRP4 (Extended Data Fig. 8a). As a positive control, we gener- after injection of MuSK-CAART or CART-19 in mice with target cells. ated a CAAR comprising Wise, which interacts with LRP4 in an 29,30 Liver and lung sections from high-dose MuSK-CAART-treated mice show agrin-independent manner , and verified Wise-CAART cytotox- lymphocytic infiltration without cytotoxic effect, which could repre- icity against U87-MG cells (Extended Data Fig. 8b,c). In contrast to sent target cells or engraftment of the human T cell product (Fig. 5b). Wise-CAART, coincubation of MuSK-CAART with U87-MG cells ± agrin CART-19-treated mice demonstrated focal hepatocellular necrosis did not induce cell lysis or IFNγ production (Extended Data Fig. 8c–e). in some liver sections, consistent with xenogeneic graft-versus-host Additionally, screens of seven primary human or induced pluripotent disease (xGVHD), and focal pulmonary thrombus consisting of histo- stem cell (iPSC)-derived cells representing skin, vascular tissue, heart, cytes and fibrosis, attributed to the intravenous (i.v.) route of injection brain/nerves, lung, liver and kidney, which were validated for MMP16 (Fig. 5b). Specific off-target cytotoxic effects of MuSK-CAART relative or LRP4 expression, did not identify specific cytolysis or IFNγ produc- to CART-19-treated mice were not observed. tion (Fig. 5f–i). Because in vivo biodistribution studies of human cellular To evaluate whether MuSK-CAART interferes with AChR clustering immunotherapies are confounded by xGVHD and may not identify or causes muscle cell cytolysis due to potential trans-interaction with off-target interactions against human proteins, recombinant MuSK-Fc LRP4, mouse C2C12 myotubes were cocultured with MuSK-CAART or protein was used to screen an MPA comprising approximately 5,300 CART-19, which is known not to cause muscle cytotoxicity in humans. human membrane proteins overexpressed in mammalian cells, which MuSK-CAART did not affect agrin-induced AChR clustering in C2C12 identified binding to matrix metalloproteinase-16 (MMP16) in ini- myotubes relative to CART-19 (Fig. 6a,b). Primary human muscle cells tial screening (Fig. 5c). Validation screens indicated overall low-level were also differentiated into myotubes, and MuSK activation by agrin MuSK-Fc binding to MMP16 relative to positive and PD-1-Fc isotype was verified by MuSK phosphotyrosine enzyme-linked immunosorbent controls (Fig. 5d,e), which was interpreted as negative (less than assay (ELISA) (Fig. 6c). IFNγ was not elevated in MuSK-CAART or NTD-T twofold higher than vector control) in one validation screen coculture supernatants, whereas a significant increase in human IFNγ and positive (more than twofold higher than vector control) in a was detected in Wise-CAART cocultures (Fig. 6d). Nature Biotechnology Fluorescence Intensity (pixels) Fold change of phospho-MuSK –1 IFNγ (ng ml ) Article https://doi.org/10.1038/s41587-022-01637-z Discussion To further evaluate MuSK-CAART in a preclinical autoimmune We report the development of a novel precision cellular immunother- model, we expressed human MuSK-CAAR in murine T cells and com- apy for autoantigen-specific B cell depletion in MuSK MG. MuSK-CAART pared its efficacy to anti-CD19 1D3-CART in an immunocompetent demonstrated specific cytotoxicity against B cells targeting each of syngeneic MuSK EAMG model. Similar to human MuSK MG, anti-MuSK the four MuSK domains (Fig. 1). By incorporating the complete MuSK B cells comprised <0.2–1.5% of splenic IgG B cells and targeted all ectodomain, MuSK-CAART is designed to eliminate autoimmune B four MuSK ectodomains (Extended Data Fig. 7). Unlike human MuSK cells targeting a broad range of MuSK epitopes. MG, both anti-MuSK IgG1 and IgG2c were induced in this model A major difference in the application of MuSK-CAART and CART-19 (Extended Data Fig. 7b). Murine IgG1 is functionally analogous to to clinical practice is the presence of soluble autoantibodies that could human IgG4 (ref. ). Murine IgG2c fixes complement, which might have varying effects on MuSK-CAART function. Human anti-MuSK anti- mediate CAART destruction or cause toxicities that may not occur bodies are predominantly IgG4, which are functionally monovalent and in patients with MuSK MG. Despite these limitations, MuSK-CAART do not fix complement or activate antibody-dependent cellular cyto- specifically reduced anti-MuSK IgG but not total IgG or total B cells, 17,31,32 toxicity . Data in Fig. 2a indicate that anti-MuSK antibodies mod- indicating antigen-specific B cell depletion without previous lym- estly inhibit MuSK-CAART cytotoxicity in vitro, although cytotoxicity phodepletion (Fig. 4). Anti-MuSK IgG reductions by MuSK-CAART increases with longer coincubation times and higher effector to target and 1D3-CART were comparable, despite lower percentages of CD45.1 ratios, while remaining specific. Cytotoxicity in the presence of soluble T cells in MuSK-CAART versus 1D3-CART-treated mice (Fig. 4f ). Dif- autoantibody was further confirmed in an EAMG model (discussed ferences in T cell engraftment were in part due to number of injected further below). Anti-MuSK IgG4 mAbs as well as IgG from patients T cells in a subset of mice, but may also reflect IgG2c-mediated clear- with MuSK MG activate MuSK-CAART to produce IFNγ and/or prolife- ance of MuSK-CAART or the lower abundance of anti-MuSK versus rate (Fig. 2 and Extended Data Fig. 4). Anti-MuSK antibodies did not CD19 B cells, which induces less expansion of MuSK-CAART relative mediate indirect lysis of Fc-receptor-expressing cells by MuSK-CAART to 1D3-CART. (Fig. 2e,f ). Collectively, these data suggest that soluble autoantibod- Multiple complementary approaches were used to screen for ies could be beneficial by amplifying the infused MuSK-CAART dose potential MuSK-CAART off-target effects (Figs. 5 and 6). Biodistribu- (Figs. 2d and 3b and Extended Data Fig. 4c) and providing a survival tion studies in mice may identify unexpected organ cytotoxicity due signal in vivo; however, autoantibodies could also induce cytokine to cross-reactivity with mouse proteins in their native context, but release syndrome (Fig. 2b,c and Extended Data Fig. 4b) and/or inhibit are limited by xGVHD and nonhomology with human proteins. MPAs cytotoxicity (Fig. 2a and Extended Data Fig. 4a). In preclinical studies of allow high-throughput screening of thousands of human cell-surface desmoglein 3 (DSG3)-CAART for the treatment of mucosal pemphigus proteins that may not otherwise be expressed in primary human cells vulgaris, similar induction of CAART proliferation and IFNγ production or mice, but may identify irrelevant targets due to artifacts of protein 21,33 by soluble autoantibodies occurred , although data from the first overexpression. One of two MPA screens identified MuSK-Fc inter- four cohorts in the ongoing DSG3-CAART clinical trial (NCT04422912) action with MMP16 (Fig. 5c–e), although follow-up screens against indicate no cytokine release syndrome or dose-limiting toxicities, as MMP16-expressing cells did not confirm MuSK-CAART cytotoxicity well as a dose-related increase in DSG3-CAART persistence throughout (Fig. 5f–i and Extended Data Fig. 8). These studies also indicated that the 28 days following infusion , suggesting that soluble autoantibod- MuSK, which physiologically interacts with LRP4 in cis, does not inter- ies do not mediate adverse events or prevent DSG3-CAART engraft- act with LRP4 in trans on muscle and other primary cell types. ment. Nevertheless, to mitigate risk, dose escalation is planned in the These data contributed to an investigational new drug application MuSK-CAART phase 1 clinical study design. for MuSK-CAART as a novel precision cellular immunotherapy for the To investigate the in vivo efficacy of MuSK-CAART, we used two treatment of MuSK autoantibody-positive MG and informed a phase complementary approaches. The NSG Nalm-6 xenograft model is a 1 clinical study design (NCT05451212). CAAR T cells may represent a well-defined model that allows (1) evaluation of the cytolytic efficacy platform technology that could be applied to numerous autoimmune and engraftment of the human clinical product, MuSK-CAART, in com- and alloimmune B cell-mediated conditions. parison to clinically approved CART-19; (2) inclusion of anti-MuSK Nalm-6 cells that bind a broad range of epitopes relevant to MuSK Online content MG and (3) sensitive real-time analysis of target cell burden by biolu- Any methods, additional references, Nature Portfolio reporting sum- minescence imaging. These studies indicate that MuSK-CAART dem- maries, source data, extended data, supplementary information, onstrates in vivo efficacy comparable to CART-19, including cytolysis acknowledgements, peer review information; details of author con- of 13-3B5/anti-Ig1 Nalm-6 cells in the presence of matching soluble tributions and competing interests; and statements of data and code autoantibody that could inhibit MuSK-CAART cytotoxicity (Fig. 3). availability are available at https://doi.org/10.1038/s41587-022-01637-z. However, nearly 100% of Nalm-6 cells are MuSK-reactive, whereas MuSK-reactive B cells in patients with MuSK MG are rare (reported References to be less than 0.15% of circulating IgG B cells ). Additionally, rare 1. Rodriguez Cruz, P. M., Cossins, J., Beeson, D. & Vincent, A. BCR-negative or BCR-low Nalm-6 cells, which persist despite enrich- The neuromuscular junction in health and disease: molecular ment for anti-MuSK BCR expression, proliferate and mediate delayed mechanisms governing synaptic formation and homeostasis. escape from MuSK-CAART in the NSG xenograft model (Extended Data Front. Mol. Neurosci. 13, 610964 (2020). Fig. 6). In humans, BCR-negative plasma cells would not be targeted by 2. Stathopoulos, P. et al. Mechanisms underlying B cell immune MuSK-CAART. However, MuSK autoantibody titers markedly decrease dysregulation and autoantibody production in MuSK myasthenia after therapy with rituximab, which also does not deplete plasma cells, gravis. Ann. N. Y. Acad. Sci. 1412, 154–165 (2018). suggesting that most anti-MuSK antibodies are produced by short-lived 3. Borges, L. S. & Richman, D. P. Muscle-specific kinase myasthenia plasma cells that are continuously replenished from the CD20 (and gravis. Front. Immunol. 11, 707 (2020). + 11–13 BCR ) memory B cell pool . Additionally, studies have shown that BCR 4. Kim, N. et al. Lrp4 is a receptor for Agrin and forms a complex with signaling and internalization are mutually exclusive and that anergic B MuSK. Cell 135, 334–342 (2008). 35,36 cells downregulate their BCR by continuous recycling , suggesting 5. Zhang, B. et al. LRP4 serves as a coreceptor of agrin. Neuron 60, that either direct cytolysis of anti-MuSK B cells or selective pressure by 285–297 (2008). MuSK-CAART to downregulate anti-MuSK BCRs may result in favorable 6. Zong, Y. et al. Structural basis of agrin-LRP4-MuSK signaling. therapeutic outcome. Genes Dev. 26, 247–258 (2012). Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z 7. Bartoccioni, E. et al. Anti-MuSK antibodies: correlation with 25. Viegas, S. et al. Passive and active immunization models of myasthenia gravis severity. Neurology 67, 505–507 (2006). MuSK-Ab positive myasthenia: electrophysiological evidence for 8. Huijbers, M. G. et al. Longitudinal epitope mapping in MuSK pre and postsynaptic defects. Exp. Neurol. 234, 506–512 (2012). myasthenia gravis: implications for disease severity. J. 26. Ulusoy, C. et al. Preferential production of IgG1, IL-4 and IL-10 in Neuroimmunol. 291, 82–88 (2016). MuSK-immunized mice. Clin. Immunol. 151, 155–163 (2014). 9. Huijbers, M. G. et al. MuSK IgG4 autoantibodies cause myasthenia 27. Ulusoy, C. et al. Dental follicle mesenchymal stem cell gravis by inhibiting binding between MuSK and Lrp4. Proc. Natl administration ameliorates muscle weakness in MuSK-immunized Acad. Sci. USA 110, 20783–20788 (2013). mice. J. Neuroinflammation 12, 231 (2015). 10. Koneczny, I., Cossins, J., Waters, P., Beeson, D. & Vincent, A. 28. Kochenderfer, J. N., Yu, Z., Frasheri, D., Restifo, N. P. & Rosenberg, MuSK myasthenia gravis IgG4 disrupts the interaction of LRP4 S. A. Adoptive transfer of syngeneic T cells transduced with a with MuSK but both IgG4 and IgG1-3 can disperse preformed chimeric antigen receptor that recognizes murine CD19 can agrin-independent AChR clusters. PLoS ONE 8, e80695 (2013). eradicate lymphoma and normal B cells. Blood 116, 3875–3886 11. Marino, M. et al. Long-lasting rituximab-induced reduction of (2010). specific-but not total-IgG4 in MuSK-positive myasthenia gravis. 29. Ohazama, A. et al. Lrp4 modulates extracellular integration of Front. Immunol. 11, 613 (2020). cell signaling pathways in development. PLoS ONE 3, e4092 12. Stathopoulos, P., Kumar, A., Nowak, R. J. & O’Connor, K. C. (2008). Autoantibody-producing plasmablasts after B cell depletion 30. Karner, C. M. et al. Lrp4 regulates initiation of ureteric budding identified in muscle-specific kinase myasthenia gravis. JCI Insight and is crucial for kidney formation–a mouse model for 2, e94263 (2017). Cenani-Lenz syndrome. PLoS ONE 5, e10418 (2010). 13. Zografou, C., Vakrakou, A. G. & Stathopoulos, P. Short- and 31. de Taeye, S. W. et al. FcgammaR binding and ADCC activity of long-lived autoantibody-secreting cells in autoimmune human IgG allotypes. Front. Immunol. 11, 740 (2020). neurological disorders. Front. Immunol. 12, 686466 (2021). 32. Lilienthal, G. M. et al. Potential of Murine IgG1 and Human IgG4 14. Jiang, R. et al. Single-cell repertoire tracing identifies to inhibit the classical complement and fcgamma receptor rituximab-resistant B cells during myasthenia gravis relapses. activation pathways. Front. Immunol. 9, 958 (2018). JCI Insight 5, e136471 (2020). 33. Ellebrecht, C. T. et al. Reengineering chimeric antigen receptor 15. Maude, S. L. et al. Tisagenlecleucel in children and young adults T cells for targeted therapy of autoimmune disease. Science 353, with B-cell lymphoblastic leukemia. N. Engl. J. Med. 378, 439–448 179–184 (2016). (2018). 34. Chang, D. J. et al. A Phase 1 trial of targeted DSG3-CAART cell 16. Schuster, S. J. et al. Tisagenlecleucel in adult relapsed or therapy in mucosal-dominant pemphigus vulgaris (mPV) patients: refractory diffuse large B-cell lymphoma. N. Engl. J. Med. 380, early cohort data. Mol. Ther. 30, 373 (2022). 45–56 (2019). 35. Blery, M., Tze, L., Miosge, L. A., Jun, J. E. & Goodnow, C. C. 17. Huijbers, M. G. et al. MuSK myasthenia gravis monoclonal Essential role of membrane cholesterol in accelerated BCR antibodies: valency dictates pathogenicity. Neurol. internalization and uncoupling from NF-kappa B in B cell clonal Neuroimmunol. Neuroinflamm. 6, e547 (2019). anergy. J. Exp. Med. 203, 1773–1783 (2006). 18. Takata, K. et al. Characterization of pathogenic monoclonal 36. Hou, P. et al. B cell antigen receptor signaling and internalization autoantibodies derived from muscle-specific kinase myasthenia are mutually exclusive events. PLoS Biol. 4, e200 (2006). gravis patients. JCI Insight 4, e127167 (2019). 19. McConville, J. et al. Detection and characterization of MuSK Publisher’s note Springer Nature remains neutral with regard to antibodies in seronegative myasthenia gravis. Ann. Neurol. 55, jurisdictional claims in published maps and institutional affiliations. 580–584 (2004). 20. Matthews, I. et al. Muscle-specific receptor tyrosine kinase Open Access This article is licensed under a Creative Commons autoantibodies–a new immunoprecipitation assay. Clin. Chim. Attribution 4.0 International License, which permits use, sharing, Acta 348, 95–99 (2004). adaptation, distribution and reproduction in any medium or format, 21. Lee, J. et al. Antigen-specific B cell depletion for precision therapy as long as you give appropriate credit to the original author(s) and the of mucosal pemphigus vulgaris. J. Clin. Invest. 130, 6317–6324 source, provide a link to the Creative Commons license, and indicate (2020). if changes were made. The images or other third party material in this 22. Ramakrishna, S. et al. Modulation of target antigen density article are included in the article’s Creative Commons license, unless improves CAR T-cell functionality and persistence. Clin. Cancer indicated otherwise in a credit line to the material. If material is not Res. 25, 5329–5341 (2019). included in the article’s Creative Commons license and your intended 23. Jha, S. et al. Myasthenia gravis induced in mice by immunization use is not permitted by statutory regulation or exceeds the permitted with the recombinant extracellular domain of rat muscle-specific use, you will need to obtain permission directly from the copyright kinase (MuSK). J. Neuroimmunol. 175, 107–117 (2006). holder. To view a copy of this license, visit http://creativecommons. 24. Mori, S. et al. Antibodies against muscle-specific kinase impair org/licenses/by/4.0/. both presynaptic and postsynaptic functions in a murine model of myasthenia gravis. Am. J. Pathol. 180, 798–810 (2012). © The Author(s) 2023 Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Methods Evaluation of MuSK monoclonal antibody and IgG titers from Design and construction of plasmids patients with MG Lentiviral plasmids . Lentiviral plasmid pTRPE (provided Relative titers of each recombinant anti-MuSK monoclonal anti- by the Penn Center for Cellular Immunotherapies) was modified body or purified plasma IgG from patients with MG (MG3 and MG5) to express MuSK-CAAR constructs and anti-MuSK BCRs as was evaluated using a Luminex-based assay. In brief, purified MG3 −1 −1 follows: IgG (0.85 mg ml ) and MG5 IgG (2.2 mg ml ) were diluted 1:10, 1:50 The human MuSK ectodomain (representing amino acids 24–495) and 1:100. Recombinant anti-MuSK monoclonal antibodies were −1 was synthesized (Integrated DNA Technologies) with flanking 5′ evaluated at 0.1, 0.2, 0.5 and 1.5 µg ml concentrations. Diluted BamHI and 3′ NheI restriction sites. Gene fragments were digested samples were added to MuSK ectodomain (aa 24–495)-coupled micro- and purified using a PCR purification kit (Qiagen), then ligated into spheres and incubated for 1 h at room temperature. After washing 33 −1 the pTRPE-DSG3-CAAR vector upstream of sequences encoding a samples, 5 µg ml antihuman IgG-Biotin was added and incubated glycine-serine linker, CD8α transmembrane, CD137 costimulatory and for 1 h at room temperature, followed by incubation with 100 µl of −1 CD3ζ signaling domains. Wise (amino acids 24–206, UniProt Q6X4U4) streptavidin-PE (4 µg ml ) for 45 min. After washing, beads were resus- was synthesized (Integrated DNA Technologies) and subcloned into pended in 100 µl of washing buffer and 60 µl was analyzed using a pTRPE vector. Luminex 100TM/200TM analyzer according to the manufacturer’s Anti-MuSK BCR 189-1 (ref. ) (also known as MuSK 1A, anti-Ig2) recommendations. and 13-3B5 (ref. ) (anti-Ig1) was produced by synthesizing (Inte- grated DNA Technologies) the variable heavy and variable light chain In vitro transduction and expansion of CAR/CAAR T cells sequences with flanking BamHI/NheI and XhoI/Bsu36I restriction In vitro transduction and expansion of human CAR/CAAR T cells. sites. Gene fragments were digested, purified (PCR purification Bulk (mixed CD4/CD8) primary human T cells (from Penn Human kit, Qiagen) and ligated into a pRRL4.IgG4 vector following previ- Immunology Core or leukapheresis (Stem Express)) were cultured 33 −1 ously published methods , then subcloned into lentiviral plasmid in human T cell culture media supplemented with 100 IU ml rhIL-2 pTRPE to generate pTRPE.IgG4.Lambda.189-1. Anti-MuSK BCRs 4A3 (Proleukin) (CTS OpTmizer media (Invitrogen, A1048501) plus 5% 12 12,18 (ref. ) (anti-Fz), 3-28 (refs. ) (anti-Ig2) and 192-8 (human anti-Fz IgM, human AB serum (Gemini Bio-Products, 100–512) or Roswell Park sequences provided by K.C.O.) were produced similarly, except that the Memorial Institute (RPMI) media supplemented with 10% FBS, 10 mM kappa variable region was synthesized with flanking XhoI/BsiWI HEPES, 1% penicillin/streptomycin and 1% GlutaMax). T cells were sites, and the kappa constant region was synthesized and cloned activated/selected with anti-CD3/CD28-coated paramagnetic beads into a pGEM-T Easy vector (Promega) before ligation into pTRPE to (Thermo Fisher Scientific, 40203D) at a 3/1 bead/cell ratio. Lentivirus generate pTRPE.IgG4.Kappa (4A3, 3-28 and 192-8). An anti-Ig3 mouse was added at 24 h after activation, and cells were expanded in either hybridoma (24C10) was produced by immunization of mice with the static culture or a Xuri Bioreactor (GE Healthcare Lifesciences) until day human MuSK ectodomain (Genscript) and the variable heavy and light 9 to day 10 after activation, with media changes approximately every chain genes sequenced (Genscript) and synthesized (Integrated DNA 2 days and magnetic bead removal before cryostorage. Expression of Technologies). Gene fragments were digested using BamHI/NheI (for anti-CD19 CAR or MuSK-CAAR on human T cells was detected using the variable heavy chain) and XhoI/BsiWI (for the variable light chain), CD19-Biotin with Streptavidin-PE, CD19-PE or recombinant anti-MuSK purified (PCR purification kit, Qiagen) and ligated into pTRPE.IgG4. antibodies (189-1, 24C10 or 4A3) with antihuman IgG4-APC or anti- Kappa vector. mouse IgG1-PE. Packaging plasmids pRSV-Rev and pGAG/POL, plus envelope plasmid Pcl VSVg (Nature Technology Corporation) were used with In vitro transduction and expansion of mouse CAR/CAAR T cells. Lipofectamine 2000 (Life Technologies) for lentiviral preparation in Mouse T cells (CD45.1 C57BL6/J, Jackson Laboratory, strain 002014) 293T cells or Lenti-X 293T cells (Takara, 632180). were purified using a CD3 T cell enrichment kit (R&D Systems) and cultured using mouse T cell culture media (RPMI-10 media Antibody plasmids. Variable heavy chain or variable light chain supplemented with 10% FBS, 10 mM HEPES, 1% penicillin/strepto- sequences of MuSK-specific antibodies were cloned into AbVec vec- mycin, and 1% GlutaMAX). T cells (10 cells per ml) were activated tors (IgG4 heavy chain, kappa light chain or lambda light chain). A 1:1 for 36 h with Dynabead Mouse T-Activator CD3/CD28 (Thermo −1 mixture of variable heavy chain and variable light chain plasmids was Fisher Scientific, 11452D) in media supplemented with 50 IU ml −1 cotransfected into 293T cells to produce recombinant monoclonal rhIL-2 and 10 ng ml rhIL-7 (BioLegend, 581902), plus fresh 20 µM antibody. mAbs were purified from 293T culture supernatants using 2-mercaptoethanol. Polystyrene nontreated plates were coated −1 protein A chromatography (Invitrogen) according to the manufac- with 30 µg ml RetroNectin (Takara) at 4 °C overnight then blocked turer’s recommendations. with mouse T cell culture media for 30 min at room temperature, followed by centrifugation with retroviral supernatant by centri- Retroviral plasmids. pMSGV1.1D3-28Z.1-3 mut was obtained from fugation at 3,000 g for 2 h at 4 °C. Transduction was conducted on Addgene (107227) . pMSGV1.MuSK-CAAR was generated by replacing days 1 and 2 after activation by adding T cells directly onto the 1D3-28Z.1-3 mut insert with the MuSK-CAAR sequence. Then 30 µg retrovirus-coated wells. Plates were centrifuged at 300 g for 10 min of each plasmid was transfected into the Plat-E (Cell Biolabs) packaging at room temperature, then placed in a cell-culture incubator over- cell line to produce retroviruses. night (first round transduction). Second round transduction was performed similarly, except T cells were incubated 4–6 h after sec- Patient samples ond round transduction in mouse T cell culture media supplemented −1 −1 Patient characteristics. MG3, a 57-year-old female, chronic active with 10 ng ml rhIL-7 and 10 ng ml rhIL-15 (BioLegend, 570302) for disease. MG5, a 34-year-old female, chronic active disease. Venipunc- an additional 3 days. Media was replaced the next day and as needed ture was performed under a protocol approved by the University of to maintain the cell concentration between 1 and 2 × 10 cells per Pennsylvania Institutional Review Board. ml. Expression of anti-CD19 CAR or MuSK-CAAR on mouse T cells was detected on day 4 after activation using antimouse IgG(H+L)-APC MuSK MG IgG purification. IgG was purified from plasma from ( Jackson ImmunoResearch) or APC-conjugated recombinant patients MG3 and MG5 by protein A chromatography (Invitrogen) anti-MuSK antibody (189-1). Mouse T cells were injected on day 5 after according to the manufacturer’s recommendations. activation. Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Pharmacologic and toxicologic effects of soluble anti-MuSK Nalm-6 3-28, Nalm-6 24C10 and Nalm-6 192-8 or Nalm-6 4A3 cells, antibodies (2) Nalm-13-3B5 and (3) Nalm-6 13-3B5*. Nalm-6 13-3B5* cells were MuSK-CAART cytotoxicity in the presence of MuSK MG IgG. generated by introducing 13-3B5 IgG4 heavy chain (without a mem- Donor-matched MuSK-CAART and NTD-T were incubated with brane anchor) into Nalm-6 13-3B5 expressing 13-3B5 IgG4 heavy chain BCR-negative Nalm-6 control cells or mixture of Nalm-6 MuSK target (membrane-bound form) and 13-3B5 light chain, resulting in 13-3B5 cells (1:1:1:1 ratio of Nalm-6 13-3B5, Nalm-6 3-28, Nalm-6 24C10 and IgG4 antibody-secreting cells that retained cell-surface 13-3B5 BCR Nalm-6 192-8) at E:T ratio of either 1:1 or 10:1 for 24 h. Purified IgG expression (Extended Data Fig. 5). Donor-matched frozen human from two patients with MuSK MG (MG3 and MG5) was added at a final T cells (NTD-T, CART-19 and MuSK-CAART) were thawed 1 day before −1 concentration of 10 mg ml IgG before coincubation. MuSK-CAART the treatment in human T cell culture media supplemented with 100 IU cytotoxicity was evaluated at 8 and 24 h using luciferase-based killing of rhIL-2. On day 4 after target cell injection, 10 human T cells were assay. Coculture supernatants were harvested after completing the injected via tail vein. Two different infusion products from the same final plate reading at 24 h and stored at −20 °C for IFNγ ELISA. donor were used in mixed (Fig. 3a) or 13-3B5/13-3B5* (Fig. 3b) Nalm-6 experiments. IVIg was injected every 2–3 days in mixed and 13-3B5 IFNγ production and proliferation of MuSK-CAART by soluble Nalm-6-engrafted mice. antibodies. Soluble anti-MuSK antibody-induced MuSK-CAART activa- tion and proliferation was evaluated by IFNγ ELISA or Cell Trace Violet Bioluminescence imaging. Bioluminescence was measured with a (CTV) cellular labeling, respectively. For IFNγ ELISA, donor-matched Xenogen IVIS Lumina S3 (Caliper Life Sciences) from day 1 after target MuSK-CAART and NTD-T were incubated with a mixture of recombinant cell injection and every 2–3 days thereafter by injecting d-Luciferin −1 anti-MuSK monoclonal antibodies (1:1:1:1 of 13-3B5, 3-28, 24C10 and potassium salt (Gold Bio) intraperitoneally at a dose of 150 mg kg . 192-8) for 24 h. For the CTV cellular labeling, T cells were labeled with Mice were anesthetized with 2% isoflurane and luminescence was CTV Cell Proliferation Kit (Invitrogen) and activated with a mixture serially measured at 1 min intervals for 7 min or until signals start of anti-MuSK monoclonal antibodies for 96 h before analysis by flow to decrease in an automatic exposure mode. Total flux in the peak cytometry. image was quantified using Living Image 4.4 (PerkinElmer) by draw- ing rectangles from head to 50% of the tail length. Radiance unit of −1 2 −1 IncuCyte assay. Monocytes and NK cells mixed with green caspase-3/7 p s  cm  sr  = number of photons per second per square centimeter dye were incubated in media containing relevant IgGs, as well as effec- that radiate into a solid angle of one steradian. tor cells. The normal human IgG amount (negative control) and UCHT1 (a positive control, anti-CD3 antibody) was matched to the equivalent Human anti-MuSK antibody ELISA. Serum samples in NSG mice total amount of mixed anti-MuSK monoclonal antibodies or puri- with Nalm-6 13-3B5* were collected at day 5 and day 15 after target fied plasma IgG from patients with MG. Monocytes and NK cells were cell injection in K EDTA tubes for MuSK antibody ELISA. To detect cocultured with MuSK-CAART or NTD-T at a 5:1 E:T ratio. Anti-MuSK human anti-MuSK IgG4, the histidine-tagged recombinant extracel- monoclonal antibodies 13-3B5, 3-28, 192-8 were mixed at 1:1:1 ratio. lular domain of human MuSK (aa 24–495, R&D Systems, catalog no. Cocultures were monitored for 48 h using an IncuCyte S3 system 10189-MK) was coated on ELISA plates in PBS overnight at 4 °C at a −1 (Sartorius) and images were taken every 2 h with a ×20 objective, then concentration of 5 µg ml . Plates were washed with washing buffer (Inv- dead cells (green positive) were counted at each timepoint from four itrogen, catalog no. 00-0400-59), and blocked with Pierce Protein-Free fields of images per each well. The number of caspase-positive cells is (PBS) Blocking Buffer (Thermo Scientific, catalog no. 37572). Mouse equal to the mean number of green cell counts in each imaging field. serum samples were evaluated at a dilution of 1:50 to 1:100 in com- parison to a 13-3B5 purified recombinant human monoclonal IgG4 Luciferase-based in vitro cytotoxicity assay antibody as a reference standard for quantitation. Antihuman IgG Luciferase-based killing assay. Click-beetle green luciferase express- (H+L) HRP (Bethyl, catalog no. A80-119P) was used to detect human ing cells or luciferase U87-MG cells (ATCC, HTB14Luc2) were cocul- antibodies. Plates were protected from the light and placed in the tured with engineered T cells or donor-matched NTD-T at an indicated dark for 2 h. After washing plates three times, 100 µl of TMB (Thermo effector:target (E:T) ratio. At 3 h after coculture, luciferase substrate Scientific, 34028) was added for 30 min. Plate reading were conducted (d-luciferin potassium salt, GoldBio) was directly added to each well and using ELISA reader (Tecan, Infinite F-50) within 15 min after adding stop emitted light was measured on a luminescence plate reader (BioTek, solution (Invitrogen, SS04). Synergy HTX microplate reader) at indicated timepoints. The percent- age of specific lysis was calculated using the luciferase activity of 5% Flow cytometry analysis. Lymphocytes were isolated from cranial SDS-treated cells as maximum cell death and media alone as spontane- bone marrow using a previously reported protocol . In brief, the cal- ous cell death using the formula: specific lysis (%) = 100 × ((experimen- varia was cut into small pieces using sterile scissors and dissociated in tal data − maximum death data)/(maximum death data − spontaneous PBS + 2% FBS with a pestle. Spleens were harvested from mice, washed death data)). in PBS and cut into 0.5 mm cubes in ice-cold PBS. Spleen or bone mar- row isolates were transferred to a 70 µm cell strainer (Falcon, catalog Ethics statement for animal research no. 352350); cells were washed with PBS and resuspended in red blood All studies involving animals were performed under a protocol cell lysis buffer (BioLegend, catalog no. 420301). Cells were stained approved by the University of Pennsylvania Institutional Animal Care for 30 min on ice using the following antibodies: anti-CD3 BV711 or and Use Committee. anti-CD3-AF647 (clone okt3, BD Biosciences, catalog nos. 750983 and 566686)), anti-MuSK PE (clone 189-1 or 24C10), antihuman Ig light chain In vivo MuSK-CAART evaluation using NSG Nalm-6 xenograft λ PE (clone MHL-38, BioLegend, catalog no. 316608), antihuman Ig light models chain κ APC (clone MHK-49, BioLegend, catalog no. 316510), antihuman scid tm1Wjl Target cell and T cell injection. NSG (NOD.Cg-Prkdc IL2rg / IgG PE (BD Biosciences, catalog no. 555787) and/or antimouse IgG APC −1 SzJ) mice received 600 mg kg i.v. immunoglobulin (Privigen, IVIg) (clone A85-1, BD Biosciences). via tail-veil injection on day −2 and day −1 before target cell injection to prevent Fc-mediated Nalm-6 clearance. On day 0, three cohorts of Direct immunofluorescence. The diaphragm was collected at the NSG mice each received 10 Nalm-6 target cell line(s) via tail-vein injec- time of tissue harvest and embedded in optimal cutting tempera- tion as follows: (1) Nalm-6 mixed (a 1:1:1:1 mixture of Nalm-6 13-3B5, ture medium (Tissue-Tek); tissue blocks were frozen on dry ice before Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z storage at −80 °C. Next, 9 µm sections were cut onto Superfrost/Plus Wise-CAART were added and cocultured for 22–26 h. For some experi- glass slides (Fisher Scientific) and stored at −80 °C. Before staining, ments, neural agrin was added at least 30 min before T cell coculture. slides were equilibrated to room temperature, washed twice in PBS Evaluation of AChR clustering in C2C12 myotubes was performed (Gibco) and blocked in PBS containing 2% BSA. Slides were stained using by Invivotek, LLC. C2C12 mouse myoblast cells (ATCC 70024392) were FITC-conjugated antihuman IgG (BioLegend) at a dilution of 1:200 in maintained in DMEM supplemented with 20% FBS and 1% penicillin/ blocking solution and washed with PBS. Binding of IgG was visualized streptomycin. To induce myotube differentiation, C2C12 cells (80–90% with a Keyence imaging system (BZ-X710) and software (BZ-X Viewer). confluent) were cultured in DMEM plus 2% horse serum and 0.5% peni- cillin/streptomycin. Agrin-induced AChR clustering was examined by Syngeneic MuSK EAMG model incubating C2C12 myotubes in 1:1 DMEM differentiation media and Immunization and boosting schedule. CD45.2 C57BL/6J mice RESGRO serum-free culture medium (EMD Millipore, SCM002) sup- ( Jackson Laboratory, strain 000664) were immunized with MuSK plemented with varying concentrations of agrin (R&D Systems, catalog Ig1–Ig2 ectodomain fragment on day 0 and boosted with MuSK Ig1-Fz no. 550-AG/CF) for 14 h at 37 °C. AChR staining was performed using −1 full-length ectodomain protein on day 26. In initial experiments, 1 µg ml AlexaFluor 488-labeled α-bungarotoxin (Invitrogen catalog −1 20 mg kg busulfan was injected on day 34 in a subset of mice to evalu- no. B13422). Myotubes were fixed with 4% paraformaldehyde for 20 min ate effects on induced antibody titer and MuSK-CAART activity and at room temperature and imaged under mounting medium (Vector 6 6 engraftment. On day 35, 8 × 10 MuSK-CAART, 16 × 10 anti-CD19-CART Laboratories) using a Leica TCS SP8 multiphoton confocal micro- 6 6 (1D3) or 8 × 10 NTD-T in set 1 or 20 × 10 cells each in set 2 were admin- scope. Fluorescence was quantified (Fiji-Image J) as corrected total istered via i.v. injection. Analyses in Fig. 4a,b,f were performed 2 or cell fluorescence = integrated density − (area of selected cell × mean 4 weeks after T cell treatment (set 2 or 1, respectively). fluorescence of background). ELISA for total and subclass-specific mouse anti-MuSK IgG Off-target toxicity against primary human cells. MuSK-CAART reac- and total mouse IgG. Blood samples were collected weekly. To tivity against two different donor batches of primary or iPSC-derived detect mouse anti-MuSK antibodies in the syngeneic MuSK EAMG human cells was performed by Charles River Discovery Research model, mouse plasma (diluted 1:100 in PBS) was incubated on MuSK Services, including iPSC-derived iCell cardiomyocytes (Fuji Cellu- protein-coated plates and subsequently detected with antimouse lar Dynamics no. R1007/R1106), iPSC-derived iCell GABANeurons IgG-HRP (diluted 1:5,000, abcam, ab7061). Mouse anti-MuSK mono- (Fuji Cellular Dynamics, no. R1013/R1011), human bronchial epithelial clonal antibody (clone 4A3) was used as a reference standard control cells (HBEC, Lonza, no. CC-2540), primary hepatocytes (InnoProt, no. across experiments. Goat antimouse IgG1, IgG2b, IgG2c or IgG3-HRP P10651), human renal epithelial cells (HREC, InnoProt, no. P10664), (SouthernBiotech) was used as a secondary antibody reagent to deter- normal human epidermal melanocytes (NHEM, Lonza no. CC-2586 and mine anti-MuSK IgG subclasses. Total mouse IgG was measured by LifeLine no. FC-0030) and human umbilical vascular endothelial cells ELISA following manufacturer’s protocols (Invitrogen, CAT), after (HUVEC, Lonza, no. CC-2517). BCR Nalm-6 cells and Nalm-6 3-28 cells diluting sera 1:10,000 in PBS. were used as a negative control and a positive control, respectively. Primary or iPSC-derived human cells were cocultured for 24 h with two ELISpot. The frequency of anti-MuSK B cells in the spleen of sets of donor-matched NTD-T and MuSK-CAART at an E:T ratio of 5:1. MuSK-immunized mice were conducted using Mouse IgG ELISpot- MuSK-CAART cytotoxicity was detected at 24 h using either an basic kit (Mabtech, 3825-2H) according to the manufacturer’s proto- HCA with Nalm-6 cells, cardiomyocytes, hepatocytes and HBEC or flow col. Briefly, splenocytes were prestimulated with a mixture of R848 cytometry analysis with Nalm-6 cells, HREC, HUVEC and NHEM, respec- −1 −1 (1 µg ml ) and rmIL-2 (10 ng ml ) for 48 h. After prestimulation, cells tively. 1 µM staurosporin or 10 µM bortezomib (for HREC) was used were washed and resuspended in medium, then 10,000 or 100,000 B as a toxin control. For HCA, CAR/CAAR T cells were stained with Cell- −1 cells were plated in ELISpot wells precoated either with anti-IgG anti- Tracker Deep Red dye for 24 h, followed by incubation with 10 µg ml −1 −1 bodies (total IgG B cells) or MuSK protein (10 µg ml ), respectively. Hoechst and 4 µg ml propidium iodide in PBS supplemented with 0.5% BSA for 10 min at room temperature. Cells were imaged using the Flow cytometry analysis. Cells were isolated from the spleen and GE Healthcare IN Cell Analyzer 6000 (×10 magnification). Brightfield, lymph nodes after red blood cell lysis and stained with anti-CD45.1-FITC ultraviolet, dsRed and Cy5 channels were used to image brightfield, (BioLegend, 110706), anti-CD45.2-PECy7 (BioLegend, 109830), Hoechst staining, propidium iodide staining and CEllTracker staining, anti-CD3ε-BV421 (BioLegend, 100227) and anti-CD19-APC (BioLegend, respectively. Cell type-specific HCA was used to quantify live target 115512) for 30 min on ice. cells (IN Cell Developer Toolbox software (v.1.9.1)). Nuclear area of nonviable target cells was determined based on propidium iodide Evaluation for off-target interactions of MuSK-CAART staining and subtracted from total target cell nuclear area. For flow Calcein-AM staining. U87-MG cells, cultured to 70–90% confluency cytometry analysis, target cells were stained with 1 µM CellTracker −1 in a 12-well plate, were stained with 0.1–1 µM of Calcein-AM following Deep Red dye then incubated with 0.5 µg ml propidium iodide in the the manufacturer’s protocol (BD Pharmingen), then coincubated for presence of precision count beads before flow cytometry (Agilent 20–24 h with 1 × 10 MuSK-CAART, Wise-CAART or NTD-T. Then 5 nM NovoCyte Quanteon, FlowJo software v.10). Absolute count of live of neuronal agrin was added to U87-MG culture supernatants at least target cells was normalized to absolute bead count. 30 min before coculture. Measurement of IFNγ. Secreted IFNγ in coculture supernatants was Off-target toxicity against differentiated muscle cells. Primary detected by ELISA (BioLegend) or with a Luminex bead array platform human skeletal muscle cells (ZenBio, SKB-F) were differentiated for (Thermo Fisher) according to the manufacturer’s instructions. All 6–7 days using skeletal muscle cell differentiation medium (ZenBio, samples were analyzed in triplicate and compared against multiple SKM-D). Next, 5 nM of neural agrin (R&D Systems, catalog no. 550-AG/ internal standards, with a seven-point standard curve. CF) was added into differentiated primary human skeletal muscle cells for 16 h. MuSK phosphorylation was detected by ELISA (RayBiotech, MPA screen. MPA was performed by Integral Molecular. In brief, a flow catalog no. PEL-MUSK-Y-1) according to the manufacturer’s protocol. cytometry assay was used to assess the binding targets of MuSK extra- Optical density (OD ) values were normalized by total protein con- cellular domains (amino acids 24–495) linked with human Fc (MuSK-Fc 450/570 centration in each sample. Donor-matched NTD-T, MuSK-CAART and Chimera) among 5,300 human membrane proteins overexpressed in Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z 293T cells (permeabilized before MuSK-Fc incubation and flow cytom- research award from Cabaletta Bio (A.S.P.), the National Institute of etry detection). To validate potential off-targets of MuSK-Fc Chimera, Allergy and Infectious Diseases of the National Institutes of Health human embryonic kidney 293T cells were transfected with plasmids through grant awards to K.C.O. (award numbers R01-AI114780 and expressing the respective targets, vector alone (negative control), R21-AI142198) and a sponsored research subaward from the University PD-1-Fc isotype control or membrane-bound protein A construct and of Pennsylvania, the primary financial sponsor of which is Cabaletta anti-MuSK BCR (4A3 clone) (positive controls). After confirming tar- Bio. S.O. was supported by the Basic Science Research Program get protein expression, titration assays were conducted to validate a through the National Research Foundation of Korea, funded by the potential off-target of MuSK-Fc Chimera at different concentrations. Ministry of Education (2019R1A6A3A03033057). The content is solely the responsibility of the authors and does not necessarily represent Biodistribution assay. A GLP-compliant biodistribution study to the official views of the National Institutes of Health. evaluate the safety of MuSK-CAART in NSG mice was performed by Pharmaseed Ltd (Ness Ziona, Israel). On days −4 and −3, mice were Author contributions pretreated with i.v. administration of i.v. immunoglobulin (IVIg). Target S.O., C.T.E., S.L.K., D.P.R., K.C.O., U.H., G.K.B., M.C.M., S.B. and A.S.P. cells (1 × 10 Nalm-6 3-28 cells) were administered i.v. on day −2. After conceived the study. S.O., S.B. and A.S.P. designed the experiments. target cell engraftment, IVIg was administered i.p. every 2–3 days up S.O., X.M., S.M.-V., J.L., D.P., E.J.C., A.A., E.C.-T., D.M. and P.Y.T. performed to day 18. Two days after target cell administration, assigned day 1, the experiments. S.O., X.M., S.M.-V., J.L., D.P., S.B. and A.S.P. analyzed 6 7 the mice were injected i.v. with either vehicle (n = 4), 3 × 10 or 1 × 10 the data. S.O. and A.S.P. wrote the paper. All authors reviewed and/or 7 7 MuSK-CAART (n = 24 for each dose), 1 × 10 CART-19 (n = 24) or 1 × 10 revised the paper. NTD-T (n = 8) at a dose volume of 200 µl per mouse. Male and female mice were distributed equally. Effector cell (NTD-T, MuSK-CAART or Competing interests CART-19) administration was performed on day 1 and mouse harvest for S.O. is involved with patent licensing from Cabaletta Bio. J.L., D.P., pathologic evaluation, serum chemistry and complete blood count was A.A., E.C.-T., U.H., G.K.B. and S.B. are employed by Cabaletta Bio. C.T.E. performed on days 15, 36 and at study termination on day 61. Weights is involved with equity and patent licensing from Cabaletta Bio and and clinical observations were performed biweekly. Nalm-6 distribu- patent licensing from Novartis. S.L.K. is a consultant for Catalyst, tion was detected using bioluminescence imaging. Organs were fixed Alexion and Argenx. D.P.R. obtained a research grant from Cabaletta in formalin for hematoxylin and eosin staining and for pathologist Bio. K.C.O. is involved with equity and obtained a research grant from evaluation (Pharmaseed). Histopathological changes of MuSK-CAART Cabaletta Bio; is involved with research support, is a consultant and and control mice were described and scored using semiquantitative has received speaker fees from Alexion/AstraZeneca; has provided grading of five grades (0–4): grade 0, normal; grade 1, minimal; grade 2, research support and received speaking fees from Viela Bio/Horizon mild; grade 3, moderate and grade 4, severe. Therapeutics; and is a consultant for and has received speaking fees from Roche and received speaking fees from Genentech and UCB. Reporting summary M.C.M. is involved with equity and has received payment, a research Further information on research design is available in the Nature Port- grant and patent licensing from Cabaletta Bio, as well patent licensing folio Reporting Summary linked to this article. from Novartis and Tmunity, and is involved with equity and has received patent licensing from Verismo. A.S.P. is involved with equity, Data availability has received payment, research grant and patent licensing from The data that support the findings of this study are available from Cabaletta Bio and patent licensing from Novartis, and is a consultant the corresponding author upon reasonable request. Source data are to Janssen. X.M., S.M.-V., E.J.C., D.M. and P.Y.T. declare no competing provided with this paper. interests. References Additional information 37. Cugurra, A. et al. Skull and vertebral bone marrow are myeloid Extended data is available for this paper at https://doi.org/10.1038/ cell reservoirs for the meninges and CNS parenchyma. Science s41587-022-01637-z. 373, eabf7844 (2021). Supplementary information The online version contains Acknowledgements supplementary material available at https://doi.org/10.1038/s41587- We are grateful to A. Secreto, J. Glover, D. Dopkin, J. Frye, T. Hunter, 022-01637-z. E. Radaelli and C.A. Assenmacher for assistance with animal experiments and necropsy analysis. We thank S. de Munnik for Correspondence and requests for materials should be addressed to conduct and oversight of primary human cell studies conducted at Samik Basu or Aimee S. Payne. Charles River Laboratories, R. Fong for conduct and oversight of the Integral Molecular MPA screen, R. Nazan-Eraslan for technical insights Peer review information Nature Biotechnology thanks the anonymous and oversight of studies performed at Invivotek, as well as A. Nyska reviewers for their contribution to the peer review of this work. for expertise in histopathologic analysis and F.D. Arditti for technical insights and oversight of studies performed at Pharmaseed Ltd. Reprints and permissions information is available at Research reported in this publication was supported by a sponsored www.nature.com/reprints. Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 1 | Anti-MuSK target cell characterization. a) Schematic confirm the domain mapping. d, e) BCRs in primary human IgG B cells and diagram of individual MuSK domain CAARs with a cytoplasmic linker to green Nalm-6 cells expressing each MuSK domain-specific BCR were stained with PE fluorescent protein (GFP), generated for epitope mapping. FL, full-length; SP, mouse anti-human IgG. f) BCR density was calculated by dividing the number of 2 + signal peptide; TMD, transmembrane domain; BBζ, CD137(4-1BB)-CD3ζ. b) PE molecules/cell by the surface area (µm ). Mean ± standard deviation of IgG Summary of recombinant anti-MuSK B cell receptor (BCR) or mAb specificities. B-cells from three individual experiments is shown. c) MuSK domain CAAR Jurkat cells were stained with anti-MuSK mAbs to Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 2 | MuSK-CAAR directs specific cytolysis of anti-MuSK (anti-MuSK Fz) cell lines at indicated E:T ratios. Cytotoxicity was evaluated at B-cells. NTD-T cells, anti-CD19 CART (CART-19), and MuSK-CAART were co- 24 hours using a luciferase-based assay. Error bars indicate mean ± standard incubated with Nalm-6 control, Nalm-6 3−28 (anti-MuSK Ig2), and Nalm-6 4A3 deviation of triplicates. Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 3 | Relative titers of anti-MuSK mAbs and MG IgG. Anti- (b) were incubated with MuSK ectodomain-coupled microspheres and stained MuSK antibody titer was evaluated using a Luminex-based assay. Recombinant with anti-human IgG-biotin. Mean fluorescence intensity (MFI) was normalized anti-MuSK mAbs (a) or purified IgG from two MuSK MG patients (MG3 and MG5) per 50 beads. Error bars indicate mean ± SEM of duplicates. Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 4 | Evaluation of soluble anti-MuSK monoclonal antibody incubated with each anti-MuSK IgG4 mAb (0.2, 1, 5, or 25 µg /mL) and human effects on MuSK-CAART cytotoxicity, IFNγ production, and proliferation. a) IFNγ was quantitated by ELISA in cell culture supernatants after 24 hours. Error Non-transduced T cells (NTD-T) or MuSK-CAART were co-incubated with Nalm-6 bars indicate mean ± standard deviation of triplicates. c) Proliferation of NTD-T control or individual MuSK domain-specific Nalm-6 target cells at an E:T ratio of (top) and MuSK-CAART (bottom) was evaluated 96 hours after the addition of 10:1 in the absence (0 µg /mL) or presence of matching soluble anti-MuSK IgG4 the indicated anti-MuSK mAbs using Cell Trace Violet (CTV) cellular labeling dye mAb at 1.25 or 6.25 µg /mL, in a total of 10 mg /mL normal human IgG. Cytotoxicity dilution by flow cytometry. (a-c) Representative plots from two (b,c) or three (a) was evaluated at 24 hours using a luciferase-based assay. Error bars indicate individual experiments using different donor T-cells are shown. mean ± standard deviation of triplicates. b) NTD-T or MuSK-CAART cells were Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 5 | Generation and validation of 13-3B5* antibody- Nalm-6 13-3B5* (green) were stained with PE-conjugated mouse anti-human IgG secreting Nalm-6 cells. a) Nalm-6 13-3B5* anti-Ig1 antibody-secreting cells were to quantify BCR density. The mean fluorescence intensity of BCR expression is generated by transducing soluble 13-3B5/anti-Ig1 antibody heavy chain plasmid shown by histogram (b) and bar graph (c). d) Nalm-6 13-3B5 or Nalm-6 13-3B5* into Nalm-6 13-3B5 BCR−expressing cells. Jurkat cells expressing individual cells were co-incubated with either NTD-T or MuSK-CAART cells at indicated E:T MuSK extracellular domains linked with GFP were stained with cell-culture ratios for 8 hours. MuSK-CAART cytotoxicity was measured using a luciferase- supernatants from Nalm-6 13-3B5* for epitope mapping. Soluble 13-3B5 antibody based assay. Error bars indicate mean ± standard deviation of triplicates. binding was detected using anti-human IgG4-APC. b, c) Nalm-6 13-3B5 (red) and Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 6 | Evaluation of Nalm-6 outgrowth in a subset of MuSK- Nalm-6 cells (dotted line indicates cutoff for positive surface IgG expression). CAART-treated mice. a, b) Bioluminescence images from Fig. 3a, b. c) Enlarged g) Graphs indicate the percent of residual Nalm-6 cells that are IgG BCR + and bioluminescence image from (a, red box). d, e) T-cell and Nalm-6 cell percentage the MFI of IgG BCR expression in residual Nalm-6 cells in NTD-T (n = 2) and in the cranial bone marrow in NTD-T treated mice (n = 2) and MuSK-CAART MuSK-CAART (n = 3) treated mice. Error bars indicate mean ± SEM. Unpaired treated mice (n = 5) were analyzed by flow cytometry. f) Representative plot t-test (two-tailed): ns, p > 0.05; *, p < 0.05; **, p < 0.01. showing the mean fluorescence intensity (MFI) of IgG BCR expression in residual Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 7 | Immunologic characterization of the MuSK EAMG followed by MuSK full length (FL) boost, using Jurkat CAAR T-cells expressing syngeneic MuSK-CAART treatment model. a) Splenocytes were harvested individual MuSK domain CAARs linked to GFP. Representative FACS plots are from select MuSK-immunized mice, and purified B-cells were evaluated using shown from three independent experiments. d) Schematic of pMSGV1-1D3- a MuSK-specific and total IgG B-cell ELISpot assay to quantitate the MuSK- CAR (mouse CD8α signal peptide, 1D3 anti-mouse CD19 single chain variable specific B-cell frequency. Dots from duplicated wells were summarized after fragment (scFv), mouse CD28 hinge (HD)/transmembrane domain (TMD)/ normalization with seeded cells (dots per 100,000 cells). MuSK-specific B-cell costimulatory domain, and a mouse CD3ζ.1-3mut domain to confer enhanced frequency is calculated by dividing anti-MuSK B-cells by total IgG B-cells. b) IgG persistence. Schematic of pMSGV1-MuSK-CAAR (human MuSK extracellular subclasses of anti-MuSK antibodies were detected in sera from MuSK-immunized domains (amino acids 24-495), glycine-serine (GS) linker, hCD8α TMD, and mice in reference to sera from negative controls (non-immunized C57BL/6 and human CD137-CD3ζ). e, f) Transduction efficiency of 1D3-CAR and MuSK-CAAR in + + Rag2IL2rγ-deficient mice) and median value was plotted. c) Epitope mapping of primary mouse T-cells (Set 1 and Set 2) and g) percentage of CD4 /CD8 T-cells in serum from a mouse immunized and boosted with MuSK Ig1-Ig2, or MuSK Ig1-Ig2 Set 1 was evaluated on day 4 after T-cell activation (one day prior to injection). Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 8 | Off-target cytotoxicity of MuSK-CAART against n = 5; NTD-T+ agrin, n = 4; MuSK-CAART, n = 5; MuSK-CAART+ agrin, n = 5; Wise- + + MMP16 LRP4 U87-MG cells was not detected. a) LRP4 (blue) and MMP16 (red) CAART, n = 2). d) U87-MG cells were stained with Calcein-AM before co-culture expression in U87-MG cells was confirmed using flow cytometry. b) MuSK-CAAR with T-cells. Viable cells (GFP ) were detected by fluorescence microscopy and Flag-tagged Wise-CAAR expression were confirmed in primary human at 16 hours after co-culture with T-cells in the presence or absence of agrin. T-cells. c) Luciferase U87-MG cells were co-incubated with human T-cells at the Representative images are shown from two individual experiments. e) Human indicated E:T ratio in the presence (dashed line) or absence (solid line) of agrin. IFNγ production was detected by ELISA in 16-hour co-culture supernatants from Cytotoxicity was measured using a luciferase-based killing assay. Error bars two individual experiments. One-way ANOVA with Holm-Sidak test for multiple indicate mean ± standard deviation from two individual experiments (NTD-T, comparisons: ***, p < 0.001. Nature Biotechnology μ μ http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Biotechnology Springer Journals

Precision targeting of autoantigen-specific B cells in muscle-specific tyrosine kinase myasthenia gravis with chimeric autoantibody receptor T cells

Download PDF
26 pages

Loading next page...
 
/lp/springer-journals/precision-targeting-of-autoantigen-specific-b-cells-in-muscle-specific-GKum2t0cit

References (40)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2023. corrected publication 2024
ISSN
1087-0156
eISSN
1546-1696
DOI
10.1038/s41587-022-01637-z
Publisher site
See Article on Publisher Site

Abstract

nature biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Precision targeting of autoantigen-specific B cells in muscle-specific tyrosine kinase myasthenia gravis with chimeric autoantibody receptor T cells 1 1 1 2 Received: 22 March 2022 Sangwook Oh    , Xuming Mao    , Silvio Manfredo-Vieira    , Jinmin Lee , 2 1 2 2 Darshil Patel , Eun Jung Choi , Andrea Alvarado , Ebony Cottman-Thomas , Accepted: 8 December 2022 1 1 1 3 Damian Maseda    , Patricia Y. Tsao    , Christoph T. Ellebrecht , Sami L. Khella , 4 5 2 2 David P. Richman , Kevin C. O’Connor    , Uri Herzberg , Gwendolyn K. Binder , Published online: xx xx xxxx 6 2 1 Michael C. Milone , Samik Basu & Aimee S. Payne    Check for updates Muscle-specific tyrosine kinase myasthenia gravis (MuSK MG) is an autoimmune disease that causes life-threatening muscle weakness due to anti-MuSK autoantibodies that disrupt neuromuscular junction signaling. To avoid chronic immunosuppression from current therapies, we engineered T cells to express a MuSK chimeric autoantibody receptor with CD137-CD3ζ signaling domains (MuSK-CAART) for precision targeting of B cells expressing anti-MuSK autoantibodies. MuSK-CAART demonstrated similar efficacy as anti-CD19 chimeric antigen receptor T cells for depletion of anti-MuSK B cells and retained cytolytic activity in the presence of soluble anti-MuSK antibodies. In an experimental autoimmune MG mouse model, MuSK-CAART reduced anti-MuSK IgG without decreasing B cells or total IgG levels, reflecting MuSK-specific B cell depletion. Specific off-target interactions of MuSK-CAART were not identified in vivo, in primary human cell screens or by high-throughput human membrane proteome array. These data contributed to an investigational new drug application and phase 1 clinical study design for MuSK-CAART for the treatment of MuSK autoantibody-positive MG. Muscle-specific tyrosine kinase myasthenia gravis (MuSK MG) is a which leads to MuSK phosphorylation and formation of high-density chronic autoimmune disorder caused by MuSK autoantibodies that acetylcholine receptor (AChR) clusters that are essential for neuro- 1–3 6 result in potentially life-threatening muscle weakness . MuSK is a trans- muscular junction synaptic transmission . membrane receptor that interacts with lipoprotein receptor-related MuSK MG disease severity correlates with anti-MuSK antibody 4,5 7 protein 4 (LRP4) in complex with the neuronal proteoglycan agrin , titers , particularly IgG4 autoantibodies targeting the MuSK Ig1 1 2 Department of Dermatology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA. Cabaletta Bio, Philadelphia, PA, USA. 3 4 Department of Neurology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA. Department of Neurology, University of 5 6 California – Davis, Davis, CA, USA. Departments of Neurology and Immunobiology, Yale School of Medicine, New Haven, CT, USA. Department of Pathology and Laboratory Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA.  e-mail: [email protected]; [email protected] Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z 8–10 domain, which disrupt MuSK-LRP4 interactions . B cell depletion coincubation (Fig. 2a). MuSK-CAART and NTD-T produced similarly low with rituximab results in reduction of anti-MuSK IgG relative to total levels of IFNγ when cultured with Nalm-6 control and IgG from patients IgG , indicating that MuSK autoantibodies are primarily produced by with MuSK MG (Fig. 2b, left panel), whereas a significant increase in 12,13 short-lived plasma cells . Disease relapse after rituximab is attributed IFNγ production was observed in MuSK-CAART relative to NTD-T when to incomplete B cell depletion , requiring repeated rituximab infusions cocultured with anti-MuSK Nalm-6 cells (Fig. 2b, right panel, black for disease control, but chronic B cell depletion risks serious infections. versus red bars). IgG from patients with MG did not significantly affect Therapy should ideally eliminate only the pathogenic anti-MuSK B IFNγ levels, although a trend toward lower levels was observed (Fig. 2b, cells and spare healthy B cells to achieve durable remission of MuSK right panel, red bars). Coincubation of MuSK-CAART with monoclonal MG without generalized immunosuppression. Nalm-6 cells and matching soluble anti-MuSK mAb showed similar Here, we report the design and functional validation of a chimeric results, except 192-8/anti-Fz IgG4 potentiated MuSK-CAART cytotoxic- autoantibody receptor (CAAR) comprising the MuSK autoantigen, ity at higher concentrations (Extended Data Fig. 4a). tethered to tandem CD137-CD3ζ signaling domains. MuSK-CAAR To determine direct effects of anti-MuSK antibodies on expression in T cells directs cytotoxicity toward B cells expressing an MuSK-CAART, MuSK-CAART was incubated with anti-MuSK mAbs anti-MuSK surface autoantibody or B cell receptor (BCR). MuSK-CAAR (13-3B5(anti-Ig1)/3-28(anti-Ig2)/24C10(anti-Ig3)/192-8 (anti-Fz)). T cell (MuSK-CAART) technology is based on a clinically approved Anti-MuSK mAbs, mixed or individually, induced IFNγ production anti-CD19 chimeric antigen receptor T cell (CART-19) therapy that has and MuSK-CAART proliferation in a concentration-related manner 15,16 led to complete and durable remissions of B cell malignancies . We (Fig. 2c,d and Extended Data Fig. 4b,c). investigated MuSK-CAART efficacy and safety in preclinical models, To evaluate whether Fc-gamma receptors (FcγRs) or neonatal which support MuSK-CAART as a precision cellular immunotherapy Fc-receptor (FcRn) mediates indirect lysis by MuSK-CAART after with potential to induce complete and durable remission of MuSK MG. binding anti-MuSK antibodies, primary human monocytes (which express high-affinity FcγR/CD64, CD32 and FcRn) and natural killer Results (NK) cells (which express low-affinity FcγR/CD16) were cocultured MuSK-CAAR targets disease-relevant anti-MuSK B cell with MuSK-CAART and normal human IgG, mixed anti-MuSK IgG4 epitopes mAbs, purified plasma IgG from patients with MuSK MG or anti-CD3 MuSK is a transmembrane tyrosine kinase whose ectodomain comprises positive-control antibody (Fig. 2e,f ). After coincubation with three immunoglobulin-like (Ig1–Ig3) and frizzled-like (Fz) domains. An MuSK-CAART or NTD-T in the presence of anti-MuSK IgG4 mAbs or estimated 100, 58 and 23% of sera from patients with MuSK MG recog- purified plasma from IgG from patients with MG, caspase-3/7-positive nize Ig1, Ig2 and Ig3-Fz domains, respectively . To target anti-MuSK B monocytes or NK cells were similar to counts observed in the pres- cells, we designed a MuSK-CAAR comprising the complete MuSK ecto- ence of normal human IgG and less than those observed with anti-CD3 domain, linked to CD137-CD3ζ costimulatory and activation domains, positive-control antibody (UCHT1). and confirmed expression on primary human T cells (Fig. 1a,b). To validate MuSK-CAART cytotoxicity, we generated Nalm-6 B BCR and CD19-targeted lysis show similar in vivo efficacy cells expressing anti-MuSK domain-specific BCRs isolated from three We next evaluated whether anti-MuSK BCR-targeted cytolysis patients with MuSK MG or three MuSK-immunized mice. Epitope map- demonstrates comparable efficacy as CD19-targeted cytolysis in ping and previous literature confirmed MuSK domain specificity: eliminating anti-MuSK B cells in vivo, using a well-characterized nod 17 18 anti-MuSK Ig1 (13-3B5, ref. ), anti-MuSK Ig2 (189-1, ref. and 3-28, scid gamma (NSG) Nalm-6 xenograft model. NSG mice were engrafted 12,18 12,18 ref. ), anti-MuSK Ig3 (24C10) and anti-MuSK Fz (4A3, refs. and with mixed luciferase-expressing 13-3B5/3-28/24C10/192-8 or 4A3 192-8) (Extended Data Fig. 1a–c). Anti-MuSK BCR density on engineered Nalm-6 cells (anti-Ig1/Ig2/Ig3/Fz), followed by treatment with Nalm-6 cells was within twofold of IgG BCR density on primary human MuSK-CAART, NTD-T or CART-19. In parallel experiments to evaluate B cells (Extended Data Fig. 1d–f ). the potential neutralizing effect of anti-MuSK IgG on MuSK-CAART MuSK-CAART demonstrated specific lysis of Nalm-6 cells target- in vivo, NSG mice were engrafted with 13-3B5/anti-Ig1 or 13-3B5* ing each MuSK domain, but not control Nalm-6 cells (Fig. 1c–g and antibody-secreting Nalm-6 cells (described in Methods and Extended Extended Data Fig. 2), with increased specific cytotoxicity at 24 versus Data Fig. 5). 5 hours of coculture. Donor-matched nontransduced (NTD) T cells Bioluminescence imaging from all four experiments indicated (NTD-T) showed no cytotoxicity. IFNγ was detected in supernatants that CART-19 and MuSK-CAART significantly reduced Nalm-6 out- of MuSK-CAART cocultured with anti-MuSK Nalm-6 but not control growth relative to NTD-T (Fig. 3a,b), although recurrence of biolumines- Nalm-6 cells (Fig. 1h). cence flux signal was observed in a subset of MuSK-CAART-treated and CART-19-treated mice. Nalm-6 recurrence in CART-19-treated mice Anti-MuSK Abs have varying effects on MuSK-CAART activity engrafted with Nalm-6 13-3B5 cells was not due to loss of T cells, as MuSK autoantibodies might block MuSK-CAAR engagement with T cells were detectable in most CART-19-treated mice (Fig. 3c). Similarly, anti-MuSK BCRs, but they could also potentiate cytotoxicity by acti- Nalm-6 outgrowth in MuSK-CAART-treated mice engrafted with mixed vating MuSK-CAART. Anti-MuSK IgG concentrations in sera from Nalm-6 (192-8) cells was not due to failure of MuSK-CAART trafficking −1 patients with MuSK MG range from 0.5 to 49.5 nM (0.16–7.4 µg ml ), to cranial bone marrow (Extended Data Fig. 6c,d). In the subset of 11,19,20 assuming one antibody bound per MuSK molecule . To deter- mice (n = 3) with residual Nalm-6 target cells in cranial bone marrow, mine soluble anti-MuSK antibody effects on MuSK-CAART activity, the percentage of IgG BCR Nalm-6 cells was comparable between we performed cytotoxicity assays in the presence of a physiologic MuSK-CAART- and NTD-T-treated mice, but IgG BCR expression level −1 concentration (10 mg ml ) of polyclonal IgG from patients with MG, was significantly reduced in MuSK-CAART-treated mice (Extended or individual or mixed monoclonal antibody (mAb), at concentrations Data Fig. 6c–g). The low BCR density of Nalm-6 192-8, which is greater within or exceeding the expected range for autoantigen-specific IgG than two standard deviations lower than the mean density on pri- −1 + (0.2–25 µg ml ). Relative binding of MuSK IgG from patients MG3 mary human IgG B cells (Extended Data Fig. 1f ), may explain Nalm-6 and MG5 and anti-MuSK IgG4 mAbs appear in Extended Data Fig. 3. recurrence, since low target antigen density negatively affects CART MuSK-CAART cytotoxicity against mixed Nalm-6 target cells (13-3B5/3- cytotoxic efficacy . In MuSK-CAART-treated mice engrafted with 28/24C10/192-8 (anti-Ig1/Ig2/Ig3/Fz)) was partly inhibited by poly- mixed Nalm-6 (4A3) cells, all target cells were eliminated, comparable clonal IgG from patients with MuSK MG, although specific cytotoxicity to CART-19-treated mice, and no target cell recurrence was observed generally increased with higher effector to target ratios and longer (Fig. 3a, right). Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z a Gating strategy Extracellular domains TMD 0.023% **** Native MuSK Ig1 Ig2 Ig3 Fz Intracellular domain 72.3% MuSK CAAR CD137-CD3ζ Ig1 Ig2 Ig3 Fz FSC-A FSC-A Anti-MuSK-APC Linker CD8α TMD NTD-T MuSK-CAART c Nalm-6 control d Nalm-6 13-3B5 (anti-Ig1) e Nalm-6 189-1 (anti-Ig2) f Nalm-6 24C10 (anti-Ig3) 120 120 120 120 100 100 100 100 80 80 80 80 60 60 60 60 40 40 40 40 20 20 20 20 0 0 0 –20 –20 –20 –20 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 E:T ratio E:T ratio E:T ratio E:T ratio g h NTD-T Nalm-6 192-8 (anti-Fz) 5 hours MuSK-CAART CART-19 NTD-T 100 20 *** **** *** MuSK-CAART *** 80 **** **** **** CART-19 40 **** **** **** 24 hours –20 NTD-T 0 10 20 30 40 0 MuSK-CAART Nalm-6: Control 13-3B5 189-1 24C10 Control 192-8 E:T ratio CART-19 Fig. 1 | MuSK-CAAR expression on primary human T cells directs specific Nalm-6 cells (c) and Nalm-6 anti-MuSK target cells 13-3B5/anti-Ig1 (d), 189-1/anti- cytolysis of anti-MuSK B cells that target unique epitopes. a, Native MuSK Ig2 (e), 24C10/anti-Ig3 (f) and 192-8/anti-Fz (g) was measured using a luciferase- is a transmembrane tyrosine kinase whose ectodomain comprises three based cytotoxicity assay at 5 h (dashed line) and 24 h (solid line) after coculture immunoglobulin-like (Ig1–Ig3) and frizzled-like (Fz) domains. MuSK-CAAR with NTD-T (black), MuSK-CAART (red) and CART-19 (blue). The effector to target comprises the native MuSK ectodomain, followed by a glycine/serine-rich (E:T) ratio is based on total T cell number. h, Human IFNγ was measured in NTD-T, linker, CD8α transmembrane domain (TMD) and CD137-CD3ζ intracellular MuSK-CAART or CART-19 coculture supernatants (10:1 E:T, 24 h, experiments run costimulatory and activation domains. b, Primary human T cells were transduced using different T cell batches are shown in separate plots). One-way analysis of with MuSK-CAAR lentivirus or NTD-T, and MuSK-CAAR expression was detected variance (ANOVA) with the Holm–Sidak test for multiple comparisons. For c–h, using anti-MuSK 4A3 or 189-1 antibody. MuSK-CAAR transduction efficiency in error bars indicate mean ± s.d. of triplicate cocultures and are representative of six different donor T cell batches (NTD-T and MuSK-CAART). Error bars indicate 2–4 independent experiments. NS, P > 0.05; *P < 0.05; ***P < 0.001; ****P < 0.0001. mean ± s.d. Unpaired t-test (two-tailed). c–g, Cytolysis of wild-type (control) Higher MuSK-CAART percentages were observed in mixed MuSK-CAART efficacy in a syngeneic MuSK EAMG model Nalm-6 and 13-3B5*-engrafted mice (Fig. 3c), potentially due to solu- To determine whether MuSK-CAART demonstrates efficacy in an immu- ble mAb-induced expansion or proliferation of MuSK-CAART in vivo nocompetent mouse model that recapitulates features of autoim- after target cell encounter. MuSK-CAAR expression remained stable mune pathophysiology, including rare anti-MuSK B cells and polyclonal in T cells isolated from spleen (Fig. 3d) compared to MuSK-CAART autoantibodies, we evaluated MuSK-CAART efficacy in a syngeneic infusion product. However, anti-CD19 CAR expression was lower in MuSK experimental autoimmune myasthenia gravis (EAMG) model. spleen T cells (Fig. 3d) compared to CART-19 infusion product, which Classic MuSK EAMG models involve immunization of mice with rat 23,24 may explain Nalm-6 recurrence in CART-19-treated mice. MuSK ; use of human MuSK for EAMG induction is associated with In 13-3B5*-xenografted mice, anti-MuSK antibody titer increased variable symptom onset, although symptomatic mice demonstrate in mice treated with NTD-T from day 5 to 15, whereas titers in CART- reduction in miniature endplate potentials and postsynaptic AChR 25–27 19- and MuSK-CAART-treated mice were significantly reduced com- density, consistent with an MG phenotype . After day 0 immuniza- pared to NTD-T-treated mice by day 15, 11 days after T cell injection tion and day 26 boost of C57BL/6 (CD45.2 ) mice with human MuSK (Fig. 3e). Anti-MuSK IgG binding to diaphragm muscle cells in CART- ectodomain fragments, anti-MuSK IgG B cells comprised <0.2–1.5% of 19- and MuSK-CAART-treated mice was also reduced compared to total splenocyte IgG B cells gathered on day 34 (Extended Data Fig. 7a). NTD-T-treated mice (Fig. 3f ). Serum anti-MuSK antibodies were predominantly of the IgG1 and Taken together, these data indicate that MuSK-CAART can target IgG2c subclasses (Extended Data Fig. 7b). Anti-MuSK antibodies tar- MuSK Ig1/Ig2/Ig3/Fz domain-specific cells with comparable efficacy geted all four MuSK domains (Extended Data Fig. 7c). Immunized mice to CART-19. were treated on day 35 with murine CD45.1 T cells expressing human Nature Biotechnology Specific lysis (%) Specific lysis (%) MuSK-CAART NTD-T SSC-A –1 IFNγ (ng ml ) FSC-H FSC-H MuSK CAAR (%) MuSK mAb mixture MuSK mAb mixture Article https://doi.org/10.1038/s41587-022-01637-z 8 hours 24 hours NTD-T MuSK-CAART NTD-T MuSK-CAART Nalm-6 control 100 100 + medium + MG3 IgG 50 50 50 + MG5 IgG Nalm-6 mixed target 0 0 0 + medium + MG3 IgG + MG5 IgG –50 –50 –50 –50 (E:T ratio) 1:1 10:1 1:1 10:1 1:1 10:1 1:1 10:1 Nalm-6 control Nalm-6 mixed (anti-Ig1/Ig2/Ig3/Fz) (E:T ratio) 1:1 10:1 1:1 10:1 10 10 NS NTD-T *** NS NS 8 8 NS MuSK-CAART ** ** 6 6 ** ** ** 4 4 NS NS 2 2 0 0 MG3 IgG: – – + + – – – – + + – – – – + + – – – – + + – – MG5 IgG: – – – – + + – – – – + + – – – – + + – – – – + + c d NTD-T MuSK-CAART 3 Medium Unstained –1 0.2 μg ml Medium –1 –1 1 μg ml 0.2 μg ml –1 –1 1 μg ml 5 μg ml –1 5 μg ml –1 25 μg ml –1 25 μg ml CTV NTD-T MuSK-CAART e f Monocytes Monocytes NK cells NK cells 8 NTD-T 8 6 MuSK-CAART MuSK-CAART NTD-T 6 6 4 4 4 4 2 2 2 2 0 0 0 0 0 20 40 60 0 20 40 60 0 20 40 60 0 20 40 60 Time (h) Time (h) Time (h) Hours –1 –1 –1 –1 0.25 –g ml NH IgG 0.25 –g ml mixed mAbs 10.0 mg ml NH IgG 0.25 –g ml UCHT1 No antibody –1 –1 –1 –1 2.5 –g ml NH IgG 2.5 –g ml mixed mAbs 9.5 mg ml MG3 IgG 2.5 –g ml UCHT1 –1 –1 –1 –1 25 –g ml NH IgG 25 –g ml mixed mAbs 10.0 mg ml MG5 IgG 25 –g ml UCHT1 Fig. 2 | Evaluation of soluble anti-MuSK antibody effects on MuSK-CAART mAb concentration shown) and human IFNγ was quantitated by ELISA after cytotoxicity, IFNγ production and proliferation. a, NTD-T and MuSK- 24 h in duplicated samples (c), or proliferation of NTD-T and MuSK-CAART was CAART were coincubated with Nalm-6 control or mixed target cells (13-3B5/3- evaluated using CTV dye dilution by flow cytometry at 96 h (d). Representative 28/24C10/192-8 (anti-Ig1/Ig2/Ig3/Fz)) at 1:1 or 10:1 E:T ratios in the presence of flow plots from two individual experiments are shown. e,f, NTD-T or MuSK- −1 purified IgGs (10 mg ml ) from two patients with MuSK MG (MG3 and MG5; CAART were coincubated with monocytes (e) or NK cells (f) at a 5:1 E:T ratio in details in Methods) or medium alone. Cytotoxicity was evaluated at 8 and 24 h the presence of normal human IgG, mixed anti-MuSK mAbs (13-3B5/3-28/192-8 using a luciferase-based cytotoxicity assay. Error bars indicate mean ± s.d. of (anti-Ig1/Ig2/Fz), total mAb concentration shown), purified polyclonal IgG from triplicates. b, Human IFNγ was measured in NTD-T or MuSK-CAART coculture plasma from patients with MuSK MG (MG3 and MG5) or an anti-CD3 positive- supernatants (24 h) in two independent experiments. Two-way ANOVA with control mAb (clone UCHT1) for 48 h. Monocyte/NK cell death was detected by Tukey’s test for multiple comparisons: NS, P > 0.05; **P < 0.01; ***P < 0.001. incorporation of caspase-3/7 dye over time. Fold change of caspase cells relative c,d, NTD-T and MuSK-CAART were incubated with an equimolar mixture of to the 0 hour timepoint is plotted. Error bars indicate mean ± s.d. of triplicates. anti-MuSK IgG4 mAbs (13-3B5/3-28/24C10/192-8 (anti-Ig1/Ig2/Ig3/Fz), total MuSK-CAAR, a 1D3/anti-CD19 CAR with modified murine CD28-CD3ζ numbers of CAR/CAAR-transduced cells and matching numbers of signaling domains that confer greater in vivo efficacy as a positive con- NTD-T (Extended Data Fig. 7e). 1D3-CART and MuSK-CAART exhibited trol or NTD-T (Extended Data Fig. 7d). Mice were dosed with equivalent similar CD4:CD8 T cell ratios (Extended Data Fig. 7f ). Nature Biotechnology –1 –1 Caspase (fold change) Specific lysis (%) IFNγ (ng ml ) IFNγ (ng ml ) ND ND ND ND ND ND Caspase (fold change) Percentage of maximum **** **** ** **** NS **** * ** NS Article https://doi.org/10.1038/s41587-022-01637-z a b Nalm-6 mixed: anti-Ig1/Ig2/Ig3/Fz Nalm-6 mixed: anti-Ig1 13-3B5/3-28/24C10/4A3 13-3B5* (Ab-secreting) 13-3B5/3-28/24C10/192-8 13-3B5 11 11 10 10 10 10 10 10 9 9 10 10 8 8 10 10 7 7 10 10 6 6 10 10 0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25 Days after target cell injection Days after target cell injection Days after target cell injection Days after target cell injection c d IP Spleen IP Spleen IP Spleen IP Spleen 120 120 120 * * ** * NS NS NS NS 100 NS 100 100 NS NS **** NS ** NS NS *** *** 80 80 80 60 60 60 40 40 40 20 20 20 0 0 0 + 192-8 + 4A3 13-3B5 13-3B5* Nalm-6 mixed target Nalm-6 target Nalm-6 mixed target Nalm-6 target Nalm-6 mixed target Nalm-6 target e f **** NTD 10 **** NTD-T NS CART-19 ** CART-19 6 MuSK-CAART 0.2 MuSK-CAART 0.1 0 50 �m Neg control Day 5 Day 15 Fig. 3 | Targeting of anti-MuSK B cells through the BCR with MuSK-CAART cells relative to the infusion product (IP) (d). Error bars indicate mean ± s.e.m. demonstrates comparable efficacy as anti-CD19 CAR-mediated cytolysis, in One-sample t-test. Two MuSK-CAART-treated mice in Nalm-6 13-3B5/13-3B5* and −1 the presence or absence of soluble anti-MuSK antibody. a,b, Total flux (p s , two CART-19-treated mice in Nalm-6 13-3B5 experiments that were used for long- photons per second) after injection of 1:1:1:1 mixed 13-3B5/3-28/24C10/192-8 term follow-up were excluded from analysis. e, Anti-MuSK antibody titer in blood or 13-3B5/3-28/24C10/4A3 (anti-Ig1/Ig2/Ig3/Fz) (a), 13-3B5 (anti-Ig1) or 13-3B5* samples was measured on days 5 and 15 after target cell injection and quantitated (anti-Ig1, antibody-secreting) Nalm-6 cells (b), followed 4 days later by treatment relative to a 13-3B5 IgG4 mAb standard. Two-way ANOVA with Dunnett’s test for with 1 × 10 NTD-T (black, n = 5), CART-19 (blue, n = 5) or MuSK-CAART (red, n = 5). multiple comparisons. f, Direct immunofluorescence analysis of diaphragm Bioluminescence images appear in Extended Data Fig. 6a. One-way ANOVA with muscle harvested on day 25, stained with antihuman IgG to detect 13-3B5 IgG4 the Holm–Sidak test for multiple comparisons, day 23. c,d, Splenocytes were binding to MuSK on the muscle cell surface. Diaphragms from 13-3B5/NTD- analyzed on days 24 or 25 after target cell injection for CD3 T cell frequency T-treated mice served as a negative control. Scale bar, 50 µm. Representative (c) (median values are indicated; Kruskal–Wallis test with Dunnett’s test for images are shown from two independent staining experiments. NS, P > 0.05; + + multiple comparisons), and percentage of MuSK-CAAR and anti-CD19 CAR T *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. After treatment, few to no B cells were detected in the spleen in addition to the rarity of anti-MuSK B cells relative to CD19-expressing and lymph nodes of 1D3-CART-treated mice, whereas NTD-T- and B cells in immunized mice and/or the differing signaling and costimula- MuSK-CAART-treated mice showed comparable B cell frequency (Fig. 4b). tory domains of MuSK-CAAR and 1D3-CAR (Fig. 4f and Extended Data Anti-MuSK antibody titer was comparable across treatment groups Fig. 7a,d). before treatment (Fig. 4c) and progressively increased in NTD-T-treated These data indicate that in an EAMG model with rare anti-MuSK mice, whereas titers in MuSK-CAART- and 1D3-CART-treated mice sig- target B cells and circulating anti-MuSK antibodies, MuSK-CAART nificantly decreased, starting 1 week after injection (Fig. 4d). Unlike treatment results in antigen-specific IgG depletion without total 1D3-CART, which significantly decreased total serum IgG, MuSK-CAART B cell depletion and does not require previous lymphodepletion for did not reduce total serum IgG levels relative to NTD-T-treated mice therapeutic effect. (Fig. 4e). CD45.1 T cells were detected in spleen and lymph nodes with relatively lower percentages of engrafted T cells in MuSK-CAART-treated Specific off-target toxicity by MuSK-CAART was not observed mice, and higher T cell percentages in 1D3-CART-treated mice, the lat- To evaluate for potential off-target interactions of MuSK-CAART, ter in part due to the twofold higher number of T cells injected in set 1, we performed (1) comprehensive organ histopathology in an NSG Nature Biotechnology + 192-8 + 4A3 13-3B5 13-3B5* + 192-8 + 4A3 13-3B5 13-3B5* –1 –1 Total flux (p s ) T cell (%) Anti-MuSK Ab (�g ml ) MuSK CAAR (%) –1 Total flux (p s ) CD19 CAR (%) *** NS *** ** NS ** Article https://doi.org/10.1038/s41587-022-01637-z a b Gating strategy Spleen Lymph nodes NS 120 NS ** ** FSC-A FSC-A CD45.2-PE NTD-T 1D3-CART MuSK-CAART NTD-T 1D3 MuSK NTD-T 1D3 MuSK CART CAART CART CAART CD19-APC c NS d e Anti-MuSK Ab titer Total mouse IgG 2.5 NS NS Set 1 + Set 2 Set 1 + Set 2 2.0 2 8 1.5 1.0 –1 4 –2 0.5 –3 –4 2 0 1 2 3 4 5 0 1 2 3 4 5 Pretreatment Weeks after treatment Weeks after treatment Spleen Lymph nodes 2.5 Set 1 Set 2 Set 1 Set 2 2.0 Set 1 Set 2 NTD-T 1.5 1D3-CART 1.0 MuSK-CAART 0.5 −1 Fig. 4 | MuSK-CAART reduces anti-MuSK IgG but not total IgG or B cell counts (µg ml ). Kruskal–Wallis with Dunnett’s test for multiple comparisons. in a syngeneic MuSK EAMG model. CD45.2 C57BL/6J mice were immunized d,e, Anti-MuSK antibody titer and total mouse IgG were measured in mouse with MuSK Ig1-2 protein (30 µg in complete Freund’s adjuvant) on day 0 and blood samples drawn weekly after treatment. Graphs indicate fold change of anti- boosted with MuSK Ig1-Fz protein (30 µg in incomplete Freund’s adjuvant) on day MuSK antibody titer (d) or total mouse IgG (e) relative to week 1 after treatment 26. Results of two independent experiments are shown (total numbers NTD-T, (NTD-T, n = 8; 1D3-CART, n = 7 (d) and n = 6 (e); MuSK-CAART, n = 8 to include all n = 12, 1D3-CART, n = 8, MuSK-CAART, n = 12). Equivalent numbers of transduced mice with longitudinal samples through week 4 after treatment; one 1D3-CART + + CD45.1 T cells or a matching number of NTD CD45.1 cells were injected on day mouse was excluded in e due to low blood sample volume precluding analysis). + - + 35. a,b, Host B cells (CD45.2 CD3 CD19 ) were analyzed in spleen and lymph Error bars indicate mean ± s.e.m. Multiple linear regression-coefficient test for nodes at days 49–63 (2–4 weeks after treatment). Representative flow plot (a) difference between the slopes. f, Frequency of CD45.1 T cells were analyzed in + + and the frequency of CD45.2 CD19 B cells in the spleen and lymph nodes (b) are the spleen and lymph nodes on day 49–63. Statistical analysis was not performed shown. Kruskal–Wallis test with Dunnett’s test for multiple comparisons. c, Anti- since absolute number of CD45.1 T cells varied among treatment groups to MuSK antibody titer was measured in individual mouse blood samples drawn on achieve the same transduced cell dose in set 1. NS, P > 0.05; *P < 0.05; **P < 0.01; the day of treatment, normalized to 4A3 mouse antihuman MuSK mAb standard ***P < 0.001. 7 7 xenograft model, (2) screening of a high-throughput human mem- followed by treatment with vehicle, 1 × 10 NTD-T, 1 × 10 CART-19 or 6 7 brane proteome array (MPA), (3) primary human cell screens and (4) 3 × 10 –1 × 10 MuSK-CAART donor-matched cells. Representative targeted investigations based on potential MuSK-interacting proteins. bioluminescence images and graphs of total bioluminescence flux MuSK-CAART biodistribution was evaluated in 146 NSG mice allo- indicate control of Nalm-6 cell outgrowth in CART-19 and high-dose cated to 19 groups, injected with 1 × 10 3-28 Nalm-6 or no target cells, MuSK-CAART-treated mice, and delayed outgrowth in low-dose Nature Biotechnology CD45.1 (%) Normalized anti-MuSK Ab CD3-BV421 SSC-A FSC-H Fold change (relative to week 1) FSC-H + + CD45.2 CD19 (%) Fold change (relative to week 1) Article https://doi.org/10.1038/s41587-022-01637-z a MuSK-CAART b Vehicle Day 36, liver Day 36, lung Vehicle (n = 6) High-dose Low-dose CART-19 NTD-T 7 6 7 7 (10 cells) (3 × 10 cells) (10 cells) (10 cells) MuSK-CAART high-dose (n = 24) 3.0 MuSK-CAART low-dose (n = 24) CART-19 (n = 24) NTD-T (n = 8) ns 2.0 200 μm 200 μm 1.0 ** 7 10 ** (x10 ) 1 4 8 14 Days after target –1 2 –1 Radiance (p s cm sr ) cell injection 4 7 500 μm 1,000 μm Color scale: min = 4.17 × 10 , max = 3.04 × 10 c d Validation screen e Validation screen (positive controls removed) 9,000 8,000 Protein A MMP16 8,000 7,000 4A3 MMP16 7,000 MMP16 6,000 Vector Vector 6,000 5,000 5,000 4,000 4,000 3,000 3,000 2,000 2,000 1,000 1,000 20.0 5.0 1.3 0.3 20.0 5.0 1.3 0.3 20.0 5.0 1.3 0.3 –1 –1 –1 MuSK-Fc (μg ml ) MuSK-Fc (μg ml ) PD-1-Fc (μg ml ) Membrane proteome array f 800 g NTD-T Target only MuSK-CAART **** **** **** + Staurosporin + NTD-T + MuSK-CAART Nalm-6 control Nalm-6 3-28 (anti-Ig2) h i 160 200 Target only Target only ** ** *** ** 120 + Staurosporin 150 *** ** *** + Staurosporin or bortezomib + NTD-T 80 100 + MuSK-CAART + NTD-T 40 50 + MuSK-CAART HREC HUVEC NHEM Fig. 5 | Off-target cytotoxic interactions of MuSK-CAART were not identified curve overlaps with vector control). A representative graph from two validation in mouse tissue or using human MPAs. a, Representative bioluminescence screens is shown in e. e, Positive controls are removed in validation screens shown images from the MuSK-CAART biodistribution study in mice injected with 3-28/ in d and the y axis is rescaled. One of two validation screens confirmed MuSK-Fc anti-Ig2 Nalm-6 cells, then treated with vehicle only (n = 6), NTD-T (n = 8), CART- binding to MMP16, defined as MFI at least twofold higher than isotype (PD-1-Fc) 19 (n = 24) or MuSK-CAART (high- and low-dose, n = 24 per each dose). Graph control at two or more concentrations. f–i, Cytotoxicity of MuSK-CAART from indicates total bioluminescence flux for all mice in each treatment group. Error two donor T cell batches was measured after coincubation for 24 hours (5:1 E:T bars indicate mean ± s.e.m. One-way ANOVA with the Holm–Sidak test for multiple ratio) with Nalm-6 wild-type (negative control), Nalm-6 3-28/anti-Ig2 (positive comparisons, day 14. b, Example images from liver and lung on day 36 after control) or seven primary human cell types from each of two different donors. treatment (MuSK-CAART, n = 8; CART-19, n = 8). Liver and lung sections from high- Representative results are shown. IFNγ production was not detected in MuSK- dose MuSK-CAART-treated mice show lymphocytic infiltration without cytotoxic CAART cocultures with primary human cells (f). Viability was analyzed using effect (black arrows). CART-19-treated mice demonstrated focal hepatocellular high-content imaging analysis (HCA) for Nalm-6 control and anti-MuSK 3-28 necrosis and focal pulmonary thrombus (black arrows). c, Human MPA screened cells (g) and human-derived cells (h) or by flow cytometry (i) at 24 hours, using with MuSK-Fc protein identified a potential binding signal with MMP16. d, MuSK- staurosporin or bortezomib as toxicity controls. Error bars indicate mean ± s.d. of Fc binding to MMP16 demonstrated low mean fluorescence intensity (MFI) relative triplicates (f–i). Multiple t-test (two-tailed), Holm–Sidak correction for multiple to protein A and anti-MuSK 4A3 positive controls in validation screening (MMP16 comparisons. NS, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Nature Biotechnology T cell only Neurons Cardiomyocytes Hepatocytes HBEC HREC HUVEC NHEM Nalm-6 control Nalm-6 3-28 1,000 2,000 3,000 4,000 5,000 6,000 Neurons Cardiomyocytes Hepatocytes HBEC Live target cells Normalized target binding –1 Day 14 Day 8 Day 4 Day 1 IFNγ (pg ml ) (% of target only) MFI (×1,000) –1 Total fux (p s ) Live target cells Live target cells (% of target only) (% of target only) MFI (×1,000) CART-19 MuSK-CAART 7 7 (1 × 10 ) (1 × 10 ) Article https://doi.org/10.1038/s41587-022-01637-z a MuSK-CAART CART-19 –1 Agrin (ng ml ): 0 3 10 0 3 10 b c d Agrin –1 2.0 0 ng ml 5 × 10 120 –1 NTD-T 3 ng ml NS **** **** 1.8 –1 MuSK-CAART 10 ng ml **** 4 × 10 **** ***** **** Wise-CAART 1.6 **** **** 3 × 10 1.4 2 × 10 1.2 NS NS 1 × 10 1.0 0 0.8 0 MuSK-CAART CART-19 Medium Agrin Non Agrin −1 Fig. 6 | MuSK-CAART off-target effects on muscle are not observed. a, Effects agrin (5 ng ml ) or medium alone. Agrin-induced MuSK phosphorylation was of MuSK-CAART or CART-19 (E:T 10:1) on agrin-induced AChR clustering in C2C12 confirmed by phospho-MuSK ELISA. OD value relative to medium-alone 450/570 mouse myotubes were visualized by α-bungarotoxin staining (×25 magnification, control is shown. Mann–Whitney U-test (two-tailed). d, Human myotubes top row, plus bottom inset (red boxes); scale bar, 50 µm). Representative images were coincubated with NTD-T, MuSK-CAART or Wise-CAART in the presence are shown from two individual experiments. b, AChR clustering was quantitated or absence of agrin for 24 hours. IFNγ production was measured in cell-culture by fluorescence intensity, measured at six different sites in each image (error bars supernatants by ELISA. Error bars indicate mean ± s.d. of triplicates. One-way indicate mean ± s.e.m.). One-way ANOVA with the Holm–Sidak test for multiple ANOVA with the Holm–Sidak test for multiple comparisons. NS, P > 0.05; comparisons. c, Differentiated primary human myotubes were incubated with *P < 0.05; ****P < 0.0001. MuSK-CAART-treated mice relative to vehicle and NTD-T-treated mice second validation screen. MuSK-CAART cytotoxicity was subsequently (Fig. 5a). Comprehensive organ histopathologic analysis indicated lym- evaluated against U87-MG glioma cells, which express MMP16 as well phocytic infiltration in multiple organs, most notably at late timepoints as LRP4 (Extended Data Fig. 8a). As a positive control, we gener- after injection of MuSK-CAART or CART-19 in mice with target cells. ated a CAAR comprising Wise, which interacts with LRP4 in an 29,30 Liver and lung sections from high-dose MuSK-CAART-treated mice show agrin-independent manner , and verified Wise-CAART cytotox- lymphocytic infiltration without cytotoxic effect, which could repre- icity against U87-MG cells (Extended Data Fig. 8b,c). In contrast to sent target cells or engraftment of the human T cell product (Fig. 5b). Wise-CAART, coincubation of MuSK-CAART with U87-MG cells ± agrin CART-19-treated mice demonstrated focal hepatocellular necrosis did not induce cell lysis or IFNγ production (Extended Data Fig. 8c–e). in some liver sections, consistent with xenogeneic graft-versus-host Additionally, screens of seven primary human or induced pluripotent disease (xGVHD), and focal pulmonary thrombus consisting of histo- stem cell (iPSC)-derived cells representing skin, vascular tissue, heart, cytes and fibrosis, attributed to the intravenous (i.v.) route of injection brain/nerves, lung, liver and kidney, which were validated for MMP16 (Fig. 5b). Specific off-target cytotoxic effects of MuSK-CAART relative or LRP4 expression, did not identify specific cytolysis or IFNγ produc- to CART-19-treated mice were not observed. tion (Fig. 5f–i). Because in vivo biodistribution studies of human cellular To evaluate whether MuSK-CAART interferes with AChR clustering immunotherapies are confounded by xGVHD and may not identify or causes muscle cell cytolysis due to potential trans-interaction with off-target interactions against human proteins, recombinant MuSK-Fc LRP4, mouse C2C12 myotubes were cocultured with MuSK-CAART or protein was used to screen an MPA comprising approximately 5,300 CART-19, which is known not to cause muscle cytotoxicity in humans. human membrane proteins overexpressed in mammalian cells, which MuSK-CAART did not affect agrin-induced AChR clustering in C2C12 identified binding to matrix metalloproteinase-16 (MMP16) in ini- myotubes relative to CART-19 (Fig. 6a,b). Primary human muscle cells tial screening (Fig. 5c). Validation screens indicated overall low-level were also differentiated into myotubes, and MuSK activation by agrin MuSK-Fc binding to MMP16 relative to positive and PD-1-Fc isotype was verified by MuSK phosphotyrosine enzyme-linked immunosorbent controls (Fig. 5d,e), which was interpreted as negative (less than assay (ELISA) (Fig. 6c). IFNγ was not elevated in MuSK-CAART or NTD-T twofold higher than vector control) in one validation screen coculture supernatants, whereas a significant increase in human IFNγ and positive (more than twofold higher than vector control) in a was detected in Wise-CAART cocultures (Fig. 6d). Nature Biotechnology Fluorescence Intensity (pixels) Fold change of phospho-MuSK –1 IFNγ (ng ml ) Article https://doi.org/10.1038/s41587-022-01637-z Discussion To further evaluate MuSK-CAART in a preclinical autoimmune We report the development of a novel precision cellular immunother- model, we expressed human MuSK-CAAR in murine T cells and com- apy for autoantigen-specific B cell depletion in MuSK MG. MuSK-CAART pared its efficacy to anti-CD19 1D3-CART in an immunocompetent demonstrated specific cytotoxicity against B cells targeting each of syngeneic MuSK EAMG model. Similar to human MuSK MG, anti-MuSK the four MuSK domains (Fig. 1). By incorporating the complete MuSK B cells comprised <0.2–1.5% of splenic IgG B cells and targeted all ectodomain, MuSK-CAART is designed to eliminate autoimmune B four MuSK ectodomains (Extended Data Fig. 7). Unlike human MuSK cells targeting a broad range of MuSK epitopes. MG, both anti-MuSK IgG1 and IgG2c were induced in this model A major difference in the application of MuSK-CAART and CART-19 (Extended Data Fig. 7b). Murine IgG1 is functionally analogous to to clinical practice is the presence of soluble autoantibodies that could human IgG4 (ref. ). Murine IgG2c fixes complement, which might have varying effects on MuSK-CAART function. Human anti-MuSK anti- mediate CAART destruction or cause toxicities that may not occur bodies are predominantly IgG4, which are functionally monovalent and in patients with MuSK MG. Despite these limitations, MuSK-CAART do not fix complement or activate antibody-dependent cellular cyto- specifically reduced anti-MuSK IgG but not total IgG or total B cells, 17,31,32 toxicity . Data in Fig. 2a indicate that anti-MuSK antibodies mod- indicating antigen-specific B cell depletion without previous lym- estly inhibit MuSK-CAART cytotoxicity in vitro, although cytotoxicity phodepletion (Fig. 4). Anti-MuSK IgG reductions by MuSK-CAART increases with longer coincubation times and higher effector to target and 1D3-CART were comparable, despite lower percentages of CD45.1 ratios, while remaining specific. Cytotoxicity in the presence of soluble T cells in MuSK-CAART versus 1D3-CART-treated mice (Fig. 4f ). Dif- autoantibody was further confirmed in an EAMG model (discussed ferences in T cell engraftment were in part due to number of injected further below). Anti-MuSK IgG4 mAbs as well as IgG from patients T cells in a subset of mice, but may also reflect IgG2c-mediated clear- with MuSK MG activate MuSK-CAART to produce IFNγ and/or prolife- ance of MuSK-CAART or the lower abundance of anti-MuSK versus rate (Fig. 2 and Extended Data Fig. 4). Anti-MuSK antibodies did not CD19 B cells, which induces less expansion of MuSK-CAART relative mediate indirect lysis of Fc-receptor-expressing cells by MuSK-CAART to 1D3-CART. (Fig. 2e,f ). Collectively, these data suggest that soluble autoantibod- Multiple complementary approaches were used to screen for ies could be beneficial by amplifying the infused MuSK-CAART dose potential MuSK-CAART off-target effects (Figs. 5 and 6). Biodistribu- (Figs. 2d and 3b and Extended Data Fig. 4c) and providing a survival tion studies in mice may identify unexpected organ cytotoxicity due signal in vivo; however, autoantibodies could also induce cytokine to cross-reactivity with mouse proteins in their native context, but release syndrome (Fig. 2b,c and Extended Data Fig. 4b) and/or inhibit are limited by xGVHD and nonhomology with human proteins. MPAs cytotoxicity (Fig. 2a and Extended Data Fig. 4a). In preclinical studies of allow high-throughput screening of thousands of human cell-surface desmoglein 3 (DSG3)-CAART for the treatment of mucosal pemphigus proteins that may not otherwise be expressed in primary human cells vulgaris, similar induction of CAART proliferation and IFNγ production or mice, but may identify irrelevant targets due to artifacts of protein 21,33 by soluble autoantibodies occurred , although data from the first overexpression. One of two MPA screens identified MuSK-Fc inter- four cohorts in the ongoing DSG3-CAART clinical trial (NCT04422912) action with MMP16 (Fig. 5c–e), although follow-up screens against indicate no cytokine release syndrome or dose-limiting toxicities, as MMP16-expressing cells did not confirm MuSK-CAART cytotoxicity well as a dose-related increase in DSG3-CAART persistence throughout (Fig. 5f–i and Extended Data Fig. 8). These studies also indicated that the 28 days following infusion , suggesting that soluble autoantibod- MuSK, which physiologically interacts with LRP4 in cis, does not inter- ies do not mediate adverse events or prevent DSG3-CAART engraft- act with LRP4 in trans on muscle and other primary cell types. ment. Nevertheless, to mitigate risk, dose escalation is planned in the These data contributed to an investigational new drug application MuSK-CAART phase 1 clinical study design. for MuSK-CAART as a novel precision cellular immunotherapy for the To investigate the in vivo efficacy of MuSK-CAART, we used two treatment of MuSK autoantibody-positive MG and informed a phase complementary approaches. The NSG Nalm-6 xenograft model is a 1 clinical study design (NCT05451212). CAAR T cells may represent a well-defined model that allows (1) evaluation of the cytolytic efficacy platform technology that could be applied to numerous autoimmune and engraftment of the human clinical product, MuSK-CAART, in com- and alloimmune B cell-mediated conditions. parison to clinically approved CART-19; (2) inclusion of anti-MuSK Nalm-6 cells that bind a broad range of epitopes relevant to MuSK Online content MG and (3) sensitive real-time analysis of target cell burden by biolu- Any methods, additional references, Nature Portfolio reporting sum- minescence imaging. These studies indicate that MuSK-CAART dem- maries, source data, extended data, supplementary information, onstrates in vivo efficacy comparable to CART-19, including cytolysis acknowledgements, peer review information; details of author con- of 13-3B5/anti-Ig1 Nalm-6 cells in the presence of matching soluble tributions and competing interests; and statements of data and code autoantibody that could inhibit MuSK-CAART cytotoxicity (Fig. 3). availability are available at https://doi.org/10.1038/s41587-022-01637-z. However, nearly 100% of Nalm-6 cells are MuSK-reactive, whereas MuSK-reactive B cells in patients with MuSK MG are rare (reported References to be less than 0.15% of circulating IgG B cells ). Additionally, rare 1. Rodriguez Cruz, P. M., Cossins, J., Beeson, D. & Vincent, A. BCR-negative or BCR-low Nalm-6 cells, which persist despite enrich- The neuromuscular junction in health and disease: molecular ment for anti-MuSK BCR expression, proliferate and mediate delayed mechanisms governing synaptic formation and homeostasis. escape from MuSK-CAART in the NSG xenograft model (Extended Data Front. Mol. Neurosci. 13, 610964 (2020). Fig. 6). In humans, BCR-negative plasma cells would not be targeted by 2. Stathopoulos, P. et al. Mechanisms underlying B cell immune MuSK-CAART. However, MuSK autoantibody titers markedly decrease dysregulation and autoantibody production in MuSK myasthenia after therapy with rituximab, which also does not deplete plasma cells, gravis. Ann. N. Y. Acad. Sci. 1412, 154–165 (2018). suggesting that most anti-MuSK antibodies are produced by short-lived 3. Borges, L. S. & Richman, D. P. Muscle-specific kinase myasthenia plasma cells that are continuously replenished from the CD20 (and gravis. Front. Immunol. 11, 707 (2020). + 11–13 BCR ) memory B cell pool . Additionally, studies have shown that BCR 4. Kim, N. et al. Lrp4 is a receptor for Agrin and forms a complex with signaling and internalization are mutually exclusive and that anergic B MuSK. Cell 135, 334–342 (2008). 35,36 cells downregulate their BCR by continuous recycling , suggesting 5. Zhang, B. et al. LRP4 serves as a coreceptor of agrin. Neuron 60, that either direct cytolysis of anti-MuSK B cells or selective pressure by 285–297 (2008). MuSK-CAART to downregulate anti-MuSK BCRs may result in favorable 6. Zong, Y. et al. Structural basis of agrin-LRP4-MuSK signaling. therapeutic outcome. Genes Dev. 26, 247–258 (2012). Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z 7. Bartoccioni, E. et al. Anti-MuSK antibodies: correlation with 25. Viegas, S. et al. Passive and active immunization models of myasthenia gravis severity. Neurology 67, 505–507 (2006). MuSK-Ab positive myasthenia: electrophysiological evidence for 8. Huijbers, M. G. et al. Longitudinal epitope mapping in MuSK pre and postsynaptic defects. Exp. Neurol. 234, 506–512 (2012). myasthenia gravis: implications for disease severity. J. 26. Ulusoy, C. et al. Preferential production of IgG1, IL-4 and IL-10 in Neuroimmunol. 291, 82–88 (2016). MuSK-immunized mice. Clin. Immunol. 151, 155–163 (2014). 9. Huijbers, M. G. et al. MuSK IgG4 autoantibodies cause myasthenia 27. Ulusoy, C. et al. Dental follicle mesenchymal stem cell gravis by inhibiting binding between MuSK and Lrp4. Proc. Natl administration ameliorates muscle weakness in MuSK-immunized Acad. Sci. USA 110, 20783–20788 (2013). mice. J. Neuroinflammation 12, 231 (2015). 10. Koneczny, I., Cossins, J., Waters, P., Beeson, D. & Vincent, A. 28. Kochenderfer, J. N., Yu, Z., Frasheri, D., Restifo, N. P. & Rosenberg, MuSK myasthenia gravis IgG4 disrupts the interaction of LRP4 S. A. Adoptive transfer of syngeneic T cells transduced with a with MuSK but both IgG4 and IgG1-3 can disperse preformed chimeric antigen receptor that recognizes murine CD19 can agrin-independent AChR clusters. PLoS ONE 8, e80695 (2013). eradicate lymphoma and normal B cells. Blood 116, 3875–3886 11. Marino, M. et al. Long-lasting rituximab-induced reduction of (2010). specific-but not total-IgG4 in MuSK-positive myasthenia gravis. 29. Ohazama, A. et al. Lrp4 modulates extracellular integration of Front. Immunol. 11, 613 (2020). cell signaling pathways in development. PLoS ONE 3, e4092 12. Stathopoulos, P., Kumar, A., Nowak, R. J. & O’Connor, K. C. (2008). Autoantibody-producing plasmablasts after B cell depletion 30. Karner, C. M. et al. Lrp4 regulates initiation of ureteric budding identified in muscle-specific kinase myasthenia gravis. JCI Insight and is crucial for kidney formation–a mouse model for 2, e94263 (2017). Cenani-Lenz syndrome. PLoS ONE 5, e10418 (2010). 13. Zografou, C., Vakrakou, A. G. & Stathopoulos, P. Short- and 31. de Taeye, S. W. et al. FcgammaR binding and ADCC activity of long-lived autoantibody-secreting cells in autoimmune human IgG allotypes. Front. Immunol. 11, 740 (2020). neurological disorders. Front. Immunol. 12, 686466 (2021). 32. Lilienthal, G. M. et al. Potential of Murine IgG1 and Human IgG4 14. Jiang, R. et al. Single-cell repertoire tracing identifies to inhibit the classical complement and fcgamma receptor rituximab-resistant B cells during myasthenia gravis relapses. activation pathways. Front. Immunol. 9, 958 (2018). JCI Insight 5, e136471 (2020). 33. Ellebrecht, C. T. et al. Reengineering chimeric antigen receptor 15. Maude, S. L. et al. Tisagenlecleucel in children and young adults T cells for targeted therapy of autoimmune disease. Science 353, with B-cell lymphoblastic leukemia. N. Engl. J. Med. 378, 439–448 179–184 (2016). (2018). 34. Chang, D. J. et al. A Phase 1 trial of targeted DSG3-CAART cell 16. Schuster, S. J. et al. Tisagenlecleucel in adult relapsed or therapy in mucosal-dominant pemphigus vulgaris (mPV) patients: refractory diffuse large B-cell lymphoma. N. Engl. J. Med. 380, early cohort data. Mol. Ther. 30, 373 (2022). 45–56 (2019). 35. Blery, M., Tze, L., Miosge, L. A., Jun, J. E. & Goodnow, C. C. 17. Huijbers, M. G. et al. MuSK myasthenia gravis monoclonal Essential role of membrane cholesterol in accelerated BCR antibodies: valency dictates pathogenicity. Neurol. internalization and uncoupling from NF-kappa B in B cell clonal Neuroimmunol. Neuroinflamm. 6, e547 (2019). anergy. J. Exp. Med. 203, 1773–1783 (2006). 18. Takata, K. et al. Characterization of pathogenic monoclonal 36. Hou, P. et al. B cell antigen receptor signaling and internalization autoantibodies derived from muscle-specific kinase myasthenia are mutually exclusive events. PLoS Biol. 4, e200 (2006). gravis patients. JCI Insight 4, e127167 (2019). 19. McConville, J. et al. Detection and characterization of MuSK Publisher’s note Springer Nature remains neutral with regard to antibodies in seronegative myasthenia gravis. Ann. Neurol. 55, jurisdictional claims in published maps and institutional affiliations. 580–584 (2004). 20. Matthews, I. et al. Muscle-specific receptor tyrosine kinase Open Access This article is licensed under a Creative Commons autoantibodies–a new immunoprecipitation assay. Clin. Chim. Attribution 4.0 International License, which permits use, sharing, Acta 348, 95–99 (2004). adaptation, distribution and reproduction in any medium or format, 21. Lee, J. et al. Antigen-specific B cell depletion for precision therapy as long as you give appropriate credit to the original author(s) and the of mucosal pemphigus vulgaris. J. Clin. Invest. 130, 6317–6324 source, provide a link to the Creative Commons license, and indicate (2020). if changes were made. The images or other third party material in this 22. Ramakrishna, S. et al. Modulation of target antigen density article are included in the article’s Creative Commons license, unless improves CAR T-cell functionality and persistence. Clin. Cancer indicated otherwise in a credit line to the material. If material is not Res. 25, 5329–5341 (2019). included in the article’s Creative Commons license and your intended 23. Jha, S. et al. Myasthenia gravis induced in mice by immunization use is not permitted by statutory regulation or exceeds the permitted with the recombinant extracellular domain of rat muscle-specific use, you will need to obtain permission directly from the copyright kinase (MuSK). J. Neuroimmunol. 175, 107–117 (2006). holder. To view a copy of this license, visit http://creativecommons. 24. Mori, S. et al. Antibodies against muscle-specific kinase impair org/licenses/by/4.0/. both presynaptic and postsynaptic functions in a murine model of myasthenia gravis. Am. J. Pathol. 180, 798–810 (2012). © The Author(s) 2023 Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Methods Evaluation of MuSK monoclonal antibody and IgG titers from Design and construction of plasmids patients with MG Lentiviral plasmids . Lentiviral plasmid pTRPE (provided Relative titers of each recombinant anti-MuSK monoclonal anti- by the Penn Center for Cellular Immunotherapies) was modified body or purified plasma IgG from patients with MG (MG3 and MG5) to express MuSK-CAAR constructs and anti-MuSK BCRs as was evaluated using a Luminex-based assay. In brief, purified MG3 −1 −1 follows: IgG (0.85 mg ml ) and MG5 IgG (2.2 mg ml ) were diluted 1:10, 1:50 The human MuSK ectodomain (representing amino acids 24–495) and 1:100. Recombinant anti-MuSK monoclonal antibodies were −1 was synthesized (Integrated DNA Technologies) with flanking 5′ evaluated at 0.1, 0.2, 0.5 and 1.5 µg ml concentrations. Diluted BamHI and 3′ NheI restriction sites. Gene fragments were digested samples were added to MuSK ectodomain (aa 24–495)-coupled micro- and purified using a PCR purification kit (Qiagen), then ligated into spheres and incubated for 1 h at room temperature. After washing 33 −1 the pTRPE-DSG3-CAAR vector upstream of sequences encoding a samples, 5 µg ml antihuman IgG-Biotin was added and incubated glycine-serine linker, CD8α transmembrane, CD137 costimulatory and for 1 h at room temperature, followed by incubation with 100 µl of −1 CD3ζ signaling domains. Wise (amino acids 24–206, UniProt Q6X4U4) streptavidin-PE (4 µg ml ) for 45 min. After washing, beads were resus- was synthesized (Integrated DNA Technologies) and subcloned into pended in 100 µl of washing buffer and 60 µl was analyzed using a pTRPE vector. Luminex 100TM/200TM analyzer according to the manufacturer’s Anti-MuSK BCR 189-1 (ref. ) (also known as MuSK 1A, anti-Ig2) recommendations. and 13-3B5 (ref. ) (anti-Ig1) was produced by synthesizing (Inte- grated DNA Technologies) the variable heavy and variable light chain In vitro transduction and expansion of CAR/CAAR T cells sequences with flanking BamHI/NheI and XhoI/Bsu36I restriction In vitro transduction and expansion of human CAR/CAAR T cells. sites. Gene fragments were digested, purified (PCR purification Bulk (mixed CD4/CD8) primary human T cells (from Penn Human kit, Qiagen) and ligated into a pRRL4.IgG4 vector following previ- Immunology Core or leukapheresis (Stem Express)) were cultured 33 −1 ously published methods , then subcloned into lentiviral plasmid in human T cell culture media supplemented with 100 IU ml rhIL-2 pTRPE to generate pTRPE.IgG4.Lambda.189-1. Anti-MuSK BCRs 4A3 (Proleukin) (CTS OpTmizer media (Invitrogen, A1048501) plus 5% 12 12,18 (ref. ) (anti-Fz), 3-28 (refs. ) (anti-Ig2) and 192-8 (human anti-Fz IgM, human AB serum (Gemini Bio-Products, 100–512) or Roswell Park sequences provided by K.C.O.) were produced similarly, except that the Memorial Institute (RPMI) media supplemented with 10% FBS, 10 mM kappa variable region was synthesized with flanking XhoI/BsiWI HEPES, 1% penicillin/streptomycin and 1% GlutaMax). T cells were sites, and the kappa constant region was synthesized and cloned activated/selected with anti-CD3/CD28-coated paramagnetic beads into a pGEM-T Easy vector (Promega) before ligation into pTRPE to (Thermo Fisher Scientific, 40203D) at a 3/1 bead/cell ratio. Lentivirus generate pTRPE.IgG4.Kappa (4A3, 3-28 and 192-8). An anti-Ig3 mouse was added at 24 h after activation, and cells were expanded in either hybridoma (24C10) was produced by immunization of mice with the static culture or a Xuri Bioreactor (GE Healthcare Lifesciences) until day human MuSK ectodomain (Genscript) and the variable heavy and light 9 to day 10 after activation, with media changes approximately every chain genes sequenced (Genscript) and synthesized (Integrated DNA 2 days and magnetic bead removal before cryostorage. Expression of Technologies). Gene fragments were digested using BamHI/NheI (for anti-CD19 CAR or MuSK-CAAR on human T cells was detected using the variable heavy chain) and XhoI/BsiWI (for the variable light chain), CD19-Biotin with Streptavidin-PE, CD19-PE or recombinant anti-MuSK purified (PCR purification kit, Qiagen) and ligated into pTRPE.IgG4. antibodies (189-1, 24C10 or 4A3) with antihuman IgG4-APC or anti- Kappa vector. mouse IgG1-PE. Packaging plasmids pRSV-Rev and pGAG/POL, plus envelope plasmid Pcl VSVg (Nature Technology Corporation) were used with In vitro transduction and expansion of mouse CAR/CAAR T cells. Lipofectamine 2000 (Life Technologies) for lentiviral preparation in Mouse T cells (CD45.1 C57BL6/J, Jackson Laboratory, strain 002014) 293T cells or Lenti-X 293T cells (Takara, 632180). were purified using a CD3 T cell enrichment kit (R&D Systems) and cultured using mouse T cell culture media (RPMI-10 media Antibody plasmids. Variable heavy chain or variable light chain supplemented with 10% FBS, 10 mM HEPES, 1% penicillin/strepto- sequences of MuSK-specific antibodies were cloned into AbVec vec- mycin, and 1% GlutaMAX). T cells (10 cells per ml) were activated tors (IgG4 heavy chain, kappa light chain or lambda light chain). A 1:1 for 36 h with Dynabead Mouse T-Activator CD3/CD28 (Thermo −1 mixture of variable heavy chain and variable light chain plasmids was Fisher Scientific, 11452D) in media supplemented with 50 IU ml −1 cotransfected into 293T cells to produce recombinant monoclonal rhIL-2 and 10 ng ml rhIL-7 (BioLegend, 581902), plus fresh 20 µM antibody. mAbs were purified from 293T culture supernatants using 2-mercaptoethanol. Polystyrene nontreated plates were coated −1 protein A chromatography (Invitrogen) according to the manufac- with 30 µg ml RetroNectin (Takara) at 4 °C overnight then blocked turer’s recommendations. with mouse T cell culture media for 30 min at room temperature, followed by centrifugation with retroviral supernatant by centri- Retroviral plasmids. pMSGV1.1D3-28Z.1-3 mut was obtained from fugation at 3,000 g for 2 h at 4 °C. Transduction was conducted on Addgene (107227) . pMSGV1.MuSK-CAAR was generated by replacing days 1 and 2 after activation by adding T cells directly onto the 1D3-28Z.1-3 mut insert with the MuSK-CAAR sequence. Then 30 µg retrovirus-coated wells. Plates were centrifuged at 300 g for 10 min of each plasmid was transfected into the Plat-E (Cell Biolabs) packaging at room temperature, then placed in a cell-culture incubator over- cell line to produce retroviruses. night (first round transduction). Second round transduction was performed similarly, except T cells were incubated 4–6 h after sec- Patient samples ond round transduction in mouse T cell culture media supplemented −1 −1 Patient characteristics. MG3, a 57-year-old female, chronic active with 10 ng ml rhIL-7 and 10 ng ml rhIL-15 (BioLegend, 570302) for disease. MG5, a 34-year-old female, chronic active disease. Venipunc- an additional 3 days. Media was replaced the next day and as needed ture was performed under a protocol approved by the University of to maintain the cell concentration between 1 and 2 × 10 cells per Pennsylvania Institutional Review Board. ml. Expression of anti-CD19 CAR or MuSK-CAAR on mouse T cells was detected on day 4 after activation using antimouse IgG(H+L)-APC MuSK MG IgG purification. IgG was purified from plasma from ( Jackson ImmunoResearch) or APC-conjugated recombinant patients MG3 and MG5 by protein A chromatography (Invitrogen) anti-MuSK antibody (189-1). Mouse T cells were injected on day 5 after according to the manufacturer’s recommendations. activation. Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Pharmacologic and toxicologic effects of soluble anti-MuSK Nalm-6 3-28, Nalm-6 24C10 and Nalm-6 192-8 or Nalm-6 4A3 cells, antibodies (2) Nalm-13-3B5 and (3) Nalm-6 13-3B5*. Nalm-6 13-3B5* cells were MuSK-CAART cytotoxicity in the presence of MuSK MG IgG. generated by introducing 13-3B5 IgG4 heavy chain (without a mem- Donor-matched MuSK-CAART and NTD-T were incubated with brane anchor) into Nalm-6 13-3B5 expressing 13-3B5 IgG4 heavy chain BCR-negative Nalm-6 control cells or mixture of Nalm-6 MuSK target (membrane-bound form) and 13-3B5 light chain, resulting in 13-3B5 cells (1:1:1:1 ratio of Nalm-6 13-3B5, Nalm-6 3-28, Nalm-6 24C10 and IgG4 antibody-secreting cells that retained cell-surface 13-3B5 BCR Nalm-6 192-8) at E:T ratio of either 1:1 or 10:1 for 24 h. Purified IgG expression (Extended Data Fig. 5). Donor-matched frozen human from two patients with MuSK MG (MG3 and MG5) was added at a final T cells (NTD-T, CART-19 and MuSK-CAART) were thawed 1 day before −1 concentration of 10 mg ml IgG before coincubation. MuSK-CAART the treatment in human T cell culture media supplemented with 100 IU cytotoxicity was evaluated at 8 and 24 h using luciferase-based killing of rhIL-2. On day 4 after target cell injection, 10 human T cells were assay. Coculture supernatants were harvested after completing the injected via tail vein. Two different infusion products from the same final plate reading at 24 h and stored at −20 °C for IFNγ ELISA. donor were used in mixed (Fig. 3a) or 13-3B5/13-3B5* (Fig. 3b) Nalm-6 experiments. IVIg was injected every 2–3 days in mixed and 13-3B5 IFNγ production and proliferation of MuSK-CAART by soluble Nalm-6-engrafted mice. antibodies. Soluble anti-MuSK antibody-induced MuSK-CAART activa- tion and proliferation was evaluated by IFNγ ELISA or Cell Trace Violet Bioluminescence imaging. Bioluminescence was measured with a (CTV) cellular labeling, respectively. For IFNγ ELISA, donor-matched Xenogen IVIS Lumina S3 (Caliper Life Sciences) from day 1 after target MuSK-CAART and NTD-T were incubated with a mixture of recombinant cell injection and every 2–3 days thereafter by injecting d-Luciferin −1 anti-MuSK monoclonal antibodies (1:1:1:1 of 13-3B5, 3-28, 24C10 and potassium salt (Gold Bio) intraperitoneally at a dose of 150 mg kg . 192-8) for 24 h. For the CTV cellular labeling, T cells were labeled with Mice were anesthetized with 2% isoflurane and luminescence was CTV Cell Proliferation Kit (Invitrogen) and activated with a mixture serially measured at 1 min intervals for 7 min or until signals start of anti-MuSK monoclonal antibodies for 96 h before analysis by flow to decrease in an automatic exposure mode. Total flux in the peak cytometry. image was quantified using Living Image 4.4 (PerkinElmer) by draw- ing rectangles from head to 50% of the tail length. Radiance unit of −1 2 −1 IncuCyte assay. Monocytes and NK cells mixed with green caspase-3/7 p s  cm  sr  = number of photons per second per square centimeter dye were incubated in media containing relevant IgGs, as well as effec- that radiate into a solid angle of one steradian. tor cells. The normal human IgG amount (negative control) and UCHT1 (a positive control, anti-CD3 antibody) was matched to the equivalent Human anti-MuSK antibody ELISA. Serum samples in NSG mice total amount of mixed anti-MuSK monoclonal antibodies or puri- with Nalm-6 13-3B5* were collected at day 5 and day 15 after target fied plasma IgG from patients with MG. Monocytes and NK cells were cell injection in K EDTA tubes for MuSK antibody ELISA. To detect cocultured with MuSK-CAART or NTD-T at a 5:1 E:T ratio. Anti-MuSK human anti-MuSK IgG4, the histidine-tagged recombinant extracel- monoclonal antibodies 13-3B5, 3-28, 192-8 were mixed at 1:1:1 ratio. lular domain of human MuSK (aa 24–495, R&D Systems, catalog no. Cocultures were monitored for 48 h using an IncuCyte S3 system 10189-MK) was coated on ELISA plates in PBS overnight at 4 °C at a −1 (Sartorius) and images were taken every 2 h with a ×20 objective, then concentration of 5 µg ml . Plates were washed with washing buffer (Inv- dead cells (green positive) were counted at each timepoint from four itrogen, catalog no. 00-0400-59), and blocked with Pierce Protein-Free fields of images per each well. The number of caspase-positive cells is (PBS) Blocking Buffer (Thermo Scientific, catalog no. 37572). Mouse equal to the mean number of green cell counts in each imaging field. serum samples were evaluated at a dilution of 1:50 to 1:100 in com- parison to a 13-3B5 purified recombinant human monoclonal IgG4 Luciferase-based in vitro cytotoxicity assay antibody as a reference standard for quantitation. Antihuman IgG Luciferase-based killing assay. Click-beetle green luciferase express- (H+L) HRP (Bethyl, catalog no. A80-119P) was used to detect human ing cells or luciferase U87-MG cells (ATCC, HTB14Luc2) were cocul- antibodies. Plates were protected from the light and placed in the tured with engineered T cells or donor-matched NTD-T at an indicated dark for 2 h. After washing plates three times, 100 µl of TMB (Thermo effector:target (E:T) ratio. At 3 h after coculture, luciferase substrate Scientific, 34028) was added for 30 min. Plate reading were conducted (d-luciferin potassium salt, GoldBio) was directly added to each well and using ELISA reader (Tecan, Infinite F-50) within 15 min after adding stop emitted light was measured on a luminescence plate reader (BioTek, solution (Invitrogen, SS04). Synergy HTX microplate reader) at indicated timepoints. The percent- age of specific lysis was calculated using the luciferase activity of 5% Flow cytometry analysis. Lymphocytes were isolated from cranial SDS-treated cells as maximum cell death and media alone as spontane- bone marrow using a previously reported protocol . In brief, the cal- ous cell death using the formula: specific lysis (%) = 100 × ((experimen- varia was cut into small pieces using sterile scissors and dissociated in tal data − maximum death data)/(maximum death data − spontaneous PBS + 2% FBS with a pestle. Spleens were harvested from mice, washed death data)). in PBS and cut into 0.5 mm cubes in ice-cold PBS. Spleen or bone mar- row isolates were transferred to a 70 µm cell strainer (Falcon, catalog Ethics statement for animal research no. 352350); cells were washed with PBS and resuspended in red blood All studies involving animals were performed under a protocol cell lysis buffer (BioLegend, catalog no. 420301). Cells were stained approved by the University of Pennsylvania Institutional Animal Care for 30 min on ice using the following antibodies: anti-CD3 BV711 or and Use Committee. anti-CD3-AF647 (clone okt3, BD Biosciences, catalog nos. 750983 and 566686)), anti-MuSK PE (clone 189-1 or 24C10), antihuman Ig light chain In vivo MuSK-CAART evaluation using NSG Nalm-6 xenograft λ PE (clone MHL-38, BioLegend, catalog no. 316608), antihuman Ig light models chain κ APC (clone MHK-49, BioLegend, catalog no. 316510), antihuman scid tm1Wjl Target cell and T cell injection. NSG (NOD.Cg-Prkdc IL2rg / IgG PE (BD Biosciences, catalog no. 555787) and/or antimouse IgG APC −1 SzJ) mice received 600 mg kg i.v. immunoglobulin (Privigen, IVIg) (clone A85-1, BD Biosciences). via tail-veil injection on day −2 and day −1 before target cell injection to prevent Fc-mediated Nalm-6 clearance. On day 0, three cohorts of Direct immunofluorescence. The diaphragm was collected at the NSG mice each received 10 Nalm-6 target cell line(s) via tail-vein injec- time of tissue harvest and embedded in optimal cutting tempera- tion as follows: (1) Nalm-6 mixed (a 1:1:1:1 mixture of Nalm-6 13-3B5, ture medium (Tissue-Tek); tissue blocks were frozen on dry ice before Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z storage at −80 °C. Next, 9 µm sections were cut onto Superfrost/Plus Wise-CAART were added and cocultured for 22–26 h. For some experi- glass slides (Fisher Scientific) and stored at −80 °C. Before staining, ments, neural agrin was added at least 30 min before T cell coculture. slides were equilibrated to room temperature, washed twice in PBS Evaluation of AChR clustering in C2C12 myotubes was performed (Gibco) and blocked in PBS containing 2% BSA. Slides were stained using by Invivotek, LLC. C2C12 mouse myoblast cells (ATCC 70024392) were FITC-conjugated antihuman IgG (BioLegend) at a dilution of 1:200 in maintained in DMEM supplemented with 20% FBS and 1% penicillin/ blocking solution and washed with PBS. Binding of IgG was visualized streptomycin. To induce myotube differentiation, C2C12 cells (80–90% with a Keyence imaging system (BZ-X710) and software (BZ-X Viewer). confluent) were cultured in DMEM plus 2% horse serum and 0.5% peni- cillin/streptomycin. Agrin-induced AChR clustering was examined by Syngeneic MuSK EAMG model incubating C2C12 myotubes in 1:1 DMEM differentiation media and Immunization and boosting schedule. CD45.2 C57BL/6J mice RESGRO serum-free culture medium (EMD Millipore, SCM002) sup- ( Jackson Laboratory, strain 000664) were immunized with MuSK plemented with varying concentrations of agrin (R&D Systems, catalog Ig1–Ig2 ectodomain fragment on day 0 and boosted with MuSK Ig1-Fz no. 550-AG/CF) for 14 h at 37 °C. AChR staining was performed using −1 full-length ectodomain protein on day 26. In initial experiments, 1 µg ml AlexaFluor 488-labeled α-bungarotoxin (Invitrogen catalog −1 20 mg kg busulfan was injected on day 34 in a subset of mice to evalu- no. B13422). Myotubes were fixed with 4% paraformaldehyde for 20 min ate effects on induced antibody titer and MuSK-CAART activity and at room temperature and imaged under mounting medium (Vector 6 6 engraftment. On day 35, 8 × 10 MuSK-CAART, 16 × 10 anti-CD19-CART Laboratories) using a Leica TCS SP8 multiphoton confocal micro- 6 6 (1D3) or 8 × 10 NTD-T in set 1 or 20 × 10 cells each in set 2 were admin- scope. Fluorescence was quantified (Fiji-Image J) as corrected total istered via i.v. injection. Analyses in Fig. 4a,b,f were performed 2 or cell fluorescence = integrated density − (area of selected cell × mean 4 weeks after T cell treatment (set 2 or 1, respectively). fluorescence of background). ELISA for total and subclass-specific mouse anti-MuSK IgG Off-target toxicity against primary human cells. MuSK-CAART reac- and total mouse IgG. Blood samples were collected weekly. To tivity against two different donor batches of primary or iPSC-derived detect mouse anti-MuSK antibodies in the syngeneic MuSK EAMG human cells was performed by Charles River Discovery Research model, mouse plasma (diluted 1:100 in PBS) was incubated on MuSK Services, including iPSC-derived iCell cardiomyocytes (Fuji Cellu- protein-coated plates and subsequently detected with antimouse lar Dynamics no. R1007/R1106), iPSC-derived iCell GABANeurons IgG-HRP (diluted 1:5,000, abcam, ab7061). Mouse anti-MuSK mono- (Fuji Cellular Dynamics, no. R1013/R1011), human bronchial epithelial clonal antibody (clone 4A3) was used as a reference standard control cells (HBEC, Lonza, no. CC-2540), primary hepatocytes (InnoProt, no. across experiments. Goat antimouse IgG1, IgG2b, IgG2c or IgG3-HRP P10651), human renal epithelial cells (HREC, InnoProt, no. P10664), (SouthernBiotech) was used as a secondary antibody reagent to deter- normal human epidermal melanocytes (NHEM, Lonza no. CC-2586 and mine anti-MuSK IgG subclasses. Total mouse IgG was measured by LifeLine no. FC-0030) and human umbilical vascular endothelial cells ELISA following manufacturer’s protocols (Invitrogen, CAT), after (HUVEC, Lonza, no. CC-2517). BCR Nalm-6 cells and Nalm-6 3-28 cells diluting sera 1:10,000 in PBS. were used as a negative control and a positive control, respectively. Primary or iPSC-derived human cells were cocultured for 24 h with two ELISpot. The frequency of anti-MuSK B cells in the spleen of sets of donor-matched NTD-T and MuSK-CAART at an E:T ratio of 5:1. MuSK-immunized mice were conducted using Mouse IgG ELISpot- MuSK-CAART cytotoxicity was detected at 24 h using either an basic kit (Mabtech, 3825-2H) according to the manufacturer’s proto- HCA with Nalm-6 cells, cardiomyocytes, hepatocytes and HBEC or flow col. Briefly, splenocytes were prestimulated with a mixture of R848 cytometry analysis with Nalm-6 cells, HREC, HUVEC and NHEM, respec- −1 −1 (1 µg ml ) and rmIL-2 (10 ng ml ) for 48 h. After prestimulation, cells tively. 1 µM staurosporin or 10 µM bortezomib (for HREC) was used were washed and resuspended in medium, then 10,000 or 100,000 B as a toxin control. For HCA, CAR/CAAR T cells were stained with Cell- −1 cells were plated in ELISpot wells precoated either with anti-IgG anti- Tracker Deep Red dye for 24 h, followed by incubation with 10 µg ml −1 −1 bodies (total IgG B cells) or MuSK protein (10 µg ml ), respectively. Hoechst and 4 µg ml propidium iodide in PBS supplemented with 0.5% BSA for 10 min at room temperature. Cells were imaged using the Flow cytometry analysis. Cells were isolated from the spleen and GE Healthcare IN Cell Analyzer 6000 (×10 magnification). Brightfield, lymph nodes after red blood cell lysis and stained with anti-CD45.1-FITC ultraviolet, dsRed and Cy5 channels were used to image brightfield, (BioLegend, 110706), anti-CD45.2-PECy7 (BioLegend, 109830), Hoechst staining, propidium iodide staining and CEllTracker staining, anti-CD3ε-BV421 (BioLegend, 100227) and anti-CD19-APC (BioLegend, respectively. Cell type-specific HCA was used to quantify live target 115512) for 30 min on ice. cells (IN Cell Developer Toolbox software (v.1.9.1)). Nuclear area of nonviable target cells was determined based on propidium iodide Evaluation for off-target interactions of MuSK-CAART staining and subtracted from total target cell nuclear area. For flow Calcein-AM staining. U87-MG cells, cultured to 70–90% confluency cytometry analysis, target cells were stained with 1 µM CellTracker −1 in a 12-well plate, were stained with 0.1–1 µM of Calcein-AM following Deep Red dye then incubated with 0.5 µg ml propidium iodide in the the manufacturer’s protocol (BD Pharmingen), then coincubated for presence of precision count beads before flow cytometry (Agilent 20–24 h with 1 × 10 MuSK-CAART, Wise-CAART or NTD-T. Then 5 nM NovoCyte Quanteon, FlowJo software v.10). Absolute count of live of neuronal agrin was added to U87-MG culture supernatants at least target cells was normalized to absolute bead count. 30 min before coculture. Measurement of IFNγ. Secreted IFNγ in coculture supernatants was Off-target toxicity against differentiated muscle cells. Primary detected by ELISA (BioLegend) or with a Luminex bead array platform human skeletal muscle cells (ZenBio, SKB-F) were differentiated for (Thermo Fisher) according to the manufacturer’s instructions. All 6–7 days using skeletal muscle cell differentiation medium (ZenBio, samples were analyzed in triplicate and compared against multiple SKM-D). Next, 5 nM of neural agrin (R&D Systems, catalog no. 550-AG/ internal standards, with a seven-point standard curve. CF) was added into differentiated primary human skeletal muscle cells for 16 h. MuSK phosphorylation was detected by ELISA (RayBiotech, MPA screen. MPA was performed by Integral Molecular. In brief, a flow catalog no. PEL-MUSK-Y-1) according to the manufacturer’s protocol. cytometry assay was used to assess the binding targets of MuSK extra- Optical density (OD ) values were normalized by total protein con- cellular domains (amino acids 24–495) linked with human Fc (MuSK-Fc 450/570 centration in each sample. Donor-matched NTD-T, MuSK-CAART and Chimera) among 5,300 human membrane proteins overexpressed in Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z 293T cells (permeabilized before MuSK-Fc incubation and flow cytom- research award from Cabaletta Bio (A.S.P.), the National Institute of etry detection). To validate potential off-targets of MuSK-Fc Chimera, Allergy and Infectious Diseases of the National Institutes of Health human embryonic kidney 293T cells were transfected with plasmids through grant awards to K.C.O. (award numbers R01-AI114780 and expressing the respective targets, vector alone (negative control), R21-AI142198) and a sponsored research subaward from the University PD-1-Fc isotype control or membrane-bound protein A construct and of Pennsylvania, the primary financial sponsor of which is Cabaletta anti-MuSK BCR (4A3 clone) (positive controls). After confirming tar- Bio. S.O. was supported by the Basic Science Research Program get protein expression, titration assays were conducted to validate a through the National Research Foundation of Korea, funded by the potential off-target of MuSK-Fc Chimera at different concentrations. Ministry of Education (2019R1A6A3A03033057). The content is solely the responsibility of the authors and does not necessarily represent Biodistribution assay. A GLP-compliant biodistribution study to the official views of the National Institutes of Health. evaluate the safety of MuSK-CAART in NSG mice was performed by Pharmaseed Ltd (Ness Ziona, Israel). On days −4 and −3, mice were Author contributions pretreated with i.v. administration of i.v. immunoglobulin (IVIg). Target S.O., C.T.E., S.L.K., D.P.R., K.C.O., U.H., G.K.B., M.C.M., S.B. and A.S.P. cells (1 × 10 Nalm-6 3-28 cells) were administered i.v. on day −2. After conceived the study. S.O., S.B. and A.S.P. designed the experiments. target cell engraftment, IVIg was administered i.p. every 2–3 days up S.O., X.M., S.M.-V., J.L., D.P., E.J.C., A.A., E.C.-T., D.M. and P.Y.T. performed to day 18. Two days after target cell administration, assigned day 1, the experiments. S.O., X.M., S.M.-V., J.L., D.P., S.B. and A.S.P. analyzed 6 7 the mice were injected i.v. with either vehicle (n = 4), 3 × 10 or 1 × 10 the data. S.O. and A.S.P. wrote the paper. All authors reviewed and/or 7 7 MuSK-CAART (n = 24 for each dose), 1 × 10 CART-19 (n = 24) or 1 × 10 revised the paper. NTD-T (n = 8) at a dose volume of 200 µl per mouse. Male and female mice were distributed equally. Effector cell (NTD-T, MuSK-CAART or Competing interests CART-19) administration was performed on day 1 and mouse harvest for S.O. is involved with patent licensing from Cabaletta Bio. J.L., D.P., pathologic evaluation, serum chemistry and complete blood count was A.A., E.C.-T., U.H., G.K.B. and S.B. are employed by Cabaletta Bio. C.T.E. performed on days 15, 36 and at study termination on day 61. Weights is involved with equity and patent licensing from Cabaletta Bio and and clinical observations were performed biweekly. Nalm-6 distribu- patent licensing from Novartis. S.L.K. is a consultant for Catalyst, tion was detected using bioluminescence imaging. Organs were fixed Alexion and Argenx. D.P.R. obtained a research grant from Cabaletta in formalin for hematoxylin and eosin staining and for pathologist Bio. K.C.O. is involved with equity and obtained a research grant from evaluation (Pharmaseed). Histopathological changes of MuSK-CAART Cabaletta Bio; is involved with research support, is a consultant and and control mice were described and scored using semiquantitative has received speaker fees from Alexion/AstraZeneca; has provided grading of five grades (0–4): grade 0, normal; grade 1, minimal; grade 2, research support and received speaking fees from Viela Bio/Horizon mild; grade 3, moderate and grade 4, severe. Therapeutics; and is a consultant for and has received speaking fees from Roche and received speaking fees from Genentech and UCB. Reporting summary M.C.M. is involved with equity and has received payment, a research Further information on research design is available in the Nature Port- grant and patent licensing from Cabaletta Bio, as well patent licensing folio Reporting Summary linked to this article. from Novartis and Tmunity, and is involved with equity and has received patent licensing from Verismo. A.S.P. is involved with equity, Data availability has received payment, research grant and patent licensing from The data that support the findings of this study are available from Cabaletta Bio and patent licensing from Novartis, and is a consultant the corresponding author upon reasonable request. Source data are to Janssen. X.M., S.M.-V., E.J.C., D.M. and P.Y.T. declare no competing provided with this paper. interests. References Additional information 37. Cugurra, A. et al. Skull and vertebral bone marrow are myeloid Extended data is available for this paper at https://doi.org/10.1038/ cell reservoirs for the meninges and CNS parenchyma. Science s41587-022-01637-z. 373, eabf7844 (2021). Supplementary information The online version contains Acknowledgements supplementary material available at https://doi.org/10.1038/s41587- We are grateful to A. Secreto, J. Glover, D. Dopkin, J. Frye, T. Hunter, 022-01637-z. E. Radaelli and C.A. Assenmacher for assistance with animal experiments and necropsy analysis. We thank S. de Munnik for Correspondence and requests for materials should be addressed to conduct and oversight of primary human cell studies conducted at Samik Basu or Aimee S. Payne. Charles River Laboratories, R. Fong for conduct and oversight of the Integral Molecular MPA screen, R. Nazan-Eraslan for technical insights Peer review information Nature Biotechnology thanks the anonymous and oversight of studies performed at Invivotek, as well as A. Nyska reviewers for their contribution to the peer review of this work. for expertise in histopathologic analysis and F.D. Arditti for technical insights and oversight of studies performed at Pharmaseed Ltd. Reprints and permissions information is available at Research reported in this publication was supported by a sponsored www.nature.com/reprints. Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 1 | Anti-MuSK target cell characterization. a) Schematic confirm the domain mapping. d, e) BCRs in primary human IgG B cells and diagram of individual MuSK domain CAARs with a cytoplasmic linker to green Nalm-6 cells expressing each MuSK domain-specific BCR were stained with PE fluorescent protein (GFP), generated for epitope mapping. FL, full-length; SP, mouse anti-human IgG. f) BCR density was calculated by dividing the number of 2 + signal peptide; TMD, transmembrane domain; BBζ, CD137(4-1BB)-CD3ζ. b) PE molecules/cell by the surface area (µm ). Mean ± standard deviation of IgG Summary of recombinant anti-MuSK B cell receptor (BCR) or mAb specificities. B-cells from three individual experiments is shown. c) MuSK domain CAAR Jurkat cells were stained with anti-MuSK mAbs to Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 2 | MuSK-CAAR directs specific cytolysis of anti-MuSK (anti-MuSK Fz) cell lines at indicated E:T ratios. Cytotoxicity was evaluated at B-cells. NTD-T cells, anti-CD19 CART (CART-19), and MuSK-CAART were co- 24 hours using a luciferase-based assay. Error bars indicate mean ± standard incubated with Nalm-6 control, Nalm-6 3−28 (anti-MuSK Ig2), and Nalm-6 4A3 deviation of triplicates. Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 3 | Relative titers of anti-MuSK mAbs and MG IgG. Anti- (b) were incubated with MuSK ectodomain-coupled microspheres and stained MuSK antibody titer was evaluated using a Luminex-based assay. Recombinant with anti-human IgG-biotin. Mean fluorescence intensity (MFI) was normalized anti-MuSK mAbs (a) or purified IgG from two MuSK MG patients (MG3 and MG5) per 50 beads. Error bars indicate mean ± SEM of duplicates. Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 4 | Evaluation of soluble anti-MuSK monoclonal antibody incubated with each anti-MuSK IgG4 mAb (0.2, 1, 5, or 25 µg /mL) and human effects on MuSK-CAART cytotoxicity, IFNγ production, and proliferation. a) IFNγ was quantitated by ELISA in cell culture supernatants after 24 hours. Error Non-transduced T cells (NTD-T) or MuSK-CAART were co-incubated with Nalm-6 bars indicate mean ± standard deviation of triplicates. c) Proliferation of NTD-T control or individual MuSK domain-specific Nalm-6 target cells at an E:T ratio of (top) and MuSK-CAART (bottom) was evaluated 96 hours after the addition of 10:1 in the absence (0 µg /mL) or presence of matching soluble anti-MuSK IgG4 the indicated anti-MuSK mAbs using Cell Trace Violet (CTV) cellular labeling dye mAb at 1.25 or 6.25 µg /mL, in a total of 10 mg /mL normal human IgG. Cytotoxicity dilution by flow cytometry. (a-c) Representative plots from two (b,c) or three (a) was evaluated at 24 hours using a luciferase-based assay. Error bars indicate individual experiments using different donor T-cells are shown. mean ± standard deviation of triplicates. b) NTD-T or MuSK-CAART cells were Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 5 | Generation and validation of 13-3B5* antibody- Nalm-6 13-3B5* (green) were stained with PE-conjugated mouse anti-human IgG secreting Nalm-6 cells. a) Nalm-6 13-3B5* anti-Ig1 antibody-secreting cells were to quantify BCR density. The mean fluorescence intensity of BCR expression is generated by transducing soluble 13-3B5/anti-Ig1 antibody heavy chain plasmid shown by histogram (b) and bar graph (c). d) Nalm-6 13-3B5 or Nalm-6 13-3B5* into Nalm-6 13-3B5 BCR−expressing cells. Jurkat cells expressing individual cells were co-incubated with either NTD-T or MuSK-CAART cells at indicated E:T MuSK extracellular domains linked with GFP were stained with cell-culture ratios for 8 hours. MuSK-CAART cytotoxicity was measured using a luciferase- supernatants from Nalm-6 13-3B5* for epitope mapping. Soluble 13-3B5 antibody based assay. Error bars indicate mean ± standard deviation of triplicates. binding was detected using anti-human IgG4-APC. b, c) Nalm-6 13-3B5 (red) and Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 6 | Evaluation of Nalm-6 outgrowth in a subset of MuSK- Nalm-6 cells (dotted line indicates cutoff for positive surface IgG expression). CAART-treated mice. a, b) Bioluminescence images from Fig. 3a, b. c) Enlarged g) Graphs indicate the percent of residual Nalm-6 cells that are IgG BCR + and bioluminescence image from (a, red box). d, e) T-cell and Nalm-6 cell percentage the MFI of IgG BCR expression in residual Nalm-6 cells in NTD-T (n = 2) and in the cranial bone marrow in NTD-T treated mice (n = 2) and MuSK-CAART MuSK-CAART (n = 3) treated mice. Error bars indicate mean ± SEM. Unpaired treated mice (n = 5) were analyzed by flow cytometry. f) Representative plot t-test (two-tailed): ns, p > 0.05; *, p < 0.05; **, p < 0.01. showing the mean fluorescence intensity (MFI) of IgG BCR expression in residual Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 7 | Immunologic characterization of the MuSK EAMG followed by MuSK full length (FL) boost, using Jurkat CAAR T-cells expressing syngeneic MuSK-CAART treatment model. a) Splenocytes were harvested individual MuSK domain CAARs linked to GFP. Representative FACS plots are from select MuSK-immunized mice, and purified B-cells were evaluated using shown from three independent experiments. d) Schematic of pMSGV1-1D3- a MuSK-specific and total IgG B-cell ELISpot assay to quantitate the MuSK- CAR (mouse CD8α signal peptide, 1D3 anti-mouse CD19 single chain variable specific B-cell frequency. Dots from duplicated wells were summarized after fragment (scFv), mouse CD28 hinge (HD)/transmembrane domain (TMD)/ normalization with seeded cells (dots per 100,000 cells). MuSK-specific B-cell costimulatory domain, and a mouse CD3ζ.1-3mut domain to confer enhanced frequency is calculated by dividing anti-MuSK B-cells by total IgG B-cells. b) IgG persistence. Schematic of pMSGV1-MuSK-CAAR (human MuSK extracellular subclasses of anti-MuSK antibodies were detected in sera from MuSK-immunized domains (amino acids 24-495), glycine-serine (GS) linker, hCD8α TMD, and mice in reference to sera from negative controls (non-immunized C57BL/6 and human CD137-CD3ζ). e, f) Transduction efficiency of 1D3-CAR and MuSK-CAAR in + + Rag2IL2rγ-deficient mice) and median value was plotted. c) Epitope mapping of primary mouse T-cells (Set 1 and Set 2) and g) percentage of CD4 /CD8 T-cells in serum from a mouse immunized and boosted with MuSK Ig1-Ig2, or MuSK Ig1-Ig2 Set 1 was evaluated on day 4 after T-cell activation (one day prior to injection). Nature Biotechnology Article https://doi.org/10.1038/s41587-022-01637-z Extended Data Fig. 8 | Off-target cytotoxicity of MuSK-CAART against n = 5; NTD-T+ agrin, n = 4; MuSK-CAART, n = 5; MuSK-CAART+ agrin, n = 5; Wise- + + MMP16 LRP4 U87-MG cells was not detected. a) LRP4 (blue) and MMP16 (red) CAART, n = 2). d) U87-MG cells were stained with Calcein-AM before co-culture expression in U87-MG cells was confirmed using flow cytometry. b) MuSK-CAAR with T-cells. Viable cells (GFP ) were detected by fluorescence microscopy and Flag-tagged Wise-CAAR expression were confirmed in primary human at 16 hours after co-culture with T-cells in the presence or absence of agrin. T-cells. c) Luciferase U87-MG cells were co-incubated with human T-cells at the Representative images are shown from two individual experiments. e) Human indicated E:T ratio in the presence (dashed line) or absence (solid line) of agrin. IFNγ production was detected by ELISA in 16-hour co-culture supernatants from Cytotoxicity was measured using a luciferase-based killing assay. Error bars two individual experiments. One-way ANOVA with Holm-Sidak test for multiple indicate mean ± standard deviation from two individual experiments (NTD-T, comparisons: ***, p < 0.001. Nature Biotechnology μ μ

Journal

Nature Biotechnology – Springer Journals

Published: Sep 1, 2023

There are no references for this article.