BCMA/CD19 CAR T cell therapy for refractory myasthenia gravis: Proteomic signatures and single-cell transcriptomics of disease flares.

INTRODUCTION

Myasthenia gravis (MG) is a classic B cell–mediated autoimmune disorder clinically manifested by fluctuating weakness and fatigue of skeletal muscles ( 1 ). Central to its pathogenesis is the breakdown of peripheral tolerance mechanisms, which drives the maturation of autoreactive B lymphocytes into antibody-secreting effector populations that generate high-titer pathogenic autoantibodies ( 2 ). These autoantibodies predominantly target postsynaptic components at the neuromuscular junction, including the acetylcholine receptor (AChR), muscle-specific kinase (MuSK), low-density lipoprotein receptor–related protein 4, and agrin, critical regulators of AChR clustering whose dysfunction impairs synaptic transmission ( 2 ). Therefore, B cell depletion (BCD) therapy is an attractive therapeutic strategy for the management of MG. Therapeutic paradigms for MG have evolved into the biological era with novel agents complementing conventional symptomatic management and immunosuppression ( 3 , 4 ). Several monoclonal antibodies have demonstrated promising efficacy in the treatment of MG based on current clinical studies, including rituximab and inebilizumab ( 5 , 6 ). However, although these treatment approaches are effective, they are unable to specifically target or eliminate antibody-producing plasma cells and demonstrate limited efficacy against B cells residing within inflammatory niches, such as those located in the neuromuscular synapses, spleen, lymph nodes, or bone marrow ( 7 , 8 ). Thus, the inaccessibility and persistence of autoreactive B cells potentially increase the risk of disease relapse. Furthermore, their suboptimal pharmacokinetic profiles necessitate frequent infusion regimens, creating substantial financial burdens and patient adherence challenges.

Chimeric antigen receptor T cells (CAR T cells) are genetically engineered T cells that have demonstrated notable success in treating hematologic tumors ( 9 , 10 ). The therapeutic landscape of CAR T cell therapy is undergoing paradigm-shifting expansion into autoimmune disease management ( 11 , 12 ). Emerging clinical evidence demonstrates that engineered CAR T cells can achieve unprecedented disease remission in the treatment of autoimmune diseases by efficiently depleting autoreactive B cells, thereby resetting the immune system, while maintaining favorable safety profiles ( 13 , 14 ). These encouraging results have motivated in-depth exploration into expanding the application of CAR T cells to a broader spectrum of autoimmune diseases, including MG. A phase 1b/2a clinical trial demonstrated the safety and preliminary efficacy of RNA-engineered CAR T cells (rCAR T cells) targeting B cell maturation antigen (BCMA) in patients with MG ( 15 ). In a separate investigation, human autologous anti-CD19 CAR T cell therapy achieved successful clinical outcomes in a refractory MG case ( 16 ). These studies collectively underscore the therapeutic promise of CAR T cell approaches in MG management. Notably, anti-BCMA rCAR T cell regimen necessitated repeated administrations ( 15 ). Moreover, complete clinical remission was not attained in either study cohort, with persistent dependence on immunosuppressive therapies, detectable pathogenic antibody levels postintervention, and short-term observation ( 15 , 16 ).

B cells have a pivotal function in the pathogenesis of autoimmune diseases, including antigen presentation, cytokine production, and autoantibody production ( 17 ). During the progression of B cell development from pro–B cells to plasma cells, there are corresponding changes in the expression profile of cell surface molecules ( 18 ). CAR T cell therapy either targets CD19, which is expressed across most B-cell lineages, including plasmablasts but not plasma cells, or targets BCMA, which is expressed on plasma cells but fails to eliminate other pathogenic B-cell subpopulations ( 18 ). These residual cells may still lead to disease relapse. Theoretically, monospecific CAR T cell targeting either CD19 or BCMA alone is insufficient to comprehensively reset the pathogenic functions of B cells. To address this limitation, we designed BCMA/CD19 dual-targeted CAR T cell therapy for refractory MG.

On the basis of these data, we conducted a comprehensive 1-year evaluation of BCMA/CD19 CAR T cell therapy in refractory patients with MG, analyzing longitudinal changes in serum proteomic profiles and immune cell dynamics. In particular, we performed single-cell RNA sequencing (scRNA-seq) of peripheral blood mononuclear cells (PBMCs) and bone marrow samples from a posttreatment relapse case to elucidate potential recurrence mechanisms, ultimately identifying previously unknown therapeutic targets.

RESULTS

Patient characteristics

Detailed patient characteristics are shown in Table 1 . Patients were consecutively enrolled between July 2023 and May 2024. All patients (two man and four women) were aged between 32 and 64 years. All patients had generalized MG, classified as MG Foundation of America (MGFA) class II or III before CAR T cell therapy, and had previously received multiple immunosuppressants and immunomodulatory agents. Postintervention status showed unchanged or worsening, meeting the criteria for refractory MG. All patients presented with mild-to-moderate MG symptoms at study entry [patient 1 (P1): quantitative MG (QMG) 8, activities of daily living scale (ADL) 6; P2: QMG 20, ADL 9; P3: QMG 16, ADL 9; P4: QMG 12, ADL 6; P5: QMG 17, ADL 10; P6: QMG 16, ADL 12]. Among the patients, three cases were complicated with thymoma, one case was complicated with thymic hyperplasia, and two cases had normal thymus. Four patients with thymic abnormalities (including thymoma or hyperplasia) had undergone thymectomy. The percentages of B cells in lymphocytes at baseline for P1 to P6 were 1.02, 11.28, 0.00 (previous rituximab exposure), 12.71, 7.50, and 11.37%, respectively.

Clinical efficacy and expansion of CAR T cell postinfusion

Administration of CAR T cell led to remission from unchanged or worse, patients P1 to P5 achieved minimal manifestations status (MMS) and discontinued all medications by month 6 of CAR T therapy ( Fig. 1, H and I ). All patients exhibited an ADL score of 0, with P3 and P5 maintaining a QMG score of 2, while the remaining three patients showed complete normalization (QMG 0). Upon extended follow-up to 12 months, all patients sustained symptom- and medication-free remission (QMG 0 and ADL 0) ( Fig. 1, B and C ). Serological analyses revealed AChR antibody seroconversion to negative in P1, P2, and P5, whereas P3 and P4 displayed borderline titers of 0.7 and 0.86 nM, respectively ( Fig. 1D ). All patients exhibited complete peripheral BCD (0% B lymphocytes) by day 14 post-CAR T cell infusion, with B cell reconstitution (BCRC) occurring at day 180 for P1, day 270 for P3, day 60 for P4, and day 90 for P5, but P2 failed to achieve BCRC through day 360 ( Fig. 1K ), accompanied by sustained declines in serum soluble BCMA (sBCMA) levels within 30 days across the cohort ( Fig. 1L ). Following CAR treatment, all patients indicated a reduction in immunoglobulin G (IgG), IgA, and IgM levels within 60 days ( Fig. 1, E to G ).

P6 demonstrated sustained clinical improvement within 180 days following CAR T cell infusion, as evidenced by a reduction in the QMG score from 16 to 5 and a decrease in the ADL score from 12 to 3 ( Fig. 1, B and C ). CD19 + BCRC occurred on day 180 ( Fig. 1K ). The AChR antibody titer became undetectable at 0.01 nM on day 90 and showed a slight increase to 0.33 nM on day 180, remaining below the diagnostic threshold ( Fig. 1D ). However, disease recurrence was observed on day 187 with an ADL score of 6. Because of a combination of persistent symptoms, financial constraints, and geographic barriers, the patient was unable to attend follow-up visits at the hospital and was instead monitored via telemedicine-based ADL assessments over time. At the 12-month follow-up, the ADL score increased from 3 at day 180 to a maximum of 7 yet remained significantly lower than the baseline score of 12 ( Fig. 1C ).

The in vitro validation of the manufacturing process demonstrated stable transduction efficiency (53 to 73%) in CAR T cell products ( Fig. 1J ). CAR T cells exhibited rapid expansion kinetics postinfusion, peaking around day 14 in all six patients, followed by prompt contraction ( Fig. 1, M and N ). Concurrent immunophenotypic profiling revealed in vivo differentiation toward central memory T cell subsets (fig. S1B), suggesting lymphoid organ homing and tissue-penetrating capacity critical for sustained therapeutic effects.

Patient safety

Six patients exhibited good tolerance to CAR T cell therapy with a favorable safety profile ( Table 2 ). No instances of CAR T cell–related immune effector cell–associated neurotoxicity syndrome (ICANS) were observed. Among the cohort, two patients (P1 and P5) developed grade 1 cytokine release syndrome (CRS) manifesting as mild fever within 28 days postinfusion, both resolving with standard supportive care. Notably, patient P5 experienced transient grade 3 hematologic toxicities on day 7, consisting of leukopenia (3-day duration) and thrombocytopenia (4-day duration), potentially attributable to the lymphodepletion chemotherapy. A delayed adverse event occurred in patient P3, who developed grade 3 leukopenia at day 90 posttreatment. P6 presented with concurrent grade 3 leukopenia and grade 4 neutropenia at the 90-day follow-up, which rapidly resolved following administration of recombinant human granulocyte colony-stimulating factor. Both patients P1 and P3 developed grade 2 bacterial pneumonia manifesting delayed onset 28 days post–CAR T cell administration ( Fig. 2, A to E ). No other treatment-related severe adverse events including hepatic or renal functional impairments were reported during the observation period (fig. S2). Longitudinal monitoring of serum protein levels encompassing key inflammatory mediators—interleukin-1β (IL-1β), IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, interferon-α (IFN-α) and IFN-γ, and tumor necrosis factor–α (TNF-α)—revealed no continuous abnormal elevation in circulating cytokine concentrations among six patients ( Fig. 2F ).

Olink targeted proteomics

We assessed the serum expression of 92 inflammatory and 92 immune response proteins using Olink technology. The Olink inflammation panel revealed that multiple cytokines exhibited a progressive increase during the first 14 days following CAR T cell infusion, including those associated with enhanced CAR T cell migration, reduced exhaustion, and improved cytotoxic activity (CXCL9, CXCL10, TNFRSF9, IL-8, IL-10, and IFN-γ) ( Fig. 2G ). However, these cytokine levels began to decline by day 30. Notably, IFN-γ linked to cytotoxic effector responses demonstrated a sustained upward trend (fig. S2G). The Olink immune response panel revealed significant up-regulation of anti-inflammatory mediators, negative regulatory factors, and T cell activation markers [leukocyte immunoglobulin like receptor B4 (LILRB4), lymphocyte activation gene 3 (LAG3), nuclear factor of activated T cells 3 (NFATC3), CD83, and CD28], along with marked down-regulation of the proinflammatory mediator triggering receptor expressed on myeloid cells-1 (TREM1) ( Fig. 3, A and B ).

Reconstitution of B cell function

Regarding long-term outcomes, five patients achieved BCRC at a mean of 156 ± 83 days posttreatment (median, 180 days; range, 60 to 270 days; Fig. 1K ), while P2 maintained complete B cell aplasia through 12-month follow-up. Phenotypic profiling revealed that reconstituted B cells in patients P1, P3, P4, and P5 predominantly displayed a naïve B cell phenotype (CD19 + CD27 − IgD + ) ( Fig. 3C ), with concomitant down-regulation of B cell activation/antigen presentation markers [OX40 ligand (OX40L), CD70, CD80, CD86, and CD83] ( Fig. 3D ) and proinflammatory cytokines [IL-6 and granulocyte-macrophage colony-stimulating factor (GM-CSF)], alongside up-regulation of the immunoregulatory cytokine IL-10 ( Fig. 3E ). Notably, P6 experienced disease relapse at day 187. Comparative analysis of day 180 samples demonstrated significant expansion of age-associated B cells (ABCs) (CD21 − CD11c + T-bet + ) and plasma cells (CD38 + CD138 + ), coinciding with elevated expression of B cell activation markers and proinflammatory cytokines compared to baseline levels ( Fig. 3, C to E ). Furthermore, we identified a marked reduction in AChR-α–specific B cell populations postreconstitution (fig. S3, A and B). Functional analyses revealed impaired differentiation capacity toward antibody-secreting cells (ASCs) and diminished IgG production upon T cell–dependent stimulation (CD40L + IgM + IL-21) (fig. S3, C to E).

Single-cell sequencing reveals the comprehensive cellular landscape of relapsed patient

To investigate potential mechanisms of relapse, we performed scRNA-seq on paired pretreatment and day 180 posttreatment PBMCs and bone marrow specimens from P6 who experienced relapse. Elevated inflammatory scores in lymphocytes were observed in P6 at day 180 following CAR T cell therapy, accompanied by heightened cytokine activity scores in T cells and natural killer (NK) cells (fig. S4). Comparative analysis revealed notable compositional remodeling, including expansion of ABCs in peripheral blood and plasma cells in the bone marrow ( Fig. 4C ). Differential gene expression analysis identified 203 up-regulated and 109 down-regulated genes in posttreatment B cells across both compartments ( Fig. 4, D and E ). The up-regulated genes were enriched in B cell receptor (BCR) signaling, B cell activation/proliferation/differentiation, antigen presentation, and cytokine production pathways [Gene Ontology (GO) analysis] ( Fig. 4F ). Cellular interaction analysis demonstrated enhance cell networking posttreatment ( Fig. 4G ). The CellChat analysis revealed that the ligand-receptor signaling pathways [macrophage migration inhibitory factor (MIF)–CD74/CXCR4 and MIF-CD74/CD44] between the reconstructed B cells/plasma cells and other cell types were up-regulated (fig. S5). Subsequently, high-resolution clonal tracing through integrated transcriptomic and BCR repertoire analysis demonstrated that while most clones remained singleton, distinct hyperexpanded clones emerged posttreatment ( Fig. 4I ). BCR deep sequencing revealed complete absence of IgG/IgA clonotypes with predominant nonclass-switched IgM heavy chain in the repertoire ( Fig. 4J ).

scRNA-seq initially identified expanded ABCs in relapsed patients with MG. Subsequent validation in an expanded cohort (40 healthy controls versus 46 treatment-naïve patients with MG) through quantitative flow cytometry analysis confirmed significant ABC expansion in de novo MG cases ( P < 0.001; Fig. 4H ), correlating positively with QMG scores ( r = 0.431, P = 0.003; Fig. 4H ). Pseudotime trajectory reconstruction revealed sequential up-regulation of the differentiation regulator Fc receptor-like 5 (FCRL5) along ABC–to–plasma cell transition paths ( Fig. 5B ). Notably, CAR T cell responders (P1/P4/P5) demonstrated FCRL5 down-regulation postinfusion across ABCs, switched memory (SwM) B cells, plasmablasts, and plasma cells, whereas nonresponders (P6) maintained elevated FCRL5 levels ( Fig. 5D ). Furthermore, patients with MG who relapsed after CD20 monoclonal antibody (rituximab) therapy demonstrated concurrent elevation of ABCs and up-regulated FCRL5 expression ( Fig. 5D ). Analysis revealed pronounced FCRL5 overexpression throughout B cell subpopulations in these patients, with maximal expression intensity observed specifically within the ABCs compartment ( Fig. 5E ).

We systematically evaluated the dynamic effects of CAR T cell infusion on T cell homeostasis. A notable observation was the significant decrease in CD4 + T cell frequency accompanied by a compensatory increase in CD8 + T cell proportion, ultimately leading to an inversion of the CD4/CD8 ratio (fig. S6). To further dissect the functional consequences of B cell modulation, we analyzed T cell responses at three distinct stages: baseline, BCD, and BCRC. BCD refers to the period around 1 month after CAR T cell infusion when B cells are depleted, and BCRC refers to the time point around 3 months after BCRC. Our findings revealed that during BCD, key cytotoxic effector molecules in CD8 + T cells, including IFN-γ, CD107a, perforin, and granzyme B, were markedly up-regulated. While these markers generally declined upon BCRC, IFN-γ levels remained elevated ( Fig. 6G ). In contrast, CD4 + T cells exhibited a predominantly anti-inflammatory profile during BCD, characterized by increased IL-10 expression and reduced levels of proinflammatory cytokines such as IL-6, IFN-γ, and TNF-α. These cytokine levels gradually rebounded during BCRC ( Fig. 6G ). Furthermore, we assessed changes in T cell activation markers (CD25, CD69, and CD27) and inhibitory receptors [T cell immune receptor with Ig and ITIM domains (TIGIT), Lag3, and programmed death receptor 1 (PD-1)], identifying CD27 and TIGIT as the most dynamically regulated molecules. Specifically, surface expression of CD27 and TIGIT on both CD4 + and CD8 + T cells was significantly enhanced during BCD and subsequently decreased with B cell recovery ( Fig. 6H ). scRNA-seq of patient P6 corroborated these findings, showing an expansion of cytotoxic CD8 + T cells marked by granzyme H (GZMH), granzyme K (GZMK), and granulysin (GNLY), along with activated CD8 + T cells expressing human leukocyte antigen (HLA)–DR ( Fig. 6C ). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of differentially expressed genes from bone marrow and peripheral CD4 + and CD8 + T cells highlighted prominent activation of pathways associated with T helper 17 cell (T H 17 cell) differentiation, chemokine signaling, nuclear factor κB (NF-κB) signaling, apoptosis, phagocytosis, and cytotoxicity ( Fig. 6E ). T cell receptor (TCR) sequencing further demonstrated a significant increase in clonal expansion ( Fig. 6F ). During the longitudinal monitoring of NK cells and monocytes in patients undergoing CAR T cell therapy, we observed a transient expansion of these immune cell populations, followed by a gradual normalization to baseline levels ( Fig. 2C and fig. S6F). Single-cell transcriptomic analysis of the relapsed patient P6 demonstrated that the up-regulated genes in myeloid cells after CAR T cell treatment were predominantly associated with inflammatory signaling pathways, such as NF-κB signaling, TNF signaling, nonobese diabetic–like receptor signaling, and T H 17 cell differentiation (fig. S7). Of particular significance was the marked activation of the chemokine signaling pathway. Notably, CXCL2-expressing monocytes were significantly increased in both bone marrow and PBMCs (fig. S7).

DISCUSSION

MG is a prototypical B cell–mediated autoimmune disorder, and the attainment of sustained remission is of paramount importance in improving patients’ quality of life ( 19 ). Approximately 10 to 20% of patients with MG demonstrate suboptimal responses to first-line immunosuppressive therapies and experience recurrent myasthenic crises, prompting our classification of this subgroup as refractory MG ( 19 , 20 ). Despite the advent of targeted biologic therapies in MG management, including reported efficacy of emerging agents such as rituximab, telitacicept, and efgartigimod in certain refractory cases ( 21 – 25 ), a subset of patients in our study continued to exhibit resistance to these biologic treatments. These “ultrarefractory” MG cases urgently require novel therapeutic strategies. Thus, we report a clinical investigation of BCMA/CD19 dual-targeted CAR T cell therapy in patients with refractory MG, assessing both the long-term safety and efficacy profiles over a 12-month follow-up period, along with alterations in serum proteomic signatures and immune cellular dynamics. To further explore the underlying mechanisms of disease recurrence, we conducted scRNA-seq on PBMCs and bone marrow samples obtained from a patient who experienced relapse following CAR T cell treatment.

Our preliminary findings first demonstrated that BCMA/CD19 CAR T cell therapy exhibits both excellent tolerability and remarkable clinical efficacy in refractory MG. At the 12-month follow-up, complete medication-free remission (including pyridostigmine withdrawal) was sustained in five patients who responded to the treatment, with seroconversion of AChR antibodies observed in a subset of cases. Our study demonstrates notable advantages compared to other CAR T cell therapies for MG (CD19 or BCMA single-target CAR T cell or BCMA rCAR T cell) ( 15 , 16 , 26 ). First, case reports demonstrate that patients receiving CD19 or BCMA single-target CAR T cell therapy continue to require posttreatment administration of pyridostigmine or low-dose corticosteroids, while RNA-based CAR T cell regimens necessitate repeated infusions over 6 weeks. In contrast, all five patients in our cohort achieved complete discontinuation of all medications following a single infusion of BCMA/CD19 dual-targeted CAR T cells. Second, the 12-month follow-up period in our study significantly surpasses the 6-month observation window in BCMA rCAR T cell trials and the 62-day follow-up duration reported in CD19 CAR T cell studies, highlighting the sustained efficacy of our therapeutic approach. Furthermore, three patients achieved serological remission characterized by AChR antibody seroconversion, while two additional cases exhibited titers approaching the threshold value, contrasting sharply with the persistent detection of AChR antibodies in both BCMA rCAR T cell and CD19 CAR T cell treatment cohorts. Crucially, according to the MGFA postintervention status classification, all five patients attained remission, whereas comparator studies only documented MMS or improvement.

The safety profile proved favorable, with no ICANS reported and only two instances of grade 1 CRS manifesting as self-limiting low-grade fever within 28 days postinfusion, both resolving with standard supportive care. Transient hematologic toxicities were observed, and, critically, no severe treatment-related adverse events or unexpected complications emerged throughout the observation period. CRS is a well-characterized adverse effect associated with the activation and expansion of CAR T cells, which drive systemic inflammation through the recruitment of myeloid cells that secrete high levels of proinflammatory cytokines ( 27 ). Previously reported studies on BCMA/CD19 dual-targeted CAR T cell therapy for systemic lupus erythematosus have shown a 90% incidence of CRS ( 28 ), whereas in our study, only two patients experienced CRS, both of which were grade 1 events. In contrast, 92% of patients in studies on BCMA/CD19 CAR T cell therapy for multiple myeloma experienced CRS, with 8% developing grade 3 or higher CRS ( 29 ). This is consistent with evidence, suggesting that CRS severity correlates with B cell burden ( 13 ). Notably, no cases of high-grade CRS have been observed in patients with MG treated with CAR T cell therapy, which is most likely attributable to the significantly lower B cell burden in patients with MG compared to those with active B cell malignancies. Besides, other adverse events included grade 3 hematological toxicity, all of which could be rapidly improved after symptomatic treatment.

Previous studies have proposed the concept of “immune reset” in CAR T cell therapy for autoimmune diseases, which refers to the process of effectively eliminating pathogenic lymphocytes and subsequently replenishing with naïve cells ( 11 , 30 ). However, the protein and cellular-level alterations following this process remain insufficiently investigated. We conducted dynamic monitoring of serum proteome and immune cells in patients with MG treated with BCMA/CD19 CAR T cells. Olink targeted proteomics revealed a substantial up-regulation of anti-inflammatory and negative regulatory factors (IL-10, IFN-γ, LILRB4, LAG3, and NFATC3), along with a marked down-regulation of proinflammatory molecules [TREM1 and signaling lymphocytic activation molecule family member 1 (SLAMF1)], indicating a shift toward the restoration of immune homeostasis. B cells are regarded for their capacity to produce antibody, release cytokines, and present antigen ( 31 ). We monitored B cell function following reconstruction and observed a predominance of naïve B cells, along with a marked reduction in pathogenic memory B cells and ASCs (including plasmablasts and plasma cells). In addition, there was an increased proportion of IL-10–producing B cells and a decrease in the expression of costimulatory molecules on the B cell surface. There was a significant reduction in AChR-specific B cells following reconstruction, accompanied by a persistent decline in serum AChR antibody titers, even leading to their conversion to negative. These results suggest that BCMA/CD19 CAR T cell therapy may exert therapeutic effects in patients with MG by modulating the overactivated B cell–related immune microenvironment and facilitating its normalization. However, a recent study reported that, over extended follow-up periods, disease recurrence occurred after CAR T cell therapy ( 32 ). Notably, the study observed that autoimmune repertoire still emerge after CAR T cell treatment in chronic inflammatory demyelinating polyneuropathy ( 32 ). These findings indicate that the extent to which CAR T cell therapy can achieve complete immune resetting remains uncertain and warrants further long-term monitoring. In addition to alterations in B cell function, a resetting phenomenon of T cells also occurs following CAR T cell therapy, with the most prominent manifestation being the inversion of the CD4/CD8 ratio and a significant increase in the cytotoxic activity of CD8 + T cells. Furthermore, it has been observed that during the BCD phase, surface CD27 levels on T cells are markedly elevated, which may be attributed to diminished interaction with CD70 on B cells ( 33 ).

Another important discovery in our study was the first-time use of single-cell sequencing to unveil alterations in the immune landscape of patients with MG experiencing relapse following CAR T cell therapy. We observed up-regulated ABCs cells in relapsed patients, accompanied by significant FCRL5 elevation in both bone marrow and peripheral blood. Notably, this phenomenon was also detected in patients who relapsed after rituximab treatment. ABCs represent a specific effector B cell subset that exhibits numerical expansion in systemic autoimmune disorders ( 34 – 36 ). In autoimmune contexts, these cells demonstrate enriched autoreactive specificities and display hallmarks of antigen-experienced lymphocytes. Emerging evidence indicates ABCs have tissue-resident capacity and can rapidly differentiate into ASCs upon antigen rechallenge or innate immune stimulation, thereby serving as functional precursors to ASCs ( 37 , 38 ). Our study also reported elevated peripheral ABCs in patients with MG, showing positive correlation with disease severity, which suggests their potential involvement in MG pathogenesis. Furthermore, we identified that the FCRL5 gene demonstrates the highest expression abundance in ABCs of patients with MG. Pseudotime trajectory analysis revealed that FCRL5 exhibits peak expression during ABC differentiation and bone marrow plasma cell development, highlighting its promise as a previously unidentified therapeutic target. Of particular clinical significance, FCRL5-directed CAR T cell therapy has already been successfully implemented in multiple myeloma patients refractory to both BCMA- and G protein-coupled receptor class C group 5 member D (GPRC5D)–targeted CAR T cell regimens ( 39 ).

Our study also has some limitations. First, the sample size of this study is limited, which restricts the generalizability of findings and precludes comparative analyses. Second, the absence of a control group receiving standard care or placebo may introduce potential bias in outcome interpretation. Third, the exclusion of patients with positive anti-MuSK antibodies raises concerns about the extrapolation of results to other MG subtypes. Fourth, the follow-up period of this study is still limited. The long-term efficacy of CAR T cell therapy requires further evaluation through extended follow-up to determine whether it can truly induce immune system reconstitution. Fifth, some patients exhibited absent B cells at enrollment due to prior rituximab treatment before CAR T cell therapy, while others showed persistent BCD even 1 year after CAR T cell infusion. As a result, longitudinal assessment of B cell functional markers, including ABC and FCRL5, could not be fully matched before and after intervention, leading to incomplete data coverage. Sixth, our study identified FCRL5 as a potential therapeutic target for relapsed patients. However, targeted interventions such as CAR T cell therapy or bispecific T cell engagers directed at FCRL5 have not been applied to this specific relapsed patient population. Future research will focus on conducting comprehensive preclinical studies to thoroughly evaluate safety profiles and potential adverse effects, before initiating relevant clinical trials.

In conclusion, this study provides the first report of 1-year follow-up data demonstrating the efficacy and safety profile of BCMA/CD19 CAR T cell therapy in refractory patients with MG. The landmark findings include five patients achieving drug-free remission with anti-AChR antibody levels either seroconverting to negative or approaching threshold values, while maintaining a robust safety profile devoid of severe adverse events. We pioneered longitudinal monitoring of serum protein dynamics and immunophenotypic alterations post–CAR T cell infusion, validating the concept of operational immune reset by translating it from theoretical frameworks to clinical observations. Furthermore, this investigation presents the first mechanistic insights into CAR T cell relapse patterns while uncovering FCRL5 as a previously unidentified therapeutic target with translational potential in treatment-resistant populations.

MATERIALS AND METHODS

Participants

All six patients with treatment-refractory MG, who underwent CAR T cell therapy, were recruited at the Affiliated Hospital of Xuzhou Medical University. Inclusion criteria include the following: (i) participants aged 14 to 70 years with confirmed refractory MG ( 19 ), meeting all baseline severity thresholds: MG-ADL score ≥ 6, QMG scale score ≥ 8, and MGFA classes II to IV at screening; (ii) seropositivity for anti-AChR antibodies; (iii) Eastern Cooperative Oncology Group performance status 0 or 1; (iv) preserved organ function (alanine aminotransferase <3 times the upper limit of normal, aspartate aminotransferase <3 times the upper limit of normal, total bilirubin of <2.0 mg/dl, and serum creatinine of <2.5 mg/dl); and (v) adequate vascular access for apheresis without contraindications to cytoreductive therapy. Exclusion criteria comprised the following: (i) neuropsychiatric conditions impairing protocol compliance, (ii) active infections (human immunodeficiency virus, hepatitis B virus, hepatitis C virus, and Mycobacterium tuberculosis ), (iii) highly allergic constitution or history of severe allergies, (iv) recent immunomodulators use (intravenous immunoglobulin, plasma exchange, or biologic agents within 4 weeks prelymphodepletion), and (v) thymectomy within 6 months before lymphodepleting chemotherapy. The validation cohort comprised 40 healthy controls and 46 patients with new-onset MG. The new-onset patients were defined as MG at the initial onset that had not yet received any immunosuppressive therapy ( 40 ). The exclusion criteria for new-onset patients were as follows: (i) incomplete baseline records; (ii) presence of other autoimmune diseases; (iii) history of malignancy except thymoma; and (iv) severe cardiac, hepatic, renal, neurological, or hematological disorders or active severe infections. Healthy controls were individuals without pharmacological treatment and were matched to the new-onset patients with MG by age and sex (table S2). Exclusion criteria for controls included any concurrent infectious, inflammatory, autoimmune, neoplastic, or systemic diseases. In addition, the validation cohort of this study included 14 patients with MG receiving rituximab treatment (table S3). These individuals underwent regular follow-up visits at our center every 3 months, including monitoring of immunological parameters. The cohort comprised eight patients in remission and six patients experiencing relapse. The B cells reappeared in all selected patient samples during the test. The rituximab doses reinfused were all low dose, consisting of 500 to 700 mg administered over 2 consecutive days (100 mg on day 1 and the remaining dose on day 2) as induction therapy. The definition of remission was considered, as patient has no symptoms or signs of MG ( 19 ). The relapse of MG was defined as the reappearance or worsening of one or more MG symptoms or signs of muscle weakness with MG-ADL score of ≥2 compared to the previous stage ( 41 ). These constitute independent validation cohorts used for research purposes. This study was approved by the Medical Ethics Committee of Xuzhou Medical University Affiliated Hospital (approved numbers: XYFY2022-KL057-02 and XYFY2024-KL416-01), and all participants have written informed consent.

CAR T cell production

The bispecific CAR T cell products were manufactured from participants’ autologous T lymphocytes through lentiviral vector transduction, as previously reported by our studies ( 29 ) and the Supplementary Materials. The second-generation CAR construction containing the bispecific tandem single-chain variable fragments, CD8α hinge and transmembrane domains, 4-1BB costimulatory domains, and CD3ξ activation domain (fig. S1A).

Lymphodepleting chemotherapy and CAR T cell treatment

Participants received lymphodepleting chemotherapy with fludarabine (30 mg/m 2 per day) and cyclophosphamide (300 mg/m 2 per day) for 3 consecutive days (days −5 to −3). Patients received a dose of 1 × 10 6 CAR cells/kg of body weight following conditioning with cyclophosphamide and fludarabine.

Safety and efficacy monitoring

Treatment-emergent adverse events were prospectively monitored and systematically documented, with severity assessments performed using the National Cancer Institute Common Terminology Criteria for Adverse Events v5.0. Specifically, CRS and ICANS were characterized and graded using established American Society for Transplantation and Cellular Therapy consensus criteria, while other neurotoxicity effects were evaluated through published criteria ( 42 , 43 ). Participants were monitored at 1, 3, 6, 9, and 12 months posttreatment, with comprehensive documentation of adverse events throughout the follow-up period. Efficacy was evaluated by analyzing the longitudinal changes in MG-ADL and QMG scores from baseline through 12 months, with assessments conducted at predefined follow-up intervals.

Evaluation of CAR T cells expansion by real-time polymerase chain reaction

The blood samples were collected from the patient within 60 days after CAR T cells infusion. Blood DNA was extracted, and then CAR transgene copy number was detected by real-time polymerase chain reaction (PCR), according to the manufacturer’s protocol (Yeasen Biotechnology).

Enzyme-linked immunosorbent assay

Serum sBCMA levels were detected with enzyme-linked immunosorbent assay (ELISA) kits (ACRO Biosystems) according to the manufacturer’s instructions. Optical densities were recorded at 450 nm, and the concentration was calculated in accordance with the standard curve.

Analysis of cytokine

Serum samples from all patients were isolated and analyzed for IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IFN-α, IFN-γ, and TNF-α using flow cytometry fluorescence technology (Luminex 2000 system), in strict accordance with the manufacturer’s protocols.

Immunomonitoring

PBMCs were isolated from freshly collected anticoagulated blood using Ficoll-Paque density gradient centrifugation (STEMCELL Technologies). For CAR T cell detection, cells were surface stained with 7-Aminoactinomycin D (7-AAD) viability dye, CD3, CD45RA, CD27 (BioLegend), and polyglycine-serine (G4S) linker (Cell Signaling Technology). B cell subpopulation analysis included surface staining with CD19, CD27, IgD, CD24, CD38, CD138, and FCRL5 (BioLegend). ABCs were identified through sequential surface staining (CD3, CD19, CD21, CD11c, and FCRL5), fixation/permeabilization using BD Cytofix/Cytoperm reagents (BD Biosciences), and intracellular T-bet detection (BioLegend). For B cell–cytokine profiling, PBMCs were stimulated for 5 hours with 2 μl of eBioscience Cell Stimulation Cocktail (Invitrogen) before surface marker staining (CD19), intracellular fixation/permeabilization (eBioscience Buffer Set), and cytokine detection (IL-6, IL-10, and GM-CSF; BioLegend). B cell antigen–presenting and activation markers (CD19, OX40L, CD70, CD80, CD83, and CD86) and T cell functional markers (CD3, CD4, CD8, CD25, CD69, CD27, TIGIT, PD-1, and LAG3) were assessed using BioLegend antibody panels. T cell cytokine analysis involved eBioscience Cell Stimulation Cocktail (Invitrogen) stimulation, followed by surface staining (CD3, CD4, and CD8) and intracellular cytokine detection (IL-2, IL-6, IL-10, IL-17A, IFN-γ, and TNF-α; BioLegend). CD8 + T cell cytotoxicity was evaluated through CD107a surface expression and intracellular perforin/granzyme B detection postpermeabilization. The flow cytometry gating strategy for T cell function detection is shown in fig. S8. All data were acquired on a BD FACSAria III flow cytometer and analyzed using FlowJo v10 (BD Biosciences). Antibodies used in immunomonitoring are listed in table S1.

AChR-α–specific B cell detection

Following the manufacturer’s protocol, we synthesized biotin-conjugated AChR-α protein (OriGene) and subsequently incubated it with dCODE Klickmer (Immudex) to establish a B cell–targeting dCODE Klickmer scaffold specific for AChR-α recognition. For AChR-α–specific B cell detection, PBMCs were stained with the CD19 (BioLegend) and AChR-α–loaded dCODE Klickmer and then detected by flow cytometry and laser confocal microscopy (Leica Microsystems).

Functional B cell assay

B cells (10 5 per well) isolated from samples collected before CAR T cell therapy and following BCRC were seeded in 96-well plates stimulated with recombinant human CD40L (50 ng/ml; PeproTech), F(ab′)2 fragment goat anti-human IgM (10 μg/ml; Jackson ImmunoResearch) and recombinant human IL-21 (50 ng/ml; R&D Systems) for 7 days. Then, B cells were collected for antibody cocktail staining in response to ASCs differentiation [anti–CD19-allophycocyanin (APC), anti–CD27-BV421, anti-CD38–fluorescein isothiocyanate (FITC) and anti-CD138-APC-Cy7] (all from BioLegend) as described above. Meanwhile, supernatant also was retained for secreted IgG (catalog no. V01190H, VICMED) detection by ELISA kits.

Olink proteomics detection and analysis

Serum samples were analyzed using the Olink Inflammation and Immune Response Panel (Olink Proteomics AB, Uppsala, Sweden), a high-sensitivity multiplex immunoassay platform. The experimental workflow of proximity extension assay followed published methodology ( 44 ). Briefly, target proteins were simultaneously recognized by two matched DNA oligonucleotide-conjugated antibodies, enabling proximal hybridization of their attached DNA strands. Subsequent extension by DNA polymerase generated double-stranded DNA templates, which were amplified via quantitative real-time PCR using a microfluidic PCR system (Signature Q100, LC-Bio Technology Co. Ltd., Hangzhou, China). Protein expression levels were quantified as normalized protein expression (NPX) values, representing relative concentrations on a log 2 -transformed arbitrary scale. All bioinformatics were analyzed using the OmicStudio tools ( www.omicstudio.cn/tool ).

Library construction and sequencing of human scRNA-seq

scRNA-seq and V(D)J sequencing were conducted by OE Biotech Co. Ltd. (Shanghai, China) using the 10x Genomics platform. Peripheral blood and bone marrow were collected from relapsed P6 at baseline and 6 months postinfusion. Following mononuclear cells isolation via Ficoll-Paque density gradient centrifugation (STEMCELL Technologies), samples were immediately processed for sequencing analysis. Single-cell suspensions were then loaded onto the 10x Chromium platform, followed by 5′ gene expression library and immune repertoire library preparation. All libraries were then subjected to NovaSeq 6000 (Illumina, San Diego, USA) with 2× 150 paired-end reads to obtain sequence information.

Data processing for scRNA-seq and single-cell V(D)J sequencing

Raw sequencing data were aligned to the GRCh38 human reference genome and quantified using Cell Ranger (v8.0, 10x Genomics). The Seurat package (v4.4.0) was used for downstream analysis with the following quality control parameters: Cells were retained if they contained 200 to 7500 detected genes and exhibited mitochondrial gene content below 10%. After filtering, 37,744 cells were retained from the human samples. After normalization using the NormalizeData function, highly variable genes were identified through the FindVariableFeatures method. Data scaling with z -score transformation was performed using the ScaleData function before principal components analysis–based dimensionality reduction. Batch effects were corrected using the RunHarmony function. Cell clustering was conducted with the FindNeighbors and FindClusters functions. Nonlinear dimensionality reduction and visualization were performed using Run Uniform Manifold Approximation and Projection (RunUMAP). To integrate datasets from different batches for unsupervised clustering, we used the Harmony algorithm to correct for batch effects. Highly expressed genes in each cluster were identified using the FindAllMarkers function, with the top genes as marker genes. Cell types for clusters were annotated on the basis of known marker genes. For single-cell immune profiling, V(D)J sequencing data were processed through Cell Ranger V(D)J, with clonotype information subsequently integrated into the Seurat metadata. To ensure dataset specificity for repertoire analysis, we excluded cells containing both BCR and TCR sequences through doublet filtration.

Differentially expressed genes identification and pathway enrichment analysis

Differentially expressed genes of two comparison groups (baseline and 6-month posttreatment) were identified using FindAllMarkers function. Statistically significant differentially expressed genes were defined with P adj < 0.05. Pathway enrichment analysis was performed using clusterProfile v4.4.2. Up-regulated and down-regulated differentially expressed genes were separately used as input, testing against the GO and KEGG. Significant enrichment pathways were defined with P .adjust < 0.05. In addition, inflammatory and cytokines scoring for cells from patients at baseline and at 6 months after the infusion was carried out with the AddModuleScore function in Seurat v4.2.0 using a literature gene set ( 45 ).

Cell-cell communication analysis

The interaction intensity of information flow between cells was calculated using CellChat (version 1.6.1). To infer cell state–specific communications, CellChat identified differentially expressed signaling genes across all cell groups in the scRNA-seq dataset using the Wilcoxon rank sum test with a significance level of 0.05. It modeled ligand receptor–mediated signaling interactions based on the law of mass action. Communication within specified subsets was predicted using Nichenet (version 2.0.4), which assessed target genes and ligands with high activity rankings and inferred ligand-to-target signal transduction pathways.

Cell development analysis

Cell development was analyzed using Monocle (version 2.26.0). Monocle uses a pseudotime ordering strategy to place individual cells along a trajectory reflecting a biological process, such as cell differentiation, using asynchronous cell progression through these processes. CytoTRACE (version 0.3.3) assessed the differentiation status of cells based on clustering results from RunUMAP. The differentiation state, captured by the gene count signature, predicted the differentiation potential of cells based on key RNA characteristics derived from scRNA-seq data.

Statistical analysis

Individual values are presented throughout, as the sample size of this case series is small. In addition, descriptive statistics are used for reporting specific parameters between baseline and 12-month follow-up. Continuous variables are presented as means ± SD or median with interquartile ranges. For data following a normal distribution, an independent sample t test was used, whereas the Mann-Whitney U test was used for skewed data. All statistical analyses were performed using GraphPad Prism 8.0 (GraphPad Software, La Jolla, CA, USA). Statistical significance was set at a P value of less than 0.05. The bioinformatic statistical analyses are performed using R (v4.3.0, R Foundation, Vienna, Austria).

Acknowledgments

We are grateful to KingMed Diagnostic Laboratory for providing excellent AChR antibody titer detection using radioimmunoassay technology. We thank LC-Bio Technology Co. Ltd. for offering Olink sequencing technical services. We also express our gratitude to OE Biotech Co. Ltd. for providing single-cell sequencing technical services, especially J. Yu and H. Xing for bioinformatics analysis assistance.

Funding:

This work was supported by the National Natural Science Foundation of China (grant numbers 82273207 and 82471378 to M.S.), Construction Project of High Level Hospital of Jiangsu Province (grant numbers GSPJS202503 to Y.Z.), Basic Research Program of Jiangsu (grant numbers BK20231158 to Y.Z. and BK20243034 to M.S.), Young Academic Leaders of Jiangsu Qinglan Project (grant numbers 202251 to D.L.), Medical Research Project of Jiangsu Provincial Health Commission (grant numbers M2022118 to Y.Z.), and Open Project of Key Laboratory of Colleges and Universities in Jiangsu Province (grant numbers XZSYSKF2021041 to Y.Z.)

Author contributions:

Conceptualization: G.C., M.S., J.Z., and Y.Z. Methodology: X.D., D.L., X.H., M.Y., G.Q., H.L., Y.Z., X.G., J.Z., X.A., Z.Z., M.S., S.L., Z.S., T.L., T.M., G.W., J.C., and S.Z. Software: Z.W., D.L., D.P., X.H., M.Y., H.L., J.Z., X.A., Z.Z., M.S., S.L., Z.S., T.L., and J.C. Validation: X.D., D.L., D.P., X.H., M.Y., H.L., Y.Z., X.G., J.Z., X.A., Z.Z., M.S., T.L., T.M., G.W., G.C., and S.Z. Formal analysis: X.D., Y.D., Z.W., D.L., D.P., X.H., M.Y., G.Q., Y.Z., X.G., J.Z., X.A., Z.Z., M.S., S.L., T.L., T.M., and G.W. Investigation: Z.W., D.L., X.H., H.L., X.G., X.A., M.S., T.L., and S.Z. Resources: X.H., X.A., and Y.Z. Data curation: Z.W., D.L., X.H., M.Y., G.Q., X.G., J.Z., X.A., Z.Z., M.S., S.L., T.L., T.M., and G.C. Visualization: Y.D., M.S., and G.C. Supervision: X.H., Y.Z., J.Z., M.S., and G.C. Writing—original draft: X.H., D.L., Z.Z., and T.L. Writing—review and editing: Y.Z., J.Z., M.S., and G.C. Project administration: Y.Z., J.Z., M.S., and G.C. Funding acquisition: Y.Z. and M.S.

Competing interests:

The authors declare that they have no competing interests

Data and materials availability:

All data and code needed to evaluate and reproduce the results in the paper are present in the paper and/or the Supplementary Materials. scRNA-seq data from relapse P6 have been deposited at the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics, 2021) in National Genomics Data Center (Nucleic Acids Research, 2021), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences under accession numbers (GSA for Human: HRA013953 https://ngdc.cncb.ac.cn/gsa-human/submit/hra/subHRA021449/detail ) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa-human/browse/HRA013953 . The raw Olink sequencing data have been provided as Excel files S1 and S2.

Supplementary Materials

The PDF file includes:

Figs. S1 to S8

Tables S1 to S3

Supplementary Text

Legends for Excel files S1 to S9

Other Supplementary Material for this manuscript includes the following:

Excel files S1 to S9

REFERENCES

1 N. E. Gilhus , Myasthenia gravis . N. Engl. J. Med. 375 , 2570 – 2581 ( 2016 ). 28029925 10.1056/NEJMra1602678

2 N. Melzer , T. Ruck , P. Fuhr , R. Gold , R. Hohlfeld , A. Marx , A. Melms , B. Tackenberg , B. Schalke , C. Schneider-Gold , F. Zimprich , S. G. Meuth , H. Wiendl , Clinical features, pathogenesis, and treatment of myasthenia gravis: A supplement to the Guidelines of the German Neurological Society . J. Neurol. 263 , 1473 – 1494 ( 2016 ). 26886206 10.1007/s00415-016-8045-z PMC4971048

3 A. Uzawa , K. Utsugisawa , Biological therapies for myasthenia gravis . Expert Opin. Biol. Ther. 23 , 253 – 260 ( 2023 ). 36852951 10.1080/14712598.2023.2184257

4 M. C. Dalakas , Immunotherapy in myasthenia gravis in the era of biologics . Nat. Rev. Neurol. 15 , 113 – 124 ( 2019 ). 30573759 10.1038/s41582-018-0110-z

5 F. Piehl , A. Eriksson-Dufva , A. Budzianowska , A. Feresiadou , W. Hansson , M. A. Hietala , I. Håkansson , R. Johansson , D. Jons , I. Kmezic , C. Lindberg , J. Lindh , F. Lundin , I. Nygren , A. R. Punga , R. Press , K. Samuelsson , P. Sundström , O. Wickberg , S. Brauner , T. Frisell , Efficacy and safety of rituximab for new-onset generalized myasthenia gravis: The RINOMAX randomized clinical trial . JAMA Neurol. 79 , 1105 – 1112 ( 2022 ). 36121672 10.1001/jamaneurol.2022.2887 PMC9486640

6 R. J. Nowak , M. Benatar , E. Ciafaloni , J. F. Howard Jr. , M. I. Leite , K. Utsugisawa , J. Vissing , M. Rojavin , Q. Li , F. Tang , Y. Wu , N. Rampal , S. Cheng , MINT Investigators , A phase 3 trial of inebilizumab in generalized myasthenia gravis . N. Engl. J. Med. 392 , 2309 – 2320 ( 2025 ). 40202593 10.1056/NEJMoa2501561

7 F. Ramírez-Valle , J. C. Maranville , S. Roy , R. M. Plenge , Sequential immunotherapy: Towards cures for autoimmunity . Nat. Rev. Drug Discov. 23 , 501 – 524 ( 2024 ). 38839912 10.1038/s41573-024-00959-8

8 T. X. Chen , Y. T. Fan , B. W. Peng , Distinct mechanisms underlying therapeutic potentials of CD20 in neurological and neuromuscular disease . Pharmacol. Ther. 238 , 108180 ( 2022 ). 35367516 10.1016/j.pharmthera.2022.108180

9 M. Kalos , B. L. Levine , D. L. Porter , S. Katz , S. A. Grupp , A. Bagg , C. H. June , T cells with chimeric antigen receptors have potent antitumor effects and can establish memory in patients with advanced leukemia . Sci. Transl. Med. 3 , 95ra73 ( 2011 ). 10.1126/scitranslmed.3002842 PMC3393096 21832238

10 H. Huang , L. Yu , H. Weng , W. Zhang , Z. Wang , L. Wang , H. Huang , Advances in CAR T cell therapy for hematologic and solid malignancies: Latest updates from 2024 ESMO Congress . J. Hematol. Oncol. 17 , 120 ( 2024 ). 39639359 10.1186/s13045-024-01639-1 PMC11622581

11 G. Schett , A. Mackensen , D. Mougiakakos , CAR T-cell therapy in autoimmune diseases . Lancet 402 , 2034 – 2044 ( 2023 ). 37748491 10.1016/S0140-6736(23)01126-1

12 J. B. Chung , J. N. Brudno , D. Borie , J. N. Kochenderfer , Chimeric antigen receptor T cell therapy for autoimmune disease . Nat. Rev. Immunol. 24 , 830 – 845 ( 2024 ). 38831163 10.1038/s41577-024-01035-3 PMC12176013

13 A. Mackensen , F. Müller , D. Mougiakakos , S. Böltz , A. Wilhelm , M. Aigner , S. Völkl , D. Simon , A. Kleyer , L. Munoz , S. Kretschmann , S. Kharboutli , R. Gary , H. Reimann , W. Rösler , S. Uderhardt , H. Bang , M. Herrmann , A. B. Ekici , C. Buettner , K. M. Habenicht , T. H. Winkler , G. Krönke , G. Schett , Anti-CD19 CAR T cell therapy for refractory systemic lupus erythematosus . Nat. Med. 28 , 2124 – 2132 ( 2022 ). 36109639 10.1038/s41591-022-02017-5

14 C. Qin , D. S. Tian , L. Q. Zhou , K. Shang , L. Huang , M. H. Dong , Y. F. You , J. Xiao , Y. Xiong , W. Wang , H. Pang , J. J. Guo , S. B. Cai , D. Wang , C. R. Li , M. Zhang , B. T. Bu , W. Wang , Anti-BCMA CAR T-cell therapy CT103A in relapsed or refractory AQP4-IgG seropositive neuromyelitis optica spectrum disorders: Phase 1 trial interim results . Signal Transduct. Target. Ther. 8 , 5 ( 2023 ). 36596762 10.1038/s41392-022-01278-3 PMC9810610

15 V. Granit , M. Benatar , M. Kurtoglu , M. D. Miljković , N. Chahin , G. Sahagian , M. H. Feinberg , A. Slansky , T. Vu , C. M. Jewell , M. S. Singer , M. V. Kalayoglu , J. F. Howard Jr. , T. Mozaffar , MG-001 Study Team, Safety and clinical activity of autologous RNA chimeric antigen receptor T-cell therapy in myasthenia gravis (MG-001): A prospective, multicentre, open-label, non-randomised phase 1b/2a study . Lancet Neurol. 22 , 578 – 590 ( 2023 ). 37353278 10.1016/S1474-4422(23)00194-1 PMC10416207

16 A. Haghikia , T. Hegelmaier , D. Wolleschak , M. Böttcher , C. Desel , D. Borie , J. Motte , G. Schett , R. Schroers , R. Gold , D. Mougiakakos , Anti-CD19 CAR T cells for refractory myasthenia gravis . Lancet Neurol. 22 , 1104 – 1105 ( 2023 ). 37977704 10.1016/S1474-4422(23)00375-7

17 V. D. Dang , A. L. Stefanski , A. C. Lino , T. Dörner , B- and plasma cell subsets in autoimmune diseases: Translational perspectives . J. Invest. Dermatol. 142 , 811 – 822 ( 2022 ). 34955289 10.1016/j.jid.2021.05.038

18 G. Schett , G. Nagy , G. Krönke , D. Mielenz , B-cell depletion in autoimmune diseases . Ann. Rheum. Dis. 83 , 1409 – 1420 ( 2024 ). 38777374 10.1136/ard-2024-225727

19 D. B. Sanders , G. I. Wolfe , M. Benatar , A. Evoli , N. E. Gilhus , I. Illa , N. Kuntz , J. M. Massey , A. Melms , H. Murai , M. Nicolle , J. Palace , D. P. Richman , J. Verschuuren , P. Narayanaswami , International consensus guidance for management of myasthenia gravis: Executive summary . Neurology 87 , 419 – 425 ( 2016 ). 27358333 10.1212/WNL.0000000000002790 PMC4977114

20 C. Schneider-Gold , T. Hagenacker , N. Melzer , T. Ruck , Understanding the burden of refractory myasthenia gravis . Ther. Adv. Neurol. Disord. 12 , 1756286419832242 ( 2019 ). 30854027 10.1177/1756286419832242 PMC6399761

21 C. Zhao , M. Pu , D. Chen , J. Shi , Z. Li , J. Guo , G. Zhang , Effectiveness and safety of rituximab for refractory myasthenia gravis: A systematic review and single-arm meta-analysis . Front. Neurol. 12 , 736190 ( 2021 ). 34721267 10.3389/fneur.2021.736190 PMC8548630

22 Z. Zhang , Z. Wang , X. Du , X. Huang , Y. Zhang , Refractory generalized myasthenia gravis treated successfully with telitacicept: Two cases report . J. Neurol. 271 , 584 – 588 ( 2024 ). 37804335 10.1007/s00415-023-12036-y

23 J. Lin , Y. Li , M. Gui , B. Bu , Z. Li , Effectiveness and safety of telitacicept for refractory generalized myasthenia gravis: A retrospective study . Ther. Adv. Neurol. Disord. 17 , 17562864241251476 ( 2024 ). 38751755 10.1177/17562864241251476 PMC11095194

24 L. Remijn-Nelissen , M. R. Tannemaat , A. M. Ruiter , Y. J. M. Campman , J. J. G. M. Verschuuren , Efgartigimod in refractory autoimmune myasthenia gravis . Muscle Nerve 70 , 325 – 332 ( 2024 ). 38899431 10.1002/mus.28184

25 J. Moniz Dionísio , P. Ambrose , G. Burke , M. E. Farrugia , P. Garcia-Reitboeck , C. Hewamadduma , M. Hill , R. S. Howard , S. Jacob , D. Kullmann , M. I. Leite , J. Miller , A. Pinto , J. Pritchard , T. Riswick , S. Sathasivam , N. Thambirajah , S. Viegas , F. Norwood , J. Spillane , Efgartigimod efficacy and safety in refractory myasthenia gravis: UK’s first real-world experience . J. Neurol. Neurosurg. Psychiatry 96 , 322 – 328 ( 2025 ). 39798959 10.1136/jnnp-2024-334086

26 D.-S. Tian , C. Qin , M.-H. Dong , M. Heming , L.-Q. Zhou , W. Wang , S.-B. Cai , Y.-F. You , K. Shang , J. Xiao , D. Wang , C.-R. Li , M. Zhang , B.-T. Bu , G. Meyer Zu Hörste , W. Wang , B cell lineage reconstitution underlies CAR T cell therapeutic efficacy in patients with refractory myasthenia gravis . EMBO Mol. Med. 16 , 966 – 987 ( 2024 ). 38409527 10.1038/s44321-024-00043-z PMC11018773

27 A. D. Hughes , D. T. Teachey , C. Diorio , Riding the storm: Managing cytokine-related toxicities in CAR T cell therapy . Semin. Immunopathol. 46 , 5 ( 2024 ). 39012374 10.1007/s00281-024-01013-w PMC11252192

28 W. Wang , S. He , W. Zhang , H. Zhang , V. M. DeStefano , M. Wada , K. Pinz , G. Deener , D. Shah , N. Hagag , M. Wang , M. Hong , R. Zeng , T. Lan , Y. Ma , F. Li , Y. Liang , Z. Guo , C. Zou , M. Wang , L. Ding , Y. Ma , Y. Yuan , BCMA-CD19 compound CAR T cells for systemic lupus erythematosus: A phase 1 open-label clinical trial . Ann. Rheum. Dis. 83 , 1304 – 1314 ( 2024 ). 38777376 10.1136/ard-2024-225785

29 M. Shi , J. Wang , H. Huang , D. Liu , H. Cheng , X. Wang , W. Chen , Z. Yan , W. Sang , K. Qi , D. Li , F. Zhu , Z. Li , J. Qiao , Q. Wu , L. Zeng , X. Fei , W. Gu , Y. Miao , K. Xu , J. Zheng , J. Cao , Bispecific CAR T cell therapy targeting BCMA and CD19 in relapsed/refractory multiple myeloma: A phase I/II trial . Nat. Commun. 15 , 3371 ( 2024 ). 38643278 10.1038/s41467-024-47801-8 PMC11032309

30 T. Junt , T. Calzascia , E. Traggiai , A. N. da Costa , P. Gergely , G. Schett , T. Dörner , R. M. Siegel , Defining immune reset: Achieving sustained remission in autoimmune diseases . Nat. Rev. Immunol. 25 , 528 – 541 ( 2025 ). 40044810 10.1038/s41577-025-01141-w

31 C. Mauri , A. Bosma , Immune regulatory function of B cells . Annu. Rev. Immunol. 30 , 221 – 241 ( 2012 ). 22224776 10.1146/annurev-immunol-020711-074934

32 M.-H. Dong , Z.-C. Mei , L.-Q. Zhou , M. Heming , L.-L. Xu , Y.-X. Liu , X.-W. Pang , Y.-H. Chu , S.-B. Cai , H. Ye , K. Shang , J. Xiao , G. Meyer Zu Hörste , W. Wang , C. Qin , D.-S. Tian , Anti-BCMA CAR T cell therapy in relapsed/refractory chronic inflammatory demyelinating polyneuropathy . Med 6 , 100704 ( 2025 ). 40425008 10.1016/j.medj.2025.100704

33 C. Ulutekin , E. Galli , B. Schreiner , M. Khademi , I. Callegari , F. Piehl , N. Sanderson , D. Kirschenbaum , S. Mundt , M. Filippi , R. Furlan , T. Olsson , T. Derfuss , F. Ingelfinger , B. Becher , B cell depletion attenuates CD27 signaling of T helper cells in multiple sclerosis . Cell Rep. Med. 5 , 101351 ( 2024 ). 38134930 10.1016/j.xcrm.2023.101351 PMC10829729

34 E. Ricker , M. Manni , D. Flores-Castro , D. Jenkins , S. Gupta , J. Rivera-Correa , W. Meng , A. M. Rosenfeld , T. Pannellini , M. Bachu , Y. Chinenov , P. K. Sculco , R. Jessberger , E. T. L. Prak , A. B. Pernis , Altered function and differentiation of age-associated B cells contribute to the female bias in lupus mice . Nat. Commun. 12 , 4813 ( 2021 ). 34376664 10.1038/s41467-021-25102-8 PMC8355159

35 N. Claes , J. Fraussen , M. Vanheusden , N. Hellings , P. Stinissen , B. Van Wijmeersch , R. Hupperts , V. Somers , Age-associated B cells with proinflammatory characteristics are expanded in a proportion of multiple sclerosis patients . J. Immunol. 197 , 4576 – 4583 ( 2016 ). 27837111 10.4049/jimmunol.1502448

36 Y. Qin , M. L. Cai , H. Z. Jin , W. Huang , C. Zhu , A. Bozec , J. Huang , Z. Chen , Age-associated B cells contribute to the pathogenesis of rheumatoid arthritis by inducing activation of fibroblast-like synoviocytes via TNF-α-mediated ERK1/2 and JAK-STAT1 pathways . Ann. Rheum. Dis. 81 , 1504 – 1514 ( 2022 ). 35760450 10.1136/ard-2022-222605

37 I. C. Mouat , E. Goldberg , M. S. Horwitz , Age-associated B cells in autoimmune diseases . Cell. Mol. Life Sci. 79 , 402 ( 2022 ). 35798993 10.1007/s00018-022-04433-9 PMC9263041

38 K. M. Nickerson , S. Smita , K. B. Hoehn , A. D. Marinov , K. B. Thomas , J. T. Kos , Y. Yang , S. I. Bastacky , C. T. Watson , S. H. Kleinstein , M. J. Shlomchik , Age-associated B cells are heterogeneous and dynamic drivers of autoimmunity in mice . J. Exp. Med. 220 , e20221346 ( 2023 ). 36828389 10.1084/jem.20221346 PMC9997508

39 X. He , X. Zhang , Z. Yu , J. Meng , Y. Jiang , Y. Zhao , H. Lyu , X. Bai , X. Xiao , M. Zhao , Anti-FcRL5 CAR T exhibits anti-MM activity against EMM after progression of anti-BCMA and anti-GPRC5D CAR T: A case report . Blood Adv. 9 , 1159 – 1162 ( 2025 ). 39729519 10.1182/bloodadvances.2024014982 PMC11925504

40 X. Huang , H. Li , Z. Zhang , Z. Wang , X. Du , Y. Zhang , Macrophage migration inhibitory factor: A noval biomarker upregulates in myasthenia gravis and correlates with disease severity and relapse . Cytokine 175 , 156485 ( 2024 ). 38159470 10.1016/j.cyto.2023.156485

41 M. Su , X. Liu , Z. Wu , J. Song , X. Huan , H. Zhong , R. Zhao , C. Zhao , Y. Zhang , S. Luo , Nomogram for predicting pregnancy-related relapse of myasthenia gravis . Orphanet J. Rare Dis. 19 , 452 ( 2024 ). 39617942 10.1186/s13023-024-03466-6 PMC11610150

42 D. W. Lee , R. Gardner , D. L. Porter , C. U. Louis , N. Ahmed , M. Jensen , S. A. Grupp , C. L. Mackall , Current concepts in the diagnosis and management of cytokine release syndrome . Blood 124 , 188 – 195 ( 2014 ). 24876563 10.1182/blood-2014-05-552729 PMC4093680

43 D. W. Lee , B. D. Santomasso , F. L. Locke , A. Ghobadi , C. J. Turtle , J. N. Brudno , M. V. Maus , J. H. Park , E. Mead , S. Pavletic , W. Y. Go , L. Eldjerou , R. A. Gardner , N. Frey , K. J. Curran , K. Peggs , M. Pasquini , J. F. DiPersio , M. R. M. van den Brink , K. V. Komanduri , S. A. Grupp , S. S. Neelapu , ASTCT consensus grading for cytokine release syndrome and neurologic toxicity associated with immune effector cells . Biol. Blood Marrow Transplant. 25 , 625 – 638 ( 2019 ). 30592986 10.1016/j.bbmt.2018.12.758 PMC12180426

44 A. Petrera , C. von Toerne , J. Behler , C. Huth , B. Thorand , A. Hilgendorff , S. M. Hauck , Multiplatform approach for plasma proteomics: Complementarity of olink proximity extension assay technology to mass spectrometry-based protein profiling . J. Proteome Res. 20 , 751 – 762 ( 2021 ). 33253581 10.1021/acs.jproteome.0c00641

45 X. Ren , W. Wen , X. Fan , W. Hou , B. Su , P. Cai , J. Li , Y. Liu , F. Tang , F. Zhang , Y. Yang , J. He , W. Ma , J. He , P. Wang , Q. Cao , F. Chen , Y. Chen , X. Cheng , G. Deng , X. Ding , W. Feng , R. Gan , C. Guo , W. Guo , S. He , C. Jiang , J. Liang , Y. M. Li , J. Lin , Y. Ling , H. Liu , J. Liu , N. Liu , S. Q. Liu , M. Luo , Q. Ma , Q. Song , W. Sun , G. Wang , F. Wang , Y. Wang , X. Wen , Q. Wu , G. Xu , X. Xie , X. Xiong , X. Xing , X. H. Xu , C. Yin , D. Yu , K. Yu , J. Yuan , B. Zhang , P. Zhang , T. Zhang , J. Zhao , P. Zhao , J. Zhou , W. Zhou , S. Zhong , X. Zhong , S. Zhang , L. Zhu , P. Zhu , B. Zou , J. Zou , Z. Zuo , F. Bai , X. Huang , P. Zhou , Q. Jiang , Z. Huang , J. X. Bei , L. Wei , X. W. Bian , X. Liu , T. Cheng , X. Li , P. Zhao , F. S. Wang , H. Wang , B. Su , Z. Zhang , K. Qu , X. Wang , J. Chen , R. Jin , Z. Zhang , COVID-19 immune features revealed by a large-scale single-cell transcriptome atlas . Cell 184 , 1895 – 1913.e19 ( 2021 ). 33657410 10.1016/j.cell.2021.01.053 PMC7857060