Glutamic acid decarboxylase 65 (GAD65) mediated cerebellar ataxia (CA) is a distinct autoimmune disease, characterized by high titers of GAD65 antibodies in the presence of gait-predominant and chronically progressive cerebellar syndrome (1). Several in vitro and in vivo studies support the pathogenic role of GAD65, demonstrating that the passive transfer of GAD65 antibodies produces pathogenic effects in rats, while the absorption of these antibodies leads to a full reversal of neuronal functionality (2, 3). Likewise, clinical studies suggest that high titers of GAD65 lead to poor outcomes in CA while reduction of antibody titers may be associated with clinical improvement (4). This suggests that antibody depletion may be effective for the treatment of GAD65 CA. There is no standard treatment for GAD65 CA, and despite the use of B-cell directed therapies, long-term outcomes in GAD65 CA are unfavorable in two-thirds of patients (5). An increasing number of reports suggest CAR-T cells may be a safe and effective therapy for treatment-refractory autoimmune diseases, including various immune-mediated neurological disorders. However, their safety and efficacy have been studied in a very limited number of immune-mediated neurological disorders. Two case reports of standard treatment-refractory GAD65-associated stiff person syndrome treated with CD19 CAR-T cells showed good response to treatment (6, 7). There are no reports of CD19 CAR-T cell therapy use in GAD65 CA. Herein, we describe a patient with GAD65 CA treated with CD19 CAR T cells.

Case description with prospective follow-up over a 9-month period using pre-established follow-up protocols (Supplementary Document S1, Supplementary Table S2). Figures were generated using Python (version 3.11.4) and Adobe Photoshop (version 26.11, 2025).

In this report, we present the first case of treatment-refractory GAD65 antibody-mediated CA successfully treated with CD19 CAR-T cells. Over a follow-up period of 9 months, we observed a favorable toxicity profile and no treatment-related serious adverse events. Additionally, we documented a rapid and sustained serological response, with a 95% reduction in serum GAD65 antibody levels from baseline and seroconversion to negative in the CSF, along with no evidence of disease progression as indicated by various clinical and radiological parameters.

Although the antigenic target for GAD65 antibodies is intracellular, supporting T cell-mediated pathogenesis, previous studies on GAD65 neurological syndromes suggest that B cells are pivotal early in the GAD65-mediated disease course to maintain T cell autoreactivity (8, 9). Likewise, in vitro evidence suggests that T cell autoreactivity is upregulated by B cells in germinal centers within the CNS (10). Eliminating B cells within tissue-resident lymphoid follicles may therefore halt T-cell-mediated tissue injury. This hypothesis is further corroborated in reports of successful treatment of T cell-mediated diseases with CD19 CAR-T cells (11, 12). In the present context, the effectiveness of CAR T cells in treating CNS neurological disorders is attributed to their ability to penetrate the CNS (13) and modify the interaction between B and T cells through B cell depletion. This is further supported by the limited effectiveness of standard anti-CD20 monoclonal antibodies in treating GAD65-mediated neurological disorders, as these antibodies have difficulty penetrating the CNS.

The timing of treatment is an important consideration in GAD65 antibody-mediated neurological disorders. The limited therapeutic response, even to CAR-T cells in some earlier reports (6, 7) suggests that symptoms may be reversible due to inhibition of neuronal function in the early course of the disease, but a time-dependent irreversible loss of GABAergic or Purkinje neurons follows (8, 9). Likewise, previous studies have shown that treatment for CNS T cell–mediated diseases, including CA, is effective only when administered early and in patients with retained ambulation (14, 15). The early therapeutic intervention is also supported by a recent expert opinion statement suggesting that early election of CD19 CAR-T may have beneficial effects across the entire spectrum of immune-mediated neurological disorders (16), including Diacylglycerol lipase alpha (DAGLA) antibody-associated ataxia-encephalitis, a novel autoimmune disorder that is most likely T cell-mediated (17). After thorough discussion with the patient, and encouraged by recent CAR-T studies in autoimmune diseases demonstrating limited adverse events (18) we elected to administer CD19-directed CAR-T cells early in the course of CA.

Although we did not observe acute or delayed immune effector cell-associated neurotoxicity syndrome (ICANS) in our case, available data from phase I trials suggest ICANS may still develop in a small fraction of patients treated with CAR-T cell therapy (19–21). In contrast to the hematological population, algorithms to identify patients at high-risk for ICANS in the autoimmune population are lacking, complicating patient selection and potentially compromising safety. Future trials will answer these important questions. Likewise, histopathological studies of ICANS cases are essential for understanding potential mechanisms of CAR-T cell-related toxicities in oncologic as well as autoimmune populations (22).

A potential limitation of the study is that we did not perform CSF measurements early after CD19 CAR-T infusion. This limited our ability to evaluate the capacity of intrathecal CD19 CAR-T cell expansion.

Likewise, in contrast to previous reports of CD19 CAR-T cells showing a dramatic improvement of severe neurological burden in stiff person syndrome and autoimmune encephalitis (7, 12, 17), our patient had minor neurological disability resulting in minor improvement. However, this minor improvement is likely clinically significant, as it surpasses the minimal clinically important difference (MCID) of the SARA score by a factor of two, as reported in previous ataxia studies (23). Lastly, the patient reported a subjective improvement in vertigo and balance after CD19 CAR-T cell therapy, while previous therapies with two second-line medications have failed to control the condition.

In summary, the favorable serological and clinical response in our patient, along with previous reports of efficacy, support further study of CD19 CAR-T cells for GAD65-related neurological syndromes in carefully selected patients. Both short and long-term safety, patient outcomes and toxicity profiles must be studied in large-scale trials.