Undergraduate The University of Texas at Austin, Dell Medical School Austin, Texas
Rationale: Although autoimmunity has been implicated as an etiology in a subset of refractory epilepsies, exact mechanisms of how the immune system drives seizures are still being discovered. Current clinical evaluation of potential autoimmune epilepsy is focused on assessment of known neuronal antibodies. More auto-antibodies are being characterized yearly, but many suspected autoimmune patients remain sero-negative. As a result, it is challenging to treat these patients since they do not respond well to antiepileptic drugs, devices, or surgeries. Thus, it is critical to better understand all aspects of the immune response in seizures, including cell and cytokine immune components in order to develop more specific therapies for these patients. Methods: In this study, we recruited patients with (i) suspected autoimmune epilepsy who were seronegative for autoimmune epilepsy antibody panel (n=5), (ii) non-autoimmune epilepsy (n=5) (iii) known autoimmune neurological conditions (n=17), such as multiple sclerosis and neuromyelitis optica, and (iv) healthy controls (n=7). Patient sera was evaluated for a total of 65 known inflammatory and anti-inflammatory T-cell, B-cell, and macrophage cytokines. Unsupervised machine learning techniques, including hierarchical clustering and principal component analysis (PCA) were used to group patients based on cytokine expression. In a subset of patients, we also performed a gene expression study using 3’ tag-sequencing, to compare immune gene repertoires between a patient with suspected autoimmune epilepsy and a healthy familial control. Results: PCA on the cytokine multiplexing data revealed a distinct group of patients with suspected autoimmune epilepsy clustering separately from patients with non-autoimmune epilepsy, controls, and patients with treated autoimmune neurological conditions. From the panel of 65 inflammatory cytokines, GM-CSF, a cytokine known to affect blood brain barrier trafficking in other neuro-autoimmune diseases, was significantly increased in patients with untreated suspected-autoimmune epilepsy patients compared to other evaluated patient groups. Other cytokines, including IL-10, IL-13, IL-15, IL-20, and IL-23 were also elevated in patients with suspected-autoimmune epilepsy patients compared to other patient groups. Gene sequencing in one of the suspected-autoimmune epilepsy patients with noted hypogammaglobulinemia and low complement levels revealed there was elevated expression of T-cell related genes relative to B-cell related genes. Treatment of this sero-negative patient with steroids and intravenous immunoglobulins resulted in improvement in seizure control as previously reported for patients with sero-positive autoimmune epilepsy. Conclusions: The results of our study suggest that a subset of epilepsy patients who are seronegative for known auto-antibodies may experienceneuroinflammation due to cell-mediated immune responses. Identification of patients with distinct autoimmune epilepsy subtypes could accelerate the development of specific neuroimmune treatments and improve health outcomes in patients with refractory seizures due to autoimmunity. Funding: Please list any funding that was received in support of this abstract.: This work was supported by the University of Texas Dell Medical School Start Up funding to E. Melamed.