A Functional Human-on-a-Chip Autoimmune Disease Model of Myasthenia Gravis for Development of Therapeutics

Myasthenia gravis (MG) is a chronic and progressive neuromuscular disease where autoantibodies target essential proteins such as the nicotinic acetylcholine receptor (nAChR) at the neuromuscular junction (NMJ) causing muscle fatigue and weakness. Autoantibodies directed against nAChRs are proposed to work by three main pathological mechanisms of receptor disruption: blocking, receptor internalization, and downregulation. Current <i>in vivo</i> models using experimental autoimmune animal models fail to recapitulate the disease pathology and are limited in clinical translatability due to disproportionate disease severity and high animal death rates. The development of a highly sensitive antibody assay that mimics human disease pathology is desirable for clinical advancement and therapeutic development. To address this lack of relevant models, an NMJ platform derived from human iPSC differentiated motoneurons and primary skeletal muscle was used to investigate the ability of an anti-nAChR antibody to induce clinically relevant MG pathology in the serum-free, spatially organized, functionally mature NMJ platform. Treatment of the NMJ model with the anti-nAChR antibody revealed decreasing NMJ stability as measured by the number of NMJs before and after the synchrony stimulation protocol. This decrease in NMJ stability was dose-dependent over a concentration range of 0.01-20 μg/mL. Immunocytochemical (ICC) analysis was used to distinguish between pathological mechanisms of antibody-mediated receptor disruption including blocking, receptor internalization and downregulation. Antibody treatment also activated the complement cascade as indicated by complement protein 3 deposition near the nAChRs. Additionally, complement cascade activation significantly altered other readouts of NMJ function including the NMJ fidelity parameter as measured by the number of muscle contractions missed in response to increasing motoneuron stimulation frequencies. This synchrony readout mimics the clinical phenotype of neurological blocking that results in failure of muscle contractions despite motoneuron stimulations. Taken together, these data indicate the establishment of a relevant disease model of MG that mimics reduction of functional nAChRs at the NMJ, decreased NMJ stability, complement activation and blocking of neuromuscular transmission. This system is the first functional human <i>in vitro</i> model of MG to be used to simulate three potential disease mechanisms as well as to establish a preclinical platform for evaluation of disease modifying treatments (etiology).