Autoimmune polyglandular syndrome type 3 variant (APS3v) refers to an autoimmune condition in which both type 1 diabetes (T1D) and autoimmune thyroiditis (AITD) develop in the same individual. HLA-DR3 confers the strongest susceptibility to APS3v. Previously we reported a unique amino acid signature pocket that predisposes to APS3v. We found that this pocket is flexible and can trigger APS3v by presenting both thyroid (Tg.1571, TPO.758) and islet (GAD.492) peptides to induce autoimmune response. We hypothesized that blocking the specific APS3v-HLA-DR3 pocket from presenting thyroid/islet antigens can block the autoimmune response in APS3v. To test this hypothesis we performed a virtual screen of small molecules blocking APS3v-HLA-DR3, and identified 11 small molecules hits that were predicted to block APS3v-HLA-DR3. Using the baculovirus-produced recombinant APS3v-HLA-DR3 protein we tested the 11 small molecules in an binding assay. We validated 4 small molecule hits, S9, S5, S53 and S15, that could block the APS3v-HLA-DR3 pocket . We then developed a novel humanized APS3v mouse model induced by co-immunizing a peptide mix of Tg.1571, TPO.758 and GAD.492. The immunized mice developed strong T-cell and antibody responses to the thyroid/islet peptides, as well as mouse thyroglobulin. In addition, the mice showed significantly lower free T4 levels compared to controls. Using the APS3v mouse model, we showed that one of the 4 small molecules, Cepharanthine (S53), blocked T-cell activation by thyroid/islet peptides and . These findings suggested Cepharanthine may have a therapeutic potential in APS3v patients carrying the specific APS3v-HLA-DR3 pocket.