A marvel of evolution, the adaptive immune system has the capacity to respond to almost any foreign antigen in a highly specific manner. Antibodies, Y-shaped glycoproteins containing both diverse variable regions responsible for antigen binding and constant regions responsible for effector function, are a key part of this capacity. However, this vast diversity comes with several drawbacks, one of which is the fact that the immune system can deleteriously respond to self-antigens. The focus of this thesis is to characterize the role of class-switching (the changing of antibody constant regions) in the pathogenesis of autoimmune disease, and in particular to trace the lineage of antigen-specific autoreactive B cells by analyzing clonal relationships between antibodies of different constant regions. Analyzing such lineages has the potential to shed light on mechanisms of autoantibody-mediated disease pathogenesis, leading to better understanding of autoimmunity and better therapeutics. The work presented in this thesis focuses on pemphigus vulgaris, or PV, a model antibody-mediated autoimmune disease characterized by a response to the cell adhesion protein desmoglein (Dsg) 3, which holds keratinocytes together in the epidermis. An enigmatic feature of this disease is the predominance of antibodies from the IgG4 subclass during active disease, which ordinarily appears to have few effector functions and may serve as a “brake” on the immune system in the setting of continuous stimulation by antigen. PV patients also display autoantibodies of the IgG1 subclass during disease and remission, but the relationship between IgG1 and IgG4 in the disease in unclear. Because the majority of cases of PV also harbor anti-Dsg antibodies of the IgA1 and IgA2 subclasses, we sought to determine the relationships between autoantibodies belonging to each of these subclasses. First, we address whether the same anti-Dsg variable region, grafted onto either IgG1 or IgG4 constant regions, can show differing affinity or pathogenicity, in order to determine whether antibody subclass is directly modulating pathogenic effect (chapter 2). Finding that the subclass has very little effect on antibody affinity, pathogenicity, or epitope preference, we then sought to determine whether B cells expressing autoantibodies of different subclasses share lineages, indicating common pathways of development (chapter 3). Using a combination of antigen-specific antibody cloning through phage display, and next-generation sequencing of subclass specific repertoires in a panel of PV patients, we managed to trace 80 lineages of anti-Dsg B cells across all four subclasses tested. In particular, we found that anti-Dsg IgG4 B cells, which are believed to be central to disease pathogenesis, tended to not share lineages with other subclasses, and in in general do not appear to share a precursor-product relationship with anti-Dsg IgG1 B cells. We have also found that anti-Dsg IgA1 and IgA2 were tightly related and often arose directly from IgG precursors. These findings are key to understanding the role of class-switching in the pathogenesis of PV, and may shed light on the class-switch mechanisms driving other autoimmune diseases and states of chronic antigen stimulation.