Abstract V(D)J recombination generates mature B cells that express huge repertoires of primary antibodies as diverse immunoglobulin heavy (IgH) and light chains (IgL) of their B cell antigen receptors (BCRs). Cognate antigen binding to BCR variable region domains activates B cells into the germinal center (GC) reaction in which somatic hypermutation (SHM) modifies primary variable region-encoding sequences, with subsequent selection for mutations that improve antigen-binding affinity, ultimately leading to antibody affinity maturation. Based on these principles, we developed a humanized mouse model approach to diversify an anti-PD1 therapeutic antibody and allow isolation of variants with novel properties. In this approach, component Ig gene segments of the anti-PD1 antibody underwent de novo V(D)J recombination to diversify the anti-PD1 antibody in the primary antibody repertoire in the mouse models. Immunization of these mouse models further modifies the anti-PD1 antibodies through SHM. Known anti-PD1 antibodies block interaction of PD1 with its ligands to alleviate PD1-mediated T cell suppression, thereby boosting anti-tumor T cell responses. By diversifying one such anti-PD1 antibody, we derived many new anti-PD1 antibodies, including anti-PD1 antibodies with the opposite activity of enhancing PD1/ligand interaction. Such antibodies theoretically might suppress deleterious T cell activities in autoimmune diseases. The approach we describe should be generally applicable for diversifying other therapeutic antibodies. Significance Statement The immune system is capable of producing enormous varieties of antibodies to counter pathogenic infections. In this study, we devised a method to harness the power of the immune system to produce potentially therapeutic antibodies. As a test, we used this method to generate antibodies against the human PD1 molecule, a negative regulator of T cell activity. Available anti-PD1 therapeutic antibodies inhibit the function of PD1, thereby boosting T cell activity against tumor cells. Through our approach, we diversified one such anti-PD1 antibody in our mouse models. From this work, we obtained multiple new anti-PD1 antibodies, some of which can stimulate, rather than inhibit, PD1 function in vitro . Such PD1 stimulatory antibodies could potentially be used to suppress pathogenic T cell activities in autoimmune diseases. This method could be used to diversify the functions of other therapeutic antibodies.