CD28 costimulation is a central tenet of T cell activation and tolerance, yet the mechanisms by which CD28 functions are still surprisingly poorly understood and even controversial. Our data and others indicate clear differences in qualitative outcome of CD28 signaling between mouse and human CD4+ T cells, with human CD28 appearing to mediate enhanced downstream signaling and associated functional outcomes compared to murine CD28. Furthermore, we have recently demonstrated that human Th17 differentiation is strongly inhibited by CD28 costimulation, while mouse Th17 cell differentiation is unaffected. These species differences in CD28 responses create a problem in interpreting and studying the role of costimulation during Th17 development in vivo in our mouse models of autoimmunity. In a wider context of therapies that target T cells, such as cancer immunotherapy, these inter-species differences in CD28 signaling create concern when extrapolating mouse model findings to clinical trials. This was most clearly demonstrated by a disastrous CD28-superagonist clinical trial which induced near-lethal cytokine storm in healthy Phase I volunteers, despite mouse models indicating expansion of regulatory T cells with minimal cytokine induction. Mouse models of human disease provide critical and sophisticated in vivo mechanistic platforms and pre-clinical drug efficacy testing that are generally not possible with human samples. There is therefore a critical need to optimize these models so that they better reflect human pathways. The C-terminal proline-rich domain of CD28 is known to be one of two critical signaling motifs in the relatively short CD28 cytoplasmic tail. It was recently demonstrated that mouse and human CD28 differ in this domain by a single amino acid, PYAP in mouse, PYPP in human, and substituting A->P in mouse CD28 enhanced recruitment of signaling adaptors and downstream cytokine production, recapitulating the response of human CD28 ligation. We therefore propose to generate a mouse that bears `humanized' CD28PYPP using CRISPR/Cas9 to generate a single point mutation in the proline domain. Importantly, the rest of the CD28 molecule will be murine, meaning that physiological levels of expression and engagement of costimulatly ligands B7-1/B7-2 will occur. We hypothesize that this CD28PYPP mouse strain will recapitulate our findings in human Th17 cells by demonstrating CD28-mediated suppression of Th17 responses. We will test this hypothesis and validate our model using ex vivo analysis of peripheral T cell development and CD28 signaling, coupled with initial in vitro and in vivo Th17 differentiation assays. This proposal will thus provide a novel validated tool with which to interrogate the role of CD28 costimulation in Th17 differentiation under various in vivo circumstances (inflammatory and homeostatic). This tool will not only enable our own Th17-focused studies, the humanized CD28PYPP mouse has the potential to significantly advance all areas of T cell immunology research, with applications to CAR-T cell immunotherapy, transplantation tolerance and infectious disease.
CD28 is an important molecule in T cell activation, but we recently demonstrated that human Th17 cells are inhibited by CD28, while it does not play the same role in mouse cells. Similarly, other investigators have found differences in activation of signaling molecules downstream of CD28 in mouse versus humans. We therefore propose to create a mouse strain that bears a mutation that mimics human CD28 activation events, in order to study the critical functions of CD28 in a translationally-relevant model.
