Exploring protein orthogonality in immune space: a case study with AAV and Cas9 orthologs

A major hurdle in protein-based therapeutics is the interaction with the adaptive immune system, which can lead to neutralization by circulating antibodies and clearance of treated cells by cytotoxic T-lymphocytes. One method of circumventing these issues is to use human or humanized proteins which avoid the immune response by self-recognition. However, this approach limits potential protein therapeutics to those of human origin, excluding many exciting effectors and delivery vehicles such as CRISPR-Cas9 and adeno-associated viruses (AAVs). To address this issue, we propose here the sequential use of orthologous proteins whose function is constrained by natural selection, but whose structure is subject to diversification by genetic drift. This would, in principle, allow for repeated treatments by immune orthogonal orthologs without reduced efficacy due to lack of immune cross-reactivity among the proteins. To explore and validate this concept we chose 91 Type II CRISPR-Cas9 orthologs and 167 AAV capsid protein orthologs, and developed a pipeline to compare total sequence similarity as well as predicted binding to class I and class II Major Histocompatibility Complex (MHC) proteins. Interestingly, MHC binding predictions revealed wide diversity among the set of Cas9 orthologs, with 83% of pairs predicted to have non cross-reacting immune responses, while no global immune orthogonality among AAV serotypes was observed. To confirm these findings we selected two Cas9 orthologs, from S. pyogenes and S. aureus, predicted to be orthogonal in immune space, and delivered them into mice via multiple AAV serotypes. We observed cross-reacting antibodies against AAV but not Cas9 orthologs in sera from immunized mice, validating the computationally predicted immune orthogonality among these proteins. Moving forward, we anticipate this framework can be applied to prescribe sequential regimens of immune orthogonal protein therapeutics to circumvent pre-existing or induced immunity, and eventually, to rationally engineer immune orthogonality among protein orthologs.