15 A novel CD28 humanized mouse model for efficacy assessment of CD28-targeting therapies

Background Immuno-intervention through targeting of activating and inhibitory immune checkpoints (ICP), has shown promising results in the clinic over the last years. To facilitate these researches, mouse models expressing humanized ICP instead of their mouse counterparts were developed. Herein, we describe a novel CD28 humanized mouse model (hCD28 model), designed to test compounds targeting human CD28 (hCD28). Methods Human and mouse CD28 (mCD28) have different signalling responses, with hCD28 being known for inducing higher levels of pro-inflammatory cytokines upon stimulation with ligands/superagonists. This can be explained by the expression of CD28i, a hCD28 amplifier isoform which is not found in mouse. Additionally, evidences suggested that the different signalling between human and mCD28 relies on one amino acid change in the intracellular domain (ICD).1 Because the hCD28 model was developed to assess hCD28-targeting therapeutics, we decided to keep the expression of both canonical and CD28i isoforms to avoid undermining the biological effects of the testing antibodies. Although keeping the human ICD could favour the evaluation of cytokine production and therefore the safety of the test therapeutics, we decided to keep the mouse ICD to enable a proper interaction of CD28 with its signalling partners, allowing a physiological stimulation of CD28 in efficacy studies. Results hCD28 mice express hCD28 on T cells and the frequency of CD3 T cells is comparable in both WT and hCD28 mice. Stimulation of hCD28 mice-isolated T cells with hCD28 ligands and agonist antibodies resulted in T cell proliferation and cytokine production, suggesting that hCD28 is functional in mouse cells. MC38 uptake rate and kinetic of growth were comparable in WT and hCD28 mice, suggesting no major defect in the immune response in the hCD28 mice. Importantly, splenocytes and tumor draining lymph nodes cells isolated from tumor-bearing hCD28 mice showed higher production of IL-2 and IFN-gamma upon in vitro re-challenged with MC38 when compared to WT cells. Since the frequency of CD3 cells (Treg, CD4+ and CD8+) is comparable to WT mice, this could be explained by the expression of the amplifier CD28i isoform, which is absent in WT mice. Conclusions The hCD28 model described here supports the efficacy assessment of hCD28-targeting biologics, enabling PK/PD studies as hCD28 expression levels and pattern are physiological. However, after careful consideration of the CD28 biology, we decided to keep the mouse ICD, although it triggers lower pro-inflammatory cytokine production than CD28 human ICD. As such, this model is not suitable for toxicology/safety studies. Reference Porciello N, Grazioli P, Campese AF, et al. A non-conserved amino acid variant regulates differential signalling between human and mouse CD28. Nat Commun 2018; 9:1–16.