Modulation of Mesenchymal Stem Cell Immunogenicity through Forced Expression of Human Cytomegalovirus Proteins

Mesenchymal stem cells (MSC) are promising candidates for cell replacement therapy since they have the ability to differentiate, under appropriate conditions, into a broad range of specialized cell types. Furthermore, MSC have low inherent immunogenicity, immunomodulatory properties, and preferentially home to/engraft damaged tissues. However, they are not invisible to the immune system, and upon allogeneic transplantation, MSC can elicit an immune response, that results in activation of the recipient’s cytotoxic T lymphocytes (CTL) and Natural Killer (NK) cells and hence rejection of the transplanted cells. Human cytomegalovirus (HCMV), a virus ubiquitously present in humans, has developed several strategies to evade CTL and NK cell recognition. HCMV avoids CTL attack by producing proteins coded for by the unique short region (US) of the genome that downregulate MHC-I surface expression. Moreover, expression of HCMV UL18 protein, which is an MHC-I homolog or decoy, renders CMV-infected cells resistant to NK lysis. Thus, we hypothesized that simultaneous expression of HCMV US protein US2, US3, US6, or US11 with UL18 would render MSC less susceptible to CTL and NK cell recognition and killing. To this end, we started by transducing MSC with the retroviral vectors encoding: US2; US3; US6; or US11, and tested the effect of each of these proteins on expression of MHC-I and CD59, a complement membrane attack complex inhibitor whose expression has been shown to prevent destruction of transplanted cells. MSC expressing US3, US2, US11 and US6 showed substantial MHC-I downregulation of 69%, 74%, 78 % and 95% respectively, when compared to untransduced MSC. Moreover, MSC expressing either US6 or US2 proteins displayed a significant up-regulation of CD59, with US6 causing a 2.5 and US2 a 3.9 fold increase, when compared to untransduced MSC (Mean Fluorescence Intensity: 869.8; 1331.19; 344.99, respectively). Each of the US transduced MSC were then tested for their capability to activate MHC-I mismatched CTL proliferation in both an allogeneic and a xenogeneic setting, using MSC transduced with an empty retrovirus as a control. MSC expressing US2 or US6 reduced CTL proliferation by the same degree, whether CTL were mismatched human- or sheep-derived. Using MSC-US2 we obtained a reduction in CTL proliferation by 18% and 23% in human-or sheep-CTL, respectively, while using MSC-US6 resulted in a reduction in human and sheep CTL proliferation of 60% and 53%. Expression of US11 on MSC was more efficient at abrogating the sheep CTL proliferation than that of mismatched human CTL (reduction in proliferation by 53% vs 31%, respectively). Finally, expression of US3 on MSC only reduced proliferation of human CTL by 21.5%, but had no effect on sheep CTL. In conclusion, engineering MSC to over-express US6 seems to be the most effective way to enable MSC to evade the host immune system, resulting in efficient downregulation of MHC-I surface expression, significant reduction in CTL proliferation, and a reduced formation of the membrane attack complex at the end of the complement cascade, which will reduce destruction of MSC by this system. In vivo studies using a large animal sheep model are underway to demonstrate whether the results obtained in vitro using US6 HCMV protein expression will translate into improved engraftment in vivo.