Autologous transplantation of EPCs does not involve pre-transplant crossmatch typing or post-transplant immunological rejection, indicating a promising treatment method for various diseases such as myocardial ischemia and hemophilia A in the clinic. Because of the poor cell number of directly isolated EPCs from peripheral blood, ex vivo expansion and differentiation of EPCs is particularly important. We previously reported a two-step culture system for the large-scale generation of human EPCs derived from CD34+ cells from human cord blood ex vivo [26]. However, considering the therapeutic purpose, CD34+ cells from human mobilized PB are safe and more practically convenient for autologous transplantation in humans. In this study, the final numbers of proliferated and differentiated EPCs from mobilized PB were higher than from human cord blood (Fig. 1d), due to more CD34+ cells being isolated from the mobilized PB compared with the cord blood. Notably, over 2 × 108 EPCs/ECs were generated from 15–20 mL mobilized PB (about 1.51 × 106 ± 3.39 × 105 CD34+ cells), up to 350-fold higher than that reported previously [16, 32, 33]. Nonhuman primates are considered as the optimal animal model for evaluating the safety and function of cell preparations for preclinical studies. Unfortunately, strong auto-immunity rejection would be induced with the direct use of human EPCs in nonhuman primates; thus, we applied our culture system to generate EPCs derived from nonhuman primate mobilized PB CD34+ cells for evaluation of preclinical studies and potential clinical application in the future.

The endothelial cells derived from rhesus monkeys share a cell surface phenotype similar to humans and reportedly can be cultured ex vivo [34]. Our novel culture system technique from previous protocols isolated, purified, and cultured EPCs from normal PB mononuclear cells. We chose mobilized PB and enriched CD34+ cells as the initial component for expansion and induction of EPCs, as well as using nonhuman primate PB CD34+ cells. It has been reported that EPCs and hematopoietic stem cells originate from a common precursor, the hemangioblast, and share a number of cell-surface markers [35]. Eunju et al. suggested that culture of the CD34 fraction is more efficient for EPC expansion than that of mononuclear cells [36]. In addition, we creatively combined hematopoietic stem cell expansion and endothelial differentiation via the two-step culture system, significantly increasing the yield of EPCs. All the cytokines and growth factors included in the culture system were biologically endogenous factors and clinical-grade reagents that have been proved to carry no risk for cell toxicity or tumor stimulation. Therefore, this ex vivo culture system could provide sufficient autologous EPCs for transplantation treatment in the clinic.

Before entering into clinical trials, new drugs or cell products need a series of proof-of-concept studies, in vivo kinetics, and safety assessment in animal models. Most studies use an immunodeficient animal host to evaluate human cells, but this approach is not very accurate since human cells may behave differently in an animal host environment. Analogous animal models would be more suitable for efficacy and toxicity research. Nonhuman primates should provide a more accurate model for therapeutic evaluation than rodents due to a greater similarity to humans, particularly regarding life-span, hematopoietic system development, and physiological homeostasis. Therefore, we originally conducted this novel autologous transplantation of EPCs in a nonhuman primate model as an in vivo safety and efficacy evaluation. To our knowledge, this is the first report of large-scale generation of primate EPCs ex vivo, achieving an over 7000-fold increase in initial EPCs. In addition, the produced primate EPCs were identified to be similar to human EPCs in terms of the cell morphology and function as assayed by photomicrograph, lectin expression, Dil-Ac-LDL uptake, and NO release. Moreover, based on the CD31/CD144 expression and eNOS/NO levels of the produced cells being dramatically increased after 12 days of culture, it could be possible that EPCs started to differentiate into ECs from day 12; the final produced cells were still a mixture of EPCs and ECs.

Before autologous transplantation in the clinic, it is very helpful to develop a primate model of hepatic sinusoidal endothelium injury. A primate model was constructed with intraperitoneal injection of MCT, a toxic agent that can disrupt the sinusoidal endothelial barrier and stimulate the incorporation of transplanted cells into liver parenchyma. Crucially, endothelial cells play central roles in liver development, organization, repair, and function. Moreover, hepatocytes and liver sinusoidal endothelial cells are the major source of FVIII, which is deficient in hemophilia A. Follenzi and colleagues reported that transplanting healthy liver sinusoidal endothelial cells into hemophilia A mice led to the restoration of plasma factor VIII activity and corrected their bleeding phenotype [8]. If this culture system is combined with ex vivo genome editing technology such as CRISP/Cas-9, sufficient healthy EPCs could be obtained from a small volume of the patient’s own mobilized PB. This would minimize the suffering of patients and possibly provide breakthroughs for the complete cure of hemophilia A.

By using the nonhuman primate model to verify the engraftment capacity of generated EPCs/ECs, we applied the FITC-microbeads to label nonhuman primate cells. In our previous testing, these microbeads can tightly bind to the cell surface and prevent fluorescence quenching. The labeling efficiency can reach up to 70% in CD34+ cells and 100% in adherent cultured cells. With this unique technique following GFP transfection, nonhuman primate cells were marked by green florescence on the surface as well as in the nucleus. Moreover, to enlarge the signal of the GFP-positive cells, we used an anti-GFP primary antibody followed by a cy3-labeled secondary antibody (red). The double-positive stained cells greatly enhances the assay specificity and identifies engrafted cells, which can be distinguished from cells that have strong autofluorescence such as blood and hepatic cells. When analyzing the double-positive cells on day 7 and day 14 post-transplantation, the results were consistent with a previous study on mice by our group [10], showing that the double-positive cells were present on both day 7 and day 14. However, in the mouse studies, the human EPCs/ECs first adhered onto the vascular inner in the livers of NOD/SCID mice and then gradually migrated to the injured sites. In this present study, the engrafted nonhuman primate cells evenly appeared in the liver tissues from day 7. There could be two explanations: 1) the migration of autologous transplantation in primates is faster than xenotransplantation in mice; and 2) we used more differentiated or matured EPCs/ECs on day 36 of culture, which may result in higher capacities [33, 37] of adherence and more repair of the sites of injury induced by MCT in the hepatic sinusoidal endothelium.

We did not expect to obtain a large number of engrafted cells using this injury model. As we reported, 2–3% of the cells in the monkey liver tissue were transplanted endothelial cells since the transplanted cells (2 × 108) are very limited compared with the total number of hepatic cells. In fact, the hemophilia A patient needs a source that can continually release FVIII rather than for the reconstruction of endothelium. The purpose of this study was to verify that the produced endothelial cells can survive and localize in the injured site without any side effects and that the produced cell preparation ex vivo is feasible for application in the clinical setting. In the future, we plan to perform more functional studies of the produced endothelial cells in a hemophilia A model by directly evaluating their capacity to reverse bleeding. In addition, we will combine FVIII gene correction approaches to develop gene therapies for hemophilia A.