Characterization of the Procoagulant Potential of Endothelial Colonies Forming Cells and Comparison with Two Cell Therapy Products: Relevance for Vascular Cell Therapy

Endothelial colonies forming cells (ECFC) are good candidates for cell-based therapy in cardiovascular diseases. Concerns have been raised about the potential risks of ECFC-based cell therapy, in terms of thrombogenicity, particularly in inflammatory conditions. We published that cord blood ECFC (cb-ECFC) express Tissue Factor (TF) in response to TNFα. TF/FVIIa interaction did not modify in vitro the non-procoagulant properties of TNFα stimulated cb-ECFC [Cuccuini et al, JTH 2010]. We also reported, in a monocyte model, a co-expression of TF and metalloproteinase MMP-9, with possible consequences in terms of haemostasis and angiogenesis [Poitevin et al, JTH 2008].

Aim of the study: The objective of this study was: 1- to further evaluate the procoagulant potential of ECFC and the capacity to exhibit MMP activity; 2- to evaluate whether TF and MMP could be detected in two different cell therapy products (CTP).

ECFC were obtained in culture after isolation of CD34+ cells from cord blood (as previously described). CTP were obtained from an ongoing clinical trial evaluating cell therapy to improve vascularization in chronic critical limb ischemia (NIH NCT00533104). Two different CTP were analyzed: 1- Peripheral blood mononuclear cells (PB-MNC) obtained by cytapheresis without any mobilization (n= 11); 2- Bone marrow mononuclear cells (BM-MNC) (n= 11). CTP are characterized in flow cytometry, in terms of cell composition, including CD34+ cells; CD34+/CD133+/VEGF-R2+ cells were considered as endothelial progenitor cells.

cb-ECFC can express procoagulant TF in response to TNFa (10 ng/mL). The expression of TF was compared with purified monocytes, CD34+ cells, mature endothelial cells (HUVEC), PB-MNC and BM-MNC. TF mRNA were detected in CD34+ cells. As microparticles (MPs) are strongly procoagulant, we evidenced that cb-ECFC produced CD146+/Annexin V+/TF+ MPs in response to TNFa. These MPs can trigger thrombin generation. The analysis of conditioned supernatants of cb-ECFC in response to TNFa indicated that the majority of TF is borne by MP, but a small amount of soluble TF is present, in the absence of MP. This soluble TF does not induce thrombin generation. Using Q-PCR, we showed that cb-ECFC expressed Tissue Factor Pathway Inhibitor (TFPI). This expression of TFPI mRNA was not increased by TNFa. The analysis of TNFa-conditioned cb-ECFC supernatants showed that the amount of TFPI protein is a 20 fold lower than plasmatic levels of free-TFPI.

We next examined the presence of TF mRNA in both CTP. TF mRNA were estimated to a 2 fold higher in BM-MNC than in cb-ECFC (baseline). On the contrary, TF mRNA were a 5 fold lower in PB-MNC than in cb-ECFC. In both CTP, the alternatively spliced form (as-TF) represented a very minor part of TF. The percentage of CD14+ cells was respectively 17 % and 26 % in BM-MNC and in PB-MNC. The amount of TF mRNA in PB-MNC was not statistically different from unstimulated purified monocytes. On the contrary, TF mRNA were a 7 fold higher in BM-MNC in comparison with PB-MNC or monocytes. The percentage of CD34+ cells was low: 2.6 % in BM-MNC and 0.1 % in PB-MNC. TF mRNA were higher in CD34+ cells than in BM-MNC (3 fold) and in PB-MNC (27 fold).

As we previously reported a co-regulation of TF with MMP, we first analyzed the expression of MMP-2 and MMP-9 in cb-ECFC. Using Q-PCR, we found that cb-ECFC strongly expressed MMP-2 whereas the expression of MMP-9 was very weak. This profile was comparable in cb-ECFC and HUVEC. Using zymography, we confirmed the expression of MMP-2 but not MMP-9, in response to TNFa. We next analyzed CTP. In BM-MNC, MMP-2 and MMP-9 were expressed at a very high level. In PB-MNC, MMP-2 and MMP-9 could be detected but at a very low level.

Taken together, our results show that ECFC, BM-MNC and PB-MNC can express TF. The procoagulant activity of ECFC is also the consequence of MP formation in inflammatory conditions. The profile of MMP in CTP remains to be clarified in terms of cell origin and possible influence on thrombotic and angiogenic properties. In our hands, the inhibition of TF can limit procoagulant activity with no effect on proangiogenic properties in vitro. In terms of clinical relevance, one single patient treated with PB-MNC, developed a deep vein thrombosis at the site of injection. It is therefore mandatory to evaluate the procoagulant potential of CTP in the context of vascular repair.

No relevant conflicts of interest to declare.