Reprogrammed Human Endothelial Cells: A novel Cell Source for Regenerative Vascular Medicine

Nothing in nature is more elegant than the transformation of fertilized oocyte into a unique and complex individual containing some 10 trillion cells with more than 200 specialized functions. Also astonishing, however, is that a cell’s career decision is not necessarily permanent, but can be reversed. Committed specialized somatic cells can, in effect, go back and start again from scratch. The generation of induced pluripotent cells (iPSC), reported first by Yamanaka and group and now generated routinely in laboratories all over the world clearly demonstrates that differentiated somatic cells can be reprogrammed into a cell of desired phenotype ¹. This concept of transcription factor (TF) based alterations in a cell’s identity was further demonstrated by a number of reports attempting direct reprogramming of a differentiated somatic cells into another differentiated somatic cell without the need for pushing the reprogrammed cell to a pluripotent state, a method now known as “direct reprogramming” ^{2, 3}. In this issue of Circulation Research, Lee et al report their success in reprogramming human dermal fibroblasts (HDFs) into functional and mature endothelial cells (ECs) using a single EC TF ER71/ETV2.⁴ This approach opens a novel era in translational cardiovascular medicine by increasing the availability of autologous ECs to be employed in cell-based vascular regenerative therapy.

Derivation of engraftable human ECs could be beneficial to patients with vascular diseases. However, purification and expansion of adult ECs in therapeutically large numbers is technically challenging. Embryonic stem cells (ESCs) and iPSCs serve as promising alternatives for ECs generation and potentially provide great therapeutic potential, however, several problems limit the clinical applications of derivative cells including inefficient cell production, long duration of cell culture and tumorigenic potential.⁵ Recently, the ability to directly reprogram accessible human cells into disease-relevant cell types through cellular reprogramming has opened novel avenues for basic research and regenerative medicine. Direct reprogramming avoids residual pluripotent stem cells in the transplanted populations, eliminating the risk of teratoma formation. Moreover, higher yields and faster kinetics of EC production by direct reprogramming of non-vascular cells would lower the costs and reduce the delivery time to patients thereby permitting the use of autologous cells even potentially for sub-acute conditions. However, transdifferentiation potential by direct reprogramming has been minimal^{6, 7} and often consists of heterogeneous populations of reprogrammed cells.⁸

Direct reprogramming of fibroblasts to generate ECs by the use of ETS family of transcription factors has been previously attempted. ETS TF are implicated in hemato-endothelial specification at the embryonic phase.⁹ ETS TF drive the expression of genes associated with EC development and function.¹⁰ ETV2 is a member of ETS transcription factor family which is transiently expressed during embryonic development and is absent in adult ECs. ^{14, 15} Recent studies reported that ETV2 directly reprograms mature amniotic cells into functional ECs avoiding the poor proliferation and lineage instability associated with pluripotent stem cell derived EC.^{9, 11} However, the strategy for generating amniotic cell-derived ECs is very complex and requires long-term culture which is not applicable for even sub-acute vascular injury, is impractical and costly and could require immunosuppression given the allogeneic cell source. Although human adult fibroblasts are readily accessible for EC generation, facilitating autologous therapeutic angiogenesis, prior studies have shown that it is difficult to induce ECs from fibroblasts.¹¹ Recently, two studies examined the effects of ETV2 on reprogramming fibroblasts into ECs. Han el al reported that a combination of five TFs (FOXO1, ETV2, KLF2, TAL1, and LMO2) directly reprogrammed mouse fibroblasts into ECs. Reprogramming failed when ETV2 was omitted.¹² However ETV2 alone was not capable of converting mouse fibroblasts into ECs.¹² In contrast, very recently Morita et al reported that ETV2 alone could convert HDFs into ECs.¹⁰ Of note, the gene level of ETV2, which is minimally or not expressed in mammalian postnatal ECs^{13, 14} and implicated in certain vascular abnormalities,¹⁵ was expressed at high levels in the converted ECs.¹⁰

In this issue of Circulation Research, Lee et. al. tackled some of these limitations to successfully reprogram adult human dermal fibroblasts into functional ECs using ETV2 as single transcription factor. They endeavored to reprogram HDFs to ECs starting with a pool of 6 vasculogenic/endothelial TFs (ETV2, FOXC2, MEF2C, SOX17, SMAD1, HEY1).⁴ They observed that overexpression of these endothelial TFs converted HDFs into EC lineage. When they omitted ETV2 factor from the 6 factors, they found a significant reduction in expression of major endothelial genes compared to the combination including ETV2. Finally they showed that overexpression of ETV2 alone best induces expression of EC markers (CDH5, KDR, PECAM1, CD34, and TEK) in HDFs. ETV2 efficiently reprogramed HDFs into functional rECs with high efficiency (30%) without transitioning through a pluripotent state. Importantly, these data revealed a novel role of ETV2 in fate changes of differentiated HDFs into ECs.⁴ The authors identified two stages of reprogrammed ECs (rECs): early and late rECs (Figure) based on EC/fibroblast gene expression and cell features. They observed that early rECs, which appeared within a week after transduction of ETV2, are immature and consisted of cells with mixed signatures of ECs and fibroblasts. Functionally, the early rECs were capable of taking up Ac-LDL, formed tubular structures and contributed to vessel formation, suggesting these cells may potentially be applied to future autologous cell-based therapy of vascular diseases in patients. Moreover, the short duration of cell culture for early rEC generation is attractive for potential clinical applications. In addition, Lee et al also reprogrammed late rECs by a second round of stimuli of early rEC with ETV2 plus valproic acid (VPA), a class 1 HDAC inhibitor. They determined the late rECs show dramatically increased PECAM1 expression and exhibit all of the phenotypic characteristics of genuine ECs (HUVECs and HMVECs). The identification of late rECs provides possibilities for understanding the underlying mechanisms of maturation of reprogrammed ECs as well as a cell source for drug discovery. These novel findings are intriguing and may pave the way for enhancing the generation of reprogrammed ECs from fibroblasts and other somatic cells for therapeutic purposes.

Figure. Summary of direct reprogramming of human dermal Fibroblasts (HDFs) into endothelial cells by ETV2.

HDF: human dermal fibroblasts; rECs: reprogrammed endothelial cells; ETV2: E-twenty-six variant 2.

Based on current and previous findings, ETV2 is clearly crucial for reprogramming of HDFs into ECs. While the data reported in this manuscript provide a platform for better understanding of EC reprogramming, there are several questions that remain incompletely answered. The underlying molecular and epigenetic mechanisms of ETV2- activated EC genes and suppressed fibroblast genes remain to be determined. Although the authors used a doxycycline-inducible system to initiate reprogramming, the stability of the reprogrammed ECs after the withdrawal of doxycycline remains to be ascertained as do the epigenetic landscape and epigenetic memories of the reprogrammed ECs. The sub-specification of reprogrammed cells (arterial vs. venous) also remains to be determined as well as understanding the aging of the mature reprogrammed EC. Another unresolved question is how closely reprogrammed cells resemble their target cell counterparts? Moreover, the effect of ETVs on reprogramming of HDFs into ECs in vivo needs to be examined as this may have even greater potential for regenerative medicine applications. Other safety issues would still need to be addressed, including the risk of genomic integration of lentiviral constructs leading to spontaneous transformation during cell expansion in culture. For avoiding insertional mutagenesis, it is critical to develop non-integrating episomal vectors or substitute a non-genetic method for lentiviral vectors for ETV2-induced reprogramming HDFs into ECs.

Despite these questions, which are applicable generally to entire field of direct cellular reprogramming the current work of Lee et al provides strong evidence that ETV2-reprogrammed early rECs are functional, as evidenced by the fact that transplantation of these cells improved blood perfusion recovery and vessel formation in mouse ischemic hind limbs. The current work greatly advances the effort towards direct cellular reprogramming including that of ECs. Finally, the insights offered by Lee et al may inform new strategies for reprogramming other human non-vascular cells into ECs. Early rEC reprogrammed from HDFs by ETV2 may provide great potential for regenerative medicine and therapy of vascular injury. Recently, exosome-secreted paracrine angiogenic factors have been implicated in regenerative medicine.^{16, 17} Future studies elucidating whether ETV2-reprogrammed early-rEC-derived-exosomes regulate their angiogenesis/neovascularization properties and what contents in the exosome play critical role on those beneficial effects may shed further light on the mechanisms of action of reprogrammed ECs.