In Vitro Differentiation of First Trimester Human Umbilical Cord Perivascular Cells into Contracting Cardiomyocyte-Like Cells

Myocardial infarction (MI) causes an extensive loss of heart muscle cells and leads to congestive heart disease (CAD), the leading cause of mortality and morbidity worldwide. Mesenchymal stromal cell- (MSC-) based cell therapy is a promising option to replace invasive interventions. However the optimal cell type providing significant cardiac regeneration after MI is yet to be found. The aim of our study was to investigate the cardiomyogenic differentiation potential of first trimester human umbilical cord perivascular cells (FTM HUCPVCs), a novel, young source of immunoprivileged mesenchymal stromal cells. Based on the expression of cardiomyocyte markers (cTnT, MYH6, SIRPA, and CX43) FTM and term HUCPVCs achieved significantly increased cardiomyogenic differentiation compared to bone marrow MSCs, while their immunogenicity remained significantly lower as indicated by HLA-A and HLA-G expression and susceptibility to T cell mediated cytotoxicity. When applying aggregate-based differentiation, FTM HUCPVCs showed increased aggregate formation potential and generated contracting cells within 1 week of coculture, making them the first MSC type with this ability. Our results indicate that young FTM HUCPVCs have superior cardiomyogenic potential coupled with beneficial immunogenic properties when compared to MSCs of older tissue sources, suggesting that in vitro predifferentiation could be a potential strategy to increase their effectiveness in vivo.


Introduction
Following an acute myocardial infarction (MI), excessive cardiomyocyte loss is associated with a cascade of events resulting in ventricular decompensation and congestive heart failure (CHF) [1]. Despite advances in pharmacological, interventional, and surgical therapies, CHF contributes to significant morbidity and mortality worldwide. The heart is a partially self-renewing organ in which stem cells play an important restorative role: both endogenous bone marrowderived and heart tissue-derived stem cells participate in cardiac regeneration following ischemic injury [2][3][4]. However, in aged patients, this regenerative capacity is diminished and generally insufficient. The implantation of healthy donor cells into the damaged myocardium to replace lost cardiomyocytes and restore ventricular structure and function has been investigated since the 1990s. Various groups reported the facilitation of remodeling and enhancement of inherent healing mechanisms in preclinical models of myocardial infarction (MI) [3][4][5][6]. The regenerative effects of the implanted cell candidate are expected to be mediated at least in part through the emergence of new cardiomyocytes at the injury site.
In order to achieve the highest cardiomyogenic differentiation efficiency and purity, the differentiation of embryonic stem (ES) cells and pluripotent stem (PS) cells into functional cardiomyocytes has been thoroughly described and optimized at the level of inductive factors and their exact timing of action [16][17][18][19][20]. These strictly controlled in vitro differentiation methods gave rise to multiple cardiac engineering projects worldwide (EHM) [21][22][23][24]. However treating patients with ESC-or iPSC-based engineered tissues still poses unsolved challenges, including suboptimal immunological and electrophysiological features [25,26] as well as ethical and safety controversion.
Since their description in 1995, mesenchymal stromal cells (MSCs), in particular bone marrow MSCs (BMSCs), have gained substantial interest in tissue regeneration including heart repair [27]. According to the most recent consensus, MSCs are resident regenerative cells that can be found virtually in any vascularised tissue [28]. Once isolated, MSCs are easy to expand and culture, and they possess immunomodulatory and immunoprivileged properties and have extensive paracrine effects [29,30]. Several preclinical studies have demonstrated their safety and efficacy in cardiac regeneration [3,6].
MSC differentiation strategies have thus far involved the use of pharmacological agents such as 5-azacytidine [27] and DMSO [31] and growth and morphogenic factors like BMP-2 [32,33] or angiotensin-II [34]. Harder to define and control, yet physiologically more relevant strategies were based on the hypothesis that the cardiac microenvironment itself can grant the multifaceted inductive effect MSCs may need for cardiomyogenic transformation and commitment. The application of cardiac cell lysates on BMSC cultures [35], direct coculture of MSCs with cardiomyocytes in vitro [14,36], or the implantation of MSCs into the ventricular myocardium of small animals in vivo [37,38] resulted in increased cardiac marker expression in MSCs. Previous reports suggested that cardiomyogenesis occurred spontaneously after treating cardiac injuries with MSCs [39][40][41][42], although de novo cardiomyocytes also arose from the fusion of BMSCs and cardiomyocytes [43,44]. Earlier several groups reported the generation of functional, spontaneously contracting cardiomyocytes from non-human MSC-like cells [31,45] but to our knowledge the derivation of functional cardiomyocytes from human MSCs of any tissue source has not been reported.
The perivascular region of human umbilical cord tissue is a rich source of MSCs with pericyte properties [46][47][48]. The cardiomyogenic potential and advantages of human umbilical cord perivascular cells (HUCPVCs) over BMSCs have been demonstrated by several groups, both in vitro [46,49] and in vivo [50][51][52]. We have previously shown that, in comparison with term HUCPVCs, FTM HUCPVCs have increased proliferative capacity and higher multilineage differentiation capacity, including the cardiomyogenic lineage [46]. We hypothesized that young, multipotent FTM HUCPVCs have greater overall cardiomyogenic potential when compared to term HUCPVCs or BMSCs. In order to further investigate and compare the cardiomyogenic differentiation potential of HUCPVCs and BMSCs, we employed single cell suspension and aggregate-based direct cocultures of HUCPVCs and BMSCs with primary cardiomyocyte feeder layers to induce their differentiation into functional heart muscle-like cells.
All animal procedures were approved by the Animal Care Committee of the University Health Network (Toronto, Canada) and all animals received humane care in compliance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health 1996). As described previously [46], primary rat cardiomyocyte cultures were prepared from rats sacrificed at p2-5. Heart ventricles were minced and agitated at 37 ∘ C in 1.5% trypsin PBS solution. Primary cultures were treated with BrdU (16 h, 5 M) and washed prior to addition of MSCs 24 h later. Dissociated MSCs were added to cardiomyocyte monolayers to achieve 5% MSC content in each well. Alternatively MSCs were labeled with viable, nontransferable fluorescent dye (CellTracker Green, Invitrogen 5 M, 1 h) prior to transfer, in order to visualize integrating MSCs.
Aggregate formation was induced by hanging 25 L drops of cell culture media containing 500 MSCs each (passage# ≤ 6). After 3 days, aggregates were transferred onto primary rat cardiomyocyte monolayers and kept in coculture for over 7 days. Alternatively aggregates were incubated with viable, nontransferable fluorescent dye (CellTracker Green, 5 M, 1 h) prior to transfer, in order to better visualize and record developing aggregates. Aggregate cocultures were dissociated (trypsin 0.25%, 3 min) for flow cytometry (FC) or fluorescence-activated cell sorting (FACS) analysis.
ReproCardio6 human induced pluripotent stem (hiPS) cell derived cardiomyocytes (ReproCell, Japan) were used as Stem Cells International 3 positive control for cardiomyocyte marker gene and protein expression analysis.
Stained specimens were kept in mounting media (Per-maFluor, ThermoScientific) and images were acquired using a fluorescent microscope (EVOS, Life Technologies) or spinning disk confocal microscope (Quorum Zeiss AxioVert, SickKids Imaging Facility, Toronto). Images were quantified using Volocity6 (Perkin-Elmer) imaging software.

RNA Isolation and Quantitative RT-PCR.
Undifferentiated hMSC cultures or hMSCs sorted from cocultures (TRA-1-85 high ) were lysed with RLT lysis buffer (Qiagen). RNA samples were prepared using the RNAeasy kit (Qiagen), according to the manufacturer's instructions. cDNA was prepared using RNA to cDNA EcoDry Premix (Clontech), according to the manufacturer's instructions. For MY6H analysis, Taqman qPCR primer assays were used (Applied Biosystems) with IDT PrimeTime qPCR probes. For cTnT analysis, primers were synthesized (Sigma) and the reaction mixture was prepared using the Rotor-Gene SYBR Green PCR kit (Qiagen). Human cardiac RNA positive control was from Ambion. qPCR was performed on the Rotor-gene 6000 instrument (Qiagen) using 10 ng of cDNA per reaction. Gene expression levels were determined by PCR Array Data Analysis Centre (SABiosciences) using the delta-delta Ct (ΔΔCt) method and reported as fold increase compared to undifferentiated cells. GAPDH was used as a normalizer. All assays were done in triplicate for at least 3 independent experiments. Oligo sequences are as follows: CTnT F: GGC AGC GGA AGA GGA TGC TGA A; CTnT R: GAG GCA CCA AGT TGG GCA TGA ACG A; MYH6 F: GCA AAG TAC TGG ATG ACA CGC T; MYH6 R: GTC ATT GCT GAA ACC GAG AAT G.

Monocyte Cytoxicity Assay.
Human blood was obtained with written informed consent from healthy volunteers (REB number 28889, University of Toronto, Toronto, Canada). Mixed lymphocytes were isolated using Ficoll gradient centrifugation. Cell concentration, viability, and purity were determined by microscopy using trypan blue. Lymphocytes were incubated with MSC monolayers at a ratio of 10 : 1 for 72 h. Colorimetric LDH measurement (Roche) was performed using cell free medium after lymphocyte incubation and quantified with plate reader (FilterMax F5, Molecular Devices) at 490 nm.
Seven days after coculture, FC analysis using anti-TRA-1-85 antibody identified low TRA-1-85-positive (TRA-1-85 low ) cells, distinguishable from both TRA-1-85-negative (rat) and high TRA-1-85-positive (human) cell populations (Supplementary Figure 1A, gate L). ICC for human nuclear antigen (HuNu, Supplementary Figure 1B) showed that TRA-1-85 high cells had single nucleus or multiple HuNu positive nuclei, while over 70% of TRA-1-85 low cells had several nuclei with various ratios of HuNu positive and negative labeling (Supplementary Figure 1B). This phenomenon was observed in all 3 MSC types and suggests the occurrence of some cell fusion in cardiomyocyte cocultures.
MSCs from direct cocultures were counted after differentiation (day 7). Starting with equal amount of cells (10k Stem Cells International 5 cells on day 0), we found remarkable differences in cell count between MSC types at the time of extraction (day 7). MSC numbers increased in both FTM HUCPVC-and term HUCPVC-containing cocultures (27k ± 7.1k, 17k ± 9.9k, resp.) but decreased in BMSC cocultures (3.8k ± 1.7k) ( < 0.01) (Figure 1(b)) suggesting a higher survival and proliferation rate of PVCs in cocultures.

Differentiating FTM HUCPVC Aggregates Spontaneously
Contract on Rat Cardiomyocyte Feeder Layers. While no contracting term HUCPVC or BMSC aggregates were observed, 3 out of 5 FTM lines consistently produced aggregates that exhibited spontaneous pulsation after 5 days in coculture with primary cardiomyocytes (Videos 1, 2, and 3). Contracting cells were observed first in the inner parts of the aggregates as revealed by using prestained FTM HUCPVCs (CellTracker Green) (Video 1). Over 7 days the entire aggregate became motile (Videos 2-5) and preserved its activity even after the rat cardiomyocyte feeder layer stopped displaying physical activity (Video 4, Video 5). Furthermore, FTM HUCPVC aggregates sharing a common cardiomyocyte feeder layer exhibited synchronous contractions (Video 4, Video 5).
MSC differentiation led to an increased sensitivity to lymphocyte-mediated cytotoxicity, with FTM and term HUCPVCs being significantly less susceptible towards cellular cytotoxicity than BMSCs (Figure 3(d), inner panel).

FTM HUCPVCs Express Alpha-Sarcomeric Actinin after One Week of Coculture with Primary Rat Cardiomyocytes.
The emergence of contracting cells from FTM HUCPVCs suggested the expression of alpha-sarcomeric actinin after differentiation. We performed immunocytochemistry on undifferentiated and differentiated FTM HUCPVCs (Figure 4) in order to detect alpha-sarcomeric actinin (aSarc) in cells with human nuclear antigene (HuNu) positive nuclei. While undifferentiated FTM HUCPVCs did not  show alpha-sarcomeric actinin expression (Figure 4(a)), FTM HUCPVCs in direct cocultures with primary rat cardiomyocytes for 1 week (Figure 4

Discussion
Our study supports 3 main novel conclusions as follows: (1  differentiation of MSCs. Although aggregate generation and hypoxic preconditioning are established strategies for human ESC-based cardiac differentiation [16,57,58] and the control of oxygen tension during the differentiation process is an essential feature of cardiac tissue engineering [19], the relevance of such approaches for human MSC applications has not been thoroughly studied. This finding suggests that strategies enhancing aggregate formation could increase the efficiency of HUCPVC-based differentiation for optimized tissue engineering applications. To elucidate the underlying cause of the observed difference in aggregate formation potential between BMSCs and HUCPVCs, we examined the expression of CD49f (integrin alpha 6) in FTM and term HUCPVCs in comparison with BMSCs. CD49f is in direct regulatory connection with the pluripotency markers OCT4 and SOX2, denoting cells with higher stemness [53], and is shown to be crucial determinant of sphere formation potential of MSCs [53]. We previously demonstrated that FTM HUCPVCs have increased OCT4A expression when compared to term counterparts or BMSCs [46]. CD49fpositive MSCs were also reported by others to have increased cardiac regenerative capacity in a post-MI mouse model [54]. Here, we demonstrated that CD49f expression levels are increased in HUCPVCs relative to BMSCs and that this correlates with their aggregate formation capacity and cardiomyogenic differentiation potential. The immunomodulatory effect of administered cells, through both secreted factors and expression of cell surface molecules, is an important determinant of in vivo engraftment. Based on their source and age, MSCs vary in expression of CAMs, HLA molecules, and also T cell activation and response [59]. While HLA-A levels determine susceptibility towards cytotoxic T cells, the immunomodulatory protein HLA-G attenuates the response of the innate immune system [60]. Representation of these antigens on engrafted cells could define a host's cellular immune response after allogeneic transplantation. Our findings are in agreement with previous reports that mesenchymal stromal cells of umbilical cord origin display lower HLA-A and higher HLA-G expression levels when compared to adult or other extraembryonic tissue originated MSCs [61]. Immunogenic properties of BMSCs have been shown to change during differentiation and lead to rejection in allogeneic models [62]. Undifferentiated FTM and term HUCPVCs exhibit low HLA-A levels but express HLA-G that provides protection for fetal tissues in utero. Although term HUCPVCs upregulated HLA-A during aggregate-based cocultures, FTM HUCPVCs maintained low HLA-A expression, and both HUCPVC types sustained or upregulated HLA-G in the course of differentiation in vitro. Under the same experimental conditions, BMSCs exhibited the highest HLA-A expression and negligible HLA-G positivity. Furthermore a functional assay measuring T cell reaction (LDH release) showed a significantly higher susceptibility of differentiated BMSCs towards T cell mediated cytotoxicity compared to HUCPVCs.
Based on our results we propose that FTM HUCPVCs have superior cardiomyogenic differentiation potential and favorable immunogenic properties when compared to human MSCs derived from older tissue sources that make them highly promising candidates for cardiac tissue engineering and in vivo cardiovascular regenerative applications.