We have successfully isolated a highly proliferative MSC population from CV of postnatal placenta while optimizing our protocols to maximize cell yield from extraction (approximately 2.7 × 107 cells/kg wet weight of CV) and after a few weeks of in-vitro culture (approximately 7.5 × 109 cells/kg wet weight of CV, after 3.6 ± 0.3 cPD). We kept CV-MSC in culture for an extended period (138 ± 64 days, corresponding to 25 ± 3 cPD) and observed delayed onset of the earlier stages of cell senescence compared to BM-MSC. In line with Barlow et al. [23], we observed that CV-MSC proliferate faster and show greater long-term growth ability than BM-MSC.

Collagenase digestion was used for cell isolation in our study so associated costs must not be overlooked as they certainly rise above those of the concurrent explant method when digesting maximized volumes of placenta tissue is a priority. However, collagenase digestion has the advantage of allowing for higher final cell yields in a short period of time. Adapted or alternative methods for isolation of placental MSC of fetal origin have been proposed [52, 53]. The preferred isolation method and in-vitro culture conditions must be assessed according to the specific requirements of each given stem cell application.

We have extensively characterized the CV-MSC cultured fraction in early and late passages. It is obvious that CV-MSC fit all MSC defining criteria [48], as reported previously [3, 50, 51, 54], including being able to differentiate into osteoblasts, adipocytes, and chondrocytes cultured in vitro in standard differentiation cocktails. We also observed that CV-MSC present a surface marker profile consistent with increased potential for proliferation (high CD44, CD73, and CD166 combined expression), vascular localization (high STRO-1, CD146, and CD106 combined expression), and predisposition for myogenic commitment (high α-SMA, SM22α, and CD146 combined expression), unlike BM-MSC. It has been introduced that combined CD146 expression and a high α-SMA expression is associated with SMC commitment in BM-MSC [55]. Myogenic expression of CV-MSC, however, remains another topic of debate, as some studies report residual expression of α-SMA in MSC isolated from placenta tissues [56, 57] while others, such as Castrechini et al. [50], report high expression but just in CV-MSC isolated from first-trimester pregnancies.

Numerous studies report that CV-MSC isolated by collagenase digestion are prone to contamination by maternal cells, which rapidly and completely overgrow CV-MSC within one or two passages [23, 58, 59]. To address this issue we have excluded by FISH X/Y chromosome analysis the presence of maternal contamination in late-passage CV-MSC.

An often overlooked contributor to an even more complete phenotypic characterization is trophoblastic contamination in the cultures. Trophoblastic cells are the cells forming the outer layer of the blastocyst that provide nutrients to the embryo. This layer later develops into a large part of the placenta and so trophoblast-derived cells make up the majority of the chorion in the developed placenta. To determine cytotrophoblastic cells (eventually the most frequent trophoblastic cell type) in vitro, markers such as human chorionic gonadotropin, pan-cytokeratin, epidermal growth factor receptor HER2, and E-Cadherin could be investigated. Despite the controversy, most indications for trophoblastic contaminations are unlikely [20, 50].

In terms of ECM remodeling potential, CV-MSC showed the ability to produce matrix proteins (namely FN and OPN) and contract a collagen matrix in vitro similar to BM-MSC, confirming CV-MSC suitability for major stem cell-based applications.

To understand whether the observed increased potential of CV-MSC to proliferate correlated with increased telomere length, we analyzed the telomeres, repeated TTAGGG sequences at the ends of chromosomes that protect them from deterioration or fusion. Telomeres undergo progressive shortening with each cell division. Progressive telomere shortening is therefore one of the molecular mechanisms underlying aging, as critically short telomeres trigger cellular senescence and loss of cell viability [6062]. Thus, telomere length is known to decline during in-vitro and in-vivo aging. Our results showed that CV-MSC undergo minimal telomere loss as they age in vitro compared to all other MSC sources. In contrast, UC-MSC underwent dramatic telomere loss within the very few passages we were able to keep them in vitro. At the same time we observed that UC-MSC become quickly senescent (i.e., in early passages) in β-galactosidase senescence assays. Interestingly, CV-MSC escaped in-vitro senescence—showing reduced β-galactosidase-positive cell rates—for more passages than all the other cells types due to increased potential to maintain telomere length while proliferating. The increased potential of CV-MSC to maintain telomere length over the other tissue sources reflects not only different rates of telomere loss within those tissues, but also different telomere loss rates during in-vitro passaging, potentially due to oxidative stress. Given that both CV-MSC and UC-MSC originate from postnatal tissues with virtually equivalent chronological age but dramatically different telomere length maintenance potentials, it is not clear to us why telomere erosion during in-vitro culture was so dramatic in UC-MSC. It can be related to donor variability, differential telomerase activity expression, or alternative mechanisms of telomere maintenance, as discussed later.

DNAm is known to change during aging. However, some CpG sites show almost linear changes during aging and so can be used for age prediction. Weidner et al. [45] have established an EAS to predict aging with higher precision than telomere length alone. EAS revealed predicted age in CV-MSC decreased rapidly with progressive passaging, confirming delayed aging phenomena. It is worth mentioning that our results were apparently limited by the fact that the signature was designed for age estimations from blood samples and does not seem to suit cultured cells. This leads us to a discussion based on the analysis of linear regression fits and not on absolute predictions (which fall out of the acceptable chronological age range).

hTERT is a catalytic subunit of the enzyme telomerase, which together with the telomerase RNA component (hTERC) comprises the most important unit of the telomerase complex. hTERC acts as a template for the addition of telomere units by hTERT. hTERT is expressed during early development but is absent in most somatic cells, with the exception of proliferating cells and renewal tissues [63]. In highly proliferative cells of the germline, in ESC, and in the majority of cancer cells, telomerase (by adding telomeric repeats onto the chromosome ends) prevents the replication-dependent loss of telomeres and cellular senescence [61]. The causal relationship between expression of telomerase, maintenance of telomere length, and elongated life span of the human cell has been established. We have confirmed the absence of hTERT mRNA in CV-MSC. hTERT gene expression typically corresponds to telomerase activity in many multicellular organisms. This can be untrue in some cases. Izadpanag et al., Yanada et al., Zimmermann et al., and Hiyama and Hiyama [6467] reported that low levels of telomerase activity were found in MSC. Contradictory studies report no telomerase activity in MSC [68]. Therefore, a mechanism other than or in addition to telomerase—such as alternative lengthening of telomeres (ALT)—might play an important role in CV-MSC telomere maintenance. There are, for instance, hints from work done on subtelomeric DNA hypomethylation facilitating telomere elongation in mammalian cells suggesting that epigenetic modifications of chromatin might occur in MSC [69]. Work done in whole chorion tissues indicates a downregulation of telomerase activity over the gestation, also supporting the idea of a decline of primitive stem cell features with aging [70]. In order to clarify the origin of our telomere observations, investigating telomerase activity levels and ALT mechanisms in CV-MSC would thus be an interesting outlook.

Previous studies using equivalent methods for isolation of CV-MSC [79] report the presence of panels of pluripotent markers such as NANOG, OCT4, and SOX2 in those cells. Studies typically focus on gene expression level observations only and often PCR primer sequences or expression data are omitted. Without further concerns, some conclude that CV-MSC retain characteristics of pluripotent stem cells.

We have designed qRT-PCR primers binding to the DNA region encoding for the highly conserved AFMVW helix inside the HMG domain of the SOX2 protein. Additionally, we used SOX2 primers to bind the DNA region encoding for the C-terminal domain of the SOX2 region, equivalent to what was done in other studies [71, 72]. Our data show that CV-MSC express SOX2 on both the gene and protein levels, an indication of improved neurogenic potential in the light of current knowledge. SOX2, a pluripotency marker, is also known to regulate FGF4 expression, which in turn promotes neural stem cell proliferation and differentiation in the postnatal brain [73]. The improved neurogenic potential of CV-MSC compared to BM-MSC has in fact been demonstrated [18, 20].

We did not detect the presence of OCT4 variant 1 or NANOG in CV-MSC. Our data are partially supported by previous work from Jones et al. [72], who compared first-trimester to term fetal placental chorionic stem cells. They observed no detectable OCT4A variant 1 in the term fetal cells at the transcript level using primer pairs binding only to a larger DNA fragment within the same region as ours. Based on the absolute expression of NANOG reported in that work, and given the fact that one of the two primer pairs we used was equivalent to theirs, we consider the possibility of marginal but nondetectable expression of NANOG in our cells due to donor variability.

We nonetheless commit to exclude the possibility of pluripotency in CV-MSC given the corroborated absence of OCT4A variant 1. OCT4 and SOX2 were identified as the fundamental transcriptor factors underpinning naïve pluripotency [74], although the critical role of SOX2 might be to activate OCT4 [75, 76]. NANOG becomes part of the OCT4/SOS2/NANOG (OSN) triumvirate as its presence is crucial for the acquisition but not the maintenance of naive pluripotency [77, 78].

One other indisputable feature of pluripotent stem cells is the formation of teratomas in vivo. We applied the equivalent in-vitro assay, designed to assess the tumorigenic potential of cells in culture—the soft agar assay—and found no evidence for malignant transformation of CV-MSC, as suggested previously [16, 79].

In human ESC the predominant signaling pathways involved in pluripotency and self-renewal [80] are TGF-β (signaling through SMAD2/3/4, activating the MAPK and AKT pathways) and the noncanonical WNT pathway (β-CATENIN signaling). Pluripotency signaling through these pathways relies predominantly upon the key transcription factors OCT4, SOX2, and NANOG. When NANOG is inhibited, differentiation takes place via the BMP pathway (Smad1/5/9 signaling) and NOTCH intracellular domain (NICD, CSHL1). After investigating multiple pathways we found no evidence of any pathways being differentially activated/deactivated leading to pluripotency of CV-MSC.

Placental-derived MSC have been reported to be capable of neural, retinal cell, pancreatic progenitor cell, and hepatic cell differentiation [8, 81], an indication for greater plasticity. We have no evidence, however, to support the notion that a putative pluripotent stem cell population is present within CV-MSC or that CV-MSC are less differentiated than BM-MSC.