br Materials methods The derivation and culturing conditions

Materials & methods The derivation and culturing conditions of hAT-MSCs are thoroughly described in Supplementary Materials and methods section, while details regarding hiPS lines generation and validation are described elsewhere (Kyrkou et al., in press).
Results
Discussion Human AT-MSCs are well-characterised in many studies constituting a solid reference point to compare the efficiency of hiPSC-based cell therapies. The isolated hAT-MSCs complied with the criteria set by the Mesenchymal and Tissue Stem Cell Committee of the International Society for Cellular Therapy (ISCT) (Dominici et al., 2006). We generated hiPSCs from human foreskin fibroblasts (HEFs) using a modification of the footprint-free episomal strategy (Okita et al., 2011). All established hiPS clones were positive for characteristic embryonic markers, their microarray profile was identical to embryonic stem Cy3-dCTP Supplier (HUES1) and they could generate all three germ layers in vivo (Kyrkou et al., in press). Previous published studies have demonstrated the derivation of MSCs from hiPSs through the inhibition of TGF-β pathway (Chen et al., 2012), the supplementation with various growth factors (Niwa et al., 2011) or the induction with MSC-conditioned medium (Lee et al., 2014). In the present study, hiPS-MSCs were derived through alteration of the serum free hiPS medium to a serum-based specialised MSC medium. This simple method resulted in mesengenic progeny possessing the characteristic features of MSCs 20days post-commitment in vitro. These hiPS-MSCs exhibited self-renewal capacity, tripotentiality and MSC-related phenotypic profile. Direct comparison of hAT-MSC and hiPS-MSC phenotypic profiles exhibited high similarity in expression levels for all mesenchymal, pericytic, haematopoietic and endothelial markers tested. Interestingly, CD324 (E-Cadherin) is highly expressed (~72%) in hiPSC colonies and strongly downregulated (~12%) in hiPSC-MSC post-mesodermal commitment indicating that loss of E-Cadherin-mediated cell–cell contacts occurs during epithelial to mesenchymal transition leading to the acquisition of a mesenchymal phenotype, as expected (Cavallaro and Christofori, 2004; Bates and Mercurio, 2005). In parallel, CD146 (MCAM) expression is strongly increased (from ~46% to 92%) in hiPSC-MSC post-mesodermal commitment suggesting that CD146 is either an inducer of epithelial to mesenchymal transition (Zeng et al., 2012) or simply correlates with an MSC subpopulation located in perivascular tissue niches possessing a vascular smooth muscle phenotype (Espagnolle et al., 2014). Genome-wide transciptome microarray analysis of hAT-MSCs and hiPS-MSCs revealed high similarity of upregulated transcripts between these mesenchymal populations versus the basal pluripotent stem cell state (hiPS). Importantly, a recent study using human platelet serum to establish hiPS-MSCs, revealed a close relationship between MSCs and generated hiPS-MSCs in their global gene expression profile (Frobel et al., 2014). In the present study, the common top 200 (p<0.005) enriched Biological Processes of hAT-MSCs and hiPS-MSCs when compared to the pluripotent background of hiPS indicated an increased tendency of these progenitors to be involved in functional processes related to angiogenesis, ossification, connective tissue formation and epithelial to mesenchymal transition. Similarly, in a previous study hiPSCs commitment to hiPS-MSCs resulted in the upregulation of mesodermal genes (MSX2, NCAM, HOXA2) and downregulation of pluripotency genes (OCT4, LEFTY1/2) in early cultures (10days). In the same study, low-throughput microarray analysis of hiPS-MSC late cultures (passage 3) indicated increased expression of genes related to epithelial to mesenchymal transition (Chen et al., 2012). Taken together, these results point to the fact that hiPSC-MSC is a mesenchymal population closely related to hAT-MSCs. To the best of our knowledge, the simultaneous in vitro reconstruction of both fibrous and osseous zones in order to produce a viable baACL construct has not been used before. For this purpose, two growth factor cocktails have been applied, a ligament inducing (TGF-β/FGF-2) in the middle part and a bone forming (BMP-2/FGF-2) in both ends of the biomaterial. Genome-wide transciptome analysis of the ligamentous and osseous zones from hAT-MSC and hiPS-MSC baACLs allowed us to make the following statements. The differentiation of hAT-MSCs to the ligamentous and osseous parts of the biopolymer proceeds with a gene expression pattern that is surprisingly similar. Indeed, when compared to the undifferentiated hAT-MSCs, the up-regulated transcripts in the fibrous (183) and the osseous (191) parts were mostly common (179), even though the inducer was different TGF-β/FGF-2 vs BMP-2/FGF-2. The same applied for the down-regulated transcripts too (from the 51 and 56 down-regulated transcripts in the fibrous and osseous parts, respectively, 50 were common). The same was observed in the case of hiPSC-MSCs\' transcriptome at the end of fibrous and osseous differentiation on the scaffold versus the transcriptome of pre-seeded hiPSC-MSCs, except that the number of the down-regulated transcripts was much higher. Indeed, from the up-/down-regulated transcripts in the fibrous (205/337) and the osseous parts (196/285) the great majority were common (175/279). Considering that both cocktails contain FGF-2 and that BMP-2 activates SMAD1/5/8, an activation that can be also accomplished by TGF-β via ALK1, then perhaps such a similarity in the expression pattern is not so surprising. However, the presence of small quantities of BMP-2 are capable of driving the differentiation to bone under these circumstances. In addition, small up-regulation of common ligament- and osteo-related transcripts may be attributed to the specific concentrations used in TGF-β/FGF-2 and BMP-2/FGF-2 cocktails, respectively. In further studies, different growth factor concentrations could yield more striking transcriptome differences between fibrous and osseous parts.