While the effects of the OM supplements have been largely studied with BMSCs [15, 19–21], there are only limited studies with ASCs [16, 22, 39]. One of the major shortcomings in all of these in vitro studies has been the lack of comparison between different serum conditions. Due to quality and safety issues, the clinical hASC-based applications need to move from the animal-derived products to human-derived or more preferably to defined and xeno-free conditions . However, most of the in vitro studies are still conducted using FBS. Given that the serum conditions can significantly affect the cell response, it is crucial to obtain research data with more clinical relevance [26, 31].
Accordingly, we aimed to optimize the osteogenic culturing conditions for the in vitro induction of hASCs by testing different concentrations of Dex and AsA2-P in HS-based medium as well as in a defined, xeno-free medium RegES, with FBS medium functioning as a control. As hypothesized, the differential effect of FBS, HS or RegES media was evident in hASC proliferation and osteogenic differentiation. Comparison of the cell growth in FBS, HS and RegES maintenance media revealed a slightly higher growth rate in hASCs cultured in FBS MM than in HS MM, whereas the highest growth rate was achieved in RegES MM. When comparing MM and OM conditions in general, significantly higher cell number was achieved in OM than MM particularly in HS and RegES cultures. The effect of OM in cell number was also dependent of serum conditions; under FBS conditions OM3 with the highest AsA2-P concentration promoted cell growth most, but under HS conditions both OM2 and OM3 resulted in equally elevated cell numbers. High AsA2-P concentration (250 µM) has been reported to stimulate proliferation of BMSCs and osteoblast-like cells [21, 40], whereas high Dex concentration (100 nM) may inhibit proliferation . More importantly, high AsA2-P concentration may stimulate proliferation without a reciprocal loss of differentiation potency [21, 41].
The osteogenic induction capacity of OM1, OM2 and OM3 was compared by analyzing the ALP activity, mineralization and relative expression of several osteogenic markers. Contrary to expectations, the level of osteogenic differentiation was the lowest in FBS-cultured hASCs as measured by ALP activity and mineralization. Another significant finding was that the RegES MM alone was found to induce the early osteogenic differentiation as shown by the elevated ALP activity, although supplementation with Dex, AsA2-P and β-GP was required to achieve mineralization. On the whole, the inductive effect of high AsA2-P and low Dex concentration, as in OM3, was most evident in HS cultures, resulting in high ALP activity and mineralization. Prior studies have noted the importance of AsA2-P in the osteogenic differentiation of BMSCs and osteoblastic cells [21, 41–43]. In contrast, Dex in high concentrations has been shown to inhibit osteogenic differentiation [44, 45], although it seems to be necessary for the efficient osteo-induction of MSCs in low concentrations [15, 19, 20, 44].
In the present study, variation between donors could be detected particularly in ALP activity. Under HS and FBS conditions, two out of three donors exhibited a notable increase in ALP activity in response to OM2 and OM3, whereas one donor did not seem to respond to the OM supplements. Donor sample variability affects the interpretation of the results regardless of analyzing method and diminishes the statistical significance by increasing standard deviation. Others have established this problem previously with ASCs [11, 23, 46] and BMSCs [15, 47, 48]. With BMSCs, Jaiswal and co-workers showed that the basal level of ALP activity as well as the timing of the peak ALP activity varied greatly between the different donor samples irrespective of donor age . The relative fold induction in ALP activity varied 1.5- to 6.4-fold depending on donor . Similar kind of variation has been detected with ASCs as well [11, 23, 46].
The osteogenic effect of OM1, OM2 and OM3 under different serum conditions was further studied by analyzing the relative expression of osteogenic markers, runx2A, DLX5, collagen type I, osteocalcin and ALP. Overall, the qRT-PCR data demonstrated significant upregulation of runx2A mRNA under osteogenic differentiation in FBS and HS medium, and early stimulation of DLX5 under FBS conditions. The regulation of other markers, collagen type I, osteocalcin and ALP, was modest. When comparing the different OM compositions, OM2 and OM3 resulted in significantly higher expression of runx2A, collagen type I, and ALP, than corresponding MM under FBS or HS conditions. In some cases, OM2 and OM3 also resulted in higher expression of runx2A, collagen type I, and ALP than OM1. Under FBS conditions, OM3 induced significantly higher expression of runx2A when compared to FBS MM at day 14. In addition, collagen type I expression was significantly upregulated by FBS OM2 and OM3 at day 7 in comparison to both MM and OM1, and versus OM1 on day 14. Under HS conditions, OM3 resulted in significantly higher runx2A expression when compared to HS MM and OM1 at day 7. Moreover, HS OM3 resulted in higher level of ALP expression than OM1 and OM2 on day 14. Hence, greater expression of osteogenic markers can be achieved by OM with increased AsA2-P and lowered Dex, that is OM2 or OM3 composition.
In RegES medium, significant differences in the runx2A, collagen type I or ALP expression were not detected, likely due to the high basal level of expression in RegES MM. However, the level of osteocalcin mRNA was significantly lower in RegES OM1 than in MM, OM2 or OM3 at day 14. The expression of DLX5 appeared to peak already on day 7 under osteogenic induction under FBS conditions, whereas no regulation of DLX5 was detected under HS and RegES conditions.
There were also notable differences in the collagen type I and ALP mRNA expression between the serum conditions in general. The expression of collagen type I was notably lower in all RegES groups when compared to the respective FBS and HS groups. The expression of ALP, on the contrary, was higher in the RegES groups than in FBS and HS groups. This result correlates with the high ALP protein activity in RegES cultures, but in turn, not with the high ALP activity found in HS cultures. Unexpectedly, the expression of runx2, collagen type I, ALP or osteocalcin was not upregulated by time (day 7 versus 14) in any of the OM groups. Although we did not see significant upregulation by time, the culture period of 14 days has been shown to be sufficient for the detection of osteogenic gene expression [49–51]. Furthermore, considering clinical applications, the in vitro culture period should be minimized, as prolonged culture may increase the risk of contamination or genetic abnormalities. Therefore, the inductive effects in vitro should also appear within as short a time period as possible.
In contrast to ALP activity and mineralization, relatively small changes were detected in mRNA expression levels between MM and OM groups under all serum conditions. Although challenging the common conception, the lack of upregulation or even downregulation of certain osteogenic markers in ASCs upon osteogenic differentiation has been reported previously [12, 18]. For OM-induced BMSCs, similar differences have been reported between the data obtained on protein level and real-time PCR for collagen type I expression . The flow cytometric analysis showed significantly upregulated expression of collagen type I in the ECM of differentiated BMSCs, but there was no notable increase in collagen type I mRNA even after osteogenic differentiation according to real-time PCR . While this phenomenon is still poorly understood, it is evident that the route of mRNA to protein is a highly regulated and complex pathway, where even small changes at transcriptional or posttranscriptional level can have a major phenotypic effect [12, 18, 53].
One of the main scopes of the present study was to investigate whether xeno-free RegES medium could be utilized for the efficient osteogenic differentiation of hASCs. Taken together, hASCs cultured in RegES showed increased osteogenic capacity, an effect likely explained by the composition of the RegES medium (Additional file 1). Apart from the several growth factors and cytokines, the high concentration of AsA2-P in RegES MM (50 µg/ml or 170 µM) may account for the high basal level of cell growth and osteogenic differentiation of hASCs in the plain RegES medium. Taking into account the additional AsA2-P in the osteogenic media, the concentrations in RegES OM1, OM2 and OM3, raised to 220, 320, and 420 µM, respectively. The maximal advantage in growth rate and differentiation in RegES-based osteogenic media was reached by AsA2-P concentrations varying from 220 to 320 µM, correlating to the results obtained with OM3 under FBS and HS conditions. However, as already mentioned, supplementation with Dex and β-GP in addition to AsA2-P was required for efficient osteo-induction and maturation of hASCs cultured in RegES.