In the present study, we investigated the effects of both donor age and passage on murine BMSC differentiation potentials towards adipogenic, chondrogenic, and osteogenic lineages. Murine BMSCs were used to facilitate future studies performed in vivo
, where implantation of tissue engineering constructs containing fluorescently labeled progenitor cells allows histologic determination of the cell source for regenerated tissues [42
]. Additionally, the availability of transgenic mice and mouse cell lines presents the opportunity for tissue engineers to investigate emerging strategies in more clinically appropriate disease models [46
]. It is important, however, to note that transgenic animals do not uniformly express GFP, and this expression is variable between tissues with values for murine bone marrow reported at close to 90% [47
] in some studies but significantly lower in others [48
], thus necessitating a careful comparison of light and fluorescence microscopy when cell numbers are quantified. Routine flow cytometric characterization of the breeder colony used in this study has repeatedly shown over 90% of adherent marrow cells express GFP, even after multiple passages (unpublished data).
No standardized practice exists for the harvest, expansion, and in vitro
differentiation of BMSCs. As noted in the Materials and Methods section, BMSCs in this study and many others refers to the adherent marrow stromal cells [16
]. Within the adherent BMSC-containing population, our data suggest that a more rapid decline occurs in differentiation potential for osteoblastic and chondrogenic lineages relative to the decline in adipogenic differentiation.
We found that osteogenic and chondrogenic potentials are adversely affected by increased donor age across all three tested donor age groups, while adipogenic differentiation potential is maintained in all but the aged donors (1 year). These results are in agreement with previous work which found that donor age affected osteogenic differentiation of BMSCs more than it affected adipogenic differentiation [24
]. Analysis of the transcriptomes of human BMSCs at passage 2 from young donors (average age = 13) suggested that BMSCs should preferentially form bone and cartilage over adipose tissue [36
]. Peng et al. recently found similar results, noting that expression of osteogenesis-related genes peaked very early following induction in BMSCs [50
]. Other studies addressed the hypothesis that age related decreases in bone regeneration were due to BMSC aging, resulting in a decreased osteogenic potential with a concurrent increase in adipogenic potential [8
]. In this study BMSCs maintained their potential for adipogenic differentiation in early aging but exhibited decreased potential for chondrogenic and osteogenic differentiation, and, although no absolute increase in adipogenic potential was observed with increasing age, the relative differences between differentiation potentials with age would thus favor adipogenesis over osteogenesis and chondrogenesis. Age related cellular dysfunction has been hypothesized to be the cause of multiple diseases of bone and cartilage associated almost exclusively with aging including osteoarthritis and osteoporosis, and loss of progenitor cell differentiation potential could contribute to these diseases [51
]. The present study supports these hypotheses.
The effect of in vitro
culture period prior to the induction of differentiation was also investigated by expanding cells through 6 passages and comparing differentiation to cells after a single passage. For chondrogenesis, increased passage only affected cells from 1-year-old donors, rendering them equal to control groups at the level of significance (p
> 0.05). For osteogenesis, the opposite effect was observed; a difference in osteogenesis due to passage alone was only observed in BMSC cultures from postnatal mice. Previous work using human BMSCs observed a greater passage related decrease in osteogenic differentiation in cells from young donors compared to cells from aged donors [16
]. Early passage postnatal BMSCs may be preferentially inclined towards osteogenesis and this preference may be quickly eliminated with repeated passages and aging, leading to the observed differences between passage 1 postnatal BMSCs and all other groups with respect to osteogenic differentiation.
The effect of passage on adipogenesis was more obscure. Quantifying the percentage of cells stained with Oil Red O showed significant passage related differences only in cultures from aged donors, whereas quantifying the percentage area stained showed significant differences in cultures from postnatal and adult donors but not those from aged mice. A previous study investigating adipogenic differentiation with increasing BMSC passages found that the size of adipocytes decreased with increased passage [16
]. Thus for postnatal and adult derived BMSCs, passage related changes in area might be largely due to decreased cell volume rather than a decrease in the number of differentiated cells. In cultures of BMSCs from aged donors, the relatively decreased number of cells undergoing adipogenic differentiation may render this effect statistically insignificant at the designated sample size (n = 6). When characterizing adipogenic differentiation via quantification of the percentage of cells stained with Oil Red O, the statistically significant interaction effect of donor age and passage reflects that cultures from aged donors were the only group to be both significantly decreased from other age groups and to have a significant decrease in cells stained between passage 1 and passage 6 cultures. This may reflect a loss of adipogenic differentiation capacity only experienced by MSCs from aged donors that is not detected when quantifying the stained area due to passage-related decreases in adipocyte size experienced in cultures from all donors. It is also important to note that the selected adipogenic cocktail utilized indomethacin, a commonly used chemical for these applications that inhibits cyclooxygenase. Indomethacin has been shown to both positively and negatively effect PPARγ in a concentration dependent manner [52
]. Although adipogenic differentiation can be induced via PPARγ dependent and independent signaling [53
], PPARγ 2 activation may be critical in BMSC differentiation as this pathway promotes terminal differentiation and suppresses Osf2/Cbfa1
]. A PPARγ ligand such as rosiglitazone may therefore be a more ideal component for adipogenic induction media for BMSCs.
Working with human BMSCs, Banfi et al. found decreased adipogenic, chondrogenic, and osteogenic potentials when increasing from passage 1 to passage 5 and found adipogenic potential to be compromised prior to osteo- or chondrogenic potential [32
]. In the present study, passage effects were variable when considered along with specific donor ages. For example, for 6-week-old donors, osteogenesis and chondrogenesis were unaffected by passage, but adipogenesis as measured by percent area stained was significantly decreased with increased passage. These results correlate well to published studies using human BMSCs [32
]; however, it should be noted that due to the use of biochemical characterization methods in addition to histology to characterize osteogenic and chondrogenic differentiation, the sample size for these methods was smaller (n = 3) than for evaluation of adipogenic differentiation (n = 6). Passage adversely affected osteogenic potential in BMSCs from postnatal donors, while chondrogenesis was only diminished by passage in BMSCs from 1-year-old donors, suggesting differences in passage effects for different differentiation lineages at different ages. When evaluating changes associated with passage, it is important to note that the adherent marrow stroma is a heterogeneous cell population, therefore there is concern that changes attributed to altered MSC function could actually be due to the preferential proliferation of one or several type(s) of cell(s) over others. The observed variable effects of passage with age suggest that different or additional factors other than BMSC number and differences in attachment/proliferation contribute to differences in differentiation potential, as, in the case that heterogeneity and subsequent differential proliferation, one would expect effects due to frequency or proliferation to be exhibited across all lineages.