2D). Both the pharmacological AMPK inhibitor compound C (Figs. 3A, B)

and transfection with AMPK shRNA (Figs. 3C, D) also suppressed osteogenic differentiation of hDP-MSC. The shRNA silencing of AMPK early during hDP-MSC activation (day 1) prevented activation of AMPK/Raptor and restored the activity of the negative autophagy regulators mTOR/S6K, resulting in the inhibition of LC3-II increase (Fig. 3E). On the other hand, late inhibition of AMPK at day 3 by compound C completely failed to block osteogenic differentiation (day 7 ALP values: 2.07 ± 0.10 and 2.11 ± 0.06 in control and compound C-treated hDP-MSC, respectively; n = 3, p > 0.05). Similarly, autophagy inhibitors bafilomycin and chloroquine were also ineffective in preventing hDP-MSC differentiation if added at day GDC0199 Bcl-2 phosphorylation 3 (ALP values: 1.82 ± 0.15, 1.76 ± 0.10 and 1.74 ± 0.08 in control, bafilomycin and chloroquine-treated hDP-MSC; n = 3, p > 0.05). Therefore, it appears that early AMPK-dependent autophagy is required for optimal differentiation of hDP-MSC to osteoblasts. Finally, we explored the role of Akt/mTOR activation in AMPK-dependent osteogenic differentiation of hDP-MSC. The selective Akt antagonist DEBC (Figs. 4A, B), as well as pharmacological mTOR inhibitor rapamycin (Figs. 4C, D) or

transfection with mTOR siRNA (Fig. 4E), inhibited hDP-MSC differentiation to osteoblasts, as confirmed by alkaline phosphatase assay and RT-PCR/immunoblot analysis of osteocalcin, Runx2 and BMP2. Similar effect, although somewhat second less pronounced, was observed even if DEBC or Akt were added at day 3 (day 7 ALP values: 1.47 ± 0.09, 1.20 ± 0.05 and 1.28 ± 0.01 in control, DEBC- or rapamycin-treated hDP-MSC; n = 3, p < 0.05) or even day 5 of differentiation (data not shown). The suppression of Akt phosphorylation

in DEBC-treated hDP-MSC prevented activation of mTOR/S6K at day 5 of differentiation, while AMPK activation remained largely unaffected ( Fig. 5A). Both the mTOR siRNA and rapamycin reduced the phosphorylation of mTOR/S6K without affecting the activation of either Akt or AMPK ( Figs. 5A, B). Finally, AMPK downregulation with compound C or shRNA mimicked the inhibitory effects of DEBC on the activation status of Akt and mTOR/S6K in differentiating hDP-MSC at day 5 ( Figs. 5A, C), indicating AMPK as an upstream signal for Akt activation and subsequent increase in mTOR/S6K activity. These data demonstrate that the optimal osteogenic transformation of hDP-MSC requires AMPK-dependent phosphorylation of Akt and consequent activation of mTOR at the latter stages of differentiation. The present study demonstrates a central role of the intracellular energy sensor AMPK in the osteogenic differentiation program of hDP-MSC.