Figure 7.

MSX2 Directly Targets TWIST1 during Mesenchymal Differentiation

(A) Western blotting analysis of exogenous TWIST1 or GFP in H1-WT and MSX2-deleted H1 cells at day 7 of MSC induction under SMs conditions. GAPDH was used as a loading control.

(B and C) FCM analysis (B) and qRT-PCR analysis (C) of indicated MSC markers in H1 and MSX2-deleted H1 cells without or with TWIST1 overexpression at day 7 of MSC induction under SMs conditions (mean ± SEM, N = 3). ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001.

(D) Relative luciferase activity in GFP-MSX2 H1 hESCs transfected with pGL3 construct containing TWIST1 promoter (pTWIST1-0.7kb-LUC) ± DOX (3 μg/mL) for 72 hr (mean ± SEM, N = 3). ^∗p < 0.05. Values are normalized to the pGL3 group (=1).

(E) ChIP-qPCR analysis of the occupancy of MSX2 on the two potential MSX2-binding sites (MBS1, MBS2) of TWIST1 promoter in GFP-MSX2 H1 hESCs with DOX (3 μg/mL) for 72 hr. Non-specific immunoglobulin G was used as isotype control. Values are normalized to those of their corresponding input samples (mean ± SEM, N = 3). ^∗p < 0.05; NS, not significant.

(F) Relative luciferase activity in GFP-MSX2 H1 hESCs transfected with WT or MSX2-binding site mutated (MBS1-mut, MBS1-mut, MBS1/2-mut) TWIST1 promoter-luciferase reporter constructs with DOX (3 μg/mL) for 3 days. A non-specific mutant in TWIST1 5′ flanking region was used as a negative control (NC-mut). Normalized to the cells transfected with pGL3 (=1) (mean ± SEM, N = 3). ^∗p < 0.05; ^∗∗p < 0.01; NS, not significant. Values are normalized to the pGL3 group (=1).

(G) Schematic model for efficient hPSC-MSC induction and the underlying mechanism. Based on MSX2 and specific SMs cocktail, hPSCs can be directly programmed into MC-MSCs through an NCC intermediate stage. During the process, MSX2 upregulates the expressions of PRAME and TWIST1. Furthermore, TWIST1 serves as a key direct target of MSX2 and mediates its programming function.

See also Figure S7.