Fig 2: Loss of Piezo1/2 reduced Yap1 and Wnt/Ctnnb1 signaling activities and osteoblast differentiation in vivo and in vitro.(a) Immunofluorescent staining of Ctnnb1 and Yap1/Wwtr1 in the humerus sections from the P0 Prrx1-Cre driven Piezo1 and Piezo1/2 mutants and littermate controls. (b) Western blot analyses of P0 femur bone tissue lysates from P0 pups. (c) qPCR analyses of gene expression from femur bone tissues of P0 pups (n = 3, means ± SD). (d) Ccn2 expression by RNAScope analysis of humerus sections of E16.5 embryos. (e) Alkaline phosphatase staining (left) and qPCR analyses (right, n = 3, means ± SD) in the indicated BMSCs infected with Ad-GFP or Ad-Cre 5 days after osteogenic induction. (f) Alkaline phosphatase staining (left) and qPCR analyses (right, n = 3, means ± SD) in Piezo1f/f;Piezo2f/f BMSCs treated with Piezo1 agonist (Yoda1) and antagonist (Gd3+) 6 days after osteogenic induction. The schematics of the induction process is shown on the top. All scale bars: 100 µm. DAPI (blue) stained the nucleus. *p<0.05, **p<0.01, ***p<0.001, two-tailed unpaired Student’s t-test (e) or one-way ANOVA followed by Tukey’s multiple comparisons tests when ANOVA was significant (c and f) (Figure 6—source data 1).Figure 6—source data 1.Original numbers and Western blots.

Fig 3: ALDH1B1 and CTNNB1 gene expression show positive correlation in cervical cancer patients. (A,B) Overall survival curves obtained from cervical squamous cell carcinoma and endocervical adenocarcinoma patients. (C) Pair-wise correlation between ALDH1B1 and CTNNB1 using the Spearman method. Graphs were obtained from GEPIA database.

Fig 4: Model simulation and experimental validation of Wnt-pathway activation upon GSK3 inhibition.(A) β-catenin (referred to by its official gene name CTNNB1 in the figure) token levels predicted by our model with initial GSK3 token levels ranging from 0 to 5. For GSK3 = 3, 4 or 5, we observed a flat β-catenin response. For GSK3 = 0, 1 or 2 β-catenin increases to low, moderate or high levels, respectively. (B) Graph combining the results from panels C and D to allow easy comparison to the modeling results depicted in panel (A), showing dose- and time-dependent activation of a Wnt/β-catenin responsive TCF/LEF luciferase reporter in HEK293TWOO cells. For all curves with black data points (corresponding to panel C), luciferase activity was plotted relative to the vehicle control (not shown), which was set at 1 for each of the three time points (3, 8 and 24 hours). For the curve with white data points (corresponding to panel D), luciferase activity was plotted relative to the vehicle control, which was set at 1 for the t = 0 hours condition. (C) Reporter assay in HEK293TWOO cells, showing dose-dependent activation at 3, 8 and 24 hours after stimulation with CHIR99021 (same concentrations as depicted in B). (D) Reporter assay in HEK293TWOO cells, showing time-dependent activation upon treatment with 3 μM CHIR99021. Values were plotted relative to the DMSO control, which was set at 1 for t = 0 hours. (E) Western blot from the experiment depicted in (D), showing total and active (non-phosphorylated) β-catenin levels. Tubulin was used as a loading control. (F) Quantification of the Western blot shown in (E). Total and active β-catenin levels were normalized to tubulin. The increase in either total or active β-catenin levels was plotted relative to time point 0, for which the normalized levels were set to 1. Experiments were repeated two (C) or three (D-F) times. A representative experiment is shown.

Fig 5: Comparison of analysis of short and long FCS traces.Autocorrelation curves were made from nine 10s portion of the measurement (trace), or a single 90s trace from the same FCS measurement of SGFP2-CTNNB1 and subsequently fitted with the same fitting model. The graphs depict the fitted values (dots) and confidence intervals determined by the goodness of fit in the FCS analysis software (lines) for the particle number (N, panel A) and the fraction of slow component (F2, panel B) of the 10s and 90s.