Biomechanical stress regulates mammalian tooth replacement via the integrin β1-RUNX2-Wnt pathway

Abstract

Renewal of integumentary organs occurs cyclically throughout an organism's lifetime, but the mechanism that initiates each cycle remains largely unknown. In a miniature pig model of tooth development that resembles tooth development in humans, the permanent tooth did not begin transitioning from the resting to the initiation stage until the deciduous tooth began to erupt. This eruption released the accumulated mechanical stress inside the mandible. Mechanical stress prevented permanent tooth development by regulating expression and activity of the integrin β1-ERK1-RUNX2 axis in the surrounding mesenchyme. We observed similar molecular expression patterns in human tooth germs. Importantly, the release of biomechanical stress induced downregulation of RUNX2-wingless/integrated (Wnt) signaling in the mesenchyme between the deciduous and permanent tooth and upregulation of Wnt signaling in the epithelium of the permanent tooth, triggering initiation of its development. Consequently, our findings identified biomechanical stress-associated Wnt modulation as a critical initiator of organ renewal, possibly shedding light on the mechanisms of integumentary organ regeneration.

Keywords: Wnt signaling; biomechanics; organ replacement; stress.

Figure 1. Morphology and molecular map of the permanent canine germ during initiation stages in swine

1. A–G

   Hematoxylin and eosin (H&E) staining of the frontal sections of miniature pig mandible slices showing morphological changes during the permanent canine (PC) initiation stage from embryonic day 50 (E50) to postnatal day 10 (PN10); (A’–F’) are magnifications of boxed regions in their corresponding figure panel. DC, deciduous canine; PC, permanent canine; SDL, successional dental lamina; n = 3.

2. H

   In situ hybridization (ISH) showing the expression pattern of Sox2 in the initiation stage from E50 to E90; right figure panels are magnifications of boxed regions in left panels. Dashed lines mark the position of the successional dental lamina (SDL); green arrowhead indicates positive staining of Sox2 at the tip of the SDL. n = 3.

3. I

   Immunofluorescence (IF) of Ki67 at the SDL at E60, E70, and E90; right figure panels are magnifications of boxed regions in left panels. n = 3.

4. J

   Semi‐quantification and comparison of Ki67 expression levels during E60, E70, and E90 stages. n = 3.

5. K

   TUNEL assay showing the apoptosis status of the SDL at E60, E70, and E90; right figure panels are magnifications of boxed regions in left panels. n = 3.

6. L

   Diagram illustrating the initiation stage at E50 (attached SDL), E60 (detached SDL), E90 (bud stage), and PN10 (bell stage). OEE, outer enamel epithelium; IEE, inner enamel epithelium.

Data information: Data represent the means ± SEM. *P < 0.05 (one‐way ANOVA and Newman–Keuls post hoc tests). Scale bars = 100 μm.

Figure EV1. Immunostaining and in situ hybridization (ISH) of the PC primordium from E50 to E90

1. A

   Immunohistochemistry (IHC) of pan‐cytokeratin from embryonic days 50 (E50) to E90 showing the dual layers of the epithelium in dental lamina and enamel organ.

2. B

   Immunofluorescence (IF) of pan‐cytokeratin from E50 to E90 showing a similar pattern.

3. C–E

   ISH of Shh, Pitx2, and Pax9 during the initiation stage from E50 to E90. Dashed lines mark the position of the SDL or PC.

Data information: n = 3 for all panels. Right figure panels are magnifications of the boxed regions in corresponding left panels. Scale bar = 100 μm.

Figure 2. Differential growth rates of deciduous canine (DC) and alveolar socket and mechanical stress inside the mandible

1. Three‐dimensional reconstruction of serial H&E frontal sections of miniature pig mandibles at embryonic day 60 (E60) and day E90; deciduous canine (DC) in purple, permanent canine (PC) in yellow, and alveolar socket in green. The red, blue, and green arrows indicate the width of the labial alveolar socket, lingual alveolar socket, and DC, respectively. n = 3.

2. DC and labial and lingual alveolar socket widths in the horizontal plane at the bottom of the PC during E60 and E90. n = 3.

3. The proportion of DC width relative to the total alveolar socket width during E60 and E90. n = 3.

4. Micro‐CT imaging of the whole mandible at E60; boxed region is magnified in D’ (top panel). (D’) Mandible slice isolated with Geomagic software. White dashed lines mark DC.

5. 3‐D color map (left) after alignment of mandible slices before and after (sham) surgery showing a comparison of the surface points. Coronal sections through cusp tips (transparent squares, left) were selected for 2‐D comparisons (middle); right panels are magnifications of yellow‐boxed regions. The solid purple contour and dotted black contour indicate the pre‐ and post‐surgery shapes, respectively. The distance between the two contours is the colored line segments showing the distance and direction of the movement. The colored ball in the 2‐D comparison marks the position of the maximum displacement. The PC position is indicated in pink. n = 3.

6. Dissected mandible slice with a “U” shape.

7. In the cup model, the mandible slice was simplified as a cup according to its dimensions (unit, mm). Red arrows indicate the uniform force (stress) exerted on the inner wall of the mandible. The bottom of the cup was fixed to avoid rigid body motions.

8. 3‐D color map shows the extent of deformation based on the cup model established with ANSYS software.

9. Probable range of mechanical stress inside the mandible evaluated by multiple simulation tests in which serial stress and Young's modulus were inputted into the cup model; the cup was set with the outer surface fixed (left, d _r = 0 denotes that radial deformation of the outer surface equals 0) or the outer surface free (right).

10. Deformation of the mandible walls under a series of stress levels with different Young's moduli. Data were obtained from multiple tests as in (I). Gray horizontal lines indicate upper and lower limit values of mean mandible wall displacement (79.74 and 36.48 μm, respectively); dashed colored lines indicate results of the free outer surface; solid colored lines indicate results of the fixed outer surface (with consideration of good linearity of the simulation results, the corresponding result points of the simulation series were omitted and replaced by solid or dashed colored lines for clarity); the actual stress value should be between the two extreme boundary conditions. Results show that the probable stress level ranged from 3 to 20 kPa. n = 3.

Data information: Data represent the means ± SEM. Unpaired t‐tests, *P < 0.05, **P < 0.01, ***P < 0.001.

Figure EV2. Mechanical test and establishment of the cup model

1. A–D

   Testing Young's moduli of the samples. (A) Mandible bony piece (mandible side wall piece) prepared for the test (0.2–0.3 mm in thickness). (B) Diagram of the testing principle of nanoindentation for Young's modulus determination using the Piuma Chiaro Nanoindenter. The indenter (tip) applies force onto the sample (mandible bony piece in A) and records the force applied and the indentation depth (sample deformation). (C) Normal force spectrum from which valid experimental data were extracted. The red line signifies that the probe was engaging before reaching max deformation (Dmax); blue line indicates that the probe retreated after reaching Dmax. (D) Distribution of the typically measured Young's moduli of the samples. Red line indicates the fitting results of the Gaussian model.

2. E–G

   Gradient densities of the mesh (sparse, medium, and dense) were tested to determine the optimal mesh density of the cup for the computation. (E, E’) The deformation is not symmetrical although it is clear in the sparse group. (F, F’ left, G, and G’) The medium and dense groups show similar deformation after exerting a force on the inner wall. (F’) Series of Poisson's ratios, including 0.15 (not shown), 0.35 (left), and 0.48 (right), were tested. Dmax values of the three cases were 1, 1.21, and 1.37 after normalizing the data.

Figure 3. Biomechanical stress determines initiation timing of PC development in vitro

1. A

   Diagram of the dissected miniature pig mandible slice illustrating the deciduous (DC) and permanent (PC) canine and application of compression in vitro using Flexcell FX‐5000 Compression System.

2. B–D

   H&E staining of canine frontal sections from embryonic day 60 (E60) (B), after culturing for 2 days without stress (0 kPa) (C), or with stress (3 kPa) (D). (B’–D’) are magnifications of boxed regions in their corresponding figure panels. Scale bars = 50 μm. n = 6/6 in E60 group; n = 12/12 in 0‐kPa group; n = 4/8 in 3‐kPa group.

3. E

   Diagram illustrating the required time during PC initiation; regular process from resting to initiation stage (left) and when the mandible is cultured without stress (0 kPa; middle) or with stress (3 kPa; right).

4. F

   RT–qPCR of several mechanosensitive proteins (TGFB1, NFKB1, EGR1, and RUNX2) and of four mechanoreceptors (ITGB1, ITGB3, ITGAV, and ITGA2) in PC and surrounding soft tissue at E60 and after culturing for 2 days with (3 kPa) or without extra mechanical stress (0 kPa). n = 3 for each group. Data represent the means ± SEM. One‐way ANOVA (Newman–Keuls test for post hoc comparisons between two groups), *P < 0.05, **P < 0.01, ***P < 0.001; NS, not significant.

Figure EV3. Morphological comparisons of cultured mandible slices subjected to different pressures

1. Morphological comparison of 3D reconstructions from a series of H&E sections of the mandible cultured under 0, 1, 3, 10, and 20 kPa. The deciduous canine (DC) is colored purple, the permanent canine (PC) germ is colored yellow, and the mandible is colored green. Arrows in yellow, red, and blue indicate the DC, inside alveolar, and outside alveolar widths, respectively.

2. Quantitative comparison of the three parameters among five groups. n = 3 for all groups. Data represent the mean ± SEM. One‐way ANOVA (Newman–Keuls test for post hoc comparisons between two groups), *P < 0.05; **P < 0.01; NS, not significant.

Figure 4. Biomechanical stress regulates the integrin β1‐ERK1‐RUNX2 pathway in mesenchyme between DC and PC

1. A–L

   IF of integrin β1, ERK1, and RUNX2 in miniature pig canine frontal sections at embryonic days 60 (E60), E90, and E60 cultured under 3 kPa stress for 2 days and E60 cultured under 0 kPa stress for 2 days; (A’–L’) are magnifications of boxed regions in the corresponding figure panels.

2. M

   Relative IF expression levels of integrin β1, ERK1, and RUNX2 during E60, E90, and E60 cultured under 3 kPa stress for 2 days and E60 cultured under 0 kPa stress for 2 days.

3. N

   Morphology of the human PC at weeks 18–19 (H&E staining); (N’) is magnification of the boxed region in (N). DC, deciduous canine; PC, permanent canine; IEE, inner enamel epithelium; OEE, outer enamel epithelium. Mesenchyme between DC and PC is indicated with a green arrow.

4. O–Q

   ISH expression patterns of ITGB1, ERK1, and RUNX2 in human PCs. Green arrows indicate the mesenchyme between DC and PC.

5. R

   IF of RUNX2 in human PC. Mesenchyme between DC and PC is indicated with a red arrow. Red arrow indicates the mesenchyme between DC and PC.

6. S

   PC epithelium areas in E60, E90, and E60 cultured under 3 kPa stress for 2 days and E60 cultured under 0 kPa stress for 2 days.

7. T

   Illustration of force exertion on cultured dental follicle cells (DFCs).

8. U

   IF of RUNX2 and phospho‐ERK1/2 (p‐ERK1/2) upon force loading with 0 or 1.0 g/cm² for 2 h or after force was removed and the cells cultured for an additional 1, 2, and 4 h.

9. V

   Relative IF expression levels of RUNX2 and p‐ERK1/2 between groups in (U).

10. W

    Western blots of p‐ERK1/2 and RUNX2 levels after 1.0 g/cm² was applied for 2 h, with or without anti‐integrin β1 antibody and ERK1 inhibitor (U0126).

11. X

    Relative expression levels of p‐ERK1/2, RUNX2, and Actin in (W).

Data information: Data represent the means ± SEM. n = 3. Scale bars = 25 μm (A’, G’) and 50 μm (other panels). Unpaired t‐tests for (M); one‐way ANOVA (Newman–Keuls test for post hoc comparisons between two groups) for( V and X), *P < 0.05, **P < 0.01, ***P < 0.001; NS, not significant.Source data are available online for this figure.

Figure EV4. Expression patterns of RUNX2 in the third deciduous incisor during resting stage of E60

1. A, B

   The expression pattern of RUNX2 in the permanent third incisor at E60 shows similarities with that of the PC via IHC and ISH assays. The boxed regions are magnified in (A’ and B’). Dashed lines mark the epithelium of PC. n = 3. Scale bars = 100 μm.

Figure 5. RUNX2 overexpression inhibits initiation of PC germs

1. A–C

   H&E staining of miniature pig canine frontal sections from embryonic day 60 (E60 group) (A) and after the mandible slices were infected with RUNX2 overexpression lentiviral vector with a Myc tag (overexpression group) (B) or overexpression control lentiviral vector with a Myc tag (O‐control group) (C); (A’–C’) are magnifications of boxed regions in the corresponding figure panels.

2. D–F

   Immunohistochemical (IHC) staining of Myc expression in mandible slices from the E60, overexpression, and O‐control groups.

3. G–I

   IF staining of RUNX2 expression (yellow arrows) in mandible slices of the E60, overexpression, and O‐control groups.

4. J, K

   Relative expression levels of Myc and RUNX2 in the E60, O‐control, and overexpression groups.

5. L, M

   IF staining of RUNX2 expression after infecting mandible slices with scrambled shRNA lentiviral vector (knockdown control, K‐control) (L) or RUNX2 shRNA lentiviral vector (knockdown) (M); (L’–M’) are magnifications of boxed regions in the corresponding figure panels.

6. N

   Relative RUNX2 expression levels between K‐control and knockdown groups.

7. O–Q

   H&E staining of mandible slice sections from the K‐control, knockdown + force (3 kPa), and K‐control + force (3 kPa) groups. (O’–Q’) are magnifications of boxed regions in the corresponding figure panels.

8. R–T

   IF staining of RUNX2 expression (yellow arrows) of mandible slice sections from the K‐control, knockdown + force (3 kPa), and K‐control + force (3 kPa) groups. (R’–T’) are magnifications of boxed regions in the corresponding figure panels.

9. U

   Relative RUNX2 expression levels among the three groups in (R–T).

10. V

    PC epithelium areas in the overexpression, O‐control, knockdown, knockdown + force, K‐control, and K‐control + force groups.

Data information: Data represent the means ± SEM. n = 3 for all experiments. Scale bars = 100 μm (D–F), 25 μm (R’–T’), and 50 μm (other panels). Dashed lines mark the outlines of PCs. Unpaired t‐tests for (N and V); one‐way ANOVA (Newman–Keuls test for post hoc comparisons between two groups) for (J, K, and U), *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 6. Wnt modulation between the mesenchyme and epithelium regulates PC initiation

1. A, B

   ISH of miniature pig canine frontal sections showing expression patterns of RUNX2 at embryonic days 60 (E60) and E90. (A’, B’) are magnifications of boxed regions in the corresponding figure panels.

2. C

   Relative ISH expression levels of RUNX2 in epithelium (Epi.) and mesenchyme (Mes.) at E60 and E90.

3. D, E

   ISH of miniature pig canine frontal sections showing expression patterns of β‐catenin at E60 and E90. (D’, E’) are magnifications of boxed regions in the corresponding figure panels.

4. F

   Relative ISH expression levels of β‐catenin in epithelium (Epi.) and mesenchyme (Mes.) at E60 and E90.

5. G, H

   IF of Lef1 at E60 and E90. (G, H’) are magnifications of boxed regions in the corresponding figure panels.

6. I

   Relative IF expression levels of Lef1 in epithelium (Epi.) and mesenchyme (Mes.) at E60 and E90.

7. J

   FISH of β‐catenin in E60 and E90 mandible slices and E60 mandible slices after 12 and 24 h in culture via RNAscope. Boxed regions of mesenchyme are magnified in the right panels. Epithelium is indicated with a yellow arrow.

8. K

   Relative FISH expression levels of β‐catenin in the nucleus and cytoplasm of epithelium and mesenchyme at E60, E90, and E60 cultured for 12 and 24 h.

9. L, M

   IF of β‐catenin at E60 and E90; the yellow arrows indicate the localization of β‐catenin in the nucleus.

10. N

    Relative IF expression levels of β‐catenin in the nucleus of epithelium and mesenchyme at E60 and E90.

11. O, P

    FISH of β‐catenin in mandible slices infected with RUNX2 shRNA lentivirus (knockdown) or scrambled shRNA lentivirus (scramsh) via RNAscope.

12. Q

    Relative FISH expression levels of β‐catenin in the nucleus and cytoplasm of epithelium and mesenchyme in knockdown and scramsh groups.

Data information: Scale bars = 50 μm. Data represent the means ± SEM. n = 3 for all experiments. Unpaired t‐tests for (C, F, I, N, and Q); one‐way ANOVA for (K), *P < 0.05, **P < 0.01, ***P < 0.001.

Figure EV5. The Wnt/β‐catenin signaling pathway regulates permanent canine initiation

1. A–A’’

   Wnt inhibitor Sfrp1 expressed in the outside layer of the PC SDL from E50 to E90.

2. B–B’’

   Wnt inhibitor Sostdc1 expressed in the inside layer of the SDL from E50 to E90, complementary to the pattern of Sfrp1.

3. C–E

   Morphological comparisons between the control, LiCl, and Dkk1 groups showing initiation of PC in the control and LiCl groups, but inhibition in the Dkk1 group. The boxed regions are magnified in the lower panels (C’, D’, and E’).

4. F–H

   IF of Lef1 showing strong positive signals in the enlarged PC epithelium and surrounding mesenchyme in LiCl and control groups, and a weak signal in the Dkk1 group.

Data information: n = 3 for all experimental groups. Dashed lines mark the successional dental lamina or epithelium of PC. Scale bars = 100 μm (A–A’’ and B–B’’), 50 μm (C–H), and 25 μm (C’–E’).

Figure 7. RUNX2 regulates Wnt signaling in DFCs

1. A, B

   IF of Lef1 in mandible slices at embryonic day 60 (E60) infected with RUNX2 overexpression lentiviral vector (overexpression) or overexpression control lentiviral vector (O‐control). (A’–B’) are magnifications of boxed regions in the corresponding figure panels. Dashed lines mark the epithelium of PC.

2. C

   Relative IF expression levels of Lef1 in epithelium (Epi.) and mesenchyme (Mes.) of the overexpression and O‐control groups.

3. D, E

   IF of Lef1 in E60 mandible slices subjected to 3 or 0 kPa stress for 2 days. (D’, E’) are magnifications of boxed regions in the corresponding figure panels. Dashed lines mark the epithelium of PC.

4. F

   Relative IF expression levels of Lef1 in the epithelium and mesenchyme after applying 3 and 0 kPa pressure.

5. G

   Western blots of Lef1, non‐phospho‐β‐catenin, and RUNX2 after DFCs were infected with control lentiviral vector or RUNX2 overexpression lentiviral vector. (G’) Relative expression levels between control vector and RUNX2 overexpression groups.

6. H

   Western blots of Lef1, non‐phospho‐β‐catenin, and RUNX2 after DFCs were infected with scrambled shRNA (scramsh) or RUNX2 knockdown shRNA (knockdown) lentiviral vectors. (H’) Relative expression levels between scramsh and knockdown groups.

7. I

   Western blots of nuclear non‐phospho‐β‐catenin and Lamin B in DFCs of RUNX2 overexpression (O), control vector (V), RUNX2 knockdown (K), scramsh (S), compressed force (F; 1.0 g/cm²), and control (C; 0 g/cm²) groups. Sample was loaded in a gradient (7.5, 15, 30, and 60 μg) showing a linear relationship for each group.

8. J

   Western blots of nuclear non‐phospho‐β‐catenin and Lamin B in DFCs treated with IWR‐1‐endo. Relative expression levels were compared between force and no force groups and between overexpression and control groups with or without IWR‐1‐endo treatment.

9. K

   Diagram illustrating the biomechanical stress regulation of Wnt/β‐catenin signaling in the mesenchyme between the deciduous (DT) and permanent tooth (PT) via the integrin β1‐ERK1‐RUNX2‐Wnt/β‐catenin pathway.

10. L

    Diagram illustrating the biomechanical stress‐associated downregulation of RUNX2‐Wnt/β‐catenin pathway in the mesenchyme, inducing upregulation of Wnt signaling in the epithelium, which triggers PT development.

Data information: Data represent the means ± SEM. Scale bars = 100 μm. n = 3 for all experiments. Unpaired t‐tests for (C, F, G’, and H’); one‐way ANOVA (Newman–Keuls test for post hoc comparisons between two groups) for (J), *P < 0.05, **P < 0.01, ***P < 0.001; NS. not significant.Source data are available online for this figure.