Fig 2: ADGRG6 enhancer mutations are associated with poor survival and tumor angiogenesis in UBC. a Sanger sequencing detected ADGRG6 enhancer mutations in the additional 196 UBCs. b Immunohistochemical staining of UBC tumors displaying consequences of ADGRG6 enhancer mutations. ADGRG6 enhancer mutations lead to ADGRG6 protein overexpression in tumor tissues. The three different groups of UBC tumors: G > A Mut (n = 25), ADGRG6 enhancer only with 142,706,209 G > A mutation; C > T Mut (n = 25), ADGRG6 enhancer only with 142,706,206C > T mutation; No Mut (n = 25), ADGRG6 enhancer with no mutation. c CD31 immunostained microvessel in the additional 196 UBCs. The three different groups of UBC tumors: G > A Mut (n = 70), ADGRG6 enhancer with 142,706,209G > A mutation; C > T/G Mut (n = 51), ADGRG6 enhancer with 142,706,206 C > T/G mutation; No Mut (n = 94), ADGRG6 enhancer with no mutation. The three horizontal lines represent mean ± SD for the subjects. *P < 0.05, **P < 0.01, ***P < 0.001 vs. No Mut group, two-sided unpaired t test with Welch’s correction (b) or Mann–Whitney test (c). d Kaplan–Meier survival curves show that the patients with the ADGRG6 enhancer mutations were significantly associated with a shorter overall survival than those with no mutation in both nonmuscle-invasive bladder cancer (NMIBC) and muscle-invasive bladder cancer (MIBC) subcohorts. Statistical significance was determined by log-rank test. e, f Knockdown of ADGRG6 in 5637 and SW780 cells compromised their abilities to recruit human umbilical vein endothelial cells (e) and induce tube formation (f). Scale bars, 100 µm. Error bars represent the SEM. The data shown represent averages from three independent experiments and were statistically analyzed by two-sided t test. *P < 0.05, **P < 0.01, ***P < 0.001 vs. siCon

Fig 3: Ablation of Adgrg6 in dense connective tissues leads to late-onset scoliosis and compromised biomechanical properties of the tendons.(A–D) Representative X-ray images of Cre (-) control and Scx-Cre; Adgrg6f/f mutant mice. Scx-Cre; Adgrg6f/f mice were analyzed at P40 (A, A’) and P120 (B, B’). Cre (-) control and Cre (+) control mice analyzed at P120 are shown in (C) and (D), respectively. Scoliosis is indicated with red arrows in (B, B’). (E, F) Longitudinal analyses of Cobb angle values of Cre (-) control mice and Scx-Cre; Adgrg6f/f mice at P40 and P120. Cre (-) control mice, n = 9 mice at both P40 and P120; Scx-Cre; Adgrg6f/f mutant mice, n = 9 at both P40 and P120. Thresholds of scoliosis (Cobb angle >10°) are indicated with two red dot lines. (G–N) Biomechanical characterization of Cre (-) control and Scx-Cre; Adgrg6f/f mutant tendons at 12 weeks. Representative phase-contrast images of tail fascicles isolated from Cre (-) control and Scx-Cre; Adgrg6f/f mutant mice are shown in (G) and (H). Quantification of fascicles’ cross-sectional area (CSA) (I), initial length (J), elastic modulus (E Modulus) (K), failure force (L), failure strain (M), and failure stress (N) are also shown. n = 8–9 fascicles isolated from three mice. For (I) and (K–N), n = 8–9 fascicles (violin plots) isolated from three different mice (close circles). For (J), n = 8 and 9 fascicles isolated from Cre (-) control mice or Scx-Cre; Adgrg6f/f mutant mice, respectively. Bars are plotted with mean and 95% CI. Cross-sectional area (I) (mean: Cre [-] control: 0.008 [95% CI 0.0069, 0.0095]; Scx-Cre; Adgrg6f/f: 0.013 [95% CI 0.0094, 0.0169]) and elastic modulus (K) (mean: Cre [-] control: 1060 [95% CI 863.8, 1256]; Scx-Cre; Adgrg6f/f: 810.9 [95% CI 644.6, 977.3]) were significantly different between control and mutant groups. The statistical difference is evaluated by unpaired t-test with Welch's correction for (J, K, L, N), and with Mann–Whitney test for non-normally distributed data (I, M). The p-value for each comparison is shown. ns: not significant. Scale bars: 10 mm in (A, A’); 100 μm in (G, H). Figure 7—source data 1.Cobb angle measurements and biomechanical testing of mice with Adgrg6 ablation in dense connective tissues.

Fig 4: β-Galactosidase staining of Rosa26-lsl-lacZ reporter mice recombined with different Cre strains.(A–C”) Fluorescent in situ hybridization (FISH) analysis of Adgrg6 on thoracic spine sections of wild-type mice at P1. Adgrg6 signal is detected in the GP, NP, and EP (B’, B”). Adgrg6 is also expressed in TB, AF (A, C’, C’’), and the outmost AF (white arrows, C'); n = 3 mice for each group. (D) Schematic of Cre targeting outlined in this study. Generally, Bglap-Cre targets bony tissues, ATC targets cartilaginous tissues, Scx-Cre targets dense connective tissues, and Col2a1-Cre targets all these tissues. (E–F’) β-galactosidase staining of Col2a1-Cre; Rosa26-lsl-lacZ spine sections at P1 (E, E’) and P28 (F, F’). Col2a1-Cre targets NP, AF, EP, and GP, as well as some cells in the bony tissues at both P1 and P28. Note that Col2a1-Cre also targets the outmost AF (black arrows, F’) and the periosteum (Po) (green arrows, F’) at P28. (G–H’) β-Galactosidase staining of Bglap-Cre; Rosa26-lsl-lacZ spine sections at P1 (G, G’) and P28 (H, H’). Bglap-Cre targets TB at P1 (G, G’), but targets TB (H’), CB (H’), and Po (green arrows, H’) at P28. Bglap-Cre also targets hypertrophic cells within the EP and GP (H’). (I–J’) β-Galactosidase staining of ATC; Rosa26-lsl-lacZ (Dox-induced from E0.5–P20) spine sections at P1 (I, I’) and P28 (J, J’). ATC targets most cells in NP, AF, EP, and GP at both time points. Note that ATC did not target the outmost AF (black dash line and arrows, J’), nor the Po (green arrows, J’). (K–L’) β-Galactosidase staining of Scx-Cre; Rosa26-lsl-lacZ spine sections at P1 (K, K’) and P28 (L, L’). Scx-Cre targets AF at P1 (K’) and P28 (L’), and also recombines in several cells of NP, EP, and GP (L’) at P28. Note that Scx-Cre also targets some cells of the outmost AF (black arrows, L’) and the Po (green arrows, L’) at P28. An illustration (M) and a bright-field image (M’) of the dorsal side of a wild-type mouse show the supraspinous ligaments (black arrows, M, M’). (N–Q’) Whole-mount β-galactosidase staining of Rosa26-lsl-lacZ reporter mice at P1. Dorsal view of Col2a1-Cre; Rosa26-lsl-lacZ mouse is shown in (N, N’). Supraspinous ligaments are indicated with black arrows (N’). Dorsal view of ATC; Rosa26-lsl-lacZ mouse (Dox-induced from E0.5) is shown in (O, O’). Facet joints are indicated with red arrows (O’). Dorsal and sagittal views of the whole-mount Scx-Cre; Rosa26-lsl-lacZ mice are shown in (P, P’) and (Q, Q’), respectively. Supraspinous ligaments are indicated with black arrows (P’, Q’). n = 3 mice in each group. Scale bars: 100 μm in (A–C) and (E–L’); 1 mm in (M’–Q’). AF: annulus fibrosis; EP: endplate; GP: growth plate; NP: nucleus pulposus; TB: trabecular bone; CB: cortical bone; Po: periosteum; Lg: ligament; Td: tendon.

Fig 5: Loss of Adgrg6 in mature osteoblast lineages is dispensable of adolescent idiopathic scoliosis (AIS) development.(A–C) Representative X-ray images of Cre (-) control (A), Cre (+) control (B), and Bglap-Cre; Adgrg6f/f mutant (C) mice at P120. (D, E) Longitudinal analyses of Cobb angle values of Cre (-) control mice (D) and Bglap-Cre; Adgrg6f/f mice (E) at P40 and P120. Cre (-) control mice, n = 8 mice at P40 and P120; Bglap-Cre; Adgrg6f/f mutant mice, n = 8 and 7 at P40 and P120, respectively. Thresholds of scoliosis (Cobb angle >10°) are indicated with two red dot lines. No Bglap-Cre; Adgrg6f/f mice showed scoliosis at P40 (0/8) or P120 (0/7) (E). (F–J) MicroCT scanning of the thoracic region of the spine shows normal morphology of the vertebral bodies in both Cre (-) controls (F) and the Bglap-Cre; Adgrg6f/f mice (G). Transverse sections of the microCT three-dimensional reconstruction of the thoracic vertebral body show a comparable bone mass in the control (H) and mutant mice (I). The bone volume per total volume (BV/TV) of the control and mutant mice is shown in (J). n = 4 mice for each group. Bars are plotted with mean and 95% CI. The statistical difference is evaluated by a two-tailed Student's t-test. The p-value is shown. ns: not significant. (K, L) Real-time RT-PCR analysis of RNA isolated from long bone shows that the expression of Adgrg6 is very low in bony tissues compared with the expression of Col1a1 (K). However, the expression of Adgrg6 was efficiently knockdown in Bglap-Cre; Adgrg6f/f mice (L). RNA was isolated and pooled from three mice of each experimental group. Bars are plotted with mean and SD. The statistical difference is evaluated by a two-tailed Student's t-test. The p-value for each comparison is shown. Scale bars: 10 mm in (A); 1 mm in (G) and (I). Figure 5—source data 1.Characterization of mice with Adgrg6 ablation in mature osteoblast lineages.