Changes to the calvarial bone, periosteum and suture mesenchyme, and bone repair with aging and interventions

(A) Parietal Bone Volume/Total Volume (BV/TV) and (B) representative transverse micro computed tomography (microCT) sections in Young, Adult, and Aged mice. Statistical analysis: Welch’s test in n = 10 mice per condition. (C) Mean cross sectional area taken up by Endomucin (Emcn)-positive vascular cells and (D) representative immunofluorescence staining of Emcn (magenta) and DAPI (grey) in Young, Adult, and Aged mouse periosteum. Statistical analysis: Welch’s tests in n = 5 mice per condition. (E) Nuclear aspect ratio (long axis divided by short axis) of cells within the periosteum and suture mesenchyme and (F) representative DAPI-stained nuclei in Young, Adult, and Aged mice. Statistical analysis: Welch’s tests in n = 5 mice per condition. Green and red arrows highlight rounded and elongated nuclei, respectively. (G) Mean Phalloidin-488 fluorescence intensity and (H) representative images for filamentous actin (F-actin, green) and DAPI (grey) in Young, Adult, and Aged mouse periosteum and suture mesenchyme. Statistical analysis: Welch’s tests in n = 5 mice per condition. (I) Mean intensity and (J) representative images of live TMRM staining (red) and Hoechst (grey) in Young, Adult, and Aged mouse periosteum ex vivo. Statistical analysis: Welch’s tests in tissue from n = 3 mice per condition. (K) Percentage of CD90+ cells within the periosteum and suture mesenchyme, expressed relative to the total cellular number (DAPI+) in Young, Adult, and Aged mice. Representative images (N) show CD90 immunostaining (magenta) and DAPI (grey). Statistical analysis: Welch’s tests in n = 5 mice per condition. (M) Percentage of CD90+ cells within calvarial defects treated with Wnt3a-bandages (blue points) or inactive controls (white points), expressed relative to the total cellular number (DAPI+) in Young, Adult, and Aged mice. (N) Representative images show CD90 immunostaining (magenta) and DAPI (grey). Statistical analysis: ratio-paired two-tailed t tests in n = 5 mice per condition. (O) Percentage of repair and (P) representative top-down microCT images within calvarial defects treated with Wnt3a-bandages (blue points) or inactive controls (white points). Statistical analysis within age groups: ratio- paired two-tailed t tests; between age groups and Aged mice following the IF protocol or NMN or Akk supplementation: Welch’s tests; Young, Aged, Aged + IF, Aged + NMN, n = 10 mice; Aged + Akk, n = 9; Adult, n = 8. Complete statistical analysis between aged groups and treatments can be found in Supplementary Table 1. The yellow dashed circle highlights the 2 mm diameter defect size.

Transcriptomic and protein expression characterization of CD90+ cells during aging

(A) UMAP-plot showing the distribution of a subpopulation of CD90+ periosteal cells with aging, categorized as Igf1high (orange) and Igf1low (green). (B) Percentage of the total CD90+ subpopulation categorized as either Igf1high or Igf1low. Young: Igf1high (n = 425 cells), Igf1low (n = 37); Adult: Igf1high (n = 553), Igf1low (n = 421); Aged: Igf1high (n = 61), Igf1low (n = 120). (C) Volcano plot showing differentially-expressed genes in the comparison of all Igf1high cells (n = 1039) against all Igf1low cells (n = 578). Statistical analysis: Wilcoxon Rank Sum test via the FindMarkers function in the Seurat R package; upregulated (pink: p < 0.05; red: False discovery rate (FDR) -adjusted p < 0.05), downregulated (cyan: p < 0.05; teal: FDR-adjusted p < 0.05), and non-significant (ns) genes are displayed on axes of log fold expression (Igf1high over Igf1low) (y, avg_log2FC) against p transformed as its negative log value. (D) Dimensional reduction plot (Monocle package(Trapnell et al., 2014)) of suture mesenchyme CD90+ populations in Young (yellow), Adult (blue), and Aged (grey) mice. (E) Percentage of CD90+ categorized as ‘proliferating’ (inside red dashed ellipse) vs ‘non-proliferating’ (outside ellipse). Young: proliferating (n = 19 cells), nonproliferating (n = 20); Adult: proliferating (n = 5), non- proliferating (n = 57); Aged: proliferating (n = 5), nonproliferating (n = 56). (F) Normalized intensity staining for ATPB in CD90+ cells. (G) Normalized intensity staining for phospho-AMPK in CD90+ cells. (H) Proportion of CD90+ cells that are Ki67+. (I) Proportion of CD90+ cells that are p16+. (J) Ratio of Nuclear:Membrane β- catenin staining in CD90+ cells. (K) Normalized intensity staining for MMP2 in CD90+ cells. For all plots (F-K), Young (yellow), Adult (blue), Aged (grey), periosteum (3 left columns), suture mesenchyme (3 right columns), statistical testing: Welch’s tests in n = 5 mice per condition. (L) Representative images depicting CD90 (magenta), DAPI (grey/blue as indicated by column), and functional markers (green) as indicated by column. Yellow arrows indicate marker-high/positive cells; red arrows indicate marker-low/negative cells.

The impact of Intermittent Fasting on Aged bone structure and osteogenic tissues

(A) Parietal Bone Volume/Total Volume (BV/TV) and (B) representative transverse micro computed tomography (microCT) of Aged mice undergoing intermittent fasting. Statistical analysis: Welch’s test in n = 10 mice per condition. (C) Mean cross sectional area taken up by Endomucin (Emcn)-positive vascular cells and (D) representative immunofluorescence staining of Emcn (magenta) and DAPI (grey) in the periosteum of Aged mice undergoing intermittent fasting. Statistical analysis: Welch’s tests in n = 5 mice per condition. (E) Mean Phalloidin-488 fluorescence intensity and (F) representative images for filamentous actin (F-actin, green) and DAPI (grey) in the suture mesenchyme of Aged mice undergoing intermittent fasting. Statistical analysis: Welch’s test in n = 5 mice per condition. (G) Nuclear aspect ratio (long axis divided by short axis) of cells within the periosteum and suture mesenchyme and representative DAPI-stained nuclei (H) of Aged mice undergoing intermittent fasting. Statistical analysis: Welch’s tests in n = 5 mice per condition. Green and red arrows highlight rounded and elongated nuclei, respectively. (I) Percentage of CD90+ cells within the periosteum (left columns) and suture (right columns), expressed relative to the total cellular number (DAPI+) in Aged mice undergoing intermittent fasting. Representative images (J) show CD90 immunostaining (magenta) and DAPI (grey). Statistical analysis: Welch’s tests in n = 5 mice. In panels A, C, E, G and I, columns in grey are data previously shown in Fig. 1, reproduced here for reference. In panels B, D, F, H, J, Aged control representative images previously shown in Fig. 1 are reproduced here for reference.

Transcriptomic and protein expression characterization of Aged CD90+ cells during intermittent fasting

(A) Volcano plot showing differentially-expressed genes in the comparison of the Aged CD90+ population in the periosteum and (B) suture during intermittent fasting (IF) vs the ad libitum (AL) diet. Statistical analysis: Wilcoxon Rank Sum test via the FindMarkers function in the Seurat R package. Upregulated (pink: p < 0.05; red: FDR-adjusted p < 0.05), downregulated (cyan: p < 0.05; teal: FDR-adjusted p < 0.05), and nonsignificant (ns) genes are displayed on axes of log fold expression (IF over AL) (y, avg_log2FC) against p transformed as its negative log value. Periosteum: Aged IF (n = 194 cells); Aged AL (n = 242 cells). Suture: Aged IF (n = 79 cells); Aged AL (n = 61 cells). (C) UMAP-plot showing the distribution of a subpopulation of CD90+ periosteal cells during intermittent fasting in Adult and Aged mice. Ad libitum (AL) cells are data previously shown, reproduced here for reference. (D) Percentage of the total cells categorized as either Igf1high or Igf1low. Aged + IF: Igf1high (n = 104), Igf1low (n = 44). (E) Plot (Monocle package) of suture CD90+ populations. (F) Percentage of CD90+ categorized as ‘proliferating’ (inside red dashed ellipse) vs ‘non-proliferating’ (outside ellipse) in Aged mice undergoing intermittent fasting. Ad libitum (AL) cells and chart are data previously shown, reproduced here for reference. Statistical analysis: two-tailed binomial test based on expected observed proportions between groups. Adult + IF: proliferating (n = 30 cells), non-proliferating (n = 86); Aged + IF: proliferating (n = 17), non-proliferating (n = 62). (G) Normalized intensity staining for ATPB in CD90+ cells. (H) Normalized intensity staining for phospho-AMPK in CD90+ cells. (I) Proportion of CD90+ cells that are Ki67+. (J) Proportion of CD90+ cells that are p16+. (K) Ratio of Nuclear:Membrane β-catenin staining in CD90+ cells. (L) Normalized intensity staining for MMP2 in CD90+ cells. For all plots (G-L), Aged + IF (red), Aged control (grey), periosteum (2 left columns), suture mesenchyme (2 right columns), statistical testing: Welch’s tests in n = 5 mice per condition. (M) Representative images depicting CD90 (magenta), DAPI (grey/blue as indicated by column), and functional markers (green) as indicated by column). Yellow arrows indicate marker-high/positive cells; red arrows indicate marker-low/negative cells. In panels C-M, data and representative images of Aged controls from previous figures are reproduced here for reference.

The impact of Nicotinamide mononucleotide (NMN) and Akkermansia muciniphila (Akk) supplementation on Aged bone structure and osteogenic tissues

(A) Mean intensity and (B) representative images of live TMRM staining (red) and Hoechst (grey) in Adult mice controls and Adult mice dosed with 600 mg/kg NMN by oral gavage, samples of periosteum ex vivo. Statistical analysis: Welch’s test in tissue from n = 3 mice per condition. (C) Parietal Bone Volume/Total Volume (BV/TV) and (D) representative transverse micro-computed tomography (microCT) sections in Aged mice supplemented with NMN or Akkermansia. Statistical analysis: Welch’s tests in n = 10 (Aged, Aged + NMN) and n = 7 (Aged + Akk) mice. (E) Percentage of CD90+ cells within the periosteum (left columns) and suture (right columns), expressed relative to the total cellular number (DAPI+) in Aged mice supplemented with NMN or Akkermansia. Representative images (F) show CD90 immunostaining (magenta) and DAPI (grey). Statistical analysis: Welch’s tests in n = 5 mice per condition. (G) Mean cross sectional area taken up by Endomucin (Emcn)-positive vascular cells and (H) representative immunofluorescence staining of Emcn (magenta) and DAPI (grey) in the periosteum of Aged mice supplemented with NMN or Akkermansia. Statistical analysis: Welch’s tests in n = 5 mice per condition. (I) Mean Phalloidin-488 fluorescence intensity and (J) representative images for filamentous actin (F-actin, green) and DAPI (grey) in the suture mesenchyme of Aged mice supplemented with NMN or Akkermansia. Statistical analysis: Welch’s tests in n = 5 mice per condition. (K-M) nuclear aspect ratio (long axis divided by short axis) of cells within the (K) periosteum and (L) suture in Aged mice supplemented with NMN or Akkermansia, and (M) Representative DAPI-stained nuclei. Statistical analysis: Welch’s tests in n = 5 mice per condition. Green and red arrows highlight rounded and elongated nuclei, respectively. (N) The relative abundance of bacteria identified as genus Akkermansia in 16S metagenomic profiling of Aged mice undergoing intermittent fasting vs. Aged controls. Statistical analysis: Welch’s test in (Aged, n = 3; Aged + IF, n = 4) mice per condition. In panels C- M, data and representative images of Aged controls from previous figures are reproduced here for reference.

Effect of hour-based alternate serum deprivation on mitochondria in MC3T3 cells in glucose- free media.

(A) Effect of alternate serum deprivation on gene expression implied in mitochondrial dynamics (MFN1, MFN2, DRP1), unfolded protein response (mtUPR) (HSP60, HSPA9, LonP1), biogenesis (PGC1a), structure (ARBP), mitophagy (LC3II PINK1, PARKIN), and mitochondrial respiratory complex IV (MTCO1). Statistical analysis: Welch’s test, n≥4/group. (B) Representative holographic images of mitochondrial shape (orange arrowheads), confirmed by Mitotracker green staining. (C-D) Effect of alternate serum deprivation on mitochondrial length. Statistical analysis: Welch’s t-test of n=5/group, total mitochondria=425. (E) Effect of alternate serum deprivation and oligomycin on mitochondrial membrane potential. Statistical analysis: Welch’s t-test of n=6/group with at least 481 cells/group. (F) Effect of alternate serum deprivation and oligomycin on ATP level. Statistical analysis: Welch’s t-test of n≥5/group.

Effect of NMN on oligomycin-induced mitochondrial damage in glucose-free MC3T3 cells.

(A) Effect of oligomycin ± NMN on ATP level in MC3T3 cells. Welch’s test, n=6/group. (B) Effect of oligomycin ± NMN on ROS level MC3T3 cells. Welch’s test, n=6/group. (C) Effect of oligomycin ± NMN on cell density (left) and mitochondrial membrane potential (right) over time in MC3T3. Welch’s test, n=6/group. (D) Representative images of mitochondrial shape alteration by oligomycin exposure ± NMN supplementation. Dashed white line = nuclear membrane, N = nucleus. Observation by holotomography (grey levels) and confirmed with Mitotracker Green (green). Red arrowheads = Type A (“round”), yellow arrowheads = Type B (“intermediate”), and green arrowheads = Type C (“rod-like/filamentous”) mitochondria. (E) Quantification of the effect of oligomycin ± NMN on mitochondrial shape. Each data point is a cell with at least 30 mitochondria. Welch’s test n = 6 cells / group, total mitochondria=1821. (F) Effect of oligomycin ± NMN on gene expression implied in mitochondrial dynamics (MFN1, MFN2, DRP1), biogenesis (PGC1a), structure (ARBP), mitophagy (PINK1, ATG5), and unfolded protein response (mtUPR) (LonP1, HSP60, HSPA9). Welch’s test, n≥5/group. (G) Effect of oligomycin ± NMN on osteoblastic differentiation. Staining of alkaline phosphatase (ALP) activity and mineralization (Alizarin Red S; ARS) of the matrix after two weeks of differentiation. Welch’s test, n=4 (ALP) or 3 (ARS) / group.

: (A) Methodology for segmentation of micro-computed tomography (microCT) images for calculation of bone morphometry metrics: (B) Parietal Thickness (C) Cortical Tissue Mineral Density (Ct.TMD), (D) Parietal Whole Bone Mineral Density (BMD). Statistical analysis: Welch’s tests, n = 10 mice per condition, except Aged + Akk (n = 9). (E) The method for quantifying immunofluorescence-positive (red) cells. The maximum pixel intensity is determined in control tissues lacking primary antibody (no primary control). This value is used to threshold the stained sample tissue, with the remaining signal distinguishing positive cells (red arrows). DAPI (grey) is also shown for reference of cellular locations. (F) Number and percentage of blood vessels positive for Emcn and/or CD31 with representative images in the periosteum and suture mesenchyme. (G) Methodology for segmentation of Endomucin (Emcn) immunofluorescent staining in fixed tissue sections, for calculation of vascular histomorphometry metrics: (H) Emcn+ staining density (proportion of total tissue area). Data are shown previously in main Figures 1, 3, and 5, and are reproduced here for reference, (I) Mean lumen area per blood vessel, (J) Number of blood vessels per area tissue (BV.N). Statistical analysis: Welch’s tests, n = 5 mice per condition. (K) Representative brightfield images of unstained tissue sections highlighting the overlying periosteum (red P, red dashed line) and parietal bone (red B). (L) Mean periosteal thickness. Statistical analysis: Welch’s tests, n = 4 mice per condition, except Aged + Akk (n = 3).

: (A) representative immunofluorescence staining of Emcn (magenta) and DAPI (grey) and (B) mean cross sectional area of Endomucin (Emcn)-positive vascular cells in mouse sutures. Statistical analysis: Welch’s tests in n = 5 mice. (C) Representative images and (D) mean intensity of live TMRM staining (red) and Hoechst (grey) in adherent cells allowed to explant from suture mesenchyme tissue for 24 hours in vitro. Remaining mesenchyme tissue was removed prior to staining. Statistical analysis: Welch’s tests in tissue from n = 3 mice per condition. (E) Representative images and (F) mean intensity of live Hoechst staining (grey) from the TMRM experiment in Fig. 1i-j. Statistical analysis: Welch’s tests in tissue from n = 3 mice per condition. (G) Schematic demonstrating the hierarchy and selected markers of Skeletal Stem Cells (SSCs), leading to the CD90+ cell pool, which ultimately differentiates into osteoblasts and osteocytes during osteogenesis. (H) Mean Phalloidin-488 staining intensity for F-actin in the periosteum. Statistical analysis: Welch’s tests in n = 5 mice.

: (A) Characterization of Wnt3a-bandage activation of the Wnt/β-catenin pathway with Comma-D Beta cells in vitro. Presented as the proportion of GFP+ cells, normalized to the proportion of GFP+ responsive to soluble Wnt3a. Statistical analysis: two-tailed one-sample t test against the inactive control-bandage (mean = 0); n = 6 for each condition. (B) Characterization of Wnt3a-bandage activation of the Wnt/β-catenin pathway with LS/L cells in vitro. Presented as fold change of Luciferase signal over control-bandage. Statistical analysis: Welch’s test, n = 10 for each Wnt3a-bandage and Soluble Wnt3a, n = 7 for Control-bandage. (C) Representative images for Comma-D Beta cells quantified in (A). mCherry (red) is constitutively expressed, GFP (green) is expressed during Wnt/β-catenin pathway activation. Wnt3a is treated with DTT, resulting in complete chemical inactivation (iWnt3a). Bovine Serum Albumin (BSA) is used as an additional negative control. Polycaprolactone (PCL) is the polymeric material that Wnt3a is bound to during Wnt3abandage manufacture. (D) Representative Movat’s pentachrome staining of bone tissues showing the lamellar/striated structures. Blue line designates the drilled defect edge, thereby separating original intact bone B from new bone NB. The dashed blue box shows the enlarged region.

: (A) Schematic depicting the methodology for measuring the ratio of Nuclear:Membrane β-catenin in single CD90+ cells. (B) Representative images depicting CD90 (magenta), DAPI (grey/blue as indicated by column), and functional markers (green) as indicated by column). Yellow arrows indicate marker-high/positive cells; red arrows indicate marker-low/negative cells. (C) Proportion of CD90+ cells that are Ki67+ by age. Statistical testing: paired, two-tailed t tests in n = 5 mice per condition. (D) Ratio of Nuclear:Membrane β-catenin staining in CD90+ cells by age. Statistical testing: ratio paired, two-tailed t tests in n = 5 mice per condition. (E) Normalized intensity staining for MMP2 in CD90+ cells by age. Statistical testing: ratio paired, two tailed t tests in n = 5 mice per condition. For all plots (C-E), Wnt3a-bandage (dark blue points), inactive bandage (white points), Adult (blue), Aged (grey).

: UMAP-coordinate plot of Thy1 (CD90, A) and Ptprc (CD45, B) expression in clusters of cells within the periosteum. (C) Periosteum cell populations (clusters) as identified through unbiased clustering (D) The CD90+/CD45- clusters of interest (0, 1, 3), separated by mouse condition. UMAP-coordinate plot of Thy1 (CD90, E) and Ptprc (CD45, F) expression in clusters of cells within the suture. (G) Suture mesenchyme cell populations (clusters) as identified through unbiased clustering (H) The CD90+/CD45- cluster of interest (10), and also the small endothelial cluster (12), separated by mouse condition.

: UMAP-coordinate plot of the entire cell populations from Periosteum (A) and Suture Mesenchyme (B). Expression of SSC markers (purple) highlights clusters of transcriptomically-similar cells. The red asterisk is adjacent to the group of cells referred to as the CD90+ (Thy1) population, which is also CD45-(Ptprc-) ( Fig. 5a-h). Prrx1 (Prx1), Ddr2, Itgav (CD51), Cd200 (CD200), Acta2(a-SMA), Axin2, Gli1 and Ctsk (Cathepsin K) are shown.

: (A) Volcano plot showing differentially-expressed genes in the comparison of the CD90+ population in Young vs Aged periosteum; (B) Young vs Aged suture; (C) Adult vs Aged periosteum and (D) Adult vs Aged Suture. Statistical analysis: Wilcoxon Rank Sum test via the FindMarkers function in the Seurat R package. Upregulated (pink: p < 0.05; red: FDR-adjusted p < 0.05), downregulated (cyan: p < 0.05; teal: FDR-adjusted p < 0.05), and non-significant (ns) genes are displayed on axes of log fold expression (y, avg_log2FC) against p transformed as its negative log value. (E) UCell Signature Scoring of transcript expression of genes relating to Complex I electron transport chain (GO-term: 0005747) in the CD90+ clusters of the periosteum and (F) suture. Statistical analysis: Uncorrected Dunn’s tests; n ≥ 242 cells in periosteum per age group, n ≥ 39 cells in suture per age group. (G-J) UMAP-coordinate plot of the CD90+ cells in the periosteum (G, H) and suture mesenchyme (I, J) for the glycolysis and lipid UCell Signature scoring.

: (A) CellChat analysis of the relative inferred levels of outgoing and incoming signaling between subsets of periosteal cells, grouped into pathways. P2 contains the Igf1high and Igf1low populations. Green color scale shows relative intensity in outgoing (left) and incoming (right) inferred signaling patterns for pathways shown at the left edge. Red/blue/green boxes at the top show the relative sum of all pathways. (B) Three-way Venn diagrams showing the overlap of statistically-significant positively-upregulated differentially-expressed genes (p < 0.05) in comparisons between the CD90+ clusters in the periosteum and (C) suture of Young, Adult, and Aged mice. Color of shading and label shows number of genes shared. The central section shows the number of genes conserved throughout all comparisons. (D) Heatmaps of gene expression in CD90+ cells from suture and periosteum mesenchyme of young, adult, aged and aged + IF mice. Scaled expression of genes positively regulated and negatively regulated in senescent cells.

: (A) The mean daily food intake of Aged mice on ad libitum (AL) diet vs intermittent fasting (IF), averaged over the course of measurements of n ≥ 10 mice. Statistical analysis: Welch’s t-test of n ≥ 36 total food measurements taken. (B) UCell Signature Scoring of transcript expression of transcriptional targets of Runx2 in the P1 CD90+ cluster subset of the periosteum. Statistical analysis: Uncorrected Dunn’s tests; n ≥ 46 cells per condition.

: (A) The relative abundance of Firmicutes to Bacterioides bacteria and (B) Lachnospiraceae bacteria, as measured by 16S metagenomic profiling. Statistical analysis: Welch’s tests on n = 3 (Aged), n = 10 (Aged + NMN), n = 8 (Aged + Akk). (C) Representative calvarial TRAP activity staining images. (D) Number of TRAP positive multinuclear cells normalized per bone tissue surface. Statistical analysis: Welch’s tests n = 3 mice per group. The black box shows the enlarged region.

: Schematic that recapitulates the mains findings of the study.

: Statistical analysis of Fig. 1o. Stars and symbols indicate statistical significance calculated by one-way ANOVA followed by unpaired t-test with Welch’s correction, or by ratio paired t-test (non-significant, p>0.05) when comparing inactive vs active within the same group (e.g. Young inactive vs Young active). Ns: non significant, *p<0.05, **p<0.01,

: RNAseq comparison data summary