Periosteal mitochondria DNA structures drive aging-associated poor skeletal repair

After we affirmed the feasibility and reliability of in vivo model, next we assessed whether mtG4 in PPM affected bone repair (Fig. 3d, e). 3D-RµCT data showed substantially reduced and more porous new bone in Polg^Mut/Mut when compared with control group (Fig. 3dd1). Data of BV, BV/TV, BMD, Tb.Th, and Tb.N. furthermore proved the significantly decreased new bone in mutant group, and meanwhile the up-regulated Tb.Sp indicated more porous structure of the indicated regenerated bone in Polg^Mut/Mut mice (Fig. 3dd2). H&E (Fig. 3ee1), Masson’s trichrome (Fig. 3Ee2), and Safranin O (Fig. 3Ee3) results all showed that in vivo induction of mtG4 in PPM completely phenocopied the aging-associated poor skeletal repair (Fig. S2), which detailedly incudes in reduced mineralized new bone but with increased cartilage tissues in the hard callus. Biomechanical analysis results demonstrated that in vivo induction of mtG4 led to poor load-bearing function (Fig. S3C). These results together proved that in vivo induction of mtG4 in PPM causes aging-like poor bone repair.

To finally confirm the role of PPM’s mtG4 in bone regeneration, we subcutaneously transplanted the osteogenic committed PPM organoids into adult null mice (Fig. 4ee1). 3D-RµCT showed the mineralized bone mass was severely decreased in K⁺ group when compared with that in the control group (Fig. 4ee2). Quantification of BV/TV and BMD furthermore proved the impaired bone regeneration of these mtG4-accumulated PPM organoids.

Thus, we harvested the PPM organoids and carried out transmission electron microscopic measurements (TEM). TEM data demonstrated that in the mtG4 induction group the cell morphology of PPM was normal in comparison to that in the control (Fig. 5g). And mtG4 neither change the vital membrane organelles such as plasma membrane and nuclear membrane (Fig. S5C c3). However, mtG4 induction specifically caused severe damage in mitochondria (Figs. 5g, S5Cc3). In detail, in the mtG4 induction group mitochondria become much rounder and swelling, and the rarefaction or destruction of mitochondrial crista emerged, showing typical mitophagy appearances (Fig. 5g, S5C c3). Quantitative results furthermore demonstrated the decreased mitochondrial length (Fig. 5h), increased numbers of damaged mitochondria (Fig. 5i), and significantly exacerbated mitophagy of organoids’ PPM in the mtG4 induction group (Fig. S5C c4). These results corroborated the pathological mitochondrial morphology alterations observed in PPM during healthy aging (Fig. S6A). When we returned to the in vivo PPM at the stem cell infiltration stage after fracture, we also observed much more shortened dotted mitochondria in the mtG4-accumulated PPM (Fig. S6B) and whose lineage descendants in the bone callus (Fig. S6C). In vivo induction of mtG4 further showed that Polg mutant PPM in the early callus consistently demonstrated more shortened dotted mitochondria (Fig. S6D). Finally, we detected the mitochondrial functions and JC-1 data showed that K⁺-treated PPM organoids owning significantly reduced mitochondrial membrane potential (Fig. 5j, k). Direct mitophagy detections by using LC3B adenoviruses and MLIF (mtG4 and Tomm20) demonstrated that the PPM of K⁺-treated organoids having significantly exacerbated mitophagy (Fig. 5l). In vivo MLIF also verified that both in healthy aging and premature aging these mtG4-accumulated PPM after bone fracture showed substantially increased mitophagy (Fig. S7). At last, we observed that these mtG4-accumulated PPM organoids demonstrated hundreds of times up-regulated senescence-associated secreting phenotype (SASP) genes such as Serpine 1 (Fig. 5mm1), Tnf (Fig. 5mm2), and Tgfb1 (Fig. 5mm3), and pro-inflammatory cytokines such as Il-1b and Il-6 (Fig. S5Cc5). And even during the osteogenic induction these SASP genes were consistently higher in the mtG4 induction group than that in the control group (Fig. 5n). These results together comprehensively establish that mitochondrial dysfunction underlies the mtG4-caused PPM senescence.

All animal experimentation protocols were subjected to thorough review and received approval from the Ethical Committees of the West China School of Stomatology, Sichuan University (Approval No. WCHSIRB-D-2022-201). The corresponding surgical models were implemented in accordance with the established guidelines of the State Key Laboratory of Oral Diseases at West China Hospital of Stomatology. With respect to the maintaining of experimental mice, they were housed in specific-pathogen free Experimental Animal Core of West China Hospital, Sichuan University (China), which was a temperature controlled (25 °C) environment under a 12-h light/12-h dark cycle with cotton batting.

Pdgfra-CreER (Stock#018280), tdTomato (Stock#007914), and Polg^D257A (stock#017341) were purchased from JAX lab. To get the homozygous mutant of Polg^D257A, we followed the guidelines from JAX lab. In brief, the female and male heterozygous Polg^D257A mice (that originated from a C57BL/6 J female x heterozygous male cross) were crossed to get the homogenous strain. The primer design of Polg^D257A mice was shown in Table S1, and the gene identification was shown in Figure S8.

For tamoxifen (TAM) administration, mice were intraperitoneally injected with corn oil-diluted TAM with the concentration of 70 mg/kg weight.

The bone regeneration model was reported by us previously.²⁵ In brief, in all animal surgical procedures, inhalation anesthesia with isoflurane (typically 1.5%–3% in oxygen) was used for induction and maintenance. buprenorphine was given 60 min prior to surgery. One of the legs was shaved and scrubbed with 75% ethanol. Closed, nonstabilized fractures were generated by three-point bending. The mice were euthanized at d3, d5, d7, and d14 post-fracture for further analysis.

For Pdgfra⁺ periosteal MSCs isolation, d5 post-fracture, the fibrous callus component in early callus was separated and cut into pieces. Then, fibrous callus was digested in 1 mL pre-warmed digestion solution (3 mg/mL collagenase I (Sigma, USA) in αMEM) for 45 min in a water bath at 37 °C with gentle agitation every 10 min. Then, the mixture of digested minced bones and target cells were treated with red blood cell lysate on ice for 3 min. After lysis, single-cell suspension was sorted using CD45 MicroBeads (Milteny, #130-052-301, Germany) and CD104a (Pdgfra) MicroBeads Kit (Milteny, #130-101-502, Germany). Enriched CD45^–Pdgfra⁺ periosteal MSCs were collected for further experiments.

For flowcytometry analysis of periosteum at homeostasis, the femur was cut at 1 mm away from the femoral head and another cut at 1 mm away from the femoral condyle to obtain the bone shaft. Bone marrow and endosteal cells were flushed out with cold PBS through a 1 mL syringe needle. The remaining bone shaft was cut into 2 halves along the longitudinal axis. After removing the residue of endosteum, bone pieces were digested in 1 mL pre-warmed digestion solution (3 mg/mL collagenase I (Sigma, USA) in αMEM) for 45 min in a water bath at 37 °C with gentle agitation every 10 min. The subsequent procedures were the same as above.

Bone tissues were fixed overnight in 4% paraformaldehyde/PBS at 4 °C, followed by decalcification for 7 days in 12% EDTA. After decalcification the tissues were immersed in a 30% sucrose solution overnight at 4 °C. Then samples were embedded in an OCT tissue-freezing medium (Leica) to conduct frozen sectioning at 6 μm thickness using a CryoJane Tape Transfer System (Leica). Frozen sections were treated with 0.1% Triton X-100/PBS (PBST) for 15 min. After blocking w/5% BSA in PBST for 20 min, the cells were detected with Senescence Assay Kit (Beta Galactosidase, Fluorescence) (Abcam, USA), according to the manufacturer’s instructions. Then cells were incubated with TPA-mTO, which was an efficient and reliable mtG4-specific fluorescent probe,²⁴ at room temperature for 45 min. Finally, cells were washed using PBS 3 times at 10 min per time, mounted using anti-fluorescence fate mounting medium, and detected by confocal laser scanning microscopy (Olympus, Japan) following the Alexa Fluor 488 protocol for Sa-β-Gal, and Ex@543 nm for TPA-mTO channel (600–700 nm).

As for FCMs analysis, freshly prepared cells were incubated with Senescence Assay Kit (Beta Galactosidase, Fluorescence) (Abcam, USA) for 2 h at 37 °C and TPA-mTO at room temperature for 45 min. Flow cytometric analysis was performed on FACSAria SORP (BD, USA).

To form the microsphere organoids of PPM from the fractured sites at d5 post-fracture, procedures were carried out in accordance to a previously published literature⁴⁶ with modifications. Briefly, CD45^–Pdgfra⁺ cells (1.5 × 10⁵) were culture in 15 mL centrifuge tubes with non-treated surface in culture medium consisting of αMEM, 10% FBS, 1% penicillin/streptomycin for 24 h with gentle agitation every 4 h. Microspheres were used for in vitro cell experiments unless otherwise noted, and carried out the osteogenic induction, osteochondrogenic induction, and colony formation experiments using these microsphere organoids following our previous study.⁴⁷

With respect to Pdgfra⁺ periosteal MSCs transplantation experiments, procedures were carried out in accordance to our previously established protocol⁴⁸ with modifications. Briefly, microsphere organoids were collected and mixed with Matrigel (Corning, USA). Then, microspheres/Matrigel composites were subcutaneously transplanted in null mice. 28 days after transplantation the composites were harvested for further analysis.

Male SD rats at 2 weeks were used to obtain initial (P0) bone mesenchymal stem cells (BMSCs) from their long bone marrow. using direct adherence as described previously.⁴⁹ In brief, bone shaft was obtained as before. Subsequently, a disposable aseptic syringe was used to draw antibiotic supplemented αMEM medium and to repeatedly wash the bone marrow cavity to collect cells in a sterile flask. The BMSCs were cultured in 25 cm² flasks (Corning, USA) using αMEM, 10% FBS, 1% penicillin/streptomycin in a controlled environment of 5% CO₂ at 37 °C. The culture medium was refreshed every 2 days, and subculturing was performed once the cells reached 90% confluence. To examine the effects of transient K⁺ stimulation on mitochondrial function in cells, passage 3 (P3) BMSCs were utilized.

KCl was used to induce the mtG4 as described in our previously established protocol.²⁴ Briefly, cells were incubating with K⁺ in PBS at a final concentration of 200 mmol/L for 30 min at 37 °C, followed by fresh PBS washing for 3 times. The control group was treated with PBS without additional K⁺. The procedure was repeated every 3 days.

For TEM observation, microspheres were digested to single-cell suspensions and fixed in 2.5% Glutaraldehyde/PBS (1/5) for 5 min and finally detected via JEOL TEM microscope (#JEM-1400PLUS, Japan). For mitochondrial membrane potential detection, microspheres were digested to single-cell suspensions and detected using Mitochondrial membrane potential assay kit with JC-1 (Beyotime, #C2003S, China). And the flowcytometry was carried out using FACSAria SORP (BD, USA). Intracellular steady ATP amount measurements were carried out using Enhanced ATP Assay Kit (Beyotime, #S0027, China) according to the manufacturer’s instruction, and luminescence was detected via the luminometer of Varioskan Flash (Thermo Scientific, USA).

With respect to single-cell suspensions of freshly prepared CD45⁻Pdgfra⁺ cells, 30 000 cells were plated into 6-well plates and maintained in 1 mL of αMEM with 10% FBS and 1% P/S. After 10 days of culture, cells were fixed with 4% paraformaldehyde/PBS and stained with crystal violet (Beyotime, China).

Osteogenic differentiation was performed using osteogenic conditional medium consisting of αMEM, 5% FBS, 1% penicillin/streptomycin, 10 nmol/L dexamethasone, 10 mmol/L glycerophosphate, 50 μg/mL L-ascorbic acid. The medium was changed every 3 d. For chondrogenic differentiation, microspheres were cultured with chondrogenic conditional medium consisting of αMEM, 10 ng/mL transforming growth factor-β3 (Cloud-Clone, China), 100x ITS (Sigma, USA), 100 μg/mL pyruvate, 40 μg/mL proline, 50 μg/mL L-ascorbic acid (Sigma, USA), and 100 nmol/L dexamethasone (Sigma, USA).

After 8 days of differentiation, the microsphere organoids were collected to perform histomorphological staining, immunofluorescence staining, and detect the expression of osteogenic/chondrogenic marker genes.

Microsphere organoids were fixed for 2 h in 4% paraformaldehyde/PBS at 4 °C. After fixation the tissues were immersed in a 30% sucrose solution overnight at 4 °C. Then samples were embedded in an OCT tissue-freezing medium (Leica) to conduct frozen sectioning at 6 μm thickness. The sections were stained with Alizarin Red S staining (Solarbio, China) and Alcian Blue staining (Solarbio, China). Images were taken with an upright optical microscope (Leica, Germany).

Microsphere organoids were fixed in 4% paraformaldehyde/PBS for 2 h and then were blocked w/5% BSA in PBST for 20 min, the microsphere organoids were incubated with corresponding first antibodies overnight at 4 °C. At the second day microsphere organoids were washed in fresh PBS three times and then were incubated using specific second antibody, TPA-mTO, with/without DAPI for 2 h at room temperature. Finally, microsphere organoids were washed using PBS for 3 times at 10 min per time, mounted using anti-fluorescence fate mounting medium, and detected.

For RT-qPCR total RNA was extracted using the TRIzol™ (Invitrogen, USA) according to the manufacturer’s protocol. Complementary DNA was synthesized by using the HiScript III RT SuperMix for qPCR (Vazyme, China) in accordance to user manuals. Then quantitative real-time PCR was performed in triplicate by using AceQ Universal SYBR qPCR Master Mix (Vazyme, China) for PCR reactions on an iCycler Real-Time Detection System (BioRad, USA). The relative amount of mRNA was normalized to house-keeping gene Glyceraldehyde-3-phosphate dehydrogenase (Gapdh). Primers used for RT-qPCR are listed in Table. S1

With respect to μCT assessment, bone tissues were isolated from euthanized mice, then fixed in 4% paraformaldehyde/PBS and transferred to 70% ethanol for storage. All outcomes were obtained via Scanco Medical μCT 35 system (Scanco Medical AG, Bassersdorf, Switzerland) system. The previously established parameters were used as follows: X-ray intensity, 145 μA; integration time, 200 ms; X-ray tube potential, 55 kVp; and threshold, 220 mg/cm³.

The collected micro-CT data were imported into Mimics Research 19.0. The three-dimensional model containing trabecular and cortical bone was reconstructed and output into STL format. The model was then imported into 3-matic software for smoothing and refining the mesh, removing discontinuous points, and then generating volumetric and surface meshes. The volumetric mesh was imported into ABAQUS 2021 for finite element analysis. Material parameters, including elastic modulus and Poisson’s ratio, were defined, along with boundary conditions and loads. The analysis was solved to compute the volume fraction of element stress and displacement data.

Bone tissues were fixed overnight in 4% paraformaldehyde/PBS at 4 °C, followed by decalcification for 3 weeks in 12% EDTA and paraffin embedded. Histomorphological analysis was performed on 6 μm thick sections.

The sections were stained with Hematoxylin and Eosin (H&E) (Biosharp, China) staining, and images were taken with a microscope (Leica, Germany). Standard protocols were followed for Mason’s Trichrome staining (Solarbio, China), Safranin O staining (Solarbio, China), and Alcian Blue staining (Solarbio, China).

For frozen sections, bone tissues were fixed overnight in 4% paraformaldehyde/PBS at 4 °C, followed by decalcification for 7 days in 12% EDTA. After decalcification the tissues were immersed in a 30% sucrose solution overnight at 4 °C. Then samples were embedded in an OCT tissue-freezing medium (Leica) to conduct frozen sectioning at 6 μm thickness. The sections were later treated with PBST as above and then the procedures were same as the immunostaining of cultured microspheres.

As for paraffin sections requiring to be imaged with TPA-mTO and tdTomato together, after rehydrated sections were further subjected to antigen retrieval for 30 min at 95 °C. The sections were later treated as the same as above. tdTomato was labelled by Alexa Fluor 488.

We used ImageJ software for cell counting. First, we limit the region of interest (ROI) to the target area. Then we count the DAPI⁺ cells within the ROI as the total cell count. Next, mtG4⁺DAPI⁺ cells were counted as mtG4⁺ cells. Calculations were performed using the following formula:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\rm{mtG}}4}^{+}{\rm{cells\; proportion}}=\frac{{{\rm{mtG}}4}^{+}{{\rm{DAPI}}}^{+}{\rm{counts}}}{{{\rm{DAPI}}}^{+}{\rm{counts}}}\times 100 \%$$\end{document}mtG4+cells proportion=mtG4+DAPI+countsDAPI+counts×100%

All results were verified via independently triplicated experiments. All graphic illustrations and statistical analyses were obtained using GraphPad Prism software v7.03. For comparison among multiple groups when appropriate one-way ANOVA followed by Tukey post hoc test was applied. With respect to results such as the comparison between the controls and the experimental groups, two-tailed Student’s t test was used to determine the significance of difference. A P-value lower than 0.05 was considered significant. All numerical data were reported in format of mean ± SEM from at least three or more independent experiments. All fluorescent images were processed with Image Pro Plus 6.0 (Media Cybernetics, Rockville, MD, USA), Imaris v9.9.0 (Oxford Instruments, UK), or ImageJ (ImageJ software v1.51w). FCM images were processed via FlowJo v10.8.1 (FlowJo LLC, USA).