Indian Hedgehog regulates senescence in bone marrow-derived mesenchymal stem cell through modulation of ROS/mTOR/4EBP1, p70S6K1/2 pathway.

Premature senescence of bone marrow-derived mesenchymal stem cells (BMSC) remains a major concern for their application clinically. Hedgehog signaling has been reported to regulate aging-associated markers and MSC skewed differentiation. Indian Hedgehog (IHH) is a ligand of Hedgehog intracellular pathway considered as an inducer in chondrogenesis of human BMSC. However, the role of IHH in the aging of BMSC is still unclear. This study explored the role IHH in the senescence of BMSC obtained from human samples and senescent mice. Isolated BMSC were transfected with IHH siRNA or incubated with exogenous IHH protein and the mechanisms of aging and differentiation investigated. Moreover, the interactions between IHH, and mammalian target of rapamycin (mTOR) and reactive oxygen species (ROS) were evaluated using the corresponding inhibitors and antioxidants. BMSC transfected with IHH siRNA showed characteristics of senescence-associated features including increased senescence-associated β-galactosidase activity (SA-β-gal), induction of cell cycle inhibitors (p53/p16), development of senescence-associated secretory phenotype (SASP), activation of ROS and mTOR pathways as well as the promotion of skewed differentiation. Interestingly, BMSC treatment with IHH protein reversed the senescence markers and corrected biased differentiation. Moreover, IHH shortage-induced senescence signs were compromised after mTOR and ROS inhibition. Our findings presented anti-aging activity for IHH in BMSC through down-regulation of ROS/mTOR pathways. This discovery might contribute to increasing the therapeutic, immunomodulatory and regenerative potency of BMSC and introduce a novel remedy in the management of aging-related diseases.


INTRODUCTION
Eukaryotic cell aging is an irreversible loss of growth and proliferation which is maintained by intrinsic and extrinsic factors after unceasing cellular replication or surrounding stress. Unbalance in genes expressions of growth regulatory proteins such as P53, P16, and P21, morphological changes, cell cycle arrest, and senescence-associated β-galactosidase (SA β-gal) activities are the main characters in cellular senescence [1]. Usage of mesenchymal stem cells (MSC), especially bone marrow-derived MSC (BMSC), in the treatment of rheumatic diseases, and regenerative medicine has acquired both treatment and curative potential. However, MSC's early aging in vitro remains a challenge which seems to restrain the efforts of scientists and physicians in MSC research and clinical applications. It has been reported that the aging of BMSC interrupts its therapeutic activities such as antiinflammatory cytokines production reduction and decrease in ability to repair bone fractures [2][3][4][5]. One of the In vivo signs of cell aging is skewed differentiation of MSC to adipocyte at the cost of osteoblast, which is considered a core effector in osteoporosis [6]. Thus, keeping BMSC young, preventing aging, and maintaining physiologically balanced differentiation during in vitro or in vivo proliferation are important fundamental requirements in the clinical applications of BMSC and management of aging-related diseases.
Hedgehog signaling is an intracellular pathway with three protein ligands; sonic Hedgehog (Shh), Indian Hedgehog (IHH), and desert Hedgehog (Dhh). Patched1/2 (PTCH1/2) and smoothened (Smo) are trans-membrane receptors which mediate the action of hedgehog members on their target, glioblastoma gene [7]. It has been reported that Hedgehog signaling orchestrates MSC differentiation and prevents the prominent marker of aging, skewed differentiation, by inducing osteogenesis at the cost of adipogenesis [8]. Meanwhile, a core player in aging, P53 pathway, may have an effective interaction with Hedgehog signaling [9]. In addition, IHH, a Hedgehog homolog related to cartilage and bone formation is considered as an inducer in chondrogenesis of human MSC [10]. Moreover, it is reported that the IHH gene transfection could inhibit cartilage senescence [11]. On the other hand, there are a variety of factors that stimulate a cell to be in senescence mood, and these include stimulation by inflammatory cytokines, poor cell-to-cell contact, nutrient deficiency, and disturbance in the regulation of intracellular signaling pathways such as STAT3, NF-kB, Akt, and PI3K [12][13][14][15]. Thus, exploring the interaction of IHH with the above-mentioned parameters may contribute to the understanding of the senescence machinery.
Oxidative stress is one of the major inducers of aging in BMSC, where reactive oxygen species (ROS) have been reported as a promotor of adipogenic differentiation and a repressor of osteogenic differentiation in MSC [8]. Additionally, ROS decrease the ability of BMSC to maintain the hematopoiesis compartment in aged mice [16]. Indeed, it has been recommended that optimizing intracellular and mitochondrial ROS of MSC is required to obtain an optimum effect in treatment by MSC [17]. Therefore, exploration of pathways that regulate the releasing of ROS may provide clues to improve immunoregulatory properties of BMSC in tissue engineering, regenerative medicine, and treatment of autoimmune diseases.
The serine/threonine protein kinase of phosphatidylinositol-3-OH kinase (PI3K) family, mammalian target of rapamycin (mTOR) has two functional proteins, mTOR complex 1 (TORC1) and mTOR complex 2 (TORC2). The mTOR pathway is a growth and metabolism regulator through its three downstream effectors, eukaryotic translation initiation factor 4E-binding protein 1 (4EBP1), phosphorylated ribosomal S6 kinase 1/2 (p70S6K1/2) and Akt [18,19]. Although it is reported that mTOR signaling plays a role in the skewed differentiation of MSC and aging [20], the effect of its inhibition in MSC differentiation is still a subject of conflict. Moreover, the inhibition of this pathway clinically may lead to the desired output in the treatment of aging-related diseases, especially osteoporosis and arthritis [21]. However, the role of IHH and mTOR interaction in the aging of BMSC is still an open issue.
In this study, we observed that IHH has regulatory activities in BMSC senescence and differentiation via controlling the mTOR downstream substrates, 4EBP1 and p70S6K1/2. Herein, we showed that IHH down-regulates oxidative stress in order to protect BMSC from senescence by maintaining the desired differentiation.

IHH is decreased at both transcription and protein levels in senescent BMSC
H2O2-induced senescence was displayed by severe morphological changes ( Figure 1A), increased SA-βgal-stained BMSC count ( Figure 1B), increased expression of aging-related genes, p53, p16, SA-β-gal, decreased ALP-BMA activity, and increased adipogenic markers, FABP4 and LPL ( Figure 1C). Because of the importance of IHH in embryonic growth and differentiation, we proposed its involvement and impact in aging. We, therefore, searched the main members of the IHH signaling pathway in non-senescent and AGING senescent BMSC. We amplified the gene expression of cells using RT-PCR and the results revealed a significant decrease in IHH gene expression in senescent BMSC compared to non-senescent cells ( Figure 1C). Due to their core role in the Hedgehog pathway, we determined PTCH1/2 receptors expressions. Consistent with the previous results, both receptors were down-regulated in senescent BMSC ( Figure 1C). In accordance with PCR results, western blot results for IHH protein ( Figure 1D) showed that senescent BMSC was characterized by decreased IHH protein synthesis. In addition, we observed a decreased IHH gene expression in BMSC isolated from aged donors compared to young donors ( Figure 1E). These results suggested that BMSC lost their ability to synthesize IHH as they aged, and therefore IHH could have an essential task in the regulation of BMSC senescence and differentiation. were incubated with or without senescence-induction medium for 8 days. IHH, PTCH1/2, SA-β-gal, p53, p16, ALP-BMA, FABP4, and LPL genes expressions were measured by RT-PCR. GAPDH was used as a housekeeping gene. (D) Isolated BMSC (n = 4) were incubated with or without senescence-induction medium for 8 days. IHH protein expression was measured by Western Blot. β-actin was used as an internal control. (E) Isolated BMSC from young and old donors (n = 4) were analyzed for IHH gene expression by RT-PCR. GAPDH was used as a housekeeping gene. Results presented as mean ± SEM. *P <0.05, **P < 0.01.

Silencing of IHH promoted aging and altered paracrine-related genes in BMSC
To assess the potential role of IHH in the aging of BMSC, IHH expression was depleted by siRNA silencing. An obvious IHH knockdown was presented by RT-PCR and western blot (Supplementary Figure  2A, 2B). Because existing literature supported the involvement of P53, P16, mTOR, SA-β-gal, PI3K and GDF11 genes in aging, we investigated their expressions in the different states of BMSC. Surprisingly, the count of BMSC cells stained by SAβ-gal stain increased in cells treated with IHH siRNA (Figure 2A). Similarly, the results of RT-PCR revealed induced expressions for all the above-mentioned genes in BMSC transfected with IHH siRNA compared to the negative control ( Figure 2B). In a related issue, the role of paracrine secretions in immunomodulatory and regenerative capacity is a major pathway in the therapeutic usage MSC. It was therefore deemed necessary to evaluate the immunomodulatory potency of BMSC expressing reduced IHH, and to investigate their relationship with the aging process. RT-PCR results showed that paracrine-related genes COX-2, IDO, and IL-6 were significantly induced, while TGFβ was significantly down-regulated, and HGF experience moderate down-regulation in IHH siRNA transfected BMSC ( Figure 2B). For further confirmation, we selected the main markers, P53, P16, and PI3K for western blot examination. Interestingly, the results obtained supported the SA-β-gal stain and RT-PCR results as the three proteins were upregulated in IHH siRNA transfected BMSC ( Figure  2C). Morphologically, the IHH siRNA transfected BMSC showed more transparency, slight enlargement, and decreased in cell count compared to the negative control (Supplementary Figure 2D). Our results demonstrated that IHH might have an anti-aging effect, and its inhibition altered paracrine secretions of BMSC.

IHH knockdown inhibited proliferation and activated cell cycle arrest and aging-related signaling pathways in BMSC
For further assessment of the role of IHH in BMSC aging, we analyzed IHH siRNA transfected BMSC for proliferation capacity and markers of cellular pathways relevant to aging mechanisms. Fibroblast colony forming assay results revealed that IHH siRNAtransfected BMSC failed to form colonies contrary to the negative control ( Figure 3A). Consistent with colony forming assay results, propidium iodide DNA staining was employed for cell cycle assay by FACS, and G0/G1 cell cycle arrest was associated with BMSC of IHH siRNA ( Figure 3B). Molecularly, immunoblotting assay results showed phosphorylation up-regulation of AKT1, NF-κB, and STAT3 was associated with IHH shortage ( Figure 3C). Since oxidative stress was well recognized as a promoter of aging and was involved in the IHH signaling pathway, we assessed ROS generation using flow cytometry. Increased ROS generation in the transfected BMSC was observed compared to the negative control ( Figure  3D). Alternatively, ROS generation results were consistent by using immunofluorescence technique ( Figure 3E). Collectively, our findings showed that IHH could protect BMSC from aging by promoting proliferation and modulating Akt1, NF-κB, and STAT3 signaling pathways, as well as down-regulating oxidative stress

BMSC distorted differentiation was associated with IHH knockdown
Normal differentiation of BMSC is a major advantage that endows the regenerative potency in stem cells. Adipogenic and osteogenic differentiations are balanced mechanisms within physiologically permissible limits, where there is the required increase in osteogenesis than adipogenesis, it is seen as an indicator of functionally normal BMSC. On the contrary, the converse of this mechanism is a sign of aging. To assay for adipogenesis, we incubated the two kinds of BMSC in an adipogenic differentiation medium for 21 days. After that, we applied Oil-Red-O staining protocol which revealed increased red-yellow fat droplets in IHH siRNA-transfected BMSC ( Figure  4Aa, 4C). Additionally, we observed a significant increase in adipocytes within the IHH siRNAtransfected BMSC compared to the non-transfected counterpart ( Figure 4G). These outcomes suggested IHH promoted osteogenesis at the cost of adipogenesis, thereby protecting BMSC from aging.
AGING realized that rIHH couldn't affect the BMSCs CD markers expressions (Supplementary Figure 3). More important, treatment of BMSC isolated from Senescence Accelerated Mouse Prone-8 (SAMP8) mouse line by rIHH decreased aging-related genes, p16 and p53 ( Figure 6B). These findings suggested that exogenous IHH reduces senescence and corrected skewed differentiation in BMSC.

IHH depletion-induced BMSC senescence was suppressed by inhibition of mTOR and ROS
In order to identify possible ways by which IHH achieved its anti-aging effect, we selected the mTOR or/and ROS pathways as keys aging-related pathways based on cues from our previous results. We then inhibited the TORC1/2 and ROS in IHH downexpressed BMSC. The results showed that the genes expressions of P53, P16, SA-β-gal, and PIK3 were compromised after inhibition of mTOR pathway and oxidative stress despite the presence of siRNA IHH ( Figure 7A-7D). GDF11 results of RT-PCR presented down-regulation in gene expression after adding INK128, but not with DPI and NAC ( Figure 7E). In the protein phase, we observed down-regulation of P53 and P16 associated with inhibition of mTOR and ROS pathways ( Figure 8A, 8B). We also observed that BMSC morphological changes caused by IHH knockdown were improved after adding the three inhibitors ( Figure 9A). Meanwhile, western blot results showed down-regulation in the phosphorylation of PI3K, Akt1, NF-κB, and STAT3 after adding of INK128, DPI, and NAC in presence of IHH siRNA ( Figure 8C-8F). In addition, the cell cycle results showed that inhibition of mTOR and ROS pathways restricted the G0/G1 cell cycle arrest caused by IHH silencing ( Figure 8G). Moreover, the colony forming ability of BMSC caused by IHH knockdown was improved after inhibition of mTOR and ROS in the presence of siRNA IHH ( Figure 8H). Furthermore, the differentiation results revealed compromised adipogenesis with INK128, DPI and NAC treatment in IHH siRNA-transfected BMSC, presented with yellow-red fat droplets disappearance (Figure 9Ba  AGING expression of osteogenesis markers, ALP-BMA, RUNX2, and osteocalcin were up-regulated ( Figure  9G). Furthermore, the specific interaction between IHH and mTORC downstream substrates was assessed. As we expected, knock down of IHH induced 4EBP1 and p70S6K1/2 phosphorylation but rIHH protein treatment down-regulate the phosphorylation process ( Figure  10A). Since INK128 is the dual inhibitor for TORC1/2 downstream substrates, 4EBP1 and p70S6K1/2, we wondered to check whether IHH modulate aging-related proteins via dual or single mTOR inhibition. To perform that, siRNA IHH-transfected BMSC incubated with and without rapamycin, S6K protein antagonist to measure the expressions of P53 and P16 proteins. The results showed increased P53 and P16 expressions in cells treated with siRNA IHH but decreased after rapamycin treatment ( Figure 10B). Taken together, our findings explained that IHH regulated the aging process in BMSC via dual or single modulation of the phosphorylation of TORC1/2 downstream substrates, 4EBP1 and p70S6K1/2, and down-regulation of oxidative stress.

DISCUSSION
It is already known that Hedgehog signaling, especially IHH, plays a major role in the development of bone and cartilage, particularly at embryonic level [22,23]. One of the major challenges in the application of MSC in the treatment of autoimmune diseases and regenerative medicine is that MSC assumes aging mode right from the onset of their multiplication in vitro [24]. In this study, we explored for the first time the anti-aging effect of IHH ligand in BMSC through suppression of mTOR and ROS pathways. Our findings showed decreased IHH expression both at the transcription and protein levels in the senescent BMSC. In literature, Hedgehog signaling has been reported to play an important role in fracture healing in young mice [25]. Also, activation of IHH is accompanied by upregulation of trans-membranous Hedgehog receptors PTCH1/2. In the present study, we noted a declined expression of PTCH1/2 receptors, an observation which commensurate the reduced level of IHH in aged BMSC.

AGING
Consistently, PTCH1 was induced during bone formation in rat ulna [26]. The need to maintain young MSC in treatment has been underscored in a report which suggests transfusion of young MSC delays occurrence of aging in mice [27], whereas adult mice are characterized by declined progenitor frequencies [28]. These findings indicated the possible contribution of IHH signaling to the repairing and development of the skeleton, and our results demonstrated that the absence or shortage in IHH may induce early aging.
The involvement of Hedgehog signaling in protection from aging-related diseases has been reported. For example, the sonic ligand is considered a positive player in regenerative medicine, particularly in mammalian cardiac regenerative response and nerve regeneration in aged pelvic plexus [29,30]. Affirmatively, our results showed that IHH knockdown in BMSC induces expressions of agingrelated genes; P53, P16, PPARγ, mTOR, SA-β-gal and PIK3, and proteins; P53, P16, and PI3K, as well as increased SA-β-gal stained cells count. Interestingly, we found that treatment of BMSC with exogenous IHH reversed aging-related genes and SA-β-gal stained cells count. Although the role of GDF11 in aging is still an issue of conflict [31,32], we observed that IHH silencing increased GDF11 expressions. Similar to our observation, Egerman and his coworkers found that GDF11 inhibited regeneration of skeletal muscle and accelerated aging [33]. Possibly, IHH directly or indirectly puts a check on GDF11 as a means of delaying aging in BMSC. The induction of aging-related genes and proteins following IHH and without IHH for 24hours. Aging-related genes, TP53, CDKN2A, SA-β-gal, and mTOR, adipogenesis markers, PPARγ and FABP4, and osteogenesis markers, ALP-BMA, and osteocalcin genes expressions were measured by RT-PCR. GAPDH was used as a housekeeping gene. (B) BMSC from SAMP8 mice (n = 4) were incubated with and without IHH for 24hours. Aging-related genes, p16 and TP53 were measured by RT-PCR. β-actin was used as a housekeeping gene. All data obtained are indicated as mean ± SEM. *P <0.05, **P < 0.01, ***p <0.001. AGING knockdown suggests IHH plays an anti-aging role in BMSC and supports its potency in regeneration activities. The current study also revealed upexpression of COX-2, IDO, and IL-6 genes expressions and down-expression of HGF and TGFβ after IHH silencing in BMSC. Consistent with these findings, senescence messaging secretome (SMS) of MSC has been reported as a tool in aging regulation which triggers senescence through paracrine secretions [34,35]. Additionally, an increase in some paracrine secretions of BMSC contributes to osteoclastogenesis, an aging-related pathway that produces osteoporosis [36,37]. Seemingly, IHH silencing causes aging of BMSC cells population through the generation of senescence-associated secretome phenotype (SASP) which induce cell-tocell communication by altering their paracrine secretions, as shown by the differential expression of the cytokines observed in this study.
Senescence is a biological process which occurs at the late age of eukaryotes, and it is regulated through intracellular signaling pathways. Studies show that signs of aging, which include cell cycle arrest, are induced through JAK2/STAT3 and PI3K/Akt/NF-κB signaling pathways in biliary tract malignancy [38]. In addition, intracellular ROS pathway is also involved in the aging of MSC [8,16,17]. In our study, we found that transfection of BMSC by siRNA IHH induced phosphorylation of PI3K, Akt1, NF-κB, and STAT3, and ROS generation, as well as induced G0/G1 phase cell cycle arrest and inhibited colonies formation. We hypothesized that IHH regulates BMSC senescence through phosphorylation modulation of the abovementioned signaling pathways and promotes the underline mechanisms of proliferation. On the other hand, BMSC skewed differentiation is considered as one of the main causes in osteoporosis, the well-known aging-related disease due to MSC senescence [6]. In the same context, we observed induced adipogenesis associated with suppressed osteogenesis after BMSC transfection with siRNA IHH. More importantly, we discovered that exogenous IHH induces osteogenesis and suppresses adipogenesis in both senescent and nonsenescent BMSC, and also compromises aging markers in SAMP8 mice, suggesting protection from aging may Colonies were visualized after staining with 0.02% crystal violet stain. All results were normally distributed and shown as mean ± SEM. *P <0.05, **p <0.01, ***p <0.001, ****p <0.001.

Figure 9. Biased differentiation induced by IHH shortage was corrected after mTOR and ROS inhibition in BMSC. (A)
Morphology of siRNA IHH BMSC with and without INK128, DPI, or NAC. (Ba, b), (D, E) siRNA IHH-transfected BMSC (n=5) were incubated with or without INK128, DPI, or NAC for 24hours and then incubated in adipogenesis differentiation medium for 21 days. Adipocytes were visualized under inverted light microscope and adipogenesis measured by staining fat droplets using Oil-Red-O stain. (C and F) siRNA IHHtransfected BMSC (n=5) were incubated with or without INK128, DPI, or NAC for 24hours and then incubated in osteogenic differentiation medium for 21 days. Osteogenesis measured using Alizarine-Red-S stain. (G) RT-PCR for adipogenesis markers, PPARγ, LPL, and FABP4, and osteogenesis markers, ALP-BMA, RUNX2, and osteocalcin in IHH-depleted BMSC after incubation with or without INK128, DPI, or NAC for 24hours. GAPDH was used as a housekeeping gene. All results presented as mean ± SEM. *P <0.05, **P < 0.01, ***p <0.001, ****p <0.0001. AGING be exerted by IHH protein through correction of skewed differentiation.
Previous studies suggested that the mTOR and ROS signaling pathways are major factors in the eukaryotic cell aging. Nacarelli and his group reported that mitochondrial oxidative stress mediates human cells and tissues aging through activation of the mTOR/p70S6K pathway [39]. In addition, the PI3K/Akt/mTOR pathway has been reported as an activator of premature aging of preadipocytes after stimulation of ROS signaling pathway by H2O2 [40]. Moreover, mTOR and ROS signaling pathways are reported in the molecular basis of aging-related diseases especially cardiovascular diseases [41]. Furthermore, mitochondrial ROS is seen as a regulatory factor for adipogenic differentiation [42]. Recently, Robert and his co-workers reported on the implications of the mTOR signaling pathway in diseases involved in aging and differentiation [43]. In this study, we observed that aging signs in BMSC exerted by IHH silencing were compromised after inhibition of mTOR and ROS pathways by INK128 or rapamycin, and the antioxidants, DPI or NAC respectively in the presence of siRNA IHH. In addition, we discovered for the first time that the dual inhibition of 4EBP1 and p70S6K phosphorylation by the specific antagonist, INK128 in the presence of siRNA IHH normalized aging-related genes and proteins, signaling pathways, PI3K, Akt1, NF-κB, and STAT3, corrected biased differentiation, and increased BMSC cells proliferation. Also, inhibition of oxidative stress by DPI and NAC in the presence of siRNA IHH provided similar results. Taken together, IHH rejuvenates BMSC by suppression of intracellular ROS which subsequently suppresses STAT3, PI3K/Akt/NF-κB, and downregulate PI3K/Akt/mTOR pathways through interruption of 4EBP1 and p70S6K phosphorylation.
In conclusion, this study demonstrates the role of IHH as an anti-aging factor in BMSC which functions by downregulating ROS/PI3K/Akt/NF-κB/mTOR/4EBP1-p70S6K pathway. Additionally, we discovered a novel pathway which could contribute to rejuvenating BMSC cells in vitro to increase their regenerative and immunomodulatory potency. Moreover, this discovery could provide a therapeutic strategy in the management of agingrelated diseases including osteoporosis, arthritis, cardiovascular diseases, and neurological diseases. Furthermore, our study is the first to indicate the possibility of using sapanisertib in the treatment of aging-related diseases. However, the introduction of exogenous IHH for the treatment of aging-related diseases still needs further exploration in vivo.

Human and animal samples
Fifteen iron deficiency patients and six healthy donors were involved in this study (15 females/6males), age ranged from 25 years up to 81 years with a SD of 20.92 and mean of 54.52. The body weight ranged from 48kg to 78kg with SD of 8.39 and mean of 60.52kg. Bone marrow aspirations were collected from the Second Hospital of the Dalian Medical University, Dalian, China. SAMP8 mouse line was donated by physiology department of Dalian Medical University. The ethical committee of human and animal research of the Dalian Medical University has given the approval for this study.

Isolation of BMSCs
Bone marrow aspirates collected in heparin anticoagulant tube were diluted in an equal volume of PBS. The mixture was dispensed slowly along the wall into a 15ml conical tube containing a double volume of Ficoll-Hypaque density gradient (Lymphoprep, TBD Science, Tianjin, China) and then centrifuged at 800g for 20 minutes continuously. The cloud-like layer of mononuclear cells was collected from the interphase (above the gray layer of Ficoll) into a new sterile tube. The collected cells were washed with sterile PBS two times by centrifugation at 1000 RPM for 5 minutes. The sediment was reconstituted aseptically with 8ml culture medium of 90% low sugar Dulbecco's modified Eagle's medium (DMEM) (Biological Industries, USA), 10% fetal bovine serum (FBS) (Biological Industries, USA), and 100 U/ml penicillin, and 100μg/ml streptomycin (P/S) in 75cm flask and then incubated at 37°C, 5% CO2 and 95% humidity overnight. Floating cells were discarded, and adherent cells washed with PBS, and medium replaced every 2-3 days. At 80-90% confluency, the BMSCs were passaged up to 3rd to 5th passages to be used in experiments. Isolation of mice BMSC was performed using SAMP8 mouse line. The femurs were dissected; both ends were cut and rinsed with PBS, then culture as above human BMSC.

Generation of senescent BMSC
The senescent cells were generated from late passages 9-11 with incubation at 20% oxygen and repeated treatment by H2O2 (200μM) according to the standard protocol [40,44]. The senescence markers are presented in Figure 1. Results of assays on senescent BMSC were compared with young cells generated from early passages (2 to 4) without H2O2 treatment.

IHH silencing
Knockdown of IHH was performed by transfecting lipofectamine 2000 (Invitrogen, Guangzhou, China) and small interference RNAs (siRNAs) complex (Gene Pharma, Shanghai, China) sequences, shown in Table 1. Briefly, siRNA and lipofectamine were incubated separately in 150μL DMEM for 5 minutes at room temperature (RTº) and then mixed thoroughly in RNase free sterile tube. After 25 minutes of incubation at RTº, the transfection mixture was added on 60-70% confluency BMSC in 6-well plat within 1mL DMEM free FBS. The cells were used in experiments after 24 or 48 hours incubation at 37°C, 5% CO2, and 95% humidity. The transfection efficacy was tested by reverse transcription polymerase chain reaction (RT-PCR) and western blot. The whole process was validated using FAM negative control (Supplementary Figure 2C), GAPDH positive control, and siRNA negative control (Table 1) (Gene Pharma, Shanghai, China).

mRNA extraction and RT-PCR
Extraction of mRNA from BMSC was performed using RNAiso (Takara Bio, Japan) as explained in a previous article [45]. PrimeScript ™ 1 st strand cDNA Synthesis Kit (Takara Bio, Japan) was used to produce cDNA using 1μg of extracted mRNA. Primers shown in Table 2 were used in PCR amplification and glyceraldehyde 3phosphate dehydrogenase (GAPDH) was used as a housekeeping gene. PCR MasterMix 2X Power Taq (Bio Teke Corporation, China), cDNA, RNase Free ddH2O, and primers (Takara Bio, Japan) were mixed well in micro-tubes according to the manufacturer's protocol and then loaded in the thermal cycler (Bio-Rad, USA), each cycle consisted of 30s for denaturation at 95 °C, 30s of annealing at 56.0, 56.5, 57, 57.5, or 58.0 °C, and 30s for extension at 72 °C, for a total of 35 or 40 cycles. The PCR outputs of all genes were dispensed in 2% agarose gels wells for electrophoresis in tris acetic acid EDTA (TAE) buffer. PCR bands were quantified by Image Lab detection system (Bio-Rad, USA) and Image J (ImageJ2x, Rawak Software Inc., Germany).

AGING
glycerophosphate in MEM-ALPHA medium (Biological Industries, USA). Osteogenesis and mineralization were assessed by the intensity of calcium deposition which stains red-orange by Alizarine-Red-S stain (Coolaber, China).

ROS measurement
2 / ,7 / -Dichlorofluorescein diacetate (DCFHDA; Sigma-Aldrich, USA) was used to quantify ROS generation levels. 5µM DCFHDA and 70% confluency BMSC were incubated in DMEM at 37°C with 5% CO2 and humidity for 30 min. Cells were harvested after being washed two times with PBS and quantified by Accuri C6 flow cytometer (BD Biosciences, San Jose, CA, USA) at 488 nm and 538 nm wavelength. Fluorescence microscope (Olympus1X71, Japan) was also used to view and capture images of stained cells.

Cell cycle assay
Harvested washed one million BMSC suspended in 1mL PBS were fixed by adding them slowly drop by drop in 9mL 75% ethanol while vortexing. Cells in 75% ethanol were incubated overnight at 4°C and then the pallet was reconstituted after centrifugation in 200μL of a mixture containing 20 μg/mL propidium iodide (PI; BD Biosciences, USA), and 10 μg/mL RNase-A (Coolaber, China) in 0.2% Triton-x water. The stained cells were incubated 15 minutes at RT° and then analyzed using flow cytometry Accuri C6 (BD Biosciences, San Jose, CA, USA).

Fibroblast colony forming unit assay (F-CFU)
3000 BMSC were seeded in complete medium 10% FBS in MEM-ALPHA and 100 U/ml/100μg/ml P/S for 12 days at 37°C 5% CO2 and humidity. The cells were washed with PBS, fixed with 1% paraformaldehyde, and stained with 0.2% crystal violet. The average colony count was estimated from images taken from at least five different fields per well.

Statistical analysis
Analysis of data was done by GraphPad Prism Version 6.07 (San Diego, USA). Comparison of variables was achieved with either one-way analysis of variance (ANOVA) or paired-sample t-test. All experiments were performed at least in triplicates, and data presented AGING as mean ± SD. The differences between groups were considered statistically significant once p-value was < 0.05.

AUTHOR CONTRIBUTIONS
MA, XL, and WB designed the project. MA, AE, JW, WW, WL, YB, YY, YT, MM, and SA were involved in cell culture, PCR, and western blot experiment. MA and WW handled data analysis. MA drafted the manuscript, and WW validated protocols and reviewed the work. The entire study was supervised by XL, WB, and YZ. All made a significant contribution to the work, and approve the final version of the manuscript submitted for publication.