Melatonin Enhances Osteoblastogenesis from Senescent Mesenchymal Stem Cells via MMSET Mediated Chromatin Remodeling

Large numbers of elderly people have aging-associated osteoporosis, but efficient 2 approaches to ameliorate bone loss are limited due to our poor understanding of the 3 underlying mechanisms. In this study, we found that melatonin levels in bone marrow 4 decreased with age, and melatonin primarily enhanced the osteogenic potential of 5 mesenchymal stem cells (MSCs) derived from elderly donors compared with fetal- or 6 young adult-derived MSCs. Mechanistic studies indicated melatonin treatment alleviated 7 the senescence-related hypermethylation of the MMSET promoter, leading to elevated 8 expression of the histone methyltransferase NSD2, and promoted the histone H3 9 dimethylation modification at lysine 36 of the osteogenic genes RUNX2 and 10 SP7/OSTERIX as a consequence. MMSET depletion partially abolished the effects of 11 melatonin on osteogenesis in senescent MSCs in vitro . Moreover, melatonin treatment 12 promoted bone formation and alleviated the progression of osteoporosis in a mouse 13 model of aging. Clinically, severity of senile osteoporosis (SOP) in patients was 14 associated with melatonin levels in bone marrow plasma and the MMSET expression in 15 MSCs, and melatonin treatment enhanced osteoblastogenesis from MSCs derived from 16 SOP patients. Our study discovered a previously unreported epigenetic regulatory role for 17 melatonin in alleviating MSC senescence and suggests that melatonin may be a potent 18 agent for preventing aging-associated osteoporosis. 19 enhancer far the TSS regison expected, MMSET overexpression led to an increase of H3K36me2 levels on the promoters of osteogenic RUNX2 and SP7 / OSTERIX genes before the induction of osteogenesis. These results strongly indicate that MMSET may render the differentiation propensity of MSCs toward osteoblasts via an epigenetic mechanism. Although our study suggested MMSET is a critical responser to melatonin treatment in ameliorating it

Melatonin is a neurohormone synthesized and secreted predominantly by the pineal gland 2 under the rhythmic control of the suprachiasmatic nucleus and the light/dark cycle(1,2). 3 Previous studies have shown that melatonin is a key molecule in a wide variety of 4 physiological and pathological processes due to the diverse expression of melatonin and 5 its receptors (3,4). Many of the effects of melatonin are mediated directly through 6 membrane-bound melatonin receptors or indirectly via nuclear orphan receptors of the 7 RORα/RZR family (5). Accumulating evidence has also indicated that melatonin is 8 involved in bone remolding, osteoporosis, osseointegration of dental implants, and 9 dentine formation (6,7). The aging-related reduction of melatonin levels has been shown 10 to be a crucial factor in bone loss and osteoporosis with aging (8). Osteoporosis is a 11 debilitating chronic disease marked by decreased bone density and strength, resulting in 12 fragile bones (9). The loss of bone among the elderly occurs silently and progressively, 13 without obvious symptoms until a painful fracture occurs. Therefore, serum melatonin 14 levels might be utilized as a biomarker for the early monitoring and prevention of 15 osteoporosis, and a better understanding of the functional machinery of melatonin will 16 benefit the application of melatonin in alleviating the aging-related progression of 17 osteoporosis(8). 18 5 Mesenchymal stem cells (MSCs) in bone marrow are multipotent stromal cells with the 1 ability to differentiate into a variety of osteogenic, chondrogenic, adipogenic, or 2 myogenic lineages (10). Melatonin can modulate multiple signals to drive the 3 commitment and differentiation of MSCs into osteoblasts (11,12). Increased oxidative 4 stress and cell injury with aging are causal factors of reduced osteogenesis by MSCs. 5 Numerous studies have confirmed that melatonin can promote osteoblast-like cell 6 proliferation, enhance the expression of type I collagen and bone marker proteins, and 7 facilitate the formation of a mineralized matrix (13,14). A study also suggested that 8 melatonin exerts suppressive effects on osteoclasts via the upregulation of calcitonin 9 secretion by osteocytes (15). Mechanistically, through binding to the MT2 receptor, 10 melatonin elevates the gene expression of bone morphogenetic protein 2 (BMP2), BMP6, 11 alkaline phosphatase (ALP), osteocalcin, and osteoprotegerin to favor osteogenesis, and 12 simultaneously suppresses the receptor activator of NF-κB ligand pathway to attenuate 13 osteolysis (8). Intriguingly, osteoblasts from MT2 −/− mice exhibit intrinsic defects in 14 differentiation and mineralization compared with their wild-type counterparts, and the 15 mutant cells fail to respond to melatonin(16). However, despite these known phenotypes 16 and functions of melatonin on osteoblastogenesis, the substantial molecular regulatory 17 mechanisms, especially the epigenetic machinery, are still not well elucidated. 18 Osteoblasts are bone-forming cells derived from MSCs, and the stemness and 19 differentiation properties of MSCs have been shown to decline with age and cellular 20 6 senescence (17). Nevertheless, an in-depth understanding of the mechanisms involved in 1 cellular senescence remains elusive, due to the highly intrinsic heterogeneity and 2 complicated genetic or epigenetic regulatory processes in MSCs. Melatonin is an 3 effective agent for the alleviation of apoptotic factors to protect MSCs from cell 4 injury (18). A series of studies conducted by SH Lee and colleagues have demonstrated 5 that melatonin treatment enhances kidney-derived MSC proliferation and prevents cell 6 senescence, probably by upregulating PPAR, via the PrPC-dependent enhancement of 7 mitochondrial function, or by exosomes carrying microRNAs (18)(19)(20). A study showed 8 that melatonin can restore the osteoporosis-impaired osteogenic potential of bone 9 marrow-derived MSCs by preserving SIRT1-mediated intracellular antioxidation (21). 10 Moreover, incubation of bone marrow-derived MSCs with melatonin predominantly 11 enhances the expression of BCL2, but decreases the expression of BAX, to protect MSCs 12 from apoptosis (22). Despite these findings, it is unknown whether the beneficial effect of 13 melatonin on maintaining MSC regeneration is based on an aging-associated mechanism.
14 Thus, a thorough understanding of the molecular processes controlling MSC senescence 15 is crucial for identifying the drivers and effectors of age-associated MSC dysfunction and 16 to guide the translational application of MSCs in the clinical setting. In this study, we 17 examined the gene expression profiles of human bone marrow MSCs derived from fetal, 18 young adult, and elderly donors to screen for aging-related genes. We discovered a 19 previously unreported phenotype in MSCs derived from elderly donors, but not from fetal 20 or young adult donors, were more sensitive to melatonin stimulation, at least partially 21 7 through the alleviation of DNA methylation on the promoter of the histone 1 methyltransferase MMSET gene. Importantly, melatonin levels in bone marrow plasma 2 were correlated with progression of senile osteoporosis in clinic. Mechanistically, 3 MMSET upregulation facilitated the expression of the osteogenic genes RUNX2 and 4 SP7/OSTERIX by modulating the levels of histone 3 dimethylation at lysine 36, and the 5 beneficial effect of melatonin against bone loss was confirmed in a mouse model of aging. 6 Thus, our study is the first to report the in-depth epigenetic regulatory mechanism of 7 melatonin on the senescence and osteoblastogenesis-related properties of MSCs.

9
Ethic approval 10 This study was approved by the Ethic Committee of Tianjin Medical University (No.   Wisconsin, USA) and cells were then fractionated on a lymphoprep density gradient by 8 centrifugation at 800g for 25 minutes at room temperature with the acceleration at 1. 9 After centrifuge, interfacial mononuclear cells were collected, and washed with 10 phosphatebuffered saline (PBS) at 300 g for 10 minutes at room temperature twice, 11 resuspended in DMEM supplemented with 10% fetal bovine serum (FBS) (Gibco, Life 12 Technologies, Carlsbad, CA, USA), 100 UmL of penicillin, 100 µgmL of streptomycin, 13 and 2 mM L-glutamine (Gibco, Life Technologies, Carlsbad, CA, USA), seeded, and 14 incubated at 37°C/5% CO2. For MSCs isolation from femur of human fetuses (n=12, 15 16-22 weeks, median=18.5, 7 males and 5 females), bone marrow was cultured directly 16 in culture media. After 48 hours, nonadherent cells were removed by changing the 17 medium. Thereafter, the medium was changed every two days. When the cells reached 18 85%-95% confluence, they were trypsinized, counted, and plated again. Cells from     For viral infection, 2×10 5 MSCs were seeded in 1 mL new complete media for 6 hours 1 and then added 50 μL viral concentration and 8 g/mL polybrene, and cells were spin at 2 1800 rpm for 45 minutes at 20C. 12 hours after spinfection, the medium was changed 3 and cells were cultured for another 48 hours until further management.   11 MSCs were cultured in a 6-well plate with complete medium. After reaching 80% 12 confluence, the medium was changed to osteogenic differentiation medium in presence or 13 absence of 1 μmol/L melatonin for 14 days with a medium change every 3 days. The   Alizarin Red S staining quantitation assay 1 The culture medium was aspirated and the cells were washed three times with PBS. Then, 2 the cells were fixed with fresh 70% ethanol for 60 minutes at 4C or fresh 95% ethanol  Alkaline phosphatase assay and quantification 13 Osteogenic differentiation was detected by Alkaline Phosphatase staining. Briefly, the 14 culture medium was aspirated and the cells were fixed with 10% neutral formalin buffer 15 for 10 minutes, washed three times with 1×PBS, and stained with alkaline phosphatase 16 dyeing working solution (Beyotime) at room temperature in dark for 10 minutes or longer, 17 until the color developed to the desired depth. Removed the dyeing working solution and 18 wash it with deionized water for 1-2 times to stop the color reaction. The images were 19 captured with a visible light microscope. To quantify the ALP activity in control and 20 osteoblast-differentiated MSCs, we used the Alkaline Phosphatase Assay Kit 1 (Colorimetric) (BioVision) with modified protocols. Cells were cultured in under normal 2 or osteogenic induction conditions. On day 7, wells were rinsed once with PBS and were 3 fixed using 3.7% formaldehyde in 90% ethanol for 30 seconds at room temperature; then 4 fixative was removed and 50 µL of p-nitrophenyl phosphate solution was added to each 5 well and incubated for 60 minutes in the dark at room temperature until a clear yellow 6 color developed. Reaction was subsequently stopped by adding 20 µL of stop solution. 7 Optical density was then measured at 405nm using a SpectraMax/M5 fluorescence 8 spectrophotometer plate reader. The data were then analyzed by evaluating the increase in 9 ALP expression of the treated samples compared to the untreated. Each experiment was 10 performed in triplicate.

12
Melatonin level in bone marrow plasma was measured using a human MT(Melatonin) 13 ELISA Kit (Elabscience, Cat: E-EL-H2016c, Wuhan, China) accordingly. Briefly, bone 14 marrow plasma from young (n=15, aging 17-45 years, median =30, 11 males and 4 15 females) and elderly donors (n=24, aging 56-84 years, median=65.5, 15 males and 9 16 females) were diluted at 1:5 with sample dilution buffer and added into the plate with 17 primary antibody for incubation at 37°C for 45 min. Afterward, secondary antibody were 18 prepared accordingly and added into samples at 37°C for 30 min. Then, the substrate was 19 13 added to develop the signal for detection. Determine the optical density (OD value) of 1 each well at once with a micro-plate reader set to 450 nm. 2 Quantitative reverse transcription polymerase chain reaction (qRT-PCR) 3 Total RNA was extracted from cells in different treatment groups using Trizol (Life 4 Technologies, South San Francisco, CA USA) and then converted to cDNA using the 5  Western blotting 16 Protein lysates were prepared in RIPA-buffer (50 mM Tris-Hcl, pH 7.5, 150 mM NaCl, 17 10 mM EDTA, 0.5 sodium deoxycholate, 1 NP-40, 1 mM sodium ovanadate, 10 18 gmL aprotinin, 1 mM phenylmethanesulfonyl fluoride, and 10 gmL leupeptin) 19 supplemented with complete protease inhibitors (Roche, Indianapolis, IN, USA). The 20 protein concentration was determined using the BCA protein assay kit (ThermoFisher 1 Scientific, Carlsbad, CA, USA). Cell lysate (50 g) was separated by electrophoresis on 2 SDS-PAGE gel and transferred to nitrocellulose membranes (Pall Corporation, 3 Washington, NY, USA). Membranes were blocked with 5 non-fat milk for 1 hour at 4 room temperature and probed overnight at 4C with specific antibodies. Antibodies used 5 in this study were listed in the supplementary resources. Membranes were washed three 6 times in PBST the next day, then incubated with horseradish peroxidase-conjugated 7 secondary antibodies for 1 hour at room temperature, washed three times with PBST and 8 finally bands were visualized using an enhanced chemiluminescence system (Millipore, 9 Los Angeles, CA USA). The representative Western blot images for at least three   Gene expression microarray 17 Total RNA was extracted from MSC cells of different age groups using TRIzol reagent 18 (Life Technologies, South San Francisco, CA USA) according to the manufacturer's 19 instructions. Quality of the purified RNA was tested on Agilent 2100 Bio analyzer 20 (Agilent RNA 6000 Nano Kit) (BGI, Shenzhen, China). Libraries for cluster generation 1 and DNA sequencing were prepared following an adapted method from BGISEQ-500 2 platform. The low quality reads (more than 20% of the bases qualities are lower than 10) 3 were filtered to get the clean reads. Then those clean reads were assembled into Unigenes, 4 followed with Unigene functional annotation, SSR detection and calculate the Unigene 5 expression levels and SNPs of each sample. Finally, DEGs (differential expressed genes) 6 were identified between samples and do clustering analysis and functional annotations. Reduced representation bisulfite sequencing (RRBS) 8 DNA was extracted from MSC cells of different age groups using Roche kit according to 9 the manufacturer's instructions. Digesting DNA using the MspⅠ restriction enzyme, which 10 cuts DNA at its recognition site (C↓CGG) independent of the CpG methylation status, 11 then, repairing the end and ligating adapters for Illumina sequencing, selecting gel-based 12 DNA fragments with insert sizes ranging from 160 bp to 400 bp, bisulfite treated two 13 successive rounds, after which we observed 98% converted cytosines outside the CpGs, 14 the bisulfite-converted library was used to PCR amplification for 20 cycles, finally, 15 single-read sequencing for 76 cycles using an Illumina Genome Analyzer II.

16
Animal experiment and bone morphology in vivo 17 Aging C57 mice (18 months old) and adult C57 mice (2 months old) were blindly 18 randomized to mock group and melatonin treatment group, and treated with vehicle or 10 19 mg/kg melatonin respectively by subcutaneous injection for 10 weeks, twice a week. 2 20 weeks before termination, calcein (0.5 mg/mice, i.p., biw× 2) was intraperitoneally 1 injected for the last 2 weeks, twice a week. To detect the fluorescence intensity in mouse 2 femur, bone tissues were resin-embedded for hard tissue slides cutting with a 50 m 3 thickness, and pictures were taken with a fluorescent microscope. 4 To assess the in vivo bone morphology, high-resolution X-ray microtomography was 5 performed on mice femur with the SkyScan 1276 microtomograph (BrukermicroCT, 6 Kontich, Belgium). After segmentation, the 3D models were constructed with the CtAn 7 software (release 2.5, Skyscan). 3D measurements were obtained with the CtAn software 8 (release 2.5, Skyscan). Trabecular bone analysis was performed on the femur body.

19
Melatonin levels in bone marrow decrease with aging 1 Because melatonin is an important endocrine hormone regulating osteogenesis and bone 2 homeostasis, we measured its levels in bone marrow plasma from donors of different 3 ages using an enzyme-linked immunosorbent assay. The amount of melatonin in bone 4 marrow decreased in an age-associated manner, as the average level of melatonin in 5 donors aged under 45 years was approximately 400 pg/mL, but it was less than 250 6 pg/mL in donors over 60 years old (Figure 1a). Meanwhile, we assessed the expression 7 of melatonin receptors in bone marrow-derived MSCs isolated from fetal, young adult, 8 and elderly donors using real-time PCR but did not find significant changes in MT1 or 9 MT2 expression (Figure 1b, 1c). To clarify the impact of synthetic enzymes in melatonin 10 synthesis, we isolated total mRNA or mitochondrial mRNA of MSCs derived from fetal, 11 young adult, and elderly donors, and detected the expressions of two key rate-limiting 12 enzymes, arylalkylamine-N-acetyltransferase (AANAT) and 13 hydroxyindole-O-methyltransferase (HIOMT) (24). Total mRNA levels of AANAT and 14 HIOMT were declined at different ranges, but mitochondrial mRNAs were significantly 15 suppressed in MSCs from the old donors, suggesting a possible reason for melatonin 16 attenuation in bone marrow (Figure 1d, 1e). At the same time, the primary characteristic 17 of senescent cells, the activity of lysosomal β-galactosidase, increased gradually in MSCs   18 from the fetal, the young and the old donors (Figure 1f, 1g). Seemingly, melatonin levels 19 in bone marrow have a close link with aging-associated osteoporosis. 20  To screen for differentially expressed genes related to aging, we compared the differences 2 in the gene expression profiles of bone marrow MSCs derived from fetal, young adult, 3 and elderly donors, using gene chip microarray analysis. Multiple comparisons between 4 the three groups of MSCs showed that there was a larger number of differentially 5 expressed genes between the fetal and elderly groups than between the other groups, with 6 1835 genes with a greater than 2-fold change in expression; comparisons between the 7 fetal and young adult groups and between elderly and young adult groups identified 994 8 and 325 genes, respectively (Figure 2a). Among the 41 genes that overlapped in all three 9 comparisons, we found that the expression of some genes known to correlate with 10 stemness, such as ATP binding cassette subfamily G member 2 (ABCG2), insulin receptor 11 substrate 1 (IRS1), and suppressor of cytokine signaling 1 (SOCS1), together with 12 MMSET, was greatly downregulated in senescent MSCs (Figure 2b). Real-time PCR 13 confirmed that MMSET was maintained at a relatively high level in MSCs from fetal and 14 young adult donors, while its expression was dramatically reduced in senescent MSCs 15 from elderly donors (Figure 2c). Western blot analysis further revealed that MMSET 16 protein expression was significantly decreased in senescent MSCs (Figure 2d). Taken   17 together, these data indicated that MMSET may be an important factor in 18 aging-associated osteoporosis. MSCs isolated from the different age groups were cultured in osteoblast differentiation 3 medium for 14 days. ALP activity was investigated and Alizarin Red S staining was 4 performed to assess capacity for osteoblastogenesis. We found that senescence led to the 5 impaired osteogenic potential of bone marrow-derived MSCs, as indicated by decreased 6 ALP activity and mineralization staining (Figure 3a). Quantification of ALP activity and 7 the number of mineralized nodules also confirmed the gradual decline of osteogenic 8 potential with the senescence of MSCs (Figure 3b, 3c). Meanwhile, the expression of 9 osteogenic markers, including bone gamma-carboxyglutamate protein (BGLAP), 10 osteopontin (OPN), and type I collagen alpha 1 (COL1A1), was reduced in senescent 11 MSCs (Figure 3d). Remarkably, association analysis indicated a strong correlation 12 between ALP levels and MMSET expression (Figure 3e), as well as between RUNX2 and 13 MMSET expression (Figure 3f). Interestingly, MSCs from elderly donors exhibited 14 higher ALP activity and mineralized nodule formation after the induction of osteogenesis 15 if they possessed higher levels of MMSET expression (Figure 3g, 3h, 3i). 16

17
To explore the relationship between melatonin and MMSET, we treated cultured MSCs 18 from donors of different ages with exogenous melatonin (1 µM) and then examined 19 21 MMSET expression. We found that MSCs from elderly donors exhibited higher 1 sensitivity to melatonin than those from young adults, but no significant change was 2 observed when compared with fetal MSCs (Figure 4a). Western blot analysis was including the 3'-and 5'-untranslated region (UTR), non-coding regions were slightly 20 increased by 1%, and intergenic region decreased by 2% (Figure 4e). Melatonin 21 22 treatment slightly increased the genome-wide average H3K36me2 signals (Figure 4f), 1 and track profiles indicated that the change in H3K36me2 was occurring in promoter, 2 enhancer, and gene body regions, as seen by the representative tracks of the RUNX2 and 3 SP7 genes (Figure 4g). 4 MMSET favors osteogenic differentiation 5 Previous studies have identified MMSET as a histone methyltransferase that can 6 methylate histones H3 and H4. We found that the repressive status of H3K36me2 was 7 ameliorated and the activated status of H3K27me3 was reduced in MSCs from elderly 8 donors if they expressed MMSET at a high level (Figure 5a), suggesting that elevated 9 MMSET expression is a marker of melatonin-mediated osteogenesis in senescent MSCs. 10 To further ascertain the role of MMSET in the osteogenic differentiation of aged MSCs, 11 we manipulated MMSET expression using a lentivirus carrying short hairpin RNA  (Figure 5c, 5d). In contrast, MSCs with MMSET knockdown exhibited 19 lower levels of ALP activity and decreased numbers of mineralized nodules (Figure 5e). 20 Silencing of MMSET impaired the expression of RUNX2 and SP7/OSTERIX induced by 1 melatonin treatment (Figure 5f). ChIP-qPCR analysis showed that melatonin treatment 2 enhanced the recruitment of H3K36me2 to RUNX2 and SP7/OSTERIX promoters in a 3 dose-dependent manner (Figure 5g), and the recruitment of H3K36me2 to RUNX2 and 4 SP7/OSTERIX promoters was correspondingly altered when MMSET was knocked down 5 or overexpressed compared with the control groups (Figure 5h). Importantly, a high 6 concentration of melatonin failed to enhance the recruitment of H3K36me2 to RUNX2 7 and SP7/OSTERIX promoters after MMSET silencing (Figure 5i). Collectively, these data 8 showed that melatonin favored osteogenic differentiation via the MMSET-mediated 9 modification of H3K36me2 on the promoters of osteogenic driver genes. 10 Melatonin alleviates DNA methylation of the MMSET promoter 11 Subsequently, we wished to identify the mechanism underlying the downregulation of 12 MMSET in senescent MSCs. We analyzed the genome-wide methylation profiles of compared with those from fetal donors (Figure 6a). 7 A previous study indicated that the expression of DNMT1 and DNMT3a decreases in 8 aged cells, whereas the expression of DNMT3b mRNA and protein increases steadily (26). 9 In the present study, we also observed that melatonin (1 µM) treatment effectively 10 decreased DNMT3b expression but had no significant effect on DNMT1 and DNMT3a 11 levels (Figure 6b). Indeed, in MSCs from elderly donors, there was a negative 12 correlation between MMSET and DNMT3b expression (Figure 6c). Consequently, 13 treatment of senescent MSCs with the methyltransferase inhibitor 5-aza to induce DNA 14 demethylation increased MMSET levels (Figure 6d). To further ascertain the relationship 15 between DNMT3b and MMSET in aged MSCs, we knocked down DNMT3b expression 16 using shRNA (S Figure 2a). DNMT3b silencing directly rescued the expression of 17 MMSET in aged MSCs, but it did not have an obvious effect on MMSET expression in 18 MSCs treated with an increasing concentration of melatonin (Figure 6e). Notably, only 19 when DNMT3b expression was rescued in the DNMT3b-knockdown MSCs, but not 20 vector control, could restore the upregulation of MMSET and the osteogenic RUNX2 and 1 SP7 expressions upon melatonin treatment (Figure 6f), consequentially enhancing the 2 MSC derived osteogenesis reflected by significantly augmented Alizarin Red staining 3 and ALP activity (Figure 6g and 6h, S Figure 2 b and 2c). Taken together, our results 4 suggest that the decrease of MMSET levels in senescent MSCs may due to  Melatonin facilitates the osteogenesis of MSCs derived from aged mice via MMSET 8 We verified the relationship between melatonin and MMSET in aged mice. We first 9 evaluated the effects of melatonin on osteoblast-mediated bone formation in vivo. Aged  (Figure 7b); nevertheless, ALP activity and 16 matrix mineralization intensity were effectively increased in the melatonin-treated group 17 compared with the DMSO-treated control group in aged mice (Figure 7c, 7d). 18 Meanwhile, the decreased expression of MMSET in MSCs from aged mice was 19 significantly increased by melatonin treatment, but those from adult mice demonstrated a 20 limited enhancement (Figure 7e). Subsequently, the in vivo bone formation rate per bone 1 surface (BFR/BS) by calcein staining and undecalcified bone sections were imaged and 2 analyzed. The results showed that the bone formation rate was increased in the aged 3 mouse group compared with the young adult mouse group under the same treatment 4 conditions (Figure 7f, 7g). A micro-CT assay showed that melatonin significantly 5 increased the bone mass of aged mice and improved the femoral trabecular 6 microstructure (Figure 7h, 7i). Overall, these data from aged mice also supported the 7 hypothesis that melatonin can ameliorate osteogenesis and bone mass of aged mice, and 8 this effect may depend on the induction of MMSET expression. 9 Melatonin treatment recovers the osteogenic potential of MSCs derived from patients 10 with senile osteoporosis 11 Finally, we investigated the association between MMSET and senile osteoporosis (SOP) 12 in the clinical setting. MSCs were collected from bone biopsy samples from healthy 13 elderly individuals and patients with SOP and every donor was examined by radiography 14 (Figure 8a). We found that MMSET expression was significantly lower in bone marrow 15 MSCs from donors with SOP than in those from age-matched control donors (Figure 8b). 16 There was a positive correlation between bone mineral density, as measured with dual 17 energy X-ray absorptiometry, and melatonin levels in bone marrow in patients with SOP 18 (Figure 8c), and between bone mineral density and MMSET expression (Figure 8d). 19 Furthermore, MMSET expression showed a very strong correlation with melatonin levels 20 in bone marrow (Figure 8e). Accordingly, melatonin (1 µM) treatment enhanced MMSET 1 expression in MSCs derived from patients with SOP (Figure 8f), and improved their 2 differentiation efficiency after osteogenic induction, as evidenced by higher levels of ALP 3 activity, increased numbers of mineralized nodules (Figure 8g, 8h), and upregulated 4 RUNX2 and SP7/OSTERIX expression (Figure 8i). These data strongly suggest that 5 MMSET downregulation is a characteristic of aging-associated osteoporosis and imply 6 that melatonin may be an efficient therapeutic agent for patients with SOP.

8
The composition, structure, and function of bone deteriorates as a consequence of aging, 9 thereby increasing the risk of osteoporosis in elderly individuals. In this study, we report 10 that melatonin levels in bone marrow decrease in an aging-related manner, and treatment 11 with melatonin can reverse the impaired osteogenic potential of senescent MSCs through 12 an epigenetic regulatory mechanism on the histone methyltransferase MMSET (Figure 9). 13 Our study therefore supports the use of melatonin as a potent therapeutic agent for the 14 prevention of aging-associated osteoporosis. 15 Melatonin is a hormone that is secreted mainly from the pineal gland during darkness and 16 plays a vital role in circadian rhythms (27). Bone marrow in pinealectomized rats can still 17 produce high levels of melatonin, suggesting a function for melatonin in regulating the 18 bone marrow microenvironment (28). Melatonin is known to modulate bone formation 19 and osteoblast differentiation of bone marrow-derived MSCs (12). Studies have revealed 1 that melatonin can boost osteoblast differentiation by upregulating osterix protein 2 stability and expression (14), reduce autophagy in high glucose-cultured osteoblasts, and 3 alleviate diabetes-induced osteoporosis by suppressing the ERK pathway (29). In addition, 4 melatonin restores the osteoporosis-impaired osteogenic potential of bone 5 marrow-derived MSCs by preserving SIRT1-mediated intracellular antioxidation(30), and 6 SPRY4 may be partially responsible for the melatonin-mediated osteogenesis of bone 7 marrow-derived MSCs (31). However, it is generally unknown how melatonin promotes 8 the osteogenic potential of bone marrow-derived MSCs undergoing aging. In this study, 9 we found that melatonin in bone marrow was declined with aging, probably due to 10 downregulation of the key rate-limiting enzymes AANAT and HIOMT in mitochondrion, 11 which is consistent with a recent study indicating that AANAT could accelerate aging in a 12 knockout mice model (32). Our study also suggests MMSET (also known as WHSC1 or looser chromatin structure. MMSET is frequently overexpressed in patients with multiple 1 myeloma and its methyltransferase activity is crucial for clonogenicity (33). However, the 2 relationship between MMSET with aging and osteoporosis has not been reported 3 previously, and this is the first description of the importance of this mediator in MSC 4 senescence. 5 Our results established a link between MMSET expression and aging. We identified 6 about 40 genes whose expressions were altered during the senescence of MSC, most of 7 that are stemness-related or senescence-related, such as CCDC28B, SLC19A1, SOX9, 8 ABCG2, NANOG, RUNX2, and CDKN2A as expected. The results also revealed that 9 MMSET was greatly decreased in senescent MSCs derived from the bone marrow of young well-differentiated cells enter into senescence, there is a drift in DNA methylation. 20 A previous study revealed that the abundance of DNMT1 and DNMT3a decreases in 1 senescent cells, whereas DNMT3b expression increases steadily (26). In fact, DNA 2 methylation is well known for gene silencing of BMP2, because higher CpG methylation 3 in the BMP2 promoter is found in osteoporotic individuals compared with healthy 4 adults (36). Given that BMP2 is a pivotal molecule modulating bone formation, aberrant 5 methylation in its promoter region may result in impeded osteogenesis (36). Our research 6 indicates that the decrease of MMSET may be a consequence of aging due to DNA 7 methylation of the MMSET promoter. Analysis of the underlying mechanism showed that 8 the decrease of MMSET in senescent MSCs may be due to DNMT3b-mediated promoter 9 methylation, and treatment with melatonin effectively decreased the expression of 10 DNMT3b to enhance MMSET expression. Dr. Reiter has proposed that melatonin exerts 11 DNMT inhibitory effects either by masking DNMT target sequences or by blocking the 12 active site of the enzyme (37). Our study investigation also suggests that melatonin 13 suppresses the expression of NDMT3b, thereby attenuates the DNA methylation status on 14 MMSET promoter, consequentially promotes the MMSET transcription activity. However, 15 the detailed working machinery of melatonin on DNMT3b suppression has not been 16 clarified in the current study and needs further investigation. 17 Taken together, our study identified MMSET as a modulator of melatonin-mediated 18 osteogenesis. Melatonin modifies the DNA methylation of the MMSET promoter in bone 19 marrow-derived MSCs to promote bone formation. Thus, this study highlights the 20 importance of melatonin-MMSET pathway in aging-associated osteoporosis.
1 Nevertheless, the current study focuses on restoring and accentuating bone formation 2 using melatonin, but the in vivo bone dynamics comprise osteoblast-derived bone 3 formation and osteoclast-driven resorption (38). Thus, apart from the assessment of bone 4 formation in response to melatonin, the residual components of bone have not been 5 investigated in this study, although recent reports have revealed inhibitory effects of 6 melatonin on osteoclastogenesis (39,40). In addition, melatonin has additional effects 7 unconducive to the treatment of age-related bone loss, whether it is amenable to 8 therapeutic purpose remains further investigation.