A novel macrolide–Del-1 axis to regenerate bone in old age

Summary Aging is associated with increased susceptibility to chronic inflammatory bone loss disorders, such as periodontitis, in large part due to the impaired regenerative potential of aging tissues. DEL-1 exerts osteogenic activity and promotes bone regeneration. However, DEL-1 expression declines with age. Here we show that systemically administered macrolide antibiotics and a non-antibiotic erythromycin derivative, EM-523, restore DEL-1 expression in 18-month-old (“aged”) mice while promoting regeneration of bone lost due to naturally occurring age-related periodontitis. These compounds failed to induce bone regeneration in age-matched DEL-1-deficient mice. Consequently, these drugs promoted DEL-1-dependent functions, including alkaline phosphatase activity and osteogenic gene expression in the periodontal tissue while inhibiting osteoclastogenesis, leading to net bone growth. Macrolide-treated aged mice exhibited increased skeletal bone mass, suggesting that this treatment may be pertinent to systemic bone loss disorders. In conclusion, we identified a macrolide–DEL-1 axis that can regenerate bone lost due to aging-related disease.


INTRODUCTION
3][4][5][6] There is an unmet need for immune modulation therapies with some promising approaches currently in clinical development. 7,8][11][12][13] The association of periodontitis with aging does not simply reflect the cumulative effect of prolonged exposure to the dysbiotic microbiome of the periodontal pockets.5][16][17] Old age is associated with increased periodontal disease severity not only in hu mans but also in mice, which are thus a useful model to study the impact of aging on this oral bone loss disorder. 11,15,18,19][22][23][24] ll iScience OPEN ACCESS

Article
Development endothelial locus-1 (DEL-1) is a homeostatic protein secreted by tissue-resident cells in the gingiva and the PDL, endothelial and mesenchymal stromal/stem cells (MSCs) and some macrophage subsets, and contributes to inflammation resolution and tissue repair. 25pecifically, during the resolution phase of experimental periodontitis in mice, DEL-1 promotes efferocytosis and the emergence of the macrophage pro-resolving phenotype as well as stimulating alveolar bone regeneration. 26,27The pro-regenerative function of DEL-1 is largely independent of its efferocytic/pro-resolving function and involves activation of a b3 integrin-FAK-ERK1/2-RUNX2 pathway in osteoprogeni tor cells. 27Alveolar bone regeneration fails in DEL-1-deficient mice or in mice that express a DEL-1 point mutant that cannot bind b3 integ rins. 27The expression of DEL-1 is severely diminished in old age, both in mice and humans. 28,29iven that DEL-1 levels decline severely in old age and mice R18 months of age are DEL-1 deficient, 30 it is important to develop potential therapeutic approaches to stimulate DEL-1 expression, thereby restoring the levels of this important homeostatic protein in the elderly.We have recently shown that the macrolide antibiotic erythromycin (ERM)-but not other antibiotics, such as penicillin and josamycin-stimulates the production of DEL-1 in vascular endothelial cells.Specifically, erythromycin interacts with the growth hormone secretagogue receptor (GHSR) and activates JAK2 and p38 MAPK signaling, leading to C/EBPb-dependent DEL-1 expression. 31Moreover, systemic erythromycin treatment in mice increased DEL-1 expression in the PDL, which connects the tooth to the surrounding alveolar bone. 31][34][35] We show that erythromycin and other macrolides restore DEL-1 expression in old mice and promote regeneration of bone lost due to naturally occurring, aging-related periodontitis.The same treatment increased the bone mass in the femurs of old mice.Importantly, EM 523, a non-antibiotic derivative of erythromycin, that retains the ability to activate a homolog of the GHSR, motilin receptor, 36,37 reproduced the DEL-1-dependent effect of erythromycin on bone regeneration.Mechanistically, macrolide and EM-523 treatments induced the formation of new bone by upregulating alkaline phosphatase (ALP) activity and the expression of osteogenic genes in periodontal tissue while reducing the number of osteoclasts, thereby favorably influencing the osteogenesis/osteoclastogenesis balance.9][40] Therefore, the non-antibiotic compound EM-523 may represent a safe, effective, and affordable new approach to regenerate bone lost due to periodontitis in humans and perhaps for increasing the mineral content of the skeletal bone in the elderly.

Local treatment with DEL-1-Fc induces bone regeneration in 12-month-old mice
We have previously monitored DEL-1 expression in the periodontal tissue of mice of different ages (ranging from 5 weeks to 24 months of age) and observed a gradual reduction in DEL-1 expression with advancing age, with the oldest mice becoming essentially DEL-1 defi cient. 30In the present study, we first investigated whether aging affects bone regeneration in mice.Aging and its impact on tissue regen eration capacity is a gradual process, and an aging-related phenotype may manifest itself before a mouse is of formal ''geriatric age,'' R18 months. 11We first evaluated 12-month-old mice in a proof-of-concept experiment, using our previously established model, in which bone loss was induced by silk ligature placement (ligature-induced periodontitis, LIP) at the left maxillary second molar for ten days fol lowed by ligature removal for five days to allowDEL-1-dependent inflammation resolution and bone regeneration. 27Whereas young (2 month-old) mice readily regenerated bone upon ligature removal, the 12-month-old mice did not exhibit significant bone regeneration (Figures 1A and 1B).However, when the 12-month-old mice were locally microinjected in the gingiva with recombinant DEL-1-Fc, bone regeneration was observed (Figures 1C and 1D).These data suggest that DEL-1 treatment can rescue the aging-related defect in bone regeneration.

Macrolides induce bone regeneration in a DEL-1-dependent manner in 18-month-old mice with naturally occurring periodontitis
In a model of aging-related periodontal bone loss, it is not practical to perform daily microinjections of DEL-1-Fc as performed in the exper iment in Figure 1.Moreover, given the high cost of recombinant protein production, an affordable approach with a convenient delivery method would be desirable.Since we have previously shown that erythromycin upregulates DEL-1 expression in the periodontal tissue of young mice, 31 we investigated whether erythromycin and other macrolides DEL-1 can induce bone regeneration in 18-month-old mice thereby promoting the regeneration of bone lost due to aging-related periodontitis.Macrolides were administered by intraperitoneal injec tion since direct palatal injection of antibiotics can exert antimicrobial effects on the ligature-associated microbiota, thereby confounding the results. 41The microbial load extracted from the ligatures was comparable between control-and macrolide-treated groups suggesting that peritoneal injection of the macrolides did not exert an inhibitory effect on the microbiota (Figure S1).
Mice develop naturally occurring periodontal bone loss, which becomes severe at an age R15 months. 19Groups of mice, aged 19 to 20-month-old and randomly distributed into control and experimental groups were treated with erythromycin, clarithromycin, or azithro mycin (ERM, CLR, and AZM, respectively) by daily intraperitoneal injection for nine days.Macrolide-treated aged mice exhibited signifi cantly reduced bone heights (CEJ-ABC distances) relative to the control group (Figure 2A, left panel).When the data were transformed to show a change in bone levels relative to the control serving as baseline, the macrolide-treated aged mice displayed positive values, indicating bone regeneration (Figure 2A, right panel).However, this regenerative effect was not observed in macrolide-treated aged Del1 À/À mice (Figure 2B).Indeed, the CEJ-ABC distances of the macrolide-and control-treated aged Del1 À/À mice were comparable (Figure 2B, left panel), indicating no bone regeneration due to macrolide treatment (Figure 2B, right panel).To study the long-term effects of macrolide treatment, aged mice were systemically treated with a macrolide twice a week for eight weeks (Figure 2C).Similar to the shorter treatment, aged mice treated with macrolides for eight weeks showed significantly reduced bone heights relative to the con trol-treated aged mice, (Figure 2C, left panel) indicating bone regeneration (Figure 2C, right panel).These results suggest that macrolides rapidly promote bone regeneration in aged wild-type (WT) mice in a DEL-1-dependent manner.To further support this notion, we exam ined the effects of macrolide treatment on DEL-1 expression and stimulation of osteogenic activity in the periodontal tissue of aged WT mice.
Quantitative real-time PCR (qPCR) of RNA extracted from the palatal gingiva of aged WT mice with naturally occurring periodontitis showed that all macrolides tested significantly upregulated the gene expression of Del1 and anti-inflammatory gene, Il10.Macrolides sup pressed the expression of pro-inflammatory genes Il17 and Il1b (Figure 2D, upper panel).Moreover, all macrolides tested significantly upregulated key osteoblast differentiation-related genes, Runx2, Sp7, and Bglap, involved in the early, middle, and late phases of osteoblast differentiation, respectively (Figure 2D, lower panel).In contrast, macrolide treatment did not downregulate inflammatory genes (Figure 2E, upper panel) or upregulate osteoblast-related genes in the gingiva of aged Del1 À/À mice (Figure 2E, lower panel).The data shown in Figures 2A and 2B suggest that aged WT, but not Del1 À/À , mice grow new bone upon macrolide treatment.To further support this notion, we stained maxillary sections using modified Masson's trichrome stain.Histological observations revealed new bone formation in the macro lide-treated groups in aged WT mice (Figure 2F).Consistent with a requirement for DEL-1 expression, no new bone formation was observed in the tissue sections from macrolide-treated aged Del1 À/À mice (Figure 2G).
Together, these findings suggest that treatment of aged mice with macrolides promotes inflammation resolution, upregulation of oste ogenic genes, and bone regeneration in a macrolide-induced Del1-dependent manner.to new alveolar bone formation. 35,42The results showed diminished DEL-1 protein expression in the PDL of control-treated aged mice; how ever, DEL-1 protein levels in the PDL were significantly elevated in macrolide-treated aged mice compared to those in the control group (Figures 3A and 3B).Staining for DEL-1 in the PDL of aged Del1 À/À mice confirmed the absence of DEL-1regardless of control or macrolide treatment (Figure 3C).
To obtain insights into the mechanism of macrolide-induced DEL-1 expression in the PDL, We examined the effect of macrolide-treat ment on hPDLCs, which contain MSCs.All three macrolides tested, ERM, CLR, AZM, upregulated DEL1 mRNA expression above the baseline levels of the control group (Figure 3D).Additionally, hPDLCs transfected with an hDEL1-promoter-Luc-plasmid and treated with macrolides or control showed a significant upregulation of DEL1 promoter activity by all macrolides compared to the control treat ment (Figure 3E).We previously showed that the transcription factor C/EBPb regulates DEL1 transcription and that ERM enhances its bind ing to the DEL1 promoter in endothelial cells. 43To examine whether a similar mechanism operates in hPDLCs, we performed a chromatin immunoprecipitation assay (ChIP-qPCR).In all macrolide-treated groups, we detected C/EBPb enrichment at the DEL1 promoter, in contrast to control samples (Figure 3F).Together, these data suggest that the increased DEL-1 gene and protein expression in the PDL of macrolide-treated aged mice could be attributed, at least in part, to macrolide stimulation of DEL1 transcription by promoting C/EBPb binding to the DEL1 promoter.Macrolides promote osteogenic activity and reduce osteoclastogenesis in the PDL of aged WT but not Del1 À/À mice The ability of the PDL MSCs to proliferate declines with age, resulting in impaired bone regeneration. 21,24a-Smooth muscle actin (a-SMA) is commonly found in vascular smooth muscle cells and myofibroblasts, which arise from mesenchymal cells during tissue injury and repair. 44-SMA is a marker for identifying osteoprogenitor cells in MSC populations.35 Increased expression of a-SMA leads to cell contraction and serves as a crucial factor in diminishing the clonal capacity of human MSCs (hMSCs) and influencing their differentiation toward osteo blasts.45 To further study the effect of macrolide treatment on promoting osteogenic differentiation, we used a-SMA as a selective marker for osteoprogenitor cells from MSCs.IF images of maxillary tissue sections detecting a-SMA revealed increased a-SMA protein levels in all mac rolide-treated groups of aged WT mice relative to the control (Figures 4A and 4C).In aged Del1 À/À mice, only AZM, of the three macrolides tested, appeared to increase the protein expression of a-SMA in the PDL (Figures 4B and 4D). InhPDLCs ERM, CLR, and AZM upregulated the mRNA expression of the a-SMA-encoding gene (ACTA2) relative to ethanol control (Figure 4E).The association of macrolide treatment with increased mRNA and protein expression of a-SMA further supports the notion that macrolides promote osteogenic activity in the PDL.
Macrolide-induced promotion of bone regeneration in aged mice suggests that the macrolides favorably regulate the osteogenesis-os teoclastogenesis balance.To test this, we performed ALP staining (Figures 4F-4I) and tartrate-resistant acid phosphatase (TRAP) staining (Figures 4J-4M) on maxillary tissue sections to detect osteogenesis and osteoclastogenesis, respectively, on the alveolar bone surfaces in the PDL of aged WT and Del1 À/À mice treated with macrolides or control.The analyses showed that macrolide treatment resulted in increased ALP activity (Figures 4F and 4H) and decreased numbers of osteoclasts (Figures 4J and 4L) in aged WT mice but not in aged Del1 À/À mice, in which macrolide treatment had no effect (Figures 4G, 4I, 4K, and 4M, for ALP and osteoclasts, respectively).AZM appeared to strongly promote ALP activity and suppress osteoclastogenesis compared to ERM and CLR (Figures 4H and 4L).Together the bone mea surements (Figure 2A) modified Masson's trichrome staining (Figure 2F), and the increased ALP activity in macrolide-treated aged mice indi cate that macrolide treatment promotes osteoblast activity which leads to bone regeneration in aged WT mice.
The ability of macrolides to reduce the number of osteoclasts in vivo in a DEL-1-dependent manner could, in part, be attributed to their anti-inflammatory effects.Another potential mechanism may involve direct inhibitory effects on osteoclastogenesis, given that DEL-1 was shown to directly regulate osteoclast differentiation. 46To study the effect of macrolide treatment on osteoclast differentiation, we performed an osteoclastogenesis assay using the RAW264.7 mouse cell line.RAW264.7 cells were treated with macrolides at a concentration well below the cytotoxic concentration (Figure S2).Macrolide-treated RAW264.7 cells yielded significantly fewer TRAP + MNCs than control-treated RAW264.7 cells (Figures 4N and 4O).Similar inhibitory effects on osteoclastogenesis by all three macrolides were observed when we used bone marrow-derived monocyte/macrophage precursor cells 47 in a RANKL-induced osteoclast differentiation assay (Figures 4P-4R).Among the three macrolides tested, AZM displayed the most potent inhibitory effect on osteoclastogenesis (Figures 4P-4R).These data suggest that macrolides, in particular AZM, inhibit RANKL-induced osteoclastogenesis in vitro and reduce the number of osteoclasts in the aged perio dontium in vivo.
In summary, in vivo studies show that macrolides promote osteogenic and restrain osteoclastogenic activity in the PDL of aged mice in a DEL-1-dependent manner.Consistently, these macrolide effects are replicated in relevant in vitro assays.

ERM increases the bone mass and affects osteolineage cell parameters of mouse femurs
To determine whether the pro-osteogenic effects of macrolide treatment also impact the skeletal bone, we next analyzed the femurs of con trol-and ERM-treated aged mice.ERM-treated aged mice showed a significant increase in bone mass compared with age-matched controls (Figures 5A-5D).Bone morphometric analysis revealed that ERM treatment elevated bone volume fraction (BV/TV), trabecular number, and reduced trabecular separation (Figure 5E).ERM treatment also affected osteolineage cell parameters within mouse femurs.The femurs from ERM-treated aged mice exhibited an increased number of osteoblasts and decreased number of osteoclasts, while chondrocytes were not significantly affected, as compared to the control group (Figure 5F).Together, these data suggest that ERM treatment promotes bone regen eration in old mice not only in the periodontium but also in long tubular bones throughout the body.

EM-523, a non-antibiotic erythromycin derivative, promotes bone regeneration in aged mice and in vitro studies
The overuse of antibiotics and increasing incidence of antimicrobial resistance has prompted researchers to dissociate the antimicrobial ef fects of antibiotics from their anti-inflammatory and immunomodulating functions, which could be exploited for treating inflammatory dis eases.EM-523 is an erythromycin derivative that is devoid of antimicrobial properties 4,35 but can activate the motilin receptor, a homolog of GHSR, 36,37 which we have shown mediates ERM-induced DEL-1 upregulation. 31We investigated whether EM-523 could reproduce the pre viously described bone regenerative effects of ERM. that of ERM (Figure 6A).However, when EM-523 and ERM were tested in an eight-week treatment regimen, EM-523-treated aged mice ex hibited significantly greater bone regeneration relative to ERM-treated aged mice (Figure 6B).IF detecting DEL-1 in maxillary sections from experimental and control-treated aged mice showed that EM-523 and ERM comparably promoted DEL-1 protein expression in the PDL relative to control-treated aged mice (Figures 6C and 6D).Furthermore, in hPDLCs cultures, EM-523 upregulated DEL1 gene expression relative to treatment with the control (Figure 6E).Moreover, we transfected hPDLCs with a hDEL1 promoter-Luc-plasmid and treated them with ERM, EM-523, or control.Measurements of luciferase activity showed a significant upregulation of DEL1 promoter activity in EM-523-treated mice compared to both ERM and control treatments (Figure 6F).To examine whether EM-523 promotes C/EBPb binding to the DEL1 promoter in hPDLCs, ChIP-qPCR was conducted and showed that EM-523 showed higher C/EBPb enrichment at the DEL1 promoter than both the ERM-treated and the control-treated groups (Figure 6G).qPCR of RNA extracted from primary mouse PDLCs, which underwent osteogenic differentiation for nine days, revealed that EM-523 induced stronger upregulation of Del1 gene expression as compared to not only the control but to all macrolides tested (Figure 6H).Similarly, compared to the macrolides tested, EM-523 exhibited the most potent effect on promoting Runx2 and Sp7 expression.Additionally, EM-523 significantly upregulated Bglap compared to the vehicle-control group (Figure 6H).
Given that EM-523 promotes DEL-1 expression and bone regeneration, we next determined if EM-523 could promote a pro-osteogenic environment in the PDL, in side-by-side experiments with ERM.Maxillary tissue sections of aged mice treated with control, ERM, or EM-523, were stained for ALP (Figures 6I and 6J) and TRAP (Figures 6K and 6L) to detect, osteogenic and osteoclastogenic activity, respectively, on the alveolar bone surfaces in the PDL.Relative to the control treatment, but similar to ERM, the EM-523 treatment resulted in increased ALP ac tivity (Figure 6J).The EM-523 treatment reduced the number of osteoclasts in the PDL of the aged periodontium as compared to both ERM and control groups (Figure 6L).IF images of maxillary tissue sections detecting a-SMA revealed increased a-SMA protein levels in EM-523 treated aged mice relative to both ERM and control (Figures 6M and 6N).Moreover, in hPDLCs cultures, EM-523 upregulated ACTA2 gene expression relative to both ERM and control treatments (Figure 6O).
Consistent with its in vivo effects on osteoblasts and osteoclasts, EM-523 induced osteogenic differentiation in MC3T3-E1 cells (Figures 7A  and 7B), hPDLCs (Figures 7C and 7D) and human iPSC-derived MSCs (Figure 7E), as well as inhibited RANKL-induced osteoclast differenti ation from primary bone marrow-derived osteoclast progenitor cells (Figure 7F).EM-523-treated sampled showed increased calcified nodule formation in MC3T3-E1 cultures compared to ERM (Figures 7A and 7B).Similarly, EM-523 treatment increased ALP activity (Figure 7C, right panel) and upregulated the expression of osteoblast-related genes (RUNX2, SP7, and BGLAP) in hPDLC-differentiated osteoblasts (Figure 7D) (O) Cultures of hPDLCs were treated with ethanol control, ERM, or EM-523 (10 and 1 mg/mL).The mRNA levels of ACTA2 from hPDLCs were quantified by qPCR on day seven after drug treatment.Data were normalized to GAPDH mRNA and plotted relative to the aged control group, set as at lower concentrations than those used for ERM and other macrolides.Moreover, when human iPSC-derived MSCs cultured in osteogenic medium for 26 days were treated with EM-523 or other macrolides, all compounds increased calcified nodule formation.In addition, the expression of DEL1 and osteogenic genes RUNX2, SP7, and BGLAP were upregulated under the same conditions (Figure 7E).In contrast, treatment with macrolides and EM-523 did not upregulate osteoblast-related genes in osteoblastic progenitor cells derived from Del1 À/À mice (Figure 7G).In RANKL-induced osteoclastogenesis, EM-523 reduced the size of osteoclasts and percentage of the TRAP + area, similarly to ERM and DEL-1-Fc.When compared to ERM, EM-523 exerted its inhibitory effect even at very low concentrations (%0.1 mg/mL) (Figure 7F, lower panel).However, the inhibitory effect of EM-523 on osteoclastogenesis was not observed in the osteoclastogenesis assay using primary osteoclast progenitors from Del1 À/À mice (Figure 7H), indicating that the action of EM-523 is DEL-1-dependent.Under DEL-1-deficient con ditions, only DEL-1-Fc treatment could reduce the size of osteoclasts and the percentage of the TRAP + area (Figure 7H, lower panel).These data suggest that EM-523 is a stronger inducer of DEL-1 mRNA and protein expression than ERM.By increasing the expression of DEL-1, EM-523 stimulates osteoblastic activity while suppressing osteoclastogenesis, thereby promoting bone growth in aged mice.

DISCUSSION
Different approaches have been proposed to regenerate bone lost due to periodontal disease.These include surgical implantation of scaf folds, allogeneic or autologous bone material, the use of stem cells, and treatment with bone morphogenetic proteins and growth factors.9][50][51][52] An affordable and effective regenerative strategy may involve approaches that promote tissue responses to favorably modulate endogenous stem cell niches and tissue microenvironments, thereby enhancing tissue repair and regeneration, which are impaired in old age. 53In the present study, we have made a significant conceptual and translational advance by establishing a macrolide-DEL-1 axis that stimulates bone regeneration in old age.Specifically, we have shown that macrolides, and, importantly, a non-antibiotic derivative, can restore the expression of the homeostatic and pro-osteogenic protein DEL-1, 25,27,46 the expression of which declines with aging in both mice and humans. 28,290][11][12] Therefore, restoring DEL-1 protein levels with an affordable and safe approach holds a great promise for treating elderly patients.
The association of old age with increased susceptibility to chronic inflammatory conditions, including periodontal disease, may be due to alterations in the immune and inflammatory status of aging tissues combined with a reduction in the regenerative ability of PDL cells. 10,14,15,54he PDL contains a heterogeneous mesenchymal stromal and stem cell population responsible for the regeneration of periodontal tissue, including the extracellular matrix and adjacent mineralized tissues, such as the alveolar bone.6][57][58][59][60][61] Multipotent human PDL cells can undergo osteogenic, adipogenic, or chondro genic differentiation depending on the culture conditions and can regenerate periodontal tissue upon transplantation to immunodeficient mice. 32,34,62,63The ability of the PDL MSCs to proliferate and undergo osteogenic differentiation declines with age, resulting in impaired tis sue repair. 20,21,245][66][67][68][69][70] Such aging-related dysfunctional alterations may also involve the declining expression of DEL-1 by tis sue-resident cells, including vascular endothelial cells. 28,71[72][73][74] DEL-1 expressed in the PDL niche may not only contribute to the resolution of inflammation 26 but can also induce osteoblastogenesis and the for mation of new alveolar bone during inflammation resolution. 27Therefore, the aging-related deficiency of DEL-1 may contribute to the failure of the aged niche to respond to the need for tissue repair and bone regeneration.The failure of the aged tissue to properly repair and regen erate itself can be reversed, as we show in this paper, by restoring the levels of DEL-1 in the PDL using macrolides and a non-antibiotic de rivative, EM-523.In perivascular areas, the PDL contains a progenitor population, expressing the MSC markers STRO-1 and CD146, that can generate multiple mesenchymal lineages, including fibroblasts, cementoblasts, and osteoblasts. 22,32,35,62Here, we showed that macrolides and EM-523 upregulate the expression of DEL-1 not only in the PDL of aged mice, but also in cultures of hMSCs, which, in the presence of macrolides, undergo enhanced osteogenic differentiation.DEL-1, endogenously produced in hMSCs, might regulate their proliferation and/ or differentiation toward osteoblasts.Since the macrolide treatment was administered systemically, it is likely that the macrolides modulated MSCs at sites different from the PDL.Consistent with this notion, the systemic treatment of aged mice with ERM, twice a week for eight consecutive weeks, resulted in increased bone mass in the long bones, suggesting potential therapeutic benefits for additional aging-related bone loss disorders, such as osteoporosis. 13DEL-1 exhibits a key inhibitory effect on osteoclasts through its ability to repress the master regulator of osteoclastogenesis, namely the transcription of nuclear factor of activated T cells cytoplasmic 1 (NFATc1) by binding to Mac-1 integrin and elevating Bcl6 expression. 30As a result, the modulation of DEL-1 expression by macrolides holds promise as a prospective approach to mitigate osteoclast activity for ther apeutic purposes.Notably, clinical studies have reported favorable outcomes with the use of macrolides in the treatment of chronic periodon titis. 75This ERM-mediated DEL-1 expression has shown efficacy in inhibiting periodontitis progression and suppressing the expression of genes associated with osteoclast activity.Additionally, AZM effectively suppresses osteoclast resorptive activity and osteoclast formation by modulating NFATc1. 76This study reveals the capacity of macrolides and EM-523, in addition to ERM, to modulate DEL-1 expression and consequently influence osteoclasts.These macrolides, including EM-523, hold promise as potential agents with the dual capability of promoting bone regeneration while also suppressing osteoclastogenesis.
To exploit the immunomodulatory effects of macrolides, these antibiotics may need to be administered over an extended period of time.Therefore, there is a possibility that antimicrobial resistance may develop.This study indicates non-antibiotic ERM-derivative, EM-523, is an effective modulator of DEL-1 expression that promotes periodontal bone regeneration, which is a therapeutically important discovery.EM 523 effectively increased DEL-1 protein levels in the PDL and favorably modulated bone cells, osteoblasts, and osteoclasts, in lower effective concentrations than the macrolides tested.Therefore, EM-523 may have a more favorable safety profile and fewer adverse reactions (e.g., pseudomembranous colitis, as reported for erythromycin and other antibiotics), 77,78 while additionally preventing antimicrobial resistance.
In summary, this study highlights the potential use of macrolide antibiotics and specifically EM-523 as potential therapeutic approaches for regenerating lost bone and preventing further disease progression, particularly in the aging population.Further understanding of the mech anisms and downstream effectors that underlie the immunomodulatory properties of macrolides and their derivatives can lead to the devel opment of safe and effective clinical applications to improve bone regeneration in elderly patients with periodontitis and other inflammatory bone loss disorders.

Limitations of the study
This study has revealed an increase in bone mass and changes in femoral bone parameters following ERM treatment by using the microcomputed tomography (micro-CT) and dual-energy X-ray absorptiometry (DXA).While our analysis involved these techniques, they were unable to offer the level of detailed information that could be provided by dynamic bone histomorphometry.Dynamic bone histomorphom etry is a comprehensive methodology measuring bone remodeling at the level of individual events.It provides a quantitative means to assess alterations in bone microstructure, bone generation, and bone renovation, offering valuable insights into cellular events. 79,80By assessing parameters such as bone mineral density (BMD), trabecular thickness, osteoblast and osteoclast activity, and bone turnover rates, dynamic bone histomorphometry offers an in-depth understanding of cellular transformations occurring within bone tissue.Future studies should consider the incorporation of dynamic bone histomorphometry as it offers valuable insights into cellular transformations.
Our present work was performed in the mouse model and is thus uncertain whether the DEL-1-dependent beneficial effects of macrolides could be extrapolated to humans.Nevertheless, the ability of DEL-1 to inhibit periodontitis in mice was also confirmed in non-human pri mates. 812][83] Macrolides are active in both mice and humans although it remains to be established whether macrolides, or the erythromycin derivative EM-523 can upregulate DEL-1 in human tissues as they do in mouse tissues.

Figure 1 .
Figure 1.Local treatment with DEL-1-Fc promotes bone regeneration in 12-month-old mice (A) Experimental design.(B, left) Measurement of bone heights (distance from cement-enamel junction [CEJ] to alveolar bone crest [ABC]; CEJ-ABC) in groups of 2-month-old and 12-month-old mice after ten days of ligature placement (10dL) or after 10 days of ligature placement followed by five days without ligatures to enable resolution from periodontitis (10dL + 5dR).(B, middle) Data from the left panel were transformed to show bone loss at ligated (L) sites vs. unligated (U) contralateral sites.(B, right) Data from the middle panel were transformed to show bone growth relative to corresponding 10dL group (baseline).(C) Experimental design.(D) 12-month-old mice were subjected to LIP for ten days with or without ligature removal for five days to enable resolution, with or without local microinjection with DEL-1-Fc (1 mg) or equal molar amount of Fc control.Treatments were performed daily on days 10-14.(D, left) Measurement of bone heights (CEJ-ABC distance).(D, middle) Data from the left panel were transformed to show bone loss in ligated (L) sites vs. unligated (U) contralateral sites.(D, right) Data from the middle panel were transformed to indicate bone growth relative to the 10dL group (baseline).Data are means G S.D. (B, n = 5-6 mice/group; D, n = 4-7 mice/group).*p < 0.05, **p < 0.01; one-way ANOVA and Dunnett's test.

Figure 2 .Figure 3 .
Figure 2. Macrolides induce bone regeneration in a DEL-1-dependent manner in 18-month-old mice with naturally occurring periodontitis (A and B) Aged WT and aged Del1 À/À mice were administered macrolide antibiotics or vehicle control (ethanol) intraperitoneally daily for nine days.(C) Aged WT were administered macrolide antibiotics intraperitoneally twice a week for eight weeks.(A-C) The distance from the CEJ-ABC was measured (left panels).The CEJ-ABC data were transformed to show the bone changes relative to the control, which was set as the baseline (right panels).Positive values (in mm) indicate bone growth relative to the baseline (control).(D and E) Analysis of gene expression from gingiva samples collected from (D) aged WT and (E) aged Del1 À/À mice after nine days of drug treatment.The mRNA levels of the indicated genes were quantified by qPCR.Data were normalized to Gapdh mRNA and plotted relative to the aged control group, set as 1. (F and G) Coronal maxillary sections from (F) aged WT and (G) aged Del1 À/À mice were stained with modified Masson's trichrome stain, which stains the mature (old) bone blue and new bone formation red.T: tooth; PDL: periodontal ligament; OB: old bone as blue; NB: new bone as red.Scale bars, 100 mm.Data are means G SD (A, n = 8 mice/group; B, n = 4-6 mice/group; C, n = 6 mice/group; D-E, n = 5 mice/group).**p < 0.01; one-way ANOVA and Dunnett's test.

Figure 4 .
Figure 4. Macrolides promote osteogenic activity and reduce osteoclastogenesis in the PDL space of aged WT but not Del1 À/À mice Aged WT and aged Del1 À/À mice were administered macrolide antibiotics or vehicle control (ethanol) intraperitoneally daily for nine days.(A and B) Representative IF images of maxillary sections obtained by fluorescence microscopy.Frozen maxillae sections were stained for alpha-smooth muscle actin (a-SMA, green), and the nuclei were counterstained using DAPI (blue).T: tooth; PDL: periodontal ligament; B: alveolar bone.Scale bars, 100 mm.(C and D) Mean fluorescence intensity (MFI) values were measured in IF images from aged WT and Del1 À/À mice using the ImageJ software.All MFI values were normalized and plotted relative to the MFI of the control group in aged mice, set as 1. (E) Cultures of hPDLCs were treated with ethanol control or indicated macrolides (10 mg/mL).The mRNA levels of ACTA2 from hPDLCs were quantified by qPCR on day seven after drug administration.Data were normalized to GAPDH mRNA and plotted relative to the aged control group, set as 1. (F and G) Optical microscopy images of maxillary sections show PDL areas stained for ALP activity.Maxillary sections are from (F) aged WT and (G) aged Del1 À/À mice.T: tooth; PDL: periodontal ligament; B: alveolar bone.Scale bars, 100 mm.(H and I) The ALP-positive region of each area was quantified and represented as a percentage of the total area.(J and K) Optical microscopy images of maxillary sections show PDL areas stained for TRAP.Maxillary sections were obtained from (J) aged WT and (K) aged Del1 À/À mice.Arrowheads indicate TRAP + cells in the PDL space.T: tooth; PDL: periodontal ligament; B: alveolar bone.Scale bars, 100 mm.(L and M) The number of TRAP + cells in the PDL space were counted and compared between indicated groups.(N and O) RAW264.7 cells were treated with RANKL for one week to induce osteoclast differentiation.(N) Representative images of TRAP + MNCs.Scale bars, 50 mm.(O) The number of TRAP + MNCs was counted.(P-R) Bone marrow cells harvested from C57BL/6 mice were subjected to RANKL-induced osteoclastogenesis assay in the presence of macrolides (10 mg/mL) or control solvent for one week.

FFigure 5 .
Figure 5. ERM increases the bone mass and affects the osteolineage cell parameters of mouse femurs (A-F) Aged WT mice were administered with ethanol control or ERM intraperitoneally twice a week for eight weeks.(A and B) Representative images of Von Kossa (A), Toluidine blue (B, upper panel), Alcian blue (B, lower panel)-stained femur.Scale bars, 500 mm.(C and D) Bone mineral content (BMC) and BMD of femurs measured by DXA.(E) Microstructural parameters of the distal femurs.BV/TV, bone volume per total volume.(F) Osteolineage cells parameters measured from Von Kossa staining images.Data are means G SD (C and D n = 4 mice/group; E,F n = 14 mice/group).*p < 0.05, **p < 0.01; Student's t test.

Figure 6 .Figure 6 .
Figure 6.EM-523, a non-antibiotic erythromycin derivative, promotes bone regeneration in aged mice (A and B) Aged WT mice were treated with ethanol control, ERM, or EM-523 intraperitoneally (A) daily for nine days or (B) twice a week for eight weeks.The distance from the CEJ-ABC was measured (left panels).The CEJ-ABC data were transformed to show the bone changes relative to the control, which was set as the baseline (right panels).Positive values (in mm) indicate bone growth relative to the baseline (control).(C) Representative IF images of maxillary sections obtained by fluorescence microscopy.Frozen maxillae sections were stained for DEL-1 (green), CD31 (red), and the nuclei were counterstained using DAPI (blue).T: tooth; PDL: periodontal ligament; B: alveolar bone.Scale bars, 100 mm.(D) Mean fluorescence intensity (MFI) values of DEL-1 were measured from randomly selected IF images from aged WT using the ImageJ software.All MFI values were normalized and presented relative to the MFI of the control group in aged WT mice, set as 1.(E)The mRNA levels of DEL1 from human periodontal ligament cells (hPDLCs) were quantified by qPCR on day seven after drug treatment.Data were normalized to GAPDH mRNA and plotted relative to the aged control group, set as 1. (F) hPDLCs were transiently transfected with hDEL1-promoter-Luc reporter plasmid, pre-treated for 30 min with ethanol control, ERM (10 mg/mL), or EM-523 (1 mg/ mL), and analyzed for luciferase activity.A renilla luciferase construct was co-transfected as an internal control for normalization.Data were calculated as fold change relative to ethanol control treatment, which was set as 1. (G) ChIP analysis of C/EBPb binding capacity at the DEL1 promoter in hPDLCs treated for 4 h with ethanol control, ERM, or EM-523 (10 and 1 mg/mL, respectively).Non-immunoprecipitated cell extracts were used as input samples.Data are plotted as a percentage of input.(H) Gene expression analysis using qPCR of primary mouse PDLCs on day nine of the osteogenic differentiation assay for the expression of the indicated genes.Cells were treated with vehicle control, macrolides, or EM-523 (ERM, CLR and AZM as 10 mg/mL; EM-523 as 1 mg/mL).Data were normalized to Gapdh mRNA and plotted relative to ethanol-treated control, set as 1. (I-N) Aged WT mice were treated with ethanol control, ERM, or EM-523 intraperitoneally for nine days.(I) Optical microscopy images of WT maxillary sections show PDL areas stained for ALP activity.T: Tooth; PDL: periodontal ligament; B: alveolar bone.Scale bars, 100 mm.(J) The ALP-positive region of each area was quantified and represented as a percentage of the total area.(K) Optical microscopy images of WT maxillary sections show PDL areas stained for TRAP.Arrowheads indicate TRAP+ cells in PDL space.T: Tooth; PDL: periodontal ligament; B: alveolar bone.Scale bars, 100 mm.(L) The number of TRAP + cells within the PDL space were counted and compared between indicated groups.(M) Representative IF images of maxillary sections obtained by fluorescence microscopy.Frozen maxillary sections were stained for alpha-smooth muscle actin (a-SMA, green), and the nuclei were counterstained using DAPI (blue).

Figure 7 .
Figure 7. EM-523 promotes bone regeneration and suppresses osteoclastogenesis in vitro (A) MC3T3-E1 cells were cultured in osteogenic differentiation medium treated with control solvent or indicated macrolides.(A, left panel) Representative images of mineralization nodules from osteoblasts, stained with Alizarin Red S (ARS) staining after 26 days of the osteogenic differentiation assay.(A, middle panel) The mineralization area in each culture was quantified and represented as a percentage of the total area.(A, right panel) Quantification of the ARS staining using 10% cetylpyridinium chloride (CPC) on day 26 of the osteogenic differentiation assay.The absorbance was measured at 560 nm.The control group was incubated with an osteogenic differentiation medium only.(B) MC3T3-E1 cells were cultured in osteogenic differentiation medium treated with ethanol, ERM, or EM-523.(B, left panel) Representative images of mineralization nodules from osteoblasts, stained with ARS after 20 days of the osteogenic differentiation assay.(B, right panel) The mineralization area in each culture was quantified and presented as a percentage of the total area.The control group was incubated with an osteogenic differentiation medium only.(C) hPDLCs were cultured in osteogenic differentiation medium with control solvent or indicated macrolides.(C, left panel) Representative images of hPDLCs cultures, stained with alkaline phosphatase stain after 26 days of the osteogenic differentiation assay.Scale bars, 200 mm.(C, right panel) The ALP-positive region of each culture was quantified and represented as a percentage of the total area.(D) Gene expression analysis of hPDLCs on day nine of the osteogenic differentiation assay using qPCR.RUNX2, SP7, and BGLAP, as representative markers for early, middle, and late osteogenic markers, respectively.Data were normalized to GAPDH mRNA and plotted relative to ethanol-treated control, set as 1. (E) hMSCs were cultured in osteogenic differentiation medium with ethanol or indicated macrolides (10 mg/mL).(E, upper panel) Representative images of mineralization nodules from osteoblasts, stained with Alizarin Red S after 26 days of the osteogenic differentiation assay.(E, middle left panel) The mineralization area in each culture was quantified and presented as a percentage of the total area.(E, middle right panel) Quantification of the Alizarin Red S staining using 10% CPC on day 26 of the osteogenic differentiation assay.(E, lower panel) Gene expression analysis using qPCR of human MSCs on day nine of the osteogenic differentiation assay for the expression of the indicated genes.Data were normalized to GAPDH mRNA and plotted relative to ethanol control, set as 1. (F) Bone marrow cells were collected from WT mice for the osteoclastogenesis assay.Cells were treated with DEL-1-Fc (1 mg/mL), equal molar amount of Fc control, control solvent, or indicated macrolides for one week.(F, upper panel) Representative images of the culture well from each group after TRAP staining are shown.Histological images of TRAP + MNCs.Scale bars, 200 mm.(F, lower left panel) The average size of TRAP + MNCs was measured using the ImageJ software.(F, lower right panel) The percentage of TRAP + area per total area was measured using the ImageJ software.(G) Primary osteoblastic progenitor cells were collected from Del1 À/À mice.Primary osteoblastic progenitor cells were cultured in osteogenic differentiation media administered with DEL-1-Fc, equal molar amount of Fc control, control solvent, or indicated macrolides for 26 days.(G, upper panel) Representative images of mineralization nodules from osteoblasts, stained with Alizarin Red S after 26 days of the osteogenic differentiation assay.(G, lower panel) Gene expression analysis was completed on day nine of the osteogenic differentiation assay using qPCR.Runx2, Sp7, and Bglap, as representative markers for early, middle, and late osteogenic markers, respectively.Data were normalized to GAPDH mRNA and plotted relative to ethanol control, set as 1. (H) Bone marrow cells were collected from Del1 À/À mice for the osteoclastogenesis assay.Cells were treated with DEL-1-Fc, equal molar amount of Fc control, control solvent, or indicated macrolides for one week.(H, upper panel) Representative images of the culture well from each group after TRAP staining were shown.Histological images of TRAP + MNCs.Scale bars, 200 mm.(H, lower left panel) The average size of TRAP + MNCs was measured using the ImageJ software.(H, lower right panel) The percentage of TRAP + area per total area was measured using the ImageJ software.Data are means G S.D. (A-C, n = 4 sets of cultures/group; D, n = 5 sets of cultures/group; E middle and lower panel, n = 4 sets of cultures/group; F-H, n = 5 sets of cultures/group).*p < 0.05, **p < 0.01; one-way ANOVA and Bonferroni's test.