Abstract

   Impaired fracture healing is a common complication in type 2 diabetes
   mellitus (T2DM), with limited effective treatments. This study
   investigates the role of macrophages in bone repair and introduces a
   novel therapeutic strategy. Reduced glutaminase (GLS) expression and
   glutaminolysis are found in macrophages from T2DM mice and monocytes
   from T2DM patients. Specific deletion of GLS in macrophages altered
   their phenotypes and delayed fracture healing in mice. Mechanistically,
   GLS deficiency reduced α‐ketoglutarate (α‐KG) levels in macrophages,
   which impairs bone morphogenetic protein 2 (BMP2) production by
   increasing cytosine methylation on the promoter, ultimately hindering
   osteogenic differentiation of bone marrow mesenchymal stem cells.
   Importantly, while systemic α‐KG supplementation deteriorates fracture
   healing in T2DM mice, a targeted delivery of α‐KG using
   α‐KG@Cy5.5@ALN‐Liposome to macrophages in bone markedly improves
   fracture healing. These findings underscore the critical role of
   macrophage glutaminolysis in fracture healing and propose targeted α‐KG
   delivery as a promising therapeutic intervention for improving fracture
   repair in T2DM patients.

   Keywords: type 2 diabetes, macrophages, fracture healing,
   glutaminolysis, DNA methylation
     __________________________________________________________________

   The study reveals that glutaminolysis in macrophages is inhibited under
   type 2 diabetes mellitus (T2DM) conditions, which impedes fracture
   healing by reducing bone morphogenetic protein 2 (BMP2) production
   through increased cytosine methylation on the promoter. The
   administration of alendronate‐modified bone‐targeting liposomes
   (α‐KG@Cy5.5@ALN‐Lipo), which deliver α‐ketoglutarate (α‐KG) to bone
   macrophages, significantly enhances fracture healing in a T2DM mouse
   model.

   graphic file with name ADVS-12-2415667-g005.jpg

1. Introduction

   In 2021, diabetes affected ≈537 million adults globally, with 90% of
   patients having type 2 diabetes mellitus (T2DM), according to the
   international diabetes federation (IDF), and the number is expected to
   rise to 643 million by 2030. T2DM is characterized by hyperglycemia and
   insulin resistance, which cause serious complications in the heart,
   kidneys, eyes, and feet.^[ [48]^1 , [49]^2 ^] Emerging evidence also
   points to its detrimental effects on skeletal health. Although patients
   with T2DM often have maintained bone mineral density (BMD), they face a
   significantly higher risk of fractures, particularly in the hip,
   vertebrae, and femur.^[ [50]^3 , [51]^4 ^] Furthermore, T2DM is
   associated with impaired fracture healing and a higher incidence of
   non‐unions compared to non‐diabetic individuals.^[ [52]^5 , [53]^6 ^]
   While anti‐diabetic drugs are widely used to manage fractures in T2DM
   patients, their primary effect is glucose level regulation, thus the
   effectiveness remains limited and even controversial.^[ [54]^7 ^]
   Clinical evidence identifies certain medications, including
   sulfonylureas, thiazolidinediones, and sodium‐glucose co‐transporter‐2
   inhibitors (SGLT2‐i), may actually increase fracture risk.^[ [55]^7 ^,
   [56]^8 ^] This underscores the urgent need to clarify the mechanisms
   behind impaired fracture healing in T2DM and develop more effective
   therapies.

   Fracture healing is a complex process involving various cell types,
   including immune cells, bone marrow mesenchymal stem cells (BMSCs), and
   vascular endothelial cells.^[ [57]^9 , [58]^10 ^] As a key component of
   the bone microenvironment and one of the first immune cells to reach
   fracture sites, macrophages secret proinflammatory cytokines including
   TNFα and CCL2 to recruit BMSCs in the early phase.^[ [59]^11 , [60]^12
   ^] As the healing proceeds, macrophages shift to a reparative
   phenotype, releasing prorepair cytokines such as TGF‐β and BMP‐2, which
   facilitate fracture healing.^[ [61]^13 , [62]^14 ^] However, T2DM
   exhibits a systemic chronic inflammation caused by the impaired
   transition of macrophages from a proinflammatory state to a reparative
   state, and the overactivation of inflammation and deficiency of
   prorepair cytokines secretion in macrophages impair tissue repair under
   T2DM.^[ [63]^15 , [64]^16 ^] Nevertheless, the precise mechanisms
   disrupting macrophage phenotype transition and function in T2DM have
   not been fully elucidated.

   Metabolism has been reported closely associated with macrophage
   phenotype and function.^[ [65]^17 , [66]^18 ^] Glutamine, the most
   abundant non‐essential amino acid in plasma, is converted by
   glutaminase (GLS) in mitochondria into glutamate, which is further
   deaminated by glutamate dehydrogenase (GDH) into α‐ketoglutarate
   (α‐KG).^[ [67]^19 ^] α‐KG links glutaminolysis to the tricarboxylic
   acid (TCA) cycle and provides a key cofactor for ten‐eleven
   translocation hydroxylases (TETs), a kind of dioxygenase converts
   5‐methyl‐cytosine (5mC) to 5‐hydroxy‐methyl‐cytosine (5hmC) in DNA and
   thus regulates chromatin accessibility and gene expression.^[ [68]^20 ,
   [69]^21 ^] Furthermore, α‐KG controls inflammation and immune responses
   of macrophages in cancer and chronic inflammation‐related disease
   through TET enzymes mediated DNA demethylation.^[ [70]^22 , [71]^23 ,
   [72]^24 , [73]^25 ^] Moreover, lower plasma levels of glutamine and a
   reduced glutamine‐to‐glutamate ratio are associated with a higher risk
   of T2DM,^[ [74]^26 ^] and glutamine supplementation has been shown to
   alleviate inflammation in obesity by reducing glycolysis in adipocytes
   and suppressing proinflammatory adipokine release, leading to improved
   insulin sensitivity and normalized glucose levels in T2DM.^[ [75]^27 ,
   [76]^28 ^] However, the alterations of macrophage glutaminolysis under
   T2DM and its role in impaired fracture healing remain poorly
   understood.

   In this study, we reveal that inhibited glutaminolysis in macrophages
   contributes to impaired fracture healing in T2DM, which can be rescued
   by targeted delivery of α‐KG to macrophages in bone. We first observed
   that macrophage glutaminolysis is inhibited under T2DM conditions,
   including high‐glucose and palmitic acid‐treated cells, T2DM mice, and
   T2DM patients. Additionally, GLS deletion in macrophages impaired
   fracture healing in vivo and hindered the osteogenic differentiation of
   BMSCs in vitro. Mechanistically, GLS deletion disrupted macrophage
   phenotype and increased proinflammatory cytokine secretion. Moreover,
   GLS deficiency reduced α‐KG levels, leading to decreased TET enzyme
   activity and increased DNA methylation at the Bmp2 promoter, resulting
   in lower BMP2 production, and supplementation with α‐KG restored BMP2
   secretion in GLS‐deficient macrophages. Finally, although systemic α‐KG
   administration worsened fracture healing due to increased mortality,
   targeted delivery of α‐KG to macrophages in bone reversed impaired
   healing in T2DM mice. These findings identify inhibited glutaminolysis
   in macrophages as a pivotal factor in impaired fracture healing and
   suggest that targeted α‐KG delivery presents a promising therapeutic
   approach for improving fracture healing in T2DM.

2. Results

2.1. Glutaminolysis is Inhibited in Macrophages of T2DM Mice and Patients

   Previous studies have demonstrated that fracture healing is impaired in
   patients with T2DM, often resulting in non‐union.^[ [77]^29 ^] To
   validate these findings in a murine model, we performed femur fracture
   surgery on a high‐fat diet (HFD)‐streptozotocin (STZ) induced T2DM mice
   and normal control (NM) mice (Figure [78] 1A; Figure [79]S1A,
   Supporting Information). The T2DM model was successfully established,
   as evidenced by increased body weight, food intake, water consumption,
   and impaired glucose tolerance and insulin tolerance (Figure [80]S1B–F,
   Supporting Information). At 10 days post‐fracture, T2DM mice exhibited
   reduced bone volume fraction (BV/TV) and trabecular number (Tb.N),
   alongside increased trabecular separation (Tb.Sp) compared to NM mice
   (Figure [81]1B,C). Furthermore, T2DM mice displayed smaller calluses
   with diminished cartilage content (Figure [82]S2A,B, Supporting
   Information). These results confirm that fracture healing is impaired
   in T2DM mice.

Figure 1.

   Figure 1
   [83]Open in a new tab

   Glutaminolysis is inhibited in macrophages of T2DM mice and patients.
   A) Flow diagram illustrating the experimental design of fracture in NM
   and T2DM mice. B) Representative images and C) quantitative analysis of
   micro‐CT images demonstrating callus mineralization (n = 6). D)
   Experimental design for PMs isolation from NM and T2DM mice. E) Protein
   levels of GLS in PMs from T2DM and NM mice (n = 6). F) Protein levels
   of GLS in BMDMs treated with high glucose plus palmitic acid (HG+PA; 25
   mm glucose plus 400 µm palmitic acid) or low glucose (LG; 5 mm glucose
   plus 20 mm mannitol) for 48 h (n = 6). G) GLS activity in BMDMs
   subjected to HG+PA or LG for 48 h (n = 6). H) Glutamine consumption in
   BMDMs treated with HG+PA or LG for 48 h (n = 9). I) Glutamine‐targeted
   metabolomics profiling showing levels of glutamine, glutamate, α‐KG,
   and succinate in BMDMs treated with HG+PA or LG for 24 h (n = 6). J)
   Heatmap derived from the GEO dataset ([84]GSE253502), illustrating gene
   expression associated with glutamine metabolism in PBMCs from NM (N
   group) and T2DM (G group) patients (n = 22). K) Experimental design for
   monocyte isolation and subsequent assays from NM and T2DM patients (n =
   10 in NM and n = 20 in T2DM). L) Bubble chart depicting BMI, fasting
   blood glucose levels, and GLS mRNA levels in monocytes from
   participants. M) Relative mRNA levels of GLS in monocytes from
   normoglycemic versus hyperglycemic individuals (n = 10 normoglycemic
   individuals and n = 20 hyperglycemic individuals). N) Relative mRNA
   levels of GLS in monocytes from normal weight versus overweight
   individuals (n = 19 normal weight individuals and n = 11 overweight
   individuals). O) Linear regression analysis of plasma glutamine levels
   versus fasting blood glucose levels in participants (n = 30). P) Linear
   regression analysis of cellular glutamine and α‐KG levels in monocytes
   versus fasting blood glucose levels in participants (n = 30). Data are
   shown as mean ± SD; *p < 0.05, **p < 0.01, and ***p < 0.001 by unpaired
   Student's two‐tailed t‐test.

   Macrophages are crucial for fracture healing, and their function is
   influenced by glutamine metabolism.^[ [85]^22 , [86]^30 ^] To explore
   the effects of T2DM on macrophage glutaminolysis, we first isolated
   peritoneal macrophages (PMs) from NM and T2DM mice (Figure [87]1D). GLS
   protein levels were significantly lower in PMs from T2DM mice compared
   to NM mice (Figure [88]1E). Additionally, we isolated bone
   marrow‐derived macrophages (BMDMs) from 6‐week‐old C57BL/6 mice and
   treated cells with high glucose and palmitic acid (HG+PA; 25 mm glucose
   and 400 µm palmitic acid) for 48 h to imitate T2DM conditions in
   vitro.^[ [89]^31 ^] The results showed that HG+PA treatment
   significantly reduced GLS expression (Figure [90]1F). Moreover, GLS
   activity and glutamine consumption were markedly lower in HG+PA‐treated
   BMDMs compared to those treated with low glucose (LG)
   (Figure [91]1G,H). We also measured glutaminolysis metabolites,
   including glutamine, glutamate, and two TCA cycle intermediates, α‐KG
   and succinate, which are critical for macrophage function.^[ [92]^22 ,
   [93]^32 ^] The results indicated an increase in glutamine levels
   following HG+PA exposure, while downstream metabolites, including
   glutamate, α‐KG, and succinate, were significantly decreased,
   confirming inhibited glutaminolysis in macrophages during T2DM
   (Figure [94]1I).

   Furthermore, transcription profiling of peripheral blood mononuclear
   cells (PBMCs) from patients with T2DM and healthy controls
   ([95]GSE253502) revealed a significant reduction in the expression of
   genes associated with glutamine metabolism in T2DM patients
   (Figure [96]1J). We then collected plasma and monocytes from T2DM
   patients and healthy controls (Figure [97]1K), and conducted a
   comprehensive analysis of body mass index (BMI), fasting blood glucose
   levels, and GLS mRNA levels in monocytes (Figure [98]1L). The results
   indicated that GLS expression in monocytes was significantly lower in
   individuals with hyperglycemia (fasting glucose > 6.0 mmol L^−1)
   compared to those with normal blood glucose levels (fasting glucose <
   6.0 mmol L^−1, Figure [99]1M). Additionally, GLS expression was
   decreased in overweight participants (BMI > 25) relative to
   normal‐weight individuals (BMI < 25, Figure [100]1N, p = 0.0993).
   Linear regression analysis further demonstrated that plasma glutamine
   levels, along with cellular glutamine and α‐KG levels in monocytes,
   were negatively correlated with fasting blood glucose levels
   (Figure [101]1O,P). Together, these results suggest that macrophage
   glutaminolysis is inhibited in T2DM, primarily due to reduced
   expression and activity of GLS.

2.2. Specific Deletion of GLS in Macrophages Impairs Fracture Healing in Mice

   To explore the role of macrophage glutaminolysis in fracture healing,
   we generated Gls^fl/fl ;Lyz2‐Cre mice and assessed the fracture healing
   process in vivo (Figure [102] 2A). In BMDMs from Gls^fl/fl ;Lyz2‐Cre
   mice, we noted significant reductions in both mRNA and protein levels
   of GLS compared to Gls^fl/fl controls (Figure [103]S3A,B, Supporting
   Information). Micro‐CT imaging was used to evaluate mineralized callus
   formation, revealing that Gls^fl/fl ;Lyz2‐Cre mice had a smaller hard
   callus 10 days post‐fracture (Figure [104]2B). Quantitative analysis
   showed a decrease in BV/TV, Tb.Th, and Tb.N, while Tb.Sp was increased
   in Gls^fl/fl ;Lyz2‐Cre mice compared to Gls^fl/fl controls
   (Figure [105]2C). Safranin O staining indicated a significant reduction
   in cartilage content within the callus of Gls^fl/fl ;Lyz2‐Cre mice
   (Figure [106]2D). We used the callus index, defined as the maximum
   callus diameter divided by the diameter of the adjacent diaphysis, to
   evaluate size differences between Gls^fl/fl ;Lyz2‐Cre and Gls^fl/fl
   mice.^[ [107]^33 ^] Consistently, the callus index in Gls^fl/fl
   ;Lyz2‐Cre mice was smaller than that in Gls^fl/fl controls
   (Figure [108]2E). Additionally, a three‐point bending test assessed the
   mechanical properties of the callus formed 28 days post‐fracture.
   Results showed that fractured femurs from Gls^fl/fl ;Lyz2‐Cre mice had
   reduced bone strength, stiffness, and toughness compared to those from
   Gls^fl/fl mice (Figure [109]2F–H). Collectively, these findings
   underscore the critical role of macrophage glutaminolysis in fracture
   healing in mice.

Figure 2.

   Figure 2
   [110]Open in a new tab

   Specific deletion of GLS in macrophages impairs fracture healing in
   mice. A) Experimental design for Gls^fl/fl and Gls^fl/fl;Lyz2‐Cre mice
   to analyze fracture healing. B) Representative micro‐CT images and C)
   quantitative analysis showing callus mineralization 10 days
   post‐fracture in Gls^fl/fl and Gls^fl/fl;Lyz2‐Cre mice (n = 6). D)
   Representative Safranin O staining images of femur calluses in
   Gls^fl/fl and Gls^fl/fl;Lyz2‐Cre mice at 10 days post‐fracture (n = 6).
   Red indicates cartilage, and green indicates bone (Scale bars, 1 mm
   (left panels) and 0.5 mm (right panels)). E) Quantitative analysis of
   callus index in femurs of Gls^fl/fl and Gls^fl/fl;Lyz2‐Cre mice at 10
   days post‐fracture (n = 6). F) Strength, G) toughness, and H) stiffness
   of callus at 28 days post‐fracture measured by three‐point bending
   tests (n = 6). Data are shown as mean ± SD; *p < 0.05, **p < 0.01, and
   ***p < 0.001 by unpaired Student's two‐tailed t‐test.

2.3. GLS Deficiency in Macrophages Hinders Osteogenic Differentiation of
BMSCs

   Upon fracture healing, the osteogenic differentiation of BMSCs is
   regulated by cytokines released from macrophages.^[ [111]^34 ^] To
   assess the effects of GLS deletion in macrophages on the
   differentiation of BMSCs, we collected conditioned media (CM) from
   BMDMs isolated from Gls^fl/fl and Gls^fl/fl;Lyz2‐Cre mice, designated
   as WT‐CM and KO‐CM, respectively. CMs were then mixed with fresh
   osteogenic induction medium at a 1:1 ratio to culture BMSCs for
   evaluations (Figure [112] 3A). BMSCs treated with KO‐CM exhibited
   significantly lower mRNA and protein levels of osteogenic markers,
   including osteocalcin (OCN), osteopontin (OPN), and runt‐related
   transcription factor 2 (RUNX2) compared to those treated with WT‐CM
   (Figure [113]3B,C). Alkaline phosphatase (ALP) activity was also
   reduced in BMSCs exposed to KO‐CM compared to the WT‐CM group
   (Figure [114]3D). ALP staining confirmed fewer ALP‐positive cells in
   the KO‐CM group relative to the WT‐CM group (Figure [115]3E).
   Additionally, Alizarin Red S (ARS) staining revealed that KO‐CM
   inhibited calcium nodule formation, as evidenced by smaller sizes and
   fewer nodules in BMSCs treated with KO‐CM compared to the WT‐CM group,
   indicating impaired mineralization (Figure [116]3F). Collectively,
   these findings suggest that GLS deficiency in macrophages may alter the
   secretory profile, thereby hindering the osteogenic differentiation of
   BMSCs.

Figure 3.

   Figure 3
   [117]Open in a new tab

   GLS deficiency in macrophages hinders the osteogenic differentiation of
   BMSCs. A) Schematic representation of BMDMs‐CM preparation and the
   experimental design for co‐culture and assays at various time points.
   B) mRNA levels of osteogenic markers (Ocn, Opn, and Runx2) in BMSCs
   co‐cultured with CM for 7 days, normalized to β‐actin (n = 3). C)
   Protein levels of osteogenic markers (OCN, OPN, and RUNX2) in BMSCs
   co‐cultured with CM for 28 days, normalized to β‐actin (n = 6). D) ALP
   activity measured in BMSCs co‐cultured with CM for 7 days (n = 6). E)
   Representative images of ALP staining in BMSCs co‐cultured with CM for
   7, 14, and 21 days (n = 3). F) Representative images of Alizarin Red S
   staining and quantitative analysis of BMSCs co‐cultured with CM for 14,
   21, and 28 days (n = 3). Data are shown as mean ± SD; *p < 0.05, **p <
   0.01, and ***p < 0.001 by unpaired Student's two‐tailed t‐test.

2.4. Aberrant Glutaminolysis Disrupts the Macrophages' Phenotype

   Macrophages participate in all stages of fracture healing, displaying a
   variety of phenotypes and functions, which are closely tied to
   metabolic processes.^[ [118]^17 ^] To investigate how glutaminolysis
   influences macrophage phenotypes, we isolated BMDMs from Gls^fl/fl and
   Gls^fl/fl;Lyz2‐Cre mice and performed transcriptomic analysis via
   high‐throughput RNA sequencing (RNA‐Seq) (Figure [119] 4A).
   Differentially expressed genes (DEGs) were identified based on fold
   changes >1.5 and p‐values <0.05. The analysis showed that GLS deletion
   resulted in the downregulation of 195 genes and the upregulation of 217
   genes (Figure [120]4B; Figure [121]S4A,B, Supporting Information). Gene
   ontology (GO) enrichment analysis revealed significant changes in
   macrophage functions, including response to stimulus, cytokine
   production, and cellular homeostasis, between WT and KO macrophages
   (Figure [122]4C). Kyoto Encyclopedia of Genes and Genomes (KEGG)
   enrichment analysis highlighted significant enrichment of pathways
   associated with inflammation, including cytokines interactions,
   chemokine signaling, TNF signaling, and osteoclast differentiation
   among the top 20 enriched pathways (Figure [123]4D). Moreover, Gene Set
   Enrichment Analysis (GSEA) indicated an upregulation in the cellular
   response to environmental stimulus in KO macrophages compared to WT
   (Figure [124]4E). Notably, despite similar baseline levels, GLS
   deficiency led to increased mRNA levels of pro‐inflammatory markers
   (Inos, Tnfα, Il1β, and Il6) in response to lipopolysaccharide (LPS)
   stimulation compared to WT macrophages (Figure [125]4F). Conversely,
   mRNA levels of anti‐inflammatory markers (Arg1, Fizz1, Mrc, and Il10)
   were downregulated in KO macrophages upon interleukin‐4 (IL‐4)
   stimulation (Figure [126]4G). These results are consistent with
   findings from Liu et al., which indicated that glutamine deprivation
   disrupts macrophage phenotype.^[ [127]^22 ^] Correspondingly, levels of
   proinflammatory cytokines, including TNFα and CCL2, were increased in
   the supernatant of KO macrophages compared to WT under LPS stimulation
   (Figure [128]4H; Figure [129]S4C, Supporting Information). Moreover,
   GLS deficiency led to an increase in the percentage of pro‐inflammatory
   (F4/80^+CD86^+) macrophages while simultaneously decreasing the
   percentage of anti‐inflammatory (F4/80^+CD206^+) macrophages
   (Figure [130]4I,J). In summary, these results indicate that aberrant
   glutaminolysis disrupts macrophage phenotype and triggers severe
   inflammation, potentially hindering fracture healing in the early
   stage.

Figure 4.

   Figure 4
   [131]Open in a new tab

   Aberrant glutaminolysis disrupts the macrophages' phenotype. A)
   Schematic overview of RNA‐seq analysis of BMDMs isolated from Gls^fl/fl
   and Gls^fl/fl;Lyz2‐Cre mice (WT and KO BMDMs) (n = 4). B) Heatmap
   displaying differentially regulated genes in WT and KO BMDMs. C) GO
   enrichment analysis highlighting biological processes related to
   inflammation, cytokine production, and cellular homeostasis in WT and
   KO BMDMs. D) KEGG pathway enrichment analysis reveals the top 20
   pathways enriched in WT and KO BMDMs. E) GSEA of genes involved in
   cellular responses to environmental stimulus. F) Pro‐inflammatory
   cytokines (Inos, Tnfa, Il1b, and Il6) mRNA expression in BMDMs
   stimulated with 100 ng mL^−1 LPS for 6 h and G) anti‐inflammatory
   cytokines (Arg1, Fizz1, Mrc, and Il10) mRNA expression in BMDMs
   stimulated with 20 ng mL^−1 IL‐4 for 24 h (n = 3). H) Concentrations of
   TNFα in the supernatants of WT and KO BMDMs treated with vehicle or LPS
   for 6 h (n = 4). I) Representative images and J) quantitative analysis
   showing the percentage of pro‐inflammatory macrophages (F4/80^+CD86^+)
   and anti‐inflammatory macrophages (F4/80^+CD206^+) in BMDMs stimulated
   with 100 ng mL^−1 LPS for 6 h or 20 ng mL^−1 IL‐4 for 24 h (n = 3).
   Data are shown as mean ± SD; **p < 0.01 and ***p < 0.001 by one‐way
   ANOVA.

2.5. GLS Deletion in Macrophages Reduces BMP2 Production Via Enhanced
Cytosine Methylation on Promoter

   Macrophages are involved in fracture healing through cytokine secretion
   and interact with various cells including BMSCs.^[ [132]^30 , [133]^35
   ^] To investigate the role of GLS in macrophage cytokine secretion, we
   then collected supernatants from WT and KO BMDMs to conduct secretomics
   analysis, which assessed differential cytokine secretion (Figure [134]
   5A). Differentially secreted proteins were defined by fold changes >1.3
   and p‐values <0.05. GLS deletion led to the upregulation of 123
   cytokines and the downregulation of 200 cytokines (Figure [135]5B;
   Figure [136]S5A,B, Supporting Information). KEGG enrichment showed that
   cytokine interaction including ECM‐receptor interaction and
   cytokine–cytokine receptor interaction was significantly altered in KO
   macrophages (Figure [137]S5C, Supporting Information). In addition, the
   results revealed that the biological processes associated with these
   differentially secreted proteins involved bone mineralization,
   osteoblast differentiation, and ossification, all were critical to
   fracture healing (Figure [138]5C–E). These findings suggest that
   impaired glutaminolysis in macrophages disrupt fracture healing by
   altering cytokine secretion.

Figure 5.

   Figure 5
   [139]Open in a new tab

   GLS deletion in macrophages reduces BMP2 production via enhanced
   cytosine methylation on promoter. A) Experimental design for analyzing
   the secretome of supernatants collected from WT and KO BMDMs (n = 4).
   B) Volcano plot illustrating differentially secreted proteins in the
   supernatants of WT and KO BMDMs. C) GO enrichment analysis of
   biological processes related to fracture healing in macrophages. D)
   Heatmap displaying differentially secreted proteins associated with
   osteoblast differentiation and proliferation in WT and KO BMDMs. E)
   Chord diagram showing the relationships between genes and biological
   processes relevant to fracture healing. F) mRNA levels of Bmp2 in WT
   and KO BMDMs (n = 3). G) Protein levels of BMP2 in WT and KO BMDMs (n =
   6). H) BMP2 concentrations in the supernatants of WT and KO BMDMs
   treated with vehicle or DM‐α‐KG, measured by ELISA (n = 3). I)
   Experimental design for DNA methylation assays in BMDMs. J) Cellular
   α‐KG levels in WT and KO BMDMs treated with vehicle or DM‐α‐KG (n = 3).
   K) TET activity levels in WT and KO BMDMs treated with vehicle or
   DM‐α‐KG (n = 3). L) Oxidative cytosine 5mC contents and M) reduced
   cytosine 5hmC contents in WT and KO BMDMs treated with vehicle or
   DM‐α‐KG (n = 6). N) Cytosine methylation levels in the promoter region
   of Bmp2 in WT and KO BMDMs treated with vehicle or DM‐α‐KG (n = 3).
   Data are shown as mean ± SD; *p < 0.05, **p < 0.01, and ***p < 0.001 by
   one‐way ANOVA.

   In the GO Chordal graph, bone morphogenetic protein 2 (BMP2), a crucial
   cytokine for fracture healing and a member of the TGF‐beta
   superfamily,^[ [140]^36 ^] was found to correlate with nearly all
   enriched biological processes related to fracture healing
   (Figure [141]5E). Then we evaluated BMP2 expression in BMDMs and
   observed that IL‐4 exposure increased BMP2 expression, while GLS
   deletion resulted in lower BMP2 mRNA and protein levels compared to WT
   macrophages (Figure [142]5F,G). Correspondingly, although BMP2
   secretion was upregulated in BMDMs following IL‐4 treatment, KO BMDMs
   showed reduced BMP2 levels both at baseline and after IL‐4 exposure.
   Importantly, the addition of DM‐α‐KG restored BMP2 secretion in
   GLS‐deficient macrophages (Figure [143]5H).

   To investigate how GLS deficiency affects BMP2 secretion, we focused on
   α‐KG dependent dioxygenases, particularly TET enzymes, which play a
   crucial role in DNA methylation and gene expression regulation
   (Figure [144]5I). Our prior results indicated that inhibited
   glutaminolysis in macrophages decreases α‐KG levels in T2DM, then we
   measured α‐KG levels in GLS‐deficient macrophages and observed a
   significant reduction of α‐KG in KO macrophages compared to WT group
   (Figure [145]5J). As a result, TET activity was diminished in KO
   macrophages due to lower α‐KG levels, which could be restored with
   DM‐α‐KG supplementation (Figure [146]5K). Additionally, GLS deletion
   increased global DNA 5‐methylcytosine (5‐mC) levels and decreased
   5‐hydroxymethylcytosine (5‐hmC) levels in macrophages compared to the
   WT group, indicating enhanced cytosine methylation in DNA, which was
   reversed by DM‐α‐KG supplementation (Figure [147]5L,M).

   For a comprehensive and quantitative analysis of DNA methylation, we
   utilized Reduced Representation Bisulfite Sequencing (RRBS), a
   technique designed for quantifying DNA methylation. Our results showed
   that the methylation level of the Bmp2 promoter was elevated in KO
   BMDMs but decreased upon DM‐α‐KG treatment (Figure [148]5N). Together,
   these results confirm that GLS deficiency in macrophages inhibits BMP2
   production by lowering α‐KG levels and inhibiting TETs activity,
   ultimately increasing cytosine methylation at the Bmp2 promoter.

2.6. Targeted Delivery of α‐KG to Macrophages in Bone Rescues Fracture
Healing of T2DM Mice

   To assess the effect of α‐KG on fracture healing in T2DM, we induced
   femur fractures in T2DM mice and administered α‐KG orally at a
   concentration of 2% (w/v) in water (Figure [149] 6A).^[ [150]^37 ^]
   Unexpectedly, α‐KG supplementation led to weakness, weight loss, and
   increased post‐fracture mortality (Figure [151]S6A,B, Supporting
   Information).^[ [152]^38 ^] We subsequently collected femurs from the
   surviving T2DM mice and analyzed calluses using micro‐CT and Safranin O
   staining. The results revealed that systemic α‐KG supplementation
   impaired fracture healing by negatively affecting callus formation and
   mineralization (Figure [153]6B; Figure [154]S6C–G, Supporting
   Information).

Figure 6.

   Figure 6
   [155]Open in a new tab

   Targeted delivery of α‐KG to macrophages in bone rescues fracture
   healing of T2DM mice. A) Experimental design for oral α‐KG
   supplementation and subsequent analysis of fracture healing in T2DM
   mice. B) Representative micro‐CT images and quantitative analysis
   demonstrating callus mineralization at 10 days post‐fracture in T2DM
   mice treated with either α‐KG or vehicle (n = 7 for the T2DM + Vehicle
   group and n = 3 for the T2DM + α‐KG group). C) Schematic representation
   of a bone‐targeted liposome, decorated with alendronate (ALN),
   containing α‐KG and Cy5.5. D) Representative images of BMDMs treated
   with α‐KG@Cy5.5@ALN‐Lipo, indicating successful uptake of the liposome
   containing α‐KG and Cy5.5 (Scale bars, 20 µm). E) Cellular α‐KG levels
   in BMDMs treated with Cy5.5@ALN‐Lipo or α‐KG@Cy5.5@ALN‐Lipo (n = 3). F)
   Experimental design for administering Cy5.5@ALN‐Lipo or
   α‐KG@Cy5.5@ALN‐Lipo to T2DM mice, followed by the analysis of fracture
   healing. G) Representative micro‐CT images and H) quantitative analysis
   of callus formation at 10 days post‐fracture in T2DM mice treated with
   Cy5.5@ALN‐Lipo or α‐KG@Cy5.5@ALN‐Lipo (n = 6). I) Calculated callus
   index of bone in T2DM mice treated with Cy5.5@ALN‐Lipo or
   α‐KG@Cy5.5@ALN‐Lipo at 10 days post‐fracture (n = 6). J) Representative
   images of Safranin O staining in the callus of T2DM mice treated with
   Cy5.5@ALN‐Lipo or α‐KG@Cy5.5@ALN‐Lipo at 10 days post‐fracture,
   cartilage (red) and bone (green) (n = 6) (Scale bars, 1 mm (left
   panels) and 0.5 mm (right panels)). K) Representative images and
   quantitative analysis of Masson staining in the callus of T2DM mice
   treated with Cy5.5@ALN‐Lipo or α‐KG@Cy5.5@ALN‐Lipo at 28 days
   post‐fracture, mineralized bone (blue) and cartilage (transparent) (n =
   6) (Scale bars, 1 mm (left panels) and 0.5 mm (right panels)). Data are
   shown as mean ± SD; *p < 0.05, **p < 0.01, and ***p < 0.001 by unpaired
   Student's two‐tailed t‐test.

   Given these results, we next explored the targeted delivery of α‐KG to
   macrophages in bone to evaluate its effects on fracture healing. We
   selected liposomes as the drug delivery system due to their capacity
   for phagocytosis by macrophages and the potential for enhanced
   targeting efficiency through modifications, such as the decorations of
   targeting peptides.^[ [156]^39 ^] Alendronate, known for its high
   affinity for bone, was chosen as the targeting peptide for our study.^[
   [157]^40 , [158]^41 ^] We prepared bone‐targeted liposomes
   (α‐KG@Cy5.5@ALN‐Lipo) using the film‐rehydration method, combining
   DSPE‐PEG‐ALN, lecithin, α‐KG, and Cy5.5 as a labeling agent
   (Figure [159]6C). Transmission electron microscopy (TEM) confirmed the
   spherical morphology of the α‐KG@Cy5.5@ALN‐Lipo (Figure [160]S7A,
   Supporting Information). Dynamic light scattering (DLS) analysis
   indicated that liposome sizes remained stable following α‐KG loading
   (diameters: 122.7 ± 3.35 nm vs 125.7 ± 1.64 nm; polydispersity index
   (PDI): 0.199 ± 0.004 vs 0.197 ± 0.008). Additionally, the zeta
   potential of α‐KG@Cy5.5@ALN‐Lipo was significantly lower than that of
   Cy5.5@ALN‐Lipo (−25.9 ± 1.65 mV vs −15.7 ± 0.94 mV) (Figure [161]S7B–D,
   Supporting Information).

   To assess the release profile of α‐KG from liposomes, we measured α‐KG
   concentrations in phosphate‐buffered saline (PBS) containing
   α‐KG@Cy5.5@ALN‐Lipo. Cumulative release data revealed that ≈60% of α‐KG
   was released within the first three days, while the remaining α‐KG was
   released gradually over the following four days (Figure [162]S7E,
   Supporting Information). We evaluated the bone‐targeting efficiency of
   the liposomes using an in vivo imaging system (IVIS) to analyze the
   biodistribution of α‐KG@Cy5.5@ALN‐Lipo. Mice were sacrificed at 6 and
   24 h post‐tail vein injection, and major organs, including the brain,
   heart, lungs, liver, spleen, kidneys, and hind limbs, were collected.
   The results indicated a significant accumulation of liposomes in the
   bone at 24 h (Figure [163]S7F, Supporting Information). Additionally,
   Cy5.5 fluorescence signals were detected in macrophages incubated with
   α‐KG@Cy5.5@ALN‐Lipo, confirming successful uptake by macrophages
   (Figure [164]6D; Figure [165]S7G, Supporting Information). Moreover,
   macrophages treated with α‐KG@Cy5.5@ALN‐Lipo exhibited significantly
   higher intracellular α‐KG levels compared to those treated with
   Cy5.5@ALN‐Lipo, indicating successful release of α‐KG within
   macrophages (Figure [166]6E).

   To validate the effects of bone‐targeted liposomes on fracture healing,
   we administered either Cy5.5@ALN‐Lipo or α‐KG@Cy5.5@ALN‐Lipo to T2DM
   mice via tail vein injection after fracture surgery (Figure [167]6F).
   Micro‐CT analysis revealed that α‐KG@Cy5.5@ALN‐Lipo enhanced callus
   mineralization in T2DM mice at 10 days post‐surgery (Figure [168]6G).
   Quantitative analysis showed that the α‐KG@Cy5.5@ALN‐Lipo group had
   increased BV/TV, Tb.Th and Tb.N, along with a decrease in Tb.Sp
   compared to the Cy5.5@ALN‐Lipo group (Figure [169]6H). Furthermore, the
   callus index for femurs from α‐KG@Cy5.5@ALN‐Lipo‐treated T2DM mice was
   significantly greater than that of the Cy5.5@ALN‐Lipo group
   (Figure [170]6I). Safranin O staining further confirmed that
   α‐KG@Cy5.5@ALN‐Lipo increased callus size and cartilage content in T2DM
   mice compared to the Cy5.5@ALN‐Lipo group at 10 days post‐surgery
   (Figure [171]6J). Masson staining also revealed that
   α‐KG@Cy5.5@ALN‐Lipo enhanced mineralized bone formation at 28 days
   post‐surgery (Figure [172]6K). Moreover, compared to the Cy5.5@ALN‐Lipo
   group, the number of OCN^+ osteoblasts increased while that of TRAP^+
   osteoclasts decreased in the callus of α‐KG@Cy5.5@ALN‐Lipo treated T2DM
   mice (Figure [173]S8A–D, Supporting Information). In summary, our
   results suggest that bone‐targeted delivery of α‐KG effectively
   improves fracture healing in T2DM mice, primarily by ameliorating
   glutaminolysis and macrophage function through α‐KG supplementation.

3. Conclusion

   This study elucidates critical insights into the metabolic dysfunctions
   underlying impaired fracture healing in T2DM. Our research identifies
   the suppression of GLS expression and the consequent reduction in
   glutamine‐derived α‐KG in macrophages as pivotal contributors to
   delayed fracture healing in T2DM. These findings establish a novel
   mechanistic link between the inhibition of glutaminolysis and
   osteogenic deficits, advancing our understanding of the metabolic
   underpinnings of this condition. Moreover, we have demonstrated that
   glutaminolysis significantly influences the inflammatory phenotype of
   macrophages and modulates the epigenetic landscape that governs BMP2
   production. This mechanistic insight provides a fresh perspective on
   the role of macrophage metabolism in bone repair, highlighting its
   potential as a therapeutic target. In addition, our investigation into
   targeted α‐KG delivery to macrophages in bone offers a promising
   therapeutic strategy. By enhancing the specificity and efficacy of α‐KG
   treatment while mitigating the adverse effects associated with systemic
   delivery, this approach underscores the necessity of tailoring
   therapeutic interventions to the metabolic context specific to T2DM.
   Collectively, these findings not only deepen our understanding of the
   metabolic basis for impaired fracture healing in T2DM but also pave the
   way for the development of targeted interventions aimed at improving
   clinical outcomes.

   Macrophages, as key responders to environmental changes, undergo
   metabolic reprogramming in response to stimuli.^[ [174]^42 , [175]^43
   ^] In T2DM, characterized by dysregulated glucose and lipid metabolism,
   chronic inflammation persists due to excessive immune activation.^[
   [176]^44 , [177]^45 ^] However, studies on amino acid metabolism,
   particularly glutamine metabolism, remain limited. Glutamine, the most
   abundant non‐essential amino acid in plasma, plays critical roles in
   tumor progression,^[ [178]^46 ^] stem cell differentiation,^[ [179]^47
   ^] and aging‐related disorders.^[ [180]^48 ^] Clinical studies show an
   inverse relationship between the serum glutamine‐to‐glutamate ratio and
   diabetes risk,^[ [181]^26 ^] and glutamine supplementation has been
   shown to reduce inflammation in white adipose tissue.^[ [182]^27 ^] In
   this study, we demonstrated that GLS expression and activity, and
   glutaminolysis are inhibited in T2DM using monocytes from patients and
   healthy controls, PMs from T2DM and NM mice, and BMDMs exposed to high
   glucose and palmitic acid.

   Macrophages are critical to all stages of fracture healing, interacting
   with immune cells, endothelial cells, and BMSCs through the secretion
   of cytokines. Among these, BMP2 is essential for promoting osteogenesis
   of BMSCs. In T2DM, however, chronic inflammation disrupts macrophage
   function, impairing the healing process.^[ [183]^49 ^] Although
   modulating macrophage phenotypes can improve bone repair,^[ [184]^50 ,
   [185]^51 ^] the connection between macrophage glutaminolysis and
   fracture healing remains unclear. In our femur fracture model, we
   observed significantly compromised callus formation and mechanical
   strength in Gls^fl/fl;Lyz2‐Cre mice compared to Gls^fl/fl controls. In
   vitro, GLS‐deficient macrophages also inhibited osteogenic
   differentiation and mineralization of BMSCs.

   RNA sequencing revealed that GLS deletion in macrophages exacerbated
   inflammation. This is consistent with previous findings that glutamine
   deficiency drives macrophages toward the pro‐inflammatory M1 phenotype,
   while α‐KG supplementation can reverse this effect.^[ [186]^22 ^]
   Mechanistically, oxidative stress, modulated by glutamine and α‐KG,^[
   [187]^22 , [188]^52 ^] plays a role in inflammation. Glutathione (GSH),
   the primary intracellular antioxidant, relies on glutamate, a
   glutaminolysis product.^[ [189]^53 ^] GLS deficiency likely reduces GSH
   production and increases reactive oxygen species, leading to imbalanced
   macrophage phenotypes and impaired function. Moreover, Liu, P.S.,
   et al. reported that glutamine‐derived α‐KG is vital for the activation
   of M2 macrophage via metabolic reprogramming (increased fatty acid
   oxidation) and Jmjd3‐dependent epigenetic modification of M2‐related
   genes.^[ [190]^22 ^] In our results, we observed that GLS deficiency
   led to decreased levels of α‐KG, which is likely the underlying reason
   for the altered macrophage phenotype observed in GLS knockout mice,
   contributing to the shift towards a pro‐inflammatory state.

   Secretome analysis showed a marked reduction in BMP2 secretion
   following GLS ablation. BMP2, a member of the TGF‐β superfamily,
   regulates bone and cartilage formation through SMAD transcription
   factor activation.^[ [191]^54 ^] Macrophage polarization and metabolism
   influence BMP2 expression, and its secretion from M2 macrophages is
   crucial for bone repair.^[ [192]^36 , [193]^55 ^] Reduced BMP2
   secretion in GLS‐deficient macrophages appears to be a key factor in
   delayed fracture healing in Gls^fl/fl;Lyz2‐Cre mice. Epigenetic
   modifications control gene expression without changing the DNA sequence
   and are regulated by metabolites such as lactate, acetyl‐CoA, and
   α‐KG.^[ [194]^56 ^] α‐KG serves as a cofactor for TETs, a kind of
   dioxygenase, which influences gene expression by modulating DNA
   methylation.^[ [195]^23 , [196]^57 ^] In our study, glutaminolysis
   impairment due to T2DM or GLS deficiency reduced α‐KG levels in
   macrophages, inhibiting TET activity. This resulted in decreased levels
   of hydroxymethylated cytosine (5hmC) and increased methylated cytosine
   (5mC) in GLS‐deficient macrophages. To assess global DNA methylation,
   we performed genome‐wide analysis using RRBS. GLS deficiency in BMDMs
   increased methylation in the promoter region of Bmp2, a change reversed
   by α‐KG supplementation. These results suggest that glutaminolysis
   regulates macrophage cytokine secretion via DNA methylation, affecting
   bone development and regeneration.

   Previous research has shown that α‐KG supplementation can reverse
   cellular dysfunction and pathological processes caused by glutamine
   deprivation.^[ [197]^22 , [198]^47 ^] Based on this, we administered
   α‐KG to T2DM mice to examine fracture healing. Despite reported
   benefits of α‐KG in T2DM, such as improving insulin sensitivity,
   enhancing cardiac mesenchymal cell survival, and promoting
   pericyte‐endothelial interactions,^[ [199]^58 , [200]^59 ^] our
   findings showed that oral α‐KG significantly impaired fracture healing
   in T2DM mice. Furthermore, α‐KG supplementation caused weight loss and
   increased mortality, likely due to enhanced gluconeogenesis, which
   raised glucose levels and worsened the metabolic state of the mice.^[
   [201]^38 ^] To address these negative effects, we developed an
   alendronate‐decorated liposome for targeted delivery of α‐KG to
   macrophages in bone.^[ [202]^40 ^] Both in vivo and in vitro
   experiments confirmed its effective bone‐targeting properties and
   macrophage uptake. Micro‐CT and Safranin O staining showed that
   bone‐specific α‐KG delivery significantly improved fracture healing in
   T2DM mice.

   In summary, this study elucidates the critical role of macrophage
   glutaminolysis in the fracture healing process, particularly within the
   context of T2DM. By targeting this metabolic pathway, we have
   demonstrated a promising therapeutic avenue that could significantly
   improve bone repair in diabetic patients, highlighting the potential of
   metabolic modulation as a therapeutic strategy in regenerative
   medicine.

4. Experimental Section

Animals

   All animal procedures were approved by the Animal Care and Use
   Committees at the Fourth Military Medical University (Xi'an, China)
   (Ethics Committee approval number: 20241481).

   T2DM was induced in 4‐week‐old mice through an 8‐week high‐fat diet
   (HFD, 60%) (Research Diets, USA), followed by an intraperitoneal
   injection (i.p.) of STZ (40 mg kg^−1 body weight) (Sigma Aldrich,
   Germany).^[ [203]^49 ^] The successful establishment of the T2DM model
   was confirmed by measuring fasting blood glucose (FBG) levels exceeding
   11.1 mmol L^−1, along with assessments of food intake, water
   consumption, body weight, intraperitoneal glucose tolerance tests
   (IPGTT) using glucose (2 g kg^−1 body weight) after overnight fasting
   and insulin tolerance tests (ITT) using insulin (1 U kg^−1 body weight,
   Novolin R) after 6 h of fasting.

   Gls^fl/fl and Lyz2‐Cre mice, both on a C57BL/6 background, were
   acquired from the Shanghai Model Organisms Center, Inc. (Shanghai,
   China). The Gls^fl/fl mice were crossed with Lyz2‐Cre mice to achieve
   macrophage‐specific GLS deletion, resulting in the Gls^fl/fl ;Lyz2‐Cre
   genotype. Male mice aged eight to sixteen weeks were used in the
   studies, with littermates serving as controls.

   All mice were housed under specific pathogen‐free conditions at 23 ±
   2 °C and maintained on a 12‐h light‐dark cycle. Ad libitum access to
   food and water was provided for all animals.

Surgical Procedures

   To establish the femur fracture model, mice were anesthetized with
   pentobarbital sodium (0.5%) via i.p. The skin was sterilized using
   ethanol (75%), and a cross‐shaped incision was made at the knee to
   expose the patella, which was dislocated laterally to access the
   femoral condyles. A stainless‐steel intramedullary pin was inserted to
   stabilize the femur. Subsequently, a blunt dissection of the muscles
   was performed, and the femur was fractured at the midpoint with a
   dentist's microdrill. After the fracture, the muscles were
   repositioned, and the skin was sutured using a 6/0 nylon suture.
   Following the surgery, liposomes were injected intravenously into T2DM
   mice every three days (1.0 mg kg^−1).

Preparation of Bone‐Targeted Liposomes

   Bone‐targeted liposomes were prepared via the film‐rehydration
   method.^[ [204]^60 ^] DSPE‐PEG2000‐ALN was procured from Qiyue Biology
   (Xi'an, China). A mixture containing DM‐α‐KG (Sigma, USA), Cy5.5,
   DSPE‐PEG‐ALN, and egg yolk lecithin was co‐dissolved in chloroform (3
   mL) with a ratio of 5:1:5:50. This mixture was dried using a rotary
   evaporator under negative pressure at 50 °C to produce a thin film. The
   film was then hydrated with pre‐warmed (50 °C) double‐distilled water
   (3 mL) under vacuum and sonicated for 1 min (180 W, 5 s on, 5 s off) at
   room temperature. Liposomes were extruded through a 0.22 µm
   polycarbonate membrane (Millex, Ireland). Unloaded DM‐α‐KG and Cy5.5
   were removed via dialysis using a polycarbonate membrane with a 30 nm
   pore size. The morphology of the liposomes was analyzed using TEM (FEI
   Talos L120C, USA). Furthermore, the liposome size, PDI, and zeta
   potential were assessed by dynamic light scattering (DLS) using a
   Zetasizer (Malvern Nano‐ZS, UK).

Microcomputed Tomography

   Fractured femurs from mice were collected, and the adjacent muscles
   were meticulously removed. Following fixation in 4% paraformaldehyde
   (PFA), the femurs were scanned with a SkyScan 1276 scanner (Bruker,
   Kartuizersweg, Belgium) at a spatial resolution of 8 µm, utilizing an
   X‐ray energy of 55 kVp and 250 µA. Three‐dimensional reconstruction of
   the bone was performed using CTVol software, and the resulting data
   were analyzed using CTAn.

Histological Staining

   Ten days after fracture, mice were sacrificed, and femurs were
   collected. The femurs were fixed in 4% PFA for 48 h and then
   decalcified in 15% EDTA‐Na[2] (Sigma Aldrich) for 14 days. After
   decalcification, the femurs were embedded in paraffin and sectioned to
   a thickness of 5 µm. For Safranin O staining, Masson staining and TRAP
   staining, the paraffin sections were dewaxed and stained with Modified
   Safranin O and Fast Green Stain Kit (#G1371, Solarbio), Masson's
   Trichrome Stain Kit (#G1340, Solarbio) and Tartrate‐Resistant Acid
   Phosphatase(TRAP) Stain Kit (#G1492, Solarbio), respectively, according
   to the manufacturer's instructions. For immunohistochemical staining of
   osteoblasts, the sections were blocked with 1% BSA at room temperature
   for 1 h and then incubated with osteocalcin (#20277‐1‐AP, Proteintech)
   primary antibody at 4 °C overnight. Then the slides were incubated with
   an HRP‐Goat anti‐rabbit recombinant secondary antibody (RGAR011,
   Proteintech) at room temperature for 1 h. The antibody‐antigen complex
   was visualized by DAB Horseradish Peroxidase Color Development Kit
   (P0203, Beyotime). The sections were dehydrated and mounted with
   neutral resin for microscopic observation (Olympus, Tokyo, Japan). Then
   the statistics of images were performed using Image‐J.

Biomechanical Assessments

   To assess the biomechanical properties of fractured femurs, a
   three‐point bending test was performed using a mechanical testing
   machine (WDW‐100; Shanghai Bairoe Instruments Co., Ltd., Shanghai,
   China). The femurs were placed on two supporting rods and loaded until
   fracture at a displacement rate of 5 mm min^−1. Bone strength,
   toughness, and stiffness were determined from the load‐deformation
   curves.

Human Studies

   Human studies were reviewed and approved by the Medical Ethics
   Committee of the First Affiliated Hospital of the Fourth Military
   Medical University (Ethics Committee approval number: KY20213177‐1).
   Participants were informed about the research project and sample
   collection procedures, and informed consent was obtained.

   Patients diagnosed with T2DM were defined as those with fasting blood
   glucose levels of ≥ 7.0 mmol L^−1 or HbA1c levels > 6.5%. Healthy
   controls were defined by fasting blood glucose levels of ≤ 6.0 mmol
   L^−1.^[ [205]^61 ^] Exclusion criteria comprised a history of heavy
   alcohol use, pregnancy or lactation in female participants, diagnoses
   of autoimmune diseases, infectious diseases, hematological disorders,
   or cancers within the past 6 months, severe trauma or surgery within
   the past three months, and diagnoses of hepatitis, liver cirrhosis, or
   liver cancer. Individual characteristics are summarized in Table
   [206]S1 (Supporting Information).

   Blood samples ranging from six to eight milliliters were collected from
   eligible patients. PBMCs were isolated using Ficoll‐Paque PLUS
   (17144002, Cytiva, China) via density gradient centrifugation.
   Following centrifugation, serum was collected from the supernatant for
   further analysis, and PBMCs were purified using PBS. Monocytes were
   positively selected from whole PBMCs using antiCD14^+ microbeads
   (Miltenyi Biotec) for the following experiments.^[ [207]^62 ^]

Cell Culture

   All cells were cultured in a 37 °C incubator with 5% CO[2], with
   culture media supplemented with penicillin (1000 U mL^−1) and
   streptomycin (100 mg mL^−1).

   Primary BMDMs were isolated from 6‐ to 8‐week‐old mice as previously
   described.^[ [208]^63 ^] Bone marrow was flushed from the femur and
   tibias using PBS. After lysing red blood cells, the remaining cells
   were cultured in BMDM medium, consisting of DMEM (HyClone, USA)
   supplemented with 10% FBS (InCellGene, Germany), 15% L929 conditioned
   medium, and 1% penicillin‐streptomycin (Invitrogen, USA). After a 7‐day
   incubation, BMDMs were digested with Accutase (#PC‐90011, PlantChemMed)
   for subsequent experiments. To induce a pro‐inflammatory phenotype (M1
   polarization), BMDMs were treated with LPS (100 ng mL^−1) (Peprotech,
   USA) for 6 h. For the induction of an anti‐inflammatory phenotype (M2
   polarization), BMDMs were treated with IL‐4 (20 ng mL^−1) (Peprotech)
   for 24 h.

   Primary BMSCs were isolated from 4‐week‐old C57BL/6 mice using a
   procedure as previously described.^[ [209]^64 ^] The medium was
   adjusted to BMSC medium, consisting of α‐MEM (HyClone) supplemented
   with 15% FBS and 1% penicillin‐streptomycin. To induce osteoblastic
   differentiation of BMSCs for ALP activity assays or Alizarin Red S
   staining, the cells were allowed to reach 90% confluency. The BMSC
   medium was then replaced with osteogenic induction medium, which
   included α‐MEM supplemented with 10% FBS, β‐glycerol phosphate (10 mm)
   (Sigma Aldrich), ascorbic acid (50 µg mL^−1) (Sigma Aldrich), and
   dexamethasone (10^−8 m) (Sigma Aldrich) for the specified duration.

qRT‐PCR

   The total RNA of cells was isolated using the Trizol reagent (Sigma)
   following the manufacturer's instruction, and then reverse transcribed
   into cDNA using the Hifair II 1st Strand cDNA Synthesis SuperMix
   (#11120ES60, Yeasen, Shanghai, China). Next, qRT‐PCR was performed
   using Hieff qPCR SYBR Green Master Mix (#11201ES08, Yeasen) on a BioRad
   Real‐time PCR System (CFX96; BioRad, Hercules, CA, USA). The primer
   sequences are listed in Table [210]S2 (Supporting Information). All
   data were normalized to 36B4 (BMDM) or β‐actin (BMSC) expression.
   Relative mRNA levels of dividual genes were calculated by 2^−△△CT
   method.

Flow Cytometry

   BMDMs were digested with Accutase (#PC‐90011, PlantChemMed) and rinsed
   with PBS. Then the cell suspensions were washed twice with PBS, and
   stained with PE‐F4/80 (#PE‐98236, Proteintech) and FITC‐CD86
   (#FITC‐65068, Proteintech) or FITC‐F4/80 (#123107, BioLegend) and
   APC‐CD206 (#141707, BioLegend). Cells were then incubated for 15 min at
   room temperature in the dark, and analyzed by flow cytometry
   (LSRFortessa; BD, Franklin Lakes, NJ, USA). The results were analyzed
   by FlowJo 7.6.5 software (Tree Star Inc., Ashland, OR, USA).

Western Blot

   Western blot was carried out as previously described.^[ [211]^51 ^]
   Cells were lysed and protein concentration were quantified through a
   BCA protein assay kit (#23225, Thermo Fisher). Further, equal amounts
   of proteins were separated by SDS‐PAGE gel electrophoresis and
   transferred onto a polyvinylidene fluoride (PVDF) membrane. After
   blocked with 5% skimmed milk at room temperature for 1 h, the membrane
   was incubated with primary antibodies against GLS (#12855‐1‐AP,
   Proteintech), Osteopontin (#25715‐1‐AP, Proteintech), Osteocalcin
   (#20277‐1‐AP, Proteintech), RUNX2 (#20700‐1‐AP, Proteintech), BMP2
   (#66383‐1‐Ig, Proteintech), and β‐actin (AP0060, Bioworld Technology)
   as loading control at 4 °C overnight. Then the blots were incubated
   with an HRP‐conjugated secondary antibody (1:5000 dilution, Santa Cruz
   Biotech) for 1 h at room temperature. Protein levels were
   semi‐quantified by scanning densitometry and analyzed with Image J
   software.

ELISA

   The supernatants of BMDMs were collected, and centrifuged at 12 000 rpm
   for 30 min to remove the debris of cells. The concentrations of
   cytokines were measured with ELISA kit: CCL2 (#SEA087Mu, Cloud‐clone),
   TNF‐α (#SEA133Mu, Cloud‐clone), and BMP2 (#SEA013Mu, Cloud‐clone)
   following the manufacturers’ instructions.

Glutamine and α‐KG Measurements

   Plasma and cellular glutamine or α‐KG were detected using a Glutamine
   (Gln) Content Assay Kit (#BC5300, Solarbio) or Amplite Colorimetric
   α‐Ketoglutarate Quantitation Kit (#10085, AAT Bioquest) separately
   according to the manufacturer's instructions. Briefly, BMDMs were lysed
   and the extracts or plasma were added to the working solution and
   incubated at 37 °C for 1 h in the dark. The absorbance was measured at
   a wavelength of 450 nm (glutamine) or 570 nm (α‐KG). Glutamine and α‐KG
   in samples were calculated using the standard curve and equation.

Measurements of Enzymatic Activity

   Glutaminase activity was determined using a GLS Assay Kit (#BC1450,
   Solarbio) according to the manufacturer's instructions. In brief, BMDMs
   were lysed, and the extracts were added to the working solution and
   incubated at 37 °C for 1 h. The absorbance was measured at a wavelength
   of 630 nm. A unit of GLS activity was defined as 1 mg protein
   catalyzing glutamine to generate 1 µmol ammonia per hour at 37 °C. The
   data were expressed as U mg^−1 protein.

   TET activity was determined using an EpiQuik Nuclear Extraction Kit I
   (#OP‐0002, EpigenTek) and Epigenase 5mC Hydroxylase TET
   Activity/Inhibition Assay Kit (#P‐3086, EpigenTek) according to the
   manufacturer's instructions.^[ [212]^23 ^] Briefly, to prepare nuclear
   extracts, BMDMs were washed with PBS twice and scraped cells into a 15
   mL conical tube. Add the extraction Buffer and incubate on ice for 10
   minutes, then centrifuge the preparation for 1 min at 12 000 rpm, and
   carefully remove the cytoplasmic extract from the nuclear pellet. To
   detect TET activity, working solutions were prepared as instructed and
   described. Four microliters of nuclear extracts were added to the
   working solution and incubated at 37 °C for 90 min. Finally, the
   absorbance was read on a microplate reader within 2 to 10 min at 450 nm
   with an optional reference wavelength of 655 nm. TET activity was
   calculated using the formulas in the instructions.

5mC and 5hmC Detection

   DNA was isolated using FitAmp Blood and Cultured Cell DNA Extraction
   Kit (#P‐1018, EpigenTek). Levels of 5‐mC and 5‐hmC were determined by
   MethylFlash Global DNA Methylation (5‐mC) ELISA Easy Kit (Colorimetric)
   (#P‐1030, EpigenTek) and MethylFlash Global DNA Hydroxymethylation
   (5‐hmC) ELISA Easy Kit (Colorimetric) (#P‐1032, EpigenTek). All steps
   were performed according to the manufacturer's instructions. Briefly,
   100 ng DNA was used, and an absorbance at 405 nm was recorded using the
   microplate reader.

The α‐KG Release of Bone‐Targeted Liposome

   Release of α‐KG from liposomes was investigated by placing 1 mL of
   α‐KG@Cy5.5@ALN‐Lipo in dialysis bags (MWCO = 3 500 Da, Thermo, USA) as
   previously reported.^[ [213]^60 ^] The dialysis bags were incubated in
   sterile PBS (40 mL) incubated at 37 °C with agitation at 100 rpm. At
   the scheduled time points, 1 mL of PBS was taken out and replaced by
   the same volume of fresh PBS. The released medium was vortexed and
   analyzed by Amplite Colorimetric α‐Ketoglutarate Quantitation Kit
   (#10085, AAT Bioquest), the detailed method was mentioned above. The
   tests were conducted in triplicate, and the average data were
   calculated.

Biodistribution of Bone‐Targeted Liposome

   For biodistribution analysis, liposomes were i.v. injected into
   8‐week‐old C57BL6 mice (1.0 mg kg^−1 body weight);^[ [214]^65 ^] the
   kidneys, spleen, lungs, brain, heart, liver, and hind limbs were
   dissected 6 or 24 h post‐injection and were subjected to biophotonic
   imaging assay for evaluation of organ distribution of liposomes using
   the IVIS Spectrum In Vivo Imaging System (PerkinElmer).

Alkaline Phosphatase (ALP) Activity Assay And Alizarin Red Staining

   To analyze the osteogenic differentiation of BMSC, cells were induced
   for designated days with an osteogenic induction medium mixed with
   macrophage conditioned medium. ALP activity was detected using an
   Alkaline Phosphatase Assay Kit (#P0321S, Beyotime) according to the
   manufacturer's instructions. To perform staining, BMSCs were washed
   three times with PBS and fixed with 4% paraformaldehyde for 20 min at
   RT. ALP staining was performed with BCIP/NBT Alkaline Phosphatase
   Colorpment Kit (#C3206, Beyotime), and incubated in the working
   solution for 30 min at room temperature. Alizarin Red S staining was
   stained with Alizarin Red S solution (#C0138, Beyotime) for 15 min.
   Then the BMSCs were washed with deionized water and visualized using
   microscopy. The calcium nodules were resolved with 10% cetylpyridinium
   chloride and the absorbance was detected at 562 nm.

RNA Sequencing

   Total RNA was extracted from BMDMs using Trizol reagent (Sigma), and
   rRNA was subsequently removed. mRNA with a poly‐A tail was enriched
   from each sample using oligo (dT) magnetic beads and fragmented into
   smaller pieces through ultrasonic treatment. The integrity and purity
   of the RNA were assessed. The first strand of cDNA was synthesized with
   the M‐MuLV reverse transcriptase system, using the RNA fragments as
   templates and random oligonucleotides as primers. The second strand of
   cDNA was synthesized in a reaction containing dNTPs and DNA polymerase
   I. Double‐stranded cDNA was purified and modified using AMPure XP beads
   to isolate fragments of ≈200 bps. cDNA amplification was performed, and
   the products were purified again with AMPure XP beads to create the
   library. For functional enrichment analysis, GO term analysis was
   conducted, categorizing the data into three subgroups: biological
   process (BP), cellular component (CC), and molecular function (MF).
   Additionally, the KEGG database was used to identify significantly
   enriched KEGG pathways at a threshold of p < 0.05.

Secretomics Analysis

   BMDMs were seeded at a density of 5.5 × 10^6 cells in 10 cm culture
   dishes for 24 h, followed by a 48‐h incubation in serum‐free DMEM to
   collect the supernatant. Secretome profiling was conducted by Jingjie
   PTM BioLab Co. Ltd (Hangzhou, Zhejiang, China). Cellular debris was
   removed from the serum samples by centrifugation at 12 000 g for 10 min
   at 4 °C. The supernatant was subsequently transferred to a new
   centrifuge tube for concentration. For digestion, trypsin was added at
   a 1:50 trypsin‐to‐protein mass ratio for overnight digestion. The
   tryptic peptides were dissolved in solvent A (0.1% formic acid, 2%
   acetonitrile in water) and separated using a gradient of solvent B
   (0.1% formic acid in 90% acetonitrile) over 60 min. The separated
   peptides were analyzed with an Orbitrap Exploris 480 mass spectrometer
   (ThermoFisher Scientific), with the automatic gain control (AGC) target
   set at 3E6 and the maximum injection time set to Auto. The resulting
   MS/MS data were processed using the MaxQuant search engine (v.
   1.6.15.0) and analyzed against the UniProt Mus musculus database.
   Bioinformatics analysis was performed utilizing OmicStudio tools.

Statistical Analysis

   The number of independent experiments and the sample size (n) for each
   experiment were detailed in the figure legends. Statistical analysis
   was performed using GraphPad Prism v8.0 (GraphPad, La Jolla, CA, USA).
   All data were presented as mean ± SD. The methods of statistic analysis
   of significant differences between groups were described in figure
   legends. p ≤ 0.05 was considered statistically significant.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   J.W., J.C., and S.Z. contributed equally to this work as co‐first
   authors. J.W., Y.F., and J.L. conceptualized the study. J.W., J.C.,
   X.Y., X.W., W.C., W.D., T.W., and X.L. performed in vivo experiments
   including femur fracture surgery or pharmacological treatment. J.W.,
   J.C., X.W., and W.D. performed in vitro experiments. J.W., J.C., S.Z.,
   and H.C. performed RNA‐seq, secretomes, and RRBS data analysis. J.W.,
   J.C., S.Z., and T.W. analyzed the data. J.W., J.C., and S.Z. visualized
   the data. Y.F., W.L., and J.L. acquired funding to support the project.
   J.W., S.Z., and H.C. wrote the original draft of the manuscript. Y.F.,
   W.L., and J.L. reviewed and edited the manuscript.

   [Correction added on 02 May 2025, after first online publication:
   Author Contributions statement has been added.]

Supporting information

   Supporting Information
   [215]ADVS-12-2415667-s001.docx^ (2.2MB, docx)

   Supplemental Table 1
   [216]ADVS-12-2415667-s002.xlsx^ (59.9KB, xlsx)

Acknowledgements