IMP2 enhances M2 macrophage polarization via LKB1-AMPK-mediated mitochondrial dynamics and fatty acid β-oxidation to ameliorate diabetic osteoporosis

Abstract

A reduction in M2 macrophage polarization is a key pathogenic contributor to diabetic osteoporosis (DOP). In this study, we revealed that IMP2 affects mitochondrial dynamics and fatty acid β-oxidation (FAO) via the LKB1-AMPK pathway, influencing the M2 polarization pattern and thereby improving DOP. Leveraging Cre-loxp technology, we generated macrophage-specific IMP2 knockout mice. Under diabetic settings, IMP2 knockout further disrupted M2 macrophage polarization and intensified bone deterioration. Mechanistically, IMP2 enhances LKB1 mRNA stability through the N6-methyladenosine (m⁶A) pathway, activating the LKB1-AMPK signaling cascade. This pathway promotes mitochondrial fusion and boosts FAO, fueling M2 macrophage polarization. Enhanced M2 polarization, in turn bolsters the osteogenic differentiation potential of bone marrow mesenchymal stem cells (BMSCs). However, in the diabetic milieu, IMP2’s regulatory role in M2 macrophages is compromised, diminishing the osteogenic support provided by M2 macrophages to BMSCs and exacerbating bone loss.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-026-06093-5.

Introduction

DOP is a common skeletal complication of diabetes, marked by reduced bone mass, decreased bone microstructure, and a substantially heightened fracture risk [1, 2]. Over 35% of patients with type 2 diabetes mellitus (T2DM) suffer from bone loss, with roughly 20% meeting the diagnostic criteria for osteoporosis [1]. Despite its prevalence, the pathogenesis of DOP remains incompletely understood, and clinical treatment options are limited, underscoring the urgent need to identify novel regulatory targets and therapeutic strategies.

Macrophages, an important component of the immune system, play pivotal roles in bone metabolism and inflammation regulation [3, 4]. They can differentiate into functionally distinct subtypes through polarization: pro-inflammatory M1 macrophages and anti-inflammatory M2 macrophages [5, 6]. In the context of diabetes, the reduction in M2 macrophages weakens their anti-inflammatory and tissue-repair capabilities [7–9]. Our previous research has demonstrated that M2 macrophages secrete TGF-β, promoting the osteogenic differentiation of BMSCs [10]. Therefore, modulating macrophage polarization, particularly promoting M2 polarization, could offer a new avenue for improving DOP.

Epigenetics exerts crucial regulatory effects on M2 macrophages. Among epigenetic modifications, RNA modifications, especially m⁶A methylation, profoundly influence polarization process by dynamically regulating RNA fate [11–13]. As one of the most abundant chemical modifications on eukaryotic mRNA, m⁶A methylation is formed by methyltransferase complexes, which can be reversibly removed by demethylases, and recognized by “readers” proteins (such as the IGF2BP family). This process regulates the stability, splicing, transport, and translation efficiency of target mRNA, playing a central role in physiological and pathological processes like cell differentiation, metabolic adaptation and immune response [14]. Insulin-like growth factor 2 mRNA-binding protein 2 (IMP2/IGF2BP2), a key reader in the m⁶A process, is abnormally expressed in T2DM, obesity and cancer [15]. IMP2 is closely related to macrophage polarization, serving as a key regulator [15, 16]. Knocking out IMP2 reduces M2 macrophage polarization and exacerbates inflammatory bowel disease [17]; however, the mechanism by which IMP2 regulates M2 macrophages in DOP remains unclear.

Recent studies indicate that the macrophage polarization is closely linked to mitochondrial dynamics and metabolic status [6, 18]. Mitochondria, the core cellular organelles for energy metabolism, regulate cellular functional state through fusion-fission dynamics. Enhanced fusion supports the polarization of M2 macrophages, while excessive fission induces the M1 pro-inflammatory phenotype [19]. FAO, an important metabolic pathway, generates acetyl coenzyme A, replenishes the tricarboxylic acid cycle, and significantly elevates oxidative phosphorylation (OXPHOS) and ATP production, promoting M2 macrophage polarization [20]. In diabetes, high glucose levels induce mitochondrial dynamics imbalance, reducing fusion and causing functional damage [21, 22], as well as a decrease in FAO [23], thereby inhibiting M2 macrophage polarization. Although the impact of mitochondrial function and metabolic status on M2 macrophage polarization is confirmed, the common epigenetic pathways remain unclear.

The LKB1-AMPK signaling pathway is central to cellular energy sensing and metabolic regulation [24], playing crucial roles in maintaining mitochondrial function [25, 26] and lipid metabolism [27, 28]. LKB1 phosphorylates and activates AMPK, which maintains mitochondrial dynamic balance and enhances FAO [28, 29]. In the high glucose environment of diabetes, the LKB1-AMPK pathway activity is often inhibited [30], potentially contributing to mitochondrial dysfunction and metabolic disorders. Previous studies have shown that m⁶A modification is involved in regulating the expression of LKB1 [31, 32], whether IMP2 participates in metabolic and immune regulation by modulating the expression or activity of pathway-related molecules remains unknown.

Therefore, we hypothesize that in DOP, IMP2 may target the LKB1-AMPK signaling pathway to influence macrophage mitochondrial dynamic balance and FAO level, thereby promoting the polarization of M2 macrophage and ultimately improving bone metabolism imbalance. By combining in vitro and in vivo experiments, we aim to elucidate the molecular mechanism of IMP2 in DOP, providing a theoretical foundation for the development of targeted immunometabolic therapeutic strategies.

Materials and methods

Animals and tissue collection

Firstly, macrophage-specific IMP2 knockout mice were constructed using the Cre-loxp system. We crossed C57BL/6JGpt-Lyz2 em1Cin(iCre)/Gpt mice (Strain NO. T003822) with IMP2^flox/flox (IMP2^f/f) mice (Strain NO. T009552) to obtain IMP2^f/f,Cre transgenic mice (both from GemPharmatech, Nanjing, China). Littermate age-matched IMP2^f/f mice were used as controls. We selected male mice for breeding and began experiments at 5 weeks of age. We established a T2DM model and designed a total of five experimental groups (n = 5 per group [33]) to comprehensively evaluate the role of IMP2 in diabetic bone metabolism: (1) IMP2^f/f group (littermate control); (2) IMP2^f/f,Cre group; (3) IMP2^f/f+T2DM group; (4) IMP2^f/f,Cre + T2DM group; (5) IMP2^f/f,Cre + T2DM + resveratrol group. The specific methods are as follows. After 4 weeks of HFD treatment in the experimental group, (groups 3, 4, 5) streptozotocin (40 mg/kg) was intraperitoneally injected every other day for 5 injections. The fasting blood glucose was measured 7 days after the last injection, and blood glucose levels higher than 11.1 mmol/L for 3 consecutive days or random blood glucose higher than 16.7 mmol/L were considered successful establishment of the T2DM model [34]. After 8 weeks [35, 36], all mice were anesthetized with 1% pentobarbital and perfused with 4% paraformaldehyde. Femurs were harvested for CT scanning, decalcified in EDTA solution, embedded in paraffin, and sectioned into 5 μm thick slices for histological analysis.

In vivo insulin stimulation and tissue collection

Tissue collection and processing

At 15 min[37] after insulin injection, mice were euthanized, and liver and skeletal muscle tissues were rapidly harvested. Tissues were snap-frozen in liquid nitrogen and stored at −80 °C for subsequent protein extraction. For histological analysis, tissues were fixed in 4% paraformaldehyde, dehydrated through graded ethanol series, embedded in paraffin, and sectioned at 5 μm thickness. Sections were used for immunofluorescence and immunohistochemical staining to evaluate protein expression and localization.

Protein extraction and analysis total

Total protein was extracted from frozen tissues using RIPA lysis buffer supplemented with protease and phosphatase inhibitors. Protein concentration was determined by BCA assay, and equal amounts of protein were subjected to Western blot analysis to assess the expression of insulin signaling-related proteins.

Mouse insulin tolerance test (ITT)

Model and control mice were fasted for 6 h with free access to water. Basal blood glucose levels were measured from the tail vein at 0 min. Short-acting insulin was administered intraperitoneally at a dose of 0.75 U/kg body weight. Blood glucose levels were measured at 0, 15, 30, 60, 90, and 120 min after injection. The ratio of blood glucose at each time point to the basal level (time = 0 min) was calculated to assess insulin sensitivity.

Micro-CT

After the bone specimens were fixed, they were analyzed using the Micro-CT analysis system (Cruiser, Recon, Avatar).

Hematoxylin and eosin (HE)

After the paraffin sections were treated with dewaxing and hydration, they were rinsed with distilled water twice, stained with hematoxylin for 15 min, rinsed with distilled water, stained with eosin for 7 min, rinsed again with distilled water, and dehydrated, transparentized and sealed with neutral gum. The sections were then observed under an optical microscope (Olympus BX-53, Tokyo, Japan).

Immunohistochemistry (IHC) staining

After dewaxing and hydration, 0.3% H₂O₂ was used to block endogenous peroxidase for 30 min, and 1% BSA-PBS was used to block non-specific binding sites for 30 min. After overnight incubation at 4 °C with the primary antibody, the sample was washed with PBS, and incubated with the secondary antibody at room temperature for 1 h. Color development was performed using DAB system (Sigma-Aldrich), the sample was re-stained with 1% methyl green, dehydrated and transparentized conventionally, and sealed with neutral resins; finally, it was observed under the Olympus BX-53 microscope, and the optical density value (OD) was determined using Image-Pro Plus 6.2 software for quantitative analysis.

Immunofluorescence (IF) staining

The dewaxing and dehydration steps of the tissue were the same as those in IHC. The cell samples were fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, and then blocked with 5% BSA-PBS for 30 min. The primary antibodies were incubated at 4 °C overnight, and then the secondary fluorescent antibody was incubated at room temperature for 1 h. The nuclei were stained with DAPI for 5 min. The samples were captured using a confocal microscope (FV3000, OLYMPUS).

Cell culture

RAW264.7 cells (from Shanghai Cell Center) were cultured in DMEM medium (containing 10% FBS and 1% penicillin–streptomycin) at 37 °C with 5% CO₂. They were seeded at a density of 1 × 10⁵/mL in 6-well plates. After 24 h, 20 ng/mL IL-4 (MCE HY-P70653) and 20 ng/mL IL-13 (MCE HY-P70460) were added to induce for 48 h. The supernatant was collected as the conditioned medium. BMSCs were isolated from the tibia and femur of 8-week-old C57BL/6 male mice, and seeded in α-MEM medium containing 20% FBS for 5–7 days to obtain BMSCs. The BMSCs were seeded at a density of 1 × 10⁴/mL, and then the above-conditioned medium was used for culturing after 24 h. BMDMs were isolated from the same mice. After bone dissection, muscle removal and end cutting, the marrow cavity was flushed with RPMI-1640 + 10%FBS until the bones turned white. Suspension was filtered (70 μm), centrifuged (1000 rpm, 5 min), pellet resuspended in RPMI-1640 + 10%FBS + 20ng/mL M-CSF and seeded. Non-adherent cells were removed at 72 h; medium with M-CSF was changed every 2–3 days, cultured for 7–10 days to mature.

CCK8

The cells were seeded at a density of 1 × 10⁴ in 96-well plates. Each well was added with the corresponding cell culture medium and a blank control well (only with culture medium) was set up. The plates were placed in a 37 °C, 5% CO₂ incubator for cultivation until the end of the experimental treatment. Then, 10 µL of CCK-8 reagent was added to each well, and the mixture was gently shaken and mixed. The cultivation was continued for 1–4 h. Finally, the absorbance (OD value) of each well was measured at a wavelength of 450 nm using an enzyme-linked spectrophotometer. The difference in OD values between the experimental group and the control group was calculated to reflect the cell proliferation activity.

Cell transfection

The IMP2 plasmid (Igf2bp2 (NM_183029) pLV3-CMV-MCS-3xFLAGCopGFP-Puro, 65.6KD) was designed and produced by Research Cloud Biology and the small interfering RNA (siRNA) was designed and synthesized by General Biol, China. When the cell confluence reached 30%−50%, transfection was performed using Lipofectamine 8000 transfection reagent (Beyotime C0533) according to the manufacturer’s instructions. The interference efficiency was evaluated by polymerase quantitative reverse transcription polymerase chain reaction (qRT-PCR) or Western blot (WB) 48 h after transfection. The sequences of all siRNAs are listed in the supplementary Table 1.​

Table 1.

Sequence of the SiRNA

Primer Sequence (5’ to 3’)
Gene	Forward	Reverse
siLKB1−1881	5’-AGGAAGACUUGUUUGACAUTT-3’	5’-AUGUCAAACAAGUCUUCCUTT-3’
siLKB1−1197	5’-UGUAUAUGGUGAUGGAGUATT-3’	5’-UACUCCAUCACCAUAUACATT-3’
siLKB1−1442	5’-GGAUGACACCUGCCGGACAATT-3’	5’-UUGUCCGGCAGGUGUCAUCTT-3’
siIMP2−992	5’-GGGUAGACAUCCACAGAAATT-3’	5’-UUUCUGUGGAUGUCUACCCTT-3
siIMP2−1196	5’-GCAGAAACCUGAAGAAAAUTT-3’	5’-AUUUUCUUCAGGUUUCUGCTT-3’
siIMP2−1782	5’-GGAGCAGUUACUUACAGAATT-3’	5’-UUCUGUAAGUAACUGCUCCTT-3

Flow cytometry

The polarization state of macrophages and the level of mitochondrial ROS were detected by flow cytometry (BD Accuri C6 Plus). The specific steps were as follows: a 0.1% BSA/PBS cell suspension was prepared, centrifuged at 300×g for 5 min; incubated with antibodies in the dark for 30 min; washed twice with PBS; and the fluorescence intensities were detected.

Mitochondrial status detection

The mitochondria were labeled using Mito-Tracker Red (C1035, Beyotime) and JC-1 mitochondrial membrane potential detection kits (C2006, Beyotime). After incubation in the dark at 37 °C for 30 min, the morphological characteristics of the mitochondrial network and the changes in mitochondrial membrane potential were observed using an Olympus FV3000 confocal microscope, and the morphology of the mitochondria was examined using TEM. The intracellular ATP content was determined using an enhanced ATP detection kit (S002, Beyotime).

Etomoxir treatment

To investigate the role of IMP2 in regulating macrophage polarization via FAO, the FAO inhibitor Etomoxir was used for intervention. Specifically, under the M2 polarization induction model, the following two groups were established: the high glucose baseline stimulation group (HG + IL-4 + IL-13 + IMP2 overexpression) and the inhibitor intervention group (HG + IL-4 + IL-13 + IMP2 overexpression + Etomoxir), to verify whether the effect of IMP2 overexpression depends on the fatty acid β-oxidation pathway. Etomoxir (MCE, HY-50202) was used at a concentration of 100 µM [38]. Cells were pretreated for 24 h [39] with complete medium containing 100 µM Etomoxir before stimulation with IL-4/IL-13 or induction of IMP2 overexpression. Subsequently, the medium was replaced with fresh medium containing the same concentration of Etomoxir and the corresponding stimulating factors, and the cells were cultured until the predetermined experimental endpoint.

Oxygen consumption rate

The OCR of RAW264.7 cells was analyzed using the XFe96 extracellular flux analyzer (Seahorse Bioscience) according to the manufacturer’s protocol (ALS22012, Alicelligent). The data were processed and analyzed using Seahorse Wave Pro software. RAW264.7 cells were seeded at a density of 1 × 10⁴ cells per well in an XF96 microplate, with a subset of cells pretreated with Etomoxir for 24 h. Subsequently, the medium was replaced with fresh medium containing IL-4/IL-13, IMP2 overexpression, or LKB1 knockdown, and the cells were further cultured for 48 h. After treatment, the cells were equilibrated for 1 h at 37 °C under CO₂-free conditions in XF basal medium supplemented with 2 mM glutamine. Oxygen consumption rate (OCR) was measured in real time using a mitochondrial stress test kit with sequential injections of 1 µM oligomycin, 1 µM FCCP, and 0.5 µM rotenone/antimycin A, on a Seahorse XF96 analyzer.

ALP and ARS staining

To determine the osteogenic activity, bone marrow mesenchymal stem cells (BMSCs) were cultured in conditioned medium to induce osteogenesis, with the culture periods being 7 days and 21 days respectively. After fixing the cells with 4% PFA, the cells were stained with ALP staining solution (Solbo Biological, G1480) and 1% safranin O solution (Solbo Biological, 20180820), according to the manufacturer’s instructions. ALP and ARS staining images were captured using an optical microscope (IX73, Olympus, Japan), and the result was analyzed using Image-Pro Plus 6.2 software.

qRT-PCR

RNA isolation from cells, qRT-PCR was performed using standard protocols [40]. The complete list of primers used in this study was provided in supplement file Table 2.

Table 2.

Sequence of the primer used for qRT-PCR

Primer Sequence (5’ to 3’)
Gene	Forword	Reverse
IMP2(igf2bp2)	5’-GTCCTACTCAAGTCCGGCTAC-3’	5’-GCCGATGATCTCTCTCAAGT GAT-3’
CD163	5’-ACTTCAGAATCACATCATGGCACA-3’	5’-TCGTCGCTTCAGAGTCCACAG-3’
CD206	5’-AGAGCTGGCGAGCATCAAGAG-3’	5’-TTCCATAGGTCAGTCCCA ACCAA-3’
ARG-1	5’-AGACCACAGTCTGGCAGTTG-3’	5’-AGACCACAGTCTGGCAGTTG-3’
CD86	5’-GACACCCACGGGATCAATTA-3’	5’-GCCTCCTCTATTTCAGGTTCAC-3
iNOS	5’-CAAGCTGAACTTGAGCGAGGA-3’	5’-TTTACTCAGTGCCAGAAGCTGGA-3’
IL-6	5’CTGCAAGAGACTTCCATCCAG-3’	AGTGGTATAGACAGGTCTGTTGG-3’
IL-10	5’CCAAGCCTTATCGGAAATGA-3’	5’-TTTTCACAGGGGAGAAATCG-3’
TGF-β	5’-TACGGCAGTGGCTGAACCAA-3’	5’-CGGTTCATGTCATGGATGGTG-3’
MFN1	5’-GTTGTTGGGGGCGTGATTTG-3’	5’-TGCAGCTTCTCGGTTGCATA-3’
MFN2	5’-AGAGGCAGTTTGAGGAGTGC-3’	5’-ATGATGAGACGAACGGCCTC-3’
Opa1	5’-GGAGAGTATCAAGCGGCACA-3’	5’-AACACACTGCTCTTGGGTCC-3’
Cpt1A	5’-ATGGTCAAGGTCTTCTCGGGT-3’	5’-TTGACAGCAAAATCCTGGGC-3’
CPT2	5’-CAGCACAGCATCGTACCCA-3’	5’-TCCCAATGCCGTTCTCAAAAT-3’
LKB1(stk11)	5’-AAGAGGTTCTCCATCCGACAG-3’	5’-GGTCCTTAGTGTCTGGGCTTG-3’
AMPK	5’-AAGTCAAAGCCGACCCAATG-3’	5’-AGTAGCAGTCCCTGATTTGGC-3’
β-Actin	5’-GGCTGTATTCCCCTCCATCG	5’-CCAGTTGGTAACAATGCCATGT-3’

WB

The protein extraction and gel imaging (General Electric Company) of cells and tissues were performed according to the standard procedure [40]. Image Pro Plus 6·2 software (Media Cybernetics) was used to analyze and quantify grayscale values normalized to β-Actin levels. The samples were derived from the same experiment or parallel experiments, and gels/blots were processed in parallel.

Actinomycin D assay

After RAW264.7 cells were treated with 10 µg/mL of actinomycin D (HY-17559, MCE) for a specific period of time, total RNA was collected and extracted at the designated time points of 0, 3, and 6 h. The mRNA half-life was determined using qRT-PCR analysis.

m⁶A modification site prediction

From the NCBI Gene database, the mouse LKB1 (stk11) gene (Gene ID: 20869) was retrieved, and its transcript sequence was downloaded. Subsequently, the authoritative online prediction tool SRAMP (https://www.cuilab.cn/sramp) was used for m⁶A modification site prediction. The “High Confidence” mode (prediction confidence score > 0.8) was selected to improve the prediction accuracy.

Acquisition and analysis of publicly available data

The public dataset GSE90639 (containing IMP2 RIP-seq and Input sample sequencing data of HEK293T cells) was obtained from the NCBI GEO database. The number of gene reads was counted using featureCounts. Subsequently, DESeq2 was used to analyze the gene differences in the IMP2 RIP sample and Input sample, and the genes with log2(enrichment factor) ≥ 1 and adjusted P value < 0.05 were selected as mRNA bound by IMP2. A gene expression scatter plot was drawn to show the distribution of these genes, and the target bound mRNA (such as LKB1) was marked.

Data analysis

Statistical analyses were conducted using GraphPad Prism software (GraphPad Company). Data were presented as mean ± standard deviation (SD). The mean comparison between the two groups was performed using an unpaired two-tailed Student’s t-test, and multiple group comparisons were conducted using one-way analysis of variance (ANOVA) combined with the Bonferroni post-hoc test. The statistical significance threshold was set at P < 0.05 as ‘*’, P < 0.01 as ‘**’, and P < 0.001 as ‘***’.

Antibody

For IHC staining, tissues were stained and treated with Anti-ALP antibody (ab65834, Abcam), Anti-RUNX2 antibody (A2851, ABclonal), Anti-OCN antibody (ER1919-20, Huabio), Anti-CD206 antibody (HA722892, Huabio), Anti- Insulin Receptor R antibody (ET1705-22, Huabio), Anti- GLUT4 antibody (R1402-3, Huabio), Goat anti- rabbit IgG H&L antibody (ab6721, Abcam)

For IF staining, tissues or cells were stained and treated with Anti-CD206 antibody (HA72289 HA601453, Huabio), Anti-IMP2 antibody (A5189, ABclonal; TA501269, ORIGENE), Anti-MFN1 antibody (A21293, ABclonal), Anti-CPT1A antibody (HA722081, Huabio), Goat anti-rabbit IgG H&L (ab150081, Abcam), Goat anti-mouse IgG H&L (ab150119, Abcam), DAPI (ab104139, Abcam), Phalloidin (C2207, Beyotime).

For Western blot, PVDF membranes were incubated with Anti- β-ACTIN antibody (AC026, ABclonal), Anti- IMP2 antibody (A5189, ABclonal), Anti-MFN1 antibody (A21293, ABclonal), Anti-MFN2 antibody(CY6638,Abways), Anti-OPA1 antibody(ET1705-9,Huabio), Anti-CPT1A antibody (HA722081,Huabio), Anti-CPT2antibody (ET1611-64,Huabio), Anti-AKTantibody (HA721870, Huabio), Anti- p-AKT antibody (ET1607-73, Huabio), Goat anti- rabbit IgG H&L antibody (ab6721, Abcam)

For flow cytometry, cells were treated with Anti-CD206 antibody (240344D12, Proteintech), Anti-CD163 antibody (155305, biolegend), Anti-F4/80 antibody (123107, biolegend)

Results

Osteogenesis disorders in a diabetic environment were accompanied by reduced M2 macrophage271 polarization and downregulated IMP2 expression

In this study, the successful establishment of the T2DM mouse model was first verified through a series of rigorous assessments. These included an insulin resistance test (Figure S1A), IHC staining for insulin receptors in insulin-target organs (i.e., skeletal muscle and liver) (Figure S1B), immunofluorescence assays demonstrating impaired GLUT4 translocation to the plasma membrane in the liver and skeletal muscle cells of STZ-induced mice (Figure S1C), as well as WB analysis of the p-AKT/AKT ratio in the insulin signaling pathway (Figure S1D). Subsequently, the characteristic bone destruction associated with diabetes was confirmed. Histological analysis revealed that compared to the control (CON) group, femoral HE staining in the T2DM group displayed typical osteoporotic features, including a reduced number of trabecular bones, increased structural sparsity, and an enlarged bone marrow cavity (Figure 1A). The expression level of CD206, a surface marker specific to M2 macrophages, was notably diminished in the bone marrow of T2DM mice (Figure 1B). Dual immunofluorescence labeling further demonstrated that the co-expression of CD206 and IMP2 in the bone marrow tissue of the T2DM group was significantly lower than that in the CON group, a finding corroborated by Pearson correlation coefficient analysis (Figure 1C). Flow cytometric profiling demonstrated that T2DM mice exhibited286 reduced counts of IMP2-positive cells in the bone marrow, with a parallel decline in the frequency of F4/80⁺CD206⁺IMP2⁺ cells—consistent with the downregulation of IMP2 expression in M2 macrophages (Figure 1D). The expression levels of osteogenic markers, including RUNX2, ALP and OCN, were significantly downregulated in the T2DM group relative to the CON group (Figure 1E, 1F, 1F, 1G).

Fig. 1.

Osteogenesis disorders in a diabetic environment were accompanied by reduced M2 macrophage polarization and downregulated IMP2 expression. (A) HE staining showing histological morphology of bone tissue in CON and T2DM groups scale bar: 100 μm. (B) IHC for the M2 macrophage marker CD206 in bone tissue scale bar, 20 μm. (C) IF double staining showing co-localization of IMP2 (red) and CD206 (green) in bone tissue; DAPI (blue) was used for nuclear counterstaining. The histogram below shows Pearson analysis of the corresponding signalsscale bar, 20 μm (D) Flow cytometry analysis of IMP2 and surface markers CD206 on BMDMs. (E-G) IHC detection of the key osteogenic transcription factor RUNX2, ALP and OCN Scale bars, 20 μm (H) Flow cytometry analysis of CD163 expression in RAW264.7 cells. The histogram on the right shows quantitative analysis. (I) TGF-β in each group of RAW264.7 cells. (J) Carried out qRT-PCR to detect the levels of IMP2 mRNA in each group of RAW264.7 cells. (K) WB analysis of IMP2 protein expression in RAW264.7 cells. (L) IF staining showing IMP2 expression (green) and CD206 (red) in RAW264.7 cells. (M-N) Representative images of ALP and ARS staining of BMSCs, and the statistical analysis results of ALP staining and ARS staining. Scale bars, 100 μm. (A-G, n=5; H-N, n=3). Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001

In in vitro experiments, we selected 30 mM glucose based on the CCK8 assay to mimic the diabetic microenvironment (Figure S1E). Under high glucose conditions, the polarization balance between M1 and M2 macrophages was disrupted, with increased expression levels of M1 macrophage markers (CD86, iNOS and IL-6) (Figure S1F) and decreased expression levels of M2 macrophage markers (CD163,ARG-1, TGF-β, and IL-10) (Figure 1H-I, S1G), indicating impaired M2 polarization capacity. Notably, high glucose significantly downregulated IMP2 expression at both transcriptional and translational levels during M2 polarization of RAW264.7 cells (Figure 1J-L). Furthermore, the high glucose microenvironment attenuated the osteogenic-promoting capacity of M2 macrophages on BMSCs, as evidenced by a significant decrease in ALP-positive cells and a reduction in the number of mineralized nodules in the high glucose group compared with the control group. (Figure 1M-N). We hypothesize that IMP2 expression may be linked to macrophage subset regulation in the diabetic milieu, thereby influencing the osteogenic-promoting ability of M2 macrophages on BMSCs.

IMP2 knockout led to a further reduction in M2 macrophages in T2DM mice and exacerbated osteogenesis disorder in DOP mice

To delve deeper into IMP2 regulation and its impact on bone mass, we generated macrophage-specific IMP2 knockout mice using the Cre-loxp system and confirmed successful knockout (Figure S2A-B). Micro-CT and HE staining revealed an increase in bone mass in macrophage IMP2 knockout mice (Figure 2A-B). Immunohistochemistry demonstrated a notable reduction in the number of CD206 positive M2 macrophages in the femurs of IMP2^f/f, Cre mice (Figure 2C). Analysis of osteogenic-related factors showed lower expression levels of ALP and RUNX2 in the IMP2^f/f, Cre group compared to the non-knockout group (Figure 2D-E). In addition, compared to the control group, the IMP2 knockout group exhibited a decreased number of TRAP positive osteoclasts, which displayed irregular morphology, smaller volume, and reduced contact with bone trabeculae (Figure 2F). In the diabetic context, IMP2 knockout further suppressed M2 macrophage polarization and exacerbated osteogenesis disorders. Specifically, Micro-CT and HE staining revealed significantly lower bone mass and reduced bone trabeculae in IMP2^f/f, Cre+T2DM mice compared to IMP2^f/f+T2DM mice (Figure 2A-B). Immunohistochemistry results showed a reduced number of CD206-positive M2 macrophages in the IMP2^f/f, Cre+T2DM group compared to the IMP2^f/f+T2DM group (Figure 2C). The number of ALP and RUNX2 positive osteogenic cells also decreased (Figure 2D-E), although there was no significant difference in the number of TRAP positive cells (Figure 2F). Furthermore, serum samples from mice in each group were collected for qRT-PCR and ELISA analyses, while BMDMs were isolated for flow cytometry analysis. The results indicated that IMP2 knockout disrupted the M1/M2 macrophage balance, characterized by increased production of pro-inflammatory cytokines, decreased secretion of anti inflammatory cytokines, and reduced expression of M2 macrophage markers, suggesting impaired M2 polarization capacity (Figure S2C-E). Collectively, these findings demonstrate that specific knockout of IMP2 in macrophages affects M2 macrophage polarization and its ability to promote osteogenic differentiation under both normal and disease conditions. This indicates that IMP2 may play a role in regulating M2 macrophage-mediated bone formation by modulating M2 polarization, and its deficiency ultimately contributes to bone mass loss in the diabetic environment.

Fig. 2.

IMP2 knockout led to a further reduction in M2 macrophages in T2DM mice and exacerbated osteogenesis disorder in DOP mice. (A) Micro-CT images of trabecular bone in the femur. Quantitative analysis of bone parameters: BV/TV, n=5. Scale bars,1 mm. (B) HE staining of paraffin-embedded mouse femora. Scale bars, 100 μm. (C-E)Representative immunohistochemical images of CD206, ALP, RUNX2, and TRAP showed the content and quantification of BV/TV (n = 5) (n=5). Scale bars, 20 μm. (F) TRAP staining in femoral tissue with quantification (n=5). Scale bar: 20 μm.Data were presented as mean ± standard deviation. Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001

IMP2 regulated M2 macrophage polarization to improve osteogenic dysfunction induced by high glucose levels

The previous results have shown that high glucose environment suppresses IMP2 expression and impedes M2 macrophage polarization. To further elucidate IMP2’s regulatory role in M2 macrophage polarization, we first overexpressed IMP2 using a plasmid (Igf2bp2 (NM_183029) pLV3-CMV-MCS- 3xFLAGCopGFP-Puro) and confirmed the transfection efficiency at both the mRNA and protein levels (Figure 3A-D). In in vitro experiments, IMP2 overexpression alone enhanced the expression of M2 macrophage markers (CD163, CD206, ARG-1) and anti-inflammatory cytokines (TGF-β, and IL-10) (Figure S3A-C), and also promoted the osteogenesis-promoting effect of M2 macrophages on BMSCs (Figure S3D). Moreover, qRT-PCR analysis revealed that overexpression of IMP2 in a high glucose environment significantly upregulated the mRNA expression levels of M2 macrophage markers (Fig. 3E) and effectively reversed the impaired M2 macrophage function induced by high glucose, which was manifested by increased secretion of the anti-inflammatory cytokines IL-10 and TGF-β (Fig. 3J). Consistent with these findings, both flow cytometry and immunofluorescence assay results confirmed that IMP2 overexpression promoted M2 macrophage polarization. (Fig. 3L). Furthermore, ALP staining and ARS staining results revealed that compared to the conditioned medium from the high glucose treatment group, the conditioned medium from the IMP2 overexpression group substantially increased the number of ALP-positive cells and the formation of mineralized nodules (Figure 3O). This indicated that IMP2 overexpression improved the osteogenic-promoting capacity of M2 macrophages under diabetic conditions. These results further confirmed that regulating IMP2 can mitigate the impaired polarization and functional defects of M2 macrophages induced by high glucose, thereby alleviating high-glucose-induced osteogenic dysfunction.

Fig. 3.

IMP2 regulated M2 macrophage polarization to improve osteogenic dysfunction induced by high-glucose levels. (A-C) Forty-eight hours after transfection, qRT-PCR and Western blotting were used to verify IMP2 overexpression in RAW264.7 cells. (D) qRT-PCR analysis of IMP2 mRNA levels in RAW264.7 cells. (E-I) qRT-PCR was used to detect the expression levels of CD163, CD206, ARG-1, IL-10 and TGF-β mRNA. (J-K) ELISA was used to detect the expression levels of IL-10 and TGF-β. (L-M) Flowcytometry was used to analyze the expression of CD163. (N) Representative immunofluorescence staining images showed the colocalization of CD206 and IMP2. Scale bars, 20 μm. (O) Representative images of ALP staining and ARS staining of BMSCs, the results were statistically analyzed, Scale bars, 100 μm. RAW264.7 cells were used in panels (D–N). Conditioned medium (CM) from RAW264.7 cells was applied to BMSCs for ALP and ARS staining (O). Data are presented as mean ± standard deviation (n=3). Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001

IMP2 regulated mitochondrial fusion to improve its dysfunction in M2 macrophages induced by high glucose levels

Mitochondrial dynamics and function directly influence cell status and differentiation [41]. Mito-Tracker staining and TEM observations revealed distinct mitochondrial phenotypes among groups, RAW264.7 cells in the high glucose induction group exhibited significantly shortened mitochondrial length, reduced fusion, mitochondrial swelling and disruption of the inner membrane structure (Figure 4A-B, D). Consistent with these morphological abnormalities, JC-1 staining showed a marked decrease in mitochondrial membrane potential in the high glucose group (Figure 4C), and the level of mitochondrial reactive oxygen species (ROS) was also significantly increased (Figure 4E). At the molecular mechanism level, qRT-PCR and WB results demonstrated that high glucose decreased the expression of key proteins involved in mitochondrial fusion (Figure 4F-G). Notably, IMP2 overexpression alone exerted a prominent protective effect. It maintained normal mitochondrial morphology (Figure S4A) and preserved mitochondrial membrane potential at a physiological level (Figure S4B), and upregulated the expression of outer membrane fusion proteins (MFN1 and MFN2) as well as the inner membrane fusion protein OPA1 (Figure S4C). Moreover, IMP2 overexpression in the high glucose environment significantly ameliorated the aforementioned high-glucose-induced abnormalities, restoring mitochondrial morphology to a near-normal state and partially reversing the impaired membrane potential and reducing mitochondrial ROS (Figure 4A-E). Additionally, compared to the high glucose group, IMP2 overexpression further enhanced the expression of the aforementioned fusion-related factors under high glucose conditions (Figure 4F-G) .

Fig. 4.

IMP2 regulated mitochondrial fusion to improve its dysfunction in M2 macrophages induced by high-glucose levels. (A) Mito-Tracker staining showed the mitochondrial morphology of each group of RAW264.7 cells. Scale bars, 10 μm (top), 2μm (bottom). (B, D) TEM images showed the mitochondrial morphology of each group of RAW264.7 cells. Red and blue arrows indicate normal mitochondria, mitochondria showing division and edema respectively. Mitochondrial morphology was assessed through quantitative morphological measurements. Scale bars, 2 μm, 500 nm. (C) JC-1 staining was used to assessmitochondrial membrane potential in RAW264.7 cells. Scale bars, 10 μm. (E) Flow cytometry analysis of mitochondrial ROS levels (MitoSOX) in RAW264.7 cells (F) qRT-PCR analysis of MFN1, MFN2,) Mito-Tracker staining showed the mitochondrial morphology of each group of RAW264.7 cells. Scale bars, 10 μm (top), 2μm (bottom). (B, D) TEM images showed the mitochondrial morphology of each group of RAW264.7 cells. Red and blue arrows indicate normal mitochondria, mitochondria showing division and edema respectively. Mitochondrial morphology was assessed through quantitative morphological measurements. Scale bars, 2 μm, 500 nm. (C) JC-1 staining was used to assess mitochondrial membrane potential in RAW264.7 cells. Scale bars, 10 μm. (E) Flow cytometry analysis of mitochondrial ROS levels (MitoSOX) in RAW264.7 cells (F) qRT-PCR analysis of MFN1, MFN2, and OPA1 mRNA levels in each group of RAW264.7 cells. (G) Western blot 756 analysis of MFN1, MFN2, and OPA1 protein levels in each group of RAW264.7 cells. (H) Mito-Tracker staining of BMDMs from the indicated genotypes and treatments, Scale bars, 2 μm. (I) Representative immunofluorescencestaining images show the co-localization of CD206 and MFN2 in femoral tissue. Scale bars, 20 μm. Data are presented as mean ± standard deviation (A-G, n=3; H-I, n=5). Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001.

In vivo experiments further corroborated the above results. We isolated BMDMs from mice in each group for Mito-Tracker staining, and the results showed consistent mitochondrial morphological changes with those observed in in vitro experiments. Specifically, BMDMs from IMP2^f/f+T2DM mice exhibited shortened mitochondrial length, while these mitochondrial abnormalities were more severe in BMDMs from IMP2 ^f/f, Cre+T2DM mice (Figure 4H). Consistent with the morphological observations, the expression of MFN2 in the femoral bone marrow of IMP2^f/f+T2DM mice was decreased, while its expression in the IMP2^f/f, Cre+T2DM group was significantly downregulated (Figure 4I). These results collectively indicate that IMP2 regulates mitochondrial fusion to maintain mitochondrial structure and function, thereby alleviating M2 macrophage impairment in the high glucose environment.

IMP2 promotes M2 macrophage polarization by regulating FAO

Mitochondrial metabolic status has a direct bearing on macrophage polarization. Further metabolic analysis revealed that during M2 polarization induced by IL-4 and IL-13, macrophages exhibited increased ATP production (Figure 5A). Notably, IMP2 overexpression alone partially enhanced ATP generation compared to the normal glucose control group (Figure S5A). However, this metabolic characteristic was significantly suppressed under high glucose conditions, while IMP2 overexpression in the high glucose environment effectively counteracted this suppression. At the molecular level, detection results demonstrated that under high glucose conditions, the expression levels of carnitine palmitoyltransferase 1 A (CPT1A) and carnitine palmitoyltransferase 2 (CPT2) were significantly downregulated at both the mRNA and protein levels (Figure 5B-F). In contrast, IMP2 overexpression alone upregulated the expression of CPT1A and CPT2 (Figure S5B-D), and this upregulatory effect was maintained even under high glucose conditions, restoring the expression of these two proteins to near normal levels (Figure 5B-H). Immunofluorescence results further corroborated this regulatory pattern (Figure 5G-H, Figure S5E). Mitochondrial function assays revealed that FAO serves as a key target for IMP2 in regulating macrophage mitochondrial metabolism. To clarify the core role of FAO in this process, we first pretreated macrophages with etomoxir, a specific inhibitor of FAO. Oxygen consumption rate (OCR) analysis demonstrated that IMP2 overexpression significantly enhanced mitochondrial respiratory capacity and reversed the high-glucose-induced reduction in macrophage mitochondrial respiration, whereas etomoxir pretreatment abrogated these effects mediated by IMP2 overexpression (Figure 5I). Further phenotypic assays showed that etomoxir treatment also markedly reversed the promoting effect of IMP2 overexpression on M2 macrophage polarization under high glucose conditions (Figure 5 J). The in vivo results were consistent with in vitro findings. CPT1A expression was decreased in femoral M2 macrophages of IMP2^f/f+T2DM and IMP2^f/f, Cre mice, with the most significant reduction observed in the IMP2^f/f, Cre+T2DM group (Figure 5 K). Collectively, these results indicate that IMP2 regulates the FAO metabolic pathway. Under high glucose conditions, IMP2 overexpression can reverse impaired FAO and OXPHOS function, thereby promoting and maintaining M2 macrophage polarization.

Fig. 5.

IMP2 promotes M2 macrophage polarization by regulating FAO. (A) Relative ATP levels in each group. (B-C) mRNA levels of CPT1A and CPT2 in each group of RAW264.7 cells. (D-F) Protein levels of CPT1A and CPT2 in each group. (G-H) Representative immunofluorescence staining images showed the expression of CPT1A, and statistical analysis was conducted. Scale bars, 10 μm. (I) Seahorse analysis of oxygen consumption rate (OCR) under the indicated conditions (including etomoxir treatment), with quantification. (J) Flow cytometry analysis of CD163 in F4/80+ RAW264.7 cells with/without etomoxir. (K)Representative immunofluorescence staining images showed the co-localization of the M2 macrophage marker CD206 with CPT1A in femoral tissue (n=5). Scale bars, 20 μm. The RAW264.7 cell line was used in the experiments described in this section. Data are presented as mean ± standard deviation (A-J, n=3; K, n=5). Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001

IMP2 stabilized LKB1 mRNA to regulate the AMPK signaling pathway and promote M2 macrophagepolarization

Given our prior observations that IMP2 modulates mitochondrial dynamics and FAO, and considering the pivotal role of the LKB1-AMPK signaling pathway in regulating cellular metabolic homeostasis, we aimed to explore whether IMP2 exerts its regulatory effects through this axis. Experimental results revealed that high glucose induction significantly reduced the expression of LKB1 and the phosphorylation level of AMPK in macrophages (Figure 6A-B). Similarly, LKB1 expression and AMPK phosphorylation were also reduced in IMP2-deficient RAW264.7 cells (Figure 6C-D, Figure S6). To explore whether IMP2 exerts its regulatory effects through this signaling pathway, bioinformatics analysis unveiled a notable positive correlation between IMP2 and LKB1 expression, and a highly reliable m⁶A modification site was identified in LKB1 mRNA (Figure 6E-F). Actinomycin D assays confirmed that IMP2 knockdown accelerated LKB1 mRNA degradation, suggesting that IMP2 may regulate LKB1 expression by maintaining its mRNA stability. (Figure 6G). To verify this mechanism, we knocked down LKB1 using siRNA, selecting siRNA-1881 as the most effective one (Figure 6H-I). Experimental results showed that LKB1 knockdown significantly reversed the protective effect of IMP2 overexpression against high glucose-induced macrophage polarization impairment. At the ultrastructural level, Mito-Tracker staining (Figure 6K) and TEM (Figure 6M) revealed that mitochondria exhibited obvious fragmentation, accompanied by inner mitochondrial membrane edema, cristae structure disappearance, and decreased mitochondrial membrane potential after LKB1 knockdown (Figure 6J). At the molecular expression level, WB results showed that LKB1 knockdown significantly downregulated the expression of mitochondrial fusion proteins (Figure 6N). In terms of metabolic function, CPT1A immunofluorescence staining (Figure 6L, P) and ATP content measurement (Figure 6O) all indicated a significant reduction in FAO capacity following LKB1 knockdown. OCR analysis confirmed the decreased mitochondrial respiratory capacity of RAW264.7 cells (Figure 6Q). More importantly, LKB1 knockdown significantly inhibited M2 macrophage polarization (Figure 7A), and ALP staining confirmed that its ability to promote osteogenic differentiation of BMSCs was significantly impaired (Figure 7B). Consistent results were observed in in vivo experiments: the expression levels of LKB1 and p-AMPK proteins in the bone marrow of IMP2 knockout mice were decreased, with the most significant reduction in the IMP2^f/f; Cre+T2DM group (Figure 7C-E). Subsequent in vivo rescue experiments demonstrated that resveratrol, which has been reported to activate the LKB1–AMPK axis [42], restored bone mass in diabetic mice by activating the LKB1-AMPK signaling pathway, as shown by HE staining (Figure 7F). Concurrently, the population of CD206+ cells in the femoral bone marrow, which was reduced in the IMP2f/f; Cre+ T2DM group, was also significantly recovered (Figure 7G). Through qRT-PCR and ELISA tests, it was observed that the ratio of M1 to M2 macrophages gradually returned to a balanced state, and the polarization tendency of M2 macrophages was improved (Figure 7H-L). These results fully demonstrate that LKB1 serves as a key downstream effector molecule through which IMP2 regulates mitochondrial function and M2 macrophage polarization, thereby further influencing the osteogenic-promoting capacity of M2 macrophages (Fig. 8).

Fig. 6.

IMP2 stabilized LKB1 mRNA to regulate the AMPK signaling pathway and promote mitochondrial fraction. (A, C) mRNA levels of LKB1 and AMPK in each group. (B, D) Protein levels of LKB1, p-LKB1, AMPK, and p-AMPK in each group. (E) Prediction of the m⁶A site on LKB1 mRNA and the highly reliable site image. (F) Gene enrichment in the GSE90639 database showed a correlation between LKB1 and IMP2. (G) Detection of the stability of LKB1 mRNA by actinomycin D. (H-I) qRT-PCR and WB were used todetect the knockdown efficiency of LKB1 siRNA in each group. (J) Detection of mitochondrial membrane potential in each group using JC-1 dye. Scale bars, 10 μm. (K) Mito-Tracker staining showed the mitochondrial morphology in each grou. (L, P) Representative immunofluorescence staining images showed the expression of CPT1A and fluorescence quantitative analysis in each group. Scale bars, 10 μm. (M) TEM images showed the mitochondrial morphology in each group. Red and blue arrows indicated normal mitochondria and mitochondria with division and edema, respectively. Mitochondrial morphology was assessed through quantitative morphological measurements Scale bars, 2μm, 500 nm. (N) WB was used to detect the protein levels of MFN1, MFN2, OPA1, CPT1A, and CPT2 in each group. (O) Relative ATP levels in each group. (Q) Seahorse analysis of OCR, with quantification of respiratory parameters. The RAW264.7 cell line was used in the experiments described in this section. Data are presented as mean ± standard deviation (n=3). Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 7.

IMP2 regulates macrophage polarization by modulating the LKB1-AMPK signaling pathway. (A) Flow cytometry analysis of CD163 expression in each group of RAW264.7 cells. (B) Representative images of ALP staining. Scale bars, 100 μm. (C-E) Representative immunofluorescence staining images showed the co-localization of CD206 with LKB1 and p-AMPK in femoral tissue (n=5). Scale bars, 20 μm. (F) HE staining of paraffin-embedded mice femora. Scale bars, 100 μm. (G-H) Representative immunofluorescence staining images showed the CD206 in femoral tissue (n=5). Scale bars, 20 μm. (I-J) mRNA levels of CD206 and CD86 in each group of BMDMs. (K-L) ELISA quantification of IL-10 and IL-6 in BMDM supernatants from the indicated groups. Data are presented as mean ± standard deviation (A-B, n=3; C-L n=5). Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 8.

The mechanism diagram that illustrates how IMP2 affects mitochondrial dynamics and FAO via the LKB1-AMPK pathway, thereby regulating the M2 polarization pattern and improving DOP. The figure was created with BioRender.com

Discussion

This study investigated the regulatory role of IMP2 in DOP and uncovered its underlying mechanisms. IMP2 stabilizes LKB1 mRNA via m⁶A-dependent mechanisms, thereby activating the LKB1-AMPK signaling pathway. This process is central to regulating mitochondrial dynamics and FAO, ultimately improving DOP by promoting M2 macrophage polarization. The pathogenesis of DOP involves intricate interactions between metabolic disorders and bone metabolism imbalances. Abnormal regulation of bone immunity, a pivotal link between the immune system and bone metabolism, has garnered increasing attention in DOP research [2, 43]. Insulin deficiency or resistance, leading to hyperglycemia, can, disrupt the phenotypes and functions of immune cells, promote excessive release of pro-inflammatory factors, disturb the osteogenesis-osteolysis balance, and ultimately exacerbate bone loss. Our previous study demonstrated that promoting M2 macrophage polarization and TGF-β secretion plays a critical role in enhancing the osteogenic differentiation of BMSCs in DOP [10]. The findings of this study align with previous research and further revealed the reasons for impaired M2 macrophage polarization in the DOP environment and identifying IMP2 as a key regulatory factor. This study confirmed that upregulating IMP2 expression can restore M2 macrophage polarization inhibited by high glucose levels, thereby alleviating osteogenic dysfunction induced by a hyperglycemic environment. Consistent with the results of this study, IMP2 enhances the mRNA expression of TSC1, PPARγ, CSF1, MYC, etc [17, 44, 45], promoting M2 macrophage polarization. These studies collectively reveal that IMP2 operates via an m⁶A-dependent mRNA stability mechanism, primarily targeting cytokines or signaling molecules related to immune phenotypes, without involving key regulatory factors of the energy metabolism pathway. This study, however, focused on exploring IMP2’s role in regulating mitochondrial function and energy metabolism to promote M2 macrophage polarization. Recent studies have highlighted that M2 macrophage polarization is heavily reliant on the adaptive regulation of metabolic phenotypes, with OXPHOS serving as a characteristic metabolic pathway [46]. As a central regulator of energy metabolism, the AMPK signaling pathway plays a crucial role in maintaining mitochondrial functional homeostasis and promoting OXPHOS [47]. LKB1, an upstream activator of AMPK, enhances its activity by promoting phosphorylation at the Thr172 site of the AMPK α subunit. LKB1 is a core molecule regulating energy sensing and mitochondrial metabolism, and this pathway also influences macrophage functions. Existing research has demonstrated that m⁶A functions as a regulator of LKB1 expression [31, 32]. The results of this study confirm that IMP2 regulates the stability of LKB1 mRNA and protein expression and activating the LKB1-AMPK pathway. This pathway maintains mitochondrial network integrity by upregulating mitochondrial fusion-related proteins, enhances the activity of key FAO enzymes to provide sufficient substrates for OXPHOS, and ultimately establishes a metabolic foundation for M2 macrophage polarization. This mechanism not only fills a gap in the field of energy metabolism regulated by IMP2 targets, but also for the first time, reveals the role of the IMP2-LKB1-AMPK axis in macrophage polarization and bone metabolism. It proposes a new mechanism of m⁶A modification regulating LKB1 mRNA stability, and expands the research scope of RNA epigenetics in metabolic bone diseases.

Notably, this study also observed a special phenomenon: bone mass in non-diabetic mice with macrophages specific knockout of IMP2 was significantly increased. TRAP staining revealed a decreased number of osteoclasts, increased the distance from the bone trabeculae, irregular morphology, and reduced the cell volume. The expression of osteogenic-related factors, ALP and RUNX2, also decreased. This suggests that the increased bone mass results from the combined effects of osteoclast inhibition and reduced osteogenesis. However, in a diabetic environment, there was no significant difference in osteoclast status between the IMP2 knockout and non-knockout groups, but ALP and RUNX2 levels in the knockout group were further decreased, leading to exacerbated osteogenesis disorders and more pronounced bone loss. Previous studies have clearly demonstrated that T2DM does not merely affect a single cell, but rather regulates the functions of various cells in the bone metabolism network through multiple pathological factors such as hyperglycemia and inflammatory microenvironments [48–50], for instance, T2DM can directly promote the generation and activation of osteoclast, and m⁶A modification has been proven to be a key regulatory mechanism mediating osteoclast differentiation under diabetic conditions [51, 52]. This background suggests that changes in bone mass are the result of the coordinated action of multiple cells, rather than being independently regulated by a single gene or cell type. Combining the results of this study, we hypothesize that the phenotypic differences between IMP2 knockout mice in non-diabetic and diabetic states are primarily due to the special pathological environment of diabetes altering the response patterns of different cells to IMP2. Existing studies have shown that in the physiological environment of non-diabetes, IMP2 can promote the activity of osteoclast [53], thus knocking out IMP2 will inhibit the function of osteoclasts, ultimately leading to an increase in bone mass; however, in the pathological environment of T2DM, factors such as hyperglycemia might weaken the sensitivity of osteoclasts to IMP2, at this time, the regulatory effect of IMP2 on macrophage polarization might become the dominant factor affecting bone metabolism, when IMP2 is absent in macrophages, the polarization of M2 macrophages is inhibited, its promoting effect on osteogenic differentiation of BMSCs is weakened, ultimately leading to further damage to bone formation and more severe bone loss. This phenomenon suggests that the function of IMP2 varies among different cell types and pathological microenvironments, and its mechanism of action in the bone metabolism network is more complex. This study, through a conditional knockout model of macrophages, clarified the key role of IMP2 in regulating bone metabolism through modulating macrophage polarization in the T2DM environment; while the regulatory mechanisms of T2DM on other cell types such as osteoclasts and BMSCs, as well as the interactions between these cells and macrophages, are all research directions that need to be further explored in the future.

This study has certain limitations. Firstly, the in vitro experiments only used high glucose as a single factor to simulate the in vivo microenvironment of diabetes, without incorporating other pathological features of diabetes. Therefore, this method can only partially reproduce the complex in vivo pathological state. Secondly, this study only focused on the pathological changes of macrophages after the absence of IMP2, but did not further explore the direct role of IMP2 in other cells related to bone metabolism. The overall regulatory network still needs improvement, and this is also a direction that requires more in-depth research in the future.

In summary, this study demonstrates that IMP2 stabilizes LKB1 mRNA through m⁶A modification and promotes its translation, enhancing mitochondrial dynamics and FAO metabolism through the LKB1-AMPK pathway, and increasing OXPHOS levels, ultimately promoting M2 macrophage polarization and BMSCs osteogenic differentiation, thereby improving DOP. These efforts will facilitate the transition of IMP2 from basic research to clinical application and provide new insights for the prevention and treatment of metabolic bone diseases.

Supplementary Information

Below is the link to the electronic supplementary material

Author contributions

Conceptualization: LMQ, GJ, LHR, LLS.

Methodology: LLS, CYJ, JY.

Investigation: LLS, CYJ, JY, MK, LXJ, LT, SJY.

Visualization: LLS, CYJ, JY, MK, LXJ, LT, SJY.

Funding acquisition: LMQ, GJ, LHR.

Project administration: LMQ, GJ, LHR.

Supervision: LMQ, GJ, LHR.

Writing: LLS, CYJ, JY, MK, LXJ, LT, SJY.

Funding

This study was partially supported by the TaiShan Scholars of Shandong Province (No. tstp20221160) to Li M, the National Natural Science Foundation of China (No.82370999) to Guo J, the Construction Engineering Special Fund of “Taishan Young Scholars” of Shandong Province (No. tsqn202103177) and the Natural Science Foundation of Shandong Province (No. ZR202210210042).to Liu H.

Data Availability

All data are available in the main text or the supplementary materials.

Declarations

Ethics approval and consent to participate

All animal procedures in this study were conducted in accordance with the guidelines for the care and use of laboratory animals of the National Institutes of Health of the United States. All animal experiments were approved by the Experimental Animal Management and Use Committee of Shandong University School of Stomatology (Approval Number: 20240802).

Consent for publication

Not applicable.

Competing interest

The authors declare no competing or financial interests.