Targeting chaperone-mediated autophagy to regulate osteoclast activity as a therapeutic strategy for osteoporosis

Abstract

Osteoporosis is an age-related bone metabolic disease characterized by a persistent bone mass decrease and bone structure destruction. Osteoclasts, important cells in the bone remodelling process, are closely associated with the onset and progression of osteoporosis; however, the regulatory mechanisms involved remain unclear. In this study, TFE3 cytoplasmic translocation inhibited LAMP2A expression in osteoclasts and precursors in elderly individuals with osteoporosis and the downregulation of LAMP2A expression mediated the attenuation of chaperone-mediated autophagy (CMA). This inhibition prevented intracellular CCR5 degradation, increased the osteoclast differentiation of osteoclast precursor cells, and enhanced the bone resorption activity of mature osteoclasts, leading to bone loss and remodelling. In addition, we constructed osteoclast-targeted nanoparticles carrying CMA activators and demonstrated that enhancing osteoclast CMA activity in vivo inhibited the abnormal bone resorption activity of osteoclasts, thereby effectively increasing bone mass and alleviating osteoporosis progression. This study revealed that LAMP2A-mediated CMA activity in osteoclasts and their precursors negatively regulates osteoclast differentiation and bone resorption activities both in vivo and in vitro. The attenuation of LAMP2A-mediated CMA activity plays an important role in the development of osteoporosis, and enhancing LAMP2A-mediated CMA activity represents a potential therapeutic strategy for osteoporosis.

Graphical abstract

1. Introduction

Osteoclasts originate from the haematopoietic stem cell lineage and differentiate from monocytes/macrophages [1]. In addition to their role in bone resorption, osteoclasts and their precursor cells function as immune cells and possess the ability to secrete various cytokines; additionally, they directly and indirectly regulate other cellular functions in the bone microenvironment through receptor‒ligand modulation and direct cellular contact, thereby contributing significantly to the regulation of bone mass and immune function within the bone microenvironment [2]. Past studies have linked the pathogenesis of osteoporosis to various processes at the tissue, cellular and molecular levels. Osteoclasts, which are important cells in the bone remodelling process, are intricately involved in the onset and progression of osteoporosis [3]. Numerous cytokines coregulate osteoclasts and their precursors, which promotes physiological processes such as osteoclast and precursor multinucleation, activation, and proliferation [4]. The abnormal activation of osteoclast function is closely linked to a variety of bone-related diseases [[5], [6], [7]].

Osteoporosis is a metabolic bone disease characterized by persistent bone mass reduction and microstructural degradation, resulting in marked fragility and heightened susceptibility to pathological fractures [4]. However, as a metabolic bone disease in elderly individuals, the occurrence of osteoporosis is intricately linked to senescence and systemic hormonal fluctuations. In recent studies, osteoporosis has been widely attributed to an imbalance between bone formation and resorption [8,9]. The balance between bone formation and bone resorption in the body primarily depends on the mutual maintenance of two key cells in the bone microenvironment: osteoblasts and osteoclasts [10]. In osteoporosis, the rate of bone loss is faster than that of bone formation, and osteoclasts mediate the abnormal increase in bone resorption. Currently, the targeted drug used for the clinical treatment of osteoporosis is RANKL antagonists; they can effectively inhibit the abnormal functional activities of osteoclasts in patients. In long-term clinical practice, the regulation of osteoclast activities by RANKL antagonists can effectively alleviate bone loss in patients with osteoporosis [11]. The evidence presented indicates a close association between osteoclast-mediated bone loss and remodelling and the development of osteoporosis. However, the precise molecular mechanisms underlying the increased number of osteoclasts and their aberrant bone resorption in osteoporosis patients remain unclear. An in-depth exploration of these molecular pathways has significant implications for understanding osteoporosis pathogenesis and defining effective treatment strategies.

Autophagy is a lysosome-dependent degradation pathway and one of the major degradation paths of cytoplasmic proteins and organelles in cells [12]. Chaperone-mediated autophagy (CMA) is a selective degradation process dependent on the molecular chaperone protein HSC70. This degradation process occurs when the amino acid sequences of the substrates contain KFERQ-like motifs. HSC70 recognizes the KFERQ-like motif in target substrates, capturing and directing substrates to lysosomes for degradation [13]. The docking process between lysosomes and HSC70 is facilitated by lysosome-associated membrane protein type 2A (LAMP2A), a splice variant of LAMP2 that recognizes and captures CMA substrates carried by HSC70, transporting them into the lumen of lysosomes for degradation [14]. As LAMP2A and HSC70 are regulators of CMA, molecular changes in both LAMP2A and HSC70 significantly influence CMA activity. The regulation and degradation of specific proteins by CMA can effectively modulate lipid metabolism, the immune response, the cell cycle, and cell differentiation. Bone marrow mesenchymal stem cells (MSCs) are the primary source of osteoblasts in the bone microenvironment. Recent studies have highlighted the significant role of CMA in balancing the lipogenic-osteogenic differentiation of MSCs and revealed that the overexpression of LAMP2A in MSCs contributes to bone fracture healing [15]. CMA activity decreases with age [16,17], and osteoporosis is a bone metabolic disease associated with ageing. Therefore, we speculate that the decrease in CMA with age is related to bone loss in osteoporosis. The regulation of bone mass in osteoporosis primarily relies on the interaction between osteoblasts and osteoclasts in the bone microenvironment [18]. Studies have shown that PRL2, which is regulated by CMA, exerts an inhibitory effect on osteoclasts, indicating that CMA significantly contributes to regulating the balance between osteoblasts and osteoclasts in the bone microenvironment [19]. Nonetheless, the role and specific molecular mechanisms of CMA in mediating functional changes in precursor osteoclasts (pOCs) and mature osteoclasts (mOCs) in osteoporosis remain unclear.

In this study, we examined the activity of chaperone-mediated autophagy (CMA) in bone samples obtained from osteoporosis patients and a mouse model of osteoporosis. Our findings revealed significantly downregulated LAMP2A expression in osteoporotic osteoclasts, which consequently led to the suppression of CMA activity. TFE3 plays a key role in inhibiting LAMP2A expression in osteoporosis. The suppression of LAMP2A expression inhibited CCR5 degradation by weakening CMA activity, subsequently promoting osteoclast differentiation in pOCs and bone resorption in mOCs. Overall, CMA suppression enhanced the bone resorption of osteoclasts, leading to further bone loss and remodelling. Furthermore, we developed nanoparticles that specifically target osteoclasts. In vivo experiments confirmed that increasing CMA activity via a CMA activator carried by nanoparticles effectively suppressed the aberrant bone resorption activity of osteoclasts, consequently increasing bone mass in osteoporotic mice.

2. Methods

2.1. Isolation and culture of human monocytes and osteoclast differentiation

Experiments were carried out according to methods previously reported in the literature. First, human whole blood was mixed with an equal volume of phosphate-buffered saline (PBS) and then slowly added to isolated human monocytes at a 1:1 ratio. Centrifugation was performed at 400 g at room temperature for 35 min, with acceleration set to 9 and deceleration set to 1. After centrifugation, a pipette was used to aspirate the stratified greyish-white intermediate layer containing peripheral blood mononuclear cells (PBMCs). The cells were washed three times with PBS and then centrifuged at 300 g for 5 min at room temperature; this process was repeated twice. The PBMCs were counted and then added to macrophage-inducing medium (α-MEM complete medium containing 25 ng/ml M-CSF, 10 % FBS, and 1 % penicillin‒streptomycin) for culture. The medium was changed every 3 days to remove stray cells and obtain purer monocytes/macrophages. To induce osteoclast differentiation, cells were seeded into 12-well plates at a density of 0.6 × 10⁶ cells per well, and pOCs were induced by adding macrophage induction medium for 3 days and osteoclast differentiation induction medium (α-MEM containing 25 ng/ml M-CSF, 50 ng/ml RANKL, 10 % FBS, and 1 % penicillin‒streptomycin) for 3 days to induce mOCs. The cells were cultured in a cell culture incubator at 37 °C with 95 % air and 5 % CO2.

2.2. Animal models

This study was approved by the Animal Ethics Committee of Sun Yat-sen University (approval No. 2023d048). C57BL/6J mice were acquired from the Laboratory Animal Centre of Sun Yat-sen University and were housed in a pathogen-free barrier environment with a 12-h light‒dark cycle. The mice were allowed free movement and access to water and food throughout the experimental procedure. Mice were anesthetized via intraperitoneal injection of 200 mg/kg tribromoethanol (Afludine).

For the osteoporosis mouse model, 16-month-old C57BL/6J mice were used for age-related osteoporosis, and 8-week-old mice were used as young controls. Postmenopausal osteoporosis was induced in 8-week-old female C57BL/6J mice under isoflurane anaesthesia. The uterus and ovaries were removed, and in the control group, approximately 1 g of periovarian fat was removed, preserving the ovaries. The wounds were subsequently cleaned and sutured. After surgery, the mice received penicillin to prevent infection. Osteoporosis was evaluated 8 weeks postsurgery. Starting seven days after ovariectomy, the mice were given intravenous injections of nanoparticles carrying drugs every 3 days. Bone samples from femurs and the skull were collected 12 weeks later.

For femoral bone defects in mice, bone defects were created in the femurs of 8-week-old mice via an electric bone drill. Prior to surgery, the skin at the femur of each mouse was sterilized with alcohol. The skin was incised, and the subcutaneous tissue was separated to expose the femur. A 1.0 mm drill was used to create a femoral defect. Postoperatively, penicillin was administered to prevent infection, and the femur was collected for micro-CT analysis 2 weeks later.

For the LPS-induced bone destruction model, 8-week-old male mice were injected with 25 mg/kg LPS (Millipore Sigma) or PBS above the skull. Moreover, drug-carrying targeted nanoparticles were injected into the mice through the tail vein. After 2 weeks, the skull of each mouse was collected and analysed via micro-CT and tartrate acid phosphatase (TRAP) staining.

2.3. Collection of human bone samples

Osteoporosis samples were taken from elderly patients undergoing surgery for osteoporosis, while normal control samples were taken from post-accident surgery patients. Criteria for the non-osteoporotic group included men aged ≤50 years with a BMD t-score ≥ -1. Osteoporotic patient selection criteria included male patients aged >50 years with BMD t-scores ≤−2.5, or male patients with fragility fractures with BMD t-scores between > −2.5 and <-1. The study was approved by the Ethics Committee of The Eighth Affiliated Hospital, Sun Yat-sen University (Clinical ethical approval No. 2023r032). After obtaining informed consent from the patients and their agreement to publish the data, we collected femoral heads from the patients for our study. Detailed characteristics of the patients in our study are provided in the Supplementary Information (Table S6).

2.4. TRAP staining

Following three washes with PBS, osteoclasts cultured in either 24- or 96-well plates were fixed with 4 % paraformaldehyde for 20 min at room temperature and then washed three times with PBS. TRAP staining was conducted using a leukocyte acid phosphatase kit (Sigma, Australia) following the manufacturer's instructions. The cells were subsequently incubated with anti-TRAP staining solution for 30 min at 37 °C in the dark and then washed three times with distilled water. The staining results were captured via a microscope imaging system and analysed via ImageJ software.

2.5. Bone slice resorption assay

Bovine bone slices preserved in 75 % ethanol were soaked in sterile PBS for 12 h, transferred to 24-well plates and rinsed three times with culture medium. PBMCs (0.3 × 10⁶ per well) were seeded on the surface of the bone slices, followed by incubation. After appropriate interventions, osteoclast differentiation was induced as previously described. After 7 days of osteoclast differentiation, the cells were dissociated with a cell dissociation solution (Sigma‒Aldrich, Australia), and then, the surface of the bone slices was scraped with a brush. Bone fragments coated with Au-Pd were observed using a scanning electron microscope (ZEISS, Germany). The bone resorption area was measured using ImageJ software.

2.6. Western blotting

The culture fluid was removed from the cell culture plates/flasks, and the cells were rinsed three times with PBS; excess fluid was aspirated using a lance tip. The cell lysis buffer containing 1 % protease and phosphatase inhibitors was prepared and added to the cells. The cells were collected after lysis and centrifuged at 14,000 rpm for 15 min at 4 °C. The protein concentration was measured via the BCA method, the protein amount was adjusted, and the samples were stored at −80 °C. Proteins were separated by 10 % SDS‒polyacrylamide gel electrophoresis and transferred to a PVDF membrane, followed by three rinses with TBST solution. The membrane was blocked with 5 % skim milk for 1 h and then incubated overnight at 4 °C with primary antibodies against LAMP2A (Abcam Cat# ab12506), HSC70 (Abcam Cat# ab51052), TFE3 (Thermo Fisher Cat# PA5101090), CCR5 (Thermo Fisher Cat# 14-1957-82), LAMP1 (Abcam Cat# ab289548), Lamin B1 (Abcam Cat# ab 16048), NFATC1 (Proteintech Cat# 66963), c-FOS (Proteintech Cat# 66590), CTSK (Abcam Cat# ab187647), GAPDH (Cell Signaling Technology Cat# 5174), DYKDDDDK Tag (Cell Signaling Technology Cat# 8146) and HA (Abcam Cat# ab236632). The membrane was washed with TBST, incubated with secondary antibodies, and washed again. ECL solution was prepared and added to the membrane. Light signals were analysed using ImageJ software to determine protein expression levels.

2.7. Proteomic sequencing and analysis

After LAMP2A was knocked down in osteoclast precursor cells, cellular proteins were extracted as described above. LC‒MS/MS analysis was performed on a timsTOF Pro mass spectrometer (Bruker) coupled to a NanoElute (Bruker Daltonics) for 60, 120, or 240 min. Peptides were loaded onto a reversed-phase capture column (Thermo Scientific Acclaim PepMap100, 100 μm ∗ 2 cm, nanoViper C18), which was connected to a C18 reversed-phase analyser column (Thermo Scientific Easy column, 10 cm long, 75 μm inner diameter, 3 μm resin) in Buffer A (0.1 % formic acid). The samples were separated in a linear gradient with Buffer B (84 % acetonitrile and 0.1 % formic acid) at a flow rate of 300 nl/min, which was controlled by IntelliFlow technology. The mass spectrometer was operated in positive ion mode. The mass spectrometer acquired ion mobility mass spectra in the mass range m/z 100–1700 and mass range 1/k0 0.6–1.6, with a target intensity of 1.5k and a threshold of 2500. Ten PASEF MS/MS cycles were performed. Active exclusion was activated with a release time of 0.4 min. The raw MS data for each sample were combined and retrieved for identification and quantitative analysis using MaxQuant 1.5.3.17 software.

2.8. RNA extraction and quantitative real-time PCR

The medium in cell culture plates/flasks was removed, and the cells were rinsed with PBS three times; the excess liquid was aspirated. The entire RNA extraction process was conducted with enzyme-free consumables in an ice bath. TRIzol solution (1 ml per well) was added to the cells, followed by cell disrupted and RNA extraction following the manufacturer's instructions. The RNA from the samples was quantified and adjusted to equal amounts and volumes. The RNA was subsequently reverse transcribed to cDNA using the TaKaRa PrimeScript RT Reagent Kit. The samples were spiked and assayed following the instructions provided with the SYBR Green Premix Ex Taq system (TaKaRa). The PCR program was configured as follows: initial denaturation at 95 °C for 1 min, followed by denaturation at 95 °C for 30 s, annealing at 58 °C for 20 s, and extension at 72 °C for 30 s, for a total of 40 cycles, with a final elongation step at 72 °C for 5 min. The gene expression level was determined using a real-time fluorescence quantitative PCR instrument from Thermo Fisher, and the relative expression was calculated using the 2^−ΔΔCt method. The primers used for mRNA expression detection in this study are shown in Table S1.

2.9. RNA sequencing and data analysis

After LAMP2A was knocked down in osteoclast precursor cells, RNA was extracted as described above. Libraries were constructed using the ABclonal mRNA-seq Lib Prep Kit (ABclonal, China). The mRNA was purified from 1 μg of total RNA using oligo(dT) magnetic beads, fragmented with divalent cations at high temperature, and then synthesized as first-strand cDNA using random hexamer primers and reverse transcriptase (RNase H) and as second-strand cDNA using DNA polymerase I, RNase H, buffer, and dNTPs. Double-stranded cDNA fragments were ligated into paired-end libraries by junctions for PCR amplification. PCR amplification was performed, and the products were purified with the AMPure XP system. Library quality was assessed using an Agilent Bioanalyzer 4150 system. Sequencing was performed using Illumina NovaSeq 6000 or MGISEQ-T7 instruments. Data generated from the Illumina/UW Genetics platform were subsequently analysed using the in-house pipeline of Shanghai Applied Protein Technology, and differential expression was analysed using DESeq2 with a |log2Fold Change| > 1 and a Q value < 0.05.

2.10. Immunofluorescence staining

Prior to immunofluorescence staining, cells were washed with PBS and fixed in 4 % paraformaldehyde, and tissue sections were deparaffinized and subjected to antigen retrieval. Cells and tissues were permeabilized with 0.1 % Triton X-100, washed with PBS, blocked with 10 % goat serum, and incubated overnight at 4 °C with primary antibodies against LAMP2A (Abcam Cat# ab12506), HSC70 (Abcam Cat# ab223356), CTSK (Santa Cruz Cat# SC-4835), TFE3 (Thermo Fisher Cat# PA5101090), CD14 (Thermo Fisher Cat# MA5-32248), CCR5 (Thermo Fisher Cat# 14-1957-82), DYKDDDDK Tag (Cell Signaling Technology Cat# 8146) and HA (Abcam Cat# ab236632). After incubation, the cells and tissues were washed with PBS, incubated with species-specific secondary antibodies conjugated with fluorescent groups for 1 h at room temperature, and shielded from light. After incubation, the secondary antibody solution was removed, and the cells and tissues were washed three times with PBS to eliminate nonspecific staining. Cells and tissues were incubated with DAPI at room temperature for 10 min to stain nuclei. Fluorescently stained sections were observed and photographed using fluorescence or confocal microscopy and then analysed with ImageJ software.

2.11. Cell cytoplasmic and nuclear fractionation

Cytoplasmic and nuclear fractions were separated via a Nuclear and Cytoplasmic Protein Extraction Kit (Beyotime, China). Proteins were extracted from each fraction. The distribution of proteins in each cellular fraction was analysed through Western blotting. Lamin B1 served as a positive control for the nuclear fraction, whereas GAPDH served as a positive control for the cytoplasmic fraction.

2.12. Coimmunoprecipitation (Co-IP)

Co-IP was performed using a Dynabeads™ Protein G Immunoprecipitation Kit (Invitrogen). Supernatants were mixed with IP lysis buffer containing 1 % phosphatase and protease inhibitor cocktail for 30 min on ice. The samples were then centrifuged at 14,000×g for 15 min, and the supernatants were collected according to the manufacturer's protocol. The samples were precleared with magnetic beads to eliminate nonspecific binding. The desired amount of primary antibody (anti-CCR5, anti-HSC70, anti-HA and anti-DDDDK-Tag) was subsequently added to the precleared samples, which were subsequently incubated overnight at 4 °C with moderate rotation. Magnetic beads were added to the samples, which were subsequently incubated for 2 h. The beads were collected and washed three times with wash buffer. The magnetic beads were mixed with SDS‒PAGE loading buffer and boiled at 100 °C for 10 min. The immunoprecipitates were collected after removal of the magnetic beads. The samples were separated on an SDS‒PAGE gel and then stained with a Coomassie blue staining kit (Beyotime Institute of Biotechnology). Western blotting was used to assess the interaction between two proteins.

2.13. DNA pull-down assay and mass spectrometry

A biotin-labelled oligonucleotide probe targeting LAMP2 was designed and synthesized by RiboBio. DNA pull-down experiments were conducted using a Bes5004 DNA pull-down kit (BersinBio, China) following the manufacturer's protocol. Briefly, streptavidin magnetic beads were coincubated with biotinylated probes and lysis buffer at 4 °C to form protein‒DNA‒magnetic bead complexes. After being pulled down with magnets, the DNA-bound proteins were washed with buffer, and subsequently, the proteins within the complex were evaluated either by Western blotting or liquid chromatography (LC)-mass spectrometry (MS)/MS.

2.14. siRNA interference

LAMP2A, TFE3, CCR5, and negative control siRNAs were designed and synthesized by IGE (Guangzhou, China). Three different interfering sequences were designed for each target molecule, and three different siRNA sequences were screened before the formal experiments; the siRNA with the highest knockdown efficiency, exceeding 70 %, was selected. The cells were transfected with Lipofectamine™ RNAiMAX (Thermo Fisher, USA) when they reached a confluence of 60–80 %, following the manufacturer's protocol. The knockdown efficiency was assessed via Western blotting or qRT‒PCR after 48 h. For experiments with a longer differentiation induction period, siRNA re-knockdown was performed on day 6. The sequences of the siRNAs used in this study are shown in Tables S2–5.

2.15. Construction and transfection of plasmids and lentiviruses

The plasmids used were pcDNA3.1(+)-Flag-HSC70, pcDNA3.4(+)-HA-CCR5, and pcDNA3.4(+)-HA-CCR5#, which were designed and constructed by IGE (Guangzhou, China). Upon reaching a confluence of 70 %–80 %, 293T cells were transfected with Lipofectamine 3000 and P3000 (Thermo Fisher, America), followed by a 10–15 min incubation at room temperature. The transfection mixture (2 ml) was combined with an equal volume of complete medium, followed by thorough mixing. The mixture was added to the cells, followed by incubation at 37 °C for 24 h, after which the transfection medium was replaced with fresh medium for subsequent experiments.

LAMP2A and TFE3 full-length overexpression lentiviruses were designed and synthesized by OBiO (Shanghai, China). Upon reaching a confluence of 60 %–70 %, transfection was performed, and the viral titre was determined on the basis of the number of cells and the viral titre (12-well plate: 0.6 × 10⁵/well). The viral mixture, polybrene (5 μg/ml), and complete DMEM were mixed and added to the cells. After 24 h, the medium containing the lentivirus was replaced with fresh medium for subsequent experiments and assays.

2.16. Immunohistochemical staining and H&E staining

Freshly isolated bone tissue samples were fixed in 4 % paraformaldehyde for 24 h, decalcified with EDTA, embedded in paraffin, and sectioned. The tissue sections were deparaffinized at 60 °C, followed by sequential immersion in xylene, gradient alcohol, and deionized water. The sections were stained with haematoxylin, which was added dropwise, for 5 min, followed by a rinse with PBS for 10 min.

For H&E staining, the sections were stained with eosin for 5 min, rinsed with PBS, dehydrated with ethanol, blocked with neutral resin, and sealed. For immunohistochemical staining, the sections were placed in a beaker with sodium citrate buffer and heated in a water bath at 60 °C for 12 h for antigen retrieval. The stained area was circled with an immunohistochemical pen, and the sections were sealed with goat serum for 30 min. Then, a diluted primary antibody was added, and the sections were incubated overnight at 4 °C. After incubation, the sections were washed three times with PBS. HRP-labelled secondary antibody was then added to the sections, which were subsequently incubated for 15 min at room temperature and washed three times with PBS. DAB chromogenic solution was prepared and applied to the sections for colour development, followed by washing with PBS to stop the reaction. The sections were stained with haematoxylin for 5 min, differentiated with 1 % hydrochloric acid in alcohol, rinsed in tap water, dehydrated with ethanol, and sealed with neutral resin. The stained sections were observed and photographed under a light microscope after sealing.

2.17. Preparation and characterization of the nanoparticles

The raw materials for the targeted nanoparticles were purchased from Tanshtech (China). The synthesis of (Asp)8-PEG-PCL was carried out through two main steps: (1) MAL-PEG-PCL block copolymers were synthesized by ring-opening polymerization of ε-CL and MAL-PEG at 130 °C in the presence of Sn(Oct)2, and the resulting products were dialyzed using dialysis membranes (MWCO 3500Da, BioSharp, China) and then lyophilized; (2) (D-Asp)8 and MAL-PEG-PCL were stirred in PBS at a molar ratio of 2:1 for 24 h at room temperature to synthesize (D-Asp)8-PEG-PCL nanoparticles; the products were collected by lyophilization after dialysis using a dialysis membrane (MWCO 5000Da, BioSharp, China) for 2 days. Finally, 5 mg of activator or 5 mg of inhibitor were mixed with 20 mg of lyophilized product, and the mixture was dried in a tower dryer, rehydrated with saline, and sterilized through a 0.22 mm sterile filter. Mice were intravenously injected with (D-Asp)8-PEG-PCL, (D-Asp)8-PEG-PCL@AR7, and (D-Asp)8-PEG-PCL@QX77 solutions 100 μl at a concentration of 500 μg/ml into the tail vein. The morphology of the nanoparticles was observed and photographed via transmission electron microscopy. The size and zeta potential of the nanoparticles were determined using dynamic light scattering (DLS). Small animal in vivo imaging and immunofluorescence staining were used to assess the distribution of the nanoparticles in the mice. In addition, the cytotoxicity of the nanoparticles in vivo and in vitro was assessed by HE staining of tissue sections from various organs in mice and a Cell Counting Kit-8 (CCK-8) assay. For targeted treatment, C57 mice were given different nanoformulations (PEG-PCL, (D-Asp)8-PEG-PCL, (D-Asp)8-PEG-PCL@AR7 and (D-Asp)8-PEG-PCL@QX77) at a dose of 1 mg/kg via intravenous injection.

2.18. Computational simulation of protein docking

The AlphaFold structure of the HSC70 protein (UniProt ID: P11142) and the crystal structure of the CCR5 protein (UniProt ID: P51681) were downloaded from the RCSB Protein Data Bank (http://www.rcsb.org/). The optimized protein structures were carefully treated in several steps, including residue repair, protonation, and partial charge assignment in the AMBER ff14SB force field. Protein‒protein blind docking with HSPA8 and CCR5 was initiated using the HDOCK server (https://hdock.phys.hust.edu.cn/) [20]. The top 10 predictions of the complexes were obtained. On the basis of the scoring and binding modes, the top-ranked complex was selected for binding mode analysis.

2.19. Microcomputed tomography (micro-CT)

C57BL/6J mouse femurs and calvarias were fixed in 10 % neutral buffered formalin and then subjected to microcomputed tomography (Skyscan 1276) at 55 kV, 200 μA, and 460 ms. The two-dimensional and three-dimensional structures were produced by reconstructing image slices using NRecon software. Bone density and trabecular morphology were analysed by determining the bone volume/total volume (BV/TV), trabecular thickness (Tb.Th), cortical thickness (Ct.Th), trabecular number (Tb.N) and trabecular spacing (Tb.Sp) using CTAn software.

2.20. Statistical analysis

The data in this study were analysed using SPSS 26.0 software (Chicago, USA), and the results are presented as means ± standard deviations (SDs). Significant differences between two groups were assessed using Student's t-test, whereas significant differences among more than two groups were assessed using one-way analysis of variance (ANOVA) followed by Bonferroni correction. A significant difference was defined as P < 0.05.

3. Results

3.1. Reduced levels of LAMP2A-mediated CMA in osteoporosis

Osteoporosis is an age-related degenerative bone disease. Previous studies have suggested that the CMA activity level is closely related to changes in bone mass and that ageing is an important factor affecting CMA activity levels. To further investigate the relationship between CMA activity levels and osteoporosis in elderly individuals, we obtained bone tissue specimens from healthy individuals and elderly patients with osteoporosis in the clinic. We also generated models of young (8-week-old) and elderly osteoporotic (18-month-old) mice. Immunofluorescence of LAMP2A and HSC70, key molecules involved in CMA, revealed that CMA activity was significantly decreased in osteoporosis. There was no significant difference in HSC70 levels, but LAMP2A expression was significantly downregulated in the osteoporosis group (Fig. 1A–B). The decrease in CMA activity in the bone microenvironment was confirmed to be caused by the downregulation of LAMP2A expression. A study by Gong demonstrated that CMA positively regulates MSC osteogenic differentiation and negatively regulates lipogenic differentiation in senile osteoporosis [21]. Osteoclasts are crucial for maintaining bone homeostasis, and no study has reported changes in CMA activity in osteoclasts in osteoporosis. To clarify whether the reduced CMA activity in osteoclasts and their precursors in osteoporotic bone tissue is mediated by the downregulation of LAMP2A expression, we measured LAMP2A levels in osteoclast precursor cells from osteoporosis patients and healthy controls and in young and old osteoporotic mice via Western blotting. We examined LAMP2A and HSC70 expression levels in osteoclast precursor cells via Western blotting. The results suggested that LAMP2A expression was significantly downregulated in osteoporosis, whereas HSC70 expression was not significantly different (Fig. 1C–D). To further clarify differences in the expression of key CMA molecules in bone tissue osteoclasts and their precursors, we labelled mature osteoclasts using the marker CTSK. Immunofluorescence confirmed that LAMP2A expression was downregulated in osteoporotic osteoclasts, whereas HSC70 levels were not significantly different between the osteoporotic and control groups (Fig. 1E–F). Additionally, we labelled monocytic precursors with CD14 in humans and F4/80 in mice observed the same downregulation of LAMP2A expression in these cells (Fig. 1G–H). These results suggest that the downregulation of LAMP2A expression mediates changes in the osteoclastic differentiation capability of osteoclast precursor cells, the bone resorption function of mature osteoclasts, and the progression of osteoporosis.

Fig. 1.

Reduced levels of LAMP2A-Mediated CMA in osteoporosis. (A) Immunofluorescence image showing the expression of Lamp2a and Hsc70 in young and old C57BL/6 mice; yellow puncta (Lamp2a⁺ & Hsc70⁺) indicating CMA activity. Scale bar = 50 μm. (B) Immunofluorescence image showing the expression of LAMP2A and HSC70 in osteoporotic and nonosteoporotic patients; yellow puncta (Lamp2a⁺ & Hsc70⁺) indicate CMA activity. Scale bar = 50 μm. (C) Western blot showing the expression of Lamp2a and Hsc70 in the bone marrow monocytes of young and aged C57BL/6 mice. (D) Western blot showing the expression of LAMP2A and HSC70 in the peripheral blood monocytes of osteoporotic and nonosteoporotic patients. (E) Immunofluorescence analysis of Lamp2a expression in Ctsk⁺ cells from young and old C57BL/6 mice. Scale bar = 50 μm. (F) Immunofluorescence showing the expression of LAMP2A in CTSK + cells from osteoporotic and nonosteoporotic patients. Scale bar = 50 μm. (G) Immunofluorescence showing the expression of Lamp2a in F4/80⁺ cells from young and aged C57BL/6 mice. Scale bar = 50 μm. (H) Immunofluorescence showing the expression of LAMP2A in CD14⁺ cells from osteoporotic patients and non-osteoporotic patients. Scale bar = 50 μm. All data are presented as the means ± SD, n = 6 per group. Statistical differences were determined using Student's t-test. ns = not statistically significant, ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

3.2. Low expression of LAMP2A enhances osteoclast differentiation and bone resorption by attenuating CMA activity

We subsequently explored the role of LAMP2A in osteoclast precursor cell differentiation and mature osteoclast bone resorption capacity in vitro. Osteoclasts differentiate into three distinct phases (Fig. 2A): monocytic precursors (MPs), pOCs, and mOCs. The immunofluorescence staining results suggested that CMA activity and the intracellular LAMP2A level decreased during osteoclast differentiation (Fig. 2B). Quantification of the protein levels of LAMP2A and HSC70 during osteoclast differentiation (Fig. 2C) indicated that the attenuation of CMA activity was mediated by the downregulation of LAMP2A expression. These results indicate that LAMP2A-mediated CMA activity is negatively correlated with osteoclast function. To further investigate the effect of LAMP2A-mediated CMA on osteoclast differentiation and bone resorption, we constructed a knockdown siRNA and overexpression lentivirus for LAMP2A and modulated the intracellular levels of LAMP2A through transfection. We examined the expression of proteins related to osteoclast differentiation, adhesion, and bone resorption (CTSK, NFATC1, and c-FOS) using Western blotting. The results revealed significant downregulated expression of these markers in the overexpression group compared with those in the control group, whereas the expression of these markers was significantly upregulated in the knockdown group (Fig. 2D). Osteoclast differentiation was assessed via immunofluorescence and TRAP staining. Immunofluorescence and TRAP staining revealed a significant reduction in the number and proportion of osteoclasts in the LAMP2A-overexpressing group, whereas the number of fused multinucleated giant cells decreased in the LAMP2A-overexpressing group and significantly increased in the LAMP2A-knockdown group (Fig. 2E–F). Additionally, we assessed changes in the bone resorption ability of mature osteoclasts. We observed that the bone resorption ability of the overexpression group was weaker than that of the control group, whereas the resorption ability of the knockdown group was greater than that of the control group (Fig. 2G). These results suggest that LAMP2A-mediated CMA negatively regulates osteoclast precursor cell differentiation, mature osteoclast adhesion, and bone resorption.

Fig. 2.

Low expression of LAMP2A enhances osteoclast differentiation and bone resorption by attenuating CMA activity. (A) Schematic diagram of osteoclast differentiation at different stages. (B) Immunofluorescence staining shows the expression and distribution of LAMP2A and HSC70 during osteoclast differentiation at different stages; LAMP2A⁺ & HSC70⁺ indicates CMA activity. Scale bar = 25 μm. (C) Western blot showing the protein levels of LAMP2A, HSC70 and osteoclast differentiation markers during different stages of osteoclast differentiation. (D) Western blot showing the protein levels of LAMP2A, HSC70 and osteoclast differentiation markers after LAMP2A knockdown and overexpression. (E) Immunofluorescence showing the morphology of multinucleated osteoclasts after LAMP2A knockdown or overexpression. Scale bar = 100 μm. (F) TRAP staining showing the number of TRAP⁺ osteoclasts after LAMP2A knockdown or overexpression. Scale bar = 100 μm. (G) Scanning electron microscopy showing bone resorption of osteoclasts after LAMP2A knockdown or overexpression. Scale bar = 20 μm. All the data are presented as means ± SDs, n = 6 per group. Statistical differences were determined using Student's t-test or ANOVA. ns = not statistically significant, ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

3.3. TFE3 cytoplasmic translocation mediates the downregulation of LAMP2A expression in osteoporosis

To further investigate the specific mechanism of the downregulation of LAMP2A expression in osteoclast precursor cells in osteoporosis, we constructed a DNA probe for LAMP2, enriched the binding protein of LAMP2 DNA in osteoclast precursor cells via a DNA pull-down assay, and detected and analysed the pull-down product via LC‒MS/MS. We identified 242 proteins that are associated with and bound to LAMP2A from the LC-MS/MS results of the DNA pull-down assay. Since the objective of our study was to identify the upstream regulatory proteins of LAMP2A, we performed GO enrichment analysis on these proteins, which revealed that 10 were associated with transcriptional regulation. Based on existing research, we further examined these 10 proteins and discovered that TFE3 has been identified in relevant studies [22] as being closely involved in the transcriptional regulation of the lysosome-associated protein. (Fig. 3A–C). We subsequently constructed siRNAs for TFE3 and knocked down TFE3 in cells; qPCR and Western blot results revealed that the expression of LAMP2A was also downregulated (Fig. 3D–E). These findings confirmed that TFE3 is an important regulator of LAMP2A expression in osteoclast precursor cells. To explore the association between TFE3 levels and osteoporosis, we examined TFE3 levels in the osteoporosis and nonosteoporosis groups using Western blot and immunofluorescence experiments. The results revealed that TFE3 levels were indistinguishable between the osteoporosis and control groups (Fig. 3F–G). Notably, as a member of the MiT/TFE transcription factor family, TFE3 tends to play a regulatory role in gene expression through cytoplasmic/nuclear spatial translocation [23]. Therefore, in our subsequent experiment, we focused on the distribution ratio of TFE3 in the cytoplasm and nucleus. Immunofluorescence staining revealed that the colocalization of TFE3 with the nucleus gradually decreased during osteoclast differentiation (Fig. 3H). We subsequently isolated osteoclast precursor cells from the osteoporosis and control groups, extracted protein after nuclear/cytoplasmic separation, and assessed the nuclear and cytoplasmic TFE3 levels via Western blotting. The distribution ratio of TFE3 in the nucleus of osteoporotic osteoclast precursor cells was lower than that in the control group, and the distribution ratio of cytoplasmic TFE3 increased significantly compared with that in the control group (Fig. 3I). In addition, immunofluorescence staining revealed that TFE3 knockdown inhibited LAMP2A-mediated CMA activity. These results suggest that the cytoplasmic translocation of TFE3 mediates the downregulation of LAMP2A expression in osteoporotic osteoclasts.

Fig. 3.

TFE3 cytoplasmic translocation mediates the downregulation of LAMP2A expression in osteoporosis. (A) Venn diagram showing the number of protein differences detected by LC‒MS/MS after DNA pull-down of LAMP2A and control DNA probes. (B) GO analysis of LAMP2A-binding proteins. (C) GO analysis showing the molecular function-related terms associated with LAMP2A-binding proteins. (D) qRT‒PCR showing the mRNA levels of LAMP2A and TFE3 after TFE3 knockdown and overexpression. (E) Western blot showing the protein levels of LAMP2A and TFE3 after TFE3 knockdown and overexpression. (F) Western blot showing the protein levels of TFE3 in osteoclast precursor cells from osteoporosis patients and nonosteoporosis patients, young and old C57BL/6 mice. (G) Immunofluorescence showing TFE3 expression in CTSK + cells from bone tissue samples from osteoporotic patients and nonosteoporosis patients, young and old C57BL/6 mice. Scale bar = 50 μm. (H) Immunofluorescence showing the distribution of TFE3 during osteoclast differentiation at different stages. Scale bar = 10 μm. (I) Western blot showing cytoplasmic, nuclear and overall TFE3 protein levels in osteoclast precursor cells from osteoporotic patients and nonosteoporosis patients, young and old C57BL/6 mice. (J) Immunofluorescence staining showing the expression and distribution of LAMP2A and HSC70 after TFE3 knockdown; LAMP2A⁺ & HSC70⁺ indicates CMA activity. Scale bar = 20 μm. All the data are presented as means ± SDs, n = 6 per group. Statistical differences were determined using Student's t-test or ANOVA. ns = not statistically significant, ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

3.4. LAMP2A-mediated CMA regulates osteoclasts and osteoclast precursor functions via CCR5

LAMP2A-mediated CMA is an important intracellular protein regulatory system. After LAMP2A was knocked down in osteoclast precursor cells, we used proteomics to identify proteins whose expression significantly differed from that in the control group (Fig. 4A–B). Since LAMP2A-mediated CMA activities target proteins, we combined proteomics and transcriptome sequencing for a joint analysis and found that most of the significantly different proteins did not coincide with significantly different RNAs (Fig. 4C–G). We identified CCR5 as a downstream target protein that was not significantly different at the RNA level but presented the greatest content variation at the protein level. The effect of CCR5 on osteoclast differentiation was investigated in a preliminary study [24]. We further verified the sequencing analysis results by Western blot and qPCR, which revealed that the suppression of CMA activity by LAMP2A knockdown had no significant effect on CCR5 RNA expression (Fig. 4H), whereas CCR5 protein levels were significantly increased (Fig. 4I). To further investigate whether CCR5 is an important downstream molecule of LAMP2A that regulates osteoclast bone resorption and precursor differentiation, we overexpressed CCR5 while knocking down LAMP2A and found that the differentiation of osteoclast precursor cells and the bone resorption ability of mature osteoclasts were both significantly regulated (Fig. 4J–K). Taken together, these results suggest that CCR5 is a key downstream molecule in the LAMP2A-mediated inhibition of autophagic activity, which regulates functional changes in osteoclasts and their precursors.

Fig. 4.

LAMP2A-Mediated CMA regulates osteoclasts and osteoclast precursor functions via CCR5. (A) Heatmap showing differentially expressed proteins between the si-LAMP2A group and the si-NC group. (B) Volcano plot showing differentially expressed proteins between the si-LAMP2A group and the si-NC group. (C) Heatmap showing differentially expressed genes between the si-LAMP2A group and the si-NC group. (D) Volcano plot showing genes that were differentially expressed between the si-LAMP2A group and the si-NC group. (E) Heatmap of the correlation analysis between transcriptomics and proteomics data. (F) Venn diagram of the correlation analysis between transcriptomics and proteomics data. (G) Correlation analysis between the transcriptomic and proteomic data. (H) qRT‒PCR showing the mRNA levels of LAMP2A and CCR5 after LAMP2A knockdown and overexpression. (I) Western blot showing the protein levels of LAMP2A and CCR5 after LAMP2A knockdown and overexpression. (J) Western blot showing the protein levels of osteoclast differentiation markers after CCR5 and LAMP2A knockdown with siRNA. (K) TRAP staining showing changes in the number of TRAP⁺ osteoclasts after CCR5 and LAMP2A knockdown with siRNA. Scale bar = 100 μm. All the data are presented as means ± SDs, n = 6 per group. Statistical differences were determined using Student's t-test or ANOVA. ns = not statistically significant, ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

3.5. CCR5 is a key substrate of LAMP2A-mediated CMA activity in osteoclasts and osteoclast precursor

CCR5 is an important downstream protein that regulates the function of osteoclasts and their precursors, but it remains unclear whether the elevated CCR5 protein level after the downregulation of LAMP2A expression is caused by CMA activity. CCR5 levels significantly increased after the addition of the lysosomal activity inhibitors Chloroquine (CQ) and Leupeptin (Leu), an effect not observed with proteasome inhibitors MG132 (Fig. 5A), suggesting that the regulation of intracellular CCR5 levels may be related to the lysosomal pathway. Therefore, we assessed the effects of the macroautophagy activation inhibitor 3-methyladenine (3-MA) and the CMA-specific activators AR7 and QX77. AR7 and QX77 activated CMA through the specific upregulation of LAMP2A expression (Fig. S3). The experimental results revealed that CCR5 levels were significantly altered only when the intervention was specific for CMA activity (Fig. 5B). To further validate the role of CCR5 as a substrate in CMA activity, we used Leu along with LAMP2A knockdown and observed changes in CCR5 protein levels in the lysosomal substrate pool via Western blotting [25,26]. These results suggest that CCR5 is a lysosomal substrate for molecular chaperone autophagy (Fig. 5C). During the CMA degradation process, the chaperone protein HSC70, lysosomal membrane protein LAMP2A, and substrate protein CCR5 interact. To further explore their roles and interactions, we extracted endogenous proteins from osteoclast precursor cells and confirmed the binding relationship between HSC70 and the substrate CCR5 by Co-IP (Fig. 5D). Immunofluorescence staining during different periods of osteoclast differentiation revealed that HSC70, LAMP2A, and CCR5 interacted at various stages of differentiation (Fig. 5E). Additionally, we constructed full-length plasmids of HSC70, CCR5, and LAMP2A, transfected them into 293T cells, extracted the proteins, and assessed the nonendogenous binding relationships among the three proteins using a Co-IP assay (Fig. 5F). The results indicated that CCR5 has a reciprocal relationship with both the chaperone protein HSC70 and the lysosomal membrane protein LAMP2A. Using the KFERQ motif finder prediction website [27], we identified the amino acid sequence ¹⁹⁴QTLKI¹⁹⁷ of CCR5 as a potential binding site for the molecular chaperone (Fig. 5G). Molecular docking revealed that CCR5 and HSC70 may bind near this site (Fig. 5H). We subsequently designed a mutant plasmid to alter the amino acid sequence to ¹⁹⁴AALKI¹⁹⁷(Fig. 5I). Co-IP experiments revealed that CCR5 with the mutated KFERQ motif could not bind normally to the chaperone protein HSC70 (Fig. 5J). Immunofluorescence revealed that the colocalization of the mutated CCR5 with HSC70 was altered compared with that of the wild-type CCR5 protein (Fig. 5K). Taken together, these results confirm that CCR5 is a key substrate in the CMA activity of osteoclasts and their precursors and that the protein level of CCR5 is regulated by CMA activity and dependent on specific KFERQ sequences.

Fig. 5.

CCR5 is a key substrate of LAMP2A-Mediated CMA activity in osteoclasts and osteoclast precursor. (A) Western blot showing the protein levels of CCR5 after treatment with the proteasome inhibitor MG-132 (1 μM) as well as the lysosomal inhibitors CQ (25 μM) and Leu (100 μM). (B) Western blot showing the protein levels of CCR5 after treatment with the macroautophagy inhibitor 3-MA (500 μM) and the CMA activators AR7 (10 μM) and QX77 (10 μM). (C) Western blot showing the protein levels of lysosomes isolated from osteoclast precursor cells treated with Leu (100 μM). (D) Immunoprecipitation assay showing the binding relationship between HSC70 and CCR5 in osteoclast precursor cells. (E) Immunofluorescence showing the colocalization relationship between LAMP2A and CCR5 and between HSC70 and CCR5 during osteoclast differentiation at different stages. Scale bar = 10 μm. (F) Immunoprecipitation assay showing the binding relationship between HA-CCR5 and Flag-HSC70 in 293T cells. (G) CCR5 KFERQ-like motif types and amino acid sequences were identified via the KFERQ finder. (H) Molecular docking of 3D structures was used to predict the binding of HSC70 with CCR5. (I) Schematic diagram showing the mutation site of the CCR5 mutant plasmid and its nucleic acid and amino acid sequences. (J) Immunoprecipitation assay showing the binding relationship between Flag-HSC70 and HA-CCR5 WT and between Flag-HSC70 and HA-CCR5 Mut in 293T cells. (K) Immunofluorescence showing the colocalization relationship between Flag-HSC70 and HA-CCR5 WT and between Flag-HSC70 and HA-CCR5 Mut in 293T cells. Scale bar = 10 μm. All the data are presented as means ± SDs, n = 6 per group. Statistical differences were determined using Student's t-test or ANOVA. ns = not statistically significant, ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

3.6. Preparation and characterization of the osteoclast bone resorption interface-targeted delivery system (D-Asp)8-PEG-PCL

To further clarify the role of osteoclast CMA activity in bone mass and bone resorption in vivo and minimize intervention effects on other bone microenvironment-associated cells, we constructed nanoparticles with a small-molecule peptide, D-Asp8, which had an osteoclast-targeted delivery function to the bone resorption interface [28,29]. These peptides were incorporated on the surface of PCL-PEG nanoparticles, and the nanoparticles were loaded with the required activator (Fig. 6A). Using transmission electron microscopy, we directly observed the morphology of the nanoparticles with and without D-Asp8, both before and after drug loading (Fig. 6B). Zeta potential and particle size analyses revealed the particle size of the nanoparticles (Fig. 6C). The cumulative release rate and Zeta potential size analyses of the nanoparticles was showed (Fig. 6D–E). Immunofluorescence revealed that the colocalization of Cy3 with osteoclast markers was greater in the (D-Asp)8-PEG-PCL group than in the PEG-PCL group, confirming that nanoparticles decorated with D-Asp8 have a strong osteoclast-targeting effect (Fig. 6F). To assess the targeting of nanoparticles after binding D-Asp8, we injected nanoparticles carrying Cy5.5 into C57BL/6J mice via the tail vein. In vivo imaging revealed that the Cy5.5 content in the (D-Asp)8-PEG-PCL group was significantly greater than that in the PEG-PCL group in bone tissue. Additionally, the clearance level via the hepatic route was significantly lower in the (D-Asp)8-PEG-PCL group than in the PEG-PCL group (Fig. 6G). Bone tissue samples from the mice were obtained after nanoparticle injection. We subsequently tested the cytotoxicity of the nanoparticles in vitro and in vivo. The CCK8 assay results revealed that the synthesized materials were not significantly toxic to the cells (Fig. 6H). HE staining of heart, liver, spleen, lung, and kidney tissue sections from C57BL/6J mice revealed no significant toxicity to major organs from each type of nanoparticle (Fig. 6I). These results indicate that (D-Asp)8-PEG-PCL nanoparticles are safe and reliable for use in animal and cellular experiments. They have a strong ability to target osteoclasts and thus can play a significant role in life science and clinical applications.

Fig. 6.

Preparation and characterization of the osteoclast bone resorption interface-targeted delivery system (D-Asp)8-PEG-PCL. (A) Schematic diagram showing the preparation and loading of (D-Asp)8-PEG-PCL. (B) Scanning electron microscopy showing the morphological characteristics of the nanoparticles. Scale bar = 100 nm. (C) Dynamic light scattering of the nanoparticles. (D) Profile showing release rate in vitro. (E) Zeta potential of the nanomaterials. (F) Immunofluorescence showing the distribution of nanoparticles in Ctsk + cells from mouse bone tissue samples. Scale bar = 50 μm. (G) Fluorescence image showing the biodistribution of the nanoparticles in the heart, liver, spleen, kidneys and hind limb after tail vein injection. (H) CCK8 assay showing the effect of nanoparticles on cell viability. (I) HE staining of heart, liver, spleen and kidney sections showing the effects of nanoparticles after tail vein injection into C57BL/6J mice for 48 h. Scale bar = 100 μm. All the data are presented as the means ± SDs, n = 3 per group in (D–F), n = 6 per group in (H). Statistical differences were determined using Student's t-test or ANOVA. ns = not statistically significant, ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

3.7. Activation of osteoclast LAMP2A-mediated CMA activity in vivo inhibits bone resorption activity and reduces bone loss

The differentiation of osteoclast precursors and the bone resorption capacity of osteoclasts play crucial roles in osteoporotic bone loss. In vitro experiments demonstrated that the downregulation of LAMP2A expression inhibits CMA activity, promotes the differentiation of osteoclast precursors and enhances the bone resorption capacity of mature osteoclasts. However, the role of CMA activity in regulating osteoclasts in vivo remains unclear. Therefore, we constructed mouse models of acute and chronic bone loss and femoral defects. We injected osteoclast-targeted nanoparticles, (D-Asp)8-PEG-PCL@AR7 and (D-Asp)8-PEG-PCL@QX77, containing CMA activators via the tail vein and subsequently collected femur and skull samples for further examination (Fig. 7A). Micro-CT 3D reconstruction of the skulls revealed that the injection of (D-Asp)8-PEG-PCL@AR7 and (D-Asp)8-PEG-PCL@QX77 reduced the distance and width of the cranial suture in a mouse model of LPS-induced acute bone loss (Fig. 7B). TRAP staining revealed that the (D-Asp)8-PEG-PCL@AR7 and (D-Asp)8-PEG-PCL@QX77 groups presented a lower TRAP positivity rate than did the LPS-induced acute bone loss control group (Fig. 7C). These results indicated that CMA activation effectively inhibited bone destruction by osteoclasts and promoted bone regeneration in the context of bone reconstruction homeostasis. To further test the effects of CMA activation on the osteogenic‒osteoblastic relationship and bone volume, we analysed the femurs of OVX-induced osteoporosis model mice. TRAP staining of femur bone tissue sections revealed a lower rate of TRAP positivity in the bone trabeculae of OVX osteoporosis model mice than in the OVX control mice. HE and Masson staining of the femur sections revealed a significant increase in the number of bone trabeculae in the (D-Asp)8-PEG-PCL@AR7 and (D-Asp)8-PEG-PCL@QX77 groups than in the OVX control group (Fig. 7D). Micro-CT scanning and bone trabeculae analysis revealed that the injection of (D-Asp)8-PEG-PCL@AR7 and (D-Asp)8-PEG-PCL@QX77 effectively restored the bone mass of OVX mice (Fig. 7E–F). To gain a deeper understanding of the effect of CMA activity on the balance between bone destruction and bone regeneration, we constructed a bone defect model in mice and applied interventions accordingly. The CT results revealed that the bone regeneration area in the (D-Asp)8-PEG-PCL@AR7 and (D-Asp)8-PEG-PCL@QX77 groups was significantly greater than that in the control group (Fig. 7G). These results indicate that CMA inhibits osteoclast differentiation and osteoclast bone resorption in vivo, positively regulating bone mass. The activation of osteoclast CMA activity can effectively inhibit bone destruction in osteoporosis and delay its progression.

Fig. 7.

Activation of osteoclast LAMP2A-Mediated CMA activity in vivo inhibits bone resorption activity and reduces bone loss. (A) Schematic diagram of the mouse model construction and experimental intervention. (B) Micro-CT 3D reconstruction showing the effects of nanoparticles on mouse calvarial defects at 1 week after PBS or LPS injection. (C) TRAP staining showing the number of osteoclasts in mouse calvaria treated with nanoparticles. (D) TRAP, HE and Masson staining results showing the number of osteoclasts and trabeculae in the femurs of sham and ovariectomized mice treated with nanoparticles. Scale bar = 100 μm. (E) Micro-CT image and 3D reconstruction showing the femurs of sham and ovariectomized mice treated with nanoparticles. (F) Quantitative analysis of bone microstructural parameters in the femurs of sham and ovariectomized mice treated with nanoparticles. (G) 3D reconstruction of micro-CT showing femoral defects in C57BL/6J mice treated with nanoparticles. All the data are presented as means ± SDs, n = 6 per group. Statistical differences were determined using Student's t-test or ANOVA. ns = not statistically significant, ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

4. Discussion

Bone homeostasis is dependent primarily on the balance between bone formation and resorption, with osteoblasts and osteoclasts playing pivotal roles in regulating these processes. The dysregulation of osteogenic and osteoclastic activities disrupts bone homeostasis, leading to a cascade of bone-related diseases [10]. In osteoporosis, the overactivation of osteoclast differentiation from pOCs leads to a significant increase in the number of osteoclasts at the bone resorption interface and a significant increase in the bone resorption capacity of mOCs, leading to sustained bone loss and microstructural deterioration [3]. However, the precise molecular mechanisms governing the aberrant function of osteoclasts and their precursors in osteoporosis remain unclear.

Autophagy is categorized into three types in current research: macroautophagy, microautophagy, and chaperone-mediated autophagy. These variants operate around the lysosome, a central organelle responsible for degradation. The key autophagy molecule P62 and TSC1 regulate osteoclast differentiation by modulating the classical macroautophagy in osteoclasts, indicating that macroautophagy in the autophagy pathway plays a crucial role in osteoclast differentiation [30,31]. Owing to its mechanism, CMA, wherein the molecular chaperone HSC70 identifies protein-specific KFERQ motifs and delivers substrates to lysosomes for degradation, is highly selective compared with other autophagy pathways. Therefore, the molecular mechanisms by which CMA and traditional autophagy regulate osteoclast differentiation may be significantly different, and further research is needed. Early studies closely linked autophagy to the ageing of an organism, and autophagy is a key regulator of lifespan [32]. The autophagic activity of an organism gradually diminishes with age, and diminished autophagic activity leads to the development of age-related diseases, including neurodegenerative diseases and osteoarthritis [33,34]. Similarly, age-dependent declines in CMA activity are observed in various tissues and organs of both humans and mice [[35], [36], [37], [38]]. In both aged humans and elderly C57BL/6 mice, CMA activity is markedly reduced in MSCs and osteoblasts of the femur [21]. Osteoblasts and osteoclasts are pivotal regulators of bone mass and structure within the bone microenvironment. Our study validated the suppression of CMA activity in osteoclasts and their precursors in both aged osteoporotic mice and osteoporotic patients, indicating a potential association between reduced CMA activity in bone tissue and age-related bone loss in osteoporosis. Notably, the expression levels of HSC70 and LAMP2A, pivotal molecules in CMA, dictate their activity levels. Previous research has demonstrated that decreased levels of the lysosomal membrane protein LAMP2A frequently underlie diminished CMA activity [39], which is often attributed to age-related alterations in lysosomal function [[39], [40], [41]]. Changes in LAMP2A levels mediate the variation in CMA activity observed in osteoclasts and their precursors between aged individuals and aged model mice [21]. Our findings indicate that reductions in LAMP2A contribute to the changes in CMA activity observed in osteoclasts and their precursors. Thus, we propose that LAMP2A regulates CMA activity in the bone microenvironment of elderly osteoporosis patients, potentially correlating with the development of osteoporosis.

Neurodegenerative diseases such as Parkinson's disease, Alzheimer's disease, frontotemporal lobar degeneration, and amyotrophic lateral sclerosis are closely related to CMA [[42], [43], [44], [45], [46]]. Osteoporosis, an age-related skeletal degenerative disease, may also contribute to the pathogenesis of LAMP2A-mediated attenuation of CMA activity. While attenuated CMA levels are frequently noted in age-related diseases, the precise reasons for this attenuation and its specific role in osteoporosis remain poorly understood. Although the molecular mechanisms driving the decline in CMA activity with age are uncertain, it is evident that the age-related decrease in LAMP2A expression significantly impacts CMA regulation [47]. Our findings indicate that reduced LAMP2A expression contributes to the decline in CMA activity in osteoporosis, with downregulated LAMP2A expression during osteoclast formation being regulated by TFE3. TFE3, belonging to the MiTF/TFE family of transcription factors, binds to CLEAR elements, promoting lysosomal biosynthesis and autophagy upon activation [22]. Our study revealed that the expression of LAMP2A in osteoclasts and their precursors is regulated by the transcription factor TFE3. These findings suggest that, despite its specificity, CMA relies on lysosomes as degradation hubs, which are regulated by lysosomal regulatory factors. Furthermore, our study revealed a significantly greater cytoplasmic distribution of TFE3 in aged individuals with osteoporosis than in control individuals, further supporting the pathogenic role of TFE3-regulated autophagic activity in age-related degenerative diseases. Notably, TFE3 transcription factors also regulate other autophagy pathways, including macroautophagy [48,49]. Thus, TFE3 regulation may be common to various autophagy pathways involving lysosomal degradation, implying that the decreased activity of CMA and other autophagy pathways with age may involve similar regulatory mechanisms.

Previous studies have demonstrated that CMA activity promotes bone formation, supporting the hypothesis that reduced CMA activity contributes to a sustained decrease in bone mass in osteoporosis. In a study by Akel et al., [50] Lamp2a knockout (KO) mice presented lower bone mass than did wild-type mice, indicating an inhibitory effect of LAMP2A knockout on the differentiation of osteoclasts. Currently, there is limited research on the impact of CMA activity on the function of osteoclasts and their precursors. However, in our study utilizing LAMP2A knockdown and overexpression, we demonstrated that reduced CMA activity promotes osteoclast formation and enhances the osteolytic function of osteoclasts. Consistent with previous studies highlighting the positive regulatory effects of CMA on bone mass during osteogenesis, our findings suggest that the inhibition of osteoclast formation and its osteolytic function by CMA is associated with increased bone mass. Physiological CMA activity maintains the balance between osteogenesis and osteolysis to preserve bone homeostasis. Reduced CMA activity disrupts this balance, favouring osteoclastogenesis and resulting in persistent bone loss, ultimately leading to osteoporosis development.

CMA functions primarily at the protein level within the cell. Through proteomic and transcriptomic analysis, we identified CCR5 as a key downstream regulatory protein of CMA in osteoclast precursor cells. The KFERQ-like motif of the substrate protein CCR5 was identified using the KFERQ motif analysis platform (KFERQ finder) developed by Kirchner et al., [27] and we confirmed that HSC70 mediates CMA degradation by binding to the ¹⁹⁴QTLKI¹⁹⁸ motif of CCR5. CCR5 serves as a crucial cytokine receptor. CCR5 deficiency inhibits osteoclast formation and impairs cell adhesion, migration, and bone resorption by mature osteoclasts [24,[51], [52], [53]]. Consequently, in osteoporosis, reduced CMA activity leads to impaired CCR5 degradation by the cytokine receptor. This results in increased and overactivated levels of the CCR5 receptor, promoting osteoclast formation and enhancing osteolytic activity, thereby exacerbating the imbalance between osteogenesis and osteolysis in the body.

In Lamp2a and Lamp2c KO mice, both young and adult mice present decreased bone mass, indicating that CMA positively regulates bone mass in vivo [50]. However, whole-cell knockout in mice does not reflect the changes in bone mass mediated by the changes in CMA levels regulated by LAMP2A in osteoclasts, which may affect bone mass via other cells in the bone microenvironment, including osteoblasts. Therefore, in our study, we constructed an osteoclast-targeting nanomaterial with the D-ASP8 peptide, whose osteoclast-targeting effects have been validated in previous studies. In an animal model with abnormally activated osteoclasts that target osteoclasts with CMA activators, increasing CMA activity in osteoclasts in vivo attenuated the bone loss caused by abnormal osteoclast activation in osteoporosis.

Therefore, Lamp2A-mediated inhibition of CMA activity is an important molecular biological mechanism for the transformation of osteogenic and osteoclastic equilibrium to abnormal osteoclastic function in osteoporosis and is an important cause of bone loss in osteoporosis (Fig. 8). Therefore, the regulation of CMA activity via LAMP2A is expected to be an important therapeutic approach for alleviating bone loss in osteoporosis.

Fig. 8.

Schematic representing the role of LAMP2A-mediated CMA activity attenuation in osteoclasts and precursors in osteoporosis. TFE3 cytoplasmic translocation inhibited LAMP2A expression, thereby increasing CCR5 levels by inhibiting osteoclast CMA-dependent degradation activity, promoting osteoclastic differentiation of osteoclast precursor cells, and increasing the bone resorption capacity of mature osteoclasts. Finally, this led to an imbalance between osteogenesis and bone resorption, ultimately causing osteoporosis.

CRediT authorship contribution statement

Yunhui Zhang: Writing – review & editing, Writing – original draft. Quanfeng Li: Writing – original draft, Conceptualization. Xiaoshuai Peng: Writing – original draft, Conceptualization. Pengfei Ji: Formal analysis, Data curation. Yibin Zhang: Investigation, Formal analysis. Jiahao Jin: Methodology, Investigation. Zihao Yuan: Project administration, Methodology. Jianan Jiang: Resources, Project administration. Guangqi Tian: Software, Resources. Mingxi Cai: Supervision, Software. Pei Feng: Validation, Supervision. Yanfeng Wu: Writing – review & editing, Writing – original draft. Wenjie Liu: Writing – review & editing, Writing – original draft. Peng Wang: Writing – review & editing, Writing – original draft, Conceptualization.

Ethics approval and consent to participate

The animal experiments were approved by the Animal Ethics Committee of Sun Yat-sen University (approval No. 2023d048). The clinical study was approved by the Ethics Committee of The Eighth Affiliated Hospital, Sun Yat-sen University (Clinical ethical approval No. 2023r032).

Funding

This study was supported by the National Natural Science Foundation of China (Grant No. 82372372, 82172349, 82302661), the Guangdong Provincial Clinical Research Center for Orthopedic Diseases (Grant No. 2023B110001), the Guangdong Natural Science Foundation (Grant No. 2023A1515010568), the Shenzhen Science and Technology Program (Grant No. KCXFZ20230731092959001, 202205303001458), the Futian Healthcare Research Project (Grant No. FTWS2022022, FTWS2023072) and the Excellent Medical Innovation Talent Program of the Eighth Affiliated Hospital of Sun Yat-sen University (Grant No. YXYXCXRC202101).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Data availability

Summary data are available from the paper and supplementary materials. The raw data sets generated and analysed during this study are available from the corresponding author.