Targeting APJ drives BNIP3-PINK1-PARKIN induced mitophagy and improves systemic inflammatory bone loss

Graphical abstract

Highlights

• •APJ is associated with M1 polarization of macrophages in inflammatory bone loss.
• •Activation of APJ can inhibit M1 polarization of macrophages.
• •APJ alleviates inflammatory bone loss by inhibiting osteoclast activation.
• •APJ mainly regulates mitophagy through AMPK/BNIP3 pathway, reducing ROS accumulation and NLRP3 inflammasome activation.

Abstract

Introduction

Inflammatory diseases, such as diabetes mellitus, rheumatoid arthritis, and inflammatory bowel disease, lead to systemic immune microenvironment disturbances, contributing to bone loss, yet the mechanisms by which specific receptors regulate this process in inflammatory bone loss remain poorly understood. As a G-protein-coupled receptor, the Apelin receptor plays a crucial role in the regulation of inflammation and immune microenvironment. However, the precise mechanisms governing its role in inflammatory bone loss remain incompletely understood.

Objective

This study aims to investigate how APJ regulates macrophage polarization to mitigate inflammatory bone loss.

Methods

Lipopolysaccharide induced systemic inflammatory bone loss model in mice was used to explore the relationship between bone loss and osteoclast activation, macrophage polarization and APJ. In vitro studies, Bone marrow derived macrophages and siRNA were used to elucidate the regulatory influence of APJ on the immune microenvironment and osteoclast differentiation, while high-throughput sequencing is leveraged to uncover the underlying mechanisms through which APJ modulates macrophage polarization.

Results

Our study established a link between APJ and macrophage M1 polarization in systemic inflammatory bone loss mice. The activation of APJ effectively mitigated M1 polarization in macrophages, suppressed excessive osteoclast activation, and alleviated systemic inflammatory bone loss. In vitro high-throughput sequencing analysis revealed that APJ modulates macrophage polarization, linking to mitochondrial autophagy and the NOD-like receptor signaling pathway and the involvement of the AMPK and MAPK signaling pathways in signal transduction after APJ activation was also suggested. Subsequent experiments substantiated that APJ predominantly enhances mitophagy and diminishes the accumulation of reactive oxygen species by regulating the AMPK/BNIP3/PINK1/PARKIN axis, thereby suppressing the activation of macrophage M1 polarization and osteoclastogenesis.

Conclusion

This study elucidated the underlying mechanism by which APJ modulates macrophage polarization, thereby proposing a new therapeutic target for addressing inflammatory bone loss.

Introduction

The maintenance of healthy bone tissue involves a continuous process of dissolution and regeneration. The delicate equilibrium between bone resorption and formation is involved by various factors. Notably, inflammation has been found to be associated with bone loss closely. The presence of inflammatory factors triggers the activation of inflammatory cells, which in turn disrupts the bone immune microenvironment and facilitates the differentiation of osteoclasts, ultimately resulting in significant bone loss. [1] However, the bone loss resulting from inflammation is frequently overlooked and can ultimately result in disability. Currently, the clinical approach to managing systemic inflammation primarily involves symptomatic treatment, such as the administration of glucocorticoids for conditions like rheumatoid arthritis (RA), chronic obstructive pulmonary disease (COPD), and other systemic inflammatory disorders. Despite the notable anti-inflammatory properties of glucocorticoids, they not only fail to alleviate inflammatory bone loss but exacerbate it. [2], [3] Hence, the investigation of inflammatory bone loss treatment remains a significant field of research, necessitating the prompt identification of effective therapeutic targets and medications.

Inflammation activation serves to eliminate pathogens and facilitate tissue repair. Nevertheless, an overabundance of inflammatory activation leads to excessive immune cell activation, resulting in the release of numerous inflammatory factors. Consequently, this uncontrolled and excessive inflammation disrupts the autoimmune microenvironment. [4] Macrophages, as the primary immune cells in the human body, play a crucial role in phagocytosing foreign bodies and eliminating pathogens, thereby contributing significantly to human innate immunity. Upon exposure to various modes of stimulation, macrophages can differentiate into distinct phenotypes, specifically, classically activated M1 macrophages and selectively activated M2 macrophages. [5] Resting macrophages (M0) can be induced and activated into M1 macrophages by lipopolysaccharide (LPS) and IFN-γ, facilitating a robust immune response and contributing to the initiation and sustenance of sterilization and inflammation. M2 macrophages play pivotal roles in suppressing inflammation, promoting wound healing, fibrosis, tissue repair, vascular regeneration, as well as facilitating tumor growth and infiltration. Under normal circumstances, the polarization of M1 and M2 macrophages serves distinct functions during various stages of inflammation, encompassing the elimination of foreign pathogens and the restoration of damaged tissues. [6], [7], [8] However, the persistent inflammatory state resulting from macrophage M1 polarization leads to ongoing tissue cell damage and the release of numerous inflammatory factors. Several proinflammatory cytokines have been demonstrated to stimulate osteoclast activation and suppress osteoblast function. [9], [10], [11] Consequently, the regulation of immune microenvironment stability is crucial in addressing inflammatory bone loss. Previous researches have shown that mitophagy plays a crucial regulatory role in regulation of immune microenvironment, macrophage polarization and its mechanism in inflammatory bone loss needs further exploration.[12], [13].

The Apelin receptor (APJ), a G protein-coupled receptor located, is activated by its endogenous ligand Apelin (APLN). Apelin has many subtypes with different lengths, and Apelin-13 is the most active subtype. The APLN-APJ system is present in numerous tissues throughout the human body. Previous researches have investigated its involvement in the central nervous and cardiovascular systems. [14], [15] However, recent studies have revealed the crucial role of APJ in regulating inflammation. The activation of APJ has been shown to exert inhibitory effects on the release of inflammatory factors and macrophage M1 polarization during the acute phase of myocardial infarction. Furthermore, it has been observed to activate the AMPK-NLRP3 signaling pathway, leading to a reduction in neuroinflammation and oxidative stress, thereby providing relief from early brain injury. Additionally, APJ has been found to regulate mitophagy, thereby promoting the proliferation of vascular smooth muscle cells and mitigating oxidative stress injury. Moreover, it has been demonstrated to facilitate the proliferation of bone marrow-derived mesenchymal stem cells. [16], [17], [18] However, the extent to which APJ can exert regulatory influence on the progression of inflammatory bone loss uncertain yet.

In this study, we reported the key role of APJ in the treatment of inflammatory bone loss. We revealed that the activation of APJ enhanced BNIP3-PINK1-PARKIN-mediated mitophagy, diminished the accumulation of reactive oxygen species (ROS), and impeded the production of the NLRP3 inflammasome, ultimately inhibiting macrophage M1 polarization and mitigating inflammatory bone loss. The results proposed that the regulation of mitophagy by APJ can effectively suppress inflammatory bone loss, thereby presenting a novel avenue for the research and treatment of this condition. (Fig. 1).

Fig. 1.

Schematic illustration of the regulatory mechanism of APLN-APJ on macrophage polarization through mitophagy. (Created with BioRender.com).

Materials and methods

Drugs

Alpha modification of Eagle’s medium (α-MEM), and fetal bovine serum (FBS) were bought from VivaCell. Recombinant Mouse receptor activator of nuclear factor-κB ligand (RANKL) and Recombinant Mouse macrophage colony stimulating factor (M−CSF) were bought from R&D Systems. The primary antibodies, including INOS (A14031, 1:1000), CD86 (A1199, 1:1000), AMPK (A12718, 1:1000), p-AMPK (AP0116, 1:1000), P38 MAPK (A14401, 1:1000), p-P38 MAPK(AP0057, 1:1000), JNK (A4867, 1:1000), p-JNK (AP0631, 1:1000), ERK (A4782, 1:1000), p-ERK (AP0234, 1:1000), PINK1 (A7131, 1:1000), PARKIN (A0968, 1:1000), SQSTM1/P62 (A7758, 1:1000), LC3B (A19665, 1:1000), BAX (A0207, 1:1000), CASPASE3 (A2156, 1:1000), NLRP3 (A5652, 1:1000), IL-1β (A16288, 1:1000), ASC/TMS1 (A1170, 1:1000), MMP9 (A0289, 1:1000), ACTB (AC026, 1:50000) were bought from ABclonal and BCL2 (ab182858, 1:2000) were bought from Abcam, APJ (20341–1-AP, 1:1000) were purchased from Proteintech and HRP-labeled Goat Anti-Rabbit IgG(H + L)) (A0208, 1:1000) were bought from Beyotime.

Cell culture

Male C57BL/6 mice aged 6–8 weeks (20–25 g) obtained from Soochow university experimental animal center were killed after cervical dislocation. The femur and tibia of the mice were removed and rinsed with sterile PBS and 75 % ethanol in turn and aspirated with a syringe containing α- MEM with 30 ng.ml⁻¹ M−CSF to wash bone marrow several times. Then, put the medium containing cells after washing into the cell culture dish, put it into the cell culture incubator to stand for three days, replace half of the medium in the dish after three days, and the cells were allowed to stand for 2–3 days. After that, bone marrow derived macrophages (BMs) were firmly attached to the bottom of the cell culture dish, discarded the supernatant, and rinsed the cell culture dish with a pipette to remove other cells that did not adhere well. Then, trypsin (Beyotime, C0203) was used to digest BMs, which was collected for experiments. Cells from two mice in each group were extracted at one time.

Inflammatory bone loss model induced by LPS

The mouse experiments were granted permission from the Ethics Committee of Soochow University (SUDA20230625A07). The C57BL/6 male mice aged 8–10-weeks (n = 18) were divided into 3 groups randomly: a PBS treatment group (sham group), an LPS injection + PBS treatment group (LPS group) and an LPS injection + 1 mg.kg⁻¹ Apelin-13 treatment group (LPS + Apelin-13 group). In brief, mice in the sham group and LPS group received an intraperitoneal injection of 0.1 ml of PBS solution and 0.1 ml of PBS-diluted LPS (5 mg.kg-1) solution respectively for 7 consecutive days, and mice in the LPS + Apelin-13 group received an intraperitoneal injection of 0.1 ml of PBS-diluted LPS (5 mg.kg⁻¹) + Apelin-13 (100 µg.kg⁻¹) solution for 7 consecutive days. On the 8th day, all mice were killed due to cervical dislocation, and femurs were collected and half of the femurs was used to extract cells, and the other half was soaked in 4 % paraformaldehyde.

Tartrate resistant acid phosphatase (TRAP) staining

BMs (10 × 10⁴ cells/well) were implanted in 12 well plates and then 50 ng.ml⁻¹ RANKL was used to stimulate the BMs and place them statically in the cell incubator. Osteoclast differentiation was observed microscopically after 5 days, then the cells were stained with a TRAP staining kit (BZ OC02, BZ Biotechnology Co., Ltd). The results of staining were obtained by upright microscope (Zeiss, Germany). Osteoclasts generally have more than two nuclei.

Si-RNA transfection

BMs were implanted in 12 well plates. After the cells adhered to the wall, the cells were transfected with 20 nm si-RNA and GP-transform-mate (GenePharma) according to the instructions, and the transfection efficiency of si-APJ was verified by Western blot. SiAPJ-1: sense, 5- GCGCUCAUUCCUGCCAUCUACAUGU-3; antisense, 5- ACAUGUAGAUGGCAGGAAUGAGCGC; siAPJ-2: sense, 5- GACUUUGCCACUGUGGGCCACUUAU-3; and antisense, 5- AUAAGUGGCCCACAGUGGCAAAGUC-3; siAPJ-3: sense, 5- CCACUUAUACCUACCGGGAGUUUGA-3; and antisense, 5- UCAAACUCCCGGUAGGUAUAAGUGG-3; si-NC: sense, 5- GCGUACCCUCGUUACAUCACCUUGU-3; and antisense, 5- ACAAGGUGAUGUAACGAGGGUACGC-3.

F-actin staining

Briefly, fixed the BMs in the well plates with 4 % paraformaldehyde, then, staining the cytoskeleton and nuclei of BMs with phalloidin (yeasen, 40762ES75) and DAPI (Beyotime, C1005). Finally, images were achieved by inverted fluorescence microscope (Zeiss, Germany) and ImageJ software was used to quantify.

Flow cytometry

The cells after different treatment conditions were collected and made into single-cell suspension. Then specific primary antibodies were added. All antibodies (PE-CD86, APC-CD11b) were purchased from Biolegend Company (California, USA). Finally, after washing with PBS, then resuspended the BMs with PBS and analyzed by FACS caliber flow cytometer (BD, USA). FlowJo10 software (Tree Star, Inc., BD) was used for quantitative analysis.

Western blot assays

The total proteins of BMs were extracted with RIPA lysis buffer. Separated equal amounts of proteins using SDS-PAGE and transferred them onto nitrocellulose (NC) membranes. Then the NC membrane corresponding to the analytical amount was incubated with the primary antibody and the secondary antibody coupled with horseradish peroxidase. QuickChemi 5200 chemiluminescent imaging system (Monad) was used to calculate the grey level by BeyoECL Plus (Beyotime, P0018S), and then images were quantified by ImageJ software (Bethesda, USA).

RT–qPCR

The total cellular RNA of BMs in different groups was obtained by Trizol reagent (Vazyme, r401-01). After measuring the concentration of RNA, the same amount of RNA was reverse transcribed, mixed with dNTPs (Vazyme, r323) and primers according to the instructions, and then performed RT-qPCR in CFX96™ thermal cycler (Bio-Rad Laboratories). The primer sequences of the target genes are listed in Table S1.

High-throughput sequencing

RNA was collected and purified using Trizol Extraction Reagent (Invitrogen, #15596018CN). For RNA quality control, the presence of RNA degradation and DNA contamination was analyzed on agarose gel, and the purity of RNA was detected by Nanophotometer spectrophotometer (Implen). Subsequently, the TruseqTM RNA sample prep Kit (Illumina) was used to construct a strand specific library for transcriptome sequencing. Then, these libraries were sequenced on the Hiseq2000 Truseq SBS kit v3-HS (Illumina) sequencing platform. Based on the Illumina Novaseq platform with peer-to-peer reading mode, the NGS QC Toolkit was used to process the original data, delete the reads containing adapter and the reads containing ploy-N and low-quality reads, and obtain clean data. Hisat2 software was used to compare the clean data with the reference genome. After confirming that the comparison rate was higher than 85 %, follow-up analysis was carried out. Stringtie software is used to analyze the above comparison results, and calculate the FPKM value for each gene, which is used as the gene expression in the sample. The differentially expressed genes were identified by DESeq2 software (R Package, version 4.1.0). The p value was corrected by multiple hypothesis test using Benjamin and Hochberg methods. The corrected p value < 0.05, and the absolute log2 (difference multiple) ≥ 0.5 were used as the threshold of significant differential expression.

Transmission electron microscope (TEM)

BMs were digested using trypsin and collected in centrifuge tubes and fixed at 4 ℃ for 12 h by adding 2.5 % glutaraldehyde fixative. Subsequently, fixed all the samples with 1 % osmium tetroxide for 2 h and dehydrated sequentially with an ethanol gradient (30 %, 50 %, 70 %, 90 % and 100 %) for 20 min, Then all the samples were soaked in pure acetone for 20 min and embedded with resin, and sectioned on a Leica EM UC7 ultratome (Leica, Germany). Finally, uranyl acetate and alkaline lead citrate staining (5 min) and TEM observation (FEI Tecnai Spirit, USA) were performed.

ROS detection

Dihydroethidium (DHE) staining was used for fluorescence quantitative analysis. In short, after the BMs in different groups were treated, the fluorescence images were obtained by inverted microscope (Zeiss, Germany) based on maintaining the cell activity according to the method in the DHE kit instructions (Beyotime, S0063). DCF-DA (Yeasen, 50101ES01) was used for flow cytometry detection. After the BMs were treated according to the method in the manual, the detection was carried out in FITC channel according to the above flow cytometry steps.

Measurement of mitochondrial membrane potential

Mito-Tracker Deep Red FM (Beyotime, C1032) and Mitochondrial membrane potential assay kit with JC-1(Beyotime, C2006) were used to detect mitochondrial membrane potential (MMP). In brief, after the intervention, the BMs attached to the cell climbing sheet were stained according to the method of the kit instructions on the basis of maintaining the cell activity. After washing with PBS, the images were acquired under an inverted microscope (Zeiss, Germany). The red fluorescent probe of Mito-Tracker labels active mitochondria according to the MMP. JC-1 produces red fluorescence when the MMP is high; when the MMP is low, green fluorescence is generated.

Micro-CT analysis

After the mouse femurs were removed, they were fixed with paraformaldehyde for one week and used the micro-CT device (Skyscan 1176, Aartselaar, Belgium) to scan the femurs. Two-dimensional (2D) was reconstructed by SkyScan1176 software, and CTAn software was used to analyze bone parameters, including bone mineral density (BMD g/cm³), bone volume to bone volume fraction (BV/TV, %), trabecular number (Tb. N, 1/mm), trabecular thickness (Tb. Th, mm), bone surface/volume ratio (BS/BV, 1/mm) and trabecular pattern factor (Tb. Pf, 1/mm), and Mimics Medical 21.0 was used for three-dimensional reconstruction.

Statistics

All the data are shown as mean ± standard deviation (SD). All the statistical analysis was processed by GraphPad Prism v. 8.0. Student's t test was used to analyze the differences between two groups. For the differences between more than two groups, after one-way ANOVA was used to find out the significant differences, Tukey's post hoc test was used to compare the significant differences between multiple groups. The error bar line represents the standard deviation in the figures, and the probability values < 0.05 were defined as statistically significant.

Results

Macrophage M1 polarization is associated with APJ in vivo

We simulated systemic inflammation in mice to explore the relationship between APJ and macrophage polarization. The results of CT reconstruction and H&E staining (Fig. S1A–C) showed that the number of trabeculae in the distal femur of mice in LPS group was significantly reduced than in sham group, indicating the successful modelling of inflammatory bone loss in mice. Subsequently, we examined BMs from mice with systemic inflammation. The protein expression of INOS, a classical indicator of macrophage M1 polarization, exhibited a substantial increase in the BMs of mice induced by systemic inflammation (Fig. 2 A-C). The results indicated that the administration of LPS via intraperitoneal injection induced macrophage M1 polarization in mice. Additionally, it was unexpectedly observed that the expression of APJ increased in response to systemic inflammation. Previous researches have evidenced that the increased expression of APJ may be caused by the stimulation of hypoxic conditions, which is regulated by hypoxia inducible factor (HIF). [19], [20] The results of immunohistochemistry (IHC) analyses shown in Fig. 2 D-F also demonstrate a substantial increase in INOS-positive macrophages as well as the expression of APJ in mice with systemic inflammation. Then, we conducted immunofluorescence staining (IF) of APJ and INOS in tissue sections, as illustrated in Fig. 2 G and S2 A, the expression of APJ exhibited significant colocalization with INOS-positive M1 macrophages. Consequently, we postulated that macrophage M1 polarization may be associated with APJ in vivo.

Fig. 2.

APJ is associated with increased macrophage M1 polarization in systemic inflammation. (A-C) The results and the assessment of protein levels in BMs of the femur in mouse were quantified. (D) Representative images of IHC staining of INOS and APJ in the distal femur. Scale bars: 0.1 mm. (E-F) Quantitative analysis of APJ and INOS positive cells in the distal femur. (G) Representative images of IF in the distal femur; red (INOS), green (APJ), and blue (nuclei). Scale bars: 0.5 mm. (H) Representative images of IF; green (APJ), and blue (nuclei). Scale bars: 50 μm. (I) Quantitative analysis of the average fluorescence intensity of APJ. (J) The results of protein levels of APJ and INOS in BMs. (ns means no significance, in vitro n = 3; in vivo n = 6, * P < 0.05, ** P < 0.01).

Subsequently, we employed BMs, extracted from femurs of healthy mice, for in vitro cytological experiments to further investigate the correlation between APJ and macrophage polarization. The results presented in Fig. 2 H-I demonstrate a significant expression of APJ in BM, which was further enhanced following LPS intervention. The Western blot results (Fig. 2 J and S2B-C) provided additional validation, indicating that as macrophage M1 polarization increased, so did the expression of APJ, aligning with our in vivo findings. These results suggested that APJ is significantly correlated with macrophage M1 polarization in inflammatory bone loss, so we speculated that APJ may be involved in the regulation of macrophage M1 polarization.

The activation of APJ inhibits macrophage M1 polarization

We performed in vitro cell experiments using Apelin, an endogenous ligand of APJ receptor, to observe the effect of APJ activation on macrophage polarization, and a concentration of 10 nM of Apelin-13 was selected for further interference with BMs (Fig. S3A and B). The results obtained from Western blot and RT-qPCR analyses (Fig. 3 A-E) suggested a prominent increase in the expression of INOS and CD86 in the LPS + si-NC group than in the control group. Treatment with Apelin-13 attenuated the expression of INOS and CD86, then, we employed si-APJ to suppress the expression of APJ (Fig. S4A and B). The inhibitory effect of Apelin-13 was reversed in the APJ knockdown group. Additionally, flow cytometry was employed to determine the proportion of M1-polarized cells in macrophages (Fig. 3 F-G), with CD86 and CD11b serving as membrane surface marker proteins for M1 macrophages and BMs, respectively. The proportion of M1 macrophages increased significantly in LPS group compared with control group. Apelin-13 alleviated the M1 polarization of macrophages. However, after blocking APJ, the proportion of macrophage M1 polarization increased. The results obtained from IF (Fig. 3 H and S4 C) were consistent with those obtained from Western blot and flow cytometry analysis.

Fig. 3.

Macrophage M1 polarization in inflammatory state can be inhibited by the activation of APJ. (A-C) The results and the assessment of protein levels in BMs after LPS, Apelin-13 and si-APJ induction were quantified. (D-E) The mRNA expressions of Nos2 and Cd86 were analysed in BMs after LPS, Apelin-13 and si-APJ induction. (F-G) Flow cytometry analysis of CD86 and CD11b upon exposure to each group. (H) Representative images of IF; green (INOS), and blue (nuclei). Scale bars: 50 μm. (I) Representative images of IHC staining of INOS in the distal femur. Scale bars: 0.1 mm. (J-K) The results and the assessment of protein levels in BMs of the femur in mice were quantified. (in vitro n = 3; in vivo n = 6, * P < 0.05, ** P < 0.01).

Subsequently, Apelin-13 was administered to mice with LPS-induced systemic inflammatory bone loss to assess its in vivo effects. We observed the variations in macrophage polarization levels in mice subsequent to APJ activation. The results of IHC, and Western blot analysis of BMs (Fig. 3 I-K and S5 A) demonstrated a decrease in macrophage M1 polarization in the treatment group, suggesting that the activation of APJ can also inhibit macrophage M1 polarization in vivo. The results of the study suggested that the activation of APJ can effectively inhibit and reduce macrophage M1 polarization in vitro and in vivo.

The activation of APJ inhibits LPS induced osteoclasts activation and alleviates inflammatory bone loss

Previous studies have indicated that the activation of osteoclasts contributes significantly to bone loss resulting from disorders in the bone immune microenvironment. [21] Then, we extracted BMs from the femoral bone marrow cavity of mice to further explore the effect of activating APJ on immune microenvironment and osteoclast activation. We collected conditioned mediums (CMs) from groups C, LN(LPS + si-NC), LA(LPS + si-NC + Apelin-13), and LAS(LPS + si-APJ + Apelin-13), and detected the expression levels of TNF-α and IL-10 in CMs by ELISA. As shown in Fig. 4 A-B, the addition of LPS resulted in a prominent increase in TNF-α levels in CMs. Apelin-13 exhibited inhibitory effects on the release of inflammatory factors, which could be negated by si-APJ. The anti-inflammatory factor IL-10 showed the opposite trend, which revealed that activating APJ could regulate the immune microenvironment and inhibit the release of inflammatory factors.

Fig. 4.

The activation of APJ can alleviate inflammatory bone loss. (A-B) ELISA detection of the expression levels of inflammatory factors in conditioned medium of each group. (C) The osteoclast differentiation of BMs was analyzed by TRAP staining. Scale bars: 0.1 mm. (D) F-actin staining of BMs in each group. Scale bars: 0.1 mm. (E-H) The results and the assessment of protein levels in BMs of the femur in mice were quantified. (I) Representative 2D and 3D reconstruction images of the distal femur. (J-O) Bone parameters, including BMD (g/cm³), Tb. N (1/mm), Tb. Pf (1/mm), BV/TV (%), Tb. Th (mm), and BS/BV (1/mm). (P) Representative images of H&E staining of the distal femur. Scale bars: 0.5 mm. (Q) TRAP staining was used to analyze osteoclast differentiation. Scale bars: 0.1 mm. (R) Number of TRAP-positive osteoclasts on the trabecular bone surface (N.Oc/BS, 1/mm) in distal femora from the Sham group, LPS group, and Apelin-13 group. (in vitro n = 3; in vivo n = 6, * P < 0.05, ** P < 0.01).

Subsequently, we observed the osteoclastic differentiation function of BMs in CMs supplemented with RANKL and macrophage colony-stimulating factor (M−CSF). The results of TRAP staining (Fig. 4 C and S5 B) demonstrated an increase in the number of activated osteoclasts in the LN group than in the control group. Although the LA group exhibited a lower number of osteoclasts compared to the LN group, the decrease in the number and size of osteoclasts was observed upon the knockdown of APJ. The outcomes of the cytoskeleton F-actin ring (Fig. 4 D and S5 C) were consistent with the findings from TRAP staining analyses. Through Western blot analysis (Fig. 4 E-H), we observed a significant increase in the expression of osteoclast-related proteins in the LN group that in the control group. In the LA group, the expression of osteoclast-associated proteins was lower than that in the LN group. However, in the LAS group with APJ knockdown, the therapeutic effect of Apelin-13 was reversed. This result suggested that the disruption of the bone immune microenvironment caused by macrophage M1 polarization may result in the excessive activation of osteoclasts.

We then evaluated the bone mass of mice in vivo. Micro-CT scanning results (Fig. 4 I) revealed a significant decrease in femoral bone mass in the intervention group compared to the sham group. However, the treatment group exhibited restored bone mass compared to the intervention group. Quantitative analysis of bone parameters (Fig. 4 J-O) revealed a significant decrease in BMD in LPS group compared to the control group, while the treatment group effectively suppressed the decrease in BMD. In addition, compared with the control group, BV/TV, Tb. N, Tb. Th in the intervention group were significantly decreased, while the treatment group could reverse the reduction in these bone parameters. BS/BV and Tb. Pf showed the opposite trend to BMD. Additionally, H&E staining demonstrated a significantly lower number of trabeculae in the LPS group, while the Apelin-13 treatment group exhibited a higher number of trabeculae than the LPS group (Fig. 4 P). These results demonstrated that the activation of APJ can alleviate inflammatory bone loss. Next, we evaluated the differentiation of osteoclasts in paraffin sections of the distal femur of mice. The results of TRAP staining (Fig. 4 Q-R) revealed a decrease in the average number of osteoclasts per bone surface (N.Oc/BS) in the Apelin-13 treatment group compared with LPS group. These findings align with the outcomes of experiments in vitro, demonstrating that activating APJ can suppress osteoclast activation, and ultimately mitigate inflammation-induced bone loss.

High-throughput sequencing verified the mechanism by which APJ regulates macrophage polarization

Subsequently, we conducted a high-throughput sequencing analysis to further investigate the underlying mechanism through which APJ modulates macrophage polarization (Fig. S6 A-B). As depicted in Fig. 5 A-B, after the intervention of Apelin-13, a total of 1538 genes exhibited upregulation, while 1746 genes displayed downregulation. Across all gene analyses, differentially expressed genes were notably enriched in immune regulation processes as illustrated in Fig. S6 C-D, including macrophage differentiation and the inflammatory response. Gene expression analysis of macrophage differentiation and inflammatory response revealed significant downregulation of macrophage M1 polarization (Nos2, Tnfaip1, Il1a, Nfkb1) and inflammatory indicators, along with upregulation of anti-inflammatory factors (Il10, Il6). Additionally, genes related to osteoclast differentiation were also significantly downregulated (Nfatc1, Mmp9, Tnfrsf11a) (Fig. 5 C). The high-throughput sequencing results demonstrated that the treatment with Apelin-13 effectively inhibited macrophage M1 polarization and osteoclast differentiation induced by LPS. Furthermore, the results, as depicted in Fig. 5 D, demonstrated the analysis of the top 30 KEGG enrichment pathways associated with downregulated genes in the Apelin-13 group. The results revealed a significant enrichment of APJ activation in the NOD-like receptor (NLRP) signaling pathway. The results of the gene set enrichment analysis (GSEA) further demonstrated the inhibitory effect of Apelin-13 on the NLRP signaling pathway (S6 E-F).

Fig. 5.

High throughput sequencing analysis of BMs after Apelin-13 intervention. (A) Volcano plot analysis. (B) Gene heat map analysis. (C) Gene heat map of inflammatory cytokines and osteoclast differentiation related genes. (D) The top 30 of KEGG enrichment of gene expression profiles. (E) Go enrichment of immune-related gene expression profiles. (F) Gene heat map of mitophagy related genes. (G-K) The mRNA levels of cybb, nlrp3, il1, lc3b and bnip3 were analyzed in BMs after LPS and Apelin-13 induction. (in vitro n = 3, * P < 0.05, ** P < 0.01).

The analysis revealed a significant enrichment of genes related to mitochondria in Gene Ontology (GO) analysis (Fig. 5 E). Mitochondria, known for their ability to regulate intracellular ROS production and clearance, play a crucial role in macrophage M1 polarization. [22] Consequently, mitochondria may serve as a pivotal organelles through which Apelin-13 modulates macrophage polarization. Subsequently, when we analyzed the mitochondria related genes, we found that the mitophagy related genes (Map1lc3b, Bnip3, Becn1, Arfip2, Phb2, Gba) changed significantly (Fig. 5 F). Considering the findings from the GO analysis, it is hypothesized that mitophagy might have a crucial role in the activation of APJ. We used RT-qPCR to further verify the sequencing results, and the results (Fig. 5 G-K) showed that the activation of APJ could effectively reduce the expression of inflammation related phenotypes (Cybb, Nlrp3, Il1) and promote the expression of mitophagy related phenotypes (Lc3b, Bnip3). To further elucidate the mechanisms by which the membrane receptor APJ transmits signals upon activation, we conducted a KEGG pathway analysis on the sequencing data. This analysis suggested potential involvement of the AMPK and MAPK signaling pathways in the signal transduction process following APJ activation. Prior research has demonstrated a significant correlation between the AMPK and MAPK signaling pathways and the activation of APJ, as well as mitophagy. [23], [24] To substantiate these findings, we performed Western blot analyses on proteins associated with the AMPK and MAPK pathways. The results showed in Fig. S7 A-E indicated a significant reduction in phosphorylated AMPK (p-AMPK) levels in the LPS-treated group, whereas there was a marked increase in the expression of phosphorylated p38 MAPK (p-p38 MAPK), phosphorylated JNK (p-JNK), and phosphorylated ERK (p-ERK). The results indicated that under inflammatory conditions, the AMPK pathway experiences substantial inhibition, whereas the MAPK pathway is markedly activated. Upon activation of APJ with Apelin-13, there was significant alterations in the expression level of p-AMPK, which was subsequently reversed following si-APJ blockade. Notably, the expression changes of p-JNK and p-ERK were inapparent, and p-p38 MAPK did not exhibit variations. Consequently, it is hypothesized that the AMPK pathway predominantly mediates signal transduction after APJ activation. Therefore, we speculated that activating APJ could promote AMPK mediated mitophagy and play a role in regulating macrophage polarization.

In summary, it is postulated that APJ may impede the activation of the NLRP signaling pathway through the facilitation of mitophagy, thereby contributing to the anti-M1 polarization of macrophages.

APJ inhibits macrophage M1 polarization by promoting mitophagy

BNIP3 and LC3B play a key role in the regulation of mitophagy, BNIP3 can bind to PINK1, stabilize it on the outer mitochondrial membrane, inhibit the degradation of PINK1, and then initiate PINK1-PARKIN mediated mitophagy.[25] To verify whether APJ plays a role by regulating BNIP3-PINK1-PARKIN-mediated mitophagy, we first performed cellular immunofluorescence co-staining of LC3B and TOM20, the mitochondrial marker protein. As shown in Fig. S8 A-C, the fluorescence intensity of TOM20 in LPS group was higher than that in treatment group, while the fluorescence intensity of LC3B was significantly lower than that in treatment group. In LPS group, mitophagy was inhibited, resulting in abnormal accumulation of mitochondria, while activating APJ could promote mitophagy and reduce abnormal accumulation of mitochondria. Next, we performed immunofluorescence co-staining of BNIP3 and PINK1. At high magnification, we found that the immunofluorescence intensity of BNIP3 and PINK1 in the treatment group was enhanced compared with that in the LPS group, and the degree of colocalization of BNIP3 and PINK1 in the treatment group was significantly higher than that in the LPS group (Fig. S9 A-C).

To further verify whether mitophagy can inhibit macrophage M1 polarization, we employed cyclosporine A (CsA) to suppress the level of mitophagy. This inhibition of mitophagy in the CsA intervention group was visually evident under TEM, as depicted in Fig. 6 A, where the damaged mitochondria were unable to be promptly cleared. As shown by the mito-tracker results in Fig. S10 A-B, the observed decline in the MMP following mitochondrial damage in LPS group, its subsequent recovery upon APJ activation, nevertheless, the effect of elevated the MMP was eventually blocked by CsA. These results suggested that APJ can promote mitophagy, reduce the accumulation of damaged mitochondria and restore MMP. On the other hand, the results of Western blot (Fig. 6 B-D) showed that macrophage M1 polarization was higher in the CsA group compared to the APJ activated group. These results suggested that inhibition of mitophagy can promote macrophage M1 polarization.

Fig. 6.

Apelin-13 can promote PINK1-PARKIN mediated mitophagy to inhibit macrophage M1 polarization. (A) The structure of damaged mitochondria and autolysosomes in BMs was observed via TEM. Red arrows indicate damaged mitochondria and yellow arrows indicate autolysosomes. Scale bars: 4 μm. (B-D) The results and the assessment of protein levels in BMs were quantified. (E) Representative images of JC-1 staining; red (J-aggregates), green (JC-1 monomers). Scale bars: 50 μm. (F) Quantitative analysis of the average fluorescence intensity of J-aggregates. (G) The structures of normal mitochondria, damaged mitochondria and autolysosomes in BMs were observed via TEM. Green arrows indicated normal mitochondria without damage, red arrows indicate mitochondria with damage and yellow arrows indicate autolysosomes. Scale bars: 4 μm. (H) The results of mitophagy related protein levels in BMs were quantified. (in vitro n = 3, * P < 0.05, ** P < 0.01).

Next, we used siRNA to knock down the expression of APJ to further explore the role of APJ in regulating mitophagy. As shown in Fig. 6 E-F, the results of JC-1 staining indicated that the MMP of LPS group exhibited a significant decrease compared to the control group. However, after the intervention of Apelin-13, the MMP was restored. The utilization of si-APJ resulted in a remarkable inhibition of the increase in the MMP observed in the LA group, which was the same as the effect of CsA as mentioned earlier. In addition, the use of TEM, as shown in the Fig. 6 G, allows for visual confirmation of the typical oval or short rod-shaped morphology of healthy mitochondria in the control group, the mitochondrial cristae exhibited a deep and regular structure. Conversely, in the LN group, the mitochondrial cristae were either reduced or absent, and the density of the mitochondrial bilayer membrane was increased. Activation of APJ leads to the observation of numerous autophagic lysosomes within the visual field, where damaged mitochondria are enveloped. The number of damaged mitochondria were significantly reduced. However, when APJ was suppressed, the mitophagy induced by Apelin-13 is significantly hindered, resulting in a decrease in the number of autophagic lysosomes compared to the previous state and some damaged mitochondria fail to initiate mitophagy, which suggested that si-APJ and CsA have similar effects in inhibiting mitophagy. The results from cellular IF analysis (Fig. S11A-B) revealed that LC3B, a crucial protein involved in the initiation of mitophagy, exhibited a similar trend to MMP. The results indicated that activating APJ can promote mitophagy and restore mitochondrial function.

We then proceeded to carry out an in-depth validation analysis of the mitophagy pathway. The Western blot analysis conducted in this study (Fig. 6 H and S12 A-D) revealed a significant decrease in the expression of PINK1, PARKIN, and LC3B Ⅱ: LC3B Ⅰ in LPS group. Conversely, the expression of SQSTM1/P62, a mitochondrial phagocytic substrate, was increased. The observed increase in mitophagy-related proteins in the LA group suggested that the activation of APJ may enhance PINK1-PARKIN mediated mitophagy. However, si-APJ can block PINK1-PARKIN mediated mitophagy, causing a decrease in PINK1, PARKIN and LC3B, as well as the accumulation of SQSTM1/P62.

Prior research has indicated that ROS serve as a connecting factor between mitophagy and macrophage polarization.[26] Hence, the initial assessment of ROS expression levels in BMs following intervention was conducted, as depicted in Fig. S13 A-C. DHE staining revealed a positive correlation between the fluorescence intensity of ROS and macrophage M1 polarization. The fluorescence intensity of ROS in the LN group was increased with the enhancement of macrophage M1 polarization. Additionally, the decrease in ROS fluorescence intensity observed in the Apelin-13 treatment group was reversed by the knockdown of APJ. These results indicated that activating APJ can promote PINK1-PARKIN mediated mitophagy and inhibit macrophage M1 polarization through the modulation of ROS expression.

Furthermore, previous research has demonstrated that the inhibition of mitophagy leads to an augmentation in macrophage M1 polarization, resulting in a notable increase in apoptosis. [27] The high-throughput sequencing results above further verified this result (Fig. 5 D). As shown in Fig. S14 A-D, we validated the involvement of these crucial apoptosis-related proteins, as evidenced by the upregulation of pro-apoptotic proteins BAX and CASPASE3, as well as the downregulation of the anti-apoptotic protein BCL-2, thereby providing additional evidence for this phenomenon. The results obtained from Annexin V/Propidium iodide (PI) staining (Fig. S15 A-B) also demonstrated a significant decrease in early apoptosis due to the activation of APJ exposure, which was mitigated by the presence of Apelin-13. However, the protective effect of Apelin-13 on cells was impeded when mitophagy was inhibited. In conclusion, our results provided evidence for the protective function of APJ in facilitating PINK1-PARKIN mediated mitophagy and suppressing macrophage M1 polarization and apoptosis.

NLRP3 inflammasome was inactivated by the activation of APJ

The analysis of high-throughput sequencing results indicated the significant involvement of the NLRP signaling pathway in the modulation of macrophage M1 polarization by APJ. As shown in Fig. 7 A-D, NLRP3, mature-IL-1β, and ASC in the LPS group exhibited a significant increase, and the expression of these crucial proteins was downregulated by the activation of APJ. The reversal of the therapeutic effect of Apelin-13 upon si-APJ knockdown indicated that the activation of APJ could effectively inhibit inflammasome activation. This phenomenon was further supported by the results of cellular IF and quantitative analysis (Fig. 7 E-H). In order to establish a stronger association between NLRP3 and mitophagy, we employed CsA to impede the mitophagy induced by APJ activation. The results demonstrated that the inhibition of APJ on the inflammasome was hindered when mitophagy was blocked. (Fig. S16 A-D).

Fig. 7.

APJ regulates the immune microenvironment by inhibiting the activation of NLRP3 inflammasome. (A-D) The results and the assessment of protein levels in BMs were quantified. (E-F) Representative images of IF; green (NLRP3 & IL-1β), and blue (nuclei). Scale bars: 50 μm. (G-H) Quantitative analysis of the average fluorescence intensity of NLRP3 and IL-1β. (I) The results of protein levels in BMs were quantified. (J) Representative images of IF in the distal femur; red (INOS) and blue (nuclei). Scale bars: 0.5 mm. (in vitro n = 3; in vivo n = 6, * P < 0.05, ** P < 0.01).

Subsequently, we proceeded to examine the impact of APJ activation on the NLRP signaling pathway in vivo through Western blotting and tissue IF. As depicted in Fig. 7 I and S17 A-C APJ activation effectively suppressed the activation of the NLRP signaling pathway in BMs within the femurs of mice, consequently impeding the maturation of IL-1β. Furthermore, the outcomes obtained from tissue IF (Fig. 7 J and S17 D) provided additional evidence supporting the notion that APJ activation mitigates the LPS-induced activation of the NLRP signaling pathway and rectifies the disorder of bone immune microenvironment. The results indicated that the activation of APJ has the potential to impede the activation of the NLRP signaling pathway and diminish the production of the NLRP3 inflammasome through the promotion of mitophagy.

Discussion

Bone loss resulting from systemic inflammation is frequently underrecognized in clinical practice. The prolonged administration of antibiotics and hormone therapy can contribute to antibiotic resistance and other complications, including osteoporosis. [28] Prior research has indicated that the overexpression of M1 macrophages, which activates osteoclasts, plays a significant role in systemic inflammatory bone loss. [29], [30], [31] APJ, a G protein-coupled receptor ubiquitously expressed in various cell types throughout the body, plays a critical regulatory role in inflammatory responses, making it a promising target for pharmacological research. [32], [33], [34] Nonetheless, prior research has just primarily identified this phenomenon without thoroughly investigating the precise mechanisms through which APLN/APJ modulates macrophage activity. In this study, we demonstrate that APJ inhibits macrophage M1 polarization and mitigates both excessive osteoclast activation and bone loss. Furthermore, we confirm that activated APJ predominantly facilitates BNIP3-PINK1-PARKIN-mediated mitophagy through the enhancement of the AMPK signaling pathway. This process reduces ROS accumulation and inhibits the production of NLRP3 inflammasomes and the release of inflammatory cytokine IL-1β, thereby modulating the immune microenvironment. Consequently, APJ is anticipated to be a promising target for the management of systemic inflammatory diseases.

In the human body, ROS function as a double-edged sword. [35] On one hand, ROS are crucial products of the inflammatory response, playing a significant role in immune regulation and microbial elimination. Conversely, excessive ROS activation can disrupt intracellular homeostasis, resulting in organelle damage and potentially leading to cell death. [36] While elevated ROS levels are necessary for the elimination of tumor cells during cancer therapy, they can adversely affect macrophages, which are essential for immune regulation, by inducing cell cycle arrest and activating inflammatory pathways, ultimately leading to M1 polarization and massive release of inflammatory factors in macrophages. [37], [38], [39], [40] Inflammatory mediators subsequently enhance the production of ROS, initiating a cascade reaction that results in the substantial accumulation of ROS. This process ultimately disrupts the immune microenvironment. The inflammatory and immune milieu within bone tissue can markedly promote the differentiation and maturation of osteoclasts. [41], [42] Prior research has identified the aberrant activation of osteoclasts as the primary contributor to bone loss in inflammation-induced osteoporosis. Over expressed ROS can further stimulate downstream signaling pathways, including MAPK and NF-κB, by upregulating the expression of RANKL, M−CSF and NFATc1. This process accelerates osteoclast differentiation and enhances their bone resorption capacity. [43], [44], [45] Our results demonstrate that inflammatory conditions lead to an aberrant activation of osteoclast differentiation in vitro and a significant reduction in bone density in the distal femur of mice with systemic inflammation. Activation of the APJ receptor can mitigate the abnormal activation of ROS, ameliorate the inflammatory immune microenvironment, decrease osteoclast activation both in vitro and in vivo, and consequently alleviate inflammation-induced bone loss.

Mitochondria serve as the primary locus for the generation of ROS within cells. During the electron transport chain, electrons may escape and react with molecular oxygen, resulting in the formation of the highly oxidative superoxide anion (O₂–), which is subsequently converted into hydrogen peroxide (H₂O₂). [46], [47] However, the inhibition of mitophagy results in the failure to timely eliminate damaged mitochondria, leading to the accumulation of ROS within cells. This accumulation further exacerbates mitochondrial damage and inhibits mitophagy, ultimately inducing apoptosis. [48] Mitophagy, a crucial mechanism for maintaining mitochondrial quality, has garnered significant attention since its inception in 2005 due to its pivotal role in aging, neuropathy, inflammation and macrophage polarization. [26], [49], [50], [51], [52], [53] Our results indicate that APJ facilitates the clearance of damaged mitochondria through the promotion of mitophagy, subsequently interrupting ROS cascade and thereby mitigating the activation of the NLRP3 inflammasome. This process ultimately inhibits the M1 polarization and apoptosis of macrophages.

Previous research has indicated that the mitophagy pathway can be primarily categorized into two distinct types: the PINK1/PARKIN-mediated ubiquitin-dependent pathway and the non-ubiquitin-dependent pathway, which is mediated by mitophagy receptors such as BNIP3, NIX, and FUNDC1. [54], [55] When mitochondrial depolarization occurs, the mitochondrial membrane potential diminishes, preventing PINK1 from translocating through the outer membrane, resulting in its accumulation on the outer membrane surface. The accumulated PINK1 subsequently undergoes phosphorylation and activation, facilitating the recruitment and activation of PARKIN in the cytoplasm. Upon activation, PARKIN facilitates the ubiquitination of mitochondrial outer membrane proteins. Subsequently, autophagy-related proteins, such as SQSTM1/P62, detect the “eat me” signals on the membrane surface. [56], [57] In contrast, the non-ubiquitin-dependent pathway is comparatively straightforward, wherein autophagy receptors like BNIP3 can directly interact with LC3 to trigger mitophagy. [55] While earlier investigations have predominantly studied these two pathways in isolation, however the fact is that that biochemical processes within cells are highly intricate, and proteins may not function entirely independently during these reactions. Our results propose that BNIP3 could play a role in the PINK1/PARKIN-mediated mitophagy within the context of APJ. Consequently, we posit that BNIP3, as a crucial regulatory protein of apoptosis and autophagy located on the mitochondrial outer membrane, facilitates ubiquitin-dependent mitophagy by stabilizing the accumulation of PINK1 on the mitochondrial outer membrane. Our results elucidate the specific mechanism through which APJ modulates mitophagy, while also underscoring the intricate and integrated regulatory functions of mitochondria and even cellular processes.

APJ, as a regulatory target for disorders of the bone immune microenvironment, possesses intrinsic advantages. The endogenous ligand Apelin, associated with APJ, is encoded by the APLN gene located on human chromosome Xq25 ∼ 26.1. In humans, Apelin is predominantly synthesized by adipocytes. [58], [59] The presence of yellow adipose tissue within the bone marrow provides convenience for Apelin synthesis and enables its binding to the APJ receptor located on the surface of bone marrow macrophage membranes, thereby contributing to the regulation of bone immunity. Azelapra / AMG986 (BGE-105), a novel APJ-specific potent agonist under development for the amelioration of muscle aging and heart failure, has shown promising biocompatibility in Phase I clinical trials (NCT03276728, NCT03318809, NCT06141889). [60], [61], [62], [63], [64] Furthermore, Phase II clinical trials are in progress to evaluate its efficacy in obesity management (NTC06515418). While clinical investigations into the role of APJ agonists in immune regulation have yet to be conducted, APJ continues to be a promising target for modulating the bone immune microenvironment.

Conclusion

In summary, our study has demonstrated the significant involvement of APJ in the regulation of the immune microenvironment. Furthermore, we have enhanced the understanding of the underlying mechanism through which APLN-APJ modulates the immune microenvironment. These findings offer promising avenues for therapeutic interventions and identify potential targets for addressing inflammatory bone loss.

Animal experiments have been granted approval by the Ethics Committee of Soochow University (SUDA20230625A07). All applicable international, national, and institutional guidelines for the care and use of animals were followed.

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: