Harmine ameliorates steroid-induced osteonecrosis of the femoral head by modulating metabolism and ferroptosis via HIF1-α

Abstract

Steroid-induced osteonecrosis of the femoral head (SONFH) often leads to disability in the terminal stage, and there are still no effective drugs for prevention. Harmine (Hm), which is extracted from Peganum harmala L., is a β-carboline alkaloid with a number of biological and pharmacological properties. In this study, we discovered the protective effect of Hm on SONFH in rats. In vitro studies showed that Hm could ameliorate oxidative stress and facilitate osteogenic differentiation in bone mesenchymal stem cells (BMSCs), which play an important role in bone development and repair. Bioinformatics analysis revealed that ferroptosis and the HIF1-α pathway played an important role in SONFH pathogenesis. Mechanistically, Hm treatment alleviated ferroptosis by promoting HIF1-α expression in BMSCs. When shHif1a was used to inhibit HIF1-α expression, the protective effects of Hm disappeared. Meanwhile, we further found that Hm intervention promoted HIF1-α expression to increase GSH production and the GSH/GSSG ratio, which further promoted the scavenging of free radicals. In summary, Hm may be a promising candidate for the treatment of SONFH.

Graphical abstract

1. Introduction

Osteonecrosis of the femoral head (ONFH) is a well-known joint disease that leads to femoral head collapse, discrepancy in leg length, and limited hip mobility [1]. Generally, ONFH can be divided into the following two categories: traumatic osteonecrosis and nontraumatic osteonecrosis. Among nontraumatic ONFH, steroid-induced osteonecrosis of the femoral head (SONFH) is the most common form and has attracted much attention due to its high incidence and associated significant bone microstructure deterioration [2,3]. Whereas the precise pathogenesis of this illness is only partially elucidated, the progression of the disease can cause serious irreversible injury to the femoral head and lower the quality of life. Recently, total hip arthroplasty has become a safer and more straightforward treatment for the terminal stage of SONFH. Nevertheless, SONFH tends to occur in younger patients, which could significantly increase the probability of revision surgery and cause a heavy economic and psychological burden [4]. Therefore, early treatment is particularly important in patients with SONFH to maintain their quality of life.

Usually, long-term or high-dose use of glucocorticoids (GCs) is the most common cause of SONFH [5]. The excessive use of GCs can lead to microcirculation perfusion disorder and the death of bone cells and bone marrow cells, which results in an imbalance in bone homeostasis [6]. In this process, inhibition of osteoblast activity by a high dose of GCs plays a critical role [6,7]. There are three distinct stages of osteoblasts involved in the process of bone formation as follows: osteoprogenitor, preosteoblast, and osteoblast. Furthermore, mature osteoblasts secrete extracellular matrixes and transform into osteocytes [8,9]. Bone mesenchymal stem cells (BMSCs) are precursors of osteoblasts and have an important influence on the supplementation of osteoblasts during bone formation and osteonecrosis repair [10]. Previous studies suggested that a high dose of GC inhibits BMSC proliferation, stimulates reactive oxygen species (ROS) production, and even leads to cell death [11,12]. In addition, it has been reported that GC can also inhibit osteogenic gene transcription and translation, which further hinders osteogenesis and bone tissue repair [13]. Promoting osteogenic function is an important strategy to treat multiple bone diseases [14,15]. Osteoblasts play a critical role in the imbalance of bone homeostasis caused by high-dose use of GCs, and the promotion of osteoblastogenesis and bone formation is a potentially promising method to treat SONFH.

Ferroptosis, a type of programmed cell death that is different from apoptosis, has been reported to induce BMSC damage during the progression of multiple orthopedic diseases [16]. Ferroptosis is driven by a dysregulated redox state and characterized by iron-dependent cell death, during which mitochondrial destruction and lipid peroxidation are significant [17]. Increasing studies claim that an uncontrolled redox state in BMSCs promotes the osteonecrosis process of SONFH, and treatment with antioxidants can effectively mitigate this process [[18], [19], [20]]. In addition, an increase in intracellular Fe²⁺ and ROS levels was observed in the rat SONFH model [21]. These results suggest that ferroptosis may also be involved in reduced osteogenic differentiation. Although the relationship between osteoblastogenesis and ferroptosis in SONFH and the underlying mechanism are not fully explained, targeting BMSC ferroptosis may be an effective way to treat SONFH.

Cellular metabolism is closely related to cell survival and cellular activity. Metabolic pathways, such as mitochondrial respiration, fatty acid metabolism and balance of iron, are directly related to lipid peroxidation and ferroptosis [22]. Additionally, redox-related metabolites, such as glutathione/glutathione oxidized (GSH/GSSG), NAD(P)H/NAD(P)+ and the ubiquinone axis, play an important role in neutralizing oxidative damage to cellular membranes [22,23]. Since cell metabolism has emerged as an important cause of ferroptosis-related diseases and since it provides a variety of pharmacological targets for the treatment of diseases, there has been great interest in the relationship between metabolism and the ferroptosis process. Due to the complexity of metabolic studies, the mechanisms between metabolism and ferroptosis remain unclear.

Currently, pharmaceutical drugs have little effect when used for the treatment of SONFH, and most SONFH patients have to undergo total hip arthroplasty (THA). Therefore, clarification of the SONFH mechanism and the development of new drugs to prevent SONFH progression are urgently needed. Exploring therapeutic drugs for SONFH using already approved drugs will help accelerate clinical translation. Harmine (Hm), a natural tricyclic β-carboline alkaloid, is extracted from numerous plants, including Peganum harmala (Syrian Rue) and Passiflora incarnata (Passionflower) [24]. It has multiple pharmacological effects, including monoamine oxidase inhibition, improvement of insulin sensitivity, vasodilatation, and antioxidant activity [25]. Hm has been reported to inhibit the differentiation of preosteoclasts into multinucleated osteoclasts, promote osteoblastogenesis and reduce bone loss in osteoporotic mice [26]. Our laboratory group previously reported that Hm treatment could inhibit wear particle-induced inflammation and facilitate osteogenic differentiation [27]. These results showed that Hm has a good immunomodulatory effect on bone homeostasis; however, the effects of Hm on ferroptosis in osteoblasts based on GC intervention, its activity in preventing SONFH, and the possible underlying mechanisms remain unclear.

Considering the important role of bone homeostasis imbalance in the pathogenetic process of SONFH, it makes sense that Hm may be a promising candidate for treatment. In this study, we used methylprednisolone sodium succinate (MP), one of the commonly used glucocorticoids for autoimmune disease, to establish a rat SONFH model and assess whether Hm treatment could suppress bone loss and ameliorate the morphology of SONFH in vivo. We also assessed the impact of Hm on ferroptosis, as well as its involvement in osteoblast formation and differentiation by MP intervention using BMSCs. In addition, we investigated the latent mechanism of Hm treatment in vitro.

2. Materials and methods

2.1. Animal experiments

All animal experiments adhered to the ARRIVE guidelines and the ethical guidelines of the Laboratory Animal Center of Soochow University. And the animal experiments were carefully performed in all possible steps to avoid animal suffering. All animal experiments were approved by the Ethics Committee of Soochow University (approval No. SUDA20220925A02).

A total of 48 10-week-old male Sprague Dawley (SD) rats (weight: 300 ± 25 g) were used. The rats were randomly divided into the following four groups (n = 12/group) to establish SONFH model: (1) Control, (2) MP, (3) Hm-low (5 mg/kg), and (4) Hm-high (20 mg/ kg). Day 26, four rats were randomly selected from each group, and serum was collected for detecting bone-turnover markers (BTMs) after euthanasia. The rats were anesthetized with an intraperitoneal injection of pentobarbital sodium (40–50 mg/kg), and blood was collected from the right ventricle. After blood collection, pentobarbital sodium was added to 100 mg/kg for euthanasia. The blood sample was left at 4 °C for 2 h, and the serum was collected by centrifugation at 3000 rpm for 20 min. In addition, on day 40, serum was collected from all the rest of the rats under anesthesia, and their bilateral femoral heads were collected after euthanasia. For detailed information, refer to the supplementary materials.

2.2. Micro-CT scanning

High-resolution micro-CT (SKYSCAN 1176, Bruker, Konitch, Belgium) was performed with the following scanning parameters: 18 µm, 70 kV, and 141 mA. The trabecular bone parameters, including bone mineral density (BMD, mg/cm3), bone volume per tissue volume (BV/TV,%), trabecular thickness (Tb. Th, mm) and trabecular separation (Tb. SP, mm) were assessed with CT Analyzer analysis software (CTAN, Bruker).

2.3. Histology and bone histomorphometry analysis

Hematoxylin and eosin (H&E) staining was performed to investigate the appearance of specimens from different groups. Masson staining was performed to identify new immature collagen (blue), which indicated new bone and fibrous tissue formation. HIF1-α antibody (Abcam, UK, #ab114977) was utilized for immunohistochemistry (IHC) staining to determine HIF1-α expression.

2.4. Cell extraction and culture

As previously described, BMSCs were collected from the femurs and tibias of 6-week-old healthy rats [28]. Cells were cultured in alpha minimum essential medium (α-MEM, HyClone, Logan, UT, USA) containing 10% fetal bovine serum under environmental condition of 37 °C and 5% CO₂.

2.5. Cell cytotoxicity assay

Briefly, BMSCs were cultivated in 96-well plates at a density of 5 × 10³ cells per well with different concentrations of MP or Hm for 24 h, 48 h, and 72 h. BMSC activity was detected by CCK-8 assay using a microplate reader (BioTek, Winooski, VT, USA) at 450 nm. For detailed information, refer to the supplementary materials.

2.6. ROS detection assay

BMSCs were cultured in 24-well plates (cell density: 1.5 × 10⁴ cells per well) for 48 h with 10 µM, 100 µM, or 1000 µM MP in three replicate wells. The level of intracellular ROS was evaluated using an ROS assay kit (2′,7′-dichlorodihydrofluorescein diacetate, DCFH-DA) (Beyotime Biotech, China).

2.7. Alkaline phosphatase (ALP) and alizarin red s (ARS) staining

BMSCs were cultured in 24-well plates (cell density: 1.5 × 10⁴ cells per well) with different treatments in 3 replicate wells. After 10 days of culture with osteogenic induction medium (a-MEM supplemented with 10% fetal bovine serum, 50 µM ascorbic acid, 10 mM glycerol β-phosphate and 100 nM dexamethasone), the BMSCs were stained for ALP activity. After 18 days of culture with osteogenic induction medium, the BMSCs were stained with ARS.

2.8. Western blotting analysis

The cells were cultured in 6-well plates (cell density: 3 × 10⁵ cells per well) with different treatments for 48 h, and the cells were collected for western blotting analysis. The primary antibodies used were HIF1-α (1:1000, Abcam, UK, #ab179483) and β-actin (1:5000, Abcam, UK, #ab8226).

2.9. Cell immunofluorescence staining

BMSCs were seeded in 24-well plates at 1.5 × 10⁴ cells per well. After treatment with osteogenic induction medium for 5 days, BMSCs were incubated with primary antibodies against osteocalcin (OCN) (1:200, ABclonal, China: A6205) or runt-related transcription factor 2 (Runx2) (1:200, Abcam: #ab192256) and then stained with immunofluorescence secondary antibodies.

2.10. Flow cytometry

BMSCs were cultured in 6-well plates (cell density: 5 × 10⁵ cells per well) with different treatments for 48 h. Then, cells were collected for flow cytometry analysis according to the instructions of different detection kits, including the V-FITC/PI apoptosis detection kit (Vazyme Biotech Co., Ltd., Nanjing, China), the C11-BODIPY 581/591 detection kit (ABclonal, Inc., China, #RM02821), the JC-1 detection kit (Yeasen Biotechnology, Shanghai, China, #36222ES76), the ROS detection kit (Beyotime Biotech, China, S0033 M), and the Fe²⁺ detection probe FerroOrange (Dojindo, Japan, #F374). A flow cytometer (Attune NxT, Thermo Fisher, USA) and FlowJo software version 10.6.2 (TreeStar, San Carlos, CA, USA) were utilized for data collection and processing.

2.11. Bioinformatic analysis

The datasets analyzed in this study were acquired in our previous study, which involved integrated analysis of the transcriptome and proteome of SONFH patients [29]. A total of six SONFH patients and six femoral neck fracture patients who underwent total hip arthroplasty (THA) surgery were selected as the SONFH and control groups. A total of 758 differentially expressed genes (DEGs) in the transcriptome and 1097 differentially expressed proteins (DEPs) in the proteome were identified. Then, Gene Ontology (GO), gene set enrichment analysis (GSEA) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were carried out by using online analysis tools (www.omicshare.com/tools).

2.12. Knockdown of HIF1-α

The shHif1a lentivirus vectors were constructed according to the protocols of Lentivirus Production and Designing shRNA Oligos for pLKO.1 (www.addgene.org/protocols). Then, the cells were transfected with shHif1a lentivirus, and total RNA and proteins were harvested to validate knockdown efficiency. A total of 3 shHif1a oligos were designed for the knockdown of HIF1-α, and we chose the most effective oligo for subsequent experiments, as shown in Fig. 4d and 4e. The target sequences of the shRNAs were as follows: shHif1a-1 (sh-1) AGACTCAAGCAACTGTTATAT, reverse complement sequence ATATAACAGTTGCTTGAGTCT; shHif1a-2 (sh-2) CACAGCCTCGATATGAAATTT, reverse complement sequence AATTTCATATCGAGGCTGTG, shHif1a-3 (sh-3) GCCGTTCAATTTATGAATATT, reverse sequence AATATTCATAAATTGAACGGC.

Fig. 4.

Harmine relieved MP-induced ferroptosis by increasing HIF1-α expression. (a) Venn diagram of intersection between DEPs (1097) in proteomics and DEGs (758) in transcriptomics. (b) KEGG enrichment analysis of the intersection results. (c) Western blot analysis of HIF1-α expression in BMSCs treated with 100 µM MP and 10 µM Hm or 50 µM Hm. (β-actin was selected as an internal reference.) (d) qPCR analysis of HIF1-α to confirm shHif1a knockdown efficiency. (β-actin was selected as an internal reference.) (e) Western blot analysis to confirm shHif1a knockdown efficiency. (β-actin was selected as an internal reference. sh-1: shHif1a-1, sh-2: shHif1a-2, sh-3: shHif1a-3.) (f) MDA levels were checked to detect lipid peroxidation. (g) JC-1 staining was applied to observe the change in mitochondrial membrane potential. (h) Annexin V-FITC/PI staining. (All bar graphs are the mean ± SD; *p < 0.05 and **p < 0.01 compared with the 100 µM MP treatment group; ns: no significant difference; #p < 0.05 and ##p < 0.01; n = 3 per group).

2.13. Real-time quantitative PCR analysis

Total RNA was extracted using TRIzol (Invitrogen, USA), and HiScript III RT SuperMix (Vazyme Biotech Co., China: #R323–01) was utilized to synthesize cDNA. Real-time quantitative PCR was performed using Taq Pro-Universal SYBR qPCR Master Mix (Vazyme Biotech Co., China: #Q712–02) and a CFX96 Touch Real-Time PCR Detection System (Bio-Rad, USA). The cycling system was as follows: 95.0 °C for 10 min at first, and then 40 cycles of 95.0 °C for 5 s, 60.0 °C for 30 s and 72 °C for 30 s. Primers were purchased from Sangon Biotech (Shanghai, China): β-actin, F: TCAGGTCATCACTATCGGCAAT, R: AAAGAAAGGGTGTAAAACGCA; Hif1a, F: CAAGCAGCAGGAATTGGAACG, R: CTCATCCATTGACTGCCCCA.

2.14. Detection of metabolites

Glucose uptake assays were performed using Glucose Uptake Probe-Green (DOJINDO Laboratories, Japanese: #UP02). The NADP+/NADPH ratio was detected with an NADP+/NADPH Assay Kit (Beyotime Biotech, China: #S0179). The GSH/GSSG ratio was detected with a GSH and GSSG Assay Kit (Beyotime Biotech, China: #S0053). All experiments were performed according to the manufacturer's instructions.

2.15. Statistical analysis

All results were expressed as the mean ± standard deviation and analyzed using GraphPad Prism 9.0 software (GraphPad Software Inc. CA, USA). Homogeneity of variances tests and one-way analyses of variance (ANOVA) were used to measure the group variation, and Tukey's test was used to conduct all post hoc analyses. A p < 0.05 was considered statistically significant.

(More detailed methods can be found in the supplementary materials.)

3. Results

3.1. MP damages the ros scavenging system and inhibits the osteogenic differentiation of BMSCs in vitro

MP is a commonly used GC in the clinic. In subsequent experiments, we used MP to investigate the negative effects of GC in vitro and in vivo. The CCK-8 results showed that 10 µM MP had no obvious effect on BMSC proliferation, and 100 µM MP significantly inhibited BMSC proliferation after 48 h (Fig. 1a).

Fig. 1.

MP intensified cellular oxidative stress and inhibited the osteogenesis of BMSCs. (a) CCK-8 analysis was performed with various concentrations of MP in BMSCs for 24 h, 48 h, and 72 h. (b) BMSCs were stained with PI to detect cell death and were analyzed by flow cytometry. (c) Immunofluorescence images of intracellular ROS, which were stained with 2,7-dichlorofluorescein diacetate (Scale bar: 200 µm). (d) Intracellular ROS and (e) mean fluorescent intensity (MFI) of ROS were detected by flow cytometry. (f) ALP staining (Scale bar: 200 µm). (g) ALP activity was detected with an ALP activity kit. (h) ARS staining (Scale bar: 200 µm). (i) ARS was dissolved in 10% cetylpyridinium chloride, and the solution was analyzed with a spectrophotometer at 562 nm. (All bar graphs are the mean ± SD; *p < 0.05 and **p < 0.01 compared with the 0 µM MP group; ns: no significant difference; n = 3 per group).

To clarify the mechanism, we treated BMSCs with a concentration gradient of MP. Propidium iodide (PI) staining showed that the rate of cell death increased after treatment with MP in a concentration-dependent manner, especially at 100 µM and 1000 µM (Fig. 1b). At the same time, we observed that the ROS levels in BMSCs were positively correlated with the rate of cell death (Fig. 1c-e). The delicate balance between ROS production and the ROS scavenging system is pivotal for maintaining normal physiological function. These results are consistent with previous studies and verify that the antioxidant capacity of BMSCs is weakened after MP treatment in a concentration-dependent manner [19].

In addition, we also found that the osteogenic ability of BMSCs was suppressed. Alkaline phosphatase (ALP) and alizarin red S (ARS) staining suggested that MP treatment downregulated osteogenesis compared with the control treatment, especially at high concentrations (100 µM and 1000 µM) (Fig. 1f-i).

3.2. Hm effectively alleviated MP-induced damage in BMSCs in vitro

Hm has been reported to inhibit the differentiation of preosteoclasts into multinucleated osteoclasts and promote osteoblastogenesis [26]. In addition, some studies have suggested that Hm has antioxidant activity. Therefore, we explored whether Hm could protect against MP-induced damage in BMSCs.

In this study, less than 100 µM Hm showed no significant toxic effect on BMSCs after 72 h of treatment in the CCK-8 experiment (Fig. 2a). Then, BMSCs were subjected to MP at a concentration of 100 µM, and were supplemented with different concentrations of Hm (Hm-low: 10 µM or Hm-high: 50 µM) for 48 h. The results showed that Hm treatment at both doses attenuated ROS production in BMSCs, especially in the Hm-high group (Fig. 2b, c).

Fig. 2.

Harmine restored BMSC activity under MP inhibition. (a) Structural formula of Hm and CCK-8 analysis with various concentrations of Hm in BMSCs. (b) Immunofluorescence images of intracellular ROS with 100 µM MP and 10 µM Hm (Hm-low) or 50 µM Hm (Hm-high) treatment. (Scale bar: 200 µm) (c) The mean fluorescence intensity of ROS was detected by flow cytometry. (d) PI staining was analyzed by flow cytometry. (e) Runx2 was detected with immunofluorescence staining. (Runx2 stained with Alexa Fluor® 647. Scale bar: 25 µm.) (f) Immunofluorescence staining of OCN. (OCN stained with Alexa Fluor® 488 and F-actin stained with rhodamine. Scale bar: 50 µm.) (g) PI positive cells in flow cytometry analysis. (h) MFI of Runx2. (i) MFI of OCN. (All bar graphs are the mean ± SD; *p < 0.05 and **p < 0.01 compared with the 100 µM MP treatment group; ns: no significant difference; #p < 0.05 and ##p < 0.01 compared with the control group; n = 3 per group).

Furthermore, PI staining showed that in the Hm treatment groups, MP-induced cell death was prevented in a concentration-dependent manner (Fig. 2d, g). Additionally, immunofluorescence staining of osteogenesis-related genes, including Runx2 and OCN, showed that the osteogenic potential of BMSCs markedly recovered after Hm supplementation (Fig. 2e-i). Finally, we also performed ALP and ARS staining on BMSCs treated with Hm alone to observe the direct effect of Hm on osteogenic differentiation, and Hm presented a significant osteogenesis-promoting function (Fig. S1a, b). These results indicated that Hm not only alleviated oxidative damage in BMSCs but also promoted the expression of osteogenesis-related genes to relieve the inhibitory effect of MP on osteogenesis.

3.3. Hm treatment reduced the involvement of ferroptosis in the pathogenesis of SONFH in vitro

To clarify the mechanism of Hm, bioinformatics methods were performed. In our previous study, we conducted transcriptomic and proteomic analyses on 12 total hip arthroplasty patients, including 6 SONFH patients and 6 femoral neck fracture patients (the control group)[29]. Proteomic analysis showed 1097 DEPs, and KEGG enrichment analysis illustrated that ferroptosis, a type of programmed cell death characterized by a dysregulated redox state, might be involved in the pathogenetic process of SONFH (Fig. 3a). In addition, GSEA further confirmed that ferroptosis played a role in the pathogenesis of SONFH (Fig. 3b).

Fig. 3.

Harmine protected BMSCs from MP-induced ferroptosis. (a) KEGG enrichment analysis of DEPs identified in the proteome of SONFH patients. (b) Gene set enrichment analysis (GSEA) of proteomics data from SONFH patients. (c) MDA levels were measured to detect lipid peroxidation after 100 µM MP and 10 µM Hm or 50 µM Hm treatment. (d) Flow cytometry analysis of C11-BODIPY staining. (Excitation: 500 nm, emission: 510 nm) (e) Flow cytometry analysis of JC-1 staining. (Excitation: 585 nm, emission: 590 nm) (f) Flow cytometry analysis of Annexin V-FITC/PI staining. (All bar graphs are the mean ± SD; *p < 0.05 and **p < 0.01 compared with the 100 µM MP treatment group; ns: no significant difference; #p < 0.05 and ##p < 0.01; n = 3 per group).

Ferroptosis is characterized by mitochondrial destruction and uncontrolled lipid peroxidation. Therefore, MDA and C11-BODIPY 581/591 were used to detect lipid peroxidation, JC-1 was used to detect mitochondrial potential, and FerroOrange, a Fe²⁺ detection probe, was used to detect Fe²⁺ levels. Both the MDA and C11-BODIPY results indicated that the state of lipid peroxidation was significantly increased after 100 µM MP treatment, whereas Hm treatment attenuated this state in a dose-dependent manner (Fig. 3c, d). The results of the mitochondrial potential detection experiment with JC-1 illustrated that MP treatment decreased the mitochondrial potential, which indicated that the mitochondria were destroyed (Fig. 3e). When the BMSCs were treated with Hm, the mitochondrial potential was recovered (Fig. 3e). Moreover, Fe²⁺ levels increased in the MP group and were recovered by supplementation with Hm (Fig. S2). To observe the anti-ferroptosis capacity more intuitively, we selected ferrostatin-1 (Fer-1), an inhibitor of ferroptosis, for comparison with Hm. The Annexin V-FITC/PI staining results showed that both Fer-1 and Hm (50 µM) effectively protected against high doses of MP-induced injury (Fig. 3f). In addition, we selected the ferroptosis inducer errastin (5 µM) to induce ferroptosis in BMSCs, and Annexin V-FITC/PI staining results showed that Hm treatment effectively reduced errastin-induced ferroptosis in BMSCs (Fig. S3). These results verified that ferroptosis was involved in MP-induced BMSC damage and that Hm could reduce the level of ferroptosis involvement in the pathogenesis of SONFH.

3.4. Hm alleviated MP-induced ferroptosis via the HIF1-α signaling pathway

To clarify the key genes involved in steroid-induced ferroptosis, Venn analysis was performed between DEGs identified by transcriptomics and DEPs identified by proteomics. In total, 44 genes significantly changed in both DEGs and DEPs (Fig. 4a and Table S1). Then, KEGG enrichment analysis was performed on these 44 genes, and the results also showed that ferroptosis-associated pathways were enriched (Fig. 4b). Meanwhile, the HIF1-α signaling pathway was identified as one of the top 10 pathways in KEGG analysis, which led us to speculate that HIF1-α was involved in MP-induced ferroptosis (Fig. 4b). HIF1-α works as a key regulator in response to hypoxia by promoting the transcription of many genes, including those involved in energy metabolism, programmed cell death, and ROS scavenging. Western blotting analysis also verified that expression of HIF1-α was inhibited by MP treatment, and Hm treatment significantly increased the expression of HIF1-α in a dose-dependent manner (Fig. 4c). However, when BMSCs were treated with Hm alone, the HIF1-α level in BMSCs was not significantly different between the Control and Hm treatment groups (Fig. S4).

In addition, rescue experiments were performed to further confirm HIF1-α function during MP-induced ferroptosis in BMSCs. Then, shHif1a was used to knockdown HIF1-α expression in BMSCs, and the empty vector plasmid group was set as shC. A total of three shHif1a sequences were constructed, including shHif1a-1, shHif1a-2, and shHif1a-3. The quantitative PCR and Western blotting results showed that shHif1a-2 and shHif1a-3 effectively reduced the expression of HIF1-α (Fig. 4d, e). Thus, shHif1a-2 was selected as the lentivirus sequence for subsequent experiments. The results of MDA detection indicated that lipid peroxidation was reduced by Hm, but HIF1-α knockdown eliminated this protective effect (Fig. 4f). The JC-1 staining results showed that Hm treatment led to recovery of the mitochondrial potential, and shHif1a significantly reduced the mitochondrial potential (Fig. 4g). Annexin V-FITC/PI staining further verified the protective role of HIF1-α, and the results showed that shHIF1-α increased ferroptosis, even though the BMSCs were treated with Hm (Fig. 4h). In summary, Hm protected BMSCs from ferroptosis induced by MP through the upregulation of HIF1-α expression.

As shown in Fig. 5a and b, Runx2 and OCN expression in BMSCs increased in the Hm group compared with the MP group, but Runx2 and OCN expression decreased after HIF1-α knockdown. Moreover, compared with the increased osteogenesis after Hm supplementation, ALP activity and the mineralized nodules of ARS staining were decreased when HIF1-α was knocked down, confirming the important role of HIF1-α in mediating osteogenic differentiation (Fig. 5c, d). These data further indicate that HIF1-α plays vital roles in Hm-regulated oxidative stress and osteogenic differentiation.

Fig. 5.

Harmine recovered the osteogenic differentiation potential of BMSCs through the HIF1-α pathway. (a) Runx2 expression was detected with immunofluorescence staining. (Runx2 stained with Alexa Fluor® 647. Scale bar: 25 µm.) (b) OCN expression was detected with immunofluorescence staining. (OCN stained with Alexa Fluor® 488 and F-actin stained with rhodamine. Scale bar: 50 µm.) (c) ALP staining. (ALP activity was detected with an ALP kit. Scale bar: 200 µm) (d) ARS staining. (ARS was dissolved in 10% cetylpyridinium chloride, and the solution was analyzed with a spectrophotometer at 562 nm. Scale bar: 200 µm) (All bar graphs are the mean ± SD; *p < 0.05 and **p < 0.01 compared with the MP group; ns: no significant difference; #p < 0.05 and ##p < 0.01; n = 3 per group).

3.5. Hm-induced HIF1-α accumulation altered the metabolic status and promoted GSH synthesis

KEGG pathway annotation indicated that metabolic pathways were significantly enriched (Fig. 6a, b). According to these data, we speculated that Hm-induced HIF1-α upregulation might reprogram metabolism to increase the GSH/GSSG ratio; this reprogramming allows the maintenance of glutathione peroxidase (GSH-Px) activity, including the scavenging of hydrogen peroxide, free radicals, and lipid peroxidation. The potentially affected metabolic pathways are shown in Fig. 6c.

Fig. 6.

HIF1-α maintained the NADPH/NADP+ and GSH/GSSG ratios to maintain redox balance. (a) KEGG pathway annotation analysis of the 1097 DEPs in proteomics. (b) KEGG pathway annotation analysis of the 44 genes in the intersection result between DEPs (1097) in proteomics and DEGs (758) in transcriptomics. (c) Graphical abstract for subsequent metabolite detection assays. (d) Glucose Uptake Probe-Green was used to detect the glucose uptake ability of BMSCs. (e) Detection of the NADPH/NADP+ ratio in BMSCs. (f) Detection of the GSH/GSSG ratio in BMSCs. (g) Detection of total GSH in BMSCs. (All bar graphs are the mean ± SD; *p < 0.05 and **p < 0.01 compared with the 100 µM MP treatment group; ns: no significant difference; #p < 0.05 and ##p < 0.01; n = 3 per group).

In subsequent experiments, a glucose uptake probe-green kit was utilized to observe the variations in glucose uptake capacity. The glucose uptake ability of BMSCs decreased after the high-dose MP intervention, while the glucose uptake rate recovered in the Hm-supplemented group (Fig. 6d); however, shHif1a intervention eliminated the effect of Hm supplementation on glucose uptake (Fig. 6d). Glutathione reductase (GR) has the ability to maintain the level of reduced glutathione to protect cells from ROS damage. In the presence of NADPH, GR reduces GSSG to GSH, and then NADPH turns into NADP+. Therefore, the NADPH/NADP+ and GSH/GSSG ratios were detected to assess intracellular redox homeostasis. As shown in Fig. 6e, the MP-induced decrease in the NADPH/NADP+ ratio was reversed by supplementation with Hm, while knockdown of HIF1-α caused a decrease in the NADP/NADP+ ratio. Similarly, the GSH/GSSG ratio decreased in the MP intervention group and increased in the Hm treatment group, and finally, shHif1a intervention eliminated the protective effect of Hm (Fig. 6f). Total GSH levels were also reduced in the MP group and recovered after Hm treatment, but shHif1a significantly reduced total GSH. (Fig. 6g). In addition, 2-deoxy-d-glucose (2-DG), a glucose 6-phosphate inhibitor, and PX-478 2HCl, an HIF1-α inhibitor, were used to further investigate the effect of Hm on cellular redox homeostasis. Despite Hm treatment, 2-DG (0.5 mM) intervention significantly reduced the NADPH/NADP+ and GSH/GSSG ratios (Fig. S5). Moreover, PX-478 (15 µM) intervention also blocked the protective effect of Hm and caused a reduction in the NADPH/NADP+ and GSH/GSSG ratios (Fig. S6).

In summary, these results confirm the hypothesis (Fig. 6c) that Hm can regulate metabolic processes via the HIF1-α pathway to promote glucose uptake, which increases the NAPDH/NAPD+ ratio to enhance GR activity and restore the GSH/GSSG levels in BMSCs. Maintaining high levels of GSH is vital for GSH-Px function, which includes scavenging free radicals, lipid peroxidation, and even ferroptosis.

3.6. Hm attenuated MP-induced bone loss in the SONFH rat model

Animal experiments further validated the protective effect of Hm. The SONFH model was established as shown in Fig. 7a. At the end of the intervention, we measured the incidence of SONFH in each group, and the results showed that Hm administration attenuated the incidence rate in a concentration-dependent manner (Fig. 7b). In addition, BTMs in serum were detected after 2 and 4 weeks of Hm intervention, and we found that procollagen I N-terminal propeptide (PINP), an osteoblast metabolite, was lower in the MP group and restored in the Hm group (Fig. 7c). In contrast, C-terminal telopeptides of type I collagen (CTX), a marker of osteoclast activity, was higher in the MP group and decreased in the Hm intervention group in a concentration-dependent manner (Fig. 7d). These results indicated that a high dose of MP could induce bone loss and that Hm had therapeutic effects. Micro-CT analysis further confirmed the therapeutic effect of Hm in SONFH (Fig. 7e-k). Trabecular changes in the subchondral area of the femoral heads were measured for subsequent analysis. Osteonecrosis was not observed in the control group, whereas obvious osteonecrosis was observed in the MP group, and Hm administration attenuated osteonecrosis (Fig. 7e-g). In addition, parameters associated with bone quality, including BMD (Fig. 7h), BV/TV (Fig. 7i), and Tb. Th (Fig. 7j) was markedly lower in the MP group. In contrast, Hm treatment alleviated osteonecrosis and increased the BMD, BV/TV, and Tb. Th (Fig. 7e-k).

Fig. 7.

Harmine alleviated bone loss in SONFH rats. (a) Graphical abstract of the SONFH modeling procedure. (b) Ratio of the rats with significant changes in the femoral head. (n = 8 per group) (c) Serum procollagen I N-terminal propeptide (PINP). (n = 4 per group) (d) Serum C-terminal telopeptides of type I collagen (CTX). (n = 4 per group) (e) 3D reconstruction of femoral heads by micro-CT. (f) Transverse-section view of femoral heads. (g) Coronal-section view. (h) Bone mineral density (BMD, mg/cm3). (n = 8 per group) (i) Bone volume/tissue volume (BV/TV,%). (n = 8 per group) (j) Trabecular thickness (Tb. Th, mm). (n = 8 per group) (k) Trabecular separation (Tb. Sp, mm). (n = 8 per group) (All bar graphs are the mean ± SD; *p < 0.05 and **p < 0.01 compared with the MP group; ns: no significant difference; #p < 0.05 and ##p < 0.01).

Consistent with the micro-CT results, histomorphological analysis with H&E and Masson staining further demonstrated that Hm alleviated SONFH in vivo. H&E staining revealed more bone loss and fewer bone trabeculae in the MP group than in the control group (Fig. 8a, b). Bone area decreased in the MP intervention group and increased in the Hm-low group; furthermore, this protective effect was more significant in the Hm-high group (Fig. 8a, b). Masson staining also showed that Hm could restore the weakened bone-forming ability (Fig. 8c, d). IHC results illustrated that HIF1-α expression decreased in the MP group and was recovered by Hm treatment (Fig. 8e, f). These results suggested that Hm-mediated HIF1-α signaling may be a promising target for preventing femoral head necrosis caused by GC. Finally, H&E staining of the major organs showed that Hm did not show significant toxicity to the lung, liver, kidney, and heart (Fig. S7).

Fig. 8.

Harmine restored bone formation in SONFH rats. (a) H&E staining of the femoral head. (The red dotted line indicates the load-bearing zone in the femoral head. Scale bar: 800 µm.) (b) The ratio of bone area/tissue area was calculated from H&E staining of the femoral head. (c) Masson staining of the femoral head. (Scale bar: 300 µm.) (d) Newly formed bone. (In Masson staining, blue tissue indicates newly formed bone, and red shows mature bone tissue) (e) Immunohistochemical staining of HIF1-α. (Scale bar: 300 µm.) (f) Integrated option density (IOD)/area of HIF1-α. (All bar graphs are the mean ± SD; *p < 0.05 and **p < 0.01 compared with the MP group; ns: no significant difference; #p < 0.05 and ##p < 0.01; n = 8 per group).

4. Discussion

Patients with SONFH often exhibit symptoms such as hip pain, limited mobility and even becoming bedridden in the end stages [30]. Usually, THA is the only method to treat the terminal stage of SONFH, and this treatment has a heavy economic and psychological burden. Therefore, early treatment is particularly important in patients with SONFH to maintain their quality of life. Recent studies have shown that GC-induced osteogenesis abnormalities may be the critical factor in disease development [31,32]. Therefore, the promotion of osteoblastogenesis and bone formation is a potentially promising method to treat SONFH.

Hm, a natural tricyclic β-carboline alkaloid, has multiple pharmacological effects [25]. Our laboratory group previously reported that Hm treatment could inhibit wear particle-induced inflammation and facilitate osteogenic differentiation; these results showed that Hm has a good immunomodulatory effect and helps maintain bone homeostasis [27]. Therefore, we performed experiments to test whether Hm could prevent SONFH in rats. The micro-CT results showed that Hm increased the subchondral bone mass of the femoral heads in SONFH rats and alleviated the decreases in BMD, BV/TV, and Tb. Th. The results of histomorphological analysis also demonstrated that Hm had a protective effect on osteonecrosis with MP intervention. These results suggested that Hm may be a promising candidate for preventing femoral head necrosis caused by GC.

Previous studies have suggested that a high dose of GC inhibits BMSC proliferation, stimulates ROS production, and even leads to cell death, indicating that BMSCs play a critical role in the pathogenesis of SONFH. Based on the results of Masson staining in vivo, Hm treatment greatly enhanced new bone formation. In vitro, ALP, ARS, and immunofluorescence staining showed that the osteogenic potential of BMSCs was markedly recovered after Hm supplementation. Collectively, these data showed that the protective effect of Hm on MP-induced SONFH may be achieved by promoting osteoblastogenesis and enhancing osteogenic formation.

Ferroptosis is a novel type of iron-dependent cell death that is characterized by lipid peroxidation, mitochondrial dysfunction, and eventually, uncontrolled oxidative stress [33,34]. In this study, bioinformatic analysis showed that ferroptosis played a critical role in SONFH. Mitochondrial potential, lipid peroxidation, and cell death detection assays also indicated the occurrence of ferroptosis in BMSCs after high-dose MP intervention. In addition, antioxidants have been used for the treatment of many diseases [35]. Then, we successfully protected BMSCs from ferroptosis by using Hm. With further analysis, we found that DEGs were enriched in the HIF1-α signaling pathway, which is involved in energy metabolism, programmed cell death, and the regulation of gene expression to maintain the redox state when adapting to hypoxia [36]. Western blotting analysis verified that HIF1-α levels were decreased after MP intervention and increased after Hm treatment. However, a previous study showed that Hm decreased HIF1-α accumulation in HepG2 and Hep3B cell lines under hypoxic conditions [37]. We speculated that there may be several reasons for this difference. The main difference is culture conditions: the previous study cultured the cells under hypoxic conditions, while we cultured the cells under normoxic and MP-treated conditions. In addition, the previous study used HepG2 and Hep3B cell lines for their experiments, and we used BMSCs from the femurs and tibias of 6-week-old healthy rats. Finally, harmine was dissolved in DMSO in the previous study, and we used harmine hydrochloride dissolved in double-distilled water to prevent the effects of organic solvents on BMSCs. However, more experiments are needed to further verify the mechanisms responsible for this difference. In subsequent experiments, shHif1a was applied to knockdown HIF1-α for rescue experiments. The results were in line with our expectations and showed that knockdown of HIF1-α could eliminate the protective effect of Hm. These data verified that ferroptosis was closely involved in MP-induced BMSC damage and that Hm suppressed ferroptosis through HIF1-α signaling. However, when BMSCs were treated with Hm alone without using MP, there was no significant change in HIF1-α levels compared with the control group. Therefore, HIF1-α in BMSCs is not affected by Hm under normal conditions, and Hm can promote HIF1-α expression to attenuate injury during high-dose MP intervention. We hypothesized that high doses of MP caused significant changes in multiple signaling pathways, including inhibitory and excitatory pathways of HIF1-α expression. When treated with MP only, the inhibitory signaling pathways were more active, and HIF1-α expression was suppressed; in contrast, Hm treatment reversed the MP-activated HIF1-α inhibitory pathway or the MP-inhibited HIF1-α excitatory pathway. Based on the currently available literature, there are many signaling pathways involved in HIF1-α expression, such as VHL, p53, PI3K, etc. [38,39]. Therefore, future studies need to clarify the interaction network of Hm in BMSCs, and if possible, the use of combined transcriptomic and proteomic analyses would be very useful.

HIF1-α regulates multiple genes, including erythropoietin, glucose transporter protein, glycolytic enzymes, and vascular endothelial growth factor, which contribute to increased oxygen delivery or promote hypoxic metabolism. Previous studies found that in Hif1a–/– mouse embryonic fibroblasts, the fibroblasts die from the failure to switch metabolic status, which leads to excessive production of ROS [40,41]. In addition, HIF1-α has the ability to affect the pentose phosphate pathway (PPP), which further impacts the levels of NADPH and GSH required for ROS scavenging [38]. Therefore, we further elucidated the therapeutic mechanism of Hm by identifying the metabolic status. Although previous studies have not combined metabolic changes with ferroptosis in SONFH, it is known that changes in metabolism are extremely important in the process of ferroptosis and in the pathogenesis of SONFH [42,43]. GSH plays an important role in the maintenance of cellular redox homoeostasis and can reduce lipid peroxidation via the GSH-Px-related peroxide scavenging system [44]. In the current study, MP-induced decreases in total GSH and the GSH/GSSG ratio were restored by Hm, and shHif1a blocked the protective effects of Hm. Moreover, the reduction of GSSG to GSH requires glutathione reductase, which works in the presence of NADPH. Therefore, the NADPH/NADP+ ratio was also measured, and the results were also in line with the GSH changes as follows: Hm increased the ratio of NADPH/NADP+ to promote GSH production.

Huang et al. reported that Hm can promote the secretion of platelet-derived growth factor-BB (PDGF-BB) by preosteoclasts to stimulate endothelial cell angiogenesis in OVX mice [26]. In the current study, IHC results showed that HIF1-α expression was recovered by Hm treatment in SONFH rats and that HIF1-α plays an important role in angiogenesis. According to the in vitro study by Huang et al., direct treatment with Hm inhibited angiogenesis in endothelial cells. Therefore, we speculate that Hm has the ability to enhance angiogenesis in SONFH rats, mainly by promoting the function of other cells and indirectly stimulating angiogenesis. Furthermore, during bone formation, osteogenesis and angiogenesis are coupled together in a closely interdependent manner. BMSCs, osteoblasts, and regulatory T-cell (Treg) cells have been reported to have pro-angiogenic capacity, and these cells play an important role in regulating bone formation [[45], [46], [47]]. Bernard Khor and his colleagues showed that Hm can regulate the differentiation of Treg cells [48]. Therefore, future studies are needed to clarify the relationship between Hm-induced angiogenesis and the functional changes in Treg cells and BMSCs during the treatment of SONFH with Hm.

In addition, there are still unconfirmed parts to be explored. Although Hm increased HIF1-α expression and knockdown of HIF1-α blocked the protective effects of Hm in vitro, conditional knockout of HIF1-α animal models need to be established to further validate the protective effects of Hm in vivo. In view of the complex mechanism of Hm effects, transcriptome analysis and proteome analyses are needed to further clarify the interaction network of Hm.

5. Conclusion

In conclusion, our results demonstrated that Hm can alleviate osteonecrosis of the femoral head in a rat SONFH model. Hm is not only a plant extract but also has an outstanding ability to maintain redox balance and is therefore studied as a potential therapeutic drug for SONFH. Mechanistically, Hm attenuated GC-induced ferroptosis and promoted osteogenic differentiation by promoting HIF1-α expression to regulate the GSH/GSSG and NADPH/NADP+ ratios. Therefore, Hm might be a suitable candidate for the prevention and treatment of SONFH.

CRediT authorship contribution statement

Conceptualization and study design: Wenming Li; Methodology: Wei Zhang; Acquisition of data: Wenming Li, Wenhao Li, Yi qin; Analysis and interpretation of data: Houyi Sun; Writing-review & editing and/or revision of the manuscript: Liangliang Wang, Jiaxiang Bai, Dechun Geng; Funding acquisition: Liangliang Wang, Dechun Geng.

Declaration of competing interest

The authors declare that they have no conflicts of interest in this work.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (82002321, 82072425, 82072498, 82272567), the Natural Science Foundation of Jiangsu Province (BK2021650), the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), the Jiangsu Commission of Health (H2019024, LK2021017), the Gusu Health Talent Program of Suzhou (GSWS2022002), Jiangsu Medical Research Project (ZD2022014), National and Local Engineering Laboratory of New Functional Polymer Materials (SDGC2205), the Special Project of Diagnosis and Treatment Technology for Key Clinical Diseases in Suzhou (LCZX202003), the Changzhou High-Level Medical Talents Training Project (2022CZBJ060), the funding from Young Talent Development Plan of Changzhou Health Commission (CZQM2021012), the National Postdoctoral Program for Innovative Talents (BX20230350), Research Funds of Centre for Leading Medicine and Advanced Technologies of IHM (2023IHM02007), the Foundation of National Center for Translational Medicine (Shanghai) SHU Branch (SUITM-202301), and China Orthopaedic and Traumatology Young Physicians Innovation and Development Program (GSKONJJ-2023-003).

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, and further inquiries can be directed to the corresponding authors.

Biographies

Wenming Li received his B.Sc. in clinical medicine from Tongji Medical College, Huazhong University of Science and Technology in 2019, and is pursuing master's degree at Soochow University. His current research interests focus on (1) research on the pathogenesis and treatment of osteoporosis, and (2) application and development of traditional Chinese Medicine onomer in osteoporosis.

Wei Zhang received his B.Sc. in clinical medicine from Jiangsu University in 2014. He obtained his M.D. in surgery from Soochow University in 2022. He then joined Joint Degeneration and Injury Laboratory in Institute of Orthopaedics at Soochow University. His current research interests focus on (1) research into the pathogenesis and treatment of osteoporosis, and (2) the prevention and treatment of peri‑prosthetic osteolysis.

Houyi Sun (BRID: 01339.00.39506) received Ph.D. degree from Soochow University in 2022. Then he joined Qilu Hospital of Shandong University and start his post-doctoral work. His research focuses on (1) metabolic disorders and therapeutic strategies in glucocorticoid-induced osteonecrosis of femoral head; (2) bone and cartilage repair in osteoarthritis; (3) novel ideas in navigation-assisted arthroplasty.

Liangliang Wang (BRID: 07100.00.71336) received his B.Sc. in clinical medicine from Soochow University in 2015. He obtained his M.D. in surgery from Soochow University in 2018. He then joined Department of Orthopedics, The Affiliated Changzhou Second People's Hospital of Nanjing Medical University. His current research interests focus on (1) research into the pathogenesis and treatment of peri‑prosthetic osteolysis, and (2) the prevention and treatment of rheumatoid arthritis.

Jiaxiang Bai received his B.Sc. in clinical medicine from Bengbu Medical College in 2017. He obtained his Ph.D. in Surgery from Soochow University in 2022. He then joined Department of Orthopedics, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China. His current research interests focus on (1) research into the pathogenesis and treatment of degenerative diseases, and (2) the prevention and treatment of implant-associated infections.

Dechun Geng received Ph.D. degree from Soochow University in 1999 and joined Department of Orthopedics, the First Affiliated Hospital of Soochow University. Currently Dr. Geng is the principal investigator of Joint Degeneration and Injury Laboratory in Institute of Orthopaedics at Soochow University with an H-index of 30. He won the second prize of The State Scientific and Technological Progress Award in 2017, and presided over more than 10 various topics including the National Natural Science Foundation of China and Distinguished Youth Foundation of Jiangsu Scientific Committee. His research focuses on (1) signaling pathways in bone metabolism and targeted therapeutic strategies; (2) bone immune regulation in bone metabolic diseases; (3) implants and osseointegration.