To the Editor: Steroid-induced osteonecrosis of the femoral head (SIONFH) is a debilitating orthopedic disorder characterized by progressive hip pain, irreversible femoral head collapse, and eventual disability,1 imposing considerable physical and socioeconomic burdens on patients. Although its precise pathogenesis remains incompletely understood, accumulating evidence highlights two interrelated drivers: excessive osteoclast activity and overproduction of reactive oxygen species (ROS). A previous study2 has reported elevated osteoclast activation and reduced expression of antioxidant enzymes in necrotic regions of the femoral head. This imbalance disrupts bone remodeling homeostasis, while insufficient antioxidant defense fails to mitigate ROS-induced cellular injury, thereby accelerating necrosis and collapse. Under oxidative stress, Kelch-like ECH-associated protein 1 (Keap1), a redox-sensitive biosensor, undergoes degradation, enabling nuclear factor erythroid 2-related factor 2 (Nrf2) to translocate into the nucleus3 and bind to antioxidant response elements (AREs), subsequently inducing transcription of antioxidant genes to restore ROS-mediated osteoclast homeostasis.4 Given this context, preserving Keap1 function to sustain Nrf2-mediated antioxidant responses has emerged as a promising therapeutic strategy for SIONFH. Loureirin B (LrB), a natural compound derived from Sanguis draxonis, has previously been shown to inhibit receptor activator of nuclear factor-κB ligand (RANKL)-induced osteoclast differentiation and reduce intracellular ROS levels.5 However, its precise molecular target and in vivo therapeutic role in SIONFH have remained unclear. This study systematically investigated the effects of LrB on SIONFH, focusing on its interaction with Keap1 in regulating ROS-mediated osteoclastogenesis. To evaluate the effects of LrB on osteoclast formation and function, primary bone marrow macrophages (BMMs) were isolated from C57BL/6 mice and seeded onto sterilized cow bone slices. Cells were stimulated with 50 ng/mL of RANKL to induce osteoclast maturation, with LrB administered at 0, 5, or 10 μmol/L. Tartrate resistant acid phosphatase (TRAcP) staining revealed that LrB reduced the number of multinucleated osteoclasts in a dose-dependent manner, with the 10 μmol/L group showing an ~71.6% reduction Supplementary Figure 1A and B, https://links.lww.com/CM9/C773, while MTS assay indicated no significant cytotoxicity Supplementary Figure 1C, https://links.lww.com/CM9/C773. Following TRAcP staining, osteoclasts were removed and resorption pits were visualized under an inverted microscope. LrB treatment markedly reduced both the size and number of resorption pits (yellow arrows, Supplementary Figure 1D, https://links.lww.com/CM9/C773). F-actin ring formation, a cytoskeletal structure essential for osteoclast adhesion and bone resorption, was further examined by confocal microscopy after costaining with rhodamine-phalloidin (F-actin), anti-vinculin antibody (adhesion sites), and Hoechst 33258 (nuclei). Compared with the RANKL group, which exhibited large, intact F-actin rings with multiple nuclei, 10 μmol/L LrB significantly diminished F-actin ring size and nucleus number Supplementary Figure 1E and F, https://links.lww.com/CM9/C773. Consistently, Western blotting of total protein from osteoclast precursors, multinucleated osteoclasts, and mature osteoclasts confirmed that 10 μmol/L of LrB suppressed expression of nuclear factor of activated T cells 1 (NFATc1) (a master transcriptional regulator) and cathepsin K (CTSK) (a key resorptive enzyme) Supplementary Figure 1G, https://links.lww.com/CM9/C773. Together, these findings demonstrate that LrB inhibits osteoclast formation and bone resorptive function without cytotoxicity. Given that ROS overproduction is a key driver of osteoclast hyperactivity in SIONFH, this study examined whether LrB modulates intracellular ROS levels and antioxidant responses. Osteoclast precursors were stimulated with RANKL in the presence or absence of LrB for 48 h. Intracellular ROS levels were measured by flow cytometry using the fluorescent probe H2DCFDA (10 μmol/L). Compared with untreated controls, RANKL stimulation induced an ~3-fold increase in fluorescence intensity, whereas LrB treatment reduced ROS to near baseline levels Supplementary Figure 2A and B, https://links.lww.com/CM9/C773. To elucidate the molecular basis of this antioxidant effect, reverse transcription quantitative polymerase chain reaction (RT-qPCR) analysis was performed. LrB dose-dependently downregulated the expression of osteoclastic genes (Acp5, Nfatc1, Ctsk, and Atp6v0d2) and upregulated antioxidant genes (HO-1 and Cat) relative to the RANKL group Supplementary Figure 2C and Supplementary Table 1, https://links.lww.com/CM9/C773. Consistently, Western blotting confirmed that LrB enhanced heme oxygenase 1 (HO-1) and catalase expression during osteoclast differentiation Supplementary Figure 2D and E, https://links.lww.com/CM9/C773. Together, these findings indicate that LrB suppresses osteoclastogenesis by reducing ROS levels and promoting antioxidant enzyme expression. To identify the specific molecular target of LrB underlying its antioxidant and anti-osteoclast effects, a combination of surface plasmon resonance (SPR) and liquid chromatography mass spectrometry (LC-MS) was used for target fishing. LrB was immobilized on a 3D photo-cross-linking SensorChip, and BMM lysates were perfused over the surface. Real-time binding was monitored by SPR, and LC-MS analysis identified 92 putative target proteins. KEGG pathway enrichment revealed 15 ROS-related candidates Supplementary Figure 3A, https://links.lww.com/CM9/C773, and the top 10 were selected based on SPR binding intensity. The 3D structures of these proteins were obtained from the RCSB Protein Data Bank (https://www.rcsb.org/) and processed using PyMOL (Schrödinger, LLC., New York, United States). The 3D structure of LrB was constructed in Chem3D (PerkinElmer Informatics, Inc., Waltham, Massachusetts, United States), optimized via energy minimization, and saved in mol2 format. Rigid molecular docking with LeDock (LePhar, Inc., Shanghai, China) demonstrated binding energies <0 for all 10 proteins, with Keap1 showing the lower binding energy and stronger affinity for LrB Supplementary Figure 3B, https://links.lww.com/CM9/C773. Molecular dynamics simulations further confirmed the stability of the LrB–Keap1 complex Supplementary Figure 3C, https://links.lww.com/CM9/C773. These findings suggest that LrB may competitively bind Keap1, thereby modulating the Keap1/Nrf2 signaling pathway. To validate whether LrB directly modulates the Keap1/Nrf2 axis, a series of mechanistic experiments was conducted. RT-qPCR of BMMs treated with RANKL and LrB for 48 h showed that LrB upregulated both Keap1 and Nrf2 mRNA levels Supplementary Figure 4A, https://links.lww.com/CM9/C773. A luciferase reporter assay further confirmed that LrB activated ARE-driven transcription Supplementary Figure 4B, https://links.lww.com/CM9/C773. To assess the dynamic regulation of Keap1 and Nrf2 proteins, time-course Western blotting was performed at 0, 4, 8, 12, 24, 36, and 48 h. RANKL stimulation induced Keap1 degradation at 4 h (stabilized after 36 h) and reduced Nrf2 levels at 4 h (with recovery by 36 h). In contrast, LrB treatment reversed Keap1 degradation and progressively increased Nrf2 protein levels over time Supplementary Figure 4C and D, https://links.lww.com/CM9/C773. Subcellular localization of Keap1 and Nrf2 was examined by immunofluorescence costaining. BMMs seeded on confocal dishes were stimulated with RANKL and LrB for 4 h and 36 h, followed by fixation, permeabilization, and blocking. Cells were incubated with Keap1 and Nrf2 antibodies, then probed with Alexa Fluor-488- and Alexa Fluor-596-conjugated secondary antibodies, and counterstained with DAPI. Confocal microscopy confirmed that LrB prevented Keap1 degradation and promoted nuclear translocation of Nrf2 Supplementary Figure 4E, https://links.lww.com/CM9/C773. Collectively, these findings demonstrate that LrB stabilizes Keap1 protein and enhances Nrf2 accumulation and nuclear activity. To confirm the therapeutic effect of LrB in vivo, a rat model of SIONFH was established by subcutaneous administration of imiquimod (30 mg/kg) and intragluteal injection of methylprednisolone (20 mg/kg). The model + LrB group received intraperitoneal injections of LrB (4 mg/kg) every 2 days for 6 weeks, whereas the vehicle and model groups were given equivalent volumes of normal saline Supplementary Figure 5A, https://links.lww.com/CM9/C773. Weekly weight monitoring revealed no significant differences among groups Supplementary Figure 5B, https://links.lww.com/CM9/C773. After sacrifice, femurs were harvested for micro-CT analysis of bone microarchitecture. The region of interest was defined as a 0.25-mm-thick segment located 0.5 mm above the femoral head epiphysis Supplementary Figure 5C, https://links.lww.com/CM9/C773. Two-dimensional reconstructions were generated using CTvol and Data-viewer softwares (Bruker, Kontich, Antwerp, Belgium), and bone morphometric parameters were quantified with CT Analyzer software (Bruker, Kontich, Antwerp, Belgium). The model group displayed severe bone microarchitectural destruction, characterized by reduced trabeculae and increased empty spaces, whereas the model + LrB group exhibited marked improvement Supplementary Figure 5D and E, https://links.lww.com/CM9/C773. Following micro-CT analysis, left femurs were fixed, decalcified, paraffin-embedded, and sectioned at 5 μm. Hematoxylin and eosin staining of subchondral bone revealed numerous empty lacunae (yellow arrows, a hallmark of osteocyte necrosis) in the model group, while the model + LrB group demonstrated a ~90% reduction Supplementary Figure 5F, https://links.lww.com/CM9/C773. Together, these results confirm that LrB effectively attenuates SIONFH-induced bone destruction and osteocyte necrosis in vivo. To further elucidate the in vivo mechanism of LrB, osteoclast activity, ROS levels, and protein expression in femoral heads and serum were analyzed. TRAcP staining of femoral head sections demonstrated a marked reduction in both osteoclast number and bone surface area occupied by osteoclasts in the model + LrB group compared with the model group Supplementary Figure 6A and B, https://links.lww.com/CM9/C773. ROS levels were assessed by intraperitoneal injection of dihydroethidium, followed by femur collection and preparation of 5 μm frozen sections counterstained with DAPI. Confocal microscopy revealed strong red fluorescence in the subchondral bone of the model group, accompanied by hypochromatosis (yellow arrows), whereas the model + LrB group displayed significantly weaker fluorescence and less hypochromatosis Supplementary Figure 6C and D, https://links.lww.com/CM9/C773. Protein expression analyses further supported these findings. Western blotting of femoral head powder showed that LrB decreased CTSK levels while increasing HO-1 and catalase expression (Supplementary Figure 6E and F, https://links.lww.com/CM9/C773). In addition, ELISA assays of serum demonstrated significantly reduced Acp5 and CTSK, along with elevated Keap1, NFE2L2, and HO-1 levels in the model + LrB group compared with the model group Supplementary Figure 6G, https://links.lww.com/CM9/C773. In conclusion, excessive ROS production and osteoclast hyperactivity are key contributors to the pathogenesis of femoral head necrosis. These in vivo findings, consistent with in vitro results, demonstrate that the natural small-molecule compound LrB binds to Keap1, establishing a positive feedback mechanism that suppresses ROS-mediated osteoclast activity. These data highlight the therapeutic potential of LrB for the treatment of SIONFH Figure 1.Figure 1: The molecular mechanism of LrB mitigating SIONFH. Hyperactive osteoclasts triggered by elevated ROS levels accelerate the progression of SIONFH. As a natural molecular drug, LrB can bind with Keap1 protein to block its degradation and concurrently facilitate the nuclear translocation of Nrf2. This promotes the expression of antioxidant proteins, effectively reducing ROS and ROS-mediated osteoclast activities. Consequently, this mechanism helps counteract the progression of SIONFH. ARE: Antioxidant response element; Keap1: Kelch-like ECH-associated protein 1; LrB: Loureirin B; Nrf2: nuclear factor erythroid 2-related factor 2; OC: Osteoclast; ROS: Reactive oxygen species; SIONFH: Steroid-induced osteonecrosis of the femoral head.Funding This research was supported by grants from the National Natural Science Foundation of China (No. 82104883), Guangdong Basic and Applied Basic Research Foundation (No. 2025A1515012848), Guangzhou Science and Technology Bureau (No. 2024A04J9998), University-Hospital Joint Fund Project of Guangzhou University of Chinese Medicine (No. GZYZS2024G08), the talent cultivation program from the Guangzhou Association for Science and Technology (No. QT-2024-033), the talent cultivation program from the Guangdong Association for Science and Technology (No. SKXRC2025259) and the talent cultivation programs from the First Affiliated Hospital of Guangzhou University of Chinese Medicine (Nos. ZYYY2023-186 and ZYYY2024-87). Conflicts of interest None. Data availability The data that support the findings of this study are available from the corresponding authors upon reasonable request.
Liu et al. (Wed,) studied this question.