Dexamethasone reduces osteogenic potential through suppressing retinoic acid signaling via NF-κB/CYP26B1 axis in mesenchymal stem cells Authors: Ai-Hua Ye, Yu-Mei Wang, Xin-Yue Wan, Jia Liu, Ni Tang, Wen-Ting Liu, Jie Cai, Dong-Mei He, Bai-Cheng He, Wen-Ge He, Liang Chen Published: 2026 ABSTRACT AbstractObjectiveGlucocorticoid-induced osteoporosis (GIOP) is the most common secondary osteoporosis, which characterized by the decreased bone strength and increased fracture risk. But the pathogenic causes of GIOP remains not completely yet. In this study, we investigated the effect of Cyp26b1, a key enzyme for all-trans retinoic acid (ATRA) degradation, on the osteogenic inhibitory effect of dexamethasone (DEX) in mesenchymal stem cells (MSCs) and revealed the possible mechanism through which DEX regulates the expression of Cyp26b1.MethodsMSCs were induced with osteogenic induction medium in vitro to establish the osteogenic model, and high concentration (10− 6 M) of DEX was utilized to treat MSCs to exert its osteogenic inhibitory effect. ATRA content was detected with ELISA, real-time polymerase chain reaction (RT-PCR), western blot, and immunofluorescence staining were employed to detect hall markers for osteogenic differentiation. Chromatin immunoprecipitation (ChIP) assay was used to detect the binding of promoter region of target gene. The rat GIOP model was constructed by intraperitoneal injection of DEX, and the adeno-associated virus was injected via tail vein after the model was constructed. The bone samples were subjected with µ-CT scanning, H&E, and Masson trichrome staining, respectively.ResultsDEX (10− 6 M) decreased the levels of osteogenic markers induced by osteogenic induction medium (OIM) in MSCs, and it was almost reversed by ATRA. But this effect of ATRA was obviously eliminated by overexpression of Cyp26b1. The mRNA expression level of Cyp26b1 was significantly increased by DEX (10− 6 M), and the content of retinoic acid was decreased in cells. Osteogenic markers decreased by DEX (10− 6 M) were recovered by Cyp26b1 knockdown, whereas Cyp26b1 overexpression exerted the opposite effect. DEX (10− 6 M) increased the activation of NF-κB signaling, and the osteogenic inhibitory effects of DEX were ameliorated by NF-κB inhibitors which was almost abolished by Cyp26b1 overexpression. The mRNA level of Cyp26b1 was up-regulated by DEX (10− 8 M and 10− 6 M), but the degradation of Cyp26b1 was promoted by DEX (10− 8 M) only via the lysosomal pathway.ConclusionThe osteogenic inhibitory effect of high concentration of DEX may be resulted from the disturbance of retinoic acid signaling. The expression of Cyp26b1 can be up-regulated by either low or high concentration of DEX, but its degradation can only be inhibited by high concentration of DEX. The effect of DEX on Cyp26b1 expression was partially mediated by activating NF-κB signaling in MSCs.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s13287-026-05131-0.Keywords: Glucocorticoid-induced osteoporosis, Retinoic acid signaling, Cytochrome P450 family 26 subfamily B member 1, Nuclear factor-kappa B, Mesenchymal stem cells Introduction Bone homeostasis depends on a dynamic balance between bone formation and bone resorption, which is well regulated and osteoporosis may happen if the balance was disrupted. Glucocorticoids (GCs) are widely used as an anti-inflammatory drug for the treatment of autoimmune diseases, asthma, rheumatoid arthritis, and many other conditions [1, 2]. However, long-term use of GCs can cause serious adverse effects, such as glucocorticoid-induced osteoporosis (GIOP), the most common secondary osteoporosi [3–5]. The pathogenic causes of GIOP are very complex, and the main pathogenesis is characterized by a persistent decrease of bone formation and an increase of bone resorption [5, 6]. The biological effects of GCs are mainly mediated through their receptor (Glucocorticoid receptor, GR) [7]. For bone metabolism, low concentrations of GCs, including DEX, have been reported to enhance the ALP activity and matrix mineralization level in primary osteoblast, whereas high concentration of GCs exhibits the opposite effect [8–10]. In additional, the inhibitory effect of high concentration of GCs on bone formation may also contribute to inhibit the proliferation, induce apoptosis of osteoblasts [11–13], and promote differentiation and maturation of osteoclast by increasing the ratio of receptor agonists for nuclear factor-κβ ligand to osteoprotegerin to promote bone resorption [14–16]. The current conventional therapy strategy for GIOP is not optimal. Thus, the pathogenic causes of GIOP need to be further elucidated, so as to improve the treatment for GIOP. All-trans retinoic acid (ATRA) is one of the major active metabolites of vitamin A. Our previous study demonstrated that it can almost rescue the osteogenic detrimental effect of DEX in vivo and in vitro [17], although the concrete mechanism remains unclear. Cyp26b1, a key metabolizing enzyme for converting ATRA to hydroxyretinoic acid, plays an important role in embryonic development and multiple disease processes [18]. Its expression state is closely associated with skeletal abnormalities [19]. During the development of zebrafish and mice, Cyp26b1 can affect the osteoblast maturation and ossification by regulating retinoic acid concentration [20]. The BMP9-induced osteogenic markers were significantly reduced by Overexpression of Cyp26b1 in MSCs [21]. Although Cyp26b1 plays an important role in regulating bone metabolism, its role in DEX-induced osteogenic impairment, a representative model of glucocorticoid-induced bone damage, remains unclear. Nuclear factor-kappa B (NF-κB) signaling pathway is widely involved in various physiological and pathological processes, such as inflammatory response, cell proliferation, differentiation and apoptosis [22, 23]. The activation of NF-κB signaling is closely related to the IκBα and p65 subunits. When the cell receives a stimulus, the IKK complex is activated and phosphorylates IκBα, and the p65 subunit is released. Then, the released p65 subunit is translocated to the nucleus, and regulate the expression of downstream targets by incorporating with other factors [24–26]. The role of NF-κB signaling in GIOP is not well defined. As reports, tuna powder can alleviate GIOP by regulating NF-κB and Wnt/β signaling pathways, α-lipoic acid can block GIOP and promote bone formation by inhibiting the GC-induced NF-κB activation [27]. Atractylenolide I can inhibit the RANKL-induced osteoclast formation and block the glucocorticoid-induced bone loss via inhibition of NF-κB signaling [28]. In addition, Klotho can also prevent the DEX-induced apoptosis of MC3T3-E1 cells through NF-κB signaling [29]. Therefore, NF-κB activation may be involved in DEX-induced osteogenic impairment and the pathological process of GIOP. In this study, we investigated the role of Cyp26b1 in DEX-induced osteogenic impairment and explored the underlying molecular mechanism, aiming to identify potential therapeutic targets for improving the treatment of GIOP. Materials and methods Cell culture and chemicals HEK293, C3H10T1/2, C2C12, MC3T3-E1, and 3T3-L1 cells were ordered from ATCC and stored in our laboratory. Cells were cultured in DMEM containing glucose (4.5 g/L), 10% fetal bovine serum, penicillin (100 U/mL), and streptomycin (MC3T3-E1 and 3T3-L1 were cultured in α-MEM) at 37 °C with 5% CO₂. Primary antibodies targeting β-Actin (AC038), p65 (A19653), p-p65 (AP0124), IKBα (A1187) and p-IKBα(A11397) were purchased from ABclonal (Wuhan, China); RUNX2 (sc-101145), OPN (sc-21742) and Cyp26b1 (sc-193493) were procured from Santa Cruz Biotechnology. Dexamethasone (ST1254) and BAY11-7082 were bought from Beyotime (Shanghai, China). Reverse transcription and polymerase chain reaction (RT-PCR) assay Total RNA was extracted with RNA-Quick Purification kit (RN001, ES Science, Shanghai, China), and the quality and concentration of RNA were detected using NanoDrop one instrument (Thermo Fisher Scientific, USA). The cDNA was obtained by reverse transcription using RT Kits (RR037A, Takara), and the product obtained was diluted 10-fold as template for the following PCR assay, utilizing a 2xSYBR Green qPCR Master Kit (B2120, Selleck, USA) and a BioRad CFX Connect System (Bio -Rad, Chicago, USA). Thermal cycling conditions were set as following: 95 °C for 3 min, followed by 40 cycles, each cycle including 95 °C denaturation for 10 s, 58 °C annealing for 30 s, and 72 °C extension for 10 s. The relative gene expression levels were finally calculated based on the 2−ΔΔCt method and normalized with β-actin. The primer sequences for the target genes are shown in Table 1. Table 1Primers used in this studyGenePrimerSequence (5’−3’) β-actin FCCACCATGTACCCAGGCATTRCGGACTCATCGTACTCCTGC Runx2 FGCCAATCCCTAAGTGTGGCTRAACAGAGAGCGAGGGGGTAT Opn FTGCACCCAGATCCTATAGCCRCTCCATCGTCATCATCATCG Cyp26a1 FATACCAGATCCCCAAGGGCTRTTGCCTACACAGCTCCGAAG Cyp26b1 FTGAAAAAGGTGCATGCTGCCRAGAAATCTGGGCCGCATCAA Cyp26c1 FCCTCTTCTCACTGCCCTTGGRATCAAGAGCAGGGCATCACC Raldh1 FGGCCAAATCATCCCCTGGAARATGTTTACCACGCCAGGAGG Raldh2 FGCTGACTTGGACTACGCTGTRTTGTGGGATCAAAGGGGCTC Raldh3 FGGAGCAACCTGAAGAGGGTCRGACCTGCTCTTCCACGAACA(F: forward, R: reverse) Primers used in this study (F: forward, R: reverse) Protein collection and immunoblot assay Cells were first seeded in six-well plates and different treatment reagents were added according to the experimental design. At the end time points, plates were washed twice with cold phosphate-buffered saline (PBS, 4°), and placed on ice. Cells were lysed for 10 min by adding 300 µL RIPA buffer containing 1% 1PMSF. For bone tissue, the bone masses were cut up with bone shears and washed with cold saline to remove blood. Then, liquid nitrogen was added to freeze the species for pulverization, and the bones were ground with a mortar and pestle until pulverized. The lysis buffer was added at RIPA lysate (ml): bone weight (g) = 10:1, and lysed on ice for 10 min. The lysates were then collected in 1.5 ml EP tubes, loading buffer was added to each tube at a ratio of 4:1, and the samples were boiled in a metal bath at 100 °C for 15 min. Samples were electrophoresed on 10–12.5% SDS-PAGE gels and subsequently transferred to PVDF membranes, and then the membranes were blocked with protein-free rapid blocking solution. The corresponding primary antibodies were incubated overnight at 4 °C. The membranes were washed four times with TBST and then incubated with the corresponding horseradish peroxidase-coupled secondary antibody for 1 h. After four additional washes, the membranes were developed using a chemiluminescent substrate kit. Alkaline phosphatase (ALP) activity assay Cells were seeded in 24-well plates with the appropriate treatment reagents for 5 or 7 days. At the end time points, cells were washed twice with PBS and fixed with 4% paraformaldehyde for 10 min. The ALP activities were subsequently detected using the BCIP/NBT ALP Chromogenic Kit (C3206, Beyotime, Shanghai, China). The plates were scanned with a scanner (Phentom, China). Images were acquired under microscope (IX53, Olympus, Tokyo, Japan). Each assay was repeated at least in three independent experiments. Quantitative analysis was performed using ImageJ software. Mineralization assay Cells were seeded in 24-well plates and treated with osteogenic induction medium and other corresponding reagents for 14 or 21 days. The medium was discarded and the cells were gently rinsed twice with PBS (pH 4.2). Then, the cells were fixed with 4% paraformaldehyde for 10 min. Discarded the paraformaldehyde solution, the cells were rinsed once more with PBS, and the cells were stained with 0.4% alizarin red solution (G8550, Solarbio, Beijing, China) for 20 min. After washing with PBS, the plates were scanned with with a scanner (Phentom, China), and images were captured under microscope (IX53, Olympus, Tokyo, Japan). Each assay was repeated at least in three independent experiments. Quantitative analysis was performed using ImageJ software. Chromatin immunoprecipitation (ChIP) assay ChIP assay was performed using the ChIP assay kit (P2083S, Beyotime, Shanghai, China). Briefly, cells were seeded in T75 and treated with the corresponding reagents for 24 h. Cells were crosslinked with 1% formaldehyde for 10 min, and terminated by adding glycine to a final concentration of 125 mM. Subsequently, pre-cooled protease inhibitor-containing PBS was added. Then, the cells were collected with a cell scraper and centrifuged, and the supernatant was discarded. DNA fragments of 250 bp to 500 bp were obtained by sonication of the samples. DNA-protein complexes were enriched using GR primary antibody. Rabbit lgG was used as a negative control. PCR was performed to detect the promoter fragment of Cyp26b1. PCR primers used for this assay are listed in the Table 2. Table 2Primer sequences used for ChIP assay in this studyGenePrimerSequence (5’−3’) Primer 1 FCCTGCCTGCCTGTACATGAARACCTTGTCTCCACCAACGTC Primer 2 FCACTCTATCCCCTGTCCCCARAGCCAGCTCTCTTTAGGGGA Primer 3 FCTGGCGTAGGGACTGGAAAGRTGGTGTCACAATCTGCCTCC(F: forward, R: reverse) Primer sequences used for ChIP assay in this study (F: forward, R: reverse) Immunofluorescence analysis Coverslips were first placed in 12-well plates, and then cells were seeded on it and treated with the appropriate reagents. After 24 h, the medium was discarded, washed three times with PBS, and then fixed with paraformaldehyde for 20 min. Subsequently, the cells were permeabilized and blocked with QuickBlockTM Immunostaining Blocking Solution for 20 min, and then washed three times with PBS. Then, cells were incubated with primary antibody overnight at 4 °C, and then washed three times with PBS. Cells were then incubated with second antibody conjugated with Alexa Fluor 488-labeled goat anti-mouse lgG (GB25301, Servicebio, Wuhan, China) or Alexa Fluor 594-labeled goat anti-rabbit lgG (GB28301, Servicebio, Wuhan, China) for 1 h at room temperature with light protection. Then, cells were washed three times with PBS and stained with 4,6-diamidino-2-phenylindole (DAPI) for 10 s. Finally, the samples were washed three times with PBS and blocked with anti-fluorescence quencher (P0126, Beyotime, Shanghai, China). Images were taken with a microscope (Ci-L, Nikon, Tokyo, Japan). Construction of rat GIOP model and delivery of AAV The animal experiments in this study were approved by the Ethics Committee of Chongqing Medical University (No. IACUC-CQMU-2025-0049) and were conducted in accordance with the ARRIVE guidelines 2.0. Twenty 6-week-old male Sprague Dawley (SD) rats (150–180 g, n = 20) were ordered from the Experimental Animal Center of Chongqing Medical University. The rats were randomly divided into four groups and five rats each group (one for control group and three for GIOP model). The rats were housed in a specific pathogen-free (SPF) environment with a 12-hour light/dark cycle, controlled temperature (25℃), humidity (40–60%), and free access to food and water. After one-week acclimatization, the rats of model groups were injected dexamethasone sodium phosphate (5 mg/kg) intraperitoneally, once a day for one month; the rats of control group were injected with an equal volume of saline at the same time. Four weeks after the model was established successfully, the GIOP experimental group was further divided into PBS group, AAV-vector group (virus control group), and AAV-siCyp26b1 group. 200 µl AAV-Cyp26b1-RNAi (2 × 1013 v.g/ml) was injected intravenously from the tail of rat for the AAV-siCyp26b1 group, and 200 ul AAV-NC (2 × 1013 v.g/ml) was injected intravenously from tail for AAV-vector group. 200 ul of PBS was injected intravenously for the GIOP group and the control group. Four weeks later, euthanasia was performed on the SD rats via cervical dislocation under deep anesthesia induced by 5% isoflurane. The depth of anesthesia was confirmed by the absence of the plantar reflex. Death was verified by the absence of spontaneous respiration, heartbeat, and corneal reflex for at least 2 min. Subsequently, serum, femurs, and tibias were collected for subsequent experiments. ELISA assay The serum was collected by centrifugation of the collected blood from SD rats at 12,000 g for 10 min to collect the serum. For the cells, the suspension was diluted with PBS and treated by repeating freeze and thaw in order to disrupt and release intracellular components. Then, centrifugation was performed at 3000 g for 20 min to collect the supernatant. Finally, ELISA assay was performed according to the manufacturer’s instructions of the Kit (Saimike, China). Histological evaluation and staining Samples were fixed with 4% paraformaldehyde for three days, decalcified with 10% EDTA solution for two months, dehydrated in ethanol, and then embedded in paraffin. Samples were cut into serial 5 μm thick slices, and subjected to regular trichrome Masson staining or immunofluorescent staining. Micro–computed tomography (µCT) analysis Rat femur specimens were subjected to micro-computed tomography (µCT) analysis using a SkyScan 1276 system (Bruker, Germany). Before scanning, the femurs were carefully dissected, cleaned of surrounding soft tissues, and fixed in 4% paraformaldehyde for 48 h. The specimens were then scanned without decalcification. Data were reconstructed and quantified with software provided by the manufacturer (software version 1.20.3.0, Bruker, Germany). The quantitative parameters include bone mineral density (BMD), bone volume fraction (BV/TV), bone surface/bone volume ratio (BS/BV), number of trabeculae (Tb.Tn), and spacing of trabeculae (Tb.Sp). Statistical analysis Data in this study were expressed as mean ± standard error, and each assay was repeated at least three times independently. GraphPad Prism 9.5.0 software was utilized to analyze the data, and images were quantified with Image J software. The two-tailed Student’s t-test was introduced to analyze the differences between two groups, and one-way ANOVA was utilized for the comparison between two or more groups. The difference is considered statistically significant when p value less than 0.05 is available. Results Effect of ATRA on the osteogenic inhibitory activity of DEX in C3H10T1/2 cells Glucocorticoids (GCs) are widely used in the clinic for anti-inflammatory, but long-term use may induce osteoporosis. Dexamethasone (DEX), one of the most common GCs, exhibits a dose-dependent effect on osteogenic differentiation effect of MSCs. In this study, we choose the concentration of 10− 6 M DEX for the in vitro study to analyze its osteogenic inhibitory effect. We firstly recaptured the effect of DEX (10− 6 M) on the osteogenic differentiation in MSCs. To evaluate osteogenic differentiation at different stages, RUNX2 was examined on day 1 and 2 as an early osteogenic marker and key transcription factor involved in osteogenic commitment, whereas OPN was examined on day 9 and 11 as a later-stage matrix-associated osteogenic marker. RT-qPCR and western blot assay results exhibit that DEX can significantly reduce the mRNA and protein levels of RUNX2 (Fig. 1A, B). Consistently, alkaline phosphatase staining showed that DEX reduced ALP activity (Fig. 1C). In addition, DEX decreased the mRNA and protein levels of OPN (Fig. 1D, E), and Alizarin Red S staining results exhibited that the matrix mineralization was reduced (Fig. 1F). CCK8 assays confirmed that DEX at 10⁻⁸ M, 10⁻⁷ M, and 10⁻⁶ M did not significantly affect cell viability at 24–48 h. At 72 h, 10⁻⁶ M DEX caused a modest decrease, but no significant difference was observed between the 10⁻⁶ M and 10⁻⁷ M DEX groups, the latter being the standard osteogenic induction condition, indicating that the osteogenic inhibition by 10⁻⁶ M DEX is not attributable to nonspecific cytotoxicity (s-Fig. 1A). Next, we determined the effect of ATRA on its osteogenic inhibitory effect of DEX. RT-qPCR and western blot assay results exhibit that ATRA can partially reverses the decrease of RUNX2 (Fig. 1G, H). Alkaline phosphatase staining further showed that ATRA alleviated the DEX-induced reduction in ALP activity (Fig. 1I). Furthermore, ATRA partially reversed the downregulation of OPN mRNA and protein levels (Fig. 1J, K), and attenuated the reduction in matrix mineralization (Fig. 1L) in C3H10T1/2 cells. These results indicate that the osteogenic inhibitory effect of DEX can be partly reversed by ATRA in MSCs. Effect of ATRA on the osteogenic inhibitory activity of DEX in C3H10T1/2 cells. A PCR results show the effect of DEX (10− 6 M) on Runx2 mRNA levels. B Western blot results show the effect of DEX (10− 6 M) on RUNX2 protein levels. C ALP staining results show the effect of DEX (10− 6 M) on ALP activities. D PCR results show the effect of DEX (10− 6 M) on Opn mRNA levels. E Western blot results show the effect of DEX (10− 6 M) on OPN protein levels. F Alizarin red S staining results show the effect of DEX (10− 6 M) on matrix mineralization. G PCR results show the effect of DEX (10− 6 M) and/or ATRA (0.5 µM) on Runx2 mRNA levels. H Western blot results show the effect of DEX (10− 6 M) and/or ATRA (0.5 µM) on RUNX2 protein levels. I ALP staining results show the effect of DEX (10− 6 M) and/or ATRA (0.5 µM) on ALP activities. J PCR results show the effect of DEX (10− 6 M) and/or ATRA (0.5 µM) on Opn mRNA levels. K Western blot results show the effect of DEX (10− 6 M) and/or ATRA (0.5 µM) on OPN protein levels. L Alizarin red S staining results show the effect of DEX (10− 6 M) and/or ATRA (0.5 µM) on matrix mineralization. *p < 0.05, **p < 0.01 Effects of DEX on the expression of Cyp26b1 in MSCs and bone of GIOP rats ATRA can reverse the osteogenic inhibitory effect of DEX, which imply that the concent of ATRA may be affected by DEX. Thus, we next determined whether DEX (10− 9 M to 10− 5 M) can affect the level of ATRA in MSCs. The ELISA assay results exhibit that the contents of ATRA were increased when the concentrations of DEX are less than 10− 7 M, or it was reduced (Fig. 2A). In particular, statistically significant differences were observed between the DEX 10− 7 M group and the DMEM, DEX 10− 9 M, DEX 10− 8 M, and DEX 10− 6 M groups. Therefore, We next determined the effect of DEX (10− 6 M) on the ATRA-related synthetic enzymes, as well as metabolic enzymes. PCR results exhibit that no substantial effect on ATRA synthetic enzymes were found, but the expression of Cyp26b1 was notably increased (Fig. 2B); meanwhile, Cyp26b1 was expressed in BMSCs (bone marrow mesenchymal stem cells), C3H10T1/2 cells (a mesenchymal stem cell line), C2C12 cells (a myoblast cell line), MC3T3-E1 cell s (a pre-osteoblast cell line), 3T3-L1 cells (a pre-adipocyte cell line), and MEFs (mouse embryonic fibroblasts) (Fig. 2C). The immunofluorescence staining results exhibit that the level of CYP26B1 was reduced by DEX (10− 8 M) compare to the control, but increased obviously by high concentration of DEX (10− 6 M) (Fig. 2D). Based on its consistency with the low-dose DEX condition used in our previous osteogenic induction study and the ability to reproduce the major low-dose-associated changes observed in the present study, DEX (10− 8 M) was selected as the representative low-dose condition for the subsequent experiments. Immunohistochemical staining and western blot assay results showed that the level of CYP26B1 was increased greatly in GIOP rats bone tissues (Fig. 2E–F). These results suggest that the retinoic acid signaling may be disturbed by high concentration of DEX via upregulating CYP26B1 in MSCs. Effects of DEX on the expression of Cyp26b1 in MSCs and bone of GIOP rats. A ELISA assay results show the retinoic acid content in cells treated with different concentrations of DEX. B PCR assay results show the mRNA level of retinoic acid-related synthases and metabolizing enzymes in cells treated with DEX (10− 6 M). C PCR assay results show the mRNA levels of Cyp26b1 in different indicated cell types. D Immunofluorescence staining results show the effect of DEX on the expression of Cyp26b1 (Scar bar: 50 μm). E Immunohistochemical staining results show the levels of CYP26B1 in the tibia of GIOP rats (Scar bar: 200 μm). F Western blot assay results show the levels of CYP26B1 in tibia of GIOP rats. G Images show the transfection rate recombinant adenoviruses in C3H10T1/2 cells (treated for 24 h). H, J Western blot assay results show the effects of AdGFP or AdCyp26b1 on the protein levels of CYP26B1. I, K Western blot assay results show the effects of AdGFP or AdsiCyp26b1 on the protein levels of CYP26B1. *p < 0.05, **p < 0.01 To further investigate the functional role of CYP26B1 in DEX-induced osteogenic inhibition, we next validated the efficiency of the recombinant adenoviruses for Cyp26b1 overexpression and knockdown. As shown in Fig. 2G, high infection efficiency was obtained in C3H10T1/2 cells 24 h following viral infection. Further Western blot analysis confirmed that level of CYP26B1 was significantly increased by using AdCyp26b1 compared with the AdGFP control (Fig. 2H–J), whereas decreased markedly byAdsiCyp26b1 (Fig. 2I–K). These results indicate that these recombinant adenoviruses were successfully established. Effects of CYP26B1 and/or ATRA on the osteogenic inhibitory effect of DEX in MSCs As DEX can up-regulate Cyp26b1, so we next determined whether the reversal effect of ATRA on the osteogenic inhibitory effect DEX can be affected by CYP26B1. PCR and western blot assay results exhibit that the effects of ATRA on reversal the DEX-reduced RUNX2 and OPN were all offset by the overexpression of Cy26b1(Fig. 3D–E, F–H). ALP and Alizarin Red S staining also exhibit the similar results (Fig. 3A–C, I–J). These results suggest that Cyp26b1 may play an important role in the osteogenic inhibitory effect of high concentration of DEX in MCSs. Effects of CYP26B1 and/or ATRA on the osteogenic inhibitory effect of DEX in MSCs. A PCR results show the effect of DEX (10− 6 M), ATRA (0.5 µM), and/or AdCyp26b1 on Runx2 mRNA levels. B Western blot results show the effect of DEX (10− 6 M), ATRA (0.5 µM), and/or AdCyp26b1 on RUNX2 protein levels. C Quantitative results of western blot assay show the effect of DEX (10− 6 M), ATRA (0.5 µM), and/or AdCyp26b1 on RUNX2 protein levels. D ALP staining results show the effect of DEX (10− 6 M), ATRA (0.5 µM), and/or AdCyp26b1 on ALP activities. E Quantitative results of ALP staining show the effect of DEX (10− 6 M), ATRA (0.5 µM), and/or AdCyp26b1 on ALP activities. F PCR results show the effect of DEX (10− 6 M), ATRA (0.5 µM), and/or AdCyp26b1 on Opn mRNA levels. G Western blot results show the effect of DEX (10− 6 M), ATRA (0.5 µM), and/or AdCyp26b1 on OPN protein levels. H Quantitative results of western blot assay show the effect of DEX (10− 6 M), ATRA (0.5 µM), and/or AdCyp26b1 on OPN protein levels. (I) Alizarin Red S staining results show the effect of DEX (10− 6 M), ATRA (0.5 µM), and/or AdCyp26b1 on matrix mineralization. J Quantitative results of Alizarin Red S staining show the effect of DEX (10− 6 M), ATRA (0.5 µM), and/or AdCyp26b1 on matrix mineralization. *p < 0.05, **p < 0.01 Effects of CYP26B1 on the osteogenic inhibitory effect of DEX in MSCs CYP26B1 can offset the reversal effect of ATRA on the osteogenic inhibitory effect DEX, we then determined whether the effect of DEX on osteogenic differentiation can be affected by CYP26B1. PCR and western blot assay results exhibit that the expression levels of RUNX2 and OPN were all decreased by DEX, but increased by Cyp26b1 knockdown; meanwhile, the inhibitory effect of DEX was almost reversed by Cyp26b1 knockdown (Fig. 4A–B, D–E). ALP and Alizarin Red S staining exhibit the similar results (Fig. 4C, F). On the contrary, the expression of RUNX2 and OPN were decreased by overexpression of Cyp26b1, and the inhibitory effects of DEX on the expression of RUNX2 and OPN were strengthened by overexpression of Cyp26b1 (Fig. 4G–H, J–K). These effect of overexpression of Cyp26b1 were recapitulated by the ALP and Alizarin Red S staining results (Fig. 4I, L). These results also suggest that CYP26B1 may play an important role in mediating the osteogenic inhibitory effect of DEX in MSCs. Effects of CYP26B1 on the osteogenic inhibitory effect of DEX in MSCs. A PCR results show the effect of DEX (10− 6 M) and/or AdsiCyp26b1 on Runx2 mRNA levels. B Western blot results show the effect of DEX (10− 6 M) and/or AdsiCyp26b1 on RUNX2 protein levels. C ALP staining results show the effect of DEX (10− 6 M) and/or AdsiCyp26b1 on ALP activities. D PCR results show the effect of DEX (10− 6 M) and/or AdsiCyp26b1 on Opn mRNA levels. E Western blot results show the effect of DEX (10− 6 M) and/or AdsiCyp26b1 on OPN protein levels. F Alizarin Red S staining results show the effect of DEX (10− 6 M) and/or AdsiCyp26b1 on matrix mineralization. G PCR results show the effect of DEX (10− 6 M) and/or AdCyp26b1 on Runx2 mRNA levels. H Western blot results show the effect of DEX (10− 6 M) and/or AdCyp26b1 on RUNX2 protein levels. I ALP staining results show the effect of DEX (10− 6 M) and/or AdCyp26b1 on ALP activities. J PCR results show the effect of DEX (10− 6 M) and/or AdsiCyp26b1 on Opn mRNA levels. K Western blot results show the effect of DEX (10− 6 M) and/or AdCyp26b1 on OPN protein levels. F Alizarin Red S staining results show the effect of DEX (10− 6 M) and/or AdCyp26b1 on matrix mineralization. *p < 0.05, **p < 0.01 Effects of Cyp26b1 knockdown on DEX-induced GIOP in rats We next established a DEX induced GIOP model in rats to further determine the effect of Cyp26b1 knockdown on bone formation (Fig. 5A). ELISA assay results exhibit that the serum ATRA levels were significantly reduced in the GIOP group, but significantly elevated in the GIOP-incorporated Cyp26b1 knockdown adeno-associated virus group (Fig. 5B). Reconstruction results of micro-CT scanning exhibit that bone trabeculae were sparse in the model and the virus vector group, compared with the control group; whereas the number of bone trabeculae was significantly increased in the GIOP-incorporated AAV-siCyp26b1-treated group (Fig. 5C). These results were consistent with the quantitative analysis results of micro-CT scanning (Fig. 5D). Masson trichrome staining exhibit that bone trabeculae and the collagen production were decreased in the GIOP and the GIOP plus empty virus group, compared with the control group. while the collagen production and trabecular density were increased substantially in GIOP combined with Cyp26b1 knockdown group (Fig. 5E). These results suggest that Cyp26b1 knockdown may ameliorate bone loss in the DEX-induced GIOP rat model. Effects of Cyp26b1 knockdown on GIOP in rats. A Schematic diagram of GIOP model construction for drug delivery. B Reconstruction results of micro-CT scanning show the effect of Cyp26b1 knockdown on the bone density and trabeculae in GIOP rats. C ELISA assay results show the effect of Cyp26b1 knockdown on the serum retinoic acid levels in GIOP rats. D Quantitative analysis results of the micro-CT scanning show the effect of Cyp26b1 knockdown on the DEX-induced GIOP in rats. E Masson trichrome staining results show the effect of Cyp26b1 knockdown on collagen fiber synthesis in the femur of GIOP in rats. *p < 0.05, **p < 0.01 Effects of NF-κB and/or CYP26B1 on the osteogenic inhibitory effect of DEX in C3H10T1/2 cells It was reported that NF-κB signaling plays a crucial role in GIOP; however, the precise relationship between them remains unclear. Western blot and immunofluorescence staining results exhibit that DEX exerted no apparent effect on the total protein of p65 and IκBα, but increased the level of p-p65 and p-IκBα concentration dependently; in addition, high concentration of DEX can promote the nucleus translocation of p-p65 (Fig. 6A–B). To determine whether NF-κB activation correlates with DEX-induced osteogenic suppression, we performed immunofluorescence co-staining for RUNX2 and p-p65. DEX at 10⁻⁸ M increases RUNX2 expression without markedly altering p-p65, whereas DEX at 10⁻⁶ M decreases RUNX2 expression but significantly elevates p-p65 levels and induces its nuclear translocation (s-Fig. 1B). To further confirm the nuclear translocation of p-p65 induced by high-concentration DEX, we performed nuclear and cytoplasmic fractionation followed by western blot analysis. The results show that DEX treatment markedly increases nuclear p65 and p-p65 levels, with p-p65 signal being barely detectable in control nuclei. In the cytoplasmic fraction, total p65 levels remain unchanged, whereas cytoplasmic p-p65 levels are decreased, consistent with its translocation into the nucleus (s-Fig. 1C–E). The immunofluorescence staining assay also exhibit that the p-p65 level was increased in the GIOP bone tissue (Fig. 6C), and this result was confirmed by the western blot assay to determine the level of p-p65 and p-IκBα in GIOP bone (Fig. 6D–E). The above data suggest that retinoic acid signaling may incorporate with NF-κB signaling to regulate the osteogenic inhibitory effect of DEX in MSCs. Thus, we further determined how CYP26B1 and NF-κB affect the osteogenic inhibitory effect of DEX. PCR and western blot assay results exhibit that the inhibitory effect of DEX on RUNX2 partially reversed by the utilization of NF-κB inhibitor (BAY11-7082) compare to the control group, but it was almost abolished by the overexpression of Cyp26b1 in C3H10T1/2 cells (Fig. 6F–G). Consistently, ALP staining results show that the DEX-induced reduction in ALP activity was partially reversed by BAY11-7082, and this protective effect was almost eliminated by Cyp26b1 overexpression (Fig. 6H). Furthermore, PCR and western blot assay results exhibit that the inhibitory effect of DEX on OPN was also partially reversed by BAY11-7082, but almost abolished by Cyp26b1 overexpression (Fig. 6I–J). Alizarin Red S staining results also exhibit that inhibition of NF-κB can reverse the inhibitory effect of DEX, which was almost eliminated by overexpression of Cyp26b1 in C3H10T1/2 cells (Fig. 6K). These results indicate that the activation of NF-κB signaling may be involved in mediating the osteogenic inhibitory effect of DEX through regulating retinoic acid signaling in MSCs. Effects of NF-κB and/or CYP26B1 on the osteogenic inhibitory effect of DEX in C3H10T1/2 cells. A Western blot results show the effect of different concentration of DEX on the level of p65, p-p65, IκBa, and p-IκBa. B Western blot results show the effect of different concentration of DEX on the level of p65, p-p65, IκBa, and p-IκBa at different time points. C Immunofluorescence staining results show the levels of p-p65 in ferum of GIOP rats (Scar bar: 100 μm). D Western blot results show the levels of the level of p65, p-p65, IκBa, p-IκBa in femurs of GIOP rats. E Quantitative results of western blot assay show the levels of the ratio of p-IκBa to IκBa, and p-p65 to p65 in femurs of GIOP rats. F PCR results show the effect of DEX (10− 6 M), BAY11-7082, and/or AdCyp26b1 on Runx2 mRNA level. G Western bot assay results show the effect of DEX (10− 6 M), BAY11-7082, and/or AdCyp26b1 on RUNX2 protein level. H ALP staining results show the effect of DEX (10− 6 M), BAY11-7082, and/or AdCyp26b1 on ALP activities. I PCR results show the effect of DEX (10− 6 M), BAY11-7082, and/or AdCyp26b1 on Opn mRNA level. J Western bot assay results show the effect of DEX (10− 6 M), BAY11-7082, and/or AdCyp26b1 on OPN protein level. K Alizarin Red S staining results show the effect of DEX (10− 6 M), BAY11-7082, and/or AdCyp26b1 on matrix mineralization. *p < 0.05, **p < 0.01. BAY11-7082: inhibitor of NF-κB Effects of NF-κB and/or DEX on the expression of Cyp26b1 in C3H10T1/2 cells We have demonstrated that NF-κB may cross-talk with retinoic acid signaling in the osteogenic inhibitory of DEX, but the concrete mechanism remains unknown. Therefore, we next investigate whether NF-κB signaling can regulate the effect of DEX on Cyp26b1 expression. RT-PCR assay results exhibit that the expression of Cyp26b1 was increased by DEX when the concentration is lower than 10− 5 M, but reduced by BAY11-7082 (Fig. 7A), and the effect of DEX (10− 6 M) on increasing the mRNA level of Cyp26b1 was abolished by the combination with BAY11-7082 (Fig. 7B). Western blot assay results exhibit that the protein level of CYP26B1 was reduced by DEX when the concentration is lower than 10− 6 M, or it was increased (Fig. 7C, E). The protein level of CYP26B1 was increased by DEX (10− 6 M), but reduced by NF-κB specific inhibitor; the effect of DEX (10− 6 M) on increasing the protein level of CYP26B1 was attenuated apparently by BAY11-7082 (Fig. 7D, F). IP and western blot assay results exhibit that p-p65 can interact with GR (Fig. 7G). The ChIP assay results exhibit that GR can interact with the promoter of Cyp26b1 (Fig. 7H). These results suggest that NF-κB signaling may promote the effect of DEX on increasing the mRNA level of Cyp26b1, although this was not consistent with their effect on the protein level of CYP26B1. Effects of NF-κB and/or DEX on the expression of Cyp26b1 in C3H10T1/2 cells. A PCR assay results show the effect of different concentration of DEX on Cyp26b1 mRNA level. B PCR assay results show the effect of different concentration of DEX (10− 6 M) and/or BAY11-7082 on Cyp26b1 mRNA level. C Western bot assay results show the effect of different concentration of DEX on CYP26B1 protein level. D Western bot assay results show the effect of DEX (10− 6 M) and/or BAY11-7082 on CYP26B1 protein level. E Quantitative results of western bot assay show the effect of different concentration of DEX on CYP26B1 protein level. F Quantitative results of western bot assay show the effect of DEX (10− 6 M) and/or BAY11-7082 on CYP26B1 protein level. G IP and western bot assay show the interaction between p-p65 and GR. H ChIP assay shows the interaction GR and the promoter of Cyp26b1. I Western bot assay results show the effect of DEX (10− 8 M) and/or BAY11-7082 on the degradation of CYP26B1. J Quantitative results of western bot assay show the effect of DEX (10− 8 M) and/or BAY11-7082 on the degradation of CYP26B1. K Western bot assay results show the effect of DEX (10− 6 M) and/or BAY11-7082 on the degradation of CYP26B1. L Quantitative results of western bot assay show the effect of DEX (10− 6 M) and/or BAY11-7082 on the degradation of CYP26B1. M Western bot assay results show the effect of DEX (10− 8 M) and BafA1/or MG132 on the levels of CYP26B1. N Western bot assay results show the effect of DEX (10− 6 M) and BafA1/or MG132 on the levels of CYP26B1. O–P Western bot assay and quantitative results show the effect of DEX (10− 8 M) and BafA1 on the levels of CYP26B1 at different time points. Q–R Western bot assay and quantitative results show the effect of DEX (10− 6 M) and BafA1 on the levels of CYP26B1 at different time points. In panel A, C, and E, DMEM indicates cells cultured with regular DMEM supplemented with 10% fetal bovine serum without DEX treatment and was used as the basal control. In panel B, D, F, and G–R, Control indicates cells cultured with osteogenic induction medium containing 10⁻⁷ M DEX, which served as the osteogenic induction condition. *p < 0.05, **p < 0.01. (GR: glucocorticoids receptor, BAY11-7082: inhibitor for NF-κB signaling. BafA1: bafilomycin A1, MG132: proteasome inhibitor, DEX: dexamethasone) s-Fig. 1 Effects of DEX on cell viability, RUNX2 expression, and p65/p-p65 subcellular distribution in C3H10T1/2 cells. A CCK8 assay results show the cell viability in C3H10T1/2 cells treated with DEX at 10⁻⁸ M, 10⁻⁷ M, or 10⁻⁶ M for 24, 48, and 72 h. B Immunofluorescence staining results show the levels of RUNX2 (green) and p-p65 (red) in C3H10T1/2 cells treated with DEX at 10⁻⁸ M or 10⁻⁶ M. Scale bar: 50 μm. C Western blot assay results show the levels of p65 and p-p65 in the cytoplasmic fraction of C3H10T1/2 cells treated with DEX at 10⁻⁶ M. D Western blot assay results show the levels of p65 and p-p65 in the nuclear fraction of C3H10T1/2 cells treated with DEX at 10⁻⁶ M. E Quantitative analysis of the p-p65/p65 ratio in the cytoplasmic and nuclear fractions. p* < 0.05, ***P < 0.001 Although low-dose DEX showed only a modest effect on Cyp26b1 mRNA expression, its protein levels were markedly decreased. This discrepancy may be attributed to accelerated lysosomal degradation of CYP26B1 induced by low-dose DEX. In contrast, high-dose DEX increased Cyp26b1 transcription and impaired CYP26B1 degradation, thereby leading to CYP26B1 protein accumulation. Therefore, we next examined whether NF-κB and DEX could affect CYP26B1 protein degradation. Western blot assay results exhibit that the degradation of CYP26B1 was accelerated by the low concentration of DEX (10− 8 M), which was blocked by BAY11-7082 (Fig. 7I–J); on the contrary, the degradation of Cyp26b1 was blocked by the high concentration of DEX (10− 6 M), which was reduced by BAY11-7082 significantly (Fig. 7K–L). Since the degradation of protein is main carried out through ubiquitin-proteasome pathway or autophagy-lysosome pathway, we then analyzed how the degradation of Cyp26b1 was affected by the different concentrations of DEX, utilizing a proteasome inhibitor MG132 and an autophagy inhibitor bafilomycin A1 (BafA1). Western blot results exhibit that the level of CYP26B1 was decreased by the low concentration of DEX (10− 8 M), either BafA1 or MG132 alone exert no substantial effect on the level of CYP26B1; but the decrease of CYP26B1 was attenuated by the combination of BafA1, and no changes were found when DEX (10− 8 M) combined with MG132 (Fig. 7M). On the contrary, the protein level of CYP26B1 was increased by the high concentration of DEX (10− 6 M) comparing to the control group, which was enhanced by BafA1 and not changed by MG132 (Fig. 7N). These results suggest that low-dose DEX reduces CYP26B1 protein mainly by enhancing autophagy-lysosome-dependent degradation, whereas high-dose DEX increases CYP26B1 protein by suppressing this degradation process. In contrast, MG132 failed to reverse these effects, indicating that the ubiquitin–proteasome pathway may not be the primary underlying mechanism. The effect of DEX on the protein level of CYP26B1 was concentration- and time-dependent (Fig. 7O–R). These results indicate that NF-κB signaling can promote the effect of DEX on upregulating the expression of Cyp26b1, but the degradation of CYP26B1 can be blocked by high concentration of DEX, which was strengthened by the activation of NF-κB signaling. These results indicate that low-dose DEX increased Cyp26b1 mRNA but did not increase CYP26B1 protein, whereas high-dose DEX increased Cyp26b1 mRNA and promoted CYP26B1 protein accumulation. Discussion It is well known that long-term or excessive administration of glucocorticoids (GCs) can disrupt bone homeostasis and lead to glucocorticoid-induced osteoporosis (GIOP). Although retinoic acid signaling plays an important role in regulating the lineage commitment of MSCs, the intrinsic relationship between GIOP and retinoic acid remains unclear completely. In this study, we found that disturbance of retinoic acid signaling homeostasis through upregulation of Cyp26b1 may be one of the pivotal mechanisms underlying the osteogenic inhibitory effect of DEX. Our findings may provide another potential target to ameliorate GIOP. GCs are widely used for inflammatory or immune-related diseases, their dose- and duration-dependent adverse effects remain a major clinical challenge. Population-based studies indicate that 1%–2% of adults worldwide use oral GCs long term, and 30%–50% of these users develop GIOP [30]. Pediatric populations are also vulnerable, and GIOP-related fracture is an important cause of morbidity in children with rheumatic disorders [31]. Mechanistically, GIOP is closely associated with impaired osteoblast function by down-regulating key transcription factors such as Runx2 and Osterix [32], while also skewing MSCs lineage commitment toward adipogenesis [33]. In addition, GCs promote osteoclast activation by increasing the RANKL/OPG ratio [34], disturb chondroblastic, osteoblastic, and B-cell lineage crosstalk during bone development [35]; meanwhile, it can also modulate epigenetic events related to osteoblast differentiation, such as DNA methylation of osteogenic genes [36]. However, the molecular basis of GIOP remains incompletely understood. Retinoic acid signaling is closely related to skeletal development and bone homeostasis, although it acts as a double-edged sword in a context- and dose-dependent manner [37, 38]. Given that ATRA may also regulate osteoclastogenesis in a dose- and context-dependent manner, the present study primarily focuses on the osteoblast-lineage commitment in MSCs. All-trans retinoic acid (ATRA), a ligand for retinoic acid receptors and retinoid X receptors, regulates cell differentiation, proliferation, and apoptosis. Our previous studies showed that low concentrations of ATRA promoted osteogenic differentiation and enhanced BMP9-induced osteogenesis, possibly through activation of Wnt/β-catenin signalin [39–41]. We also demonstrated that ATRA attenuated DEX-induced osteogenic disruption [17], a finding that was recapitulated in the present study. Biologically, intracellular ATRA levels are determined by both synthesis and degradation. ATRA is synthesized from retinol through alcohol dehydrogenases, retinol dehydrogenases, and retinaldehyde dehydrogenases, whereas its degradation is mainly mediated by CYP26 family enzymes, including CYP26A1, CYP26B1, and CYP26C1 [42–44]. Therefore, disruption of ATRA homeostasis may represent an important but underexplored mechanism contributing to GIOP. In this study, we found that intracellular ATRA content was increased by low concentrations of DEX, but decreased by high concentrations of DEX, especially at 10− 6 M. Although DEX (10− 6 M) exerted no substantial effect on the expression of Raldhs, Cyp26a1, or Cyp26c1, it greatly increased the expression level of Cyp26b1. Further analysis showed that the effect of ATRA in reversing the osteogenic inhibitory effect of DEX (10− 6 M) was almost diminished by Cyp26b1 overexpression, and the osteogenic inhibitory effect of DEX was almost offset by Cyp26b1 knockdown but strengthened by Cyp26b1 overexpression. Since CYP26B1 is a major retinoic acid-metabolizing enzyme that determines intracellular ATRA availability, its degradation may prevent excessive ATRA catabolism and thereby maintain an endogenous retinoid level favorable for osteoblast differentiation [42–44]. This may explain why low-dose DEX, especially 10− 8 M, can be compatible with or supportive of osteogenic induction, as observed as report [17]. In contrast, high-dose DEX promoted CYP26B1 accumulation, which may accelerate ATRA clearance and contribute to impaired osteogenic differentiation. Therefore, the opposite biological effects of low- and high-dose DEX may be at least attributed to the different effects on the expression of CYP26B1. Our data showed that the changes in Cyp26b1 mRNA expression were not completely parallel with CYP26B1 protein levels under different DEX concentrations. This discrepancy suggests that CYP26B1 may be regulated by DEX not only at the transcriptional level, but also at the post-translational level. Although low-dose DEX, especially 10− 8 M, increased Cyp26b1 mRNA expression, it simultaneously promoted CYP26B1 protein degradation through the autophagy-lysosome pathway, thereby preventing CYP26B1 protein accumulation. In contrast, high-dose DEX increased Cyp26b1 transcription and inhibited CYP26B1 degradation, resulting in CYP26B1 protein accumulation. Therefore, the CYP26B1 protein level appears to be determined by the balance between transcriptional induction and protein degradation. This dose-dependent regulation of CYP26B1 protein stability may explain the apparent inconsistency between Cyp26b1 mRNA and protein level. NF-κB signaling plays an important role in the regulation of immune response, cell proliferation, differentiation, and apoptosis [45, 46]. In recent years, the role of NF-κB signaling in osteoporosis, heterotopic ossification, osteoarthritis, and other bone-related diseases has attracted increasing attention [47–50]. Consistent with its relevance to bone metabolism, previous studies reported that modulation of NF-κB signaling can alleviate glucocorticoid-induced bone loss or DEX-induced osteoblast injur [27–29]. In this study, we recapitulated that DEX (10− 6 M) can activate NF-κB signaling and demonstrated that the osteogenic disruption effect of DEX was largely reversed by a specific NF-κB inhibitor. Thus, NF-κB appears to partially contribute to DEX-induced Cyp26b1 upregulation. Further analysis showed that p65 may interact with the glucocorticoid receptor to upregulate the expression of Cyp26b1, thereby linking DEX/GR signaling to the suppression of retinoic acid signaling. Moreover, high concentrations of DEX also inhibited the degradation of CYP26B1, which was offset by the NF-κB specific inhibitor, suggesting that NF-κB may regulate CYP26B1 at both the transcriptional and protein-stability levels. Therefore, activation of the NF-κB/CYP26B1 axis may represent an important mechanism by which high-dose DEX disturbs ATRA homeostasis and impairs osteogenic differentiation. However, the precise mechanisms driving this non-canonical interaction between DEX/GR signaling, NF-κB activation, and CYP26B1 protein stabilization in MSCs need further investigation. From a translational perspective, our findings suggest that CYP26B1 may be a potential therapeutic target for GIOP. CYP26 enzymes are responsible for the oxidative catabolism of retinoic acid and have been considered pharmacologically tractable targets [42, 43, 51]. Several retinoic acid metabolism-blocking agents (RAMBAs), including liarozole and talarozole, have been developed to inhibit CYP26-mediated retinoic acid degradation and enhance endogenous retinoid signaling [52, 53]. Compared with direct systemic administration of exogenous ATRA, targeting CYP26B1 may offer a more physiological strategy by preserving or restoring endogenous ATRA availability within local tissues. This may be of particular importance for GIOP, wherein high-dose DEX, a prototypical synthetic glucocorticoid, impairs osteogenic differentiation and bone formation. However, given the context- and dose-dependent effects of retinoic acid on skeletal homeostasis [37, 38], CYP26B1-targeted intervention should be interpreted cautiously. Future studies using bone-targeted delivery systems, CYP26B1-selective inhibitors, or osteoblast-lineage-specific genetic models will be necessary to determine whether modulation of the CYP26B1/ATRA axis can safely and effectively prevent bone loss from GIOP in vivo. Although DEX was used as a representative synthetic glucocorticoid in this study, additional work is warranted to clarify whether other GCs can regulate the NF-κB/CYP26B1/ATRA axis in the identical fashion. Although the 5 mg/kg DEX used in our study is a highly standardized regimen to establish a robust pharmacological GIOP model [54, 55], the lack of direct measurement of local DEX concentrations in bone tissue remains a limitation. Given the substantial discrepancies between in vivo pharmacokinetic profiles and in vitro experimental settings, additional studies including tissue concentration quantification are needed to establish an accurate correlation between local bone DEX concentrations and the high-dose phenotypes observed in vitro. Conclusion NF-κB signaling activation contributes critically to DEX-regulated CYP26B1 expression by stimulating its transcription while suppressing its protein degradation. Thus, our results provide an inherent link between DEX-induced osteogenic suppression and retinoic acid signaling disturbance, laying a mechanistic foundation and identifying a prospective therapeutic target for the treatment of GIOP. Supplementary Information Supplementary Material 1. Supplementary Material 2. Acknowledgements We sincerely thank technical Chen Chen (Key Laboratory of Biochemistry and Molecular Pharmacology of Chongqing, Chongqing Medical University) for providing recombinant adenovirus for this experiment.The authors declare that they have not use AI-generated work in this manuscript. Author contributions LC, WGH, and BCH conceptualized and designed the study. AHY and YMW led the experimental work and performed the majority of data analysis. XYW, JL, NT, and WTL provided support with animal model establishment and in vivo sample collection. JC and DMH assisted with in vitro experiments and statistical analyses. AH and YMW drafted the manuscript, while BCH, WGH and LC contributed to manuscript editing and formatting revisions. All authors reviewed and approved the final version of the manuscript. Funding The present study was funded partially by Chongqing Science and Technology Bureau (CSTB2024NSCQ-MSX0411 to Baicheng He), Science and Technology Research Project of Chongqing Municipal Education Commission (KJQN202300125 to WenGe He), and Chongqing Innovative Medical Device Application Demonstration Project (CQEIC2024MDAD-010 to Liang Chen). Data availability All data for this study are available upon reasonable request from the corresponding author. Declarations Ethics approval and consent to participate This study was approved by the Ethics Committee of Chongqing Medical University (CQMU). The approved project title was “Glucocorticoids inhibit osteogenic differentiation of mesenchymal stem cells by upregulating CYP21B1”, with the approval number IACUC-CQMU-2025-0049, granted on February 5, 2025. No additional human subjects or embryonic materials were involved in this study. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Ai-Hua Ye and Yu-Mei Wang have contributed equally to this work. Contributor Information Bai-Cheng He, Email: bche@cqmu.edu.cn. Wen-Ge He, Email: cqmu_hwg1995@163.com. Liang Chen, Email: Chen.liang.1979@cqu.edu.cn. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Data Availability Statement All data for this study are available upon reasonable request from the corresponding author.