TLR4-dependent upregulation of IL-22-producing T cells in the bone niche links to the imbalance of bone homeostasis

Abstract

Background

Toll-like receptor 4 (TLR4) activation within the bone marrow (BM) microenvironment regulates proinflammatory cytokines and chemokines, which influence immune activity and bone remodeling. However, the immune cell factors mediating TLR4-dependent cross-talk in the bone/BM microenvironment remain insufficiently defined.

Methods

We provide a transcriptomic analysis of significant cellular and immune changes in total BM cells of wild-type mice following lipopolysaccharide (LPS)-induced TLR4 activation in vivo compared with TLR4 knockout (KO) mice. To obtain direct evidence that interleukin (IL)-22 contributes to TLR4-driven bone destruction, we analyzed their bone phenotypes following LPS injection into IL-22 KO mice. In addition, co-culture of pre-osteoblasts (OBs) and BM cells was conducted to evaluate the effect of IL-22 in osteoclast (OC) formation in a pre-OB-dependent manner.

Results

Differentially expressed gene analysis from BM cells of wild-type (WT) mice revealed an increase in IL-22-producing cells, along with increased IL-22 production in a TLR4-dependent manner. Flow cytometry confirmed that LPS enhanced the number of CD4⁺IL-22⁺ cells in the BM and decreased in TLR4-deficient BM. LPS also enhanced the differentiation of CD4⁺IL-22⁺ T cells and promoted IL-22 induction in T cells. IL-22 KO mice showed minimal LPS-induced bone loss, whereas WT mice exhibited severe bone destruction with increased formation of tartrate-resistant acid phosphatase-positive OCs. Interestingly, while IL-22 did not directly affect OC differentiation, IL-22 enhanced OC differentiation only when BM cells were co-cultured with pre-OBs through the elevation of the receptor activator of nuclear factor-κB ligand (RANKL) in pre-OBs during osteogenesis. Mechanistically, IL-22 induced cyclooxygenase-2 in pre-OBs through extracellular signal-regulated kinase and p38 signaling, driving prostaglandin E2 production and subsequent RANKL induction.

Conclusions

TLR4 activation enhanced IL-22 production in T cells, which stimulates RANKL production in pre-OBs to further induce OC formation. These results support that IL-22 acts as a TLR4-driven mediator to shift bone remodeling toward catabolic bone degeneration.

Supplementary Information

The online version contains Supplementary Material available at https://doi.org/10.1186/s11658-026-00994-w.

Introduction

The bone marrow (BM) microenvironment serves as a specialized niche for mesenchymal stromal cells (MSCs) and hematopoietic stem cells, providing signals for survival, self-renewal, and differentiation into osteoblasts (OBs) and osteoclasts (OCs) [1, 2]. This niche regulates bone homeostasis through a network of cellular and molecular interactions among BM resident cells, including stromal cells and OBs, and various immune cells such as T cells, B cells, macrophages, and innate lymphoid cells (ILCs) [1, 2]. Bone homeostasis relies on the balance between bone-forming OBs and bone-resorbing OCs [3]. OC differentiation is influenced by the maturation state of OBs: Immature OBs (pre-OBs) promote osteoclastogenesis by producing receptor activator of nuclear factor-κB (NF-κB) ligand (RANKL), whereas mature OBs inhibit this process by secreting osteoprotegerin (OPG), a decoy receptor for RANKL [4]. Inflammation disrupts this balance by altering the differentiation of OBs and OCs, leading to bone destruction [5, 6]. In response to inflammatory signals, the BM resident cells secrete immunomodulatory factors that regulate bone cell differentiation [7, 8]. Prominent inflammatory cytokines such as tumor necrosis factor-α (TNF-α), IL-1β, and IL-6, which are produced by activated BM immune cells, directly enhance OC formation and suppress OB formation [9–11]. These interactions are central to the pathogenesis of inflammatory bone diseases such as rheumatoid arthritis (RA), psoriasis, and osteoporosis, which are characterized by progressive bone loss and structural deformity [12–14].

Toll-like receptor 4 (TLR4) is a pattern recognition receptor that induces inflammatory responses to microbial stimuli [15]. Furthermore, TLR4 is expressed in bone cells, including OBs, OCs, and MSCs, which give rise to osteoprogenitor cells [16]. TLR4 signaling contributes to catabolic and anabolic processes of bone remodeling [17]. Although TLR4 activation promotes MSC proliferation during bone repair [18], it also disrupts bone homeostasis [17] by decreasing osteogenic differentiation and increasing adipogenesis-related gene expression. This shift promotes MSC commitment to adipogenic rather than osteogenic cells, as observed in osteoporosis and periodontitis [19, 20]. For example, high glycolipid conditions induce TLR4 expression and subsequently suppress osteogenic differentiation in OBs via S100B/extracellular signal-regulated kinase (ERK) [21]. TLR4 activation also stimulates the secretion of inflammatory cytokines such as TNF-α, IL-1β, and IL-6, through NF-κB activation in immune cells. These cytokines enhance OC activation and exacerbate bone destruction in inflammatory conditions, including osteomyelitis, periodontitis, and RA [22–26]. Although TLR4-mediated inflammation is associated with the development of a catabolic, bone-degrading environment, the specific BM-derived factors induced by TLR4 that mediate this imbalance remain insufficiently defined.

IL-22 is a member of the IL-10 cytokine family and shares approximately 20% amino acid homology with IL-10. IL-22 was originally termed the IL-10-related T cell-derived inducible factor [27]. It is primarily produced by activated T helper (Th) 17 cells [28], Th22 cells [29], γδ T cells, natural killer cells [30, 31], and ILC3s [32], a subset of tissue-resident innate immune cells that lack rearranged antigen-specific receptors but produce effector cytokines [33, 34]. Binding of IL-22 to its receptor complex activates Janus kinase (Jak) 1 and tyrosine kinase 2, leading to downstream signaling through mitogen-activated protein kinases (MAPKs), including ERK, mitogen-activated protein kinase kinase, c-Jun N-terminal kinase (JNK), and p38 kinase and signal transducer and activator of transcription (STAT) 1, STAT3, and STAT5 [35]. IL-22 is recognized as a key mediator in inflammatory diseases such as colitis and hepatitis [36, 37]. Increasing evidence also links IL-22 to bone loss in autoimmune disorders [38–42]. Interestingly, IL-22-deficient mice display distinct phenotypes across autoimmune models—showing increased susceptibility to experimental autoimmune encephalomyelitis but reduced susceptibility to collagen-induced arthritis [43, 44]—suggesting context-dependent, non-immune roles in bone regulation. IL-22 also modulates osteoprogenitor proliferation and migration, as well as calcium production under inflammatory conditions, implicating its involvement in postinflammatory osteogenesis in spondyloarthropathies [45]. However, the mechanisms by which IL-22 influences the bone microenvironment during inflammation remain poorly defined.

In this study, the TLR4-dependent increase in IL-22 production by CD4⁺IL-22⁺ T cells enhanced the migration of osteoprogenitor cells and promoted their differentiation into pre-OBs within the bone niche. It was further demonstrated that IL-22 is crucial for inducing RANKL in pre-OBs by activating cyclooxygenase (COX)-2/prostaglandin E₂ (PGE2) pathway through ERK and p38 signaling. This induction ultimately stimulated OC formation in in vitro co-culture systems and an in vivo inflammatory bone loss model. Together, these findings reveal a non-immune function role of IL-22 in regulating osteoprogenitor cell function, advancing our understanding of its contribution to bone homeostasis and the pathogenesis of bone-destructive diseases.

Materials and methods

Animal experiments

All animal procedures followed protocols approved by the Institutional Committee for the Care and Use of Laboratory Animals at Ulsan University (2016–13315). Wild-type (WT) and TLR4 knockout (KO) C57BL/6J mice were procured from Jackson Laboratory (Bar Harbor, ME, USA). IL-22 KO mice were obtained from UC Davis MMRRC (Davis, CA, USA) and backcrossed with C57BL/6J mice for eight generations.

Micro-computed tomography (micro-CT) analysis

An inflammatory bone-loss murine model was generated by systemic lipopolysaccharide (LPS) administration as previously described [46]. WT, TLR4 KO, or IL-22 KO mice (8 weeks old, female) received intraperitoneal LPS injections (5 mg/kg) or phosphate-buffered saline (PBS) on days 1 and 4. All mice were euthanized 3 days after the second injection. Long bones were collected and analyzed by micro-computed tomography using a 1072 Microtomograph system (Skyscan, Kontich, Belgium; 15-μm pixel size, 50 kVp, 200 μA, 0.5-mm AI filter). Datasets were reconstructed with NRecon software using a Feldkamp-based algorithm and segmented into binary images by adaptive thresholding. Bone volume was quantified from 300 tomographic slices using CTAn software (version 1.6) within a predefined region of interest in the distal femoral metaphysis. Three-dimensional surface models were generated using CTAn and visualized using CTVol (Bruker micro-computed tomography).

OB differentiation

Primary OB precursor cells were isolated from the calvariae of 1-day-old mice of both sexes following removal of the attached periosteum, blood vessels, and interstitial cartilage, as previously described [47]. Calvarial tissues were digested in Hank’s balanced salt solution containing 0.1% collagenase type II (GIBCO, Invitrogen, Carlsbad, CA, USA) and 0.2% dispase (Roche, Penzberg, Germany) for 20 min at 37 °C with agitation. This digestion was repeated six times, and cells from the final five digestions were pooled for culture. The collected cells were cultured for 3 days, and only adherent cells were used for subsequent experiments. Human MSCs (hMSCs) were procured from Lonza (Basel, Switzerland) and maintained in MSCGM^® Bullet Kit (Lonza) complete medium according to the manufacturer’s instructions. OB differentiation was induced by culturing cells in osteogenic medium (alpha modification of Eagle’s minimum essential medium, 10% fetal bovine serum, 1% penicillin–streptomycin, 10 mM β-glycerophosphate, and 50 mg/ml ascorbic acid) for 2 weeks, with medium changes twice weekly. Differentiation and mineralization were evaluated using alkaline phosphatase (ALP) and Alizarin red (AR).

OC differentiation

BM cells were isolated from the hind limbs of 6-week-old WT male and female mice, as previously described [46]. Cells were plated in 100-mm dishes and cultured for 24 h in alpha modification of Eagle’s minimum essential medium supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin. Nonadherent cells were collected and cultured for 3 days with macrophage colony-stimulating factor (M-CSF) (30 ng/ml). The adherent cells were considered to be BM-derived monocytes/macrophages and were used as OC precursors. To achieve OC differentiation, BM monocytes/macrophages were cultured with M-CSF (30 ng/ml) and RANKL (100 ng/ml) in the presence or absence of IL-22 for 4–5 days.

Tartrate-resistant acid phosphatase (TRAP) staining

Fixed OCs were stained for TRAP using a leukocyte acid phosphatase assay kit (Sigma-Aldrich, St. Louis, MO, USA) and examined under a light microscope. TRAP-positive multinucleated cells containing more than three nuclei were classified as mature OCs. For co-culture experiments, BM cells (2 × 10⁵ cells) and calvarial pre-OBs (2 × 10⁴ cells) were seeded in 48-well plates and cultured with 1α,25-dihydroxyvitamin D₃ (10^–8 M) and PGE2 (10^–6 M) for 4–7 days, with medium changes every 3 days [48, 49]. Cultures were then fixed and subjected to TRAP staining to assess OC differentiation. For in vivo analyses, hind limbs from WT and IL-22 KO mice treated with LPS (5 mg/kg) or PBS, administered twice weekly, were fixed in 4% paraformaldehyde and decalcified in 10% ethylenediaminetetraacetic acid for 30 days. Tissues were embedded in paraffin, sectioned at 5 μm, and stained with hematoxylin and eosin (H&E) or TRAP, followed by counterstaining with fast green as previously described [17]. TRAP-positive areas were assessed using ImageJ.

RNA-sequencing and analysis

RNA was isolated from BM cells of WT and TLR4 KO male mice treated with LPS (5 mg/kg) or PBS, administered twice weekly (n = 3 mice per group), using TRIzol reagent (Life Technologies, Carlsbad, CA, USA), followed by purification with the Direct-Zol™ RNA Miniprep Plus kit (Zymo Research, Irvine, CA, USA), according to the manufacturer’s instructions. Purified RNA was used for complementary DNA (cDNA) synthesis, indexing, and quantification. Transcripts were quantified using Salmon [50]. Quantified sf files were analyzed using DESeq2 (version 1.48.2). Gene Ontology enrichment analysis was performed using ClusterProfiler (version 4.16.0). Cytokine messenger RNA expression profiles were visualized using MultiExperiment Viewer (MeV version 4.9.0).

Quantitative real-time polymerase chain reaction (PCR)

RNA was isolated from cells using QIAzol reagent (Qiagen, Hilden, Germany). A quantity of 1 µg of total RNA was reverse-transcribed into cDNA using the RevertiAid First Strand cDNA Synthesis kit (Thermo Scientific, Vilnius, Lithuania) according to manufacturer’s instructions. Quantitative real-time PCR was performed using the LightCycler 480 SYBR Green 1-Step kit and the LightCycler^® 480 Instrument II Real Time PCR System (Roche Applied Science; Mannheim, Germany). Gene expression levels were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as an internal control. Relative expression was calculated using the standard curve method based on the cycle threshold (Ct) values of each target gene and GAPDH.

Western blotting

Immunoblotting was performed as previously described [17]. Briefly, the pre-OBs were treated with IL-22 (10 ng/ml) for the indicated durations, washed with ice-cold PBS, and lysed in Pro-pre™ Protein Extraction Solution (iNtRON Biotechnology, Seongnam, Korea) supplemented with protease and phosphatase inhibitors. Lysates were centrifuged at 12,000g for 5 min, and protein concentrations were determined using the Pierce BCA Protein Assay kit (Thermo Scientific, Waltham, MA, USA). Equal amounts of protein (10 µg per sample) were separated on 10–12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis gels and transferred onto polyvinylidene difluoride membranes (Cytiva, Marlborough, MA, USA). Membranes were blocked for 1 h with 5% skim milk in Tris-buffer saline (20 mM Tris/HCl, pH 7.6, 150 mM NaCl, and 0.1% Triton X-100), followed by overnight incubation at 4 °C with primary antibodies against p-ERK (cat. no. 9101S, Cell Signaling Technology, 1:1000), ERK (cat. no. 9102S, Cell Signaling Technology, 1:1000), p-p38 (cat. no. 9211S, Cell Signaling Technology, 1:1000), p38 (cat. no. 9212S, Cell Signaling Technology, 1:1000), p-JNK (cat. no. 4668S, Cell Signaling Technology, 1:1000), JNK (cat. no. 9252S, Cell Signaling Technology, 1:1000), p-STAT1 (cat. no. 9167L, Cell Signaling Technology, 1:1000), STAT1 (cat. no. NBP1-81,578, Novus Biologicals, 1:1000), p-STAT3 (cat. no. 9131S, Cell Signaling Technology, 1:1000), STAT3 (cat. no. NBP2-22,471, Novus Biologicals, 1:1000), p-STAT5 (cat. no. 9359S, Cell Signaling Technology, 1:1000), STAT5 (cat. no. 9363S, Cell Signaling Technology, 1:1000), COX-2 (cat. no. ab179800, Abcam, 1:1000), and actin (cat. no. ab8226, Abcam, 1:10,000). Membranes were then washed and incubated with horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. Immunoreactive signals were detected using an ECL kit (Cytiva, Marlborough, MA, USA). Protein band intensities were quantified by densitometry using ImageJ.

Enzyme-linked immunosorbent assay (ELISA)

Soluble mouse IL-22 was measured using mouse IL-22 ELISA kits (R&D Systems, Minneapolis, MN, USA). Recombinant mouse IL-22 serially diluted in assay diluent served as the standard. Control, standard, and samples were added to triplicate wells and incubated for 2 h at room temperature. After washing with PBS containing 0.05% Tween 20, wells were incubated with biotinylated anti-mouse IL-22 detection antibodies for 2 h at room temperature, followed by streptavidin–horseradish peroxidase for 30 min. Color development was achieved with tetramethylbenzidine solution and stopped with 0.1 M sulfuric acid. Absorbance at 450 nm was recorded using a Bio-Rad microtiter plate reader (Bio-Rad Laboratories, Hercules, CA, USA), and blank values were subtracted. Levels of mouse RANKL, OPG, and PGE2 in culture supernatants were quantified using specific and parameter assay kits (R&D Systems) according to the manufacturer’s instructions.

Flow cytometry and cell sorting

BM cells were isolated from male WT and TLR4 KO mice, with or without LPS injection, by flushing femurs and tibiae with Hanks’ balanced salt solution containing 2% trypsin–EDTA, 3% penicillin–streptomycin, and 1% collagenase type II. Following incubation at 37 °C for 20 min, red blood cells were lysed using BD Pharm Lyse (BD Biosciences, San Jose, CA, USA). Cells were filtered through a 70-μm cell strainer and incubated with TruStain FcX (cat. no. 101320, Biolegend, 1:50) for 15 min at 4 °C in FACS buffer (calcium- and magnesium-free PBS containing 1% bovine serum albumin, 4% fetal bovine serum, and 0.5 M ethylenediaminetetraacetic acid). Cells were stained for 30 min with PE-conjugated anti-mouse CD4 antibody (cat. no. 116006, Biolegend, 1:100) and PE/cy7-conjugated anti-mouse CD8 antibody (cat. no. 100722, Biolegend, 1:100), washed with PBS, fixed, and permeabilized using the Cytofix/Cytoperm kit (BD Biosciences) according to the manufacturer’s instructions. Intracellular IL-22 staining was performed using an Alexa Fluor 647-conjugated anti-mouse IL-22 antibody (cat. no. 516406, Biolegend, 1:20). Splenocytes were isolated from 10-week-old C57BL/6J male mice and stained with Zombie Aqua fixable viability dye (cat. no. 423101, Biolegend), Percp/Cy5.5-conjugated anti-mouse CD3 antibody (cat. no. 100218, Biolegend, 1:50), and PE-conjugated anti-mouse CD4 antibody (cat. no. 116006, Biolegend, 1:100) for fluorescence-activated cell sorting to purify CD3⁺CD4⁺ T cells. Sorted cells were cultured for 3 days with or without LPS (1 μg/ml) in the presence of anti-mouse CD3 antibody (cat. no. 100201, Biolegend, 1:500) and anti-mouse CD28 antibody (cat. no. 102102, Biolegend, 1:100). Cells were then stimulated with ionomycin (0.5 μg/ml), phorbol 12-myristate 13-acetate (10 ng/ml), and brefeldin (3 μg/ml) for 3 h. Following stimulation, CD3⁺CD4⁺ T cells were harvested, stained with Zombie Aqua viability dye (cat. no. 423101, Biolegend), fixed, permeabilized, and stained for IL-22 using IL-22 antibody (cat. no. bs-2623R, Bioss, 1:50) for 30 min. Cells were then washed and stained with Alexa Fluor 488-conjugated anti-rabbit antibody (cat. no. A11008, Invitrogen, 1:500). The culture supernatants were analyzed using IL-22 ELISA kits (R&D Systems).

Wound-healing assay

hMSCs passage 3 (0.25 × 10⁶ cells) were seeded in six-well plates and cultured to confluence. Cells were stained with CellTracker™ Green (Thermo Fisher Scientific, Waltham, MA, USA), and scratch wounds were made using a sterile 200-μl pipette tip. Detached cells were removed by washing with PBS, followed by the addition of fresh growth medium. Cells were imaged using a fluorescence microscope at 0 and 24 h after wounding. Wound areas were quantified using ImageJ.

Propidium iodide-annexin V apoptosis assay

Apoptosis was assessed using the BD Fluorescein Isothiocyanate Annexin V Apoptosis Detection Kit I (BD Biosciences) according to the manufacturer’s instructions. Briefly, hMSCs were treated with IL-22 (50 ng/ml), a concentration previously shown to reduce cell growth in murine thymic epithelial cells [51], then harvested and incubated with fluorescein isothiocyanate-annexin V and propidium iodide in the dark for 15 min. Labeled cells were analyzed using a Canto II cytometer (BD Biosciences).

Statistical analysis

Experiments included three to five mice per group for in vivo studies and three wells per group for in vitro studies, and the data are representative of at least three independent experiments. Between-group differences were assessed using the unpaired Student’s t-test. Comparisons among three or more groups were analyzed using one-way analysis of variance (ANOVA), followed by Tukey’s post hoc analysis where appropriate. Comparisons between WT and KO mice under PBS or LPS stimulation were performed using two-way ANOVA followed by Tukey’s post hoc analysis. All statistical analyses were performed using GraphPad Prism 8.02 (La Jolla, CA, USA), and p-values < 0.05 were considered statistically significant.

Results

Transcriptome analysis in mouse BM following TLR4-dependent bone destruction

In the inflamed bone microenvironment, cytokines produced by BM-resident and circulating immune cells modulate the bone niche and alter bone homeostasis by regulating bone cell differentiation [5, 17]. To assess transcriptional changes associated with TLR4 activation, the gene expression profiles of BM cells from WT and TLR4 KO mice following TLR4-driven bone destruction after PBS or LPS administration were compared using messenger RNA sequencing. Micro-CT analysis revealed that LPS stimulation induced severe bone loss with a reduction of bone mineral density, bone volume per total volume, and trabecular thickness in WT mice, whereas it was mild in TLR4 KO (Fig. 1A, B). Under these bone-destruction conditions, the majority of genes upregulated or downregulated by LPS treatment in BM of WT mice were not significantly changed in KO mice, as shown in Fig. 1C and D, demonstrating TLR4-dependency in LPS-induced signaling pathway activation. This phenomenon was also illustrated in principal component analysis (PCA) plots of gene expression data of 12 samples (Fig. 1E). Of note, samples from TLR4 KO mice clustered separately from WT samples, suggesting that TLR4 KO induces distinct gene expression changes regardless of LPS treatment. We then analyzed differential gene expression by LPS treatment with DESeq2 [52]. In WT mice, LPS stimulation induced 2548 differentially expressed genes (DEGs) (fold change > 2, adjusted p-value < 0.05) and 1398 upregulated and 1150 downregulated genes (Supplementary Table S1; volcano plot shown in Fig. 1F). In contrast, the transcriptomic response was markedly attenuated in TLR4 KO mice, with only 191 genes identified as DEGs, including 99 upregulated and 92 downregulated under the same criteria (Supplementary Table S2; volcano plot shown in Supplementary Fig. S1A). We also sought to identify common biological functions of DEGs induced by LPS treatment. We performed Gene Ontology (GO) enrichment analysis from DEGs between samples from PBS and LPS treatment in WT mice using ClusterProfiler [53] (Fig. 1G). As expected, GO terms related to immune cell function and hematopoiesis were the most significantly enriched biological functions, and GO terms related to cell proliferation (mitotic cell cycle phase transition, chromosome segregation) were highly influenced by LPS. Intriguingly, T cell-related GO terms (regulation of T cell activation, T cell differentiation) were also included. Notably, GO enrichment analysis of DEGs in TLR4 KO mice predominantly identified cytoskeleton-related GO terms (Supplementary Fig. S1B). We then compared the expression of T cell marker genes such as Cd3g, Cd3dCD3D, Cd3e (T cells), Cd4 (Th cells), Cd8a (cytotoxic T cells), and Foxp3 (regulatory T cells) to investigate whether T cell infiltration occurred in the inflamed BM. LPS treatment significantly upregulated Cd3g (p < 0.001) and Cd3d (p = 0.012) expression, whereas Cd3e showed a similar increase that did not reach statistical significance (p = 0.055) (Fig. 1H).

Fig. 1.

LPS treatment induces transcriptional changes in mouse bone marrow (BM). A Femurs from WT and TLR4 KO mice injected with PBS or LPS (5 mg/kg) were isolated and fixed in 4% paraformaldehyde. Trabecular bone structure was analyzed by micro-computed tomography. B Bone mineral density (BMD) and three-dimensional morphometric parameters, including bone volume per tissue volume (BV/TV) and trabecular thickness (Tb.Th.) were quantified (three to five mice per group). p-Values were calculated using two-way analysis of variance (ANOVA). Data are presented as the median with 95% confidence interval. **p < 0.01, ***p < 0.001 versus control. C, D Heat map visualization of messenger RNA sequencing data from BM cells of WT and TLR4 KO mice injected twice weekly with PBS or LPS (5 mg/kg). Genes upregulated in LPS-treated WT mice (C) and genes downregulated in LPS-treated WT mice (D) are shown. E PCA plot of RNA sequencing data from 12 samples from BM cells of WT and TLR4 KO mice, treated with PBS or LPS. F Volcano plot showing log fold change and −log₁₀-adjusted p-values of DEGs between PBS or LPS treatment in WT mice analyzed using DESeq2. G GO enrichment analysis results from DEGs using ClusterProfiler between PBS or LPS treatment in WT mice. H T cell marker gene expression changes with LPS treatment in WT mice. Values are normalized count with DESeq2 vst function

TLR4-dependent increase of IL-22-producing T cells in the mouse BM and induction of IL-22 production from T cells by TLR4 stimulation

Analysis of cytokine-related transcriptional changes identified 19 upregulated and 9 downregulated genes in WT mice after LPS stimulation (fold change > 3.5). Among these, nine genes showed no significant change in TLR4 KO mice, confirming their TLR4-dependent regulation (Fig. 2A). Notably, IL-22 and interferon- γ (IFN-γ) were significantly upregulated by LPS in WT (Fig. 2B, C) mice but remained unchanged in TLR4 KO mice. Although IFN-γ is known to interact with TLR4 activation in regulating bone homeostasis during inflammation [54, 55], the role of IL-22 downstream of TLR4 activation in the bone niche remains unclear. Therefore, this study focused on elucidating the role of IL-22 in this context. Since multiple immune cell types—including B cells, macrophages, ILCs, and T cells—circulate within the BM microenvironment [56, 57], the elevated IL-22 levels observed in the BM may reflect an increase in IL-22-producing cells. To examine this, the expression of genes associated with Th17, Th22, and ILC3 subsets, all of which are known sources of IL-22, were evaluated. LPS stimulation upregulated genes related to Th17/Th22 differentiation (Fig. 2D). Consistently, the frequency of IL-22⁺CD4⁺ T cells increased in the BM after LPS administration, indicating that enhanced IL-22 levels were attributable to increased IL-22-producing T cells (Fig. 2E, F). Protein analysis confirmed elevated IL-22 in BM fluid from LPS-treated WT mice, whereas no changes were detected in TLR4 KO mice (Fig. 2G). To determine whether LPS directly induces IL-22 production in T cells, CD3⁺CD4⁺ splenic T cells isolated from C57BL/6J mice were stimulated with LPS for 3 days (Fig. 2H). LPS increased IL-22 expression, as indicated by higher MFI, and elevated secreted IL-22 levels, as measured by ELISA (Fig. 2I, J). Together, these findings demonstrate that TLR4 activation enhances IL-22 production in BM by increasing IL-22⁺ T cell populations in WT mice.

Fig. 2.

LPS stimulation increases IL-22-producing T cells in the BM and enhances IL-22-producing capacity of T cells. A Venn diagram showing immune response-related genes upregulated or downregulated in LPS-treated WT mice relative to PBS controls (blue circle), and in LPS-treated TLR4 KO mice relative to LPS-treated WT mice (yellow circle). Contra-regulated genes appear in the intersecting region. B Heat map of genes exhibiting opposing expression patterns between LPS-treated WT versus PBS groups and LPS-treated TLR4 KO versus LPS-treated WT groups. C Scatter plot of immune response-related gene expression (axis: average of normalized data log₂). D Clustered heat map of messenger RNA sequencing data showing the expression of genes associated with Th17, Th22, and ILC3 in BM from WT and TLR4 KO mice following PBS or LPS (5 mg/kg) injection twice over a week. E, F Flow cytometric analysis of IL-22⁺CD4⁺ T cells in the BM of WT and TLR4 KO mice treated with PBS or LPS (E), with corresponding quantification of IL-22⁺CD4⁺ cells (four to five mice per group) (F). G IL-22 protein levels in BM fluid from WT and TLR4 KO mice treated with PBS or LPS, measured by enzyme-linked immunosorbent assay (ELISA) (four to five mice per group). F, G p-Values were calculated using two-way ANOVA. H–J CD3⁺CD4⁺ T cells were isolated from spleens by flow cytometric sorting and cultured with or without LPS (1 μg/ml) in the presence of anti-CD3 and anti-CD28 antibodies for 3 days (H). IL-22 production was assessed by flow cytometry and quantified as mean fluorescence intensity (n = 3 wells per group) (I). IL-22 levels in culture supernatants were measured by ELISA (n = 3 wells per group) (J). I, J p-Values were calculated using Student’s t-test. Data are presented as the median with 95% confidence intervals. *p < 0.05 versus control

Decreased inflammatory bone destruction by limiting OC formation in IL-22-deficient mice

Systemic LPS administration is a well-established murine model of inflammatory bone loss, characterized by bone inflammation followed by progressive bone destruction [17]. As TLR4 activation increased IL-22 in the BM, it was hypothesized that IL-22 may be a cross-talk factor mediating inflammatory bone destruction. To directly test this hypothesis, the in vivo bone phenotype following TLR4 activation was compared in WT and IL-22 KO mice injected with PBS or LPS. In WT mice, LPS stimulation resulted in markedly reduced trabecular bone density, a phenotype that was largely absent in IL-22 KO mice (Fig. 3A). Consistent with this, bone mineral density, bone volume per total volume, trabecular thickness, and trabecular number were significantly decreased, whereas trabecular spacing and total porosity were significantly increased in LPS-injected WT mice (Fig. 3B). TRAP staining of tibial sections further confirmed enhanced TRAP⁺ area and reduced cancellous bone mass in LPS-treated WT mice compared with PBS controls, while IL-22 KO mice showed no such differences after LPS stimulation (Fig. 3C). Together, these findings suggest that IL-22 induced by LPS regulates OC activity and contributes to inflammatory bone destruction.

Fig. 3.

IL-22 deficiency attenuates bone loss induced by toll-like receptor 4 activation in vivo. A Femurs from WT and IL-22 KO mice injected with PBS or LPS (5 mg/kg) were isolated and fixed in 4% paraformaldehyde. Trabecular bone structure was analyzed by micro-computed tomography. B BMD and three-dimensional morphometric parameters, including BV/TV, Tb.Th., trabecular number (Tb.N.), trabecular separation (Tb.Sp.), and total porosity were quantified (four to five mice per group). C Tibiae from WT and IL-22 KO mice treated with PBS or LPS (5 mg/kg) were processed for hematoxylin and eosin (H&E) or TRAP staining. TRAP⁺ areas were quantified (four to five mice per group). Scale bar = 100 μm. B, C p-Values were calculated using two-way ANOVA. Data are presented as the median with 95% confidence intervals. *p < 0.05, **p < 0.005 versus control

Stimulation of RANKL production by IL-22 in pre-OBs during osteogenesis

We next sought to obtain a mechanistic explanation for how IL-22 acts on BM-resident cells in the BM and ultimately disrupts bone homeostasis. As IL-22 has been implicated in regulating MSC proliferation, migration, and differentiation under inflammatory conditions [45], its direct effects on MSCs were first examined. IL-22 treatment did not alter apoptotic cell death in hMSCs (Fig. 4A). In contrast, IL-22 markedly increased hMSC migration into the wound site after 24 h, indicating enhanced hMSC migration (Fig. 4B), a prerequisite for OB differentiation during bone repair [58]. The osteogenic potential of IL-22 was then evaluated by assessing ALP activity, an early marker of osteogenic commitment, and AR, a marker of late mineralization [59]. hMSCs cultured in osteogenic media with IL-22 showed significantly higher ALP staining than controls (Fig. 4C), and this effect extended to enhanced AR staining, indicating increased mineral deposition (Fig. 4D). Similarly, IL-22 promoted osteogenesis in pre-OBs isolated from mouse calvaria, as shown by elevated ALP-positive and AR-positive mature OBs (Fig. 4E), along with increased expression of osteogenic marker genes (Runx2, Alp, and Bglap) (Fig. 4F). Notably, IL-22 induced robust secretion of soluble RANKL with a concurrent reduction in OPG, resulting in a significantly elevated RANKL/OPG ratio (Fig. 4G–I). These findings suggest that IL-22 enhances osteoclastogenic potential by stimulating RANKL production in pre-OBs during osteogenesis. We thus reasoned that IL-22 may promote OC differentiation from BM monocytes/macrophages in co-culture condition with pre-OBs. IL-22 treatment significantly increased OC formation in a dose-dependent manner in the co-culture system using calvarial pre-OBs (Fig. 5A), as evidenced by a significant increase in the number of TRAP⁺ multinucleated cells (Fig. 5B). In contrast, IL-22 had no effect on the OC formation induced by RANKL (Fig. 5C, D), indicating that IL-22 does not directly influence OC differentiation. Collectively, these findings indicate that LPS-induced bone loss is dependent on IL-22 and support a model in which IL-22 indirectly promotes OC-mediated bone destruction by stimulating pre-OB-derived RANKL.

Fig. 4.

IL-22 promotes osteoblastic differentiation and modulates RANKL/OPG expression in pre-OBs. A Apoptosis of MSCs treated with IL-22 (50 ng/ml) was assessed using a propidium iodide-annexin V assay kit (n = 3 wells per group). A p-Values were calculated using Student’s t-test. B MSC motility following IL-22 stimulation (10, 20, or 50 ng/ml) for 24 h was evaluated using a wound-healing assay. Quantification represents the percentage of wound closure, visualized by fluorescence microscopy (n = 3 wells per group). Scale bar = 200 μm. C, D MSCs were cultured in an osteogenic medium with or without IL-22 (10, 20, and 50 ng/ml) for 2 weeks. Osteoblastic differentiation was assessed by alkaline phosphatase (ALP) staining (C) and alizarin red (AR) staining (D) (n = 3 wells per group). (B–D) p-Values were calculated using one-way ANOVA. E Calvarial pre-OBs were cultured in an osteogenic medium with or without IL-22 (10 ng/ml) for 2 weeks, followed by ALP and AR staining to assess osteogenic differentiation (n = 3 wells per group). F Messenger ribonucleic acid expression levels of Runx2, Alp, and Bglap in pre-OBs stimulated with IL-22 (10 ng/ml) were analyzed (n = 3 wells per group). G–I Culture supernatants from pre-OBs treated with or without IL-22 (10 ng/ml) were analyzed by enzyme-linked immunosorbent assay to determine RANKL (G) and OPG (H) levels, and the RANKL/OPG ratio (I) (n = 3 wells per group). E–I p-Values were calculated using Student’s t-test. Data are presented as the median with 95% confidence intervals. *p < 0.05, **p < 0.01, ***p < 0.001 versus control. All representative data from three independent experiments are shown

Fig. 5.

IL-22 promotes OC differentiation by stimulating RANKL production in pre-OBs. A, B BM cells were co-cultured with calvarial pre-OBs from WT mice with PBS or IL-22 (10, 25, and 50 ng/ml) along with VtD₃ (10⁻⁸ M) and PGE2 (10⁻⁶ M) for 5 days (A). Cells were fixed and stained for tartrate-resistant acid phosphatase (TRAP), and TRAP⁺ multinucleated cells (MNCs) containing ≥ 3 nuclei were quantified (n = 3 wells per group) (B). Scale bar = 100 μm. C, D BM monocytes/macrophages were cultured for 4 days with PBS or IL-22 (10, 25, or 50 ng/ml) in the presence of M-CSF (30 ng/ml) and RANKL (100 ng/ml) (C). Cells were fixed with 4% paraformaldehyde, stained for TRAP, and TRAP⁺ MNCs containing ≥ 3 nuclei were counted under a light microscope (n = 3 wells per group) (D). Scale bar = 100 μm. B, D p-Values were calculated using one-way ANOVA. Data are presented as the median with 95% confidence intervals. *p < 0.05, **p < 0.005 versus control. B and D shows images representative of a minimum of three independent experiments

Induction of RANKL production in pre-OBs via IL-22-mediated COX-2/PGE2 pathway activation

We next determined the mechanism by which IL-22 induces RANKL production in pre-OBs. IL-22 activates the Jak1–STAT signaling pathway, employing Jak1 and tyrosine kinase 2 to mediate downstream MAPK signaling pathways and STAT1, STAT3, and STAT5 transcription factors [60]. In pre-OBs, IL-22 strongly induced ERK activation and gradually increased p-p38 levels, whereas JNK activation remained unchanged (Fig. 6A). IL-22 did not affect STAT1 or STAT3 phosphorylation and only modestly affected STAT5 phosphorylation (Fig. 6B), indicating that IL-22 primarily activates ERK and p38 signaling in pre-OBs. Furthermore, inhibition of ERK (PD98059) and p38 (SB202190) pathways significantly reduced IL-22-induced RANKL expression, whereas JNK (SP600125) inhibition had no effect (Fig. 6C, D). COX-2, a key enzyme in prostaglandin synthesis, is known to modulate RANKL production in OBs, thereby influencing osteoclastogenesis and bone remodeling [3, 61]. To elucidate the mechanism by which IL-22 promotes RANKL expression in pre-OBs, COX-2 expression following IL-22 treatment was evaluated, and it was found that IL-22 upregulated COX-2 expression in pre-OBs (Fig. 6E, F). Since MAPK and STAT pathways contribute to COX-2 upregulation, a key regulator of PGE2-mediated RANKL production [3], the effects of pathway-specific inhibitors were next examined. IL-22-induced COX-2 expression and PGE2 production were significantly suppressed by ERK and p38 inhibition, whereas JNK inhibition had no effect (Fig. 6G, H). In addition, the IL-22-induced increase in PGE2 was blocked by the COX-2 inhibitor (NS398) (Fig. 6I). Consistently, COX-2 inhibition suppressed IL-22-induced RANKL expression at the messenger RNA and protein levels (Fig. 6J, K). Together, these findings suggest that IL-22 increases RANKL production in pre-OBs via COX-2-dependent PGE2 production, thereby promoting OC formation.

Fig. 6.

IL-22 enhances OC differentiation via pre-OB RANKL secretion mediated by COX-2/PGE2. A, B Pre-OBs were serum-starved for 2 h and stimulated with IL-22 (10 ng/ml) for the indicated times. Total protein extracts were analyzed by western blotting using antibodies against pERK, ERK, p-p38, p38, p-JNK, and JNK (A) and p-STAT1, STAT1, p-STAT3, STAT3, p-STAT5, and STAT5 (B). C, D Pre-OBs were treated with IL-22 (10 ng/ml) in the presence of PD98059 (PD, 10 μM), SB202190 (SB, 10 μM), and SP600125 (SP, 10 μM). RANKL expression was assessed by quantitative polymerase chain reaction (qPCR) C and enzyme-linked immunosorbent assay (ELISA) D (n = 3 wells per group). C, D p-Values were calculated by one-way ANOVA. E, F IL-22 (10 ng/ml) treatment for 24 h induced cyclooxygenase-2 (COX-2) expression, as measured by qPCR (E) and western blotting (F), with densitometric quantification shown (n = 3 wells per group). E, F p-Values were calculated using Student’s t-test. G, H Pre-OBs treated with PD (10 μM), SB (10 μM), or SP (10 μM) were analyzed for COX-2 mRNA expression by qPCR (G) and prostaglandin E₂ (PGE2) expression by ELISA (H) (n = 3 wells per group). I-K Pre-OBs were treated with or without NS398 (1 μM) in the presence or absence of IL-22 (10 ng/ml) for 24 h. PGE2 levels were measured by ELISA (I), RANKL mRNA expression by qPCR (J), and secreted RANKL levels by ELISA (K) (n = 3 wells per group). G–K p-Values were calculated using one-way ANOVA. Data are presented as the median with 95% confidence intervals. *p < 0.05, **p < 0.01, ***p < 0.001 versus control

Discussion

It has been suggested that T cells in the BM may contribute to bone loss by releasing inflammatory cytokines such as TNF-α and expressing RANKL, thereby stimulating OC to degrade bone in pathological conditions in the context of disease such as osteoporosis or certain cancers [62, 63]. Elevated Th22 frequencies in the BM of patients with multiple myeloma correlate with a dysregulated cytokine environment and poor prognosis [64]. In osteoporosis condition, increased T cell production of TNF-α due to increased BM T cell numbers has been suggested as a major mechanism by which estrogen deficiency induces bone loss in vivo [62]. Additionally, glucocorticoid (steroid) treatment induces T cell accumulation in the BM by increasing RANKL expression and contributes to bone loss, but the underlying mechanism by which glucocorticoid upregulates RANKL in T cells remains unknown [65]. Although Th22 and Th17 cell frequencies are higher in the BM than in the peripheral blood of patients with RA, these associations [66] do not establish a causal role for IL-22 in inflammatory bone destruction. In the present study, direct evidence was provided that TLR4 activation increases IL-22 levels through increased infiltration of IL-22-producing cells into the inflamed BM, thereby promoting bone destruction in vivo (Figs. 2, 3). These findings suggest a pathological role for increased IL-22 as a key mechanism in inflammatory systemic bone destruction.

A lack of TLR4 on bone marrow cells has been shown to ameliorate arthritis symptoms by decreasing attraction of γδ T cells to inflamed joints and subsequent IL-17 production, highlighting the importance of the relationship between BM and TLR4 in T cell responses in disease pathogenesis [67]. The findings of this study demonstrate that IL-22-producing T cells increase within the BM niche under inflammatory conditions and can directly respond to local inflammatory signals—such as TLR4 agonists—to drive severe inflammatory bone destruction. IL-22 is primarily produced by lymphoid cells, including CD4⁺ T cell subsets such as Th17 and Th22 [68, 69], natural killer cells, and ILC3 [70, 71]. Nonlymphoid cells, including neutrophils, mast cells, macrophages, and fibroblasts, can also secrete IL-22, exerting either protective or pathological effects, depending on the disease context [72–75]. For example, necrotic cell-derived TLR4 agonists stimulate intrarenal mononuclear cells to secrete IL-22, promoting tubular regeneration and recovery in acute kidney injury [76]. Consistent with this notion, we speculate that the increase in IL-22 in the BM in response to TLR4 activation is due to an increase in the number of infiltrated IL-22-producing T cells and an upregulation of the amount of IL-22 produced by each active T cell upon TLR4 stimulation (Fig. 2).

Previous studies have shown that IL-22 can directly promote osteoclastogenesis from human monocytes [77, 78]. In contrast, our findings revealed that IL-22 had no direct effect on OC differentiation from murine BM-derived monocytes/macrophages (Fig. 5D), while exerting a protective effect against inflammatory bone loss in vivo (Fig. 3A, B). This discrepancy likely reflects fundamental species-specific differences in OC precursor biology. Human OCs are derived from relatively mature circulating monocytes that comprise heterogeneous subtypes previously exposed to systemic inflammatory signals, thereby displaying variable responsiveness to osteoclastogenic stimuli [79]. By contrast, murine OC precursors used in this study comprised relatively homogeneous BM-derived monocyte populations isolated under controlled conditions using established methods [46]. Such biological and developmental differences may explain the divergent findings between species. Importantly, our experimental model enabled evaluation of the specific effects of IL-22 on RANKL-induced OC differentiation while minimizing confounding influences from other factors. Notably, our findings indicate that IL-22 indirectly regulates bone homeostasis by modulating RANKL expression in pre-OBs, thereby promoting OC formation under co-culture conditions (Figs. 4G and 5B). A similar mechanism has been reported in RA, where elevated IL-22 levels in synovial fluid induce RANKL expression in synovial fibroblasts [80]. However, unlike previous studies focused primarily on localized autoimmune inflammation, our findings identify a novel TLR4–IL-22 axis operating within the BM niche that links systemic inflammation to dysregulated bone remodeling. In addition, IL-22 receptor expression was observed throughout the osteogenic lineage, including MSCs, pre-OBs, and mature OBs, across multiple stages of differentiation (data not shown). These observations suggest that OB-lineage cells are the primary responders to IL-22. This distinction is important, as it indicates that IL-22 does not directly stimulate OC precursors during systemic inflammatory conditions but instead promotes bone resorption indirectly through modulation of the RANKL/OPG ratio in pre-OB. Collectively, these findings define a regulatory axis through which systemic inflammation may contribute to pathological bone loss.

Mechanistically, IL-22 strongly induced phosphorylation of ERK and p38 in pre-OBs, whereas no appreciable activation of STAT1 or STAT3 was detected, and STAT5 activation was minimal (Fig. 6A, B). Although IL-22 is known to activate STAT3 [37, 81], previous studies have also shown that it can engage protein kinase B and MAPK pathways [35, 60]. Consistent with this signaling profile, pharmacological inhibition using PD98059, SB202190, and SP60025 demonstrated that only the ERK and p38 pathways were required for IL-22-mediated responses in pre-OBs. Furthermore, IL-22 significantly upregulated COX-2 expression in pre-OBs (Fig. 6E, F). COX-2 is a rate-limiting enzyme in PGE2 biosynthesis and plays a critical role in OB differentiation and RANKL production, thereby promoting OC formation [3, 61]. Although COX-2 and PGE2 are crucial for bone homeostasis, persistent PGE2 signaling contributes to bone resorption; accordingly, COX-2 inhibitors have been shown to increase bone mineral density in postmenopausal women by inhibiting continuous PGE2-driven bone loss [82]. Moreover, increased COX-2 expression has been observed in the BM of patients with Staphylococcus aureus osteomyelitis, where it promotes OC activity and contributes to inflammatory bone erosion [83]. Given that MAPK signaling is a well-established inducer of COX-2, which in turn regulates PGE2-mediated RANKL production, our findings suggest that IL-22 promotes COX-2/PGE2-mediated RANKL expression through ERK- and p38-mediated signaling pathways. Notably, opposing effects of IL-22 on RANKL and OPG expression were observed (Fig. 4G–I). OPG is regulated by multiple factors, including transforming growth factor-β, estrogen, and Wnt ligands, which enhance its expression, whereas PGE2 and glucocorticoids suppress it [84–87]. Therefore, IL-22 may reduce OPG expression by inducing COX-2 expression and subsequent elevation of PGE2 levels. These findings highlight the complexity of IL-22-mediated signaling within the bone microenvironment; however, the precise mechanisms underlying OPG suppression require further experimental validation.

Interestingly, IL-22-deficient mice were protected against inflammatory bone destruction (Fig. 3A, B), despite our in vitro findings showing that IL-22 promoted osteogenic differentiation of MSCs (Fig. 4C, D). BM osteoprogenitor cells, including MSCs, play essential roles in bone regeneration by migrating to sites of injury and differentiating into osteogenic cells. This process can be enhanced by inflammatory cytokines released during the early phase of fracture healing or tissue injury, including IL-22, IFN-γ, and TNF-α [45]. However, the biological effects of these cytokines are highly concentration-dependent. Although IL-22 promotes osteogenic differentiation of MSCs, this effect may be significantly attenuated in the presence of excessive levels of IFN-γ and TNF-α [45]. Low concentrations or transient exposure to TNF-α can stimulate osteogenesis through activation of signaling pathways such as NF-κB in osteoprogenitors or MSCs [88]. In contrast, high concentrations of TNF-α, which are typically present in inflamed BM environment [89], are well recognized to drive pathological bone loss by enhancing OC differentiation while simultaneously suppressing OB differentiation and function [90]. These findings underscore the dual and context-dependent role of TNF-α in bone remodeling. Given that IFN-γ and TNF-α expression increased following TLR4 activation (Supplementary Table S1), it is possible that the direct osteogenic effects of IL-22 were overridden by the in vivo microenvironment rich in these proinflammatory cytokines. Consequently, the cytokine-rich microenvironment induced by TLR4 activation may disrupt the intricate interaction between OBs and OCs through the IL-22 axis, thereby shifting bone remodeling toward net catabolic bone loss. Accordingly, the TLR4–IL-22 pathway provides mechanistic insight into the pathogenesis of inflammation-induced bone destruction and may represent a relevant target in inflammatory bone diseases. For example, ankylosing spondylitis (AS) is associated with elevated IL-22 levels [91] and is characterized by fluctuating inflammatory activity, with recurrent phases of exacerbation and resolution linked to sequential processes of inflammation-driven bone resorption followed by pathological new bone formation [92]. In contrast to RA, which is primarily driven by synovial inflammation mediated by fibroblast-like synoviocytes [93], AS predominantly involves inflammation at the enthesis and adjacent bone, where OB-lineage cells contribute to syndesmophyte formation [94]. Within this context, IL-22 may exert dual and stage-dependent effects in AS. During periods of active inflammation, particularly in the present of proinflammatory cytokines such as IFN-γ and TNF-α, IL-22 may enhance OC indirectly by inducing RANKL expression in pre-OBs. Conversely, during phases of lower inflammatory activity, IL-22 may favor OB-mediated bone formation through its osteogenic properties (Fig. 4E). Therefore, therapeutic targeting of IL-22 may offer the potential to suppress both RANKL-induced osteoclastogenesis during active inflammation and IL-22-mediated pathological bone formation during resolution phases. However, given the pleiotropic and context-dependent functions of IL-22, its overall therapeutic value in AS requires further comprehensive investigation.

Transcriptomic analysis is an established technique for the detection of the full range of RNA transcripts in a sample to infer the functions of different cell types, such as their ability to self-renew, differentiate into other cell types, or secrete cytokines [95]. Although single-cell RNA sequencing (scRNA seq) is a more specialized technique that analyzes the RNA of individual cells, here we adopted total mRNA sequencing to provide a snapshot of which genes are active in BM cells and focused on the alteration and function of T cells. We recognize, however, that the signaling networks that interact among various immune cells altered by TLR4 activation remain to be determined. Therefore, additional scRNA seq analysis may help elucidate the various physiological and clinical implications of the increased diversity of immune cells in the inflamed BM, which could further advance the field of osteoimmunology.

In summary, and as shown in Fig. 7, it is proposed that the increase in TLR4-dependent IL-22-producing T cells in the BM promotes the migration of osteoprogenitor cells and subsequent osteogenesis, potentially contributing to anabolic bone formation. However, it is noteworthy that IL-22 also drives catabolic bone destruction in vivo, despite not directly inducing OC formation, as demonstrated in this study. Notably, IL-22 enhances RANKL expression in pre-OBs while suppressing OPG production, thereby elevating the RANKL/OPG ratio and promoting OC formation within the bone microenvironment. IL-22 activates the ERK and p38 pathways, thereby upregulating COX-2/PGE2 signaling and inducing RANKL production. Herein, direct evidence and mechanism insight demonstrating that IL-22 promotes bone destruction in vitro and in vivo has been provided. Furthermore, the findings of the study broaden the current understanding of TLR4-mediated regulation of bone homeostasis by identifying IL-22 as a novel mediator within the inflamed bone niche. This may be significant in providing a novel strategy using the TLR4–IL-22 axis for the prevention and treatment of inflammatory bone destruction.

Fig. 7.

Schematic representation of IL-22 function in bone homeostasis upon LPS stimulation. LPS activates TLR4, leading to increased IL-22 production from infiltrating cluster of differentiation CD4⁺IL-22⁺ T cells in the BM. Increased IL-22 in the BM niche stimulates pre-OBs to upregulate COX-2 via ERK and p38 signaling, resulting in enhanced PGE2 production. PGE2 subsequently induces RANKL expression in pre-OBs. RANKL interacts with RANK, on monocytes/macrophages, promoting their differentiation into OCs and contributing to inflammatory bone destruction

Conclusions

LPS-induced TLR4 activation stimulates IL-22 production in T cells, which activates osteoblastic differentiation of osteoprogenitor cells. Enhanced IL-22 in the inflamed bone niche stimulates RANKL production in pre-OBs by promoting COX-2-dependent PGE2 production via ERK and p38 signaling pathways, which results in enhanced OC formation and bone destruction. These results demonstrate a comprehensive understanding of the role of IL-22 in TLR4-dependent bone destruction.

Abbreviations

TLR4

Toll-like receptor 4

BM

Bone marrow

LPS

Lipopolysaccharide

KO

Knockout

IL

Interleukin

OC

Osteoclast

OB

Osteoblast

RANKL

Receptor activator of nuclear factor-κb ligand

MSC

Mesenchymal stromal cell

ILC

Innate lymphoid cell

pre-OB

Immature osteoblast

NF-κB

Nuclear factor-κb

OPG

Osteoprotegerin

TNF-α

Tumor necrosis factor-α

RA

Rheumatoid arthritis

ERK

Extracellular signal-regulated kinase

Th

T helper

Jak

Janus kinase

MAPK

Mitogen-activated protein kinase

JNK

c-Jun N-terminal kinase

STAT

Signal transducer and activator of transcription

CD

Cluster of differentiation

COX-2

Cyclooxygenase-2

PGE2

Prostaglandin E2

WT

Wild type

PBS

Phosphate-buffered saline

hMSC

Human MSC

ALP

Alkaline phosphatase

AR

Alizarin red

M-CSF

Macrophage colony-stimulating factor

TRAP

Tartrate-resistant acid phosphatase

cDNA

Complementary DNA

PCR

Polymerase chain reaction

GAPDH

Glyceraldehyde-3-phosphate dehydrogenase

Runx2

Runt-related transcription factor 2

Bglap

Bone gamma-carboxyglutamate protein

mRNA

Messenger RNA

ELISA

Enzyme-linked immunosorbent assay

PCA

Principal component analysis

DEG

Differentially expressed gene

GO

Gene Ontology

IFN-γ

Interferon γ

scRNA seq

Single-cell RNA sequencing

Ctsd

Cathepsin D

Klf2

Kruppel-like factor 2

Ifit1bl2

Interferon induced protein with tetratricopeptide repeats 1B-like 2

Kif6

Kinesin family member 6

Abcc3

ATP-binding cassette subfamily C member 3

Ccpg1

Cell cycle progression 1

Ube2l6

Ubiquitin-conjugating enzyme E2 L6

Il7r

Interleukin 7 receptor

Foxp3

Forkhead box P3

Jun

Jun proto-oncogene

Tpsb2

Tryptase beta 2

F3

Coagulation factor 3

IL-22

Interleukin 22

Epha2

Eph receptor A2

IL-33

Interleukin 33

Ctsw

Cathepsin W

Adam19

A disintegrin and metalloproteinase domain-containing protein 19

Bmpr2

Bone morphogenetic protein receptor type II

Igf1r

Insulin-like growth factor 1 receptor

Scpep1

Serine carboxypeptidase 1

Gzma

Granzyme A

Ccl5

C–C motif chemokine ligand 5

Rorc

RAR-related orphan receptor C

Lst1

Leukocyte-specific transcript 1

Tnfsf

Tumor necrosis factor superfamily

S100a4

S100 calcium-binding protein A4

Nr1d1

Nuclear receptor subfamily 1 group D member 1

mIL

Mouse interleukin

FACS

Fluorescence-activated cell sorting

BMD

Bone mineral density

BV/TV

Bone volume per tissue volume

Tb.Th.

Trabecular thickness

Tb.N.

Trabecular number

Tb.Sp.

Trabecular separation

H&E

Hematoxylin and eosin

OM

Osteogenic medium

FITC

Fluorescein isothiocyanate

MNCs

Multinucleated cells

VtD₃

1α,25-Dihydroxyvitamin D₃

p

Phosphorylated

PD

PD98059

SB

SB202190

SP

SP600125

NS

NS398

EP

Prostaglandin E receptor

RANK

Receptor activator of nuclear factor-κB

MFI

Mean fluorescence intensity

AS

Ankylosing spondylitis

ANOVA

Analysis of variance

Ct

cycle threshold

Author contributions

SWK and EJC conceptualized the study; SWK and EJC supervised the study; JEK, JYG, EJL, and YJL curated the data; JEK, EJL, and EJC wrote the original draft; JEK, JYG, SWK, and EJC reviewed and edited the manuscript; JEK, JYG, EJL, SMK, DWJ, HWN, MSH, YJL, YLS, MKJ, YJH, and EWL performed the investigations. All authors read and approved the final manuscript.

Funding

This work was supported by the Basic Science and Engineering Research Program (2022R1A2C1009522) funded by the Korean government (MSIT) to Eun-Ju Chang and supported by grant (2025IP0005) from the Asan Institute for Life Sciences, Asan Medical Center to Seong Who Kim and Eun-Ju Chang.

Data availability

The datasets used and/or analyzed during the present study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

All animal procedures followed protocols approved by the Institutional Committee for the Care and Use of Laboratory Animals at Ulsan University (approved no. 2016–13-315, 9 January 2017). All studies were conducted in accordance with the relevant guidelines of the International Council for Laboratory Animal Science (ICLAS).

Consent for publication

All authors read and approved the final manuscript.

Competing interests

The authors declare no competing interests.