Osteocytic Lipocalin-2 regulates bone formation locally through iron-dependent ferroptosis and Wnt suppression
Vivek Khanal, Madeline Carroll, Fatemeh Moradi, Jayden Carter, Ying Zhong, Chikkamagaluru G. Shashank, Amy Y. Sato, Ryan M. Allen, Umesh D. Wankhade, Neha S. Dole
TL;DR
This study shows that the protein Lipocalin-2 in bone cells causes local bone weakness by triggering cell death and blocking bone growth signals.
Contribution
The novel finding is that LCN2 regulates bone formation locally via ferroptosis and Wnt suppression, independent of systemic metabolism.
Findings
LCN2 promotes iron accumulation and ferroptosis in osteocytes via SLC22A17.
Deleting LCN2 improves mitochondrial function and increases bone connectivity.
LCN2 deletion enhances bone formation by suppressing Wnt antagonists DKK1 and SOST.
Abstract
Osteocytes, the most abundant bone cells, are central regulators of bone remodeling that also exert endocrine control over systemic metabolism. Among the factors they produce, Lipocalin-2 (LCN2) has emerged as a cytokine linking bone and energy homeostasis, yet its local role within the skeleton remains elusive. Here, we identify that LCN2 promotes intracellular iron accumulation, mitochondrial dysfunction, and lipid peroxidation through its receptor SLC22A17, and drives ferroptotic cell death. Dmp1-Cre–mediated deletion of Lcn2 preserves mitochondrial integrity, reduces intracellular iron and lipid peroxidation, and enhances osteocyte dendricity and lacunocanalicular connectivity. Mechanistically, loss of Lcn2 suppresses Wnt antagonists DKK1 and SOST, thereby promoting Wnt/β-catenin signaling and stimulating osteoblast-mediated bone formation. Notably, Dmp1-Cre-mediated deletion of…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9- —https://doi.org/10.13039/100000057U.S. Department of Health & Human Services | NIH | National Institute of General Medical Sciences (NIGMS)
- —https://doi.org/10.13039/100007917United States Department of Agriculture | Agricultural Research Service (USDA Agricultural Research Service)
- —https://doi.org/10.13039/100000090United States Department of Defense | United States Army | Army Medical Command | Congressionally Directed Medical Research Programs (CDMRP)
- —https://doi.org/10.13039/100006108U.S. Department of Health & Human Services | NIH | National Center for Advancing Translational Sciences (NCATS)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsFerroptosis and cancer prognosis · Bone Metabolism and Diseases · Wnt/β-catenin signaling in development and cancer
Introduction
Lipocalin-2 (LCN2) is a multifunctional secreted glycoprotein with broad biological roles in immunity, metabolism, and cellular stress regulation [1, 2]. Within the skeleton, LCN2 is one of the few osteokines known to influence systemic energy balance [3, 4]. Osteoblast-derived LCN2 crosses the blood–brain barrier to activate hypothalamic melanocortin 4 receptor, thereby suppressing appetite and regulating body weight [3–6]. Despite its well-defined endocrine role, however, LCN2’s local skeletal function remains poorly understood. Global deletion of Lcn2 leads to osteopenia with impaired osteoblast differentiation and reduced bone formation in one study [7], but not in another [3]. Similarly, osteoblast-specific Lcn2 overexpression induces bone loss by suppressing osteoblast activity and enhancing osteoclastogenesis in one model [5], yet has no detectable skeletal effect in another [4]. These conflicting results may reflect context-dependent actions of LCN2 in bone and/or differences in the underlying genetic models.
Beyond its endocrine effects, LCN2 contributes to innate immunity by sequestering iron-loaded siderophores and limiting bacterial growth [2]. While iron is essential for normal cellular function, its excess catalyzes the Fenton reaction, generating reactive oxygen species (ROS) that damage lipids, proteins, and DNA. In chronic kidney disease and cancer, elevated LCN2 has been implicated in exacerbating iron-driven oxidative stress [8–10]. In these settings, LCN2 delivers non-transferrin-bound iron (NTBI) into cells via its receptor SLC22A17, promoting intracellular iron accumulation and oxidative stress that ultimately leads to ferroptosis, an iron-dependent form of regulated cell death. Apart from increased intracellular iron load, ferroptosis is also characterized by excessive lipid peroxidation and reduced expression and function of antioxidant defense genes, particularly glutathione peroxidase 4 (Gpx4) and system Xc⁻(Xct) [11]. Importantly, ferroptosis has been implicated in osteoporosis associated with aging, glucocorticoid exposure, and diabetes [12–14], conditions that also show elevated circulating LCN2 [4, 15–17]. Whether LCN2 directly regulates ferroptosis in bone cells and in maintaining skeletal homeostasis remains elusive.
In this study, we focused on osteocytes, which comprise more than 95% of bone cells [18] and were recently identified as a source of LCN2 [19]. We found that LCN2 promotes intracellular iron accumulation, mitochondrial dysfunction, and lipid peroxidation through its receptor SLC22A17. Deletion of Lcn2 in Dmp1-Cre-targeted cells preserved mitochondrial integrity, prevented lipid peroxidation, and enhanced osteocyte dendricity and lacunocanalicular network area. Mechanistically, loss of LCN2 suppressed Wnt antagonists DKK1 and SOST and enhanced Wnt/β-catenin signaling, leading to increased osteoblast-mediated bone formation. Importantly, Lcn2 deletion with Dmp1-Cre did not alter systemic energy balance, indicating a predominantly local skeletal function of LCN2 in bone.
Results
LCN2 promotes osteocyte ferroptosis
LCN2 expression by pre-osteoblasts and osteoblasts has been well documented [3]; however, whether osteocytes continue to abundantly express LCN2 remains unclear. To address this, we first used in vitro osteogenic differentiation models to define Lcn2 mRNA levels across the osteoblast-osteocyte lineage. Murine bone marrow stromal cells (BMSCs) were induced to undergo osteogenic differentiation using β-glycerophosphate and ascorbate. mRNA levels of osteoblast differentiation marker Runx2 and early osteocyte markers Phex and Dmp1 significantly increased by week 2, whereas mature osteocyte markers Fgf23 and Sost were detected only at weeks 3 and 4 of differentiation (Fig. 1A). Notably, Lcn2 mRNA levels progressively increased throughout differentiation, with mature osteocytes at weeks 3 and 4 exhibiting robust Lcn2 expression (Fig. 1A). These findings were further validated in OCY454 osteocyte-like cells, a well-established in vitro model of osteocytes [20]. Differentiation of OCY454 cells for two weeks resulted in a marked increase in osteocyte markers Dmp1, Phex, and Fgf23 (Fig. 1B), confirming their osteocytic identity. Consistently, Lcn2 mRNA levels were upregulated during the differentiation of OCY454 cells (Fig. 1B), indicating that osteocytes retain and even augment LCN2 expression along the osteogenic lineage.Fig. 1LCN2 promotes ferroptosis by increasing iron accumulation, oxidative stress, lipid peroxidation, and cell death in osteocytes.qPCR analysis of Lcn2 and osteoblast and osteocyte markers (Phex, Runx2, Dmp1, Fgf23, Sost) during osteogenic differentiation of A BMSCs and B OCY454 osteocyte-like cells. p < 0.05 vs. week 0, n = 4 biological replicates per time point. C–I OCY454 cells were treated with recombinant murine LCN2 (rmLCN2, 100 ng/mL) ± the ferroptosis inhibitor Deferoxamine (DFO, 100 µM) for 24 h. C Quantification of FerroOrange staining showing elevated Fe²⁺ accumulation; D, E total and Fe³⁺ iron concentrations; F, G DCFDA fluorescence showing increased ROS; H, I C11-BODIPY fluorescence indicating lipid peroxidation. DFO co-treatment rescued all parameters. n = 3–4 biological replicates per group. J Annexin V/PI flow-cytometric analysis showing increased cell death with LCN2, fully rescued by DFO but not by inhibitors of apoptosis (DEVD), necroptosis (Necrostatin-1), or pyroptosis (VX-765), confirming ferroptosis as the predominant mode of cell death. K–P Loss-of-function validation in OCY454 cells transfected with Lcn2 shRNA or scrambled control. K, L, M qPCR and Western blot confirming knockdown efficiency; N, O, P flow-cytometric analysis showing reduced ROS (DCFDA, N), lipid peroxidation (C11-BODIPY, O), and cell death (Annexin V/PI, P) following Lcn2 silencing. n = 3–4 biological replicates per group. Data are presented as mean ± SD. *p < 0.05 vs. control or scrambled shRNA; $ p < 0.05 vs. LCN2 alone; one-way ANOVA with Newman-Keuls post hoc test or Student’s t test as appropriate. Data in (A–P) are representative of three independent experiments.
Since LCN2 binds iron-siderophore complexes in multiple tissues [2, 21–25], we next investigated whether recombinant LCN2 increases intracellular iron levels under basal conditions. FerroOrange staining revealed a significant rise in Fe²⁺ levels in LCN2-treated cells (Fig. 1C). Likewise, LCN2 markedly increased both total and ferric iron (Fe³⁺) content in osteocytes (Fig. 1D, E). Given that excess intracellular iron can drive oxidative stress via the Fenton reaction, we examined whether LCN2 also promotes oxidative stress and lipid peroxidation in osteocytes. Flow cytometry analysis of DCFDA (2′,7′-dichlorodihydrofluorescein diacetate) fluorescence showed a robust increase in ROS generation in LCN2-treated osteocytes compared with controls (Fig. 1F, G, Supplementary Fig. S1A). Similarly, BODIPY 581/591 C11 staining, which detects oxidized membrane lipids, revealed a strong increase in oxidized BODIPY fluorescence following LCN2 treatment, indicating elevated lipid peroxidation (Fig. 1H, I, Supplementary Fig. S1B). Importantly, all of these effects were abolished when osteocytes were co-treated with the iron chelator and ferroptosis inhibitor deferoxamine (DFO) (Fig. 1C–I), demonstrating that LCN2 promotes intracellular iron accumulation, oxidative stress, and lipid peroxidation–hallmarks of ferroptosis.
To determine whether these oxidative changes culminate in cell death, we performed Annexin V/PI staining followed by flow cytometry. Compared with the untreated group, LCN2 treatment markedly increased the proportion of Annexin V-positive osteocytes, indicating enhanced cell death (Fig. 1J, Supplementary Fig. S1C, D). This cytotoxic effect of LCN2 was significantly reduced by DFO, but not by inhibitors of necroptosis (Necrostatin-1, Nec-1), apoptosis (DEVD), or pyroptosis (VX-765). The finding that only DFO prevented LCN2-induced death supports a mechanism of iron-dependent oxidative damage consistent with ferroptosis.
To further validate LCN2’s role in ferroptosis, we transiently knocked down Lcn2 in OCY454 cells using shRNA. Compared with scrambled control shRNA, Lcn2 shRNA reduced LCN2 mRNA and protein levels by approximately 50% (Fig. 1K–M). Suppression of endogenous Lcn2 alleviated oxidative stress, as evidenced by reduced DCFDA fluorescence (Fig. 1N, Supplementary Fig. S2A), and decreased lipid peroxidation, reflected by diminished BODIPY fluorescence (Fig. 1O, Supplementary Fig. S2B). Lcn2 knockdown also lowered Annexin V/PI-positive cells (Fig. 1P, Supplementary Fig. S2C), confirming that loss of Lcn2 reduces osteocytes vulnerability to ferroptotic stress. Together, these findings establish that LCN2 promotes ferroptosis in osteocytes by driving iron accumulation, oxidative stress, and lipid peroxidation, and that reducing Lcn2 mitigates these ferroptotic hallmarks, underscoring its critical role in osteocyte survival and iron homeostasis.
LCN2 drives osteocyte ferroptosis via SLC22A17 under basal and iron overload conditions
LCN2-iron-siderophore complexes are internalized through receptor-mediated endocytosis [26], prompting us to examine whether osteocytes express the LCN2 receptor SLC22A17. We assessed Slc22a17 mRNA levels in murine BMSCs undergoing osteogenic differentiation and differentiated OCY454 osteocyte-like cells used in Fig. 1A, B. During osteogenic differentiation of BMSCs, Slc22a17 mRNA levels progressively increased, with a sharp upregulation at weeks 3 and 4, coinciding with the transition of osteoblasts into mature osteocytes (Fig. 2A). Similarly, differentiated OCY454 cells exhibited a marked increase in Slc22a17 expression (Fig. 2B).Fig. 2SLC22A17 mediates LCN2-induced iron accumulation, oxidative stress, lipid peroxidation, and ferroptosis in osteocytes.qPCR analysis of Slc22a17 mRNA levels in (A) primary osteocytes derived from BMSCs and B differentiated OCY454 osteocyte-like cells, confirms presence of this receptor in osteocytes. *p < 0.05 vs. week 0, n = 4 biological replicates/time point. C qPCR analysis of Slc22a17 mRNA in OCY454 cells transfected with scrambled control (Scr) siRNA or Slc22a17 siRNA for 72 h shows knockdown efficiency. n = 4 biological replicates/group. Quantification (D) and representative image (E, scale bar = 40 µm) of FerroOrange staining shows increased Fe²⁺ accumulation in OCY454 cells treated with rmLCN2 (100 ng/ml) for 48 h, that is attenuated by Slc22a17 knockdown. n = 4 biological replicates/group. F Flow cytometric analysis of DCFDA fluorescence shows increased oxidative stress (ROS) in rmLCN2-treated OCY454 cells, which is reduced by Slc22a17 knockdown. G Flow cytometric quantification of C11-BODIPY fluorescence indicating increased lipid peroxidation in rmLCN2-treated OCY454 cells, which is attenuated by Slc22a17 knockdown. n = 3 biological replicates/group. H Flow cytometric analysis of Annexin V/PI staining showing increased ferroptotic cell death in rmLCN2-treated OCY454 cells, which is reduced by Slc22a17 knockdown. n = 3 biological replicates/group. Under iron overload conditions (FAC, 2.5 mM), Slc22a17 knockdown reduces Fe²⁺ accumulation (I, J) and ferroptotic cell death (K) in OCY454 cells treated with LCN2, n = 3–4 biological replicates/group, scale bar in J = 40 µm. Data are presented as mean ± SD. *p < 0.05 vs. control treatment; ^$^p < 0.05 vs. LCN2-treated Scr siRNA, determined using one-way ANOVA with Newman-Keuls post hoc correction (A, B, D), Student’s t test (C), or two-way ANOVA with Newman-Keuls post-hoc correction (F–K). Data in (A–K) were derived from three independent experiments. A, B contain data obtained from the experiment reported in Fig. 1A, B.
To determine whether SLC22A17 is required for LCN2-driven ferroptosis, we transiently silenced endogenous expression of Slc22a17 using siRNA (Fig. 2C) and then treated cells with recombinant LCN2. Cells transfected with scrambled siRNA showed a robust LCN2-induced increase in intracellular Fe^2+^ levels, and that Slc22a17 knockdown significant blunted this effect of LCN2 (Fig. 2D, E). Similarly, LCN2-induced oxidative stress, lipid peroxidation, and cell death were markedly reduced upon Slc22a17 silencing (Fig. 2F–H). These findings indicate that LCN2-driven iron accumulation, oxidative stress, lipid peroxidation, and cell death are mediated by SLC22A17 in osteocytes.
Because LCN2 alone was sufficient to trigger ferroptosis under basal conditions, we next examined whether excess iron overload synergizes with LCN2 to exacerbate osteocyte ferroptosis. Using ferric ammonium citrate (FAC), an established iron overload model [27, 28], we assessed LCN2’s impact on Fe²⁺ accumulation and cell death. FerroOrange staining confirmed that FAC alone increased Fe²⁺ levels by ~2-fold compared to basal conditions (Fig. 2D, E vs. 2I, J). Addition of LCN2 further enhanced Fe²⁺ accumulation beyond FAC alone (Fig. 2I, J). Consistently, FAC alone induced ferroptotic death in osteocytes, but co-treatment with LCN2 markedly increased the susceptibility of osteocytes to FAC-induced ferroptosis, indicating a compounding effect (Fig. 2K). Notably, Slc22a17 knockdown attenuated these effects, reducing Fe²⁺ accumulation and ferroptotic death under both basal and iron overload conditions. Together, these findings establish that LCN2 exacerbates ferroptotic death under both basal and iron-overload conditions and that SLC22A17 is a key mediator of LCN2-driven Fe²⁺ uptake, oxidative stress, and ferroptosis in osteocytes.
Deletion of LCN2 in late osteoblasts and osteocytes reduces their intracellular iron accumulation, enhances antioxidant defenses, and mitigates ferroptosis susceptibility in vivo
To investigate the role of LCN2 in regulating biological processes within late osteoblasts and osteocytes vivo, we generated Dmp1-Cre; Lcn2^fl/fl^ mice. qPCR analysis confirmed a significant reduction in Lcn2 mRNA expression in bones of Dmp1-Cre; Lcn2^fl/fl^ mice compared to control littermates (Cre-negative, wild-type, WT) (Fig. 3A). Consistently, immunofluorescence staining demonstrated a marked decrease in LCN2-positive osteocytes in Dmp1-Cre; Lcn2^fl/fl^ mice (Fig. 3B, C), validating efficient gene ablation in osteocytes.Fig. 3LCN2 deletion in osteocytes reduces iron accumulation and ferroptosis susceptibility in vivo.A qPCR analysis of RNA obtained from humeri of 13-week-old Dmp1-Cre; Lcn2^fl/fl^ and WT male mice confirms reduced Lcn2 mRNA levels in Dmp1-Cre; Lcn2^fl/fl^ bones compared to WT controls. n = 8–10 mice/group. B, C Immunohistochemistry was used to quantify for LCN2 expression in osteocytes from Dmp1-Cre; Lcn2^fl/fl^ and WT tibiae. n = 3–6 mice/group, 4 ROI/mouse. Representative immunofluorescence images show LCN2 (red) and DAPI (blue)-positive osteocytes in tibial sections from Dmp1-Cre; Lcn2^fl/fl^ and WT mice. D, E Principal component analysis (PCA) (D) and volcano plot (E) of RNA-seq data from humeri of Dmp1-Cre; Lcn2^fl/fl^ and WT mice, show distinct clustering and differential gene expression (FDR < 0.05, FC ≥ 1.5). n = 4 mice/group. F, G Overlap between differentially expressed genes (1144 upregulated, 1415 downregulated) and ferroptosis-related genes from FerrDb, is shown as a Venn diagram (F), highlighting ferroptosis drivers and suppressors. Heatmap (G) shows significant changes in ferroptosis driver and suppressor genes in WT and Dmp1-Cre; Lcn2^fl/fl^ mouse bones. H Representative images of tibial sections from Dmp1-Cre; Lcn2^fl/fl^ and WT mice shows Perls’ Prussian blue staining and immunohistochemistry for ferroptosis markers SLC7A11, GPX4, and MDA in osteocytes. I−L Quantification of Perls’ Prussian blue staining (I) and percentage of SLC7A11 (J), GPX4 (K), and MDA (L) positive osteocytes indicate that in Dmp1-Cre; Lcn2^fl/fl^ tibiae, osteocytes show both reduced Fe³⁺ deposits (blue) and increased resilience against ferroptosis. White arrows indicate osteocytes positive for DAPI and the respective markers. n = 4 mice/group, 4 ROI/mouse. M qPCR analysis shows upregulated ferroptosis suppressors (Slc7a11, Gpx4, Nrf2) and downregulated ferroptosis driver (Alox5) in Dmp1-Cre; Lcn2^fl/fl^ humeri. n = 8–10 mice/group. Data are shown as mean ± SD, *p < 0.05 compared to WT, determined using Student’s t test (A, B, I–M). Scale bar for C and H = 50 µm.
To evaluate how osteocytic Lcn2 deficiency affects global gene expression in bone, we performed bulk RNA-seq on osteocyte-enriched bone from both WT and Dmp1-Cre; Lcn2^fl/fl^ mice. Principal component analysis (PCA) revealed distinct clustering between WT and Dmp1-Cre; Lcn2^fl/fl^ samples (Fig. 3D), indicating divergent transcriptional profiles. Differential expression analysis identified 2593 genes as differentially expressed (DEGs; FDR < 0.05, fold-change > 1.5) between genotypes (Fig. 3E).
Given our in vitro evidence that LCN2 promotes osteocyte ferroptosis, we next examined whether Lcn2 ablation in osteocytes alters expression of ferroptosis-related genes (FRGs) in vivo. Cross-referencing DEGs with FRGs curated from FerrDb, a database of validated ferroptosis regulators [29], revealed that 21 ferroptosis “driver” genes were downregulated in Dmp1-Cre; Lcn2^fl/fl^ bones, while 20 ferroptosis “suppressor” genes were upregulated (Fig. 3F). A heatmap of differentially expressed ferroptosis-associated genes further highlighted this coordinated shift toward a ferroptosis-resistant transcriptional program (Fig. 3G). These data suggest that LCN2 deficiency may enhance osteocyte resilience to ferroptotic stress.
To experimentally validate these transcriptional findings, we examined whether LCN2 ablation affects ferroptotic susceptibility in vivo. We first examined iron deposition in bone using Perl’s Prussian blue staining [12], which detects Fe^3+^ deposits. Dmp1-Cre; Lcn2^fl/fl^ bones displayed a notable reduction in Fe^3+^ deposits within cortical and trabecular bone osteocytes compared with WT controls (Fig. 3H, I), indicating that Lcn2 deletion limits iron loading in osteocytes–a key determinant of ferroptotic vulnerability.
We then evaluated whether LCN2 ablation affects antioxidant defenses that counteract ferroptosis, focusing on the cystine/glutamate antiporter System Xc⁻ (encoded by Slc7a11/SLC7A11/xCT) and glutathione peroxidase (GPX4), which together mitigates oxidative stress by promoting cystine import and glutathione synthesis [11, 30]. Immunohistochemistry of cortical bone revealed a significant increase in SLC7A11-positive and GPX4-positive osteocytes in Dmp1-Cre; Lcn2^fl/fl^ mice compared to WT controls (Fig. 3H, J, K), indicating enhanced cystine transport and antioxidant capacity [30]. In cultured OCY454 cells, recombinant LCN2 reduced the Slc7a11 and Gpx4 expression (Supplemental Fig. S3A, B), confirming that LCN2 dampens these antioxidant pathways. Interestingly, loss of its receptor SLC22A17 did not reverse this effect, suggesting that LCN2 modulates System Xc^-^-GPX4 pathway through mechanisms independent of SLC22A17.
LCN2-deficient osteocytes also exhibited reduced levels of malondialdehyde (MDA), a product of lipid peroxidation and a marker of ferroptotic activity (Fig. 3H, L). This decrease supports the notion that deleting Lcn2 alleviates oxidative damage in osteocytes. Consistent with this, qPCR analysis showed higher expression of Slc7a11, Gpx4, and Nrf2–key regulators of the antioxidant response–in LCN2-deficient bone. Alox5, which encodes arachidonate 5-lipoxygenase, an enzyme that promotes polyunsaturated fatty acid (PUFA) peroxidation, was suppressed in Dmp1-Cre; Lcn2^fl/fl^ bones relative to WT controls (Fig. 3M). Because PUFA oxidation is central to ferroptosis, lower Alox5 levels provide further evidence that Lcn2 deletion limits susceptibility to ferroptosis. Although Alox5 was differentially expressed in the RNA-seq dataset, Slc7a11, Gpx4, and Nrf2 did not meet significance thresholds, likely reflecting differences in assay sensitivity or the limited RNA-seq sample size. Overall, these data show that loss of LCN2 in osteocytes reduces iron accumulation, strengthens antioxidant defenses, and diminishes lipid peroxidation, thereby protecting osteocytes from ferroptotic stress.
LCN2 ablation enhances mitochondrial function in osteocytes
Mitochondria are central to cellular energy metabolism and redox balance, relying on iron for these process. Consequently, they are particularly vulnerable to disruption in iron homeostasis and oxidative stress [31–34]. Given that LCN2 functions as an iron-trafficking protein and that excess iron disrupts mitochondrial integrity, leading to membrane depolarization, dysregulated oxidative phosphorylation (OXPHOS), and ATP depletion, we investigated whether Lcn2 deletion in osteocytes affects mitochondrial function under basal conditions.
To explore the molecular basis for these potential changes, we cross-referenced 1178 upregulated DEGs from Dmp1-Cre; Lcn2^fl/fl^ versus WT bone RNA-seq with MitoCarta 3.0, a curated database of mitochondrial-localized genes [35]. This analysis revealed an enrichment of pathways associated with mitochondrial dynamics, transcription and translation, protein import/homeostasis, OXPHOS, and metabolism (Fig. 4A). Strikingly, nearly half of all genes related to OXPHOS were upregulated in bones of LCN2-deficient mice, including genes of electron transport chain (ETC) complexes I–V (Fig. 4B). Additionally, many genes associated with mitochondrial integrity, including stability, fusion, and mitophagy, were increased in bones of Dmp1-Cre; Lcn2^fl/fl^ mice (Supplemental Fig S4A).Fig. 4LCN2 deletion enhances mitochondrial function in osteocytes by preserving membrane potential and oxidative phosphorylation.A Bulk RNA-seq analysis of humeri from Dmp1-Cre; Lcn2^fl/fl^ and WT mice shows upregulation of mitochondrial-related pathways, cross-referenced with MitoCarta3.0. Venn diagram highlights differentially expressed genes across mitochondrial pathways, including central dogma, protein import, oxidative phosphorylation (OXPHOS), metabolism, small molecule transport, signaling, and mitochondrial dynamics. n = 4 mice/group. B Heatmap of differentially expressed genes in respiratory complexes I–V of the electron transport chain (ETC), shows broad upregulation of mitochondrial respiration-related genes in Dmp1-Cre; Lcn2^fl/fl^ bones compared to WT mice. C Representative JC1 staining images of OCY454 osteocyte-like cells treated with control (αMEM), rmLCN2 (100 ng/ml), or FCCP (positive control) for 24 h. Scale bar = 40 µm. D Quantification of JC1 fluorescence shows decreased mitochondrial membrane potential (ΔΨm) in rmLCN2-treated osteocytes, similar to the mitochondrial uncoupler, FCCP. n = 4 biological replicates. E LCN2 shRNA rescues mitochondrial membrane potential, as indicated by JC1 fluorescence quantification in OCY454 cells transfected with scrambled or Lcn2 shRNA. n = 3 biological replicates. F Knockdown of Slc22a17 prevents LCN2-induced mitochondrial depolarization in OCY454 cells transfected with scrambled control (Scr) or Slc22a17 siRNA and treated with rmLCN2. n = 4 biological replicates. G, H Seahorse extracellular flux analysis shows that LCN2 treatment does not significantly alter basal respiration (G) but impairs ATP production (H) which is restored upon Slc22a17 knockdown. n = 6 biological replicates/group. Data are shown as mean ± SD. *p < 0.05 vs. control treatment; ^$^p < 0.05 vs. LCN2-treated Scr siRNA, determined using either one-way ANOVA with Newman-Keuls post hoc correction (D) Student’s t test (E) or two-way ANOVA with Newman-Keuls post hoc correction (F–H). Data in D–H were reproduced in three independent experiments.
To determine whether these transcriptomic changes observed in vivo correspond to functional changes in osteocyte mitochondria, we assessed mitochondrial membrane potential (ΔΨm) using JC1 dye staining, a widely used indicator of mitochondrial integrity. Recombinant LCN2 treatment of OCY454 cells induced mitochondrial depolarization, as indicated by a shift from JC1 aggregates (red) to JC1 monomers (green) (Fig. 4C). Quantification confirmed that LCN2 reduced JC1 red/green fluorescence ratio, similar to the effect of mitochondrial uncoupler FCCP (carbonyl cyanide-p-trifluoromethoxy phenylhydrazone) (Fig. 4D). Consistent with these findings, LCN2 also impared mitochondrial OXPHOS function (Supplementary Fig. S4B). Indeed, Lcn2 knockdown via shRNA restored mitochondrial membrane potential, as evidenced by increased JC1 fluorescence (Fig. 4E). Together, these findings suggest that LCN2 regulates mitochondrial function in osteocytes even under basal conditions.
Next, we examined whether LCN2-induced mitochondrial dysfunction occurs via SLC22A17. JC1 staining showed that knockdown of Slc22a17 significantly attenuated LCN2-induced mitochondrial depolarization (Fig. 4F). Consistent with this, Seahorse extracellular flux analysis revealed that LCN2 reduced basal respiration (showing a strong trend toward reduction; p < 0.08, Fig. 4G) and significantly decreased ATP production (Fig. 4H), whereas Slc22a17 knockdown blunted the effects (Fig. 4G, H). Collectively, these data identify LCN2-SLC22A17 signaling as a novel regulator of mitochondrial function and suggest that dysregulation in this pathway may adversely affect mitochondrial activity at baseline.
LCN2 suppresses Wnt/β-Catenin signaling in osteocytes
Recent studies have linked mitochondrial biogenesis to Wnt/ β-catenin signaling. Activation of mitochondrial OXPHOS in bone marrow stromal cells has been shown to enhance their osteogenic differentiation by promoting β-catenin acetylation and activity [36]. Similarly, overexpression of mitochondrial transcription factor A, Tfam, augments Wnt-induced osteogenesis [37]. Given our observation that LCN2 deficiency improves mitochondrial integrity and function in osteocytes, we hypothesized that Lcn2 deletion may enhance Wnt/β-catenin signaling, a critical pathway for bone formation.
To test this, we analyzed transcriptomic data from Dmp1-Cre; Lcn2^fl/fl^ and WT bones for differentially expressed genes related to Wnt/β-catenin signaling. Notably, negative regulators of Wnt signaling, Sclerostin (Sost) and Dickkopf-1 (Dkk1), were significantly downregulated in bones of mice deficient in LCN2 (Fig. 5A). We validated reduced Sost and Dkk1 mRNA levels in bone samples of Dmp1-Cre; Lcn2^fl/fl^ mice relative to WT littermates (Fig. 5B, C). At protein level, immunohistochemistry showed significantly fewer SOST-positive (Fig. 5D, E) and DKK1-positive (Fig. 5D, F) osteocytes in Dmp1-Cre; Lcn2^fl/fl^ bones. Consistent with decreased Wnt antagonists, LCN2-deficient bones exhibited a significant increase in active β-catenin positive osteocytes (Fig. 5D, G), indicating elevated Wnt/β-catenin signaling in the absence of osteocytic LCN2.Fig. 5LCN2 regulates canonical Wnt/β-catenin signaling by modulating SOST and DKK1 expression in osteocytes.A Heatmap of RNA-seq data from humeri of Dmp1-Cre; Lcn2^fl/fl^ and WT mice shows differential expression of Wnt signaling-related genes. n = 4 mice/group. B, C qPCR analysis of Sost (B) and Dkk1 (C) mRNA expression in humeri of Dmp1-Cre; Lcn2^fl/fl^ and WT mice. n = 8–10 mice/group. D-G Representative immunohistochemistry images and their quantification. shows decreased SOST-positive (E, SOST+ve) and DKK1-positive (F, DKK1+ve), and increased active β-catenin-positive (G) osteocytes in tibial sections of Dmp1-Cre; Lcn2^fl/fl^ mouse bones compared to WT. n = 5 mice/group, 4 ROI/mouse, scale bar = 50 µm. H qPCR analysis of Dkk1 mRNA expression in OCY454 osteocyte-like cells treated with control (αMEM) or rmLCN2 (100 ng/ml) for 24 h. n = 3 biological replicates/group. I Representative western blot of β-catenin and DKK1 protein expression in OCY454 cells treated with rmLCN2 or Wnt3a. Quantification of western blot data shows increased DKK1 (J) and reduced active β-catenin (K) protein levels following rmLCN2 treatment. n = 3 biological replicates. Data are presented as mean ± SD. *p < 0.05 vs. WT mice, determined using Student’s t test (B, C, E–H, J, K). Data in (H–K) were reproduced in at least two independent experiments.
To determine whether LCN2 directly inhibits Wnt signaling in a cell-autonomous manner, we treated OCY454 cells with recombinant LCN2. qPCR revealed a strong trend toward increased Dkk1 mRNA following LCN2 treatment (Fig. 5H, p = 0.0592), while immunoblotting showed that LCN2 increased DKK1 protein (Fig. 5I, J) and reduced active β-catenin protein (Fig. 5I, K) levels. These data further support our in vivo findings that LCN2 is an inhibitor of Wnt/β-catenin pathway.
Overall, our data indicate that LCN2 promotes DKK1 and SOST to inhibit Wnt/β-catenin signaling, a mechanism that may impair anabolic bone formation.
LCN2 deletion specifically enhances trabecular bone mass by increasing bone formation
To directly assess the consequences of osteocytic Lcn2 deletion on bone architecture, we performed micro-computed tomography (µCT) on Dmp1-Cre; Lcn2^fl/fl^ and WT mouse bones. Loss of Lcn2 in Dmp1-Cre targeted cells significantly increased trabecular bone volume fraction (BV/TV) and trabecular number (Tb. N) (Fig. 6A–C). Dmp1-Cre; Lcn2^fl/f^ bones also showed reduced trabecular spacing (Tb. Sp) and structure model index (Tb. SMI) suggesting a shift toward a more plate-like trabeculae, which are mechanically advantageous for trabecular microarchitecture (Table 1, Fig. 6D, E). Trabecular connective density (Tb. Conn-Den) and bone mineral density (Tb.BMD) were also significantly increased in Dmp1-Cre; Lcn2^fl/fl^ mice relative to WT littermates (Table 1, Fig. 6F–H).Fig. 6LCN2 deletion enhances trabecular bone mass, osteoblast activity, and osteocyte dendritic connectivity.A–G μCT analysis of femurs from 13-week-old male Dmp1-Cre; Lcn2^fl/fl^ and WT mice shows trabecular bone volume fraction (BV/TV, A), trabecular thickness (Tb.Th, B), trabecular number (Tb.N, C), trabecular spacing (Tb.Sp, D), structure model index (Tb.SMI, E), connectivity density (Tb.Conn-Den, F), and trabecular bone mineral density (Tb.BMD, G). Representative 3D reconstruction of trabecular bone is shown (H). n = 8–10 mice/group, scale bar = 100 µm. I−L Dynamic histomorphometry of tibias from Dmp1-Cre; Lcn2^fl/fl^ and WT mice shows bone formation rate (BFR, I), mineral apposition rate (MAR, J), and mineralizing surface per bone surface (MS/BS, L). Representative calcein and alizarin red double labeling images is shown (K). n = 8–10 mice/group, scale bar = 50 µm. M–P Static histomorphometry quantifies osteoblast number per bone perimeter (N.Ob/B.Pm, M), osteoclast number per bone perimeter (N.Oc/B.Pm, O), and eroded surface per bone surface (ES/BS, P) in toluidine blue and TRAP-stained tibial sections. Representative images of osteoblasts (N, black arrows) and osteoclasts (red cell on bone surface) are shown. n = 6 mice/group, 16 ROIs/mouse, scale bar = 50 µm. Q–S Phalloidin staining of osteocytes shows dendrite length (Q) and dendrite number (R) in Dmp1-Cre; Lcn2^fl/fl^ tibiae. Representative phalloidin (green) and DAPI (blue) staining shown (S). n = 4–5 mice/group, 4 ROIs/mouse, scale bar = 20 µm. H&E staining was used to quantify viable osteocyte number per bone area (T), while Ploton silver nitrate staining was used to visualize the lacunocanalicular network (U, V), n = 4–5 mice/group, 4 ROIs/mouse, scale bar = 20 µm. Data are presented as mean ± SD. p < 0.05 vs. WT mice, determined using Student’s t test.Table 1µCT analysis of femurs harvested from 13-week-old WT and Dmp1-Cre; Lcn2^fl/fl^ male mice shows significant changes in trabecular but not cortical bone parameters.ParametersWTDmp1-Cre; Lcn2^fl/fl^p-valueTrabecular BoneTV, mm^3^2.7 ± 0.2712.818 ± 0.2090.283BV,mm^3^0.407 ± 0.1010.548 ± 0.0820.003BS, mm^2^21.004 ± 3.88126.889 ± 3.0340.001BV/TV, %15.011 ± 3.11419.413 ± 2.3150.002BS/TV, mm^−1^7.753 ± 0.9289.535 ± 0.6930.000BS/BV, mm^−1^52.707 ± 6.55349.414 ± 3.2890.152Tb.N, mm^−1^4.871 ± 0.3435.548 ± 0.3080.000Tb.Th, mm0.051 ± 0.0050.052 ± 0.0030.557Tb.Sp, mm0.205 ± 0.0160.178 ± 0.0140.001SMI2.071 ± 0.3981.648 ± 0.2350.008Conn.D, mm^−3^139.831 ± 24.002192.429 ± 17.5970.000BMD, mgHA/ccm201.115 ± 26.85234.598 ± 22.8910.008TMD, mgHA/ccm1045.875 ± 14.8971064.225 ± 18.698*0.032Cortical BoneCt.Th, mm0.192 ± 0.0110.201 ± 0.0110.112Ct. Ar, mm^2^0.957 ± 0.0471.013 ± 0.070.062Ma.Ar, mm^2^1.25 ± 0.1451.253 ± 0.1120.950Tt.Ar, mm^2^2.206 ± 0.1712.266 ± 0.1440.408Ct.Ar/Tt.Ar, %43.496 ± 2.49344.75 ± 2.470.282Ps.Pm, mm5.789 ± 0.1725.922 ± 0.2110.155Ec.Pm, mm4.594 ± 0.4564.535 ± 0.2730.720BMD, mgHA/ccm1219.126 ± 23.1171238.694 ± 22.4490.075TMD, mgHA/ccm1290.49 ± 24.5921310.578 ± 19.2810.056
Since, osteocytes are critical regulators of osteoblasts and osteoclasts, we assessed whether increased trabecular bone mass in Dmp1-Cre; Lcn2^fl/fl^ mice was a result of enhanced osteoblast or reduced osteoclast activity. Dynamic histomorphometry showed that Dmp1-Cre; Lcn2^fl/fl^ bones exhibited increased bone formation rate (BFR), mineral apposition rate (MAR), and mineralized surface per bone surface (MS/BS) (Fig. 6I–L). Static histomorphometry showed increased osteoblast number per bone perimeter (N.Ob/B.Pm) (Fig. 6M, N), while osteoclast number and activity remained unchanged (Fig. 6N–P), suggesting that LCN2 deletion enhances trabecular bone mass primarily by promoting osteoblast function.
We further examined LCN2’s impact on osteocyte number, connectivity, and lacunocanalicular network integrity. Phalloidin and Ploton silver nitrate staining showed no change in osteocyte dendrite length (Fig. 6Q), but increased dendrite number (Fig. 6R, S) in bones of Dmp1-Cre; Lcn2^fl/fl^ mice. Despite the same number of osteocytes larger lacunocanalicular network area (LCN area, Fig. 6T–V). (Fig. 6T–V) was observed in Dmp1-Cre; Lcn2^fl/fl^ bones compared to WT littermates.
Cortical bone analysis revealed a trend toward increased cortical bone area (BA, Fig. 7A, B) and total area (TA, Fig. 7C) in Dmp1-Cre; Lcn2^fl/fl^ mice, although these differences did not reach statistical significance (p = 0.0617 and p = 0.4076, respectively). Consequently, cortical area fraction (BA/TA, Fig. 7D, E) remained unchanged (p = 0.2821). Similarly, cortical thickness (Ct. Th, Fig. 7F) showed a modest but non-significant increase (p = 0.1124), and marrow area (MA, Fig. 7G), periosteal perimeter (Ps. Pm, Fig. 7H), and endocortical perimeter (Ec. Pm, Fig. 7I) were comparable between genotypes. Overall, LCN2 deletion had no significant effect on cortical bone morphology (Table 1). A similar pattern—selective enhancement of trabecular but not cortical bone—was observed in tibias from Dmp1-Cre; Lcn2^fl/fl^ mice (Supplementary Fig. S5).Fig. 7LCN2 deletion does not alter cortical bone structure or mechanical properties.A Schematic representation of the cortical bone cross-section showing analyzed parameters. B–H μCT analysis of femurs from 13-week-old male Dmp1-Cre; Lcn2^fl/fl^ and WT mice shows cortical bone area (BA, B), total area (TA, C), cortical bone mass (BA/TA, D), cortical thickness (Ct.Th, E), marrow area (MA, F), periosteal perimeter (Ps.Pm, G), and endocortical perimeter (Ec.Pm, H). Representative μCT cross-sectional images of cortical bone are shown (I). n = 8–10 mice/group, scale bar = 100 µm. J–S Three-point bending mechanical testing of femurs from 17-week-old Dmp1-Cre; Lcn2^fl/fl^ and WT male mice was performed to report structural and material properties of bone including stiffness (J), ultimate force (K), yield force (L), post-yield displacement (M), energy to yield (N), energy to ultimate force (O), energy to failure (P), ultimate stress (Q), modulus (R), and toughness to ultimate load (S), which remained comparable between genotypes. Data are presented as mean ± SD. n = 8–10 mice/group, *p < 0.05 vs. WT mice, determined using Student’s t test.
To determine whether deletion of Lcn2 in osteocytes affects cortical bone mechanical properties, we conducted three-point bending tests. Structural properties, including stiffness (Fig. 7J), ultimate force (Fig. 7K), yield force (Fig. 7L), post-yield displacement (Fig. 7M), energy to yield (Fig. 7N), and energy to ultimate force (Fig. 7O) were similar between genotypes. Material properties, including energy to failure (Fig. 7P), ultimate stress (Fig. 7Q), elastic modulus (Fig. 7R), and toughness to ultimate load (Fig. 7S), remained unchanged (Table 2). The lack of a cortical bone phenotype in Dmp1-Cre; Lcn2^fl/fl^ mice may suggest that osteocyte-intrinsic LCN2 regulates bone remodeling indirectly through osteoblasts, which are more abundant in trabecular bone. Since cortical bone undergoes slower remodeling and is more responsive to mechanical loading, future studies incorporating mechanical load could reveal a more pronounced cortical bone response to LCN2 ablation.Table 2. Flexural strength analysis of femurs harvested from 17-week-old WT and Dmp1-Cre; Lcn2^fl/fl^ male mice show no apparent changes in structural and material properties of bone.ParameterWTDmp1-Cre; Lcn2^fl/fl^p-valueStructural PropertiesEnergy to failure, mJ7.663 ± 2.8947.585 ± 1.6910.941Stiffness, N/mm75.838 ± 12.30471.621 ± 12.1590.453Ultimate Force, N14.096 ± 2.1113.6 ± 1.1660.513failure (breaking) force, N11.722 ± 2.48310.907 ± 1.7590.402Yield Force, N9.033 ± 2.4738.114 ± 1.3830.307Postyield Displacement, mm0.585 ± 0.2360.606 ± 0.1860.825Preyield displacement, mm0.133 ± 0.0260.133 ± 0.0340.954Postyield energy to failure, mJ6.945 ± 2.7356.948 ± 1.7560.998Energy to Yield, mJ0.718 ± 0.3250.637 ± 0.270.551Energy to Ultimate Force, mJ4.292 ± 1.5843.686 ± 0.7010.268Material PropertiesUltimate Stress, MPa43.213 ± 4.2646.56 ± 10.7060.391Modulus, MPa1826.336 ± 342.2242023.56 ± 596.7310.392Toughness to ultimate load, mJ/m^3^1.711 ± 0.6651.452 ± 0.3190.481Toughness by energy to failure, mJ/m^3^3.098 ± 1.2342.887 ± 0.550.896Toughness by energy to yield, mJ/m^3^0.284 ± 0.1240.255 ± 0.110.657Post-Yield Toughness, mJ/m^3^2.814 ± 1.1692.632 ± 0.5850.857
Dmp1-Cre-targeted LCN2 ablation does not alter systemic metabolism
Given that bone-derived LCN2 has been previously implicated in regulating systemic energy balance, we next asked whether the skeletal phenotype observed in Dmp1-Cre; Lcn2^fl/fl^ mice might be secondary to changes in systemic energy balance. 13-week-old male Dmp1-Cre; Lcn2^fl/fl^ mice and littermate controls were subjected to comprehensive metabolic phenotyping. Body composition analysis revealed no significant differences in fat or lean mass between genotypes (Fig. 8A). Fasting glucose and serum insulin levels were comparable, and glucose and insulin tolerance tests revealed similar glycemic index between the two genotypes (Fig. 8B, C).Fig. 8. Dmp1-Cre mediated Lcn2 deletion does not alter systemic energy metabolism or body composition.A Body composition analysis shows lean and fat mass percentages in 13-week-old male WT and Dmp1-Cre; Lcn2^fl/fl^ mice. B, C Glucose tolerance test (GTT, B) and insulin tolerance test (ITT, C) in 13-week-old mice show comparable glucose handling between genotypes. D−I Indirect calorimetry (Promethion) measurements show oxygen consumption (VO₂, D), carbon dioxide production (VCO₂, E), respiratory exchange ratio (RER, F), energy expenditure (EE, G), activity counts (H), and daily food intake (I) during light and dark cycles in WT and Dmp1-Cre;Lcn2^fl/fl^. J Serum LCN2 levels measured by ELISA confirm reduced circulating LCN2 in Dmp1-Cre; Lcn2^fl/fl^ mice. K Longitudinal body weight tracking showing similar growth curves between WT and Dmp1-Cre; Lcn2^fl/fl^ mice from 14 to 30 weeks of age. Data are presented as mean ± SD. n = 8–10 mice/group for body composition, GTT, ITT, and body weight; n = 4 mice/group for Promethion metabolic measurements (D–I); n = 6 mice/group for serum LCN2. p < 0.01 vs. WT, Student’s t test.
Indirect calorimetry demonstrated no significant differences in oxygen consumption (VO₂), carbon dioxide production (VCO₂), respiratory exchange ratio (RER), or energy expenditure (EE) between Dmp1-Cre; Lcn2^fl/fl^ mice and control littermates (Fig. 8D–G). Likewise, locomotor activity (pedmeters) and food intake were indistinguishable between groups (Fig. 8H, I). To confirm that the absence of metabolic differences was not due to incomplete gene targeting, we measured circulating LCN2 by ELISA and found that serum LCN2 concentrations were significantly reduced in Dmp1-Cre; Lcn2^fl/fl^ mice compared with controls (Fig. 8J). Longitudinal monitoring up to 30 weeks of age showed that Dmp1-Cre; Lcn2^fl/fl^ mice-maintained same body weight as their littermate controls (Fig. 8K).
Together, these data demonstrate that although Dmp1-Cre-targeted Lcn2 deletion effectively lowers systemic LCN2 levels, it does not alter whole-body energy metabolism, activity, or body weight. These findings support the notion that LCN2 primarily acts locally within bone rather than via systemic endocrine mechanisms at least during homeostasis.
Discussion
Although LCN2 has long been recognized as an endocrine regulator of appetite and energy metabolism [3, 4, 7], its skeletal role has remained poorly defined. The present work demonstrates that Dmp1-Cre–mediated deletion of Lcn2, which targets late osteoblast and osteocytes, increases trabecular bone formation without altering body weight or systemic metabolic parameters. These findings indicate that osteocyte- and late osteoblast-derived LCN2 primarily acts as a local regulator of bone formation under basal conditions. This represents a shift from the current paradigm and highlights that LCN2’s influence on bone arises from intrinsic actions within the osteocyte network, largely independent of its endocrine metabolic effects.
A key mechanistic insight from this study is that LCN2 integrates iron metabolism with redox signaling pathways in osteocytes. Iron homeostasis is essential for skeletal integrity [38], yet the mechanisms by which non-transferrin-bound iron (NTBI) enters bone cells have been poorly characterized. Our findings identify the LCN2-SLC22A17 complex as a candidate pathway for iron import, operating alongside the classical transferrin receptor (TFR1) system. Unlike TFR1, which internalizes transferrin-bound Fe³⁺ via endocytosis, LCN2 is thought to transport siderophore-bound Fe³⁺ through its receptor SLC22A17 [26, 39]. Although the precise intracellular fate of this complex remains to be fully defined, it is plausible that Fe³⁺ internalized via SLC22A17 undergoes endosomal reduction and contributes to the cytosolic labile iron pool. This potential mechanism provides a framework to explain how bone cells may accumulate iron under physiological or inflammatory conditions, linking elevated LCN2 levels to local iron overload in bone.
The primary consequence of LCN2-mediated iron loading in osteocytes is disruption of redox homeostasis, leading to oxidative stress and cell death. By facilitating iron import while simultaneously suppressing antioxidant defenses, including the System Xc^–^ and GPX4, LCN2 elevates lipid peroxidation and ROS, priming osteocytes for a pro-ferroptotic environment. Although under basal conditions, the level of ferroptosis stress induced by LCN2 appears largely sublethal for osteocytes, it is sufficient to interfere with anabolic signaling. In particular, LCN2-induced oxidative stress and lipid peroxidation attenuated Wnt/β-catenin signaling, which normally supports osteoblast differentiation and bone formation. In this setting, LCN2 acts as a local rheostat that tunes osteocyte redox state and indirectly constrains bone formation. In contrast, deletion of LCN2 restores antioxidant capacity and enhances osteoblast activity, resulting in greater bone mass. In iron overload conditions, the LCN2–SLC22A17 pathway drives overt ferroptotic cell death—an effect fully mitigated by SLC22A17 ablation. Together, these findings suggest that LCN2 helps set the threshold between adaptive and maladaptive redox signaling in bone.
LCN2’s influence also extends to mitochondrial quality control (Fig. 4). Mitochondria are both regulators and victims of iron metabolism, and their dysfunction is a recognized feature of ferroptotic priming. LCN2 deficiency has been linked to changes in mitochondrial biogenesis, membrane fusion, and cardiolipin remodeling, consistent with improved oxidative resilience and metabolic flexibility in other cell types. Functionally, we found that LCN2 compromises mitochondrial integrity in osteocytes, exacerbating oxidative stress and reducing viability, whereas osteocytes lacking LCN2 showed increased expression of mitochondrial biogenesis genes and improved mitochondrial function. RNA-seq analysis further reveals upregulation of Mtfp1, a key regulator of mitochondrial fusion and inner membrane stability, as well as genes involved in cardiolipin biosynthesis [40]. Interestingly, LCN2 ablation has been associated with elevated cardiolipin species enriched in long-chain polyunsaturated fatty acids (LC-PUFAs) and reduced levels of those containing monounsaturated fatty acids [41]. These shifts in cardiolipin composition may influence mitochondrial membrane properties and activate compensatory pathways such as peroxisomal metabolism and mTOR signaling, thereby promoting cellular resilience to metabolic stress [24, 41]. Although the metabolic programming of osteocytes remains poorly defined, emerging evidence suggests that mitochondrial integrity within osteocytes plays a critical role in regulating their function in mechanotransduction and angiogenesis [42, 43]. Given that LCN2 expression is responsive to mechanical stimuli [44], the suppression of mitochondrial integrity by LCN2 could underlie reduced mechano-adaptive responses in aging or inflammatory states, an idea that warrants further investigation.
In addition to its role in osteocyte survival and connectivity, LCN2 appears to suppress bone formation by inhibiting Wnt/β-catenin signaling. The Wnt/β-catenin pathway is a crucial regulator of osteoblast differentiation and bone mass, and its disruption plays a causal role in driving osteoporosis [45]. Elevated expression of Wnt inhibitors, particularly Dickkopf-1 (DKK1), has been implicated in postmenopausal and age-related bone loss [46, 47], and therapies targeting Wnt inhibitors, such as sclerostin (SOST) and DKK1 have been pursued for their anabolic effects on bone in many skeletal pathologies [48]. Our findings demonstrate that LCN2 upregulates both SOST and DKK1, thereby suppressing Wnt signaling and reducing osteoblast activity. Previous studies have reported a correlation between increased LCN2 and DKK1 serum levels in osteoporotic patients [47]; our study provides experimental validation of this relationship. Although we provide mechanistic evidence that LCN2 is linked to bone formation, it is unclear if LCN2’s effects occur through direct signaling in osteocytes or through osteocyte-driven paracrine modulation of osteoblasts. Future studies targeting receptors of LCN2, namely SLC22A17 and LRP2 in specific bone cell populations will be needed to dissect these interactions and determine how LCN2-Wnt crosstalk contributes to skeletal fragility.
The skeletal effects of LCN2 appear compartment-specific and that its deletion primarily increased trabecular bone, likely reflecting higher remodeling turnover and greater osteoblast responsiveness. Notably, LCN2 ablation enhanced osteocyte connectivity and lacunocanalicular network density, however, this had no impact on cortical bone structure. Because cortical bone remodeling is strongly load-dependent, LCN2’s inhibitory effects in cortical bone may become more apparent during mechanical loading, a possibility that warrants further investigation.
A key distinction between our study and previous work is the choice of Cre driver used to delete LCN2. In prior studies, Ocn-Cre-mediated Lcn2 deletion increased appetite and body weight without affecting bone [3]. In contrast, Dmp1-Cre-mediated Lcn2 deletion in our study resulted in a starkly different phenotype: enhanced bone formation without any changes in systemic energy balance (Figs. 6 and 8). These divergent outcomes likely reflect differences in the timing and lineage specificity of Cre-mediated recombination. Dmp1-Cre targets late osteoblasts and osteocytes, while Ocn-Cre acts earlier in differentiation and also targets osteocytes [49–51]. Since most bone growth is complete by approximately three months of age in mice, it is possible that Ocn-Cre-mediated Lcn2 deletion may remove LCN2 from too few osteocytes to produce a detectable skeletal phenotype. This pattern is consistent with other examples in which Cre driver timing and cell specificity leads to distinct skeletal outcomes for the same gene. For instance, β-catenin activation via Col1a1-Cre increases bone mass by reducing resorption, whereas activation with Dmp1-Cre mediated β-catenin activation enhances bone formation by stimulating osteoblast activity [52, 53]. These cases highlight the importance of confirming gene function in bone using multiple Cre drivers; convergent results across independent models strengthen evidence for a bona fide cell-autonomous role of LCN2 in bone.
Our findings should be interpreted in light of several limitations. Although this study was conducted under basal physiological conditions, it establishes a mechanistic framework by which osteocytic LCN2 regulates iron handling, redox balance, and Wnt signaling in bone. Pathological states such as aging, obesity, and chronic inflammation, where LCN2 expression is elevated [2, 4, 15], may engage these same pathways to amplify ferroptotic stress and suppress bone formation. Investigating these disease contexts will be an important next step toward understanding the clinical relevance of LCN2 signaling in bone. Furthermore, as the present work focused on young adult male mice, it remains to be determined whether LCN2 exerts sex-specific or age-dependent effects. Because both LCN2 expression and iron metabolism are hormonally regulated, these variables are likely to influence skeletal outcomes and merit further study.
In summary, this study identifies LCN2 as a bone-intrinsic regulator linking iron handling, ferroptotic signaling, mitochondrial metabolism, and Wnt pathway suppression to the control of bone remodeling. Rather than acting solely as a circulating hormone, LCN2 functions locally in osteocytes by binding to SLC22A17 receptor to modulate redox balance and osteoblast activity. These findings broaden the functional repertoire of LCN2 and establish a mechanistic basis for its contribution to age- and inflammation-associated skeletal fragility. Because circulating LCN2 increases with chronic inflammation and aging, therapeutic strategies targeting the LCN2-SLC22A17 axis may hold promise for preserving bone health and preventing iron-driven skeletal decline.
Methods
Cell culture and treatments
OCY454 osteocyte-like cells [20] (obtained from Dr. Paola Divieti Pajevic) were cultured in αMEM (Gibco; Cat# 12571-063) supplemented with 10% heat-inactivated FBS and 1% Antibiotic-Antimycotic. Undifferentiated OCY454 cells were maintained at 33 °C and 5% CO_2_ and routinely tested for mycoplasma contamination. Cells were seeded and allowed to attach overnight in full-serum media. Prior to treatment, cells were serum-starved for 1 h, and the treatment media were prepared in low-serum (0.5% FBS). Treatment with recombinant mouse Lipocalin 2 (rmLCN2, 100 ng/ml, R&D Systems; Cat# 1857-LC) was performed for 24–48 h. Lcn2 expression was knocked down using shRNA (10 nM, VectorLabs Cat# VP-RNAi-7273) via Lipofectamine 3000 (Invitrogen; Cat# L3000001). Slc22a17 expression was knocked down using siRNA (10 nM, Dharmacon; Cat# M-059399-01-0005). 16–18 h post-transfection, cells were treated with rmLCN2 (L) or control (C) for an additional 24 h. For iron overload, OCY454 cells were cultured in the presence of ferric ammonium chloride (FAC, 250 μM, Sigma; Cat# 158040) for 24 h, while rmLCN2 treatment was reduced to 6 h. To assess cell death pathways, cells were treated with rmLCN2 (100 ng/mL) in the presence of selective inhibitors: the ferroptosis inhibitor deferoxamine (DFO, 100 μM, Tocris Cat# 14595), the necroptosis inhibitor necrostatin-1 (Nec-1, 20 μM, Tocris Cat# 11658), the apoptosis inhibitor DEVD (10 μM, Tocris Cat# 14414), and the pyroptosis inhibitor VX-765 (10 μM, Tocris Cat#28825).
Seahorse extracellular flux assay
Undifferentiated OCY454 cells were seeded in XFe96/XF Pro microplates (Agilent Seahorse Bioscience; Cat# 101085-004) at a density of 10,000 cells/well and cultured overnight [54]. Cells were serum-starved for 1 h in 0.5% FBS-αMEM media and then treated with rmLCN2 for 24 h. Mitochondrial respiration and glycolysis were assessed using a Seahorse Extracellular Flux Analyzer (Agilent). Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were recorded during a Mito Stress Test following sequential injection of oligomycin (Millipore; Cat# 495455-10MG, 2 μM), FCCP (Carbonyl cyanide-p-trifluoromethoxyphenylhydrazone, Sigma; Cat# C2920-10MG, 2 μM), rotenone (Fluka; Cat# 45656-250MG, 1 μM), and antimycin A (Sigma; Cat# A8674-25MG, 1 μM). After the assay, cells were lysed in RIPA buffer, and total protein/ well was quantified. OCR and ECAR values were normalized to protein concentration in each well using Wave Desktop software (v 2.6, Agilent). Data were obtained from three independent biological experiments, each with at least 6 technical replicates. Representative data from one of the experiments is shown.
Cellular ROS analysis
Intracellular ROS levels were measured using the DCFDA/H₂DCFDA (2’,7’ –dichlorofluorescein diacetate) Cellular ROS Assay Kit (Abcam; Cat# ab113851) following the manufacturer’s instructions [54]. Cells were incubated with 20 μM DCFDA in 1X assay buffer for 30 min at 37 °C, washed in 1X buffer, and analyzed by flow cytometry on the LSR FORTESSA instrument (BD Biosciences, San Jose, CA; Cat# 647586) using FACS Diva software (BD; Cat# 661618). Flow cytometry data were further analyzed, and plots and histograms were generated, using FlowJo software v.10 (BD; Cat# 446593). Each experiment included three technical replicates, and reproducible data from three independent experiments is reported.
Mitochondrial membrane potential analysis
Mitochondrial membrane potential (ΔΨm) was evaluated using the potentiometric dye JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethyl benzimidazole-carbocyanine iodide; Carlo Erba, Cat# 235425000) [54]. OCY454 cells were seeded on collagen-coated Ibidi chambered slides (Ibidi, Cat# 80826) at 40,000 cells/ well density, allowed to attach overnight, and then treated for 24 h with 100 ng/ml of rmLCN2. Cells were incubated in the dark for 30 min at 37 °C with fresh media containing JC1 (10 μg/ml). After incubation, JC1 containing media was removed, and cells were washed and incubated in reduced serum media (αMEM supplemented with 0.5% FBS) prior to imaging under confocal microscope. Zeiss Airyscan confocal microscope with a 63x/1.4 oil objective was used to conduct live-cell imaging and acquire confocal z-stacks of cells in chamber slides (5% CO_2_, 37 °C). Image analysis was performed using ImageJ (NIH; RRID: SCR_003070) with sum projections of red and green channels. The red-to-green fluorescence intensity ratio (R/G) was calculated by dividing red intensity by green intensity. These values were normalized to the control condition, and the fold change in mean R/G fluorescence was reported. For each treatment, 3–4 images were collected per experiment, and representative data from one experiment is shown. The results were reproducible across three independent experiments performed by two different investigators.
Lipid peroxidation analysis
Intracellular lipid peroxidation levels were measured using the C11-BODIPY 581/591 kit (Invitrogen; Cat# D3861) following the manufacturer’s instructions [55]. Cells were seeded on 24-well plates (Corning; Cat# 3526) at 40,000 cells/well density, allowed to attach overnight, and treated for 24 h with 100 ng/ml of rmLCN2. After treatment, cells were washed with PBS, subjected to trypsin treatment, and a single cell suspension was further stained in α-MEM containing 5 ng/ml C11-BODIPY for 30 min at 37 °C. Cells were analyzed by flow cytometry using LSR FORTESSA, with the reduced and oxidized forms of C11-BODIPY measured at 581–591 nm and 488–510 nm, respectively. The ratio of oxidized to reduced C11-BODIPY stained cells was evaluated using FlowJo software v.10. At least three technical replicates were used, and reproducible data from three independent experiments are reported.
FerroOrange staining and iron assay
The labile iron pool was assessed using FerroOrange (Dojindo; Cat# F374) [56]. OCY454 cells were seeded at 10,000 cells/well in 8-well Ibidi chambered slides (Ibidi; Cat# 80826) and allowed to adhere overnight. Cells were treated with rmLCN2 (100 ng/ml) for 24 h, washed with PBS, and stained in α-MEM containing 1 μM FerroOrange for 30–45 min at 37 °C in the dark. After staining, cells were washed with Hank’s Balanced Salt Solution (HBSS, Gibco; Cat# 14025092), and counterstained with DAPI (1 μg/ml; Thermo Fisher; Cat# D1306). Cells were immediately covered with base medium and transferred to a live-cell chamber of Zeiss Airyscan confocal microscope for imaging under 20x and 40x/1.4 oil objectives. Fluorescence intensities from 3 to 4 fields/well were analyzed using ImageJ (NIH; RRID: SCR_003070) and normalized to the control condition. Representative data from one experiment is shown, and results were reproducible across three independent biological experiments conducted by two independent investigators. Total iron, Fe2+, and Fe3+ in cell lysates were measured using the Abcam Iron assay kit (#ab83366) after rmLCN2 (100ng/ml, 24h) treatment of OCY454 cells. Three technical and three biological replicates were analyzed per group.
Cell Death Assessment
Cell death was assessed using the Annexin V-FITC/Propidium Iodide (PI) Flow Cytometry Kit (Invitrogen; Cat# V13242) according to the manufacturer’s protocol. OCY454 cells were seeded in 24-well plates (Corning; Cat# 3526) at 40,000 cells/well density, allowed to adhere overnight, and treated with rmLCN2 (100ng/ml) for 24 h. Cells were then trypsinized, centrifuged at 3000 rpm for 3 min, and reconstituted in 1X Annexin binding buffer (Invitrogen; Cat# V13242). Annexin V-FITC and PI were added, and cells were incubated for 15 min at room temperature before analysis by flow cytometry (LSR FORTESSA, BD Biosciences; Cat# 647586). The percentage of early and late apoptotic cells was quantified using FlowJo software v.10 (BD; Cat# 446593). Each experiment included 3–4 technical replicates, and data from three independent experiments are reported.
Western blots
Whole-cell lysates of OCY454 cells were collected in RIPA lysis buffer containing 50 mM Tris (pH 7.4), 1% NP-40, 0.25% sodium deoxycholate, 150 mM NaCl, and 1 mM EDTA, supplemented with phosphatase inhibitor (Pierce; Cat# A32957), protease inhibitor (cOmplete Mini, Roche; Cat# 11836153001), and 1 mM PMSF (Sigma; Cat# P7626) [54]. Lysates were sonicated on ice using a probe sonicator (five 15-s pulses, 45 s rest between pulses) and centrifuged at 12,000 rpm for 10 min at 4 °C. Protein concentration was determined using the Pierce BCA Protein Assay Kit (Thermo Scientific; Cat# 23225).
For immunoblotting 25 µg of total protein per sample was loaded onto 10% SDS-polyacrylamide gels, separated by electrophoresis, transferred to a nitrocellulose membrane (LI-COR; Cat# 926-31092), and blocked with Intercept® (PBS) Blocking Buffer (LI-COR; Cat# 927-70001). Membranes were incubated with the following primary antibodies prepared in Intercept blocking buffer with 0.1% Tween-20. Following primary antibodies were used in this study: anti-β-actin (1:2500, Abcam; Cat# ab8226), anti-β-tubulin (1:1000, Novus Biologicals; Cat# NB100-1612), anti-Lipocalin 2 (1:1000, Abcam; Cat# ab216462), anti-DKK1 (1:2000, ProteinTech; Cat# 21112-1-AP), and anti-active-β-catenin (1:1000, Abcam; Cat# ab305261). Blots were then incubated with IRDye 680- or 800- conjugated secondary anitbodies anti-mouse and anti-rabbit antibodies conjugated to fluorophores (anti-mouse,1:15,000, Cat# 926-68070; anti-rabbit, 1:10,000, Cat# 926-32211; LI-COR Biosciences). Signal was detected using the Odyssey infrared imaging system (LI-COR Biosciences; Cat# 9140-00), and band intensities were quantified using Image Studio Lite v5.2 (LI-COR Biosciences).
Membranes were stripped and re-probed for different antibodies, and the efficiency of the stripping was verified by the absence of residual secondary antibody signal. Protein expression was normalized to β-actin or β-tubulin, and fold changes were calculated relative to control groups. Data are presented as mean ± SD, with representative blots shown from experiments using three biological replicates per group, pooled from two independent experiments.
Mouse husbandry
Lcn2 floxed mice (Lcn2^fl/fl^; JAX Strain #031034) were bred with 9.6 kb Dmp1-Cre mice (JAX Strain #023047). The resulting offspring included Dmp1-Cre + /-; Lcn2^fl/fl^ mice (denoted as Dmp1-Cre; Lcn2^fl/fl^) and Dmp1-Cre-/-; Lcn2^fl/fl^ littermate controls (denoted as Wild-type, WT Control), which were used for subsequent experiments. 50 ng of genomic DNA extracted from tail biopsies was used for genotyping. The presence of the Dmp1-Cre transgene was confirmed using the following primers: Forward: 5’-TTG CCT TTC TCT CCA CAG GT-3’, Reverse: 5’-CAT GTC CAT CAG GTT CTT GC-3’. The presence of the Lcn2 floxed allele was confirmed with the following primers: Forward: 5’-CCT CAA GGA CGA CAA CAT CA-3’ and Reverse: 5’-GAG GAA GCT TGG ACA GGA ATC-3’. Thirteen-week-old WT and Dmp1-Cre; Lcn2^fl/fl^ male mice were used for the qPCR, bulk RNA-seq, immunohistochemistry, bone histomorphometry, and micro-computed tomography. Seventeen-week-old WT and Dmp1-Cre; Lcn2^fl/fl^ male mice were used for mechanical testing. Mice were allocated to experimental groups based on genotype (WT vs. Dmp1-Cre; Lcn2^fl/fl^); therefore, no additional randomization procedures were applied.
Quantitative real time PCR (qPCR)
RNA was extracted from mouse humeri using the miRNeasy Mini Kit (Qiagen, Valencia, CA) following the manufacturer’s protocol [54, 57]. The marrow and periosteum were carefully removed prior to flash-freezing the bones in liquid nitrogen. Bones were then homogenized in QIAzol, and RNA was extracted. RNA from cell lysates was extracted using the Direct-Zol RNA Miniprep Kit (Zymo Research). cDNA was synthesized using iScript cDNA Synthesis Kit (Bio-Rad) followed by qPCR with iQ SYBR Green Supermix (Bio-Rad) or TaqMan probes (Applied Biosystems, Thermo Fisher Scientific), according to the manufacturer’s instructions. β-Actin or 18S ribosomal RNA were used as internal controls for normalization. Relative mRNA expression was calculated using the 2^−ΔΔCT^ method and is presented as fold change in mRNA levels. Primer sequences and IDs are listed in Supplemental Table 1.
Bulk RNA-seq analysis
Total RNA was isolated from humeri using the miRNeasy Minikit (Qiagen, Valencia, CA). RNA purity was assessed spectrophotometrically (A_260_/A_280_ ratio >1.9) and RNA integrity was evaluated using Bioanalyzer (Agilent). RNA samples with RNA integrity number (RIN) > 7 were used for library preparation and sequencing. Total RNA was extracted from two pooled humeri per mouse, and equal amounts of RNA were used to generate biologically distinct replicates for each group (n = 4 mice/group).
Stranded mRNA-seq libraries were prepared using the NEBNext® Ultra™ II RNA library Prep Kit for Illumina® (New England Biolabs, NEB #E7770). First- and second-strand cDNA synthesis, end-filling with Klenow fragment, and dA-tailing were performed per manufacturer’s instructions. Illumina paired-end adapters for multiplexed sequencing were ligated in a 50 μl reaction volume for 30 min at room temperature using 1 μl of T4 DNA ligase and 0.3 μM of annealed adapters. Ligated products were size-selected using a high-resolution 2% agarose gel, and fragments around 200 bp ( ± 50 bp) were excised and purified with the Qiagen Gel Extraction Kit. Size-selected cDNA libraries were amplified using indexed primers for 12–14 cycles in a 50 μl PCR containing 29 μl of template, 1 μl of forward and reverse primers (25 μM each), and 1 U of Phusion high-fidelity DNA polymerase (New England Biolabs). PCR products were purified using the QIAquick PCR purification kit (Qiagen) and eluted in 30 μl of nuclease-free water. A small aliquot (~1 μl) was analyzed with a DNA1K chip (Experion, Bio-Rad) to confirm the absence of primer-dimers or spurious products. RNA-seq libraries were quantified using the Qubit dsDNA HS Assay kit.
Libraries were sequenced on the Illumina NextSeq 2000 platform using single-end 100 bp reads. Raw FASTQ files were obtained directly from the Illumina BaseSpace environment. Initial quality control (QC) was performed using the DRAGEN FASTQC app within BaseSpace, assessing metrics such as per-base sequence quality, per sequence GC content, adapter content, and sequence duplication levels. Only samples meeting QC thresholds were retained for further analysis. FASTQC outputs were aggregated with MultiQC to generate a consolidated QC report and identify outlier samples or systematic issues.
High-quality reads were processed using the DRAGEN RNA Pipeline in BaseSpace. This pipeline performs alignment and quantification using the DRAGEN Bio-IT Platform. Reads were aligned to the mouse reference genome (GRCm38_v100) using the DRAGEN aligner, and gene- and transcript-level quantification was performed using a GTF annotation. Expression levels were reported as raw read counts, TPM (transcripts per million), and FPKM (fragments per kilobase of transcript per million mapped reads), as applicable.
Aligned BAM files from the DRAGEN RNA pipeline were used for downstream differential expression analysis and data visualization. BAM files were analyzed in SeqMonk v1.47.2 for transcript quantification. Genes with fewer than 25 counts were excluded prior to differential expression analysis. Differentially expression genes (DEGs) were identified using the DESeq2, with significance defined as false discovery rate (FDR) ≤ 0.05 and absolute log_2_ fold change ≥1.5. Ferroptosis-related genes were obtained from the FerrDb database [29]. Additionally, DEGs were mapped to mitochondrial pathways using the MitoCarta database, to annotate genes involved in mitochondrial structure and function (e.g., genes related to mitochondrial central dogma, protein import and sorting, protein homeostasis, oxidative phosphorylation (OXPHOS), metabolism, small molecule transport, signaling, and mitochondrial dynamics and surveillance).
Immunohistochemistry and staining
Paraffin-embedded bone sections (5 µm thick) were deparaffinized, rehydrated, subjected to antigen retrieval, blocked, and permeabilized with 0.5% Triton X-100 (Innovex) [54]. Sections were then incubated overnight at 4 °C with primary antibodies, including anti-xCT (1:250, Abcam, ab307601), anti-GPX4 (1:50, Abcam, ab125066), anti-LCN2 (1:250, Abcam, ab216462), anti-MDA (1:200, Abcam, ab243066), anti-SOST (1:100, R&D Systems, BAF1589), anti-DKK1 (1:100, ProteinTech, 21112-1-AP), and anti-active-β-catenin (1:500, Abcam, ab305261). The next day, sections were incubated for 1 h at room temperature with fluorescent secondary antibodies, including Donkey Anti-Goat IgG (1:1000, Jackson ImmunoResearch, 705-606-147, Alexa Fluor® 647) and Goat Anti-Rabbit IgG H&L (1:1000, Abcam, ab150080, Alexa Fluor® 594). Corresponding nonimmune IgGs (Abcam, ab172730 for rabbit and R&D systems, AB-108-C for goat) were used as negative controls. H&E staining was performed to assess viable osteocytes.
Bone sections were stained for Perl’s Prussian Blue (Abcam, ab150674) to detect Fe³⁺ deposits, following the manufacturer’s protocol [12]. Sections were incubated with 1:2 HCl- Potassium-Ferrocyanide (Sigma, P-3289) for 30 min, rinsed in distilled water, counterstained with nuclear fast red (Sigma, 60700) for 1 min, and examined by light microscopy. Staining interpretation: nuclei appears red, cytoplasm appears pink, and iron deposits appear bright blue.
Lacunocanalicular networks were visualized using Ploton silver staining [54, 58], with images acquired on a ZEISS Axio Imager M2 or Olympus BX53 and quantified using ImageJ. Osteocyte dendrites were observed using 10 µm thick frozen sections prepared from femurs embedded in O.C.T. Compound (Fisher, 23-730-571). Phalloidin (Alexa Fluor 594, 5 μg/ml; Invitrogen, R37110) and DAPI (10 μg/ml; Invitrogen, P36935) were used to stain F-actin and osteocyte dendritic networks [59]. Confocal images were acquired using a Zeiss LSM880 microscope with a 100× oil objective. The Simple Neurite Tracer plugin for ImageJ was used to analyze osteocyte dendrite number and length [60]. Z-stack images were converted to greyscale, and dendrites were semi-automatically traced to obtain average length and number per osteocyte. Four mice per group were analyzed, with three images per mouse, assessing 2–3 osteocytes per image within the full field of view.
Bone histomorphometric analysis
Dynamic histomorphometry was performed to assess mineralizing surface to bone surface (MS/BS), mineral apposition rate (MAR), and bone formation rate normalized to bone surface (BFR/BS) [61]. Mice received intraperitoneal injections of calcein green (G, 30 mg/kg) and alizarin red (A, 50 mg/kg) in a G-A sequence at 10 and 3 days before euthanasia, respectively. Left tibias were fixed in 10% buffered formalin and embedded in methyl methacrylate. Thick cross-sections at the mid-diaphysis were prepared using a diamond-embedded wire saw (Histosaw, Delaware Diamond Knives, Wilmington, DE, USA) and ground to a final thickness of 30–40 µm for periosteal and endosteal bone formation measurements.
For static histomorphometry, longitudinal sections of distal femurs were stained for TRAPase and counterstained with toluidine blue. TRAPase positive multinucleated osteoclasts (≥3 nuclei) attached to the cancellous bone region (350–1750 μm from the distal growth plate) were quantified. Histomorphometric analysis was conducted using the OsteoMeasure High-Resolution Digital Video System (OsteoMetrics, Decatur, GA) interfaced with an Olympus BX51 fluorescence microscope (Olympus America Inc., Center Valley, PA). All analyses followed ASBMR Histomorphometry Nomenclature Committee guidelines and histomorphometric analysis were performed in a blinded manner.
Micro-computed tomography (µCT)
Femurs from 13-week-old male WT and Dmp1-Cre; Lcn2^fl/fl^ mice were dissected, cleaned of soft tissue, and fixed in 10% neutral buffered formalin for 48 h before being stored in 70% ethanol for scanning [61]. µCT analysis was performed using the VivaCT-80 (Scanco Medical, Switzerland) with the following settings: E = 70 kVp, I = 114 µA, integration time = 200 ms, and an isotropic voxel size of 10.4 µm. For trabecular analysis, a Gaussian filter (sigma = 0.8, support = 1) was applied, and a lower threshold of 450 mg/cm³ was used. Trabecular bone in the distal femur was analyzed beginning 10 slices away from the growth plate to exclude the primary spongiosa, extending for a total of 150 slices. Cortical bone was analyzed using a lower threshold of 580 mg/cm³. Mid-diaphyseal cortical bone parameters were assessed over a 40-slice region centered at the femoral midpoint. All nomenclature, symbols, and units adhered to established µCT guidelines, and n = 8–10 mice per group were analyzed.
Mechanical testing
Structural and material properties of femurs from 17-week-old WT and Dmp1-Cre; Lcn2^fl/fl^ male mice were assessed using three-point bending [54, 57, 61]. Femurs were stored at −20 °C in HBSS-soaked gauze and thawed to room temperature before testing. Testing was conducted on an Electroforce TA 5500 system (TA Instruments, New Castle, DE) [62] with a loading rate of 3 mm/min and an lower support span of 8 mm that recorded the application force (N) and displacement (mm) in real-time. Femoral length was measured with a digital caliper prior to force application. The polar moment of inertia and minor diameter (anterior to posterior diameter) at the femoral midshaft were determined by µCT (VivaCT-80, Scanco Medical) with a 10 µm voxel resolution. Mechanical properties were normalized for bone size and used to calculate material properties using standard equations as previously published [63, 64].
Glucose and insulin tolerance test
Glucose tolerance test (GTT) and Insulin Tolerance test (ITT) were performed in 13-week-old WT and Dmp1-Cre; Lcn2^fl/fl^ male mice maintained on standard chow diet. For GTT, mice were fasted overnight; for ITT mice were fasted for 4 h. Glucose (1.5 mg/kg of body weight) or insulin (0.75 U/kg of body weight, Humulin, Lilly USA) was administered intraperitoneally, and blood glucose levels were measured at 0, 15, 30, 45, 60, and 120 min after injection using an Accu-Chek Aviva meter (Roche Diabetes Care, Inc.).
Indirect calorimetry
Indirect calorimetry was performed using an eight-channel Promethion system (Sable Systems International). Individual cages (interior dimensions of 31.5 × 15.5 (floor), up to 34.5 × 19.0 (ceiling) × 13.0 cm tall) included a ceiling-mounted, “hut” into which mice can climb, for automated body mass measurements, a food hopper (3 mg resolution), and water spigot. X- and Y-axis (horizontal plane) photoelectric beam motion detectors are positioned around the cage. A running wheel accessory was available but excluded from the current study. Airflow through the chamber was set to 2000 ml/min (negative pressure), and gas analyses were recorded once per minute. Cages are housed inside of thermally controlled cabinets that are maintained at ~22-25 °C, to match room temperature.
Although data were recorded for four days (Monday to Friday), only three day data (Tuesday to Friday) were used for statistical analyses to allow one day of animal acclimation. Metabolic rate was estimated from rate of oxygen consumption (VO2) and carbon dioxide production (VCO2), and respiratory exchange ratio (RER). Food intake, physical activity, and energy expenditure parameters were also measured.
ELISA for LCN2
Serum LCN2 levels were measured using the DuoSet ELISA for Mouse Lipocalin 2 (DY1857-05, R&D Systems, Bio-Techne), according to the manufacturer’s protocol.
Study approval
Mice were housed in groups in a pathogen-free facility at 22 °C with a 12-h light/dark cycle and fed their irradiated diets and water ad libitum. All studies were approved by the Institutional Animal Care and Use Committee of the University of Arkansas for Medical Sciences and conducted in accordance with institutional guidelines.
Statistical analysis
Sample sizes were determined based on a power calculation that provides an 80% chance of detecting a significant difference (p < 0.05). Technical replicates and biological replicates (n) used for all experiments are described in the figure legends. Data are reported as mean ± S.D. Statistical analyses were performed using Prism 10.0 (GraphPad Software, Inc.). For comparisons between two groups, two-tailed Student’s t-tests were used. For comparisons involving multiple groups, one-way ANOVA followed by the Newman–Keuls post hoc test was used. To analyze the effects of two variables and their interaction, two-way ANOVA followed by the Newman–Keuls post hoc test was applied. Experiments using primary cell cultures were performed with at least three replicates and independently repeated twice. Experiments using OCY454 cells included 3–4 technical replicates per condition and were repeated at least three times by two investigators. Investigators were blinded to group allocation with respect to mouse genotype during data collection, and genotype information was revealed only during data analysis.
Supplementary information
Supplemental data_ clean copy Supplemental Figure 1 Supplemental Figure 2 Supplemental Figure 3 Supplemental Figure 4 Supplemental Figure 5 Dataset 3- qPCR raw values Original Data- Raw western blot images
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Owen HC, Roberts SJ, Ahmed SF, Farquharson C. Dexamethasone-induced expression of the glucocorticoid response gene lipocalin 2 in chondrocytes. Am J Physiol Endocrinol Metabol 2008; 294. 10.1152/ajpendo.00586.2007.10.1152/ajpendo.00586.200718381927 · doi ↗ · pubmed ↗
- 2Rucci N, Capulli M, Piperni SG, Cappariello A, Lau P, Frings-Meuthen P et al. Lipocalin 2: A new mechanoresponding gene regulating bone homeostasis. J Bone Miner Res. 2015;30. 10.1002/jbmr.2341.10.1002/jbmr.234125112732 · doi ↗ · pubmed ↗
- 3Tu X, Delgado-Calle J, Condon KW, Maycas M, Zhang H, Carlesso N et al. Osteocytes mediate the anabolic actions of canonical Wnt/β-catenin signaling in bone. Proc Natl Acad Sci USA. 2015;112. 10.1073/pnas.1409857112.10.1073/pnas.1409857112 PMC 432127125605937 · doi ↗ · pubmed ↗
- 4Dole NS, Betancourt-Torres A, Kaya S, Obata Y, Schurman CA, Yoon J et al. High-fat and high-carbohydrate diets increase bone fragility through TGF-β-dependent control of osteocyte function. JCI Insight. 2024;9. 10.1172/jci.insight.175103.10.1172/jci.insight.175103 PMC 1134360839171528 · doi ↗ · pubmed ↗