Estrogen Loss Activates Memory T-Cells to Compromise Bone Integrity Through Distinct Cortical Compartments in Mice

Abstract

Fragility fractures are a significant cause of morbidity and mortality in postmenopausal women. Menopause leads to a significant decline in bone mass and quality, with over half of women sustaining fragility fractures without reaching the osteoporotic threshold (T-score < −2.5), underscoring the pivotal role of bone quality in fracture risk. Previous studies have shown that estrogen (E₂) deficiency following ovariectomy (OVX) in mice activates memory T-cells (T_M) to produce TNFα and IL-17A, resulting in trabecular bone loss. This study extends these findings to cortical bone, revealing that under habitual load osteoclasts are predominantly localized on the posterior endosteal surface. Post-OVX, mice exhibited enlarged lacunae indicative of osteocytic osteolysis and reduced dendrite density in osteocytes (Ocy) adjacent to T-cells, with these effects being more pronounced on the posterior side where osteoclast–T-cell interactions are heightened. Additionally, osteoblast (OB) function analysis revealed that while bone formation at the mid-diaphysis remained unchanged, the collagen matrix became more disorganized, particularly in the posterior cortical compartment. Importantly, OVX increased bone fragility without altering cortical thickness or mineral density. These detrimental changes were absent in OVX mice lacking TNFα and IL-17A expression in TM cells (IL15RA^ΔT), suggesting that these cytokines specifically impair the osteolineage (Ocy and OB), compromising bone quality in ways undetectable by μCT. Our findings reveal a novel mechanism where T-cell–mediated inflammation reduced cortical bone quality by targeting the osteolineage, leading to disrupted matrix organization and Ocy dendrite density. Clinically, these results highlight the potential of targeting T-cell responses to maintain bone quality and strength in estrogen-deficient states. Additionally, estrogen loss adversely affects endosteal bone quality in distinct cortical compartments without impacting bone mass, a deficit that may remain undetected by DXA scans.

Lay Summary

Menopause decreases estrogen levels, affecting bone strength and mass. Our study in mice shows that estrogen loss activates long-lived T-cells, which produce inflammatory signals. These signals target bone cells: osteoclasts that break down bone, osteocytes that help sense mechanical stress and load on bone, and osteoblasts that build bone. Forces from normal body weight are not distributed evenly over bone. Notably, the effect was more pronounced on the side of the bone under greater strain, under regular weight. As a result, bones become more fragile even though their density remains unchanged, making fractures harder to detect with standard scans. Blocking these inflammatory signals prevented bone weakening, suggesting new ways to protect bone health during menopause.

Introduction

Osteoporosis is the most common progressive bone disease characterized by loss of bone mass and compromised microarchitecture that disproportionally affect postmenopausal women, posing a high public health and economic burden.⁽¹⁾ Postmenopausal osteoporosis significantly increases the risk of fragility fractures that occur with minimal trauma such as a fall from standing height. Osteoporosis is routinely screened and diagnosed by dual-energy X-ray absorptiometry (DXA scan)⁽²⁾. However, accumulating data shows that only 44% of patients who experienced a fragility fracture had a T-score of less than or equal to −2.5⁽³⁾. 35% of fragility fracture patients had osteopenia, and 21% had normal bone mineral density (BMD). This observation indicates that, while correlating highly with fracture risk, bone mass alone is insufficient in predicting fractures. Increasingly, bone quality is recognized as an independent contributor to bone strength. ⁽⁴⁾ Adequate bone remodeling is essential to maintaining good bone quality, a function carried out by the osteolineage. Osteoblasts (OB) produce a collagen rich extra-cellular matrix (ECM) that provides structural flexibility and serves as scaffolding for calcium hydroxyapatite deposition. The nature of the ECM and the mineral content determine the biomechanical properties of bone⁽⁵⁾. During bone formation some OBs become embedded in the bone and differentiate into osteocytes (Ocy). Ocy are the most abundant cells in the bone, acting as mechanostats which sense both mechanical load and microdamage⁽⁶⁾. Ocy regulate both osteoclasts (OC) and OB to direct bone remodeling in accord with load (e.g., Wolff’s law) and to repair damaged bone⁽⁷⁾.

Multiple factors determine the risk of fragility fractures in postmenopausal women. In addition to age, medications, diet, physical activity, exposure to sunlight, smoking and alcohol consumption are all known to increase fracture risk (reviewed in⁽⁸⁾). Comorbidities such as rheumatoid arthritis, chronic kidney disease, diabetes and other metabolic diseases as well as some viral infections also contribute to development of osteoporosis and to risk of fracture. ⁽⁸⁾ Genetics plays a key role in attainment of peak bone mass, which establishes the baseline in an individual from which bone loss occurs and thus the risk of osteoporosis and fracture. All the above factors can be partitioned into bone intrinsic (e.g., attaining peak bone mass), overall health (e.g., vestibular issues that can affect stability) and immune mediated factors (i.e., infections and smoking). In the context of immune factors, T-cells have been shown to play a critical role in the pathophysiology postmenopausal osteoporosis⁽⁹⁾. Prior studies of OVX-induced osteoporosis, including our own, have primarily focused on assessing trabecular bone near the metaphysis. However, fragility fractures of the hip and long bones tend to initiate from the cortical bone. ⁽¹⁰⁾ While increased cortical porosity and decreased cortical thickness are observed in osteoporotic patients, their contributions to increased fracture risk are not well understood. ⁽²⁾

Estrogen (E₂) has been shown have a pleiotropic effect on bone cells, including altering the ratio of RANKL to OPG, and influencing osteoblast lifespans⁽¹¹⁾. Our previous studies showed that E₂-loss post-ovariectomy (OVX) leads to the production of TNFα and IL-17A by memory T-cells (T_M) to promote loss of bone mass. ⁽¹²⁾ The T_M were activated by the combined effect of IL-7 and IL-15, which we identified were produced by dendritic cells. To distinguish the effect of TNFα and IL-17A expressed by T-cells from the direct effect of estrogen loss on bone cells, we generated a T-cell—specific knockout that does not produce TNFα and IL-17A post-OVX. We achieved this by crossing a T-cell—specific-CRE mouse (LCK-CRE) to the IL-15RA-floxed mouse. When we ovariectomized the 12-week-old females from the resulting crosses, only the heterozygous (i.e., flox/+ and CRE⁺ = IL15RA⁺) females lost trabecular bone, while no bone loss was observed in the homozygous (flox/flox CRE⁺) littermates. Furthermore, no increase in TNFα or IL-17A was observed in the homozygous OVX mice. The results validate that IL-15 is needed to induce TNFα and IL-17A and that these cytokines are needed for the decrease in bone mass. However, the effect of T-cell mediated inflammation post OVX on bone quality is unknown.

Previous studies using rodent OVX models have primarily focused on trabecular bone as a surrogate for bone fragility. Since trabecular bone is highly responsive to estrogen loss, significant resorption occurs within 2 to 4 weeks post-OVX. In contrast, cortical bone loss progresses more slowly and is typically assessed at 15 weeks or later. To gain a better understanding of cortical bone fragility and the relationship between porosity, collagen fibrils, to its biomechanical properties, we ovariectomized 14-week-old mice and examined cortical bone at an earlier time point—5 weeks post-OVX. With an aging population, it is of great importance to understand the parameters other than bone mass by that determine fracture risk from low-impact trauma. To gain insights into how E₂ loss and inflammation target cortical bone cells, we focused on assessing biomechanical and microarchitecture property changes in cortical bone post-OVX in mice. In addition, we determined whether these changes were due directly to estrogen-loss or to TNFα and IL-17A expressed by T_M.

Materials and Methods

Mice

All mice used in this study are bred and maintained in-house and are on a C57BL/6J background. Wild-type C57BL/6J mice (model 000664) were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). The IL15RA^ΔT mice were previously generated⁽¹²⁾ by crossing IL15RA-floxed mice (model 022365)⁽¹³⁾ and Lck-Cre Tg540-I mice (model 006889)⁽¹⁴⁾.

OVX

14-week-old female mice were randomly assigned to sham surgery or OVX groups. Each cage had at least one sham and one OVX mouse to account for the microbiome between groups. Mice were anesthetized and maintained on 2.5 % isoflurane. Both ovaries were accessed through a single incision from the back and exteriorized through the peritoneal membrane on each side. Each ovary was clamped using a hemostat and removed by a single cut. Surgical staples (cat# DS-15; 3M, Maplewood, MN, USA) were used to close the skin incision. To minimize discomfort during and after surgery, a single dose of 1 mg/kg buprenorphine extended release (ER) was administered subcutaneously prior to surgery. No adverse responses or effects were observed in any of the mice used in these experiments.

Animal use statement

All animals were maintained in the Department of Comparative Medicine, Saint Louis University School of Medicine in accordance with institutional and Public Health Service Guidelines. The mice were housed in microisolator caging, tested, and found to be specific pathogen free (SPF). The mice used in experiments described in this study were maintained in 12-hour day/light schedule on rodent chow (5LOB, LabDiet, Richmond, IA, USA). Saint Louis University School of Medicine Institutional Animal Care and Use Committee (IACUC) approved all procedures performed on mice (protocol numbers 2072 and 2184).

Flow Cytometry

Anti-mouse antibodies for FACS used in this study are listed in Table 1. For FACS, bone marrow cells (BMC) were isolated from the one of the long bones and red blood cells (RBC) were lysed using BD Pharm Lyse^™ (cat# 555899; BD Biosciences, Franklin Lakes, NJ, USA). BMC were resuspended in 50 μL of BD Horizon^™ Brilliant Stain Buffer (cat# 566349, BD Biosciences) and stained for 30 minutes at room temperature with fluorophore-conjugated antibodies protected from light. For intracellular staining (ICS), cells were washed after surface staining, fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton^® X-100 (cat# T8787, Sigma-Aldrich, Saint Louis, MO, USA) and stained overnight at 4°C. Cells were washed, fixed with 1% paraformaldehyde, and analyzed on LSRII instrument with FACS Diva 9.0 (BD Biosciences) software. Gates were determined using combination of single-color and fluor-minus-one controls. FlowJo software (version 10.8.1; FlowJo, LLC, Ashland, OR USA) was used for data analysis.

Table 1:

Antibodies used for flow cytometry

Target	Clone	Fluorochrome
CD45	30-F11	FITC
CD3	17A2	PE-Cy7
CD4	RM4–5	PerCP-Cy5.5
CD8	53–6.7	ACP-CY7
CD44	IM7	APC
CD62L	MEL+14	PacBlue
TNFα	MP6-XT22	PE
IL-17A	TC11–18H10	BV7585

Primary Osteoclast Cultures

Femurs and tibias were isolated from 12-week-old female mice. The proximal end of the femur and the distal end of the tibia were removed, and whole bone marrow (BM) was isolated by spinning at maximum speed for 30 seconds using a tabletop centrifuge. Osteoclast precursors were isolated as previously described. ⁽¹⁵⁾ The cell population was filtered through a 40 μm cell strainer (BD Biosciences), pelleted, resuspended and maintained in αMEM growth medium (αMEM supplemented with 10% heat-inactivated FBS, Penicillin, Streptomycin, 2 mM glutamine, 1 mM sodium pyruvate, non-essential amino acids and 55μM β-mercaptoethanol; all from Sigma-Aldrich) and recombinant murine macrophage colony stimulating factor (M-CSF; cat# 315-02, Peprotech, Cranbury, NJ, USA) at 20ng/ml. Osteoclasts were generated by addition of recombinant murine soluble RANKL (sRANKL; cat# 315-11C, Peprotech) to a final concentration of 20 ng/ml. M-CSF and sRANKL were added every 48 to 72-hours. Bone marrow precursors and osteoclasts were maintained on Petri dishes (non-tissue culture treated) and cells were harvested by a 10 to 15 min treatment with versene (cat# 15040066, Gibco^™, Thermo Fisher Scientific, Waltham, MA, USA).

Chemotaxis assays

Chemotaxis assays were performed using the 24-well Costar transwell system (5μm polycarbonate membrane; Corning). Briefly, bone marrow cells were plated at 2×10⁶ cells/well in 175μl containing growth media with 20ng/ml RANKL and incubated for 48 hours at 37°C, 5.5% CO2. 5×10⁶ freshly harvested splenocytes were added in the top insert of the transwell in 175μl. Control wells had 100ng/ml SDF-1α (cat# 300-28A, Peprotech) in 175 μl αMEM growth medium in the bottom well. Cells in the bottom well and insert were collected using versene (Gibco) after three-hours. Cells were stained with anti-CD3, anti-CD8 and anti-CD4 antibodies and analyzed by FACS for cell counts.

Quantitative PCR (qPCR)

For cortical bone, all femurs and tibias were harvested 4 to 5 weeks after OVX. Bone marrow cells were removed by centrifugation (21,000 g × 30 sec) at room temperature. The bones were cut at 2 mm from both the proximal and distal end to remove the metaphysis. Tissues were flash-frozen in liquid nitrogen and pulverized using a mortar and pestle. The bone powder was transferred to an Eppendorf tube pre-chilled and liquid nitrogen and total RNA was isolated using RNeasy Kit per manufacturer’s instructions (74104, Qiagen, Hilden, Germany). For primary osteoblast cultures, total RNA was isolated using Agilent Total RNA Isolation Mini Kit (5185–5999, Agilent Technologies, Santa Clara, CA, USA) accordioning to manufacturer protocol. Gene expression changes were measured by qRT-PCR using KAPA SYBR^® FAST One-Step Universal (KK4652, Roche, Mannheim, Germany). Primers used are listed in Table 2.

Table 2:

PCR primers used in this study

Primer	Sequence
GAPDH Forward	CTGGAGAAACCTGCCAAGTA
GAPDH Reverse	TGTTGCTGTAGCCGTATTCA
Cx43_F	TTCAATGGCTGCTCCTCACC
Cx43_R	TGGGAGTTGGAGATGGTGCT
RANKL_F	CCCATCGGGTTCCCATAAAGT
RANKL_R	CCCGATGTTT CATGATGCCG
Col1A1_F	GCTCCTCTTAGGGGCCACT
Col1A1_R	CCACGTCTCACCATTGGGG
ALP_F	CCAACTCTTTTGTGCCAGAGA
ALP_R	GGCTACATTGGTGTTGAGCTTTT
OCN_F	CTGACCTCACAGATCCCAAGC
OCN_R	TGGTCTGATAGCTCGTCACAAG

Histological staining of bone sections

For dynamic histomorphometry, calcein green (20 mg/kg) and alizarin red (30 mg/kg) were given intraperitoneally 5 days and 2 days prior to harvesting bone, respectively. All femurs and tibias were harvested 4 to 5 weeks after OVX. Forparrafin embedding, femurs and tibias were fixed in 10% neutral buffered formalin overnight, washed briefly with PBS, placed in 70% ethanol, and delivered to the Musculoskeletal Research Center Histology and Morphometry Core (MRC Core C) at Washington University in Saint Louis (WUSTL) for decalcification and sectioning. Picrosirius red (PSR) and TRAP staining were performed in the core according to their establish protocol⁽¹⁶⁾. For immunofluorescence (IF) staining, 5 μm paraffin embedded sections were washed with Xylene and rehydrated in decreasing concentrations of ethanol (100% to 30%) in Coplin jars. All slides underwent antigen retrieval overnight in citrate buffer (C9999, Sigma-Aldrich) at 55°C. Background Sniper (BS966, BioCare Medical) was used for blocking according to the manufacture’s protocol. PBS containing 0.0025% Triton^® X-100 (Sigma-Aldrich) was used as wash buffer, and PBS containing 0.1% Triton X-100 was used for permeabilization for ICS. Antibodies used as Primary antibodies were incubated overnight at 4°C and secondary antibodies were incubated for 1 hour at room temperature. The following primary antibodies were used: rat anti-mouse CD3 (1:100, Ab11089, Abcam, Cambridge, MA, USA) and rabbit anti-mouse TNFα (1:100, Ab183218, Abcam). The following fluor-conjugated secondary antibodies were used: goat anti-rat-AF647 (1:500, A21247, Invitrogen) and goat anti-rabbit-AF549 (1:500, Ab150080, Abcam). Both primary and secondary antibody dilutions were made in 1.5% BSA in PBS. For phalloidin staining, 50 μm frozen sections were prepared according to a previous publication⁽¹⁷⁾. Briefly, bones were fixed in 4% paraformaldehyde overnight and submitted to MRC Core C for CryoJane frozen sectioning. Sections were collected in 24 wells and stored at −80°C. Prior to staining, sections are transferred to the 48 wells, washed with PBS, and incubated with AF647-conjuated phalloidin (Ab176759, Abcam) for 48 hr. All stained slides were mounted with ProLong^™ Gold antifade (P36934, Invitrogen) under a #1.5 coverslip and cured overnight.

Image analysis

TRAP staining, PSR staining, and double labeled staining images were all taken on a Keyence All-in-One Fluorescence Microscope BZ-X800 (Keyence, Itasca, IL, USA) at 40X. Phalloidin staining was imaged by confocal microscopy on a Leica TCS SP8 (Leica, Wetzlar, Germany). Osteoclast and dynamic histomorphometry were analyzed using BioQuant Osteo Software (Nashville, TN, USA). The average length of the femurs at 19-weeks of age were 12.5 (± 0.4) mm. All image quantifications were done by an operator blinded to surgery groups.

Analysis of lacunae:

A 2 mm region at the mid-diaphysis was selected and regions of interest (ROIs) of 1800 by 900 pixels were drawn for all analysis (Fig. 3). Vascular channels are significantly larger than osteocyte lacunae and were excluded. OVX increased the average lacunae size only on the posterior side. A mean of five ROIs from each side was used with 10 mice per group. TRAP⁺ osteocyte cell counts and lacunae analysis were obtained using Image J/Fiji (NIH, Bethesda, MD, USA). Osteocyte dendrite density was quantified as previously described ⁽¹⁸⁾ using Image J/Fiji. Osteocyte dendrites were visualized using phalloidin staining on transverse sections at 37% to 42% length of the femur. N= 10 mice/group with 4 sections/mouse.

Figure 3. OVX leads to increased lacunae through osteocytic osteolysis on the posterior surface of endocortical bone and to decreased osteocyte dendrite density.

(A) Picrosirius Red (PSR) staining highlights osteocyte lacunae as empty holes. (B) TRAP staining revealed increased TRAP⁺ lacunae in the cortical bone on the posterior side. Insert shows a higher magnification of TRAP⁺ Ocy. Data was obtained from 1 mm section from each side of the mid-diaphysis from 10 mice/group. (C) Osteocyte dendrite density decreased on both anterior and posterior sides post-OVX. 97.5% (P < 0.0001) of the observed variance was attributed to OVX by ANOVA. (D) qRT-PCR analysis of total RNA isolated from the cortical bone at the mid-diaphysis showed an increase in RANKL expression but no change in Cx43 expression. Gene expression data are from 5 mice/group and mean ± SD are plotted. Statistical significance was calculated using an unpaired two-tailed Mann-Whitney test (**p < 0.1, **p < 0.01, ***p < 1×10⁻³).

Collagen fibrils:

Collagen fibril analysis on PSR images were done using CT-FIRE with default settings.⁽¹⁹⁾ All images were rotated so that the long axis of the femur is at 90°. A 2 mm region at the mid-diaphysis was selected and ROIs of 1800 by 900 pixels were analyzed using CT-FIRE. The collagen matrix organization is measured by the frequency of fibrils oriented at 90°, number of fibrils extracted and the average fibril length. There is an inherent difference in the fibril orientation between the anterior and posterior sides. OVX lead more fewer fibrils at 90° and increased the number of fibrils extracted due to decreased average fibril length. The anterior and posterior regions analyzed are highlighted in the yellow and magenta boxes, respectively(see Fig. 4A).

Figure 4. OXV leads to disorganized collagen matrix at the mid-diaphysis without affecting bone formation.

(A) Compared to polarized light, fibrils in all orientations are visible under fluorescence (red channel). On sagittal sections, the clear distinction between the anterior and posterior cortical bone can be appreciated. A representative collagen fibril extraction by CT-FIRE is shown. (B) Double labeled staining confirmed that both mineral apposition rate (MAR) and bone formation rate normalized to bone surface (BFR/BS) were mostly unaffected by OVX. There is a slight decrease in BFR on the posterior side compared to the anterior side post-OVX, but it was not significantly different compared to SHAM. The sides accounted for 20 – 40% of the variance for both MAR and BFR by ANOVA. Graphs are shown was mean ± SEM. For both panels A and B data are from 10 mice/group with 5 ROI/side/mouse. Statistical significance was calculated using an unpaired two-tailed Mann-Whitney test (**p < 0.1, **p < 0.01, ***p < 1×10⁻³, ****p < 1×10⁻⁴).

Micro-CT data collection and analysis

Femurs were scanned in μCT 50 (Scanco Medical) at 70 kVp, 57 μA, 4 W, at a resolution of 10 μm. Analysis was done based on following protocol ⁽²⁰⁾, and default analysis scripts were used without any adjustments. For cortical analysis, the midpoint of the femur was identified, and 50 slices were selected before and after, for a total of 100 slices (1 mm) to evaluate the cortical compartment.

Biomechanical testing

Femurs are harvested, wrapped in gauze and stored in PBS at −80°C until testing. The femurs were delivered to MRC Structure and Strength Core for three-point bending analysis. The operator was blinded to the study.

Statistical analysis

Statistical significance (p values) was assessed in all cases using unpaired two-tailed Mann-Whitney U test because no assumption of normal distribution is assumed. ANOVA and p values were calculated in GraphPad Prism 9.5.0 (GraphPad Software, Inc., La Jolla, CA, USA).

Results

Ovariectomy does not affect the mass of cortical bone but degrades the biomechanical properties.

To understand the early steps that promote cortical bone loss we examined femurs at midshaft from sham-operated and OVX mice. The femurs were analyzed by μCT 5 weeks post-surgery when mice were 19-week-old. We focused on cortical femur midshaft because this region of the bone is measured in the three-point-bending assay. The femur is a relatively uniform cylindrical structure, making it more suitable for standardized three-point bending tests. As is customary for cortical bone, the overall thickness (CT.th) and tissue mineral density (TMD) were examined (Fig. 1A). No statistically significant difference in either parameter was observed between the two groups. However, when the femurs were subject to a three-point bending assay, statistically significant decreases in stiffness, yield-load and post-yield displacement were observed (Fig. 1B; Table 3). The quantification of difference in bone thickness on the anterior and posterior sides in sham and OVX mice is shown in Fig. 1C. However post-OVX, the width on the posterior side decreased but did not change significantly on the anterior side. The side contributed 54.5% (P < 0.0001), surgery contributed 12.9% (P = 0.0008) of the observed variance, with 15.7% (P = 0.0003) due to the interaction between side and surgery by two factor ANOVA.

Figure 1. Ovariectomy (OVX) leads to change in biomechanical and physical properties of cortical bone but not average bone mineral density.

(A) No change in cortical thickness (Ct.Th) or tissue mineral density (TMD) was detected at the mid diaphysis of the femur at 10 μm resolution. (B) Three-point bending assay showed that OVX in WT mice decreases the stiffness, yield load, and post-yield displacement (also please see Table 3). (C) The width of cortical bone is different on the anterior and posterior sides sham-operated mice. The width of cortical bone on the anterior and posterior sides normalized by the diameter of the bone. Graphs show mean ± SD. Data is from 10 mice/group, the same femur is used for both μCT and three-point bending. Pairwise statistical significance was calculated using an unpaired two-tailed Mann-Whitney test (*p < 0.1, **p < 0.01).

Table 3.

Summary of Cortical, Histomorphometry, Biomechanical and Material Properties Data in WT and IL15RA^ΔT mice

Experiment	Parameter	WT (C57BL6/J)		IL15RAΔT (IL15RAf/f × Lck-Cre)
		SHAM (n=10)	OVX (n=10)	SHAM (n=10)	OVX (n=10)
μCT (femur)	BV/TV	0.1259 ± 0.01212	0.07691 ± 0.006515 ****	0.1038 ± 0.05300	0.1149 ± 0.07363 NS
Trabecular	Tb.N	3.680 ± 0.1059	2.984 ± 0.2873 **	3.894 ± 0.2293	3.677 ± 0.3572 NS
	Tb.Th (mm)	0.03735 ± 0.002528	0.03026 ± 0.0008895 **	0.04039 ± 0.0002978	0.03995 ± 0.007224 NS
	BMD (mgHA/cm3)	116.4 ± 7.251	83.91 ± 4.794 ***	111.7 ± 6.855	108.19 ± 4.157 NS
Cortical	TA (mm2)	1.633 ± 0.04968	1.604 ± 0.06164 NS	1.632 ± 0.05911	1.521 ± 0.06802 NS
	BA (mm2)	0.8155 ± 0.05008	0.7592 ± 0.05541 NS	0.7286 ± 0.03068	0.7443 ± 0.04959 NS
	MA (mm2)	0.8771 ± 0.05789	0.9293 ± 0.04748 NS	0.9666 ± 0.03226	0.8857 ± 0.08974 NS
	Ct.Th (mm)	0.1990 ± 0.01226	0.1858 ± 0.01692 NS	0.1843 ± 0.01218	0.1811 ± 0.006984 NS
	TMD (mgHA/cm3)	1116 ± 12.16	1108 ± 8.683 NS	1111 ± 10.77	1120 ± 6.162 NS
Static Histomorphometry (mid-diaphysis femur)	Oc.S/BS (%) Anterior Posterior	ND 6.859 ± 2.007 †	ND 21.03 ± 3.254 † ****	ND 6.950 ± 1.472 †	ND 23.23 ± 4.331 † ****
	Oc.N/BS (#) Anterior Posterior	ND 3.428 ± 1.413 †	ND 12.42 ± 1.917 † ****	ND 3.418 ± 0.6555 †	ND 12.11 ± 2.891 † ****
Dynamic Histomorphometry (mid-diaphysis femur) Endosteal	Anterior MS/BS Posterior	0.7519 ± 0.1650 0.2985 ± 0.1645 †	0.5923 ± 0.2101 NS 0.4130 ± 0.1820 NS	0.6187 ± 0.1452 0.3191 ± 0.1461 †	0.7124 ± 0.2316 NS 0.3292 ± 0.1277 †
	MAR (μm/day) Anterior Posterior	1.530 ± 0.1391 1.244 ± 0.3148	1.470 ± 0.263 8 NS 1.246 ± 0.2960 NS	1.621 ± 0.3105 1.551 ± 0.2734	1.5981 ± 0.2361 NS 1.553 ± 0.3014 NS
	BFR/BS (μm/day) Posterior Anterior	1.075 ± 0.06464 0.7336 ± 0.3803 †	0.9383 ± 0.1320 NS 0.5861 ± 0.2004 † *	1.522 ± 0.05702 0.6257 ± 0.07768 †	1.482 ± 0.08120 NS 0.5768 ± 0.06233 †
Periosteal	MS/BS Anterior Posterior	0.1770 ± 0.1875 0.4981 ± 0.1774 †	0.06165 ± 0.06180 NS 0.1367 ± 0.1177 ***	0.1486 ± 0.1273 1.628 ± 0.3005 †	0.08294 ± 0.07265 NS 1.628 ± 0.3005 †
	MAR (μm/day) Anterior Posterior	0.1067 ± 0.3539 1.129 ± 0.5922 †	0.1559 ± 0.4126 NS 0.2288 ± 0.3981 **	0.1277 ± 0.4274 0.9675 ± 0.6122 †	0.09512 ± 0.5792 NS 1.037 ± 0.4695 †
	BFR/BS (μm/day) Anterior Posterior	0.04702 ± 0.1560 0.6026 ± 0.3503 †	0.03194 ± 0.07824 NS 0.06062 ± 0.1151 **	0.04223 ± 0.0771 0.5354 ± 0.2901 †	0.04634 ± 0.06811 NS 0.6104 ± 0.3740 †
Biomechanical and Material Properties (mid-diaphysis femur)
	Stiffness (N/mm)	83.70 ± 4.779	66.77 ± 4.396 **	76.17 ± 3.197	73.82 ± 4.392 NS
	Yield Load (N)	10.86 ± 1.467	7.610 ± 0.5160 **	10.05 ± 2.169	9.486 ± 1.330 NS
	Maximum Load (N)	14.95 ± 0.7652	14.44 ± 1.349 NS	14.51 ± 1.271	13.35 ± 0.5386 NS
	Post-Yield Displacement (mm)	1.230 ± 0.4494	0.3172 ± 0.04424 **	1.222 ± 0.3667	1.210 ± 0.3765 NS
	Work-to-Fracture (N*mm)	11.16 ± 2.243	6.682 ± 1.510 **	10.55 ± 2.761	10.81 ± 1.805 NS
	Young’s Modules (N/mm2)	4546 ± 915.0	4446 ± 787.3 NS	4870 ± 982.7	4642 ± 637.8 NS
	Yield Stress (N/mm2)	119.0 ± 19.47	104.4 ± 23.52 NS	110.8 ± 14.57	107.7 ± 23.21 NS
	Ultimate Stress (N/mm2)	137.2 ± 9.821	128.6 ± 6.504 NS	138.3 ± 7.166	133.4 ± 12.42 NS

TV = tissue volume; BV = bone volume; BMD = bone mineral density; Tb.n = trabecular number; Tb.Th = trabecular thickness; TA = tissue area; BA = bone area; MA = medullary area; Ct.Th = cortical thickness; TMD = tissue mineral density; Oc.S/BS = osteoclast surface per bone surface; Oc.N/BS = osteoclast numbers per bone surface; MS = mineralized surface; BS = bone surface; MAR = mineral apposition rate ; BFR/BS = bone formation rate normalized by bone surface.

All p values were calculated using an unpaired, two-tailed Mann-Whitney test, comparing (1) OVX to SHAM and (2) anterior to posterior within the same group.

Mean ± SD is shown for all parameters. ND = none detected

NS = not statistically significant

^*p < 0.1

^**p < 0.01

^***p < 1 × 10⁻³

^****p < 1 × 10⁻⁴

^†denotes statistical significance between anterior and posterior

The endosteal surface on the posterior side of the femur is enriched in resorbing osteoclasts

To determine the whether the decreased thickness on the posterior cortical bone was due to increased bone resorption, we performed TRAP staining at the mid-diaphysis of the femur in otherwise healthy 14-week-old C57BL6/J female mice. We examined both transverse and sagittal sections to ensure the appreciation of the three-dimensionality of bone. For quantitation, the sagittal cuts were used because it was easier to measure a 2 mm flat bone surface consistently (the denominator BS) in BioQuant software. OC were only found on the posterior endosteal surface (Fig. 2A; Table 3) in sham-operated mice. We next evaluated OC numbers at each endocortical surface 5 weeks post-OVX. We found that OVX further increased OC numbers (Fig. 2B; Table 3). Two factor ANOVA indicated that side accounted for 59.4% (P < 0.0001) and surgery accounted for 18.4% (P < 0.0001) of the observed variance in OC.S/BS and the interaction between surgery and side was 18.4% (P< 0.0001). OC.N/BS showed similar variance by two-way ANOVA. As our studies have shown that T-cells produce TNFα and IL-17A post-OVX, we stained for CD3⁺ cells using immunofluorescence and found T-cells located proximal (i.e., within 60 μ) to the posterior endosteal surface (Fig. 2C). Both the total number T-cells and percent of TNFα⁺ T-cells increased post-OVX and were enriched near the posterior endosteal surface (Fig. 2C). To determine the mechanism for the co-localization we used a trans-well chemotaxis assay. OC were generated from bone marrow cells (BMC) using M-CSF and RANKL. We first measured chemokines produced by OC cultures in collected media using multiplexed ELISA. Next, T-cells were added to an insert with a 5μm pore membrane, that was placed over OC at the bottom of the well. The number of cells that migrated from top to bottom in three hours well were enumerated by FACS and are displayed as a percent of the total number of T-cells added. We found that OC produce multiple chemokines (Fig. 2D), that recruit both CD4⁺ and CD8⁺ T-cells (Fig. 2D). Taken together, results in Fig. 2 show that the localization of OC is spatially distinct in cortical bone in healthy mice, a feature exaggerated by OVX. Post-OVX there is increased number of OC on the posterior endosteal surface. As TNFα increases osteoclastogenesis, the increased OC on the posterior surface is most likely due to recruitment of proinflammatory T-cells post-OVX.

Figure 2. Resorbing osteoclasts have distinct spatial distribution at the mid-diaphysis and colocalize with T-cells.

(A) TRAP staining revealed that osteoclasts are enriched on the posterior surface, both on sagittal and transverse sections. (B) OVX resulted in increased osteoclast number (OC.S/BS) on the posterior endosteal surface. (C) Immunofluorescence staining of CD3⁺ (red arrow) and TNFα⁺ (grey arrow) T-cells in the bone marrow near the endosteal surface (black dashed line). A 2 mm section at the mid-diaphysis on the sagittal plane was analyzed and CD3⁺ cell that are within 60 μm from the endosteal surface were counted. OVX increased the percentage of TNFα⁺ T-cell largely on the posterior side. The average of 4 fields/slide are shown for 5 slides per mouse per group. (D) Primary osteoclasts produce multiple T-cell attracting chemokines in vitro as measured by ELISA, and recruit both CD4⁺ and CD8⁺ T-cells in trans-well assays. Data is representative of two independent experiments with 3 technical replicates per condition. Graphs are shown was mean ± SD. Intergroup statistical significance was calculated using an unpaired two-tailed Mann-Whitney test (*p < 0.05, **p < 0.01, ***p < 1×10⁻³, ****p < 1×10⁻⁴).

OVX increased lacuna size and osteocyte resorption as well as decreased dendrite density

We sought to further assess whether the increased number of T-cells also leads to cortical porosity in mice post-OVX. We used a picrosirius red (PSR) staining to visualize lacunae. As seen in Fig. 3A, OVX increased the average lacunae size on the posterior side at the femoral midshaft. ANOVA showed that 21.0% (P < 0.0001), 18.2% (P = 0.0002) and 21.6% (P < 0.0001) of the observed variance was attributed to side, surgery and the interaction between the two variables, respectively. To address whether this increase in cortical porosity is due to increases resorption by Ocy, we focused on TRAP staining in the cortical bone and found increased TRAP⁺ lacunae that were only on the posterior side (Fig. 3B). The sidedness of lacuna explained 21.0% (P < 0.0001), and E₂ loss explained 18.2% (P = 0.0002) of the observed variance, while 21.6% (P < 0.0001) was attributed to the interaction between side and surgery by two-factor ANOVA.

Ocy, embedded within bone, detect mechanical load and strain. Ocy sense changes in load and aberrant strain through their dendritic processes, and direct OC and/or OB to alter bone properties to maintain strength. ^(21,22) We hypothesized that recruitment of proinflammatory T-cells in proximity to endosteal bone would affect Ocy. To assess Ocy dendrites phalloidin staining on thick (50 μm) frozen sections was performed. ⁽¹⁷⁾ We analyzed the tibias for osteocyte dendrites because the femurs from the same animal were used for three-point bending, ΤRAP and PSR staining. The anterior and posterior regions where osteocyte dendrites were analyzed are highlighted in yellow and magenta boxes, respectively, in a representative μCT image of a transverse view of the tibia at 37% length of the diaphysis (Fig. 3C). We found that dendrite density decreased by ~50% on both sides post OVX (Fig. 3C), consistent with increase of TNFα+ T-cells on both sides. Results in Fig. 3 show that, while increased osteocytic osteolysis was only present on one side, Ocy dendrite density decreased globally post OVX resulting in disrupted Ocy communication.

OVX leads to disorganized collagen matrix and altered bone formation at the mid-diaphysis

Αs Ocy direct bone remodeling by sensing strain or load, disruption of Ocy dendrites can also affect bone formation by osteoblasts, particularly the quality of the ECM. To assess OB activity in cortical bone, sections were stained with picrosirius red and imaged under fluorescence light to visualize the collagen fibrils. PSR staining of normal healthy bone showed that collagen fibrils in the anterior cortical bone are lamellar. In contrast, bone on the posterior side is partially woven (Fig. 4A). To further assess quality of the ECM, we used CT-FIRE software to analyze collagen fibrils. ⁽¹⁹⁾ CT-FIRE assesses the degree of collagen matrix organization as the percentage of fibrils oriented at 90° (anisotropy), as well as the number and length of the fibrils. Prior to analysis all images were rotated so that the long axes of the femurs were at 90° and equal numbers of 1800 by 900 pixels ROIs were selected on both the anterior and posterior side (Fig. 4A). In sham-operated mice about 60% of fibrils were lamellar (e.g., aligned) on the anterior side but only 52% on the posterior side (Fig 4A). There was an increase in the total number of fibrils extracted post OVX, likely due to the shortening of the fibril length (Fig. 4A). Interestingly, at the mid-diaphysis, dynamic histomorphometry revealed that bone formation was predominantly occurring on the anterior endosteal surface (Fig. 4B). While there was no difference in mineral apposition rate (MAR), the bone formation rate (BFR) was reduced on the posterior endosteal surface post OVX. Indeed, OVX also significantly reduced OB number on the posterior periosteal surface (Fig. 4B). Together, results in Fig. 4 showed that OVX leads to a disorganized collagen matrix on both anterior and posterior endosteal regions. OVX also reduces OB activity on periosteal surface. Collagen disorganization is known to reduce yield load and stiffness observed in the three-point bending assay⁽²³⁾.

Bone fragility post-OVX is due to TNFα and IL-17A expressed by T-cells:

OVX leads to loss of estrogen, which triggers memory T-cell (T_M) to secreted TNFα and IL-17A To determine if T-cell derived cytokines caused the observed compartment-specific effects on cortical bone, sham and OX surgeries were performed on IL15RA^ΔT mice. We observed that OVX did not increase the production of TNFα and IL-17A in T_M (Fig. 5A) in IL15RA^ΔT mice as previously described⁽¹²⁾. The number and distribution of OC observed in IL15RA^ΔΤ was same as in WT mice (Table 3). The collagen matrix organization, sham-operated mice retain the inherent difference between the anterior and posterior cortical bone and OVX did not affect fibril alignment, the number of fibrils extracted or average fibril length (Fig. 5B). There was no change in average lacunae size (Fig. 5C), the percent of TRAP⁺ Ocy (Fig. 5D; Table 3) and Ocy dendrite density (Fig. 5E) post OVX in IL15RA^ΔT mice. Three-point bending assay showed no compromised in biomechanical properties and μCT showed no difference in cortical bone parameters post OVX in IL15RA^ΔT mice (Fig. 5F and Fig. 5G; also see Table 3). The above data established that the Ocy and OB defects post OVX is driven by TNFα and IL-17A produced by T_M.

Figure 5. IL-15 signaling in T-cell is required for diminished bone quality post-OVX.

(A) Flow cytometry of bone marrow cells showed that OVX in IL15RA^ΔT mice does not increase expression of TNFα and IL-17A in T_M compared to WT mice, in agreement with our previous findings.⁽¹²⁾ (B) The inherent differences in the collagen fibril orientation between the anterior and posterior were present in IL15RA^ΔT mice, and OVX did not further increase the matrix disorganization. There was no change in (C) average lacunae size, (D) TRAP⁺ osteocytes or (E) osteocyte dendrite density post-OVX in IL15RA^ΔT mice. (F) Three-point bending assay showed that in IL15RA^ΔT mice, OVX did not result in diminished biomechanical properties. (G) No change in cortical thickness (Ct.Th) or tissue mineral density (TMD) was detected at the mid-diaphysis on μCT at 10 μm resolution. All analyses were performed using 10 mice/group with identical numbers as shown for corresponding WT mice in Figs. 1 through 4. Graphs are shown as mean ± SD. Statistical significance was calculated using an unpaired two-tailed Mann-Whitney test.

Discussion

Fractures associated with bone fragility represent a major public health concern owing to high morbidity and mortality post-fracture in postmenopausal women.⁽²⁴⁾ Current clinical assessment of bone health predominantly relies on BMD measurements, typically obtained via DXA scans. However, BMD reflects bone mass rather than bone quality, which is equally critical in determining fracture risk⁽²⁵⁾ Even with the development of tools like FRAX score that incorporates additional factors to predict fracture probability, a significant gap in assessing bone quality remains. We designed this study to gain further insights into how estrogen (E₂) loss affects bone quality and biomechanical properties. We tested whether we could observe changes in the biomechanical properties as well as cortical bone microarchitecture and BMD in ovariectomized (OVX) mice.

As most common serious fractures, such as intertrochanteric or femoral neck fractures are of the cortical bone leaving patients bed-ridden, in the present study we examined changes in the cortical bone post OVX. ⁽²⁶⁾ Femoral neck is estimated to have 25–50% trabecular component, leaving 50–75% as cortical bone. ⁽²⁶⁾ Femoral hip fractures are associated with high mortality: one-year survival of 20–25% and rising to 50–70% at 3 years post fracture. ⁽²⁷⁾ Unlike prior studies, where cortical bone loss is analyzed at 15 weeks⁽²⁸⁾ or later, post-OVX, we found that at 5 weeks post-OVX cortical BMD does not change (Fig. 1A). Despite the lack of change in cortical BMD, at this early time-point we observed a reduction in stiffness, yield-load and post-yield displacement (Fig. 1B). Unexpectedly, analysis of μCT data showed most of loss of bone mass was in the posterior surface of the cortical bone in the femur (Fig. 1C), which explains why the total BMD does not change.

To understand the reason(s) for the “sidedness” in change of thickness, we stained for osteoclasts (OC) on each side of the femur at mid-diaphysis. Surprisingly, we found that TRAP⁺ OC were entirely on the posterior endosteal surface (Fig. 2A) in non-surgically operated female mice. When we compared our findings of OC localization to previously published strain maps developed using strain gauges and finite element analysis, we found that there is a remarkable concordance with the sites under high strain. ⁽²⁹⁾ Our results indicate that the posterior side of cortical bone, under habitual load due the positioning of the femur in mice, has increased bone resorption.

OVX further exacerbated the difference in TRAP⁺ OC between anterior and posterior endocortical surfaces (Fig. 2B). Our laboratory has previously shown that TNFα and IL-17A secreted by memory T-cells are required for bone loss post-OVX.⁽¹²⁾ To assess whether T-cells are in proximity of TRAP⁺ OC, we stained for CD3⁺ T-cells at the midshaft. We observed that T-cells were enriched on posterior side and colocalized (less than 60 μm away) with resorbing OC (Fig. 2C). Both the total number of T-cells and TNFα⁺ T-cells increased after OVX which corresponds to increased OC resorption (Fig. 2C). Cell culture experiments show that OC produced chemokines that recruit T-cells (Fig. 2D). Both TNFα and IL-17A have been shown to drive osteoclastogenesis and favor bone resorption^(30,31). OVX creates a vicious cycle where resorbing OC recruit pro-inflammatory T-cells to the posterior endosteal surface which drives further bone resorption. Although, TNFα⁺ T-cells also increase on the anterior side, there was no bone resorption because are no observed OC on the anterior endosteal surface.

A prominent feature of osteoporosis patients is the increase in cortical porosity.^(32,33) Studies by Sharma et al. ⁽³⁴⁾ showed increased cortical porosity near the metaphysis post-OVX in rats, due to both expanding vascular channels and lacunae-canalicular network (LCN). In this study, we examined cortical porosity at the mid-diaphysis using picrosirius red (PSR) staining. OVX in mice increased the average lacuna size only on the posterior side (Fig. 3A) and this is due to increased osteocytic osteolysis as revealed by examining the TRAP-stained cortical bone (Fig. 3B). Osteocytic osteolysis has been previously characterized in lactation but this is the first time it has been observed post OVX^(35,36).

Prior studies have shown that Ocy dendrite density decreases in the context of aging ⁽¹⁸⁾. Here, we provide evidence that this also occurs post OVX, in which Ocy dendrite density was decreased on both sides (Fig. 3C). Since, osteocytic osteolysis only occurred on the proximal region, and only on the same side as the OC, we hypothesize that this is due to communication between OC and Ocy, perhaps due to RANKL. Additional studies are needed to elucidate the mechanism for both the increased osteocytic osteolysis and loss of dendrite density post-OVX.

As Ocy direct OC and OB activity by sensing aberrant loads and microdamage to the bone, we examined OB activity on the posterior and anterior sides of cortical bone. We chose to examine the work performed by OB by examining collagen fibril organization and bone formation rates. PSR staining and analysis using CT-FIRE⁽¹⁹⁾ showed that the collagen fibrils are more lamellar on the anterior side of cortical bone even in sham-operated mice (Fig. 4A). The collagen fibril organization was decreased by OVX on both sides of cortical bone, and we observed an increased in the total number of fibrils being extracted, in large part due to a decrease in the average fibril length (Fig. 4A). The details of how the collagen fibrils are organized to best adapt to load remains unclear, and further studies are necessary to elucidate the pathways that are affected post-OVX. No difference was observed in mineral apposition rate (MAR) between anterior and posterior endosteal surfaces in sham-operated mice, and OVX did not have a significant effect on MAR (Fig. 4B). However, we observed a significant diminishment of MAR and BFR on the periosteal posterior surface of cortical bone. Bone formation rate (BFR) was not statistically different between posterior and anterior endosteal surfaces in sham-operated mice. OVX disproportionately reduced BFR on the posterior side (Fig. 4B). This result highlights both the difference the rate of trabecular and cortical remodeling, as well as the sensitivity of each compartment to E₂ loss.

We have previously shown that loss of bone mass post-OVX is through TNFα and IL-17A produced by memory T-cells (T_M) in a IL-15 dependent mechanism.⁽¹²⁾ As many cells express E₂ receptors, the E₂ loss may affect cellular function(s) directly or through T_M secreted TNFα and IL-17A. To distinguish whether osteolineage dysfunction is mediated by TNFα and IL-17A, we ovariectomized IL15RA^ΔT mice and examined changes in OB and Ocy. Using the IL15RA^ΔT mice, we found that all effects on OC, OB and Ocy result from T_M expressed TNFα and IL-17A. These results indicate increased OC recruit proinflammatory T-cells⁽¹²⁾ to posterior endosteal surface, which then leads to dysfunction of both OB and Ocy. This leads to disrupted Ocy communication and processive OB making shorter, more disorganized collagen fibrils on the posterior surface of cortical bone, both contributing to deteriorated biomechanical properties and fragile bone.

The current study extends the previous findings by characterizing spatial localization of OC and T-cells within cortical bone. Prior studies using loading and finite element analysis showed distinct distribution of compressive and tensile strains due to the geometry of the bone, where the anterior femur and posterior tibia experience compression and the posterior femur and anterior tibia experience tension. ^(37,38) Here, we observed TRAP⁺ OC on the posterior endosteal surface, a region under tensile stress. Post-OVX, the TRAP⁺ OC increased on the posterior endosteal surface, and they colocalize with both TNFα⁺ T-cells and with osteolytic osteocytes. The OC The effect of T-cell mediated inflammation causative as evidenced by the results from OVX of IL15RA^ΔΤ mice.

We recognize that habitual load in mice is not the same as in humans (as bipedal), but in principle and based on the shape of the bone, can be anticipated to have a sidedness. Other studies have addressed the correlation between habitual loading and biomechanical properties that increase the risk of fracture in humans and animal models. For instance, Thomas et al. showed that patients with low-trauma hip fractures had reduced cortical thickness on the posterior side.⁽²⁶⁾ Several other studies using CT or HR-pQCT have also noted lower cortical thickness on the side under load in women with fragility fractures. ^(39,40) Our study may provide a mechanism for these observations.

A limitation of our study is the fact that bone is three-dimensional, and histological investigations are two-dimensional. While we were able to validate our observations on sagittal planes using transverse sections, future studies are required further dissect the distinct spatial effects of T-cell mediated inflammation on the Ocy and OB. Whole bone clearing, and spatial transcriptomic approaches are needed to further elucidate the hyper-localized defects in bone quality. A related technical limitation is that we use decalcified sections to identify lacunae size (Figs. 3A and 5C) which may affect measurements. A more accurate method for quantifying lacunae size is back-scatter electron microscopy (BSEM) of undecalcified bone. However, as we only observed enlarged lacunae on the posterior side and no changes on the anterior side of the same physical bone section in each of 10 bones examined, we believe that the data is reasonably accurate as, it is internally controlled. We also have controls from sham-operated WT mice and from sham-operated and OVX IL15RA^ΔT mice that showed no difference between anterior and posterior sides.

This study is the first to demonstrate the distinct spatial and asymmetric localization of pro-inflammatory T-cells and their direct impact on osteolineage cells in an OVX mouse model. We used biomechanical testing, μCT, dynamic histomorphometry with analysis of functional, cellular and molecular changes to characterize the degradation of bone quality of in cortical bone post-E₂ loss. Detailed studies of fracture mechanics (angle of falls, impact forces, models of fracture etc.) show that when force is applied to cortical bone, breaks begin at sites of compromised microarchitecture known as microcracks. ^(41,42) Catastrophic failure occurs as the break coalesces with other local defects and propagates through fissures by shear forces^(43,44). Thus, a high concentration of microcracks likely indicates a region with greater propensity for fragility fracture. Healthy Ocy sense damage with their dendrites, through highly connected canicular networks, directing OC and OB, in a coupled fashion, to repair the damage and maintain bone quality⁽⁴⁵⁾.

In summary, our results show that TNFα and IL-17A affects the anterior and posterior sides of the bone asymmetrically. These cytokines lead to reduction in the number of dendrites per Ocy in OVX mice. The reduction in Ocy dendrites is also observed in humans. ^(46,47) We anticipate reduced Ocy dendrites will lead to accumulation of unrepaired microcracks bone, and consequently to increased risk of fracture. As Ocy initiate coupled bone remodeling, the suppression of bone formation by TNFα and IL-17A creates an imbalance in remodeling dynamics on the posterior endosteal region. Asymmetric bone remodeling, in combination with impaired Ocy mechanosensing, will lead to bone that is maladapted to normal load further increasing fracture susceptibility. Lastly, on the anterior endosteal compartment the accumulation of disorganized extracellular matrix (ECM) also will reduce the biomechanical properties of the bone. We note, none of the current therapies on the market, whether anti-resorptive or anabolic, directly targets inflammation. Future studies will investigate whether these drug classes improve osteolineage functions. Additionally, we will examine the relationship between localized bone quality defects and whole bone strength and develop new screening tools to prevent fractures in patients.

Data Availability Statement:

The authors confirm that the data supporting the findings of this study are available within the article. Any additional data and/or methods will be made available from the corresponding author, RA, upon reasonable request.