Structural role of osteocalcin and its modification in bone fracture

Abstract

Osteocalcin (OC), an abundant non-collagenous protein in bone extracellular matrix, plays a vital role in both its biological and mechanical function. OC undergoes post-translational modification, such as glycation; however, it remains unknown whether glycation of OC affects bone's resistance to fracture. Here, for the first time, we demonstrate the formation of pentosidine, an advanced glycation end-product (AGE) cross-link on mouse OC analyzed by ultra-performance liquid chromatography. Next, we establish that the presence of OC in mouse bone matrix is associated with lower interlamellar separation (distance) and thicker bridges spanning the lamellae, both of which are critical for maintaining bone's structural integrity. Furthermore, to determine the impact of modification of OC by glycation on bone toughness, we glycated bone samples in vitro from wild-type (WT) and osteocalcin deficient (Oc^−/−) mice, and compared the differences in total fluorescent AGEs and fracture toughness between the Oc^−/− glycated and control mouse bones and the WT glycated and control mouse bones. We determined that glycation resulted in significantly higher AGEs in WT compared to Oc^−/− mouse bones (delta-WT > delta-OC, p = 0.025). This observed change corresponded to a significant decrease in fracture toughness between WT and Oc^−/− mice (delta-WT vs delta-OC, p = 0.018). Thus, we propose a molecular deformation and fracture mechanics model that corroborates our experimental findings and provides evidence to support a 37%–90% loss in energy dissipation of OC due to formation of pentosidine cross-link by glycation. We anticipate that our study will aid in elucidating the effects of a major non-collagenous bone matrix protein, osteocalcin, and its modifications on bone fragility and help identify potential therapeutic targets for maintaining skeletal health.

I. INTRODUCTION

The remarkable fracture properties of bone originate from its complex hierarchical structure. It is becoming increasingly apparent that the nanoscale architecture and chemical interaction of bone's constituents are critical to the understanding of its fracture behavior.^1–3 In this context, non-collagenous proteins (NCPs) constitute only 10% of the total organic content of bone, but exhibit multifunctional roles known to affect its biological function and resistance to fracture.^4,5 Studies have demonstrated that NCPs act as structural molecules and imbue mineralized collagen fibrils with superior resistance against interfibrillar sliding during crack propagation.^6–8 In addition, NCPs contribute to toughening mechanisms such as crack deflection, microcrack-toughening, diffuse damage, and uncracked ligament formation, as well as sacrificial bonding, all of which operate across the scales of hierarchy enhancing energy dissipation by bone.^9–11 Given the seminal structural and interfacial “glue-like” role of NCPs in bone,^7,9 it is understandable that any modification in NCP structure will likely have severe implications for the resistance of bone to fracture.

Osteocalcin (OC), also known as bone γ-carboxyglutamate protein, is the most abundant NCP in bone.¹² This small protein comprises ∼1% of total body protein¹³ and has a molecular weight of ∼5.6 kDa with three ɣ-carboxyglutamic (GLA) residues, three α-helices, a disulfide bond, a C-terminal hydrophobic core, and an unstructured N-terminus.^14,15 The GLA residues allow OC to bind specifically and strongly to hydroxyapatite (HA), the principal mineral phase of bone, and participate in mineral crystal growth and maturation.¹⁶ OC is produced during bone formation exclusively by osteoblasts and is commonly used as a clinical marker of bone turnover.^17,18 During bone remodeling, OC acts as a chemoattractant and differentiation agent for osteoclasts.¹⁹ Much of the newly synthesized protein is adsorbed to bone, while a small amount may be released in circulation. As such, OC in circulation has also been proposed to act as an endocrine hormone with pleiotropic effects on multiple end organs and tissues regulating functions ranging from bone mass accumulation to body weight, adiposity, glucose and energy metabolism, male fertility, brain development, and cognition.^17,20–23 Mouse bone lacking OC has altered mineralization including smaller and less mature mineral crystals with lower type B carbonate substitutions compared to wild-type mice.^24–26 OC does not directly affect collagen assembly.^25,27 However, OC together with osteopontin protein links extrafibrillar mineral aggregates and contributes to energy dissipation and fracture toughness.^9,24,25,27

Osteocalcin remains deposited within the skeleton and has been highly conserved throughout evolution. OC is only replaced with newly synthesized material during bone remodeling.²⁸ However, long-lived proteins are highly susceptible to posttranslational modifications, such as non-enzymatic glycation (NEG).^29,30 NEG is the spontaneous reaction of sugars with alpha- and epsilon-amino groups on proteins in vivo and in vitro.^31,32 The reaction occurs by the initial formation of a Schiff base or aldimine, which then undergoes an Amadori rearrangement to a more stable ketoamine.^33,34 The extent of this process is dependent upon the concentration, reactivities, and pKa values of the protein's amino groups, the average concentration of sugar in the surrounding medium, and the life span of a given protein. Typically, glycation of proteins has significant effects on their function because it induces conformational changes and/or decreases their solubility and resistance to enzymatic digestion.³⁵ This aspect has been well demonstrated in diabetes, where modifications to bone matrix proteins such as collagen type-I²⁹ lead to lowered bone tissue fracture toughness and increased risk of fracture.^36–38

It has been recognized that OC in human and bovine bone can undergo glycation at the N-terminus.³⁹ It was established that the level of glycated osteocalcin increased linearly with age.^39,40 OC may therefore play a role in the pathogenesis of diseases affecting the skeleton. Recently, Thomas et al.⁴¹ induced glycation on synthetic human osteocalcin peptide in vitro and confirmed through fragmentation that the N-terminus is the general region of glycation on the peptide. However, the exact location on the N-terminus was not identified and the authors postulated that glycation could occur either on the N-terminal tyrosine or the arginine residue at position 19 on synthetic human osteocalcin. Interestingly, the results of a recent molecular dynamic simulation, performed on healthy and glycated human OC using imagined glucosepane-like cross-link between arginine and N-terminus, demonstrated that α-helices in healthy OC have high affinity for HA and they allow for significant energy dissipation during deformation, through stick-slip behavior, sacrificial bonding, and HA interaction.⁴² Tavakol et al.⁴² established that glycation disrupted OC's α-helical structures through covalent cross-linking between Arginine and N-terminus regions, causing reduced affinity of OC for HA and thereby lowered energy dissipation. Given the diverse roles of OC in bone, it is possible that OC glycation may also impair the structure of bone matrix and consequently tissue-level mechanical responses.

In this study, we investigated the impact of the loss (i.e., using bone samples from osteocalcin deficient OC^−/− mouse) and modification (i.e., ribosylation) of OC on bone toughness.

To accomplish this, we first demonstrated that the level of OC in human bone declines with age. Next, using mouse models, we identified, for the first time, the glycation of OC by the formation of pentosidine (PEN) that is likely to contribute to age-related fragility. Thus, to gain understanding of the contribution of OC to bone toughening, we then mechanically tested wild-type (WT) and Oc^−/− mouse bone specimens and determined that various toughening mechanisms were affected by the loss of OC, particularly interlamellar separation, crack bridging, and crack deflection. Finally, to isolate the effects of glycated OC on bone mechanical competence, we performed in vitro ribosylation of WT and Oc^−/− mouse bones and compared the change in the accumulation of advanced glycation end-products (AGEs) and fracture toughness with and without OC in bone matrix.

II. RESULTS

A. Age-related loss of osteocalcin in human bone

Mineral-bound proteins were extracted from tibial cortical bone of male donors (n = 14, ages 19–85), and total protein concentration was quantified using a colorimetric detection assay. A solid-phase Enzyme Amplified Sensitivity Immunoassay (EASIA) was used to quantify the amount of OC in 25 μL of protein extract. It was established that the level of OC significantly declined with age (r² = 0.74, p < 0.001) (Fig. 1). OC content increased until ∼45 years of age and then plateaued, and a relatively steady level of OC occurred for ∼10 years of human life. After reaching age 55, a decline in OC content was observed. Given that OC reflects the activity of osteoblasts, the significant decline in OC level seen here is consistent with reports on reduced bone formation.⁴³

FIG. 1.

Osteocalcin levels in organic matrix of bone determined for male human donors representing each decade of human life (i.e., from second to ninth decade). The power calculated using G*Power 3.1.9.7 software is 0.9514 (or 95%). Outliers are indicated by +.

B. First evidence of osteocalcin modification by pentosidine (PEN) in mice

To determine the nature of osteocalcin modification in a mouse model, a combination of sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and ultra-high performance liquid chromatography (UPLC) methodology was used. Total bone matrix proteins were extracted from cortical bone pieces of WT and Oc^−/− mouse tibiae (n = 3/group). The proteins were separated followed by gel staining and processing of protein spots. OC was cut from the gel in the case of WT samples and from the matching control areas of the gel were isolated from Oc^−/− samples. Next, the OC samples were hydrolyzed in 6 N HCl and analyzed using an acuity UPLC machine. The UPLC chromatograms revealed the presence of PEN on OC protein isolated from WT mouse samples (Fig. 2), which was not detected in Oc^−/− mouse samples. Thus, for the first time we demonstrate that PEN is formed intramolecularly in mouse OC.

FIG. 2.

UPLC chromatogram demonstrating the presence of pentosidine (PEN) in the wild-type mouse sample (a) but not in the control sample from the genetically modified Oc^−/− mouse that is missing the OC protein due to its deletion (b).

C. Deletion of osteocalcin influences bone fracture and toughening mechanisms

Femoral bone from Oc^−/− and WT mice (n = 8/group) was mechanically tested to measure toughness (reported in Ref. 9). We found that the mean fracture toughness of Oc^−/− mice was significantly lower than WTs (4.75 MPa m^0.5 ± 0.38 vs 5.80 MPa m^0.5 ± 0.82, p < 0.01). Next, to determine the contribution of OC to fracture toughening mechanisms in bone, the fracture surfaces of WT and Oc^−/− mouse bones were analyzed using scanning electron microscopy. Here, our analyses, using one-way analysis of variance (ANOVA) with post hoc Student–Newman–Keuls (SNK) test, indicated that the lamellar separation in Oc^−/− mice (2.21 ± 0.34 μm, p < 0.001) was higher than in WT mice (1.54 ± 0.25 μm) [Fig. 3(a)]. Bridges that span lamellae are crucial in maintaining bone matrix integrity. Thickness measurements of the bridges showed that WT mouse bone (0.51 ± 0.20 μm) had thicker bridges compared to Oc^−/− bones (0.28 ± 0.11 μm, p < 0.001) (Fig. 4). The bridges in the WT specimens were also rectangular edged [Fig. 4(a)] as opposed to conical shaped bridges in the Oc^−/− specimens [Fig. 4(b)]. In addition, the non-fibrillar bone matrix from WT mice showed the presence of intricate void networks (indicative of dilatational band formation—a nanoscale energy dissipation mechanism in bone⁹), which were absent in Oc^−/− mice [Figs. 4(c) and 4(d)]. These findings suggest that decreased crack bridging and increased lamellar spacing in the absence of OC contribute to disrupted energy dissipation mechanisms and lower fracture toughness.

FIG. 3.

The mean and standard deviation is reported based on 15 measurements conducted on randomly selected areas of bones from each group (N = 3 bones/group). Significant differences in the interlamellar spacing (a) and interlamellar bridging thickness (b) were seen between the WT and Oc^−/− bones. Interlamellar spacing (a) increased in the Oc^−/− mice, indicating easier lamellar separation during the fracture process. Bridges spanning cracks and lamellar interfaces (b), were thinner in the knockout specimen. Statistical significance was tested using the t-test; ^*** denotes p < 0.001; SD denotes standard deviation.

FIG. 4.

Fractography images of the fractured surfaces in WT and Oc^−/− mouse bones (n = 8/group). Oc^−/− mouse bone (b) exhibited increased lamellar separation as compared to WT shown between the arrows (a). Bridges (encircled in yellow) in WT mice (c) were rectangular in morphology as opposed to those in Oc^−/− mouse bone that were conical in shape (d). The presence of extensive fibrillar matrix can be seen in the WT mice (c). Interlamellar separation can be considered akin to delamination in laminates (layered composites). Delamination indicates poor bonding between the layers of the composites culminating in failure of the composite. In this case, the lamellae within the Oc^−/− bone, as compared to WT, exhibit greater separation. It is postulated that changes to OC through glycation (or its depletion) alter the bonding between lamellae, leading to the increased separation, as indicated by the SEM images. In turn, these microstructural flaws ease the propagation of cracks in Oc^−/− bones.

D. Determination of advanced glycation end-products (AGEs) related to osteocalcin and the association of glycated osteocalcin with loss of bone toughness

To capture the effects of OC glycation on bone fracture toughness, we performed in vitro glycation of whole femora from WT and Oc^−/− mice, followed by mechanical testing. Given that glycation modifies the entire organic matrix including OC, we compared the differences in AGEs and fracture toughness caused by glycation of the organic matrix with (WT-glycated minus WT-control, WT-delta) and without OC (Oc^−/−-glycated minus Oc^−/−-control, Oc^−/−-delta). Fracture toughness values of glycated samples are normalized to their respective non-glycated controls and reported as percent change in fracture toughness. Glycation resulted in higher total fluorescent AGEs formation, primarily PEN over other AGEs,³⁰ in both genotypes compared to the respective non-glycated controls (WT-glycated vs WT-non-glycated p ≤ 0.001, Oc^−/−-glycated; vs Oc^−/−-non-glycated p < 0.001). More importantly, the increase in fluorescent AGEs, attributable predominantly to pentosidine, was significantly different between the groups where WT-delta was greater than Oc^−/−-delta (p = 0.016) (Fig. 5).

FIG. 5.

Mean total fluorescent AGEs (left) and change in AGEs (right) due to in vitro glycation for WT and Oc^−/− groups. ^* indicates significance at p < 0.05. Eighteen six-month-old male C57Bl/6 wild-type (WT) and Oc^−/− mice were used in this study (n = 9 animals per group) where one femur from each group was in vitro glycated and other treated as control (see Sec. IV for details on treatment).

Glycation caused lower fracture toughness in both genotypes (WT and Oc^−/− groups) when compared to their respective non-glycated controls (WT-glycated vs WT-non-glycated p = 0.004, Oc^−/−-glycated; vs Oc^−/−-non-glycated p = 0.002; Fig. 6). Glycation of wild-type bones (WT-glycated) reduced bone toughness to similar levels of bones depleted of osteocalcin (Oc^−/−-glycated and Oc^−/−-non-glycated). More importantly, the loss in toughness was significantly different between the groups where WT-delta was greater than Oc^−/−-delta (p = 0.027) (Fig. 6).

FIG. 6.

Mean propagation toughness (left) and percent change in propagation toughness (right) due to in vitro glycation for WT and Oc^−/− groups. * indicates significance at p < 0.05. Eighteen six-month-old male C57Bl/6 wild-type (WT) and Oc^−/− mice were used in this study (n = 9 animals per group) where one femur from each group was in vitro glycated and other treated as control (See Sec. IV for details on treatment).

III. DISCUSSION

Osteocalcin, the most abundant non-collagenous bone matrix protein,⁴⁴ plays an essential role in directing bone remodeling, mineralization, cell–matrix interactions, and resistance to fracture. In this study, we demonstrated that the level of osteocalcin declines with age in human bone, and this decline is associated with disrupted energy dissipation mechanisms. Using WT and OC-deficient mouse models, for the first time, we identified the formation of pentosidine as one of the glycation-driven post-translational modifications of OC. Of importance to this study is that pentosidine is a well-established fluorescent cross-link known to predict fracture risk in bone.^45–48 Consistent with the formation of pentosidine, the accumulation of fluorescent AGEs was also significantly greater in mouse bone matrix with OC (also known as WT) and showed greater loss of toughness than Oc^−/−.

Osteocalcin is a useful clinical marker of bone remodeling, and the level of OC in combination with bone mineral density (BMD) can be used to assess osteoporosis and predict fracture risk. For example, serum OC in postmenopausal women is inversely correlated with BMD^18,49 and is higher in patients with bone fractures compared to non-fracture controls.^50,51 A reduction in OC content was also observed in cortical bone from patients with femoral neck fractures compared to patients with osteoarthritis.⁵² Despite being a promising clinical marker, currently no routine OC tests are performed for patients.

OC plays a role at the organic–mineral interface of bone by providing energy dissipative properties, thereby increasing its fracture resistance.^6,9,53 Mice lacking OC have reduced propagation toughness⁹ and increased bone hardness.²⁵ Glycation of OC may produce additional crosslinks, which can result in altered mechanical properties.

There are very few studies that provide a direct link between glycated OC and bone fracture. Glycated OC has been shown to increase linearly with age³⁹ and aging is known to be associated with increased fragility fractures.^54–56 Glycation of total bone matrix proteins (including OC) measured as total fluorescent AGEs has been shown to reduce bone's mechanical competence and certain disease pathology also accelerate the accumulation of AGEs.^47,57,58 For example, serum OC levels were lower in postmenopausal women with type 2 diabetes⁵⁹ compared to postmenopausal women without diabetes.⁵⁹ The serum OC from these patients was negatively correlated with glycated hemoglobin,^59,60 indicating that poor glycemic control is related to impaired bone formation. It also suggests that hyperglycemia could modify levels of bone matrix proteins such as OC and contribute to increased fracture risk observed in individuals with T2D.^61–63

Consistent with the above notion, in our study, the increase in total fluorescent AGEs was indeed associated with a ∼20% reduction in propagation toughness for glycated WT bones while a ∼6% reduction was observed for glycated Oc^−/− bones when compared to their respective non-glycated controls. The change in propagation toughness due to glycation was greater in WT than Oc^−/− bone samples. Therefore, our results indicate that glycated OC (present in WT bone) contributes to a greater loss of bone toughness. Our findings are consistent with the literature data^57,64,65 because like others, we also observed a significant reduction in fracture toughness and higher AGE content with glycation.

The hallmark of osteocalcin's structure is the presence of three γ-carboxyglutamic acid residues which are involved in the binding of the protein to hydroxyapatite.^14,66 The discovery of pentosidine in mouse osteocalcin presented in this study led us to the reanalysis of its amino acid sequence. Consequently, it is likely that in addition to two arginine residues, mature mouse osteocalcin contains two lysine residues as pentosidine is known to form between these two amino acids.³⁰ We further posit that one arginine residue is located on α1-helix and the second arginine is on α3-helix of mature mouse OC. Both lysine residues are located on α3-helix.⁶⁷ Furthermore, the analysis of the mature mouse OC (amino acids 13–49) suggests that the formation of pentosidine most likely occurs between arginine-20 of α1-helix and lysine-38 of α3-helix (this is because formation of pentosidine takes place between glucose, arginine, and lysine both in vivo and in vitro).³⁰ The size of glucose molecule, as measured by the length of the open-chain form, is approximately 9 Å (0.9 nm).⁶⁸ Thus, for pentosidine to be formed, a distance of 7.5 Å or less between the target arginine residue and the ε-amino group of lysine is considered to be sufficient.⁶⁹ These structural requirements are fulfilled by the aforementioned arginine-20 of α1-helix and lysine-38 of α3-helix of mouse OC. Based on the above layout, it is likely that an alteration to OC structure or its interaction to mineral via the GLA residues will disrupt the energy dissipation mechanisms in bone as was found in our study.

The GLA residues on OC are located at amino acid positions 17, 21, and 24 and a disulfide bond, between the cysteine residues at positions 23 and 29.⁷⁰ The intracellular carboxylation of GLA residues (in osteoblasts) is mediated by vitamin K.⁷¹ Undercarboxylated OC is released into the circulation while carboxylated OC exhibits a strong affinity to bone mineral HA.⁷¹ Upon secretion into bone matrix, carboxylated OC undergoes a conformational change that allows its GLA residues to align with the Ca²⁺ ions on HA. Thus, all bonds, which osteocalcin forms (3 in number),¹⁴ are with HA mineral. Based on the values of bond constants reported in the literature, a total energy of 2.1 eV/molecule is dissipated per OC molecule.^72,73 An experimentally obtained value of 1 eV has been reported by Gupta et al.⁷⁴ Furthermore, these estimates consider only Ca²⁺ interactions. If other interactions including the disulfide and hydrogen-bonds are considered, higher amounts of energy may indeed be dissipated per molecule. Each α-helix contains four H-bonds.¹⁵ Based on the molecular structures, non-glycated OC contains three helices totaling six turns (24 H-bonds).⁷⁵ With 0.125 eV per H-bond we have a yield of 3 eV/OC.^75,76 The disulfide bond in OC also contributes ∼0.6 eV/OC (60 kJ/mol).¹⁵ Therefore, OC is likely to dissipate at total of 5.7 eV/molecule.¹⁴ A recent molecular dynamics simulation of glycated and non-glycated OC suggests energy dissipation of ∼2.6 eV per OC (non-glycated).⁴² This value, however, is expected to be substantially reduced, once OC has been glycated.

In our study, we show that the presence of PEN cross-linking discovered on OC supports the statement that glycation leads to changes in OC's tertiary structure. A recent in silico study indicated that glycation disrupted the α-helical structure of OC, which prevented the binding of GLA residues on OC to Ca²⁺ ions on HA mineral,⁴² and thus, this work further confirms our observation stated above. Specifically, we deducted the contribution of 2.1 eV, leaving us with 3.6 eV of energy dissipation in glycated OC. Still, one should note that this is a very conservative assumption and if disruption of helices also affects internal H-bonds, then we can deduct a further 3 eV/molecule. Therefore, we can consider 37%–90% loss in energy dissipation capacity of OC due to glycation (Fig. 7).

FIG. 7.

Schematic showing a potential effect of glycation on osteocalcin energy dissipation. In a healthy, control bone, osteocalcin (OC) binds to the [100] face of hydroxyapatite with the three Gla residues (positions 17, 21, and 24 in OC amino acid sequence) on its alpha‐1 helix and is able to stretch during loading. Upon glycation, the formation of the cross-linking AGE. Pentosidine, between lysine 38 and arginine 20, impairs OC's stretching ability, thereby reducing energy dissipation. The model of the OC structure was prepared using mouse OC PB 86546 and the SWISS-MODEL Interactive Workspace (https://swissmodel.expasy.org.). Schematic was assembled in BioRender 2022.

To corroborate our molecular deformation and fracture mechanics model with our experimental data on the fracture toughness of glycated and control WT and Oc^−/− mouse bones (Table I), we make the following inferences. The difference in mean values of toughness between WT and Oc^−/− control groups (0.97 MPa m^0.5 or a ∼20% loss in toughness due to depletion of OC) can be construed as the contribution of OC to bone matrix fracture toughness.⁹ Similarly, the difference in mean values of toughness between WT glycated group and Oc^−/− glycated group (0.33 MPa m^0.5 or ∼9%) can be construed as the contribution of glycated OC to bone matrix. On comparing the values noted above, one can infer that with glycation, there is a 66% reduction in OC's contribution to energy dissipation. This number coincides well with an estimate of 37%–90% obtained in our physical models that take into consideration various energy dissipation mechanisms affected via the glycation of OC.

TABLE I.

Glycation reduces OC's contribution to bone toughness from 0.97 to 0.33 MPa m^0.5, or a reduction of 66%, as per experimental data.

Propagation toughness (MPa m0.5)	Non-glycated	Glycated
WT	4.55	3.68
Oc−/−	3.58	3.35
WT toughness—Oc−/− toughness	0.97 (contribution of OC to bone toughness)	0.33 (contribution of glycated OC to bone toughness)

Our study is not without limitations. Here, we show that OC content declines with age in human cortical bone, and we have previously mechanically tested a subset of these samples and show that fracture toughness of human bone also declines with age.³⁷ However, we lack statistical power to determine the association between OC content and fracture toughness in human samples with aging. Thus, as an alternate OC/age experiment, we compared bones with (WT) and without OC (Oc^−/−) in bone matrix. These studies allow us to demonstrate the principal effects of loss of OC and its modification by total fluorescent AGEs (such as PEN) on bone toughening which have not been previously shown. In addition, here we propose a model that considers various mechanisms, including the results of our in vitro whole bone glycated studies, to obtain a closer approximation of the contribution of glycated OC to bone fracture.

In summary, this study shows for the first time that the accumulation of advanced glycation end-products was significantly greater in mouse bone matrix with OC. Furthermore, pentosidine, a specific and clinically relevant AGE linked to increased skeletal fragility, was discovered in mouse osteocalcin. Loss of osteocalcin disrupted bone toughening mechanisms through increased interlamellar separation and decreased interlamellar bridge thickness. Using a molecular deformation and fracture mechanics model which considers multiple potential deformation mechanisms, we propose that glycated OC leads to a 66% loss in bone toughness. We anticipate that our study will aid in elucidating the effects of major bone matrix protein modifications on bone fragility and help identify potential therapeutic targets for maintaining skeletal health.

IV. METHODS

A. Assessment of OC and its modifications in bone

1. Bone samples

Cortical bone from the posterior quadrant of fourteen male human cadaveric tibiae (ages 19–85) were obtained from the centralized National Disease Research Interchange (NDRI) biobank. The specimens were known to be free of osteoarthritis, diabetes, and other metabolic bone diseases as well as human immunodeficiency virus (HIV) and hepatitis B virus (HBV). Tibiae were also harvested from C57BL/6 wild-type and Oc^−/− mice (n = 3/group). Bone samples were cleaned of soft tissue and the ends removed using a slow-speed diamond blade (Buehler, Lake Bluff, IL). All bone pieces were repeatedly washed in cold distilled water until the washings were free of blood and then defatted using isopropyl ether. After freeze-drying, the specimens were stored at −80 °C until their use.

2. Protein isolation

Bone pieces (approximately 0.050–0.055 g in size) were washed in cold distilled water, defatted using isopropyl ether, immersed in liquid nitrogen, and grounded into smaller particles (40–60 μm). Bone powders were transferred into eppendorf tubes that had a hole melted in the lid and then extraction buffer (0.05 M EDTA, 4 M guanidine chloride, 30 mM Tris-HCl, 1 mg/ml bovine serum albumin (BSA), 10 μl/ml Halt Protease Inhibitor, pH 7.4) was added. The tube was covered with a dialysis membrane (Spectra Por^® 3 Dialysis Membrane; Spectrum Laboratories, Inc., CA, USA) and closed allowing the membrane to line the hole. The micro-dialysis was conducted at 4 °C against several changes of the PBS buffer pH 7.4. After measurement of protein concentration, supernatants were either used directly for 2D SDS-PAGE [i.e., supernatants contained an equal (∼200–250 ng) total-protein contents] or freeze-dried and stored at −80 °C until use.

3. Quantitation of osteocalcin using ELISA

Protein concentration was measured using a Bradford assay according to the manufacturer's manual [Pierce™ Coomassie Plus™ (Bradford) Protein Assay; Thermo Fisher Scientific, Waltham, MA]. A bovine serum albumin (BSA) included with the kit was used to make a standard curve. The hOST-EASIA (kit # IB79146, Germany distributed by ARP Inc., Belmont, MA) was used to quantify the levels of osteocalcin (OC) in protein extracts according to the manual included with the kit. The hOST-EASIA is a solid phase enzyme amplified sensitivity immunoassay performed using microtiterplates (MT-plates). The assay uses monoclonal antibodies directed against distinct epitopes of human osteocalcin. Calibrators and samples react with the OC capture monoclonal antibody (primary antibody) coating wells of MT-plate as well as with a monoclonal antibody labeled with an enzyme, horseradish peroxidase (HRP; secondary antibody). HRP performs a chromogenic reaction on a substrate 3,3′,5,5′-tetramethylbenzidine (TMB). The amount of TMB oxidized by HRP is proportional to the osteocalcin concentration and is measured spectrophotometrically.

4. Separation of bone-matrix proteins using 2D SDS-PAGE

The 2D SDS-PAGE is based on standard (8 × 8 cm²) polyacrylamide gels (thickness of 1 mm or less) and small-size apparatuses.⁷⁷ Separation of protein extracts was conducted using the NuPAGE gels (NuPAGE Technical Guide IM-1001, version E; Invitrogen, Carlsbad, CA). Gel staining and processing of protein spots after their cutting from the gel were performed according to the protocols included with the “SilverQuest” silver staining kit compatible with mass spectrometry (Invitrogen Life Sciences, Carlsbad, CA). Briefly, to extract osteocalcin and the corresponding control area from the gel, the selected spots were cut out from the gel and placed into a 1.5 ml sterile microcentrifuge tube. Next, 50 μl of Destainer A and 50 μl of Destainer B were added to each microcentrifuge tube and incubated for 15 min at room temperature. The solutions were transferred into new tubes. The extraction procedure was repeated two times by adding 200 μl of ultrapure water. The extracts were centrifuged at 10 000 × g for 10 min, the supernatants collected into new microcentrifuge tubes, concentrated by overnight lyophilization, and used for acid hydrolysis.

5. Ultrahigh-Pressure Liquid Chromatography (UPLC) for pentosidine (PEN) identification in mouse osteocalcin

Acid hydrolysis of mouse OC samples was performed in 6 N HCl (100 μl/OC present in a gel spot) at 110 °C for 16 h. After cooling, the hydrolysates were centrifuged, and the respective supernatants were divided into portions, transferred to clean tubes, concentrated/dried, and stored at −80 °C until use. Before the UPLC analysis, each hydrolysate was dissolved in 1% n-heptafluorobutyric acid (HFBA). The hydrolysates were analyzed using an Acquity UPLC machine (Waters Corp., Milford, MA, USA) equipped with the reverse-phase Acquity UPLC HSS T3 column (1.8 μm; 2.1 × 100 mm). The column flow rate and temperature were 0.667 ml/min and 40 °C, respectively. The solvents used for the UPLC analysis of PEN comprised: (1) solvent A containing 0.12% hepatafluorobutyric acid (HBFA) in 18 ohms pure water, and (2) solvent B composed of 50: 50 (v: v) mixture of solvent A: acetonitrile. Prior the use, the column was equilibrated using 10% solvent B. Gradient of 10 to 31% of solvent B was used for the separation of the collagen crosslinks. The elution of PEN was monitored for fluorescence emission at 385 nm after excitation at 335 nm.

6. Chemicals and reagents

All reagents used for UPLC separations were of the UPLC grade. The chemicals used for collagen isolation and SDS-PAGE were molecular biology grade. Acetonitrile and acetic acid were purchased from Fisher Scientific (Morris Plains, NJ, USA). Heptafluorobutyric acid was purchased from Sigma (St. Louis, MO, USA). The PEN standard was obtained from Dr. Vincent Monnier (CWRU, Cleveland, OH, USA). The Halt Protease Inhibitor was purchased from Pierce/Thermo Scientific, Rockford, IL, USA.

B. Fractography of non-glycated Oc^−/− and WT bones

Femora from six-month-old Oc^−/− and WT mice (n = 8 per group) were notched in the anterior mid-diaphyseal region using a slow speed diamond blade (Buehler) and soaked in saline for an hour prior to testing. The bending tests were done to failure in the displacement feedback mode (Elf 3200, EnduraTEC) at crosshead rate of 0.001 mm/s. Propagation toughness was computed from test data. After the toughness tests, fracture surfaces of knockout and wild-type bones were investigated using scanning electron microscopy on a Carl Zeiss Supra SEM (Carl Zeiss Microscopy, Thornwood, USA) equipped with InLens SE detector, Everhart-Thornley SE detector and a Robinson BSE detector, with voltage and current limits of 30 kV and 20 nA, respectively. Secondary electron images were acquired at voltages between 2.5–5 kV at a working distance of 10 mm. Samples were coated with a thin platinum layer prior to imaging to avoid surface charging and related artifacts. First, the notch was identified at low magnification. Areas adjacent to the notch (i.e., regions of stable crack growth) were then imaged at higher magnifications to study the fracture surface in detail, both quantitatively and qualitatively. Interlamellar distance and bridging thickness were quantified using ImageJ software. Toughening mechanisms, including micro-cracking, ligament bridging, delamination, and nanoscale void formation were qualitatively assessed.

C. In vitro glycation and fracture toughness testing of Oc^−/− and WT bones

1. Sample Preparation and Bone Geometry

Eighteen six-month-old male C57Bl/6 wild-type (WT) and Oc^−/−mice were used in this study (n = 9 animals per group). The femoral head and condyle of each sample were cut off (to allow uniform glycation of cortical bone), and a notch with precrack induced by razor blade was created on the anterior mid-diaphysis region of all samples using a slow speed diamond blade saw (IsoMet Low Speed Saw, Buehler) to mimic natural loading conditions by three-point bending tests. The notch with precrack represents a preexisting crack that will initiate and propagate into a large-scale fracture.⁷⁸ This technique more accurately assesses the fracture resistance of bone matrix. The samples were then rinsed with 10× phosphate-buffered saline and stored in physiological saline-soaked gauzed at −80 °C until use.

The bones were scanned in Eppendorf tubes with saline using micro-computed tomography (μCT) at high resolution 10.5 μm voxel size, 70 kVp, 114 mA, and 200 ms integration time (vivaCT 40, Scanco Medical AG, Bassersdorf, Switzerland). Cortical thickness, and internal and external diameters were determined at approximately 0.40 mm below the notch using a 2D slice of the cross section, and the notch angle is measured from the corresponding 3D image (ImageJ).

2. In vitro glycation

One femur from each animal was randomly selected for in vitro glycation while the contralateral limb served as the non-glycated control. The glycation solution consisting of 25 mM ε-amino-n-caproic acid, 30 mM hepes, 5 mM benzamidine, 10 mM N-ethylmaleimide, and 0.6 M ribose solubilized in hanks buffer.⁵⁷ All non-glycated samples were placed in a similar buffer solution without added sugar. The ribosylation protocol was carried out for 10 days with pH adjustments using reagents described above (pH 7.2–7.6), and then the samples were rinsed with nanopure water to remove trace reagents.

D. Mechanical testing

All samples stored at −80 °C and hydrated in saline prior to testing and were loaded in three-point bending (Fig. 8) in wet conditions until failure at a loading rate of 0.001 mm/s (Elf Enduratec 3200). The distance between the lower supports was 7 mm, and the resulting load displacement curve was used to calculate a single-value fracture toughness K_c at maximum load (propagation toughness) for each sample.⁷⁸ Propagation toughness more comprehensively captures the fracture behavior of bone than initiation toughness. Toughness measured here is dependent on the material thus reflecting the changes due to glycation.

FIG. 8.

Schematic of three-point bending setup of mouse bone. The femoral head and condyle were removed for uniform glycation of cortical bone. The bone sits on two supports of span length “S” with the notch created in the mid-diaphysis directly below the loading point. Red arrows indicate that the anterior surface is loading in tension while the posterior surface is in compression. This razor thin micrometer size notching technique produces a consistently sharp notch with a root radius of ∼10 μm representing sharp crack for biological materials, such as bone; this method is at the limit of validity according to ASTM Standards.

1. Measurement of total fluorescent advanced glycation end-products (AGEs)

After mechanical testing, a piece of bone from each sample (∼13–15 mg) was defatted using isopropyl ether and lyophilized (freeze-dried) overnight. Mouse bones were then hydrolyzed overnight in 6 N hydrochloric acid (HCl). The fluorescence of the hydrolysates and quinine sulfate standard (stock: 10 μg/ml quinine per 0.1 N sulfuric acid) was measured at 360 nm/460 nm excitation/emission using a microplate reader (Infinite 200, Tecan). The collagen content in each sample was calculated using a hydroxyproline standard of increasing concentration (stock: 2000 μg/ml L-hydroxyproline per 0.001 N HCl). Total fluorescent AGEs were calculated as the amount of quinine per unit of collagen.⁴⁶

2. Statistical analysis

Following the normality tests, paired samples t-test was used to determine differences in fracture toughness and AGEs between the groups (WT-glycated vs WT-non-glycated; Oc^−/−-glycated; vs Oc^−/−-non-glycated). Because NEG modifies the organic matrix including type-I collagen and OC, we compared the change in AGEs and propagation toughness caused by glycation of the organic matrix with (WT-glycated minus WT-non-glycated) and without osteocalcin (Oc^−/−-glycated minus Oc^−/− non-glycated) by independent samples t-test. All analyses were conducted using IBM SPSS 2.

AUTHOR DECLARATIONS

Conflict of Interest

The authors have no conflicts to disclose.

Author Contributions

Stacyann Bailey, Atharva A. Poundarik, and Grazyna E. Sroga contributed equally to this work.

Stacyann Bailey: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Atharva A. Poundarik: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Grazyna E. Sroga: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Deepak Vashishth: Conceptualization (lead); Funding acquisition (lead); Investigation (supporting); Project administration (lead); Resources (lead); Supervision (lead); Writing – review & editing (equal).

DATA AVAILABILITY

The data that support the findings of this study are available from the corresponding author upon reasonable request.