Scholarly article on topic 'Effects of 18-month treatment with bazedoxifene on enzymatic immature and mature cross-links and non-enzymatic advanced glycation end products, mineralization, and trabecular microarchitecture of vertebra in ovariectomized monkeys'

Effects of 18-month treatment with bazedoxifene on enzymatic immature and mature cross-links and non-enzymatic advanced glycation end products, mineralization, and trabecular microarchitecture of vertebra in ovariectomized monkeys Academic research paper on "Clinical medicine"

CC BY-NC-ND
0
0
Share paper
Academic journal
Bone
OECD Field of science
Keywords
{Bazedoxifene / "Collagen cross-links" / "Cancellous bone" / "OVX monkeys" / "Advanced glycation end products" / Pentosidine}

Abstract of research paper on Clinical medicine, author of scientific article — Mitsuru Saito, Yoshikuni Kida, Tetsuro Nishizawa, Shotaro Arakawa, Hinako Okabe, et al.

Abstract Bazedoxifene (BZA) is used for the treatment of post-menopausal osteoporosis. To elucidate changes in collagen, mineralization, and structural properties and their relationship to bone strength after treatment with BZA in ovariectomized (OVX) monkeys, the levels of collagen and enzymatic immature, mature, and non-enzymatic cross-links were simultaneously examined, as well as trabecular architecture and mineralization of vertebrae. Adult female cynomolgus monkeys were divided into 4 groups (n=18 each) as follows: Sham group, OVX group, and OVX monkeys given either 0.2 or 0.5mg/kg BZA for 18months. Collagen concentration, enzymatic and non-enzymatic pentosidine cross-links, whole fluorescent advanced glycation end products (AGEs), trabecular architecture, mineralization, and cancellous bone strength of vertebrae were analyzed. The levels of enzymatic immature and mature cross-links, bone volume (BV/TV), and trabecular thickness (Tb.Th) in BZA-treated groups were significantly higher than those in the OVX control group. In contrast, the trabecular bone pattern factor (TBPf), the structure model index (SMI), the enzymatic cross-link ratio, and the levels of pentosidine and whole AGEs in BZA-treated groups were significantly lower than those in the OVX control group. Stepwise logistic regression analysis revealed that BV/TV, Tb.Th, TbPf, and pentosidine or whole AGEs independently affected ultimate load (model R2 =0.748, p<0.001) and breaking energy (model R2 =0.702). Stiffness was affected by Tb.Th, enzymatic immature cross-link levels and their ratio (model R2 =0.400). Treatment with BZA prevented OVX-induced deterioration in the total levels of immature enzymatic cross-links and AGEs accumulation and structural properties such as BV/TV, Tb.Th, and TbPf, which contribute significantly to vertebral cancellous bone strength.

Academic research paper on topic "Effects of 18-month treatment with bazedoxifene on enzymatic immature and mature cross-links and non-enzymatic advanced glycation end products, mineralization, and trabecular microarchitecture of vertebra in ovariectomized monkeys"

Bone

ELSEVIER

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/bone

Original Full Length Article

Effects of 18-month treatment with bazedoxifene on enzymatic immature and mature cross-links and non-enzymatic advanced glycation end products, mineralization, and trabecular microarchitecture of vertebra in ovariectomized monkeys^

Mitsuru Saito a,*< Yoshikuni Kida a, Tetsuro Nishizawa a, Shotaro Arakawa a, Hinako Okabe a, Azusa Sekib, Keishi Marumo a

a Department of Orthopaedic Surgery, Jikei University School of Medicine, Japan b Tsukuba Research Center, HAMRICo., Ltd., Ibaraki, Japan

ARTICLE INFO

ABSTRACT

Article history: Received 29 April 2015 Revised 26 August 2015 Accepted 14 September 2015 Available online 16 September 2015

Keywords: Bazedoxifene Collagen cross-links Cancellous bone OVX monkeys

Advanced glycation end products Pentosidine

Bazedoxifene (BZA) is used for the treatment of post-menopausal osteoporosis. To elucidate changes in collagen, mineralization, and structural properties and their relationship to bone strength after treatment with BZA in ovariectomized (OVX) monkeys, the levels of collagen and enzymatic immature, mature, and non-enzymatic cross-links were simultaneously examined, as well as trabecular architecture and mineralization of vertebrae. Adult female cynomolgus monkeys were divided into 4 groups (n = 18 each) as follows: Sham group, OVX group, and OVX monkeys given either 0.2 or 0.5 mg/kg BZA for 18 months. Collagen concentration, enzymatic and non-enzymatic pentosidine cross-links, whole fluorescent advanced glycation end products (AGEs), trabecular architecture, mineralization, and cancellous bone strength of vertebrae were analyzed. The levels of enzymatic immature and mature cross-links, bone volume (BV/TV), and trabecular thickness (Tb.Th) in BZA-treated groups were significantly higher than those in the OVX control group. In contrast, the trabecular bone pattern factor (TBPf), the structure model index (SMI), the enzymatic cross-link ratio, and the levels of pentosidine and whole AGEs in BZA-treated groups were significantly lower than those in the OVX control group. Stepwise logistic regression analysis revealed that BV/TV, Tb.Th, TbPf, and pentosidine or whole AGEs independently affected ultimate load (model R2 = 0.748, p < 0.001) and breaking energy (model R2 = 0.702). Stiffness was affected by Tb.Th, enzymatic immature cross-link levels and their ratio (model R2 = 0.400). Treatment with BZA prevented OVX-induced deterioration in the total levels of immature enzymatic cross-links and AGEs accumulation and structural properties such as BV/TV, Tb.Th, and TbPf, which contribute significantly to vertebral cancel-lous bone strength.

© 2015 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Bazedoxifene (BZA) is a novel indole-based third-generation selective estrogen receptor modulator (SERM) with unique structural characteristics compared with other SERMs (e.g. raloxifene and tamoxifen) [1,2]. BZA decreased the vertebral fracture risk in osteoporosis with a smaller increase in bone mineral density (BMD) [3,4,5]. Because BMD and bone qualities are important determinants of bone strength [6],

☆ Conflict of interest: Mitsuru Saito received research grants and/or consulting or speaking fees from Pfizer Inc., Eli Lilly, Chugai, Dai-ichi Sankyo, Asahikasei Pharma, Astellas Pharma, Taisho Toyama Pharma, Teijin Pharma, and Ono Pharma. Yoshikuni Kida, Tetsuro Nishizawa, Shotaro Arakawa, Hinako Okabe, Azusa Seki, and Keishi Marumo declare that they have no conflict of interest.

* Corresponding author at:Department of Orthopaedic SurgeryJikei University School of Medicine,3-25-8, Nishi-Shinbashi,Minato-ku,Tokyo 105-8461 Japan.

BZA may show favorable effects on bone quality such as collagen cross-link formation, mineralization, and microarchitecture. However, little is known about the effects of BZA on bone quality and the degree to which changes in bone qualitative properties induced by BZA treatment contribute to bone strength.

The determinants of bone material properties are, among others, the degree of mineralization, the collagen concentration, and the amount of collagen cross-linking, including enzymatic divalent immature, triva-lent mature, and non-enzymatic senescent types of cross-linking such as advanced glycation end products (AGEs). Furthermore, the relative ratio of hydroxylysine to lysine-derived enzymatic cross-links, as well as the levels of cross-links, independently affects bone strength via the regulation of collagen fibrillogenesis [7-13].

Stabilization of newly formed collagen fibers is initially achieved by the formation of covalent cross-links between neighboring collagen

http://dx.doi.org/10.1016/j.bone.2015.09.006

8756-3282/© 2015 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

molecules. Collagen cross-links can be divided into lysyl oxidase and lysine hydroxylase mediated enzymatic cross-links and glycation- or oxidation-induced non-enzymatic cross-links, AGEs. Enzymatic crosslinks are first formed as divalent immature cross-links such as dihydroxylysinonorleucine (DHLNL), hydroxylysinonorleucine (HLNL), and lysinonorleucine (LNL) between the non-helical domains and the helical domain of adjacent collagen molecules [7-13]. A portion of divalent immature cross-links spontaneously react to form trivalent mature cross-links such as pyridinoline, deoxypyridinoline, pyrrole, and deoxy-pyrrole. Enzymatic cross-link formation has positive effects on bone mechanical properties within a beneficial range without brittleness [11-16]. In contrast, AGE cross-links lead to a 'stiffer' (not brittle) collagen network, which leads to alterations in microdamage formation and morphology and increased bone fragility [14,15,18-23]. AGE cross-links are formed by oxidation or glycation reactions in a time-dependent manner, which is regulated by tissue turnover rate [20], the degree of oxidative stress [19,23,24], or glycation level [15,17,18,23,24]. Pentosidine is a well-established intermolecular cross-linking AGE [8, 9,12-15,17-20,23-25] and is used as a surrogate marker of whole AGEs [8,9,23,26]. In this study, we determined the levels ofpentosidine and total fluorescent AGEs (tfAGEs).

The purpose of this study was to clarify the long-term effect of BZA treatment on trabecular structural properties, degree of mineralization, and the levels of collagen enzymatic immature and mature cross-links, pentosidine, and total AGEs in OVX monkeys. The significant contributing factors to vertebral cancellous bone strength after treatment with BZA were analyzed.

2. Materials and methods

2.1. Animals and experimental design

All animal procedures were as previously reported by Smith SY et al. [27]. In this study, we used and evaluated bone strength, bone material, and structural properties of the 12th thoracic vertebra (Th12) in the previously reported bone turnover and BMD study [27]. The original study [27] was sponsored by Wyeth Research, which was acquired by Pfizer Inc. in October, 2009. The local affiliate, Pfizer Japan Inc., provided frozen bones (T12) as well as funding for micro-CT analysis in the current study.

Briefly, skeletally mature female colony-bred cynomolgus monkeys of Chinese origin (The Research Center of Primate Laboratory Animals, Guangxi Forestry Province, Nanning, Guangxi, China), between 10 and 13 years of age and weighing 2.6-5.8 kg, were socially housed, generally in pairs, with environmental conditions targeted at 24 °C (±3 °C), 50% (±20%) relative humidity, and under controlled 12 h light/dark cycles each day. Approximately 100 g of standard certified pelleted commercial primate food (PMI Certified Laboratory Fiber-Plus® Monkey Diet Jumbo 5050; PMI Nutrition International, St. Louis, MO, USA) was provided twice daily to animals in each housing unit, as well as daily fruit and fruit-flavored cookies (PrimaTreats, Bio-Serv, Frenchtown, NJ, USA). Municipal tap water that had been softened, purified by reverse osmosis, and sterilized by ultraviolet light was provided ad libitum. This study was conducted in compliance with the Good Laboratory Practice Regulations and with US and Canadian Council on Animal Care (CCAC) regulations for animal care and use.

Animals were acclimated for at least 6 weeks at the study site prior to a 2-month baseline monitoring period. The monkeys underwent surgical procedures (sham or OVX surgery) at the end of the baseline monitoring period. For animals randomly assigned to undergo OVX, both ovaries and surrounding tissues were excised completely. Animals in the sham vehicle control group underwent a sham surgical procedure in which their ovaries were extruded, but remained intact. Daily treatment with study drug or vehicle started the day following the sham operation or OVX and continued for 18 months.

In the original study, OVX monkeys receiving the study drug were randomly assigned into five groups (n = 18 each) and each group was treated with 0.2,0.5,1.0, 5.0 or 25.0 mg/kg/day of BZA (Wyeth Research, Collegeville, PA, USA) by oral gavage for 18 months [27]. We selected the lowest two doses for the current study since they affect bone strength and qualities. Unanesthetized animals were hand-restrained during oral gavage. To acclimate the animals to the daily gavage procedure, approximately 2 weeks before starting treatment all animals received daily oral gavage of deionized water (5 mL/kg) using a rubber catheter attached to a plastic syringe. During active treatment, the daily dose was delivered using the same oral gavage technique in a 1-5 mL/kg suspension at approximately the same time each day. Two additional groups of 18 animals (one sham-operated and one OVX) served as controls and these received a 5 mL/kg solution of vehicle. BZA was diluted with vehicle [1.0% polysorbate 80, NF; 0.5% methylcel-lulose (4000 cps)] in deionized water and the pH was adjusted to 2.9-3.1 with 1.0 N acetic acid as needed.

All experiments were approved by the Experimental Animal Ethics Committee at our institution and were conducted in accordance with guidelines concerning the management and handling of experimental animals.

2.2. Preparation of bone

Following necropsy, the isolated Th12 vertebrae were cleaned of adherent soft tissue, wrapped in saline-soaked gauze, and kept in a sealed plastic bag. The samples were kept at — 20 °C prior to testing.

2.3. Micro-computed tomography measurements of bone trabecular architecture

Cone-beam X-ray micro-computed tomography (micro-CT; MCT-CB100MF, Hitachi Medico Technology, Kashima, Japan) was used to take tomography images of the Th12 vertebral bodies with the following settings: tube voltage, 55 kV; tube current, 0. mA; slice thickness of 32 |jm; and pixel size of 32 |jm2. The specimens were located in the axial direction and cross-sectional images (140 slices, approximately 4.5 mm high) were taken at the central portion of each Th12 vertebral body. Three-dimensional images were reconstructed from these tomo-grams and analyzed three-dimensionally using TRI/3D-BON software (RATOC System Engineering, Tokyo, Japan). The following parameters were obtained: cancellous bone volume (BV/TV, %), trabecular thickness (Tb.Th, |am), trabecular number (Tb.N, 1/mm), trabecular separation (Tb.Sp, |jm), trabecular bone pattern factor (TBPf, 1/mm), and structure model index (SMI) [28].

2.4. Mechanical properties of vertebral cancellous bone

The cranial and caudal end plates of the Th12 vertebra were removed with a bone saw (Labocutter MC-120, Maruto, Tokyo, Japan) to obtain a vertebral body specimen with two parallel surfaces. The specimens were placed on a lower platen with the cranial side facing up and the cancellous region of each specimen was compressed with an upper reduced-platen, 4 mm in diameter, using a material testing machine (MZ-500S, Maruto, Tokyo, Japan) at a constant speed of 10 mm/min over a displacement of 3 mm. Because the diameter of the platen is 4 mm, the platen compressed the area of cancellous bone [13] (Fig. 1).

The size ofthe reduced-platen was determined using cross-sectional images obtained by micro-CT with the diameter as large as possible, but well within the endocortical perimeter. The load and displacement curves were recorded and the following extrinsic parameters were calculated by the testing machine software (CTRwin, System Supply, Nagano, Japan): ultimate load (N), stiffness (N/mm), and breaking energy (mJ).

2.6. Calcium and phosphorus concentration in cancellous bone

Cortical bone

Cancellous bone

Fig. 1. Compression test of vertebral cancellous bone.

2.5. Characterization of enzymatic, non-enzymatic cross-linl<s and AGEs

After mechanical testing, crushed cancellous bone was obtained from the Th12 vertebral body. The cancellous bone samples were cleaned of bone marrow, frozen in liquid nitrogen, and pulverized in liquid nitrogen as previously reported [13,22,29]. Measurement of crosslinks was carried out as previously described [29]. Briefly, bone powder was demineralized with 0.5 M EDTA in 50 mM Tris buffer (pH 7.4) for 96 h at 4 °C. Demineralized bone residues were suspended in potassium phosphate buffer (pH 7.6) and reduced at 37 °C with sodium borohy-dride (NaBH4; Sigma-Aldrich, St. Louis, MO, USA). The specimens were hydrolyzed in 6 N HCl at 110 °C for 24 h. Hydrolysates were analyzed for cross-links and hydroxyproline levels on a Shimadzu LC9 HPLC fitted with a cation exchange column (0.9 x 10 cm, Aa pack-Na; JASCO, Ltd., Tokyo, Japan). It was assumed that collagen weighed 7.5 x the measured weight of hydroxyproline, with a molecular weight of 300,000 Da [29]. The resulting data were used to calculate cross-link values as mol/mol of collagen. We determined the levels of enzymatic immature reducible and mature non-reducible pyridinium cross-links such as pyridinoline (Pyr), deoxypyridinoline (Dpyr), and AGE cross-linking (pentosidine). Reducible immature cross-links (deH-DHLNL, deH-HLNL, and deH-LNL) were identified and quantified according to their reduced forms (DHLNL, HLNL, and LNL, respectively). Immature cross-links and hy-droxyproline were detected with O-phthalaldehyde derivatization, whereas Pyr, Dpyr, and pentosidine were detected by natural fluorescence. Our established HPLC system enabled us to determine enzymatic and non-enzymatic cross-link concentration within a linear range from 0.2 to 600 pmol in bone specimens. Because the degree of lysine hydrox-ylation, which is a precursor of enzymatic cross-links, affects not only the cross-link pattern, but also the biological characteristics of collagen such as the mineralization process and fibrillogenesis [10], the ratio of high to low hydroxylated cross-links was estimated as (DHLNL + HLNL)/LNL or Pyr/Dpyr [10,12,13]. The levels of tfAGEs were determined by the method of Tang et al. [30] and Ural et al. [31]. Briefly, AGEs levels were determined using a fluorescence reader (JASCO FP6200) at wavelengths of 370 nm excitation and 440 nm emission and normalized to a quinine sulfate standard. Pentosidine is a useful surrogate marker of all fluorescent types of crosslinking AGEs in bone. Karim et al. [26] reported a significant relationship between pentosidine concentration and the bulk fluorescence used to detect all AGEs by immunohistochemistry. We also showed a similar relationship between the concentration of total fluorescent AGEs and pentosidine in primate vertebral bone [23]. Crosslinking AGEs such as pentosidine, crossline, and pyrropyridine, have natural fluorescence, whereas non-crosslinking AGEs do not fluoresce except GA pyridine. Because there are no reports regarding the existence of GA pyridine in bone, determining the concentration of total fluorescent AGEs is suitable for estimating accumulation of crosslinking AGEs.

The bone powder of cancellous bone samples was dried in a drying oven at 105 °C for 3 h, and weighed as dry weight. Bone powder was pre-heated at 300 °C for 5 min using an electric heater, heated at 600 °C for 5 h in an electric furnace, and weighed as ash weight. Each ashed bone powder was dissolved in 15 ml of 30% (vol/vol) nitric acid solution and heated at 80 °C for 20 min. 500 ^L of Yttrium (1000 ppm) was added as an internal standard. Calcium and phosphorus concentrations of the bone solutions were measured using an ICP atomic emission spectrometer (ULTIMA-2, Horiba, Tokyo, Japan) [13,25]. The amounts of calcium and phosphorus were expressed as a percent of bone dry weight.

2.7. Statistical analysis

All values are listed as mean ± standard deviation (SD). All groups were compared to each other to confirm statistical significance by ANOVA with a Tukey-Kramer post-hoc test. A Spearman's rank correlation coefficient test was used to compare bone mechanical strength in the OVX and BZA-treated groups with the following variables: bone mass, trabecular architecture, mineral concentration, collagen levels, collagen cross-links, and tfAGEs. Stepwise multiple regression analysis was performed to predict bone strength from the various parameters. All p-values less than 0.05 were considered significant. All analyses were performed using JMP software, Version 10 (JMP Institute Inc., Cary, NC, USA).

3. Results

3.1. Degree of mineralization

Regarding the degree of mineralization, the concentration of calcium (-3.3%, p = 0.010) and phosphorus (-3.1%, p = 0.017) in the OVX group was significantly lower than that of the SHAM group. There was no significant difference between the OVX and BZA groups (Table 1).

3.2. Trabecular architecture

The parameters of trabecular architecture in the OVX group significantly differed from those in the SHAM group (Table 1). The BV/TV (-14%, p = 0.004) and Tb.Th (-11%, p = 0.003) were significantly lower than those in the SHAM group. In contrast, Tb.Sp (21%, p = 0.002), TBPf (p < 0.001), and SMI (45%, p < 0.001) were significantly higher than those in the SHAM group (Table 1).

The 0.2 (BZA0.2) and 0.5 (BZA0.5) mg/kg/day BZA groups showed higher BV/TV (14%, p = 0.040 and 15%, p = 0.039), Tb.Th (10%, p = 0.024 and 6%, p = 0.332) and lower Tb.Sp (14%, p = 0.030 and 18%, p = 0.006), TbPf (p = 0.002 and p = 0.003), and SMI (- 30%, p = 0.002 and 32%, p = 0.002) than the OVX group (Table 1). There was no significant difference in Tb.Th between the OVX and the BZA0.5 groups. Tb.N was not significantly different among the groups.

3.3. Collagen concentration and cross-links in cancellous bone

OVX significantly lowered the concentration of collagen ( — 7%, p = 0.011) and total immature and mature cross-links, and the sum of immature and mature cross-links (—15%, p < 0.001; —18%, p = 0.002; and 15%, p < 0.001, respectively).

Conversely, the ratio of (DHLNL + HLNL)/LNL (200%, p < 0.001), the ratio of Pyr/Dpyr (175%, p < 0.001), pentosidine (155%, p < 0.001), and tfAGEs (116%, p = 0.011) were significantly higher in the OVX group than in the SHAM group (Table 1, Fig. 2). There was no significant difference in the ratio of mature to immature cross-links between the SHAM and the OVX groups (Table 1, Fig. 2).

Table 1

Comparison of collagen content, collagen cross-links, mineral content, and micro-CT among the experimental groups.

SHAM (n = 18) OVX (n = 18) BZA0.2 (n = 18) BZA0.5 (n = 18)

Ultimate load (N) 389.4 ± 86.5 350.9 ± 86.5 421.2 ± 121.6b 434.5 ± 62.8b

Stiffness (N/mm) 203.0 ± 54.3 204.6 ± 42.1 272.1 ± 63.0a,b 264.5 ± 60.8a,b

Breaking energy 761.2 ± 185.8 687.6 ± 122.1 887.7 ± 291.7b 907.0 ± 115.5b

Collagen content (% of tissue weight) 21.1 ± 1.9 19.6 ± 1.9a 21.4 ± 1.8b 21.2 ± 1.4b

Immature cross-links (mol/mol of collagen) 0.908 ± 0.082 0.774 ± 0.080a 1.061 ± 0.123b 1.074 ±0.163b

Mature pyridinium cross-links (mol/mol of collagen) 0.145 ± 0.023 0.118 ± 0.024a 0.155 ± 0.020b 0.146 ± 0.030b

Immature + Mature pyridinium cross-links (mol/mol of collagen) 1.053 ± 0.091 0.891 ± 0.086a 1.216 ±0.130b 1.221 ± 0.150b

Mature/Immature cross-links 0.160 ± 0.026 0.153 ± 0.035 0.139 ± 0.028 0.175 ± 0.043

Pentosidine (mmol/mol of collagen) 1.141 ± 0.201 1.763 ± 0.490a 0.768 ± 0.329b 0.856 ± 0.298b

Total fluorescent AGEs (ng quinine/mg of collagen) 144.8 ± 28.1 168.1 ± 35.4a 117.0 ± 15.7a b 120.7 ± 23.3b

(DHLNL + HLNL)/LNL 51.4 ± 4.6 102.8 ± 26.6a 59.8 ± 7.6b 56.1 ± 15.3b

Pyr/Dpyr 4.10 ± 0.77 7.61 ± 4.20a 3.70 ± 0.55a,b 3.65 ± 0.62a,b

Calcium content (mg/g of tissue dry weight) 233.5 ± 7.6 225.9 ± 8.2a 226.8 ± 8.0a 226.6 ± 10.2a

Phosphorus content (mg/g of tissue dry weight) 103.5 ± 3.4 100.3 ± 4.0a 101.3 ±3.2a 100.2 ±4.7a

BV/TV (%) 34.5 ± 4.8 29.5 ± 5.2a 33.6 ± 5.6b 33.8 ± 4.1b

Trabecular thickness (Tb.Th: |im) 145.1 ± 19.5 129.9 ± 12.6a 143.5 ± 11.6b 137.6 ± 9.8b

Trabecular number (Tb.N: 1/mm) 2.35 ± 0.34 2.20 ± 0.32 2.35 ± 0.35 2.42 ± 0.28

Trabecular separation (Tb.Sp: | m) 271.4 ± 39.0 328.6 ± 65.4a 281.3 ± 59.4b 269.4 ± 37.6b

Trabecular bone pattern factor (TbPf: 1/mm) -0.202 ± 0.747 1.003 ± 0.851a 0.009 ± 0.904a,b 0.023 ± 0.723a,b

Structure model index (SMI) 0.762 ± 0.281 1.102 ± 0.273a 0.771 ± 0.301b 0.752 ± 0.213b

Values are expressed as mean ± standard deviation (SD). Immature cross-links: the sum of DHLNL, HLNL, and LNL. Mature pyridinium cross-links: the sum of Pyr and Dpyr. a p < 0.05 vs. SHAM. b p < 0.05 vs. OVX.

The administration of BZA restored collagen concentration and the cross-link profile to similar levels as the SHAM group (Table 1). The actual levels of enzymatic immature and mature pyridinium cross-links was significantly higher in the BZA0.2 group (immature cross-links: 37%, p < 0.001; mature cross-links: 31%, p < 0.001) and the BZA0.5 group (immature cross-links: 39%, p < 0.001; mature cross-links: 24%, p < 0.001) than in the OVX group (Table 1, Fig. 2). There was no difference in the ratio of mature pyridinium to immature cross-links between OVX- and BZA-treated groups (Table 1). The deterioration of the high to low hydroxylated lysine-derived enzymatic cross-link ratio, such as DHLNL + HLNL/LNL and Pyr/Dpyr, was prevented in the BZA0.2 (- 42%, p < 0.001 and - 51%, p < 0.001, respectively) and BZA0.5 ( — 45%, p < 0.001 and — 52, p < 0.001, respectively) groups and the ratios were at the same level as the SHAM group (Table 1, Fig. 2).

Pentosidine and tfAGEs levels were significantly lower in the BZA0.2 ( — 56%, p < 0.001 and — 30%, p < 0.001, respectively) and BZA0.5 (— 51%, p < 0.001 and — 28%, p < 0.001, respectively) groups than in the OVX group (Table 1, Fig. 2).

3.4. Contributors to the improvement of cancellous bone strength by BZA treatment.

The mean values of ultimate load and breaking energy in the OVX group were lower than those in the SHAM group, although these differences did not reach statistical significance (Table 1 and Fig. 3). These parameters of bone strength were higher in the BZA0.2 (ultimate load: 20%, p = 0.066, stiffness: 33%, p = 0.003, breaking energy: 29%, p = 0.014) and BZA0.5 groups (ultimate load: 24%, p = 0.031, stiffness: 29%, p = 0.013, breaking energy: 32%, p < 0.001) than in the OVX group (Fig. 3). Stiffness was significantly higher in the BZA0.2 (34%, p = 0.006) and BZA0.5 (30%, p = 0.019) groups than in the SHAM group (Table 1, Fig. 3).

BV/TV, Tb.Th, and immature cross-links were significantly and positively correlated with ultimate load, stiffness, and breaking energy by Spearman's rank correlation analysis (Table 2). The mature pyridinium cross-links and total enzymatic cross-links estimated by the sum of immature and mature cross-links showed a significant positive correlation

Fig. 2. Effect of BZA at two doses on the concentration of collagen cross-links in vertebral bone. (a) Enzymatic cross-links: the sum of DHLNL, HLNL, LNL, Pyr, and Dpyr. (b) Non-enzymatic cross-link: pentosidine. (c) Total fluorescent AGEs. Values are the mean ± SD. ap < 0.05: vs. SHAM. bp < 0.05: vs. OVX. BZA0.2 and BZA0.5 means 0.2 and 0.5 mg/kg/day of BZA, respectively.

Fig.3. Effect ofBZA at two doses on Th12 vertebral cancellous bone strength. (A) Ultimate load; (B) Stiffness; (C) Breaking energy. Values are the mean ± SD. ap < 0.05: vs. SHAM. bp < 0.05: vs. OVX BZA0.2 and BZA0.5 means 0.2 and 0.5 mg/kg/day ofBZA, respectively.

to stiffness (Table 2). In contrast, Tb.Sp, TBPf, SMI, pentosidine, tfAGEs, the ratio of (DHLNL + HLNL)/LNL, and Pyr/Dpyr showed a significant negative correlation to bone strength (Table 2). There was no correlation between bone strength and the concentration of collagen, calcium, and phosphorus.

We performed stepwise multiple regression to reveal associations between bone strength and their determinants such as material and structural parameters (Table 3). BV/TV, Tb.Th, TbPf, and the levels of pentosidine associated significantly and independently with ultimate load (model R2 = 0.785). Breaking energy (model R2 = 0.702) was explained by BV/TV, Tb.Th, TbPf, the levels of tfAGEs, and the ratio of (DHLNL + HLNL)/LNL. The levels of enzymatic immature cross-links and the ratio of (DHLNL + HLNL)/LNL were significant determinants of stiffness that were independent of Tb.Th (model R2 = 0.400).

4. Discussion

This study demonstrated that BZA treatment reversed the deleterious changes in collagen concentration, the non-enzymatic AGEs cross-

Table 2

Spearman rank correlation coefficient (r) between the parameters ofbone mass, structural and material determinants and bone biomechanical properties.

Stiffness r Breaking energy r Ultimate load r

Collagen content 0.138 0.100 0.096

Immature cross-links 0.475b 0.314a 0.211a

Mature pyridinium cross-links 0.553a 0.077 0.034

Immature + Mature pyridinium cross-links 0.343a 0.112 0.009

Mature/Immature cross-links 0.131 0.023 0.066

Pentosidine - 0.337a - 0.373b - 0.356a

Total fluorescent AGEs - 0.292a - 0.229a - 0.220a

(DHLNL + HLNL)/LNL - 0.353a - 0.386b - 0.310a

Pyr/Dpyr - 0.313a - 0.202 - 0.191

Calcium content - 0.136 0.032 0.062

BV/TV 0.425b 0.725c 0.829c

Tb.Th 0.353a 0.615c 0.670c

Tb.N 0.252 0.434b 0.520c

Tb.Sp - 0.417b -0.571c - 0.640c

SMI - 0.283a - 0.592c - 0.691c

TbPf - 0.306a - 0.683c - 0.794c

Immature cross-links: the sum of DHLNL, HLNL, and LNL. Mature pyridinium cross-links: the sum of Pyr and Dpyr. a p < 0.05. b p < 0.01. c p < 0.001.

link pentosidine, enzymatic immature and mature cross-links, and trabecular architecture in OVX monkeys.

4.1. Trabecular architecture following BZA treatment

OVX monkeys are commonly used as a non-human primate model of post-menopausal osteoporosis [13,23,27]. In this study, we evaluated Th12 vertebrae from a previously reported bone turnover and BMD study [27]. Smith et al. [27] demonstrated that OVX resulted in significant increases in biochemical markers such as bone-specific alkaline phosphatase (BAP), osteocalcin (OC), C-telopeptide (CTx), and urinary N-telopeptide (NTx) as well as cancellous bone formation and resorption as evaluated by dynamic histomorphometric evaluation of long bone diaphyses and femoral neck cancellous bone. Based on data from bone of the same monkey [27], BZA treatment partially suppressed the OVX-induced 2- to 3-fold increase in bone turnover markers, although BZA had no significant effect on OVX-induced changes in static or dynamic histomorphometric parameters in femoral neck trabecular bone. There is no information regarding the histomorphometric parameters in vertebral cancellous bone in OVX monkeys. The plausible explanation of preservation of trabecular parameters to similar levels as the sham group may be due to suppression of bone turnover although the actual histomorphometric data of vertebral cancellous bone is needed in the future study (Table 1). In terms of degree of mineralization, a further histomorphometric analysis is needed in same bone.

Table 3

Prediction models of mechanical bone strength by structural and material properties using stepwise regression analysis.

3 p value Model r2

Ultimate load

BV/TV 362.7 <0.001

Tb.N -126.5 <0.001

TbPf 98.8 0.071

Pentosidine - 26.1 0.040

Stiffness

Tb.Th 43.2 0.014

Immature cross-links 33.5 0.050

(DHLNL + HLNL)/LNL - 55.5 0.004

Breaking energy

BV/TV 946.1 <0.001

Tb.N - 377.2 <0.001

TbPf 332.9 0.035

Total fluorescent AGEs - 62.3 0.080

(DHLNL + HLNL)/LNL - 174.5 0.009

The results of the structural and material properties are shown in Table 1. Immature cross-links: the sum of the actual amount of DHLNL, HLNL, and LNL.

42. Collagen cross-linking and AGEs by BZA treatment

It is unknown whether the OVX monkeys show a similar deterioration to human post-menopausal osteoporosis regarding the actual levels of immature divalent and mature trivalent enzymatic crosslinks and AGEs cross-linking. We showed that the OVX group exhibited collagen abnormalities similar to those reported in our bone biopsy data from post-menopausal hip fracture cases and OVX rats (Table 1, Fig. 2) [7,8,12,17,25,33,34]. These results are consistent with our previously reported OVX monkeys [13,23]. Ozasa et al. [35] demonstrated in OVX rats that the activity of lysyl oxidase-catalyzed enzymatic cross-link formation decreased by 25% three days after OVX, but activity was completely rescued by estradiol injection. This indicates that estrogen may be a regulatory factor for enzymatic cross-link formation. In this study, estrogen deficiency induced by OVX resulted in a decrease of the total amount of enzymatic cross-link formation. Ireland et al. [36] showed that estrogen stimulates type I collagen synthesis via the induction of estrogen receptors and accelerates mineralization in vitro using human bone-derived osteoblasts. The other candidate of impaired enzymatic cross-link formation by estrogen deficiency is oxidative stress. It is well known that post-menopausal decreases in estrogen increase oxidative stress as well as age-dependent renal dysfunction [36-38]. It is thought that oxidative stress reduces lysyl oxidase activity [39,40]. We showed that increased oxidative stress decreases enzymatic cross-link formation in bone from mice deficient in cytoplasmic copper/zinc superoxide dis-mutase (CuZn-SOD, encoded by the Sod1 gene; Sod1 —/—)[24].Because this result was evident without OVX in the knockout mice, increased ox-idative stress may deteriorate enzymatic cross-linking in bone. These results indicate that impaired enzymatic cross-links in the OVX group might be attributed to estrogen deficiency via the interaction with the estrogen receptor and increased oxidative stress. In this study, there was a significant negative correlation between the concentrations of enzymatic crosslinks and tfAGEs (R2 = 0.313, p < 0.0001). AGEs are thought to be formed by lysine residues, which are essential sites of enzymatic cross-linking in collagen, and result in competitively inhibiting formation of enzymatic cross-links [7,33]. Therefore, the other candidate for impaired enzymatic cross-link formation may be excessive AGEs formation.

In terms of AGEs accumulation in bone, formation of AGEs occurs in a time-dependent manner in physiological conditions. However, AGEs formation is markedly increased during excessively elevated ox-idative stress, even when bone turnover rates and glycemic control are within normal or high ranges [8,19,41]. We demonstrated that cortical [12] and cancellous [22] bone from patients with primary osteoporosis without diabetes or renal failure, but with a high turnover rate as estimated by urinary DPD levels, contained significantly higher pentosidine levels in both low and high mineralized bone tissue. Therefore, the significant increased accumulation of pentosidine and tfAGEs in the OVX group may be due to oxidative stress induced by estrogen deficiency.

Ozasa et al. [35] and Sanada et al. [42] demonstrated in a rodent model that reduced lysyl oxidase activity induced by OVX was completely improved by estradiol injection. Khastgir et al. [43] showed that estrogen hormone replacement therapy (HRT) increased enzymatic cross-links in bone from elderly post-menopausal women with osteoporosis. Paschalis et al. [44] also demonstrated the beneficial effects of HRT and RLX on enzymatic cross-link ratios assessed by FTIRI when paired biopsy specimens were compared before and after 2 years of HRT. These results were consistent with our previous study, which demonstrated the favorable effects of raloxifene (RLX), a second generation SERM, on collagen enzymatic and non-enzymatic AGEs cross-links in OVX rabbits with or without hyperhomocysteinemia [19]. RLX increased the concentration of enzymatic cross-links, while decreasing AGEs cross-links after 16 weeks of treatment. The relatively short treatment duration is only 16 weeks in this report suggests that the marked increase in crosslink concentration cannot be explained by remodeling

alone. Not all enzymatic crosslink precursors, such as Lys and Hyl residues in collagen molecules, are cross-linked. Therefore, if lysyl oxidase activity is increased by estrogen or SERMs, the Lys and Hyl may form additional crosslinks in preexisting collagen fibers independent of bone remodeling. Gallant et al. [45] reported that ex vivo exposure of nonviable bone to raloxifene improves intrinsic toughness. These effects are cell independent and appear to be mediated by an increase in matrix-bound water. The hydroxyl groups (OH) on raloxifene were shown to be important for the increase of both water content and toughness. Similar results were reported by Allen et al. [46] using ultra-short echotime magnetic resonance imaging (UTE-MRI) in vivo. These results lead us to believe that there is also a change in matrix-bound water and an increase in the crosslinks ofpreexisting collagen. In this study, we demonstrated that a clinical dose of BZA for 18 months prevented the OVX-induced deleterious changes in immature and mature pyridinium cross-links and excessive hydroxylation of cross-links, pentosidine, and tfAGEs. Such changes in cross-link formation significantly contributed to the preservation of bone strength (Tables 2 and 3). BZA is well known to act as an estrogen agonist in bone via the binding to the estrogen receptor in bone cells and this interaction reduces oxidative stress [1,47]. Such positive effects of HRT and RLX on collagen cross-links may provide a possible explanation for the prevention of OVX-induced deterioration of collagen crosslink formation such as the amount of enzymatic cross-links, pentosidine and tfAGEs that was detected in the BZA group.

Our study had several limitations. First, we focused on vertebral can-cellous bone so we cannot definitively state whether the observed changes in bone mass, quality, and strength are applicable to whole bone strength including cortical bone. Second, we did not evaluate other determinants of material properties such as matrix-bound water and non-collagenous proteins, which may also affect bone strength. Third, enzymatic mature pyrrole cross-links may be as important as pyridinium cross-links [7,33]. However, since pyrrole cross-links are unstable during acid hydrolysis, we could not measure them using our HPLC method. Thus, we analyzed conventional immature divalent cross-links, mature pyridinium cross-links, pentosidine, tfAGEs, and hy-droxyproline analysis after acid hydrolysis. Since the major determinant of the total amount of immature divalent cross-links, which is the predominant type in bone collagen, and mature cross-links, such as pyridinium and pyrrole, is lysyl oxidase activity, the overall formation of pyrrole cross-links might be similar to other lysyl oxidase-controlled cross-linking. There is no doubt that not only pyridinium cross-links, but also pyrroles independently affect tissue strength [7, 33]. Indeed, human bone appears to contain pyrrole and pyridinium cross-links to the same extent [7,33]. Third, glucosepane is the major type of non-fluorescent AGEs. Glucosepane accumulates in the human extracellular matrix with aging in the skin and glomerular basement membrane [48]. Biemel et al. [49] demonstrated the structural similarities between pentosidine and glucosepane, which suggests a parallel mechanism in the respective formation pathways. However, to date, glucosepane in bone and changes in aging and estrogen deficiency have not been reported. Thus, the presence of glucosepane in bone collagen as well as skin and basement membrane should be confirmed and we should attempt to clarify the role of glucosepane in human and primate bone in future studies. Fourth, bone strength parameters were not significantly different between OVX and SHAM, in spite of the differences in micro-CT microarchitecture parameters. Although the mean values of ultimate load and breaking energy in the OVX group were lower than those in the SHAM group, these differences did not reach statistical significance. However, all parameters ofmicroarchitecture estimated by micro-CT correlated significantly to ultimate load, stiffness, and breaking energy in SHAM and OVX groups by Spearman's rank correlation coefficient analysis excluding the BZA-treated groups. Thus, microarchitecture affects bone strength. In previous reports using OVX monkeys, there was no significant reduction in bone strength in spite of the differences in microarchitecture and bone mineral density,

which may be due to individual differences [13,50-52]. Thus, this may be a limitation of OVX monkeys.

In conclusion, the prevention of OVX-induced deterioration of the actual amount of enzymatic immature and mature cross-link formation, the ratio of hydroxylysine-derived enzymatic cross-links, the non-enzymatic cross-link, pentosidine, and tfAGEs by BZA for 18 months in cynomolgus monkeys coincided with preservation of vertebral cancel-lous bone strength. The levels of enzymatic immature cross-links, pentosidine, and whole fluorescent AGEs after treatment with BZA significantly contributed to vertebral cancellous bone strength.

Acknowledgments

The authors are grateful to Ms. Mika Imamura and Ms. Kazumi Hirakawa (Research Assistants, Jikei University School of Medicine, Japan) for aiding in specimen preparation and testing.

References

[1] B.S. Komm, Y.P. Kharode, P.V. Bodine, H.A. Harris, C.P. Miller, C.R Lyttle, Bazedoxifene acetate: a selective estrogen receptor modulator with improved selectivity, Endocrinology 146 (2005) 3999-4008.

[2] C.P. Miller, M.D. Collini, B.D. Tran, H.A Harris, Y.P. Kharode, J.T. Marzolf, RA. Moran, RA Henderson, R.H. Bender, RJ. Unwalla, L.M. Greenberger, J.P. Yardley, M.A Abou-Gharbia, C.R Lyttle, B.S. Komm, Design, synthesis, and preclinical characterization of novel, highly selective indole estrogens, J. Med. Chem. 44 (2001) 1654-1657.

[3] O. Bruyére,J. Detilleux, A Chines, J.Y. Reginster, Relationships between changes in bone mineral density or bone turnover markers and vertebral fracture incidence in patients treated with bazedoxifene, Calcif. Tissue Int. 91 (2012) 244-249.

[4] B.S. Komm, A.A. Chines, Bazedoxifene: the evolving role of third-generation selective estrogen-receptor modulators in the management of postmenopausal osteoporosis, Ther Adv Musculoskelet Dis. 4 (2012) 21-34.

[5] A.G. Ellis, J.Y. Reginster, X. Luo, AG. Bushmakin, R Williams, S. Sutradhar, S. Mirkin, J.P. Jansen, Indirect comparison of bazedoxifene vs oral bisphosphonates for the prevention of vertebral fractures in postmenopausal osteoporotic women, Curr. Med. Res. Opin. 30 (2014) 1617-1626.

[6] E. Seeman, P.D. Delmas, Bone quality-the material and structural basis of bone strength and fragility, N. Engl. J. Med. 354 (2006) 2250-2261.

[7] L. Knott, A.J. Bailey, Collagen cross-links in mineralizing tissues: a review of their chemistry, function, and clinical relevance, Bone 22 (1998) 181-187.

[8] M. Saito, K. Marumo, Collagen cross-links as a determinant of bone quality: a possible explanation for bone fragility in aging, osteoporosis, and diabetes mellitus, Osteoporos. Int. 21 (2010) 195-214.

[9] S. Viguet-Carrin, P. Garnero, P.D. Delmas, The role of collagen in bone strength, Osteoporos. Int. 17 (2006) 319-336.

[10] K. Uzawa, W.J. Grzesik, T. Nishiura, S.A Kuznetsov, P.G. Robey, D.A. Brenner, M. Yamauchi, Differential expression of human lysyl hydroxylase genes, lysine hydrox-ylation, and cross-linking of type I collagen during osteoblastic differentiation in vitro, J. Bone Miner. Res. 14 (1999) 1272-1280.

[11] X. Banse, T.J. Sims, A.J. Bailey, Mechanical properties of adult vertebral cancellous bone: correlation with collagen intermolecular cross-links, J. Bone Miner. Res. 17 (2002) 1621-1628.

[12] M. Saito, K. Fujii, S. Soshi, T. Tanaka, Reductions in degree of mineralization and enzymatic collagen cross-links and increases in glycation-induced pentosidine in the femoral neck cortex in cases of femoral neck fracture, Osteoporos. Int. 17 (2006) 986-995.

[13] M. Saito, K. Marumo, Y. Kida, C. Ushiku, S. Kato, R Takao-Kawabata, T. Kuroda, Changes in the contents of enzymatic immature, mature, and non-enzymatic senescent cross-links of collagen after once-weekly treatment with human parathyroid hormone (1-34) for 18 months contribute to improvement of bone strength in ovariectomized monkeys, Osteoporos. Int. 22 (2011) 2373-2383.

[14] X. Wang, X. Shen, X. Li, C.M. Agrawal, Age-related changes in the collagen network and toughness of bone, Bone 31 (2002) 1-7.

[15] M. Saito, K. Fujii, Y. Mori, K. Marumo, Role of collagen enzymatic and glycation induced cross-links as a determinant of bone quality in spontaneously diabetic WBN/Kob rats, Osteoporos. Int. 17 (2006) 1514-1523.

[16] H. Oxlund, M. Barckman, G. Ortoft, T.T. Andreassen, Reduced concentrations of collagen cross-links are associated with reduced strength of bone, Bone 17 (1995) 365S-371S.

[17] D. Vashishth, The role of the collagen matrix in skeletal fragility, Curr. Osteoporos. Rep. 5 (2007) 62-66.

[18] P. Garnero, O. Borel, E. Gineyts, F. Duboeuf, H. Solberg, M.L. Bouxsein, C. Christiansen, P.D. Delmas, Extracellular post-translational modifications of collagen are major determinants of biomechanical properties of fetal bovine cortical bone, Bone 38 (2006) 300-309.

[19] M. Saito, K. Marumo, S. Soshi, Y. Kida, C. Ushiku, A. Shinohara, Raloxifene ameliorates detrimental enzymatic and nonenzymatic collagen cross-links and bone strength in rabbits with hyperhomocysteinemia, Osteoporos. Int. 21 (2010) 655-666.

[20] M. Saito, S. Mori, T. Mashiba, S. Komatsubara, K. Marumo, Collagen maturity, glycation induced-pentosidine, and mineralization are increased following 3-year treatment with incadronate in dogs, Osteoporos. Int. 19 (2008) 1343-1354.

[21] D. Vashishth, G.J. Gibson, J.I. Khoury, M.B. Schaffler, J. Kimura, D.P. Fyhrie, Influence of nonenzymatic glycation on biomechanical properties of cortical bone, Bone 28 (2001) 195-201.

[22] S.Y. Tang, D. Vashishth, Non-enzymatic glycation alters microdamage formation in human cancellous bone, Bone 46 (2010) 148-154.

[23] M. Saito, M.D. Grynpas, D.B. Burr, M.R. Allen, S.Y. Smith, N. Doyle, N. Amizuka, T. Hasegawa, Y. Kida, K. Marumo, H. Saito, Treatment with eldecalcitol positively affects mineralization, microdamage, and collagen crosslinks in primate bone, Bone 73 (2015) 8-15.

[24] H. Nojiri, Y. Saita, M. Saito, K. Kobayashi, C. Tsuda, T. Miyazaki, M. Saito, K. Marumo, I. Yonezawa, K. Kaneko, T. Shirasawa, T. Shimizu, Cytoplasmic superoxide causes bone fragility due to low turnover osteoporosis with impaired collagen cross-links, J. Bone Miner. Res. 26 (2011) 2682-2694.

[25] M. Saito, K. Fujii, K. Marumo, Degree of mineralization-related collagen cross-linking in the femoral neck cancellous bone in cases of hip fracture and controls, Calcif. Tissue Int. 79 (2006) 160-168.

[26] L. Karim, S.Y. Tang, G.E. Sroga, D. Vashishth, Differences in non-enzymatic glycation and collagen cross-links between human cortical and cancellous bone, Osteoporos. Int. 24 (2013) 2441-2447.

[27] S.Y. Smith, J. Jolette, L. Chouinard, B.S. Komm, The effects of bazedoxifene in the ovariectomized aged cynomolgus monkey, J. Bone Miner. Metab. 32 (2014) 161-172.

[28] A. Laib, J.L. Kumer, S. Majumdar, N.E. Lane, The temporal changes of trabecular architecture in ovariectomized rats assessed by MicroCT, Osteoporos. Int. 12 (2001) 936-941.

[29] M. Saito, K. Marumo, K. Fujii, N. Ishioka, Single-column high-performance liquid chromatographic-fluorescence detection of immature, mature, and senescent cross-links of collagen, Anal. Biochem. 253 (1997) 26-32.

[30] S.Y. Tang, U. Zeenath, D. Vashishth, Effects of non-enzymatic glycation on cancellous bone fragility, Bone 40 (2007) 1144-1151.

[31] A. Ural, C. Janeiro, L. Karim, T. Diab, D. Vashishth, Association between non-enzymatic glycation, resorption, and microdamage in human tibial cortices, Osteoporos Int. 26 (2015) 865-873.

[33] S.P. Robins, Biochemistry and functional significance of collagen cross-linking, Biochem. Soc. Trans. 35 (2007) 849-852.

[34] M. Saito, A. Shiraishi, M. Ito, S. Sakai, N. Hayakawa, M. Mihara, K. Marumo, Comparison of effects of alfacalcidol and alendronate on mechanical properties and bone collagen cross-links of callus in the fracture repair rat model, Bone 46 (2010) 1170-1179.

[35] H. Ozasa, T. Tominaga, T. Nishimura, T. Takeda, Lysyl oxidase activity in the mouse uterine cervix is physiologically regulated by estrogen, Endocrinology 109 (1981) 618-621.

[36] D.C. Ireland, S. Bord, S.R. Beavan, J.E. Compston, Effects of estrogen on collagen synthesis by cultured human osteoblasts depend on the rate of cellular differentiation, J. Cell. Biochem. 86 (2) (2002) 251-257.

[37] S.C. Manolagas, From estrogen-centric to aging and oxidative stress: a revised perspective of the pathogenesis of osteoporosis, Endocr. Rev. 31 (2010) 266-300.

[38] V.H. Urbieta-Caceres, F.A. Syed, J. Lin, X.Y. Zhu, K.L. Jordan, C.C. Bell, M.D. Bentley, A. Lerman, S. Khosla, L.O. Lerman, Age-dependent renal cortical microvascular loss in female mice, Am. J. Physiol. Endocrinol. Metab. 302 (2012) E979-E986.

[39] J.Y. Chen, P.J. Tsai, H.C. Tai, R.L. Tsai, Y.T. Chang, M.C. Wang, Y.W. Chiou, M.L. Yeh, M.J. Tang, C.F. Lam, S.C. Shiesh, Y.H. Li, W.C. Tsai, C.H. Chou, L.J. Lin, H.L. Wu, Y.S. Tsai, Increased aortic stiffness and attenuated lysyl oxidase activity in obesity, Arterioscler. Thromb. Vasc. Biol. 33 (2013) 839-846.

[40] B. Raposo, C. Rodriguez, J. Martinez-Gonzalez, L. Badimon, High levels of homocysteine inhibit lysyl oxidase (LOX) and down regulate LOX expression in vascular endothelial cells, Atherosclerosis 177 (2004) 1-8.

[41] A.D. McCarthy, S.B. Etcheverry, L. Bruzzone, G. Lettieri, D.A. Barrio, A.M. Cortizo, Non-enzymatic glycation of a type I collagen matrix: effect on osteoblastic development and oxidative stress, BMC Cell Biol. 2 (2001) 16.

[42] H. Sanada, J. Shikata, H. Hamamoto, Y. Ueba, T. Yamamuro, T. Takeda, Changes in collagen cross-linking and lysyl oxidase by estrogen, Biochim. Biophys. Acta 541 (1978) 408-413.

[43] G. Khastgir, J. Studd, N. Holland, J. Alaghband-Zadeh, T.J. Sims, A.J. Bailey, Anabolic effect of long-term estrogen replacement on bone collagen in elderly postmeno-pausal women with osteoporosis, Osteoporos. Int. 12 (2001) 465-470.

[44] E.P. Paschalis, A.L. Boskey, M. Kassem, E.F. Erisksen, Effect of hormone replacement therapy on bone quality in early postmenopausal women, J. Bone Miner. Res. 18 (2003) 955-959.

[45] M.A. Gallant, D.M. Brown, M. Hammond, J.M. Wallace, J. Du, A.C. Deymier-Black J.D. Almer, S.R. Stock, M.R. Allen, D.B. Burr, Bone cell-independent benefits of raloxifene on the skeleton: a novel mechanism for improving bone material properties, Bone 61 (2014 Apr) 191-200, http://dx.doi.org/10.10167j.bone.2014.01.009 Epub 2014 Jan 24.

[46] M.R. Allen, P.R. Territo, C. Lin, S. Persohn, L. Jiang, A.A. Riley, B.P. McCarthy, C.L. Newman, D.B. Burr, G.D. Hutchins, In vivo UTE-MRI reveals positive effects of raloxifene on skeletal bound water in skeletally mature beagle dogs, J Bone Miner Res (Jan 31 2015), http://dx.doi.org/10.1002/jbmr.2470 ([Epub ahead of print]).

[47] L.P. Hemachandra, H. Patel, R.E. Chandrasena, J. Choi, S.C. Piyankarage, S. Wang, Y. Wang, E.N. Thayer, R.A Scism, B.T. Michalsen, R Xiong, M.I. Siklos, J.L. Bolton, G.R Thatcher, SERMs attenuate estrogen-induced malignant transformation of human mammary epithelial cells by upregulating detoxification of oxidative metabolites, Cancer Prev. Res. (Phila.) 7 (2014) 505-515.

[48] X. Banse, T.J. Sims, A.J. Bailey, Mechanical properties of adult vertebral cancellous bone: correlation with collagen intermolecular cross-links, J. Bone Miner. Res. 17 (2002) 1621-1628.

[49] K.M. Biemel, O. Reihl, J. Conrad, M.O. Lederer, Formation pathways for lysine-arginine cross-links derived from hexoses and pentoses by Maillard processes: unraveling the structure of a pentosidine precursor, J. Biol. Chem. 276 (2001) 23405-23412.

[50] D.B. Burr, T. Hirano, C.H. Turner, C. Hotchkiss, R Brommage, J.M. Hock, Intermittently administered human parathyroid hormone(1-34) treatment increases intracortical bone turnover and porosity without reducing bone strength in the

humerus of ovariectomized cynomolgus monkeys, J. Bone Miner. Res. 16 (2001) 157-165.

[51] J. Fox, MA. Miller, M.K. Newman, C.H. Turner, R.R. Recker, S.Y. Smith, Treatment of skeletally mature ovariectomized rhesus monkeys with PTH(1-84) for 16 months increases bone formation and density and improves trabecular architecture and bio-mechanical properties at the lumbar spine, J. Bone Miner. Res. 22 (2007) 260-273.

[52] J. Fox, M.K. Newman, C.H. Turner, R.E. Guldberg, A. Varela, S.Y. Smith, Effects of treatment with parathyroid hormone 1-84 on quantity and biomechanical properties of thoracic vertebral trabecular bone in ovariectomized rhesus monkeys, Calcif. Tissue Int. 82 (2008) 212-220.