1154–1163www.elsevier.com/locate/bone

Bone 42 (2008)

Onion decreases the ovariectomy-induced osteopenia in young adult rats

Tsang-Hai Huang a, Roman C. Mühlbauer g, Chih-Hsin Tang e, Hsiun-Ing Chen b,Guan-Liang Chang c, Yi-Wei Huang d, Yu-Ting Lai d, Hsin-Shi Lin a,

Wei-Ting Yang c, Rong-Sen Yang f,⁎

a Institute of Physical Education, Health and Leisure Studies, National Cheng-kung University, No. 1, University Road, Tainan City 701, Taiwanb Department of Physiology, National Cheng-kung University, No. 1, University Road, Tainan City 701, Taiwan

c Institute of Biomedical Engineering, National Cheng-kung University, No. 1, University Road, Tainan City 701, Taiwand Department of Life Sciences, National Cheng-kung University, No. 1, University Road, Tainan City 701, Taiwan

e Department of Pharmacology, School of Medicine, China Medical University, No. 91, Xue-shi Road, Taichung City 404, Taiwanf Department of Orthopaedics, National Taiwan University and Hospital, No. 7, Chung-Shan South Road, Taipei City 100, Taiwan

g Bone Biology Group, Department of Clinical Research, University of Bern, Murtenstrasse 35, CH-3010 Bern, Switzerland

Received 21 July 2007; revised 15 November 2007; accepted 30 January 2008Available online 29 February 2008

Edited by: Robert Recker

Abstract

It has been suggested that fruit and vegetable consumption are associated with good bone health. Onion, in particular, has been verified in itsefficacy in bone resorption activity. In this study, we further investigated the effects of an onion-containing diet on ovariectomy-induced boneloss using methods of serum marker assay, histomorphometric analysis and biomechanical tests. Sixty-four female Wistar rats (14-week-old)with sham operations or ovariectomy were assigned to 6 groups: CON, sham-operated control group; OVX, ovariectomized group; ALN,ovariectomized rats treated with alendronate (1 mg/kg/day, p.o.); and 3% ON, 7% ON and 14% ON, ovariectomized rats fed with dietscontaining 3%, 7% and 14% (wt/wt) onion powder, respectively. Animals were sacrificed after a six-week treatment course. In the serum markerassay, alendronate and all three onion-enriched diets significantly decreased serum calcium level (pb0.05). Both 14% ON group and the ALNgroup even showed similarly lower level of serum osteocalcin ( pb0.05), suggesting a down-regulation of bone turnover. Thehistomorphometric analysis showed that ovariectomy markedly decrease bone trabeculae. The ALN and 14% ON rats were 80% and 46%higher, respectively, in BV/TV than the OVX rats ( pb0.05), and the rats fed with onion-enriched food showed a lesser ovariectomy-inducedbone loss in a dose-dependent manner. Additionally, both ALN and 14% ON groups had significantly more trabecular number, less separatedtrabeculae, and fewer osteoclasts ( pb0.05), but the protective efficacy from the 14% onion-enriched diet was slightly inferior to that ofalendronate. Ovariectomy also significantly decreased tissue weight and biomechanical strength in the OVX group ( pb0.05). The ALN and14% ON groups equivalently showed a lesser decrease in tissue weight, though the difference was not significant. On the other hand, both theALN and 14% ON groups represented similar biomaterial properties of femurs, and both reduced the ovariectomy-induced decrease in bendingload and bending energy ( pb0.05). The present study further verified that an onion-enriched diet could counteract ovariectomy-induced boneloss and deterioration of biomechanical properties.© 2008 Elsevier Inc. All rights reserved.

Keywords: Osteoporosis; Phytotherapy; Alendronate; Osteoclast; Menopause

⁎ Corresponding author. Department of Orthopaedics, College of Medicine,National Taiwan University and Hospital, No. 7, Chung-Shan South Road,Taipei (100), Taiwan. Fax: +886 2 23936577.

E-mail address: rsyang@ntuh.gov.tw (R.-S. Yang).

8756-3282/$ - see front matter © 2008 Elsevier Inc. All rights reserved.doi:10.1016/j.bone.2008.01.032

Introduction

The increasing incidence of osteoporosis and its related frac-tures are important global health issues. Osteoporosis affectsabout 75 million people in Europe, USA and Japan [1]. Inaddition, one-third of women, as well as one-fifth of men over 50,may develop osteoporotic fractures in their lifetimes [2–4]. It was

1155T.-H. Huang et al. / Bone 42 (2008) 1154–1163

predicted that, by 2050, more than 50% of worldwide osteo-porotic hip fractures would occur in Asia [5]. Among thepharmacological treatments, hormone therapy (HT), selectiveestrogen receptor modulators (SERMs) and bisphosphonate havebeen shown to be effective either in increasing bone mineraldensity (BMD) and/or reducing fracture rates [6]. However, theside effects of these medications, including gastrointestinaltolerance problems in bisphosphonate and the potential malig-nancies in HT [7,8], may preclude their long-term use. Growingevidence of the benefits of natural foods for bone health pro-vide an alternative option for prevention and/or treatment ofosteoporosis.

Cross-sectional studies have shown that people, includingpostmenopausal women, healthy women, and elderly men,consuming more fruits and vegetables accumulated a higherbone mineral density (BMD) [9–15]. In addition, increased fruitand vegetable consumption showed benefits for bone health andbone development in young people [15–17]. In animal studies,it had been demonstrated that short-term treatment with diversefruits, vegetables and herbs could effectively reduce boneresorption activity [18–20]. One recent study showed the effectsof 5%, 15%, and 25% prune-enriched diets in reversing boneloss in young adult ovariectomized rats [21]. The 25% prune-enriched diet even showed an equally protective effect on bonevolume ratio (BV/TV, %) as 17β-estradiol E2. Similarly, asummary of animal studies strongly suggested the positiveeffects of onion on good bone health. A short-term treatment(10 days) with onion (0.03–1.5 g/rat/day) has been shown todecrease the ovariectomy-induced bone resorption rate in adose-dependent manner [22]. A dose at 1 g/rat/day has evenserved as a positive control treatment for down-regulating boneresorption [19,20]. The aim of this study was to investigatewhether an onion-enriched diet could prevent estrogendeficiency induced osteopenia. For this purpose, serum bonemarkers and long bone biomechanical testing were measuredand tissue histomorphometry was performed to investigate thedose-dependent bone protective effects of onion on 14-week-old ovariectomized Wistar rats for a period of six weeks.

Fig. 1. Experimental design. Animals were fed semi-purified AIN-76A powder start14 weeks old, treatments with onion diet or alendronate for each group were begun 3CON, control group; ON, onion-treated groups; OVX, ovariectomized group.

Materials and methods

Animals

Female Wistar rats (n=64) were housed under controlled conditions includinga room temperature of 22±1 °C with a 12:12 h light–dark cycle. All animals werefed with standard Purina Rodent Chow 5001 (Labdiet®, Richmond, IN, U.S.A.)containing 0.95% calcium and 1.07 % phosphate (wt/wt of dry food) and tap waterad libitum. The procedures of the animal study, including the raising, feeding, andthewhole surgical process, followed the APS's “Guiding Principles in the Care andUse of Animals” andwere approved by the Committee of Animal Study at NationalCheng Kung University (Document Serial No. 940204).

Experimental design and treatments

The use of rats younger than 3months is not suitable for ovariectomy due to theconcurrence of bone remodeling and bone growth [23]. Thus, we used adult virginfemale Wistar rats with an age of 14 weeks. One week before the ovariectomy orsham operation, all animals (13 weeks old) were given demineralized water and adiet of “semi-purified” diet powder (AIN-76A, TestDiet®, Richmond, IN, USA)with animal protein (20% casein) containing 0.52% calcium and 1.12%phosphate.Three days after surgery, sham-operated animals (n=8) were set as the controlgroup and ovariectomized rats (n=56) were randomly assigned to five groups(Fig. 1): the OVX group (n=11), fed with the AIN-76A diet without any addition;theALNgroup (n=12), fedwith theAIN-76Adiet+alendronate (CalbiochemCo.,Merck Taiwan Ltd, Taiwan) at a dose of 1 mg/kg b.w./day (p.o.), which has beenused by others in either growing or aged ovariectomized rats [24,25]; and threeonion treatment groups (n=11 for each), the 3%ON, 7%ON and 14%ON groups,fed with an AIN-76A diet containing 3%, 7% and 14% (wt/wt) onion powder,respectively. The onion powder (Chiseng®, Taipei, Taiwan), whichwasmade fromfresh onion without any additives, is commercially available. The treatment coursefor all groups was six weeks (Fig. 1).

The female Wistar rats ate, on average, 15–20 g/day of AIN-76A powderaccording to the evaluation in the week before ovariectomy surgery. Since three orfour rats were housed per cage, we used ad libitum feeding. Thus, the diet for eachgroupwas freshly prepared everyday andwas given at amean amount of 25 g per rat.

Serum bone marker assay

After the treatments were completed, the animals were sacrificed under deepanesthesia with a high dose of sodium pentobarbital (intraperitoneally, 65 mg/kg)and bloodletting via carotid arteries. Blood samples were collected andcentrifuged (4 °C) at 1500 ×g for 20 min. Serum was separated immediatelyandwas stored at−75 °C for future bonemetabolicmarker assay. Serum calcium,and phosphorus were determined with an automatic chemistry analyzer

ing at 13 weeks old; sham or ovariectomy (OVX) operations were performed atdays after surgery and maintained for 6 weeks. ALN, alendronate-treated group;

Table 1Body weight gain (gm) during the treatment period

CON OVX ALN 3% ON 7% ON 14% ON Power

1st week 13.9±1.3c

35.0±1.8a

28.8±2.3b

32.1±2.0ab

31.5±1.5ab

28.0±2.5b

1.000

2nd week 27.1±2.9b

50.7±3.3a

46.9±3.6a

53.4±2.9a

48.6±3.0a

44.5±4.5a

0.990

3rd week 32.1±3.4b

60.5±5.1a

58.3±3.9a

65.5±3.6a

65.0±4.7a

58.7±5.0a

0.994

4th week 36.8±4.4b

67.9±5.1a

63.8±4.1a

71.2±3.4a

61.7±4.5a

61.6±5.1a

0.991

5th week 36.1±4.2b

76.8±5.7a

73.2±3.2a

79.5±4.1a

67.5±5.4a

66.9±4.8a

1.000

6th week 44.9±4.6b

81.5±5.4a

76.3±5.0a

79.5±3.9a

68.2±5.9a

75.1±5.8a

0.984

Data were presented as mean±SEM. Power, power values calculated usingα=0.05; a–cMean±SEM values within a group not sharing a common super-script are significantly different ( pb0.05).

1156 T.-H. Huang et al. / Bone 42 (2008) 1154–1163

(Beckman Synchron LX20, Diamond Diagnostics, Holliston, MA, USA).ELISA kits, including RatTRAP™, Rat-MID™ and Rat/Mouse IGF-I (NordicBioscience Diagnostics, Herlev, Demark), were used to analyze serum tartrate-resistant acid phosphatase (TRAP) activity, osteocalcin, and insulin-likegrowth factor-I (IGF-I), respectively. Assay procedures followed the instruc-tions included in package of each kit. A microplate photometer (MultiskanAscent, Thermo Fisher Scientific, Inc., Waltham, MA, USA) was used tomeasure the absorbance of each well after a series of sample preparations andreactions.

Bone samples preparation

The left tibiae were removed, fixed with a 4% neutral paraformaldehyde inPBS solution ( pH 7.4) for 48 h and then decalcified with a 10% ethylenediaminetetraacetic acid (EDTA) solution (pH 7.4) at 4 °C for 4 weeks. After decal-cification, each bone sample was cut along the coronal plane and embedded withparaffin for further tissue section and histological staining. For geometric andbiomechanical measurement, the left femora were removed and the soft tissueswere cleaned. Then, the bone tissues were wrapped in 0.9% sodium chloridesoaked gauze and aluminum foil, and stored at −20 °C for future biomechanicaltesting.

Bone histomorphometric analysis

Histological staining, including hemotaxylin and eosin (H & E) staining andtartrate-resistant acid phosphates (TRAP) staining, was performed on twononconsecutive sections (5 μm) for each sample, which were the 1st and the 11thsections in sequence. H & E staining was performed according to our previousstudies [26]. The histomorphometric study of the metaphysis of the proximaltibiae was performed as previously described with image analysis software(Image Pro Plus 6.1 for Windows; Media Cybernetics, Silver Spring, MD, USA)[27]. Parameters measured included the bone volume ratio (BV/TV, %); meanthickness of the trabeculae (Tb.Th, μm); trabecular number (Tb. N., 1/mm); andtrabecular separation (Tb. Sp., mm). TRAP staining with a standard kit (Sigma-Aldrich Co., St. Louis, MO, USA) was used to determine the presence ofosteoclasts. The TRAP-positive multinucleated osteoclast along the metaphy-seal trabecular bone surfaces were counted, and the number of osteoclasts perunit trabecular tissue area (N. Oc/T. A, 1/mm2) was measured as an index ofbone resorption [27].

Femur weight measurements

Before biomechanical testing, femora were thawed at room temperature andweighed as wet weight. After the completion of biomechanical testing, thefractured bone specimens were immersed in a solvent (2 vol of chloroformcombined with 1 vol of methanol) for 1 week and then dried at 80 °C for 24 h[28]. The fat-free dry weight (FFDW) of each femur was then measured.

Biomechanical three-point bending testing and geometry measurements

Mechanical properties of long bone tissues were determined by a three-pointbending test [29]. The material testing system was designed by our lab with a50 lb load cell (MDB-50, Transducer Techniques, Inc., Temecula, CA, USA),which was connected to a data acquisition system (Model: iNstruNet, GWInstruments, Inc. Somerville, MA, USA) in a personal computer. The span of thetwo support points was 20 mm, the deformation rate was 0.05 mm/sec andsampling frequency was 50 Hz. Load-displacement data were used for furthercomputing of the extrinsic material properties, including max load, ultimate loadto failure (fracture load), energy to maximal load, energy to fracture load, andlinear stiffness. The failure sites of all bone specimens were photographed undera ×25 magnification by a high-resolution digital camera (COOLPIX 4500,Nikon, Japan) [29]. Cross-sectional parameters, including cortical bonethickness and cross-sectional area of cortical bone (CSA), were measuredfrom the photographs using the software Image Pro Plus 6.1 for Windows(Media Cybernetics, Silver Spring, MD, USA). The cross-sectional moment ofinertia (CSMI) of the failure site was measured by the methods of Turner andBurr for irregular cross-sections [30]. Based on the assumption that a bone cross-

section is made up of many rectangular elements (pixels), each with a height of hand a width of w, the equation is shown as follows:

I ¼Xn

i¼1

wh3=12þ whd2i� �

where I is CMSI, n is the number of pixels, d is the distance from the center ofthe element of the area to a given axis on the cross-section. All of theseparameters were obtained using the software Image Pro Plus 6.1 for Windows(Media Cybernetics, Silver Spring, MD, USA).

Statistical analysis

The data are presented as means±SEM. For the data of serum markers,histomorphometry and body weight gain, one-way analysis of variance(ANOVA) was used to evaluate the differences among groups. For geometricand biomechanical data, one-way analysis of covariance (ANCOVA) was usedto evaluate the differences among groups, and body weight data was set as acovariate. When significant levels ( pb0.05) were revealed, pair-wisecomparisons between groups were made using the Fisher's least significantdifference (LSD) method. Power values regarding the size of groups werecalculated using α=0.05. Statistical analysis software, SPSS (11.5 version,SPSS, Chicago, IL), was adopted for processing the data of the present study.

Results

Body weight gain (one-way ANOVA)

After surgery, body weight gain was typically and consistentlyhigher in the ovariectomized rats in each group when comparedwith the sham-operated control rats (pb0.05) (Table 1). Exceptfor the slightly lower body weight gain of the ALN and 14% ONgroup in the first week, the interventions we used, includingalendronate and three onion-enriched diets, did not significantlyaffect body weight gain throughout the experimental period.

Serum bone markers (one-way ANOVA)

After the six-week treatment, the ALN and three onion-enriched groups similarly showed significantly lower serumcalcium levels as compared to the CON and OVX groups(pb0.05) (Table 2). In osteocalcin assay, the ALN group showed

Table 2Serum markers of bone metabolism

CON OVX ALN 3% ON 7% ON 14% ON Power

Calcium(mg/dl)

9.68±0.07a

9.43±0.1a

8.96±0.09b

8.90±0.11b

8.83±0.15b

8.83±0.10b

1.000

Phosphorus(mg/dl)

4.73±0.25

5.17±0.28

4.73±0.42

5.68±0.29

5.49±0.35

4.85±0.18

0.546

TRAP(U/L)

1.84±0.43

1.47±0.12

1.55±0.17

1.95±0.12

1.68±0.14

1.75±0.19

0.314

Osteocalcin(ng/ml)

123.4±23.2bc

166.7±17.0ab

109.2±8.7c

199.3±13.9a

150.6±21.0abc

129.1±19.8bc

0.927

IGF-I(ng/ml)

931.9±74.8

1020.2±58.7

913.9±46.9

1003.2±78.4

976.2±49.5

899.0±65.6

0.221

Data were presented as mean±SEM. Power, power values calculated usingα=0.05; a–cMean±SEMvalues within a group not sharing a common superscriptare significantly different (pb0.05).

1157T.-H. Huang et al. / Bone 42 (2008) 1154–1163

the lowest value and was significantly lower than the OVX and3% ON groups (pb0.05). The 14% ON groups also showedsignificantly lower serum osteocalcin level when compared to the3% ON group (pb0.05). The TRAP and IGF-I assays in bothALN group and 14% ON group also showed a numeric lowerlevel as compared with CON group or OVX group, but notreaching statistical significance.

Histomorphometry study (one-way ANOVA)

As shown in Figs. 2 and 3, the ovariectomy procedure causedsignificantly less bone volume ratio (BV/TV, %), less trabecularnumber (Tb. N., 1/mm), and increased trabecular separation(Tb. Sp., mm) in the OVX group as compared with the CONgroup ( pb0.05). In addition, the CON rats had a significantlyhigher BV/TV than the rats in the ALN and three onion-enriched diet groups ( pb0.05) (Figs. 2A and 3). Alendronate-treated rats showed a significantly higher BV/TV as comparedto the rats in the OVX, 3% ON and 7% ON groups ( pb0.05),but not the rats in the 14% ON group. The 14% ON group wassignificantly higher in BV/TV than the OVX and 3% ONgroups ( pb0.05). Linear regression analysis showed that theBV/TV data were significantly related to the ratio of onion-contained in the diet (r=489, pb0.05) (Fig. 2B).

Further histomorphometric analysis of architecture para-meters showed that trabecular sections were 14–17% thicker inrats fed 7% and 14% onion-enriched diets than those in the ALNgroup, but not at a significant level (Fig. 2C). Though the 14%ON group was significantly less in trabecular number than theCON and ALN groups, the rats in those three groups hadsignificantly higher trabecular number and less trabecular sepa-ration as compared to the OVX, 3% ON and 7% ON groups( pb0.05) (Fig. 2D and E). In osteoclast number per trabeculararea, the OVX group was significantly higher than the other fivegroups ( pb0.05) (Fig. 2F).

Femur weight, geometry, and biomechanical testing(one-way ANCOVA)

In order to eliminate the effects of body weight on size-related parameters and extrinsic biomechanical properties, ananalysis of covariance was used to adjust and analyze the mean

values using body weight as a covariate [29]. The CON groupwas significantly higher in femoral wet weight than the OVX,3% ON and 7% ON groups ( pb0.05) (Table 3). The ALN andthe 14% ON groups were numerically higher in wet weight thanthe three other ovariectomized groups, but did not show asignificant difference. The CON group was also significantlyhigher in femoral fat-free dry weight (FFDW) than allovariectomized groups ( pb0.05). However, the ALN groupshowed a smaller decrement of FFDW and was higher than theOVX ( p=0.057) and 3% ON groups ( p=0.007). As comparedto the CON group, the five ovariectomized groups had lowervalues for three cross-sectional geometric measurements,ranging from −2.0% to −13.6% lower (Table 3), but showedno statistical significance. Also, the ALN and 14% ON groupsshowed less decrease in CSA, cortical thickness and CSMI ascompared with OVX group.

Regarding the femoral three-point bending test, the CON ratsshowed the highest maximal load value and were significantlydifferent from rats in the OVX, 3% ON and 7% ON group(Fig. 4A). In fracture load value, the CON group was also thehighest, but only significantly higher than the 3% ON group.Treatments with alendronate and 14% onion-enriched diets couldprevent the deterioration of bending strength; the ALN and 14%ON groups were significant higher in values of maximal load andfracture load than the 3% ON group (pb0.05) (Fig. 4A and B).Paralleling loading values, the CON group showed the highestenergy tomaximal load and fracture loadwith significantly higherdata than those of the OVX group (pb0.05) (Fig. 4C and D).Again, the alendronate and 14% onion-enriched diets respectivelypreserved the properties of bone bending energy. The ALN groupshowed significantly higher energy to maximal load than theOVX and 3% ON groups, and significantly higher energy tofracture load than theOVXgroup.On the other hand, the 14%ONgroup also had a higher bending energy than the OVX group inenergy to maximal load (p=0.032) and fracture load (p=0.059),though a statistically significant difference was not revealed in thelater measurement (Fig. 4D). In bending stiffness, the CON groupwas significant higher than all ovariectomized groups (pb0.05)except for the 14% ON group. The rats treated with 14% ONshowed a slightly higher stiffness and were significantly higherthan the 3% ON group (Fig. 4E).

Discussion

In this study we showed that onion-enriched diets couldcounteract the ovariectomy-induced osteopenia in young adultrats in a dose-dependent manner. The efficacies of the highestdose of onion (14%) in preventing bone loss were similar tothose of alendronate treatment in many measurements.

Body weight gain

Previous work showed that onion is effective in intact malerats and does not affect specific body weight [19,20]. A previousstudy using rutin, an extract from onion, as a treatment showedno effect on ovariectomy-induced weight gain [31]. Therefore,there is no reason to assume that onion displays an estrogen-like

Fig. 2. Histomorphometric analysis of proximal tibiae among different groups. A. bone volume ratio (BV/TV, %); B. linear regression between BV/TV and dietary onion consumption ratio; C. mean thickness of thetrabeculae (Tb.Th, μm); D. trabecular number (Tb. N., 1/mm); E. trabecular separation (Tb. Sp., mm). F. number of osteoclasts per unit tissue area (N. Oc/T. A, 1/mm2). Ovariectomy caused much decrease of BV/TV(%), and trabecular number, increased trabecular separation and number of osteoclasts. However, the ALN and 14% ON groups could counteract the ovariectomy-induced changes. Power, power values calculated usingα=0.05; a–dMean±SEM values within a group not sharing a common superscript are significantly different ( pb0.05).

1158T.-H

.Huang

etal.

/Bone

42(2008)

1154–1163

Fig. 3. Histological sections of metaphysis of the proximal tibiae in rats of different groups. The ovariectomy caused serious loss of trabeculae in the proximal tibiae ofthe OVX group, but alendronate and 14% onion-contained diet significantly protected bone from ovariectomy-induced osteopenia. An onion-containing diet preservedmore trabeculae close to the endo-cortical surface, especially in the 7% ON group. Trabeculae near by cortical bone were thicker (arrows) than those located in thecentral of medullar canal (arrow heads); Bar, 1 mm.

1159T.-H. Huang et al. / Bone 42 (2008) 1154–1163

effect. In order to minimize the interference from body weight-induced mechanical loading on bone, alendronate was chosenas positive control treatment due to its nominal effect onovariectomy-induced body weight gain [32]. During the six-week treatment period, all five ovariectomized groups showedsimilar trends of body weight gain. Neither the alendronate, northe three different onion-enriched diets showed any obviousmodification in the increment of body weight gain induced byovariectomy and, thus, reduced the interference from bodyweight on bone metabolism among different treated groups.

Serum bone markers

Except for serum TRACP 5b activity, the ALN group showedlower levels in serum mineral level and markers of boneformation. It has been previously verified that alendronateprevents bone loss through down-regulating bone turnover [33–

Table 3Tissue weight, geometry measurements

CON OVX AL

Wet weight, mg 1165.2±40.2a 1047.5±33.5b 110(−10.1%) (−4

FFDW, mg 704.5±19.2a 594.3±16.0bc 638(−15.6%) (−9

CSA, mm2 6.40±0.28 5.81±0.23 6.0(−9.2%) (−4

Cortical bone thickness of the CS, mm 0.70±0.02 0.66±0.02 0.6(−5.9%) (−2

CSMI mm4 5.55±0.44 4.83±0.37 5.2(−13.1%) (−6

Data were presented as means±SEM. FFDW, fat-free dry weight; CSA, cross-sectionData in parentheses were percentage differences relative to the CON group. Power, psharing a common superscript are significantly different ( pb0.05).

35]. Since the high turnover may lead to loss of bone, such adown-regulation of bone turnover may be beneficial to the bonemetabolism. The 14% ON group's equivalent results suggestedthat a high onion-enriched diet could counteract bone loss througha similar pathway. In contrast with previous studies investigatingthe effects of dried plums [21,36], our serum markers seemed notsuggest enhanced bone formation activity from onion-enricheddiets. This might be due to the different status of subjects. SinceArjmandi et al. and Deyhim et al. treated their subjects after aperiod of estrogen deficiency [21,36], their studies demonstrated amajor effect of dried plums in bone restoration. However, oursubjects were treated immediately after ovariectomy. Moreover,onion and dried plums might act on bone metabolism throughdifferent pathways, which need to be further clarified.

Regarding bone resorption markers, TRACP 5b has beenproven as a sensitive marker for detecting bone resorption;however, this sensitivity seemed to be limited within the initial

N 3% ON 7% ON 14%ON Power

9.0±32.6ab 1020.1±36.3b 1045.8±35.6b 1079.1±33.8ab 0.556.8%) (−12.5%) (−10.2%) (−7.4%).6±15.6b 571.9±17.3c 593.5±17.0bc 603.7±16.1bc 0.992.4%) (−18.8%) (−15.6%) (−14.3%)9±0.22 5.77±0.25 5.86±0.25 6.20±0.23 0.290.8%) (−9.8%) (−8.4%) (−3.1%)8±0.02 0.66±0.02 0.65±0.02 0.69±0.02 0.344.1%) (−6.3%) (−7.8%) (−2.0%)1±0.36 4.80±0.40 5.12±0.39 5.28±0.37 0.156.2%) (−13.6%) (−7.8%) (−5.0%)

al area of cortical bone in fracture site; CSMI, cross-sectional moment of inertia.ower values calculated using α=0.05; a–cMean±SEM values within a group not

Fig. 4. Evaluations of the midshaft femur biomechanical properties by using a three-point bending test. A. loading force to maximal load; B. loading force to tissuefracture; C. energy to maximal load; D. energy to tissue fracture; E. bending stiffness of tissue. Ovariectomy significantly decreased the bending load, bending energy,and stiffness as compared with controls. However, the ALN and 14% ON groups showed effects to counteract the deteriorations of these parameters. Power, powervalues calculated using α=0.05; a–cMean±SEM values within a group not sharing a common superscript are significantly different ( pb0.05).

1160 T.-H. Huang et al. / Bone 42 (2008) 1154–1163

two weeks after ovariectomy or orchidectomy [37]. SinceTRACP 5b is released from osteoclasts, the serum activity ofTRACP 5b could be partially affected by the absolute numberof osteoclasts. Therefore, it was speculated that the serumTRACP 5b activity turn back to sham level due to the decreasein trabecular area and absolute osteoclast number as well [37].This might be able to explain the inconsistency between the

serum TRACP activity and histological osteoclast number pertrabeculae area (N. Oc/T. A, Fig. 2F).

Histomorphometry analysis

Previous studies showed that onion at a dose of 1 g/rat/daycould inhibit bone resorption shown as about a 20% decrease of

1161T.-H. Huang et al. / Bone 42 (2008) 1154–1163

[H3]-tetracycline urinary excretion [18–20]. Furthermore, ahigher dose of onion (1.5 g/rat/day) could blunt [H3]-tetracyclineurinary excretion up to 26% in ovariectomized rats [22]. In thepresent study, we further investigated the effects of onion onovariectomy-induced loss of trabecular bone tissue. As a positivecontrol treatment, alendronate significantly prevented trabecularbone loss in bone volume ratio (BV/TV) and kept better archi-tecture as compared to the OVX group (Figs. 2 and 3). However,the dose of alendronate (1 mg/kg/day) used in the present studycould not fully counteract the effects of ovariectomy whencompared to control rats (Fig. 2).

On the other hand, the rats treated with onion-enriched dietsalso showed a good preservation of BV/TV. A positive linearrelationship between BV/TV and onion-enrichment ratio sug-gested the impact of onion is dose-dependent (Fig. 2A and B).The highest onion-treated (14%) group showed similar effects tothe alendronate-treated group. Only the highest onion doseprevented a decrease in the number of trabeculae (Fig. 2D).Interestingly, the 7% ON and 14% ON groups both showednumerically higher average trabecular thickness (Fig. 2C). Sincethe trabeculae near the endo-cortical surface were thicker onaverage than those located in the center of the medullar canal, thepreservation of more trabeculae in the central medullary canal inthe ALN rats made their thickness of trabeculae (Tb.Th) onaverage slightly lower (Fig. 2C). A decreased number of TRAP-positive osteoclasts in onion-treated groups and the ALN groupindicated that onion-enriched diets could act similarly toalendronate through a pathway of bone resorption inhibition(Fig. 2F). This decreased number of osteoclasts confirmed theonion's anti-resorptive effects on bone, which has previouslybeen suggested [18–20]. The 3% ON group failed to showpreservation of BV/TV, which might be because the lower doseof onion treatment was only able to inhibit osteoclastic activityduring the later phase of estrogen deficiency, when the boneresorption rate was slower.

Tissue weight, geometry and strength

Typically, ovariectomy induced a body weight gain in thepresent study. According to Frost's mechanostat theory [38], thisincrement of body weight causes a higher loading on bone, thuspreserving bone strength to a certain extent. In addition, it hasbeen suggested that the ovariectomy-induced higher body weightmight partially compensate for the deterioration of bone quality[39]. Several studies failed to reveal the ovariectomy-inducedloss of bone strength in the midshaft of the femur, which might bedue to the interference of this increased body weight [21,40,41].Therefore, we compared the cross-sectional parameters and longbone three-point bending data among groups using a one-wayanalysis of covariance (ANCOVA) with body weight data as acovariate. Again, alendronate, as well as a 14% onion-enricheddiet, better maintained bone quality in biomaterial properties(Fig. 4A–D), and the 14% ON group even showed a slightlyhigher stiffness than other ovariectomized groups (Fig. 4E).

On the other hand, the effect of ovariectomy on tissue weightwas relatively modest, and moreover, only numeric differenceswere shown between the CON group and the OVX group in

cross-sectional cortex measurements (Table 3). Nevertheless,the ALN and 14% ON groups still showed tendencies towardminor decreases in tissue weight or cross-sectional cortex mea-surements. Cross-sectional parameters, such as cortical thick-ness, cortical area or moment of inertia, would contribute to longbone mechanical properties [30]. However, the significant dif-ferences shown in three-point bending test were not revealed incross-sectional measurements. This inconsistent data mightbe explained by the following: (1) Geometric parameters inmidshaft femur were less affected by ovariectomy, which mightbe due to the lack of a Haversian system and vessels in corticalbone of young adult rats [42–44]; (2) According to the formulaof three-point bending, a certain amount of change in CSMI andcross-sectional diameter would cause a fourfold change inloading values of three-point bending when the tested samplewas an ideal cylinder with an equilibrium material property [30].Therefore, a slight change in cross-sectional parameters wouldcause significant variation in loading values; (3) Other thangeometric factors, it has been suggested that the strength of bonetissue could also be related to compositional measures, such astissue density, collagen content or ash fraction, as well as themicrostructure parameters, including collagen cross-linking orfiber orientation [45,46]. Further studies could be useful indetermining the preventive effects of an onion-enriched diet oncomposition and microstructure of osteoporotic bone.

Possible mechanisms

Limited evidence was available to explain the actions of onionsand other vegetables and fruits. One possible mechanism of bonerestoration by those natural foods might be related to their baseexcess [10,14], which is believed to retain calcium in bones.However, Mühlbauer et al. showed that onion could enhance bonemineral through a pharmacological inhibition on bone resorptionrather than its base excess [19]. The pharmacological inhibitioneffects might be from various extracts of onion, including rutin,and several phytoestrogens (coumestrol, isoflavones). Theseextracts have also been proven to have individual benefits forthe bone of ovariectomy animals [31,47–49]. Recently, anotherextract of onion, gamma glutamyl peptide, has been proven toinhibit osteoclastic activity in vitro, and further animal study isawaited [50]. Aside from the effects of a specific extract, thesynergistic effects of extracts from natural foods could be alsoimportant. As noted by Mühlbauer [51], the working dose of rutinused by Horcajada-Molteni et al. [31] is six times higher than itscontent in 1 g of onion and 700–800 times higher than its content in1 g of leeks or wild garlic, which implies that those extracts mightbe more effective when taken together, rather than separately.

Limitations and applicability in human diet habits

Animals in this study were kept in a controlled environmentand fed with a standard semi-purified diet, which is far from thepractice of real human life. Natural inhibitors contained in ourdaily diets and various lifestyle factors (ex. alcohol consump-tion, smoking etc.) would counteract the benefits from onionand vegetables as well.

1162 T.-H. Huang et al. / Bone 42 (2008) 1154–1163

Though we verified onion as another whole food that mayprevent bone loss in animal osteoporotic model, it would notbe applicable to suggest that people consume large amountsof a specific fruit or vegetable as a strategy to prevent boneloss. Notwithstanding those limitations mentioned above, thereare still ways and clues that encourage us to apply a potentiallyanti-resorptive diet in our lives. Indirect evidence from humanstudies has suggested that eating more fruits and vegetableswould benefit bone health [9–17]. Besides, plenty of vegetablesand herbs commonly consumed by humans have been men-tioned about their similarly inhibitive effects on bone resorptionin animal studies [18–20,22]. Therefore, inclusion of an appro-priate amount of these vegetables and herbs would keep varietyin the daily diet and could be an effective and more practicalway to decrease the incidence of osteoporosis in humans.

Conclusion

In summary, our study showed that high onion-enriched dietsare able to decrease ovariectomy-induced osteopenia anddeterioration of biomechanical strength. Further investigationswould be valuable in investigating the cellular mechanisms ofonion on bone homeostasis. Although it remains unclear howthose natural foods benefit bone health, increasing the intake offruits and vegetables might offer a potentially alternative way toimprove bone health.

Acknowledgments

This study was supported by a grant from the NationalScience Council (NSC-94-2314-B-006-033, Taiwan). Deepappreciation for English editing assistance from Miss Jae Cody.

References

[1] Who are candidates for prevention and treatment for osteoporosis?Osteoporos Int 1997;7:1–6.

[2] Kanis JA, Johnell O, OdenA, Sembo I, Redlund-Johnell I, Dawson A. Long-term risk of osteoporotic fracture inMalmo. Osteoporos Int 2000;11:669–74.

[3] Melton III LJ, Atkinson EJ, O'ConnorMK, O'FallonWM,Riggs BL. Bonedensity and fracture risk in men. J Bone Miner Res 1998;13:1915–23.

[4] Melton III LJ, Chrischilles EA, Cooper C, Lane AW, Riggs BL. Perspective.How many women have osteoporosis? J Bone Miner Res 1992;7:1005–10.

[5] Gullberg B, Johnell O, Kanis JA. World-wide projections for hip fracture.Osteoporos Int 1997;7:407–13.

[6] Goltzman D. Discoveries, drugs and skeletal disorders. Nat Rev DrugDiscov 2002;1:784–96.

[7] Reid IR. Pharmacotherapy of osteoporosis in postmenopausal women:focus on safety. Expert Opin Drug Saf 2002;1:93–107.

[8] Yeh IT. Postmenopausal hormone replacement therapy: endometrial andbreast effects. Adv Anat Pathol 2007;14:17–24.

[9] Chen YM, Ho SC, Woo JL. Greater fruit and vegetable intake is associatedwith increased bone mass among postmenopausal Chinese women. Br JNutr 2006;96:745–51.

[10] New SA, Robins SP, Campbell MK, Martin JC, Garton MJ, Bolton-SmithC. Dietary influences on bone mass and bone metabolism: further evidenceof a positive link between fruit and vegetable consumption and bonehealth? Am J Clin Nutr 2000;71:142–51.

[11] Okubo H, Sasaki S, Horiguchi H, Oguma E, Miyamoto K, Hosoi Y.Dietary patterns associated with bone mineral density in premenopausalJapanese farmwomen. Am J Clin Nutr 2006;83:1185–92.

[12] Prynne CJ, Mishra GD, O'Connell MA, Muniz G, Laskey MA, Yan L.Fruit and vegetable intakes and bone mineral status: a cross sectional studyin 5 age and sex cohorts. Am J Clin Nutr 2006;83:1420–8.

[13] Tucker KL, Chen H, Hannan MT, Cupples LA, Wilson PW, Felson D.Bone mineral density and dietary patterns in older adults: the FraminghamOsteoporosis Study. Am J Clin Nutr 2002;76:245–52.

[14] Tucker KL, Hannan MT, Chen H, Cupples LA, Wilson PW, Kiel DP.Potassium, magnesium, and fruit and vegetable intakes are associated withgreater bone mineral density in elderly men and women. Am J Clin Nutr1999;69:727–36.

[15] Vatanparast H, Baxter-Jones A, Faulkner RA, Bailey DA, Whiting SJ.Positive effects of vegetable and fruit consumption and calcium intakeon bone mineral accrual in boys during growth from childhood to ado-lescence: the University of Saskatchewan Pediatric Bone Mineral AccrualStudy. Am J Clin Nutr 2005;82:700–6.

[16] McGartland CP, Robson PJ, Murray LJ, Cran GW, Savage MJ, WatkinsDC. Fruit and vegetable consumption and bone mineral density: theNorthern Ireland Young Hearts Project. Am J Clin Nutr 2004;80:1019–23.

[17] Tylavsky FA, Holliday K, Danish R, Womack C, Norwood J, Carbone L.Fruit and vegetable intakes are an independent predictor of bone size inearly pubertal children. Am J Clin Nutr 2004;79:311–7.

[18] Muhlbauer RC, Li F. Effect of vegetables on bone metabolism. Nature1999;401:343–4.

[19] Muhlbauer RC, Lozano A, Reinli A. Onion and a mixture of vegetables,salads, and herbs affect bone resorption in the rat by a mechanism inde-pendent of their base excess. J Bone Miner Res 2002;17:1230–6.

[20] Muhlbauer RC, Lozano A, Reinli A, Wetli H. Various selected vegetables,fruits, mushrooms and red wine residue inhibit bone resorption in rats.J Nutr 2003;133:3592–7.

[21] Deyhim F, Stoecker BJ, Brusewitz GH, Devareddy L, Arjmandi BH.Dried plum reverses bone loss in an osteopenic rat model of osteoporosis.Menopause 2005;12:755–62.

[22] Muhlbauer RC, Li F, Lozano A, Reinli A, Tschudi I. Some vegetables(commonly consumed by humans) efficiently modulate bone metabolism.J Musculoskelet Neuronal Interact 2000;1:137–40.

[23] Kalu DN, Liu CC, Hardin RR, Hollis BW. The aged rat model of ovarianhormone deficiency bone loss. Endocrinology 1989;124:7–16.

[24] Azuma Y, Oue Y, Kanatani H, Ohta T, Kiyoki M, Komoriya K. Effects ofcontinuous alendronate treatment on bone mass and mechanical propertiesin ovariectomized rats: comparison with pamidronate and etidronate ingrowing rats. J Pharmacol Exp Ther 1998;286:128–35.

[25] Ito M, Azuma Y, Takagi H, Komoriya K, Ohta T, Kawaguchi H. Curativeeffect of combined treatment with alendronate and 1 alpha-hydroxyvitaminD3 on bone loss by ovariectomy in aged rats. Jpn J Pharmacol 2002;89:255–66.

[26] Huang TH, Yang RS, Hsieh SS, Liu SH. Effects of caffeine and exercise onthe development of bone: a densitometric and histomorphometric study inyoung Wistar rats. Bone 2002;30:293–9.

[27] Parfitt AM, Drezner MK, Glorieux FH, Kanis JA,Malluche H,Meunier PJ.Bone histomorphometry: standardization of nomenclature, symbols, andunits. Report of the ASBMRHistomorphometry Nomenclature Committee.J Bone Miner Res 1987;2:595–610.

[28] Honda A, Umemura Y, Nagasawa S. Effect of high-impact and low-repetition training on bones in ovariectomized rats. J Bone Miner Res2001;16:1688–93.

[29] Huang TH, Lin SC, Chang FL, Hsieh SS, Liu SH, Yang RS. Effects ofdifferent exercise modes on mineralization, structure, and biomechanicalproperties of growing bone. J Appl Physiol 2003;95:300–7.

[30] Turner CH, Burr DB. Basic biomechanical measurements of bone: atutorial. Bone 1993;14:595–608.

[31] Horcajada-Molteni MN, Crespy V, Coxam V, Davicco MJ, Remesy C,Barlet JP. Rutin inhibits ovariectomy-induced osteopenia in rats. J BoneMiner Res 2000;15:2251–8.

[32] Ogawa K, Hori M, Takao R, Sakurada T. Effects of combined elcatoninand alendronate treatment on the architecture and strength of bone inovariectomized rats. J Bone Miner Metab 2005;23:351–8.

[33] da Paz LH, de Falco V, Teng NC, dos Reis LM, Pereira RM, Jorgetti V.Effect of 17beta-estradiol or alendronate on the bone densitometry, bone

1163T.-H. Huang et al. / Bone 42 (2008) 1154–1163

histomorphometry and bone metabolism of ovariectomized rats. Braz JMed Biol Res 2001;34:1015–22.

[34] Frolik CA, Bryant HU, Black EC, Magee DE, Chandrasekhar S. Time-dependent changes in biochemical bone markers and serum cholesterol inovariectomized rats: effects of raloxifene HCl, tamoxifen, estrogen, andalendronate. Bone 1996;18:621–7.

[35] Iwamoto J, Seki A, Takeda T, Sato Y, Yamada H, Yeh JK. Comparativeeffects of alendronate and alfacalcidol on cancellous and cortical bonemass and bone mechanical properties in ovariectomized rats. Exp Anim2006;55:357–67.

[36] Arjmandi BH, Khalil DA, Lucas EA, Georgis A, Stoecker BJ, Hardin C.Dried plums improve indices of bone formation in postmenopausalwomen. J Womens Health Gend Based Med 2002;11:61–8.

[37] Alatalo SL, Peng Z, Janckila AJ, Kaija H, Vihko P, Vaananen HK. A novelimmunoassay for the determination of tartrate-resistant acid phosphatase5b from rat serum. J Bone Miner Res 2003;18:134–9.

[38] Frost HM. Bone “mass” and the “mechanostat”: a proposal. Anat Rec1987;219:1–9.

[39] Peng ZQ, Vaananen HK, Zhang HX, Tuukkanen J. Long-term effects ofovariectomy on the mechanical properties and chemical composition of ratbone. Bone 1997;20:207–12.

[40] Cai DJ, Zhao Y, Glasier J, Cullen D, Barnes S, Turner CH. Comparativeeffect of soy protein, soy isoflavones, and 17beta-estradiol on bonemetabolism in adult ovariectomized rats. J Bone Miner Res 2005;20:828–39.

[41] Chachra D, Lee JM, Kasra M, Grynpas MD. Differential effects ofovariectomy on the mechanical properties of cortical and cancellous bonein rat femora and vertebrae. Biomed Sci Instrum 2000;36:123–8.

[42] Bagi CM, Ammann P, Rizzoli R, Miller SC. Effect of estrogen deficiencyon cancellous and cortical bone structure and strength of the femoral neckin rats. Calcif Tissue Int 1997;61:336–44.

[43] Martiniakova M, Grosskopf B, Omelka R, Vondrakova M, Bauerova M.Differences among species in compact bone tissue microstructure ofmammalian skeleton: use of a discriminant function analysis for speciesidentification. J Forensic Sci 2006;51:1235–9.

[44] Ruth EB. Bone studies. II. An experimental study of the Haversian-typevascular channels. Am J Anat 1953;93:429–55.

[45] Friedman AW. Important determinants of bone strength: beyond bonemineral density. J Clin Rheumatol 2006;12:70–7.

[46] van der Meulen MC, Jepsen KJ, Mikic B. Understanding bone strength:size isn't everything. Bone 2001;29:101–4.

[47] Arjmandi BH, Alekel L, Hollis BW, Amin D, Stacewicz-Sapuntzakis M,Guo P. Dietary soybean protein prevents bone loss in an ovariectomized ratmodel of osteoporosis. J Nutr 1996;126:161–7.

[48] Draper CR, Edel MJ, Dick IM, Randall AG, Martin GB, Prince RL.Phytoestrogens reduce bone loss and bone resorption in oophorectomizedrats. J Nutr 1997;127:1795–9.

[49] Ye SF, Saga I, Ichimura K, Nagai T, Shinoda M, Matsuzaki S. Coumestrolas well as isoflavones in soybean extract prevent bone resorption inovariectomized rats. Endocr Regul 2003;37:145–52.

[50] Wetli HA, Brenneisen R, Tschudi I, Langos M, Bigler P, Sprang T. Agamma-glutamyl peptide isolated from onion (Allium cepa L.) by bioassay-guided fractionation inhibits resorption activity of osteoclasts. J Agric FoodChem 2005;53:3408–14.

[51] Muhlbauer RC. Rutin cannot explain the effect of vegetables on bonemetabolism. J Bone Miner Res 2001;16:970–1.