Bone Has a Sexually Dimorphic Response to Aromatase Deficiency

JOURNAL OF BONE AND MINERAL RESEARCH Volume 15, Number 3, 2000 r 2000 American Society for Bone and Mineral Research

Bone Has a Sexually Dimorphic Response to Aromatase Deficiency ¨ Z,1 JOSEPH E. ZERWEKH,2 CAROLYN FISHER,3 KATHY GRAVES,4 LYDIA NANU,1 ORHAN K. O RITA MILLSAPS,2 and EVAN R. SIMPSON5

ABSTRACT Aromatase synthesizes estrogen from androgen precursors. To better understand the role of estrogen in skeletal metabolism and growth, we have assessed long bone growth and histomorphometry in aromatase-deficient (ArKO) mice. The age range for the animals was 5–7 months. At this age mice have already achieved peak bone density but continue slow bone growth. Femur length, an index of long bone growth, showed decreased growth in ArKO males compared with wild-type (wt) littermates but no significant difference in females. Radiographically, compared with age- and sex- matched littermates both ArKO males and females showed osteopenia in the lumbar spine. Histologically, both ArKO males and females showed an osteoporotic-type picture, characterized by significant decreases in trabecular bone volume and trabecular thickness. However, compared with wt littermates female ArKO animals showed a bone remodeling picture consistent with increased bone turnover, much like early postmenopausal osteoporosis in humans. On the other hand, male ArKO animals showed decreases in both osteoblastic and osteoclastic surfaces compared with wt littermates, similar to age-related osteopenia. These findings suggest that osteoporosis seen in aromatase-deficient mice may arise from different bone remodeling activities between males and females. These results also show that the ArKO model exhibits the expected results of estrogen deficiency and may be a good model for investigating sex-specific responses to estrogen deficiency. Furthermore, they imply that estrogen is important for attaining peak bone mass in male as well as in female mice. (J Bone Miner Res 2000;15:507–514) Key words:

aromatase, estrogen, osteoporosis, mice, bone

INTRODUCTION

T

dynamic organ in which mineralized bone is resorbed continuously by osteoclasts and new bone is formed by osteoblasts. Hormones, such as sex steroids, are known to be important regulators of bone cell function, and alterations in their levels can lead to abnormal bone metabolism and growth. In particular, estrogen deficiency caused by natural or surgically induced menopause results in increased bone turnover predisposing to osteoporosis and is the single most common cause of osteoporosis. Osteoporosis is characterized by low trabecular bone mass HE SKELETON IS A

and microarchitectural deterioration with a consequent increase in bone fragility and susceptibility to fractures. The exact mechanism of bone loss in estrogen deficiency is not known nor are the molecular mechanisms for the protective effects of estrogens known. The importance of estrogen in maintaining bone mass in women is firmly established. However, until recently, a similar role for estrogen in men has received little consideration. The importance of estrogen in bone metabolism of men has been dramatically highlighted by the descriptions of a man with estrogen receptor (ER) deficiency (estrogen resistance) and two men with aromatase deficiency (estrogen

1Department

of Radiology, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, U.S.A. of Internal Medicine, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, U.S.A. 3School of Pharmacy, Texas Tech University Health Sciences Center, Amarillo, Texas, U.S.A. 4Department of Biochemistry, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas, U.S.A. 5Victorian Breast Cancer Consortium, Prince Henry’s Institute of Medical Research, Clayton, Victoria, Australia. 2Department

507

¨ Z ET AL. O

508

deficiency).(1–3) All three of these individuals exhibited failure of epiphyseal closure and sustained linear growth through and beyond puberty into their mid-to-late 20s. Furthermore, careful analysis of growth curves from these individuals suggested that either estrogen deficiency or estrogen resistance also was associated with an absence of the pubertal growth spurt.(4) One of the latter males has been treated with estrogen replacement therapy for 3 years, which has resulted in skeletal maturation with an increase in bone mass, further documenting an important role for estrogen in male bone metabolism.(5) In the other, treatment with regular intramuscular testosterone produced no benefit but treatment with transdermal estradiol resulted in skeletal maturation as shown by a rapid increase in lumbar bone mineral density (BMD) and epiphyseal plate closure.(6) In a study on the effects of androgen supplementation in eugonadal men treated with testosterone, changes in spine BMD significantly correlated with a change in serum estradiol but not with a change in serum testosterone.(7) In addition, recent epidemiological studies of aging men have shown that circulating estrogens or bioavailable serum estrogen levels were more strongly associated with BMD than androgen levels.(8–10) Thus, evidence is accumulating that estrogen is important for normal bone growth and mineralization in human males. Unfortunately, more detailed study of the effects of estrogen in such men is limited by the rarity of the defects. In addition, human studies are always limited by the ability to recruit patients and follow them long term. Thus, animal models are important to further understanding of estrogen’s effect(s) on bone. Historically, such models have included animals with surgical gonadectomy, ER-␣ gene inactivation or chemical inhibition of aromatase, the enzyme responsible for synthesis of estrogens from androgen precursors. Surgical models are limited in that the surgery is nontrivial and they do not account for the effects of estrogen before surgery. Recently, a second form of ER has been identified, namely the ER-␤.(11) Furthermore, there are now known to be at least two splice variants of the ER-␤.(12–14) Therefore, the ER-␣ gene inactivation model is now known not to be a model of complete loss of estrogen function in bone. Furthermore, the results of using an aromatase inhibitor in animal studies are complicated by possible toxicity to the drug as evidenced by the animals’ decreased growth.(15) Recently an aromatase-deficient mouse model (ArKO) has been created by means of homologous recombination.(16) Because this model represents estrogen deficiency from conception it is ideally suited for studies on the effects of estrogen deficiency on bone growth, peak bone mass, and metabolism. We hypothesized that aromatase deficiency would alter long bone growth and cause osteoporosis. To this end, we measured femur length as an index of long bone growth, quantitated markers of bone formation and resorption to assess global bone metabolism, and performed quantitative static and dynamic histomorphometry to evaluate cancellous bone metabolism in adult ArKO animals, aged 5–7 months. At this age mice have already achieved peak bone density but continue very slow bone growth.(17)

MATERIALS AND METHODS Animals The targeted disruption of the aromatase gene and initial phenotypic characterization of the mice has been previously reported.(16) Aromatase-deficient mice and their wild-type (wt) littermates, aged 5–7 months, were obtained from our colony maintained at the University of Texas Southwestern Medical Center at Dallas, TX, U.S.A.. In this age range we could observe the consequences of estrogen deficiency on long bone growth while studying its effects on peak bone density and remodeling with minimal confounding effects from bone growth.(17) All animals were maintained and used in accordance with the National Institutes of Health guidelines on the care and use of laboratory animals. The aromatase inactivation was maintained on the hybrid J129/ Sv/C57BL/6J background. F2 generation animals were used in this study.

Plain film radiography Small focal spot imaging of sex-matched ArKO and wt littermates, with attention to the spinal column, was performed on a mammography unit located in the Department of Radiology at Parkland Memorial Hospital, Dallas, TX, U.S.A. Animals also were size-matched as closely as possible to minimize contributions of soft tissue to apparent differences in radiographic density.

Static and dynamic histomorphometry Ten days before death, animals were given a single intraperitoneal dose of tetracycline at 30 mg/kg. Six days later a second dose was given. After another 4 days the animals were killed for analysis. Uteri were harvested for weight determination. Femurs were harvested for bone length determinations using calipers. For histological comparison of trabecular bone volume, one femur was demineralized in 10% buffered EDTA for 2 weeks. Subsequently, the samples were dehydrated in a graded series of alcohol, embedded in paraffin and 4 µm sections were made for hematoxylin and eosin staining. The entire lumbar spine was harvested for histomorphometry and stored in 70% ethanol until analysis. After dehydration in a graded series of alcohol, the bones were processed undecalcified in methylmethacrylate as previously described.(18) The histomorphometric examination was performed with a computer and a digitizing tablet interfaced to an Aus Jena microscope with a drawing tube attachment and bone histomorphometry software (BioQuant-R&M Biometrics, Nashville, TN, U.S.A.). Static measurements of cancellous bone volume and thickness and cellular parameters were made on toluidine blue– stained sections. Fluorochrome-based indices of bone formation were measured in unstained sections. Measurements were performed on at least two vertebral bodies for each animal. The terminology used is that recommended by the histomorphometry nomenclature committee of the American Society for Bone and Mineral Research.(19) Data were collected without knowledge of the sex or genotype of any of the animals.

BONE PHENOTYPE OF AROMATASE-DEFICIENT MICE

Analytical methods Serum estrogen levels were measured using a commercially available radioimmunoassay (RIA) kit (Diagnostic Products Corp., Los Angeles, CA, U.S.A.). To enhance the sensitivity of the method for murine samples, ether extraction of serum was performed according to manufacturer’s instructions. The large volume required (500 µl) necessitated pooling of samples in most cases. To test this assay for specificity, an aqueous extract was made from an amount of mouse chow equivalent to daily consumption. Before analysis this aqueous solution was subjected to the same ether extraction procedure as the serum samples. Serum testosterone levels were measured in the university’s clinical laboratory using an Automated Chemiluminescence System (Chiron Diagnostics, East Walpole, MA, U.S.A.). Serum estrone levels were determined by double antibody RIA (Diagnostics Systems Laboratories, Inc., Webster, TX, U.S.A.). The lowest standard available for this assay is 15 pg/ml. Serum osteocalcin levels were measured using an RIA kit specific for mouse osteocalcin (Biomedical Technologies, Inc., Stoughton, MA, U.S.A.). Deoxypyridinoline (DPD) crosslink levels in 24-h urine samples were measured with a Pyrilinks-D ELISA kit, which uses antibody that crossreacts with the mouse analyte (Metra Biosystems, Mountain View, CA, U.S.A.).

Statistical analysis All data are expressed as mean ⫾ SD. For the femur length measurements and metabolic marker levels, significance of difference between wt and aromatase-deficient littermates was assessed by a two-tailed t-test. Histomorphometric data, when normally distributed, were analyzed by a pooled variance or separate variance t-test for normal or unequal variances, respectively. For nonnormally distributed data, the Mann-Whitney test was utilized. In all instances, p ⬍ 0.05 was taken to be significant.

RESULTS Serum steroid hormone levels Serum hormone levels reported in the original report on the ArKO animals were obtained on younger animals than in our study.(16) Therefore, we determined estradiol, estrone, and testosterone levels in the age range of animals in our study. For estradiol determination in animal specimens the manufacturer recommends an extraction procedure to decrease the interference from nonestrogen components in the serum and achieve an assay matrix for the samples that is the same as that of the standards. Following this recommendation, we analyzed serum from three pools of wt females, three pools of ArKO females, two pools of wt males, and two pools of ArKO males. We did not attempt to bias the data by sampling wt females only when they were in the high estradiol state of proestrous phase of the ovarian cycle. As reported previously, most samples were at the limits of detection (Table 1).(16) We found detectable levels in one pool of wt females, one pool of wt males, and ArKO males.

509

TABLE 1. ESTRADIOL LEVELS IN WILD-TYPE AND ArKO MICE Pool

N (animals in pool)

Concentration (pg/ml)

2 2 2 2 4

36 ⬍5 ⬍5 5.2 ⬍5

5 3 1 2 —

5.4 15 9 7.5 18

Females Wild-type Wild-type Wild-type ArKO ArKO Males Wild-type Wild-type ArKO ArKO Chow Extract

TABLE 2. ESTRONE LEVELS IN WILD-TYPE AND ArKO MICE Genotype

Concentration ⫾ SD (pg/ml)

p*

29.1 ⫾ 5.9 23.7 ⫾ 3.3

0.10

Females Wild-type (n ⫽ 7) ArKO (n ⫽ 3) Males Wild-type (n ⫽ 5) ArKO (n ⫽ 6)

16.0 ⫾ 6.3 19.9 ⫾ 2.4

0.25

*By t-test.

More importantly, when we prepared an extract from the estimated mass of chow one animal consumes per day, we found an ‘‘estradiol’’ concentration exceeding that obtained for all our specimens except for the one pool of wt females (Table 1). Mean serum estrone levels were lower in female ArKO animals compared with wt animals but not yet significantly different (Table 2). The estrone levels from male animals were at the level of the lowest standard and were not significantly different. The soy isoflavone metabolite equol reacts with this assay and likely accounts for the measurable levels of estrone in the ArKO females (E. Simpson, personal communication, March 1999). As another test of estrogen deficiency, we measured uterine weights in females. The mean uterine weight in the ArKO females was decreased by 75% of the wt uterine weight (Table 3), providing in vivo evidence of estrogen deficiency. As before(16) mean testosterone levels were higher in ArKO females. Although mean testosterone levels of ArKO males were higher than wt animals, this was not significantly different because of the wide variation in values (Table 4).

Bone growth: femur length Femur length was used an index of long bone growth. Animals in our study had attained most of their anticipated long bone growth. Mean femur length of male ArKO

¨ Z ET AL. O

510

TABLE 3. UTERINE WEIGHTS IN WILD-TYPE AND ArKO MICE Genotype (n)

Mean ⫾ SD (mg)

p-value*

Wild-type (n ⫽ 5) ArKO (n ⫽ 6)

128.6 ⫾ 30.9 32.6 ⫾ 5.9

0.002

*ArKO significantly different from wild-type by t-test.

TABLE 4. TESTOSTERONE LEVELS IN WILD-TYPE AND ArKO MICE Genotype Females Wild-type (n ⫽ 3) ArKO (n ⫽ 2) Males Wild-type (n ⫽ 8) ArKO (n ⫽ 7)

Concentration ⫾ SD (pg/ml)

p*

128.3 ⫾ 10.4 2050 ⫾ 636.4

ND

21026.5 ⫾ 46517.3 37908.1 ⫾ 50974.9

0.5

ND, not determined.

TABLE 5. MEAN FEMUR LENGTH IN ADULT MICE Genotype

Length ⫾ SD (mm)

Males Wild-type (n ⫽ 15) ArKO (n ⫽ 49) Females Wild-type (n ⫽ 17) ArKO (n ⫽ 23)

16.3 ⫾ 0.5 14.9 ⫾ 0.5† 16.3 ⫾ 0.13 16.3 ⫾ 0.14

*p ⬍ 0.001 for ArKO versus wild-type by the t-test.

animals was 9% shorter than wt littermates (p ⬍ 0.001). There was no significant difference between ArKO females and their wt littermates (Table 5).

Bone density: radiographic examination To detect whether there might be gross differences in bone density, plain film radiography was performed. Images obtained in the lateral projection showed osteopenia in age-matched male and female ArKO animals compared with wt littermates (Fig. 1).

Bone metabolism: histomorphometric examination and metabolic markers To determine whether the radiographic osteopenia reflected osteoporosis, the spines were evaluated further. Quantitative static and dynamic histomorphometry were undertaken to further characterize the changes in trabecular bone structure, surface, and dynamic parameters. Compared with wt littermates, ArKO animals of both sexes showed decreases in trabecular bone volume and thickness as well as reductions in cortical thickness (Figs. 2A–2D and Tables 6

FIG. 1. Plain film radiography in wt and ArKO Mice. Radiographs of mice centered on the spinal columns, taken in the lateral projection, are shown. (A) Male wt; (B) male ArKO; (C) female wt; (D) female ArKO. Notice that lumbar vertebral bodies in ArKO animals are less dense (not as white and there is greater distinction between the vertebral bodies and their end plates).

and 7). The difference in mean trabecular bone volume for the male animals was 14% (p ⫽ 0.036) compared with a 39% (p ⬍ 0.001) difference in the female animals. Compared with wt littermates, mean trabecular thickness was decreased by 13% in male ArKO mice (p ⫽ 0.027) and 14% for female ArKO animals (p ⫽ 0.001). Mean cortical thickness was reduced by 7% in ArKO male mice (p ⫽ 0.630). In female ArKO mice, mean cortical thickness was decreased by 25% as compared with wt littermates (p ⫽ 0.001). Surface based parameters also disclosed significant differences between male ArKO animals and their wt littermates. Male ArKO animals had significant reductions in osteoblastic, osteoid, and eroded surfaces (Table 6; illustrated in Figs. 3A,a and 3B,b). Although mineral apposition rates were comparable between male ArKO and wt littermates, the ArKO animals showed significant reductions in the extent of mineralizing surfaces as observed from tetracycline uptake. This was true for both the total and double-labeled surfaces. On the other hand, there was a general tendency for the surface based parameters to be increased for female ArKO animals as compared with wt littermates (Table 7; illustrated in Figs. 3C,c and 3D,d). The only significant difference was observed for doubled labeled surfaces. Mean serum osteocalcin concentration in male ArKO animals was significantly less than that observed for wt littermates (p ⫽ 0.003, Table 8). There was no significant

BONE PHENOTYPE OF AROMATASE-DEFICIENT MICE

511

TABLE 6. HISTOMORPHOMETRIC FINDINGS IN MALE WILD-TYPE MICE AND THEIR ArKO LITTERMATES

TABLE 8. METABOLIC MARKERS

Genotype Parameter

Wild-type (n ⫽ 13)

ArKO (n ⫽ 18)

C. Th (µm) BV/TV (%) OV/BV (%) Tb.Th (µm) OS/BS (%) Ob.S/BS (%) ES/BS (%) Oc.S/BS (%) MS/BS8 (%) MS/BS (2X%) MAR (µm/day) O. Th (µm) Age (months)

92 ⫾ 16.2 ⫾ 2.1 0.34 ⫾ 0.44 32 ⫾ 3 3.0 ⫾ 1.5 2.2 ⫾ 1.6 3.6 ⫾ 1.3 0.6 ⫾ 0.5 12.0 ⫾ 4.7 2.7 ⫾ 2.0 1.5 ⫾ 0.3 3.3 ⫾ 1.2 5.8 ⫾ 1.2

86 ⫾ 27 13.9 ⫾ 3.6 0.30 ⫾ 0.41 28 ⫾ 5 1.4 ⫾ 1.3 1.1 ⫾ 1.1 2.4 ⫾ 0.9 0.3 ⫾ 0.30 8.9 ⫾ 9.7 0.3 ⫾ 0.5 1.3 ⫾ 0.5 3.2 ⫾ 0.6 6.0 ⫾ 1.2

43a

p 0.630 0.036b 0.201 0.027c 0.007d 0.043d 0.013d 0.125 0.003d 0.0001d 0.334 0.083

8 Denotes sum of both single- and double-labeled surfaces. aAll values expressed as mean ⫾ standard deviation. bArKO significantly different from wild-type by unequal variance t-test. cArKO significantly different from wild-type by equal variance t-test. dArKO significantly different from wild-type by Mann-Whitney nonparametric test.

TABLE 7. HISTOMORPHOMETRIC FINDINGS IN FEMALE WILD-TYPE MICE AND THEIR ArKO LITTERMATES Genotype Parameter

Wild-type (n ⫽ 14)

ArKO (n ⫽ 16)

C. Th (µm) BV/TV (%) OV/BV (%) Tb. Th (µm) OS/BS (%) Ob. S/BS (%) ES/BS (%) Oc. S/BS (%) MS/BS8 (%) MS/BS (2X%) MAR (µm/day) O. Th (µm) Age (month)

118 ⫾ 25a 18.9 ⫾ 4.7 0.46 ⫾ 0.44 29 ⫾ 2 4.0 ⫾ 2.8 2.9 ⫾ 2.3 4.2 ⫾ 1.3 0.85 ⫾ 0.76 13.2 ⫾ 6.5 2.2 ⫾ 1.2 2.1 ⫾ 1.6 3.5 ⫾ 0.9 5.9 ⫾ 1.1

89 ⫾ 20 11.5 ⫾ 7.6 0.92 ⫾ 1.0 25 ⫾ 2 5.9 ⫾ 3.5 4.1 ⫾ 3.4 4.7 ⫾ 2.3 1.14 ⫾ 1.06 17.2 ⫾ 6.3 6.0 ⫾ 3.6 1.6 ⫾ 0.2 3.2 ⫾ 0.6 6.2 ⫾ 0.9

p 0.001b 0.001b 0.075 0.001b 0.073 0.271 0.473 0.545 0.101 0.014b 0.273 0.393

8 Denotes sum of both single- and double-labeled surfaces. aAll values expressed as mean ⫾ standard deviation. bArKO significantly different from wild-type by Mann-Whitney nonparametric test.

difference in mean serum osteocalcin levels between wt and ArKO females. Mean urinary DPD levels were not significantly different between male and female ArKO animals and their respective wt littermates.

Males Wild-type ArKO Females Wild-type ArKO

Serum osteocalcin (ng/ml)

Urine DPD (nmol/mg creatinine)

70 ⫾ 27 (n ⫽ 11) 38 ⫾ 15 (n ⫽ 17)*

68 ⫾ 16 (n ⫽ 17) 79 ⫾ 36 (n ⫽ 12)

105 ⫾ 41 (n ⫽ 12) 105 ⫾ 35 (n ⫽ 10)

128 ⫾ 19 (n ⫽ 8) 154 ⫾ 55 (n ⫽ 13)†

*p ⫽ 0.003 for ArKO versus wild-type by t-test. †p ⫽ 0.12 for ArKO versus wild-type by t-test.

DISCUSSION Previous studies of the role of sex hormones on bone remodeling have focused on the relationship between circulating levels of these steroid hormones and osteoclast/ osteoblast activity. Aromatase, the enzyme that catalyzes the conversion of androgens to estrogen, recently has been reported to be expressed in osteoblast-like cells and osteoclast-like cells as well as in normal human bone tissue ¨ z, unpublished results March 1999).(20–25) This raises (O. O the possibility that peripheral conversion of androgens to estrogen or local production of estrogen in bone also may have an autocrine or paracrine role in the maintenance of normal bone growth and mass. In the present study, targeted disruption of the aromatase gene resulted in a bone phenotype best described as osteoporosis. Osteopenia was evident on plain film radiographs and spinal osteoporosis was confirmed by quantitative histomorphometry. Similar histological findings also have been found in the distal femur metaphysis in a limited number of male and female animals that we have examined suggesting this bone loss to be a generalized skeletal response to aromatase deficiency. Reductions in cancellous bone volume and in cortical and trabecular thickness were evident in both genders of ArKO animals. What was most surprising was the sexually dimorphic difference by which aromatase deficiency resulted in loss of bone mass. For males, aromatase deficiency contributed to a clear suppression of bone formation as shown by significant reductions in osteoblastic, osteoid, and mineralizing surfaces. This finding also was supported by the significant decrease in serum osteocalcin level observed for the ArKO males. In contrast, urinary DPD concentrations were not significantly different between wt and ArKO males. This observation might be considered inconsistent with the histological data that support decreased bone turnover. However, it should be pointed out that the skin in the ArKO mice is abnormal with increased thickness of the epidermis and decreased subcutaneous adipose tissue (E. Simpson, unpublished observations, February 1997). It might be speculated that abnormal dermal collagen metabolism, including post-translational modification, may be contributing to urinary DPD in the ArKO males resulting in elevated urinary concentrations of this marker. Note that rodent skin has been shown to contain DPD.(26) Interestingly, a similar

512

pattern of a decrease in the concentration of a bone formation marker, bone specific alkaline phosphatase, and an increase in bone resorption markers N-telopeptide and C-telopeptide cross-links was recently reported in older men treated with the aromatase inhibitor anastrozole.(27) Thus, in this respect, our results for the biochemical markers in the male ArKO mouse mimic those observed for the chemically induced aromatase-deficient human male. Female ArKO mice showed a histological picture more consistent with increased bone turnover. Although most of the surface based changes in female ArKO animals were not significant, the consistent trend for these parameters to all be increased, as well as the nearly 3-fold increase in double-

FIG. 2. Histologic comparison of trabecular bone volume and thickness in wt and ArKO mice. Coronal sections of decalcified femurs stained with hematoxylin and esoin to demonstrate bone are shown. (A) Male wt; (B) male ArKO; (C) female wt; (D) female ArKO. Notice that compared with their wt littermates, metaphysis from ArKO males and females have decreased trabecular thickness and trabecular bone volume. Magnification is ⫻62.

¨ Z ET AL. O

labeled surfaces support this contention. The failure to reach significance for the majority of the surface based parameters and for the DPD concentrations in the female mice may be a reflection of small animal numbers in our study. It also could be speculated that our wt females had already entered a stage of increased bone remodeling thereby decreasing differences between the two groups. Alternatively, perhaps there is a blunted response to estrogen deficiency in our ArKO females. Such a possibility might be secondary to phytoestrogens in the diet, which have been shown to prevent bone loss in ovariectomized female rodents.(28) Static and dynamic histomorphometry on bones of animals maintained on a no-soy (low phytoestrogens) diet as well as evaluation of younger animals are underway to resolve these issues. Differences in long bone growth also were noted with ArKO males showing decreased femur length, whereas the female ArKO animals did not. The shortened femur length in adult male ArKO animals versus ArKO females suggests that estrogen may have sex-specific effects on bone growth as well as on bone remodeling dynamics. Further, it is different from observations for aromatase-deficient humans in which adult males were tall but skeletally immature. It is important to note that these patients were not tall because of excess growth but rather because of failure to close their epiphyseal plates. More recently, it has been suggested that estrogen deficiency and estrogen resistance actually result in absence of the pubertal growth spurt.(4) Thus, they continue to grow at the slow prepubertal rate throughout adulthood. Therefore, if estrogen-deficient patients were to close their growth plates, we would expect them to have shorter long bones as in our ArKO males or, at best, attain the height of their genetic potential. Why ArKO females do not have shorter bones is not clear. One possibility is that the high androgen levels present in young ArKO animals enhance bone growth in females but decrease it in males.(16) This hypothesis is currently being tested by castration and androgen inhibition experiments in younger animals.

FIG. 3. Histologic comparison of surface parameters in wt and ArKO mice. (A–D) Low and (a–d) high power magnifications of trabecular bone surfaces are shown to demonstrate osteoblasts (ob) with subjacent osteoid (os) and osteoclasts (oc) covering the bone surface. (A,a) Male wt; (B,b) male ArKO; (C,c) female wt; (D,d) female ArKO. Notice that while compared with male mice, females have more bone cells covering surfaces, the female ArKO animals show the greatest osteoblastic and osteoclastic surfaces (panel D,d). Note that ArKO males show few active trabecular surfaces compared with wt males. (A–D) Magnification ⫻250; (a–d) magnification ⫻1000.

BONE PHENOTYPE OF AROMATASE-DEFICIENT MICE

Estrogen deficiency in postmenopausal women and in ovariectomized animals results in trabecular bone volume loss. The loss is thought to result from bone resorption exceeding bone formation. At the cellular level there is an increase in both osteoblastogenesis and osteoclastogenesis that results in increases in both osteoblastic and osteoclastic surfaces. This is the phenotype observed in adult female ArKO animals. The skeletal effects of estrogen deficiency in males are less well characterized. In human males with aromatase deficiency osteopenia and delayed skeletal maturation are observed.(2,3) These are corrected with estrogen replacement therapy but not androgen therapy.(5,6) In one of these patients a transiliac biopsy was reported to show evidence of increased remodeling but the biochemical markers were normal.(6) In the other patient no biopsy was reported but serum alkaline phosphatase and osteocalcin were elevated suggesting a high turnover state.(5) Our ArKO male animals also showed osteopenia secondary to osteoporosis but did not show evidence of a high turnover state. In contrast serum osteocalcin concentration was significantly decreased in ArKO male mice consistent with the low turnover histological picture. This difference between aromatase-deficient men and ArKO male mice may be attributable to species differences. Alternatively, our ArKO male mice may have represented a later stage of disease. We are testing this possibility by studying younger animals. There are several reports that androgens have stimulatory effects on bone density and mechanical strength in female rats, nonhuman primates, and in humans (reviewed in Refs. 29 and 30). Thus, one could argue that the ArKO mice bone phenotype, as it relates to trabecular bone, is secondary to androgen resistance. Arguing against androgen resistance in ArKO animals is the fact that ArKO males have enlarged prostate glands, as might be expected from androgen stimulation.(16) Therefore, the ArKO bone phenotype appears to be that of estrogen deficiency not of androgen excess coupled with androgen resistance. In conclusion, ArKO male and female adult mice show an osteoporotic phenotype and should be a good model for studies on mechanisms of estrogen-deficiency–mediated bone loss. A strength of the model is that it is amenable to modulation of the bone phenotype by exogenous sources of estrogen, a benefit not possible with current estrogenresistant models. This will allow us to identify factors regulated by the estrogen status in both male and female mice and how such regulation may differ between genders.

ACKNOWLEDGEMENTS The authors thank Dr. Thomas Lane and Dr. Jon Anderson of the Department of Radiology University of Texas Southwestern Medical Center at Dallas for assistance in defining the optimal conditions for the plain film imaging. This study was supported by USPHS grant R37AG08174 and, in part, by funds from the Effie and Woffard Cain Chair endowment, institutional funds from the Center for Mineral Metabolism and Clinical Research, and by a grant from the Victorian Breast Cancer Research Consortium.

513

REFERENCES 1. Smith EP, Boyd J, Frank GR, Takahashi H, Cohen RM, Specker B, Williams TC, Lubahn DB, Korach KS 1994 Estrogen resistance caused by a mutation in the estrogenreceptor gene in a man. N Engl J Med 331:1056–1061. 2. Morishima A, Grumbach MM, Simpson ER, Fisher C, Qin K 1995 Aromatase deficiency in male and female siblings caused by a novel mutation and the physiological role of estrogens. J Clin Endocrinol Metab 80:3689–3698. 3. Qin K, Fisher CR, Simpson ER, Serpente S, Faustini-Fustini M, Carani C 1996 Oligospermia and persistent linear growth of a male subject caused by a missense mutation in the gene encoding aromatase. Proceedings of the IVth International Aromatase Conference, Tahoe City, CA, U.S.A. 4. MacGillivray MH, Morishima A, Conte F, Grumbach M, Smith EP 1998 Pediatric endocrinology update: An overview. The essential roles of estrogens in pubertal growth, epiphyseal fusion and bone turnover: Lessons from mutations in the genes for aromatase and the estrogen receptor. Horm Res 49:(Suppl 1):2–8. 5. Bilezikian JP, Morishima A, Bell J, Grumbach MM 1998 Increased bone mass as a result of estrogen therapy in a man with aromatase deficiency. N Engl J Med 339:599–603. 6. Carani C, Qin K, Simoni M, Faustini-Fustini M, Serpente S, Boyd J, Korach KS, Simpson ER 1997 Effect of testosterone and estradiol in a man with aromatase deficiency. N Engl J Med 337:91–95. 7. Anderson FH, Francis RM, Selby PL, Cooper C 1998 Sex hormones and osteoporosis in men. Calcif Tissue Int 62:185– 188. 8. Slemenda CW, Longcope C, Zhou L, Hui SL, Peacock M, Johnston CC 1997 Sex steroids and bone mass in older men: Positive associations with serum estrogens and negative associations with androgens. J Clin Invest 100:1755–1759. 9. Ongphiphadhanakul B, Rajatanavin R, Chanprasertyothin S, Piaseu N, Chailurkit L 1998 Serum oestradiol and oestrogenreceptor gene polymorphism are associated with bone mineral density independently of serum testosterone in normal males. Clin Endocrinol 49:803–809. 10. Khosla S, Melton LJ III, Atkinson EJ, O’Fallon WM, Klee GG, Riggs BL 1998 Relationship of serum sex steroid levels and bone turnover markers with bone mineral density in men and women: A key role for bioavailable estrogen. J Clin Endocrinol Metab 83:2266–2274. 11. Kuiper GG, Enmark E, Pelto-Huikko M, Nilsson S, Gustafsson JA 1996 Cloning of a novel receptor expressed in rat prostate and ovary. Proc Natl Acad Sci U S A 93:5925–5930. 12. Hanstein B, Liu H, Yancisin MC, Brown M 1999 Functional analysis of a novel estrogen receptor-␤ isoform. Mol Endocrinol 13:129–137. 13. Chu S, Fuller PH 1997 Identification of a splice variant of the rat estrogen receptor beta gene. Mol Cell Endocrinol 132:195– 199. 14. Petersen DN, Kalcevic GTT, Koza-Taylor PH, Turi TG, Brown TA 1998 Identification of estrogen receptor beta2 a functional variant of estrogen receptor beta expressed in normal rat tissues. Endocrinology 139:1082–1092. 15. Vanderschueren D, van Herck E, Nijs J, Ederveen AGR, de Coster R, Bouillon R 1993 Aromatase inhibition impairs skeletal modeling and decreases bone mineral density in growing male rats. Endocrinology 138:2301–2307. 16. Fisher CR, Graves KH, Parlow AF, Simpson ER 1998 Characterization of mice deficient in aromatase (ArKO) because of targeted disruption of the cyp19 gene. Proc Natl Acad Sci U S A 95:6965–6970. 17. Beamer WG, Donahue LR, Rosen CJ, Baylink DJ 1996 Genetic variability in adult bone density among inbred strains of mice. Bone 18:397–403. 18. Zerwekh JE, Sakhaee K, Breslau NA, Gottschalk F, Pak CYC 1992 Impaired bone formation in male idiopathic osteoporosis: Further reduction in the presence of concomitant hypercalciuria. Osteoporos Int 2:128–134.

514 19. Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H, Meunier PJ, Ott SM, Recker RR 1987 Bone histomorphometry: standardization of nomenclature, symbols and units. J Bone Miner Res 2:595–610. 20. Shozu M, Simpson ER 1998 Aromatase expression of human osteoblast-like cells. Mol Cell Endocrinol 139:117–129. 21. Nawata J, Tanaka S, Tanaka S, Takayanagi R, Sakai Y, Yanase T, Ikuyama S, Haji M 1995 Aromatase in bone cell: assocation with osteoporosis in postmenopausal women. J Steroid Biochem Mol Biol 53:165–174. 22. Eyre LJ, Bland R, Bujalska J, Sheppard M, Stewart PM, Hewison M 1998 Characterization of aromatase and 17␤hydroxysteroid dehydrogenase expression in rat osteoblastic cells. J Bone Miner Res 13:996–1004. 23. Shozu M, Zhao Y, Simpson ER 1997 Estrogen biosynthesis in THP1 cells is regulated by promoter switching of the aromatase (CYP19) gene. Endocrinology 138:5125–5135. 24. Schweikert H-U, Wolf L, Romalo G 1995 Oestrogen formation from androstenedione in human bone. Clin Endocrinol 43: 37–42. 25. Sasano H, Uzuki M, Sawai T, Nagura H, Matsunaga G, Kashimoto O, Harada N 1997 Aromatase in human bone tissue. J Bone Miner Res 12:1416–1423. 26. Grazioli V, Massimo A, Stenico A, Casari E 1996 Urinary output of hydroxylysine glycosides and pyridinium cross-links in detecting rat bone collagen turnover rate. FEBS Lett 388:134–138.

¨ Z ET AL. O 27. Taxel P, Raisz LG 1998 The effect of an aromatase inhibitor on bone turnover in older men. Bone 23(Suppl):S178 (abstract). 28. Arjmandi BH, Alekel L, Hollis BW 1996 Dietary soybean protein prevents bone loss in an ovarectomized rat model of osteoporosis. J Nutr 126:161–167. 29. Orwoll ES 1996 Androgens. In: Bilezikian JP, Raisz LG, Rodan GA (eds.) Principles of Bone Biology. Academic Press, San Diego, CA, U.S.A., pp. 563–580. 30. Vanderschueren D, Boonen S, Bouilllon R 1998 Action of androgens versus estrogens in male skeletal homeostasis. Bone 23:391–394.

Address reprint requests to: ¨ z, MD, Ph.D. Orhan K. O Advanced Radiological Sciences Department of Radiology University of Texas Southwestern Medical Center at Dallas 5323 Harry Hines Boulevard Dallas, TX 75235-9153 U.S.A. Received in original form February 19, 1999; in revised form August 24, 1999; accepted October 12, 1999.