Neonatal Estrogen Exposure Results in Biphasic Age-Dependent Effects on the Skeletal Development of Male Mice

Abstract

Peak bone mass, one of the most important predictors for fracture risk later in life, is attained during puberty and adolescence and influenced by neonatal and pubertal sex-specific gonadal hormones and GH-IGF-I secretion patterns. This study examined the effects of brief neonatal estrogen (NE) exposure on growth and skeletal development in C57BL/6J mice. A single injection of 100-μg estradiol or vehicle was administered on the first day of life. Growth parameters were monitored and skeletal phenotyping performed at 16 weeks in female mice and at 4 and 16 weeks in male mice. NE exposure negatively impacted adult femoral length in both sexes, but adult body weight, areal bone density, and bone strength in female mice were unaffected. In contrast, somatic growth was attenuated in estrogen-exposed male mice throughout the study period. At the prepubertal time point, the estrogen-exposed males exhibited higher bone mineral density, cortical volume, and cortical thickness compared with controls. However, by the time of peak bone mass acquisition, the early skeletal findings had reversed; estrogen-exposed mice had lower bone density with reduced cross-sectional area, cortical volume, and cortical thickness, resulting in cortical bones that were less resistant to fracture. NE exposure also resulted in reduced testicular volume and lower circulating IGF-I. Male mice exposed to estrogen on the first day of life experience age-dependent changes in skeletal development. Prepubertal animals experience greater endocortical bone acquisition as a result of estrogen exposure. However, by adulthood, continued developmental changes result in overall reduced skeletal integrity.

Peak bone mass, achieved during adolescence and early adulthood, is one of the most important determinants of bone strength and fracture risk later in life. There is strong evidence that peak bone mass is under considerable heritable control. However, the genetically determined trajectory of bone mass acquisition is remarkably sexually dimorphic; compared with women, men achieve a cortical bone structure with greater resistance to mechanical forces and thus experience fewer fragility fractures. These differences in bone development begin to emerge during puberty, with gonadal hormones and the GH-IGF-I axis implicated as primary factors in their generation (1–3). Interestingly, there is growing evidence that programming of these patterns, leading to the adult skeletal phenotype, takes place very early in development.

A transient activation of the hypothalamic-pituitary-gonadal axis occurs in the early postnatal period, resulting in a surge of reproductive hormones mimicking the levels observed in puberty (4). This minipuberty of infancy is critical for normal development of gonadal and reproductive function and differentiation of sexual behavior (5–7). Rodents demonstrate a similar postnatal surge of reproductive hormones and are a useful model to study the role of neonatal sex steroid exposure on developmental processes. Neonatal hormones exert their effects via direct action on target tissues but also act via effects on neural differentiation in sexually dimorphic neuroendocrine structures (8). This critical developmental window in the early postnatal period is characterized by heightened sensitivity to hormones (9, 10).

Bone is a target tissue for estrogens and androgens. Both estrogen receptors and androgen receptors are present in bone tissue, and sex steroids play important roles in both bone modeling and remodeling (11–13). Because males and females have similar levels of estrogen receptor and androgen receptor in bone tissue, there is ongoing interest in the role estrogen plays in the growth and maintenance of the male skeleton. To some extent, this interest emerged from concerns about the potential detrimental effects of environmental estrogens in the male. Studies have begun to investigate the effects of excess estrogen during the critical neonatal developmental window (14–16). A previous study demonstrated that exposing neonatal male mice to estrogenic compounds produced changes in somatic growth when evaluated at 20 weeks of age (17). However, mechanisms involved in the sex-specific effects of neonatal estrogen (NE) remain unclear. In addition, the effect of NE on skeletal parameters in prepubertal animals had not previously been investigated. The aim of the present study was to more fully characterize the impact of brief NE exposure on somatic growth, skeletal development, and the GH-IGF-I axis in laboratory mice.

Materials and Methods

Animals

All C57BL/6J mice used in these experiments were bred under identical conditions at the Portland Veterans Affairs Veterinary Medical Unit. On the day of birth, pups were injected sc with either 100-μg β-estradiol 3-benzoate (Sigma) dissolved in 0.01-mL sesame oil or 0.01-mL sesame oil alone. At the time of weaning, mice were group housed by sex (4 animals per cage) and maintained with ad libitum water and laboratory rodent chow (Purina Diet 5001; PMI Feeds, Inc) in a 12-hour light, 12-hour dark cycle (6 am to 6 pm). Monthly measurements included body weight (BW), body length (defined as the distance from the tip of the nose to the base of the tail), anogenital distance (defined as the distance between the posterior base of the penis and the anterior lip of the anus), and whole-body bone mineral density (BMD). All animal procedures were approved by the VA Institutional Animal Care and Use Committee and were performed in accordance with National Institute of Health guidelines for the care and use of animals in research.

Tissue collection

Male mice were killed at either 4 or 16 weeks of age by decapitation under isoflurane anesthesia. Female mice were killed at 16 weeks of age. Blood was collected for hormone analysis by cardiac puncture before decapitation. Serum was isolated and stored at −80°C until used for biochemical analyses. Pituitary glands were carefully dissected and stored at −80°C until tissue processing. Testes were carefully dissected and weighed. Both femora were harvested immediately from each mouse. Fifth lumbar (L5) vertebrae were harvested from 16-week animals. The right femur was measured from the greater trochanter to the external condyle using a digital caliper. After ex vivo dissection of soft tissue and fat, the excised bones were wrapped in sterile gauze soaked in PBS and stored at −20°C until subsequent analyses.

Body composition and BMD

Body composition (lean and fat mass), areal bone mineral density (aBMD) measurements (mg/cm²) of the whole body, femora, and L5 vertebrae were determined by dual energy x-ray absorptiometry using the PIXImus instrument (Lunar Corp). Densitometric analysis of the whole body (defined as the whole-body image minus the calvarium, mandible, and teeth) was performed monthly on animals under isoflurane anesthesia. Routine calibration was performed daily with a defined standard (phantom). Previous experience with this instrument indicates a precision error (expressed as the coefficient of variation) for bone mineral content of 0.99 ± 0.51% and for BMD of 1.71 ± 0.33%.

Femoral shaft cross-sectional geometry

In the initial study, cortical femoral shaft bone geometry was examined with a desktop x-ray microtomographic (μCT) scanner (SkyScan Model 1074). Images were analyzed with Optimas software (version 6.2; Media Cybernetics) as previously described (18).

Cortical and cancellous bone microarchitecture

Three-dimensional microarchitectural changes in cortical and cancellous bone from 4- and 16-week-old male mice were assessed using quantitative x-ray μCT. Femora and L5 vertebrae were scanned in 70% ethanol using a Scanco μCT40 scanner (Scanco Medical AG) at a voxel size of 12 × 12 × 12 μm (55-kV_p x-ray voltage, 145-μA intensity, and 200-ms integration time). Filtering parameters sigma and support were set to 0.8 and 1, respectively. Bone segmentation was conducted at a threshold of 237 for femora and 245 for lumbar vertebrae (scale 0–1000) determined empirically. Cortical bone was evaluated in the midshaft of the femur and cancellous bone was evaluated in the distal metaphysis and lumbar vertebrae. Cortical and cancellous bone compartments were defined using manual contouring. Twenty slices (240 μm) at the midshaft of the femoral cortical diaphysis were evaluated for total cross-sectional volume (cortical and marrow volume, μm³), marrow volume (μm³), and cortical thickness (μm). Forty-two slices (504 μm) of trabecular (cancellous) bone at the femoral metaphysis were evaluated for measurement of bone volume/tissue volume (BV/TV) (%), trabecular number (Tb.N) (mm^{− 1}), and trabecular thickness (Tb.Th) (μm). Analysis of lumbar vertebrae included the entire region of cancellous bone in the vertebral body between the cranial and caudal growth plates (175 ± 2 slices, 2100 ± 24 μm). Direct cancellous bone measurements included cancellous bone volume fraction (BV/TV, %), Tb.Th (μm), Tb.N (mm⁻¹), and trabecular spacing (μm).

Biomechanical strength

To assess femoral structural properties, the right femur was tested to failure in 3-point bending with a high-resolution materials test apparatus (Model 4442; Instron Corp). The loading fixture consisted of 2 fixed lower supports, placed at a span length of approximately 7 mm, and an upper loading point attached to a moving actuator. The upper loading point contacted the specimen at its midpoint, which was coincident with the center of the span. System software was used to displace the actuator at a strain rate of 0.5%/s until failure occurred. Load and displacement data were recorded and failure load, representing the energy absorbed before breaking, and stiffness, calculated from the linear portion of the load vs displacement curve, were determined using system software.

Measurement of biochemical parameters

Serum IGF-I was measured using a commercial mouse IGF-I ELISA kit (ScienCell Research Laboratories), which was reported to have no detectable cross-reactivity with IGF-II. All samples and standards were measured in duplicate. Serum estradiol was measured using a commercial mouse estradiol ELISA kit (catalog number 2046Z; Diagnostic Automation, Inc).

RNA preparation and quantitative PCR

Total RNA was extracted from pituitary glands using QIAGEN RNeasy kits (QIAGEN, Inc). DNA was removed from total RNA using ribonuclease-free deoxyribonuclease (QIAGEN). RT reactions were prepared using a TaqMan Reverse Transcription kit (Applied Biosystems, Inc). For each reaction, cDNA synthesis was prepared using 100-ng RNA in a reaction containing 1-μL 10× RT buffer, 2-μL 25mM MgCl₂, 1.6-μL 10mM deoxynucleotide triphosphates, 0.6-μL 50μM random hexamers, 0.4-μL ribonuclease inhibitor, and 0.6-μL MultiScribe RT, as much as suffices to 10 μL with nuclease-free water. RT reactions were performed on an Eppendorf Mastercycler (Eppendorf AG) programmed for 25°C for 10 minutes, 37°C for 1 hour, and 95°C for 5 minutes. Samples were diluted with 40-μL nuclease-free water and stored at 4°C until RT-PCR was performed.

Real-time PCR (RT-PCR) was performed on an ABI 7300 Real-Time PCR System using mouse-specific primer probe set obtained from Applied Biosystems (GH Mm00433590_g1). Each RT-PCR contained 50-μL TaqMan Universal PCR Master Mix, 0.5-μL primer probe, and 5-μL cDNA. Samples and endogenous controls (eukaryotic 18s rRNA) were run in duplicate to ensure repeatability. Auto comparative threshold values were calculated using 7300 relative quantity Study Software version 1.3 and verified. Gene expression is reported as fold change relative to the oil-treated control group using the 2^−ΔΔCt method. Statistical analyses were performed on the Δ cycle threshold values.

Statistical analysis

All data were analyzed using Prism5 software (GraphPad Software, Inc). Main effect differences between the 2 groups were assessed by unpaired two-tailed t tests. All data are expressed as mean ± SEM.

Results

Sex-specific effects of NE on somatic growth, attainment of peak BMD, and biomechanical strength

Transient NE exposure resulted in dramatic sex-specific effects on the acquisition of adult BW and bone development in C57BL/6J mice. At 16 weeks of age, BW, as well as whole-body, femoral, and vertebral aBMD of female mice were unaffected by NE exposure (Table 1). However, adult NE male mice exhibited reduced BW (−20%), whole-body (−7%), vertebral (−14%), and femoral (−15%) aBMD as compared with the control oil-injected male (CM) mice. Consistent with previous reports (16, 19), transient NE exposure negatively impacted femoral length in both sexes (−5% in females and −8% in males), but μCT evaluation of cross-sectional femoral shaft geometry demonstrated that NE exposure resulted in strikingly different effects on the structural configuration of adult cortical bone between male and female C57BL/6J mice. As compared with the CM group, adult male NE mice exhibited a 21% reduction in overall femoral cross-sectional area (FCSA) (P < .001), 7% reduction in cortical area (P < .001), and 31% reduction in marrow area (P < .001). Although NE exposure resulted in 10% greater cortical thickness, the overall reduction in femoral cortical diameter and moment of inertia observed in the NE group predicted the femoral shafts of the NE group to be less resistant to bending forces. Ultimate force, representing the amount of energy absorbed by the bone before breaking, was indeed 17% lower in the male NE group (P < 0 0.01) compared with controls. Bone stiffness, which represents the elastic rigidity of the bone and takes into account the amount of mineral present in the bone tissue, was 11% lower in NE males compared with controls (P < .001) (Table 1). In contrast, overall cross-sectional area, moment of inertia, and biomechanical measures of femoral bone strength (failure load and stiffness) of adult female mice were unaffected by transient neonatal exposure to estrogen. However, the female NE mice did exhibit a 7% increase in femoral cortical area (P < .001) and a 14% reduction in marrow area (P < .001) and a 15% increase in cortical thickness.

Table 1.

Somatic and Skeletal Measures of Adult C57BL/6J Mice Exposed to Estrogen on the First Day of Life

	Female C57BL/6J Mice			Male C57BL/6J Mice
	Control (n = 13)	Estradiol (n = 14)	P Value	Control (n = 12)	Estradiol (n = 14)	P Value
Whole body
Weight, g	22.65 ± 0.4	24.9 ± 1.4	n.s.	28.3 ± 0.3	22.6 ± 0.7	<.0001
aBMD, mg/cm2	52.0 ± 0.2	50.8 ± 0.7	n.s.	52.1 ± 0.4	48.2 ± 0.7	<.001
L5 vertebrae
aBMD, mg/cm2	43.0 ± 0.5	41.0 ± 1.5	n.s.	42.5 ± 0.7	36.4 ± 0.6	<.0001
Whole femur
Length, cm	15.6 ± 0.1	14.7 ± 0.2	<.001	15.7 ± 0.1	14.5 ± 0.1	<.0001
aBMD, mg/cm2	53.9 ± 0.5	52.6 ± 1.2	n.s.	57.2 ± 1.5	48.6 ± 0.6	<.0001
Femoral shaft geometry
Cross-sectional area, mm2	1.623 ± 0.021	1.535 ± 0.058	n.s.	1.870 ± 0.047	1.474 ± 0.031	<.0001
Marrow area, mm2	0.933 ± 0.019	0.798 ± 0.055	<.05	1.111 ± 0.029	0.768 ± 0.033	<.0001
Cortical area, mm2	0.691 ± 0.005	0.737 ± 0.007	<.0001	0.759 ± 0.022	0.706 ± 0.009	<.05
Cortical thickness, μm	166 ± 1	191 ± 5	<.001	167 ± 3	185 ± 4	<.01
Moment of inertia, mm4	0.107 ± 0.005	0.103 ± 0.004	n.s.
Femoral shaft biomechanics
Ultimate failure load, N	17.1 ± 0.4	17.0 ± 0.6	n.s.	19.3 ± 1.1	16.0 ± 0.5	<.01
Stiffness, N/mm	118.4 ± 4.2	119.8 ± 3.2	n.s.	117.1 ± 3.9	104.8 ± 2.6	<.05

n.s., nonsignificant by t test.

Thus, transient estrogen exposure of C57BL/6J mice during early development resulted in lasting, sex-specific effects. In broad terms, male adult mice experienced an attenuation of somatic growth and compromised bone mass and strength as a consequence of the NE exposure, whereas the attainment of adult BW, aBMD, and bone strength in female mice was, for the most part, unaffected by NE exposure.

Longitudinal effects of NE on somatic growth and attainment of peak BMD and biomechanical strength in male C57BL/6J mice

To better understand the consequences of NE exposure in male C57BL/6J mice, we performed a second experiment examining BW, body composition, bone density, and 3-dimensional cortical and cancellous bone architecture at 4 weeks of age (an age before the influence of pubertal levels of sex steroid hormones on growth and skeletal development in mice) and again at 16 weeks of age (when the acquisition of adult bone mass is complete).

By 4 weeks of age, the NE-exposed males had already deviated from CMs in BW (22% reduction; P < .0001), whole-body length (12% reduction; P < .001), and percent fat mass (12% increase; P < .0001) (Table 2 and Figure 1). Following the same somatic trend as that of whole-body length, femoral length of the prepubertal NE males was reduced by 8% compared with CM (P < .05). However, despite the attenuation of somatic growth, whole-body BMD measures were similar between the 2 groups of mice, and femoral aBMD was 32% greater in the prepubertal NE males compared with their CM counterparts (P < .001) (Table 2).

Table 2.

Somatic and Skeletal Measures of Male Mice

	Prepubertal 4-week-old mice			Adult 16-week-old mice
	Control (n = 10)	Estradiol (n = 7)	P value	Control (n = 17)	Estradiol (n = 11)	P value
Whole body
Length, cm	6.9 ± 0.13	6.1 ± 0.13	<.001	8.9 ± 0	8.4 ± 0.1	<.0001
Weight, g	13.5 ± 0.4	10.4 ± 0.3	<.0001	27.3 ± 0.3	22.6 ± 0.4	<.0001
% Body fat	10.4 ± 0.2	11.6 ± 0.3	<.001	13.0 ± 0.3	13.7 ± 0.6	n.s.
aBMD, mg/cm2	29.9 ± 0.5	31.0 ± 0.6	n.s.	51.6 ± 0.3	47.8 ± 0.5	<.0001
Whole femur
Length, cm	10.7 ± 0.3	9.9 ± 0.2	<.05	15.9 ± 0	14.9 ± 0.1	<.0001
aBMD, mg/cm2	27.7 ± 0.9	36.5 ± 2.3	<.001	57.7 ± 0.7	49.0 ± 0.5	<.0001
Femoral metaphysis
BV/TV, %	18.0 ± 2.9	57.8 ± 6.4	<.0001	23.2 ± 3.4	5.2 ± 1.5	<.0001
Tb.N, 1/mm	6.6 ± 0.7	11.8 ± 0.6	<.001	6.1 ± 0.2	4.1 ± 0.2	<.0001
Tb.Th, μm	42.1 ± 1.8	72.3 ± 7.0	<.01	53.4 ± 7.0	33.1 ± 7.2	<.001
Femoral diaphysis
Cross-sectional volume, mm3	0.316 ± 0.012	0.240 ± 0.005	<.0001	0.460 ± 0.014	0.355 ± 0.018	<.001
Marrow volume, mm3	0.218 ± 0.01	0.099 ± 0.022	<.001	0.198 ± 0.019	0.122 ± 0.037	<.001
Cortical volume, mm3	0.098 ± 0.003	0.141 ± 0.018	<.05	0.240 ± 0.014	0.208 ± 0.013	n.s.
Cortical thickness, μm	111 ± 3	214 ± 36	<.05	216 ± 6	236 ± 13	n.s.
Ultimate failure load, N	n.d.	n.d.		20.9 ± 0.6	17.6 ± 0.8	<.01
Stiffness, N/mm	n.d.	n.d.		113.5 ± 2.1	102.2 ± 2	<.001
IGF-I	1016 ± 234	272 ± 77	<.01	1498 ± 102	1272 ± 141	n.s.

n.d., not done; n.s., nonsignificant by t test.

Figure 1. Somatic growth and body composition of male mice exposed to NE vs controls.

A, Whole-body weight in CMs (□) and NE-exposed (estradiol) males (■) at 4, 6, 8, 10, 12, and 16 weeks. Also shown, control (oil) females (▴) at 16 weeks. B, Whole-body length, defined as the distance from the tip of the nose to the base of the tail, in CMs (□) and NE-exposed (estradiol) males (■) at 4, 6, 8, 10, 12, and 16 weeks. Also shown, CFs (▴) at 16 weeks. C, Whole-body size of CMs (□) and NE-exposed (estradiol) males (■) at 16 weeks of age. D, Total-body lean mass of CMs and NE-exposed (estradiol) males at 4 and 16 weeks of age (*, P < .0001).

To more fully assess these early changes in skeletal development, femoral microarchitecture was determined with quantitative μCT. Analysis of cortical bone at the femoral mid-diaphysis demonstrated a 24% reduction in femoral cross-sectional volume (P < .001) and 55% reduction in marrow volume (P < .001) in prepubertal NE mice compared with CM mice. However, cortical volume was 44% greater (P < .05) and cortical thickness was 93% greater (P < .01) in prepubertal NE mice compared with their CM counterparts (Table 2 and Figure 2). The trabecular bone compartment is more metabolically active and less influenced by mechanical forces than cortical bone. However, similar findings were observed in the trabecular bone compartment (femoral metaphysis). Despite the substantial difference in body mass at this age, bone volume fraction of the prepubertal NE mice was increased more than 3-fold above that of the controls (P < .01). Both Tb.N and Tb.Th were also higher in the NE group compared with CM (77% and 71%, respectively; P < .01) (Table 2 and Figure 2). Thus, NE exposure resulted in increased bone mass in both the cortical and trabecular compartments of 4-week-old male mice.

Figure 2.

Changes in femoral and vertebral microarchitecture assessed by quantitative x-ray μCT in CM mice and NE-exposed (estradiol) male mice at 4 and 16 weeks.

In order to determine whether any of the observed effects on measured parameters could be the result of residual estradiol lingering from the single exposure weeks earlier, serum estradiol was measured in a separate cohort of 3-week-old male mice injected on the first day of life in an identical manner with either 100-μg β-estradiol 3-benzoate dissolved in sesame oil or sesame oil alone. The assay revealed no measureable estradiol in either group of mice (data not shown).

As predicted by our initial studies, the inhibitory effect of transient NE exposure on somatic growth persisted throughout the 16-week experiment (Table 2 and Figure 1). By adulthood (16 wk of age), the NE group was 22% lower in weight (P < .001) and 12% shorter in body length (P < .001) in comparison with the CM (Table 2). As depicted in Figure 1, the BW and length of the adult NE males more closely approximated that of adult control females (CFs) than that of CMs. Notwithstanding their smaller size, body composition analysis demonstrated adult NE males exhibited a higher whole-body percent fat mass (Table 2). Coupled with their overall reduced body mass, NE exposure thus resulted in adult male mice with 20% lower lean mass than their controls (P < .001) (Figure 1D).

In contrast to that observed in the prepubertal animals, adult NE male mice again exhibited reduced whole-body and femoral aBMD values as compared with the CM mice (Table 2). Femoral midshaft cortical bone volume and overall cross-sectional volume increased between 4 and 16 weeks of age in both groups of mice. However, cortical thickness only significantly increased in the CM mice. As compared with the CM group, adult NE mice exhibited a 23% reduction in femoral cross-sectional volume (P < .001) and 39% reduction in marrow volume (P < .001). Biomechanical measures of whole bone strength were again reduced in the NE mice (16% reduction in failure load and 10% reduction in stiffness) compared with the CM mice.

Consistent with the cortical bone findings, adult NE male mice also accumulated significantly less trabecular bone compared with the CM mice as evidenced by a 77% reduction in bone volume fraction at the femoral metaphysis (P < .001) and 31% reduction in the lumbar vertebra (P < .001). The NE group demonstrated a 32% decrease in Tb.N (P < .001) and 38% decrease in Tb.Th (P < .001) at the femoral metaphysis and a 10% reduction in Tb.N (P < .01) and 29% reduction in Tb.Th (P < .001) in the lumbar vertebra (Table 2 and Figure 2). Thus, by the time the animals reached skeletal maturity, NE exposure had produced a phenotype consisting of less cortical and trabecular bone, leaving bones that were smaller and more prone to fracture compared with controls.

Alterations in testicular development

Testicular weight was measured to determine whether NE exposure had an effect on testicular development. The testicular weight, corrected for BW of 4-week-old NE mice, was 82% (P < .001) less than that of that of their CM counterparts (Figure 3). At 16 weeks, there was a less prominent but persistent negative impact on testicular weight (corrected for BW) than the NE group (NE, 6.1 ± 0.3 mg/g BW vs CM, 6.8 ± 0.3 mg/g BW; P = .2). Despite the apparent inhibitory effect of transient NE exposure on gonadal size, anogenital distance did not differ between the 22 groups of male mice at either age (4 wk NE, 0.39 ± 0.02 cm vs CM, 0.35 ± 0.01 cm; 16 wk NE, 0.62 ± 0.03 cm vs CM, 0.61 ± 0.02 cm).

Figure 3. Testicular growth.

A, Testicular weight of CM mice and NE-exposed (estradiol) male mice at 4 weeks of age (*, P < .001) and at 16 weeks of age (†, P < .05). B, Appearance of testes removed from CM mice and NE-exposed (estradiol) male mice at 4 weeks of age.

Changes in the GH-IGF-I axis

To examine the impact of NE on the central GH axis, we employed real-time PCR to assess steady-state pituitary GH mRNA levels. In the prepubertal animals, steady-state GH mRNA levels were 41% lower in the NE males relative to CM (P < .15), but this difference diminished by adulthood (Figure 4A). We observed the same trend in circulating serum IGF-I levels (Figure 4B). Prepubertal NE males exhibited 73% lower circulating IGF-I compared with the CM group (P < .01), but by 16 weeks of age, no difference in serum IGF-I was discernible between the 2 groups.

Figure 4. Changes in GH and IGF-I axis.

A, GH mRNA relative expression measured by RT-PCR from CM mice and NE-exposed (estradiol) male mice at 4 and 16 weeks of age. No statistical difference was detected in GH mRNA steady-state levels at either time point. B, Serum IGF-I measured in CM mice and NE-exposed (estradiol) male mice at 4 weeks of age (*, P < .01) and at 16 weeks of age (p = nonsignificant by t test).

Discussion

As evidenced by our experimental findings, the neonatal hormone milieu plays a critical role in determining life-long sex-specific patterns in murine growth and skeletal development. Male adult mice experienced an attenuation of somatic growth and compromised bone mass and strength as a consequence of the NE exposure, whereas female mice attained expected BW, areal bone density, and bone strength. Although overall FCSA was unchanged, the increased femoral cortical area and reduced marrow area suggest an effect of NE exposure to augment endosteal bone growth in the NE female mice. In contrast to that observed in females, adult NE male mice weighed less than their CM counterparts with reduced aBMD and femoral bone strength values. Similar to the females, transient NE exposure resulted in reduced femoral bone length. However, these femora also exhibited reduced overall FCSA (due to reductions in both cortical and marrow areas), suggesting some limitation of periosteal bone growth in addition to the putative positive effects on endosteal bone growth seen in the female mice.

Our findings in adult C57BL/6J mice are in line with previous studies investigating the effects of neonatal exposure to diethylstilbesterol (DES) (a synthetic nonsteroidal estrogen) on skeletal development. Using outbred CD-1 mice, Migliaccio et al (16) found that exogenous DES exposure, during gestation as well as administered in the first 5 days of life, resulted in shorter femur length and higher bone mass in female mice. Kaludjerovic and Ward (17) reported that male CD-1 mice exposed to DES on each of the first 5 days of life exhibited lower BW, diminished cortical bone area and thickness, lower BMD, and decreased bone stiffness and strength compared with control males at 4 months of age. However, the current study is the first to investigate prepubertal effects of transient NE exposure on developmental changes in skeletal phenotype.

Our findings demonstrate that the changes in somatic growth and bone development resulting from a single NE exposure begin to emerge well before mice reach skeletal maturity. At 4 weeks of age, the NE males had increased bone mass in both the cortical and cancellous compartments. The structural changes that occurred by 4 weeks of age, however, were not sustained throughout the remainder of growth. By the time the mice reached skeletal maturity, NE males exhibited bone density and size and a reduced ability to sustain a bending force before fracture compared with the CM mice. The NE males demonstrated feminized growth patterns and adopted body composition parameters in adulthood similar to CFs, with higher percent fat and lower BW compared with control males (Table 2).

BW and mechanical loading have an important role on bone growth and development. Larger body mass, and thus higher mechanical strains, stimulate bone formation (20). This concept may have had a role in the changes seen in our postpubertal NE males. NE exposure resulted in slower weight gain throughout the study interval. Upon reaching skeletal maturity, the lower body mass of NE males may have contributed to lower BMD, reduced periosteal expansion and cortical area, and decreased bone strength. However, this speculation does not hold true in the prepubertal cohort. At 4 weeks of age, despite significantly lower BW, NE males had higher cortical and cancellous bone mass. The structural arrangement in the cortical compartment suggests that increased endocortical bone apposition in the femur occurred in the NE males during the first 4 weeks of life, without the concurrent periosteal expansion or length extension observed in the 4-week controls.

In humans and rodents, sexual dimorphism in growth patterns and skeletal development appear during puberty. Postpubertal differences between males and females are primarily established by sex-specific levels of gonadal hormones and GH/IGF-I secretion patterns (3, 21–24). Studies in animal models and humans have demonstrated that males, under the influence of testosterone and a male-specific GH secretion pattern during puberty, undergo greater periosteal bone expansion than females, resulting in larger bones with greater resistance to bending (3, 11, 20, 25). Neonatal gonadectomy of male mice has been shown to result in slower weight gain throughout growth and decreased bone formation compared with intact males in adulthood, despite pubertal testosterone replacement (14, 15, 21, 26). In addition, the inhibition of bone growth is more pronounced when gonadectomy is performed in the neonatal period compared with gonadectomy during puberty (22–24). Androgen-deficient hypogonadal (hpg) adult mice, which do not experience the normal postnatal testosterone surge, demonstrate shortened adult bone length and disordered bone turnover (27). Many of these changes persist even with peripubertal exogenous testosterone administration, indicating that the perinatal male-specific hormone environment is necessary to achieve normal adult male bone size and biomechanics. In contrast, estrogen exerts little influence on periosteal bone apposition but stimulates endocortical bone formation by lowering the modeling threshold on endosteal bone surfaces (28). Free estradiol levels were inversely related to endosteal circumference of the radius in young healthy men (29), and estrogen treatment of an adult with complete androgen insensitivity syndrome resulted in endosteal contraction, with increasing cortical thickness (+3%), cortical area (+5%), and unchanged periosteal circumference (30). Although puberty is the time during which these patterns emerge, there is growing evidence that the neonatal period is a critical time point for programming of the usual sex-specific patterns of growth and skeletal development. Our study demonstrates that disruption of the expected male neonatal hormone milieu with brief estrogen exposure results in significant age-dependent changes in skeletal phenotype.

Disruption of the neonatal hormone environment is also known to affect reproduction and fertility. In male rodents, exposure to excess estrogen during this critical neonatal period has detrimental effects on testicular development and spermatogenesis (7, 31–33). This results in a reduction in testicular weight, which corresponds with lower testosterone levels in adulthood as well as Sertoli cell dysfunction. Our study produced similar findings, with NE males demonstrating lower testicular weights than control males at both 4 and 16 weeks with a presumed reduction in circulating serum testosterone. The feminized growth pattern, body composition, and skeletal findings in the adult group suggest that diminished pubertal testosterone levels are implicated in these observed changes. However, the skeletal changes observed in the prepubertal cohort likely occurred independently from changes in circulating testosterone, because testosterone is normally suppressed in the first 4 weeks of life in mice. We hypothesize that the early exposure to estrogen was directly responsible for the endocortical bone expansion present by 4 weeks of age. Transgenic male mice overexpressing the human aromatase gene have unaffected periosteal circumference but increased cortical thickness as a result of reduced endosteal circumference (34). Furthermore, male senescence-accelerated mouse prone 6 mice exhibit marrow expansion due to impaired endocortical bone formation at 4 weeks of age that can be reversed with estradiol (35). Because the estrogen exposure in the present experiment was transient in nature, we believe the direct effects of estrogen on endocortical bone formation at 4 weeks of age eventually diminished to be replaced by the detrimental effects of NE exposure on the testosterone status of these growing male mice resulting in a hypogonadal skeletal phenotype at 16 weeks of age.

Sex-specific GH and IGF-I production can be altered by manipulation of the neonatal sex steroid hormone environment. Rodents demonstrate sexual dimorphism in GH secretory patterns at the time of puberty, with males demonstrating a low-frequency, high-amplitude secretory pattern producing low or undetectable GH levels between pulses. Females, on the other hand, demonstrate high-frequency, low-amplitude GH pulses with higher baseline GH levels (2, 36). These differences in GH secretion influence the skeletal dimorphism observed at the time of puberty, with the male GH secretory pattern having a more effective role in stimulating growth. This sexual differentiation of the GH secretory pattern is largely regulated by differences in pubertal hormones but is also dependent on the neonatal sex steroid environment (23, 24, 37, 38). There are sex-specific differences in hypothalamic GHRH and somatostatin neurons that can be altered by manipulating the neonatal sex steroid environment (2, 38–41). In this article, we investigated whether the feminized growth pattern and skeletal phenotype in the estrogen-exposed males occurred via alterations in the expected male-specific GH secretion. We observed markedly reduced circulating serum IGF-I and a trend towards lower pituitary GH mRNA at 4 weeks in the NE males compared with controls, but neither of these differences persisted at 16 weeks. Further studies are needed to elucidate whether NE exposure causing changes in somatostatin or GHRH secretory dynamics may be part of the explanation of the changes that we observed in the GH-IGF-I axis and growth and skeletal phenotype.

There are limitations to the present study. The dose of estradiol used here produced changes in reproductive tissues in mice when given in the neonatal period (Figure 3) and was higher than what would be expected as exposure to environmental estrogens. Further studies will be needed to determine whether there is a dose-response element of estrogen exposure on skeletal development and whether the observed changes are dependent on the timing and duration of this exposure. We presume that skeletal changes were, in part, related to reduced circulating serum testosterone, given the dramatic changes that we observed in testicular volume. However, we did not measure serum testosterone directly. Measuring serum testosterone and sex hormone binding globulin at 4 and 16 weeks by mass spectrometry would help delineate the role of circulating serum testosterone on the skeletal changes observed. Several studies have demonstrated that exposing male rodents to exogenous estrogen in the neonatal period results in a significant decrease in adult testicular weight, germ cell volume per testis, and serum testosterone levels (7, 42). We suspect that our findings of a reduction in testicular weight in the estrogen-treated group are in line with these previous studies and would correlate with a concurrent reduction in serum testosterone. Future studies employing targeted transgenic mouse models can determine whether the phenotypic changes occurring after early estrogen exposure are due to a direct effect on bone or via centrally mediated mechanisms.

We have demonstrated that brief transient exposure of male mice to estrogen in the neonatal period reverses and feminizes the expected trajectory of somatic growth, skeletal morphology, and bone biomechanics. The perinatal and neonatal period is a time of developmental plasticity, during which exogenous factors can significantly alter programming of growth and hormonal patterns. Exogenous estrogen, given to male mice in the neonatal period, appears to alter developmental programming, resulting in early stimulatory effects on endocortical bone formation. However, by the time the mice reach skeletal maturity detrimental changes in skeletal integrity, with smaller bones more prone to fracture, are the end result of the early estrogen exposure. These findings suggest that a pathophysiologic contributor to male osteoporosis risk could, in part, lie in aberrant developmental programming of hormonal secretion patterns during the neonatal and perinatal periods.