Several studies demonstrate that androgens are important in males. Orchidectomy decreases longitudinal growth and radial cortical growth in the long bones of rodents (Turner, R. T et al (1990) J. Orthop Res. 8,612-617; Turner, R. T et al (1989) J Bone Miner Res. 4,557-563; Sandstedt, J et al (1994) Endocrinology 135,2574-2580; Ornoy, A. et al (1994) Bone Miner 24,43-58). Furthermore, androgen treatment stimulates growth in orchidectomized growing rats and mice (Turner R. T. et al (1990) supra; Ornoy, A. et al (1994) supra; Jansson, J. O. et al (1985) Endocrinology 117,1881-1889) as well as in growing boys (Richman, R. A. & Kirschm, L. R. (1988) N Engl. J. Med. 319,1563-1567). These effects may either be direct via the stimulation of androgen receptors or indirect via aromatization of androgens into estrogen and thereafter stimulation of estrogen receptors.

Recently it was demonstrated by Vanderschueren et al that the conversion of androgens into estrogen is required for normal body growth in male rats, indicating that indirect effects of androgens, mediated by estrogen, are important (Vanderschueren, D et al (1997) Endocrinology 138,2301-2307).

In addition to the growth-related effects of gonadal deficiency, orchidectomy also decreases bone mass in adult rodents (Turner R. T et al (1989) supra; Vanderschueren, D et al (1997) supra; Koh, E T et al (1996) Magnes Res. 9,13-21). This effect is at least partly dependent on the androgen receptor as treatment with non-aromatizable androgens restores bone mass (Vanderschueren D et al (1992) supra; Wakley, G. K. et al) 1991) J Bone Miner Res. 6,325-330). On the other hand several clinical studies have demonstrated a strong relationship between serum estrogen levels and BMD in males (Slemenda, C. W. et al (1997) J Clin Invest 100,1755-1759; Gillberg, P et al (1999) Calcif Tissue Int. 64, 209-213; Ongphiphadhanakul, B. et al (1995) Clin. Endocrinol (Oxf) 43,727-733; Ongphiphadhanakul, B. et al (1998) Clin. Endocrinol (Oxf) 49,803-809). Furthermore, aromatase deficiency in humans (Morishima, A et al (1995) J. Clin. Endocrinol. Metab. 80, 3689-3698) as well as aromatase inhibition in rats (Vanderschueren, D. et al (1997) supra; Vanderschueren, D. et al (1996) Calcif Tissue Int. 59,179-183) is associated with osteopenia, suggesting that androgens may also regulate adult bone metabolism, either directly by stimulation of androgen receptors, or indirectly via aromatization and subsequent stimulation of estrogen receptors.

The cloning of the novel estrogen receptor, ER (3, suggested that there may exist alternative mechanisms of action for estrogen (Kuiper, G. G., et al (1996) Proc. Natl. Acad Sci. USA 93,5925-5930). We and others have demonstrated that ERP is expressed in growth plate chondrocytes and osteoblasts, indicating a possible role for ERD in the regulation of longitudinal bone growth and/or adult bone metabolism (Onoe, Y., et al (1997) Endocrinology 138,4509-4512, Arts, J., Kuiper, G. G., et al (1997) Endocrinology 138, 5067-5070, Vidal, O., et al (1999) JBone Miner Res In press, Nilsson, L. O., et al (1999) J Clin Endocrinol Metab 84,370-373; Windahl own unpublished results). However, the physiological role of ERP in the regulation of growth and bone metabolism is still unknown. In humans, evidence for the importance of ERa for mediating effects of estrogen in the skeleton comes from a case report describing a young male with estrogen resistance due to a mutation in the human ERa gene (Smith, E. P. et al (1995) N. Engl. J.

Med. 331,1056-1061). This male was reported to suffer from osteoporosis at the age of 28.

Mice lacking a functional ERa gene, ERa Knockout mice (ERKO), have been generated (Couse, J. F. et al (1995) Mol. Endocrinol. 9,1441-1454) and more recently ERP Knockout mice (BERKO) have also been described (Krege, J. H. et al (1998) Proc. Natl.

Acad. Sci USA 95,15677-15682). At present the skeletal phenotype of male ERKO mice is unclear (Kimbro, K. et al (1996) J. Bone Miner Res. 11, S 125; Schmidt, A. et al (1999) J Bone Miner Res. 14, S456; Ederveen. A. et al (1999) J Bone Miner Res. 14, S170).

Furthermore, we recently demonstrated that male BERKO mice do not exhibit osteopenia (Windahl, S. H. et al (1999) J Clin Invest. 104,895-901). This has raised the question concerning the relative importance of estrogen receptor subtypes in the skeleton of male mice. In order to investigate the estrogen receptor specificity in the regulation of growth and adult bone metabolism in male mice, we have generated Double-ER-Knockout mice (DERKO). In the present study we have compared the skeletal phenotypes of male WT, ERKO, BERKO and DERKO mice.

We have recently generated mice devoid of functional ERP protein and reported that ERP is essential for normal ovulation efficiency, but is not essential for female or male sexual development, fertility, or lactation (Krege, J. H., et al (1998) Proc Natl Acad Sci USA 95, 15677-15682).

The molecular mechanisms of action for ERa versus ERP have recently been investigated.

ERa and Ergs have almost identical DNA-binding domains and studies in vitro have demonstrated that the two receptors have similar affinities for estrogenic compounds (Kuiper, G. G. et al (1996) Proc Natl Acad Sci U S A 93,5925-5930, Kuiper, G. G., et al (1997) Endocrinology 138,863-870, Tremblay, G. B., et al (1997) Mol Endocrinol 11, 353-365). The amino-acid sequence of ERP differs from ERa in the N-and C-terminal trans-activating regions. Therefore the transcriptional activation mediated by ERP may be distinct from that of ERa (Paech, K., et al (1997) Science 277,1508-1510). Considering the great similarities in ligand-and DNA-binding specificity it has been speculated that a differential tissue distribution of estrogen receptors may be important for mediating tissue specific responses to estrogens (Kuiper, G. G., and Gustafsson, J. A. (1997) FEBS Lett 410, 87-90). Thus, the unique transactivating domains of the two receptor subtypes, in combination with differential tissue-distribution, or differential cell-type distribution within a tissue, could be important factors to determine the estrogen response in target tissues.

The hormone testosterone is required for the pubertal growth spurt and the acquisition of normal bone density in mammals. These effects of testosterone may be direct via stimulation of the androgen receptor, or indirect via aromatisation of testosterone and thereafter stimulation of estrogen receptors. In the present study, the inventors have looked at the role of estrogen receptor subtypes for pubertal growth and adult bone metabolism in male mammals, particularly male mice.

The effect of androgens on the male skeleton may either be direct via a stimulation of androgen receptors or indirect via aromatization of androgens into estrogen and thereafter stimulation of estrogen receptors. Possible direct effects of androgens are illustrated by skeletal abnormalities in androgen resistant humans and rodents (Bertelloni. S. et al (1998) Horm. Res. 50,309-314, Vanderschueren, D. et al (1993) J. Bone. Miner. Res. 8,801-809).

However, several studies have clearly demonstrated that the effect of androgens on the male skeleton, at least partly, is dependent on the conversion of androgens into estrogen. In the present study, we demonstrate that estrogen resistance in the male mouse, due to loss of all known estrogen receptors, results in decreased skeletal growth. ERKO and DERKO but not BERKO mice display similar growth phenotypes, demonstrating that ERa but not ERP is the estrogen receptor mediating the effects of estrogen on skeletal growth in the male mouse. The shortening of the long bones in ERKO and DERKO mice was associated with decreased growth plate width in the proximal tibia. Similar findings have also been reported in orchidectomized mice and rats (Turner, R. T. et al (1989) supra; Sandstedt, J. et al (1994) supra). Furthermore, Ornoy et al showed that orchidectomy in mice decreases growth plate area measured in the proximal tibia and that low-dose estrogen treatment increases the same parameter (Ornoy, A et al (1994) supra). These findings demonstrate that physiological levels of estrogen have a stimulatory effect on longitudinal growth in male rodents. Similarly, estrogens are required for the pubertal growth spurt in boys (MacGillivray, M. H. et al (1998) Horm. Res. 49 Suppl 1,2-8). Estrogen regulates final height in humans by a stimulatory effect on the pubertal growth spurt, followed by closure of the epiphyseal growth plates at the end of puberty. In humans with estrogen deficiency or estrogen resistance growth plate fusion never occurs. This results in continuous slow growth even after puberty (Morishima, A. et al (1995) J. Clin. Endocrinol. Metab. 80, 3689-3698, Smith, E. P et al (1994) N. Engl. J. Med 331,1056-1061). In rodents, on the other hand, growth plate closure does not occur. Therefore rodent species grow continuously throughout life. Thus, the decreased growth observed in ERKO and DERKO is caused by a lack of estrogen stimulated growth, whereas the tall stature in the previously described estrogen resistant adult male was caused by a lack of growth plate closure (Smith, E. P. et al (1994) supra). Therefore, the male ERKO mouse is not a good model for postpubertal growth in humans but it may be a model for skeletal growth during adolescence.

It is a well-established fact that orchidectomized rodents as well as hypogonadal humans develop osteopenia (Vanderschueren, D (1996) Horm. Res. 46,95-98; Seeman, E et al (1983) Am. J. Med. 75 977-983; Stanley, H. L. et al (1991) J Am. Geriatr. Soc. 39 766-771). Although androgen replacement restores bone mass in gonadectomized male rats (Wakley, G. K. et al (1991) supra) it has also been demonstrated that estrogen, at least partly, reverses bone loss caused by orchidectomy (Ornoy. A. et al (1994) supra; Vanderschueren, D. et al (1992) supra). The present study, with estrogen insensitivity due to inactivation of both ERa and ERD, supports the notion that estrogen exerts important effects on the male skeleton. The phenotype of the male DERKO mouse is similar to what has earlier been described for aromatase inhibited male rats (Vanderschueren, D. et al (1997) supra). Both male DERKO mice and aromatase inhibited male rats demonstrate decreased femoral BMC, areal BMD, as well as decreased cortical dimensions and moment of inertia, without any effect on cortical volumetric BMD or cortical thickness. These two studies demonstrate that the conversion of androgens into estrogen is important in male rodents and that skeletal maturation in such species is estrogen dependent. In the present study, BMC measured by DXA and adjusted for body weight, was significantly reduced in ERKO and DERKO males, demonstrating that the effects on BMC were specific and not only reflected a general growth inhibition. This finding clearly demonstrates that ERa alone, and not ER, is the mediator of estrogenic effects in the skeleton of mammals such as the male mouse. Interestingly, a slight decrease in the relative weights of the heart and lung was also seen in ERKO but not in BERKO mice, indicating that ERa may exert specific effects in these two organs as well. In contrast, the relative weights of liver, kidney and brain were unchanged in ERKO, BERKO and DERKO males.

ERKO and DERKO mice demonstrated a decreased diaphyseal cross sectional area and periosteal circumference of femur, resulting in a pronounced decrease of the area moment of inertia. When the quality of the bone is unchanged, the area moment of inertia is normally proportional to the mechanical strength of the bone determined by three-point-bending (Ferretti, J. L. et al (1996) Bone 18,97-102). The maximal load was decreased in male ERKO mice but it was not decreased more than suggested by the changes in area moment of inertia. Therefore, the amount of bone, but not the mechanical quality of the bone, was decreased in ERa inactivated male mice.

Aromatase inhibition of male rats resulted in a small decrease in trabecular BMD (Vanderschueren, D. et al (1997) supra). In the present study, neither the pQCT technique nor bone histomorphometry detected any significant changes in cancellous bone density in male ERKO, BERKO or DERKO mice. Thus, our experiments indicate that neither ERa nor EX (3 is essential for the maintenance of cancellous bone mass in the male mouse. This finding raises the question whether other estrogen receptor subtypes exist or whether other hormones may compensate for estrogen resistance in the skeleton of male DERKO mice.

Androgens prevent cancellous osteopenia in orchidectomized rats. Therefore androgens could compensate for loss of estrogen receptor activity in ERKO, BERKO and DERKO males. Interestingly, ERKO males have somewhat increased serum levels of testosterone (Eddy, E. M et al (1996) Endocrinology. 137,4796-4805).

Bone loss following gonadal deficiency is normally associated with increased bone turnover. Surprisingly, osteocalcin, a marker for bone formation, was decreased in ERKO males. This finding and the pronounced cortical osteopenia seen in ERKO and DERKO males led us to seek other explanations to the skeletal phenotype in these mice. Over-all size and cortical radial growth are parameters that are highly sensitive to changes in the GH/IGF-I axis (Andreassen, T. T. (1995) J. Bone. Miner Res. 10 1057-1067, Ohlsson, C. et al (1998) Endocr. Rev. 19 55-79; Rosen, H. N. et al (1995) J. Bone. Miner. Res. 10, 1352-1358). Because these parameters were altered in ERKO and DERKO males, serum IGF-I was measured to investigate if the GH/IGF-I axis was affected in ERKO and DERKO males. Serum IGF-I levels were decreased in ERa inactivated mice. We also found a strong correlation between serum IGF-I levels and affected skeletal parameters in the ERKO and DERKO mice, including length, BMC of femur, periosteal circumference and maximal load in the femur diaphysis. These findings do not prove, but indicate, that changes in the GH/IGF-I axis could partly explain the skeletal phenotype seen in male ERKO and DERKO mice. GH and IGF-I are known to increase serum osteocalcin (Ohlsson. C. et al (1998) Endocr. Rev. 19,55-79). Therefore, the decreased serum osteocalcin levels in male ERKO mice may be caused by reduced serum IGF-I levels. This is also supported by the finding that aromatase inhibited male rats have decreased serum IGF-I levels and reduced levels of serum osteocalcin (Vanderschueren, D. et al (1997) supra). An effect of estrogen on the GH/IGF-I axis in males is also supported by several clinical as well as experimental studies. Circulating GH and IGF-I concentrations increase during normal male puberty (Miller, J. D. et al (1982) J. Clin. Endocrinol Metab. 55, 989-994; Mauras, N. et al (1987) J. Clin Endocrinol Metab 64 596-601; Martha Jr. P M. et al (1989) J. Clin. Endocrinol. Metab 69,563-570; Weissberger. A. J. et al (1989) Horm Res. 32,148-150). These changes appear to be secondary to the pubertal rise in testosterone concentrations since they are also observed in prepubertal and hypogonadal boys undergoing induction of puberty with exogenous testosterone (Miller, J. D. et al (1982) supra; Link, K. et al (1986) J. Clin. Endocrinol Metab. 62,159-164). The mechanism whereby testosterone interacts with the somatotropic axis may either be direct, mediated by androgen receptors, or indirect through the action of estrogen on estrogen receptors. The possibility that estrogen mediates the effects of testosterone on the somatotropic axis has been suggested in a previous study showing a significant correlation between circulating levels of estrogen, but not testosterone, and GH secretion in men (Ho. K. Y. et al (1987) J.

Clin. Endocrinol Metab. 64,51-58). Furthermore, it has also been demonstrated that testosterone plays an important role in the modulation of the somatotropic axis in adulthood and this effect is, at least partly, dependent on the conversion of testosterone to estrogen (Weissberger, A. J. et al (1993) J. Clin. Endocrinol. Metab 76,1407-1412).

The effects of androgens in the skeleton of the male mouse are summarised in Fig. 8.

Others have presented studies indicating that androgens, directly via interaction with the androgen receptor, exert effects on the male skeleton. In the present study we have confirmed that part of the effect of androgens is dependent on aromatization. Furthermore, the present study clearly demonstrates that ERa, but not ER (3, mediates the effect of estrogen on the skeleton in the male mouse. In conclusion, we have generated DERKO mice, which are fully viable despite the fact that they are devoid of all known estrogen receptors. Male ERKO and DERKO mice have decreased body weight, reduced longitudinal bone growth and a pronounced cortical osteopenia. Our findings demonstrate that ERa but not ERD mediates the effect of estrogen in the male skeleton. We propose that some of the skeletal effects seen in ERa inactivated male mice may be due to an inhibition of the GH/IGF-I axis.

According to one aspect of the invention, there is provided a method of treating growth disorders in a mammal, the method comprising treating the mammal with an ERa-specific agonist.

According to another aspect of the invention, there is provided a method of treating growth disorders in a mammal, the method comprising treating the mammal with an ERa-specific antagonist.

The ERa ligand/modulator used in the method of the invention may be a SERM (Selective Estrogen Receptor Modulator) i. e a compound having a tissue-selective mixed agonist/antagonist activity. SERMs include tamoxifen, raloxifene, drolixifene and tamoxifen methiodide.

The mammal may be male or female and is preferably pre-pubesent.

The ERa agonist or antagonist used in the method may have a binding affinity of less than lOnM for ERa.. Preferably, the ERa agonist or antagonist has a binding affinity of 0.0001 to 10 nM for ERa.

According to another aspect of the invention, there is provided the use of an ERa selective agonist in the preparation of a medicament for the treatment of a growth disorder.

According to another aspect of the invention, there is provided the use of an ERa selective antagonist in the preparation of a medicament for the treatment of a growth disorder.

In such uses the ERa antagonist may have a binding affinity of ERa of less than 10 nM, preferably 0.0001 to 10 nM.

According to another aspect of the invention, there is provided a pharmaceutical composition suitable for treating or preventing growth disorders in a mammal, the composition comprising an ERa antagonist or agonist. Preferably, the ERa agonist or antagonist has a binding affinity for ERa of less than 10 nM, most preferably 0.0001 to 10 nM.

Pharmaceutical compositions of this invention comprise any of the compounds of the present invention, and pharmaceutically acceptable salts thereof, with any pharmaceutically acceptable carrier, adjuvant or vehicle. Pharmaceutically acceptable carriers, adjuvants and vehicles that may be used in the pharmaceutical compositions of this invention include, but are not limited to, ion exchangers, alumina, aluminum stearate, lecithin, serum proteins, such as human serum albumin, buffer substances such as phosphates, glycine, sorbic acid, potassium sorbate, partial glyceride mixtures of saturated vegetable fatty acids, water, salts or electrolytes, such as protamine sulphate, disodium hydrogen phosphate, potassium hydrogen phosphate, sodium chloride, zinc salts, colloidal silica, magnesium trisilicate, polyvinyl pyrrolidone, cellulose-based substances, polyethylene glycol, sodium carboxymethylcellulose, polyacrylates, waxes, polyethylene-polyoxypropylene-block polymers, polyethylene glycol and wool fat.

The pharmaceutical compositions of this invention may be administered orally, parenterally, by inhalation spray, topically, rectally, nasally, buccally, vaginally or via an implanted reservoir. The pharmaceutical compositions of this invention may contain any conventional non-toxic pharmaceutically-acceptable carriers, adjuvants or vehicles. The term parenteral as used herein includes subcutaneous, intracutaneous, intravenous, intramuscular, intra-articular, intrasynovial, intrastemal, intrasternal, intrathecal, intralesional and intracranial injection or infusion techniques.

The pharmaceutical compositions may be in the form of a sterile injectable preparation, for example, as a sterile injectable aqueous or oleaginous suspension. This suspension may be formulated according to techniques known in the art using suitable dispersing or wetting agents (such as, for example, Tween 80) and suspending agents. The sterile injectable preparation may also be a sterile injectable solution or suspension in a non-toxic parenterally-acceptable diluent or solvent, for example, as a solution in 1,3-butanediol.

Among the acceptable vehicles and solvents that may be employed are mannitol, water, Ringer's solution and isotonic sodium chloride solution. In addition, sterile, fixed oils are conventionally employed as a solvent or suspending medium. For this purpose, any bland fixed oil may be employed including synthetic mono-or diglycerides. Fatty acids, such as oleic acid and its glyceride derivatives are useful in the preparation of injectables, as are natural pharmaceutically-acceptable oils, such as olive oil or caster oil, especially in their polyoxyethylated versions. These oil solutions or suspensions may also contain a long-chain alcohol diluent or dispersant such as Ph. Helv or a similar alcohol.

The pharmaceutical compositions of this invention may be orally administered in any orally acceptable dosage form including, but not limited to, capsules, tablets, and aqueous suspensions and solutions. In the case of tablets for oral use, carriers which are commonly used include lactose and corn starch. Lubricating agents, such as magnesium stearate, are also typically added. For oral administration in a capsule form, useful diluents include lactose and dried corn starch. When aqueous suspensions are administered orally, the active ingredient is combined with emulsifying and suspending agents. If desired, certain sweetening and/or flavoring and/or colouring agents may be added.

As the skilled artisan will appreciate, lower or higher doses than those recited above may be required. Specific dosage and treatment regiments for any particular patient will depend upon a variety of factors, including the activity of the specific compound employed, the age, body weight, general health status, sex, diet, time of administration, rate of excretion, drug combination, the severity and course of the disease, the patient's disposition to the disease and the judgment of the treating physician.

According to another aspect of the invention, there is provided a method of selecting compounds for the regulation of body growth in mammals, the method comprising selecting a compound on the basis of its ability to antagonise agonist-dependent ERa activity.

According to a further aspect of the invention, there is provided a method of selecting compounds for the use in the treatment of growth disorders, the method comprising testing the compound in a mammal which is wholly or partially ERa deficient or in cells derived from such an animal.

According to the invention, there is also provided a method of treating a bone mineral density disorder in a mammal, the method comprising treating the mammal with an ERa -specifc agonist. Alternatively, the invention provides a method of treating a bone mineral density disorder in a mammal, the method comprising treating the mammal with an ERa -specific antagonist. The ERa-specific ligand/modulator may be a SERM, the mammal may be a male or female and may be pre-pubesent. The ERa agonist or antagonist may have a binding affinity of less than 10 nM, preferably 0.0001 to 10 nM or ERa.

The invention also provides the use of an ERa selective agonist in the preparation of a medicament for the treatment of a bone mineral density disorder.

Alternatively the invention provides the use of an ERa selective antagonist in the preparation of a medicament for the treatment of a bone mineral density disorder.

The invention also provides a pharmaceutical composition suitable for treating or preventing bone mineral density disorders in a mammal, the composition comprising an ERa antagonist or agonist.

The invention also provides a method of selecting compounds for the regulation of bone mineral density in mammals, the method comprising selecting a compound on the basis of its ability to antagonise agonist-dependent ERa activity. In particular, compounds are selected for the regulation of adult bone mineral density disorders.

Other aspects of the invention are apparent from the claims.

Methods in accordance with the invention will now be described, by way of example only, with reference to the accompanying drawings, Figures 1 to 8 in which: Fig 1 shows the results of experiments on weight gain in male mice; Fig, 2 illustrates the body weight in wild type (WT), ERKO, BERKP and DERKO mice at different ages.

Fig. 3 illustrates the results of experiments of Organ weights/Body weight expressed as % of wild type mice (WT) at 4 months of age in wild type (WT), ERKO, BERKO and DERKO (n=6 for WT, n=9 for ERKO, n=5 for BERKO and n=5 for DERKO).

Fig. 4 illustrates the results of experiments of Length of femur (A) and crown rump (B) and width of the proximal tibial growth plate (C) in wild type (WT), ERKO, BERKO and DERKO mice (n=6 for WT, n=9 for ERKO, n=6 for BERKO and n=5 for DERKO).

Fig. 5 shows the results of tests on bone mineral density in rats.

Fig 6 illustrates the results of experiments of DXA measurements of bone parameters in wild type (WT), ERKO, BERKO and DERKO mice (n=6 for WT, n=9 for ERKO, n=6 for BERKO and n=5 for DERKO); Fig. 7 are representative DXA scans (A) and mid-diaphyseal pQCT scans (B) of femora in adult wild type (WT), ERKO, BERKO and DERKO mice. (bottom 4A: High = high bone mineral density and Low = low bone mineral density); Fig. 8 illustrates the effects of androgens in the male mouse skeleton. AR = androgen receptor, ERa = estrogen receptor-a, ERß = estrogen receptor-p.

1. Generation of Knockout Mice Male double heterozygous (ERa+'~ ß +'~) mice were mated with female double heterozygous (ERa+'-3+'-) mice resulting WT, ERKO, BERKO and DERKO offspring. All mice were of mixed C57BL/65/129 backgrounds.

The animals were maintained under standardised environmental conditions, with free access to food and water. Genotyping of tail DNA was performed at 3 weeks of age. The ERa-gene was analysed with the following primer pairs: Primers AACTCGCCGGCTGCCACTTACCAT and CATCAGCGGGCTAGGCGACACG for the WT gene, correspond to flanking regions in the targeted exon no. 2. They produce a fragment of approximately 320 bp. Primers TGTGGCCGGCTGGGTGTG and GGCGCTGGGCTCGTTCTC for the KO gene, correspond to part of the NEO-cassette and the flanking exon no. 2. They produce a 700 bp fragment. Genotyping of the ERßgene has been previously described (30).

2. Body Growth Male wild type (WT) as well as estrogen receptor (BERKO) mice demonstrated a pubertal growth spurt as measured with body weight gain/day (Fig. 1) In contrast, no pubertal growth spurt was seen in estrogen receptor a-/-mice (ERKO) or in mice devoid of both estrogen receptors (DERKO).

In the results shown in Fig. 2, values are given as means. The bodyweights in WT and ERKO mice at different ages were first analysed by a two-way analysis of variance followed by Student Newman Keuls multiple range test. P values versus WT mice are indicated. Body weight was unchanged in ERKO, BERKO and DERKO at the prepubertal stage when compared to WT littermates (Fig 2, day 17, one-way ANOVA). Late pubertal and adult weight was decreased in ERKO and DERKO but not in BERKO when compared to WT mice (Fig 2, day 46-81, two-way ANOVA).

Growth of the appendicular-as well as the axial-skeleton was followed using repeated X-ray measurements. In the results shown in Fig. 3, values are given as means SEM.

Data at different ages were first analysed by a two-way analysis of vairance (A and B) or by a one-way analysis of variance (C) followed by Student Newman Keuls multiple range test.

P values versus WT mice are indicated in A and B. **p<0.05 versus WT (C). The length of the femur was chosen as a measure of appendicular growth whereas crown-rump (CR) length was used as a measure of axial growth. The length of the femur was unchanged at the prepubertal stage (Fig 3A, day 31, one-way ANOVA). Thereafter ERKO and DERKO demonstrated a gradual decrease in growth rate, resulting in a decreased femoral length at the adult stage (ERKO-5.7%, DERKO-4.4% versus WT, Fig 3A, SA). The decreased growth of the long bones in ERKO and DERKO was associated with a decreased growth plate width measured in the proximal tibia (Fig 3C). The CR length was also decreased in ERKO and DERKO compared with WT (Fig 3B).

3. Dual X-Ray Absorptiometry (DXA) Areal Bone mineral density (Areal BMD; BMC/cm2) and bone mineral content (BMC) were measured with the Norland pDEXA Sabre (Fort Atkinson, WI) and the Sabre Research software (3.6) as previously described (30).

In vivo measurements of animals were performed in order to determine total body, spine, femur and cranium BMC (medium resolution scan with line spacing set at 0.05 cm). Three mice were analysed at a time. A mouse, which was sacrificed at the beginning of the experiment, was included in all the scans as an internal standard in order to avoid inter-scan variations.

Ex vivo measurements of the left femur and tibiae were performed on excised bones placed on a 1 cm thick plexiglass table. All bones compared were measured in the same scan (high-resolution scan with line spacing set at 0.01 cm).

4. Peripheral Quantitative Computerized Tomography (pQCT) Computerized tomography was performed with the Stratec pQCT XCT Research M (Norland, software version 5.4B) operating at a resolution of 70 llm as previously described (30).

Mid-diaphyseal pQCTscans of femora and tibiae were performed to determine the cortical volumetric bone mineral density (volumetric BMD), cortical cross sectional area, periosteal and endosteal circumference and the cross sectional moment of inertia. The mid-diaphyseal region of femora and tibiae in mice contains only cortical bone.

Metaphyseal pQCT scans of left femora and tibiae were performed to measure trabecular volumetric BMD. The scan was positioned in the metaphysis at a distance from the distal growth plate corresponding to 4 % of the total length of the femur (an area containing cortical as well as trabecular bone). The trabecular bone region was defined by setting an inner threshold to 45% of the total area. The inter-assay coefficients of variation (CV) for the pQCT measurements were less than 2%.

The DXA technique gives the areal BMD whereas the pQCT gives the true volumetric BMD. Therefore a factor regulating the outer dimensions of a bone, will affect the areal BMD (DXA) but not the volumetric BMD (pQCT).

5. Histological examination and Histomorphometry Growth plate measurements. Right and left tibiae were fixed in 4% formaldehyde, embedded in paraffin and sectioned at a thickness of 4 m. The width of the growth plate was measured, after staining with Alcian blue/Van Gieson, using an image-processing system (Easy Image, Bergströms Instrument, Stockholm, Sweden) coupled to a microscope. The average of 20 growth plate measurements (2 sections, 10 measurements/section) was calculated for each tibia.

Bone Histomorphometry : The areas of trabecular bone within a reference area of the proximal tibia were measured in sections stained with Hematoxylin/Eosin. Measurements were performed on printed copies by point counting using a square lattice (1 and 2 cm).

Three fields of vision on three sections from each animal were used for the analysis. Data is presented as the ratio of trabecular bone volume (BV) to total volume (TV).

6. Radioimmunoassay Serum IGF-1 levels were measured by double antibody IGF binding protein-blocked radio immunoassay according to Blum and Breier (31).

7. Statistical Procedure Dynamic measurements were first analysed by a two-way analysis of variance (ANOVA) followed by Student Newman Keuls multiple range test. Static measurements (at the time of sacrifice) were first analysed by one-way ANOVA followed by Student Newman Keuls multiple range test.

8. Bone mineral status as determined by DXA BMC (g) and areal BMD (mg/mm2) were measured with DXA. In the results depicted in Fig. 5, the BMC (A) and BMC/Body weight (B) of the whole skeleton (total), femur, spine and cranium were measured using DXA technique as described in Methods. Values are given as means SEM. Data at different ages were first analysed by a two-way analysis of variance followed by Student Newman Keuls multiple range test. P values versus WT mice are indicated. BERKO demonstrated unchanged BMC and areal BMD (Fig 5A, table 1).

Furthermore, in ERKO and DERKO no effect was seen on BMC and areal BMD at the prepubertal stage (day 31, one-way ANOVA). However, later on ERKO and DERKO presented a marked decrease in total body BMC (Fig 5A). In addition regional measurements of BMC in the femur and spine also showed a significant decrease (day 118; total body: ERKO-21%, DERKO-22%; femur: ERKO-23%, DERKO-20%; spine: ERKO-23%, DERKO-19%, versus WT; Fig 5A, 6A). In the results in Fig. 6A, High = high bone mineral density and Low = low bone mineral density. Only a small effect was seen in the cranium (ERKO-7% versus WT, Fig 5A). Total body areal BMD was slightly decreased in ERKO at the adult stage. Both ERKO and DERKO displayed a decreased adult areal BMD in the femur (Table 1).

Table 1. Areal BMD as Measured using DXA WT ERKO BERKO DERKO (n=6) (n=9) (n=6) (n=5) Total Body BMD (mg/cm') Day 31 48. 70.8 50.60.5 Day 65 58.60.2 Day 118 65.80.8 2-way ANOVA P<0.05 NS NS Femur BMD (mg/cm') Day 31 35.51.2 Day 65 58.21.3 55.01.7 51.02.6 Day 118 60.82.1 2-way ANOVA P<0. 01 NS P<0.05 Spine BMD (mg/cm') Day 31 37.20.7 Day 65 51. 50.8 49.51.9 Day 118 60.83.0 2-way ANOVA NS NS NS Values are given as means SEM Data at different ages were first analysed by a two-way analysis of variance followed by Student Newman Keuls multiple range test. P values versus WT mice are indicated. NS = non significant.

To determine if the decrease in BMC in ERKO and DERKO males was greater than that associated with retarded body growth, BMC/body weight was calculated for the whole skeleton and for individual bones. Interestingly, in adult mice total body BMC/body weight was decreased in ERKO (-18%) and DERKO (-22%) when compared to WT. This was also the case for femur (ERKO-20; DERKO-19%) and spine (ERKO-21%; DERKO -18%; Fig SB).

9. Cancellous [what does this term mean?] bone density The pQCT technique was used to measure trabecular volumetric BMD in the metaphysis of the distal femur and in the proximal tibia. Results are shown in Table 2 Table 2. Trabecular volumetric BMD and Cortical Bone Parameters of Femur as Measured using pQCT WT ERKO BERKO DERKO (n=6) (n=9) (n=6) (n=5) Trabecular density (mg/mm') 0.2850.019 Corticaldensity (mg/mm3) 1.1840.011 Cortical area (mm2) 0.910.02** 1.010.04 0.920.04* Cortical bone mineral content (mg/mm) 1.090.05* Cortical periosteal circumference (mm) 5.260.12* Cortical endosteal circumference (mm) Values are given as means SEM. Data were first analysed by a one-way analysis variance followed by followed by Student Newman Keuls multiple range test. * p< 0.05. ** p< 0.01 versus WT.

In addition, histomorphometry was performed in the metaphysis of the proximal tibia, where trabecular bone volume/total volume (BV/TV) was measured. Neither the pQCT technique (Table 2, and data not shown) nor bone histomorphometry (BV/TV: WT ERKO BERKO DERKO one-way ANOVA) detected any significant changes in cancellous bone density.

10 Cortical bone parameters Cortical bone parameters were studied in detail in mid-diaphyseal pQCT scans of femora and tibiae (Table 2, Fig 6B and data not shown). The cortical BMC in the mid-diaphyseal section of femur was decreased in ERKO (-14%) and DERKO (-14%) compared with WT and this decrease was mainly due to a decreased cross-sectional bone area whereas cortical volumetric density was unchanged (Table 2). The decrease in cross sectional area in ERKO and DERKO was associated with decreased periosteal and endosteal circumference (Fig 6B and Table 2).

11 Mechanical testing of the femur diaphysis The size and position of the cortical cross-sectional bone area in ERKO and DERKO resulted in a pronounced decrease of cortical cross-sectional moment of inertia (ERKO -29%, DERKO-24% versus WT, Table 3).

Table 3. Mechanical testing of Femur Diaphysis WT ERKO BERKO DERKO (n=6) (n=9) (n=5) (n=5) AreaMomentofInertia (mm4) 02** Maximal Load (N) 24.31.6 Elastic Modulus (MPa) 3.50.1 Maximal Stress (GPa) 1328 1405 12210 1406 Values are given as means SEM. Data were first analysed by a one-way analysis variance followed by followed by Student Newman Keuls multiple range test. * p< 0.05. ** p< 0.01 versus WT.

Changes in area moment of inertia are often directly correlated to changes in the mechanical strength of the bone. Therefore, mechanical strength was tested by three-point-bending at the mid-diaphysel region of femur. ERKO demonstrated a significantly decreased maximal load whereas a tendency to decrease was seen in DERKO (ERKO-18%, DERKO-15%) compared with WT (Table 3). Other bone parameters, including maximal stress and elastic modulus, reflecting the quality of the bone, were not statistically changed (Table 3).

12. Biochemical Bone markers and IGF-I in serum Osteocalcin, a marker of bone formation was measured in serum at 110 days of age.

Osteocalcin was decreased in ERKO (Osteocalcin-25%, Table 4) and a tendency to decrease was seen in DERKO (Osteocalcin-9%, Table 4).

Table 4. Biochemical Bone Markers and IGF-I in Serum WT ERKO BERKO DERKO (n=6) (n=9) (n=6) (n=5) Osteocalcin (ng/ml) 943 713* 936 8611 IGF-I (ng/ml) 33736 2508* 31312 2646 Values are given as means SEM. Data were first analysed by a one-way analysis variance followed by followed by Student Newman Keuls multiple range test. * p< 0.05. ** p< 0.01 versus WT.

Overall size and cortical radial growth are parameters, which are highly sensitive to changes in the GH/IGF-I axis. Because these parameters were altered in ERKO and DERKO males serum IGF-I levels were measured to investigate if the GH/IGF-I axis was affected in the ERKO and DERKO. Serum IGF-I levels were decreased in ERKO (-26%) and there was a tendency to a decrease in DERKO (-22%, Table 4). Serum IGF-I levels were statistically correlated with length, BMC, BMC/weight, cortical cross sectional area, periosteal circumference, moment of inertia and ultimate load of femur (Table 5).

Table 5. Correlation with serum IGF-I r Femur Length 0.60** BMC 0.70*** BMC/weight 0.59** Trab vol BMD 0.22 Cortical cross sectional area 0.55** Endosteal circumference 0.47* Periosteal circumference 0.54** Moment of Inertia 0.55** Ultimate Load 0.50* Liver Weight 0.18 Kidney Weight 0.04 Heart Weight 0.39 Lung Weight 0.07 Brain Weight-0.17 <BR> <BR> <BR> <BR> <BR> <BR> Correlations were calculated using Pearsons correlation coefficient (r). * p<0.05, **<BR> <BR> <BR> <BR> <BR> <BR> p< 0.01,*** p<0. 001.

13. Organ weights The weights of several other organs were measured to see if the effect on the skeleton in ERKO and DERKO was tissue specific. To compare the relative growth of different organs the individual organ weights were divided with the total body weight. The weights of the liver, kidney, brain and testis were not significantly changed in any group. However, the weights of heart and lung were decreased in the ERKO compared with WT (heart-15%, lung-17%, Fig 7). In the results shown in Fig. 7, values are given as means SEM. Data were first analysed by a one-way analysis of variance followed by Student Newman Keuls multiple range test. * p<0.05 versus WT. These experiments demonstrate that ERa but not ER (3 is involved in the regulation of pubertal growth and adult bone mineral density in male mammals such as mice.