0013.7227/92/1312-0883$03.00/0 Endocrinology Copyright 0 1992 by The Endocrine
Vol. 131, No. 2 Printed in U.S.A.
Society
Mechanism of Action of Estrogen on Intramembranous Bone Formation: Regulation of Osteoblast Differentiation and Activity* RUSSELL T. TURNER, PEGGY BACKUP?, PAMELA GLENDA L. EVANS, AND THOMAS C. SPELSBERG
J. SHERMAN&
Departments of Orthopedic Surgery and Biochemistry and Molecular Medicine and Mayo Foundation, Rochester, Minnesota 55905
ESTHER
HILLS,
Biology, Mayo Graduate
School of
ABSTRACT Dynamic bone histomorphometry, 13H]thymidine radioautography, and Northern analysis for bone matrix proteins and insulin-like growth factor-I (IGF-I) were performed in calvariae of ovariectomized (OVX) and estrogen-treated OVX rats. Treatment of OVX rats with diethylstilbestrol (DES) for 2 weeks reduced the periosteal mineral apposition rate, osteoblast number, and osteoblast size in calvarial periosteum. DES treatment also reduced the number of preosteoblasts in the S phase of the cell cycle, suggesting that the decrease in osteoblast number was due in part to inhibition of proliferation of osteoprogenitor cells. One week after ovariectomy, there were small increases in mRNA levels for pre pro+2 (I) subunit of type I collagen (collagen), ostcocal-
tin, and osteonectin and a large increase in the mRNA level for IGF-I. DES treatment resulted in rapid decreases (3 h) in the mRNA levels for osteonectin, osteocalcin, and IGF-I. In contrast, mRNA levels for collagen were virtually unchanged after short term DES treatment. Uterus and liver served as positive and negative control tissues, respectively, for the effects of DES on IGF-I mRNA levels in OVX rats; mRNA levels were increased in uterus and decreased in liver after hormone treatment. We conclude from these studies that estrogen reduces periosteal bone formation by inhibiting both the differentiation and activity of osteoblasts. Furthermore, down-regulation of mRNA levels for IGF-I and bone matrix proteins precedes the changes in dynamic bone histomorphometry. (Endocrinology 131: 883-889, 1992)
G
creasedduring growth by intramembranous bone formation exclusively, whereas changesin bone volume in long bones are a summation of several processes,including longitudinal growth, endochondral ossification, modeling, and secondary intramembranous bone formation. Studies in long bones may, therefore, be complicated by estrogen having different actions on periosteal addition (5-8), endosteal and cancellous bone modeling (10, 1l), longitudinal growth (12), and endochondral ossification (12-15). Furthermore, in long bones there are possibleinteractions via locally produced signaling molecules (growth factors) among the various bone compartments, including periosteum, endosteum, metaphysis, growth plate, and epiphysis. Becausethe effects of sex steroids on bone cell metabolism in flat boneshave been largely unexplored, our present study is intended to provide initial characterization of the actions of estrogen on osteoblast differentiation, expression of mRNA for selected bone proteins, and bone matrix deposition in calvarial periosteum from growing female rats.
ONADAL hormones are implicated in the sexual dimorphism of the rat skeleton, including the flat bones of the skull (1, 2), pelvis (3), and tubular long bones (4-7). Ovariectomy (OVX) results in increased periosteal bone formation in tibiae of growing rats (5-8), whereas orchiectomy results in decreased periosteal bone formation (5, 6). Although gonadectomy reduces the sex differences in radial bone growth, these differences are reestablished by the administration of estrogens to ovariectomized (OVX) females (7) and androgens to orchidectomized males (6). In long bones, the inhibitory effects of estrogen on periosteal bone formation are accompanied by decreasesin mRNA levels for bone matrix proteins and alkaline phosphatase,but not glyceraldehyde-3-phosphate dehydrogenase, suggesting that the hormone regulates the expression of specific bone cell proteins (9). Estrogenic effects on long bone periosteum raise the question of whether the changesin gene expressionoccur in other skeletal sites. The potential effects of estrogen on flat bones are of interest because in these bones bone volume is in-
Materials
Received January 30, 1992. Address all correspondence and requests for reprints to: Russell T. Turner, 3-71 Medical Sciences Building, Mayo Clinic, Rochester, Minnesota 55905. *This work was supported by Grants AR-35651, AR-41418, and AG04875 from the NIH, the Mayo Foundation, and the National Osteoporosis Foundation. t Recipient of a National Osteoporosis Foundation Student Research Fellowship. $ Current address: Brown University Medical School, Providence, Rhode Island 02912. 5 Current address: University of Miami, Miami, Florida 33103.
and Methods
Animals Ten-week-old sham-operated and OVX female Sprague-Dawley rats were obtained from Harlan Sprague-Dawley (Indianapolis, IN). The rats were maintained in group cages (three or four rats per cage) and fed rat chow and water ad libitum.
Experiments Exp 1 characterized the histology periosteum-free calvariae, and excised
of intact calvarial periosteum, periosteum. A total of 10 OVX
883
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
ESTROGEN
884
ACTION
ON INTRAMEMBRANOUS
rats were studied. The intact calvariae were harvested from 1 group of 5 animals and fixed in 70% ethanol. The calvarial periosteum was harvested from rats in the second group as described below, and the excised periosteum and periosteum-free calvariae were fixed in 70% ethanol. Exp 2 was a histological study to determine the effects of 2 weeks of diethylstilbestrol (DES) treatment on periosteal bone formation in calvariae. A total of 24 OVX rats were divided into 2 equal groups. The OVX control group consisted of OVX rats implanted SC on the back of the neck with drug-free pellets; the estrogen-treated group consisted of OVX rats implanted with pellets containing DES. The pellets were rate (Innovative designed to deliver DES (2.5 mg/3 weeks) at a uniform Research, Toledo, OH). Tetracycline (tetracycline hydrochloride, Sigma Chemical Co., St. Louis, MO) was administered (20 mg/kg, ip) to each of the rats 7 days after implantation of the DES or drug-free pellets and again 13 days after the implantation. Tetracycline becomes deposited at the mineralizing front of newly deposited bone matrix and is used as a marker to determine the extent of appositional bone growth during a defined interval. All of the rats were killed with CO* gas 24 h after the second tetracycline labeling. After death, the calvariae were excised and fixed in 70% ethanol before processing for measurement of dynamic bone histomorphometry. The uteri were trimmed, excised, and weighed after gently expelling fluid from their lumen. Exp 3 was a cell kinetics study to determine the effects of estrogen on the number of osteoprogenitor cells in the S phase of the cell cycle. A total of 10 OVX rats were divided into 2 groups of 5 animals each. The OVX control group was implanted with drug-free pellets; the estrogen-treated group consisted of OVX rats implanted with DES. Six days after implantation, all of the rats were given an ip injection of [methyl-3H]thymidine (Amersham, Arlington Heights, IL; 45 Ci/mmol) at 1 pCi/g BW. Eighteen hours later, the rats were killed, and the intact calvariae were harvested and processed, as described below. Exp 4 characterized the effects of OVX and short term treatment with DES on mRNA levels for bone matrix proteins and insulin-like growth factor-I (IGF-I) in periosteum from calvariae. Five groups of 10 rats were studied. Two of the groups consisted of sham-operated animals, whereas the rats in the remaining 3 groups were OVX. One week after surgery, 1 group of OVX rats was given a single ip injection of DES (5 pg delivered in 0.1 ml 0.05% ethanol) and killed 3 h later. The remaining groups of sham-operated and OVX animals received carrier only. The calvarial periosteum was harvested after death, as described below, and the total cellular RNA was extracted. Exp 5 and 6 were carried out to verify the changes in mRNA levels in calvarial periosteum after short term DES treatment. Twenty OVX rats were studied in each of these experiments. One week after surgery, DES was administered to half of the rats, and the animals were killed about 3 h later. The calvarial periosteum, liver, and uteri were rapidly excised and frozen on dry ice before extraction of total cellular RNA.
Isolation
of total cellular RNA from calvarial periosteum
The skin on the head was moistened with water, shaved, and removed to expose the cranium. The superficial delicate fascia was incised on both sides of the midsagittal and coronal sutures, over the frontal and parietal bones, and parallel to the insertion of the temporalis muscularis. The periosteum, excluding the sutures, was then gently lifted off the bone surface using a dental probe and either fixed in 70% ethanol for histology or frozen on a block of dry ice before isolation of total cellular RNA. The frozen periosteum was pooled, homogenized, and extracted in guanidine hydrochloride (Gdn-HCl) (16), and the total cellular RNA was recovered (1.5-2 pg/calvaria) after CsCl gradient centrifugation using a SW50.1 rotor (Beckman Instruments, Berkley, CA) at 36,000 rpm for 16 h at 22 C (17). The frozen liver and uteri were homogenized and treated thereafter in the same manner as calvariae.
Electrophoresis The total cellular RNA from calvarial periosteum was resuspended water and separated electrophoretically on a 1% (wt/wt) agarose Each lane contained 15 pg total cellular RNA, which was denatured incubation for 1 h at 52 C in a solution of 1 M glyoxal dimethylsulfoxide
in gel. by
Endo. 1992 Voll31 *No 2
BONE
50% (vol/vol) in 10 mM sodium phosphate (pH 7.0). Electrophoresis was performed in 10 mM sodium phosphate (pH 7.0) at 60-70 V for about 4 h (17).
Northern
blot analyses
The RNA was transferred overnight from the gel to Magna 66 nylon membranes (MSI, Fisher Scientific, Pittsburg, PA) by capillary action in 20 x SSC (3 M NaCl and 0.3 M trisodium citrate), as described previously for the transfer of DNA (18). The filters were baked for 2 h at 80 C and prehybridized overnight. The 32P-labeled cDNA probes were prepared by random sequence hexanucleotide primer extension using Multiprime DNA labeling systems from Amersham or equivalent reagents. [32P]Deoxy-CTP (SA, -3000 Ci/ml; New England Nuclear Research Products, Boston, MA) was used to radiolabel the cDNA. Specific activities of about lo9 cpm/ pg DNA were achieved. The following templates were used for probe production: 1) rat osteocalcin (BGP), a gift from Dr. S. Rossi-Langer, Genetics Institute (Cambridge, MA); this full-length probe is pR 22-11, which is an EcoRI insert into pSI’65 (19); 2) human collagen, a gift from Drs. H. Hivanlem and G. Tromp, Jefferson Medical College (Philadelphia, PA); this is PHUCI, which is a full-length cDNA clone for the prepro-cu2 (I) chain of human type 1 precollagen inserted into the EcoRI site (20); 3) mouse 185 ribosomal RNA (21); this is an 1150-basepair (bp) cDNA insert into the BamHI site; 4) human osteonectin (ON), a gift from Dr. G. Long, University of Vermont (Burlington, VT) (22); this is pHVON-9-2 plasmid DNA containing a 546.bp human ON cDNA insert; and 5) human IGFI, a gift from Dr. G. Bell, Howard Hughes Medical Institute Research Laboratories, University of Chicago (Chicago, IL); this is plasmid phigf 1 containing an insert of about 660 bp, which extends from the second nucleotide of the codon for amino acid -15 of the single peptide of the IGF-I precursor to the poly(A) tract (24). The blots were hybridized overnight at 43 C with the 3ZP-labeled cDNA probes. After hybridization, the blots were washed under conditions of increasing stringency, up to 0.1 X SSC and 0.1% sodium dodecyl sulfate at 65 C, and then exposed to Kodak X-Omat AR-5 film (Eastman Kodak, Rochester, NY).
Densitometry
of the autoradiographs
The Northern blots were quantitated by densitometric autoradiographs, using a Shimadzu flying spot scanner (Shimadzu Scientific Instruments, Columbia, MD).
analysis of the model CS-9000
Histomorphometry After fixation for 2 days, the fixed tissues from Exp 1 and 2 were dehydrated in a series of increasing concentrations of ethanol, embedded without demineralization in a mixture of methylmethacrylate:2hydroxyethyl-methacrylate (12.5:1) to retain the tetracycline, and sectioned at an indicated thickness of 5 pm. The intact calvariae, periosteum-free calvariae, and excised periosteum from Exp 1 were stained either with toluidine blue or for alkaline phosphatase (25). The intact calvariae from Exp 2 were viewed unstained with a fluorescence microscope (Olympus BH-2, New Hyde Park, NY) to detect the tetracycline labeling. Histomorphometric procedures were carried out using a SMI-Microcamp semiautomatic image analysis system (Southern Micro Instruments, Inc., Atlanta, GA), consisting of a Compaq computer with Microcamp software interphased with a microscope and image analysis system. In this system, a high resolution color video camera mounted on an Olympus BH-2 microscope displays the image of the specimen on a color video monitor. The movement of a pen on a graphics tablet superimposes a tracing on the image of the specimen on the video screen. By this method, the region of interest is traced, and the line length and the area bounded by the tracing are calculated. The periosteal mineral apposition rate provides a quantitative assessment of bone matrix deposition and was determined as the mean of the distance between the two tetracycline labels measured every 50 pm (minimum of 50 measurements/section) along the periosteal perimeter divided by
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
ESTROGEN
ACTION
ON INTRAMEMBRANOUS
the labeling interval, which was 6 days. Osteoblast number and area were determined in toluidine-stained sections. pH]Thymidine
radioautography
Intact calvariae from Exp 3 were demineralized in 5% formic acid in 10% formalin for 4 days, dehydrated in a series of increasing concentrations of ethanol, embedded in glycolmethacrylate (Sorval, Norwalk, CT),
885
BONE
TABLE 1. The effects of DES on bone and uterine measurements in OVX rats Measurement Exp 1 Calvaria Mineral apposition rate tw/day) Uterus Wet ti tg) Exp 2 Calvaria Osteoblast no. (cells/ mm) Osteoblast area (lm*/ cell) Labeled preosteoblasts (cells/mm)
ovx
OVX + DES % Change
3.11 f 0.17 1.46 + 0.16”
-53
0.15 f 0.01 0.74 rfr 0.08”
+393
104 + 6
72 f 2’
-31
117 j, 10
55 zk 18*
-53
3.10 & 0.47 0.02 + 0.02”
-99
Values are the mean + SEM [n = lo-14 (exp 1) and 4-5 (Exp 2)]. ’ P < 0.001, by Student’s t test. *P < 0.01, by Student’s t test. and sectioned at an indicated thickness of 5 pm. The sections were attached to unsubbed acid-cleaned slides, dipped in the dark into melted (50 C) Ilford k50 nuclear tract emulsion, and exposed for 1 week. The sections were developed in Kodak D-19 and stained with toluidine blue before viewing with an Olympus BH-2 microscope to detect [3H]thymidine-labeled preosteoblasts. A cell with six or more grains per nucleus was considered labeled. Cells one or more cell layers removed from the osteoblast monolayer lining the periosteal perimeter and within the cambial layer were considered to be preosteoblasts, and the labeled cells were enumerated. We demonstrated in preliminary studies that with increasing time after the administration of radiolabeled thymidine, the number of labeled osteoblasts increased, and there was a corresponding decrease in the labeled putative preosteoblast population (data not shown). Approximately 10,000 cells were scanned per group, and the data were expressed as labeled preosteoblasts per mm periosteal perimiter.
Results
FIG. 1. The morphology of the calvarial periosteum. A, An intact calvaria stained with toluidine blue, showing the organization of the periosteum. Bone occupies the bottom portion of this panel. A monolayer of osteoblasts (large arrows) lines the mineralized surface of the bone. Less differentiated osteoprogenitor cells (small arrows) are found in this cellular layer, bounded by osteoblasts and the fibrous region. The outer fibrous region is composed chiefly of fibroblasts, but also includes fat cells, nerve cells, and epithelial cells. B, Isolated periosteum, showing the in situ localization of alkaline phosphatase activity (arrows) in the osteoblast cell layer. The tissue section was counterstained with methyl green thionine. The orientation is similar to that in A, with no bone present. C, Periosteum-free calvaria, showing the absence of alkaline phosphatase activity after complete removal of periosteal cells. The section was counterstained with methyl green thionine. The orientation is similar to that in A, with no cellular layer present. The horizontal bar is 25 pm in length.
The histology of an intact calvaria, a periosteum-free calvaria, and the harvested periosteum is shown in Fig. 1. The periosteum of an intact calvaria (Fig. 1A) is divided anatomically into two developmentally related, but spatially and metabolically distinct regions. The outer fibrous region is composed chiefly of fibroblasts, bundles of collagenous fibers, and fat cells. The inner cellular region contains cells of the osteoblast lineage, the preosteoblasts and osteoblasts. Osteoblasts are present in the innermost (or cambial) layer of the cellular portion, closely opposed to the bone surface, and are separated from the fibrous periosteum by one or more layers of preosteoblasts.Preosteoblastsdiffer morphologically and functionally from fibroblasts of the fibrous portion of the periosteum, in that the former contain more cytoplasm per cell and are committed in an osteogenicdirection. Osteoclasts are rarely found on the periosteal surface and when observed are associatedwith developing vascular spaces.The procedure for harvesting the periosteum for RNA analysis separated the cellular periosteum (Fig. 1B) from the remainder of the calvaria (Fig. 1C) at the osteoid seamand was effective in recovering the osteoblastslining the periosteal surface. Osteoblasts were identified in intact calvariae (Fig. 1A) and in isolated periosteum (Fig. lB), but not in periosteum-free calvaria (Fig. lC), by criteria which included
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
886
ESTROGEN
ACTION
ON INTRAMEMBRANOUS
BONE
Endo. 1992 Vo1131.No2
BGP123456 FIG. 3. Representative Northern blot of periosteal and liver cell total cellular RNA. Each lane represents RNA-pooled from 10 animals. The blot was successivelv nrobed with the 32P-labeled cDNAs for ON (ton band), 18s ribosomaiRNA (middle band), and BGP (bottom baidj. The bands were then identified based on previously reported sizes of the RNA transcripts for these genes, which hybridized with specific cDNA probes complementary for the mRNA species. The calvariae were from rats that-were intact (lane l), OVX (lanes 2 and 3), or OVX and treated with DES for 3 h before death (lane 4 and 5). The livers were from OVX rats (lane 6). Note that the expression of dN and BGP was below the detection limits for liver.
FIG. 2. [3H]Thymidine incorporation into nuclei of preosteoblasts. A, Radioautograph of intact calvarial periosteum (sectioned and stained with toluidine blue) of an OVX rat. Bone occupies the bottom portion of this panel. A preosteoblast in the S phase of the cell cycle is identified by grains over the nucleus (small arrow). The periosteal perimeter is lined bv a monolaver of osteoblasts (lar,e arrows). B, Radioautomaph of calvarial periosieum (sectioned and stained with toluidine blue)-of an OVX rat treated with DES for 6 days. Bone occupies the bottom portion of this panel. No labeled preosteoblasts are observed in this photograph, they were rarely observed in DES-treated rats. The periosteal perimeter is lined by smaller, less robust cells (arrows) than those in A. The horizontal bar is 25 pm in length.
size, location, intense basophilic staining with toluidine blue, a prominent Golgi apparatus, and high levels of alkaline phosphatase activity. The effects of 2 weeks of DES treatment on bone measurements and uterine wet weight in OVX rats are shown in Table 1. DES treatment resulted in a highly significant 53% decreasein the mineral apposition rate of calvarial periosteum between days 7 and 13 of treatment. In contrast, DES treatment of OVX rats resulted in a dramatic 393% increase in uterine weight. Seven days of treatment of OVX rats with DES resulted in decreasesin osteoblast number and size as well as a dramatic decreasein the number of [3H]thymidinelabeled preosteoblasts. Representative [3HJthymidine radioautographs of the calvarial periosteum are illustrated in Fig. 2. The periosteal perimeter of the calvariae of OVX rats
(Fig. 2A) was typically lined by plump cells (large umws) that expressedosteoblastcharacteristics, including basophilic staining, an asymmetrically located nucleus, and prominent Golgi apparatus; [3H]thymidine-labeled preosteoblasts(small arrow) were observed. In contrast, the periosteum of DEStreated rats (Fig. 2B) was typically lined by much smaller cells (arrows) that did not express features associatedwith active osteoblasts,and the putative preosteoblastswere rarely labeled with [3H]thymidine (Table 1). Single bands were detected by Northern analyses at the expected sizes for BGP, ON, and 18s ribosomal RNAs after hybridization with the radioactively labeled cDNA probes. The expected doublet was observed for type 1 collagen. As expected, four transcripts were identified for IGF-I. These were located at 7.5, 3.9, 1.8 and 0.7-1.2 kilobases(kb). The most prominant transcript expressedin liver, uteri, and calvariae was at 7.5 kb. Representative Northern blots are shown for BGP, ON, 18s ribosomal mRNA, and IGF-I mRNA (Figs. 3 and 4). Northern blots for collagen are not shown, but have been described previously (9). The effects of OVX on mRNA levels for selected bone proteins in calvarial periosteum are summarized in Fig. 5. OVX resulted in large increasesin mRNA levels for IGF-I and smaller increases in mRNA levels for ON, BGP, and collagen. The effects of short term estrogen treatment on mRNA levels for bone proteins are shown in Fig. 6. Administration of DES for about 3 h resulted in decreasesin mRNA levels for IGF-I, BGP, and ON and a negligible change in collagen.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
ESTROGEN
ACTION
ON INTRAMEMBRANOUS
BONE
887
12 BGP
~
/
1234567 FIG. 4. The Northern blot analyses shown in Fig. 3 were washed and rehybridized with the cDNA for IGF-I, as described in Fig. 3. Liver expressed higher mRNA levels for IGF-I then periosteum. The most prominent IGF-I transcript was located at 7.5 kb. Other transcripts were noted at 3.9, 1.8, and 0.7-1.2 kb.
Interestingly, after DES treatment, IGF-I mRNA levels were increasedin uterus and decreasedin liver (Fig. 7). Discussion Previous studieshave shown that OVX results in increases in periostealbone formation and apposition rates in the tibia1 diaphysis of growing rats (5-8). In contrast, OVX had little effect on the tetracycline-labeled perimeter (5-8, 12). These findings were originally interpreted to mean that ovarian hormone deficiency results in increased osteoblast activity, with little change in osteoblast number. However, it was assumedthat the tetracycline-labeled perimeter was directly proportional to osteoblastnumber, an assumptionwhich this study found to be incorrect for calvarial periosteum. Although 100% of the periosteal perimeter was labeled with tetracycline in calvariae from DES-treated rats, there was a 3 1% decreasein the number of osteoblastsper mm periosteal perimeter and a 53% decreasein osteoblast size. Thus, the estrogen-induced decreasein periosteal bone formation seen in this study was very likely due to a combination of reduced osteoblastnumber and reduced osteoblastactivity. Estrogen has similar effects on periosteum in long bones (5-9) and calvariae to inhibit bone formation. Studies of the tibia1 metaphysis have also shown that estrogen inhibits the formation of cancellous bone (10). In spite of similarities in
12 ON
FIG. 5. The effects of OVX on mRNA levels for IGF-I and bone matrix proteins. One week after OVX, rats were killed, and total cellular RNA was extracted, separated by size, and hybridized with [32P]cDNA probes, and the hybridization products were identified by radioautography. mRNA levels were determined after densitometry of the x-ray films and normalization to 18s ribosomal RNA to correct for differences in total RNA. The data for OVX rats are expressed as a percentage of the value in the ovary-intact control group. The results for two separate determinations (labeled 1 and 2) are shown. The dotted line represents the value for the intact control groups. COLL, Collagen.
the respective responses of cancellous and cortical bone compartments to estrogendeficiency, there are alsoprofound differences; there is a net increase in bone added to the periosteal surface of the tibia1 diaphysis after OVX, whereas there is a pronounced decreasein cancellousbone volume in the metaphysis (26, 27). BecauseOVX produces an increase in bone formation at both sites,the different results observed are probably due to the relative amount of bone resorption in the respective skeletal compartments. Osteoclastsare common in cancellous bone, and osteoclastnumber is increased after OVX (11, 26). In contrast, osteoclastsare uncommon to the periosteal surface of long bones and calvariae and are limited to the developing vascular spaces. Erosion of the periosteal surface does not occur in OVX rats, indicating that bone resorption is not increased by estrogen deficiency at this site. Prior studies in long bones were limited by the procedure used to isolate periosteal cells, which involved prolonged treatment of the boneswith collagenaseat 37 C. This method precluded investigation of rapid changes in gene regulation (9). Furthermore, we found that collagenase treatment of liver from OVX rats obliterates expressionof IGF-I mRNA in a tissue that expresseshigh levels of the message(unpublished data). The novel method used to separate periosteum from calvariae described in this manuscript now allows detection of mRNAs for rapidly regulated cell-signaling molecules, such as IGF-I, as well as for structural proteins. We show for the first time that OVX increases mRNA levels for bone matrix proteins and IGF-I in a skeletal tissue. The changes in mRNA levels for the bone matrix proteins
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
ESTROGEN
888
ACTION
ON INTRAMEMBRANOUS
BONE
Endo. Voll31.
1992 No 2
1.25
12
3
4
5
6
7
8
7. Representative Northern blot of liver and uterus total cellular RNA fromOVX rats. The blot was probed with the 32P-labeled cDNAs for IGF-I. as described in Fie. 3. Lanes l-4 were loaded with RNA from liver, whereas lanes 5-8 we& loaded with RNA from uterus. The oddnumbered lanes had RNA from solvent-treated rats, and the evennumbered lanes had RNA from DES-treated animals. Interestingly, DES treatment down-regulated IGF-I steady state mRNA levels in liver and up-regulated the message in uterus.
FIG.
O.o(
IGF-I
COLL
BGP
ON
6. The effects of DES treatment on mRNA levels for IGF-I and bone matrix proteins. One week after OVX, rats were given 5 pg DES or solvent. The DES-treated and control rats were killed and assayed as described in Fig. 5. The data are expressed as a percentage of the control value. The values are the mean ? SEM (n = 3). **, P < 0.01, by Student’s t test. COLL, Collagen. FIG.
are modest and consistent with the increase in bone formation determined by histomorphometry (5-7). We have previously shown an excellent correlation between indices of bone formation determined histologically and mRNA levels for collagen (28). The changes in mRNA levels after OVX are probably due to estrogen deficiency, rather than deficiency of other ovarian hormones, because DES treatment reversed the direction of the changes. The current studies support and extend earlier studies in long bones which showed that long term (l- to 2-week) treatment with estrogen reduces bone matrix synthesis and mRNA levels for bone matrix proteins and alkaline phosphatase (7, 9). The effects of estrogen on expression of IGF-I mRNA have not previously been investigated in periosteum; indeed, IGF-I has not previously been shown to be expressedin this skeletal tissue in viva Further studies will be necessary to identify the periosteal cell population(s) that expressesIGF-I mRNA and to verify that the changes in messagelevel reflect peptide synthesis. The present study demonstratesthat in calvariae, estrogen has rapid effects on mRNA levels for BGP, ON, and IGF-I. Interestingly, OVX and DES had minimal effects on mRNA levels for collagen. Although DES treatment ultimately reduced collagen mRNA levels in periosteum from tibiae and femora, these changes were not detected during the first week of estrogen treatment (9). The effect of estrogen on osteoblastsin other skeletal compartments is controversial.
The majority (11, 29), but not all (30), histomorphometric studies in humans and animals suggeststhat estrogen also inhibits the turnover of cancellous bone. The present study in calvariae strengthens the argument that estrogen affects osteoblasts similarly regardless of their location. The net effect of estrogen on bone volume would, however, depend upon the relative abundance of osteoblastsand osteoclasts. The dramatic decreasein labeled preosteoblastsin DEStreated OVX rats suggeststhat reduced proliferation of preosteoblasts with a consequent decreasein the production of osteoblasts was responsible for the reduction in osteoblast number in calvarial periosteum. The effects of estrogen on DNA synthesis may be direct or, alternatively, may be coupled to changes in bone cell activity, perhaps through decreased expression of paracrine signaling molecules. The results of this study are consistent with an indirect effect, in that IGF-I mRNA levels were found to be reduced in estrogen-treated calvariae. Estrogen has variable effects on cultured osteoblast-like cells and has been reported to increase, decrease, and not alter cell proliferation and mRNA and peptide levels for bone matrix proteins and IGF-I (31-33). IGF-I is believed to be an important mediator of estrogenic action on uterus (34), liver (35), and breast cancer cells (36) and may act asan autocrine/paracrine factor in skeletal tissue as well (37). Estrogen had opposite effects on IGF-I mRNA levels in uterus and liver, a finding that agreeswith previous studies (34, 35). The inhibitory effects of estrogen on intramembranous bone formation and proliferation of osteoprogenitor cells contrast markedly with the ability of the hormone to promote
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
ESTROGEN
ACTION
ON INTRAMEMBRANOUS
uterine growth. Interestingly, estrogen has been shown to up-regulate IGF-I mRNA levels in uterus (34), the opposite of its effects on calvarial periosteum and liver (35). Further studies will be necessary to determine the precise role of changes in expression of IGF-I in mediating the effects of estrogen on bone matrix synthesis and differentiation of osteoblasts. Acknowledgments The authors this manuscript.
thank
Ms. Carolyn
Blankenship
for typing
and proofing
References 1. Dahinten SL, Pucciarelli HM 1986 Variation in sexual dimorphism in the skulls of rats subjected to malnutrition, castration and treatment with gonadal hormones. Am J Physical Anthropol7:63-67 2. Riesenfeld A 1974 Endocrine and biochemical control of craniofacial growth: an experimental study. Hum Biol 46:531-572 3. Riesenfeld A 1972 Functional and hormonal control of pelvic morphology in the rat. Acta Anat 82:231-253 4. Hansson LI, Menander-Sellman K, Stenstrom A, Thorngren K-G 1972 Rate of normal longitudinal bone growth in the rat. Calcif Tissue Res 10:238-251 5. Turner RT, Hannon KS, Demers L, Buchanan J, Bell NH 1989 The differential effects of gonadal insufficiency on bone histomorphometry in male and female rats. J Bone Mineral Res 4:557-563 6. Turner RT, Wakley GK, Hannon KS 1990 Differential effects of androgens on cortical bone histomorphometry in gonadectomized male and female rats. J Orthop Res 8:612-617 RT, Vandersteenhoven JJ, Bell NH 1987 The effects of 7. Turner ovariectomy and 17 beta estradiol on cortical bone histomorphometry in growing rats. J Bone Mineral Res 2:115-122 8. Turner RT, Wakley GK, Hannon KS, Bell NH 1987 Tamoxifen prevents the altered bone turnover resulting from ovarian hormone deficiency. J Bone Mineral Res 2:449-456 9. Turner RT, Colvard DS, Spelsberg TC 1990 Estrogen inhibition of periosteal bone formation in rat long bones: down regulation of gene expression for bone matrix proteins. Endocrinology 127:13461351 10 Wronski TJ, Walsh CC, Ignaszewski LA 1986 Histological evidence for osteopenia and increased bone turnover in ovariectomized rats, Bone 7:119-123 JT, Cintron M, Doherty SL, Dann LM 1988 Estrogen 11 Wronski treatment prevents osteopenia and depresses bone turnover in ovariectomized rats. Endocrinology 123:681-686 12 Jansson J-O, Eden S, Isaksson 0 1983 Sites of action of testosterone and estradiol on longitudinal bone growth. Am J Physiol 244:E135El40 S, Sanchez P, Sweet DE, Loriaux DL, 13 Ren SG, Malozowski Cassorla F 1989 Direct administration of testosterone increases rat tibia1 epiphyseal growth plate width. Acta Endocrinol (Copenh) 121:401-405 14 Budy AM, Urist MR, McLean FC 1952 The effect of estrogens on the growth apparatus of the bones of immature rats. Am J Path01 28:1143-1167 15 Frasier SD, Smith Jr FG 1968 Effect of estrogens on mature height in tall girls: a controlled study. J Clin Endocrinol Metab 28:416-419 16 Chomcznski P, Sacchi N 1987 Single-step method of RNA isolation by acid guanidinium theocyanate-phenol-chloroform extraction. Anal Biochem 162:156-159 17 Fink KL, Wieben ED, Woloschak GE, Spelsberg TC 1988 Rapid regulation of c-myc proto-oncogene expression by progesterone in
BONE
889
the avian oviduct. Proc Nat1 Acad Sci USA 85:1976-1980 GK, Carmichael GG 1977 Analysis of single and double18. McMaster stranded nucleic acids on polyacrylamide and agarose gels by using glyoxal and acridine orange. Proc Nat1 Acad Sci USA 74:4835-4838 19. Celeste Al, Rosen V, Buecker JL, Kriz R, Wang EA, Wozney JM 1986 Isolation of the human gene for the bone gla protein utilizing mouse and rat cDNA clones. EMBO J 5:1885-1890 20. Kuivaniemi H, Tromp G, Chu M-L, Prockop DJ 1989 Structure of a full-length cDNA clone for the pre pro o( 2 (1) chain of human type I procollagen. Comparison with the chicken gene confirms unusual patterns of gene conservation. Biochem J 252:633-640 21. Bowman LH, Rabin B, Schlessinger D 1981 Multiple ribosomal RNA cleavage pathways in mammalian cells. Nucleic Acids Res 9:4951-4966 22. Villarreal XC, Mann KG, Long GL 1989 Structure of human osteonectin based upon analysis of cDNA and genomic sequences. Biochemistry 28:6483-6491 23. Deleted in proof 24. Bell GI, Merryweather JP, Sanchez-Pescador R, Stempien MM, Pristely L, Scott J, Rall LB 1984 Sequence of a cDNA clone encoding human pre pro insulin-like growth factor II. Nature 30:775-777 25 Liu C, Sanghri R, Burnell JM, Howard GA 1987 Simultaneous demonstration of bone alkaline phosphatase and acid phosphatase activities in plastic-embedded sections and differential inhibition of activities. Histochemistry 86:559-565 26 Turner RT, Lifrak ET, Beckner M, Wakley GK, Hannon KS, Parker LN 1990 Maintenance of normal serum dehydroepiandrosterone reduces cancellous bone osteopenia in growing ovariectomized rats. Am J Physiol 258:E673-E677 27 Turner RT, Wakley GK, Hannon KS, Bell NH 1988 Tamoxifen inhibits osteoclast mediated resorption of trabecular bone in ovarian hormone-deficient rats. Endocrinology 122:1146-l 150 28 Turner RT, Spelsberg TC 1991 Correlation between mRNA levels for bone cell proteins and bone formation in long bones of maturing rats, Am J Physiol 261:E348-E353 FT 1969 Effect of 29 Riggs BL, Jowsey J, Kelly PJ, Jones JD, Maher sex steroids on bone in primary osteoporosis. J Clin Invest 48:10651072 TJ 1991 High concen30 Tobias JH, Chow J, Colston KW, Chambers tration of 170.estradiol stimulate trabecular bone formation in adult female rats. Endocrinology 128:408-412 31 Ernst M, Heath JK Rodan GA 1989 Estradiol effects proliferation, messenger ribonucleic acid for collagen and insulin-like growth factor-I, and parathyroid hormone-stimulated adenylate cyclase activity in osteoblastic cells from calvariae and long bones. Endocrinology 125:825-833 32 Keeting PE, Scott RE, Colvard DS, Han IK, Spelsberg TC, Riggs BL 1991 Lack of a direct effect of estrogen on proliferation and differentiation of normal human osteoblast-like cells. J Bone Mineral Res 6:297-304 33 Chen TL, Liu F, Bates RL, Hintz RL 1991 Further characterization of insulin-like growth factor binding proteins in rat osteoblast-like cell cultures: modulation by 17@-estradiol and human growth hormone. Endocrinology 128:2489-2496 34 Murphy LJ, Ghahary A 1990 Uterine insulin-like growth factor-l: regulation of expression and its role in estrogen-induced uterine proliferation. Endocr Rev 11:443-452 35. Murphy LJ, Friesen HG 1988 Differential effects of estrogen and growth hormone on uterine and hepatic insulin-like growth factor 1 gene expression in the ovariectomized hypophysectomized rat. Endocrinology 122:325-332 36. Mizukami Y, Nonomura A, Yamada T, Kurumaya H, Hayashi M, Koyasaki N, Taniya TM, Nakamura S, Matsubara F 1990 Immunohistochemical demonstration of growth factors TGF-n, TGF-0, IGF-I and new oncogene product in benign and malignant human breast tissues, Anticancer Res 10:1115-1126 37. McCarthy TL, Centrella M, Canalis E 1989 Insulin-like growth factor and bone. Connect Tissue Res 20:277-282
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
Vol. 131, No. 2 Printed in U.S.A.
Society
Mechanism of Action of Estrogen on Intramembranous Bone Formation: Regulation of Osteoblast Differentiation and Activity* RUSSELL T. TURNER, PEGGY BACKUP?, PAMELA GLENDA L. EVANS, AND THOMAS C. SPELSBERG
J. SHERMAN&
Departments of Orthopedic Surgery and Biochemistry and Molecular Medicine and Mayo Foundation, Rochester, Minnesota 55905
ESTHER
HILLS,
Biology, Mayo Graduate
School of
ABSTRACT Dynamic bone histomorphometry, 13H]thymidine radioautography, and Northern analysis for bone matrix proteins and insulin-like growth factor-I (IGF-I) were performed in calvariae of ovariectomized (OVX) and estrogen-treated OVX rats. Treatment of OVX rats with diethylstilbestrol (DES) for 2 weeks reduced the periosteal mineral apposition rate, osteoblast number, and osteoblast size in calvarial periosteum. DES treatment also reduced the number of preosteoblasts in the S phase of the cell cycle, suggesting that the decrease in osteoblast number was due in part to inhibition of proliferation of osteoprogenitor cells. One week after ovariectomy, there were small increases in mRNA levels for pre pro+2 (I) subunit of type I collagen (collagen), ostcocal-
tin, and osteonectin and a large increase in the mRNA level for IGF-I. DES treatment resulted in rapid decreases (3 h) in the mRNA levels for osteonectin, osteocalcin, and IGF-I. In contrast, mRNA levels for collagen were virtually unchanged after short term DES treatment. Uterus and liver served as positive and negative control tissues, respectively, for the effects of DES on IGF-I mRNA levels in OVX rats; mRNA levels were increased in uterus and decreased in liver after hormone treatment. We conclude from these studies that estrogen reduces periosteal bone formation by inhibiting both the differentiation and activity of osteoblasts. Furthermore, down-regulation of mRNA levels for IGF-I and bone matrix proteins precedes the changes in dynamic bone histomorphometry. (Endocrinology 131: 883-889, 1992)
G
creasedduring growth by intramembranous bone formation exclusively, whereas changesin bone volume in long bones are a summation of several processes,including longitudinal growth, endochondral ossification, modeling, and secondary intramembranous bone formation. Studies in long bones may, therefore, be complicated by estrogen having different actions on periosteal addition (5-8), endosteal and cancellous bone modeling (10, 1l), longitudinal growth (12), and endochondral ossification (12-15). Furthermore, in long bones there are possibleinteractions via locally produced signaling molecules (growth factors) among the various bone compartments, including periosteum, endosteum, metaphysis, growth plate, and epiphysis. Becausethe effects of sex steroids on bone cell metabolism in flat boneshave been largely unexplored, our present study is intended to provide initial characterization of the actions of estrogen on osteoblast differentiation, expression of mRNA for selected bone proteins, and bone matrix deposition in calvarial periosteum from growing female rats.
ONADAL hormones are implicated in the sexual dimorphism of the rat skeleton, including the flat bones of the skull (1, 2), pelvis (3), and tubular long bones (4-7). Ovariectomy (OVX) results in increased periosteal bone formation in tibiae of growing rats (5-8), whereas orchiectomy results in decreased periosteal bone formation (5, 6). Although gonadectomy reduces the sex differences in radial bone growth, these differences are reestablished by the administration of estrogens to ovariectomized (OVX) females (7) and androgens to orchidectomized males (6). In long bones, the inhibitory effects of estrogen on periosteal bone formation are accompanied by decreasesin mRNA levels for bone matrix proteins and alkaline phosphatase,but not glyceraldehyde-3-phosphate dehydrogenase, suggesting that the hormone regulates the expression of specific bone cell proteins (9). Estrogenic effects on long bone periosteum raise the question of whether the changesin gene expressionoccur in other skeletal sites. The potential effects of estrogen on flat bones are of interest because in these bones bone volume is in-
Materials
Received January 30, 1992. Address all correspondence and requests for reprints to: Russell T. Turner, 3-71 Medical Sciences Building, Mayo Clinic, Rochester, Minnesota 55905. *This work was supported by Grants AR-35651, AR-41418, and AG04875 from the NIH, the Mayo Foundation, and the National Osteoporosis Foundation. t Recipient of a National Osteoporosis Foundation Student Research Fellowship. $ Current address: Brown University Medical School, Providence, Rhode Island 02912. 5 Current address: University of Miami, Miami, Florida 33103.
and Methods
Animals Ten-week-old sham-operated and OVX female Sprague-Dawley rats were obtained from Harlan Sprague-Dawley (Indianapolis, IN). The rats were maintained in group cages (three or four rats per cage) and fed rat chow and water ad libitum.
Experiments Exp 1 characterized the histology periosteum-free calvariae, and excised
of intact calvarial periosteum, periosteum. A total of 10 OVX
883
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
ESTROGEN
884
ACTION
ON INTRAMEMBRANOUS
rats were studied. The intact calvariae were harvested from 1 group of 5 animals and fixed in 70% ethanol. The calvarial periosteum was harvested from rats in the second group as described below, and the excised periosteum and periosteum-free calvariae were fixed in 70% ethanol. Exp 2 was a histological study to determine the effects of 2 weeks of diethylstilbestrol (DES) treatment on periosteal bone formation in calvariae. A total of 24 OVX rats were divided into 2 equal groups. The OVX control group consisted of OVX rats implanted SC on the back of the neck with drug-free pellets; the estrogen-treated group consisted of OVX rats implanted with pellets containing DES. The pellets were rate (Innovative designed to deliver DES (2.5 mg/3 weeks) at a uniform Research, Toledo, OH). Tetracycline (tetracycline hydrochloride, Sigma Chemical Co., St. Louis, MO) was administered (20 mg/kg, ip) to each of the rats 7 days after implantation of the DES or drug-free pellets and again 13 days after the implantation. Tetracycline becomes deposited at the mineralizing front of newly deposited bone matrix and is used as a marker to determine the extent of appositional bone growth during a defined interval. All of the rats were killed with CO* gas 24 h after the second tetracycline labeling. After death, the calvariae were excised and fixed in 70% ethanol before processing for measurement of dynamic bone histomorphometry. The uteri were trimmed, excised, and weighed after gently expelling fluid from their lumen. Exp 3 was a cell kinetics study to determine the effects of estrogen on the number of osteoprogenitor cells in the S phase of the cell cycle. A total of 10 OVX rats were divided into 2 groups of 5 animals each. The OVX control group was implanted with drug-free pellets; the estrogen-treated group consisted of OVX rats implanted with DES. Six days after implantation, all of the rats were given an ip injection of [methyl-3H]thymidine (Amersham, Arlington Heights, IL; 45 Ci/mmol) at 1 pCi/g BW. Eighteen hours later, the rats were killed, and the intact calvariae were harvested and processed, as described below. Exp 4 characterized the effects of OVX and short term treatment with DES on mRNA levels for bone matrix proteins and insulin-like growth factor-I (IGF-I) in periosteum from calvariae. Five groups of 10 rats were studied. Two of the groups consisted of sham-operated animals, whereas the rats in the remaining 3 groups were OVX. One week after surgery, 1 group of OVX rats was given a single ip injection of DES (5 pg delivered in 0.1 ml 0.05% ethanol) and killed 3 h later. The remaining groups of sham-operated and OVX animals received carrier only. The calvarial periosteum was harvested after death, as described below, and the total cellular RNA was extracted. Exp 5 and 6 were carried out to verify the changes in mRNA levels in calvarial periosteum after short term DES treatment. Twenty OVX rats were studied in each of these experiments. One week after surgery, DES was administered to half of the rats, and the animals were killed about 3 h later. The calvarial periosteum, liver, and uteri were rapidly excised and frozen on dry ice before extraction of total cellular RNA.
Isolation
of total cellular RNA from calvarial periosteum
The skin on the head was moistened with water, shaved, and removed to expose the cranium. The superficial delicate fascia was incised on both sides of the midsagittal and coronal sutures, over the frontal and parietal bones, and parallel to the insertion of the temporalis muscularis. The periosteum, excluding the sutures, was then gently lifted off the bone surface using a dental probe and either fixed in 70% ethanol for histology or frozen on a block of dry ice before isolation of total cellular RNA. The frozen periosteum was pooled, homogenized, and extracted in guanidine hydrochloride (Gdn-HCl) (16), and the total cellular RNA was recovered (1.5-2 pg/calvaria) after CsCl gradient centrifugation using a SW50.1 rotor (Beckman Instruments, Berkley, CA) at 36,000 rpm for 16 h at 22 C (17). The frozen liver and uteri were homogenized and treated thereafter in the same manner as calvariae.
Electrophoresis The total cellular RNA from calvarial periosteum was resuspended water and separated electrophoretically on a 1% (wt/wt) agarose Each lane contained 15 pg total cellular RNA, which was denatured incubation for 1 h at 52 C in a solution of 1 M glyoxal dimethylsulfoxide
in gel. by
Endo. 1992 Voll31 *No 2
BONE
50% (vol/vol) in 10 mM sodium phosphate (pH 7.0). Electrophoresis was performed in 10 mM sodium phosphate (pH 7.0) at 60-70 V for about 4 h (17).
Northern
blot analyses
The RNA was transferred overnight from the gel to Magna 66 nylon membranes (MSI, Fisher Scientific, Pittsburg, PA) by capillary action in 20 x SSC (3 M NaCl and 0.3 M trisodium citrate), as described previously for the transfer of DNA (18). The filters were baked for 2 h at 80 C and prehybridized overnight. The 32P-labeled cDNA probes were prepared by random sequence hexanucleotide primer extension using Multiprime DNA labeling systems from Amersham or equivalent reagents. [32P]Deoxy-CTP (SA, -3000 Ci/ml; New England Nuclear Research Products, Boston, MA) was used to radiolabel the cDNA. Specific activities of about lo9 cpm/ pg DNA were achieved. The following templates were used for probe production: 1) rat osteocalcin (BGP), a gift from Dr. S. Rossi-Langer, Genetics Institute (Cambridge, MA); this full-length probe is pR 22-11, which is an EcoRI insert into pSI’65 (19); 2) human collagen, a gift from Drs. H. Hivanlem and G. Tromp, Jefferson Medical College (Philadelphia, PA); this is PHUCI, which is a full-length cDNA clone for the prepro-cu2 (I) chain of human type 1 precollagen inserted into the EcoRI site (20); 3) mouse 185 ribosomal RNA (21); this is an 1150-basepair (bp) cDNA insert into the BamHI site; 4) human osteonectin (ON), a gift from Dr. G. Long, University of Vermont (Burlington, VT) (22); this is pHVON-9-2 plasmid DNA containing a 546.bp human ON cDNA insert; and 5) human IGFI, a gift from Dr. G. Bell, Howard Hughes Medical Institute Research Laboratories, University of Chicago (Chicago, IL); this is plasmid phigf 1 containing an insert of about 660 bp, which extends from the second nucleotide of the codon for amino acid -15 of the single peptide of the IGF-I precursor to the poly(A) tract (24). The blots were hybridized overnight at 43 C with the 3ZP-labeled cDNA probes. After hybridization, the blots were washed under conditions of increasing stringency, up to 0.1 X SSC and 0.1% sodium dodecyl sulfate at 65 C, and then exposed to Kodak X-Omat AR-5 film (Eastman Kodak, Rochester, NY).
Densitometry
of the autoradiographs
The Northern blots were quantitated by densitometric autoradiographs, using a Shimadzu flying spot scanner (Shimadzu Scientific Instruments, Columbia, MD).
analysis of the model CS-9000
Histomorphometry After fixation for 2 days, the fixed tissues from Exp 1 and 2 were dehydrated in a series of increasing concentrations of ethanol, embedded without demineralization in a mixture of methylmethacrylate:2hydroxyethyl-methacrylate (12.5:1) to retain the tetracycline, and sectioned at an indicated thickness of 5 pm. The intact calvariae, periosteum-free calvariae, and excised periosteum from Exp 1 were stained either with toluidine blue or for alkaline phosphatase (25). The intact calvariae from Exp 2 were viewed unstained with a fluorescence microscope (Olympus BH-2, New Hyde Park, NY) to detect the tetracycline labeling. Histomorphometric procedures were carried out using a SMI-Microcamp semiautomatic image analysis system (Southern Micro Instruments, Inc., Atlanta, GA), consisting of a Compaq computer with Microcamp software interphased with a microscope and image analysis system. In this system, a high resolution color video camera mounted on an Olympus BH-2 microscope displays the image of the specimen on a color video monitor. The movement of a pen on a graphics tablet superimposes a tracing on the image of the specimen on the video screen. By this method, the region of interest is traced, and the line length and the area bounded by the tracing are calculated. The periosteal mineral apposition rate provides a quantitative assessment of bone matrix deposition and was determined as the mean of the distance between the two tetracycline labels measured every 50 pm (minimum of 50 measurements/section) along the periosteal perimeter divided by
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
ESTROGEN
ACTION
ON INTRAMEMBRANOUS
the labeling interval, which was 6 days. Osteoblast number and area were determined in toluidine-stained sections. pH]Thymidine
radioautography
Intact calvariae from Exp 3 were demineralized in 5% formic acid in 10% formalin for 4 days, dehydrated in a series of increasing concentrations of ethanol, embedded in glycolmethacrylate (Sorval, Norwalk, CT),
885
BONE
TABLE 1. The effects of DES on bone and uterine measurements in OVX rats Measurement Exp 1 Calvaria Mineral apposition rate tw/day) Uterus Wet ti tg) Exp 2 Calvaria Osteoblast no. (cells/ mm) Osteoblast area (lm*/ cell) Labeled preosteoblasts (cells/mm)
ovx
OVX + DES % Change
3.11 f 0.17 1.46 + 0.16”
-53
0.15 f 0.01 0.74 rfr 0.08”
+393
104 + 6
72 f 2’
-31
117 j, 10
55 zk 18*
-53
3.10 & 0.47 0.02 + 0.02”
-99
Values are the mean + SEM [n = lo-14 (exp 1) and 4-5 (Exp 2)]. ’ P < 0.001, by Student’s t test. *P < 0.01, by Student’s t test. and sectioned at an indicated thickness of 5 pm. The sections were attached to unsubbed acid-cleaned slides, dipped in the dark into melted (50 C) Ilford k50 nuclear tract emulsion, and exposed for 1 week. The sections were developed in Kodak D-19 and stained with toluidine blue before viewing with an Olympus BH-2 microscope to detect [3H]thymidine-labeled preosteoblasts. A cell with six or more grains per nucleus was considered labeled. Cells one or more cell layers removed from the osteoblast monolayer lining the periosteal perimeter and within the cambial layer were considered to be preosteoblasts, and the labeled cells were enumerated. We demonstrated in preliminary studies that with increasing time after the administration of radiolabeled thymidine, the number of labeled osteoblasts increased, and there was a corresponding decrease in the labeled putative preosteoblast population (data not shown). Approximately 10,000 cells were scanned per group, and the data were expressed as labeled preosteoblasts per mm periosteal perimiter.
Results
FIG. 1. The morphology of the calvarial periosteum. A, An intact calvaria stained with toluidine blue, showing the organization of the periosteum. Bone occupies the bottom portion of this panel. A monolayer of osteoblasts (large arrows) lines the mineralized surface of the bone. Less differentiated osteoprogenitor cells (small arrows) are found in this cellular layer, bounded by osteoblasts and the fibrous region. The outer fibrous region is composed chiefly of fibroblasts, but also includes fat cells, nerve cells, and epithelial cells. B, Isolated periosteum, showing the in situ localization of alkaline phosphatase activity (arrows) in the osteoblast cell layer. The tissue section was counterstained with methyl green thionine. The orientation is similar to that in A, with no bone present. C, Periosteum-free calvaria, showing the absence of alkaline phosphatase activity after complete removal of periosteal cells. The section was counterstained with methyl green thionine. The orientation is similar to that in A, with no cellular layer present. The horizontal bar is 25 pm in length.
The histology of an intact calvaria, a periosteum-free calvaria, and the harvested periosteum is shown in Fig. 1. The periosteum of an intact calvaria (Fig. 1A) is divided anatomically into two developmentally related, but spatially and metabolically distinct regions. The outer fibrous region is composed chiefly of fibroblasts, bundles of collagenous fibers, and fat cells. The inner cellular region contains cells of the osteoblast lineage, the preosteoblasts and osteoblasts. Osteoblasts are present in the innermost (or cambial) layer of the cellular portion, closely opposed to the bone surface, and are separated from the fibrous periosteum by one or more layers of preosteoblasts.Preosteoblastsdiffer morphologically and functionally from fibroblasts of the fibrous portion of the periosteum, in that the former contain more cytoplasm per cell and are committed in an osteogenicdirection. Osteoclasts are rarely found on the periosteal surface and when observed are associatedwith developing vascular spaces.The procedure for harvesting the periosteum for RNA analysis separated the cellular periosteum (Fig. 1B) from the remainder of the calvaria (Fig. 1C) at the osteoid seamand was effective in recovering the osteoblastslining the periosteal surface. Osteoblasts were identified in intact calvariae (Fig. 1A) and in isolated periosteum (Fig. lB), but not in periosteum-free calvaria (Fig. lC), by criteria which included
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
886
ESTROGEN
ACTION
ON INTRAMEMBRANOUS
BONE
Endo. 1992 Vo1131.No2
BGP123456 FIG. 3. Representative Northern blot of periosteal and liver cell total cellular RNA. Each lane represents RNA-pooled from 10 animals. The blot was successivelv nrobed with the 32P-labeled cDNAs for ON (ton band), 18s ribosomaiRNA (middle band), and BGP (bottom baidj. The bands were then identified based on previously reported sizes of the RNA transcripts for these genes, which hybridized with specific cDNA probes complementary for the mRNA species. The calvariae were from rats that-were intact (lane l), OVX (lanes 2 and 3), or OVX and treated with DES for 3 h before death (lane 4 and 5). The livers were from OVX rats (lane 6). Note that the expression of dN and BGP was below the detection limits for liver.
FIG. 2. [3H]Thymidine incorporation into nuclei of preosteoblasts. A, Radioautograph of intact calvarial periosteum (sectioned and stained with toluidine blue) of an OVX rat. Bone occupies the bottom portion of this panel. A preosteoblast in the S phase of the cell cycle is identified by grains over the nucleus (small arrow). The periosteal perimeter is lined bv a monolaver of osteoblasts (lar,e arrows). B, Radioautomaph of calvarial periosieum (sectioned and stained with toluidine blue)-of an OVX rat treated with DES for 6 days. Bone occupies the bottom portion of this panel. No labeled preosteoblasts are observed in this photograph, they were rarely observed in DES-treated rats. The periosteal perimeter is lined by smaller, less robust cells (arrows) than those in A. The horizontal bar is 25 pm in length.
size, location, intense basophilic staining with toluidine blue, a prominent Golgi apparatus, and high levels of alkaline phosphatase activity. The effects of 2 weeks of DES treatment on bone measurements and uterine wet weight in OVX rats are shown in Table 1. DES treatment resulted in a highly significant 53% decreasein the mineral apposition rate of calvarial periosteum between days 7 and 13 of treatment. In contrast, DES treatment of OVX rats resulted in a dramatic 393% increase in uterine weight. Seven days of treatment of OVX rats with DES resulted in decreasesin osteoblast number and size as well as a dramatic decreasein the number of [3H]thymidinelabeled preosteoblasts. Representative [3HJthymidine radioautographs of the calvarial periosteum are illustrated in Fig. 2. The periosteal perimeter of the calvariae of OVX rats
(Fig. 2A) was typically lined by plump cells (large umws) that expressedosteoblastcharacteristics, including basophilic staining, an asymmetrically located nucleus, and prominent Golgi apparatus; [3H]thymidine-labeled preosteoblasts(small arrow) were observed. In contrast, the periosteum of DEStreated rats (Fig. 2B) was typically lined by much smaller cells (arrows) that did not express features associatedwith active osteoblasts,and the putative preosteoblastswere rarely labeled with [3H]thymidine (Table 1). Single bands were detected by Northern analyses at the expected sizes for BGP, ON, and 18s ribosomal RNAs after hybridization with the radioactively labeled cDNA probes. The expected doublet was observed for type 1 collagen. As expected, four transcripts were identified for IGF-I. These were located at 7.5, 3.9, 1.8 and 0.7-1.2 kilobases(kb). The most prominant transcript expressedin liver, uteri, and calvariae was at 7.5 kb. Representative Northern blots are shown for BGP, ON, 18s ribosomal mRNA, and IGF-I mRNA (Figs. 3 and 4). Northern blots for collagen are not shown, but have been described previously (9). The effects of OVX on mRNA levels for selected bone proteins in calvarial periosteum are summarized in Fig. 5. OVX resulted in large increasesin mRNA levels for IGF-I and smaller increases in mRNA levels for ON, BGP, and collagen. The effects of short term estrogen treatment on mRNA levels for bone proteins are shown in Fig. 6. Administration of DES for about 3 h resulted in decreasesin mRNA levels for IGF-I, BGP, and ON and a negligible change in collagen.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
ESTROGEN
ACTION
ON INTRAMEMBRANOUS
BONE
887
12 BGP
~
/
1234567 FIG. 4. The Northern blot analyses shown in Fig. 3 were washed and rehybridized with the cDNA for IGF-I, as described in Fig. 3. Liver expressed higher mRNA levels for IGF-I then periosteum. The most prominent IGF-I transcript was located at 7.5 kb. Other transcripts were noted at 3.9, 1.8, and 0.7-1.2 kb.
Interestingly, after DES treatment, IGF-I mRNA levels were increasedin uterus and decreasedin liver (Fig. 7). Discussion Previous studieshave shown that OVX results in increases in periostealbone formation and apposition rates in the tibia1 diaphysis of growing rats (5-8). In contrast, OVX had little effect on the tetracycline-labeled perimeter (5-8, 12). These findings were originally interpreted to mean that ovarian hormone deficiency results in increased osteoblast activity, with little change in osteoblast number. However, it was assumedthat the tetracycline-labeled perimeter was directly proportional to osteoblastnumber, an assumptionwhich this study found to be incorrect for calvarial periosteum. Although 100% of the periosteal perimeter was labeled with tetracycline in calvariae from DES-treated rats, there was a 3 1% decreasein the number of osteoblastsper mm periosteal perimeter and a 53% decreasein osteoblast size. Thus, the estrogen-induced decreasein periosteal bone formation seen in this study was very likely due to a combination of reduced osteoblastnumber and reduced osteoblastactivity. Estrogen has similar effects on periosteum in long bones (5-9) and calvariae to inhibit bone formation. Studies of the tibia1 metaphysis have also shown that estrogen inhibits the formation of cancellous bone (10). In spite of similarities in
12 ON
FIG. 5. The effects of OVX on mRNA levels for IGF-I and bone matrix proteins. One week after OVX, rats were killed, and total cellular RNA was extracted, separated by size, and hybridized with [32P]cDNA probes, and the hybridization products were identified by radioautography. mRNA levels were determined after densitometry of the x-ray films and normalization to 18s ribosomal RNA to correct for differences in total RNA. The data for OVX rats are expressed as a percentage of the value in the ovary-intact control group. The results for two separate determinations (labeled 1 and 2) are shown. The dotted line represents the value for the intact control groups. COLL, Collagen.
the respective responses of cancellous and cortical bone compartments to estrogendeficiency, there are alsoprofound differences; there is a net increase in bone added to the periosteal surface of the tibia1 diaphysis after OVX, whereas there is a pronounced decreasein cancellousbone volume in the metaphysis (26, 27). BecauseOVX produces an increase in bone formation at both sites,the different results observed are probably due to the relative amount of bone resorption in the respective skeletal compartments. Osteoclastsare common in cancellous bone, and osteoclastnumber is increased after OVX (11, 26). In contrast, osteoclastsare uncommon to the periosteal surface of long bones and calvariae and are limited to the developing vascular spaces. Erosion of the periosteal surface does not occur in OVX rats, indicating that bone resorption is not increased by estrogen deficiency at this site. Prior studies in long bones were limited by the procedure used to isolate periosteal cells, which involved prolonged treatment of the boneswith collagenaseat 37 C. This method precluded investigation of rapid changes in gene regulation (9). Furthermore, we found that collagenase treatment of liver from OVX rats obliterates expressionof IGF-I mRNA in a tissue that expresseshigh levels of the message(unpublished data). The novel method used to separate periosteum from calvariae described in this manuscript now allows detection of mRNAs for rapidly regulated cell-signaling molecules, such as IGF-I, as well as for structural proteins. We show for the first time that OVX increases mRNA levels for bone matrix proteins and IGF-I in a skeletal tissue. The changes in mRNA levels for the bone matrix proteins
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
ESTROGEN
888
ACTION
ON INTRAMEMBRANOUS
BONE
Endo. Voll31.
1992 No 2
1.25
12
3
4
5
6
7
8
7. Representative Northern blot of liver and uterus total cellular RNA fromOVX rats. The blot was probed with the 32P-labeled cDNAs for IGF-I. as described in Fie. 3. Lanes l-4 were loaded with RNA from liver, whereas lanes 5-8 we& loaded with RNA from uterus. The oddnumbered lanes had RNA from solvent-treated rats, and the evennumbered lanes had RNA from DES-treated animals. Interestingly, DES treatment down-regulated IGF-I steady state mRNA levels in liver and up-regulated the message in uterus.
FIG.
O.o(
IGF-I
COLL
BGP
ON
6. The effects of DES treatment on mRNA levels for IGF-I and bone matrix proteins. One week after OVX, rats were given 5 pg DES or solvent. The DES-treated and control rats were killed and assayed as described in Fig. 5. The data are expressed as a percentage of the control value. The values are the mean ? SEM (n = 3). **, P < 0.01, by Student’s t test. COLL, Collagen. FIG.
are modest and consistent with the increase in bone formation determined by histomorphometry (5-7). We have previously shown an excellent correlation between indices of bone formation determined histologically and mRNA levels for collagen (28). The changes in mRNA levels after OVX are probably due to estrogen deficiency, rather than deficiency of other ovarian hormones, because DES treatment reversed the direction of the changes. The current studies support and extend earlier studies in long bones which showed that long term (l- to 2-week) treatment with estrogen reduces bone matrix synthesis and mRNA levels for bone matrix proteins and alkaline phosphatase (7, 9). The effects of estrogen on expression of IGF-I mRNA have not previously been investigated in periosteum; indeed, IGF-I has not previously been shown to be expressedin this skeletal tissue in viva Further studies will be necessary to identify the periosteal cell population(s) that expressesIGF-I mRNA and to verify that the changes in messagelevel reflect peptide synthesis. The present study demonstratesthat in calvariae, estrogen has rapid effects on mRNA levels for BGP, ON, and IGF-I. Interestingly, OVX and DES had minimal effects on mRNA levels for collagen. Although DES treatment ultimately reduced collagen mRNA levels in periosteum from tibiae and femora, these changes were not detected during the first week of estrogen treatment (9). The effect of estrogen on osteoblastsin other skeletal compartments is controversial.
The majority (11, 29), but not all (30), histomorphometric studies in humans and animals suggeststhat estrogen also inhibits the turnover of cancellous bone. The present study in calvariae strengthens the argument that estrogen affects osteoblasts similarly regardless of their location. The net effect of estrogen on bone volume would, however, depend upon the relative abundance of osteoblastsand osteoclasts. The dramatic decreasein labeled preosteoblastsin DEStreated OVX rats suggeststhat reduced proliferation of preosteoblasts with a consequent decreasein the production of osteoblasts was responsible for the reduction in osteoblast number in calvarial periosteum. The effects of estrogen on DNA synthesis may be direct or, alternatively, may be coupled to changes in bone cell activity, perhaps through decreased expression of paracrine signaling molecules. The results of this study are consistent with an indirect effect, in that IGF-I mRNA levels were found to be reduced in estrogen-treated calvariae. Estrogen has variable effects on cultured osteoblast-like cells and has been reported to increase, decrease, and not alter cell proliferation and mRNA and peptide levels for bone matrix proteins and IGF-I (31-33). IGF-I is believed to be an important mediator of estrogenic action on uterus (34), liver (35), and breast cancer cells (36) and may act asan autocrine/paracrine factor in skeletal tissue as well (37). Estrogen had opposite effects on IGF-I mRNA levels in uterus and liver, a finding that agreeswith previous studies (34, 35). The inhibitory effects of estrogen on intramembranous bone formation and proliferation of osteoprogenitor cells contrast markedly with the ability of the hormone to promote
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
ESTROGEN
ACTION
ON INTRAMEMBRANOUS
uterine growth. Interestingly, estrogen has been shown to up-regulate IGF-I mRNA levels in uterus (34), the opposite of its effects on calvarial periosteum and liver (35). Further studies will be necessary to determine the precise role of changes in expression of IGF-I in mediating the effects of estrogen on bone matrix synthesis and differentiation of osteoblasts. Acknowledgments The authors this manuscript.
thank
Ms. Carolyn
Blankenship
for typing
and proofing
References 1. Dahinten SL, Pucciarelli HM 1986 Variation in sexual dimorphism in the skulls of rats subjected to malnutrition, castration and treatment with gonadal hormones. Am J Physical Anthropol7:63-67 2. Riesenfeld A 1974 Endocrine and biochemical control of craniofacial growth: an experimental study. Hum Biol 46:531-572 3. Riesenfeld A 1972 Functional and hormonal control of pelvic morphology in the rat. Acta Anat 82:231-253 4. Hansson LI, Menander-Sellman K, Stenstrom A, Thorngren K-G 1972 Rate of normal longitudinal bone growth in the rat. Calcif Tissue Res 10:238-251 5. Turner RT, Hannon KS, Demers L, Buchanan J, Bell NH 1989 The differential effects of gonadal insufficiency on bone histomorphometry in male and female rats. J Bone Mineral Res 4:557-563 6. Turner RT, Wakley GK, Hannon KS 1990 Differential effects of androgens on cortical bone histomorphometry in gonadectomized male and female rats. J Orthop Res 8:612-617 RT, Vandersteenhoven JJ, Bell NH 1987 The effects of 7. Turner ovariectomy and 17 beta estradiol on cortical bone histomorphometry in growing rats. J Bone Mineral Res 2:115-122 8. Turner RT, Wakley GK, Hannon KS, Bell NH 1987 Tamoxifen prevents the altered bone turnover resulting from ovarian hormone deficiency. J Bone Mineral Res 2:449-456 9. Turner RT, Colvard DS, Spelsberg TC 1990 Estrogen inhibition of periosteal bone formation in rat long bones: down regulation of gene expression for bone matrix proteins. Endocrinology 127:13461351 10 Wronski TJ, Walsh CC, Ignaszewski LA 1986 Histological evidence for osteopenia and increased bone turnover in ovariectomized rats, Bone 7:119-123 JT, Cintron M, Doherty SL, Dann LM 1988 Estrogen 11 Wronski treatment prevents osteopenia and depresses bone turnover in ovariectomized rats. Endocrinology 123:681-686 12 Jansson J-O, Eden S, Isaksson 0 1983 Sites of action of testosterone and estradiol on longitudinal bone growth. Am J Physiol 244:E135El40 S, Sanchez P, Sweet DE, Loriaux DL, 13 Ren SG, Malozowski Cassorla F 1989 Direct administration of testosterone increases rat tibia1 epiphyseal growth plate width. Acta Endocrinol (Copenh) 121:401-405 14 Budy AM, Urist MR, McLean FC 1952 The effect of estrogens on the growth apparatus of the bones of immature rats. Am J Path01 28:1143-1167 15 Frasier SD, Smith Jr FG 1968 Effect of estrogens on mature height in tall girls: a controlled study. J Clin Endocrinol Metab 28:416-419 16 Chomcznski P, Sacchi N 1987 Single-step method of RNA isolation by acid guanidinium theocyanate-phenol-chloroform extraction. Anal Biochem 162:156-159 17 Fink KL, Wieben ED, Woloschak GE, Spelsberg TC 1988 Rapid regulation of c-myc proto-oncogene expression by progesterone in
BONE
889
the avian oviduct. Proc Nat1 Acad Sci USA 85:1976-1980 GK, Carmichael GG 1977 Analysis of single and double18. McMaster stranded nucleic acids on polyacrylamide and agarose gels by using glyoxal and acridine orange. Proc Nat1 Acad Sci USA 74:4835-4838 19. Celeste Al, Rosen V, Buecker JL, Kriz R, Wang EA, Wozney JM 1986 Isolation of the human gene for the bone gla protein utilizing mouse and rat cDNA clones. EMBO J 5:1885-1890 20. Kuivaniemi H, Tromp G, Chu M-L, Prockop DJ 1989 Structure of a full-length cDNA clone for the pre pro o( 2 (1) chain of human type I procollagen. Comparison with the chicken gene confirms unusual patterns of gene conservation. Biochem J 252:633-640 21. Bowman LH, Rabin B, Schlessinger D 1981 Multiple ribosomal RNA cleavage pathways in mammalian cells. Nucleic Acids Res 9:4951-4966 22. Villarreal XC, Mann KG, Long GL 1989 Structure of human osteonectin based upon analysis of cDNA and genomic sequences. Biochemistry 28:6483-6491 23. Deleted in proof 24. Bell GI, Merryweather JP, Sanchez-Pescador R, Stempien MM, Pristely L, Scott J, Rall LB 1984 Sequence of a cDNA clone encoding human pre pro insulin-like growth factor II. Nature 30:775-777 25 Liu C, Sanghri R, Burnell JM, Howard GA 1987 Simultaneous demonstration of bone alkaline phosphatase and acid phosphatase activities in plastic-embedded sections and differential inhibition of activities. Histochemistry 86:559-565 26 Turner RT, Lifrak ET, Beckner M, Wakley GK, Hannon KS, Parker LN 1990 Maintenance of normal serum dehydroepiandrosterone reduces cancellous bone osteopenia in growing ovariectomized rats. Am J Physiol 258:E673-E677 27 Turner RT, Wakley GK, Hannon KS, Bell NH 1988 Tamoxifen inhibits osteoclast mediated resorption of trabecular bone in ovarian hormone-deficient rats. Endocrinology 122:1146-l 150 28 Turner RT, Spelsberg TC 1991 Correlation between mRNA levels for bone cell proteins and bone formation in long bones of maturing rats, Am J Physiol 261:E348-E353 FT 1969 Effect of 29 Riggs BL, Jowsey J, Kelly PJ, Jones JD, Maher sex steroids on bone in primary osteoporosis. J Clin Invest 48:10651072 TJ 1991 High concen30 Tobias JH, Chow J, Colston KW, Chambers tration of 170.estradiol stimulate trabecular bone formation in adult female rats. Endocrinology 128:408-412 31 Ernst M, Heath JK Rodan GA 1989 Estradiol effects proliferation, messenger ribonucleic acid for collagen and insulin-like growth factor-I, and parathyroid hormone-stimulated adenylate cyclase activity in osteoblastic cells from calvariae and long bones. Endocrinology 125:825-833 32 Keeting PE, Scott RE, Colvard DS, Han IK, Spelsberg TC, Riggs BL 1991 Lack of a direct effect of estrogen on proliferation and differentiation of normal human osteoblast-like cells. J Bone Mineral Res 6:297-304 33 Chen TL, Liu F, Bates RL, Hintz RL 1991 Further characterization of insulin-like growth factor binding proteins in rat osteoblast-like cell cultures: modulation by 17@-estradiol and human growth hormone. Endocrinology 128:2489-2496 34 Murphy LJ, Ghahary A 1990 Uterine insulin-like growth factor-l: regulation of expression and its role in estrogen-induced uterine proliferation. Endocr Rev 11:443-452 35. Murphy LJ, Friesen HG 1988 Differential effects of estrogen and growth hormone on uterine and hepatic insulin-like growth factor 1 gene expression in the ovariectomized hypophysectomized rat. Endocrinology 122:325-332 36. Mizukami Y, Nonomura A, Yamada T, Kurumaya H, Hayashi M, Koyasaki N, Taniya TM, Nakamura S, Matsubara F 1990 Immunohistochemical demonstration of growth factors TGF-n, TGF-0, IGF-I and new oncogene product in benign and malignant human breast tissues, Anticancer Res 10:1115-1126 37. McCarthy TL, Centrella M, Canalis E 1989 Insulin-like growth factor and bone. Connect Tissue Res 20:277-282
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 06:26 For personal use only. No other uses without permission. . All rights reserved.
