Brain edema in meningiomas is associated with increased vascular endothelial growth factor expression

OBJECTIVE: Vascular permeability factor/vascular endothelial growth factor (VPF/VEGF), an endothelial cell-specific cytokine, induces proliferation of endothelial cells and increases vascular permeability dramatically. All gliomas secrete significant amounts of VEGF, whereas meningiomas are variable in expression. Thus, we sought to determine whether the extent of VPF/VEGF expression in meningiomas correlated with differences in brain edema associated with these tumors. METHODS: Meningioma tissue samples from 37 patients (15 men, average age 65 +/- 13 yr; 22 women, average age 60 +/- 10 yr) who underwent surgery at or were referred to the University of Alabama Hospital were examined retrospectively for the extent of expression of immunoreactive VPF/VEGF. Additionally, peritumoral edema was assessed on a blinded basis radiographically from preoperative magnetic resonance imaging scans. Selected specimens were examined by in situ hybridization to document the source of VPF/VEGF. RESULTS: The predominant meningioma subclassifications were transitional (57%) or meningothelial (27%) subtypes. VPF/VEGF immunoreactivity ranged from 0 to 3.5, with a median value of 2 on a subjective 5-point scale; magnetic resonance imaging-assessed edema ranged in extent from 0 to 4 (subjective 5-point scale), with a median value of 2.5. The correlation of determination (R2) of magnetic resonance imaging-assessed tumor edema rating and VPF/VEGF staining intensity rating was 0.6087 (r = 0.78; P = 0.0001). In situ hybridization localized VPF/VEGF messenger ribonucleic acid in meningioma cells and not in normal parenchymal brain cells. CONCLUSION: These data suggest that meningioma-associated edema may be a result of the capacity of meningioma cells to produce VPF/VEGF locally, leading to increased tumor neovascularization and enhanced vascular permeability.     

Growth hormone receptor activity is stimulated by insulin-like growth factor binding protein 5 in rat osteosarcoma cells

Osteoblast-like UMR-106.01 rat osteosarcoma cells express high affinity growth hormone (GH) receptors (GHRs). Because osteoblasts secrete insulin-like growth factor binding protein-5 (IGFBP-5), we evaluated whether it also modulates GH binding and GHR expression in UMR cells. Human recombinant intact IGFBP-5 stimulated 125I-hGH binding in a dose-dependent manner (dose range 300-3000 ng/ml), inducing an increase to 193.6 +/- 2.1% of control binding at 3000 ng/ml (P < 0.001). Carboxy-truncated IGFBP-5 also stimulated GH binding but with less potency (125 +/- 2.7% of control at 3000 ng/ml, P < 0.01). GHRs identified by chemical crosslinking of 125I-hGH to cell monolayers increased after treatment with IGFBP-5 and decreased in response to insulin-like growth factor-I (IGF-I). GHR mRNA levels, as quantitated by a solution hybridization RNAse protection assay, increased up to 3 to 7-fold in a time-dependent manner by intact IGFBP-5 but not by carboxy-truncated IGFBP-5. An antiserum to IGFBP-5 reduced basal GH binding to 56.7 +/- 4.3% of control value at a concentration of 0.5% (P < 0.001), showing that IGFBP-5 produced by the cells is a strong regulator of GH binding. IGFBP-5 antiserum also decreased GH binding to 85.9 +/- 0.9% of IGFBP-5 stimulated value (P < 0.001), showing the specificity of IGFBP-5 stimulation. To determine whether the GHR upregulation was physiologically significant, cell proliferation was evaluated after coincubation of IGFBP-5 with low, non-stimulatory concentrations of GH. IGFBP-5 (1000 ng/ml) induced cell proliferation to 116.2 +/- 3.2% of control levels, and coincubation with hGH at 10 ng/ml induced an increase to 133.3 +/- 0.1% of control levels. We conclude that exogenous and endogenous IGFBP-5 upregulate GHR mRNA levels and GH binding and this interaction potentiates GH-stimulated mitogenesis in osteoblastic cells.     

Stimulation of intestinal growth is associated with increased insulin-like growth factor-binding protein 5 mRNA in the jejunal mucosa of insulin-like growth factor-I-treated parenterally fed rats

Surgically stressed rats maintained with total parenteral nutrition (TPN) exhibit jejunal atrophy, which can be attenuated by insulin-like growth factor-I (IGF-I) but not by growth hormone (GH) treatment. In order to understand the basis for the selective action of IGF-I, the levels of mRNAs encoding IGF-I, IGF-binding proteins (IGFBPs), IGF-I receptor, and GH receptor/binding protein (GHR/GHBP) were determined in rats given TPN and treated with GH, IGF-I, or GH + IGF-I. GH treatment significantly stimulated hepatic IGF-I mRNA. IGF-I treatment did not alter liver IGF-I mRNA, nor was there any evidence for interaction between GH and IGF-I. Jejunal mucosa IGF-I mRNA was extremely low and was not altered by TPN or by any of the hormonal treatments. The inability of GH to stimulate jejunal growth was not associated with a deficiency in GHR/GHBP mRNA. In jejunal mucosa, IGF-I and GH treatment independently and synergistically stimulated IGFBP-3 mRNA. IGF-I stimulated jejunal IGFBP-5 mRNA, but GH had no effect on IGFBP-5 mRNA. The levels of IGF-I receptor and IGFBP-1, 2, 4, and 6 mRNAs were extremely low and/or were not altered by any of the treatments. These results suggest that the ability of exogenous IGF-I, but not GH, to induce IGFBP-5 mRNA in jejunal mucosa may lead to the selective growth-promoting effect of IGF-I. Jejunal mucosa IGFBP-3 mRNA levels were not correlated with altered growth. We postulate that IGFBP-5 positively modulates the anabolic effects induced by exogenous IGF-I in the jejunum.     

Controlling insulin-like growth factor activity and the modulation of insulin-like growth factor binding protein and receptor binding

The insulin-like growth factors (IGF) and insulin perform seemingly unique roles by causing the same metabolic effect: cellular hypertrophy. Although overlapping, there are different consequences to cellular hypertrophy induced by IGF and that induced by insulin. The IGF enhance the cell hypertrophy that is requisite for cell survival, hyperplasia, and differentiation, and insulin enhances cell hypertrophy primarily as a means to increase nutrient stores. The effects of IGF and insulin are controlled by the segregation of their receptors between different cell types. A model is discussed that describes the need for three hormones (IGF-I, IGF-II, and insulin) to control nutrient partitioning. Insulin receptor localization, as well as an episodic mode of secretion, evolved to perform the short-term action of clearing excess nutrients from the circulation. In contrast, a complex and interactive set of factors ensure that maximal IGF activity occurs only when conditions are optimal for growth. A relatively invariant rate of secretion and the IGF binding proteins serve to maintain a large mutable pool of IGF. This pool exists to ensure a constant supply of IGF to maintain the basal metabolic rate and to ensure that, once a cell begins to proliferate or differentiate, adequate exposure is available to complete the process even after severe short-term physiological insults. The IGF concentrations only change in response to prolonged differences in protein and energy availabilities, environmental and body temperatures, and external stress. Also, evidence is now emerging that describes a discrete role for trace nutrients in the regulation of IGF activity. In this latter regard, zinc has the notable role of targeting IGF binding proteins to the cell surface. New data are presented showing that zinc also changes the affinity of the type 1 IGF receptor and cell-associated IGF binding proteins to optimize IGF activity.     

Expression of insulin-like growth factor binding protein-4 and -5 mRNAs in adult rat forebrain

Accumulating evidence indicates that the insulin-like growth factors (IGFs) can act as neurotrophic factors. A family of at least six IGF binding proteins (IGFBPs) has been characterized. The IGFBPs prolong the half-life of IGFs in plasma and may modulate IGF action in a cell- or tissue-specific fashion. Two recently characterized IGFBPs, IGFBP-4 and -5, have been shown by northern blot hybridization to be expressed in rat brain, but their cellular sites of synthesis are poorly characterized. Because IGFBP-4 and IGFBP-5 could potentially modulate IGF actions in the brain, we used in situ hybridization histochemistry and 35S-labeled IGFBP-4 and IGFBP-5 riboprobes to localize sites of IGFBP-4 and -5 mRNA expression in adult rat brain. The two IGFBP mRNAs are abundantly expressed within discrete regions of brain. The expression patterns of the two genes are largely nonoverlapping. Notably, IGFBP-4 mRNA is highly expressed within hippocampal and cortical areas, whereas IGFBP-5 mRNA is not detected above background in these areas. Within the hippocampus, abundant IGFBP-4 mRNA expression is detected in pyramidal neurons of the subfields of Ammon's horn and the subiculum and in the granule cell layer of the anterior hippocampal continuation. In the cortex, IGFBP-4 mRNA is widely expressed in most areas and layers. In contrast, IGFBP-5, but not IGFBP-4, mRNA is detected within thalamic nuclei, leptomeninges, and perivascular sheaths. The distinct expression patterns of IGFBP-4 and -5 mRNAs within the brain suggest that these IGFBPs may modulate paracrine/autocrine actions of the IGFs in discrete brain regions or compartmentalization of the IGFs within the brain.     

Role of insulin-like growth factor binding proteins in the control of IGF actions

The insulin-like growth factor binding proteins have been shown to modify IGF actions. IGFBP-5 binds to extracellular matrix (ECM) and its ability to potentiate IGF activity is dependent upon the amount that is ECM associated. To determine the specific regions of IGFBP-5 that are required for ECM association, site directed mutagenesis has been used to prepare several forms of IGFBP-5. Mutants that have had the amino acids between positions 201 and 218 altered have been useful. Mutation of the lysine 211 resulted in no change in the affinity of IGFBP-5 for ECM or heparin Sepharose; however, it resulted in a major reduction in affinity for IGF-I following heparin binding. Other mutations which disrupted heparin binding also resulted in loss of this affinity shift. Most distruptive were mutations of amino acids 211, 214, 217 and 218 and 202, 206 and 207. Mutation of residues 201 plus 202 had some effect, but substitution for 207, 211, 217 and 218 had no effect. When binding to intact ECM was analyzed, similar results were obtained. This suggests that amino acids 202, 206 and 214 are definitely involved in heparin and ECM binding. When binding to proteoglycans such as tenascin and heparin sulfate proteoglycan was analyzed, similar results were obtained. IGFBP-5 also binds to other proteins in ECM, including type IV collagen and plasminogen activator inhibitor-I. Specific antisera for plasminogen activator inhibitor-1 can coprecipitate IGFBP-5. IGFBPs are degraded by specific proteases. Three proteases that degrade IGFBP-2, -4 and -5 have been characterized. They are serine proteases that cleave these proteins at basic residues. Although several well characterized serine proteases cleave IGFBP-4 or -5, the proteases in cell conditioned media appear to be distinct.     

Inhibition of insulin receptor activation by insulin-like growth factor binding proteins

The insulin-like growth factors (IGFs) are transported by a family of high-affinity binding proteins (IGFBPs) that protect IGFs from degradation, limit their binding to IGF receptors, and modulate IGF actions. The six classical IGFBPs have been believed to have no affinity for insulin. We now demonstrate that IGFBP-7/mac25, a newly identified member of the IGFBP superfamily that binds IGFs specifically with low affinity is a high-affinity insulin binding protein. IGFBP-7 blocks insulin binding to the insulin receptor and thereby inhibiting the earliest steps in insulin action, such as autophosphorylation of the insulin receptor beta subunit and phosphorylation of IRS-1, indicating that IGFBP-7 is a functional insulin-binding protein. The affinity of other IGFBPs for insulin can be enhanced by modifications that disrupt disulfide bonds or remove the conserved COOH terminus. Like IGFBP-7, an NH2-terminal fragment of IGFBP-3 (IGFBP-3((1-87))), also binds insulin with high affinity and blocks insulin action. IGFBPs with enhanced affinity for insulin might contribute to the insulin resistance of pregnancy, type II diabetes mellitus, and other pathological conditions.     

In vivo hormonal regulation of insulin-like growth factor binding protein-5 mRNA in the immature rat ovary

OBJECTIVE: Despite the potential importance of insulin-like growth factor binding protein-5 (IGFBP-5) to follicular development, the hormonal regulation of this antigonadotropic IGFBP has not been investigated. Therefore, it was the objective of this study to eludicate the role of gonadotropins and estrogen in the in vivo regulation of IGFBP-5 mRNA expression. METHODS: Two models of follicular development in immature rats were used. Specifically, rats were hypophysectomized and treated with FSH and/or diethylstilbestrol (DES). Alternatively, terminal follicular development was induced in intact immature rats by pregnant mare serum gonadotropin (PMSG) and hCG. The IGFBP-5 mRNA in whole ovarian RNA was assayed by Northern blot hybridization. Localization of expression in PMSG and hCG-stimulated ovaries was further assessed by in situ hybridization. RESULTS: Expression of IGFBP-5 mRNA was increased in ovaries from hypophysectomized rats. Treatment with FSH and/or DES did not alter the abundance of this mRNA. Treatment with PMSG induced a transient increase in IGFBP-5 expression that was localized in a subset of alpha-inhibin-negative follicles. At later times after PMSG, IGFBP-5 expression persisted in the surface epithelium but was not detected in large preovulatory follicles. In vitro studies affirmed the antigonadotropic action of IGFBP-5. CONCLUSION: In vivo expression of IGFBP-5 in the rat ovary is moderated by hormonal treatment both in terms of total expression and follicular localization.     

Pregnancy-dependent growth of mammary tumors is associated with overexpression of insulin-like growth factor II

We demonstrate that although IGF-II gene expression is approximately 3-fold higher in 9,10-dimethyl-1,2-benzanthracine (DMBA)-induced rat mammary tumors (MTs) than in nonneoplastic breast tissue, IGF-II mRNA abundance in DMBA-induced MTs is approximately 130-fold higher in pregnant as compared to nonpregnant hosts. This correlated with accelerated tumor growth in pregnant hosts. Immunohistochemical studies of DMBA-induced MTs with an anti-IGF-II antibody showed an intense staining of tumor cells for IGF-II, whereas a very low staining signal was observed for normal epithelial cells in the lobules. A similar immunostaining pattern was observed in three of three human ductal cancers and adjacent normal breast tissue obtained during pregnancy. DMBA-induced MTs expressed high levels of type I receptor for IGFs as determined by Northern blots. In vitro studies confirmed that IGF-II is a mitogen for neoplastic epithelial cells derived from DMBA-induced MTs. These results demonstrate that hormonal changes associated with pregnancy accelerate breast cancer cell proliferation in the DMBA-induced MT model and suggest that this acceleration is mediated by up-regulation of IGF-II expression within neoplasms.     

Differential spatio-temporal expression of the insulin-like growth factor genes in regenerating sciatic nerve

BACKGROUND: Prognosis following locoregional recurrence of breast cancer after mastectomy often is described as fatal. However, certain subgroups with better prognosis are supposed. We analysed established prognostic factors for their influence on post recurrence survival in order to discriminate favourable from unfavourable subgroups. PATIENTS AND METHODS: Between 1979 and 1989 163 patients with a local or regional recurrence of breast cancer following mastectomy were treated at the Department of Radiation Oncology of the University of Würzburg. One hundred and forty had an isolated recurrence, without evidence of distant disease at the time of recurrence. Median follow up for patients alive at the time of analysis was 102 months from diagnosis of recurrence. Thirteen prognostic factors were tested. RESULTS: Out of the 140 patients 94 (58%) developed distant metastases within the follow-up period. Metastatic-free rate was 42% at 5 years and 38% at 10 years following recurrence. Recurrences occurred in 50% of patients within the first 2 years from primary surgery, in 83% within 5 years. In univariate analysis statistically significant influence on survival rates was found for pT, pN-status, lymphatic vessel invasion, blood vessel invasion, tumor necrosis, hormonal receptor status, presence or development of distant metastases, time to recurrence and site and extension of recurrence. Two- and 5-year survival rates ranged from 64% to 81% and from 40% to 60%, respectively in the favourable subgroups compared to a survival rate ranging from 15% to 44% at 2 years and 0% to 29% at 5 years in the unfavourable subgroups. In patients with involved axillary lymph nodes, the absolute number of nodes did not prove to have significant influence on overall survival. Histopathological grading did not reach statistical significance levels although an influence on survival was observed. Preceding adjuvant radiotherapy did not influence post-recurrence survival rates. Also preceding adjuvant systemic therapy showed no significant impact on survival. Multivariate analysis demonstrated that primary axillary status correlated most strongly with overall survival (p < 0.001) followed by tumor necrosis (p < 0.01). CONCLUSIONS: The mentioned prognostic factors may be useful in determining the adequate (local and systemic) therapy and the best time for it. Our data support previous findings, that certain subgroups with favourable prognostic features exist and they might still have a chance for cure by an adequate local treatment, whereas subgroups of patients with unfavourable prognostic factors have to receive systemic therapy immediately following local therapy because of the forthcoming systemic progression.     

Competitive binding assay for determination of rat insulin-like growth factor binding protein-3

We describe a novel competitive assay for rat insulin-like growth factor (IGF)-binding protein-3 (rIGFBP-3) based on the ability of IGFBP-3 to form a ternary complex with the acid labile subunit (ALS) in the presence of IGF. Human (h)ALS was bound to test tubes pre-coated with anti-human ALS antibody. The assay depends on competition between a covalent complex of 125I-hIGF-I and hIGFBP-3, added as tracer, and hIGFBP-3 or rIGFBP-3 in standards and test samples, for binding to the immobilized hALS. Purified natural hIGFBP-3 served as standard. Human IGFBP-3 and rIGFBP-3 were able to compete for tracer binding in the presence, but not in the absence, of IGF-I. Before assay, rat serum samples were acidified to denature endogenous ALS. Standards ranged from 0.10 (lower detection limit) to 20 ng/tube. Rat serum, semipurified rIGFBP-3, human serum and purified hIGFBP-3 diluted in parallel. The level of rIGFBP-3 was 1.63+/-0.28 mg/l (mean+/-SEM) in young rats and increased to 3.41+/-0.26 mg/l (p < 0.05) in old rats (n = 5-6). Fasting for 3 days reduced rIGFBP-3 from 2.41+/-0.27 to 1.33+/-0.14 mg/l (p < 0.05). Levels of rIGFBP-3 were reduced in hypophysectomized (0.16+/-0.04 mg/l; p < 0.05) and diabetic rats (1.04+/-0.30 mg/l; p < 0.05), and normal in insulin-treated diabetic rats (2.49+/-0.04 mg/l; ns), when compared to controls (2.79+/-0.22 mg/l). Changes in levels of IGFBP-3 parallelled those of immunoreactive rALS. We conclude that this assay provides a novel method of quantitating functional IGFBP-3 in rat serum.     

Free and total insulin-like growth factors and insulin-like growth factor binding proteins during 14 days of growth hormone administration in healthy adults

The objective was to investigate the effect of growth hormone (GH) administration on circulating levels of free insulin-like growth factors (IGFs) in healthy adults. Eight healthy male subjects were given placebo and two doses of GH (3 and 6 IU/m2 per day) for 14 days in a double-blind crossover study. Fasting blood samples were obtained every second day. Free IGF-I and IGF-II were determined by ultrafiltration of serum. Total IGF-I and IGF-II were measured after acid-ethanol extraction. In addition, GH, insulin, IGF binding protein 1 (IGFBP-1) and IGFBP-3 were measured. Serum-free and total IGF-I increased in a dose-dependent manner during the 14 days of GH administration. After 14 days, serum-free IGF-I values were 610 +/- 100 ng/l (mean +/- SEM) (placebo), 2760 +/- 190 ng/l (3 IU/ m2) and 3720 +/- 240 ng/l (6 IU/m2) (p = 0.0001 for 3 and 6 IU/m2 vs placebo; p = 0.004 for 3 IU/m2 vs 6 IU/m2). Total IGF-I values were 190 +/- 10 micrograms/l (placebo), 525 +/- 10 (3 IU/m2), and 655 +/- 40 micrograms/l (6 IU/m2) (p < 0.0001 for 3 and 6 IU/m2 vs placebo; p = 0.04 for 3 IU/m2). There were no differences in the levels of free or total IGF-II during the three study periods. Insulin-like growth factor binding protein 1 was decreased during GH administration (p = 0.04 for placebo vs 3 IU/m2; p = 0.006 for placebo vs 6 IU/m2). In conclusion, fasting serum free IGF-I increased dose dependently during GH administration and free IGF-I increased relatively more than total IGF-I. This may partly be due to the decrease in IGFBP-1.     

Apoptosis in the pancreatic islet cells of the neonatal rat is associated with a reduced expression of insulin-like growth factor II that may act as a survival factor

Islet cell ontogeny will define adult beta-cell mass and will consist of a balance of islet cell birth and death. We have investigated the ontogeny of factors that may be related to developmental apoptosis in the islets, insulin-like growth factor II (IGF-II) and inducible nitric oxide synthase (iNOS), in pancreata of young Wistar rats. Pancreata were collected from rats of 21 days gestation to 29 days postnatal age. In situ hybridization and immunohistochemistry showed that IGF-II was expressed and present in fetal and neonatal islet cells, but declined rapidly 2 weeks after birth. Little IGF-I was associated with fetal or postnatal islets. Apoptosis in islet cells was visualized by molecular histochemistry for DNA breakage in tissue sections. Apoptosis was low in the fetus, but increased in incidence postnatally so that 13% of islet cells were undergoing apoptosis on postnatal day 14, with the incidence declining thereafter. Immunohistochemistry for iNOS showed that it was expressed within beta-cells and was most abundant 12 days after birth. When islets were isolated from rat pancreata 20-22 days after birth, islet cell viability, DNA synthetic rate, and insulin release were reduced after incubation with interleukin-1beta, tumor necrosis factor, or interferon-gamma. An increased rate of islet cell survival was found after simultaneous incubation with IGF-I or -II. Cytokine-mediated islet cell death involved the induction of apoptosis. Islets isolated from neonatal rats were not killed after exposure to these cytokines at the same concentrations, but cytokine-induced cell death was seen when neonatal islets were incubated with a neutralizing antibody against IGF-II. These experiments show that a peak of islet cell apoptosis that is maximal in the rat pancreas 14 days after birth is temporally associated with a fall in the islet cell expression of IGF-II. IGF-II was shown to function as an islet survival factor in vitro. The induction of islet cell apoptosis in vivo may involve an increased expression of iNOS within beta-cells.     

Islet cell proliferation and apoptosis in insulin-like growth factor binding protein-1 in transgenic mice

Transgenic mice which overexpress insulin-like growth factor binding protein-1 (IGFPB-1) demonstrate fasting hyperglycemia, hyperinsulinemia and glucose intolerance in adult life. Here we have examined the ontogeny of pancreatic endocrine dysfunction and investigated islet cell proliferation and apoptosis in this mouse model. In addition we have examined pancreatic insulin content in transgenic mice derived from blastocyst transfer into non-transgenic mice. Transgenic mice were normoglycemic at birth but had markedly elevated plasma insulin levels, 56.2 +/- 4.5 versus 25.4 +/- 1.5 pmol/l, p < 0.001, and pancreatic insulin concentration, 60.5 +/- 2.5 versus 49.0 +/- 2.6 ng/mg of tissue, P < 0.01, compared with wild-type mice. Transgenic mice derived from blastocyst transfer to wild-type foster mothers had an elevated pancreatic insulin content similar to that seen in pups from transgenic mice. There was an age-related decline in pancreatic insulin content and plasma insulin levels and an increase in fasting blood glucose concentrations, such that adult transgenic mice had significantly less pancreatic insulin than wild-type mice. Pancreatic islet number and the size of mature islets were increased in transgenic animals at birth compared with wild-type mice. Both islet cell proliferation, measured by 5-bromo-2'-deoxyuridine labeling, and apoptosis, assessed by the in situ terminal deoxynucleotidyl transferase and nick translation assay, were increased in islets of newborn transgenic mice compared with wild-type mice. In adult mice both islet cell proliferation and apoptosis were low and similar in transgenic and wild-type mice. Islets remained significantly larger and more numerous in adult transgenic mice despite a reduction in pancreatic insulin content. These data suggest that overexpression of IGFBP-1, either directly or indirectly via local or systemic mechanisms, has a positive trophic effect on islet development.     

Effect of protein restriction on the messenger RNA contents of bone-matrix proteins, insulin-like growth factors and insulin-like growth factor binding proteins in femur of ovariectomized rats

It has been reported that loss of ovarian oestrogen after menopause or by ovariectomy causes osteoporosis. In order to elucidate the effect of dietary protein restriction on bone metabolism after ovariectomy, we fed ovariectomized young female rats on a casein-based diet (50 g/kg diet (protein restriction) or 200 g/kg diet (control)) for 3 weeks and measured mRNA contents of bone-matrix proteins such as osteocalcin, osteopontin and alpha 1 type I collagen, insulin-like growth factors (IGF) and IGF-binding proteins (IGFBP) in femur. Ovariectomy decreased the weight of fat-free dry bone and increased urinary excretion of pyridinium cross-links significantly, although dietary protein restriction did not affect them. Neither ovariectomy nor protein restriction affected the content of mRNA of osteopontin and osteocalcin; however, ovariectomy increased and protein restriction extensively decreased the alpha 1 type I collagen mRNA content in bone tissues. Ovariectomy increased IGF-I mRNA only in the rats fed on the control diet. Conversely, protein restriction increased and ovariectomy decreased the IGF-II mRNA content in femur. Furthermore, the contents of IGFBP-2, IGFBP-4 and IGFBP-5 mRNA increased, but the content of IGFBP-3 mRNA decreased in femur of the rats fed on the protein-restricted diet. In particular, ovariectomy decreased the IGFBP-2 mRNA content in the protein-restricted rats and the IGFBP-6 mRNA content in the rats fed on the control diet. These results clearly show that the mRNA for some of the proteins which have been shown to be involved in bone formation are regulated by both quantity of dietary proteins and ovarian hormones.