Part of this work was presented at the International Conference on Bone and Mineral Metabolism in Vienna, Austria in April, 1998 and the ASBMR-IBMS Second Joint Meeting in San Francisco, California, U.S.A. in December, 1998.
Thyroid hormone (T3) and insulin-like growth factor I (IGF-I) are critical regulators of skeletal function. T3 increases IGF-I production in bone. To assess the potential role of IGF-I as a mediator of T3 actions, we characterized phenotypic markers of osteoblast activity in two osteoblast models, normal mouse osteoblasts and MC3T3-E1 cells, exposed to T3 alone or under conditions that interfere with IGF-I actions. T3 significantly increased osteoblast 3H-proline incorporation, alkaline phosphatase (ALP), and osteocalcin. Both αIR3, a neutralizing monoclonal antibody to the IGF-I receptor, and JB1, an IGF-I analogue antagonist, attenuated the stimulatory effects of T3. T3 effects also were decreased in cells transfected with antisense oligonucleotide (AS-ODN) to the IGF-I receptor gene. Both IGF-I and T3 had mitogenic effects that were inhibited by the antagonists. IGF-I by itself did not stimulate 3H-proline incorporation, ALP, and osteocalcin in the models used, revealing that although IGF-I is essential for the anabolic effects of T3, it acts in concert with other factors to elicit these phenotypic responses. (J Bone Miner Res 2000;15:188–197)
Thyroid hormones (T3/T4) are critical regulators of skeletal development and maturation. Evidence demonstrating effects of T3 to promote both the formation and breakdown of bone can be found in clinical observations as well as in vivo investigations in animals.(1,2) Long-standing juvenile hypothyroidism can cause severe growth retardation when thyroid hormones are not replaced.(3) T4 treatment in children with congenital hypothyroidism resulted in a positive correlation between the bone age and the dose of T4 administered.(4) Although in children and in young animals excess T3 causes enhanced bone growth,(5,6) it leads to bone loss in adults.(7,8) It is well established that both bone formation and resorption are markedly increased in hyperthyroidism.(9) Greater increases in the resorption markers than the formation markers suggest an imbalance between resorption and formation, leading to a net loss of cortical and trabecular bone volume.(10)
Despite the clinical importance of the regulatory actions of T3 on bone, little is known about the mechanisms by which this occurs. Recent studies have identified the presence of thyroid hormone receptors in osteoblast-like cells and immortalized osteosarcoma cells and have further demonstrated that these receptors are functional.(11,12) The expression of certain thyroid hormone receptor isoforms has been shown in osteoclasts.(13) The expression of many genes in a variety of tissues, including bone, is regulated by T3.(14) Specifically, production of several osteoblast phenotypic markers such as alkaline phosphatase (ALP), osteocalcin, and collagen are affected by T3 in a number of osteoblastic cells.(15,16) Furthermore, there is evidence suggesting that certain effects of T3 on bone remodeling are mediated through the production of local factors. Factors such as prostaglandins and interleukin-1 (IL-1) have been implicated as mediators of the resorptive effects of T3 on bone.(17,18)
As one of the most abundant growth factors present in bone, insulin-like growth factor I (IGF-I) represents a likely mediator of the anabolic actions of systemic hormones. It is found in the plasma and also is produced in many tissues, where it has both autocrine and paracrine action.(19,20) IGF-I stimulates bone production by increasing osteoblast proliferation and matrix synthesis.(21,22) T3 causes an increase in IGF-I production in UMR-106 osteoblastic cells and bone organ cultures.(23,24) IGF-I messenger RNA (mRNA) in MC3T3-E1 osteoblastic cells also is increased after T3 treatment.(25) The IGF-I gene is regulated by several different hormones. Growth hormone (GH) is the prototypic regulator of the IGF-I gene, especially in the liver. T3 has been shown to interact with GH and, together, these hormones regulate IGF-I production by the liver and other tissues.(26) Levels of GH and IGF-I are reduced in hypothyroidism, but treatment with GH alone failed to restore normal growth in hypothyroid children.(27) Although the biological significance is still unclear, estrogen can induce IGF-I mRNAin osteoblast cell lines.(28) In addition, parathyroid hormone (PTH) also increases IGF-I mRNA in bone cell cultures.(29) IGF-I appears to mediate specific anabolic effects of PTH on bone; interference with IGF-I blocks the anabolic actions of PTH.(22)
The purpose of this study was to determine whether IGF-I activity is important for the skeletal effects of T3 on bone formation. Because both IGF-I and T3 have anabolic actions on bone and T3 also increases IGF-I production in bone, it was important to determine whether IGF-I could be a mediator of T3 actions on bone. We used several different complementary approaches to interfere specifically with the IGF-I actions in osteoblasts and to determine whether these interventions would prevent the anabolic effects of T3. Osteoblastic functions assessed included total protein synthesis based on proline incorporation and the production of the phenotypic markers of osteoblastic activity, osteocalcin, and ALP. We also measured the mitogenic activity of the osteoblasts. Normal mouse osteoblasts and/or MC3T3-E1 cells, a widely used clonal osteoblast cell line, were used for the studies.
Our data show that interference with IGF-I actions attenuates T3-stimulated osteoblastic functions and thus support a model in which IGF-I represents a critical factor for the anabolic effects of T3.
Primary osteoblast cultures were prepared from calvarial bones obtained from 6–7-day-old neonatal mice.(30) Mice were housed and killed in accordance with policy at the Northwestern University Animal Care and Use Facility. After the calvarial bones were dissected, cells were released from the bone by six sequential 20-minute collagenase digestions. A pool from digestions 2–6 was collected and seeded into 75-cm2 flasks for culture in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 5% fetal bovine serum (FBS) and 100 μg/ml penicillin and streptomycin. This primary culture was allowed to grow to confluence and was trypsinized for subculture onto appropriate experiment dishes (either 24- or 96- well multi-well plates). This subculture (technically passage 2) expressed a consistent osteoblastic phenotype based on ALP, osteocalcin (OCN), and 3H-proline incorporation, and it was used to study osteoblast responses to agonist stimulation.
Mouse osteoblastic MC3T3-E1 cells were cultured in a modified essential medium (α-MEM) supplemented with 5% FBS and 50 μg/ml gentamycin. Cells were passaged every 7 days by harvesting with 0.1% trypsin-EDTA and reseeding in 75-cm2 flasks. For experiments, MC3T3-E1 cells between passages 3 and 14 were seeded into 60-mm dishes or 24- or 96-well cluster dishes. Assessment of mature osteoblastic phenotypic markers such as ALP and OCN in MC3T3-E1 cell was performed in cells after 18 days of culture time. The long culture time was required for MC3T3-E1 cells to develop into more mature osteoblasts with the ability to express detectable levels of ALP.(31) The cells were seeded and grown in normal serum media for 3 days and the cells then were cultured in T3-free medium for 18-days before treatment with agonists. Medium was changed every 3 days.
For experiments assessing mitogenic activity, once the cells reached 60–80% confluence after seeding, cells were made quiescent by incubation for 18 h in medium supplemented with 0.1% bovine serum albumin (BSA) and antibiotic (hereafter designated as serum-free medium). For experiments involving effects of hormonal manipulation on osteoblast functions, cells were cultured in hormone-depleted medium (designated as T3-free medium) after the cells reached confluence in normal medium. Treatment with Dowex AF-1-X-10 resin (Sigma, St. Louis, MO, USA) was used to remove T3 and T4 hormones from serum with reasonable selectivity. This method, originally reported by Samuels et al., reduced T3 from 150 ng/dl to <2.0 ng/dl and T4 from 6.8 μg/dl to 0.08 μg/dl in calf serum.(32) T3-free medium contained 5% resin-treated FBS, 50 μg/ml ascorbic acid, and appropriate antibiotics.
Osteoblasts were plated in 24-well cluster dishes with normal medium at 20,000 cells/well. They were made quiescent as described above and then cultured in serum-free medium with either T3 or IGF-I for a predetermined treatment period. The concentrations of T3 or IGF-I used were the lowest concentrations that caused a significant response and they were determined based on preliminary studies. [3H]Thymidine (0.5 μCi/ml) was added to the cells during the last hour of the treatment in the continued presence or absence of the agonists. At the termination of treatment, [3H]thymidine-containing medium was removed. The cells were washed once with ice-cold phosphate-buffered saline (PBS), once with 10% trichloroacetic acid (TCA), incubated on ice for 15 minutes with 10% TCA, and after one wash with ice-cold 95% ethanol, radioactivity was extracted with 1N NaOH for assay by liquid scintillation.
Osteoblasts were subcultured onto 24-well cluster dishes with normal medium at 35,000 cells/well. After the culture reached confluence, cells were hormone-depleted by culturing in T3-free medium for 48 h before treatment. Cells were treated for 72 h unless otherwise indicated. For measurement of proline incorporation, cells were incubated for the last 2 h of the treatment period with [3H]proline.(1 μCi/ml) Samples were prepared similarly to the procedures described for [3H]thymidine incorporation. Incorporation of the label into the TCA-precipitable fraction was measured by liquid scintillation counting.
Osteoblasts were seeded onto 96-well cluster dishes at 16,000 cells/well in normal medium. Confluent cells then were cultured in T3-free medium before agonist stimulation of predetermined length. Media from the final 72 h of treatment were collected and stored at −20°C until assay for OCN. OCN secreted by the osteoblasts was measured by a commercial radioimmunoassay (RIA; Biomedical Technologies, Stoughton, MA, USA). This assay uses goat anti-mouse OCN and mouse 125I-OCN tracer. This assay is specific for mouse OCN and recognizes total OCN. The limit of detection of the assay is 1.56 ng/ml.
Osteoblasts were seeded in 96-well culture cluster dishes at 16,000 cells/well in normal medium. Confluent monolayer cells then were cultured in T3-free medium before agonist stimulation. Cells were treated with agonists for 72 h before ALP measurement. ALP in cell lysates was measured by the production of p-nitrophenol from p-nitrophenyl phosphate using a protocol modified from Schlossman et al.(33) Briefly, medium was removed from the treated cells and the monolayers were washed with cold PBS. After the PBS was aspirated, 100 μl diethanolamine (50 mM, pH 10.5) was added to disrupt the monolayer and 50 μl of 2.5 nM p-nitrophenyl phosphate in glycine buffer (100 nM glycine, 2 mM MgCl2) was added as substrate. The entire complex was incubated at 37°C for 30 minutes, and the reaction was stopped by the addition of 0.1N NaOH. Absorbance was read on a Dynatech (Chantilly, VA, USA) MR5000 spectrophotometer at 410 nm. Activity was calculated in reference to a standard curve of p-nitrophenol. To assess the effect of IGF-I receptor down-regulation on either T3- or IGF-I–stimulated ALP activity, MC3T3-E1 cells precultured in T3-free medium for 18 days were transfected with 1.66 ng/well of either antisense oligonucleotide (AS-ODN) or mismatch oligonucleotide (MS-ODN) for 4 h using the protocol described below. After the transfection period, cells were returned to T3-free medium and subsequently exposed to either IGF-I or T3 for 72 h. ALP activity was measured at the end of the treatment period.
Antagonists to IGF-I receptor
To determine the effects of IGF-I receptor antagonists on IGF-I- or T3-induced cellular functions, experiments were performed in which the normal mouse osteoblasts or MC3T3-E1 cells were exposed to the agonists alone or in the presence of antagonists. αIR3 (Calbiochem, San Diego, CA, USA), a monoclonal antibody specific against the IGF-I receptor and JB1 (Peninsula Laboratories, Inc., San Carlos, CA, USA), a peptide analogue of IGF-I, were used as antagonists. Both agents have specificity for the IGF-I receptor and neutralizing properties that block the IGF-I signaling pathway.(34,35) A mouse myeloma nonspecific immunoglobulin G (IgG) antibody also was used to confirm specificity in the osteoblast model.
Antisense IGF-I receptor oligonucleotides
The AS-ODN, 5′-CCC TCC TCC GGA GCC-3′, is complementary to a sequence around the AUG-site of the mouse IGF-I receptor mRNA sequence; MS-ODN is a scrambled version of AS-ODN, 5′-CCC GGA TCC TCC GCC-3′. Dr. Sergei Gryaznov (personal communication, 1997) kindly provided these sequences. To assess the efficacy of the AS-ODN, confluent MC3T3-E1 osteoblasts seeded in 60-mm dishes were transfected with either AS-ODN or MS-ODN (3.22 μg DNA/dish) using Lipofectamine-PLUS reagent following the manufacturer's protocol (Gibco, Grand Island, NY, USA). The cells were transfected in serum-free medium for 4 h and the medium then was replaced with normal growth medium. Levels of IGF-I receptor in transfected cells were assessed 48–72 h posttransfection.
IGF-I receptor expression was determined by Western immunoblotting using methods described previously.(30) Briefly, an aliquot of the whole cell lysates of MC3T3-E1 cells was used for determination of protein using the method of Lowry.(36) Fifteen micrograms of protein was applied per lane, and electrophoresis was performed under denaturing conditions on a 10% polyacrylamide-sodium dodecyl sulfate (SDS) gel. After overnight wet transfer to nitrocellulose membranes (Schleicher & Schuell, Keene, NH, USA), the blot was probed with a rabbit IGF-I receptor β-subunit antibody (Santa Cruz, Santa Cruz, CA, USA) at a 1:500 dilution and followed by subsequent incubation with a goat anti-rabbit, peroxidase-conjugated secondary antibody (Sigma, St. Louis, MO, USA) at a 1:2000 dilution. The blot was visualized by enhanced chemiluminescence (ECL, Amersham, Piscataway, NJ, USA) using Kodak X-OMAT LS film. For normalization purposes, the same blot also was probed with a rabbit β-actin antibody (Sigma, St. Louis, MO, USA). Immune complexes were quantified by densitometry using a Bio-Rad imaging densitometer (BioRad, Herculus, CA, USA).
Data are expressed as mean ± SEM. Statistical analysis was carried out by one-way analysis of variance. Significance of differences among means was determined by post hoc testing, using Tukey's method.
Effects of blockade of IGF-I receptors on mitogenic activity of T3 and IGF-I in osteoblasts
Three nM T3 increased [3H]thymidine incorporation in quiesced normal mouse osteoblasts (Fig. 1A). This result also was observed with MC3T3-E1 cells treated with 10 nM T3 (Fig. 2A). As shown, 12-h T3 treatment increased [3H]thymidine incorporation in normal mouse osteoblasts and MC3T3-E1 cells by 33% and 19%, respectively. Treatment with IGF-I significantly increased [3H]thymidine incorporation in both cell models. A 59% increase in [3H]thymidine incorporation was seen in normal osteoblasts treated with 1 nM IGF-I for 12 h (Fig. 1B). MC3T3-E1 cells treated with 10 nM IGF-I for 12 h showed an 82% increase compared with control cells (Fig. 2B). The concentrations of T3 and IGF-I used were the lowest levels that showed a significant effect during preliminary studies.
When normal mouse osteoblasts or MC3T3-E1 cells were coincubated with T3 and a neutralizing antibody to the IGF-I receptor (αIR3), the stimulatory effect of T3 on [3H]thymidine incorporation was blocked (Figs. 1A and 2A). Treatment with T3 in the presence of 1.5 μg/ml of αIR3 attenuated the stimulatory effect of T3 to the level of nonstimulated control cells. The same concentration of αIR3 also blocked the increased [3H]thymidine incorporation stimulated by IGF-I in both cell models (Figs. 1B and 2B). Treatment with αIR3 alone had no effect on basal proliferative responses. A nonspecific, control IgG antibody failed to inhibit either IGF-I or T3-stimulated responses in MC3T3-E1 cells or normal mouse osteoblasts.
Effects of IGF-I receptor antagonists on T3-stimulated osteoblastic functions
[3H]proline incorporation: The effects of 72h T3 treatment on 3H-proline incorporation in confluent normal mouse osteoblasts precultured in T3-free medium are illustrated in Fig. 3A. Ten-nM T3 caused a 41% increase in osteoblast [3H]proline incorporation. In this model and under the conditions used, 10 nM IGF-I elicited no stimulation. There was no synergism in the observed responses when the cells were treated with both T3 and IGF-I (data not shown). When cells were treated with 10 nM T3 and in the presence of either αIR3 or peptide analogue of IGF-I, JB1, the T3-stimulated [3H]proline incorporation was attenuated to the level of control. Although both αIR3 and JB1 appear to have slight stimulatory effects on basal [3H]proline incorporation, these changes were not statistically significant.
The 10 nM T3 also stimulated [3H]proline incorporation in confluent MC3T3-E1 cells grown in T3-free medium. As shown in Fig. 3B, 10 nM T3 caused a 43% increase in [3H]proline incorporation. As in normal mouse osteoblasts, no significant stimulation was observed with 10 nM IGF-I treatment. The stimulatory effect of T3 on [3H]proline incorporation was attenuated by αIR3. The αIR3 alone did not change 3H-proline incorporation when compared with control.
OCN: For these studies, normal mouse osteoblasts, grown to confluence and precultured in T3-free medium for 48 h, were incubated with 1 nM T3 for 72 h. Higher concentrations of T3 did not elicit greater stimulation and we chose the lowest concentration that elicited a significant response for the antagonist studies. The culture media were collected and assayed for OCN production using an RIA. In media from control cells, OCN was 2.3 ± 2.6 ng/ml (mean ± SEM, n = 6). As shown in Fig. 4A, 1 nM T3 induced an approximately 50-fold increase in OCN production by the osteoblasts. Similar to the findings on 3H-proline incorporation, 10 nM IGF-I did not stimulate detectable OCN production under the conditions used. To determine whether IGF-I was required for the T3-stimulated OCN production, the cells were treated with T3 in the presence of the IGF-I receptor antibody, αIR3. The αIR3 (0.75–3.0 μg/ml) significantly decreased T3-stimulated OCN production. The αIR3 alone had no effects on OCN production.
T3 also stimulated OCN production in MC3T3-E1 cells. As shown in Fig. 4B, the maximal effect was obtained with 10 nM T3. OCN from control MC3T3-E1 cells was 0.72 ± 1.5 ng/ml (extrapolated value, below the detection limit of the assay) and 10 nMT3 stimulated OCN production to 11 ± 5 ng/ml. Although OCN stimulation in MC3T3-E1 cells treated with T3 was statistically significant compared with that in the control cells, the OCN concentration in MC3T3-E1 cells was significantly less than levels observed in normal osteoblasts. For these reasons, we focused our efforts on the normal osteoblasts and did not study the effects of IGF-I antagonists on T3-stimulated OCN production in the MC3T3-E1 cells. As was found in normal mouse osteoblasts, IGF-I did not increase OCN production in the MC3T3-E1 cells.
ALP: We attempted to study the effects of T3 on ALP activity normal mouse osteoblasts. As shown in Fig. 5A, the expression of ALP in these cells did not show any stimulation after 72-h treatment with T3 or IGF-I. The baseline ALP level in confluent mouse osteoblasts was four times the baseline ALP level in confluent MC3T3-E1 cells.
A stimulatory effect of T3 on ALP in MC3T3-E1 cells was evident when confluent MC3T3-E1 cells grown for 18 days were treated with 10 nM T3 for 72 h (total culture time of 21 days). T3-treated cells showed a 34% increase in ALP activity compared with nontreated cells (Fig. 5B). The 10 nM IGF-I treatment did not change ALP levels significantly. Cotreatment with 10 nM IGF and 10 nM T3 did not produce stimulation significantly different from the response to 10 nM T3 alone (data not shown).
The T3-stimulated ALP activity was attenuated in MC3T3-E1 transfected with an AS-ODN that is complementary to mouse IGF-I receptor mRNA (Fig. 5B). In these cells, T3 failed to elicit a significant increase in ALP. Cells transfected with 1.66 μg/ml of the MS-ODNs did not cause an attenuated response to T3. Western immunoblotting for IGF-I receptor in MC3T3-E1 cells confirmed the specificity of AS-ODN for the IGF-I receptor. As shown in Fig. 6, MC3T3-E1 cells transfected with 1.66 μg/ml AS-ODN showed a 40% decrease in IGF-I receptor expression. Cells transfected with same amount of MS-ODN did not show decreased receptor expression. Transfection of the MC3T3-E1 cells with either AS-ODN or MS-ODN had no effect on the levels of actin protein expression.
The importance of T3 to maintain normal skeletal physiology is well accepted. However, the mechanism of T3 effects on bone remodeling remains to be elucidated. We focused our project specifically on studying the anabolic actions of T3 on bone. These anabolic responses included osteoblast proliferation and osteoblast phenotypic markers such as ALP activity, OCN production, and collagen synthesis. An improved understanding of this basic anabolic process elicited by T3 could be beneficial in the management of skeletal problems in patients with altered thyroid status caused by either disease processes or of iatrogenic origins and is thus of clinical importance.
Results from previous studies that showed T3 stimulates IGF-I production in bone models suggest the possibility that IGF-I could be responsible for some of the anabolic effects of T3 in bone tissues. Both T3 and IGF-I have skeletal anabolic effects in vivo: hypothyroid patients, if untreated, often show a bone age delayed more than 2 standard deviations from their chronological age; and with subsequent thyroid replacement therapy, the bone age positively correlates with the concentration of serum T4.(4) A patient with a T3 receptor mutation (homozygous deletion of threonine-337 in TRβ gene) showed similar abnormal skeletal development.(37) Similarly, the importance of IGF-I to skeletal systems is evident from the delayed bone development in IGF-I and IGF-I gene knock-out mice; mice with IGF-I receptor mutations show significantly delayed or arrested ossification in the developing skeleton.(38) These observations, taken together, further suggest the likelihood at IGF-I may be a candidate to mediate the stimulatory fects of T3 on bone formation or be essential for these imulatory effects. We therefore tried to determine whether ocking IGF-I action could influence T3 anabolic effects in steoblast cell lines.
Normal osteoblasts and MC3T3-E1 osteoblastic cells ere used as model systems for our studies because these ells have the capacity to express phenotypic markers of steoblast function and to synthesize IGF-I under basal and 3-stimulatedconditions.(2,39) We selected [3H]proline incororation, OCN production, and ALP activity because these arameters generally are accepted as osteoblast phenotypic arkers. We chose to measure 3H-proline incorporation into tal protein and not to distinguish between incorporation to collagenase-digestible protein and noncollagen protein ecause previous studies have shown that T3 stimulates H-proline incorporation into both collagen and noncollagen rotein to a similar extent in bone.(40) The choice of the cell type to use for a particular arameter was dictated by both cell characteristics and ethodological feasibility. Under our experimental condions, we were able to measure adequately the baseline H]thymidine and 3H-proline incorporation in both normal ouse osteoblasts and MC3T3-E1cells and these parameters ere responsive to T3 stimulation. Although OCN production could be detected in both cell models after T3 treatment, the response was much greater in normal osteoblasts. OCN levels in untreated cells were within the working range of the RIA in normal osteoblasts only; OCN levels in control of MC3T3-E1 cells were below the limit of detection of the assay. Therefore, the effect of the IGF-I antagonist on T3-stimulated OCN production was studied only in normal osteoblasts. Assessment of the effect of the IGF-I antagonist on ALP activity was limited to MC3T3-E1 cells because no stimulatory effects of T3 on ALP were observed in the normal osteoblasts. In previous studies, the effects of T3 on ALP have been variable, with some studies showing stimulatory effects and others reporting inhibitory or no response. For example, T3 failed to affect ALP activity in a study using UMR-106 cells.(41) In normal rat osteoblastic cells, T3 had dose-dependent biphasic effects on ALP.(42)
We used three different independent approaches to interfere specifically with the IGF-I actions in bone. Both the neutralizing antibody αIR3 and the IGF-I peptide analogue JB1 act as antagonists to the IGF-I receptor to prevent binding of IGF-I and the AS-ODN prevents the synthesis of IGF-I receptor. Both αIR3 and JB1 were shown to block specifically IGF-I-stimulated [3H]thymidine incorporation in various cell lines.(34,35) The inhibitory effects of αIR3 also were seen in our cell models where we observed attenuated [3H]thymidine incorporation in response to IGF-I stimulation in cells cotreated with the antagonists. The specificity of effects achieved by αIR3 was further confirmed by the inability of a control IgG antibody to attenuate IGF-I stimulation. A third approach was to inhibit IGF-I action by decreasing IGF-I receptor number using AS-ODNs specific for IGF-I receptor mRNA. Experiments using transfection of oligonucleotides were performed only in MC3T3-E1 cells because we were not successful in transfecting normal osteoblasts. The effectiveness of the AS-ODN to reduce IGF-I receptor protein was demonstrated using Western blotting for the IGF-I receptor. Using multiple approaches to interfere with IGF-I actions made it less likely that the observed responses were results of nonspecific side effects from any one of the antagonists.
Our data demonstrated mitogenic effects of T3. Other studies have shown that T3 has a dose-dependent, biphasic effect on osteoblast growth, with stimulation at low and inhibition at high concentrations.(43) T3 elicited significant stimulation of [3H]thymidine incorporation in both normal osteoblasts and MC3T3-E1 cells. IGF-I also stimulated [3H]thymidine incorporation in both models. Consistent with our proposed hypothesis, treatment with neutralizing antibody to IGF-I receptor blocked both IGF-I and T3-stimulated [3H]thymidine incorporation.
Our findings also showed anabolic effects of T3 on the osteoblast systems. T3 treatment caused a significant increase in phenotypic markers of osteoblasts. Normal osteoblasts treated with T3 for 72 h showed significant increases in both 3H-proline incorporation and OCN production. To determine whether endogenously produced IGF-I was necessary for T3-stimulated osteoblastic responses, we tested the effect of a neutralizing anti-IGF-I receptor antibody on T3-induced 3H-proline incorporation and OCN production. In addition, an IGF-I peptide antagonist JB1 was tested for its effect on T3-induced [3H]proline incorporation. Both the antibody and the peptide antagonist had significant attenuating effects on the T3-stimulated responses in normal osteoblasts, indicating a requirement for IGF-I in T3-induced total protein and OCN synthesis. Treatment with αIR3 antibody also attenuated T3-stimulated [3H]proline incorporation in the MC3T3-E1 cells.
Dependence on IGF-I also was observed for T3 effects on ALP activity in MC3T3-E1 cells. Treatment with 10 nM T3 greatly stimulated ALP expression; however, cells transfected with AS-ODN did not show increased ALP activity with the same T3 treatment. It is interesting that even a partial reduction in the expression of IGF-I receptor resulted in a significant attenuation in response. Similar results also were reported in a recent study by Muller et al. using ovarian cancer cells transfected with IGF-I receptor AS-ODN.(44) They showed that IGF-I receptor protein was reduced by the AS-ODN to 53% of control value and that transfected cells showed decreased proliferation in response to IGF-I stimulation. The control MS-ODN, which we demonstrated by Western immunoblotting to have no effect on IGF-I receptor expression, did not have an attenuating effect on T3-stimulated ALP activity. This result further confirmed that the IGF-I pathway is necessary for T3-stimulated responses.
Given our results showing that agents that interfere with IGF-I actions can block T3-stimulated mitogenic and anabolic responses in osteoblasts, it was expected that treatment with IGF-I alone would increase both mitogenesis and the osteoblast phenotypic markers. Unexpectedly, in both models, treatment with IGF-I alone caused only a stimulation in mitogenic activity and did not produce a stimulation in 3H-proline incorporation, OCN production, or ALP expression; cells that were cotreated with both IGF-I and T3 did not produce responses that were significantly greater than cells treated with T3 alone. These data were unexpected because previously published studies demonstrated IGF-I to have anabolic effects on bone. IGF-I stimulated both collagen and noncollagen protein production in cultures of fetal rat calvaria and human osteoblasts.(21,45–47)
A major and the most significant difference between our models and those in which IGF-I alone had effects is that others did not use resin-stripped serum, which we used in all of our anabolic studies. It is possible that other factors could have been present and were removed by the stripping procedure; these factors might complement IGF-I to stimulate anabolic actions in bone. This possibility is supported by the findings of Pirskanen et al., who showed that IGF-I alone stimulated mitogenesis in MG-63 cells but failed to stimulate OCN in the absence of other factors such as calcitriol.(48) T3 treatment might induce the synthesis of these other factors in its pleiotropic effects on bone.
It may be significant that other studies were performed in different osteoblast systems than those we used. These cell-based differences in response to an IGF stimulus may reflect receptor population and receptor cross-reactivity and may depend on cell type species and osteoblast lineage and maturation.(49,50) Recent studies have shown differential effects of thyroid hormones on several parameters, including IGF-I, at different skeletal sites.(51–53) Malpe et al. further demonstrated differential levels of IGF-I/II and IGF binding proteins (IGFBPs) 3, 4, and 5 at five different skeletal sites.(54) Because both of the cell models used in our study were calvarial derived, it would be of interest in future studies to determine whether the IGF-I dependence of T3 effects observed in the current study also are found in cells from other skeletal sites.
Interactions with various IGFBPs also can affect the observed responses. Both T3 and IGF-I can increase the production of IGFBPs.(24) IGFBPs can modulate IGF actions in bone. Certain IGFBPs inhibit IGF-I actions and others have been shown to have enhancing effects.(50,55,56) The relative amounts of the IGFBPs produced by T3 and IGF under different experimental conditions and in different models could play a role in the different responses found, because the final osteoblast response depends on presence or absence of endogenous IGFs and other IGFBPs.
In summary, different complementary methods of blocking IGF-I actions all attenuated T3-stimulated responses in osteoblasts. From our findings we conclude that IGF-I is a necessary component for T3-stimulated anabolic osteoblast functions and that the stimulation of IGF-I production by T3 is thus essential for the anabolic effects of the hormone. The observation that IGF-I alone did not elicit the same responses as T3 on 3H-proline incorporation, OCN production, or ALP activity (although it stimulated mitogenesis) suggests that IGF-I is a necessary but not a sufficient factor for T3 responses. Other T3-stimulated local factors that are essential for IGF-I action factors also may contribute to the anabolic effects of T3 on bone.
This study was supported by the U.S. Army Medical Research and Material Command grant DAMD17–96–1–6304 (P.H.S and L.D.M.) and a Knoll Thyroid Research Advisory Council grant (P.H.S). B.K.H. is a recipient of a Howard Hughes Medical Institute Research Fellowship for Medical Students. The authors thank Dr. Sergei Gryaznov and Dr. Yasphal S. Kanwar for providing technical advice on the antisense studies and Dr. Gryaznov for providing the oligonucleotide sequences used. They also thank Dr. William Lowe and Dr. Guo-guang Du for critical discussion of the manuscript.