THE ROLE OF THE mTOR NUTRIENT SIGNALING PATHWAY IN CHONDROCYTE DIFFERENTIATION

Abstract

The mammalian Target of Rapamycin (mTOR) is a nutrient-sensing protein kinase that regulates numerous cellular processes. Physilogical fetal rat metatarsal explant cultures were used, to study the effect of mTOR inhibition on chondrogenesis. Insulin significantly enhanced bone growth. Rapamycin significantly diminished the growth response to insulin, selectively on hypertrophic zone. Cell proliferation (BrdU incorporation) was unaffected by rapamycin. Similar observation was noted in In vivo injection of rapamycin to E19 fetal rats. In ATDC5 chondrogenic cell line, rapamycin inhibited proteoglycan accumulation and collagen X expression. Rapamycin decreased Indian Hedgehog (Ihh) accumulation, a regulator of chondrocyte differentiation. Furthermore, addition of Ihh to culture media was able to reverse of effect of rapamycin. We conclude that modulation of mTOR signaling contributes to chondrocyte differentiation, perhaps through its ability to regulate Ihh expression. Our findings are consistent with the hypothesis that nutrients, acting through mTOR, directly influence chondrocyte differentiation and long bone growth.

Introduction

Linear growth is largely a function of endochondral bone elongation, a process that occurs mainly in the cartilaginous growth plate (Karsenty 2003). Growth during childhood follows a predictable pattern; rapid growth during the first year of life, pre-pubertal slowing, pubertal acceleration and cessation of growth at the end of puberty (Spagnoli and Rosenfeld 1996). This pattern can be accounted for by the complex regulation of chondrocyte proliferation, differentiation and senescence. Disorders that disrupt chondrocyte proliferation and differentiation manifest as abnormal linear growth. Many chronic diseases, including those associated with malnutrition, are well-documented causes of impaired linear growth (De Luca 2006;Allen 1995). The suppression of growth is generally accepted as arising from down-regulation of insulin-like growth factor-I (IGF-I) acting as a circulating endocrine factor or, perhaps more importantly, acting locally in a paracrine manner (Thissen et al. 1994). Impaired IGF-I production in undernutrition is at least in part associated with impaired hepatic GH sensitivity, reflected by decreased hepatic GH receptor expression, hepatic IGF-I mRNA and circulating IGF-I levels (Noguchi 2000).

Recent years have seen progress in our understanding of the molecular mechanisms by which nutrients, including amino acids, can act as signaling factors (Kimball and Jefferson 2006a). Central to the control of cell growth and metabolism by amino acids is the conserved Ser/Thr signaling kinase referred to as the mammalian target of rapamycin (mTOR) (Martin and Hall 2005). The functions of mTOR include regulation of translation, transcription, ribosome biogenesis, nutrient transport and autophagy in response to nutrient availability. In addition, mTOR has a role in cell size determination (Wullschleger et al. 2006). At the signaling level, the nutrient-responsive mTOR pathway and the signaling pathways downstream from insulin and IGF-I are functionally interconnected (Hay and Sonenberg 2004).

Endochondral bone formation, starting from mesenchymal condensation, chondrocyte proliferation, chondrocyte hypertrophy and the formation of osteoblasts, is a result of an orchestrated series of gene expressions events and induction of numerous transcription factors (Kobayashi and Kronenberg 2005). The expression of Sox-9, Sox-5 and Sox-6 is induced in chondrogenic mesenchymal cells and persists in proliferating chondrocytes (Kobayashi and Kronenberg 2005;Lefebvre and Smits 2005). As chondrocytes proceed toward terminal differentiation, the small, round cells become flattened and organized into parallel, longitudinal columns (Smits et al. 2001;Provot and Schipani 2005). At this stage, prehypertrophic chondrocytes undergo irreversible growth arrest, a process for which the parathyroid hormone-related peptide (PTHrP) and Indian Hedgehog (Ihh) signaling pathways appear to have important roles (Lefebvre and Smits 2005;Gao et al. 2001).

The differentiation of chondroblasts into hypertrophic chondrocytes represents a major phenotypic switch (Provot and Schipani 2005;Lefebvre and Smits 2005). During this transition, the hypertrophic cells upregulate Collagen X expression and cease to express early cartilage markers (Lefebvre and Smits 2005;Kobayashi and Kronenberg 2005). In addition to exiting the cell cycle, the small, round cells progressively increase their cytoplasmic volume up to ten times (Lu et al. 1993). Given the established role of mTOR in the positive regulation of ribosome biogenesis and cell size (Martin and Hall 2005), we hypothesized that mTOR inhibition would affect chondrocyte growth and differentiation. We designed our studies to take advantage of the highly specific mTOR inhibitor, rapamycin. We studied the effect of mTOR inhibition in the well characterized ATDC5 chondrogenic cell line to elucidate the mechanism by which mTOR affects chondrogenesis. The effect of rapamycin on chondrocyte growth and differentiation was examined in a physiological model, fetal rat metatarsal explants. These studies on the role of mTOR in the development of metatarsal chondrogenesis was extended to an in vivo in which fetal rats were injected with rapamycin in situ then allowed to develop further for an additional day. Results obtained using these complementary models support the conclusion that mTOR has an important role in chondrocyte proliferation and differentiation.

Materials and Methods

Materials

Antibodies directed towards phosphorylated ribosomal protein S6 (ser 235/236) were obtained from Cell signaling Technology Inc. (Beverly, MA). Immunohistochemistry reagents were purchased from Vector Laboratories, Inc Burlingame, CA). Indian hedgehog peptide was obtained from R&D Systems (Minneapolis, MN). Indian hedgehog antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies against mouse type X collagen were kindly provided by Drs. William Horton and Greg Lunstrum (Oregon Health & Science University, Portland, OR). Other reagents were obtained from sources used in previous studies (Phornphutkul et al. 2004;Phornphutkul et al. 2006).

Cell Culture Studies

The ATDC5 cells from the Riken Cell Bank (Tsukuba, Japan) were cultured and analyzed for differentiation and proliferation as described previously (Phornphutkul, Wu, Yang, Chen, and Gruppuso 2004). Western immunoblotting was performed using standard methods (Phornphutkul, Wu, Yang, Chen, and Gruppuso 2004). ATDC5 cell proliferation was determined as the incorporation of ³H-thymidine into DNA (Curran, Jr. et al. 1993).

Organ Culture Studies

Metatarsal bones were dissected from fetal rats on embryonic day 19 (E19). All animal studies were approved by the Rhode Island Hospital Institutional Animal Care and Use Committee. Metatarsals were cultured as described previously (Phornphutkul, Wu, and Gruppuso 2006). Metatarsal length and surface area was measured daily and each experiment was repeated at least 3 times.

For morphometric analyses, three 5 μm longitudinal sections were obtained from each metatarsal and stained using hematoxylin and eosin. From each of the three sections, the heights of the proliferative zone, hypertrophic zone and the zone of mineralization were measured by an observer who was blinded to the experimental conditions. The hypertrophic zone was defined as that region containing cells greater than 9 μm in diameter (Wu and De Luca 2004;Bagi and Burger 1989).

Cell proliferation in the metatarsal bone explants was assessed by measuring BrdU incorporation (Phornphutkul, Wu, and Gruppuso 2006). Sections were counter-stained with hematoxylin AS for total cell counts. Image analysis was performed using the IPLab imaging program, Scanalytics Inc (Fairfax, VA).

For immunohistochemistry studies, explants were fixed overnight in 10% buffered formalin then embedded in paraffin. Sections were washed in xylene and rehydrated in graded ethanol. Endogenous peroxidase activity was quenched by incubation in 3% H₂O₂ for 15 min. After pre-incubation with 1.5% serum for 30 min, the sections were incubated for 30 min at room temperature with goat Ihh polyclonal antibody (1:600) or rabbit polyclonal antisera raised against mouse type X collagen peptides, anti-NC2 domain (1:200). Phopho-S6 antibody was diluted 1:100 prior to overnight incubation. Sections were counter stained with hematoxylin AS for total cell count. Images were acquired and image analysis was performed using the IPLab imaging program.

In Vivo Studies on Metatarsal Growth

Cesarean section was performed on E19. The uterus was exposed and a mixture containing BrdU (20 μg/g) plus rapamycin (2.5 μg/g) or DMSO (control) was administered to the fetuses by intraperitoneal injection through the uterine wall (Boylan et al. 2001). DMSO and rapamycin injections were given to alternating fetuses along both uterine horns. The uterus was placed back into the abdominal cavity and the pups were delivered 24 hours later. Fetal metatarsal bones were harvested as described above and fixed in formalin for further investigation.

Polymerase Chain Reaction (PCR)

For determination of collagen X, PTHrP, Ihh and BMP-2 expression in ATDC5 cells, total RNA was prepared using Tri Reagent purchased from Molecular Research Center, Inc. (Cincinnati, Ohio). Primer sequences used for detection of type X collagen mRNAs were those used by Negishi et al. (Negishi et al. 2000). Primer sequences are given in Table 1. After reverse transcription, PCR was performed as previously described (Phornphutkul, Wu, Yang, Chen, and Gruppuso 2004).

Table 1.

Gene	F- Primer	R- Primer
Col X	5′-AAAGCTTACCCAGCAGTAGG-3′	5′-ACGTACTCAGAGGAGTAGAG-3′
Ihh	5′-TGGATATCACCACCTCAGAC-3′	5′-GATTGTCCGCAATGAAGAGC-3′
PTHrP	5′-CAAGTCCTTGGAAGATCTTC-3′	5′-TGGATATCACCACCTCAGAC-3′
BMP-2	5′-GGTCTTTGCACCAAGATGAAC-3′	5′-CAACCCTCCACAACCATGTC-3′
Actin	5′-CACCCTGTGCTGCTCACCGAGGCC-3′	5′-CCACACAGATGACTTGAGCTCAGG-3′

Statistical Analysis

Data are expressed throughout as mean and standard deviation except where noted. The significance of differences between groups was determined by one-way analysis of variance followed by a Tukey post-hoc test.

Results

Effects of mTOR inhibition on longitudinal bone growth and growth plate chondrogenesis

Metatarsal bones were harvested and cultured in serum free medium in the absence or presence of 10 μg/ml insulin and 50 nM rapamycin (n, 12–20 per group). The length of each metatarsal was measured daily for 3 days (Fig. 1A). Explants cultured in the presence of insulin showed enhanced linear growth, an increment of approximately 60% from baseline. When rapamycin was added to media containing insulin, the magnitude of the insulin-dependent growth stimulation was significantly diminished. Rapamycin alone had no effect on long bone growth over a period of 3 days.

Fig. 1.

The effect of rapamycin on the growth of fetal rat metatarsal explants. Panel A: The total length of metatarsal explants was measured after 1, 2 and 3 days in culture. Growth was measured as the cumulative increase in length expressed as a percent of starting length. Control (no insulin, no rapamycin; open circles), 10 μg/ml insulin (unfilled diamonds), 50 nM rapamycin (filled circles) or 10 μg/ml insulin and rapamycin (filled diamonds). Results are shown as the mean +/− standard deviation for 12 to 20 samples per data point. *, P<0.05 versus all other groups. P <0.05 comparing insulin and insulin with rapamycin. Panel B: Bones were cultured in the absence or presence of 10 μg/ml insulin or 50 nM rapamycin for 72 hr. During the last 12 hr of the incubation period, BrdU was added to the culture medium. At the end of the experiment, all bones were fixed in 10% buffered formalin, embedded in paraffin, and 5 μm longitudinal sections were prepared. Following staining for BrdU incorporation, the BrdU-labeling index was calculated as the number of BrdU-labeled cells within a fixed area divided by the total number of cells in that same area. The areas that were analyzed contained on average approximately one hundred cells. For each treatment group, eight bones were sampled and two growth plate sections of each bone were analyzed for n=16 per group. There were no significant differences between groups. Panel C: The effect of rapamycin on zone-specific growth of fetal rat metatarsal explants was examined. Formalin-fixed explants were sectioned and stained using hematoxylin and eosin. The lengths of the explant and of the hypertrophic zone were measured by an observer who was blinded to the experimental conditions. The hypertrophic zone was defined as that region containing cells greater than 9 μm in diameter. Results for the length of the hypertrophic zone, expressed as percent of total length, are shown as mean +1 SD for n = 9–12 per condition. *, P<0.05 versus all other conditions

To assess chondrocyte proliferation in the explants, we performed immunohistochemical staining for BrdU incorporation (Fig. 1B). Insulin stimulated chondrocyte proliferation. The addition of rapamycin to media containing insulin showed a trend toward decrease in BrdU incorporation, but this change was not statistically significant. There was no effect of rapamycin alone.

Given the relative lack of an effect of rapamycin on chondrocyte proliferation despite decreased linear growth, we examined the histology of explants cultured under the conditions employed above in order to assess the effect on the hypertrophic zone (Fig. 1C). After 3 days in culture, the proportion of total bone length occupied by the hypertrophic zone was significantly increased by insulin. The addition of rapamycin to media containing insulin showed compete inhibition of the insulin effect, while the presence of rapamycin in the absence of insulin again had no effect.

On H&E staining (Fig. 2A), the growth plates of the metatarsals exposed to insulin were generally larger, both in length and width, relative to bones grown in the absence of insulin. The size of the hypertrophic zone was significantly enhanced (Fig. 2B). The growth plates of the metatarsals exposed to rapamycin appeared to have a modest but consistent degree of disorganization in the columnar proliferating zone. This occurred in metatarsals cultures under basal conditions as well as in the presence of insulin. Although we were unable to quantify this effect, it was noted by multiple observers across multiple experiments.

Fig. 2.

Effect of rapamycin on fetal metatarsal explants morphology and S6 phosphorylation. Panel A shows H&E staining of E19 metatarsal bones after 72 hours in culture. Representative photomicrographs show bones cultured under control conditions or in the presence or absence of 10 μg/ml insulin and 50 nM rapamycin (X4). Panel B shows representative photomicrograph at higher magnification (X20). Panel C shows representative analyses of metatarsal explants from the same experimental groups stained for phospho-S6 (X10).

Given the absence of an effect on BrdU incorporation at the end of the experimental period, we sought to confirm that rapamycin was absorbed into the growth plate and that its effect was persistent. To do so, we performed immunohistochemistry for phospho-S6 (Fig. 2C). Phospho S6 staining was most intense in the prehypertrophic and hypertrophic zones, staining was also observed in the proliferative zone. Rapamycin (50 nM) completely abolished phospho-S6 staining at the end of a 3-day incubation period.

We evaluated the effects of rapamycin on chondrocyte differentiation by immunohistochemical detection of collagen X in intact metatarsals. No change was seen (data not shown). However, staining was minimal, perhaps indicating that the in vitro conditions used for the explants and/or the time in culture was inadequate to allow for significant accumulation of collagen X. In contrast, Ihh was readily detected. We did not observe a consistent effect of rapamycin on staining for Ihh. Quantitation of Ihh staining did not show differences between the four groups (data not shown).

Effects of mTOR inhibition on growth plate chondrogenesis in vivo

In order to perform complentary, physiological experiments, we turned to an approach that our group has used previously to manipulate mTOR signaling in vivo, the administration to late gestation fetal rats in situ (Boylan, Anand, and Gruppuso 2001). Cesarean sections were performed on E19 and alternating fetuses were injected with rapamycin plus BrdU or DMSO plus BrdU (8 fetuses per group). The fetuses were delivered 24 hours later and the second, third and fourth fetal metatarsal bones were harvested. To confirm that the fetus had received rapamycin, we performed immunohistochemistry staining for phospho-S6 from each fetus (representative result shown in Fig. 3A). Similar to the explant model, phospho-S6 staining was most intense in the prehypertrophic and early hypertrophic chondrocytes, while minimal staining was observed in the proliferating zone. Staining was markedly reduced in both metatarsals and liver (not shown) from fetuses administered rapamycin. These results were confirmed by immunoblotting of liver homogenates for phospho- and total S6 of the corresponding animals (not shown).

Fig. 3.

Effect of in vivo rapamycin administration on fetal metatarsals. Cesarean sections were performed on E19 and alternating fetuses were injected with rapamycin plus BrdU or DMSO plus BrdU (8 fetuses per group). The fetuses were delivered 24 hours later and the second, third and forth fetal metatarsal bones were harvested from each fetus. Metatarsals underwent routine histological processing and sectioning (5–7 μm). Representative results are shown for immunohistochemical staining for phospho-S6 (Panel A), Ihh (Panel B), BrdU (Panel C) and collagen X (Panel E). Panel D shows the mean plus standard deviation for BrdU labeling index. Data represent the results for 8 bones, two growth plate sections per bone, for each experimental condition. Panel F shows the quantitation of collagen X staining, measured as the percent of the total area that was stained. *, P<0.05 versus control group.

Results of immunohistochemical staining for Ihh protein (Fig. 3B) was also unaffected by rapamycin administration. To study the effect of rapamycin on chondrocyte proliferation in vivo, we examined BrdU incorporation by proliferative zone chondrocytes (Fig. 3C, D). There was no difference between the control and rapamycin groups. However, collagen X was decreased in response to rapamycin (Fig. 3E, F).

Effects of mTOR inhibition on ATDC5 cell differentiation and proliferation

We chose to study the effect of mTOR inhibition on chondrocyte proliferation and differentiation using the ATDC5 cell line. Our purpose was to elucidate the molecular mechanism by which mTOR affects chondrogenesis. ATDC5 cells maintained in differentiating media were exposed to rapamycin at 0, 2, 20 or 200 nM for 12 days. The media was changed every 2 days so as to maintain consistent rapamycin concentrations. At the end of the experiment, day 12, the degree of differentiation was assessed by measuring Alcian Blue staining. We observed a decline in Alcian Blue staining in cells exposed to 20 or 200 nM rapamycin (Fig. 4A). We also studied the effect of rapamycin on the expression of collagen X, a known marker for hypertrophic chondrocytes. It has previously been shown that collagen X expression is readily detectable by day 6–8 in culture (Chen et al. 2005). Cell lysates were collected after maintaining cells in differentiating media for 8 days. Rapamycin at 2, 20 and 200 nM reduced collagen X expression (Fig. 4B). Similar results were obtained in a separate experiment using 5, 10 and 50 nM rapamycin (data not shown). The effectiveness of rapamycin in inhibiting the mTOR pathway over a full 14-day experimental period was confirmed by Western immunoblotting for S6 phosphorylation (Fig. 4C).

Fig. 4.

The effect of rapamycin on ATDC5 cell differentiation. ATDC5 cells were grown in DMEM (Cont) or differentiating medium (+ Insulin) containing 0, 2, 20 or 200 nM rapamycin. Two markers of differentiation were assessed. Panel A: Cells were harvested after 12 days and stained for proteoglycan accumulation using Alcian Blue. The absorbance of the dye extracted from each well is shown as the mean + 1 SD of triplicate measures for each condition. Similar results were obtained in an additional experiment. * P<0.001 versus + Insulin, 0 Rapamycin. Panel B: A similar experiment was performed in which cells were maintained in differentiating medium with either 0, 2, 20 or 200 nM rapamycin for 8 days. Total RNA was isolated and analyzed by semi-quantitative RT-PCR. Upper panel is a representative analysis for collagen X expression. The lower panel shows β-actin expression, which was used as a control. Panel C: Cells grown under control condition, or grown in the presence of insulin or insulin plus 20 nM rapamycin were used after 14 days for the preparation of cell lysates. The lysates were analyzed by Western immunoblotting for phosphorylated S6 (upper panel) and total S6 (lower panel).

We used Neutral Red staining to assess the affect of rapamycin on total cellular mass. No effect was seen (Fig. 5A). However, ATDC5 cell proliferation was inhibited approximately 25% by rapamycin, as determined by ³H-thymidine incorporation (Fig. 5B). This effect was dose-independent but highly potent, showing the same level of inhibition at 2, 20 and 200 nM concentrations.

Fig. 5.

The effect of rapamycin on ATDC5 cell mass and proliferation. ATDC5 were grown under conditions identical to those used to obtain the data in Figure 4A. Panel A: Cells were harvested on day 12 and stained with Neutral Red. The absorbance of the dye extracted from each well is shown as the mean + 1 SD of triplicate measures for each condition. Similar results were obtained in an additional experiment. Panel B: Subconfluent ATDC5 cells were serum starved overnight, then treated with [Arg³]IGF-I and various concentrations of rapamycin for 24 h. During the last 6 h of the incubation period ³H-thymidine (1 μCi/well) was added. Results for ³H incorporated into DNA are shown as the mean + 1 SD of triplicate measures for each condition. * P<0.001 versus + IGF-I, 0 Rapamycin.

To study the critical time period during which rapamycin affects chondrocyte differentiation, we added the drug at a concentration of 20 nM to differentiating chondrocytes on days 0, 2, 4, 6, 8 or 10 during the differentiation process. The cells were harvested at 12 days and assessed for Alcian blue accumulation. Results (Fig. 6) showed that Alcian blue accumulation was significantly decreased when rapamycin was added within the first 2 days of the differentiation process. If added after 4 days, Alcian blue accumulation was unaffected by rapamycin.

Fig. 6.

The effect of rapamycin on ATDC5 cell differentiation. ATDC5 cells were induced to differentiate in the absence or presence of 20 nM rapamycin. Rapamycin was added beginning on day 0, 2, 4, 6, 8 or 10. Cell lysates were collected at day 12 and assayed for proteoglycan accumulation using Alcian Blue. The absorbance of the dye extracted from each well is shown as the mean + 1 SD of triplicate measures for each condition. Similar results were obtained in a second, replicate experiment. * P<0.001 versus cells grown without rapamycin.

The same experimental design was used to assess the effect of rapamycin on the expression of a group of genes known to have critical roles in chondrocyte differentiation (Fig. 7A). When rapamycin is added to differentiating ATDC5 cells, the expression of collagen X, Ihh and BMP-2 decreased, PTHrP expression increased marginally, and Sox-6 and Runx-2 accumulation was unaffected (data not shown for the latter two genes). The effect of rapamycin on collagen X expression occurred only if the drug was added to the culture media within the first 4 days. In contrast, the effects on Ihh and PTHrP expression (the respective decrease and increase) were observed only if rapamycin was added within the first 2 days of the differentiation period. Immunoblotting for phospho-S6 at various stages of ATDC5 differentiation showed a pattern whereby S6 was highly phosphorylated during the first 2 two days, decreased by 50% on day 4 and completely disappeared by day 8 (Fig. 7B). Levels of total S6 relative to protein content were constant (Fig. 7B). We attempted to perform a corollary experiment in which rapamycin was added to ATDC5 cultures for limited periods of time then “washed out.” These experiments proved to not be feasible because the effect of rapamycin on S6 phosphorylation persisted for up to 10 days after its removal (data not shown).

Fig. 7.

The effect of rapamycin on ATDC5 cell differentiation. Panel A: ATDC5 cells were induced to differentiate in the absence or presence of 20 nM rapamycin. Rapamycin was added beginning on day 0, 2, 4, or 6. Total RNA was isolated at 8 days and analyzed by semi-quantitative RT-PCR for collagen X, Ihh, PTHrP and BMP-2. Triplicate samples were analyzed for each time-point. The lower panel shows β-actin expression, which was used as a control. Panel B: Cells grown under control conditions, or grown in the presence of insulin or insulin plus 20 nM rapamycin were collected after 0, 2, 4, 8 and 10 days for the preparation of cell lysates. The lysates were analyzed by Western immunoblotting for phosphorylated S6 (upper panel) and total S6 (lower panel).

Based on the results shown in Figure 7, we focused our attention on Ihh, a protein known to be expressed predominantly in prehypertrophic chondrocytes. We hypothesized that rapamycin may affect early-phase chondrocyte differentiation by attenuating Ihh expression, thereby limiting the pro-differentiating effects of this signaling protein. We tested this hypothesis by the direct addition of Ihh to the differentiating medium to assess its ability to rescue ATDC5 cells from the inhibitory effect of rapamycin (Fig. 8A and 8B). Ihh protein was added to differentiation medium containing either 0 or 20 nM rapamycin at a concentration of 0, 5, 10 or 20 nM. We measured Alcian blue accumulation and collagen X expression at 11 days (Fig. 8A and 8B). Consistent with our other results, ATDC5 cells exposed to rapamycin showed inhibition of both differentiation markers. The addition of Ihh to the culture medium resulted in a dose-dependent reversal of the rapamycin-mediated inhibition of Alcian blue accumulation. When 20 nM Ihh was added to the differentiating ATDC5 cells grown in media containing rapamycin, the inhibitory effect of rapamycin on Alcian blue accumulation (Fig. 8B) was completely reversed. Similar results were seen for BMP-2 and collagen X accumulation (Fig. 9). In addition, PTHrP expression showed a marginal decrease (Fig. 9). In absence of rapamycin, the addition of Ihh had no effect on any indicators of differentiation (data not shown).

Fig. 8.

The effect of Ihh on ATDC5 cell differentiation. ATDC5 were grown in DMEM or exposed to differentiating medium (+ Insulin) containing 20 nM rapamycin. Different concentrations of Ihh were added at the beginning of the differentiation period. Two markers of differentiation were assessed. Panel A: Cells were harvested after 11 days and stained for proteoglycan accumulation using Alcian Blue. The absorbance of the dye extracted from each well is shown as the mean + 1 SD of triplicate measures for each condition. * P<0.01 versus + Insulin, + Rapamycin and −Ihh. **P<0.0001 versus + Insulin, + Rapamycin and −Ihh. Panel B: A similar experiment was performed using higher Ihh concentrations. * P<0.0001 versus + Insulin, + Rapamycin and −Ihh.

Fig. 9.

A similar experiment was performed ATDC5 cells were maintained in differentiating medium with 20 nM rapamycin for 11 days. Total RNA was isolated and analyzed by semi-quantitative RT-PCR. Upper panels show representative analyses for collagen X, Ihh, PTHrP and BMP-2 expression. The lower panel shows β-actin expression, which was used as a control.

Discussion

The present studies were aimed at characterizing the role of mTOR in the regulation of chondrocyte growth and differentiation. The overarching purpose of these studies was to test the hypothesis that this key nutrient-sensing pathway may provide for a mechanism whereby nutrients can directly modify linear growth via effects at the growth plate. In order to directly examine the effects of mTOR inhibition on bone growth, we used a fetal rat metatarsal explant culture system. We were able to confirm an inhibitory effect of rapamycin on the linear growth of the explants. This inhibition was seen only when bones were stimulated to grow by the inclusion of pharmacological insulin concentrations in the culture medium. We (Phornphutkul, Wu, Yang, Chen, and Gruppuso 2004) and others (Pandini et al. 2002) have demonstrated that the effects of high concentration insulin are likely to be mediated by IGF-I receptors and their downstream signaling pathways. In the fetal rat metatarsal explant model, similar growth was observed in bones exposed to 10 μg/ml insulin or 10 nM of IGF-I. The concentration of the latter is considered physiological, mimicking the in vivo environment (Smith et al. 1988;Luna et al. 1983).

Using the conditions employed for the present studies, the growth inhibitory effect of rapamycin was precisely that stimulated by the addition of insulin. This reduction was largely accounted for by a decrease in the proportion of bone length contributed by the hypertrophic zone. As has been shown in other studies {Wilsman, 1996 129/id;Breur, 1991 10/id}, chondrocyte hypertrophy has an important role in longitudinal bone growth. In rapidly growing bones, while 9% of bone length is contributed by proliferating cells and 32% by matrix synthesis throughout the growth plate, the process of hypertrophy accounts for approximately 60% of bone elongation. DNA synthesis in the explants was minimally inhibited by rapamycin, and the results did not reach statistical significance even though we studied 16 growth plates per condition.

The use of an in vivo model for the exposure of developing metatarsals to rapamycin was employed to permit studies in a physiological system with a normal macroenvironment for chondrocyte development. The primary disadvantage of this system is that longitudinal measurements of bone growth over time are not possible. Nonetheless, the model did allow us to see a direct effect on differentiation. The context for rapamycin effect in the in vivo system was the stimulation of bone growth by endogenous factors rather than the addition of growth factor to culture media, further supporting a physiological role for mTOR signaling in chondrocyte differentiation.

We noted a subtle effect of rapamycin on the architectural organization of the columnar proliferating and prehypertrophic cells in the explant cultures. This led us to examine the expression of Ihh, a protein that is expressed in prehypertrophic cells (Lefebvre and Smits 2005). Ihh is known to be involved in spatial orientation of developing chondrocytes (Lai and Mitchell 2005;Kobayashi et al. 2002). We did not observe an effect of rapamycin on the expression of Ihh at the protein level. However, immunohistochemical staining can hardly be considered a precise, quantitative analytical method when applied to explant sections. We therefore employed a cell culture model.

Using the well characterized ATDC5 chondrogenic cell line, we observed a modest inhibitory effect of rapamycin on ATDC5 cell proliferation. The role of rapamycin as an anti-proliferation agent has been established in many cell types (Law et al. 2006;Shi et al. 2005;Nelsen et al. 2003). However, inhibition of proliferation by rapamycin was unlikely to account for decreased chondrocyte differentiation. In fact, we showed previously that the inhibition of ATDC5 proliferation by the addition of the Erk inhibitor PD98059 results in enhanced differentiation (Phornphutkul, Wu, Yang, Chen, and Gruppuso 2004). In addition, total cell mass as determined by Neutral Red staining was not significantly affected by rapamycin.

Our studies in ATDC5 cells further supported a role for mTOR signaling in the chondrocyte differentiation program. We went on to determine that the critical time point for mTOR inhibition in ATDC5 cells occurs during the first 2 days of differentiation. During that time, numerous genes undergo induction of their expression, Ihh being one of them (Chen, Fink, Zhang, Ebbesen, and Zachar 2005). Ihh is a key contributor to chondrocyte proliferation and differentiation (Lai and Mitchell 2005). Mutations in Ihh gene have been linked to skeletal developmental defects, brachydactyly type A-1 and acrocapitofemoral dysplasia (Gao, Guo, She, Shu, Yang, Tan, Yang, Guo, Feng, and He 2001;Hellemans et al. 2003). Mice with a chondrocyte-specific deletion of Ihh show a decrease in chondrocyte proliferation and maturation (St Jacques et al. 1999). Ihh not only directly affects chondrocyte growth and differentiation, but it has been shown to modulate the expression of PTHrP, which itself is a critical regulator of chondrocyte growth and differentiation.

Rapamycin exposure of ATDC5 cells was associated with inhibition of Ihh expression and, conversely, augmented expression of PTHrP. More importantly, exposure of ATDC5 cells to Ihh could rescue them from rapamycin-induced inhibition of differentiation, as measured by collagen X expression and increased proteoglycan accumulation. We interpret these data as indicating that mTOR-mediated induction of Ihh expression is a significant component of the mechanism by which mTOR signaling promotes differentiation.

Despite its established role in chondrocyte differentiation, the mechanisms that regulate the expression of Ihh are not well characterized. Several mechanisms, including those involving Runx-2, BMPs, retinoic acid, and the Erk and p38 MAP kinases have been shown to activate Ihh transcription (Yoshida et al. 2004;Seki and Hata 2004;Yoshida et al. 2001;Lai et al. 2005). Rapamycin did not affect Runx2 expression in our studies. On the other hand, FGF and PTH have been shown to inhibit Ihh expression. mTOR is not known to be involved in signaling downstream from PTH, FGF or BMP.

mTOR has been shown to be a key regulator of mRNA cap-dependent translation initiation and of ribosomal protein translation in numerous biological systems (Martin and Hall 2005;Jaeschke et al. 2002). These functions manifest in some systems as the direct regulation of cell size (Wullschleger, Loewith, and Hall 2006;Ruvinsky and Meyuhas 2006). The differentiation of proliferating chondrocytes is accompanied by marked cell hypertrophy, with the volume of the cells increasing more than ten-fold (Lu, Iwamoto, Fanning, Pacifici, and Olsen 1993). Cell hypertrophy, in general, is tied to ribosome biogenesis and activation of global protein translation (Tee et al. 2002;Tee and Blenis 2005). Thus, it is not unexpected that inhibition of mTOR might affect long bone growth via an inhibition of cell hypertrophy. Rapamycin is widely used as a potent immunosuppressive drug to prevent graft rejection in transplantation (Webster et al. 2006;Mota 2005). In this context, the drug has not been shown to have side effects specific to skeletal growth. One ramification of our studies might be the evaluation of rapamycin effect on linear growth in the case of pediatric usage.

Another implication of our findings, given mTOR’s established function as a nutrient sensor (Kimball and Jefferson 2006a;Wang et al. 2005;Avruch et al. 2005), is that limited nutrient availability may have a direct effect on long bone growth at the level of the growth plate chondrocyte. mTOR has been shown to be responsive to branched-chain amino acids, leucine in particular, as well as other nutrients (Kimball and Jefferson 2006b). While the role of nutrition in modulating the GH-IGF-I system is established (Oldham and Hafen 2003;Kimball and Jefferson 2004), our results demonstrate that inhibition of mTOR, a component of the IGF-I signaling network, can also modulate bone growth. Similarly, the direct effect of insulin on growth plate chondrocyte differentiation that we demonstrated previously (Phornphutkul, Wu, and Gruppuso 2006) may be bypassed or amplified by direct mTOR-mediated actions of nutrients. In the aggregate, the effects of nutrient availability on the GH-IGF-I system, insulin secretion and signaling and direct signaling through mTOR may account for the profound overall effect of nutritional status on linear growth.