Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Cell Metab. Author manuscript; available in PMC 2009 Oct 5.
Published in final edited form as:
PMCID: PMC2756713

ATF4 mediation of NF1 functions in osteoblast reveals a nutritional basis for congenital skeletal dysplasiae


The transcription factor ATF4 enhances bone formation by favoring amino acid import and collagen synthesis in osteoblasts, a function requiring its phosphorylation by RSK2, the kinase inactivated in Coffin-Lowry Syndrome. Here, we show that in contrast, RSK2 activity, ATF4-dependent collagen synthesis, and bone formation are increased in mice lacking neurofibromin in osteoblasts (Nf1ob−/− mice). Independently of RSK2, ATF4 phosphorylation by PKA is enhanced in Nf1ob−/− mice, thereby increasing Rankl expression, osteoclast differentiation, and bone resorption. In agreement with ATF4 function in amino acid transport, a low-protein diet decreased bone protein synthesis and normalized bone formation and bone mass in Nf1ob−/− mice without affecting other organ weight, while a high-protein diet overcame Atf4−/− and Rsk2−/− mice developmental defects, perinatal lethality, and low bone mass. By showing that ATF4-dependent skeletal dysplasiae are treatable by dietary manipulations, this study reveals a molecular connection between nutrition and skeletal development.


The molecular elucidation of human skeletal dysplasiae has greatly contributed to our current knowledge of skeletogenesis (Olsen et al., 2000). Indeed, many of the transcription factors implicated in this process such as Msx1, Sox9, Runx2, Twist-1 and Atf4 are mutated or their activity is affected in several dysplasiae (Bi et al., 2001; Bialek et al., 2004; Ducy et al., 1997; Satokata and Maas, 1994; Wagner et al., 1994; Yang et al., 2004). Conversely, the identification through molecular means and/or mouse genetics of transcription factors regulating chondrocyte and osteoblast differentiation led, in few cases, to a better molecular understanding of the pathogenesis of human skeletal dysplasiae. A potential implication of these advances is that one could propose and test, at least in mice, adapted therapies for some of these diseases based on the knowledge of the molecular mode of action of these transcription factors.

In the vast majority of cases, however, molecular elucidation of skeletal dysplasiae has not led yet to a better understanding of the disease or to a more rational treatment. Neurofibromatosis type I (NF1), a disease caused by loss of function mutations in NF1, a gene encoding the Ras GAP (GTPase Activating Protein) neurofibromin (Dasgupta and Gutmann, 2003; Klose et al., 1998; Parada, 2000) belongs to this latter category. Indeed, besides well-described tumors of the nervous system, NF1 patients often display well-documented skeletal abnormalities whose cellular and molecular pathogenesis is not fully elucidated, in part because no animal model has been available to study them (Alwan et al., 2005; Jacquemin et al., 2003; Kuorilehto et al., 2004, 2005; Kwok et al., 2002; Ruggieri et al., 1999; Sigillo et al., 2002; Simsek et al., 2003; Stevenson et al., 1999). As a result of the paucity of molecular knowledge of how neurofibromin affects bone biology the only available treatment for these often debilitating manifestations remains surgery.

ATF4, an osteoblast-enriched member of the CREB family of transcription factors, is necessary for the latest phases of osteoblast differentiation, for extracellular matrix (ECM) synthesis by osteoblasts, i.e., bone formation, and for osteoclast differentiation (Elefteriou et al., 2005; Yang et al., 2004). These different functions require phosphorylation of ATF4 by distinct kinases. RSK2, the kinase mutated in Coffin-Lowry syndrome (CLS), regulates the first two activities of ATF4 while PKA regulates its third function. Remarkably, ATF4 favors bone formation through two distinct mechanisms, both of which depend on its phosphorylation by RSK2. On the one hand, it positively regulates osteoblast-specific genes at the transcriptional level, and on the other hand, it promotes synthesis of Type I collagen, the main component of the bone ECM, by favoring amino acid import (Yang et al., 2004). In that respect the fact that the skeletal manifestations of CLS worsen overtime (Hunter, 2002; Touraine et al., 2002) raises the hypothesis that it may be due, primarily, to a lack of collagen synthesis during skeletal growth. This hypothesis is testable using mouse models. In any case the central role played by ATF4 in regulating osteoblasts main functions suggests that it may be implicated in the development of additional bone diseases beside CLS.

In the present study, we generated mice lacking Nf1 only in osteoblasts (Nf1ob−/− mice) with the primary goal to elucidate the transcriptional events accounting for NF1 skeletal manifestations. This conditional deletion bypasses the embryonic lethality of a somatic deletion of the gene allowing us to analyze the function of Nf1 in bone (Zhu et al., 2001). Through molecular and genetic analyses of this mouse model we demonstrate here that ATF4 plays a crucial role in mediating neurofibromin signaling in osteoblasts. Surprisingly, further use of this mouse model and of Atf4−/− and Rsk2−/− mice led to the conclusion that, at least in mice, one can reverse and even prevent skeletal manifestations that are secondary to abnormal phosphorylation of ATF4 by RSK2 through simple dietary manipulations. Thus, these results not only provide a molecular basis for NF skeletal manifestations but also identify a hitherto unanticipated connection between amino acid intake and skeletal development.


Multiple bone phenotypes in Nf1ob−/− mice

To identify molecular events accounting for NF skeletal manifestations we studied neurofibromin functions in osteoblast, the bone cell type where Nf1 is expressed at the highest level (Figures 1A and 1B). Because Nf1-deficient embryos die before skeletogenesis is initiated (Jacks et al., 1994) and Nf1+/− mice do not display any appreciable bone abnormalities (Yu et al., 2005), we generated a mutant mouse strain lacking Nf1 only in osteoblasts (Nf1ob−/− mice) by crossing mice harboring a floxed allele of Nf1 (Zhu et al., 2001) with mice expressing Cre recombinase specifically in osteoblasts beyond E14.5 (α1(I) Collagen Cre transgenic mice) (Dacquin et al., 2002; Zhu et al., 2001). Excision at the Nf1 locus exceeded 90% in primary osteoblast culture originating from 4-day-old Nf1ob−/− mice while no excision could be detected in any other cell type tested including osteoclasts, the bone-resorbing cells (Figure 1C).

Figure 1
Increased bone formation, osteoid thickness, and bone resorption in Nf1ob−/− mice

Nf1ob−/− mice were born at the expected Mendelian ratio, did not display any skeletal patterning abnormalities and had a normal life span (Figure S1 in the Supplemental Data available with this article online, and data not shown). Histological and 3D microtomography analyses performed on vertebrae and long bones of male and female Nf1ob−/− mice at 2, 3, and 6 months of age showed a progressive increase in bone volume, i.e., a high bone mass phenotype (Figures 1D and 1E, S1C, and data not shown). Static and dynamic histomorphometric analyses revealed a marked increase in osteoblast surface and number and in the bone formation rate (Figure 1D) indicating that this high bone mass phenotype was secondary to an increase in bone formation.

This histomorphometric analysis also uncovered two other and unanticipated abnormalities in Nf1ob−/− mice. The first one was, in all mutant bones analyzed, a significant increase in osteoid volume, i.e., in the amount of bone ECM that is not mineralized (Figure 1F). Second, Nf1ob−/− mice displayed an increased bone resorption characterized by an increase in the number of tartrate resistant acid phosphatase (TRAP)-positive multinucleated osteoclasts, and in urinary elimination of deoxypiridinoline (Dpd), a biomarker of osteoclasts activity (Figure 1G). A similar increase in Dpd urinary elimination has been reported in NF1 patients, underscoring the relevance of the Nf1ob−/− mice as a model of NF1 bone disease (Stevenson et al., 2005).

Consistent with the osteoblast-specific gene inactivation performed, coculture of Nf1−/− osteoblasts with bone marrow macrophages (BMMs), originated from either WT or Nf1ob−/− mice yielded a significantly higher number of multinucleated differentiated osteoclasts than identical cocultures of BMMs (of either genotypes) with WT osteoblasts (Figure 1H). Despite the role of Nf1 signaling in osteoblast-dependent osteoclast differentiation, vitamin D was still required for osteoclast differentiation in this ex vivo assay. In addition, spleen cells isolated from WT or Nf1ob−/− mice differentiated equally well into multinucleated osteoclasts in the presence of RANKL and m-CSF excluding an autonomous defect in osteoclasts from Nf1ob−/− mice (Figure 1I). Thus Nf1 inactivation in osteoblasts affects osteoclast differentiation in addition to bone formation.

Molecular bases of the increased osteoid volume in Nf1ob−/− mice

To elucidate the molecular mechanisms whereby neurofibromin deletion in osteoblasts affects altogether bone formation, bone resorption and ECM mineralization, we used this latter phenotype as an entry point. Type I Collagen and tissue nonspecific alkaline phosphatase (Tnap), which are together necessary and sufficient to induce bone ECM mineralization in mice, were normally expressed in Nf1ob−/− mice as were Runx2, Osterix and Atf4 mRNA, the three main transcription factors governing osteoblast differentiation (Stein et al., 1996; Murshed et al., 2005; Wagner and Karsenty, 2001; Yang et al., 2004) (Figures 1J and 1K). The only molecular abnormality observed in Nf1−/− osteoblasts was an increase in the expression of Osteocalcin, a gene whose inactivation or overexpression in osteoblasts does not affect ECM mineralization (Ducy et al., 1996; Murshed et al., 2004). Furthermore, there was no overt abnormality of calcium and phosphate metabolism that could explain this defect in ECM mineralization (Figure S2). Taken together, these data ruled out that the increase in osteoid thickness in Nf1ob−/− mice was due to an abnormal expression of genes involved in osteoblast differentiation or ECM mineralization.

We also asked whether the ECM mineralization abnormality observed in Nf1ob−/− mice could be secondary to posttranscriptional events. To that end we measured Type I collagen production and TNAP activity in WT and Nf1ob−/− bones and osteoblasts. Bone collagen content, as assessed by total trabecular thickness and by Western blot was increased in Nf1ob−/− bones and osteoblasts compared to WT controls (Figures 1F and 1L). In contrast, TNAP activity was normal in Nf1−/− osteoblasts as was TNAP serum levels in Nf1ob−/− mice (Figure 1M). This increase in collagen secretion was not observed in osteoblasts isolated from another mouse model characterized by an increased osteoid thickness such as the Hyp mice (Eicher et al., 1976) indicating that it was a specific consequence of Nf1 deletion in osteoblasts (Figure 1L). Lastly, mineralization lag time, i.e., the time period between collagen synthesis and the initiation of mineralization, was increased in Nf1ob−/− mice, suggesting the existence of an additional function of Nf1 in bone mineralization (Figure 1F). Taken together, these observations indicated that a posttranscriptional increase in collagen synthesis as well as an increase in osteoblast number along with a delay in mineralization contributed to the increase in osteoid amount in Nf1ob−/− mice.

ATF4 mediates neurofibromin signaling in osteoblasts

The increase in collagen content in the face of a normal expression of Type I collagen along with the increase in Osteocalcin expression is the mirror image of what is seen in mice lacking Atf4. ATF4 is an osteoblast-enriched member of the CREB family of transcription factors regulating Osteocalcin expression, osteoclast differentiation and Type I collagen synthesis via its ability to favor amino-acid import (Elefteriou et al., 2005; Harding et al., 2000, 2003; Yang et al., 2004). This observation led us to test whether an increase in ATF4 activity contributed to the skeletal abnormalities in Nf1ob−/− mice.

To regulate Collagen synthesis and Osteocalcin expression ATF4 must be phosphorylated at serine 251 by RSK2, a substrate of the Ras/MEK/ERK signaling pathway (Dalby et al., 1998; Fisher and Blenis, 1996). Accordingly, RSK2 and ERK1/2 phosphorylation were increased in Nf1−/− osteoblasts and Ras activity was higher in Nf1−/− than in WT osteoblasts (Figure 2A). These alterations were specific of the ERK pathway as AKT and p38 phosphorylation was not affected in Nf1−/− osteoblasts. As expected the increase in RSK2 activation enhanced ATF4 phosphorylation at serine 251, conversely treatment of Nf1−/− osteoblasts with the MEK inhibitor U0126 reduced 251S-ATF4 phosphorylation, leading to a reduction of Osteocalcin expression (Figures 2A–2C). Likewise, treatment of Nf1ob−/− mice with a MEK inhibitor for 5 weeks reduced collagen synthesis and osteoid thickness without altering Type I collagen mRNA levels and bone volume (Figures 2D and 2E and data not shown). Taken together, these results established that, in osteoblasts, neurofibromin is a negative regulator of the RSK2-mediated phosphorylation of ATF4, thereby a negative regulator of Osteocalcin expression and of Type I collagen synthesis, the most abundant protein component of the bone ECM.

Figure 2
Nf1 deficiency in osteoblasts activates ATF4 via the MAPK pathway

On the other hand, to favor Rankl expression and osteoclast differentiation, ATF4 must be phosphorylated by PKA on serine 254, which is a RSK2-independent event (Elefteriou et al., 2005). PKA activity was increased in Nf1−/− compared to WT osteoblasts (Figure 2F), consequently ATF4 was more highly phosphorylated at serine 254 and Rankl expression was higher in Nf1−/− than in WT osteoblasts, while expression of Opg, a decoy receptor for RANKL, was unaffected (Figures 2G–2I). The PKA inhibitor H89 blunted the increase in Rankl expression in Nf1−/− osteoblasts further supporting the notion that neurofibromin-dependent PKA phosphorylation of ATF4 contributes to the osteoblast-mediated increase in bone resorption observed in Nf1ob−/− mice (Figure 2I).

Genetic evidence of ATF4 involvement in Nf1ob−/− mice bone phenotypes

To provide further support to the notion that ATF4 mediates neurofibromin signaling in osteoblasts we next used mouse models. We first constructed and analyzed transgenic mice overexpressing Atf4 specifically in osteoblasts (α1(I) Collagen-Atf4 mice) (Figures 3A and S3). α1(I) Collagen-Atf4 mice displayed the same three bone phenotypes observed in Nf1ob−/− mice: an increase in bone formation and bone mass, in bone resorption and in osteoid thickness (Figures 3B and 3D). At the molecular level the similarities between the two models were also striking and included an increase in Ocn and Rankl expression, two target genes of ATF4, and an increase in collagen content (but not its mRNA expression) in α1(I) Collagen-Atf4 mice (Figures 3E and 3F). The increase in Rankl expression explains the increase in bone resorption observed in these transgenic mice while the increase in collagen content together with an increase in osteoblast number provides an explanation for their bone formation and osteoid thickness abnormalities. This is consistent with the known role of Atf4 as a regulator of osteoblast differentiation and collagen synthesis (Yang et al., 2004).

Figure 3
In vivo evidence of ATF4 involvement in Nf1ob−/− bone phenotype

Next we generated and analyzed Nf1ob−/− mice lacking one allele of Atf4 or of Runx2 or Osterix, two other osteoblast-specific transcription factors. Six-month-old Nf1ob−/−;Atf4+/− mice displayed a significant reduction in collagen synthesis that was not observed in Nf1ob−/− mice lacking one allele of either Runx2 or Osterix (Figure 3G). Accordingly, osteoid thickness was decreased in Nf1ob−/−;Atf4+/− mice but not in Nf1ob−/−;Runx2+/− or Nf1ob−/−;Osx+/− mice (Figure 3H and data not shown). Taken together, these results provide genetic support to the notion that ATF4 mediates neurofibromin signaling in osteoblasts.

Low-protein diet corrects Nf1ob−/− mice bone formation abnormalities

What could be the therapeutic implications of these molecular findings? It is known that ATF4 promotes amino acid import into cells by activating genes encoding membranes transporters (Harding et al., 2003). Consequently, Atf4−/− osteoblasts are subject to amino acid depletion, a phenotype corrected by addition of nonessential amino acid to their culture medium (Yang et al., 2004). Thus we asked whether the ATF4-mediated increase in intracellular amino acid availability plays a role in the development of bone abnormalities in Nf1ob−/− mice.

WT and Nf1ob−/− mice were fed with either a normal diet (ND) or a LPD from birth to 4 months of age. To determine if the LPD affected amino acid pools in bone cells, we measured the activity of GCN2, a kinase that is activated by uncharged transfer RNAs and that phosphorylates serine 51 on the α subunit of the translation initiation factor 2 (eIF2α) resulting in reduced translation initiation and reduced protein synthesis (Dong et al., 2000; Harding et al., 2000). Levels of activated GCN2 and phosphorylated eIF2α were both increased in bones of LPD-fed WT mice but not in bones of LPD-fed Nf1ob−/− mice (Figure 4A). These observations suggested that dietary manipulation can affect amino acid pools in bone cells and that, as predicted these pools are greater in Nf1ob−/− mice. Accordingly, the amount of Type I collagen detectable by Western blot was markedly decreased in bones of Nf1ob−/− mice fed a LPD (Figure 4B). Despite the fact that GCN2 remained inactive in mutant mice, histomorphometric analyses showed that the high bone mass, increase in bone formation parameters and in osteoid thickness characteristic of Nf1ob−/− mice fed a normal diet were all corrected by the LPD (Figure 4C). Importantly, the LPD did not affect any of these parameters in WT mice nor the overall weight of mice or of their internal organs regardless of their genotypes (Figure S4). In addition, LPD did not influence bone resorption parameters in Nf1ob−/− mice indicating that the increase in PKA activity observed in Nf1-deficient osteoblasts was not secondary to the increase in RSK2 activity (Figure 4C). The LPD also normalized bone mass, osteoid thickness and bone formation parameters in 2-month-old α1(I) Collagen-Atf4 mice further demonstrating the involvement of ATF4 as a mediator of neurofibromin signaling in osteoblasts (Figure 4D).

Figure 4
Low-protein diet corrects RSK2-dependent bone abnormalities in Nf1ob−/− mice

High-protein diet corrects Rsk2−/− and Atf4−/− bone phenotypes

The remarkable efficacy of a LPD in rescuing skeletal manifestations of a disease caused by an increase in ATF4 activity suggested that, conversely, a high-protein diet (HPD) could rescue the skeletal manifestations observed in Rsk2-deficient mice, a mouse model of Coffin-Lowry syndrome (CLS) characterized by a decrease in ATF4 function.

To test this hypothesis, Atf4+/− and Rsk2+/− mothers were fed from the first day of gestation with a ND, a high-fat diet (HFD) or a HPD. After weaning, pups of each genotype were kept on this same diet up to 28 days of age. As previously reported (Tanaka et al., 1998), when Atf4+/− mothers and Atf4−/− pups received a ND only 10% of the pups recovered at 1 month of age were Atf4−/−. This perinatal lethality was not corrected by a HFD (Figure 5A), in contrast, the proportion of 1-month-old Atf4−/− mice alive doubled when Atf4+/− mothers and Atf4−/− pups were fed a HPD. This rescue was specific of the Atf4 mutation since HPD did not affect the perinatal lethality of mice lacking Runx2 or Osterix, the two other osteoblast-specific transcription factors whose deletion prevents osteoblast differentiation (Komori et al., 1997; Nakashima et al., 2002; Otto et al., 1997). That phosphorylation of GCN2 as well as eIF2α was decreased in HPD-fed Atf4- and Rsk2-deficient bones suggested a state of amino acid sufficiency in mutant osteoblasts (Figure 5B). In agreement with this assumption, histomorphometric analyses showed that HPD, but not HFD, fully rescued the low bone formation and low bone mass otherwise observed in 1-month-old Atf4−/− and Rsk2−/− mice (Figures 5C and S5A) and normalized Type I collagen synthesis in Atf4−/− and Rsk2−/− but not in WT bones without affecting Type I collagen expression (Figure 5E). Return to normal diet following 1 month of HPD reversed the HPD-induced bone mass rescue observed in Atf4−/− mice (Figure 5C), indicating that nutritional manipulations of ATF4's function are effective postnatally as well.

Figure 5
High-protein diet corrects the bone abnormalities in Atf4−/− and Rsk2−/− mice

A surprising feature of Atf4- and Rsk2-deficient mice fed a HPD was the normalization of their osteoblasts number (Figure 5C). This suggested that this diet may have also rescued the delay in osteoblast differentiation characterizing these mutant embryos. To address this possibility, we analyzed Atf4−/− and Rsk2−/− embryos whose mothers were fed either a ND or a HPD at the time osteoblast differentiation occurs. Molecular analysis of E15.5 and E18.5 mutant embryos showed that feeding pregnant mothers with HPD but not HFD sufficed to correct the lack of bone trabeculae otherwise observed in Atf4−/− and Rsk2−/− embryos (Figure 5D and S5B). This rescue was specific of the RSK2-ATF4 pathway as the HPD did not correct the osteoblast differentiation defect observed in embryos lacking Runx2 or Osterix, two transcription factors not phosphorylated by RSK2 (Yang et al., 2004) (Figure 5D). The restoration of the expression of Bone sialoprotein (Bsp) and Osteocalcin, two target genes of ATF4, in E15.5 and E18.5 Atf4−/− embryos to levels seen in WT embryos verified that osteoblast differentiation has been rescued by the HPD (Figure 5F). Taken together, these results demonstrate that ATF4 regulation of osteoblast differentiation and bone formation during development and postnatally occurs mainly through its ability to favor amino acid import and thereby protein synthesis.


This study identifies ATF4 as a crucial transcriptional mediator of NF1 signaling in osteoblasts, the bone cells in which Nf1 is expressed at the highest level and identifies neurofibromin as a negative regulator of bone remodeling acting via distinct pathways on bone formation and bone resorption (Figure 6). Moreover, further use of Nf1ob−/− mice elicited by the knowledge surrounding ATF4 molecular mode of action showed that development of ATF4-dependent skeletal dysplasiae can be corrected by simple diet manipulations thus uncovering a novel mechanism regulating skeletal development and suggesting a simple treatment for a relatively frequent skeletal dysplasia.

Figure 6
Roles of neurofibromin in the regulation of bone remodeling

A mouse model of NF1 skeletal manifestations

Pleiotropic skeletal manifestations ranging from bowing of long bones to pseudoarthrosis are observed in a significant portion of NF1 patients. Remarkably, these manifestations often worsen with age (Ruggieri et al., 1999; Hermanns-Sachweh et al., 2005; Kuorilehto et al., 2004; Lammert et al., 2005). Little was known about the cellular and molecular mechanisms accounting for NF1 skeletal manifestations since no animal model has been available until now. Indeed, Nf1−/− mice die in utero before skeletogenesis is initiated and Nf1+/− mice, unlike human patients, do not have any overt skeletal manifestations (Yu et al., 2005). Thus, to elucidate the histological, cellular, and molecular bases of NF1 skeletal manifestations, we generated mice lacking NF1 only in the bone-specific cell in which Nf1 expression is the highest, namely the osteoblast. We show here that in mice, through expression in osteoblasts, Nf1 affects both bone formation and bone resorption. In addition, it also affects the overall degree of mineralization of bone ECM. These results define an animal model of NF1 skeletal manifestations that may contribute to elucidate the pleiotropic nature of NF1 skeletal manifestations in humans. They also uncovered a molecular mechanism of action for neurofibromin in osteoblasts.

ATF4, a trancriptional mediator of neurofibromin signaling in osteoblasts

The only molecular abnormalities observed in Nf1ob−/− mice and bones were an increase in collagen synthesis and in Osteocalcin expression. Those features were the mirror image of what was seen in mice lacking one of the major osteoblast-enriched transcription factor, ATF4 (Yang et al., 2004). Several lines of evidence of molecular, genetic or pharmacological nature establish that ATF4 is a mediator of neurofibromin signaling in osteoblasts. First, RSK2 phosphorylation is increased as is ATF4 phosphorylation on serine 251 in Nf1−/− osteoblasts, thus explaining the increase in collagen synthesis and in Osteocalcin expression; second, PKA activity is increased as is ATF4 phosphorylation on serine 254 and as a result, Rankl expression, osteoclast differentiation, and bone resorption in Nf1ob−/− mice. That neurofibromin can affect PKA–dependent signaling events has been shown in other cell types such as schwann cells (Kim et al., 2001; Xu et al., 2002). Third, transgenic mice overexpressing Atf4 in osteoblasts have a phenotype similar to the one of Nf1ob−/− mice; fourth and more importantly, removing one allele of Atf4 sufficed to correct, at least partially, Nf1ob−/− mice bone phenotypes while removing one allele of either Runx2 or Osterix could not; fifth, a low-protein diet corrected all RSK2-dependent abnormalities, i.e., the increase in bone formation and the defect in ECM mineralization in both Nf1ob −/− and α1(I) Collagen-Atf4 mice.

These findings, however, do not exclude that other transcription factors whose activity can be modulated by RSK2 and/or PKA may also contribute to Nf1ob−/− mice skeletal phenotypes. In that respect it is worthwhile to note that CREB phosphorylation was increased in Nf1-deficient osteoblasts (data not shown). The possible involvement of CREB (or of other transcription factors) in this signaling cascade will have to await the generation of mutant mice lacking CREB only in osteoblasts.

Rescue of ATF4-dependent skeletal dysplasia through diets in mice

ATF4 is a peculiar player in the field of transcriptional control of bone formation as it has at least two mechanisms of action. It regulates osteoblast differentiation and the expression of osteoblast-specific genes such as Osteocalcin, however, Osteocalcin deletion does not delay bone formation (Ducy et al., 1996; Yang et al., 2004). At the same time, it promotes amino acid import in osteoblasts by activating genes encoding membrane transporters (Harding et al., 2003). The observation that NF1 skeletal manifestations worsen overtime and reappear after surgical treatment in patients suggested to us that the former mode of action of ATF4 may explain this feature of NF skeletal manifestations. Consistent with this hypothesis, reducing protein intake normalized all skeletal manifestations related to the ability of ATF4 to favor amino acid import in cells of Nf1ob −/− mice. Conversely, increasing protein intake in a mouse model of Coffin-Lowry Syndrome rescued the skeletal manifestations observed in adult mice. Remarkably, these latter manipulations also rescued all skeletal developmental abnormalities in both Atf4−/− and Rsk2−/− embryos, thus revealing that food intake affects skeletal development or at least ATF4-dependent events in this process.

These findings raise several questions beyond NF1 skeletal manifestations per se; first, are other manifestations in NF1 secondary, at least in part, to an increase in amino acid import in other cell types? That Rsk2 and other members of this family of kinases as well as other members of the CREB family of transcription factors are expressed in neurons give some credence to this hypothesis that can easily be tested. Second, one can wonder whether development of other skeletal dysplasiae may not be affected by amino acid intake. Although this is a valid hypothesis, we should stress here that this notion applies only to ATF4-related diseases. More generally, our results illustrate how the precise knowledge of the molecular mechanism of action of a given gene involved in the pathogenesis of human genetic disease can potentially translate into therapeutic interventions.

Experimental procedures

Generation of mutant mice and animal treatments

The generation of 2.3kb α1 (I) collagen cre mice, Nf1 flox, Atf4−/−, Rsk2−/−, Runx2−/−, and Osx−/− mice have been described (Dacquin et al., 2002; Nakashima et al., 2002; Otto et al., 1997; Yang et al., 2004; Zhu et al., 2001). WT are cre(−); Nf1flox/Nf1flox, Nf1ob−/− are cre(+); Nf1flox/Nf1flox. Alpha(I) collagen-Atf4 construct was generated by cloning the Atf4 genomic sequence 3′ of the 2.3 a1(I) Collagen promoter and 5′ of a mp1 polyA signal in the pJ251 vector. All mice used in this study were backcrossed on a C57BL6 background. Four week-old Nf1ob−/− mice were administered PD198306 daily (20 mg/kg/day) for 5 weeks by oral gavage. Normal diet (Picolab-Roden 20% #5053, TEKLAD, Madison, WI) contained 23.6% protein, 64.5% carbohydrate and 11.9% fat. High-protein diet (TD94266, TEKLAD) contained 53.3% protein, 33.6% carbohydrate and 13.1% fat. High-fat diet (DI12331, Research Diets, New Brunswick, NJ) contained 16% protein, 26% carbohydrate, and 58% fat. Low-protein diet (TD92208, TEKLAD) contained 10.8% protein, 76.0% carbohydrate, and 13.2% fat. All animal protocols were approved by the Animal Care Committees of Baylor College of Medicine.

Morphometric measurements

Static and dynamic histomorphometry measurements were performed as previously described in accordance with standard nomenclature, using the Osteomeasure Analysis System (Osteometrics, Inc) (Ducy et al., 2000; Parfitt et al., 1987; Takeda et al., 2002). 3D microtomographic measurements were performed at 12microns resolution under the growth plate (Scanco Medical, Bassersdorf, Switzerland). Four to 9 animals were assigned per group.

Cell cultures

Calvaria osteoblasts were extracted by triple collagenase/trypsine digestion from 4 day-old CD1 pups and differentiated with ascorbic acid as previously described (Ducy et al., 2000). Cre-mediated recombination efficiency was performed with differentiated calvarial osteoblasts. PCR was performed as originally described {Zhu, 2001 #1650}. Osteoclast differentiation assays were performed using Percoll gradient-purified BMM isolated from WT and Nf1ob−/− bones, in the presence of M-CSF (30 ng/ml) and RANKL (50 ng/ml) for 4 days. Osteoblast/osteoclasts cocultures were performed in triplicate in 48-well plates using WT and Nf1−/− calvaria osteoblasts and BMM in the presence of 1,25(OH)2 vitamin D3 (10−8M, Calbiochem) for 4 days (Takahashi et al., 1988). In vitro TRAP staining was performed at day 4 using the Acid Phosphatase Leukocyte kit (Sigma). Only TRAP-positive multinucleated (n > 3 nuclei) were counted. Van Gieson and alkaline phosphatase staining was performed using established protocols using buffered 10% formalin-fixed osteoblasts.

Gene expression studies

Osteoblasts were treated in serum-free medium with vehicle, U0126 (30 μM, Promega) or H89 (10 μM, Sigma) for 24 hr. Total RNA were extracted with Trizol (Invitrogen). cDNA were generated using the ABI Reverse transcriptase system and random hexanucleotide primers. Real-time PCR was performed using Taqman ABI probes/primers on a 9700 ABI cycler and 18S rRNA expression as endogenous control. Northern blot analyses were performed using 15 μg of total RNA and 32P-labeled probes. PCR primers used for transgenic and endogenous Atf4 expression are available upon request. In situ hybridization analyses were performed using 4% PFA-fixed embryos embedded in paraffin, according to standard procedures (Ducy et al., 1997).

Biochemical studies

Osteoblasts were treated in serum-free medium with vehicle or U0126 (30 μM) for 6 hr. Lysates from primary osteoblasts or crushed frozen bones were prepared in RIPA buffer in the presence of protease and phosphatase inhibitors. Twenty to 60ug of proteins were separated by SDS-PAGE in reducing conditions and transferred on nitrocellulose membrane using standard protocols. Membranes were incubated with primary antibodies including total or anti-Phospho ATF4 (Elefteriou et al., 2005; Yang et al., 2004), total and anti-Phospho GCN2 (Harding et al., 2003), eIF2α (Cell Signaling), RSK2 (Dumont et al., 2005), CREB (Upstate), AKT, p38 (Santa Cruz) antibodies. Type I collagen from crushed frozen bones (20 mg) or calvaria primary osteoblast was extracted twice with pepsin (1 mg/ml) in 0.5 M acetic acid for 72 hr at 4°C. Collagen was precipitated by NaCl (0.7% final concentration) overnight and collected by centrifugation. Collagen pellets were resolubilized in 0.5M Tris-HCl (pH 7.8) and separated on a 7.5% SDS-PAGE in presence of 2 M urea, followed by Western blot analysis using anti-collagen type I antibody (Rockland, Gilbertsville, PA) (Bashey et al., 1989). Anti-actin antibody (Santa Cruz Biotechnology, Inc.) was used for loading. Protein kinase A assay (Stressgen) was performed in the presence of 10% FBS. Ras assay (Cell Signaling) was performed using differentiated WT and Nf1−/− osteoblasts following overnight starvation and 2 min stimulation with 10% FBS and according to manufacturer's recommendations. In short, cells were washed and lysed in the presence of protease and phosphatase inhibitors, supernatants were assessed for protein concentration and kinase reaction was performed in the presence of ATP. PKA activity values were corrected for protein concentration. All biochemical assays were performed using WT and Nf1−/− osteoblasts grown and treated simultaneously. Serum measurements for PTH (Immunotopics), calcium, phosphate, and alkaline phosphatase (Sigma) were performed according to the manufacturer recommendations. Deoxypyridinoline cross-links and Creatinine were measured using the Pyrilinks-D and creatinine kits (Metra Biosystems).

Statistical analyses

Data are expressed as mean ± SEM. Statistical significance was assessed by Student's test. Values were considered statistically significant when p < 0.05.

Supplementary Material

sup 1

sup 2

sup 3

sup 4

sup 5


We thank Dr. B. de Crombrugghe for the Osterix−/− mice, D.W. Owens (Pfizer Inc.) for PD198306, and Dr. A. Hanauer for the Phospho-Rsk2 antibody. This work was supported by grants from National Institutes of Health (G.K. and L.P.); The Children's Tumor Foundation (F.E.); the Japan Osteoporosis Foundation, The Uehara Memorial Foundation, Kanzawa Medical Research Foundation, and The Mochida Memorial Foundation for Medical and Pharmaceutical Research (H.S.).


Supplemental data: Supplemental Data include five figures and can be found with this article online at http://www.cellmetabolism.org/cgi/content/full/4/6/441/DC1/.


  • Alwan S, Tredwell SJ, Friedman JM. Is osseous dysplasia a primary feature of neurofibromatosis 1 (NF1)? Clin Genet. 2005;67:378–390. [PubMed]
  • Bashey RI, Cox R, McCann J, Jimenez SA. Changes in collagen biosynthesis, types, and mechanics of aorta in hypertensive rats. J Lab Clin Med. 1989;113:604–611. [PubMed]
  • Bi W, Huang WH, Whitworth DJ, Deng JM, Zhang Z, Behringer R, de Crombrugghe B. Haploinsufficiency of Sox9 results in defective cartilage primordia and premature skeletal mineralization. Proc Natl Acad Sci USA. 2001;98:6698–6703. [PMC free article] [PubMed]
  • Bialek P, Kern B, Yang X, Schrock M, Sosic D, Hong N, Wu H, Yu K, Ornitz DM, Olson EN, et al. A twist code determines the onset of osteoblast differentiation. Dev Cell. 2004;6:423–435. [PubMed]
  • Dacquin R, Starbuck M, Schinke T, Karsenty G. Mouse alpha1(I)-collagen promoter is the best known promoter to drive efficient Cre recombinase expression in osteoblast. Dev Dyn. 2002;224:245–251. [PubMed]
  • Dalby KN, Morrice N, Caudwell FB, Avruch J, Cohen P. Identification of regulatory phosphorylation sites in mitogen-activated protein kinase (MAPK)-activated protein kinase-1a/p90rsk that are inducible by MAPK. J Biol Chem. 1998;273:1496–1505. [PubMed]
  • Dasgupta B, Gutmann DH. Neurofibromatosis 1: closing the GAP between mice and men. Curr Opin Genet Dev. 2003;13:20–27. [PubMed]
  • Dong J, Qiu H, Garcia-Barrio M, Anderson J, Hinnebusch AG. Uncharged tRNA activates GCN2 by displacing the protein kinase moiety from a bipartite tRNA-binding domain. Mol Cell. 2000;6:269–279. [PubMed]
  • Ducy P, Amling M, Takeda S, Priemel M, Schilling AF, Beil T, Shen J, Vinson C, Rueger JM, Karsenty G. Leptin inhibits bone formation through a hypothalamic relay: A central control of bone mass. Cell. 2000;100:197–207. [PubMed]
  • Ducy P, Desbois C, Boyce B, Pinero G, Story B, Dunstan C, Smith E, Bonadio J, Goldstein S, Gundberg C, et al. Increased bone formation in osteocalcin-deficient mice. Nature. 1996;382:448–452. [PubMed]
  • Ducy P, Zhang R, Geoffroy V, Ridall AL, Karsenty G. Osf2/Cbfa1: a transcriptional activator of osteoblast differentiation. Cell. 1997;89:747–754. [PubMed]
  • Dumont J, Umbhauer M, Rassinier P, Hanauer A, Verlhac MH. p90Rsk is not involved in cytostatic factor arrest in mouse oocytes. J Cell Biol. 2005;169:227–231. [PMC free article] [PubMed]
  • Eicher EM, Southard JL, Scriver CR, Glorieux FH. Hypophosphatemia: mouse model for human familial hypophosphatemic (vitamin D-resistant) rickets. Proc Natl Acad Sci USA. 1976;73:4667–4671. [PMC free article] [PubMed]
  • Elefteriou F, Ahn JD, Takeda S, Starbuck M, Yang X, Liu X, Kondo H, Richards WG, Bannon TW, Noda M, et al. Leptin regulation of bone resorption by the sympathetic nervous system and CART. Nature. 2005;434:514–520. [PubMed]
  • Fisher TL, Blenis J. Evidence for two catalytically active kinase domains in pp90rsk. Mol Cell Biol. 1996;16:1212–1219. [PMC free article] [PubMed]
  • Harding HP, Novoa I, Zhang Y, Zeng H, Wek R, Schapira M, Ron D. Regulated translation initiation controls stress-induced gene expression in mammalian cells. Mol Cell. 2000;6:1099–1108. [PubMed]
  • Harding HP, Zhang Y, Zeng H, Novoa I, Lu PD, Calfon M, Sadri N, Yun C, Popko B, Paules R, et al. An integrated stress response regulates amino acid metabolism and resistance to oxidative stress. Mol Cell. 2003;11:619–633. [PubMed]
  • Hermanns-Sachweh B, Senderek J, Alfer J, Klosterhalfen B, Buttner R, Fuzesi L, Weber M. Vascular changes in the periosteum of congenital pseudarthrosis of the tibia. Pathol Res Pract. 2005;201:305–312. [PubMed]
  • Hunter AG. Coffin-Lowry syndrome: a 20-year follow-up and review of long-term outcomes. Am J Med Genet. 2002;111:345–355. [PubMed]
  • Jacks T, Shih TS, Schmitt EM, Bronson RT, Bernards A, Weinberg RA. Tumour predisposition in mice heterozygous for a targeted mutation in Nf1. Nat Genet. 1994;7:353–361. [PubMed]
  • Jacquemin C, Bosley TM, Svedberg H. Orbit deformities in craniofacial neurofibromatosis type 1. AJNR Am J Neuroradiol. 2003;24:1678–1682. [PubMed]
  • Kim HA, Ratner N, Roberts TM, Stiles CD. Schwann cell proliferative responses to cAMP and Nf1 are mediated by cyclin D1. J Neurosci. 2001;21:1110–1116. [PubMed]
  • Klose A, Ahmadian MR, Schuelke M, Scheffzek K, Hoffmeyer S, Gewies A, Schmitz F, Kaufmann D, Peters H, Wittinghofer A, Nurnberg P. Selective disactivation of neurofibromin GAP activity in neurofibromatosis type 1. Hum Mol Genet. 1998;7:1261–1268. [PubMed]
  • Komori T, Yagi H, Nomura S, Yamaguchi A, Sasaki K, Deguchi K, Shimizu Y, Bronson RT, Gao YH, Inada M, et al. Targeted disruption of Cbfa1 results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell. 1997;89:755–764. [PubMed]
  • Kuorilehto T, Nissinen M, Koivunen J, Benson MD, Peltonen J. NF1 tumor suppressor protein and mRNA in skeletal tissues of developing and adult normal mouse and NF1-deficient embryos. J Bone Miner Res. 2004;19:983–989. [PubMed]
  • Kuorilehto T, Poyhonen M, Bloigu R, Heikkinen J, Vaananen K, Peltonen J. Decreased bone mineral density and content in neurofibromatosis type 1: lowest local values are located in the load-carrying parts of the body. Osteoporos Int. 2005;16:928–936. [PubMed]
  • Kwok ES, Sawatzky B, Birch P, Friedman JM, Tredwell SJ. Vertebral scalloping in neurofibromatosis type 1: a quantitative approach. Can J Surg. 2002;45:181–184. [PMC free article] [PubMed]
  • Lammert M, Kappler M, Mautner VF, Lammert K, Storkel S, Friedman JM, Atkins D. Decreased bone mineral density in patients with neurofibromatosis 1. Osteoporos Int. 2005;16:1161–1166. [PubMed]
  • Murshed M, Harmey D, Millan JL, McKee MD, Karsenty G. Unique coexpression in osteoblasts of broadly expressed genes accounts for the spatial restriction of ECM mineralization to bone. Genes Dev. 2005;19:1093–1104. [PMC free article] [PubMed]
  • Murshed M, Schinke T, McKee MD, Karsenty G. Extracellular matrix mineralization is regulated locally; different roles of two gla-containing proteins. J Cell Biol. 2004;165:625–630. [PMC free article] [PubMed]
  • Nakashima K, Zhou X, Kunkel G, Zhang Z, Deng JM, Behringer R, de Crombrugghe B. The novel zinc-finger-containing transcription factor Osterix is required for osteoblast differentiation and bone formation. Cell. 2002;108:17–29. [PubMed]
  • Olsen BJ, Reginato AM, Wang W. Bone Development. Annu Rev Cell Dev Biol. 2000;16:191–220. [PubMed]
  • Otto F, Thornell AP, Crompton T, Denzel A, Gilmour KC, Rosewell IR, Stamp GW, Beddington RS, Mundlos S, Olsen BR, et al. Cbfa1, a candidate gene for cleidocranial dysplasia, is essential for osteoblast differentiation and bone development. Cell. 1997;89:765–771. [PubMed]
  • Parada LF. Neurofibromatosis type 1. Biochim Biophys Acta. 2000;1471:M13–M19. [PubMed]
  • Parfitt AM, Drezner MK, Glorieux FH, Kanis HA, Malluche H, Meunier PJ, Ott SM, Recker RR. Bone histomorphometry: standardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphometry Committee. J Bone Miner Res. 1987;6:595–610. [PubMed]
  • Ruggieri M, Pavone V, De Luca D, Franzo A, Tine A, Pavone L. Congenital bone malformations in patients with neurofibromatosis type 1 (Nf1) J Pediatr Orthop. 1999;19:301–305. [PubMed]
  • Satokata I, Maas R. Msx1 deficient mice exhibit cleft palate and abnormalities of craniofacial and tooth development. Nat Genet. 1994;6:348–355. [PubMed]
  • Sigillo R, Rivera H, Nikitakis NG, Sauk JJ. Neurofibromatosis type 1: a clinicopathological study of the orofacial manifestations in 6 pediatric patients. Pediatr Dent. 2002;24:575–580. [PubMed]
  • Simsek E, Yavuz C, Ustundag N. Atypical meningioma and extensive calvarium defects in neurofibromatosis type 1. Pediatr Radiol. 2003;33:551–553. [PubMed]
  • Stein GS, Lian JB, Stein JL, Van Wijnen AJ, Montecino M. Transcriptional control of osteoblast growth and differentiation. Physiol Rev. 1996;76:593–629. [PubMed]
  • Stevenson DA, Birch PH, Friedman JM, Viskochil DH, Balestrazzi P, Boni S, Buske A, Korf BR, Niimura M, Pivnick EK, et al. Descriptive analysis of tibial pseudarthrosis in patients with neurofibromatosis 1. Am J Med Genet. 1999;84:413–419. [PubMed]
  • Stevenson DA, Pasquali M, Moyer-Mileur L, Hoff CJ, Schwarz EL, Hartford D, Viskochil DH. Abstract book of the Children's Tumor Foundation International Consortium for the Molecular Biology of NF1, NF2 and Schwannomatosis. 2005. Evidence of increased bone resorption in neurofibromatosis type 1 using urinary pyridinium crosslink analysis.
  • Takahashi N, Akatsu T, Udagawa N, Sasaki T, Yamaguchi A, Moseley JM, Martin TJ, Suda T. Osteoblastic cells are involved in osteoclast formation. Endocrinology. 1988;123:2600–2602. [PubMed]
  • Takeda S, Elefteriou F, Levasseur R, Liu X, Zhao L, Parker KL, Armstrong D, Ducy P, Karsenty G. Leptin regulates bone formation via the sympathetic nervous system. Cell. 2002;111:305–317. [PubMed]
  • Tanaka T, Tsujimura T, Takeda K, Sugihara A, Maekawa A, Terada N, Yoshida N, Akira S. Targeted disruption of ATF4 discloses its essential role in the formation of eye lens fibres. Genes Cells. 1998;3:801–810. [PubMed]
  • Touraine RL, Zeniou M, Hanauer A. A syndromic form of X-linked mental retardation: the Coffin-Lowry syndrome. Eur J Pediatr. 2002;161:179–187. [PubMed]
  • Wagner EF, Karsenty G. Genetic control of skeletal development. Curr Opin Genet Dev. 2001;11:527–532. [PubMed]
  • Wagner T, Wirth J, Meyer J, Zabel B, Held M, Zimmer J, Pasantes J, Bricarelli FD, Keutel J, Hustert E, et al. Autosomal sex reversal and campomelic dysplasia are caused by mutations in and around the SRY-related gene SOX9. Cell. 1994;79:1111–1120. [PubMed]
  • Xu Y, Chiamvimonvat N, Vazquez AE, Akunuru S, Ratner N, Yamoah EN. Gene-targeted deletion of neurofibromin enhances the expression of a transient outward K+ current in Schwann cells: a protein kinase A-mediated mechanism. J Neurosci. 2002;22:9194–9202. [PubMed]
  • Yang X, Matsuda K, Bialek P, Jacquot S, Masuoka HC, Schinke T, Li L, Brancorsini S, Sassone-Corsi P, Townes TM, et al. ATF4 is a substrate of RSK2 and an essential regulator of osteoblast biology; implication for Coffin-Lowry Syndrome. Cell. 2004;117:387–398. [PubMed]
  • Yu X, Chen S, Potter OL, Murthy SM, Li J, Pulcini JM, Ohashi N, Winata T, Everett ET, Ingram D, et al. Neurofibromin and its inactivation of Ras are prerequisites for osteoblast functioning. Bone. 2005;36:793–802. [PubMed]
  • Zhu Y, Romero MI, Ghosh P, Ye Z, Charnay P, Rushing EJ, Marth JD, Parada LF. Ablation of NF1 function in neurons induces abnormal development of cerebral cortex and reactive gliosis in the brain. Genes Dev. 2001;15:859–876. [PMC free article] [PubMed]
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Gene
    Gene records that cite the current articles. Citations in Gene are added manually by NCBI or imported from outside public resources.
  • GEO Profiles
    GEO Profiles
    Gene Expression Omnibus (GEO) Profiles of molecular abundance data. The current articles are references on the Gene record associated with the GEO profile.
  • HomoloGene
    HomoloGene clusters of homologous genes and sequences that cite the current articles. These are references on the Gene and sequence records in the HomoloGene entry.
  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem chemical substance records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records.

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...