Cell growth, proliferation, differentiation and survival are influenced by the availability of oxygen. The effect of hypoxia on embryonic cells and the underlying molecular mechanisms to maintain cellular viability are still poorly understood. In this study, we show that hypoxia during Xenopus embryogenesis rapidly leads to a significant developmental delay and to cell apoptosis after prolonged exposure. We provide strong evidence that hypoxia does not affect somitogenesis but affects the number of mitotic cells and muscle-specific protein accumulation in somites, without interfering with the expression of MyoD and MRF4 transcription factors. We also demonstrate that hypoxia reversibly decreases Akt phosphorylation and increases the total amount of the translational repressor 4E-BP, in combination with an increase of the 4E-BP associated with eIF4E. Interestingly, the inhibition of PI3-kinase or mTOR, with LY29002 or rapamycin, respectively, triggers the 4E-BP accumulation in Xenopus embryos. Finally, the overexpression of the non-phosphorylatable 4E-BP protein induces, similar to hypoxia, a decrease in mitotic cells and a decrease in muscle-specific protein accumulation in somites. Taken together, our studies suggest that 4E-BP plays a central role under hypoxia in promoting the cap-independent translation at the expense of cap-dependent translation and triggers specific defects in muscle development.
Oxygen (O2) is an environmental and developmental signal regulator involved in several events such as energy homeostasis, development and progenitor cell differentiation. Hypoxia can arise from physiological events (e.g. exercise, travel at high altitude) as well as from pathophysiological conditions (e.g. tissue ischemia, inflammation, solid tumours) (Papandreou et al., 2005; Wouters et al., 2005). Understanding cellular responses to hypoxia is of great importance since hypoxia impacts normal development and the malignant progression of a solid tumour (Rivard, 1986; Wouters et al., 2005; Sharma et al., 2006; Simon and Keith, 2008; Wouters and Koritzinsky, 2008; Dunwoodie, 2009; Rodricks et al., 2010).
The cellular response to hypoxia depends on the severity of the hypoxic insult and may result in complex gene expression changes to maintain cellular viability through transcriptional and post-transcriptional events. This response is largely driven by a transcriptional program initiated by stabilization of hypoxia-inducible factors (HIF) (Guillemin and Krasnow, 1997; Semenza, 2003; Cummins and Taylor, 2005; Wouters et al., 2005). HIF-1 and HIF-2 promote transcription of more than 60 putative downstream genes that affect hypoxia tolerance, energy homeostasis, angiogenesis and tumour growth (Semenza, 2003).
Besides HIF stabilization, the PI3K/Akt signalling pathway is an important contributor to hypoxia tolerance (Mazure et al., 1997; Barry et al., 2007). Akt is a serine/threonine kinase that resides downstream of PI3K and phosphorylates a variety of substrates (Kandel and Hay, 1999; Brazil and Hemmings, 2001). Akt phosphorylates and inhibits the tuberous sclerosis complex (TSC2), a negative regulator of the mammalian target of rapamycin (mTOR). Consequently mTOR is activated and stimulates the cap-dependent translation (Hay and Sonenberg, 2004).
Cap-dependent translation requires the assembly of an active eIF4F (eukaryotic initiation factor 4F) complex at the m7GpppN cap structure present at the 5′ terminus of the mRNA (Hay and Sonenberg, 2004; Kapp and Lorsch, 2004; Holcik and Sonenberg, 2005; Sonenberg and Hinnebusch, 2009). The eIF4F complex consists of the cap-binding protein, eIF4E, a scaffolding protein, eIF4G and an ATP-dependent helicase, eIF4A. Formation of eIF4F facilitates the recruitment of the pre-initiation complex, 43S, which is necessary to initiate translation. The assembly of the cap-binding complex eIF4F is a key step of the translational regulation and is regulated through a set of eIF4E-binding proteins (4E-BPs) controlled by mTOR-mediated phosphorylation (Hay and Sonenberg, 2004; Sonenberg and Hinnebusch, 2009). 4E-BPs reversibly bind to eIF4E in their hypophosphorylated form obstructing the interaction between eIF4E and eIF4G and consequently preventing the formation of eIF4F complex (Pause et al., 1994; Haghighat et al., 1995). Interestingly, several cellular mRNAs use an alternative mechanism of translation initiation, involving an internal ribosomal entry site (IRES) at the 5′ untranslated region and are translated efficiently when cap-dependent translation is impaired under a variety of stress conditions including serum deprivation, irradiation, apoptosis and hypoxia (Holcik et al., 2000).
Hypoxia results in a rapid inhibition of protein synthesis through the repression of the initiation step of mRNA translation (Wouters et al., 2005). This can be regulated by a hypoxia dependent inhibition of eIF4F as evidenced by a reduction of eIF4E/eIF4G association and a corresponding increase of eIF4E/4E-BP (Koritzinsky et al., 2006). In the same way, hypoxia rapidly and reversibly triggers hypophosphorylation of mTOR and its effector 4E-BP in human embryonic kidney cells (Arsham et al., 2003). In addition, hypoxia rapidly induces a reversible increase in 4E-BP protein levels in sea urchin embryos (Le Bouffant et al., 2006) and FoxO activates 4E-BP mRNA transcription in Drosophila embryos undergoing oxidative stress or after nutrient starvation (Tettweiler et al., 2005). While the 4E-BP translational inhibitory functions and mechanisms involved in its regulation have been extensively studied in vitro, the biological role of 4E-BP in developing organisms is unclear to date. Few studies have been performed regarding the effects of hypoxia on embryonic development, even though O2 has been involved in the development of several pathologies of human pregnancy (Unger et al., 1988; Julian, 2011). Xenopus laevis embryos contain a 4E-BP protein (Fraser et al., 1999) but little is known about its role during embryonic development, particularly during hypoxic stress.
Our objective was to determine the effects of hypoxia on vertebrate embryonic development and muscle formation and to examine the involvement of 4E-BP in mediating these hypoxic effects. Here, we show that hypoxia causes oxygen-dependent growth retardation, apoptosis and lethality after prolonged exposure. Exposure to hypoxic conditions also decreases Akt phosphorylation and induces a reversible increase in 4E-BP downstream protein amount that associates with eIF4E in hypoxia. Moreover, analysis of muscle development provides strong evidence that hypoxia does not affect somitogenesis but affects the accumulation of muscle-specific proteins in somites, independently of mRNA variation. Finally, our data show for the first time that overexpression of the non-phosphorylatable 4E-BP protein causes effects similar to hypoxia on somite cells. Taken together, our studies suggest that 4E-BP plays a key role in mediating the hypoxia-induced defects in muscle formation, without affecting the expression of myogenic transcription factors.
Hypoxia causes growth retardation, lethality and apoptosis
To investigate the effects of hypoxia on early development, we exposed Xenopus embryos to several levels of oxygen from the end of the gastrulation (stage 13) to late tailbud stages (stage 42). Control embryos were exposed to 21% O2 corresponding to normoxia and hypoxic embryos were exposed to 10% or 5% O2 that represent moderate and strong hypoxic stress, respectively. These conditions were validated by measurements of dissolved oxygen level and by quantitative real-time PCR (qRT-PCR) to determine HIF-1α transcription level which is tightly controlled by cellular oxygen tension (Gorlach, 2009) and its known VEGF target gene. The oxygen partial pressure in the medium was proportional to the oxygen level in the hypoxic chamber and HIF-1α and VEGF mRNA expression increased significantly in relation to oxygen availability (supplementary material Fig. S1). Moreover, we performed western blot analyses with an antibody against the phospho-p38-MAPK, a member of the class of mitogen-activated protein kinases that are responsive to stress stimuli. As shown in Fig. 1A,B, phospho-p38-MAPK was undetectable at stage 21 and was detected at very low level at stage 25 in control embryos. When embryos were subjected to 10% O2, levels of phospho-p38-MAPK were not significantly higher as compared to the control at any stage (Fig. 1A,B). However, levels were dramatically higher at stages 21 and 25 in embryos exposed to 5% O2 when compared to the control embryos and to embryos subjected to 10% O2 (Fig. 1A,B). These results show that exposure to 5% O2 causes a strong cellular stress response during embryonic development whereas 10% O2 seems to have less of an effect.
External embryo morphology was observed and compared to Nieuwkoop and Faber table stage illustrations (Nieuwkoop and Faber, 1994). As shown in Fig. 1C,D, both moderate and strong hypoxia slowed down embryonic development rate. Hypoxia for 42 h caused a significant retardation in Xenopus development without affecting external embryo morphology (Fig. 1D). When normoxic embryos have reached stage 30, embryos exposed to 10% O2 have reached stage 28 and embryos exposed to 5% O2 have reached stage 25. Moreover, hypoxic embryos maintain their delayed growth through development (Fig. 1C).
Depigmentation of embryos and dissociation of tissues were used as indicators to assess hypoxia-related lethality. The number of dead embryos was counted (Fig. 1E). Xenopus embryos exposed for 24 h to 10% O2 had a survival rate of 77% while those exposed for 24 h to 5% O2 had a stark decrease in survival represented by a 48% survival rate. Since hypoxia causes significant embryonic death depending on oxygen availability, the apoptosis patterns in relation to the oxygen concentration was analysed. Embryos (n = 15) were fixed at stage 25 when the cellular stress marker phospho-p38-MAPK level was higher and sections were treated with TUNEL. TUNEL-positive cells were counted on frontal sections of control and hypoxic embryos, and the average of positive cells per embryo is presented in Fig. 1F. Apoptosis was 77% higher in embryos exposed to strong hypoxia (5% O2) compared to the control. In embryos exposed to 10% O2 the number of apoptotic cells was 43% higher compared to the control embryos suggesting that hypoxia causes apoptosis depending on the oxygen level.
Strikingly, embryos exposed to hypoxia and especially to 5% O2 did not make any movement when pricked with fine pliers compared to control embryos which ran off (data not shown). This observation suggests that hypoxia caused strong motility defects probably by affecting muscles. To further investigate the effect of hypoxia on muscle development, experiments were performed on embryos from stage 21 to stage 25, after the formation of the first somites (Nieuwkoop and Faber, 1994) and before death of the majority of embryos exposed to strong hypoxia.
Hypoxia decreases Akt phosphorylation in Xenopus embryos
The serine/threonine protein kinase Akt plays a critical role in metabolism, cell survival, mRNA translation, proliferation and cell cycle regulation. In addition, Akt signalling pathway has been shown to regulate myoblast proliferation and differentiation (Bodine et al., 2001; Wilson and Rotwein, 2007; Wu et al., 2011). Based on these data, the Akt signalling activation was analysed by western blotting using two anti-phospho-Akt antibodies (Fig. 2).
Akt was phosphorylated on Thr308 and Ser473 in control embryos suggesting an activation of Akt protein during the normal development. Interestingly, in embryos exposed to hypoxia, the level of phospho-Akt decreased in an oxygen concentration-dependent manner at stages 21 and 25. The phosphorylation status of Akt protein was significantly weaker under strong hypoxia (5% O2) in both stages compared to control and to 10% O2-exposed embryos, although each batch present variability and in some cases hypoxic embryos already presented phospho-Akt at stage 25 (Fig. 2). Therefore, the Akt signalling protein is inhibited in hypoxia during early embryonic development particularly in 5% O2 context. Otherwise, the Akt phosphorylation level was more affected at stage 21 than stage 25.
Hypoxia increases the amount of 4E-BP protein, which binds to eIF4E
The translational repressor 4E-BP is known to be a downstream substrate of the Akt signalling pathway (Gingras et al., 1998). Akt mediated phosphorylation of the translational repressor 4E-BP can subsequently lead to inactivation or degradation of 4E-BP (Elia et al., 2008). We then performed western blot analyses to determine 4E-BP protein levels in embryos exposed to hypoxia. As shown in Fig. 3A,B, 4E-BP protein levels were significantly higher at stages 21 and 25 in hypoxia-exposed embryos compared to the controls maintained under normoxic conditions. Moreover, 4E-BP protein levels were significantly higher in embryos subjected to 5% O2 than to 10% O2, suggesting that hypoxia-induced overexpression of 4E-BP depends on oxygen availability. To know if the increase of 4E-BP protein was due to transcriptional regulation, we performed qRT-PCR. As shown in Fig. 3D, the relative 4E-BP mRNA level did not change significantly under hypoxia.
In mammals, 4E-BP has been shown to bind to eIF4E in its hypophosphorylated form. Therefore, decrease of Akt activity induced by hypoxia should consequently increase the amount of 4E-BP associated with eIF4E. To validate this hypothesis, we used m7GTP affinity beads to test if overexpressed 4E-BP is able to bind its target, eIF4E. Protein extracts were incubated with m7GTP-conjugated Sepharose beads and bound proteins were analysed by western blotting with antibodies against 4E-BP and eIF4E. After m7GTP purification, western blot analysis revealed that the amount of 4E-BP associated with eIF4E increased significantly in embryos exposed to 5% and 10% O2 (Fig. 3C). Taken together, these data demonstrate that hypoxia reduces Akt phosphorylation level and leads to 4E-BP accumulation, which is able to bind to eIF4E.
It was then important to show links between 4E-BP amount regulation and Akt signalling pathway activity in Xenopus embryos. Indeed, if an active Akt signalling pathway maintains a low level of 4E-BP, it is then conceivable that embryos treatment with LY294002, a well-established inhibitor of PI3-kinase acting upstream of Akt, will modify the amount of 4E-BP protein in normoxic condition. To test this hypothesis, 4E-BP protein amount was analysed by western blotting in control and treated embryos. As expected, a 2-hour treatment with 50 µM of LY294002 affected the Akt phosphorylation state and triggered an increase in the amount of 4E-BP protein (Fig. 3E). In addition, since mTOR acts downstream Akt, we then tested the effect of rapamycin, a mTOR activity inhibitor. Interestingly, this drug induced a dose-dependent accumulation of 4E-BP in embryos treated for 3 h (Fig. 3F).
Taken together, these data suggest that the 4E-BP protein level is dependent of the PI3K/Akt/mTOR signalling pathway in Xenopus embryos.
Effects of hypoxia exposure are reversible upon reoxygenation
We investigated whether hypoxic effects are reversible by exposing embryos to 5% O2 until stage 21 and then subjecting a subset of embryos to normoxia (i.e. reoxygenation; Fig. 4A). Protein levels of 4E-BP as well as the phosphorylation of p38-MAPK and Akt were analysed by western blotting. As shown in Fig. 4B, phospho-p38-MAPK level decreased in reoxygenated embryos showing that the cellular stress was abolished. Moreover, embryonic death was prevented in the subset of embryos subsequently exposed to normoxia (9%) compared to the subset of embryos kept under hypoxic condition (54%). Interestingly, both phospho-Akt and 4E-BP protein levels returned to basal levels observed in controls (normoxia) suggesting that this signalling pathway was reactivated.
We have shown that the reoxygenation of the embryos triggers a decrease of the 4E-BP protein level and that 4E-BP amount is under a rapamycin-sensitive pathway. We then hypothesized that the decrease of 4E-BP amount triggered by reoxygenation of the embryos should be altered after rapamycin treatment. To test this idea, embryos were exposed to 5% O2 until stage 21 and shifted to normoxia in absence or presence of rapamycin. After 15, 30 and 45 min of reoxygenation, the level of 4E-BP was analysed by western blotting. As shown in Fig. 4C, 4E-BP amount was more important in reoxygenated embryos cultured for 30 and 45 min in presence of rapamycin compared to those cultured in absence of the drug. However, 4E-BP amount decreased in both cases suggesting that this time of exposure to rapamycin (less than 1 hour) seems to be not sufficient to induce the maximal effect of this drug and to maintain a high level of 4E-BP. In agreement with this, we show that 3 hours treatment with rapamycin was required to induce an increase of the 4E-BP amount in normoxia (supplementary material Fig. S2).
Altogether, these results show that effects of hypoxia are reversible when oxygen is available, suggesting a correlation between oxygen concentration and embryonic survival as well as a correlation between oxygen concentration, the PI3K/Akt/mTOR signalling pathway and the stability of 4E-BP.
Hypoxia specifically decreases the number of mitotic cell in somites
Our data show that hypoxia decreased Akt activation and increased the levels of translational repressor 4E-BP. Since Akt have been shown to be involved in cell proliferation and that control of polypeptide synthesis plays an important role in this mechanism, we used phosphorylated histone H3 (phospho-H3) as a mitosis marker. Immunodetection was performed with the anti-phospho-H3 antibody on cryostat sections of stage 25 embryos (n = 8). The number of phospho-H3-positive cells was determined and the average of positive cells per tissues is presented in Fig. 5. The number of mitotic cells in the neural tube, the notochord and the epidermis was not significantly affected in strong hypoxia. However, the number of positive cells was strongly reduced (90%) in somites of embryos exposed to 5% O2 as compared to control.
These results suggest that hypoxia specifically induces a decrease in the number of mitotic cell in somites, suggesting a decrease in somite cell proliferation.
Hypoxia does not affect somitogenesis but affects the accumulation of specific proteins in muscle cells
In vertebrate embryos, somites are regular transient structures repeated along the anterior/posterior axis of the embryo, which then differentiate into a part of the dermis, bone, cartilage and skeletal muscles. During somitogenesis, somitic cells are stacked in blocks and undergo a 90° rotation relative to the anteroposterior axis of the embryo to form myotome fibres that are aligned parallel to the notochord (Keller, 2000). Movements of nuclei were followed during progressive alignment and, at the end of rotation, nuclei are arranged in regularly ordered stripes so that one stripe corresponds to one somite.
In this study, we previously observed that embryos exposed to hypoxia did not make any movement compared to control embryos, suggesting muscle defects. In addition, we showed that hypoxia specifically decreases mitotic cell in somites and negatively regulates the serine/threonine protein kinase Akt that has been involved in muscle cell differentiation (Bodine et al., 2001; Wilson and Rotwein, 2007; Wu et al., 2011). To further investigate the effects of hypoxia on somitogenesis and muscle formation, we sectioned embryos and analysed the morphology of somites by immunostaining with the F59 antibody against MHC (myosin heavy chain) protein and myotome/skeletal muscle-specific monoclonal 12/101 antibody (Kintner and Brockes, 1984). On frontal sections, blocks of somitic cells were discernible, any changes in somite number was observed and segmentation of the presomitic mesoderm did not appear to be affected in hypoxia (Fig. 6A–D). Interestingly, we noticed that both the F59 (Fig. 6A,B) and the 12/101 (Fig. 6C,D) antibody stainings were dramatically weaker in the embryos exposed to hypoxia as compared to the control. These data were confirmed by western blot analysis showing that the 12/101 protein level was weaker in embryos exposed to 10% and 5% O2 as compared to the control (Fig. 6E).
To investigate whether hypoxic effects are reversible, embryos were exposed to 5% O2 until stage 21 and then a subset of them was subjected to normoxia. As shown in supplementary material Fig. S3, the 12/101 staining was higher in reoxygenated embryos compared to hypoxic embryos, indicating that some hours of reoxygenation partially reversed the phenotype in muscles.
Since the expression of MHC protein and the 12/101 muscle marker were decreased in hypoxic embryos, it is possible that hypoxia affects the somite fate by inhibiting the expression of genes coding for transcription factors involved in muscle cell differentiation. Therefore, we investigated whether hypoxia interfered with the myogenic signalling pathway by analysing the expression patterns of XMyoD and XMRF4, early and late markers of muscle cell differentiation, respectively. The expression of these specific proteins was analysed by immunostaining of frontal sections. Comparison between hypoxic and control embryos did not reveal any significant differences in the expression of both proteins. Indeed, anti-XMyoD and anti-XMRF4 antibodies revealed identical nuclear stainings in both normoxic (21% O2) and hypoxic (5% O2) embryos at stage 25 (Fig. 6F–Q).
Interestingly, the decrease of MHC protein amount induced by hypoxia was independent of Myh1, 4, 8 and 6 mRNAs variation (encoding for the MHC protein recognized by the F59 antibody) as shown by qRT-PCR analysis (Fig. 6R; supplementary material Fig. S4). Moreover and as expected, qRT-PCR analysis also showed no significant variations of MyoD mRNA level for stage 25 under hypoxic conditions (Fig. 6S).
In the epidermis of hypoxic embryos, the expression of p63 transcription factor, one of the earliest markers of epidermis differentiation, as well as the acetylated tubulin protein, a late marker of ciliated cells differentiation, were not affected (Fig. 6T–W).
Together, these results show that hypoxia does not affect somitogenesis and the expression of at least two transcription factors involved in muscle cell differentiation. However, hypoxia affects the accumulation of muscle-specific proteins in somites without affecting stability or transcription of mRNAs.
Overexpression of a non-phosphorylatable 4E-BP protein affects the accumulation of muscle-specific proteins in somites
The results above show that hypoxia increases functional 4E-BP protein levels, as indicated by eIF4E protein binding. In addition, hypoxia affects the accumulation of muscle-specific proteins. We then hypothesized that maintained eIF4E sequestration by 4E-BP is involved in the muscle cell protein accumulation. To validate this possibility, both blastomeres of two-cell stage embryos were injected with mRNA encoding for full-length 4E-BP protein (WT-4E-BP) and western blot analysis was used to establish 4E-BP protein levels at stage 25. Unexpectedly, in this condition, only a very small amount of WT-4E-BP was detectable (Fig. 7A), suggesting that the overexpressed protein was rapidly degraded. Since it has been recently suggested that phosphorylation of 4E-BP may promote its degradation (Elia et al., 2008), we hypothesized that the injection of mRNA encoding for a non-phosphorylatable 4E-BP protein should maintain high overexpression of 4E-BP protein and efficiently sequester eIF4E. Most of the data available to date indicates that mTOR is the main kinase of 4E-BP (Brunn et al., 1997; Burnett et al., 1998). Therefore, both blastomeres of two-cell stage embryos were then injected with mRNA encoding for a 4E-BP mutated protein (4A-4E-BP), in which all the four mTOR regulated residues in 4E-BP were mutated to alanine (reviewed in Martineau et al., 2012). Abolishing the possibility of phosphorylation on these four serine/threonine-to-alanine substitution has been shown to render 4E-BP constitutively active in sequestrating eIF4E (Oulhen et al., 2009). As expected, 4A-4E-BP mRNA injection induced a significant increase in 4E-BP protein levels without exposure to hypoxia (Fig. 7A). Interestingly, overexpression of 4E-BP did not significantly affect developmental stage, apoptosis and embryo survival (data not shown).
We then checked that the 4A-4E-BP mutated protein associated efficiently with the endogenous eIF4E using m7GTP affinity beads. Protein extracts were incubated with m7GTP-conjugated Sepharose beads and bound proteins were analysed by western blotting with antibodies against 4E-BP and eIF4E. As shown in Fig. 7B, WT-4E-BP and 4A-4E-BP proteins bind eIF4E, indicating that the mutation of the phosphorylation sites of 4E-BP did not alter its functionality other than stabilizing the protein.
To investigate the effects of overexpressed 4E-BP protein in somites, embryos were unilaterally injected at the four-cell stage into the presumptive somite region with mRNA encoding for 4A-4E-BP or WT-4E-BP as a control. Immunostaining of sectioned embryos revealed that the 4A-4E-BP protein was unilaterally detected in somites (Fig. 7C). Both the 4A-4E-BP- and WT-4E-BP-mRNA-injected embryos developed normally according to the external morphology (Fig. 7D–E). However, Fig. 7F shows a significant decrease in the number of mitotic cells in somites on the 4A-4E-BP-mRNA-injected side compared to the control side of embryos, suggesting that overexpression of non-phosphorylatable 4E-BP protein induces a decrease in somite cell proliferation. Immunohistochemical analyses with 12/101 antibody and Hoechst were performed. Interestingly, the 12/101 staining was dramatically weaker in the 4A-4E-BP-mRNA-injected side compared to the control side as well as to the WT-4E-BP-mRNA-injected embryos (Fig. 7G,H) while qRT-PCR analyses show that the injection of 4A-4E-BP mRNA did not alter the Myh1, 4, 8 and 6 mRNAs expression level (Fig. 7I; supplementary material Fig. S4). Then, immunohistochemistry and qRT-PCR analysis of MyoD expression showed that this transcription factor is correctly expressed in 4A-4E-BP-mRNA-injected embryos at both protein and mRNA levels (Fig. 7J–L,M, respectively).
These results show that overexpression of the non-phosphorylatable 4E-BP protein leads to similar muscle effects compared to those observed in embryos exposed to hypoxia on the number of mitotic cells, the expression of muscle transcription factor and on the accumulation of muscle-specific proteins in the somite cells, independently of mRNAs variation.
Our data show that exposure to hypoxia for 24 hours significantly reduced growth and delayed embryonic development in an oxygen-dependent manner. These effects are similar to those observed in mouse, chicken and sea urchin embryos (Le Bouffant et al., 2006; Sharma et al., 2006; Ream et al., 2008). Hypoxia also affected embryonic survival depending on the oxygen availability. During human pregnancy, acute hypoxia observed during retroplacental haemorrhage, intoxication with CO, and acute decompensation of maternal diabetes mellitus, has been linked with the sudden death of the foetus. Our model is much more linked with chronic hypoxia of the foetus. In such cases, chronic hypoxia leads to intrauterine growth retardation (IUGR; Vandenbosche and Kirchner, 1998) defined by a foetal weight corresponding that less than 10% of predicted foetal weight for gestational age. Its annual incidence is estimated between 4% to 7% of the total gestation number and the major conditions that can give rise to IUGR are maternal hypertension, maternal diabetes mellitus, maternal systemic lupus, smoking, certain infectious diseases, and lastly foetal chromosomal abnormalities like 21 trisomy or Turner syndrome. Interestingly, since Xenopus embryos develop externally, we show that growth retardation and embryonic death were independent on the maternal behaviour and placenta formation. Moreover, the growth retardation of embryos may reflect defects in timing of development suggesting alterations of developmental timer. However, this feature of hypoxia was not explored.
In the same way, we demonstrate that both strong and moderate hypoxia induce apoptosis. While the p38-MAPK has been shown to be involved in apoptosis (Bulavin et al., 1999; She et al., 2000; Zarubin and Han, 2005), in our hands there is no clear correlation between apoptotic patterns and phospho-p38-MAPK levels. Therefore, hypoxia may activate other mechanisms involved in apoptosis induction, such as the activation of pro-apoptotic factors. Since Akt was shown to be involved in cell survival (Dudek et al., 1997), the regulation of this kinase could be a good candidate to activate apoptosis in hypoxia. Indeed, Bcl-2-Associated Death promoter (BAD) protein, caspase 9, and FoxO have been identified as targets of Akt. Activated Akt can phosphorylate and inactivate these factors, whereas the inhibition of Akt induces apoptosis (Datta et al., 1997; Kandel and Hay, 1999; Brazil and Hemmings, 2001). In agreement with this hypothesis, we show that Akt is reversibly dephosphorylated in hypoxia suggesting a link between this signalling pathway and the hypoxia-induced apoptosis in Xenopus embryos. However, we cannot exclude the involvement of other signalling pathways in the induction of apoptosis. Interestingly, the pro-apoptotic factor Bmf (Bcl-2-Modifying Factor), a member of the Bcl-2 family, has been shown to act as a sensor for stress that associates with the repression of the cap-dependent translation (Grespi et al., 2010).
Cap-dependent translation is correlated with the availability of the eukaryotic initiation factor eIF4E, which constitutes a major checkpoint for mRNA translation regulation (Richter and Sonenberg, 2005). The availability of eIF4E has been implicated in hypoxic responses in part by the phosphorylation of 4E-BP, a downstream target of the Akt signalling pathway (Wouters et al., 2005). Strikingly, we demonstrate that in Xenopus embryos hypoxia-induced overexpression of a functional 4E-BP protein able to sequester eIF4E. The fact that 4E-BP association with eIF4E increases in hypoxia is in agreement with the Akt inhibition, since the 4E-BP/eIF4E dissociation complex is dependent of an active Akt signalling pathway and the phosphorylation state of 4E-BP. While the phosphorylation control of 4E-BP has been extensively studied (Armengol et al., 2007), little is known about the significant role of the variation in 4E-BP levels as a means of control of the availability of eIF4E (Cormier et al., 2003). The first demonstration that modification of 4E-BP levels may influence cell fate has been obtained in sea urchin where 4E-BP is fully degraded shortly after fertilization to allow embryonic development (Salaun et al., 2003). Moreover, hypoxia was shown to result in a reversible increase in 4E-BP protein levels in sea urchin embryos (Le Bouffant et al., 2006). In agreement with this, we show that the amount of the 4E-BP protein increases dramatically in Xenopus embryos following hypoxia therefore supporting the hypothesis that, besides its phosphorylation state, 4E-BP overexpression is also important to inhibit eIF4E availability and subsequent cap-dependent translation.
However, little is known about the mechanisms that control 4E-BP expression since the amount of protein in the cell reflects both protein synthesis and degradation. Therefore, a model is proposed in Fig. 8 to illustrate mechanisms that could be involved in 4E-BP overexpression during embryonic development in hypoxia. On the one hand, deactivated Akt in hypoxia could prevent the phosphorylation of FoxO members resulting in nuclear translocation of these transcription factors and expression of FoxO-regulated genes such as 4E-BP, as shown in Drosophila embryos under dietary restriction and oxidative stress (Tettweiler et al., 2005). It is important to note that in pancreatic cells, 4E-BP mRNA transcription is regulated by the transcription factor Smad4 after TGFβ treatment (Azar et al., 2009) and consequently, it cannot be excluded that 4E-BP mRNA transcription is under the control of different transcription factors. However, qRT-PCR analyses did not show any significant variation in 4E-BP mRNA levels in embryos exposed to 10% and 5% O2, suggesting that the 4E-BP gene transcription was not responsible for the increase in 4E-BP protein amount at stages 21 and 25. On the other hand, deactivated Akt in hypoxia prevents the activation of its downstream mTOR protein and therefore prevents the phosphorylation of 4E-BP. Since it was shown that inhibition of proteasomal activity enhances the level of hypophosphorylated 4E-BP in Jurkat cells, 4E-BP protein can be activated and stabilized to accumulate in hypoxic cells via the inhibition of its degradation pathway. In agreement with this hypothesis, we show that inhibition of the PI3K/Akt pathway or mTOR activity, using LY294002 or rapamycin respectively, leads to an increase in 4E-BP amount in normoxic embryos, suggesting that the 4E-BP level is dependent of the Akt signalling pathway and under a rapamycin-sensitive pathway. Interestingly, since the inhibition of mTOR, known to phosphorylate 4E-BP, leads to its accumulation in normoxic embryos and slows its degradation in reoxygenated embryos, our data also suggests that the phosphorylation state of 4E-BP is associated with its stability, probably due to the inhibition of its degradation pathway. However, an additional increase in 4E-BP mRNA translation could also participate in the accumulation of 4E-BP in hypoxic cells (Fig. 8).
Importantly, the inhibition of the initiation step of mRNA translation by 4E-BP has been proposed to act as a cellular survival mechanism involving decreased protein synthesis in order to conserve oxygen because protein synthesis is extremely energy costly. Teleman et al. (Teleman et al., 2005) and Tettweiler et al. (Tettweiler et al., 2005) provide support for the role of 4E-BP in mediating Drosophila embryonic survival under unfavourable conditions. In agreement with this idea, we demonstrate that effects of hypoxia exposure are reversible when oxygen is available and embryonic death was reduced with reoxygenation concomitant with the decrease in 4E-BP level. Moreover, inhibition of 4E-BP mRNA translation with specific morpholino injection in two-cell stage embryos strongly induced lethality in normoxia, suggesting that a basal level of 4E-BP is required to regulate protein synthesis and ensure survival of embryos during the early embryonic development (data not shown).
In this study, we also observed that embryos exposed to hypoxia did not make any movement. In vertebrate embryos, somites are regular transient structures repeated along the anterior/posterior axis of the embryo, which then differentiate into a part of the dermis, bone, cartilage, tendon-cell lineages and skeletal muscles. Since the sclerotome is formed at stage 37 and the dermatome is a single layer of cells under the epidermis, somites of embryos at stages 21 and 25 are mainly composed of myotome cells, which will give rise to skeletal muscle. Because premyoblast proliferation is known to be an early step of the myogenesis, one hypothesis was that hypoxia exposure affects cell proliferation in developing somites. In agreement with this, we found that hypoxia induces a 90% decrease in mitotic cell in somites. Interestingly, 4E-BP is known to be implicated in the inhibition of cell proliferation (Armengol et al., 2007; Azar et al., 2009) and, in the same way, we show that overexpression of a non-phosphorylatable 4E-BP protein in somites induced a decrease in cell proliferation. Consequently, hypoxia-induced 4E-BP overexpression is in agreement with the inhibition of myoblast proliferation. However, how is this only cell type affected by hypoxia remains an open question. One hypothesis is that the hypoxia-induced 4E-BP overexpression occurs selectively in myoblast precursors. Therefore, it would be interesting to perform in situ analysis of protein distribution in the embryos exposed to hypoxia.
In addition, we observed that the 12/101 monoclonal antibody, which recognizes an unidentified epitope on the differentiated muscle cells (Kintner and Brockes, 1984), produced a weaker signal in hypoxic embryos. We demonstrate that hypoxia does not affect mechanisms involved in the segmentation of the presomitic mesoderm but affects accumulation of muscle-specific proteins during the embryonic development. However, hypoxia does not disturb the protein expression pattern of both XMyoD and XMRF4, early and late markers of muscle cell fate respectively, suggesting that oxygen does not interfere with the events leading to the expression of the myogenic transcription factors in somite nuclei. Therefore, hypoxia could affect cap-dependent translation of mRNA produced downstream MyoD and MRF4 activation. The fact that MyoD and MRF4 were not affected following hypoxia treatment or 4E-BP overexpression lead us to propose that the translation of these two transcription factors mRNAs is under cap-independent control which is mostly mediated by mRNA structural element called IRES (Hellen and Sarnow, 2001). Interestingly, IRES have been identified in several mammalian mRNAs, mainly in control genes such as transcription factors and growth factors (Vagner et al., 2001). For instance, functional IRESs have been recently identified in Fibroblast Growth Factor 1 (Martineau et al., 2004; Conte et al., 2009), which plays a crucial role in myoblast differentiation. It is noteworthy that a number of genes including the hypoxia-inducible genes encoding HIF-1α remain preferentially translated when eIF4E is sequestered by its repressor 4E-BP (Wouters and Koritzinsky, 2008) and we show that HIF-1α mRNA levels increased depending on the oxygen availability. Therefore, 4E-BP overexpression could represent an oxygen-sensitive pathway that occurs upstream changes in gene expression mediated by HIF-1 in hypoxia. This regulation of the translational machinery by hypoxia in Xenopus embryos might support mRNA-selective translation. In agreement with this hypothesis and with our data, it has been recently published that translation facilitates genes classes like transcription and signal transduction during acute hypoxia, whereas it represses the expression of genes involved in cell growth and protein metabolism (van den Beucken et al., 2011). Therefore, determination of the number and the nature of mRNAs translated in hypoxia condition is now an important goal.
Finally, an important step in muscle development is the abundant synthesis of specific proteins required for muscle cell contraction including actin and myosin that approximately occupy 80% of the cytoplasmic volume. To determine whether the translation machinery and particularly the availability of the cap binding protein eIF4E was involved in muscle defects, two constructs were used. The first one codes for full-length sea urchin 4E-BP protein (WT-4E-BP). The second one codes for 4E-BP mutated protein (4A-4E-BP) in which four phosphorylation sites of 4E-BP were mutated dramatically affecting the release of 4E-BP from eIF4E (Oulhen et al., 2009). Our data support that the phosphorylation state of the 4E-BP protein is crucial to ensure the stability of the protein in Xenopus embryo since the 4E-BP protein accumulate in 4A-4E-BP-injected embryos but not in WT-4E-BP. Moreover, since mTOR is the major kinase of these phosphorylatable sites, this is in good agreement that the regulation of 4E-BP amount is under the PI3K/Akt/mTOR axis. We also observed that the 12/101 monoclonal antibody produced a weaker signal although MyoD was expressed in 4A-4E-BP-injected embryos. We therefore demonstrate that overexpression of the non-phosphorylatable 4E-BP protein leads to similar effect on muscle proteins than in embryos exposed to hypoxia (i.e. decrease in 12/101 signal but not in MyoD expression). The data suggest for the first time that the hypoxia-induced 4E-BP could be responsible for defects in muscle formation during the early vertebrate development of the myotomes, without affecting specification by the myogenic transcription factors but preventing the abundant synthesis of muscle-specific proteins.
Materials and Methods
Xenopus laevis embryos, hypoxia exposure and measurement of dissolved oxygen
Xenopus laevis embryos were obtained as previously described (Hidalgo et al., 2009). After the gastrulation events [stage 13 according to Nieuwkoop and Faber (Nieuwkoop and Faber, 1994)], embryos were grown in 0.1× MMR with several oxygen levels (10% and 5% O2) in hypoxic chambers. The oxygen partial pressure in the medium was measured at several oxygen levels with an apparatus used to measure blood gases (Radiometer ABL520, Copenhagen, Denmark).
Plasmid constructs and mRNA injection experiments
The 342-bp fragment corresponding to sea urchin wild-type 4E-BP or to mutant 4A-4E-BP was isolated from the pGex4T1 plasmid (Oulhen et al., 2009) and inserted into the pCS2+ vector digested by BamH1 and Xho1. Synthetic capped mRNAs were made by in vitro transcription as described by Djiane et al. (Djiane et al., 2000). All blastomeres or two blastomeres of two-cell stage or four-cell stage embryos were injected depending on the experiment.
Isolation of eIF4E and associated proteins
Xenopus embryos were collected and homogenized in lysis buffer (50 mM Tris HCl, 100 mM NaCl, pH 7.5, 2 mM EDTA, 2 mM EGTA, 50 mM β-glycerophosphate, 50 mM sodium fluoride, 1 mM Na3VO4, 120 nM okadaic acid, 1 mM PMSF, 1% Triton X-100) supplemented with protease inhibitors. Lysates were centrifuged for 15 minutes at 14,000 g at 4°C. Isolation of eIF4E and its partners from embryo extracts was performed using m7GTP beads (Amersham Biosciences). Briefly, 2 mg proteins were incubated with 25 µg of m7GTP Sepharose beads for 1 hour at 4°C. The beads were washed with 2 ml of 1× binding buffer containing 100 mM NaCl. Laemmli sample buffer was added directly to the beads and denatured at 95°C for 5 minutes before western blot analysis.
Western blot analysis
Xenopus embryos were collected and homogenized in lysis buffer (50 mM Tris HCl, 100 mM NaCl, pH 7.5, 2 mM EDTA, 2 mM EGTA, 50 mM β-glycerophosphate, 50 mM sodium fluoride, 1 mM Na3VO4, 120 nM okadaic acid, 1 mM PMSF, 1% Triton X-100) supplemented with protease inhibitors. Protein samples, separated by 12% SDS-PAGE, were transferred to nitrocellulose membranes (Hybond) as described by Towbin (Towbin et al., 1979). The membranes were blocked in 5% non-fat dry milk and incubated with the following primary antibodies: anti-α-tubulin (T0926, Sigma), anti-4E-BP1 (9452, Cell Signaling), anti-eIF4E (9742, Cell Signaling), anti-p38-MAPK (9212, Cell Signaling), anti-phospho-p38-MAPK (9211, Cell Signaling), anti-phospho-Akt Ser473 (9271, Cell Signaling), anti-phospho-Akt Thr308 (9275, Cell Signaling), anti-Akt total (9272, Cell Signaling), anti-sea urchin 4E-BP [alpha-69 (Oulhen et al., 2010)], 12/101 (DSHB). After washing they were incubated with the appropriate secondary antibodies: anti-rabbit HRP conjugated (111-036-045, Jackson ImmunoResearch), anti-mouse HRP conjugated (115-036-062, Jackson ImmunoResearch), and detected by chemiluminescence. All experiments were repeated at least three times. For quantification, blots were exposed to X-ray film for various time points and the films were scanned and analysed using the ImageJ software from the NIH.
Embryos were embedded with 15% cold-water fish gelatin (FLUKA, Biochemika) and 15% sucrose. Tissues were sectioned at 14 µm thickness by a cryostat (Leica CM 3050S). Sections were blocked in 20% goat serum and the following primary antibodies were used: mouse monoclonal 12/101 (DSHB; 1:2000), F59 (DSHB; 1:100), alpha-69 [(Oulhen et al., 2010) 1:100], mouse anti-MyoD (D7F2, DSHB), guinea pig anti-MRF4 (kindly provided by B. Della Gaspera, Université Paris 5, France; 1:50), mouse anti-P63 (Abcam; 1:50), mouse anti-acetylated tubulin (Sigma 6-11B-1; 1:500). After washing they were incubated with the appropriate secondary antibodies: anti-mouse CY3 conjugated (Sigma C2181; 1:100), anti-rabbit FITC conjugated (Jackson ImmunoResearch 111-095-144; 1:40), anti-guinea pig Alexa-Fluor-488 conjugated (Invitrogen A11073; 1:500). Nuclei were stained with Hoechst H33258 (Sigma; 1:1000). Sections were washed and then mounted in Immunomount (Thermo Electron Corporation).
All imaging was done at room temperature. Immunofluorescent staining was imaged using a Nikon Eclipse E800 microscope equipped with a QEi Evolution camera (Media Cybernetics) (ARC No. 7867). A 4× (Plan 0.1 NA), 10× (Plan-apochromat 0.45 NA) or 20× (Plan-apochromat 0.75 NA) were used and the image acquisition software was Image-Pro Plus.
Total RNAs were isolated from 6 embryos (stages 21 and 25) using the RNeasy mini kit (Quiagen) according to the manufacturer's instructions. All primers were designed using Primer Express Software. PCR reactions were carried out using SYBR green (Applied Biosystems) on a StepOnePlus (Applied Biosystems). All experiments were repeated at least 3 times on separate experiment and the real-time PCR was also performed in duplicate. The results were analysed using the 2-ΔΔCt method (Livak and Schmittgen, 2001). The relative expression of genes is shown normalized to the expression of the house-keeping gene RPL13 although several house-keeping genes were tested to validate the results. The real-time PCR primers are given supplementary material Table S1.
TUNEL staining and proliferation assays
Embryos (n = 15) were fixed and sectioned. TUNEL (terminal deoxynucleotidyl transferase-mediated deoxyuridinetriphosphate nick end-labelling) staining was carried out following the protocol as previously described (Hidalgo et al., 2009). Sections were mounted in Immunomount (Thermo Electron Corporation) and positive cells were counted, on all sections of embryos in several tissues, by microscopy at a magnification of 100×, in non-overlapping fields.
Proliferation assays were performed on sections of embryos (n = 8) stained with the rabbit anti-human phospho-histone H3 antibody (Ser 10, mitosis marker, Euromedex H5110-14B; 1:500) and the anti-rabbit alkaline phosphatase-conjugated antibody (Jackson ImmunoResearch 111-055-144; 1:5000). Positive cells were counted, on all sections of embryos in several tissues, in both the control and the hypoxic embryos by microscopy at a magnification of 100×, in non-overlapping fields.
We are grateful to B. Della Gaspera (Université Paris 5, France) for providing the XMRF4 antibody. Anti-MyoD (developed by J. Gurdon and H. J. Standley), 12/101 (developed by J. P. Brockes) and F59 (developed by F. E. Stockdale) antibodies were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA 52242. We also thank Martin Catala, Thomas M. Onorato and Daniel J. Berg for reading and commenting on the manuscript.
This work was supported by a predoctoral fellowship from P13 University to M.H.; the Association Française contre les Myopathies [grant number 13955 to M.B. and T.D.]; and Réseau Picard of the P. and M. Curie University to M.H. and T.D.
Supplementary material available online at http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.097998/-/DC1
- Accepted May 15, 2012.
- © 2012. Published by The Company of Biologists Ltd