In recent years, metabolism has re-emerged as an active effector of physiological changes rather than a housekeeping process in development. Embryogenesis is underpinned by shifts in cellular energy requirements, as well as the use of anabolic precursors (Gándara and Wappner, 2018). However, we still know little about how cellular processes maintain this metabolic balance to build an adult organism from a single cell (Ghosh et al., 2023).
One of the key players in this process is glucose metabolism. Under normal oxygen levels and in the presence of glucose, cells primarily rely on mitochondrial cellular respiration for energy generation. Alternatively, cells can opt for aerobic glycolysis (Warburg, 1956), where glycolysis produces ATP even in the presence of oxygen, bypassing mitochondrial oxidative phosphorylation. Although aerobic glycolysis is a less efficient method to generate ATP from glucose, it meets the increased energetic and anabolic demands associated with processes such as cell signalling, proliferation and differentiation (Miyazawa and Aulehla, 2018).
Besides the function of cellular metabolic pathways in supplying energy and carbon chains, intermediates often participate in epigenetic regulation by serving as substrates or cofactors for epigenome-modifying enzymes (Gándara and Wappner, 2018). Therefore, metabolism is a highly regulated process that can be rewired not only to maximize energy production but also to meet the varying energetic and growth demands, while concurrently influencing gene regulation. Following this growing appreciation for metabolic control over fate specification, a body of recent preprints have considered metabolic regulation, specifically glucose metabolism, as a possible signalling axis in the control of gastrulation (Cao et al., 2023 preprint; Dingare et al., 2023 preprint; Stapornwongkul et al., 2023 preprint; Villaronga Luque et al., 2023 preprint).
In the first of these preprints, the Sozen lab looked at cell-type specific programming of glucose handling in developing mouse embryos at embryonic day (E) 6.5-E7.25 (Cao et al., 2023 preprint). They studied the expression of a glucose transporter (Glut1) and traced glucose using 2NBDG, a fluorescent tracer used for monitoring glucose uptake into living cells. They observed two distinct waves of glucose uptake before primitive streak formation in the epiblast and again in the mesoderm during lateral migration. They then revealed the role of glucose metabolism at these stages. By blocking specific enzymes at different steps of glucose metabolism, they showed that the hexosamine biosynthetic pathway (HBP) in epiblast cells impairs mesoderm fate specification, epithelial-mesenchymal transition and distal elongation. In contrast, inhibition of glycolytic enzymes of downstream steps, such as the phosphofructokinase PFKFB3 or pyruvate kinase M2, impacted lateral migration of the mesoderm. Finally, the authors aimed to elucidate a mechanism for this disruption by interrogating the FGF signalling pathway because it is important for primitive streak (PS) formation (Ciruna and Rossant, 2001) and can be disrupted by blocking protein-modifying enzymes that use glycolytic substrates (Moussaieff et al., 2015; Shimokawa et al., 2011; Oginuma et al., 2018). Here, they used dephosphorylated ERK signal and a live ERK reporter as a readout of FGF signalling and showed that FGF signalling follows a similar pattern to the previously observed wave of glucose uptake. This signal was lost with general glucose metabolism inhibition by 2DG treatment (which prevents glucose transport into the cell) and specifically with HBP inhibition by Azaserine treatment both in embryos and in in vitro mesoderm differentiation of mouse embryonic stem cells. Finally, they related the previously reported FGF-inhibition phenotype of posterior streak build-up of cells to their glycolysis-inhibited embryos (via the small molecule YZ9). Quantifications of the treatment indicate an increase in PS size and dividing cells, leading the authors to conclude that FGF signalling may regulate late stages of glycolysis within mesodermal cells.
Moving fully in vitro, two subsequent preprints from the Steventon (Dingare et al., 2023 preprint) and Trivedi (Stapornwongkul et al., 2023 preprint) labs follow similar lines of thought looking into glucose-based metabolic control over the emergence of posterior lineages in a 3D mouse embryonic stem cell-based gastrulation model (Beccari et al., 2018). Dingare and colleagues specifically focussed on mesoderm specification, demonstrating that several glucose transporters are present in the elongating pole of the gastrulation model and overlap with known markers of neuro-mesodermal progenitors and mesoderm (T/Bra). The authors then show that inhibiting glucose uptake with 2DG before symmetry breaking, presumed gastrulation in this model, impairs mesoderm specification and elongation of the stem cell aggregates. Concordantly, Stapornwongkul and colleagues showed similar results at an earlier stage in their differentiation, with 2DG inhibition of glucose pathways at 24-48 h of differentiation. Next, both papers tested whether glucose deprivation would have a similar effect on aggregate elongation as glucose inhibition with 2DG. When the gastrulation model was cultured in media lacking glucose at 24-48 h by Stapornwongkul and colleagues, loss of PS markers and elongation was observed in agreement with 2DG treatment; however, Dingare and colleagues saw no effect of glucose deprivation at 72-96 h of their differentiation. Further, Stapornwongkul and colleagues showed that this early deprivation of glucose reduces the proportion of endoderm and mesoderm specified, and that the effects of early glycolytic inhibition can be rescued by the addition of activin A to the culture to induce Nodal signalling. Meanwhile, Dingare and colleagues demonstrated that glucose deprivation alone did not have an effect at later stages (72-96 h). Instead, they showed through metabolomics that 2DG inhibition has a distinct metabolic signature in comparison to glucose depletion. They also demonstrated that supplementation of mannose to their 2DG glycolysis-inhibited system rescued the defect in mesoderm differentiation, highlighting that, at this later stage in the gastrulation differentiation, it is not glucose-driven glycolysis but an alternate pathway through mannose metabolism that is regulating fate specification. Overall, like Cao and colleagues, these two preprints show that within pluripotent stem cell models of gastrulation different processes divert the uses of glucose into specific metabolic pathways that may not be directly related to maximizing the energetic potential of glucose through glycolysis. Thus, glucose metabolism appears to be temporally adjusted during the emergence of primitive streak lineages and mesoderm specification. Importantly, the in vitro findings in the preprints match in vivo findings, which argues that stem cell models may be a good system to dissect the role of metabolic pathways during development.
Finally, the last preprint we discuss, from the Veenvliet lab (Villaronga Luque et al., 2023 preprint), did not directly start out investigating glycolysis. However, like the previously discussed preprints, it generated similar conclusions in an unbiased manner about glucose pathway control of gastrulation. The motivation behind Villaronga Luque, Savill and colleagues was to determine the source of heterogeneity in their in vitro model of axial elongation (∼E7.5-E9) (Veenvliet et al., 2020), which demonstrates the structural organization of the developing trunk but with seemingly stochastic variation. Integrating staged molecular (single-cell RNA sequencing) and morphogenic (imaging-based quantitative phenotypic profiling) information throughout their differentiation, the authors identified early prognostic features at around 72-96 h that resulted in defined morphological differentiation outcomes at 120 h. Leveraging the predictive power, Villaronga Luque, Savill and colleagues then examined pseudo-bulk transcriptome data and found that, at this ‘predictive stage’, there was higher expression of genes related to glycolysis and lower expression of genes related to oxidative phosphorylation in structures predicted to be successful. This led them to hypothesize that divergent metabolic states are a key driver for the main differences in morphologies at the end time point. Subsequently, they used similar inhibitors to the previously discussed preprints (2DG and rotenone) at equivalent time scales to Dingare et al. (2023 preprint) (72-96 h or 96-120 h) to block glucose metabolism or boost specifically glycolysis. These treatments perturbed the proportions of cell types and structure morphology of the endpoint in their differentiation. In this preprint, the findings underscore the involvement of metabolic pathways in altering cell proportions during gastrulation, especially during axial elongation, aligning closely with the in vivo observations made by Cao and colleagues. Furthermore, they showcase how early metabolic status of stem cell differentiations could be used as a predictive tool for end-point outcomes (Nikitina et al., 2023; Maughon et al., 2022).
These preprints, regardless of approach, all point to potential connections of glucose metabolism with signalling pathways in gastrulation. Cao and colleagues link glucose uptake waves to FGF signalling, whereas Villaronga Luque, Savill and colleagues demonstrate correlated gene expression involving FGF, glycolysis and Wnt activity. In Stapornwongkul and colleagues’ study, changes in cell lineage proportions were rescued by Nodal signalling pathway (activin A) and partially by CHIR (Wnt), and Dingare and colleagues link the function of specific metabolites to signalling pathways by hypothesizing that mannose rescue of glycolysis inhibition highlights the need for mannosylation of proteins involved in lineage specification. Expanding on these works, are there other mechanisms to link the function of specific metabolites to signalling pathways? Glycolytic intermediates may also affect post-translational modifications, such as glycosylation (Oginuma et al., 2018; Tian et al., 2012; Czajewski and Van Aalten, 2023). Lipid or sugar modifications on key signalling molecules and their receptors, such as Wnt or FGFR, are essential for their synthesis, transport and extracellular properties, or to act as co-receptors and activate the downstream response (Mehta et al., 2021; Ghabrial, 2012; McGough et al., 2020). It is also possible that glycolytic metabolites mediate epigenetic changes through changes in acetyl-CoA and histone acetylation (Gándara and Wappner, 2018; Moussaieff et al., 2015), which could affect signalling and gene regulation. We need to consider that metabolic regulation may impact the pace of gastrulation as well. Glycolysis regulates the segmentation clock period either through signalling via the glycolytic metabolite FBP upstream of Wnt signalling in vivo (Miyazawa et al., 2024 preprint) or altering the bioenergetics of cells via the NAD+/NADH ratio in vitro (Diaz-Cuadros et al., 2023). Future work will need to assess the role of metabolic regulation in the pace of biological processes. Moreover, it would be good to investigate the effect of the glycolytic perturbations in terms of anabolism and how these effects may be linked with signalling. These same glycolytic intermediates funnel into multiple biosynthetic pathways to generate ATP or to synthesize de novo nucleic acids, phospholipids and amino acids to cater for the demands of new biomass, which makes the decoupling of bioenergetics from signalling increasingly complicated.
Some of the preprints dissect the contribution of glucose to the production of energy through cellular respiration. Stapornwongkul and colleagues found that blockade of oxidative phosphorylation with sodium azide reduced growth but did not impact changes in T/Bra expression; Villaronga Luque, Savill and colleagues did not describe significant elongation defects associated to blocking glucose metabolism, whereas inhibiting mitochondrial complex I corrected neural bias and aberrant morphogenesis. Altogether, these results suggest that, in the absence of functional respiration in mitochondria, cells can adapt by rewiring metabolic pathways to produce enough energy through glycolysis, impacting neural proportions. However, the preprints do not explore the balance between energetics and signalling in this context. For example, one could measure ATP production after blocking the steps that generate ATP in glycolysis. As the field moves into considering the signalling potential of glycolysis in-depth, we need to be able to establish controllable perturbations to decouple energetics from signalling. Ideally, we would want to understand the energetic budget at a cellular resolution as well as the impact of specific metabolites on cell fate and morphological changes.
Looking forward, research in the field could aim to address multiple aspects of metabolic control. The role and regulation of post-translational modifications on key signalling molecules, the regulative potential of sugars or lipids through perturbing key enzymes implicated in these biosynthetic pathways or targeting the role of metabolism in epigenetics and gene regulation. Finally, as the field of development and metabolism continues to expand, we can begin to piece together how alternative metabolic pathways beyond the central glucose pathway could continue to fine tune a developmental system.
Footnotes
Funding
A.L. is supported by a long-term postdoctoral fellowship awarded by the European Molecular Biology Organization (ALTF 149-2020). T.R. is supported by an Institute Strategic Programme Grant (BBS/E/B/000C0421) and a Pioneer Award (BB/Y513076/1) funded by the Biotechnology and Biological Sciences Research Council, and an ERC Starting Grant funded by the Engineering and Physical Sciences Research Council (EP/X021521/1).
References
Competing interests
The authors declare no competing or financial interests.