Metabolism across species: using transcriptome profiling to investigate glycolytic and oxidative flux in mammalian embryos

Mammalian embryology research relies on animal models with different developmental timelines and implantation modes. Here, we present a cross-species analysis of 15 single-cell RNA-seq datasets showing conserved glycolysis and oxidative phosphorylation trends in pre-gastrulation mammalian embryos.

Like Comment
Read the paper

Cells of the early mammalian embryo face a unique challenge: to produce large numbers of progeny quickly whilst regulating appropriate cell differentiation in a short amount of time. The regulation of energy metabolism is crucial for these processes, not only to supply energy for proliferation and cell movements and shape changes, but also to generate substrates to synthesise biomolecules such as lipids, amino acids, and nucleotides. The main two energy metabolism processes are glycolysis and oxidative phosphorylation (OxPhos). Crucially, metabolic flux can influence the pool of substrates required for writing epigenetic modifications, reflecting its crucial role in embryonic development.1

Various mammalian species are currently used as model organisms to elucidate the mechanisms of early development. They all differ in the length of embryogenesis pre-gastrulation and pre-implantation, what extraembryonic tissues they produce, and at what time, as well as the mode of implantation they undergo. Whereas the metabolic changes in selected species have been empirically studied during pre-implantation by quantifying metabolites in culture2, it is much more difficult to study these trends in vivo and in post-implantation tissues. To address this knowledge gap, we became interested in how glycolytic and oxidative flux states in embryos of various mammalian species differ from each other.  

Beginning this project during the lockdown period allowed us the time necessary to think about and implement bioinformatic meta-analyses. The group’s recent work on spatial profiling produced a large single-cell RNA-seq dataset of marmoset early embryogenesis.3 At the start of the project, we decided to investigate how metabolism changes during the development of the marmoset embryo, and for comparison, we have started to assemble and analyse other mammalian datasets. In the end, we have compiled 15 publicly available single-cell RNA-seq datasets from six mammalian species (human, cynomolgus monkey, marmoset, mouse, pig, and opossum). Importantly, these species represent different modes of implantation and differ in the time of development pre-gastrulation. We have annotated and stage-matched all datasets to reflect a timeline approximating Carnegie stage one (zygote) through to seven (gastrula).

Employing the ‘module score’ analysis of single-cell datasets available through the Seurat suite, we have used the transcription levels of glycolytic and oxidative genes as a proxy for the metabolic flux in pre-gastrulation embryos. We show that the six mammalian species exhibit a surprising level of conservation in the glycolysis and oxidative phosphorylation trends, regardless of up to 160 million years of evolution separating them. This indicates that mammalian embryogenesis is intrinsically regulated and independent of the timing and mode of implantation. 


Figure 1. Summary of metabolic trends in 6 mammalian species.

Based on our analysis, we can summarise the metabolic trends of the early mammalian embryogenesis as follows: the early cleavage stages in mammalian embryogenesis are characterised by low levels of metabolic activity, which start increasing with a rise in oxidative phosphorylation during the formation of the blastocyst (Figure 1). As cavitation of the blastocyst is a highly energy-consuming process due to numerous cell rearrangements and lumen formation, high oxidative flux can provide the ATP required for those processes. Glycolysis is only highly increased following the blastocyst stage, at which point OxPhos levels start to drop. This reflects the requirement for cell growth and therefore production of macromolecules. For instance, directing the energy flux through the glycolytic pathway rather than the oxidative pathway can channel the substrates to the pentose phosphate pathway, which generates metabolites required to produce lipids and nucleotides. Crucially, all species show a direct switch from a bivalent to a mostly glycolytic metabolism as the embryonic disc is formed, even in animals where embryonic disc formation occurs before implantation, such as the pig and the opossum.

Our use of single-cell transcriptomic profiling allowed us also to investigate the metabolic differences between the embryonic and extraembryonic tissues of various mammalian species. We have discovered that the hypoblast and trophoblast as well as the tissues derived from them at a later stage largely follow the metabolic conditions of the embryonic counterparts at the same stage. Interestingly, visceral endoderm during the formation of the embryonic disc (CS5) showed consistently decreased metabolic flux in comparison to the embryonic cells.

Finally, we have also investigated the energy metabolism trends of in vitro models of embryonic development, confirming well-established metabolic characteristics of embryonic and epiblast-like stem cells. We have also compared available datasets of in vitro cultures to the corresponding in vivo embryonic tissue. As expected, we have observed that culturing the embryonic cells in vitro has an extensive effect on their metabolic flux. In the cynomolgus monkey samples, where a direct comparison between stages was possible, we found that metabolic scores for both OxPhos and glycolysis were increased in the in vitro cultured embryos. Moreover, different types of embryo culture also influence metabolic scores, as embryos cultured in Matrigel showed an increase in glycolysis in contrast to embryos cultured in a regular liquid medium. We hypothesize that these differences are due to the composition of the culture media, which may provide metabolites in excess, thus boosting metabolic flux. All in all, our results indicate that it is crucial to carefully consider the effect of embryo culture on the metabolic flux, which can have major effects on cell differentiation and proliferation.

In summary, we have compiled several single-cell RNA-seq datasets, using transcriptional profiling to infer metabolic changes in early mammalian embryos. Our work has shown that the dynamics of energy metabolism in the early mammalian embryo are surprisingly similar across six different species, despite different timing and mode of implantation. We hope that our work will serve as a resource for researchers studying metabolic changes in mammalian embryogenesis.


  1. Krisher, R. L. & Prather, R. S. A role for the Warburg effect in preimplantation embryo development: Metabolic modification to support rapid cell proliferation. Molecular Reproduction and Development 79, 311–320 (2012).
  2. Leese, H. J. Metabolism of the preimplantation embryo: 40 years on. Reproduction 143, 417–427 (2012).
  3. Bergmann, S., Penfold, C.A., Slatery, E. et al. Spatial profiling of early primate gastrulation in uteroNature (2022).

Anna Malkowska

PhD Student, University of Cambridge