Background

All animals, big and small, begin their lives the same way: as a fertilised egg, and then an embryo. Yet the amount of time different animals devote to embryonic development varies wildly between different species. The human pregnancy lasts 9 months, and mouse gestation is just 21 days long, yet the blue whale is pregnant for 10-12 months – so this clearly does not correlate with size. The record for the longest gestation goes to the alpine salamander, developing for 2-3 years at low altitudes, or 3-5 years in the cold climates of the Alps!

What governs the rate of embryo development, if it isn’t size or complexity? The alpine salamander gives us a clue: in cold temperatures, the mother’s metabolism drops, leading to a longer gestation. Metabolic rate is a key determinant in the overall pace of development.

Mice have a higher metabolic rate than humans, and their embryos develop much faster – despite needing to build much the same set of body parts (whiskers aside). Somitogenesis is a good example: the rhythmic formation of the segments that will become our vertebrae and skeletal muscle proceeds at double the speed in mice. However, unlike the salamander, this can’t be explained by temperature. Mouse and human cells develop at different rates even when grown in a dish in the same conditions, hinting that metabolic rate must be species-intrinsic.

Why I chose to highlight this preprint

I chose to highlight this preprint because I’m fascinated by this huge amount of variation across the animal kingdom. We know about the metabolic differences, but there is a clear gap in our knowledge, as we don’t know the determining factors for species-specific metabolic rate. Here, the authors use directly comparable stem cell culture systems for both mouse and human. This allowed for causal relationships between the genetics, metabolic rate, and the pace of development to be uncovered.

Metabolism 101

If you’re anything like me, you know that metabolism is central to all biology, but struggle to remember all the different terms and pathways. Here is a reminder of the parts you need to understand for this preprint.

Key findings

Redox metabolism controls developmental rate

The ratio of NAD+ to NADH acts as a gauge of the cell’s metabolic state: high NAD+ is similar to a build-up of empty batteries that are ready to accept electrons, while high NADH indicates a build-up of charged batteries, which causes glycolysis to slow down. Mirroring the results of previous work, the authors of this preprint could demonstrate that this ratio is linked to the speed of somitogenesis.

To do this, they used a stem cell-derived model of human and mouse presomitic mesoderm (PSM) – cells destined to form the somites – expressing a fluorescent Hes7 reporter, a gene that switches on and off cyclically as each new segment forms, driving the speed of the segmentation clock. The mouse clock ticked twice as fast as the human clock. They then expressed a NAD+/NADH sensor called SoNar in the cells, causing them to make a protein that fluoresces differently depending on whether NAD+ or NADH is bound to it. This showed that the mouse PSM had a higher NAD+/NADH ratio than human PSM.

They then took some bacterial genes that can modify redox state and expressed them in human cells, and found that when NAD+ was increased, the clock ticked faster, and the PSM cells induced MSGN1 – the ‘master regulator’ of somite development – earlier, linking the NAD+/NADH ratio directly to developmental tempo.

GPD1L: the missing enzyme

How are these different redox states achieved in mouse and human? As I mentioned earlier, the regeneration of NAD+ from NADH is important for continued glycolysis. The cell has three main ways of achieving this: the G3P shuttle (described in Figure 1), a malate-aspartate (MAS) shuttle, and the LDH reaction, which makes lactate without involving the mitochondria.

The authors analysed transcriptomic datasets of mouse and human cells. Genes related to the MAS shuttle and the LDH reaction, as well as GPD2 – the enzyme that takes electrons from G3P – were expressed similarly in both mouse and human. The difference came down to GPD1L, which was found at high levels in mouse, but barely expressed in human cells (Figure 1H in the preprint).

It makes sense, then, that human cells were found to contain less G3P than mouse cells, but the authors still needed to check that this wasn’t a result of differences earlier in glycolysis. They performed two experiments, one with ‘heavy’ carbon and one with ‘heavy’ hydrogen, which labelled the atoms and allowed them to trace the fractions incorporated at each point in the pathway. They found that in human cells, the G3P – but not the DHAP – contained a smaller fraction of both heavy carbon (Figure 2D in the preprint) and heavy hydrogen, indicating the that the conversion step of DHAP to G3P was less efficient, making the G3P shuttle less active in humans than in mice.

Mouse cells burn faster

Embryonic cells, like cancer cells, rely heavily on aerobic glycolysis. Glucose is converted to lactate, allowing for faster ATP production and supplying the building blocks for new cells more quickly, bypassing the mitochondria. But, continued glycolytic flux requires NADH oxidation back to NAD+, and a build-up of NADH can slow this down.

To see whether this was limiting human cells, the authors ran a glycolytic stress test on a ‘Seahorse Extracellular Flux Analyser’. It sounds like something from Back To The Future, but all it does is measure the acidification of the cell culture medium as the cells produce lactate. The mouse cells had a bigger and faster increase in acidification, confirming elevated glycolytic flux. The authors backed this up with transcriptomic data, showing that the mouse PSM expresses higher levels of genes involved in glycolysis.

Speeding up human development

The authors then put what they had learned into practice: could they make human development faster by introducing mouse-like metabolic features?

They first used CRISPRa to activate transcription of the human GPD1 or GPD1L genes in human PSM cultures (Figure 3A in the preprint). This led to an increase in G3P shuttle activation, an elevated NAD+/NADH ratio, elevated glycolytic flux, and importantly, faster segmentation clock oscillations – shortening the clock period by 40-50 minutes (Figure 3D in the preprint). This was accompanied by a faster cell cycle; however, it didn’t change the timing of MSGN1 induction.

Finally, the authors asked whether these findings apply to other cell types. When mouse and human stem cells were differentiated into definitive endoderm or neural progenitors using equivalent protocols, the mouse cells consistently expressed SOX17 (endoderm) or PAX6 (neural) earlier than human cells. However, when human cells were made to overexpress GPD1/1L, this led to cells switching on their lineage markers earlier.

Altogether, this preprint gives convincing evidence that GPD1/1L expression, by controlling a critical step of glycolysis, helps set the rate of embryonic development across multiple tissues.

Tags: developmental rate, embryo, human, metabolism, mouse, redox, segmentation, somitogenesis, tempo