In vitro gametogenesis and reproductive synbio

In vitro gametogenesis takes a somatic cell, reprogrammes it to pluripotency, and then walks it forward along the germline route: first a transient mesoderm-like state, then primordial germ cell-like cells induced largely by BMP signalling, then a long maturation inside a reconstituted gonadal environment. In mice this route has been completed. Cultured cells have produced functional oocytes and sperm that yielded fertile offspring, and the female side has been run entirely in a dish. The mouse result is what makes the field credible; the distance between mouse and human is what makes it hard.

The epigenetic reset

The germline is the only lineage that erases and rewrites the genome’s methylation marks wholesale. After specification, primordial germ cells undergo genome-wide DNA demethylation to a few per cent CpG methylation, which wipes parental genomic imprints and, in females, reactivates the silenced X chromosome. Sex-specific imprints are then laid down again during gametogenesis proper — maternal marks during oocyte growth, paternal marks in prospermatogonia.

This is the true difficulty. An imprinting error is not a growth defect visible in culture; it is a silent lesion whose consequences appear in a developing embryo, and the disorders caused by imprinting failure at specific loci are well documented. There is no non-destructive assay for the epigenome of the single cell you intend to use. Any characterisation kills the gamete, so quality is inferred from siblings in the same batch. The only complete test of the product is offspring.

Why humans are not slow mice

Human germ cell specification runs on a different transcriptional circuit from the mouse — SOX17 and PRDM1 rather than the mouse’s SOX2-dependent programme — so induction protocols could not simply be transferred, and were rebuilt. Timing is the second gap: human germ cells take months in vivo where mouse cells take days, and cultures must be sustained on that scale. Human primordial germ cell-like cells can be induced reliably and, in long co-culture with gonadal somatic cells, progress to an oogonia-like state with much of the demethylation accomplished. They then stall.

They stall before meiosis, which is the third barrier and a genuinely different kind of problem. Meiosis begins with deliberate double-strand breaks made by SPO11, which must be repaired by recombination between correctly paired homologues held by the synaptonemal complex. Surveillance checkpoints eliminate cells with unrepaired breaks or unsynapsed chromosomes. These mechanisms are protective — they are why gametes are not routinely aneuploid — and in vitro they behave as a filter that few cells pass. Driving cells through meiosis is not the goal; driving them through with correct segregation and recombination is, and forcing entry without satisfying the checkpoints produces exactly the aneuploidy the checkpoints exist to prevent.

Where the limit is not technical

Even a competent human gamete would meet a regulatory boundary independent of its quality. Heritable modification of the human germline is prohibited or unapprovable in many jurisdictions, and embryo research is time-limited under rules such as the UK’s statutory fourteen-day limit. Clinically nearer-term work uses the same somatic-support biology to mature oocytes rather than to create them — a real application, and a different claim.