The events within a crucial four-week period of human development have long eluded scientists. Stem-cell models are finally revealing them – and raising new ethical questions.
Every human life begins as a single cell, formed in a brief, dazzling moment of fusion as sperm merges with egg. Already, this microscopic sphere, no wider than the thickness of a strand of hair, bears the full set of instructions for a unique human being.
Within hours, the cell begins to divide. First into two, then four, then eight – a rhythm of replication that continues with clockwork precision.
After three days, the dividing cluster resembles a microscopic blackberry. And remarkably, despite their genetic uniformity, the cells begin to behave differently: some tuck inward, others spread to the surface. At around five days, fluid is pumped between them, creating an internal cavity.
Inside this delicate, fluid filled sphere, a small cluster of cells gathers at one side, out of which the embryo proper will grow: every organ, tissue, and feature. The surrounding layer begins preparing for another essential step: burrowing into the lining of the mother’s womb to establish the placenta that will sustain it.
Over centuries, researchers have gradually worked out how to watch these events unfold – the grainy opening frames of the movie of life, played on repeat.
But even with today’s techniques, there’s a gap – a crucial section of the movie that’s particularly blurry and hard to decipher.
By two weeks, “we’ve learnt a fair amount about how life’s complexity begins to emerge,” says Naomi Moris, who leads the Crick’s Developmental Models Lab. By this point, the structure resembles a “double bubble”: an upper sphere that will ultimately form the amniotic sac, and a lower yolk sac, with the embryo forming a thin layer of cells sandwiched between.
Fast forward to six weeks, and so too have researchers begun to unravel how a tadpole like fetus, no bigger than a lentil, goes on to develop into a human baby.
Between these points – weeks two and six – a mysterious and breathtaking metamorphosis occurs. Somehow, that flat layer of cells folds and thickens, forming three layers in a process called gastrulation, before curving into the recognisable outline of an embryo.
Yet precisely how this transformation happens remains shrouded in mystery – a “black box” of human development. “We can piece together what might happen, for example from studies of animal embryos,” she says, “but for humans there’s so much we don’t know.”
Understanding these missing moments could reveal clues to infertility, miscarriages and a better understanding of congenital conditions. But there are huge challenges in doing so. Technical, legal and ethical limits prevent donated human embryos, left over from IVF procedures, from being grown in the lab beyond the two week mark.
Meanwhile, embryos from terminations – the mainstay of much historic research – are rarely donated before a pregnancy’s sixth week and almost never before its fourth.
So, in search of answers, researchers like Naomi are using the incredible self organising ability of stem cells to rerun and decipher the events within the ‘black box’. Could these so called stem cell based embryo models further prise it open, and let us see the missing scenes in our own story? And what ethical considerations might arise as they do?
Naomi Moris in her office at the Crick. Behind her hang historical images of early embryos. Credit: Dave Guttridge.
Embryology dates back to ancient Greece, when Aristotle cracked open chicken’s eggs to reveal that embryos grow in complexity over time – a concept he called epigenesis. But only after the invention of the microscope could scientists watch life begin from a single cell. Even then, the technical and moral challenges of working with human embryos confined such studies to laboratory animals such as worms, fruit flies, frogs, fish, chickens and mice.
“What we’ve learnt from animal studies has been incredible,” Naomi says – not least in developing and refining experimental techniques. “The mouse has been particularly important since, like humans, it’s a mammal – it has structures like a placenta to support the development of the embryo.”
Yet differences remain between mice and men. Early mouse embryos resemble a cup shape, while ours begin as a flat disc. Genes switch on at different times. The two species have radically different placentas. And a mouse pregnancy lasts just 20 days. As the saying goes, ‘all models are wrong; some are useful.’ So, researchers have needed ways to model ourselves directly.
For decades, human embryology relied on collections of thousands of embryos, often gathered post-mortem or after hysterectomies. One of the most famous, assembled in the early 20th century, is the Carnegie Collection in Washington, DC, where US researchers amassed thousands of donated embryos, preserved and photographed in fine detail – the first windows into how our organs and tissues form.
Their limitation, Naomi notes, was that they only allowed observation, not experimentation. “They’ve been great for cataloguing developmental stages,” she says, “but they don’t let us probe what controls progression between them.”
Much changed with the arrival of in vitro fertilisation in the 1970s. IVF built on decades of animal work: first achieved in rabbits in 1959, by the early 1970s researchers had grown human embryos to five days in the lab, paving the way for the birth of Louise Brown in 1978.
But IVF wasn’t just a fertility breakthrough – it was a turning point for developmental biology, enabling direct study of early human embryos. And the society wide ethical discussion it kickstarted led to the UK’s influential Warnock Committee of 1982.
After careful consultation, the Committee ruled that suitably credentialled scientists could study donated IVF embryos for up to 14 days – a point, pragmatically chosen, after which an embryo can no longer form twins (and so can be regarded as the beginning of individuality), but before its nervous system forms.
After this point, the study must conclude. This became UK law as the 1990 Human Fertility and Embryology Act, and many other countries subsequently adopted versions of what became known as ‘the 14-day rule’.
Naomi’s colleague at the Crick, developmental biologist Robin Lovell-Badge, witnessed this process first hand as his mentor, Warnock Committee-member Anne McLaren, inspired his career-long engagement with ethics and science policy.
“In the UK, research involving human embryos is regulated through the Human Fertilisation and Embryology Authority (HFEA),” he says. “It’s proven to be a very robust mechanism.”
Even today, no one has cultured a human embryo beyond 13 days, albeit as much for technical as for ethical or legal reasons: after a week, a human embryo normally implants into the womb. An isolated lab-grown embryo starts to diverge after this point, making it hard to trust experimental results. As a result, using this method to study the events in the ‘black box’ has been largely impossible.
But if the arrival of IVF embryos was one revolution, another was also brewing: stem cells.
First identified in mice’s bone marrow in the 1960s, stem cells can morph, or ‘differentiate’, into another type of cell as they divide. Those found in specific tissues help the body grow and repair itself, but these adult stem cells can only replenish the tissue in which they reside.
Then, in the early 1980s, researchers discovered cells in early-stage mouse embryos that had an innate ability to renew themselves indefinitely, and also differentiate into any cell type – a phenomenon called pluripotency – and by 1998, researchers had created the first human embryonic stem cell line – opening up entirely new avenues of research.
“We could finally study these cells and explore what makes them tick,” says Naomi. Soon, researchers began to reveal the intricate choreography of gene activation and cell communication driving early development.
Then, in 2007, another breakthrough: scientists discovered how to reprogram ordinary adult cells into pluripotent ones. These ‘induced’ pluripotent stem cells didn’t come from hard-to-obtain IVF embryos, so researchers didn’t need an HFEA licence to create them. It democratised stem-cell research, rapidly spreading new methods around the world. And since they could be derived from patients with inherited diseases, this field, in particular, was transformed.
Soon, scientists had discovered chemical cues that pushed stem cells to specialise in different ways, and were coaxing them into miniature 3D tissues. Called organoids, these miniature lab‑grown replicas of organs such as the gut, brain and kidney have become invaluable tools for studying conditions such as Alzheimer’s and cancer.
As organoid research flourished, researchers began to wonder: what if, instead of steering stem cells towards organs, they were allowed just to run their innate developmental programmes? Could a more ‘hands‑off’ approach replay life’s opening moments – and further illuminate the black box?
“The embryo model field, and the techniques we use, grew out of the organoid field,” says Naomi.
The first basic models were developed around a decade ago. In one landmark study, researchers in New York persuaded human embryonic stem cells to self organise into concentric rings representing the three layers characteristic of gastrulation.
Around the same time, a team in Cambridge grew mouse cells into what they called ‘gastruloids’ – tiny, self organising, 3D structures that spontaneously ‘broke symmetry’, developing a left, a right; a front and back; a top and bottom.
Soon, researchers around the world were adapting these techniques to model a range of developmental processes: the field of stem cell based embryo models had been born.
“Essentially, to make an embryo model, we take pluripotent stem cells growing in a flat layer, lift them up and dissociate them,” says Naomi. After adding chemicals to help them survive, “we put a small number – about 400 – in a little round-bottomed well, and, with the right conditions, they clump together... and then we let them do their thing.”
By tweaking the conditions, or number or type of cells, researchers can replay different sequences from within that missing four weeks. Hundreds of variations of these models now exist, capturing processes ranging from pre implantation development to early organ formation – all driven by the cells’ own self organisation.
At the Crick, Naomi’s lab focuses on two longstanding biological mysteries: how the cells that form our sperm and eggs – germ cells – first arise; and how the embryo develops its segmented backbone – a process that goes awry in certain congenital conditions.
“Primordial germ cells appear very early,” Naomi says. “In a sense, they’re the most important cells in the embryo – the only ones that pass life on.”
As they develop, these cells migrate along the embryo’s nascent gut tube, or endoderm, to its gonads. Here, they somehow ‘read’ signals coming from gonad cells and – depending on whether they’re destined to be testes or ovaries – develop into sperm or eggs.
Naomi is using mouse gastruloids to probe how germ cells arise, migrate and interact with surrounding tissues. “It’s a very simple model, and develops over time in a similar way to a mouse embryo,” she says. Crucially, they’ve shown that gastruloids spontaneously generate early germ cells, which then interact with endoderm tissue, and this is helping them identify the molecular signals involved.
As is so often the case, the mouse is a staging post to probe human biology. “It’s not at all clear yet where our own germ cells come from, nor how they mature,” she says, “so we’re laying the foundation in the mouse before making the jump into human models”.
Ultimately, such insights could transform fertility treatment – perhaps even bypassing today’s invasive egg-retrieval procedures. “Could we one day simply make hundreds of eggs from someone’s skin cells?” she muses.
A 'human trunk-like structure (hTLS)' – an embryo model used to study how the spine and other features develop. Different cell types are tagged with coloured antibodies, revealing their identities. Credit: Naomi Moris.
In Naomi’s office hang three historical paintings of early human embryos – a reminder that developmental studies have advanced human health in tangible ways, from IVF to the discovery that folic acid prevents spina bifida. Future benefits, Naomi says, could include insights into why miscarriages occur, and into rare diseases, and congenital abnormalities such as heart defects.
Ethical questions are never far from the discussion, yet public attitudes evolve when the medical stakes become clear – IVF faced opposition until it helped infertile couples; stem-cell research gained acceptance as its therapeutic promise emerged. Effective regulation, born of well-conducted dialogue between scientists and the public, has provided confidence that research is ethical, while also helping guide researchers’ work. And broadly speaking, Robin and Naomi agree that, so far, the UK has got regulation about right.
“The 1990 HFE Act has had several updates – one in 2001 to cover stem cell and cloning research,” says Robin, “plus a major 2008 overhaul to extend the types and purposes of research that could be undertaken, including genetic alterations, as well as changes around clinical treatments and patient care.”
But recent advances, such as embryo models, mean another update is needed. There is a governance gap, Naomi points out. “While human embryo research is very heavily regulated, there’s currently no legislation, anywhere in the world, specifically for embryo models – and no legal reason to stop experiments reaching stages that might raise ethical concerns.”
The principle underpinning the 14-day limit in embryos, she says, has been hugely helpful: it set a publicly agreed, biologically relevant marker that was – at the time – far ahead of what was possible technically. “It stopped the clock before you have to worry,” she says.
For embryo models, the solution so far has been voluntary self-regulation through guidelines, including those produced by the highly respected International Society of Stem Cell Research. These are broadly agreed to be well thought-through and reasonable. However, such is the speed of progress, that they too have struggled to keep up. A 2021 update covering embryo models was followed swiftly by another in 2025. Nevertheless, these guidelines are broadly maintaining ethical standards.
“Although they have no teeth in law,” says Robin, “most scientists want to be respected by their peers, while journals and funders can insist that you follow guidelines.”
Recently, both Robin and Naomi have proposed new frameworks for regulating embryo model research, guided by ethical principles, including that no model should ever develop the ability to experience sensory input such as pain.
It is clear, however, that the UK’s HFE Act needs another overhaul. “Science moves much faster than we can change the law, so the Act desperately needs to be updated,” says Robin. "Policymakers I talk to are definitely engaged and very willing to take this all on, but we’re a bit frustrated at the pace, because it first needs to be opened up by government, and currently their attention appears to be elsewhere.”
Naomi agrees: “No scientist wants to push beyond the boundaries of acceptability,” she says.
That said, she is very clear about what stem cell based embryo models are not. “These aren’t yet embryos,” she says. In her own research she uses relatively simple models that would never be able to grow into a fetus. Even more complex models are, at the moment, a long way from being able to develop further.
As a result, discoveries made with current models need careful validation in real embryos to be sure they’re genuine discoveries, rather than artefacts of that particular system. As Naomi says, quoting German embryologist Viktor Hamburger, “the embryo is always right.”
But this need for validation focuses attention back on the 14-day rule, and to embryos themselves, which researchers are gradually learning how to keep growing in the lab with greater fidelity.
Consequently, many – including the HFEA itself – are starting to consider whether the 14-day limit itself needs to be carefully reconsidered.
“So far, the public broadly supports both embryo models, and extending the 14-day rule – with some crucial caveats,” says Robin.
Postdoctoral research scientist Ignacio Rodriguez Polo imaging a 3D embryo model using advanced microscopy techniques. Credit: Dave Guttridge.
“Researchers need to be asking important scientific questions that we can’t answer by other methods – and we need to use as few embryos as required to give statistically valid results.”
Both experts agree that, thanks to public engagement work to date, researchers in the UK have the benefit of strong public backing. Retaining this, through effective dialogue and regulation – particularly with harder‑to‑reach groups – is vital as science advances.
A key point for Robin is that any future regulatory set up retains as much flexibility as possible. “We don’t know where the science will lead, nor how fast – but what’s certain is that progress will happen faster than the ability to legislate,” he says.
Clearly, there are some big, society level discussions to be had. But do we have the tools and understanding to make these decisions?
Reflecting on her own experiences of pregnancy, Naomi recalls apps comparing the size of a growing baby to an orange, or a watermelon, and the anatomy of the embryo at those stages – but never the early disc of cells as it folds into human form.
“It does make me wonder whether people currently have the tools to understand what an embryo is at 14, or 20, or 25 days,” she says.
As researchers gain unprecedented power to model these hidden stages, public engagement will need to catch up. The science is racing ahead. The question is whether the rest of us can keep pace.
