Julian Hitchcock argues that self-regulation may be the best way of regulating the creation and use of models of human embryos that have developed from stem cells.
At the end of her reflection on her front-page Guardian report of Magdalena Zernicka-Goetz’s speech to the 2023 meeting of the International Society for Stem Cell Research (ISSCR), Hannah Devlin made the following remarks about the development of entities assembled from stem cells that approximate human embryos:
"As these models mirror the pathway of natural embryos ever more closely, their scientific value increases. But there is also a growing need for clear ethical and legal boundaries, which do not currently exist. It is crucial that scientists proceed cautiously and that legislators, including in the UK, take urgent steps to catch up."
I'd like to consider the reference to "legislators". In particular, I want to challenge the assumption that for every area in which something appears to be ungoverned, statute is the answer. At least, in one very particular area.
We all do it. Flick the ethical tendon and the "we need regulation" reflex kicks into action. Don't get me wrong: it is entirely right that we should reflect on the governance of areas that may invoke ethical considerations. But governance is not the same as legislation, and in some cases public confidence may be better served by robust, trustworthy systems of self-regulation. I'm very far from being a deregulatory zealot, but a career spent dealing with the regulation of life sciences has persuaded this lawyer that statutes are not always the solution, assuming there's any problem in need of one. In my view, the position of stem cell based embryo models (SCBEMs) is one such case.
Objects in a legal void
Let's get our bearings. First, SCBEMs are not embryos. Open any embryology textbook in the world and look at the canonical first steps of development. Sperm meets egg. The fertilised egg divides into two. The two cells become four. Four become eight. Eight become sixteen etcetera. Because all this cleavage takes place within the gelatinous shell of the egg (the zona pellucida) the cells are increasingly compacted. And it’s not just three-dimensional geometry: the cells are not all the same and, all unseen, the molecular landscape within the cleaving ball is far from uniform. SCBEMs don’t end up looking like real embryos because they aren’t real embryos. Instead of beginning with egg and sperm, researchers place pluripotent stem cells together in a 3D matrix and study how those cells interact with one another as they lay down the foundations of the tissues and organs that keep us alive. Traces of primordial blood are propelled by bunches of cells like early heart muscle. Clusters of early brain cells organise themselves in space. But an embryo it is not.
Second, SCBEMs can never become human beings. Even if they could develop to the extent that, on the evidence of morphology and RNA transcripts from each individual cell, they were point-for-point identical to a human embryo - even then they could not be implanted. Indeed, there is absolutely no interest in doing so, not least because embryos are better at being embryos than knock-offs. Let’s get a little technical: not all SCBEMs are equal. Most SCBEMs don't have the extra-embryonic cells needed for making a placenta, and for normal embryonic development. These have no potential to develop into a foetus, let alone a human being. Some SCBEMs, so-called “integrated” SCBEMs, do include extra-embryonic material. But because they cannot be implanted, and therefore lack the active involvement of a uterus, there is no possibility of developing, Brave New World like, into a human being.
Caring about stem cell based embryo models
So, if an SCBEM isn’t an embryo, why should we care? Because the more successfully we recapitulate the development of a human embryo, the closer we get to the ethical considerations that precipitated the rules for embryo research that are now set out in the Human Fertilisation and Embryology Act. In its thoughtful editorial (18 June 2023) on SCBEMs, the Guardian concluded that:
“The birth of Louise Brown, the world’s first IVF baby, led to the British government setting up the Warnock committee to investigate whether the technology was acceptable and how it should be regulated. Perhaps it’s time for a new panel to convene and find an ethical consensus.”
The Guardian editorial framed the issue as one of public trust in science, which had increased with the Covid-19 pandemic, warning that science that might be seen as dehumanising could undermine this trust, and asking ministers to “be alert to the forces and practices taking us down this road and vigilantly steer another course”. Sure. However, the example it cites is not that of SCBEMs, but of Hank Greely’s suggestion that “most people in developed countries will cease reproduction through sex, using sex exclusively for pleasure, and instead will rely on reproduction through pluripotent stem-cell-derived gametes". Now I am a great admirer of Professor Greely, but I have to respectfully disagree on this one. I mean, it’s intellectually interesting and all that, but is it actually going to happen? Bioethics benefits from thinking through extreme examples like this, but we must keep our feet firmly on the ground. SCBEMs are here and now. They are not going to grow up and shake you by the hand, but they still have profound human significance. If "the proper study of mankind is man" (Alexander Pope), then SCBEMs are undoubtedly part of that project of self-understanding. However, as a former Lord Chancellor and the founder of modern scientific method observed, “Science discovery should be driven not just by the quest for intellectual enlightenment, but also for the relief of man’s estate”. Bacon was talking about public funding for science, not restrictive rules, but the point remains relevant to SCBEMs.
Relieving Man’s Estate
So, do SCBEMs have the capacity to relieve human suffering?
The answer is “yes”. Park the philosophy. Park the fact that SCBEMs provide an alternative, good-enough platform for a field of biomedical research that is hampered by a dearth of human embryos. You can even park the obvious opportunity to understand why pregnancies fail or succeed. Think instead about the development of therapeutic interventions. To do this, we have to understand the earliest stages of disease development. Unfortunately, human studies rarely capture this and, with limited opportunity for experimental intervention, it is hard to gain insights into what may be happening before the pathology makes itself known. We owe a lot to mice, but our dependence on animal models relies on evolutionary equivalence: the presumption that the biological pathways involved in human disease are shared despite our evolutionary divergence. Given that the last common ancestor of mice and humans died before the Chicxulub meteorite ended the era of non-avian dinosaurs (at least 65 million years ago) you can see the issue. So the problem is this: with such limited ability to recreate human-specific processes in human or animal subjects, how can we bridge the gap? Organoids are beginning to do so, and SCBEMs may allow us to do it better.
Organoids are in vitro cellular systems derived from stem cells or progenitor cells that co-develop, through interaction and self-organisation, in a manner approximating development in vivo, leading to a 3D cell architecture, comprising multiple organ-specific cell types, approximating in vivo-developed tissue that recapitulates specific functions of the in vivo organ (e.g. contraction, neural activity, endocrine secretion, filtration, excretion). If you try to develop a kidney or lung organoid, your focus is on redeveloping that organ as best you can in 3D. Advances are impressive: though miniscule, organoids are becoming more histologically relatable to the real thing. However, no-one should underestimate the histological complexity of organs such as the kidney or lung. Organoids may be good enough for many purposes, and have already succeeded as functional animal grafts, but laboratory organogenesis simply can’t produce organoids that are as organ-like as real organs. That is because actual organs do not develop in isolation, but as part of an embryo. By contrast, embryos gets organogenesis right every time, unless there's a problem. To understand healthy or problem organs, you might think about an SCBEM approach.
For most purposes, you can think of SCBEMs as “body-oids”. By applying the same general “organoid” approach (of using emergent properties of cells to elicit the development of functional tissue) up to the level of the whole body, it should be possible to derive organoids from primordial SCBEM tissues that are more likely to continue their development in a way that a kidney, lung, gut or heart would do naturally. If that's correct, then the result should be organoids that are more histologically and functionally equivalent. Better organoids strengthen the bridge between human studies and animal models, providing a platform from which to develop new diagnostics and therapeutics; even, perhaps, better material for clinical transplantation.
SCBEMs, then, have the capacity to relieve “man’s estate”. If we block their creation or use, humans will not have access to healthcare benefits that would otherwise accrue. Equity of access is an important issue, as with other advanced biomedical technologies, but equity is not served by prohibition.
Still, the fact that SCBEMs pass Francis Bacon’s test of a good technology does not mean that some may not consider their development or use to be, as the Guardian worries, “dehumanising”. It is important, in developing rules for their governance, that we listen to those with ethical concerns that we may not share. What if SCBEM-developed tissues felt pain, for example, or neural organoids learnt to process natural language? This is extraordinarily unlikely, but it reasonably falls within public contemplation. What if SCBEMs were used for testing chemicals or cosmetics? This is not an ethical slam dunk. As Hannah Devlin and the Editor of the Guardian suggest, it is important to take public concerns seriously. My question is whether the correct response to such concerns is a statute or a trustworthy, transparent governance system.
Legislation or governance?
Now you might object that statutory regulations are trustworthy and transparent. Surely, you might say, the only way to achieve public confidence is by means of an Act of Parliament or a regulation made under the authority of such an Act. I'm not so sure. There are reasons to be cautious about the regulatory reflex. I not only think that a self-regulatory approach can be used to regulate the development and use of SCBEMs; I think it may very well be the best way.
The first reason is that statutory regulations may be made without public accountability. Hannah Devlin's news story was serendipitously knocked off the front page by a story about the UK legislature: the improper use of it by an ex-Prime Minister. Less reported by the media is an approach to law-making that constitutional lawyers loath, and which has become a commonplace in recent years: the Henry VIII clause. This is a provision of an Act of Parliament that grants ministers powers to escape the scrutiny of Parliament. In effect, MPs are whipped into ceding their legislative power as democratic representatives to ministers, now and in future. The more power MPs grant to ministers, the less ministers are accountable to the public.
As it happens, delegated powers are genuinely important. Consider, for example, the approval of a new medicine or medical device. MPs are not fit to undertake technology assessments, but the Acts of Parliament they’ve approved provide mechanisms for experts to do that work on the basis of demanding technical standards. However, when non-technical matters are delegated to non-expert ministers, we might have pause for concern. Don't worry about the details: the minister will sort it out.
At the other extreme, a delegated power may compromise the public interest by involving MPs. Supposedly the poster-child of delegated reproductive regulation, the powers under the 2015 Mitochondrial Donation Regulations, are a case in point. If a minister wishes to permit an improved process for implementing mitochondrial transfer, they must secure the consent of both Houses of Parliament. Parliamentarians, who lack technical expertise and had anyway already agreed the purpose of any such technology, are not given enough Parliamentary time to do the technical assessment. It’s not a government priority, so it doesn’t get listed. The regulatory framework doesn't respond to scientific advance, and the effect is that improved treatments remain inaccessible.
Back to SCBEMs. They occupy the twilit world between the technical and the ethical. They raise a surprising constitutional question. What are the limits of Parliamentary competence over matters of scientific ethics? The answer appears obvious at first: Parliament is supreme. Period. But we have already seen that Parliament can delegate technical matters in the public interest. Is there a public interest in delegating ethical matters arising from complex matters of science? Much will depend on how, and to whom, that power is delegated and the standards of governance under which it is exercised, but I think the answer is “yes”.
Let’s compare the world of the early 1980s with that of the early 2020s. Specifically, let’s look at the deliberations of the Warnock Committee: the group appointed by Parliament to consider the kind of issues we’re looking at today. It was the Committee's 1982 report that laid the foundations for the UK’s Human Fertilisation and Embryology Act 1990. The report is exemplary and its considerations were exhaustive. However, seen with the benefit of 41 years of, frankly, amazing science, the Committee's vision of embryology inevitably appears myopic. Its recommended prohibition on cultivating embryos beyond 14 days was a pragmatic compromise* designed to balance the interests of different groups, but was inevitably predicated by the limited scientific understanding of the time. Fourteen days was the best guess for the last point at which individual (as opposed to multiple) bodies could develop from a single embryo. Whatever one thinks of the prohibition, the rationale is moored in the morphological. Today, we can look at other aspects of development, most notably, the genomic. In particular, we can look at the genomic activity of individual cells of the developing embryo or, for that matter, SCBEM. These and other advances enable us to compare embryos and SCBEMs, not just in terms of bits that look like other bits, but in terms of what cells and architectures are doing in four dimensions. If we apply the same reasoning as that of the Warnock Committee, then we might find that the last point at which individual bodies could develop from a single embryo arises earlier than 14 days. It wouldn’t change the law - that is firmly pegged to the morphology of the embryo - but what it does do is challenge us to reconsider the basis of our future laws. In particular, it asks us to consider the ethical consequences of particular genomic activities, natural and induced. It even asks us to look beyond the embryonic frame of reference to, say, the use of brain organoids. And because any number of potential genomic interventions may be used to understand development in health and disease, it is neither reasonable nor practical to expect MPs and Members of the House of Lords to appraise them all and their ethical consequences, let alone to screw into law statutes which are destined, by the nature of advancing science, to become outdated before the ink has dried.
This presents a problem. Who is going to set the rules for SCBEMs? Members of the public are unlikely to be too interested in the technical details of 3D culture systems or scRNA-seq analyses, but they may be reasonably concerned about work that mimics the subject of stringent regulation. For some, SCBEMs might look like a loophole technology: something prescribed by a lawyer like me to get around the law. They are not (and I didn’t): at the risk of repeating myself, SCBEMs are not embryos. Even so, if something that isn’t a human embryo is intended to model key aspects of one, the public might expect it to be treated with a degree of responsibility and seriousness, and to include any red lines not already existing as a matter of law. How, then, are we to establish rules to govern the development and use of these embryo lookalikes, who is going to do the governing, and why should we trust them?
The scientific community often looks back admiringly to the example of the Asilomar conferences on molecular cloning (“genetic engineering”) in the mid-1970s. This was when, with no prompting from government or any other quarter whatsoever, scientific and legal heads banged together to forge biosafety rules to contain and limit the use of recombinant technologies. It remains a great achievement of scientific responsibility, not least that of the late Paul Berg. The fact that it produced a consensus statement setting out containment rules for different categories of risk is remarkable. But the scope of Asilomar was narrow, and it is certainly not a paragon of self-governance. In the intervening years, the characteristics of a good scheme of biological self-governance have become better recognised. At its broadest, its aim must be to prevent misuses of biology in the field of research, while being responsive to scientific development. It must uphold the rule of law, and be both impartial and transparent. It must also be participatory, in the sense that it must be informed by the views of scientific, academic and research institutions, ethics organisations, leading scientific journals, regulators, and funders, including those providing public funding. And it must be both effective and, by virtue of its setting within publicly-funded institutions, publicly accountable.
Can the UK achieve this for SCBEM research? I think it can. It is important to have a sense of proportion. SCBEMs do not pose a biohazard risk. Nor are many SCBEMs or organoids of concern likely to arise. For this reason, the rules and their implementation should not and need not be burdensome. However, insofar as there may be public concern, it may be reasonable to expect there to be a trustworthy scheme for the development and use of this unique research platform.
As it happens, for the present the only way to control the creation and use of SCBEMs is by means of self-governance. There just isn’t room in what remains of the current Parliament to introduce, let alone debate, a Bill. This puts UK’s scientific community on its mettle. Even before the announcements at last week’s ISSCR meeting, the University of Cambridge and Progress Educational Trust had established an independent SCBEM working party to produce guidelines and a governance framework for SCBEM research. However, the media response to the announcements of Magdalena Zernicka-Goetz, Jacob Hanna and Jitesh Neupane accelerated the project’s public launch, with its Chair, Roger Sturmey, giving an interview on the Today programme (at 0:23:14) and editorials appearing in the Times, Guardian and elsewhere. The working party aims to produce a draft version of the SCBEM guidelines by the end of this year. A future Parliament may choose to bring SCBEM rules under the wing of government, but for the present the public should be confident that those minded with the responsibility of developing those rules are tackling them with the seriousness it would expect.