Models are important tools: they resemble, they mimic, they imitate something to a greater or lesser extent. How similar models are to the 'real thing' is usually a challenging issue. And it's a big issue with stem-cell derived models of the human embryo.
These embryo models, models of the embryo's 8-cell stage, of the blastocyst or of the gastrula are emerging and they are ones that labs can use to characterize the molecular and physiologic events that take place during early embryogenesis. My story in Nature Methods about some of these embryo models is here.
For this story, I spoke with Christine Mummery , a researcher in the anatomy and embryology department at Leiden University Medical Center.
You can listen to the podcast here, it's also on Google podcasts, Apple podcasts, Spotify and other podcast streaming platforms in a series called Conversations with scientists. A transcript is pasted below.
Note: These podcasts are produced to be heard. If you can, please tune in. Transcripts are generated using speech recognition software and there’s a human editor. But a transcript may contain errors. Please check the corresponding audio before quoting.
If I'm claiming this is a liver cell, what does it have to show? And this is a tricky, tricky thing.
That’s developmental biologist Dr. Christine Mummery, who is in the anatomy and embryology department at Leiden University Medical Center in Leiden, The Netherlands. You will hear more from her about her in a moment.
Hi and welcome to Conversations with scientists, I’m Vivien Marx. Just briefly about this podcast. for my stories, I interview scientists around the world and these podcasts are a way to share more of what I find out. You can find some of my work in Nature journals and elsewhere. I wrote a story recently about embryo models, that means having a model of the human embryo in the lab. Dr Mummery has plenty to share on this subject.
I like to ask scientists about their research and also the impact of COVID-19 on their work. Here’s Dr. Mummery commenting on that
Christine Mummery [1:01]
It's really hit the lab quite hard. Particularly the ones where we, you know, within the hospital, so we're within a hospital, and we have to regard patients. So we are still quite sharp on access. We're not allowed to have visitors from outside, you know, things like that. Not allowed to have large meetings. And for a long time, we've been pretty masked up and closed down. So we are less open than the rest of the, let's say, the community. And I think that's true for all university hospitals.
So, I mean, we have eight in the Netherlands. So they're all pretty much the same. Because the last thing you want is to have cross-contamination. So some of our, let's say, patients that you know, on a drip or something, they go to the general canteen to see some other surroundings. But if we go there without our masks and stuff like that, it's not a good idea. So yeah.
And aside from that, the trainees have been very stressed about their project's time ticking away, without them producing the papers. In some cases for a thesis, you have to have submitted a paper or more. So they've been really worried. We've had a few on sort of, also postdocs edging on burnout, or actually gone on burnout. We've had PIs feeling the pressure to report stuff on grants, things like that. We've had, we have a couple of cases of long COVID in the lab. Nobody knows where it comes from. These are people who are normally sporty, robust, fit, mentally extremely stable. So it's all a bit challenging.
COVID is hard on everyone. For researchers there are all kinds of clocks, a thesis that has to be finished, publications and grants, the tenure clock. None of those clocks readily stop due to COVID. It’s difficult for sure.
And, of course, doing science is difficult, too. So back to developmental biology.
When sperm meets egg, well that’s a biological fact of life, sometime,s but not always development continues.
How a fertilized egg continues along its developmental path, that’s a process that could be intriguing to watch in detail and informative, too. But, of course, you can’t just watch.
Sure, at in vitro fertilization clinics around the world the process of very early development is watched. A fertilized egg becomes a multi-celled blastomere, than a few days later a blastocyst when it might be implanted into a person’s uterus, someone who is trying to become pregnant. And development continues on to gastrulation stages and beyond.
Once inside the person’s body, the process of development cannot be studied in detail. And when fertilization happens inside a person’s body—you know the classic way--these things happen then it cannot be watched from even those very early stages.
Given that there are, for example, disorders that occur during early human development, scientists would like to better understand when things go right and what is happening when things do not go right. But of course it is not technically or ethically ok to experiment on an embryo in a person’s body.
So labs have begun using stem cells to study early human development. These are cells that might start out as skin cells. In the lab, the cells are cultivated and then treated in different ways to turn back time.
The cells are then no longer specialized cells, like heart cells or liver cells or muscle cells.They can become any number of cell types, they have been reprogrammed to a time early in development.
And then when they become stem cells this way, they can be handled and treated in such a way that they can be coaxed to be like a blastocyst for example. They can mimic early development. It’s really important to say these are not real embryos they are models. They resemble, they are similar to, for example a blastocyst. Here’s Christine Mummery on that.
Christine Mummery [5:20]
Saying similar to is of course much easier to saying identical to. I mean, a blastoid-like structure, whatever you want to call it, is never identical can never be identical to a blastocyst, because it's not been derived by fertilization. But at what point does it get so similar, that it's indistinguishable from a blastocyst, is the point. And I would say in general, some are a bit closer than others, or some have features that resemble blastocyst extremely strongly, but lack other features. And in other models, it might be the other way around.
So there's a lot of different features that actually say this is a blastocyst and some of them we don't even know what they are in humans. So all they can really say is: the structure we've made is as similar as possible in these features to a real blastocyst as far is as known.
A number of labs have developed blastocyst-like structures, sometimes they’re called blastoids. Discussions are ongoing about which tests are needed to establish exactly how similar a blastoid is to a blastocyst.
Researchers told me they might look at shape, cell number, they might look at an aspect such as check which genes are expressed using techniques such as single cell transcriptomics. It might look like a blastocyst in some ways but not many.
Or it might look and test like a blastocyst in all kinds of ways, which is when it might even be indistinguishable from a human blastocyst. Ii this context the guidelines of the International Society for Stem Cell Research come into play.
Christin Mummery offers some examples and some thinking on this and this is also about the updated ISSCR guidelines.
The link to these guidelines is in the show notes, which also has a transcript. https://www.isscr.org/policy/guidelines-for-stem-cell-research-and-clinical-translation
These guidelines as lab work on what one needs to measure and show, to determine how similar a blastoid is to blastocyst, or a gastruloid is to a gastrula. Scientists want to establish this in order to figure out how well their embryo models might model a process they are trying to study.
Christine Mummery [7:45]
I mean, that's sort of one of the important points, it’s a bit missing in the literature. So ironically, just yesterday, we submitted a paper a review, where we actually asked this question. To what extent are quantitative measures that you have an in vitro system, to what extent are they similar or the same as real values you measure in a real tissue. Just and just by very nature of what you can measure in vitro versus in vivo, nine times out of 10, they don't match.
So if I would look, if I would say, like, I'm looking at the barrier function of the vascular wall of a vein, or an artery. And what I measure in vitro, is extremely difficult to measure in vivo, in living animal. But there are other features, the way we do diagnostics in patients, there can be, let's say, secreted products in the bloodstream, at a certain level, which say, this person is having a myocardial infarction.
Now, if I see that same set of markers popping up in my model, which I stressed, I say, Okay, this model is mimicking a myocardial infarction. And then the question is, what about the level of those markers? Is it the same per millimeter of fluid? Or microliter whatever you want to unit you want to use? Or is it 10, or 100 times lower, or yeah, bigger. So then it gets very hard because you're talking then about the size of an organ, versus the amount of blood in the bloodstream. So what is needed is often a set of triggers, which gives you the relative increase in response to trigger X, Y, or Z.
And this is exactly what we we've been trying to figure out what data is out there that we could use to benchmark our models? What is what are we claiming about, let's say, the electrical behavior of the heart, which is measured entirely different ways. But the response to certain drugs is comparable. So there's a whole sort of list of benchmarking criteria.
And it's also what the ISSCR is working on. To try and establish standards. If I'm claiming this is a liver cell, what does it have to show? And this is a tricky, tricky thing. And what it has to show may only be what's necessary for your particular assay. So it has to, you know, if the liver it has it has to show P450 expression, because that's a measure of toxicity, for example.
If it's measuring that, and it's responding to known liver toxic compounds in the way you'd expect, then it's fit for purpose. If you're looking, let's say to transplant a liver cell into a patient, the question will be different. How close does it have to be to native tissue to be integrated into the organ?
So there are different questions. And again, is my model fit for purpose? So if you're looking at blastoids the ultimate test is: can it develop into an embryo, and if it can, then by definition, it is a blastocyst. But for humans, those experiments are not allowed, according to also ISSCR guidelines. So if you did those experiments in a mouse, and if you got a fetus and a healthy newborn, then adults then you could say, my model have a blastoid in mouse mice is the same as a blastocyst in every respect.
People like to make the analogy. Okay, my mouse blastoid look looks like my human blastoid. Therefore, it is a blastocyst. But it isn't. Because we don't know what, you know, it should look like in humans at those stages beyond implantation. So we're only beginning to get that information. What is it? Does it actually go through a gastrulation process? And so, the reason, the ISSR categorized blastoids in a higher, ethically risk group, then of course, gastruloids, because with blastoids, you can potentially add in the trophectoderm, which would make it all the cellular components of a blastocyst. So you're very close, actually, to having an embryo.
The updated guidelines from the International Society for Stem Cell Research are out and definitely worth studying even if you are not a stem cell biologist. As Christine Mummery and her co-authors point out in a paper about the guidelines, the new guidelines recommend research be permitted on all stem-cell based embryo models.
Previously embryo models could be used up to 14 days if they had organismic potential, which can be difficult to define. What will replace this is a case by case review in some categories. Integrated models such as blastoids, which have all the cell lineages that can sustain embryo development have to be seen in a different review category for scientific and ethical review. Non-integrated models that do not need this kind of review are for example gastruloids that only mimic certain aspects of human embryo development and usually do not have the cell lineages that would sustain development. And it’s perhaps too late in development for a gastruloid to get the tissues it would need. Please consult the guidelines on any of this in detail of course. And I asked Christine Mummery about this, also how gastruloids differ from blastoids.
Christine Mummery [14:10]
A gastruloid looks more like an embryo. But it doesn't have any of the trophectoderm or extra embryonic tissues. And it's too late, actually, we believe to get them.
Looking like the so-called real thing isn’t enough. A model of the gastrula, a gastruloid likely cannot be simply given the components it is missing to then become an embryo. It’s all about the developmental time windows and the fact that an embryo develops in a holistic way.
Christine Mummery [14:40]
So the extraembryonic amount of tissues and trophectoderm develops as well. And it needs the interaction with the embryo proper to do that. So if you were just chucking in some trophectoderm cells into a gastruloid, it wouldn't be at the right stage anymore, it's lost its window of opportunity to develop with it.
So this is this is why ethically, we have less problem with gastruloids, even though they look more like an embryo. So, but then the gastruloids can do something else, they could make germ cells. So that you're starting making germ cells, then they get themselves into another slightly more ethically sensitive area, even if they make sperm or eggs. Also, no problem. There's a lot of people trying to do that.
But if you try to fertilize them, or use them to fertilize, as has been done in mice, and has recently been done in that new Science paper for rats, then you're getting into a different area. But that's not allowed for humans.
The guidelines lay out for labs what is permissible and what isn’t. And of course this all can be challenging for scientists. What is certainly challenging is natural variability. For instance, the knowledge of what a typical blastocyst is, in detail. Because this is not entirely known and that’ is what can make it hard to gauge a model.
For example, one scientist I interviewed, Ge Guo from the University of Exeter told me about some differences that matter here. It’s important to remember that the starting point for a human embryo, the fertilized egg, is unlike the starting point for blastoids, which begin from a group of stem cells at a certain stage of development.
A blastoid seems to differ from the human blastocyst in the fact that it has a variable number of hypoblast cells, which are a certain cell type. But it’s also true that in human blastocysts, the number of hypoblasts can vary as well from one blastocyst to another as published data show. This means benchmarking is not straightforward.
Christine Mummery [17:00]
It is indeed because there is natural variability. As you probably know, when you do, let's say, biopsies from an eight-cell stage embryo to do a pre-implantation genetic diagnosis. It can miss one eighth of its cells, no problem. Right? And so it puts them back so there is absolute no absolute so the indeed the hypoblasts can have different numbers of cells. But in the end, and the baby
An eighth it can lose. That's kind of extraordinary, isn't it? Wow.
Absolutely. I mean, that's how pre-implantation genetic diagnosis is done, and people didn't actually realize it was going to be so safe. So you can, you know, that's a substantial bit of, you know, an embryo of that very early stage and it goes fine. And everything develops normally. So there's no reason to think those embryos are abnormal.
They somehow later on in development, they get a feeling of space, they know, it's all to do with HOX genes and things like that. So they know where they are, they know what the neighbors are doing. And if there's not enough neighbors, they'll wait until there are enough neighbors to do their next thing. So if the hypoblasts different sizes, then yeah, it may not do its next thing at the same time, or maybe slightly different. It may wait. So there's all these different variables. But there's a certain range of flexibility where everything sort of works out fine. Anyway. I think that’s what the what people mean, that, you know, nature isn't absolute.
These discussions about similarity and resemblance and mimicking the real thing are not only relevant for working with embryo models in basic research. In biomedicine for example, one can imagine stem cell-derived approaches to explore how one might, in the future, address injury to the heart. For example, find new ways to repair damage done due to a heart attack. I asked Christine Mummery what happens when, for example, cardiologists approach stem cell biologists and ask: when might they have developed something that can help heart attack patients. And stem cell biologists possibly say something like: yes we are working on this.
Here's Christine Mummery talking such conversations and more generally about where this work on models is going and where it might lead.
Christine Mummery [19:45]
Yes, so those conversations do take place. But they take place. They've changed over the years. So, you know, cardiologists who work on mice or rats, or humans, absolutely don't like stem cell models. They don't like them, because they say the cardiomyocytes are immature. They look at it like 20 weeks of gestation. Well, we all know that.
Nevertheless, they can be extremely useful for predicting the potential toxicity of drugs. And we do have drugs bought on the market withdrawn because of toxic effects in humans. And if we'd had even our simple stem cell derived human cardiomyocytes, we could have prevented people dying.
Nonetheless, that's no defense. If you want to mimic a human adult human disease, you're going to need much more adult cells. So a substantial group of stem cell biologists have worked for many years trying to get what they call cardiomyocyte maturation. And if you work in the neuro field, you've been trying to get neural maturation, kidney maturation, whatever field you work in, you'd be trying to get more mature cells.
So what happens then, there are many ways to Rome, and you have a plethora of papers, who say I have got--doing X, Y, and Z-- mature cardiomyocytes, neurons, kidney cells. And they may have looked at one or two features, they may have looked, in our case, of electrophysiology, they may have looked at sarcomere alignments, but very rarely have they looked at a defined set of parameters that say: this is a mature cardiomyocyte.
So in our work, we've always tried to say, 'Okay, what does a mature cardiomyocyte look like?’ And to what extent does our cell resemble it? And what do we have to do to get it closer to the adult state?
Now, that's what I mean by benchmarking. Now, if you see a report, or somebody says, ‘I've got better sarcomere alignments with Factor X. As a referee, I always ask, but what does the electrophysiology look like? And then they say, ‘We don't have an electrophysiologist. So we can't measure it. ‘And then I say, very sorry, you can't backup your claim.
And so if, that's true for every organ, If you have to define what you want it to look like and see to what extent your particular thing gets it closer to the end product.
Now, I would predict that there are maybe at the moment, maybe five to 10 claims of it's going to be this gene pathway or that gene pathway. And I bet you, all of them are important, they may be upstream or downstream of one another, they may be in parallel. And we may have to add them all up to get to where we want. Right? So there's nobody right or wrong. It's just, we're still working on it.
Now, if you would, if those same people were a cardiologist looking to transplant something into a patient. So the problem there is we don't know how mature those cells have to be, to actually be happy in the adult human heart.
We know from animal experiments, if you transplant an adult cardiomyocyte into an adult heart, it will die more or less immediately.
We know if we take fetal-like cells and put them into an adult heart, they survive, but they cause arrhythmias.
So how far do we have to be along that pathway of maturity, so a cardiomyocyte is entirely mature, it stops beating? That's what it should do. It needs a pacemaker. If it's earlier, it's beating away, and you've seen it in hundreds of talks, these happily beating cardiomyocytes, Why are they beating is because they're immature, they have intrinsic pacemaker-like activity, but then they don't have a pacemaker cell.
So you have to find out how far along the pathway they are, you can do that with single-cell transcriptomics. There's lots of lots of ways of doing it. But as long as we don't know the answer to that question, how mature does it have to be? It's really hard to know what to answer to this cardiologist, when will you be ready to give us something. Can give you something tomorrow? It might be highly dangerous? Would I advise you using it? Probably not because we're chicken. But that's the way it's going.
So there are people I would say, people like Chuck Murray, Michael Laflamme, maybe Joe Wu, those kinds of people, they're seriously looking at this issue. And, of course, they're transplanting into relatively young animals, sometimes non-human primates that have only had a myocardial infarction, they're not having smoked, they haven't got high blood pressure.
That's as good as it gets at the moment.
So they are asking this very question. What's best to take? very immature, slightly mature, highly mature, should I inject not just cardiomyocytes, but blood vessels ready to go? Should I inject pre organized bits of tissue in carrier proteins, extracellular matrix, hydrogels, whatever?
Nobody knows the answer to that question. For other organs, the question isn't as complicated. So you will have noticed perhaps that there are dopaminergic neurons being injected into Parkinson patients, that is proving extremely interesting and ready to go.
And the reason is putting in immature neurons allows them to apparently to grow out their neurites and do their thing. Or they're still plastic enough to adapt to that environment.
The newest work on let's say, pancreas cells, from Vertex? So that was Doug Melton's you could say, the technology he developed. It's amazing that they've got the cells into patients, I think they're in, you know, a sort of thing that protects them from being immune-attacked in some way. But the point is, when the paper was published, the patient who got it was an extreme diabetic, and the insulin reduction, I think, was in the order of 50%. And that patient, rumor has it to be confirmed, is now completely off insulin. Can you imagine?
Note: This interview took place before the ISSCR annual meeting at which Vertex announced the is indeed insulin free.
Nobody had any idea it would be that successful.
I see. So that just shows the potential of
These are immature pancreas cells going in, so we don't even for each organ we don't know whether it's helped, do the cells have to be. And sometimes you get lucky, apparently. And sometimes you don't.
So this is all very empirical stuff. But you do need a benchmark of these tissues at every stage of their development, to be able to say what state it's in, because that will help you define for other circumstances. So if, let's say, this Vertex study is successful, and it turns out that at this stage of development or maturation they're working. Then the regulatory authorities can say, Okay, if you get your pancreas cells from stem cells that far, you're ready to go. Because the regulatory authorities also don't know what they should ask people to show.
A lot will still unfold with stem cell models and it seems that as the science unfolds, these models might be able to let scientists show plenty. And perhaps, down the road, they can be used to develop therapies. It is too early to say any of this for certain of course. But these models can now be used to study the detailed impact, for example of injury to the heart. And congenital disorders of the heart or the brain might be possible to study with embryo models, too. Especially gastruloid models. Here’s Christine Mummery.
Christine Mummery [28:20]
Oh, yeah, absolutely. I mean, I'm a very strong believer in gastruloid models. Blastoids, I'm not entirely sure, it's very, very early. And, you know, a blastocyst that goes wrong often leads to a very early miscarriage. It's when you get congenital, slightly later, congenital defects, microcephaly, the side of the heart swaps. So we're very interested in gastruloids to look at, you know, altered sidedness in the heart.
What you seem to be able to get, at least in mouse gastruloids, and maybe soon in human is that you can get left-right asymmetry. And that's one of the important things that go wrong in development. So these very early stages are almost impossible to study when you know when you already can see on an echo that the heart's the wrong way around, or is missing one of the ventricles. Yeah, you
I suppose a heart could be on the other side, right?
Okay, but that should be
Can be on the other side, but it can also be swapped around and it can also miss one of the major chambers. So there's, there's a single ventricle heart, which people will survive to birth.
But, you know, we don't really know why, why it occurs. And if you could pick it up early, you might be able to put in a balloon catheter or something to make it grow or something like that.
So I think we're super-interested in how early we can pick this up. And also, which genes or you know, I am saying plural, we know some single genes that cause it. But what we need to know is which multiple genes or predisposition factors or SNPs, whatever, give a higher likelihood, and then we’d screen individuals properly and early.
So I think it's super interesting for sidedness. It's also very interesting perhaps for somite formation. So that gives rise to the muscle and limb development. So if the somites go wrong, you don't get limbs properly developing.
The same with the brain. If the brain doesn't develop properly early on, then you have a major problem. So I think for, to study aspects of gastrulation. Gastrulation is, I mean, Wolpert said it's more important, most important event of your life. And I think these opportunities to study it without actually using gastrula-stage embryos, which of course, are virtually impossible to get, is super exciting.
What matters in development is not just that, for example an organ is developing but what is important is where it is developing and when it is developing. And that the neighbors this organ needs in the body are interacting with this developing organ. So something has to be present at the right place and at the right time.
Christine Mummery [31.35]
Both. Both is important. I mean, if you know if something, if you're looking at regions that have to signal to each other, if one region has not developed, before the other needs to signal, it won't work.
A human embryo has a dynamic, it changes constantly. Embryo models can help scientists explore these changes. They are not human embryos they resemble human embryos, they are models. Even though thezyare models there are guidelines governing how they can be used in labs. A link to the ISSCR guidelines is in the story notes.
And that was Conversations with scientists. Today’s episode was about embryo models and was with Dr. Christine Mummery, a developmental biologist in the anatomy and embryology department at Leiden University Medical Center in Leiden in The Netherlands.
And I just wanted to say, because there is confusion about these things sometimes, Leiden University Medical Center did not pay to be in this podcast. This is independent journalism that I produce in my living room. I’m Vivien Marx thanks for listening.