Jeff Coller on mRNA, RNA Therapeutics, and the Future of Bespoke Medicine

Jeff Coller

Great. Thank you for the invitation, Sean.

Jeff Coller

Yeah, sure. Just to get everybody on the same page about what we mean when we use the word mRNA. In your body, in every cell in your body, you have DNA. DNA is basically the blueprint of life. It's like a giant recipe book that tells your body how to be you. And within that recipe book are hundreds and hundreds of different recipes. An mRNA is an individual recipe. For each gene in the human body, there will be mRNAs made that transmit information. These are messages, recipes that have to go from the library, if you will, which is where the DNA is, to the kitchen, which is a ribosome. The ribosome is the cook that makes the individual dish that's going to be made that day. So these mRNAs tell the cell what to make. And what's really important is that after that mRNA is read, it's destroyed. That way the cook doesn't keep making the exact same recipe over and over again.

Jeff Coller

One per gene, that's exactly right. We have approximately 25,000 genes in the human body, so there will be about 25,000 mRNAs — a little bit more than that, but about 25,000.

Jeff Coller

There are lots of inputs that decide when you're going to turn a gene on or off. A good example: when we eat, if you eat something very sugary versus very fatty, you don't necessarily need the proteins that would digest sugar if you're eating a bunch of fat, and vice versa. So you turn on genes in response to certain stimuli. A stimulus might be a food source, but it could also be a growth hormone, or a signal from a particular part of the body that says, "Hey, it's time to make this." For example, if you have a bunch of sugar in your bloodstream, it would be time to make more insulin so you can take that sugar into your cells. A lot of the time these stimuli are external signals that say it's time to turn that gene on or that gene off. When you turn on a gene, you're basically making an mRNA — a copy of that gene as a messenger that then gets read by the cook.

Jeff Coller

Essentially, yes. There are building blocks within your cells. They really are kind of sugars in and of themselves, stitched together to make an mRNA. They're very similar to DNA. mRNA is almost identical to DNA, with one tiny little difference, but other than that it's pretty much a direct copy.

Jeff Coller

There are really only two differences, but the major one is that RNA has a single oxygen on it that DNA doesn't have. DNA is deoxyribonucleic acid, and RNA is ribonucleic acid — it doesn't have the deoxy part. There's also a more nuanced difference: RNA has a unique letter. In DNA there are four letters — A, T, C, and G. In RNA, it doesn't have the T. In RNA it's A, U, C, and G. Those are the major differences.

Jeff Coller

Yeah. It's often believed by most scientists now, at least molecular biologists, that RNA was the predecessor to DNA — that life evolved on Earth first using RNA as a genetic material rather than DNA. And it's really that little difference I talked about, that oxygen present within RNA, that's the key that led scientists to think it could be the precursor to life, because it gives RNA the ability to do unique chemistry that DNA cannot do. Scientists like Tom Czech and Sydney Altman back in the 80s were able to show that RNA could do enzymatic catalysis — it had the ability to stimulate a chemical reaction, where DNA cannot. And that's a necessity of life. You have to be able to catalyse reactions, and RNA can do that.

Jeff Coller

That's right. DNA tends to be more stable, again for the same reason — it doesn't have that little oxygen. So it tends to last a lot longer.

Jeff Coller

Yeah. And the other part of this, in terms of the RNA world hypothesis, is that RNA can do chemistry but can also adopt very unique, complex shapes. DNA really only exists famously as the double helix — that's how you find it. RNA can get into all sorts of contorted shapes and sizes, and have this ability to do chemistry. So it really does become an ideal molecule to see how life could have evolved on Earth before a DNA world.

Jeff Coller

Doing different purposes. And it goes back to what I said about being able to form different structures. The largest set of RNAs in your cell are your ribosomes. The ribosome — that cook I talked about at the beginning that reads the mRNA — is actually RNA itself. It's half protein, half RNA. But the part that actually reads the genetic code is in fact RNA. So it's called a ribosomal RNA. It forms this really large, very complicated three-dimensional structure that's consistent through all of life, and it's the machine that reads the genetic code. There are a bunch of other RNAs around the cell that do different work, but mRNA in and of itself is a direct copy of a gene within the DNA.

Jeff Coller

It's fascinating to see, because it all works in such beautiful harmony with itself. What you realise when you dive deeper into the structure of these complexes is that it evolved very early on and has diversified from there. For example, that ribosomal RNA — the ribosomes within the human body are not much different from ribosomes you'll find in fungus, that you'll find in bacteria, and in some of the most primordial life forms on Earth. In fact, the study of ribosomes was one of the reasons that led a guy named Carl Woese to propose that RNA was the precursor to all life. It's as if there was a primordial life form that eventually evolved into a ribosome, and everything evolved from there.

Jeff Coller

A lot of this happened in the early 1960s. Famously you had the double helix in the 50s, with Watson and Crick discovering the structure of DNA — they didn't really discover that it was the genetic material, but they discovered the structure. About the same time, it was hypothesised that there had to be an intermediary between DNA and proteins. We knew DNA was a genetic material in the 1950s, we knew protein was a very important part of biology, and a guy named Sydney Brenner was one of the first people to propose that there must be this theoretical intermediary between them, which then, about 1964 or so, was shown to be messenger RNA. So the discovery of messenger RNA dates back 60, 70 years, to the 60s. And as molecular biology has developed over the last 70 years, we've learned more and more about the complexity of RNA within the cell — there are dozens of different types.

Jeff Coller

Yeah. Your nucleus contains your nuclear DNA, and there's a ribosome used in the cytoplasm to decode the mRNAs that come from your nucleus. But there's a separate ribosome present within the mitochondria. For your listeners, it's often thought that the way evolution proceeded is that there was a bacterial invasion of a cell — at some point a cell went inside a cell and became permanently fixed. That is what your mitochondria is. If you look at your mitochondria, it actually looks very bacteria-like, including those ribosomes. The ribosomes specifically made to decode the mRNAs from your mitochondria are bacterial-looking, not human-looking.

Jeff Coller

That's right. It is kind of a mess, because the thing about evolution is that there's really no selection for simplicity within the cell. It just gets built on top of itself, to the point where you can't reconstruct it to make it something more elegant and simplistic.

Jeff Coller

There's thousands. In a given human cell, there are probably about half a million ribosomes. So it's quite a bit.

Jeff Coller

Within one cell. Half a million ribosomes within one cell. When it really comes down to it, your cell is just a big bag of ribosomes.

Jeff Coller

Yeah. Going back to the 50s and 60s, we knew DNA contained the genetic information, and deciphering that information — learning what the code was — was important. DNA contains these four letters, A, T, G, and C. If you look at billions of base pairs of DNA, it has a billion different combinations of those four letters, and those letters make up what we now call the genetic code. What DNA does is make mRNA, and that mRNA makes a protein. There are 20 building blocks to a protein, and these are called amino acids. So you have four letters that have to encode 20 meanings, and that's really the simplicity of the genetic code. If you do the mathematics of it, you need 64 words to have 20 meanings. That's the only way you can do it mathematically.

Jeff Coller

You can't cover it. And then, by having three letters per word, you get 64, which is too many. But there's nowhere in between. So you have four letters arranged three at a time, and that leads to 20 amino acids. It's 64 words basically. Does that make sense?

Jeff Coller

That's exactly right. We call that the degeneracy of the genetic code. To use a specific example, there's an amino acid called arginine. Arginine in fact has five words in the genetic code that mean arginine. Of these 64 words, there's some repetition. That's the way mother nature did it.

Jeff Coller

Not really. There are two amino acids that only have a single word. That is very important for one of them, because every single protein always starts with that first word, and that word is the nucleotides ATG. That word means methionine. Every protein always starts with methionine, and it has to be that way, because then the ribosome knows that the first time it sees the word methionine, it's the start of a sentence. It's just like the capital letter of a sentence. And then there are three words that are a punctuation mark, a period — the ending of the sentence. Those two parts are very important for the ribosome, because it knows how to read the genetic code: it knows when to start and when to end. For the rest of the amino acids, it's relatively random. In fact, Francis Crick called it a frozen accident — this is just what happened, it got frozen at some point in evolution, and now all organisms use the exact same genetic code, for the most part.

Jeff Coller

Yeah, there are. But for the most part, life uses those 20.

Jeff Coller

We had room to use 61 — we still needed those three periods, those punctuation marks. So 61 was theoretically possible, but we don't use it.

Jeff Coller

There are people that work on unique chemistries, with both the adapter molecules that read the genetic code — these are RNAs called transfer RNAs, little tiny RNAs that read each of those words, with one of the amino acids bound to them. There are synthetic biologists who look at that. There are amino acids out there that are not used routinely in biology — could you stitch them into a protein, and what consequence would that have? Who knows?

Jeff Coller

I think it has more to do with their natural abundance in the environment. But there's a lot we don't understand — why these 20, why are they here on Earth, why are they abundant on Earth? Some of these amino acids have been seen in outer space; you can see them in nebulae and other places. So they might actually have a common natural origin within the universe that makes them a good property to use.

Jeff Coller

It's getting it all organised. The actual building blocks of DNA, RNA, and protein tend to form naturally with the ingredients out there in the universe, and with some heat and radiation they come together. But stitching it all together in a way that gives life meaning — where life is really a chemical reaction that perpetuates itself — that's something we don't understand yet, how that happened.

Jeff Coller

That's exactly right.

Jeff Coller

They don't seem to be, and that's something my lab has studied. As I said, these codons have degeneracy, meaning you can encode the same amino acid with multiple words. But the way the ribosome reads those words uses that other little RNA, the transfer RNA, and it turns out that the abundance of that transfer RNA dictates how fast you read the word. A simple way to think about this: the ribosome is the cook reading the recipe, and it has to interpret 64 of these words. Some words it will stumble upon, because it doesn't readily find them. There are these transfer RNAs floating around in a cloud around the ribosome, and how concentrated that tRNA is dictates how fast the ribosome can read it. It's no different from a bucket of bingo balls all coloured blue with one red one — the red one would be very hard to find if you were sampling randomly. That's the same thing with the ribosome randomly sampling transfer RNAs to decode the genetic code. It hesitates over some words and goes fast over others, and that speed is dictated by the genetic code. That's what my lab has studied — how fast the genetic code gets read — because it has serious implications for the stability of the mRNA. I mentioned that mRNAs are recipes that get read and, as soon as they're read, destroyed. How fast they're destroyed is a consequence of how well they're read. If it's something you can read very easily, you keep that message around a lot longer than if it's a message that's really hard to read.

Jeff Coller

It's active. Very active.

Jeff Coller

That's exactly right. Destroy upon reading.

Jeff Coller

Sure. This is a very old history with mRNA being used in medicine, despite what the press probably has as an idea. The fact that mRNA exists — we know this is the recipe that sends information to the cook. There's an mRNA for every one of your genes, because it's telling your cell what to make at any given time. When we discovered mRNAs in the 60s, this was exciting, because it means that if we can design mRNAs, we can get the cell to make any protein we want, regardless of whether it naturally exists or not. This is a natural system — mRNAs are natural products of your body. We understand the genetic code, we understand a lot about how proteins work, and so we can design mRNAs that we stick into a cell to have the cell make any protein we want. That's what really electrified scientists in the late 90s and early 2000s: could we design mRNAs to be used in medicine? The place where people really got excited was cancer, and I imagine we'll talk more about that. But the other was vaccines. Vaccines are usually a protein — a non-native protein, something that comes from a virus or a bacterium, a foreign invader. Your body has the ability to tell who you are; it knows every protein in you. And if there's a foreign protein, it attacks it. So, could we use mRNA to make those foreign proteins, so the body could recognise them and launch an immune reaction? That's really the genius of using mRNA in vaccines. There are three things about it. One, it's natural — a natural product, and the body destroys it quickly, within a matter of hours. The other side is that you can deploy these really fast. We can design an mRNA in just a matter of hours. When it comes to preparedness for a pandemic or an emergent infection, you can develop it very quickly, where the classic way of making vaccines takes 10 years or more. That speed of manufacturing is what has really electrified the community.

Jeff Coller

Oh, it's infinite. There's an infinite number of possible protein signatures out there. We all know this from seasonal allergies — that's basically an immune reaction against a foreign protein, usually a pollen. There are tree or plant proteins that you're breathing in, and it's irritating your immune system.

Jeff Coller

That's actually not far-fetched, because the fact that mRNA is this emissary of the genetic code means we can use it to do all sorts of biology. One of the places people are looking is in immune reactions — or autoimmunity, which is a big issue. You have the ability to say what proteins in your body are yours and what are not, but sometimes your immune system goes haywire and starts attacking your own body. People would recognise arthritis as a good example. If you can learn to downregulate the immune system through mRNA technology, which people are investigating, then it's one way to protect from autoimmunity, or from the overstimulation of the immune system you get with allergy reactions.

Jeff Coller

Yeah. Before COVID hit in 2019, there were over a hundred FDA trials currently in humans for different mRNA technologies. Most of these were around cancer and vaccination. At the time COVID hit, Moderna, for example, was testing a vaccine for influenza. That allowed them, because they already had that program built, to quickly change the influenza sequence out for the COVID sequence — they had all that infrastructure already in play. It's important for the public to realise that. While in 2019 this was a new technology to them, it wasn't new to the scientists. We had over 20 years of in-human testing already conducted with mRNA technologies.

Jeff Coller

Inside a human being. That's exactly right.

Jeff Coller

That's exactly right. That's the beauty of the mRNA technology too. You're letting the body make a protein that is foreign, and the body recognises it and goes, "This protein doesn't belong, so I'm going to mount an immune reaction." And the amazing thing is the mRNA then disappears. It goes away. The only thing that remains is the immune system's memory of that so-called foreign invader, and that's what gives you the immune reaction that carries forward for months, hopefully years.

Jeff Coller

The main difference is what they are. Most vaccines are protein, where the mRNA vaccine is the mRNA that then makes the protein. From a product standpoint, that's the major difference — the MMR vaccine, or the yearly flu vaccine, are proteins that elicit the immune reaction. With the mRNA, we're just giving the instructions to make the protein. That has serious implications in two ways. One is the speed at which you can develop it, because it's actually very difficult, even in 2026, to make proteins at scale — large amounts of proteins. To be honest with you, we would still use an mRNA even if we weren't doing it in the body. The second thing is cost, because making a protein is very costly. Usually, to make a protein vaccine, you have to program something like a chicken egg with a virus, or a DNA or an mRNA, to make that protein, then grow it up. You'd have to have millions of eggs to do this, and then purify that protein from those eggs. That's a costly project.

Jeff Coller

It's slow, and it could take 15 years to develop. Where an mRNA you can develop in just a matter of weeks in terms of manufacturing, and you can develop it in silico, in the computer, within a matter of hours, then scale it in a matter of weeks. So the cost is significantly lower to be passed on to the public.

Jeff Coller

It wouldn't be that hard, to be honest. I have a lot of inside baseball on this one. When COVID hit in December of 2019, the Chinese researchers had sequenced and identified SARS-CoV-2 as the virus responsible for COVID-19. They sequenced its genome and then released it publicly around the world. The day that came out was in early January, and the very next day the sequence had been downloaded. My graduate student, who was the lead designer of the COVID vaccine at Moderna, downloaded it, put it into his program, and had designed the vaccine that went into millions and millions of human beings within just a matter of hours.

Jeff Coller

His name was Vlad Presnyak.

Jeff Coller

In fact, when he got the information, because we were all still not completely aware what was going to hit us, he didn't really know the implications of what he had just done.

Jeff Coller

In this case, what was done for the COVID vaccines is they took a particular protein from SARS-CoV-2, which is called the S protein. The S protein is used by the virus to help it stick to a human cell and get inside. What's key here is we don't make an S protein — humans don't have one. It's a large protein in SARS-CoV-2. They put that entire sequence for the S protein into the vaccine. There wasn't much trickery to how you designed it. It was just, we'll take the S protein and stick it into an mRNA. There's a little bit more nuance — some optimisation of the sequences — but for the most part it is the S protein that's naturally occurring in SARS-CoV-2, with a few little changes.

Jeff Coller

That's right. S protein is short for spike.

Jeff Coller

The spike protein, which is the reason it's called a coronavirus, because it looks like a crown.

Jeff Coller

Yeah, that's pretty much what we do. If you identify a new virus — everybody's now talking about this hepta virus that people are worried about — it wouldn't be that hard to develop a putative vaccine for it. If you know the genetic sequence of that virus, you identify a protein that might be a good candidate for a vaccine, and you'd clone it, put it into an mRNA, and then you'd have to do testing.

Jeff Coller

It's all done in a test tube, using what we call an in vitro procedure — in vitro being in a test tube. You make a DNA template, and from that template you can use a couple of enzymes and amplify up mRNA in huge volumes. If you think about this from a manufacturing standpoint, traditional vaccines based on proteins need what are called bioreactors — basically a vat, a giant pot you're cooking it in. A bioreactor for a traditional vaccine could fill half of an aircraft hangar; you'd put thousands and thousands of litres of these vats into a room to make your vaccine. With an mRNA, your bioreactor can be about as big as your body, for the entire population on Earth.

Jeff Coller

Some of the newer technologies I've seen — I've talked to investigators who have bioreactors no bigger than a two-litre bottle of soda that would hold enough mRNA to inoculate the entire planet.

Jeff Coller

Yeah. Your body doesn't like to take up strange nucleic acids, as it shouldn't, because viruses are in fact nucleic acids like RNA and DNA. And to be honest, you eat RNA, mRNA, and DNA every single day — every single bite of organic food, whether it's a steak or a salad or a bean, has bean RNA in it, cow DNA in it, cow mRNAs in it. You can't just readily take those things up, because they would work just like any mRNA in your body. So the way we get the body to take these up is we encapsulate them in little fat bubbles, and that's called a lipid nanoparticle. Your cells are made out of lipids, and we make these little fat bubbles out of lipids that we keep the mRNA in. There's an old saying in chemistry that like dissolves like. Water and fat don't go well together — everybody who cooks knows that if you put oil in a pot of water, it separates. But if you put oil on top of oil, it mixes. That's what a lipid nanoparticle is. Your cells are lipids, we put the mRNA in a lipid, and when they touch each other they just fuse and join together, and that mRNA enters into the cell.

Jeff Coller

No, it doesn't. In fact, it's a feature that it doesn't go to the nucleus. mRNA works in the cytoplasm, so it never goes near the DNA, and it's read by the cook, the ribosome, in the cytoplasm.

Jeff Coller

That's exactly right. All through these natural processes.

Jeff Coller

It is the way we have now, but there are lots of people working on improving delivery in different ways, because these lipid nanoparticles lack what we call tropism. What I mean is they work really well if you stick it into your arm, because you have immune cells there that take it up. They'll read it and communicate that information to your other immune cells, and that spreads naturally through your body. But the only other place we really have good ability to deliver to right now is the liver. That's because everything goes to the liver in your body — it's your detoxifying organ. If you inject anything into your body, whether it's Tylenol or aspirin or whatever, everything goes to your liver. So right now we can target mRNAs to those immune cells around the injection site, and to your liver. But investigators are working quite seriously on developing delivery mechanisms to other cell types — the pancreas, the brain, the heart, the kidneys — so that we can tap into the potential of delivering these medicines to the places they really need to be.

Jeff Coller

The mRNA doesn't really spread far. It's basically at the injection site, because that's where the cells are taking up the lipid nanoparticle and the mRNA. Those cells are the ones expressing the foreign protein, and then the immune system is learning from that, and that's what spreads. The mRNAs themselves are very unstable — they have half-lives on the order of hours. So even by the time you drive home from the pharmacist, most of it's gone.

Jeff Coller

The brain's tough, because of the blood-brain barrier. Your listeners may not know this, but there's very little exchange of material between the bloodstream and the brain. The brain is a very protected space, which is called the blood-brain barrier, the BBB. Getting things across it is difficult, and there's no real good way to do that through an injection in your arm. In the type of gene therapies developed for devastating brain disorders — and I'm not talking about mRNA, I'm talking about other gene therapies — those usually are injected directly into the brain, so you can go right through the barrier.

Jeff Coller

Yeah. This is what the world is very excited about right now — it's the amalgamation of several incredible technologies, two of which are Nobel Prize-winning. CRISPR is basically a system discovered from bacteria. Bacteria have this primordial immune system — a way to know whether they've been infected by viruses. CRISPR was harnessed by researchers to target specific areas within your DNA and then make cuts. That's what CRISPR does. It's like a GPS signal: find this piece of DNA and then make a cut. Why is that important? Because you have billions and billions of base pairs of DNA, and if you want to specifically change one tiny region, you need a GPS locator to find it. Then, on top of that, researchers at the Broad — a guy named David Liu — took CRISPR and hooked it up with another set of proteins that could make changes to the DNA. So now, with CRISPR you can localise to a very specific region, GPS-target it, and if there's an error there, you could make a correction. That's what base editing is. CRISPR fused with base editing is this surgical way to go in where there's a mutation, a change in the genetic code, find that region, and correct it so it's normal again. Those are the two big pieces of technology. But the problem is it's not perfect. You can make a change there, but if there's any opportunity for that CRISPR to wander around the DNA, it might make a change somewhere else that could be bad. You don't want that machine always present in your body. And that's where this has been married to mRNA technology. You take that CRISPR base editor protein and make it as an mRNA. Think about it this way: the CRISPR base editor is a surgical instrument to change the DNA — think of that as the new recipe your body needs to make. How are you going to deliver it? With that recipe card, the mRNA. And the reason you want to do this is the key I've been saying during the entire podcast: mRNA is destroyed after reading. So you introduce it, help it do what it needs to do, make the little correction, and then destroy it so it's no longer there. The only thing left is the change that was made — the correction, which has now gone back to a normal piece of DNA. It's just like a surgical instrument.

Jeff Coller

No, this is not a vaccine. There have been a handful of these done in vivo, meaning in the human body. The first example wasn't using an mRNA — they were done on blood cells. They would take blood out, the cells that need to be changed for something like sickle cell anaemia, and do the change ex vivo, outside the body, then put those corrected cells back in the body. That has worked. But in that case, it always had to be done outside the body. By combining the three technologies — the CRISPR, the base editor, the mRNA — it allows you to do the correction in the body. And we only have one example of this.

Jeff Coller

The one example is this young baby, born last year — 2024 — in Pennsylvania, named KJ Muldoon. KJ was born with an ultra-rare genetic condition that affects about one in 1.3 million children. It's a deficiency in an enzyme in his liver, a protein called CPS1. What that protein has to do is help metabolise proteins, and if you don't have it, you build up toxic ammonia in your body. It causes severe neurological damage, because ammonia is poisonous to us. When KJ Muldoon was born with this CPS1 deficiency, it was clear something was going wrong. The only option for him was a liver transplant. But he was way too young and too delicate to survive one. And a liver transplant is not a great option anyway — it's really significant surgery, it builds on the tragedy of someone else, and you can't always guarantee a liver is going to be compatible, again because of the immune system. So what researchers at the Children's Hospital of Philadelphia decided to do was create this never-before-done technique of taking CRISPR base editors with an mRNA and going into his body — because it was the liver, and we can get things to go to the liver — to see if they could make a correction in his DNA that would restore his CPS1 gene back to normal. They moved heaven and earth to do this. In an unprecedented set of exercises, they were able to secure FDA approval within a matter of weeks — a week — and treat him within a few months of his birth. I think it was about six months. And essentially, it corrected his mutation. Now he's thriving. He's no longer in danger, he's metabolising proteins appropriately, he's walking, he's talking. He still has to be monitored — he's not cured, we can't say he's cured — but he doesn't need a liver transplant. And the incredible thing is that because he had an mRNA-based technology used on him, and that mRNA goes away over time, he could be redosed again in five years if we figured he needed a little more correction, because his liver is going to grow and expand. He still has some cells with the mutant copy of the DNA, so maybe he'll need another round of correction to build up more normal cells, and we can do it with that approach. That has everybody super excited, because it's an ultra-rare genetic disorder — one in a million — and he has a personalised therapy that goes in like a surgical instrument, changing his DNA back to normal, and he's thriving. If we can do it for him, we can do it for any ultra-rare genetic disorder, or more common genetic disorders in the future, if we invest and have the right technologies in play.

Jeff Coller

Absolute science-fiction stuff, right? They fixed his DNA, which is ultimately where we've always wanted to be in any kind of therapeutic. There are over 7,000 genetic diseases that humans suffer from. We often call these rare genetic disorders, but one in 13 people suffers from some form of genetic disease. So it's not that rare — they're just rare as an individual indication. Some will have cystic fibrosis, some might have muscular dystrophy, some may have Alzheimer's or dementia or sickle cell. All of these are the result of an abnormality in their DNA. If we have the power to go in and correct these, then it's a potential path toward improved lives for these individuals.

Jeff Coller

That's always a challenge for any particular disease — how much do you have to correct? For the most part we don't know for a lot of diseases. In the case of baby KJ, he was dosed three times with this corrector, in stages. They gave him a little bit, then a little bit more, waited, gave him a little bit more, and then he started to show signs of improving. To this day we don't really know how much correction he got, because we can't tell — we're not going to go in and do a biopsy on him. But with other diseases, with gene therapy, sometimes you may only need maybe 10 or 15% improvement over the baseline to get some sort of benefit. It would be a miss to say we're curing people. We're not there yet. But if we're improving their lives, making their lives more tolerable, having them live longer, reducing their medical needs, all of that is a good outcome for any therapeutic. In some cases it may only be 10 or 15% above what they already have to give a meaningful change to their lives.

Jeff Coller

There's a whole group of researchers — thousands of researchers and thousands of small biotech companies — that have tried to make gene therapies for every disease under the sun. For all the ones that have a genetic origin, which is the vast majority, if we could make the correction in their DNA, that's the solution. You need to do a few things. You need to be able to get to those cells, and that is a limitation — we have to be very honest about that. Right now we can reach the liver. We don't have the ability to go to a lot of other places. But there also hasn't been a lot of incentive, because we haven't had technologies this powerful. And humans are really good at responding when there's a need. Necessity is the mother of invention. If we have a technology that could reach the brain, or make significant changes to patients with genetic disorders of the central nervous system, we're going to figure it out. We will, in time.

Jeff Coller

Absolutely. The technology that keeps us back right now is really the delivery component — how do you get it to the right cell types? But the ability to identify a mutation in a patient, we've had for many years now. We can sequence a person's genome for a couple hundred bucks now. The Human Genome Project cost over a billion dollars; we can sequence a human's DNA in a day now for a couple hundred dollars. That's incredible technology in 30 years. So we can identify the genetic basis of most diseases, and now we have these CRISPR mRNA base editors to make changes. Those are improving — they're going to continue to iterate and get much better. But delivery is going to be the challenge, and every organ has a unique challenge. Cystic fibrosis is a great example — it's the lung, and your lung is coated with mucus, because that's how it functions. Part of what that mucosal layer does is help trap foreign things: you breathe in a virus or a bacterium and it gets trapped so it doesn't get into your body. The same would be true of a gene therapy — it can't get through that thick mucosa.

Jeff Coller

It would just stick to the mucus and never really penetrate in. But there are people working on getting it past that mucosal layer, and if we can do that, we can go after diseases of the lung. The point is every organ has a unique challenge. My gut feeling is that in 20 years, maybe less, we'll have a toolbox of all these delivery mechanisms to get CRISPR base editors to different cell types and tissues, and it will be the standard of care for genetic medicine.

Jeff Coller

Yeah, in different ways. Cancer is a complex disorder, but where people are really getting excited — not with the CRISPR base editor, but with the mRNA technology again. mRNA is this natural product that delivers information to your body, and we tap into that. What was identified a number of years ago by James Allison is that when cancer cells develop — cancer is basically your cell going crazy — it tends to rearrange your genome. When it does that, it's making unique proteins. And when you make a unique protein your body doesn't normally see, that's a target for your immune system. So if you can train your immune system to attack your cancer cells, your body can attack them and lead to a remission of that cancer. What investigators are doing with the mRNA technology is trying to identify those unique protein signatures within a cancer, stitch them together into an mRNA, inject it into the tumour, and train the immune system to attack that tumour. A great example came out of Sloan Kettering in 2022 — a paper in the journal Nature that showed, in a small clinical trial, a personalised cancer therapy, what's called a neoantigen therapy, using an mRNA for patients with pancreatic cancer. Pancreatic cancer is scary — it's like 95% mortality within a year. They injected these personalised neoantigen therapies into these patients, and 50% of them responded and are alive today. The study started in 2020, so it's been about six years, and they're pretty much cancer-free.

Jeff Coller

50% is not 100%, but much better than a 95% chance of death within a year. And that's just out of the box. As we get better at using mRNAs to go after cancer, we're learning the rules that can revolutionise our treatment of some of the worst cancers we've ever had to deal with — melanoma, glioblastoma, pancreatic cancer. We're making headway there.

Jeff Coller

It's one of those things where a lot of these incredible technologies are being put together. He had a Nobel Prize-winning idea, the mRNA is a Nobel Prize-winning idea — you put peanut butter and chocolate together and it makes a great candy bar.

Jeff Coller

Well, everything can be used for the force of bad, I suppose. But I would say the force of good is more prevalent than bad. The one thing that keeps me up at night is some of the policy decisions made in the last few years around a technology that's been demonised — by people who want to demonise it because it's a way to get people scared. There was a study a few months ago using artificial intelligence to design viruses. These were viruses that would infect bacteria, so they're not human pathogens. Using AI programmed with millions of viral sequences for bacteria, the researchers were able to develop novel viruses that had never been seen before by mother nature. And 16 of those performed better than the viruses that had been seen.

Jeff Coller

Meaning they were able to kill the bacteria more readily. These are viruses that affect bacteria, so, okay, who cares? But those researchers were American researchers, so they're bound by ethics regulations and couldn't program their AI algorithms with human viruses. As AI gets more and more powerful, a bad actor anywhere in the world with access to a simple AI algorithm could do the same thing programmed with human viruses. So you potentially could create a human pathogen using artificial intelligence that could be quite deadly. The only countermeasure we have to that threat is mRNA-based vaccines. It's the only thing we could ever leverage and deploy to our troops, or to a population, at a speed that would be a natural deterrent to that threat. That bothers me, because we can't throw away a technology that — in national defence we call this deterrence by denial. You don't allow the other party to develop a technology, because you've got the countermeasure that can match it with equal speed and performance. So we've got to protect this technology rather than throw it away. It is an American invention.

Jeff Coller

There already is. China is developing about 46% of all mRNA-based vaccines now, not the United States.

Jeff Coller

And this is not far-fetched science fiction. There was a time back in the Middle Ages when they used to launch infected people over the walls of the castle — if they had anthrax, they would launch their bodies into the castle to spread it within the walls. This is something we know how to do. The other side is that the incapacitation of troops with pathogens is much more effective than kinetic weapons. We saw this during COVID with the USS Theodore Roosevelt — coronavirus went rampant through it, and it was taken completely out of commission within a few days because of the infection. Something a kinetic weapon was incapable of doing. A virus took down an entire ship.

Jeff Coller

We should invest in education about what they actually are, and in the infrastructure we created in the United States. There's been an unnecessary demonisation of a technology that was very much an American invention, that really did help end the pandemic and saved millions of American lives. We need to keep building that infrastructure rather than run this very concerted mis- and disinformation campaign, which is based on falsehoods and not science.

Jeff Coller

Oh yeah. The regulatory issues are huge, because the FDA and the regulatory agencies are set up to deal with blockbuster drugs. When you have a drug like a GLP-1 inhibitor that costs hundreds of millions of dollars to create, but is needed by hundreds of millions of people around the world, you can distribute that cost among all those patients. That's not true when you deal with ultra-rare genetic disorders. It still costs $100 million to develop, but now you have five patients worldwide. The economic model just doesn't work. The FDA is really designed on that blockbuster-drug mentality — a drug used in millions of patients, not dozens. We need to change that regulatory framework now, because we have technologies that are personalised, whether that's the cancer approach I talked about with pancreatic cancer, because that is personalised — each one of those would be an individualised product. And the same is true with a bespoke gene therapy for a baby like KJ. This is an individualised, personalised medicine for him. It's not useful for any other person; it's useful for him, and it cost $8 million. So we have to change our regulatory and commercial enterprise to meet this moment, where we now have the capacity to treat individuals one by one, to their unique genetic signature.

Jeff Coller

Absolutely. And it's not unrealistic that that might happen in the next five years.

Jeff Coller

Yeah, thank you. Thank you for your time.