For a man playing God, George Church certainly looks the part. Over the past 45 years, the Harvard geneticist and his bushy white beard have published hundreds of papers and earned dozens of patents expanding our ability to read, write, and edit DNA, the code of life. He was among the first to apply the gene editing tool CRISPR to mammalian cells (he tied with his former postdoc). Church and his eclectic lab have pushed bioengineering in multiple directions, showing how it can be used to resurrect mammoths, eradicate malaria-carrying mosquitos, produce atmosphere-cleansing bacteria, and even detect bits of dark matter pelting us from space. He once stored 70 billion copies of his book, Regenesis, in a drop of DNA the size of a period after translating it into the As, Ts, Cs, and Gs of DNA’s double helix. But we wanted to talk with Church about the future of humanity, and to that end, he mused on pig organs, dating apps, brains in a dish, and artificial intelligence. This conversation has been edited for brevity and clarity.
Are you hopeful about humanity?
George Church: If you look out into space, so far all we see are dead rocks. I’m very concerned that that’s not a great alternative, and we need to do everything we can to preserve our species until we get some evidence that there’s something out there that’s as good. It’s not even clear that, if we disappeared, the apes or the squids would take our place. I guess more to your point, what are my odds that the species will make it 100 years? I just have no idea.
Lots of people want to extend their lifespans. What’s your approach to increasing longevity?
There are nine pathways, according to review articles. It would be naive to think that there is going to be one magic bullet, some simple food or drug or absence of food that’s going to do it by itself. One key pathway is the reduction of inflammation. There are anti-inflammatory proteins that can be delivered either locally or systemically. Another pathway is restoring function in mitochondria, the powerhouses of the cell. TFAM is an example of a protein that we’ve shown, if you increase it, you get high levels of a small molecule used by mitochondria. And as we age, there’s a shortening of telomeres, the DNA protecting the tips of chromosomes. TERT is the enzyme that adds base pairs to the end of chromosomes. So if that’s weak as you age, then you can add excess. The list goes on and on, and many of these have been shown to work in mice. I like the gene therapy approach.
As a technologist, I see things a few years before other biologists do, and certainly before the bioethicists do.
How do gene therapies work, and what else can they be used for?
Proteins and genes can be delivered intravenously to humans. If a particular protein is missing or damaged, you can either put that protein in or you can put the gene that codes for that protein in. There are about 2,000 gene therapies in clinical trials. A gene called RPE65 makes essential proteins in the retina, and mutations in this gene lead to vision loss in a condition called Leber congenital amaurosis, or LCA, but the gene can be replaced with gene therapy. There are other things you can do with gene therapy and protein therapy that aren’t necessarily fixing some mutation. You’ve got a very sophisticated machine, which is a protein, which you can engineer to do new things in the body. You can go through the human species and find rare individuals who have exceptionally interesting proteins. For example, you can find people who are highly exposed to HIV but don’t get AIDS, even without treatment. And you can find out why it is they don’t, and you can turn that into either a protein or a gene therapy.
Another approach to longevity is organ transplants for those who need them, but spare human organs are scarce. Where are you with modifying the pig genome to grow human-compatible organs?
We have begun trials in nonhuman primates of organs from engineered pigs. Some people say, “Oh, you shouldn’t do enhancement,” but the thing is we do enhancement all the time — to some extent, all aging reversal is enhancement. Vaccines are enhancement. Even though it’s hard to come up with a strong argument for enhancing a healthy human being, a lot can be said for enhancing an organ that’s going into an unhealthy human being, because you want that organ to be as good as it can be. You want it to be pathogen-resistant; you want it to be cancer-resistant. You might want it to be resistant to cryopreservation if you want to store the organ.
Have you experienced much ethical backlash from people who don’t like the idea of mixing species?
I’ve found that my worst critic is actually me, so I have a lot of backlash from George Church. And part of that is because as a technologist, I happen to see things a few years before even other biologists do, and certainly before the bioethicists do, so I get first dibs, in a way, on all the ways things can go wrong. Some of my colleagues worry that regulations are going to slow things down. I have never seen a regulation slow down my research. But I have seen fields slowed down by inadequate regulations. For example, there weren’t enough regulations in Europe on thalidomide, which caused miscarriages and severe birth defects, or in the U.S. on Vioxx, the painkiller that increased the risk of fatal heart attacks.
I’m also pro-discussion of things that are outside the FDA’s purview. They’re not responsible for economics and equity, making sure a particular treatment is available to all people of the world, for instance. I’ve spent a huge fraction of my career trying to bring down the price of technologies. The one that’s most dramatic is DNA sequencing. We brought that price down by 10 million-fold over the past 14 years, and I’m hoping that this year we’ll bring it down below zero dollars. Hopefully we’ll be paying people to get their genome sequenced.
You’ve said everyone should have their genome sequenced. Why?
Everyone should have the ability to get sequenced. We could save hundreds of billions of dollars, because about 5 percent of all babies are born with very severe genetic disorders. They die young, and there’s a lot of pain along the way and a lot of money spent.
The idea is that people would avoid marrying others whose genes would combine with theirs to produce sick babies?
The same carrier status, yeah. This has already been proven to work for a very small number of diseases in a very small population. Tay-Sachs disease, for example, was almost completely eliminated by matchmaking. So matchmaking isn’t even FDA regulated, and it’s not even medicine, but it’s extremely powerful in its medical consequences. And there’s almost no downside of matchmaking. You’re going to lose 5 percent of the people on your social network that you would otherwise be considering as a mate.
I don’t think CRISPR is such a big deal.
Do you foresee people adding their genome sequences to dating profiles and setting them to automatically filter out genetically incompatible mates?
That’s totally feasible to do with software. Or you could text someone and say, “Hey, the software says we’re compatible. Would you like to go out and get some coffee?” It’s a little bit edgy, in the sense that you’re talking about procreating. Kind of a funny opening comment for a pre-date.
What’s the future of sequencing?
The next stage will probably be three kinds of sequencing. One is the lower-cost version of what we’re doing right now, which is fluorescent sequencing. The basic idea is you place your DNA under a microscope and add chemicals that attach to A, C, and G, and fluoresce with four distinct colors. A camera records which dots are in which order. Second, there’ll be fluorescent in situ sequencing, where we can actually see what the name of all the expressed genes are and where they’re located in a tissue section.
Like augmented reality?
It could be ordinary, just seeing it on your computer screen, but the point is right now we throw away that 3D information by just blasting open the cells and randomly distributing their guts on a slide. We could retain that information. You’re right, once we have that 3D information, you would need a way of visualizing it, and probably that would involve 3D imaging like virtual reality or augmented reality.
And the third one I think is very exciting: wearable sequencing, where the device is so small and so fast that it can almost keep up with where you are, and it can tell you whether the environment around you has allergens or pathogens, and maybe even identify the animals and people that are near you.
As for gene editing, you played a big role in the application of CRISPR to mammalian cells. Could you briefly explain what it is, and then talk about why it’s such a big deal?
Well, actually, let me start by saying I don’t think CRISPR is such a big deal. I should say it’s a big deal because I’ve probably personally profited from it more than anybody else on the planet, but there are about eight different ways of doing editing. CRISPR is just one of them. It’s a fairly small increment over the previous method, and all these other editing techniques still work. CRISPR is maybe a threefold improvement over the previous one in certain categories, and it’s worse in other categories. But there’s no question that the whole thing is a revolution.
How many startups are you part of now, and what’s the appeal of working with industry?
I’ve co-founded about 25 startups, and I’ve advised an additional similar number over the years, and some big companies as well. The attraction is that if you invent something in the academic ivory tower, you can publish it, but that doesn’t necessarily have any impact. A few other groups might read it. But to actually turn an invention into a fully debugged product and manufacture it at scale and make sure that it’s accompanied by adequate instructions and training, that really requires that you get it out of academia and into the real world via companies.
Some people are disgusted by the idea of messing with nature. For example, they don’t want to eat GMOs. Why?
In terms of GMOs, we want to draw red lines in the sand, but there’s a tendency before we understand things to draw those lines with the wrong criteria in mind. For example, I don’t think germline changes — changes made to embryos or eggs or sperm-producing cells that are transmitted over many generations — are the right place to draw a line. Or recombinant DNA—combining genes from different species.
If you change an A to a T or a C to a G or any of those things, that’s so close to what happens in nature that it’s almost impossible to tell the difference between a change that occurred due to CRISPR from one that occurred due to nature. It seems to me that we should be regulating the outcome, not how we get there. To me, I’m scared about GMOs, but I’m scared about plants being engineered the way they were throughout the history of agriculture, where you introduce thousands of random mutations, any one of which could be an allergen. I don’t think random is necessarily good. My experience with randomness is it tends to be bad. Engineering gives us a chance to ask whether a change is safe and effective.
Why is germline engineering the wrong line to draw in the sand?
You should say, what is it that we don’t want to happen in germline? If you have a genetic mutation in your family, the standard medical practice is to get an abortion or to use in vitro fertilization. Engineering sperm is potentially much safer than engineering embryos or even IVF and abortion.
You’re turning stem cells into human brain tissue in the lab. How far along are you?
The largest structures we’ve made are on the order of half a billion cells, which is larger than a mouse brain. It’s not really a macho size thing yet; it’s just exploiting the ability to get flow through capillaries. We can now get most of the major cell types of the brain. We can make oligodendrocytes, which wrap the myelin sheath that insulates neural connections. We can make endothelial cells, which is really critical, because if you don’t have endothelium, you don’t have capillaries. We’re trying to make larger, more complex organoids that have good blood flow.
And what are the applications?
There are three applications, in order of proximity: One is testing genetic variants. Does any particular mutation cause a particular brain dysfunction? We even have an Alzheimer’s model. If you take cells from late-onset Alzheimer’s patients and age-matched controls and use them to grow brain organoids, you can see how they develop differently. The second application is testing new drugs or new therapies. They could be electronic devices; they could be gene therapies. And then the third category is transplants.
What would be the use of a brain transplant?
The main use that’s been discussed so far is for Parkinson’s. Dopaminergic neurons die, leading to movement and motivational problems, and we’d like to replace them. But they could also be for cases of epilepsy. Any part of the brain where you can introduce some structure that’s capable of making new connections would be a plus. Another use is regeneration across breaks that normally would cause paralysis.
Okay, so we’re not talking about entire brain transplants. There’s a joke that the only organ that’s better to donate than to receive is the brain.
No, no, no, just pieces.
Might people add brain tissue for extra IQ points?
For it to be used in healthy people, it has to be exceptionally safe. But I could imagine that being quite safe.
I think doing experiments on humanlike artificial intelligence would be unethical.
Are there applications of these brain organoids to artificial intelligence?
Oh, that’s the fourth category. The human brain is pretty far ahead of any silicon-based computing system, except for very specialized tasks like information retrieval or math or chess. And we do it at 20 watts of power for the brain, relative to, say, 100,000 watts for a computer doing a very specialized task like chess. So, we’re ahead both in the energy category and in versatility and out-of-the-box thinking. Also, Moore’s law is reaching a plateau, while biotechnology is going through super-exponential growth, where it’s improving by factors of 10 per year in cost/benefit.
Currently, computers have a central processing unit (CPU), often accompanied by specialized chips for particular tasks, like a graphical processing unit (GPU). Might a computer someday have an NPU, or neural processing unit — a bit of brain matter plugged into it?
Yeah, it could. Hybrid systems, such as humans using smartphones, are very valuable, because there are specialized tasks that computers are very good at, like retrieval and math. Although even that could change. For example, now there’s a big effort to store information in DNA. It’s about a million times higher-density than current silicon or other inorganic storage media. That could conceivably in the future be something where biological systems could be better than inorganic or even hybrid systems.
At what size should we think about whether lab brains deserve rights?
All of these things will at some point be capable of all kinds of advanced thinking. I think doing experiments on humanlike artificial intelligence would be unethical as well. There’s this growing tendency of computer scientists to want to make them general purpose. Even if they’re what we would call intellectually challenged, they would have some rights. We may want ways of asking them questions, as in a Turing test, but in this case, to make sure we’re not doing something that would cause pain or anxiety.
Will we ever develop into something that calls itself a new species? And could there be branching of the species tree?
It’s a little hard to predict whether we’ll go toward a monoculture or whether we’ll go toward high diversity. Even if we go toward high diversity, they could still be interbreedable. You look at dogs, for example. Very high diversity, but in principle, any breed of dog can mate with any other dog and produce hybrid puppies. My guess is that we will go toward greater diversity and yet greater interoperability. I think that’s kind of the tendency. We want all of our systems to interoperate. If you look into big cities, you’re getting more and more ability to bridge languages, to bridge cultures. I think that will also be true for species.
Do you think your greatest contribution to humanity will be something you’ve done, or something you’ve yet to do?
Well, I hope it’s something I have yet to do. I think I’m just now getting up to speed after 63 years of education. Aging reversal is something that will buy me and many of my colleagues a lot more time to make many more contributions, so you might consider that a meta-level contribution, if we can pull that off. The sort of things we’re doing with brains and new ways of computing could again be a meta thing. In other words, if we can think in new ways or scale up new forms of intelligence, that would lead to a whole new set of enabling technologies.