WORLD SCIENCE FESTIVAL-CRISPR BRAVE NEW WORLD

ROBERT BENEZRA, CANCER BIOLOGIST: Good afternoon. Thank you all for coming on this beautiful Saturday afternoon. I’m sure we’re going to make it worth your while as we discuss this exciting new technology that’s revolutionizing our fields. And we’ll discuss both the science and the ethics of CRISPR technology with our expert panel. Just a show of hands how many people were at the science party last night? Any? Some.About a quarter. So we’ll go through the basic biology to describe it as well as some of the newer translational biology attempts to use this to correct genes in various organisms. And I think this is an incredibly powerful technology. With great power comes responsibility and in some cases, a little bit of danger. And I think these are all important issues to discuss and that’s what we’re here to do today. So without further ado let me introduce our expert panel. Our first guest is a molecular biologist and passionate advocate of citizen science. In 2009, she co-founded Genspace NYC, the world’s first community biotechnology laboratory. In 2011, Genspace’s groundbreaking programs were awarded the prize for best social study in synthetic biology at SB5.0, which is the leading international synthetic biology conference. Please welcome Ellen Jorgensen. Also with us tonight is post-doctoral research associate at the Rockefeller University. He has adapted genome editing techniques including CRISPR to investigate the genes and the neural circuits that control behaviors in mosquitoes that carry diseases such as dengue and zika virus.

Please welcome Ben Matthews. Our next participant is a core faculty member and Assistant Investigator at the Genome Center and the assistant professor at the Department of Biology and the Center for genomics and systems biology at NYU. He’s a bioengineer creating new tools to understand the impact of genetic changes on the nervous system and cancer evolution. He’s a recipient of the NIH pathway to independence award and is a next generation leader for the Paul Allen Institute for Brain Science. Please welcome Neville Sanjana. Neville. Also joining us is an Associate Professor of law and affiliated faculty at the Innovation Center for Law and Technology at New York Law School. His focus is on how scientific developments affect patent law and litigation. He was also a fellow at the Center for Law and Bio Sciences at Stanford Law School and a patent litigator at Gibson Dunn & Crutcher in New York. Please welcome Jacob Sherkow. And last but not least, our final guest today is director of the International Referral Center for Inherited Retinal Degeneration. He has been working with embryonic stem cells since 1992. And in 1995, he created the first mouse model for a recessive form of retinitis pigmentosa by applying genome engineering to ESL technology. He has co-author of 147 peer-reviewed publications and is the editor of a book on the uses of genome engineering in the field of stem cell research. Please say hello to Stephen Tsang.

OK so let’s begin our discussion at the very beginning. CRISPR, as some of you may know, is actually an ancient immune system that was developed in bacteria to protect- to protect against their main predators, which are viruses. And we have been able to harness that into a powerful gene that editing technology. Ellen can you walk us through the story of how CRISPR was first discovered and how we decided to use it for editing genomes?

ELLEN JORGENSEN, MOLECULAR BIOLOGIST: Well actually, the first mention of it was way back in ’87. And you know, a Japanese scientists, it was a sequencing an E.coli gene. And at the very, very end of his paper, there was a little note saying, well I saw these weird little repeats when I was doing the sequencing. I had no idea what they are. And we were talking about it. You look at that and you kind of get chills because you realize what was happening and the trajectory that it was going to follow when you see something like that. And then it really is a story that covers several continents. A lot of the early work was done in Europe by a scientist named Mojica who was studying Archaea in a very salty bay near where his research institute was. And he noticed these things. And he ended up compiling many, many sequences of single celled organisms and finding these repeats in all of them.

[00:04:47] JORGENSEN: And when you see something happening over and over and over again in nature, you start to think well maybe there’s something important about this. And he was actually the first to propose that it was a bacterial immune system. So then, people started dissecting it. And it turned out that there were three different types of CRISPR systems. The one we ended up hijacking is the simplest. And it takes a certain amount of energy, some proteins, to go through the motions of taking pieces of DNA from invading bacteriophages, plasmids, anything that would threaten the life of the bacterium, and insert it between these little repeat regions. And that’s the targeting part of CRISPR in the wild. So that= that system ends up making two separate RNA molecules that are then processed which takes more proteins and end up coming together and then attaching to the Cas9 protein, which is the one that does the actual cutting of the genome as we saw in the movie. So there’s a couple of different components to the system and it can be fairly complicated. However if you just artificially make this targeting RNA, you don’t need any of those enzymes. All you need is essentially, if you think of it, as a guided missile, the Cas9 is the warhead and the guide RNA is the targeting system. And if you make it artificially, you don’t have the need for all of that- that other machinery. And it becomes very, very, very simple. You need one gene to make the Cas9 and one translational unit to make the guide RNA. And voila you’ve got the- the whole thing in one package.

BENEZRA: Can you say a little bit about how the protein and the guide are delivered in various systems? What types of vehicles do we use to get the introduced?

JORGENSEN: Well see that that’s the thing, it is, you know, it’s-it’s a great system but the devil is always in the details. So if you’re going to modify something like yeast, which is a unicellular organism and there are many-many ways to coax yeast to take up DNA, you can put these two things on a loop of DNA called a plasmid, put them into the yeast, and the machinery will transcribe and make the Cas protein and make the guide RNA. And you’re off to the races. It gets a little more difficult if you start talking about higher cells. If you’ve got them in a petri dish, you can do the same sort of thing. There are systems if any of you are doing that sort of stuff in your labs, you know, you can transact with lipofectamine, you can do various different ways of getting this plasmid into the cell. The difficulty comes if you’re trying to now do it to either a germ line. So an embryo, or trying to create, say a genetic defect in a fully grown person or animal, because then you have to actually get the CRISPR into the target cells. And I think there are some people here who are going to talk about the places in the body that are easily targeted at this- at this time.

BENEZRA: But at the same time, people are building mice that express this enzyme in all tissues. Right. So they could potentially deliver the guy and have an alteration in whatever tissue they want as a model for looking at the consequences. I think that’s- that’s…

JORGENSEN: That’s always a little dicey if you’ll pardon the pun. Because the system isn’t perfect yet.

BENEZRA: Correct.

JORGENSEN: And so if you’ve got it on all the time, there’s a low level of off target stuff.

BENEZRA: We will talk a lot about target. That’s a very important point. But I’ll open this question to all of you. I mean are you surprised how quickly this is sort of revolutionized the field? I mean after all, there were other gene editing technologies around. Yet this has captured the imagination of scientists so rapidly and is being implemented so rapidly. Can some of you describe why you think that might be why is that happening? Ben?

BEN MATTHEWS, BIOLOGIST: Then I can jump in real quick. So I work on a mosquito. Aedes aegypti, which is sort of a nontraditional genetic model organism which just means that up until five or ten years ago, we did not have tools available to easily edit its genome and so we would turn to other insects to model the traits that we were interested in studying. And I went to a meeting at which people were discussing how to use some of the first generation of genome editing tools like Zinc Fingers and TALEN and between the time that they announced the meeting and the time that the meeting actually happened, there were about three or four papers that came out on this new kid on the block this CRISPR tool. And I remember going to this meeting and they completely reshuffled the schedule. They invited all these new speakers and everybody I talked to who had done it in cells said it’s too easy for you not to try. And I think that was really the lead for me. Was zinc fingers and TALEN were great, except they relied on engineering a protein for every specific gene that you wanted to target. And to do that involved either a lot of labor or a lot of money and a lot of time to outsource that to other people. And this involves one single gene, as Ellen said, and then an RNA molecule which we can make in the lab in, you know, hours rather than days, weeks, months. And so it truly was too easy not to try. And so we tried it and it basically worked first time. So I think speaking as somebody who is applying these tools after our colleagues have developed them in systems where you can do this much more quickly, the ease and the efficiency of it was the big driver for us. Simplicity. And then also this issue of working in essentially every organism that’s been tried.

BENEZRA: It’s amazing. It’s quite remarkable. Yeah. You want to jump in there?

[00:10:33] STEPHEN TSANG, CLINICIAN-SCIENTIST: We can talk about the clinical use in humans as a little bit later. But I saw research used in my lab to generate knockout mice. It used to take at least nine months and you need to pray that your embryonic stem cell can become the sperm, germ line.

BENEZRA: Just say what a knockout mouse is. Just in case some people don’t know.

TSANG: So we have to do tissue culture to- to generate the mutations in one of the two chromosomes of a gene. And then you hope that these embryonic stem cells can become a sperm and then past the mutation onto the germ line. The germ line means that you either are- or what side of sperm they in the germ line- but then- so it- so the stem cell to become a sperm, that takes about the whole process of making chimeras, and all this takes about like nine months. And there’s a lot of- when I was a graduate student, a lot of things can go wrong in between that-your cell will just not become a mouse-will not become a sperm. But with CRISPR, you can on Monday, you can design your guide. And then on Wednesday, you can inject the oocytes and now you get- because you do not need to use embryonic stem cells for some of the manipulations. So in three weeks instead of- and you can multitask. So there’s no- it has been shown that you can do seven knockouts in the same mouse. We can do- There’s no reason you cannot do 15, 30 in the same oocyte. And then- and it can. So-so by Wednesday, you inject all this oocyte. In three weeks, you have your knockout mice already, as opposed to nine months. So they had really had that- now you can do multiplex, complex traits. To look at different genes interacting in the same-same mouse as opposed to we do one gene at a time, taken nine months to knockout one gene.

BENEZRA: Right. So I think that brings up an important point which is that much of the conversation has revolved around treating disease. And ultimately perhaps editing genes in human embryos, which is a very controversial point which will we’ll come to. But many researchers are saying the real revolution is at the bench and that we can learn things that we- very rapidly from model systems, model organisms that will teach us about disease and disease processes that ultimately will lead to drugs. As opposed to using this editing technology to correct the genetic defect which has- which has its difficulties. So let’s spend a little bit of time maybe each of us can describe the significant ways science- scientists are using CRISPR at the bench. We’ve talked about some means-do you want to add a little bit more to…

NEVILLE SANJANA, BIOENGINEER: Sure. Yeah. I’m happy to talk about what I’ve done so. I just want to echo the point that you made, which is that really the quantum leap forward in my mind with CRISPR and CRISPR proteins like Cas9 is that genome engineers- like when I started my post-doc, post-doctoral work in a bio engineering lab about five years ago, I was using one of those previous generations of genome engineering technologies that you mentioned, TALEN nucleus, which come from plants or- sorry from plant pathogens, and so with these-there’s a long history, actually, of genetic modeling for science for using model organisms like mice to model human diseases to get some-some insight into these diseases. And really in the last 20 years that we’ve had programmable nucleuses, of which CRISPR is one kind of programmable nucleus. It’s always been kind of out of the grasp of most biological scientists to use these tools easily.

And the reason is that just what you said. Making proteins, protein engineering is hard. So imagine the genome, which like the human genome is three billion base pairs long. So every time you need to target a different region of the genome, you need to build a new protein. And that, you know, you can just take my word for it that at the bench, when you do this, especially with the programmable nucleus, like zinc fingers or TALENS. It is quite difficult. These repetitive protein domains and rearranging them is just not super easy. The quantum leap forward with CRISPR Cas9, is that the nucleus component Cas9 is actually generic. The thing that actually comes onto the DNA and cuts it and it’s guided to a specific location in that big genome by a small piece of DNA or RNA. And one technology that we have really well- well done in our labs, actually we don’t even do in our labs, we just send it out to a company and the next day it just arrives in a tube. It’s overnight. You can get these-these reagents that can take that generic Cas9 nucleus and move it around. So that-so that’s really I think the if you, you know, take away one thing, it’s that compared to, you know, this field of genome engineering actually existed for a long time. The real difference is it just got a whole lot easier in the last few years with this CRISPR-with the CRISPR systems. And so I’ll give like the briefest bit of what I do with CRISPR. So one of the- there’s, as you’ll think here from the from folks here, there are many, many things that are possible with genome engineering tools.

[00:15:28] SANJANA: One of the things that I’m very excited about is, because it’s so easy to build these little DNA guides, something that I and others in the lab I came from in John’s lab and in my lab now at the New York Genome Center in NYU, is we’re developing large libraries of thousands to hundreds of thousands of CRISPR reagents to target many different places in the genome. And so the kinds of questions that we can ask when we can easily build genome engineering reagents are just a different scale of question. So one of the first things we did is we took melanoma cells, so you know there’s people get skin cancers, which are called melanomas, and in when these-these skin cancers are treated with certain drugs, often what happens, what’s observed in the clinic, is that there’s resistance to the drug. And so what we did is we took a library of CRISPR reagents that can knock out every gene in the human genome and we combine- we basically said what specific gene mutations of all the genes in the genome, this kind of question you couldn’t ask before. Of all the genes in the genome, what specific genetic mutations could lead to resistance to this drug, which was called Dabrafenib, it was approved in 2011 by the FDA for melanoma. And so and what- you know because people have been observing this resistance to Dabrafenib in the clinic, it was really helpful to try and get this overall view instead of collecting all these patients, you know maybe hundreds of patients, to try and then sequencing to see OK, what how- how did their tumor evolve to get resistance to the drug? Here we could try and prospectively predict this. So-so that just kind of one thing you can do when it’s easier to do genome engineering than with those previous things like zinc fingers and TALEN where making new proteins is really hard. There was a very long answer to a simple question.

BENEZRA: But a great-but a great answer. I think delivery systems are evolving as well. In my laboratory, we actually used adenovirus to deliver CRISPR the protein as well as the guide RNA. And we can put this into accessible organs. One thing the adenovirus causes the common cold, to get rid of a common cold virus and it can actually be used in to insert gene or to manipulate genes in the lungs of mice. If you have them inhale with adenovirus. So we could look at potential causes of lung cancer for example. And as Kate mentioned, we also use it in the brain. We can actually inject these viruses into the brains of mice and ask whether or not certain genetic lesions that we see in humans can actually cause or drive brain cancer in the molecular systems. So I think these are important.

TSANG: So there’s one more use, say in medical genetics where you frequently see patients and now they with the next generation sequencing technology. The most common result they use for sequencing patients, is a variant of unknowns significance on some new gene. So before we can validate on whether or not this new gene, or even list of genes, can get a five out of 10 candidates from- frequently we see patients with only, maybe say, private mutation in that family or private mutation has not been described before. But now with the CRISPR Cas9 of the system, either you can test -get functional test validation. These vary of unknown significance to turn them into significant if you have had functional as a either in patients’ cell lines or in mice. So before we take nine months to make a knockout mouse, it’s not very practical. But three weeks is practical. Sometimes you can make the mice in patients’ cell lines to get a functional as a.

BENEZRA: Ben, you want to say a few words about how you are using CRISPR technology in the mosquito as well?

MATTHEWS: Sure absolutely. So I think we’ll talk a little bit later about something that’s been in the news, which is called gene drive. So we’ll save that. But what we mostly do, and I work in Leslie Vosshall’s lab at the Rockefeller University, and she’s been interested in the smell of-or the taste and smell systems of insects for the majority of her career. And so I joined the lab and became interested in the sense of taste and smell in the mosquito. Aedes aegypti. And so this is the- what we think is the primary vector of Zika virus, as well as chikungunya, dengue, yellow fever and other viruses in that camp. And so what we do is we take the ease of these new generation of genome editing tools and we’re kind of working our way systematically through different gene families that we think are involved in the mosquitoes’ ability to target humans because these mosquitoes are very anthropophilic, which means if you’re at a zoo, these mosquitoes will bite you over every other warm blooded animal in that zoo.

[00:20:13] MATTHEWS: That’s just what they’ve done. And they know that we are an easy source of protein and the female mosquito takes a blood meal in order to develop eggs. So this is a critical component of their lifecycle. So evolution has acted on their chemo sensory systems to make them extremely good at identifying and biting human beings. And so just to back up, what we’re doing is we’re now moving beyond a scenario where we could target one or two or three genes in order to ask questions about what the genetic pathways are that control taste and smell. And now for example, I can mutate 10 or 12 or 15. And ultimately, we’re going to go much higher than that. So what we do is, we make stable genetic mutants. So the equivalent of a knockout mouse, we make a knockout mosquito and then we test them in various sensitive behavioral lapses to see whether or not they have any deficits in their ability to find and target human beings.

And so that’s kind of the very basic application of CRISPR which is just to delete a gene or inactivate a gene. But the other thing that I’m really excited about now is we’re actually using it to insert new DNA into the genome of the mosquito. And if we do this in the right place, if we do it at a location in the genome where a gene that is expressed in their olfactory system sits, then we can actually use the regulatory elements of that gene to now express proteins in the cells that that endogenous gene would normally be expressed in. And we can control neural circuits through light, for example, or your temperature. And then we can ask questions that go beyond just pure genetic lesions and say, “Are these particular sets of brain cells, these neurons, are they involved in the mosquitoes’ kind of behavioral output in response to human odors?” So the ease, I’m just going to echo what we’ll probably say this all afternoon, but it really is the ease of this system that lets us take these experiments, which we’ve thought about for a long time and our colleagues who work in organisms like the fruit fly have been doing for a long time, and I used to work in the fruit fly and they’re not that interesting in terms of their behavior. And if you’re interested in something like disease transmission, you know,

BENEZRA: Someone might take issue with that.

MATTHEWS: I love fruit flies, I do. But if you’re interested in the transmission of an arbovirus virus, like zika virus, you cannot model that in a fruit fly. And so now we’re taking these classes of experiments that we could only dream of 5 or 10 years ago and now we’re actually starting to implement the reagents that we need to ask those questions.

BENEZRA: Thank you Ben. So Jake I want to bring you in on the ethics. But before I-but before I do, I want to introduce the concept of manipulating the human genome. So that’s obviously a very important point to discuss. Do we want to do that? How do we want to do it? I think there are two general categories. One is manipulating the genome of let’s say a patient who has a genetic abnormality and we want to mutate and we want to change the genetics of a particular cell type within that person versus the idea of manipulating the human germline, where we introduce change into a very early embryo and that person will carry that change throughout their life as well as pass it on to their progeny. OK so those are two very different categories of human genetic manipulation. And there are potential problems with both. Does anybody want to tackle the question of what the problems might be in doing that? Let’s say first in the somatic cell and then ultimately in the germline?

JORGENSEN: Well, I mean as you know there are technical problems first of all. So the second half of the CRISPR story is that what this-this guided missile does is it makes a double strand break in DNA. And that can be catastrophic for a cell. So there’s all sorts of mechanisms to fix this and there are two major pathways and they both can make mistakes. One just joins the two broken ends together and sometimes deletes a letter or adds a letter. And that usually messes up the gene where it happens. And then the other one goes, “Oh my God there’s a break here. Is there anything else I can use to fix it that looks kind of like this?” And it uses a homologous piece of DNA and literally copies it. And so if you use that second pathway, what you have to do along with the CRISPR is put in another piece of DNA that’s like a Trojan horse. The two ends of it have DNA similar to where you make the break and then you can put whatever you want in the middle. And there are size limitations to that. But there’s no limitation as to what that thing in the middle might be. And these two pathways operate in cells and some types of cells like one pathway better than the other. And so sometimes you can’t control that. Because the one you want is the Trojan Horse one, because then you could do just about anything. So there’s that and then there’s also, as you said, this-this percentage right now there’s a tremendous amount of work going into making the system more accurate and also bringing our level of sensitivity to seeing if you can detect an off target effect.

[00:25:34] BENEZRA: So by off target, Ellen means changes that happen at some place other than the intended target and the consequences of that change. We don’t know necessarily. We don’t know where else the changes being made or what are  the long term consequences of that. So Jake, who do you think should be. I know it’s a tough question, but who should be responsible for deciding when it’s OK to try to introduce these changes. Let’s say first let’s stick with somatic cells just in our body as opposed to the germ line. Should the scientists be making the decisions? Are there ethical considerations that others need to- need to be thinking about?

JACOB SHERKOW, LAW PROFESSOR: Yeah I mean that’s-that’s essentially what’s happening now. I mean, you know we have various regulatory layers in place to ensure that to the extent we’re doing experiments on humans, that we’re doing them in the most ethically responsible way. We have institutional review boards that are, you know, these are boards typically of scientists at any given institution that review whether or not professors who are doing this kind of work are again kind of doing them in ethically appropriate ways. We have several regulatory agencies in the United States, such as for example, the Food and Drug Administration National Institute of Health. That kind of also take a look at this to ensure that if you’re either receiving federal funding to do some of this work or if you want to try to push a drug or push a therapy through clinical trials, because you want to be able to sell that. That you’re that you’re not only doing it in a way that appears to be therapeutically effective, but you’re doing it in a way that ensures that it’s kind of being done with the highest sense of ethics in mind.

SHERKOW: Thus far, the system has worked. I would say, extraordinarily well. That’s not to say that there haven’t been lapses and there haven’t been some pretty critical lapses where people have died. But if you look at all of the clinical trials that are being conducted with kind of all of the work including gene therapy work that has been you know that has been done over the past five years 20, 30, 40 years. The vast majority of these cases go forward without any incident. So I think with that in mind, I think the perfect is the enemy of the good. I don’t necessarily think that we should kind of kind of uproot the entire system of regulatory approval that we have just because CRISPR’s really used to use. Just because it’s really powerful, just because it seems to be effective. I think that the system that we have of ethical review, if you want to think of it that way, has worked well this far. And until there is some evidence, or kind of some model to suggest that for whatever reason that’s not going to work here, I don’t necessarily think we should reinvent the wheel.

JORGENSEN: Can I jump in for a second? If you notice the two things that you mentioned were selling and government funding. So there is no law in the United States against engineering human embryos. You cannot use NIH funding without review to engineer human embryos. But there’s- in certain countries it’s banned, like Canada. It’s not banned in the U.S. So the control in the U.S. is all around-

BENEZRA: Wait, what type of engineering are you talking about in terms of…

JORGENSEN: Germline engineering.

BENEZRA: Germline engineering of a human embryo.

JORGENSEN: Of  human embryo is not illegal in the United States. It’s- you can’t use government funding to do it. It’s like the stem cell story. At one point, you couldn’t use anything except a very small number of stem cell lines. But the state of California set up its own fund. And because there was no law against it in California or in the United States, but the funding was cut off to everyone because of the laws. The other thing is the idea of selling. So if you’re not selling anything, then you don’t -the FDA doesn’t interact with you. I mean the FDA- their mandate is that if someone tries to put a product and get it approved for use, that it’s safe. But if you’re not trying to make a product, I mean, I just came back from a two day meeting at the National Academy of Sciences around this, but you know it’s famous that a lot of things on engineering plants are slipping through the regulations. So it’s interesting that human embryos while in Europe and Canada it’s it’s outlawed without permission from the government.

BENEZRA: But let’s be clear. I mean nobody is using CRISPR yet to modify a human embryo or, that we know of.

[00:30:03] JORGENSEN: But wait a minute. Someone in the UK is doing research and they are the first person to get, in November I think it was, of this year or maybe even later. A researcher in the UK got permission to engineer early stage human embryos.

BENEZRA: And this brings us to a very important ethical consideration, because in my mind, there’s a very thick line drawn in the sand that we have to discuss and it was discussed a little bit last night. Jennifer Doudna, who you saw who was one of the early adopters or one of the early inventors of CRISPR, has said that one of her greatest nightmares is to wake up one day and find out some CRISPR baby, some embryos had been modified with CRISPR, was born. And she would be very disturbed by that. And George Church, on the other hand, felt well you know we’ve been through this before. You know people were afraid of in vitro fertilization and they were worried what would happen with those children and that turned out to be perfectly fine. Well, I think it’s a different order of magnitude actually because even if that CRISPR baby is born, it might look perfectly fine when it’s a day or two old but who knows what’s going to happen 10 years from now. Right? And how do we ever do that first child? I mean I think that’s an ethical consideration that people really have.

JORGENSEN: Why is that different from test tube babies?

BENEZRA: Well because I think manipulating the genome is very different in my mind than putting together a healthy sperm with a healthy egg and then developing. Yes, there are potentially inherent risks, but it seems to me inherently safer than manipulating a piece of the genome. The long term consequences of which, yes, we know the effect. Even if you can get rid of all of the off target effects. How do we know that that specific change won’t have long term consequences in the human that we cannot model.

JORGENSEN: What if you’re just bringing back one base pair sickle cell case or something?

BENEZRA: Do you know-can you say unequivocally that that won’t effect chromosome dynamics 20 years from now in some cell type?

JORGENSEN: Well we know what it is in normal people. And if you just bring it back to what it is in people that don’t have the disease.

BENEZRA: Right. But potentially- off target effects then still become still become an issue.

JORGENSEN: Yes, they do.

BENEZRA: And we can do that. So the question becomes should we do that with CRISPR, or should we do it by screening embryos for the correct-for an embryo that has you know all the correct wheels, assuming it’s a recessive mutation. And the parents-the parents couldn’t possibly contribute to a mutant phenotype. Do we do that or do we do CRISPR? I think that’s an important consideration. Anybody else want to jump in here?

SHERKOW: Yeah. I mean so

BENEZRA: I put you on the spot.

SHERKOW: Sure, sure. So I suppose that kind of weight of ethics falls on me. Fine. I mean- I mean so-so you know look, the technical concerns, I think, are serious. And when we’re talking about risks, I think one of the things to think about is that even if it is true, and it is likely true, that the risks are small the harm, the potential harm is great. And so I’m not necessarily sure that I agree with the notion that well you know IVF totally fine and normal and it was fine that when we tried it the first time back in the 1970s. But oh man CRISPR edited humans let’s, you know, kind of order of magnitude separate. Again these are instances in which the risk profile may be different, but the harms are still, the potential harms I should rather say, are still pretty substantial. I think that we are I hate to use this word. We are lucky that IVF has worked out with such grandeur the way that it has to the point where there’s a substantial fraction of the population that has been born through IVF. In fact, looking out at this room and kind of sadly needing to exclude anybody who is roughly over 45 or so, you know assuming that we have more than you know, let’s say 30 to 50 young people out here.

One of you is that statistically likely to have been born through the IVF process. At least here in the United States. We’re lucky that has worked. And again, you know, to the extent that someone is going to engage in CRISPR edited humans, and to be clear, I I think that a you’re ethically crazy to try and b, somebody is going to try and I will, you know, I will-I will take bets with anybody in the audience that it is going to happen soon. Just because again, it is so easy and there are scientists out there whose I’m not sure how to put this,  who’s ethical bar is lower than your median scientists in the United States. So it’s kind of, you know, that is going to that is going to happen. So I mean I think it is dependent upon scientists here in the United States to be the people who are the vanguard. You know, saying that the risks are high, the potential risks are high, and therefore this is something that we should think about before we engage it.

BENEZRA: So that’s one of the potential dangers. Any other potential dangers? People think of terrorism as a possibility. Does anybody think of that as a realistic possibility, using CRISPR as a terrorist weapon? It’s been in the news. I think it’s unlikely. But anybody…?

[00:35:12] JORGENSEN: You’re talking not in humans

BENEZRA: Well, releasing a virus that would modulate genes of whoever it infected. An unlikely scenario? Not something we need to worry about?

TSANG: There’s no difference from 30 years ago. And when recombinant DNA comes out. Right?

BENEZRA: Right. So there was a big discussion of recombinant DNA. People were worried about that. And scientists actually called their own moratorium on that research. They’re not doing that today. Right? There’s no moratorium as far as I know. Obviously, manipulating the human embryos. Yes there is essentially an ethical moratorium, but it’s still possible to do.

JORGENSEN: There was a big meeting earlier this year where a lot of people actually some people from the original Asilomar conference weighed in at that meeting and said it was very similar.

BENEZRA: And how did the discussion go?

JORGENSEN: Well they came out with an actual set of guidelines from that meeting.  So

BENEZRA: Do you want to outline what some of those were.

JORGENSEN:I don’t know in-I don’t know in detail. I mean obviously, we need to start talking about Gene drives if we’re talking about doomsday scenarios.

BENEZRA: That brings us to Ben. Ben, tell us about gene drives.

JORGENSEN: Our doomsday guy.

BENEZRA: Our gene drive guy.

MATTHEWS: So gene drive, you may have heard about. And just to be clear, this is not something that we personally do in our lab but it’s a class-maybe I have you know one of my graduate student underlings at the bench right now. But the basic idea is since CRISPR originated as a bacterial immune system. So in essence a defense system against invading pathogens. And we are already repurposing it in terms of targeting sequences that we want to target. The next step is to now genetically introduce the components of CRISPR. In this case, it would be the Cas9 nucleus. The gene encoding it, if you put that into the genome of a mosquito, for example, along with the appropriate guide RNA and then whatever other payload you want. And so we have colleagues who were doing this in mosquitoes that carry malaria. And what they do is they also along with the Cas9 and the guide RNA, they express the genes that express antibodies against the plasmodium which is the malaria parasite. And so the key here is that when you put this whole package into the genome of the mosquito, and then which they haven’t done yet but they talk about doing, if you release it into the environment, this now becomes something that is going to break the rules of mendelian inheritance. So the one animal you release into the environment will mate with a wild type animal and normally that animal would have you know a 50/50 chance of inheriting the copy of this trench that you put in. But the gene drive system actually will convert every heterozygous animal into a homozygous animal. Meaning that within that lifespan, you will change the genome of the offspring of this mating so that they now have two copies of the transgene rather than a 50 percent chance of having one copy of the transgene.

And so when you do this, the mathematical modelers tell us that this will drive the gene through the population and something like an insect has a very quick life cycle. So it can be as quick as two or three weeks. And so if you were to start with a small release of these genetically modified animals, you could conceivably convert the genomes of the wild type population in this area to carry the transgenes that you’ve introduced. And so the good- the promise of this is I think pretty obvious. You could potentially kill a certain species, or you could express transgenes that would essentially render the animals sterile, or you could have them express antibodies against plasmodium or against the viruses that they carry. And the risks are that it’s a very hard thing to rein back in. Once-once you’ve released them into the environment. And you know we can talk. There are potential risks to that ecosystem if we could achieve our goals of actually locally extincting an entire species, which is what some people would like to use us for. And we just-once you start pulling out those threads in the ecosystem, we don’t know what other insects would come over to take over that niche. We don’t know if these are the food source for various other insects. You know you can have cascading effects that we don’t completely understand.

JORGENSEN: What do you think about the idea of sending another gene drive after the first one if there is something that’s wrong? Because that’s what George was talking about last night.

MATTHEWS: And I mean, that sounds like a sci-fi movie in the making and it really does. And you know I like to think of-you know there have been these experiments, particularly in geographically isolated places like islands. You know so there will be invasive species. You know in the Galapagos, they have goats that people dropped off and now they’re flying helicopters around shooting at the goats in the helicopters because they’re destroying native habitats for native animals. And Hawaii, you know you had this kind of cascading you know release of various invasive animals. Australia has done the same thing where you release one animal to predate on the first one that you released to have a good effect. So I think if we don’t fully understand the risks of the first gene drive, I would be very uncomfortable sending another gene right after that first one to try and roll it back.

[00:40:33] BENEZRA: But George did talk about releasing these into a contained environment and now I ask, is that a prospect?

MATTHEWS: Absolutely. Yeah, yeah.

BENEZRA: So describe what he was talking about?

MATTHEWS: Yeah. So I think he used the word “dome.” So he envisions building a village that is isolated from the rest of the environment physically. The kind of the built in way to do that is to find an island somewhere. And you know, mosquitoes can fly, but they can’t fly that far. So if it’s isolated enough, you can-they can hop on boats. That’s how they’ve made it all the way around the world. But the idea is that we will absolutely need to test this in some sort of isolated situation before. But I always get back to the idea that we can never fully model an ecosystem. We could try. We can try our best and we can try and show that this will be effective, at least, and safe in so far as we can. You know as a safety. But there’s something uncomfortable about sending something out that you don’t have a really easy way of reining back.

JORGENSEN: We never had this kind of power before.

MATTHEWS: I guess that’s true. Yeah, yeah that’s true.

BENEZRA: So let’s-let’s get back to something a little less controversial. Maybe manipulate- manipulating various organs that-where we could help cure disease. So Stephen, you’re involved in retinal diseases. Is there a way CRISPR could be used safely you think?

TSANG: So I know- I showed the video. These are the results of the first retinal human gene therapy trial. And now the data is-so this is a optic nerve in the eye. So someone can inject viruses, delivery system, on the DNA. So this is- injected a big bubble in the retina. And this is a pre-op before surgery. Before I get the-this is This is the adeno-associated virus that will most likely will be approved sometime next year for the first gene therapy drug. This is the patient. The task is to go from one side to the other side of the room. This is before surgery. So the patient bump into the thing and then time how long does it take from this side of the room to go the other side. This is a patient at Moorfields Eye Hospital in London. There’s also about 30 something patients in U.S. who’s getting the same gene therapy virus. And those states are currently being reviewed by FDA. This is a phase one trial. So these patients more advanced in the phase three trial some of the children who are enrolled in the phase three. So patient bump into going to the wall, get disoriented.

So the task is to go from this side of the room to the other side of the room and they’re being timed, how long it takes. So sometimes they get disoriented. This is monogenic retinal degeneration, a form of retinitis pigmentosa. And I’ll show you how patients look like-how they see the world with retinitis pigmentosa after this. So it takes a long time from this side of the room to go outside. In Philadelphia, the effect, the Children’s Hospital in Philadelphia, the effect peaks around six months. In London, the effect peaks around one year. So now, the same patient a few months after surgery. So they take-they time how long it takes. So it takes about 77 seconds from this side of room to go to the outside. So now this is the same patient six months after surgery. From here and the same patient from here to here.

BENEZRA: And this patient received stem cells?

TSANG: Gene therapy.

BENEZRA: Gene therapy delivered by injection. Into the-into the eye.

TSANG: So the patient went through. So that’s my-most likely your tax dollar, Medicare would cover this. And the price will be, after the data, we can discuss what you think the price of gene therapy. And the reason that-there is the CRISPR pharmaceuticals, I guess I know two of them. They- one of them would claim that 2017, next year, they would do the eye. Part of the reason, this is a blood vessel. And this is the only part in the human body, the only part in the central nervous system you can count individuals cells. The individual light sensing neuron. You can see them. It takes about 10 minutes to get this picture. So this is why all the CRISPR companies are interested to do eye, the same reason why the only embryonic stem cells trial in US is also done in the eye.

[00:45:14] TSANG: Most of the gene therapy trials now is also done in the eye, because you have two eyes. Right. So you don’t have two brains, you don’t have two spinal cords, or two hearts. So there’s already a control already. And then they said it more than 100-a century of history of brain-sensitive psycho- physical testing, almost as a single cell level. So ophthalmologists usually do not trust what patients tell them, what they see. But we can actually test them objectively and then show this imaging. So you can scan this out for gene therapy, CRISPR, or stem cells. You need to-you can count the number of cells to see whether or not they are-because for neuro-degeneration, you’re not likely to improve people. But you want to see if the number of cells survive longer. So you can just count by counting. The surgery I just showed you earlier is outpatient procedure. It takes about one hour. So the eye surgery is outpatient, much easier. Topical- local anesthesia. So local delivery you can- and that’s making my DNA making vectors, making cells, you just need 10 to the 6 cells for stem cell transplantation. The production of cells, bio production is much easier. And maybe not so radical, CRISPR for stem cell transplantation, the eyes here’s an advantage. So for cornea transplantation, we do not do any matching. The eye express a lot of very high level ligand a lot of T-cells just get killed. Next slide.

So again, these are the histological levels. These are light sensing neurons and most of the time, in patients with retinitis pigmentosa, macular degeneration. In macular degeneration, these kind of cells died away. In retinitis pigmentosa, these kind of cells died away. And you can count the number of cells and this is done in histology. The next slide. This is done in histology, as in human retina. And this is done in the life imaging in the same patient take about- less than five minutes apart. So you can almost get histological level in live people. Next slide.

BENEZRA: This is stem cells or this is just returning a gene in…

TSANG: So in the eyes, both embryonic stem cell transplant and gene supplementation, gene addition-

BENEZRA: Not CRISPR. CRISPR has not been used.

TSANG: It’s not approved by the FDA yet. But they think 2017.

BENEZRA: Right. So the gene that was being corrected, can you say a word. How does the gene therapy being done? What was…

TSANG: So CRISPR is a game changer in the sense that before we can think about recessive conditions, because we can only add a gene. Now CRISPR, in principle, you can either destroy a bad copy selectively of a dominant condition. So before CRISPR, gene therapy is directed mostly at recessive conditions, had not been a vision for dominant conditions.

BENEZRA: And the trial that you showed us, that was just a…

TSANG: Recessive.

BENEZRA: Recessive.

TSANG: So retinitis pigmentosa, these people get tunnel vision. They-they will pass the New York State and New Jersey driving test.

BENEZRA: That doesn’t give us much…

TSANG: So they are- of course from a pharmaceutical point of view, is to retinitis pigmentosa is just a test case essentially for them. Macular degeneration, one third of the people in this room will get some form of macular degeneration by 75. So by the time macular degeneration now expect more than 10 million people. By 2020, the number of cases of macular degeneration would double and probably that’s the reason that they are going out currently monogenic disorder. And the big prize, at least for-in neuroscience, will be too cure-do CRISPR for macular degeneration.

BENEZRA: Thank you. So we have a few more minutes before I’m going to open it up to questions from the audience so I want everybody thinking about their questions. But Neville, I guess on this day in which Mohammed Ali passed away of Parkinson’s, I was wonder if you could tell us is there any likelihood that CRISPR technology might be used to replace dopaminergic neurons.

SANJANA: Yeah I mean that’s a really nice kind of inspiring thing to think about because it’s one of these diseases that’s very difficult to treat. So I was going to take your human gene-editing question and maybe a different direction. Because I think people mostly, when you say human gene editing, you just it just sounds kind of devious and evil in some way and so I’ll give you an example of saying that we’re doing the lab in a neuroscience context. It’s not Parkinson’s, but it’s a neuroscience context. That I think is personally like a very exciting direction with human gene editing and so as all of you probably know that if we’re interested in neurons and neuropsychiatric or nerve developmental diseases, it’s very hard for us to go to patients and just take out these neurons and observe them. Nobody is going to approve that. Nobody is going to allow that. You just can’t do that. And so but you know, as a neuroscientist, we really want to study the cells that have the disease, the neurons. And so we have all these fancy techniques for kind of looking at it from the outside of the brain and that’s kind of at the end of the day, it can only be satisfying in a limiting way and so one of the revolutions that’s happened over the last 15 years also has been- we’ve kind of talked a little bit about embryonic stem cell technology. Whether it’s pluripotent- I should say pluripotent stem cell technology, whether it’s in a mouse context or a human context. And what the word “pluripotent” means and the reason why everybody’s excited is it means that this cell is capable of generating any other cell type that’s there in the body.

[00:50:40] SANJANA: And so when we think about cells like neurons, where we just can’t go up to the patient and grab the neurons out, we can use stem cells, pluripotent stem cells to actually make neurons, even in a very quick time period. We do it in the lab in just one or two weeks and we have these things that fire action potentials. They spike like neurons do. They have these electrical depolarizations. And so how does this intersect with CRISPR? Let’s try and I guess bring it back there, which is that one of the nerve developmental diseases that I’m very interested in is autism. And one of the reasons that I think everybody should be interested autism right now is one, the incidence rate is rising quite a bit. But the other thing is that due to genome sequencing techniques, we’re getting we’re doing a better job of characterizing the genetic differences between patients who have- people who have autism and what we might just call like a control population. And my interest is more in the very severe forms of autism. There’s a constellation of symptoms that are not just social disorders, but maybe also include like an I.Q. deficit. Very severe forms. And so what we do is we sequence more and more people and this list of different gene variants for all the 20,000 genes in the genome kind of grows and grows the list that’s associated with autism.

And at the end of the day as a scientist, it’s not super-we’re not convinced when we just say this might be associated with the disease, because we’re taught that correlation does not equal causation. So what we really want to do is say, “does this gene cause the disease?” And so one thing we can do in the lab is that we can take these pluripotent stem cells, stick in the mutation, if we have some easy way of putting in targeted mutations using either of the two pathways you talked about the non-malicious enjoining pathway which is used for knockout. Or homologous recombination, which is used for precise gene editing to put in a precise change. So we have this list of these changes that we observe in these patients. And what we’re now able to do, enabled by CRISPR, we can put them into these otherwise healthy stem cells and we produce an identical twin stem cell. But this one just differs in one place from its twin. It has the mutation that we think might cause autism and then we can differentiate these in the lab into neurons. This cool cell type that we just can’t get any other way. And these are human neurons with the human genome. And so we can look at them at a molecular level, at a genetic level, at the level of neural networks when they’re connected to each other. And this is just, you know, I started as a neuroscience Ph.D. student a little while ago. Now this-I mean, if you came to me in my first year of grad school in 2001 and described the experiment that I just described to you. I would have you know I’ve just been like this is impossible, this is you know-many different things didn’t exist. Maybe the stem cell technology, the ability to differentiate neurons, the ability to do quick genome editing. And so this is really going to impact-it’s kind of like what you said about the variance of unknown significance. We’re great at reading DNA, but not so good at writing DNA. And so as we read all this DNA, we find all these disease associated variants and I care about autism. But there are many diseases. You can pick any disease and this you can use this as a general platform for generating disease relevant cell types that we can then use to do drug screens. All these important things that we think at the end of the day are going to lead us to one, understand disease and two, develop therapeutics to address it in a place like autism we really have no therapeutics.

BENEZRA: So is that done with ES cells or IPF cells? And maybe you can explain the difference between the two.

SANJANA: Yeah so. So without getting too deep into the stem cell holes since I’m, you know I use stem cells, but I don’t consider myself a stem cell biologist. So there are two kinds of stem cells commonly used in both in mice and in humans and those those stem cells are referred to as embryonic stem cells, which are isolated from the inner cell mass of the blastocyst early stage embryo or the or the real. Like just amazing breakthrough that was recognized with the Nobel Prize, induced pluripotent stem cells, IPS cells. And this is, if you think about it, there’s a basic fact of biology you might have learned in high school that actually has wide ranging implications that were only recently realized, which is that all the cells in your body whether they’re neurons, or liver cells, or skin cells, they all have a complete copy of the instructions to make you.

[00:55:01] SANJANA: They all have the same three billion base pair genome. But why do the cells look so different? The liver cell, the hepatocyte, makes fat, it does all this metabolic stuff. The brain cell thinks. Why-you know they all have the same genome, but they’ve evolved in different ways. And this kind of gives you-hints at another question. Is it possible to interconvert, do some kind of cellular alchemy to make like skin cells into neurons or make neurons into liver cells. And it turns out that only very recently we discovered that it’s true. There’s the same copy of the instructions of life in every cell and you can you can interconvert them so there’s work by done by Yamanaka and colleagues that was recognized by a Nobel Prize a few years ago now where they show that they could take fully differentiated skin cells, fiberglass, take a little sample, just kind of you know scratch you a little bit. And by expressing a certain set of genes, they could kind of kick the cell back into its pluripotent state to make it again that kind of cell that can make any other any other cell type. And this is-again it’s obvious right? All cells have the same full copy of the DNA. But the fact that this was just only recently realized, the true potential of it is amazing.

BENEZRA: The real advantage of this is that if you take a skin cell from a patient and you want to induce it to make another cell type, and then replace that cell, it has the exact same histocompatibility. It won’t be rejected by the patient because it has the exact same genetic…

SANJANA: This actually brings up a really good point, if you don’t mind me go back and forth a little pushback. So you might say well why do we need this genome editing? If you have this IPS, this induced pluripotent stem cell, why don’t we just collect all these autistic patients, we just get a little bit of skin from them that’s not so painful. And why don’t we just go ahead and make neurons from those stem cells right?  That should be possible. You don’t need CRISPR. And this is an approach many folks have pursued in the field, especially before CRISPR and a lot of productive science has come from it. But if you think about it, if you take a bunch of unrelated people who you’ve classified as having a disease like autism, they also have a totally different genetic backgrounds like are- even though maybe we’re normal in social behavior, hopefully. You know we look very different. We have a genetic back-we have a difference in genetic background. And so those can be, for scientists, confounding factors.

So the approach that I described is called isogenic kind of stem cell disease modeling. And this is to be contrasted with getting those IPS cells from a bunch of different people. Here what we can do is take one healthy stem cell, maybe you know from me or from you, an IPS cell, and put in all these different autism-so we have a list of a 100 mutations that neither you nor I have, but we can put these guys in. And then each cell is kind of an identical twin to the other cells. Hopefully aside from just this one locus, this one genetic region that that we’ve modified, and so scientists really like this idea of a well controlled, tightly controlled experiment, same genetic background. We love using identical twins a lot in genetic studies. And so this shows you a key difference between just using IPS cell technology alone and adding on this genome engineering. And so I guess my encouragement to you overall is when you hear things about human gene editing, I really think about all the positive things that we can do at the bench kind and hopefully extend it to the clinic too. And really I think there’s a lot of fun science fiction-y things that we can think about also that maybe aren’t, to me, aren’t as relevant because the stuff that’s right in front of us I think is just going to have such a tremendous impact on human health that no nobody in this room-I can certainly not me- can even predict where we’re going to be five years from now in terms of therapy development and disease modeling ability. What I just told you about.

BENEZRA: Stephen, you want to add a little bit on to that?

TSANG: Yeah. So one of you have mentioned now-when we had embryonic stem cells, the current embryonic stem cells trial in U.S. for the retina is that people is on immunosuppressant, because embryonic stem cell does not come from the patient. So the idea is to take your own skin cells, turn them into stem cells, and transplant them back to the retina. So this is being approached now. In Japan, they did one patient, but then if they- so this is before monogenic disorder. Some with retinitis pigmentosa. You still need to correct the mutation before go back to the same patient or else their cells will degenerate again. So before gene editing as a whole and CRISPR, in particular, because of the ease of doing gene editing to correct-gene repair then it’s now possible to envision that you can take the patient’s skin, turn them into stem cells, correct the mutation, and put it back in the same positions. The only thing preventing from them doing that is the cost and the regulation.

JORGENSEN: What about bone marrow? Couldn’t you- for blood cell disorders. If you take it out, engineer it, put it back.

BENEZRA: Several companies are doing this.

[1:00:01] SHERKOW: Yeah absolutely. Yeah even I think prior to that, there are some companies here based in New York that were doing similar work but with B-cell lymphoma. So I mean, that’s you know part of a long lineage of experiments like this.

SANJANA: There are folks working on knocking out the receptor with previous things like zinc finger or nucleus or TALEN nucleus, the receptor…

JORGENSEN: CCR5?

SANJANA: Right, CCR5, CXCR4.

BENEZRA: Yeah right. So we have about 20 minutes left and I should tell you that I can judge our success by the quality of the questions that come from people. Now the pressure is on. If there are any questions in the audience please. Now would be your opportunity. I think we have microphones that can come around and…ethical, scientific or otherwise. There is one. Thank you.

AUDIENCE: This is more of an ethical question but in a world where CRISPR becomes more available to both scientists and everyday people, do you think that our future with CRISPR in it, could suffer more from overregulation from governmental authorities or under regulation?

SHERKOW: Yeah. So I mean I’m happy to step in. If the rest of the panel think that I’m wrong, which is fine. So I think when it comes to new therapies that are developed with the new technology, the historical trend has been to regulate those. To, in other words, add on additional regulations that we didn’t previously see to the current regulations that we have. And my personal opinion, with some notable exceptions in the gene therapy areas especially in gene therapy circa the late 1990s. This hasn’t really seemed to affect the progress of clinical trials that we’re conducting here in the United States to develop these to develop some of the therapies using new technologies that we’ve had. So while the historical trend is kind of, you know, increasing incremental regulation for some of the stuff, I don’t necessarily think it’s been kind of you know it doesn’t really quash all of the technology so far. But that’s not to say that things are going to be different for CRISPR here.

I can envision one world in which are- in which people in Congress think, oh my god this is such a powerful technology and you know we’re just going to kind of let anybody use it. And that’s particularly problematic. And so I can see Congress attaching a bunch of additional requirements to laboratories or receive federal funding to not engage in this type of research much in the same way that we’ve done with stem cells. Or I can also envision a Congress that can’t get their act together and does nothing at all. Which is not that shocking to anybody here, in which case essentially the status quo wins with respect to CRISPR is very lightly regulated if it’s regulated at all,  depending upon what your definition of that is. Ellen I’m sure that you have more thoughts on this.

JORGENSEN: I do just because I just came back from two days in Washington. There’s a National Academy of Sciences panel that has been tasked with that question, basically, is to think about these things that are going to come through the regulatory system with the next 5 or 10 years. And do we need more regulation? Obviously there were a lot of corporate representatives there that were fighting for sort of the viewpoint that really this is no different in terms of the edits that we make. It’s just we’re making them faster so it doesn’t really require a new level of regulation besides the regulation that we have now. And then there were people from other viewpoints, more sort of citizen groups that were sort of -it was a public meeting, by the way, it was streamed live and I think it’s probably going to be online somewhere at the National Academy of Sciences, if you’re interested. But one of the reasons they brought me in is that I run this citizen science lab in Brooklyn and I teach the general public to do CRISPR. So if you come and you take our class, you will be in the lab, you will be modifying yeast with CRISPR, we will teach you how to make the guide RNA’s and it’s essentially just like a class that you would get in a school.

Although I don’t know if anyone’s teaching CRISPR classes. We’re probably teaching classes before the schools are. It’s something that you’d learn at the bench from your from your PI. So right now in the United States, genetic engineering outside of a conventional institution is not illegal. In Europe, for example, you have a site license. In the United States, the only law that sort of overarching is the Homeland Security Act that says it has to be done for useful and peaceful purposes. So what we find at Genspace, because one of the things we are is sort of at pre-incubator space, we have people that pay $100 a month to use our lab, which is a fully functioning molecular biology lab in Brooklyn. And a lot of them are starting companies. And a lot of them have professional backgrounds, or are partnering with people that do and they have ideas. And the idea is do you want to regulate away innovation by- for example the companies that sell DNA are at this point being held to develop very strict guidelines on what sequences they sell. You know they have to screen them. Is this something that could be used as a weapon?

BENEZRA: These genes, things that cause cancer. They’re prevent from selling or…?

JORGENSEN: No, it’s more like select agents like if any part of the smallpox virus or anything like that and they have these very elaborate software things. But there are actually pushing for those- they’re right now there are voluntary guidelines, but they’re pushing for them to become law, because in essence that would give them a not a monopoly on the DNA synthesis business, because it’s expensive to do that screening. And so if you’re a small startup trying to make a DNA synthesis company, you couldn’t compete with them. So there are a lot of very interesting ins and outs- it’s a very good question.

[01:05:39] BENEZRA: Other questions out there? Yeah. Microphone.

AUDIENCE: So you mentioned before that the CRISPR in comparison to older techniques is a lot faster, but a lot more cost effective. But I didn’t sort of lay out like how much. How much would it actually cost to do like some experiments with CRISPR?

SANJANA: It depends on the scale of…I mean, I’m sure many people can answer this here- depends on the scale of the DNA synthesis, to get back to DNA synthesis. You know at kind of the singleplex level, you know something that molecular biologists are very used to doing in the lab is buying short pieces of DNA to do something called PCR, polymerase chain reaction, which is a way we routinely amplify DNA since the late 1980s, just an in-vitro reaction. And so the cost of those little DNA pieces that we used for PCR. It’s pretty much the same thing, same size piece of DNA that’s used to build a CRISPR-ied sequence. We mentioned that the protein component of the CRISPR system is generic and it’s guided to its target by a little piece of nucleic acid. So to just to give you an actual number, I mean if you’re just like, man off the street calls up IDT, Integrated DNA Technologies, and you ask for one of these to be shipped overnight, $8, $5?

JORGENSEN: Well, it’s $40 bucks when you add the shipping.  

SANJANA: You have to negotiate with IDT is what I really recommend.

JORGENSEN: Oh, you probably get the institutional discount.

SANJANA: I think most-most universities try and get a discount.

BENEZRA: Wait so you have to grow the cells and there are other expenses. So it’s a little bit more…but it’s not on that scale.

MATTHEWS: Let me just jump too, because it’s completely depends on what you’re trying to modify right. So we’ve gotten to the point now where in the mosquito, the cost of the DNA synthesis and the reagents to actually do the CRISPR is negligible. You know it’s a drop in the bucket. We have to get these into the embryos of mosquitoes. So that involves microscopes and needles and we have to rear these animals, we have to feed them over time. We have to employ technicians to keep the insect facility running. And so at this point, I would say there’s nothing we could do to make CRISPR any cheaper that would bring down-that’s not the rate limiting step for us anymore. And that’s kind of amazing, because back when I started five years ago, the cost of a single zinc finger nucleus from Sigma was $25,000. So we’ve gone five years more than that. Exactly so yeah. So as somebody who just jumped into this game, I’ve already seen it come down orders of magnitude. And so unless you’re talking about genome wide scale. So if you want to take 20,000 genes, then it’s going to add up.

SANJANA: Well I was going to say the price is even cheaper because…

MATTHEWS: Per gene.

SANJANA: because we do oligo array synthesis. It’s a different scale of synthesis and there it becomes pennies basically. But as you said, the rate limiting step I don’t think is the DNA. It’s the other aspects of just doing- being in a biology lab. You need to grow the cells, grow the organisms

MATTHEWS: House of the mice.

SANJANA: House the mice.

MATTHEWS: …more for mice than you do for CRISPR now.

JORGENSEN: There is actually, and this is very controversial. There is an Indiegogo campaign right now that some guy who claims to be promoting citizen science is running. And I think it’s $160 and you get some plates of yeast. You get a CRISPR Cas9 plasmid that has a knock out- a specific guide RNA that knocks out a gene in the adenine synthesis pathway, an intermediate will build up and the yeast will turn pink. So you’ll know whether or not you’ve done it, and a couple of reagents to help you make the yeast take the DNA, and some plates and some pipets for 160 bucks. Easy-Bake crisper I mean.

[01:10:10] TSANG: The thing is, it is time. Time is the most expensive. So for CRISPR-on Monday you have designed a guide and Wednesday you’re ready to inject. But with TALEN, which even is the most-you have to develop to make it easier to use. It takes about a month from conception to the TALEN is ready to assemble. And then before for the zinc finger, if you design it, there’s like 80 percent failure rate. You pay Sigma $20,000. Then they say 90 percent success rate.

BENEZRA: Other questions. Yes. We have a microphone just if you can wait just a sec.

AUDIENCE: I wonder if anyone on the panel can comment on the applications in agriculture and livestock farming and marine development.

BENEZRA: That’s a good question. That was discussed heavily last night. You guys feel like tackling that or…not my area of…

JORGENSEN: Well the plant thing is-is really interesting. Did anyone see the new story about the mushroom that people or- the GMO mushroom that people are upset about because it doesn’t. It’s not GMO according to the FDA. So if you use CRISPR to knock a gene out, the CRISPR- there are ways that you can do it such that the CRISPR system doesn’t stay in the organism- it’s-it’s the organism loses it. Because if you don’t give it a selective advantage to keep it, it’ll just lose it. And so you make this break, the normal system comes in and tries to repair it. And now that repair job, if it’s botched, can knock out a gene if it’s that non homologous enjoining. And it’s really- you can’t tell whether that was a natural -natural phenomenon or genetically modified. And apparently, a lot of very valuable phenotypes in plants you can get to by knocking genes out. So now you have a whole array of technology that you could use that doesn’t really interact with the regulatory system right now. Because you’re not putting foreign DNA in. If you use a gene gun, then you’re not using plant pathogens and yeah. So-so-so actually the plant people are very, very excited about CRISPR.

BENEZRA: Very excited about getting much more plants per acre and feeding a hungry world. And that was discussed last night.

SHERKOW: Yes. So there’s also the vector that’s being used for some of these. So the thing that’s actually transformed in some of these cases is not the plant itself. It’s some of the bacteria that attached themselves around the root of the plant. So in the mushroom that was recently approved, or I think that’s definitely the wrong word, that was determined not to be under the regulatory authority of either the USDA or the FDA. The thing that was modified was this bacteria called agro fasiens right?

JORGENSEN: Oh really I didn’t know it was the bacteria. I thought it was the mushroom.

SHERKOW: No, no, no, no. So what they do is they modify the agro fasiens, this is work that’s being done at Penn State right now. And that attaches to the root of the shroom and that’s the thing where the DNA from the bacteria itself is uptake by the plant in that particular case. The reason why this is important is it is much, much, much easier to do gene editing in bacteria than even things like plants. So again, a lot of the agricultural scientists are just kind of over the moon at the possibilities here.

BENEZRA: So could somebody explain to me is there any real rational reason to worry about GMO? I mean honestly, is there anything that we should really be concerned about in terms of what could potentially happen?

SHERKOW: We have done a fantastic natural experiment. You know for the past almost 20 years, if you have eaten soy period since roughly 1997, you have eaten a GMO.

BENEZRA: In fact if you’ve eaten yogurt, you’ve also eaten bacteria have been defending themselves against viruses that would ruin the bacteria that are needed to make the yogurt for many, many years.

JORGENSEN: With CRISPR.

BENEZRA: With CRISPR, exactly. Question. Yeah?

AUDIENCE: Just back to the ethics again. You know as a layperson, I hear folks like you and lots of people talk about the dangers, the possible dangers in the germline research and the off-target effects and so on. But I can’t help thinking about communities like the Huntington’s community, Tay-Sachs CF, BRAC gene. What do you say to those people and how do you resist their pleas for help when something is out there that could knock these diseases out permanently as I understand it?

BENEZRA: Absolutely. A very, very good question and. It’s an ethical question that we all have to struggle with. And you know, physicians take an oath to do no harm. Yes. They want to do good. But at the same time they can’t- you cannot take the risk that there will be some worse debilitating disease downstream-now, we might be able to get to the point. It’s not hopeless. We might be able to get to the point where the benefits will absolutely outweigh the risks. I personally just don’t think we’re there yet, quite frankly. And I understand the passion and I understand the plea and I think it’s an important, very important ethical issue, but one that both scientists and laypeople have to struggle with for just that reason. You see what I mean.

[01:15:36] SANJANA: But we shall we should say that for most of the diseases you mentioned I think there are NIH funded and other scientists all over the world that are using a variety of tools, including genome at editing to one, to create models of those diseases that they can easily handle in the lab. And then two, to- they actually are working on- I mean this is a product of all your tax dollars, funding the NIH which is a very tiny portion of the budget. Could be bigger. But it’s you know this is the great stuff that’s happening. So it’s not like I don’t think there’s even though maybe I don’t know if any of us work on any of the diseases you mentioned, but there are a lot of people who do…

BENEZRA: The study, but not necessarily to introduce into the human germline at this point.

SANJANA: Maybe not to introduce in the human germline, but certainly to think about therapeutic strategies, like you talked about a lot of hematopoietic diseases. Anything in the blood system is quite easy, because our blood system really renews itself quite quickly using the stem cells that are in our bone marrow. And so you can think of X- you know not editing anything inside people, but taking out some of these bone marrow cells, maybe correcting some mutation, and then offering some something where like sickle cell anemia. A horribly painful disease where there’s basically no treatment. We use the same treatment that’s been- that was first prescribed in like the 60s or 70s and it’s not, for these people, it’s not and.

BENEZRA: But again, the question was germline therapy as opposed to somatic cell therapy.

BENEZRA: That’s a very clear line that has to be drawn. And it’s an incredibly important question that many people are struggling with and we need to struggle with and we have to have these types of discussions to try to answer that question rationally and reasonably.

TSANG: But that ethical dynamic won’t go away with a highly effective somatic gene therapy treatment for Huntington’s, right?

BENEZRA: Yeah. Well obviously if you’re isolated to a particular cell type, yes, but if it’s more disseminated, then it becomes much more difficult.

TSANG: But there’s a good somatic gene therapy delivery to the brain. Right. Then you don’t need to worry about the germline editing.

BENEZRA: Yes. Another question.

AUDIENCE: Autism and using IPSCs and I just wondered if you can say a little bit more about sort of where CRISPR combined with IPSCs is actually going to, say in neuroscience in the next 5 or 10 years, potentially affect the ability for us to create new treatments. Because as you know for autism and pretty much across a lot of neuroscience in psychiatry, there haven’t really been any fundamental improvements in you know many, many decades. And it’s sort of shocking and in a way when you think about it. So I just wondered if there are some with some of these complex disorders, some specifics of where you see the combination of CRISPR and IPSCs actually changing you know screening and development of therapeutics.

BENEZRA: That’s a great question. Neville, you want to…

SANJANA: Sure I can start.  First, I completely agree with you that especially like kind of in neuroscience diseases, that it’s been kind of abysmal, our progress. And I think a lot of drug companies, though now things are changing, but have exited kind of neuroscience therapies over the last 10, 15 years, because the progress has been so poor and so I think the-the very- the problem with, as I see, it is a very personal view. The reason why in cancer we’ve had I think quite nice progress in the last few years, and the pipeline looks pretty good in cancer actually for the next several years, is that the disease phenotypes are so clear in cancer does it does the cell die does it survive? Does it travel doesn’t metastasize or does it stay put? Does the tumor get bigger? Is the tumor resistant to the drug?

SANJANA: These are like life and death kind of things. They’re very easy like presence or absence kind of tests. But in neuroscience, like the phenotypes tend to be very, very subtle like you know what is the how. You know we don’t even know really what are the right cellular correlates of a lot of these social or psychiatric deficits that we observe in the- in the patients. And so I think getting the right understanding- because we think that a lot of these behaviors-this is the major principle of neuroscience is these behaviors we observe and the people are caused by things happening in here at the cellular and molecular level. And so the question then becomes, how do we create great models of what’s hap how do we understand what’s happening at the cellular and molecular level? And I think that is where CRISPR, I mean we’re just in the early days here like we’re talking about just a couple of years after the scientific community started to decide…

[01:20:02] JORGENSEN: 2012. I mean three and a half years.

SANJANA: This is like the Stone Age of CRISPR, I guess is the right way to put it. But you know imagine once the neuroscientist have a little runway to really move with these-I think the clearest answer to your question is that people will be able to create accurate models that they can then apply the whole wide array of molecular biology and cellular biology tools and assets that we’ve developed you know over the last hundred years or something but they just haven’t been able. They don’t have the relevant maybe cell type with the relevant mutations or what they think is the putative causal variant in there. And now we can just try out. We can we can easily create these. Whether it’s in a mouse or in a human cell in a dish very, very quickly. And so I think that is really going to fuel things like genome wide screens small molecules screens for drugs things like that.

BENEZRA: To put this in a time scale, 65 years ago, we didn’t not know what the genetic material was. We didn’t. We did not know what DNA was. I mean, the speed with which this is happening is really quite remarkable. Maybe one or two last questions. We have another minute or so.

TSANG: This is with Dr. Tsang. A question about RP, retinitis pigmentosa, does CRISPR change how you would treat the disease or be able to cure the disease? So in so we-in the clinic, this depends on what is available by FDA. Right? So. So in 2000- in 2016, we still-there’s no proven effective treatment for RP in 2016. In 2017, one of the CRISPR pharmaceutical said they would treat one form of the RP. RP is heterogeneous. there are at least 60 genes that can cause RP. And then there is probably another 60 genes still being identified. So in the long run, you want it to be-there’s nothing more better than precision medicine to treat the patient’s own specific mutation and then say you are applying for CRISPR to deliver in the eye. But fortunately, most of the gene therapy companies pharmaceutical and CRISPR pharmaceutical, they pick the eye as the target. This is the most-highest level of precision medicine. You need to know the mutation of that patient and then put the gene is repaired by CRISPR. But also just to put it together and emphasize what Neville say also, they also bring together the same as in autism. There’s no disease that we cannot treat in medicine if we understand the mechanisms. So you can also envision that CRISPR identify a pathway, a target. Then you may not need to know the specific gene. For example, CRISPR is being applied to the vascular and grow factor for a form of macular degeneration. So instead of people getting antibody injection, $2000 a month, some patients got a hundred injections already, and maybe you can do CRISPR and do it once. Because the pathway for macular-for macular degeneration is quite understood as supposed to for retinitis pigmentosa.

BENEZRA: Okay, they’re signaling our time is up. I can tell from the questions we did our job well. Thank you all for coming and thanks to the panel.