share interactive transcript request transcript/captions live captions
|
MyPlaylist
ANNOUNCER: This is the production of Cornell University.
BUD JERMY: Good evening. And welcome to the last summer session lecture, regular lecture. Tomorrow night, we have a bonus. Joyce Carol Oates will be here and at the same time. And I would invite you back for that.
Please silence all electronic devices. And I want to also thank Kathryn Boor, the Dean of the College of Agriculture and Life Sciences, for the use of the hall. She's been very generous with us this year. My name is Bud Jermy, and I'm from the School of Continuing Education and Summer Sessions. And we're glad to have you here.
Ailong Ke is a professor in Cornell's Department of Molecular Biology and Genetics. He is a member of the graduate field of biochemistry, molecular and cell biology, the graduate field of physics, and the graduate field of chemistry and chemical biology. His research focuses on CRISPR interference and RNA-guided defense mechanism in bacteria and prokaryotic microorganisms. And his work has been published in Nature, Science, and Cell.
In 2018, he was the recipient of the RNA Society of Mid-Career Award, and also the inaugural Provost Research Innovation Award in the Life Sciences at Cornell. Ailong has an undergraduate degree in biology from the University of Science and Technology of China and the PhD in biophysics from the Johns Hopkins University School of Medicine.
Before coming to Cornell, he was a postdoctoral researcher with Jennifer Doudna at the University of California, Berkeley. Ailong is a leader in developing a new type of gene-editing CRISPR system that he and his colleagues have used in human cells for the first time. This method is a major advance in the field, and that is what he will be telling us about tonight. CRISPR gene editing moves out of the laboratory and into human testing.
[APPLAUSE]
AILONG KE: Thank you, Bud. So I was told this is going to be a diverse set of audience. And so I really broadened my talk. And so my own research is going to be a fraction of it.
And so I want to touch base more on basically putting CRISPR into the general picture and giving you an overview of where the industry is and ethical issues related to editing. Maybe some of you are more interested in hearing that. And so without further ado, I guess, I'll move on to the intro.
And so this is my attempt to make it to the general audience. I might have gone too far. So if I did that, excuse me. So I'm a sci-fi fan. So I started-- I want to start with this movie by Steven Spielberg, basically a adaptation of Joyce Wells' novel The War of the Worlds.
So, the storyline was that the aliens in one day basically started their pre-meditated invasion to Earth. They were beamed into those tripods, which were hidden on Earth. And then they fired these machines up. And this is when they make that loud horn. And within 10 seconds, they start to fire their primary weapon and start to basically zap humans to vaporize humans.
And it was a pretty desperate situation. They were grazing on Earth, picking up human beings as candies and converting us to fertilizers. And no human weapons were effective against them. So it was really a hell kind of a situation.
And so what I liked the most was the abrupt ending at the end when just all in a sudden, they dropped dead. And so they seemed to have died from diarrhea or other kind of diseases. And so that's the interesting part. And so they said that from the moment they invade, these invaders arrive, start to breathe our air, eat and drink, they're doomed. So destroyed by the tiny creatures on Earth, which as a biologist, I assume were the microbes.
And so what I'm trying to allude to is why do we earn our right to survive? So I want to basically point out that word. We have immunity. So over the billions of years at the expense of billions of lives, we developed our immunity and our right basically to survive among this planet's infinite organisms.
So here, "we" in the movie refers to human beings. And so basically all vertebrates share the same type of immunity. So this refers to a subset of white blood cells, the B cells and T cells. And they have an island in their genome that seems to evolve really fast, give rise to all kinds of diversity. And that gives rise to different kinds of proteins.
So in the B cells, these are the antibodies displayed or secreted into the environment. And then for B cells, these are the receptors on the surface. And the B cells were able to use antibodies to neutralize the pathogens, and the bacterias, the viruses. And the T cells were able to basically kill all the barren cells in our body.
And we really benefited from these cells, and not only to await the war against the aliens. But a rough understanding of the system already give us the technology, the vaccination. So we start to gain memories against pathogens that we have never seen before. And the hardest therapy in town against cancer is this immunotherapy.
So essentially what I want to say is that immunity system is really, really important for multicellular organisms. And it turns out that "we" not only refers to these kind of organisms. In microbes-- so these microbes, the prokaryotes, also rely on adaptive immunity system to survive.
This is because they also face a very tough environment. So if you just take a drop of water from sea or from Beebe Lake and you stain for nuclear acids, what you will see is that basically DNA is from eukaryotes, bacterias. And then in the background, you see those tiny dots and they come from viruses against bacteria.
So on average, each given bacteria is surrounded by viruses in a ratio of 10 to 1. And so it's a very tough environment for the bacteria to survive. And so here's a picture of the bacteria surrounded by a swarm of bacteriophages waiting to gain access into the resources inside and basically multiply into more viral particles. And so here are two bacterial cells kissing each other. A simple exchange of genetic material could lead to a transmission of a mobile genetic element or a prophage that cause death in the recipient cells.
And so over the years, that bacteria also evolved a adaptive immunity system. And this turns out to be not a protein-based immunity system, but a nuclear acid-based immunity system. So that's the thing. A long twist, but this is where I'm trying to get at. This involves a CRISPR locus constant repeats in their space by variable spacers. And there's a close CRISPR-associated operon involving proteins that contain activities, nucleuses, RNA-binding proteins, et cetera, et cetera.
So when I started my independent career here at Cornell, that was 2005. And the name CRISPR was never mentioned in-- well, at least I didn't hear about it in the literature. And it was one day my colleague Matt DeLisa knocked on the door in 2008.
So he was trying to engineer a metabolic pathway. And his genes kept being silenced by some mysterious bacteria systems. And he was diligent enough to have done a genetic screening and stumbled onto this CRISPR-Cas systems. And he did some literature search and found some very interesting papers. And so he knocked on my door and said, maybe we should collaborate and study those systems because you are an RNA expert.
So, the paper he was referring to was this paper by Eugene Koonin So he's an informaticist. And so based on various primitive observations that the variable spacers in the CRISPR array here seem to match the sequences in bacteriophages, the viruses, and mobile genetic elements against phages, and these CRISPR associate operons contain all sorts of genes working on nuclear acids, and so just based on these primitive observations, he came up with the hypothesis that this entire system is an RNA-based immunity system in prokaryotes against basically foreign genetic acids-- foreign nuclear acids.
And the basic principle he hypothesized is similar to the eukaryotic RNA interference system. So that was really, really interesting. And so if he didn't point me to this paper, I would have never read that journal Biology Direct. So that was really serendipitous for me.
And so at a time when we did literature search, there were maybe 36 papers in the entire Pubnet. So it was really wild west in this subject. And so there was just one paper published in 2007 providing experimental evidence that CRISPR indeed is an adaptive immunity system that provides protection for prokaryotes against viruses. So that was reassuring.
And then around half a year later, there was a paper by Erik Sontheimer and Marraffini, saying that interestingly, the CRISPR were targeting double-stranded DNA rather than single-stranded RNA. So that was unexpected because we were looking at RNA interference, which target single-stranded RNA.
And it turns out that implementations already reported three different times of CRISPRs. So it appears that the CRISPR systems are quite diverse. So they may have evolved from different origins.
And so many here probably knew the name Cas9 here. At the time, it was not called Cas9. It was something called csm1. And it was one of these CRISPRs. So, it's a very diverse system. And somehow it was able to acquire new DNAs into the array and acquire new memories.
And based on the analogy to the RNA interference, we were expecting that some of the proteins here in the Cas operon will form a factor complex with the guide RNA. And together, they should be able to find a target and somehow destroy it.
The burning question at the time was that, so if it were RNAI, we would have understood it from day one that basically it involves [? wasper ?] pairing between the guide and the target. But our target here is a double-stranded DNA. How does the RNA find a matching sequence that's buried in a duplex DNA? So that was very puzzling at the time.
So the other puzzling question was that basically if you look at the guide RNA, so it has a perfect match in the foreign DNA, but also a perfect match in the CRISPR array. So how does it distinguish what's foreign DNA and what self DNA and prevent auto-immunity from killing-- cleaning the self DNA and cause basically death?
So the same question was answered first. And so but informaticists basically said that the foreign target is always flanked by something called protospacer adjacent motif. The short sequence identifies-- basically earmarks the target. And the equivalent position in the CRISPR array essentially disallow target searching in that array. So then this is self, and that's foreign.
So the second question was solved by biochemists. Essentially they demonstrated that in order to find the target, the effector complex has to unwind double strand DNA and basically promote base pairing between the guide RNA and the target strand DNA. And this effectively loops out the non-target strand DNA. So, we're forming something called an R-loop intermediate. And so that's an essential process in every CRISPR system that targets DNA.
And it just turns out that some time later, people started to realize that the nucleuses in the CRISPR systems almost exclusively target single-stranded DNA. So essentially you have to open the R-loop, expose DNA in the single-strand formed, then cleave them. And this is a quality control mechanism to prevent aberrant targeting-- a premature targeting of a near target.
And things become really interesting in 2012 when Jennifer Doudna and Emmanuel Charpentier published this paper in Science essentially reported the biochemical reconstitution of the CRISPR/Cas9-based DNA targeting in test tube. And then demonstrated that you can change the guide and program the Cas9 to target a different guide. And essentially you have a programmable nucleus that you can use to target different regions. And that formed the basis for-- opens the doors to do gene editing.
And that was a tough biochemistry experiment. And they have to identify that there are two different non-coding RNAs that's involved. And it has to target PAM. And there are two single-strand nucleuses. And when everything comes together, you have DNA cleavage. And then they were creative enough to fuse the two guided RNAs and give you a single guide RNA that can be efficiently used in gene-targeting experiments.
And so synthetic biologists at Boston quickly followed up. So John and George Church came up with back-to-back papers only half a year later and basically reporting the usage of CRISPR/Cas9 in human cells. And did they demonstrated that they all come from CRISPR/Cas9-guided targeting human cells frequently result in indel formation, so-called insertion and deletion-- small insertion/deletions as the result of targeting events.
And they came up with many concepts that are still being discussed in the literature. They basically formulate the delivery strategy of how to multiply this system, target multiple sites, how to use just a single nucleus activity rather than double nucleus-- double-stranded nucleus activity, use two guides to generate a deletion rather than a double-strand break, and the repair pathways involved in repair the targeting outcomes, and the concept of the off-target events. And so these are really a brilliant set of biologists that really shaped the field.
And from these important resource, we essentially-- this is the beginning of an era of efficient genome editing. So this is not the first genome-editing tool. So before that, we have zinc fingers. And our colleague at Cornell, Adam Bogdanov-- I hope I pronounced this right-- so basically pioneered the usage of TALEN in genome editing here. So there are two tools preceding CRISPR.
So there are certain traits in CRISPR that just make it really, really popular among the researchers. So this includes the fact that it's really, really easy to use. An average biology lab can just pick it up. And nowadays, undergraduates can be trained to use it fairly easily.
And this CRISPR/Cas9 tool is super efficient. And especially the S. pyogenes Cas9 that's reported in Jennifer's paper, this remains the most effective Cas9 today. So essentially, they hit the jackpot from day one.
And this tool is very versatile. And you can adapt to different systems and get different outcomes. And it really allows the researcher to break the barrier of having to work on model organisms where genetic tools are available. And now they can chase interesting biological questions and go to exotic model organisms to do genetics and understand the outcomes.
And there are some limitations of this tool. And that becomes obvious very quickly as well. And so the Cas9 in particular is favoring efficiency over accuracy. So it's prone to off-targeting effect. And when it comes to therapeutics, this is a concern. So you don't want to edit a size that's not intended to be edited.
And in order to achieve editing, as I mentioned, you have to have a PAM sequence last earmarking the target site. And so the PAM sequence restriction here really limits the targetable sequence space. So we had-- if you want to repair some place, it has to be flanked by a PAM.
And there are also some other limitations. So it's hard to predict which site can be efficiently edited, whereas others are not so efficiently edited. It seems to be-- there's some randomness there. And the outcome of the editing is hard to predict. And I'll explain that in details.
I won't go into details about all the mechanistic work. But there was a huge line of-- there is huge interest from the community. A lot of labs contributed to important studies to understand the structure, the function, the mechanisms, and how to make good use of the system, so high resolution structures, single-molecule studies of this dynamic behavior, and directly evolution to make it more useful, et cetera, et cetera.
So there are interesting studies to understand the PAM recognition, which is not depicted here. But based on that structure study, they were able to come up with rationally-designed new PAM codes for Cas9 to some extent. And they were able to come up with creative basically ideas or rational designs to generate high fidelity versions of Cas9 that's more accurate, and there's less off-target effects. And they were able to come up with new activities from that. And so I won't go into the details.
So I want to reiterate that the default activity of CRISPR/Cas9 is to cleave and destroy DNA. And that's the outcome of the RNA-based immunity system. And in Cas9 system, this is the formation of a double-strand break here.
So the final outcome is the outcome of the repair by the host cells. So in human cells, the default repair pathway to deal with a double-strand break in most cases is a process called non-homologous end joining process. And so basically a double-strand break is a crisis situation in the cells. And the cells try to patch it up by polishing the ends to a blunt end and then quickly ligate them together.
So this process will trigger a lot of error. And so if the targeted site is in the protein coding region, these errors will result in auto-frame insertions and deletions and loss of function in the targeted protein. And so in many cases, the uses of CRISPR/Cas9 is to cause out-of-frame mutations in the targeted genes.
So if you want to achieve very precise editing, you want to provide a template and promote the cells to use the homology-directory repair pathway. So essentially this requires resection of the DNA, and invading into the template DNA, and copying off the information from that. And so that's not the default repair pathway in the cells. You really have to do a lot of work to essentially coax the cells to use this pathway. And so usually it's not very efficient.
So if you combine all these things together, the take-home message is that essentially, in most cases, if you do CRISPR/Cas9-based editing, the outcome is not so predictable. But it's usually a disruption, a loss of function editing outcome.
And another drawback in using this tool with the idea of to do therapeutics is the problem of the off-target effect. So I mentioned that Cas9 is not inherently the most accurate at editing enzymes. So it favors efficiency over accuracy.
So there are some other enzymes that are inherently more accurate, but the trade-off is that they're not as efficient. And so the off-target effect basically refers to the modifications of unwanted sites. Usually these are highly homologous sequences that are a few base pairs different from the intended target.
So, if you have the luxury to target not just one site but a stretch of sites, then you can use informatics to choose the site that's quite unique and avoid the sites that could trigger an off-target effect. And so there are some creative ways to engineer high fidelity Cas9s that solve the problem partially.
And it became obvious that the longer you expose human DNA to Cas9, the more likely you will get off-target effect. So really reducing the contact time is an important strategy to avoid off-targeting effect. So you don't want to deliver Cas9 permanently. You want to have a transient delivery effect. And if you can avoid delivering DNA, then delivering mRNA is better. Or better yet, if you can deliver RNA protein complexes, which will only last a few hours in the cells, then the off-target effect become dramatically reduced.
So there are many creative approaches being created to circumvent this off-target effect. And yet one approach involves introducing a shut off switch, the so-called Anti-CRISPRs that's naturally occurring and was used by the bacteriophages to inactivate the CRISPR systems. And so here the researchers introduced the Anti-CRISPR and basically shut off the Cas9 after a certain exposure to the human cells, so a lot of creative research.
So I want to switch to therapeutics, I guess, with the idea that there are limitations to the CRISPR/Cas9 tool. It's a very, very powerful tool. But we understand that it has off-target possibility. And the editing outcome is kind of unpredictable in many cases.
So then with these understandings, there are ethical concerns when you use them in human patients. So, it's one thing to use it on somatic cells without the possibility of passing the edited cells' genetic information to the next generation, so-called somatic editing. And so you can do it either ex vivo or in vivo.
Ex vivo is even safer. It involves taking the cells away from the patient, do the editing experiment, and then putting it back to the patient. So the in vivo editing really involves really injecting the editors into the bloodstream or localized areas and try to achieve editing outcome.
This stands in contrast to so-called the germline editing, so where you're really editing an embryo or the gametes. So here, this is really, really dangerous because we're really passing the genetic information to the next generation. If we cannot control the editing outcome, we're basically passing whatever the failed experiments to the next generation. And from the patient's benefit from the species as a whole, human being species, this is something we don't want to do if we cannot do it safely.
And the ethical concerns varies depending on the situation. So if is a life-threatening disease, then it's easier to gain ethical approval to carry out a genome editing experiment. So, for example, here, this is a picture of Adrian Krainer, a biologist who invented an anti--sense oligo therapeutic strategy against, I think it's spinal muscular dystrophy. So these patients will succumb to the disease at a very young age.
But he was able to rescue that disease phenotype. And these patients were able to walk rather than being paralyzed in wheelchairs. And so this researcher's really, really happy about the outcome.
So in those kinds of situations, it's understandable. Even if the editing outcome is not a perfect fix, it's a workaround. But it was able to-- it was able to alleviate a really detrimental situation.
So this stands in contrast to the situations where the editing was mostly geared towards some kind of enhancement, so as a cosmetic editing where people were trying to gain better attributes, physical attributes, or something from the editing experiment. And so these-- it's hard to gain ethical approval in today's atmosphere.
So essentially, if it's a treatment strategy against a serious disease, and it only involves somatic editing, the current ethical background is that, yes, please proceed with the therapeutic approach, but with caution. But it is an enhancement type of editing, the common understanding is that you should not have proceeded it in any way. And germline editing should not been proceeded because of the safety concerns. So that's the common understanding. But we'll discuss some violations and situations.
So in the next few slides, I'm going to basically give you an overview of where the industry is, kind of a cutaway of the forefront of the CRISPR industry. And it sounds like I'm doing advertisement for this industry, but I'm learning it myself. And so I'm stealing slides from the company website. So I think these are the appropriate use of the CRISPR technology.
So it appears that the early pioneers, each one basically set up their own companies. And so George Church is this guy at Harvard who has maybe hundreds of different companies. And so one of them is called eGenesis. And this involves doing genome editing on pigs with the ultimate goal of maybe using these pigs as organ donors for human patients.
And so it's well understood that organs are hard to come by. And a lot of patients die in waiting for the organ transplantation. And it's well known that pigs has a physiology that's very similar to human. Also they're a great target.
But they-- you can't just take the pig organ and put it into human. And they will be rejected very soon. So there are a lot of surface markers that has to be removed in order to have the pig organ recognized as a self organ rather than a foreign object. So that involved using CRISPR tools to knock out these kind of markers.
And pigs have some viruses that are potentially harmful for humans, like retroviruses. And so they're basically diligently removing these viruses one by one from the pig genome. And so initially it was claimed as that they have a very ambitious goal. They were trying to reach the human trials within two years. That was 2016 or something like that. I haven't heard anything from them on human trials. So I think they're still doing animal testing maybe first in big animals, primates, or something before it's considered safe enough to try on humans.
So another company is set by Emmanuel Charpentier who co-published that seminal paper with Jennifer Doudna. And so her company is called CRISPR Therapeutics. And they seem to be the first one who came out-- roll out the human clinical trials.
And so first CRISPR clinical trial, and this involves ex vivo somatic genome editing using CRISPR/Cas9. And it's a very popular target. And so many companies are targeting this one, sickle cell anemia. So if you studied biochemistry, it's a textbook example of a molecular disease.
So the phenotype is that the patient's cells have this characteristic sickle cell shape-- sickle shape. And this is because of a single nucleotide mutation in the hemoglobin gene that changed the identity of a single nuclear acid-- a single amino acid, which causes essentially an aggregation of hemoglobin in the patient. And this is the first molecular disease being deciphered by researchers, so by Linus Pauling and others.
So a perfect fix would be to convert this mutated nuclear acid back to the wild-type sequence. But because I mentioned that Cas9 is good at destroying things rather than fix things precisely, this is hard to achieve in a precise fashion. So most of the companies try to find a workaround.
So this company, the workaround was based on this observation. So, each human being has two different versions of hemoglobin genes. So we have the adult version and the fetal version.
So at birth, there is a transcription switching going on. The adult version is switched on, and the fetal version of the gene is switched off. So in some individuals, this gene persisted and kept expressing in adult individuals. And there's only-- there's essentially not much detrimental phenotype in those individuals. So that was the therapeutic strategy used by CRISPR Therapeutics.
And so their reasoning is that rather than targeting the diseased allele and trying to achieve a perfect fix or a gain of function rescue, we could target this repressed allele and try to essentially disrupt the genetic-- disrupt the regulatory circuits that shut off the allele and allow the gene to express in a constitutive fashion.
So what they did basically was based on this paper, they disrupted the transcription factor binding site and allowed this gene, the fetal version, to express constitutively. And that was-- and so they were able to rescue the disease phenotype at the cell level. And so they're now testing it in humans.
And this is not a naturally-occurring mutation in humans cells. And you don't want it to pass-- to be passed to the next generation. So it makes sense that they're doing somatic genome editing.
And to be more prudent, they're doing it ex vivo. So they're extracting the bone marrow from the patient, and they're going to do the genetic experiment in test tubes, amplify the bone marrow, and then for the patient, they're going to basically eradicate all the bone marrow cells in the patient and then infuse the edited cells back to the individual, and basically undergoing a bone marrow transplantation but using the individual's own cells that have been modified. So everything seems to make sense. And so hopefully, they were able to achieve a cure for these patients.
So here is another clinical trial that has just been rolled out two days ago. And this is another pioneer from John's Lab. He has this Editas company that attracted a lot of attention. And they have a long list of therapeutic targets.
Interestingly, the first one was a very-- it's a rare genetic disease that has a pretty small patient base. But I guess they targeted this disease and put out the clinical trials because it's considered a safer therapeutic strategy. So they're basically doing localized in vivo human cell editing. And they're doing it in human eyes and targeting this congenital eye disease.
So what I learned from their website was that basically this is a mutation in a structural protein inside the eyes. And this protein will cause the collapse of the photoreceptor cone. So essentially, if not rescued-- if not fixed, this mutation will cause blindness in the patient within a few years after birth.
And their strategy was to deliver an adenovirus containing Cas9-- encoding Cas9 and guide RNAs locally into the patient's eyes. And the hope was that this editing event will rescue the function of the deceased gene. And we're going to get the normal photoreceptor cone and the normal function of the eye. And we're going to get improvement from the patients.
So if we really look at the-- while we have a basis for this therapy, it's again a workaround from achieving a perfect fix. So here is the mutation. And this is a mutation existing in the introns of a gene.
So somehow, this mutation generated a splice site and basically causes apparent splicing event that introduced a exon in the patient. And this exon contains a premature stop codon that essentially caused the gene to stop here rather than splice and translate into the full-length proteins. And that causes the disease.
And so the rescue was to not directly fixing the mutation, but use Cas9 programmed with a pair of guides targeting sequences around the mutation site, and try to generate a double cut here and trigger either a deletion event or an inversion event. In either cases this is going to basically prevent this mutation as being recognized as an exon. This is going to prevent the inclusion of a bearing exon here. And that will cause the normal splicing and a normal function of a gene.
So again, this is not a mutation naturally occurring in the human population. And that's why they choose to do somatic genome editing. But this is the only allowed editing anyway.
But it makes sense. And it's a localized injection, so we're not exposing patients to unwanted editing events in other places in the body. So everything seems to be rational in design and makes a lot of sense. So there are a lot of considerations into the ethics.
But there are also cases where it was not well thought through, and there was the poor way of executing science. And essentially because the tool is so easy to use and is so potent, when it's in the wrong hands, it really causes unwanted outcomes.
So we've heard in the news that people start to mail order CRISPR kits. And we've heard of people claiming that they're going to do garage genome editing and so essentially, make them more muscular. And that's what-- I saw a video online. They're like, I'm going to edit my muscles and I'm going to be-- and I don't have to work out as much.
But nothing was more perturbing when at the end of last year when news came. So there was a real editing experiment in embryo that causes the-- done by He Jiankui, a scientist in China. Essentially, this led to the actual birth of a twin CRISPR baby, so something that people have feared, and the scientific community has been self-policing to avoid doing. And someone just pushed the envelope and did it all in once.
So that was very perturbing. It caused an outcry in the scientific community, and including the scientists in China because they don't know this guy. This guy was not a CRISPR expert. And so he just came out of blue and did the experiment and thought that he's doing something heroic.
So let's review the guidelines for doing those genome editing experiments in germ lines. So there are two versions of regulation. And one is a 2015 statement from the first human genome editing consortium essentially.
And it's hard for me to read. Basically, recognize the limitations of the technology and said that they should never have been-- it would not-- it would be irresponsible to proceed with any clinical use of germline editing. But it relies on the scientists to self-police and it left on back doors. So it says that we're going to revisit the issue and see whether it is more acceptable in the audience-- in the general public, and whether the technology becomes safer.
And there is another statement in 2017 by the US National Academy of Scientists. And so basically this is a panel discussion involving experts in the US and representatives from all over the world, like major countries in China, UK, Russia, and others. So they came out with a guideline for genome editing experiments.
And so essentially-- whoops-- if you read the guidelines and try to follow these to design the germline editing experiment, none of the experiments would have passed those guidelines. It's very strict. And again, it relies on scientists to follow the guidelines.
So the editing experiments done by He Jiankui was on this CCR5 gene. So this is a gene that's basically used by-- there's a surface receptor gene. And the HIV targets this gene as one of its co-receptors to gain cell entry. So interestingly, there was a naturally-occurring allele. And especially the allele frequency is higher in European population.
And so this mutation, CCR5 delta 32 is an out-of-frame mutation that will cause a premature stop codon and essentially a truncated protein being expressed. And this protein can no longer interact with HIV surface markers. And HIV, it can now gain entry into the cells if the individual is homozygous for CCR4 delta 32.
So if you have one wild-type allele and one mutant, they can still use the wild-type allele. A virus can still use that to enter. Double mutant seems to be immune against HIV. And there was-- this was discovered in a serendipitous situation when an HIV patient received a bone marrow transplant from a donor that was essentially a homozygous-- carries a homozygous mutation for CCR4 delta 32. And it was found that not only did it cure his leukemia, but also cured the HIV.
So one can imagine that in certain disease situations, it is a good target maybe for genome editing experiments. But nowadays there are so many ways to prevent the transmission of HIV. So if one uses the genome editing of CCR5 as a preventive treatment, then the ethical background for that is very shaky because there are so many alternatives. And it's considered not necessary.
So this news leaked out two days before the second human genome editing meeting. And because of the news attention, it was-- his talk was broadcasted online. And I was among the 1 million viewers around the globe basically watching and hearing what he said. So these are the reflections I remember at the time. And it gave me the impression that he really did it.
And there was a situation where-- the scenario was that the dad was the HIV positive, and the mom was negative. And so in such situations when actually the transmission of HIV can be prevented in many ways, including washing the sperm before the inception and fertilization, that would be usually sufficient to prevent the viral infection. And so then it really begs the question, where was the unmet medical need? So there is just no need to do this genome editing experiment.
And then the next question is, how did it achieve the clinical-- the ethics evaluation? How did it pass that? How did it-- how was he able to achieve the approval?
And I mentioned that in most cases, the editing outcome is messy and unpredictable. And in this case, it really showcased that this is the case. One should not have used it very lightly.
And so it was a twin girl being born. And one of the individual carries mutations in both alleles of the gene. It kind of mimics the CCR5 delta 32 mutation, but not the exact mutation. So it's not the exact delta 32 fix. But it mimics the premature truncation situation.
The other individual has one wild-type allele. And the other allele carries just a short deletion. And it didn't fully knock out this area.
And so this particular unnatural variant probably would not have conferred HIV resistance. And with the wild-type allele, it definitely will not provide resistance. So at least for this individual, it's a failed experiment. And for this one, there's unpredictable fitness consequences.
So in both cases, I would say it's not-- it's a very shaky experimental outcome. And it really begs the question-- the further questions, so these two individuals never signed any consent forms. And they were brought into the world without any consultation. So how do we protect their own interests? And how do we prevent similar reckless attempts?
So there are a lot of discussions going on in the community, including talks about moratoriums, I guess. But they're opinions from both sides. And so this is being heatedly debated right now in the scientific community.
And so that study still has rippling effects after half a year. And so this is a new study coming from Nation Medicine just last month. And so essentially talking about the fitness consequences of having the CCR5 delta 32 indentation.
And so as a statistical analysis, basically the observation was that individuals with homozygous mutation tends to be under-represented in senior population. So the interpretation is that maybe there's a fitness consequence. Their life expectancy is not as long as a wild-type individual. So there might be a fitness cost for being HIV-resistant.
And so there was an early study published in Cell claiming that there might be some beneficial effect from having this mutation. So the observation was that when there's brain damage, for example, a stroke, in a person, a human being with the CCR5 mutation is more resilient and seems to recover with a better outcome. So it seems like there is some gain of function phenotype.
Then it becomes-- so you can interpret these scenarios in different-- with different sentiments. So, for example, if we use genome editing really liberally, and so we inadvertently generated some gain of function phenotype, some genome editing outcome was a super human kind of a phenotype, then the average person becomes-- has to undergo this debate. Are we under pressure to also adapt? So there are some ethical concerns there as well. So it just underlines the importance that we really need to understand the outcome of the editing experiment before we do anything seriously with clinical research.
So I don't know a how-- what was the-- how many?
MAN: You're OK. Go ahead.
AILONG KE: OK, good. So I want to spend maybe 10 minutes talking about my own research.
[CROWD CHUCKLES]
So the question is, there have been 10 years of CRISPR. Is it-- it has been at the forefront of a revolution in biology. And so the question is, is the fever over? Certainly from my perspective, working on it day by day, I feel that we're still in the wave. So every time we think that we're over the peak, there are new and very interesting biology coming out.
So, I mentioned that there are a lot of diversity in CRISPR systems. And in the early days, there are three different types. And so by now, there's six different types. And so we can really grouped them into-- the CRISPR system into two major classes.
And the star molecules in the genome editing fields, the Cas9s, the Cas12s, or Cas13s, they really belong to one class of CRISPR systems. And these are the single component CRISPR systems. And so they're easy-- they're very simple, one-effector protein, in many cases, one guide RNA. So it's easy to adapt them for human genome editing.
So there are yet another class of CRISPR systems that involve multi-component effectors, so more sophisticated CRISPRs that are hard to adapt but carry-- nonetheless carry very important-- very interesting attributes. So the major differences are, I would argue, the nucleuses.
So they basically use the same principle to look for targets. But the nucleus have different functions. So in Cas9-- in single-component systems, usually the nucleus stop at making a double-strand break. But in systems like type 1, for example, the nucleus is almost like a terminator. It destroys everything after targeting.
And so by now, we understood-- so the first step of CRISPR interference involves cutting a piece of DNA and inserting into the CRISPR array. 10 years ago, it was somehow. And nowadays we have a good understanding of the molecular mechanism. My lab was one of the labs that contributed to key insights into its molecular mechanism.
And in type I system, the effector complex is assembled from five different proteins. And this really is a beast, like 400 kb, large complex, multi-component systems, recognized PAM, opens the R-loop, similar principles. But what's different is what happens after the R-loop formation.
So in type I system, the actual nucleus is a-- it's a very sophisticated enzyme that has two enzymatic activities. Not only does it have the nucleus, but they also have a helicase activity. So basically it's a nucleus fused with a locomotor so it can move.
And so this nucleus is recruited in trans to the R-loop-forming cascade. And once recruited and activated, it's going to cleave the single-stranded DNA not base pair with the guide RNA, and then fire its engine at the expense of APB hydrolysis, and erase the DNA for a long stretch of distance.
And maybe that's the reason why this is the most popular CRISPR systems in nature. Because once targeted, it really leaves no chance for the parasitic elements to repair. It erases the genetic information completely. So that's basically-- it's going to generate a much bigger impact upon targeting.
And so this has been the focus of my lab. And so as a biochemist and structural biologists, we really want to understand the mechanism and provide the highest resolution information about the mechanism. And essentially, we're trying to generate snapshots of these complexes in different stages of action.
And part of it is serendipity, but a part of it is hard work. We're able to achieve a lot in this particular system. And so essentially, we solved the structure of the cascade, the effector complex, and it is binding to double-stranded DNA and opening an R-loop mimic. And we solved the nucleus, basically a helicase, a nucleus-fusing enzyme, and we know how it recruits the DNA and degrades it persistently.
And by applying the cutting edge structural biology tools, something called Cryo-EM, we're able to capture the cascade in motion, capturing different snapshots of this complex binding to DNA, opening an intermediate, and opening the entire R-loop, and then recruiting the nucleus forming the ternary complex, and then nick the DNA, and start to carry out its job to destroy it.
So really hard biochemisty, and the resolution determines what kind of a mechanism-- how reliable the mechanism can be. And with the resolution, we're pretty confident that we're really diving into the heart of the mechanisms. And four different snapshots. And so the mechanism is this. So essentially the goal is to scan for a match that's complementary to the guide RNA and eventually open the R-loop and recruit a nucleus to cleave it.
And so the process involves-- essentially the start of the R-loop formation involves the cascade holding the DNA and really bend it to a very uncomfortable situation and trigger the DNA to undergo a DNA bending and melting transition. So you force it to transiently breathe.
And so that moment of breathing gives the cascade effector a chance to check whether the target strand DNA is complementary to the guide RNA or not. So if it's not, then it's rejected. If yes, then it's stabilized into an intermediate.
And then the cascade attempts to further unzip the DNA, very much like when we unzip our zipper. So it's a directional DNA unwinding process. When you unwind the DNA, you need to encompass it for the energy loss by forming base pairs with the guide RNA. So then the entire process become very smooth.
And when the entire sequence checks out, it enters into a point of no return. And so then at that point, a lot of things happen at once. And the cascade is stabilized into an R-loop forming confirmation. So these seriously events involves a confirmational change in the cascade that essentially locks down both strands of DNA, a single-stranded form, and then reorganize the surface in one area of the cascade, and also traps a flexible bulge near that reorganized surface.
So this sends a signal for the nucleus to bind. And so this nucleus is the sequence non-specific. And so it only cares about whether the effector complex has found the target or not. And so once it finds the target, it generates that signal, then a binds. So once bound to the cascade-- so that flexible bulge allows the nucleus to swallow the substrate and makes a nick in the substrate.
And thus nick then spontaneously repositions the substrate into a position that essentially allows the helicase to thread the single-stranded DNA through the nucl-- through the helicase into the nucleus. So then this Cas3 enzyme is in a position to fire as the engine, burn ATP, and start to [? attracting ?] it towards itself, and eventually lead the cascade and move by myself for a long distance. Along the way, it's going to chop the DNA to pieces.
So that is the mechanism from years of biochemistry and structural biology. And we're quite confident about it. And so then the next goal-- at some point, we were like, we know enough about it, and so maybe we can test it in human cells and see now what we can do-- what kind of a genome editing we can do in human cells, and whether we can come up with creative applications from that.
So I started to go to a different genome editing meetings and hear about all the progresses. In one of the meetings, I met this brand new assistant professor Yan Zhang, who just finished training with Erik Sontheimer and started her own lab at the University of Michigan. And she saw through me, and she said, you must be thinking about doing genome editing in type I system. And so I have this perfect system. We should collaborate. And so it was a very, very productive collaboration.
And so what we provided from our lab are good mechanistic understandings and a blueprint for how we should deliver the complexes in the active form, how to target it to the nucleus where the DNA targets are, and what kind of a genome editing outcomes we might be anticipating, and how we should detect them. And from Yan Zhang's side, they have a great editing platform that we could use, and all the expertise in doing genome editing experiments with other Cas9 proteins. And so in her platform, the equivalent of a chromosomal loci, in one side, she tacked with GFP, green fluorescent protein. On the other side, she has essentially a red fluorescent protein on the other side.
So then why do you target the GFP, and you inactivate the GFP signal, the red fluorescent protein is serving as a internal control for the experiment. And so the editing outcome-- so her system is really, really sensitive. And it's the normal cells, not the cancer cells. And so we're really dealing with normal somatic cells and looking at the outcomes.
So with good understanding, that really translates to success from day one. And so the first editing experiment just worked. And we saw-- in the yellow background normal cells, we saw the appearance of red cells. And that really becomes-- so it only appears when we deliver both the targeting complex and the nucleus. We deliver individual ones, they're not effective.
And this is our cell sorting experiment. We're monitoring both the green fluorescent signal and red signal. If it's an unedited cell, the cells should populate along the diagonal lines.
If we inactivate GFP, we're going to see accumulation of red cells at this quadrant. And so that's exactly what we saw. And this is, again, programmable. If we target tdTomato, the red fluorescent protein we saw, green cells are accumulating.
And it's very obvious that the activity is limited by the target searching complex, the cascade. And so the more cascade we deliver, the more activities we saw. And so these are old data now. So we're seeing 13% editing, meaning that every 100 cells we saw, 13 cells being edited.
And now in some genome locus, we saw editing efficiencies up to 60% to 90%. And so if there seems to be some dependency on the chromosomal environment. And we still don't fully understand exactly how this plays out.
And the outcome was quite different from what one saw for Cas9. So this is a genome editor that will really impact a long stretch of DNA. And so this is the targeted site. And this will be the direction of the Cas3 helicase movement direction. And we're seeing directions-- we're seeing divisions along the Cas3 movement direction.
And the start up the deletion is a little bit random. But it's always upstream of the targeted site. And it ends in the random position. But in almost every case, we're deleting a very long stretch, like kilobases of DNA.
And so when we apply deep sequencing detection method, we're seeing that on average most of the deletions involve maybe three to five kb. And keep in mind we're doing RNA protein complex injections rather than delivering RNase and allowing to amplify inside.
So if we were to deliver a lot more RNPs or repeated deliveries, we're going to see very long deletions along this line. We're already seeing the deletions up to 30 kb, sometimes 100 kb. So this is impacting genome in a very profound way, so a true long range DNA eraser.
And so it may not be a perfect scenario for precision medicine at this point, but for a research point of view, 98% of our genome are non-coding sequences. And there are many important genetic elements in these areas. And we just don't have an efficient tool to look for those six elements and understand their function.
And I think that those CRISPR/Cas3 tool is an efficient screening tool to decipher the non-coding genomes in a high-throughput fashion. And the fact that in a single-targeting event, the targeted site is intact in most cases allows one to essentially repeatedly administer CRISPR/Cas3s into the cells. And we're going to get a gradual increase of editings and achieving different outcomes.
And we are thinking about therapeutic applications. And some might be more challenging than others. But the most obvious one we're thinking about involves targeting essentially ectopic viruses in our cells. For example, a very famous example is the herpes virus. So these cells undergo two phases of the infection. And there's the acute phase that causes sores, cold sores, something like that.
And so the herpes is famous for then entering into a dormancy and basically hide inside our nucleus almost like a plasmid, a kind of a circular DNA, and just hide there, and wait for the moment when the host, the immune system becomes weak, and they really come out again and cause a severe infection again. And there were some reports claiming that herpes in some cases cause lymphoma. And in some other cases, there is a link with the Alzheimer's disease. So all these things are ongoing.
A variant of the herpes virus, something called the Epstein-Barr virus, undergoes the same latent infection cycle. And these herpes target the immune cells. And so there's a clear link that the EBV is causing lymphoma in many patients and causing nasal cancer in other individuals. And this nasal cancer case actually is pandemic in Southeast Asia. It seems to be correlated with a specific strain of the EBV virus.
So in both cases-- so if you were to target Cas9, you're waiting for the repair enzyme to make a mistake in order to maybe inactivate the viruses. And so in our case, if we were to deliver our tool into the cells, once targeted, basically there's no chance for evasions of the virus. And so these viruses will be actively destroyed. And so in theory, on paper, we have a way to erase those viruses very efficiently.
And so there is yet another case. Hepatitis B virus, so that exists mostly as ectopic circular DNA. And there's a clear link to cause liver cirrhosis and carcinoma. And this is pandemic in many third world countries. And just a little bit more dangerous to use it against the HPV because sometimes they integrate into the genome. But nonetheless, these are viruses that cause severe human disease and could be targeted by our genomic tools.
So this is a preview of what's going on in our lab. We're still trying very hard to achieve the potential of our tools. I'll just stop here and thank the people involved in the work.
And so all the structural work was done by this-- a single individual, Yibei Xiao, in our lab. And so previously, Bob Hayes also solved some important crystal structures of the cascade. And together, these two really provided a very, very solid set of mechanistic understanding of the system that allows the eventual applications to take place.
And Adam Dolan is this smart kid, undergraduate student, who worked with Yan Zhang's lab and demonstrated usage of CRISPR/Cas3. And I didn't have time to talk about Sherwin and Jagat's work on the acquisition part of the CRISPR biology.
I was fortunate enough to collaborate with many peoples-- many people in the past, and so great collaborators. I just want to give a special shout out to the junior faculty that I collaborate with on the CRISPR biology. And this includes the EM specialist Maofu Liao at Harvard and single molecule biologist, Ilya Finkelstein at UT Austin, and then Yan Zhang for genome editing at the University of Michigan. And funding was from NIH. I guess I'll just stop here and take questions.
[APPLAUSE]
Yes.
AUDIENCE: You mentioned nasal pharyngeal cancer, and Guangzhou in particular, Southeastern Asia and China.
AILONG KE: That's right.
AUDIENCE: Well, one of the things that's been spoken about is the preponderance of people using their cell phones over there, which up regulates the Epstein Barr virus--
AILONG KE: Huh.
AUDIENCE: --at 50 Hz. And if-- I have a home in Dubai and spend a lot of time in Hong Kong and Dubai and that area. Literally young Chinese people in that area live with their phones connected to their bodies.
AILONG KE: That's an interesting observation. Although, I don't know. So this has been pandemic before the cell phone. So I don't know. There might be an enhancement from those exposures. But certainly beyond that, there is some connections already. There was pandemic in the '70s, '80s, and so.
AUDIENCE: And in your bio, you have some training in mitochondrial DNA and epigenetics. You didn't think about that at all tonight and the relationship between the 1,300 [? dB, ?] and the mitochondrial DNA, and the 24 RNA transcriptional genes in the mitochondrial DNA, which we inherit all from our mother. And I just want to know how you feel they are affected by this light versus the natural sunlight which we evolved in.
AILONG KE: Hm. Well, to be honest, I'm not an expert discussing the electromagnetic field, the effect of this field on DNA. Yeah, I think I should refrain from saying too much of that.
AUDIENCE: They're not affecting the DNA. They're actually affecting the software, which then affects the DNA in genomes.
AILONG KE: Well, let me read more into that before I make a--
AUDIENCE: This is very narrow. And managing your light is between 440 and 475. Our evolvement basically came in natural full-spectrum sunlight, 280 to 700. We're living in a much different world than we did, even our great-grandparents.
AILONG KE: I agree with that, yeah. So it's going to affect our rhythm. It's going to affect our vision. You're talking about affecting our transcriptomes.
AUDIENCE: Exactly.
AILONG KE: Yeah. So that, I need to read more. So that's something I didn't--
AUDIENCE: And it affects germline and the preponderance of autism that's happening today in the world, too. It can be passed down not even multi--generational, but maybe skip a generation from grandma to grandchild.
AILONG KE: Right. So yeah, wow, that's fascinating. I thought that the energy was too low to cause that. So let's read something before I draw conclusions.
AUDIENCE: It's not an ionizing energy, so it's not thermal.
AILONG KE: OK, yeah, please.
AUDIENCE: So, given all of the pros and cons of CRISPR that you described here, are there other kinds of techniques that are being explored that might have complimentary or different [? reactives? ?]
AILONG KE: Right. So pros and cons, which aspect? Are you talking about the accuracy part?
AUDIENCE: Yeah.
AILONG KE: OK, yeah. So, in terms of accuracy, I think so far it beats all the other alternative tools on the market. So even though it's not perfect, it's better than other tools.
One thing I didn't mention, our Cas-- so the Cas9 targets something like a window of 20 base pairs. The type I system targets a much longer region, about 30 base pairs. So potentially, it's more accurate.
So there are different alternative tools that are potentially more accurate. And there are alternative strategies. For example, without causing double-strand breaks, you can fuse it with base editors and chase DNA sequences that way.
And so there was this incredible creativity among researchers trying to make it a better tool. So I think it's just a matter of time that we're going to achieve that, yeah. But so far no alternative that's better. Yeah.
AUDIENCE: That was an excellent review. Thank you very much for the well [INAUDIBLE]. Just a comment and a question. You're absolutely right, NPC has been around for many years well before cell phones. So the head and neck surgeons, and ENT surgeons, it's a real problem that is well recognized to the EBV. So if your system can target that, it will be a huge contribution to head and neck cancers that ENT certainly encounter.
But my question is this to you, the ethical dilemma that you described reminds one of the 1970s when I read about with recombinant DNA technology--
AILONG KE: That's right.
AUDIENCE: --and the moratorium that occurred in many places. And then, because we need people into [INAUDIBLE] and people to legislate even knowing what the outcomes would be, do you think the solution's going to be waiting till it gets more precise? Or do you think something else has to happen because there's so homologous to what happened in the '70s with recombinant DNA--
AILONG KE: That's exactly-- so when the recombinant DNA technology just became, and there was like all those precautions, and we're going to-- we should burn all the DNA before we release them into the environment, that kind of thing. So yeah, it's hard to say. I think from the researchers, I think the common agreement is that right now is not so-- it's not safe not to do germline editing. That's the common agreement. But people are trying.
I think in the end, it might be that the therapeutic-- we'll accumulate data from therapeutic approaches. And it's going to reach to a point that we, with the follow-ups, we build up our confidence and to a point that we're comfortable enough to try and germline cells. Because you can hear arguments from both ways.
So somatic editing, it's almost impossible to achieve 100% editing. So whereas if you were to do it in germ lines, if it's a perfect editing, you only need to worry about two targets, and then biology will take its course, and the individual will have perfect genes in every cell. So one could argue both ways, that they could be the perfect solution, but right now it's not safe enough. Yeah.
AUDIENCE: Thank you for giving a good lecture. I was wondering, even in somatic cells, is there a possibility of epigenetic inheritance? Even though you're not, obviously, a [INAUDIBLE] genes, is there a way to affect the environment? Does that make sense.
AILONG KE: Yeah. That's hard to answer, I would say. In most cases, the epigenetics are reset in the-- when the-- in the embryonic-- in the embryos. I would say probably to-- I don't see worries on that too much at this point. I think most people are either doing it ex vivo or localized in vivo.
There are some efforts on the whole body, trying to achieve that. But I didn't hear any clinical trials on that yet. So even that I think people are perceiving with caution at this point. Yeah.
AUDIENCE: I'm sure this is not you're slice of area, but in terms of the safety of these kinds of approaches, do you deal with myeloablative conditioning? There's actually a risk of mortality just from that. So it would seem that [INAUDIBLE] to risk, germline therapy might seem more attractive. Your thoughts?
AILONG KE: Yeah, I agree. That's a painful procedure to go through. And that has the potential possibility to cause cancer and increase the risk of cancer, I would say.
Yeah, so I've read it in both ways. So if the germline editing can be done in the precise way without the possibility of off-target effects, then it could be the perfect solution. But the technology is not there yet. And the public is not with us yet.
AUDIENCE: How do you mail order a CRISPR package? And how much does it cost? Can I get one at Amazon? And how do I use it?
AILONG KE: Yeah. So if you spend $60, you can get it from Addgene. But you have to purify yourself, I guess. Well, you can get it from a company, that really comes in mail.
But yeah, it was a joke. So I'm sure these experiments will fail. That's why I'm not so worried about it. It's the one that-- the educated person who ignored all the ethical considerations, that was the most worrisome part of this.
AUDIENCE: Is it legal to sell them?
AILONG KE: I didn't check that yet. So I should check it before I answer that question.
AUDIENCE: OK.
AILONG KE: So any other people sell various stuff on eBay, so I don't know.
AUDIENCE: Yeah, I'd be interested.
AILONG KE: Yeah. Oh, seriously?
AUDIENCE: Yeah.
AILONG KE: OK.
AUDIENCE: Who else has any questions?
MAN: [INAUDIBLE].
AUDIENCE: No? OK.
MAN: Cool.
AILONG KE: OK, thank you.
[APPLAUSE]
ANNOUNCER: This has been a production of Cornell University on the web at cornell.edu.
Ailong Ke, professor in the Department of Molecular Biology and Genetics at Cornell University, presents a lecture about the growing promise of gene therapies based on RNA research on bacteria. The lecture is part of Cornell's Free Summer Events Series produced by the School of Continuing Education and Summer Sessions.
