DIR Seminar Series - Stephen Chanock

okay afternoon everyone and I did say afternoon and not good afternoon um so as to acknowledge the challenging times we find ourselves in so we will all have to take that into account uh and in spite of that uh continue to do the good work that we are all doing and good work is what my friend Steven chanik does a lot of Stephen is a leading expert in the discovery of characterization of cancer susceptibility disorders in the human genome he has received unsurprisingly numerous awards for his scientific contributions to understanding common common inherited genetic variants associated with cancer risk and outcomes including the KNE house Southworth and visen Bach award in cancer clinical genetics some of you may also remember he was honored to receive the invitation for the Jeffrey Trent lecture for our own Institute he's received the aacr a CS award for research excellence in cancer epidemiology and prevention and he's elected member a fellow pediatrician I do want to point out he's a fellow uh member of the society for Pediatric research aacr the association of American I mean AAP the association of American Physicians and the American epidemiological Society aacr on and on and on National Academy a AAS uh Steve's background is that he received his bachelor's degree which I did not know this in music uh from Princeton University and his MD from Harvard uh 1983 and he trained in Peds uh pediatric infectious disease and heon at Boston Children s Dana Farber and served and for 25 years still serving transition TR okay 25 years two and a half decades as a director of fantastic for kids with cancer um Steven started here at NIH in 2001 uh where he's a tene investigator in the genomic variation section of the pediatric oncology branch of the NCI Center for Cancer Research and 2001 he was appointed chief of the cancer genomics research laboratory and 2007 is chief of the laboratory of translational genomics within DC Steve co-led the cancer genetic markers of susceptibility project and from 2012 to 13 he was acting co-director of the NCI Center for Cancer genomics and director which he is now serving of dcag in 2013 so Stephen uh please tell us about this great work you're doing on this important [Applause] question it's a pleasure to be here um I have uh two Corrections on that one is I came in ' 91 as a investigator and I first came to I it's in 1958 and I lived just down the street and learned to ride a bicycle and lived on this campus for a good part of my life um and so um it's always a pleasure to come back to building 50 which I know has niad on the top floor and NHD on the 5ifth and I don't know who's on the other floors but those those are the those are two very important floors in my life um it's a real pleasure to be here to talk about science something that I'm not having a lot of chance to talk about recently but uh this is I have actually new data for a paper that's out for review right now that I'll show towards the end uh so those moments have been able to do science are far and few between but they're very exciting and they keep me going and they get me up in the morning so today I'm going to talk about um epidemiology and genomics in uh the in individuals who survived or were exposed following the Chernobyl accident and this is really a Michael I went the wrong way sorry next okay okay um this is a long-term commitment in dcg uh I um became the director 12 years ago Joe frani had run various versions of that for over 30 years there's been over a 30-year commitment to understanding radiation epidemiology and you can see people like Gil Bigby and Elaine Ron and Jeffrey how who started these studies and we have very dear collaborators in two institutions one is imem which is the institute for endocrinology and metabolism mikolo tronco and taana banova and then Dimitri Baza runs the nrcm which follows the cleanup worker so we have a cohort of 29,000 people who were brought in during the Soviet age and cleaned up and then left and we followed them knowing their do symmetry exactly how much radiation they were exposed to in what kinds of conditions um and there's a longstanding tradition in in dcg itself of individuals and particularly the last person Lindsey Morton has really been instrumental in the studies that I'm going to talk about right now so I think you all know about know about the Chernobyl power plant accident uh April 26 at 1:26 in the morning morning a reactor 4 exploded in 1986 this was still Soviet Union and it was the end of the Soviet Union and this was a highly economically depressed area and that's really important for the first story because the major source of nutrition for younger kids was milk from cows who fed in the fields that were contaminated this was the major source of protein for younger kids and so unfortunately the i131 that get spewed is concentrated in those areas and as you can see here just is just looking at the Ukrainian map and Belarus got a very strong uh dose as did parts of Russia but the key thing here is really looking at the rates of thyroid cancer which was the second most common cancer probably the most and I say probably is lukemia but most of those individuals died rather quickly so the time that it took to set up the INF structure to be able to follow and capture and have the you know the the tumor repositories and the light all were passed by so the aplastic anemas and leukemias were indeed missed uh shortly after the accident but thyroid is a very interesting organ that has the real propensity to develop after uh external beam radiation radiation therapy atomic bomb survivors so you can see here um that there that there's an age aspect the younger you are when you're exposed the greater the effect of whatever radiation you're getting like most things in Pediatrics once you hit adolesence it's all downhill but here it's just the opposite because once you're over adolescence you're much less susceptible to this if you're exposed as an 18-year-old is very different than an eight-year-old um in terms of the risk of thyroid cancer now ionizing radiation is a Class one iar carcinogen and as I mentioned leukemia and thyroid are the classic ones but we know from therapeutic radiation given to patients that certainly sarcomas and other kinds of skin cancers can come up animal studies for decades have shown the double strand DNA breakage with clean cuts but no one has really understood what's the specific mechanism how frequently how does that happen or not happen so that was really what's dri driven us to try and understand this because at a public health level it's a very important question of sort of as we go from very high doses to lower doses how much risk is retained as you go to lower Doses and at what point do you maybe not have risk for cancer that's a that's a big debate and so by understanding uh an endpoint like cancer which we also know for instance cardiovascular disease and particularly cataracts are very common in individuals who had exposure after um after the Chernobyl accident and cataracts are probably the biggest problem is people get older um so papillary thyroid cancer has a higher incidence in women with thank God a greater than 95% survival it's a cancer of young adults highly differentiated and the mutational burden is interestingly higher in older patients so there's an age related issue and how much of this relates to what we know is the background what we call clock mutations signatures one and five that accumulate in any given cell and mixed with that I'm giving some of the punch lines is that this is a tumor that's driven by one driver just one event that's very unusual in cancers when you look at melanoma or you look at smoking driven lung cancer there can be seven or eight separate drivers and subclones that are all uh unfortunately conspiring so mutational analysis so before had really shown that the BFF 600 and tur promoter mutations can actually be useful for therapeutic decisions so we undertook this to try and understand thyroid carcinogenesis in into radiation induced tumors is there a complex signature can we look at a tumor and say this is radiation based or radiation driven and this one is not because we know in the study that I'm going to show you that there is a background sporadic rate of thyroid cancer in young individuals and the question is which ones are truly driven by radiation and which ones were going to happen no matter what and I'll come back to that in the middle of the talk and does the profiling suggests therapeutic steps and how does this inform this is sort of the the controversial aspect of radiation epidemiology in the last 60 years basically since the studies that really crystallized after following the survivors of the atomic blasts in Nagasaki and Hiroshima which we study and are part of and I'll refer come to it at the very end for a study and we're doing a comparable study that Les has been very helpful with is this issue of when you get the very low doses of radiation there's some that argue that hormesis that a little radiation is actually advantageous PES for a cell there are some that say there's no effect and some say it's not good and there are lab models that that support each of these three with different biases but the question is the clinical epidemiology is very difficult so how do you how do you really quantitate and capture people that live near nuclear power plants or radiation uh workers or radiation technologists which we have big studies of following 40,000 people who have worked in radiation in in you know Radiology Labs or radi Radiology Suites for the last you know 40 years and this is an unanswered question and this is one that really drives the public health questions we know that um thyroid papillary thyroid uh cancer is very interesting from tcga in 2014 it was one of the first tcga papers to be published and here is What's called the classic commute plot and this shows a lot of things all scrunched together and the thing that's notable here is all of this is just the genetics so um over the objection of some of us tcga was done with virtually no real regard for what the exposures were and effort to be able to capture that as well as the clinical outcomes and we are spending millions of dollars trying to recapture retrospectively a lot of that information and it's how do I say holy not in the religious sense but filled with lots of missing data that makes it particularly difficult to be able to infer and understand things because we now know you can use sematic altered patterns you're looking at mutational signatures to work backwards in epidemiology sort of like NCIS you can go to the scene of the crime the cancer and say ah this person smoked this person drank too much this person you know was exposed to Too Much UV light or whatever and whether that actually causes the cancer is another question that um we'll get to but as you can see here what's important to realize is that about 70% of the patients had a point mutation and they're just in four genes BFF and then the three versions of Rass Gene HK andr okay and then about 15% had fusions um and they're relatively restricted their fusions in uh genes that are mapped to the the map kyes pathway and then about 25% they couldn't find them they did exom sequencing not whole genome sequencing in the original paper and then there were basically no copy numbers other than a little bit of 22q deletion um so we then decided that we wanted to comprehensively look at and really characterize the molecular profile in a cohort that had been identified and followed after the Chernobyl accident and put in What's called the Chernobyl tissue bank which used to be in the UK it's a uh International Organization where each sample each case is reviewed by six different different pathologist so you really have a common understanding of the phenotype we've actually moved it to the NCI in the last couple of years and this tissue bank had lots and lots of samples and so I'm going to talk about the first 440 cases that we have access to we have another 220 that are in the pipeline for papillary thyroid and then there are other types of cancers uh like uh papillary carcinoma and then there are also uh you know adomas that we're now looking at so this is something that you know the gift will keep on giving in our mind so we looked at these 440 individuals who had fresh frozen pathologically confirmed PTC um and we did whole genome sequencing ending up with 374 cancers that met you know the QC criteria we did mRNA sequencing on a bit more uh and then we had both non-tumor tissue from the other side of the thyroid and then we had blood you know and a few that are in these interesting trios that are allowing us to really explore what kinds of mutational events may be in the background now the thing about you know uh this particular study is um we know exactly how much radiation each person we have an incredibly good estimation of how much radiation each case got we knew exactly where they were at the time of the accident they were all diagnosed and had surgery in the same institution and only six surgeons over 20 years have done that surgery according to certain protocols and it's very very methodical so we have all these maps in Ukrainian where you know taking this part of the thyroid out this size weigh this amount and it's really I it's really unparalleled in the cancer world to have such exact detailed uh pathologic information um and that's been helpful for some other things and so um we within this 440 we were able to identify 81 individuals who were born more than 9 months after the accident from a region nearby by I don't have time to go into the population genetics but this is all from the middle of Ukraine and if we look at the pcas this is a extremely homogeneous region that uh you know all but two people are very very similar uh in terms of their population genetics background if you go to the to the east where the war is Raging on they look more like the muscovites and if you go to the West it's much more of a combination of the poles and lithuanians who for six or 700 years had large kingdoms that controlled a good part of the western part so we have these individuals who had no exposure so we in a sense have our I wouldn't call controls but our comparison group that's embedded within this so 359 exposed with 250 Migra um normally when you get therapy you get anywhere from 4 to 10 gray so this is an order of magnitude lower this is what people would get by environmental exposure or accidental exposure and this is what is really worrisome at of Public Health if there were an accident or a dirty bomb or whatever these These are the ranges that people would potentially be exposed and it's measured i131 which was one of the main um radio nucleotides it was released and i131 you know has a short halflife so that we could really map out and know exactly how much and the hard thing also is i131 is concentrated in the thyroid because the thyroid uses iodine more than any other tissue in the body so we know there's a particularly high dose to the thyroid per se so we looked at a whole series of uh genomic variables that we created from the whole genome sequence in the MRNA microrna snip methylation profile and T limmer links with colleagues at the broad and in our program uh Danielle karotti who used to be here in an hdri come over she's been terrific in this and we looked at all the all the usual suspects from single nucleotide variant to microsatellites the drivers you name it and then we have these wonderful radiation epidemiologists who have very very sophisticated programs for taking the modeling of the radiation exposure and incorporating the uncertainty because no one was measuring minute by minute but with multiple measurements what do we expect and so the uncertainty principle is really important in this so we did multivariable regression models adjusted for age at PTC and sex and so this is sort of what I've told the story in already one way that the accident the concentration in the fields the cows drink it the kids drink the milk and they get D you know what we estimate to be 20 uh breaks per uh per genome uh 20 per excuse me 20 for one gray exposure thyroid cancer we know what that looks like and it's these uh double strand breaks are really important in generating the drivers and that's what we're going to see in the next thing so here is another commute plot of r440 but this is organized in a way that no other one is that's been published if you look at the top it's by environmental exposure so the beauty of this or the really interest is the integration of epidemiology with genomics so this isn't just saying this is what we see when we sequence and we'll figure out later whether someone smoked or not so we know exactly how much radiation and you can see from no exposure to 1 to 99 migrate to 100 to over 500 and what's really striking is down on this line here with I'm showing the red are the point mutations but the higher the dose the more the ratio shifts from point mutation to fusions and the fusions are with just a handful of genes all that have at least one of the participating members from the map kyes uh um family or pathway so it's really a very very parsimonious uh kind of cancer so when we looked first at after taking the whole genome sequencing it's one of the lowest in terms of mutational Burden the Chernobyl thyroid lower than the adult lower than just about anything else four orders of magnitude less than what you see with smoking or UV where there's just constant bombard and a lot of the mutations are due to the environmental exposures some of them may generate the mutations that are important for the cancer but a lot of them are just you know spread across the background of the genome per se and we see the map kyes pathway is what's hit almost exclusively you know and then when we looked at the numbers and we were able to use the MRNA as well as the whole genome we could find drivers in 98% 93% were a single candidate driver this is really unusual there very few pediatric cancers that are driven by one particular driver now we see that in the developmental ones you see that in you know Hep blastoma you can see that in rabar saroma but not in uings you not you don't see it in in osteosarcoma ones that develop later in life per se so it's a it's it's a really interesting question developmentally that we want to look at and ask why the thyroid what is it about the thyroid what's going on in the thyroid that you just need this one mutation and you get cancer where whereas if you looked at um all of us in this room and took skin biopsies you could see many of these mutations particularly BFF in the skin of just about everybody in this room who's had any UV exposure but we all don't have melanoma we all don't have basil cell carcinoma so uh you know it's a to me that's a really interesting next question that we're we're really trying to get out so when we looked at the 176 Fusion drivers the Rett NT kr3 and BFF in PP ARG these are all things that are fall into the map kise that are part of this one particular pathway that's very important for thyroid development so now the next thing is we then decided to do a sophisticated analysis where we wanted to look at the e that's the excess odds ratio so what would you expect what does radiation do so this is where as you can see the rate of the E per 100 Migra against the dose that's exposed and we could see a very strong effect almost all of these cancers are clonal we don't see a lot of second and third clones per se as I mentioned there were a handful that had two drivers and those you know there you'd see quote unquote subclonality we didn't see the two drivers necessarily on the same strand per se and they would have you could separate uh to the best of our knowledge that and we'll come back when we talk about the metastases fusions primarily in map kyes but again these single clonal drivers are or a hit at the wrong time so the cut of the DNA is taking place and if particularly since fusions are over represented the higher the dose of radiation it's that single dose that hits and creates that particular driver and this is something that really really validated what there were lots of early reports in the late 90s of like eight or 10 or 12 uh cases that were brought out of bellarus and and out of Ukraine where they saw these retti mut s or our Fusion drivers but they were really anecdotal they were important but we had no idea how and what breadth and what comprehensive aspect of that so that's that's an important part of this story in our minds um so we then looked at the single based substitutions thinking is there radiation exposure signature and the answer is no we see the clock the so-called signature five which we still don't understand the origin of whereas one we know is the SE you know a very a very set interesting you know biochemical change and otherwise sbs8 which is unknown which shows up in most other cancers as well particularly pediatric cancer so it raises an interesting developmental question we then looked at the small events anything more than one base when you go to those and we could see that id8 and id5 certainly showed up when were very strong insertions and deletions so then when we said this question all right radiation Associated molecular characteristics Mike Stratton and Company had looked at 12 tumors of mixed types that were clearly driven by radiation and they saw lots of inversions and deletions the same time we published our paper there was a paper from a German group looking at 190 post- radiation recurrent gomas that's kind of brain tumor they saw the same thing an increase in deletions more than inversions per se again a small size and then in RS we could clearly see that as well and we could do this here I'm showing the data normalized against the background of snvs sort of trying to take out the clock mutations because there are occasional uh inversions and deletions in the clock uh program per se but they're a very small fraction so but we presented this as particularly a you know a normalized fashion um so the structural variance we could see these increase SVS with quite substantial and they look to be mostly simple balanced SVS okay so in Pediatrics about 40% of the drivers that are fusioned are very complex where it's a very large part of a chromosome gets connected to a second and there may be a third one attached to that or that if they're just two parts of a chromosome or from two different chromosomes coming together there's a deletion or insertion of hundreds to thousands of bases here again it's this notion of the clean cut that we could clearly see evidence for so we can't see evidence for these complex or the simple unbalanced um and this was really important because then we said all right what's we have the fusion from the RNA we have the DNA fully sequence we could go look at each and every one of the fusions and say what does the characteristic look at will that help us to understand what particular mechanism is actually disrupted and the answer is yes we could do that we could see enrich for non-homologous and joining with the vast majority of the simple balanced having very very small events around the way they would literally fit together and so you know knowing that there are multiple variations on the repair mechanisms um I'll I'll run to that but you could see for less than 20 base pairs of intervening loss we had a high fraction had four or less bases and there was a mixture of a few with four to 19 but was really quite striking to see that and when we looked at this again again having this opportunity of having the epidemiology and the do symmetry knowing exactly I mean we all we also know you know the incubation period from the time of all these people were exposed April 26 to some point in May and then when their cancers developed and or at least when they were diagnosed with their cancer as a as an end point when the surgery took place we could look at that and you could see that the the microhomology analysis for the what we saw in the less than five base pair events really correlated with the radiation so with the more radiation the more of these events are going on throughout the genome but particularly they're important for those circumstances where they generate an actual driver and then um we you know with this primary analysis we you know we know that double strand uh braks are important and no radiation dose relationship for SV and SPS and Driver genes um or particularly we were surprised with transcriptomic and epigenomic profiles I'm not going to go into that in detail but we really couldn't see it it turns it's either Ras or raft that drives what it looks like so once the tumor starts the tumor takes over and if there was thetis of the radiation it would be left behind you wouldn't see it in the methylation pattern and you certainly wouldn't see it in the expression pattern which is a little bit different from what you see for for instance we have this big program in Sherlock lung where we're sequencing 2,000 non-smoking and then there are about 5,000 in the literature of smoking and you can see the residual of smoking in the RNA expression uh and in the methylation here these are such parsimonious uh hard driving tumors so to speak okay um so the question is coming back here is there a way to determine whether there's a radiation characteristic or set of those those so we then decided to look at specific tumors caused by radiation can we find Reliable biomarkers we could see these differences in gene rearrangements versus the sporadic in those tumors yes but that's not going to answer at all we could see this from the Japanese data following uh following the atomic blast but not Fukushima I'll come back to that a little bit later um and then we could sort of see how these different kinds of things really were laid out in in the past here okay uh from previous studies that I'd mentioned so the age the tumor of initiation the clonal clock we know clonal clocks are going all through life in every cell per se so the question is can we use those to be able to identif with the mutations that we have and the insertion delions can we sort of develop a profile that would help us to say with with greater and greater certainty that this is a radiation induced versus a non-radiation induced the one thing I neglected to say earlier was fusions are also more common the younger patient even in the sporadic cases so if you go to the Philadelphia series of Andy Bower and Amy Franco of 135 cases of PTC the younger the child the greater the likelihood they can have fusions not at the rates that we see necessarily in you know post Chernobyl but there is an age element to this that if you're going to develop PTC and you're younger uh there's a there's a greater propensity to have a fusion driver as a opposed to the bff or the RAS point mutations driving it so here we know that these events certainly can occur around at the time of you know the tumor initiation that's what we think of as driving the event due to radiation but it's also plausible that some of these events the radiation may have taken place and then at a later time they're developing a Cancer and can we separate that or similarly did they already have a Cancer and then the radiation occurs and we're looking for you know that was a sporadic case it was already started and it just happens to be they were in you know the Chernobyl region and can we use these you know these set of of markers particularly the clonal clock one and five to be able to determine that the answer is yes we can so if we look at those individuals who have two breaks less than 20 base pair gain or loss you know what I mentioned before is really radiation related we can see the age of the exposure in relation ship to um their clonal clock mutations and the accumulation so this makes sense so to speak if we then go to the more complicated versions of of mutations that are seen and granted there in smaller numbers and particularly the B rafts we can't see the distribution of clonal clock in any way statistically lines up in other words as a as a regular accumulation either the tumor it was at the time of the tumor or just shortly before within a year or so that someone was starting that way and the radiation may have actually induced that per se so that's important and I already mentioned the transcriptomic analysis that we can see these things are all driven by the driver not by the radiation dose and similarly the methylation I forgot that I left those in so I can't come to NHD and not talk about the germine for a minute it won't be very long on this project but when I get to the other projects it'll all be about germline okay so so don't there will be germine discussion um so we looked at the PRS interestingly enough thyroid is one of the the diseases that has a very high um heritability so if you look at the gwos the genomewide association studies for common variant it's like number three out of 35 cancers that the you know can be explained by Snips per se you know CLL and testicular cancer are the numbers one and two a multiple myom is not far behind and there others like esophageal squis and um and bladder cancer that you we we can infer from our larger studies and you know and and projecting based on a large Goos that we're not going to explain much of the heritability by Snips per se these are diseases that are going to be driven by other things but the PRS is very powerful in thyroid cancer and so what was interesting is we took the known PRS there were only 12 and what was interesting is is in inverse relationship the PRS was lower in individuals who had more radiation now this isn't statistically that exciting in my mind but what it what it suggests is to me that that that the genetic burden could be important for the sporadic but not for the PRS is not operating or in any way responsive to the radiation per se now these 440 individuals we combed looking for any and every possible secondary finding any possible Gene that had been implicated of 150 cancer predisposition genes not to mention 400 other genes that have been implicated in radiation biology you hear about them when I talk about clonal hematopoesis in about 10 minutes but we couldn't find germline mutations in any of these individuals granted it's only 440 individuals um and we have like another 300 that are in the pipeline right now and with whole genome sequencing it'll be interesting to see that one of the other things that was really remarkable with this with the imem group is just the care of the Clin the clinical care and the followup allowed us to look at metastases so the clinical annotation all these patients are followed for life in this particular place and I will tell I'll digress for a second but when the war broke out in 22 um the there was a terrible disruption of the pharmacies in the Ukraine and we have these contract s with our ukrainians and they follow you know several thousand particularly young women middle-aged women with thyroid cancer who' had thyroid cancer and they didn't have thyroxin so we took out our credit cards bought in in Poland $9,500 worth of thyroxin twice and had it cross the border and go there so that for a period of time they had thyroxin but it's a you know they we have a very good relationship with them and they're very good at following all of the the details and we could see that you know Advanced disease is associated with larger tumor size in multifocality but when we then had 47 Mets available to us from the same individual we were able to sequence them and they looked pretty much identical to the primary there's one that wasn't and that's because they're about 2 to 3% are multifocal and the met that we actually chose happened to be from uh a different primary you know there was a different primary that when we went and looked at it we could at least see that particular driver mutation but it's an interesting thing that when Mets occur in the in the cervical region they they really are a spread of this particular tumor so we don't see some of the things that people are interested in in in brain tumors interestingly enough the Mets are particularly we see them with fusions um uh more often than we see them with point mutations and this is something we we published last year and we also know that the genomic architecture allows us to really see that the the Mets occur for different reasons sometimes there's a primary monoclonal that is metastic monoclonal others that is primary monoclonal where the metastasis looks to be polyclonal where in other words there's subclonality in the samples that we had and then the primary polyclonal leading to a monoclonal spot somewhere else we have a few instances of those and we're following up on that right now we have a fair amount of sequencing going forward on that so our priorities are to go deeper into the characterization of radiation and PTC characteristics we have another 215 ffp and another 100 fresh frozen that are in various stages of our pipeline analysis and we're working closely with the groups in in Toronto and and Philadelphia about this age issue because this is we can see a strong radiation effect but we can't say that age is not a a contributing factor to this and we you know have to be realistic about that um so single cell RNA attack is going on right now because one of the things we're really interested in is why and what's the vulnerability of the thyroid why this one mutation why is radiation hitting there and doing that whereas there are many other organs that get radiation and don't necessarily go on to sort of monogenic uh sematic mutations that are driving particular cancers um so we are also looking at sematic alterations in the normal tissue with um double stranded sequencing with um right now with lud Mill alexandr out in um UCSD and then we've just started with a number of centers looking at those kids who got radiation therapy the older pediatricians will remember when the kids with Al would get treated with cranos spinal 1 12200 or 1800 RADS uh as prophylaxis against uh CNS disease that's gone its own way we also have uh we're just finishing a study of about 150 individuals with PTC from the World Trade Center cohorts in New York where that we don't know of radiation of i131 but again as as being a significant issue per se it was so hot that some people think that there were different radio nucle tiddes and there is some debate about that but we've jumped in to at least do the genomic care ization to see do we see features that would help someone sort of epidemiologically think backwards on this or and be asking the question of what would be the driving force in that um so that's it for the PTC and particularly Lindsay and Danielle and my former oxcam uh graduate student Eric Dawson who you know now at Nvidia making obscene amounts of money programming things and coming back and helping us I'm going to talk about something these help us you know out of the goodness of his heart I mean it was a wonderful Ox cam I know elen you used to run that so he's one of we've had three or four terrific oxcam people so the next thing I want to talk about is in this setting we designed a family study to address this question of transgenerational effects of protracted radiation exposure following the Chernobyl accident so this is a study we wanted to look at the novo variant sorry I thought I corrected all this sles less cuz we had this conversation in the fall and this one I missed I'm sorry denovo should be denovo variants not denovo mutations and adult children born to Liquidators in evacuees so we had a family design um where which we published in the American Journal of epidemiology so we've now collected over two 290 families but the first paper that we published uh four years ago was 105 families with 130 children all born more than One calendar year after they left the Chernobyl area we know exactly how much the mothers and or fathers were exposed we did short read very high coverage adx I didn't we didn't want to miss anything and now we have the 290 with the clonal hematopoesis study in progress I'll talk about but we've now extended this a bit more we know denovo variants are very important for risk for a number of pediatric diseases autism certainly top one of the top of there but certainly some of the catastrophic pediatric developmental disorders and certainly the icelanders have been you know out front in developing some of the pipelines it's an it's an essential type of variant that due to random mutations for the most part now there are you know there are these ve interesting group about select genes where there may be a higher predisposition for a mutation per se but that you know that's a again a very small fraction of that but these variants are generated by classical Pipeline and that's how the icelanders and how Ben Solomon when he was with Wendy Wong and company over iova did it and would see between 40 and 100 uh that would increase with the age at conception newly appearing variants are known as denovo variants and to do this you have to see both the parents and then say this only exists in the child and that's a lot of QC activity and it's a critical step in evolution because this is how new variants are introduced and can be selected in populations or deleted per se um and and particularly when you think about all of the you know um the mutational signatures and and in Primary Pediatric diseases that we see Den Nova variants are really important in pediatric cancer it may be that 20 25% of bone tumors osteosarcoma are due to denovo variants arising in families with very strong pathogenic mutations all right um prior data on the transgeneric generational effect due to radiation certainly has been there with animal studies where they would get two to four gray which is a pretty high dose and evidence of DNA uh repair non-homologous enjoining with some structural and chromosomal events and we know that fish is is used in the commercial world for following uh people that work in the nuclear uh Reg at in nuclear power plants that they have blood drawn periodically and if they have chrom there are like six chromosomes if they by fish have rearrangements in those then they're considered to have been exposed too much and either stop or they pause for a period of time but um there had been no real good human studies using whole genome sequencing particularly for lower protracted dose exposure and particularly the ability to look at in micro satellites there was originally a paper of Nagasaki of Three Families highly anecdotal very interesting and we'll come back to that so when we looked at our 105 families we could see the distribution followed exactly what had been published in all the other papers from Denmark from Virginia from Iceland you um you know and and the real favoring of a c te which is you know the really the Hallmark of of mutational signature one which occurs most fre frequently because of the biochemical thing but we were able because we had the radiation exposure to look at the age of the parents obviously and saw that the paternal effect is very strong and the maternal was just borderline overall the radiation we did six different ways I'm just showing you the one that seems to make most of the radiation epidemiologists happiest that's what we put in the primary table but in the paper there are five other ways in the supplements we presented it with three and of course reviewers wanted two other ways and we put them in you know it's always you want to have a happy family when you do that and we looked at whether smoking or alcohol was also part of this and we did not see that so there was no difference in mutational distribution between the Chernobyl trios and populationbased studies when we looked at the different types of events from clusters and complexes to indels to micro satellites snvs we couldn't see any of those subcategories so to speak of genetic events enriched due to radiation exposure you know they all had the the paternal effect uh to varying degrees and this we were too small in this study and we've been corresponding with the icelanders about sort of asking the question is that paternal effect more important for micro satellites versus snvs per se and micro satellites have gotten more interesting but when you can look at the parent of origin estimates based on flanking Snips per se we could actually see now the maternal effect and we know that about 42% of the dnvs we can assign to either mother or father and the more recent study just about to come out of um Iceland suggests that 18 to 20% of the events are are maternal and this is you know this the old shth that you know the SP spermatogenesis is going on continually whereas the eggs are basically set at the time of birth well there is obviously some turnover if there is this these kinds of events taking place and so it raises a really fundamental interesting question that I don't have an answer for so we had done this by going through maternal paternal child going through a series of um of QC steps and we actually reviewed every single one we looked at the igv of all of them so Meredith jger and Mike Dean looked at hundreds of them you know excluding some and they each did it independently not knowing the radiation and then crossed their their data and they were about 96 97% congruent between the two of them which is reassuring for looking at this kind of data so we then have now turn this into a pipeline that we're just about to publish and we've done this on a different study of I'm not going to talk about right now a smaller study in Brazil there was an accident where this uh basically this vessel with cesium dropped off the back of a truck in a relatively poor part of a city and a number of people came and picked it up and children played with this and so there we I we've been sent a number of those families and we are doing the same kind of study and so we use the new Pipeline and what we're able to do is it helps us to be able to identify indels and micro satellites at a higher rate per se so we've gone back and we've done this on our Chernobyl 105 we see no difference statistically radiation is not in our minds associated with a higher risk for any of these kinds of events it's just we now particularly the micro satellites we can get and and as people are now starting to use pack bio sequencing and we're going to be doing that on some of the families where we see some very complex things just to be able to better sort of pull those apart we we will do that but uh we're very excited that this is now more automated than it was before because the 105 families took Mike and Meredith literally four or five months of just continually looking at igvs and that well I'd be crosseyed if I did that uh but uh God bless them for doing that so our study considerations I think it's important to realize that we've looked at peripheral blood and this is in adult children so there is a Survivor bias but not much of one I mean again we're looking at parents who had doses that were less by and large less than one gray extended over a period of time so we didn't have parents who who had the leukemia or a plastic anemia because they got four or five gray went to Moscow and died within a short period of time but then but conceived a family beforehand so uh there is a bias but this is an important Public Health message for particularly Japan and I'll come to that in a minute so um we're we're you know we're moving forward with this and we can see that the risk of Del arious dnvs is probably very small if at all I mean if there is something if there is a pediatric developmental thing it's probably more a random as opposed to radiation driven we had adequate power to stratify substantive increases in DMVs in adult children um and we've done epigenetics as well and we really don't see much in the way of what happens in the family the doses were lower and extended and this is important because when we put our paper into science there was one view were who insisted that we put one sentence in and he or she wouldn't allow it to go by until we said there's an interesting alteration in the balance between new gonadal DMVs and DNA repair and this may be you know there may have been some repair taking place and and it's it's an interesting question animal models this is a highly debated issue so since then the in Britain they looked at 60 families who were nuclear test veterans who had been in some of the islands in the Bikini atal etc for the British testing you know and they saw no increase in DNM um and this has been really an important study once we publish this I still go to aacr I get multiple requests for interviews from the Japanese press because of Fukushima because remember Fukushima had doses that are about one tenth of what these individuals in the Chernobyl saw and what's interesting though is one of the major reasons that people have not moved back into Fukushima which now about 12 13 years ago is parents of or individuals of potential Parenthood were worried about mutational events and didn't want to go back and have children because they would have mutations and you know this is something that's really been hard to convince the Japanese because the right-wing press had seized upon this and it's really hard to dislodge that uh the timing of the radiation could be critical and we have more families that uh don't really show as significant change I um there were just sort of a final stage long read NGS of lower of highest and lower doses is going on and the I already mentioned the Brazil but we have this big study that been trying for 12 years at RF and God bless Les to actually go there and assure them that they can have secondary return of results which is a really hard thing for the Japanese to get their hands around but there's a study of five 80 families where they are sequencing 60x and uh it will be very interesting there's different radiation Fallout from that and the exposure curves look a bit different but it will be very interesting to see this and they're in the process of doing their first sequencing we're looking at some of the QC of that so um we see this as really a a very important kind of study so the last thing I'm going to talk about for just about three or four minutes is acquired sematic mutations Al mopis and chromal mosaicisms two ends of the extreme I mean we were we were the first out of gooth to show the chromal chromosomal uh aberration back in 2010 and they occurred about half to 2% of the population boy sorry um oh this is I'm trying to go back in time sorry um and what I want to say is that we knew that inherited environmental factors increase particularly clonal hematopoesis and so Kelly Bolton another terrific graduate of the oxcam who was in my program 10 years ago was one of the leaders of clonal Mapes has been uh instrumental in in these studies she had published a study looking in nature genetics in the slown ketering experience andin seen increase in in clinal map poesis of indeterminant potential in the expected genes but increased in individuals who'd either received radiation therapy or high therapeutic doses so in other words the you know these very strong environmental exposures increase the risk of Chip so we wanted to know at lower doses as you would expect in Chernobyl can we see that so we went to our Liquidators and we have a target of over 2,000 that were doing this but we started with these 290 families with the mothers and fathers we also have the children um and the objective was to look at the frequencies and types of CH and does environmental radiation exposure increase either the chips or the uh Mosaic chromosome elaborations and are they risk factors for cardiovascular disease cataracts and diabetes because we follow these cohorts and we can look at non-cancer outcomes and that's something that as the numbers get larger we'll be able to do so we've now done 581 of the target 2000 higher doses in the fathers than the mothers and um this is a THX of 620 genes all the expected known CH genes and anything ever implicated in radiation biology we put on that list wondering if any of them would do this and then we also did the GSA version 3 Cho so the association here we have red marrow doses okay which is different you know do symmetry has to think about different body parts getting different kinds of radiation exposure and um in this study we've done the uh red marrow dose and what's interesting is we don't see any real effect of the red marrow dose and having an increase or decrease by E so far with our first uh 580 individuals we I'm presenting the fathers separately I the mothers I didn't they have lower doses because by and large 90% of the Liquidators were men who were basically conscripts sent in by the Soviet government to go clean up not told much not trained well and had extensive exposure whole body exposure so this is of particular interest to us so when we looked at the children we saw really no appreciable CH or MCA and we think this is informative because there really haven't been studies of young adults and children for clonal hematopoesis and the the big question in clonal hematopoesis are these events early and they're suppressed because you were really good at holding back some something or is it because your DNA repair as you get older starts falling apart and you generate them later in life so the fact that we don't see anything with thousand neck coverage on you know 330 30 year olds um to speak of is interesting and we're going to continue to do this and and you know that's an important sort of pediatric observation I'm not saying it either makes or breaks the story but it's an important thing that to fit into our trying to understand how and why clonal homat leases emerges in in what age group and and to what degree so having the Thousand X gives us those vals okay so for this work certainly Meredith jger and Mitch mcka and Amy have been marvelous but many others and um I'll finish by just saying that here in dcg we do wonderful training and we have 100 120 fellows at any time and I guess their questions and um dcg does lots of exciting things we love to collaborate okay a I don't have any online questions yet so if there's questions in the room please I would have thought I would have predicted that chip was not going to be responsive to the low radiation was that your prediction also um yes but you know we we need to is a sufficiently large number just like our transgenerational there people say well why are you doing this if you don't see the effect this is where genomics and public health really do intersect we need to know this because there is lots of lowd dose radiation exposure and there may be more either planned or unplanned and we need to know what that really means for subsequent Generations but I I I agree I I didn't expect to see a whole lot I wondered whether we'd see some different genes per per say as opposed to the usual you know methylation uh D repair you know p-53 T 2 axle I was surprised that we didn't see anything but what has already been described in the adult population thank you very cool um very cool actually about getting the Synthroid to them that's really neat um what are you doing with secondary findings in in the ukrainians how how in the world are those being followed up well it's an interest it's an interesting question so this is we started the study like in 2017 2018 and the IRB there viewed this as a purely research finding so we're having new conversations about this because because at least the nci's position is if it's an IRB and it's a truly research study then the return of secondary results is not necessarily required okay and this is hotly debated in different places I realize and the Ukrainian no no and we're talking about unless you and I may not go to the Ukraine for a while once it settle down we may go I mean I go periodically but haven't been since 2020 and I have no intention of going till I know the missiles are not firing but I think the it it is a an important question the Japanese are clearly doing this because they're doing it in a clinical way their their IRB protocol is telling the patients you're going to get this back whereas um you know in the Ukraine it's much more post Soviet you know you how much information you get what you give back is much more restricted I mean we have studies in the bellarus where they don't tell them anything and we say we'd like you to be able to share the overall results of thyroid and now we have a big breast cancer women who are exposed to the Chernobyl accident seem to be having an increase in breast cancer because lactating breasts take up a lot of iodine and you and so you know part women at that time could be a real risk they won't allow us and they won't go back and tell anyone that I mean it's a it's it's a very different world but I I think we will get to the point of uh return of secondary results but it'll take some time I do have another question that's completely different um so given that there are some experts in cancer in dogs and radiation therapy I wonder how how have you convin we've talked Elaine has come talk to usal dog exactly and uh absolutely I mean the the Chernobyl wild dogs are very interesting because they survivorship and the generations are so short you know we're now looking what six or eight Generations out 15 sorry okay yeah so the question is a transgenerational effect would have been selected you it would be hard to know what the denovos are at this point dogs have lower denovo rates which is an interesting thing I I yeah I I learned that really relatively recently someone wrote me a question and it Purdue what the wonderful woman de deap yeah hi Paul very nice talk uh so I have a question about the chip that part of study yeah so you mention those don't have chip but have you look at overall mutation low mutation burden in those patients I mean in those well well we're looking at we're looking we're looking at 620 genes with you know targeted sequenc in Thousand X right we don't I mean in terms of having background like uh uh synonymous variants Etc well yeah yeah I you know it's a great question I I don't know the answer they may have already generated that but we haven't talked about in our data in the paper we sent in we didn't talk about synonymous and you know like I showed you with or or other G not in those 600 G well but those are the only ones that we have a THX coverage we have eight ex coverage in 440 cases of of thyroid cancer and then we're getting data from um couple population geneticists of five or 600 whole genomes across Ukraine just because they're interested in looking at some some interesting questions of the population genetics and could the radiation be related to l1s and things like that right because the contact is radiation may cause overall mutation increase but there selection because as you that these are younger yeah individuals so they may be selected against I mean in the we don't see a lot of point mutations that that's the thing the smv load is very low you know I mean it's basically clock background mutation that you would see in any normal population at basically that rate another question yes I just have a brief question on how you defined adult children for the study what the cut off was there there are 30 years old and older basically because the study you left between 87 and ' 92 and we started recruiting this entire study in 2017 and so the latest anyone had a child I think the youngest was 28 or 29 so adult children from these family trios but not kids as a pediatrician I let me make that very clear sorry if I misspoke it's all right I I get it now thank you my other question is I know your focus is on thyroid cancer but what what I'm interested in why not other cancer right it they for example some of these children were exposed at really young age right but it's interesting they're not other types of cancer correct now the so after the the accident and it was seen pretty quickly based on sort of old Israeli and US Israeli data from giving uh kids with seic dermatitis radiation to the head and then the all experience of cranos spinal prophylaxis that there was a focus on thyroid so we have a bias in that the IM only was looking at cases that were a number of them were screened multiple times or followed there were three different cohorts uh focused on thyroid now the background rate of pediatric cancer is relative is far too low to have to have anything appreciable show up other than than anecdotally one in or two in in these kinds of numbers there was no sort of pediatric cohort set up after that as I mentioned there were a number of kids who had alll and a plastic anemia some of them were sent to Moscow and some of them never made it um out of preiato that area just because this 1986 and the the medical system was pretty uh pretty Spartan and this is a Soviet Union falling apart thank you great thanks everyone I appreciate that and and you more questions afterwards yeah meeting with with train [Applause]