Colon Cancer Genetics

today"s speaker is a long termfriend. that is the case today, as you have the -- as i have the opportunity to introduce, who i am known for 25 years, as we have laned together in a very exciting field called genomics as it's applied to lots of humans conditions, especially as

you'll hear about today his contributions to cancer are adding to our wonders at what we can learn about this disease. his education began enindia, he got his undergraduate and master's degree, then he went opto get a ph.d. in genetics at the university of illinois in

issue bana after postdoc at yale. he, then, option to several academic institutions, princeton, university of illinois, alpert einstein college of medicine and most recent harvard, the professor of genetics, professor of medicine

at harvard medical school. his healthcare has covered many irises of human -- areas of human genetics, we spent a lot of times together in those days. he has done significant work, in particular, human genetic disorders. he's been a very significant

contributor to the cancer genome atlas, tcga, joint efforts between the cancer institute and the genome institute. he has given of himself to many other important endeavors. currently a member of the national advisory council for the national human genome

researchã±r institute, and also a member of the presidential commission for the study of bioethics, where he contributes important insights to the dilemmas to read genomos and other interesting aspects of our personhood. i'm delighted he's here 250

speak to you about the genomics of cancer. police help me give -- please help me give him a warm welcome. [applause] >> francis, thank you very much. i think i'm not going to stand behind the podium. when i do, nobody can see me.

it's a great pleasure for me to be here. i spent many, many times here at the nih in a variety of different ways. and, of course, grateful for the nih for all of the financial support that i have received over is years.

so i want to tell you a little bit about how our knowledge about cancer is changing the way that we think about therapeutic approaches to cancer. so this is the tale of two presidents. first of all. many of you probably recognize

that it was during the time when president roosevelt was in the white house, that the national cancer institute was established. and the time that the nci was established, it was recognized that cancer is abimportant disorder.

it was not very clear as to whether or not we would really be able to do anything about it. at the time that the president roosevelt signed the bill, it was just beginning to understand about cancer. over the years things have really changed.

when president nixon was in the white house, this was the declaration of the war on cancer. whether or not the war is successful can be debated. one thing that cannot be debated is the bill that president nixon signed increased the amount of

money that the national cancer institute has had. many of the people like me who have been receiving support and have benefited from this. obviously, the reason why the people of the united states give us money is really to try to solve the problems and try to

cure disease. so it is important to be able to ask what has happened to this important disorder? and what kind of progress have we, as a community, have made? and as you could see, here, i was able to find the data only not the last five years, but

five years ago. as you could see here, in all of these deeps of cancers -- types of cancers, the overall survival rates have begun to get better. but we're not quite there. if you think about certain types of cancers, lung cancer, survival rates -- if we want to

be able to increase the survival rates and make cancer history, what do we need to do? and so one notion is that we need to be able to try to gain a better understanding about and so how can we think about that? and there are many, many

different approaches to understand this problem. but one approach is a genetic problem. and we all know that cash is agen -- carp -- cancer is a genetic disease. the reason, germ line mutations in a number of genes cause

cancer predisposition syndromes, in many of these cases the pen transplants with which these individuals develop cash is very high. the other piece of evidence is that many of the genes involved in this cancer predisposition subject as the gene that causes

the predisposition for colin cancer, germ lime mutations are present in all colin cancer. and over these last many years we have identified a large number of tumor suppressing and ang aegeans and mutations are found in cancers. and we're beginning to recognize

that it is not just a mutation in genes that are important, but varieties of other types of changes such a copy number changes, dna methylation changes, changes to the levels of expressions of genes and proteins are also very important.

so the notion was that would it be possible to be able to try to do a comprehensive understanding of all of these genetic changes. if we do gain such an understanding, that it would help think about therapeutic approaches. and so this endeavor now really

became possible from the human gene project. and as frances mentioned, i had the privilege of being part of the human gene project. all have our recognize that this -- you recognize that this provided the blue print of the human genome but also developed

new technologies side with that, and all of these and thetechnology advancements: relatively lower expenses of being able to sequence the possible togenomes is makingfx be able to undertake a comprehensive understpading about the genetics and jokes of

cancer -- genomics of cancer. so accomplish this goal, the institute and national cancer institute established the program called the cancer genome atlas. and the particular name is attributed to be given to this program by francis, and so the

tcga stands for the cancer genome atlas. the goal of this program is to examine large numbers of solid tumors. in each of these cases, to get 500 tumors from patients when untreated, and'o nucclayic acids from them, also

individuals and conducteok same varieties of analyses for all of these. and one of the wonderful things is that you would be able to do -- apply varieties of different types of genomic methodologies with the same samples.

this is an unprecedented program to try to gain such an understanding. to accomplish this, a number of different types of centers that have been established, and one of those kinds ofxd centers are so-called cancer genomic centers, and i haveã±r the

privilege ofxd running one ofthem at harvard. and so i won't tell you about all of thexd different thingsthat the tcga has been doing. i'll take a little bit ofã‘imostly the work from our group but a little bit from other places to tell youym how things are

changing, and i want to emphasize today is to how this important is actually helping us to think about therapeutic approaches for cancer. so one of the approaches for this is to be able to do sequencing of dna. there are varieties of different

types of sequencing that you can do. one, look at whole genome sequencing, this is ha we do in my group. it is also possible to be able sequence the exomes from the tumors and also possible to be able to sequence all of the

transcript ohms from each of these sets of genes. so let's take some of these things, how do they work? first one is be able to try to use this information to determine copy number changes. this is done is a very simple fashion, illustrated here on the

upper left-hand side. and the way it is done is you sequence the tumor dna and normal dna, and segment the genome into small bite sizes pieces and just count the number of sequence tags that are present here. so when you see the cases isnq4ã·

controlled. approximate number of leads that are present in the tumor, that means thatã§ã³ that region has not undergone any types of copy number changes. if the number of reads that are present in the number increase relative to normal, there is an

increase in copy number, and the reveres of this, there would be copy laws, you would be able to detect that. the picture on the bottom really shows for one tumor, this is a colon tumor, as to what the profilejf for the entire genome is.

you can see large numbers of the genome that are unchanged, and there are some segments of the genome where there is increase encopy number, and there are others that are reduction in copy number. so you take one tumor, what it looks leak.

and here are the data from 500 tumors, and what they look like. and in this slide, all of the red and blue that you would beabling to see, though represent -- regions of amplication or loss of a copy of the gene. and i just want to point out

that basically, there are a bunch of tumors seen at the top of this slide. they don't have any color. turns out these are tumors that do not have much copy number changes. what is interesting, on the right-hand side it shows the

no.o somatic mutations that are present in each of these tumors. and what is really interesting is that all of these tombingers that have -- tumors that have these stable copy number changes have a very high level of mutation. so where do these high mutations

come from?xd and do they have any clinical significance? so you can ask that question. and he'll illustrate this point by looking at data from colon cancer, in this case, the number ofã± r. 224ã±r tumors are plotted.

and as you could see, there are some tumorsym xdxd that arew3ã‘iextreme right-handã§ã³ side, that has a a fewi]p, t(q mutatr nã‘ asoneã±rxd per megabase of dew point that is sequenced. there is a continuous range ofxd them, on the left-hand side, you would be ablatives there are

some tumors that have as many as 800 mutations. [technical difficulties] so why would a genome containveryã‘pdigh numbers of mutations and what are the possible types of mechanisms that are result in types of changes? there are basically one

mechanism that is well understood, as to how tumors are acquired high levels of mutations. that is shown on the right-hand side. as you all may know, every cell, mammalian cells or other organisms, have a system to

repair mismatched of dna that might occur either due to changes and there is axd system, the dna mismatch repair system that is capablet( of recognizing these types ofc changes, and capable of repairing thoseã‘i mutaâ«ns.㺠d turns out that germ lineand itã‘ir

mutations inã‘i these genes can occur. germ line mutations cause cancer predisposition syndromes. somatic will also result in high levels of mutation frequencies. second possible way in whichã‘i mutationslp can occur isq actually throughc

and women who's breast cancers have ampplication -- they show excellence response with this drug. when this discovery was made in the case of colon cancer, it was suggested that because it is also amplification or expression, these patients might

also benefit from treatment. so what is also interesting is that this ampplication, discovered in breast cancer is found to be present in small portions of not only colon cancer but also lung cancers, and they can indeed bothriad in a clinical trial -- be tried in

a clinical trial with this. now, using these approaches it is possible to look at the types of pathways that are effected. one of those pathway ways known to be involved earn colon cancer is the win signaling pathway. the connondle pathways that are known to be present in colin

cancer are the gene i mentioned to you, apc, a large proportion of the tumors have apc we were able to go and look at, not only the apc gene but many members of the win signaling family. and all of the genes shown here on the -- are found to be

altered in colon cancer. so what this actually tells us is that it is not just one gene that is important, but there are many tiff ways of -- in this -- by which this pathway can be altered, and the cancer cells use all of those mechanisms to change the pathway.

and these results that i showed you here about the genes that are involved in them are copy number changes, and mutations that are present. and also turns out that there is another very interesting mechanism by which this pathway is altered, and that is shown

here. this whole genome sequencing, when you look at them, it turns out that the way that it is done is that a piece of dna usually 300 base pairs, is sequenced from both ends. and both ends that are sedweebsers neek cleo tides, so

map to the genome, so you would expect those ends of those sequenced fragments would map about 3kb apart. occasionally, they do not. there is an example of one of those. and during sequence, all of those gray regions are

sequences. and the red dots really represent is the so-called read. and this particular fragment one maps to this gene tcg7-l2. the other end maps to a gene 1a. so what is -- what is the effect of this? in this particular case it turns

out that this rearrangement that involves these two genes is rather complicated. but the ultimate goal of what happens through this rearrangement is shown at the bottom that the red here, all of the transcripts are shown here. the red shows increased level of

expression. this rearrangement now resulted in overexpression of this gene. what is gtf7-l2. this is a transcription factor that is present in the nucleus, under normal circumstances, a transcriptional represser. when the win signaling pathway

is turned on, it's binding and results in the transcriptional activation of many oncogenic and growth regularity pathways. including mic. this is a completely different way in this pathway gene is altered. and so the question is if you

find such -- are they important? do they have biological and my colleagues matt meyer son owned adam at harvard of done the experiment of taking one of these cell lines that contains such a translocation, and use rna to inhibit that. shows that this is a

[inaudible]. so particular ones are present in different studies, different frequencies. 8% of the colon tumors have these deeps of changes. so that is actually a small subset. in a bunch of other deeps of

genes, transportation lour indications that involve members of the win signaling pathway, all of them in colon cancer. and so don't have to look at another one of them separately. but you could see that there are a large number of genes that are turned off.

or tissued on, that -- turned on that increase in the pathway. now you ask what is the importance of these? and if you -- this is also -- each of the tumors on the vertical access, see all the changes present. most of the tumors have

mutations in apc, well established, a whole bunch of other types of changes are present. and what this result would suggest to us, the fact that some of these types of translocations are shown to be functionally important, is that

more than one way in which you would be able to alter this win signaling pathway as a selective advantage in these particular sets of tumors. the other thing is when you're now thinking about developing win signaling pathway inhibitors, it's not sufficient

to look at the genes and genetic status, but you need to look at all of these different sets of genes and understand their status, otherwise you would we able to relieve some critical patients out of [inaudible]. so this is also would suggest that the old way of trying to

look for mutations in one gene or two is not going to be sufficient, and you need to be able to use modern approaches to try to analyze the genes and genomes of these tumors before you make clinical decisions. is this true just for a win signaling pathway?

it is true for all the pathways that are involved and implicated in cancer. and one other pathway that we all know very well pi3 kinase signaling pathway and the wrath signaling pathway. and leak mutations in care-as and n-ras and b of ras are

different in many tumor types. we show in this picture genetic changes in these tumors. but what i want to show you in this next slide is that it turns out that if you look at these pathways, many, many of these genes are now also subject to rearrangements, like the ones

shown, like in the case of ras genes. all of those things shown in the red, for example, they're all involved in chromosomal translocations. in many cases it can directly show that the -- that is though such translocations result in

changes and expression of the gene, suggesting that they have functional significance. the other feature that's also very interesting and suggested these would have functional significance, is this mutual exclue sievety analysis shown at the bottom, at the top of the

two lines, all the tumors that have h care end ras mutations. the bottom line shows the tumors that have structural aberrations. you can see in those cases that have these types of structural aberrations shown on the right-hand side, they do not

have point mufftations. so that suggestions there is alienate way to be -- alternate way to change the expression and the patterns of these tumors. this is -- there are other types of things that you would be able to see. and there are some results from

lung cancer, structural aberrations that are found. and the interesting features about them is that when you do this type of analysis, certain types of structural such as eml, l-4, that is well-known for which there is a drug that's available right now, is

detected. but there are a whole bunch ofory types of things that are related in kinases that are involved in these structural all of them result in activation of these genes. again would suggest for those cases where there is -- because

of the kinases and susceptible por drug development. there are opportunities for new drug development using these types of targets. now, how could all this information be used in thinking about therapeutic approaches? and this slide actually shows

the -- emphasizing the point i was trying to make earlier about understanding these pathways. so what we did for all of these different tumors that have been analyzed, we looked at plotted which of the pathways are actually altered? and again, each bar represents

one tumor. and so as you could see, there are certain types of tombers in this particular case, that had their win signaling pathway in a ras acvating, pi3 kentucky nation -- kinase mutations, or loss of function mutations you can see different types

so the importance of this, if you're actually thinking about therapeutic approaches for these patients, you need to know this information. for example, it has been known in animal models if that you take a mouse that has an activation in the ras map kinase

pathway and tried to use a mec inhibitor, it turns out that those mice would develop transient response but they become resistant. when you look at them, turns out they have now turned on the pi3 kinase pathway. so one needs to know, then, what

that means is that if you take the ras mutations in these patients and tried to give them in a mec inhibitors, they're not going to work. you need to understand that some of these patients, you need to inhibit both ras map kinase pathway as well as the pi3

kinase pathway, to have. so that's obviously important. this kind of information is -- but the other interesting piece of information about this is that this type of feature of -- you know, overlapping sets of genes that are mutated, is not common to all cancers.

this is somewhat unique to colon in lung cancer, the pathways are unique as i'll show you in a second. in this particular case it's what it means is that we need to be able to look at all of the genes involved in each of these pathways to understand whether

they're turned on or off, those pathways, before you could think of an appropriate therapeutic approach. now, i want to just give a one more illustration about the importance of these structural and that comes from this study with the melanoma.

again, we look at large numbers of melanoma patients. as you expect it turns out a significant number of them have b.ras mutations, nras and so on. shown by the red regions. there are tumors at the right-hand corner that do not have any of these types of

what could be the reason? why? what is the driver for those? it turns out that if you look at these, all of those tumors at the upper right-hand end, contains some type of structural aberration that are shown here. and all of those structural

aberrations shown on the bottom, they're an important gene. mdm2. regulator of p53, akt, important component of the pi3 kinase pathway. ras genes which are important in these constructeral aberrations turn on these different set of

genes suggesting that those tumors have turned these functions on, and we know only about one of these right now in a functional assay turns out to be an activating change. most likely, the other ones are, too. it would suggest that these

types of structural changes that has been recognized to be important, in malignancies are not as well as appreciated in solid tumors but they also play a very important role. all right. so one other feature that is actually important from these

studies using these types of genome sequencing efforts, is that the way that the analysis is done is all of the [indiscernible] obtained map back to the upgenome. anything that doesn't back to the human genome is left. turns out that the dna that is

left aside that doesn't align to the human genome has anonymous amount of tremendously valuable and one of them is mitobeyond creial dna. even at low coverage, you have huge coverage of might cond creial dna, and small amount of genome present.

one of the interesting things about the analysis of mitochondrial dna a couple examples, they're all as you know, that the energetics of the cell are normally using the phosphorylation system, and the other energetic system is that if this fails, use the cycle.

and that is also known owes [indiscernible]. owned it turns out that in a large number of these tumors that you see, actually turns out that one or more of the genes in the mitocondria are tougher and also, most interesting, it turns out that these mitocond

dria turn out as heteroplasmic. in the tumors, those that contain the mutations are homo plastic. a strong selection for them. and this is true in a bunch of cancers that have been examined, and other types of jeeps are also -- genes are also -- genes

known to be involved and cause the -- causing childhood disorders such as mil.syndrome, those germ line mutations, mitochondrial, somatic mutations of these are found in these tumors, also clinically significant, because there are -- it is possible to be able

to develop synthetic [indiscernible] if the cell is completely dependent on using the cycle inhibition of that pathway, would essentially abroke gate the energetic functions of the cell you can kill them. there are several such drugs.

so given all of this piece of information, now, what are the possibilities for colon cancer? to understand the therapeutic approach of colon cancer, a longest period of time, and even today, there are two chemo therapeutic regiments that are most popular.

a cocktail of chemotherapy drugs that are used and virtually every patient gets them. there is egf receptor inhibitors that are used in these patients, and and dro genesis inhibitors also approved for this. now, when we look at all of these different types of we

find, it turns out that there are currently, in drugs in development for win signal pathway and 95% of the tumors have changes in win signaling pathway, ras mutations, 40% of the tumors, have ras, pathway dewrath mutations, 15% of these tumors bras.

they constitute about 50% of the amp pliication and so on and so forth. so now, a cancer type for which there wasn't really much of a therapeutic opportunity, and very little targeted therapies, now we don't have the opportunity to be able to think

about a large trial. and there is currently the national clinical trials network is contemplating and planning on doing a clinical trial in which it plans to recruit about 3,000 patients. and examine them for all of these different sets of genes.

and then assign individual patients to specific therapeutic approaches passed upon the genetic and as you know, there is a similar trial here at the nih called the match trial for all types of cancers and being able to do these types of analysis.

and there is also other trial for lung, squamous cell carcinoma. also these types of analysis. so in a few years that we would be able to have information about how to deal with it. so i just want to close by saying that what i told you

primarily focused on colon cancer but this is not restricted to colon cancer. lung -- lung cancer, 11 classifications are known. every one of these categories, there is either an approved drug or a drug in development. so at our institution, all of

these patients, lung cancer patient are tested for all these different changes. and the nature of the treatment that is provided to those patients is dependent upon the genetic profile of the tumors. and a paper that is michiganed by my colleague, he shows that

it is just not the lung, adeno of squamous, breast, all these different cancers can be classified into the subcategories or stratify these patients based on genetic differences. interestingly, it turns out that all of these types of genetic

changes that are seen, they have either drugs that are approved or currently being developed. so we can anticipate, then, the flex few years that as some of these drugs get approved they would be able to increase our capacity to be able to work through one of them.

to summarize. our knowledge about cancer as a result of this large scale effort is increasing, and the knowledge the increasing at a very rapid pace. we're beginning to understand the role of many, many different types of genetic an genomic

changes that play a very important role. and we also begin to better understand the different ways by which the different pathways can be altered. most importantly, many of these changes are associated with approvinged or emerging

therapies. and i want to close by acknowledging ail of the members of the tcga, and many of them are involved. and my own colleagues at harvard and the program that started, with [inaudible], she now moved to md anderson. thank you very much.

'4 >> we have time for some questions. if you have a question, come to the microphone where people can hear, including those through will watching by video. we seem to have a question hazard.

yes, please. >> thank you. i have a fundamental concern about using specific cancers and identifying mutations in the middle of a very chaotic the structure or on the structure. and identifying too many

different genes, which are mutated. and, of course, they should be mutated when they are growing to cancerous cells. why aren't we focusing on the right focus, which is, to see what causes the tumor verses tumorogenic pathways of immune

or cancer cell virulence, which has been known a hundred years ago, and we accidentally bumped into it when we were inin atry diseases. -- inflammatory diseases. why aren't we looking to see what causes immune dysfunction? other than, there is dysfunction

in [indiscernible], owned so forth and so on. many receptors are there. and, of course, you would expect that the changes in the balance between wound healing, growth factors, and epstoickis should. >> great yes. thank you.

yes, there -- my talk just deals with genetic and genomic but actually, as you may know, there are very exciting developments in understanding the immune item and the role they place in tumor surveillance and so on. there are a number of genes that

are known to be involved in the immune check points that have now been identified. owned some of those i, like the ctla4, and antibodies have been approved for therapy in the case of melanoma. they are showing fantastic results.

and there are many other molecules, like p1, pdl, many of them are in clinical trials currently. they're also immune check points also showing really extremely promising result. many people think about the way to approach them, is to not just

to think about one approach. identify the genetic and genomic changes, try to stratify the patients based upon that and try to get a drug that would be able to inhibit the pathway. and get the immune system to work at the same time. and using both of those

approaches might ultimately turn out to be the very best way in which we would be able to care for cancer. >> you talked about the interesting phenomenon of [indiscernible], a chromosome gets shattered into pieces. you showed the circus plots that

showed in those same tumors, most of the chromosomes aren't effective. seems to be a specific chromosome that just gets blasted. and it's hard to imagine what's happening mechanistically there, that happens on a chromosomal

level. not even a regional level. but the whole chromosome seems to be effective. do you have an idea what happens there? is this one step or an original rearrangement that destabilizes the chromosome and over many

cell divisions, ends up in a scramble? what do we know about that? >> so this has been considered to be a catastrophic event that occurs once, and the mechanism is proposed poon the following. under normal circumstances, there is normal disjunction of

mitosis but occasionally, some chromosomes lack and enter into these micronuclei. what happens is the -- each of the nuclei start application processes, so these one chromosome that is sitting in one micro newicallyious, dna synthesis is supposed to be in a

completely discombolulated from the rest of them. that is proposed as a mechanism. but we don't have evidence to direct -- direct evidence to support that, except from these kinds of studies that show clearly what is happening, that region of the chromosome under

goes a significant amount of [indiscernible] >> you haven't caught one in the act. >> that's correct. >> over here. >> have you found any distinct association correlating with your diseased with history tone

methylation? has epigenetics entered into your equation? if everything is blasted, there has to be something controlling the chromosomes in the first place. as you know, epigenetics is a big deal.

mistones a -- history tones are a big deal. this would play into what you're talking about? >> good question. there is significant amount of role that dna methylation plays in all of these tumors. variety of ways.

for example, these hyper mutation that i talked about, many of those cases that actually result doctor the methylation, and the methylation enactvation of mlh1 gene. so that's very common. in many of these tumor types, many of the genes involved in

the methylation and maintenance are actually mutated, so dna methylation plays a very >> one other question. you're talking about cancer per se but you weren't talking about well knowing, and important precancerous states. for example, endometrial cancer,

you have hyper malaysia, a tim-- that would be very helpful. displasic polyps, in terms of colon cancer, that would be very helpful if there office good marker. you have -- if there was a good you have hyper malaysia in breast lesion.

2340 one can tell which ones become cancer or not. a big question mark. in terms of lung cancer, you have bronchial dismalaysiaia, epithelial displacia. those are meaningful clinical questions need to be addressed via this in terms of diagnostic

testing. >> you're absolutely right. so first of all, with regard to tcea, one of the constraints was that we had to get tumors from patients who were untreated. and the tumor should be a primary tumor. and it should be an excised

tumor. and it should provide enough nucelaic acids to be distributed to all these different centers to do all these different types of analysis. so we were constrained by that. and many of these other types of things that you're talking

about, tremendously interesting to look at. and to be able to see when does -- these changes occur, and whether interventions at early stages are important or not important or what are the kinds of interventions that we're going to have?

and so one of the things, some of us in the scientific community have been proposing is not look at the 10,000 tumors that we look at. that we should have a program to look ought a hundred thousand or more tumors of all stages and all types.

and as i know, all of them are constrained by the amount of money that we have. >> there is that. well, on that note, i'd like to certainly invite all of you to a reception in the medical library where you can continue the conversation, but please, let's

thank our speaker once phenomenon for a wonderful present -- one again for a wonderful presentation. [applause].