Biological Cancer Treatment

>> i hope you had a happy thanksgiving. we're back in class on traco. we've talked about cancer treatment with chemo therapeutic, kinase inunderrors, radiation therapy, but the best way is if you can prevent cancer by detecting it early when it's

still in a reversible stage. today we have eva szabo from the difference of cancer prevention talking about nonsmall cell lung cancer. >> thank you, terry. i'm going to talk about nonsmall cell lung cancer,'s a treatment and as prevention.

i have to say this was strange, i know there are a lot of people on the phone, but seeing four people, wow, i'm going to be looking at you guys. so just to give a little bit of background, i've been doing this for 15 years or whatever. the estimates continue to go up.

number of new cases for 2014 is 224,000, about almost 160,000 deaths are expected. you see that ratio, most people who develop lung cancer die. that's the bottom line. this is the leading cause of cancer deaths, more so than the next three leading causes of

cancer deaths in the u.s. put together. however, the good news is that the death rates have been decreasing since the '90s. finally decreasing in women, that has taken a long time. and you can see that here, this is years, 1990s, you really

have the peak in the number of deaths in men, and then you have a fairly steep decline, although it's still a huge number of deaths of lung cancer, these are all the other cancers, in women the number of lung cancer deaths is not quite as out of proportion, it's only now

starting to go down. the bottom line, five-year survival is dismal, 16%, it's gone up only a little bit in 60 years, you don't find it until it's generally spread. this is one of my favorite slides, it's old but it's still the title, radio radiographicuse

linking tobacco use to lung you have this cancer here, the muddy lung, these little things are emphysema. there's the pack of cigarettes on the x-ray. this sums up 85% of lung cancer, by no means all. so the risk factors are tobacco,

secondhand smoking, of course not as much but still a significant component, but it's mainly active smoking. prior aero digestive malignancy, copd, emphysema, chronic bronchitis, pulmonary disease generally associated. those are the main causes.

there are other exposures, asbestos, indoor radon, that's why you have your house tested before you buy it. a number of less common exposures. and some genetic predisposition. and others, there are susceptibilities, you can see

the nicotinic acetylcholine. staging has to do with the size and the regional and distance spread. the smallest cancer, these are stage 1, 3 centimeters or less in size. five-year survival is stillome 61%.

it's not all due to lung cancer, people smoke, people are older, the prime lung cancer patient has other cardiovascular diseases, emphysema, so on. but a large part of it is lung and even from these smallest tumors it goes downhill rapidly. that's the reason terry said

much better to prevent than to treat. having said that i'm going to talk a certain amount about treatment. there's been a revolution as to how we approach lung cancer, personalized cancer therapeutics, and i think this

has -- everybody should know it. not all lung cancer is the same. about 80% is what we call nonsmall cell lung cancer which used to be grouped together, but it's not small cell lung cancer. you will have a lecture on that. and the grouping historically is because small cells for the past

30 years we've known it's very chemotherapeutic responsive. nothing curable in most cases but still very responsive, and nonsmall cell classically is not as responsive. it used to be grouped together because it was approached the same way.

that is changing. but the majority of are adenocarcinomas, the majority of people who are never smokers, they generally fall into this adenocarcinoma category. however, there's a number of cases. squamos cell cars carcinomasused to

be the most common, they have to do with how cigarettes are smoked and made, you get deeper penetration into the periphery of the lungs, which is where you find adenocarcinomas, out here, as opposed to squamos and small cells. large cell is sort of an orphan

lung cancer disease. small numbers. and then there's carcinoids. small cell is aggressive, it is very rapidly growing, so i'm going to talk about nonsmall cells today. and prevention of nonsmall cell. having said that, i sort of want

to set up how i'm going to discuss the rest of this talk. so when you start with your normal epithelia, you expose it to whatever terrible things such as a cigarette. you could also have a variety of histologic and molecular changes such as dysplasia, the

epithelium become more atypical, eventually going through and you have the basal cancer. you have a very long time to prevent tumors. you have a shorter time to detect it early in the early invasive stage.

and you have about three years treated if it's metastatic a little bit less time. let's start with treatment. the treatment is anatomically based. lung cancer, nonsmall cell, early stage 2, small lung. if there's lymph node

involvement, or other prognostic indicators, adjuvant therapy is after that. once you start to spread to the central lung media stinal lymph node you're out of the surgery alone category and we tend to use radiation therapy with chemotherapy.

sometimes with surgery. and once you're past that, you have metastatic disease, either within the lung, multiple nodules, outside the lung, you generally have to do systemic therapy, chemotherapy with radiation and even reception of isolated metastatsis as needed

for control. small cells chemotherapy based, plus or minus radiation scenarios. so as i said, we used to think of all nonsmall cell as being the same but we now now that's not true. and where we really know it is

in the case of lung cancer. 40 to 50% of all nonsmall cells you can find various molecular abnormalities, which this piechart is growing. there may be 40% of nonsmall cell lung cancers that don't have some of these nonmolecular aberrations.

many of which can be targeted for treatment. let me show you some data where treatment targeting these specific abnormalities can become standard first response therapy. the poster child of treatment, it is an important growth factor

signaling pathway involving many integral processes involved in carcinogenesis, growth, survival, antiapoptosis, you name it. i'm sure the people here have heard about it elsewhere. this is for a small subgroup, 10%, about 50% of never smokers,

so it's a small but extremely important subgroup, where targeting the up determinal growth factors receptor with a specific kinase inhabitor. actually, results in very high response rates, prolonged disease stability, and probably longer survival, say probably

those studies have been harder to do, mainly because it doesn't matter whether you give that treatment up front or past the chemotherapy, as long as you give it to somebody with the proper genotype, egfr in the tumor, you can get a significant clinical response and people

live 30 months or so, some much longer. this is the very important therapeutic change. really in the past ten years. and so erlotinib is approved as a single agent, completely receptive populations.

a second very important molecular abnormality is the eml 4-alk fusion gene. this abnormality was first identified about seven years ago. it's about 5% of all nonsmall cells in the united states,

again almost primarily never smokers. and there is an inhibitor, crizotinib, really developed as an inhibitor of the oncogene, but it also is excellent inhibitor of this gene product and has similar response rates to erlonitib, about a 60%

response rate. the majority of the rest of the people have stable disease. the length of response is again 8 to 12 months. now there's a second line agent that has gone through rapid development, ceritinib, also approved for people who have

progressed or can't tolerate it. this is a rapidly moving field. people who have these abnormalities, there's a lot of work being done and lots of options, at least two other drugs in development here. again, this is something that is a fairly new thing, ros is a

tyrosine kinase member. rearrangements are similar to alk, present in a smaller number of patients. 1.7%, less than 2% of all people with nonsmall cell lung cancer. multiple different partners but crizotinib is extremely effective agent with this

particular rearrangement, with a very long median duration of response. so this is the small subgroup. this is actually the good news and the bad news, depends on how you look at it, that we're finding these people who have these driver mutations, that can

be targeted with specific drugs, but as you can see small subgroups, now we're starting to really have to do a lot of looking when somebody's first identified with nonsmall cell to see whether they have the appropriate abnormalities that should be targeted.

those are the three main ones, two of which are approved, with targeted therapies, a third is ros1. her 2 is not as easy to target as the others, although there are there are data suggesting mutations in a small number of people you can have good

efficacy. raf is similar, braf are effective. ret, the fusion genes, and there are reports now of small numbers of cases. so we're making small incremental progress. squamos cell carcinoma, a lot

harder -- excuse me. a lot harder to target. and there are no approved therapies there. these two, fgfr receptor 1 amplification is frequently found. 22% of squamos cell carcinoma, the smokers, everything else was

namely never smokers, that small group of people. but there are drugs that target this abnormality that are in development, whether they will prove to be effective, you have many different abnormalities, more genetically complex tumors, that remains to be seen.

ddr2 is an oncogene found in a small number of squamos cells, at least in vitro there are drugs which are sensitive but i've not seen results in clinical studies. as you can imagine, we're talking about 4% of squamos cells, 20% of all nonsmall lung

cancer, you need multiple studies to do appropriate small trials. so the other thing that has really changed the whole landscape of treating lung cancer is immunotherapy. melanoma is the poster child for immunotherapy, and nonsmall cell

is following in the footsteps. so the checkpoint of inhibitors which -- to make it pathetically simple, immunotherapy, which essentially undo the blockade of t-cell activity in peripheral tissues. the check point inhibitors, pd-l 1 have been shown to be

effective in melanoma, the first drug is approved, and there are excellent data that nonsmall cell is close, close second, and i think the expectation is that the drugs, one drug, will be approved probably in the next few months for nonsmall cell. so there's the ligand as well as

receptor, pdl 1 is the ligand, the response rate is using antibodies or to the receptor, on the order of 50 to 25%. the pd-l 1 drugs, receptor drugs, seem to be a little bit better tolerated than -- i'm sorry, the ligand inhibitors are better tolerated.

and what's really exciting about this is not so much the 20% response rate because that's okay, but hardly a home run, but the fact that for some people who respond, that response goes on and on and on. i told you that the immunotherapies, the response

here if you allowed for error, more than a year, stable or minimal. so there is a lot of excitement that may be appropriate. so that's -- what did i just do? nothing. okay. so that's sort of the treatment

story in a very pathetic, small nutshell. i just want to -- you have to know that setting to understand why some of us spend our lives doing this. it's because even with the best treatments, you don't cure it. maybe you'll cure a small

subset. lung cancer is a death sentence. how do you reduce mortality? okay, from lung cancer? well, prevent. that's what we're going to talk about next. so this is that long phase from the insult that leads to the

initiation of the carcinogenic process, through the early basis when you get the progressive histologic and molecular abnormalities. so the best prevention, because 85% of nonsmall cell is caused by smoking, would be to never smoke, right?

is it as good to stop smoking as it is even today, maybe 22, 23% of of the population in the united states will continue to smoke. is it -- should we be telling people to stop smoking? obviously the answer is yes, but the data for decreasing cancer

incidents are not as striking as one would like. this is the lung health study, which was really copd study, where the interventions have to do with copd and with quitting smoking, and what you see here is after 14 -- well, almost 15 years of follow-up, people who

quit had about half the mortality of people who continued to smoke. people who quit intermittently had a decrease but it was not as profound as people who had stopped smoking. but the point that i want to make here is that there's still

a fair amount of lung cancer deaths. these are the 15th year follow-up data. if i would show you the graphs for five years after quitting, there's no change in the number of lung cancer deaths. so, in other words, you need to

have a long time to see a perceptible difference. as you continue to smoke, you continue to increase your risk. when you stop smoking, that damage stops, but you still have the risk that you already got, so you really need to go out many years.

in other words, the graphs sort of look like this. if you smoke, your risk goes up. if you stop smoking, it levels off. it doesn't go down, it levels yes, smoking cessation is very important. the sooner, the better.

never start smoking, even better. all these people who remain at risk, what can you do with them? that's where my day job comes in. cancer chemoprevention. that's going to be the bulk of what i talk about for the rest

of this lecture. so chemoprevention is a concept really popularized in the 70s, 40 years ago, that is defined as the use of natural synthetic agents, could even be immunotherapeutic interventions, to suppress or reverse the process of carcinogenesis.

we're not longer talking about be a anatomically defined cancer nodule, we're talking about the process that leads to that anatomically defined cancer nodule. it's regressing existing pre-neoplastic lesions, preventing the developing of

new, suppressing the recurrence, if there had been neoplastic lesions there previously. the rationale is straightforward. you can't cure metastatic i think i can say that honestly. it's probably for the rest of my career, unfortunately.

we do know that cancer is preventible from other model systems. for instance, breast cancer prevention with tamoxifen is well documented. multiple animal model systems show you can do this. obviously animals are not man

and are much less complex. also the knowledge that we have this long pre-clinical phase during which there are increasing abnormalities, therefore we can identify populations at risk, hopefully by demographic and biomarker profiles so that one can

actually start to prevent lung that's the rationale. okay? the execution has been a little bit less easy. so, again, a little bit more on the rationale when is the best time to intervene? obviously if you intervene this

100% efficacy with metastatic lung cancer, that's where you want it to be, but we believe, and maybe it's true, maybe it's not, i think it is, that early cancer or early pre-cancer is more amenable to interventions than late cancer. for instance, early stage lung

cancer is more curable than metastatic, maybe stage 1 we think precursor legions are less dramatically complex, maybe more amenable therapy, if you could prevent the dna damage or get rid of abnormal clones early on that would be the best. but the worry is that there are

actually multiple pathways to carcinogenesis. i showed you in never smokers there are many molecular you have one or the other, mutually exclusive. so there are probably multiple pathways of carcinogenesis which may require multiple ways of

dealing with this abnormal process. the other thing is that once you start to intervene you are going to get toxicity from intervention. aspirin is not completely safe. tylenol is not completely safe. it has a side effect.

and anything stronger than that is also going to have a side effect. so the risk benefit, toxicity profile, needs to be carefully looked at. the more people at risk, the larger the target population, so there's more than 90 million

current and former smokers in the united states right now. we're not going to target 90 million people just for the prevention of cancer. you can't do it, even if we knew how to do it. we need to figure out how to target the highest risk

subjects. of course, anything you do has a cost, resource-wise, psychologically, turning people into patients because of drugs, et cetera. and it's very important to weigh the risks and benefits. benefits in terms of efficacy to

prevent not just cancer but also cancer-associated morbidity, of course mortality. but then there's risks, side effects, that increase with the mortality, for instance some of you may be familiar with the vioxx story. vioxx was used for pain.

it's a nonsteroidal, until it was shown to increase cardiovascular disease, risk of myocardial infarction, shown to do that in a cancer prevention trial, not for lung cancer. effective in regressing polyps, also effective in causing mis, went right off the market.

major morbidities and mortalities are important in order to substitute one's disease for another, when you try to prevent. and then there's the minor morbidity and tolerability. soming from the oncology point of view, if you tell me that i

give you drugs, let's see if i have bouts of diarrhea, i can deal with that, but are you going to take the drug for years if it's going to give you diarrhea every day? i don't think i would. so tolerability becomes very important when you're talking

about prevention in the setting of people thinking they are healthy, essentially feeling healthy, but it's not as much of an issue for cancer treatment. so how do we go about that, preventing or finding these agents for cancer prevention? knowledge of mechanism, okay?

which actually i submit to you we don't really know the mechanisms of early lung cancer development very well. so we often go to pre-clinical data, in vitro and animal models, and we expose those models for what could happen to people, although as i said a

mouse is not a man. so the translation from animal models to human beings can be difficult and is often inexact. we go to the literature, cohort and case-control studies, some is cases have been very effective in identifying effective agents, not so much

for lung cancer, nothing jumps out that easily. of course we'll look at the clinical trials. for instance, drugs used with adjuvant treatment to see the secondary primary cancer, the target organs prevented, for lung cancer so far adjuvant

chemotherapy is not going to be something you can use for prevention. i'm going to tell you three short stories of how we've tried to develop a lung cancer preventive agent, what kind of data, and what we did. this is a work in evolution

still. so one is the inflammation story, steroids. lots of data in the literature, going back to the '70s, showing that immune suppression is effective, or are effective, in animal models. skin cancer animal models, lung

cancer animal models, and you can even use inhaled to prevent cancer, also the work of one of the grand daddies of the field. because they are used for copd or asthma, you can test that in humans. epidemiology for this kind of approach was plus-minus, i would

say most of the studies with asthma or copd were of short duration, although there is one study that does keep using veterans administration cohort showing people with copd use inhaled steroids had a much decreased risk of lung cancer compared to people who did not.

it is what it is. it's a problematic study, but it is what it is. so just to show how we do some of these studies, this is one such example where animals were given biocarbonate, in mice you could count the lesions. what you can see here, it's not

a steroid, in the diet get an 80% decrease in the number of tumors, and the tumors are adenoma primarily as opposed to carcinoma, you shift to a more benign cell. what you're doing is delaying the conversion to carcinoma. so data such as this that caused

us to work with the british columbia cancer agency in canada to do what i would consider sort of a classic standard phase 2 lung cancer lung dysplasia trial, so what stephen did, he screened about a thousand people, those who are the abnormalities, atypia were

invited to undergo a broncoscopey, targeted biopsies, and identified about 560 people who underwent broncoscopey, 112, had dysplasia, agreed to go on, randomized and reached inbudesonid or placebo and underwent multiple biopsies of

the sites biopsies previously and any new sites and underwent helico ct to see if there are nodules in the lungs. to make a long story short, this was clearly a huge amount of work so kudos for doing it, out of this 1040 screened, 13 did not go on.

when all was said and done after six months, the complete response rate were about the same, about 30% of people had progression of all the lesions, about 45 to 50% had progression to higher grades of abnormalities of new lesions, no difference between the two

groups. so it is what it is. what was interesting though was ct-detected lung nodules, there was a statistical significant decrease in the number of nodules. now, the animal data i showed you looked at adenoma, and

adenocarcinoma. this was looking at the central airways where squamos cell appeared. this was the only way they knew there was a little bit of a mismatch between the animal and human data. so maybe all is correct.

so the next study that we elected to do with the ct screening entered the therapeutic landscape, imaging landscape, was to actually ask, well, given these data with the progressions with budesonide if you only look at those with peripheral nodules, presumably

adenocarcinoma precursors, can you find an effect? this is a study performed at the european institute of oncology which had and has a ct screening program. and they took smokers with persistent ct-dependent lung nodules, any nodule not felt to

be cancer. randomized for a year of inhaled budesonide and placebo, and the primary end point was shrinkage of nodules. i won't pretend to say this was the perfect study. what are the problems? well, we see what the nodules

are but you don't know what they are. they are too small to biopsy. they need to be biopsied, those people were excluded. so what we found was that overall there was no response. if you start to look at different types of nodules,

there was a differential response, secondary not pre-planned analysis. what we found was the solid nodule, the small, 25% of people have these, smokers, those did not change at all. but the nonsolids, i'll show you a picture in a minute, there we

started to see a difference with budesonide versus placebo. we followed for one year, followed the nonsolid nodules over the next five years, continued screening protocol, the budesonide treated people go with a gradual decrease and a significant difference in nodule

size. this is only in the nonsolid now, what does this mean? what are these nonsolid modules? this is what they look like. they are not solid, okay? this is solid. this is something that you feel like you could see through it,

and where some of these lesions are is this. this is a typical adenomatous hyperplasia. you have the cubeoidal cells, as opposed to the type 1 cells that cover most of this. and in various studies that looked at removed lesions, these

are primarily big lesions, more than a centimeter, aah is found at 25 to 50% of ground glass opacities. some of these are the precursors of lung adenocarcinoma, it has been hypothesized. one study shows especially in smokers the path these nonsolid

nodules, they tend to grow over time. and i'll show you one example, who is actually one of our patients, not part of the screening study. this is in our clinic. but somebody, a woman, middle-aged woman found to have

this nodule when she went to the e.r. for chest pain. they did a ct and found this. and the follow-up, three months later is still there. she did go for additional cat scans. six years later, it's still there.

it's a little bit bigger now but it's still nonsolid. a year later, here it is, a little bit bigger yet, and now solid. and when that was taken out, basic adenocarcinoma. this is not proof but this is in addition to the adenocarcinoma,

there were areas of adjacent aah. so this suggests to us some of these lesions -- it takes a long we never looked at those lesions before, ct, so we don't know what's actually been going on in people's lungs, but i think that some of these lesions are what

we need to pay attention to. that's what we're now targeting. i'll show you a little bit more data. this is from the national lung screening trial, ct screening study, where we asked the question -- i'll show you its outcomes, decreased death from

lung cancer, we asked for people with these nonsolid nodules. are they at increased risk of developing cancer? it so this study was done in such a way that people had three screenings, three cts, at yearly intervals.

we went back and asked five years later how many people had lung cancer, we can tell did they correlate with the lesions they had, and for ground glass nonsolid nodules, the risk of cancer went way up at later points, early on it was actually down.

those are noncancer, at the time you first see them. but years later, people who have those nodules had a significantly increased risk of lung cancer and it was in the lobe the original nodule was in. that's the best data that we could come up with to follow oh

the natural history. so this again suggesting that ground glass nodules have increased risk of turning into lung cancer, some of them are actual precursors. so how do we move forward? what do we do? we now think that the way to

look at peripheral lung adenocarcinoma prevention is to focus on this, the new studies. what drugs do we want to use? well, we're still in the antiinflammatory mindset, okay? and instead of going to the inhaled steroids where there's some question as to how well

they get out to the periphery, they are optimized really more for asthma. i'd like to bring to your attention the study by peter rothwell, a series of am sis looking at aspirin and cancer mortality. what you see here is that people

who took aspirin versus those who did not, at least for intervention trials, that the risk of lung cancer deaths is about 30 to 40% decreased, but it's primarily later on. so you don't see the effect until about five years from the time they started.

but this is the effect on death. it's adenocarcinoma only. not a squamos cell. so our thinking is that five years later, your precursors, your invasive lung cancer, would have occurred somewhere around here. your precursor lesions would

have been present here, and the aspirin people prevent development, so the study ongoing at the end with our colleagues is to take smokers who are undergoing ct screening, who have persistent -- at this time only nonsolid nodules, nonsolid or part-solid nodules,

the lesions we think are enriched for the atypical adenoma hyperplasia to see whether aspirin will change the progression of cancer or regression. this is just started. it will probably take us about two years to develop.

i'll tell you a second story, these will be much quicker. so myo-inositol, a source of several second messengers, signaling molecules found in rice and various other things, a fairly long history in small studies in the literature. found to be quite effective in

animal studies, again the work of rossberg, inhibited carcinogenesis in smoking induced and carcinogen induced models, and importantly this was a drug that is glass, regarded as safe by fda terminology, you could just do it. he don't need permission.

what made this intolerable? 18 grams a day gave some level of diarrhea. what he found was that there was a decrease in the regression of dysplasia. 50% of people with res, i showed you that data. here only everybody has

expression of dysplasia. we were lucky to work with boston university who took samples that stephen had been collecting and started to ask, what are the pathways that are abnormal in people who have dysplasia? he did it not by looking at the

lesions, by looking at the epithelia, normal epithelia. he found the pathway is deregulated adjacent, and also in people who have adjacent areas of dysplasia. in other words the precursor. and you can inhibit this signal. so providing a possible route

forward as to how we test, by looking at global gene expression analysis, looking at signatures associated, specific pathways, so this gives us the potential for identifying people at risk because they have p13k signaling as well as dysplasia which we may not pick up as

easily, allows us to look at things. more obviously this needs testing, everybody underwent this profile, this gene expression signature from normal bronchial epithelium in addition to having the effect of myo-inositol tested on

we're finishing up the study now, and looking to see how well the gene expression and whether myo-inositol is effective. next year we'll have the data. last story, something a little bit different. drug called peroxisome proliferator-activated receptor,

target for diabetic drugs. it is used for type 2 diabetes. and there's a lot of data, pre-clinical data, that ligands, which are used in diabetes, induce growth arrest and differentiation of cell types, there are animal data for not small cell.

we chose to first go to the head and neck cancer model, precursor to oral cancer, this yellowish thing. the reason we went to this is shared etiology, in the lung, you can get to the lesions easier, biopsy them with these in situ and ask does the drug

work? so this was a phase 2 trial, university of minnesota, 22 people were treated for three months, the lesions were measured again at and biopsied again. the response was 80% shrinkage of lesions, although there was

not a -- 30% decrease in the level of dysplasia. what stayed was same level of so the phase 2 trial, which is 100 persons, is now again being looked at, although it was not 100 people so the ability to test i think is going to be limited.

but these are all preliminary data that are leading us to think should we look at things that are more difficult. these are animal models, tumors, two studies, we're looking at the effect on the normal field as well as tumors, markers within the tumors, just to get a

little bit more data before going to the trial. the last couple minutes, three minutes, five minutes, house keeping, due respect to early detection, where we made the most headway. let me just do that. so like everything else,

screening is not straightforward, okay? you would think that if you find a cancer on screening modality, you would have done something good. lead time bias, in other words by looking, you diagnose earlier.

you don't actually postpone death. you just pick it up earlier. length bias, you're only diagnosing -- the going to happen at any rate in between the screening intervals, so you're only detecting the less aggressive disease with better

prognosis. and overdiagnosis, identifying lesions that are actually unimportant. the person dies, they are going to die from it, whether they knew they had the cancer or not, particularly true with prostate cancer, autopsy studies shows

the majority of elderly men have prostate cancer. they don't from prostate cancer. that's overdiagnosis. so these are things you need to be aware of, and that's why cancer screening studies really need to be done very carefully and in a randomized obviously

not placebo controlled fashion. this is the plco chest x-ray randomized trial. that's in contrast to ct screening, much more sensitive, where the national lung screening trial, a randomized

trial, 53, 000-plus current and former smokers. and showed a 20% risk of lung cancer deaths. if you show decrease in the risk of death, there was some overdiagnosis there, not just in lung cancer deaths but all cause

mortality shows that lung cancer mortality is a significant driver of all cause deaths. actually ct screening is recommended by the u.s. preventive services task force, and medicare won't cover it, with some caveat, that's the latest data.

this is just a visual showing the cumulative number of lung cancer cases, and the number of deaths declined more with ct. how do we put this to use in cancer prevention? maybe that's the best way to move forward.

well, whether you find the lung cancer cases, you find the nodules, it's going to be those ground glass capacities, those will actually be more further along the road to cancer. the british columbia cancer agency developed a model that can reasonably identify people

most likely to develop cancer in the next two years, and now this gives us this validated tool for identifying people for highest risk we should have moving to cancer prevention trials. so to summarize, huge progress has been made, understanding the process of lung cancer genesis,

precision medicine is applicable to significant but small subset of advanced stage patients. it's still the early days of immunotherapy but there is a lot of excitement with prolonged survival as a single modality. so we have made lots of progress in that capacity with early

detection. all of this together is helping us figure out how cancer i would be happy to answer questions if anybody has them. [applause] [low audio] >> those were not cancers. those were precursors, small

lesions that we saw on ct. it was simply the size, correct. right. that was not a transition to tumor. none of those. the rate of cancer was about 2% per year, so very small number of cases, not the difference

between the placebo and intervention. >> we're, we're pleased to have yves pommier, md and ph.d. from university of paris in france, at nih since 1981, chief of the developmental therapeutics branch, serves on many committees at nci, won

numerous awards, he's here to talk to us today about topoisomerase inhibitors in >> thank you. one of the indications of small cell lung cancer, used in first line therapy for the topoisomerase inhibitors. it's an interesting coincidence.

so it's not a simple topic. the topoisomerases are very complex, hopefully not that complex. it's not easy to get everything in one review, usually you have to do many, many different places because there are six genes, different types of drugs,

some are anticancer drugs. the review, we give you a synopsis of the whole thing where in one review i took the challenge to put take all the topoisomerases, the topoisomerase inhibitors, useful for and how they work. many of the slides come from the

book, it was out a couple years ago, where you could really find a lot of chapters if you're really interested on topoisomerase biology. and so the complexity is due to the fact that are actually in humans six genes, and they are divided in three groups.

they are what we call the top 1, top 2 gene and top3 gene. if you assume dna like any of these long wires you have entanglements. the topoisomerase would resolve these, you need a break to get one strand to move around the

other, you need to cut one to get the other one through it, by breaking the backbone. they are like gates into the dna. type 1 and 3, the other numbers break one strand at a time, type 2 even number break two strand at a time.

that's the way they were numbered, by order of appearance. so top 1, type 1b, type 1a. this is a top 1, top2, this is a dimer, top 1 and top3 are one strand. top there is 2 is two strands in

concert. if you think the enzyme -- it's like a can opener, it's going to open the backbone, and it forms on one side, this is what definds the topoisomerase. that's where the top 3 are different from the top1. the enzymes, top 1 nucleus, top

1 is is for the might condominiumry al genome. r y al gene -- mitochondrial genome. if you compare humans and e. coli, in humans you have this six genes. in e. coli it's simpler.

you only have two types, type 1a and type 2a. so what e. coli doesn't have is this one, type 1b. so why do we care about the different enzymes? it's because top 1 is the anticancer target of cancer camptothecins.

this is a very big spectrum in terms of pharmacology and medicine. the two enzymes are critical. here you see the whole. the type 1a there is no drug targeting type 1a, neither in vector here or in humans. and this is probably something

that the pharmaceutical industry is waiting to go into antibiotics is clearly one opportunity. to go more into detail about opening the dna backbone, this is a reparation of the topoisomerase, sugar here and one phase, another phase,

topoisomerase we approach the problem cleaving dna, we use a tyrosine. and the tyrosine then would act as the nucleophile, this is why the topoisomerases are tyrosyl-using enzymes. the way that we cleave those, the polarity, if you wish to go

that, would have been on the type of the enzyme related to the fact top 1 we link as it cleaves, it will link to this. so all the type 1 has this characteristic. the other topoisomerase, top 2 and top 3, link to this. in bacteria there's no such

thing for many bacteria such as top 1. the linkage is from the family and it is possible the top 1 used in vertebrates is actually something that derived from the recombinant enzyme, or a combination of enzymes, but in humans top 1s are used as dna

relaxation. let's discuss a little bit about top 1. top 1 was discovered as the first topoisomerase probably about 40 years ago, by using an extract from murine aclls.ur ats. the te, it was possible to look at dna, one can

differential rate from relaxed dna and what was discovered was that by adding a drug of similar extract, the dna became fully relaxed within minutes. we call this dna untwisting enzyme, it would untwist the dna and the way this is going is by introducing the transient strand

break which enable the the swivelling of the broken strand around the strand to remove the supercoils and topoisomerase 1 is therefore essential for any dna transaction. you can well imagine if you have a duplex dna and if you open it up, what will happen because of

the helical structure, you will overwind the dna on one side of the bubble and if i have another side here to actually unwind on the other side, and for the transcription and to go on, you need to remove or you could not separate. it would be all bundled.

knocking out top 1 in mice or in flies is lethal, there's no embryo. in yeast it's permissible because there's competition by the other topoisomerases, but this is an aging phenotype. so another way to think of this, what i just told you, is that as

transcription machinery, as remodeling takes place, you have to separate the two strands, which as a consequence will be ahead of the unwinding, dna is not free to swivel very long and this needs to be resolved, the resolution of the super coil is carried out by top 1.

and then dna becomes relaxed, as a result of which then everything can go on. so the closing reaction shown here, provides a swivel point for this strand to move around. when it's fully relaxed, realigns and top 1 is eliminated.

it's a reversible reaction. that's what's shown here. after the dna is realigned this attacks back and then it reverses. so it's so perfect, as long as topoisomerase don't get stuck in the middle of their activity. in reality they do get stuck.

they get stuck in different conditions. think get stuck if you introduce a drug such as an anticancer drug, if the dna itself is not perfect, so it misaligns and doesn't realign properly, so based on introducing cleavage complexes, top 1 cleavage

complexes, and then in addition to that this is another process that's not been fully understood, it's probably a useful thing for the cells to mark all these lesions on dna and put top 1 on it to go to full apoptosis. so the two human top 1s are

nuclear top 1, the first discovered, the dnn twisting enzyme, the untwisting enzyme is made of three parts. this is the head of the molecule which contains the signal, and then all of this is the machinery. the tyrosine at the c terminus

here. in 2000, when there were only 5 topoisomerase in humans, we decided to look for another one and we discovered the second top 1, and when we looked at it, we realized if you compare the nuclear versus the new top 1 we had identified, that they were

very, very conserved in all the catalytic motifs, the main difference was the n terminus, and the n terminus contains a very specific mitochondrial sequence. it's a duplicating of the ancestral gene in vertebrates, are a nuclear top 1 which we

name top 1 and then to avoid having to change the whole nomenclature we call that top1mt. you can see in some ways it's somewhat similar to the virus that encodes its own top 1, it has to have a top 1 to replicate, you can see the gene

is quite similar to the stripped down version. this is again the mechanism, so if you assume you will have this swivel, you will relax dna and get full relaxation. many years before top 1 was actually became known, the nci had a very active drug discovery

effort. the idea or the principle was that extracts from plants from all over the planet were put into the leukemia model in mice and looked for activity, one such activity was found potent from the bark of the chinese tree campthodecin.

this was purified, tornado identified and named camptothecin, in ehe 1970s it went into clinical trial. it provided some responses to patients but was deemed difficult to use and toxic and nobody knew how it worked, the drug was shelf and put aside,

the single trials were terminated in 1975. in 1990, it was discovered as a specific top 1 inhibitor and drug companies went back. knowing that it was very nonwater soluble, now they are used and fda approved, it was developed in the

united states and irinothecan was developed. it is used for lung cancer, colon cancer. what was interesting, when this came about, the fact that they were inhibiting top 1, the mechanism was -- the way it works, this is the chemical

structure, a small molecule, quasiplanar. when you take purified topoisomerase 1 we can close the dna, you see very also knitting. you have mostly this, very little of that. when you put it in a purified system, you see a lot of the

cleavage complexities, which led to the hypothesis it was binding at the interface of the enzyme and the dna in the cleavage site, which we then called interfacial inhibition. they blocked the complex. it's not by competing with anything.

they distort. they take advantage of a cavity, come in and prevent. this was confirmed by crystal structure, and the structure of topoisomerase in top 1 dna complex is shown here. so this is the surface presentation, top 1 in gray, if

you take the surface away you can see there the drug, the break in the dna backbone, that's exactly what's drawn the drug is just bound at interface of the enzyme of the this binding is very specific to it doesn't bind into top 2 cleavage complexes, so nature

has selected over the evolution this particular chemical to be a toxin produced in a number of trees, probably to kill insects that are feeding. so we also found that we developed over the years, we have exactly the same type of crystal structure.

these are clinical trials here, they act in the same way. they bind, you see the broken strand here, the tyrosine and the drug, it's a perfect sandwich and makes a number of interactions with the polypeptide protein. now we have a crystal structure

for all top 1 inhibitors, they all work the same way, they trap the cleavage complex, normally it's reversible. you put the drug in, it traps it, i cannot reverse. you may wonder how does the plant actually -- how do the plants manage to produce

camptothecin without killing themselves? it has the same top 1. a group in japan sequenced the plant that produced camptothecin, they found all the plants have a mutation, next to the catalytic tyrosine, it's a mutation n722s, published some

time ago. what was most remarkable when they published this, then they wrote me an e-mail saying that it was exactly the same mutation that they had observed in a human leukemia cell line that has been generated by selection by exposure to camptothecin.

the single amino acid residue change confers total resistance, totally targeted by definition. that's how the plants actually managed to grow. so in about ten years ago, now even more i would say, we wondered whether we should develop new top 1 inhibitors,

and the rationales were simple. we knew they were effective as anticancer drugs, therefore top 1 was validatedded. we also new something else, even if you take a molecular target, you could have two drugs that have exactly the same effect on the target in cells or in

animals or humans, the pharmacological effect would be different. one example is colchicine, you could say it equals vinblastine but you will not treat a leukemia patient with colchicine, and you will not treat with vinblastine.

if you produce another drug it's likely to have a different clinical profile. and there was also a more rational idea, camptothecin, first trials were not well tolerated, doctors are difficulties with toxicity, hard to overcome.

bone marrow toxicity, intestinal toxicity almost killed the drug. drug efflux substrates, and thorough toxic by nature but in patients they immediately inactivate. this is the active form. as soon as you put camptothecin, within minutes, it wings open.

only a small fraction is active at any time, it varies across people but it's a few percent. you're getting a lot of drugs. that was a real problem. we decided to look, also because it reverses, they are very nice, you treat cells, they immediately form cleavage, wash

the cells away, cleavage complexes reverse in minutes, you could do great experiments but in patients you have to have more exposure. the nci, 60 cell lines are used to screen drugs, when the model was put aside, because it was good for leukemia, it was

decided to screen for drugs using cell base models. a pioneer enterprise, the cells are derived, and you can see if you take any given compounds through the bars and if a cell line is in the middle it means it's average sensitivity.

if it's on the left it means it's a cell line which is hyper resistant. if it's on the right, it means it's more sensitive than the average. what you see now is sort of a key, you've generated a key. using that key, we use this,

which was designed by our colleague, we look for the half a million compounding the database, anything that looks like campotothecin. this is the chemical structure. we found this molecule deposited by a colleague of mine mark kushman.

it was a by-product of something else he was trying to synthesize, it was not what he wanted but he put it in the database, ten years data, we found it and i called mark and we made a lot of effort to improve the activity. this is the initiating compound.

in 1998 was the first discovery, and it took ten years to go and screen about 500 of them. there's three that are in clinical trials, all any here, and here, finishing phase 1 for the last two, and this one is also in clinical trials with the two others all over the united

states, in veterinarian, and we're seeing quite good activity with them, with reasonable toxicity. so they are tolerated, we killed no one, fortunately in the phase 1 clinical trial and we see activity. what's going on in this stage,

beyond the cancer testing, the indenoisoquinoline could be viewed as second generation top 1 inhibitor, there's another drug, so the indenoisoquinoline, as i said, they are selected, they are important. they are chemically stable. these molecules have no outside

drug, we believe the clinical spectrum of activity should be different. what we had to do for phase 1 nowadays is difficult in phase 1 with just the path of determining maximum dose. phase 1 you have to have some kind of biomarker telling you

the drug has activity somewhere. we decided to take two biomarkers, h2ax and top 1. so the reason we chose h2ax, you're probably aware, everybody has heard of it, gamma h2x was discovered by a colleague in the same lab, we're in the same lab, william boner discovered h2ax,

the modification which is the phosphorylation of h2ax which he called gamma-h2ax and now it's taken for granted, we assume it's always been there but it's not been always there. with bill boner and james doroshow, also in our branch, we decided to validate hdax as a

biomarker, it's easier to say, much harder to do, go back to murine models to make sure it could be used as a biomarker in patients when we do the phase 1, and we had to go back to the reference drug. these are examples of mice that

have been treated with the indenoisoquinoline, the gamma h2x here, these are untreated, when you treat with top 1 inhibitor, the indenoisoquinoline, very strong response, which you could quantify, modified here, and under these conditions what was

nice is the mice had tumor this is the tumor response in the untreated mice. this is the response with the indenoisoquinoline treated mice. very large. and here are the responses with the mice, and they lose 20% of

their weight. here they are okay. we felt we had a better drug, and the dose and the time here, we had to make sure we could generate gamma h2x signal apreviousated here, and you could see we could generate these at about a third of the

maximum tolerated dose, a fifth of the dose, and the time was important because if you do biopsy after injection in the patient, when do you do it? you can't do it much for time. if you do it in the tumor. and we realize you had to wait four or seven hours after the

one hour infusion to actually rebiopsy the patient to look at it. that was generated top 1. when the cell -- when the topo inhibitors are acting, it disappears. top 1 gets degraded. here a function of

concentration. this is visible also at four or seven hours. so we have two biomarkers in the phase 1 trial. so these drugs now are moving along, hopefully we'll keep moving. they are potent.

they have specific inhibitors of top 1, both in vitro. they can be used in biochemical assays and in cells. they are chemically stable. the cleavage complexes are persistent. they are antitumor, less toxic

than camptothecin. i'll move on to top 2. what i've shown you now is that top 1 is one strand break. top 1 is linked to this. in the case of top 2, these are more complicated. each monomer will cleave one strand.

the two subunits are cleaving. now what you're generating is two breaks. now, the difference between top 1 and top 2 are many. they are structuralally unrelated. the only thing in common -- top

2 has to burn atp to carry out the cleavage. top 2 also requires magnesium, top 1 doesn't. it can work at zero degrees. top 1 does not. etoposide, they don't cross over. and quinolones for top 2.

so there are two forms of top 2 in humans, and all vertebrates. one is called alpha, the other one is called beta. alpha was the first discovered, beta was discovered later. two different genes. you could see they are very, very conserved.

the degrees of similarities are very high. the c terminus and n terminus, you can see they are somewhat related to the top 2 rate, in e. coli if the genes are made, you have a and b, so it is made and these enzymes also have a

high degree of conservation and use methyls to coordinate with the dna. this is the dna backbone, these are the residues of the enzymes highly conserved, that's how they are formed using the magnesium, to coordinate the dna with the enzyme.

now, these two enzymes in humans have different expression, different function. top 2 are found as associated with reputation. it's essentially highly expressed in cancer cells. breast cancers, for example, overexpress.

top 2 beta is expressed in all cells, including proliferating if you take the heart, for example, it has top 2 beta but does not express. it has the gene but does not express. the brain does not express. so these are division of

functions. they are two different genes. the top 2 has a broad range of reaction. to to describe, it acts like a gate. top 2 will only let you enter, or let let you if you are a strand of dna, if you go through

the gate, it closes back and goes through the second. let's look at -- you have two strands of dna, i'm trying to get that strand to go across that one, okay? so what top 2 will do, it will open the first gate, top 2 is is a dimer.

it will bind one strand of dna. so that would be the strand, and one strand at the top. just the strand in the top, you've gotten in the first part, you're inside the enzyme now with the other strand. the strand, passing strand goes through, and uses atm magnesium

and goes to the other end and this is actually opens up, it's a totally safe reaction, two-gate mechanism. open up, get one strand in, close, the strand is in, re-open, the strand is out. so it's a strand passage, what we call a strand passage.

as a result of this it's not a very fast reaction, this is very effective for a lot of reaction. when you replicate two circles, you are going to end up with this. top 2 is absolutely required to do this. top 1 cannot do that.

you have to open two strands up. this is what we call -- it can be reversed. it will deplete and let one strand go. one enzyme can even generate. it will generate. in invertebrates there's no such thing as a gyrate.

so the drug that acts on top 2 are many. there are two classes. for your topo in your cell, it's the anticancer drugs. the anticancer drugs, the most specific top 2 inhibitor, if you want to use it as a tool or reagent in your

experiment, is a topocide. doxorubicin, therefore they have many other effects. they also generate oxygen radicals so therefore these are top 2 are interpolators. the antibiotics are here, some

are well known, and the newer quinolone, and those are only targeted to the bacterial top 2 gyrase. the drugs are so selective for the enzymes. the way the top 2 inhibitors block the top 2 or trap the top 2 cleavage complexes are similar

in fact to what happens with top 1. so the paradigm that was first developed for top 1 inhibitor has been extended to top 2 inhibitors. this is the structure of the cleavage complex, this is the molecule, etoposide, drawn with

3d, and this is the crystal structure of the homodime. you can see the monomer and the other monomer, the drug is right in here, you can see by transparency the broken -- the dna that's broken by the top 2 is in here. now you can see better the dna,

the drug molecules inside the cleavage site, just inside. they bind at the interphase. you can see the interphase. like the top 1 inhibitor this type of agent is interfacial inhibitors such as you saw in the natural products, but it's been extended, not only true for

top 2 inhibitor, also for the antibacterial. so this is the structure of the bacterial top 4 and this is made of four parts, but the four parts came out similar to the two parts of the human top 2, you can see each symmetry, the dimer here, dimer there, and

quinolone blocks the bacterial top 2 very much the same way, so the principle of inhibition extends to the topo inhibitors. if you want to read more, this is not limited to topo inhibition. this is a paradigm that extends to many, many natural products

and even hiv -- anti-hiv drugs. then the question, we'll finish on this, why in the end these drugs would be of any interest for cancer, knowing that topoisomerase are present. they could be more sensitive because have more topoisomerase, the drugs would make more dna

damage but that's not that satisfactory. so in fact, what you would say that the use of top targeted drugs would be for patients with high top level, but especially in defective dna repair because what's really happening is that cancer arises, benefitting from

dna repair deficiencies that enable adaptation and these are an achilles heel. if you have that, the cells become more sensitive to dna damaging agent, suggesting top 1 and top 2. you could see here the pathways that confer particularly

sensitivity to top 1 inhibitor and top 2 inhibitor, and this is a panel of the different cell line, published this year, but you can see all these cells are single knockout for brca genes, much more sensitive than wild-type. and other cells are very

sensitive to top 1 and the top 2 side you can see the sensitivity here, the brca cell, so there are a number of predisposing diseases, cancer related. so the way this complex is repaired is actually better known nowadays. if you have a top 1 cleavage

complex, you generate a complex here, the tyrosine linked and this is a massive lesion, if you have a top 2 cleavage complex, linked, and what has emerged over the last ten years, cells going down to this, that enzyme discovered here at nih removes the tyrosyl-dna, a surgical

enzyme removes it and enables cells to repair. the fact it's conserved telling us it's formed spontaneously and more recently a second identified in humans and most vertebrates, tdp2, this is for the repair of top 2 complex. the cells use nuclease, and you

see cells such as mre 11, xpg, and it could tell you right away if a cancer cell is deficient in this pathway, it's absolutely dependent on that pathway because of the redonancy, so it means that the achilles heel of cancer cell may be they are defective somewhere in these

pathways, that's what renders them more susceptible to topoisomerase inhibitors, that's we would like to know in patients, which of these genes. this is another way to draw it. yeah, this is the same. and i'll take any questions, if you have any at this time.

and i'll be happy to answer e-mail if you have a question. >> i think in this stage, what we have to look at is sequence and find which tumors have mutations, so it's not like we have huge, so mutation, expression. tdp1, unexpected, we found two

lung cancers, and i've looked at other samples, other lung cancer cells, we need to know when they are turned off and if they are then presumably these would be. [low audio] other clinical trials, so they convolute these. it's a huge issue, for a drug to

exhibit activity on phase 1 is extremely challenging. you could go back and resequence and find out why the candidate gene, is there anything that really speaks. you have to have first the acute responder in phase 2. if you go to phase 2, especially

at nih, research institution, you have to have that. if you have responders, you can't crack it. even if you have all the genomics, you may not crack it but at least you have the tools. thank you.