Blood Test To Detect Ovarian Cancer

>> okay, we're going to get started. and if you want to sign up for the course, this is the last day. as many of you know, there's a potential government shut down tomorrow, so we won't be assigning any course until the

government is fully in action. and also, if there is a shut own and it lags on to next monday, if the government is closed, traco is closed. so we hope if there is a shut down it's a very short one. our speaker today is christina annunziata, she got a degree

from georgetown and came to nih and worked with -- in multiple myeloma. and now a principal investigator working on ovarian cancer. the title of her talk, ovarian cancer in the genomic sphere. >> thank you very much. do you put up my slides or do i

put up my own slides. you will get them, okay. well thank you for having me give this talk. i've done this for a couple years now but i kind of update the slides every year. so we should see a little bit more.

i used to talk about angiogenesis in ovarian cancer but now i'm talking about genomic which is a little bit more modern although angiogenesis is still a very interesting target. i was first going to start talking about ovarian cancer

from a clinical standpoint so we can sort of have some background. are we up yet? i'm not pushing anything. okay, here we go. okay, great. so, ovarian cancer can be defined as four stages as most

cancers can be. stage one of ovarian cancer is confined to the ovaries. stage two is confined to the pelvis and stage three is spread beyond the pelvis into the peritoneum or the intraperitoneal nodes and stage four is the solid metastasis.

so as you can see, should i use this pointer here? most of the ovarian cancer is diagnosed in a late stage of disease. and so that's why the survival is so poor in ovarian cancer. if ovarian cancer can be detected in early stage,

survival would be much better. and this is sort of where we get diagnosed in breast cancer, most of the breast cancers are stage one ask stage two as opposed to ovarian cancer which is more diagnosed as stage three and stage four disease when the survival is much worse.

the prognosticrs otherthan the stage for ovarian cancer is the extent of cytoreduction and that's the amount that can be taken out during surgery and we can talk about that in a little more depth later. the histology of the ovarian cancer, it can be mostly

searous -- tends to have a worse prognosis. and then the grade when means more prolific activity, undifferentiated morphology of the cells that also portend the worst prognosis. the performance status is a measure of how well the patient

is doing. so if the woman with ovarian cancer is in poor health at the time of diagnosis, she'll have a worst outcome as opposed to a woman who is in good health. p53 status although this may not be an independent predictor because it's very closely tied

to the histology. vital organ function is related to -- and physiological age. also we're going to talk about platinum and taxing is primary therapies if people can get all six psych owls of their primary therapy this is better outkut come.

intraperitoneal therapy we'll talk about and the rca and veg f production we'll talk about as well. what is the treatment for newly diagnosed ovarian cancer? the first step of course is complete surgical staging. the next is optimal reduction of

the tumors based on surgery. the next would be chemotherapy and then of course clinical trials. so complete surgical staging. this involves full assessment of the abdomen with random biopsies even of visually negative areas. and that's to make sure there's

no microscopic disease in areas outside the pelvis or in abdominal organs. lymph nodes dissection done in stage one but of course it cannot be in stage one if the random sampling has not been done. so where does ovarian cancer go.

you can imagine it starts first in the ovary here. of course there's newer literature that says it could start that distal fallopian tube. let's just say for purposes of this picture that it starts in the ovary.

where does it go next? so the ovarian cancer does not spread primarily by lymph nodes or blood such as in breast cancer or other solid tumors. the first spread of ovarian cancer is through the ascites which is the peritoneal fluid. so it's the fluid that's

circulating and bathing the abdominal organ. the first place it goes it kind of releases off the surface of the ovaries as you posed to going through the blood semen and invading that way. the way it goes is that it goes into the peritoneal fluid and

circulates in this sort of clockwise direction because that's the typical way of circulating the peritoneal fluid. you can imagine if the cells float off of the ovary, they're going to go here and the first place they'll stick is on the

top or the outside of the liver or stick in between the liver and the diaphragm. so is it stuck right on the outside of there and not invading, it's considered a stage three because it has not invaded the liver, still sitting on the outside of the peritoneum

which is the tissue that coats, that covers the abdominal organs. it has not invaded. if it were to invade the liver then it would be a stage four. so once this surgery is then, once the patient has been staged then the next step in that same

operation to perform as much cytoreduction as possible. in stage one and stage two of course the goal is to remove all of the disease. in stage three and four, the less disease you leave, the better. and of course what's considered

optimal nowadays is less than one centimeter but that's probably going to change and be actually no visible disease. that's based on that. so this is an historical graph of survival just based on the type, the amount of surgical cytoreduction that could have

been done. so you can see here if patients were, and this is a very old, this is very old, this is like from 20 years ago. if patients were left with two centimeters of the disease, greater than two centimeters of disease their outcome is much

worse than if they were left with no visible disease at the time of finishing up the surgery. if it was zero to one, that was you know, better than if there's two but still not as good as zero. so as of now, optimal

cytoreduction is considered less than one centimeter so it would be kind of like on this half as opposed to this half. but like i said that's changing because surgeons are becoming more and more aggressive and we're getting better at getting zero centimeters left.

this is kind of tricky because here you can have, are it's one centimeter not total disease left, it's one, like the largest blob left is one centimeter. so if you have a seat of tumors on your peritoneum, that's still considered optimal so that's why the survival is so much worse.

than if you had nothing left. now, this is also a question of is it the type of, is it the biology of the tumor that's different. so this is also up for discussion. if you can get everything, is that sort of a different

biology. for example if it grows in individual nodules, is that a different type of tumor than if it's growing in sheaths and you can't get everything. that's up for discussion. but just to make the point, if you can leave nothing, you have

a better prognosis. so, after surgery, the next step would be to perform, to give the patient chemotherapy, and the standard chemotherapy is a platinum and taxane and that's usually carbon owe platen or sic platen in twinnation with pacitax el or doce tax el.

the way you give it is a discussion and the trials going on now is how you do it. so the history of this, the survival of ovarian cancer actually was markedly improved with the addition of platinum in the 70's. before the 1970's, no one with

ovarian cancer lived five years. in the 1970's, platinum agents were introduced and we got 5% of the people to live for more than five years. then with the addition of the taxane, so paxy taxel was the first taxane and survival rate jumped to 35%.

that was a big improvement in survival here. since then we have gotten slightly better. this peritoneal is not a chemotherapy agent it's a different way to give the chemotherapy. they are given both iv and

directly into the abdomen right after surgery. that had a significant improvement in overall survival at five years. but not a huge jump like the addition of the taxane. other agents, chemotherapiations chemotherapy --s

have been campaigned. the thought is this kind of a key thing to notice here. the paci tax elf -- so all these other things in combination require dose reductions of the chemotherapy because they add to the toxicity of the platinum which is most commonly your

blood count's going down. so red blood cells, white blood cells and platelets going down. but for some reason the paci tax el -- so the doses of both the -- and the paci tax el are able to be maintained at full dose because of that. and the new standard of care is

the interperitoneal therapy based on this trial which was done by the gynecology oncology group. this is a group sponsored by the national cancer institute. this was a multicenter clinical trials testing. the standard therapy with

intravenous -- passe tax el given into therapy. this was given interperitoneally -- bioneerly 100%. however the -- so paclitaxel when you give it interperitoneally does not get into the circulation, it has to

be given iv as well. so you can see the total dose that you're giving the patient is actually higher in this arm. only 42% of patients were able to complete all six cycles of therapy. however there's still even with that caveat there was overall

improvement. this was a journal in 2006 where the median survival of the standard arm which was just intravenous was just 49 months. whereas the overall survival of the interperitoneal in the arm was over five years so it was 65.6 months.

that's a dramatic increase in overall survival. now when i say that we take this with a grain of salt is that this was actually only with stage three disease so not stage four. so it's not clear whether stage four, patients with stage four

disease will benefit from this but patients had to have optimal cytoreduction. so remember i showed you before, optimun means less than one centimeter of disease remaining. you must fall into that category of stage three optimal cytoreduction to actually have

qualified for this to be administered. the other caveat is that it is actually a very difficult regimen to tolerate. as i mention only 42% of the patients on the study completed all six cycles. so it's not very well received

by women with even the perfect amount of disease and the perfect scenario, it's still not very well received. so it's not been widely implemented as the new standard of care. even though this was shown to have a dramatic improvement in

the overall survival. so, where do we go from here. so what about genomics of ovarian cancer. so how can we improve on this five year median overall survival with interperitoneal this is the translational oncology course.

what is translational oncology. we start with an observation, get a hypothesis, can do an experiment or clinical trial, get a result, analyze the result or come up with a new hypothesis. so we're going to try to use genomics to go through this

scenario. in 2011, the cancer genome atlas published their comprehensive genomic analysis of ovarian cancer in nature. the background here is that the cancer genome atlas had 500, nearly 500 clinically annotated high grade serous.

high grade serous this is subtype ovarian canal. they identified molecular abnormalities which would influence pathophysiology, affect outcome and constitute therapeutic targets. this is the goal of the genomic analysis.

so what did they do? they looked at gene expression. so rna expression. they looked at micro rna exfretion. they looked at dna copy number and dna promoter and they do sequencing in 316 of those samples.

the criteria was they had to be newly diagnosed patients. they had to have serous subtypes of ovarian cancer. no prior treatment. they could have any surgical stage or grade. no, they had to be high grade. stage.

they had to have frozen tumor specimens with a normal companion normal. in some cases the normal companion specimens were abnormal -- although there were a few cases where they had normal ovarian serous epithelium from the non-involved ovary.

the clinical data actually was accessed so this is how long the patients have survived, whether they were resistent chemotherapy initially, whether they responded to chemotherapy and then recurred. also, the demographics of the patients, all this can be

accessed through the website, actually it's not this website name. it's cancer genome.gov is the website. they have demographics, they have the pathologic information, they have treatment details and outcomes, parameters.

so the very complex analysis because there were multiple centers involved with the cancer genome atlas and this probably applies to the other histologies of cancer, the other types of cancers included in the genome atlas as well. so first there was a preliminary

pathology reviewed to sort of find the cases. they started out with 1,020 cases. they did a pathology review and actually they took out 200 cases here, 200 cases failed the pathology review. they went through analysis of

the quality of the specimens and took out another 243 that were poor quality. and then they also collected the clinical data elements of these. and then they sent them to seven characterization centers. the data coordinating centers and three sequencing centers.

these were 518 cases here but then additional 46 which had been subsequently included are not included on the initial so it's a very, it was a very large effort on the part of the cans team atlas for ovarian cancer as well as for the other cancer.

so, what did we find? so they found, not we, i was not involved, with the ovarian cancer specifically with the sequencing, this is in the 316 samples that were sequenced, of course the highest mutation was, the highest rating mutation was found in the p53 gene.

so now based on this nearly all of them had p53 mutation or a loss of p53. high grade serous ovarian cancer are sort of thought to be kind of universally dependent on the p53 mutation. another 10%, almost 10% had a brca one or two, one or two

mutations. and then these other genes that were found to be mutated in the relatively high frequency. but still, i mean there's not a lot. there's nothing really that jumps out at you other than the p53 and the brca and maybe these

other ones in here. the rb and this one are in sequence. so there's not a high amount of frequency of mutations in high grade serous ovarian cancer. and we'll get to that in a second. which is right here.

so if you can see here, so in the glial blastoma, this is comparing the dna copy number here on this axis here are the individual chromosomes. blue is deleted. white is this neutral and red is this amplified. so you can see here in glial

blastoma which is actually the first tumor that was analyzed through the cancer genome atlas. in glial blastoma, there's very distinct areas of amplification, distinct areas of deletion. and then most of the other parts of the genome are relatively unaffected.

as in contrast to the ovarian cancer where pretty much the whole genome is affected. there's not really one area that's particularly white. everything is either blue or red. so this is showing you that there's just a high level of dna

copy number variation throughout the chromosomes, throughout the whole genome of serous ovarian there were molecular subtypes identified by jean profiling. this was a reference published in upal years before in 2008 where they had 245 tumors here are 489 tumors.

they used 1,000 genes to subgroup into four different subgroups based on sort of the genes that were the networks that were found in those subgroups. so one of the subgroups they call differentiated, one of them was immunoreactive, one was

parenchymal and the other was proliferative. they tried to say that okay we validated this in another related data set which is the 245 tumors in reference to the 25 that there was a similar break down of tumors in these four categories.

so what does that mean? so putting this together, i'm jumping a lot again, this is a very detailed paper published in nature so i'm really just skipping to the highlights here. but what does this mean? so they're saying well we have a high level of genomic chaos,

amplifications, overexpression, under expression, how can we put this altogether. so they tried to group it into pathways that were altered. so what pathways did they find? they found there were 67 merls percent of cases had some alterations in the rv signaling

pathway and this is bloodily defined as cyclin cdk and 2a. cyclin e which is ctne. cyclin d1 and d2 and rv. so you could see that all of these kind of aberrations would lead to cell cycle progression with these being down regulated the inhibitor down regulated and

the activator being up regulated. another 45% of cases had alterations in the -- ras pathways by things being deleted, amplified or mutated in this case with very rare mutations in v ras. notch signaling came out with

22% of cases or am -- amplifications or mutations, amplifications of genes in this pathway. the most common thing that was altered was probably the homologous recombination. so here is what they're calling homologous recombination

alterations. either brca1 germline mutation in purple. somatic mutation in green or epi genetic silencing meaning methylation of the gene in this blue and gray box. so in these cases, in 33% of the cases they're saying there might

be a defect in dna damaged repair. the dna damage repair pathway, and this was the homologous recombination pathway. i'm going to talk about this in a little more detail later but the thought was that 51% of the cases have this alteration in

this pathway. and actually the ones with the brca mutations had a better survival than ones with even -- epi genetic sequencing or wild type. this might get to the question we saw at the very beginning which was that all of the cases

had these crazy genomic patterns of genomic amplifications and nothing really that specific but kind of whole genome might be because because there's defects in dna repair. this is kind of what came out of the ttga is that number one p53 is mutated and number two, there

is some global defect in repairing dna. again, these are pathways, this is kind of another way of representing it the dna repair pathway was altered and the cell cycle progression pathways were also altered. so again, just to summarize, the

ttga was a large scale integrated view of aberrations in high great serous ovarian the mutational spectrum was emphasized as being surprisingly simple where p53 mutations were in 96% of the cases. brca one and two were in 22% of the cases and there were other

significant mutated against in 6% of cases. this makes high grade serous ovarian cancer distinct from the other histologic subtypes of ovarian cancer that's been published in the literature. so for example there's a subtype of ovarian cancer called clear

cell -- there's recurrent one a and there's recurrent -- kinase there's frequent data -- and very few k53 again. there's some k-ras -- ovarian cancer and again distinct from high grade serous ovarian so again, to summarize, there was a remarkable degree, this is

quotations from the actual manifest in nature. remarkable degree of genomic disarray, striking contrast to -- findings in glial blastoma. mutations and promoter methylation -- may explain the high prevalence of copy number

so what next? so how can we use this information to try to improve the overall survival of women with high grade serous, with any type of ovarian cancer? 50 percent of the patients i don't there are drugs developed called parp inhibitors which are

better at targeting cancers with defects in dna repair. there are mutated pathways such as rb, ras -- fox m1 and notch and perhaps we can find molecules that target these pathways specifically and give those to women who show aberrations in those type of

signaling pathways. and they did mention there are inhibitors for 22 genes that are in regions of recurrent amplification. i think the scene here is that not one gene can be targeted but we probably need to target network and direct the therapies

toward the network to have better improvement in survival. so now we get to clinical so what are we going to use for targeted therapies in individual clinical trials. and you can imagine that one of the things that we could use is a parp inhibitor.

so in actually, this predated cancer genome atlas in 2009, there was a discussion at the asco to discuss the phase two trial ofn't inhibitor called olaparib in ovarian cancer. what were parp inhibitors. we need to understand the pathway of dna repair.

this is very simplistic schematic diagram of dna repair but i'll put it in context for the parp inhibitor. with normal cellular metabolism and environmental he can portions, there can be single stranded breaks delivered to the dna that need to be repaired.

parp is one of the enzymes that identifies and sort of highlights the single strand dna breaks and the way it does that is the parp stands for poly adp ribose polymerase which is an enzyme that binds to these single strand breaks, recruits other elements and puts polymers

of adp ribose on to sort of flag it. if the single strand break is not repaired, that single strand break out of replication fork will become, will be converted into a double strand break because the replication machinery won't be able to get

through there. it will make the break but it won't be able to repair the rest of the strands. so if you have a parp inhibitor, in a normal cell, this double strand break will become, will get repaired by a pathway called so homologous recombination

requires the brca protein in order to successfully complete. so if the brca proteins are there, homologous combination will ensue and you will have this repaired and surviving. cells with brca consistency there's no repair of there. there are other repair

mechanisms but they're not as efficient and they're not, and they're much more error prone. so if it doesn't, if the single strand breaks, if the double strand machinery gets overwhelmed, the yell will -- cell will not repair the dna and they will undergo cell death.

so olaparib is an oral parp inhibitor that was shown in cell culture to have sent thetic -- homozygous cell lines. so the primary aim of this clinical trial was to test the efficacy of olaparib in confirmed brca1 or two mutation carriers with advanced

refractory ovarian cancer. in 57 patients 39 of them had brca1 mutation and 18 had br ca2. there were 33 available at a dose of 400 milligrams twice a day and 24 at a dose of 100 milligrams twice a day. the overall response rate was

higher in the 400 milligram arm as expected. however, the clinical benefit rate was -- sorry, clinical benefit rate was defined by response so shrinkage of the tumor or a 50% declean -- decline in a marker we call -- 25.

there was a higher clinical rate at 24 milligrams than at 100 milligrams. the survival was 5.8 plunltsd and the median duration of that was .6 months. the toxicity was actually quite similar to chemotherapy in that grade one or two nausea so

that's mild nausea, fatigue, anemia and low white blood cells which is the leukopenia. this is what shows you the best response of each patient on the clinical trial. so this means that the tumor grew and then below zero means that the tumor shrunk.

in three patients there was a clinical response. this was quite good because in refractory and recurrent refractory ovarian cancer we rarely get complete responses and a lot of these women had had multiple lines of chemotherapy prior.

this was actually quite remarkable that there was most of the patients had some sort of shrinkage and actually some of them even had complete resolution of the tumor. this is a graphical representation of what i already mentioned with the 400 milligram

dose is better than the 100 milligram dose. the conclusions from this in 2009 was that olaparib was in -- deficient ovarian cancer. greater activity at the higher dose. toxicity was similar to that seen in non-carriers.

it's not a direct outcome of this trial but that's compared to pretrials. and they quoted a positive proof of the concept of the activity and tolerablity of genetically defined targeted therapy with olaparib in brc deficient ovarian cancers.

so this is actually quite a big step. this is now being tested in randomized placebo controlled clinical trials for women actually in first -- ovarian cancer to see if we can get that overall survival bumped out beyond five years.

so now i'm going to tell you a little bit about where we've been going in the lab here. we're looking at nfkb did he regulation and targeting. the other thing, the other olaparib story took us from the genomic a braition to the pathway alter ation and we're

looking at a pathway and genetic alterations we can find that highlight women who have aberrations. this is not specifically ovarian cancer you can see it's far more complex than this but you can start with tnf signaling. you have this try maker complex

of -- gamma which activates the nfkb proteins that are actually transcription factors. i kappa v alpha which is the inhibitor needs to be phosphorylated and degraded. and then these transcription factors usually function in dimers of heterodimers of either

p55 and p50 or another components could be p52 and -- b. once those are released they can travel into the nucleus, bind to the dna and promote transcription of target genes. one of the key function here is that there's a molecule called

ciap, and that stands for inhibitor of apoptosis. that can be involved in activating nfkb by promoting the heterotrimeric -- inhibiting apoptosis. we looked to define ovarian cancer in kappa b signaling. we looked at an ikk beta

inhibitor. we wanted to define an ikk beta or gene signature -- to see whether this was an important pathway and whether we could find if it had a survival difference or how often it was expressed. we used an ikk beta inhibitor

which was a drug or we used ikk beta -- to knowing down ikk beta in order to look for genes that were consistently down regulated across both drug and srhna. in order to define a signature to probe large data sets. so we found that these nine genes were consistently

deregulated upon treatment with either the inhibitor or the shrna. we define this as our nine gene ikk beta signature in order to use in further study. so we looked in a bunch of data sets showing that these genes actually are co-regulated where

red is over expressed and green is under expressed. and this is a set of 185 ovarian cancer samples. each colon represents an individual patient and then each row is a particular gene in the so what you can see just kind of looking at this as an overview,

there are a set of patients which have high coordinated expression of these genes. here is the correlation with each gene with its average and this is a very highly significant correlation across these 185 genes, and i didn't include the p value here but

they're all less than .0001. so what this tells you is that because they're coordinated it suggests a biological function of these genes, not just you know random, some are up, some are down in various cases. when they're up, they're all down.

when they're down, they're all that implies there's a biological significance linking these genes together. and we did verify these in other data sets as well. so when we looked at these genes in overall survival, there was a significant, not a huge

difference, but there was a significant difference. if the patients with the very high expression of the nine genes had a worse overall survival than those with very low expression of the nine gene. so we wanted to find out okay, we found out, we convinced

ourselves that -- is active in there are patients who have a high level of -- activity and therefore perhaps if we block nf kappa b signaling in these patients, this might be a good therapeutic target for these patients. but we wanted to find out how is

the, how can we target it and how could we, and what other pathways we would be affecting. how is nf kappa b actually turned on in ovarian cancer. so what we did was shrna library screen with a library that's targeting 500 kinases in human kinome.

we combine the shrna screen with or without an ikk beta in order to see which pathways actually would enhance specifically the effect of the ikk beta inhibitor without, while not looking at things that were just broad we toxic across all ovarian cancer cells.

so the way this works is that the shrnas were transsected into a cell line. they are bar coded so each of them can be detected at the end by a bar code. infected into the cancer cells. we induced the shrna in all the cells.

and then treated half of them with ikk beta inhibitor and half of them with just the vehicle. and what we were looking for in the readout were shrnas that were lost only in the ikk beta group in the inhibitor-treated group and not in the ikk beta negative group.

in the inhibitor group in the vehicle controlled group. and if it was lost in both equally we didn't look at that at this stage. we were just looking for things that enhanced the toxicity of the ikk beta inhibitor. what we found actually was the

gene -- when that was knocked down, that actually was specifically toxic to cells that were treated with the ikk beta so how can we use -- to help to try to treat ovarian cancer. so what i had told you about the, the ikk, i'm sorry the nf kappa pathway with ikk beta

requires this signaling through the inhibitor of apoptosis or the iap molecule. now, in normal cells, there is a protein released from the mitochondria to promote apoptosis which is called smack which stands for second mitochondrial activator of -- so

when smack is released from the mitochondrion it binds to iap and gets rid of iap and that main function is to allow signaling to activate the apoptosis pathway. but then also, what happens when you get rid of iap, you also shut down nf kappa b signaling.

where does -- sit in this testing, it's actually activated here during this pathway in the proapoptotic arm. in tumor cells, it's thought that apoptosis is disregulated due to low levels of smack or to overexpression of iap's. so this might be a pathway we

could target in ovarian cancer. and we thought we would look first in situations where there was high immuno-- expression of tnf alpha -- signaling in serum concentration and tnf alpha -- trigger apoptosis in its absence. so our new hypotheses moving

forward from a this point was that intersection of the nf kappa b pathways provides a novel point of therapeutic intervention where prosurvival signals can be switched to pro apoptotic signals to kill kashes luckily there was a smack mimetic so there's a drug that

mimics the protein smack that's released from the mitochondrion -- we thought we would look to see in a panel of ovarian cancer cell lines whether this correlated with high nf kappa b activities. and it did. the blue lines are ones that

have high nf kappa b activity specifically -- which was the line we used for the library screen. so this line which has high dependence on nf kappa b signaling was the most sensitive to this smack mimetic. other cell lines were less

sensitive to treatment with the smack mimetic either by itself and then we also enhanced the killing actually by adding tnf. so tnf can stimulate kappa b signaling in cancer and other cancer and it can simulate what we found was two additional cell lines became sensitive to

the smack mimetic killing when we turned on the signaling through the tnf pathway. this was at a very very low concentration of tnf which did not kill the cell just by adding itself. we did a preliminary look at the proteins is what's happening

here and you can see of course the ciap was completely down regulated or degraded after smack mimetic treatment. casp --ally was up, casp interestingly went up and i kappa b phosphorylation which is this upper band did go away with an ikk beta inhibitor and with a

smack mimetic. so where are we now? we are going ahead with the clinical trial with the smack mimetic, relapse ovarian cancer. it's a drug that's given weekly for three weeks out of four. so this is the drug administration.

what we're doing to confirm, to confirm the activity of the drug we're taking a biopsy of the tumor before we treat the patient and we give the patient two cycles of the smack mimetic. and take another biopsy to look at whether the iap is degraded number one.

and then also to look at whether we turned off the nf kappa b pathway and turned on the apoptotic pathway. and then from a clinical standpoint we do cat scan imaging after every t cycle. our primarynd point of this trial is objective response or

progression-free survival. but we also have several correlative studies meaning the translational experiments and we're also collecting peripheral blood as well and serum cytokine. we'll be looking at the chemistry to measure tnf alpha

trail and gene expression profiling to look at that nine gene nf kappa b signature to see if this was altered by the smack -- we'll look at the apoptosis signature and look at a lot of the proteins in the biopsies like i mentioned the iap1 and 2, casp base cleavage

as well as parp cleavage. so again we made an observation in our patients that there were potentially a group that had high nf kappa b activity. we came up with the hypotheses that they might benefit from inhibitors of nf kappa b. we are currently in the process

of doing what we did a couple experiments, we're currently in the process of doing the trial. and then once we get the results we will analyze them and come up with new hypotheses. and i can give you some insight into where we are going. there are a number of drugs that

can augment apoptosis when given to cancer cells in vitro and in vivo. and these are listed here. so our thought is that we can promote apoptosis by treating the patients with a proapoptotic agent. and then also give them the

smack mimetic to augustment the response to this smack mimetic. so what we're doing in the lab right now is actually combining it with -- which is one of those drugs that can change the -- and promote apoptosis through release of cytochrome c. so our thought is perhaps we

could prime the cells by giving them the dose pax ill first and then give theming the smack mimetic. we will do the smack first or vice versa to see if we can look at protoemic end points and efficacy end points. to the cell line and then moving

to clinical trials. so our goal for individualized therapy is that therapeutic agents can target each critical pathway and they can be tested for their ability to overcome molecularly defined cancers. i really think this is the way we're going.

we know ovarian cancer is extremely heterogenus. i don't think it's a good bullet for all patients with ovarian cancer so i think what we need to move forward is defining subsets of ovarian cancer even within the histologic subtypes to sue if we can find -- pathway

dependent, target that pathway in that specific group of patients without having to expose everyone to a particular drug in order to get it into the clinic. the signature direct therapy will be assessed by prospectively profiling

individual patients and allocating therapy based on the molecular defect that we identify at the bench. this is mainly the clinical team that i work with. in the lab we have excellent scientists -- working on this from a laboratory standpoint.

and with that, i will finish and take any questions. >> [indiscernible]. >> so our trial is specifically for women who have shown resistance to platinum therapy. the reason for that is because we think that this pathway might be able to overcome the

resistance to platinum because a lot of time the platinum might be due to insufficient we have sensitive women in combination with other agents and then we are going to do the dose pax ill in combination with smack mimetic and that could be for anyone platinum sensitive or

platinum resistent. >> [indiscernible] >> that's an interesting question. there hadn't been anything in the lab that's shown the brca mutation patients are more sensitive. there have been in the clinic,

there's three or four smack my metics in the clinics and in one of them they did show, it's coming out that there is more sensitivity in the brca now whether or not that's just because those cells are just generally primed for apoptosis, that i don't know.

so i think that's still up for questions. >> not particularly, no, no. not particularly. in again the cell lines are kind of messed up, you know. the cell line are messed up they just have poly -- so from there isn't any impairity defect not

specifically like there are in the brca pathways. okay, thank you. >> our next speaker is sonia jakowlew and she got a ph.d. in rutgers university. subsequently she did a post doctoral fellowship in france and then she came to nci and

she's been here wow, probably many three decades now. so her title, transforming growth factor beta and lung tumor genesis. sonia. >> thank you for inviting me to talk to your class about transforming -- and lung tumor

genesis. and i know that dr. moody has spoken to you recently about lung cancer so i'll start my talk by highlighting brief specifics about lung cancer in this country in -- most recent -- for which this data is

available -- most common cause of death in the united states among both men and women. in this year, there have been, there were more than 200,000 diagnosed new cases of lung and over 150,000 new deaths due to this disease. most cases now occur in former

smokers because of the decline in smoking, the primary cause of lung cancer in this country. but lung cancer is thought to be a problem in the future in other countries because of the increase of smoking, particularly in the asian countries.

and right now, the five year survival rate for this disease is still less than 15%. so this is a very important disease that needs to be studied and needs people having these problems need hip. now my lab was interested in transforming the -- is a

multifunctional regular cell growth -- co-inhibitor of most normal epithelial cells in culture and shows wide spread tissue expression. it's been shown to play a pivotal role in maintaining epithelial homeostasis and it's associated with various types of

cancers, including human lung it shows context inhibition or stimulation of cell proliferation in neo plastic transformation, depending on the cell type. so we think tgf-beta is a candidate for new therapeutic intervention approaches and lung

now, i can take you back to the beginning. tgf-beta -- another growth factor called -- this growth factor was identified by -- in the 70's and it's a polypeptide that was found to be -- sarcoma virus transformed mouse fibroblast that stimulates

normal rat fibroblasts -- used to identify it and follow it. following this two classes of tgf's were isolated from this msv-transformed cells by anita roberts and michael sporn in the early 80's at the nci. and one class was found to be compete with even determinal

growth factor for receptor binding. and this was tgf alpha. the second class was found to not be able -- tgf binding but instead had colony forming activities that was enhanced by tgf. and that was the -- so following

this tgf-beta was -- and bovine [indiscernible]. now to give you some idea of the scale of this purification from bovine kidney typically present agent started with a hundred grams of bovine kidney obtained from the slaughterhouse. there were eight liters of acid

ethanol. it's expected with eight liters of the mixture of -- and then precipitated over night in a cold room with about 50 liters of ether ethanol mixtures. after this it's dissolved in two liters of -- acid and then applied to an 80 liter bio gel

p-60 column where one liter fractions were collected and processed for better chromatography. typically yield was about five to six micrograms of purified tgf-beta 1. and to give you some idea the columns that we used for this

purification, this is a typical preparation with using the -- column and here is given a 55 gallin drum for reference. now the essays for tgf-beta employed the growth of normal rag kidney cells -- typically a mixture of the media serum, normal kidney -- and the

portrait of the samples -- these were incubated for one week and then stained. the colonies -- so it's no colonies were present. no tgf betas present that would expect any colonies to be present. if tgf-beta is present in the

sample then they should be colonies that would be counted. so the final -- purification of a -- and is a poly it lien dose gel showing a -- on the gel. this was purified tgf-beta. michael -- roberts at the nci are credited for birthing this beta.

now following the purification the amino -- called the peptide -- and the tgf-beta containing, consisting of 112 amino acids. -- beta one in humans. i will point out the -- that are kip all of to summarize this is in 25,000 molecularly cell

binded -- pretty simple of tgf-beta includes the platelets, bones -- and is usually created in what is called the latent inactive form which must be activated in order for the molecule to perform various activity -- contains not only tgf-beta but the -- factors.

now following the amino sequence of tgf-beta the critical structure of tgf-beta was established at the nci and tgf-beta was -- a diermt of two identical monomers and contains a -- which is needed for interaction with this receptor molecule.

and what is called the tgf-beta super family. tgf-beta is the protocol member of this family. tgf-beta wants to -- exist --4 and 5 have been found in birds and amphibians and my thought has been in the mammalians in addition to the tgf-beta we

have these genetic proteins and the gross differentiation factor as well as the -- inhibitors of this that are part of this super family. now tgf-beta super family is thought to place a central control role for many biological processes including such

important processes as development, immune system function, reproduction, angiogenesis, aging, tissue repair, metabolic regulation and homeostasis regulation. there are many processes that are regulated by tgf-beta. among the most important are the

tgf-beta able to inhibit proliferation of most normal cells as well as some tumor it also is able to regulate such important processes such as apoptosis, differentiation and immune cell functions. also stimulates the accumulation of extracellular matrix and

promotes chemotaxis. now on model four how signaling occurs by tgf-beta involves the tgf-beta receptor complex. and this is worked out in the early 90's. so tgf-betas, the live-ins shown in blue binds to the type 2 tgf-beta re reresource which is

instingively phosphorylated -- type one receptor in order for transsection to occur in activities such as cell cycle arrests and gene activation and a whole host of other processes may occur. now, in addition to the receptors, there are receptor

activated inhibitors that are involved in this signal to inspection process. for tgf-beta -- two and three are the principal players that transduce the signal -- one, five dependent on the transcriptor model -- these then transduce the signal to a common

mass 4 which then is able to transport the signals signals into the cytoplasm into -- in addition to the inhibitors six and seven that can short-circuit the whole process. now, to put this altogether we have the tgf-beta which is the -- and then combines with

the type 2 receptor, recruits the type 1 receptor and then transphosphorylates, let me show you a sample, phosphorylates. then the -- depending on the molecule two or three are recruited, and they transduce the signal from the -- complex through the proteins.

the common -- then is recruited and it is the -- 4 that is able to transduce the signal into the nucleus whereby it can interact with various factors, growth factors and turn on or turn off p targeting. in addition -- seven is the inhibitor which can inhibit the

entire process. now clinically this has been shown to be a tumor suppressor. this is based on a loose lines of evidence. i'm giving you three lines of evidence here. jermine mutations have been found in pathway components that

cause familial predisposition to some cancers. smad4 in juvenile poly posis syndromes -- in some human cancers and this includes the -- receptor in human non-poly posis poly rectal cancers as well as -- pancreatic cancers. and thirdly, reduces expression

of some tgf-beta signaling pathway components or over expression of some pathway inhibitors have been associated with cancer progression. and it's been shown that the type one and type two tgf-beta are involved here as well as the inhibitors.

and another -- which was able to interact with tgf-beta. now, in addition clinically tgf-beta has also been shown to be a tumor promoter -- elevated in many advanced human tumors and it's been shown to correlate with either the metastasis or the prognosis of this disease.

and this has been shown in several types of human cancers including breast, colon, stomach and liver -- adeno carcino me which has been the for tgf-beta one. and tgf bayton one is found to silt in the interplay between tumor parenchyma and the micro

environment. so, what role does tgf-beta play in carcinogenesis. is it a hero or villain. it's a proximate effector of the malignant phenotype and it's also a potent growth inhibitor and humor suppressor and a prometastatic factor -- well the

hypotheses has been put forward that can withstand the test of time where tgf-beta is able to switch from being a tumor presser under normal circumstances to become more pro oncogenic factor during cancer progression. so how does it do this?

well, we have the normal epithelium and the tumor suppressor activities of tgf-beta can dominate, it can suppress the growth of normal now it -- and epi genomic as it progresses. we have a potential decrease in the responsiveness of tgf-beta.

we also have a possible increase in expression and/or activation of molecules. so what we have is that tgf-beta, the prooncogenic activities start to dominate as cancer progresses to metastasis. so the tumor suppressor actually goes from being a tumor

suppressor to more like a tumor promoter or pro-oncogenic factor. now in addition to the smad which we refer to the as the smad-independent pathway because it depends on the -- there are various smad independent pathways by which tgf-beta can

also operate. including the mac -- pathways, the mac p38 pathways, the ras kinase pathway the -- pathway and the -- kinase pathway. now, this seems very complicated and that's part of the problem. so we decided to work on one aspect of this complex problem

involving the role of tgf-beta in the ras kinase pathway. because we're interested in long cancer carcinogenesis and it shows an activational mutation in 25, at least 25-50% of human lung adeno carcinomas. it's been shown that mutation of even one allele of k-ras

increases appearance of long lesions in -- and it's been shown there's definite cross talk between smad dependent pathway and the ras signaling activation of a ras pathway has been shown to be able to modulate tgf-beta signaling with smad pathway.

and in-vitro studies have shown that tgf-beta 1 can dominate over the mito genic effects of ras. but when ras is activated by mutation, it can override the anti-proliferative effect of tgf-beta one. how can it do this?

it can do this by at least four different mechanisms that we can think of. there can be a decrease expression or production of the type two receptor. it can short circuit this smad pathway and lead to other progressions that can lead to

tumor progression. there can be activation through the pathways to activate ras. this is -- and pull it away into other activities that -- which again pulls everything into other pathway for tumor promotion. and there can be other changes

in the progressive arm of tgf-beta that can lead away from the smsd to other pathways -- there are probably other activities that can -- as well but we're interested mostly in the role of the ras map kinase pathway in our work. so the role of my laboratory was

to determine the role of tgf-beta to determine the role of transformation in epithelial this was done in the epithelial carcinogenesis section cell lab and cancer biology branch -- we have two objectives which is and the effect of tgf-beta 1 deletion and k-ras mutation

alone and in combination on lung tumor incidence and pathology. secondly to determine the early events in the development of lung lesions and their and thirdly, to identify potential signals, transdukion pathway changes that occur with advancing tumor genesis.

so for our work we employed four mouse model systems and i'll describe basically the aj mouse system -- the c57 -- heterozygous mouse, the -- and the tgf-beta one heterozygous k-ras -- mouse and i'll probably be speaking -- to shorten what i have to tell you.

so we asked two questions. does lung tumor genesis affect the tgf-beta signaling pathway and the receptical question does affect lung tumor genesis. we started with the aging model system because the lymph system has been -- and lung tumors that develop in a time-dependent

manner and goes through the stages of hyperplasia, adenoma and then carcinoma. and the carcinoma that develops have been shown to be very histologically similar to human lung adenocarcinoma. and the same molecular mutations have been found in both human

and mouse lung tumors. for instance over expression of the ras and loss of p53 expression. so we took these, we took two month old aging mice and injected them with thatal carbonate carcinogen also known as -- and then sacrificed them

in intervals for up to a year. and shown here is the tumors that was -- for t imsm f beta one, the type one receptor and type two receptor in two month four months and eight months sample. the hyperplasia adenoma and carcinoma.

and we see distinct brown staining for the tgf-beta one ligand as well as the type one receptor in all stages of tumor however, when we look for staining for the type two receptor protein, we see decreased staining for this protein at all stages in the

tumors. in comparison to the type one tgf-beta tester. now this is shown more clearly in this picture where we see distinct brown staining for the type one receptor and the tumor, as well as in the surrounding normal bronchial tissues.

now, in comparison, we look at -- chemical staining. this should be -- didn't survive the slides. okay. so when we looked at the type two receptor staining, we see reduced staining on the tumor compared to the type one

receptor. with you normal brown staining in the surrounding bronchial tissues. so we also looked at the messenger rnas for this -- receptor in mouse lung tumor derived cell lines and we see various levels of expression of

the type one [ -- and the type two receptor -- by an ethyl carbonate lung tumors and this will show expression of the type one -- but very low if any expression of the type two so this goes in good correlation with what we're seeing in our mouse model system.

now, we also looked at this staining for the type tgf-beta 1 and the tumor receptor and -- tumors as well, and we see brown staining in the tumors for tgf-beta one and the type one receptor as well as brown staining in the surrounding bronchial tissues.

we look at hybridization -- for the corresponding mrnas we see distinctions expression for tgf-beta one as well as the type one receptor. however when we look at the type two receptor for tgf-beta we see decrease staining in the tumor or the protein as well as for

the messenger rna with mueller -- staining -- so the type two receptor is being affected in this model system. we see decrease expression of the type two receptor and this correlates with lung tumor formation in mouse systems. now in this question -- does

deletion of tgf-beta one affect lung tumor genesis. for this, we use the c57 black tgf-beta one mouse. the beta one knock out mouse was generated in the early 80's and it survived. it looks a lot different smaller -- than its wild type --

but it can be worth it after birth. however that's 21 days after birth, it begin to succumb to a general wasting syndrome where all the organs were to shut so it's not a very good model system for looking at prolonged carcinogenesis.

but when working with these mice, it was noted that when these tgf-beta heterozygous mouse in which only one allele tgf-beta was released instead of two. when these mice were challenged with liver carcinogen, they showed a response in increased

liver tumors in these mice. now surprisingly this also shows increased lung tumor genesis in these mice as well. so we're interested in lung in order to make this system better for studying lung carcinogenesis, we cross the c57 -- heterozygous mouse with

the aging mice -- wild cards for tgf-beta one -- mice that had two phenotypes in the f1 generation. the heterozygous and the wild type genotype. our plan was to challenge them with carcinogens to generate lung tumors.

so this slide shows -- beta one in the heterozygous mouse -- expression of the protein as well as the messenger rna for tgf-beta one compared to the wild type -- when we looked at this expression of the mrna [indiscernible] we see decrease expression in the heterozygous

genotype -- we did the he can pression on the tgf-beta one -- in the heterozygous mice compared to the wild type as we expected. so the system is working. next we injected two month old mice with ethyl carbonate as we did before and then sacrificed

these mice over a span of 12 months. and shown here on this graph we show increase tumor incidence in multiplicity and decrease tumor latency in the heterozygous mouse compared to the wild type -- if you look at hyperplasia -- there's

hyperplasia and -- occurring one month after challenge with the carcinogen. it's not until two months and four months that we begin to see hyperplasia in the wild type mice. most dramatically we see carcinomas appearing by four

months in the heterozygous mouse, mice, whereas it takes at least 12 months or longer before we see tumors appearing in the wild type mice. now, we stained these tumors for tgf-beta one in both -- and we see -- tgf-beta one in the wild type and heterozygous mouse and

hyperplasia and adenoma. however when we get to the carcinoma advanced stage, thing pression of the type two receptor for tgf-beta is very reduced in the heterozygous mice compared to the wild type mice. we looked at the corresponding mrna for this as well and we see

decreasing levels of the messenger rna for the type two receptor was increasing tumor general os as we proceed from hyperplasia to adenoma and on to next, we ask the question does the deletion of tgf-beta one and mutation of k-ras in combination and for those who use the

tgf-beta one heterozygous k-ras -- so we, to study the interplay of tgf-beta one with k-ras, we generated mice by crossing tgf-beta one heterozygous mice in p56 -- was k-ras activatable -- mit and they tained what is called latent activatable mutation

where the k ras mutation is there but it was for another event to act rate the mutation and get tumor genesis going. these mice generates four different geno types. activatable k-ras mice -- single mutant mouse and the wild type mouse that's controlled.

and shown here are some photographs of the resulting tumors that we found in these lungs at three months after the challenge with carcinogens. and the double -- we see nodules on the surface which are the -- and this we see as well in the single mutant k-ras mice.

in the tgf-beta one heterozygous and the wild type mice we see only a rare if any tumor appear even after prolonged time. we looked at the effect of tgf-beta one gene deletion in k-ras mutation on mouse survival, and we -- from genotype and showed that the

tgf-beta 1 k-ras double mutant and the k-ras single mutants had decreased life span in comparison to the tgf-beta one heterozygous mouse in the wild type mice. next we looked at the pathology of some of these lung lesions and we see increased numbers of

hyperplasia and adenomas, and the single k-ras mutant mice. increased numbers of carcinomas in -- mice. next, we performed immuno-- staining in type two receptor in -- adenomas and carcinomas. we see staining in proteins in the adenomas.

advanced carcinoma stage and the double mutant mice we see various decreased reduced staining -- protein as well as the protein for type two in the adenocarcinomas. and next we decided to look at the protein levels of some of the tgf-beta components in these

mice and also a few slides of results. and when we look at the type two receptor in the mutant mice compared to the single k-ras mutant mice we see an expedited reduction in the level of type two receptor protein compared at two months compared to three

months in a single mutant k-ras and we see now virtually no changes in the other two genotypes. when we looked at the smad three proteins which is a receptor smad, we see as well an increase in the time at which you see increase production of the smad

two protein in double -- mouse mutant mice. so there's an expedited increase in the smad protein and an expedited decrease reduction in the type two receptor protein. next we looked at the smad 4 and both k-ras double mutants and single mutants there seems to be

a reduced expression of the type 4 protein particularly in the double mutant mouse compared to the single mutant mouse. we looked at smad 7 and surprisingly very low if anything in both genotypes in -- and the wild type heterozygous mouse.

so this implies these no inhibitor protein to counteract the system with the k-ras mice. next we look at the corresponding mrna for some of the smad and the type two essentially we find mass of rna for the smads and the adenomas and reduced as well as in some

of the carcinomas -- the protein level. since we're interested in the k-ras system, we looked at expression of k-ras in raf1 and in the double mutant mouse compared to the wild type latent activatable mouse we see an expedited increase in the k-ras

and raf protein by about a month compared to the single mutant k-ras mouse. the wild type we don't see any change. since tgf-beta is important in apop toafsz we saw reduced apoptosis in this mice. there were reduced levels of

apoptosis in the double mutant k-ras mice compared to the single mutant k-ras mouse. this is also affected when we do tunel assay apoptosis. in the double mutant mouse we see decrease levels of apoptosis staining in the adenomas and hyperplasias to the mutant k-ras

now virtually we can't -- apoptosis in the advanced carcinoma phase. so putting this altogether. in the single mutant k-ras mice, we have tgf-beta, smad inhibitory pathway in decreased at the same time we have k-ras which is the bully and they look

to push the progression of what is occurring because it is so, such an active molecule. a lot of the -- activity and the smad independent activity in order to accelerate tumor but in contrast with the tgf-beta 1 heterozygous k-ras double mutant mouse, we have,

with this advancing at the tgf-beta smad dependent and -- pathway is debilitated because it is -- earlier. in combination we had the k-ras activated k-ras -- the same smad independent pathway effect which lead to carcinogenesis. in addition, we shown in our

studies, there is very much reduced levels of smad 4 in so this also helps to reverse the inhibitory pathway to encourage more the prooncogenic effects of tgf-beta. so we think these mouse model systems are the way to go to try to understand what's happening

to this dreaded disease. and i think that -- and other people have now constructed better k-ras mouse models where you can look at these for longer times and over time periods to really find out what is happening. so in our mouse model system

we're showing that you see decrease levels of the type 2 receptor for tgf-beta -- one tumor promotion. we also see activated ras -- pathways that also correlates with lung tumor promotion. our system shows decreased levels of smad 4 in our system.

and that seems to also correlate with lung tumor promotions. and finally, compromised apoptosis in a model system also correlates as one tumor so i would like to thank the people who contributed to this work that have since moved on to a bigger and better position.

jerry angdisen. yang kang -- again i credit tyler jacks for the k-ras -- and now we can answer any questions if you have any and time permits. thank you. [applause] >> antagonists are being

developed, okay. unfortunately they're very difficult to produce and they showed minimal effects on lung cancer cells. but they're in the works. have i overwhelmed you? well thank you. thank you so much.