Harnessing Cross Talk Between Signaling Pathways to Improve Cancer Tr

I get to take over as care of this section and introduce the next speaker which i assume is you yes this is our co-chair so it’s a very it’s a seamlessly organized meeting that I involves minimal excess energy so mighty aathi needs no introduction at all Mike I’ve known since he was a postdoc in lieu Cantonese lab i will say that i have an introduction here that he rode out no offer no actually I didn’t but I just want to say one thing about Mike Mike is both remarkable and crazy okay so Mike and I are both md PhDs and most mere mortal md PhDs after completing their training clinical training if they do clinical training and their research training pick one thing or the other to engage in but Mike is mike is one of those rare individuals whose a triple threat player he stays involved clinically and not only does he stay involved clinically he’s a surgeon which is really nuts so he’s a really remarkable I don’t think he sleeps maybe he writes himself prescriptions for adderall with the that’s good oh okay I’m never getting invited to MIT again first talk here so anyway so without any further ado I will have my okay thank you been I used to not need sleep along good I’m like I used to not need sleep but that was before I had kids okay so what I’d like to tell you about today in the next 20 minutes or hour 25 minutes or so is some work that we’ve been doing on how you can use dynamic rewiring your signaling networks the same type of thing that Joan broody talked about earlier in order to optimize our ability to design rational accommodation therapies for cancer and what we’ve been interested in particularly is the combinations of either chemotherapy drugs or chemotherapy and radiation because those are of course the mainstays of how we treat cancer and the thing that my lab has been focused on primarily has been on studying the signaling networks that are involved in DNA damage because that’s the mechanism by which chemotherapy and radiation are used to cure cancer but more recently it’s been the intersection of the DNA damage singling network with protein kinases and cytokines for example a little bit now in cancer metabolism and asking how these networks work together in the type of approach that the Glavine burger talked about in order to control what the fate of cells is now before I get there I just want to put in one quick plug for this new journal called science signaling it’s the first offshoot of science that was that there is and it’s of course focused on the topic that’s near and dear to all of our hearts and that’s signaling and I would encourage all of you to send your very best papers to science that we forget nature forget cell you want to send them to science signaling the chief editor of science signaling is a nice guy and the editor the overall editor Nancy Gough is here at the meeting and so I encourage you afterwards during the drinks to go up and say hi okay i want to make three main points in my talk if you don’t if you sleep through everything else i hope you’ll remember these three points the first is they’re targeting mono therapies for cancer the stuff that that you’ve heard about from from jeff settlement today and you heard about from earlier speakers they’re fantastic things but they don’t cure the disease they result as you’ve seen in impressive remissions of the tumor and sometimes for prolonged periods of time but ultimately the disease recurs in all the patients if they live long enough the second thing I want to tell you is that most forms of combination chemotherapy aren’t actually synergistic now this is different than what I was taught because I was taught that the reason we gave combination chemotherapy like chop for leukemia lymphoma was because the drugs somehow acted synergistically but the data that’s begun to emerge suggests that in fact that’s not the case at all that the reason that those drugs work is because they target heterogeneity in the tumor cell population so cytoxan for example targets one set of drugs and vincristine urban blasting targets different tech targets a different set of cells and so in fact it’s not personalized medicine I would argue it’s d personalized medicine now the advantage of the fact that we use these different drugs they don’t have overlapping toxicities but the disadvantages we’re really not making the best use that we could of combination chemotherapy and if we want to cure cancer what we have to do is come up with an effective way to identify synergistic combinations and I’m going to make the plea that the answer comes from systems biology the idea is that we can use on i’m going to show you this time dependent network rewiring to sensitize cells in a predictable way to drugs we can use this in essence to funnel cells down a pathway that we want them to go down and then clobber oh okay and this may be particularly important for killing cancer stem cells whatever it is that you think these stem-like cells are whether those are cells that have moved backwards of the type that Rudy talked about or the type that Joan talked about on the outside of those spheroids but I think this is an approach it’s going to be particularly useful for that the

other advantage I’ll show you is that we can use these systems insights not necessarily to indicate that there are new drugs we have to go out there and discover but maybe we can take the same old drugs we already have and use them in more effective ways ok now since I’m going to make a plea about systems biology I have to tell you what it is in a great thing about systems biology the thing I like best about it is it’s whatever you want it to be whatever you do if you want to call it systems biology it’s okay with me the approach the definition I like best is the one that craig Venter made and he said basically it’s we got to get out of that one postdoc one protein three-year mindset and change it to studying what does he call it here he says we have to study integration of thousands of proteins I don’t we do thousands but at least more than one in a dynamically changing environment okay now the way that we’ve been focused on this at MIT inspired largely by the work of dog laughs and burger is to think of biological circuits much in an analogy to electrical circuits now obviously this is an electrical circuit and this is our biological circuit and of course since my lab is is really focused on signaling and protein kinase is in particular I’m going to make the argument that protein phosphorylation basically works like electrons flowing through wires to communicate and that’s the current of signaling okay now if you wanted to study what was happening in an electrical circuit you wouldn’t sit there with a voltmeter and measure the voltage at one point and say I understand how this complicated circuit works because I’ve measured the voltage at one point but somehow it’s okay if we measure Kirk activity right in one cell line in response to one stimulus and we pretend we understand it what you really want to do is you want to take a circuit and you want to measure using what’s called a bed of nails testers the voltage is in currency thousands of nodes and at least what we want to do with biological circuits is to measure metaphorically the activities of multiple signaling proteins Mac and Eric and EGF for example junk p38 caspase-8 and as many points in time as we can under realistic conditions ok and we’re going to and again I you know my lab and I think I speak and speak fairly for forest white slab and peter sauber slab we’ve all drunked the same kool-aid the Doug cooks up and that is that and that is the belief that we can understand thanks to a multivariate network model the so-called Q signal response model and the idea is that if we stimulate cells I’m sorry this seems to be dying here if we can if we stimulate cells with DNA damaging agents and we measure their response in particular how well can we kill them that the death we measure must be some function of the input as mediated through the signaling network so the idea is let’s measure as many signals as we can and let’s make the assumption that the response we see is some function of the signals and all we have to do is decode what the function is that relates the signals to the responses and then we’ll know exactly what we need to do oh it’s fantastic thank you ah wow wow that’s powerful so you ended when you run the Koch Institute I guess you can have that kind of power I’m going to give this to you in a very concrete example I can’t retrieve an example and this is an example that Mike Lee in the lab has worked on part of this is published and I’ll show you new date at the end that is not now Mike was very interested in trying to ask could we improve the treatment of breast cancer and the particular type of breast cancer that Mike was interested in treating is a type called triple negative breast cancer while this band knows that and many of you may know breast cancers historically before people study done by gene expression Alice were identified by a pathologist is what type they were they were called either while we didn’t call luminol they were called either estrogen and progesterone receptor positive tumors that if they had hormone receptors or if they didn’t they overexpress the EGF receptor family member her too and if they were negative for estrogen receptors and negative for progesterone receptors and they didn’t overexpress her too while they had three strikes so they were called triple negative breast cancers now the problem with triple negative breast cancers is they affect young women primarily and they have the absolutely worst prognosis these we know what to do with at least a little bit we can use a at agganis of estrogen and progesterone receptors these guys we can use her to directed therapies but there’s no good therapy for triple negative breast cancer now a subset of breast cancers overexpress the EGF receptor and that’s somewhat over-represented about 50 to 75 percent of triple negative breast cancers so maybe there’s something we can do with this and so a rationale was what we would do would be to combine the things that we were interested in DNA damaging drugs or treatments for example like ionizing radiation or camptothecin or cisplatin umrah top aside with specific signaling pathway inhibitors like that be easy drug that you heard about earlier and we would ask could we find some combination of drugs that we could use that might be particularly effective and that the combination that I’ll tell you about ah one thing made this different so of course all of you that are in industry and all of you that are interested in this from an academic point of view have already probably fall up and tried these experiments where you mix drugs together but we reasoned because of some work we had done over the years on systems biology that maybe there was a little tweak we could do that would make things different maybe what we could do was instead of just mixing the drugs and seeing what happened we could give one drug and we could wait and then later on we could

give the second drug or we could give the give this green drug first for example and then later on give the blue drug and we reasoned that maybe things would happen over time between what we gave the first drug in the second drug that might change the outcome and the two drugs that I’m going to tell you about today are doxorubicin adriamycin a dna-damaging drug that causes double strand breaks and are a lot maybe need GF receptor inhibitor and the point I want to make first is you know doxorubicin and erlotinib are both clinically out there used for the treatment of triple negative breast cancer now when we first started to do these experiments all of my who treat patients with breast cancer said to me Mike you’re wasting time what do you want to do this for it’s already been done in culture and there’s very little there’s very little benefit from mixing the two drugs together and there have been three clinical trials where people have given patients or lot nib and doxorubicin and the benefit is really minimal and that’s what I’m showing you here this is the percentage of a pot pot excels this is a triple negative breast cancer cell line in a dish and we’re looking at a pop ptosis if we give them erlotinib 48 or 36 hours where we treat them with doc service and/or we put the two drugs in together and you can see if we mix the drugs there’s a little bit more a pop ptosis but certainly nothing impressive but what Mike noticed was that if he simply waited if he gave the erlotinib first and waited and then somewhere between 1 and 48 hours later gave the doxorubicin he could find conditions where the cell death increased five hundred percent if you reverse the order he damaged the DNA first with doxorubicin then he gave her a lot I’ve in fact not only was it not beneficial but he got less death than he would have gotten if he’d even given doxorubicin alone this was antagonistic okay now to make life easy I’m very fortunate because doxorubicin starts with D and damages DNA so wherever you see the letter D think DNA damage and erlotinib which targets the EGF receptor happens to also begin with e so wherever you see II think EGF receptor and where there’s an arrow it means I’m going to do one thing first wait and then I’m going to do the other things so er OD means block the EGF receptor wait damage DNA okay and this is the result I’ve shown you in fact if you do a child to Lily analysis this is synergistic you can increase the death remarkably if you simply give a drug that’s already out there wait and then give the second drug but it only works in triple negative breast cancer cells if you do the very same thing in a different cell line this is a cell line that over expresses the her2 oncogene not only don’t you get an increase in death in fact you get a decrease you get this is antagonistic you’re worse off here than if you just gave doxorubicin alone in luminal cells the estrogen and progesterone receptor expressing cells if you do this time stagger treatment you get a little bit more death but it turns out not to be synergistic and in a cell line that’s annotated as being normal breast cells you get almost no death notice the maximal is five percent nothing happens so this is treatment that somehow specific to these triple negative breast cancer cells now if you’re a signaler there’s a problem that’s going to bother you about this and here’s the problem the problem is if I look at what happens to EGF receptor activity by phosphorylation of the EGF receptor or phosphor activation of erk activity all that activities over in 15 minutes but I showed you to get maximal effect I had to wait eight hours so if the EGF receptor inhibition happens in minutes why do I have to wait hours to see the synergistic killing well let me dismiss one possibility right away you might say well that’s easy Mike it’s an off-target effect it has nothing to do with the EGF receptor at all you gave her a lot nib and all kinds of weird stuff happened it was independent of the EGF receptor so what mike lee did was he used RNAi against EGF receptor he knocked down the levels of the EGF receptor and if you knock down the Jeffers afternoon to give doxorubicin you get just essentially almost as much killing as you get with the time stager treatment and if on top of that you now add or a lot and if you don’t you you really don’t get any benefit so it’s going through the EGF receptor for the sake of time I’m going to going to simply tell you the punchline of the next part which is if you because it took hours it must mean something is different and so I did what we put in my lab we always say you do when you run out of alternatives and that’s you do gene expression analysis and you ask okay what what what can I get out of this and this turned out to be very revealing what it showed us in fact was that about 20 different pathways we’re all being down regulated but the overall signature of gene expression that is what all those pathways came together to tell us was this was a signature of the wrath unka gene so it meant in fact that the EGF receptor signaling seemed to be driving the expression of a rass onca gene expression signature and that these triple negative breast cancer cells might be Agha gene addicted to that wrath on gene expression signature because of the EGF receptor and when we blocked it chronically but not acutely that is four to eight hours we could wean the cells off of their oncogene addiction and basically wean them off of the expression signature that goes along with rasikas gene addiction in fact if you use the GSE a data set at the braud that what it looks like

is everybody who does an experiment where they knock down wrasse or they knock down Mick gets a signature that looks like what we saw when we chronically suppress the EGF receptor okay so of course we’re signalers and this symposium is about signaling so let me tell you the signaling punch line what’s the signaling mechanism ah finally a sewing Network ok so what we wanted to understand was what was the relationship between for example these growth factor receptor pathways and the DNA damage we were inducing with doxorubicin and how did this somehow feed into things that were controlling cell death or proliferation or DNA damage repair or not a pathetic death or autophagy and so mike lee set out two measures much of this stuff as he could and he worked out assays over the course of a few months they could measure basically everything that you see here shown everything you could measure everything shown here in white and about these days about a third of the things shown here in gray so we could measure a reasonable amount of signals that were in all of these different pathways now this is a terrible slide and I say it’s a terrible slide because what I’ve just done is reduced about nine months of Mike’s life to one slide two slides actually and what I’m showing you here is a summary of measurements of every one of those signaling molecules that he could measure in three different cell lines triple negative breast cancer cells her to over expressing cells luminal cells in which under all six different conditions so every one of these panels this panel right here represents for example this is Kirk activity this is phosphorylation of work in those cells in the triple negative breast cancer cells treat with doxorubicin or a lot nib were the three different combination stocks and are live together Doc’s first and then a lot never Lawton it first and then doxorubicin and to make life easy i’m just going to color code them for you if this if the signal went up really it’s shaded green if it went up late its shaded red if it went up and back down again we call that transient shaded in yellow if it was high and went down that shaded in blue and the strength is obviously the height of the bar so this is all the signals that we could measure and he did this using a combination of micro western blotting that was very quantitative spent a long time making sure that if we said the increase in arc was one point eight fold that was really one point eight fold it wasn’t two point eight fold and it wasn’t one point three fold it was really 1.8 fold plus or minus some small variance you also did this using reverse phase lysate arrays and now we have this big compendium of signals that we might be able to use to understand why these cells were sensitive to a particular treatment but these other cell types weren’t Mike also measured responses so what he measured of course was a pop ptosis proliferation where they were in the cell cycle an apology and this again was all relatively straightforward you could measure cell survival for example by measuring ATP content in the cells you could look at where they were in the cell cycle by using facts and if you cut if you stayed for fossil histone h3 by fax then the cells that were here with 4n DNA content but up here with fossil histone h3 these were mitotic cells so you could quantitate how many cells were in g1 how many were in g2 how many were in em and how many were in s and you could stay in the cells for cleve parp and cleaved caspase-3 to figure out how many were a pop otic and then you could transfer the cells with gfp lc3 and count puncta after DNA damages a marker of otology so now we had signals and we had responses and we could ask what signals correlated with what responses and how did that account for why these cells appear to die when we use that time stagger treatment and again we’re going to I’m going to follow Doug’s lead here and tell you that the way we did this was rather than looking at things the way I was used to looking at things as a signaler which would be plotting things like art activity or junk activity as a function of time building a signaling space in which each axis was a different signal so junk might be this access or might be this access h2ax would be this axis and asking what direction can I walk now of course we have 35 signals so it’s 35 dimensional signaling space and we want to ask what direction do i walk in thirty five dimensional signaling space that most correlates with a pop ptosis or most correlates with cell cycle arrest or most correlates with otology and the best predictor of course what we’re doing of course is we’re finding the principal components the directions we walk to capture the signal and I can tell you that a simple two component model did a pretty good job of measuring a lot of the things particularly apoptosis so the first principal component turned out to correlate with whether cells were surviving or dying the dead cells were over here the surviving and proliferating cells were over the second principle component I don’t really understand it must have something to do with the cell cycle because then phase cells are up here and g2 cells are here and s and jiwon cells are here but it’s hard to put a simple biological meaning on it nonetheless if you take these two principal components you can do a phenomenally good job at predicting cell death under any of those DNA damage conditions it’s a it works really well if you say up I’m going to give the cells this much a lot bib and I’m going to wait and give them this much doxorubicin how much death will I get and you build a model with conditions

that you haven’t tested so you take the data from the condition you’re not going to examine and you build a model and then you do take the condition you’re interested in you make the measurements you dump it in and you say with these signals how much death do you predict you’ll get how much death do you measure you get you see it’s bang one really works well so the model works but more importantly the model can give you biological insight you can go back to the model and say why did this model work what is special in triple-negative breast cancer cells about pre treating them with erlotinib and then clobbering and with doxorubicin it makes the cells die and so what I’m showing you here is something called the variable importance and projection I’m going to take all those signals and I’m going to a ask for bt20 cells if you would just focus on the blue lines I’m going to rank the signals from most important to least important in predicting death and what the model says is the most important signal it’s important in predicting the death you get with that time staggered approach is cleaved caspase-3 see furthermore not only is it the most important signal for the triple negative cells for all those other cell types that rip the hair to overexpress errs and the luminal cells shown in red and green you see it’s the least important predictor it’s only important in those triple negative breast cancer cells this was great we were all very excited till we looked in the literature because it doesn’t make sense and the reason I say it doesn’t make sense is because when you damaged cells with doxorubicin and you cause intrinsic DNA damage that’s supposed to signal through the intrinsic death pathway which is mediated by caspase-9 and caspase-3 and I just told you that the model said the critical thing was caspase-8 caspase has nothing to do with the intrinsic death pathway caspase 8 is involved in an extrinsic death pathway that you see here and so I said to Mike Lee who done the work I said Mike that’s great really exciting except it can’t be right because them after all that’s not what the literature says so Mike said let’s see what the model says let’s make a model of those triple negative breast cancer cells that captures the cell death we saw and let’s take caspase-8 out and if you take caspase-8 out what the model says is the biggest difference you’re going to see is if you do that pretreatment the thing that killed all the cells but what the model says is if you give the two drugs together you’re not going to see a big change and furthermore the model says that if you look at those other cell types to hurt to over expressing cells with caspase-8 without caspase-8 not a big difference Mike then used RNAi to test it he knocked down caspase-8 in both of these cell types and he made the measurements and to my surprise in the light I think you can see here that they’re signaling data for the that we observe for the triple negative breast cancer cells perfectly mimic what the model said we specifically lost the increased death that we saw with the pretreatment now this is important i would argue because it means we now have a biomarker it means not only do we have a mechanism that explains the death we have a biomarker that we can use if we decide to do this in patients for following if they respond we can look at the tumors after tripping and ask do we see cliff caspase-8 if we do then we know it must be working through this mechanism now is this the cure-all for all triple negative breast cancers I was excited we wrote the paper up we sent it off we were all excited the views came back said that’s great you can cure bt20 cells now in a dish show it show us that it’s true in general so we looked at ten different triple negative breast cancer cells and I’d like you to just focus please on this black the black bar which is the condition I showed you where you give the EGF receptor inhibitor you wait and then you damage the DNA and in every case except one in every case except one it’s the most effective at killing but it’s only synergistic if you do a child to layla analysis in forty percent of the cell lines only these four shows synergistic killing between time staggered EGF receptor inhibition in DNA damage why what special about these four cell types compared to the six that wasn’t synergistic in and I would have said well I bet it have to do with EGF receptor expression levels maybe it only works in cells that have a lot of EGF receptor expression after all there have been many many laboratories some quite famous and national laboratories at that who are devoted to showing you that we should be measuring mrna levels of these things and using that to tell which patients ought to be on which drugs and so what we did was to look at the cells the amount of death that we saw at eight hours in response to the different treatments and I’m showing you them to you hear what I want to do is I want to normalize these to the amount of death we see relative to doxorubicin alone so if when we add the erlotinib we get no difference compared to erlotinib alone will make that light yellow and if we get the most synergy will make that black and I’ve ranked these cells from those that show the most synergy to those that show the least energy and here I’m showing you the levels of EGF receptor that are expressed in the cells I’d like you to notice that there’s no correlation there’s no correlation if you use egfr protein levels there’s no correlation if

use egfr mrna levels and so everyone that tells you you’re going to able to figure out what drug to put your patient 1 by measuring their mrna levels of signaling molecules sorry i’m going to tell you they’re they’re lying to you they’re selling you the same bill of goods that Romney is now sorry the only political crack I’ll make but if instead of looking at EGF receptor levels you look at phosphorylation of the receptor you look at how much is that cell line signaling through the EGF receptor in the absence of any stimulus it’s a beautiful correlation it’s those cells that under basal conditions are signaling the most through the EGF receptor those are the guys that are addicted and those are the ones who show not only the most apoptosis with the combination time stager treatment but they’re also the ones that show the cleavage of caspase-8 okay and so I think what this gives us is a biomarker for patient selection we can select the patients who would benefit from this time staggered approached by by asking which patients tumors have the highest levels of Faso a GF are now I learned a lot of things from Tyler at the Koch Institute and one of the things that I learned the most about was that you can’t cure cancer with cells in a petri dish you have to do it in a mouse right if you can do it in a mouse you’re one step better now and so we set out to do that experiment we took we took a triple negative breast cancer cells in this case it’s not it’s not the perfect mouth that Tyler would have used it’s as even agree after the nude Mouse but it’s at least a step in the right direction and we tried treating them either either with the drugs in the combination i showed you or giving the drugs together and we looked at what happened now in these experiments what i’m showing you is tumor growth is a function of time in these Xena graphs and when the tumors get to be about this size we can’t at this point our animal care people say the tumors are too big you have to stop the experiment this is what the tumors do by themselves and where you see this arrow we’re going to give these mics one dose one single dose of doxorubicin and when we do that the tumors regress a little bit and then they pick up and keep growing again if we give the two drugs together just like we’ve done in all the clinical trials we give doxorubicin and our light up at the same time we got a little bit more of a reduction in the tumor size but it picks up and keeps growing but in this experiment if we gave a dose of erlotinib in 12 hours later we came back and we gave those mice one dose of doxorubicin not only did the tumor shrink but at least over the course of the experiment it never grew back would it grow back if we went out further probably but certainly this was very encouraging data and so the model at this point is that we think that a subset of triple-negative breast cancer cells are addicted through the EGF receptor to an oncogene signature that masks this caspase-8 death pathway so when we use DNA damaging chemotherapies we can only kill them through that intrinsic death pathway but that we can dynamically rewire those signaling pathways just like Joan talked about so that if we chronically suppress the EGF receptor with erlotinib we can wean them off there are Jean addiction unmask this caspase-8 pathway and now the same amount of DNA damage gives us increased death because there are two death pathways that we can activate let me leave you with a few very quick slides first is this only true and triple negative breast cancer cells no fact it’s true in lung cancer what I’m showing you here are our egfr mutant lung cancer cells treated with doc service and a drug fact that we don’t use in the clinic because it’s not very effective but in fact you can see it’s dramatically effective if you simply pretreat these these triple negative these lung cancer cells with erlotinib and then come back with doxorubicin later the mechanism seems to be the same we see caspase-8 activated only in the pretreatment condition and we dramatically reduce the synergistic death that we see if we knock out caspase-8 using RNAi so it looks like it’s true in this subset of lung cancer cells here’s an even more paradoxical response now these are a549 cells now these cells have a wild-type EGF receptor in fact they have a rass mutation I would have thought rass being downstream of the EGF receptor these would show no synergy but in fact it’s even more impressive here’s a cell line doesn’t do anything if you treat with doxorubicin you give it a lot not much happens give the two drugs together not much happens but if you give the or lot they’ve been you wait and you allow the cells to rewire now doc service and results a dramatic increase in cell death and again it’s only in this condition that you activate caspase-8 cleavage and you can dramatically reduce that synergy if you get rid of caspase-8 or reduce it using RNAi well maybe this is something that’s unique to the EGF receptor maybe it’ll only work with EGF receptor let me convince you or try to convince you that this is true in an even broader sense maybe this is true with receptor tyrosine kinase signaling in general so this is the data that I showed you before this was the data where I showed you the triple negative breast cancer cells could be synergistically killed in this black bar if we pretreated with the EGF receptor inhibitor wait come back with doxorubicin and this was the data I showed you earlier that if we took her to over expressing cells there if we

pretreated we got less death it was antagonistic and we reasoned that maybe that was because her two cells instead of being addicted to the EGF receptor were addicted to her too so what if we use the her2 inhibitor instead of an EGF receptor inhibitor in that case we saw that we could dramatically could synergistically increase the amount of death we got and again what’s interesting here is you get a fair amount that even if you just mix lapatinib and doxorubicin but the best killing you get is if you do that time staggered approach you block the her2 you block her too and egfr in this case with Adam wait and then come back with doxorubicin very last slide last piece of that I want to tell you now you might say well that’s all well and good but you know pharmacokinetics and patience is a lot harder than being able to add things to media and cell culture how are you ever going to get conditions right in a patient where you’re going to be able to get the concentration of erlotinib in the tumor at the right amount at the right time to be able to then hit them with her with a dot with doxorubicin and so recently we’ve teamed up with Paula Hamid a chemical engineer in the Koch Institute to try to build nanoparticle-based drugs that have an outer shell that releases in EGF receptor inhibitor and an inner core that then releases doxorubicin so we can program this time staggered release directly in a single drug so the way that we’ve been doing this is to use a combination of use a plga scaffold in which the erlotinib is free floating in the doxorubicin is both free and tethered to the scaffold so that you’ve got basically an outer shell of erlotinib an inner shell of doxorubicin linked to the plga and what you get in essence is rapid release of erlotinib followed by a lower dose sustained release of doxorubicin and again this is the result I showed you earlier this is cell death overtime when you use that time staggered approach where you give their lot and they’ve also called tarceva that’s why it says T and then you follow that with doxorubicin if we use these nanoparticles alone where we use nanoparticles that only release the EGF receptor inhibitor or nanoparticles that only contain doxorubicin you don’t see very much if we use particles that have doxorubicin and we give oral i nip to the mice sorry we give a lot of you know this isn’t much we give her a lot up to the mice we got a little bit of death but the best death we get is if we use a particle that simultaneously releases or latinum quickly and doxorubicin later so we can recapitulate what we saw using two drugs time staggered by one drug one nano particle that has time delayed release of doxorubicin with earlier release of her laudanum okay and again it’s only with those particles that we see this activation of caspase-8 so what I’ve told you about today is how I think we can use context dependent rewiring of signaling pathways to get results that you wouldn’t see with one drug own or a second drug alone if you use the drugs in an intelligent way that you can get enlightened by by systems biology I told you the EGF receptor pathway Cross talks with the DNA damage pathway in a subset of triple negative breast cancer cells and you can use this this cross talk ordinarily limits the efficacy of cytotoxic chemotherapy you can rewire dynamically rewire those pathways in a manner that suppresses the signaling through the EGF receptor to bring up a new response cleavage of caspase-8 and you can get insight into that from systems biology and this idea of dynamic rewiring isn’t just limited to the EGF receptor but i think is a general approach that you can use to discover novel combination therapies and potentially use this to find new ways of doing drug delivery so that creates both new scientific challenges and can create new IP for old drugs of course what we have to do now is we have to test these systems based insights about dynamic rewiring we have to test this it in in other things like cancer stem cells better mouse models of the types of tyler’s lab generates and potentially in the clinic since both of the drugs I talked about are already approved for the treatment triple negative breast cancer I’ll stop here thank you for your attention and simply tell you that all the work i told you about was really done by one fantastic post doc Mike Lee together with the help from people shown here and it built on worked at alexs gordinho who’s in the audience andreea Tanner and Jerry Alzheimer had done thank you very much I just want to ask one question so do you know what the death receptor is that’s triggered so I we thought that this was due to that extrinsic pathway through a death receptor and we can’t actually find evidence that that’s true um I can tell you that when you look at gene expression it doesn’t look like death receptors in particular are upregulated and furthermore when we take media from each from erlotinib pretreated cells and then put that on fresh cells and then give dr. person it doesn’t cause death so if it’s through that pathway it’s probably intracellular rather than release of a ligand into the music density I don’t know we haven’t tried did you actually check whether doxorubicin gets into the new clothes so for many of the cell lines show dr Wilson actually doesn’t get into nucleus

at also an alternative explanation is that a lot in a pretreatment might simply down regulate the mechanism that keeps doxorubicin I mean that was you know so you can make use of the fact that dr. person is fluorescent and so you can follow the fluorescence intensity increase over time and if you compare dr. was in treatment alone with or without pretreatment it’s exactly the same the dotted curves for one hour the solid curves for four hours we thought about that so there’s no difference in dr. wissen treatment in fact there’s no difference in the amount of DNA damage that it causes and there’s no difference in where the cells are in the cell cycle with the pretreatment so it’s in fact it’s not a difference in the signaling I’m sorry it’s not a difference in the cell cycle stage that or the extent of DNA damage it’s the signaling between the DNA damage in the response that seems to be different it actually question okay i think that one quick question did you try overexpression fad or downregulation so we’re trying to see we’re trying to we’re trying see flip right now I suspect that some of this has to do with c-flip regulation but the problem is every time we knock down si si flip is really tricky because the levels of it you can either get more death or less death when you overexpress it or not get down depending on the extent to which you do it and so these experiments have proved to be much trickier than I thought they would but that’s exactly what we’re looking at haven’t looked grids a good point we haven’t low okay thanks mates