epithelioid cells in mesothelioma

welcome dr elaine fuchs developed this methodology making the breakthrough that stem cells often rely upon their neighbors in order to be able to ploy the necessary factors that are important in controlling their stemness. and so howard green made the use of a

fibroblast feeder layer that then supported the growth and propagation of human epidermal stem cells. when my laboratory first started out one of my graduate students began to work with three dimensional cultures, this is back in the early 1980s where we could pretty

much from scratch recapitulate cultured epidermal cells where they castrate guy. these systems work very well. howard green went on to develop this methodology for one of the first stem cell therapies. we often speak of therapies as something of the future but here is

30-year success. of one of these examples. this is using cultured stem cells where simply took a section of the good skin of the burn victim, basically made sheets and sheets of cultured epidermal cells then engrafted them on to the patient. these

patient's epiderms give us 30 year standing of saying, what happened, with all that propagation going on in culture. the skin of these patients never showed signs of cancers or perturbations. that tell us that stemness can be maintained in propagated conditions under

where the cells do not acquire serious mutations that would compromise. that r that told us early on stem cells that exist one of the things that my laboratory has done is systematically work our way to identify the resident stem cells of sebaceous glands, more

recently sweat glands, also hair follicles. so, the foundation of cultured epidermal cells also served as the foundation for the culturing of human embryonic stem cells and mouse embryonic stem cells. it was five years later that gail martin and her co-workers

adapted what howard green had developed the use of fibroblast feeder layer to embryonic stem cells which are also epithelial-like in nature. this then of course opened the door for mouse genetics or human embryonic stem cells now for ips cells.

a variation on the theme came ten years after howard green's discovery from another post doctoral fellow of howard green's laboratory who made the realization that you could take corneal epithelial cells if there was a patient who had only one blind eye from, for

instance, industrial burn accident, you could take stem cells from the good eye, culture those and this is the blind eye before treatment and this after treatment. and after ten years, 100 patients who were treated in this way ways i cannily had --

basically had their vision restored as consequence of the stem cell therapy. again, another success rate in the field of stem cells for regenerative medicine. not only do stem cells rely upon heterologous neighbors but in fact stem cells of most adult

tissues reside in some kind of a niche. that's true for the hair follicle, intestine and hematopoietic systems. they typically exist in two distinct states. hyperproceed live russ state where primed or activated state

as well as a quiescent. depending on the particular tissue the needs of the particular tissue we find different degrees of quia essence cuiescent. going from one state to another to guy essent state to active -- quiescent

state there has to be a change in the micro environment. stem cells are niches there have to be a change in the signals that come from those niches. that change in signal can come from many different inputs that leads to an activated stem cell which then typically gives rise

to short lived progenitor or amplifying cell but then goes on to differentiate to make the bulk of the tissue. most of the tissue that you see is basically not coming from expansion of the stem cells but rather expansion of their short-lived progenitors.

for me this problem really made stones study it in the hair follicle. because not only is the hair beautiful but it really is the ideal for setting what -- how stem cells are turned on and what turns them off again. because the hair under goes

cyclical bouts of activity. this is a normal homeostasis which might interfere with our interpretation. are the red ones here he, during quiescent state known as the bulge that is arresting state. the after what can be even months sitting in a quiescent

state when the activating of the niche are overriding the inhibitory cues now stem cells from the very base undergo proliferation may generate transit amplifying population now shown in green here which as the cells are growing downward and forming the new hair

follicles the stimulus is being pushed further away from the this is where the bulk of the activity for tissue development. you don't yet know but we've been studying lately the destructive phase, all of a sudden there is apoptosis that

happened in the progenitors and tire system retracts until we get back to this stage again and as i said, the resting stage then can last for weeks upon months. fortunately our resting stage gets longer as we get older and just recently published study on

why that is. again, it deals with this issue of balancing positive and inhibitory cues. to be able to mark these stem cells and really study them in detail we developed some years ago a post-chase animal that we could turn on the florescent

histone. if we turn it on embryo genesis what we find is that initially before turning it off all the skin epithelial cells have their nuclei with the florescence if we turn it off by simply adding tetracycline to the animal's diet wait four weeks that's

where we see this florescent population of cells. dividing next to their proliferative or differentiating neighbors. so this allowed us to segue in to monitoring the stem cells and their niche as i told you when stem cells get activated they do

so at their base. these are the label retaining cells now in green after that histone post-chase. we look at the new emerging hair follicles what we can do ask to monitor the division by factor of two delusion in histone h to be gfp we can see all of the

cells in the emerging hair follicle are florescent they just have less florescence due to the fact that they're dividing now but more florescence than anyone where else can september for the bulge region which is the resident of these stem cells.

these quiescent cells in fact the cells at the base of the quiescent niche, are basically the ones that are giving lies to this early new hair follicle. we now know that both these populations are contributing to the new hair growth. they are both in quiescent state

for weeks or months before getting activated. now we can deal with the transition between stem cell quiescence to activation because in the mouse the initial hair cycles are synchronized. this allows to you look now with short pulses of nucleotide at

what happens when now the new hair follicle is beginning to cycle we see proliferation at the pace close to the sill my husband those are proliferating. so what this tells us is that these two signals, inhibition of bmp signals are critical in driving the new hair cycle also

tells us that even though there is quiescence that there is a considerable cross stalk going on changing the input that basically this stem cells are responding to until finally the activating cues overcome the inhibitory bmp signals that are in the niche.

also know that get too much of a good thing overactivation of nwnts. after creating various different knock-out mice. when we have these cells now florescently labeled we can purify them and profile them and we studied quite a bit the

changes that occur stem cells go from quiescent state to activated state to state where they're fully committed to making transit amplifying progeny. we know that relative to their progeny that there are high levels of variety of different

transcription factors, lgl5 -- dash relative to activated amplifying progeny. we look at the niche in the quiescent state, signals are down and bmp signals. there are overriding cues counteracting that have to build up to tip the balance.

what we also know is that if we were to knock out some of these various different ups transcription factor as my laboratory has done it turns out not to be quite so simple as it seems in the embryonic stem cells where all of the transcription factors are acting

in this case the we knock out lhx2 what happens is that the entire stem cell niche turns out become a sebaceous gland instead of a stem cell niche for the hair follicle f. we repress tcs3 for four signalling or basically repressing wnt signals. there's very high levels of

nuclear downstream target of bmp signalling and calcium. we know that they govern quiescence. the protein methylase govern the switch between transit amplifying cells. we know that without -- basically never form and the --

we don't get hair, epidermal repair we don't get sebaceous glands forming. these teamingly very disparity functions for not only the various different transcription factors but also for the various different epigenetic mark now leading us to wonder exactly how

these regulatory circuits are controlling various different sets of genes. we're digging ever deeper in to this problem. we have real advantage because we can isolate sufficient quantities of stem cells and quiescent state, activated state

and transit amplifying state to be able to do such detailed technologies as chip seq analysis. we're doing this in vivo in stem cells, in their native niche as they're undergoing transitional states. another thing that i think is

interesting that we often don't consider is that while signal very important roles in variety of different stem cells that they don't simply act on the numberly just to change transcription. we know for many years at cell junctions only recently learned

that actually when you lose beta could teen anyone a real defectthat arise in inter--cell could you already junctions. and over activation of -- can increase. now even further. we know that there are also affects now going to cytoplasmic

effects which is glock gsk. there are many kinase targets beyond that of simply beta catenin. we know that some of those can really have an impact on stem cell biology as a doctor has demonstrated through showing that there's one scale regulator

that is regulated by gsk kinase of wnt signals the kinase is inhibited. that allows the stem cells to migrate properly in to a wound site. so again, i think one of the reasons why we're finding so many different outcomes of wnt

signalling depending on whether we look in the intestine or in the hair follicle or whether we look in variety of different tissues is that these factors and signalling not only are regulated by whether there's a positive or negative transactivate or inter-actor for

beta catenin but also by the variety of different gsk targets that are affected or impacted as a consequence. another area that i think is interesting, again i'm trying to give you vignettes of what has emerged in the field over the last ten years and how does our

work fit in to this. is that one thing that we realize very early on as we were purifying these stem cells is that if we looked in their normal native state in homeostasis these hair follicle stem cells will only regrow the hair follicle that's what they

do. and that's what they do in homeostasis. there are other resident stem cells that make these other tissues. what we find if we purify these stem cells takingem out of context and we engraft them in

to a hairless mouse we not only get these nice tufts of hair but also get epidermis and sebaceous glands. something that these stem cells do not do in normal homeostasis. so what is responsible for the difference. based on what i have told you so

far i think you'll not be surprised to realize and reflect that there are two major differences that -- from merely refreshing existing tissue. one is the stem cell niche is missing we're putting this in to a foreign environment and if there are other proper

heterologous cells those cells are capable of inter-acting with the stem cells and eventually reforming the niche. but it takes awhile for that to happen. a lot of things that can happen in the meantime. another interesting

consideration is tissue progeny themselves are missing. and in fact one of the things that my lab has learned is that in fact tissue progeny can impact dramatically on the behavior of the stem cells themselves. these are basically downstream

from emerging from the stem cell when you consider the whole aspect, for instance of wound repair, where we get a wound and of course we know that the wound can be repaired, our stem cells get activated. don't expect to see a tumor growing at the wound site.

something that is telling the stem cells to stop making tissues, and what better way to do that than to have tissue progeny downstream feeding back to the stem cells to say, okay, stop. i don't need any more of your activated cells.

in fact we discovered that differentiated cells can hone back to that niche and provide bmp signals to further raise the threshold. think about whether this applies to many different systems i think that if you look at the literature over the last couple

of years there are now a number of emerging cases where it looks like this type of feedback mechanism is coming in to play both from the heterologous signals but also from autonomous signals. just to give you one example what studies fly hematopoiesis.

is a downstream differentiation signals that goes back to the cell in the fly that tells the stem cell to stop its activation. that is very similar to what i've just described for the hair another interesting aspect is that this system is perturbed in

cancer and in fact you can think about cancerous tissue now this system doesn't work properly, when this system doesn't work properly these signals aren't provided back. in fact we realize several years ago, in fact for every time we started to look at what turns

stem cells on and what turns them off again, begin to realize when we over activated up getting cancer and under activated the inhibitory queue, is that gave us cancer. we began to realize in fact tumor initiating cells have pretty much hijacked the natural

mechanisms the stem cells use to control their behavior to be able to control tumor genesis. thiole us to the question of where are the cancer stem cells what do they look like, we tackled this problem by starting to systematically dissect out fact purify various different

populations of the heterogeneous squamous cellars know mas to got down at the single cell level in serial transplant assay. a single cell was sufficient to be able to introduce it in to a host recipient mouse get a cell that resembled the parent. this can be done over and over

again that tells you that you've got cancer stem cell and then we wondered what are these cancer stem cells look like. well, light normal stem cells they reside at the tumor stroma interface. so they, too, have their own they express high levels of

integration like normal stem cells and they express high levels of self renewing. like normal stem cells. in fact we've even learned the difference between what controls the balance between those two quiescent states. enjoy muss cell there is high

levels of t gi beta. and long as the cancer stem cells can respond to tgf beta that makes them less proliferative. if the cancer stem cells lose their ability to respond to tgf beta signalling now the number of cancer stem cells goes way

up, goes up by factor of tenfold and we lose those differences between proliferative states of the two cancer stem cell types. but if we look at the similarities that's about where the similarities end. in fact if we profile the cancer stem cells we find that there

are more than 700 differences that exist between the cancer stem cell and either hair follicle or epidermal. they don't look anything like normal stem cells. it has a different stroma, immune inflammatory. coming around the towel or they

have different niche also have different progeny they're going to have different signals and different properties. in fact when we look at this there are many different cancer genes, tgf alpha is high. veg f is high. what we find as a twist -- the

transition is high. catenins are down regulated. it's not a characteristic of as we began to look at all of these different aspects we began to realize that we can't systematically take a look at these 750 or some odd genes are really relevant.

can we take our conventional genetics approach, i don't think so. certainly not to be able to go through the functional significance of all of these different changes. so, just really became a challenge for us as we started

to move to the human genome sequencing era. and began to deal or try to grapple with all these hundreds of changes. if you look at the cancer sequencing genome products. the most life threatening of all of the joy ma cell carcinoma s.

but there are hundreds of coding mutations occurring more than two patients or more than one patient in these head and neck cancers. they have the same kind of problem ever how do you sift through the hundreds of mutations to find out which of

these are actually drivers for the cancer and which ones are mir bystanders. we can reference the two data sets, we can reference the human sequencing efforts and we can reference our epigenetic changes we still get hundreds of changes.

let's compare the methods. we get human data if we look at rna sequencing as we've done with cancer stem cells versus normal that gives us epigenetic changes as well as mutations so that expands the repertoire of types of gene changes that we get but of course this is

focusing on mouse, you could legitimately argue there are differences between mouse and human cancers. still of course the different species. the challenge really for both of us is how do we actually sift through these hundreds of

changes to really identify which ones are important and really to stop cherry picking which we can do, we're likely to overlook something simply because we don't know enough biology. ideally what we'd like to do is harness the power of fly and worm genetics for the mouse to

be able to get our cake and eat it, too. where we don't have to go to a worm we're trying to figure out if we knock out a gene how does this relate to human joy muss several years ago. a post doctoral fellow who is now at the fred hutchinson

cancer center and another who is now post doc in scott lowe's laboratory, seemed together. to knock down genes rapidly we ride rnrnas then came to working with human culture or mouse actual tougherred stem couldn't gel oligo nucleotides in.

viruses, viruses worked pretty good. in fact when they came on the put genes in to the cells. then we wanted to get them in vivo. started to scrape the skin and apply lenti virus. that didn't work very well.

tried needles, the virus went in to the very first cells that it saw. louis started working with organ transplant cultures. none the other layers got transduced. it went in to the bottom layer. none of the other layers got

until it finally dawned on us one day that right after the embryo exists as -- epithelium exists as single layer that covers the embryo surface. perhaps if we could simply expose the embryo injecting the amniotic sac and leaving the animal in fact f. we can do that

in vivo might be able to simply put mom under anesthesia and inject the sac with virus and get the transsection we were seeking. that remarkably worked well. this shows you six days later the embryo surface. we get very good transsection

greater than 90% in the head region. then once lentivirus. it's completely untouched is below the surface epithelium. so this now without having to use any -- gives us tissue specific gene express as

consequence of learn tie viral specific in sec tift and the learn tie virus. so, where can we take this. well, we could put any hairpin we want now in to the lentivirus and knock down any gene that we want. that started to give us good

results. we published several different papers in the early 2010 decade, 2011, 2010. that illustrated the principle and demonstrated that once the lentivirus goes on once it integrates and stabley propagate that these different genes that

are harbored within the lentivirus don't turn off so we can monitor them in the adult. reason for that is that the mechanisms for balancing viral genome basically is happening much earlier than 9 1/2 days of so, that worked nicely. we can knock down genes and

genetic pathways. we started to look at asymmetric cell division. we could see for instance if we knock down a gene it affects the p5 pathway we can knock down p5. we can start to do very complicated genetics, even a year ago we couldn't even dream

to think about before we had this technology. but then i was dealing with a post doc who trained in slide g, they think big. the slide geneticist thought maybe take this one step further. the first thing that we wanted

to do was to see just how uniform is growth on the embryo surface. i thought for sure growth would be very variable on different regions of the embryo. you put in a low tighter gfp tag lentivirus, i thought followed expansion basically -- turns out

that doesn't happen. within a factor of one to two cell divisions when embryo is born all the clone sizes are similar to one another. clonal expansion is remarkably uniform. if you take 100 stem cells and you put them in to culture we

don't get 100 cole he's of all the same sizes. some colonies never grow. some grow big, some grow small. the reason is that you're subjecting the cells taken out of context to all sorts of artificial conditions, submerging them.

growth factors, they are subjected to enormous amount of stress you have variety of different stress responses turned on. some stem cells make it and others don't. but here we're subjecting the skin surface to all the natural

systemic, heterologous,, niche environment responses that these stem cells normally experience. and they grow at remarkable uniform rate where we're not subjecting the surface to a lot of stress. so my post doc thought, why not simply take entire mouse genome

and set up an rna i screen using lentiviral technologies taking the entire 78,000 lentiviruses from the using of course if you screen for genes that negatively affect growth, if you were to screen for that in just one animal, in a normal wildtype animal.

what you get are lots of ribosomal rna genes. lots of housekeeping so-called genes. if you screen for differences for lentiviruses or hairpins that preferentially inhibit oncogenic growth versus normal growth which is what you'd

ideally like if you were designing a cancer drug then effectively that worked beautifully all the hairpin -- all of the housekeeping genes basically fall out of that equation. what you are left with then is a handful of some very interesting

genes that turn out to be relevant. you have to obviously do years of controls, setting the viral titers, setting the representation within the pooled library. you have to make sure that your coverage is good that is to say

enough of the cells on the embryo surface get exactly the same hairpin. you have to make sure that on average only a cell gets a single hairpin. when do you that, you can take a look at three different replicates, the top ten hits

were exactly the same in all three replicates. if you take the top 400 hits turns out that more than 70% of those are in -- coming up in the same list in all three replicates. that was surprising. i never thought it would work as

well as it did. obviously there are going to be off-target a picks but taking advantage of the fact that we have five hairpins per gene and so what we're looking for are cases where there are more than one hairpin giving us the exact same differences and none of the

other three hairpins give us the opposite difference, we're also then looking for dose responsiveness, of course you also ultimately have to check to make sure that the genes that you're choosing are genes that are relevant. our top hits by the way was beta

catenin which is well-known oncogene we went on to study that and demonstrate that it was physiological relevant as was another. more systematically now pursuing this. now with technology that makes our knock out demonstrating for

conditional knock out much more accelerated for the future which has been rate limiting step of basically getting this far once you get your hit you have to validate them. so, now i want to move in the last few minutes then to this new work that we've been doing,

i'm going to cover it relatively quickly, it's coming out in the next week or so. i encourage you to take a look at the paper or watch for the paper if you're interested. this gets back to this issue of the human genome sequencing project and our problem with the

different cancer stem cells. so what we did was we said, why not simply make a smaller pool, not 78,000 lentiviruses, but let's simply screen a thousand only now we wanted to look for tumor suppressors and you don't really want to look in the embryo where it's growing very

fast, great model system to look at negative regulators of growth not good one to look at for positive regulators. in this case what we did was we used a lower tighter, smaller number of genes basically we did a control library versus the normal -- our test library we

wanted to put it on a tumor prone background we used tower tgf beta receptor null mice which don't show any signs of squaus cells carcinoma. and then we simply said, wait for tumors to develop on these animals where they don't develop in the control and basically

let's just clone out and find out which hairpins are being selected for. i'll give you the punch line here. the punch line is that myosin, this is the -- myosin 2a turns out to have hairpins that are present in greater than 30

squamous cell carcinoma, is that came out of our library screen and four of our five hairpins came out in these different tumors. when we went back and checked them they showed ghost responsiveness. the most tumors is with the most

potent myosin. then we tested these individually. they showed dose dependent tumor control. that's just example of some of the tumors when tested i didn't believe it at this stage because we've been working

on this for a long time. i didn't think myosin would be included in this kind of list. let's make the conditional knock-out mouse which we did and we tested this. sure enough. it turns out that myosin 2a is acting as tumor receptor.

in fact even at the heterozygous level it has some potency which turns out to be i think quite interesting. then we looked in other tumors that came out you start with the single layer epithelium that gives rise to the mammary glands, for instance.

what we found tumors coming out and if we stain four myosin 2a we found dramatic reduction, normally i don't know if the controls to show you there would be much stronger, you can see that these are invading tumor cells and that they basically are expressing carotin 14.

we looked at the various different tumors, used the different tumor markers, getting squamous cell carcinomas what is going on here. i say, maybe just tumor background. let's do it on dmba -- tpa classically -- our myosin

knocked down animals treat them now leaving tgf beta out of the question. and again we're seeing that relative to controls we're getting higher levels of tumors. the tumor frequency is increased as well. then we did it on oncogenic rat

background, same thing. on p53 we didn't get any change in tumor genesis. we wondered what is going on there. maybe in the same pathway somehow. maybe operating in the same pathway.

it was one possibility we could test it relatively quickly. we know how it's degraded in normal homeostasis we know how to induce p53 in the dna damage we could easily do these dna damaged tumor responses what we found is that actually if you do that what you find is that when

you knock down buy is in in every case we knocked down myosin then did a dna damage inducibility what we found that the normal increase of p53 doesn't map. basically got very impressive if you look at p53 target that, too, doesn't happen.

we looked at article whole slew of them. then i thought let's just do it on the knock out. maybe this is not going to work we took our knock-out mouse cells, same thing. if do you a dna damage response basically p53 doesn't stabilize

in response to the dna damage response pathway in p21 is not up regulated. here is the whole group of them. basically show that p53 target genes are not getting activated. in the absence of high is in 2a in this case we're dealing with the tumor genetic model, is that

we generated. so now we looked at what downstream -- where is myosin acting downstream and to make a long story short, it turns out two things. one is that if atp ace domain is important we look actually i can see that nuclear accumulation of

p32 in response does not happen properly in the absence of myosin or myosin atpase blockers. we looked at the pathway. we thought maybe this is just culture artifact. let's look in vivo here is in vivo if we take mouse you

irradiate it what you can see nuclear p53 if you take the myosin conditional knock-out mouse doesn't happen. i was surprised. in any ways we said, what is the relevance, is there any relevance to human squamous cell carcinoma i knew my cancer

friends were basically going to challenge saying that's nice for your mouse but what difference does it make if it's not we started to look at human it's really interesting that actually there's wide variation with regards to myosin expression.

we did the carotin expression controls which i don't have on this picture. a wide variation, carotin 48 is and then we looked at 54% of all of our squamous cell carcinomas had no or weak myosin to staining 79% of head and neck then what does that have to do

with prognosis or pore survival turns out that the ones if you look at which ones have no staining of myosin then you look at prognosis that basically very poor prognosis correlates with low levels of myosin staining. then we said, let's go back to looking at the mutations so,

that's valid. i don't mind. we have -- we're relative newcomers to this. it's true that when you look, myosin is big gene. so if you look at the p value it looks great then look at the false discovery rate.

and if you just look at all the myosin mutations, it's not so great. that would make someone at the institute probably, immediately going to say, see, i told you it's not statistically significant or not relevant. if it has a high false discovery

rate. wait a second. let's take a look at what we know about myosin and let's take a look where those mutations reside. then if you look where those mutations reside turns out that they -- turns out that they

cluster in the atpase domain. one of these human patients has amy sakes and -- in residue. myosin 2a has to make that mutation as conserved residue. it compromises atpase domain. i like this story. if you like it, you can read all the details in the next couple

of weeks. but that's what i wanted to tell you. your the first audience of guinea pigs you can barrage me with a lot of criticism on this but i think it's really cool. this is kind of where we've come over the last years and we're

thinking differently about how we study skin, skin stem cells and how we deal with the myriad of differences and changes, is that francis collins created for us some decade ago or so. in completing the human genome project. francis, thanks, you made us

come up with new ideas how we're going to get around these problems. i've given them credit for the ones responsible. but i'd like to thank francis again for the invitation, thank julie and basically happy to take questions.

thanks. >> exciting work. do they have -- so that they could participate in the generations. you have shown nicely in the model, is that they do generate. what happens to the human. >> in humans the stem cell --

you mean are there equivalent stem cells in human hair follicles or what happened to the pun knee hair follicles on your body surface. can you folk is the question -- happy to answer. >> that we could transplant the cells and grow --

>> why can't we do that in vivo. now you are asking very good so, if you can take purified stem cells and you can get hair in the mouse why can't we do it in a burn patient, for instance. there i think what it's telling us that we're missing a critical niche signal.

and because otherwise it's very hard to explain. otherwise you're dealing with fundamental difference between human stem cells and house stem cells versus a missing niche signal that is also present that it was also lost in the burn which in a nude mouse the nude

mouse has everything, ready to make hair. you just have to give it a functional stem cell that can do it. where as burn patient you're missing a lot of other parts of the determine muss. what we're now doing to try to

gap well that issue is to come f all the various differenting niche signals and trying to identify those. , i think resultswe also have some pretty which i didn't talk about. but that kind of give us some queues about what might be -- what might be happening in

aspects of hair loss, too, which i think could also be i think in the end it will apply. just that with regards to burn patience i think -- experiments are telling us that we're missing. >> very nice.

by way of confirmation, my lab has knocked out -- i wanted to ask you there is -- not uncommon called myosin 2a related to these people -- these mutations included including atpase activity. have you been able to look to see if these patients develop

increase in squamous cells? >> very good suggestion. if it's myosin 2a one of the -- other thing that i didn't looked at all three different myosin 2s. turns out that the p5 affect is we don't know yet whether the differences, part of the reason

not knowing exactly what the difference in mechanism is -- don't know what domain is playing a role in this process. we don't know how. it could be part of the natural mechanism of active myosin that is influencing nuclear export which could be interesting.

but until we know that it's to be hard to really address those kinds of questions. just finding out some of these other diseases which are myosin disorders, whether there's actually predisposition to any kind of tumor genesis going to

be interesting to take a look at. >> over here. >> a question from your last part of the talk. lead to the cancer cells. do you think any factor -- >> one of the things that we're surprised by, actually quite a

considerable level of myosin 2a that shows up in the numberly just. our fractions are pretty clean. and so first i thought, just contamination. nothing is coming up with the it's nuclear myosin. we have been very cautious about

saying whether that's what's happening, what we know it's nuclear export. we know that with regards to p53. p5 there for nuclear inhibitor on myosin 2a background then we find that now p5 can respond to dna damage, it can stabilize, it

can accumulate. and it can activate downstream target. checking in to whether that might have therapeutic values. i think that's a curious possibility. but in terms of its other potential affects for us the

only thing that clued us in to looking at this was that as we're trying it out, i thought p 53 and didn't have any additional affect. which is what led us to think that maybe we're dealing with some aspect that's in the same so it could be that myosin if it

is nuclear is everything multiple different affects or it could be that something like nf cap type activating mechanism that is happening in the cytoplasm that could have additional affects. but this in and of itself i think is enough to explain the

tumorigenic phenotypes that we're seeing whether it's all that explains it is completely open question right now. >> one more question. >> i think that this large tumor suppress or isn't that connected to myosin 2. >> yes.

there are a lot of suspicious characters out there that are connected. i think again being interested in a skeleton my whole career i think that the whole field i think is kind of gotten tied up so much with all of those nuclear factors for so many

years. and i think there is going to be a lot more to meet the eye than what we have seen. i think the other aspect that i find very fascinating about the genome sequencing efforts, and etch genomic efforts is that these afford the opportunity of

looking for low penetrants players in the game. it may not be a common player that allows to you look for low penetrants players where we start from a functional assay and then look forward as opposed to the traditional way to look at something that's very

prevalent in cancer. there you can really start with the genetics and pick that apart. >> an actin binding that is associations with pancreatic all kinds that have been ignored. >> good point.

>> fascinating presentation. thank you for the hot new science you've allowed us to see for the first time. let's thank dr. fuchs again. [applause] now be reception in the library for those of you who like to continue the conversation with

coffee and cookies and dr. fuchs will wander over there just a moment.