Cycle de conférences 2023-2024 : Exploring the World of Protein Post-translational Modifications
Conférence du 27 mai 2024 : New Insights into Intercellular Crosstalk between Different Cell Populations in Pancreatic Tumors That Can Be Translated into the Clinic
Conférencier : Tony Hunter, Salk Institute, La Jolla, California, USA
Retrouvez les enregistrements audios et vidéos du cycle et son texte de présentation :
https://www.college-de-france.fr/fr/agenda/conferencier-invite/exploring-the-world-of-protein-post-translational-modifications
Chaire Oncologie cellulaire et moléculaire
Professeur : Hugues de Thé
Retrouvez tous les enseignements du Pr Hugues de Thé :
https://www.college-de-france.fr/chaire/hugues-de-the-oncologie-cellulaire-et-moleculaire-chaire-statutaire
Le Collège de France est une institution de recherche fondamentale dans tous les domaines de la connaissance et un lieu de diffusion du « savoir en train de se faire » ouvert à tous.
Les cours, séminaires, colloques sont enregistrés puis mis à disposition du public sur le site internet du Collège de France.
Découvrez toutes les ressources du Collège de France :
https://www.college-de-france.fr
Soutenir le Collège de France :
https://www.fondation-cdf.fr/faire-un-don
Suivez-nous sur :
Facebook : https://www.facebook.com/College.de.France
Instagram : https://www.instagram.com/collegedefrance
X (ex-Twitter) : https://x.com/cdf1530
LinkedIn : https://fr.linkedin.com/company/collègedefrance
[Music] [Music] so good afternoon everyone it’s my pleasure and honor uh to introduce Tony hunter from the sulk Institute for his fourth and La last lecture uh today on uh P pancreatic cancer Tony thank you for being here and the floor is yours thanks you well it’s down to my last lecture and this time I really won’t be spending much time talking about post transation modifications except in passing but I thought I would start this lecture as I’ve started every other lecture with my little bit of personal life history here I am was a new graduate student in 1965 some evidence I actually did experiments and then once I got my PhD in 1969 I started to grow my hair out and in 1971 I went to the S Institute to do a post-doctoral study with Walter ehart on polyoma virus and after 2 years I came back and there had been a dramatic transition uh I grew my my beard the story is that I last shaved when I walked down into the Grand Canyon to join a rafting trip in July of 1972 so my beard turned 50 a couple of years ago now I looked for job in England but failed to find one but luckily they’d offered me a position at the so Institute and I came back in 1975 and that’s where I have been ever since now I thought I’d just continue for one more slide on the beard theme these are two images um made by former post dos of mine Randy poon who worked on the cell cycle in my lab a very talented oil artist actually and this is a cartoon by Dave midlas uh you probably can’t read this necessarily but I can phosphorate therefore what does it say therefore I am apologist to decart here this is actually phosphotyrosine here in in the red in case you’re wondering what it is and then final beard picture I went to the Galapagos a couple of years ago now and this is the Darwin Museum on the Galapagos um Darwin was was at my college in Cambridge and so this is somehow appropriate that there I am standing next to his um mannequin here I’m trying to get some inspiration anyway as I said I wouldn’t really be talking about post transation modifications but for the last three lectures I’ve been talking about fos foration of cine stenine and tyrosine and histadine and also extensively about ubiquit alation of Lysine now the point in showing this is that there are other post transational modifications out there waiting to be discovered and so um even though there were over 400 known um certainly new ones could be discovered and there will plenty to do to understand how they work but instead what I want to do today is talk about our work on pancreatic cancer which began about almost 15 years ago now when I joined a standup to cancer pancreatic Dream Team to study pancreatic cancer and the reason for studying I mean I’ve been studying cancer my whole career but not really at a translational level but the reason for studying pancreatic cancer is that it has a grim prognosis it’s almost always diagnosed too late you can see here over 50% are diagnosed when the when the cancer is already metastasized and it’s the third leading cause of cancer death cancer related death at the moment even though it’s only the 10th most common cancer and this is anticipated to increase to the second most uh second most caused reason for cancer death unless a better therapy can be uh developed currently the one one year survival is only 20% of patients and 5e survival is a miserable 11% now we know quite a lot about the genetics of pancreatic cancer it under goes a series of both morphological and genetic changes over years probably many years the initiating mutation is thought to be an activating mutation of the kras enor protein and then the this is followed by loss of function mutations in p16 AR cdk2 n2a the cdk inhibitor and the p53 and smad for uh tumor suppressor proteins and you can see that there’s additional mutations are required the morphology um there’s an EOD ductal metaplastic transition um through a stages of panins pancreatic cancer in situ and then finally Beyond panin 3 cells begin to leave the primary tumor and invade into the local tissue uh where they can be disseminated throughout the body by intravasation into the bloodstream and then finally extravasation to form a metastasis in this case most commonly um in the liver now I should remind you then that solid tumors are a community of cells they’re like a a tissue if you if you want which contains not only the tumor cells but many types of stromal cells immune cells endothelial cells the blood vessels that supply the oxygen and nutrients to the tissue and it’s becoming realized then that these cells these different types of cells talk to one another to maintain this this tumor tissue both in the primary tumor and in subsequent metastatic tumors and we started working on a A specialized cell type in in the pancreas called the pancreatic stellate cell because of their star shape the quiescent cells resident in the normal pancreas so you can see they may be up to 4% of all pancreatic cells and they’re sort of Prof professional surveillance of the tissue they are looking for damage inflammatory signals and cancer and in response to such signals these stellate cells or PSC transform into a myofibroblast like cell uh sometimes called a MF or myofibroblastic cancer Associated fiberblast which then begins to secrete Matrix mostly collagen based Matrix but whole range of cyto into the local environment around around the U incipient tumor so there’s been a lot of interest in these cells which give rise to to one class of cancer Associated fiberblast so the question is why is pancreatic cancer so hard to treat well it’s a very stromal tumor this is sometimes called a desmoplastic response and probably the tumor cells are only 10% of the mass of the actual ual tumor tissue the tumor a nest of tumor cells and is surrounded by this dense Matrix contributed by the cancer Associated fiberblast and one consequence is that the key immune cells the cd8 positive tea cells that are excluded from the tumor tissue and this is really a wound healing response that would occur say for instance in response to pancreatitis and it’s often been said that cancer is the wound that never heals so this is really a wound response it’s it’s Walling off the T the tumor and if you in fact if you eliminate these surrounding calfs genetically the tumor gets much more aggressive it’s been shown in Mouse models so they are serving a purpose but they also make it difficult to deliver therapeutic agents antibodies and small molecule drugs to the tumor cells so in a little more detail then we have uh tumor cells in the middle often forming a lumen there are um INF infiltrating immune cells um macrofagos neutrophils and myo derived suppressor cells but as I said the cytotoxic tea cells are excluded and these tumors are poorly vascularized here so we want to understand the cross talk between the different cells in in the pancreatic tumor environment and particularly we’d like to know what factors are secreted by the cancer Associated fiber blast and and the immune cells in in the tumor tissue that lead amongst other things to exclusion of of tea cells and the generation of an immunosuppressive micro environment and sadly it’s true that pancreatic tumors almost never respond to checkpoint antibody therapy so we set out to try and identify specifically soluble protein factors that were generated by the cancer associated fiberblast in pancreatic cancer that would act on the tumor cells and potentially other cells in the micro environment to promote tumor Genesis we already knew there are factors made by the T tumor cells such as TGF beta and pdgf that act on the CATE cells to promote their transition into myofiber blast and the story I’m going to tell you right at the in the beginning part of my talk is our discovery that leukemia inhibitory factor that I’ll be calling LIF plays a driver role at least in Mouse pancreatic cancer models so to study what secreted proteins might be important Yushi a postdoc in my lab and his high school classmate ran Tian who at the time was a postdoc in Tony Paulson’s lab set out to identify the or characterize the secretomes of a pancreatic human pancreatic tumor cell line M Paka 2 and Cals culturing these cells in serum free medium and then collecting all the protein and carrying out um Mass Spec analysis in in duplicate so you can see there were 2400 proteins in the M Paka 2 secretome around 1,400 proteins in the calf secretome and then looked for proteins unique to either one or the other Secreto they’re around 400 that were unique to the secret home of the tumor cells around 170 unique to the fibroplastic cells and then we focused on proteins that we thought would serve as signals like growth factors cyto kindes chemokines but I should point out that many of the unique proteins secreted by the calfs are proteins involved in extracellular Matrix functions but really we focus then on factors unique to the to the myofiber blast and in particular we focused on LIF but you can see for instance that pdgf secreted by um the tumor cells which we know acts on the fiberblast and TGF beta secreted by both cells which also acts on the fiberblast now why do we why do we focus on CF well like all cyto kindes like I six of which it’s a a family member it induces stat three signaling which we knew knew was elevated in pancreatic cancer and is fairly easy to study uh immunohistochemically but more importantly LIF is a stem cell factor it’s well known as a factor that can maintain Pur potency in embryonic stem cells and this seemed like a property that would be useful to um a tumor and in its normal function it creates an IM supressive micro environment so what we showed was then that the condition medium from the calfs was able to stimulate phosphorilation of stat 3 in the tumor cells in culture and we showed that the condition medium elicits binding of stat 3 to the lift receptor and its co-receptor and that a neutralizing anti-lift monoclonal antibody can block most of the stat 3 uh phosphine 705 signal induced by the condition medium and just to show you that piece of data here um we have the U this P I think it’s kp4 cells another pancreatic tumor cell line stimulated either with lift or condition medium or nothing and you see a strong signal here and now if you include a new neutralizing anti-lift antibody in the conditioned medium uh or in the medium added with lift you can see essentially all of the signal goes away so this suggested to us that at least one of the factors in the condition medium that was activating stat signaling was um was Li now it turns out we were somewhat fortunate here because the two cell lines we’d studied pancreatic cancer cell lines we’ studied for their secretomes Mia Paca 2 and pank one turn out to have Express very low levels of endogenous lift but there are certainly are other pancreatic cancer cell lines like aspc1 and kaen one which have rather high levels of lift and I’ll come back to that in a minute uh you can see that um the cell cells the calfs have have high levels of lip so in addition to these properties I’ve just told you about if you look at the um tcga catalog of RNA expression in human tumor types the 20 human tumor types that they surveyed pancreatic cancer or known as PAAD here has much the highest level of lift of all the pan of all the human tumors this is at the RNA level now particularly compared to the normal pancreas which has very low levels and another thing that we noted was that high LIF RNA expression um from early tumors these are tumors that probably had not metastasized uh corelates with poor early stage prognosis and so you can see with LIF RNA High very poor prognosis compared to LIF RNA low so what what is lift known for well it got its name in the late 1980s through its ability to inhibit the differentiation of M1 leukemia cells suggesting it has a role somehow in the immune system in regulating differentiation but it’s probably not the best name that could have been chosen but this is the name we’re we’re stuck with and it’s a highly plot Tropic cyto kind that does many things to many cells um and it’s expressed by many types of cell in the body including stem cells themselves seyes Myas Etc including myoid cells as I’ll show you towards the end now surprisingly uh lift knockout mice are viable both males and females and the main uh phenotype they show are fertil fertility defects so the the female lift knockout mice produce fertile eggs they can’t implant because LIF expressed by the endometrial cells is needed for blast assist implantation it actually induces an invasive phenotype in in the embryo or the blasticus to penetrate the the wall and it also creates an immunosuppressive micro environment to prevent attack of the mother’s immune system on the embryo the lift knockout males are viable but infertile they don’t produce sperm as I said lift promotes self-renewal and plur potency of many stem cell types now the lift receptor is a little more complicated not only is it a co-receptor so it binds lift but needs a co-receptor to Signal which it does largely through jack stat and the KK map kyes Cascades but the lift knockout mice die perinatally so it isn’t essential Gene but you can get there are conditional lift knockout which we used in fact um so as I said lift is part of a heterodimeric receptor complex but it’s the receptor not only for LIF but also several other cyto kindes cntf ct1 CLC and alason M which I’ll talk about at the very end uh which also has a separate receptor the osmr so all of these side signal through the lift receptor complex and activate Jack stat signaling and what we don’t really understand I think this is true for many cyto kindes is how different cyto kindes can use the same receptor system and yet generate a different signal in the cell something I think we all we need to think about more and if one wants to try and antagonize lift lift receptor signaling targeting the lift receptor um might come with liabilities because it might be blocking signals through the other side of kinds that can recognize it so this is what the family looks like um these are all the ones in plus lift that use the lift receptor itself but each of these receptors has a unique Lian binding subunit like is 6r and I 11R that heterodimer izes with with gp130 and the lift receptor but all of them really do the same thing they activate um Jack stat signaling uh here by binding the Jack kinases phosphor the receptor to induce a stat s stat um binding through its sh2 domain and then dimerization of the phosphos stat and lift signaling to the map kyes pathway and the pi3 k a pathway can be transduced through uh ship two binding it’s interesting that the LIF recept has a much longer tail with several tyrosines in it and we wonder whether that’s important for what it does in the cell so with this this evidence in hand then we decided to test whether lift was playing a role in pancreatic cancer and for this purpose we used the mouse model that uh Dave terson’s lab Dev developed in 2003 um hurani as the first author and we chose to use this version of the model in which a lock stop locks g12d oncogenic Al of Kass can be activated by cre exision of the lock stop locks and where both alals of tp53 are flocked so that when is expressed in the pancreatic an lagaga um kras g12d expression is activated and p53 is deleted these mice um develop a very fulminant pancreatic cancer it’s already evident when the mice are born that there is something happening in the pancreas um and their half half time of survival is 50 days after birth so a very aggressive disease but we wanted to see if neutralizing lift could um improve survival of these mice now the first thing we did then was to look to see where lift and lift receptor are expressed both in the normal Mouse pancreas and in the tumors so we’ve used RNA scope to color Inu RNA Inu hybridization so if you just focus really on the bottom panels This is the uh signal for Lift lift re for Lift itself and you can see that most of the lift signal is in cells surrounding the blue keratin 19 positive tumor cells although there are lift positive tumor cells consistent with what’s been found in cell lines the lift receptor in contrast is primarily expressed in the uh tumor cells and the not in the blue periostin positive uh cells surrounding the tumor cells so that fits with our model and the LIF is being made by stromal cells and acting on the tumor cells which have lift receptor so this is the model we were pursuing then that the stellate cells in this case the myofiber blast um are making lift that’s acting on the tumor cells which may also make autocrine lift and then the tumor cells make other factors that act on the stellate cells to convert them into myofiber blast now importantly for what we wanted to do ultimately the same pattern of lift expression is seen in human pancreatic tumor samp so this is a a tumor sample again using RNA scope 2 color Inu hybridization for cartin 19 which marks the tumor cells and and lift and you can see that there is some lift signal in the tumor cells here in the blow up as well but the most of the lift signal is coming from the surrounding Str stromal cells so with that uh information we went ahead and set up a a preclinical test in the KPC Mouse model that I described to you where we treated the mice with 25 migs per kig three times a week with this neutralizing anti-lift antibody d25 which was actually made by Genentech as you can see a long time ago now for another purpose Al together and they put it into the public domain so it’s available uh and the control was a normal ice type matched HG so here you can see in the control strong nuclear staining of cells um but the tumor cells and actually some surrounding cells um for phosphat 3 uh in the mice treated with lift antibody you can see much of the nuclear stat signal has has gone away consistent with lift driving some of the stat 3 phosphor stat 3 signal and the uh one can neutralize it with this antibody so that led us then to do this sort of preclinical treatment model where we treated the mice um with the lift monocon antibody 25 migs per kig starting treatment at 32 days of age and at the same time included with or without gem cabine which was a standard of care nucleoside analog therapy for pancreatic cancer it really doesn’t work very well but we included it as a as what you might decide to do when you’re treating a human cancer patient and what we found was that um here’s the control again the mice dying around 50 days if we just use the lift antibody we get a modest extension but that if we include a com in the combination of gemide of being an anti-lif antibody we get quite a healthy increase in survival I agree it’s not dramatic itic but it is it is an increase in in a model it’s difficult to um to treat so that that’s what we concluded then that the lift antibody can slow tumor progression and potentially enhance the response to gem cabine we carried out a second type of of preclinical treatment regimen in which we pre-treated the mice for one week with a combination of pxel a tubulin drug G cabine and CIS platin a DNA drug which is what is commonly used in patients and Then followed that with the gem cabine anti-lift antibody um treatment protocol and now you can see that particularly with the gem cidab beine plus anti-lift antibody we get a quite strong quite a doubling in the increase in survival so that’s consistent then with the earlier slide I showed you and suggest that maybe lift blockade could be of therapeutic benefit now we wanted to be certain that this was really dependent on on the lift receptor and so we used the lift receptor knockout mice which we bred into the KPC background and now you can see that just using gem cabine alone when both Ali of the lift receptor are knocked out locally in the pancreas in the in the tumor cells um you again get this increase in survival we’ve not done the experiment I must admit with the conditional lift knockout because there was no lift flocked Al available then there is in fact now so this really recapitulates um conclusion we did with we had with the antibody now just a couple more things we we did in terms of analyzing what effect the antibody had on on tumor progression we carried out um pathological measurements and showed that gem cabine plus the antibody resulted in in a an increase in areas of the tumors as gauged by three independent Pathologists that were moderately uh differentiated suggesting that one thing that lift does is potentially to slow um dedifferentiation we found that there were fewer cells tumor cells epcam positive tumor cells that had cd133 or alh stem cell markers quite a significant decrease and perhaps more importantly that um epcam positive cells were severely deficient in their ability to form spheroids in culture or to to form flank tumors when reintroduced into into mice suggesting that one thing that that um lift antagonism is doing is to reduce the number of stemlike cells um in in the tumor we also showed from RNA sequence analysis that lift blockade suppresses EMT expression remember EMT is a conversion to a my a fiid plastic like cell a mesenchimal cell which allows the cell to become invasive um and interestingly also increases the expression of these two transporter proteins which um transport gem cabine into the cell tumor cells which may be the reason why the Synergy between um the two the two drugs now if we look at a progression we find that the levels of lift rise significantly note this is a log or logarithmic scale here rise significantly um during the course of seven weeks in the mice and by week five we can actually begin to detect circulating lift in in the serum of these mice which I’ll get back to in just a second this was done by aluminex Elisa assay so this the next step one then was then to ask whether the are similar changes in lift expression in human pancreatic cancer and so we collaborated with several groups who um treated pancreatic cancer patients and who did um either biopsies or autopsies on on pancreatic cancer patients to analyze for levels of of lift in tumor samples compared to chronic pancreatitis samples and normal pancreas and you can see that the majority of pancreatic cancer samples had high levels of high levels of of lift which were actually confirmed independently by using PRM ASPC analysis targeting lift done by Ren Tian Now using a a higher sensitivity digitaly of the quantx system we could detect um high levels of lift again this is now a log scale high levels of lift in in many pancreatic cancer patients but I must admit these are not matched samples so they’re different patients but so this suggested maybe lift could be used as a as a diagnostic or even a prognostic marker so as was the case in mice you can see that if you look now at these samples where pathology had indicated they were either well moderately or poorly differentiated you can see the poorly differentiated tumors have high levels of lift and there is a you can see as a Trend here not a very strong Trend but there is a trend in addition because of we were part of this standup to cancer Dream Team who were treating pancreatic cancer patients we collaborated with UT borans and Dan Vanhoff at honor Health in in Phoenix where they pre-treated um patients with a braine which is the NAB pxel tubulin drug J havein cisplatin paricalcitol um this case an anti pdl1 checkpoint antibody for two weeks and then uh analyze whether there was objective uh an objective response or whether there was progression and the um you can see there were there were Falls in in lift levels uh the majority of patients in the serum and in fact LIF scores better in this uh recis grade uh Roc analysis than ca99 in in predicting um the response to therapy so in the long term it might be possible to to use lift in in combination with ca99 which is the uh conventional uh marker for pancreatic cancer serum marker for pancreatic cancer so this sort of summarizes what um what we’d found in our studies which were published in in 2019 um so lift plays a driver role at least in the mouse model of pancreatic cancer lift levels are high not only in the mouse model but also in human pancreatic cancer we know lift acts on the tumor cells but it may well act on other cell types in in the cancer and I’ll I’ll get back to that in a minute and Li could serve as as a as a biomarker so the question is then can lift be used therapeutically or can it be targeted therapeutically in human pancreatic cancer and Northern biologics in Toronto small biotech company um developed for other reasons actually an anti-lift neutralizing humanized anti-lift neutralizing monocon antibody and together with cell Gene initiated a phase one trial um in highly refractory cancer patients and they added in nine pancreatic cancer patients because of our results and it was completed successfully there was um no um maximum tolerated dose reached 1500 migs uh K was the 1500 migs was the highest dose that they used um the results have been published and there were some progression free survival responses but nothing significant so what do we know about LIF and LIF aring cancer in general um so it’s been implicated in several types of cancer non-sm cell lung cancer breast cancer coloral cancer expression of lift and LIF R is upregulated in the breast cancer um although there is one report suggesting LIF might be a breast cancer metastasis suppressor Joan say own group showed that uh now about 12 years ago so showed that um TGF beta increases self renewal of of Goma initiating cells by induction of lift particularly in in in GBM and then they’ve gone on to show that um in Mouse uh xenograph models of um with uh gleo blastoma cells and ovarian cancer cells that that LIF induces the recruitment of um macrofagos and acts on macrofagos in the tumors to repress expression of cxcl9 a chemine that normally mediates the recruitment of cd8 T cells into the tumor and so this is one mechanism through which lth is playing a role in imos supression in lung cancer um Pap papag acopos go I could I could pronounce it uh recently reported that autocrine lift is um signaling is important in this genetic subtype of non-sm cell lung cancer lkb um mutant lung cancer and that um blocking lift in a mouse model increases the number of intratumoral CD4 and cd8 T cells interestingly in pancreatic cancer tomasini group showed that lift actually drives neural remodeling VI LIF on peripheral nerves and this is important because actually often the first symptom of pancreatic cancer is back pain back pain and that’s because presumably LIF is now recruiting neurons into the somat sensory neurons into the tumor which is then ultimately leading to Pain tu’s Group has an interesting story that aisle one can induce lift expression and jackstack signaling to generate another class of calfs inflammatory calfs and it was Frack momic group around the same time as our paper showed that lift expression is induced in some pancreatic tumor cell Lines by oncogenic kasas so there are multiple ways to um to induce lift for it to act in the tumors and just to keep you awake this is what you have to do to raise uh funds for pancreatic cancer research um this was a a fundraising walk obviously so I’ve been talking about one way to neutralize LIF um in cancer LIF LIF R in cancer using uh receptor rather anti-lift antibodies but you could also Target the lift receptor and this has been um attempted or is being attempted by a couple of companies I know about so whether this will have unwanted side effects I think is unclear one could also engineer a lift receptor-based lift trap protein like the tnf alpha trap protein Remicade that’s used in uh uh inflam as a treatment for inflammation I want to tell you this story in a single slide because the work was carried out by our older son sha when he was a graduate student at Stanford was in Jennifer Cochran’s lab and a protein engineering lab and so what sha did was to take these three domains of the lift receptor and Carry Out multiple rounds of random mutagenesis and surface display on yeast to generate uh what he called an e lifar engineered lifar with an improved affinity for um for for human LIF and he got a protein with about a 30-fold increase in affinity for lift and he fused this to a human FC to generate Eli RFC and we helped him do the experiments to test whether this new uh lift trap would work in a in a tumor model and in this case um these uh nude mice bearing kp4 human pancreatic cancer cell xenographs treated either with PBS in this case or the Eli RFC and you can see that the El if RFC significantly slows the growth of these tumors at 10 migs per K and I should say these are T tumor cells that make autocrine LIF there are there is at least one drug EC 359 which is claim to block the interaction of lift and lift receptor and I know of other companies trying to make such molecules and um this this is in clinical trials in fact not for pancreatic cancer though then finally another way that one might be able to uh abrigate lift expression would be to use epigenetic drugs like htac Inhibitors or bet tomain Inhibitors because it turns out that lift is the most prominent super enhancer in pancreatic cancer human pancreatic cancer tissue so super enhancers are defined by high levels of H3 k27 AC Asal K lysine 27 and if we look at the top 10 enhancers in this list it’s lift and osm which I’ll talk about in a minute so clearly lift is highly turned on in in pancreatic cancer through this super enhancer and assumably that involves histone acetalation so in collaboration with Ron evans’s lab which and this was mostly done in Ron evans’s lab um particularly by gaang leang a postto in the lab they have tested anat a class one hdac Inhibitors so it inhibits hdac 1 2 and three and um again in the KPC model uh they tested and anat gemcitabine or a combination of anat and gemcitabine and it’s the combination again that seems to give you the strongest increase in in survival and they also found that the antenna stat treatment led to um an increased in increase in differentiation score in in these tumors and what gang was able to show was that if one takes human calfs um treats them with anat and then acquires the conditioned medium now the conditioned medium from the antenna treated cells shows a significant decrease in the ability of the conditioned medium to induce stat 3 phosphorilation and then tenat treatment of these cells significantly reduces lift RNA expression and lift secreted lift protein expression so clearly um histone acetalation particularly uh reversible Hy acetalation is important for LIF expression importantly what they found was that in mice treated with anat there was a significant increase in the infiltration of C8 positive T cells you can see a strong increase here um suggesting that a combination of an hdac inhibitor with a chemotherapeutic drug could be therapeutically beneficial so I’ve talked about ways that we can um prevent lift function in pancreatic cancer but what we really like to do is to be able to use neutralization antagonism of lift to improve the uh immune response now I’ve told you that Northern biologic developed um a humanized neutralizing LIF antibody and there program was acquired by astroica um and it’s now called a 0171 and they have just finished a phase 2 trial in which a 0171 was combined with dulab and their anti pdl1 antibody gem cider be and na pacet taxel in advanced pancreatic cancer and this is going to be unblinded sometime this year so it will be very interesting to see what the outcome is they have just opened another trial with a 0171 in non-s small cell lung cancer as a Neo aivant or aivant therapy so early receptable lung cancer probably based on um phes Pap penopolis uh new data so what about LIF LIF R and the and the immune system I don’t I don’t want to take you through this but obviously the the original discovery that it affected myoid cell differentiation State uh s was suggestive Li is known to increase the level of t-regs um and lift positive tumors have high levels of tumor resident macrofagos and oid derives suppressor cells and I already told you about this cxcl9 story so our own work using single cell RNA seek on cd45 so immune cell immune cells from KPC um pancreatic tumors show that uh lift is very highly expressed by one subtype of myoid cells here which are M1 like Pro inflammatory macrofagos I they also present Express is 6 uh and osm now we’ve not followed up on Isis 6 largely because our Mass Spec data um didn’t show very high levels of I6 in in human pancreatic cancer although it is there so one shouldn’t rule it out but um it’s interesting that all three of these family members are highly highly expressed and in terms of LIF receptor some many of these myoid cells also myoid lineage cells also uh have have lift receptor so there’s one piece of evidence unpublished data from Ron evans’s lab which we helped with as well particularly work of when bin Lee who’s been trying to test whether lift antagonism will improve te- cell killing uh of pancreatic tumor cells and the way that they’ve done this is to generate or acquire uh pedac tumor cells um Mouse tumor cells um expressing gp33 which is glycoprotein from lcmv virus which is very immunogenic um and are killed with when uh people b14 cd8 T cells Mouse T cells are activated with this gp33 peptide and you can see that as you increase the t- cell the tumor cell ratio you get very strong killing here at a 4:1 ratio um so what they found is that if you isolate tumor interstitial fluid by squeezing an orthotopic pancreatic cancer so you get liquid out of it and then um adds Tiff to this killing assay with an 8:1 or 6:1 ratio here Tiff totally abolishes t- cell killing um but if you um just have I’m getting this confused now um but I think that’s that’s the most important finding here so the question is is the is one of the factors in in Tiff which is causing this um killing blockade of killing lift and the answer appears to be yes because if you um use the antibody neutralization to pre-treat Tiff uh and add anti-lift at the same time you can see that the ability of the um te- cells to kill is now significantly increased and so one possibility is that that LIF is actually directly somehow antagonizing the ability of cd8 te cells to um kill pancreatic uh tumor cells and so this is obviously something that they’re they’re following up on now whoops so finally then in the last 5 minutes or so we’re interested in finding other possible secreted Lian receptor signaling Loops that can be targeted therapeutically in pancreatic cancer and we’ve been collaborating with Rin Tian again who’s now at sustc in Shenzhen who is a a mass spectrometry expert to carry out masspec analysis on human pancreatic tissue samples and in particular to enrich for proteins that are glycosilated and he’s developed a spin tip method that can analyze just a few hundred cells isolated by laser capture uh micro dissection but he’s also carried out um phosphotyrosine proteomics on um pancreatic cancer cells treated with um sorry the other way around of uh pancreatic cancer cell condition medium used to treat calfs and and looked at the phosphotyrosine signal so here are the tumor cells conditioned medium acting on on the stellate cells and two of the strongest signals that one gets in this case are the phosphorilated pdgf receptor tyrosine 84 49 and um the tyrosine 542 in in ship 2 which is um a tyin phosphotyrosine phosphatase that is needed for many receptor tyin KY signaling Pathways and using um Inhibitors of either um this pgf no that’s the next slide uh you can see that lift uh expression is strongly increased by uh treatment with pdgfbb and he’s used a panel of inhibitors for this type of mass spec analysis um in cells treated with pdgfbb in the pdgf receptor inhibitor decreases secreted lift expression as does an inhibitor of uh ship 2 uh and Jack one and um finally he’s been he’s he’s used um sna treatment to downregulate fce which is commonly Downstream of this pathway which is as you can see blocked by a Mech inhibitor an KK inhibitor here and um fans that knock down of f also leads to decreased expression of lift and so the model is then that one of the one of the Pathways in the cancer Associated fiberblast that is activated to produce lift is via pdgf secreted by the tumor cells uh that results in uh KK pathway dependent C Force activation of lift expression but I should say then that lift expression can be driven in many ways probably because the super enhancer contains many different response elements and the others are reported fat 3 itself can drive lift expression kreb one the sight leg mp uh pathway transcription Factor driven by crtc2 co-activator and fox M1 and srf and so it’s likely then that there are many ways to activate lift expression in calfs and and in other cells such as myoid cells that need all to need to be worked out I think using a single cell um spatial transcriptomics and proteomics so one can really develop them better targeted therapies so the question is are there other other combinations Factor combinations that could um be targeted and waya GA and ruen have carried out a global analysis of secreted glycoproteins and receptors gly glycosilated receptors in about 30 human pancreatic tumor samples of different stages of normal pancreas and um pancreatitis samples and they’ve identified identified over 1200 secreted glycoproteins and nearly 3,000 membrane glycoproteins and 90% of the receptors have a paired liand expressed in the tissue for instance gas 6 produced by the calfs and Axle art the axle SE tyosin KY produced on the on the tumor cells so there clearly are going to be other secreted Lian recepted Loops that can be targeted and um we’ve chosen to investigate osm and osm receptor because of the data I I showed you earlier and because we already know a lot about cocine signaling Pathways so on satanam is also quite highly expressed um in pancrea IC cancer again using the same tcga data set um and it’s significantly higher than in normal pancreas the black there so that’s encouraging and as I’ve already shown you the myoy cells Express very high levels of oncostatin M so oncostatin M uses both the lift receptor as I told you and also its own dedicated receptor um and so one has to sort out the different functions of these two receptor systems but yanii uh in the lab has begun to study onm in in P in human pancreatic tumors using Multiplex imuno staining and what interestingly if you look at the osm signal which is the red signal here and this is a this is a tumor Nest here you can see that this this the these tumor cells have high levels of oncostatin M receptor whereas there are other tumor nests which seem to have very low levels and that’s something we need to try and try and work out why that should be but interestingly he’s knocked out anatin M receptor in the kp4 uh tumor cells and shown that this strongly reduces proliferation and migration and that oncostatin M receptor reexpression now uh restores the growth and the migration of these cells so Enos this is in the ontin M receptor in the absence of any added ontin m is clearly playing some interesting role in in these tumor cells and we have similar data for other pancreatic tumor cell lines and I’ll finish with data I showed um last time in lecture three which was my phosy phosphohistidine lecture because what what I showed you was that the are high levels of phosphohistidine in this uh ATP citrate lias protein which has a phosphohistidine enzyme intermediate that is needed for expression of AAL COA and what Natalie laler in the lab had shown was that um there are high levels of phosphohistidine in red in the in the tumor green tumor cells here but a lot of the phosphohistidine turns out to be in stromal cells and what she showed was that this was mostly in the cd45 cells and her work and the work of a new postto Aubry mitchi has shown that this is largely present in the myoid derived suppressor cells so hopefully by understanding the role of ACL we may be able to um devise a therapy that would boost the immune response to pancreatic cancer by reducing see the effectiveness of the Malai deriv suppress C so with that and I think I’ll skip this slide since I’ve just run out of time I’ll thank the people who did all the work on pacreatic cancer in my lab Yushi yui Natalie lutal and Aubry Mishi um gaang langang and wemi Lee and Morgan truet in ronal evans’s lab tily rentan who’s done all the mass spec analysis and helped with Nikki lightl and and this r at UCSD and Andy Loy uh Tim donaho Dan Vanhoff and H bansy uh our son at Stanford and Jennifer Cochran’s lab and I particularly want to thank the college to France for inviting me to give this series of lectures and for generously supporting my my visit to Paris and particularly hug D for sponsoring me and and being my host and finally arianto [Applause] [Music]