Cancer Treatment Institute

>> good afternoon. i'm larry samuelson, privileged to fill in for francis collins, who couldn't be here today to introduce today's speaker, dr. carl june. before i do that, just to remind you that there will be a reception in the library after

the talk, sponsored by the fas. we'll have questions afterwards, please come to the microphones to ask your questions. you probably have introductory material about dr. june. i'll pick highlights to emphasize for reasons that will be clear to you.

carl's currently richard vague professor in immunotherapy and director of the center for cellular immunotherapy and director of parker institute for cancer at the university of pennsylvania in philadelphia. he has deep local roots. he was a graduate of the naval

academy in annapolis and went to baylor for medical school, came back to what was then called bethesda naval hospital, west for fellowship in seattle, and then came back in a faculty position and between 1986 and 1999 was across the street again, faculty at uniformed

services university, literally going up in the ranks from assistant to full professor and captain, in the navy. remained in the navy. this was a very important time for a number of us. i had a very productive fun collaboration with carl at that

time, and looking at his c.v. i'm surprised but not surprised that nearly 20 of us p.i.s here in all institutes of the nih had a similar very productive collaborative relationship. so nonetheless, good things come to an end.

we were sad when he went to philadelphia where he's been ever since. he has unique ability to combine basic science skills, knowledge working here, since with a passionate goal in curing chronic serious diseases such as chronic hiv infection and

cancer, and he developed methods for increasing very efficiently growing cells in vitro and using those as he worked on this very powerful chimeric antigen receptor therapy, culminating and continuing but culminating in the 2011 study on patients with leukemia, i'm sure what he

will be talking about today. so carl has received a number of awards. i note that these are all named for distinguished deceased immunologists, donald, thomas, ehrlich and whites man and rosenberg of ccr, nci, novartis prize, well deserved, so it's

with great pleasure that i welcome carl and give him a good round of applause. [applause] >> i want to very much thank larry for the introduction and ability to come back and see so many friends that i see that provided me mentorship in my

early days. i'll talk about leukemia and what we've learned from there. i think those of who knew bill paul know about the unmet medical needs in leukemia, and i'll say that in his memory. so my work is sponsored by novartis, i'll discuss off label

use of car bcma. we're at a great time in many ways. and where cancer is becoming an epidemic as we age. you can see from this data that finally heart disease has been controlled enough that the major source and growing form of

morbidity and mortality is cancer in the u.s., not true in third world countries. i'm going to discuss some data about car t-cells which is really synthetic biology. cd19 cars and bcma directed cars in myeloma, cars targeting glycoapatopes.

i've listed some major publications with human adoptive therapy. there were not gene modified. in 1992, in patients after transplant when they have no immune system a major source of death was disseminated cmv infection.

phil and stan showed in 1992 that you could cure that with a single infusion cnv specific. steve rosenberg and his branch showed with populations of cells lymphodepletion not cancer directed but really against the host immune system augmented the engraftment and persistence of

transfer transfer autologous cells. we could augment vaccine responses in patients with myeloma by giving autologous t cells. asel hoos reviewed the tool set for cancer immunothey were. it began most people think in

human terms with the vaccine from dendrion in 2010 showing you could get fda approval for autologous dendritic cell or cell product, in this case castrate resistant prostate cancer, and checkpoint inhibitors approved in 2011, and have been many more on the way.

this year in 2015 replication, oncolytic vector approved for metastatic melanoma, herpes simplex virus engineered to encode gm-csf, directed directly into tumors. it's anticipated cars will be approved in 2017 for cd19 specific malignancy, the first

gene transfer with modified immune systems that way. many, many combinations will be possible, in the near future, how do we combine these to prevent resistance? we reviewed the kinds of adoptive therapy now in advanced development.

and there's really three forms. one being the till therapy developed for metastatic melanoma. it's more complicated because surgical excision of an accessible mass is required. but then cells are growing and given via intravenous infusion

usually, the gene modified cells with chimeric antigen receptors with lymphocytes, cells are genetically modified in the lab and returned to patients, usually after some sort of conditioning, chemotherapy. so cars are synthetic molecules, don't exist naturally.

but they are a great example how basic science can drive translational science. the first car was made by brian irving and arthur wise in 1991, and sorries including rick klausner's group, the cd4, cd8 molecule shown here activating ras pathways, replicating most

signaling of the t-cell receptor, the single molecule and progressively more complicatedded designs developed, the clinical trials that i'll talk about are using now single chain variable fragments to direct specificity. this would be the tumor cell and

intracellular part of the t cell down here. they have hard wired co-stimulation built into the cars. we did the first car trials here with this cd4 zeta molecule with both niaid personnel and the army.

the surgical branch, first generation cars for cancer patients, and by a group in europe, and they didn't work. they found different survival and engraftment in first generation patients compared to hiv. so this is the first trial we

did with niaid and cell genesis using cd4 zeta car, gave infusions at monthly intervals, 30 to 60 days, cells infused with or without il-2. they could be measured. at this level they are 10% of the cells, we could see them in the circulation, but you see

transient and then until they were reinfused. we then changed our manufacturing, so how we grew the cells with using beads bruce levine made in my lab where we used immobiled and grew the cells and gave three infusions over a month.

six months, there was no -- there was permanentengraftment, cd4 and cd8 cells. with the fda we follow patients for a year, every year for 15 years until the transgene goes away. we found in 2012 that in fact 36 out of 39 patients we treated

still had cars, out past a decade, high levels and stable. very similar to what we found at six months so that they are both cd4 and cd8 that can last with a half-life projected in excess of 17 years. so genetically cells can persist, we think they persist

longer than natural t cells and discussed today maybe because they have an extra copy of zeta from this integrated transgene with competitive advantage and no side effects. we treated as of may 368 patients, between when we started in bethesdas with the

cd4 zeta car now to what we're doing now with novartis, 224 patients. all this observation time we've never had a genotoxic event. no one has had transformation t cell leukemia or lymphoma, now 1500 patient years of safety, safer than chemotherapy or

radiotherapy. when one looks at adoptive therapy, it's the host and how they have been conditioned and which lymphocyte subsets are infused with cars and pre-clinical models, cd4 and cd8 are the best. we don't know are th1 better

than t h 17. how you optimize cells, do the manufacturing, the shorter the manufacturing the better. and progressively we're getting shorter and shorter. now down to we have 4 to 5 days ex vivo time. finally sin synthetic biology,they

are all important. we've been now looking at our cars, compare metabolic properties to natural t cells. erica pierce reviewed data as t cells get activated they go from resting quiescent, and they have massive glycolysis, akt activation, and memory cells

have more exploratory capacity, mitochondrial bass, look different metabolically. we can measure this using the assay and measuring oxygen consumption rates. this is work done by kawalekar in my lab. if you vary the co-stimulatory

receptor, 4bb and zeta, activate through surrogate molecule cd19 they triple in cell volume in three or four days, nutrients required biosynthetically device. seven days after stimulation there's much more oxygen maximal ocr consumption compared to

cd28, preserved out to here on 21 days after. there's plasticity in the car metabolic profile of the car t-cells can be changed as a function of its signaling domain. and this is looking at both by dyes, electron microscopy,

number of mitochondria in the car t-cells. you can see 14 days later, this car t cell is packed with mitochondria, this is not. this is cytosolic, glycolysis, many more mitochondria here that we see as well. the biogenesis changes depending

on that signaling after the encounter with the surrogate antigen. this is a schematic of a model we've come up with, which is we can make cells, i didn't show the data. there's more central memory cells in the cells that have the

mitochondrially driven car, compared to the one here that's glycolysis. memory cells in immunosurveillance. now using the cd19 car now called the unpronounceable name by novartis. first patient in 2010, six years

follow-up, they had cll, and then we started treating pediatric and adult all in 2012 less follow-up there. with results shown here, 14 patients were treated, these were all patients were incurable cll that was advanced, and we had an overall response rate of

57%, four patients have crs, one notable aspect, we've not had any patients relapse. and so it's effective be in a subset of patient. if you look at patients, response, array 14 patients here, these are the patients non-responding at the bottom, we

can measure the cars in blood and bone marrow by genetic sequence of 4nbb and zeta, and this is a log plot on the y axis. and cr patients have massive expansion of a car cell, it's not uncommon 20 to 90% of peripheral blood cells become

when antigen goes into no more detectable tumor or b-cells they have a collapse of a couple orders of magnitude and persistence now to six years of the non-responding patients have engraftment, much less proliferation for reasons we don't understand, a cell

intrinsic issue with host immune system or tumor microenvironment. last patient here had delayed peak in cars, 50 days after we treated him. and this patient lets us conclude as i'll show you one cell is enough to eradicate the

tumor, the daughter cells, from the patient. this patient was delayed as you can see compared to the rest when you look at pk of the cars over time. and he was a 77-year-old man that had advanced cll, also large bulk mass around the aorta

in the mediastinum, and developed macrophage activation syndrome, 50 days after treatment, 21 days when he had his first follow-up, very few cars in the blood and had persistent leukemia, so that will basically a month later are full blown tumor lysis syndrome,

progressessive clearance, now on cr with no detectable tumor. and we asked how did this happen? this late proliferation of these and turns out his cars on day 51 were all cd8, usually we key mixture of cd4 and cd8 when we stain.

these are car cells in his infusion product, beta 5.1, 13.1 here. but over time, the time he was having tumor lysis syndrome, all beta 5.1, we couldn't find any other tcrs, they are the daughter cells of a single cell and we found that out by cloning

those cd8 cells sorted on day 51, but this is looking at conality of cars by the 23v beta, day 28 cars were all families represented, day 51 they are all one, no index, tumor lysis syndrome, no entropy. he went in remission, began to

have multiple different cars, that were subdominant at this time, but at the time of tumor lysis they were all one car. it's integrated as a single copy on chromosome 4, q 24. we modify cells using moi, integrated into chrome season 4, locus of tet 2, integrated

between exons 9 and 10. we made primers to the car around this area, ltc, central poly purine tract. looked at the patient cells rna, three chimeric splicers, one through the vector, the car, and then created two codons, similar at the splice variants.

it knocked out at least one allele of tet 2. tet 2 controls this degradation of 5 methyl cytosein. we were in the unique position taking day 50 cells, staining with antibody, and with antibody that infects, these are cars, companion control cells that are

not cars, a massive different in the amount of hydroxy methyl cytocine. one cell doubling to what we calculate between day 30 and 50, enough to eradicate leukemia, we don't know if this tet2 disruption was a driver or

passenger here. gain of function or not? tet2 in the leukemia world has been shown to increase stem cell self renewal. although not sufficient for transformation. and we don't know what the role of the tcr is in this patient.

chris garcia, we put it through his system and couldn't find any peptides that bind. it may not be in his library. there are now several trials that are underway for registration of cd19 cars, for various leukemias and lymphomas. this is a novartis trial, juno

trials, some at nci, jim is doing one of these. some others now even that aren't listed. the results we've seen in all have been more dramatic than cll, young adults 24 and less that had refractory all, 93% c.r. rate in 60 patients.

green is patients who had c.r.s. blue means b-cell. this is our first, emily white. her igh is germline. she has no, as far as we can detect, rearranged b-cell receptors. and then some patients who have

lost b-cell hiplasi here, relapse of tumor in red. so often it heralds, but most patients remain in long-term remission. those who do relapse, it's loss of b-cell hiplasia or loss of target. cd19 has 14 exons.

and if you look at a refseq, cd19, and in the patients who relapse with cd19 negative leukemia, they often have splicing out of exon 2. so the car no longer can see its cd19 escape happens in the later patients, and we think that needs to be dealt are, with

combinations of cars and trials at our institution to try to prevent target loss. interestingly this doesn't happen with any appreciable frequency in clls or lymphomas like in all, more immature tumor. first all tumor was emily

whitehead, now more than four years out. and this was her cytokine levels where fold elevation over 10,000 fold or a thousand fold elevation of il-2 primarily interleukin 6 and gamma, interferon gamma here. cytokine release syndrome, came

as we give adults, pediatric immune system appears on a per cell basis, and she had very severe cytokine release syndrome, that responded to il-6 blockade, she did not respond to high dose steroids or tnf receptor block aid. b-cell aplasia was predicted,

correlate biomarker in a.l. l and cll. apparently not in diffuse b-cell limb foam a tumor lysis syndrome occurs in patients with high bulk, the more blasts in the marrow a higher probability patients will have crs and tumor lysis syndrome.

low tumor burden could be treated outpatient, high tumor one-third require care for cytokine release. and some get a biochemical signature, high levels of serum fer tin and d dimers, and both of these respond to tocilizumab. these were non-infectious, he

was treated with tocilizumab which binds il-6 and blocks signaling, you can see immediate reset of temperature. we have ongoing a trial asking if we can preemptively give il-blockade or should we wait and have inflammation. it's not been really easy to

model in mice because they they don't get cytokine release. novartis has a trial in u.s. and internationally, results discussed at ash, this is in refractory all. what about non-b cell malignancy? myeloma is a major unmet medical

need. we have a trial, treated and reported two patients with cd19 car with myeloma, cd19ing in, asking hypothesis precursor cell shares the same cloneotype, might have an effect on myeloma and do see treatment effects and have a phase 2 trial underway

with that. and then b-cell maturation antigen, three trials as of a couple months ago, including starting here at the nci that jim is doing, and then full myeloma, and myeloma expresses general fairly high levels of bcma, cart 138 tested in china.

this is an ongoing trial testing bcma car, bcma is a receptor for, viability factor for myeloma cells. and the first cohort of patients they got no conditioning chemotherapy and injection of cars, this is pk again, what we see is similar to what we see

in all with cd9. two patients here, patients on high level cars are clearance of paraprotein made by myeloma cells in the patient with less car proliferation did not have complete response. we see cytokine release syndrome similar to all and cll, it

happens outside targets of cd19. 9there's some data on solid tumors today have not had any massive success, unlike bone marrow derived tumors. the main issues being there around dream targets like cd19. most solid tumors have targets on the surface essential for

other cells, not restricted. we knew with cd19 two things. we knew that congenital immunodeficiency of b-cells was fully compatible with good lifestyle and with b-cell replacement therapy. with solid tumors, target molecules would be -- wouldn't

have that luxury. one aspect we're looking at is targeting glycosylation that's abberant, with o-linked found on proteins, in the case of o-link glycans they began with a single addition to sero3 with a chaperone molecule called cosmic.

george springer wrote this in "science," these antigens, those are normally covered with longer chain branches of sugars, so that immune system has never seen those so we're not tolerant to these. he wrote in 1984 they are masked, not accessible to the

immune system, precursors to normal complex carbohydrates that make tumors have antigens as recognized by foreign by the patient's immune system. a number of academics and biotechs tried to vaccinate people against carbohydrate structures without success, but

they have been safe today. so one aspect of these, it's intriguing there are clinical syndrome when patient don't develop antibodies to tn and tn images, one is iga nephropathy, and hemolytic anemia, and found by rick cummings in his lab, published

in "nature" in 2005 to get core glycosylation you need cosmc, this chaperone function doesn't occur and these antigens appear when normally they are not. they are often overexpressed on pancreatic cancer cells. you can see it's widely expressed here.

now there are isogenic tumor lines in cells where cosmc is knocked out. you can so it gives gain of function with various oncogenes for reasons that are unexplained, at least to me. so avery posey in my lab, this doesn't bind to naturally

glycosylated muc1, normally heavily glycosylated, more than its molecular weight, the sugars, if you have an antibody that targets normally glycosylated with cosmc, the antibody is made by henrik clausen in copenhagen, you can

with recombinant assays make these, it does not bind to muc1. here is 5e10, it binds do breast cancer and normal breast and lactating breast, the car antibody binds to cancer in this case triple-negative breast cancer and not to normal breast or lactating breast.

and we found that jurkat cells have a knockout of cosmc, there's two mutations, and jurkat t cells have massive overexpression of cosmc and no t synthase activity. so avery asked if they could be a target. in fact if you take jurkat

leukemia cells, make this 5e5 car t-cells, it's an allogeneic control, but 5e5 criminals jurkat, these are engrafted with human leukemia survival with 5e5 backup not other car treatments. showing in restoration experiment if you transfects or reexpress cosmc in jurkat cells,

and then takes away cytotoxicity so it's specific as what we're seeing here. and finally, pancreatic cancer, using pancreatic cancer cell line from atcc, also labeled with luminescence, we can see here after giving the cars very specific -- in fact this is

a neomodel with pancreatic cancer, very strong survival effect. so target multiple cancer histotypes with carbohydrate directed are cars but specifically expressed on tumors but not normal cells and tissues.

so in summary, i've shown you that at least in some cases or one case, one car t cell can eliminate leukemia, so that one aspect of these cells is that they are living drug, they are highly proliferative and they need to have proliferative capacity.

another is that cd19 and bcma-directed cars have poet president activity. and there would be toxicity remains unknown. there are a number of challenges and opportunities. in 2017 we'll see this enter commercial availability,

initially done at high end cancer centers and it will take -- difficulties in getting this implemented at community hospitals because physicians aren't trained and this is initially starting with people who really have basically bone marrow transplanters who

understand a lot of this. and so there's going to be that type 2 translational issue. another is more manufacturing, scaleout issue. we have a scale-issue because most proliferation of cells occurs in patients, but we have scaling out, making -- it's a

similar problem facing allogeneic transplant patients, they said it would not be possible to do bone marrow transplants, only done at hopkins and seattle in the '80s, and now two years ago one millionth bone marrow transplant was done, they are

more complicated than car t-cells, we need robotic and fully automatic cell culture, we have systems that still depend on academic-based manufacturing systems, basically requiring highly trained personnel. so industry is going to i think solve this over some time frame

which is do make it automated and done without human hands. until that happens, the scaleout issue will be a limiting issue. because we don't have the trained technicians and scientists sufficient to do that. another major issue is whether

or not universal third party cells can work. i didn't discuss the work we have on this and others. but is it possible to use off the shelf cells from either cord blood or other healthy donors, and until pre-clinical experiments the answer is yes,

there was a -- still unpublished, the london baby experiment came out in the press last year where a child was given cells that had genetically edited out the t-cell receptor, cd19 car inserted, bridged into a transplants so it may be possible to use third party

cells with issues on how to deal with engrafted before becoming rejected by the host immune system. and then i mentioned adenocarcinomas really have issues with having adequate surface targets. t cell receptors have been shown

to work when we have targeting of either cancer antigens, on high pressure target effects, and very good data here on synovial cell sarcomas, glaxosmithkline taking it advance the trials. intracellular proteome have a much better potential than

cars do for surface targets in solid cancers. long run, what we're having now will be orthogonal approaches can crispr technology, which presents really the opportunity to do many things to make synthetically enhanced t cells to overcome issues with

checkpoint resistance with replicating senescence, and also endogenous repertoire can all be modulated with these kinds of approaches. this is where we're going to see a lot of development on the next several years over the first generation of cells that we have

at this point. finally, car t-cells when we first treated patients in 2010 there were three centers with trials open. and now last, as of june, if you do the search term chimeric antigen receptor in clinicaltrials.gov, it came up

with 110 trials open with most in the u.s. and then increasing number in china and when you look at the rate of change which we've done now for a year, in about a year china will have more trials than the u.s. will. i think this points to a need for scientific priorities.

the chinese are invested at higher levels than we are. and so that's emergent. there's relatively less, geographic disparity, lots of pre-clinical science in europe that has been difficult for regulatory reasons for the european scientists to get these

so the u.s. has i think very good relationships with the fda, more straightforward getting the process in europe is more cumbersome. a number of people are involved in these trials. and first i began working with bruce levine in 1992, and he

moved to university of pennsylvania with me and coordinates all the cell manufacturing. david porter has led the clinical team. and michael malone designed the cd19 car we use, henry clausen on the carbohydrate cars,

stephen grupp with children's group at philadelphia, and collaborators at novartis. and i'd like to thank you for your attention. [applause] >> getting back to the -- what you initially talked about with the hiv car t-cells that

persisted for such long periods, can you say more about both with those trials and thoughts towards the future, why they didn't work in terms of anti-viral effect? >> so hiv, those patients then were placed on potent anti-retro virals, low antigen, cd4 zeta

cars redirected towards gp 41, the envelope product of hiv. so it's possible low levels of, you know, latent pool maybe stimulate cells, it's also possible zeta chain alone gives them, and then fully human car. there's no cd4 is human, so the only thing that might be

antigenic would be the coding joint of a car. so what we know is when we compared vaccine trials and so on in patients with chronic disease, cars survived longer than when people measure response after vaccine. and i don't -- whether or not

one needs a co-stimulatory receptor i think is an open question. but the co-stimulatory receptor at least in cancer added, no one compared whether it maybe must be copies of zeta or other co-stimulatory molecules. >> do you foresee any

combination type therapies with td 1 or ctla-4 like a combo that would synergize and work together? >> yeah, phil darcy in melbourne has done some studies in a mouse where they have human her-2. made a her-2 herceptin based car, the mouse tumors had her-2

antigen and there there's potent synergy by giving anti-pd1 antibody, maybe pd-l1 would be better. car is synergistic, if they had a complete response, checkpoint therapy, we've seen in two cases a very potent reawakening of the now, we have a trial that will

start early next year where we're -- with crispr knocking out pd1, cell intrinsic, that may increase the initial burst of cells. so i think it's dependent on tumor micro environment. some lymphomas have hard-wired pd-l1 expression, translocation

and pd-l1 locus, very bright for and so one way to overcome that is checkpoint resistance or make the t cell itself checkpoint resistant. >> thank you. >> after car therapy do you need to give bone marrow transplantation?

>> that's a big question. i'll tell you, so should the t cell therapy be used as bridge to transplant, it depends on clinical setting, whether that needs to be done. so in all and most of our patients have not had a transplants.

many relapsed after a transplant and bone marrow chimera at the time we harvest, graft with donor cells here and in our place people like that have been used, they are really getting donor t cells made with cars, and the response rate if anything is higher than people

fully autologous. now, you know, the argument for a transplant is that one could broaden out specificities through a gbh gvl effect, and but i think, you know, overall it would be nice to not have to do a transplant and to build an immunotherapy that's autologous

and op obviate the need forthat. i think it will be dependent on diseases. aml right now is a disease where the cars work well in pre-clinical models but most of the aml car targets right now also cause aplastic anemia because they have some targeting

on this hematopoietic stem cell, so a transplant would overcome >> okay, thank you. >> reception again. thank you very much, carl.