Talks@12: Immunotherapy: An Answer to Cancer?

Is this working? Can you hear? Can you hear me? No? There we go Good afternoon Thank you for being here We clearly have tremendous interest in this talk on Immunotherapy– An Answer To Cancer And we are really privileged to have two very special presenters here today I’m Gina Vild I’m the Chief Communications Officer for Harvard Medical School And in addition to all of you here with us, we are live-streaming this around the globe And so I want to welcome all of you who are here watching us from afar After the discussion, we will be taking questions We will have a microphone circulating here in the auditorium, but if you are watching from outside of Harvard Medical School, please post your question on the Facebook page, or if you are watching, you can also post on Twitter at Talks at 12 T-A-L-K-S A-T 1-2 Talks at 12 And we’re really hoping we have time for discussion So immunotherapy is the treatment that uses the body’s own defense to combat cancer Although treatments of this type were first attempted more than a century ago, it’s only recently that immunotherapy has shown promise as a means of treating cancer and other autoimmune diseases Ralph Waldo Emerson said, “Do not go where the path may lead Go instead where there is no path and leave a trail.” And we are really privileged to have two trailblazers with us today Arlene Sharpe is the George Fabyan Professor of Comparative Pathology and Head of the Division of Immunology and interim co-chair, as you may know, of the Department of Microbiology and Immunobiology at Harvard Medical School Gordon Freeman is Professor of Medicine at the Dana-Farber Cancer Institute in HMS You may know that they run separate labs, but not everyone knows they also are married And so there is collaboration at many different levels Dr. Sharpe and Dr Freeman’s discoveries have yielded key insights into how cancer evades the immune surveillance and thwarts the body’s immune defenses Their insights helped launch a generation of new drugs that unleashed the immune system to attack tumors and halt the spread of cancer These discoveries have made immunotherapy an exciting new strategy for improving cancer treatment, and importantly, for extending lives In partnership with Harvard Medical School, the Warren Alpert Foundation each year recognizes the world’s foremost scientists for their seminal discoveries– discoveries that hold the promise to change understanding of disease and our ability to treat it And I’m really privileged to tell you that Dr. Sharpe and Dr. Freeman were among the five recipients of the 2017 Warren Alpert prize awarded last month We’re honored you can both join us Thank you for being here [APPLAUSE] Can you hear me? A little louder How’s that? Well thank you, Gina, for that wonderful introduction And in our talk this afternoon, Dr. Freeman and I are going to give a brief introduction to Cancer Immunology and then discuss cancer and immunotherapy focusing on our work on PD-1 that has been translated to therapy for cancer and is beginning to lead to a paradigm shift for cancer care These are our disclosures So as Gina already mentioned, immunology has offered hope for curing cancer for over 100 years but now is just beginning to deliver on this promise What is different now is the strategy We’re beginning to appreciate that the tumor micro-environment has many layers of immunosuppression Tumors evolve to evade immune attack The immune system can survey our bodies for tumors and can eliminate many early tumors Elegant studies of Dr. Robert Schreiber But as tumors grow, they evolve and evade the immune system

This new strategy for cancer immunotherapy blocks pathways that are used by tumors to inhibit anti-tumor immunity And this strategy is called checkpoint blockade Since this audience has a variety of backgrounds, I’m going to start by telling you a little bit about what we know how the immune system fights cancer The immune system has the remarkable ability to recognize a diversity of substances It’s been estimated that the immune system of a person can recognize somewhere between 10 million to 100 million different substances This enables the immune system to defend us against the diversity of the microbial world– from the smallest virus to worms, which can be many meters in length Immunologists call these substances antigens And these antigens are recognized by white blood cells that are called lymphocytes And lymphocytes have, on their cell surface, receptors that can recognize these antigens Now we all have a limited number of lymphocytes that can recognize a particular antigen. But upon activation, these lymphocytes can expand many fold For example, it’s been estimated that T lymphocytes can expand over 50,000 fold within one week, and such a rapid response is needed to defend us against infections The same mechanisms that are used to defend us against microbes also can defend us against tumors There are certain types of T lymphocytes that can kill cancer cells, as well as infected cells, and these can expand rapidly following activation, so this activation process needs to be carefully regulated For T cells to become activated, they need to receive two signals that are delivered by antigen presenting cells, such as dendritic cells The first signal confers specificity to an immune response and involves recognition of antigen. The antigen is presented by a cell such as the dendritic cell to the T cell But in order for optimal and complete activation, the T cell also needs to receive a second signal known as a co-signal or co-stimulatory signal And when the T cell receives both Signal 1 and Signal 2, it can become optimally activated However, we now know that in addition to these co-stimulatory signals, there are also negative second signals that can inhibit T cell activation Our early work in the 1990s, early 1990s, identified that the B7-2 CD28 pathway is the major positive second signal, providing critical signals for T cell activation We identified B7-2 as an early activator of T cells B7-2 engages CD28, and together this signal with CD28 together with a T cell receptor signal promotes T cell activation The signal through B7 CD28 leads to an increase in growth factors so the T cells can grow, survival factors, as well as the bio-energetics So T cells are able to proliferate and then differentiate into effector cells, such as cells that can kill tumor cells Our discovery of B7-2 lead us to look for other molecules that looked like B7 This work in the early 1990s was before we had the entire human genome, so we conducted high homology searches and looked for cousins of these B7 molecules Two molecules that we discovered were PD-L1 and PD-L2 And we discovered that these were ligands for PD-1 Let me use this cartoon to introduce the players The PD-1 receptor is up regulated on T cells upon their activation And when PD-1 is engaged by either PD-L1 or PD-L2, it becomes phosphorylated on tyrosine motifs and it’s cytoplasmic domain This leads to the association of protein tyrosine phosphatases,

such as SHP-2, which then can dephosphorylate kinases downstream of the T cell receptor or CD28 As a result, there is reduced signaling through the T cell receptor and reduced T cell responses Now PD-1 is an acronym for Program Death 1 The PD-1 receptor was cloned in 1992 by the laboratory of Dr. Honjo in Japan from a CD3-activated T cell hybridoma that was undergoing cell death But despite its name, PD-1 does not directly cause cell death, or apoptosis It doesn’t directly activate caspases So it’s not like a molecule such as Fas that can cause cell death The effect of PD-1 on cell death is indirect because PD-1 can reduce cytokine production and survival factors that T cells need for their activation and responsiveness Ever since the discovery of PD-L1 and PD-L2, we’ve been fascinated by their expression What we learned is that PD-L1 was very broadly expressed on a wide variety of hematopoietic cells and non-hematopoietic cells, whereas PD-L2 was much more restricted in its expression PD-L1 could be expressed on dendritic cells, macrophages, B cells and T cells, and on a variety of non-hematopoietic cells, including vascular epithelial cells, epithelium, muscle cells, liver cells, islets in the pancreas, and it’s sights of immune privilege, including the placenta and the eye The expression of PD-L1 on a variety of tissues suggested to us that PD-L1 may control responses locally, T cell responses, within tissues PD-L2, in contrast, is more restricted in its expression It’s expressed mainly on dendritic cells and macrophages However, it can be expressed on several types of non-hematopoietic cells, in particular, cells in the lung on airway epithelial The PD-L2 molecule has an important role in controlling responses in the lung micro-environment Interferons, alpha, beta, and gamma, are potent stimuli for up regulating expression of PD-L1 And interferons can also up regulate PD-L2, but 1L-4 and GM-CSF are more potent stimuli for up regulating PD-L2 To So different types of inflammatory stimuli up regulate this pathway To understand the functions of the PD-1 pathway, we developed tools to probe its function Gordon’s lab made antibodies to PD-1 and it’s ligands to block functional interactions My lab made knockout mice lacking PD-1 or it’s ligands to understand how elimination of these molecules from the immune system would impact immune responses Using these complementary approaches, we learned that the PD-1 pathway has an important role in controlling tolerance to self and preventing auto immunity We also learned that this pathway has an important role in controlling resolution of inflammation Sometimes you can have too much of a good thing and inflammatory responses can damage tissues And the PD-1 pathway has an important role in preventing tissue damage in response to an immune stimulus, as well as tuning down the immune response after elimination of disease We also learned that this pathway is a key mediator of T cell dysfunction This pathway has been exploited over and over again by tumors and microbes, who I think are the smartest immunologists, and they’ve used this pathway to evade eradication by the immune system PD-1 contributes to T cell dysfunction during cancer and chronic infections During T cell dysfunction, there are T cells that highly express PD-1 and PD-1 inhibits their function Another special feature of this pathway is that the tumors themselves can express the ligands for PD-1, PD-L1, and PD-L2 This is from a paper that we published in 2001 where we showed that PD-L1 was expressed on breast cancer cell lines There was an article in the Dana-Farber Cancer Newsletter that named this discovery as cancer’s shield

against the immune system, recognizing that PD-L1 expression on a tumor may be a means for the tumor to evade the immune response Indeed, using antibodies that we developed and working together with colleague Scott Rodig at Sabina Signoretti and David McDermott we determined that PD-L1 can be expressed on a variety of tumors We now know that PD-L1 can be expressed on the surface of about 30% of solid tumors and certain hematologic malignancies, and that PD-L1 on tumor cells can inhibit T cell responses This is just one example of a human kidney tumor and a human non-small cell lung tumor showing PD-L1 expression PD-L1 expression here is visualized using an anti-PD-L1 antibody And this brown staining here is showing expression of PD-L1 on the rim of the tumor cells, bringing here brown in the kidney tumor, as well as you can see there is very high expression on the tumor cells in the non-small cell lung cancer So in the setting of cancer then, we have– in the tumor micro-environment– CT8 T cells that highly expressed PD-1 And these are dysfunctional PD-L1 can be expressed on tumor cells themselves, or hematopoietic cells, or stromal cells in the tumor micro-environment And the antibody that’s used as a drug blocks anti-PD-1 using anti-PD-1 or anti-PD-L1 to block the interactions between PD-1 and PD-L1, and as the results of this can unleash a potent anti-tumor immune response, increasing killing by these CD8 cytolytic T cells, as well as cytokine production As a scientist, you always hope that something you discover will make a difference, never really knowing whether that will be the case And it turns out that the PD-1 pathway is a major way that cancer turns off the immune response And this has been translated to therapy using PD-1 pathway blockade to block the inhibitory functions of this pathway And I’ll turn it over to Gordon [APPLAUSE] Thank you for the invitation to come and talk to you If the people on the back row, if I’m not loud enough, just wave your hands and wave a little if you can’t hear me So I wanted to show you now how the basic scientific research at Harvard Medical School developed into the effective cancer therapies that have come from this work This basic science was taken up by scientists at [INAUDIBLE] Alan Korman and Nils Lonberg, who developed antibodies against PD-1 And there are now five FDA approved antibodies which have progressed through clinical trials to FDA approval These have names like Nivolumab, Pembrolizumab, Atezolimumab, Durvalumab, and Avelumab Now I’m a PhD immunologist, but some very good clinicians have provided me some clinical slides And this is how patients progress on a PD-1 therapy trial So this is in kidney cancer And there are 34 patients Now the drug is generally tolerable The side effects include fatigue, rash, and pruritus Now each line on this graph indicates the progression of one patient and the change in the tumor size in that patient So clearly you can see here that the tumor increases in size This patient hasn’t had a successful response and goes on to a different therapy About 29% of the patients had what’s called an objective response, which means that the tumor has shrunk more than 30% in size So here you can see these objective responses Now what was remarkable in this trial is that after about two years, all the patients stopped the drug And that after the drug was stopped, there wasn’t rapid regrowth of the tumor There is off treatment survival PD-1 antibody can lead to durable responses

Now the PD-1 and PD-L1 drugs have been used in a large variety of clinical trials, leading to FDA approval in each of the cancer types indicated in dark blue So for instance, at the Dana-Farber, Margaret Ship and Philip Armand, have shown that Hodgkin’s lymphoma can have an 87% objective response rate The first FDA approval came with melanoma in 2014 with about a 35% to 40% response rate Lung cancer accounts for 21% of cancer deaths And the PD-1 drugs have been approved there Now in the lighter blue is clinical trials which have not yet led to FDA approval, but it probably will be approved in some Now clearly the PD-1 antibodies don’t work in all tumor types At the bottom, you can see that there’s very low responses to colon cancer or pancreatic That doesn’t mean that there’s absolutely no hope For instance Ira Mehlman’s group has shown that if you combine a MEK inhibitor with PD-1 in colorectal cancer, you can get a 17% response rate So even though discoveries over the last year, 20 years, have emphasized that cancer is many different diseases, to the immune response, cancer is much more one disease This can lead to some very notable , successes such as a 90-year-old with metastatic melanoma and four brain metastases The brain metastases were treated with radiation He also received a PD-1 antibody, and is now self-reported as to be doing well Now I wanted to emphasize that PD-1 immunotherapy is really different from chemotherapy It’s not a poison You’re not trying to directly kill the tumor cell Instead, you’re trying to let the immune system activate and attack the tumor cell So with PD-1 therapy, there’s just a little nausea, no hair loss, no neutropenia or blood cell count declines, has a reasonably good safety profile The PD-1 drug is delivered by a half hour intravenous infusion This is fairly simple, and I think can become community hospital medicine Now the adverse events seen with immunotherapy are very different than those seen with chemotherapy They all end with names like itis They’re basically on target side effects of the immune system going somewhat awry and attacking self organs So these have names like pneumonitis, basically immune attack in the lung, or myocarditis, attack on the heart, which can be very serious The serious grade three or four events occur in about 12% of patients Now clearly physicians are– there’s an education process going on And physicians are going to have to learn to handle a very different range of adverse events than seen in chemotherapy The other thing that’s notable about PD-1 immunotherapy is, when it works, the quality of life can really be good So for instance, this is a graph of the quality of life in lung cancer patients You can see, if treated with chemotherapy, the quality of life of a cancer patient who’s responding to chemo is still the quality of life of a lung cancer patient In contrast, if PD-1 works in lung cancer, the quality of life can actually return to approximately normal And I know some patients at the Dana-Farber who’ve been off their PD-1 drugs and are now riding their bicycles into checkups It certainly doesn’t happen for every patient The tremendous challenge now is that responses are seen in about 20% of patients, and the tremendous challenge is to bring benefit to the remaining approximately 80% of patients Now one of the reasons we have this enthusiasm for immunotherapy is shown in this slide The first checkpoint blockade was CTLA-4 And these show results in melanoma patients, from Steve Hodi and Schadendorf’s work Now historically, melanoma is a terrible disease Most patients died within two years, as shown in the red line Now in the green line is the response to the Ipilimumab, a CTLA-4 antibody

About 20% of the patients respond at approximately two years But what’s remarkable is that if you follow survival across many years, these patients who have responded are alive at years three, four, five, out to the 10 years tracked here And they only received the drug in the first year Now the PD-1 drugs have not been around as long, and this graph follows them out for about six years What’s impressive is that the curve does flatten at about two years at about 35% survival The other– in contrast on the right, is the response to a targeted kinase inhibitor, Vemurafenib, which hits Braf mutants Now the wonderful thing about this targeted kinase inhibitors is, if you have the mutation, you’re almost 100% likely to respond But the problem is, because this hits just one target, the tremendous mutational heterogeneity of cancer learns a way to get around it and evade it So after a year or so, patients become unresponsive to this targeted kinase inhibitor Some more recent targeted kinase inhibitors last more years But if you target just one thing, cancer will find a way around Instead, what I’d emphasize is cancer is an evolving system It has hundreds of pattern recognition molecules, billions of T cell receptors and antibodies, and it can evolve and change as the tumor changes It can attack many points, so evasion of just one doesn’t lose you the effectiveness The immune system also has a memory So T cells that have expanded remember what they’ve seen Now here, the CTLA-4 antibody works in melanoma, but unfortunately as a single agent, hasn’t worked in well enough in other tumors to receive FDA approval In contrast, the PD-1 antibody works about twice as well in melanoma with about half the side effects, and has been extended to many other tumor types Now critical questions for us now are how do we identify who will respond to PD-1 blockade What are the mechanisms of failure to respond, and how do we bring benefit to the patients who aren’t responding? What are the mechanisms of resistance? And how do we develop combination therapies for non-responders to PD-1? So this can begin with what does the immune system see to attack in a tumor? And what it sees is changes Indeed, what the immune system sees is changes Now if you’re a micro-organism or a virus, those are totally foreign And the immune system sees them as foreign Now a tumor has many mutations Some of these, such as pictured in the bull’s eye here, are the driver mutations, such as in Ras or Raf, which drive proliferation of the tumor But there are many other passenger mutations, some of which contribute to the tumor, many of which do not And these can change the protein coding information And these changes are seen as foreign by the immune system And they’re called neoantigens for new antigens So one realization from this is that the tumor is an evolving thing and that there’s also an evolution of immune evasion The tumor is learning to evolve and evade the immune system by expression of PD-1, IDO, and other mechanisms The other thing this emphasizes is that every tumor to the immune system is different, because each person’s tumor has different mutational changes than another person’s Now the response rates to PD-1 therapy appear to correlate with the mutational frequency Work at the Broad Institute and Farber has shown us that different tumor types have different levels of mutations The tumors with the highest mutation rates are things like melanoma from sunlight, or lung cancer from tobacco mutagens And this graph shows you how mutated each tumor type is And over on the far right, you can see melanoma or lung cancer, also bladder cancer And these are the tumor types that respond very well

to PD-1 immunotherapy The tumor types over on the left have lower mutations, fewer changes for the immune system to see, historically are not as good responders Now Rizvi has shown that the higher mutation burden rate in some tumors can be associated with better responses to PD-1 immunotherapy And indeed, understanding the immunology and the genetics has identified groups that respond particularly well to PD-1 or PD-L1 immunotherapy Work by Louis Diaz at the Johns Hopkins has shown that certain tumor types have deficits in mismatch repair And these tumors have many more mutations than the average run of the mill tumor And immunotherapy turns these mutations on their head All these mutations let the tumor advance, but they give much more targets for the immune system, once you block the tumor from evading the immune system So these mismatch repair defense deficit tumors have about a 62% overall response rate At the Dana-Farber, Margaret Shipp has pioneered and shown that there are genetic amplifications in PD-L1 and PD-L2 in literally every Hodgkin’s lymphoma Now in all, about 4% of tumors have mismatch repair, particularly those about 15% of colorectal, 20% of endometrial, a range of other tumors The Hodgkin’s have amplifications, but 1% or 2% of broad range of tumors have genetic amplifications of PD-L1 So one of the things that I think there’s enthusiasm developing for is for any cancer patient who walks in the door to be analyzed to see whether they have these deficits, which make them particularly amenable to PD-1 immunotherapy Now other things I’d emphasize is that the immune system sees differences So tumors caused by viral antigens, such as Merkel cell polyoma or HPV in head and neck can be particularly well-recognized and have good response rates Now how do we identify who will respond to PD-1 blockade? And these are predictive biomarkers Now the first biomarker examined was PD-L1 expression It does suggest who is likely to respond to PD-1 therapy, but it’s not a highly confident prediction Some patients who don’t express PD-L1 on their tumor still respond Definitely every tumor with PD-L1 on their tumor does not respond So we need better biomarkers And different studies have shown that the CD8 infiltration and proliferation or an interferon gamma signature, neoantigen burden, all can suggest higher response rates to PD-1 immunotherapy But there’s a lot of work that remains to be done And we may need combinations of markers to confidently predict response So the failures of response can include no good neoantigens, or the expression of other immunoinhibitors, like Tim 3 And there are clearly many more cancer immunotherapy targets than just PD-1 These include all of the molecules shown here on this slide, including CTLA-4, TIM-3, TIGIT, and many others And indeed, academic labs and pharma are developing agents to test each of these So other failures of response can include the failure of immune cells to infiltrate into the tumor And this slide basically illustrates this idea Over on the left here is a tumor with a lot of infiltrating lymphocytes and PD-L1 expression And these have a good response to PD-1 immunotherapy And over on the far right is what’s characterized as an immune desert– a tumor that doesn’t have any infiltrating T cells or lymphocytes And this doesn’t respond well So what this does say is that where PD-1 is working, we’re taking a smoldering immune response and releasing it And the big challenge is to find ways to get T cells into these immune deserts Now the future is clearly combination therapy And indeed, this list is just a small list

of the possibilities There are publications of over 300 different agents that can be combined with PD-1 CTLA-4 to get improved results So CTLA-4 and PD-1 combination has been approved in melanoma and is showing promising results in lung and kidney cancer, though it’s more toxic than either agent alone Other agents in advanced clinical trials include IDO, TIM-3 and LAG3 TIGIT is coming Other possibilities include combining PD-1 with immunostimulators And immunology’s first approach to treating cancer was to try to stimulate the immune response with vaccines or adjuvants And that didn’t work But once you block tumor immunivation, a lot of these tumors stimulation strategies are enabled and can work So you can combine PD-1 with anti-Ox40, IL-2, TLR ligands, or STING Now chemotherapy in its simplest form is basically designed to kill proliferating cells And the basics of the immune system is proliferating cells So a lot of chemotherapy you might expect would knock out the immune response But not all So academics and scientists have carefully looked at which targeted therapies and chemotherapies can combine with PD-1 immunotherapy And ones that can work include Braf inhibitors, MEK inhibitors, cisplatins So certain chemotherapies can be combined with PD-1 immunotherapies and work better Others possibilities include angiogenesis blockade Five years ago, if you asked me how radiation worked, I would have said it just kills tumor cells It does do that But when you kill a tumor cell, it’s taken up and presented by the immune system So you can get the immune system engaged HDAC inhibitors, affecting epigenetic changes, can also enhance the immune response against tumors Other possibilities include PD-1 blockade plus cancer vaccines, such as being pioneered by Kathy Wu at the Dana-Farber, or oncolytic viruses, such as being pioneered by Sean Lawler, Martuza, and Nino Chioza at the Brigham Or PD-1 therapy can also enhance the activity of CAR-T cells There are other more nonstandard possibilities, such as PD-1 plus certain nutrients For instance, vitamin D is not good It can actually increase PD-L1 expression So you don’t want to just take a simple-minded mega vitamin Now recent papers are also emphasizing the importance of your microbiome The presence of certain microbes in your gut can enhance anti-tumor immune responses So I hate to say it, but you may go to your drugstore and get a probiotic which is purported to increase anti-cancer immunity And careful studies are needed there to find the right, real ones So the goal is basically illustrated in this slide We want to take PD-1 therapy and combine it with some of these combinations Many of these combinations have been shown to work better in mice Many of them are progressing to human clinical trials And we’ll see how they do So we hope to increase the response rate and to increase the durability of these combination immunotherapies I think the future of cancer immunotherapies is going to start with cancer genome sequencing If a cancer is identified to have mismatch repair or PD-L1 amplification or viral genomes, that patient is a good candidate to progress directly to PD-1 therapy It will also identify what oncogenes are drug targets And so what tumors are good targets for a Braf targeted kinase inhibitor, plus PD-1 It will tell us what tumor, what mutations, are immunogenic and can be developed into a vaccine The pathologist is going to look at the tumor and show us how it’s evading So it’ll show us how much PD-L1 expression, how much IDO, what’s the right strategy to take And this will allow us to choose the best immunotherapy and combine it with the best vaccine or targeted therapy So it’s really an exciting time to be a researcher or an oncologist, because it’s a better time for patients

And there’s tremendous opportunity for all of you to do additional research and improve these therapies The PD-1 is a good foundation to work with because it has a moderate percentage of responders and is reasonably safe It gives us a foundation to build on And with this success, a tremendous amount of human creativity has been unleashed on the problem So as of October last year, there were five FDA approved drugs 15 more are in clinical trials I think now over 1,000 clinical trials ongoing, with over 166,000 patients slots Clearly, there’s a tsunami of data coming And bioinformaticians are badly needed We’ll learn what works, what safe, how it works, and who it’ll work for And I think this will really change cancer therapy So I’d like to acknowledge over the years the collaboration between Arlene and myself, and all the members of the Freeman and Sharpe labs who have contributed to this work, particularly Ivette Latchman and Julia Brown in the early stages Many others thereafter All the patients, nurses, and physicians who have participated in these clinical trials, and all our collaborators over the years, including David Reardon, Peter Hammerman, Tony Chuari, Kwok Wong, Nick Haning, Glen Dranoff and Margaret Shipp At the Brigham and Women’s, wonderful pathology by David Dorfman, Sabina Signoretti, and Scott Rodig Great work on TIM-3 by Vijay Kutru Really critical early biological studies, as well as tremendous signaling and metabolic studies by Vicki Bouceiatis at the Beth Israel and the Dana-Farber Kidney cancer collaboration with David McDermott and Michael Atkins Wonderful collaborations on the idea of T cell exhaustion with Rathi Ahmed and John Whery, and really the early discovery work that opened the door on the PD-1, PD-L1 pathway, with Tusuko Honjo and Clive Wood at Genetics Institute Thanks for coming We’re happy to take questions [APPLAUSE] So we have some time for questions, but Gina just told me that we have a number of people watching So I want to thank the many people who are watching And it’s truly amazing We have people participating, many people from India, Algeria, Brazil, Moscow, Vietnam, Peru, the UK, Dominican Republic, Cambodia, Spain, Mexico, and Tanzania Truly amazing You can take those question or we can ask questions from the audience Anybody have a question? Here are two of the questions Could you have given this talk four years ago, or how much of it? Could you hear him? No Could you have given this talk four years ago? How much has changed in four years? So some of the principles– could we have given this talk four years ago? Some of the principles were known, but what’s amazing is the astonishing rate at which the clinical translational is occurring The advances in the clinical trials, the FDA approvals of PD-1– this year alone, there have been four or five new FDA approvals for therapy As more clinical trials are occurring, more types of questions arise in terms of how we want to answer them, going back from the bench to the bedside, trying to understand mechanisms of response and mechanisms of resistance PD-1 therapy really was approved only in 2014 So in the past three years, there’s been a tremendous change We’re going to take one of the questions from the thing It’s said tumors evolve for survival Does this also increase the immune resistance of cancer cells? Well, they’re evolving for survival by changing survival factors, like BCL-2 and evolving mutated oncogenes But at the same time, the immune system is trying to attack and destroy the tumor

The genome of tumors are unstable and pattern recognition molecules can also recognize the tumor So the tumor has to evolve to evade the immune system by expression of PD-L1, IDO, and other mechanisms So they’re parallel pathways of evolution But in addition, there are some patients who do initially respond to PD-1 and then become resistant And we’re beginning to learn about those types of mechanisms that tumors no longer express PD-L1, for example, or have mutations and MHC molecules or processing pathways And so this has also led to great interest in combination therapies So you’re combining several different types of targets at the same time Hello So I’d like to ask, since this therapy PD, PD-L1, has showed such a high success in primary tumors such as lung or melanoma, what’s your thoughts about how successful do you think this therapy could be, for example, in brain metastases? Since, you know, the majority of brain metastases are precisely from the lung or melanoma tumors Would you repeat the last– the disease type you said at the last? Do you think that this therapy will be– could be successful in brain metastases, since it’s showing such promising results in the primary tumor, such as lung and melanoma? There are publications that the immunotherapy can work on brain metastases Certainly not every brain metastasis Many of the brain metastases are treated with focused radiation, in addition to the immunotherapy So there may be some synergy between the two Beautiful work Congratulations I wonder if it could be a numbers game, though, that there might not be enough T cells to kill all the tumor cells How many tumor cells can one T cell kill? Any information on that? [LAUGHTER] That’s a very good question I don’t know how many tumor cells one T cell can kill in vivo In a tissue culture dish, we probably can estimate that But in vivo, I think we focused on talking about T cells, but in addition, there’s other types of cells There are different types of T cells that can participate in killing tumor cells In addition, there are natural killer cells So there’s a whole armamentarium of cells that can be unleashed and really T cells are the first ones Thank you very much for the very inspiring work So my question is, do you think for the future we can, instead of going to the direction of blocking, all the checkpoint blockades, can we engineer the immune system not to express the checkpoint blockades, for example I don’t know It can be used SIRA 2 shot down, CTLA-4 expression, PD-L1 expression Can we engineer the immune system? Do you think that would be a promising future direction? Well, these inhibitory signals that are expressed by these checkpoints have important physiologic roles in regulating our immune responses I think the PD-1 pathway is the paradigm for how inhibitory signals are used for proper functioning of the immune response Signals through PD-1 and CTLA-4 are important So we have tolerance We don’t have autoimmune disease In addition, if you eliminate these– so if you eliminate these, you could get increased autoimmune disease You also could get increased tissue damage with inflammation So these pathways have very important physiologic roles So you wouldn’t want to eliminate them completely Though in strategies where there are engineering of T cells, such as car T cells, there is work ongoing to think about how you might eliminate these cells completely or temporally Now one of the questions on the card was, is immunotherapy good for glioblastoma, which is a serious brain cancer?

And the answer there is, as a monotherapy, the response rate was about 4%, which isn’t very high So the initial answer is no Glioblastoma is going to need combination immunotherapies Some of the complicating factors there are edema in the brain, and swelling is clearly a serious side effect in glioblastoma And that’s often treated with corticosteroids, which can reduce immune responses So probably finding ways both to use less immunosuppressants and to also alleviate brain edema, such as research we’re doing with Rakesh Jain are needed to make immunotherapy safe and effective in the brain Thank you very much for both the talks It was a very inspiring and very informing as well I have a question about, if I take a look at the bigger picture, ideally it would move towards a form of cancer therapy where you have [INAUDIBLE] So you have destruction of tumor cells You alleviate the micro environment immunosuppression, such as the PD-1 therapy And you would like to stimulate the immune system and make it stronger to have a stronger response to a tumor So first of all, am I missing something there in terms of the bigger picture? And the second question would be, are there other mechanisms in the microenvironment of the tumor that can be targeted to avoid immunosuppression? So I think you summed up accurately the good take home messages from our talk in terms of the tumor using checkpoint blockade as one means to inhibit these suppressive mechanisms that are in the tumor microenvironment But as was mentioned, the PD-1 pathway is only one type of immune suppression There are many other inhibitory receptors that can be expressed locally within a tissue microenvironment In addition, there are stromal cells There are different types of myeloid suppressor cells, adenosine, many different elements that we’re learning about that can be suppressive And all of these different types of suppressive mechanisms can become targets for immunotherapy And in fact, they are targets They’re being analyzed now individually and in combination I also wanted to expand on the glioblastoma answer And in mouse models of glioblastoma, we and others have shown that the immune response can get into the brain, attack glioblastoma, and eradicate it So it is susceptible to immunotherapy There are also a number of combination therapies being advanced, such as oncololytic viruses, viruses that rather specifically attack, kill tumors, and bring in the immune response These are being pioneered at the Mass General and Brigham and Women’s And so there are combination therapies relevant for glioblastoma Thank you very much for the talk It was a beautiful love story between the industry and academia I don’t know how much of the profit the academia will see from that, but my point question is whether the T cells or the system develops a memory after the treatment, because it seems the treatment is rather brief And then the effect is prolonged If you have any data on that or any insight on it So you raise a very good point And I think one of the special features of immunotherapy for cancer is the durability that can be seen And so from the earliest days, the work that Steve Hodi did using anti-CTLA-4 ipilimumab, it was beginning to see a tail on a curve where you could see that there was long-lived durability in a subset of patients, about 20% of patients And now these patients who have shown benefit are over a decade out With PD-1, it’s a shorter experience, but they’re also a group of patients, about 30%, who also are showing durable responses So the excitement here is much the same way when we see responses to infections where we can see immune memory, that are we generating immune memory to these cancers? And that may explain some of the durability

Hi Thank you for the beautiful talk During the talk, you have described these PD-1 inhibitors as a means to unleash T cell responses So since T cells are involved in many growing inflammatory diseases, my question will be whether by using this therapy, are you risking to actually trigger an inflammatory [INAUDIBLE] in these patients? If that makes sense I don’t know Yes The adverse events seen in PD-1 immunotherapy are indeed increased immune activation, attacking self tissues– something like pneumonitis is the immune system attacking lung And again, this is serious in about 12% of patients The other– as you said, this PD-1 expression is broadly seen in other chronic viral and infections So for instance, in, AIDS, tuberculosis, malaria, hepatitis C, all of these are diseases where the immune system has tried to attack, not succeeded, and then reached a compromise to muddle along, not destroy the liver of a hepatitis patient, but not let the virus go wild either And PD-1 is a major regulator of this attenuated dysfunctional immune response So the PD-1 immunotherapies may be relevant in chronic viral diseases, such as AIDS and hepatitis In addition, there is at least one paper saying that PD-1 therapy in a mouse model of Alzheimer’s was therapeutic, as long as you kept giving it the PD-1 antibody This, perhaps, could be through enhanced macrophage activation and clean up of brain garbage But that’s an intriguing area of future work In addition, when we first got into this field, we were very interested in autoimmunity And if you could– all the PD-1 agents that are drugs now are blockers But what if you could engage the pathway and increase its activation? That could be therapeutic in multiple sclerosis or rheumatoid arthritis, and is really an open, but challenging, area to develop agents that increase the activity So there’s a lot of work remaining to be done So I think, as Gordon mentioned, that developing agonists to trigger these inhibitory signals to shut down immune responses in type 1 diabetes or multiple sclerosis– and because these pathways can control responses locally within the target tissue, these are areas of active investigation And hopefully soon we’ll see some agents that can deliver Since the PD-1 therapy work in many type of tumors, it looks like [INAUDIBLE] or antigen tumor [INAUDIBLE] antigen. So I’m curious what are your comments for antigen specific therapy for [INAUDIBLE] a tumor independent of this PD-1 therapy Well, I think the big idea of the last four years is the realization of the importance of neoantigens, the new antigens in the tumor caused by mutations in the protein coding regions So the immune system recognizes those neoantigens and attacks the tumor But each person’s tumor has a different spectrum of neoantigens than every other patient Each patient is really– has an individual spectrum of changes which are recognized Is there any actual success at all in the dendritic basic immunotherapy? Could you repeat? Is there any success at all in the dendritic basic immunotherapy? I’m just curious because there are plenty of studies describing T cells and all other cell types, but dendritic cells are kind of interesting People are not– I think there are clear, small clinical trials successes with dendritic therapies I think the– they’re just more complicated than not

I think the Cardiff therapies, I think, are very exciting They work well in some of the hematologic malignancies And the tremendous challenge now is to make the Cardiff therapies relevant and effective for solid tumors, where they haven’t yet shown significant success Hi So in your slide, you showed that there are a higher response rate to PD-1 than CTLA-4 So why is that so, when both CTLA-4 and PD-L1 was in a similar manner to directly immuno response? I think part of it is timing, that both of these pathways can control initial T cell activation But the PD-1 pathway can control effector T cell responses and responses in target organs And it’s a key mediator of this T cell dysfunction And because it’s expressed locally and many tumors can express the ligands, that when you block this pathway within a tissue, you can stimulate an immune response So I think it’s partly timing CTLA-4 acts early PD-1 can act early, but it also can act later during an immune response One last question [INAUDIBLE] So the question is, which anti-PD-1 immunotherapy drug has shown the most success for adenocarcinoma? [LAUGHTER] I don’t think I’m going to get into the business of trying to choose between there Last question When you talked about T cell infiltration, you talked about immune deserts or some term like, that meant that T cells, areas the T cells cannot get in at all How common is this, and it should be a hot area of research Do we know anything about the mechanisms of blocking T cells from completely getting into an area of tumors, or how does this work? So there are a variety of different types of tumor microenvironments, some that are inflamed called hot tumors, and others that are called cold tumors that are these immune deserts And these types of environments are more common than the ones where you have an inflamed environment And there can be many reasons why you may have this type of environment And so there are a number of strategies underway to try to stimulate T cell activation and migration into these tumor environments So the strategies for success for cancer immunotherapy may be different for a cold tumor or an immune desert than for an immunogenic tumor And with the realization that there are different requirements, then we can begin to tailor strategies to these different types of tumor microenvironments Thank you very much [APPLAUSE]