Abstract
Certainly. I'm the CEO and director of Scientific Affairs at the Cancer Research Institute. We've been around since 1953, and we have had one persistent mission—to understand how your immune system could be used to treat, control, and potentially cure all cancers. Seventy years ago not many people believed in this and when I came to CRI 37 years ago, there was still a lot of skepticism. Many people felt it had been tried and disproved, and I really believe we were the lone voice in the wilderness, always supporting excellent science. Lloyd Old was the head of our Scientific Advisory Council for 40 years, and it was his vision that said we need to understand how the immune system functions before you could intelligently apply it to the cancer question. That's exactly what CRI did. No one doubts today that your immune system can be used to treat cancer, albeit we know that it's not the panacea … but I think we now have certainly the acceptance by medical oncologists that you can use immunotherapy to treat different cancers. I'm very proud that CRI can really look back and say we kept the field alive. We have trained multiple generations of immunologists who are now the leaders in the field.
Jim Allison is our director of the Cancer Research Institute Scientific Advisory Council and he won the Nobel Prize for CTLA4. His key findings over the last decade around checkpoint blockades have been transformative. That's only one type of immunotherapy and there is some education needed around immunotherapies. I think the general public and some treating physicians think about immunotherapy in terms of just checkpoint blockades. But there's many different types of immunotherapies, and the challenge now is progressing all of those and figuring out combinations of various immunotherapies or immunotherapies with standard-of-care treatments like radiation and chemotherapy to deliver a more precise treatment for patients. This isn't a one-size-fits-all approach.
I became the head of the organization in 1993, and I'm very proud to say that I've been involved in developing many of the programs that we have at CRI. When I first came we had a postdoc fellowship program—back then there wasn't the feeling that young scientists needed to be trained as immunologists. We've always been an organization that's very responsive to the field. We have a board of directors and a volunteer Scientific Advisory Council, and we listen to them in terms of needs, challenges, and opportunities. Since I've been there, we have developed an entire continuum of programs that really deal with training, but also basic immunology, basic cancer, translational and clinical trials … I hear over and over again, even though we have successes in the clinic, that there is still a lot of biology we do not understand and we cannot abandon supportive basic research. It is in basic research where the real discoveries are produced, but then we need to figure out how to translate that in the right way into successful drug development. I think we have done an excellent job in that area.
I think you're right. It's wonderful for patients that there are percentages that do respond, and respond with really long-term effects in many cases. But as you've said the majority of patients still do not really benefit from the current checkpoint blockades. Then we have CAR T cells and as you mentioned, Emily Whitehead, where these only work in hematological malignancies and not in solid tumors. A lot of work has gone on in therapeutic cancer vaccines and there's decades and decades of failures there. But we're now seeing glimmers of hope in that area, so I think we're in a different place. As I said, my perspective is over decades. We are now at a place that we have responders and non-responders and that wasn't the case 20 years ago. So now we're able to interrogate, what is the difference between that? Can we pinpoint why some patients with the exact same type of cancer, seemingly the exact stage of cancer … one gets a great response and one doesn't.
I think we really need to double down on making sure we get samples of blood and tissue prior to treatment, during treatment, after treatment…precious patient samples that will allow us now to correlate those with clinical responses. What was the difference in the patient that responded and the patient that didn't respond? I think that's key for advancing the field. What's happening in the tumor microenvironment is what we need to understand … there's been such an explosion in technologies that are allowing us to interrogate these blood and tissue samples in ways that never could have been done before.
All of the multiomics and spatial approaches—understanding exactly what different classes of immune cells are in the tumor environment, where they are in relationship to one other, really gives insight into saying what is happening here. When you look at that in a responder versus a non-responder, that gives you a very good hint of what's the effective cell communication that needs to happen to get an effective response. But we really need to understand mechanism. It's that communication between the lab and the clinic and the clinic and the lab, that circular iterative process that I think is really important. That's why we fund usually small clinical trials. I think you can learn a lot from a trial of 10 or 15 patients, a well-designed patient trial where you're getting clinical signals. When I first started at CRI and we started funding clinical trials, we were not getting clinical responses to any of the immunotherapies that were being tested. You saw you were getting an immunological response but it wasn't really developing into a clinical response. Now, we're at a different place. You have a clinical response. You can now correlate these changes at the cellular and the molecular level with response.
Yes, it's a very complicated system. This isn't like developing one targeted therapy, one single small molecule that affects one pathway in the immune system. It requires a systems biology approach … I think there was a naivete, especially among people that were not immunologists. Once the checkpoint blockade had these remote responses in melanoma, in non-small cell lung cancer, people thought, “The only thing we have to do now is take a checkpoint blockade and combine it with my favorite molecule, and we're going to get great responses.” Well, we've seen over the last decade that that hasn't played out. I think we have to be smarter, and I think we have to be science-driven. That's why we have this large array of different programs that we fund. And we have a technology impact award. We are not funding just biology, but realizing technologies really lead to step changes in research. It allows you to do things. We are funding startup ideas around new technologies that hopefully will end up being effective and being distributed.
I always think back to when there was so much expectation and so many high hopes and how we witnessed such disastrous declines. Everybody was saying let's throw the baby out with the bathwater because the immune system doesn't work.
I think we were very lucky that the first checkpoint blockades with CTLA4, and then with PD-1, had such remarkable responses. I think if they hadn't the field wouldn't be where it is today. It changed not only medical oncologists' views, but it changed investors views and funding agencies' reviews. I mean, when you look back … I hear all the time back in the eighties and the seventies … you couldn't get your grant funded by NIH or NCRI … it was very high risk. Now there's a lot of certainly more funding available and it's great. On the flip side of that I worry…once we saw these checkpoints the whole world [thought] we were going to get a 100% response rate, so all of these treatments are going to work. It hasn't. So you do worry about the venture capital dollars that have come in and the startups. Are they going to be scared off now? They didn't get their returns as quickly as they thought they were going to. I hope we don't go into this period of people saying, “PD-1 was a one-shot thing and it's never going to happen again.” I don't think that's true, but I think you need people that have an understanding of the complexity of the immune system, and how difficult it is mechanistically to understand what is truly a productive immune response in every patient.
We all know that pancreatic cancer is one of the hardest cancers to treat. Response rates are pretty low to any of the existing treatments. We partnered with the Parker Institute for Cancer Immunotherapy and started with the PRINCE trial. This was comparing a combination of standard chemotherapy, plus one or two different immunotherapies. In mouse models it looked like the four-drug combination was going be a success and surprisingly, the four-drug combination was not a success. But we did see significant clinical responses with either a checkpoint blockade or with a CD-agonist. By doing multiomics analysis of the patient samples, we were able to find blood-based signatures in patients before treatment that would define who would respond to which combination therapy. Now that is the value of doing this correlative science, because I think we all know it is not going to be one-size-fits-all.
Obviously most pharmaceutical companies would love that to be an off-the-shelf thing where you just treat everybody with pancreatic cancer the same way. Well, it's not the way. But this is where we need biomarkers of response. In immunotherapy we don't really have great biomarkers. Obviously PD-1 is considered a biomarker, but what does that mean? Where's your cut-off? What level do you [need]? I think multiple parameters have to be built in. With multiomics, you can get a genomic signature of the type of patient that's going to respond to one treatment or another. But if you understand mechanistically what's happening, you can tailor the treatment so that the right patient gets the right therapy. I think that's where the field is going. I have been challenged when I talked about this before … can we really personalize medicine? Can that really be scalable? We're talking about an n of 1 scenario, i.e., single patient, single trial. But I think we're going to be seeing archetypes of patients. There's groups of patients that look this way or have this genomic signature that respond one way or another, and I think that's where the field is going in terms of development of immunotherapies. I think we're very lucky that sometimes PD-1 works. But that's not 80% of the patients. We need to understand the defects in the immune responses of the non-responders, because clearly taking the breaks off the immune system with PD-1 at the tumor site was not enough. So what's the problem? Are there no T cells there to begin with? Why are there no T cells? Are there no antigens being presented to the T cells? Or are there physical barriers in the stroma of the tumor margin, and T cells can't get in? There's all these different things that we have to understand. Doing the analysis of responders and non-responders at the level of genomics will really aid the field.
We actually completed a study and we're just about to fund the second follow-on. We partnered with the Canadian Cancer Trials Group, which acts as our clinical trial management. The lead was out of Johns Hopkins and PGDx was the company that measured ctDNA. I think that with the reports coming out in Nature Medicine, we will see a ctDNA measurement from non-small cell lung cancer patients being treated with pembrolizumab. We've been able to show in these 50 patients that measuring the amount of ctDNA was a better and earlier indicator than this gold standard of radiology scans that could show if the patients were responding or not to immunotherapy. This is really important because I think the earlier we know if people are responding to PD-1, [the earlier we can tell them] they should continue that. If they're not having a response, they need to go on to some other treatment. We're now expanding this to a much larger trial. But it looked like the ctDNA blood-based assay was better at predicting overall survival than scans. We need non-invasive, quicker, earlier, faster ways to be able to really know and make clinical decisions about how the patient should stay or not stay on an immunotherapy. The trial we're going to be funding next is actually a registration trial and maybe a way to really change practice.
Right now I'll say we're not doing much in AI. I think there is a future for that. I think it's early days. I don't think we know how to use it, but I think it will be important. Over the last year we kept hearing how important computational approaches are in analyzing all the research, from the clinical research to laboratory research.
You need people that can not only do the analysis but understand it in the context of the immune system. Just like back in the early seventies when we needed to train young immunologists before we could expect to get cancer immunotherapies … we need to train people in a dual way to be immunologists and data scientists. We've started an immunoinformatics fellowship this year that allows training … it's an adjunct to our postdoctoral fellowship program. It's for someone that may have a PhD in immunology that wants to do their postdoc in a data science lab or a computational biologist that wants to learn more about immunology. I think this is serving a challenge and an opportunity. At any given time we have about 100 postdoctoral fellows that we're supporting around the world. We did survey them to ask, do they need data scientists? Do they understand how to do these multiomics analyses? Do they understand single cells? They understand spatial [analysis] and many of them need it but don't rely on a core lab, or they rely on a friend in the lab. But they weren't really trained. This spring we're going do a boot camp for our fellows … a week-long immersion course on the analysis of bulk DNA, bulk RNA, single cell and spatial [analysis]. I think this is important. I think data is here to stay. They're going to generate more data, we need to understand that. Down the line this is where AI comes in, because the volume is so large that we're going to have to develop algorithms that can help it. But it has to be done without blinders on, the mathematical has to be contextualized.
Sure. Over the last few years we have been funding a group of scientists at Sage and the Institute of Systems Biology outside Seattle. It's an interactive web-based platform that houses immunogenomic data that originally started out of TCGA (The Cancer Genome Atlas Program) data on the 33 different cancer types. We're expanding it to accept single-cell data. It also has information on published clinical trials where you can get the immunogenomic data. Our vision is that this becomes a centralized immunogenomic repository for data that is open science … that it allows people to come in and compare their data to what's in there, interrogate and generate hypotheses, and hopefully push the field forward. Our CRI iAtlas is an ongoing project and we still have a ways to go, but it is evolving and responding to what's needed in the field.
The majority of our programs are really funding programs for academics and are done through open access. Applications come in and we use our Scientific Advisory Council, which I really believe is the who's who of the field in immunology, to review, rank, and refund those programs. We created what we thought was an academic industry nonprofit incubator. This is a little bit different from our other programs. It's not open applications, but rather we have a core of academics who we consider the brain trust, who come up with questions they would like to ask that they feel are not getting asked by pharma nowadays … focusing mostly on Phase II trials. We bring them together and engage in discussion with our clinical accelerator team to look across the field. What trials are on already? We're looking for the white spaces. We don't want to use our philanthropic dollars to duplicate an existing trial. Then our team goes out and tries to gain access to the drugs that the academics feel are the right drugs to affect different mechanisms. Then we broker relationships with pharma and biotech to gain access to the drugs and get some co-funding. They also agree that we can combine their drug with a drug from another company, and then we fund that study. We think of it as a kickstart to a development path. This may not be prime in their development path, but it gives the pharma company another shot on goal by working with our brain trust. We can show why this would work and we share the data and hopefully, if it's important enough and it aligns with where they're doing, we can get this into the drug development path. Then hopefully we get a return on investment. So that's kind of the model we've been pursuing for almost a decade now.
I'm very proud of where we are now. I think a lot of what we do is limited by funding. I think our programs are really important and we need to act more as this link between academia and industry. There's a role to be played there. I think excellence is everywhere and if we can fund and support this on a global scale, bringing all these people together and allowing really smart people to have the tools and the resources they need, then I truly believe science will create great things. I would love to take a stronger role in deciding what actually goes into clinical trials because I think we have a lot of duplicative trials. There's thousands of trials going on with immunotherapy and I think that's a waste of resources and a waste of patients. This year we're thinking of instituting a way of working more closely with companies and academics that are doing the clinical trials and funding the translational piece. There's a lot of great trials going on, but the investment is lacking.
I think there is a role for data. We have to figure out how to enable scientists to better analyze data, share data, learn from data. We put our toe in the water this year with our informatics boot camp and our iAtlas. We have to think about how to expand. We will need to train the next generation of scientists with our fellowship programs. We still have to support translational research with our CLIP grants and we need to continue early phase clinical trials, and invest in a data-driven immunology ecosystem over the next 5–10 years.