Last Updated on July 19, 2021 by Zion Miller
Salveen Richter (SR)
Andrew Allen (AA)
SR: Hi good morning everyone my name is Salveen Richter, a Senior Biotech Analyst at Goldman Sachs thank you for joining us. We are pleased to have Gritstone Bio with us with Andrew Allen co-founder, president, and CEO. With that Andrew, to start the company recently changed its name from Gritstone Oncology to Gritstone Bio given expansion outside of oncology, could you discuss this shift and what it really means for the future of the company?
AA: Certainly, well good morning Salveen, and thanks for having us here at your conference. Yea so Gritstone is coming up on six years old and when we began we were focusing on the new challenge of oncology that presented itself at the time which was how to figure out antigen-based immunotherapy for cancer. And we were born really on the heels of this exciting publication in the New England Journal from Tim Chan and Naiyer Rizvi. They demonstrated that in patients with solid tumors who were responding to immunotherapy the targets for the T cells seem to be tuned to neoantigens-these mutated proteins that can be present surface and tumor cells and we also knew that some patients could respond to simple checkpoint therapy which is fantastic and obviously has transformed the field, but unfortunately, most patients with solid tumors do not respond to simple checkpoint inhibitors. The hypothesis that we articulated collectively back then was that the reason for this was that there were many patients who have tumors that contain neoantigens, but the tumor has successfully hidden from the adaptive immune system and there are no neoantigen-specific CD8 T cells in those patients. So, if you treat them with a checkpoint inhibitor which essentially non specifically activates T cells then there is no substrate and no basis for them to work and all that could happen is that you will stimulate T cells and generate autoimmunity and of course that’s exactly what we’ve observed. So the hypothesis was if we can identify the neoantigens for these patients and put them into a vaccine that could then drive a strong CD8 T cell response in humans then we can convert these patients from being check-point unresponsive perhaps to being check-point responsive and so we built the company to solve the two core scientific challenges. The first one how do I know which bits of a mutant tumor cell are going to be presented as neoantigens that the immune system can recognize and secondly how do I then administer them in a vaccine that drives a strong CD8 T cell response which at that point was a major challenge for the vaccinology field and we’ve essentially solved both of those problems and having done so we now have a platform that can identify antigens from any substrate and of course that includes tumor cells but also infectious diseases and infectious pathogens. Interestingly, perhaps one day they might even include autoimmune antigens to start thinking about auto-immunity. And then our chief scientific officer at the time and her team developed a very potent vaccine system that drives strong CD8 T-cell responses consistently in humans which had not been shown before and that’s a powerful tool that can be also used for certain pathogens where CD8 T cell responses are important. And so having really focused on oncology as the basis for solving those problems, having then developed solutions we, of course, realized that we had applications outside of oncology most obviously in infectious disease and at that point, the name Gritstone Oncology seemed rather anachronistic and so we evolved to Gritstone Bio to reflect the broader possible scope of the platforms that we’ve built.
SR: Could you walk us through just given the various programs that you have you know how we should think about data readouts for the remainder of 2021?
AA: Yeah we’re excited by the second half of this year we have a lot of interesting data coming through so perhaps the flagship data in some sense is the granite personalized neoantigen program so this is a very elegant test of the fundamental hypothesis that we articulated six years ago and obviously, we began with a blank sheet of paper and we built a company that includes technology platforms that have been published patented, it includes our own biomanufacturing facility that’s been generating GMP material now for a couple of years so a lot of work that brings us to this point where we’ve been dosing patients with solid tumors with a personalized neoantigen immunotherapy we call that GRANITE. So what we do is we take a tumor biopsy from each patient, we sequence it, we identify all of the mutations in the tumor. Every mutation is then given a probability of being a neoantigen. Most mutations have extremely low probability lots of zeros and then a one and basically are highly unlikely to be neoantigens. But every so often about 1 to 2% will have a probability that’s meaningful and so we aggregate all of the higher probability neoantigens into a vaccine. We put 20 neoantigens into that vaccine and then we administer that to each patient and is given with an adenovirus prime which is the most potent CD8 priming vector that we know and then we boost periodically with a self-amplifying RNA construct containing the same neoantigens and at the same time as these vaccinations patients are receiving a PD1 antibody and subcutaneous low dose CTLA4 antibody ipilimumab. So this is the GRANITE program Phase 1 and then early Phase 2 in advanced disease patients I think it’s always worth remembering these are probably the last patients you would want to begin with if you have a vaccine-based immunotherapy because it takes something like 8 to 12 weeks for the T cell response to really mount and some of our patients are progressing and dying within a handful of weeks clearly where we don’t have the right therapeutic modality to help some of those patients and we obviously accept that that’s the way the system works that’s appropriate nonetheless within these patients we’ve seen clear evidence and we presented some early data in July of last year that we’re changing the trajectory of the disease. Now we’ve been building that data set and most importantly we’ve been following patients over time and we will be sharing new data at ESMO in September and that will be then a much more complete and longitudinal set of data from around 20 patients who’ve been treated with the GRANITE immunotherapy platform and some of those patients obviously we treated over a year ago so importantly, of course, we’re looking for this longitudinal follow up and we have a particular interest in colorectal cancer and even talking about 12 this is actually meaningful in its own right. Colorectal cancer I’m talking here, the standard microsatellite stable form of colorectal cancer which is the very common variety we know that it doesn’t really respond to immuno therapy at all and indeed third line patients which is the context in which we’re treating patients have been studied many times by many different programs and depressingly the median PFS is two months so most patients progressed at the first scan the median survival is around six to seven months and those data very tight and in colorectal third line trials, Ill give you an example the Bayer stivarga trial was running third like colorectal cancer so the same population that we’re studying and it’s a drug that’s approved because there was a significant impact on PFS and OS nonetheless they out of 760 patients randomized in that pivotal trial 2:1 to Stigarva or best supportive care the number of patient that were alive and progression free at 12 months was zero so there are no good actors in this disease everybody does poorly unfortunately but that means of course signals should be fairly easy to see and so we’re excited to be presenting these longer-term data from a few more patients at ESMO in September and we’ll have a corporate event to go into those data in more detail and behind those data we’re launching two randomized phase two studies in earlier stage disease and we’ve not been specific about those but will only give a lot more detail and the rationale will become obvious at September so that’s an important event for us around the core program for the company which is the GRANITE program and in the background we’ve been continuing to innovate in terms of biomanufacturing and our manufacturing process is becoming faster, cheaper, and higher throughput because of course making personalized products as we all know is nontrivial this is a lot easier than cell therapy we don’t have live cells to deal with, but nonetheless you have to have a highly rigorous robust process for making such products at scale and that’s something else we do aggressively in the background so that’s GRANITE. Associated with GRANITE is our off the shelf neoantigen immunotherapy program called SLATE and we presented some early data last year and we showed that in fact we probably made a mistake when we were designing the antigenic cassette we put a lot of KRAS neoantigens together into a vaccine and some P53 neoantigens and we showed that these P53 neoantigens could indeed be shared between patients. The truth is they’re actually pretty rare individual P53 mutations and quite rare. It’s very common to have lots of P53 but there are many ways to lose function in a gene so there are lots of different mutations so if you like the phenotype is common but individual genotypes are rare anyway we put those together turns out P53 extremely strong antigen and it seems to crush- to use the jargon it’s immunodominant over KRAS and that crushed down the magnitude of the KRAS responses. That said, we have seen some interesting data and will be giving an update probably in the summer on SLATE as we move to what we’ve referred to as version 2 where we’ve removed the P53 and it’s a very clean KRAZ focused product off the shelf that can be used for the many patients who have KRAS mutations and have the right HLA type so in the summer will be updating on the final data with the version 1 as well as launching the version 2 which will be happening in the summer we expect to be dosing first patients around then so that’s the SLATE program and then later in the year will be launching a third program which will be using a novel as yet undisclosed neoantigen that we’re very excited about this is a shared neoantigen it’s in the lung cancer space we’re very excited about that program will disclose more there over time so that’s it on the oncology side if we switch over to the infectious disease we’ve of course built this now fairly well known second generation COVID-19 program and of course we started this work early last year and happily for all of us the first generation vaccines that primarily generate neutralizing antibodies to spike have been remarkably effective and very safe and that’s obviously good for us all there is however some continuing interest in two different questions left I think on the table. The one that we started with which is a little deemphasized right now but it still looms out there like the fog hanging off the coast of San Francisco is the notion of new variants. We’ve obviously seen several new variants arising happily some of them only have driven the infectivity of the virus but actually seem to respond just as well to antibodies generated by the first generation vaccines so I’m referring here to sometimes called the UK variant now with the new WHO alpha variant or it’s B117 if you prefer the numbers so that variant obviously is highly infectious it is now dominant around in many countries around the world happily it doesn’t seem to be significantly more pathogenic and there is still some debate on that question but the good news is that the antibodies the first generation vaccines generate clearly neutralize it just as effectively as they do the original variant of SARS-CoV-2. There are two variants that do seem to have relative resistance to the antibodies that the current vaccines generate that’s the so-called South African variant, the Brazil variants this is the beta and gamma, and now the Indian variant delta and the latest data on delta which is spreading rapidly through the UK is it has about a five to six full lower sensitivity to neutralizing antibodies generated by the first generation vaccines and that’s we can manage that so if it’s five or six-fold lower potency all we’ve got to do is make sure that we all have decent antibody titers and so boosting with the first Gen vaccines looks like it will give us good protection against the delta or Indian variant so so far we’ve been able to deal with the variants that have arisen so the question remains will there be another variant that actually are vaccines can’t deal with and it’s selection pressure rises because we’ve all been vaccinated or naturally infected will this new variant arise that might have greater escape from the immunological control affected by first-gen vaccines. No one knows the answer to that and if that happens it would be good to have a product available that can work and that’s where our program potentially would be relevant. What we did that’s a little different was to take that basic premise I described at the outset, in other words, a vaccine that makes really strong CD8 T cells and we applied that to SARS-CoV-2. Now we didn’t spend a little time on antibodies at the outset and we’ve done that work over the last year we showed that our platform can certainly could generate strong neutralizing antibody titers both in mice but importantly in nonhuman primates which seemed to reflect humans much more accurately than rodents do so we have very good preclinical data suggesting this will generate good antibody titers also, of course, it will generate strong CD8 titers. The reason that’s important is that the neutralizing antibodies all recognize spike and in particular, the neutralizing antibodies recognize that S1 domain is the sort of outer domain of spike and if there is going to be an escape variant it’s going to involve changes in that spike protein. The T cell epitopes in SAS-CoV-2 there are some in Spike but there are lots from other genes which are largely ignored in all the other vaccines they just don’t put them in because they’re not relevant to antibodies but they’re very relevant to T cells because as you know CD8 T cells recognize virally infected cells and the antigens they recognize are very distinct they recognize short peptide fragments displayed on the surface of an infect itself totally different from a spike protein on the outside of a virus and so changes in spike protein probably won’t change the peptides recognized by T cells and so our vaccine is what we refer to as a chimera so we have the whole spike protein to make the antibodies but also we have regions of other viral proteins that provide strong antigens for the T cells, for the CD8 T cells and that may give us much broader coverage antigenically much broader coverage in terms of immune response antibody and CD8 T cell and that could provide enduring protection that might be constant through significant variation in spike. So that’s potentially one opportunity for our COVID vaccine. The second that’s actually current is immunocompromised subjects. There’s been a whole flurry of manuscripts in the literature showing that there is a large number of Americans who aren’t making very good antibody responses to the first generation vaccines so if you think of it you think of patients on dialysis generally have some blanket immunosuppression so they’ll make some response some patients won’t have much of a response some it’ll just be attenuated so there’s a relatively low response in that group relative immunocompromised. Organ transplants are heavily immunosuppressed so those patients are really at risk from COVID-19. And perhaps the cleanest group to contemplate are patients who’ve received the CD 20 antibody so of course a patient with leukemia or lymphoma who’s been treated with Rituxan is B-cell depleted that’s how Rituxan works it depletes B cells so those patients often make extremely poor responses to a vaccine that’s primarily trying to generate an antibody response not surprisingly and of course we could immunize those patients with a potent T cell vaccine to give them T cell protection against the virus and a paper as she was published online yesterday from Johnson and Johnson and Dan Barouch in Beth Israel suggesting that with the JNJ vaccine which uses an adenovirus that does generate good antibodies but also some degree of CD8 response it’s not optimized for CD8 response but there is more of a CD8 response than probably with the mRNA vaccines against the South Africa variant, the beta variant the antibody based protection is diminished but the clinical protection against severe disease was preserved and the article interrogates the T cell responses in the vaccine recipients and suggests that those are fully preserved in those subjects which is continuing evidence that T cells while they wait may not prevent you from actually getting infected they can perhaps prevent you from getting sick which of course we care about a lot so the notion of T cell vaccination potency cell vaccination for a Rituxan recipient is a pretty attractive idea so that’s something that we’re also exploring as you can imagine in our own clinical program so that’s the backdrop to why we have a COVID program in terms of clinical data NIH/NIAID who has been running a phase one study that started in March and so we should have data coming off of that platform sometime in third quarter and then Gritstone is beginning our own sponsored studies also in third quarter pursuing some of the ideas that I just described in terms of obviously we’re not looking for naive subjects here in the US but we’re thinking about boosting patients who may have received a different vaccine but still have persistent unmet need or may benefit from a broader T cell response than the first generation vaccines can offer so that’s the data from the COVID platform and then finally we have the collaboration with Gilead around HIV obviously a program we are very excited about working with the world’s probably leading HIV company and that is a timeline that’s under their control so I don’t anticipate any clinical data this year but obviously we hope to have some milestone progress in that platform towards the end of the end of the year.
SR: Andrew thank you that was a really comprehensive overview. You talked about seeing first data from the work you’ve done with the NIH in the third quarter. Maybe help us understand what you’re really looking for when we see this data set. Then with regards to your internal run program is the construct the same for the vaccine for both the NIH program and yours?
AA: Yeah so with the NIAID study obviously we’re looking for antibodies so we’re looking for neutralizing antibody which obviously now there are some standards using standard assays that are fairly comparable from program to program and of course the differentiation that we’re looking for is the generation CD8T cells and so those are the data that will be exploring in samples coming off of the NIAID patients and there are different ways to analyze CD8 T cells. It is technically more challenging to quantify and qualify CD8 T cells which is why you see less data around this topic but obviously can be done and so that’s important work that there will be starting soon on the samples off of the NIAID subjects. We have some background data from some of the competitors that we will be able to indirectly compare against but the most standard assay that’s run is the so-called EliSpot Assay and that is a very simple assay you literally just take blood from patients, separate off the red blood cells, and isolate their B cells, T cells, NK cells monocytes that the so-called PBMC’s user jargon, and they literally just put them in a dish add some peptide and essentially count proliferating cell clusters the next day so there’s no manipulations its a very standard I say you can compare pretty cleanly from one trial to another trial as long as everyone is doing the same thing and Johnson and Johnson have published data and as I say they you know they’ve been very sophisticated in their development and they selected a variant of the vaccine that biased towards antibody generation away from CD8 even though they’re using the adenovirus and so they’ve published data the we’ve obviously looked at and in the elderly for example only about 30% of subjects generate what they qualify as a meaningful CD8 response and of course the elderly they’re at risk of severe disease and mortality from SARS-CoV-2 in particular and so there’s room I think for significant increase there so that will be one of the elements that will be looking at. In terms of the construct we put in the very first construct we put in with the NIH because it was available then released actually as a spike only product so similar to to what others have done and then more recently we’ve begun using a spike plus T cell epitope construct so that’s the one obviously that is meaningfully differentiated because it has material viral material that the other vaccines don’t have and it broadens the substrate for the T cell response so obviously that’s the one we’re all really looking at now we’ve continued to iterate we actually have another version of that but that’s not going to be in the NH study but obviously if we get to a point where we’re thinking about pivotal trials will probably move to a slightly modified version where we’ve taken all of the recent learnings around T cells and capture those within the construct as well we we have a collaboration with Alex Sette’s and his team at the La Jolla Institute for Immunology and they’ve been in the vanguard of really looking at human T cell responses in patients infected with natural SARS Co V2 and that of course is an incredibly valuable data resource for selecting which bits of the virus do you want to put in a vaccine if you’re trying to drive a strong CD8 T cell response alongside the spike antibody response and the latest learnings from that are captured in a newer version as I say that we can potentially put into a program that might be launched later this year or early next.
SR: Andrew is the data we’re gonna get in the third quarter from the NIH is that just from the spike-only construct or is that from the second construct as well?
AA: And just go so I to just to answer that, its a little hard for me to know that. We’re not the sponsors of this study so I don’t have direct control over that so I’m not sure is the simple answer.
SR: And then with your study do you given that you’re going to be using in patients who’ve been exposed to the previous vaccines do you have a sense of the endpoints in the clinical development path?
AA: Yeah it’s a great question so you know Dr. Fauci talks about correlate protection meaning what is the immunological result that we are confident will provide clinical benefit clinical protection and for neutralizing antibodies there’s a lot of data out now and I think you know the FDA has spoken to this they haven’t got numbers down but they I think they obviously are collating data suggesting that they’ve now got pretty good data supporting a particular neutralizing antibody threshold so using a standard assay here is the number that we want you know the population to achieve and they’ll have some criteria I’m sure around what fraction of the population must exceed that number and for how long to believe that this product has clinical value so for neutralizing antibodies I think we’re well on that road it’s harder with T cells so again if you think about these patients with who’ve received Rituxan and they’re not going to make very good antibody responses if at all but CDA protection is likely to be extremely helpful and again JNJ have been in the vanguard of that movement clinically. So how do we set a correlate for CD8T cells? Today that’s an open question. We have pre-clinical data we do not yet have robust clinical data, at least not in the public domain. There may be data the FDA has that that I’m not aware of but that’s kind of the direction we’d be going in where we clearly have a population that is at risk you know solid organ transplants where data emerged pretty early those patients have really high mortality from SARS-CoV-2 infection and some elective transplant programs were shut down because the mortality was so very high in those subjects. So there is a clear unmet need and the question is how do you develop and approve a product where you are relying on CD8s as the effector mechanism given the limited data to help us set immune correlates. That is obviously a dialogue that will need to have with the agency the unmet need is clearly there but this is establishing an immunologic correlate, a threshold is something that lies in our near future I think
SV: And then turning to the cancer vaccine pipeline we’re going to see data from your personalized cancer vaccine or GRANITE program at ESMO this year. Help us understand you know when will be able to really get a true sense of proof of concept? I think we’ve seen early proof of concept but then it probably takes a lot more patients duration and so forth to better answers so how should we be thinking about how to interpret this data coming at ESMO?
AA: Yeah so it’s obviously a critical question and it’s challenging particularly because we use checkpoints alongside our vaccines. So the simple thing to do so well don’t use checkpoints but obviously we’re trying to generate neoantigen reactive CD8 T cells that express PD1 and we know that tumors can escape T cell destruction by simply upregulating PDL1 and shutting down the T cells so it really makes little sense to not have PD one antibody in the mix so immediately you have this question well how do I dissect out the contribution of the vaccine versus the contribution of the PD1. It is this reason that we chose colorectal cancer as an important substrate for clinical trials because we know the checkpoints in the microsatellite stable colorectal population really don’t do anything they’ve been studied extensively the response rate the objective response rate is less than 1% and there’s no significant impact on survival or even progression-free survival and again this has been studied many times. So um we like colorectal for that reason there’s a lot of unmet need there’s a lot of these patients who are struggling out there and so it’s not a terribly competitive field to lung cancer or Melanoma so for many reasons colorectal cancer is interesting. So if I think about colorectal cancer I can probably say with reasonable comfort that if I see anything it probably isn’t attributable to the checkpoint so that’s an important step so then what am I looking for well of course traditionally we’ve all jumped on RECIST response and RECIST response at the first scan has become sort of widely sort of sought after as the early marker of efficacy but the challenge we have is that what we’re trying to do does not reflect itself likely in early research response ’cause if you think about it research was developed for cytotoxic chemotherapy in other words a treatment that just kills tumor cells and I extrapolated to say if I measure tumor masses on radiology and if I assume that the tumor mass is basically all tumor cells which we know is not entirely true but let’s assume that if my chemotherapy works it will shrink the mass and so recist is just a standardized way of measuring mass size on the assumption that the treatment if it works willl shrink the mass because it’s killing tumor cells and it obviously works well and we’ve used it extensively. However, think about what we’re trying to do we’re taking cold tumors that have very limited if any T cells in them we’re trying to generate an immune response which takes 8 to 12 weeks so an early response I think is unlikely and then we’re trying to drive those T cells into the tumors the lesions and the T cells will then meet their antigens and should start proliferating so we’re introducing a very very challenging dimension into the use of RECIST because we now want proliferation of a cell type within a lesion, of course, we want to add destruction at the same time but I have no idea of how to separate out the if I’m just measuring a gray blob on the CT scan and it gets bigger is that good because that’s T cells proliferating and tumor cells are being destroyed or is it bad because actually there’s no T-cell proliferation and its all tumor expansion? I just can’t tell. So, RECIST responses in the short term may be unhelpful. So how can I get through that? Well, the technique that’s now of great interest is circulating tumor DNA and in the checkpoint field, we’ve seen now several examples whereby you treat with a checkpoint and acutely you see sometimes you see legion expansion for the reasons I’ve described but CT DNA goes down in those patients if they subsequently the tumor diminishes that’s called pseudoprogression the CT DNA actually going down will help you differentiate between pseudoprogression and true progression so it works very effectively. so cDNA something we’ve been studying quite carefully and you can think of it as molecular response and so seeing molecular response is a way of sort of fighting through the fog of RECIST radiology so that’s something that we’re very interested in and again the the natural history of this disease is that when patients begin a trial in third line as I said they typically progress at the first scan in other words their CT DNA is likely just to be rising relentlessly so if we start to see evidence that the patients are not progressing aggressively and they’re getting CT DNA reduction that would be extremely encouraging data showing that we’re changing the course of their disease so you combine that with radiology and of course we do scans and we follow them long term and maybe overtime you can start to see lesions disappearing we don’t know this is a journey that no one’s really been on before that’s why longitudinal data matters so if the median survival is as I say six or seven months, if nobody in a 760 patient randomized trial made it alive and free of progression to 12 months, and if we’re going to be showing data for some patients at that 12 months point of course I think you will be able to discern whether there is a signal here or not. We’ll present that full data set; we’ll present radiology, CT DNA, tumor markers of course (CEA is another marker that’s relevant), so we’ll have that whole data set, and I think it’ll become pretty clear even with single-arm study with all the challenges and usual caveats I think this will be interpretable data.
SR: Great, now maybe just one last quick question here you have the off-the-shelf slate platform as well and have been working to optimize the version two cassette just remind us where that optimization stands?
AA: Yes so the IND is being filed imminently and so we should be dosing the first patient in the summer. At that point, we will probably make that announcement and give an update on what we’ve observed in the two-week version one which is not uninteresting.
SR: Great. And then maybe one more follow on your infectious disease platform outside of COVID how does that work with Gilead and I guess your internal work progressing?
AA: Yeah so the Gilead program obviously is under their control, they are the sponsors and they run that program so that’s something that obviously I can’t really speak to that’s that’s their program. We obviously are interested in potentially other infectious diseases of course and we have some internal work exploring some different options and obviously will be making announcements when appropriate later this year.
SR: I’m gonna try and throw one other quick one. On the cancer vaccine side when you look at the work that we’re seeing at the mRNA vaccines with Moderna and BioNTech in cancer and yourselves how do you think in the end that the three, the two technologies may play out in the market in terms of the profile of these drugs? Are they going to be much differentiated?
AA: Yeah the mRNA vaccines are clearly spectacular at driving antibody responses they struggle on CD8s and obviously basically everything we know about cancer in therapy and neoantigens suggests if you’re trying to predict Class 1 presented neoantigens the final effector mechanism of tumor destruction will be CD8 T-cells so I think if you orient around the axis of CD8 T cell generation you’re going to be in the right zip code and that’s obviously what we’ve been very focused on and the data for the mRNA vaccines look very different both qualitatively and quantitatively. If you just look at the CD8 responses to spike for example across the different datasets you’ll see clear differences between the products so I think differentiation is highly likely and these should not be aggregated as a neoantigen vaccine; it is extremely different depending upon the neoantigens I’m putting in as in am I good at predicting which mutations create neoantigens that’s a hugely important step, very complex, we are very good at it and then secondly does my vector elicit CD8 T cells in large numbers in humans. If you don’t do both of those things well I suspect it’s going to be hard to demonstrate efficacy in solid tumor immunotherapy
SR: Great, well with that thank you so much and we really appreciate the time today.
AA: Thank you Salveen, take care.