Adaptive Biotechnologies Corporation (ADPT) Earnings Call Transcript & Summary

We are starting our next panel, discovery of novel TCRs and making personalized therapies. It's my pleasure to introduce Sharon Benzeno; Dr. Benzeno, Chief Business Development Officer of Adaptive Biotechnologies. Please welcome.

Thanks, Janet. Good afternoon. My name is Sharon Benzeno. I'm Chief Business Development Officer at Adaptive, and I lead our drug discovery group. It's a pleasure to present to you today. Adaptive is a publicly traded clinical-stage company with the mission of translating the genetics of the adaptive immune system to transform the diagnosis and treatment of disease. The adaptive immune system, as you may know, detects and treats most diseases in exactly the same way, be it in cancer, autoimmune or infectious diseases. And it does so because of the inherent diversity of the adaptive immune system built -- composed of, for example, trillions of T cell receptors that can recognize and respond to millions of disease-causing antigens. To give a sense of scale, this is the size of the human genome made up of 3 gigabases of information. The transcriptome on the exome are even smaller. In contrast, the adaptive immune system made up of T and B cells is massive with 10 gigabases of information containing 100 million genes. This is because lymphocytes, T cells and B cells meet to recognize millions of antigens that can't be known in advance. And major solution to this has been to create this massive diversity of millions of antigen-binding receptor that can have randomized specificity. Over the past decade, Adaptive has built a robust immune medicine platform to be able to screen and quantify this diversity. The biology we leverage is the dynamic and random rearrangements of the (V) variable, (D) diversity and (J) joining segments within the TCRs and BCRs. As an example, here's the rearrangement that occurs creating unique CDR3 regions, which is the sequence across which we leverage our technology and can quantify. Once identified, T cell receptor sequences, in this case, looking at the TCR beta locus can be tracked over time. And so once identified, we can track a TCR beta sequence through its nucleotide or amino acid protein level, but also quantify its frequency as well as other standard metrics that we measure, such as clonality and T cell fraction. Recognizing that there's a lot of information that's locked up in this single chain sequence, we set out to expand our immune profiling platform. From left to right, immunoSEQ is our standard assays that are the foundation for single chain sequencing of T cell and B cell receptors. And MIRA is our multiplex approach to map TCRs to antigens at scale, more on this technology in a bit. immunoSEQ Dx is our technology that allows at a high throughput to pair the alpha and beta chains of T cell receptors, and the same technology is also used to accurately pair the heavy and light chains of B cell receptors for antibody discovery. And then our cellular immunology capabilities and expertise allows us to fully characterize T cell receptors through our end-to-end true TCR approach as well as antibodies through our TruAB discovery process. And so for the TCR beta sequence shown here, we can go beyond a single chain sequence to actually match what that T cell receptor recognizes in terms of an antigen, which is a permanent record for that antigen TCR match. And then, of course, given our pairing technology, we can also identify the cognate alpha to be able to fully reconstitute and synthesize this T cell receptor. Our TCR discovery approach is based on the power of being able to access the massive diversity of the naive repertoire to identify naturally occurring T cell receptors that are optimized for potency and safety. We do this an unprecedented throughput and scale. And then specifically, we're interested in high avidity TCRs shown at the tail end of this distribution. It's relatively easy to find many TCRs that have low-to-moderate avidity. It's a whole other realm to be able to really hone in on these unique T cell receptors for therapeutic use. This is our TruTCR approach. We start with an optimal number of cells, either from healthy donors or cancer patients that contain millions of T cell receptors, and we can screen across hundreds of antigens within 1 experiment. Our MIRA technology allows this first step to match TCRs to antigens. The first step allows us to identify thousands of TCR antigen hits. We then apply our pairSEQ technology to pair the cognate alpha, and we also upfront apply a proprietary titration message to be able to hone in on the high avidity TCRs that are the best binders and like the optimal clinical candidate. We then invest in further characterization of those TCRs regarding functional avidity, cell killing as well as a battery of safety steps to derisk a given TCR, including alanine license can, homology modeling and assessment of primary cells, ultimately to evaluate any off-target effects. To date, we've been able to synthesize more than 4,000 antigen-specific T cell receptors across hundreds of clinically relevant targets. This section, we'll go into a little bit of a deeper dive regarding our MIRA approach, mainly because that's the foundation of identifying TCRs for cellular therapy as well as, for example, informing the design personalized vaccine. So we start off with 4 underlying principles, really basically identifying the antigens that are of interest or putative antigens to go after. These span the categories of new antigens, tumor associated antigens, viral antigens or even self antigens for autoimmune diseases. We then incubate the antigens with immune cells and proceed to sort the antigen-specific T cells using any marker of interest. This could be an activation or proliferation. We then follow with our deep sequencing immunoSEQ assay to match the T cell clones to the specific antigens of interest. And so for simplicity, here's an example of our MIRA design using 3 antigens. The key is to assign each antigen a unique address. Shown here antigen 1 in red is in pools A and B of PBMCs with ADCs, but not in pool C. Antigen 2 is in pools B and C, but not A. And so on and so forth for the rest of the antigens, which we stimulate with the Ale coated PBMCs and then proceed to sorting, in this case, using a CD137 activation marker. And then we deep sequence using our immunoSEQ assay, the antigen-specific T cell pools so that it becomes a deconvolution problem. When we identify sequence X in red in pools A and B, but not C, we confidently infer these TCR sequences, recognized and respond to antigen 1, which was also only in pools A and B. And we do this through the full inventory of TCR beta sequences and match them to their cognate antigen that they respond to. But doing this, 1 antigen or even a few antigens at a time can get tedious, so we've modified our MIRA approach to be able to multiplex and look at hundreds of antigens in -- simultaneously in 1 experiment. Here's an example of 270 query antigens, as you can see across viruses, pathogens, tumor-associated antigens and some autoimmune antigens. Here's the experimental setup when this scales. Again, each antigen has assigned a unique address as denoted by the X. And by design, there are many unassigned addresses showed up in the gray boxes. And the approach is, in essence, based on TCR identification based on clone enrichment within the assigned peptide pools. These are highlighted in the orange box, all these peptides were assigned addresses. And as I mentioned, by design, there are many unassigned addresses. This allows us to statistically power and differentiate signal-to-noise. So we can unambiguously identify a T cell response and its magnitude by the number of TCR clones responding to a given antigen. Here's another MIRA panel looking at 360 peptides that have been previously published and known to be associated with ovarian cancer. On the x-axis are plotted the specific peptides. And on the y-axis is the TCR beta cone frequency on a log scale. Each dot is a unique T cell receptor, and this is the result of screening against 1 repertoire. The beauty of our approach is the throughput and scale that we can quickly screen against additional repertoires and rapidly amass a comprehensive map of T cell receptors responding to a given antigen. For example, these TCRs highlighted in brown, recognize and respond to MUC16. Of note, the red dotted line highlights the sensitivity threshold of conventional immune assays, such as ELISPOT, which has a limit of detection of about 1 in 10,000 cells. Many of the clones below this line would have likely been missed. In contrast, MIRA's sensitivity is at least 1 in 20 million, if not 1 in 30 million cells. Here's yet another MIRA panel looking at about 200 peptides against 90 relatively common somatic mutations. And as you can see, we can identify, again, many T cell receptors against these new antigens. But the reality is that not all these TCRs are optimal as clinical therapeutic candidates, and it would take a lot of time and resources to characterize fully all of these. And so upfront, we apply our proprietary titration method that simplifies the problem from these many TCRs to the subset of TCRs that we prioritize for further investments and characterization. And so today, we standardized our MIRA approach in parallel running both a peptide based MIRA as well as a transgene based MIRA to allow us to identify and validate quickly what's naturally processed and presented in terms of epitopes. The top panel shows that many TCRs can respond to different peptides, but there's no guarantee that these are actually naturally processed and presented. And so with our transgene based MIRA, we biologically and functionally confirm which is -- which epitope is naturally processed and are able to basically categorize TCRs responding to those peptides or rather epitopes. And this really allows us to full target validation, and we've built a library of these validated targets. Our MIRA responses in terms of the T cells are also a part in the HLA context. And so shown here is another MIRA panel looking at neoantigens and tumor-associated antigens expressed by transgene, screening against the repertoire of 2 separate donors. Shown here is donor 1 eliciting a strong T cell response to neoantigen 5 in this donor CO 501 HLA context, which is not present in donor 2. Donor 2 responds strongly to neoantigens 3 and 10 in its HLA context of B1525, which is absent in donor 1. So once we identify a TCR beta to its antigen, of course, we can reconstitute the full T cell receptor and further characterize in terms of cell killing. Shown here is an example of target cell killing of one of our lead tumor-associated antigen-specific TCRs, shown in this blue curve relative to 3 available benchmarks. This experiment uses T2 cells loaded with the peptide, and you can see nice reactivity in terms of lysis at low peptide concentration. Next, we set out to confirm that this T cell receptor also kills cancer cell lines that express endogenously the antigen. In this first experiment in the gray bars, we also compared our TCR shown in blue relative to the benchmark when K562s are loaded exogenously with the peptide. And you can see that both TCRs performed similarly. However, only our T cell receptor is able to show robust cell killing of K562s that endogenously expressed either low or medium levels of the antigen. The benchmark performed similarly to the no TCR control. We're continuing to characterize a number of other attractive tumor-associated antigens but over the past year, we've really scaled our efforts to identify potent neoantigen-specific TCRs. Shown in the top panel are examples for -- against PTEN mutation, PDGFR and CDKN2A mutations in terms of functional lividity using CD69 activation. And you can see each T cell receptor shows nice reactivity at low peptide concentration. On the bottom panel, these are neoantigens. So we wanted to make sure that the TCR indeed recognizes only the mutation. And in this experiment, looking at cytokine secretion, you can see that each TCR nicely responds to the peptide that contains the mutation or the transgene express version of the mutant epitope, neither TCR recognizes either with a wild-type or no peptide control. And so we're continuing to characterize numerous attractive neoantigens specific TCRs to be able to consider, including them as cellular therapy and treat patients with a variety of solid tumor types. And so the MIRA results have shown you -- allow us to screen using blood from healthy donors, but MIRA can also importantly be used from the blood of a cancer patient. This case study is in collaboration with Memorial Sloan Kettering, where we had access to a lung cancer patients PBMCs as well as tumor. We basically designed MIRA looking at the 240 or so mutations this patient had. And you can see we can identify a number of new antigen-specific T cell receptors from this patient's blood. A subset of these neoantigen-specific TCRs were also seen to be clonally expanded in this patient's tumor, which is a nice cross validation. And then we took 2 of these new antigen-specific TCRs and fully reconstituted them. And these are initial characterizations showing nice functionality of these patient specific new antigen TCRs recognizing low peptide concentrations of the epitope. Lastly, we're partnered with Microsoft to combine our immune medicine platform with their machine learning models and algorithms. And the goal is to generate and build a TCR antigen map disease by disease, starting in oncology but also in autoimmune and infectious diseases. In fact, this week, we published a paper showing that we've fully mapped TCR antigen to the SARS-CoV-2 genome. Specifically to the cancer antigens, the goal here is to be able to predict TCR antigen binding from the TCR sequence alone. And as you can see in these preliminary results on the left, we have good accuracy, often north of 25%, up 75%, sorry, for a number of tumor-associated antigens and neoantigens. And then on the right is encouraging data showing that our prediction algorithms and models improve with larger data sets and training sets as well. So we're beginning to incorporate these prediction models and algorithm in our drug discovery workflows. In summary, our TCR discovery approach is versatile, high throughput and sensitive to be able to identify naturally occurring TCRs that are optimized for potency and safety as clinical candidates for various applications. This is what attracted Genentech to select Adaptive's platform to advance differentiated TCR based cellular therapies in oncology. We're focused with Genentech on advancing 2 products in the cellular therapy space in oncology. The first is the shared product shown on the upper workflow here, where we're leveraging Adaptive's true TCR library of TCRs that are fully characterized against common cancer antigens that are present in many cancer patients. Genentech selects the TCR to the antigen of interest and then manufactures and develop cellular therapy to be infused in a given patient that over expresses that antigen. On the bottom panel here is the workflow for the fully personalized cellular therapy approach or private product, whereby Adaptive will be conducting real-time TCR screening from a cancer patient's own blood, much like I showed the example for the lung cancer patient sample from MSK. And the real-time screening will allow us to identify a subset of the best T cell receptors against that patient's unique tumor-specific mutations. Genentech then again manufactures and develops the cellular therapy and infuses into the patient. And this has the potential we hope to be hopefully a curate -- a potentially curative approach. The -- we are very much looking forward to providing updates, particularly for our first shared product in the next 9 to 12 months as we aim for speeds to the clinic. And in closing, here's just an overview of future directions as well as drug discovery growth opportunities for Adaptive. Today, we can leverage our antigen map to be able to annotate TCRs to antigens in oncology, but also other diseases. And this has the potential to also inform target discovery for future therapeutics. Given our validated library of immunogenic antigens, we can also inform the design of next-generation vaccines and be able to monitor vaccine induced responses as well as persistence. On the TCR discovery side in terms of other drug modalities aside from cell therapy in oncology, we're particularly interested in advancing soluble TCRS, bispecifics and TCR mimetics as well as identifying antigen-specific T regs for cellular therapy applications in autoimmune diseases, such as multiple sclerosis as an example. And finally, this year, we announced the platform extension into our ability to discover antibodies. Our first application is to identify potent neutralizing antibodies against SARS-CoV-2. We're actively characterizing a number of these antibodies sourced from COVID-19 patients that have either the tail end of an acute infection or recovered, with the goal of identifying a number of potent antibodies that have nonoverlapping mechanisms of action to inform a cocktail. I'd like to thank our San Francisco Group, who generated all the data that I presented to you today as well as our Computational Biology and bioinformatics group in Seattle that did all the analysis. Thank you for your time. And if there's time, I'm happy to take any questions.

Sharon, I think we can -- I can pepper in a question here before the next talk, but hope to have additional discussion at the end. Are you doing any maturation to the natural TCRs you find, maybe trying to fine-tune the affinity when you try to make it into a kind of drug candidate?

We're not. And that is a differentiator. Given the sensitivity, specificity and the depth of sequencing, we really go after the ultra-low frequency potent, naturally occurring T cell receptors that we confirm are efficacious and safe without the need for further enhancements.

Very interesting. Well, thanks, Sharon. I think we'll come back at the end and obviously have plenty of discussion on picking the right antigen and such. But next would like to move on to Dr. Karin Jooss, the Chief Scientific Officer and Executive Vice President of Research at Gritstone Oncology. Talk about how they're identifying these neoantigens as well.

Hi. Thank you very much for the kind introduction. My name is Karin Jooss. I am Chief Scientific Officer at Gritstone. I'm going to make forward-looking statements. This is Gritstone overview. We are currently in 2 Phase I clinical trials with neoantigen based therapies. GRANITE is our fully individualized personalized vaccine where we demonstrate consistent strong neoantigen-specific CD8 T cells post vaccination that we find in all patients with clear signals of clinical benefit entering Phase II in the second half of this year. SLATE, the second program is on off-the-shelf vaccine, where we demonstrate strong CD8 T cell responses, specifically to the p531 of the epitopes in the cassette Phase II. Patients are currently enrolling, and we have optimized the product for KRAS mutants, which will enter the Phase II part of the study in the first half of '21. Our [indiscernible] is here exemplified in the center. We have a bispecific antibody program. The uniqueness is that we are targeting with the bispecific MHC peptide complexes that we identified with our epitope algorithm, and we have a clinical candidate at the end of this year. We still continue our collaboration with Bluebird Bio for tumor-specific targets as well as natural TCR -- T cell receptors for their cellular therapies. We have our in-house manufacturing facility in Pleasanton, California. And as of end of June, we have about $93 million cash. EDGE is our neural network model that allows us with over70% positive predictive value to predict HLA complexes, peptide complexes on the cell surface of tumors. This work has been published at the end of '18 -- 2018 in nature by technology, and this is currently being utilized in our clinical studies. We are continue to advance the Class II prediction. Currently, that is close to 400,000 peptides that part of our training data set with over 50 HLAs covered. The positive predictive value for our Class II prediction is close to 40%, which compares favorable to below 10% utilizing the public tools. Gritstone is using for the vaccine platform, heterologous prime-boost and has shown to try durable and potent T cell responses. For the prime vaccination, we are utilizing a chimpanzee adenoviral vector, which has shown over and over to consistently drive and activate also naive CD8 T cell responses specifically. And in order to keep the immune pressure on the tumor, we are boosting the neoantigen T cells utilizing its self amplifying RNA, which is formulated in the context of lipid nanoparticles in both of these vaccines administered intramuscularly. The unique feature about the self amplifying RNA is that it starts replicating once it's being released in the muscle cell, therefore, driving very high expression of the antigen, which usually leads to durable and high T cell tires. GRANITE and SLATE programs are quite similar in their design. They are fully personalized GRANITE program is currently being assessed in patients with lung gastric, MSS colorectal and bladder cancer, and SLATE, we are, of course, focusing on patients with high-frequency KRAS mutations, which is lung pancreatic also MSS, colorectal, and of course, all cancers with positive mutations in the tumor. The design is, as I mentioned, similar, we keep the chimpanzee adenoviral vector dose at 1/12 particle steady throughout the study, we dose escalates itself amplifying RNA vaccine, boost a vaccine from 30 to 100 up to 300. My programs, all patients are receiving anti-PD-1 and at the higher dose cohorts, we are introducing low dose subcutaneous ipilimumab to expand the T cells. The vaccines to date are well tolerated with treatment-related adverse events indicative of immune responses with fever being the only serious adverse event associated with GRANITE. What we see in SLATE can be contributed to checkpoint inhibitors that are part of our vaccine regimen and no other clear patterns evolving. I have now dive a little bit more in detail on GRANITE. This ELISPOT data from the first 8 patients, unfortunately, we couldn't perform or look at the new response in the fifth patient because we didn't get access to the blood, but what you can see here, this is an overnight. It is part looking at the T cell induction or responses in the blood of patients post vaccination. You see that we consistently induce T cell responses in all patients, which increased over time. These patients are further long compared to G8, for example. But the consistency of picking up T cell responses post vaccination in an ex vivo ELISPOT without amplifying the T cell response was our objective, and we hit that goal. Important for us to answer was the T cells that are being induced in the blood of the patients, do they track it to the tumor, which is a necessary step to provide therapeutic benefit to patients with cancer. And the answer is, yes. This is GRANITE patient 3 examples where we identified 27 TCR betas in the blood and 5 of which we also found in the tumor. And we have performed that now in 2 of the patients utilizing tumor biopsies post-treatment and find trafficking of the vaccine induced T cells into the tumor. This is an overview of GRANITE to date. As I mentioned, no DLTs have been observed many times, the treatment extends beyond apparent radiologic progression. Dose level 1 to dose level 4 here, and I would like to draw your attention to, for example, GRANITE patient 2. It's a stage IV gastric cancer patient. That patient was 8 months on chemotherapy before entering our trial at the time of study start vaccination the patient had no evidence of disease and remains without evidence of disease 12 months into the study, suggesting that the vaccine can provide therapeutic benefit -- durable therapeutic benefit to patients with minimal residual disease such as adjuvant setting. I also would like to go into detail in the next couple of slides for patient G8, which is MSS colorectal cancer patient. This patient is a 50-year-old female with MSS colorectal cancer. The patient was for 15 months on chemotherapy, FOLFOX and Avastin. Had progressive disease, received then FOLFIRI + AVASTIN for 6 months and then came on to our study. She is clinically feeling well. It's more than 112 days on our study. Her best overall response to date is stable disease at week 16, 1 liver lesion is stable, the other 1 is shrinking. We did see a very nice induction of T cell responses that are specific to the neoantigen delivered in the vaccine cassette. Interestingly, also we measured ctDNA in the blood of the patient. And what we see here is with the induction of the T service bonds, we actually see a reduction of ctDNA with some of the neoantigen-specific signals going down below baseline. And also coinciding with actually a mirror image of the kinetics of the ctDNA. We also observed with the tumor biomarker CEA, we see, first, a spike at 4 weeks post vaccination and then a very strong reduction of the CEA levels, which is currently below the level that she had at diagnosis. Importantly, since she had liver metastatic lesions, we follow ALT and AST liver functions and there's a normalization of the liver functions. So all suggestive of therapeutic benefit provided by the vaccine to this patient. These are some of her scans. These are lung lesions. And here, I want to demonstrate to you this is baseline 8 week and 16 week post initiation of vaccination. What we do see here at the first skin on 8 weeks, we see an increase kind of flaring up of the lesions before we actually see a contraction, a reduction in size. And that is actually a very consistent observation throughout all of her metastatic lesions in the lung. For the liver lesion, she has a mixed response. Here, there is a lesion that clearly reduces in size over the treatment and this is the lesion which is steady over time. So I don't want to go in detail through it. GRANITE data here is very well tolerated. We see transcend fevers, which are self resolving. We see very strong and consistent induction of lytic CD8 T cell responses in all of the patients and for 2 patients, where we looked at these T cells are actually also trafficking into the tumor. For dose level 3, we have encouraging early clinical signals, and we are pursuing MSS colorectal cancer also in the fourth dose cohort currently. Now coming to SLATE. For SLATE, we see the induction of CD8 T cell responses consistently. Here, there is 2 patients, patient 2 and 4 where we actually see them in an ex vivo ELISPOT, this patient had a existing T cell response to KRAS G12c, which actually increased on the vaccine, and this is a de novo induction of T cell responses against the KRAS 261H. For the other patients, we performed IVs and post IVs, we find very strong T cell responses against all of the KRAS mutations. An interesting observation that we made in our patients was the immune dominance of p53 in our SLATE cassette. Here you see with ex vivo ELISPOT, a very strong induction of the T cell responses against multiple p53s that are in our concept suggesting immune dominance of these epitopes in this SLATE cassette. Patients with small -- non-small lung carcinoma, all of whom had prior IO, seem to have the largest degree of clinical benefit in our study currently and are going to be pursued in the future. Two examples from SLATE. This is SLATE patient 2. The 84-year-old female with non-small cell lung carcinoma, KRAS G12C mutation. The patient was on pembro before for 4 months had progressive disease then received antibody for 10 months followed by chemo, which only for 2 months, she couldn't tolerate it and then she entered our study. She was 168 days in our study, then declined actually further treatment. She felt fatigue. The best overall response, tumor response was a stable disease with a 20% reduction from baseline. We saw a very nice induction of the T cells, as I shared with you on the previous slide, which coincided with the trough of ctDNA in the blood. In this patient, this patient had a cold tumor at baseline and post vaccination -- 8 weeks post vaccination, we turn the tumor into a hot tumor. With an increase of CD8 T cell responses or T cells in the tumor, about 7%. The third SLATE patient is a 55-year-old male with non-small cell lung carcinoma, again, KRAS G12C positive, who was in pembro and chemo before, had progressive disease then entered our study. The patient is doing very well, more than 196 days on study. The best tumor response is stable disease, again, with a reduction of 15% from baseline and induction of T cell response already 2 weeks post prime against the KRAS G12C and no T12C mutation is currently being detected in the blood with ctDNA. So the conclusion for SLATE is it's well tolerated. The side effect profile is mirroring that of a CPI into in related adverse events. We have identified the recommended Phase II dose, which is our highest dose, dose level 4. We see strong and consistent CD8 T cell responses to p53, which appears to be immune dominant in the cassette. Less consistent induction of the CD8 T cells to KRAS. We do identify them and detect them after in vitro stimulation. And we actually have identified or isolated a better epitope cassette that will into clinic shortly. And a clear evidence of a therapeutic benefit in -- especially non-small cell lung carcinoma patients. Where do we take these studies? So there's 2 clinical studies starting in the second half of 2020. For GRANITE, we are, of course, pursuing MSS colorectal cancer as well as gastric cancer. For MSS colorectal, we select patient post FOLFOX or FOLFIRI, existing going study 10 patients, and then we also continue to assess the GRANITE potency in gastric cancer patients, second line post chemotherapy, again, 10 patients to be initiated in second half of 2020. For SLATE, the observation that our vaccine induces a quite potent T cell responses to p53 led us to actually -- we will initiate in the second half of 2020, a study selecting patients with p53 mutations. It's ovarian and other cancers, again and of 10. And then for the current SLATE cassette, we are focusing on non-small lung carcinoma patients post IO or post chemo, but resistance. And then we are also moving in the first half of next year, the improved cassette forward with focus on trying interspecies immune responses for which we have evidence in preclinical studies to increase T cell responses significantly, especially to KRAS. This is now the overview, what are the data readouts over the next 18 months for GRANITE Phase I dose level 4 data will be coming in, in the second half of 2020. We won't nominate the clinical candidate for the bispecific program. Then for SLATE Phase II, focusing on p53 as well as KRAS and that is our current vaccine. We anticipate data readout in the first half of '21 and the 2 Phase II studies GRANITE in MSS colorectal and gastric cancer, we expect data in the second half of 2021. And the -- also in the second half of 2021, we expect data for the SLATE Phase II with the optimized cassette, specifically targeting the KRAS antigens. And with this, I would like to thank you very much for your attention.

Thanks, Karin. I appreciate that. And we'll try next speaker. Janet would you like to introduce. Janet, you are on mute.

It's my pleasure to introduce Dr. Alfred Slanetz, President and CEO of Geneius Biotechnology. Please welcome, Alfred.

Thank you, Janet, for that kind introduction. So do you see the slides, I just wanted to make sure…

Not yet, Alfred. If you could share your screen?

You do or you don't?

Not yet.

Okay. Yes. All right. We may need to do it differently, let's see. We're on Zoom again, share screen. Okay. This one here, share, there we go. Does that help?

Yes. Well, put it in presentation mode.

Yes.

Here we go.

All right. Good, good. So thank you, Matt, and Janet for inviting me to this. So Geneius has a little bit of a different focus than the other 2 companies, really a different approach, trying to target cancer with a product that's really the ideal cancer killing cells. We target tumor-specific antigens, whether they're viral or neoantigens or other antigens that are on the tumor. We use high affinity T cell receptors from the patient's own blood. And in 4 weeks, being to vein, we can create this product at a cost of goods, which is under $10,000. And therefore, it can be accessible to people. And up to 80% of the cells in our products are 2 of the antigens that we want with virtually no exhaustion, which is also really important. And very high memory, about 40% central and stem cell memory, which for the long-term durable responses. We target multiple antigens all at the same time because we think that this is also important in trying to combat cancer and avoid the immunoescape and hit it from multiple angles and have a highly, high expression of extravasation homing and trafficking receptors. So when the cells are infused, they can get into and spread across the tumor and create a very pro-inflammatory state across all of that. Not all T cell products are created equal, and we focus very, very hard on developing a T cell product that has all these attributes all at once. And we do this really based upon a technology platform that we have developed, which is really refocusing on previously minor parts of the immune response in making those the new major box in response. I mean these cancers have a lot of antigens, obviously, present in them. But we want to basically reeducate. So instead of focusing on the same antigens that already allowed the tumor to grow, that we can actually target all at once on multiple antigens to overcome this. Very much what happens in the first immune response that you have to a target for recapitulating that. And then essentially expanding those cells in an environment ex vivo away from the bulky tumor away from that microenvironment to actually create an effective response. So one that doesn't have any T regulatory cells, one that has the proper type of help on something called a Th1 that helps a cytotoxic response with interferon gamma and TNF alpha. One that actually can leave the bloodstream and enter the tissue. And then one, of course, that locks that in with long-term memory as opposed to what could be occurring in an escaping tumor. We do this really through not engineering the T cell product genetically as I used to do to our old company, Bluebird Bio, where I was CEO a number of years now ago. We're really reengineering the T cell population, all at once, kind of without the genetic engineering. So from a single blood draw, we can isolate the PBMCs. We can then have the viral or neoantigens or even tumor-associated antigens and stimulate against these antigens. And expand these T cells to increase the killing home easternization of memory to have a very optimized and broadly applicable product. And what we're really doing, which is, again, I think, sort of a philosophically different approach and different approaches are good, right, in science, particularly in some of those important cancer, we're all in there to treat patients, and each patient is going to respond differently to different approaches. But we're really leveraging millions of years of evolution. The full genetic diversity of the MHC, the full T cell repertoire of that individual, which, by the way, has been negatively selected as well for certain TCRs that might have been self reactive, right? So all of that's very, very specific to an individual. But we're leveraging that to basically select the best antigen targets that are specific in the tumor and create a T cell product targeting those and it's really very unbiased and works for CD4 just as well the CD8s in terms of selecting a cells. So our first product really is focused on viral targets. Epstein-Barr virus, positive lymphoma. And as you can see, we're in Phase 0 now and within a year or so, we'll be in Phase I clinical studies. And then there are other viral targets. But I think for today's talk, we're primarily going to focus on the neoantigens. And our first target there is really lung cancer. Because we think that neoantigen T cells have been demonstrated by PD-1 antagonist to be very important actually in terms of treating them. And then to be able to expand it to other solid tumors from there and more of a basket study approach. So T cells to date in academic studies have been demonstrated to have clinical effectiveness. This is a patient -- 21 patient study with relapsed and stage 4 EBV positive lymphoma, with multiple other therapies, 70% overall response, 50% complete, half of them durable at 5 years with only mild flu-like symptoms, and no chemotherapy or conditioning required. And this is another study with EBV T cells actually in nasopharyngeal. This time as an adjuvant therapy to prior gemcitabine carboplatin in patients who was recurrent head and neck candidates or the nasal fringe carcinoma. And you can see that in this, you had 62% overall survival versus 30% or 40% when you only combine additional chemotherapy. And importantly, very, very low toxicity, no grade 3, 4 events with cells, whereas, of course, when you have chemotherapy, you have overlapping toxicity. So we feel that it's definitely an approach, which is effective if one can match this and get it right. So what we did was using our manufacturing process, which I told you, is really all about creating the greatest on T cells. We were able to meet our GMP release specifications in 90% of the patients. And successfully producing T cells from 90% as opposed to about 15% in those academic studies that I just showed you, which obviously, no cancer patients going to want to wait around if it's only a 15% chance they're going to be treated. And we lowered the time of production to 3 weeks as opposed to 10 weeks. And as I mentioned, 40% memory cells versus 3% and 30% in those 2 studies by the same market methodology. So we would consider ourselves comparable to there and then a slightly higher effector function relative to those academic studies. So we think this virus indication is a very high likely to success. And when we use our manufacturing process to expand T cells, we'll see in a patient something like this, where before our process, you'll have a very low percentage of T cells responding to these antigens of the virus. When we have the antigen present in the process, you can see that 6% or 30% or 10% of the cells in this product will actually be responding by the end to the antigens that we want it to, which is very important. And they have the right phenotype. 98% of the cells in the product are CD3. CD4 and CD8s, we believe both are really important in the immune response. We have 60% at CD8, 40% CD4 is on average. CCR7, a marker of memory, this 1 is about 30%. And then as I said, CXCR 3, which leads to homo extravasation about 20% or so. And we can get from a single blood drop of 100 ml, we can get greater than 2 billion cells at the end of our process, which is enough for many, many different administrations. The cells kill. You can see that when there's antigens not present, you don't see a lot of reactivity. But when the antigen is present, you get very big killing of the blast from these patients. And in terms of where we are in the clinic, so we have our Phase 0, which has been completed. We are planning a joint Phase Ia in EBV positive cancers to develop the safety. And then an extension into EBV positive lymphoma to have 20 patients basically treated, where we hope to at least replicate this result that I was showing you in the academic studies with T cells made by our process. But the real ultimate product is obviously to target multiple personal neoantigens. And as the other speakers have spoken about, these are very, very prevalent in a number of different types of cancer, and this is an older paper from the early days of the which we all were preparing. But you can see that there's up to 500 mutations in certain types of very, very common solid cancers, all of which could be exquisite targets for the immune risks and as we mentioned before, in lung cancer, the PD-1 basically is effective in high mutational burden lung cancer based by [indiscernible] original study. About 63% overall response rate versus really 0 in low mutational burden. And why is this? Well, one possibility is that it's the number of T cells that are present in a state where they can be re-unleashed, I guess, by the PD-1 to be active and you reach that critical threshold of T cells, and therefore, you have a response. So one of the things that we're really good at, right, is growing lots of T cells to the antigen. So kind of a key study that we want to do is to -- in patients with lower mutational burden, expand large numbers of T cells and fuse them in and see, do we actually increase the overall response rate using these on T cells to this product. And that's truly important because as we know, in lung cancer, 20% or so of the cases can be treated with PD-1. Melanoma, a lot of patients can be but then beyond that, for other solid tumors, it's the MSI-high ones, maybe the 2% or so of colorectal cancer, for example, that are MSI high. And yes, everybody is getting ped one. And in combination, it certainly works with some things. But to be able to get something which expands to the other, let's say, 80% or plus is really what I think we're all after. And multiple different approaches will probably be needed to achieve that, but we'd like to see, can we do this, for example, in the low mutational burden lung cancer patients. Let me start with the next-generation sequencing panel. And we can use our T cells to probe, which of the mutations, which are specific are actually the important ones. And therefore, all of the targets are potentially actionable. As opposed to today, where you're only predicting maybe a particular targeted therapy that may be approved for different tumor or a potential clinical trial that the patient can participate in. We use the full diversity of the genetic makeup of the population. And here's what we see. You can see that when you have T cells have been expanded in this way, they don't react with wild types. You don't have the cell reactivity risk, which is very, very important. But in the mutated peptides, you can see you have about the same percentage responding in this patient to the neoantigen peptides as you have with a whole viral protein, which is astounding because the viral protein is an infection and it's immunologically completely foreign. It means that 1 mutation, the immune response are 1 group of mutations, right? And these neoantigens can generate that strong response, which we think is very, very important and bodes well for clinical application of a T cell therapy, such as ours in the clinic. And we want to target primarily clonal antigens that are across all of the lesions. And those are really the early troncal driver mutations, which are in all of the tumors from day 1 and not really the subclonal ones, right? And it turns out that if you look at the study that we did in collaboration with Karolinska now a number of years ago, that's very interesting, when you look at till, so tumor-infiltrating lymphocytes for their responses to a series of neoantigens mutations, which are in the tumor. You have a much, much more focused group in the tumor as opposed to in the blood, which is reflective of everything. Even though it's at a lower precursor frequency for each one, right, because you're in the blood, it's more representative of all of the lesions. So this is one of the reasons why we decided to start expanding out of the blood. And we've really, I mean, gotten it, so 90% of the time it works patients as opposed to when he started it out, which was a lot easier to grow hills than it is to grow PBMC based T cells effectively against multiple antigens. So for a neoantigen program, as I mentioned, we start out in lung cancer. We're focused on a Phase Ia in that. And then into the high mutational and low mutation burden in tumors plus or minus PD-1, and then a basket studies in other tumor, which we'll expand out in terms of the potential utility of the cancer product. We have IP that's early. We've filed our earliest patents back in 2010 really in the early days of this field, and it expanded them beyond that. So obviously, we had 23 different patent applications now filed broadly covering the technology broadly. And what we really envision is a whole new way of treating cancer and starting with the blood draw local community clinic, right, not necessarily only tertiary medical centers for this, which is really important because that's where 85% of people get their care. Then sequencing, and expansion in our GMP facility to tumor-specific mutations. And then QC release with 2 frozen bags shipped to the site and do 30 minute infusions, the Saline, and that's a 30-day treatment cycle. So very, very much like a cell therapy. It's almost like Apple iPhones. As we're trying to be cell therapy just like chemotherapy or infused antibodies, which is what these oncologists are really used to and to try to bring great technology to a local community for the patient. So in summary, then, we have the ideal cancer killing phenotype in 90% of patients. This is what we've achieved. Our lead product candidate is backed by clinical data from -- with a high likelihood of high complete remission rate. Solve the cost of goods for personalized T cells, making products that are going to be $500,000 or $1 million is great, and I believe, in high-value and certainly, having been at companies like Bluebird and Genentech, I think the world of this. But I think we also have to try to get as many of these products out to patients to try to benefit them. And one of them is about access and trying to do our best anyway early on to think about that as 1 of the components we look at. And we're routinely expanding diverse cells to neoantigens now in the blood and clinical trials are anticipated within the year. So thank you, and welcome any questions.

Thanks, Alfred. We'll go ahead and start a little Q&A discussion. I think a couple of different approaches here all to go after, hopefully, identifying the right antigens and developing therapies for patients. Sharon, I'm wondering as you're looking doing this deep sequencing of patients, both their TCR repertoire and kind of, I guess, the patient tumor sample as well. How confident are you that when you find a good TCR that maps to a specific mutation, that will translate to other patients where the processing might be different? Or is it not still working out? What do you think about that?

Yes, that's a great question. So of course, it's the ensuring, of course, that the patient has the right HLA that is available to create the complex of the TCR peptide MHC. As it relates to sort of the uniqueness of other sort of micron -- tumor microenvironment that may hinder the key is, obviously, of course, enrolling in patient selection to ensure that there's a threshold above which there's overexpression in that patient's neoantigen expression. And going after attractive driver mutations, I think, is key. Ultimately, the reason for cellular therapy and specifically the personalized approach, our belief with Genentech is that what's going to really be differentiated is several neoantigen-specific TCRs that target that patient's mutations.

Okay. So again, thinking about a mix maybe in the final product. Is that...

That's right. Multiple TCRs that are fully characterized and specific that target multiple mutations.

It seems like all -- everybody here is focused a little bit on trying to make sure you're not putting your eggs on 1 antigen basket, which can help with in tumor heterogeneity. Karin, question for you. Is the EDGE system primarily built both on point mutation identification or also looking at translocations and fusions or other types of mutations?

Maturity of mutations we currently have in our cassette point mutations, but we are, of course, looking at [indiscernible], et cetera, all of the above, yes.

And Sharon, I guess, maybe you can comment on if you see different frequencies of how often epitope is a good epitope based on whether it's a point mutation or translocation fusion, any type of other...

We can also look across the board. We've specifically focused for characterizing TCRs against point mutations, but we can go after deletions, et cetera, translocations. We also go after post translationally modified epitopes. So we're agnostic in that context.

Alfred, you guys, I think, ability to cover all?

Yes. No, we're able to cover all of them. And we look at all MHCs and all T cell receptors at the same time, of course, with our approach. It's natural T cell approach from that patient. So we start out with the sequencing, right, to determine something that's tumor specific, and it could be a mutation, it could be a rearrangement. It probably won't be a deletion because, obviously, if it's not expressed, we may not see it. We won't see it. But the other types of changes we look at and we multiplex. We look at them all at the same time across all the MHC and that patient's entire T cell repertoire to select the best ones. And generally, there may be about 1/3 of the specific mutations where there's really, really big responses against different MHCs. And it's about tumor heterogeneity, I think, particularly important as you get into more solid tumor rare films. But also, I just think the thought of hitting the tumor from multiple angles. It's like the ultimate, if you think about combo chemotherapy, it's the ultimate combination of immunotherapy then the tumor just can't get away, right? Once you hit it.

Yes. One thing that comes up a lot, I think, in neoantigen prediction, and I'm sure you guys have been asked this before, but the thinking of Class I versus Class II antigens getting cross presentation. Do you need a CD4 T cell response to help boost the CD8 T cell response. So curious, I guess, Karin, maybe you guys are thinking about it, and then we'll go to the other speakers.

So we initially focused on the edge prediction for Class I epitopes. And so basically, our EDGE prediction was actually -- our vaccine was matched to the EDGE prediction because the heterologous prime-boost was very strong viral vectors, our goal was to induce and boost very high CD8 T cell responses. On the Class II, we have universal Class II epitopes in the cassette. We have viral vectors that bring a lot of Class II help and now as a next -- we do find primarily induction of CD8 responses, which I'm delighted because we predict Class I epitope. So there is a clear match. We do find CD4s at a lower level. And now with the Class II EDGE prediction, we are anticipating to build cassettes going forward mixing Class I and Class II epitope. So we hope that we will have with the translational signs, actually, an answer to your question in the future.

Great. And Sharon , I know you touched very briefly on potentially looking at this mere platform to go after regulatory T cells down the line. So I guess you guys are already kind of looking at this, I understanding it correctly?

Yes, absolutely. Just like Karin said, we've also historically focused on class and validated our peptide based and transient based approach to Class I for obvious reasons. But Class II is something that we're -- we have identified and are validating appropriate epitopes. In oncology, it is a system and so there is benefits to going after both Class I and Class II, particularly some state mechanisms and some solid tumors might show advantages in going after Class II. And then ultimately, in the autoimmune space, we're building the TC antigen map with Microsoft, but also applications of our TCRs in the autoimmune space, Class II is definitely becoming a major focus of ours.

Got it. And I guess, Alfred, are you selecting out the T cell types in the process?

Yes. We actually, in all of our products have both CD4 and CD8 at the same time. So we believe in help and particularly Th1 types of help, which are the help of the cytotoxic T lymphocytes in the case of that. And in fact, because as you develop these things, I'm sure our colleagues there can say the same thing. You learn a lot about what not to do, right? And we were trying to avoid the T regulatory cells so in fact, part of our early work in IP is also on T regulatory cells though we're primarily focused in cancer. So we haven't really pushed that, but we're very good also at growing those types of T cells. When we, well, when we don't want to say you were now to do something sometimes. So we have both actually in our IP, but we're focused exclusively on really CD8 killing cells and CD4 Th1 types of cells to target multiple antigens in cancer.

Karin and Sharon, can I push you a little bit. Thinking about the industry as a whole, everyone's chasing after neoantigens, hoping to identify I mean they're the 1 that works across tumor types or makes patient-specific drugs. And less of a focus on maybe a viral antigen that we know can drive tumor genesis or it can be targeted. I mean I know Gilead has worked a little bit on HPV specific TCR. But does it almost seem like that's the place to start and making sure you have everything else figured out to get good responses where it's completely foreign antigen. And then hopefully, the learnings from that build up into the neoantigens that are closer to self maybe. Not sure who wants to go first. Sharon?

I'm happy to go. You're actually correct, there's ongoing efforts. HPV is a great target. We've identified TCRs against HPV, obviously, head and neck, cervical, et cetera, where that's expressed. So viral antigens are certainly viable target, even EBV driven tumors, for example. Certainly, EBV in the context of autoimmune diseases like lupus. So absolutely, these are valid targets in oncology and ultimately to make a dent in solid tumors, it's the neoantigens that make a lot of sense. In the personalized approach, you could imagine we're going after what's relevant in that patient's specific tumor. So it might be an HPV TCR plus set of neoantigens. So they're definitely not excluded, and so we categorize them in the bucket, if you will, viral antigens, tumor-associated antigens differentiating from specific mutations.

Got it. Okay. Karin?

May I add to what Sharon said, I completely agree. As an immunologist, as long as the epitope is foreign to the immune system, the vaccine approach usually works extremely well. Our preclinical work in monkeys was with SIV epitopes. Now we are targeting neoepitopes and see a similar induction of CD8 T cell responses in humans as we have seen in the past in monkeys. The power of neoantigens is we are delivering 20 neo epitopes in the cassette of the vaccine. We are going after press of many targeting many targets on the tumor. Often virally, derived targets have you, you usually do not have 20 targets on the tumor. So -- but they potentially are more immunogenic, right? So I think both approaches valid in use robust T cell responses. Yes.

I know for now you guys are interested in this?

Yes. No, absolutely. In fact, that's where we started, really, it was with viral antigens, such as EBV. We've also worked on when we grow HPV specific T cells, hepatitis B specific T cells for cellular cancer. So -- and I think, I wholeheartedly agree that combining viral T cells with neoantigen T cells is really -- gives you great big power in terms of targeting tumors, for sure.

Okay. And I think one last question from me for the group. When you're especially trying to translate data from 1 -- what's an observation of 1 patient to another and make predictions. How are you trying to rule out any cross reactivity with a healthy antigen? I mean we all know the kind of titan experience that plagued the field a little bit initially with some CARDIOTOX. Is that -- what do you think the risks of that are as you start kind of taking seeing you observing 1 patient and moving it to another. Sharon, I guess so you again.

Sure. We certainly -- that's a big emphasis and effort in terms of TCRs to derisk them and ensure that they don't have off-target effects. Mostly on the tumor-associated sites that these are self antigens. We do rigorous assessments across all normal tissues to derisk. Neoantigens are by definition foreign. But even so, we ensure, as I showed some data on, that the TCRs recognize only the mutation and not the wild-type version. So absolutely, that is a very, very crucial, ultimately, especially because TCRs -- so TCRs you can't really do, if you will, like the card therapy, mouse models are preclinical, so you really need to have solid safety package that passes muster and before you obviously advance that in the clinic.

Karin, I guess from a vaccine perspective, hopefully, you would get a self regulation, but...

We do get self regulation. But the whole field focus is concerned about recognition of self. We are focused on it. We look at it. We pulled out KRAS G12C specific TCRs. Looked for cross reactivity with wild type. It was very clean. Killing was very specific for the new peptide. And we look carefully in our bispecific program, again, going after shared neoantigens, for example, we look for we actually use the EDGE prediction program in modifying specific positions within the target epitope and then looking for target liabilities with the EDGE prediction throughout the human genome. And we identify them, and then we engineer around them until we have a clean binder. So I think we are very aware in the vaccine space, probably it's less concerning TCRs and bispecific. We all pay attention to it, and we prove that it's specific genomication. Yes.

Yes. So what we do, obviously, is we're starting with T cells from the individual the patient. So the immune system, the natural positive and negative selection of the T cell repertoire has already occurred. So there isn't a high degree of risk that you're going to get something as self pre active. That said, in our neoantigen program, we always screen against a wild-type gene as well as the mutation and look for that level of specificity. And thankfully, to date, we haven't seen cross reactivity but again, we always do test for it, even though it's the patient's own cells stimulate against the neoantigen and set of neoantigens.

Great. Okay. Well, with that, I think we'll wrap up. I appreciate you all taking the time. It's great work across the field on advancing these therapies. And hopefully, obviously, they make a difference.

Thanks for having us.