Hi everyone, my name is Mitchell Kapoor, and I’m a senior biotech analyst at H.C. Wainwright. Today, I’m pleased to welcome Carisma Therapeutics to present at our Cell Therapy Day. We’re going to have a fireside chat, a bit of Q&A back and forth. I’m pleased to welcome from the company Michael Klichinsky, the co-founder and Chief Scientific Officer of Carisma. Michael, thank you for joining us today. Thanks for having me, Mitchell. Great. Just maybe to start off, for those who may not be familiar with Carisma, if you could just give us a brief background to the company and the current pipeline efforts that Carisma is working on. Absolutely. I co-founded Carisma Therapeutics with Saar Gill out of the University of Pennsylvania almost ten years ago now. I think it's about nine years in. The company was initially formed around the concept of chimeric antigen receptor macrophages, introducing CARs into myeloid cells to potentially overcome some of the limitations that T cell therapies and then NK cell therapies specifically have in the solid tumor setting. We'll come back to that, I'm sure. Over the years, the company has expanded both in terms of the technology and in terms of the potential indications that we're going after with engineered macrophages. I'll talk today a little bit about our work in liver fibrosis, where macrophages in MASH patients have a defect in the ability to efferocytose or clear apoptotic hepatocytes. From a technology perspective, we've expanded from an ex vivo macrophage approach initially to an ex vivo monocyte approach, which shortened our manufacturing time down to a single day and allowed us to automate the process. We've expanded into a direct in vivo approach, both in partnership with Moderna using LNP mRNA to deliver CARs to myeloid cells in vivo for oncology, and on our own, delivering mRNA encoding efferocytosis receptors to hepatic macrophages for advanced MASH and liver fibrosis. From a pipeline perspective, we had a HER2 program, which I'll talk about some of the key learnings. We have a mesothelin targeted CAR monocyte cell therapy, CT-1119, that's set to initiate a Phase 1 study in the coming months and read out this year for patients with advanced metastatic mesothelin positive ovarian and pancreatic cancers. We have a lead preclinical program in liver fibrosis where we're delivering TIM4 to Kupffer cells using an off-the-shelf LNP mRNA. This program is CT-2401, and that's intended for patients with advanced MASH F3 disease or compensated F4 cirrhosis. In partnership with Moderna, we're developing in vivo CAR-M for oncology, where the lead program is a GPC3-targeted off-the-shelf LNP in vivo CAR-M for patients with advanced hepatocellular carcinoma. That's wonderful. Thank you for setting the stage. To go a little, I guess, the next step deeper, a lot of folks have heard about different cell therapies, more familiar with the T cell, the NK cell maybe. Could you tell us a little bit about the macrophage, the potential of the macrophage, and why it might be preferential to these other modalities? Absolutely. I'll focus on oncology initially. T cells are exquisite killers, and they have the capacity to expand upon antigen engagement. They have the capacity to persist. This has been clear in the context of B cell disease and in some cases multiple myeloma. The responses in solid tumors with CAR T cells, regardless of the CAR generation, the T cell phenotype, the method of T cell transduction, the dosing regimen, the route, the indication, combination therapy, etc., responses in solid tumors with CAR T cells have been very limited. It's not that the T cells can't kill solid tumor cells. There's ample preclinical evidence that they can. There's a defect in the ability to infiltrate the tumor. There's a defect in the ability to overcome the immunosuppressive tumor microenvironment. Critically, a CAR targets a single antigen. In some cases, folks have developed CARs that target two antigens or even three antigens, which are much, much earlier in development. Solid tumors are very heterogeneous, and there is no equivalent antigen to CD19 in the solid tumor landscape where you can eradicate the entire population of diseased cells. Take HER2 as an example of a well-validated antigen that's known to be overexpressed in solid tumors. If you look at a single, if you look at a biopsy of a single lesion, not all of the cells will express HER2. There will be significant heterogeneity. Some cells will be HER2 negative. Some cells will be HER2 overexpressing. If you only kill those antigen high cells, the antigen dim or antigen negative cells will rapidly overtake the disease, and you'll have antigen negative relapse. This has been seen time and time again. Macrophages and monocytes, in addition to being able to kill, inflame the tumor microenvironment, recruit T cells, and act as antigen presenting cells. If we have a CAR against HER2, that macrophage with the CAR will phagocytose a HER2 overexpressing tumor cell, but then process and present tumor derived neoantigens to T cells that are patient specific, polyclonal, and enable a polyclonal anti-tumor T cell response that's unique to that patient's own antigen. You can think of this approach almost as an inherent combination of a CAR cytotoxic therapy and an in situ vaccine that also inflames the TME to enable a more robust T cell response. As I mentioned, we started with macrophages. We then moved to monocytes. The mechanism of action does not change between these cells. Macrophages are differentiated monocytes. When we administer CAR monocytes, those cells become CAR macrophages directly within the tumor, within the patient, rather than ex vivo during the manufacturing process. They carry out the same exact mechanism of action. Phagocytosis, TME inflammation and modulation, T cell recruitment, and critically antigen presentation. T cells, NK cells, NK T cells, gamma delta T cells are not professional antigen presenting cells like monocytes and macrophages. Very helpful. We have some clinical data for those who may not be up to speed with the story. Could you just walk us through those different clinical data for the macrophage and monocyte platforms? What have we seen that's encouraging for these platforms? At some point, you had decided to discontinue these programs. Could you just tell us, contextualize all of this information for us so we understand better about what we've seen and where we're going? Yeah, absolutely. I'll walk you through some of the key results with the first program, which was CT-0508, an anti-HER2 CAR-macrophage. We ran a Phase 1 multicenter study for patients with HER2 overexpressing advanced metastatic solid tumors, primarily breast and gastroesophageal cancer patients. These patients were heavily pretreated. They all had multiple lines of other HER2 targeted therapies prior to coming to the study. We had patients up to 12th line. These patients have been pretty beat up through multiple lines of treatment. Nevertheless, we were able to manufacture the cells successfully, which was a key question going into this trial. This is the first in human CAR-M experience. I should say these results were all just published in Nature Medicine last week, if anyone's interested in digging in a little deeper. We were able to manufacture with a high success rate, high CAR expression. The cells took on the desired M1 pro-inflammatory phenotype. We administered cells without any chemotherapy lymphodepletion. We do not think lymphodepletion is necessary for a myeloid cell therapy. The key question was around safety. Is there a dose limiting toxicity? Is there a maximally tolerated dose? Do these cells have the same issues as other cell therapies where black box warnings of severe CRS, neurotoxicity, etc., are common? The cells were well tolerated. No DLTs, no MTD identified. We did not have any cases of severe CRS. Patients had low-grade CRS shortly after infusion, grade one or grade two, which is essentially a low-grade fever that lasted 24 to 48 hours and was either self-limiting or treated with Tylenol and standard care. We did not see any evidence of on-target, off-tumor toxicity. We did not see evidence of attacking HER2, which is known to be expressed at low levels on cardiopulmonary tissue. We did not see any cases of neurotoxicity. The cells were well tolerated. In terms of efficacy, our best overall response was stable disease. We saw a heightened rate of anti-tumor activity in patients that had HER2 3+ disease. When we biopsied patients, we screened for HER2 levels. We only enrolled patients that had HER2 2+ by IHC with FISH amplification of HER2 or patients that had HER2 3+ disease. All of the clinical benefit that we saw was in patients with HER2 3+ disease, suggesting that at this low dose level with this particular chimeric antigen receptor, there is a certain threshold of antigen that's required to trigger activation of the CAR-M. The most interesting finding, in my view, was from circulating tumor DNA. We tracked circulating tumor DNA in the peripheral blood of all patients. ctDNA is a robust biomarker that allows you to track what's happening essentially in real time, as opposed to imaging where you take an image at two months, and at four months, you have no idea what happens in between. The ctDNA revealed that in the eight patients that had HER2 3+ solid tumors, six of them had robust anti-tumor activity, but it was transient in nature. Six out of eight patients had a significant decrease. A median decrease was about 83% in circulating tumor DNA, up to 93%. The patient had the best effect, which was an inflammatory breast cancer. Interestingly, this patient had visceral, she had a cancerous rash on her chest that improved upon treatment significantly, and the timing aligned with the improvement in circulating tumor DNA. Now, what that tells us is we're able to infiltrate the tumor. We took biopsy at baseline and at week four approximately in patients to look at the TME remodeling. Is there any impact on the tumor infiltrating myeloid cells? Is there any impact on tumor infiltrating lymphocytes? We found that CT-0508 monotherapy with a single dose significantly remodeled the tumor microenvironment. It recruited T cells. Based on T cell clonality, we expanded the TCR repertoire within the TME, and the most expanding T cell clones took on an activated cytotoxic CD8 T cell phenotype, which suggests that we're not just recruiting these cells, but they're becoming activated against antigens that are present in the tumor. Difficult to imagine it would be anything other than tumor associated antigens, though we do not know exactly what those T cells are acting against. The key limitations, I believe, were the dose. We only gave these patients a single dose. While that paradigm is currently the case for CAR-19 and in multiple myeloma with BCMA CARs, I do not think that is ultimately going to be the path for cell therapy in solid tumors. There is no comparison where a single agent is given as a monotherapy with a single dose, and you expect to see durable effect. Our plan moving forward is to repeat dose patients every three weeks. We saw peak response by ctDNA at three weeks, and that tells us that we should be dosing approximately every three weeks. In our next program, CT-1119, which the Phase 1 is kicking off imminently, patients will be dosed every three weeks for up to five cycles, 2 billion cells per dose, and will be combining with an anti-PD-1, tislelizumab. Now, why did we discontinue the HER2 program? We made this decision primarily due to commercial and practical reasons. This was not driven by any safety findings, and it was not driven by any efficacy findings. Our goal all along was to get to a monocyte therapy, a CAR monocyte, which enables better trafficking and persistence and higher dose with repeat dosing in combination with PD-1. That was our plan for the follow-on HER2 program, CT-0525. It turns out that the antibody drug conjugate in HER2 developed by Daiichi, which is a fantastic drug and quite potent, leads to significant HER2 downregulation or HER2 loss at progression. At this stage, in a Phase 1 exploratory trial, every patient that we get will be coming off of an HER2. We were finding that about 75%-80% of patients were screen failures, meaning they historically had HER2 overexpressing tumors. When we biopsy and stain for HER2, the antigen is either lost or negative. In fact, folks from MD Anderson Cancer Center just presented at San Antonio Breast that in their hands, about 60%-65% of patients had HER2 loss or downregulation on an HER2. The HER2 market for overexpression is essentially disappearing. To enroll a study, one would require the ability to go international, to have dozens of sites, and for a smaller company, that simply was not feasible. With our next program, we're continuing with our plan to run a CAR monocyte with repeat dosing in combination with a T cell checkpoint inhibitor. We moved to CT-1119, which is a mesothelin targeted product. There are no approved mesothelin targeted therapies, unlike HER2, where there are, I believe, nine currently on the market. This issue of antigen loss should not occur. That would allow streamlined enrollment. The study will be conducted with a collaborator in China. That allows us to move directly to the target regimen, which is five cycles every three weeks of 2 billion cells in combination with an anti-PD-1. Great. Very helpful. Yeah, moving into 1119, can you just summarize what are the compelling features about this program and how you can overcome setbacks from the prior programs? I think investors are focused on how we can see success here and move past discontinuations. Yeah, absolutely. There are a few things. Number one, the persistence of a fully differentiated macrophage we found in the context of a solid tumor is limited. What we're administering in CT-1119 is a more naive cell, a monocyte. The longer that you culture the cells ex vivo, the less naive these cells get, and the shorter their ultimate persistence in vivo will be. We developed a process where the cells are essentially minimally touched. The cells are isolated rapidly through CD14 positive selection. It's a monocyte surface marker. The cells are immediately transduced. There's no culture, there's no activation, and the cells are then washed and cryopreserved. The entire process takes less than 24 hours. For an ex vivo monocyte approach, for an ex vivo CAR-M approach, this is as naive as a myeloid cell could be. In addition to having improved persistence, preclinically, it's about a 10x increase in the half-life of a CAR monocyte versus a CAR macrophage. These cells infiltrate tumors more actively. They express more chemokine receptors, more integrins. They're smaller spherical cells that don't get trapped in the lung or in fenestrated organs such as the liver. These cells infiltrate with about a 40x increase in tumor accumulation preclinically. Importantly, in the macrophage trial, we were able to produce about 1.6 billion cells on average. With the monocyte process, we can go up to about 10 billion cells. That gives us not only the ability to dose higher, it gives us the ability to create a bank for each patient so we can give them multiple doses without having to do multiple aphereses or multiple rounds of manufacturing. Additionally, the chimeric antigen receptor itself that was used in the HER2 program was a first-generation CAR. In the mesothelin program, the CAR that's utilized is our most potent CAR that we've developed over the last eight years, where we've screened about 2,500 CARs with our discovery team. It has an optimized hinge, an optimized transmembrane, optimized signaling domain that leads to significantly enhanced myeloid cell activation, cytokine release, killing, etc., that's still all target dependent. Additionally, we found that CD47, the do-not-eat-me signal, which is a checkpoint to monoclonal antibody-based therapies that have activity through Fc receptors on macrophages. The same checkpoint stands for CAR-M, not surprisingly. What we've employed is SIRPα knockdown. SIRPα is the inhibitory receptor for CD47. We're incorporating shRNA against SIRPα in the CAR expression cassette. We're simultaneously delivering the CAR and knocking down SIRPα. Think of it like knocking out PD-1 in a CAR T cell. We're removing the key inhibitory pathway. We've improved the CAR. We're overcoming the key direct checkpoint. We'll be repeat dosing, and we'll be combining with an anti-PD-1. This is the regimen that we've been driving toward over the years. We think that mesothelin is the right antigen. Patients will be earlier in line because there are no approved mesothelin targeted therapies. The study will be for patients with mesothelin overexpressing ovarian and pancreatic cancers. In those two indications, mesothelin is overexpressed in about 70%-90% of patients. The antigen is robust, and we think we'll be able to enroll the study quickly and get a result by the end of this year. Great. A couple of things that dovetail into the next question. 70%-90% of mesothelin patients express mesothelin. There are no mesothelin approved products. What else is risk mitigating and validating to the mesothelin target? Why mesothelin? Yeah, absolutely. Mesothelin is highly overexpressed on certain solid tumors, including ovarian, pancreatic, and a number of others. Normal tissue expression of mesothelin is very low, and it is limited to mesothelial cells, for example, cells lining the pleura or the peritoneum, where we know that CAR-M will not have access to those tissues. There have been a number of mesothelin targeted therapies explored at various stages, including antibodies, antibody drug conjugates, vaccines, CAR T cells, and CAR NK cells targeting mesothelin have all been in the clinic. This is one of the most well-validated solid tumor antigens. From a safety perspective, we think we have a clear therapeutic window, and we know that this antigen is highly expressed, and we expect that it will be able to trigger the CAR and lead to anti-tumor activity. Great. Okay. Briefly on the trial design, what should we know about that? What tumor type might enroll the fastest, do you think, based on is it based on expression or unmet need? What do you think we could see in terms of enrollment? Yeah. We will be screening for mesothelin expression by immunohistochemistry. We expect that the majority of patients will overexpress the antigen based on historic numbers. In terms of ovarian versus pancreatic, we anticipate that it'll be a mix of the two. I can't say one will enroll faster of the other. This particular center that we're using is a major center for both of these indications. We anticipate that it'll be roughly a mix of the two types. There is no dose escalation or dose run-in. We are moving directly to the desired regimen, which is repeat dosing together with an anti-PD-1. Patients will receive 2 billion cells, 2 billion CAR positive cells intravenously every three weeks for up to five cycles. What do I mean by up to five cycles? If a patient produces its autologous, so the manufactured number of cells will vary from patient to patient. If a patient's apheresis results in 10 billion CAR monocytes, they'll get five cycles. If it's eight, they'll get four. If it's nine, they'll get four cycles of two, and then a fifth of one, and so on and so forth. We anticipate that the average yield will be around 8-9 billion CAR monocytes. Patients will start on anti-PD-1 together with the first dose of the cells. Both the cells and anti-PD-1 will be given every three weeks. Once the patient exhausts their cell bank, they will continue on anti-PD-1 monotherapy for at least a year or until progression. Great. Okay. Moving to in vivo, could you talk about the Moderna partnership, what we've seen out of that so far, what are the economics, and just what have we seen with in vivo efficacy? Yeah, absolutely. This is very exciting, of course. The same mechanism of action that I described of engineering myeloid cells with CARs to target tumor-associated antigens to drive phagocytosis, TME activation, and antigen presentation to T cells. Rather than doing this with an ex vivo cell therapy approach, this is a direct in vivo approach with LNP encapsulating mRNA encoding CARs. The LNP is myeloid tropic, so it will primarily transfect macrophages without transfecting other cell types to a significant degree, arming them with CARs to attack the cancer. The lead program is a GPC3 targeted in vivo CAR-M for hepatocellular carcinoma. We presented data at SITC, and for those interested, there is a poster on the presentation section of our website. We showed very clearly that we are primarily generating CAR-M with this LNP in preclinical models. It's well tolerated, and it's driving significant anti-tumor activity in models of metastatic liver cancer, where we take animals with metastatic disease. We administer CAR-LNP or control LNP once a week, and we're seeing an improvement in survival. We're seeing a regression of overall tumor burden and an almost complete clearance of metastatic nodules in the liver, which was quite impressive. The program is in Moderna's hands as we had the development candidate nomination, and they're driving it forward. Unfortunately, I can't guide to a particular timeline at this stage as the guidance will come from them. We had a DC nomination, and it's moving toward IND for an HCC Phase 1 trial. We're also working on additional targets in oncology. As that one moves forward, we've turned our attention to follow-on targets which haven't publicly been disclosed at this stage. In terms of the economics, when we entered into the collaboration a few years back, Moderna paid Carisma an upfront of $80 million. They provide research funding on top of the upfront for all activities. There are $3 billion in potential milestones across 12 programs and a sales royalty for each of the programs. Can you just tell us a little bit about the GPC3 target and how that's risk mitigating? that's risk mitigating? What's risk mitigating for GPC3? Absolutely. GPC3 is highly overexpressed on the majority of hepatocellular carcinoma. In normal tissues, GPC3 is absent. There are some fetal tissues in which GPC3 can be found. As far as solid tumor antigens go, it is one of the cleanest antigens out there. Like mesothelin, there have been a number of approaches targeting GPC3, including most recently a CAR-T experience. That is AstraZeneca and AbelZeta C-CAR 031, I think, where they treated patients with pretty high doses of GPC3 CAR T cells autologous. They saw that the cells were well tolerated. There is no evidence of on-target off-tumor toxicity. You see your typical CAR-T-like toxicities of CRS or lymphodepletion-associated events, which we do not anticipate with a CAR-M approach. GPC3 is well validated. It has been explored with multiple modalities, relevant modalities, and it has a favorable protein expression profile. Okay, great. Briefly on the liver fibrosis program, another one that you all are highly focused on. Could you just touch upon that? What should we know from a high level? Absolutely. We have expanded into liver fibrosis, specifically advanced MASH, where macrophages in the liver are key culprits in the progression of a fatty liver to a fibrotic liver. In MASH, the presence of fibrosis is what drives organ dysfunction, ultimately organ failure, progression to cirrhosis, development of hepatocellular carcinoma. There is currently one approved drug in MASH from Madrigal. It has a metabolic mechanism of action. All of the follow-on agents, the GLP-1s, the FGF21 agonists, they are all metabolic in mechanism of action. Any anti-fibrotic effect that is seen is primarily from an overall improvement in the health of the liver. In other words, you reduce the lipid profile, fat accumulation, and naturally see some fibrosis disappearing. Our goal is to directly target fibrosis, initially in the liver and then beyond. Macrophages are responsible for a process called efferocytosis. Efferocytosis is the clearance of apoptotic dead cells. In MASH, these dead cells accumulate, and they cause fibroblast activation and collagen production. They cause macrophage activation and inflammation. This is due to the loss of a specific receptor called TIM4, which we systematically identified together with a collaborator at Columbia University, is lost in patients with advanced MASH. Our approach is to restore TIM4 expression in Kupffer cells, which are hepatic macrophages, to restore the ability to clear dead cells and reverse the progression of MASH. In preclinical models, we've seen robust improvement in fibrosis, in inflammation in the liver. We see regeneration of the liver. We see improvement in metabolic pathways. We're quite excited about this approach. It's novel. This concept of loss of efferocytosis is not unique to liver fibrosis. Alveolar macrophages in the lung have defective efferocytosis and idiopathic pulmonary fibrosis. Cardiac macrophages and atherosclerosis and heart failure. This seems to be an emerging central node. There's a defect in macrophages, and with our platform, we can solve it directly. We're doing this with an off-the-shelf approach. We're delivering LNP mRNA encoding TIM4 to Kupffer cells in the liver to restore natural TIM4 expression, which is lost in these MASH patients. In terms of the program, we expect to have a development candidate nominated in the first quarter, so very shortly. We believe we can drive this program into the clinic with an IND or IND equivalent in the first half of 2026, with initial data later in the year in 2026. Great. Thank you, Michael. The last thing before we wrap up, just you can tell us about the catalyst for the next 6 to 12 months. What should we be looking for from Carisma? Yeah, absolutely. On the fibrosis front, we'll have a development candidate nominated in the first quarter of this year. We will have a regulatory filing in the first half of 2026. On the oncology front, CT-1119 is entering a Phase 1 trial in the first half of this year. We expect to have 12 patients of safety and efficacy follow-up data by the end of this year in 2025. Those are the key milestones on those key programs. The Moderna partnered programs are moving forward, but unfortunately, I can't provide clear timelines on those as it's in their hands at this stage. Makes sense. Michael, thank you so much. This has been incredibly helpful to us. Thank you to all of our investors who have dialed in as well. We really appreciate your time. Thank you, Mitch. Thanks for having me.