Welcome, everyone, joining today's Cellectis Third Quarter 2026 Earnings Conference Call. At this time, all participants are in a listen-only mode. Later, you will have the opportunity to ask questions during the question-and-answer session. To register to ask a question at any time, please press star one on your telephone keypad. Please note, this call is being recorded, and we are standing by if you should need any assistance. It is now my pleasure to turn the meeting over to Arthur Stril, Chief Financial Officer and Chief Business Officer. Please go ahead.
Good morning, and welcome, everyone, to Cellectis Third Quarter 2026 Business Update and Financial Results Conference Call. Joining me on the call today are Dr. André Choulika, our Chief Executive Officer, and Dr. Adrian Kilcoyne, our Chief Medical Officer. Yesterday evening, Cellectis issued a six-K and press release reporting our financial statements for the nine-month period ended September 30, 2026, and a business update. The report and press release are available on our website at cellectis.com. As a reminder, please note that this presentation contains forward-looking statements within the meaning of applicable securities laws, including statements regarding our strategy, development programs, financial condition, and expected cash runway. These forward-looking statements are based on current expectations and assumptions and are subject to various risks and uncertainties that could cause actual results to differ materially from those expressed or implied by such statements.
For a discussion of these risks and uncertainties, please refer to our filings with the U.S. Securities and Exchange Commission and our annual financial report prepared in French. Except as required by law, we undertake no obligation to publicly update or revise any forward-looking statements contained in this presentation. I would now like to turn the call over to André.
Thank you, Arthur. Good morning, and thank you, everyone, for joining us today. Two months ago, we announced the strategic transformation of Cellectis into an in vivo gene editing company. Since September, our team has been fully committed to this new chapter. We're starting with severe dyslipidemias, where the unmet medical need remains high. In severe hypertriglyceridemia, persistently elevated triglyceride levels can expose patients to serious complications, including acute pancreatitis. While in severe hypercholesterolemia, prolonged exposure to very high LDL cholesterol substantially increases the risk of cardiovascular disease. For many of these patients, treatment can mean daily pills or repeated injections over the long term, without always achieving sufficient lipid control. Our approach is different, a single injection designed to deliver a potential long-lasting effect. This is what HEAL-101 and HEAL-201 are being developed to achieve.
Let me start with .HEAL-101, our in vivo base editing candidate targeting the APOC3 gene for severe hypertriglyceridemia. At the European Society of Gene and Cell Therapy in October, we presented an expanded preclinical dataset. After a single administration of .HEAL-101, edited up to 70% of APOC3 gene, with activity restricted to the liver and constantly reduced triglycerides by 76% in a diseased model. Taken together, these results strengthen our confidence in .HEAL-101 as we continue to advance the program toward the clinic. Adrian will walk you through the details in a moment. The initiation of the investigator-initiated trial in China is on track, with preliminary clinical data expected in the second half of 2027. Turning now to the second lead program, .HEAL-201, an in vivo gene editing candidate targeting PCSK9 gene for severe hypercholesterolemia. It is designed to durably silence PCSK9 without altering the DNA sequence.
Our preclinical data are encouraging. In vitro, in human hepatic cell lines, our PCSK9 epigenetic modulator achieved high on-site epigenetic editing, with a mean above 90%, and no off-site changes in gene or protein expression detected. In vivo, it reduced mean plasma PCSK9 levels by more than 90% in a humanized liver mouse model. Building on these results, we are now preparing the next step, which we plan to be phase I, with preliminary clinical data expected in the first half of 2028. Now, I would like to say a word about our legacy cell therapy programs. Stopping the development of lasme-cel and eti-cel was not an easy decision. Behind these programs are patients and families, as well as the investigators and site teams who have been part of these programs over the years. We are deeply grateful to all of them.
Our priority now is clear, to advance .HEAL-101 and .HEAL-201 toward the clinic while maintaining the financial discipline required to reach the key milestones ahead. Alongside the strategic focus, our partnership with AstraZeneca, Servier, Allogene, and Iovance remain an important source of value for Cellectis. With that, I will hand over to Adrian, our Chief Medical Officer, who will take you through the .HEAL-101 data in more detail.
Thank you, André. Good morning, everyone. Let me start with a brief reminder of how .HEAL-101 works. .HEAL-101 is a TALE-based editor designed to make a single C-to-T conversion in the APOC3 gene. This creates a stop codon that prevents production of the APOC3 protein without cutting DNA or creating a double-strand break. The data we presented at ESGCT built on what we first shared in September and addresses four key questions. One, how efficiently .HEAL-101 edits its target? Two, how precisely that editing is? Three, whether it translates into a functional effect? And finally, where that activity occurs. In a human hepatic cell line, .HEAL-101 achieved a mean APOC3 editing of approximately 90%, while generating minimal indel and no detectable DNA double-strand breaks at the target site.
Additionally, no detectable editing was observed at the top putative off-target sites, and this editing resulted in a marked reduction in APOC3 protein secretion. In a normal lipidemic humanized liver mouse model, a single intravenous dose resulted in a mean APOC3 editing of approximately 65%, with editing levels reaching up to 70%. This was accompanied by lower circulating APOC3 levels and lower triglycerides, demonstrating the effect from gene editing to protein reduction and ultimately to the lipid phenotype. Importantly, we observed no significant increase in ALT levels compared with untreated controls. In a hypertriglyceridemic human APOC3 transgenic mouse model, which more closely reflects the disease setting, HEAL-101 reduced APOC3 levels by 67%. This translated into a 76% reduction in triglycerides, consistent with the efficacy we first reported in September. Importantly, editing was detected in the liver with no significant editing detected in the other tissues analyzed.
This liver-restricted activity is consistent with the intended profile of our liver-directed LNP delivery approach and is an important consideration as we advance HEAL-101 towards the clinic. Taken together, these preclinical data show a consistent profile, efficient and precise editing, a functional effect on triglycerides, and editing activity restricted to the target organ. As André mentioned, our goal with HEAL-101 is to develop a single administration therapy with the potential to provide a lasting reduction in triglycerides. The initiation of the investigator-initiated trial in China is on track, with preliminary clinical data expected in the second half of 2027. We look forward to sharing more details on the trial as it progresses. With that, I'll turn the call over to Arthur, our Chief Financial Officer and Chief Business Officer, for our third quarter financial results.
Thank you, Adrian. I will now walk you through our financial results for the third quarter of 2026. As André highlighted, this is a transition quarter. As part of our strategic transformation, the exit of Cellectis lasme-cel and eti-cel clinical programs is proceeding as planned. Enrollment has stopped, site closeout activities are underway, and treated patients will continue to be monitored in accordance with protocol-specified follow-up requirements. Furthermore, the realignment of the company's operating model is ongoing. In connection with this initiative, the company currently expects its global headcount to be reduced by up to 70% from its pre-transition level over the course of 2026 and 2027, subject to the completion of the applicable French Works Council consultation process and compliance with applicable French and U.S. labor law requirements.
Going forward, we expect Cellectis cash spending to be primarily focused on advancing our pipeline, including HEAL-101, HEAL-201, and potential new product candidates, as well as funding one-off restructuring costs. As of September 30, 2026, our cash equivalents, restricted cash, and fixed-term deposits classified as current and non-current financial assets amounted to $149 million, compared to $211 million as of December 31st, 2025. The $62 million change was primarily driven by payments to supplier of $37.9 million, payroll-related payments, salaries, bonuses, and social charges of $37.5 million, lease liability payments of $7.3 million, repayment of $4.0 million under the PGE loan, and capital expenditures of $1.9 million, partially offset by $21.4 million of cash received from customers and $5.4 million of interest received from our financial and cash equivalent instruments.
Consolidated revenues and other income were $19.9 million for the nine-month period ended September 30th, 2026, compared to $67.4 million for the nine-month period ended September 30th, 2025. The $47.5 million decrease year-over-year was primarily driven by the level of activities performed under the research plans of the AstraZeneca joint research collaboration agreements during the period. Consolidated R&D expenses were $108.8 million for the nine-month period ended September 30th, 2026, compared to $69.1 million for the nine-month period ended September 30, 2025. The $39.7 million increase was primarily driven by a $34.9 million impact from the strategic transformation announced in September 2026, including a $26.6 million one-off impairment loss, $6.2 million of termination benefits, and $2.1 million of costs related to the closeout of the lasme-cel and eti-cel programs.
Excluding the one-off impacts, research and development expenses increased by $4.8 million in 2026 compared to 2025, primarily driven by a $1.4 million increase in non-cash stock-based compensation expense and $2.1 million increase in purchases and external expenses, mainly attributable to the higher clinical development costs incurred before the strategic transformation. Consolidated SG&A expenses were $17.5 million for the nine-month period ended September 30th, 2026, compared to the $15.0 million for the nine-month period ended September 30th, 2025. The $2.5 million increase was primarily driven by a $1.1 million impact of the strategic transformation announced in September 2026, comprising $0.9 million of termination benefits. Excluding the one-off impacts, selling general and administrative expenses increased by $1.4 million, mainly due to higher stock-based compensation expense.
Turning to our financial results, the consolidated net gain for the nine-month period ended September 30th, 2026, was $5.2 million, compared to a net financial loss of $25.6 million for the nine-month period ended September 30th, 2025. The $30.8 million difference reflects a $10.8 million increase in financial income and a $20.0 million decrease in financial expenses. The $10.8 million increase in financial income was primarily attributable to a $14.3 million increase in non-cash gains recognized from fair value measurements, mainly reflecting a gain from the fair value measurements of the tranche A, B, and C warrants issued to the European Investment Bank, EIB, during the period, compared with a fair value loss in the same period in 2025.
Partially offset by a $2.7 million decrease in interest income earned on cash equivalents, and financial assets, and a $1.0 million decrease in foreign exchange gains. The $20.0 million decrease in financial expenses was primarily attributable to a $23.7 million decrease in foreign exchange losses, mainly resulting from the appreciation of the US dollar against the euro, a $5.9 million decrease in losses from fair value measurements, as the tranche A, B, and C warrants issued to the EIB generated a fair value gain over the period, and a $9.7 million increase in other financial expenses, mainly due to the expected credit loss allowance related to a financial warranty. For details on our consolidated net loss attributable to shareholders for the nine months ended September 30th, 2026, I invite you to refer to our press release.
As André mentioned, once our realignment is fully implemented as planned, we will have a cash runway into the second half of 2028. Importantly, this is anticipated to take us beyond the expected first clinical data from HEAL-101 in the second half of 2027 and from HEAL-201 in the first half of 2028. Our priorities are clear: to execute our strategic transformation with financial discipline and advance HEAL-101 and HEAL-201 towards their key clinical milestones. With that, we welcome your questions.
Thank you. If you would like to ask a question, please press star one on your keypad. To leave the queue at any time, press star two. Once again, that is star and one to ask a question. We will take our first question from Amin Makarem with Jefferies. Please go ahead. Your line is open.
Hi. Thank you for taking our questions. Couple of questions from us. For .HEAL-101, what level of triglyceride lowering would be enough to move into a global development? Does it need to be closer to that roughly 80% reduction we see from plozasiran, or would a 50%-60% reduction be enough? For .HEAL-201, what LDL-C reduction did you see alongside the greater than 90% PCSK9 suppression? Those would be helpful if you can address those two.
Thanks for the question, Amin. This is Adrian. I'll take the first part of that. Our anticipation is that the efficacy we will see will be at least as good as the current treatments, and I think that's our expectation. There is another aspect to it, which I think we like to talk about in that. One is the percentage reduction, the other is the proportion of patients hitting target, and that's incredibly important. We want to really optimize the number of patients hitting under that target of 500 for a number of reasons. One, the risks associated with hypertriglyceridemia are really based on it being over 500. That is critical. Certainly in the high-risk patient population and previous acute pancreatitis, once you get under that 500, your risks essentially are eliminated. That's important to also consider.
We also know that between the 200 and 500, when we look at any outcomes, we're not able to see a real benefit. So, two aspects. We want to optimize the number of patients who hit below target, but we have no reason to believe, based on our current data package, that we would be anything other than at least as good as the current therapies available. Again, with the .HEAL-101 program, to remind people that this is being run in China. The regulations in China have changed rather dramatically this year, and this is supported by a very robust preclinical package. For the PCSK9 90% reduction, we haven't gone public on the LDL-C reductions yet. That will be coming in June for us. But we'll certainly follow up with you.
Helpful. Thank you.
Thank you. We will move next with Jack Allen with Baird. Please go ahead. Your line is open.
Great. Thanks for taking the questions, and congrats to the team on the progress made over the course of the quarter. Two quick ones from our end. The first I wanted to ask on HEAL-101 again. I just wanted to, I guess, poke the management's brains a little bit here on how they selected APOC3 versus ANGPTL3. I know there are gene editors going after both these targets. I am just curious what drove you towards APOC3 and what attracted you to that target. Then briefly on 101 and the study startup activities, have you identified the sites you are going to go into in the investigator-led study in China? Then, any color you can provide on the dose escalation plans there? Do you expect to start at efficacious dose, and how many doses could you escalate through? Thanks so much.
Great. Well, I can start on that end. I think there are three questions in there. You may have to remind me of some of those we are going through, Jack. The first, why APOC3 versus ANGPTL3? Given that our indication was hypertriglyceridemia, the first thing we did is actually extensive market research to see what is required in the marketplace. Specifically for hypertriglyceridemia, the general feeling, based on that extensive market research across cardiologists, endocrinologists, internists, was that optimal reduction in triglycerides is important. The general feeling was that APOC3, in terms of specifically targeting triglycerides, was still superior to ANGPTL3 in that regard. So, it seemed to be the right indication, the right target for specifically hypertriglyceridemia. In terms of the sites in China, yes, we have chosen where we are going to be running this in China.
We have not got agreement to make that public with the Chinese in China, but once we have that, we will share them with you. But suffice to say that we are very satisfied that we will be able to recruit at a pace, given the access to a huge number of patients. And it is the PI in the trial would be considered probably the number one treater in this space in China, and has done all the other gene therapy-related programs in the past. So, extensive experience. And they specifically wanted to do this program because they like the technology. They perceive it to be potentially safer based on no nicks in DNA. The third question was dose escalation. That is really important to us.
We have, again, in choosing the sites in China, the reason we chose the PI was their desire to, yes, start at a safe dose, but also an incredible desire to start at an effective dose. They have developed their own modeling specific for these types of treatments that allows them to look at, again, the very strong preclinical package we have and do extensive modeling to allow us to start at essentially a slightly higher starting dose. We do still anticipate dose escalation, of course. It is a phase I program. However, we do anticipate the dose at which we start will be effective. Now, it may not achieve the optimal efficacy, but it does lead into what the expectations for data readout are for 2027.
Once we start the dose escalation in Q1, that first cohort, we anticipate, will have a level of efficacy, and we would anticipate then by the end of year, you will start to see the durability of that effect, which is a unique proposition from Cellectis. Indeed, if we start recruiting patients in Q1, by the end of the year, we should be starting to see the nine-month type of durability of effect.
One thing I'd like to add, hi, Jack, it's André, is APOC3 is a way more effective metabolic intervention under reduction of triglyceride than ANGPTL3, which is effective, but not as effective. Our goal here is not to have a product that lowers triglyceride and LDL, but essentially to try to effectively treat potential risk of pancreatitis. We think this is the best adapted target, and there is a clear POC with ASOs, RNAs, on this target that is very valid. There is a real unmet medical need on this side of a very drastic lowering because of the metabolic action.
Great. Thanks so much for all the color.
Thank you. Our next question comes from Yanan Zhu with Wells Fargo. Please go ahead.
Oh, great. Thanks for taking our questions. On the APOC3 program, could you potentially compare the APOC3 and triglyceride reduction between your base editing approach with nuclease editing approach? Just wondering whether there could be any potential differences. Could you also touch on the LNP that you're using and your level of confidence in terms of liver safety? Thank you.
Well, thank you for the question, Yanan. This is André. The effectiveness of either making base editing versus cutting DNA with a double strand break, is like the deepness of action, is very similar with a slight upside for base editing. The difference essentially is that you need to have a replication inside the cell, but liver cells replicate quite intensively, and therefore, you have a very effective base editing. We decided to go for in vivo gene editing with base editors and epigenetic modifiers because the risk of snapping DNA and having what's called a non-homologous end joining that would induce a mutation in the gene, is of high risk of potentially having a very, very low rate, but still a rate of translocation, or chromatolysis, or loss of heterozygosity.
It means you can lose potentially an arm of the chromosome or potentially having the chromosome recombine somewhere else, et cetera. If you just use a deaminase, this will just switch one base for another, and after this, there's replication. So the efficiency is similar, if not slightly better with the base editor, but the risk is so much lower that there is actually the main reason to use a nuclease or double strand break, such as a TALEN, N standing for nuclease, it would be essentially to do either gene insertions or gene replacements or large portion of deletion. But if you just want to interrupt a gene, you can either interrupt it through methylation, the deacetylation, or essentially base editing, inducing either a mutation in the coding sequence or at splice sites.
Thank you.
I don't know if this answers the question.
Yeah, that's helpful. Thanks, André. How about the LNP liver safety?
Yes, it's definitely a safety approach.
Yeah, as I say, we do have NHP data, we have mouse data. Suffice to say, the liver toxicity has been very reassuring.
Great. Thanks for the color.
Thank you. Our next question comes from Lukas Shumway with Barclays. Please go ahead.
Hi, thanks for taking the question. Real quick on the data reported in the preclinical poster and here today versus September for .HEAL-101. Can you talk about what the difference is in the dataset in the disease model there?
Well, actually, there are two different type of models that have been used. The first one is essentially a model that is a human liver that has been grafted on a NOD mice. You have the human gene, and of course, the mouse gene would not work in this setting, and so you need to really test the compound on a real human gene. It has been tested in the liver. The problem is that these mice do not develop SHTG, so they're not sick. They're differently sick, actually. They have an immune deficiency, but they're not really sick. The model that has been used, the second model that has been used, essentially a model that is a transgenic mice, so it's a mice liver, but that have a transgene in multiple copies.
I think it's approximately over 50 copies of the human APOC3 gene. Therefore, as there's a lot of APOC3 that is secreted by the mice, the mice develop an SHTG. Therefore, it allows us to show an effective reduction of triglyceride in these types of models, and you can make a clear correlation, and that's what the poster shows, between the reduction of APOC3, which goes back to the first question on the choice of APOC3 versus ANGPTL3. We see that even over 50 copies in a single hepatocyte can be very efficiently edited by the TALE-based editor and reduce the triglyceride. There is a clear correlation between the number of edits and also the reduction in the drop in triglycerides. So that's essentially what has been shown essentially working between a human liver and really diseased mice.
Sorry. Okay. What I meant was, in that disease model, the overexpression model, what's the difference between the data now and the data in September? Because the numbers are different.
Well, essentially the correlation that you can do between the reduction triglycerides, there was more points of measures that have been done and more dose escalation and the way it has been shown. It's not, of course, drastic, and we should say that it's essentially this type of, the work that we presented in September versus the poster is not that much meaningful, but there is more of a ground of data that has been shown there.
Okay. Thanks.
Thank you. Once again, that is star and one on your telephone keypad if you would like to join the queue. We will move next with Silvan Tuerkcan with Citizens JMP. Please go ahead.
Hi, this is Joshua Werman on for Silvan. Thanks for taking my question, and congrats on the updates here. Maybe just going back to the LNP briefly, is this a delivery system that has been used clinically before, either by Cellectis or elsewhere? Then how can maybe that help us be informed on the dose escalation? Additionally, building off the decision to target maybe APOC3 over ANGPTL3, how are you thinking about the ultimate addressable population for 101? If the POC is robust, where do you think the biology could support expansion beyond SHTG? Thank you.
Well, first of all, the first question was?
Was on the LNP.
Oh, the LNP. Yes. Sorry about that. I was thinking about the second question. Yes, the answer is yes. We've been working currently with a biotech company, not a CDMO. We're going to transfer to a CDMO after working with the LNP, but a biotech company that has been using these LNPs in the clinic for transduction of liver cells, but not with mRNA, or RNA, not especially messenger, sorry, with RNA for liver transduction with very good safety profile but very high efficiency. It's like a next generation type of charged type of lipids that have been put and nicely decorated too, that gives a very good transduction level in other type of settings in the liver with RNA, and that are currently in the clinic.
Therefore, this is why we choose this, and there is a very nice modeling that has been done with animal models, such as NHP and mice, that we're currently conducting, that would allow us to translate this and arrive when we'll start in the clinic with probably the best, most optimized type of dose leveling at a time. So, in between now and the time where we'll have the IIT active, then this modeling will be between the human data that has been obtained so far, plus the animal models, NHP and mice, we'll be able to do a very good dosing with this. But they've been really well proven for lack of toxicity.
I suppose the other thing is, because it has been used clinically before, we have the biotech's NHP data, which we've been able to recapitulate, which is very helpful, but also it allows us to go to a dose level, which I think is part of your second question, it allows us to think about a potentially higher starting dose as well, based on the modeling. So, it's given us that flexibility, which has been helpful. The second part of the question is the size of the market, I think, ANGPTL3 versus APOC3, and where could you go beyond? Well, our belief is an APOC3 target is very much a hypertriglyceridemia target, but this market is incredibly large. So if we look at the target population that we're looking at, it's about 1.1 million across U.S. and EU5, or EU4 plus U.K.
This is an incredibly large market for an APOC3 target. Again, we've done a lot of market research to see where would this product fit within that market, and the view is, based on the potential durability of response, this would be a best-in-class product. We would be launching this into an established market, being best in class.
Yeah, the comparison between ANGPTL3 and APOC3 is that when you're treating severe hypertriglyceridemia, you want a product that totally focus on this, and you will choose the most efficient one. If you're treating, for example, LDL, so hypercholesterolemia, which is also active for ANGPTL3, I think you'll probably use PCSK9. I know that this product might cover both indications, but as the reduction, it's not the highest performer in both, but these are two different type of diseases.
Yeah.
Usually it doesn't. The mix between the two, the overlap is quite small. You're better used to a very specialized drug in each indication.
Yeah, and two other important point when you think about peak sales is we talk about addressable market, but you also have to think about pricing. If, as we are thinking about these to be long-lasting treatments, and if you look at the pricing per year per patient of other therapies in the SHTG space, I think proving durability in years will allow us to extract bigger pricing in terms of the per year cost of current therapies. Then finally, when you think about cost of goods, this is LNP, mRNA formulated in LNP, which is extremely cheap. The gross margin potential of these products is absolutely tremendous. Don't only think about the addressable market in terms of number of patients, but I think these therapies will unlock pricing and cost of goods opportunities.
Absolutely. Great. Thanks, Simon.
I think we can add. There's a final thing to add, sorry, which was we have looked at the market research in this regard, regarding the positioning of an ANGPTL3, and the view was this was more a mixed dyslipidemia product. The number of patients that have overlapping LDL-C with hypertriglyceridemia is actually relatively small. So they see the ANGPTL3 probably preferentially moving into that hypercholesterolemia mixed dyslipidemia space. They believe that APOC3 is superior in terms of triglyceride lowering.
Great. Thank you.
Thank you. At this time, there are no further questions in queue. I will now turn the meeting back to André Choulika for closing comments.
Well, thank you everyone for attending this quarterly financial statement for Cellectis and update for the quarter. As we said, we keep on the same timing as we said in September, so we are really excited about what is going to happen in 2027. So keep updated on the company on a regular basis, and Cellectis will be present at many conferences if you want to follow up on the discussion. In the meantime, we will continue the execution, and I would like to thank you for your time.
Thank you. This brings us to the end of today's meeting. We appreciate your time and participation. You may now disconnect.