Sana Biotechnology, Inc. (SANA) Earnings Call Transcript & Summary

Thank you all for joining. Pleasure to have Steve join us. I still -- I always mention this. Steve did the very first fireside I ever did at any conference, which was back as CFO of Juno, and we've been talking about cell therapy since. And I feel like some of the themes we've been discussing even at Sana are now very prominent in vivo CAR-T in particular, across a lot of companies and a lot of players emerging.

So Steve, why don't you kick it off? And I think maybe just remind us about the platform, but also remind us about how Sana is playing in vivo CAR-T because it's fashionable and everyone is playing in there, but I feel like you guys are not necessarily just starting up in there, for example.

Yes. First of all, thank you, everybody, for joining. Thank you to Evercore for having us. I'm sure you guys recognize, we may make a few forward-looking statements, so feel free to check out our 10-Q for the risk factors. The company actually -- we were founded on these couple of ideas. One is to be able to hide cells from immune detection and make cells to replace missing cells. And the other is to be able to deliver payloads to specific cells in vivo. And both of these seem to be working right now. And so our -- probably the asset that people care the most about or we spend most time on is a drug called SC451. And the goal of this is a functional cure of type 1 diabetes. Take a step back. Type 1 diabetes is a disease of about 9 million people. It's growing very rapidly, about 15 billion by 2040. They have -- what happens is a person's immune system attacks and kills the pancreatic beta cell, that's known, right? The beta cell is the only cell in the body that makes insulin. So up until 102 years ago, a patient would die pretty quickly once they were diagnosed. Their cells would just starve to death. And with the invention of insulin over the last 100 years, it's gradually gotten better, but it's still a disease where with the best care, a person is going to live 10 to 15 years less. It is a life where you're making 140 executive decisions every day about what to eat and how many carbs am I having and how much insulin should I take. And it's a disease where the whole time, you're worried about too low of blood sugars, which can lead to death very quickly or too high of blood sugars, which lead to long-term problems of blindness, amputation, heart attack, stroke, kidney failure, all kinds of things. And so it's a giant problem. And there -- what we've learned over the last 20 years is that a group led by James Shapiro in Canada showed that you could take islets -- pancreatic beta cell is a missing cell and islet is a beta cell plus a support structure, let's just call it that, right? So you can take isolate islets from a pancreas for someone who recently died and transplant them. And if you give enough of those islets and you give immunosuppression like you would get with an organ transplant, patients can live for a decade plus without any insulin. Normal blood sugar is off insulin. The problem is it's not a scalable supply source. It's a very variable supply source. And there aren't that many people for whom lifelong immunosuppression is better than lifelong insulin, right? So thousands of people have gotten this, but it's not a curative. It's not a scalable solution. Over the last couple of years, several groups have shown that you can take stem cells, pluripotent stem cells and grow them in a manufacturing process into islets and transplant them. And they really do work. It's a more scalable solution. It's a more predictable solution, but you still have the problem of immunosuppression. So it's not very likely to be something that's broadly utilizable. What we showed beginning through the course of this year has been that we can transplant -- we can gene-modify islets, transplant them into a person and these cells persist and function with no immunosuppression. It's actually a transformative result. I've never seen allogeneic cells transplanted without immunosuppression. And it gives us all the pieces put together to make a functional cure for type 1 diabetes, which is a single treatment with no more insulin, no more monitoring, no more glucose and no immunosuppression. So that program is most of the focus of the company. It can be a very -- I think, a very substantial product. We have an IND that will hopefully be occurring next year. We've talked about IND and beginning the study as we move into 2026.

So these 2 patients with data, that was not with an IND because I couldn't find them on...

Yes. So what we did before was with a -- we took a cadaveric islet and gene modified it at a low dose. And so what we were doing here is we're gene modifying a pluripotent stem cell and growing it into many, many doses and make it into islet and then it becomes a replicable scale process. So the goal is with the first drug was -- the first trial was just to see could you see these cells survive and function. From here on out, our goal is to cure the patient, right? It is to see the cells survive function and eliminate their need for insulin going forward.

So -- but just so I'm clear, the 2 patients' worth of data...

Just 1 patient.

That is the same patient. Okay.

That patient was done with...

It was at day 28 and week 12, I think.

Well, we've done weeks, month 6.

Okay.

We have month 12...

But the 1 patient was done through a trial or at a...

It was done at a clinical study run out of Uppsala University in Sweden.

I see. Okay. Okay. Okay. So it's one patient worth of data. So I remember last year, when we had this chat, I remember you left it at either we engraft or we don't engraft. So it's going to be pretty straightforward. And I think as we sit here today, it looks like you were able to hit all your...

If it did engraft, you'd either see if it worked. If it were graft, you see if they evade the immune system or not, and they do.

Correct.

So now we're in the process of making the scalable solution.

Makes sense. So the product and the process that was used in that patient, how meaningfully does it differ or not versus what's being...

How many what?

The product that was used in that 1 patient, how does it differ or not versus what's going into the IND?

How does it differ?

How does it differ or is there a differ at all?

That was a -- a person died, a donor person died, donated their pancreas. The -- so those islets were isolated, they were gene modified. So about 40% of all the cells were fully gene modified. We knock 2 genes out, we knock in. And the rest of them were not. And so it was done at a low dose, and that's it. So now the main -- the real drug, so you start with a pluripotent stem cell from a single donor that will be your product forever. We do the gene modifications. From that point forward, there's no more gene editing. There's a master cell bank that never gets changed, and it's our drug, hopefully, in 2050 for millions of people. And that single cell will be made into trillions of cells.

Got it. I see.

And you will take those stem cells, you will differentiate them into islets and every patient is getting the exact same product, hopefully, right?

Makes sense. Makes sense. And then also remind us, Steve, what's the data currently in the marketplace on stem cell-derived islet transplants that are out there. I think they occur, but the engraftment rate is not the same and it's not as successful.

So there is -- so Vertex has done this very successfully published in England Journal of Medicine. They were 12 out of 12 patients who reached a year are insulin-free. They put it in the portal vein. It worked. patients are immunosuppression, but it works, right? And there's been another group in China they actually took a patient -- they took a person who had a type 1 diabetes and a liver transplant. So they were already on immunosuppression. They then made the patient's own blood cells back into pluripotent stem cell, grew them up into islets and transplant them with immunosuppression. So it was an autologous pancreas. And again, that worked. That patient is doing very well. But again, they had to be on profound immunosuppression to overcome the autoimmune reaction. So what we have done is we've done these gene modifications that prevent the patient's autoimmune system from seeing this and prevents their allogeneic system from seeing it.

Right. Got it. So in the Swedish data set that was generated, you focused a lot on graft survival on MRI and then you also focus a lot on C peptide. Are those the 2 parameters we'll be tracking? Or will MRI not be as...

Well, again, the previous study, you want to see did they survive and function. So you look at MRI, we look at PET MRI, so you could see that they were insulin secreting cells, right? You know you don't have those in your arm. We then looked at -- so when islet's beta cell makes insulin, actually makes pro-insulin. And it's secreted -- when it's secreted, it's cleaved in the C-peptide and insulin. So when you have C-peptide in your blood, it's a 1:1 number with how much insulin you're making, right? So these are people who had diabetes and had no C-peptide, now they have stable C-peptide in this person. And then when this person eats, their C-peptide goes up. Again, you see this stably and happening over 6 months. That's not our goal going forward. Our goal going forward is to give a high enough dose that you want to see those things, you want to see the cell surviving, you want to see C-peptide, but you want to have enough of a dose that they have a normal blood sugar and they are completely off insulin and they get to live the life that I live.

Right. So what is that dose that you think will enable you to get there? And could you remind us what was the dose used in that suite?

So in that study, fully edited cells is probably 60 million, 70 million cells. If you look at the data from just the field, it usually takes around 1 billion cells to get to euglycemia. So that's around the right number.

Okay. So the goal -- so as you're growing it, you need $1 billion per patient.

Pretty much, yes, more or less.

Okay. And where is the production right now in terms of getting ready for IND in terms of the cell bank that you're populating?

So we've had multiple meetings with regulators around the world, multiple meetings with the FDA in alignment of what we need to do going forward. So we have to do 2 things to get to human testing. One is complete tech transfer and GMP manufacturing. Two is complete our GLP tox stuff. They will happen. We've done all the studies that will need to be done already. It need to be done in a GLP setting with the current lock kind of process and off we go.

Got it.

So I mean we could face slip-ups along the way. That happens all the time in these things. But generally, we'll get through them. It's a matter of when, not if. And I think we're far enough along and confident enough of what we have that we think that there's a high probability, no guarantees that we'll able to get this done in 2026 and start the study next year. And data come very quickly once you start the study.

I mean, presumably, you can get proof of concept -- I mean, you kind of already have it, but you'll get it super fast on the actual sort of insulin...

[indiscernible] very straightforward.

So you go...

So the challenge in this will be scaling manufacturing. We need to take a Phase I manufacturing process and modify it and something that is really good enough for early commercial launch, right? Because you have to have a registration study, you won't be able to change it. And again, I think that, that's something we can get done. And then I think the actual clinical trial, if you look at, again, others in the field, they're doing a Phase I, II, III program of 50 patients in aggregate.

So I guess where are the bottlenecks then on the scale up? And where is it that you're spending most time on that? Or what is it that the company is still trying to figure out or working its way through?

I think the hardest part of -- so you're -- what we know like this is like a protein biologic. You're taking a cell and you're getting to spit out a bunch of protein. And all you care about is that protein and the antibody is the same every time, right? What we're trying to do is make the exact same cell, which is a very complicated function each time. And there are 2 challenges, I think, as you're going through this that are the biggest to scale. One is as you go through all of the manufacturing, let's say that you end up with a lot of the risk of genomic instability. So you'll see problematic mutations arise. Let's just say, mutations in genes that encode for DNA repair is the most common thing. And you probably don't want to transplant as an example, 1 billion cells with p53 missing, right? It makes sense you wouldn't want to do that. So genomic stability is part one. The second is product purity. So it's actually not that hard to kind of make these things. But when you're making them, you're going through normal development pathways. You go like you start out with an pluripotent stem cell, it then becomes like endoderm, then it becomes foregut and then it becomes -- if you end up with some stomach or some GI tract, those aren't terminally differentiated cells. They'll just keep growing in the patient. And if this is going to be there for 10, 15, 40 years, you don't want a bunch of stomach growing in some other or wherever we will transplant this in muscle. You want to grow in the muscle. And so genomic stability and product purity are going to be -- are the biggest challenge as you get into making many, many, many, many cells.

So on the genomic stability, I guess, once you have like -- once you're sort of able to boil down to the cell, which you want to then grow off of, stability would just need time to be able to prove that. No?

Well, it took us years to make a cell line. I'm not aware of -- I think it has been a problem for the field broadly, where you gene edit it and maintain genomic stability with a pluripotent stem cell. It's been super hard. I think that we got there a lot because of the cell line we're using, a lot because of the process that we're using those 2 things. And I think it took a little bit of luck. I don't think there's any way around that. I think you need all 3 to make it happen.

Got it. Does FDA want to see a certain amount of time duration, once these...?

There's a lot of testing that goes into this. It's not time, it's many, many, many divisions. So I'm not going to get into what you do, but there's a lot of genomic testing that's done that you're looking for, first of all, the gene edits, are they on target, right? The second is, do you have chromosomal rearrangements. Those happen that just as you make changes, you grow these stem cells, little parts get clipped off or 15 ends up on 13, whatever it is. Then you have to look for the risk of these cancer genes popping up. And you have to look for -- do your cells do what they're supposed to do genomically, do you have any mutations, any gene that matters for the function of the cell you're going after?

Got it.

It's rigorously, rigorously tested. And that's what a big part of what we spent the last few years on is getting that done. And then last year, you're getting alignment with global regulators on what you need to have to release a master cell bank in an area like this.

Got it. How should I think about cell viability once the transplant does happen? How should I think about that?

Think about what?

Cell viability and...

Hopefully, it's for decades and decades.

Okay.

We've seen that happen in the cadaveric islet transplant field. Competitors or others in the field are out several years with stem cell. I don't see the reason why they'd be less. In fact, they might be more.

And remind me, I think you mentioned in mice no MHC 1, no MHC2, right?

What's that?

No MHC 1, no MHC2. So there shouldn't be NK cell attackable.

So if you just knock out MHC class I and class II, NK cells will kill them. So -- and that's been the challenge of the field. And our insight was that overexpressing CD47 in the context of knocking out MHC class I and class II protect cells from NK cells, it protects cells from T cells, it protects cells from B cells, protect cells from macrophages. So you overcome both the innate immune system and the adaptive immune system. And again, we've shown this in all kinds of in vitro assays. We've shown this in mice, humanized mice, monkeys. We've now shown it in humans across several different diseases and cell types. This works, right? And so now what we need to do is just put it in -- put the whole system together and put it into humans safely and then see it scale.

Got it. My last question. Based on some of the data you can generate next year, could it form the basis of a breakthrough designation?

Could what?

Could it form the basis of a breakthrough designation?

I don't think that, that's very challenging for us.

Being able to get to that.

Yes. Yes. Without getting -- yes, I think that, that...

I only say because, look, it may not be challenging given the biology that exists and as long as it proves through on an in vivo basis. I'm asking because just from a market perspective, oftentimes, that's a trigger for a lot of people to start paying attention to that who may have been on the sidelines. So it's a realistic possibility of Phase I data.

True. I mean I don't know when, but I mean I think these data would be -- they're transformative -- they fit all the criteria.

Okay. How many patients worth of manufacturing do you have ready for Phase I? How many patients can you dose in Phase I?

Phase I, our goal -- so if you look at the field, you would think that the Phase I would be 12, 15 patients. That would be the goal.

And you're ready for that by, let's say, March?

I'm not saying when. We said next year, 2026. But my guess, you're not going to dose everybody -- you're going to have some little stagger to begin with, right? And then we'll treat patients.

Got it. Excellent.

So in vivo CAR-T, I'll go through very quickly.

All right.

So in vivo CAR-T, what we do is we do cell-specific delivery of some type of genetic payload. And what we're doing in the case of the CAR-T is we're delivering a DNA plasmid, right? So we made 2 fundamental assumptions at the beginning of this program. If they're both true, I think we have a best-in-class. Number one is that cell specificity matters, you do not want to get off-target cells. So we believe that because it improves your manufacturability, it improves your immunogenicity risk and it improves your safety profile, right? And pretty much everything else I've looked at some of the cells over the liver and other places, right? The second is that you want to put in a payload that will integrate in the DNA. So you could do mRNA. But the math problem here is you might make 100 million CAR T cells, and you need to make even 500 million, but you might make 1 billion, but you're going to try to get rid of hundreds of billions of cells. And so you can't redose yourself out of that problem. And so you need to see logarithmic or exponential growth of your CAR T cell. And so that was basically the 2 assumptions. If we turn out not to be -- those 2 things aren't true, others have made other platforms that will be simpler to make and people kind of prefer mRNA to integrate DNA if they don't need it. So we will have made things very complicated. If they both turn out to be true, I think we have best-in-class data. We can show in nonhuman primates, which is a very good system that we can deliver specifically the T cells that we make a CAR T cell, they expand like you'd expect to see with an autologous. It looks like an autologous cell. You give no lymphodepletion, though. You then see all the B cells go away in the blood. You biopsy lymph nodes, there are no B cells left. When the B cells come back, you see what you saw with George Chet's data in lupus that you now are full of just naive T cells. So you've seen that B cell reset. So that's a very good biomarker of a very deep effective safe B-cell depletion. So I'm very optimistic about this. It needs to come into humans. What's that?

What's your -- you're not a viral vector, are you?

It's a VLP. It's a virus-like particle.

It's a virus-like particle.

So it's -- it has -- the manufacturing is very similar to what you would utilize for like a lentivirus.

Got it.

The targeting and things like that are different.

Got it. Last question, I guess, again, on this topic, just to wrap it up. There's been some feedback that on some in vivo CAR-T approaches, you dose and let's say, you dose 10 patients, but 2 just don't develop any CAR-T at all.

A what?

You dose 10 patients, but 2 out of 10 may not develop any CAR-T at all. And this has happened on some programs, and this is like anecdotal feedback from pharmas that have looked at everybody's programs. Why would -- in your opinion, why would that be happening? Or does that just explain some of the limitations of how tricky this stuff is in practice?

So I don't know that -- we have a very predictable effect in nonhuman primates. What you have with T cells is, first of all, in a patient, particularly someone who's been sick, who's been on many, many chemotherapies or many, many immunosuppressants, they may have different fitness of their T cell. They actually may have actually expression of different restriction enzymes that allow it to happen. They may have different immune systems that have some type of immunogenicity to what's already been put in there. I think all of those things could happen without knowing details of this...

Okay. [indiscernible]

It's certainly -- I don't know of many medicines that work at 100% of people, right? And so we don't -- and I wouldn't know without the details. But those 3 things all seem very viable.

Okay.

And T cell fitness would be a big one.

T cell fitness, T cell fitness is key. Fantastic. Well, thank you so much. This was very helpful.

Thank you.

I'm really looking forward to staying in touch to your Phase I stuff.