Inside Arbor Biotech's In Vivo Gene Therapy Approach

A conversation with Devyn Smith, Ph.D., and Pam Stetkiewicz, Ph.D. — Arbor Biotechnologies

If you want to make an in vivo gene therapy and avoid a biodistribution quagmire, you probably want to encapsulate it in a lipid nanoparticle (LNP) and target liver disease.

That’s the approach Arbor Biotechnologies is taking with its lead candidate, ABO-101. The company aims to treat primary hyperoxaluria type 1 (PH1), a rare genetic disorder that often affects children and is characterized by overproduction and buildup of oxalate, which leads to large, agonizing kidney stones.

Kicking off its Phase 1/2 trial following IND and CTA clearance, Arbor dosed its first patient in June and reported a second patient close behind who is undergoing screening. The company is one of several to follow the promise of LNP-mediated gene therapy for treating liver disease. ABO-101 uses LNPs to deliver a Type V CRISPR Cas12i2 nuclease and a guide RNA to knock out the HAO1 gene.

To talk about the specific challenges in producing an in vivo gene therapy, we asked Chief Executive Officer Devyn Smith and Chief Operating Officer Pam Stetkiewicz about their approach, particularly now as they take ABO-101 into the clinic for the first time. Their answers are edited for brevity and clarity.

Let’s start at the top by looking at Arbor's internal in vivo gene therapy pipeline and your lead candidate, ABO-101, in particular. Can you give us a quick overview?

Smith: The technology we have at Arbor is a wholly owned and developed gene editing technology.

We have the ability, within a cell, to change the DNA. We can turn off a gene; we can turn on a gene. We can also repair a mutation, so you have a normal, functioning gene. Some diseases have too much of a repeat — these repeat regions that expand and cause disease, like in Huntington's — and we can cut that out to restore normal cellular homeostasis. And then finally, there may be times when you want to add in a particular gene to replace something. We can do that.

There's a lot of versatility in what we can do within the cells of a person. The other important piece is actually delivering the gene editing machinery to the appropriate cells in a safe manner. The approach we've taken with PH1 is using a lipid nanoparticle. Fat globules form a little fat droplet, which contains the RNA encoded gene editor made up of the nuclease and a guide.

The gene editor has what's called a PAM (protospacer adjacent motif). It's a three- to five-letter code that tells the nuclease or editor where it could bind on the DNA. Our technology allows you to bind to nearly 100% of the DNA given the number of different editors we have. The other part of a gene editing machine that gives specificity is called the guide, and that tells it where to bind and what to do from an editing perspective. Together, the nuclease and guide make up the gene editor.

Stetkiewicz: Our lead program is ABO-101 to treat primary hyperoxaluria type 1 (PH1) in which the patient has a mutated enzyme in the liver, causing them to produce too much oxalate, which binds to calcium and creates massive kidney stones.

They're irregularly shaped and very large, and they basically shred the kidney and lead to kidney failure. When the disease is progressed, it can also lead to systemic oxalosis; patients get depositions throughout their body, It’s a big problem for these patients.

We want to augment the pathway with a validated target HAO1, which has been prosecuted by an Alnylam ASO to provide clinical benefit to PH1 patients. What we really are excited about is, as Devyn mentioned, our ability to encapsulate our gene editing components in these LNPs. We can give a one-time administration via IV dosing that targets a large number of cells in the liver called hepatocytes and knocks out the HAO-1 gene, leading to a reduction in oxalate.

It should be a one-and-done treatment for these patients, so, unlike the standard of care today, they don't have to worry about going in for treatment every couple of weeks to months. They don't have to live near a center of excellence. Also, many parents express fear of losing their health insurance to cover this very expensive therapeutic for their kids. We're really excited to be able to provide patients with an alternative one-time potentially curative therapy.

With in vivo methods, complications with variability, biodistribution, and durability become magnified. How are you approaching these issues? 

Smith: The reason most, if not all, the genetic companies are in the liver today is because LNPs target the liver. They typically bind to the LDL receptor, which is very prominent in hepatocytes. You can usually get pretty good saturation of the hepatocytes with an LNP-based delivery.

The LNP does travel to a few other tissues at a very, very low percentage, but that's something you have to manage by making sure you've got a clean safety profile. Targeting the liver is pretty straightforward with an LNP as we have seen so far in the many patients who have been dosed and the positive data published across several gene editing clinical trials (not ours) — primarily by Intellia and Verve Therapeutics.

They've seen a small amount of variation. If you're looking at a biomarker readout, one of the advantages of gene editing over say, an siRNA or an ASO approach, is that for those methods, in any given cell, you could be knocking down anywhere from zero to 100% of the RNA. So that cell may produce 10% of the protein or 50%.

In the gene editing example, depending on what you're doing, it's binary for every cell. You're either eliminating all the bad stuff or fixing all the mutation in that cell or none of it.

That's been the experience so far, which is why it's such a powerful approach. Pam can talk about how we think about off-targets and safety, both off-targets within a cell and off-target tissue.

Stetkiewicz: The guide RNA allows us to give a lot of specificity. That guide RNA is about 20 nucleotides long and looks at a very specific region that we're going to anchor. We do a lot of compute biology to make sure that the sequence isn't repeated in other areas of the genome.

We do an exercise called “off-target editing.” We ask: are we hitting the genome anywhere else that we don't intend to?

We could look at that from a computational standpoint to ask, for example: if we had one mismatch out of the 20, what would happen? Looking at computational and some cell and biochemical assays, we compile a list of a couple thousand potential off-targets.

Then, we make primers and probes to these, and we look in a human hepatocyte to see if we edit any the DNA outside the intended target. It is quite exquisite, and we've had very few real off-targets. And the off-targets we have identified were in introns, so we believe that they're non-coding DNA.

We really want to do this in human primary cells because we're concerned about specificity of our drug on the human genome. We use non-human primates to look at the whole body to ensure that we're not editing tissues outside the liver.

One of the interesting things about the target we've chosen, this HAO1 gene, is that it's only really expressed in the liver and at very low levels in some skin cells. So, theoretically, it doesn't have a function outside the liver, providing another level of safety. We also want to make sure we're not hitting germline cells. We did extensive studies around that to make sure we're not editing those.

Presumably there are some frontier activities that you're doing here to make this work. In vivo therapies are not as well characterized as their ex vivo counterparts. Can you talk about the process as you're currently gearing up for Phase 1/2?

Stetkiewicz: CMC is very complex for these genetic therapies. We have two drug substances, which we encapsulate within the LNP. It's a complicated workstream.

We’ve got to have GMP for each of the individual drug substances that we then formulate through mixing into the LNPs and then vial it and store it in the cold storage chain.

On the upside of CMC, there are more vendors in the ecosystem that are capable of doing this, especially after COVID. Some of the time-consuming issues are getting release assays validated with your vendor around all these parameters. Once each individual drug substance and the drug product come out, you basically have three rounds of batch record approvals.

Our lessons learned are twofold: pick your vendors carefully and make sure your assays are in place.

We did a lot of work with our regulatory agencies around our specifications so that we were aligned with them before we submitted our regulatory documents. I think that was very useful. Different countries still have very different views of CMC, especially across the EU, despite attempts to have EU harmonization.

You mentioned three rounds of batch record verification. Are you using product to test during that time? It seems like you would be consuming quite a substantial amount of drug product.

Stetkiewicz: Yes, we do utilize a substantial amount of drug substance and drug product for release testing and stability. We have stability studies because traditional LNPs have a shelf life that might be shorter than AAVs. We have quality take a good portion of our samples for quality check and stability.

We've done an engineering run at scale, the same scale that we do the GMP and in the same suites, and we use the engineering material for stability testing at least. But of course, quality batch release testing needs to come out of the GMP.

About The Experts:

Devyn Smith, Ph.D., is chief executive officer of Arbor Biotechnologies. Before that, he was chief operations officer and head of strategy at Sigilon Therapeutics. Previously, he led business operations and strategy for one of Pfizer’s R&D divisions. Earlier, he worked as principal consultant for the Frankel Group. He received his Ph.D. from Harvard Medical School.

Pam Stetkiewicz, Ph.D., is chief operating officer at Arbor Biotechnologies. Previously, she worked at Flagship Pioneering Ventures as a senior vice president, and before that, she was vice president of program and alliance management where she led the team that filed the first IND for an in vivo CRISPR drug. She has worked at Novartis Institute of Biomedical Research. She received her Ph.D. from Johns Hopkins University.