A Clinical Trial Failure Derails a Promising Technology

In February, Synlogic cut 90 percent of its workforce as it ended a pivotal study of its lead experimental therapy to treat the rare, metabolic condition phenylketonuria. The decision came in response to results of an internal review that indicated the trial was unlikely to meet its primary endpoint. Synlogic, which had been developing a new class of therapies using synthetic biology, is now weighing its strategic options. We spoke to Neal Sondheimer, outgoing head of clinical for Synlogic and adjunct associate professor of pediatrics and molecular genetics for The Hospital for Sick Children in Toronto, about PKU, the treatment options for people with the condition, and the consequences of the surprising results from the Synlogic study.

Daniel Levine: Neal, thanks for joining us.

Neal Sondheimer: Sure, it’s a pleasure to be here.

Daniel Levine: We’re going to talk about the rare metabolic condition, PKU, Synlogic, and the challenges of clinical development of therapies for rare diseases. Let’s start with PKU. For listeners not familiar with the condition, what is it?

Neal Sondheimer: Sure. PKU and its full name is phenylketonuria, is an inherited condition impacting the way that people metabolize the specific component of protein, which is the amino acid phenylalanine. It is one of the most well-known and one of the most early discovered inborn errors of metabolism. So it was a disease that was beginning to be understood back in the 1930s. And rudimentary forms of treatment were being developed by mid-century. It’s also the disorder that led to the adoption of newborn screening. So newborn screening programs were really developed because phenylketonuria was a disorder that was both detectable at the time of birth, and it was treatable with good effect. That is, we could make changes to the way that people ate that could prevent the consequences of PKU. And it was recognized that it was an important public health program to do exactly that—that universal screening of all children who were born was actually a way of both preventing the consequences of disease, but actually saving money in public healthcare systems because the outcomes of people who were treated with PKU were so much better than for folks who weren’t treated.

Daniel Levine: How does the condition manifest itself and progress, and what are the risks over time?

Neal Sondheimer: In its untreated state, which is very, very rarely seen nowadays, the disorder has a slow onset during the first year of life with developmental impairment, and that developmental impairment becomes quite static. Most children who have untreated phenylketonuria will miss many of their milestones later in the first year and have great difficulty with learning to speak or learning to walk. So it causes very profound impairment in its untreated state. Nowadays, in the United States and in Canada and in Europe, and really in most of the world where healthcare systems are sufficiently robust to run newborn screening, you don’t see these impacts anymore. Children are recognized very close to the time of birth before they’ve suffered consequences of PKU. They’re treated for PKU by control of their diet and they don’t develop these symptoms. We view it as a treatable but not curable condition at this time. It is a genetic condition, so it can’t simply be erased and the therapies that we provide for it really get around the metabolic problem of being unable to metabolize phenylalanine. Phenylalanine is a part of normal protein that’s found in food and it’s in excess for everyone. Most of us will consume the phenylalanine that we eat and turn it into other amino acids or into fuel. Some phenylalanine is incorporated into our own body protein, and so phenylalanine is actually required. It’s essential. You can’t give someone no phenylalanine, but the way in which people are treated with PKU is that they have a very restricted intake of phenylalanine, and they do this in general by eating a small amount of regular protein and much more protein where all the phenylalanine has been removed, sometimes known as metabolic formula or incomplete protein. The risks over time are complex. It’s recognized that people with PKU who can’t control their phenylalanine levels over time are at increased risk for problems with what’s called executive function, the ability to make decisions and especially complex decisions. And these can have impacts on their ability to hold jobs, to develop stable relationships. It can impact mood, and better control of phenylalanine levels is associated with better quality of life.

Daniel Levine: You wear multiple hats. You run the largest PKU clinic in Canada, Sick Kids. I think it’s easy for people to not understand the seriousness of a disease like PKU because the condition is not visibly detectable and people may assume it’s no big deal. You just have to avoid certain foods. I remember when one of the drugs that won approval, one of the powerful moments in the regulatory review was when a leading expert made the reviewers taste the medicinal foods that people with PKU had to live on. And I’m wondering how challenging is it for people, particularly young people, to adhere to such a restrictive diet that consists largely of medicinal foods?

Neal Sondheimer: So, it’s certainly true that people with PKU aren’t visibly detectable. If you were to walk into the waiting room for my clinic when I was seeing PKU patients, you would just see normal looking children of various ages. There’d be nothing about them that you would see as different until they started to eat. And eating is the problem, and it sounds like something that’s trivial or not difficult or, “Oh, you just have to adhere to a vegetarian or vegan diet. It can’t be that difficult.” Many people choose to do that, but that’s not the case at all. There is no way for most people with PKU around the use of medical foods and medical foods, as you note in that anecdote, are very unpleasant. Protein is designed well, not designed exactly, but it tastes good. Intact protein tastes good, whether it comes from animal or vegetable sources, it’s tasty to us. We’re engineered to enjoy it. But if you chop protein into its constituent amino acids, which you have to do to get rid of phenylalanine, it doesn’t taste good. Amino acids by themselves when they’re not linked together as protein taste terrible. And so it’s very challenging because food is part of our lives and our rituals and our social interactions in a deeply ingrained way that people don’t even recognize. One of the really challenging things for people with young kids is the birthday cake scenario, which their children either host or are invited to birthday parties and they can’t consume the cake and ice cream that everybody else is eating because it has too much phenylalanine for them. And that may sound like a small thing, but imagine yourself in the situation where that occurs. It’s really not. And then imagine that situation occurs every day for that person at work and later in their school settings when they want to go out and enjoy meals with friends or romantic partners. It’s always there. It’s always a challenge to maintain phenylalanine levels, and it’s the low grade pressure day to day. That’s the biggest obstacle for patients and families. It’s actually not hard to manage in infants. It’s actually quite easy. Infants have a very poor sense of taste and don’t recognize the difference in the taste of metabolic foods versus regular foods. But as people get older, their perception of taste improves. And in addition, parents are extremely good at making sure that their children’s phenylalanine value is low. They want them to develop well. They want them to achieve everything that they can. Your mother and father are a better protector of you than you are yourself, and that’s a well-recognized problem with many adults in their interaction with the medical system. So the challenges of maintaining stable phenylalanine level and protecting yourself from the consequences of disease greatly amplify as kids go through adolescence and really increase in adults, and particularly in adult men, to the extent that most adult men with PKU have very limited control of their disease.

Daniel Levine: How effective are existing therapies to treat the condition?

Neal Sondheimer: So existing therapies are very effective if they’re used. The cornerstone of therapy is dietary control and dietary control is sufficient to control any person with PKU, no matter how severe their PKU is, they can be controlled with diet. Phenylalanine only comes from food. It can’t be generated within the body. If you don’t feed somebody phenylalanine, whether they have PKU or not, their phenylalanine level in their body will drop. So it can be very effectively controlled. And in addition, there are two approved forms of therapy that can provide additional support. So, there’s a set of vitamin- based therapies. The generic term is sapropterin, and sapropterin is just a vitamin like any other vitamin. And sapropterin supports the enzyme that degrades phenylalanine and can help about a quarter of people who have PKU to maintain control. And then there’s another therapy that was developed over the last 15 years called palynziq, which is an injectable medication that is a non-human enzyme that can degrade phenylalanine directly. So those two therapies are both available. Again, both of them have their advantages and disadvantages that make it so that they aren’t universally used. So palynziq is used by about 10 percent of individuals in the United States who have PKU, much less throughout the rest of the world, and sapropterin is used by about a quarter of individuals in the United States and somewhat less throughout the world.

Daniel Levine: You didn’t mention Kuvan, is that still being used?

Neal Sondheimer: Kuvan and sapropterin are the same thing and has been joined by generic forms of sapropterin. Kuvan was the original drug, and Kuvan is just the trade name for sapropterin that’s manufactured by one drug company. And there are now generic forms of sapropterin.

Daniel Levine: It would seem that this would be a disease well suited for an enzyme replacement therapy. But the approach that’s been taken has been to do what’s more accurately called an enzyme substitution using a bacterial enzyme to break it down. Why haven’t we seen enzyme replacement therapies for this condition?

Neal Sondheimer: So, enzyme replacement therapies are challenging to produce. You have to make an enzyme outside the body. You have to apply the enzyme to the patient in such a way that it can get to the tissue where it needs to work and into the specific location within the cells of the body that it needs to function. And it has to do that in a state where it’s active. It’s been very successfully used for disorders called lysosomal storage disorders where a targeting trick is used to get the enzyme into the correct compartment within the body. But that problem was never been solved for the enzyme that’s deficient in PKU, which is called PAH or phenylalanine hydroxylase. People did work on this problem. People were hopeful that ERT or enzyme replacement therapy with PAH would work, but the efforts were ultimately abandoned in the early 2000s. And the principal reason seems to have been, and I didn’t work on this myself, the principal reason seems to have been the inability to make a stable form of PAH. That is you could make a form of PAH that was active, but once it was released into the body, it couldn’t get into the liver where it needed to function in a form where it was sufficiently stable for it to work.

Daniel Levine: As a clinician who works with these patients, how big a need remains for more therapeutic options?

Neal Sondheimer: So, there will be a continuing need for more options until everyone with PKU is able to tolerate a normal or nearly normal diet, or at least a diet that doesn’t include medical protein. Medical protein is so unpleasant for people with PKU that folks will never stop searching for alternatives until they don’t have to use that. And right now the approved therapies just don’t cover all patients who need support. So there is still a huge continued drive. It’s a major focus of the family organizations that are devoted to PKU advocacy to develop more options for this disorder.

Daniel Levine: A lot has happened since we first planned this interview. Synlogic, the company where you’ve been working, viewed interim results for its phase 3 study in PKU. Synlogic had been developing an experimental therapy, the lead in a class of therapies, it calls synthetic biotics. What are these?

Neal Sondheimer: So, a synthetic biotic is an engineered form of a bacteria that would naturally live inside RGI tract. So it is developed from a probiotic bacterium, and again, we’re immensely colonized. We have many, many trillions of bacteria living within us at all times. And a synthetic biotic is one where we’ve taken that organism and we’ve introduced genes that we want into the organism so that they’ll be functional once they are given to the patient. That is once they ingest the synthetic biotic, the synthetic bio goes in and does a specific task.

Daniel Levine: And what was Synlogic’s approach to treating the condition?

Neal Sondheimer: So Synlogic has developed many synthetic biotics. The one that was being studied in the phase 3 was called SYNB1934. 1934 was selected based on an important date in the history of phenylketonuria, not for any other reason. which is an injectable medication that is a non-human enzyme that can degrade phenylalanine directly, not for any other reason. And SYNB1934 was synthetic biotic where the change is made, included giving it a transporter for phenylalanine, so that the bacteria could take up phenylalanine from food that had been ingested and could also take up phenylalanine from the small pool of phenylalanine that recirculates that comes out of people’s bloodstream into the GI tract and is then reabsorbed, so it could pick up phenylalanine. And then it was also given an enzyme to degrade phenylalanine. The enzyme that it was given was actually the same enzyme that’s in palynziq, it’s phenylalanine ammonia lyase. Phenylalanine ammonia lyase converts phenylalanine into something called TCA and TCA is harmless and can be easily excreted or converted by the body. So the goal of SYNB1934 was to allow people with PKU to lighten the restriction of their diet and maintain control of phenylalanine by basically breaking apart that phenylalanine before the patient ever had a chance to absorb it from their food. This was a very unique approach to therapy. There are no other approved synthetic biotics and Synlogic was really leading the way in understanding how synthetic biotics could be used to treat inherited metabolic disease.

Daniel Levine: For those of us who may be crushed to learn that there weren’t 1,933 other compounds in development, that was the year that phenylketonuria was discovered.

Neal Sondheimer: That’s correct.

Daniel Levine: Okay. In February, the company announced it was discontinuing its pivotal phase 3 study of SYNB1934 after reviewing interim data ahead of an independent data monitoring committee, the decision wasn’t related to any safety or tolerability concerns. What happened when the data was reviewed.

Neal Sondheimer: So, during the course of a phase 3 study, and the phase 3 study lasts for many years, we blind ourselves to different forms of the data at different times. And we do this so that we don’t tip our hand about what’s going to happen in the study, so that we aren’t aware of how efficacious the medication is. We do always look at safety data all the way along. So we were never blinded to looking at what was happening to patients. We were always sure that they were safe, but we were approaching, actually in March the first time, where we were going to have to show our data and we were going to have to show our data to an outside collection of experts—experts known as a data monitoring committee. And this is a standard measure taken in phase 3 studies. The purpose of a data monitoring committee is to ensure that for the participants in the trial, the benefits of taking an investigational medication outweigh any risks of taking that medication. And so those committees are basically there for the protection of participants. So we were at the point where we needed to test our systems for generating these outputs. It would be lovely if it was as simple as pressing a button and everything came out in a completely comprehensible form right away. But studies are very, very tricky. We have to, for instance, encode the identity of individuals so that we don’t know who they are. We collect lots and lots of different data about individuals at lots of different time points. And all of those things need to match up when we create outputs. So we were doing what was called a dry run. And in a dry run, we were simply doing a test to see if our outputs were going to show the data correctly. What happened when we looked at the outputs is that we saw that there was an unanticipated problem with the efficacy of the drug with its ability to reduce phenylalanine levels. That was not expected. What we saw is that phenylalanine levels were not being reduced. Not only were they not coming down, there was also not a relationship between the dose of SYNB1934 that we were giving to people and the effect on phenylalanine. So the absence of those two signals was a very serious problem at that time. As the medical monitor of the study, I was concerned that the risks of participating in the trial were not being outweighed by any potential benefit. The company’s leadership met, we evaluated the data together and came to an agreement that the study had to be closed at that time. We wanted to do this even before the DMC meeting occurred, because we were confident that the drug was not being effective in the way that we had anticipated, and there was no purpose to continuing the trial. Unfortunately, as a result of this, it did mean the end of the company. This was our drug that was furthest along in development, and we were really relying on this drug making it to market approval in order to maintain the financial health of the company. And since that was not going to occur, it also meant that we had to wind down the operations of the company.

Daniel Levine: This is an experimental therapy that had promising results in earlier studies. How surprising was it to see the phase 3 unblinded?

Neal Sondheimer: It was very surprising. It was very disappointing. We had all worked very hard to design this trial, writing out how it was going to be performed, picking the sites, designing the systems for producing the drug and analyzing the data. This is all that we had worked on for quite some time. So it was a huge, huge blow and very surprising. We had completed a phase 2 study two years earlier. That study had been published. I had actually worked on that publication and had shown that the drug was effective in PKU patients. So it was a tremendous shock. But unfortunately, phase 3 studies are an experiment and not all experiments work the way they’re intended.

Daniel Levine: Well help people understand that. How does a drug that works in a phase 2 study, and there’s a reason that we do phase 3 studies, but how does a drug that works in a phase 2 study not work in a phase 3 study?

Neal Sondheimer: So, you can think of a phase 3 study as being a real world test for a drug. We are required by the FDA to make the conditions under which the drug is used as similar to the way as we anticipate it being used when it’s a marketed product as we possibly can. And the better way to look at it perhaps, is to think about what is artificial about a phase 2 study. Phase 2 studies are done on small populations of individuals who have the disease that you’re targeting for short durations of time and allow you to perform studies and measurements that wouldn’t be typically used. So, in our phase 2 study, we had 20 participants who used synthetic biotics. There were actually two synthetic biotics in that study. One was an earlier generation of SYNB1934, and we were actually comparing the two of them. So, there were really only half the population who was using SYMB1934. The exposure in that study was only for two weeks in length. We were doing things to control participants’ fee intake, like giving them prescribed menus. So we were telling them exactly what we wanted them to eat during the trial. And we were also using measures that are not typical in medicine. So, for instance, we were giving people labeled forms of phenylalanine as an oral dose and then measuring the excretion of that labeled phenylalanine in urine and levels of it in blood. And that’s something that you wouldn’t do with somebody who had PKU. So, the first thing to know is that phase 3 [trials] are just different and they’re really supposed to test whether the drug is going to work in real practice. So, why did it fail? And we’ve painfully had to think quite a bit about this and try to understand it. And I don’t have any answers for you. I can tell you the two leading ideas that we have about it. The first has to do with the product itself. SYMB1934 is a type of medication that’s not been used before. There’s no other synthetic bacteria that is given to people to promote changes in their health. So, we didn’t have a roadmap for many of the decisions that we had to make, and we had to make changes in our product between phase 2 and phase 3. For example, one of the biggest changes was scale. We had to make hundreds of times as much drug to supply the phase 3 study because we were going to have 150 participants taking it rather than 10, and they were going to be taking it for years instead of two weeks. So, this scale up forced a lot of changes in the way that we grew the bacteria, and it forced changes in the way we provided it to patients. So, here’s a trivial example. In the phase 2 study, we took the raw bacteria, combined it with the excipients, with the other things that went in it, and basically hand filled it at our laboratory in Cambridge. But we couldn’t do that for this study because there were just too many packets of the medication to fill. So we had to use automated filling processes. Now that should be fine. This should be harmless to do. Using machines to fill little envelopes with a powder shouldn’t impact the way the powder works when it’s used by a participant. The idea and the way that we did this was approved for the study by the FDA, but it’s possible that something about the automated processes physically impacted SYNB1934. To be clear, we tested the activity of our drug all the way along to make sure that it was still active. We even did animal studies to make sure that it would work in a biological system. What we didn’t do and couldn’t do until the phase 3 was give it to people with PKU to see if it worked because that was the experiment that we were doing. So we had to wait until this study to do that. And it is possible that there’s something about the way in which SYMB1934 was manufactured in this study that caused it not to be effective in people. The other set of ideas that we have is more of an unknown question of biology. And to understand this, you have to know something about our product in comparison to probiotics. So many people take probiotics or eat yogurt, for instance, to eat helpful bacteria. When you take a probiotic or when you eat yogurt, the bacteria that’s in that product is alive like SYNB1934. But it’s also capable of dividing so it can colonize you. That’s what you’re trying to do with a probiotic is get a good bacteria to outcompete other bacteria that live in your intestine. Our bacteria is alive, but it’s unable to divide. And the reason we did that is twofold. First, living bacteria that live inside our body generally live very low down in the intestine in the region where phenylalanine metabolism and phenylalanine uptake doesn’t occur. So there was no advantage for us to colonize people. But the other issue is that since no medication like this had ever been used, the colonization was something that’s been genetically modified, was something that people could have potentially had concerns about. And we wanted to be able to say that our product, once you stopped taking it would go away. So this may have led us to a problem that we couldn’t have anticipated, which is does the body get better at attacking bacteria when you’re eating it over time? That is, could we have missed in the two week phase 2 study that over time people get progressively better at killing bacteria that they were eating especially in their stomach. And if that occurred, again there would’ve been no way we could have foreseen it except to run the phase 3 study. So this is a question that we would really like to know the answer to. Unfortunately, it’s a very difficult question to answer,

Daniel Levine: Although if it was the body attacking the bacteria, would you have seen an evidence of an immune response?

Neal Sondheimer: So that’s a great question. We’re not talking about exactly an immune mechanism here. We’re actually talking about the secretion of lytic factors that can directly destroy the bacteria. You don’t have the type of immune system that actually attacks food within your intestine or bacteria within your intestine. We actually can’t have an immune system that does that. Your immune system is very capable of dealing with things that try to invade through the intestine. It’s also very capable of dealing with antigens that are in your food that you’re trying to absorb. But the fact that we have a microbiome filled with bacteria that otherwise in any other compartment of our body would be found harmful and our immune system would attack it shows that we don’t have exactly that type of defense system.

Daniel Levine: And what’s happened to patients in the study? Have they been transitioned to alternatives or are they just back to medicinal foods?

Neal Sondheimer: So, no patient who entered the study was asked to change the way that they were treated. We allowed people into the study even if they were using, for example, Kuvan or sapropterin. We also told participants in the study specifically not to change their diet. Participants were really never taken off their therapies. So, at the conclusion of the study for folks who were in it, they were instructed to simply continue the therapies that they were already taking.

Daniel Levine: I know this is an industry where failure is the norm. This is not only the loss of a drug, but the discontinuation of a company that I thought was developing rather compelling technology. As you look around as a clinician, are there newer therapies advancing in people’s pipelines that you think will provide improved outcomes for PKU patients improve their quality of life?

Neal Sondheimer: So, there are all the time. The truth is that a lot of people like to work on PKU for a rare disease. It’s fairly common. It’s also a bit of an advantageous disorder to work on for drug development because phenylalanine is something that’s very easy to measure and has such a good correlation with outcomes. So there are several approaches out there right now that are in later stage trials. One product that’s in very late stage trials right now is a drug that’s actually quite similar to sapropterin called sepiapterin. And sepiapterin works by the same mechanism, but the folks who are developing sepiapterin feel and have shown reasonable data that it’s more effective than sapropterin is. The other approach that is moving towards later stage trials is a drug that interferes with the kidney’s ability to take phenylalanine back up. This drug, which is made by a company called Jnana, is an inhibitor of a transporter in the kidney that you use to basically keep you from peeing out your amino acids. Our bodies are very good at holding amino acids in because of course they’re food, so we don’t want to get rid of them. And so, this drug is an attempt to block that process and allow people to basically pee out excess phenylalanine. That drug is now moving to later stage clinical trials. People have worked for many years on gene therapy programs. Gene therapy is, of course, something that as a geneticist we would like to develop for lots of disorders. We’re at a stage in the history of science, biology, and medicine where gene therapies are just starting to become available for disorders finally as clinical products. Unfortunately, many of the gene therapy programs that are out there for PKU have been shut down. There was a company called Homology that was working on PKU gene therapy that also closed. But many of the programs have been ended, and PKU may not be a great target for gene therapy. And part of the reason is that if you’re going to do gene therapy for something, it better be something where you get a massive and tangible benefit that you couldn’t get from any other therapy. So, there is, for example, an approved medication for a form of blindness called Luxturna that allows people who haven’t been able to see with a certain genetic form of blindness to be able to see. And that’s an example of something where people are going to be willing to use gene therapy. I’m not sure that people with PKU or parents of children with PKU will be willing to try gene therapy.

Daniel Levine: And is that a function of cost or risk?

Neal Sondheimer: It’s a function of risk. Costs are a complicated matter. Costs are really what the payers of the world, the insurance programs and the government programs for health are willing to pay. It’s not always a question of what individuals are willing to pay. Very few people pay for any of the drugs that I’ve discussed in this call. Even Kuvan, which has been on the market for many, many years, is far too expensive for people to afford with their earnings, they have to pay for it using insurance. Or in Canada, they rely on the public healthcare system to pay for it. It’s really risk. When I talk to people with PKU and parents of children with PKU, I find that if I talked to someone who is the parent of a newborn infant about whether they would be willing to do gene therapy, they say, of course they would. Of course they’d be willing to do it. They’d do anything to get rid of the disorder. But if I talk to the parent of a five-year-old, they’ll almost always say no. And the reason is that PKU is scary, but after five years, it’s something that they know. It’s something that they understand and it’s something that they know how to treat. Whereas gene therapy is something that’s unknown and is potentially frightening to them. And they know the history of gene therapy and the history of gene therapy has been beautiful at times, and it’s been very ugly at times. And so they think about that and they’re not willing to be the first over the line. And I think that’s a problem for PKU gene therapy.

Daniel Levine: Neal Sondheimer, outgoing vice president and head of clinical for Synlogic and adjunct associate professor of pediatrics and molecular genetics for the Hospital for Sick Children in Toronto. Neil, thanks so much for your time today.

Neal Sondheimer: It’s been a pleasure.

This transcript has been lightly edited for clarity and readability.

The RARECast podcast is made possible through support from the Global Genes’ Corporate Alliance. The members of the Corporate Alliance support Global Genes’ mission and programs, work to meet the vital needs of people with rare diseases, and address inequities they face. To learn more about the Corporate Alliance or how your organization can become a member, click here.