Encephalitis Can Change How the World Thinks About Complex Neuropsychiatric Disorders

Help Pharmacy Times welcome Jay Lichter, PhD, the president and chief executive officer at Arialys Therapeutics. Lichter discusses anti-NMDA receptor encephalitis (ANRE), a rare neurologic condition that is often mistaken for schizophrenia and still not widely understood. Lichter discusses how ANRE is actually an immune-modulated condition, promising preclinical data on a new therapeutic agent called ART5803, and urges pharmacists to relinquish the dogma that diseases of the central nervous system (CNS) will never be immune related.

PT Staff: Thank you for being here. Now, I wanted to start by asking about patients with anti-NMDA receptor encephalitis, or ANRE for short. For those who have this (or another related condition) what causes the immune system to produce the destructive autoantibodies that go against, what should be, a healthful system?

Jay Lichter, PhD: Yes. You know, it's not entirely known. Probably [for] about a third of women who get encephalitis, it's because they have a teratoma, an ovarian teratoma. So, we get a tumor coming off of the ovaries and then that tumor produces NMDA on the surface. And then the B cells learn and make autoantibodies from that. Yes, so definitely about a third of the women [get this condition from the ovaries]—so none of the men obviously will get an ovarian teratoma—and two-thirds of the women don't have an ovarian.

So the ovaries just produce a target. So, you can make those antibodies [which] get into the brain and bind the NMDA receptor. So when they bind this receptor in the brain, they crosslink 2 that are close together and then those 2 internalize. So, what you get is essentially a reduction in NMDA activity; this is like, for example, any antagonists of NMDA receptor: ketamine, or phencyclidine (PCP). Those are very selective known drugs that bind and inhibit that receptor.

And we know what happens when patients take those drugs. At low doses, they get a little bit of euphoria, I mean, at sort of medium-ish doses, they start to hallucinate, and at high doses they'll have neurological symptoms. And then when they overdose, they typically die of respiratory failure. So you can see that spectrum of symptoms in patients with encephalitis, [whereby] in the early stages of the disease they start with a psychotic break and then that ultimately progresses into neurological symptoms. And usually, they land in the intensive care unit (ICU). So it's not well known. What we do know is that the binding site that we're attacking, both with a pathogenic antibody and our drug, is the same epitope (same binding site). That has identity to a gene in a parasite called toxoplasmosis gondii, which is toxoplasmosis. So we don't have any proof other than observation that the sequences are the same. It's possible that some patients get an infection and can make antibodies that way. But for 75% of patients, it's unknown.

PT Staff: Can you explain a little more about what this disease is and why it's often misdiagnosed? Why is it sometimes masked as something else?

Jay Lichter, PhD: There has been a lot of variation in the different types of clinical expression of the disease. It's only been characterized molecularly within the last decade. So it's a relatively newly characterized disease. If you're at a major medical center, you'll get properly diagnosed pretty quickly.

The definitive diagnosis is taking a measure of a sample of the cerebral spinal fluid (CSF) and sending that out for a test to see if there's antibodies in the CSF. This method is a definitive test normally, and for the experienced neurologists, they can quickly have a guess that it's ANRE by relatively rapid things. You can do magnetic resonance imaging (MRI), sometimes it can help just to find an immune function or immune sort of abnormality. The rapid onset of the symptoms and the sort of breadth of the types of symptoms [can] usually give a hint. So typically what you're going to see (and some of the hardest patients to diagnose) are the ones that either have no neurological symptoms or very light neurological symptoms that are maybe hard to see. And so they may get misdiagnosed as schizophrenia, when in fact, they have encephalitis, because you couldn't detect the sort of subtle neurological or movement changes that these patients have.

PT Staff: And how does that manifest differently than schizophrenia? Does it impact a different part of the brain?

Jay Lichter, PhD: So schizophrenia is a lot of diseases. And we think that with maybe 5% of schizophrenics, that the root of their disease are autoantibodies against the NMDA receptor—its somewhere around 5%, and the rest of people have schizophrenia because of a different cause.

Now, I kind of think of this as the beginning of, let's say, where cancer was around 100 years ago. It was where you had cancer and didn't matter where it was or what its molecular characteristics were, you didn’t know if it was one [type of] cancer [versus another]. Now, of course, with breast cancer there are a dozen different types of breast cancer, a dozen different types of lung cancer, and I think probably over the next 10 or 15 years, you're going to see more and more people finding molecular tools to differentiate how these complex psychological and neurological diseases are broken out. And this will be the first piece we believe, where we can find a subset of patients where there's a molecular trigger that we can measure in a laboratory. And that's really what these diseases have been dying for, is something that's a really a definitive test and you don't have to have the patient meet with a neurologist or psychiatrist for several sessions to come up with the diagnosis, you can have an answer. So, I think it's this, you can run the test and in 10 days, [and] you can find the answer. So, I think it's the beginning of a change, a change in the way people think about complex neuropsychiatric disorders, with our doorway in [being] encephalitis.

PT Staff: That's incredible to hear, because when talking about issues of psychosis, it's such an umbrella. It's so broad [and] you're right, there's not so many biomarkers in that area.

Jay Lichter, PhD: There's virtually none. And… it's hard because it's the brain [and] it's more cerebral, so it's harder to get samples from it. Part of it is, if you think about these diseases and how the therapeutic approach of these diseases started, it started with making modulators that impact how the different pharmacological receptors work in the brain. So there's been a big push to modulate these receptors by small molecules with the hope that you can have a lot of people having a huge amount of benefit. Unfortunately, you get some patients with some benefit and a lot of patients with no benefit. And so, it just hasn't… it's just not easy to sample, right? No one's going to give you a brain biopsy unless you're dead. So it's not like, “Oh, I'll take a liver biopsy and figure out what kind of tumor it is” or these things that you can do with other organs. You just can't really do with the brain.

PT Staff: Well, it’s incredible that this is thing that is opening the floodgates, if you will.

Jay Lichter, PhD: At least identifying the floodgates.

PT Staff: [Laughing and squinting into distance]: It’s like, “Okay, we see [the floodgates], We see them! What do we do next?” OK, so can you please describe the therapeutic agent ART5803 for ANRE and its pharmacodynamics?

Jay Lichter, PhD: So it's a single arm antibody. The pathogenic antibodies are 2-armed, and 1 arm binds to 1 NMDA receptor and the other one binds the other NMDA receptor— they're next door [to each other], and those 2 (now crossing) can go into ours as a single. So ART5803 will bind to a receptor and then it'll block the pathogenic autoantibody from binding to it.

It's really a very simple mechanism. You take this 100 Kilodalton protein (which is our drug) and it binds to the same place on the NMDA receptor. Therefore, it prevents the pathogenic antibodies. And so we see that routinely—we have a really interesting marmoset model [and] we've seen a lot of in vitro settings that we have data for—the pharmacodynamics are really very clean. Our molecule is much more potent than the naturally occurring autoantibodies, so a much lower dose can displace the pathogenic autoantibodies. Then having to do, for example, on equal molar dose, so we'll learn the precise doses once we start looking in healthy volunteers and ultimately when we get to patients with disease. Our expectation is that we'll need a relatively large dose because only a small fraction gets into the brain. Once it's in the brain, it'll be very, very active because it's very potent and selective.

PT Staff: So from my understanding, when you put the dose in the brain, it stays there. You don't have to keep upping the ante on how much you add?

Jay Lichter, PhD: That’s a good question. We can't entirely answer that. What I can say is that when you inject it peripherally, and so you do an intravenous (IV) injection, it'll go into the CSF, and this CSF turns over every few days. And so where does that drain? That drains back into serum. So it goes in the CSF and goes back to serum and you have this loop, where over time you get more and more drug accumulating.

Your antibodies typically have a half-life of 2 to 3 weeks in humans, so you have this accumulation. Now CSF is not the site of action for the drug. The site of action is in brain tissue, where the NMDA receptors are in the cortex and other regions. And so what we don't know is how long the drug stays in the brain tissue. We don't know that. And it's very difficult for us to assess. We have some marmoset tissue that will probably be looking at, to determine how long the drug stays in brain tissue— not just in the fluid in the brain. And so we'll start to have those answers at least in a small primate, like a marmoset. But human data is virtually impossible to get. We'll probably get it though, we’ll do our toxicology study in cynomolgus monkeys and rats, and so we'll probably get the data in those monkeys as well as in rats. And to the extent that predicts human, we don't know. But we'll use that data as we can.

PT Staff: Can you describe findings from the preclinical proof-of-concept data?

Jay Lichter, PhD: So we bought this molecule from a Japanese company called Astellas, and what they did, about 4 years ago, was clone a human pathogenic antibody. So they have the sequence we know that made the antibody, and they use an intracerebroventricular injection (ICV). So you can basically inject it straight into the brain. So you put it into the brain of these monkeys and over 2 weeks, they get very sick. They're not feeding themselves and grooming themselves, and they have anxiety. They're unsure, they can't make little decisions, such as “Do I eat food, or do I drink water?” They can't decide. They're very confused and anxious. And then what Astellas did is co-administer our drug into its brain through a direct injection.

At the end of 2 weeks, all those monkeys had recovered, basically back to baseline. They completely recovered versus a placebo group that they hadn't recovered. So they did that they did internally with their monkeys at Astellas in Japan. And then we just recently repeated this lead in July of this year, where we do the same sort of ICV direct injection into the brain of the pathogenic antibody. The monkeys get sick, but then we inject [our drug] peripherally (either IV or IP). So here, we can simulate what it's going to look like in human. And when we did that, again, we cured the monkeys. This time we saw monkeys cured at 1 week and 2 weeks post-dose.

So it's extremely rapid onset and it's it’s a very significant response. And you may not know but these behavioral models are very, very difficult. And we make this as hard as possible. We went to a different monkey colony, a contact research organization (CRO) not internal group. We had a different administration, a different batch of drugs, a different protocol, different animal handlers, and we still got a very robust result. And for CNS— except for something like Valium— you have to take drugs for months before they have an impact.

This drug is impacting within a few days, and that is consistent with the mechanism. So, we just stopped the internalization of NMDA receptors, new NMDA receptors come to the surface, and they're not inhibited. So, after a few days, when these proteins are turning over, you start to get back to normal synapse in terms of the density of NMDA receptor.

PT Staff: What is the possible scope of this treatment, beyond just treatment for ANRE? Could it be used elsewhere, say for schizophrenia for other CNS diseases?

Jay Lichter, PhD: That's certainly the hope. Our focus is encephalitis, 2000 to 3000 patients a year in the United States, at least right now, are being diagnosed, maybe there are more. You know, 25% of these patients will have a second episode sometime in their life. It’s really horrible. With this disease, you'll get it maybe within 2 or 3 months and then you're in ICU. You can be in ICU for weeks to months, and then for most patients, it's 18 to 24 months for recovery (and not more than half of them get back to baseline). Half of them end up with a persistent neurological psychological disorder that they will never get over. So it's a really horrible disease.

We're expecting that we'll be able to address that population as a whole, as well provide some maintenance therapy so that if they get it again, it hopefully won't be as serious and it won't be as rapid onset. Now, we do believe that in and around 5% of schizophrenics have this disease, as I mentioned, because of the NMDA receptor autoantibodies. We are planning on doing a trial there as well, along with get data for other [diseases] (i.e., certain dementias, major depression, bipolar epilepsy, and even anxiety) and NMDA autoantibodies.

Now some of that is a little bit speculative and early, and we will continue to generate nonclinical data to address that. But when I think big about this product opportunity, it could be 5% of many major psychiatric disorders. And there are some folks who believe that even if you don't have NMDA auto antibodies, this drug may help stabilize that receptor in some way, so it functions better.

PT Staff: Is there something that you would want pharmacists or providers to know about diseases of the central nervous system? And how they should approach patients who have it, whether it is schizophrenia or other?

Jay Lichter, PhD: So it's a great question. There’s been dogma in the industry (in the community) for decades— that the CNS is immune-privileged, which means there's no immune cells there and there's no immune activity. Caregivers, whether you're a pharmacist or physician or nurse practitioner, or whatever you're doing to see patients, you… have a dogma that CNS is immune-privileged, and it's just not true. There's dozens and dozens and dozens of papers showing that there are B cells, and T cells and antibodies in the CNS; sometimes tney are doing nothing, sometimes doing good things, sometimes doing bad things. But I think just getting people who see patients to consider that there could be an immune component to the disease is really critical.

There was a paper, an article, in The Washington Post a couple months ago about the neuroimmune, or the autoimmune, in psychosis. And they talked about one woman who was institutionalized catatonic for decades (2 decades). And they couldn't do anything, there was nothing and the physician said, “Well, let's try something. Let's try rituximab (Rituxin, Genentech; Biogen), intravenous immunoglobulin (IVIG), and steroids and see if this patient does better. After 6 cycles, the patient was clearly not catatonic. They weren't back to baseline, but they were able to interact with people, have relationships, and do things that is normal in society. So that was really a breakthrough.

And I knew that case study—I'd seen that as a case report, and then it came out of The Washington Post. So I think that, again, it's mindset when you are treating patients. You see a patient with neurological symptoms or psychiatric symptoms, and you immediately go to the typical toolbox (which is certainly the right thing to do), but if you run into trouble and that toolbox isn't working, think about if there could be an immune modulatory way that could do this. And it's easy. Steroids, IVIG and rituximab are all available. And if nothing else is working, let's try it, and you might get some patients to respond. So I just think that the dogma needs to be fought against because it's no longer true. It was never true, but we didn't have the data. Now we have the data to know that it's not true.