In this podcast, Jonathan Meyer, MD, clinical professor of psychiatry at the University of California in San Diego, discusses the role of D2 antagonism in schizophrenia treatment, lessons learned about schizophrenia treatment from clozapine, and emerging treatments for schizophrenia and their therapeutic targets.
Dr. Meyer spoke about these topics and more during his session at the Psych Congress 2020 "21st Century Psychopharmacology: Bridging Symptoms, Neurobiology and Treatment" virtual preconference.
Read the transcript:
Hi, this is Dr. Jonathan Meyer. I'm a clinical professor of psychiatry at the University of California in San Diego.
D2 antagonism really has been the cornerstone of schizophrenia treatment since Arvid Carlsson unraveled the connection between what chlorpromazine and haloperidol do and the development of antipsychotic‑induced Parkinsonism. Once we understood that those drugs worked by dopamine blockade, it became the springboard for developing a series of compounds which all really had the same mechanism of action, which was dopamine D2 blockade.
Now this proved very effective, but the downside is, as people say, the collateral damage, meaning that while we want the D2 blockade to happen in the associative striatum, or what some people call the mesolimbic dopamine pathway, unfortunately, the drug goes everywhere, including to the dorsal striatum or the nigrostriatal pathway. The development of Parkinsonism, akathisia, dystonia, unfortunately, was part and parcel of the early D2 antagonists.
A typical antipsychotic, or the second‑generation agents, had a built‑in mechanism in the form of serotonin 2A antagonism to help mitigate some of these neurological adverse effects, but they weren't absent.
Nonetheless, D2 antagonism has been very helpful in mitigating the positive symptoms of psychosis and schizophrenia. Even the partial agonist antipsychotics, such as aripiprazole, brexpiprazole, and cariprazine all work at D2.
But we would like to have things possibly which don't do anything direct with D2, mostly, so we don't have to spend time trying to worry or manage the adverse effects such as akathisia, Parkinsonism, and dystonia which result from nigrostriatal dopamine D2 blockade.
Clozapine was really the first exception to the rule that high levels of D2 antagonism are necessary to be an effective antipsychotic.
If you are a D2 antagonist, we know from imaging studies that somewhere between 65 and 80 percent D2 receptor occupancy in the striatum is the sweet spot and associated with reduction in positive symptoms of psychosis with relatively lower risk for neurological adverse effects.
When clozapine came out, it became very clear, and most people never even got to 50 percent D2 receptor occupancy. People spent a lot of time and are still spending a lot of time trying to replicate clozapine because it is the only antipsychotic that works for treatment‑resistant schizophrenia.
If you have people who are treatment‑resistant, offering them anything else but clozapine is really considered below the standard of care and substandard medical practice. The problem is that clozapine is clearly a very complex molecule.
Initially, people thought that the serotonin 2A antagonism was the key to clozapine's unique properties. We now understand that serotonin 2A antagonism by itself can be an antipsychotic mechanism.
There's a drug out there, pimavanserin, which is approved for Parkinson's disease psychosis, which only works at serotonin 2A, and thus doesn't have any risk for the neurological adverse effects related to D2 blockade. But that mechanism may not be sufficient for people with schizophrenia, although potent serotonin 2A antagonism and saturation of that receptor can certainly make a weaker D2 antagonist more effective.
People have done studies where they added pimavanserin to low doses of risperidone. That made it a bit more effective, but when you added pimavanserin to haloperidol, which already gives you a lot of D2 blockade, it did not make it more effective.
People have also looked at mechanisms related to its metabolite, which we'll talk about in a second. Other interesting properties of clozapine may be having to do with effects at the glycine site on the glutamate NMDA receptor.
One thing that clozapine really opened our eyes to was the possibilities that there are other mechanisms which contribute to its unique efficacy in treatment‑resistant schizophrenia.
Which of these mechanisms is crucial for this property is unclear, but we understand now that low levels of D2 occupancy combined with high levels of serotonin 2A antagonism is not necessarily the key to being an effective drug for treatment‑resistant schizophrenia.
Because clozapine really did elaborate a problem, how do we figure out what it is doing, and what is special about it?
People have been looking very hard for other mechanisms which are not dependent on D2 antagonism. One thing which came out from studies of the metabolite norclozapine is the fact that it appears to be a cholinergic agonist.
We now understand through both animal models, as well as some newer medications, that animals which have knockouts of their muscular M4 receptor, meaning a cholinergic receptor, M4 subtype, have a phenotype which resembles psychosis. And so we understand now that perhaps one of clozapine's unique properties is actually conferred by its metabolite, norclozapine, which is an agonist at M4 receptors.
Using that insight, people have actually studied M4 agonist drugs, one of which is called xanomeline. This is spelled with an X. It has been in study almost for 15 years. There's a paper published in 2008 in The American Journal of Psychiatry, which showed that it worked for schizophrenia, but the downside is that it caused a lot of GI side effects because it is a cholinergic agonist.
Now, a new company has taken this up. They're combining xanomeline with a peripherally acting anticholinergic to mitigate its GI adverse effects. Again, it shows a lot of interest.
Other targets which are emerging have come from new forms of animal discovery methods. Part of the issue with the second-generation drugs is we have a lot of “me‑too” compounds. People are replicating what has been done before. This does not necessarily allow you to discover new therapeutic targets.
People have gone back, in many ways, to the older methods. They have actually used machine learning to train computers to look at animals being exposed to drugs with known properties to learn what is the phenotype, based upon a range of behaviors. They have 2,000 outcomes in terms of behaviors that could distinguish, for example, an antipsychotic from an anxiolytic, from an antidepressant.
You put the animal in this thing called a SmartCube, which is a small, clear cage, if you want to think of it that way, which has food and water. They give an animal a series of tasks to perform. Then, they train the computer based upon a number of cameras which are observing this animal. What does it look like when an animal gets an antipsychotic or an antidepressant?
The machine learning aspect of the algorithm eventually learns to distinguish these agents very well. Then, they can give it test compounds to test its accuracy. Once they are happy with the accuracy, they can now give it novel unknown compounds.
The advantage of this approach is that you don't have to have an idea about what is the way to treat schizophrenia, that it has to work at D2, or it has to do something at serotonin 2A, or glutamate, or whatever.
All you can say is we have a compound. We think it has psychotropic properties. We're going to give it to a series of mice. Put them in this SmartCube. The machine will tell us “what does this thing look like?”.
From that, we've had some novel therapeutic targets that have emerged, some of which show a lot of promise. One of them is TAAR1, which is trace amine‑associated receptor type 1.
This is an intracellular receptor. It's been hard to study. The endogenous ligands are things like octopamine and tyramine, which you may have heard of a little bit. We still don't fully understand everything about this system. It has both presynaptic and postsynaptic locations.
Through this SmartCube system, a TAAR1 compound was found to have antipsychotic properties. This drug, which has a code name SEP‑856, is now in phase 3 studies, having shown good results in phase 2.
The interesting aspect is that this was a receptor which was not well‑known. It was only really characterized in the last 15 years but now shows potential as a therapeutic target for schizophrenia.
Another company, of course, once they got the idea about TAAR1, is also studying this, as well.
Another interesting mechanism which came out of the SmartCube system is phosphodiesterase 10A or PDE10A. This shows promise primarily for the negative symptoms of schizophrenia. But again, it was something that was developed via machine learning using this SmartCube system.
So we now have a range of possible targets which are not directly acting at D2. The beauty of these is, for one thing, they may be complimentary with D2 antagonists. That remains to be seen.
Certainly, when you have agents that don't work directly at D2 to any great degree, you avoid a lot of the neurological adverse effects, which have really been the bane of our existence when using the traditional compounds, the first‑generation agents, the earlier second‑generation agents, and even the partial agonists, as well.
So, this is an exciting time to be in psychiatry. We're learning a lot about ways to treat schizophrenia. D2 still may be quite important, but we'll see where all this takes us. It may not always be necessary, or the amount which you block D2 receptors may not be necessarily very high in order to get an effective compound which works for schizophrenia.
So stay tuned for more of these clinical results, and keep your eye on some of these new, novel targets for schizophrenia treatment.