Wednesday, April 18, 2012

Refried serotonin lunch

That title sounds like the prequel to a William Burroughs novel. I wish I'd come up with it myself, but I'm actually plagiarizing almost word for word from Dr. Steven Hyman, Director of the Stanley Center for Psychiatric Research at the Broad Institute of MIT and Harvard. Last week, Emory awarded Dr. Hyman this year's Neuroscience and Ethics Award. Dr. Hyman spoke on "Addiction as a Window on Volitional Control", which shouldn't be surprising, given his molecular and genetic studies of the dopaminergic system.

Earlier in the day, Dr. Hyman stopped by the Emory Center for Ethics to have lunch with faculty and students from several schools and programs. He held forth on the state of translational neuropsychiatric research for an hour and a half, while the rest of us prevented him from eating by constantly barraging him with questions. I have a feeling I'm not the only one who found what he had to say both thought-provoking and provocative. He liberally spiced up his comments with colorful phrases like the aforementioned "refried serotonin meal" (which I will put in context at the end of this post).


Before I talk about what Hyman talked about, though, I have to put the man himself in context. Here's an interview with him and a talk he gave at a benefit, if you want to check them out. And here's the Cliff's Notes version of his career: Hyman started college as a philosopher, and then realized that neuroscientists were the ones working on the questions he wanted to answer, and so he snuck into the field through the back door of med school. Eventually he established his own lab. He then went on to head the National Institutes of Mental Health (NIMH) for five years, and followed that with a stint as provost at Harvard before becoming director of the Stanley Center. Let's relate all that to what he had to say at lunch. I guess his background explains, in part, his ability to turn a phrase. And his C.V. explains why those phrases tend to cast psychiatry and big pharma in a pretty harsh light.


Hyman leapt right into why we've had such a hard time curing mental health disorders. According to him, there's two main reasons. One is that research, for a long time, focused on drugs--the "paradigm dominate[d] the field". He conceded that research on drugs initially made sense: the drugs helped people with mental health problems when nothing had before. As he pointed out, Thorazine was just one of many antihistamines, which had great sedative properties, until doctors noticed it ameliorated some psychotic symptoms. The other reason we’ve had such a hard time curing mental health disorders, according to Dr.Hyman, is that psychoanalytics still casts a long shadow over psychiatry. He pointed to the DSM-III, released in 1980 (if you don't know, that's Diagnostic and Statistical Manual of Mental Disorders), which he said was based on "1970s science". With the DSM-III, he said, you had a case of extremely "bright people becom[ing] guardians of the holy writ". Even now, as the American Psychological Association puts together the DSM-V, "people will kill each other over the placement of a semi-colon".


The end result of the DSM is that disorders are defined in terms of symptoms. The problem with this is that human beings, especially those struggling with mental disorders, are not likely to be easily divided into groups based on their symptoms. No one who is depressed thinks, "Oh, it's been two weeks and I've only had three episodes of suicidal thoughts--I'd better get some stinkin' thinkin' goin’ to meet my quota." Yet, this is exactly the way that scientific studies were being framed when Hyman became head of the NIMH, and, because of the DSM, they’re still being framed that way. Hyman tried to change that. “I felt I was responsible for two billion dollars of taxpayer money--and if we are [funding] biological characterization of something that is not a natural kind, the study [was] stillborn.” In other words, we’ll never come up with a therapy for a mental disorder that doesn’t really exist.

In contrast, Hyman feels that “the genetics of neuropsychiatric disease is going to work”. The cost of sequencing genomes has come down something like 10-million fold. “It makes Moore’s law look pretty lazy.” He pointed out that we know that mental health issues run in families, even if the symptoms don’t always look the same. There are shared genes that put people at risk for mental health disorders in general: bi-polar patients show up in families with high incidence of schizophrenia. By the same token, Hyman said, all this genetic data has shown that phenotyping—i.e., looking at symptoms—is “worthless”. For example, there’s no correlations between genes and symptoms of schizophrenia that show that we can divide the disease into sub-classes. And while we do see a clear correlation between certain genes and the presence or absence of mental health problems in general in families, we still have to deal with 100s or 1000s of genes of small effect. “The hard problem is the genes to biology problem” in Hyman’s view.


There were a lot of other topics Dr.Hyman covered during the hour and a half that he talked with us. A couple of people asked questions about epigenetics, and I thought his answers came off as a bit glib. In case you don’t already know, the field of epigenetics studies changes in gene expression that are inherited but don’t require changes in DNA sequence. For example, DNA methylation patterns can be transferred from one generation of cells to the next, and these patterns affect transcription of genes. Epigenetics provides a way for the environment to affect the genome. In other words, this field could explain why one man suffers from a mental disorder when his identical twin doesn’t. Shouldn’t we be studying that, and worrying a little less about those 1000s of “genes of small effects”?

I also was a little taken aback to hear someone in Hyman’s position say that big pharma had woken up from its “refried serotonin meal”. I mean, I thought SSRIs were not so efficacious when I was watching my depressed friends in high school crush them up and smoke them, but I wasn’t head of the NIMH at the time. (And I guess I won’t be after writing this paragraph.) With the benefit of age and wisdom and stuff, I realize that they represent about the best we can do right now. So I have to agree with Dr. Hyman when he says “no one’s improved on Lithium,” in terms of efficacy, “and that was 1949.” To improve our therapies, we’ll have to overcome a lot of the baggage left over from that time in psychology when Freud stopped studying crawfish and started prescribing cocaine. (That’s my cynicism you hear in that sentence, not his.) Hyman’s take-home message is that we have an opportunity right now to make big leaps forward in treating psychiatric disorders. I think we can all agree on that.

--David Nicholson Neuroscience Graduate Student, Sober Lab



Want to cite this post?
Nicholson, D. (2012). Refried serotonin lunch. The Neuroethics Blog. Retrieved on , from http://www.theneuroethicsblog.com/2012/04/refried-serotonin-lunch.html

2 comments:

Laura E. Mariani said...

Nicely summarized, David, although I was disappointed that you didn't say more about the lunch. That baba ghanoush was good.

I think a lot of scientists share Hyman's qualms with the DSM's attempt to define psychiatric diseases through sometimes arbitrary checklists, and I agree with him that we're in the midst of learning a ton of useful stuff from genetic studies. (I'd better agree with that point or they'd have to kick me out of the Department of Human Genetics.) But, I'm hesitant to be totally dismissive of the DSM's symptom-based categorizations, even if they don't agree with patient genotyping, if it turns out that grouping patients by symptoms is still helpful in terms of treatment outcomes. I guess upon further reflection, I don't find the point about phenotyping being "worthless" because it doesn't align with genotyping to be so damning. It can be true that people with the same symptoms have very different etiologies for their disease, and thus require different treatments to correct different underlying neurobiological problems. But I'm also remembering what we learned about predictive validity way back in first-year Systems Neuroscience -- like how the rodent forced swim test is wacky and not really measuring "depression," but it's good at showing which antidepressant drugs are going to be effective, so we roll with it as a model. I'm not saying the categories in the DSM are currently so awesome at predicting which patients will respond best to which therapies, but it may turn out that assessing how patients feel and behave is more helpful than looking at their DNA sequence or at some biomarker when we're dealing with a disease caused by many genes of small effect. This is something that will have to be determined empirically, though, rather than by a committee of semicolon fanatics.

Of course, the really interesting question is WHY genotype and phenotype don't always match up, which is where your point about epigenetics (and also stuff like brain plasticity in the face of different environments/experiences, and any number of other things that fall under "nurture" and not "nature") comes in.

David Nicholson said...

Thanks for your comments, Laura. I feel like I could write a couple more posts about different things that were said at the lunch, and one of them was Hyman's statement that the forced swim test was a "sham" (or remarkably strong words to that effect). Like you, I remember being taught that the test has "predictive validity" because the same antidepressant drugs that get rats to swim longer also help people that are diagnosed with depression (even if the drugs act right away in rats but take months to help people feel better...I hope any neuroscientists reading will forgive my informal description of the forced swim test, but I just gave myself an "acute administration" of caffeine which is preventing me from thinking of multisyllabic words).

And I totally agree that we shouldn't just ignore patients' self-reported systems just because results from sequencing studies don't line up with diagnoses in the DSM. As far as the giant gap between genotype and phenotype goes, I hope people that are running giant institutes are doing more than just acknowledging that "there will be an industry of research studying epigenetics", as Hyman said at the lunch. They're missing a huge opportunity if they don't start doing that research themselves.