Overweening Generalist

Wednesday, October 2, 2013

The Drug Report: Crisis In Psychopharmacology

It's been at least 30 years since a truly new drug has hit the market that addresses the needs of patients suffering from depression, anxiety, manic depression (now rather bloodlessly called "bipolar disorder"), and schizophrenia. Any "new" drugs in the last 30 years have been basically some variation on an older, established drug (called "Me Too" drugs), in an effort of competing drug companies to keep up with the competition. These non-new "new" drugs are almost always marketed as "blockbuster" or "revolutionary" therapeutics, touting less side effects than older, competing drugs. They are not new and the side effects are just different, not less. 50 or so psychiatric drugs bring in $25 billion a year in Unistat alone. And they're pretty lousy.

(I know, I know: you'd be far worse off without the one that worked for you. Hey: they do some good. For some people. I want better drugs for you, is all. And we were promised them with the 2000 mapping of the human genome. So...where are they? Later.)

                                                        serotonin

The drugs people use - by every estimate I've seen between 20% to 25% of the Unistat population takes  at least one of these - were discovered by accident. By serendipity. In the 15 years after 1945. In 1952 a tuberculosis drug didn't work for TB, but iproniozid sure elicited euphoria when tested! Bingo: the first antidepressant. The drug that became Tofranil was supposed to work for schizophrenics, but it didn't help them, only make them run naked into town, laughing. Another antidepressant. In 1949 lithium was discovered, by accident, to treat manic depression. In 1957 Leo Sternbach was about ready to give up his research into a class of antihistamines, things were looking like a dead-end, when he stumbled onto the benzodiazepines: your Valium, Xanax, Lorazepam, Klonopin, etc: an empire of anti-anxiety drugs, and a huge influence on the tonality of culture in the West in the latter half of the 20th century.

With better technics, we learned much more about neurons and neurotransmitters. The SSRIs seemed to treat depression and anxiety. They were really the last big breakthrough. Ever since then, clinical trials that have made it to Stage III have been nothing but huge, sad, very expensive wastes. And so Novartis, Glaxo-Smith-Kline, Astra Zeneca, Pfizer, Sanofri and Merck have by and large quit trying. They've halted clinical trials, moved onto research that shows more promise. The pipeline for new psychopharmacological drugs is dry.

                                    psilocybin, very much like serotonin in structure

Wait a minute: with more neuroscientists than ever before, far better imaging devices, a tremendous acceleration of knowledge about the human brain over the past 30 years...why? And mental health takes an increasing toll on us. If not you, someone you know. Why is this so difficult? Is it because what R.D. Laing called "the medical model" finally showed its hand? (A pair of nines?)

Again: our technology to map with ever finer-grains our cells, genes, and organs is greater than ever. We now have a deeper understanding of the human genome, an explosive discovery of the complexity of the epigenome, increasing understanding of how our environment and microbes interact with us...why don't we have a drug that will cure depression by now? Are we simply too complex to understand? Were we destined to be granted a brief window of time in which a few "happy accidents" would yield up as good as it gets, and it all ended 30 years ago? What about our computing power and pharmacological knowledge? Isn't it also subject to Moore's Law: a doubling roughly every 18 months? Shouldn't we have had a bevy of breakthroughs by now?

What are we doing wrong?

In 2011 Eli Lilly thought they had a breakthrough for schizophrenia. They'd given PCP to mice, then their new drug and...the mice calmed down! Everything went well. They got to Stage III clinical trials (humans) and 18 months later the drug was dead. Placebos worked just as well. Lilly is another company that has all but given up now too.

                                    LSD: like psilocybin and serotonin, structurally

Some New Ways of Thinking and Genuine Promise 
Steven Hyman of Harvard and M.I.T. knows this field well. He was quoted in an article I read as admitting of his colleagues, "People are tired of curing mice."

Let's go back to the last breakthough: Prozac and all its cousins.

It had been assumed that, when those happy accidents occurred, there must be a theoretical basis. Pharmacologists have always acted like they were on top of what was going on, but the trade secret was they were faking it: when a drug worked, it went on the market, people used it and they "worked" well enough, but at first the chemists and psychiatrists had no idea why. With better understanding of the brain, they found the ancient model of the imbalance of humors as an explanatory scheme. Only they juiced it up: they found  these drugs altered neurotransmitters. Therefore, the lack of the neurotransmitter caused the disease! It seemed quite plausible, and very much like the hardcore finding that insulin works for diabetics.



Nassim Nicholas Taleb says this is a classic case of the "reverse-engineering problem": drop an ice cube on the floor and then go play cards with your friends in the other room. Can you visualize the cube breaking down into a tiny pool of water? Of course you can. You walk back into the kitchen and see a tiny pool of water where you had dropped the cube. It's pretty straight-forward. Now: imagine walking down the street and coming upon a tiny pool of water. A little spot of wet. How many ways can you dream up the cause of this spot?

A cop comes upon a drunken man looking for his keys, at night, under a streetlight. The cop asks the drunk why he keeps looking under the streetlight, and the drunk says it's because the light is so much better there.

Obviously, even our best researchers have been looking where the light was bright. And the reverse-engineered explanation of our not-all-that-great/we-can-do-better psychopharmacological drugs? Human. All-too human.



The neurotransmitters are not the cause of the mental illness. They merely point at the underlying cause; neurotransmitters (dopamine, serotonin, norepinephrine, etc) are tangential and partial. Reverse-engineering to allow more serotonin to remain in the synaptic gap between neurons was a genius move; too bad there are a handful of studies that show SSRIs work little better than placebos. (For some people they have worked well enough; I don't want to slight this!) All in all, there's a "truthiness" about depression drugs.

We treat everyone the same in studies, while knowing they have variable epigenomes. This is receiving some major research and seems quite promising, to my eyes. We have a semantic problem with experts dealing with a patient, making observations and tests, then naming the disease they "have," which is a major problem: people and diseases do not fall into our socially-constructed and convenient categories as well as we'd like. This problem is now far more acknowledged than ever, which seems promising to me. One example is the Research Domain criteria: we map behavioral abnormalities and symptoms and link them to specific causes in the brain, without the label of "schizophrenia" or "panic disorder." Why is this approach better? Because it's more targeted. Instead of looking at one or two neurotransmitters that "cause" schizophrenia, we try to find out specifically what causes people to hear voices, or become catatonic.

The idea that we must take 18 years from conception through clinical trials is being re-thought. Even more crucially for mental disease: non-human animal studies long ago reached diminished returns. Now the idea is small-scale, carefully controlled studies on humans will speed up the process and may yield breakthroughs in shorter periods.

Another area of promise: when a drug failed, it often worked for a few people. But our gold standard of drug testing: double-blind and placebo-controlled? The rules were that if the placebo worked as well as the drug, throw out the drug. But the people who were helped probably should have told us something.

Along those lines, there is a strong call to restore abandoned or "invisible" clinical trials to correct the scientific record. We may learn some very interesting things from "failed" trials.

The techniques surrounding stem cells have accelerated at an incredibly dizzying pace upward and for the better: now researchers can test cells and drugs in a a dish and make very good guesses as to whether a compound would have some efficacy.

With the mapping of human genome in 2000, hundreds of utopian promises were made that now seem embarrassing or outright quackery. But there was reason to be optimistic. We thought because we were very complex, we'd have the most genes, but instead of 100,000 we only had about 21,000. Grapes have more genes than us: this was nothing like what we'd expected. Worse: 13 years later we now know that a "bigger" system - in terms of complexity - governs the genome: the epigenome. It turns out that RNA plays a far, far bigger part than we'd thought. The complexity can seem overwhelming.

In 2002 researcher Andrew Hopkins came up with an eye-opening paper, the "druggable genome": Okay: we'd thought we had 100,000 genes. We have closer to 21,000. He estimated that only about 10% of those genes coded for proteins that could bind to small molecules, which is how drugs work, basically. So: about 2,100 genes. But he estimated that, of those, only about 20% would be likely to involve diseases. So now we're down to about 420 possibilities for targets. And then he guessed we'd already discovered 50% of those (probably accidentally?). We only had 210 targets left? For all diseases, not just mental illnesses? Not exactly a rosy scenario. But...

Cheminformatics! This is a burgeoning discipline using the aforementioned computational doubling: there are tens of thousands of compounds in digitized libraries. Do you test them all? Two guys wrote  an algorithm to teach a computer to sift through a welter of data on TB, which is becoming antibiotic-resistant. A Big Deal, quite threatening to all of us, potentially. Their algorithm said: find all compounds that are like the drugs that used to work on tuberculosis. So you get that data set. Then the algorithm says, throw out every compound known to be toxic to mammalian cells. You have a smaller set, but a safer one to work with. The algorithm discovered a 40-year old drug that was shown to have anti-TB properties but had been forgotten.




Even more interesting and promising: researchers in Cambridge, MA have taken messenger RNA (mRNA), an ultra fragile molecule which, when injected activates the body's immune response, tweaked a couple of "letters" in its nucleotide sequence, and made a non-fragile mRNA that does not turn on the immune system. What this could do is take the information from the DNA in a gene and make it "fix" missing or broken proteins in another cell, in effect causing a patient with a (probably inherited?) protein abnormality to make a drug inside their own cells!

Nessa Carey, a gifted explainer of how epigenetics works in our bodies, has urged us to be cautious about getting too excited over drugs based on DNA-RNA, because so far, "One of the major problems with this kind of approach therapeutically may sound rather mundane. Nucleic acids, such as RNA-DNA, are just difficult to turn into good drugs. Most good existing drugs - ibuprofen, Viagra, antihistamines - have certain characteristics in common. You can swallow them, they get across your gut wall, they get distributed around your body, they don't get destroyed too quickly by your liver, they get taken in by cells, and they work their effects on the molecules in or on the cells. Those all sound like really simple things, but they're often the most difficult things to get right when developing a new drug."

Finally, there is a very real call to combine all our new technologies with an active looking for happy accidents, like in the 1945-60 period. We find as many compounds that could possibly have efficacy, get people willing to be guinea pigs to try them (we have far better ways to guess at what's likely to have horrendous side effects or death-dealing qualities, but we're by no means "covered" here), and see what happens! Yes, the dark side is that the poor will probably be the ones to sign up...How do we find new things to try? "Scientists Map All Possible Drug-Like Chemical Compounds." It turns out the drunk looking for his keys was far more accurate an analogy than we might've guessed. Or wanted to guess. Check out all the unexplored chemical "space" yet to be charted! It reminds me of the incredible number of phenethylamines and tryptamines that Alexander Shulgin mapped: but a drop in the ocean? (Shulgin deserved the Nobel Prize for Chemistry: just read-up on his career! It's almost criminal he didn't get the Prize.) It's like looking for signs of life in the Milky Way! Or more prosaically: like geologists learning how to more profitably drill for oil. It's also about algorithms and possibilities and adventure and hellacious mistakes yet to be made.

To all of us looking for better living through chemistry: Bon appetite! I do think we may make it through this bottleneck to a whole new world of more sophisticated drugs that will make all the ones we've had since 1945 look primitive. Maybe?

Some Of The Works Consulted:
The Epigenetics Revolution by Nessa Carey
"No New Meds," by Laura Sanders:
http://www.sciencenews.org/view/feature/id/348115/description/No_New_Meds
Happy Accidents: Serendipity In Modern Medical Breakthroughs, by Morton A. Meyers
"The Psychiatric Drug Crisis" by Gary Greenberg:
http://www.newyorker.com/online/blogs/elements/2013/09/psychiatry-prozac-ssri-mental-health-theory-discredited.html
PIHKAL: A Chemical Love Story, by Alexander and Ann Shulgin
"Where Are All The Miracle Drugs?" by Brian Palmer:
http://www.slate.com/articles/health_and_science/human_genome/2013/09/human_genome_drugs_where_are_the_miracle_cures_from_genomics_did_the_genome.single.html
"Messenger RNAs Could Create a New Class of Drugs," by Susan Young:
http://www.technologyreview.com/news/512926/messenger-rnas-could-create-a-new-class-of-drugs/
"Faster, Smarter and Cheaper Drug Discovery":
http://www.sciencedaily.com/releases/2013/03/130321131920.htm
Serendipity: Accidental Discoveries In Science, by Royston Roberts
Hope or Hype: The Obsession With Medical Advances and the High Cost of False Promises, by Richard A. Deyo and Donald L. Patrick
"Experts Propose Restoring Invisible and Abandoned Trials to 'Correct the Scientific Record'":
http://www.sciencecodex.com/experts_propose_restoring_invisible_and_abandoned_trials_to_correct_the_scientific_record-114055
The Black Swan: The Impact of the Highly Improbable, by Nassim Nicholas Taleb

7 comments:

Eric Wagner said...

Terrific piece. I can't think of much of a response except "420 possibilities."

The 1960's and 1970's seemed to present a fifth circuit revolution. The 1990's and 2000's presented a sort of "Cyber-Electronic Piloting Quantum Electronic Info" revolution, using Leary's name for the sixth circuit. Perhaps the coming years will present a "Cyber-Genetic Piloting DNA/RNA Data" revolution (using Leary's words again). Perhaps a lot of kids will study chemistry after watching "Breaking Bad."

Anonymous said...

Nicely done, a good packaging of a
number of complex conundrums.

From my view there is a real great
misunderstanding of the how/what of
the human nervous system. Tinkering
around with the synapses isn't going
to fix things like schizoidal acts.

My money is on Ramachandran and his
students to pin down voices in the
head to a specific regional module.
Then you can target the region but
drugs might not be the way to do
that.

There's a bit of tension between
science and medicine. My last
clash with my doctor was over his
blanket statement about all of the
studies that proved something.

I had just read the paper that
pointed out that an enormous
number were hand selected or never
followed up for peer reviews and
verification. It took awhile but
he ended up in agreement. This is
not a fault on his part, they don't
have time to keep up with much
outside (or inside) their field.
Doctors used to make soothing and
uplifting noises while patients
cured themself, that worked well
until the pesky science boys began
intervening with drugs that worked.

The economics of becoming rich off
the cash cow drug and spending on
propaganda instead of boring old
research isn't new to pharma. The
same methods killed American Steel
turned Detroit into the disneyland
of the midwest industrial belt.

I pay a lot more attention to the
nanotech people than the DSM types.

You'd think the world would be a
lot happier more cheerful place
with all those psych drugs being
taken. Maybe their inner worlds
are but the external hasn't been
given much benefit.

michael said...

Eric- I like the 5th-6th-7th circs mapped onto historical trends. I thought the 420 was funny too...it was my extrapolation from the math given in a couple of articles, not by the authors themselves. I tried to find Hopkins's original "druggable genome" paper, but kept running up against a paywall. How fucking stupid: a science paper from 2002 and they want me to pay how much to read it? There's all kinds of problems with scientific journals: they're too biased, too expensive for libraries to buy...it's all crap...fodder for another blog?

I have read a few articles about how Breaking Bad has romanticized chemistry and made some students suddenly interested. Same with the CSI shows: lots more students interested in forensic science now. I think there's something to embedding a discipline in a dense narrative and making it seem mysterious, wonderful, powerful and potentially empowering. (Fodder for another blog?)

Back to the genetic archives: Leary died far too soon. He would have been blown away by what we've found in our genes already!

michael said...

Anon-

My spew was once again far too long, and I had to leave out some really interesting stuff being done with fine-tuned implants and pulses of current to stave off things like hearing voices. No drugs at all.

I try to read everything Ramachandran puts out. Not only is he an interesting neuroscientist, but he's articulate enough to shed light on ancient Qs in philosophy via neuroscience.

I've had similar run-ins with doctors. When they find out you've read a lot of the literature, they drop the facade and get real...except on psychiatrist I knew. He seemed totally sold on SSRIs as miraculous things. Me: But what about the welter of studies that show they're little better than a placebo? And in a few studies patients did BETTER on the placebo?"

Shrink: There are all kinds of Scientologists and other kooks who want you to jettison science for their snake oil.

re: Drs who can't keep up: I had one doctor tell me she keeps up by watching shows like 20/20. And she was serious. Primary care doctors are soooo busy they really don't have the time to keep up with the literature. The best delivery of info: their "group" has decided some new procedure is "best practice," so they go along: some theorist higher up (probably allied with insurance groups) has convinced the local group of doctors that some new wrinkle is better for care. Secondly: doctors get new info at conferences, which are heavily funded by private interests. Videos are made to try to get to doctors, but apparently they don't watch them very often. (Would you?)

There was this amazing savant named Roy Porter, who died a few years ago. He's written some anthropology stuff that blew me away. Then I found he wrote a history of medicine, The Greatest Benefit To Mankind, around 900 pages. Actual scientific "medicine" is still in its infancy, as you know. The hopeless, galling, ridiculous quackery that passed for "medicine"! Amazing book. For most of the last 2000 years, it was a far better choice to go to one of the "weird" women who lived on the outskirts of town...

The inklings of a new paradigm for finding meds - not only for mental illness but for all sorts of maladies - will go Big Money as soon as some group of researchers around the universities hit on something big: they'll sell it to Eli Lilly or Pfizer, and their investors will be there. Or: maybe not.

I'm still trying to get a line on nanotechnology and its medical promises, which are jaw-droppingly great. It's difficult to tell when this stuff will cash out for...us. I hazard a guess that it will be at first flush with success in two or three specific diseases that it's good at. When it comes to inherited diseases, it seems the mRNA and similar approaches will be used...unless you can teach the nanoparticles to get into a faulty cell and fix it. Until them, I see it as a sort of super-augmentation of the immune system, and as very-efficient phage systems...eating away a tumor and leaving the healthy tissue behind.

But when?

One thing about epigenetics that's related to happiness: the environment that a genome finds itself in has a heavy influence on what genes become expressed...and oddly, this becomes heritable in many instances. This adds a profound level of support to social theorists who say that if you improve the environment you improve people...which would seem to imply that you allow them to become happier.

Poverty seems like one of the most violent crimes in the world, in this scenario. The inner world reflects the outer, and vice-versa, of course!

Eric Wagner said...

Coincidentally, I just reached page 420 in rereading Vico's The New Science, wherein he discusses pimps.

michael said...

On pimps in Plautus? Coincidentally, 420 makes those passages seem even funnier; or rather: It works for me.

Drew Zi said...

"Nassim Nicholas Taleb says this is a classic case of the "reverse-engineering problem": drop an ice cube on the floor and then go play cards with your friends in the other room. Can you visualize the cube breaking down into a tiny pool of water? Of course you can. You walk back into the kitchen and see a tiny pool of water where you had dropped the cube. It's pretty straight-forward. Now: imagine walking down the street and coming upon a tiny pool of water. A little spot of wet. How many ways can you dream up the cause of this spot?"


Popper, although the piece was criticised, did claim in conjectures and refutations, and even though the argument he made for it formally fails, it is still a good insight, is that "normal" scientific practice tries to find explanations that are plausible (which I think he claimed was some kind of confirmation bias), whereas his theory was that something in its strongest form is not only testable but also should strive to be improbable, because if it is improbable based on our current theories/prejudices, it has the potential to point out something more significant about our earlier theories even if in the long run it turns out to be false.