Mind bending drugs

Honoré de Balzac wrote his novels at night, on coffee. He described taking it ground and dry, on an empty stomach.

He described the result himself, in an 1839 essay on modern stimulants. The ideas came up like battalions on a battlefield. Memories charged in with flags flying.

Balzac had a workplace stimulant. His account tells us what it felt like to write on coffee, not whether coffee made him clever.

A century later a Belgian drug company would sell the clever part. The word it used was nootropic, a term now used to sell "brain supplements" online.

For a study published online in 2020, Pieter Cohen and colleagues tested ten supplements bought in September 2019. They selected products whose labels named particular unapproved drugs, not a random sample of the market. All ten contained omberacetam, sold in Russia as Noopept but not FDA-approved. The highest amount per serving was about four times the usual single dose cited by the researchers. Nine of twelve declared quantities were inaccurate.

Those bottles brought together chemically different drugs under one sales pitch. The racetams are one family within that trade, not a name for everything it sells.

Where did the word come from?

In 1964 chemists at UCB, outside Brussels, made a ring-shaped cousin of GABA, the brain's main braking chemical. The file number was 6215. The usual account says it was meant as a sedative. In animals it sedated almost nothing. It had almost no side effects. In some tests the animals learned faster.

The pharmacologist on the project, Corneliu Giurgea, spent the next eight years deciding what that meant. In 1972 he published his answer in a French journal. The compound was the first of a new class. He called the class nootrope, from the Greek for mind and for turning. He set out criteria: better learning, protection against brain insults, almost no toxicity, none of the sedation or stimulation of other psychiatric drugs.

The molecule was piracetam. The word was the product.

In Britain, piracetam is a prescription medicine licensed for cortical myoclonus, a movement disorder. That is a different indication from improving memory. It has never been approved by the FDA for anything, and the agency's position is that it cannot be sold in the United States as a dietary supplement.

The ring stayed. Over the next forty years companies in Belgium, Japan, the Soviet Union and elsewhere hung new side chains on it and tried again.

What did Japan find?

Japan prescribed drugs for these problems under a blunter name. The Japanese category was

脳循環代謝改善剤nō junkan taisha kaizen-zai: drugs that improve cerebral circulation and metabolism.

Through the 1980s and 1990s Japanese doctors prescribed them after strokes, mostly for apathy and memory trouble. Idebenone alone went to an estimated eight million patients between 1986 and 1998.

Then the health ministry ran a re-evaluation. In May 1998 it withdrew four drugs after new placebo-controlled trials failed to establish efficacy: idebenone, propentofylline, indeloxazine and bifemelane. These were not racetams; they belonged to the wider Japanese prescribing category. That August five Japanese authors wrote to The Lancet questioning the approvals.

The letter criticised subjective global-improvement ratings without structured criteria. Trials tested 30–100 endpoints without accounting for multiple comparisons, and excluded so many patients that most endpoints were assessed in only half those eligible. The authors argued that the reported benefits did not reproduce consistently.

The racetams were on the same shelf. Aniracetam was Roche chemistry, sold in Japan as Draganon by Nippon Roche and as Sarpul by Toyama Chemical. Its label covered anxiety, agitation and low mood after a stroke. It got its own post-marketing trial against placebo. A pharmacy notice dated 24 July 2000 reported that both manufacturers were recalling their products voluntarily. The trial had not shown sufficient efficacy; the notice said no new safety problem had prompted the recall.

Other Japanese racetams reached clinical trials without establishing treatments for those indications. Daiichi's nefiracetam was studied for dementia and post-stroke apathy. Nippon Shinyaku took fasoracetam into Phase 3 for vascular dementia, then discontinued that programme. That was not the end of research on the molecule: a small ADHD study published in 2018 examined a different indication. It was exploratory, not proof of routine clinical benefit.

What about the one that worked on choline?

One Japanese branch did something new. Mitsubishi's MKC-231, later named coluracetam, has the racetam ring and a large fused ring system on the other end. Its target was high-affinity choline uptake, the slowest step in making acetylcholine, the messenger that memory research had focused on since the 1970s.

In rats whose cholinergic neurons had been chemically damaged, the drug improved performance on a memory task. A three-paper series from Mitsubishi in 2008 found something odder. After repeated dosing, the improvement was still measurable 24 hours after the last dose, when the drug was no longer detectable in the brain. The researchers proposed that more choline transporters had moved to the neurons' surface.

On 10 August 2006 a San Diego start-up, BrainCells Inc., announced it had licensed the compound. Its release said Mitsubishi had suspended the program because the drug met safety expectations and did not reach its efficacy endpoints. BrainCells renamed it BCI-540 and pursued it for depression with anxiety. The licensing announcement did not establish a benefit in patients. Coluracetam later appeared for sale online.

Does piracetam touch anything real?

It does. It is a weak effect at a real receptor.

In 1990 a team in Japan showed that aniracetam boosts currents through what are now called AMPA receptors, which carry much of the brain's fast excitatory signalling. Twenty years later Ahmed and Oswald at Cornell got crystal structures of piracetam sitting on the same receptor, at a site no drug had been seen using before. Their paper title says piracetam "defines a new binding site." It binds weakly. The structures are in the Protein Data Bank.

A binding site is a long way from a better memory. A 1976 double-blind study in healthy volunteers did report improved verbal learning after fourteen days. That finding deserves to be in the record. It does not, by itself, establish a reliable benefit across people and everyday tasks, and the receptor structures cannot settle that question.

What did Moscow do with it?

The Soviet branch hung a phenyl ring on piracetam's backbone. The result is phenylpiracetam, also listed as fonturacetam and, in sport, as carphedon. It is sold in Russia as a prescription drug. It is on the World Anti-Doping Agency's list of banned stimulants.

The story sellers tell is that it was developed for cosmonauts. That story lives in vendor copy. Pharmakon has not found a primary document that puts it on a flight.

The Russian pharmacy shelf also carries drugs sold for cognition that are not racetams. Omberacetam is a dipeptide derivative, sold as Noopept. It was the drug found in every product in Cohen's sample, and it has no FDA approval.

How did the epilepsy drugs emerge?

UCB never stopped working on the ring. In 1966 its chemists made etiracetam, an ethyl cousin of piracetam. In 1977 they separated its two mirror-image forms. The active one was later named levetiracetam.

Through the 1980s UCB tested it in patients with cognitive impairment. Those trials did not establish it as a treatment.

In 1992 a UCB pharmacologist, Alma Gower, published its anticonvulsant profile. The drug powerfully protected mice bred to convulse at loud noises, and her paper called its pattern of activity unique. In 1993 Wolfgang Löscher's group in Hanover showed that it blocked kindled seizures, a model used to study focal epilepsy. But it was inactive in two conventional screening tests: maximal electroshock and seizures induced by a subcutaneous dose of pentylenetetrazol. UCB's 1998 study confirmed that unusual profile. Relying on those two tests alone would have missed its potential.

The FDA approved levetiracetam for epilepsy in 1999. Its molecular binding target had not yet been identified. UCB had found a specific binding site in brain tissue in 1995. In 2004 a UCB team identified the protein: SV2A, which sits on the small sacs neurons use to release their messengers. The drug had been on the market for five years.

UCB was already designing compounds to bind that site more tightly. Its January 2004 discovery paper described the candidate that became brivaracetam, before the SV2A identification paper appeared. Brivaracetam binds SV2A more tightly than levetiracetam. The FDA approved it in 2016.

A compound developed for memory became an epilepsy medicine.

So what is in the bottle?

In 2001 the Cochrane Collaboration reviewed the piracetam trials in dementia and cognitive impairment. Most were old and badly reported. Doctors' global impressions sometimes improved. Specific measures of memory and thinking mostly did not. The reviewers concluded the evidence did not support clinical use.

In February 2019 the FDA and the Federal Trade Commission sent warning letters to online sellers of racetams. That November, Pieter Cohen's group reported piracetam on sale in American supplements anyway, labeled as such. The 2020 study followed.

The claims, checked

Timeline

Pharmakon