Seaside Therapeutics has recently announced that they would be withdrawing their Fragile X drug Arbaclofen (STX209) from further clinical trials (see here). They had already reached Phase 3 and the drug showed promise in some patients but probably not enough to secure enough funding to continue or guarantee FDA approval. See the New York Times story for some personal accounts of the impact of this decision. The drug is a GABA-B agonist and is similar to the drug Baclofen, which is used to treat muscle spasms. Fragile X syndrome, which has symptoms similar to autism, is caused by a mutation to the FMR1 gene that silences the production of the FMRP protein. Like most proteins, it is not exactly clear what FMRP does except that it may involve protein translation and affects synaptic plasticity in mouse models. One hypothesis of the cause of autism and Fragile X is that there is an overabundance of synaptic excitation.
My paper with Shashaank Vattikuti explored the effects of such imbalances on a cortical circuit model and showed that it could reproduce some psychophysical experiments (see here for summary of the paper). It is thus a plausible hypothesis that a GABA-B agonist, which are inhibitory transmitters, may alleviate some of the symptoms and I believe that it does in some patients. However, such a blunt instrument would probably not work in all patients. One reason is that not all imbalances between excitation and inhibition are necessarily equal, i.e. too much excitation may not be the same as too little inhibition. A neural circuit with very high excitation and inhibition balancing each other could behave very differently from one with low amounts of each balancing each other. The high circuit would have “high gain” and be very responsive to perturbations while the low circuit would have low gain. It is also not clear that simply increasing inhibition everywhere will result in a net decrease in inhibition because of the multiple feedback loops in the system. Increasing inhibition between inhibitory neurons could decrease the net inhibition on excitatory neurons.
The trials seem to show that about a third of the patients improved with Arbaclofen. They are probably the ones that have too little inhibition and increasing inhibition helps. I think this case suggests that we may need a new model for FDA approval of drugs. Perhaps we should not insist that drugs only treat specific illnesses but should also be approved if they are shown to have some biological effect and do not cause harm. I believe that there are many drugs that have failed to obtain FDA approval that actually do work and could help some patients. Instead of waiting until we can figure out ahead of time which patients will benefit from a given drug before we approve of it, we can just approve of it for restricted use and try it on patients to see what happens. The danger of course is that it may be difficult to know if a drug works and desperate patients and especially parents will insist on using a drug even if the physician believes it has no effect. This could cause harm and increase the cost of medical care. One of the things we could do is to have the government or nonprofit companies take over failed but safe drugs and provide them at low cost under some regulation. Actually, I think we need to completely revamp how drugs are developed but I need to leave that to a future post.