The incredibly rapid improvement in depression (hours) produced by ketamine is unprecedented and surely is telling us something vitally important about depression. If only we could figure out what it is. Clinicians were used to waiting weeks for antidepressants of all sorts to work. As a neurologist, I’d see it work in a week or so in my MS patients depressed due a relapse.
Two recent papers show just how hard it is going to be [Neuron vol. 104 pp. 182 – 182, 338 – 352 ’19 ]. First off you have to accept the idea that even though animals (usually mice) can’t tell us how they feel, we still have reasonable animal models of depression (tail suspension test, forced swim test). We can at least get a handle on anhedonia using the sucrose preference test.
Throw ketamine at an animal and measure the biochemical or the neurophysiologic effect of your choice. There are zillions of them. Throw just about anything at the brain, and all sorts of things change. The problem is showing that the change is relevant. Is the known blockade of NMDA receptors by ketamine how it helps depression. Give enough and you get out of body experiences and all sorts of craziness, not an antidepressant effect.
Homer1a is a protein found at the synapse, and like all scaffold proteins, it interacts with a bunch of different proteins. It links another type of glutamic acid receptor (mGluR1 and mGluR5) to inositol 1, 4, 5 trisphosphate receptors (IP3Rs) on the endoplasmic reticulum. It also links mGluR1 and mGluR5 to NMDARs and other ion channels.
So what?
Other work by the authors showed that knockdown of Homer1a (using small interfering RNA – siRNA) in the medial prefrontal cortex (mPFC) abolished the antidepressant effects (in animal models) to ketamine. Well that’s good, but even better is that knockdown also abolished the antidepressant effects of a tricyclic antidepressant (imipramine).
The present work showed that increasing the expression of Homer1a (the protein comes in various isoforms) in the frontal cortex reduced depression in the various models.
Pretty good — all we have to do is increase Homer1a expression to have a treatment of depression.
Don’t get your hopes up, and this is why depression research is so — well depressing.
Increasing Homer1a expression in another brain region (the hippocampus) has exactly the opposite effects.
The four hour cure for depression: what is KetaminTe doing?
The drugs we use for depression aren’t great. They don’t help at least a third of the patients, and they usually take several weeks to work for endogenous depression. They seemed to work faster in my MS patients who had a relapse and were quite naturally depressed by an exogenous event completely out of their control.
Enter Ketamine which, when given IV, can transiently lift depression within a few hours. You can find more details and references in an article in Neuron ( vol. 101 pp. 774 – 778 ’19) written by the guys at Yale who did some of the original work. However, here’s the gist of the article. A single dose of ketamine produced antidepressant effects that began within hours peaked in 24 – 72 hours and dissipated within 2 weeks (if ketamine wasn’t repeated). This occurred in 50 – 75% of people with treatment resistant depression. Remarkably one third of treated patients went into remission.
This simply has to be telling us something very important about the neurochemistry of depression.
Naturally there has been a lot of work on the neurochemical changes produced by ketamine, none of which I’ve found convincing ( see https://luysii.wordpress.com/2019/10/27/how-does-ketamine-lift-depression/ ) untilthe following paper [ Neuron vol. 106 pp. 715 – 726 ’20 ].
In what follows you have to have some basic knowledge of synaptic structure, but here’s a probably inadequate elevator pitch. Synapses have two sides, pre- and post-. On the presynaptic side neurotransmitters are enclosed in synaptic vesicles. Their contents are released into the synaptic cleft when an action potential arrives from elsewhere in the neuron. The neurotransmitters flow across the very narrow synapse to bind to receptors on the postsynaptic side, triggering (or not) a response of the postsynaptic neuron. Presynaptic terminals vary in the number vesicles they contain.
Synapses are able to change their strength (how likely an action potential is to produce a postsynaptic response). Otherwise our brains wouldn’t be able to change and learn anything. This is called synaptic plasticity.
One way to change the strength of a synapse is to adjust the number of synaptic vesicles found on the presynaptic side. Presynaptic neurons form synapses with many different neurons. The average neuron in the cerebral cortex is post-synaptic to thousands of neurons.
We think that synaptic plasticity involves changes at particular synapses but not at all of them.
Not so with ketamine according to the paper. It changes the number of presynaptic vesicles at all synapses of a given neuron by the same percentage — this is called synaptic scaling. Given 3 synapses containing 60 50 and 40 vesicles, upward synaptic scaling by 20% would add 12 vesicles to the first 10 to the second and 8 to the third. The paper states that this is exactly what ketamine does to neurons using glutamic acid (the major excitatory neurotransmitter found in brain). Even more interesting, is the fact that lithium which treats mania has the opposite effects decreasing the number of vesicles in each synapse by the same percentage.
I found this rather depressing when I first read it, as I realized that there was no chemical process intrinsic to a neuron which could possibly work quickly enough to change all the synapses at once. To do this you need a drug which goes everywhere at once.
But you don’t. There are certain brain nuclei which send their processes everywhere in the brain. Not only that but their processes contain varicosities which release their neurotransmitter even where there is no post-synaptic apparatus. One such nucleus (the pars compacta of the substantia nigra) has extensively ramified processes so much so that “Individual neurons of the pars compact are calculated to give rise to 4.5 meters of axons once all the branches are summed” — [ Neuron vol. 96 p. 651 ’17 ]. So when that single neuron fires, dopamine is likely to bathe every neuron in the brain. We think that something similar occurs in the locus coeruleus of the lower brain which has only 15,000 neurons and releases norepinephrine, and also in the raphe nuclei of the brainstem which release serotonin.
It should be less than a surprise that drugs which alter neurotransmission by these neurotransmitters are used to treat various psychiatric diseases. Some drugs of abuse alter them as well (Cocaine and speed release norepinephrine, LSD binds one of the serotonin receptors etc, etc.)
The substantia nigra contains only 450,000 neurons at birth, so you don’t need a big nucleus to affect our 80 billion neuron brains.
So the question before the house, is have we missed other nuclei in the brain which control volume neurotransmission by glutamic acid? If they exist, could their malfunction be a cause of mania and/or depression? There is plenty of room for 10,000 to 100,000 neurons to hide in an 80 billion neuron brain.
Time to think outside the box people. Here is an example: Since ketamine blocks activation of one receptor for glutamic acid, could there be a system using volume neurotransmission which releases a receptor inhibitor?
Addendum 7 July — I sent a copy of the post to the authors and received this back from one of them. “Thank you very much for your kind words and interest in our work. Your explanation is quite accurate (my only suggestion would be to replace “vesicles” with “receptors”, as the changes we propose are postsynaptic). Reading your blog reassures us that our review article accomplished its main goal of reaching beyond the immediate neuroscience community to a wider audience like yourself.”