It isn’t every day that a paper comes along which gives you a new way to study neurodegenerative diseases in which synapses shrink (which is essentially all of them). The paper also ties together two proteins, each intensively studied for decades on which tens of thousands of neurochemical, neuropharmacological and neurophysiological papers have been written. I speak of Neuron vol. 112, pp. 2708 – 2720 ’24.
First: the meat of the paper assuming you’ve been reading the research literature for years.
Second: explanation and elaboration of what the paper is about for the rest of us
The meat: A 108 amino acid breakdown product of the amyloid precursor protein (APP) called AETA (A neta) binds to an ion channel receptor for glutamic acid (NMDAR) known to be involved in learning, shutting down its ability to conduct ions across the postsynaptic membrane and in some way (unknown at present) and causing synapses to shrink.
Explanation and elaboration: The amyloid precursor protein (APP) has been studied for years and years because one of its breakdown products (the aBeta peptide of around 40 amino acids) is the main constituent of the senile plaque of Alzheimer’s disease. APP is big with from 695 to 770 amino acid proteins formed from the gene (depending on the splice variants chosen).
While still in the cell, APP is broken down by 3 enzymes called secretases (because the breakdown products are secreted). The secretases are alpha, beta and neta. Neta secretase appears to be another name for the very well studied gamma secretase and the term is idiosyncratic being found only in this paper with a Google search turning up nothing (this took me about an hour).
The first cut of APP leading to the abeta peptide occurs within the transmembrane segment of APP and is due to beta secretase action. This produces a (amino terminal) fragment which is subsequently cleaved by gamma secretase to form the aBeta peptide. Cleavage of APP can alternatively occur due to alpha secretase, which prevents subsequent beta secretase cleavage, preventing a beta peptide formation when gamma secretase gets to work on the product. It was a quite logical move to inhibit beta secretase as a potential treatment of Alzheimer’s disease, but it hasn’t worked. The paper under discussion may tell us why.
Onward and upward to NMDAR, one of the many (18 genes) receptors for glutamic acid, the main excitatory neurotransmitter used in the brain. NMDARs are tetramers of these gene products. Activation of NMDARs also requires binding of one of two other transmitters (D-serine, glycine) to a site different than the glutamic acid binding site. There is plenty of room for multiple binding sites. The tetramer contains over 3,400 amino acids. NMDAR activation also requires membrane depolarization (to relieve a Magnesium ion blocking the ion channel of NMNDAR) making it a Hebbian coincidence detector, and many experiments have shown that NMDARs are crucial for the best neurophysiologic model of learning we have (Long Term Potentiation, aka LTP).
Well, usually activation of an ion channel means opening the channel so ions flow across the neuronal cell membrane in which the channel is embedded. This is where Neuron vol. 112, pp. 2708 – 2720 ’24 gets interesting. The paper describes evidence that NMDARs can produce synapse depression and spine shrinkage, completely independently of their ability to conduct ions.
Moreover, the paper describes a culprit shutting down ion conduction of NMDAR, namely a 108 amino acid product of APP cleavage called AETA (A neta) by Secretase neta (whatever that is). AETA competes with coAgonists D-serine and glycine for their binding site. 108 amino acids is a large enough protein to bring two intracellular tails of two of the proteins making up the tetramer together shutting down ion conduction.
The paper shows that production of AETA increases as neuronal activity increases, making it a brake on excessive neuronal activity and excitotoxic cell death.
The best laid plans of mice and men department. Recall that inhibiting beta-secretase seemed like a logical way to treat Alzheimer’s (it decreased aBeta peptide production). This work shows that another side effect of beta-secretase (BACE1) inhibition is increased AETA levels. It also notes that increased AETA levels cause synapses to shrink and promotes long term depression (LTD), the opposite of memory formation; just what Alzheimer’s disease brains don’t need. The mechanism by which this happens isn’t known at present. Finding out just what the mechanism is and figuring out ways to inhibit it might be a way to treat Alzheimer’s disease.
Even better, NMDAR function has been ‘implicated’ (a weasel word if there ever was one implying accusation without proof) in just about every neurodegeneration known. This implies that AETA production and its effects on NMDARs should be studied in all neurodegenertions.
So this paper is really a blockbuster giving us a new way to look at neurodegeneration, something desperately needed given our minimal therapeutic progress so far.
