Category Archives: Neurology & Psychiatry

The evolutionary construction and magnification of the human brain

Our brains are 3 times the size of the chimp and more complex.  Now that we have the complete genome sequences of both (and other monkeys) it is possible to look for the protein coding genes which separate us.

First some terminology.  Not every species found since the divergence of man and chimp is our direct ancestor.  Many banches are extinct.  The whole group of species are called hominins [Nature vol. 422 pp. 849 – 857 ‘ 03 ].  Hominids are species in the path between us and the chimp — sort of a direct line of descent.  However the terminology is in flux and confusing and I’m not sure this is right.   But we do need some terminology to proceed.

Hominid Specific genes (HS genes) result which result from recent gene duplications in hominid/human genomes.  Gene duplication is a great way for evolution to work quickly.  Even if one gene is essential, messing with the other copy won’t be fatal.  HS genes include >20 gene families that are dynamically expressed during the formation of the human brain.  It was hard for me to find out just how many HS genes there are.

Here are some examples. The human-specific NOTCH2NL genes increase the self-renewal potential of human cortical progenitors (meaning more brain cell can result from them).  TBC1D3and ARGHAP11B, are involved in basal progenitor amplification (ditto).

A recent paper [ Neuron vol. 111 pp. 65 – 80 ’23 ] discusses CROCCP2 (you don’t want to know what the acronym stands for) which is one of several genes in this family with at least 6 copies in various hominid genomes.  However, CROCCP2 is a duplicate unique to man.   It is highly expressed during brain development and enhances outer Radial Glial Cell progenitor proliferation.

The mechanism by which this happens is detailed in the paper and involves the cilium found on every neuron, mTOR, IFT20 and others.

But that’s not the point here, fascinating although these mechanisms are.   We’re watching a series of at least gene duplications with subsequent modifications build the brain that is unique to us over relatively rapid evolutionary times.  The split between man and chimp is thought to have happened only 8 million years ago.

What should we call this process?  Evolution?  The Creator in action? The Blind Watchmaker?   It is certainly is eerie to think about.  There are 17 more HS genes to go involving in building our brains remaining to be worked out.  Stay tuned

Overblown Stock Market Reaction to Simufilam results

The stock market reaction to Simufilam’s 1 year open label results is extremely overblown.  Here is a link to the results — https://www.cassavasciences.com/news-releases/news-release-details/cassava-sciences-announces-positive-top-line-clinical-results

First: no other therapy has shown improvement in Alzheimer patients.  The best they can claim is a slower rate of decline.

Second: In 30+ years of clinical neurologic practice, I never saw anyone with Alzheimer’s get better after a year. One or two remained stable for a year, but everyone else got worse. Cassava’s results are impressive (with nearly half improving at one year)  and unique. There is little reason to doubt them, given the way the data has been handled.

Third: even though not a controlled study, a placebo effect is extremely unlikely given my clinical experience with Cognex (Tacrine) when it came out — for details please see — https://luysii.wordpress.com/2023/01/25/why-cassavas-simufilam-results-are-not-a-placebo-effect/

Fourth: the realities of clinical practice.  Assuming that Simufilam is released with data similar to the 1 year results, as a physician I would be remiss if I didn’t offer a drug with nearly a 50% chance of improvement at one year, given the current miserable therapeutic landscape.  Back in the day no patient refused trying Cognex.  Then there is the likelihood of being sued for NOT giving Simufilam, as people were sued for not giving tissue plasminogen activator for stroke, a therapy with minimal evidence for it when it came out — for details please see — https://luysii.wordpress.com/2015/09/02/reproducibility-and-its-discontents/

Fifth: the fact that not everyone responds to Simufilam is irrelevant to eventual FDA approval.  Given all the illnesses we are heir to, even the best drug for any particular illness among the many does not work for everyone with it.  For more on these thoughts please see – https://luysii.wordpress.com/2023/01/26/the-fact-that-not-everyone-responds-to-simufilam-is-irrelevant-to-its-eventual-fda-approval/

The fact that not everyone responds to Simufilam is irrelevant to its eventual FDA approval

A very intelligent friend does not share my optimism about Simufilam.

“Is the data really that positive? ADAS-Cog mean scores changed minimally over 1 year in patients with mild-to-moderate Alzheimer’s disease.  47% of patients improved ADAS-Cog over 1 year by 4.7 points. But 23% of patients declined by <5 points. Mild patients responded better than patients with moderate Alzheimer’s.”

Why are these thoughts irrelevant to the eventual approval of Simufilam by the FDA?

First: no drug for anything works for everyone with the condition

Second: The assumption that Alzheimer dementia is a single disease is based on just that: an assumption.

An example: When I was running a muscular dystrophy clinic in MonrN (’71 – ’87), we saw something called limb girdle muscular dystrophy , in which the patients were weak primarily in muscles about the shoulders and hips. Now we know that there are at least 13 different genetic causes of the disorder.

If the clinical picture of Alzheimer’s disease is due to multiple causes, it is unsurprising that Simufilam doesn’t help all of them.

Also it is time for some humility about our knowledge about Alzheimer’s disease.  We have misunderstood what the senile plaque of Alzheimer’s disease really is for 111 years — see the following post written 12/22 — https://luysii.wordpress.com/2022/12/13/111-years-of-study-of-the-alzheimer-plaque-still-got-it-wrong-until-now/

Third (and probably the most relevant for FDA approval):  Less that perfect drugs will be approved if every other treatment is worse.

The example of immune checkpoint blockade therapy for cancer is particularly relevant.

Some absolutely spectacular results for the therapy has led to the approval of 6 different drugs in this class (all of them monoclonal antibodies against proteins involving the immune system).

One example [ Cell vol. 162 pp. 1186 – 1190 ’15 ]:  “20% of metastatic melanoma patients are cured with Ipilimumab, a fully humanized anti-CTLA4 monoclonal antibody.”

Would that results like this were the rule not the exception. Unfortunately — [ Nature vol. 565 pp. 43 – 48 ’19 ] “Most patients with cancer either do not respond to immune checkpoint blockade or develop resistance to it.”

So what.

Immune checkpoint blockade, despite being less than perfect,  is  still being offered to cancer patients, just the way Simufilam with its nearly 50% chance of improvement at 1 year should be offered to Alzheimer patients.  

Why Cassava’s Simufilam results are not a placebo effect

Any open label study without controls is subject to the reasonable criticism that any benefits seen are due to the placebo effect.  Neurologic and Psychiatric disease trials can have large (33%) placebo effects (e.g. migraine, depression).

This is very unlikely to be the case with the 1 year results of the open label trial of Simufilam in 200 patients with Alzheimer’s disease.  “47% of patients improved on ADAS-Cog over 1 year, and this group improved by 4.7 points”

It is based on my clinical experience with a drug for Alzheimer’s disease released 30 years ago — Tacrine (Cognex).  Initially it was touted as helping Alzheimer’s disease (e.g. improving thinking and memory) although later it was held to slow the decline.   The local medical school advertised it aggressively (primarily as a marketing device).

So I put my Alzheimer patients on Cognex.  I wanted them to get better. They wanted to get better, and their families and caregivers certainly did.  Just about all of them thought it might have helped on followup visits in the first month.  I couldn’t see much difference.  By the second month, they weren’t sure, and later in the first year they didn’t think it helped, and most weren’t using the drug after 1 year.

Not only that, but I was in a call group with 4 other neurologists, and they saw exactly the same thing.  I was practicing in an area with a catchment area of over a million people, and every local neurologist I talked to had the same experience. People thought that “Cognex helped” for a month or two and then they didn’t

This a classic example of a placebo effect.  Moreover it occurred in a therapeutic trial for Alzheimer’s disease.  Crucially, the placebo effect was quite transient and  absent at 1 year.

This is why the Simufilam results mentioned above are not a placebo effect.

Those not interested in neuropharmacology can stop at this point.  There were excellent clinical and theoretical reasons for the use of Tacrine.

Clinically it was apparent that drugs that blocked the effects of the neurotransmitter acetyl choline on one type of its receptors (muscarinic) profoundly impaired memory. Scopolamine is one such drug.  One of the earliest and most invariable symptoms of Alzheimer’s disease is deficient memory.

That’s the clinical part.  Here’s the theory.  So logically, increasing acetyl choline should help memory.  How to do this?  Well there are enzymes that break acetyl choline down (the acetyl cholinesterases).  So by inhibiting cholinesterases, acetyl choline levels in the brain should increase, and memory should be improved.

Impeccable logic and theory.  Unfortunately, like many such it didn’t work.

The synapse is not a oneway street

Back in the day when I was a Med student at Penn ’62 – ’66 and not that far away from the first programmable computer (The ENIAC) built at Penn in the 40’s using vacuum tubes (the transistor was far in the future) we ‘knew’ that information flowed across the synapse in one way, the same way that current flowed just one way in vacuum tubes.  Our conception of the synapse that way was that the presynaptic side was the master and the post synaptic side was the slave.  It worked well as a model of how the brain computes things initially, with current flowing through boolean diagrams of neurons.  It still works pretty well today with neural nets where information flows pretty much in one way through the layers of ‘neurons’.

Well that’s not the way we think about the synapse today, and an excellent paper [ Neuron vol. 110 pp. 4144 – 4161 ’22 ] brings it home.  It shows that the same axon releases its neurotransmitter (glutamic acid) differently depending on what post-synaptic neuron it is innervating.

When the axon from a pyramidal neuron synapses on a fast spiking parvalbumin containing interneuron there is a high probability of glutamic acid release with each nerve impulse (action potential).  The same axon when innervating another type of interneuron (containing somatostatin), the release probability is 10fold lower.

So it all depends on what is being innervated.  The somatostatin interneuron expresses Elfn1 (you don’t want to know what this acronym is for) which activates a presynaptic receptor for glutamic acid (mGluR7) which reduces synaptic release probability.

This isn’t seen in the parvalbumin interneurons.  Essentially the post-synaptic neuron is partially controlling the presynaptic neuron.

The work was done in the hippocampus, where the neuroanatomy and neurochemistry has been worked out better than most areas of the brain.  How widespread differential release of transmitter is, will have to wait until we understand the neuroanatomy of other brain areas better.

If you want to read about the building of the ENIAC at Penn, have a look at the following post — https://luysii.wordpress.com/2022/09/18/book-review-proving-ground-kathy-kleiman/.  Trigger warning — the women who figured out how to program the ENIAC were not treated particularly well but decades later received the recognition they deserved.

Finally, an article in the press that’s not a hit piece on Cassava

Cassava Biosciences has had the worst press imaginable with hit pieces in the Wall Street Journal, Science magazine, the New Yorker and the New York Times.  Finally Nature News has a balanced article showing how the shorts have been attacking the company and its drug — https://www.nature.com/articles/d41586-023-00050-z.

I’d written about this before and that post can be found after the ***

The Nature article discusses concerns by Elizabeth McNally editor of the Journal of Clinical Investigation, that journals are being manipulated by short sellers claiming that an article is fraudulent.

“Typically, when a whistle-blower contacts a journal about concerns over manipulated images or otherwise questionable data, the allegations are taken on good faith, McNally told Nature. The idea that whistle-blowers could be doing this for their own financial gain “was very eye-opening to me”, she says.”

One particular criticism of Cassava found in the Nature article is rather amusing. “Amid the allegations about Cassava’s data, researchers have expressed concern over how Simufilam works. Aside from the preliminary studies by Cassava and its collaborators, the strategy of stablilizing filamin-A to tackle Alzheimer’s hasn’t been on anyone’s radar, says George Perry, an Alzheimer’s researcher at the University of Texas at San Antonio. “The fact that it hasn’t been widely studied means that it hasn’t been confirmed.”

The fact that filamin-A hasn’t been on anyone’s radar is actually in its favor, since aBeta, the great white whale of Alzheimer’s research has been impaled with multiple expensive harpoons, with minimal benefit to patients.

The Nature article notes that some of the FDA petitioners wanted the Simulfilam studies stopped, something any drug company with a competing product for Alzheimer’s might wish, but should never ask for.

****

The copy of this post was changed to respond to the valid criticisms of Dr. Elizabeth Bik.

 

Cassava shorts should be worried

Yesterday, 1 November ’22, a blockbuster  article was published in the Journal of Clinical Investigation (JCI) written by its editor Elizabeth McNally — https://www.jci.org/articles/view/166176.

It is just over a year ago since the first of the articles attacking Cassava Sciences appeared.  The first was in the New Yorker which profiled Jordan Thomas as the second coming of Christ for exposing supposed fraudulent data published by Cassava principals —

Radden Keefe P. The Bounty Hunter. The New Yorker. Updated January 17, 2022. Accessed October 11, 2022. https://www.newyorker.com/magazine/2022/01/24/jordan-thomas-army-of-whistle-blowers.

There were similar articles in Science — 2022;377(6604):358–363

and the New York Times https://www.nytimes.com/2022/04/18/health/alzheimers-cassava-simufilam.html.

They relied on the same assertions given to the FDA asking that the clinical trials be stopped because of ‘danger’ to the patients.

It’s worth reading McNally’s article completely.  It isn’t very long.

A few highlights (“the Journal” refers to the JCI)

“Throughout 2022, the Journal has been repeatedly contacted to comment on the 2012 JCI paper. Although we cannot be certain, there now appear to be new “short and distorters.” A recent round of emails was sent simultaneously to multiple journals and editors, identifying 25 articles with potential problems and providing recommendations on how the journals should respond. Importantly, these accusatory emails do not identify any financial conflicts of interest on the part of the whistleblowers. The emails insist that an investigation begin within 24 hours and request that the journals update them on investigative progress. As an editor, I am expressing concern because this represents a new means of manipulating the scientific publishing industry.”

So journal editors are like docs. They talk to each other to find out what’s really going on.  It is likely that McNally called up other journal editors to find out if her experience was common.

Here is why those sending the eMails should not sleep well of a night.

“Last, if the Journal uncovers allegations made for the purposes of stock manipulation, with evidence of misinformation, the JCI may elect to express its concern to the US Securities and Exchange Commission or the Department of Justice.”

It’s about time.

Whether the ‘whistle-blowers’ are guilty of anything will be determined by the suits (from investors losing money on Cassava, or perhaps Cassava itself) which are almost sure to follow.

As some of you know, I think Cassava’s data is even better than they realize. Be warned the following link is long, detailed and will require your concentration  — https://luysii.wordpress.com/2021/08/25/cassava-sciences-9-month-data-is-probably-better-than-they-realize/

Silent synapses

For about the past 20 years we’ve been able to observe dendritic spines forming synapses  in the living (rodent) brain  — for months ! ! In 1970, if you told me that, I’d have said you were smoking something.  The surprising finding is that dendritic spines are a work in progress, being newly formed and removed all the time.  The early literature (e.g. 10 years ago) is contentious about how long a given spine lasts, but most agree that spine plasticity is present every time it’s looked for.  Here are a few references [ Neuron vol. 69 pp. 1039 – 1041 ’11, ibid vol. 49 pp. 780 – 783, 877 – 887 ’06 ].

It is yet another reason why a wiring diagram of the brain wouldn’t help you understand it.   For much more on this please see — https://luysii.wordpress.com/2021/04/25/the-wiring-diagram-of-the-brain-takes-another hit

Not only that, but not all of these new synapses are functional, e.g. stimulating the presynaptic side doesn’t result in a response in the post-synaptic side.  These are the silent synapses. This is thought to be due to a lack of postsynaptic ion channels which can respond to released neurotransmitter.  In particular AMPAR ion channels which respond to glutamic acid are thought to be absent in the silent synapse.  Only after stimulation of NMDAR ion channels (which are thought to be present) are AMPAR ion channels inserted into the postSynaptic membrane converting it to an active synapse.

Obviously, in the fetal brain most synapses are newly formed, hence likely to be silent. It was thought that silent synapses are few and far between in the adult brain.

Not so says Nature vol. 612 pp. 323 – 327 ’22.  They used superResolution protein imaging to study some 2,234 synapses in layer V pyramidal neurons in the adult mouse primary visual cortex (probably the best studied piece of cortex in the brain).   Amazingly some 25% of these synapses lacked AMPARs and were presumably silent.  Most of them were found where you’d expect — at the tips of dendritic filopodia, which are moving around looking to form a new synapse.

If this is generally true of the cerebral cortex, it helps explain our ability to learn.  In this sense the brain is both similar and not similar to neural nets, which learn by increasing or decreasing efficiency (weights) of connections between ‘neurons’ as they are exposed to stimuli with feedback.  The connections (synapses) are fixed in neural nets, but the individual synapses are not fixed in the human brain.  However, if you think of all the connections between two neurons in our brains as a ‘synapse’ then clearly efficiency is clearly being adjusted, as synapses form and die.

111 years of study of the Alzheimer plaque still got it wrong (until now)

The senile plaque of Alzheimer’s disease has been known for 111 years  which is when Alzheimer’s first patient died and he studied her brain. For the past 60 or so years, we’ve studied it using every technique at our disposal.  We know its chemistry fairly  well, and understand many of the mutations that cause the familial forms of Alzheimer’s disease.

However, we’ve still been interpreting its structure incorrectly until this month.  In addition to the amorphous gunk of the plaque, electron microscopy has described swollen ‘dystrophic neurites’ in and surrounding the plaque.  The semantics of neurites implies a small nerve process which led us all down the garden path to assume that they are dendrites (which are usually smaller than axons).  Wrong, wrong, wrong, they are axons as a recent paper proves conclusively [ Nature vol. 612 pp. 328 – 337 ’22 ].

It took a lot of technology to reach this point.  First was development of the 5XFAD mouse which gets plaques galore, because it contains 5 mutations spread over two proteins, the amyloid precursor (APP) protein from whence the aBeta peptide of the senile plaque and PSEN1 a protein which helps to process APP into aBeta.  Second was the ability to observe dendrites and axons in the living (mouse) brain for long periods using specialized microscopic techniques and a variety of dyes and fluorescent proteins.  They allow us to watch action potentials pass along axons without sticking an electrode into them (by measuring rapid changes in local calcium concentration).

Each senile plaque contained hundreds of axons with focal swellings (the dystrophic neurites).  Most were present for months, but some disappeared without axon loss.  When an action potential got to a focal swelling (also known as a spheroid) it slowed down (the swelling acts as a sink for the current  due to its ability to store ions  (higher capacitance).  Random slowing of nerve conduction is murder for information processing.  It’s old technology but just think of what happens when you play  of  a 33 rpm record at 78 rpms.  It’ s also why the random demyelination (which changes action potential velocity)  of nerve fibers in MS raises hob with information transmission hence neurologic function.

Why did electron microscopy miss this?  Because it is just a two dimensional (very thin) slice of dead brain.

The paper has a lot more about what’s in the swelling — large endolysosomal vesicles, and a possible way to treat Alzheimer’s — genetic ablation of phospholipase D3 (PLD3) was able to reduce the average size of the dystrophic neurities and improve axon conduction.

It’s actually a hopeful paper, because we’ve been assuming that the dystrophic neurites were either dead, severed  or nonfunctional, and here they are intact and conducting nerve impulses.

Like all great scientific papers, it raises more questions than it answers.  Is the swelling due to extracellular aBeta?  Is the swelling an attempt to internalize aBeta and destroy it?  Is there a way to inhibit PLD3 ?   Genetic ablation of a gene in a living human is at or beyond our current technology.

Why the results of the open label trial of Simufilam will be misinterpreted

Cassava Sciences said they will release the results of the 200 person open label trial this year.  I think the results are likely to be misunderstood.

First: a disclaimer.  I have no inside information about the results of the trial.  I have known Lindsay Burns since she was in high school, as my wife and I were friendly with her parents when I practiced neurology in Montana.  Lindsay comes from smart .people, father Horatio went through Harvard in under 4 years getting a degree in physics.

Lindsay and I have been contact for over 10 years, mostly concerning the science behind Alzheimer’s disease and Simufilam,.  We do schmooze a bit about Montana and our time there and the vituperation she and Cassava Sciences have been exposed to.

Lindsay and I are well aware about the use of inside information, particularly since both of my wife’s parents had lifelong careers at the SEC, beginning in the depression.

Let’s assume that the results in the open label trial on 200 patients for 1 year are similar to those released in August of 2021 on the first 50 patients in the study to have been on Simulfilam for 9 months.

10% of the patients likely had at least a 50% improvement in their ADAS-Cog-11 score, and over 50% had some improvement.  Some, of course got worse, so the overall improvement of the group  at 9 months as a whole was slight but real.

Let’s say the results on the whole 200 are similar, with 20 or so patients showing similar 50% improvement, but the group overall showing only slight improvement.

These results, even though open label and unblinded, would be unprecedented.  People with Alzheimer’s have good and bad days, but NONE are better after a year.  Throw in all the studies with monoclonal antibodies against aBeta, and you won’t see results like this.  The best they have to offer is a slightly slower (25 – 50%) rate of decline.  This is also true for the demented patients I saw in over 30 years of clinical practice.  So even though the study is open label, we have a ton of controls outside the study.

Even if Simufilam significantly helps 10% of those receiving it, these are results worth having and should lead to early adoption of Simufilam

6 months down the road the Cognition Maintenance Study (which is blinded and placebo controlled) should give a more definitive answer, leading to early adoption.

Here is a link to a more lengthy analysis of the first 50 cases to go 9 months

Cassava Sciences 9 month data is probably better than they realize

Here is a link to a description of the Cognition Maintenance study

Cassava’s Cognition Maintenance Study may prove Simufilam works

What AlphaZero ‘knows’ about Chess

I’m a lousy chess player.  When I was 12, my 7 year old brother beat me regularly.  Fortunately chess ability doesn’t correlate with intelligence, as another pair of brothers will show.   The younger brother could beat the older one at similar ages, but as the years passed, the older brother became a Rhodes scholar, while the severely handicapped younger brother (due to encephalitis at one month of age) is currently living in a group home.

A fascinating paper [ Proc. Natl. Acad. Sci vol. 119 e2206625119 ’22 ] opens the black box of AlphaZero, a neural net which is currently the world champion, to see what it ‘knows’ about chess as it relentlessly plays itself to build up expertise.

The paper is highly technical and I don’t claim to understand all of it (or even most of it), but it’s likely behind a paywall so you’ll have to content yourself with this unless you subscribe ($235/year for the online edition). The first computer chess machines used a bunch of rules developed by expert chess players.  Neural nets require training.  For picture classification they required thousands and thousands of pictures, and feed back about whether they got it right or wrong.  Then the probability of firing between elements of the net (neurons) was adjusted up if the answer was correct, down otherwise.  This is supervised learning.

Game playing machines are unsupervised, they just play thousands and millions of games against themselves (AlphaZero played one million).  Gradually they get better and better, so they beat humans and earlier rule based machines.  A net that has played 32,000 games beats the same net that has played 16,000 games 100 games out of 100 games.  However the 128,000 beats 64,000 only 64 times.

They they had a world chess champion (V.K.) analyze how the machines were playing.

Between 16,000 and 32,000 plays the net began to understand the relative values of the pieces (anything vs. pawns, queen vs. rook etc. etc.)

Between 32,000 and 64,000 king safety appeared

Between 64,000 and 128,00 games which attack was most likely to succeed appeared.

Showing that there is no perfect strategy, separate 1,000,000 runs of the machine settled on two variants of the (extremely popular) Ruy Lopez opening.

They studied recorded human games (between experts or they wouldn’t have been recorded) in the past 500 years.  Initially most people played the same way, with variants appearing as the years passed.  The neural net was just the opposite, trying lots of different things initially and subsequently settling on few approaches.

All in all, a fascinating look inside the black box of a neural net.