Matt Kaeberlein--Two Promising Drugs for Alzheimer’s & Aging (Science Backed)

Let Us Flatter Ourselves
Good video on lithium and rapa. For us the money shot is:

“What I’m trying to say is that these lines between the totally healthy, very rich, high IQ person, I mean these are not random. The people that are taking RAPA off label are not normal people. They’re TORC2 upregulated. Biology, you know, behavior is not random. You see amazing referral bias. Sure. I mean, I would agree with that. I’m interested to know why you say they’re TORC2 upregulated. That’s an that’s a they have a high IQ…Biology, you know, behavior is not random. You see amazing referral bias. Sure. I mean, I would agree with that.”

This clip starts at the “money” shot. (click twice to get to relevant section)

5 Likes

Sadly, I can’t find much data that supports it…

From CGPT5:

Short answer: I can’t find any human study showing that “more intelligent people” have up-regulated mTORC2.

What we do have:

  • In animals, mTORC2 (via Rictor) is required for consolidating long-term memory; deleting Rictor impairs L-LTP and long-term memory, and pharmacologically boosting mTORC2 can enhance memory in mice. That’s mechanism, not a human inter-individual intelligence finding. (DOI)
  • Reviews note that mTORC2’s role in plasticity is real but comparatively less mapped than mTORC1, underscoring gaps in human data. (PMC)
  • Very large human GWAS of intelligence/educational attainment identify thousands of loci, but there isn’t a widely reported, specific association tying higher cognitive ability to up-regulated mTORC2/RICTOR signaling. (These papers don’t measure brain mTORC2 activity; they just map common variants.) (Nature)

Related/nuance:

  • In human disease models, reducing excessive mTORC2 can normalize phenotypes (e.g., PTEN-linked neurodevelopmental abnormalities; seizure susceptibility), so “more mTORC2 = better cognition” is unlikely to be a general rule. (PMC)

Bottom line: there’s solid preclinical evidence that mTORC2 supports memory formation, but no peer-reviewed evidence that people with higher intelligence have measurably up-regulated mTORC2. If you’re exploring this mechanistically, the best proxy readouts are typically AKT Ser473 phosphorylation and RICTOR-dependent signaling in brain tissue—metrics that aren’t available at population scale in living humans. (DOI)

3 Likes

It looked like Matt was a bit surprised and skeptical of that claim as well. The idea that mTORC1 activity correlates with certain aspects of cognition is plausible, but equating it directly with “intelligence” seems to be an oversimplification.

More likely, the guest meant that people who are already more intelligent—or perhaps more resourceful and health-conscious—are the ones who tend to discover and experiment with rapamycin, rather than that mTORC1 activation itself produces intelligence. In other words, it’s more about the pre-existing characteristics of the population drawn to such interventions than the biochemical pathway per se.

This “seeking-out” pattern isn’t new. When Linus Pauling popularized vitamin C megadosing, that movement also drew a self-selecting crowd—often more educated or scientifically curious than average. I’m not aware of any formal studies comparing “smart supplement seekers” across eras, but it would make for an interesting sociological or behavioral topic.

AI-assisted searches suggest supplement use in general skews female, while the “edgier” domains—rapamycin, peptides, experimental longevity drugs—are predominately male. As AI describes that phenotype: higher openness to risk, novelty, and abstract reasoning, but lower conformity and agreeableness..

I still cannot get the transcript from youtube directly, but I have got it via a transcript site and asked Claude 4.5 to tidy it up, produce a summary and critique.

Claude was quite slow in producing the transcript, but that may be because I have switched it into more analysis mode. Also it is long so posted in chunks.

Podcast Transcript Analysis: Lithium, Rapamycin, and Chronic Fatigue Syndrome

TIDIED TRANSCRIPT

Host: All right, John, thanks for coming back down to OptiSpan HQ for the podcast. It’s been a while. I can’t remember when we had our first go-around at this, but a couple of topics I want to dive into with you today. One, the paper that you were a co-author on about rapamycin and chronic fatigue syndrome, which I’ve talked about on the podcast before. And the other, of course, being lithium, which you touched on in our first interview. But a lot has happened in the lithium conversation recently, in particular related to Alzheimer’s disease. So I’m excited to get into it. I think maybe let’s start with lithium.

You’re a psychiatrist by training, so you probably thought a lot about lithium and used it in that context. What first got you to think about lithium beyond the sort of traditional way that it is used in psychiatry?

John: Well, lithium was always a metabolic drug. If you avoid getting lost in terminology, psychiatrists generally think about using uppers and downers. We use uppers—stimulants and antidepressants—for low activity syndromes and hibernation, and we use downers for manic symptoms and psychosis. It’s all cell surface signaling on nerve cells.

Then suddenly we have this very odd drug, lithium, that doesn’t fit within that framework. Lithium has been the graveyard for psychiatry for 20-30 years because we didn’t have a technology to understand what it was really doing. There were hypotheses that it affected glycogen synthase kinase, inositol signaling, and other things, but we just didn’t have a granular technology. Yet when you see lithium work in patients, it’s an amazing drug. You can have extremely ill people be essentially completely normal when you see them every couple of years, and that’s the way it rolls.

Then the data came out in dribs and drabs over the last 10 or 20 years that people on lithium had longer telomeres, had less atrophy on MRI, had decreased death rates, decreased rates of Alzheimer’s disease. It became clear that for all of us who care about biochemistry—biochemistry is life and life is biochemistry—you can’t disentangle the two. So I made it a real effort to try to go into the source code and understand lithium. I went a lot to the mitochondrial conferences and did a lot of my own clinical work looking at mitochondrial measures.

Very recently, the field has shifted. The Nature paper just came out maybe a month ago where a very well-funded Harvard group looking at a well-accepted model of Alzheimer’s showed that changes in lithium concentration dramatically reduced markers of Alzheimer’s and possibly changed the inflammatory response to amyloid and tau tangles. That’s fantastic.

Host: That was in mice, but they also had human data, right? In human brains of people with Alzheimer’s disease, you saw lithium depletion in certain regions and binding by amyloid. This is really a major breakthrough, right?

John: It’s a breakthrough because there’s this lack of interest in a generic drug that costs 40 bucks a year. I’ve talked with the drug reps about these treatments for Alzheimer’s disease—they’re $70,000 a year. They require multiple MRIs, they require a specialty team. The waiting list at Johns Hopkins is reportedly two years. The waiting list here at the UW to see a specialty neurologist in Alzheimer’s is six months.

Now we have interest that is perceived as legitimate and not fringe—that lithium is worth looking at again. That’s a major advance.

Host: It sounds like from your perspective, you’re not surprised by this, but probably pleased to see the attention.

John: I need all the help I can get. There’s no ego involved in just getting your patients better.

Host: We’ve spent billions of dollars on Alzheimer’s drugs, most of which have failed. We have a few now that are approved that don’t work particularly well, but at least they have some impact. Do you think lithium could actually succeed where these drugs have all failed?

John: We talked last time about this new biomarker, the ATN profile, where you’re looking at proteins that leak out of a stressed brain that has Alzheimer’s. I was very excited to have those proteins available because it allowed me to manage the patients that I’ve seen for 20 to 25 years and be with them at the next stage of their life.

The Chinese have some data out looking at these markers, and they estimate that about 25% of people between the age of 72 and 77 are going to start having personality change associated with Alzheimer’s, and that may not necessarily be cognitive impairment. Alzheimer’s is not a disease just of learning and memory—it’s of mood. One of the first signs of Alzheimer’s is actually anxiety.

We’re left in this terrible world where 25% of the population is going to be in psychic pain because of Alzheimer’s even if they don’t have cognitive impairment.

Host: I guess this is where the definition of Alzheimer’s gets murky as well, right? Is it necessarily the case that all of those people will go on to develop at least classical Alzheimer’s as it’s been defined historically, or is this really a whole spectrum of conditions?

John: Everything in biology is one web with no cut-off points. This is the problem in medicine when you recruit individuals who are good at memorization. Medicine is enriched for highly autistic people that like clear cut points. Biology laughs at that. Biology is a continuum.

What I’m seeing in my clinic now is that when you look for people above the age of 65, you can see stress markers very commonly in these patients. For some of those patients, if you give them lithium, my God, they feel better. Sometimes their Alzheimer’s markers get better within months.

Host: By Alzheimer’s markers, you mean in particular the ATN profile?

John: Yeah. Amyloid, tau, and neurofilament light (NFL).

Host: The question which is incredibly relevant is: what’s the dose?

John: There’s this sort of fantasy that there’s going to be some large study done by Harvard or done by a multi-center group, and it’s going to give us a protocol and it’s going to say you need this dose, and that’s the optimal dose to prevent Alzheimer’s. Biology laughs at that.

Host: Let me ask you a question. The study in Nature landed on lithium orotate as a particular salt of lithium that was resistant to the binding by amyloid. Is that does that line up with your experience, or does it matter, or maybe the dose is dependent on the salt of lithium?

John: I want to go really broad. I think it’s really important to think about the issue of locality. It’s maybe the difference between Newtonian physics and quantum mechanics. In Newtonian physics, you see these large objects and you have very clear equations that tell you exactly what’s going to happen. Quantum mechanics is fast time scales, very small objects, and it’s even worse—it’s a probability distribution.

What’s really great about this Nature paper is that it brings up the concept of locality—that there’s a distinction in medicine between endocrine signaling, where you have a chemical like thyroid hormone or insulin that affects every cell in the body at the same time, versus paracrine signaling, where there’s a chemical that’s excreted that is only meant to have a local effect.

They bring up this concept of paracrine signaling and locality in terms of the activation in the microglia and the activation of the synapses that are associated with Alzheimer’s. If you just look at CSF or if you just look at blood, are you looking at a global variable and you’re not really able to see the local signaling?

Host: In your experience, in your patients, do you use lithium orotate as a preferred form, and do you see any difference in efficacy?

John: Let’s look at their data. When you look at their data, they make an argument that the dissociation constant is different between an inorganic ion and an organic ion. On the market we have lithium citrate, which is available and essentially equivalent to orotate in terms of their raw data in terms of dissociation constants. I have not yet compared and contrasted the differences between an inorganic ion, which is typically lithium carbonate, with the citrate.

But there’s this kind of odd physical chemistry. In their paper they say, “Well, psychiatry typically produces lithium levels that are a thousand times the physiological level.” So they’re arguing the dissociation constant—maybe a hundred times different with the carbonate versus the orotate. If you increase the reactant concentration high enough, it doesn’t really matter whether it’s carbonate—you can overcome it.

Do you get more off-target side effects? Again, dose finding, dose finding, dose finding. How much available lithium is there in the microenvironment? Do you need 450 mg of lithium carbonate and 150 mg of lithium citrate? This is animal data. It’s a great idea, but until you hit the clinic, you don’t know anything.

Host: It sounds like maybe we just don’t have the data yet.

John: This paper just dropped. I don’t know enough about clinical practice to know whether lithium orotate or lithium citrate were commonly used. Orotate is usually used by the naturopaths because they’re probably prescribing much lower doses—40 to 50 mg, that kind of stuff.

I had an interesting conversation that your listeners might find thought-provoking. Somebody watched our podcast last year and they’re in Belgium and they wanted me to advise on the use of lithium. I can’t manage a patient in Europe for all sorts of reasons, but I talked to one of their representatives in England who was essentially a primary care doctor. She’s like, “Oh, it’s illegal for me to prescribe lithium in England.” It’s only allowed with psychiatrists. Lithium carbonate is illegal if you’re not a psychiatrist, but if it’s orotate, you can prescribe it.

Host: That’s crazy. I was just in Germany and I asked about lithium. There, they can prescribe it, but you can’t buy it over the counter like you can here. You can buy lithium orotate on Amazon in the U.S.

John: When lithium first came out in Denmark, it was actually illegal because they didn’t know how to dose it. This was maybe in the '20s or the '30s. Then I think Mogens Schou figured out you just need to dose a little lower. This whole issue of dose finding is literally a hundred years old with all sorts of interesting regulatory starts and stops. It’s just human frailty at large.

Host: You mentioned the ATN profile. I wanted to dive a little bit deeper on that. Is it your experience that you feel like you see consistent changes in that biomarker with lithium, or at least in some people you see changes in ATN?

John: I think the problem is that I’m so busy and I’m such a generalist. I’m doing this to some degree because these are people I’ve worked with for 20-25 years. I have a responsibility for them, and I’m not going to pretend that there’s nothing to do and send them off to Harvard as if Harvard’s going to save them. It won’t. I know better.

What we see in our patients is that we see many people that have elevations of the amyloid or they have elevations of phosphorylated tau, and they don’t yet have cell death. So they don’t have neurofilament light—they don’t have fragments of the axons in their blood. I think it’s going to turn out that if you don’t have elevations of NFL, you tend not to have atrophy on MRI. That’s a really lovely place to be. Theoretically, your disease is reversible because there’s no cell death. Nobody wants to get Alzheimer’s and nobody wants to get older, but to know that you have time—that’s such a gift.

Then the Harvard paper comes out and suddenly it’s no longer wackiness to try that bipolar drug that crazy people take.

Host: Is it your guess that the people who are in that sort of stage—prior to significant neuron loss—they’re the ones who are going to get the best outcome from something like lithium?

John: Well, once cells are dead, they’re dead. You’re not going to bring those back, and you have a damaged brain. It doesn’t repair quite as well as something like stroke. You see this in clinical practice all the time—quite significant strokes, but if the general brain is healthy, you rewire very well. But you’ve got an Alzheimer’s brain and you get cell death—boy, that’s a tough road to go. You’ve got to really reverse the underlying process. Obviously you want to catch this early.

Host: Given what at least what we think we understand about how lithium is probably working here, it’s going to be much more effective prior to significant neuron loss.

John: We just want to fix the underlying problem. If people are having stress markers leaking out of their brain into their blood more than the 95th percentile, that’s a problem. Do I see a reasonable group of people try lithium and they’re on lithium and they sure feel better and they tell you right away? They’ll say, “I’m thinking clearer. I handle stress better,” and they don’t have bipolar. These are not bipolar patients, but they’re on lithium. It’s health span rather than necessarily lifespan.

Is something like the ATN a marker of somebody who is like, “Yeah, maybe you should try a little bit of lithium and see whether you just feel better”? Now, if the markers reverse, that’s a plus. But we haven’t had time yet to do the kind of studies that say, “What if they just stop?” That’s still a plus, right? If you don’t get cell death and they don’t progress and you’re feeling better on a drug that costs 40 bucks a year that’s been used for more than a hundred years, my God, that’s a win.

That’s what the Harvard paper is telling us—that there may be a $20 bill that’s just on the ground that anybody can pick up who has some reasonable degree of medical skill.

Host: When you and I talked a year ago, I read in your voice like, “Oh, this is kind of an unusual claim that lithium is potentially one of the best anti-aging drugs.” Things have changed, certainly for Alzheimer’s. We could talk about the anti-aging potential effects of lithium as well. Alzheimer’s is one age-related disease, but I think there’s reason to think lithium can go beyond Alzheimer’s.

Although, if we look specifically at the Nature paper, at least the mechanism they propose seems to be very specific for Alzheimer’s disease in the sense of amyloid depleting lithium in specific brain regions. Again, could be a lot more going on than what they proposed in that paper, but that was the mechanism that they landed on.

John: There was a couple things that really bothered me about that paper. Number one, when you looked at their pathway analysis in their supplement, they talked a lot about the Wnt pathway or the wingless pathway, which implies that when you block glycogen synthase kinase-3 beta, which is a downstream product to mTORC2, you’re making cells less differentiated. I think that fits with a lot of what we know in terms of how lithium works in bipolar patients—that it differentiates the astroglia in bipolar patients, so it becomes a speed governor on excitatory nerve cells.

The other major pathway is autophagy. Autophagy was massively upregulated with lithium, and they don’t even mention that in the paper. I thought that’s really strange because autophagy is everybody’s—it’s like motherhood. Everybody’s into autophagy, and then they left it out of the paper.

Host: Well, maybe they’ll watch the podcast and respond. I mean, when you do multi-omics, it’s really hard to see the forest for the trees. Often times the reality is to get your paper into a top tier journal like Nature or Science, you really do have to pick one relatively simple story and stick to it.

John: That’s sort of what I was getting at—I think there’s a lot going on with lithium. I think this model of amyloid beta depleting lithium in specific brain regions makes sense. It fits the data, but there’s probably a lot more to it. Fifty percent of Alzheimer’s genes are microglia, but I’m not sure that’s the proximal cause. I think the proximal cause is the excitatory nerve cells.

Host: I just had a really interesting podcast conversation with Corey Funk from Institute for Systems Biology. They’re doing a lot of work as well on cholesterol metabolism and membrane fluidity as being a key driver of at least early components of Alzheimer’s disease as well.

John: Yes, I think that’s right. The data on lithium impacting age-related processes independent of Alzheimer’s disease is also pretty interesting. There could be multiple mechanisms at play here. It’s an mTORC2—it’s a downstream mTORC2 inhibitor. It’s going to affect every cell in the body in different ways.

Host: There could be multiple ways. Why would we just focus on microglia?

John: We have another technology that’s out that’s commercially available that supplements the lithium story. We can now do a commercial RNA—microRNA—assay of about 400-500 species. The assay generates a vector of RNA abundance in blood, and the company has a very deep phenotype matrix that they’ve gathered over 5-10 years, 2,000-3,000 people.

What they’re seeing is a really large signal for lithium in a variety of psychiatric conditions and also in the aging area as well. We were stuck with proteomics that had low dimensionality—amyloid and tau and NFL. Now we get to blow up the dimensionality a little bit more. It’s a large enough dimensionality you can do gene set enrichment.

Host: So this is data from a variety of individuals, some of whom have been taking lithium. Is that what you mean?

John: No, let me go over the protocol. This was a group out of Indiana that followed let’s say 2,000 people, and they did serial microRNA estimates in these patients. They looked at things like suicide attempts and memory loss with age and divergence from biological age versus the aging clock stuff, hospitalization for bipolar disorder, chronic pain measures, ER visits for chronic pain.

Now they have an AI that has a large enough data set that you can start to do cluster analysis and you can cluster phenotypes. But you can also break the AI and look at each RNA species and say, “Hey, what drugs out of the 2,200 FDA approved drugs are going to move this vector towards the origin?”

I’ve done about 20 of these so far, and what just comes up over and over and over again is lithium. This is a prediction based on what we know about lithium that it would move the gene expression profile in the direction of health.

I think what they’re showing is that there’s a group of humans that have a gain of function in mTORC2. We’re all sort of comfortable with that in the rapamycin world—there’s a whole bunch of diseases where mTORC1 is upregulated, mitochondrial disorders and gain of function in mTORC1. But I’m looking at this data and saying there’s an mTORC2 gain of function that is part of the human condition, and lithium is a targeted drug across so many phenotypes because we really are hitting source code.

Host: That’s interesting. As you know, a lot less has been done on mTORC2 compared to mTORC1 in terms of understanding the biology of that complex.

John: I was at the aging conference and I was talking to some people about, “Hey, I need to know more about mTORC2.” They said, “Oh, John, you should go to a Gordon conference.” I’m like, “I’m a psychiatrist seeing patients.”

Host: Okay, so let’s go back kind of big picture with Alzheimer’s disease. Best case scenario for lithium, how much do you think it will move the needle for Alzheimer’s disease? Are we talking a year, five years, complete cure? What’s your best guess?

John: I think thinking in aggregates is just so dangerous. I have no doubt there is someone out there on lithium who is going to live 10 years longer because they’re on lithium. It’s an mTORC2 gain of function—you downregulate mTORC2 and the cells don’t go into senescence as quickly, and they die more cleanly. They die with apoptosis rather than screaming with their head on the spike because they’re overregulated.

I think most—if you really look granularly in the psychiatry literature, you would see this data—that there are certain bipolar patients you put on lithium and their lifespan goes up a lot. Not even their health span, but just the brain works better. They exercise more, they drive more safely. It’s not suicide attempts—it’s just the brain acts on the environment and they have a brain that works.

Are there people that lithium has absolutely nothing to do with Alzheimer’s? 100%. I have some people that have markers and I give them this psychiatrically therapeutic, homeopathic dose—75 milligrams—and that’s all they can tolerate. Now, they’re still better, but it’s not anywhere near the typical psychiatric dosing.

I just can’t say it often enough: there isn’t a right dose and a right drug for aggregate populations. This is what you’re doing at Ora Biomedical—saying look, it’s not one drug, it’s a unique cocktail for each individual person.

Host: Do you think in the case of Alzheimer’s that’s because Alzheimer’s is a collection of many different diseases and lithium only works for some of them, or at least the cause related to mTOR2 is different in different people?

John: I think what it is—and I may even be deluding myself—a patient with Alzheimer’s disease is in a complicated n-dimensional biochemical space. That space is not linear. It has ridges and valleys, and none of our drugs are moving patients to the origin in a linear way. We’re in this very complicated nonlinear world.

It’s not that it’s a collection of diseases—it’s that our tools are not linear tools. When we combine them, we may be in a nonlinear world which I think biologists have to get their head around. I think the AI people understand that better than biologists do. The AI guys are like, “Hey look, if I don’t design my GPT or my ChatGPT somewhere, I’m going to hallucinate.” Whereas in biology, even the best people are just doing multivariate discriminant analysis and they just assume linearity, and that’s a mistake.

Host: What about side effects? I think lithium scares a lot of people. What are your thoughts there in terms of how concerned should people be—thyroid, kidney—at the doses that might be effective for Alzheimer’s?

John: Let’s talk about clinically who does poorly with lithium and what does it mean to do poorly. There was a study in the early '70s looking at lithium and temporal lobe epilepsy, and they had to stop the study because patients became too sick. There is a seizure model of lithium and pilocarpine exposure in rats where you can cause seizures.

In the temporal lobe epileptics, if you drive growth, you push toward the mTOR distribution—you dedifferentiate cells. Do you drive excessive cell growth in mesial temporal sclerosis and the temporal lobe epilepsy syndrome? There’s a problem differentiating cells in some patients that is related to excessive nerve biogenesis.

This is just to slap someone out of this idea—“Oh, it causes neurogenesis. That’s a good thing.” That’s really a dumb comment. Brains are—nerve cells are postmitotic for a reason. There’s only a few areas of the brain that have new nerve cells, the temporal lobe and the olfactory lobe, olfactory bulb. And those are the most epileptogenic regions in the brain. It’s not always obvious that you want to grow new nerve cells. That’s a problem in some people, and it can take months for seizure-like events from lithium to develop. It takes a lot of technical skill to pick that up.

Host: How frequent is that in people, would you say? These are high doses, right? Do we have any data on the lower doses of lithium?

John: Again, it’s a dimension, and the phenotypes are subtle. I once had a guy—he was in the prison, he was on lithium, and he was getting a great response to it. He says, “John, my nose is dripping on one side and I feel a little confused.” It turns out that in temporal lobe epilepsy monitoring units when someone’s having a seizure, you get neurogenic rhinorrhea on one side, ipsilateral to the seizure, and you can tell which side the seizure is on by just which side of the nose the patient wipes their nose.

So temporal epilepsy is a problem—this issue of excessive neurogenesis. But again, is that going to be primarily at the high psychiatric doses?

Host: We don’t know, right? It’s a continuum.

John: The other issue is in the migraine patients. Lithium with migraine is really tough. We had a real fantastic migraine specialist who’s retired since now who was just like, “Oh, John, please never give my patients lithium,” because not always, but often you make migraine worse.

Host: What’s the cause of that? Is that differentiation of the astroglia with fuel supply to the brain? Is it excessive autophagy?

John: That’s another interesting article because there is some cellular data that suggests that mitophagy and autophagy are distinct, and if you get too much autophagy, you drop the mitochondria below this energetic threshold, and that may be one of the lithium toxicities that happens.

Those are the two—you’ve got to be really careful with those syndromes. But migraine is 25% of women have migraine. One percent of the population has epilepsy. Ten percent of the population will have an epileptic seizure at some point in their life.

Host: I didn’t realize it was that high.

John: It’s high. And what if it takes four or five years? You’re just on the wrong gradient.

Host: So based on that, do you have concerns with the fact that people are going out and buying lithium on Amazon and taking it based on this Nature paper? Or do you think that the doses are low enough most people are not going to experience these kinds of side effects?

John: Saying I have concerns implies that I have some judgment on anybody. I don’t want to have judgment on anybody. This is an illness that’s going to kill a lot of people and it’s going to steal a lot of joy. You can’t find a specialist because the waiting list is two years, and there aren’t a lot of people like me.

I think we have to let a thousand flowers bloom and just do our best to have everybody work together and exchange information and try to learn in a logical way.

Host: Let me maybe frame it a different way. If people want to start taking lithium based on what they’ve heard and the latest research, what should they be paying attention to as potential side effects that might be worth watching? Let’s assume they don’t have access to you or somebody with your experience who can give them medical care. They’re buying it on Amazon. What should they pay attention to?

John: We’re in a desert. We’re in a specialist desert and people have to make their own decisions because they don’t have a choice. England makes it illegal—crazy stuff.

I think what we see in psychiatry with people that have bipolar disorder is you increase the dose until the risk becomes greater than the benefit. In general, the more lithium—this is well known—fewer hospitalizations, fewer suicide attempts, less mood swings. But once you get a blood level higher than 0.7 in a population of bipolar patients—and they’re not average, they have upregulated mTORC2—but the cut point is around 0.7 where you start getting tremor.

You start getting urinary—decreased urine concentration. You start getting brain fog which is different than Alzheimer’s, just sort of a fuzziness. Low-grade migraine symptoms are very common—just sort of photosensitive and kind of this soft cephalic pressure, maybe a little nausea. That’s kind of chronic daily headache, not migraine, but it’s definitely lithium-induced migraine.

In terms of thyroid, lithium is a blocker of the final step of adding iodine to the thyroxine molecule. So it’s a synthetic block. It doesn’t damage your thyroid—you just add thyroid back in. A medical student can do that. It’s a real side effect, but it’s not difficult to treat.

The kidney issue is probably again the “wet roads cause rain” story where the rates of spontaneous kidney failure in bipolar patients are elevated. When you stratify kidney failure in bipolar patients, lithium doesn’t have a higher rate of kidney failure than antipsychotics. This is classic age-related kidney failure, which actually correlates very highly in community samples with neurofilament light. That’s a $20 bill where we need to get our nephrologists in touch with our neurologists and really look at that population carefully.

Host: And directionality there could go either way or both ways, right?

John: Both ways. I’m assuming it’s a shared causality.

Host: What do you think is reasonable?

John: I think it’s reasonable if a primary care doc who has been managing illness for 20 or 30 years and has a generalist intuition and you want to work with a primary care doc—there’s nobody else out there, you don’t really have any choices—I don’t think anyone has the moral authority to say, “Oh, you shouldn’t do this.” But you want to have some sort of experienced sage advice, who can just say, “Hey, maybe you’re pushing this a little too hard.”

Host: Anything we haven’t talked about with lithium that you want to dive into before we shift gears to rapamycin?

John: No, I think we’re good.


RAPAMYCIN AND CHRONIC FATIGUE SYNDROME

Host: So, I want to talk about this chronic fatigue-rapamycin connection. I know you’re a co-author on the—I don’t think it’s been peer review published yet, but it’s in proofs.

John: It’s in. I just got an email about the proofs two days ago.

Host: Great. So, this first came across my radar as a pre-print. My recollection of this study—and obviously I’d like you to tell me the details because you were involved in the study—the take-home being that in a population of people with chronic fatigue syndrome who were being given rapamycin off label, a significant fraction—something like 75% of them—showed improvements in symptomology, which is striking. That seemed to be enriched for people whose chronic fatigue was associated with a prior viral infection, which I also think is pretty interesting given rapamycin’s connection to antiviral gene expression and chronic inflammation.

Then the other part of the paper that I thought was interesting was what was not seen, which was changes in glucose, insulin, lipids, which are side effects that people talk a lot about with rapamycin but that typically I think the data don’t actually show as being at least significant side effects across a population.

John: Great jumping off point. This paper is really a great paper because the core argument is 100% true. There’s a group of individuals who have a hyperinflammatory response to foreign RNA or DNA. It doesn’t matter whether it’s COVID, it doesn’t matter whether it’s flu, it doesn’t matter whether it’s Moderna or Pfizer. It doesn’t matter whether it’s a defect in the lamin A nuclear membrane. Too much DNA and RNA when exposed to certain individuals activates the STING pathway and other pathways, and you get a hyperinflammatory response.

Host: Let me just clarify—make sure I understood you correctly. When you said doesn’t matter if it’s Moderna or Pfizer—some of these people can be responding to the vaccine itself, the RNA, rather than the virus that the vaccine is protecting against, right?

John: Which is why—neither one of us are anti-vax. Vaccines work for what they’re designed to do. I don’t give a—but let’s be clear, vaccines usually work for what they’re designed to do, and they can in some people have unanticipated hyperinflammatory consequences. I just want to make sure that I understood you correctly that the vaccine itself could in some cases be the cause of the inflammatory—

Again, it’s referral bias. What you see as a clinician depends on the patients you see. If you’re a pediatrician, to some degree kids don’t die and you see a lot of ear infections. If you’re a primary care doc in today’s world, you see a lot of type 2 diabetes. If you’re a psychiatrist who has an interest in treatment-resistant depression and chronic pain, I saw a lot of vaccine injury.

That doesn’t mean that I want to write public policy. I’m not about that.

Host: No, and I wasn’t suggesting that people should not get vaccinated for the things where vaccines are quite useful. I’m just trying to make sure I understand your take on the biology and that we are being clear with the variety of things that can cause these types of hyperinflammatory—

John: It’s looking like there’s a group of patients—maybe that have a gain of function in mTORC1—and it’s part of the phenotype if they’re hyperinflammatory when they are exposed to things that activate the innate immune system.

Host: It doesn’t even have to be viral in nature, right? It could be bacterial from the gut, it could be mitochondrial damage and mitochondrial leakage, it could be a defect in the nuclear membrane and you’re having DNA leaking out. There are multiple causes of this inflammatory response.

John: The thing I hadn’t appreciated before reading this paper is in chronic fatigue, the high fraction of people whose chronic fatigue can be associated at least with some sort of severe viral infection preceding it.

Host: Yes. So this is a terribly interesting question: what are the death components in lymphoid cells and myeloid cells? In other words, you have an infection, a cell becomes activated, and then there should be a dampening of the inflammatory response that turns off the effector components in the lymphoid compartment and maybe to some degree in the myeloid compartment.

What are those dampening components? One of them is a global signal which is steroid from the adrenals. Everybody knows—it took us too long to figure this out in COVID—if you’re an 85-year-old and you get COVID, after two to three days, you need a round of steroids because you’re not dying from COVID, you’re dying from the hyperinflammatory response. Every internist in the world knows this—if you have a patient who’s septic, you need to give them steroids sometimes, otherwise their blood pressure goes to zero and they die.

So steroids are one of the things that turn off the hyperinflammatory immune system. The other thing that turns off the hyperinflammatory immune system is the T-regulatory lymphocyte, and rapamycin massively upregulates T-regs relative to the effector cells. This has been seen in essentially every autoimmune study that’s ever looked at it.

Host: Maybe less so in neurological disorders that are related to mitochondrial metabolism because the T-regs use mitochondrial metabolism rather than glycolysis?

John: Right. Are we looking at not so much of a defect in steroid production, although we could be, but are we looking at a defect—a gain of function—in terms of too much effector cell relative to the T-reg? The T-reg is an mTORC2 cell. It’s hyper-specialized. You’re stuck in this mTORC1 world and you need to get shoved into the specialized world which downregulates the inflammatory response.

This has been seen in lupus. When you use rapamycin in lupus—and one of the causes of lupus is just constant DNA exposure to the STING pathway in the cytosol—you can see remarkable improvement in lupus.

What is this study saying to us? Very simply, they’re saying we’re looking at people that look inflamed. The etiology in this case, they’re kind of easy to pick out because they’re a stair-step drop of function—somewhat normal, they get a virus, they never recover. Do they have clonal selection for effector T-cells and effector myeloid cells that are sort of persisting in a sort of benign cancerous state? If we pulse with rapamycin or if we give rapamycin, can we change the equilibrium distribution from an effector immune state to a tolerant immune state?

That’s not a very ambitious hypothesis when you really think about it. It’s kind of like, “Well, yeah, no shit.” It’s kind of crazy that it’s taken this long for this study to be done.

Host: Yeah.

John: What did they say? They saw this whole broad phenotype get better. Fatigue got better. Pain got better. Non-refreshing sleep got better. Orthostatic intolerance got better, and it got better at a very low dose. I want to come back to that because that’s really important, and that’s one of the flaws of the study—the dosing is flawed.

When they looked at biomarkers—and they didn’t have any money. I talked with the PI quite a bit, and he’s like startup central.

Host: Well, again, this is the problem with why we don’t have more data on rapamycin at this point—no company’s going to invest in it. I shouldn’t say no company, but few are. And for whatever reason, grant funders haven’t prioritized these kinds of studies.

John: I’m trying to be fair to the grant funder. It’s just do you write a little check or a big check? You don’t know. You have to give them a startup project and show whether they can deliver before you write a big check.

Host: Maybe, although I would suggest the data on rapamycin is so deep and so broad that there is no question that these kinds of studies should be funded.

John: There and they’re probably good ideas. You and I are alike in that way. We come up with all sorts of good ideas.

Host: Well, no, but I never get funded.

John: No, I get it. But see, the difference is we’re not just reading papers. You’ve done your animal work and I’ve done my human work. We don’t care about the guidelines anymore because we know that they’re based on nothing.

So one of the things that they showed is that serum markers of autophagy tended to correlate with response. Again, I think that speaks to the limitations of their funding. They didn’t have money to look at a lot of biomarkers. They looked at one marker which is autophagy, and they said, “Hey look, when you increase autophagy in blood, people tend to get better.” And then again, throwing us backward to the Nature paper with lithium—that autophagy story.

Host: Yeah. It would be interesting—if you want to speculate—would lithium work in this context?

John: Lithium never works in this context.

Host: Interesting. Why?

John: Observationally, lithium doesn’t reverse pain in the chronic pain patient. It may help with the agitation or the anxiety, but I’ve never seen it work. It makes migraine worse. Psychiatrists don’t prescribe lithium for pain, and rheumatologists don’t prescribe lithium for pain. It just doesn’t seem to work.

Host: There’s interesting biology there. We don’t understand. There’s some clues here that are worth looking at.

John: That’s all the money they had was to look at autophagy. And then again, I want to go back to this concept of locality. Whenever you do blood work, you’re looking at non-local. How much of it are you actually looking at the underlying disease process?

Blood is coming like Athena, fresh out of Zeus’s head, glowing and sparkling with beautiful armor right out of the bone marrow. It’s not really in an mTORC2 world. It’s almost a stem cell. So only a small fraction of the cells in the blood should actually be differentiated. Most of them are in stem mode. When you do these sorts of omnibus metabolomics in blood, how much is that representing the local tissue that you’re really interested in?

Host: Do we know in chronic fatigue what the local tissue is that you would want to look at?

John: Biology is like one giant conversation. There are myokines from the muscle that are talking to the brain and to the liver. It’s just one big web. This particular condition—it’s hard to know where if you had all the money in the world, it’s non-obvious to me where you would specifically look. Obviously you could do a lot more in terms of biomarkers and assessments, and you could do much deeper immune phenotyping other than the immune aging hypothesis, which I think all of us kind of in the area like—there’s something to this. The immune cells are particularly sensitive, and that seems to be the case.

Host: So let’s talk about dose because you said the dosing was flawed in this study. What did you mean by that?

John: I want to make an argument to the aging community to get out of their silo a little bit. We have this amazing amount of data that we got with death and pain and blood and suffering, which is: what is the dose of rapamycin that is necessary to prevent graft versus host disease?

That information didn’t come for free. This has been a lot of hard work with patients and doctors over the last 10 to 20 years. If you look at the population estimates of what it takes to prevent graft versus host disease, it ranges between 3 to 20 nanograms per ml—trough levels. That’s a lot.

Let’s think about what is graft versus host disease. Let’s say you have leukemia and someone kills your bone marrow with radiation and a cytotoxic drug, and they give you somebody else’s bone marrow. That bone marrow is going to kill you because it looks at your body as a cancer, as a virus, or DNA that should be killed.

Transplant doctors have learned that you can’t control this graft versus host disease with low doses of rapamycin. You often need high doses. You often need another drug.

Host: They’re almost always taking other drugs in addition to rapamycin.

John: We should talk—we don’t have time—but they’ve learned that there’s a distribution of hyperinflammatory response in the population that is defined by the dose level of 3 to 20. We know this—that’s the population distribution.

Most of these aging studies are done with one milligram a day. That is not representative of the population hyperinflammatory distribution.

Host: Let me comment. First of all, most of the aging studies—I mean there aren’t very many to begin with in humans—and they either fall into one of two camps. They either do the one milligram a day, which to me is a dumb dose for a variety of reasons, or they do something like six milligrams a week, which is I think what was done in this study if I remember correctly.

John: They were nice to me. I said I don’t want to do pulsing. I did once a day, but I think the rest of the people did six milligrams pulsing.

Host: There’s a reason for that in the aging literature—that’s Joan Mannick’s work with vaccine response. I know you know that. All of that’s just to say there is some rationale for that once a week dosing.

But the point I wanted to make is I think the use of rapamycin for something like health span extension in people who are not chronically ill—and we could debate whether aging is a chronic illness, but let’s just leave that for now. They don’t have an existing chronic disease. The risk tolerance of course is going to be much lower than in somebody whose quality of life is dramatically impaired by something like chronic fatigue syndrome or who’s going to die if they don’t get an organ transplant rejection. So there is a pragmatic piece here to the dosing.

John: I understand that. This is not meant to criticize the aging people. What I’m trying to say is that these lines between the totally healthy, very rich, high IQ person—these are not random. The people that are taking rapamycin off label are not normal people. They’re mTORC2 upregulated.

Host: I mean, I would agree with that. I’m interested to know why you say they’re mTORC2 upregulated.

John: They have a high IQ. In order to have a high IQ, you must upregulate mTORC2.

Host: Okay. This is going to—we definitely should have a conversation about that. I’m intrigued.

But I think the point that I’m making is that we’re all just scrambling for clues about how to use rapamycin at the right dose for the right patient.

John: I agree.

Host: Lack of biomarkers is one of the challenges among many.

John: So what do we see in the transplant world? What we see in the transplant world is that in order to keep people from getting graft versus host disease, you generally need to be running people two, three, four milligrams daily with other agents.

Host: Daily. Yes. And that’s because they have a foreign bone marrow that will kill them.

John: If you have an autoimmune disease, you have a foreign bone marrow that will kill you. You have a subpopulation of a myeloid or a lymphoid population that thinks you are the cancer rather than yourself.

Host: Yeah.

John: So the mapping between graft versus host disease and classical autoimmune disease is much tighter than these categories suggest.

Host: And that makes sense.

John: Right. So when you have a very sick population—and these people are very sick—is one milligram—it’s a pilot study, I get that—but is one milligram too low? Are we going to see better outcomes with dosages that are more typically used in hyperinflammatory syndromes?

Host: And no one is doing this, right?

John: Because the rheumatology world is entirely funded by very expensive biologics—$3,000 to $4,000 a month with lots of marketing, lots of junior faculty members writing guidelines and professional societies that are co-opted by the pharma money.

Host: Let me also suggest that one of the reasons people aren’t doing these kinds of studies is because when you go to FDA to get approval for your clinical trial, they will push back at using rapamycin at organ transplant doses because of their concerns about safety. I know this because I’ve been involved in clinical trials that have tried to do even lower.

So I mean there are pragmatic reasons as well which—right or wrong—I don’t disagree with you. I think for certain conditions organ transplant dosing is going to give a better efficacy signal than one milligram a day or six milligrams a week.

John: Well, I know it’s the case because I’ve seen it. I have people that simply don’t respond at one milligram.

We’ll get to the Joan Mannick argument here in a second because it’s important, and the side effects go up. So again, this is all about balancing risk and reward. Of course. But remember the mortality rate in lupus is 25%.

Host: So condition matters. Endpoint matters. I agree.

John: Yes. We’re—I mean I know you and I agree—but I just want to be clear about why it’s not obvious that for everybody going to four milligrams of rapamycin a day is the way to go.

Host: Of course not.

John: Of course not. But what happens I think sometimes is you get caught in the blinders of just like, “Well, I don’t want to go higher than 1 milligram,” and what are we seeing in general in the longevity literature? Six milligrams a week or that dose—generally it’s pretty safe. We’re not seeing a lot of problems.

And yet if you try to do a clinical trial of rapamycin at six milligrams per week, FDA is going to want you to do a dose escalation where you start at one milligram a week, two milligrams a week, four milligrams a week. I know because Joan Mannick had to do—

Host: As stupid as that sounds, right?

John: Yeah, it’s ridiculous. It really is ridiculous. There’s a couple of branch points from that. Number one, in the world of small molecules and rapid development, the FDA is gone. Does it die in 5 years? Does it die in 10 years? Does it die in 15 years? It can’t live in the world of multivariate biochemical space. It just can’t.

We’re going to have an AI. I am using an AI right now. I have an AI that says you should use this combination of drugs. It’s a terrible AI. It’s got problems—and that’s no criticism of the founders, it’s hard. But once we have an AI where patients are located in an n-dimensional space, what do we need the FDA for? They’re useless. They don’t help us. They have no value, and we want to invent a tool. We don’t need the FDA.

That’s why we must—I just keep preaching this—if we don’t get rapidly into the multivariate world, we’re going to be stuck with these ridiculous FDA guidance and we can’t do it.

Host: I don’t want to get too far down a rabbit hole on FDA. I’m not quite so confident that FDA is not serving a useful or couldn’t serve a useful purpose in that world, and that we don’t need somebody to create guard rails on companies, entities that want to develop new medications but may be more motivated by profit than by safety, which is just a reality.

John: I think the idea of regressing on a point is really the fundamental FDA flaw. You have an indication—you have lupus or you don’t. You have high blood sugar or you don’t.

Host: But that could change. Again, I think we could see changes in the way that clinical trials are designed, run, integration with AI tools, biomarkers that can be used for as surrogate endpoints. We already see this happening in certain conditions.

John: They got a long way to go. So I agree.

Host: No, I mean you and I think are mostly in alignment here. I don’t see a future—maybe it’ll happen—I don’t see a future where we just allow anybody to bring any drug to market and start making claims and selling it to people without safety—

John: But you have to have—no, no, no. Again, there’s always got to be a credentialing mechanism.

Host: 100%.

John: And maybe it’s not FDA. Maybe it’s something—I think something will replace the FDA that is more intellectually fluid.

Host: That should happen regardless of whether it happens within FDA or as a—

John: The conceptual model is dead.

Host: Yes. I think it needs to change for sure.

John: We—okay. So let’s talk about the mTORC1-mTORC2 pulse versus non-pulse data.

Host: Yes.

John: I was listening to Joan give her talk about the virus in some rapamycin series and she said, “Oh, we just need a pure mTORC1 inhibitor because we don’t want this off-label toxicity.”

Host: Maybe worth just so everybody who’s not a rapamycin officianado knows what we’re talking about. So Joan’s model—and I mean I think the data more or less back this up—is that chronic inhibition of mTOR complex 1, which is what high dose daily rapamycin will do, leads to indirect effects on mTOR complex 2. The model from that though is that it’s these off-target effects on mTOR complex 2 that lead to all of the negative side effects associated with—

John: But she said something that I thought was really kind of amusing from my end which is she said, “And if it gets high enough you start inhibiting PI3 kinase.” I’m not sure I understand why that would make any difference.

Host: Which is interesting as well because RTB-101, the drug that they ended up using in their—this is resTORbio now—in their pivotal phase 3 is a dual kinase inhibitor—

John: So they picked a drug—

Host: Yeah, exactly. So what do you—

John: Sorry, Joan. So what do you see clinically? I have a rough estimate of PI3 kinase gain of function, which is the insulin level divided by the BMI. This has been seen in people that have PI3 kinase—so it is a phenotype that you see in people who have gain of function PI3K. It’s insulin sensitivity. How much bang for your buck do you get for your insulin?

So you have this very low insulin level and you have a very high BMI.

Host: Does body composition matter or is it just BMI at this point?

John: My guess is—I haven’t studied this enough, but I have this data that I’m ready to put in an AI model whenever I can get the funds. We have a rough estimate of PI3K. So large people who have low insulin levels.

Host: Yes.

John: But it doesn’t have to be large.

Host: There’s some play in this that’s really interesting, I think, in terms of developmental—when does steroid turn on and off, this kind of stuff.

John: And in genetic PI3K activation, you’re going to have that developmental effect.

Host: Yeah, yeah. I know. There’s good stuff here.

John: Let’s say we have a population that’s got a PI3K gain of function, and they’re insulin sensitive, and they have autoimmune disease, and they may have autistic traits. I know that exists. I have a cluster of those people. You give them rapamycin, their sugars don’t move at all. Zero. You don’t see changes in triglycerides. You don’t see changes. You may see a little bit of insulin. You may see a trace of glucose, but they’re on four milligrams of rapamycin. You don’t see anything.

Host: Interesting.

John: Okay. You have other people you put on one milligram of rapamycin. I have a guy I just talked to last week. He’s like, “Oh, I really like the way I felt on it, but my sugars went up 30 or 40 points.” One milligram.

Host: Right. Yeah.

John: So are there people who need a dual kinase inhibitor and the search for the holy grail of the pure mTORC1 is perhaps a distraction?

Host: Yeah. I mean, again, I don’t disagree with that. I’ve certainly said many times over the years that while it’s a nice model, we don’t really know that this idea that all the good stuff comes from mTORC1 inhibition and all the bad stuff comes from off-target effects on mTORC2. Obviously, that’s too simplistic anyways, but it’s going to be very individual, which I think is what you’re getting at, right?

John: I mean, there’s data in PI3 kinase mice—use interfering RNA for mTORC2, you treat the autism and the epilepsy. So you may—I mean again this is dose, dose, dose, right? We need a lot more thought about this. There may be people that need 3 to 4 milligrams not because you’re trying to inhibit mTORC1—you’re trying to hit mTORC2.

And I inhibit mTORC2 all the time. I’ve been inhibiting mTORC2 for 30 years because I use lithium and I use high-dose tamoxifen, the PKC inhibitor. We do this all the time and we’re comfortable. We love it. We have patients that love us, they’re happy with us.

Is rapamycin going to be a mood stabilizer in the treatment refractory bipolar patient?

Host: What’s your gut feeling?

John: I have people that—I mean 30 years of depression—they’re like, “I’m on rapamycin, 70% better.”

Host: Yeah.

John: Right.

Host: And how often does that happen?

John: Not often enough. I want to generalize. But I think you saw it in your paper as well.

Host: Yeah. We definitely had a signal. I mean, there were some people like, “Oh my god, it’s not just fatigue. It’s not just pain. It’s like I’m not suicidal anymore and I’m really excited about life.”

John: Right. Right.

Host: So let’s come back to the paper for a minute because I want to understand a little bit better what the structure here was and what your role was. So some of the patients in this paper were your patients, right?

John: I think I had about 20.

Host: Okay. And was are you able to share? Are you willing to share? How did your group differ from the rest? Dosing was different it sounds like.

John: Yeah. As a psychiatrist, I just live in this world of human frailty and I just can’t stand asking people to dose once a week because it’s just—people are creatures of habit. Just dose every day. Keep things simple. It’s too many meds. So I just said, “Hey, look guys, will you let me dose once a day?” They said, “Sure, no problem.” So I’m dosing once a day.

Host: And was the cumulative dose about the same?

John: About the same. It’s close enough.

Host: Okay. Yeah.

John: What I heard from the PI, he says, “Your people look a little different. Their succinate levels were different.” And of course, that really got my attention, but I haven’t heard anything more about it. I was like, “Please, can I have my succinate levels please?”

It turns out succinate’s one of the predictors of alcoholism, which is its own interesting—one of the chief genetic risk factors.

Host: Let me just tangent on that for a second. Is it a predictor or it’s a responder? Like—

John: It’s genetic. Genetic gain of function. It predicts alcoholism.

Host: Okay. Yeah. And that would be a podcast alone—to talk about succinate because that’s a good topic.

Anyway, so my people were one milligram a day and I probably had a broader referral bias than the other members. The other members were probably very much like, “Oh yeah, I was really doing well. I got COVID and I never recovered,” or “I get infections and I’m always wiped out for three months,” that kind of stuff. My group was a little broader because I was just looking more in sort of hyperinflammatory syndromes.

I think I had two or three people that were poisoned by PCBs and they had activation of the aryl hydrocarbon receptor, and so you get all sorts of just very strange downstream immune activation. This is an area of huge litigation and you can’t trust anybody because everybody’s dirty, but nobody will treat them. So I had some people in that, and it was a broader group.

Host: Okay. And the way this was designed—the study had been funded, there were psychiatrists, medical doctors—I’m guessing generally internists?

John: Internist. I was probably the only psychiatrist who agreed to participate. When somebody came in with chronic fatigue symptoms, they would be—if eligible—put on rapamycin for I think it was very simple. You meet the entrance criteria and I would just sort of send people and let them say, “Hey, is that close enough to the entrance criteria?”

Host: And was there a placebo group or everyone in the study—

John: Just everyone got—everyone got around a milligram a day or six milligrams a week, that kind of stuff. And they were followed for four months. People who got benefit kept taking the drug. People that didn’t get benefit dropped out. I think some people dropped out because they couldn’t afford the lab work or they couldn’t afford—

Host: Oh, so the patients paid for the medication. There was a patient out of pocket expense and that was a real problem.

John: Yeah. Unfortunately that’s the world we live in with rapamycin—underfunded studies.

Host: So I looked in my group—and in my group was about 20 people—I think at the end of four months, four to five were still taking it. So it was a pretty small number needed to treat—was four. But that was at one milligram. I didn’t go higher. I didn’t go higher. Personally, in medicine, if you can get a number needed to treat of four for a chronic illness that has no treatment, you’re doing pretty good.

Host: Yeah.

John: I’ll give you an example of one of these guys that I’m seeing now. He says, “I get a viral infection every two weeks. It always wipes me out. This has been my whole life. I’ve been on rapamycin”—and I think he’s up to two or three milligrams now. He said, “I haven’t gotten the viral infection in two years.”

I mean the Joan Mannick story just writ large. I just don’t get sick anymore. He’s still getting infected. I mean, we’ve been infected with viruses since—

Host: Just not having the hyperinflammatory response.

John: Right. Yeah. And again, going back to COVID, which is really—it’s just a fascinating—COVID’s been such a wonderful lesson for all of us.

Host: It’s been—if you’re religious, God is chastising you and making you a better person.

John: Chastening my soul.

But there was this issue—people didn’t get—there was a group of people that never got infected with COVID because they didn’t have antibodies. I think some people looked and said, “Oh no, their T-cell response was so good, there was never enough innate immune activation to actually drive B cell antibody maturation.”

Host: Interesting. Yeah.

John: Right. So if your immune system—if your innate immune system is cooking with gas and your T cells are cooking with gas, this antibody testing may not even be relevant. It’s kind of cool to think about being that wonderful person that never gets sick—the immune system is so perfectly tuned.

Host: Yeah. I often look back at the whole resTORbio timeline relative to COVID and just imagine how different the situation could have been because they shut down that pivotal clinical trial in November 2019.

John: I think another topic that we should really think about in the area is the issue of high-dose pulse rapamycin.

Host: Okay. Yeah. Let’s talk about that. So what do you mean by high dose in this context?

John: Okay. So again, the transplant people are giving us lessons and we’re not listening. What they do when they give someone a bone marrow transplant is they 3 to 4 days after a transplant they give a high dose—I think of cyclophosphamide—which is a DNA damaging agent and it kills rapidly dividing cells. Those rapidly dividing cells are literally the T-cells and B cells that are antigenically specific for the host.

So they’re kind of cleaning out the most dangerous part of the graft versus host disease. So there’s this role of killing rapidly dividing immune cells when there is an injury. You’re clonally deleting these cells. You’re making the stem cell population cleaner, and we know rapamycin can do that.

You think about what we do in medicine all the time. Someone comes in with migraine, back pain, rheumatoid arthritis flare, post-respiratory asthma—you’ll give them 40 milligram pulse of prednisone, 30, 20, 10, and just sort of crush the inflammation. You’re trying to knock it down. But it may be more than just crushing the inflammation. Steroid kills—steroid causes apoptosis in certain lymphoid cells.

So the body naturally does clonal adverse selection for overly mTORC1 growth cells that grow too much, have too much BCL-2, they don’t die the way they’re supposed to, they just sit around and you kind of have—if you will—a low-grade multiple myeloma or lymphoma. You’re just always kind of there with this clonally selected hyperinflammatory cell. And the body does this all the time with steroids. You get a steroid pulse from the adrenals when you’re infected. Everybody in medicine is comfortable with this. Medical students do this.

Is there a role for a rapamycin pulse?

Host: Yeah. That’s interesting. So you’re talking like a very high acute dose one time maybe?

John: Yeah. One time. Yeah. Just cause massive apoptosis.

Host: I don’t know. I mean, would it cause massive apoptosis? I think—but there’s this interest—like I’m seeing women that are like going to get pregnant, and there’s all this beautiful data on ova quality when you give pregnant pups rapamycin. For a month or so, you clean out the bad eggs and you leave the good—somehow you select.

Host: Yeah. I don’t think we know the mechanism there. I mean again—and maybe it would be different with a high acute dose—typically rapamycin is not cytotoxic. It’s cytostatic and it will—it may increase mitochondrial mitophagy.

John: Absolutely. So you’re going to clear out some damage in cells. Let’s talk about an egg. An egg is just full of mitochondria. It’s huge. Let’s say you pulse a woman who wants to get pregnant with a high dose of rapamycin and you’re cleaning out heteroplasmy that shouldn’t be there. That’s an interesting concept.

Host: Yeah.

John: And that wouldn’t be that hard to do in terms of people who are doing IVF. You randomize some women who are getting IVF and maybe they get two milligrams a day for a month or a week. And then you have a control group and then you do your standard embryos and you just say, “Hey, what’s my embryo quality look like?”

Host: Related—I’m going to see Esu at Columbia in about a week. She’s involved with Zev Williams leading the rapamycin premature ovarian failure trial. So these are in fact women who would like to get pregnant and have premature ovarian failure. I believe they’re doing six milligrams once a week or four milligrams once a week in their trial. I haven’t seen any data from that trial yet, but it’s an interesting concept—that maybe a single acute dose or a single week of dosing would have a beneficial effect on egg quality.

John: I had a gal that came to me and we did CSF metabolomics and she had a marker for early Parkinson’s disease that was like seven z-scores up. I think a family member had Parkinson’s disease, and so she got on rapamycin. After 3 months she said, “Oh yeah, I don’t have any more endometriosis. It’s gone.” And she stopped rapamycin and never came back.

Host: Yeah.

John: Can we do a single pulse of rapamycin and do adverse clonal selection and it’s somewhat curative for quite a while in terms of stem cell quality? How we do that in a safe way, that’s a real challenge.

Host: I think that’s going to take a lot of thoughtful—

John: Why is it a challenge? Are you more worried about safety from a single acute dose than you are from chronic dosing?

Host: How much is enough to really clear out that—is it do you need a pulse of 20 milligram? Do you need a pulse of 40 milligram? Where do you start getting—I mean, I think there are basic fundamental questions that need to be researched. And again it took the cancer guys a long time to understand—this is the dose of cyclophosphamide that I need in order to have adverse clonal selection in these bone marrow transplants. But everybody’s too afraid of this topic to even start. Somebody has to start.

Host: Well yeah, and again I don’t think we have very much data on acute dosing with rapamycin.

John: Although some of the people who’ve been using rapamycin off label have definitely pushed it into the 20-40 milligram range routinely and it’s all anecdotal without significant toxicity.

Again I have people on like 3 to 4 milligrams a day. There’s zero change in their glucose and their triglycerides.

Host: Are they more susceptible to infection—bacterial infections—in your experience at those doses? I mean that would be the thing most people would be worried about—it’s primarily immune suppression, and in particular bacterial—

John: I think Joan Mannick’s point is a really good point—is there’s a group of people and this is well known in immunology—this is actually a known cause of tuberculosis which is an overgrowth syndrome in the cytotoxic T cells and they just grow so fast in response to antigen that they die. They turn senescent or they become exhausted, and it’s a risk factor for actually getting TB because you have a hyperactive immune system.

So if we had better biomarkers—

Host: No, I’m in agreement. I was just asking in your experience, do your patients who are on three to four milligrams a day typically seem to get more bacterial infections at a rate that you pick up on?

John: I clearly have dropouts. There’ll be some people that are on one milligram a day, they’re getting benefit. You go up to two milligrams, they’re saying, “Oh, I’m getting some thrush,” or “I’ve had three urinary tract infections in the last week” or something. But then again, it’s all quality of life.

Host: Yeah, of course.

John: There’s a million people on transplants right now that’ll be like, “Hey, if I need antibiotics twice a year and my quality of life is fantastic, no problem.”

Host: Yeah. No, I’m in complete agreement with you. And I think that what you said is really important that we always need to balance the side effects with the benefit. That’s going to be very individual.

Mostly I was just wondering from my own knowledge and experience base at those doses how prevalent are these kinds of side effects in terms—because it’s my perception that six milligrams a week it’s pretty low, like barely detectable if there is a signal, it’s very small for bacterial infections—

John: What have I seen? I’ve seen capillary leak at 1 milligram. I’ve seen the mouth sores. The mouth sores is definitely—

Host: Sure. There’s I think 15% ballpark of people.

John: I don’t see them that often. Again, I think that’s referral bias. It’s really interesting. Referral bias means something. I think it’s related to Behçet’s syndrome, which is when you push the immune system in one direction, the other portion can kind of poke out.

But I think last time you and I talked, I was like, “I don’t get these mouth sores.” I get these mouth sores more. I think this is closer to Behçet’s. This is where we need really experienced rheumatologist immunologists to be part of these teams.

Host: Yeah. Absolutely.

John: More than just the transplant people. Because the transplant people talk about how you manage it symptomatically. What we’re really looking for are the source code people.

Host: Cool. Okay. Anything that we haven’t touched on that you really feel like we should talk about?

John: I think just going back to—I think I sent you an email on ketamine. Going back to the talk that we had last year about how ketamine probably is one of its targets is the M2 microglia, and there’s a little bit more to that story since we’ve talked.

Host: Okay.

John: So first of all the microRNA testing that we talked about—when they looked at these patients with chronic pain and they did gene set enrichment in a large enough group of things, what did they see? They saw pathways associated with TNF-alpha, IL-1 beta, RNA polymerase 2 negative regulation. Super interesting. And then the clinical pathways were related to psoriatic arthritis, graft versus host disease. And the drugs that they recommended in the chronic pain patient were ketamine, prednisone, rapamycin.

So it’s nice to see that not only can we start with biomarkers which is the RNA that say, “Hey, you should try this drug,” but you get into things like RNA polymerase 2 which is one of the checkpoints for efferocytosis—digesting the corpses of dead synapses. It’s one of these sort control functions. And in liver, RNA polymerase 2—40% of it is stalled because it’s seeing DNA damage and it just sits there. It gets hung up on the DNA damage.

So this whole issue of just going deeper and deeper into the source code because we have better tools—right now we have a tool that says, “Oh we should start looking at RNA polymerase 2 as one of these causal factors for chronic pain.” That’s a major advance. And we have a commercial product that patients can get and we don’t have to wait for a grant to get funded.

Host: And just so I understand that the drug recommendation—ketamine, prednisone, rapamycin—that’s an AI based recommendation based—

John: It’s an AI based. There’s no human data.

Host: That’s intriguing.

John: Yeah. Yeah, it is. And then that’s one portion I thought was very interesting that we have something that’s actually logical in terms of saying—and it’s commercially available.

Host: What’s the test called?

John: It’s called MindX-1.

Host: MindX-1. Okay. And is there anything new? Because we talked last time about the ketamine plus rapamycin data. Anything new there that’s worth—

John: Well I think I sent you an email on this and I think it’s really fascinating which is going back to Phil Morgan’s work on anesthetic sensitivity. Luca Turín and Nick Lane have been looking at this concept that the reason anesthetics work is because they affect the polarization of the electron flux through complex 1, and that a polarized electrical field generates a magnetic field.

So if you’re thinking about this very important issue of chaperones—alpha-synuclein is a cardiolipin chaperone—the electron flux through complex 1 organizes the shape of the mitochondria. The ferric sulfide clusters are going to align in relation to complex 1.

So what do anesthetics do? Something like xenon. Xenon has no chemical binding. How in the world could it act as an anesthetic other than it interrupts the electron flow, the polarization of the electron flow?

So then what happens? Then you get a collapse of the mitochondrial structure in the presynaptic neuron, and you get a mitochondrial unfolded protein response. So you get a short-term effect which is you get an anesthetic, but you get a longer-term effective effect which is the mitochondrial unfolded protein response.

And now we could start to think about this issue of what are anesthetics doing in terms of subacute issues, and what does it mean when somebody is really sensitive to ketamine? Like I have some people that are on 25 milligrams of ketamine—that’s all they can take and they’re too sleepy. This is not a microglia phenomenon. This is somebody that has a defect in the presynaptic glutamatergic neuron, very similar to the ND4.

Host: Which is not—which is in the membrane. It’s not one of the actual charge carriers.

John: That’s right. Yep. It’s an accessory or stabilizing part.

Host: Yeah.

John: And so now we have this terribly interesting question of, “Oh ketamine’s doing more and telling us more than simply it’s an anti-inflammatory.” If you listen to your patients carefully, you’re going to hear echoes of the ND4. And I mean, again, $20 bill just right on the ground. And we need people to pick it up. And there aren’t a lot of us.

Host: Yeah. Yeah. Well, and getting funding to do that kind of research is again going to be challenging for something like ketamine.

John: Yeah. Pretty tough.

Host: So, clinically, again, I’m intrigued by this interaction though between rapamycin and ketamine that I think you’ve talked about and that has been published on just a little bit, but I haven’t been following it, but I haven’t really seen any new studies there.

John: No. Is there any—is there anything there or do you feel like—oh the interaction between the two?

Host: Well, or a combination when you combine them in patients—

John: Oh, I probably have at least 20 people on the combination. Yeah, at least.

Host: And how do you choose who to put on the combination or what would be the typical traits?

John: Again, I’m thinking—I’m just trying to summarize because people are all individuals. Usually what happens is you have a patient who has absolutely treatment resistant pain or depression. They’ve tried everything. And I get people sent to me typically by other physicians in the community saying, “Hey, we’re done. I don’t know what to do.”

People start with ketamine because it’s relatively cheap and there’s been a lot of ancillary data from the IV people, so it’s credible. It has face credibility. Rapamycin doesn’t have face credibility yet. And you have to spend a lot of time doing informed consent. And some pharmacies will refuse to fill it, which really—

Host: Yeah. Interesting.

John: Amazon pharmacy doesn’t.

Host: No, I’ve had a pharmacist refused to fill it. I actually sent a pre-print of that paper saying, “Hey, I’m in—I’ve been a multi-center trial with Mayo and it would be nice if you filled the patient’s medication.” “Oh, no. We’re not going to do that.”

So it has less face validity. The combo—again we have to go to the basic science. I think when you do RNA analysis of the M2A and the M2C microglia—so the M2C is activated by steroid and IL-6 and the M2A is activated by IL-4—rapamycin tends to do—looks like an IL-4 M2A microglia and ketamine’s an M2C. Their transcription profile looks a little different.

And I’m just not good enough to really know what that means from a physiological perspective. It looks like the M2A have more arginase. So, paracrine signaling—they’re breaking down the precursors for nitric oxide which is produced by the M1.

Host: Right.

John: So these cells are competing against each other and they may even be competing within themselves in a paracrine way for—am I going to be an M1 or am I going to be an M2A? And then I think the M2C has HIF-1-alpha—

Host: Hypoxia sensing.

John: Right. I mean there’s a lot of local physiology.

Host: Yeah. Sure. Of course.

John: I mean but ultimately—I agree with you—figuring out those mechanisms is the next step for the deep science. But ultimately does a person respond or not?

Host: From a patient perspective.

John: Yeah. And they respond. But I draw people the graph in my clinic. I say, “Look, let’s look at a 2D space. One component—x-axis—is your rapamycin dose. The other component is your ketamine dose. You and I will both be dead before we find the global optimum, because we don’t have an easy biomarker. We can guess but our address is not going to be very precise.”

And it just goes back again and again to the readers—to the listeners—dosing really matters.

Host: So in the people you’ve got who are on both, is that typically—it’s still a daily dose of rapamycin like one milligram—

John: Yeah. I hate to pulse people. It’s just—they’ve got other medical conditions and it’s just so hard.

Host: Sure. That makes sense. Yeah. Yeah. Interesting.

All right. Well, thank you so much for coming by today. We’ve covered a lot of ground, a little bit of a deep dive, a lot of concepts. And I think I’m obviously going to continue to follow the lithium story, see where that goes. I think there’s going to be a lot more research on it coming out soon.

Any parting words you want to leave with our listeners?

John: I think you need to join OptiSpan. Come on down to OptiSpan. These are good guys. They really are doing their part.

Host: Thanks, John. I appreciate that. And as always, love being able to spend time with you and talk about all of this cool biology.

SUMMARY

This conversation between two researchers discusses emerging research on lithium and rapamycin, focusing particularly on their potential roles in treating Alzheimer’s disease and chronic fatigue syndrome.

Lithium and Alzheimer’s Disease

Key Findings:

  • A recent Nature paper from Harvard showed that lithium dramatically reduced markers of Alzheimer’s in mouse models and found lithium depletion in specific brain regions of human Alzheimer’s patients
  • People on lithium historically show longer telomeres, less brain atrophy on MRI, decreased death rates, and decreased rates of Alzheimer’s disease
  • Lithium appears to work by inhibiting mTORC2 and promoting autophagy, affecting multiple cellular pathways

Clinical Applications:

  • The ATN profile (amyloid, tau, and neurofilament light) provides biomarkers to identify early-stage Alzheimer’s risk
  • Early intervention before significant neuronal death appears most promising
  • Patients report feeling “clearer” and handling stress better even without bipolar disorder
  • Cost is approximately $40/year for generic lithium, compared to $70,000/year for newer Alzheimer’s drugs

Dosing and Safety:

  • Therapeutic dose varies significantly by individual (“biology laughs at” standardized dosing)
  • Different lithium salts (carbonate, citrate, orotate) may have different efficacy profiles
  • Main side effects at higher doses (>0.7 blood level): tremor, decreased urine concentration, brain fog, mild migraine symptoms
  • Thyroid effects are manageable with supplementation
  • Concerns about kidney failure appear confounded by underlying disease progression
  • Specific populations to monitor: temporal lobe epilepsy and migraine patients

Rapamycin and Chronic Fatigue Syndrome

Study Results:

  • Approximately 75% of chronic fatigue patients showed improvement on rapamycin
  • Benefits included: reduced fatigue, improved pain, better sleep quality, reduced orthostatic intolerance
  • Particularly effective in patients whose chronic fatigue followed viral infections
  • Notably, no significant changes in glucose, insulin, or lipid markers were observed

Mechanism:

  • Hyperinflammatory response to foreign RNA/DNA (viral, bacterial, or vaccine-related)
  • Rapamycin upregulates T-regulatory lymphocytes, dampening excessive immune responses
  • Increases autophagy and shifts immune balance from effector to tolerant state
  • Similar mechanism explains benefits in autoimmune conditions like lupus

Dosing Controversies:

  • Study used 1 mg/day or 6 mg/week
  • Discussion suggests higher doses (3-4 mg/day) may be more effective based on transplant medicine data
  • Graft-versus-host disease prevention requires 3-20 ng/ml trough levels
  • Individual response varies dramatically—some show metabolic changes at 1 mg, others show no changes at 4 mg
  • Number needed to treat: approximately 4 patients

mTORC1 vs mTORC2:

  • Debate over whether “off-target” mTORC2 inhibition is problematic or beneficial
  • Some patients may specifically need mTORC2 inhibition
  • PI3 kinase gain of function patients may benefit from dual kinase inhibition
  • Weekly pulsing vs. daily dosing remains controversial

Additional Insights

Ketamine Combination:

  • 20+ patients on ketamine + rapamycin combination
  • Different microglial activation profiles (M2A vs M2C)
  • Used for treatment-resistant depression and chronic pain
  • Dosing requires individual titration in “n-dimensional biochemical space”

MicroRNA Testing:

  • New commercial test (MindX-1) analyzes 400-500 RNA species
  • AI analysis consistently recommends lithium across multiple phenotypes
  • Suggests mTORC2 gain of function as common feature in certain patient populations
  • May help identify responders and optimal drug combinations

Regulatory and Research Challenges:

  • Generic drugs lack financial incentive for large pharmaceutical trials
  • FDA requirements for dose escalation slow research progress
  • Specialist waiting times: 6 months to 2 years for Alzheimer’s clinics
  • Need for multivariate, personalized approaches rather than population-level protocols

High-Dose Pulse Therapy:

  • Theoretical approach: single high dose to cause clonal deletion of hyperactive immune cells
  • Analogous to cyclophosphamide use in bone marrow transplant
  • Anecdotal success in endometriosis treatment
  • Potential application for improving egg quality in IVF
  • Safety profile and optimal dosing remain unexplored

CRITIQUE

Strengths

Clinical Experience and Practical Insights: The conversation provides valuable real-world clinical experience with both lithium and rapamycin that extends beyond published literature. John’s 20-25 years of patient follow-up offers longitudinal data rarely captured in formal studies. His observation that “people on lithium can be essentially completely normal” when seen every couple of years represents the kind of sustained outcome data that clinical trials typically don’t capture.

Systems-Level Thinking: The discussion demonstrates sophisticated understanding of biological complexity, particularly the concept of “n-dimensional biochemical space” and nonlinear drug responses. The acknowledgment that “biology is a continuum” and “laughs at clear cut points” represents mature thinking about personalized medicine that moves beyond simplistic dosing protocols.

Mechanistic Depth: The conversation integrates multiple levels of biological understanding—from molecular signaling (mTORC1/2, PI3K, autophagy) to cellular responses (T-regs, microglia activation) to clinical phenotypes. The discussion of locality and paracrine signaling versus systemic biomarkers shows nuanced understanding of measurement challenges.

Identification of Knowledge Gaps: Both participants honestly acknowledge uncertainty and areas needing further research, from optimal dosing strategies to the relative importance of different mechanisms.

Weaknesses and Concerns

Anecdotal Evidence Base: Much of the clinical discussion relies on anecdotal observations from John’s practice without systematic data collection, peer review, or controlled conditions. Statements like “I have people that… 30 years of depression… 70% better” lack the rigor of formal assessment tools, blinding, or comparison groups. Referral bias is acknowledged but may be more influential than recognized.

Dosing Recommendations: The casual discussion of people “buying lithium on Amazon” and self-dosing raises significant safety concerns. While John acknowledges this is happening because of specialist shortages, there’s insufficient emphasis on the serious risks of lithium toxicity, drug interactions, and the need for blood level monitoring. The blood level of 0.7 mentioned as a rough guideline needs much more context about individual variation, kidney function monitoring, and drug interactions.

Vaccine Injury Discussion: The statement “I saw a lot of vaccine injury” based on patients presenting with chronic fatigue following vaccination requires much more careful framing. Without controlled studies, it’s impossible to determine:

  • Baseline rates of chronic fatigue in comparable populations
  • Whether vaccination was causal or coincidental
  • The denominator (how many patients were vaccinated without problems)
  • Whether these patients had predisposing conditions

The concern is that such statements, even when prefaced with “I’m not anti-vax,” can fuel vaccine hesitancy without providing sufficient epidemiological context.

Regulatory Critique: The dismissal of the FDA as “useless” and “dead” oversimplifies the complex role of regulatory oversight. While valid criticisms exist about dose escalation requirements and adaptiveness to personalized medicine, the FDA serves important functions in preventing dangerous or fraudulent products from reaching market. The enthusiasm for AI-driven individualized medicine doesn’t address who validates the AI, monitors for harm, or protects vulnerable populations.

Publication Bias: The chronic fatigue study is discussed extensively before peer review publication. While pre-prints serve important functions, definitive claims about efficacy should be tempered until peer review is complete. The “number needed to treat of 4” is presented as impressive, but without placebo control, natural history data, or standardized outcome measures, this figure may be misleading.

Mechanistic Overinterpretation: Several mechanistic explanations are presented as established fact when they remain speculative:

  • The link between high IQ and mTORC2 upregulation
  • The specific role of lithium in clonal deletion
  • The interpretation of autophagy signals in blood as representative of tissue-level processes
  • The distinction between M2A and M2C microglia activation with different drugs

Temporal Lobe Epilepsy Risks: The discussion of lithium potentially causing or exacerbating seizures in temporal lobe epilepsy patients, while important, lacks quantification of risk. The anecdote about rhinorrhea as a seizure indicator, while interesting, shouldn’t be promoted as a monitoring strategy without validation.

Off-Label Prescribing Ethics: While John acknowledges informed consent requirements, the casual discussion of prescribing 3-4 mg/day rapamycin for depression or chronic pain—far outside standard indications—raises ethical questions about:

  • Adequate informed consent when evidence base is limited
  • Patient vulnerability when desperate for relief
  • Financial conflicts (though not apparent here, off-label prescribing can be lucrative)
  • Documentation and monitoring requirements

Migraine Contraindication: The statement that lithium “makes migraine worse” in many patients is clinically important but presented without quantification, mechanism explanation, or guidance on management. Given that 25% of women have migraines, this represents a significant population who might attempt lithium self-treatment.

High-Dose Pulse Therapy: The speculation about high-dose pulse rapamycin therapy for clonal deletion is interesting scientifically but presented without adequate safety discussion. Questions include:

  • What dose would be required?
  • What are acute toxicity risks?
  • How would efficacy be measured?
  • What patient populations might benefit?
  • What are the ethical considerations for experimentation?

The anecdote about endometriosis resolution is intriguing but represents a single case without follow-up, mechanism confirmation, or replication.

Missing Context

Health Equity: No discussion of access disparities. While lithium is inexpensive, monitoring requires blood tests, kidney function tests, and thyroid monitoring—costs that may be prohibitive. The mention that chronic fatigue study participants dropped out due to inability to afford lab work highlights this but isn’t explored further.

Patient Selection: John’s practice likely represents a highly selected population—sophisticated, educated, willing to try experimental approaches. Generalizability to broader populations is unclear.

Comparative Effectiveness: Limited discussion of how lithium compares to existing Alzheimer’s interventions beyond cost. The approved amyloid antibodies, while expensive and modest in benefit, have been through rigorous trials. Similarly, rapamycin isn’t compared to existing chronic fatigue or autoimmune treatments.

Long-term Safety: Decades-long lithium use has established safety profile for bipolar disorder, but this is at therapeutic serum levels with regular monitoring. Long-term effects of lower doses for Alzheimer’s prevention in aging populations remain unclear.

Scientific Validity Concerns

MicroRNA AI Test: The MindX-1 test is presented uncritically. Important questions include:

  • Validation studies?
  • False positive/negative rates?
  • How was the AI trained?
  • Were drug recommendations tested prospectively?
  • Cost-effectiveness?
  • Replication by independent groups?

Succinate and Alcoholism: The casual mention of succinate as “one of the chief genetic risk factors” for alcoholism needs citation and context. This appears to conflate a metabolic marker with genetic causation.

Complexity vs. Clarity: While the emphasis on biological complexity and individual variation is intellectually honest, it can become an excuse for lack of systematic study. “Everything is connected” can prevent hypothesis testing and controlled experimentation.

Positive Elements Often Overlooked

Hypothesis Generation: The conversation generates numerous testable hypotheses:

  • Lithium orotate vs. carbonate comparative trial
  • Rapamycin dose-response in chronic fatigue
  • High-dose pulse therapy pilot studies
  • PI3K genotype predicting rapamycin response
  • Combination therapy protocols

Pattern Recognition: John’s longitudinal experience allows pattern recognition across patients that might not be captured in shorter studies. The observation about personality changes preceding cognitive decline in Alzheimer’s is clinically valuable.

Multidisciplinary Integration: The conversation bridges psychiatry, neurology, immunology, metabolism, and aging research in ways that formal studies often don’t.

Recommendations for Interpretation

For Patients:

  1. This conversation represents frontier clinical experimentation, not standard of care
  2. Do not self-prescribe based on this discussion
  3. Any lithium use requires medical supervision and monitoring
  4. Rapamycin for chronic fatigue remains experimental
  5. Individual responses vary dramatically

For Clinicians:

  1. Consider the hypotheses generated but require supporting evidence
  2. Recognize the limitations of anecdotal clinical experience
  3. Maintain careful documentation if pursuing off-label approaches
  4. Ensure robust informed consent
  5. Monitor carefully for both benefits and harms

For Researchers:

  1. The clinical observations suggest important research directions
  2. Biomarker development for patient selection is critical
  3. Dose-finding studies are urgently needed
  4. Combination therapy deserves systematic study
  5. Registry studies of off-label use could provide valuable data

Overall Assessment

This conversation represents a fascinating glimpse into cutting-edge clinical thinking about aging, neurodegeneration, and chronic illness. The integration of mechanistic biology with clinical observation is valuable, and the hypotheses generated deserve investigation.

However, the presentation risks overstating certainty about clinical benefits while understating risks and knowledge gaps. The enthusiasm for personalized medicine and rejection of “one-size-fits-all” approaches is justified, but shouldn’t abandon systematic evaluation of safety and efficacy.

The discussion would be strengthened by:

  • More explicit acknowledgment of evidence levels
  • Clearer distinction between hypothesis and established fact
  • Greater emphasis on monitoring and safety
  • More critical evaluation of biomarker validity
  • Discussion of failed approaches and negative results
  • Consideration of health equity and access

This represents important exploratory clinical work that should inform future research, but listeners should not interpret it as a treatment protocol or sufficient evidence for self-experimentation.