Rapamycin *worsens* osteoarthritis in non-human primates (BioRxiv) by Adam Salmon, etc

#1

Inhibition of mTORC1 by rapamycin results in feedback activation of AktS473 and aggravates hallmarks of osteoarthritis in female mice and non-human primates

Paper summary: https://www.biorxiv.org/content/10.1101/2024.05.14.594256v2

Full PDF: https://www.biorxiv.org/content/10.1101/2024.05.14.594256v2.full.pdf

I hadn’t seen this posted here. All I found is a 202 thread saying that Rapa delays osteoarthritis (Rapamycin Delays Age-related Osteoarthritis)

I think the study abstract says everything you need to know, so I just copy-paste it here:

ABSTRACT
Purpose: Genetic deletion of mTOR has protected against post-traumatic osteoarthritis (OA) in
male mice, however, effects of pharmacological mTOR-inhibition are equivocal and have not been
tested in aging models nor in female subjects. Therefore, the goal of this study was to determine
if mTOR-inhibition by rapamycin can modify OA pathology in aging non-human primates and
female mice.

Methods: Common marmosets were administered oral rapamycin (1mg/kg/day) or vehicle starting near mid-life until death. Five-month-old, female C57BL/6J mice were treated with vehicle or
rapamycin (IP, 2mg/kg, 3x/week) for 8-weeks following non-invasive ACL rupture. Knee OA
pathology was assessed via microCT and histology. Phosphorylation of mTORC1 (p-RPS6S235/36)
and mTORC2 (p-AktS473, p-NDRG1T638, p-PKCαT348) substrates were evaluated via western blot
in articular cartilage, meniscus, and/or infrapatellar fat pad. ATDC5 cells were cultured with
rapamycin to determine time and dose effects on mTORC1/2 signaling.

Results: In marmosets, rapamycin did not impact age-related radiographic OA severity or
cartilage pathology but increased medial meniscus calcification and lowered lateral tibia
subchondral thickness, particularly in females. In female mice, rapamycin worsened ACLR-
induced meniscus calcification and cartilage pathology. In marmoset and mouse joint tissues,
rapamycin inhibited mTORC1 and increased p-AktS473 but not p-NDRG1T638 or p-PKCαT348

This mTOR signaling pattern was replicated in ATDC5 cells during exposure to low concentrations of
rapamycin.

Conclusions: Rapamycin attenuated mTORC1 signaling with feedback activation of AktS473 in
articular cartilage, meniscus, and/or infrapatellar fat pad and was accompanied by deleterious
effects on meniscus calcification and/or cartilage pathology in female mice and common
marmosets.

So this looks like a negative result from a good quality study. I like that they show every data point, and there’s a lot of biological variability, which looks very realistic and truthful to me. (When I see super tight error bars on papers, I am suspicious). They also do validate that Rapa was hitting the intended targets, suppressing mTOR etc.

What I will say is that it looks to me like the data were affected by a few strong outliers in the Rapamycin group. If you look at Figure 2B and 2D, this is very clear that 4 monkeys had huge calcification whereas 3 had zero. Whether those are some individual “bad responders”, I am not sure.

The marmoset dose (1mg/kg/day) looks incredibly high, which I know has been discussed in detail on this forum, but I don’t think we ever reached a consensus or had an explanation for why this was chosen. Still, the paper is still a sign that Rapa is not protective of OA in this model.

5 Likes
Why Smart People Fall for False Information and What to do About It (UCSF)
#2

When I interviewed Adam, I asked him about that 1mg/kg daily figure. They were modeling it after the ITP’s 14 PPM study. See here:

Speaker 2 15:40
I think in your 2015 paper, you started off doing one milligram per kilogram dosing. That’s a pretty aggressive dose. What was their rationale behind that?

Speaker 1 15:52
So part of it was based on the ITPs first 14 parts per million. Its in the ballpark of 14 ppm.

Speaker 1 16:20
So that was one part we were trying to talk target that. And I think following up from that, both from the ITP, from Matt’s lab, it’s shown that they can, mice can tolerate much higher doses and potentially added benefit. The other thing that we drove that on is we did an early pharmacology study pharmacokinetics, and we wanted to go lower doses for, I think, good reasons, right? And then with the side effects, we don’t want to get translational potential benefit, people say, a big, huge, high dose. But what we found is that much lower than one made per kilogram, the trough levels of grapple were ordering on undetectable. 24 hours after we gave them that dose, they were already out of that clinical even

Speaker 2 17:01
Don’t Marmosets have basically the same pharmacokinetics as humans, in terms of half life or not,

Speaker 1 17:10
you would think, but no, it’s, it’s probably more on par with the mouse estimated half life because we didn’t do full but it’s probably more on par with mice,

You can listen to the interview here - Source: Adam Salmon's Marmoset Longevity Study - #4 by RapAdmin

From ChatGPT:

Here’s a structured summary and analysis of the attached paper:

:pushpin: Summary of the Paper

Title: Inhibition of mTORC1 by rapamycin results in feedback activation of AktS473 and aggravates hallmarks of osteoarthritis in female mice and non-human primates

Objective:
To test whether pharmacological inhibition of mTORC1 with rapamycin can alter osteoarthritis (OA) pathology in female mice and aging non-human primates (common marmosets).

Methods:

  • Marmosets: Midlife animals received oral rapamycin (1 mg/kg/day) until death. OA severity assessed via microCT and histology.
  • Mice: Female C57BL/6J mice underwent non-invasive ACL rupture and were treated with rapamycin (2 mg/kg, IP, 3x/week, 8 weeks).
  • Cell culture: ATDC5 chondrocytic cells treated with rapamycin to examine dose/time effects on mTORC1/2 signaling.
  • Molecular assays: Western blots for mTORC1 (p-RPS6) and mTORC2 (p-AktS473, p-NDRG1, p-PKCα) signaling.

Key Findings:

  • Marmosets: Rapamycin did not reduce cartilage degeneration but increased medial meniscus calcification and decreased lateral tibial subchondral bone thickness, especially in females.
  • Mice: Rapamycin worsened meniscus calcification and cartilage degeneration after ACL rupture. Blood glucose was also elevated in treated mice.
  • Molecular signaling: In both species, rapamycin inhibited mTORC1 but caused feedback activation of AktS473, a signaling pattern replicated in ATDC5 cells at low rapamycin concentrations. Other mTORC2 substrates (NDRG1, PKCα) were inhibited or unaffected.
  • Conclusion: Rapamycin may aggravate OA pathology in females by enhancing Akt signaling despite suppressing mTORC1.

:white_check_mark: Strengths

  1. Translational model choice:

    • Using both mice and non-human primates (marmosets) strengthens translational relevance, bridging rodent and human biology.

  2. Focus on females:

    • Most prior OA/mTOR studies emphasized males; this paper fills a critical gap by testing female animals.

  3. Multi-level approach:

    • Combined in vivo (mice + primates) and in vitro (cell culture) systems, allowing mechanistic interpretation of signaling dynamics.

  4. Robust methodology:

    • Use of microCT, histopathology, and molecular assays provides convergent evidence for OA pathology and signaling effects.

  5. Novel mechanistic insight:

    • Demonstrates that pharmacological mTORC1 inhibition can paradoxically activate AktS473, suggesting a mechanism for OA worsening.

:x: Weaknesses / Limitations

  1. Cross-sectional primate design:

    • Treatment began at midlife in animals with likely pre-existing OA, limiting the ability to assess rapamycin as a preventive intervention.

  2. Sex imbalance in findings:

    • Major effects observed in females, with unclear generalizability to males. Authors acknowledge males had lower baseline propensity for meniscus calcification.

  3. OA induction mismatch:

    • In mice, OA was artificially induced (ACL rupture), whereas in primates it developed naturally with aging. This complicates direct comparison.

  4. Site mismatch in signaling analysis:

    • Mouse signaling assays were performed on shoulder cartilage (due to antibody reactivity), not on the knee (the OA site).

  5. Dosing relevance:

    • Systemic rapamycin exposure may differ from clinical or targeted intra-articular administration. Authors note that low tissue penetration may drive Akt feedback effects.

  6. Lack of functional outcomes:

    • No direct measures of pain, mobility, or functional decline—only structural and molecular markers.

:bar_chart: Overall Analysis

This study provides compelling evidence that systemic rapamycin, while known for longevity benefits, may exacerbate osteoarthritis in females by inducing Akt feedback signaling. The translational inclusion of non-human primates is a major strength, but limitations (cross-sectional primate design, lack of functional readouts, OA model differences) restrict generalizability.

The results suggest that:

  • mTOR inhibition alone may not be a viable OA therapy.
  • Sex-specific effects are important in OA research.
  • Alternative delivery methods or dual mTORC1/mTORC2 targeting may be required to avoid Akt overactivation.
3 Likes
#3

Do people read these summaries of papers/videos? I’m actually not sure.

3 Likes
#4

On average, when presented with an AI generated summary of a research paper in these forums, do you tend read them, or tend to avoid reading them?

  • Tend to Read the AI Summaries
  • Tend to Avoid reading the AI Summaries

0 voters

1 Like
#5

I feel in this instance, the study is fairly simple, and the study abstract was excellently written and gave the main points in a very clear and concise way. That’s why I pasted it here.

However, they can be useful, particularly for more complicated studies.

1 Like
#6

But you are using it in “agent” mode, with a back-and-forth discussion with the AI about the paper then, no?

#7

What I mean is an AI-generated summary like an abstract, but simpler. For example, those big data-mining type papers where the methods and conclusions are not always obvious.

However, this study is simple, basic research, and I think the AI summary posted above is longer and less substantiative than the simple study abstract.

Anyway - what do we think about Rapa and potentially worsening OA?

1 Like
#8

FWIW, the PEARL trial showed no negative signal reported for OA in women (or men), after 48 weeks. At 3mg/1-week, women have reported some positive muscoskeletal affects - in fact, it was the only measurable effect in that trial.

Two thoughts here. Obviously, dosage is an issue. Marmosets (and mice) do appear to metabolize rapamycin rather differently (and clear it far more rapidly) than humans (humans, based on the literature related to organ transplants). Therefore one should be careful translating the dosage effects from mice and marmoset models to humans. In humans, far lower dosages seemed beneficial for the muscoskeletal system in women. For men, possibly higher dosages are needed to see any effects - whereas notably in animal models here males were not negatively affected by rapamycin in the respect of OA. My thought here is that at the dosages commonly used by people for life/health extension, these findings re: OA, are not relevant to humans. This might be confirmed by the fact that there are no such OA effect reports in humans, regardless of sex in the literature, including the organ transplant literature.

Bottom line, I am not concerned that rapamycin would have a negative impact on OA in humans in dosages typically used for life/health extension. YMMV.

4 Likes
#9

I’m admittedly an outlier but my ADD brain finds trouble getting through any of them, even if it’s a summary of an video, and it’s the same reason I have trouble getting through a study.

1 Like
#10

I developed a mild form of osteoarthritis on both of my hands. Started last year rather abruptly, which at first was dismissed as overusing my hands holding the wheel when on a long driving trip. Unfortunately the pain didn’t go away and recent X-ray showed mild osteoarthritis. I am not fully convinced that it’s rapamycin to blame which I’ve been taking for 15 years. It could be age and wear and tear factor, genetics, transplanted kidney which may contribute to calcium accumulation. The strange thing was that it came suddenly and changes in mobility occurred very quickly within days. I’m on a minimum rapamycin dose for kidney transplant.

5 Likes
#11

Someone help me out because I’m not good at converting dosages from marmosets into humans.

Do these daily dosages in marmosets translate to human intermittent dosing? If they do, it’s a relevant data point. If they don’t, I’d throw the study away.

4 Likes
#12

The 1mg/kg dosing used in marmosets is a high, but not impossible, dosing for humans. It approaches the 10mg/day dosing used in some cancer studies I’ve seen.

Here are the dose translations from the other rapamycin studies - mice metabolize rapamycin much faster. I would say we really don’t know the details of how the ITP and other trials showing the longevity impacts of rapamycin on mice lifespans translate to humans… the tests just haven’t been done (and may never, given that rapamycin is a generic medication so unless a government or nonprofit steps up to do the clinical trials, there is no financial incentive for a company to do the clinical trials).

My belief is that the best we can do is test different dosing levels and timing approaches on our own bodies, track blood test and functional results closely - and modulate in a way that we think will optimize results.

If you have not done it yet, I recommend you read the entire “Rapamycin Dosing” thread - it has a lot of good information in it and feedback from different people: What is the Rapamycin Dose / Dosage for Anti-Aging or Longevity?

I think daily dosing in mice is roughly equivalent to about once every 4 days or so in human terms given the speed that mice metabolize rapamycin is about 4 times faster.

Sirolimus
Dose		 Mouse
mg/kg/day
Dose		 Mouse:
Blood/Sirolimus
Level		 Human
mg/kg/day
Dose		 Dose for 60kg Human Daily Dose adjusted for longer half-life (/4)
4.7ppm ∼2.24 3 to 4 ng/mL 0.182 mg/kg 10.92 mg 2.73 mg
14ppm ~6.67 9-16 ng/mL 0.542 mg/kg 32.54 mg 8.135 mg
42ppm ~20 23-80 ng/mL 1.626 mg/kg 97.56 mg 24.39 mg
126ppm ~60 4.878 mg/kg 292.68 mg 73.17 mg
378ppm ~180 45 to 1800 ng/mL 14.634 mg/kg 878.04 mg 218 mg
Sirolimus
Dose		 mg/kg/day
Dose		 Blood/Sirolimus
Level		 Male Median LS Increase Female Median LS Increase
4.7ppm ∼2.24 3 to 4 ng/mL 3% 16%
14ppm ~6.67 9-16 ng/mL 13% 21%
42ppm ~20 23-80 ng/mL 23% 26%

Based on the FDA animal to human dosing conversion guide here.

Note: ½ life for sirolimus in mice is approx. 15 hours, vs. approx. 62 hours in humans. So, mice metabolize sirolimus approximately 4 times faster than humans.

Related discussions:

5 Likes
#13

Folks - I emailed Adam Salmon. Do people here think it would be worthwhile to do another interview at this time with Adam, to cover this paper and get an update on the Marmoset Longevity study?

Let me know questions you have, if you would like to see another interview of Adam.

He says they are almost done with the study (most marmosets have now died, I guess), but they are having delays on the pathology reports. It’s not clear when the Marmoset longevity paper will be published, but from the sounds of it at least another 6 months out.

11 Likes
#14

Lara, the same sudden arthritis takeover happened to me. After a few months off, I had been back on 3 mg rapamycin with grapefruit juice every 10 days for about a month. I decided to take magnesium – it was probably 500 mg – threw the bottle away, so don’t know for sure – and full-body arthritis hit. One day I was normal and the next day I had pain all over.

What had been mild arthritis in knees and shoulders, overnight became painful in all joints – hips wrists, elbows, everything. My hands swelled and I couldn’t close my fingers. All movement hurt. Rolling over in bed or getting out of bed was an ordeal. I stopped the magnesium, gave it a few days, but nothing changed. Stopped all supplements – no rapamycin, multivitamin, fish oil, nothing.

Gave that 3 months, but still no change. Guessing that it was probably inflammation, started back with a 3 mg plus grapefruit juice dose. 3 hours later, the pain started easing. 3 days later it started coming back. Took 3 mg and grapefruit juice again. This time it was 4 days between doses, and pain was much less when it did come back. Today is the 6th day since the last dose, and remnants of the pain and swelling are still around, but nearly back to pre-magnesium days. I’m going to see if I can get back on a 3 mg plus grapefruit juice dose every 10 days, but won’t wait if the pain and swelling start to come back.

Since I was on rapamycin when it started, I’m puzzled that it didn’t prevent the problem. But there’s no question that it’s stopping it.

I’m 80 and can expect decline, but this was so sudden and dramatic that it really got my attention. Looking online, I see that others have similar problems with magnesium.

5 Likes
#15

Wow! I never thought to connect it to magnesium. When the pain started I was on magnesium twice a day. I am still on the same dose of 150 mg. Do you think it could be magnesium? My pain is localized only to my thumb/wrist junction. More on the right and less on the left.

3 Likes
#16

I don’t know if it was the magnesium, but that was the first time I had ever taken it. The timing and reading about the experiences of others made me think that was the cause.

The whole thing is bizarre. What I know for sure is that rapamycin is fixing it.

5 Likes
#17

Definitely.
I would be interested in finding out how the arthritis phenotype was different from human age-rated arthritis.
Isnt it true that Marmoset have higher and faster back ground musculoskeletal ageing and OA to begin with as compared to humans? Rodents have more robust cartilage biology.
What clinical and radiological/imaging signals should we be looking for in humans?

2 Likes
#18

So that case study mentioned OA improvement in a Rapa study, should be evaluated for the other medications and supplements involved.

I wonder if the individual from the case study was on taurine.

1 Like
#19

Do you suspect that taurine could be at play here?

1 Like
#20

Possibly. We have anecdotal reports of it causing recovery similar to what was seen in the case study. The case study mentioned the patient was taking unlisted supplements for longevity. Possible Taurine was on the list.

2 Likes