18 replies to this topic

[Michael]

My long-promised, long-delayed post on the opposing outcomes of the nonhuman primate CR studies at the Wisconsin National Primate Research Center (WNPRC) and at the National Institute on Aging (NIA) is finally up at the SENS Research Foundation CSO Blog:
CR in Nonhuman Primates: A Muddle for Monkeys, Men, and Mimetics

[Fixed link, and deleted AgeVivo's correction post].

Edited by Michael, 11 May 2013 - 11:54 AM.

[James Cain]

Holy crap that is epic. I'm very interested in your thoughts and will spend the weekend reading it. Thanks for the time and effort!

[AgeVivo]

Dear Michael,

The most compelling of these(40) reported a very surprisingly high level of such variability amongst 41 recombinant inbred mouse strains, with CR shortening lifespan in more strains than it lengthened it. Analysis of the relationship between substrain response to CR and other genetically-linked traits found that successful adaptation to CR was associated with greater ability to maintain fat mass under the diet,(41) which is consistent with much a more modest effect of adipose maintenance observed in previous studies in individual animals within CR cohorts.(43-45) However, the generalizability of the high level and opposing directions of response to CR in this study(40) is rendered unlikely by the inclusion of the DBA/2 strain

Perhaps there is something I do not understand: I wouldn't say that the specific case of DBA/2 would make a consistently observed link unlikeliky to use as a first estimation? It seems to me that DBA/2 look more like an exception (perhaps some other explanation may exist for that strain) given the consistency you indicate.

*If* we were to use the consistently observed link (between fat mass maintenance and CR adaptation success) as a rule of thumb estimation,
- to what degree would it explain the CR adaptation difference between WNPRC and NIA studies?
- would humans be more like the Chinese or Indian monkeys?
Thank you!

Edited by AgeVivo, 11 May 2013 - 09:44 AM.

[AgeVivo]

reducing protein intake to approximately the Recommended Daily Allowance of 0.8 g/kg allowed IGF-1 levels to fall to within the lower end of the reference range, or even beneath it (59, and author's congruent anecdotal observation)

Perhaps I read badly: is there some end of sentence lacking?

the inclusion of the relatively large number of old-onset animals in such a pooled analysis of the NIA animals may mask the true potential of CR against cancer

In any of the 2 interpretations that you consider the most likely, you think that if the design of the NIA study had been more optimal in terms of age-at-start and past husbandry, we would have seen a better effect of CR, right ?

Edited by AgeVivo, 11 May 2013 - 09:43 AM.

[Michael]

The most compelling of these(40) reported a very surprisingly high level of such variability amongst 41 recombinant inbred mouse strains, with CR shortening lifespan in more strains than it lengthened it. Analysis of the relationship between substrain response to CR and other genetically-linked traits found that successful adaptation to CR was associated with greater ability to maintain fat mass under the diet,(41) which is consistent with much a more modest effect of adipose maintenance observed in previous studies in individual animals within CR cohorts.(43-45) However, the generalizability of the high level and opposing directions of response to CR in this study(40) is rendered unlikely by the inclusion of the DBA/2 strain

Perhaps there is something I do not understand: I wouldn't say that the specific case of DBA/2 would make a consistently observed link unlikeliky to use as a first estimation? It seems to me that DBA/2 look more like an exception (perhaps some other explanation may exist for that strain) given the consistency you indicate.

Sorry: I don't understand your question.

*If* we were to use the consistently observed link (between fat mass maintenance and CR adaptation success) as a rule of thumb estimation,
- to what degree would it explain the CR adaptation difference between WNPRC and NIA studies?

Because they aren't comparable studies at the level of the CR itself, there's really no connection there: the WNPRC "CR" animals weren't really on CR -- they merely prevented from overeating. However, differential maintenance of body fat might, hypothetically, explain a differential response to CR between the Indian- and Chinese-sourced animals (or individual animals irrespective of broad geographical-genetic background), if such a differential response exists. Again, we don't know whether such a differential response exists, because the NIA authors have not broken down outcomes by background of the animals, and because of the severe confounders of the lack of background info on the Chinese/military-sourced animals and of the fact that Chinese/military-sourced females had so much endometriosis (and the specter of other health problems, absent an understanding of why they were differentially affected).

- would humans be more like the Chinese or Indian monkeys?

I've never been housed on a military base and subjected to unknown experimentation . Look: we don't know what, if any, impact Chinese or Indian origin per se had on survival or disease outcomes, so we certainly don't know if humans are more like the Chinese or Indian monkeys -- or, indeed, whether humans of particular ethnic origin have differential response to CR similar to what might be hypothesized to have happened with Chinese vs Indian monkeys.

reducing protein intake to approximately the Recommended Daily Allowance of 0.8 g/kg allowed IGF-1 levels to fall to within the lower end of the reference range, or even beneath it (59, and author's congruent anecdotal observation)

Perhaps I read badly: is there some end of sentence lacking?

No: you're being referred to reference (59) of the original post, and to my own congruent anecdotal observation.

the inclusion of the relatively large number of old-onset animals in such a pooled analysis of the NIA animals may mask the true potential of CR against cancer

In any of the 2 interpretations that you consider the most likely, you think that if the design of the NIA study had been more optimal in terms of age-at-start and past husbandry, we would have seen a better effect of CR, right ?

No: in the first interpretation ("Diminishing Returns"), it doesn't matter how well you tweak the CR protocol at NIA, because CR just doesn't translate to primates: once you get a basic, lean physique, there's no advantage to further reducing Calorie intake.

[AgeVivo]

Dear Michael,

1. sorry I'm reading my questions above and realize... how tired I was at 3 am; hem..

- the answer of my first question was in your text (I was asking why you thought that DBA/2 makes the conclusion of (40) unlikely generalizable, it is because all crosses studied in that article were 'contaminated' by DBA/2)
- the next questions were in fact to ask how this generalization of the conclusion of (40) would comparely predict the impact of CR (true CR: no war sugar burger considerations here) in indian monkeys/ chinese monkeys/ humans, based on known ability to maintain fat under a diet. But I understand why (40) is likely inapropriate for that and indeed can we have comparable "fat maintainance under diet" between monkeys and humans...
- ...I had not seen the semicolon before "reducing"

2. The only potential valid point is the last one, which was floating in my head and therefore badly explained:
Somehow you make the point that the NIA data statistically rejects the "Disminishing Returns" Hypothesis :

young-onset CR afforded very substantial protection against malignant neoplastic disease at NIA (6/46 AL cases, of which 5 were fatal, vs. 0/40 cases in CR), while old-onset CR did not (6/18 vs 7/17).

So you statistically make the point for cancer. Should the point be valid for lifespan as well? Then I guess, at the risk of caricaturing,
out of the 2 studies we are mostly left with one likely interpretation, a "Dose-Response" one: CR extends lifespan in young-onset monkeys, at least via cancer protection
If I haven't bothered you too much with my 0-cent questions I'd be interested in knowing if that makes senses to you who have the many contextual details in mind.

[Michael]

1. sorry I'm reading my questions above and realize... how tired I was at 3 am; hem..

The only potential valid point is the last one, which was floating in my head and therefore badly explained:
Somehow you make the point that the NIA data statistically rejects the "Disminishing Returns" Hypothesis :

young-onset CR afforded very substantial protection against malignant neoplastic disease at NIA (6/46 AL cases, of which 5 were fatal, vs. 0/40 cases in CR), while old-onset CR did not (6/18 vs 7/17).

So you statistically make the point for cancer.

We-e-el, that's not really a statistical point, as I imagine you realize: it just looks, prima facie, 'suspicious,' and when you combine it with the long latency period of cancer, it seems a reasonable hypothesis that CR "worked" against cancer in these animals — given enough time. (One fly in that ointment, it should be said (and I maybe should'a said in the post) is that Spindler found the late-life effect against cancer to be fairly rapid-onset in mice.(1,2) There might be differences on that front because the sheer number of cells required for a tumor to kill a mouse is smaller than what's needed to cause problems for humans or primates, and it needn't metastasize — but, of course, that could affect the efficacy of CR as well as the time-course.

Should the point be valid for lifespan as well? Then I guess, at the risk of caricaturing,
out of the 2 studies we are mostly left with one likely interpretation, a "Dose-Response" one: CR extends lifespan in young-onset monkeys, at least via cancer protection

We can't make he point for lifespan, because even when you look just at younger-onset animals, the survival curves don't seem to have any kind of signal in them:

Total (a) and age-related (b) mortality.

... if anything, the CR animals kind of seem to have fared more poorly, tho' we don't have the numbers and they tell us there wer no significant differences. It would sure be nice to have one without the 'contamination' of the 'Aberdeen' monkeys, of course. But it would really be nice to have had more animals and more rigorous CR from the get-go.


References
1: Dhahbi JM, Kim HJ, Mote PL, Beaver RJ, Spindler SR. Temporal linkage between the phenotypic and genomic responses to caloric restriction. Proc Natl Acad Sci U S A. 2004 Apr 13;101(15):5524-9. Epub 2004 Mar 25. PubMed PMID: 15044709; PubMed Central PMCID: PMC397416.

2: Spindler SR, Dhahbi JM. Conserved and tissue-specific genic and physiologic responses to caloric restriction and altered IGFI signaling in mitotic and postmitotic tissues. Annu Rev Nutr. 2007;27:193-217. Review. PubMed PMID: 17428180.

[scottknl]

... if anything, the CR animals kind of seem to have fared more poorly, tho' we don't have the numbers and they tell us there wer no significant differences. It would sure be nice to have one without the 'contamination' of the 'Aberdeen' monkeys, of course.

I was under the impression that the aberdeen group 20 early deaths were statistically removed from the analysis results. Is this one of the instances when the data was included? Or were you talking about the remaining 6? (20 out of 26 females die before the age of 12)

It seemed apparent that this cohort was differentially affected in terms of long-term health, and thus, an indicator variable ... was included in most analyses to control statistically for the effects of these animals on the outcomes of interest.

[kismet]

Well done!

Two questions.

1.
You speculate that the diagnostic criteria for diabetes differed between the studies? Were the authors really unable or unwilling to confirm or refute this?
This is *really* important when assessing morbidity and authors from both studies should be aware of that problem. IMHO the NIA researchers should have reported any such discrepancy when doing their between-study-comparison of morbidity, this is VERY bad style to put it most kindly.


2.
You speculated that the NIA researchers may have a hard time detecting a modest, real effect over the noise. I would go even farther:
Subspecies differences, divergent life histories, early life pair-housing, mixing of juvenile- and young-onset animals, natural ingredient base of their diet, uneven CR implementation (estimating energy requirements by formula) and higher CR-level* all conspire to decrease the power of the NIA cohort and increase variability. The authors were already worried by statistical power in their 2003 interim report, although, perhaps for other reasons. (Mattison 2003). In the worst case this could render the data useless.

*if some of the animals were starving, see below

Whereas the WNPRC researchers have done well with their interventionist approach: a “5% body fat rule [my term]“ to prevent animals from starving and drug treatment as needed (e.g. treating endometriosis). Conversely, this also implies high variability in the NIA cohort.

Keeping with this idea, how do you explain that some animals were starving in the WNPRC cohort even though all other evidence points to a modest level of restriction?
(ramsey 2000):
“Most of the animals adapted well to this DR regimen. For two animals, however, the 30% restriction was too severe and their body fat decreased to 3% in association with hair loss and a clinically ill appearance. As a result of this experience and following discussions with the Wisconsin Regional Primate Research Center veterinary staff, it was concluded that greater than 3% body fat is required to maintain good health. Thus, in March 1996, the decision was made to increase the food intake of any animal that fell to this level. To allow a margin of safety, we also estimated the amount of fat (above 3%) that would be required to meet energy demands for three days and calculated a minimum fat requirement for each animal. This resulted in a minimum body fat requirement of approximately 5%.”

Would this not imply that even more animals were starving in the NIA cohort with their hands off approach and lower absolute caloric intakes? Even though the level of restriction might have been too low on average?!

Evidence of this can be seen in the higher SEM (standard error of the mean) of reported kcal-intakes and bodyweights in the NIA cohort, but unfortunately this might be due to different measurement methods…

(I skimmed/read your post twice and ctrl+F'd for "starvation", but I did not see you address this issue. I am too excited by this question to focus on re-reading your post)

Edited by kismet, 17 May 2013 - 07:04 PM.

[kismet]

3. this needs sorting out, have not looked in detail:
...one possible background combination is the C57BL/6xDBA/2 F1 hybrid, consisting of two genetically diverse lines. Of the seven genetically placed inbred groups of laboratory mice (Mus musculus), C57BL/6 is from group 4 and DBA/2 is from group 6 (Petkov et al., 2004). This hybrid has been shown to be long lived, with a significant increase in lifespan on a caloric restricted diet (Turturro et al., 1999).

http://onlinelibrary...09.00491.x/full

[mwestbro]

This is a fascinating discussion, but I'd like to ask an obvious question. Michael, has this changed your personal practice of CR? Is CR still worth it to you? I don't believe these results disprove the value of CR for human beings, but they do make the calculus of benefits vs. costs a lot more uncertain. Are you reconsidering the value of CR for you personally?

Mike

[kismet]

This is a fascinating discussion, but I'd like to ask an obvious question. Michael, has this changed your personal practice of CR? Is CR still worth it to you? I don't believe these results disprove the value of CR for human beings, but they do make the calculus of benefits vs. costs a lot more uncertain. Are you reconsidering the value of CR for you personally?

Mike

I believe that was answered in part(?) on the cr-list: http://arc.crsociety...ad.php?2,215643

There is also some more discussion at the link.

Edited by kismet, 28 May 2013 - 06:48 PM.

[mwestbro]

Thanks, Kismet. And to Michael, for the original post.

Mike

[scottknl]

I talked to some people involved in these studies recently and found out a few interesting things.
1) If you didn't include the "aberdeen" group of sick females in the NIA study, the CR results would have been much better for overall mortality, but may still not have reached statistical significance because of the size of the group.

2) The level of actual CR really produced 3 levels of restriction if you take the results of the two studies. WNPRC CR, NIA Controls and NIA CR from least to most with all 3 groups under some kind of CR stress.

3) This species of monkey was awful to work with and was ill suited to be a subject for a CR study because of it's aggressive behavior and thus it's need to be expensively housed in separate cages.

4) If you didn't include the "aberdeen" group of sick females, the overall statistical power of the study would have made it difficult to come to any conclusion at all.

5) The behavior of the subjects resulted in a level of CR that changed significantly throughout the lifespan of the subjects for both CR and control groups.

6) It was a mistake to include 2 different strains of monkeys in the study and a suggestion was made that the level of CR may have been too severe for one of the species.

7) The scientists were surprised at what they perceived as unfair treatment in the popular press
eg. saying CR doesn't work when in fact it showed significant protection vs cancer, heart disease and diabetes in addition to better muscle preservation, better brain preservation and better immune system preservation.

8) Most surprising of all was that the NIA study was intended to be only 3 - 10 years study of bio-markers of aging and was later extended to include more animals and so some of the aspects of the study were not optimized simply because they were not planned for.

Much of this has been said before in what Michael wrote and in other discussion, but it was a bit different to hear it directly from people who have spent a big chunk of their lives working with these studies.

Edited by scottknl, 12 June 2013 - 05:19 AM.

[kismet]

6) It was a mistake to include 2 different strains of monkeys in the study and a suggestion was made that the level of CR may have been too severe for one of the species.

Do you have any details? Was this off the record? This would support my speculation in 2.

4. comment:
Why were the studies designed so terribly? Did we really know so little in the 80s & 90s? I presume we had some of the best people in the field working on these studies? Nonetheless, there is so little standardization and cooperation between the NIA & WNPRC group (e.g. in studied outcomes). Furthermore, the data is dispersed and almost impossible to understand for most non-biogerontologists working in related fields.

That is in addition to the methodological issues already mentioned.

[Mind]

The studies have changed the calculus for me. I am 42. I have practiced CR a little over the last 10 years - for a few months at a time. I will probably rely more on maintaining a lighter BMI and exercise to stay healthy, rather than meticulous CR.

Despite the very valid criticisms of these two studies, CR was supposed to be so powerful (proclaimed by the practitioners) that I would have expected a lifespan extension no matter the crappy design. It seems likely that CR is beneficial for primates, but not to the extent of "lower animals".

[scottknl]

6) It was a mistake to include 2 different strains of monkeys in the study and a suggestion was made that the level of CR may have been too severe for one of the species.

Do you have any details? Was this off the record? This would support my speculation in 2.

4. comment:
Why were the studies designed so terribly? Did we really know so little in the 80s & 90s? I presume we had some of the best people in the field working on these studies? Nonetheless, there is so little standardization and cooperation between the NIA & WNPRC group (e.g. in studied outcomes). Furthermore, the data is dispersed and almost impossible to understand for most non-biogerontologists working in related fields.

That is in addition to the methodological issues already mentioned.

Yes, this was an off the record conversation and taking some of the pieces of a presentation to the CR Society last week. The reasoning goes that the WNPRC study worked out better because the level of CR could be tuned to the "CR-tolerance" of the single species of primate they used. The NIA study was never intended to go as a full term longevity study - instead they just wanted to test out some biomarkers vs aging over a 3 - 10 year time-frame. Later on they secured funding to extend the study and increase the number of animals. At this point they had to move fast to spend the money that was allocated, and wound up with a different species of primate (chinese vs indian origin I think).

There were no apologies for the diet in the WNPRC study which is the chief deficiency in it. Neither study really seem to place a lot of emphasis on the female subjects living much less than the male subjects (like a different kind of species really). The NIA study with it's aberdeen group, chinese vs indian sourced primates and varying level of CR for the controls and CR groups was pieced together with changing objectives over time and varying levels of funding and thus expectations. I think the data was poor in the NIA study because the design was poor and changed a lot over time.

The only takeaway's from these studies for me is that CR delays the diseases of aging in both studies effectively. It also delays degeneration of muscle, teeth, brain and immune system function. Whether CR has an effect on longevity was shrouded by the issues in each of the studies and one can't conclude anything definitive.

[scottknl]

The studies have changed the calculus for me. I am 42. I have practiced CR a little over the last 10 years - for a few months at a time. I will probably rely more on maintaining a lighter BMI and exercise to stay healthy, rather than meticulous CR.

Despite the very valid criticisms of these two studies, CR was supposed to be so powerful (proclaimed by the practitioners) that I would have expected a lifespan extension no matter the crappy design. It seems likely that CR is beneficial for primates, but not to the extent of "lower animals".

You're still a bit young. Once you get some real evidence of the speedup in aging that happens in the 40's and 50's maybe you'll reevaluate.

In the NIA study the graph of age vs calorie consumption has the lines converging as the animals age. One can't expect CR results if one doesn't test CR.
You can get a secondary confirmation by looking at the body weights of the subjects. NIA CR and Control groups were much closer in body weight than in the WNPRC study. And the female body weight lines in the NIA study were almost on top of one another.



I don't doubt that there is some degradation between the level of CR that can be sustained by one long lived species vs another, but it's all we have except for maybe rapamyacin. Even a 20% extension of lifespan is considerable in humans and if one has to limit by 30% to get it, I think the trade off is worth it if it comes with good health and compression of morbidity.

Attached Files

•  NIA_Calories2.jpg   124.76KB   16 downloads

[Mind]

The studies have changed the calculus for me. I am 42. I have practiced CR a little over the last 10 years - for a few months at a time. I will probably rely more on maintaining a lighter BMI and exercise to stay healthy, rather than meticulous CR.

Despite the very valid criticisms of these two studies, CR was supposed to be so powerful (proclaimed by the practitioners) that I would have expected a lifespan extension no matter the crappy design. It seems likely that CR is beneficial for primates, but not to the extent of "lower animals".

You're still a bit young. Once you get some real evidence of the speedup in aging that happens in the 40's and 50's maybe you'll reevaluate.

In the NIA study the graph of age vs calorie consumption has the lines converging as the animals age. One can't expect CR results if one doesn't test CR.
You can get a secondary confirmation by looking at the body weights of the subjects. NIA CR and Control groups were much closer in body weight than in the WNPRC study. And the female body weight lines in the NIA study were almost on top of one another.



I don't doubt that there is some degradation between the level of CR that can be sustained by one long lived species vs another, but it's all we have except for maybe rapamyacin. Even a 20% extension of lifespan is considerable in humans and if one has to limit by 30% to get it, I think the trade off is worth it if it comes with good health and compression of morbidity.

Each of us has a different calculus. I applaud the CR practitioners and their advocacy, I really do.

For me, I don't carry around much white adipose tissue (body fat 10%), and I get plenty of exercise. I know you follow this thread about exercise and aging, but did you see this amazing study? So I am slim and might even be slightly CR (can't say for sure because I don't tabulate my calories every day) and I exercise. Considering my age, technological progress, and the powerful effects of exercise, and the "muddle" of the primate CR studies.....being religiously devoted to CR is not in my "calculus". In contrast to many CR people who find the practice very pleasurable, it makes me cold, irritable, and hungry. So there are some downsides on quality of life.