I agree that the hypothesis is shaping up nicely, I would just like to add some nuance here and there. First of all, I believe that lipid composition either is or is not involved in aging, be it aging of ad libitum or CR-animals; it must be a general mechanism. And, it follows, that in principle it must be testable and provable in the ad lib state, even if this is more difficult than for CR animals.
Regarding the comparative approach: not as clear as it looks
All the comparative and correlative data linking increased unsaturation with decreased lifespan is (perhaps critically) flawed as explained by Speakman (2005):
Larger animals are generally more long-lived. Now, if we imagine that larger animals have decreased unsaturation to regulate, for instance, metabolic rate*, this then would produce a spurious relationship. And, indeed, the correlations do not persist after adjustment for bodyweight (1, 2). The counterpoint has always been that such adjustment is nonsensical given small effects and sample sizes because "statistical adjustment for mass dramatically reduces the statistical power for the regression or correlation." (e.g. ref. 3)
Nonetheless, I'd certainly describe the correlative data as still controversial.
Researchers in the field, Barja, Hulbert (presumably) and my colleague Valencak are sitting on swaths of data, which taken together might one day allow us to overcome these statistical problems. They just need to realize how important cooperation on a pooled analysis is.
* reasons to reduce unsaturation do exist since higher unsat. may lead to higher metabolic rate. And most naively and succinctly put: “If an elephant’s metabolism per unit volume were the same as that of a mouse, the elephant’s skin would burn up.”
Lack of Experimental Validation: direct contradiction of the hypothesis?
I have blogged about this study (4) before:
Your counterpoint is, of course, well taken and appears to have been validated by López-Domínguez et al. 2014 (emphasis added):
An update on mitochondrial peroxidation index, DHA content and lifespan
Valencak and Azzu just recently published a paper on this topic (5). It contains – what appears at first glance to be* – the largest pooled or combined analysis of mitochondrial lipids with lifespan. A few noteworthy points:
Only mitochondrial DHA content shows any associations with lifespan at all, and this is only true for muscle and not liver. Of course, it’s striking how well this fits our working hypothesis: “post-mitotic tissues accumulate mitochondrial deletions, that's where membrane generated ROS in close proximity to the mitochondrial genome would matter most“. It remains to be seen if this is more than idle speculation. It looks like studying liver samples just because they are easily accessible might have been quite foolish!
*I say at first glance because there is no overlap between the references you give on this topic and the ones Valencak and Azzu used as per Table 2; what exactly is going on here? Your references are liver, their table is not clear on this point AND they claim liver showed no association(!), this may require more reading than I have done so far.
In principle, though I do not favor this hypothesis, the correlative data might be coincidental and not causal. The feeding studies could be artifactual due to giving unphysiologic amounts of n3 and/or n6 fatty acids that interfere with other aspects of physiology.
1. Correlations between physiology and lifespan--two widely ignored problems with comparative studies. Speakman JR. Aging Cell. 2005 Aug;4(4):167-75. Review.
2. Aging Cell. 2007 Feb;6(1):15-25. Epub 2006 Dec 5. N-3 polyunsaturated fatty acids impair lifespan but have no role for metabolism. Valencak TG1, Ruf T.
3. J Gerontol A Biol Sci Med Sci. 2014 Feb 12. [Epub ahead of print]Fibroblasts From Long-Lived Rodent Species Exclude Cadmium.Dostál L1, Kohler WM, Penner-Hahn JE, Miller RA, Fierke CA.
4. Lifelong treatment with atenolol decreases membrane fatty acid unsaturation and oxidative stress in heart and skeletal muscle mitochondria and improves immunity and behavior, without changing mice longevity. Gómez A, Sánchez-Roman I, Gomez J, Cruces J, Mate I, Lopez-Torres M, Naudi A, Portero-Otin M, Pamplona R, De la Fuente M, Barja G. Aging Cell. 2014 Jun;13(3):551-60. doi: 10.1111/acel.12205. Epub 2014 Feb 26.
5. Longev Healthspan. 2014 Mar 3;3(1):3. doi: 10.1186/2046-2395-3-3.
Making heads or tails of mitochondrial membranes in longevity and aging: a role for comparative studies.
Valencak TG1, Azzu V.
Edits for clarity