The search for compounds that mimic the health and longevity benefits of CR without requiring actual calorie reduction is an attractive proposition in current nutrition research. Although previous studies have suggested that metformin supplementation provides a transcriptional response highly consistent with actual CR (4
), the results from this study do not support the proposition that metformin is a CRM, at least at the one dose investigated. Previous studies published over the past three decades have shown that metformin and other related biguanides can extend life span in rodent models of disease, including cancer, hypertension, and Huntington’s (23
). However, the use of a normal, highly researched rodent model for basic life-span assessment was indicated. The previous reports of life-span extension with biguanide supplementation may be related to a reduction of early mortality associated with the disease condition in strains heavily prone to disease, especially cancer, a situation resulting in relative life-span extension compared with untreated controls. A recent study suggests that this may not be the sole explanation as life span was extended in female outbred Swiss-derived SHR mice, even without a reduction in tumor onset or incidence in the treated mice, hinting at the possibility of metformin altering aging pathways independent of disease states (24
). Even so, the use of this particular rodent strain (male F344 rats) may have contributed to the lack of an observed benefit. However, no other reports are available that conclusively demonstrate life-span extension in a “normal” rodent model with metformin supplementation. Importantly, the dose of metformin used in the present study was efficacious for disease prevention and life-span extension in previous rodent studies, although it should be noted that this dose is ~10-fold higher than the maximum daily dose used in human treatment (milligrams per kilogram body weight) (33
). Recently, metformin supplementation (50 mM dose) was shown to increase the median life span (but not maximum) of Caenorhabditis elegans
, although 10 or 100 mM doses showed no significant life-span benefits (43
One limitation of the current study is the lack of a robust CR response for extension of maximum life span (; and ). CR initiated at 6 months of age in male F344 rats was previously reported to increase life span (44
). Therefore, the possibility that the strain of rodent used does not possess the ability to respond to CR seems unlikely. Rather, the reduced CR response could arise from a sampling artifact resulting in a reduced significance of the life-span effect for this particular set of rats. As such, the reduced efficacy of CR in the current study might provide a partial explanation for the lack of a significant increase in life span with MET treatment. Additionally, the type of diet used in the current study (NTP-2000) was optimized for health and longevity benefits when used in AL feeding but has not yet been reported for life-span extension during CR. Comparing the observed life-span measures with previous studies, the mean and maximum life span of the AL CON group is longer, whereas the mean and maximum life span of the CR group is shorter than was previously reported () (45
). Although these were male F344 rats, a different supplier and diet were used in the previous report compared with the current study. Thus, there is some question as to whether the reduced protein concentration of the diet would be adequate for long-term studies in a restricted state (38
There were two early deaths in the CR group within 3 weeks of group randomization and CR initiation (Days 20 and 21). The life-span results included these two early deaths (–). Censoring these two animals had no effect on the significance of the life-span comparison with the exceptions of a significant increase in the 50th quantile of life span in the CR group compared with CON and a significant increase in the mean life span of the CR group (901.6 ± 112.9 days after censoring; ). When including the two early deaths in the CR group, CR still delayed early mortality compared with CON (; —25th quantile), whereas MET did not. Therefore, any MET life span response is not as robust as CR in the current investigation. Importantly, although life span was not extended with chronic metformin supplementation, the MET group life span was also not reduced compared with CON (; ). Regarding the Gompertz α parameter, it is important to note that the small sample size makes the model susceptible to early deaths observed in the CR, MET, and PF–MET groups (). This causes the “aging rate” to appear elevated in the CON group, despite there being no significant difference in mean, maximal, or overall life span among groups ().
In addition to the dampened CR response, metformin supplementation did not significantly affect glucose/insulin levels in the current study. As such, the tested concentration may have been insufficient to trigger a full CR-like response, resulting in altered life span. Despite this, previous research using similar or lower concentrations of biguanides has reported beneficial effects (24
), although the method of administration differed between studies. Additionally, the metformin concentration utilized in the diet is approximately 10 times that of the highest dose used in human treatments (33
), implying that any increase necessary to observe life-span benefits is questionable for a human application.
Metformin supplementation resulted in reduced body weight, despite similar food intake, during the middle of the study (A and B). A reduction of adipose tissue is proposed to contribute to longevity-associated benefits of CR (47
). Body composition measures were not acquired in the present study, making it impossible to determine if reduced body fat percentage accompanied the reduction in weight. Future studies may benefit from the inclusion of lean and fat mass measures to more fully explore the relationship between metformin supplementation on body composition in relationship to longevity.
The four aspects of a CRM were previously outlined (13
) as (i) metabolic, hormonal, and physiological effects similar to CR, (ii) no significant reduction of long-term food intake, (iii) activated stress response pathways similar to CR, and (iv) beneficial effects on longevity and reduction of age-related disease. Using these criteria to assess the effect of metformin supplementation in this study, there were no significant reductions in insulin, glucose, or CBT with metformin ( and ). However, the CR group was also not consistently reduced in these same measures ( and ). Long-term food intake was not significantly reduced with metformin supplementation, although body weight was slightly reduced (). There were no specific measures of stress response or protection performed in the present study, and necroscopic findings were similar among groups (coded “mass”: CON-8, CR-9, and MET-11). Finally, with multiple life-span analyses, no significant life-span extension was observed with metformin supplementation, a critical component for the validation of a CRM. In contrast with previous reports using disease models in rodents and transcriptional profiling, these results challenge the proposition of metformin supplementation at the specified dosage acting as a bona fide CRM.