Technical Comments

Comment on "Long-Lived Drosophila with Overexpressed dFOXO in Adult Fat Body"

Science  04 Feb 2005:
Vol. 307, Issue 5710, pp. 675
DOI: 10.1126/science.1104721

Aging can be retarded when insulin-like signals are reduced in a limited set of tissues in several animal models (1). Current research aims to determine which tissues are necessary and sufficient for mediating effects on life span (24). Giannakou et al. (5) assert that the adult fat body is the target tissue in Drosophila melanogaster. Giannakou et al. analyzed survival of adults containing the conditional P{Switch} galactose 4 (Gal4) driver S1106, where RU486 can activate Gal4 in the abdominal and thoracic fat body (3, 6), and UAS-dFOXO, the insulin-like responsive transcription factor (79). Giannakou et al. (5) state, “Induced expression of dFOXO in the fat body from the onset of adulthood increased life span... by 20 to 50%.” This conclusion is based on a demographic error, which is obscured because the data are presented as cumulative survivorship. In two of three replicate experiments, treatment to induce dFOXO did not increase life span; rather, young control flies experienced excessive mortality in a single week, and this inadvertently depressed their survivorship. Control flies depicted in figures 1B and 1C in (5) died at an unusually elevated rate between days 19 and 27. For instance, 57 control females died on day 22. Because these trials were conducted at the same time (10), this mortality in control females appears to represent a period artifact. The cause of excess deaths is unknown but may have occurred in controls alone because these flies were housed separately and fed a unique diet. Before day 19 and after day 27, there were no meaningful differences in mortality between control and treatment females. Furthermore, when days 19 to 27 are censored, there are no significant differences in overall mortality among the control and treatment cohorts [log-rank test on data from (5): figure 1B, χ2 = 2.84, P = 0.092; figure 1C, χ2 = 3.07, P = 0.080]. Likewise, differences in life span are small and are not significant. [Median from eclosion, with lower and upper 95% confidence interval: figure 1B, control, 38 days (33, 40), +RU486, 40 days (38, 41); figure 1C, control, 36 days (32, 39), +RU486, 40 days (38, 43). Mean from eclosion, with standard error: figure 1B, control, 38.1 days (0.90), +RU486, 39.9 days (0.79); figure 1C, control, 37.0 days (1.09), +RU486, 40.2 days (0.88)]. Thus, contrary to the assertion of Giannakou et al., the data from these replicate trials provide no evidence that longevity is conferred by overexpression of dFOXO via the tissue-specific driver S1106. Differences as reported for these data arise because an early failure in the control cohort reduced their mean longevity and cumulative survivorship.

Fig. 1.

Mortality rate (lnμx) calculated from data of Giannakou et al. (5). (A) Corresponds to figure 1B in (5). (B) Corresponds to figure 1C in (5). Shaded box indicates the approximate period of excessive mortality in control (–RU486) females.

If additional trials from Giannakou et al. indicate that induction of UAS-dFOXO by P{Switch}S1106 increases life span because it reduces mortality in treated flies, unlike the cases reported in (5), it will be important to provide criteria for how to select among these conflicting results. In addition, if other trials are thought to be effective, it will be necessary to collect under these conditions new data on males, fecundity, and stress resistance. Data on these variables presented in (5) were collected when there was no evidence for a positive impact of dFOXO on female longevity.


Navigate This Article