nine equal meals delivered every
160 min (CR-spread), thus mini-
mizing the chance of prolonged
fasting. To assess the contribu-
tion of the length of fasting, they
delivered the same CR ration
within a 2-hour window (CR-2h
or 22-hour fasting) or split into
eight equal meals delivered ev-
ery 90 min over 12 hours (CR-12h
or 12-hour fasting). To test the
impact of the CR ration fed dur-
ing the day or night, they deliv-
ered the CR-2h and CR-12h in
the day (CR-day-2h, CR-day-12h)
or night (CR-night-2h, CR-night-
12h) (see the figure).
Despite all CR groups receiv-
ing the same quality and quan-
tity of diet, their life spans dif-
fered. In the CR-spread group,
the median life span was 10%
longer than that of the AL group.
All other CR groups that had 12-
hour or 22-hour fasting lived lon-
ger than the CR-spread group,
demonstrating that fasting can
boost the life-extending effect
of CR. However, the median life
span of CR-day-2h or CR-day-12h
mice fed during their daytime
rest period was extended by 20%,
whereas mice in the CR-night-2h
and CR-night-12h groups (when
they are active) lived almost 35%
longer than mice in the AL group
( 2 ). Thus, the life-extending effect of fasting
is further boosted when it overlaps with the
circadian sleep period.
Despite eating less, all mice at middle age
in the CR groups were more active than mice
in the AL group. CR and fasting increase the
production of ketone bodies from the liver,
which can act on the circadian clock in the
brain to increase food-seeking activity ( 5 ).
Physical activity and exercise exert pleiotro-
pic health benefits in mice and humans ( 6 ).
Accordingly, as the mice aged, more-active
mice also lived longer. However, among the
CR groups, the CR-spread mice were more
active in older age and yet lived the short-
est amount of time. The CR-spread mice had
slightly more daytime activity, coinciding
with their mealtime, and there was a trend
toward daytime activity and reduced life
span. The CR-day mice also had slightly more
daytime activity, and they did not live as long
as the CR-night mice. Although sleep was not
measured, being more active during the day
likely disrupts sleep in nocturnal rodents.
Altogether, these results imply that maintain-
ing a robust sleep-wake and fasting-feeding
cycle aligned with the circadian clock, can
boost the life span–extending effect of CR.
The fasting-feeding cycle in all CR groups
except CR-spread likely sustained better cir-
cadian rhythms. Although the light-dark
cycle entrains the central circadian clock and
supports sleep ( 7 ), the fasting-feeding cycle
sustains robust circadian rhythms in periph-
eral organs ( 8 ). Hence, optimum alignment
of fasting and feeding with the light-dark
cycle can maintain robust circadian rhythms,
which in turn activates numerous pathways
in different organs at the optimum time ( 9 ).
Aligning fasting-feeding with the circadian
clock–programmed sleep-wake cycle to gain
health benefits has some precedence. In day-
active fruit flies, nightly fasting leads to life-
span extension compared with daytime fast-
ing ( 10 ). Mutations in circadian clock genes
disrupt feeding and sleeping patterns and
also dampen CR-dependent life-span exten-
sion in fruit flies and mice ( 11 , 12 ). Even when
mice are fed an isocaloric diet, those fed dur-
ing the active phase have improved health
and extended life span ( 8 ).
To investigate the cause of death, Acosta-
Rodríguez et al. examined mice that were
found dead of old age or moribund mice that
were euthanized. They found that cancer was
the major cause of death in all groups andthat liver cancer was the most
prevalent type, suggesting that CR
delayed the onset or severity of
cancer. To gain clues about the un-
derlying mechanism, the authors
probed the liver transcriptomes
of all cohorts. CR attenuated the
age-associated gene expression
changes observed in AL mice and
maintained a gene expression sig-
nature indicative of better meta-
bolic homeostasis. The CR-night
group predominantly exhibited
changes in immune gene expres-
sion, suggesting reduced inflam-
mation. This group also had more
genes expressed in alignment
with circadian rhythms compared
with CR-day groups. Circadian
gene expression temporally op-
timizes cellular processes, which
may partly explain the longer life
span of CR-night groups.
Beyond the liver transcrip-
tomes, there was little clue about
life-span effects in different CR
groups. Body weight and body
composition were comparable
among all CR groups. Likewise,
in humans, 25% CR with or with-
out 8-hour time-restricted eating
(8-hour CR or 16-hour fast) for 1
year led to similar weight loss ( 13 ).
Conversely, in weight-stabilized
humans, 6-hour time-restricted
eating improved cardiometabolic
health compared with habitual eating within
12-hours ( 14 ). This implies that changes in
body weight or composition in response to
dietary interventions that involve CR, fasting,
or meal timing may not accurately predict the
overall effectiveness of these interventions, al-
though perhaps trials with larger cohorts and
more restrictive timing are needed. Studies
of plasma and tissues from people undergo-
ing CR, fasting, or meal-time restriction hold
untapped potential for understanding the
system-wide impact of these dietary interven-
tions and predicting the extent to which they
can prevent or manage chronic diseases. j
REFERENCES AND NOTES- L. K. Heilbronn, E. Ravussin, Am. J. Clin. Nutr. 78 , 361
(2003). - V. Acosta-Rodríguez et al., Science 376 , 1192 (2022).
- V. A. Acosta-Rodríguez et al., Cell Metab. 26 , 267 (2017).
- V. D. Longo, S. Panda, Cell Metab. 23 , 1048 (2016).
- R. Chavan et al., Nat. Commun. 7 , 10580 (2016).
- J. A. Sanford et al., Cell 181 , 1464 (2020).
- M. Hatori, S. Panda, Trends Mol. Med. 16 , 435 (2010).
- E. N. C. Manoogian et al., E n d o c r. Rev. 43 , 405 (2022).
- L. S. Mure et al., Science 359 , eaao0318 (2018).
- M. Ulgherait et al., Nature 598 , 353 (2021).
- S. D. Katewa et al., Cell Metab. 23 , 143 (2016).
- S. A. Patel et al., FAS E B J. 30 , 1634 (2016).
- D. Liu et al., N. Engl. J. Med. 386 , 1495 (2022).
- E. F. Sutton et al., Cell Metab. 27 , 1212 (2018).
10.1126/science.adc8824INSIGHTS | PERSPECTIVES
GRAPHIC: N. DESAI/SCIENCEscience.org SCIENCEAd libitumMedian life span Maximum life spanFood availabilityCR-spread 10% CR effectCR-day-12h 10% 9% Fasting effectCR-day-2h 10% 11%CR-night-12h Circadian alignment
10% 24% and fasting effectCR-night-2hLife span (days)600 700 800 900 1000 1100 1200 130010% 25%Meal times and calorie restriction affect longevity
Circadian alignment of feeding to the most active part of the day (nighttime in
mice) and fasting boosts the life span–extending effects of calorie restriction (CR).1160 10 JUNE 2022 • VOL 376 ISSUE 6598