Science - USA (2022-06-10)

maintained most activity during the nighttime

(Fig. 1, C and D). Therefore, feeding and loco-

motor activity are misaligned with daytime

feeding in these animals throughout their life

span,whichwouldbeexpectedtoleadtoad-

verse metabolic consequences ( 25 , 26 ). Ac-

tivity of the mice declined as they aged ( 27 ),

with AL mice having the lowest activity levels

compared with the CR groups between 6 and

18 months of age [Fig. 1G and fig. S5; two-way

analysis of variance (ANOVA); age,P < 0.0001;

feeding,P < 0.0001; interaction nonsignifi-

cant (NS)].

Life-span extension by CR depends on

feeding time

We investigated the contribution of calories,

feeding time, and fasting period on longevity.

CR was sufficient to extend the median life

span in male mice, but the range of this ex-

tension depended on when the food was con-

sumed (Fig. 2A). The percentage of life-span

extension varied across conditions. Consistent

with other reports, AL mice had a median life

span of 792 days ( 24 , 28 ). CR-fed mice lived

10 to 35% longer than AL mice depending on

the CR group. The CR-spread group, which

had a 30% reduction in calories but with feed-

ing spread throughout the day-night cycle, had

a median life span of 875 days, which is 10.5%

longer than that of AL mice, demonstrating

that CR alone without time restriction or fast-

ing is sufficient to extend longevity. The CR-

day-12h and CR-day-2h groups had median

life spans of 942 and 959 days, respectively,

which are 18.9 and 21.1% longer than the life

spans of AL mice, respectively. Thus, in addi-

tion to the reduction in calories, a minimum

of 12 hours of fasting induces its own ben-

efits on longevity. There were no significant

differencesinlifespanwhenday-fedmice

fasted for ~22 hours versus 12 hours, indicat-

ing that 12 hours of fasting is sufficient. CR-

night fed mice outlived both CR-day groups:

The CR-night-12h and CR-night-2h mice had

median life spans of 1058 and 1068 days, re-

spectively, which are 33.6 and 34.8% longer

than the life spans of AL mice, respectively.

Again, there was no additional benefit of

~22 hours of fasting compared with 12 hours

of fasting in these groups fed at night, indi-

cating that 12 hours of fasting is sufficient for

prolonging life span. There was a significant

extension of life span by CR-night-2h mice

(34.8% extension) over CR-day-2h mice (21.1%

extension) (log-rank Mantel-Cox,P <0.05),

which differed in the relative phase of food

consumptionbythemice.Itispossiblethat

sleep disruption due to misaligned feeding

could contribute to the difference in life span.

Further studies are required to determine

whether sleep is affected and, if so, if potential

sleep disruptions can contribute to the differ-

ences in life span observed between CR-day

Acosta-Rodríguezet al., Science 376 , 1192–1202 (2022) 10 June 2022 3of11

35%

10%

20%

CR

Median lifespan

increase vs AL

0 300 600 900 1200

0

50

100

Age (days)

Survival(%)

0.0001

0.0061

0.0002

<0.0001

<0.0001

CR-night-02h

CR-spread

CR-day-02h

CR-night-12h

CR-day-12h

AL^792

Median

Lifespan(days)

Log-rank

Mantel-Cox (p-value)

942 ***

(^875) **
959 *
1068 **
1058 ****
A
B
C
2mo 12mo6mo 18mo 24mo 30mo 36mo
(^4008001200400800120040080012004008001200400800120040080012004008001200)
0
20
40
0
20
40
0
20
40
0
20
40
0
20
40
0
20
40
Lifespan
Total activity (count/min)
CRnight2h
CRspread
CRday2h
CRnight12h
CRday12h
AL

• 
  

  ---

  
  

  ---

  
  

  ---

  
  

  ---

  
  

  ---

  
  

  **
  AL CR-
  spread
  CR-
  day
  12h
  CR-
  day
  2h
  CR-
  night
  12h
  CR-
  night
  2h
  Adenoma 13% 9% 22% 16% 17% 4%
  Hepatocellular
  carcinoma 8% 19% 19% 29% 23% 15%
  Histiocytic sarcoma 21% 38% 44% 26% 47% 33%
  Lymphoma 3% 0% 0% 0% 3% 4%
  Neoplasia 16% 16% 4% 10% 3% 4%
  Acidophilic macrophage
  pneumoniaBronchoalveolar 3% 9% 7% 6% 3% 19%
  carcinomaBronchoalveolar 3% 0% 4% 6% 7% 7%
  adenoma 13% 16% 7% 16% 17% 19%
  Histiocytic sarcoma 5% 16% 19% 16% 13% 19%
  Glomerulonephritis 24% 3% 11% 10% 17% 7%
  Histiocytic sarcoma 3% 3% 0% 0% 7% 0%
  Lymphoma 3% 0% 0% 0% 7% 7%
  Histiocytic sarcoma 3% 9% 15% 13% 17% 4%
  Lymphoma 3% 0% 0% 3% 7% 0%
  Liver
  Lung
  Kidney
  Spleen
  0 100 200
  other
  Heart
  Adrenal
  Gland
  Lymph
  nodes
  Spleen
  Kidneys
  Lungs
  Liver
  AL
  CR-spread
  (^) CR-day-12h
  (^) CR-day-2h
  (^) CR-night-12h
  CR-night-2h
  Frequency
  Fig. 2. Extent of CR-mediated increases in longevity depend on feeding time.(A) Survival curves
  for each group (n =43forALandn = 36 for each of the CR groups) are shown in the left panel and
  the median life span (days) is shown in the inset. Right panel summarizes the results, showing the
  increase in life span from timed feeding with the largest increase when food is restricted to night.
  (B) Correlation plots comparing life span (days) for each mouse with its daily averaged total activity
  (counts/min) at different ages. Increased activity significantly correlates with longer life span in old,
  but not young, mice in all groups (see asterisks, Spearman correlation). (C) Necropsy followed by
  histopathology results showing pathologies and diseases (left) and tissues mostly affected (right) at the
  time of death for each feeding condition.
  RESEARCH | RESEARCH ARTICLE