Science - USA (2022-06-10)

identified as cycling in the CR-day-2h group.
Therefore, we do not recommend putting too
much weight on the number of cycling genes
called but rather emphasize the phase and
amplitude of gene expression.

Discussion

Classic CR protocols not only reduce energy
intake but also lead to severe behavioral time-
restricted feeding behavior and prolonged
fasting intervals ( 7 – 9 ). Because time-restricted
feeding and fasting are beneficial to health,
these two factors may contribute to life-span
extension in classic CR experiments. In the

work presented here, we deconvolved the

effects of calories, fasting, and circadian align-

ment on longevity. We compared five different

CR feeding groups that differed only in the

daily pattern of food consumption without

any changes in food composition or energy

content. We used an automated feeding sys-

tem ( 7 ) to test whether the timing and fasting

period between meals affected life span in male

mice under 30% CR by allowing food access

only during the day, at night, or evenly dis-

tributed throughout 24 hours. By spreading

out food through 24 hours, with no day-night

feeding pattern in the CR-spread group, we

found only an ~10% extension of median life

span compared with AL-fed mice. The CR-day

fed groups that had an ~12- or ~22-hour fast-

ing interval lived ~20%longer than AL control

mice. The degree of life-span extension was

significantly longer when food was consumed

during the nighttime, which is the normal feed-

ing time in nocturnal rodents (~35% versus

20% compared with AL, log-rank Mantel-Cox

P < 0.0001, and ~10% night versus day, log-

rank Mantel-CoxP <0.05)(Fig.2A).Although

potential sleep disruption needs to be carefully

studied in CR-day versus CR-night groups, re-

cent evidence shows that sleep homeostasis is

Acosta-Rodríguezet al., Science 376 , 1192–1202 (2022) 10 June 2022 8of11

1,718

694

1,507

Y

O

U

NG AGED

B

0

4

8

12

16

20

24

0 4 8 12162024

Phase 06mo_AL (h)

Phase 19mo_AL (h)

C

5

10

Fold change 19mo_AL

Fold change 06mo_AL

5 10

Time (h)

550 (20%)

267 (22%)

435 (20%)

373 (21%)

564 (20%)

372 (21%)

238 (22%)

244 (23%)

723 (20%)

45 (47%)

signaling

small molecule metabolic pr...

regulation of biological qu...

regulation of signaling

cell communication

regulation of cell communic...

response to oxygen−containi...

homeostatic process

cellular response to stimulus

cell−cell adhesion via plas...

0246

GO:BP

493 (17%)

178 (20%)

201 (20%)

562 (17%)

43 (30%)

589 (17%)

55 (29%)

510 (18%)

550 (18%)

633 (17%)

cell communication

negative regulation of RNA ...

negative regulation of cell...

developmental process

circadian rhythm

multicellular organismal pr...

rhythmic process

negative regulation of cell...

negative regulation of biol...

cellular response to stimulus

0123

−log10(adjusted_p_value)

GO:BP

Young AL

Aged AL

A

D

−2

−1

0

1

2

Young AL Aged AL

02444

0 24 44

Time (h)

E

Genes

young (6mo) aged (19mo)

Slope = 0.592 3

p < 2 e-16

0

4

40

160

0

60

0

1000

0

15

012243648

0

25

mRNA Level (RPKM)

(^012243648012243648)
Arntl
Per1
Nr1d1
Per2
Gys2
Pck1
Fig. 5. Circadian rhythms in liver gene expression are blunted during aging
in AL mice.(A) Gene expression patterns from mRNA-seq were analyzed for
circadian rhythms using the ARSER, JTK_CYCLE (from Metacycle R Package),
and RAIN circadian algorithms. Heatmaps (top) are sorted by phase of gene
expression. Each row is one gene with expression level inz-score at 12 time
points (columns). Venn diagram (bottom) shows the number of rhythmic genes
in young (gray) and aged (red) AL livers using stringent criteria (significantly
cycling according to three algorithms; Benjamini-HochbergP and q < 0.05 and
log 2 FC > 0.3) to define rhythmicity. (B) Examples of circadian profiles of genes
that are rhythmic in both young and aged AL livers. Black indicates young
AL livers, and red indicates aged AL livers. (C) Comparison of phase (left, hours)
and amplitude (right, daily fold change) of the 694 genes that were rhythmic
in both age groups. The red correlation line (Spearman) and linear regression
(slope is statistically different from 1;P <2×10–^16 )inthefoldchange
comparison indicates that aged animals showed overall reduced amplitude
of rhythmic genes. (D andE) GO terms of genes that are cycling in either
young (D) or aged (E) AL mice. Represented are 10 nonredundant terms of the
top 25 most significant enriched terms.
RESEARCH | RESEARCH ARTICLE