Reduced caloric intake and periodic fasting independently contribute to metabolic effects of caloric restriction

Abstract

Caloric restriction (CR) has positive effects on health and longevity. CR in mammals implements time-restricted (TR) feeding, a short period of feeding followed by prolonged fasting. Periodic fasting, in the form of TR or mealtime, improves metabolism without reduction in caloric intake. In order to understand the relative contribution of reduced food intake and periodic fasting to the health benefits of CR, we compared physiological and metabolic changes induced by CR and TR (without reduced food intake) in mice. CR significantly reduced blood glucose and insulin around the clock, improved glucose tolerance, and increased insulin sensitivity (IS). TR reduced blood insulin and increased insulin sensitivity, but in contrast to CR, TR did not improve glucose homeostasis. Liver expression of circadian clock genes was affected by both diets while the mRNA expression of glucose metabolism genes was significantly induced by CR, and not by TR, which is in agreement with the minor effect of TR on glucose metabolism. Thus, periodic fasting contributes to some metabolic benefits of CR, but TR is metabolically different from CR. This difference might contribute to differential effects of CR and TR on longevity.

Keywords: caloric restriction; circadian rhythms; fasting; gene expression; glucose homeostasis; insulin sensitivity; longevity; metabolism.

Figure 1

TR did not change daily pattern of food intake and body weight. (a) Scheme of feeding protocol. Mice were randomly assigned into three groups: Ad libitum feed (AL) had unlimited access to the food around the clock, 30% caloric restriction (CR 30%), and time‐restricted feed (TR) for 8–10 weeks on the respective diet, at the start of the experiments. Analysis was performed around the clock at times indicated by yellow arrows. (b) Weekly average food intake (AL (n = 8 cages), CR (n = 8 cages), and TR (n = 6 cages)) and (c) daily average food intake (AL (n = 8 cages), CR (n = 8 cages), and TR (n = 6 cages)). Daily average food intake for AL group for the first ten days was not measured; only weekly measurements were performed; and hence, AL is represented by dotted lines. For first 10 days food intake (c): AL—blue dotted line, blue triangles; TR—green solid line, green solid diamonds; CR—red solid line, red squares; for (b): AL—blue solid line, blue triangles; TR—green solid line, green solid diamonds; CR—red solid line, red squares. (d) Hourly food intake for mice on different diets. The data were normalized to daily food intake. The total daily food intake was set up as 1.0. AL—blue solid line, TR—green solid line, CR—red box. Data were double plotted for illustrative purposes. AL (n = 7 cages), CR (n = 6 cages), and TR (n = 5 cages). (e) Percent of daily food intake during the dark phase for mice on different diets. AL (n = 7 cages), CR (n = 6 cages), and TR (n = 5 cages). (f) Percent of daily food intake during the first meal. AL (n = 7 cages), CR (n = 6 cages), and TR (n = 5 cages). (g) Amount of food consumed as first meal. AL (n = 7 cages), CR (n = 6 cages), and TR (n = 5 cages). For (e‐f) AL—blue bars, TR—green bars, and CR—red bars. The time of the day when the food was provided for CR and TR mice is indicated by the red arrow. All data represented as Mean ± SD. One‐way ANOVA with Bonferroni correction for multiple comparison was performed. Letters indicate significant effect of the diet (p < .05); a—AL versus CR, b—AL versus TR, c—CR versus TR. Light was turned on at ZT0, and light was turned off at ZT12. Light and dark bars indicate light and dark phases of the day

Figure 2

Periodic fasting did not improve glucose homeostasis. (a) Around the clock blood glucose (4‐hr resolution) in AL, CR, and TR mice. The blood glucose was assayed in mice that were not fasted before the experiments. (b) Blood glucose around the feeding time (1‐hr resolution) in AL, CR, and TR mice (n = 6–9 per time point per diet). (c) Fasting blood glucose in mice, the food was removed at ZT16, and blood glucose was measured every 4 hr (n = 6–8 per time point), (d and e) Intraperitoneal glucose tolerance test (ip‐GTT). Blue solid line—AL mice (n = 7) were fasted for 12 hr before the test; blue dotted line—AL mice were fasted for 22 hr before the experiment (n = 4); red solid line—CR mice (n = 9); green solid line—TR mice (n = 10). (f) Area under the curve, the quantification of ip‐GTT data presented in (d and e). The same symbols indicate no statistical significant difference between the diets; different symbols indicate statistical significant difference (p < .05) between diets. (g‐m) mRNA expression of rate‐limiting gluconeogenesis genes (g) Pcx, (h) Pck1, (i) Fbp1, and (j) G6pc; glycolytic genes (k) Gck and (l) Pfk1; and glycerol metabolism gene (m) Gk in the liver of AL, CR, and TR mice (n = 4 per time point per diet). The time of the day when the food was provided for CR and TR mice is indicated by the red arrow. All data represented as Mean ± SD. One‐way ANOVA with Bonferroni correction for multiple comparison was performed. Letters and asterix indicate significant effect of the diet (p < .05); a—AL versus CR, b—AL versus TR, c—CR versus TR. *p ≤ .05, **p ≤ .01, ***p ≤ .001 and ****p ≤ .0001. Light was turned on at ZT0, and light was turned off at ZT12. Light and dark bars indicate light and dark phases of the day

Figure 3

Periodic fasting contributes to CR induced improved insulin homeostasis and downregulation of mTORC1. (a) Around the clock plasma insulin levels. AL (upper panel), CR (middle panel), and TR (lower panel) mice (n = 4 per time point per group). The insulin was measured in blood obtained from tail vein. Mice on all three diets were not fasted prior to the experiment. (b‐c) Intraperitoneal Insulin tolerance test (ip‐ITT). Blue dashed line—AL mice fasted for 6 hr (n = 5); blue solid line—AL mice fasted for 12 hr (n = 5); blue dotted line—AL mice fasted for 22 hr (n = 4); green solid line—TR mice (n = 5); red solid line—CR mice (n = 8). (d) Area above the curve, the quantification of ip‐ITT data presented in (b and c). The same symbols indicate no statistical significant difference between the diets; different symbols indicate statistical significant difference (p < .05) between diets. (e) Circadian rhythms of ribosomal protein S6 phosphorylated on serine 235/236 in the liver of mice on different diets; (e) representative Western blot and (f) quantification (n = 4 per time point per diet). The time of the day when the food was provided for CR and TR mice is indicated by the red arrow. All data represented as Mean ± SD. One‐way ANOVA with Bonferroni correction for multiple comparison was performed. Letters and asterix indicate significant effect of the diet (p < .05); a—AL versus CR, b—AL versus TR, c—CR versus TR. *p ≤ .05, **p ≤ .01, ***p ≤ .001 and ****p ≤ .0001. Light was turned on at ZT0, and light was turned off at ZT12. Light and dark bars indicate light and dark phases of the day