Obesogenic Diet Cycling Produces Graded Effects on Cognition and Microbiota Composition in Rats

Abstract

Scope: The effects of diet cycling on cognition and fecal microbiota are not well understood.

Method and results: Adult male Sprague-Dawley rats were cycled between a high-fat, high-sugar "cafeteria" diet (Caf) and regular chow. The impairment in place recognition memory produced by 16 days of Caf diet was reduced by switching to chow for 11 but not 4 days. Next, rats received 16 days of Caf diet in 2, 4, 8, or 16-day cycles, each separated by 4-day chow cycles. Place recognition memory declined from baseline in all groups and was impaired in the 16- versus 2-day group. Finally, rats received 24 days of Caf diet continuously or in 3-day cycles separated by 2- or 4-day chow cycles. Any Caf diet access impaired cognition and increased adiposity relative to controls, without altering hippocampal gene expression. Place recognition and adiposity were the strongest predictors of global microbiota composition. Overall, diets with higher Caf > chow ratios produced greater spatial memory impairments and larger shifts in gut microbiota species richness and beta diversity.

Conclusion: Results suggest that diet-induced cognitive deficits worsen in proportion to unhealthy diet exposure, and that shifting to a healthy chow for at least a week is required for recovery under the conditions tested here.

Keywords: cafeteria diet; cognitive impairment; diet cycling; gut microbiota; obesity.

Figure 1

Experimental design. The four groups in Experiment 1 received chow (control) or 16 days of Caf diet, with two groups withdrawn to chow for 4 or 11 days. The four groups in Experiment 2 received 16 days of Caf diet in 2, 4, 8, or 16‐day cycles, with each cycle separated by 4 days of chow. The four groups in Experiment 3 received chow (control), or 24 days of Caf diet provided continuously, or in 3‐day cycles separated by 4 or 2 days of chow. The introduction of Caf diet was staggered so that final tests could be conducted at the same time across groups.

Figure 2

Energy intake in Experiment 1. The schematic indicates when groups were exposed to Caf (black) or chow (open) diets. Long, Short, and CAF groups received 16 days of Caf diet, with groups Long and Short subsequently withdrawn to chow for 11 (Long) or 4 days (Short). A) Caf diet increased energy intake relative to controls. B) Average daily chow intake after Caf withdrawal was lower in groups Long and Short than the control group. n = 4 cages per group. Data analyzed by one‐way analysis of variance (ANOVA) followed by Tukey post‐hoc; *p < 0.05.

Figure 3

Body weight gain, fat, and lean mass in Experiment 1. The schematic indicates when groups were exposed to Caf (black) and chow (white) diets. Long, Short, and CAF groups received 16 days of Caf diet, with groups Long and Short subsequently switched to chow for 11 (Long) or 4 days (Short). Body weight A) and percent fat mass B) were significantly greater in group CAF and group Short than Controls, and intermediate in group Long. Lean mass C) was significantly elevated in group CAF relative to group Control. n = 11–12 per group. The ^ symbol in panel A denotes timing of body composition assessed by EchoMRI. Data were analyzed by one‐way analysis of variance (ANOVA) followed by Tukey post‐hoc; *p < 0.05.

Figure 4

Place and object recognition in Experiment 1. The schematic indicates when groups were exposed to Caf (black) and chow (white) diets. Baseline tests (left panels) showed no group differences in performance on either test. Test 2 was held after 16 days of Caf diet and withdrawal to chow for 11 (Long) or 4 (Short) days. Place recognition memory (top right) was impaired in groups Short and CAF relative to group Long, and in group CAF relative to controls, with no difference in object recognition (bottom right). n = 11–12 per group. Dashed line at 0.5 indicates impaired short‐term memory; i.e., equal exploration of both objects. Data were analyzed by one‐way analysis of variance (ANOVA) followed by Tukey post‐hoc; *p < 0.05.

Figure 5

Energy intake in Experiment 2. The schematic indicates when groups were exposed to Caf (black) and chow (white) diets. The four groups received 16 days of Caf diet in 2, 4, 8, or 16‐day blocks, separated by 4 days on chow. A) Energy intake on Caf diet days did not differ significantly between groups. B) Energy intake during chow cycles was significantly lower for 8CAF:4CHOW and 16CAF:4CHOW groups relative to 2CAF:4CHOW and 4CAF:4CHOW groups, averaged across the experiment. n = 4 cages per group. Data were analyzed by one‐way analysis of variance (ANOVA) followed by Tukey post‐hoc; *p < 0.05.

Figure 6

Body weight gain, fat, and lean mass in Experiment 2. The schematic indicates when groups were exposed to Caf (black) and chow (white) diets. The four groups received 16 days of Caf diet in 2, 4, 8, or 16‐day blocks, separated by 4 days on chow. There were no group differences in body weight A), percent fat mass B), or lean mass C) on Caf day 16. n = 12 per group. Data were analyzed by one‐way analysis of variance (ANOVA); “Baseline” in panel A refers to the rats yet to commence Caf diet cycling. BL1, time of first baseline EchoMRI measure; BL2, time of second EchoMRI measure, taken on Caf day 8 for group 2CAF:4CHOW and prior to Caf day 1 for other groups. *p < 0.05 versus other groups; Tukey post‐hoc.

Figure 7

Place recognition memory in Experiment 2. The schematic indicates when groups were exposed to Caf (black) and chow (white) diets. The four groups received 16 days of Caf diet in 2, 4, 8, or 16‐day blocks, separated by 4 days on chow. A) Place recognition memory declined significantly in all groups from baseline to Caf day 16, and was stable 4 days later on chow. B) Averaged across Tests 2 and 3, place recognition memory was significantly lower in group 16CAF:4CHOW than 2CAF:4CHOW, and not significantly different in other groups. n = 12 per group. Dashed line at 0.5 indicates equal exploration of both objects and an impairment in short‐term memory. Data were analyzed by one‐way analysis of variance (ANOVA) followed by Tukey post‐hoc; *p < 0.05.

Figure 8

Energy intake in Experiment 3. The schematic indicates when groups were exposed to Caf (black) and chow (white) diets. Control rats fed chow were compared with groups fed Caf diet for 24 days continuously (CAF) or in 3‐day cycles separated by 4 days (3CAF:4CHOW) or 2 days (3CAF:2CHOW) of chow. A) Caf diet intake, measured on the first day of each Caf cycle (or equivalent), was elevated in all Caf‐fed groups relative to chow, with no differences between continuous and cycled groups. B) Cycled groups suppressed energy intake during chow cycles relative to controls fed chow. n = 4 cages per group. Data were analyzed by one‐way analysis of variance (ANOVA) followed by Tukey post‐hoc; *p < 0.05.

Figure 9

Body weight and adiposity in Experiment 3. The schematic indicates when groups were exposed to Caf (black) and chow (white) diets. Control rats fed chow were compared with groups fed Caf diet for 24 days continuously (CAF) or in 3‐day cycles separated by 4‐day (3CAF:4CHOW) or 2‐day (3CAF:2CHOW) chow cycles. Terminal body weight A) and percent fat mass B) were significantly greater in CAF and 3CAF:2CHOW groups than the control group, while group 3CAF:4CHOW did not differ from the 3CAF:2CHOW or control groups. C) lean mass did not vary between groups. n = 12 per group. The ^ symbol in panel A denotes the timing of body composition assessment by EchoMRI. Data were analyzed by one‐way analysis of variance (ANOVA) followed by Tukey post‐hoc; *p < 0.05; n = 11–12.

Figure 10

Place and object recognition in Experiment 3. Baseline tests showed no group differences in place A) or object recognition C). At Test 2, held on Caf diet day 22–23, place recognition memory was impaired in all three Caf‐fed groups relative to the control group B), with no difference in object recognition D). n = 11–12 per group. Data were analyzed by one‐way analysis of variance (ANOVA) followed by Tukey post‐hoc; *p < 0.05; n = 11–12.

Figure 11

Caf diet exposure is negatively associated with place recognition memory. Group means (n = 11–12) for post‐diet place recognition (left), object recognition (middle), and place exploration time (right) are plotted against the percentage of time on Caf diet. A significant negative association was found for place recognition but not for object recognition or total exploration time. There are fewer data points in the middle panel as object recognition was not tested in Experiment 2.

Figure 12

Effects of Caf diet cycling on microbiota alpha diversity. Relative to control rats fed chow, Margalef's richness (left) was significantly reduced by higher percent time exposed to the Caf diet, with the 50% Caf group not significantly different to any other group. The only significant pairwise difference in Pielou's evenness (middle) was between 60% and 67% Caf groups, whereas Shannon diversity (right) differed between groups fed 43%–60% Caf relative to 67% Caf diet. Data were analyzed by one‐way analysis of variance (ANOVA) followed by Tukey post‐hoc tests; *p < 0.05; n = 11–12. Groups 33%, 50%, 67%, and 80% Caf from Experiment 2; Control, 43%, 60%, and 100% Caf from Experiment 3.

Figure 13

Nonmetric multidimensional scaling plot of gut microbiota composition in rats exposed to various schedules of cycling Caf diet. Relative to control rats fed chow (unfilled circles, left), increasing the proportion of Caf diet access (% days with Caf diet available) progressively shifted microbiota composition toward that of rats fed Caf diet continuous (100% Caf diet, top right). Permutational multivariate analysis of variance (PERMANOVA) analysis and pairwise comparisons indicated that all groups differed significantly except for three pairs of groups given cycling Caf (67% versus 80%, 33% versus 50%, 43% versus 60% Caf diet). Data points represent individual rats and proximity reflects microbiota similarity. Plot derived from a Bray–Curtis similarity index at the OTU level, n = 11–12 per group.