Examining the role of olfaction in dietary choice

Abstract

Obesity is frequently caused by calorie-rich dietary choices across the animal kingdom. As prandial preference toward a high-fat diet develops in mice, an anti-preference or devaluation of a nutritionally balanced but less palatable standard chow diet occurs concomitantly. Although mechanistic insights underlying devaluation have been observed physiologically in the brain, it is unclear how peripheral sensory processing affects food choice. Because olfactory cues and odor perception help coordinate food preference and intake, we determine the role of smell in the targeted consumption of a high-fat diet and simultaneous devaluation of a standard chow diet. Using inaccessible food and loss-of-function manipulations, we find that olfactory information is neither sufficient nor necessary for both the acute and chronic selection of high-fat diet and coincident diminished value of standard diet. This work suggests alternative means are behind the immediate and sustained consumption of high-fat diet and concurrent standard diet devaluation.

Keywords: anosmia; devaluation; high-fat diet; obesity; olfaction.

Figure 1.. Devaluation of SD occurs only when HFD is accessible for consumption

(A) Experimental timeline and group schematic for home-cage measurements. (B–D) Weekly measurements of (B) weight (repeated measures [RM] two-way ANOVA, week × group: F (22, 253) = 17.10, p < 0.0001); (C) total number of calories consumed (RM two-way ANOVA, week × group: F (22, 253) = 55.85, p < 0.0001); and (D) number of SD calories consumed (RM two-way ANOVA, week × group: F (22, 253) = 145.9, p < 0.0001). (E) Experimental timeline and group schematic for fast-refeed test with 1-h SD access. (F) Within-subject comparison of 1-h SD fast-refeed consumption across testing sessions (RM two-way ANOVA, time × group: F (8, 88) = 23.91, p < 0.0001, Tukey’s multiple comparisons). Dotted lines in (B)–(D) delineate window of HFD availability or inaccessible HFD. B, baseline; W, withdrawal. Shaded blue area in (F) represents HFD home-cage availability. All groups had a mix of males and females. n = 8–9 per group. All error bars and shaded regions of (B)–(D) represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 2.. Long-term accessibility, not just exposure, of HFD is necessary for continued devaluation of SD

(A) Experimental paradigm and group schematic for home-cage measurements in mice undergoing various inaccessible exposure times to HFD. (B–D) Weekly measurements of (B) weights (RM two-way ANOVA, week × group: F (18, 168) = 17.70, p < 0.0001); (C) total number of calories consumed (RM two-way ANOVA, week × group: F (18, 168) = 27.79, p < 0.0001); and (D) number of calories consumed in SD only (RM two-way ANOVA, week × group: F (18, 168) = 251.3, p < 0.0001). (E) Daily measurements of the number of calories consumed in SD post-withdrawal (RM two-way ANOVA, week × group: F (18, 168) = 10.00, p < 0.0001). (F) Number of SD calories consumed during a 1-h fast-refeed session at baseline, at 1 and 4 weeks post-HFD exposure/baseline, and at 5 weeks after a 1-week withdrawal from HFD (RM two-way ANOVA, week × group: F (9, 84) = 22.22, p < 0.0001). (G) Experimental paradigm and group schematic for home-cage measurements in mice undergoing various accessible exposure times to HFD. (H–J) Weekly measurements of (H) weights (RM two-way ANOVA, week × group: F (18, 168) = 19.77, p < 0.0001); (I) total number of calories consumed (RM two-way ANOVA, week × group: F (18, 168) = 33.62, p < 0.0001); and (J) number of calories consumed in SD only (RM two-way ANOVA, week × group: F (18, 168) = 163.5, p < 0.0001). (K) Daily measurements of the number of calories consumed in SD right after the baseline period ends (RM two-way ANOVA, week × group: F (21, 196) = 24.31, p < 0.0001). (L) Number of SD calories consumed during a 1-h fast-refeed session at baseline, at 1 and 4 weeks post-HFD exposure/baseline, and at 5 weeks after a 1-week withdrawal from HFD (RM two-way ANOVA, week × group: F (9, 84) = 19.80, p < 0.0001). Shaded blue area in (F) and (L) represent HFD home-cage availability. All groups had a mix of males and females. n = 8 per group. All error bars and shaded regions of (B)–(E) and (H)–(K) represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 3.. Anosmia does not protect against HFD-induced weight gain, HFD preference, or SD anti-preference

(A) Representative photographs of mice that underwent either a sham or olfactory ablation surgery resulting in bilateral removal of the olfactory bulb. Scale bar represents 2 mm. (B) Experimental paradigm and group schematic for buried food assay. (C) Latencies (s) to discovering the food object (unpaired t test [two-tailed’, p < 0.0001), n = 20–22 per group. (D) Experimental timeline and group schematic for home-cage measurements. (E–I) Weekly measurements of (E) weight (RM two-way ANOVA, week × group: F (18, 228) = 12.13, p < 0.0001), n = 9–13 per group; (F) total number of calories consumed (RM two-way ANOVA, week × group: F (15, 190) = 10.87, p < 0.0001), n = 9–13 per group; (G) number of SD calories (RM two-way ANOVA, week × group: F (15, 190) = 74.59, p < 0.0001), n = 9–13 per group or (H) HFD consumed (RM two-way ANOVA, week × group: F (3, 63) = 0.6864, p = 0.5637), n = 9–13 per group; and (I) number of HFD calories consumed in sated animals during a 1-h novel exposure to HFD (unpaired t test [two-tailed], p = 0.4024), n = 9–12 per group. All groups had a mix of males and females. Dotted lines in (E)–(H) delineate window of HFD availability. All error bars and shaded regions of (E)–(H) represent mean ± SEM. ****p < 0.0001.

Figure 4.. Selective SD devaluation occurs independent of physiological hunger state in both anosmic and olfaction-intact mice

(A) Experimental timeline and group schematic for fast-refeed test with 1-h SD access. (B) Within-subject comparison of 1-h SD fast-refeed consumption across testing sessions (RM two-way ANOVA, time × group: F (9, 75) = 6.978, p < 0.0001, Tukey’s multiple comparisons), n = 6–8 per group. (C) Experimental timeline and group schematic for fast-refeed test with 1-h SD and HFD access. (D) Within-subject comparison of 1-h SD (RM two-way ANOVAs, time × group: F (3, 36) = 0.6966, p = 0.6966, Tukey’s multiple comparisons), n = 5–9 per group; and HFD (RM two-way ANOVAs, time × group: F (2, 24) = 0.06694, p = 0.9354, Tukey’s multiple comparisons) fast-refeed consumption across testing sessions, n = 5–9 per group. All groups had a mix of males and females. Shaded blue area in (B) and (D) represent HFD home-cage availability. All error bars represent mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.