Time-Restricted Feeding Reduces Atherosclerosis in LDLR KO Mice but Not in ApoE Knockout Mice

Abstract

Background: Dyslipidemia increases cardiovascular disease risk, the leading cause of death worldwide. Under time-restricted feeding (TRF), wherein food intake is restricted to a consistent window of <12 hours, weight gain, glucose intolerance, inflammation, dyslipidemia, and hypercholesterolemia are all reduced in mice fed an obesogenic diet. LDLR (low-density lipoprotein receptor) mutations are a major cause of familial hypercholesterolemia and early-onset cardiovascular disease.

Methods: We subjected benchmark preclinical models, mice lacking LDLR-knockout or ApoE knockout to ad libitum feeding of an isocaloric atherogenic diet either ad libitum or 9 hours TRF for up to 13 weeks and assessed disease development, mechanism, and global changes in hepatic gene expression and plasma lipids. In a regression model, a subset of LDLR-knockout mice were ad libitum fed and then subject to TRF.

Results: TRF could significantly attenuate weight gain, hypercholesterolemia, and atherosclerosis in mice lacking the LDLR-knockout mice under experimental conditions of both prevention and regression. In LDLR-knockout mice, increased hepatic expression of genes mediating β-oxidation during fasting is associated with reduced VLDL (very-low-density lipoprotein) secretion and lipid accumulation. Additionally, increased sterol catabolism coupled with fecal loss of cholesterol and bile acids contributes to the atheroprotective effect of TRF. Finally, TRF alone or combined with a cholesterol-free diet can reduce atherosclerosis in LDLR-knockout mice. However, mice lacking ApoE, which is an important protein for hepatic lipoprotein reuptake do not respond to TRF.

Conclusions: In a preclinical animal model, TRF is effective in both the prevention and regression of atherosclerosis in LDLR knockout mice. The results suggest TRF alone or in combination with a low-cholesterol diet can be a lifestyle intervention for reducing cardiovascular disease risk in humans.

Keywords: atherosclerosis; bile acids; fatty liver; intermittent fasting; lipoproteins; sterols.

Figure 1:. Time-restricted feeding of an isocaloric high fat high cholesterol diet prevents body weight gain in LDLR-KO mice.

A. Schematic of the experimental design. B-C. Evolution of body weight (B) (n=40/group initially) and cumulative food consumption (C) (n=8 cages (40 mice)/group). (Feeding ns P-value = 0.16, Time x Feeding *** P-value = 0.0001/ ns P-value = 0.37) D-E. Profile of food consumption over 48h of recording in metabolic chambers (D) and associated quantification of daily percentage of food eaten during the dark and light phase (below) and total food consumption (E) (n=6 mice/group). (Feeding ns P-value = 0.34/ 0.13) F-G. Activity profile over 48h of recording in metabolic chambers (F) and associated quantification of daily percentage during the dark and light phase (below) and total activity (G). (n=6 mice/group). (Feeding ns P-value = 0.14/ * P-value = 0.01, Time x Feeding * P-value = 0.02/ ns P-value = 0.38) H. Energy expenditure profile over 48h of recording in metabolic chambers (n=6 mice/group). (ns P-value = 0.52) I-J. Fecal output in g/day (I) and in kcal/day as determined by bomb calorimetry (J) (n=4 mice/group). (ns P-value = 0.052/ 0.23) K-L. Body composition in grams (K) and as a % of body weight (L) after 12 weeks on the feeding protocol (n=11–13 mice/group). (** P-value = 0.004) Statistics: (B,C) Mixed-effects model (REML) (Feeding & Time; inset) and Sidak’s multiple comparisons tests (on graph). (E,G) Two-way ANOVA (Feeding & Time; inset) and Sidak’s multiple comparisons tests (on graph). (D,F,H) Two-way Repeated Measures ANOVA (Feeding & Time; inset) and Sidak’s multiple comparisons tests (on graph). (I-L) Unpaired t-test. For all, * p<0.05, ** p<0.01, *** p<0.001

Figure 2:. Time-restricted feeding of an isocaloric high fat high cholesterol diet attenuates atherosclerotic plaques in LDLR-KO mice

A-B. Serum cholesterol (A) triglyceride concentration (B) (n=20 mice/group). (* P-value = 0.01) C-D. Cholesterol (C) and triglycerides (D) concentration in indicated fraction of FPLC-fractionated serum (n=3 pools of serum from 3 mice/group). (Feeding ns P-value = 0.18) E-F. Representative Oil Red O staining of dissected aorta (E) and total quantification of the % of the lesioned aorta (F) (n=13 mice/group). (* P-value = 0.02) G-H-I. Representative Van Gieson staining of section #4 out of 10 in the aortic leaflet (G), total quantification of the lesion volume (n=7–8 mice/group) (H) and % necrotic core area (I). Lumen, V: ventricle, and NC: necrotic core. (** P-value = 0.007, * P-value = 0.01) Statistics: (A, B, F, H, I) Unpaired t-test, * p<0.05, ** p<0.01, *** p<0.001 (C, D) One-way ANOVA with Tukey’s multiple comparisons test For all, * p<0.05, ** p<0.01, *** p<0.001

Figure 3:. Time-restricted feeding of an isocaloric high fat high cholesterol diet fails to prevent atherosclerosis in ApoE-KO mice.

A. Schematic of the experimental design. B-C. Evolution of body weight (B) (n=40 mice/group initially) and food consumption (C) (n=8 cages/group). (Feeding ** P-value = 0.001/ * P-value = 0.02, Time x Feeding ns P-value = 0.33) D-E. Profile of food consumption over 48h of recording in metabolic chambers (D) and associated quantification of daily percentage of food eaten during the dark and light phase (below) and total food consumption (E) (n=6 mice/group). (Feeding ns P-value = 0.09) F-G. Activity profile over 48h of recording in metabolic chambers (F) and associated quantification of daily percentage during the dark and light phase (below) and total activity (G). (n=6 mice/group). (Feeding ns P-value = 0.96/ 0.94, Time x Feeding ns P-value = 0.63/ 0.99) H. Energy expenditure profile over 48h of recording in metabolic chambers (n=6 mice/group). (Feeding ns P-value = 0.44) I-J. Fecal output in g/day (I) and in kcal/day as determined by bomb calorimetry (J) (n=4 mice/group). (ns P-value = 0.06, ns P-value = 0.28) K-L. Body composition in grams (K) and as a % of body weight (L) after 12 weeks on the feeding protocol (n=19–23 mice/group). (*** P-value = 0.0007, ** P-value = 0.005) M-N. Serum cholesterol (M) (n=9 mice/group) and triglyceride concentration (N) (n=10 mice/group). (ns P-value = 0.69, ns P-value = 0.31) O-P. Cholesterol (O) and triglycerides (P) concentration in indicated fraction of FPLC-fractionated serum (n=3 pools of serum from 3 mice/group). (ns P-value = 0.86, ns P-value = 0.94) Q-R. Representative Oil Red O staining of dissected aorta (Q) and total quantification of the % of the lesioned aorta (R) (n=12–13 mice/group). (ns P-value = 0.73) S-T-U. Representative Van Gieson staining of section #4 out of 10 in the heart leaflet (S), total quantification of the lesion size (T) and spatial distribution of the lesions (U) (n=3 mice quantified/group). Lumen, V: ventricle, and NC: necrotic core. (ns P-value = 0.27, 0.99) Statistics: (B,C) Mixed-effects model (REML) (Feeding & Time; inset) and Sidak’s multiple comparisons tests (on graph). (E,G) Two-way ANOVA (Feeding & Time; inset) and Sidak’s multiple comparisons tests (on graph). (D,F,H) Two-way Repeated Measures ANOVA (Feeding & Time; inset) and Sidak’s multiple comparisons tests (on graph). (I-L, M,N,R,T) Unpaired t-test. (O,P) One-way ANOVA with Tukey’s multiple comparisons test. For all, * p<0.05, ** p<0.01, *** p<0.001

Figure 4:. Time-restricted feeding reduces hepatic VLDL production.

A. Schematic of hepatic regulation of serum TRLs. B. Serum lipoprotein lipase activity (n=4 mice/group). (ns P-value = 0.13) C. TRLs absorption assay: Serum TG concentration after IV tyloxapol (0.5mg/g) and corn oil gavage (n= 8mice/group). (Feeding ns P-value = 0.51, Time x Feeding ns P-value = 0.60) D. TRLs clearance assay: Serum [³H]retinol (cpm/10μL plasma)(n=5–6 mice/group). (Feeding ns P-value = 0.78, Time x Feeding ns P-value = 0.58) E. VLDL secretion assay: Serum TG concentration after IV tyloxapol (0.5mg/g) (n=4 mice/group. (Feeding ns P-value = 0.07, Time x Feeding * P-value = 0.01) (Feeding ns P-value = 0.09, Time x Feeding ## P-value = 0.002) F. RER over 48h of recording in metabolic chambers (n=6 mice/group) in LDLR-KO mice. G. Serum level of 3-beta-hydroxybutyrate (n=4–5 mice/group) in LDLR-KO mice at subjective fast (ZT12) or fed state (ZT18). H. RER over 48h of recording in metabolic chambers (n=6 mice/group) in ApoE-KO mice. I. Serum level of 3-beta-hydroxybutyrate (n=4–5 mice/group) in ApoE-KO mice at subjective fast (ZT12) or fed state (ZT18). J. Representative images of liver Hematoxylin and Eosin staining at 20X magnification. K. Hepatic triglyceride quantification (n=6 mice/group). (*** P-value = 0.0004, Feeding * P-value = 0.01, Feeding x Genotype *** P-value = 0.0003) Statistics: (B) Unpaired t-test (on graph). (C,D,E,F,H) Two-way Repeated Measures ANOVA (Feeding & Time; inset) and Sidak’s multiple comparisons tests (on graph). (G,I) One-way ANOVA with Tukey’s multiple comparisons test (on graph). (K) Two-way ANOVA (Feeding & Genotype; below) and Tukey’s multiple comparisons tests

Figure 5:. Time-restricted feeding affects the expression of hepatic transcripts implicated in lipid and cholesterol metabolism.

Schematic of significant changes in mRNA expression of genes involved in lipid and glucose metabolic pathway in the liver of LDLR-KO mice on ad lib (LA) or TRF (LT) at ZT10 “fasted” and ZT16 “fed” (3h post feeding in TRF mice) (n=6/group). Y-axis is TMM normalized counts per million (CPM). Also see Data Set DS1 for expression values and statistical tests.

Figure 6:. Time-restricted feeding modulates liver cholesterol and bile acid metabolism to promote fecal loss.

A. Hepatic cholesterol quantification (n=6–12 mice/group). (*** P-value = 0.0006, Feeding * P-value = 0.02, Feeding x Genotype *** P-value = 0.0003) B-C. Cholesterol absorption assay: evolution of serum fluorescence after gavage of a fluoresterol:corn oil mixture (n=4 mice/group) in LDLR-KO (B) and ApoE-KO (C) mice. (Feeding ns P-value = 0.62/ 0.21, Time x Feeding ns P-value = 0.79/ 0.28) D-E. Serum bile acid concentration (n=4–5 mice/group) in LDLR-KO (D) and ApoE-KO (E) mice. (** P-value = 0.002, * P-value = 0.01) F-G. Fecal cholesterol (F) and bile acid concentration (G) (n=6 mice/group). (** P-value = 0.004, * P-value = 0.02, ns P-value = 0.97) Statistics: (A,F,G) Two-way ANOVA (Feeding & Genotype; below) and Tukey’s multiple comparisons tests (on graph). (B,C) Two-way Repeated Measures ANOVA (Feeding & Time; inset) and Sidak’s multiple comparisons tests (on graph). (D,E) Unpaired t-test (on graph).

Figure 7:. In a regression therapeutic model, time-restricted feeding and diet quality act additively to attenuate atherosclerotic risk in LDLR-KO mice.

A. Experimental design of regression therapeutic model. B. Evolution of body weight over 13 weeks (n=9–10 mice/group). (** P-value = 0.001, *** P-value = 0.0001) C-D. Body composition at week 10 (4 weeks post intervention) (C) and fat mass as a percentage of body weight (n=5 mice/group) (D). (* P-value = 0.02) E-F. Serum triglyceride (E) and cholesterol concentration (n=7–10 mice/group) (F). (* P-value = 0.01, ** P-value = 0.006, *** P-value = 0.0008, ns P-value = 0.22) G-H. Fecal total bile acid (μM) (G) and cholesterol content (mg/dL) (n=4 mice/group) (H). (** P-value = 0.002, *** P-value = 0.005) I-J. Representative Oil Red O staining of dissected aorta (I) and total quantification of the % of the lesioned aorta (J) (n=8–9 mice/group). (** P-value = 0.001) K. Total quantification of plaque volume in the heart leaflet. (n=9–10 mice/group). (* P-value = 0.02, ns P-value = 0.52) Statistics: (B) Mixed-effects model (REML) (Feeding & Time; inset) and Tukey’s multiple comparisons tests (results at the last time point only are shown on the graph). (D) One-way ANOVA with Tukey’s multiple comparisons test. (E-F) One-way ANOVA with Sidak’s multiple comparisons test for indicated pairs of comparisons. (G-K) One-way ANOVA with Dunnett’s multiple comparisons test (relative to L-FCA). For all, * p<0.05, ** p<0.01, *** p<0.001