Dietary citrate supplementation enhances longevity, metabolic health, and memory performance through promoting ketogenesis

Abstract

Citrate is an essential substrate for energy metabolism that plays critical roles in regulating cell growth and survival. However, the action of citrate in regulating metabolism, cognition, and aging at the organismal level remains poorly understood. Here, we report that dietary supplementation with citrate significantly reduces energy status and extends lifespan in Drosophila melanogaster. Our genetic studies in fruit flies implicate a molecular mechanism associated with AMP-activated protein kinase (AMPK), target of rapamycin (TOR), and ketogenesis. Mice fed a high-fat diet that supplemented with citrate or the ketone body β-hydroxybutyrate (βOHB) also display improved metabolic health and memory. These results suggest that dietary citrate supplementation may prove to be a useful intervention in the future treatment of age-related dysfunction.

Keywords: dendritic spine; hippocampus; insulin; lifespan.

FIGURE 1

Physiological status of w¹¹¹⁸ Drosophila melanogaster treated with various concentrations of citrate. (a, b) Lifespans of male (a) and female (b) w¹¹¹⁸ flies treated with different concentrations of citrate. (c–f) Food intake (c), body weight (d), locomotor activity (e), and fecundity (f) of flies treated with different concentrations of citrate for 10 days. Detailed statistical analyses for the lifespans are shown in Table S2. The other data are expressed as mean ± SEM (n = 10–42 vials, 10 flies per vial). Not significant (n.s.), **p < 0.01, ***p < 0.001, compared to the control group by one‐way ANOVA with Fisher's LSD post hoc test

FIGURE 2

Citrate‐induced lifespan extension is associated with the AMPK and TOR pathways. (a–c) The levels of hemolymph glucose (a), total triglyceride (b), and ATP/ADP ratio (c) of male w¹¹¹⁸  flies treated with vehicle or 0.1% citrate for 10 days. (d) Representative Western blots and quantitative analyses showing that phosphorylated AMPK and phosphorylated S6K are upregulated and downregulated, respectively, in male w¹¹¹⁸  flies treated with 0.1% citrate for 10 days, compared with vehicle‐treated controls. (e, f) Lifespans of male mutant flies with RU486‐induced systemic inhibition of AMPK (e, Tub‐GS>UAS‐AMPK RNAi) and dTOR (f, Tub‐GS>UAS‐dTOR^TED ), treated with vehicle or 0.1% citrate. (h, i) Lifespans of male mutant flies with RU486‐induced fat body‐specific inhibition of AMPK (h, S106>UAS‐AMPK RNAi) and dTOR (i, S106>UAS‐dTOR^TED ), treated with vehicle or 0.1% citrate. (g, j) Lifespans of male mutant flies with RU486‐induced systemic (g, Tub‐GS>UAS‐GFP) and fat body‐specific (j, S106>UAS‐GFP) overexpression of GFP, treated with vehicle or 0.1% citrate. Detailed statistical analyses for the lifespans are shown in Table S2. The other data are expressed as mean ± SEM (n = 6–9 samples). *p < 0.05, ***p < 0.001 compared to the control group by Student's t test

FIGURE 3

Ketogenesis mediates citrate‐induced lifespan extension. (a, b) Lifespans of male mutant flies with RU486‐induced systemic inhibition of PGC‐1α (a, Tub‐GS>UAS‐PGC‐1α RNAi) and Hmgcl (b, Tub‐GS>UAS‐Hmgcl RNAi), treated with vehicle or 0.1% citrate. (c, d) Lifespans of male mutant flies with RU486‐induced fat body‐specific inhibition of PGC‐1α (c, S106>UAS‐PGC‐1α RNAi) and Hmgcl (d, S106>UAS‐Hmgcl RNAi), treated with vehicle or 0.1% citrate. (e, f) Citrate treatment for 10 days increases βOHB levels in male genetic control flies (Tub‐GS/S106>UAS‐GFP), but not in male mutant flies with systemic and fat body‐specific overexpression of AMPK RNAi (Tub‐GS/S106>UAS‐AMPK RNAi), dTOR^TED (Tub‐GS/S106>UAS‐dTOR^TED ), PGC‐1α RNAi (Tub‐GS/S106>UAS‐PGC‐1α RNAi), and HMGL RNAi (Tub‐GS/S106>UAS‐Hmgcl RNAi). (g) Dietary βOHB supplementation for 10 days dose‐dependently increases βOHB in male flies. (h) Lifespans of male w¹¹¹⁸  flies treated with various concentrations of βOHB. Detailed statistical analyses for the lifespans are shown in Table S2 and S3. The other data are expressed as mean ± SEM (n = 9–10 samples). *p < 0.05, **p < 0.01, ***p < 0.001 compared to the control group by Student's t test or one‐way ANOVA with Fisher's LSD post hoc test

FIGURE 4

Citrate administration improves metabolic health in mice fed on a high‐fat diet. Mice fed on a high‐fat diet were treated with various concentrations of citrate from 10 weeks of age (arrow). Metabolic health of mice was examined between 20 and 28 weeks of age. (a–d) Body weight (a), rate of weight gain (b), food intake (c), and water consumption (d) were monitored and analyzed. (e, f) Energy expenditure (e), and spontaneous locomotor activity (f) were analyzed in mice receiving vehicle and citrate treatments. (g, h) Glucose tolerance test (g) and insulin tolerance test (h) were conducted in mice receiving vehicle and citrate treatments. (i) Body composition was analyzed in mice receiving vehicle and citrate treatments. (j–m) Mice receiving 1% citrate treatment showed reduced liver weight (j, k) and decreased hepatic lipid accumulation, including reduced oil red O staining (j, lower panels) and lower levels of triglyceride (TG, l) and cholesterol (m). (n) Citrate administration dose‐dependently enhanced βOHB levels in the blood. (o) mRNA measurements of genes associated with de novo lipogenesis, cholesterol biosynthesis, and ketogenesis, from livers of mice receiving vehicle and citrate treatments. ATP citrate lyase (Acly), acetyl‐CoA carboxylase (Acaca), fatty acid synthase (Fasn), HMG‐CoA synthetase 1 (Hmgcs1), HMG‐CoA reductase (Hmgr), mevalonate diphosphate decarboxylase (Mvd), HMG‐CoA synthetase 2 (Hmgcs2), HMG‐CoA lyase (Hmgcl), β‐hydroxybutyrate dehydrogenase 1 (Bdh1). Data are expressed as mean ± SEM (n = 8 mice for each group in the whole‐body metabolic analyses, n = 10 mice for each group in all the other tests). Not significant (n.s.), *p < 0.05, **p < 0.01, ***p < 0.001 compared to the control group by one‐way ANOVA with Fisher's LSD post hoc test. Scale bar = 200 μm

FIGURE 5

Citrate administration affects mouse cognitive function. Mice fed on a high‐fat diet were treated with different concentrations of citrate from 10 weeks of age. Behavioral tests of mice were carried out between 16 and 20 weeks of age. (a–j) Mice were subjected to the open field test (a, representative moving path and total travel distance), rotarod test (b), elevated plus maze test (c), tail suspension test (d), forced swim test (e), buried food test (f), 3‐chamber social test (g–i), and novel object recognition test (j). (k–o) Structural analyses of hippocampal DG granule cells from vehicle‐ and citrate‐treated mice were analyzed at 28 weeks of age. DG granule cells were reconstructed (k), and the dendritic profile (l) and dendritic length (m) were analyzed. Representative micrographs of proximal (<50 μm from the soma) and distal (>150 μm from the soma) dendrites (n), and quantitative spine density (o) of DG granule cells. Data are expressed as mean ± SEM (n = 10 mice for behavioral tests, n = 29–30 neurons, and 58–60 dendritic segments for structural analyses). Not significant (n.s.), *p < 0.05, **p < 0.01, ***p < 0.001 compared to the control group by one‐way ANOVA with Fisher's LSD post hoc test. Scale bar = 10 μm

FIGURE 6

βOHB administration improves metabolic health and memory performance of mice. Mice fed on a high‐fat diet were treated with different concentrations of βOHB and/or vehicle from 10 weeks of age (arrow). Metabolic health and behavioral tests were carried out between 16 and 24 weeks of age. (a–c) Circulating βOHB (a), body weight (b), and food and water intake (c) were monitored and analyzed. (d, e) Energy expenditure (d), and spontaneous locomotor activity (e) were analyzed in mice receiving vehicle and βOHB treatments. (f, g) Glucose tolerance test (f) and insulin tolerance test (g) were conducted in mice receiving vehicle and βOHB treatments. (h, i) Memory performance of mice was evaluated using a 3‐chamber social test (h) and novel object recognition test (i), respectively. (j–m) DG granule cells were reconstructed (j), and the dendritic profile (k), dendritic length (l), and spine density of distal dendrites (m) were analyzed. Data are expressed as mean ± SEM (n = 4–5 for whole‐body metabolic analyses, n = 8–10 mice for behavioral and the other tests, n = 26–39 neurons, and 30–31 dendritic segments for structural analyses). Not significant (n.s.), *p < 0.05, **p < 0.01, ***p < 0.001 compared to the control group by Student's t test