High-fat diet impairs intermediate-term memory by autophagic-lysosomal dysfunction in Drosophila

Abstract

High-fat diet (HFD) is considered a risk factor for age-related memory impairments such as Alzheimer's disease. However, how HFD affects memory formation remains unclear. In this study, we established a model of memory defects caused by HFD in Drosophila. Our results revealed that the HFD impaired intermediate-term memory (ITM), but not short-term memory (STM), produced by classical aversive olfactory conditioning, and decreased autophagic activity in the heads of the HFD-fed flies. Transient reduction in autophagic activity also impaired ITM, but not STM. Genetic enhancement of autophagic activity in neurons effectively restored ITM performance in the HFD-fed flies. Mechanistically, HFD impairs lysosomal function by downregulating the expression of lysosome-related genes, leading to impaired fusion of autophagosomes with lysosomes. These findings suggest that HFD impairs ITM by reducing autophagic activity and lysosomal dysfunction in the neurons.

Fig 1. HFD decreases ITM and alters metabolic homeostasis.

(A) Food treatment condition and feeding paradigm. The control diet (ND) and high-fat diet (HFD) were treated for 7 days in Canton-S flies. (B) Whole body triacylglycerol (TAG) levels in flies fed with ND and HFD. The TAG accumulated in the whole body significantly increased under the HFD condition compared to the ND condition. Error bars indicate SD (n = 7, 3. Student’s t-test, ** p < 0.01). (C) Relative circulating glucose levels. HFD-fed flies significantly increased basal glucose levels compared to ND-fed flies. Error bars indicate SD (n = 4. Student’s t-test, * p < 0.05). (D) Hemolymph was collected from ND-fed and HFD-fed flies at each timepoint. HFD-fed flies showed a significant rise of blood glucose level immediately after glucose loading compared to the blood glucose level after starvation (n ≥ 5, Two-way ANOVA, ** p < 0.01), while ND-fed flies showed a moderate rise of glucose level. The error bars show standard deviation. (E) Short-term memory (STM) and intermediate-term memory (ITM) in 10-day-old flies fed with ND and HFD. ITM was significantly decreased in the HFD group. Error bars indicate SEM (n = 25, 23, 34, 29. Two-way ANOVA, n.s., not significant, ** p < 0.01). (F) Long-term memory (LTM) was significantly decreased in the HFD group. Error bars indicate SEM (n = 12, 11. Student’s t-test, * p < 0.05).

Fig 2. HFD impairs autophagic activity in fly brains.

(A) HFD significantly increased the protein levels of Ref(2)p protein in the whole body, head, and body samples. The quantification of band intensity is shown below. Error bars indicate SD (n ≥ 5. Students’ t-test, * p < 0.05). (B) Western blot analysis of Atg8a lipidation in the samples of whole body, head, and body of ND or HFD-fed flies. The lower band represents lipidated Atg8a (Atg8a-II), and the upper band represents non-lipidated Atg8a (Atg8a-I). To obtain clearer Atg8a-II protein bands, the images were subjected to high-exposure processing (high exp). The high-exposure processed bands were placed below the original figures. The quantification of band intensity is shown below. Error bars indicate SD (n ≥ 3. Students’ t-test, * p < 0.05). (C) Representative images of Ref(2)p immunostaining in the area of cell bodies of mushroom body neurons (MBNs) at 10-day-old Canton-S flies fed with ND, ND + CQ, HFD, and HFD + CQ. Scale bar: 10µm. (D) Quantification of Ref(2)p puncta. Error bars indicate SD (n = 10, 9, 8, 10. One-way ANOVA, *** p < 0.001, **** p < 0.0001).

Fig 3. Transient suppression of autophagy in neurons impairs ITM.

(A) STM was significantly impaired in Elav-Gal4/UAS-Atg5 RNAi flies compared to control groups. Error bars indicate SEM (n = 22, 16, 15. One-way ANOVA, **** p < 0.0001). (B) ITM was significantly impaired in Elav-Gal4/UAS-Atg5 RNAi flies compared to control groups. Error bars indicate SEM (n = 10, 12, 10. One-way ANOVA, ** p < 0.01). (C) Representative images of Ref(2)p immunostaining in the cell body area of MBNs of Elav-Gal4/UAS-Atg5 RNAi and control groups. Confocal sections through the MBNs were shown. Scale bar: 10µm. (D) The experimental timeline shows the time of RU486 feedings and memory tests in 10-day-old flies. (E) STM in flies with transient knockdown of Atg1 in neurons (Elav-GS/UAS-Atg1-RNAi RU+) was not changed compared with other genotypes. Error bars indicate SEM (n = 12, 13, 9, 9. Two-way ANOVA, n.s., not significant). (F) The ITM was significantly reduced in flies with transient knockdown of Atg1 (RU+) flies compared to control groups. Error bars indicate SEM (n = 11, 11, 10, 11. Two-way ANOVA, * p < 0.05, ** p < 0.01).

Fig 4. Elevated autophagic levels in neurons restore HFD-induced memory impairment.

(A) The experimental timeline shows the time of RU + HFD feedings and memory tests in 10-day-old flies. (B) Representative images of Ref(2)p immunostaining in brains of 10-day-old Elav-GS/ + , Elav-GS/UAS-Rubicon RNAi, and Elav-GS/UAS-Atg1 fed with RU + HFD. Confocal sections of the area of MB cell bodies were shown. Scale bar: 10µm. (C) Quantification of Ref(2)p puncta. Error bars indicate SD (n = 18, 6, 11. One-way ANOVA, * p < 0.05, ** p < 0.01). (D) The STM performance of Elav-GS/+ and Elav-GS/UAS-Rubicon-RNAi flies. Error bars indicate SEM (n = 9, 8, 9, 6. Two-way ANOVA, n.s., not significant). (E) The ITM performance of Elav-GS/+ and Elav-GS/UAS-Rubicon-RNAi flies. HFD induces memory decline in control groups but not in Rubicon knockdown flies. Error bars indicate SEM (n = 22, 15, 16, 11. Two-way ANOVA, * p < 0.05, n.s., not significant). (F) The STM performance of Elav-GS/+ and Elav-GS/UAS-Atg1 flies. Error bars indicate SEM (n = 7, 8, 10, 7. Two-way ANOVA, n.s., not significant). (G) The ITM performance of Elav-GS/+ and Elav-GS/UAS-Atg1 flies. HFD induces memory decline in control groups but not in Atg1 overexpression flies. Error bars indicate SEM (n = 16, 15, 15, 9. Two-way ANOVA, * p < 0.05, n.s., not significant). (H) The ITM performance in 10-day-old flies fed with ND, ND+Rapamycin (Rap), HFD, and HFD + Rap. Error bars indicate SEM (n = 20, 15, 12, 14. Two-way ANOVA, * p < 0.05, ***p < 0.001, n.s., not significant).

Fig 5. HFD changes lysosomal activity.

(A) Representative images of Kenyon cell layers of flies expressing the UAS-GFP-mCherry-Atg8a transgene exclusively in the neuron systems (Elav-GS). The autophagy flux in Kenyon cell layers of brains suggested that HFD increased autophagosomes (AP, yellow arrows), whereas the number of autophagic lysosomes (AL, red arrows) remained unchanged. (B) Quantification of mCherry-GFP-Atg8a puncta. Error bars indicate SD (n = 10, 9, 8, 10. Two-way ANOVA, n.s., not significant, **** p < 0.0001). (C) Representative images of LysoTracker Red staining in MB at 10-day-old flies fed with ND, HFD, and ND + CQ. Lysosomes were marked by the LysoTracker Red signal. Confocal sections through the cell body of MBNs were shown. Scale bar: 10µm. (D) Quantification of LysoTracker Red in Kenyon cell layers. Error bars indicate SD (n = 5, 9, 6. One-way ANOVA, *p < 0.05, **p < 0.01, n.s., not significant). (E) Pie chart illustrating the distribution of 444 differentially expressed genes (DEGs), including 87 upregulated genes and 357 downregulated genes. (F) KEGG analysis of downregulated genes in ND and HFD head samples.

Fig 6. Inhibition of lysosomal activity causes memory impairment.

(A)Heatmap of significantly down-regulated genes associated with the lysosomal signaling pathway between ND and HFD. (B–D) HFD-fed fly heads showed decreased expression of LManII (B), CG4847 (C), and sap-r (D). qPCR measures each gene’s relative expression levels in ND and HFD head samples. The normalized expression level is shown (n = 6. Mann-Whitney test for LManII; other genes use Student’s t-test, * p < 0.05). (E) The ITM performance in flies with LManII knockdown in the whole body (Tub-GS/UAS-LManII RNAi, RU+) is unchanged compared with other groups. Error bars indicate SEM (n = 9, 10, 8, 10. Two-way ANOVA, n.s., not significant). (F) The ITM performance in flies with CG4847 knockdown in the whole body (Tub-GS/UAS-CG4847 RNAi, RU+) is unchanged compared with other groups. Error bars indicate SEM (n = 9, 10, 8, 6. Two-way ANOVA, n.s., not significant). (G) The ITM performance in flies with sap-r knockdown in the whole body (Tub-Gal4/UAS Sap-r RNAi) is unchanged compared to other groups. Error bars indicate SEM (n = 8, 14, 12. Two-way ANOVA, n.s., not significant). (H) The ITM performance was significantly reduced in CQ-fed flies compared to control groups. Error bars indicate SEM (n = 8. Student’s t-test, * p < 0.05).