APOE4 drives widespread changes to the hepatic proteome and alters metabolic function

Abstract

Apolipoprotein E (APOE) is essential for lipid homeostasis and has been extensively studied in Alzheimer's disease (AD). Individuals carrying an APOE4 allele have an increased risk of AD and exhibit deficits in energy metabolism, including glucose utilization and mitochondrial dysfunction. While the role of APOE in the liver is well characterized, the impact of APOE genotype on hepatic health and metabolism remains poorly understood. We sought to investigate this using young APOE3 and APOE4-targeted replacement mice and isogenic-induced pluripotent stem cell (iPSC)-derived hepatocyte-like cells (iHLCs). Proteomic and functional assays show that APOE4 causes extensive changes to liver mitochondrial function in a sex-specific manner in mice and alters glucose and lipid metabolism. APOE4 also impairs mitochondrial function in iHLCs, shifts metabolism towards glycolysis, increases reliance on fatty acid utilization, and drives lipid accumulation. Together, these findings show that APOE genetic variation causes mitochondrial dysfunction and rewires hepatic metabolism.

Keywords: Biochemistry; Human metabolism; Proteomics.

Graphical abstract

Figure 1

Whole liver proteomics in female and male APOE3 and APOE4 mice (A) Study schematic showing primary outcomes in APOE-targeted replacement (TR) mice. n = 3 mice per group for all mouse proteomic outcomes. (B) APOE protein expression from whole liver measured via ELISA in male and female APOE3 and APOE4 mice. Statistical significance was determined by two-way ANOVA. Data are presented as mean ± SD. n = 8 mice per genotype and sex. (C) The number of differentially expressed (DE) proteins in male and female APOE3 and APOE4 mice. (D) Volcano plot showing upregulated and downregulated proteins in female APOE4 vs. APOE3 mice. (E) IPA pathway showing top 5 upregulated and downregulated pathways in female APOE4 vs. APOE3 mice. (F) Volcano plot showing upregulated and downregulated proteins in male APOE4 vs. APOE3 mice. (G) IPA pathway analysis showing top 5 upregulated and downregulated pathways in male APOE4 vs. APOE3 mice. (H) Heatmap of proteins involved in fatty acid metabolism, cholesterol and bile acid metabolism, lipid transport, and lipid storage showing upregulated and downregulated proteins between APOE4 and APOE3 mice. (I) Venn diagram showing the number of shared upregulated proteins between comparisons of female APOE4 vs. APOE3 mice and male APOE4 vs. APOE3 mice. (J) Venn diagram showing the number of downregulated proteins shared between comparisons of female APOE4 vs. APOE3 mice and male APOE4 vs. APOE3 mice. (K) Violin plot showing normalized Pgam1 protein expression from liver proteomics in male and female APOE3 and APOE4 mice. Statistical significance was determined by two-way ANOVA with Fisher’s LSD post hoc. ∗p < 0.05. G, genotype. n = 3 mice per genotype and sex. See also Figure S1.

Figure 2

Isolated liver mitochondrial proteomics in male and female APOE3 or APOE4 mice (A) Western blot analysis assessing purity of liver mitochondrial isolates. Blots were probed for HDAC1, HK1, and CS. (B) Pie chart showing the proportion of proteins identified in isolated liver mitochondrial samples that are annotated as mitochondrial in the MitoCarta3.0 database. (C) Volcano plot showing upregulated and downregulated proteins between female APOE4 and APOE3 mice. (D) Volcano plot showing upregulated and downregulated proteins between male APOE4 and APOE3 mice. (E) IPA pathway analysis showing top upregulated and downregulated pathways between female APOE4 and APOE3 mice. (F) IPA pathway analysis showing top upregulated and downregulated pathways between male APOE4 and APOE3 mice. (G) Heatmap of oxidative phosphorylation complex I–V components between APOE4 and APOE3 female and male mice. n = 3 for proteomic outcomes. (H) Heatmap comparing whole liver to isolated mitochondria results of proteins involved in mitochondrial protein import and degradation between APOE4 and APOE3 female mice. (I) Heatmap of proteins involved in TCA cycle and pyruvate metabolism between APOE4 and APOE3 female and male mice. See also Figure S2.

Figure 3

Mitochondrial respiration from isolated liver mitochondria in male and female APOE3 and APOE4 mice (A–D) Fatty acid-driven respiration (palmitoyl-carnitine; PCoA) measured at state 2 (A), state 3 (B), state 3S (C), and uncoupled respiration (D). (E–H) Carbohydrate-driven respiration (pyruvate and malate; PM) measured at state 2 (E), state 3 (F), state 3S (G), and uncoupled respiration (H). (A–H) Statistical significance was determined by two-way ANOVA with Fisher’s LSD post hoc. Data shown as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. S = sex, G = genotype, I = interaction. n = 8 mice per genotype and sex.

Figure 4

Proteomics analysis of APOE isogenic iHLCs (A) iHLC study schematic. Two pairs of isogenic iPSCs (A and B) homozygous for either APOE3 or APOE4 were used to generate iHLCs and examine proteomic and bioenergetic outcomes. n = 5 per group for all iHLC proteomic outcomes. (B) Violin plots of APOE protein expression between two sample batches in pair A isogenics. Statistical significance was determined by unpaired t test. ns, not significant, ∗∗∗p < 0.001. (C) Volcano plot showing upregulated and downregulated proteins between APOE4 and APOE3 iHLCs from batch 1. (D) Volcano plot showing upregulated and downregulated proteins between APOE4 and APOE3 iHLCs from batch 2. (E) Heatmap of the top 20 most significant upregulated and top 20 most significant downregulated proteins (p < 0.05) shared between batch 1 and batch 2. (F) IPA pathway analysis showing top 12 upregulated and top 12 downregulated pathways shared in both batch 1 and 2 between APOE4 and APOE3 iHLCs. (G) STRING network analysis of the most significant upregulated and downregulated proteins shared between batch 1 and batch 2 in APOE4 vs. APOE3 iHLCs. (H) Cellular component Gene Ontology (GO) analysis of the top 20 upregulated and top 20 downregulated proteins in APOE4 vs. APOE3 iHLCs. See also Figure S3.

Figure 5

Mitochondrial function in APOE isogenic iHLCs (A) Mitochondrial stress test (MST) tracing from isogenic pair A. (B) Quantification of basal respiration, maximal respiration, proton leak, and ATP-production linked respiration from isogenic pair an MST. n = 13–14 per group. (C) MST tracing from isogenic pair B. (D) Quantification of basal respiration, maximal respiration, proton leak, and ATP-production-linked respiration from isogenic pair B MST. n = 12–15 per group. (E) Electron transport chain (ETC) flux through complex I, II, III, and IV in isogenic pair A. n = 7–22 per group. (F) ETC flux through complex I, II, III, and IV in isogenic pair B. n = 14–22 per group. (G) Representative tetramethyl rhodamine, ethyl ester, perchlorate (TMRE) images from isogenic pair A. Scale bars, 50 μm. (H–K) Quantification of TMRE (H), MitoSOX (I), Amplex Red (J), and Rhod-2 AM (K) fluorescence intensities from isogenic pair A. n = 8–16 per group. (L) Representative TMRE images from isogenic pair B. Scale bars, 50 μm. (M–P) Quantification of TMRE (M), MitoSOX (N), Amplex Red (O), and Rhod-2 AM (P) fluorescence intensities from isogenic pair B. n = 8–16 per group. (B, D, E–F, H–K, and M–P) Statistical significance was determined by unpaired t test. Data are shown as mean ± SD. ns = not significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 6

Glycolytic function and glucose/pyruvate oxidation in APOE isogenic iHLCs (A) Glycolytic stress test (GST) tracing from isogenic pair A. (B) Quantification of basal glycolysis, glycolytic capacity, glycolytic reserve, and non-glycolytic acidification from isogenic pair A GST. n = 11 per group. (C) GST tracing from isogenic pair B. (D) Quantification of basal glycolysis, glycolytic capacity, glycolytic reserve, and non-glycolytic acidification from isogenic pair B GST. n = 10–16 per group. (E) Glucose/pyruvate oxidation stress test from isogenic pair A. (F) Quantification of basal respiration, acute response, and maximal respiration from isogenic pair A glucose/pyruvate oxidation stress test. n = 7–10 per group. (G) Glucose/pyruvate oxidation stress test from isogenic pair B. (H) Quantification of basal respiration, acute response, and maximal respiration from isogenic pair B glucose/pyruvate oxidation stress test. n = 7–11 per group. (I) Significant DE proteins involved in glucose metabolism pathway from IPA. (J) Glucose metabolism protein network from STRING analysis. (B and D) Statistical significance was determined by unpaired t test and (F and H) two-way ANOVA with Fisher’s LSD post hoc. Data are shown as mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. D, drug, G, genotype.

Figure 7

Fatty acid oxidation and LipidTOX staining in APOE isogenic iHLCs (A) Long chain fatty acid (LCFA) oxidation stress test from isogenic pair A. (B) Quantification of basal respiration, acute response, and maximal respiration from isogenic pair An LCFA oxidation stress test. n = 6–8 per group. (C) LCFA oxidation stress test from isogenic pair B. (D) Quantification of basal respiration, acute response, and maximal respiration from isogenic pair B LCFA oxidation stress test. n = 7–10 per group. (E) Representative LipidTOX images from isogenic pair A. Scale bars, 50 μm. (F) Quantification of lipid droplet (LD) numbers per cell and diameter (μm) from isogenic pair A. n = 24 per group. (G) Representative LipidTOX images from isogenic pair B. Scale bars, 50 μm. (H) Quantification of LD numbers per cell and diameter (μm) from isogenic pair B. n = 24 per group. (I) Comparison analysis showing significantly altered proteins in SREBF2 network from proteomics. (B and D) Statistical significance was determined by a two-way ANOVA with Fisher’s LSD post hoc and (F and H) unpaired t test. Data are shown as mean ± SD. ns = not significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. D, drug, G, genotype.

Figure 8

Comparative proteomic analysis of whole mouse livers and iHLCs (A) Venn diagram depicting overlap of significant DE proteins in APOE4 vs. APOE3 across whole mouse livers and iHLCs (batch 1 and 2). (B) Gene Ontology Biological Process (GO-BP) and KEGG pathway analysis of the 70 proteins shared across mouse liver and iHLCs, showing protein counts per category. (C) Heatmap of proteins consistently upregulated or downregulated in APOE4 vs. APOE3 across whole mouse liver and iHLCs (batch 1 and 2). (D–E) Western blot analysis of CYP27A1, FDFT1, and PGAM1 in APOE4 and APOE3 whole mouse livers, with relative densitometry quantification. n = 16 per group. (F–G) Western blot analysis of CYP27A1, FDFT1, and PGAM1 in APOE4 and APOE3 isogenic iHLCs, with relative densitometry quantification. n = 6 per group. (E and G) Statistical significance was determined by unpaired t test. Data are shown as mean ± SD. ∗p < 0.05.