Loss of carnitine palmitoyltransferase 1a reduces docosahexaenoic acid-containing phospholipids and drives sexually dimorphic liver disease in mice

Abstract

Background and aims: Genome and epigenome wide association studies identified variants in carnitine palmitoyltransferase 1a (CPT1a) that associate with lipid traits. The goal of this study was to determine the role of liver-specific CPT1a on hepatic lipid metabolism.

Approach and results: Male and female liver-specific knockout (LKO) and littermate controls were placed on a low-fat or high-fat diet (60% kcal fat) for 15 weeks. Mice were necropsied after a 16 h fast, and tissues were collected for lipidomics, matrix-assisted laser desorption ionization mass spectrometry imaging, kinome analysis, RNA-sequencing, and protein expression by immunoblotting. Female LKO mice had increased serum alanine aminotransferase levels which were associated with greater deposition of hepatic lipids, while male mice were not affected by CPT1a deletion relative to male control mice. Mice with CPT1a deletion had reductions in DHA-containing phospholipids at the expense of monounsaturated fatty acids (MUFA)-containing phospholipids in whole liver and at the level of the lipid droplet (LD). Male and female LKO mice increased RNA levels of genes involved in LD lipolysis (Plin2, Cidec, G0S2) and in polyunsaturated fatty acid metabolism (Elovl5, Fads1, Elovl2), while only female LKO mice increased genes involved in inflammation (Ly6d, Mmp12, Cxcl2). Kinase profiling showed decreased protein kinase A activity, which coincided with increased PLIN2, PLIN5, and G0S2 protein levels and decreased triglyceride hydrolysis in LKO mice.

Conclusions: Liver-specific deletion of CPT1a promotes sexually dimorphic steatotic liver disease (SLD) in mice, and here we have identified new mechanisms by which females are protected from HFD-induced liver injury.

Keywords: Fatty acid metabolism; Lipid droplets; Lipolysis; Mitochondria; Triglycerides.

Graphical abstract

Figure 1

Liver-specific CPT1a Deletion Has No Effect on Body Weight in Response to HFD-Feeding. Male and female control and LKO mice were fed a HFD for 15-weeks. (A–C) Liver CPT1a RNA and protein levels were measured by qPCR (A; n = 6–8) and western blot (B; n = 5) followed by densitometry (C; n = 5), respectively. Vinculin is used as a loading control. (D) Cpt1b RNA levels were measured by qPCR from livers of HFD-fed male and female control and LKO mice (n = 6–8). For all qPCR analyses, housekeeping genes Tbp and Hprt were averaged and used for normalization. (E) β-hydroxybutyrate levels were measured from the plasma of fasted mice (n = 6–8). (F, G) Percent body weight (F) and fat (G) mass were recorded throughout the duration of the study (n = 8–14). Significance was determined by two-way ANOVA with Tukey's multiple comparison post hoc analysis. ∗P < 0.05; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001.

Figure 2

Female LKO Mice Develop Exacerbated Steatosis and Liver Dysfunction in Response to HFD-Feeding. Male and female control and LKO mice were fed a HFD for 15-weeks. (A–C) Liver weights normalized to body weight (A; LW:BW, %; n = 8–14), and hepatic triglycerides (B; n = 8) and total cholesterol (C; n = 8) levels were quantified enzymatically. (D) Serum ALT levels were quantified (n = 7–8). (E, F) Representative Oil Red O staining to assess neutral lipid accumulation across groups. The scale bar (50 μm) is embedded within the lumen of the portal (E) or central (F) veins in the corner of each image. Total number of lipid droplets (# lipid droplets/nuclei in 100X field) and their respective diameters were quantified in periportal (G) and pericentral hepatocytes (H). A total of 17,473 lipid droplets were quantified across 19 mice. (I) Transmission electron microscopy was completed on livers collected from HFD-fed female control and LKO mice (scale bars = 10 μm on left image and 1 μm on right image). Significance was determined by two-way ANOVA with Tukey's multiple comparison post hoc analysis. ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001.

Figure 3

Female LKO Mice Have Increased PE and MUFA-Containing Phospholipids in the Liver. Male and female control and LKO mice were fed a HFD for 15-weeks. (A) Whole cell liver lysates were subjected to LC-MS/MS based lipidomics. Phospholipid species (PE, PC, PS, PI, PG, PA) displayed as percent (%) of total phospholipids (n = 6–8). (B) A ratio of total PC to PE across the four groups (n = 6–8). (C) Total read counts of Pemt from bulk RNA-sequencing data provided in Figure 6 (FPKM = fragments per kilobase million; n = 6). (D, E) Immunoblotting (D) followed by densitometry (E) for PEMT protein (n = 5). Vinculin is used as a loading control. (F, G) The fatty acyl composition of PE (F) and PC (G) from female (top) and male (bottom) mice. All data are presented as % of total phospholipid (n = 7–8). Significance was determined by an unpaired Student's t-test (panels A, F, G) or by a two-way ANOVA with Tukey's multiple comparison post hoc analysis (panels B, C, E). ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001.

Figure 4

Lipid Droplets Isolated from LKO Mice Have Increased MUFA-Containing Phospholipids. Male and female control and LKO mice were fed a HFD for 15-weeks. (A) Non-LD and LD fractions were isolated from livers and immunoblotted for PLIN2, PLIN5, VDAC, and GAPDH (n = 3). (B, C) Lipid droplet fractions were subjected to targeted semi-quantification of PC (B) and PE (C; n = 3–5). (D, E) MALDI-MSI was completed on whole livers from female control and LKO mice (n = 4). A representative spatial map and pixel intensity are shown for 36:3 PE (D) and 38:6 PE (E). Significance was determined by unpaired Student's t-tests. ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001.

Figure 5

PPAR Signaling and Lipid Droplet Genes Are Elevated in LKO Mice. Bulk RNA sequencing was completed on the livers of male and female HFD-control and LKO mice. (A, B) Volcano plots in males (A) and females (B) highlighting all genes increased (in blue) or decreased (in red) with Cpt1a-deficiency. The horizontal black bar denotes the significance cutoff FDR = 0.05. The vertical black bars denote a minimal threshold for the effect size of 1.5 (Log2 fold change = ±0.58). (C, D) KEGG analysis dot plots for the top 15 dysregulated pathways in WT and LKO male (C) and female (D) mice. (E, F) Total read counts (in FPKM) for genes involved in the biosynthesis of PUFAs (E; Elovl5, Fads1, Elovl2) and fatty acyl remodeling of phospholipids (F; Acsl4, Tlcd1). Significance was determined by a two-way ANOVA with Tukey's multiple comparison post hoc analysis (panels E, F). ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001.

Figure 6

PKA Signaling and Triglyceride Hydrolysis Are Impaired in LKO Mice. A serine-threonine kinome analysis was completed on pooled liver samples from male HFD-fed control and LKO mice. A total of 6 mice per genotype were pooled and ran in triplicate on the PamGene PamStation. (A–C) PKA peacock (A), MEOW (B), and substrate waterfall plots (C) are presented comparing male control and LKO mice. The waterfall plot shows all known peptides that are phosphorylated by PKA and the extent by which their phosphorylation, in real-time, is increased or decreased (in red) with Cpt1a-deficiency. (D, E) Immunoblotting (D) followed by densitometry (E) for proteins (PLIN2, PLIN5, ATGL, HSL, MGL, CGI-58, G0S2) associated with regulating the size and turnover of LDs (n = 5). Vinculin is used as a loading control. (F) Triglyceride hydrolysis was completed using liver lysates from HFD-fed control and LKO mice using resorufin ester as substrate (n = 8). Significance was determined by a two-way ANOVA with Tukey's multiple comparison post hoc analysis. ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ∗∗∗∗P < 0.0001.