Disturbed neurotransmitter homeostasis in ether lipid deficiency

Abstract

Plasmalogens, the most prominent ether (phospho)lipids in mammals, are structural components of most cellular membranes. Due to their physicochemical properties and abundance in the central nervous system, a role of plasmalogens in neurotransmission has been proposed, but conclusive data are lacking. Here, we targeted this issue in the glyceronephosphate O-acyltransferase (Gnpat) KO mouse, a model of complete deficiency in ether lipid biosynthesis. Throughout the study, focusing on adult male animals, we found reduced brain levels of various neurotransmitters. In the dopaminergic nigrostriatal tract, synaptic endings but not neuronal cell bodies were affected. Neurotransmitter turnover was altered in ether lipid-deficient murine as well as human post-mortem brain tissue. A generalized loss of synapses did not account for the neurotransmitter deficits, since the levels of several presynaptic proteins appeared unchanged. However, reduced amounts of vesicular monoamine transporter indicate a compromised vesicular uptake of neurotransmitters. As exemplified by norepinephrine, the release of neurotransmitters from Gnpat KO brain slices was diminished in response to strong electrical and chemical stimuli. Finally, addressing potential phenotypic correlates of the disturbed neurotransmitter homeostasis, we show that ether lipid deficiency manifests as hyperactivity and impaired social interaction. We propose that the lack of ether lipids alters the properties of synaptic vesicles leading to reduced amounts and release of neurotransmitters. These features likely contribute to the behavioral phenotype of Gnpat KO mice, potentially modeling some human neurodevelopmental disorders like autism or attention deficit hyperactivity disorder.

Figure 1

Ether lipid deficiency manifests as neurotransmitter deficits in the murine cerebrum. The levels of dopamine (A), norepinephrine (B), serotonin (C), GABA (D), glycine (E), glutamate (F) and taurine (G) were determined in cerebral homogenates from WT and Gnpat KO mice (A–H, n = 6 for both genotypes) by using HPLC. Statistical analysis was performed using two-tailed Student’s t-tests. Graphs depict individual data together with group mean ± SD. (H) For control reasons, the levels of several proteinogenic amino acids were analyzed. Statistical analysis using two-tailed Student’s t-tests did not reveal any significant differences (no correction for multiple comparisons due to the absence of significant results). Bars represent group mean ± SD. ^***P < 0.001; ^**P < 0.01; ^*P < 0.05; n.s., not significant. The displayed dopamine and serotonin data were used for the calculation of the neurotransmitter-to-metabolite ratios presented in Table 1.

Figure 2

A synapse-related problem accounts for the dopamine deficit in the ether lipid-deficient mouse brain. The cartoon at the top illustrates the nigrostriatal pathway with cell bodies of dopaminergic neurons in the substantia nigra and their axon terminals in the striatum. Dopamine levels were determined in the striatum (left panel; n = 7/genotype) and the substantia nigra (right panel; WT: n = 12, Gnpat KO: n = 11) of WT and Gnpat KO mice by using HPLC. Graphs depict individual data together with group mean ± SD. Statistical analysis was performed using two-tailed Student’s t-tests. ^*P < 0.05; n.s., not significant. The displayed data were used for the calculation of the neurotransmitter-to-metabolite ratios presented in Table 1.

Figure 3

No general synapse loss in ether lipid-deficient mouse brains. (A) Western blot analysis of cortical membrane–enriched extracts (5 μg per lane) derived from WT and Gnpat KO mice (n = 4/genotype) testing for the amounts of synaptotagmin I (Syntag, A) and synaptophysin (Synphys, B). Immunoblots were stripped and reprobed with actin (A) and the membrane marker transferrin receptor (TfR, B) as loading controls. Densitometric quantification is shown as individual data together with group mean ± SD. Statistical analysis was performed using two-tailed Student’s t-tests. ^**P < 0.01; n.s., not significant. Full blots for the presented data can be found in Figure S2.

Figure 4

Norepinephrine release from brain slices is impaired upon ether lipid deficiency. (A) Hippocampal slices were loaded with [³H]-norepinephrine, and neurotransmitter release was induced by two consecutive electropulse series (100 pulses, 10 Hz; black arrows) and exposure to 40 mM KCl for 30 s (green arrow). The reuptake inhibitor desipramine was added after the first electrical stimulus (shaded area). The results of a representative experiment are shown in the upper panel. Each data point indicates the mean ± SD of six brain slices. A summary of all experiments (n = 10) for each of the three peaks is provided in the lower panels. Connected data points derive from the same experiment and bars indicate means. Statistical analysis was performed using paired Student’s t-tests. (B) Hippocampal brain slices were processed like in (A) and stimulated by three consecutive strong electropulse series (500 pulses, 50 Hz; black arrows). The results of a representative experiment are shown in the upper panel. Each data point indicates the mean ± SD of six brain slices. A summary of all experiments (n = 10) for each of the three peaks is provided in the lower panels. Connected data points derive from the same experiment and bars indicate means. Statistical analysis was performed using paired Student’s t-tests. ^**P < 0.01; n.s., not significant.

Figure 5

Binding assays targeting key proteins in synaptic transmission in cortical membrane and striatal vesicle preparations from WT and ether lipid-deficient mice. (A) Radioactive binding assays using the specific ligand [³H]-nisoxetine provided the levels of NET in cortical membrane fractions from WT and Gnpat KO mice (n = 5/genotype). The left graph shows the results of one representative experiment expressed as mean ± SD of technical triplicates. The numbers on the x-axis indicate individual mice. The right graph summarizes all results with individual data points representing means derived from two independent experiments. In addition, the group mean ± SD of all analyzed animals is depicted. (B) Radioactive binding assays using the specific ligand [³H]-imipramine provided the levels of SERT in cortical membrane fractions from WT and Gnpat KO mice (n = 5/genotype). Individual data (means of two independent experiments) and group mean ± SD of all analyzed animals are depicted. (C) Radioactive binding assays using the specific ligand [³H]-WIN 35,428 provided the levels of DAT in striatal membrane fractions from WT and Gnpat KO mice (n = 5/genotype). Individual data and group mean ± SD of all analyzed animals are depicted. (D) Radioactive binding assays using the specific ligand [³H]-MK-801 provided the levels of NMDA receptor in cortical membrane fractions from WT and Gnpat KO mice (n = 10/genotype). For technical reasons, all mice could not be analyzed in the same experiment; thus, data were normalized to the mean of all animals within each experiment. Individual data representing means of two independent experiments and group mean ± SD of all analyzed animals are depicted. (E) Radioactive binding assays using the specific ligand [³H]-DTBZ provided the levels of VMAT2 in striatal vesicular preparations from WT (n = 4) and Gnpat KO (n = 3) mice. Data are shown individually and as group mean ± SD. Statistical analysis for all panels was performed using two-tailed Student’s t-tests. ^*P < 0.05; n.s., not significant.

Figure 6

Ether lipid-deficient mice show hyperactivity and impaired sociability. (A) Results of WT (n = 13) and Gnpat KO (n = 11) mice in the closed-wheel test are shown as the time active during a 30-min period (left graph) and the average velocity while running (right graph). Box plots are drawn according to Tukey’s method, and statistical analysis was performed using two-tailed Student’s t-tests. The same cohort of mice was exposed to the open field paradigm for a total of 61 min, and vertical movements (B), ambulatory movements (C) and non-ambulatory movements (D) were recorded. Vertical movements were counted throughout the whole 61-min period, and box plots are drawn according to Tukey’s method. Statistical analysis was performed using Mann–Whitney U-test. For ambulatory and non-ambulatory movements, the first minute was analyzed separately and the remaining time was divided into three phases (phase 1: minute 2–21, phase 2: minute 22–41, phase 3: minute 42–61). Results are shown as mean ± SD, and statistical analysis was performed using two-tailed Student’s t-tests followed by Holm–Sidak correction for the repeated measurements. (E) Sociability was assessed by the three-chamber social interaction test, and the time spent interacting with an unfamiliar mouse (‘stranger’) or a neutral, novel object (‘object’) was quantified (WT: n = 11; Gnpat KO: n = 6). Results are shown as mean ± SD. Statistical analysis was performed using paired, two-tailed Student’s t-tests (for the comparison of the preference for stranger or object within each genotype) and repeated measures two-way ANOVA (for comparison of the genotypes; interaction effect: F(1,15) = 6.533; P = 0.022). (F) The contextual fear conditioning paradigm was applied to WT (n = 11) and Gnpat KO (n = 6) mice, and the duration of the freezing response is depicted as mean ± SD. Statistical analysis was performed using two-tailed Student’s t-test. ^***P < 0.001; ^**P < 0.01; ^*P < 0.05; n.s., not significant.