Imaging decreased brain docosahexaenoic acid metabolism and signaling in iPLA(2)β (VIA)-deficient mice.

Ca2+-independent phospholipase A2β (iPLA2β) selectively hydrolyzes docosahexaenoic acid (DHA, 22:6n-3) in vitro from phospholipid. Mutations in the PLA2G6 gene encoding this enzyme occur in patients with idiopathic neurodegeneration plus brain iron accumulation and dystonia-parkinsonism without iron accumulation, whereas mice lacking PLA2G6 show neurological dysfunction and neuropathology after 13 months. We hypothesized that brain DHA metabolism and signaling would be reduced in 4-month-old iPLA2β-deficient mice without overt neuropathology. Saline or the cholinergic muscarinic M1,3,5 receptor agonist arecoline (30 mg/kg) was administered to unanesthetized iPLA2β−/−, iPLA2β+/−, and iPLA2β+/+ mice, and [1-14C]DHA was infused intravenously. DHA incorporation coefficients k* and rates Jin, representing DHA metabolism, were determined using quantitative autoradiography in 81 brain regions. iPLA2β−/− or iPLA2β+/− compared with iPLA2β+/+ mice showed widespread and significant baseline reductions in k* and Jin for DHA. Arecoline increased both parameters in brain regions of iPLA2β+/+ mice but quantitatively less so in iPLA2β−/− and iPLA2β+/− mice. Consistent with iPLA2β’s reported ability to selectively hydrolyze DHA from phospholipid in vitro, iPLA2β deficiency reduces brain DHA metabolism and signaling in vivo at baseline and following M1,3,5 receptor activation. Positron emission tomography might be used to image disturbed brain DHA metabolism in patients with PLA2G6 mutations.


Chemical analysis
Arterial blood samples collected before, during, and after [1-14 C]DHA infusion were centrifuged immediately (30 s, 18,000 g ). For each sample, total lipids were extracted ( 37 ) from plasma (5 µl) with chloroform-methanol (1 ml, 2:1, v/v) and 0.1 M KCl (0.5 ml). Radioactivity was determined in an organic phase aliquot (100 µl) by liquid scintillation spectrometry. After the 3 min [1- 14 C]DHA infusion, at least 98% of total plasma radioactivity was unmetabolized [1-14 C]DHA, and 95% of the total brain radioactivity was in the form of esterifi ed [1-14 C]DHA in phospholipid ( 25 ). Concentrations of unlabeled unesterifi ed DHA also were determined in arterial plasma (100 µl) to calculate J in . Total lipids were extracted ( 37 ) and separated by thin layer chromatography on silica gel-60 plates by using the solvent system heptane:diethylether:glacial acetic acid (60:40:3, v/v/v). Unesterifi ed fatty acids were scraped from the plate and converted to methyl ester derivatives (1% H 2 SO 4 in methanol, 3 h, 70°C), which then were analyzed by gas chromatography with fl ame ionization detection and quantifi ed relative to an internal standard, heptadecanoic acid (17:0).

Quantitative autoradiography
Frozen brains were cut in serial 20-m-thick coronal sections on a cryostat at Ϫ 20°C, then placed for 4 weeks with calibrated [ 14 C]methylmethacrylate standards (Amersham, Arlington Heights, IL) on Ektascan C/RA fi lm (Eastman Kodak Co., Rochester, NY). Radioactivity (nCi/g wet weight brain) in 81 identifi ed regions ( 38 ) was measured bilaterally six times by quantitative densitometry by using the public domain NIH Image program 1.62. Regional DHA incorporation coeffi cients k* (ml/s/g wet weight brain) were calculated as ( 8,23 ): where * brain c (nCi/g wet brain weight) is radioactivity of brain lipid at time 20 min (time of termination of experiment), * plasma c (nCi/ml plasma) is the arterial plasma concentration of labeled unesterifi ed DHA, and t (min) is time after beginning [1-14 C] DHA infusion. Integrated plasma radioactivity due to unesterifi ed [1-14 C]DHA (input function) was determined by trapezoidal integration and used to calculate regional values of k*.
In the present study, we imaged k* and J in for DHA in brains of unanesthetized iPLA 2 ␤ Ϫ / Ϫ , iPLA 2 ␤ +/ Ϫ , and iPLA 2 ␤ +/+ mice ( 24 ) at baseline and following administration of arecoline ( 7,10,33,34 ). Based on the in vitro evidence cited above that iPLA 2 ␤ selectively hydrolyzes DHA from phospholipid, we predicted that brain DHA signaling would be reduced at rest and following arecoline in the iPLA 2 ␤ -defi cient compared with wild-type mice. To minimize the effects of neuropathology that appear in older iPLA 2 ␤ Ϫ / Ϫ mice, we studied 4-month-old mice free of signifi cant histopathology or neurological abnormalities ( 35 ). An abstract of part of this work has been published ( 36 ).

Surgical procedures and tracer infusion
A mouse was anesthetized with 2-3% halothane in O 2 , and PE 10 polyethylene catheters were inserted into the right femoral artery and vein. The wound site was closed with 454 Instant Adhesive (Loctite Corp., Hartford, CT), and the animal was wrapped loosely with the upper body remaining free in a fast-setting plaster cast taped to a wooden block and allowed to recover from anesthesia (3-4 h) in a warm environment. in iPLA 2 ␤ +/+ mice injected with arecoline compared with saline. The arecoline-induced elevations appear reduced or absent in the iPLA 2 ␤ -defi cient mice.
Baseline. In a one-way ANOVA with Tukey's post hoc test, we compared baseline values of k* for DHA among the three genotypes. Partial and total iPLA 2 ␤ deletion signifi cantly decreased baseline k* by 20-45% in 60 and 70 of 81 brain regions, respectively, compared with baseline k* in iPLA 2 ␤ +/+ mice (data not shown). The baseline decreases were comparable in iPLA 2 ␤ Ϫ / Ϫ and iPLA 2 ␤ Arecoline activation. Mean DHA incorporation coefficients k* at baseline and following arecoline in each of 81 brain regions were compared among experimental groups and conditions using a two-way ANOVA. Thirty-fi ve of the 81 regions did not have a statistically signifi cant genotype × drug interaction, indicating that the iPLA 2 ␤ genotype (heterozygous or homozygous) did not alter the arecoline response (data not shown). These regions included the median eminence, white matter, lateral and anterior arcuate nucleus, periventricular hypothalamus, mammillary nucleus, medial and lateral nuclei of the septum, nucleus accumbens, amygdala, CA1 to CA3 areas of the hippocampus, prefrontal cortex layers I and IV, primary olfactory cortex, globus pallidus, habenular nuclei, medial geniculate nucleus, substantia nigra, ventroposterior medial, and paraventricular and parafascicular thalamic nuclei. In the 35 regions, the main effect of arecoline was statistically signifi cant, which means that increments in k* following arecoline were equally robust in the three genotypes. Increments (compared with saline) ranged from 32% in the median eminence to 170% in prefrontal cortex layer IV (mean = 78 ± 34%).

Regional DHA incorporation rates
Because the mean total (labeled and unlabeled) unesterifi ed plasma DHA concentration did not differ signifi cantly among genotypes (see above), the statistical signifi cance Regional incorporation rates of unesterifi ed unlabeled DHA from plasma into brain, J in (nmol/s/g), were calculated as: where C plasma equals unesterifi ed unlabeled plasma DHA.
Regional DHA incorporation coeffi cients k* Figure 1 presents color-coded coronal autoradiographs representing k* for DHA from brains of iPLA 2 ␤ -defi cient and wild-type mice injected with saline or arecoline. iPLA 2 ␤ +/ Ϫ and iPLA 2 ␤ Ϫ / Ϫ mice apparently had lower baseline (following saline) values of k* (Eq. 1) than did iPLA 2 ␤ +/+ controls. Regional values of k* appear elevated overlap with sites having high densities of postsynaptic M 1,3,5 receptors ( 41,42 ). These regions include neocortical projection regions, parts of the hippocampus, and the caudate-putamen. Low M 1,3,5 receptor densities are reported in the thalamus, brainstem, and hypothalamus, where genotype × drug interactions often were statistically insignifi cant. The arecoline-induced increments in DHA incorporation in these latter regions may have represented downstream effects of direct activation elsewhere ( 43 ). Because DHA cannot be synthesized de novo in vertebrates ( 30 ) and only a negligible amount (<0.1%) is elongated in brain from precursor ␣ -LNA ( 8,28,31,32 ), the lower values of k* and J in at baseline and following arecoline in the iPLA 2 ␤ Ϫ / Ϫ and iPLA 2 ␤ +/ Ϫ compared with iPLA 2 ␤ +/+ mice represent reduced brain DHA consumption under resting (steady-state) and agonist stimulation conditions, respectively. In iPLA 2 ␤ Ϫ / Ϫ mice, these reductions are associated with a reduced DHA concentration in brain ethanolamine glycerophospholipid (Y. Cheon, A. Taha, H. Y. Kim, and S. I. Rapoport, unpublished observations). A role for iPLA 2 ␤ in regulating brain DHA metabolism is consistent with evidence that brain DHA turnover is reduced, as are the brain DHA concentration and iPLA 2 ␤ mRNA, protein, and activity levels, in rats fed a low n-3 PUFA diet lacking DHA ( 31,44 ). Increased incorporation of labeled unesterifi ed DHA from plasma into the sn -2 position of synaptic membrane phospholipids of brain has been demonstrated directly by chemical analysis in unanesthetized rats given arecoline ( 7,10 ). Our new data suggest that a congenital absence of of group differences in incorporation rates J in corresponded generally to the differences in regional values of k*, because J in is the product of k* and the unesterifi ed unlabeled plasma DHA concentration (Eq. 2). In iPLA 2 ␤ +/+ mice, baseline J in ranged from 311 ± 49 × 10 Ϫ 4 nmol/s/g in the piriform cortex to 898 ± 196 × 10 Ϫ 4 nmol/s/g in the inferior colliculus. In iPLA 2 ␤ Ϫ / Ϫ mice, the range was 267 ± 38 × 10 Ϫ 4 nmol/s/g in the internal capsule to 581 ± 55 × 10 Ϫ 4 nmol /s/g in the inferior colliculus. In iPLA 2 ␤ +/ Ϫ mice, baseline J in ranged from 166 ± 28 × 10 Ϫ 4 nmol/s/g in the periventricular of the hypothalamus to 487 ± 128 × 10 Ϫ 4 nmol/s/g in the inferior colliculus. Similarly, in response to arecoline, means for J in decreased signifi cantly in iPLA 2 ␤ Ϫ / Ϫ and iPLA 2 ␤ +/ Ϫ compared with iPLA 2 ␤ +/+ mice.
In summary, a congenital partial or complete absence of iPLA 2 ␤ in 4-month-old mice reduced brain DHA signaling and metabolism at baseline and following M 1,3,5 receptor activation. Studies in 13-month-old iPLA 2 ␤ -defi cient mice with neurological and behavioral impairments that correlate with brain accumulation of ubiquitin-containing tubulovesicular membranes ( 35, 60 ) may identify additional brain lipid metabolic disturbances. A detailed analysis of brain enzyme activity, lipid composition ( 68 ), and PUFA metabolism of mice at both ages could be informative and clinically relevant. iPLA 2 ␤ reduces this incorporation, as well as baseline signaling. The effects did not seem to depend on whether the mice had a partial or full iPLA 2 ␤ deletion, possibly because of additional compensatory neuroplastic adaptive responses associated with the full deletion ( 45,46 ).
Although we considered iPLA 2 ␤ deletion effects on brain DHA signaling in this paper, based on in vitro evidence of the enzyme's selectivity for DHA (1)(2)(3)(4)6 ), AA signaling may be disturbed as well, because brain activities of sPLA 2 and cPLA 2 -IV, which release AA from phospholipids ( 59 ), are elevated in iPLA 2 ␤ Ϫ / Ϫ mice, whereas the brain AA concentration in phospholipid is reduced (Y. Cheon, A. Taha, H. Y. Kim, and S. I. Rapoport, unpublished observations). These enzyme changes would represent additional indirect compensatory responses in the full knockout condition ( 45,46 ). iPLA 2 ␤ is reported to modulate apoptosis, cell proliferation, membrane fusion, behavior, memory, and motor function ( 14,(60)(61)(62). Some of these effects may be due to its infl uence on brain DHA metabolism and signaling. iPLA 2 ␤ mRNA is reduced in the hippocampus of aged rats ( 63 ), but its brain mRNA and protein levels are increased in multiple sclerosis patients and in mice with experimental autoimmune encephalomyelitis (for which pretreatment with a selective iPLA 2 inhibitor was helpful) ( 64 ).
Because DHA is a precursor of antiinfl ammatory neuroprotectins and resolvins, the reduced brain DHA metabolism in the iPLA 2 ␤ -defi cient mice may increase their vulnerability to neuroinfl ammation ( 9,65,66 ). Our observations also suggest that brain DHA signaling and metabolism would be altered in patients with a PLA2G6 mutation