Phospholipid abnormalities in postmortem schizophrenic brains detected by 31P nuclear magnetic resonance spectroscopy: a preliminary study

Psychiatry Research: Neuroimaging Section 106 Ž2001. 171᎐180

Phospholipid abnormalities in postmortem schizophrenic brains detected by 31 P nuclear magnetic resonance spectroscopy: a preliminary study Richard A. Komoroskia,b,c,d,f,U , John M. Pearce a , W. Sue T. Griffin e, Robert E. Mrak b,f , Masao Omori g , Craig N. Karsonb,c,f,1 a

Department of Radiology, Uni¨ ersity of Arkansas for Medical Sciences, Little Rock, AR 72205 USA Department of Pathology, Uni¨ ersity of Arkansas for Medical Sciences, Little Rock, AR 72205 USA Department of Psychiatry, Uni¨ ersity of Arkansas for Medical Sciences, Little Rock, AR 72205 USA d Department of Biochemistry, Uni¨ ersity of Arkansas for Medical Sciences, Little Rock, AR 72205 USA e Department of Geriatrics, Uni¨ ersity of Arkansas for Medical Sciences, Little Rock, AR 72205 USA f Central Arkansas Veterans Healthcare System, Little Rock, AR 72205 USA g Department of Neuropsychiatry, Fukui Medical Uni¨ ersity, Matsuoka, Fukui 910-1193, Japan b

c

Received 31 August 2000; received in revised form 26 February 2001; accepted 19 March 2001

Abstract It has been hypothesized that schizophrenia arises from cell membrane abnormalities due to changes in phospholipid ŽPL. composition and metabolism. We have used high resolution, in vitro 31 P nuclear magnetic resonance ŽNMR. to characterize the PLs in left frontal cortex Žgray matter. of postmortem brain from four schizophrenics and five controls. High resolution 31 P NMR spectra were obtained in an organic-solvent system to resolve PL classes Žheadgroups. and in a sodium-cholate, aqueous dispersion system to resolve phosphatidylcholine ŽPC. molecular species. Multivariate analysis which included the major PC molecular species and phosphatidylinositol ŽPI. showed a significant difference between schizophrenics and controls. Analysis of specific interactions showed that the PI was significantly higher in the schizophrenic group than in the control group. There were no differences between the two groups for other individual PL classes, or for individual PL subclasses determined by the linkage type at the sn-1 position on glycerol. There was a trend for total PL content to be higher in schizophrenics than in controls. There was no evidence for elevated lysophosphatidylcholine or lysophosphatidylethanolamine in schizophrenia. The intensity of the PC peak representing molecular species with one saturated and one unsaturated Žone or two

U

Corresponding author. NMR Lab Ž151KrNLR., VA Medical Center, 2200 Fort Roots Drive, North Little Rock, AR 72114, USA. Tel.: q1-501-257-1810; fax: q1-501-257-1811. E-mail address: [email protected] ŽR.A. Komoroski.. 1 Present address: VA Boston Healthcare System, 150 South Huntington Avenue, Boston, MA 02130, USA. 0925-4927r01r$ - see front matter 䊚 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 9 2 5 - 4 9 2 7 Ž 0 1 . 0 0 0 8 1 - 6

172

R.A. Komoroski et al. r Psychiatry Research: Neuroimaging Section 106 (2001) 171᎐180

double bonds. acyl chain was higher for the schizophrenic group than for the control group. Although these results are not in complete agreement with previous studies, they support the idea that PL abnormalities occur in the brain in schizophrenia and that fatty acid metabolism may be abnormal. 䊚 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Cell membrane; Fatty acids; Molecular species; Phosphatidylinositol; In vitro

1. Introduction It has been hypothesized that schizophrenia arises from cell membrane abnormalities due to changes in phospholipid ŽPL. composition and metabolism ŽHorrobin et al., 1994; Peet et al., 1999; Fenton et al., 2000.. Numerous studies on erythrocytes and platelets suggest that alterations in membrane PL composition may play a role in the pathophysiology of schizophrenia. Previous findings have been mixed, but generally suggested differences in phosphatidylserine ŽPS., phosphatidylcholine ŽPC., phosphatidylethanolamine ŽPE., or phosphatidylinositol ŽPI. concentrations between schizophrenics and controls ŽRıpova ´ ´ et al., 1997.. Pangerl et al. Ž1991. found increased phospholipase A 2 ŽPLA 2 . activity and lysophosphatidylcholine ŽLPC a . in platelets in schizophrenia. Although erythrocytes and platelets are convenient models for some aspects of neurometabolism, changes measured in peripheral tissue may not reflect changes in neuronal membranes and may be influenced more readily by extraneous factors. Membrane PL composition varies greatly among different body tissues, consistent with widely different functions. Thus, it is essential to measure PL composition in the brain itself, which currently is only possible postmortem. In vivo 31 P NMR can be used to measure compounds involved in PL metabolism, including low molecular weight, phosphomonoester ŽPME. precursors and phosphodiester ŽPDE. degradation products. Most in vivo 31 P NMR studies suggest alterations of PL metabolism in the frontal lobes in schizophrenia ŽWilliamson and Drost, 1999.. Because PLs are solid-like substances which have reduced molecular mobility in cell mem-

branes, their NMR resonances are very broad and typically not visible in vivo. Tissue PLs can be characterized using 31 P NMR in vitro after organic solvent extraction and suspension in a special reagent to prevent aggregation ŽLondon and Feigenson, 1979; Glonek, 1994.. Acyl-side-chain unsaturation Žmolecular species. can be partially probed by 31 P NMR analysis of solubilized, aqueous PLs ŽPearce and Komoroski, 1993.. Although high performance liquid chromatography ŽHPLC. is often used for PL analysis, this approach is labor intensive and can require substantial sample modifications and multiple analyses. High resolution 31 P NMR has the advantages of simultaneous detection and quantitation of numerous PLs in a single assay, the detection of unsuspected and unidentified PLs and the non-destructive nature of the technique. Because the number of P-atoms per molecule is known for generic PLs, molar concentrations can be determined for PLs of unknown side chain composition. We recently developed techniques for the combined analyses of PL classes Žheadgroups., subclasses Žlinkage type at the sn-1 position of glycerol., and molecular species in the human brain ŽPearce and Komoroski, 2000.. These techniques are used here to detect PL abnormalities in postmortem schizophrenic brains.

2. Methods The protocol was approved by the Human Research Advisory Committee of the University of Arkansas for Medical Sciences. Autopsies were obtained with the informed consent of next of kin. Written informed consent was obtained from each subject assessed premortem. Postmortem

R.A. Komoroski et al. r Psychiatry Research: Neuroimaging Section 106 (2001) 171᎐180

brain tissue was obtained from four schizophrenics and five controls who had previously been studied by in vitro 1 H NMR spectroscopy ŽOmori et al., 1997.. Diagnoses were made postmortem using the Diagnostic Evaluation after Death ŽDEAD. protocol ŽSalzman et al., 1983.. A previous study ŽKarson et al., 1993. confirmed that schizophrenia diagnoses made by DEAD agreed with those made according to DSM-III-R ŽAmerican Psychiatric Press, 1987.. Table 1 summarizes the pertinent information on the subjects studied here. Samples were taken from Brodmann’s area 10 Žfrontal pole. in the left brain. The section was anterior to a vertical cut 2 cm posterior to the apex of the frontal pole and superior to the orbitofrontal gyrus. For each extraction, as much white matter as possible was removed. Additional details have been given previously ŽOmori et al., 1997.. Based on a visual inspection of the partially thawed surface of each frozen sample, 0.2-0.5 g of primarily gray matter was isolated and deposited into a centrifuge tube on ice. Cold hexane-isopropanol Ž3:1. ŽHIP. was then added at a volume ratio of 18:1, based on an assumed tissue density of 1.0 grml. The mixture was homogenized in the tube by manual grinding with a teflon rod for 5 min, followed by sonication in a water-cooled Ž; 15⬚C. cup-horn flow cell assembly ŽFisher Model 50 Sonic Dismembrator. for 5 min. Homo-

173

genates were then centrifuged at 2500 = g for 5 min and the supernatants removed and stored at y20⬚C with 0.1% BHT prior to 31 P NMR analysis. The supernatants from subject numbers 4, 5 and 9 in Table 1 were each subjected to an additional washing step using aqueous Na 2 SO4 Ž1 gr15 ml H 2 O, half volume of organic phase.. The resulting supernatant was combined with the supernatant obtained by re-extraction of the aqueous wash phase with HIP Ž7:2, double the volume of the aqueous phase.. Samples were prepared for analysis in two solvent systems ŽPearce and Komoroski, 2000.. The CHCl 3 ᎐CH 3 OH᎐H 2 O system ŽPL cm ., normally best for PL class and subclass analysis, was employed first. Typically 0.30 ml of a 2.00 mgrml solution of lysophosphatidylglycerol ŽLPGa . as internal quantitation reference in CHCl 3 ᎐CH 3 OH Ž2:1. was added to the centrifuged supernatant, which was then dried under a stream of N2 gas. ŽAlternatively, for subject numbers 4, 5 and 9 in Table 1, a comparable amount of aqueous, sodium cholate-solubilized 16:0r16:0-phosphatidylglycerol was added as a quantitation reference during later sample preparation Žsee below. for quantitative analysis of the aqueous bile salt spectrum only.. Trace H 2 O removal was facilitated by small additions of CHCl 3 near the endpoint andror by lyophilization. To the dried extract, 2.00 ml of CDCl 3 and 0.80 ml of CH 3 OH were

Table 1 Characteristics of Subjects Subjectr group

Age Žyears.a r sex

Post-mortem interval Žh.b

Cause of death

Duration of psychiatric illness Žyears.

Psychiatric medication

1rcontrol 2rcontrol 3rcontrol 4rcontrol 5rcontrol 6rschiz. 7rschiz. 8rschiz. 9rschiz.

74rM 49rF 68rM 70rM 73rM 63rM 42rF 82rM 65rM

7 7 4 5 4.5 7.5 23 9 16

Lung cancer Pulmonary embolism Pulmonary embolism Lung cancer Pneumonia COPDc Pulmonary embolism Myocardial infarction Pneumonia

᎐ ᎐ ᎐ ᎐ ᎐ 33 21 37 32

᎐ ᎐ ᎐ ᎐ ᎐ Trifluoperazine Clozapine Trifluoperazine Haloperidol

a

Schizophrenic subjects Ž63.0" 16.4. not significantly different from controls Ž66.8" 10.2. Ž t-test, Ps 0.68, d.f.s 7.. Schizophrenic subjects Ž13.9" 7.1. significantly different from controls Ž5.5" 1.4. Ž t-test, Ps 0.035, d.f.s 7.. c Chronic obstructive pulmonary disease. b

174

R.A. Komoroski et al. r Psychiatry Research: Neuroimaging Section 106 (2001) 171᎐180

immediately added and the extract was dissolved by vortexing briefly. Finally, 0.40 ml of deionized H 2 O containing 0.1 M Cs᎐EDTA at pH 6.0 was added. The mixture was vortexed again and the two phases were allowed to settle for 15᎐60 min before transferring them, lower phase first, to a 10-mm NMR tube. The 31 P NMR spectrum was acquired as described below. Following the PL cm experiment, the same sample was prepared for analysis in an aqueous dispersion of a bile salt micelle ŽPL ad . ŽPearce and Komoroski, 2000.. The dissolved extract was concentrated under a stream of N2 gas, then lyophilized to remove residual H 2 O. Sodium cholate Ž240 mg. and 99.9 at.% D 2 O Ž2.1 ml. were then added, followed by 1᎐5 min of gentle vortexing and immersion in a 50-W bath sonicator at 35᎐45⬚C. Sample pD was adjusted to approximately 9 with 0.1 N KOD prior to filtration through a 0.7-␮m glass-fiber-syringe filter into a 10-mm NMR tube, where the final pD was ad-

justed to 9.3᎐9.4 and the volume to 3.0 ml using KOD and D 2 O, respectively. The optically clear sample was then briefly resonicated, bubbled with N2 gas for 1᎐2 min, capped and equilibrated at the desired temperature in the spectrometer. High resolution 31 P NMR spectra were acquired at 121.65 MHz on a General Electric GN-300WB spectrometer Ž7.05 T. equipped with a 10-mm broadband probe. The PL cm samples were spun at 10᎐12 rps and the temperature was controlled at 25⬚C. In a one-pulse experiment with a 90⬚ pulse width of 16 ␮s and a pulse-repetition time of 10 s, 1 K complex points were collected over a spectral width of "256 Hz for 1200 acquisitions. For PL ad , a higher temperature Ž27 or 32⬚C. with more acquisitions Ž1600. and a wider spectral width Ž"512 Hz. was used. Because 31 P spin-lattice relaxation times ŽT1 . in these systems are typically 1.5᎐2.0 s ŽPearce and Komoroski, 1993., saturation of resonance intensities was minimal and no intensity corrections

Table 2 Phospholipid concentrations ŽmM. in left frontal cortex of schizophrenics and controls a Phospholipid

Controls Ns5

Schizophrenics Ns4

Cardiolipin, CL Alkylacyl-phosphatidylethanolamine, PEe Phosphatidylethanolamine plasmalogen, PEp Phosphatidylethanolamine, PEa Phosphatidylserine, PS Sphingomyelin, SM Phosphatidylinositol, PI Phosphatidylcholine Žboth chains saturated., dsPC Phosphatidylcholine Žunknown, possibly PCp ., PCu Phosphatidylcholine wone saturated q one unsaturated chain Žms 1,2.x, suL PC Phosphatidylcholine wsuH Žm) 3. q both chains unsaturated Žms 1,2.x ŽsuH q duL .PC Phosphatidylcholine plasmalogen, PCpe Alkylacyl-phosphatidylcholine, PCe e Total phospholipidf

0.48" 0.13 0.72" 0.19 8.68" 1.00 7.28" 0.86 5.52" 1.08 3.50" 0.62 1.34" 0.18 4.10" 0.84 1.36" 0.25 8.14" 0.69

0.50" 0.18 0.72" 0.21 10.28" 2.34 7.72" 0.81 6.38" 1.10 4.00" 0.24 1.65" 0.24b,c 3.82" 0.87b 1.55" 0.30 9.20" 0.37b,d

2.46" 0.57

2.65" 0.59b

1.22" 0.19 0.25" 0.17 43.7" 3.3

1.80" 0.67 0.31" 0.21 49.0" 4.3g

a

Mean " S.D. Significant difference between schizophrenics and controls by MANCOVA Ž Ps 0.009, Wilks’ lambda s 0.024.. c Significantly different from controls Ž Ps 0.05, F Ž1,6. s 5.97.. d Significantly different from controls Ž Ps 0.045, F Ž1,6. s 6.34.. e Quantified in the PL cm spectrum via SM Žsee text.. f Includes SM and CL, as well as PC p and PC e quantified in the PL cm spectrum, but not PC u . g Different from controls Ž Ps 0.085, F Ž1,6. s 4.25.. b

R.A. Komoroski et al. r Psychiatry Research: Neuroimaging Section 106 (2001) 171᎐180

175

Fig. 1. The 31 P NMR spectra at 121.65 MHz of PLs in left frontal cortex Žgray matter. from a control subject Ž噛2 in Table 1.: Žbottom. using the PL cm system; Žtop. using the PL ad system. The compound 1᎐16:0-LPGa , which isomerizes partially to 2᎐16:0-LPGa , was added as a quantitation standard.

were necessary. WALTZ-16 broadband proton decoupling was used during acquisition only Žinverse-gated decoupling., yielding spectra with no nuclear Overhauser enhancement. Free induction decays ŽFIDs. were processed using the NUTS 2D software package ŽAcorn NMR, Fremont, CA. running on an IBM-type PC. The FIDs were zero-filled once, multiplied by a Lorentzian᎐ Gaussian composite function Žy0.6 Hz, maximum at 0.4= acquisition time., Fourier transformed and the resulting spectra phased and fit with Lorentzian᎐Gaussian lineshapes using a Simplex optimization algorithm. The chemical shift scales in Figs. 1 and 2 were determined relative to y0.51 ppm for 16:0r16:0-PC. Most PLs were quantified using the PL ad spectrum. However, PC p and PC e , which are resolved and assigned in PL cm , were quantified in PL cm relative to the concentration of SM determined in PL ad ŽTable 2.. Additional details concerning sample preparation and analysis can be found elsewhere ŽPearce and Komoroski, 2000..

M u ltiva ria te a n a lysis o f c o va ria n c e ŽMANCOVA., univariate analysis of covariance ŽANCOVA. and Pearson and Spearman correlations were performed using Statistica ŽStatsoft, Inc., Tulsa, OK.. Because some samples were processed with a wash step and others were not, this factor was entered as a covariate in the analyses of variance. The statistical significance of individual comparisons was confirmed by post hoc analysis using the Scheffe ´ test. 3. Results Although the schizophrenic and control groups were matched for age and sex, they were not matched for postmortem interval ŽPMI. ŽTable 1.. However, we have shown that the PL composition of rat brain, as measured by 31 P NMR, does not depend on PMI from approximately 0.25 to 18 h ŽPearce and Komoroski, 2000.. None of the present results for either group, or for all subjects combined, correlated with PMI ŽPearson correla-

176

R.A. Komoroski et al. r Psychiatry Research: Neuroimaging Section 106 (2001) 171᎐180

Fig. 2. The 31 P NMR spectra at 121.65 MHz of PLs in left frontal cortex Žgray matter. from: Žtop. a schizophrenic patient Ž噛9 in Table 1.; Žbottom. a control Ž噛5 in Table 1.. 1-Acyl-LPC a was not observed in the top spectrum.

tion, all P) 0.08, all r 2 - 0.83, d.f.s 8, not significant.. Thus, PMI should not be a factor in the present comparison of the schizophrenic and control groups. The schizophrenic patients were on a variety of antipsychotic medications at the time of death ŽTable 1.. The results reported here for the schizophrenic group did not depend on the medication used, as determined by lack of a significant Spearman rank-order correlation between any of the metabolite levels and the identity of the medication Žall P) 0.05, d.f.s 3, not significant.. In Fig. 1 are the 31 P NMR spectra of PLs in left frontal cortex from a control subject Ž噛2 in Table 1.. The bottom spectrum was acquired using the PL cm system, while the top spectrum was acquired using the PL ad system. Peak assignments are those determined previously ŽPearce and Komoroski, 2000.. Abbreviations are given in Table 2. The pattern for PC molecular species has been determined previously ŽPearce and Komoroski, 1993, 2000.. It arises from PCs with two saturated ŽdsPC., one saturated and one unsaturated Žone

or two double bonds. Žsu L PC., and two unsaturated Žone or two double bonds per chain. Ždu L PC. acyl side chains. PCs with one saturated and one highly unsaturated chain Žfour or more double bonds. Žsu H , such as arachidonoyl, 20:4., which occurs to a significant extent in the human brain, probably co-resonate with du L PC, based on our previous result for 18:0r20:4-PC ŽPearce and Komoroski, 1993.. The distribution of PC molecular species seen here is in reasonable agreement with that determined previously ŽPearce and Komoroski, 2000.. In Fig. 2 are 31 P spectra acquired using the PL ad method for samples from left frontal cortex of a schizophrenic Ž噛9 in Table 1, top. and a control Ž噛5 in Table 1, bottom.. In Table 2 are the PL concentrations Žmean " S.D.. for the two groups. Most concentrations were determined from the PL ad spectra as described previously ŽPearce and Komoroski, 2000.. Because of PL partitioning into the aqueous phase, the PL cm spectra acquired here are not suitable for quantitation of many PLs. However, PC p and PC e , which are resolved and assigned in PL cm , could

R.A. Komoroski et al. r Psychiatry Research: Neuroimaging Section 106 (2001) 171᎐180

be quantified in PL cm relative to the concentration of SM determined in PL ad ŽPearce and Komoroski, 2000.. Given the number of variables observed, a full MANCOVA analysis was not possible. MANCOVA which included the major PC species groups wdsPC, su L PC, Žsu H q du L .PCx and PI, with wash step as covariate, showed a significant group effect Ž Ps 0.009, Wilks’ lambda s 0.024.. Detailed analysis of specific interactions showed that su L PC w Ps 0.045, F Ž1,6. s 6.34x and PI w P s 0.05, F Ž1,6. s 5.97x were significantly higher in schizophrenics than in controls. In addition, there was a trend for total PL content to be higher in schizophrenics than in controls w Ps 0.085, F Ž1,6. s 4.25x. Statistical significance of the individual comparisons was confirmed by post hoc analysis using the Scheffe ´ test. By ANCOVA there were no significant differences between the two groups for PS, PE a , PE p , PE e , SM, CL, dsPC, Žsu H q du L .PC, PC p , and PC e . The last two PLs were observed using the PL cm method Žsee Fig. 1.. In some PL ad samples, resonances from LPE a and LPC a were observed. As these were not observed using the PL cm method and grew with time in the aqueous dispersion, they were attributed to a small degree of sample hydrolysis, which occurred under the PL ad conditions used. Resonances were typically not resolved for molecular species of PE a and PE p in the PL ad system in this study. With precise adjustment of conditions, some species resolution can be obtained for these cases ŽPearce and Komoroski, 2000..

4. Discussion Studies on erythrocytes and platelets suggest that alterations in membrane PL composition may play a role in the pathophysiology of schizophrenia. Early findings have been recently summarized ŽRıpova ´ ´ et al., 1997.. These findings were mixed, but generally suggested decreases in PC and PE, and increases in PS and SM in schizophrenics relative to controls. Decreases in PC, PE and PS were seen for skin fibroblasts of schizophrenics ŽMahadik et al., 1994..

177

Results are also mixed for PI in peripheral tissue. Keshavan et al. Ž1992., Rıpova ´ ´ et al. Ž1997. and Mahadik et al. Ž1994. found decreased PI in erythrocytes, platelets and fibroblasts, respectively, of schizophrenics relative to normal controls. However, Demish et al. Ž1992. found increased PI in platelets from schizophrenics. Others have found differences in platelet phosphoinositide turnover in schizophrenia ŽYao et al., 1992.. Although treatment with neuroleptics did not appear to affect PL composition ŽLautin et al., 1982., Essali et al. Ž1990. found that neuroleptic treatment produced a long-term modification of phosphoinositide turnover. In a study of the fatty-acid composition of PLs in frontal cortex, Horrobin et al. Ž1991. found reductions in the concentrations of several polyunsaturated fatty acids ŽPUFAs. in the PE fraction for schizophrenics relative to controls. No changes were seen for the fatty acids in the PC fraction. Very recently, Yao et al. Ž2000a. found significantly lower concentrations of PC and PE in schizophrenia for postmortem tissue from the caudate. Trends were also noted for a reduction in total PL content and for increases in PI and SM. In the same study, a decrease in total membrane fatty acids was noted, as were decreases in both saturated and polyunsaturated Ž n-6. fatty acids, particularly arachidonic acid. Our results support the notion that PL abnormalities occur in the brain in schizophrenia ŽHorrobin et al., 1994; Peet et al., 1999; Fenton et al., 2000., although not to the extent seen in many previous studies in vitro. Our results differ from previous work in many respects. For example, we see no significant diff erences between schizophrenics and controls for most individual PLs, including PC and PE, unlike Yao et al. Ž2000a. in the caudate. These authors also see a reduction in total PLs, in contrast to our results. However, comparisons are limited by the fact that, except for Horrobin et al. Ž1991., no previous studies were of frontal cortex. Our results for PI support the work of Demish et al. Ž1992., where elevated PI was seen in platelets for schizophrenics, and the work of Yao et al. Ž2000a. in the caudate, where a trend toward increased PI was noted. Our present results

178

R.A. Komoroski et al. r Psychiatry Research: Neuroimaging Section 106 (2001) 171᎐180

for PI are also consistent with some previous work in our laboratory. In an in vitro 1 H NMR study of metabolites in postmortem brain ŽOmori et al., 1997., we observed a trend toward elevated myo-inositol, a metabolic precursor to PI, in left frontal cortex in schizophrenia. However, in a localized, in vivo 1 H study involving medicated patients ŽHeimberg et al., 1998., myo-inositol was not significantly elevated in the same brain region for schizophrenia. The su L PC difference seen here suggests a disturbance in fatty acid composition of PC in schizophrenia ŽHorrobin et al., 1994.. The su L PC resonance arises from molecular species such as 16:0r18:1-PC, 18:0r18:1-PC, 16:0r18:2-PC, and 18:0r18:2-PC. Based on previous reports ŽHorrobin et al., 1991; Yao et al., 2000a., changes would more likely be expected in the Žsu H q du L .PC or dsPC resonances. However, changes in membrane composition in schizophrenia may involve small changes in the amounts of specific molecular species. The effects of such changes may be difficult to observe in the multicomponent 31 P-NMR resonances of PC species. Gattaz and coworkers found increased PLA 2 activity and the expected product, LPC a , in serum or platelets from drug-naive schizophrenics ŽGattaz et al., 1990, 1995; Pangerl et al., 1991.. Shortterm administration of neuroleptics appeared to reduce the increased PLA 2 activity to normal. More recently, Ross et al. Ž1997. identified the increased PLA 2 activity, seen in serum for schizophrenics treated long-term, as being associated with a calcium-independent form of the enzyme. The increased concentration of LPC a expected from increased PLA 2 activity would be expected to alter the physical properties of the cell membrane, and hence the functioning of membrane proteins and receptors ŽLundbaek and Andersen, 1994.. In the 31 P spectra we observe very low levels of LPC a and LPE a , which do not differ between neuroleptic-treated schizophrenics and controls. The concentrations are substantially lower as a fraction of total PC than those reported by Pangerl et al. Ž1991. for platelets. We attribute the small amounts of LPC a and LPE a observed here to slight hydrolysis which can occur in the

PL ad samples. Our observations provide no evidence for increased PLA 2 activity or LPC a in schizophrenic brain. It is reasonable to expect PL metabolism and concentrations to be altered by antipsychotic medications. Not surprisingly, we detected no dependence on antipsychotic medication in our very limited sample. Using the techniques applied here, this question could possibly be tested postmortem in a much larger sample of patients treated with a variety of antipsychotic medications. Recently, Fukuzako et al. Ž1999. found by in vivo 31 P NMR that haloperidol reduced elevated PDEs in the left temporal lobe in schizophrenics. NMR spectroscopy can be used to measure various metabolites in the brain in vivo in schizophrenia ŽKegeles et al., 1998.. In vivo 31 P NMR can be used to measure low molecular weight compounds involved in PL metabolism, including PME precursors, such as phosphocholine and phosphoethanolamine and PDE degradation products, such as glycerophosphocholine ŽGPC. and glycerophosphoethanolamine ŽGPE.. Pettegrew et al. Ž1991. found reduced PMEs and elevated PDEs in the dorsal prefrontal cortex of drug-naive schizophrenics. Since that report, several studies have observed lowered PMEs. Less consistently, some have observed either elevated or lowered PDEs for frontal cortex of medicated patients. Changes are sometimes observed in other brain regions. Although the findings of the various studies are not totally consistent, in general, in vivo 31 P NMR studies suggest alterations of PL metabolism in the frontal lobes in schizophrenia ŽWilliamson and Drost, 1999.. Recently, Yao et al. Ž2000b. found significant positive correlations between the concentrations of erythrocyte membrane polyunsaturated fatty acids and PMEs or PMErPDE as measured by in vivo 31 P NMR for combined left and right frontal cortex in schizophrenics. Because PLs are solid-like and have reduced molecular mobility in cell membranes, their resonances are typically very broad and not expected to be visible in in vivo NMR studies. However, a signal arising from relatively mobile PLs can be partially detected under the proper conditions Žshort delay time to acquisition. by 31 P NMR in

R.A. Komoroski et al. r Psychiatry Research: Neuroimaging Section 106 (2001) 171᎐180

vivo ŽKilby et al., 1991.. This signal, which has been attributed to PLs in vesicles andror proteolipids, has been studied in schizophrenia ŽStanley et al., 1997; Potwarka et al., 1999.. In an in vivo 31 P NMR study without 1 H decoupling, Stanley et al. Ž1997. saw reduced total PDEs, which they attributed to a reduction in the mobile-PL component, in schizophrenics relative to controls. In a 1 H-decoupled in vivo 31 P NMR study, which provides better resolution of individual components of the PME and PDE peaks, Potwarka et al. Ž1999 . saw increased mobile PLs in schizophrenia, with no change in GPC or GPE. Also, they saw reduced phosphocholine and unchanged phosphoethanolamine Žthese peaks are now resolved in the 1 H-decoupled experiment. in schizophrenia. In contrast to the above, Bluml ¨ et al. Ž1999. saw elevated GPC and GPE in parietal cortex in schizophrenia by 1 H-decoupled 31 P NMR in vivo. However, their measurement of total PDE concentration, which was substantially larger than the sum of GPC and GPE, was not different between schizophrenics and controls. This implies that the mobile-PL peak Žwhich these authors do not address explicitly. is reduced in schizophrenia. However, Bluml ¨ et al. Ž1999. acquired their spectra with an echo time of 12 ms, which is substantially longer than the 1.6᎐3.1-ms delays used in the other studies cited above. This would substantially reduce the extent to which the mobile-PL component was detected. Given this difference in signal detection and the fact that Bluml ¨ et al. Ž1999. studied parietal cortex, it is hard to compare the results of this study with those of Stanley et al. Ž1997. and Potwarka et al. Ž1999.. Our in vitro observation of a trend toward increased total PL content in frontal cortex in schizophrenia indirectly supports the findings of Potwarka et al. Ž1999. of an increased mobile-PL peak in vivo. However, additional work is necessary to sort out how observation of the mobile-PL peak in vivo depends on the details of scanning, decoupling and postprocessing conditions. Our present observations have a number of advantages. Firstly, they pertain directly to the frontal cortex and not to the peripheral tissue or

179

body fluids. Secondly, they contain information on many PLs, including PL subclasses and molecular species. Lastly, they are relatively rapid and quantitative. Disadvantages of the present study are those often associated with the use of postmortem tissue. These include: Ž1. the subjects were old; Ž2. the patients had been treated with antipsychotic medications for a long time; and Ž3. the possible presence of perimortem factors. In summary, our in vitro 31 P results support the general notion of abnormalities in PL and membrane fatty acid compositions in the frontal cortex in schizophrenia. However, our results differ in many details from those of previous authors for both brain and peripheral tissue. Further studies of brain PLs and the corresponding low molecular weight metabolites seen in vivo are warranted in a larger sample of subjects.

Acknowledgements We thank Prof. Perry Renshaw for helpful discussions. R.A. Komoroski is a Core Investigator of the VISN 16 Mid-South Mental Illness Research, Education, and Clinical Center ŽMIRECC. of the US Veterans Healthcare System. References American Psychiatric Association, 1987. Diagnostic and Statistical Manual of Mental Disorders, 3rd revised edition. Author, Washington, DC. Bluml, ¨ S., Tan, J., Harris, K., Adatia, N., Karme, A., Sproull, T., Ross, B., 1999. Quantitative proton-decoupled 31 P MRS of the schizophrenic brain in vivo. Journal of Computer Assisted Tomography 23, 272᎐275. Demish, L., Heinz, K., Gerbaldo, H., Kirsten, R., 1992. Increased concentrations of phosphatidylinositol ŽPI. and decreased esterification of arachidonic acid into phospholipids in platelets from patients with schizoaffective disorders or atypic phasic psychoses. Prostaglandins, Leukotrienes and Essential Fatty Acids 46, 47᎐52. Essali, M.A., Das, I., de Belleroche, J., Hirsch, S.R., 1990. The platelet polyphosphoinositide system in schizophrenia: the effects of neuroleptic trreatment. Biological Psychiatry 28, 475᎐487. Fenton, W.S., Hibbeln, J., Knable, M., 2000. Essential fatty acids, lipid membrane abnormalities and the diagnosis and treatment of schizophrenia. Biological Psychiatry 47, 8᎐21.

180

R.A. Komoroski et al. r Psychiatry Research: Neuroimaging Section 106 (2001) 171᎐180

Fukuzako, H., Fukuzako, T., Kodama, S., Hashiguchi, T., Takigawa, M., Fujimoto, T., 1999. Haloperidol improves membrane phospholipid abnormalities in temporal lobes of schizophrenic patients. Neuropsychopharmacology 21, 542᎐549. Gattaz, W.F., Hubner, C.v.K., Nevalainen, T.J., Thuren, T., ¨ Kinnunen, T.J., 1990. Increased serum phospholipase A 2 activity in schizophrenia: a replication study. Biological Psychiatry 28, 495᎐501. Gattaz, W.F., Schmitt, A., Maras, A., 1995. Increased platelet phospholipase A 2 activity in schizophrenia. Schizophrenia Research 16, 1᎐6. Glonek, T., 1994. 31 P NMR in the analysis of extracted tissue phospholipids. In: Quin, L.D., Verkade, J.G. ŽEds.., Phosphorus-31 NMR Spectral Properties in Compound Characterization and Structural Analysis. VCH, New York, pp. 283᎐294. Heimberg, C., Komoroski, R.A., Lawson, W.B., Cardwell, D., Karson, C.N., 1998. Regional proton magnetic resonance spectroscopy in schizophrenia and exploration of the drug effect. Psychiatry Research: Neuroimaging 83, 105᎐115. Horrobin, D.F., Manku, M.S., Hillman, H., Iain, A., Glen, M., 1991. Fatty acid levels in the brains of schizophrenics and normal controls. Biological Psychiatry 30, 795᎐805. Horrobin, D.F., Glen, A.I.M., Vaddadi, K., 1994. The membrane hypothesis of schizophrenia. Schizophrenia Research 13, 195᎐207. Karson, C.N., Casanova, M.F., Kleinman, J.E., Griffin, W.S.T., 1993. Choline acetyltransferase in schizophrenia. American Journal of Psychiatry 150, 455᎐459. Kegeles, L.S., Humaran, T.J., Mann, J.J., 1998. In vivo neurochemistry of the brain in schizophrenia as revealed by magnetic resonance spectroscopy. Biological Psychiatry 44, 382᎐398. Keshavan, M.S., Mallinger, A., Panchalingam, K.S., Pettegrew, J.W., 1992. Erythrocyte membrane phospholipid alterations in schizophrenia. Biological Psychiatry 31, 62A᎐63A. Kilby, P.M., Bolas, N.M., Radda, G.K., 1991. 31 P-NMR study of brain phospholipid structures in vivo. Biochimica et Biophysica Acta 1085, 257᎐264. Lautin, A., Cordasco, D.M., Segarnick, D.J., Wood, L., Mason, M.F., Wolkin, A., Rotrosen, J., 1982. Red cell phospholipids in schizophrenia. Life Sciences 31, 3051᎐3056. London, E., Feigenson, G.W., 1979. Phosphorus NMR analysis of phospholipids in detergents. Journal of Lipid Research 20, 408᎐412. Lundbaek, J.A., Andersen, O.S., 1994. Lysophospholipids modulate channel function by altering the mechanical properties of lipid bilayers. Journal of General Physiology 104, 645᎐673. Mahadik, S.P., Mukherjee, S., Correnti, E.E., Kelkar, H.S., Wakade, C.G., Costa, R.M., Scheffer, R., 1994. Plasma membrane phospholipid and cholesterol distribution of skin fibroblasts from drug-naive patients at the onset of psychosis. Schizophrenia Research 13, 239᎐247. Omori, M., Pearce, J., Komoroski, R.A., Griffin, S.T., Mrak, R.E., Husain, M.M., Karson, C.N., 1997. In vitro 1 H-mag-

netic resonance spectroscopy of postmortem brains with schizophrenia. Biological Psychiatry 42, 359᎐366. Pangerl, A.M., Steudle, A., Jaroni, H.W., Rufer, R., Gattaz, ¨ W.F., 1991. Increased platelet membrane lysophosphatidylcholine in schizophrenia. Biological Psychiatry 30, 837᎐840. Pearce, J.M., Komoroski, R.A., 1993. Resolution of phospholipid molecular species by 31 P NMR. Magnetic Resonance in Medicine 29, 724᎐731. Pearce, J.M., Komoroski, R.A., 2000. Analysis of phospholipid molecular species in brain by 31 P NMR spectroscopy. Magnetic Resonance in Medicine 44, 215᎐223. Peet, M., Glen, I., Horrobin, D.F., 1999. Phospholipid Spectrum Disorder in Psychiatry. Marius Press, Carnforth, UK. Pettegrew, J.W., Keshavan, M.S., Panchalingam, K., Strychor, S., Kaplan, D.B., Tretta, M.G., Allen, M., 1991. Alterations in brain high-energy phosphate and membrane phospholipid metabolism in first-episode, drug-naive schizophrenics. A pilot study of the dorsal prefrontal cortex by in vivo phosphorus 31 nuclear magnetic resonance spectroscopy. Archives of General Psychiatry 48, 563᎐568. Potwarka, J.J., Drost, D.J., Wiliamson, P.C., Carr, T., Canaran, G., Rylett, W.J., Neufeld, W.J., 1999. A 1 H-decoupled 31 P chemical shift imaging study of medicated schizophrenic patients and healthy controls. Biological Psychiatry 45, 687᎐693. Rıpova, ´ ´ D., Strunecka, ´ A., Nemcova, ´ V., Farska, ´ I., 1997. Phospholipids and calcium alterations in platelets of schizophrenic patients. Physiological Research 46, 59᎐68. Ross, B.M., Hudson, C., Erlich, J., Warsh, J.J., Kish, S.J., 1997. Increased phospholipid breakdown in schizophrenia: evidence for the involvement of a calcium-independent phospholipase A 2 . Archives of General Psychiatry 54, 487᎐494. Salzman, S., Endicott, J., Clayton, P., Winokur, G. ŽEds.., 1983. Diagnostic Evaluation After Death ŽDEAD.. NIMH Neurosciences Research Branch, Rockville, MD. Stanley, J.A., Panchalingam, K., Miller, G., McClure, R.J., Pettegrew, J.W., 1997. A new method to quantify the broad component under the phosphodiester resonance and its application to study first-episode never medicated schizophrenics. Proceedings of the International Society of Magnetic Resonance in Medicine 5, 1408. Williamson, P.C., Drost, D.J., 1999. 31 P magnetic resonance spectroscopy in the assessment of brain phospholipid metabolism in schizophrenia. In: Peet, M., Glen, I., Horrobin, D.F. ŽEds.., Phospholipid Spectrum Disorder in Psychiatry. Marius Press, Carnforth, UK, pp. 45᎐55. Yao, J.K., Yasaei, P., van Kammen, D.P., 1992. Increased turnover of platelet phosphatidylinositol in schizophrenia. Prostaglandins, Leukotrienes and Essential Fatty Acids 46, 39᎐46. Yao, J.K., Leonard, S., Reddy, R.D., 2000a. Membrane phospholipid abnormalities in postmortem brains from schizophrenic patients. Schizophrenia Research 42, 7᎐17. Yao, J.K., Stanley, J.A., Reddy, R.D., Keshavan, M.S., Pettegrew, J.W., 2000b. Correlations between RBC fatty acids and 31 P MRS brain measures in schizophrenia. Biological Psychiatry 47, 41S.