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

It has been hypothesized that schizophrenia arises from cell membrane abnormalities due to changes in phospholipid (PL) composition and metabolism. We have used high resolution, in vitro 31P nuclear magnetic resonance (NMR) to characterize the PLs in left frontal cortex (gray matter) of postmortem brain from four schizophrenics and five controls. High resolution 31P NMR spectra were obtained in an organic-solvent system to resolve PL classes (headgroups) and in a sodium-cholate, aqueous dispersion system to resolve phosphatidylcholine (PC) molecular species. Multivariate analysis which included the major PC molecular species and phosphatidylinositol (PI) 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 (one or two 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.

Analysis of phospholipid molecular species in brain by (31)P NMR spectroscopy.

Techniques are described for the (31)P NMR analysis of glycerophospholipid (PL) headgroup and molecular species in brain. The (31)P NMR spectrum of PLs from human temporal cortex, solubilized in aqueous Na cholate, typically showed 3 major resonances, assigned to phosphatidylcholine (PC) molecular species containing 0, 1, or 2 fully saturated acyl chains. Less species resolution was obtained for the other PL headgroups under these conditions. Alkylacyl- and alkenylacyl-PC were readily discerned using the CHCl(3)-CH(3)OH-H(2)O solvent method. The chain-length, temperature, and species dependences of the (31)P NMR chemical shifts were explored in model PLs. Assignments of signals from phosphatidylethanolamine (PE) subclasses were confirmed in the sodium-cholate system by lipase-mediated selective hydrolysis of bovine-brain PE. The utility of (31)P NMR to monitor enzymatic PL oxidation was further demonstrated. Possible changes in PL composition with postmortem interval (PMI) in rat brain were examined. No significant changes were seen in PL headgroup or PC species composition with PMI at up to 18 hours. Where comparable, the Na-cholate-solubilization and solvent methods gave similar quantitative results for headgroup analysis on the same samples. The present work demonstrates the feasibility and utility of the dual system for analysis of PLs in brain. Magn Reson Med 44:215-223, 2000.

Contribution of trifluoperazine metabolites to the in vivo (19)F NMR spectrum of rat brain.

Fluorine-19 NMR spectra were acquired from extracts of tissues from heads of rats given the antipsychotic drug trifluoperazine (TFP). Contributions to the in vivo (19)F spectra from tissues other than brain were negligible. The in vivo (19)F resonance at -62.3 ppm from CCl(3)F consisted of 6-8 resolved resonances in vitro. Some in vitro resonances were assigned to previously identified TFP metabolites. Multiple resonances in vitro partially explain the relatively large line width seen in vivo for TFP. Unidentified metabolites were observed at about -74 to -75 ppm in a number of spectra of extracts of brain and muscle.

The use of nuclear magnetic resonance spectroscopy in the detection of drug intoxication.

The use of nuclear magnetic resonance (NMR) spectroscopy as a method for drug analysis has the advantages of reduced pre-analytical preparation time and the potential to detect and quantitate drug conjugates and metabolites simultaneously. NMR was investigated as a method to screen for organic substances (and metabolites) in 25 patients who presented to the Emergency Department with clinical indications of a drug overdose. Urine specimens were examined by 1H NMR spectroscopy at 300 MHz and the results compared with gas chromatography-mass spectrometry (GC-MS) results. There was a 56% concordance (14 of 25 samples) between NMR and GC-MS. NMR identified acetaminophen, ibuprofen, aspirin, valproate, carbamazepine, and pseudoephedrine as parent compounds or metabolites. For a patient for whom GC-MS results were negative, NMR strongly suggested the presence of erythromycin. NMR was most successful in identifying analgesics and antiepileptic drugs (sensitivity 83-100%). In 10 patients, signals from 1,2-propanediol, a common vehicle for some pediatric medications, were observed by NMR spectroscopy. NMR had 0% sensitivity in identifying tricyclic antidepressants and antipsychotic drugs. In these samples, GC-MS detected a variety of compounds, including tricyclic antidepressants and their metabolites and chlorpromazine. In addition, other substances that had not been disclosed as having been ingested, such as caffeine, diphenhydramine, and nicotine, were detected by GC-MS. NMR spectroscopy represents an emerging supplementary analytical technique that is applicable to a wide range of possible intoxicants and to the evaluation of the intoxicated patient, particularly when larger amounts of the intoxicant (> 200 mg) are ingested.

Effects of antipsychotic drugs on metabolite ratios in rat brain in vivo.

Localized in vivo proton magnetic resonance spectroscopy at 4.7 T was used to examine the brains of rats that were given the antipsychotic drugs haloperidol, clozapine, or olanzapine for 1 week. Spectra were collected before and during treatment. The ratios of N-acetylaspartate (NAA) to creatine (Cr) and choline to Cr were determined from the spectra. No significant differences in these ratios were seen among the rats given the various antipsychotic medications or between the control rats and the treated rats. No significant time-dependent changes were seen in most cases, except for a small reduction of NAA/Cr after 7 days of olanzapine administration. These results suggest that differences in brain metabolite ratios in vivo in schizophrenics relative to controls, at least for short-term treatment, arise from the disease, and not as a metabolic effect of the medication.

Applications of (7)Li NMR in biomedicine.

The applications of (7)Li NMR spectroscopy and imaging in biology and experimental medicine have been progressing steadily. The interest derives primarily from the clinical use of Li salts to treat mania and manic-depressive illness. One area of investigation is ionic transport across the cellular membrane and compartmentation, so as to elucidate the mechanism(s) of therapeutic action and toxicity in clinical practice. The second is the development of a noninvasive, in vivo analytical tool to measure brain Li concentrations in humans, both as an adjunct to treatment and as a mechanistic probe. Here we review progress to date in this area.

Effects of gender and region on proton MRS of normal human brain.

Localized, in vivo 1H magnetic resonance spectroscopy has been performed in a number of brain regions of neuropsychiatric interest in male and female control subjects to determine if gender and region affect the measured metabolite ratios. In contrast to some previous reports, no significant differences were seen in any region for any metabolite ratio between males and females. As expected, significant variations with brain region were seen for metabolite ratios for the total group of subjects.

Regional proton magnetic resonance spectroscopy in schizophrenia and exploration of drug effect.

Schizophrenia is a disorder with an unclear pathophysiology, despite numerous attempts to elucidate its etiology. We have employed proton magnetic resonance spectroscopy in vivo to explore the neurochemistry of several brain regions (left frontal and temporal cortices, left basal ganglia, and left and right thalamus) in patients with schizophrenia and in normal control subjects. We have also examined patients in different medication states. A trend toward a decreased level of inositol/creatine was found in the left temporal lobe of patients with schizophrenia, as was a trend toward a reduced level of N-acetylaspartate/creatine in the left thalamus of patients. In schizophrenic patients treated with atypical antipsychotics, decreased levels of choline were found in the left basal ganglia, while increased levels of N-acetylaspartate were found in the left frontal cortex. These results suggest altered metabolism in patients with schizophrenia, and imply that further study is needed to clarify the effects of the more recently available antipsychotics.

Lithium visibility in rat brain and muscle in vivo by 7Li NMR imaging.

The apparent concentration of lithium (Li) in vivo was determined for several regions in the brain and muscle of rats by 7Li NMR imaging at 4.7 T with inclusion of an external standard of known concentration and visibility. The average apparent concentrations were 10.1 mM for muscle, and 4.2-5.3 mM for various brain regions under the dosing conditions used. The results were compared to concentrations determined in vitro by high-resolution 7Li NMR spectroscopy of extracts of brain and muscle tissue from the same rats. The comparison provided estimates of the 7Li NMR visibility of the Li cation in each tissue region. Although there was considerable scatter of the calculated visibilities among the five rats studied, the results suggested essentially full visibility (96%) for Li in muscle, and somewhat reduced visibility (74-93%) in the various brain regions.

In vivo proton magnetic resonance spectroscopy of the medial temporal lobes of subjects with combat-related posttraumatic stress disorder.

Recent findings using volumetric MRI techniques have revealed that patients with combat-related and noncombat-related posttraumatic stress disorder (PTSD) have reductions in right hippocampal volume. Twenty-one veterans with PTSD and eight age-matched control veterans were studied using proton magnetic resonance spectroscopy to test the hypothesis that the N-acetyl-L-aspartic acid/creatine (NAA/Cr) ratio would be decreased in the right medial temporal lobe structures of patients with PTSD compared to controls. Patients with PTSD displayed significantly lower NAA/Cr ratio for the right medial temporal lobe relative to the left (P < or = 0.011). Patients with PTSD also had lower NAA/Cr in right medial temporal lobe (P < or = 0.013) and lower choline/Cr in left medial temporal lobe (P < or = 0.030) compared to control subjects. Because NAA is regarded as an indicator of neuronal density, this finding suggests that the neuronal density of right-sided medial temporal structures in patients with combat-related PTSD may be decreased.

In vitro 1H-magnetic resonance spectroscopy of postmortem brains with schizophrenia.

Some evidence suggests that thalamic dysfunction could explain some of the signs and symptoms of schizophrenia. We measured the absolute concentrations of amino acid metabolites in thalamus, frontal pole, and cerebellar vermis in extracts of postmortem brains from 8 schizophrenics and 10 controls using high-resolution 1H-magnetic resonance spectroscopy. The concentrations of N-acetyl aspartate, glutamate, and valine tended to be reduced in the thalamus of the schizophrenic group. Although it is difficult to ascribe significance to the "tendencies," these data may tend to support other data suggesting decreased thalamic volume or neuronal number in schizophrenia.

The distribution of lithium in rat brain and muscle in vivo by 7Li NMR imaging.

The regional distribution of lithium (Li) in vivo was determined on a relative basis in the brain and muscle of rats by 7Li NMR imaging. Both high resolution 7Li NMR spectroscopy and atomic absorption spectrophotometry were performed on extracts of brain and muscle tissue from the same rats. The average in vivo results were generally in good agreement with the corresponding average in vitro results, despite the relatively low signal-to-noise ratio of the in vivo images. There was good agreement between the two in vitro methods of analysis, both on average and for individual animals. Significant differences were found in vivo among Li concentration ratios for the various brain regions and muscle by 7Li imaging, as well as by the in vitro methods.

Regional metabolic alterations in Alzheimer's disease: an in vitro 1H NMR study of the hippocampus and cerebellum.

The concentrations of selected metabolites in the hippocampus and cerebellum of 13 Alzheimer's diseased (AD) and four nondemented postmortem brains were measured using high resolution 1H NMR spectroscopy. For both the hippocampal region and the cerebellum, the putative neuronal marker N-acetyl aspartate (NAA) was significantly lower in AD brains relative to the nondemented brains. For the hippocampal region, the NAA concentration correlated inversely with semiquantitative assessments of neuronal loss and neurofibrillary tangles. The gamma-aminobutyric acid levels in both hippocampus and cerebellum of an age- and a postmortem interval-matched subset of AD brains were lower than those of the controls. Because the cerebellum is generally thought to be unaffected by AD, the NAA decrease in the Alzheimer cerebellum may be due to lesions of either the Alzheimer or non-Alzheimer type in contralateral cerebrum.

An in vitro 1H nuclear magnetic resonance study of the temporoparietal cortex of Alzheimer brains.

The concentrations of selected metabolites in the posterior temporoparietal cortex of 13 Alzheimer's diseased (AD) and four nondemented postmortem brains (of individuals between the ages of 63 and 95) were determined using high-resolution 1H nuclear magnetic resonance (NMR) spectroscopy. The estimates for glutamate and inositol for AD brains did not show any statistically significant difference (P > 0.05) from those for the nondemented brains. The putative neuronal marker N-acetyl aspartate (NAA), creatine, and GABA were decreased in AD brains compared with the nondemented brains. The estimates for creatine, glutamate, and GABA showed significant linear correlations with those of NAA. Creatine, glutamate, GABA, and NAA appeared to be negatively correlated with the neurofibrillary tangles. Our results support a neuronal loss in the posterior temporoparietal cortices of AD brains.

n-Amyl alcohol partitioning in synaptic plasma membranes.

Nuclear magnetic resonance and fluorescence polarization techniques were used to determine n-amyl alcohol partitioning between, and effects on, lipid microdomains of isolated rat cerebral synaptic plasma membranes. n-Amyl alcohol binding to the hydrophobic membrane core had an unchanging binding constant over an aqueous alcohol concentration range of 2.5-22.5 mM, indicating a linear relationship between membrane core and aqueous alcohol concentrations. Binding to the membrane surface, in contrast, was cooperative with a steadily increasing binding constant over this alcohol concentration range. Membrane lipid order was determined using various fluorescent probes with preferences for the membrane core, for the mid-acyl regions of the exofacial or cytofacial bilayer leaflets and for ordered or bulk microdomains. All these probes showed steady decreases in membrane order with increasing alcohol concentration, at least for the nanosecond time scale sampled by this technique. These results further demonstrate the complexity of interaction between natural membranes and membrane disordering agents.

In vivo 19F spin relaxation and localized spectroscopy of fluoxetine in human brain.

Fluorine-19 NMR spectroscopy was used to monitor the anti-depressant drug fluoxetine (and its metabolite norfluoxetine) in vivo in human brain. A quadrature birdcage head coil, developed for operation at 60.1 MHz, yielded a signal from the head 2 to 4 times stronger than for surface coils. It was used to measure the in vivo 19F spin-lattice relaxation time (T1) of fluoxetine for five patients by the inversion-recovery technique. The individual T1s varied from 149 to 386 ms, which was attributed in part to interindividual differences based on the reproducibility of a phantom T1. The individual T1 correlated weakly with approximate brain concentration. A lower limit of 3 to 4 ms was found for the spin-spin relaxation time from line width measurements. Low resolution 4-dimensional spectroscopic imaging confirmed that the single in vivo 19F resonance for fluoxetine arose primarily from brain. The spectrum of a cerebral hemisphere (in formalin) obtained at autopsy from a patient on 40 mg/day of fluoxetine for 19 weeks was comparable with that seen for patients in vivo. The in vivo signal arose about equally from fluoxetine and the active metabolite norfluoxetine, as demonstrated by the in vitro 19F NMR spectrum of the lipophilic extract of a small section of brain. In vitro quantitation of frozen samples from three brain regions yielded combined fluoxetine/norfluoxetine concentrations of 12.3 to 18.6 micrograms/ml, which is higher than typically determined in vivo, and suggests that the fluorinated drugs may not be 100% visible in vivo.

Analysis of saturated phosphatidylcholine in amniotic fluid by 31P NMR.

A technique using solubilization with sodium cholate to resolve 31P NMR resonances of phospholipid molecular species was applied to amniotic fluid samples from 16 subjects. Gestational ages from 25 to 40 weeks were represented, and two subjects were sampled sequentially. Fitting of the partially resolved 31P NMR signal of phosphatidylcholine (PC) generated an estimate of percent disaturated acyl PC (%dsPC) which correlated more highly with gestational age than did several other potential indices of fetal lung maturation, such as the ratio of PC to inorganic phosphate from the solubilized spectra, or PC to sphingomyelin from extract spectra. In a few cases, enzyme-catalyzed PC hydrolysis limited the precision, but did not appear to affect the accuracy, of the %dsPC estimates. Resolution of palmitoyl and oleoyl lyso-PC species was observed for both the 1- and 2-acyl isomers. Upfield shifts due to the presence of cis double bonds in the lone acyl chain of the lyso-PCs were analogous to those observed for the diacyl PCs.

In vivo 7Li nuclear magnetic resonance study of lithium pharmacokinetics and chemical shift imaging in psychiatric patients.

New data are presented on the application of 7Li in vivo nuclear magnetic resonance (NMR) spectroscopy to human studies. The technique was used to monitor the between-dose pharmacokinetics of lithium (Li) in brain for three patients on Li therapy. Brain Li concentrations were at their highest from 0 to 2 hours after the peak occurred in serum concentration. Elimination from brain tissue took longer than elimination from muscle, and no signal could be detected from brain at 10 days after termination of therapy. A birdcage radiofrequency coil for 7Li was constructed and used to measure the 7Li spin-lattice relaxation time of 4.6 seconds in vivo in human head, and to acquire preliminary spectroscopic images of a phantom and human brain.