Incorporation of [3H]valproic acid into lipids in GT1-7 neurons

Article abstract of DOI:10.1016/S0006-2952(98)00148-8

Valproic acid (2-propylpentanoic acid, valproate, VPA), an 8-carbon, branched chain fatty acid, is effectively used in the treatment of mania and epilepsy. The biochemical mechanisms by which this drug has its therapeutic effects are not yet established. The purpose of this study was to partially characterize the incorporation of [³H]VPA into phospholipids of GT1-7 neurons, an immortalized hypothalamic cell line. GT1-7 neurons were grown to confluence in culture dishes, and then were incubated with various concentrations of [³H]VPA between 10 and 400 μg/mL for various times up to 20 hr. Total lipids were extracted and phospholipids were separated from neutral lipids using TLC. Our results indicate that [³H]VPA (10 μg/mL) was incorporated into phospholipids of GT1-7 neurons in a time-dependent and saturable manner over 300 min. Subsequent separation of the lipid fraction by TLC indicated that 44.4% of the radioactivity taken up by the cells was incorporated into phospholipids and neutral lipids. One of the phospholipids migrated with a slightly lower R(f) value than authentic phosphatidylcholine. Our results show that the incorporation of VPA into phospholipids and glycerides was linear with VPA concentrations from 10 to 400 μg/mL. Finally, we synthesized 1-acyl-2-valproyl-sn-glycero-3-phosphocholine and validated its structure with nuclear magnetic resonance and electrospray mass spectrometry to verify the structure of this compound, confirming that this compound is structurally possible. We conclude that VPA is incorporated into lipids in GT1-7 neurons and discuss the possible effects of valproyl phospholipids on neuronal functional properties. Copyright (C) 1998 Elsevier Science, Inc.

Full text of DOI:10.1016/S0006-2952(98)00148-8

Biochemical Pharmacology, Vol. 56, pp. 207–212, 1998.
1998 Elsevier Science Inc. All rights reserved.
ISSN 0006-2952/98/$19.00 ϩ 0.00
©
PII S0006-2952(98)00148-8
3
Incorporation of [ H]Valproic Acid into Lipids in
GT1–7 Neurons
Athanassia Siafaka-Kapadai,† Marinis Patiris,† Charles Bowden* and Martin Javors*‡
*
DEPARTMENTS OF PSYCHIATRY AND PHARMACOLOGY, THE UNIVERSITY OF TEXAS HEALTH SCIENCE CENTER, SAN
ANTONIO, TX 78284, U.S.A.; AND †DEPARTMENT OF CHEMISTRY, UNIVERSITY OF ATHENS, ATHENS, GREECE
ABSTRACT. Valproic acid (2-propylpentanoic acid, valproate, VPA), an 8-carbon, branched chain fatty acid,
is effectively used in the treatment of mania and epilepsy. The biochemical mechanisms by which this drug has
its therapeutic effects are not yet established. The purpose of this study was to partially characterize the
3
incorporation of [ H]VPA into phospholipids of GT1–7 neurons, an immortalized hypothalamic cell line.
GT1–7 neurons were grown to confluence in culture dishes, and then were incubated with various
3
concentrations of [ H]VPA between 10 and 400 ␮g/mL for various times up to 20 hr. Total lipids were extracted
3
and phospholipids were separated from neutral lipids using TLC. Our results indicate that [ H]VPA (10 ␮g/mL)
was incorporated into phospholipids of GT1–7 neurons in a time-dependent and saturable manner over 300 min.
Subsequent separation of the lipid fraction by TLC indicated that 44.4% of the radioactivity taken up by the
cells was incorporated into phospholipids and neutral lipids. One of the phospholipids migrated with a slightly
lower R value than authentic phosphatidylcholine. Our results show that the incorporation of VPA into
f
phospholipids and glycerides was linear with VPA concentrations from 10 to 400 ␮g/mL. Finally, we synthesized
1
-acyl-2-valproyl-sn-glycero-3-phosphocholine and validated its structure with nuclear magnetic resonance and
electrospray mass spectrometry to verify the structure of this compound, confirming that this compound is
structurally possible. We conclude that VPA is incorporated into lipids in GT1–7 neurons and discuss the
possible effects of valproyl phospholipids on neuronal functional properties. BIOCHEM PHARMACOL 56;2:
2
07–212, 1998. © 1998 Elsevier Science Inc.
KEY WORDS. valproic acid; phospholipid; lipid; neuron; hypothalamus; GT1–7; epilepsy; mania
VPA§ is a branched, short-chain fatty acid with a wide
spectrum of activities against different types of seizures and
is the drug of choice for treatment of absence epileptic
seizures [1]. Recently, Bowden et al. [2] reported that VPA
was significantly more effective than a placebo in reducing
the symptoms of acute mania and that the efficacy of the
divalproex form of valproate appeared to be independent of
prior responsiveness to lithium.
The pharmacodynamic and pharmacotherapeutic effects
of VPA and the possible mechanisms of action have been
studied extensively, but the pharmacodynamic basis of its
anticonvulsant and antimanic action remains unknown [1].
It has been suggested that VPA acts through a combination
of several mechanisms. VPA increases GABA turnover,
reduces the release of the epileptogenic amino acid ␥-hy-
droxybutyrate, and exerts a direct action on ion channels
[3]. VPA is known to affect GABA and norepinephrine
metabolism through the inhibition of GABA transaminase
and aldehyde reductase [4], respectively. Chen et al. [5]
reported that chronic sodium valproate selectively de-
creases the ␣ and ε, but not the ␦ and ␨ isoforms of protein
kinase C in C6 rat glioma cells. Petty et al. [6] reported that
plasma GABA predicts acute response to valproate treat-
ment in mania. Also, VPA inhibits the synthesis of major
phospholipids [7], decreases the fluidity of brain mitochon-
drial membranes [8], interferes with embryonic lipid syn-
thesis [9], may affect transport of phospholipid precursors
across the cell membrane [10], and inhibits the secretion of
triacylglycerols from rat hepatocytes [11].
The results reported here were obtained using GT1–7
neurons, an immortalized neuronal cell line. These adult
hypothalamic neurons, developed from tumors in mice
[12–14], synthesize and secrete gonadotropin releasing hor-
mone (GnRH) and express several receptor transducing
‡
Corresponding author: Martin A. Javors, Ph.D., Department of Psychi-
systems, including GABA receptors [15–20].
A
atry, The University of Texas Health Science Center, 7703 Floyd Curl
Drive, San Antonio, TX 78284. Tel. (210) 567-5396; FAX (210)
The purpose of this study was to determine whether VPA
is incorporated into membrane phospholipids in a neuronal
cell line, to determine the time frame and amount of any
incorporation, and to identify specific phospholipids into
which VPA is incorporated. A long-term goal of this
5
67-3759; E-mail: javors@uthscsa.edu.
Abbreviations: GABA, ␥-aminobutyric acid; PAF, platelet-activating
factor; PC-VPA, 1-oleoyl-2-valproyl-sn-glycero-3-phosphocholine; TNS,
§
6
-(p-toluidino)-2-naphthalenesulfonic acid; and VPA, valproic acid.
Received 24 November 1997; accepted 24 March 1998.

2
08
A. Siafaka-Kapadai et al.
research is to test the hypothesis that valproyl-phospholip-
ids modify neuronal activity and that this modification
plays a role in the antiepileptic and antimanic-depressive
effects of the drug.
in vials with 10 mL of Beckman Ready Proteinϩ scintilla-
tion fluid; then radioactivity was quantitated using a liquid
scintillation counter (Beckman LS7500) [23].
Synthesis and Structure Verification
of Valproylphosphatidylcholine
MATERIALS AND METHODS
Culture dishes were obtained from Fisher Scientific. Cul-
ture media were products of Life Technologies. All other
VPA was converted to valproyl chloride by treatment
overnight at room temperature with 0.35 mL of freshly
distilled thionyl chloride in a calcium chloride and NaOH
desiccator. Unreacted thionyl chloride was evaporated at
80° under nitrogen. A 10-␮L aliquot of the resultant
valproyl chloride was added to a mixture of 1-oleoyl-2-lyso-
glycerophosphocholine (5 mg) and N,NЈ-dimethylamin-
opyridine (17.5 mg) in dry chloroform (0.1 mL) [24]. The
mixture was stirred at room temperature for 1.5 hr, and the
PC-VPA formed was purified using preparative TLC with
three different solvent systems: a basic solvent system
containing chloroform:methanol:28% ammonium hydrox-
ide (65:35:8, by vol.), a chloroform:acetone (9:5, v/v)
system, and a neutral solvent system containing chloro-
form:methanol:water (65:35:7, by vol.).
reagents were purchased from the Sigma Chemical Co.
3
[
H]VPA, sodium salt in 80% ethanol, 1 mCi/mL, 15
Ci/mmol, was purchased from ARC. TLC plates were
purchased from Whatman.
Cell Culture of GT1–7 Neurons
GT1–7 neurons were plated at a density of approximately
6
5
ϫ 10 cells/dish in Dulbecco’s Modified Eagle’s Medium
(DMEM, Life Technologies No. 12800, glucose 4.5 g/L)
supplemented with 5% horse serum, 5% fetal bovine serum,
penicillin (50 IU/mL), and streptomycin (50 ␮g/mL). Cells
were grown to confluence in 100-mm-diameter culture
dishes at 37° in an atmosphere of 95% air/5% CO for
approximately 5 days.
The chromatographic profile of the synthesized com-
pound was assessed using the three solvent systems de-
scribed above and an acidic solvent system that contained
chloroform:methanol:acetic acid:water (65:35:8:4, by vol.).
The structure of PC-VPA was verified with electrospray
mass spectrometry using a Finnigan Quadruple MAT
SSQ700 with an Analytica Bradford electrospray system at
a voltage of Ϫ3500 V for positively charged molecules. The
purified compound was solubilized in chloroform:methanol
(4:1, v/v) at a concentration of 1.5 mg/mL. An aliquot was
diluted 20-fold in the same solvent and a few microliters
injected into the spectrometer. The flow rate was 2 ␮L/min.
NMR was performed using a Varian NMR Unity Plus, 300
MHz.
2
3
Incorporation of [ H]VPA into Lipids of GT1–7
Neurons and Extraction of Lipids
GT1–7 neurons were incubated with various concentra-
3
tions of [ H]VPA (1 ␮Ci/mL) for various times (see legends
to figures for details of specific experiments). The medium
was removed and cells were washed twice with 5 mL of
ice-cold normal saline. Next, 3 mL of methanol and 1.2 mL
of water were added and the cells were scraped from the
plates. This procedure was then repeated to ensure the
removal of all the cells. The scraped fractions were trans-
ferred to glass tubes containing 3 mL of chloroform. After
vortexing, 3 mL of chloroform and 3 mL of water were
added to the mixture to achieve phase separation [21]. The
lower phase, containing total lipids, was transferred to
another tube and evaporated under nitrogen to dryness.
Lipids were dissolved in chloroform:methanol (4:1).
RESULTS
3
GT1–7 neurons were incubated with 10 ␮g/mL of [ H]VPA
(1 ␮Ci/mL) for 5, 15, 30, 90, and 300 min. Total lipids were
extracted and subjected to TLC using a solvent system
(
chloroform:acetone, 9:5) in which total phospholipids
remained at the origin. VPA, R Х 0.5, and neutral lipids
f
Lipid Fractionation by TLC and Measurement
of Radioactivity
(
mono-, di-, and triglycerides) migrated into the plates and
were separated as shown in the representation of a TLC
plate in Fig. 1A. We performed a radioactivity scan of lane
f of the TLC plate, which contained the extract from the
GT1–7 neurons that had been incubated with 10 ␮g
Total lipids were subjected to TLC (K6 Whatman Silica
Gel Plates) using two solvent systems: 1) chloroform:
acetone (9:5) for the separation of total phospholipids from
various neutral lipids and VPA; and 2) chloroform:meth-
anol:water (65:35:7) for the separation of individual phos-
pholipids. Lipids were visualized on the plates using TNS
spray [22]. Phospholipids were located with ammonium
molybdenum spray. Radioactivity in each lane was mea-
sured using a thin-layer scanner coupled with a data system
3
3
[ H]VPA/mL for 300 min. [ H]VPA was incorporated into
phospholipids as well as into neutral lipids (Fig. 1B).
The TLC plates were next exposed to a second solvent
system (chloroform:methanol:water, 65:35:7) for separation
of individual major phospholipid classes. Neutral lipids
migrated near the solvent front with the second solvent
system. Our results indicate that radioactivity was present
(BioScan System 200 Imaging Scanner, BioScan). Addi-
tionally, fractions were scraped from the plates, and placed
at R positions that were consistent with migration of
f

Incorporation of VPA into Lipids in GT1–7 Neurons
209
3
FIG. 1. Incorporation of [ H]VPA into
lipids in GT1–7 neurons. GT1–7 neu-
rons were grown in culture dishes, and
then were incubated for 300 min with 10
3
␮
g/mL of [ H]VPA as described in Ma-
terials and Methods. Total lipids were
extracted from GT1–7 neurons with the
Bligh–Dyer procedure [21] and then were
subjected to TLC using a chloroform:
acetone (9:5, v/v) solvent system that
separated phospholipids from VPA and
other neutral lipids. Panel A is a repre-
sentation of the TLC plate for the 300-
min incubation. All lipids were located by
staining with TNS (open and solid spots).
Phospholipids were located by staining with
ammonium molybdenum (solid spots). Key:
(
a) triglycerides, (b) diglycerides, (c) VPA,
(
d) monoglycerides, (e) phosphatidylcho-
line, and (f) GT1–7 neuronal extract. Panel
B is a radioactivity scan of lane f of the TLC
plate for the 300-min incubation with
VPA.
phosphatidylcholine, phosphatidylethanolamine, sphingo-
lipids, and possibly phosphatidylinositols (Fig. 2B). The
migration of phospholipids in the control and sample lanes
was visualized with ammonium molybdenum, and the
presence of radioactivity was measured to indicate the
lipids that were extracted from the GT1–7 neurons into the
organic fraction of the Bligh-Dyer procedure [21]. The
incorporation of radioactivity, presumably associated with
3
[ H]VPA, into phospholipids of GT1–7 neurons proceeded
in a time-dependent fashion over 20 hr (Fig. 3). The
incorporation was saturated by about 5 hr when the cells
3
presence of incorporated [ H]VPA. A scan of radioactivity
3
3
for a control sample of [ H]VPA is shown in Fig. 2C.
were incubated with 10 ␮g/mL of [ H]VPA. Approximately
Figure 3 describes the time frame of the incorporation of
radioactivity into phospholipids as the percentage of total
50% of the incorporated radioactivity was associated with
the phospholipid fraction. The results in Fig. 2 suggest that
3
FIG. 2. Incorporation of [ H]VPA into phospholipids in GT1–7 neurons. The TLC plate shown in Fig. 1 was developed further using
a second solvent system, chloroform:methanol:water (65:35:7, by vol.), which was designed to separate individual phospholipids. All
lipids were located by staining with TNS (open and solid spots). Phospholipids were located by staining with ammonium molybdenum
(
solid spots). Panel A is a representation of the TLC plate. The lanes contained: (a) phosphatidylcholine, (b) sphingomyelin,
phosphatidylethanolamine, VPA (in order from origin to solvent front), (c) GT1–7 neuronal extract, and (d) lysophosphatidylcholine,
3
VPA. Panel B is a radioactivity scan of lane c of the TLC plate. Panel C is a radioactivity scan of lane d in which [ H]VPA was tested.

2
10
A. Siafaka-Kapadai et al.
3
FIG. 3. Time course of incorporation of [ H]VPA
into total phospholipids in GT1–7 neurons. This
experiment was performed similarly to the experi-
ments described in the legends to Figs. 1 and 2. The
cells were incubated for 5, 15, 30, 90, 300, and
3
1
200 min with 10 ␮g/mL of [ H]VPA, and then
were washed to remove extracellular radioactivity.
Lipids were extracted from the cells and analyzed
by TLC as described in Materials and Methods.
3
The incorporation of [ H]VPA into phospholipids
is expressed as a percentage of the radioactivity
associated with the total lipid fraction, which was
approximately 15,000–20,000 dpm at each time
point. Each symbol represents the mean and SEM
for two experiments in duplicate.
3
[
H]VPA was incorporated into phosphatidylethanolamine,
trospray mass spectrometry (data not shown). As expected,
the electrospray-MS spectra indicated that the phosphati-
dylcholine with a valproyl residue gave intense signals of
phosphatidylcholine, and sphingomyelin. The remainder of
the radioactivity in the organic fraction of the Bligh-Dyer
procedure was probably free VPA and neutral lipids (Fig. 2,
lane b, near the solvent front).
ϩ
ϩ
[M ϩ H] at m/z 648.3 and of [M Ϫ 144.2 ϩ H] at m/z
504.1. The molecular weight of VPA is 144.2. Two char-
acteristic fragment ions of choline phospholipids were
The data in Fig. 4 indicate that the incorporation of
H]VPA into neutral lipids and phospholipids in GT1–7
3
ϩ
[
observed at m/z 184 ([phosphocholine] ) and 104.1 ([cho-
ϩ
neurons increased in a linear fashion with concentrations of
VPA up to 400 ␮g/mL. In this experiment, the cells were
incubated with VPA for 180 min. This linear relationship
was observed for total uptake of VPA into GT1–7 neurons
and for incorporation into phospholipids as well as di- and
triacylglycerols. The fraction labeled di- and triacylglycer-
ols could also include small amounts of other neutral lipids.
A comparison of the slopes of the linear regressions indi-
line] ). The NMR results confirmed those obtained with
electrospray MS (data not shown).
We synthesized PC-VPA both to establish that this
compound was stereospecifically possible and to aid in the
3
identification of compounds into which [ H]VPA was
incorporated. Our results show that the Rf value for
synthesized PC-VPA was identical to one of the phospho-
lipids in the GT1–7 extract into which the radioactivity
was incorporated. This result suggests that PC-VPA was
synthesized by GT1–7 neurons during our experiments.
3
cates that 20.6 and 23.8% of the [ H]VPA taken up by the
cells was incorporated into phospholipid and neutral lipid
(di- and triglycerols) fractions, respectively.
We synthesized PC-VPA from lysophosphatidylcholine
DISCUSSION
and valproyl chloride in the presence of dimethylaminopy-
ridine and dry chloroform. VPA and thionyl chloride were
used for the synthesis of valproyl chloride, which was then
purified using TLC with three different solvent systems.
The structure of PC-VPA was proven by NMR and elec-
Our results show that VPA was incorporated into lipids of
GT1–7 neurons in a time-dependent manner that reached
equilibrium within about 5 hr at 10 ␮g/mL. Incorporation
into total lipids, phospholipids, and neutral lipids was linear
FIG. 4. VPA concentration curve for in-
corporation of VPA into lipids of GT1–7
neurons. This experiment was performed
similarly to the experiments described in
the legends to Figs. 1 and 2. The cells were
incubated for 180 min with 10, 50, 100,
3
2
00, and 400 ␮g/mL of [ H]VPA, and
then were washed to remove extracellular
radioactivity. Lipids were extracted from
the cells and analyzed by TLC. The incor-
poration of VPA into the phospholipid
(
open circles) and neutral lipid (solid dia-
monds) fractions is expressed as picomoles
incorporated. Also reported are picomoles
of VPA taken into the cells (solid circles).
Each symbol represents the mean and SEM
for two experiments in duplicate.

Incorporation of VPA into Lipids in GT1–7 Neurons
211
with the concentration of VPA. We accomplished the first
synthesis of PC-VPA, thereby establishing that this com-
monocarboxylyl groups have been isolated from a bovine
brain lipid extract, apparently formed by lipid peroxidation
[24].
pound is stereospecifically possible and stable. The R value
f
for PC-VPA was identical to the R value for one of the
principal phospholipids into which [ H]VPA was incorpo-
The function of membrane-associated enzymes may be
altered by the presence of asymmetric phospholipids. For
example, when rat hepatocyte plasma membranes were
enriched with semisynthetic phosphatidylcholines with
short acyl chain moieties of 10–12 carbons at the sn-2
position, the activity of endogenous membrane phospho-
f
3
rated, supporting the likelihood that PC-VPA is formed in
neuronal membranes with valproate administration.
Previous studies have failed to identify the incorporation
of VPA into phospholipids in the brain. Aly and Abdel-
3
Latif [25] did not detect incorporation of [ H]VPA into
lipase A was decreased while the activity of acyl-CoA:
2
brain phospholipids 30 min after the i.p. administration of
the drug to rats. Becker and Harris [26] found no evidence
for the accumulation of valproyl-CoA in rat brain 30 min
after the i.p. administration of valproate, although it
occurred in liver tissue. Perhaps the 30-min time period of
these experiments was insufficient for measurable incorpo-
ration because of the time required for distribution of VPA
into the brain and of other factors such as enzyme turnover
rate that could further delay the incorporation. In this
study, with direct application of VPA to the GT1–7
neurons in culture, the time required for saturation of VPA
incorporation was approximately 5 hr at a concentration of
lysophospholipid acyl transferase was increased [29]. Intro-
duction of VPA, a branched short chain fatty acid, into a
membrane phospholipid reduces the symmetry of the re-
sultant molecule. The incorporation of adequate amounts
of an asymmetric phospholipid into neuronal membranes
thus may alter membrane structure and, consequently,
function. Also, Bola n˜ os and Medina [7, 30] have reviewed
the effects of valproate on metabolism in the CNS. They
concluded that valproate interfered with brain lipid syn-
thesis and altered the structure of neuronal membranes.
These findings may be explained, in part, by the incorpo-
ration of VPA into phospholipids, but this possibility was
not discussed in these reports.
1
0 ␮g/mL. At 30 min, only about 30–40% of maximal
incorporation had been achieved (Fig. 3). It is possible that
an even longer time is necessary for detectable synthesis of
valproyl-CoA or incorporation of VPA into rat brain
phospholipids.
The linear relationship between the concentration of
VPA and the incorporation into lipids in GT1–7 neurons
was unexpected, because it was not a saturable process. Our
3
data show that the incorporation of [ H]VPA into phos-
Becker and Harris [26] reported that VPA did not change
cerebral acetyl-CoA content and that no valproyl-CoA was
found in brain (as mentioned above), hippocampal prisms,
or brain mitochondria. These data indicate that the pre-
cursor of VPA incorporation into phospholipids did not
accumulate at detectable levels in these experiments. Nev-
ertheless, the activation of fatty acids of three and more
carbons has been shown in rat brain homogenates [27].
Although we did not measure valproyl-CoA itself, our
experiments suggest that valproyl-CoA could be formed in
GT1–7 neurons, so it is not clear at this time why the levels
were not detectable in Becker’s experiments.
The possibility of the incorporation of VPA into lipids as
a potential mechanism for its biochemical and physiologi-
cal effects is heuristically interesting. It has been established
that most of the hydrophobic lipid moieties of neuronal
membranes are very long chain, highly unsaturated, and, in
some cases, hydroxylated residues. Each lipid class contains
characteristic molecular species. The contribution of these
compounds to the functional properties of neuronal mem-
branes such as excitability or transmembrane receptor-
mediated signal conduction is not well elucidated. In
general, biological membranes contain species of phospho-
lipids that possess hydrocarbon chains of similar length at
the sn-1 and sn-2 positions. However, asymmetric molec-
ular species with acyl moieties of widely differing lengths
have been observed. PAF is an example of an asymmetric
phospholipid with a wide spectrum of biological activities
pholipids was linear over the concentrations of VPA tested,
5–400 ␮g/mL. This range of concentrations was chosen
because the concentration of VPA in plasma that is
associated with therapeutic antiepileptic and antimanic
responses is 50–100 ␮g/mL. Studies in laboratory animals
establish that the pharmacokinetics of plasma and brain
levels of VPA correspond [1]. It may be that a critical level
of valproyl-containing lipids in biological membranes exists
for alteration of signal transduction pathways critical to
VPA’s pharmacodynamic properties fundamental for its
antiepileptic or antimanic properties. Studies to character-
ize the rate of turnover and the effects of valproyl-phos-
pholipid incorporation on membrane function are required
to test for and understand this possibility.
We thank Dr. Donald J. Hanahan in San Antonio for valuable
discussions and support in the completion of this project, Dr. Susan
Weintraub in San Antonio for performing the electrospray mass
spectrometry, Dr. Demetrios Argyropoulos in Athens for performing
the nuclear magnetic resonance, Dr. Akis Froussios in Athens for help
with the phospholipid synthesis, and Dr. Xiaoying Chang in San
Antonio for help with the GT1–7 neuron cell cultures.
References
1
. Chapman A, Keane PE, Meldrum BS, Simiand J and Ver-
nieres JC, Mechanism of anticonvulsant action of valproate.
Prog Neurobiol 19: 315–359, 1982.
. Bowden CL, Brugger AM, Swann AC, Calabrese JR, Janicak
PG, Petty R, Dilsaver SC, Davis JM, Rush AJ, Small JG,
Garza-Trevino ES, Risch SC, Goodnick PJ and Morris DD,
2
[28]. Also, asymmetric phospholipids with a short chain,
sn-2 positioned monocarboxylyl, dicarboxylyl, or hydroxy-

2
12
A. Siafaka-Kapadai et al.
Efficacy of divalproex vs lithium and placebo in the treatment
of mania. JAMA 271: 918–924, 1994.
17. Findell PR, Wong KH, Jackman JK and Daniels DV, ␤-Ad-
renergic and dopamine (D )-receptors coupled to adenylyl
1
3
. Loscher W, Effects of the antiepileptic drug valproate on
metabolism and function of inhibitory and excitatory amino
acids in the brain. Neurochem Res 18: 485–502, 1993.
. Javors MA and Erwin VG, Effects of benzodiazepines and
valproic acid on brain aldehyde reductase and a proposed
mechanism of anticonvulsant action. Biochem Pharmacol 29:
cyclase activation in GT1 gonadotropin-releasing hormone
neurosecretory cells. Endocrinology 132: 682–688, 1993.
1
8. Hales TG, Kim H, Longoni B, Olsen RW and Tobin AJ,
Immortalized hypothalamic GT1–7 neurons express func-
tional GABA type A receptors. Mol Pharmacol 42: 197–202,
4
1
992.
1
703–1708, 1980.
1
9. Hales TG, Sanderson MJ and Charles AC, GABA has
excitatory actions on GnRH-secreting immortalized hypotha-
lamic (GT1–7) neurons. Neuroendocrinology 59: 297–308,
5
6
. Chen G, Manji HK, Hawver DB, Wright CB and Potter WZ,
Chronic sodium valproate selectively decreases protein kinase
C ␣ and ε in vitro. J Neurochem 63: 2361–2364, 1994.
. Petty F, Rush AJ, Davis JM, Calabrese JR, Small JG, Swann
AC, Kramer GL, Orsulak PJ, Blake M and Bowden CL,
Plasma GABA predicts acute response to divalproex in
mania. Biol Psychiatry 39: 278–284, 1996.
1
994.
0. Lee A, Talley E, Rosin DL and Lynch KR, Characterization of
␣2A-adrenergic receptors in GT1 neurosecretory cells. Neu-
2
2
2
roendocrinology 62: 215–225, 1995.
1. Bligh EG and Dyer WJ, A rapid method of total lipid
extraction and purification. Can J Biochem Physiol 37: 911–
7
. Bola n˜ os JP and Medina JM, Effect of valproate on lipogenesis
in neonatal rat brain. Biochem Pharmacol 45: 1283–1288, 1993.
. Dellantonio R, Guratola G, Angeleri F, Paladini D, Rec-
chioni MA, Sirocchi G, Angeleri VA and Giannini L, Effect
of valproic acid on fluidity of brain mitochondrial mem-
branes. Adv Epileptol 17: 101–104, 1989.
9
17, 1959.
8
2. Siafaka-Kapadai A and Hanahan DJ, An endogenous inhib-
itor of PAF-induced platelet aggregation isolated from rat
liver has been identified as free fatty acid. Biochim Biophys
Acta 1166: 217–221, 1993.
9
. Brown NA, Clarke DO and McCarthy A, Adaptation of
postimplantation embryos to culture: Membrane lipid synthe-
sis and response to valproate. Reprod Toxicol 5: 245–253,
23. Siafaka-Kapadai A, Hanahan DJ and Javors MA, Oleic
2
ϩ
acid-induced Ca mobilization in human platelets: Is oleic
acid an intracellular messenger? J Lipid Mediat Cell Signal 15:
215–232, 1997.
1
991.
1
1
1
0. Binaglia L, Effect of pyridoxal-5Ј-phosphate and valproic acid
on phospholipid synthesis in neuroblastoma NA. Biochem
Pharmacol 38: 3407–3413, 1989.
24. Tanaka T, Minamino S, Unezaki H, Tsukatani H and
2
ϩ
Tokumura A, Formation of PAF-like phospholipids by Fe
/
1. Bellringer ME, Rahman K and Coleman R, Sodium valproate
inhibits the movement of secretory vesicles in rat hepato-
cytes. Biochem J 249: 513–519, 1988.
ascorbate/EDTA-induced lipid peroxidation. Biochim Biophys
Acta 1166: 264–274, 1993.
5. Aly MI and Abdel-Latif AA, Studies of distribution and
metabolism of valproate in rat brain, liver, and kidney.
Neurochem Res 5: 1231–1242, 1980.
6. Becker C-M and Harris RA, Influence of valproic acid on
hepatic carbohydrate and lipid metabolism. Arch Biochem
Biophys 223: 381–392, 1983.
7. Aas M, Organ and subcellular distribution of fatty acid
activating enzymes in the rat. Biochim Biophys Acta 231:
2
2
2
2
2. Mellon PL, Windle JJ, Goldsmith PC, Padula CA, Roberts JL
and Weiner RI, Immortalization of hypothalamic GnRH
neurons by genetically targeted tumorigenesis. Neuron 5:
1
–10, 1990.
1
1
3. Mellon PL, Windle JJ and Weiner RI, Immortalization of
neuroendocrine cells by targeted oncogenesis. Recent Prog
Horm Res 47: 69–93, 1991.
4. Mellon PL, Wetzel WC, Windle JJ, Valenca MM, Goldsmith
PC, Whyte DB, Eraly SA, Negro-Vilar A and Weiner RI,
Immortalized hypothalamic gonadotropin-releasing hormone
neurons. Ciba Found Symp 168: 104–126, 1992.
3
2–47, 1971.
8. Koltai M, Hosford D, Guinot P, Esanu A and Braquet P,
Platelet activating factor (PAF): A review of its effects,
antagonists, and possible future clinical implications. Drugs
1
1
5. Krsmanovic LZ, Stojilkovic SS, Balla T, al-Damluji S, Weiner
RI and Catt KJ, Receptors and neurosecretory actions of
endothelin in hypothalamic neurons. Proc Natl Acad Sci USA
4
2: 9–29, 1991.
2
9. Koumanov KS, Momchilova-Pankova AB, Markovska TT,
Quinn PJ and Wolf C, Phospholipids with a short acyl-chain
modulate phospholipase and acyltransferase activities. J Lipid
Mediat Cell Signal 12: 59–76, 1995.
88: 11124–11128, 1991.
6. Milenkovi c´ L, D’Angelo G, Kelly PA and Weiner RI,
Inhibition of gonadotropin hormone-releasing hormone re-
lease by prolactin from GT1 neuronal cell lines through
prolactin receptors. Proc Natl Acad Sci USA 91: 1244–1247,
30. Bola n˜ os JP and Medina JM, Effect of valproate on the
metabolism of the central nervous system. Life Sci 60:
1933–1942, 1997.
1
994.