Báo cáo khoa học: Brain lipid composition in postnatal iron-induced motor behavior alterations following chronic neuroleptic administration in mice potx

Brain lipid composition in postnatal iron-induced motor
behavior alterations following chronic neuroleptic
administration in mice
Giorgis Isaac
1
*, Anders Fredriksson
2
, Rolf Danielsson
1
, Per Eriksson
3
and Jonas Bergquist
1
1 Department of Analytical Chemistry, Biomedical Center, Uppsala University, Sweden
2 Department of Neuroscience, Psychiatry Ullera
˚
ker, Uppsala University, Sweden
3 Department of Environmental Toxicology, Evolutionary Biology, Uppsala University, Sweden
The physical properties and function of biological
membranes are mediated to a large extent by their
lipid composition. Not only lipid classes such as phos-
phatidylcholine (PC), cholesterol, phosphatidylethanol-
amine, sphingomyelin (SM) but also the molecular
species composition of the various phospholipids (PLs)
and SM have to be determined in the characterization
of lipid membranes [1].
Phospholipids comprise a major part of the cell
membrane, giving the membrane its structure and
integrity. SM is an important constituent in nervous
tissue and plasma membrane of higher animals and is
a phosphocholine ester of ceramide [2,3]. A typical

structure of a PL consists of three parts: a glycerol
backbone, a polar head group and two fatty acid
chains esterified at the sn-1 and sn-2 positions. The
Keywords
Neuroleptics; neurodegenerative and
psychiatric disorders; phosphatidylcholine;
postnatal iron treatment; sphingomyelin
Correspondence
J. Bergquist, Department of Analytical
Chemistry, Uppsala University, Box 599,
SE-751 24 Uppsala, Sweden
Fax: +46 18 4713692
Tel: +46 18 4713675
E-mail:
*Present address
Division of Biology, Kansas State University,
Manhattan, KA, USA
(Received 22 December 2005, revised 23
Feburary 2006, accepted 20 March 2006)
doi:10.1111/j.1742-4658.2006.05236.x
Several studies have shown that deficient uptake or excessive break down
of membrane phospholipids may be associated with neurodegenerative and
psychiatric disorders. The purpose of the present study was to examine the
effects of postnatal iron administration in lipid composition and behavior
and whether or not the established effects may be altered by subchronic
administration of the neuroleptic compounds, clozapine and haloperidol.
In addition to motor activities such as locomotion, rearing and activity, a
targeted lipidomics approach has been used to investigated the brains of
eight groups of mice (four vehicle groups and four iron groups) containing
six individuals in each group treated with vehicle, low dose clozapine, high

dose clozapine and haloperidol. Lipids were extracted by the Folch method
and analyzed using reversed-phase capillary liquid chromatography coupled
on-line to electrospray ionization mass spectrometry (LC ⁄ ESI ⁄MS). Identi-
fication of phosphatidylcholine (PC) and sphingomyelin (SM) molecular
species was based on their retention time, m ⁄ z ratio, head group specific
up-front fragmentation and analysis of the product ions produced upon
fragmentation. A comparison between the Ve-groups and Fe-groups
showed that levels of PC and SM molecular species and motor activities
were significantly lower in Fe-Ve compared to Ve-Ve. The effects of neuro-
leptic treatment with and without iron supplementation were studied. In
conclusion our results support the hypothesis that an association between
psychiatric disorders and lipid and behavior abnormalities in the brain
exists.
Abbreviations
AD, Alzheimer’s disease; EFA, essential fatty acid; IS, internal standard; LC ⁄ ESI ⁄ MS, liquid chromatography coupled on-line to electrospray
ionization mass spectrometry; PD, Parkinson
’
s disease; PCA, principal component analysis; PL, phospholipids; PLA
2
, phospholipase A
2
.
2232 FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS
specific hydrophilic head group varies and includes
choline, ethanolamine, serine and inositol. As the fatty
acid chains can vary in length and degree of unsatura-
tion, each natural PL and SM contain numerous
molecular species. The fatty acid tails can range from
16 to 24 carbons; typically one chain is saturated
(sn-1) and the second (sn-2) contains one or more dou-

ble bonds [4].
Brain lipids have many physiological functions. The
phospholipid structure of neuronal membrane is essen-
tial for normal functioning of the nervous system.
Numerous hypotheses have been proposed to concep-
tualize the pathophysiology of neurodegenerative and
psychiatric disorders, e.g. schizophrenia. One of these
gaining a major research interest is the study of mem-
brane composition and function, the phospholipid
membrane hypothesis, originated by Feldberg [5] and
Horrobin [6]. The phospholipid hypothesis suggests
that in some neurodegenerative and psychiatric dis-
orders the metabolism and structure of membrane
phospholipids are abnormal, not just in brain but also
in other tissues, e.g. red blood cells [7–9]. Of particular
interest in the context of these disorders are the pro-
found effects that membrane structure can exert on
receptor function. The specific essential fatty acid
(EFA) content of synaptic membrane can modify neur-
onal functions and produce clinical effects through at
least two mechanisms: first, changes in EFA content
alter the microenvironment and hence structure and
function of membrane receptors, ion channels and
enzymes. Second, EFAs contribute to cell regulation
by acting as a source of precursor for second messen-
gers in intra- and intercellular signal transduction
[4,10,11]. The same membrane protein can behave
quite differently when embedded in a membrane com-
posed of saturated fatty acids or one in which unsatur-
ated fatty acids predominate [4,8]. A single change in

membrane structure could produce changes in the
behavior of all the receptor types associated with that
membrane [4,7,8,11] and thus alter the behavior and
response of any neurotransmitter system. In general,
membrane hypotheses are appealing because of their
apparent ability to account for a range of disparate
findings related to various disorders [4].
Iron, the most abundant nonalkaline metal in the
human body and brain [12], is involved in several
metabolic processes and is essential for a normal neu-
rological development [13], and iron deficiency during
critical periods of development is associated with dis-
ruption of behavioral performance, e.g. in learning and
memory tasks [13]. Nevertheless, there is accumulating
evidence that excessive iron deposits in the brain,
which may generate cytotoxic free radical formation
[14,15], and alterations in iron metabolism play an
important role in many neurologic diseases [16–18].
Iron-overload and ⁄ or defects of the iron-dependent
enzyme complex are implicated in the pathogenesis of
neurodegenerative disorders, e.g. Parkinson
’
s disease
(PD) and Alzheimer,s disease (AD) [19–21] and in the
Hallervorden–Spatz syndrome, involving aberrant
brain iron metabolism [22]. Several studies have dem-
onstrated that elevations of iron are found in the sub-
stantia nigra of patients afflicted with PD [23–26]. As
demonstrated in several previous studies [27–31], post-
natal administration of Fe

2+
(7.5 mgÆkg
)1
, on days
10–12) increased the level of iron in the adult mice and
rat brain together with alterations in motor behavior
performance. This postnatal iron administration can
be used as a model for oxidative stress induced brain
damage and neurodegenerative disorders.
Haloperidol, a butyrophenone, first developed as an
analgesic, has found wide use as an antipsychotic agent
due to its neuroleptic action. It is known to block the
dopamine receptors of the D2 type which is accepted
as a mechanism for its neuroleptic action [32]. Cloza-
pine is rather novel and unique prototype atypical,
dibenzodiazepine derivative, antipsychotic agent. It has
been proven effective and significantly superior to pla-
cebo, as well as to conventional neuroleptics. Approxi-
mately 30–60% of all schizophrenic patients who fail
to respond to typical antipsychotics may respond to
clozapine [33].
Membrane PL abnormalities have been studied previ-
ously in humans and animals using conventional chro-
matographic methods in brain [11,34–36], platelets and
red blood cells [36,37] and using
31
P-magnetic resonance
spectroscopy (P-MRS) [36,38–41]. Conventional analy-
sis methods of PL and SM molecular species require
laborious procedures including separation by column,

argentation thin-layer chromatography or liquid chro-
matography (LC) after pre or postcolumn derivatization
[42,43]. A general drawback of these approaches is that
they preclude the analysis of metabolic studies and the
derivatization may subject the lipid to the danger of
rearrangement of the fatty acyl chains on the glycerol
backbone. It is also more time consuming to obtain the
derivatives compared to analyzing a total lipid extract
directly without any derivatization [44]. Although the
in vivo determination of metabolic processes in the brain
by P-MRS is a powerful method, individual molecular
species of phospholipid membrane components can not
be measured using present technology [37]. Therefore
brain extracts have been used to study the pattern and
abnormalities of individual molecular species in the cel-
lular level in more detail using LC coupled on-line
to mass spectrometry (MS) previously developed by
G. Isaac et al. Brain lipid composition in postnatal iron-induced mice
FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS 2233
Isaac et al. [45]. The separation power of LC into either
different lipid classes or molecular species within a class
together with selective MS detection makes the simulta-
neous determination of protonated molecules and the
identification of several structural elements possible
[45,46].
The purpose of the present study was to examine
both the effect of postnatal iron administration in lipid
composition and motor activities and whether or not
the established effects may be altered by subchronic
administration of the neuroleptic compounds, cloza-

pine and haloperidol. We investigated the PC and SM
molecular species in brains from eight groups of mice
treated with vehicle, low dose clozapine, high dose
clozapine and haloperidol.
Results and Discussion
Lipids make up a very high proportion of the brain
(50–60% of dry weight) [47]. A major proportion of
these lipids are EFAs mainly bound to PLs [47]. The
membrane is a complex structure composed primarily
of phospholipids and their constituent fatty acids, that
provide scaffolding for many key functional systems,
including neurotransmitter receptor binding, signal
transduction and transmembrane ion channels. Thus,
the dynamic state of all membranes, including those of
neurons and glia, is dependent on their composition,
such that small changes in key phospholipids or the
polyunsaturated fatty acids that make up phospholi-
pids can lead to a broad range of membrane dysfunc-
tion. The interest in the analysis of lipids in general
and PL and SM in particular is continuously increas-
ing due to the importance of these molecules related to
various disorders. Depending on the polar head group,
fatty acid chain length and number of double bonds
numerous individual lipid molecular species are present
in a cell. These species differ greatly in their chemical
and biological properties under pathological conditions
and their identification and quantitation are of great
interest. Recent advances in LC ⁄ MS have greatly
enhanced the identification and profiling of individual
molecular species. These advances have made possible

the field of lipidomics, which aims to identify all
endogenous lipids [48–50]. In this study a targeted lipi-
domics approach using LC ⁄ ESI ⁄ MS was applied to
the different treatment groups in order to examine
alteration in PC and SM molecular species (Table 1).
LC/ESI/MS of lipid extracts
First direct infusion of crude lipid extracts from mouse
brain was tested as suggested by Han et al. [51]. How-
ever this approach did not yield acceptable signals
probably reflecting signal suppression by other compo-
nents in the crude lipid extract. In order to reduce this
suppression effect, the LC ⁄ ESI ⁄ MS method previously
developed [45] for the analysis of PC and SM mole-
cular species was applied with minor modifications.
The LC column was changed to a larger internal dia-
meter (0.5 mm) in order to avoid frequent plugging of
the column, variation in retention time from run to
run and to accommodate the large number of runs
from the crude lipid extracts. Identification of PC and
SM molecular species was based on their retention
time, m ⁄ z ratio, head group specific up-front fragmen-
tation and analysis of the product ions produced upon
fragmentation. Shown in Fig. 1 are representatives of
the extracted normalized ion chromatograms of the
most abundant protonated molecular ions of PC and
SM molecular species from crude mouse brain extract
spiked with deuterium labeled 1,2-dipalmitoyl-sn-
glycero-3-phosphocholine (d4–16 : 0⁄ 16 : 0) internal
standard (IS). The actual peak area of individual
molecular species normalized against the IS in the

vehicle group and iron group after three weeks of low
dose clozapine (1 mgÆkg
)1
), high dose clozapine
(5 mgÆkg
)1
) and haloperidol (1 mgÆkg
)1
) treatment are
shown in Table 2.
Effect of neuroleptics and iron treatment
on lipid membrane
A principal component analysis (PCA) of all lipid
results (11 variables for 90 samples) resulted in loa-
dings for the first two components (explaining 76%
and 10%, respectively, of the total variance) according
to Fig. 2A. The axes were scaled in proportion to the
square root of the explained variance; hence the influ-
ence of a variable is directly related to its distance
from the origin. It is seen that all the variables have
about the same influence on the lipid variations
explained by the two components. Moreover, the cor-
relation between two variables is the cosine of the
angle between their connection lines to the origin. It is
seen that most of the variables are clustered, indicating
Table 1. The table shows the eight treatment groups. For details in
treatment see Experimental procedures.
Treatment Ve-groups Fe-groups
Vehicle Ve-Ve Fe-Ve
Clozapine (1mgkg

)1
) Ve-C1 Fe-C1
Clozapine (5mgkg
)1
) Ve-C5 Fe-C5
Haloperidol (1mgkg
)1
) Ve-H1 Fe-H1
Brain lipid composition in postnatal iron-induced mice G. Isaac et al.
2234 FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS
similar variation pattern, although with some deviation
for PC5, PC6 and especially PC3. The correlation coef-
ficient between PC1 and PC4 (bracketing the cluster) is
0.71, while that between PC1 and PC3 (the least con-
cordant variables) is only 0.16.
According to Fig. 2A, the first principal component
is a linear combination with positive contributions
(loadings) from all variables, but to some lesser degree
from PC3 which then dominates the second compo-
nent. The score values for the two components,
together accounting for 86% of the variance, were then
analyzed by nested anova to separate the contribu-
tions from different treatments, individual animals and
replicate analyses. Only the scores for the first compo-
nent showed significant difference between treatment
groups, which is the systematic variation we are look-
ing for. The random variation in the first component
(not related to treatment) was mainly due to individual
animals with less contribution from the repeated analy-
sis (87% and 13%, respectively).

Justified by these findings, the further data analysis
was performed on the scores for the first principal
components rather than on the separate lipid concen-
trations. For each animal the average score from the
two runs was taken as an over-all measure of the lipid
concentration. In a few cases only one measurement
was available, but as the repeatability error was com-
paratively low the single score value could be used. A
one-way anova as well as Kruskal–Wallis’ nonpara-
metric test showed highly significant variation between
the treatment groups (P<0.001). The mean values
(least squares means) for the different groups are
shown in Fig. 2B. The neuroleptic treated groups are
shown with error bars related to the least significant
difference. The vehicle treated groups, Ve-Ve and
Fe-Ve, define reference levels depicted as error bands.
Hence there is a significant difference between two
means at approximately 5% significance level when the
error bars or bands are not overlapping. The signifi-
cance is further confirmed with an ordinary two-sided
t-test between two means, and the P-value is reported.
The confirmation of anova and least significant differ-
ence (LSD) comparisons with other tests is due to the
slightly significant inhomogeniety of variance accord-
ing to Bartlett’s test (P ¼ 0.034).
First, it can be noted that iron supplementation
results in a significant decrease in the amount of PC
and SM molecular species (nonoverlapping reference
bands, P < 0.025). The effect of subsequent neurolep-
tic treatment for three weeks can be examined by com-

paring the corresponding error bars with the Fe-Ve
band. Both low and high dose clozapine (Fe-C1 and
Fe-C5) treated mice did not show significant differ-
ences in lipid concentration compared to Fe-Ve. How-
ever haloperidol (Fe-H1) significantly increased the
lipid concentration compared to Fe-Ve (P ¼ 0.009)
and to both low and high dose clozapine treated mice.
(P ¼ 0.004 and P < 0.001, respectively).
The intrinsic effect of the neuroleptic treatment was
studied by comparing Ve-Ve against Ve-C1, Ve-C5
and Ve-H1. As can be seen in Fig. 2B there is no signi-
ficant change of the PC and SM molecular species
when low dose clozapine (Ve-C1) is given compared to
vehicle (Ve-Ve). When high dose clozapine (Ve-C5) or
15
Normalized Intensity
0
8
10 20 25 30 35 40 45 50
PC1
PC3
PC8
PC6
PC7
IS
SM1
PC9
0
0.25
10 15 20 25 30 35 40 45 50

SM2
PC2
Time (min)
Fig. 1. Extracted normalized ion chromat-
ograms of the most abundant protonated
molecular ions of PC and SM molecular
species. The peak intensities are normal-
ized to the intensity of the deuterium
labeled 1,2-dipalmitoyl-sn-glycero-3-phos-
phocholine (d4–16 : 0 ⁄ 16 : 0) internal
standard (IS). The molecular species
were resolved and identified based on
their retention time, m ⁄ z ratio, head
group specific up-front fragmentation
and analysis of the product ions pro-
duced upon fragmentation. The inserted
chromatogram shows the low intensity
species. Peak identification as in Table 2.
G. Isaac et al. Brain lipid composition in postnatal iron-induced mice
FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS 2235
haloperidol (Ve-H1) is given there is a significant
decrease in lipid concentration compared to Ve-Ve
(P ¼ 0.037 and p ¼ 0.008, respectively). It was also
shown by Singh et al. that haloperidol reduced biosyn-
thesis of various phospholipids in rat brain by follow-
ing incorporation of
32
P into individual classes of
phospholipid [32]. The traditional neuroleptic halo-
peridol decreased the PC and SM molecular species

concentration in comparison to clozapine.
A comparison was also made between the vehicle
and iron groups to examine the difference between
iron and vehicle supplemented groups after subchronic
administration of the neuroleptic compounds. An
interesting phenomenon is that lipid concentrations
were lower in Fe-C1 compared to Ve-C1 (P ¼ 0.038),
no significant difference with high dose clozapine and
actually higher for mice treated with haloperidol
(P<0.001).
Effect of neuroleptics and iron treatment
on behavior
Motor activities such as locomotion (L), rearing (R)
and activity (A) were measured for short (20 min) and
long (60 min) terms as mentioned under experimental
section. The motor activities were analyzed by PCA in
a similar way as the lipid data. The loading plot
(Fig. 3A) for the first two components, explaining
83% and 9%, respectively, shows that the behavioral
variables contribute almost equally to the dominating
first component. The second component seems to
reflect the difference between short-term and long-term
behavior. The lowest correlation coefficient, 0.58, is
found between L60 and R20.
As for the lipid data, one way anova (as well as
Kruskal–Wallis’ test) showed a highly significant dif-
ference between the treatment groups (P<0.001) for
the score values along the first component. Barlett’s
test indicated slightly inhomogenuos variance (P ¼
0.025), which gives cause for the confirmation with

alternative tests. Also for the second component the
scores showed a significant variation between the
groups (P ¼ 0.012). However, in view of the lower sig-
nificance and the small proportion of explained vari-
ance (9%) the scores for the second component were
not further analyzed.
The scores for the first principal components were
taken as a single-valued measure of the behavior (loco-
motion, rearing and activity) for each animal. The
mean value for each group is shown in Fig. 3B in
the same way as for lipid concentrations (cf. Fig. 2B).
The clear gap between the reference bands (Ve-Ve,
Fe-Ve) shows the highly significant decrease in motor
Table 2. Data given as peak area of the molecular species normalized against the peak area of the internal standard [IS (mean values ± SEM)]. Each treatment group contains 6 animals
except treatments Fe-Ve and Fe-C1 which contain 5 animals due to losses during the extraction procedure. The duplicate measurements for each animal were averaged, and for each
group the mean values and the corresponding standard errors of the mean were calculated.
Treatment Ve-Ve Ve-C1 Ve-C5 Ve-H1 Fe-Ve Fe-C1 Fe-C5 Fe-H1
Number of animals 6 · 26· 26· 26· 25· 25· 26· 26· 2
PC1 (16 : 0 ⁄ 22 : 6, m ⁄ z 806.6) 2.70 ± 0.03 3.15 ± 0.02 2.59 ± 0.05 2.30 ± 0.08 2.38 ± 0.11 2.41 ± 0.06 2.60 ± 0.10 3.13 ± 0.02
PC2 (16 : 0 ⁄ 18 : 2, m ⁄ z 758.6) 0.61 ± 0.02 0.63 ± 0.01 0.52 ± 0.02 0.47 ± 0.02 0.42 ± 0.02 0.50 ± 0.03 0.48 ± 0.03 0.59 ± 0.03
PC3 (18 : 2 ⁄ 18 : 2, m ⁄ z 782.6) 0.79 ± 0.05 0.77 ± 0.01 0.91 ± 0.04 0.76 ± 0.05 0.54 ± 0.04 0.59 ± 0.01 0.72 ± 0.11 0.80 ± 0.08
PC4 (18 : 1 ⁄ 20 : 4, m ⁄ z 808.6) 1.13 ± 0.05 1.36 ± 0.01 1.17 ± 0.06 0.95 ± 0.05 0.86 ± 0.02 0.90 ± 0.05 1.02 ± 0.08 1.25 ± 0.06
PC5 (16 : 0 ⁄ 16 : 0, m ⁄ z 734.6) 9.60 ± 0.32 9.88 ± 0.10 9.23 ± 0.32 8.53 ± 0.22 7.19 ± 0.18 8.13 ± 0.37 8.21 ± 0.78 9.79 ± 0.43
PC6 (16 : 0 ⁄ 18 : 1, m ⁄ z 760.6) 16.11 ± 0.60 16.49 ± 0.20 15.64 ± 0.60 15.01 ± 0.51 13.42 ± 0.27 14.63 ± 0.81 14.49 ± 1.26 16.52 ± 0.87
PC7 (18 : 0 ⁄ 18 : 2, m ⁄ z 786.6) 3.72 ± 0.11 4.14 ± 0.05 3.55 ± 0.07 3.21 ± 0.04 2.97 ± 0.09 3.47 ± 0.11 3.29 ± 0.16 3.98 ± 0.15
PC8 (18 : 0 ⁄ 20 : 4, m ⁄ z 810.6) 5.36 ± 0.20 6.03 ± 0.07 5.19 ± 0.14 4.61 ± 0.15 4.18 ± 0.14 4.76 ± 0.17 4.71 ± 0.24 5.72 ± 0.26
PC9 (18 : 0 ⁄ 18 : 1, m ⁄ z 788.6) 10.19 ± 0.44 11.62 ± 0.18 9.69 ± 0.37 8.67 ± 0.30 7.86 ± 0.16 9.37 ± 0.35 9.01 ± 0.41 11.11 ± 0 .44
SM1 (18 : 0, m ⁄ z 731.6) 2.46 ± 0.09 2.88 ± 0.03 2.39 ± 0.07 2.00 ± 0.06 1.78 ± 0.05 2.25 ± 0.19 2.19 ± 0.14 2.67 ± 0.10
SM2 (24 : 1, m ⁄ z 813.6) 0.59 ± 0.03 0.71 ± 0.02 0.57 ± 0.04 0.50 ± 0.03 0.43 ± 0.01 0.57 ± 0.04 0.56 ± 0.04 0.69 ± 0.02
Brain lipid composition in postnatal iron-induced mice G. Isaac et al.

2236 FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS
activities with iron supplementation (P<0.001). Sub-
sequent treatment with low dose clozapine (Fe-C1)
had no significant effect in the activities compared to
Fe-Ve. Haloperidol (Fe-H1) and the higher dose of
clozapine (Fe-C5) elevated the motor activities signifi-
cantly compared to Fe-Ve ( P<0.001). The activities
even raised to the same level as Ve-Ve, thus the be-
havioral effect of iron supplementation was abolished
by the neuroleptic treatment.
The intrinsic effect of neuroleptic treatment on
motor activities was studied by comparing Ve-Ve
against Ve-C1, Ve-C5 and Ve-H1. As can be seen in
Fig. 3B there is no significant difference for low (Ve-
C1) and high (Ve-C5) dose clozapine treated mice
compared to Ve-Ve but a significant increase with
haloperidol (Ve-H1, p ¼ 0.035).
The difference in motor activities induced by iron
supplementation (Ve-Ve compared to Fe-Ve) was
about the same after low dose clozapine treatment
(Ve-C1 versus Fe-C1). However, there was no signifi-
cant difference related to iron supplementation for the
groups treated with high dose clozapine (Ve-C5 versus
Fe-C5) or haloperidol (Ve-H1 versus Fe-H1).
The combined effects in motor activities and lipid
concentrations are visualized in Fig. 4, where the X-axis
represents the lipid concentrations and the Y-axis repre-
sents activities. The ovals represent error regions based
on least significant difference similarly to the error
L60

A60
R60
L20
A20
R20
1st comp
2nd comp
AB
-
Activities
+
Ve-
Fe-
Ve-Ve
Fe-Ve
-C1 -C5 -H1
Fig. 3. (A) Loading plot for the behavioral variables measured for 20 and 60 min. Locomotion (L), rearing (R) and activity (A). (B) Mean scores
for the treatment groups from behavioral data. The neuroleptic treated groups are shown with error bars, while the vehicle treated groups,
Ve-Ve and Fe-Ve, define reference levels depicted as error bands. Overlapping error bars (bands) indicate significant differences according to
the LSD.
B
PC1
PC2
PC4
PC3
PC6
PC5
PC5
PC8
PC7

SM2
SM1
1st comp
2nd comp
A
Ve-Ve
- Lipid conc +
Ve-
Fe-
-C1 -C5 -H1
Fe-Ve
Fig. 2. (A) Loading plot for the lipid concentration variables. Phosphatidylcholine (PC) and sphingomyelin (SM) molecular species identification
as in Table 2. (B) Mean scores for the treatment groups from lipid data. The neuroleptic treated groups are shown with error bars, while the
vehicle treated groups, Ve-Ve and Fe-Ve, define reference levels depicted as error bands. Overlapping error bars (bands) indicate significant
differences according to the least significant difference (LSD).
G. Isaac et al. Brain lipid composition in postnatal iron-induced mice
FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS 2237
bands and error bars in Figs 2 and 3. Hence, when two
ovals are not overlapping there is a significant difference
at approximately 5% significance level. Clearly, the iron
supplemented mice treated with vehicle or low dose
clozapine are similar, standing out from the others.
The immediate postnatal period is critical for estab-
lishment of normal iron content in the adult brain and
its regional distribution. Investigations of cerebral iron
uptake indicate that both iron transport and transfer-
rin, the iron mobilization protein, binding sites are
maximal during the postnatal period of rapid brain
growth, essentially during the second week post par-
tum in rats and mice [52], and maximal brain iron

uptake occurs in 15-day-old rats [53]. Dwork et al.
[54] showed that iron acquired by the brain during this
period of development is retained in the brain without
being returned to plasma sites. In many mammalian
species, a period of rapid brain growth occurs during
perinatal development, termed ‘brain growth spurt’
[55]. In humans this period begins during the third
trimester of pregnancy and is maintained throughout
the first years of life whereas in rats and mice the cor-
responding period occurs during the first three-to-four
weeks of postnatal life. During this critical period of
neuronal development the essential processes of brain
structure and function are established [56] and funda-
mental sensory-motor faculties are acquired [57,58].
Thus, it appears that there may exist a critical neo-
natal period of brain development associated with the
establishment of normal iron content in the adult
brain.
Several studies by Ben-Shachar et al. [14,59,60] have
related the effects of chronic administration of neuro-
leptic compounds to the involvement of iron in the
expressions of neuroleptic-induced dopamine (DA)
supersensitivity. These studies suggest that the mobil-
ization of iron from peripheral tissues into the brain
may exert a role in the mechanism of action of neuro-
leptic compounds. Haloperidol has been shown to alter
the property of the blood–brain barrier by enhancing,
normally restricted, iron transport into the brain
whereas clozapine prohibits iron uptake into the brain
[59]. Furthermore, Ben-Shachar and Youdim [60]

examined the possibility that neuroleptic-induced DA
D
2
receptor supersensitivity involves an alteration in
brain iron content whereby liver nonheme iron stores
were depleted in rats treated chronically with haloper-
idol (5 mgÆkg
)1
daily for 14 or 21 days) as compared
with values in control rats. Ben-Shachar et al. [14]
showed that FeCl
2
(5 mgÆkg
)1
per day over 21 days)
administered to rats treated with chlorpromazine
(10 mgÆkg
)1
per day over 21 days) prevented the
expressions of DA D
2
receptor supersensitivity, both
biochemically and behaviourally (apomorphine-
induced activity).
Moreover, neuroleptics may alter brain phospholipid
metabolism and composition by regulating phospholi-
pase A
2
(PLA
2

) activity. Possible effect of neuroleptics
such as haloperidol on membrane phospholipid meta-
bolism and composition are under investigation.
Several studies have shown that various neuroleptics
reduce the break down of phospholipid membrane by
inhibiting PLA
2
activity [10,38,61–63]. Gattaz et al. [61]
found platelet PLA
2
activity to be greater in schizophre-
nic patients than in healthy controls, and this elevation
was reduced by haloperidol treatment. Trzeciak et al.
[63] reported a decrease in PLA
2
activity after a single
or four week administration of chlorpromazine, triflu-
operazine, haloperidol and sulpiride. Fluphenazine and
thioridazine caused an increase of PLA
2
activity in rat
brain both after a single dose and long-term administra-
tion [63]. Clozapine is good antioxidants and decreases
membrane lipid peroxidation [64] whereas haloperidol
has been reported to have the opposite effect [65].
Conclusions
In summary, a targeted lipidomics approach has been
used to identify the PC and SM molecular species
alteration coupled with behavioral data to investigate
the effects of postnatal iron supplementation. The

effect of subsequent subchronic treatment with neuro-
leptics was also studied. A comparison between the
Ve-groups and Fe-groups showed that levels of PC
Ve-H1
Ve-C5
Ve-Ve
Ve-C1
Fe-C5
Fe-H1
Fe-Ve
Fe-C1
Activities +-
Lipid conc. +-
Fig. 4. Mean scores for the treatment groups from lipid and behavi-
oral data. The oval shapes represent confidence regions related to
the LSD. Two nonoverlapping groups show significant difference
(P<0.05) considering both motor activities and lipid concentrat-
ions. Filled ovals represent mice supplemented with iron (Fe-) and
open ovals those supplemented with vehicle (Ve-).
Brain lipid composition in postnatal iron-induced mice G. Isaac et al.
2238 FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS
and SM molecular species and motor activities were
significantly lower in Fe-Ve compared to Ve-Ve. A
comparison of Fe-Ve with Fe-C1, Fe-C5 and Fe-H1
showed no significant change in PC and SM molecular
species for low and high dose clozapine treated mice
while haloperidol administration significantly increased
the molecular species. Low dose clozapine (Fe-C1) has
no significant effect in motor activities but Fe-C5 and
Fe-H1 elevated the activities significantly compared to

Fe-Ve. The intrinsic effects of neuroleptic treatment in
the Ve-groups (i.e. by comparing Ve-Ve with Ve-C1,
Ve-C5 and Ve-H1) were demonstrated. No significant
change in PC and SM molecular species were found
when the mice were treated with low dose clozapine
while high dose clozapine and haloperidol treated mice
showed significant change in PC and SM molecular
species. There is no significant change in motor activit-
ies for Ve-C1 and Ve-C5 treated mice but with halo-
peridol (Ve-H1) there was a slight increase in activities.
An interesting phenomenon is that both lipid concen-
trations and activities were decreased in Fe-C1 com-
pared to Ve-C1 and for high dose clozapine there is no
significant difference. When the mice were treated with
haloperidol the lipid concentration is significantly
higher in Fe-H1 compared to Ve-H1 while there was
no change in activities.
Experimental procedures
Chemicals and reagents
Lipid standards of l-a-phosphatidylcholine (PC from egg
yolk) and N-acyl-d-sphingosine-1-phosphocholine (SM from
bovine brain) were purchased from Sigma Chemical Co.
(St Louis, MO, USA). A deuterium labeled 1,2-dipalmitoyl-
sn-glycero-3-phosphocholine (d4–16 : 0 ⁄ 16 : 0) was pur-
chased from Avanti Polar Lipids (Alabaster, AL, USA).
Ultrapure water was prepared by Milli-Q water system
(Millipore, Bedford, MA, USA). Other organic solvents and
reagents were of the highest purity commercially available.
Clozapine (Sandoz, Basel, Switzerland) and Haloperidol
(Leo, Ha

¨
lsingborg, Sweden). FerromynÒ (Iron succinate:
3.7 mg Fe
2+
per mL
)1
,ABHa
¨
ssle, Go
¨
teborg, Sweden).
Animals
Pregnant C57 Bl ⁄ 6 mice were purchased from B & K, Sol-
lentuna, Sweden. Each litter adjusted within 48 h to 8–10
mice and to contain offspring of either sex in about equal
number, was kept together with its respective mother in a
plastic cage in a room at temperature of 22 ± 1 °C and a
12 ⁄ 12 h constant light ⁄ dark cycle (lights on between 06.00
and 18.00 h). The male offspring only were used in this
study. At the age of 4 weeks the mice were weaned and the
males were placed and raised in groups of 4–6 animals in a
room maintained for male mice only.
Male mice, postnatal days 10–12, were administered
Fe
2+
(see below) or saline. At 61 days of age and onwards
for three weeks, these mice, weighing 20–24 g, were admin-
istered subcutaneously either a neuroleptic compound
[Clozapine (1 or 5 mgÆkg
)1

) or Haloperidol (1 mgÆkg
)1
)] or
Vehicle (Tween-80). Free access to food and water was
maintained throughout. They were housed in groups of
6 animals and tested only during the hours of light (08.00–
15.00 h). Behavioral testing was initiated three weeks
following the start of treatment with the neuroleptic com-
pounds or vehicle. All testing was performed in a normally
lighted room. This test room, in which all 12 ADEA acti-
vity test chambers, each identical to the home cage, were
placed, was well-secluded and used only for this purpose.
Each test chamber (i.e. motor activity test cage) was placed
in a sound-proofed wooden box with 12 cm thick walls and
front panels and a small double-glass window to allow
observation; each box had a dimmed lighting. Mice were
killed by cervical dislocation within 2 weeks after comple-
tion of behavioral testing. The brains from each group were
dissected and stored in )80 °C until extraction. Experi-
ments were carried out in accordance with the European
Communities Council Directive of 24th November 1986
(86 ⁄ 609 ⁄ EEC) after approval from the local ethical com-
mittee (Uppsala University and Agricultural Research
Council), and by the Swedish Committee for Ethical
Experiments on Laboratory Animals (license S93 ⁄ 92 and
S77 ⁄ 94, Stockholm, Sweden).
Behavioral measurements and apparatus
Activity test chambers
An automated device, consisting of macrolon rodent test
cages (40 · 25 · 15 cm) each placed within two series of

infra-red beams (at two different heights, one low and one
high, 2 and 8 cm, respectively, above the surface of the saw-
dust, 1 cm deep), was used to measure spontaneous motor
activity (RAT-O-MATIC, ADEA Elektronik AB, Uppsala,
Sweden). The distance between the infra-red beams was as
follows: the low level beams were 73 mm apart lengthwise
and 58 mm apart breadth wise in relation to the test cham-
ber; the high level beams, placed only along each long side
of the test chamber were 28 mm apart. According to the
procedures described previously [66], the following para-
meters were measured: locomotion was measured by the
low grid of infra-red beams. Counts were registered only
when the mouse in the horizontal plane, ambulating around
the test-cage. Rearing was registered throughout the time
when at least one high level beam was interrupted, that is
the number of counts registered was proportional to the
amount of time spent rearing. Total activity was measured
G. Isaac et al. Brain lipid composition in postnatal iron-induced mice
FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS 2239
by a sensor (a pick-up similar to a gramophone needle,
mounted on a lever with a counterweight) with which the
test cage was constantly in contact. The sensor registered
all types of vibration received from the test cage, such as
those produced both by locomotion and rearing as well as
shaking, tremors, scratching and grooming. All three
behavioral parameters were measured over three consecu-
tive 20-min periods. The motor activity test room, in which
all 12 ADEA activity test chambers, each identical to the
home cage, were placed, was well-secluded and used only
for this purpose. Each test chamber (that is activity cage)

was placed in a sound-proofed wooden box with 12 cm
thick walls and front panels, and day-lighting. Motor activ-
ity parameters were tested on one occasion only, over three
consecutive 20-min periods, at the age of 3–4 months.
Extraction and Analysis
Brain lipids were extracted by Folch method and PC and
SM molecular species were analyzed with minor modifica-
tion of the previously developed LC ⁄ ESI ⁄ MS method [45].
Design and Treatment
Eight treatment groups were derived from four groups hav-
ing received Fe
2+
as pups on postnatal days 10–12 was
administered orally via a metallic gastric tube in a volume
of 10 mLÆkg
)1
(7.5 mgÆkg
)1
) body weight [Fe-groups] and
four groups having received vehicle (saline) at that same
period [Ve-groups]. Saline was used as vehicle and to pre-
pare the dose of Fe
2+
. On day 61 after birth and for three
weeks following, two Fe-groups and two Ve-groups
received clozapine (1 or 5 mgÆkg
)1
), one Fe-group and one
Ve-group received haloperidol (1 mgÆ kg
)1

), one Fe-group
and one Ve-group received vehicle administration to pro-
vide the groups shown in Table 1. Spontaneous motor be-
havior was initiated 24 h following the final injection of
clozapine, haloperidol or vehicle. Locomotion, rearing and
total activity counts were registered over 60 min.
Statistical Analysis
With principal component analysis (PCA) the concentra-
tions for the different lipid compounds were linearly com-
bined into score values, and the scores rather than single
concentrations were investigated with one-way anova to elu-
cidate the variations between the groups. The aim of this
type of analysis was to enhance the systematic variation,
which could be expected to show up in the few first compo-
nents. A loading plot also reveals to what degree the individ-
ual variables (lipids) contribute to the variation pattern.
Variables with similar loading values are more or less indica-
tors of the same phenomena, which can be confirmed by a
high correlation coefficient. The mean score (one for each
principal component) for a group represents the lipid state
of the group, thus condensed into a single value if only the
first component needs to be considered. If the mean scores
show significant variation between the groups, pair-wise
comparisons may be made using the least significant differ-
ence (LSD). Visually this was accomplished by representing
the means as intervals according to mean ± tÆs ⁄Ö(2n), where
s is the pooled standard deviation and n is the number of
replicates in the treatment group. When the intervals for two
groups are nonoverlapping, there is a significant difference
between the means according to the LSD test. This approach

is based on the approximation Ö (1 ⁄ n
1
+1 ⁄ n
2
)  1 ⁄Ö(2n
1
)
+1 ⁄Ö(2n
2
), which is exact for n
1
¼ n
2
and holds within
0.1% for n
1
¼ 6 and n
2
¼ 5 (the only unequal numbers in
this study). However, the application of anova and the LSD
comparison test requires equal variances as well as normal
distribution for the variations within the groups. Therefore,
when significance was found it was confirmed by Kruskal–
Wallis’ nonparametric test as a robust alternative to anova.
The significant pair-wise group differences were confirmed
by application of an ordinary t-test for the means with
unequal variances, a test that also is fairly insensitive to
deviations from normality distribution. Bartlett’s test for
equal variances was applied in order to find out whether the
confirmation with alternative tests should be performed.

The same multivariate approach was taken for the six
behavioral variables, enabling test of significant differences
in total activities between the groups. Finally, a plot of
lipid scores versus behavioral scores, with each group depic-
ted as a confidence region, was made in order to visualize
the overall influence of postnatal iron supplementation as
well as neuroleptic treatments. All basic calculations were
performed with Microsoft excel 2002 SP3, anova with
minitab release 14 (Minitab Inc., State College, PA, USA)
and PCA with the unscrambler v7.6 SR1 (CAMO Tech-
nologies Inc., Woodbridge, NJ, USA).
Acknowledgements
Financial support from SIDA ⁄ SAREC Foundation
and Uppsala University (GI), the Swedish Society for
Medical Research (J.B), and the Swedish Research
Council (621-2002-5261, 629-2002-6821 JB.) is grate-
fully acknowledged. The analytical instrumentation
was supported by a grant from the Swedish Founda-
tion for Strategic Research (Karin E. Markides). Jonas
Bergquist has a senior research position at the Swedish
Research Council.
References
1 Brouwers JFHM, Gadella BM, van Golde LMG &
Tielens AGM (1998) Quantitative analysis of phosphati-
Brain lipid composition in postnatal iron-induced mice G. Isaac et al.
2240 FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS
dylcholine molecular species using HPLC and light scat-
tering detection. J Lipid Res 39, 344–353.
2 Mano N, Oda Y, Yamada K, Asakawa N & Katayama
K (1997) Simultaneous quantitative determination

method for sphingolipid metabolites by liquid
chromatography ⁄ ionspray ionization tandem mass
spectrometry. Anal Biochem 244, 291–300.
3 Murphy RC, Fiedler J & Hevko J (2001) Analysis of
nonvolatile lipids by mass spectrometry. Chem Rev 101,
479–526.
4 Fenton WS, Hibbeln J & Knable M (2000) Essential
fatty acids, lipid membrane abnormalities, and the diag-
nosis and treatment of schizophrenia. Biol Psychiatry
47, 8–21.
5 Feldberg W (1976) Possible association of schizophrenia
with a disturbance in prostaglandin metabolism: a phy-
siological hypothesis. Psychiatry Med 6, 359–369.
6 Horrobin DF (1977) Schizophrenia as a prostaglandin
deficiency disease. Lancet 1, 936–937.
7 Horrobin DF (1998) The membrane phospholipid hypo-
thesis as a biochemical basis for the neurodevelopmental
concept of schizophrenia. Schizophr Res 30, 193–208.
8 Horrobin DF, Glen AIM & Vaddadi K (1994) The
membrane hypothesis of schizophrenia. Schizophr Res
13, 195–207.
9 Horrobin DF, Manku MS, Hillman H, Iain A & Glen
M (1991) Fatty acid levels in the brains of schizophre-
nics and normal controls. Biol Psychiatry 30, 795–805.
10 Ross BM (2003) Phospholipid and eicosanoid signaling
disturbances in schizophrenia. Prostaglandins Leukot
Essent Fatty Acids 69, 407–412.
11 Yao JK & Van Kammen DP (2004) Membrane phos-
pholipids and cytokine interaction in schizophrenia. Int
Rev Neurobiol 59, 297–326.

12 Connor JR, Pavlick G, Karli D, Menzies SL & Palmer
C (1995) A histochemical study of iron-positive cells in
the developing rat brain. J Comp Neurol 355, 111–123.
13 Youdim MB, Ben-Shachar D & Riederer P (1991) Iron
in brain function and dysfunction with emphasis on
Parkinson’s disease. Eur Neurol 31, 34–40.
14 Ben-Shachar D, Pinhassi B & Youdim MB (1991) Pre-
vention of neuroleptic-induced dopamine D2 receptor
supersensitivity by chronic iron salt treatment. Eur J
Pharmacol 202, 177–183.
15 Ben-Shachar D & Youdim MB (1991) Intranigral iron
injection induces behavioural and biochemical parkin-
sonism in rats. J Neurochem 57, 2133–2135.
16 Evans PH (1993) Free radicals in brain metabolism and
pathology. Br Med Bull 49, 577–587.
17 Olanow CW, Cohen G, Perl DP & Marsden CD (1992)
Role of iron and oxidation stress in the normal and par-
kinsonian brain. Ann Neurol 32, S1–S145.
18 Strong R, Mattamal M & Andorn A (1993) Free Radi-
cals in Aging. In Free Radicals, the Aging Brain, and
Age-Related Neurodegenerative Disorders. CRC Press,
Boca Raton, FL, USA.
19 Dexter DT, Carayon A, Javoy-Agid F, Agid Y, Wells
FR, Daniel SE, Lees AJ, Jenner P & Marsden CD
(1991) Alterations in the levels of iron, ferritin and
other trace metals in Parkinson’s disease and other
neurodegenerative diseases affecting the basal ganglia.
Brain 114, 1953–1975.
20 Kienzl E, Puchinger L, Jellinger K, Linert W, Stachel-
berger H & Jameson RF (1995) The role of transition

metals in the pathogenesis of Parkinson’s disease.
J Neuro Sci 134 (Suppl.), 69–78.
21 Reichmann H, Janetzky B & Riederer P (1995) Iron-
dependent enzymes in Parkinson’s disease. J Neural
Transm Supplement 46, 157–164.
22 Swaiman KF (1991) Hallervorden–Spatz syndrome and
brain iron metabolism. Arch Neurol 48, 1285–1293.
23 Dexter DT, Holley AE, Flitter WD, Slater TF, Wells
FR, Daniel SE, Lees AJ, Jenner P & MC (1994)
Increased level of lipid hydroperoxides in the parkinso-
nian substantia nigra: an HPLC and ESR study. Mov
Disord 9, 92–97.
24 Dexter DT, Wells FR, Agid F, Agid Y, Lees AJ, Jenner
P & Marsden CD (1987) Increased nigral iron content
in postmortem parkinsonian brain (letters). Lancet 2,
1219–1220.
25 Drayer BP, Olanow W, Burger P, Johanson GA, Herf-
kens R & Riederer S (1986) Parkinson plus syndrome:
diagnosis using high field MR imaging of brain iron.
Radiology 159, 493–498.
26 Rustin P, von Kleist-Retzow JC & Munnich A (1998)
Iron overload and mitochondrial diseases. Lancet 351,
1286–1287.
27 Fredriksson A, Schroder N & Archer T (2003) Neuro-
behavioural deficits following postnatal iron overload.
I. Spontaneous motor activity. Neurotoxicity Res 5,
53–76.
28 Fredriksson A, Schroder N, Eriksson P, Izquierdo I &
Archer T (2001) Neonatal iron potentiates adult MPTP-
induced neurodegenerative and functional deficits.

Parkinsonism Relat Disord 7, 97–105.
29 Fredriksson A, Schroder N, Eriksson P, Izquierdo I &
Archer T (2000) Maze learning and motor activity defi-
cits in adult mice induced by iron exposure during a cri-
tical postnatal period. Brain Res Dev Brain Res 119,
65–74.
30 Fredriksson A, Schroder N, Eriksson P, Izquierdo I &
Archer T (1999) Neonatal iron exposure induces neuro-
behavioural dysfunctions in adult mice. Toxicol Appl
Pharmacol 159, 25–30.
31 Schroder N, Fredriksson A, Vianna M, Roesler R,
Izquierdo I & Archer T (2001) Memory deficits in adult
rats following postnatal iron administration. Behav
Brain Res 124, 77–85.
G. Isaac et al. Brain lipid composition in postnatal iron-induced mice
FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS 2241
32 Singh SP & Shankar R (1996) Effect of haloperidol on
phospholipid biosythesis in rat brain. Indian J Exp Biol
34, 111–114.
33 Iqbal MM, Raham A, Husain Z, Mahmud SZ, Ryan
WR & Feldman JM (2003) Clozapine: a clinical review
of adverse effects and management. Ann Clin Psychiatry
15, 33–48.
34 Schmitt A, Wilczek K, Blennow K, Maras A, Jatzko A,
Petroianu G, Braus DF & Gattaz WF (2004) Altered
thalamic membrane phospholipids in schizophrenias: a
postmortem study. Biol Psychiatry 56, 41–45.
35 Yao JK, Leonard S & Reddy RD (2000) Membrane
phospholipid abnormalities in postmortem brains from
schizophrenic patients. Schizophr Res 42, 7–17.

36 Yao JK, Stanley JA, Reddy RD, Keshavan MS & Pet-
tegrew JW (2002) Correlation between peripheral poly-
unsaturated fatty acid content and in vivo membrane
phospholipid metabolites. Biol Psychiatry 52, 823–830.
37 Schmitt A, Maras A, Petroianu G, Braus DF, Scheuer
L & Gattaz WF (2001) Effect of antipsychotic treatment
on membrane phospholipid metabolism in schizophre-
nia. J Neural Transm General Sect 108, 1081–1091.
38 Fukuzako H, Fukuzako T, Kodama S, Hashiguchi T,
Takigawa M & Fujimoto T (1999) Haloperidol
improves membrane phospholipi abnormalities in tem-
poral lobes of schizophrenic patients. Neuropsychophar-
macology 21, 542–549.
39 Jensen JE, Al-Semaan YM, Williamson PC, Neufeld
RWJ, Menson RS, Schaeffer B, Densmore M & Drost
DJ (2002) Regional-specific changes in phospholipid
metabolism in chronic, medicated shizophrenia: 31P-
MRS study at 4.0 Tesla. Br J Psychiatry 180, 39–44.
40 Komoroski RAMPJ, Griffin WST, Mark RE, Omori M
& Karson CN (2001) Phospholipid abnormalities in
postmortem schizophrenic brains detected by
31
P mag-
netic resonance spectroscopy: a preliminary study. Psy-
chiatry Res Neuroimaging 106, 171–180.
41 Volz H-P, Ro
¨
bger G, Riehemann S, Humber G, Maurer
I, Wenda B, Rzanny R, Kaiser WA & Sauer H (1999)
Increase of phosphodiester during neuroleptic treatment

of schizophrenics: a longitudinal
31
P-magnetic resonance
spectroscopic study. Biol Psychiatry 45, 1221–1225.
42 Hsu FF, Bohrer A & Turk J (1998) Formation of
lithiated adducts of glycerophosphocholine lipids facili-
tates their identification by electrospray ionization tan-
dem mass spectrometry. J Am Soc Mass Spectrom 9,
516–526.
43 Kim HY, Wang TC & Ma YC (1994) Liquid chromato-
graphy ⁄ mass spectrometry of phospholipids using elec-
trospray ionization. Anal Chem 66, 3977–3982.
44 Olsson NU & Salem N Jr (1997) Molecular species
analysis of phospholipids. J Chromatogr B 692, 245–
256.
45 Isaac G, Bylund D, Ma
˚
nsson J-E, Markides KE &
Bergquist J (2003) Analysis of phosphatidycholine and
sphingomyelin molecular species from brain extracts
using capillary liquid chromatography electrospray
ionization mass spectrometry. J Neurosci Meth 128,
111–119.
46 Karlsson AA, Michelsen P & Odham G (1998) Molecu-
lar species of sphingomyelin: determination by high-per-
formance liquid chromatography ⁄ mass spectrometry
with electrospray and high-performance liquid chroma-
tography ⁄ tandem mass spectrometry with atmospheric
pressure chemical ionization. J Mass Spectrom 33,
1192–1198.

47 Berger GE, Wood SJ, Pantelis C, Velakoulis D, Wellard
RM & McGory PD (2001) Implication of lipid biology
for the pathogenesis of schizophrenia. Aus NZ J Psy-
chiatry 35, 355–366.
48 Lagarde M, Geloen A, Record M, Vance D & Spener F
(2003) Lipidomics is emerging. Biochim Biophys Acta
1634, 61–61.
49 Piomelli D (2005) The challenge of brain lipidomics.
Prostaglandins 77, 23–34.
50 Spener F, Lagarde M & Record M (2003) What is lipi-
domic? Eur J Lipid Sci Technol 105, 481–482.
51 Han X & Gross RW (2003) Global analysis of cellular
lipidomes directly from crude extracts of biological sam-
ples by ESI mass spectrometry: a bridge to lipidomics.
J Lipid Res 44, 1071–1079.
52 Taylor EM & ME (1990) Developmental changes in
transferrin and iron uptake by the brain in the rat.
Brain Res Dev Brain Res 55, 35–42.
53 Taylor EM, Crowe A & Morgan EH (1991) Transferrin
and iron uptake by the brain: effects of altered iron sta-
tus. J Neurochem 57, 1584–1592.
54 Dwork AJ, Lawler G, Zybert PA, Durkin M, Osman
M, Willson N & Barkai AI (1990) An autoradiographic
study of the uptake and distribution of iron by the
brain of the young rat. Brain Res 518, 31–39.
55 Davison AN & Dobbing J (1968) Applied Neurochemis-
try. Blackwell, Oxford, UK.
56 Rozenzweig MR, Leiman AL & Breedlove SM (1999)
Biological Psychology: an Introduction to Behavioral,
Cognitive and Clinical Neuroscience. Sinauer Assoc, Inc,

Sunderland, MA, USA.
57 Bolles RG & Woods PJ (1964) The ontogeny of beha-
viour in the albino rat. Anim Behav 12, 427–441.
58 Campbell BA, Lytle LD & Fibiger HC (1969) Ontogeny
of adrenergic arousal and cholinergic inhibitory
mechanisms in the rat. Science 166, 635–637.
59 Ben-Shachar D, Livne E, Spanier I, Zuk R & Youdim
MB (1993) Iron modulates neuroleptic-induced effects
related to the dopaminergic system. Isr J Med Sci 29,
587–592.
60 Ben-Shachar D & Youdim MB (1990) Neuroleptic-
induced supersensitivity and brain iron. I. Iron defi-
ciency and neuroleptic-induced dopamine D2 receptor
supersensitivity. J Neurochem 54, 1136–1141.
Brain lipid composition in postnatal iron-induced mice G. Isaac et al.
2242 FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS
61 Gattaz WF, Schmitt A & Maras A (1995) Increased
platelet phospholipase A2 activity in schizophrenia.
Schizophr Res 16, 1–6.
62 Ross BM, Hudson C, Erlich J, Warsh JJ & Kish SJ
(1997) Increased phospholipid breakdown in schizophre-
nia. Arch General Psychiatry 54, 487–494.
63 Trzeciak HI, Kalacinski W, Malecki A & Kokot D
(1995) Effect of neuroleptics on phospholipase A2 activ-
ity in the brain of rats. Eur Arch Psychiatry Clin Neu-
rosci 245, 179–182.
64 Dalla Libera A, Scutari G, Boscolo R, Rigobello MP
& Bindoli A (1998) Antioxidant properties of cloza-
pine and related neuroleptics. Free Radic Res 29, 151–
157.

65 Sawas AH & Gilbert JC (1985) Lipid peroxidation as
a possible mechanism for the neurotoxic and nephro-
toxic effect of a combination of lithium carbonate
and haloperidol. Arch Int Pharmacody Ther 276, 301–
312.
66 Archer T, Fredriksson A, Jonsson G, Lewander T,
Mohammed A, Ross S & Soderberg U (1986) Central
noradrenaline depletion antagonizes aspects of
d-amphetamine-induced hyperactivity in the rat.
Psychopharmacology 88, 141–146.
G. Isaac et al. Brain lipid composition in postnatal iron-induced mice
FEBS Journal 273 (2006) 2232–2243 ª 2006 The Authors Journal compilation ª 2006 FEBS 2243