mque
a r t i c l e i n f o
Article history:
Received 9 February 2009
Received in revised form 31 March 2009
Accepted 7 April 2009
Available online 13 May 2009
Conclusions: Our findings suggest a role for lipid metabolism and immune response genes in
Journal of Affective Disorders 120 (2010) 24–31
Contents lists available at ScienceDirect
Journal of Af ect
j ourna l homepage: w w.e© 2009 Elsevier B.V. All rights reserved.
1. Introduction
The association between low levels of serumcholesterol and
suicidal behaviour has been reported in numerous studies
utilizing different methodological approaches (Kunugi et al.,
1997; Modai et al., 1994; Alvarez et al., 1999; Zureik et al., 1996;
Ellison andMorrison, 2001; Golomb,1998; Golomb et al., 2000;
Atmaca et al., 2008). These studies have led to a range of ideas
about how serum cholesterol might play a role in influencing
suicide risk, and this has widened the scope of investigations to
examine other aspects of lipid metabolism in relation to
suicidality. Correlations have been found between other
circulating lipids and suicidal behaviour, in particular fatty
acids (Tanskanen et al., 2001; Huan et al., 2004; Sublette et al.,
2006; Garland et al., 2007) and leptin (Atmaca et al., 2002a,b,
2003; Westling et al., 2004)—both of which are also correlated⁎ Corresponding author. McGill Group for Suicide St
HealthUniversity Institute, 6875 LaSalle Blvd., Verdun,Qu
Tel.: +1514 761 6131x2369; fax: +1 514 762 3023.
E-mail address: gustavo.turecki@mcgill.ca (G. Ture
0165-0327/$ – see front matter © 2009 Elsevier B.V.
doi:10.1016/j.jad.2009.04.007depressed suicide completers and lend further support to the relationship between lipidmetabolism
and suicidality.a b s t r a c t
Background: Studies investigating the association between low cholesterol and suicidality have
generated a range of ideas about how cholesterol might play a role in influencing suicide risk,
extending studies to other aspects of lipid metabolism, as well as immune response, in relation
to suicide.
Methods: We performed large-scale microarray gene expression analysis using the Affymetrix
HG-U133 chipset and focused our investigation on the expression profile of genes related to
lipid metabolism and immune response in post-mortem brains from suicide completers and
comparison subjects. We used tissue from three regions of the frontal cortex (Brodmann areas
(BA) 8/9, 11, and 47) from 22 male suicide completers, 15 of whom were diagnosed with major
depressive disorder, and 13 male comparison subjects.
Results: Fatty acid desaturase (FADS1), leptin receptor (LEPR), phosphoinositide-3-kinase (class
2 alpha; PIK3C2A) and stearoyl-CoA desaturase (SCD) were consistently down-regulated in all
three regions of the frontal cortex of depressed suicides compared to comparison subjects, and
were among the genes for which significant correlations were observed between our
microarray and real-time PCR data.
Limitations: Given the absence of a non-suicidal depressed comparison group in this study, it
cannot be ascertained whether the gene expression changes identified are associated with
depression or suicide.Keywords:
Cholesterol
Cytokines
Microarray
Post-mortem
Depression
Suicidea McGill Group for Suicide Studies, Douglas Mental Health University Institute, McGill University, 6875 LaSalle Blvd., Verdun, Québec, Canada H4H 1R3
b Douglas Mental Health University Institute, McGill University, 6875 LaSalle Blvd., Verdun, Québec, Canada H4H 1R3Research report
Altered expression of lipid metabolis
frontal cortex of suicide completers
Aleksandra Lalovic a, Tim Klempan a, Adolfo Seudies, Douglas Menta
ébec, CanadaH4H1R3
cki).
All rights reserved.f
wl
.and immune response genes in the
ira a, Giamal Luheshi b, Gustavo Turecki a,⁎
ive Disorders
l sev ie r.com/ locate / j adwith low levels of serum total cholesterol. Low serum leptin
levels have also been observed in patients with major
depression (Jow et al., 2006; Kraus et al., 2001). Similarly,
many studies have reported a link between fatty acids and
depression (Edwards et al., 1998; Maes et al., 1996, 1999; Peet
et al., 1998).

25A. Lalovic et al. / Journal of Affective Disorders 120 (2010) 24–31The evidence from studies showing altered blood levels of
cholesterol, leptin and fatty acids in suicidal individuals is
mounting, but an explanation for how these substances in the
peripheral circulation might affect behaviour and elevate the
risk of suicide is lacking. One hypothetical means through
which cholesterol might influence behaviour is by altering the
fluidity of brain cell membranes, affecting the function of
receptors (e.g. serotonin receptors) (Engelberg, 1992). Alter-
natively, a deficient supplyof cholesterol in thebrain could limit
synaptic plasticity, and hence, have important behavioural
effects (Pfrieger, 2003). Both of these explanations rely on the
idea that low brain cholesterol levels would also be observed in
suicide. We recently found that violent compared to non-
violent suicide completers had lower cholesterol content in the
frontal cortex (Lalovic et al., 2007b). Fatty acids were subse-
quently also examined in the brains of suicide completers with
and without major depressive disorder (MDD), but no
significant alterations were discovered (Lalovic et al., 2007a).
One hypothesis to explain the putative role of peripheral fatty
acid alterations inmoodandbehaviour is that an increased ratio
of omega-6/omega-3 fatty acids in the blood, in combination
with increased stress levels, could lead to increased production
of pro-inflammatory cytokines, which in turn, has been linked
to depression and suicidal behaviour (Logan, 2003; Maes et al.,
1993, 1995, 1997).
How cholesterol, fatty acids, leptin and other cytokines are
related to one another and to suicidal behaviour is unclear.
Investigating proteins in the brain that may be related to the
transport of lipids into and within the brain, enzymes
involved in lipidmetabolism, or proteins thatmay be involved
in regulating common aspects of lipid metabolism and
immune response might provide additional clues. Given the
enormous number of proteins involved in these processes,
identifying the ones to focus on is a complex task. One
valuable tool that can serve as a starting point in this respect
is microarray technology, which allows for the simultaneous
assessment of the expression levels of thousands of genes.
Lipid metabolism and inflammation genes that are unveiled
as differentially expressed in the brains of suicide completers
can then be investigated for alterations at the protein level
that may be involved in mediating the cholesterol–suicide
connection. Therefore, we used the Affymetrix HG-U133
chipset to investigate the expression of genes related to
lipid metabolism and immune response in brain tissue from
suicide completers with and without MDD and non-suicidal
comparison subjects.
2. Methods
2.1. Subjects
A total of 35 male subjects of French Canadian origin were
included in this study. Brain tissue samples were obtained in
collaborationwith theQuebecCoroner'sOffice from theQuebec
Suicide Brain Bank (QSBB) in Montreal. Cause of death was
ascertained by the coroner's office. Toxicological screening of
body fluid or tissue samples were used to aid in the coroner's
assessment and thus information on alcohol and drug use was
available for some subjects. All deaths were sudden and
without a prolonged agonal state. Macroscopic examination of
brains did not uncover evidence of neurodegeneration or anyother abnormalities in any of the brains. Information on
sociodemographic variables and clinical data, including psy-
chiatric and medical history, was obtained through structured
interviews with one or more informants of the deceased,
supplemented by a review of coroner's notes and medical
records, as previously described (Kim et al., 2003; Lesage et al.,
1994). This information, including Structured Clinical Interview
for DSM-IV Axis I (SCID-I) data, was assessed by a panel of
psychiatrists to obtain best-consensus DSM-IV Axis I diagnoses.
Approval for this study was granted by our local Institutional
Review Board, and written informed consent was obtained
fromeachparticipating family of thedeceasedprior to inclusion
in the study. The sample consisted of 15 suicides with a current
(within the sixmonthsprior to death) diagnosis ofMDD(group
‘D’) and 7 suicide completers without MDD (group ‘S’). The
comparison group (group ‘C’) consisted of 13 psychiatrically
healthy individuals with no history of suicidal behaviour that
died suddenly from causes that had nodirect influence on brain
tissue.
2.2. Brain tissue processing
Whole brains were processed using standard procedures
and brain tissue sections were snap frozen and stored at
−80 °C. The average period between death and brain tissue
collectionwas approximately 25 h. Brain tissue specimens used
for microarray and real-time PCR analyses were dissected from
three regions of the frontal cortex from the left hemisphere of
all subjects, namely the dorsolateral prefrontal cortex (BA 8/9),
orbitofrontal cortex (BA 11) and ventral prefrontal cortex (BA
47). Following from our biochemical studies of lipids in brain
tissue from suicide completers (Lalovic et al., 2007a,b), these
three frontal cortical regions were chosen given the evidence
implicating these areas in suicide (Arango et al., 1997; Turecki
et al., 1999; Mann et al., 2001).
2.3. Microarray analysis
Brain tissue sample processing, RNA extraction, RNA quality
control, and microarray gene expression analyses were per-
formed atGene Logic Inc. (Gaithersburg,MD, USA, http://www.
genelogic.com). RNA was processed on the HG-U133 chipset,
which consists of two GeneChip arrays with approximately
45000 probe sets that represent more than 39000 transcripts
drawn from about 33000 characterized human genes (full list
available at: http://www.affymetrix.com). RNA samples were
assessed for evidence of degradation and integrity, and samples
included in themicroarrayanalyses had aminimumA260/A280
ratio of 1.9, and a 28S/18S ratio of at least 1.6 (2100-Bioanalyzer,
Agilent Technologies, Palo Alto, CA, USA). Analysis of the
GeneChip data was done using Genesis 2.0 software (Gene
Logic Inc., Gaithersburg, MD, USA), Microarray Analysis Suite
Version 5.0 (MAS 5.0), Data Mining Tool 2.0, and Microarray
database software (available at: http://www.affymetrix.com).
All of the transcripts represented on the GeneChip were
normalized and scaled to a signal intensity of 100. RNA integrity
indicatorswere used to filter the sample and ensuremicroarray
quality control. Quality control parameters included: noise
(RawQvalues), present calls for genes across arrays, andβ-actin
and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) 5′/
3′ signal ratios. Samples with β-actin and GAPDH 5′/3′ signal