Assessment of Cytochrome C Oxidase Dysfunction in the

Substantia Nigra/Ventral Tegmental Area in

Schizophrenia
Matthew W. Rice, Kristen L. Smith, Rosalinda C. Roberts, Emma Perez-Costas., Miguel Melendez-Ferro*.
Department of Psychiatry and Behavioral Neurobiology, University of Alabama at Birmingham, Birmingham, Alabama, United States of America

Abstract
Perturbations in metabolism are a well-documented but complex facet of schizophrenia pathology. Optimal cellular
performance requires the proper functioning of the electron transport chain, which is constituted by four enzymes located
within the inner membrane of mitochondria. These enzymes create a proton gradient that is used to power the enzyme ATP
synthase, producing ATP, which is crucial for the maintenance of cellular functioning. Anomalies in a single enzyme of the
electron transport chain are sufficient to cause disruption of cellular metabolism. The last of these complexes is the
cytochrome c oxidase (COX) enzyme, which is composed of thirteen different subunits. COX is a major site for oxidative
phosphorylation, and anomalies in this enzyme are one of the most frequent causes of mitochondrial pathology. The
objective of the present report was to assess if metabolic anomalies linked to COX dysfunction may contribute to substantia
nigra/ventral tegmental area (SN/VTA) pathology in schizophrenia. We tested COX activity in postmortem SN/VTA from
schizophrenia and non-psychiatric controls. We also tested the protein expression of key subunits for the assembly and
activity of the enzyme, and the effect of antipsychotic medication on subunit expression. COX activity was not significantly
different between schizophrenia and non-psychiatric controls. However, we found significant decreases in the expression of
subunits II and IV-I of COX in schizophrenia. Interestingly, these decreases were observed in samples containing the entire
rostro-caudal extent of the SN/VTA, while no significant differences were observed for samples containing only mid-caudal
regions of the SN/VTA. Finally, rats chronically treated with antipsychotic drugs did not show significant changes in COX
subunit expression. These findings suggest that COX subunit expression may be compromised in specific sub-regions of the
SN/VTA (i.e. rostral regions), which may lead to a faulty assembly of the enzyme and a greater vulnerability to metabolic
insult.

Citation: Rice MW, Smith KL, Roberts RC, Perez-Costas E, Melendez-Ferro M (2014) Assessment of Cytochrome C Oxidase Dysfunction in the Substantia Nigra/
Ventral Tegmental Area in Schizophrenia. PLoS ONE 9(6): e100054. doi:10.1371/journal.pone.0100054
Editor: Tim Douglas Aumann, Florey Institute of Neuroscience & Mental Health, Australia
Received November 22, 2013; Accepted May 21, 2014; Published June 18, 2014
Copyright: ß 2014 Rice et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Work was supported by NIH R01 grant MH066123 to MMF, EPC and RCR. The funders had no role in study design, data collection and analysis, decision
to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* Email: [email protected]
. These authors contributed equally to this work.

Introduction lation. The brain is a high energy-demanding organ, which

obtains the majority of its energy from oxidative phosphorylation
Schizophrenia is a devastating mental illness that affects [9–11] and disruptions of this pathway could account for some of
approximately 1% of the world population [1]. Currently, most the metabolic anomalies observed in schizophrenia. As an
studies on schizophrenia concentrate on the study of pathologies of example, decreased concentrations of ATP have been observed
either neuronal circuitry or molecular mechanisms at the cellular in the frontal lobe of schizophrenia subjects [12], which is
and subcellular levels. This includes the assessment of cellular indicative of a deficit in oxidative phosphorylation. The synthesis
metabolism and mitochondrial function. of ATP requires proper functioning of the electron transport chain
One of the first studies that implicated mitochondrial dysfunc- (ETC), which consists of a series of four enzyme complexes located
tion in the pathology of schizophrenia was performed by within the inner membrane of the mitochondria [13,14]. These
Takahashi and Ogushi [2], which revealed reduced aerobic enzymes transfer electrons between electron donors and acceptors,
glycolysis in schizophrenia post-mortem brain tissue. Since then, establishing a proton gradient that is ultimately used to power the
perturbations in metabolism have become a well-documented, if
enzyme ATP synthase [15–18]. Adequate production of ATP is
complex, facet of schizophrenia pathology. This is also supported
crucial for neuronal plasticity, intracellular signaling, calcium
by studies showing changes in mitochondrial density and increased
buffering, and neurotransmission [19–23]. Anomalies in any single
mitochondrial morphological anomalies in several brain regions,
individual complex of the ETC can be sufficient to cause a
including the prefrontal and limbic cortex, the striatum and the
disruption in cellular metabolism [24,25]. However, the activity of
substantia nigra [3–8].
a given complex is not contingent on the proper functioning of the
Some of the most thoroughly studied metabolic anomalies in
other complexes [24,26].
schizophrenia are related to disruptions in oxidative phosphory-

PLOS ONE | www.plosone.org 1 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

In schizophrenia, anomalies have been reported in individual Materials and Methods

components of the ETC, including complexes I, III, and IV
[14,27–34]. Complex IV or cytochrome c oxidase (COX) is the Ethics Statement
terminal enzyme of the ETC, and its role is to catalyze the Human postmortem brain samples used in this study were
oxidation of cytochrome c, transferring electrons to molecular obtained from the Maryland Brain Collection (Maryland Psychi-
oxygen in order to produce a molecule of H2O [35–37]. COX has atric Research Center, University of Maryland School of
been recognized as a major regulation site for oxidative Medicine) with permission from the Maryland Brain Collection
phosphorylation, and anomalies in this enzyme are some of the Steering Committee. The Maryland Brain Collection obtains
most frequent causes of mitochondrial pathology [38–40]). X-ray written consent from the next-of-kin for the use in research of all
crystallography has shown that COX is composed of 13 subunits, human brain samples collected. This brain collection has full
three of which (COX I, II, and III) are encoded by the approval from the Institutional Review Board (IRB) at the
mitochondrial genome, with the remaining subunits (COX IV, University of Maryland School of Medicine (DHMH protocol # 07–
Va, Vb, VIa, VIb, VIc, VIIa, VIIb, VIIc, and VIII) encoded by 23). In addition, all postmortem human brain experimental
the nuclear DNA [41–42]). Mitochondrial DNA-encoded subunits procedures were conducted following a protocol approved by
are commonly known as the ‘‘catalytic core’’ of the enzyme due to the University of Alabama at Birmingham Institutional Review
their role in electron transfer and proton expulsion [43]. Board (protocol # N110505002), and in accordance with the
Mutations within subunits of the catalytic core result in disruptions principles expressed in the Declaration of Helsinki.
of both structure and function of the entire COX enzyme [40,44– No animals were handled, treated or sacrificed for this study.
47]). Among the nuclear DNA-encoded subunits, COX subunit However, brain protein extracts from adult rats from a stock
IV is especially important for the assembly and proper functioning maintained in our laboratory were used. In strict accordance with
of the enzyme [37]. COX subunit IV is the largest of the nuclear the National Institutes of Health Policy on Humane Care and Use
DNA-encoded subunits and may be involved in the modulation of of Laboratory Animals, the samples used in this study are
electron transfer between components of the catalytic core [36– considered cadaveric tissues, and were not prepared or collected
37]. Suppression of subunit IV has been linked to a reduced specifically for this study. This protein stock was the remnant from
function in overall COX activity and an increased susceptibility to a previously published study [58] performed by some of the
apoptosis [36–37]. authors of the present work at the University of Maryland School
of Medicine, and on that previous study, all research was
In schizophrenia COX activity is differently affected depending
conducted following an approved IACUC protocol (IACUC
on the region of the brain studied. Reduced COX activity has
protocol # 0705010). All animal housing, care, and experimental
been reported in the frontal and temporal cortex, as well as in the
procedures were done in strict accordance with National Institutes
caudate nucleus [27,30], while an increase in activity was observed
of Health Guidelines regarding the care and use of animals for
in the putamen [31], and no changes in activity have been
experimental procedures.
reported in brain tissue from the nucleus accumbens, globus
pallidus, thalamus, cerebellum, or mesencephalon [27,31]. On the
other hand, Whatley et al. [48] did not find changes in COX Postmortem Human Brain Samples
activity within the frontal cortex, adding to the complexity of the All human brain samples were obtained from the Maryland
pathology of COX within the schizophrenia brain. Although some Brain Collection (Maryland Psychiatric Research Center, Univer-
studies have reported increases [26,31] and decreases [49] in sity of Maryland School of Medicine). For all postmortem human
COX activity due to the use of antipsychotic medication, the cases, two independent psychiatrists established DSM-IV diagno-
majority of studies have concluded that COX activity is not ses based on the review of available medical and autopsy records,
and data obtained from the Structured Clinical Interview for
affected by antipsychotic drugs [50–54]. Taken together, these
DSM-IIIR [59] and DSM-IV Axis I disorders with the next-of-kin.
findings suggest that COX anomalies are likely an intrinsic feature
Table 1 contains demographic data, as well as information on
of schizophrenia.
antipsychotic use, cause of death, and other relevant parameters
The substantia nigra/ventral tegmental area complex (SN/
for the samples used in this study.
VTA) contains the largest group of dopaminergic neurons in the
Human brain samples were dissected in two different ways: 1)
brain, and is the origin for several major dopamine pathways [55–
samples that contained the entire rostro-caudal extent of the SN/
56]. Dopaminergic anomalies are a well-known component of
VTA (n = 15; Table 1) or 2) samples that contained only the mid-
schizophrenia pathology [57], and previous findings from our
caudal regions of the SN/VTA (n = 14; Table 1). Cases from these
laboratory have shown significant reductions in tyrosine hydrox-
two different dissections were analyzed separately due to the
ylase (TH) protein levels within rostral regions of the SN/VTA in
heterogeneous nature of the subnuclei within the SN/VTA. All
schizophrenia [58].
samples were frozen in dry ice and stored at 280uC prior to
The aim of this study was to assess if metabolic anomalies linked sectioning. Using a cryostat, five parallel (i.e. adjacent) series of
to COX dysfunction might contribute to SN/VTA pathology in 16 mm thick sections were obtained in the coronal plane following
schizophrenia. To accomplish this we tested for differences in the longitudinal axis of the brainstem. Sample blocks were
COX activity within the SN/VTA in human postmortem samples trimmed to reduce the presence of cell groups not related to our
from schizophrenia subjects compared to non-psychiatric controls. study. All samples were sectioned in their entirety. Four series were
In addition, we tested the hypothesis that alterations in the subunit collected on charged slides, while the fifth series was collected in a
composition of COX could contribute to schizophrenia pathology vial for protein extraction. Series #1 was stained with thionin to
in the SN/VTA. To test this we analyzed protein expression for all assess proper morphological preservation, and only those cases
the COX subunits that constitute the catalytic core, as well as for that presented good preservation of the SN/VTA were included in
subunit IV, which is critical for the assembly and function of the the study. Exclusion criteria included the presence of extensive
enzyme. autolysis of cells, cell shrinkage and poor morphological preser-
vation. These criteria were equally applied to select the
schizophrenia and control cases included in the study. Rostral

PLOS ONE | www.plosone.org 2 June 2014 | Volume 9 | Issue 6 | e100054

 Table 1. Demographic and clinical data of the cases used for the study.

Schizophrenia: rostro-caudal samples (n = 8). All cases used for both COX activity and western blots of subunits

Case # Age (years) race gender COD PMI (hours) pH SZ subtype APD type TD

SZ1 25 W M Suicide (hanging) 20 5.90 Chronic paranoid Risperidone (A) Yes

SZ2 53 W M ASCVD 14 6.01 Undifferentiated Quetiapine (A) Yes
SZ3 48 AA M Narcotic intoxication 24 7.07 NOS Quetiapine (A) un

PLOS ONE | www.plosone.org

SZ4 54 W M Suicide (bleeding) 19 6.58 NOS Quetiapine (A) No
SZ5 83 W M Electrocution 18 6.98 NOS Quetiapine (A) un
SZ6 77 W M ASCVD 17 6.84 CUT Risperidone (A) un
SZ7 46 W F DVT 25 6.18 NOS Clozapine (A) un
SZ8 42 AA F ASCVD/MI 13 5.86 Chronic paranoid Fluphenazine+Quetiapine (T+ A) Yes
Mean 53.50 18.75 6.43
SD 18.72 4.27 0.50

Schizophrenia: mid-caudal samples (n = 8). *Indicates case used only for COX activity

Case # Age (years) race gender COD PMI (hours) pH SZ subtype APD type TD

SZ9 48 W F Respiratory arrest 21 7.80 CUT Clozapine (A) No

3
SZ10 42 AA M Suicide (stabbing) 10 7.10 Chronic paranoid Haloperidol (T) Yes
SZ11 77 AA F ASCVD 17 6.20 NOS UNK un
SZ12* 67 AA M ASCVD 7 7.17 CUT UNK un
SZ13 37 W M Suicide (electrocution) 14 6.68 Chronic paranoid Risperidone+Fluphenazine (T+A) No
SZ14 42 AA M Pulmonary embolism 16 7.10 Chronic paranoid Haloperidol+Chlorpromazine (T) Yes
SZ15 49 AA F ASCVD 23 7.20 NOS Haloperidol (T) Yes
SZ16 69 AA M ASCVD 14 6.71 CUT Haloperidol (T) No
Mean 53.88 15.25 7.00
SD 14.93 5.28 0.47

Controls: rostro-caudal samples (n = 7). All cases used for both COX activity and western blots of subunits

Case # Age (years) race gender COD PMI (hours) pH

C1 33 AA M Drowning 21 7.20
C2 49 AA F Asthma 19 6.19
C3 51 AA M ASCVD 19 7.18
C4 45 W F Cardiac arrhythmia 16 6.69
C5 77 W M ASCVD 16 6.80
C6 62 W M Cardiac arrest 19 6.48
C7 67 AA M ASCVD 17 6.68
Cytochrome C Oxidase in Schizophrenia

June 2014 | Volume 9 | Issue 6 | e100054

 Table 1. Cont.

Schizophrenia: rostro-caudal samples (n = 8). All cases used for both COX activity and western blots of subunits

Case # Age (years) race gender COD PMI (hours) pH SZ subtype APD type TD

Mean 54.86 18.14 6.75

SD 14.79 1.86 0.36

PLOS ONE | www.plosone.org

Controls: mid-caudal samples (n = 6). *Indicates case used only for western blots of subunits

Case # Age (years) race gender COD PMI (hours) pH

C8 35 W M MVA 4 6.70
C9 72 W F MVA 17 6.83
C10 45 W F Pulmonary embolism 12 5.74
C11 41 AA F Cardiac disease 21 6.73
C12 48 W M Aneurysm 19 6.51
C13* 28 W M Gunshot wound 7 6.82
Mean 44.83 13.33 6.56
SD 15.12 6.83 0.42

Abbreviations: A, atypical antipsychotic; AA, African American; APD, antipsychotic drug; ASCVD, arteriosclerotic cardiovascular disease; C, control case; COD, cause of death; CUT, chronic undifferentiated type; DVT, deep vein

4
thrombosis; F, female; M, male; MI, myocardial infarction; NOS, not otherwise specified; MVA, motor vehicle accident; PMI, postmortem interval; SD, standard deviation; SZ, schizophrenia case; SZ subtype, schizophrenia subtype; T,
typical antipsychotic; TD, tardive dyskinesia; un, unknown; UNK, unknown medication at time of death; W, white.
doi:10.1371/journal.pone.0100054.t001
Cytochrome C Oxidase in Schizophrenia

June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

Table 2. Note that experiments were run independently for each outcome measure, thus, independent multiple regression tests
were run for each outcome measure.

Table 2. Effect of demographic and sample variables

ROSTRO-CAUDAL SN/VTA SAMPLES MID-CAUDAL SN/VTA SAMPLES

Effect of Demographic
Effect of Demographic Effect of Sample variables (age, gender, Effect of Sample
OUTCOME MEASURE variables (age, gender, race) variables (pH, PMI) race) variables (pH, PMI)

COX activity p = 0.1040 p = 0.9769 p = 0.0982 p = 0.5743

R2a = 0.2566 R2a = 20.1621 R2a = 0.3148 R2a = 20.0740
COX subunit I p = 0.1746 p = 0.8477 p = 0.0925 p = 0.5974
R2a = 0.1744 R2a = 20.1350 R2a = 0.3247 R2a = 20.0825
COX subunit II p = 0.1528 p = 0.3665 p = 0.7831 p = 0.9402
R2a = 0.1967 R2a = 0.0131 R2a = 20.1893 R2a = 20.1853
COX subunit III p = 0.2365 p = 0.0894 p = 0.4934 p = 0.8659
R2a = 1.208 R2a = 0.2198 R2a = 20.0347 R2a = 20.1659
COX subunit IV-I p = 0.4938 p = 0.6648 p = 0.4801 p = 0.7671
R2a = 20.0326 R2a = 20.0899 R2a = 20.0268 R2a = 20.1380
COX subunit IV-II p = 0.5753 p = 0.6191 p = 0.5272 p = 0.6193
R2a = 20.0704 R2a = 20.0771 R2a = 20.0542 R2a = 20.0903

Analyses were carried independently for rostro-caudal and mid-caudal samples. Demographic and sample variables were tested in separated tests. Abbreviations: R2a,
adjusted R squared. pH, brain pH. PMI, postmortem interval.
doi:10.1371/journal.pone.0100054.t002

and caudal boundaries for samples containing the entire extent of the COX incubation medium, which consisted in a solution of
the SN/VTA were determined using the Paxinos and Huang [60] 22.4 mg cytochrome c (Sigma-Aldrich, St. Louis, MO, USA;
atlas of the human brainstem. Mid-Caudal SN/VTA samples C3131), 115.23 mg diaminobenzidine (DAB, Sigma-Aldrich,
were defined as those transversally dissected caudal to the D5637), 4.5 g sucrose (Sigma-Aldrich, S0389), 12.51 ml of a 1%
transition between the red nucleus parvocellularis and magno- nickel ammonium sulfate solution (Sigma-Aldrich, A1827), all
cellularis. mixed in 100 ml of 0.1M Hepes buffer (Sigma-Aldrich, H3375)
pH 7.4 [62,63]. To terminate the COX enzymatic reaction,
Animal Brain Protein Samples sections were immersed in 4% paraformaldehyde in 0.1M
A stock of protein extracts maintained in the laboratory, which phosphate buffer (PB) pH 7.4 for one hour at room temperature.
was obtained from rats previously treated with antipsychotics [58], Sections were then rinsed in PB, dehydrated in ethanol, cleared in
was used to assess the effects of antipsychotic medication on COX xylene, and coverslipped using Eukitt (Electron Microscopy
subunit protein expression. Treatment procedures for these Science, Hatfield, PA, USA; 15322). Negative controls were done
animals have been previously described in detail in [58]. Briefly, in adjacent tissue sections by prior immersion in a solution of 4%
animals were randomly assigned to one of three treatment groups paraformaldehyde in 0.1M PB pH 7.4 for one hour at room
(n = 9 per group) consisting of haloperidol (1.5 mg/kg/day), temperature, followed by rinses in PB, and then immersion in a
olanzapine (6 mg/kg/day), or control (i.e. water). Antipsychotics solution of 10 mM sodium azide in PB for 17 hours at room
were administered for three weeks in drinking water, adjusting the temperature. The aim of the negative controls is to subtract from
doses to body weight and daily water consumption as described in the analysis the presence of any background not specifically due to
Perez-Costas et al. [61]. Animals were euthanized and brains were COX activity (e.g. possible random precipitation of DAB or nickel
immediately removed from the skull, frozen in dry ice, and stored ammonium sulfate). Thus, the prior treatment with paraformal-
at 280uC. Brains were sectioned on a cryostat at 220uC in the dehyde and sodium azide was used to irreversibly inactivate COX
coronal plane, and 16 mm thick sections through the SN/VTA in the negative control sections. After inactivating the COX
were collected in five parallel series. Sample blocks were trimmed enzyme in the negative control, these sections were incubated
to reduce the presence of cell groups not related to our study. Four simultaneously with the samples in the COX incubation medium.
series were collected on charged slides, while the fifth series was In order to more accurately assess the differences in COX
collected in a vial for protein extraction. Series #1 was stained activity among our groups, we applied a novel methodology
with thionin to assess morphology and general quality of the tissue developed in our laboratory [64]. Briefly, cytochrome c oxidase
collected. No morphological anomalies were found in any of the from bovine heart (Sigma-Aldrich, C5499) was diluted in Tris-
samples used in this study. Buffered-Saline (TBS) to create a series of standards that contained
a known quantity of pure COX. Each of these known standards
Cytochrome C Oxidase Histochemistry in Human SN/VTA was loaded onto a nitrocellulose membrane, using vacuum and a
Slides containing human SN/VTA sections were removed from slot-blot microfiltration apparatus (Bio-Rad, Hercules, CA, USA;
280uC storage and fan-dried for 25 minutes at room temperature. 170-6542). For all COX histochemical assays, negative controls
Sections were then incubated in the dark for two hours at 37uC in

PLOS ONE | www.plosone.org 5 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

Figure 1. Analysis of COX activity in samples containing the entire rostro-caudal extent of the human SN/VTA. A) Representative
images of thionin-stained sections and COX histochemistry in the SN/VTA and red nucleus from schizophrenia and control cases. Dashed lines in the
thionin-stained sections show the masks used for the delineation of the region of interest. B) Top panel shows the logarithm of optical density (log
OD) for COX activity in the SN/VTA (geometric mean with bar representing the 95% confidence interval). The mid panel shows a scatter plot of the
individual data. The horizontal bar in each scatter plot indicates the geometric mean OD. The bottom panel shows COX activity in the red nucleus.
doi:10.1371/journal.pone.0100054.g001

and COX standards were incubated together with sections to be for 15 minutes at 4uC, the supernatant was collected, and total
analyzed for COX activity. protein concentration was measured using a modified Lowry
technique (Bio-Rad, DC Protein Assay; 500-0113 and 500-0114).
Protein Extracts for COX Subunit Analysis in Human and Known concentrations of bovine serum albumin were used as
Rat SN/VTA standards. Aliquots of both rat and human SN/VTA protein
Human and rat brain protein extracts were obtained by extracts were stored at 280uC until use.
sonication of tissue in a cold lysis buffer mixture (1:5 w:v)
containing 50 mM Tris pH 8.0, 5 mM EDTA, 150 mM NaCl, Gel Electrophoresis and Western-blot
1% SDS, and 10 ml/ml of a protease inhibitor cocktail (Sigma- Western-blot assays were used to test for differences in COX
Aldrich; P8340). Homogenates were then centrifuged at 15000 g subunit protein expression in both animal and human SN/VTA

PLOS ONE | www.plosone.org 6 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

Figure 2. Analysis COX activity in samples containing only the mid-caudal regions of the human SN/VTA. A) Representative images of
thionin-stained sections and COX histochemistry in the SN/VTA and red nucleus from schizophrenia and control cases. Dashed lines in the thionin-
stained sections show the masks used for the delineation of the region of interest. B) Top panel shows the logarithm of optical density (log OD) for
COX activity in the SN/VTA (geometric mean with bar representing the 95% confidence interval). The mid panel shows a scatter plot of the individual
data. The horizontal bar in each scatter plot indicates the geometric mean OD. The bottom panel shows COX activity in the red nucleus.
doi:10.1371/journal.pone.0100054.g002

samples. For gel electrophoresis, samples were diluted 1:1 in (TBS-T) for one hour at room temperature. Membranes were then
loading buffer (62.5 mM Tris-HCl pH 6.8, 25% glycerol, 2% incubated with the appropriate primary antibody at the concen-
SDS, 0.01% bromophenol blue, and 5% b-mercaptoethanol) and trations listed below, in a solution of 1% non-fat dry milk in TBS-
heated to 95uC in a dry bath for 5 minutes before loading on a 4– T for 19 hours at 4uC. Following that, membranes were rinsed in a
20% polyacrylamide gradient gel (Lonza, Basel, Switzerland; solution of 1% non-fat dry milk in TBS-T prior to incubation with
58505). For rat samples, 25 mg of SN/VTA total protein extract the appropriate secondary antibody at the concentrations listed
were loaded in each lane, while for human samples 60 mg of SN/ below, in a solution of 1% non-fat dry milk in TBS-T for one hour
VTA total protein extract were loaded. In addition, a molecular at room temperature. After incubation with the secondary
weight marker (Lonza; 50550) was loaded for molecular weight antibody, membranes were rinsed first with TBS-T and then with
reference. Proteins were resolved using a constant 150V current TBS. Membranes were developed using an Immun-Star alkaline
and transferred onto polyvinylidene fluoride (PVDF) membranes phosphatase substrate chemiluminescence kit (Bio-Rad; 170-
(Bio-Rad; 162-0714) under a constant current of 30V for 21 hours 5018), exposing Kodak Biomax XAR films (Kodak, Rochester,
at 4uC. NY, USA; 166-0760).
For western-blot assays, membranes were initially rinsed in For each subunit tested a minimum of two separate western-blot
0.05M TBS, followed by blocking in a solution of 5% non-fat dry experiments were carried out. For human SN/VTA, rostro-caudal
milk (Bio-Rad; 170-6404) in TBS containing 0.1% Tween 20 and mid-caudal samples were run in the same experiment, which

PLOS ONE | www.plosone.org 7 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

Figure 3. Analysis of COX subunits I, II, and III in samples containing the entire rostro-caudal extent of the human SN/VTA. For each
subunit, western blot images of all the samples assessed in schizophrenia and control cases are shown at the top of each panel, together with an
image of the same western blot membrane reincubated for the detection of actin. Outliers are indicated with a red box surrounding the case number.
Bar graphs indicate the mean calibrated optical density (OD) and standard deviation. Scatter plots are shown on the right with a horizontal bar
indicating the mean. A) Analysis of COX Subunit I expression. Subunit I protein expression did not differ significantly between schizophrenia and
control cases (OD mean 6 SD: schizophrenia: 0.618260.07451; control: 0.601460.08532). B) Analysis of COX Subunit II expression. There were
significant (*) differences for subunit II expression between schizophrenia and controls (OD mean 6 SD: schizophrenia: 0.2171960.02887; control:
0.475160.1954). C) Analysis of COX Subunit III expression. Subunit III protein expression did not differ significantly between schizophrenia and
control cases (OD mean 6 SD: schizophrenia: 0.815360.3878; control: 0.538860.1281).
doi:10.1371/journal.pone.0100054.g003

required the simultaneous incubation and development of 3 Antibodies

western-blot membranes per experiment. Three western-blot The following primary antibodies were used for western-blot
membranes were also necessary to accommodate all rat samples assays: mouse monoclonal anti-COX subunit I (MitoSciences,
in a single experiment. Eugene, OR, USA; ab14705, diluted 1:1000 for rat samples),
rabbit polyclonal anti-COX subunit I (Abcam, Cambridge,

PLOS ONE | www.plosone.org 8 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

Figure 4. Analysis of COX subunits IV-I and IV-II in samples containing the entire rostro-caudal extent of the human SN/VTA. For
each subunit, western blot images of all the samples assessed in schizophrenia and control cases are shown at the top of each panel, together with
an image of the same western blot membrane reincubated for the detection of actin. An outlier is indicated with a red box surrounding the case
number. Bar graphs indicate the mean calibrated optical density (OD) and standard deviation. Scatter plots are shown on the right with a horizontal
bar indicating the mean. A) Analysis of COX Subunit IV-I expression. Subunit IV-I protein expression showed significant (*) differences between the
two groups (OD mean 6 SD: schizophrenia: 0.4354160.20501; control: 0.8441660.56772). B) Analysis of COX Subunit IV-II protein expression.
Subunit IV-II protein expression did not differ significantly between schizophrenia and control cases (OD mean 6 SD: schizophrenia: 1.34060.2261;
control: 1.35460.3104).
doi:10.1371/journal.pone.0100054.g004

England; ab90668, diluted 1:1000 for human samples), mouse 1:1000), and alkaline phosphatase-conjugated donkey anti-goat
monoclonal anti-COX subunit II (MitoSciences; ab110258, (Millipore, AP180A, diluted 1:15,000).
diluted 1:1000 for human samples), goat polyclonal anti-COX
subunit II (Santa Cruz Biotechnology, Dallas, TX, USA; sc-23984, Image Acquisition and Analysis
diluted 1:1000 for rat samples), rabbit polyclonal anti-COX Complete series of slides containing SN/VTA sections that were
subunit III (Mitosciences; ab138956, diluted 1:2000 for rat assessed using COX histochemistry, membranes containing COX
samples and 1:1000 for human samples), rabbit polyclonal anti- standards, and western-blot films, were digitized using a flatbed
COX subunit IV-I (Sigma-Aldrich; HPA002485, diluted 1:2000 scanner (Epson Perfection V750-M Pro; Seiko Epson Suwa,
for rat and human samples), rabbit polyclonal anti-COX subunit Japan), and saved at a resolution of 600 dpi as uncompressed
IV-II (Sigma-Aldrich; SAB4503384, diluted 1:1000 for rat TIFF files. Images of sections assessed by COX histochemistry
samples), and rabbit polyclonal anti-COX subunit IV-II (Sigma- were acquired as 16 bit grayscale, while images of western blot
Aldrich, HPA029307, diluted 1:2000 for human samples). For all films were obtained as 8 bit grayscale. NIH Image J software
western-blot experiments membranes were reblotted for actin (version 1.46 r; http://rsbweb.nih.gov/ij/) was used to measure
using a mouse monoclonal anti-actin antibody (Millipore, Billerica, optical density (OD) for all experiments.
MA, USA; MAB1501, diluted 1:40,000 for both rat and human For COX activity measurements, masks defining the region of
samples). All primary antibodies used in the study were previously interest were created by delineating the SN/VTA on the
tested for specificity by the manufacturers and produced consistent corresponding scanned Nissl-stained section within each case. A
bands at the expected molecular weight. standard curve was created using OD values from the COX
Secondary antibodies used included alkaline phosphatase- standards [64]. In addition, as an internal control for the validity
conjugated goat anti-mouse (Millipore; AP124A, diluted of the assay, COX activity was also measured in the red nucleus
1:15,000), alkaline phosphatase-conjugated goat anti-rabbit (Vec- within the same section, since no studies have reported changes in
tor Laboratories, Burlingame, CA, USA; AP-1000, diluted COX activity for the red nucleus in schizophrenia. Using a

PLOS ONE | www.plosone.org 9 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

Figure 5. Analysis of COX subunits I, II, and III protein expression in the mid-caudal human SN/VTA. For each subunit, western blot
images of all the samples assessed in schizophrenia and control cases are shown at the top of each panel, together with an image of the same
western blot membrane reincubated for the detection of actin. Outliers are indicated with a red box surrounding the case number. Bar graphs
indicate the mean calibrated optical density (OD) and standard deviation. Scatter plots are shown on the right with a horizontal bar indicating the
mean. A) Analysis of COX Subunit I. Subunit I protein expression did not differ significantly between schizophrenia and control cases (OD mean 6 SD:
schizophrenia: 0.635560.1680; control: 0.599860.1213). B) Analysis of COX Subunit II. Subunit II protein expression did not differ significantly
between schizophrenia and control cases (OD mean 6 SD: schizophrenia: 0.933260.3837; control: 1.35260.7247). C) Analysis of COX Subunit III.
Subunit III protein expression did not differ significantly between schizophrenia and control cases (OD mean 6 SD: schizophrenia: 0.522860.08916;
control: 0.840860.5324).
doi:10.1371/journal.pone.0100054.g005

random number generator (random.org), twelve slides (for rostro- placed within the mask delineating the region of interest (i.e. SN/
caudal cases) or eight slides (for mid-caudal cases) representing VTA or red nucleus).
different sub-regions of the SN/VTA were randomly chosen for For western-blot analysis of human and rat samples, an OD
analysis. For each section measured, using Image J rectangle tool, standard curve was created using a step calibration tablet of known
five non-overlapping boxes of 0.1960.19 cm were randomly calibrated OD (Stouffer Industries Inc., Mishawaka, IN, USA;

PLOS ONE | www.plosone.org 10 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

Figure 6. Analysis of COX subunits IV-I and IV-II protein expression in the mid-caudal human SN/VTA. For each subunit, western blot
images of all the samples assessed in schizophrenia and control cases are shown at the top of each panel, together with an image of the same
western blot membrane reincubated for the detection of actin. Bar graphs indicate the mean calibrated optical density (OD) and standard deviation.
Scatter plots are shown on the right with a horizontal bar indicating the mean. A) Analysis of COX Subunit IV-I. Subunit IV-I protein expression did not
differ significantly between schizophrenia and control cases (OD mean 6 SD: schizophrenia: 1.33660.4237; control: 1.78960.4741). B) Analysis of COX
Subunit IV-II. Subunit IV-II protein expression did not differ significantly between schizophrenia and control cases (OD mean 6 SD: schizophrenia:
1.27960.3895; control: 1.09760.1950).
doi:10.1371/journal.pone.0100054.g006

T2120, series #130501). After calibration, OD measurements normalized over actin by dividing the observed ‘‘raw’’ calibrated
were obtained using Image J rectangle tool to delineate each band. OD value of each measure by its respective standardized actin
In all cases, measurements of digitized images were performed calibrated OD value.
by two different researchers blind to diagnosis or treatment group. Human postmortem samples containing the entire rostro-caudal
Each researcher performed three measurements for each sample extent of the SN/VTA and those containing only mid-caudal
assessed. Measures obtained from the same experiment were regions were analyzed separately. In both cases, outcome measures
averaged for analysis. obtained from schizophrenia and non-psychiatric control samples
were assessed using unpaired Student’s t-test or Mann-Whitney
Statistical Analysis (non-parametric) tests after assessing normality and the possible
Schizophrenia and non-psychiatric control cases were matched lognormal distribution of the data. Since COX activity has been
for demographic variables (age, gender, race), and sample previously shown to best fit an exponential curve [64], all COX
variables (postmortem interval, brain pH). We used multiple activity data were transformed to their common logarithm prior to
regression to assess the possible influence of demographic and analysis.
sample variables on our outcome measures (i.e. COX activity and For COX subunit expression, the data were assessed using the
subunits expression) independent of disease status. Demographic following scheme: Data sets were assessed for normality, and when
samples presented significant deviations from normality, the data
variables (i.e. age, gender, race), and sample variables (i.e.
were transformed to their common logarithms in order to test if
postmortem interval and pH) where tested in separated multiple
they were a good fit for a lognormal distribution. If the lack of
regression analyses. For the analysis of demographic variables a
normality persisted after logarithmic transformation, the original
dummy coding was used to numerically interpret gender and race.
(untransformed) data were tested for the presence of outliers and
For protein measures, actin values were standardized by dividing
used for analysis. One-way ANOVA was used to assess the
the observed actin value of each case by the average actin value of
outcome measures on protein samples from animals treated with
each individual film. COX subunit protein measures were then
antipsychotic medications. Assumptions of the parametric tests

PLOS ONE | www.plosone.org 11 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

Figure 7. Analysis of COX subunits I, II, and III in the SN/VTA of animals treated with antipsychotic drugs. Western blot images of the
expression of each COX subunit in all animals from the three treatment groups are shown at the top of each panel. Each western blot membrane was
reincubated for the detection of actin. Bar graphs indicate the mean calibrated optical density (OD) and standard deviation. Scatter plots are shown
on the right with a horizontal bar indicating the mean. A) Analysis of COX subunit I. Subunit I protein expression did not differ significantly among the
three treatment groups (OD mean 6 SD: control: 0.475360.06210; haloperidol: 0.514560.1158; olanzapine: 0.473160.1025). B) Analysis of COX
subunit II. Subunit II protein expression did not differ significantly among the three treatment groups (OD mean 6 SD: control: 0.474060.1118;
haloperidol: 0.553560.1312; olanzapine: 0.556460.1121). C) Analysis of COX subunit III. Subunit III protein expression did not differ significantly
among the three treatment groups (OD mean 6 SD: control: 1.48060.4543; haloperidol: 1.60960.2989; olanzapine: 1.45560.4438).
doi:10.1371/journal.pone.0100054.g007

were addressed in all analyses, including normal distribution of the (ROUT) test for outlier detection [65], with the coefficient Q set to
residuals in the dependent variable, homogeneity of variance among 1%. When detected, outliers were removed from the analysis and
the groups, and independence of errors. The presence of outliers graphical representations, but are still present in the western-blot
was determined using the ‘‘Robust regression and OUtlier Removal’’ images. Shapiro-Wilk and Kolmogorov-Smirnov normality tests

PLOS ONE | www.plosone.org 12 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

Figure 8. Analysis of cytochrome c oxidase subunits IV-I and IV-II in the SN/VTA of animals treated with antipsychotic drugs.
Western blot images of the expression of each COX subunit in all animals from the three treatment groups are shown at the top of each panel. Each
western blot membrane was reincubated for the detection of actin. The bar graphs indicate the mean calibrated optical density (OD) and standard
deviation. Scatter plots are shown on the right with a horizontal bar indicating the mean. A) Analysis of COX subunit IV-I. Subunit IV-I protein
expression did not differ significantly among the three treatment groups (OD mean 6 SD: control: 1.08460.2609; haloperidol: 1.02160.3270;
olanzapine: 1.09860.3559). B) Analysis of COX subunit IV-II. Subunit IV-II protein expression did not differ significantly among the three treatment
groups (OD mean 6 SD: control: 0.682960.1375; haloperidol: 0.810060.1611; olanzapine: 0.755160.08140).
doi:10.1371/journal.pone.0100054.g008

were used to assess normality in all data sets. Significant deviations multiple corrections approach, the COX subunits that form the
from normality, and the presence of outliers, when present, are catalytic core (i.e. subunits I, II, III) were analyzed together.
reported in the results section. Subunits IV-I and IV-II were also analyzed using this correction.
The experimental design was done considering each subunit Subunits I, II and III constitute the catalytic core of the enzyme,
outcome independent (i.e. not corrected for multiple comparisons). are encoded by the mitochondrial genome, and their transcription
It has been shown that specific subunits of the COX enzyme can and translation occur within the mitochondria, while subunits IV-I
be independently regulated and targeted by specific pharmaco- and IV-II are encoded by the nuclear genome and their synthesis
logical treatments [70–74]. For example, mutations in subunit I occurs in the cytoplasm [43,68–69]. Thus, mitochondrial and
[74] as well as specific decreases in the expression of subunit II nuclear subunits were analyzed separately.
[72] or subunit III [73] affect the activity of the complex without Mean and standard deviation data are reported in the figure
altering significantly the expression of the other subunits of the legends. Statistical analysis and graphical representation of the
complex [70–74]. On the other hand, it has been shown that results obtained were done using GraphPad Prism 6.0 software (La
multiple comparison corrections reduce the risk of type I errors, Jolla, CA, USA).
but greatly increase the risk of type II errors, that is, not detecting
relevant differences due to the reduction of power produced by the Results and Discussion
correction [66,67]. However, due to the lack of agreement in the Anomalies in metabolism are a well-established aspect of
field of statistics on the use (or not) of corrections for multiple schizophrenia pathology. Reported anomalies include decreased
comparisons [67] data were also analyzed using multiple t-test overall metabolic activity in the cortex, changes in mitochondrial
correction (Holm-Sidak correction with alpha 0.05). Thus, results density within the caudate nucleus and putamen, and reductions
using independent t-test (i.e. uncorrected) as well as corrected for in ATP concentration within the frontal lobe [3–7,12]. Despite the
multiple comparisons, are both reported (see results). For the potential role of metabolic anomalies in the pathology of the

PLOS ONE | www.plosone.org 13 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

dopamine system in schizophrenia, few studies have addressed the In addition, analysis of COX activity just for the rostral region
metabolic status of the SN/VTA in this disorder. Interestingly, of the SN/VTA in these same cases also yielded non-significant
there have been reports of mitochondrial hyperplasia within differences in activity (unpaired t-test p = 0.5816, t = 0.5652,
presynaptic neurons that synapse onto dopamine neurons in the df = 13). For this sub-analysis of the rostral region, COX activity
substantia nigra compacta [8]. In the present study we assessed the was measured at 4 different rostro-caudal levels within the rostral
metabolic health of the SN/VTA complex by quantifying COX region. The geometric mean for COX activity in the rostral SN/
activity, and by measuring protein expression for key subunits of VTA for the schizophrenia group was 1.435 (95% CI: 1.259,
the COX enzyme. Activity measures provided a ‘‘snapshot’’ of the 1.635) and for the control group was 1.340 (95% CI: 1.060, 1.695).
functional status of the SN/VTA, while assessing the protein COX activity in Mid-Caudal samples. In samples con-
expression of key subunits of the COX enzyme allowed us to infer taining only mid-caudal regions of the SN/VTA unpaired t-test
information about the subunit composition of the enzyme in the revealed non-significant differences between schizophrenia and
SN/VTA in schizophrenia. control cases (p = 0.8737, t = 0.1624, df = 11) [Figure 2]. The
geometric mean for the schizophrenia group was 1.083 (95% CI:
Demographic and Sample Features 0.9102, 1.290) and for the control group was 1.076 (95% CI:
Postmortem human samples selected for this study were 0.9021, 1.285). For the red nucleus, unpaired t-test revealed non-
carefully matched between schizophrenia and non-psychiatric significant differences for COX activity (p = 0.5145, t = 0.7011,
control groups for 1) demographic (age, gender, race) and 2) df = 5). The geometric mean for the red nucleus for the
sample variables (brain pH and PMI) (see Table 1). In addition, schizophrenia group was 1.252 (95% CI: 0.8047, 1.947) and for
multiple regression was used to assess if 1) demographic or 2) the control group was 1.127 (95% CI: 0.8010, 1.585) [Figure 2].
sample variables, had a significant contribution to our dependent Anomalies of COX activity in schizophrenia have been studied
measures. The effect of these demographic and sample variables in various regions of the brain including frontal and temporal
was tested independently for COX activity, as well as for the cortex, caudate nucleus, putamen, nucleus accumbens, globus
expression of each COX subunit assessed in our study (i.e. COX pallidus, thalamus, and cerebellum [27,30,31,48]. These studies
subunits I, II, III, IV-1, IV-II). In all cases, multiple regression revealed that COX activity is differently affected depending on the
analysis showed that these variables did not have significant region examined. In the present study we tested for differences in
COX activity in two different types of samples from the SN/VTA
predictive value. Table 2 summarizes the results obtained from
(i.e. rostro-caudal and mid-caudal). Our analysis of COX activity
the multiple regression tests performed independently for the
did not yield any significant differences for either region of the
rostro-caudal and mid-caudal SN/VTA cases, and reports the
SN/VTA. This is in agreement with a previous study by Prince
exact p value and adjusted R2 for each independently tested
et al. [31] that also found no changes in COX activity within the
outcome measure (see Table 2). In addition we did not find any
mesencephalon. An alternative explanation for our findings is that
significant colinearity among the demographic or sample vari-
metabolic pathology, and more specifically anomalies in COX
ables.
activity, may affect discrete subpopulations of dopaminergic
neurons within the SN/VTA.
Cytochrome C Oxidase Activity in Schizophrenia Versus
Non-psychiatric Controls Cytochrome C Oxidase Subunit Analysis in the SN/VTA
COX activity was measured independently for samples Subunit composition of the COX enzyme was assessed in
containing the entire rostro-caudal extent of the SN/VTA and schizophrenia and non-psychiatric control cases, as well as in
those containing only mid-caudal regions. For both types of animals chronically treated with antipsychotic medications, by
dissections COX activity was also measured for the red nucleus for measuring protein expression. Subunits I, II, and III of the COX
each case as an internal control for the validity of the assay (see enzyme were selected for study because these mitochondrial DNA-
methods). As stated in methods, COX activity has been shown to encoded subunits constitute the catalytic core of the enzyme. In
best fit an exponential curve [64], thus, all OD data for COX addition, COX subunit IV, which is encoded by nuclear DNA,
activity were transformed to their common logarithm [y = 21*log was selected for study due to its essential role in the assembly of the
(y)], prior to analysis. COX activity is shown in all graphs as the enzyme and in respiratory function [37]. COX subunit IV has two
logarithm of OD units, and the mean value reflects the isoforms (i.e. COX IV-I and COX IV-II) in the human, rat, and
geometrical mean of the lognormal distribution. After logarithmic mouse [36], thus, protein expression of both isoforms was assessed
transformation, all COX activity data were normally distributed in our study.
(Shapiro-Wilk and Kolmogorov-Smirnov tests in all cases p.0.1), COX subunits in Rostro-Caudal SN/VTA
which further supports the exponential curve fit of COX activity. samples. COX subunit I expression did not differ significantly
COX activity in Rostro-Caudal samples. COX activity between schizophrenia and non-psychiatric control samples
measurements for the entire rostro-caudal extent of the SN/VTA (unpaired t-test, p = 0.6895, t = 0.4086, df = 13) [Figure 3A].
were assessed using unpaired t-test to compare the geometric Interestingly, COX subunit II was significantly decreased in
means of the schizophrenia and control groups. This analysis schizophrenia subjects versus non-psychiatric controls (unpaired t-
yielded non-significant differences in COX activity between test with Welch’s correction, p = 0.0332, adjusted t = 2.716,
schizophrenia and controls for the rostro-caudal SN/VTA adjusted df = 6.305). Welch’s correction was used to address the
(p = 0.6001, t = 0.5373, df = 13). The geometric mean for schizo- significant differences in variance between the two groups (F-test
phrenia group was 1.389 (95% CI: 1.212, 1.592) and for the p = 0.0007). Two outliers were identified in the schizophrenia
control group was 1.328 (95% CI: 1.130, 1.561) [Figure 1]. For group (SZ1 and SZ6). Notably, the mean expression of COX
the red nucleus, unpaired t-test revealed non-significant differences subunit II in schizophrenia compared to the control group
for COX activity (p = 0.2011, t = 1.347, df = 13). The geometric presented a 42.77% reduction [Figure 3B]. COX subunit III did
mean for the red nucleus for the schizophrenia group was 1.513 not present significant differences between the two groups
(95% CI: 1.426, 1.605) and for the control group was 1.409 (95% (unpaired t-test with Welch’s correction, p = 0.0925, adjusted
CI: 1.255, 1.581) [Figure 1]. t = 1.884, adjusted df = 8.92). Welch’s correction was used to

PLOS ONE | www.plosone.org 14 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

address the significant differences in variance between the two [Figure 8A], and COX subunit IV-II [one-way ANOVA
groups (F-test p = 0.0267). One outlier (C2) was identified for p = 0.1403, F (2,24) = 2.134] [Figure 8B]. These data support
COX subunit III in the control group [Figure 3C]. COX subunit that antipsychotic drugs do not affect COX subunits expression in
IV-I expression data were not normally distributed for the the SN/VTA.
schizophrenia group (Shapiro-Wilk, p = 0.0018; Kolmogorov- Unlike COX activity, little research has been conducted on the
Smirnov p = 0.0141), and one outlier was identified in the subunit composition of cytochrome c oxidase in regard to
schizophrenia group (SZ6). There were significant differences in schizophrenia pathology. Furthermore, to the authors’ knowledge,
COX subunit IV-I expression between schizophrenia and non- no studies have examined how antipsychotic medication may
psychiatric controls (Mann-Whitney test, p = 0.048, U = 9) affect the individual subunits of this enzyme. For this study we
[Figure 4A]. However, no significant differences were observed concentrated our efforts on examining COX subunits that
between the two groups for COX subunit IV-II (unpaired t-test, contribute directly to electron transfer, proton pumping, and
p = 0.9234, t = 0.0981, df = 13) [Figure 4B]. enzyme stability. To this end, we studied COX subunits I, II, and
Multiple t-test Holm-Sidak correction for the analysis of III (i.e. the ‘‘catalytic core’’) and the two isoforms of COX subunit
subunits I, II, III (i.e. catalytic core) produced similar results to IV (i.e. COX IV-I and COX IV-II). Our study revealed that COX
the independent analysis of these subunits, yielding significant subunit II presented robust significant reductions (i.e. regardless of
differences in the expression of subunit II between the schizo- the type of analysis used) of its expression in schizophrenia when
phrenia and control samples (p = 0.0291, t = 2.5082, df = 11), while the entire rostro-caudal SN/VTA region was present. In addition,
subunits I and III did not present significant differences between a significant reduction of subunit IV-I was also found for SN/VTA
the two groups (subunit I: p = 0.6904, t = 0.4074, df = 13; subunit containing the entire SN/VTA in the independent analysis, but
III: p = 0.1218, t = 1.6649, df = 12). For subunits IV–I and IV–II not when the Holm-Sidak correction for multiple t-test compar-
(i.e. nuclear genome encoded) Holm-Sidak corrected t-test yielded isons was applied. All the other subunits assessed (i.e. subunits I, III
non-significant differences between schizophrenia and control and IV-II) did not yield any significant differences between the two
groups (subunit IV–I: p = 0.9212, t = 0.0386, df = 12; subunit IV– groups for samples containing the entire SN/VTA. For samples
II: p = 0.0291, t = 0.1008, df = 13). containing only mid-caudal SN/VTA regions, none of the
COX subunits in Mid-Caudal SN/VTA samples. COX subunits assessed presented significant differences between schizo-
subunit I expression did not differ significantly between schizo- phrenia and non-psychiatric controls. Even though all the samples
phrenia and control groups for mid-caudal SN/VTA samples used in the study were trimmed to reduce the presence of other cell
(unpaired t-test, p = 0.6736, t = 0.4327, df = 11) [Figure 5A]. groups, we cannot discard small contributions of surrounding cell
Despite the significant differences found in rostro-caudal SN/VTA groups to our results.
samples for COX subunit II protein expression, in mid-caudal Although we did not find significant changes in COX activity,
samples no significant differences were found between the two the alterations in the expression of key subunits of the COX
groups (unpaired t-test, p = 0.2093, t = 1.334, df = 11) [Figure 5B]. enzyme could still have important functional implications. It has
For COX subunit III, an outlier was detected in the schizophrenia been demonstrated that mitochondria have a ‘‘reserve capacity’’ of
group (SZ14), and no significant differences were found between energy production that is critical in determining cellular metabolic
the two groups for this subunit (unpaired t-test with Welch’s response to insults such as metabolic or xenobiotic stress (e.g.
correction, p = 0.2057, t = 1.443, df = 5). Welch’s correction was oxidative stress and hypoxia) [75–76]. It is highly plausible that
used to address the significant differences in variance between the disruptions in subunit composition could affect the proper
two groups (F-test p = 0.0013). [Figure 5C]. No significant assembly of the COX enzyme, thus lowering the reserve capacity
differences were found for either of the two isoforms of COX of mitochondria, and making them more vulnerable to insult. In
subunit IV (subunit IV-I, unpaired t-test, p = 0.0963, t = 1.818, other words, a non-optimally assembled COX enzyme may be
df = 11; subunit IV-II unpaired t-test, p = 0.3249, t = 1.031, df = 11) able to maintain proper basal COX activity and ATP production,
[Figure 6A–B]. but may have an impaired capacity to respond to higher energy
Multiple t-test Holm-Sidak correction also yielded non-signif- demands. Supporting this idea, it has been shown that specific
icant results for the analysis of the catalytic core subunits (subunit mutations in subunits of the catalytic core produce severe
I: p = 0.6742, t = 0.4318, df = 11; subunit II: p = 0.2096, t = 1.3327, impairment of mitochondrial function [44–47].
df = 11; subunit III: p = 0.1796, t = 1.4429, df = 10), as well as for Subunits II and IV of the COX enzyme are crucial for the
the nuclear genome encoded subunits (subunit IV-I: p = 0.0960, proper functioning of the COX complex as a whole [42,45–46].
t = 1.8203, df = 11; subunit IV-II: p = 0.3232, t = 1.0343, df = 11). COX subunit II is responsible for the binding of cytochrome c and
Effect of antipsychotic medication on COX subunit the subsequent electron transfer to subunit I of the COX enzyme
expression. Since all schizophrenia samples included in our [43]. Interestingly, anomalies in COX subunit II mRNA
study were obtained from individuals that were medicated with expression have been previously reported in the frontal cortex in
antipsychotic drugs at time of death, we tested in rats if schizophrenia without significant changes in COX activity [45]. In
antipsychotic treatment could have an impact in the expression addition, reduced expression of COX subunit II has been shown
of the COX subunits included in our study. All rat samples to correlate with a higher susceptibility to neuronal loss, and with
included the entire rostro-caudal extent of the SN/VTA. an increased susceptibility to kainic acid-induced epilepsy in mice
One-way analysis of variance (ANOVA) was used to compare deficient in DNA mismatch repair [77]. COX subunit IV is
the expression of each subunit among the 3 treatment groups. We encoded by nuclear DNA and is not a component of the catalytic
did not find any statistically significant differences among the three core, although several studies support that subunit IV is necessary
treatment groups for the expression of any of the COX subunits for proper electron transfer and enzyme stability [36–37,42]. In
tested. COX subunit I [one-way ANOVA p = 0.5971, F addition, subunit IV has a crucial role in the correct assembly of
(2,24) = 0.5269] [Figure 7A], COX subunit II [one-way ANOVA the COX enzyme, and the binding of subunit IV to subunit I is the
p = 0.2668, F (2,24) = 1.397] [Figure 7B], COX subunit III [one- first step for the assembly of the entire COX complex [15,37,42].
way ANOVA p = 0.6903, F (2, 24) = 0.3764] [Figure 7C], COX Furthermore, suppression of COX subunit IV expression has been
subunit IV-I [one-way ANOVA p = 0.8636, F(2,24) = 0.1475] shown to increase cell susceptibility to apoptosis [37]. It is

PLOS ONE | www.plosone.org 15 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

important to note that subunits encoded by both the mitochondria cortex in schizophrenia, which are some of the main projection
and nuclear genomes were affected in our study. As far as we are areas for the neurons of the rostral SN/VTA. Decreases in the
aware, mutations in the genes encoding these subunits have not expression of these subunits could produce a faulty stoichiometry
been reported in schizophrenia. Deficiencies in the expression of in the assembly of the COX enzyme, leading to a greater
these subunits in the SN/VTA in schizophrenia could be due to a vulnerability to metabolic insults. Additional studies will be needed
faulty transcription or translation, although anomalies in mRNA to fully elucidate the meaning of these findings, including if the
expression for subunit II have been previously reported in the anomalies observed in subunits II and IV-I of the COX enzyme
cortex for this disorder [48]. occur at the level of transcription or translation, and if other
Another interesting finding of our study is that significant components of the metabolic machinery may be compromised
decreases in protein expression for subunits II and IV-I of the within the rostral SN/VTA.
COX enzyme were found only in schizophrenia samples that
contained the entire rostro-caudal extent of the SN/VTA. These Acknowledgments
findings suggest that anomalies in the expression of key subunits of
the COX enzyme could affect specifically rostral regions of the The authors wish to thank the staff of the Maryland Brain Collection,
SN/VTA, which is in strong agreement with our previous report University of Maryland School of Medicine for their help in obtaining the
on deficits in tyrosine hydroxylase (i.e. the rate limiting enzyme for samples used in this study. Drs. Emma Perez-Costas and Miguel Melendez-
Ferro contributed equally to this work.
the production of dopamine) in the same rostral region of the SN/
VTA [58]. This holds important neuroanatomical and clinical
significance since rostral areas of the SN/VTA mainly contain Author Contributions
dopaminergic neurons that contribute to the mesolimbic and Conceived and designed the experiments: MMF EPC. Performed the
mesocortical pathways [56,78–81]. Interestingly, anomalies in the experiments: MMF EPC MWR KLS. Analyzed the data: MMF EPC
expression of tyrosine hydroxylase [82] and subunit II of the COX MWR KLS RCR. Contributed reagents/materials/analysis tools: MMF
enzyme [48] have also been reported in the prefrontal and frontal EPC RCR. Wrote the paper: MMF EPC MWR RCR.

References
1. American Psychiatric Association (2000) Diagnostic and Statistical Manual of 18. Campbell NA, Williamson B, Heyden RJ (2006) Biology: Exploring Life. Boston:
Mental Disorders-IV Text Revision. Washington, DC: American Psychiatric Pearson Prentice Hall.
Press. 19. Babcock DF, Hille B (1998) Mitochondrial oversight of cellular Ca2+ signaling.
2. Takahashi Y, Ogushi T (1953) On biochemical studies of schizophrenia. I. An Curr Opin Neurobiol 8: 398–404.
enzymological study on brain tissue and serum of schizophrenic patients; choline 20. Brodin L, Bakeeva L, Shupliakov O (1999) Presynaptic mitochondria and the
esterase. Folia Psychiatr Neurol Jpn 6: 244–261. temporal pattern of neurotransmitter release. Philos Trans R Soc Lond B Biol
3. Uranova NA, Aganova EA (1989) Ultrastructure of the synapses of the anterior Sci 354: 365–372.
limbic cortex in schizophrenia. Zh Nevropatol Psikhiatr Im S S Korsakova 89: 21. Mattson MP, Culmsee C, Yu Z, Camandola S (2000) Roles of nuclear factor
56–59. kappaB in neuronal survival and plasticity. J Neurochem 74: 443–456.
4. Uranova N, Orlovskaya D, Vikhreva O, Zimina I, Kolomeets N, et al. (2001) 22. Li Z, Okamoto KI, Hayashi Y, Sheng M (2004) The importance of dendritic
Electron microscopy of oligodendroglia in severe mental illness. Brain Res Bull mitochondria in the morphogenesis and plasticity of spines and synapses. Cell
55: 597–610. 119: 873–887.
5. Kung L, Roberts RC (1999) Mitochondrial pathology in human schizophrenic 23. Verstreken P, Ly CV, Venken KJT, Koh T-W, Zhou Y, et al. (2005) Synaptic
striatum: a postmortem ultrastructural study. Synapse 31: 67–75. mitochondria are critical for mobilization of reserve pool vesicles at Drosophila
6. Somerville SM, Lahti AC, Conley RR, Roberts RC (2011) Mitochondria in the neuromuscular junctions. Neuron 47: 365–378.
striatum of subjects with schizophrenia: relationship to treatment response. 24. Barksdale KA, Perez-Costas E, Gandy JC, Melendez-Ferro M, Roberts RC, et
Synapse 65: 215–224. al. (2010) Mitochondrial viability in mouse and human postmortem brain.
7. Somerville SM, Conley RR, Roberts RC (2012) Striatal mitochondria in subjects FASEB J 24: 3590–3599.
with chronic undifferentiated vs. chronic paranoid schizophrenia. Synapse 66: 25. Lee Y, Oh SB, Park HR, Kim HS, Kim MS, et al. (2013) Selective impairment
29–41. on the proliferation of neural progenitor cells by oxidative phosphorylation
8. Kolomeets NS, Uranova NA (1997) Synaptic contacts in schizophrenia: study disruption. Neurosci Lett 535: 134–139.
with immunocytochemical identification of dopaminergic neurons. Zh Nevrol 26. Prince JA, Yassin MS, Oreland L (1997) Neuroleptic-induced mitochondrial
Psikhiatr Im S S Korsakova 97: 39–43. enzyme alterations in the rat brain. J Pharmacol Exp Ther 280: 261–267.
9. Orth M, Schapira AH (2001) Mitochondria and degenerative disorders. 27. Maurer I, Zierz S, Moller H (2001) Evidence for a mitochondrial oxidative
Am J Med Genet 106: 27–36. phosphorylation defect in brains from patients with schizophrenia. Schizophr
10. Rollins B, Martin MV, Sequeira PA, Moon EA, Morgan LZ, et al. (2009) Res 48: 125–136.
Mitochondrial variants in schizophrenia, bipolar disorder, and major depressive 28. Rosenfeld M, Brenner-Lavie H, Ari SG-B, Kavushansky A, Ben-Shachar D
disorder. PLoS One 4: e4913. (2011) Perturbation in mitochondrial network dynamics and in complex I
dependent cellular respiration in schizophrenia. Biol Psychiatry 69: 980–988.
11. Hall CN, Klein-Flugge MC, Howarth C, Attwell D (2012) Oxidative
29. Manji H, Kato T, Di Prospero NA, Ness S, Beal MF, et al. (2012) Impaired
phosphorylation, not glycolysis, powers presynaptic and postsynaptic mecha-
mitochondrial function in psychiatric disorders. Nat Rev Neurosci 13: 293–307.
nisms underlying brain information processing. J Neurosci 32: 8940–8951.
30. Cavelier L, Jazin EE, Eriksson I, Prince J, Bave U, et al. (1995) Decreased
12. Volz HR, Riehemann S, Maurer I, Smesny S, Sommer M, et al. (2000) Reduced
cytochrome-c oxidase activity and lack of age-related accumulation of
phosphodiesters and high-energy phosphates in the frontal lobe of schizophrenic
mitochondrial DNA deletions in the brains of schizophrenics. Genomics 29:
patients: a (31)P chemical shift spectroscopic-imaging study. Biol Psychiatry 47:
217–224.
954–961.
31. Prince JA, Blennow K, Gottfries CG, Karlsson I, Oreland L (1999)
13. Saddar S, Dienhart MK, Stuart RA (2008) The F1F0-ATP synthase complex
Mitochondrial function is differentially altered in the basal ganglia of chronic
influences the assembly state of the cytochrome bc1-cytochrome oxidase schizophrenics. Neuropsychopharmacology 21: 372–379.
supercomplex and its association with the TIM23 machinery. J Biol Chem 32. Ben-Shachar D, Karry R (2008) Neuroanatomical pattern of mitochondrial
283: 6677–6686. complex I pathology varies between schizophrenia, bipolar disorder and major
14. Manatt M, Chandra S (2011) The effects of mitochondrial dysfunction in depression. PLoS One 3 (11): e3676.
schizophrenia. J Med Genet Genomics. 3: 84–94. 33. Ben-Shachar D (2009) The interplay between mitochondrial complex I,
15. Fontanesi F, Soto IC, Horn D, Barrientos A (2006) Assembly of mitochondrial dopamine and Sp1 in schizophrenia. J Neural Transm 116: 1383–1398.
cytochrome c-oxidase, a complicated and highly regulated cellular process. 34. Rosenfeld M, Brenner-Lavie H, Ari SG, Kavushansky A, Ben-Shachar D (2011)
Am J Physiol Cell Physiol 291: C1129–C1147. Perturbation in mitochondrial network dynamics and in complex I dependent
16. Fukui H, Moraes CT (2008) The mitochondrial impairment, oxidative stress and cellular respiration in schizophrenia. Biol Psych 69: 980–988.
neurodegeneration connection: reality or just an attractive hypothesis? Trends 35. Verkhovsky MI, Jasaitis A, Verkhovskaya ML, Morgan JE, Wikstrom M (1999)
Neurosci 31: 251–256. Proton translocation by cytochrome c oxidase. Nature 400: 480–483.
17. Michel H, Behr J, Harrenga A, Kannt A (1998) Cytochrome c oxidase: structure 36. Huttemann M, Kadenbach B, Grossman LI (2001) Mammalian subunit IV
and spectroscopy. Annu Rev Biophys Biomol Struct 27: 329–356. isoforms of cytochrome c oxidase. Gene 267: 111–123.

PLOS ONE | www.plosone.org 16 June 2014 | Volume 9 | Issue 6 | e100054

 Cytochrome C Oxidase in Schizophrenia

37. Li Y, Park JS, Deng JH, Bai Y (2006) Cytochrome c oxidase subunit IV is 60. Paxinos G, Huang XF (1995) Atlas of the human brainstem. San Diego:
essential for assembly and respiratory function of the enzyme complex. Academic Press.
J Bioenerg Biomembr 38: 283–291. 61. Perez-Costas E, Guidetti P, Melendez-Ferro M, Kelley JJ, Roberts RC (2008)
38. Kadenbach B, Huttemann M, Arnold S, Lee I, Bender E (2000) Mitochondrial Neuroleptics and animal models: feasibility of oral treatment monitored by
energy metabolism is regulated via nuclear-coded subunits of cytochrome c plasma levels and receptor occupancy assays. J Neural Transm 115: 745–753.
oxidase. Free Radic Biol Med 29: 211–221. 62. Divac I, Mojsilovic-Petrovic J, Lopez-Figueroa MO, Petrovic-Minic B, Moller M
39. Barrientos A, Barros MH, Valnot I, Rotig A, Rustin P, et al. (2002) Cytochrome (1995) Improved contrast in histochemical detection of cytochrome oxidase:
oxidase in health and disease. Gene 286: 53–63. metallic ions protocol. J Neurosci Methods 56: 105–113.
40. Shoubridge EA (2001) Cytochrome c oxidase deficiency. Am J Ned Genet 106: 63. Poeggeler B, Rassoulpour A, Guidetti P, Wu HQ, Schwarcz R (1998)
46–52. Dopaminergic control of kynurenate levels and N-methyl-D-aspartate toxicity
41. Tsukihara T, Aoyama H, Yamashita E, Tomizaki T, Yamaguchi H, et al. (1996) in the developing rat striatum. Dev Neurosci 20: 146–153.
The whole structure of the 13-subunit oxidized cytochrome c oxidase at 2.8 A. 64. Melendez-Ferro M, Rice MW, Roberts RC, Perez-Costas E (2013) An accurate
Science 272: 1136–1144. method for the quantification of cytochrome C oxidase in tissue sections.
42. Nijtmans LG, Taanman JW, Muijsers AO, Speijer D, Van den Bogert C (1998) J Neurosci Methods 214: 156–162.
Assembly of cytochrome-c oxidase in cultured human cells. Eur J Biochem 254: 65. Motulski HJ, Brown RE (2006) Detecting outliers when fitting data with
389–394. nonlinear regression- a new method based on robust nonlinear regression and
43. Taanman JW (1997) Human cytochrome c oxidase: structure, function, and the false discovery rate. BMC Bioinformatics 7: 123.
deficiency. J Bioenerg Biomembr 29: 151–163. 66. Rothman KJ (1990) No adjustments are needed for multiple comparisons.
44. Bruno C, Martinuzzi A, Tang Y, Andreu AL, Pallotti F, et al. (1999) A stop- Epidemiology 1: 43–46.
codon mutation in the human mtDNA cytochrome c oxidase I gene disrupts the 67. Motulsky HJ (2010) Intuitive biostatistics: A nonmathematical guide to statistical
functional structure of complex IV. Am J Hum Genet 65: 611–620. thinking. New York: Oxford University Press. 331–342 pp.
45. Clark KM, Taylor RW, Johnson MA, Chinnery PF, Chrzanowska-Lightowlers 68. Pel HJ, Grivell LA (1994) Protein synthesis in mitochondria. Mol Biol Rep 19:
ZM, et al. (1999) An mtDNA mutation in the initiation codon of the cytochrome 183–194.
C oxidase subunit II gene results in lower levels of the protein and a 69. Richman TR, Davies SM, Shearwood AM, Ermer JA, Scott LH, et al. (2014) A
mitochondrial encephalomyopathy. Am J Hum Genet 64: 1330–1339. bifunctional protein regulates mitochondrial protein synthesis. Nuclei Acids Res
46. Rahman S, Taanman JW, Cooper JM, Nelson I, Hargreaves I, et al. (1999) A 42: 5483–5494.
missense mutation of cytochrome oxidase subunit II causes defective assembly 70. Gattermann N, Retzlaff S, Wang YL, Hofhaus G, Heinisch J, et al. (1997)
and myopathy. Am J Hum Genet 65: 1030–1039. Heteroplasmic point mutations of mitochondrial DNA affecting subunit I of
47. Tiranti V, Corona P, Greco M, Taanman JW, Carrara F, et al. (2000) A novel cytochrome c oxidase in two patients with acquired idiopathic sideroblastic
frameshift mutation of the mtDNA COIII gene leads to impaired assembly of anemia. Blood 90: 4961–4972.
cytochrome c oxidase in a patient affected by Leigh-like syndrome. Hum Mol 71. You KR, Wen J, Lee ST, Kim DG (2002) Cytochrome c oxidase subunit III: a
Genet 9: 2733–2742. molecular marker for N-(4-hydroxyphenyl)retinamise-induced oxidative stress in
48. Whatley SA, Curti D, Marchbanks RM (1996) Mitochondrial involvement in hepatoma cells. J Biol Chem 277: 3870–3877.
schizophrenia and other functional psychoses. Neurochem Res 21: 995–1004. 72. Abril J, de Heredia ML, Gonzalez L, Cleries R, Nadal M, et al. (2008) Altered
49. Maurer I, Moller HJ (1997) Inhibition of complex I by neuroleptics in normal expression of 12S/MT-RNR1, MTCO2/COX2, and MT-ATP6 mitochondrial
human brain cortex parallels the extrapyramidal toxicity of neuroleptics. Mol genes in prostate cancer. Prostate 68: 1086–1096.
Cell Biochem 174: 255–259. 73. Ivanova MM, Luken KH, Zimmer AS, Lenzo FL, Smith RJ, et al. (2011)
50. Burkhardt C, Kelly JP, Lim YH, Filley CM, Parker WD (1993) Neuroleptic Tamoxifen increases nuclear respiratory factor 1 transcription by activating
medications inhibit complex I of the electron transport chain. Ann Neurol 33: estrogen receptor beta and AP-1 recruitment to adjacent promoter binding sites.
512–517. FASEB J 25: 1402–1416.
51. Whatley SA, Curti D, Das Gupta F, Ferrier IN, Jones S, et al. (1998) Superoxide, 74. Arnold SR, Sun Q, Sun CQ, Richards JC, O’Hearn S, et al. (2013) A inherited
neuroleptics and the ubiquinone and cytochrome b5 reductases in brain and heteroplasmic mutation in mitochondrial gene COI in a patient with prostate
lymphocytes from normals and schizophrenic patients. Mol Psychiatry 3: 227– cancer alters reactive oxygen, reactive nitrogen and proliferation. Biomed Res
237. Int (Epub ahead of print) doi: 10.1155/2013/239257.
52. Balijepalli S, Boyd MR, Ravindranath V (1999) Inhibition of mitochondrial 75. Perez J, Hill BG, Benavides GA, Dranka BP, Darley-Usmar VM (2010) Role of
complex I by haloperidol: the role of thiol oxidation. Neuropharmacology 38: cellular bioenergetics in smooth muscle cell proliferation induced by platelet-
567–577. derived growth factor. Biochem J 428: 255–267.
53. Balijepalli S, Kenchappa RS, Boyd MR, Ravindranath V (2001) Protein thiol 76. Mitchell T, Chacko B, Ballinger SW, Bailey SM, Zhang J, et al. (2013)
oxidation by haloperidol results in inhibition of mitochondrial complex I in brain Convergent mechanisms for dysregulation of mitochondrial quality control in
regions: comparison with atypical antipsychotics. Neurochem Int 38: 425–435. metabolic disease: implications for mitochondrial therapeutics. Biochem Soc
54. Streck EL, Rezin GT, Barbosa LM, Assis LC, Grandi E, et al. (2007) Effect of Trans 41: 127–133.
antipsychotics on succinate dehydrogenase and cytochrome oxidase activities in 77. Francisconi S, Codenotti M, Ferrari Toninelli G, Uberti D, Memo M (2006)
rat brain. Naunyn Schmiedebergs Arch Pharmacol 376: 127–133. Mitochondrial dysfunction and increased sensitivity to excitotoxicity in mice
55. van Domburg PH, ten Donkelaar HJ (1991) The human substantia nigra and deficient in DNA mismatch repair. J Neurochem 98: 223–233.
ventral tegmental area. A neuroanatomical study with notes on aging and aging 78. Crosby EC, Woodburne RT (1943) The nuclear pattern of the non-tectal
diseases. Adv Anat Embryol Cell Biol 121: 1–132. portions of the midbrain and isthmus in primates. J Comp Neurol 78: 441–482.
56. Haber SN, Fudge JL (1997) The primate substantia nigra and VTA: integrative 79. Marin O, Smeets WJ, Gonzalez A (1998) Evolution of the basal ganglia in
circuitry and function. Crit Rev Neurobiol 11: 323–342. tetrapods: a new perspective based on recent studies in amphibians. Trends
57. Perez-Costas E, Melendez-Ferro M, Roberts RC (2010) Basal ganglia pathology Neurosci 21: 487–494.
in schizophrenia: dopamine connections and anomalies. J Neurochem 113: 287– 80. Joel D, Weiner I (2000) The connections of the dopaminergic system with the
302. striatum in rats and primates: an analysis with respect to the functional and
58. Perez-Costas E, Melendez-Ferro M, Rice MW, Conley RR, Roberts RC (2012) compartmental organization of the striatum. Neuroscience 96: 451–474.
Dopamine pathology in schizophrenia: analysis of total and phosphorylated 81. Augustine JR (2008) The brain stem. In: Human Neuroanatomy. London:
tyrosine hydroxylase in the substantia nigra. Front psychiatry 3: 31. Academic Press. 49–68.
59. Spitzer RL, Williams JB, Gibbon M, First MB (1992) The Structured Clinical 82. Akil M, Pierri JN, Whitehead RE, Edgar CL, Mohila C, et al. (1999) Lamina-
Interview for DSM-III-R (SCID). I: History, rationale, and description. Arch specific alterations in the dopamine innervation of the prefrontal cortex in
Gen Psychiatry 49: 624–629. schizophrenic subjects. Am J Psychiatry 156: 1580–1589.

PLOS ONE | www.plosone.org 17 June 2014 | Volume 9 | Issue 6 | e100054