Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Psychiatry Research journal homepage: www.elsevier.com/locate/psychres

An updated meta-analysis of oxidative stress markers in bipolar disorder Nicole C. Brown a, Ana C. Andreazza a,b, L. Trevor Young a,b,n a b

Department of Pharmacology and Toxicology, University of Toronto, Toronto, ON, Canada Department of Psychiatry, University of Toronto, Toronto, ON, Canada

art ic l e i nf o

a b s t r a c t

Article history: Received 18 September 2013 Received in revised form 22 January 2014 Accepted 2 April 2014

Despite its debilitating symptoms, the pathophysiology of bipolar disorder (BD) remains unclear. One consistently compelling finding, however, has been the presence of oxidative stress. In the present investigation, we conducted a meta-analysis of studies that measured oxidative stress markers in BD patients compared to healthy controls. Search terms and selection criteria were determined a priori to identify and include all studies that measured a marker of oxidative stress in BD compared to healthy controls. Eight markers were included: superoxide dismutase, catalase, protein carbonyl, glutathione peroxidase, 3-nitrotyrosine, lipid peroxidation, nitric oxide, and DNA/RNA damage. A meta-analysis of standardized means was conducted using a random-effects model with generic inverse weighting. Between-study heterogeneity, publication bias, and sensitivity analyses were also examined for each marker. Twenty-seven papers were included in the meta-analysis, which comprised a total of 971 unique patients with BD and 886 healthy controls. Lipid peroxidation, DNA/RNA damage, and nitric oxide were significantly increased in BD patients compared to healthy controls. Additionally, the effect size for lipid peroxidation was very high. Publication bias was not detected for any of the markers. The main limitations in this meta-analysis are the high degree of heterogeneity between studies and the small number of studies used in the analysis of some markers. Additionally, the sensitivity analysis indicated that some results are not very robust. The results from this meta-analysis support the role of oxidative stress in bipolar disorder, especially to DNA, RNA, and lipids. & 2014 Elsevier Ireland Ltd. All rights reserved.

Keywords: Bipolar disorder Oxidative stress Antioxidant enzymes Lipid peroxidation Post-mortem brain

1. Introduction Psychiatry, unlike most other fields of medicine, lacks specific and reliable biomarkers to diagnose and monitor illness. Although clinician observation is important in many branches of medicine, most also utilize diagnostic tests. Bipolar disorder can be difficult to diagnose because of symptom overlap with other mood and psychotic disorders such as major depressive disorder and schizophrenia. Genetic epidemiology findings have also provided evidence of shared genetic risk factors between bipolar disorder, schizophrenia, and major depressive disorder (Craddock and Owen, 2005). There may be a long delay (up to 10 years) between illness onset and a diagnosis of bipolar disorder in which time a misdiagnosis may lead to ineffective treatment and worse outcomes. For example, a misdiagnosis of BD as unipolar depression may lead to inappropriate prescriptions, such as the use of

n Correspondence to: Department of Psychiatry, University of Toronto, 250 College Street, Toronto, ON Canada M5T 1R8, Room: Ste. 835. Tel.: þ 1 416 979 6948. E-mail address: [email protected] (L.T. Young).

antidepressants without a mood-stabilizing drug, which may lead to mania and poor clinical and functional outcomes (Phillips and Kupfer, 2013). The development of a biomarker for bipolar disorder would improve diagnostic accuracy and potentially allow intervention at early stages of the illness, which may be critical to lowering the lifetime illness burden (Perry et al., 1999; Miklowitz et al., 2013). The complexity of bipolar disorder makes the identification of its pathophysiology a challenge. One consistently compelling finding of biological alterations in BD is oxidative stress damage. A recent positional paper from the biomarkers network from the International Society for Bipolar Disorder (ISBD-BIONET) included oxidative stress markers, among others, as potential biomarkers for BD (Frey et al., 2013). Although many oxidative stress markers have been investigated in BD, the findings are not always consistent; some studies have identified oxidative damage to DNA, RNA, proteins, and lipids in BD subjects, while others report that altered levels of some antioxidant enzymes are altered. These results are supported by evidence such as mitochondrial DNA mutations and decreased levels of proteins from the mitochondrial electron transport chain. A meta-analysis from our group in 2008

http://dx.doi.org/10.1016/j.psychres.2014.04.005 0165-1781/& 2014 Elsevier Ireland Ltd. All rights reserved.

Please cite this article as: Brown, N.C., et al., An updated meta-analysis of oxidative stress markers in bipolar disorder. Psychiatry Research (2014), http://dx.doi.org/10.1016/j.psychres.2014.04.005i

N.C. Brown et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

2

showed a statistically significant increase in lipid peroxidation and nitric oxide in BD (Andreazza et al., 2008). Since then, there have been many additional studies and therefore it is the objective of this analysis to incorporate these new results and to identify any new oxidative stress markers in BD.

2. Methods 2.1. Search strategy A prospective protocol for this study was developed a priori with search terms and inclusion criteria chosen in an attempt to include all relevant publications. Web of Science, BIOSIS, and MEDLINE databases were searched for the term bipolar disorder with the following: oxidative stress, reactive oxygen species, free radicals, antioxidant, nitric oxide, lipid peroxidation, TBARS, protein carbonyl, 3-nitrotyrosine, catalase, glutathione, DNA oxidation, DNA damage, or DNA fragmentation. References cited in publications found using these search terms were also reviewed for any relevant studies not already identified and all searches were conducted prior to May 2013 with no time span specified. 2.2. Selection criteria One reviewer screened all abstracts of potentially relevant publications. Studies were included if they met the following criteria: (1) measured levels of one or more of the following oxidative stress markers in both patients with bipolar disorder and healthy controls: superoxide dismutase, catalase, glutathione peroxidase, protein carbonyl, 3-nitrotyrosine, nitric oxide, DNA/RNA damage, and lipid peroxidation; (2) were reported in an original research paper in a peer-reviewed journal; and (3) if they adequately described their samples (e.g. diagnostic criteria, source of samples, and storage) and methods such that the experiments could be replicated (or included appropriate references). Studies were retained regardless of the measurement method or sample type (peripheral or post-mortem brain). Additionally, authors were contacted for mean values and standard deviations when their methods were appropriate but data was expressed in a graph or figure only (Andreazza et al., 2009; Wang et al., 2009; Che et al., 2010; Mustak et al., 2010; Gawryluk et al., 2011; Gigante et al., 2011; Andreazza et al., 2013). For all included studies, the disease state of BD patients, number of drug-free patients, sample type, type of assay/measurement, and results were recorded. 2.3. Statistical analysis The meta-analysis of pooled standardized mean differences was conducted using Review Manager software (Version 5.2, Copenhagen) from The Cochrane Collaboration. The effect sizes for the standardized mean differences were expressed through Hedges's G and a Z-score; a p-value of o 0.05 for Z was considered statistically significant. A random-effects model was used and studies were weighted by the generic inverse variance method. The between-study heterogeneity was determined using the Cochrane's Q statistic and expressed using I2 and τ2. Publication bias was assessed by visually inspecting funnel plots and applying Egger's regression test with p o 0.1 as statistically significant (Egger et al., 1997) using the software program Comprehensive Meta-analysis (Borenstein et al., 2005). A one-study removed sensitivity analysis was performed for each oxidation marker by manually excluding each study included in the analysis to determine robustness. In cases where patients were separated into subgroups (i.e. manic, depressed, or euthymic), the means and standard deviations were pooled to compare all bipolar groups with healthy controls; 15 out of 27 studies included information about the patient disease state. All comparisons were two-tailed and 95% confidence intervals (CI) are expressed where applicable.

3. Results In total, 226 studies were screened and 29 fit the selection criteria. Of the 226 screened papers, 68 were review articles, 48 were animal or cell studies, 51 did not measure an included marker of oxidative stress, 28 were genetic studies, and two did not include a healthy control group. Twenty-seven studies were included in the meta-analysis out of the 29 that fit the selection criteria; two studies were excluded due to missing means and standard deviations (Benes et al., 2003; Buttner et al., 2007). All diagnoses, except for one, were established based on DSM-IV criteria; the one exception was published by Abdalla et al. in 1986 and used ICD-9 (International Classification of Diseases)

criteria, which was deemed appropriate for inclusion. After pooling the included studies, there were a total of 971 unique BD patients and 886 healthy controls. Table 1 outlines the characteristics of these studies including the disease state of BD patients, number of drug-free patients, sample type, type of assay, and overall results. A total of eight oxidative stress markers were included in this analysis: superoxide dismutase, catalase, glutathione peroxidase, protein carbonyl, 3-nitrotyrosine, nitric oxide, DNA/RNA damage, and lipid peroxidation. Table 2 outlines the pooled statistics and meta-analysis for the oxidative stress markers in patients with BD and controls. In total, three out of these eight oxidative stress markers showed a statistically significant change in BD patients compared to healthy controls: lipid peroxidation, nitric oxide level, and DNA/RNA damage. Forest plots of all standardized mean differences and 95% confidence intervals are shown in Fig. 1. Given the small number of studies, we performed a one-study removed sensitivity analysis by excluding each study individually. The Z-value remained significant for DNA/RNA damage and lipid peroxidation and the effect size for SOD and GPx remained essentially unchanged in direction and magnitude after the removal of each study individually. The sensitivity analysis of protein carbonyl, 3-nitrotyrosine, catalase, and nitric oxide showed that these results are not very robust and should be interpreted cautiously: (1) for protein carbonyl, the removal of Andreazza et al. (2009) caused the Z-value to increase from 1.19 (p ¼0.23) to 2.13 (p¼ 0.03); (2) for 3-nitrotyrosine, the removal of Andreazza et al. (2013) caused the Z-value to increase from 1.72 (p ¼0.09) to 4.32 (p ¼0.000015); (3) for catalase, the removal of Machado-Vieira et al. (2007) caused the Z-value to decrease from  1.65 (p ¼0.10) to  3.26 (p ¼0.024); and (4) for nitric oxide, the removal of Ozcan et al. (2004) caused the Z-value to increase from 2.06 (p ¼0.04) to 5.39 (p o0.00001). Publication bias, measured by Egger's regression test, was negative for all markers: SOD (95% CI¼  38.7 to 6.0; p ¼0.13), catalase (95% CI ¼  43.0 to 24.9; p ¼0.45), GPx (95% CI ¼  9.8 to 7.4; p ¼0.74), lipid peroxidation (95% CI¼  7.2 to 8.5; p ¼0.87), protein carbonyl (95% CI¼  50.7 to 58.4; p ¼0.79), nitric oxide (95% CI ¼ 53.7 to 55.9; p ¼0.95), 3-nitrotyrosine (95% CI ¼  146.4 to 130.6; p ¼0.60), and DNA/RNA damage (95% CI¼  2.7 to 14.4; p¼ 0.12).

4. Discussion This meta-analysis further supports the presence of oxidative damage in BD; specifically, our results showed increased lipid peroxidation, increased DNA/RNA damage, and increased levels of nitric oxide in BD patients compared to healthy controls. Many lines of examination in the pathophysiology of BD converge on oxidative stress and an underlying abnormality in oxidative energy generation. Mitochondria are intracellular organelles that are responsible for ATP production through oxidative phosphorylation by the electron transport chain. Alterations to this pathway could lead to increased reactive oxygen species which may overwhelm antioxidant systems and cause damage to lipids (cell and organelle membranes), proteins (receptors, transcription factors, and enzymes, etc.) and DNA. The involvement of mitochondrial dysfunction in BD is supported by several lines of evidence such as reduced expression of several mitochondrial electron transport chain subunits, increased mtDNA deletion and mutation, reduced pH, and decreased levels of high-energy phosphates in the brain of BD patients (Clay et al., 2011). Further studies are needed to determine the longitudinal effects of oxidative stress in BD. In this meta-analysis there is a very strong effect size of lipid peroxidation in BD compared to healthy controls and this

Please cite this article as: Brown, N.C., et al., An updated meta-analysis of oxidative stress markers in bipolar disorder. Psychiatry Research (2014), http://dx.doi.org/10.1016/j.psychres.2014.04.005i

Reference

Number (patients/ controls)

Bipolar patients

Sample

Marker

Assay

Results of BD compared to healthy controls

Manic Depressed Euthymic Firstepisode

Drugfree

Peripheral Post-mortem brain

Abdalla et al. (1986) 20/58

NA

NA

NA

NA

NA

RBC

–

SOD GPx

Nitroblue tetrazolium nmol NADPH oxidized/min

Increased NS

Andreazza et al. (2007a)

32

21

32

0

0

Serum

–

SOD

Adrenochrome

Increased in depressed and manic patients

CAT GPx Lipid per.

μmol of H2O2 consumed/min nmol NADPH oxidized/min TBARS

Decreased in euthymic and manic patients Increased in euthymic patients Increased in manic, decreased in euthymic patients

85/32

Andreazza et al. (2007b)

32/32

NA

NA

NA

NA

0

Whole blood

–

DNA/RNA dam.

Comet assay

DNA damage increased

Andreazza et al. (2009)

30/30 (early BD) 30/30 (late BD)

NA

NA

NA

NA

0

Serum

–

GPx

nmol NADPH oxidized/min

NS

PCC 3-NT

DNPH reaction ELISA

NS Increased in early and late stage patients

PCC

DNPH reaction

Increased

3-NT

ELISA

Increased

Andreazza et al. (2010)

15/15

NA

NA

NA

NA

0

–

PFC (BA10)

Andreazza et al. (2013)

16/26

NA

NA

NA

NA

0

–

PFC (BA10)

Lipid perox. PCC 3-NT

Lipid hydroperoxides assay kit and 4-HNE increased in synaptosomal section, no 4-HNE ELISA difference in LPH DNPH reaction Increased in synaptosomal proteins Immunoblotting Increased in mitochondrial proteins

Banerjee et al. (2012)

73/35

0

0

48

25

NA

Serum

–

Lipid perox.

TBARS

Increased in all BD groups

Benes et al. (2003)a

10/18

NA

NA

NA

NA

4

–

ACC

DNA/RNA dam.

Klenow method

NS

Buttner et al. (2007)a

14/14

NA

NA

NA

NA

2

–

ACC (BA24)

DNA/RNA dam.

Klenow method

Scission increased in non-GABAergic cells only

Che et al. (2010)b

15/15

NA

NA

NA

NA

3

–

Anterior hippo.

DNA/RNA dam.

Immunohistochemistry

RNA damage increased in patients more than DNA

Gawryluk et al. (2011)

14/12

NA

NA

NA

NA

NA

–

PFC (BA10)

GPx

Immunoblotting

NS

Gergerlioglu et al. (2007)

29/30

29

0

0

0

0

Serum

–

SOD

Nitroblue tetrazolium

Decreased

NO

Greiss reaction

Increased

Gigante et al. (2011) 35/35

NA

NA

NA

NA

NA

–

Dorsolateral PFC SOD (BA9)

Immunoblotting

NS

Kapczinski et al. (2011)

60/80

20

20

20

0

0

Serum

–

Lipid perox. PCC

TBARS

Increased in manic and depressed patients

Kuloglu et al. (2002) 23/20

NA

NA

NA

NA

NA

RBC

–

SOD GPx

DNPH reaction

Increased in manic and depressed

Nitroblue tetrazolium nmol NADPH oxidized/min TBARS

Increased NS Increased

N.C. Brown et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Please cite this article as: Brown, N.C., et al., An updated meta-analysis of oxidative stress markers in bipolar disorder. Psychiatry Research (2014), http://dx.doi.org/10.1016/j.psychres.2014.04.005i

Table 1 Selected characteristics of all studies included in the meta-analysis of oxidative stress markers in bipolar disorder patients compared to healthy controls.

3

4

Reference

Number (patients/ controls)

Bipolar patients

Sample

Manic Depressed Euthymic Firstepisode

Drugfree

Marker

Assay

Results of BD compared to healthy controls

Peripheral Post-mortem brain Lipid perox.

Kunz et al. (2008)

83/32

32

19

32

0

NA

Serum

–

SOD Lipid perox.

Adrenochrome TBARS

Increased in manic and depressed patients Increased in all BD groups

Machado-Vieira et al. (2007)

45/30

45

0

0

0

30

Serum

–

SOD

Adrenochrome

Increased in drug-free manic patients

CAT Lipid perox.

μmol of H2O2 consumed/min TBARS

Increased Increased in drug-free manic patients

Lipid perox. PCC

TBARS

NS

DNPH reaction

Increased

Magalhaes et al. (2012)

53/89 (lipid perox.) 48/ 11 75 (PCC)

42

0

NA

44

Serum

–

Mustak et al. (2010) 10/8

NA

NA

NA

NA

10

–

Multiple brain regions

DNA/RNA dam.

Klenow method and incorporation Increased single and double stranded DNA breaks of 3[H]-dTTP

Ozcan et al. (2004)

16

2

0

0

0

RBC

–

Serum

–

SOD CAT GPx Lipid perox. NO

Nitroblue tetrazolium μmol of H2O2 consumed/min nmol NADPH oxidized/min TBARS

30/21

NS Decreased in all BD groups Decreased in pretreatment group Increased in pre- and post-treatment groups

Plasma

–

Greiss reaction

Decreased in pretreatment group

Raffa et al. (2012)

30/40

8

5

17

0

NA

RBC

–

SOD CAT GPx

Pyrogallol μmol of H2O2 consumed/min nmol NADPH oxidized/min

NS Decreased NS

Ranjekar et al. (2003)

10/31

NA

NA

NA

NA

NA

RBC

–

SOD

Nitroblue tetrazolium

NS

CAT GPx Lipid perox.

μmol of H2O2 consumed/min nmol NADPH oxidized/min TBARS

Decreased NS NS

Savas et al. (2002)

44/21

44

0

0

0

0

Plasma

–

NO

Greiss reaction

Increased

Savas et al. (2006)

27/20

0

0

27

0

0

Serum

–

SOD NO

Nitroblue tetrazolium Greiss reaction

Increased Increased

Selek et al. (2008)

30/30

0

30

0

0

0

Serum

–

SOD NO

Nitroblue tetrazolium Greiss reaction

Decreased Increased

Soeiro-de-Souza et al. (2013)

50/50

26

24

0

NA

50

Whole blood

–

DNA/RNA dam.

ELISA

Increased hydroxylated guanine in DNA

Versace et al. (2013) 24/18

0

0

24

0

0

Whole blood

–

Lipid perox.

Lipid hydroperoxides assay kit

Increased

Wang et al. (2009)

15/15

NA

NA

NA

NA

NA

–

ACC

Lipid perox.

Immunnohistochemistry

Increased

Yanik et al. (2004)

43/31

43

0

0

0

0

Plasma

–

NO

Greiss reaction

Increased

a b

Not included in meta-analysis due to missing data. Patient info obtained from Dowlatshahi et al. (1999).

N.C. Brown et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Please cite this article as: Brown, N.C., et al., An updated meta-analysis of oxidative stress markers in bipolar disorder. Psychiatry Research (2014), http://dx.doi.org/10.1016/j.psychres.2014.04.005i

Table 1 (continued )

N.C. Brown et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

5

Table 2 Pooled statistics and meta-analysis of standardized mean group differences for oxidative stress markers in BD compared with healthy controls. Marker

Lipid peroxidation n Nitric oxide n DNA/RNA damage n Superoxide dismutase Catalase Protein carbonyl Glutathione peroxidase 3-Nitrotyrosine

Number of studies

12 6 4 12 5 5 8 3

Total N

Effect

Heterogeneity

BD

Con

Hedges's g (95% CI)

Z

P(Z)

τ2

I2 (%)

517 203 117 440 200 199 272 90

426 153 113 376 154 255 273 100

1.62 0.93 3.13 0.12  1.58 0.62  0.05 1.17

5.31 2.06 3.59 0.26  1.65 1.19 0.26 1.72

o 0.00001 0.04 0.0003 0.80 0.10 0.23 0.80 0.09

1.07 1.13 3.14 2.71 4.42 1.28 0.27 1.28

93 93 94 97 98 96 79 93

(1.02, 2.22) (0.05, 1.82) (1.42, 4.84) (  0.82, 1.07) (  3.46, 0.30) (  0.40, 1.64) (  0.47, 0.36) (  0.16, 2.50)

Abbreviations: BD, bipolar disorder; Con, controls. n

Statistically significant (p o 0.05).

increased lipid peroxidation is shown consistently in both serum and post-mortem brain samples. Lipids are very prone to oxidative damage due to their large size and high number of unsaturated bonds. Oxidative damage to these lipids disrupts cell membranes and the end products of peroxidation are toxic. Since lipids account for about 70% of the dry weight of myelin, the main component of white matter, this damage may play a role in the pathophysiology of BD. Interestingly, a recent paper examined whether peripheral lipid peroxidation levels were associated with white matter abnormalities and showed that 59% and 51% of fractional anisotropy and radial diffusivity differences, respectively, could be explained by variation in lipid hydroperoxide levels (Versace et al., 2013). There is evidence that lipid peroxidation in serum is decreased with medication (Aliyazicioglu et al., 2007); however this was not accounted for in this analysis and the effect size was strong. Lipid peroxidation is a promising potential marker since it can be measured in serum and holds promise to reflect brain alterations. If validated, there is a possibility for markers of lipid peroxidation to be used as a prognostic biomarker along with neuroimaging tests. Furthermore, the widely used TBARS or LPH assays for quantification do not require specialized skills or equipment beyond that in a normal diagnostic laboratory. There are many pathways through which increased oxidative stress can damage DNA or RNA including scission or breaks and base modifications; these two types of oxidative damage were included in this analysis. This is the first time a meta-analysis has examined DNA and RNA oxidative damage in BD and our results show damage was increased in all studies, which includes postmortem brain samples and peripheral samples (Andreazza et al., 2007b; Che et al., 2010; Mustak et al., 2010; Soeiro-de-Souza et al., 2013). This increase in DNA scission and base hydroxylation may lead to increased cell necrosis and subsequent inflammation of nearby tissues (Kim et al., 2001). Oxidative stress to cells may also induce epigenetic changes through different mechanisms including DNA hypomethylation and histone acetylation (Gu et al., 2013). Two of the included studies measured base modifications in DNA and RNA (Che et al., 2010; Soeiro-de-Souza et al., 2013). The two important nucleoside oxidation targets are guanosine and cytosine. Guanosine is the most readily oxidizable base and its hydroxylation to 8-hydroxy-2-deoxyguanosine (8-OHdG) is often considered an indicator of overall DNA and RNA damage. Cytoplasmic RNA is especially vulnerable to this hydroxylation, and damage to mRNA causes improper translation and protein aggregation (Shan et al., 2003). Hydroxylation of guanosine bases may also promote hypomethylation through conformation changes in the DNA that may affect the ability of methyl binding proteins to recognize their CpG island target. The oxidation of 5methylcytosine (5-mC) to 5-hydroxymethylcytosine (5-HmC) is an important step for epigenetic regulation and is normally

controlled by the enzyme ten-eleven-translocation (TET) oxidase (Matarese et al., 2011). This hydroxylation step ultimately leads to DNA demethylation and, therefore, often an increase in gene expression (Klug et al., 2013). Upon review of the literature, it appears that DNA/RNA oxidation damage is very region specific in post-mortem brain samples. For example, the 2010 study by Mustak et al. (2010) found increased single- and double-stranded breaks to genomic DNA in the parietal, temporal and occipital lobes, thalamus, cerebellum, hypothalamus, medulla, pons, and frontal cortex, but not in the hippocampus of bipolar patients. Another study that used post-mortem hippocampus suggested that damage in this region occurs predominantly in the cytoplasm of cells and thus affects RNA more than DNA (Che et al., 2010). One study, not included in this meta-analysis, measured the methylation patterns of monozygotic twins discordant for BD and found differences in four of the 10 explored regions (Kuratomi et al., 2008). Clearly, these oxidative modifications to DNA and RNA may impact the heritability of bipolar disorder and should be further investigated. The two products of oxidative protein damage included in this meta-analysis, 3-nitrotyrosine and protein carbonyl content, were not significant. 3-Nitrotyrosine is a product of protein nitrosative damage that occurs when peroxynitrite/carbon dioxide-derived radicals attack the hydroxyl group of tyrosine residues. Similarly, protein carbonylation occurs when peroxide or oxygen radicals attack amine groups in amino acid side-chains, often through a metal–cation catalyzed reaction. Oxidative damage to proteins in BD is likely very transient due to the cell's ability to remove these products and, therefore, it is vital to study patients at different stages of the illness and in different disease phases in order to fully determine the role protein damage may play in BD. Nitration of proteins is dependent on levels of nitric oxide and an increase of nitric oxide in BD patients was found in our meta-analysis. Nitric oxide is a widely used signaling molecule in the nervous system; however, it can react with the free oxygen radical, superoxide, to form the more unstable peroxynitrite. When the antioxidant system is overwhelmed, peroxynitrite and its derivatives may cause damage to cellular lipids, proteins, and nucleic acids. The increased nitric oxide levels in BD patients are discussed further in the previous meta-analysis (Andreazza et al., 2008) and no relevant papers have since been published. The antioxidant enzymes examined in the meta-analysis (GPx, SOD and catalase) did not show any overall significant changes in BD compared to healthy controls, however, it still remains a possibility that there are changes to larger antioxidant systems. SOD breaks the highly reactive and damaging superoxide anion (O2 ) into molecular oxygen (O2) and hydrogen peroxide (H2O2) through a copper-catalyzed redox reaction. The enzymes GPx and catalase can then remove hydrogen peroxide from cells through

Please cite this article as: Brown, N.C., et al., An updated meta-analysis of oxidative stress markers in bipolar disorder. Psychiatry Research (2014), http://dx.doi.org/10.1016/j.psychres.2014.04.005i

6

N.C. Brown et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 1. Forest plots of standardized mean differences and 95% confidence intervals for oxidative stress markers in patients with bipolar disorder compared to healthy controls. (1) 4-Hydroxynonenal. (2) Lipid hydroperoxides. (3) Single-stranded DNA breaks. (4) Double-stranded DNA breaks. Note: Mustak et al. (2010) used the same study population single-stranded DNA breaks and double-stranded DNA breaks and Andreazza et al. (2013) used the same study population for 4-hydroxynonenal and lipid hydroperoxides. In the meta-analysis each study was weighted as one, despite having two relevant measurements, to prevent one sample population from being overrepresented.

Please cite this article as: Brown, N.C., et al., An updated meta-analysis of oxidative stress markers in bipolar disorder. Psychiatry Research (2014), http://dx.doi.org/10.1016/j.psychres.2014.04.005i

N.C. Brown et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

further reduction. Two studies found that the ratio of SOD/ GPx þcatalase was increased in manic and depressed patients but not in euthymic patients (Andreazza et al., 2007a, 2007b). Consistent with these observations, the mood stabilizer lithium that is typically effective in BD patients significantly decreased the SOD/catalase ratio in healthy subjects (Khairova et al., 2012). Furthermore, a genetic study showed a significant interaction between SOD and GPx haplotypes which increased risk for BD (Fullerton et al., 2010). All these antioxidant enzymes (SOD, catalase, and GPx) form complicated relationships, and despite not being independently significant in this analysis, they may still play an important role in the overall pathophysiology of BD. The main limitations in this meta-analysis are the high degree of heterogeneity between studies and the small number of studies used in the analysis of PCC, RNA/DNA damage, and 3-NT. The sensitivity analysis also revealed that the results of the analysis for catalase, NO, 3-NT, and PCC are not very robust. For catalase, the one study removed sensitivity analysis showed that the lack of statistical significance was weak; removing the study by MachadoVieira et al. (2007) caused the negative effect size to become significant which would indicate that BD patients have a lower activity of peripheral catalase. One potential cause of this sensitivity could be the large drug-free population in the study by Machado-Vieira et al. (2007) and this result may indicate the effect of medication use on catalase activity. In the sensitivity analysis of NO, removal of the study by Ozcan et al. (2004) caused a drastic increase in effect size and significance level. There are no apparent differences in methods used since all included studies used the Greiss reaction to measure NO, however there is a lot of heterogeneity between patient samples which is likely the main contributor to the sensitivity of this analysis. The sensitivity in the results from 3-NT is likely due to the small number of studies and the sensitivity in PCC is likely due to heterogeneity between patient populations. The considerable between-study heterogeneity in this meta-analysis may be a reflection of the heterogeneity in BD itself. Few papers report length of illness, age of onset, number of mood episodes, illness phase, or BD phenotype; however, there is considerable evidence that these are important factors in the level of oxidative stress (Andreazza et al., 2007a, 2009; Kapczinski et al., 2011). There is also evidence that drug treatment may partly alleviate increased oxidative stress, which is unaccounted for in this meta-analysis due to few papers reporting drug status (Ozcan et al., 2004; Aliyazicioglu et al., 2007; Frey et al., 2007; MachadoVieira et al., 2007). Laboratory methodology is also another source of heterogeneity between studies in this meta-analysis. In addition, studies were conducted in different geographical locations which may add confounding factors such as diet. Due to these limitations, all interpretations of this meta-analysis must be considered cautiously. In conclusion, the results of this meta-analysis further confirm the presence of oxidative stress in BD patients. Compared to healthy controls, BD patients had higher levels of nitric oxide, more DNA and RNA damage, and increased lipid peroxidation. Determining the cause and effects of BD and its biological progression will lead to more effective treatments and care. Furthermore, through the use of very large studies or metaanalyses, a biomarker may be detected to aid in the diagnosis and treatment of BD. The large effect size and robustness of increased lipid peroxidation in BD patients shown in this metaanalysis make it a good candidate as a potential biomarker for BD. References Abdalla, D.S.P., Monteiro, H.P., Oliveira, J.A.C., Bechara, E.J.H., 1986. Activities of superoxide-dismutase and glutathione-peroxidase in schizophrenic and manicdepressive patients. Clinical Chemistry 32, 805–807.

7

Aliyazicioglu, R., Kural, B., Colak, M., Karahan, S.C., Ayvaz, S., Deger, O., 2007. Treatment with lithium, alone or in combination with olanzapine, relieves oxidative stress but increases atherogenic lipids in bipolar disorder. Tohoku Journal Experimental Medicine 213, 79–87. Andreazza, A.C., Cassini, C., Rosa, A.R., Leite, M.C., de Almeida, L.M.V., Nardin, P., Cunha, A.B.N., Cereser, K.M., Santin, A., Gottfried, C., Salvador, M., Kapczinski, F., Goncalves, C.A., 2007a. Serum S100B and antioxidant enzymes in bipolar patients. Journal of Psychiatric Research 41, 523–529. Andreazza, A.C., Frey, B.N., Erdtmann, B., Salvador, M., Rombaldi, F., Santin, A., Goncalves, C.A., Kapczinski, F., 2007b. DNA damage in bipolar disorder. Psychiatry Research 153, 27–32. Andreazza, A.C., Kapczinski, F., Kauer-Sant'Anna, M., Walz, J.C., Bond, D.J., Goncalves, C.A., Young, L.T., Yatham, L.N., 2009. 3-Nitrotyrosine and glutathione antioxidant system in patients in the early and late stages of bipolar disorder. Journal of Psychiatry and Neuroscience 34, 263–271. Andreazza, A.C., Kauer-Sant'Anna, M., Frey, B.N., Bond, D.J., Kapczinski, F., Young, L. T., Yatham, L.N., 2008. Oxidative stress markers in bipolar disorder: a metaanalysis. Journal of Affective Disorders 111, 135–144. Andreazza, A.C., Shao, L., Wang, J.F., Young, L.T., 2010. Mitochondrial complex I activity and oxidative damage to mitochondrial proteins in the prefrontal cortex of patients with bipolar disorder. Archives of General Psychiatry 67, 360–368. Andreazza, A.C., Wang, J.F., Salmasi, F., Shao, L., Young, L.T., 2013. Specific subcellular changes in oxidative stress in prefrontal cortex from patients with bipolar disorder. Journal of Neurochemistry. Banerjee, U., Dasgupta, A., Rout, J.K., Singh, O.P., 2012. Effects of lithium therapy on Na þ –K þ -ATPase activity and lipid peroxidation in bipolar disorder. Progress in Neuro-Psychopharmacology and Biological Psychiatry 37, 56–61. Benes, F.M., Walsh, J., Bhattacharyya, S., Sheth, A., Berretta, S., 2003. DNA fragmentation decreased in schizophrenia but not bipolar disorder. Archives of General Psychiatry 60, 359–364. Borenstein, M., Hedges, L., Higgins, J. and Rothstein, H., 2005. Comprehensive Meta Analysis, Version 2. Englewood, NJ: Biostat. Buttner, N., Bhattacharyya, S., Walsh, J., Benes, F.M., 2007. DNA fragmentation is increased in non-GABAergic neurons in bipolar disorder but not in schizophrenia. Schizophrenia Research 93, 33–41. Che, Y., Wang, J.F., Shao, L., Young, T., 2010. Oxidative damage to RNA but not DNA in the hippocampus of patients with major mental illness. Journal of Psychiatry and Neuroscience 35, 296–302. Clay, H.B., Sillivan, S., Konradi, C., 2011. Mitochondrial dysfunction and pathology in bipolar disorder and schizophrenia. International Journal of Developmental Neuroscience 29, 311–324. Craddock, N., Owen, M.J., 2005. The beginning of the end for the Kraepelinian dichotomy. British Journal of Psychiatry 186, 364–366. Dowlatshahi, D., MacQueen, G.M., Wang, J.F., Reiach, J.S., Young, L.T., 1999. G protein-coupled cyclic AMP signaling in postmortem brain of subjects with mood disorders: effects of diagnosis, suicide, and treatment at the time of death. Journal of Neurochemistry 73, 1121–1126. Egger, M., Smith, G.D., Schneider, M., Minder, C., 1997. Bias in meta-analysis detected by a simple, graphical test. British Medical Journal 315, 629–634. Frey, B.N., Andreazza, A.C., Houenou, J., Jamain, S., Goldstein, B.I., Frye, M.A., Leboyer, M., Berk, M., Malhi, G.S., Lopez-Jaramillo, C., Taylor, V.H., Dodd, S., Frangou, S., Hall, G.B., Fernandes, B.S., Kauer-Sant'Anna, M., Yatham, L.N., Kapczinski, F., Young, L.T., 2013. Biomarkers in bipolar disorder: a positional paper from the International Society for Bipolar Disorders Biomarkers Task Force. Australian and New Zealand Journal of Psychiatry 47 (4), 321–332, http: //dx.doi.org/10.1177/0004867413478217. Frey, B.N., Andreazza, A.C., Kunz, M., Gomes, F.A., Quevedo, J., Salvador, M., Goncalves, C.A., Kapczinski, F., 2007. Increased oxidative stress and DNA damage in bipolar disorder: a twin-case report. Progress in NeuroPsychopharmacology and Biological Psychiatry 31, 283–285. Fullerton, J.M., Tiwari, Y., Agahi, G., Heath, A., Berk, M., Mitchell, P.B., Schofield, P.R., 2010. Assessing oxidative pathway genes as risk factors for bipolar disorder. Bipolar Disorders 12, 550–556. Gawryluk, J.W., Wang, J.F., Andreazza, A.C., Shao, L., Young, L.T., 2011. Decreased levels of glutathione, the major brain antioxidant, in post-mortem prefrontal cortex from patients with psychiatric disorders. International Journal of Neuropsychopharmacology 14, 123–130. Gergerlioglu, H.S., Savas, H.A., Bulbul, F., Selek, S., Uz, E., Yumru, M., 2007. Changes in nitric oxide level and superoxide dismutase activity during antimanic treatment. Progress in Neuro-Psychopharmacology and Biological Psychiatry 31, 697–702. Gigante, A.D., Andreazza, A.C., Lafer, B., Yatham, L.N., Beasley, C.L., Young, L.T., 2011. Decreased mRNA expression of uncoupling protein 2, a mitochondrial proton transporter, in post-mortem prefrontal cortex from patients with bipolar disorder and schizophrenia. Neuroscience Letters 505, 47–51. Gu, X., Sun, J., Li, S., Wu, X., Li, L., 2013. Oxidative stress induces DNA demethylation and histone acetylation in SH-SY5Y cells: potential epigenetic mechanisms in gene transcription in Abeta production. Neurobiology of Aging 34, 1069–1079. Kapczinski, F., Dal-Pizzol, F., Teixeira, A.L., Magalhaes, P.V., Kauer-Sant'Anna, M., Klamt, F., Moreira, J.C., de Bittencourt Pasquali, M.A., Fries, G.R., Quevedo, J., Gama, C.S., Post, R., 2011. Peripheral biomarkers and illness activity in bipolar disorder. Journal of Psychiatric Research 45, 156–161. Khairova, R., Pawar, R., Salvadore, G., Juruena, M.F., de Sousa, R.T., Soeiro-de-Souza, M.G., Salvador, M., Zarate, C.A., Gattaz, W.F., Machado-Vieira, R., 2012. Effects of

Please cite this article as: Brown, N.C., et al., An updated meta-analysis of oxidative stress markers in bipolar disorder. Psychiatry Research (2014), http://dx.doi.org/10.1016/j.psychres.2014.04.005i

8

N.C. Brown et al. / Psychiatry Research ∎ (∎∎∎∎) ∎∎∎–∎∎∎

lithium on oxidative stress parameters in healthy subjects. Molecular Medicine Reports 5, 680–682. Kim, H.J., Soh, Y., Jang, J.H., Lee, J.S., Oh, Y.J., Surh, Y.J., 2001. Differential cell death induced by salsolinol with and without copper: possible role of reactive oxygen species. Molecular Pharmacology 60, 440–449. Klug, M., Schmidhofer, S., Gebhard, C., Andreesen, R., Rehli, M., 2013. 5-Hydroxymethylcytosine is an essential intermediate of active DNA demethylation processes in primary human monocytes. Genome Biology 14, R46. Kuloglu, M., Ustundag, B., Atmaca, M., Canatan, H., Tezcan, A.E., Cinkilinc, N., 2002. Lipid peroxidation and antioxidant enzyme levels in patients with schizophrenia and bipolar disorder. Cell Biochemistry and Function 20, 171–175. Kunz, M., Gama, C.S., Andreazza, A.C., Salvador, M., Cereser, K.M., Gomes, F.A., Belmonte-de-Abreu, P.S., Berk, M., Kapczinski, F., 2008. Elevated serum superoxide dismutase and thiobarbituric acid reactive substances in different phases of bipolar disorder and in schizophrenia. Progress in NeuroPsychopharmacology and Biological Psychiatry 32, 1677–1681. Kuratomi, G., Iwamoto, K., Bundo, M., Kusumi, I., Kato, N., Iwata, N., Ozaki, N., Kato, T., 2008. Aberrant DNA methylation associated with bipolar disorder identified from discordant monozygotic twins. Molecular Psychiatry 13, 429–441. Machado-Vieira, R., Andreazza, A.C., Viale, C.I., Zanatto, V., Cereser, V., Vargas, R.D., Kapczinski, F., Portela, L.V., Souza, D.O., Salvador, M., Gentil, V., 2007. Oxidative stress parameters in unmedicated and treated bipolar subjects during initial manic episode: a possible role for lithium antioxidant effects. Neuroscience Letters 421, 33–36. Magalhaes, P.V., Jansen, K., Pinheiro, R.T., Colpo, G.D., da Motta, L.L., Klamt, F., da Silva, R.A., Kapczinski, F., 2012. Peripheral oxidative damage in early-stage mood disorders: a nested population-based case-control study. Internation Journal of Neuropsychopharmacology 15, 1043–1050. Matarese, F., Carrillo-de Santa Pau, E., Stunnenberg, H.G., 2011. 5-Hydroxymethylcytosine: a new kid on the epigenetic block? Molecular Systems Biology 7, 562. Miklowitz, D.J., Schneck, C.D., Singh, M.K., Taylor, D.O., George, E.L., Cosgrove, V.E., Howe, M.E., Dickinson, L.M., Garber, J., Chang, K.K.D., 2013. Early Intervention for Symptomatic Youth at Risk for Bipolar Disorder: a randomized trial of family-focused therapy. Journal of the American Academy of Child and Adolescent Psychiatry 52, 121–131. Mustak, M.S., Hegde, M.L., Dinesh, A., Britton, G.B., Berrocal, R., Subba Rao, K., Shamasundar, N.M., Rao, K.S., Sathyanarayana Rao, T.S., 2010. Evidence of altered DNA integrity in the brain regions of suicidal victims of Bipolar Depression. Indian Journal of Psychiatry 52, 220–228. Ozcan, M.E., Gulec, M., Ozerol, E., Polat, R., Akyol, O., 2004. Antioxidant enzyme activities and oxidative stress in affective disorders. International Clinical Psychopharmacology 19, 89–95.

Perry, A., Tarrier, N., Morriss, R., McCarthy, E., Limb, K., 1999. Randomised controlled trial of efficacy of teaching patients with bipolar disorder to identify early symptoms of relapse and obtain treatment. British Medical Journal 318, 149–153. Phillips, M.L., Kupfer, D.J., 2013. Bipolar disorder diagnosis: challenges and future directions. Lancet 381, 1663–1671. Raffa, M., Barhoumi, S., Atig, F., Fendri, C., Kerkeni, A., Mechri, A., 2012. Reduced antioxidant defense systems in schizophrenia and bipolar I disorder. Progress in Neuro-Psychopharmacology and Biological Psychiatry 39, 371–375. Ranjekar, P.K., Hinge, A., Hegde, M.V., Ghate, M., Kale, A., Sitasawad, S., Wagh, U.V., Debsikdar, V.B., Mahadik, S.P., 2003. Decreased antioxidant enzymes and membrane essential polyunsaturated fatty acids in schizophrenic and bipolar mood disorder patients. Psychiatry Research 121, 109–122. Savas, H.A., Gergerlioglu, H.S., Armutcu, F., Herken, H., Yilmaz, H.R., Kocoglu, E., Selek, S., Tutkun, H., Zoroglu, S.S., Akyol, O., 2006. Elevated serum nitric oxide and superoxide dismutase in euthymic bipolar patients: impact of past episodes. World Journal of Biological Psychiatry 7, 51–55. Savas, H.A., Herken, H., Yurekli, M., Uz, E., Tutkun, H., Zoroglu, S.S., Ozen, M.E., Cengiz, B., Akyol, O., 2002. Possible role of nitric oxide and adrenomedullin in bipolar affective disorder. Neuropsychobiology 45, 57–61. Selek, S., Savas, H.A., Gergerlioglu, H.S., Bulbul, F., Uz, E., Yumru, M., 2008. The course of nitric oxide and superoxide dismutase during treatment of bipolar depressive episode. Journal of Affective Disorders 107, 89–94. Shan, X., Tashiro, H., Lin, C.L., 2003. The identification and characterization of oxidized RNAs in Alzheimer's disease. Journal of Neuroscience 23, 4913–4921. Soeiro-de-Souza, M.G., Andreazza, A.C., Carvalho, A.F., Machado-Vieira, R., Young, L. T., Moreno, R.A., 2013. Number of manic episodes is associated with elevated DNA oxidation in bipolar I disorder. International Journal of Neuropsychopharmacology, 1–8. Versace, A., Andreazza, A.C., Young, L.T., Fournier, J.C., Almeida, J.R., Stiffler, R.S., Lockovich, J.C., Aslam, H.A., Pollock, M.H., Park, H., Nimgaonkar, V.L., Kupfer, D. J., Phillips, M.L., 2013. Elevated serum measures of lipid peroxidation and abnormal prefrontal white matter in euthymic bipolar adults: toward peripheral biomarkers of bipolar disorder. Molecular Psychiatry. Wang, J.F., Shao, L., Sun, X., Young, L.T., 2009. Increased oxidative stress in the anterior cingulate cortex of subjects with bipolar disorder and schizophrenia. Bipolar Disorders 11, 523–529. Yanik, M., Vural, H., Tutkun, H., Zoroglu, S.S., Sava, H.A., Herken, H., Kocyigit, A., Keles, H., Akyol, O., 2004. The role of the arginine-nitric oxide pathway in the pathogenesis of bipolar affective disorder. European Archives of Psychiatry and Clinical Neuroscience 254, 43–47.

Please cite this article as: Brown, N.C., et al., An updated meta-analysis of oxidative stress markers in bipolar disorder. Psychiatry Research (2014), http://dx.doi.org/10.1016/j.psychres.2014.04.005i