Molecular Psychiatry (2006) 11, 547–556 & 2006 Nature Publishing Group All rights reserved 1359-4184/06 $30.00 www.nature.com/mp FEATURE REVIEW The role of A2 in schizophrenia MH Law1, RGH Cotton1 and GE Berger1,2 1Genomic Disorders Research Centre, Melbourne, VI, Australia and 2ORYGEN Research Centre, Melbourne, VI, Australia

A range of neurotransmitter systems have been implicated in the pathogenesis of schizophrenia based on the antidopaminergic activities of antipsychotic medications, and chemicals that can induce psychotic-like symptoms, such as ketamine or PCP. Such neurotransmitter systems often mediate their cellular response via G-protein-coupled release of (AA) via the activation of phospholipases A2 (PLA2s). The interaction of three PLA2s are important for the regulation of the release of AA – A2 Group 2 A, Group 4A and phospholipase A2 Group 6A. variations of these three key have been associated with schizophrenia with conflicting results. Preclinical data suggest that the activity of these three enzymes are associated with monoaminergic neurotransmission, and may contribute to the differential efficacy of antipsychotic medications, as well as other biological changes thought to underlie schizophrenia, such as altered neurodevelopment and synaptic remodelling. We review the evidence and discuss the potential roles of these three key enzymes for schizophrenia with particular emphasis on published association studies. Molecular Psychiatry (2006) 11, 547–556. doi:10.1038/sj.mp.4001819; published online 4 April 2006 Keywords: review; schizophrenia/genetics; PLA2/genetics; arachidonic acid; common diseases; PLA2/schizophrenia

Introduction (PLA2GVIA, PLA2G6A), dopamine, serotonin, G- protein-coupled receptor, and phospho- Neurotransmitters, such as dopamine, serotonin and lipids. Crosschecking of references was used to glutamate often mediate their cellular response via identify further papers of interest. G-protein-coupled activation of second messengers. One such common second messenger is arachidonic acid (AA) that is released from membrane phospho- PLA2s – naming conventions lipids via the G-protein-coupled activation of phos- pholipase A2 (PLA2).1 Phospholipase A2 activity and When analysing PLA2 studies it is essential to AA-based signalling is also important for normal properly define which enzymes are being examined. brain development and synaptic functioning.2–5 This Phospholipases A2 are a large family of enzymes that review outlines how PLA2s, in particular PLA2G2A, specifically deacylate fatty acids from the sterospeci- PLA2G4A and PLA2G6A, may be relevant for the fically numbered second carbon atom (sn2, thus understanding of schizophrenia pathogenesis and its PLA2) of the triglyceride backbone of membrane treatments.6 , producing a free fatty acid and a lyso- (Figure 1).7 Historically, PLA2s were named by activity location, that is pancreatic, Method cytosolic or secretory. Later naming systems also included calcium requirements, although this can be Medline and Google Scholar (www.scholar.google.- misleading. Some calcium-dependent PLA2s require com) were used to identify articles from 1970 to 2005 calcium for catalytic activity, whereas others are using the following key words: schizophrenia, essen- constitutively active and calcium promotes binding tial fatty acids (EFA), polyunsaturated fatty acid to phospholipid membranes. The more structured (PUFA), AA, PLA2, phospholipase A2 Group 4A classification system based on genetic relationships (PLA2GIVA, PLA2G4A), phospholipase A2 Group 2A will be used in this review.7 In this system each PLA2 (PLA2GIIA, PLA2G2A), phospholipase A2 Group 6A is assigned to one of (currently) 11 groups; each group may contain multiple homologues further assigned a Correspondence: Dr G Berger, ORYGEN Research Centre/ORYGEN letter. For example, phospholipase A2 Group 4A Youth Health, 35 Poplar rd, Parkville, Victoria 3052, Australia. E-mail: [email protected] (PLA2G4A) and phospholipase A2 Group 4B Received 14 December 2005; accepted 30 January 2006; published (PLA2G4B) are evolutionarily related homologues (A online 4 April 2006 and B) within Group 4 (G4). PLA2s and schizophrenia MH Law et al 548

Figure 1 De-acylation of arachidonic acid (AA) from a phospholipid by the activity of a PLA2 . Figure 2 Arachidonic acid release in response to receptor activation. (a) Initial or early stage AA release requires PLA2s and AA recruitment of PLA2G4A. (b) To promote and maintain AA release PLA2GA recruits PLA2G2A. PLA2s regulate both the availability of free lysopho- spholipids for AA membrane incorporation, and the cleavage of AA from membranes for signalling 8,9 G2A, PLA2G4A and PLA2G6A in both brain matura- (Figures 1 and 2). The polyunsaturated fatty acid tional processes and neurotransmission. (PUFA) AA itself is a potent signalling molecule, and also the precursor for a range of messengers (eicosa- 10,11 Neurotransmission noids) necessary for normal neuronal function. To The receptors for neurotransmitter systems impli- limit aberrant signalling AA is usually bound to cell cated in schizophrenia activate downstream re- membranes phospholipids. In addition to creating a sponses via a family of guanosine triphosphate reservoir for later release of AA in response to binding proteins (G-proteins), which in turn appropriate signals, the ratio of saturated to unsatu- activates PLA2G4A-mediated release of AA 26,27 rated fatty acids (such as AA) also defines membrane and recruit PLA2G2A to further enhance signalling fluidity, which alters the activity of membrane bound 5,11,17,28 12,13 (Figure 2). The colocalisation of monoaminer- proteins including neurotransmitter receptors. gic receptors with PLA2G4A, PLA2G2A and eicosa- PLA2G4A has a 50-fold preference for phospholi- noid-synthesising enzymes further emphasise their pids containing AA over any other PUFA, whereas importance for monoaminergic neurotransmis- PLA2G6A and PLA2G2A show no fatty acid pre- 15,29–31 14 sion. The proper functioning and availability ference. PLA2G4A initiates AA release and eicosa- of these PLA2s are therefore crucial for G-protein- noid production, whereas PLA2G2A enhances the coupled neurotransmission. production of AA (Figure 2).15–18 Mice with the for PLA2G4A and PLA2G2A nullified (knocked out) Memory formation (long-term potentiation) do not produce eicosanoids, have reduced incorpora- Long-term potentiation (LTP), the proposed mechan- tion rates of free AA into cell membranes, yet normal 19,20 ism for memory formation, is associated with AA levels. G-protein-coupled glutamate receptor (AMPAR) acti- PLA2G6A cleavage of phospholipids is primarily vation and release of AA.32–35 PLA2 inhibitors block for the purpose of cellular membrane remodelling, by AMPAR-associated LTP, while increasing PLA2 activ- altering phospholipids/fatty acid ratios and modify- 34 21,22 ity enhances LTP. PLA2 enhancement of LTP is ing membrane fluidity. Loss of PLA2G6A function halted by sequestering free AA, but not by removing leads to significant reductions in the amount of AA 34,36 23 calcium or blocking production. Gen- incorporated into the cell membrane. eral inhibition of brain PLA2 activity, and particularly specific inhibition of PLA2G6A activity decreased PLA2 in neurobiology memory formation in rats and reduced membrane fluidity, revealing the requirement of PLA2G6A in The neurodevelopmental hypothesis and the neuro- memory formation.37 transmitter theories of schizophrenia propose that abnormal genetic and environmental processes inter- Brain maturation, cortical development and synaptic fere with normal brain maturation and result in remodelling dysfunctional monoaminergic neurotransmission as- Brain maturation, cortical development and synaptic sociated with the phenotype of schizophrenia.24,25 remodelling involves removal of excess brain cells or The following evidence outlines the roles of PLA2- parts of cells such as dendrites or axons. Key regulator

Molecular Psychiatry PLA2s and schizophrenia MH Law et al 549 proteins involved in these processes are the caspases, tion between peripheral AA levels and brain phos- in particular caspase 3. Caspases manage apoptotic pholipid metabolism. cellular metabolism via the cleavage of certain proteins such as PLA2G4A and PLA2G6A into Niacin insensitivity – a marker for an abnormal AA alternative forms, altering their behaviour.22 Cleaved metabolism PLA2G4A proteins have a dominant-negative func- The vitamin nicotinic acid (niacin) induces the tion, halting cellular PLA2G4A activity. Truncation of release of inflammatory eicosanoids such as prosta- PLA2G6A upregulates its activity, allowing rapid glandin D2, which are produced from AA metabo- remodelling of the cellular membrane and generation lism.60 Depending on the method/definitions used, of free AA required for apoptosis.38 The remodelled 40–80% of patients with psychotic disorders and membranes generates phagocytic attraction signals to 10% of controls show an impaired sensitivity to allow correct disposal of the cell remnants created niacin.61–64 Niacin sensitivity is also impaired in resulting from the apoptotic process.22,39–41 healthy first degree relatives of probands, suggesting While not required within apoptotic neurons, it is a marker for underlying genetic risk.65 Niacin complete loss of PLA2G4A and PLA2G2A abolishes insensitivity remains associated with schizophrenia following central nervous system (CNS) after exclusion of environmental factors modifying injury, and reduces neuronal apoptosis.19 Studies inflammation.66–69 Niacin insensitive patients have have shown how PLA2G4A is expressed in adjacent significantly lower AA levels when compared to glial cells postinjury, but not within apoptotic niacin sensitive patients.70 Prospective monitoring neurons.42,43 Together this indicates a primary role revealed conversion from niacin insensitive to sensi- for PLA2G6A activity in neuronal remodelling by tive was associated with restoration of AA levels and apoptosis, while PLA2G4A and PLA2G2A act periph- improvement in symptomatology.70,71 Interestingly, erally to manage inflammation.44,45 the abolishment of AA release and resultant eicosa- noid production reported previously in mice with the genes for PLA2G4A and PLA2G2A knocked out is PLA2s in schizophrenia applicable to niacin insensitivity.8,20 Niacin sensitiv- Horrobin46,47 suggested that schizophrenia might be a ity could serve as a readily measurable biological deficiency disease and consequently marker of altered AA metabolism indicative of altered formulated the Membrane Phospholipid Hypothesis PLA2 activity. of schizophrenia. This hypothesis can be examined in the light of the combined functions of these three Direct evidence of altered PLA2 activity in PLA2s. schizophrenia PLA2s and reduced PUFAs in schizophrenia To evaluate the Membrane Hypothesis of Schizophre- Red blood cell (RBC) and CNS cell membranes of nia a number of studies have measured PLA2 activity medicated and drug naı¨ve patients with schizophre- in patients with schizophrenia. Most studies measur- nia have significantly lower levels of PUFAs, in ing PLA2 activity in psychotic disorders have particular AA 48–52 suggesting that this is not only used either a radiometric or a fluorometric method a drug effect. The change in AA levels correlates (Table 1). The fluorometric method measures all to decreased membrane fluidity and to psychosis calcium-independent PLA2s, including PLA2G6A, severity12,13,49,51,53 with some conflicting results.54 AA whereas the radiometric method detects the activity reductions could not solely be explained by environ- of calcium-dependent PLA2s, encompassing PLA2- mental factors such as smoking, diet or medication.55–57 G4A and PLA2G2A.72 Additionally, a recent study has As loss of PLA2G6A function results in reduced cell reported on a direct measure of the PLA2G4A protein, membrane AA levels, variations in this gene may be and found increased levels of PLA2G4A in schizo- responsible for these results.23 phrenia.73 The method used by Noponen et al.74 to detect increased PLA2 activity was novel, with the PLA2s and in vivo phospholipid metabolism target PLA2s unconfirmed, and thus will be excluded 31-Phosphorus magnetic resonance imaging (31P from further discussion. Owing to the limitations in MRS) allows measurement of membrane phospholi- methodology activity studies will be examined in pid dynamics in the living brain.45 31P MRS of drug- terms of PLA2 calcium requirements. naı¨ve first episode schizophrenia patients reveals significantly increased breakdown and reduced crea- Calcium-independent PLA2 activity tion of cellular membranes in never treated patients Increased calcium-independent PLA2 activity in with schizophrenia.58,59 Interestingly, a significant patients with schizophrenia has been replicated both relationship between reduced peripheral RBC AA peripherally and in post-mortem human brain tis- levels and in vivo brain membrane breakdown sue.62,72,75–78 Higher calcium-independent PLA2 ac- products has been demonstrated in patients with tivity has been correlated to greater severity of schizophrenia.48 Changes in the activity of these psychopathology72 though other positive studies PLA2s, which are important for apoptosis and have not seen this association.62,75 Katila et al.79 membrane maintenance would explain the correla- reported normal calcium-independent PLA2 activity

Molecular Psychiatry PLA2s and schizophrenia MH Law et al 550 Table 1 Biochemical analysis of phospholipase activity

Article Diagnosis and Medicationb Controls Analysis Source Resultc numbersa

Gattaz et al.75 SZ 20, 6 Psych 8 DN, AF 21 Flourometric Serum m CI Gattaz et al.76 SZ 14, Psych, 8 6 DN, AF 20 Flourometric Serum m CI Albers et al.81 SZ/SZF 10, Psych 25 All DN 10 Radiometric Serum = CD Gattaz et al.80 SZ 31, Psych 31 Y, AF 31 Radiometric Platelet m CD Hudson et al.91 SZ 23 Y, AF 30 Niacin sensitivity, Serum k CD Radiometric Katila et al.79 SZ 34, Psych, 28 11 DN, AF 62 Flourometric Plasma = CI Ross et al.72 SZ 24 Y 33 Flourometric, Serum m CI, = CD Radiometric Ross et al.78 SZ 10, Bipolar 8 Y, AF 12 Flourometric Brain m CI, k CD Tavares et al.62 SZ 38 19 DN, AF 28 Niacin sensitivity, Serum m CI Flourometric Lasch et al.77 SZ/SZF 26 11 DN, AF 26 Flourometric Serum m CI Macdonald et al.73 SZ 29 Y, AF 27 Protein measurement RBC m PLA2G4A

aSZ, schizophrenia, SZF, schizophreniform, Psych., psychiatric Illnesses other than schizophrenia. bY, yes, DN, drug naı¨ve, AF, medication state accounted for during statistical analysis. cCI, calcium-independent PLA2 activity; CD, calcium-dependent PLA2 activity.

in schizophrenia, though Lasch et al.77 suggest this PLA2 activity of niacin insensitive patients did not may have been a methodological problem. differ from controls.64

Calcium-dependent PLA2 An initial report of increased peripheral calcium- Genetic profiles of PLA2G2A, PLA2G4A and dependent PLA2 activity in schizophrenia has not PLA2G6A been replicated.64,72,80,81 Analysis of post-mortem As we are reviewing the possible involvement of brain tissue found calcium-dependent PLA2 activity genetic variation in PLA2G2A, PLA2G4A and PLA2- was decreased, rather than increased.78 G6A, it is worthwhile briefly summarising their gene structure in Table 2. Regions implicated in schizo- Could response to antipsychotic medication be due to phrenia in genomewide linkage studies (GWL) are interactions with PLA2s? recorded here, divided by the maximum 10 cM range Eight weeks of antipsychotic medication reduced the suggested as evidence of gene–signal relationship.82– high calcium-independent PLA2 activity of patients 84 A published variant is defined as having minimum with schizophrenia, replicating a result seen in verification when it has either been reported sepa- the original study of calcium-independent PLA2 rately at least twice, or has been genotyped in a large activity.62,75 Three positive studies, and one negative enough population to ensure it is truly poly- study measuring calcium-independent activity morphic.85 Untranslated regions are exons that are utilised a portion of drug naı¨ve patients.62,75,76,79 transcribed to the messenger RNA (mRNA) and not Statistical analysis has found no correlation between translated into the protein while coding variants are medication state and calcium-dependent PLA2 activ- in exons translated into amino acids. A synonymous ity.64,72,80 A small population of drug naı¨ve patients coding variant does not change the amino acid coded had normal calcium-dependent PLA2 activity.81 for, while non-synonymous variants alter the result- ing protein structure. Niacin insensitivity and PLA2 activity The increased calcium-independent PLA2 activity associated with schizophrenia is greater still in the Association analysis of PLA2s and schizophrenia niacin insensitive sub group.62 As reported, 8 weeks of medication reduced the abnormally high calcium- Interest in the Membrane Phospholipid Hypothesis independent PLA2 activity. Those patients who has fuelled association studies between PLA2 genes converted to niacin sensitive showed further normal- and schizophrenia, as summarised in Table 3. isation of calcium-independent PLA2 activity.62 Although Hudson et al.64 initially found no differ- PLA2G4A ences in calcium-dependent PLA2 activity, niacin Deletion of a promotor CA microsatellite (numerous sensitive (normal) patients had significantly lower repeats of the bases cytosine paired with an adenine) calcium-dependent PLA2 activity compared to con- increases PLA2G4A mRNA expression by 50%.86 One trols and insensitive patients. Calcium-dependent CA allele, (140 repeats) has a frequency of 0.96,

Molecular Psychiatry PLA2s and schizophrenia MH Law et al 551 Table 2 Available data for PLA2 genes on dbSNP85

Gene Names Locus Exons GWL < 10 cM GWL > 10 cM Variants/ Coding/ Untranslated/ Verified Verified Verified

PLA2G2A sPLA2 1q35 6 No 1q42 50/40 8/4 21/17 PLA2G4A cPLA2 1q25 18 No 1q21–23 726/407 5/5 63/27 PLA2G6A iPLA2 22q13.1 16 No 22q11–12 494/299 10/8 91/47

Abbreviations: GWL, genomewide linkage studies; PLA2, phospholipases A2; PLA2G2A, phospholipase A2 Group 2 A; PLA2G4A, phospholipase A2 Group 4 A; PLA2G6A, phospholipase A2 Group 6 A.

Table 3 Genetic analysis of PLA2 genes in psychotic disorders

Article Modela Case no. Ethnicityb,c Control no. Analysis methodd Associated?

Hudson et al.90 CC 65 Ca/C, USA/C 65 Poly(A) Yes Hudson et al.91 T 44 Ca/C, R/C 88 Poly(A) Yes Hudson et al.91 CC 20 Ca/C, USA/C 20 Poly(A), Niacin sensitivity Yes Price et al.93 CC 58 S/C 56 Poly(A) No Doris et al.54 CC 35 NR 40 Poly(A) No Wei et al.95 CC 193 NR/C 101 BanINo Wei et al.95 T 50 NR/C 100 BanIYes Peet et al.94 CC 36 I 27 BanIYes Ramchand et al.89 CC 52 I 48 CA No Chowdari et al.100 T 86 USA/C, USA/Af 130 Poly(A), BanINo Chowdari et al.100 CC 86 USA/C, USA/Af 94 Poly(A), BanINo Chowdari et al.100 T 159 I 283 BanINo Frieboes et al.92 T 328 E/C. M, T/A 426 Poly(A) No Junqueira.et al.101 CC 240 B 312 BanINo Junqueira et al.101 CC 240 B 312 AvrII Yes Wei and Hemmings96 T 118 UK/C 236 BanIYes Yu et al.99 T 168 M/M 336 BanINo Pae et al.97 CC 97 K/K 117 BanIYes Tao et al.102 T 240 M/M 480 BanINo Wei and Hemmings98 T 132 UK/C 262 BanIYes aCC, case/control test; T, transmission disequilibrium test. bNR, not recorded in paper. cEthnicity is expressed as country of origin/ethnicity. (a) Country of origin shorthand: USA = United States of America, Ca = Canada, S = Scotland, M = China, T = Taiwan, E = Europe, R = Italy, P = Poland, M = China, I = India, K = Korea, UK = - United Kingdom. (b) Ethnicity shorthand – C = Caucasian, A = Asian, I = Indian, Af = African, B = Brazilian, M = Chinese, K = Korean. dPoly(A), adenine repeat PLA2G4A promotor variant, CA, CA microsatellite in promotor of PLA2G4A, BanI, rs10798059 in first intron of PLA2G4A, AvrII, rs4375 in fourth intron of PLA2G6A. Abbreviation: PLA2, phospholipases A2.

Table 4 Range of Poly(A) alleles

Studies Method Alleles No. (A) repeats Result

Hudson et al.90,91 Electrophoresis vs known allele 10 41–60 Yes Price et al.93 Electrophoresis vs absolute size marker and sequencing. 24 22–57 No Doris et al.54 Electrophoresis vs absolute size marker 10 136–156 No Chowdari et al.100 Direct sequencing 26 17–52 No Frieboes et al.92 Electrophoresis vs absolute size marker 14 ? No

Abbreviation: Poly(A), PLA2G4A promotor adenine repeat variant. requiring large sample sizes to achieve sufficient Hudson et al.90,91 reported the PLA2G4A promotor statistical power to detect association.87,88 The one adenine repeat variant (Poly(A)) (an adenine repeat) reported study of the CA microsatellite found no occurred as 10 size alleles, termed A1–A10, ranging significant association with schizophrenia (Table 4).89 from 41 to 60 repeats in length. The small differences

Molecular Psychiatry PLA2s and schizophrenia MH Law et al 552 between alleles, except for five bases between A6 (d) Increased calcium-independent PLA2 activity has and A7 made genotyping difficult, so analysis been replicated in both peripheral and CNS tissue, was performed using A1–A6 vs A7–A10 groupings. and in drug naı¨ve patients (Table 1).62,72,76,78 Association between schizophrenia and the A7–A10 (e) Higher calcium-independent PLA2 activity corre- group was detected under a case/control and a lates significantly with greater severity of psycho- transmission model.90,91 Additionally, niacin insensi- pathology, suggesting a direct relationship tivity was greater in patients with the A7–A10.91 Later between the degree of PLA2G6A activity and methods detected completely differing alleles, and disease progression.72 found association between schizophrenia and Poly(A) (f) Though numbers were small, the work of Tavares (Tables 3 and 4).54,92,93 et al.62 supports the concept of antipsychotic Alternative studies have used an adenine (A1, not efficacy through modulating PLA2 activity. cut by PLA2G4A, rs10798059, (g) Niacin sensitivity requires PLA2G4A- and PLA2- intron one adenine/guanine variant (BanI)) or guanine G2A-mediated AA release from cellular mem- (A2, cut by BanI) variant in the first intron of branes. Membrane AA levels are maintained by PLA2G4A, rs10798059 (BanI).94 The A2 allele of PLA2G6A, and the separate associations of BanI has been associated with schizophrenia under reduced AA levels and increased calcium-inde- both case/control and transmission disequilibrium pendent activity with niacin insensitivity is test (TDT) models (Table 3).94–98 Five additional intriguing.20,23 studies have not replicated any association (h) The strengthening of association between schizo- between the BanI locus and schizophrenia.96,99–102 phrenia and a variant in the PLA2G4A gene after Testing of markers across the region stratification by niacin sensitivity is promising, containing PLA2G4A make it unlikely that associa- but not replicated.90 Yet increased calcium-depen- tion is due to linkage disequilibrium (LD) with a dent PLA2 activity, which includes PLA2G4A, nearby gene.96 was not associated with niacin insensitivity. Rather it was the patient group with normal PLA2G2A and PLA2G6A sensitivity to niacin that possessed increased A variant in the fourth intron of PLA2G6A rs4375 has calcium-dependent activity.64 Low participant been used to test for association with schizophrenia. numbers may explain this unexpected result, or Rs4375 is either a cytosine base that creates a cut site as reported by the authors, the niacin insensitive for the AvrIII restriction enzyme or a noncut thymine. group had the greatest activity variation, with a Both a case/control and a TDT model in this Brazilian standard deviation twice that of the other population indicated association between the cy- groups.64 A repeat of this experiment in a larger steine allele and schizophrenia.101 Although studies population using direct measures of PLA2G4A have used alternative PLA2s, such as PLA2G1B, there protein level is required before further conclu- are no published association studies of PLA2G2A in sions can be made. psychotic illnesses.90,92,101 That there are no GWL peaks for schizophrenia within the suggested 10 cM of PLA2 loci is not absolute proof of noninvolvement (Table 2).104 Discussion The population needed to detect GWL increases dramatically as the relative risk (RR) conferred by The three PLA2s regulate the levels of bound and free variants at that loci decrease.87 The RR estimated AA that has key functions in processes such as brain from replicated associations of variants in dopamine maturation, memory formation and synaptic remodel- and serotonin receptors as well as the neuregulin 1 ling all processes suggested to be important for gene range from 1.2 to 2.4.105,106 If the RR of PLA2 schizophrenia.5,6,11,103 An increasing body of evidence variants is two or more, then thousands to tens suggests that AA is altered in schizophrenia in a of thousands of families are required to detect GWL, causative manner: and if the RR is lower millions of families (a) In vivo 31P-MRS and post-mortem measurements are needed.87 All that can be concluded is that of AA levels of patients with schizophrenia these PLA2 genes, like neuregulin or dopamine confirm that AA alterations are present in the receptors, do not contribute a high (4 þ ) risk for brain and peripheral tissues.48,52 schizophrenia. (b) AA changes are present in drug naı¨ve patients The different allele profiles found for the Poly(A) indicates that the relationship is independent of site in the PLA2G4A promotor may be due to method drug treatment and associated with the onset of of size detection, as there was the greatest consistency disease.49–51 in allele profiles when direct sequencing was used (c) Chronic haloperidol administration in rats re- (Table 4). Interestingly the Poly(A) and BanI variants duced AA levels specifically in dopaminergic are in significant LD, supporting the concept that neurons, possibly due to interactions with dopa- association is due to an adjacent causative var- mine receptors and their downstream targets, the iant(s).100 Although the absence of the repressive CA PLA2 enzymes.57 microsatellite has a functional effect, it is not clear if a

Molecular Psychiatry PLA2s and schizophrenia MH Law et al 553 change in CA repeat numbers also modifies expres- membrane saturated:unsaturated fats ratio, altering sion. Genotyping in sufficiently large populations to neurotransmitter receptors dynamics.12,13 Thus, PLA2 overcome the high frequency of one of the alleles genetic variants with altered activity have implica- would be interesting. tions in both the release and the reception of neuronal The association between schizophrenia and either signals. Environmental modification of AA metabo- of PLA2G4A variants in populations of different lism could, in a background of deleterious PLA2 ethnicity and alternative statistical models is promis- variants, lead to a run away failure of brain fun- ing (Table 4). Yet many studies have not replicated ction and development. Detection of these variants these associations. There are a number of possible will ultimately require models that take into account reasons for this inconsistency. both rare and common risk variants. If sampling of variation occurs in a large study population, associa- (a) The associations so far may have been false tion tests using the variants found should allow positives due to population heterogeneity. evaluation of both the CDCV and CDRV hypotheses. (b) The tested variants may not be in perfect LD with In summary, the PLA2G4A gene has been asso- the causative variant(s), increasing false negatives. ciated with schizophrenia and reduced niacin sensi- Using a population large enough to detect associa- tivity. There are as yet to be replicated studies tion under varying models of LD and RR, and utilising indicating PLA2G4A protein levels and specific niacin sensitivity or RBC AA levels to reduce activity are increased in, and the PLA2G6A gene is heterogeneity should limit these problems. associated with, schizophrenia. That mice null for Additionally, replication inconsistency may be due PLA2G2A and PLA2G4A genes are viable could to the underlying spectrum of disease variants. The explain the relative high frequency of schizophrenia, association studies discussed have followed the as negative variants in these genes would be under common disease, common variant (CDCV) hypothesis reduced selective pressure due to compensation by that for each ‘disease’ gene there are one or few risk other PLA2s. The lack of normal inflammation in variants with minor allele frequency > 10%.87,107 The these mice indicates that this compensation is alternative common disease, rare variant (CDRV) incomplete, and such PLA2 activity shortfalls in hypothesis proposes that there a large pool of rare humans may result in schizophrenia. (minor allele frequency < 10%) or even unique risk variants.108–110 Interestingly, association between a common variant and a disease can be due to LD with Abbreviations high penetrance rare variants present in only a few of AA, arachidonic acid; AMPAR, glutamate receptor; the total study population.111 BanI, PLA2G4A, rs10798059, intron one adenine/ It is likely that both paradigms contribute, as a guanine variant; bp, DNA base pairs; CDCV, common spectrum of common and rare deleterious variants has disease common variant; CDRV, common disease rare been seen in both Mendelian and complex disor- variant; CNS, central nervous system; EFA, essential ders.112–114 However most studies do not characterise fatty acids; G-proteins, guanosine triphosphate bind- the rare variants present in the study population nor ing proteins; GWL, genomewide linkage studies; LD, have the statistical power to test them.87 linkage disequilibrium; LTP, long-term potentiation; mRNA, messenger RNA; PLA2, phospholipase A2; Conclusion PLA2GIVA, PLA2G4A, phospholipase A2 Group 4 A; PLA2GIIA, PLA2G2A, phospholipase A2 Group 2 A; The collected evidence discussed point towards the PLA2GVIA, PLA2G6A, phospholipase A2 Group 6 A; involvement of AA and PLA2s in the pathogenesis of Poly(A), PLA2G4A promotor adenine repeat variant; schizophrenia. The reported aberrations in AA levels, PUFA, polyunsaturated fatty acid; 31P MRS, 31- PLA2 enzyme activities and in vivo brain membrane phosphorus magnetic resonance imaging; RBC, red breakdown specific to schizophrenia are biological blood cells; TDT, transmission disequilibrium test. processes mediated by PLA2G6A, PLA2G4A and PLA2G2A. As AA levels and metabolism is tightly associated with the activity of PLA2G4A, PLA2G2A Acknowledgments and PLA2G6A, measurement of AA may reduce We wish to thank R Purcell for editing and proof- schizophrenia heterogeneity (‘a PLA2s-AA endophe- reading assistance. notype’). The topical Niacin skin flush test or RBC AA level measurements have the potential to identify Web Resources 45,61,64,115,116 such an endophenotype. The URLs for data presented herein is as follows As reported, these three PLA2 enzymes lie down- dbSNP, http://www.ncbi.nlm.nih.gov/SNP/ stream of activation of neurotransmitter pathways implicated in schizophrenia, such as dopaminergic, serotoninergic or glutamatergic systems. It is plausi- References ble that antipsychotic efficacy can be modulated by 1 Piomelli D, Pilon C, Giros B, Sokoloff P, Martres MP, Schwartz JC. PLA2 genetic variants. Feedback may occur, as Dopamine activation of the arachidonic acid cascade as a basis for changes in AA release by PLA2s will alter the D1/D2 receptor synergism. Nature 1991; 353: 164–167.

Molecular Psychiatry PLA2s and schizophrenia MH Law et al 554 2 Meiri KF, Saffell JL, Walsh FS, Doherty P. Neurite outgrowth 20 Rosenberger TA, Villacreses NE, Contreras MA, Bonventre JV, stimulated by neural cell adhesion molecules requires growth- Rapoport SI. Brain lipid metabolism in the cPLA(2) knockout associated protein-43 (GAP-43) function and is associated with mouse. J Lipid Res 2003; 44: 109–117. GAP-43 phosphorylation in growth cones. J Neurosci 1998; 18: 21 Balsinde J, Dennis EA. Function of calcium-independent phos- 10429–10437. pholipase A2 in arachidonic acid metabolism in P388D1 3 Negre-Aminou P, Pfenninger KH. Arachidonic acid turnover and macrophages. Adv Exp Med Biol 1997; 407: 99–103. phospholipase A2 activity in neuronal growth cones. J Neuro- 22 Atsumi G, Murakami M, Kojima K, Hadano A, Tajima M, Kudo I. chem 1993; 60: 1126–1136. Distinct roles of two intracellular phospholipase A2s in fatty acid 4 St-Gelais F, Menard C, Congar P, Trudeau LE, Massicotte G. release in the cell death pathway. Proteolytic fragment of type Postsynaptic injection of calcium-independent phospholipase A2 IVA cytosolic phospholipase A2alpha inhibits stimulus-induced inhibitors selectively increases AMPA receptor-mediated synap- arachidonate release, whereas that of type VI Ca2 þ -independent tic transmission. Hippocampus 2004; 14: 319–325. phospholipase A2 augments spontaneous fatty acid release. J Biol 5 Garcia MC, Kim HY. Mobilization of arachidonate and docosa- Chem 2000; 275: 18248–18258. hexaenoate by stimulation of the 5-HT2A receptor in rat C6 23 Balsinde J, Balboa MA, Dennis EA. Antisense inhibition of group glioma cells. Brain Res 1997; 768: 43–48. VI Ca2 þ -independent phospholipase A2 blocks phospholipid 6 Mounier CM, Ghomashchi F, Lindsay MR, James S, Singer AG, fatty acid remodeling in murine P388D1 macrophages. J Biol Parton RG et al. Arachidonic acid release from mammalian cells Chem 1997; 272: 29317–29321. transfected with human groups IIA and X secreted phospholipase 24 Weinberger DR. Implications of normal brain development for the A(2) occurs predominantly during the secretory process and with pathogenesis of schizophrenia. Arch Gen Psychiatry 1987; 44: the involvement of cytosolic phospholipase A(2)-alpha. J Biol 660–669. Chem 2004; 279: 25024–25038. 25 Murray RM, Lewis SW. Is schizophrenia a neurodevelopmental 7 Six DA, Dennis EA. The expanding superfamily of phospholipase disorder? Br Med J (Clin Res Ed) 1987; 295: 681–682. A(2) enzymes: classification and characterization. Biochimica et 26 Gilman AG. G proteins: transducers of receptor-generated signals. Biophysica Acta 2000; 1488: 1–19. Annu Rev Biochem 1987; 56: 615–649. 8 Balsinde J, Dennis EA. Function and inhibition of intracellular 27 Clapham DE, Neer EJ. G protein beta gamma subunits. Annu Rev calcium-independent phospholipase A2. J Biol Chem 1997; 272: Pharmacol Toxicol 1997; 37: 167–203. 16069–16072. 28 Balsinde J, Balboa MA, Dennis EA. Functional coupling between 9 Winstead MV, Balsinde J, Dennis EA. Calcium-independent secretory phospholipase A2 and cyclooxygenase-2 and its phospholipase A(2): structure and function. Biochimica et regulation by cytosolic group IV phospholipase A2. Proc Natl Biophysica Acta 2000; 1488: 28–39. Acad Sci USA 1998; 95: 7951–7956. 10 Shimizu T, Wolfe LS. Arachidonic acid cascade and signal 29 Murakami M, Nakatani Y, Kuwata H, Kudo I. Cellular compo- transduction. J Neurochem 1990; 55: 1–15. nents that functionally interact with signaling phospholipase 11 Kurrasch-Orbaugh DM, Parrish JC, Watts VJ, Nichols DE. A A(2)s. Biochim Biophys Acta 2000; 1488: 159–166. complex signaling cascade links the serotonin2A receptor to 30 Li S, Okamoto T, Chun M, Sargiacomo M, Casanova JE, Hansen phospholipase A2 activation: the involvement of MAP kinases. SH et al. Evidence for a regulated interaction between J Neurochem 2003; 86: 980–991. heterotrimeric G proteins and caveolin. J Biol Chem 1995; 270: 12 Yao JK, van Kammen DP. Red blood cell membrane dynamics in 15693–15701. schizophrenia. I. Membrane fluidity. Schizophr Res 1994; 11: 31 Carman CV, Lisanti MP, Benovic JL. Regulation of G protein- 209–216. coupled receptor kinases by caveolin. J Biol Chem 1999; 274: 13 Farkas E, de Wilde MC, Kiliaan AJ, Meijer J, Keijser JN, Luiten 8858–8864. PG. Dietary long chain PUFAs differentially affect hippocampal 32 Wolf MJ, Izumi Y, Zorumski CF, Gross RW. Long-term potentia- muscarinic 1 and serotonergic 1A receptors in experimental tion requires activation of calcium-independent phospholipase cerebral hypoperfusion. Brain Res 2002; 954: 32–41. A2. FEBS Lett 1995; 377: 358–362. 14 Clark JD, Milona N, Knopf JL. Purification of a 110-kilodalton 33 Barbour B, Szatkowski M, Ingledew N, Attwell D. Arachidonic cytosolic phospholipase A2 from the human monocytic cell line acid induces a prolonged inhibition of glutamate uptake into glial U937. Proc Natl Acad Sci USA 1990; 87: 7708–7712. cells. Nature 1989; 342: 918–920. 15 Murakami M, Kambe T, Shimbara S, Yamamoto S, Kuwata H, 34 Bernard J, Chabot C, Gagne J, Baudry M, Massicotte G. Melittin Kudo I. Functional association of type IIA secretory phospholi- increases AMPA receptor affinity in rat brain synaptoneuro- pase A(2) with the glycosylphosphatidylinositol-anchored hepar- somes. Brain Res 1995; 671: 195–200. an sulfate proteoglycan in the cyclooxygenase-2-mediated 35 Gaudreault SB, Chabot C, Gratton JP, Poirier J. The caveolin delayed prostanoid-biosynthetic pathway. J Biol Chem 1999; scaffolding domain modifies 2-amino-3-hydroxy-5-methyl-4-iso- 274: 29927–29936. xazole propionate receptor binding properties by inhibiting 16 Murakami M, Shimbara S, Kambe T, Kuwata H, Winstead MV, phospholipase A2 activity. J Biol Chem 2004; 279: 356–362. Tischfield JA et al. The functions of five distinct mammalian 36 Tournois C, Mutel V, Manivet P, Launay JM, Kellermann O. Cross- phospholipase A2S in regulating arachidonic acid release. Type talk between 5-hydroxytryptamine receptors in a serotonergic IIa and type V secretory phospholipase A2S are functionally cell line. Involvement of arachidonic acid metabolism. J Biol redundant and act in concert with cytosolic phospholipase A2. Chem 1998; 273: 17498–17503. J Biol Chem 1998; 273: 14411–14423. 37 Schaeffer EL, Gattaz WF. Inhibition of calcium-independent 17 Kuwata H, Yamamoto S, Miyazaki Y, Shimbara S, Nakatani Y, phospholipase A(2) activity in rat hippocampus impairs acquisi- Suzuki H et al. Studies on a mechanism by which cytosolic tion of short- and long-term memory. Psychopharmacology phospholipase A2 regulates the expression and function of (Berlin) 2005; 181: 392–400. type IIA secretory phospholipase A2. J Immunol 2000; 165: 38 Cao Y, Pearman AT, Zimmerman GA, McIntyre TM, Prescott SM. 4024–4031. Intracellular unesterified arachidonic acid signals apoptosis. 18 Han WK, Sapirstein A, Hung CC, Alessandrini A, Bonventre JV. Proc Natl Acad Sci USA 2000; 97: 11280–11285. Cross-talk between cytosolic phospholipase A2 alpha (cPLA2 39 Shinzawa K, Tsujimoto Y. PLA2 activity is required for nuclear alpha) and secretory phospholipase A2 (sPLA2) in hydrogen shrinkage in caspase-independent cell death. J Cell Biol 2003; peroxide-induced arachidonic acid release in murine 163: 1219–1230. mesangial cells: sPLA2 regulates cPLA2 alpha activity that is 40 Kim SJ, Gershov D, Ma X, Brot N, Elkon KB. I-PLA(2) activation responsible for arachidonic acid release. J Biol Chem 2003; 278: during apoptosis promotes the exposure of membrane lysopho- 24153–24163. sphatidylcholine leading to binding by natural immunoglobulin 19 Bonventre JV, Huang Z, Taheri MR, O’Leary E, Li E, Moskowitz M antibodies and complement activation. J Exp Med 2002; 196: MA et al. Reduced fertility and postischaemic brain injury in 655–665. mice deficient in cytosolic phospholipase A2. Nature 1997; 390: 41 Lauber K, Bohn E, Krober SM, Xiao YJ, Blumenthal SG, 622–625. Lindemann RK et al. Apoptotic cells induce migration of

Molecular Psychiatry PLA2s and schizophrenia MH Law et al 555 phagocytes via caspase-3-mediated release of a lipid attraction signal transduction in schizophrenia? Biol Psychiatry 1997; 41: signal. Cell 2003; 113: 717–730. 507–513. 42 Clemens JA, Stephenson DT, Smalstig EB, Roberts EF, Johnstone 62 Tavares H, Yacubian J, Talib LL, Barbosa NR, Gattaz WF. EM, Sharp JD et al. Reactive glia express cytosolic phospholipase Increased phospholipase A2 activity in schizophrenia with A2 after transient global forebrain ischemia in the rat. Stroke absent response to niacin. Schizophr Res 2003; 61: 1–6. 1996; 27: 527–535. 63 Shah SH, Vankar GK, Peet M, Ramchand CN. Unmedicated 43 Stephenson D, Rash K, Smalstig B, Roberts E, Johnstone E, Sharp schizophrenic patients have a reduced skin flush in response to J et al. Cytosolic phospholipase A2 is induced in reactive glia topical niacin. Schizophr Res 2000; 43: 163–164. following different forms of neurodegeneration. Glia 1999; 27: 64 Hudson CJ, Gotowiec A, Seeman M, Warsh J, Ross BM. Clinical 110–128. subtyping reveals significant differences in calcium-dependent 44 Gilroy DW, Newson J, Sawmynaden P, Willoughby DA, Croxtall phospholipase A2 activity in schizophrenia. Biol Psychiatry JD. A novel role for phospholipase A2 isoforms in the checkpoint 1999; 46: 401–405. control of acute inflammation. FASEB J 2004; 18: 489–498. 65 Waldo MC. Co-distribution of sensory gating and impaired niacin 45 Berger GE, Wood SJ, Pantelis C, Velakoulis D, Wellard RM, flush response in the parents of schizophrenics. Schizophr Res McGorry PD. Implications of lipid biology for the pathogenesis of 1999; 40: 49–53. schizophrenia. Aust NZ J Psychiatry 2002; 36: 355–366. 66 Turenne SD, Seeman M, Ross BM. An animal model of nicotinic- 46 Horrobin DF. Schizophrenia as a prostaglandin deficiency acid-induced vasodilation: effect of haloperidol, caffeine and disease. Lancet 1977; 1: 936–937. nicotine upon nicotinic acid response. Schizophr Res 2001; 50: 47 Horrobin DF. The membrane phospholipid hypothesis as a 191–197. biochemical basis for the neurodevelopmental concept of 67 Ross BM, Hughes B, Turenne S, Seeman M, Warsh JJ. Reduced schizophrenia. Schizophr Res 1998; 30: 193–208. vasodilatory response to methylnicotinate in schizophrenia as 48 Yao JK, Stanley JA, Reddy RD, Keshavan MS, Pettegrew JW. assessed by laser Doppler flowmetry. Eur Neuropsychopharma- Correlations between peripheral polyunsaturated fatty acid col 2004; 14: 191–197. content and in vivo membrane phospholipid metabolites. Biol 68 Smesny S, Berger G, Rosburg T, Riemann S, Riehemann S, Psychiatry 2002; 52: 823–830. McGorry P et al. Potential use of the topical niacin skin test in 49 Yao JK, van Kammen DP, Welker JA. Red blood cell membrane early psychosis – a combined approach using optical reflection dynamics in schizophrenia. II. Fatty acid composition. Schizophr spectroscopy and a descriptive rating scale. J Psychiatr Res 2003; Res 1994; 13: 217–226. 37: 237–247. 50 Khan MM, Evans DR, Gunna V, Scheffer RE, Parikh VV, Mahadik 69 Smesny S, Rosburg T, Klemm S, Riemann S, Baur K, Rudolph N SP. Reduced erythrocyte membrane essential fatty acids and et al. The influence of age and gender on niacin skin test results – increased lipid peroxides in schizophrenia at the never-medi- implications for the use as a biochemical marker in schizo- cated first-episode of psychosis and after years of treatment with phrenia. J Psychiatr Res 2004; 38: 537–543. antipsychotics. Schizophr Res 2002; 58: 1–10. 70 Glen AI, Cooper SJ, Rybakowski J, Vaddadi K, Brayshaw N, 51 Arvindakshan M, Sitasawad S, Debsikdar V, Ghate M, Evans D, Horrobin DF. Membrane fatty acids, niacin flushing and clinical Horrobin DF et al. Essential polyunsaturated fatty acid and lipid parameters. Prostag Leukot Essent Fatty Acids 1996; 55: 9–15. peroxide levels in never-medicated and medicated schizophrenia 71 Horrobin DF. Schizophrenia: a biochemical disorder? Biomedi- patients. Biol Psychiatry 2003; 53: 56–64. cine 1980; 32: 54–55. 52 Horrobin DF, Manku MS, Hillman H, Iain A, Glen M. Fatty acid 72 Ross BM, Hudson CJ, Erlich J, Warsh JJ, Kish SJ. Increased levels in the brains of schizophrenics and normal controls. Biol phospholipid breakdown in schizophrenia. Evidence for the Psychiatry 1991; 30: 795–805. involvement of a calcium-independent phospholipase A2. Arch 53 Yao JK, van Kammen DP, Gurklis J. Red blood cell membrane Gen Psychiatry 1997; 54: 487–494. dynamics in schizophrenia. III. Correlation of fatty acid 73 Macdonald DJ, Boyle RM, Glen AC, Ross BM, Glen AI, Ward PE et abnormalities with clinical measures. Schizophr Res 1994; 13: al. The investigation of cytosolic phospholipase A2 using ELISA. 227–232. Prostag Leukot Essent Fatty Acids 2004; 70: 377–381. 54 Doris AB, Wahle K, MacDonald A, Morris S, Coffey I, Muir W et 74 Noponen M, Sanfilipo M, Samanich K, Ryer H, Ko G, Angrist B et al. Red cell membrane fatty acids, cytosolic phospholipase-A2 al. Elevated PLA2 activity in schizophrenics and other psychia- and schizophrenia. Schizophr Res 1998; 31: 185–196. tric patients. Biol Psychiatry 1993; 34: 641–649. 55 Strassnig M, Singh Brar J, Ganguli R. Dietary fatty acid and 75 Gattaz WF, Kollisch M, Thuren T, Virtanen JA, Kinnunen PK. antioxidant intake in community-dwelling patients suffering Increased plasma phospholipase-A2 activity in schizophrenic from schizophrenia. Schizophr Res 2005; 76: 343–351. patients: reduction after neuroleptic therapy. Biol Psychiatry 56 Hibbeln JR, Makino KK, Martin CE, Dickerson F, Boronow J, 1987; 22: 421–426. Fenton WS. Smoking, gender, and dietary influences on 76 Gattaz WF, Hubner CV, Nevalainen TJ, Thuren T, Kinnunen PK. erythrocyte essential fatty acid composition among patients with Increased serum phospholipase A2 activity in schizophrenia: a schizophrenia or schizoaffective disorder. Biol Psychiatry 2003; replication study. Biol Psychiatry 1990; 28: 495–501. 53: 431–441. 77 Lasch J, Willhardt I, Kinder D, Sauer H, Smesny S. Fluorometric 57 Myers CS, Contreras MA, Chang MC, Rapoport SI, Appel NM. assays of phospholipase A2 activity with three different sub- Haloperidol downregulates phospholipase A(2) signaling in rat strates in biological samples of patients with schizophrenia. Clin basal ganglia circuits. Brain Res 2001; 896: 96–101. Chem Lab Med 2003; 41: 908–914. 58 Pettegrew JW, Keshavan MS, Panchalingam K, Strychor S, Kaplan 78 Ross BM, Turenne S, Moszczynska A, Warsh JJ, Kish SJ. DB, Tretta MG et al. Alterations in brain high-energy phosphate Differential alteration of phospholipase A2 activities in brain of amd membrane phospholipid metabolism in first-episode, drug- patients with schizophrenia. Brain Res 1999; 821: 407–413. naive schizophrenics. A pilot study of the dorsal prefrontal 79 Katila H, Appelberg B, Rimon R. No differences in phospholi- cortex by in vivo phosphorous 31 nuclear magnetic resonance pase-A2 activity between acute psychiatric patients and controls. spectroscopy. Arch Gen Psychiatry 1991; 48: 563–568. Schizophr Res 1997; 26: 103–105. 59 Fukuzako H, Fukuzako T, Hashiguchi T, Kodama S, Takigawa M, 80 Gattaz WF, Schmitt A, Maras A. Increased platelet phospholipase Fujimoto T. Changes in levels of phosphorus metabolites in A2 activity in schizophrenia. Schizophr Res 1995; 16: 1–6. temporal lobes of drug-naive schizophrenic patients. Am J 81 Albers M, Meurer H, Marki F, Klotz J. Phospholipase A2 activity Psychiatry 1999; 156: 1205–1208. in serum of neuroleptic-naive psychiatric inpatients. Pharma- 60 Morrow JD, Awad JA, Oates JA, Roberts II LJ. Identification of copsychiatry 1993; 26: 94–98. skin as a major site of prostaglandin D2 release following oral 82 Brzustowicz LM, Hodgkinson KA, Chow EW, Honer WG, Bassett administration of niacin in humans. J Invest Dermatol 1992; 98: AS. Location of a major susceptibility locus for familial schizo- 812–815. phrenia on chromosome 1q21–q22. Science 2000; 288: 678–682. 61 Hudson CJ, Lin A, Cogan S, Cashman F, Warsh JJ. The niacin 83 Gurling HM, Kalsi G, Brynjolfson J, Sigmundsson T, Sherrington challenge test: clinical manifestation of altered transmembrane R, Mankoo BS et al. Genomewide genetic linkage analysis

Molecular Psychiatry PLA2s and schizophrenia MH Law et al 556 confirms the presence of susceptibility loci for schizophrenia, on polymorphisms and schizophrenia among two independent 1q32.2, 5q33.2, and 8p21–22 and provides support family-based samples. Psychiatr Genet 2001; 11: 207–212. for linkage to schizophrenia, on chromosomes 11q23.3–24 and 101 Junqueira R, Cordeiro Q, Meira-Lima I, Farid Gattaz W, Vallada H. 20q12.1–11.23. Am J Hum Genet 2001; 68: 661–673. Allelic association analysis of phospholipase A2 genes with 84 Lewis CM, Levinson DF, Wise LH, DeLisi LE, Straub RE, Hovatta I schizophrenia. Psychiatr Genet 2004; 14: 157–160. et al. Genome scan meta-analysis of schizophrenia and bipolar 102 Tao R, Yu Y, Zhang X, Guo Y, Shi J, Xie L et al. Cytosolic PLA2 disorder, part II: schizophrenia. Am J Hum Genet 2003; 73: 34–48. genes possibly contribute to the etiology of schizophrenia. Am J 85 Sherry ST, Ward M, Sirotkin K. dbSNP-database for single Med Genet B Neuropsychiatr Genet 2005; 137: 56–58. nucleotide polymorphisms and other classes of minor genetic 103 Farooqui AA, Yang HC, Rosenberger TA, Horrocks LA. Phospho- variation. Genome Res 1999; 9: 677–679. A2 and its role in brain tissue. J Neurochem 1997; 69: 86 Dolan-O’Keefe M, Chow V, Monnier J, Visner GA, Nick HS. 889–901. Transcriptional regulation and structural organization of the 104 Niculescu III AB, Segal DS, Kuczenski R, Barrett T, Hauger RL, human cytosolic phospholipase A(2) gene. Am J Physiol Lung Kelsoe JR. Identifying a series of candidate genes for mania and Cell Mol Physiol 2000; 278: L649–L657. psychosis: a convergent functional genomics approach. Physiol 87 Risch N, Merikangas K. The future of genetic studies of complex Genom 2000; 4: 83–91. human diseases. Science 1996; 273: 1516–1517. 105 Craddock N, O’Donovan MC, Owen MJ. The genetics of 88 Tay A, Maxwell P, Li Z, Goldberg H, Skorecki K. Isolation of schizophrenia and bipolar disorder: dissecting psychosis. J Med promoter for cytosolic phospholipase A2 (cPLA2). Biochim Genet 2005; 42: 193–204. Biophys Acta 1994; 1217: 345–347. 106 Stefansson H, Sarginson J, Kong A, Yates P, Steinthorsdottir V, 89 Ramchand CN, Wei J, Lee KH, Peet M. Phospholipase A2 gene Gudfinnsson E et al. Association of neuregulin 1 with schizo- polymorphism and associated biochemical alterations in schizo- phrenia confirmed in a Scottish population. Am J Hum Genet phrenia. Phospholipid Spectrum Disorder in Psychiatry. Carn- 2003; 72: 83–87. forth: Marius Press, 1999; 1999: 31–38. 107 Reich DE, Lander ES. On the allelic spectrum of human disease. 90 Hudson CJ, Kennedy JL, Gotowiec A, Lin A, King N, Gojtan K et Trends Genet 2001; 17: 502–510. al. Genetic variant near cytosolic phospholipase A2 associated 108 Fearnhead NS, Wilding JL, Winney B, Tonks S, Bartlett S, with schizophrenia. Schizophr Res 1996; 21: 111–116. Bicknell DC et al. Multiple rare variants in different genes 91 Hudson CJ, Lin A, Horrobin DF. Phospholipases: in search of a account for multifactorial inherited susceptibility to colorectal genetic base of schizophrenia. Prostag Leukot Essent Fatty Acids adenomas. Proc Natl Acad Sci USA 2004; 101: 15992–15997. 1996; 55: 119–122. 109 Cohen JC, Kiss RS, Pertsemlidis A, Marcel YL, McPherson R, 92 Frieboes RM, Moises HW, Gattaz WF, Yang L, Li T, Liu X et al. Hobbs HH. Multiple rare alleles contribute to low plasma levels Lack of association between schizophrenia and the phospholi- of HDL cholesterol. Science 2004; 305: 869–872. pase-A(2) genes cPLA2 and sPLA2. Am J Med Genet 2001; 105: 110 Pritchard JK. Are rare variants responsible for susceptibility to 246–249. complex diseases? Am J Hum Genet 2001; 69: 124–137. 93 Price SA, Fox H, St Clair D, Shaw DJ. Lack of association between 111 Hopper JL, Hayes VM, Spurdle AB, Chenevix-Trench G, Jenkins schizophrenia and a polymorphism close to the cytosolic MA, Milne RL et al. A protein-truncating mutation in CYP17A1 phospholipase A2 gene. Psychiatr Genet 1997; 7: 111–114. in three sisters with early-onset breast cancer. Hum Mutat 2005; 94 Peet M, Ramchand CN, Lee J, Telang SD, Vankar GK, Shah S et al. 26: 298–302. Association of the Ban I dimorphic site at the human cytosolic 112 Wen G, Mahata SK, Cadman P, Mahata M, Ghosh S, Mahapatra phospholipase A2 gene with schizophrenia. Psychiatr Genet NR et al. Both rare and common polymorphisms contribute 1998; 8: 191–192. functional variation at CHGA, a regulator of catecholamine 95 Wei J, Lee KH, Hemmings GP. Is the cPLA2 gene associated with physiology. Am J Hum Genet 2004; 74: 197–207. schizophrenia? Mol Psychiatry 1998; 3: 480–481. 113 Wright AF, Hastie ND. Complex genetic diseases: controversy 96 Wei J, Hemmings GP. A study of a genetic association between the over the Croesus code. Genome Biol 2001; 2: COMMENT2007.1– PTGS2/PLA2G4A locus and schizophrenia. Prostag Leukot COMMENT 2007.8. Essent Fatty Acids 2004; 70: 413–415. 114 Knoblauch H, Bauerfeind A, Toliat MR, Becker C, Luganskaja T, 97 Pae CU, Yu HS, Lee KU, Kim JJ, Lee CU, Lee SJ et al. BanI Gunther UP et al. Haplotypes and SNPs in 13 lipid-relevant genes polymorphism of the cytosolic phospholipase A2 gene may explain most of the genetic variance in high-density lipoprotein confer susceptibility to the development of schizophrenia. Prog and low-density lipoprotein cholesterol. Hum Mol Genet 2004; Neuropsychopharmacol Biol Psychiatry 2004; 28: 739–741. 13: 993–1004. 98 Wei J, Hemmings GP. A study of the combined effect of the 115 Nakatani N, Uozumi N, Kume K, Murakami M, Kudo I, CLDN5 locus and the genes for the phospholipid metabolism Shimizu T. Role of cytosolic phospholipase A2 in the production pathway in schizophrenia. Prostag Leukot Essent Fatty Acids of lipid mediators and histamine release in mouse bone-marrow- 2005; 73: 441–445. derived mast cells. Biochemical Journal. 2000; 352(Part 2): 99 Yu YQ, Tao R, Wei J, Xu Q, Liu SZ, Ju GZ et al. No association 311–317. between the PTGS2/PLA2G4A locus and schizophrenia in a 116 Fujishima H, Sanchez Mejia RO, Bingham III CO, Lam BK, Chinese population. Prostag Leukot Essent Fatty Acids 2004; 71: Sapirstein A, Bonventre JV et al. Cytosolic phospholipase A2 is 405–408. essential for both the immediate and the delayed phases of 100 Chowdari KV, Brandstaetter B, Semwal P, Bhatia T, Deshpande S, eicosanoid generation in mouse bone marrow-derived mast cells. Reddy R et al. Association studies of cytosolic phospholipase A2 Proc Natl Acad Sci USA 1999; 96: 4803–4807.

Molecular Psychiatry