The authors have no conflict of interest to declare.
Article first published online: 14 MAR 2012
Copyright © 2012 Wiley Periodicals, Inc.
American Journal of Medical Genetics Part B: Neuropsychiatric Genetics
Volume 159B, Issue 4, pages 392–404, June 2012
How to Cite
Verbrugghe, P., Bouwer, S., Wiltshire, S., Carter, K., Chandler, D., Cooper, M., Morar, B., Razif, M. F.M., Henders, A., Badcock, J. C., Dragovic, M., Carr, V., Almeida, O. P., Flicker, L., Montgomery, G., Jablensky, A. and Kalaydjieva, L. (2012), Impact of the Reelin signaling cascade (Ligands–Receptors–Adaptor Complex) on cognition in schizophrenia. Am. J. Med. Genet., 159B: 392–404. doi: 10.1002/ajmg.b.32042
Phebe Verbrugghe and Sonja Bouwer contributed equally to this work.
How to Cite this Article: Verbrugghe P, Bouwer S, Wiltshire S, Carter K, Chandler D, Cooper M, Morar B, Razif MFM, Henders A, Badcock JC, Dragovic M, Carr V, Almeida OP, Flicker L, Montgomery G, Jablensky A, Kalaydjieva L. 2012. Impact of the Reelin Signaling Cascade (Ligands–Receptors–Adaptor Complex) on Cognition in Schizophrenia. Am J Med Genet Part B 159B:392–404.
- Issue published online: 7 MAY 2012
- Article first published online: 14 MAR 2012
- Manuscript Accepted: 17 FEB 2012
- Manuscript Received: 26 OCT 2011
- National Health and Medical Research Council of Australia. Grant Numbers: 37580400, 37580900, 964145, 139093, 403963, 455811
- cognitive deficit;
- Top of page
- SUBJECTS AND METHODS
- Supporting Information
Our previous neurocognitive studies of schizophrenia outlined two clusters of affected subjects—cognitively spared (CS) and cognitive deficit (CD), the latter's characteristics pointing to developmental origins and impaired synaptic plasticity. Here we investigate the contribution of polymorphisms in major regulators of these processes to susceptibility to schizophrenia and to CD in patients. We examine variation in genes encoding proteins at the gateway of Reelin signaling: ligands RELN and APOE, their common receptors APOER2 and VLDLR, and adaptor DAB1. Association analysis with disease outcome and cognitive performance in the Western Australian Family Study of Schizophrenia (WAFSS) was followed by replication analysis in the Australian Schizophrenia Research Bank (ASRB) and in the Health in Men Study (HIMS) of normal aging males. In the WAFSS sample, we observed significant association of APOE, APOER2, VLDLR, and DAB1 SNPs with disease outcome in the case–control and CD–control datasets, and with pre-morbid intelligence and verbal memory in cases. HIMS replication analysis supported rs439401 (APOE regulatory region), and rs2297660 and rs3737983 (APOER2), with an effect on memory performance in normal aging subjects consistent with the findings in schizophrenia cases. APOER2 gene expression analysis revealed lower transcript levels in lymphoblastoid cells from cognitively impaired schizophrenia patients of the alternatively spliced exon 19, mediating Reelin signaling and synaptic plasticity in the adult brain. ASRB replication analysis produced marginally significant results, possibly reflecting a recruitment strategy biased toward CS patients. The data suggest a contribution of neurodevelopmental/synaptic plasticity genes to cognitive impairment in schizophrenia. © 2012 Wiley Periodicals, Inc.
- Top of page
- SUBJECTS AND METHODS
- Supporting Information
Schizophrenia is a complex disorder characterized by significant phenotypic and genetic heterogeneity [Pulver et al., 2000; Jablensky, 2006; Allen et al., 2008]. Cognitive dysfunction, often referred to as a “core” feature of schizophrenia, involves deficits in declarative and working memory, executive function, sustained attention and learning, visuospatial ability, oculomotor control, and speed of information processing [Heinrichs and Zakzanis, 1998; Goldberg et al., 2003]. These deficits contribute to compromise the ability of the individual to cope with the demands of daily living independent of the contribution of psychotic symptoms [Mohamed et al., 2008]. However, their scope and severity show marked variation among affected individuals, ranging from pervasive deficits across multiple cognitive domains (estimated at 61–78% of cases) to mild or patchy dysfunction limited to few cognitive skills (estimated at 27–55%) [Heinrichs and Zakzanis, 1998; Reichenberg and Harvey, 2007]. The pathogenesis of cognitive impairment in schizophrenia is almost certainly polyfactorial, likely to involve multiple gene networks and interactions with the environment.
The Reelin signaling cascade is involved in brain development as well as in synaptic plasticity, long-term potentiation (LTP), learning and memory in the adult brain [D'Arcangelo et al., 1995; Trommsdorff et al., 1999; Weeber et al., 2002; Beffert et al., 2005, 2006a; Soriano and Del Rio, 2005; Herz and Chen, 2006]. In the present study, we set out to examine if genes involved in this cascade contribute to the cognitive deficits of schizophrenia patients. We focused on the genes encoding the interacting proteins at the gate of the cascade: the ligands Reelin (RELN) and Apolipoprotein E (APOE), their common receptors Very Low Density Lipoprotein Receptor (VLDLR) and APOE Receptor 2 (APOER2), and the adapter Disabled 1 (DAB1). For brevity, we refer to these as LRAC—ligand–receptor–adaptor complex. LRAC protein dysfunction is known, or suspected, to play a role in various neurological, neurodegenerative, and neuropsychiatric disorders, including brain malformations, temporal lobe epilepsy, Alzheimer's disease (AD), schizophrenia, bipolar disorder, and autism [Corder et al., 1993; Impagnatiello et al., 1998; D'Arcangelo et al., 1999; Guidotti et al., 2000; Hong et al., 2000; Costa et al., 2001; Haas et al., 2002; Eastwood and Harrison, 2003; Knable et al., 2004; Boycott et al., 2005; Fatemi et al., 2005]. The search for genetic associations in schizophrenia, targeting exclusively the RELN gene has produced conflicting or inconclusive results [Goldberger et al., 2005; Shifman et al., 2008; Wedenoja et al., 2008]. Our choice of focus on the entire complex was grounded on the physical and functional unity of its individual components, and the a priori high plausibility of its involvement in the molecular machinery underlying cognition in health and disease, with added impetus from recent research suggesting shared genetic contribution to schizophrenia and some of the conditions mentioned above [Bassett et al., 2010; Gejman et al., 2011; Gershon et al., 2011; Reichelt et al., 2012].
Our genetic association analyses are based on a discovery case–control sample drawn from the Western Australian Family Study of Schizophrenia (WAFSS), in which all participants underwent, in addition to standardized diagnostic interviews, a detailed neurocognitive assessment. The multiple cognitive tests data were statistically integrated into two composite continuous traits, one characterizing pervasive cognitive deficit across the majority of cognitive tasks, and another depicting only mild or focal deficits. Based on their psychometric affinity to one or the other of the composite trait profiles, patients were assigned to a “cognitive deficit” (CD) or a “cognitively spared” (CS) cluster [Hallmayer et al., 2005; Jablensky, 2006]. We set out to examine the differential contribution of polymorphisms in the selected genes to the risk of clinically defined schizophrenia, its cognitively defined case clusters, and to individual cognitive traits in affected subjects.
Replication analyses were performed in two independent samples: (i) the Australian Schizophrenia Research Bank (ASRB), a national sample of cognitively characterized schizophrenia cases and controls [Loughland et al., 2010]; and (ii) a sample of normal aging males, drawn from the general population - the Western Australian Health in Men Study (HIMS) [Norman et al., 2009].
SUBJECTS AND METHODS
- Top of page
- SUBJECTS AND METHODS
- Supporting Information
The WAFSS sample comprised 508 individuals (336 schizophrenia patients and 172 normal controls) of European descent (>75% Anglo-Irish ancestry) (Supplementary Table I). Patients were recruited from consecutive admissions to a psychiatric hospital or from community mental health centers and are a fairly representative cross-section of service users with severe mental disorder. The control subjects were recruited from the same area by random sampling from local telephone directories, or among Red Cross blood donors, and screened for psychopathology to exclude those with a personal or family history of psychotic illness. Written informed consent was obtained from all participants. The study was approved by the Human Research Ethics Committees of the University of Western Australia and of the North Metropolitan Health Area Service in Perth, Australia.
Diagnostic assessment was based on standardized interviews employing the Schedules for Clinical Assessment in Neuropsychiatry (SCAN [Wing et al., 1990]) and scored using the OPCRIT algorithm [McGuffin et al., 1991]. Videorecorded interviews and clinical charts were independently reviewed by two senior clinicians who assigned consensus ICD-10 and DSM-IV research diagnoses. Patients and controls were administered a battery of neurocognitive tests targeting key domains of cognition: (i) general cognitive ability (the National Adult Reading Test, NART, estimating prior or premorbid IQ [Nelson and Willison, 1991]; and the Shipley Institute of Living Scale, SILS [Zachary, 1986], assessing current IQ); (ii) episodic verbal memory (the Rey Auditory Verbal Learning test, RAVLT [Schmidt, 1996], measuring immediate recall of a 15-word list and delayed free recall after distraction); (iii) sustained attention (the visual Continuous Performance Task [Beck et al., 1956]—identical pairs version, CPT-IP, and degraded stimulus version, CPT-DS); (iv) executive function (the FAS version of the Controlled Oral Word Association Test [Benton et al., 1994]). Patients were examined during periods of relative clinical stabilization. Performance data from the multiple cognitive domains were integrated into composite continuous traits, using a latent structure method (grade of membership, GoM [Manton et al., 1994]). GoM computes simultaneously a number of multivariate response patterns, or “pure types” (PT), and quantifies every individual's degrees of similarity to each one of the PTs. Two PTs, displaying contrasting test profiles, accounted for 82.1% of the patient sample: one characterized by pervasive CD affecting verbal memory, executive function, attention and general ability, and another, CS, with only slower speed of information processing distinguishing it significantly from controls. Based on a statistically defined cut-off in the PT trait scores, patients were assigned to a CD (N = 155) or a CS (N = 121) cluster. Sixty patients, who could not be unequivocally assigned to either cluster, remained in the clinical schizophrenia sample.
The ASRB case–control sample
The ASRB is a national research facility involving a multi-site consortium of investigators [Loughland et al., 2010]. Recruitment of cases and controls was through a large-scale, multi-media advertising campaign. The replication sample (Supplementary Table II) included 308 schizophrenia cases and 129 controls. Both patients and controls were English speaking, but their ethnic background was more varied than in WAFSS. Clinical assessment was based on the Diagnostic Interview for Psychosis (DIP [Castle et al., 2006]), scored with the OPCRIT algorithm, and yielding diagnoses according to ICD-10 and DSM-IV. All participants completed neurocognitive assessment comprising prior and current IQ (Wechsler Test of Adult Reading, WTAR [Wechsler, 2001]; Wechsler Abbreviated Scale of Intelligence [Wechsler, 1999]); immediate and delayed memory, attention and language (Repeatable Battery for Assessment of Neuropsychological Status, RBANS [Randolph, 1998]); and executive function (Controlled Oral Word Association Test [Benton et al., 1994]). Compared with the discovery sample, ASRB patients had higher educational achievement and level of functioning. ASRB controls were educated significantly beyond the high school level and had IQ scores up to one SD above the population norm.
The HIMS study
HIMS is a longitudinal study of community-dwelling men aged 65 years and over, who originally participated in a screening trial for abdominal aortic aneurysm. Subsequently, the coverage of the study was extended to include common health conditions, cognitive functioning, and psychosocial elements of health and wellbeing [Norman et al., 2009; Almeida et al., 2010]. Recruitment was randomized, based on the Western Australian electoral roll. The 683 subjects participating in the present study (mean age 76.6 years; 52.9% completed secondary school or higher) had a Mini-Mental State Examination [Folstein et al., 1975] mean score of 27.6 (out of 30). Memory performance was assessed with the California Verbal Learning Test (CVLT-II) [Delis et al., 2000].
Choice of Genes
The selected genes encode the five closely interacting proteins that drive the downstream pathways of Reelin signaling (Fig. 1). The ligands Reelin and ApoE exert their effects on brain development and adult function through competitive binding to the receptors VLDLR and APOER2, which activates downstream signaling via phosphorylation of the DAB1 adapter. During development, LRAC proteins regulate neuronal migration and synaptogenesis [D'Arcangelo et al., 1995; D'Arcangelo, 2005; Trommsdorff et al., 1999; Tissir and Goffinet, 2003; Soriano and Del Rio, 2005; Herz and Chen, 2006]. In the mature brain they are involved in the synapse/spine formation, synaptic plasticity and LTP, learning and memory through regulation of a specialized post-synaptic pathway which modulates NMDA receptor function and Ca2+ flux [Weeber et al., 2002; Beffert et al., 2005, 2006a; Qiu et al., 2006]. Aging across species, including humans, is associated with a reduction in RELN expression and the number of Reelin-positive interneurons [Knuesel et al., 2009]. In age-related neuropsychiatric disorders, the ApoE ε4 isoform is the best-established genetic contributor to late-onset AD [Corder et al., 1993]. Reelin and VLDLR deficiencies cause severe neurodevelopmental disorders, such as Norman–Roberts type lissencephaly [Hong et al., 2000)] and cerebellar hypoplasia with cerebral gyral simplification [Boycott et al., 2005]. RELN and VLDLR expression is reduced in schizophrenia brains [Impagnatiello et al., 1998; Guidotti et al., 2000; Costa et al., 2001; Eastwood and Harrison, 2003; Knable et al., 2004], and haploinsufficient reeler mice show behavioral abnormalities resembling schizophrenia [Qiu et al., 2006]. Cross-talk between ApoE and Reelin signaling is increasingly recognized as a factor in the etiology of dementias [Herz and Chen, 2006; Rebeck et al., 2006], and dysregulated tau phosphorylation is a proposed convergence point of schizophrenia and AD pathogenesis [Deutsch et al., 2006]. The role of the entire complex in schizophrenia has not been systematically investigated.
Comprehensive coverage of the genomic sequence spanning the longest transcripts and ∼10 kb upstream and downstream of each gene was ensured by a panel of 526 single nucleotide polymorphisms (SNPs). These were selected from HapMap Rel21/phase II-listed variants (http://www.hapmap.org/publications.html.en) with a minor allele frequency >5%, augmented with coding SNPs from dbSNP (http://www.ncbi.nlm.nih.gov/SNP/). TagSNPs were identified using the pair-wise option of Tagger [de Bakker et al., 2005] implemented in Haploview [Barrett et al., 2005], with a threshold of r2 > 0.8.
Genotyping and Data Cleaning
Genotyping was performed using the Illumina GoldenGate technology. DNA samples from CEPH trio 1,334 (Coriell Cell Repository) served as internal controls. Primary analysis of the results with the Illumina BeadStudio Genotyping Module was followed by visual inspection and assessment of data quality and clustering. Consistency (at P > 0.001) with Hardy–Weinberg equilibrium was assessed in the control samples using PLINK [Purcell et al., 2007]. The average genotyping call rate was 99.9%; eight samples with a call rate <98% were removed. Statistical analysis was thus performed on 500 individuals (329 cases and 171 controls) for 503 markers, capturing 2,074 variants (Supplementary Table III). Genotyping in the ASRB and HIMS replication samples was performed using the Applied Biosystems (Melbourne, Australia) Taqman® and SNPlex® assays.
Association of each polymorphism with disease outcome was analyzed in the schizophrenia–controls, CD–controls, and CS–controls datasets of the discovery sample using logistic regression with 2 degrees of freedom tests of genotypic association. Significant results (at P < 0.01) were explored further to determine the best fitting genetic models as appropriate following inspection of the odds ratios (ORs). To generate the number of false positives expected under the null hypothesis of no association, we permuted (1,000 replicates) the genotype/phenotype correspondence in the data. Statistical “noteworthiness” was assessed using the false positive reporting probability (FPRP) [Wacholder et al., 2004] modeled over a range of prior probabilities of association (0.25, 0.1, 0.01, 0.001) and statistical power estimates at OR 1.5.
Analysis of association of each polymorphism with the raw test scores for general cognitive ability (NART and SILS), verbal memory (RAVLTi and RAVLTd), and attention (CPT-IP and CPT-DS) was done in the schizophrenia discovery sample using linear regression. Correction for multiple testing was done on a gene-by gene basis, for each trait and across all traits, by permuting the genotype/phenotype data (1,000 replicates).
SNPs associated with disease risk (at P < 0.01) in the schizophrenia–controls and/or CD–controls datasets of the discovery sample were analyzed in the ASRB schizophrenia/controls sample. To examine whether such variants also contribute to normal variation in cognitive function in aging individuals, they were tested for association with verbal memory in the HIMS sample.
Replication in both the ASRB and HIMS samples was also sought for SNPs significantly associated (after correction for multiple testing) with individual cognitive traits in the discovery sample.
Gene Expression Analyses
We examined mRNA expression in lymphoblastoid cell lines (LCLs) from 65 patients (21 CD, 32 CS, and 12 non-CD–non-CS) and 34 control subjects. LCL suitability as a model for neuronal expression of the genes investigated in this study was first tested by checking for presence/abundance of the transcripts, and by comparing splicing patterns to those in total adult brain mRNA (Clontech, Australia). Quantitative polymerase chain reactions (qPCR) were designed depending on the location of associated polymorphisms, bioinformatics analysis (Transcription Element Search System at www.cbil.upenn.edu/cgi-bin/tess) and predicted functional effects. Experimental details are described in the Supplementary Methods.
- Top of page
- SUBJECTS AND METHODS
- Supporting Information
Association with clinical schizophrenia and its cognitively characterized case clusters
The full set of results is shown in Supplementary Table IV as ORs, 95% confidence intervals and unadjusted P values.
The number of observed positive findings (at P < 0.01) versus those expected by chance under the null hypothesis differed between the three datasets, with the CD cluster producing the highest ratio of observed versus randomly expected significant results: schizophrenia/controls 4 observed versus 3 expected; CD/controls 5 observed versus 2.9 expected, while no significant contribution of LRAC gene polymorphisms was found in the CS cluster (2 observed/2.7 expected). In the schizophrenia/controls and/or CD/controls datasets (Table I), the associated SNPs were located in the genes encoding APOE, the two receptors, and the adaptor DAB1, with no positive results for Reelin. The ORs and P values in the total schizophrenia/control and the CD/controls datasets were similar (Table I), suggesting that the association with schizophrenia was driven by the CD cluster despite its substantially smaller sample size. These variants also showed FPRP ≤0.5 over a range of prior probabilities (Supplementary Table V), suggesting “noteworthiness.” Since current knowledge of gene function and of schizophrenia pathogenesis does not permit further judgment, all were considered equally plausible and taken to the replication stage (except rs3737983 in APOER2, which is highly correlated with rs2297660: D′ = 1.0; r2 > 0.9).
|Gene||SNP ID||WAFSS SCZ/controls||WAFSS CD-controls||ASRB SCZ/controls||HIMS episodic memory (CVLT scores)|
|OR (95% CI)||P-Value||OR (95% CI)||P-Value||OR (95% CI)||P-Value||OR (95% CI)||P-Value|
|APOE||rs439401||1.78 (1.22; 2.60) dom||0.003||2.02 (1.28; 3.2) dom||0.003||1.04 (0.81; 1.34)||0.38||−0.47 (−0.89; 0.00) delayed||0.03|
|APOER2||rs2297660a||0.49 (0.3; 0.78) rec||0.002||0.42 (0.24; 0.76) rec||0.004||0.83 (0.65; 1.06)||0.06||0.87 (0.12; 1.60) delayed||0.02|
|rs3737983a||0.53 (0.34; 0.83) rec||0.005||0.39 (0.22; 0.71) rec||0.002||NA||0.83 (0.14; 1.56) delayed||0.02|
|VLDLR||rs1454626||1.59 (1.17; 2.15) add||0.003||1.73 (1.22; 2.46) add||0.002||0.93 (0.65; 1.33)||0.22||−0.66 (−1.45; 0.13) immediate||0.10|
|DAB1||rs694060||1.71 (1.18; 2.49) dom||0.005||2.10 (1.34; 3.29) dom||0.001||0.99 (0.54; 1.81)||0.26||NA|
|rs534455||1.56 (1.18; 2.06) add||0.002||1.76 (1.27; 2.44) add||0.0006||1.52 (0.91; 2.57)||0.055||0.66 (0.00; 1.33) immediate||0.05|
Replication analyses showed a trend for rs2297660 in APOER2 and rs534455 in DAB1 in the ASRB sample (Table I). The direction of effect was identical to that in the discovery sample, and the OR (95%CI) of the DAB1 polymorphism was very similar in the two samples. The results for the remaining three SNPs were non-significant.
In the HIMS sample of normal aging men, the analysis of association with memory performance (CVLT scores) produced significant results for the APOE and APOER2 variants associated with risk of schizophrenia and CD in the discovery sample (Table I). The direction of the effects corroborated our original findings: the risk allele of APOE rs439401 was associated with poorer memory performance in HIMS participants; similarly, APOER2 rs2297660, which had a protective effect in the discovery and ASRB samples, was associated with better memory in HIMS. DAB1 rs534455 produced a marginally significant result with an opposite direction of the effect—it was associated with higher disease risk in WAFSS and ASRB but with higher CVLT scores in the HIMS replication sample.
Association with individual cognitive traits
The effect of LRAC polymorphisms on cognitive performance was examined in the WAFSS case sample, using the raw test scores for general cognitive ability, verbal memory and sustained attention as continuous traits. The full set of results (Supplementary Table VI) is shown as regression (β) coefficients representing the change in phenotype measure per copy of the minor allele, 95% confidence limits and unadjusted P values. The significant results are summarized in Table II.
|Gene||SNP ID||Cognitive trait||WAFSS discovery sample||Replication ASRB||Replication HIMSc|
|β-coefficient (95% CI)||P-Value||β-coefficient (95% CI)||P-Value||β-coefficient (95% CI)||P-Value|
|Unadjusted||Corrected per trait||Corrected across traits|
|ApoER2||rs1288480||Pre-morbid IQa||−3.24 (−4.88; −1.59)||0.0001||0.003||0.02||−1.88 (−4.23; 0.467)||0.06||0.72 (−007; 1.54)||0.08|
|ApoER2||rs1288487||−2.98 (−4.53; −1.42)||0.0002||0.005||0.03||NA||NA|
|ApoER2||rs1288493||3.01 (1.52; 4.49)||0.00009||0.003||0.02||−0.1236 (−2.427; 2.18)||0.45||NA|
|DAB1||rs12133813||Episodic memoryb||1.87 (0.97; 2.76)||0.00005||0.01||0.06||0.30 (−0.63; 1.23)||0.26||−0.60 (−1.23; 0.04)||0.03|
Three SNPs, rs1288480, rs1288487, and rs1288493, located within 13 kb in intron 2 of the APOER2 gene (D′ > 0.7 and r2 < 0.5) were associated with estimated premorbid intelligence (NART scores). They also showed association, albeit at a lower significance level, with correlated traits—current IQ, verbal memory, and sustained attention. Haplotype testing showed that the combination of the minor alleles of rs1288480 and rs1288487 and the major allele of rs1288493 is strongly associated with low premorbid IQ (β = −3.86, 95% CI −5.55 to −2.17, P = 0.00001).
A significant effect on verbal memory performance was observed for DAB1 SNP rs12133813, associated with higher RAVLTi scores in schizophrenia patients.
Replication in the ASRB sample (Table II) showed a trend of association of APOER2 rs1288480 with lower pre-morbid intelligence (WTAR scores), but came short of reaching nominal significance. The results for APOER2 rs1288493 and DAB1 rs12133813 were not significant.
In the HIMS sample, we obtained significant results for DAB1 rs12133813, again with an effect opposite to the discovery sample—higher RAVLT scores in WAFSS and lower CVLT in unaffected aging individuals (Table II). The APOER2 variants, associated with premorbid intelligence in schizophrenia, had no significant effect on memory in the HIMS replication.
The Associated Polymorphisms
SNP rs439401 is located downstream of the APOE gene, close to a peroxisome-proliferator-activated-receptor-γ (PPAR-γ) response element (PPRE) (Fig. 2) [Galetto et al., 2001]. It has been suggested to modulate the response of AD patients to treatment with rosiglitazone maleate [Li et al., 2008], a PPAR-γ agonist, which improves cognition in AD patients and in mouse AD models [Pedersen et al., 2006; Risner et al., 2006; Brodbeck et al., 2008].
To check whether the observed association was a reflection of the effects of APOE isoforms, we performed haplotype analysis of schizophrenia with the coding SNPs, rs429358 and rs7412, defining the ε alleles. We found no association of ε4, or any other isoform, with schizophrenia. Moreover, the risk-conferring minor allele of the associated rs439401 occurred mostly on an ε3 background, and combined ε allele-rs439401 haplotype analysis failed to produce significant results (Supplementary Table VII), supporting an independent rs439401 effect.
Whether rs439401 itself is a functional variant or flags another regulatory polymorphisms within the region, and whether this variation affects the expression of APOE, or plays a role in PPAR-γ-mediated neuroprotection, is currently unclear.
APOE mRNA baseline levels in lymphocyte cell lines (LCL) were significantly lower than those in brain and than required for reliable qPCR (not shown), making LCLs an inappropriate model for APOE expression studies.
APOER2 rs2297660 and rs3737983
These two synonymous coding SNPs, G419G in exon 9 and D874D in exon 17, fall within an LD block (D′ ∼ 1.0, r2 ≥ 0.9) that contains many additional polymorphisms and spans multiple exons and functionally important protein domains (Fig. 3a) [D'Arcangelo et al., 1999; Hiesberger et al., 1999; Trommsdorff et al., 1999]. In view of the focus of this study on cognitive performance, the most interesting of these is exon 19, which undergoes activity-dependent alternative splicing, and mediates Reelin signaling in the mature synapse [Beffert et al., 2005; D'Arcangelo, 2005]. Its encoded 59-amino acid sequence is responsible for the APOER2 interactions with NMDA receptors and scaffolding protein PSD-95 at the post-synaptic density. LTP, memory and learning are selectively impaired in animal models lacking exon 19 [Beffert et al., 2005]. Exon 19-containing isoforms also play a role in the protection against age-related neuronal loss [Beffert et al., 2006b].
Based on the specialized functions of exon 19 and its location in the LD block with the associated SNPs, we hypothesized that sequence variation in the region could affect splicing efficiency. Our comparative examination of alternative transcripts showed that exon 19-containing isoforms are expressed in LCLs, which can thus serve as a model for brain expression (Fig. 3b). Quantitative analysis of mRNA levels (Fig. 3c) showed down-regulation of exon 19-containing transcripts in LCLs derived from schizophrenia (P = 0.023) and CD (P = 0.016), but not CS patients. There was a trend toward higher expression level in cells homozygous for the protective SNP alleles, although that sample was small and the difference did not reach statistical significance.
DAB1 rs534455 and rs12133813
These two SNPs, only 369 basepairs apart, are located in DAB1 intron 11 (Fig. 4) with LD in the control sample of D′ = 1 and r2 = 0.42. DAB1 is a gene of unusual size and complexity, so far never implicated in human disorders but central to Reelin signaling in the developing and mature brain [Howell et al., 1997; Bar et al., 2003]. Speculative links to impaired neurodevelopment and adult brain function are suggested by the location of the SNPs close to an alternatively spliced isoform (referred to as 555*), which is highly expressed in the ventricular zone at the early developmental stages and excluded as neural differentiation progresses [Howell et al., 1997; Bar et al., 2003]. While it is possible that rs534455 and rs12133813 or a functional polymorphism in LD may affect splicing efficiency of the 555* isoform, the cell- and developmental stage-specific nature of this potential effect precluded its analysis in LCLs, highlighting their limitations as a model for testing disease-associated neuronal changes.
- Top of page
- SUBJECTS AND METHODS
- Supporting Information
This study was designed to address the likely pathogenetic heterogeneity of schizophrenia and the distinctive relevance of genes involved in CNS development and synaptic function to different subsets of cases identifiable by neurocognitive subtyping. Our choice of candidate genes was based on abundant a priori evidence documenting the major role of LRAC genes in neurodevelopment and adult synaptic function [D'Arcangelo et al., 1995; Trommsdorff et al., 1999; Weeber et al., 2002; Beffert et al., 2005, 2006a; Soriano and Del Rio, 2005; Herz and Chen, 2006], and on the salient features of the CD case cluster pointing to impairments in these processes [Hallmayer et al., 2005]. The physical and functional unity of the selected candidate genes demanded their analysis as a complex rather than in isolation, and the present study is the first examination of the entire LRAC in schizophrenia.
In agreement with the characteristics of the CD cluster and with our choice of genes, the discovery sample produced evidence of association for sequence polymorphism in 4 out of the 5 LRAC genes (APOE, APOER2, VLDLR, and DAB1). We obtained no positive results for any SNPs in the RELN gene, in agreement with previous negative findings in Finnish schizophrenia families [Wedenoja et al., 2008]. SNP rs7341475 in RELN intron 4 was found in a genome-wide association scan of an Ashkenazi Jewish sample to increase risk of schizophrenia in women [Shifman et al., 2008], with a recent replication of the findings reported by Liu et al. . In our study, this variant was represented by tag SNP rs6954479, which showed no positive results in either the association (Supplementary Table IV) or the gene–sex interaction analyses specifically targeting this SNP (P > 0.59) in any case–control sample. At the level of individual cognitive traits in WAFSS schizophrenia cases, we observed a strong effect of a set of APOER2 polymorphisms on estimated premorbid IQ, and of a DAB1 SNP on verbal memory.
The exploratory nature of our analyses, dealing with a discovery sample that is very well characterized yet modest in size, imposed the need for replication studies of the “noteworthy” findings. Replication in the ASRB case–control sample [Loughland et al., 2010] whose neurocognitive characterization (albeit relying on a different set of tests) allows a comparison to the WAFSS findings, showed a trend for association with schizophrenia and pre-morbid IQ of SNPs in APOER2 and DAB1. While the direction of effects and the OR for the DAB1 polymorphism were similar to the discovery sample, the marginally significant results and the lack of replication of the remaining SNPs might have been due to dissimilarities in sample composition and recruitment strategies. WAFSS case finding, based on consecutive hospital or clinic admissions, resulted in greater variation in the cognitive profiles of the recruited cases and in approximately equal representation of CD and CS affected subjects [Hallmayer et al., 2005]. Since ASRB recruitment relied primarily on a media advertising campaign, the sample was enriched with high-functioning schizophrenia patients corresponding to the CS cluster, which showed no evidence of association with LRAC genes.
The HIMS collection [Norman et al., 2009; Almeida et al., 2010] was included to explore parallels between the molecular mechanisms of cognitive decline in schizophrenia to those in normal aging. The replication results in this sample are of particular interest. SNP rs439401, located in close proximity to a PPAR-γ response element downstream of APOE, was associated with CD in schizophrenia patients and with poor memory performance in HIMS, both independent of ApoE4 status. The effects could be hypothetically related to dysregulation of ApoE expression and generally of the PPAR-γ pathway, which plays a major role in energy metabolism and mitochondrial biology [Galetto et al., 2001]. Similarly, correspondent results between WAFSS and HIMS were obtained for APOER2 rs2297660, which was associated with reduced risk of schizophrenia and CD in the WAFSS sample and with better memory performance in HIMS. The observed associations could be related to the production of APOER2 exon 19-containing isoforms of the receptor, which have a documented role in memory, synaptic plasticity, and protection against age-related neuronal loss [Beffert et al., 2005; Beffert et al., 2006b]. Our hypothesis was supported by the in vitro expression studies, showing a modest but significant reduction in exon 19-containing transcripts in patients' LCLs, especially in those derived from affected subjects in the CD cluster (Fig. 3c). Chronically impaired exon 19 production (even if modest, as shown in this study) could be one of the multiple factors contributing to CD in schizophrenia and age-related decline in the general population. The adaptor protein Disabled 1 is central to Reelin signaling and the integration of pathways [Bar et al., 2003]. The evidence of association with risk of schizophrenia and CD (rs534455), and with memory performance in affected subjects (rs12133813) is enhanced by the location of the two polymorphisms in the same small region of DAB1. Both SNPs were also associated with verbal memory in the HIMS sample, where, interestingly, the direction of effect was opposite to that in affected subjects. Such effect reversal is not unusual, and is compatible with the “inverted U-curve” dose–response pattern, described in studies of cortical excitability in response to varying concentrations of neurotransmitter molecules [Vijayraghavan et al., 2007; Christie-Fougere et al., 2009]. We have previously reported such differences in the direction of effect between cases and controls for NRG3, GRM3, and PRKCA [Jablensky et al., 2011; Morar et al., 2011], pointing to a likely role of an abnormal brain environment in schizophrenia and/or effects of environmental stressors.
While LRAC functioning depends on the integrity of the network, its individual components have distinct roles that allow independent contributions to synaptic dysfunction in schizophrenia through diverse mechanisms. ApoE involvement in schizophrenia could be related to its competitive binding to the APOER2 and VLDLR receptors, interfering with Reelin signaling, a mechanism also suspected to operate in AD [Herz and Chen, 2006; Frotscher, 2010; Knuesel, 2010]. This mechanism could be independent of ε isoforms and result from dysregulation of APOE expression, a scenario supported by the findings of increased APOE levels in schizophrenia brains [Dean et al., 2003; Gibbons et al., 2010]. The latter study found no changes in the expression of the APOER2 receptor in schizophrenia, however our results raise the need for investigations specifically targeting the protein domain encoded by the alternatively spliced exon 19. While, to our knowledge, no association studies of VLDLR and APOER2 in schizophrenia have been published, animal KO models reveal interesting phenotypic characteristics depending on the affected receptor molecule that appear relevant to the findings of the present study. Developmental defects in the Apoer2 KO affect the hippocampus and neocortex, while the cerebellum is mostly targeted in the Vldlr KO, and late-phase LTP is significantly affected in Apoer2- but not in Vldlr-deficient animals [Trommsdorff et al., 1999; Weeber et al., 2002]. Reelin signaling could be impaired further by inefficient splicing of the APOER2 receptor, selectively affecting memory [Beffert et al., 2005] and, hypothetically, by the presence of the “risk” alleles in DAB1 that, together with multiple other genetic and environmental effects, could contribute to impaired neurodevelopment, and/or affect phosphorylation and activation of the adaptor molecule. Importantly, all these effects would be superimposed on a background of suboptimal Reelin signaling, which arises from promoter hypermethylation and down-regulation of RELN expression in schizophrenia brains [Guidotti et al., 2000; Costa et al., 2001; Eastwood and Harrison, 2003; Knable et al., 2004], and from lower RELN expression and loss of Reelin-positive interneurons as part of normal aging [Tamura et al., 2007; Knuesel et al., 2009]. A parallelism between declining LRAC function in normal brain aging and a similar extent of LRAC dysfunction in the much younger CD cluster of schizophrenia patients is consonant with the hypothesis of “accelerated aging” in schizophrenia [Kirkpatrick et al., 2008].
The results of the present study, placed into the context of recent advances in molecular neurobiology and cognitive neuroscience, suggest that LRAC is a polyfunctional system, a veritable molecular “hub” with multiple ramifications extending all the way from neuropsychiatric disorders, such as schizophrenia, to neurodegenerative disease and normal brain aging. Finally, our results support the utility of refined cognitive phenotyping in schizophrenia, which enabled us to demonstrate LRAC impact on a particular subgroup of schizophrenia patients showing features similar to Kraepelin's dementia praecox [Kraepelin, 1971] and raising the possibility of accelerated brain aging in this subset, a prospect which merits further attention.
- Top of page
- SUBJECTS AND METHODS
- Supporting Information
We thank patients, family members, and volunteer controls for their participation in WAFFS. The study was supported by National Health and Medical Research Council of Australia grant nos. 37580400 and 37580900 to AJ and LK, and 964145, 139093, 403963, 455811 to OPA and LF. Samples and data were also received from the Australian Schizophrenia Research Bank (ASRB), which is supported by enabling grant no. 386500 of the National Health and Medical Research Council of Australia (Chief Investigators: V. Carr, U. Schall, R. Scott, A. Jablensky, B. Mowry, P. Michie, S. Catts, F. Henskens, and C. Pantelis), the Pratt Foundation, Ramsay Healthcare, the Viertel Charitable Foundation, and the Schizophrenia Research Institute.
- Top of page
- SUBJECTS AND METHODS
- Supporting Information
- 2008. Systematic meta-analyses and field synopsis of genetic association studies in schizophrenia: The SzGene database. Nat Genet 40(7): 827– 834. , , , , , , , .
- 2010. Depression, antidepressant use and mortality in later life: The Health in Men Study. PLoS ONE 5(6): e11266. , , , .
- 2003. The gene encoding disabled-1 (DAB1), the intracellular adaptor of the Reelin pathway, reveals unusual complexity in human and mouse. J Biol Chem 278(8): 5802– 5812. , , , , .
- 2005. Haploview: Analysis and visualization of LD and haplotype maps. Bioinformatics 21(2): 263– 265. , , , .
- 2010. Copy number variations in schizophrenia: Critical review and new perspectives on concepts of genetics and disease. Am J Psychiatry 167(8): 899– 914. , , .
- 1956. A continuous performance test of brain damage. J Consult Psychol 20(5): 343– 350. , , , , .
- 2005. Modulation of synaptic plasticity and memory by Reelin involves differential splicing of the lipoprotein receptor Apoer2. Neuron 47(4): 567– 579. , , , , , , , , , , .
- 2006a. Functional dissection of Reelin signaling by site-directed disruption of Disabled-1 adaptor binding to apolipoprotein E receptor 2: Distinct roles in development and synaptic plasticity. J Neurosci 26(7): 2041– 2052. , , , , , , , .
- 2006b. ApoE receptor 2 controls neuronal survival in the adult brain. Curr Biol 16(24): 2446– 2452. , , , , , , .
- 1994. Multilingual aphasia examination, 3rd edition. Iowa City: AJA Associates. , , .
- 2005. Homozygous deletion of the very low density lipoprotein receptor gene causes autosomal recessive cerebellar hypoplasia with cerebral gyral simplification. Am J Hum Genet 77(3): 477– 483. , , , , , , , , , , .
- 2008. Rosiglitazone increases dendritic spine density and rescues spine loss caused by apolipoprotein E4 in primary cortical neurons. Proc Natl Acad Sci USA 105(4): 1343– 1346. , , , , , .
- 2008. SimHap GUI: An intuitive graphical user interface for genetic association analysis. BMC Bioinformatics 9: 557. , , .
- 2006. The diagnostic interview for psychoses (DIP): Development, reliability and applications. Psychol Med 36(1): 69– 680. , , , , , , , , , .
- 2009. Calcineurin inhibition eliminates the normal inverted U curve, enhances acquisition and prolongs memory in a mammalian 3′–5′-cyclic AMP-dependent learning paradigm. Neuroscience 158(4): 1277– 1283. , , , .
- 1993. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families. Science 261(5123): 921– 923. , , , , , , , , .
- 2001. Dendritic spine hypoplasticity and downregulation of Reelin and GABAergic tone in schizophrenia vulnerability. Neurobiol Dis 8(5): 723– 742. , , , , , .
- 2005. Apoer2: A Reelin receptor to remember. Neuron 47(4): 471– 473. .
- 1995. A protein related to extracellular matrix proteins deleted in the mouse mutant reeler. Nature 374(6524): 719– 723. , , , , , .
- 1999. Reelin is a ligand for lipoprotein receptors. Neuron 24(2): 471– 479. , , , , , .
- 2005. Efficiency and power in genetic association studies. Nat Genet 37(11): 1217– 1223. , , , , , .
- 2003. Increased levels of apolipoprotein E in the frontal cortex of subjects with schizophrenia. Biol Psychiatry 54(6): 616– 622. , , , , , , , , , , , .
- 2000. California Verbal Learning test-second edition (CVLT-II) manual. San Antonio, TX: PsychCorp (Harcourt Assessment). , , , .
- 2006. Dysregulation of tau phosphorylation is a hypothesized point of convergence in the pathogenesis of Alzheimer's disease, frontotemporal dementia and schizophrenia with therapeutic implications. Prog Neuropsychopharmacol Biol Psychiatry 30(8): 1369– 1380. , , .
- 2003. Interstitial white matter neurons express less Reelin and are abnormally distributed in schizophrenia: Towards an integration of molecular and morphologic aspects of the neurodevelopmental hypothesis. Mol Psychiatry 8(9): 821– 831. , .
- 2005. Reelin signaling is impaired in autism. Biol Psychiatry 57(7): 777– 787. , , , , , , , .
- 1975. Mini-mental state. J Psychiatr Res 12(3): 189– 198. , , .
- 2010. Role for Reelin in stabilizing cortical architecture. Trends Neurosci 33(9): 407– 414. .
- 2001. Identification of a peroxisome-proliferator-activated-receptor response element in the apolipoprotein E gene control region. Biochem J 357(Pt2): 521– 527. , , , , .
- 2011. Genetics of schizophrenia: New findings and challenges. Annu Rev Genomics Hum Genet 12: 121– 144. , , .
- 2011. After GWAS: Searching for genetic risk for schizophrenia and bipolar disorder. Am J Psychiatry 168(3): 253– 256. , , .
- 2010. Low density lipoprotein receptor-related protein and apolipoprotein E expression is altered in schizophrenia. Front Psychiatry 1: 19. , , , .
- 2003. Neurocognitive deficits in schizophrenia. In: Weinberger SHD, editor. Schizophrenia, 2nd edition. Oxford: Blackwell. pp 168– 184. , , .
- 2005. Population-based and family based association study of 5′UTR polymorphism of the Reelin gene and schizophrenia. Am J Med Genet B Neuropsychiatr Genet 137B(1): 51– 55. , , , , , , .
- 2000. Decrease in Reelin and glutamic acid decarboxylase67 (GAD67) expression in schizophrenia and bipolar disorder: A postmortem brain study. Arch Gen Psychiatry 57(11): 1061– 1069. , , , , , , , , , , , .
- 2002. Role for Reelin in the development of granule cell dispersion in temporal lobe epilepsy. J Neurosci 22(14): 5797– 5802. , , , , , , , .
- 2005. Genetic evidence for a distinct subtype of schizophrenia characterized by pervasive cognitive deficit. Am J Hum Genet 77(3): 468– 476. , , , , , , , , , , , .
- 1998. Neurocognitive deficit in schizophrenia: A quantitative review of the evidence. Neuropsychology 12(3): 426– 445. , .
- 2006. Reelin, lipoprotein receptors and synaptic plasticity. Nat Rev Neurosci 7(11): 850– 859. , .
- 1999. Direct binding of Reelin to VLDL receptor and ApoE receptor 2 induces tyrosine phosphorylation of disabled-1 and modulates tau phosphorylation. Neuron 24(2): 481– 489. , , , , , , .
- 2000. Autosomal recessive lissencephaly with cerebellar hypoplasia is associated with human RELN mutations. Nat Genet 26(1): 93– 96. , , , , , , , .
- 1997. Scrambler and yotari disrupt the disabled gene and produce a reeler-like phenotype in mice. Nature 389(6652): 730– 733. , , , .
- 1998. A decrease of Reelin expression as a putative vulnerability factor in schizophrenia. Proc Natl Acad Sci USA 95(26): 15718– 15723. , , , , , , , , , , , , , .
- 2006. Subtyping schizophrenia: Implications for genetic research. Mol Psychiatry 11(9): 815– 836. .
- 2011. Polymorphisms associated with normal memory variation also affect memory impairment in schizophrenia. Genes Brain Behav 10(4): 410– 417. , , , , , , , , .
- 2008. Is schizophrenia a syndrome of accelerated aging? Schizophr Bull 34(6): 1024– 1032. , , , , .
- 2004. Molecular abnormalities of the hippocampus in severe psychiatric illness: Postmortem findings from the Stanley Neuropathology Consortium. Mol Psychiatry 9(6): 609– 620. , , , , .
- 2010. Reelin-mediated signaling in neuropsychiatric and neurodegenerative diseases. Prog Neurobiol 91(4): 257– 274. .
- 2009. Age-related accumulation of Reelin in amyloid-like deposits. Neurobiol Aging 30(5): 697– 716. , , , , , , , , , , .
- 1971. Psychiatrie. 8 Auflage. Barth Leipzig. English translation: 1971. Dementia praecox and paraphrenia. Huntington, New York: Krieger. .
- 2008. Candidate single-nucleotide polymorphisms from a genome-wide association study of Alzheimer disease. Arch Neurol 65(1): 45– 53. , , , , , , et al.
- 2010. Replication of an association of a common variant in the Reelin gene (RELN) with schizophrenia in Ashkenazi Jewish women. Psychiatr Genet 20(4): 184– 186. , , , , , , et al.
- 2010. Australian Schizophrenia Research Bank: A database of comprehensive clinical, endophenotypic and genetic data for aetiological studies of schizophrenia. Aust N Z J Psychiatry 44(11): 1029– 1035. , , , , , , , , , , , , , , , .
- 1994. Statistical applications using fuzzy sets. New York: John Wiley. , , .
- 1991. A polydiagnostic application of operational criteria in studies of psychotic illness. Development and reliability of the OPCRIT system. Arch Gen Psychiatry 48(8): 764– 770. , , .
- 2008. Relationship of cognition and psychopathology to functional impairment in schizophrenia. Am J Psychiatry 165(8): 978– 987. , , , , , .
- 2011. Neuregulin 3 (NRG3) as a susceptibility gene in a schizophrenia subtype with florid delusions and relatively spared cognition. Mol Psychiatry 16(8): 860– 866. , , , , , .
- 1991. The National Adult Reading Test (NART): Test manual, 2nd edition. Windsor, UK: NFER-Nelson. , .
- 2009. Cohort profile: The health in men study (HIMS). Int J Epidemiol 38(1): 48– 52. , , , , , .
- 2006. Rosiglitazone attenuates learning and memory deficits in Tg2576 Alzheimer mice. Exp Neurol 199(2): 265– 273. , , , , , .
- 2000. Genetic heterogeneity in schizophrenia: Stratification of genome scan data using co-segregating related phenotypes. Mol Psychiatry 5(6): 650– 653. , , , , , , , , , , , , .
- 2007. PLINK: A tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 81(3): 559– 575. , , , , , , , , , , .
- 2006. Cognitive disruption and altered hippocampus synaptic function in Reelin haploinsufficient mice. Neurobiol Learn Mem 85(3): 228– 242. , , , , , .
- 1998. RBANS manual—Repeatable battery for the assessment of neuropsychological status. San Antonio, TX: PsychCorp (Harcourt Assessment). .
- 2006. The generation and function of soluble apoE receptors in the CNS. Mol Neurodegener 1: 15. , , , , .
- 2012. The role of neurexins in schizophrenia and autistic spectrum disorder. Neuropharmacology 62(3): 1519– 1526. , , .
- 2007. Neuropsychological impairments in schizophrenia: Integration of performance-based and brain imaging findings. Psychol Bull 133(5): 833– 858. , .
- Roziglitazone in Alzheimer's Disease Study Group. 2006. Efficacy of rosiglitazone in a genetically defined population with mild-to-moderate Alzheimer's disease. Pharmacogenomics J 6(4): 246– 254. , , , , , , , , ,
- 1996. Rey Auditory and Verbal Learning test. A handbook. Los Angeles: Western Psychological Services. .
- 2008. Genome-wide association identifies a common variant in the Reelin gene that increases the risk of schizophrenia only in women. PLoS Genet 4(2): e28. , , , , , , , , , , , , , , , .
- 2005. The cells of Cajal-Retzius: Still a mystery one century after. Neuron 46(3): 389– 394. , .
- 2007. Epigenetic aberration of the human REELIN gene in psychiatric disorders. Mol Psychiatry 12(6): 593– 600. , , , .
- 2003. Reelin and brain development. Nat Rev Neurosci 4(6): 496– 505. , .
- 1999. Reeler/disabled-like disruption of neuronal migration in knockout mice lacking the VLDL receptor and ApoE receptor 2. Cell 97(6): 689– 701. , , , , , , , , .
- 2007. Inverted-U dopamine D1 receptor actions on prefrontal neurons engaged in working memory. Nat Neurosci 10(3): 376– 384. , , , , .
- 2004. Assessing the probability that a positive report is false: An approach for molecular epidemiology studies. J Natl Cancer Inst 96(6): 434– 442. , , , , .
- 1999. Wechsler Abbreviated Scale of Intelligence (WASI). San Antonio, TX: PsychCorp (Harcourt Assessment). .
- 2001. Wechsler Test of Adult Reading (WTAR). San Antonio, TX: PsychCorp (Harcourt Assessment). .
- 2008. Replication of linkage on chromosome 7q22 and association of the regional Reelin gene with working memory in schizophrenia families. Mol Psychiatry 13(7): 673– 684. , , , , , , , , , , , .
- 2002. Reelin and ApoE receptors cooperate to enhance hippocampal synaptic plasticity and learning. J Biol Chem 277(42): 39944– 39952. , , , , , , .
- 1990. SCAN: Schedules for clinical assessment in neuropsychiatry. Arch Gen Psychiatry 47(6): 589– 593. , , , , , , , , .
- 1986. Shipley Institute of Living Scale. Revised manual. Los Angeles: Western Psychological Services. .
- Top of page
- SUBJECTS AND METHODS
- Supporting Information
Additional supporting information may be found in the online version of this article.
|ajmg_32042_sm_SupplTab1.doc||38K||Supplementary Table 1|
|ajmg_32042_sm_SupplTab2.doc||26K||Supplementary Table 2|
|ajmg_32042_sm_SupplTab3.doc||588K||Supplementary Table 3|
|ajmg_32042_sm_SupplTab4.doc||624K||Supplementary Table 4|
|ajmg_32042_sm_SupplTab5.doc||24K||Supplementary Table 5|
|ajmg_32042_sm_SupplTab6.doc||771K||Supplementary Table 6|
|ajmg_32042_sm_SupplTab7.doc||28K||Supplementary Table 7|
Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.