Original Article

Molecular Psychiatry (2010) 15, 463–472; doi:10.1038/mp.2008.110; published online 21 October 2008

Polymorphisms in SREBF1 and SREBF2, two antipsychotic-activated transcription factors controlling cellular lipogenesis, are associated with schizophrenia in German and Scandinavian samples

S Le Hellard1,2,11, T W Mühleisen3,11, S Djurovic4,5, J Fernø1,2, Z Ouriaghi3, M Mattheisen3, C Vasilescu3, M B Raeder1,2, T Hansen6, J Strohmaier7, A Georgi7, F F Brockschmidt3, I Melle5,8, I Nenadic9, H Sauer9, M Rietschel7, M M Nöthen3,10, T Werge6, O A Andreassen5,8, S Cichon3,10,11 and V M Steen1,2,11

  1. 1Department of Clinical Medicine, Bergen Mental Health Research Center, University of Bergen, Bergen, Norway
  2. 2Dr Einar Martens’ Research Group for Biological Psychiatry, Center for Medical Genetics and Molecular Medicine, Haukeland University Hospital, Helse Bergen HF, Bergen, Norway
  3. 3Department of Genomics, Life and Brain Center, University of Bonn, Bonn, Germany
  4. 4Department of Medical Genetics, Ullevål University Hospital, Kirkeveien 166, Oslo, Norway
  5. 5The TOP Study Group, Institute of Psychiatry, University of Oslo, Oslo, Norway
  6. 6Research Institute of Biological Psychiatry, Sct. Hans Hospital, Boserupvej 2, Roskilde, Denmark
  7. 7Division of Genetic Epidemiology in Psychiatry, Central Institute of Mental Health, University of Heidelberg, Mannheim, Germany
  8. 8Department of Psychiatry, Ullevål University Hospital, Kirkeveien 166, Oslo, Norway
  9. 9Department of Psychiatry and Psychotherapy, Friedrich-Schiller-University of Jena, Philosophenweg 3, Jena, Germany
  10. 10Insitute of Human Genetics, University of Bonn, Wilhelmstr 31, Bonn, Germany

Correspondence: Dr S Le Hellard, Bergen Mental Health Research Center, Department of Clinical Medicine, University of Bergen, Bergen N-5021, Norway. E-mail: stephanie.le.hellard@helse-bergen.no

11These authors contributed equally to this work.

Received 9 July 2008; Revised 2 September 2008; Accepted 15 September 2008; Published online 21 October 2008.



Several studies have reported structural brain abnormalities, decreased myelination and oligodendrocyte dysfunction in schizophrenia. In the central nervous system, glia-derived de novo synthesized cholesterol is essential for both myelination and synaptogenesis. Previously, we demonstrated in glial cell lines that antipsychotic drugs induce the expression of genes involved in cholesterol and fatty acids biosynthesis through activation of the sterol regulatory element binding protein (SREBP) transcription factors, encoded by the sterol regulatory element binding transcription factor 1 (SREBF1) and sterol regulatory element binding transcription factor 2 (SREBF2) genes. Considering the importance of these factors in the lipid biosynthesis and their possible involvement in antipsychotic drug effects, we hypothesized that genetic variants of SREBF1 and/or SREBF2 could affect schizophrenia susceptibility. We therefore conducted a HapMap-based association study in a large German sample, and identified association between schizophrenia and five markers in SREBF1 and five markers in SREBF2. Follow-up studies in two independent samples of Danish and Norwegian origin (part of the Scandinavian collaboration of psychiatric etiology study, SCOPE) replicated the association for the five SREBF1 markers and for two markers in SREBF2. A combined analysis of all samples resulted in highly significant genotypic P-values of 9 × 10−4 for SREBF1 (rs11868035, odd ration (OR)=1.26, 95% confidence interval (CI) (1.09–1.45)) and 4 × 10−5 for SREBF2 (rs1057217, OR=1.39, 95% CI (1.19–1.63)). This finding strengthens the hypothesis that SREBP-controlled cholesterol biosynthesis is involved in the etiology of schizophrenia.


functional convergent genomics; myelin; 22q; 17p; lipids



Despite the estimated high heritability of schizophrenia, the identification of underlying genetic factors has been challenging. So far, more than 130 susceptibility genes have been proposed in schizophrenia by linkage or association studies, most of them awaiting robust replication and identification of functionally relevant genetic variants.1 Recent reports looking at the overall picture propose that such candidate genes could be grouped in interlinked pathways related to glutamate transmission, synaptic plasticity, oxidative stress, myelination and oligodendrocyte viability.1, 2, 3

In search of promising candidate genes, studies of gene expression in brain samples from patients with schizophrenia or various cellular and animal models have also been carried out.4 Using such a functional convergent genomics approach, we investigated transcriptional responses in cultured human glial cells exposed to antipsychotic drugs. We found that some typical and atypical antipsychotics markedly upregulated several genes involved in the biosynthesis of cholesterol and fatty acids.5, 6 These genes are all controlled by the sterol regulatory element binding protein (SREBP) transcription factors,7, 8 and we demonstrated that the psychotrophic drugs stimulated cellular lipogenesis through a direct activation of the SREBP proteins.5, 6

Interestingly, there is now substantial support for abnormalities in lipid biosynthesis and oligodendrocyte function in the brain of patients with schizophrenia, with focus on indicators of disturbed myelination of fiber tracts (for review, Davis et al., 20032 and Garver et al., 20089). Cholesterol is a major component of myelin, and in the brain, cholesterol is synthesized de novo.10 On this background, in combination with our demonstration of antipsychotic-induced transcriptional activation of the SREBP system, we considered the SREBPs as promising new candidate genes in schizophrenia. There are two different SREBP isoforms, SREBP1 (with two splice variants, SREBP1a and -1c, both encoded by the sterol regulatory element binding transcription factor 1 (SREBF1) gene on chr17p11.2) and sterol regulatory element binding transcription factor 2 (SREBP2) (encoded by the SREBF2 gene on chr22q13.2). Although the different SREBP isoforms overlap in function, SREBP1c controls the expression of fatty acid, phospholipid and triglyceride biosynthetic genes,11 SREBP2 mainly regulates cholesterol biosynthesis,8 whereas SREBP1a stimulates the expression of both cholesterol and fatty acid biosynthesis genes, with the later more efficiently activated.11

The aim of this study was to examine haplotype-tagging SNPs (htSNPs) in SREBF1 and SREBF2 for association to schizophrenia in several large case–control samples of German and Scandinavian origin.


Materials and methods

Phase I: Association study of SREBF1 and SREBF2 in a German sample of schizophrenia patients and population-based controls

Clinical samples

Psychiatric examinations were approved by the ethics committees of the Faculties of Medicine at the Universities of Bonn, Jena, Heidelberg, and Mannheim. Patients were recruited from consecutive admissions to the inpatient units of the Central Institute of Mental Health in Mannheim, and the Departments of Psychiatry at the Universities of Bonn and Jena. All patient and control subjects gave written informed consent before study participation and are of German descent. A total of 782 patients were interviewed by an experienced psychiatrist or psychologist. Diagnoses were made according to DSM-IV12 criteria based on an SCID-I13 interview, medical records and the family history method. We also used the OPCRIT14 system. Final diagnoses were reached by consensus of at least two investigators using the lifetime best estimate procedure,15 9.6% of our cases suffered from schizoaffective disorder (n=77) and were subgrouped according to the criteria used for schizophrenia. The patients’ average age was 40.6±13.6 (mean±s.d.), and their average age-of-onset was 22.9±10.1. The control sample consisted of 839 population-based individuals (481 males, 358 females) with an average age of 44.0±14.02. Genomic DNA was extracted from blood samples using a conventional salting-out protocol.16

Marker selection and genotyping

SNP genotype data covering the genomic region of each gene plus 5′ and 3′ flanking sequences were downloaded from HapMap data release 20 (phase II, NCBI build 35, dbSNP build 125) for the CEPH (Utah residents with ancestry from northern and western Europe, Central European University (CEU)) trios: SREBF1 (chr17p11.2: 17612211–17936632) and SREBF2 (chr22q13.2: 40539032–40668797). The most centromeric and telomeric HapMap markers downloaded were: rs2975005–rs11658846 (SREBF1) and rs139572–rs8140869 (SREBF2). Genotype data were subsequently uploaded into Haploview 3.32 (http://www.broad.mit.edu/mpg/haploview/index.php)17 to construct region-specific linkage disequilibrium (LD) and haplotype block maps. Pair-wise comparisons of markers >500kb apart were ignored, only markers with a minor allele frequency >0.1%, a Hardy–Weinberg P-value >0.001, and a marker call rate of >75% (Haploview defaults) were included in the analysis. Haplotype blocks were defined by the Gabriel et al.18 algorithm which creates blocks only if 95% of informative comparisons are ‘strong LD’. On block analysis, we extended the selection of markers to the complete haplotype blocks they were localized in, that is markers were selected for the whole blocks not just within the candidate gene. Haplotype tagging (ht) SNPs were selected on a block-by-block basis to represent haplotypes of frequencies >1% in the European population, resulting in 10 markers for SREBF1 and 30 for SREBF2.

Genotyping was performed with the high-throughput BeadArray platform technology from Illumina Inc. (San Diego, CA, USA). Microtitre plates contained mixtures of patients and controls; two individuals were genotyped as technical controls on each plate and genotype replicate consistency was 100%.

Copy number variants in SREBF1 and SREBF2

The Toronto copy number variant (CNV) database (http://projects.tcag.ca/variation/)19 was searched for putative CNVs localized in SREBF1 or SREBF2. According to this database, there is no CNV spanning the SREBF1 region. Iarfate et al.20 reported the identification of a deletion on the clone RP5-821D11, which contains the SREBF2 5-UTR, exon 1 and part of intron 1. Since this original report, the Toronto variations located on this clone (12263, 12264 and 27059) have been further defined to an insertion/deletion of 322–327bp in the MEI1 gene, apparently more centromeric and not overlapping with or affecting the SREBF2 gene.

In addition, we screened genome-wide genotype data interrogated by HumanHap550 BeadArrays Illumina, San Diego, USA) available from another study (genome-wide association scan of schizophrenia using about 1400 individuals of German origin, unpublished data) for the existence of putative CNVs using the QuantiSNP program.21 We did not find any evidence of CNVs in the SREBF1 and SREBF2 genes.

Phase 2: Replication study in two Scandinavian case control samples

Clinical sample

For phase 2, we used two case control samples of the Scandinavian collaboration of psychiatric etiology (SCOPE), which includes patients with schizophrenia spectrum disorders and controls originated from Norway (the TOP study, 258 cases and 221 controls) or Denmark (the Danish Psychiatric Biobank in Denmark, 473 cases and 905 controls). A total of 731 patients, and 1126 healthy control subjects were successfully genotyped. A detailed description of the SCOPE samples has recently been reported elsewhere.22

The Norwegian Scientific-Ethical Committees, the Norwegian Data Protection Agency, the Danish Scientific Committees, and the Danish Data Protection Agency all approved the study. All patients have given written informed consent before inclusion into the project.

Marker selection and genotyping

The five markers in SREBF1 (rs11654081, rs11868035, rs9907246, rs7503334 and rs7222480) and five markers in SREBF2 (rs1569451, rs133280, rs13055841, rs1052717 and rs2267443) that showed association to schizophrenia (P<0.05) in the German sample were selected for genotyping in the SCOPE samples. The markers were genotyped using the MassArray system on a Sequenom Compact MALDI-TOF device (Sequenom Inc., San Diego, CA, USA). PCR assay conditions can be obtained from the authors on request. Two percent of individuals were genotyped in duplicate, and there were no replication errors.

Phases 1 and 2: data analysis

Genotypes were quality controlled with the following criteria: individual samples with genotype call rate <90% and markers with call rate <96% were excluded from analysis.

Genotyping data were analyzed using the Helix Tree software (HelixTree Genetics Analysis Software, Golden Helix Inc., Bozeman, MT, USA, http://www.goldenhelix.com/SNP_Variation/HelixTree/index.html). Standard χ2-test of independence was used to test markers for Hardy–Weinberg equilibrium (HWE); markers with a P-value <0.001 in controls were excluded. The SNPs were then analyzed as single markers for a logistic regression where the case versus control status was the outcome predicted by the genotypes. The genotypes were also compared with an F-test where bins of one genotype vs the two others were compared with a so-called split analysis (as implemented in Helix Tree, http://www.goldenhelix.com/SN
) to test for dominant, codominant or recessive mode of transmission.

Haplotype trend regression on haplotype block basis and for 2- and 3-markers sliding window were estimated using the expectation–maximization algorithm implemented in Helix Tree (1000 iterations). In order to avoid misleading results caused by multiple rare haplotypes, haplotypes with a frequency less than or equal to 0.05 were clumped together into one group.

A 10000 shuffles permutation P-value was calculated using the regression module implemented in the Helix Tree software when a P-value <0.05 was observed.

Combined analysis of the phases 1 and 2 genotypes and samples

To assess the genetic homogeneity of the three populations, the fixation index (FST) was calculated by Arlequin 3.1 (cmpg.unibe.ch/software/arlequin3), using the control samples from Norway, Denmark and Germany. The combined analysis was performed with the same allelic and genotypic tests as for phases 1 and 2.



Phase 1: Association analysis in the German case control sample

We systematically selected 40 htSNPs from HapMap that entirely cover the SREBF1 (n=10) and SREBF2 (n=30) genomic regions, capturing all locus-specific haplotypes at a frequency >1% in the CEU population (Utah residents with ancestry from northern and western Europe). Both SNP sets were genotyped in a sample of 782 schizophrenia patients and 839 control individuals, all of German origin. For SREBF1 and SREBF2, marker call rates were 99.94 and 99.96%, respectively. Of the SREBF2 SNPs, we excluded rs5758487 because controls were out of HWE (P<0.005) and rs5758511 because of unreliable genotype clusters.

We analyzed the markers for allelic association by logistic regression and for genotypic association by a split analysis as implemented in Helix Tree (testing for dominant, codominant and recessive models). The results of the single marker analysis in the German sample along with SNP information are displayed in Table 1. We observed significant association (P-value=0.046–0.02, allelic test) between schizophrenia and five SREBF1 markers (rs11654081, rs11868035, rs9907246, rs7503334 and rs7222480). The haplotype analysis using two or three markers sliding windows over the 10 SREBF1 markers or over the two HapMap haplotype blocks did not provide additional information (data not shown). The marker rs11654081 had an allelic odd ratio (OR) of 1.18 for T allele (95% confidence interval (CI) (1.03–1.36), data not shown).

We also observed significant association (P=0.045–0.015, allelic test, see Table 1) between schizophrenia and four markers in SREBF2 (rs1569451, rs13055841, rs1052717 and rs2267443) and stronger genotypic association (P=0.032 to 0.00022, see Table 1) with three of these markers (rs13055841, rs1052717 and rs2267443) plus one additional marker (rs133280). These five markers did not show strong pair-wise LD (r2 <0.8) except for rs133280–rs1052717 (D=0.99 and r2=0.96), although these two markers are located in separate blocks (see Supplementary Figure 2_SREBF2). For the marker with the strongest allelic association (rs1569451), the allelic OR was 1.20 for the G allele (95% CI (1.03–1.39), data not shown), and for the marker with the strongest genotypic association (rs2267443; A/G), the OR for GG vs AG-AA was 1.41 (95% CI (1.15–1.72), data not shown). Considering the method of selection of the markers (htSNPs), we also performed a haplotype trend regression analysis on a per block basis based on the HapMap version used for the marker selection, with the following blocks: block 1: markers rs7364180–rs1009544, block 2: markers rs133280–rs11702960, block 3: markers rs4822063–rs13055841, block 4: markers rs5751174–rs2267443 and block 5: markers rs6002524–rs10483213. At the global haplotype analysis level, none of the blocks showed significant association (data not shown), but in block 1 the haplotype rs7364180_A, rs1569451_A, rs20066451_G and rs 1009544_C was significantly associated (P=0.005) to schizophrenia.

Phase 2: Replication in Scandinavian case control samples; and combined analysis

On the basis of the association data obtained in the German sample, with a relaxed threshold of uncorrected P-values, all markers that showed significant association (P-value <0.05) with schizophrenia were selected for replication: rs11654081, rs11868035, rs9907246, rs7503334, rs7222480 for SREBF1; and rs1569451, rs133280, rs13055841, rs1052717, rs2267443 for SREBF2.

These ten markers were genotyped in a Danish and in a Norwegian sample (469 and 198 cases, and 874 and 216 controls, respectively, after removing samples that genotyped poorly, i.e. genotype success below 0.9). The rs2267443 marker in SREBF2 had unfortunately a genotyping success rate below 85% and was therefore considered as failed in the replication sample. The other markers were analyzed for allelic and genotypic association as in phase 1.

On the basis of the calculation of a gene-based FST, we did not find evidence of significant population stratification between the Danish, Norwegian and German control populations (SREBF1 and SREBF2 FST, P-value=0.97 and 0.63, respectively). The three study samples were therefore be regarded as genetically homogenous, allowing for analysis of merged samples, and analyzed as: (i) Danish and Norwegian separately, (ii) Danish and Norwegian together and (iii) in total (Danish, Norwegian and German samples merged together).

Association between SREBF1 polymorphisms and schizophrenia in the Scandinavian and merged samples

Four of the five SREBF1 markers associated in the German sample were also significantly associated with schizophrenia in the Norwegian sample (P-values=0.042–0.00084, allelic test) but not in the Danish sample alone, nor in the Danish and Norwegian samples together (see Table 2). The diagnostic reliability of the two Scandinavian samples has been ascertained thoroughly and is unlikely to be the cause of this difference. In contrast, the samples are different in other clinical aspects that may confound our findings. Most importantly, the Norwegian patients are generally younger with a shorter duration of illness than the considerably more chronic and poorer outcome of the Danish patients.22

In the total sample (German, Danish and Norwegian), the association between SREBF1 and schizophrenia was stronger than in the German sample alone (allelic P-value=0.017–0.0015; Table 2). The markers rs11654081 and rs11868035 are in strong LD (D=1 and r2=0.75, see Supplementary Figure 1_SREBF1) and so are the markers rs9907246, rs7503334 and rs7222480 (D=0.99–1 and r2=0.97–0.99). Actually, the whole SREBF1 gene is located in a region of strong LD, within a Gabriel LD block of 247kb,17, 18 which encompasses SREBF1 as well as TOMIL2 (target of myb1 like 2), LRRC48 (leucine rich repeat containing 48), ATPAF2 (ATP synthase mitochondrial F1 complex assembly factor 2) and partially RAI1 (retinoic acid induced 1 gene) and C17orf39 (see Figure 1). For further studies we chose to focus on the markers rs11868035 and rs7222480, which had the least LD (D=0.88 and r2=0.72).

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Sterol regulatory element binding transcription factor 1 (SREBF1) genomic structure and environment with localization of polymorphisms typed. SREBF1 is located on chromosome 17p11.2 in a region of strong linkage disequilibrium (LD). In the present study, haplotype-tagging SNPs (represented by vertical bars) were selected based on the solid spine of LD structure around SREBF1, covering from marker rs3818717 (located in exon 4 of the RAI1 gene) to the marker rs9303144 (5.6kb from the 3′ end of C17orf39). The marker G09710 (represented with an arrow) has been associated to schizophrenia and neuroleptic response.23 The markers associated to schizophrenia in both phases I and II (represented by an *) were mainly located around SREBF1 and in TOM1L2, and are in strong LD with each others.

Full figure and legend (64K)

The haplotype trend regression for rs11868035–rs7222480 showed a global P-value of 0.012 and a stronger association for the rs11868035_A–rs7222480_T haplotype (P=0.0032) that remained significant after 10000 permutations (corrected P-value=0.034, data not shown).

The odds ratio for rs11868035 and rs7222480 (see Table 3), are OR=1.26 (95% CI (1.09–1.45)) for GG vs AG-AA for the marker and OR=1.25 (95% CI (1.08–1.44)) for GG vs GT-TT, respectively.

Association between SREBF2 polymorphisms and schizophrenia in the Scandinavian and merged samples

Two of the five markers associated with schizophrenia in the German sample were also strongly associated in both the Danish and the Norwegian samples (rs133280 and rs1052717, Table 2). However, these two markers are in strong LD (D=0.98, r2=0.97, see Supplementary Table 2 and Supplementary Figure 2) and are thus likely to pick up the same association signal. For further analysis, we therefore concentrated on the marker rs1052717. The association in the total merged sample (German, Danish and Norwegian) was stronger in the genotypic analysis than the regression analysis (P=0.000042 vs 0.0090), thus indicating an apparent dominant effect of the rs1052717_A allele rather than a dose-dependent effect (i.e. not linear effect). Interestingly, both the genotypic and allelic associations were stronger with increasing number of samples, that is the association in the total sample was stronger than in any of the national samples on their own, which reflects that the distribution of the associated alleles or genotypes are shifted in the same direction in all populations. On the basis of the genotypic OR for rs1052717 (Table 3), the risk of having schizophrenia was 1.39 in individuals with genotype AA or AG versus individuals with GG genotype in the total merged sample, with similarly increased ORs (ranging from 1.32 to 1.63) in each of the national samples. Haplotype Trend Regression analysis with sliding windows of two or three markers on the five markers tested in SREBF2 did not provide any additional evidence for a risk haplotype than the data obtained from the association to the single markers (data not shown).

In HapMap on the CEU trios, the associated markers rs133280 and rs1052717 are in LD (D′>0.9 and r2>0.8) with five additional markers (rs7285782, rs9607850, rs2413660, rs714015 and rs133290). Together, these seven markers span most of the gene (from 4.1kb before exon 1 to intron 11; see Figure 2). For the moment, we have not been able to narrow the associated candidate region to a smaller region or a more specific haplotype.

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Sterol regulatory element binding transcription factor 2 (SREBF2) genomic structure and localization of polymorphisms. The figure displays the chromosomal location of SREBF2 and neighboring genes on chromosome 22q13.2. The haplotype-tagging SNPs (htSNPs) (vertical bars) that were screened in the German sample are mapped from the end of CCDC134 (coil–coil domain containing 134), through SREBF2, and TNFRSF13c (Tumor Necrosis Factor Receptor SuperFamily 13C) to CENPM (CENtromer Protein M). All htSNPs significantly associated to schizophrenia in the German population (*) are located in SREBF2. The SNPs rs133280 and rs1052717 (**) were associated to schizophrenia in the three samples and more so in the combined sample. These two markers are in LD with five additional markers (displayed by arrows off rs133280). Those seven markers are located between 4.1kb 5′ to the intron 11 of SREBF2.

Full figure and legend (69K)



We have previously reported that both typical and atypical antipsychotic drugs induce transcriptional upregulation of cholesterol and fatty acid biosynthesis in cultured human glial cells through activation of the sterol regulatory element-binding transcription factors SREBP1 and SREBP2.5, 6 De novo produced cholesterol is a major component of myelin,10 and reduced myelination in the brain has consistently been implicated in the pathophysiology of schizophrenia.2 Against this background, it is interesting that we here report novel association between schizophrenia and several SNPs covering the lipogenesis-controlling SREBF1 and SREBF2 genes. The significant findings were observed in three independent case control samples of German, Norwegian and Danish origin, with several marker allele and genotype distributions for cases versus controls shifted similarly in all samples, thereby giving stronger evidence for association in the merged analysis.

SREBF1 and SREBF2 as novel loci for schizophrenia susceptibility

There is no report of linkage for schizophrenia around the SREBF1 gene per se, but a CAA repeat polymorphism in exon 4 of the neighboring RAI1 gene has been reported as associated with neuroleptic response and schizophrenia.23 Considering this finding and the strong LD in the region around SREBF1 on 17p11.2, it is possible that the association observed with the RAI1 marker is actually picking up signal from a SREBF1 variant, or vice versa. In dbSNP (http://www.ncbi.nlm.nih.gov/projects/SNP/) and the scientific literature, several SNPs have been reported to change the amino acid composition of SREBP1 or to influence the level of cholesterol synthesis,24, 25, 26 but none of these SNPs are deposited in HapMap and we have therefore not been able to conclude on their LD to the associated SNPs of the present study.

The SREBF2 gene is located on chromosome 22q13.2. There have been many reports of linkage between the long arm of chromosome 22 and schizophrenia. Recently, Condra et al.27 linked the marker D22S270 and markers around SULT4A1, to schizophrenia in a linkage scan of 27 families. These markers are located closed to SREBF2 (0.8 and 1.9Mb, respectively), and we suggest that SREBF2 could be an interesting and relevant positional candidate for this linkage. There are several known non-synonymous SNPs in SREBF2, but all have frequencies too low to be used for association tests except for rs4822063, which was typed in the phase I on the German population with no significant results.

At the moment, we are not able to explain the functional implication of the association we report, and further replications, higher density genotyping and re-sequencing in both SREBF1 and SREBF2 genes will be necessary to better characterize genetic variants in those genes that could be functionally relevant for susceptibility to schizophrenia.

Abnormalities in myelination and lipid biosynthesis in schizophrenia

There are now several lines of evidence that implicate oligodendrocyte disturbances and myelin defects in schizophrenia (for review, see Davis et al.2). For instance, in an electron microscopy study in 1972, Miyakawa et al.28 compared a small sample of postmortem biopsies from frontal cortex of patients with schizophrenia and controls. They reported increased amounts of electron dense material in the axon–oligodendrocyte interface and myelin sheats, together with lipofucin-like material in the cytoplasm of oligodendrocytes, whereas astrocytes and microglia were normal. Since this original report, other studies have shown various signs of oligodendrocyte degeneration and myelin sheath abnormalities in prefrontal cortex in brains from patients with schizophrenia.29, 30 MRI studies using diffusion tensor imaging in vivo have indicated abnormalities in white matter in prefrontal cortex and other brain areas of patients with schizophrenia, indicating the possibility of disturbed maintenance of myelination through oligodendrocytes31, 32, 33, 34 although not all studies agree.35, 36 Recently, postmortem gene expression profiling has boosted the field by showing that many myelin-related genes are downregulated in several brain regions of patients with schizophrenia compared with controls37, 38, 39, 40, 41 Furthermore, it is interesting to note that while most of the myelination occurs shortly postnatally,42 the peak of myelination in the cortical areas and especially in the prefrontal cortex occurs during the adolescence and early adulthood,43, 44 thus coinciding with a critical period for the onset of schizophrenia.45, 46

Myelin is composed of 70% cholesterol, and impaired cholesterol biosynthesis by disruption of the squalene synthase (an SREBP-controlled gene), has been shown to severely disrupt myelination.47 The import of cholesterol in the central nervous system is restricted by the blood brain barrier,10 hence de novo synthesis by glial cells is the major source of cholesterol in the brain. The transcription factors SREBP1 and SREBP2 are key players in the control of genes involved in the synthesis of cholesterol as well as fatty acids, phospholipids and triglycerides.48 Although SREBP1 and SREBP2 predominantly control the synthesis of fatty acids and cholesterol, respectively, they also overlap in function, especially for the production of cholesterol.8, 11 When the sterol level becomes low, the SREBP1 and SREBP2 transcription factors are activated in the cell, via activation of a complex formed together with the SREBP cleavage-activating protein (SCAP) and insulin-induced gene (INSIG1 and INSIG2) proteins that are essential for sensing and maintaining cholesterol homeostasis.49 To our knowledge, there is no specific studies so far about the impact of the SREBP pathway in myelination in the central nervous system, but several studies have demonstrated the importance of this pathway in the peripheral myelination of axons.50, 51 The present implication of SREBF1 and SREBF2 as genetic susceptibility factors in schizophrenia further underscores the importance of lipid biosynthesis and myelination in the pathophysiology of schizophrenia.

As myelin defects have also been observed in affective mood disorders and substance abuse (for review, see refs. 4,52,53), suggesting disturbed myelin as a common, if nonspecific, denominator for a variety of neuropsychiatric disorders, it will be interesting, in future work, to assess the association of SREBF1 and SREBF2 variants in samples of bipolar affective disorder, major depressive disorder or alcohol dependence.

Effects of antipsychotic drugs on myelination

Recent reports have shown that atypical antipsychotic drugs may increase remyelination in a mouse model of demyelination,54 and also enhance the volume of white matter in patients with schizophrenia.9, 55 In addition, other reports have shown increased levels of serum lipids in patients treated with olanzapine or clozapine.56, 57 Considering these observations together with our previous reports that several antipsychotics (and also some antidepressants) activate the SREBP pathway,5, 6, 58, 59 and that an upregulation of SREBP-controlled fatty acid biosynthesis-related genes is seen in peripheral blood cells from olanzapine-medicated patients,60 it raises the hypothesis that some antipsychotic drugs may have a beneficial effect on the negative symptoms and cognitive dysfunctions in schizophrenia through a lipogenic activation and secondary actions on the myelination in the brain. We therefore propose to direct more research toward (i) understanding how some of the existing antipsychotic drugs could act on the myelination level or synaptic function in the brain, (ii) identifying genetic or anatomical (e.g. on MRI scans) markers that may determine subgroups of patients that should benefit the most from these treatment61 and (iii) establishing genetic tests and clinical interventions to handle patients with higher risks of antipsychotic induced metabolic adverse effect.61, 62, 63



  1. Carter CJ. Schizophrenia susceptibility genes converge on interlinked pathways related to glutamatergic transmission and long-term potentiation, oxidative stress and oligodendrocyte viability. Schizophr Res 2006; 86: 1–14. | Article | PubMed | ChemPort |
  2. Davis KL, Stewart DG, Friedman JI, Buchsbaum M, Harvey PD, Hof PR et al. White matter changes in schizophrenia: evidence for myelin-related dysfunction. Arch Gen Psychiatry 2003; 60: 443–456. | Article | PubMed | ISI
  3. Karoutzou G, Emrich HM, Dietrich DE. The myelin–pathogenesis puzzle in schizophrenia: a literature review. Mol Psychiatry 2008; 13: 245–260. | Article | PubMed | ChemPort |
  4. Le-Niculescu H, Balaraman Y, Patel S, Tan J, Sidhu K, Jerome RE et al. Towards understanding the schizophrenia code: an expanded convergent functional genomics approach. Am J Med Genet B Neuropsychiatr Genet 2007; 144: 129–158. | PubMed | ChemPort |
  5. Ferno J, Raeder MB, Vik-Mo AO, Skrede S, Glambek M, Tronstad KJ et al. Antipsychotic drugs activate SREBP-regulated expression of lipid biosynthetic genes in cultured human glioma cells: a novel mechanism of action? Pharmacogenomics 2005; 5: 298–304. | Article | ChemPort |
  6. Ferno J, Skrede S, Vik-Mo AO, Håvik B, Steen VM. Drug-induced activation of SREBP-controlled lipogenic gene expression in CNS-related cell lines: marked differences between various antipsychotic drugs. BMC Neurosci 2006; 7: 69. | Article | PubMed | ChemPort |
  7. Edwards PA, Tabor D, Kast HR, Venkateswaran A. Regulation of gene expression by SREBP and SCAP. Biochim Biophys Acta 2000; 1529: 103–113. | Article | PubMed | ISI | ChemPort |
  8. Horton JD, Shimomura I, Brown MS, Hammer RE, Goldstein JL, Shimano H. Activation of cholesterol synthesis in preference to fatty acid synthesis in liver and adipose tissue of transgenic mice overproducing sterol regulatory element-binding protein-2. J Clin Invest 1998; 101: 2331–2339. | Article | PubMed | ISI | ChemPort |
  9. Garver DL, Holcomb JA, Christensen JD. Compromised myelin integrity during psychosis with repair during remission in drug-responding schizophrenia. Int J Neuropsychopharmacol 2008; 11: 49–61. | Article | PubMed | ChemPort |
  10. Dietschy JM, Turley SD. Thematic review series: brain lipids. Cholesterol metabolism in the central nervous system during early development and in the mature animal. J Lipid Res 2004; 45: 1375–1397. | Article | PubMed | ChemPort |
  11. Shimano H, Horton JD, Shimomura I, Hammer RE, Brown MS, Goldstein JL. Isoform 1c of sterol regulatory element binding protein is less active than isoform 1a in livers of transgenic mice and in cultured cells. J Clin Invest 1997; 99: 846–854. | Article | PubMed | ISI | ChemPort |
  12. American Psychiatric Association, Committee on Nomenclature and Statistics. Diagnostic and statistical manual of mental disorders, 4th edn. American Psychiatric Association: Washington, 1994.
  13. First MB, Spitzer RL, Gibbon M, Williams JBW. Structured Clinical Interview for DSM-IV Axis I Disorders. Patient ed. Biometrics Research Department, New York State Psychiatric Institute: New York, 1997.
  14. Farmer AE, Wessely S, Castle D, McGuffin P. Methodological issues in using a polydiagnostic approach to define psychotic illness. Br J Psychiatry 1992; 161: 824–830. | Article | PubMed | ChemPort |
  15. Leckman JF, Sholomskas D, Thompson WD, Belanger A, Weissman MM. Best estimate of lifetime psychiatric diagnosis: a methodological study. Arch Gen Psychiatry 1982; 39: 879–883. | PubMed | ISI | ChemPort |
  16. Miller SA, Dykes DD, Polesky HF. A simple salting out procedure for extracting DNA from human nucleated cells. Nucl Acids Res 1988; 16: 1215. | Article | PubMed | ISI | ChemPort |
  17. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005; 21: 263–265. | Article | PubMed | ISI | ChemPort |
  18. Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B et al. The structure of haplotype blocks in the human genome. Science 2002; 296: 2225–2229. | Article | PubMed | ISI | ChemPort |
  19. Zogopoulos G, Ha KC, Naqib F, Moore S, Kim H, Montpetit A et al. Germ-line DNA copy number variation frequencies in a large North American population. Hum Genet 2007; 122: 345–353. | Article | PubMed | ChemPort |
  20. Iafrate AJ, Feuk L, Rivera MN, Listewnik ML, Donahoe PK, Qi Y et al. Detection of large-scale variation in the human genome. Nat Genet 2004; 36: 949–951. | Article | PubMed | ISI | ChemPort |
  21. Colella S, Yau C, Taylor JM, Mirza G, Butler H, Clouston P et al. QuantiSNP: an Objective Bayes Hidden-Markov Model to detect and accurately map copy number variation using SNP genotyping data. Nucl Acid Res 2007; 35: 2013–2025. | Article | ChemPort |
  22. Hansen T, Olsen L, Lindow M, Jakobsen KD, Ullum H, Jonsson E et al. Brain expressed microRNAs implicated in schizophrenia etiology. PLoS One 2007; 2: e873. | Article | PubMed | ChemPort |
  23. Joober R, Benkelfat C, Toulouse A, Lafrenière RG, Lal S, Ajroud S et al. Analysis of 14 CAG repeat-containing genes in schizophrenia. Am J Med Genet 1999; 88: 694–699. | Article | PubMed | ChemPort |
  24. Eberlé D, Clément K, Meyre D, Sahbatou M, Vaxillaire M, Le Gall A et al. SREBF-1 gene polymorphisms are associated with obesity and type 2 diabetes in French obese and diabetic cohorts. Diabetes 2004; 53: 2153–2157. | Article | PubMed | ISI | ChemPort |
  25. Rios DL, Vargas AF, Torres MR, Zago AJ, Callegari-Jacques SM, Hutz MH. Interaction between SREBP-1a and APOB polymorphisms influences total and low-density lipoprotein cholesterol levels in patients with coronary artery disease. Clin Genet 2003; 63: 380–385. | Article | PubMed | ChemPort |
  26. Salek L, Lutucuta S, Ballantyne CM, Gotto Jr AM, Marian AJ. Effects of SREBF-1a and SCAP polymorphisms on plasma levels of lipids, severity, progression and regression of coronary atherosclerosis and response to therapy with fluvastatin. J Mol Med 2002; 80: 737–744. | Article | PubMed | ISI | ChemPort |
  27. Condra JA, Neibergs H, Wei W, Brennan MD. Evidence for two schizophrenia susceptibility genes on chromosome 22q13. Psychiatry Genet 2007; 17: 292–298. | Article
  28. Miyakawa T, Sumiyoshi S, Deshimaru M, Suzuki T, Tomonari H. Electron microscopic study on schizophrenia. Mechanism of pathological changes. Acta Neuropathol 1972; 20: 67–77. | Article | PubMed | ChemPort |
  29. Uranova N, Orlovskaya D, Vikhreva O, Zimina I, Kolomeets N, Vostrikov V et al. Electron microscopy of oligodendroglia in severe mental illness. Brain Res Bull 2001; 55: 597–610. | Article | PubMed | ISI | ChemPort |
  30. Uranova NA, Vostrikov VM, Orlovskaya DD, Rachmanova VI. Oligodendroglial density in the prefrontal cortex in schizophrenia and mood disorders: a study from the Stanley neuropathology consortium. Schizophr Res 2004; 67: 269–275. | Article | PubMed | ISI
  31. Buchsbaum MS, Tang CY, Peled S, Gudbjartsson H, Lu D, Hazlett EA et al. MRI white matter diffusion anisotropy and PET metabolic rate in schizophrenia. Neuroreport 1998; 9: 425–430. | Article | PubMed | ISI | ChemPort |
  32. Kanaan RA, Kim JS, Kaufmann WE, Pearlson GD, Barker GJ, McGuire PK. Diffusion tensor imaging in schizophrenia. Biol Psychiatry 2005; 58: 921–929. | Article | PubMed
  33. Kubicki M, McCarley RW, Shenton ME. Evidence for white matter abnormalities in schizophrenia. Curr Opin Psychiatry 2005; 18: 121–134. | Article | PubMed
  34. Lim KO, Adalsteinsson E, Spielman D, Sullivan EV, Rosenbloom MJ, Pfefferbaum A. Proton magnetic resonance spectroscopic imaging of cortical gray and white matter in schizophrenia. Arch Gen Psychiatry 1998; 55: 346–352. | Article | PubMed | ChemPort |
  35. van Haren NE, Hulshoff Pol HE, Schnack HG, Cahn W, Mandl RC, Collins DL et al. Focal gray matter changes in schizophrenia across the course of the illness: a 5-year follow-up study. Neuropsychopharmacology 2007; 32: 2057–2066. | Article | PubMed
  36. Molina V, Reig S, Sanz J, Palomo T, Benito C, Sánchez J et al. Increase in gray matter and decrease in white matter volumes in the cortex during treatment with atypical neuroleptics in schizophrenia. Schizophr Res 2005; 80: 61–71. | Article | PubMed | ISI
  37. Tkachev D, Mimmack ML, Ryan MM, Wayland M, Freeman T, Jones PB et al. Oligodendrocyte dysfunction in schizophrenia and bipolar disorder. Lancet 2003; 362: 798–805. | Article | PubMed | ISI | ChemPort |
  38. Hakak Y, Walker JR, Li C, Wong WH, Davis KL, Buxbaum JD et al. Genome-wide expression analysis reveals dysregulation of myelination-related genes in chronic schizophrenia. Proc Natl Acad Sci 2001; 98: 4746–4751. | Article | PubMed | ChemPort |
  39. Katsel P, Davis KL, Haroutunian V. Variations in myelin and oligodendrocyte-related gene expression across multiple brain regions in schizophrenia: a gene ontology study. Schizophr Res 2005; 79: 157–173. | Article | PubMed | ISI
  40. Aston C, Jiang L, Sokolov BP. Microarray analysis of postmortem temporal cortex from patients with schizophrenia. J Neurosci Res 2004; 77: 858–866. | Article | PubMed | ISI | ChemPort |
  41. Aston C, Jiang L, Sokolov BP. Transcriptional profiling reveals evidence for signaling and oligodendroglial abnormalities in the temporal cortex from patients with major depressive disorder. Mol Psychiatry 2005; 10: 309–322. | Article | PubMed | ISI | ChemPort |
  42. de Graaf-Peters VB, Hadders-Algra M. Ontogeny of the human central nervous system: what is happening when? Early Hum Dev 2006; 82: 257–266. | Article | PubMed
  43. Paus T, Zijdenbos A, Worsley K, Collins DL, Blumenthal J, Giedd JN et al. Structural maturation of neural pathways in children and adolescents: in vivo study. Science 1999; 283: 1908–1911. | Article | PubMed | ISI | ChemPort |
  44. Giedd JN, Blumenthal J, Jeffries NO, Castellanos FX, Liu H, Zijdenbos A et al. Brain development during childhood and adolescence: a longitudinal MRI study. Nat Neurosci 1999; 2: 861–863. | Article | PubMed | ISI | ChemPort |
  45. Bartzokis G, Beckson M, Lu PH, Nuechterlein KH, Edwards N, Mintz J. Age-related changes in frontal and temporal lobe volumes in men: a magnetic resonance imaging study. Arch Gen Psychiatry 2001; 58: 461–465. | Article | PubMed | ISI | ChemPort |
  46. Bartzokis G. Schizophrenia: breakdown in the well-regulated lifelong process of brain development and maturation. Neuropsychopharmacology 2002; 27: 672–683. | Article | PubMed | ISI
  47. Saher G, Brügger B, Lappe-Siefke C, Möbius W, Tozawa R, Wehr MC et al. High cholesterol level is essential for myelin membrane growth. Nat Neurosci 2005; 8: 468–475. | Article | PubMed | ISI | ChemPort |
  48. Horton JD, Goldstein JL, Brown MS. SREBPs: activators of the complete program of cholesterol and fatty acid synthesis in the liver. J Clin Invest 2002; 109: 1125–1131. | Article | PubMed | ISI | ChemPort |
  49. Loewen CJ, Levine TP. Cholesterol homeostasis: not until the SCAP lady INSIGs. Curr Biol 2002; 12: R779–R781. | Article | PubMed | ISI | ChemPort |
  50. Leblanc SE, Srinivasan R, Ferri C, Mager GM, Gillian-Daniel AL, Wrabetz L et al. Regulation of cholesterol/lipid biosynthetic genes by Egr2/Krox20 during peripheral nerve myelination. J Neurochem 2005; 93: 737–748. | Article | PubMed | ISI | ChemPort |
  51. Verheijen MH, Chrast R, Burrola P, Lemke G. Local regulation of fat metabolism in peripheral nerves. Genes Dev 2003; 17: 2450–2464. | Article | PubMed | ISI | ChemPort |
  52. Sokolov BP. Oligodendroglial abnormalities in schizophrenia, mood disorders and substance abuse. Comorbidity, shared traits, or molecular phenocopies? Int J Neuropsychopharmacol 2007; 10: 547–555. | Article | PubMed | ChemPort |
  53. Le-Niculescu H, McFarland MJ, Ogden CA, Balaraman Y, Patel S, Tan J et al. Phenomic, convergent functional genomic, and biomarker studies in a stress-reactive genetic animal model of bipolar disorder and co-morbid alcoholism. Am J Med Genet B Neuropsychiatr Genet 2008; 147B: 134–166. | Article
  54. Xiao L, Xu H, Zhang Y, Wei Z, He J, Jiang W et al. Quetiapine facilitates oligodendrocyte development and prevents mice from myelin breakdown and behavioral changes. Mol Psychiatry 2008; 13: 697–708. | Article | PubMed | ChemPort |
  55. Bartzokis G, Lu PH, Nuechterlein KH, Gitlin M, Doi C, Edwards N et al. Differential effects of typical and atypical antipsychotics on brain myelination in schizophrenia. Schizophr Res 2007; 93: 13–22. | Article | PubMed
  56. Huang TL, Lu CY. Correlations between weight changes and lipid profile changes in schizophrenic patients after antipsychotics therapy. Chang Gung Med J 2007; 30: 26–32. | PubMed |
  57. Procyshyn RM, Wasan KM, Thornton AE, Barr AM, Chen EY, Pomarol-Clotet E et al. Changes in serum lipids, independent of weight, are associated with changes in symptoms during long-term clozapine treatment. J Psychiatry Neurosci 2007; 32: 331–338. | PubMed |
  58. Raeder MB, Ferno J, Glambek M, Stansberg C, Steen VM. Antidepressant drugs activate SREBP and up-regulate cholesterol and fatty acid biosynthesis in human glial cells. Neurosci Lett 2006; 395: 185–190. | Article | PubMed | ISI | ChemPort |
  59. Raeder MB, Ferno J, Vik-Mo AO, Steen VM. SREBP activation by antipsychotic- and antidepressant-drugs in cultured human liver cells: relevance for metabolic side-effects? Mol Cell Biochem 2006; 289: 167–173. | Article | PubMed | ChemPort |
  60. Vik-Mo AO, Birkenaes AB, Fernø J, Jonsdottir H, Andreassen OA, Steen VM. Increased expression of lipid biosynthesis genes in peripheral blood cells of olanzapine-treated patients. Int J Neuropsychopharmacol 2008; 11: 679–684. | Article | PubMed | ChemPort |
  61. Arranz MJ, de Leon J. Pharmacogenetics and pharmacogenomics of schizophrenia: a review of last decade of research. Mol Psychiatry 2007; 12: 707–747. | Article | PubMed | ChemPort |
  62. Le Hellard S, Theisen FM, Haberhausen M, Raeder MB, Fernø J, Gebhardt S et al. Association between the insulin-induced gene 2 (INSIG2) and weight gain in a German sample of antipsychotic-treated schizophrenic patients: perturbation of SREBP-controlled lipogenesis in drug-related metabolic adverse effects?. Mol Psychiatry 2008; e-pub ahead of print: doi:10.1038.
  63. Malhotra AK, Lencz T, Correll CU, Kane JM. Genomics and the future of pharmacotherapy in psychiatry. Int Rev Psychiatry 2007; 19: 523–530. | Article | PubMed


This study is based on an initial microarray-based gene expression analysis on antipsychotic drug action, using the infrastructure provided by the Norwegian Microarray Consortium FUGE technology platform (www.microarray.no), funded by the FUGE program of the Research Council of Norway. The study was supported by grants from the Research Council of Norway (incl. FUGE grant nos. 151904 and 183327, and Psykisk Helse grant no. 175345), Helse Vest RHF and Dr Einar Martens Fund. MMN, SC received support for their work from the Alfried Krupp von Bohlen und Halbach-Stiftung. CV was supported through a Marie Curie grant as part of the Research Training Netweork ‘EUTwins’ to MMN, SC, IN, HS (EU, FP6). The Norwegian TOP study was supported by Research Council of Norway (grants nos. 167153/V50 and 163070/V50) and the Eastern Norway Health Authority. We are indebted to the patients for their participation in this study.

Supplementary Information accompanies the paper on the Molecular Psychiatry website