Analysis of a polymorphism in the promoter region of the tumor necrosis factor alpha gene in schizophrenia and bipolar disorder: further support for an association with schizophrenia

Sir – Activation of the inflammatory response system has been observed in the schizophrenia and affective disorders.1,2 Investigations of the role of the inflammatory response in these psychiatric disorders include analysis of the immune mediators at the gene level, so the genetic polymorphism of specific cytokine loci has been investigated in psychiatric disorders.3,4 Recently, Boin et al5 investigated a biallelic variant (−G308A) of the TNF-α gene directly related to increased production of these cytokines, and found a significant association with schizophrenia. Interestingly, the TNF-α gene is located in the region 6p21.1–21.3, near the HLA region, where linkage and association studies have suggested a locus of susceptibility to schizophrenia.6

In the present investigation, we performed a cross-sectional study to verify whether the −G308A TNF-α gene polymorphism was associated with a predisposition to schizophrenia and/or bipolar affective disorder (BPAD) in a Brazilian sample.

A total of 186 patients who had the DSM-III-R criteria for schizophrenia and 161 patients who had the DSM-III-R criteria for BPAD were selected for the study from the public psychiatric assistance services in the Sao Paulo metropolitan area, based on the analysis of medical records and interviews conducted by a trained psychiatrist. The demographic and clinical characteristics of the patient groups were as follows: The schizophrenic group had a mean age of 36 years (SD±8 years), 47% female, and the whole group had the following ethnic makeup: 57% Caucasian, 36% mulatto, and 7% black patients. The mean age at onset of illness was 20.6 years (SD±6 years), with the subtypes of schizophrenia being: paranoid: 59%; disorganized: 28%; undifferentiated: 6%; residual: 5%; and catatonic: 2%. In the BPAD patient group, 56% was female, mean age of 42 years (SD±11 years); and an ethnic makeup that was 59% Caucasian, 34% mulatto, and 7% black patients. The mean age of onset of illness was 21.8 years (SD±7 years), all patients had at least one manic episode, and the mean number of episodes was eight. In all, 79% of the patients had a history of psychotic symptoms.

A total of 657 controls were randomly recruited from unrelated subjects admitted to a routine check-up outpatient clinic. Venous blood samples were collected after written informed consent was obtained. The demographic characteristics of this group were: 57% females, mean age of 40 years (SD±11 years). The group was 67% Caucasian, 29% mulatto, and 4% black patients.

Genomic DNA was extracted using a ‘salting out’ method, and the primers and PCR conditions were described by Wilson et al.7 The PCR product was then digested with NcoI, which cleaved the guanine (G) containing variant at position −308 of the TNF-α gene (TNF1 allele) into two fragments, one 87 bp and the other 20 bp in length. The variant with adenine (A) at the −308 position (TNF2 allele) remained undigested and yielded a 107 bp fragment.

Different allele and genotype frequencies in each group were compared using standard chi-square (χ2) analysis. Odds ratios (ORs) were calculated using logistic regression models considering both a dominant and a recessive genetic mechanism. These models were controlled for age, gender, and ethnicity. Analyses were performed using Statview software for Windows. The population attributable fraction was estimated using the equation pd(RR−1)/RR, where pd represents the proportion of subjects with the risk factor and RR is the adjusted relative risk or adjusted OR.8

The distribution of genotype and allele frequencies of the −G308A TNF-α gene polymorphism, in the studied groups, is shown in Table 1. The TNF-α gene variant did not present a significant difference in the allele and genotype distribution between the BPAD and the control groups (χ2=2.7, 2 df, P=0.2). In contrast, a significantly higher frequency of the TNF2 allele and of the homozygous genotype was observed in the schizophrenic group in comparison with the control group (χ2=12.8, 2 df, P=0.001). Multiple logistic regression analysis adjusted for gender, age, and ethnicity revealed an OR of 3.7 (CI 95%, 1.2–12) for the association of the homozygous genotype of the TNF2 allele with schizophrenia in a recessive genetic model. The population-attributable fraction of the TNF2 homozygous genotype in schizophrenia was estimated at 3%.

Table 1 Distribution of alleles and genotypes of the −G308A TNF-α gene polymorphism, among schizophrenics, BPAD patients, and controls

The aim of the present study was to explore the association between −G308A TNF-α gene polymorphism and BPAD and/or schizophrenia. The data obtained did not indicate association of this variant with BPAD, and were in agreement with a previous study.4 However, consistent with the observations of Boin et al,5 we did find a significant association of this polymorphism with schizophrenia.

Previous studies have shown that the level of TNF-α production is under genetic control and determined in part by alleles of the −G308A TNF-α promoter polymorphism.9 TNF-α is a key factor in the inflammatory process that may exert either beneficial or deleterious effects depending on the timing, localization, and extent of its release. Boin et al5 speculate that abnormal TNF-α production, triggered by viral and environmental insult during critical stages of brain development, could increase the risk of developing schizophrenia later in life in TNF2 allele carriers.

Population differences in the genetic structure of the alleles studied could feasibly generate a false positive result, so we adjusted for age, gender, and ethnicity in our logistic regression analysis. In addition, when stratified by ethnicity, no significant difference was observed in the allele and genotype distribution among the groups (data not shown).

The TNF2 allele is known to be in strong linkage disequilibrium with the major histocompatibility complex (MHC) haplotype HLA A1-B8-DR3-DQ2.10 The MHC region is rich in polymorphic genes.11 We therefore cannot at present exclude the possibility that the association demonstrated in this study was because of other genetic polymorphisms in the same region. However, there is a good rationale for the selection of this candidate gene. In addition, two new observations of the association between this specific polymorphism and schizophrenia have been reported.12,13

In conclusion, our data showed an association between the −G308A TNF-α polymorphism and schizophrenia, similar to that observed by Boin et al. Our larger sample suggests that these findings are not false-positive results, and supports the hypothesis that the TNF-α locus plays a role in susceptibility to schizophrenia.


  1. 1

    Kronfol Z, Remick DG . Am J Psychiatry 2000; 157: 683–694.

  2. 2

    Prolo P, Licinio J . Mol Psychiatry 1999; 4: 109–111.

  3. 3

    Katila H, Hanninen K, Hurme M . Mol Psychiatry 1999; 4: 179–181.

  4. 4

    Middle F, Jones I, Robertson E, Lendon C, Craddock N . Psychiatr Genet 2000; 10: 195–198.

  5. 5

    Boin F, Zanardini R, Pioli R, Altamura CA, Maes M, Gennarelli M . Mol Psychiatry 2001; 6: 79–82.

  6. 6

    Wright P, Nimgaonkar VL, Donaldson PT, Murray RM . Schizophr Res 2001; 47: 1–12.

  7. 7

    Wilson AG, Di Giovine FS, Blackmore AL, Duff GW . Hum Mol Gen 1992; 1: 353.

  8. 8

    Rockhill B, Newman B, Weinberg C . Am J Public Health 1998; 88: 15–19.

  9. 9

    Wilson AG, Symons JA, McDowell TL, McDevitt HO, Duff GW . Proc Natl Acad Sci USA 1997; 94: 3195–3199.

  10. 10

    Wilson AG, de Vries N, Pociot F, di Giovine FS, van der Putt LB, Duff GW . J Exp Med 1993; 177: 557–560.

  11. 11

    The MHC Sequencing Consortium*. Nature 1999; 401: 921–923.

  12. 12

    Mondabon S, Schwab SG, Knapp M, Albens M, Hallmayer J, Borrmann-Hassenbach M, Lerer B, Maier W, Wildenauer DB . Am J Med Genet 2002; 114: 861 (abstract).

  13. 13

    Tan EC, Chong AS, Tan CH, Teo YY, Peng KT, Mahendran R . Am J Med Genet 2002; 114: 874–875 (abstract).

Download references

Author information

Correspondence to H Vallada.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Meira-Lima, I., Pereira, A., Mota, G. et al. Analysis of a polymorphism in the promoter region of the tumor necrosis factor alpha gene in schizophrenia and bipolar disorder: further support for an association with schizophrenia. Mol Psychiatry 8, 718–720 (2003).

Download citation

Further reading