Are KIR and HLA class I genes associated with schizophrenia?

Abstract

Schizophrenia is a chronic and very serious psychiatric disorder, the etiology and pathophysiology of which has not been firmly established. It is multifactorial, and genetic, immunologic, and environmental factors play a role (1–3). It has been shown that infections and autoimmune diseases increase the risk of schizophrenia (4, 5). Natural killer (NK) cells are elements of innate immunity, important for the defense against viral infections and oncogenesis by lysis of infected and transformed cells, respectively, and they also contribute to autoimmune diseases (6). The activity of NK cells in schizophrenia patients is increased, whichmay suggest a contribution of these cells to the disease (7, 8). Activation of NK cells depends on a balance of activating and inhibitory signals from cell surface receptors. Among these receptors, the most genetically polymorphic are killer cell immunoglobulin-like receptors (KIRs). These receptors are encoded by 15 genes on chromosome 19, but individuals differ in number and kind (activating vs inhibitory) of KIR genes, which influences their susceptibility to viral infections, neoplasms, and autoimmune diseases (9). Ligands of KIRs, when known, are specific epitopes on human leukocyte antigen (HLA) class I molecules (6): all HLA-C allotypes fall into C1 or C2 group (asparagine and lysine, respectively, at position 80) recognized byKIR2DL2/KIR2DL3 and KIR2DL1/KIR2DS1, respectively. HLA-Bw4 is an epitope present on roughly half of HLA-B allotypes, as well as on some HLA-A molecules, and is recognized by KIR3DL1. Some other KIRs bind also some HLA class I molecules (10). We hypothesized that someKIR andHLA class I genesmight be associated with schizophrenia. Here, we report the results of our comparison of KIR genes and HLA class I-encoded KIR ligand frequencies in patients with schizophrenia and healthy control individuals in the Polish Caucasian population. Our case–control study enrolled 200 unrelated schizophrenic patients (113 women and 87 men). The mean age at blood sampling was 37.8 years [standard deviation (SD): 12.25, range: 21–75]. All patients were diagnosed at the Wroclaw Medical University Hospital according to ICD-10 (11) and DSM-IV (12) criteria based on interview data and hospital case notes. In addition to unstructured interviews and review of medical records, the patients were evaluated for lifetime psychotic symptomatology using the Operational Criteria for Psychotic Illness (OPCRIT) checklist, which provides a polydiagnostic categorical and dimensional approach to diagnosis of schizophrenia (13). Patients having a history of traumatic brain injury, neurologic disorders, and substance abuse were excluded from the study by detailed medical examination. The control group consisted of 561 (326 women and 235 men) healthy subjects. They were volunteers declaring good health condition and lack of psychiatric abnormalities. The mean age at blood sampling was 39.3 years (SD: 14.67, range: 19–83). The study was approved by the Bioethics Committee of the Wroclaw Medical University. Signed informed consent was obtained from all the subjects. DNA isolation was done using Invisorb Spin Blood Midi Kit (Invitek, Berlin, Germany) following the manufacturer’s instructions, or according to Gustincich et al. (14). KIR genotyping was performed by single (15) or multiplex (16) polymerase chain reaction using sequence-specific primers. KIR ligands were established using primers for HLA-C C1 and C2 (17) as well as a commercial kit (Olerup SSP®KIRHLALigand) for HLA-CAsparagine80 – (C1), HLA-CLysine80 – (C2), HLA-ABw4+, HLA-B Bw4Isoleucine80, and HLA-BThreonine80 (Olerup SSP AB, Saltsjobaden, Sweden). Statistical analysis was performed as follows: global test for difference between two sets of k dependent proportions i.e. x1 = (p11, ... , p1k) and x2 = (p21, ... , p2k) was T = ‖x1−x2‖2 SE‖x1−x2‖2 , where distribution of T statistic was estimated numerically. Proportions were transformated to p′ = sin−1( √ p). Odds ratio and its 95% confidence interval was used as a measure of effect size. Comparison of all studied KIR gene frequencies in 200 patients with schizophrenia and 561 healthy control individuals did not reveal any significant difference (pglobal = 0.15). Similarly, no differences were found in distribution of KIR gene ligands (pglobal = 0.73 and 0.67 for the HLA-C1, -C2 and for the HLA-ABw4+, HLA-B Bw4Ile, and HLA-B Bw4Thr, respectively). These results seem contradictory to those of the recent GWAS genome-wide association study (18), where a strong association of HLA-C*01:02 allele with schizophrenia was observed. The reason for this discrepancy may lay in the fact that (i) we typed only for C1 and C2 epitopes of HLA-C but not for particular HLA-C alleles, whereas the HLA-C*01:02 allele is rare in Caucasians including Poles (19) and constitutes