PATIENTS AND METHODS

Patients and random controls

All patients were white and from southern Germany. The diagnosis of schizophrenia was performed at the Psychiatric Hospital of the University of Munich. Schizophrenia was diagnosed according to the third edition of the “Diagnostic and Statistical Manual of Mental Disorders” (DSM-III-R) in patients in panel 1 and panel 2. All patients were inpatients at the time of diagnosis. Different schizophrenic subtypes had no impact on the observed deviation of frequencies of HLA class I alleles and antigens (not shown).

Patient panel 1 (n=31) was previously collected to study HLA class I antigens and HLA class II alleles. Results of HLA class II alleles have already been published.¹³ Patient panel 2 (n=71) was collected later over a three year period and HLA-C polymorphism was investigated by sequence based typing (SBT) for best resolution. Only 60 of these 71 patients could be used for serological typing.

Samples for the random control panels were collected from members of the Institute of Immunology in Munich and from medical students at the University of Munich. All subjects were white. The frequencies of HLA antigens did not deviate from other published German HLA antigen frequencies.

Isolation of peripheral mononuclear cells (PBMC), RNA, and HLA typing

Peripheral blood mononuclear cells (PBMC) from healthy random controls and schizophrenic patients were isolated by Ficoll gradient centrifugation.

Serological HLA-A typing was performed according to the NIH method with some modifications to the original technique.¹⁴ Our own monoclonal antibodies, reagents obtained by exchange, and commercial reagents were applied. A total of 28 HLA-A and -C antigens were defined using a battery of 120 reagents. Total RNA was prepared from freshly isolated cells or after CD3 stimulation using the Micro RNA isolation kit (Stratagene, La Jolla, Ca 92037, USA). RNA was transcribed into cDNA by RAV-2 reverse transcriptase (Amersham, Braunschweig, Germany) and amplified by PCR. Sequence, length, and localisation of primers were as previously described.¹² Sequencing primers were hybridised to conservative regions of HLA alleles. Amplified PCR products were sequenced after strand separation with streptavidin coated Dynabeads M-280 (Dynal, Hamburg, Germany). Sequencing was performed using fluoroprime or indodicabocyanine labelled primers and gels were run on an automated DNA sequencer (ALF Express, Pharmacia, Freiburg, Germany). Both strands of every PCR product were sequenced. Twenty-three different HLA-C alleles were identified in patients and controls (not shown).

Statistical evaluation

HLA-A and HLA-Cw antigen frequencies of patients and random controls as well as HLA-Cw* allelic frequencies (not shown) were compared using Pearson χ² statistics with 95% confidence intervals; exact significance p values were determined two tailed (table 1). Uncorrected significant p values were confirmed in HLA-A10 in panel 2. Therefore, the p value of HLA-A10 did not require correction by investigated antigens. Data were calculated using the statistical programme SPSS (version 10.1.3, SPSS Inc).

RESULTS

Both patient panels and controls were typed for 20 HLA-A antigens and eight HLA-C antigens serologically. The following HLA-C alleles were identified by sequence based typing: HLA-Cw*0102, -Cw*02022, -Cw*0302, -Cw*0303, -Cw*0304, -Cw*0401, -Cw*0501, -Cw*0602, -Cw*0701, -Cw*0702, -Cw*0704, -Cw*0802, -Cw*1202, -Cw*1203, -Cw*1502, -Cw*1504, -Cw*1505, and -Cw*1601. The random control panel is not shown because the results of HLA-C typing of a major part of the control panel has previously been published.¹² As shown there, HLA-Cw5 was encoded only by one allele, HLA-Cw*0501. Nine patients were identified as carrying HLA-Cw*0501 and these patients also expressed HLA-Cw5 serologically. We did not find a significant difference between HLA-SBT typing and serological typing for HLA-Cw 5/HLA-Cw*0501. No significant HLA-C allelic frequency deviation was found in patient panel 2 as compared to the random controls (not shown), nor of HLA-Cw4 as previously published.¹⁵ Therefore the results were restricted to serological typings of patient panel 1 and patient panel 2 (table 1A). From 20 serologically tested HLA-A antigens (not shown), only HLA-A10 showed a significant frequency deviation. The results of HLA-A10 in patient panel 1 (p=0.025, corrected, not significant) was replicated in patient panel 2 (p=0.006, p value correction not required, legend table 1)

DISCUSSION

Many microbial factors have been implicated in the pathogenesis of schizophrenia, but so far each microbial factor has been identified in a relatively small subgroup of patients.^16–20 The heterogeneity of these microbial factors is also reflected by the associations with different HLA loci and their alleles. Polymorphic HLA alleles encode individual smell receptors²¹ and influence mating preferences.²² Polymorphic HLA molecules process, select, and present degraded microbial proteins.^23,24 The set of inherited HLA alleles determines susceptibility or resistance to particular microbes.

We recently analysed immune response genes of the HLA class II region in patient panel 1 with sequence specific oligonucleotide typing and reported an increased frequency of the narcolepsy and multiple sclerosis associated HLA class II allele DQB1*0602 in the schizophrenic patients.¹³ A former report of a strongly increased frequency of HLA-Cw4 in Czech schizophrenic patients¹⁵ inspired us to reinvestigate HLA-C polymorphism by sequencing the schizophrenia patient panel 2 (not shown). However, we did not find a significantly increased frequency of any HLA-Cw allele in the patients, rather a decreased frequency of the allele HLA-Cw*0501 (not shown). This allele is the only one to encode the serologically defined antigen HLA-Cw5.¹² Although the results of patient panels 1 and 2 showed a similarly decreased frequency of HLA-Cw5, the decrease in frequency did not reach significance (table 1A). The significant serologically defined increased frequency of HLA-A10 in patient panel 1 (p=0.025, table 1A) caused us to look at the frequency of HLA-A10 in patient panel 2. We found a nearly identical increased frequency of HLA-A10 in patient panel 2 (table 1B).

It is worth noting that higher resolution of antigenic determinants, either by HLA-Cw sequencing or further subdivision of the HLA-A10 antigen (not shown), did not strengthen the degree of association with schizophrenia. This could indicate that several different microbial factors were presented by this public HLA-A10 antigen, but we favour a lesser association of HLA-A10 with schizophrenia because of genes in the chromosomal vicinity of the HLA-A locus. Two loci in the close neighbourhood of HLA-A (fig 1) appear to be very good candidates for further investigations, HLA-G and HLA-E. Both genes are polymorphic and play an important role during embryogenesis. They are also expressed by activated immune cells in adults. Investigations of polymorphisms in this genetic region could be particularly interesting in mothers causing schizophrenic birth excesses, infected by influenza epidemics, or by Borrelia Burgdorferi.^15,25 Infected immune cell subpopulations may fail to protect embryonal brain development or fail to produce, for example, brain derived neurotrophic factor (BDNF).³ Infections of such subpopulations in adults could modulate/disturb the neurotrophic system and thereby influence neuronal adaptive responses. Restoration of such circuits by adoptive immunotherapy may have led to the substantial improvement observed in three psychiatric patients, suffering from autism, bipolar depression, or schizophrenia.²⁶ It will be worthwhile to study HLA polymorphisms further in order to identify immunogenetic dispositions, that is, HLA alleles lacking resistance against certain microbial infections. It will be important to identify in these chronically infected patients immune cell subpopulations that function in communication with the central nervous system.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Localisation of discussed non-classical class I genes within the HLA class I region. HLA-B, -C, and -A encode the classical class I MHC molecules. HLA-E, -F, and -G represent non-classical class I genes encoding HLA class Ib molecules. HLA-X, -J, and -H are at present categorised as pseudogenes.