Serum kynurenic acid is reduced in affective psychosis.

A subgroup of individuals with mood and psychotic disorders shows evidence of inflammation that leads to activation of the kynurenine pathway and the increased production of neuroactive kynurenine metabolites. Depression is hypothesized to be causally associated with an imbalance in the kynurenine pathway, with an increased metabolism down the 3-hydroxykynurenine (3HK) branch of the pathway leading to increased levels of the neurotoxic metabolite, quinolinic acid (QA), which is a putative N-methyl-d-aspartate (NMDA) receptor agonist. In contrast, schizophrenia and psychosis are hypothesized to arise from increased metabolism of the NMDA receptor antagonist, kynurenic acid (KynA), leading to hypofunction of GABAergic interneurons, the disinhibition of pyramidal neurons and striatal hyperdopaminergia. Here we present results that challenge the model of excess KynA production in affective psychosis. After rigorous control of potential confounders and multiple testing we find significant reductions in serum KynA and/or KynA/QA in acutely ill inpatients with major depressive disorder (N=35), bipolar disorder (N=53) and schizoaffective disorder (N=40) versus healthy controls (N=92). No significant difference was found between acutely ill inpatients with schizophrenia (n=21) and healthy controls. Further, a post hoc comparison of patients divided into the categories of non-psychotic affective disorder, affective psychosis and psychotic disorder (non-affective) showed that the greatest decrease in KynA was in the affective psychosis group relative to the other diagnostic groups. Our results are consistent with reports of elevations in proinflammatory cytokines in psychosis, and preclinical work showing that inflammation upregulates the enzyme, kynurenine mono-oxygenase (KMO), which converts kynurenine into 3-hydroxykynurenine and quinolinic acid.

Gene expression responses in lymphoblastoid cells after radiation exposure.

Individual differences in response to radiation are well known, but the molecular basis for these differences is not well understood, and molecular indicators that are useful in assessing individual variation are lacking. Cells from patients developing unexpected radiation responses have occasionally been analyzed for rare genetic anomalies (such as alleles of the ATM gene), but few data exist on the long-term effects of genetic variation on radiation response. We hypothesize that much of the variation in the response to radiation is due to differences in the genes that respond to radiation exposure, and that changes in gene expression may serve as surrogate markers of individual response. As a first step in developing a selection of suitable markers of gene expression, we used cDNA microarrays to identify genes that were altered in expression in lymphoblastoid cells 4 h after exposure to 1 Gy X rays. We found changes in gene expression ranging from a 10-fold repression to a 12-fold induction. Some of the responsive genes have been noted previously in other cell types, whereas others are reported for the first time. Using these data, we are beginning to characterize the range of structural, temporal and functional variations in the responsive genes. The results of this work will assist in developing response markers both for prescreening for sensitive individuals and for risk assessment.

Identification of single nucleotide polymorphisms in human DNA repair genes.

Variation in gene coding sequence represents a significant factor in predisposition to disease, including cancer. Variants of some DNA repair genes (e.g. MLH1, MSH2 and MSH6) are known to predispose to cancer. We identified single nucleotide polymorphisms (SNPs) in five DNA repair genes in 142 healthy individuals using a DNA sequencing protocol optimized for the direct detection of single nucleotide polymorphisms. This approach, called the heterozygote sequencing protocol (HSP), enables moderate-scale population surveys of SNPs. HSP uses fluorescently tagged primers and exploits the large dynamic range and low background of automated fluorescent sequencing. HSP may be used for any sequence that can be amplified by PCR. A total of 12 SNP variants in MGMT, ERCC1, CDK7, CCNH and XRCC4 were identified, 11 at polymorphic frequencies, with an average frequency of 0.22 (95% confidence interval 0.20-0.24). Among the 82 individuals for whom complete SNP profiles were available, no one person carried the GenBank reference sequence for all five genes. The extensive heterogeneity observed in these five genes is intriguing. All variants are in Hardy-Weinberg equilibrium, although the meaning of this equilibrium is unclear. Using this approach, possible associations of sequence variation, and hence of variation in DNA repair, with disease risk can be assessed.

Identification of genes responsive to BPDE treatment in HeLa cells using cDNA expression assays.

Genotoxic stresses induce cellular responses that can be observed at the level of gene expression. We have studied changes in gene expression following BPDE exposure in HeLa cells by using a cDNA expression array of 597 human genes. After a 53-hr exposure to 0.4 microM BPDE, nine genes were upregulated. The protein products of these genes are: fos-related antigen 2, apoptotic cysteine protease MCH4, DB1 (zinc finger protein 91), transcription factor ETR103, integrin alpha, interleukin-4, interleukin-6, 23-kDa highly basic protein, and ribosomal protein S9. We observed the downregulation of gene expression of three genes: heat-shock protein 27, DNA-binding protein TAX, and NADH-ubiquinone oxidoreductase B18 subunit. These results suggest unknown functions or regulatory circuits for several of the responsive genes and demonstrate the complexity of cellular responses to genotoxic insults.

Screening a human population sample for DNA repair gene deficiencies utilizing the protein truncation test.

A significant fraction of human cancers are thought to have a genetic component and several lines of evidence suggest that deficiencies in DNA repair may be a contributing factor. Little is known, however, about the frequency and distribution of variants of DNA repair genes in the general human population. The protein truncation test (PTT) was used to screen 136 healthy volunteers for protein-truncating variants of 10 DNA repair genes: APE, CDK7, ERCC1, WAF1, HOGG1, MGMT, POLB, UNG, HAAG, and CCNH. This sample consisted of 41males (30%) and 95 females (70%) with an average age of 25.3 years, ranging from 17 to 60 years of age. No truncating mutations were found in the 10 genes examined in any of the subjects. The 95% confidence interval for a proportion of 0 over the 272 alleles examined per locus is 0-0.01. The calculated frequency of truncating mutations in each of these genes, among the general population, is thus less than 1%. Among the 10 genes tested in 136 people, a single sample had no PCR product for HAAG, even though PCR products were obtained on all other genes. Total RNA dot hybridization confirmed the presence of HAAG mRNA transcripts in this sample. Despite identification of this single DNA repair variant, these results indicate a low frequency of truncating mutations in DNA repair genes in the general population.

Human DNA repair systems: an overview.

DNA repair systems act to maintain genome integrity in the face of replication errors, environmental insults, and the cumulative effects of age. More than 70 human genes directly involved in the five major pathways of DNA repair have been described, including chromosomal location and cDNA sequence. However, a great deal of information as to the precise functions of these genes and their role in human health is still lacking. Hence, we summarize what is known about these genes and their contra part in bacterial, yeast, and rodent systems and discuss their involvement in human disease. While some associations are already well understood, it is clear that additional diseases will be found which are linked to DNA repair defects or deficiencies.

Strand bias in mutation involving 5-methylcytosine deamination in the human hprt gene.

Despite being generally under-represented in the genome, CpG sequences represent a disproportionately high fraction of sites involved in mutational events leading to human genetic disease. Cytosine within CpG dinucleotides is often modified to 5-methylcytosine. Deamination of 5-methylcytosine in situ yields a thymine, which being mispaired with guanine, is potentially mutagenic. Previous reports have indicated that most mutations recovered at these sites appear to originate on the non-transcribed strand as C-->T transitions. This trend may however, reflect the lack of detectable mutant phenotypes resulting from this transition at the complementary positions on the transcribed strand. To date, there has not been a good model system in which mutations can be recovered on both strands at the same CpG site. The human hprt gene has MeCpG sites contained within arginine codons for which mutations have been recovered on both strands. From an analysis of a database of hprt mutations, a statistically significant strand bias is observed in mutations recovered at CpG sites. We describe some models for the bias of mutation distribution observed at MeCpG sites in light of this and previous work are described.