Gene expression in peripheral blood mononuclear cells from patients with chronic fatigue syndrome

BACKGROUND

Chronic fatigue syndrome (CFS) is a multisystem disease, the pathogenesis of which remains undetermined.

AIMS

To test the hypothesis that there are reproducible abnormalities of gene expression in patients with CFS compared with normal healthy persons.

METHODS

To gain further insight into the pathogenesis of this disease, gene expression was analysed in peripheral blood mononuclear cells from 25 patients with CFS diagnosed according to the Centers for Disease Control criteria and 25 normal blood donors matched for age, sex, and geographical location, using a single colour microarray representing 9522 human genes. After normalisation, average difference values for each gene were compared between test and control groups using a cutoff fold difference of expression > or = 1.5 and a p value of 0.001. Genes showing differential expression were further analysed using Taqman real time polymerase chain reaction (PCR) in fresh samples.

RESULTS

Analysis of microarray data revealed differential expression of 35 genes. Real time PCR confirmed differential expression in the same direction as array results for 16 of these genes, 15 of which were upregulated (ABCD4, PRKCL1, MRPL23, CD2BP2, GSN, NTE, POLR2G, PEX16, EIF2B4, EIF4G1, ANAPC11, PDCD2, KHSRP, BRMS1, and GABARAPL1) and one of which was downregulated (IL-10RA). This profile suggests T cell activation and perturbation of neuronal and mitochondrial function. Upregulation of neuropathy target esterase and eukaryotic translation initiation factor 4G1 may suggest links with organophosphate exposure and virus infection, respectively.

CONCLUSION

These results suggest that patients with CFS have reproducible alterations in gene regulation.

Evaluating machine learning approaches for aiding probe selection for gene-expression arrays

MOTIVATION

Microarrays are a fast and cost-effective method of performing thousands of DNA hybridization experiments simultaneously. DNA probes are typically used to measure the expression level of specific genes. Because probes greatly vary in the quality of their hybridizations, choosing good probes is a difficult task. If one could accurately choose probes that are likely to hybridize well, then fewer probes would be needed to represent each gene in a gene-expression microarray, and, hence, more genes could be placed on an array of a given physical size. Our goal is to empirically evaluate how successfully three standard machine-learning algorithms-naïve Bayes, decision trees, and artificial neural networks-can be applied to the task of predicting good probes. Fortunately it is relatively easy to get training examples for such a learning task: place various probes on a gene chip, add a sample where the corresponding genes are highly expressed, and then record how well each probe measures the presence of its corresponding gene. With such training examples, it is possible that an accurate predictor of probe quality can be learned.

RESULTS

Two of the learning algorithms we investigate-naïve Bayes and neural networks-learn to predict probe quality surprisingly well. For example, in the top ten predicted probes for a given gene not used for training, on average about five rank in the top 2.5% of that gene's hundreds of possible probes. Decision-tree induction and the simple approach of using predicted melting temperature to rank probes perform significantly worse than these two algorithms. The features we use to represent probes are very easily computed and the time taken to score each candidate probe after training is minor. Training the naïve Bayes algorithm takes very little time, and while it takes over 10 times as long to train a neural network, that time is still not very substantial (on the order of a few hours on a desktop workstation). We also report the information contained in the features we use to describe the probes. We find the fraction of cytosine in the probe to be the most informative feature. We also find, not surprisingly, that the nucleotides in the middle of the probes sequence are more informative than those at the ends of the sequence.

MAPS: a microarray project system for gene expression experiment information and data validation

SUMMARY

MAPS is a MicroArray Project System for management and interpretation of microarray gene expression experiment information and data. Microarray project information is organized to track experiments and results that are: (1) validated by performing analysis on stored replicate gene expression data; and (2) queried according to the biological classifications of genes deposited on microarray chips.

Application of complementary DNA microarray technology to carcinogen identification, toxicology, and drug safety evaluation

One major challenge facing today's cancer researchers and toxicologists is the development of new approaches for the identification of carcinogens and other environmental hazards. Here, we describe the potential impact of emerging technologies for measuring gene expression profiles on carcinogen identification and on the general field of toxicology. An example of one of these technologies is the use of cDNA microarray chips. We provide an overview to the key questions that are confronting investigators charged with determining the relative safety of natural or synthetic chemicals to which humans are exposed, followed by a discussion of how cDNA microarray technology may be applied to these questions. Gene chip technology is still a relatively new technology, and only a handful of studies have demonstrated its utility. However, as the technical hurdles to development are passed, the use of this methodology in addressing the questions raised here will be critical to increase the sensitivity of detection of the potential toxic effects of environmental chemicals and to understand their risks to humans.

Microarrays and toxicology: the advent of toxicogenomics

The availability of genome-scale DNA sequence information and reagents has radically altered life-science research. This revolution has led to the development of a new scientific subdiscipline derived from a combination of the fields of toxicology and genomics. This subdiscipline, termed toxicogenomics, is concerned with the identification of potential human and environmental toxicants, and their putative mechanisms of action, through the use of genomics resources. One such resource is DNA microarrays or "chips," which allow the monitoring of the expression levels of thousands of genes simultaneously. Here we propose a general method by which gene expression, as measured by cDNA microarrays, can be used as a highly sensitive and informative marker for toxicity. Our purpose is to acquaint the reader with the development and current state of microarray technology and to present our view of the usefulness of microarrays to the field of toxicology.

Structure-activity relationships for triphenylethylene antiestrogens on hepatic phase-I and phase-II enzyme expression

To better understand the mechanism(s) by which tamoxifen induces rat hepatic CYPIIB2 and suppresses GSTA1, structure-activity studies were performed. Compounds employed in these studies included: tamoxifen, fixed-ring tamoxifen, ethylated fixed-ring tamoxifen, pyrrolidino-tamoxifen, 4-iodotamoxifen, idoxifene, and toremifene. With respect to GSTA1 suppression, tamoxifen, fixed-ring tamoxifen, 4-iodotamoxifen, idoxifene, and toremifene were all potent suppressors of GSTA1, while ethylated fixed-ring tamoxifen and pyrrolidino-tamoxifen were completely without activity. The results suggest that the aminoethoxy side chain plays a crucial role in GSTA1 suppression, and that 4-iodination may potentiate this activity. With respect to induction of CYPIIB2, tamoxifen, fixed-ring tamoxifen, and ethylated fixed-ring tamoxifen were inducers of this enzyme, while toremifene and 4-iodotamoxifen were inactive, suggesting that the aminoethoxy side chain is not a structural determinant of CYPIIB2 induction. Because ethylated fixed-ring tamoxifen, toremifene, and 4-iodotamoxifen had differential activities in the two assays, we conclude that CYPIIB2 induction and GSTA1 suppression by triphenylethylenes are the result of two separate and distinct mechanistic pathways. Structure-activity relationships for GSTA1 suppression and CYPIIB2 induction were compared with previously published relationships for triphenylethylene: 1) estrogen receptor relative binding affinity; 2) calmodulin antagonism; 3) antiuterotrophic activity; and 4) antagonism of MCF-7 cell growth. No clear correlation was observed between the effects on CYPIIB2 and these other four activities, suggesting no relationship between the mechanisms responsible for these effects. Similarly, no precise correlation was observed between GSTA1 suppression and these other activities, although rough similarities were observed for relative binding affinity and antiuterotrophic activity. This suggests that the mechanisms responsible for CYPIIB2 induction and GSTA1 suppression are not related to the mechanisms of action for these other documented activities, and may represent different mechanistic pathways.

The effect of short-term fasting, phenobarbital and refeeding on apoptotic loss, cell replication and gene expression in rat liver during the promotion stage

Previous work from this laboratory has reported on the effects of two sequential 5 day periods of fasting and subsequent refeeding on tumor promotion in multistage hepatocarcinogenesis in the rat (Carcinogenesis, 18, 159-166, 1997). In the present extension of the earlier study, the sequential fasting-refeeding regimen was begun at later time points (28 and 54 days post-initiation) than the first study. This was done to determine whether larger-sized altered hepatic foci (AHF) exhibited a depletion similar to that of the relatively small AHF in the published experiment and to study concomitant molecular changes during the fasting periods. Groups of animals were fasted in the presence and absence of 0.05% phenobarbital (PB) in the drinking water. During the fasting periods, both body and liver weights decreased dramatically, less in the fast begun at 54 days. This change was accompanied by a significant decrease in the bromodeoxyuridine (BrdU) labeling indices of hepatocytes within AHF. Apoptotic bodies increased dramatically in the non-focal (surrounding the AHF) hepatocytes during the fasting periods. These parameters were slightly lower in hepatocytes of rats administered PB during the fasting periods, most notably during the 54-66 day period. With the nick end-labeling method, the proportion of hepatocytes undergoing apoptosis was significantly higher in cells within AHF at the end of each of the fasting periods in all but one group. Concomitantly, the number of AHF and percentage of liver volume occupied by AHF decreased dramatically during the fasting periods. Refeeding caused a marked increase in BrdU labeling in hepatocytes within and surrounding AHF during the first week or two, most notably in animals not receiving PB during the fasting period. Both the number and volume percentage of liver AHF returned to control values within approximately 2 weeks of the refeeding regimen. Assays of nuclear DNA fragmentation with samples of whole liver indicated that a 'laddering' effect was most noticeable in livers of animals subjected to the fasting-refeeding regimen when phenobarbital was not present during the fasting period. Studies of the levels of mRNA of several genes in the total liver revealed that the expression of c-myc increased 3- to 9-fold during the fasting periods but rapidly returned to normal levels after refeeding. Levels of albumin and insulin-like growth factor I mRNAs decreased significantly during the fasting period, but rapidly reappeared on refeeding. These results indicate that the extensive loss of AHF during the short-term fasting periods occurs even when the number and volume of AHF are 10- to 50-fold greater at the beginning of the fast than the values published previously. Both the decrease in insulin growth factor I and the elevation of c-myc expression during the fasting period may indicate the role of these genes in the transcriptional regulation of hepatocyte apoptosis in both normal and preneoplastic hepatocytes in the rat.

Effects of triethyl lead administration on the expression of glutathione S-transferase isoenzymes and quinone reductase in rat kidney and liver

The effects of triethyl lead chloride (TEL) on the expression of two detoxication enzyme families, glutathione S-transferases (GSTs) and NAD(P)H:quinone oxidoreductase (QR) were determined in rat liver and kidney. Fischer 344 rats were given one intraperitoneal (i.p.) injection of TEL. GST activity, GST isoenzyme levels, mRNA levels of alpha class GST isoenzymes Ya1, Ya2, and Yc1 and activity of QR were determined. Treatment of rats with TEL caused a significant increase in GST activity in kidney. In kidney, the levels of all GST subunits were significantly elevated; the largest increase was a 3.2-fold increase in GST Yb1. The levels of GST Ya1, Ya2, and Yc1 mRNA also increased after injection of TEL. In liver, TEL injection resulted in decreased GST activity and lower levels of hepatic GSTs Yb2, Yb3, Ya1, and Ya2. The largest decrease was a 40% reduction of GST Ya1. In contrast, the level of liver GST Yc1 increased from day 4 through day 14 after injection of 10 mg/kg TEL and Yp was increased 1.4-fold 4 days after injection of 12 mg/kg TEL. The levels of liver mRNAs coding for alpha class GSTs Ya1, Ya2, and Yc1 were reduced 12 h after injection of TEL. The mRNA levels of GST Ya1 and Ya2 returned to basal level while Yc1 message increased to a level higher than controls 24 h after TEL injection. The increase in Yc1 protein between days 4 and 14 is consistent with the increase in the corresponding mRNA. The activity of QR was elevated 1.5-fold in kidney and 2.7-fold in liver 14 days after the injection of TEL. This report demonstrates that administration of organic lead significantly affects GST expression and QR activity in a tissue-specific and isoenzyme-specific manner. These results indicate that GST expression and QR activity are not co-regulated.

Phase II enzyme expression in rat liver in response to the antiestrogen tamoxifen

The genotoxicity and carcinogenicity of tamoxifen have been attributed to metabolic activation of tamoxifen to an electrophile. Phase II enzymes are known to be involved in the metabolism of the drug and possibly in the formation or elimination of the active metabolite. To determine the effects of tamoxifen on phase II enzyme expression, the drug was administered to F344 rats, and hepatic glutathione S-transferase (GST), UDP-glucuronosyltransferase (UGT), and sulfotransferase (ST) expression was evaluated. Some of the tamoxifen-induced effects, including dramatic suppression of selected GST enzymes and activity, were observed at a dose in rats that is directly equivalent, on a mg/kg b.w. basis, to the doses used for breast cancer treatment. Most of the observed responses are not consistent with the previously described phenobarbital-like properties of tamoxifen and could be the result of the partial agonist activity of tamoxifen at the estrogen receptor. Northern blot analysis was performed with isozyme-specific oligonucleotide probes for rat GST, ST, and UGT. In addition, GST subunit protein levels were assayed by high-performance liquid chromatography. In females, tamoxifen treatment resulted in a 60% suppression of GST Ya1 mRNA and protein levels and a 40% suppression of GST Ya2 levels. In males, tamoxifen treatment suppressed GST Ya1 expression approximately 60%, and GST Ya2 expression was suppressed at low doses but induced above control at high doses. Male GST Yc1 was induced approximately 80% over control. The expression of all other major forms of rat hepatic GST subunit protein, including GST Yb1, Yb2, Yb3, Yp, and Yl, was unaffected by tamoxifen treatment. GST conjugation activity toward delta 5-androstene-3,17-dione, a GST Ya1- and Ya2-specific substrate, was suppressed approximately 40% in both sexes, consistent with our protein and mRNA data. Total GST activity, as measured by the rate of chlorodinitrobenzene conjugation, was not changed. Tamoxifen also produced a dose-dependent increase in UGT2B1 mRNA, a phenobarbital-inducible enzyme; mRNA levels reached 210 and 420% of control in females and males, respectively. In addition, mRNA levels for ST2A2, a female-specific ST gene, were suppressed 50% in females and induced 120% over control in males. mRNA expression for all other forms of rat liver UGT and ST isozymes that were tested was not significantly affected by tamoxifen treatment. Overall, these results demonstrate that tamoxifen has significant effects on hepatic phase II enzyme expression that may have implications for the carcinogenicity and/or therapeutic activity of the drug.

Effects of tamoxifen administration on the expression of xenobiotic metabolizing enzymes in rat liver

The nonsteroidal antiestrogen tamoxifen is widely used in breast cancer treatment and is currently under evaluation as a chemopreventive agent for individuals at high risk of contracting the disease. The effects of tamoxifen administration on the expression of xenobiotic metabolizing enzymes in F344 rat liver have been investigated. Tamoxifen administration for 7 days produced a dose-dependent increase in enzyme expression similar to that reported to be produced by phenobarbital. Increases in CYPIIB1, CYPIIB2, CYPIIIA, and microsomal epoxide hydrolase mRNA and protein levels in males and females were observed by Western and Northern blotting. The expression of CYPIA1, CYPIA2, and gamma-glutamyl transpeptidase mRNA was not significantly affected by tamoxifen treatment. Tamoxifen was approximately one-tenth as potent an inducer of combined CYPIIB1/2 mRNA compared with phenobarbital when the two drugs were administered at equimolar doses. In addition to the effects observed after short-term tamoxifen exposure, increases in CYPIIB1 and CYPIIB2 protein levels were noted after 6 and 15 months of 250 ppm tamoxifen in the diet. Taken together, these results suggest that tamoxifen is a weak phenobarbital-like inducer. However, there are significant differences in the induction profiles produced by the two drugs. Most significant of these differences was the relatively weak induction of CYPIIB1 but striking induction of CYPIIB2 by tamoxifen. In addition, females were often more sensitive than males to tamoxifen, especially at low doses. These differences suggest that tamoxifen and phenobarbital do not use identical molecular mechanisms to produce enzyme induction. It is possible that the effects of tamoxifen are a result of phenobarbital-like properties coupled with the effects of tamoxifen-induced hormonal perturbations in the animal. In sum, tamoxifen induces enzyme expression in rats at a dose comparable, on a mg/kg basis, to the dose women receive for disease management, suggesting these results may be significant for human exposure.

Effect of riboflavin-binding protein deficiency on riboflavin metabolism in the laying hen

Normal chicken eggs contain substantial amounts of riboflavin, all of which is bound to a specific, high-affinity, riboflavin-binding protein (RfBP). Two hens, genetically unable to produce RfBP and thus unable to deposit sufficient riboflavin in their eggs, were compared to two normal hens with respect to the biological half-life of [14C]riboflavin, the tissue distribution of 14C-labeled flavins, and the relative contributions of tissue and dietary riboflavin to flavins deposited in the egg. The biological half-life of [14C]riboflavin was slightly but insignificantly less in the RfBP-deficient hens (11.5 +/- 1.7 days vs 15.1 +/- 3.3 days). The 14C-labeled flavin content of a variety of tissues 3 weeks after the intraperitoneal injection of 5 microCi of riboflavin was also very similar among the four hens. In contrast, the 14C-labeled flavin content of egg yolk, egg albumen, and blood plasma from RfBP-deficient birds was less than 10% of normal. For all hens, the specific radioactivity of flavins in yolk and albumen was similar to that in liver but less than that in heart. We conclude that riboflavin deposited in egg had equilibrated with the large hepatic flavin pool and was not derived preferentially from unlabeled dietary riboflavin. Other than the inability to deposit riboflavin in their eggs, hens of the mutant strain have normal riboflavin metabolism.