Unraveling gene-gene interactions regulated by ligands of the aryl hydrocarbon receptor *.The co-expression of genes coupled to additive probabilistic relationships was used to identify gene sets predictive of the complex biological interactions regulated by ligands of the aryl hydrocarbon receptor The Aryl hydrocarbon receptor (AhR) is member of the family of basic-helix-loop-helix transcription factors. AhR is a cytosolic transcription factor that is normally inactive, bound to several co-chaperones. (Ahr). To maximize the number of possible gene-gene combinations, data sets from murine murine /mu·rine/ (mur´en) pertaining to, derived from, or characteristic of mice or rats. mu·rine adj. embryonic kidney, fetal heart, and vascular smooth muscle Vascular smooth muscle refers to the particular type of smooth muscle found within, and composing the majority of the wall of blood vessels. Vascular smooth muscle contracts or relaxes to both change the volume of blood vessels and the local blood pressure, a mechanism that cells challenged in vitro in vitro /in vi·tro/ (in ve´tro) [L.] within a glass; observable in a test tube; in an artificial environment. in vi·tro adj. In an artificial environment outside a living organism. with ligands of the Ahr were used to create predictor/training data sets. Biologically relevant gene predictor sets were calculated for Ahr, cytochrome cytochrome (sī`təkrōm'), protein containing heme (see coenzyme) that participates in the phase of biochemical respiration called oxidative phosphorylation. P450 1B1, insulin-like growth factor-binding protein-5, lysyl oxidase Lysyl oxidase An enzyme required for the crosslinking of elastin and collagen molecules to form properly functioning connective tissue; present in relatively low levels in at least some forms of cutis laxa. , and osteopontin. Transcript levels were categorized into ternary (programming) ternary - A description of an operator taking three arguments. The only common example is C's ?: operator which is used in the form "CONDITION ? EXP1 : EXP2" and returns EXP1 if CONDITION is true else EXP2. expressions and target genes selected from the data set and tested for all possible combinations using three gene sets as predictors of transitional level. The goodness of prediction for each set was quantified using a multivariate nonlinear coefficient of determination Coefficient of determination A measure of the goodness of fit of the relationship between the dependent and independent variables in a regression analysis; for instance, the percentage of variation in the return of an asset explained by the market portfolio return. Also known as R-square. . Evidence is presented that predictor gene combinations can be effectively used to resolve gene-gene interactions regulated by Ahr ligands, Key words: aryl hydrocarbon receptor, bioinformatics, gene networks, genomics. Environ Health Perspect 112:403-412 (2004). doi:10.1289/txg.6758 available via http://dx.doi.org/[Online 14 January 2004] ********** The assumption that deconstructive methodologies accurately describe the complexity of biological processes is inadequate at best. Until recently, functional genetic studies have been of limited scope and able to elucidate the role of only one or a few genes at a time. Several limitations render these techniques nonconducive to large-scale expression analysis; therefore, nucleotide hybridization hybridization /hy·brid·iza·tion/ (hi?brid-i-za´shun) 1. crossbreeding; the act or process of producing hybrids. 2. molecular hybridization 3. technologies are now used to monitor the expression of thousands of genes at a given point in time. A fundamental goal of genomics research is to understand individual gene expression patterns within the symphonic context of the transcriptome The transcriptome is the set of all messenger RNA (mRNA) molecules, or "transcripts", produced in one or a population of cells. The term can be applied to the total set of transcripts in a given organism, or to the specific subset of transcripts present in a particular cell type. and to unravel the genetic networks responsible for health and disease. However, the interactive gene networks responsible for expression of altered cellular phenotypes have not been fully defined. To date, most microarray experiments have used correlation analysis to identify common genomic responses to a particular stimulus, or multivariate methodologies to examine more complex gene-gene interactions, Univariate methodologies can be used to identify, common genomic responses to a particular stimulus but do not account for multiple influences on gene expression. Logistic regression and stepwise regression, on the other hand, are multivariate approaches successfully used to examine more complex genomic interactions but require prior knowledge of the system and assume linearity in assigning biological relatedness. To understand the complex nature of cellular transformation in cancer, computational prediction methodology has been used to examine global patterns of gene expression (Kim et al. 2000a, 2000b). This method identifies associations between the expression patterns of individual genes by determining whether knowledge of the transcriptional levels of a small gene set predicts the associated transcriptional state of another gene. Although mRNA is not the final product of a gene, transcription is a critical component in the regulatory cascade and therefore provides an ideal point of investigation. A key goal in networks analysis is the development of analytical tools to delineate how individual gene actions are integrated into complex biological systems at the organelle organelle /or·ga·nelle/ (or?gah-nel´) a specialized structure of a cell, such as a mitochondrion, Golgi complex, lysosome, endoplasmic reticulum, ribosome, centriole, chloroplast, cilium, or flagellum. , cell, organ, and organism levels. The goal of this study was to unravel biological networks regulated by ligands of the aryl hydrocarbon receptor (Ahr). Ahr is a ligand-activated transcription factor involved in the regulation of cellular growth, differentiation, and metabolism in all species examined (Carlson and Perdew 2002). Ahr is a member of the large basic helix-loop-helix-PAS (bHLH-PAS) homology domain family of transcription factors that includes proteins involved in myoblast myoblast /myo·blast/ (mi´o-blast) an embryonic cell which becomes a muscle cell or fiber.myoblas´tic my·o·blast n. A primitive muscle cell having the potential to develop into a muscle fiber. differentiation, such as myogenic myogenic /my·o·gen·ic/ (-jen´ik) 1. pertaining to myogenesis. 2. originating in myocytes or muscle tissue. my·o·gen·ic or my·o·ge·net·ic adj. 1. differentiation antigen 1; the cellular response to hypoxia hypoxia Condition in which tissues are starved of oxygen. The extreme is anoxia (absence of oxygen). There are four types: hypoxemic, from low blood oxygen content (e.g., in altitude sickness); anemic, from low blood oxygen-carrying capacity (e.g. , such as Ahr nuclear translocator (Arnt) and hypoxia-inducible factor-1; the Drosophila Drosophila: see fruit fly. drosophila Any member of about 1,000 species in the dipteran genus Drosophila, commonly known as fruit flies but also called vinegar flies. Some species, particularly D. neurogenic neurogenic /neu·ro·gen·ic/ (-jen´ik) 1. forming nervous tissue. 2. originating in the nervous system or from a lesion in the nervous system. protein Sim (single-minded), and the Drosophila circadian rhythm circadian rhythm: see rhythm, biological. circadian rhythm Inherent cycle of approximately 24 hours in length that appears to control or initiate various biological processes, including sleep, wakefulness, and digestive and hormonal activity. protein Per (period). bHLH-PAS proteins generally form heterodimeric transcription factors, of which Ahr is the only member conditionally activated in response to ligand binding. Polycyclic aromatic hydrocarbon polycyclic aromatic hydrocarbon n. Any of a class of carcinogenic organic molecules that consist of three or more rings containing carbon and hydrogen and that are commonly produced by fossil fuel combustion. contaminants and by-products of aspartate aminotransferase aspartate aminotransferase n. Abbr. AST See SGOT. aspartate aminotransferase an enzyme that catalyzes the reversible transfer of an amino group: $$\eqalign $$ metabolism are now recognized as ligands of the Ahr (Bittinger et al. 2003). After ligand binding within the PAS domain, the cytosolic Ahr undergoes a conformational change, dissociates from two 90-kDa heat shock proteins and the hepatitis B Hepatitis B Definition Hepatitis B is a potentially serious form of liver inflammation due to infection by the hepatitis B virus (HBV). It occurs in both rapidly developing (acute) and long-lasting (chronic) forms, and is one of the most common chronic virus X-associated protein 2, and translocates to the nucleus where it dimerizes with Arnt (Carver and Bradfield 1997). The Ahr/Arnt heterodimer interacts with Ahr-responsive elements (5'-TNGCGTG-3') upstream of target genes to activate/ repress re·press v. 1. To hold back by an act of volition. 2. To exclude something from the conscious mind. transcription of target genes. Several drug-metabolizing enzymes (Nebert 1994) are regulated by Ahr, but key molecular targets involved in regulation of cellular differentiation and growth have remained largely elusive. The complexity of Ahr signaling is emphasized by recent studies showing that Ahr also participates in posttranscriptional post·tran·scrip·tion·al adj. Of or relating to a substance or process, such as splicing, that occurs or is formed after transcription of RNA: posttranscriptional modification of RNA. regulation of gene expression Gene modulation redirects here. For information on therapeutic regulation of gene expression, see therapeutic gene modulation.
. (Falahatpisheh and Ramos 2003). The target genes chosen for study included Ahr, cytochrome P450 1B1 (Cyp1b1), insulin-like growth factor-binding protein-5 (Igfbp-5), lysyl oxidase (Lox lox 1 n. pl. lox or lox·es Smoked salmon. [Yiddish laks, from Middle High German lahs, salmon, from Old High German; see laks- ), and osteopontin (Opn). Gene networks were defined on the basis of the co-determination of transcriptional states resolved by statistical evaluation of data sets derived from large-scale simultaneous measurements made using Affymetrix microarray technology. Materials and Methods Model Systems Gene transcription Gene transcription The process by which genetic information is copied from DNA to RNA, resulting in a specific protein formation. Mentioned in: Gene Therapy information from three independent Affymetrix microarray experiments was used for comprehensive computational analysis. The model systems used included mouse embryonic heart, kidney, and thoracic aorta challenged with hydrocarbon ligands to activate Ahr signaling. Vascular smooth muscle cells (vSMC) and fetal kidneys were exposed in vitro, whereas fetal hearts were exposed in vivo in vivo /in vi·vo/ (ve´vo) [L.] within the living body. in vi·vo adj. Within a living organism. in vivo adv. . This approach enabled us to identify common gene sets across tissue types without regard for contextual differences in transcriptional status under basal conditions. The objective was to identify highly conserved interactions regulated by Ahr regardless of genomic context. Data set 1. Vascular smooth muscle cells were isolated from the thoracic aorta of adult C57BL6J untreated mice (Jackson Laboratories, Bar Harbor, ME) (6 weeks of age) and maintained in serial culture as described (Ramos and Cox 1993). All studies were initiated using cells seeded in 150-mm plates at 75% confluence. To induce [G.sub.0] synchronization, cultures were incubated for 72 hr in 0.1% fetal bovine serum Fetal bovine serum ( or foetal bovine serum) is serum taken from the fetuses of cows. Fetal Bovine Serum (or FBS) is the most widely used serum in the culturing of cells. In some papers the expression foetal calf serum is used. in Medium 199 (Invitrogen Corp., Carlsbad, CA). (Jells were challenged for 8 or 24 hr with dimethyl sulfoxide dimethyl sulfoxide (DMSO) Colourless, nearly odourless liquid organic compound. It mixes in all proportions with water, ethanol, and most organic solvents and dissolves a wide variety of compounds (but not aliphatic hydrocarbons). (DMSO DMSO dimethyl sulfoxide. DMSO n. Dimethyl sulfoxide; a colorless hygroscopic liquid obtained from lignin, used as a penetrant to convey medications into the tissues. DMSO, n. ) or 3 [micro]M benzo[a]pyrene (B[a]P) and RNA RNA: see nucleic acid. RNA in full ribonucleic acid One of the two main types of nucleic acid (the other being DNA), which functions in cellular protein synthesis in all living cells and replaces DNA as the carrier of genetic was isolated. B[a]P is a hydrocarbon ligand of the Ahr that modulates growth and differentiation of vascular cells (Kerzee and Ramos 2000). Normalized data from 12 chips were used for the analysis. Data set 2. Day 11.5 mouse embryos were surgically resected from C57bL/6J wild-type or Ahr knockout mice and placed in Hanks' balanced salt solution. Embryonic kidneys (approximately 0.5 mm x 1 mm) were isolated by microsurgical dissection and deposited on culture inserts. B[a]P was added to the medium at 3 [micro]M daily for 4 days, and an equivalent volume of DMSO was added to controls. On day 4, kidneys were harvested for RNA isolation. Normalized data from a total of 8 chips were used for analysis. Data set 3. After exposure to 1.5, 3, and 6 [micro]g/kg 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in utero in utero (in u´ter-o) [L.] within the uterus. in u·ter·o adj. In the uterus. in utero adv. on gestation day 14.5 (n = 4 litters/treatment), gestation day 17.5 fetal mouse hearts were surgically resected and RNA was isolated. TCDD is a potent hydrocarbon ligand that binds and activates Ahr signaling. Normalized data from a total of 20 chips were used from this study. RNA Isolation Total RNA was extracted using TRI TRI Toxics Release Inventory (US EPA) TRI Touch Research Institute TRI Taux de Rentabilité Interne (French: internal rate of return) TRI Taux de Rentabilité Interne TRI Tile Roofing Institute Reagent (Molecular Research Center, Inc., Cincinnati, OH) according to manufacturer's specifications. Affymetrix GeneChip The Affymetrix Murine Genome U74A Array (Affymetrix, Inc., Santa Clara, CA) used in these studies represents all functionally characterized sequences (approximately 6,000) in the Mouse UniGene database (http://www.ncbi.nlm.nih. gov/entrez/query.fcgi?db=unigene). In addition, approximately 6,000 expressed sequence tag An expressed sequence tag or EST is a short sub-sequence of a transcribed spliced nucleotide sequence (either protein-coding or not). They may be used to identify gene transcripts, and are instrumental in gene discovery and gene sequence determination. clusters are included. Experimental procedures including double-stranded cDNA synthesis and biotinylated cRNA preparation were conducted as recommended in the Affymetrix GeneChip Expression Analysis Technical Manual (Affymetrix, Inc. 2003). Double-Stranded cDNA Synthesis Total RNA was processed using Qiagen RNeasy kit (Qiagen, Inc., Valencia, CA) according to manufacturer's specifications. For double-stranded cDNA synthesis, 15-20 mg total RNA was first hybridized with 100 pmol T7-(dT)24 primer [5'(biotin biotin: see vitamin; coenzyme. biotin Organic compound, part of the vitamin B complex, essential for growth and well-being in animals and some microorganisms. )-GTCGTCAAAGATGCTAC-CGTTCAGCA-3'] and high performance liquid chromatography High-performance liquid chromatography (HPLC) is a form of column chromatography used frequently in biochemistry and analytical chemistry. It is also sometimes referred to as high-pressure liquid chromatography. (GENSET Corp., La Jolla, CA) purified at 70[degrees]C for 10 min. Primer hybridization was completed in a 20-mL reaction containing a final concentration of 10 mM dithiothreitol (DTT DTT Deloitte Touche Tohmatsu (Deloitte & Touch Global Operations) DTT Dithiothreitol (cytology reagent) DTT Digital Terrestrial Television DTT Discrete Trial Training ), 500 mM each of deoxyribonucleoside triphosphate triphosphate /tri·phos·phate/ (tri-fos´fat) a salt containing three phosphate radicals. tri·phos·phate n. A salt or ester containing three phosphate groups. (dNTP) mix, and 1 x first-strand cDNA buffer. The reaction was incubated at 42[degrees]C for 2 min followed by addition of 400-600 U SuperScript Any letter, digit or symbol that appears above the line. For example, 10 to the 9th power is written with the 9 in superscript (109). Contrast with subscript. II Reverse Transcriptase Reverse transcriptase Any of the deoxyribonucleic acid (DNA) polymerases present in particles of retroviruses which are able to carry out DNA synthesis using an RNA template. (Gibco Life Technologies, Rockville, MD) to synthesize first-strand cDNA. After 1 hr, second-strand cDNA synthesis was carried out by adding 200 mM each dNTP, 10 U Escherichia coli Escherichia coli (ĕsh'ərĭk`ēə kō`lī), common bacterium that normally inhabits the intestinal tracts of humans and animals, but can cause infection in other parts of the body, especially the urinary tract. DNA ligase DNA ligase /DNA li·gase/ (li´gas) a ligase that catalyzes the linkage between two free ends of double-stranded DNA chains by forming a phosphodiester bond between them, as in the repair of damaged DNA. , 40 U E. coli E. coli: see Escherichia coli. E. coli in full Escherichia coli Species of bacterium that inhabits the stomach and intestines. E. coli can be transmitted by water, milk, food, or flies and other insects. DNA polymerase I DNA polymerase I is an enzyme that mediates the process of DNA replication in prokaryotes. It is 928 residues long, and an example of a processive enzyme - an enzyme which catalyzes a series of polymerisations. , 20 U E. coli RNase H, and 1x second-strand reaction buffer in a 150-mL volume and incubated at 16[degrees]C for 2 hr. T4 DNA polymerase DNA polymerase /DNA po·lym·er·ase/ (pah-lim´er-as) any of various enzymes catalyzing the template-directed incorporation of deoxyribonucleotides into a DNA chain, particularly one using a DNA template. was added at the end of the reaction for an additional 5 min and soaked in 10 mL 0.5 M EDTA EDTA: see chelating agents. . Phase Lock Gel (Eppendorf Scientific, Inc., Westbury, NY) extraction with phenol/ chloroform chloroform (klôr`əfôrm) or trichloromethane (trī'klôrōmĕth`ān), CHCl3 followed by ethanol precipitation was subsequently performed to clean up the double-stranded cDNA. The cDNA pellet was resuspended in 12 mL RNase-free water (Ambion, Inc., Austin, TX). Biotin-Labeled cRNA Preparation Biotin-labeled cRNA target for hybridization to GeneChip Array (Affymetrix, Inc.) was prepared by in vitro transcription using BioArray High Yield RNA Transcript Labeling Kit (Affymetrix, Inc.). Briefly, 3.3-5 mL double-stranded cDNA was mixed gently with 4 mL each of 10x high-yield reaction buffer, 10x biotin-labeled ribonucleotides, 10x DTT, 10x RNase inhibitor mix, and 10x T7 RNA polymerase T7 RNA Polymerase is an RNA polymerase that catalyzes the formation of RNA in the 5'→ 3' direction. T7 RNA polymerase is extremely promoter-specific and only transcribes bacteriophage T7 DNA or DNA cloned downstream of a T7 promoter. provided by the kit and incubated at 37[degrees]C for 4-5 hr, with gentle mixing every 30 min. Labeled cRNA was then cleaned with RNeasy Mini Kit (Qiagen, Inc.) to obtain an accurate quantification of the labeled cRNA. Twenty to 30 mg labeled cRNA was then fragmented to 35-200 bp with 8 mL 5x fragmentation buffer containing 200 mM Tris-acetate, pH 8.1; 500 mM potassium acetate potassium acetate K-Vescent, K-Effervescent Pharmacologic class: Mineral, electrolyte Therapeutic class: Electrolyte replacement, nutritional supplement Pregnancy risk category C ; and 150 mM magnesium acetate in a total volume of 40 mL for 35 min at 94[degrees]C. Before hybridization onto GeneChip Array, the quality of labeling and fragmentation was verified on agarose agarosemore highly purified form of agar with similar uses to agar and widely used in the separation of nucleic acid fragments. gel, transferred onto nylon membrane, and detected with alkaline phosphatase alkaline phosphatase /al·ka·line phos·pha·tase/ (ALP) (fos´fah-tas) an enzyme that catalyzes the cleavage of orthophosphate from orthophosphoric monoesters under alkaline conditions. streptavidin and DuoLuX Chemiluminescent/Fluorescent Substrate using UltraSNAP Biotinylated Nucleic Acid nucleic acid, any of a group of organic substances found in the chromosomes of living cells and viruses that play a central role in the storage and replication of hereditary information and in the expression of this information through protein synthesis. Detection Kit (Vector Laboratories, Inc., Burlingame, CA). Hybridization to GeneChip Array After labeling and fragmentation, 15 [micro]g fragmented biotinylated cRNA was hybridized to the Affymetrix GeneChip Array in a 300-mL cocktail containing 5 mL of 3 nM control oligonucleotide B2, 15 mL 20x eukaryotic eukaryotic /eu·kary·ot·ic/ (u?kar-e-ot´ik) pertaining to a eukaryon or to a eukaryote. eukaryotic pertaining to eukaryosis. eukaryotic cells see cell. hybridization controls, and 150 mL 2x hybridization buffer provided in the GeneChip Eukaryotic Hybridization Control Kit (Affymetrix, Inc.) together with 3 mL 10 mg/mL herring sperm DNA DNA: see nucleic acid. DNA or deoxyribonucleic acid One of two types of nucleic acid (the other is RNA); a complex organic compound found in all living cells and many viruses. It is the chemical substance of genes. and 3 mL 50 mg/mL acetylated bovine serum albumin serum albumin n. See seralbumin. (BSA 1. BSA - Business Software Alliance. 2. BSA - Bidouilleurs Sans Argent. ). The hybridization was carried out at 45[degrees]C and 60 rpm in GeneChip Hybridization Oven 640 (Affymetrix, Inc.) for 16 hr. Washing, Staining, and Scanning the Array After hybridization the washing and staining procedures were carried out on a GeneChip Fluidics fluidics, branch of engineering and technology concerned with the development of equivalents of various electronic circuits using movements of fluid rather than movements of electric charge. Station 400 in conjunction with Affymetrix Microarray Suite 5.0 software (MAS 5.0; Affymetrix, Inc.). Briefly, the array was first washed with 10 cycles of 2 filling and draining cycles in a nonstringent buffer containing 6x SSPE SSPE abbr. subacute sclerosing panencephalitis SSPE subacute sclerosing panencephalitis. SSPE Subacute sclerosing panencephalitis, see there (52.9 g sodium chloride sodium chloride, NaCl, common salt. Properties Sodium chloride is readily soluble in water and insoluble or only slightly soluble in most other liquids. It forms small, transparent, colorless to white cubic crystals. ; 8.28 g sodium phosphate; monobasic monobasic /mono·ba·sic/ (-ba´sik) having but one atom of replaceable hydrogen. mon·o·ba·sic adj. 1. Having only one hydrogen ion to donate to a base in an acid-base reaction. , 2.82 g EDTA, pH 7.9) and 0.01% Tween tween n. A child between middle childhood and adolesence, usually between 8 and 12 years old. [Blend of teen1 and between.] 20 at 25[degrees]C, followed by 4 cycles of 15 filling and draining cycles in a stringent buffer containing 100 mM MES (Manufacturing Execution Software) Software that provides real time access to plant activities that include equipment, labor, orders and inventory. An MES integrates the data with enterprise resource planning (ERP) systems so that management has complete control of [2-(N-morpholine)ethanesulfonic acid], 0.1 M [Na.sup.+] and 0.01% Tween 20 at 50[degrees]C. The probe array then was first stained with a 600-mL streptavidin-phycoerythrin (SAPE SAPE Sapient Corp (stock symbol) SAPE Substance Abuse Prevention Education SAPE Survivable Adaptive Planning Experiment SAPE Sexual Assault Prevention and Education ) solution containing 1x MES stain buffer (100 mM MES, 1 M [Na.sup.+], and 0.05% Tween 20), 2 mg/mL acetylated BSA, and 10 mg/mL SAPE (Molecular Probes. Inc., Eugene, OR) for 10 min at 25[degrees]C. After the first staining the array was washed with 10 cycles of 4 filling and draining cycles in nonstringent buffer at 25[degrees]C. The stained signals were then amplified in a 600-mL antibody solution containing 1x MES stain buffer, 2 mg/mL acetylated BSA, 0.1 mg/mL normal goat IgG, and 3 mg/mL antistreptavidin biotinylated antibody (Vector Laboratories, Inc.) for 10 min at 25[degrees]C, followed by a second SAPE staining at the same temperature for another 10 min. Finally, the probe array was washed for 15 cycles of 4 filling and draining cycles in nonstringent buffer and scanned by the Agilent GeneArray Scanner (Affymetrix, Inc.) at an excitation wavelength of 570 nm. Data Analysis After scanning, each image was inspected for major chip defects or abnormalities in hybridization signal as a quality control and analyzed using MAS 5.0. The data were then normalized using a scaling factor of 500 according to the Affymetrix GeneChip Expression Analysis Technical Manual (Affymetrix, Inc. 2003). The expression of approximately 12,000 genes across 40 different Affymetrix microarrays was quantified and normalized. Each of the targets was correlated to the complete Affymetrix gene data set (MATLAB (MATrix LABoratory) A programming language for technical computing from The MathWorks, Natick, MA (www.mathworks.com). Used for a wide variety of scientific and engineering calculations, especially for automatic control and signal processing, MATLAB runs on Windows, Mac and 6.0; The MathWorks, Inc., Natick, MA). Computational Methodology After chip normalization In relational database management, a process that breaks down data into record groups for efficient processing. There are six stages. By the third stage (third normal form), data are identified only by the key field in their record. , the mean ([bar.x]), standard deviation In statistics, the average amount a number varies from the average number in a series of numbers. (statistics) standard deviation - (SD) A measure of the range of values in a set of numbers. (sd) and coefficient of variation Coefficient of Variation A measure of investment risk that defines risk as the standard deviation per unit of expected return. [1] cv = sd/[bar.x] were calculated for each gene across all samples, and the intensity level (x) for each gene was standardized across all arrays by [2] x = x-[bar.x]/sd. All genes were sorted by cv, and, because of computational constraints, the 200 clones with the greatest cv were selected for use as predictors. The genes to be predicted (i.e., targets) were selected on the basis of biological relevance and included Ahr, Cyp1b1, Igfbp-5, Lox, and Opn. A heuristic method heuristic method Decision making A form of problem-solving based, not on scientific proof but rather on plausible, possible, or creative conclusions to questions that cannot be answered in the context of, or the 'logic' of which lies outside of, a currently was used to discretize the data into ternary states that describe their behavior, with -1 used for downregulated genes, 0 for invariant (programming) invariant - A rule, such as the ordering of an ordered list or heap, that applies throughout the life of a data structure or procedure. Each change to the data structure must maintain the correctness of the invariant. genes, and +1 for upregulated genes. Each array was ordered from lowest to highest standardized values x without regard to gene identity and the means across all arrays at the 33rd and 66th percentile were used as the cutoffs for the different values [bar.x] of the standardized data set. These percentiles were chosen because they divided the data evenly to provide sufficient variation throughout the computational analysis. Our goal was to yield good estimates of the transcriptional state of the target using ternary, discrete functions. The computation of prediction strengths was completed using a stochastic model of gene intensities, defining [X.sub.i] as the random variable that represents the ternary intensity (or transcription rate) of gene [G.sub.i]. Realizations of [X.sub.i] were denoted by lower case [x.sub.i]. A microarray experiment consisted of a realization [x.sub.1] to [x.sub.v] of v random variables [X.sub.1] to [X.sub.v] representing the intensities of the v genes [G.sub.1] to [G.sub.v]. Given N microarrays, the values [x.sub.1](j) to [x.sub.v](j) represented the values for the microarray j, with j = 1 to n. Without loss of generality Without loss of generality (abbreviated to WLOG or WOLOG and less commonly stated as without any loss of generality) is a frequently used expression in mathematics. the predictive strength of genes [G.sub.1], [G.sub.2], and [G.sub.3] over gene [G.sub.4] was unknown and therefore needed to be computed. In this context, genes [G.sub.1], [G.sub.2], and [G.sup.3] were regarded as predictors, and [G.sub.4] as target. An example of ternary intensity values [x.sub.1](j), [x.sub.2](j), [x.sub.3](j), and [x.sub.4](j), j = 1 to 9 (or realizations of [X.sub.1], [X.sub.2], [X.sub.3], and [X.sub.4]), for [G.sub.1], [G.sub.2], [G.sub.3], and [G.sub.4], across nine conditions (microarrays) is illustrated in Figure 1 (rows 1-4). The function with the smaller expected error was denoted by [[psi].sub.opt]. The optimal constant function was denoted by [[psi].sub.0], and was the one with minimal error over the three possible constant functions. The optimal functions [[psi].sub.opt] and [[psi].sub.0], and their errors [epsilon][[[psi].sub.opt]] and [epsilon][[[psi].sub.0]], respectively, were estimated from the microarray data by splitting them randomly in two sets of conditions, with n conditions for test (estimation of the error) and N-n conditions for train (estimation of the function), where N was the total number of conditions. The optimal function [[psi].sub.opt] and the optimal constant function [[psi].sub.0] were estimated from the N-n examples using plug-in rule (Devroye et al. 1996). Let [[psi].sub.opt] and [[psi].sub.0] be the estimates of [[psi].sub.opt] and [[psi].sub.0], respectively; the errors [epsilon][[[psi].sub.opt]] and [epsilon][[[psi].sub.0]] were estimated using the following equations: [3] [epsilon]([[psi].sub.opt]) = 1/n [n.summation over j=1]|[[psi].sub.opt]([x.sub.1](j), [x.sub.2](j), [x.sub.3](j)) - [x.sub.4](j)| and [4] [epsilon]([[psi].sub.0]) = 1/n [n.summation over j=1]|[[psi].sub.0]([x.sub.1](j), [x.sub.2](j), [x.sub.3](j)) - [x.sub.4](j)|. In our study with N = 40, n was defined as n = N/3, so 2n was used to train the functions and n to estimate the error. The performance of the predictor set was defined by the coefficient of determination (CoD), that is, the degree to which the transcriptional state of a given set of genes improves the prediction of the transcriptional state of a target gene relative to the best possible constant function (defined by a fixed value) (Devroye et al. 1996) and was defined by [5] [theta Theta A measure of the rate of decline in the value of an option due to the passage of time. Theta can also be referred to as the time decay on the value of an option. If everything is held constant, then the option will lose value as time moves closer to the maturity of the option. ] = [epsilon]([[psi].sub.0]) - [epsilon]([[psi].sub.opt]/[epsilon]([[psi].sub.0]). The CoD was estimated from the data using the following equation: [6] [theta] = [epsilon]([[psi].sub.0]) - [epsilon]([[psi].sub.opt]/[epsilon]([[psi].sub.0]). This process defines an estimate [theta] of the CoD for the set of predictor genes [G.sub.1], [G.sub.2], [G.sub.3] and the target gene [G.sub.4]. The process was repeated 1,000 times with different random splitting and the final estimates [bar.[theta]] as the average value of [theta] over the 1,000 replications. The mean error [bar.[epsilon]]([[psi].sub.opt]) was also estimated as the average of the 1,000 estimates [epsilon]([[psi].sub.opt]). The higher the CoD, the more accurate the prediction of the target gene transcriptional state derived from the transcriptional state of the three predictor genes. All possible combinations of three predictor genes were studied for each target, where the number of combinations was on the order of 8 million sets of 3 selected from 200, and the estimated CoD [bar.[theta]] was computed for all combinations. The predictor sets (or combinations) were ordered from best to worst based on [bar.[theta]], and the analysis focused on the predictor combinations with [bar.[theta]] > 0.9 and [bar.[epsilon]]([[psi].sub.opt]) < 0.05. The cutoff used was selected heuristically to restrict the number of sets to those with good prediction potential. Because of the nature of the analysis, there was no measure of false-positive rates other than future biological experimentation and scrutiny of the literature. Each gene [G.sub.i] was ranked by [f.sub.s], the percentage of sets in this family of good combinations that contain such a gene. Network Plot The possible relationships predicted by the predictor algorithm were illustrated using the program developed by Breitkreutz et al. (2003). All three gene combinations that met the cutoff criteria on the basis of test errors and CoD values [[bar.[theta]] > 0.9 and [bar.[epsilon]]([[psi].sub.opt]) < 0.05] were linked to their respective targets. These schematics were overlaid to form a linkage diagram showing the relationship of target genes to predictors, as well as connections among target genes. The diagrams generated represent the total hypothetical interrelationships between all 200 predictor genes and the five target genes used in the analysis. Results The expression of 200 genes across 40 different Affymetrix arrays from three different tissues after activation of Ahr signaling was used to predict the behavior of selected targets. Ahr activation by a wide range of structurally divergent agents plays a major role in vascular, renal, cardiac, skin, bone marrow, liver, eye, ovary ovary, ductless gland of the female in which the ova (female reproductive cells) are produced. In vertebrate animals the ovary also secretes the sex hormones estrogen and progesterone, which control the development of the sexual organs and the secondary sexual , and immune system immune system Cells, cell products, organs, and structures of the body involved in the detection and destruction of foreign invaders, such as bacteria, viruses, and cancer cells. Immunity is based on the system's ability to launch a defense against such invaders. function, as well as in carcinogenesis car·ci·no·gen·e·sis n. The production of cancer. carcinogenesis production of cancer. biological carcinogenesis viruses and some parasites are capable of initiating neoplasia. (Denison and Nagy 2003). Table 1 ranks predictor genes by the [f.sub.i] commonly identified for each target. The Ahr receptor itself was most commonly predicted by lymphocyte antigen 6 complex, locus e (selected 18% of the time for all good three-gene combinations), Igfbp-3 (16%), and tumor necrosis factor receptor A tumor necrosis factor receptor (TNFR) is, as its name would indicate, a receptor which binds tumor necrosis factors (TNF). Because "TNF" is often used to describe TNF alpha, "TNFR" is often used to describe the receptors that bind to TNF alpha - namely, CD120. super family-member 1b (14%). Cyp1b1 was best predicted by spleen tyrosine kinase tyrosine kinase An enzyme intimately linked to signal transduction–ST, either as a receptor-type TK, which participates in transmembrane signaling, or as an intracellular TK, participating in ST to the nucleus; ↑ or ↓ TK activity is associated with (31%), squalene squalene (skwäˑ·lēn), n a popular traditional Asian remedy derived from the liver oil of sharks. epoxidase (16%), and nicotinamide adenine dinucleotide nicotinamide adenine dinucleotide and nicotinamide adenine dinucleotide phosphate: see coenzyme. Nicotinamide adenine dinucleotide (NAD) (NADH NADH the reduced form of NAD. NADH n. The reduced form of NAD. NADH, n.pr a coenzyme that incorporates niacin and involved in the Krebs cycle. ) dehydrogenase dehydrogenase /de·hy·dro·gen·ase/ (de-hi´dro-jen-as?) an enzyme that catalyzes the transfer of hydrogen or electrons from a donor, oxidizing it, to an acceptor, reducing it. de·hy·dro·gen·ase n. (Ndufc1) (14%). Igfbp-5 was most frequently predicted by Opn (14%), matrix metalloproteinase 3 (9%), and RNA binding motif-single-stranded interacting protein 1 (8%). Lox was best predicted by lymphocyte antigen 6 complex, locus H (Ly6h) (36%), single-stranded interacting protein 1 (36%), and glutamyl aminopeptidase (23%). Finally, Opn was most often predicted by brain-derived neurotrophic factor Brain-derived neurotrophic factor (BDNF) is a neurotrophic factor found in the brain and the periphery. It is a protein that acts on certain neurons of the central nervous system and the peripheral nervous system that helps to support the survival of existing neurons and encourage (Bdnf) (15%), interleukin 6 (14%), and proliferin (14%). The frequency of predictor usage was influenced by the cutoffs used, thus emphasizing the importance of arbitrary cutoff selection. For comparison, each target was also correlated to the complete 12,000 Affymetrix gene data set using normalized intensity values. The results indicated that in many instances genes frequently identified as predictors in the CoD algorithm also exhibited strong correlation to their targets. This is best exemplified for Ahr and Opn, where Ndufc1 and [alpha]-actin (Actc1) exhibited strong correlations and were good predictors. As demonstrated by CoD, however, genes exhibiting low correlation to specific targets may play an important rote in predicting target behavior. This is exemplified by the best predictor for Lox, with a frequency of 36% but a correlation coefficient Correlation Coefficient A measure that determines the degree to which two variable's movements are associated. The correlation coefficient is calculated as: of only -0.37. Thus, the CoD methodology identified interactions between targets and predictors that could potentially be missed when examined on a gene-by-gene basis. The complex nature of biological interactions between targets and predictors is illustrated in Figure 2. The linking diagram was built using the franks for each predictor gene and each target gene. The inclusion of any singular gene does not represent a causal relationship between predictor and target, but rather that gene [G.sub.i] with two other genes was also a good predictor of a specific target. To some degree, all targets were predictive of each other, a relationship not surprising given that targets were selected a priori a priori In epistemology, knowledge that is independent of all particular experiences, as opposed to a posteriori (or empirical) knowledge, which derives from experience. on the basis of their responsiveness to Ahr ligands. Several genes predominated as predictors of each target. These relationships were denoted by the thickness of the connective lines (Figure 2) and the percentage of time used in three gene combinations (Table 1). The number of unique predictors varied with target and cutoff range, with 85% of all possible predictors used to estimate Ahr, 63% Cyp1b1, 98% Igfbp-5, 22% Lox, and 53% Opn. Given our cutoff range, a total of 34 genes were resolved to provide a comprehensive association of all targets and predictors (Figure 3). The predominant linkages were between all five targets, particularly Opn as a predictor of Ahr and Cyp1b1. The best nontarget non·tar·get adj. Not being the target, as of an agent or weapon: effects of radiotherapy on nontarget cells. predictors included Bdnf, Actc1, c-myc single-strand-binding protein, Igfbp-6, and integrin integrin /in·te·grin/ (in´te-grin) any of a family of heterodimeric cell adhesion receptors, each consisting of an a and a ß polypetide chain, that mediate cell-to-cell and cell-to–extracellular matrix interactions. beta4. Figure 4A-E A-E, AE above-elbow; see under amputation. shows the use of individual gene predictor values for construction of three gene predictor set models. It should be noted that predictor genes may lie upstream or downstream from the target or be part of a chain of interaction among various intermediates within the biological network. A large number of three-clone combinations met the selection criteria [([bar.[theta]] > 0.9 and [bar.[epsilon]]([[psi].sub.opt]) < 0.05], with one or two clones identified as predominant predictors within the sample pool. A complete listing of target-predictor clone combinations is presented in Table 1 of the Supplemental Material (http://www. ehponline.org). [FIGURES 2-4 OMITTED] When examined as three-gene predictor sets, the best combinations for Ahr were Igfbp-3, lymphocyte antigen 6, and Actc1 with CoD of 0.963; Ras-related, proliferin, and Igfbp-6 with CoD of 0.96; and thrombomodulin, lymphocyte antigen 6, locus, and Igfbp-6 with CoD of 0.953 (Figure 4A). The best combinations for Cyp1b1 were Ahr, melanoma growth-simulating activity-alpha (Gro1) oncogene oncogene Gene that can cause cancer. It is a sequence of DNA that has been altered or mutated from its original form, the proto-oncogene (see mutation). Proto-oncogenes promote the specialization and division of normal cells. , and Cd44 antigen with CoD of 0.944; Actc1, Bdnf, and epiregulin with CoD of 0.939; and Ahr, Cd44 antigen, and proliferin 3 with CoD of 0.944 (Figure 4B). The best combinations for Igfbp-5 were signal transducer and activator of transcription 1 (Stat1), Opn, and procollagen-lysine, 2-oxoglutarate 5-dioxygenase 3 (Plod3) with CoD of 0.98; Opn, Igfbp-1, and interleukin 1 alpha with CoD of 0.97; and insulin II, Opn, and Igfbp-1 with CoD of 0.97. These combinations performed best when used collectively (Figure 4C). When examined as three gene predictor sets, the best predictor combinations for Lox were lymphocyte antigen 6, locus H, Ndufc1, and perpherin with CoD of 0.902; perpherin, lymphocyte antigen 6, locus H, and proliferin 2 with CoD of 0.923; and cadherin 16, single-stranded interacting protein 1, and Cd44 antigen with CoD of 0.924 (Figure 4D). For Opn the best three-gene predictor combinations were Ahr, Gro1 oncogene, and Cd44 antigen with CoD of 0.944; Actc1, Bdnf and epiregulin with CoD of 0.939; and Ahr, Cd44 antigen, and proliferin 3 with CoD of 0.944 (Figure 4E). [FIGURE 4 OMITTED] Discussion The identification of relevant components of the biological response to Ahr ligands was modeled using the transcriptional profiles of cells from murine embryonic heart and kidney or vasculature vasculature /vas·cu·la·ture/ (vas´ku-lah-chur) 1. circulatory system. 2. any part of the circulatory system. vas·cu·la·ture n. . Genes whose transcription levels exhibited coupling by CoD methodology were hypothesized to be predictive of one another, whether lying upstream or downstream within the biological network or distributed about the network in such a way that their relation to the target gene was only based on chains of interactions among various intermediate genes. Linear correlation was found to identify predominant genomic responses to a particular stimulus but was unable to identify sets of genes whose actions and interactions drive the transcriptional level of a target. This is problematic, as a gene in and of itself may not be highly correlated to a target but in combination with other genes may be predictive of the behavior of the target. For example, Ly6h was identified as the most frequently used Lox predictor but showed only modest correlation with Lox as a target. Similarly, the small proline-rich proteins 2b and 2f were only poorly correlated to Ahr but rank among the top predictors of Ahr behavior. When used individually, each predictor gene exhibited a CoD measure comparable to the value obtained using a linear correlation coefficient model. In contrast the CoD method detected multivariate nonlinear influences on gene expression in complex genetic networks and enabled the calculation of a CoD value that mathematically reflected interactions among all predictors. Use of different treatment modalities was advantageous and allowed identification of genes that move across the transcriptome in unison after activation of Ahr signaling regardless of contextual differences in gene expression or tissue-specific patterns of differentiation. For example, the constitutive and inducible expression of Cyp1b1, a gene involved in steroidogenesis steroidogenesis /ste·roi·do·gen·e·sis/ (ste-roi?do-jen´e-sis) production of steroids, as by the adrenal glands.steroidogen´ic ste·roid·o·gen·e·sis n. The biological synthesis of steroids. , is highly tissue specific. Cyp1b1 is expressed constitutively in the adrenal gland adrenal gland (ədrēn`əl) or suprarenal gland (s  prərēn`əl), endocrine gland (see endocrine system) about 2 in. (5. , ovary, and testes testesor testicles Male reproductive organs (see reproductive system). Humans have two oval-shaped testes 1.5–2 in. (4–5 cm) long that produce sperm and androgens (mainly testosterone), contained in a sac (scrotum) behind the penis. and is highly inducible by adrenocorticotropin adrenocorticotropin /adre·no·cor·ti·co·trop·in/ (-kor?ti-ko-tro´pin) corticotropin. ad·re·no·cor·ti·co·trop·in or ad·re·no·cor·ti·co·troph·in n. See ACTH. , cAMP, peptide hormones peptide hormones (pepˑ·tīd hōrˑ·mōnz), n.pl , and Ahr ligands (Buters et al. 2002). Cyp1b1 is a complex gene, as Ahr-independent induction is observed in certain tissues (Kerzee and Ramos 2001), and the gene is overexpressed as a function of transformation (Vondracek et al. 2002). Similarly, Opn is differentially expressed under constitutive and inducible conditions and interacts with surface receptors in a context-specific manner. In the normal human kidney, Cd44 is the primary receptor partner of Opn, whereas integrin binding and activation are predominant under stressful conditions (Xie et al. 2001). This dichotomy is also observed within the context of injury, as Opn promotes accumulation of macrophages Macrophages White blood cells whose job is to destroy invading microorganisms. Listeria monocytogenes avoids being killed and can multiply within the macrophage. and participates in macrophage-mediated renal injury but exerts renoprotective actions during acute ischemia. Such differences give rise to unique fingerprints that influence gene-gene interactions. Thus, combined data sets within different contextual networks increased the power of the prediction and allowed better integration of biological complexity. As a biological target, Ahr was best predicted by lymphocyte antigen be (Ly6e), a retinoic acid--responsive gene involved in T-cell activation (Mao et al. 1996; Schedlich and Graham 2002). Ahr has been implicated im·pli·cate tr.v. im·pli·cat·ed, im·pli·cat·ing, im·pli·cates 1. To involve or connect intimately or incriminatingly: evidence that implicates others in the plot. 2. in the immune suppressant activity of aromatic hydrocarbons (Kerkvliet et al. 2002) and in the regulation of retinoic acid retinoic acid /ret·i·no·ic ac·id/ (ret?i-no´ik) an oxidized derivative of retinol, believed to be the form of vitamin A that plays a role in the development and growth of bone and in the maintenance of normal epithelial structures. metabolism. Two other predictor genes for Ahr, Igfbp-3, and Igfbp-6, have also been identified as retinoic acid-regulated genes (Dailly et al. 2001; Schedlich and Graham 2002), suggesting that Ahr may couple to predictor genes through retinoic acid linkage (Andreola et al. 1997). Recent studies demonstrate that retinoids Retinoids A derivative of synthetic Vitamin A. Mentioned in: Ichthyosis retinoids (reˑ·t activate Ahr signaling (Soprano and Soprano 2003), whereas Ahr controls the expression of genes involved in retinoic acid metabolism (Gonzalez and Fernandez-Salguero 1998). Thus, computational strategies delineated connections between Ahr and retinoic acid that otherwise could not have been predicted in the absence of biological information. Cyp1b1 was predicted by syk tyrosine kinase (Syk) in all cases, a relationship consistent with work by Mounho and Burchiel (1998) showing that oxidative metabolites Metabolites Substances produced by metabolism or by a metabolic process. Mentioned in: Interactions of B[a]P increase tyrosine phosphorylation phosphorylation, chemical process in which a phosphate group is added to an organic molecule. In living cells phosphorylation is associated with respiration, which takes place in the cell's mitochondria, and photosynthesis, which takes place in the chloroplasts. of Syk. Syk participates in integrin signaling, leading to nuclear factor-kappa B activation and increased levels of cytokine Cytokine Any of a group of soluble proteins that are released by a cell to send messages which are delivered to the same cell (autocrine), an adjacent cell (paracrine), or a distant cell (endocrine). mRNAs (Lin et al. 1995) and regulation of cancer cell growth and metastasis metastasis /me·tas·ta·sis/ (me-tas´tah-sis) pl. metas´tases 1. transfer of disease from one organ or part of the body to another not directly connected with it, due either to transfer of pathogenic microorganisms or to (Coopman et al. 2000). Similarly, Ndufc1 appeared in each predictor set, with recent microarray experiments showing common responses in cancer cells for Cyp1b1 and Ndufc1 (Huang et al. 2003). Cyp1b1 catalyzes activation of molecular oxygen in an NADH-dependent electron transport electron transport n. The successive passage of electrons from one cytochrome or flavoprotein to another by a series of oxidation-reduction reactions during the aerobic production of ATP, with the electrons originating from an oxidizable substrate and pathway, so a connection with Ndufc1 and the mitochondrial mitochondrial pertaining to mitochondria. mitochondrial RNAs a unique set of tRNAs, mRNAs, rRNAs, transcribed from mitochondrial DNA by a mitochondrial-specific RNA polymerase, that account for about 4% of the total cell RNA that electron transport chain An electron transport chain associates electron carriers (such as NAD+ and FADH2) and mediating biochemical reactions that produce adenosine triphosphate (ATP), which is the energy currency of life. is consistent with the known biology. Individuals with clefts of the lip and/or palate often share genetic variations in both Ndufc1 and cytochrome P-450, thus raising intriguing links between the two enzyme systems (Johnston and Bronsky 1995). In several instances, the algorithm selected a predominant gene, such as Opn, when Igfbp-5 was chosen as a target. Opn is a secreted acidic phosphoprotein phosphoprotein /phos·pho·pro·tein/ (-pro´ten) a conjugated protein in which phosphoric acid is esterified with a hydroxy amino acid. phos·pho·pro·tein n. containing a conserved GRGDS sequence that regulates a variety of cellular processes, primarily through the [alpha]v[beta]3 integrin. Opn is a component of the human atherosclerotic plaque Atherosclerotic plaque A deposit of fat and other substances that accumulate in the lining of the artery wall. Mentioned in: Atherectomy atherosclerotic plaque that promotes adhesion and spreading of vascular cells and is a potent chemotactic factor chemotactic factor soluble molecules which attract and guide the movement of cells such as phagocytes in the inflammatory response. chemotactic factor for vSMC (Wilson et al. 2002). The strong relationship between Igfbp-5 and Opn is in keeping with studies showing that Opn binds to Igfbp-5 with high affinity (Nam et al. 2000) and that this interaction concentrates Igfbp-5 on the matrix and modulates cooperative interactions between Igf-I receptor and integrin [[alpha].sub.v][[beta].sub.3] in atherosclerotic lesions (Nichols et al. 1999). Igfbp-5 is the most conserved member of the IGFBP IGFBP Insulin-Like Growth Factor Binding Protein family and a regulator of bone, kidney, and mammary gland mammary gland, organ of the female mammal that produces and secretes milk for the nourishment of the young. A mammal may have from 1 to 11 pairs of mammary glands, depending on the species. Generally, those mammals that bear larger litters have more glands. function. Igfbp-5 plays a decisive role in tumor cell proliferation (Schedlich and Graham 2002) and together with Opn promotes IGF-I--inducible biological effects. Other genes selected as predictors have been linked to Igfbp-5, such as signal transducer and activator of transcription (Chapman et al. 1999) and insulin (Duan et al. 1996). For other targets, such as Lox, the connections between predictor and target gene were less clear, Lox is an extracellular and intracellular copper-containing enzyme that initiates cross-linking of collagen and elastin elastin /elas·tin/ (e-las´tin) a yellow scleroprotein, the essential constituent of elastic connective tissue; it is brittle when dry, but when moist is flexible and elastic. e·las·tin n. by catalyzing oxidative deamination of the epsilon-amino group in certain lysine lysine (lī`sēn), organic compound, one of the 20 amino acids commonly found in animal proteins. Only the l-stereoisomer appears in mammalian protein. and hydroxylysine residues of collagens and lysine residues of elastin. Lox is found in smooth muscle cell nuclei and is speculated to be involved in oxidative deamination of peptidyl lysine (Kuivaniemi et al. 1984; Li et al. 1997). As such, the finding that matrix-related genes such as peripherin and proliferin2 were selected as predictors is of interest. Integrins integrins (inˑ·t  ·grinz),n.pl. may link Lox and peripherin, with both collagen and peripherin modulated by the [alpha]2[beta]1 integrin (Khalsa et al. 2000). The connection with immune-related genes such as Cd44 and Ly6h is intriguing and may involve interferon. Interferon-gamma, a proinflammatory cytokine, down regulates Lox gene expression in rat aortic aortic pertaining to or emanating from the aorta. See also aortic arch. aortic aneurysm occurs most often in dogs, where it is caused by Spirocerca lupi larvae, turkeys and primates, causing dyspnea, cyanosis and coughing. smooth muscle cells (Song et al. 2000), and Ly6h has also been identified as a target of interferon (Horie et al. 1998). Lox and Ndufc1 are coregulated under sheer stress (Ando et al. 1996), so a link between these two genes may also exist. The relationship between Opn and Ahr is intriguing. Recent studies have shown that resveratrol res·ver·a·trol n. A natural compound found in grapes, mulberries, peanuts, and other plants or food products, especially red wine, that may protect against cancer and cardiovascular disease by acting as an antioxidant, antimutagen, and (3, 5, 4'-trihydroxystilbene), an Ahr antagonist, modifies the inhibitory effect of TCDD on Opn expression (Singh et al. 2000). Opn is a ligand for the Cd44 receptor and stimulates Cd44 expression on the osteoclast osteoclast /os·teo·clast/ (os´te-o-klast?) 1. a large multinuclear cell associated with absorption and removal of bone. 2. an instrument used for osteoclasis. surface (Chellaiah et al. 2003; Sodek et al. 2002). The relationship between Cd44 and Opn is well established (Denhardt et al. 2001). The analysis also predicted a relationship between Opn and Actc1, a connection consistent with reports by Shu and co-workers (2002) showing that Opn-expressing tubuli in kidney were tightly associated with a peritubular influx of alpha-smooth muscle actin. It is also not surprising that Opn expression was predicted by two growth factors such as epiregulin and Bdnf(Lee et al. 2001; Toyoda et al. 1995). The observed associations indicate that measurement of the transcriptional state of multiple gene sets can be combined to better predict the expression level of a target gene The strongest interaction between targets was the interplay between Ahr, Cyp1b1, and Opn (Figures 2 and 3). Although these biological connections are novel, they are in keeping with studies showing that Ahr signaling is critical to Opn regulation (Singh et al. 2000) and that ligands of the Ahr modulate Opn expression. For some of the target/predictor relationships, no direct linkages could be established. An example is Bdnf, a protein that allows the survival of specific neuronal populations and found to be highly predictive of Opn. As both Bdnf and Opn are involved in matrix regulation, the relationship could be an interesting area for future investigation. [FIGURES 2-3 OMITTED] Conclusions Novel putative interactions were defined to investigate gene--gene interactions regulated by the Ahr. In many instances, the relevance of the biological relationships uncovered using CoD methodology was ratified by published reports, whereas in others, novel interactions were identified. The computational approach used afforded us the ability to begin constructing gene networks that define broad-ranging interactive biological relationships. Although the biological bases for these theoretical relationships must be investigated further, the number of possible combinations is now reduced to a manageable size that can be systematically scrutinized using established molecular methodologies.
Table 1. Genes that predict selected target gene behavior. (a)
Targets Predictor Probe set ID (a)
Ahr Lymphocyte antigen 6 complex, locus e 101488_r_at
160495_at Insulin-like growth factor binding
protein 3 95082_at
Tumor necrosis factor receptor
superfamily, member 1b 94928_at
Insulin-like growth factor binding
protein 6 103904_at
Brain derived neurotrophic factor 102727_at
Secreted phosphoprotein 1 97519_at
Stratifin 96704_at
Small proline-rich protein 2f 94120_s-at
Actin, alpha, cardiac 101028_i_at
NADH dehydrogenase (ubiquinone) 1 100628_at
Fibroblast growth factor receptor 4 92937_at
Ras-related protein 97319_at
Small proline-rich protein 2b 99701_f_at
Cyp1b1 Spleen tyrosine kinase 162363_at
99979_at Squalene epoxidase 94322_at
NADH dehydrogenase 92330_r_at
Calbindin-28k 98133_at
Renin 1 structural 98480_s_at
Carboxylesterase 3 101539_f_at
Cytokine receptor-like factor 1 161046_at
Mitogen regulated protein,
proliferin 3 93929_s_at
Tumor-associated calcium signal
transducer 2 160651_at
Fos-like antigen 1 99835_at
Forkhead box a2 93950_at
Proliferin 2 93883_at
Thrombomodulin 104601_at
Fibroblast growth factor 7 99435_at
Actin, alpha, cardiac 101028_i_at
Matrix metalloproteinase 9 99957_at
Forkhead box f2 99846_at
Cytochrome p450, 1a1 94715_at
Insulin-like growth factor 2 98623_g_at
Ras-related protein 97319_at
Follistatin 98817_at
Lymphocyte antigen 6 complex, locus A 93078_at
lgfbp-5 Secreted phosphoprotein 1 97519_at
100566_at Matrix metal loproteinase 3 98833_at
RNA binding motif, single stranded
interacting protein 1 96207_at
Alpha-2-hs-glycoprotein 99862_at
Angiotensinogen 101887_at
Actin, alpha 1, skeletal muscle 100381_at
Renin 1 structural 98480_s_at
Interferon activated gene 202a 94774_at
Retinol binding protein 4, plasma 96047_at
Integrin beta 4 103305_at
Insulin-like growth factor binding
protein 3 95082_at
Hydroxysteroid 11-beta
dehydragenase 2 100493_at
Small proline-rich protein 2c 101761_f_at
Follistatin 98817_at
Integrin alpha 6 95511_at
Carbonic anhydrase 3 160375_at
Lox Lymphocyte antigen 6 complex, locus H 103487_at
160095_at Single stranded interacting protein 1 96207_at
Glutamyl aminopeptidase 102373_at
Peripherin 161482_f_at
Cadherin 16 93515_at
Coagulation factor II 101899_at
Fibroblast growth factor 7 99435_at
Forkhead box a2 93950_at
Thrombomodulin 104601_at
Alcohol dehydrogenase 1, complex 94906_at
Fos-like antigen 1 99835_at
Proliferin 2 93883_at
Forkhead box f2 99846_at
Actin, alpha, cardiac 101028_i_at
Cyclin-dependent kinase inhibitor 1a
(p21) 94881_at
High mobility group at-hook 2 99058_at
Integrin alpha 3 104211_at
Interferon gamma 99334_at
Mitogen regulated protein,
proliferin 3 93929_s_at
Phospholipase a2 group VII 101923_at
Ras-related 97319_at
Opn Brain derived neurotrophic factor 102727_at
97519_at Interleukin 6 102218_at
Proliferin 94838_r_at
Tumor necrosis factor receptor
superfamily, member 1b 94928_at
Insulin-like growth factor binding
protein 3 95082_at
Gro1 oncogene 95348_at
Epiregulin 98802_at
Ras-related protein 97319_at
NADH dehydrogenase 100628_at
Actin, alpha, cardiac 101028_i_at
Actin, alpha 1, skeletal muscle 100381_at
Cyclin-dependent kinase inhibitor 1a
(p21) 94881_at
Interferon gamma 99334_at
Angiopoietin-like 4 96119_s_at
Cd44 antigen 103005_sat
Cytokine receptor-like factor 1 161046_at
High mobility group at-hook 2 99058_at
Tumor-associated calcium signal
transducer 2 160651_at
Forkhead box a2 93950_at
Fos-like antigen 1 99835_at
Matrix metalloproteinase 3 98833_at
Targets Predictor Percentage (b)
Ahr Lymphocyte antigen 6 complex, locus e 18
160495_at Insulin-like growth factor binding
protein 3 16
Tumor necrosis factor receptor
superfamily, member 1b 14
Insulin-like growth factor binding
protein 6 13
Brain derived neurotrophic factor 13
Secreted phosphoprotein 1 9
Stratifin 7
Small proline-rich protein 2f 6
Actin, alpha, cardiac 6
NADH dehydrogenase (ubiquinone) 1 6
Fibroblast growth factor receptor 4 5
Ras-related protein 5
Small proline-rich protein 2b 5
Cyp1b1 Spleen tyrosine kinase 31
99979_at Squalene epoxidase 18
NADH dehydrogenase 14
Calbindin-28k 10
Renin 1 structural 10
Carboxylesterase 3 10
Cytokine receptor-like factor 1 9
Mitogen regulated protein,
proliferin 3 9
Tumor-associated calcium signal
transducer 2 9
Fos-like antigen 1 9
Forkhead box a2 8
Proliferin 2 8
Thrombomodulin 8
Fibroblast growth factor 7 8
Actin, alpha, cardiac 7
Matrix metalloproteinase 9 7
Forkhead box f2 7
Cytochrome p450, 1a1 7
Insulin-like growth factor 2 6
Ras-related protein 6
Follistatin 6
Lymphocyte antigen 6 complex, locus A 5
lgfbp-5 Secreted phosphoprotein 1 14
100566_at Matrix metal loproteinase 3 9
RNA binding motif, single stranded
interacting protein 1 8
Alpha-2-hs-glycoprotein 7
Angiotensinogen 7
Actin, alpha 1, skeletal muscle 7
Renin 1 structural 7
Interferon activated gene 202a 6
Retinol binding protein 4, plasma 6
Integrin beta 4 6
Insulin-like growth factor binding
protein 3 5
Hydroxysteroid 11-beta
dehydragenase 2 5
Small proline-rich protein 2c 5
Follistatin 5
Integrin alpha 6 5
Carbonic anhydrase 3 5
Lox Lymphocyte antigen 6 complex, locus H 36
160095_at Single stranded interacting protein 1 36
Glutamyl aminopeptidase 23
Peripherin 21
Cadherin 16 16
Coagulation factor II 11
Fibroblast growth factor 7 11
Forkhead box a2 11
Thrombomodulin 11
Alcohol dehydrogenase 1, complex 9
Fos-like antigen 1 9
Proliferin 2 9
Forkhead box f2 7
Actin, alpha, cardiac 5
Cyclin-dependent kinase inhibitor 1a
(p21) 5
High mobility group at-hook 2 5
Integrin alpha 3 5
Interferon gamma 5
Mitogen regulated protein,
proliferin 3 5
Phospholipase a2 group VII 5
Ras-related 5
Opn Brain derived neurotrophic factor 15
97519_at Interleukin 6 14
Proliferin 14
Tumor necrosis factor receptor
superfamily, member 1b 13
Insulin-like growth factor binding
protein 3 12
Gro1 oncogene 10
Epiregulin 9
Ras-related protein 8
NADH dehydrogenase 7
Actin, alpha, cardiac 7
Actin, alpha 1, skeletal muscle 7
Cyclin-dependent kinase inhibitor 1a
(p21) 6
Interferon gamma 6
Angiopoietin-like 4 6
Cd44 antigen 5
Cytokine receptor-like factor 1 5
High mobility group at-hook 2 5
Tumor-associated calcium signal
transducer 2 5
Forkhead box a2 5
Fos-like antigen 1 5
Matrix metalloproteinase 3 5
Targets Predictor Correlation (c)
Ahr Lymphocyte antigen 6 complex, locus e 0.77
160495_at Insulin-like growth factor binding
protein 3 0.83
Tumor necrosis factor receptor
superfamily, member 1b 0.70
Insulin-like growth factor binding
protein 6 0.72
Brain derived neurotrophic factor 0.80
Secreted phosphoprotein 1 0.87
Stratifin 0.23
Small proline-rich protein 2f 0.18
Actin, alpha, cardiac -0.84
NADH dehydrogenase (ubiquinone) 1 -0.84
Fibroblast growth factor receptor 4 0.31
Ras-related protein -0.83
Small proline-rich protein 2b 0.20
Cyp1b1 Spleen tyrosine kinase -0.44
99979_at Squalene epoxidase 0.37
NADH dehydrogenase 0.64
Calbindin-28k -0.20
Renin 1 structural -0.10
Carboxylesterase 3 0.86
Cytokine receptor-like factor 1 0.92
Mitogen regulated protein,
proliferin 3 0.92
Tumor-associated calcium signal
transducer 2 -0.12
Fos-like antigen 1 0.89
Forkhead box a2 0.86
Proliferin 2 0.93
Thrombomodulin 0.73
Fibroblast growth factor 7 0.94
Actin, alpha, cardiac -0.68
Matrix metalloproteinase 9 0.13
Forkhead box f2 0.70
Cytochrome p450, 1a1 0.01
Insulin-like growth factor 2 -0.03
Ras-related protein -0.65
Follistatin 0.73
Lymphocyte antigen 6 complex, locus A 0.37
lgfbp-5 Secreted phosphoprotein 1 -0.64
100566_at Matrix metal loproteinase 3 -0.61
RNA binding motif, single stranded
interacting protein 1 -0.57
Alpha-2-hs-glycoprotein 0.27
Angiotensinogen -0.86
Actin, alpha 1, skeletal muscle 0.68
Renin 1 structural 0.23
Interferon activated gene 202a 0.04
Retinol binding protein 4, plasma 0.40
Integrin beta 4 0.43
Insulin-like growth factor binding
protein 3 -0.40
Hydroxysteroid 11-beta
dehydragenase 2 0.24
Small proline-rich protein 2c 0.09
Follistatin -0.57
Integrin alpha 6 0.85
Carbonic anhydrase 3 0.26
Lox Lymphocyte antigen 6 complex, locus H -0.37
160095_at Single stranded interacting protein 1 0.55
Glutamyl aminopeptidase -0.55
Peripherin -0.35
Cadherin 16 -0.40
Coagulation factor II -0.30
Fibroblast growth factor 7 0.95
Forkhead box a2 0.95
Thrombomodulin 0.91
Alcohol dehydrogenase 1, complex -0.35
Fos-like antigen 1 0.97
Proliferin 2 0.93
Forkhead box f2 0.87
Actin, alpha, cardiac -0.60
Cyclin-dependent kinase inhibitor 1a
(p21) -0.53
High mobility group at-hook 2 0.85
Integrin alpha 3 0.32
Interferon gamma 0.53
Mitogen regulated protein,
proliferin 3 0.98
Phospholipase a2 group VII -0.48
Ras-related -0.56
Opn Brain derived neurotrophic factor 0.73
97519_at Interleukin 6 0.70
Proliferin 0.52
Tumor necrosis factor receptor
superfamily, member 1b 0.76
Insulin-like growth factor binding
protein 3 0.92
Gro1 oncogene 0.73
Epiregulin 0.71
Ras-related protein -0.91
NADH dehydrogenase -0.91
Actin, alpha, cardiac -0.92
Actin, alpha 1, skeletal muscle -0.85
Cyclin-dependent kinase inhibitor 1a
(p21) 0.04
Interferon gamma 0.38
Angiopoietin-like 4 0.69
Cd44 antigen 0.71
Cytokine receptor-like factor 1 0.72
High mobility group at-hook 2 0.80
Tumor-associated calcium signal
transducer 2 0.31
Forkhead box a2 0.67
Fos-like antigen 1 0.72
Matrix metalloproteinase 3 0.55
(a) Murine genome U74v2 (for complete details:
http://www.affymetrix.com). (b) Percentage representing the number
of times a given predictor was used forgiven target. (c) Linear
correlation of the predictor genes to the given target.
Figure 1. Schematic representation of gene ternary values and predictor
functions [psi]. The intensity of genes [G.sub.1], [G.sub.2], and
[G.sub.3] can be used to predict the values of the target [G.sub.4] via
a ternary function [psi]. An arbitrary ternary function [psi] defined
by [psi] ([x.sub.1], [x.sub.2], [x.sub.3]) = [x.sub.1] + [x.sub.2] -
[x.sub.3] [with the sum defined inside (-1, 0, 1)] is shown on line 5.
The quality of [psi] to explain [G.sub.4] from [G.sub.1], [G.sub.2],
and [G.sub.3] can be measured by its expected absolute error [epsilon]
[[psi]] = E[[absolute value of [psi]([x.sub.1], [x.sub.2], [x.sub.3])-
[x.sub.4]]], where | | denotes absolute value, and the expectation is
computed over the joint distribution of the random variables [x.sub.1],
[x.sub.2], [x.sub.3], and [x.sub.4]. In line 6, the difference
[absolute value of ([x.sub.1](j), [x.sub.2](j), [x.sub.3](j) -
[x.sub.4](j)], for j = 1 to 9 is shown to exemplify absolute
differences. An estimation of the value of [x.sub.4], assuming a
constant value, is shown on line 7.
j=1 j=2 j=3 j=4 j=5
[x.sub.1] (j) -1 -1 -1 1 0
[x.sub.2] (j) -1 1 -1 1 0
[x.sub.3] (j) -1 0 -1 -1 1
[x.sub.4] (j) -1 0 0 0 0
[psi] ([x.sub.1],[x.sub.2],[x.sub.3]) -1 0 -1 0 1
[absolute value of [psi] ([x.sub.1],
[x.sub.2],[x.sub.3])-[x.sub.4](j)] 0 0 1 0 1
[[psi].sub.0] ([x.sub.4]) -1 -1 -1 -1 -1
j=6 j=7 j=8 j=9
[x.sub.1] (j) -1 0 0 0
[x.sub.2] (j) -1 1 0 -1
[x.sub.3] (j) -1 -1 -1 1
[x.sub.4] (j) -1 -1 -1 1
[psi] ([x.sub.1],[x.sub.2],[x.sub.3]) -1 -1 1 1
[absolute value of [psi] ([x.sub.1],
[x.sub.2],[x.sub.3])-[x.sub.4](j)] 0 0 2 0
[[psi].sub.0] ([x.sub.4]) -1 -1 -1 -1
REFERENCES Affymetrix, Inc. 2003. GeneChip Expression Analysis Technical Manual, Rev. 4. Santa Clara, CA: Affymetrix, Inc. Ando J, Tsuboi H, Korenaga R, Takahashi K, Kosaki K, Isshiki M, et al. 1996. Differential display and cloning of shear stress-responsive messenger RNAs in human endothelial cells Endothelial cells The cells lining the inner walls of the blood vessels. Mentioned in: Von Willebrand Disease . Biochem Biophys Res Commun 225:347-351. Andreola F, Fernandez-Salguero PM, Chiantore MV, Petkovich MP, Gonzalez FJ, De Luca LM. 1997. Aryl hydrocarbon receptor knockout mice (AHR-/-) exhibit liver retinoid retinoid /ret·i·noid/ (ret´i-noid) 1. resembling the retina. 2. retinal, retinol, or any structurally similar natural derivative or synthetic compound, with or without vitamin A activity. accumulation and reduced retinoic acid metabolism. Cancer Res 57:2835-2838. Bittinger MA, Nguyen LP, Bradfield CA. 2003. Aspartate aminotransferase generates pro-agonists of the aryl hydrocarbon receptor. Mol Pharmacol 64:550-556. Breitkreutz BJ, Stark C, Tyers M. 2003. Osprey osprey (ŏs`prē), common name for a bird of prey related to the hawk and the New World vulture and found near water in most parts of the world. : a network visualization system. Genome Biol 4:R22.1-R22.4 Buters JT, Mahadevan B, Quintanilla-Martinez L, Gonzalez FJ, Greim H, Baird WM, et al. 2002. Cytochrome P450 1B1 determines susceptibility to dibenzo[a,/]lpyrene-induced tumor formation. Chem Res Toxicol 15:1127-1135. Carlson DB, Perdew GH. 2002. A dynamic role for the Ah receptor in cell signaling? Insights from a diverse group of Ah receptor interacting proteins. J Biochem Mol Toxicol 16:317-325. Carver LA, Bradfield, CA. 1997. Ligand-dependent interaction of the aryl hydrocarbon receptor with a novel immunophilin homolog hom·o·log n. Variant of homologue. in vivo. J Biol Chem 272:11452-11456. Chapman RS, Lourenco PC, Tonner E, Flint DJ, Selbert S, Takeda K, et al. 1999. Suppression of epithelial apoptosis and delayed mammary gland involution involution /in·vo·lu·tion/ (in?vo-loo´shun) 1. a rolling or turning inward. 2. a retrograde change of the body or of an organ, as the retrograde changes in size of the female genital organs after delivery. in mice with a conditional knockout of Stat3. Genes Dev 13:2604-2616. Chellaiah MA, Kizer N, Biswas R, Alvarez U, Strauss-Schoenberger J, Rifas L, et al. 2003. Osteopontin deficiency produces osteoclast dysfunction due to reduced Cd44 surface expression. Mol Biol Cell 14:173-189. Coopman PJ, Do MT, Berth M, Bowden ET, Hayes AJ, Basyuk E, et at. 2000. The Syk tyrosine kinase suppresses malignant growth of human breast cancer cells. Nature 406:742-747. Dailly YP, Zhou Y, Linkhart TA, Baylink DJ, Strong DD. 2001. Structure and characterization of the human insulin-like growth factor binding protein The Insulin-like growth factor binding protein serves as a carrier protein for Insulin-like growth factor 1. Approximately 98% of IGF-1 is always bound to one of 6 binding proteins (IGF-BP). IGFBP-3, the most abundant protein, accounts for 80% of all IGF binding. (Igfbp)-6 promoter: identification of a functional retinoid response element. Biochim Biophys Acta 1518:145-151. Denhardt DT, Giachelli CM, Rittling SR. 2001. Role of osteopontin in cellular signaling and toxicant toxicant /tox·i·cant/ (tok´si-kant) 1. poisonous. 2. poison. tox·i·cant n. 1. A poison or poisonous agent. 2. An intoxicant. adj. injury. Annu Rev Pharmacol Toxicol 41:723-749. Denison MS, Nagy SR. 2003. Activation of the aryl hydrocarbon receptor by structurally diverse exogenous and endogenous chemicals. Annu Rev Pharmacol Toxicol 43:309-334. Devroye L, Gyorffy L, Lugosi G. 1996. A Probabilistic Theory of Pattern Recognition. New York New York, state, United States New York, Middle Atlantic state of the United States. It is bordered by Vermont, Massachusetts, Connecticut, and the Atlantic Ocean (E), New Jersey and Pennsylvania (S), Lakes Erie and Ontario and the Canadian province of : Springer-Verlag. Duan C, Hawes SB, Prevette T, Clemmons DR. 1996. Insulin-like growth factor-I (IGF-I) regulates IGF-binding protein-5 synthesis through transcriptional activation of the gene in aortic smooth muscle cells. J Biol Chem 271:4280-4288. Falahetpisheh MH, Ramos KS. 2003. Ligand-activated Ahr signaling leads to disruption of nephrogenesis and altered Wilms' tumor Wilms' Tumor Definition Wilms' tumor is a cancerous tumor of the kidney that usually occurs in young children. It is named for Max Wilms, a German surgeon (1867–1918) and is also known as a nephroblastoma. suppressor mRNA splicing splicing /splic·ing/ (spli´sing) 1. the attachment of individual DNA molecules to each other, as in the production of chimeric genes. 2. RNA s. . Oncogene 10:2160-2171. Gonzalez FJ, Fernandez-Salguero P. 1998. The aryl hydrocarbon receptor: studies using the AHR-null mice. Drug Metab Dispos 26:1194-1198. Horie M, Okutomi K, Taniguchi Y, Ohbuchi Y, Suzuki M, Takahashi E. 1998. Isolation and characterization of a new member of the human Ly6 gene family (Ly6h). Genomics 53:365-388. Huang E, Ishida S, Pittman J, Dressman H, Gild A, Kloos M, et al. 2003. Gone expression phenotypic models that predict the activity of oncogenic oncogenic /on·co·gen·ic/ (-jen´ik) giving rise to tumors or causing tumor formation; said especially of tumor-inducing viruses. on·co·gen·ic or on·cog·e·nous adj. pathways. Nat Genet genet: see civet. 34:226-230. Johnston MC, Bronsky PT. 1995. Prenatal craniofacial craniofacial /cra·nio·fa·cial/ (kra?ne-o-fa´sh'l) pertaining to the cranium and the face. cra·ni·o·fa·cial adj. Of or involving both the cranium and the face. development: new insights on normal and abnormal mechanisms. Crit Rev Oral Biol Med 6:368-422. Kerkvliet NI, Shepherd OM, Baecher-Steppan L. 2002. T lymphocytes are direct, aryl hydrocarbon receptor (AhR)-dependent targets of 2, 3, 7, 8-tetrachlorodibenzo-p-dioxin (TCDD): AhR expression in both CD[4.sup.+] and CD[8.sub.+] T cells is necessary for full suppression of a cytotoxic T lymphocyte cytotoxic T lymphocyte CTL, cytotoxic T cell Immunology A subset of T cells with a CD8 receptor on the surface that recognizes and lyses malignant or virally-infected self cells bearing self, ie 'haplotype restricted', class I MHC molecules. response by TCDD. Toxicol Appl Pharmacol 185:146-152. Kerzee JK, Ramos KS. 2000. Activation of c-Ha-ras by benzo(a)pyrene in vascular smooth muscle cells involves redox redox (rē`dŏks): see oxidation and reduction. stress and aryl hydrocarbon receptor. Mol Pharmacol 58:152-158. --. 2001. Constitutive and inducible expression of Cyp1a1 and Cyp1b1 in vascular smooth muscle cells: role of the Ahr bHLH/PAS transcription factor. Circ Res 89:573-582. Khalsa PS, Zhang C, Sommerfeldt D, Hadjiargyrou M. 2000. Expression of integrin [alpha]2[beta]1 in axons and receptive endings of neurons in rat, hairy skin. Neurosci Lett 293:13-26. Kim S, Dougherty ER, Bittner ML, Chen Y, Sivakumar K, Meltzer PS, et al. 2000a. Multivariate measurement of gene expression relationships. Genomics 67:201-209. --. 2000b. General nonlinear framework for the analysis of gone interaction via multivariate expression arrays. J Biomed Optics 5:411-424. Kuivaniemi H, Savolainen ER, Kivirikko KI. 1984. Human placental lysyl oxidase: purification, partial characterization, and preparation of two specific antisera to the enzyme. J Biol Chem 259:6996-7002. Lee R, Kermani P, Teng KK, Hampstead BL. 2001. Regulation of cell survival by secreted proneurotrophins. Science 294:1945-1948. Li W, Nellaiappan K, Strassmaier T, Graham L, Thomas KM, Kagan HM. 1997. Localization Customizing software and documentation for a particular country. It includes the translation of menus and messages into the native spoken language as well as changes in the user interface to accommodate different alphabets and culture. See internationalization and l10n. and activity of lysyl oxidase within nuclei of fibrogenic cells. Proc Nat Acad Sci USA 94:12917-12922. Lin TH, Rosales C, Mondal K, Bolen JB, Haskill S, Juliano RL. 1995. Integrin-mediated tyrosine phosphorylation and cytokine message induction in monocytic cells. A possible signaling role for the Syk tyrosine kinase. J Biol Chem 270:16189-16197. Mao M, Yu M, Tong J-H, Ye J, Zhu J, Huang Q-H, et at. 1996. RIG-E, a human homolog of the murine Ly-6 family, is induced by retinoic acid during the differentiation of acute promyelocytic leukemia acute pro·my·e·lo·cyt·ic leukemia n. A severe bleeding disorder that is a form of leukemia and is characterized by low concentrations of plasma fibrigen, defective coagulation, and infiltration of the bone marrow with abnormal promyelocytes and cell. Proc Nat Acad Sci USA 93:5910-5914. Mounho BJ, Burchiel SW. 1998. Alterations in human B cell calcium homeostasis homeostasis Any self-regulating process by which a biological or mechanical system maintains stability while adjusting to changing conditions. Systems in dynamic equilibrium reach a balance in which internal change continuously compensates for external change in a feedback by polycyclic aromatic hydrocarbons: possible associations with cytochrome P450 metabolism and increased protein tyrosine phosphorylation. Toxicol Appl Pharmacol 149:80-89. Nam TJ, Busby WH Jr, Rees C, Clemmons DR. 2000. Thrombospondin and osteopontin bind to insulin-like growth factor insulin-like growth factor one of the twenty or so substances, additional to the classic bone-regulating hormones, which exert an effect on bone cell metabolism. See also somatomedin C. (IGF (Internet Governance Forum) An international organization of governments and U.N. agencies that was founded to discuss Internet issues such as security and spam. It was created at the United Nations Summit in 2005 after the U.S. )-binding protein-5 leading to an alteration in IGF-I-stimulated cell growth. Endocrinology 141:1100-1106. Nebert DW. 1994. Drug metabolism and signal transduction: possible role of Ah receptor and arachidonic acid cascade in protection from ethanol toxicity. EXS EXS Expandable Switching System (Lucent) EXS Executed Schedule 71:231-240. Nichols TC, du Laney T, Zheng B, Bellinger DA, Nickols GA, Engleman W, et al. 1999. Reduction in atherosclerotic lesion size in pigs by alphaVbeta3 inhibitors is associated with inhibition of insulin-like growth factor-l-mediated signaling. Circ Res 85:1040-1045. Ramos KS, Cox LR. 1993. Aortic endothelial endothelial /en·do·the·li·al/ (-the´le-al) pertaining to or made up of endothelium. Endothelial A layer of cells that lines the inside of certain body cavities, for example, blood vessels. and smooth muscle cell cultures. Vol 1 (Tyson CA, Frazier JM, eds). In: Methods in Toxicology: In Vitro Biological Systems. St. Louis, MO: Academic Press, 159-168. Schedlich L J, Graham LD. 2002. Role of insulin-like growth factor binding protein-3 in breast cancer cell growth. Microsc Res Tech 59:12-22. Shu Y, Hoshi S, Tomari S, Watanabe T, Negate M. 2002. Phenotypic changes and cell cycle activation in early tubulointerstitial injury of rat adriamycin nephrosis nephrosis (nəfrō`səs), kidney disease characterized by lesions of the epithelial lining of the renal tubules, resulting in marked disturbance in the filtration function and the consequent appearance of large amounts of protein (albumin) . Pathol Int 52:214-223. Singh SU, Casper RF, Fritz PC, Sukhu B, Ganss B, Girard B Jr, et al. 2000. Inhibition of dioxin effects on bone formation in vitro by a newly described aryl hydrocarbon receptor antagonist, resveratrol. J Endocrinol 167:183-195. Sodek J, Zhu B, Huynh MH, Brown TJ, Ringuette M 2002. Novel functions of the matricellular proteins osteopontin and osteonectin/SPARC. Connect Tissue Res 43:308-319. Song YL, Ford JW, Gordon D, Shanley CJ. 2000. Regulation of lysyl oxidase by interferon-gamma in rat aortic smooth muscle ceils. Arterioscler Thromb Vasc Biol 20:982-988. Soprano DR, Soprano KJ. 2003. Pharmacological doses of some synthetic retinoids can modulate both the aryl hydrocarbon receptor and retinoid receptor pathways. J Nutr 133:277S-281S. Toyoda H, Komurasaki T, Uchida D, Takayama Y, Isobe T, Okuyama T, et al. 1995. Epiregulin: a novel epidermal growth factor Epidermal growth factor or EGF is a growth factor that plays an important role in the regulation of cell growth, proliferation and differentiation. Human EGF is a 6045 Da protein with 53 amino acid residues and three intramolecular disulfide bonds. with mitogenic activity for rat primary hepatocytes. J Biol Chem 270:7495-7500. Vondracek M, Weaver DA, Sarang Z, Hedberg JJ, Willey JC, Warngard L, et al. 2002. Transcript profiling of enzymes involved in detoxification Detoxification Definition Detoxification is one of the more widely used treatments and concepts in alternative medicine. It is based on the principle that illnesses can be caused by the accumulation of toxic substances (toxins) in the body. of xenobiotics and reactive oxygen in human normal and simian virus 40 T antigen-immortalized oral keratinocytes Keratinocytes Cells found in the epidermis. The keratinocytes at the outer surface of the epidermis are dead and form a tough protective layer. The cells underneath divide to replenish the supply. . Int J Cancer 99:776-782. Wilson E, Parrish AR, Bral CM, Williams ES, Ramos KS. 2002. Collagen suppresses the proliferative phenotype of allylamine-injured vascular smooth muscle cells. Atherosclerosis 162:289-297. Xie Y, Sakatsume M, Nishi S, Narita I, Arakawa M, Gejyo F. 2001. Expression, roles, receptors, and regulation of osteopontin in the kidney. Kidney Int 60:1645-1657. Charles D. Johnson, (1,2) Yoganand Balagurunathan, (3) Mahlet G. Tadesse, (4) M. Hadi Falahatpisheh, (1,2) Marcel Brun, (1,3) Mary K. Walker, (5) Edward R. Dougherty, (3) and Kenneth S. Ramos (1,2) (1) Department of Biochemistry and Molecular Biology and Center for Genetics and Molecular Medicine, University of Louisville See also
1. ^ [1] 2. ^ [2] URL accessed on June 8 2006 3. , Louisville, Kentucky, USA; Departments of (2) Veterinary Physiology and Pharmacology, (3) Electrical Engineering, and (4) Statistics, Texas A&M University, College Station, Texas College Station is a city in Brazos County, Texas, situated in Central Texas. It is located in the heart of the Brazos Valley. The city is located within the most populated region of Texas, near to three of the 10 largest cities in the United States - Houston, Dallas, and San , USA; (5) Department of Pharmacology and Toxicology, University of New Mexico The University of New Mexico (UNM) is a public university in Albuquerque, New Mexico. It was founded in 1889. It also offers multiple bachelor's, master's, doctoral, and professional degree programs in all areas of the arts, sciences, and engineering. , Albuquerque, New Mexico “Albuquerque” redirects here. For other uses, see Albuquerque (disambiguation). Albuquerque (pronounced [ˈæl.bə.kɚ.kiː], Spanish: [al.βu. , USA Address correspondence to K.S. Ramos, Department of Biochemistry and Molecular Biology, University of Louisville Health Sciences Center, Louisville, KY 40292 USA. Telephone: (502) 852-5217. Fax: (502) 852-6222. E-mail: kenneth.ramos@louisville.edu * The online version of this article (available at http://www.ehponline.org) contains Supplemental Material. This research was supported by National Institutes of Health grants ES04849, ES09106, ES07273 to K.S.R and ES09804 and ES012072 to M.K.W. C.D.J was supported by National Research Service Award (NSRA NSRA National Small Bore Rifle Association (UK) NSRA National Shorthand Reporters Association NSRA National Snowmobile Racing Association NSRA Nova Scotia Rifle Association (Dartmouth, NS, Canada) ) ES012117, M.G.T by National Cancer Institute CA90301, and M.H.F. by NRSA NRSA National Research Service Award (US National Institutes of Health) NRSA National Remote Sensing Agency (India) NRSA Non-Revenue Space Available (airline travel) ES012542. The authors declare they have no competing financial interests. Received 23 September 2003; accepted 14 January 2004. |
|
||||||||||||||||||||
Printer friendly
Cite/link
Email
Feedback
Reader Opinion