Home Data Index Rat Study Overview

Microarray Expression Profiling of Rat Consomic Panels

Contents:
Overview
Consomic Animals
Microarray Clone Set
General Experimental Design
Data Analysis

Publications

Overview
Genetic and environmental risk factors contribute to the variability in physiological phenotypes. It is often difficult to separate the contribution of each parameter to disease susceptibility. However, by utilizing well characterized, in-bred model animals, whereby both genes and environment can be carefully controlled, the relative contribution of genetic components and environmental factors can be deciphered. To increase our understanding of the genetic basis of heart, lung, kidney, blood and vasculature responses to environmental stresses and disease processes, microarray expression profiling was performed on a number of chromosomal substitution strains and their genetically diverse parental stains. Consomic rat strains were derived by our collaborators by PhysGen (Medical College of Wisconsin), who are using chromosomal substitution strains to dissect multigenic diseases. Panels of chromosomal substitution strains of rats (consomic rat panels) are being developed and subjected to environmental stressors such as hypoxia, exercise, and high salt intake to unmask deficiencies in normal homeostatic and idiopathic mechanisms that contribute to disease. To generate a panel of consomic rats, chromosomes from one genetic strain (acceptor strain) are replaced, one at a time, with the corresponding chromosome from a second genetic strain (donor). By doing this, the contribution of genes on each chromosome can be assessed by phenotyping the consomic strain for disease traits of interest. We are utilizing these consomic panels for microarray analysis, which will offer an unprecedented means to link complex gene expression profiles to pathophysiological pathways. Comparisons between consomic strains will provide valuable insights into how differential gene expression may influence a cardiovascular trait. Towards this goal, the consomic rat panels will be characterized by PhysGen (Medical College of Wisconsin) using over 200 phenotypes specific to heart, lung, kidney, vasculature, and blood function in response to environmental stressors¹. Physiogenomic data from PhysGen, coupled with gene expression profiling data generated by TREX (The Institute for Genomic Research) will provide an unprecedented look at the role of genes and environment in disease states.

1. Cowley Jr, A. W. et al. Brown Norway chromosome 13 confers protection from high salt to consomic Dahl S rat. Hypertension 37, 456-61 (2001).

Consomic Animals
In a complete panel of consomic rats (22 strains), each chromosome from strain A is introgressed into the genetic background of strain B, so that the contribution of genes on each chromosome can be assessed by phenotyping and expression profiling. For example, the SS.13BN animal is derived from the introgression of chromosome 13 from the Brown Norway (BN) donor strain into the genetic background of the SS acceptor stain. A reciprocal panel (SS chromosome introgressed into BN background) can also be constructed so that the chromosomes from each strain are placed on each genetic background (susceptibility versus resistance). Medical College of Wisconsin will produce the SS.BN panel for chromosomes 1-20, X, Y and the FHH.BN panel (chromosomes 1-20, X, Y). The FHH.BN panel, Fawn-hooded Hypertensive (FHH) background with chromosomal introgression from BN animals, will also be available for expression profiling. For cardio-vascular studies, one of the strengths for selecting the rat is the wide strain diversity. Consomic animals can also be used to identify modifier genes¹. Environmental stressors such as hypoxia, exercise, and high salt intake can be used to unmask deficiencies in normal homeostatic and idiopathic mechanisms that contribute to disease. The three genetically diverse parental strains used for these studies are the Fawn Hooded Hypertensive (FHH), the Dahl Salt Sensitive (SS) and the Brown Norway (BN). The BN strain is assumed to be most closely related to wild rats and the complete genome from a line inbred at the MCW colony (BN/SsNHsd/Mcwi) was recently sequenced². Detailed genetic mapping has been done on these strains^3-6 and the consomic panels capture, on average, half of the genetic variation present in the most commonly used inbred rats. BN is normotensive, resistant to cardiac ischemia and is a model for asthma and other pulmonary diseases⁷. SS rats serve primarily to study salt-sensitive hypertension, insulin resistance and end-stage renal disease^8-11. FHH animals develop both systolic and pulmonary hypertension, have a platelet storage defect and are susceptible to renal disease^12,13. Consomic rat strains allow alterations in phenotypic traits to be associated with changes in gene expression profiles.

1. Kwitek-Black, A. E. & Jacob, H. J. The use of designer rats in the genetic dissection of hypertension. Curr Hypertens Rep 3, 12-8 (2001).
2. Consortium, R. G. S. P. Genome sequence of the Brown Norway rat yields insights into mammalian evolution. Nature 428, 493-521 (2004).
3. Stoll, M. et al. A genomic-systems biology map for cardiovascular function. Science 294, 1723-6 (2001).
4. Steen, R. G. et al. A high-density integrated genetic linkage and radiation hybrid map of the laboratory rat. Genome Res 9, AP1-8 (1999).
5. Thomas, M. A., Chen, C. F., Jensen-Seaman, M. I., Tonellato, P. J. & Twigger, S. N. Phylogenetics of rat inbred strains. Mamm Genome 14, 61-4 (2003).
6. Moreno, C. et al. Genomic map of cardiovascular phenotypes of hypertension in female Dahl S rats. Physiol. Genomics 15, 243-57 Epub (2003).
7. Cui, Z. H. et al. Bronchial hyperresponsiveness, epithelial damage, and airway eosinophilia after single and repeated allergen exposure in a rat model of anhydride-induced asthma. Allergy 52, 739-46 (1997).
8. Rapp, J. P. Dahl salt-susceptible and salt-resistant rats. A review. Hypertension 4, 753-63 (1982).
9. Rapp, J. P. Genetic analysis of inherited hypertension in the rat. Physiol. Rev. 80, 135-72 (2000).
10. Roman, R. J., Alonso-Galicia, M. & Wilson, T. W. Renal P450 metabolites of arachidonic acid and the development of hypertension in Dahl salt-sensitive rats. Am J Hypertens 10, 63S-67S (1997).
11. Kotchen, T. A., Zhang, H. Y., Covelli, M. & Blehschmidt, N. Insulin resistance and blood pressure in Dahl rats and in one-kidney, one-clip hypertensive rats. Am J Physiol 261, E692-7 (1991).
12. Kuijpers, M. H. & de Jong, W. Spontaneous hypertension in the fawn-hooded rat: a cardiovascular disease model. J Hypertens Suppl 4, S41-4 (1986).
13. Kreisberg, J. I. & Karnovsky, M. J. Focal glomerular sclerosis in the fawn-hooded rat. Am J Pathol 92, 637-52 (1978).

The TIGR 26k Rat Microarray Set
In 1997, we initiated a project funded by the National Heart, Lung and Blood Institute to generate 3'ESTs as part of an initiative to catalog the expressed genes in the rat genome. Rat 3'EST sequences were derived from single-pass sequencing of the 3'-end (containing the polyA tail) of randomly selected cDNA clones from 12 libraries (10 normalized and 2 non-normalized) representing 11 different tissues and 2 developmental states (fetus and adult). The normalized libraries (constructed by M. Bento Soares, University of Iowa) were from rat fetus, placenta, ovary, heart, lung, liver, spleen, kidney, skeletal muscle and adult brain. The non-normalized libraries (constructed at TIGR) were from a pheochromocytoma cell line treated with and without growth factor. To date, 3'ESTs have been generated from a set of ~52,000 cDNA clones and the resulting 3'ESTs have been assembled into overlapping tentative consensus sequences or TCs. The clone set represents 20,465 unique sequences (16,306 unique Tentative Consensus sequences plus 4,159 singleton ESTs based on the current TIGR Rat Gene Index Build 11.0. A description about the rat TCs can be accessed in TIGR's Rat Gene Index. The minimal clone set serves used as a resource for microarray experiments has been subjected to single-pass sequencing from the 5'-end (5'EST) to provide clone verification and gene identification. Annotation for the TIGR 26K Rat Microarray Set is freely available here.

General Experimental Design
Microarray expression profiling of the rat consomic panels and parental strains is documented in four tissues (heart, lung, kidney, lung), both genders, and two environmental conditions (hypoxia 12% O2; normoxia, 21% O2). Animal husbandry, tissue procurement and RNA isolation are performed at PhysGen, using standard protocols. Briefly, RNA is extracted using Trizol and further purified on a Qiagen RNeasy column. Total RNA (15 µg) is labeled with Cy dyes using an indirect labeling method and hybridized to TIGR 26K Rat Microarrays. We use a small loop design in our microarray experiments consisting of replicate dye-swap hybridizations. The replicates are derived from two independent pools of total RNA. A limited number of reference hybridizations were used to examine the parental strain, to allow comparisons to be made between the parental animals. For the Reference design, a dye-swap strategy was not used. Rather, replicate pools of total RNA were hybridized in duplicate. Of note, we do not use the loop design in data analysis, rather we use the direct comparisons contained within the design. However, the experimental design does allow indirect comparisons to be made, should individual investigators wish to do so. Physiogenomic data was generated from individual animals, independent from the animals used to generate the microarray data.

Small loop design for direct comparison between parental and consomic animals, or normoxic vs. hypoxic treated animals. In the example below, the SS parentals, male and female, can be directly compared to SS.13BN animals under normoxic or hypoxic conditions (vertical arrows). Or, hypoxic and normoxic gene expression profiles can be compared within a strain (horizontal arrows). Each rat below represents total RNA isolated from a pool of three animals. Forty-eight animals were used per study. A complete listing of completed studies is here. The same strategy is used for the entire consomic panel. Arrows indicate the direction of dye labeling of the RNA samples. Cy3 -> Cy5 or Cy5 -> Cy3.

Reference design - exclusively used for comparing cardiac gene expression profiles between the parental strains. Common reference RNAs were generated by mixing equal concentrations of commercially available RNA (Clontech) from five rat tissues: brain, heart, kidney, liver and lung. Each rat below represents total RNA isolated from a pool of three animals.

Quality Control Data Filtering.
Representative QC scatter plots of female SS lung versus SS.16BN lung hybridizations. Animals were exposed to hypoxia (12% oxygen) for 2 weeks prior to isolation of RNA. (Top) Rep A hybridizations. Log2 expression ratios with Z-scores > ± 1.96 standard deviations from the array mean of a hybridization between SS lung cDNA labeled with Cy3 versus SS.16BN lung cDNA labeled with Cy5 are plotted against a dye-swapped hybridization (i.e. SS lung cDNA labeled with Cy5 versus SS.16BN lung cDNA labeled with Cy3). Note strong correlation and slope close to 1, indicating high quality hybridizations. (Center) Rep B hybridizations. A similar dye-swap hybridization scheme was used for Rep B samples as in Rep A. Rep B hybridizations are biological replicates of Rep A. A plot of a dye-swap hybridization pair generally has a positive regression slope and a correlation > 0.65. However, where strain differences are minimal, the expression data will be compressed. This is not unexpected for highly similar gene expression profiles and approximates a self-to-self comparison (i.e. RNA sample is divided in half where one-half is labeled with Cy3 and the other half with Cy5). (Lower) The mean log2 ratios from Rep A (log2 ratios from the two hybridizations of a dye-swap pair are averaged) are compared to the mean log2 ratios from Rep B. Again, the regression slope is positive when strain differences are present. If the above analysis suggests minimal differences are present between two strains, or treatments (approximating a self-to-self comparison), the hybridizations are repeated for confirmation and posted to the external web-site.

Data Analysis
TIGR Rat Expression Datasets can be accessed by BROWSING the rat individual studies or by using the SEARCH function where gender (male or female), tissue (kidney, liver, lung, heart), stress (hypoxia or normoxia), and strain (currently > 16 consomics and 3 parental strains profiled) are selected. The output can be further limited by statistical parameters (raw, standard deviation regularization, flip-dye) and file format (Stanford, MEV, bar graph). Standard deviation regularization assumes each slide should have the same spread for the log2 ratio of the single channel intensities. The intensity pair for each gene is then scaled appropriately. Flip-dye normalization outputs genes that are consistent within a pair of flip-dye hybridizations. Statistical filters are contained in MIDAS, part of the TM4 software package.

Comprehensive Gene Lists of differentially regulated genes (t-test and SAM generated data) are posted for the convenience of novice microarray users. The Comprehensive Gene Lists are aimed at the novice data miner and contain genes which may be responsible for cardiovascular traits. For each direct comparison (e.g. SS normoxic versus SS.13BN normoxic), 4 independent hybridizations (replicate dye swaps) were performed, normalized individually by LOWESS and then subjected to standard deviation regularization. We have used standard cut-off values to pre-filter the data (one sample t-test, p < 0.05 with additional criteria that the fold change be > |0.48| or |0.58|; SAM, FDR 10%). The more savy microarray user will want to download specific datasets of interest and perform investigator specific data analysis to generate their own list of cardiovascular candidates for subsequent analysis.

Home Data Index Rat Study Overview