University of New Hampshire University of New Hampshire Scholars' Repository

New Hampshire Agricultural Experiment Station Publications New Hampshire Agricultural Experiment Station

7-1-2013

Atherosclerosis-susceptible and atherosclerosis-resistant pigeon aortic cells express different in vivo

Janet L. Anderson University of New Hampshire, [email protected]

C. M. Ashwell University of New Hampshire - Main Campus

S. C. Smith University of New Hampshire - Main Campus

R. Shine University of New Hampshire - Main Campus

E. C. Smith University of New Hampshire - Main Campus

See next page for additional authors

Follow this and additional works at: https://scholars.unh.edu/nhaes

Part of the Poultry or Avian Science Commons

Recommended Citation J. L. Anderson, C. M. Ashwell, S. C. Smith, R. Shine, E. C. Smith and R. L. Taylor, Jr. Atherosclerosis- susceptible and atherosclerosis-resistant pigeon aortic cells express different genes in vivo Poultry Science (2013) 92 (10): 2668-2680 doi:10.3382/ps.2013-03306

This Article is brought to you for free and open access by the New Hampshire Agricultural Experiment Station at University of New Hampshire Scholars' Repository. It has been accepted for inclusion in New Hampshire Agricultural Experiment Station Publications by an authorized administrator of University of New Hampshire Scholars' Repository. For more information, please contact [email protected]. Authors Janet L. Anderson, C. M. Ashwell, S. C. Smith, R. Shine, E. C. Smith, and Robert L. Taylor Jr.

This article is available at University of New Hampshire Scholars' Repository: https://scholars.unh.edu/nhaes/207

Atherosclerosis-susceptible and atherosclerosis-resistant pigeon aortic cells express different genes in vivo

J. L. Anderson ,* C. M. Ashwell ,† S. C. Smith ,* R. Shine ,* E. C. Smith ,* and R. L. Taylor Jr. *1

* Department of Animal and Nutritional Sciences, University of New Hampshire, Durham 03824; and † Department of Poultry Science, North Carolina State University, Raleigh 27695

ABSTRACT Spontaneous atherosclerosis in the White NOT were exclusive to the SR-Ar. Microarray analysis Carneau (WC-As) pigeon is inherited as a single revealed 48 genes with differential expression. Vascular disorder, and its progression closely mirrors the human endothelial growth factor and p53 binding were disease. Representational difference analysis and micro- among the 17 genes upregulated in the WC-As. Thirty- Downloaded from array were used to identify genes that were differen- one genes were upregulated in the SR-Ar including the tially expressed between the susceptible WC-As and transforming growth factor-β signaling factor SMAD2 resistant Show Racer (SR-Ar) aortic tissue. The RNA and heat shock protein 90. Genes representing several extracted from 1-d-old squab aortas was used to make biochemical pathways were distinctly different between cDNA for each experiment. Fifty-six unique genes were breeds. The most striking divergences were in cytoskel- http://ps.oxfordjournals.org/ found using representational difference analysis, with 25 etal remodeling, proteasome activity, cellular respira- exclusively expressed in the WC-As, 15 exclusive to the tion, and immune response. cytoskeletal remod- SR-Ar, and 16 nonexclusive genes having copy number eling appears to be one of the first differences between variation between breeds. Caveolin and β-actin were susceptible and resistant breeds, lending support to the expressed in the WC-As, whereas the proteasome mat- smooth muscle cell phenotypic reversion hypothesis of uration protein and the transcription complex CCR4- human atherogenesis.

Key words: atherosclerosis , pigeon , aorta , Show Racer , White Carneau at University of New Hampshire Library on July 17, 2014 2013 Poultry Science 92 :2668–2680

http://dx.doi.org/ 10.3382/ps.2013-03306

INTRODUCTION level of the arterial wall. The Show Racer (SR-Ar) pi- geon is resistant to the development of atherosclerosis Investigation of human disease relies on animal mod- under identical diet and housing conditions, and with els that parallel specific human pathologies. One of the similar blood cholesterol levels (Clarkson et al., 1959). significant human diseases that relies heavily on models Crossbreeding and backcross experiments demonstrat- is atherosclerosis, especially at the earliest stages of the ed aortic atherosclerosis susceptibility to be inherited disease. Diverse animal models exist, each of them hav- in a pattern consistent with an autosomal recessive ing unique features appropriate for investigating differ- Mendelian trait (Smith et al., 2001). ent stages of atherosclerosis. The susceptible-resistant Atherosclerosis continues to be a major factor affect- pigeon (Columba livia) model (Anderson et al., 2012a) ing American mortality. According to the Centers for has been employed to understand genetic components Disease Control (Heron, 2012), diseases of the heart of this disease. White Carneau (WC-As) pigeons de- (#1) account for 25% of the total number of deaths, velop spontaneous atherosclerosis without known risk and cerebrovascular accidents (#4) account for ap- factors (Clarkson et al., 1959; Santerre et al., 1972). proximately 5%. Atherosclerosis contributes to both of The pigeon lesions (St. Clair, 1998; Moghadasian et these leading causes of death in the United States be- al., 2001) have greater similarities to human athero- cause, depending on the location of the affected artery, sclerosis than any other animal model of heart disease. the condition can promote heart disease, stroke, and St. Clair (1983) has reviewed multiple studies clearly kidney failure. Atherosclerosis most commonly occurs demonstrating that WC-As susceptibility resides at the along branch points of the arterial tree, such as the cor- onary and carotid arteries, and the celiac bifurcation of the aorta (Bassiouny et al., 1994; Kjaernes et al., 1981). © 2013 Poultry Science Association Inc. There is a strong familial component to the disease, Received May 10, 2013. Accepted July 1, 2013. and many risk factors that contribute to lesion progres- 1 Corresponding author: [email protected] sion and plaque stability have been identified. However,

2668 IN VIVO PIGEON ATHEROSCLEROSIS GENES 2669 the earliest events of atherogenesis remain unclear, and balanced salt solution and placed on an ice-cooled nu- approximately 50% of heart attacks occur in individu- clease-free surface where extraneous lipid, blood, and als displaying none of the classical risk factors (Ridker, connective tissue were removed. After blotting on filter 2000). Until the genetic predisposition to the disease is paper, the aorta was flash-frozen in liquid nitrogen and fully understood, prevention efforts will remain limited. placed in a −20°C freezer. Multiple hypotheses have been proposed to explain The aortas from each breed were pooled and placed atherosclerotic lesion initiation. Some of these hypoth- in a liquid-nitrogen chilled mortar. They were crushed eses have been based on work in animal models. Lipid with a pestle to a fine powder and homogenized with infiltration, response to retention, response to injury, TRIzol reagent (catalog 15596–026, Life Technolo- inflammation, monoclonal origin, smooth muscle cell gies, Grand Island, NY). The resulting slurries were (SMC) phenotypic reversion, and hemodynamics have transferred into 1.7 mL of nuclease-free microcentrifuge all been explored (Anderson et al., 2013). Many of tubes and centrifuged for 10 min at 16,000 × g at room these explain the preferred site of fatty streak forma- temperature. tion, some explain the accumulation of lipid in the arte- The supernatants were transferred to 1.5 mL of rial wall and the appearance of macrophage cells, but heavy phase Lock Gels (catalog 955154151, Eppendorf, Downloaded from none of the current hypotheses describe the complete Hauppauge, NY) followed by chloroform extraction. set of events that occurs in atherogenesis. Genetic fac- The RNA was mixed with cold isopropanol and incu- tors clearly influence cell proliferation rates (Lichter, bated overnight at −20°C. Following incubation, the 2000), the mitochondrial oxidative capacity for cellular tubes were centrifuged for 15 min at 16,000 × g at room lipids (Scheckhuber, 2005; Yu et al., 2012), cholesterol temperature to sediment the RNA. The supernatant http://ps.oxfordjournals.org/ metabolism (Ordovas and Shen, 2002), and the immune was decanted and the pellets rinsed with 1 mL of cold response (VanderLaan and Reardon, 2005; Hansson 70% ethanol. After decanting, the pellets received an and Libby, 2006), thereby manifesting an underlying additional rinse with 95% ethanol at −20°C. The etha- influence on all aspects of atherogenesis that warrants nol was again decanted, the pellets dried under vacuum additional investigation. and dissolved in 20 μL of distilled, deionized water. Genetic differences between aortic SMC in pigeon Total RNA was assessed in a 3.0 mM sodium phos- breeds were previously characterized in vitro and cat- phate, monobasic buffer at wavelengths ranging from egorized into 6 general metabolic themes (Anderson 220 to 320 nm. Ratios calculated from spectrophoto- at University of New Hampshire Library on July 17, 2014 et al., 2012b). Disparities in energy metabolism and metric readings revealed no protein (OD260/OD280) SMC phenotypic markers were the most remarkable. or polysaccharide (OD230/OD260) contamination. The However, due to the compressed time frame of cell de- RNA concentration was calculated using OD260. Two velopment in culture, it was impossible to distinguish micrograms of RNA from each breed were frozen at the chronology of events. The objective of the current −80°C for the RDA experiments. study was to determine genes that are differentially ex- pressed between aortic cells in atherosclerosis-suscepti- RDA Experiment ble (WC-As) and atherosclerosis-resistant (SR-Ar) 1-d- old squabs using representational difference analysis Total RNA was extracted and converted to double- (RDA) and microarray analysis. stranded cDNA using the BD SMART PCR cDNA Synthesis kit (K1052–1, Clontech, Mountain View, CA). Four cDNA samples were selected for 2 indepen- MATERIALS AND METHODS dent RDA comparisons. The RDA protocol followed Pastorian et al. (2000), including modifications of Ty- Birds son and Shanahan (2000) and Anderson et al. (2012b). Three rounds of PCR-coupled subtractive hybridiza- White Carneau and Show Racer pigeons were ob- tions were performed in reciprocal experiments. Final tained from the Palmetto Pigeon Plant in Sumter, difference products were cloned into the BamHI restric- South Carolina, in 1962. Both breeds have been kept tion site of the Stratagene pBluescript SK+ phagemid in fly coops at ambient temperatures and allowed free vector (212205, Agilent Technologies, Santa Clara, access to water, Purina Pigeon Chow Checkers, and CA) as previously described (Anderson et al., 2012b). Kaytee Red Grit. The closed flocks were maintained ac- Clones were sequenced by the Hubbard Center for cording to the UNH Animal Care and Use Committee Genome Studies (University of New Hampshire, Dur- protocol (approval #050601). ham) on a Beckman CEQ 8000 Genetic Analysis Cap- illary System. Raw sequence data were trimmed and RNA Extraction analyzed using the Basic Local Alignment Search Tool (BLAST) against the National Center for Biotechnol- Aortas were removed from four 1-d-old squabs from ogy Information (NCBI) nonredundant nucleotide da- each breed after carbon dioxide anesthesia and exsan- tabase (ntBLASTnr) exactly as described (Anderson et guination. The aortas were then rinsed in chilled Hanks al., 2012b). 2670 Anderson et al. Microarray Experiment ber was entered into the search feature on the NCBI webpage to recover a sequence that was then subjected An aliquot of the extracted RNA was shipped frozen to BLAST in the manner previously mentioned. to the Poultry Genomics Laboratory at North Carolina To integrate the individual differ- State University, Raleigh. Isolated RNA was quantified ences into networks and pathways that result in the using a ND-1000 spectrophotometer (NanoDrop Tech- altered phenotype of the WC-As, gene lists including nologie Inc., Wilmington, DE), and RNA integrity was both those identified by microarray and RDA were ana- verified by electrophoresis on 1.5% agarose gel. Equal lyzed using the GeneGo server from Thomson Reuters amounts of RNA extracted from samples acquired from (Thomson Reuters, London, UK). The GeneGo server’s 6 replicate animals on the same embryonic day were MetaCore database is a highly curated web-based ap- pooled and adjusted to 0.5 µg/µL of concentration, re- plication for identification of processes sulting in 4 biological replicates that were indirectly in input gene sets (Nikolsky et al., 2005). The analysis labeled, 2 with Cy3 and 2 with Cy5, according to the methodology uses annotation databases and identifies experimental design (Burnside et al., 2005). On each gene ontology biological processes associated with ex- array, a combination of 33 pmol of cDNA labeled with perimentally identified differentially expressed genes, Downloaded from different CyDyes were hybridized for 16 h at 42°C. which are then ranked according to their P-value. Gene Slides were scanned on a ScanArray GX PLUS Micro- pathways and networks are identified by an overrepre- array Scanner (PerkinElmer Life and Analytical Sci- sentation of pathway or network group members in the ences, Shelton, CT) set to 65% laser power. input gene list. MetaCore was also used for network The experimental design included a dye swap inter- analysis as a pathway mapping tool. The list of differ- http://ps.oxfordjournals.org/ woven loop (Garosi et al., 2005). Spot quantification entially expressed genes identified experimentally was was carried out using ScanAlyze Software (Eisen et al., uploaded to MetaCore for analysis and construction of 1998). Raw data files were joined, transformed to a log2 any potential biological network(s). The shortest path base, and analyzed in JMP Genomics (SAS Institute, algorithm was used to map the shortest path for in- 2007). Array data were analyzed using the normaliza- teraction between the differentially expressed genes tion method (Wolfinger et al., 2001) based on the over- (Supplemental Figure 1; available at http://dx.doi. all ANOVA method (Kerr et al., 2000). The residuals org/10.3382/ps.2013-03306).

were analyzed by mixed ANOVA in JMP Genomics at University of New Hampshire Library on July 17, 2014 (SAS Institute, 2007) according to the gene-specific model: RESULTS RDA Y = µ + Line + Dye + Hyb + slide + e, The difference products from 2 reciprocal RDA ex- where Y = hybridization intensity, μ = mean intensity, periments were cloned into 4 libraries. Libraries WC175 and e = random error, with genetic line (Line) and and WC180 represent genetic transcripts that are up- Cy-Dye (Dye) as fixed effects and hybridization batch regulated in 1-d-old WC-As pigeon aortic tissue rela- (Hyb) and Slide as random effects. Mean intensities tive to the SR-Ar. The SR-Ar libraries (SR175, SR180) were compared using false discovery rate at P < 0.01. from 1-d-old aortic tissue contain expressed sequence The array probes that were found to be significant- tags (EST) that are upregulated in that breed relative ly different between the lines were identified by the to the WC-As. Ninety-six clones from each library were array annotation database. Due to limited annota- randomly selected for sequencing. The BLAST analysis tion information included in the array database, each of difference products is presented in Table 1. corresponding The Institute for Genomic Research Three hundred eighty-five EST were found in the 4 li- (TIGR) tentative consensus (TC) number was then braries. This total represents some clones that captured entered into the “Identifiers or Keywords” search fea- multiple difference products and a few that contained ture on the DFCI Chicken Gene Index webpage, edi- none. Two hundred sixty-four (69%) of these EST were tion 11.0, updated June 17, 2006 (http://compbio.dfci. identified. After examining these sequences for du- harvard.edu/tgi/tgipage.html). For some array spots, plicity, pigeon aortic tissue at 1 d of age expressed 62 an annotation was included in the information for the unique transcripts. Twenty-five genes were exclusively corresponding TC number, which was used. Otherwise expressed by the WC-As (Table 2), including β-actin the sequence that corresponded to the TC number was (ACTB), caveolin (CAV1), and 2 types of collagen placed into the BLASTn queue on the NCBI website (COL4A2, COL5A2). Fifteen genes were exclusive to using the nonredundant nucleotide (nr) database. A list the SR-Ar (Table 2) including CCR4-NOT transcrip- of matches was provided by the program, and the top tion complex (CNOT2) and proteasome maturation hit was taken as the annotation if the e-value was equal protein (POMP). We found 22 genes that were ex- to or less than e−10. pressed by both breeds, with 16 demonstrating copy If there was no TIGR TC number matching the spot number (CN) variation, and 6 that had equivalent CN on the microarray chip, the GenBank number was then in each breed (Table 3). The largest differential expres- used to determine an annotation. The GenBank num- sion for transcripts found in both breeds was NADH IN VIVO PIGEON ATHEROSCLEROSIS GENES 2671 Table 1. Representational difference analysis (RDA) products from White Carneau (WC-As) and Show Racer (SR-Ar) pigeon 1-d- old aortic cells in vivo

Identified by NCBI BLAST2 Unknown (novel) No. of RDA Library EST1 GenBank accession no. No. of EST % DP3 No. of EST % DP3 1 WC175 94 FG589856-FG589914 72 77 22 23 FL593203-FL593217 FL633614-FL633636 SR175 96 FG589807-FG589855 80 83 16 17 FL633566-FL633598 FL593143-FL593156 2 WC180 96 FG598345-FG598404 63 66 33 34 FL633637-FL633649 FL593218-FL593240 SR180 99 FG598307-FG598344 49 49 50 51 FL633599-FL633613

FL593157-FL593202 Downloaded from Total 385 264 69 121 31 1EST = expressed sequence tags. 2NCBI BLAST = National Center for Biotechnology Information Basic Local Alignment Search Tool. 3% difference products. http://ps.oxfordjournals.org/ dehydrogenase subunit 1 (ND1) in the SR-Ar and Lu- The most inclusive network, which contained the mican (LUM) in the WC-As. largest number of differentially expressed transcripts, The 56 unique genes with differential expression is depicted in Figure 1. Most of the transcripts were were subjected to pathway analysis using MetaCore upregulated in the WC-As including insulin growth (GeneGo Inc., Carlsbad, CA) to assess their collective factor (IGF1), fibulin 5 (FBLN5), cytoskeletal actin function and the relationship to atherosclerotic sus- (ACTB), and capping protein (CAPZA1). A sec- ceptibility/resistance in the pigeon. Of the pathways ond network revolves around the transcription factor at University of New Hampshire Library on July 17, 2014 presented in Table 4, remodeling, oxida- peroxisome proliferator-activated receptor (PPAR)-γ tive phosphorylation, and cell adhesion were the most (Figure 2). This network contained 7 genes that were significant. Three cytoskeleton remodeling pathways upregulated in the SR-Ar, including CNOT2, were represented in the WC-As by (DSTN), (SEPT2), and ACTA2. The WC-As upregulated genes ACTB, CAV1, COL4A2, and light chain ki- were ACTB, a direct target of PPAR-gamma, and nase (MYLK). Eukaryotic translation initiation fac- HSPA14. tor (EIF4A) was found in both breeds, but the CN was higher in the SR-Ar. Oxidative phosphorylation Microarray Analysis was the next most significant pathway, indicated by the upregulation of cytochrome oxidase II (COII), cy- Forty-eight genes were found to be differentially ex- tochrome b (CYTB), as well as NADH dehydrogenase pressed by microarray analysis. Seventeen of these were subunits 1 and 2 (ND1, ND2) in the SR-Ar aortas. upregulated in the WC-As and 31 were upregulated Two pathways were categorized as cell adhesion and in the SR-Ar (Table 5). In the WC-As, genes dem- include CAV1, ACTB, and COL4A2, all upregulated onstrating the greatest fold change include hemoglo- in the WC-As. bin β chain (HBB) and p53 binding protein (mdm1), Gene ontology (GO) processes were also ranked for whereas the same criterion in the SR-Ar highlighted significance (Table 4) by MetaCore analysis. Genetic dif- heat shock protein 90 (HSP90), acyl CoA dehydroge- ferences between the 2 breeds were apparent in system nase (ACAD8), and TGFB signaling factor SMAD2 development, blood circulation, and vascular smooth (MADH2), a transforming growth factor-β (TGFB) muscle contraction. Four processes focused on devel- transcription factor. opment and included EIF4A, α actin (ACTA2) and Putative pathways as scored by MetaCore indicate septin (SEPT2) in the SR-Ar, and COL5A2, ACTB, significant differences in the immune response, regu- and heat shock protein 70 (HSPA14) in the WC-As. lation of lipid metabolism, and cytoskeleton remodel- Circulation and vascular smooth muscle contraction ing (Table 6). The immune response is suggested by were represented by 2 processes each. The breeds dem- the MHC class I (MHC-1) and HSP90 expression in onstrated differences in actin isoforms with ACTB ex- the SR-Ar. Sterol regulatory element binding protein clusively expressed in the WC-As and ACTA2 upregu- 2 (SREBP2) was also upregulated in the SR-Ar, sug- lated in the SR-Ar. Network analysis was conducted to gesting a potential difference in overall lipid metabo- compile individual pathways and identify common foci lism between the 2 breeds. The GO processes mostly because, as expected, the data set indicated multiple reflected an immune response, with MHC-1 represent- processes and pathways. ing antigen presenting, T cell tolerance induction, and 2672 Anderson et al. Table 2. Transcripts expressed exclusively in representational difference analysis from atherosclerosis-susceptible White Carneau (WC-As) or atherosclerosis-resistant Show Racer (SR-Ar) 1-d-old aortic cells in vivo

Copy no.

Gene Gene name WC SR P-value Expressed exclusively in WC-As aortic tissue ACTB β-Actin 3 0 0.0886 CAV1 Caveolin-1 3 0 0.0886 MAP4K3 Mitogen activated protein kinase kinase kinase kinase 3 2 0 0.1652 BRE Brain and reproductive expressed/TNFRSF1A modulator 1 0 0.3274 CAPZA1 Capping protein (actin filament) muscle Z-line, α 1 1 0 0.3274 COL4A2 Collagen, type IV α 2 1 0 0.3274 COL5A2 Collagen, type V α 2 1 0 0.3274 COPB2 Coatomer protein complex, subunit β 2 (β prime) 1 0 0.3274 CYR61 Cysteine-rich, angiogenic inducer, 61 1 0 0.3274 DSTN Destrin (actin depolymerizing factor) 1 0 0.3274 DYNC2H1 , cytoplasmic 2, heavy chain 1 1 0 0.3274

HMGB3 High mobility group box 3 1 0 0.3274 Downloaded from HSPA14 Heat shock 70 kDa protein 14 1 0 0.3274 HTRA1 HtrA serine peptidase 1 1 0 0.3274 IGF1 Insulin-like growth factor 1 (somatomedin C) 1 0 0.3274 MAP1B -associated protein 1B 1 0 0.3274 MTA1 Metastasis-associated protein 1 1 0 0.3274 NOV Nephroblastoma overexpressed gene 1 0 0.3274

PFKP Phosphofructokinase, platelet 1 0 0.3274 http://ps.oxfordjournals.org/ RPL3 Ribosomal protein L3 1 0 0.3274 RPN2 Ribophorin II 1 0 0.3274 SCYE1 Aminoacyl tRNA synthetase complex-interacting multifunctional protein 1 1 0 0.3274 SPON1 Spondin 1, extracellular matrix protein 1 0 0.3274 TMEM167A Transmembrane protein 167A 1 0 0.3274 TPM1 1 (α) 1 0 0.3274 Total 30 0 Expressed exclusively in SR-Ar aortic tissue CNOT2 CCR4-NOT transcription complex, subunit 2 0 2 0.1464

POMP Proteasome maturation protein 0 2 0.1464 at University of New Hampshire Library on July 17, 2014 TKT Transketolase 0 2 0.1464 ATP1A1 ATPase, Na+/K+ transporting, α 1 polypeptide 0 1 0.3054 CWC15 CWC15 spliceosome-associated protein homolog (Saccharomyces cerevisiae) 0 1 0.3054 DCN Decorin 0 1 0.3054 DHX9 DDX9 ATP-dependent RNA helicase A 0 1 0.3054 DRG2 Developmentally regulated GTP binding protein 2 0 1 0.3054 MED17 Mediator complex subunit 17 0 1 0.3054 ND2 NADH dehydrogenase subunit 2 0 1 0.3054 PHB2 Prohibitin 2 0 1 0.3054 POSTN Periostin, osteoblast specific factor 0 1 0.3054 RUNX1T1 Runt-related transcription factor 1; translocated to, 1 (cyclin D-related) 0 1 0.3054 SEPT2 Septin 2 0 1 0.3054 TAF1 TAF1 RNA polymerase II, TATA box binding protein-associated factor, 250 kDa 0 1 0.3054 Total 0 18 along with purinergic receptor P2X (P2RX7), regula- DISCUSSION tion of T cell-mediated cytotoxicity in the SR-Ar. As with the RDA experiment, further analysis was RDA Experiment conducted to determine the potential networks func- tioning in the WC-As and SR-Ar aortas that may The genes ACTB, CAV1 (Table 2), and LUM (Ta- connect expressed transcripts. The most inclusive net- ble 3) were upregulated in the 1-d-old WC-As relative work contained 10 differentially expressed genes, most to the SR-Ar based on RDA. In the SR-Ar, CNOT2, of which were from the SR-Ar (Figure 3). Represen- POMP (Table 2), and ND1 (Table 3) genes were up- tative transcripts include MHC-1, solute carrier pro- regulated. Table 4 shows the biological pathways and tein (SLC25A6), and DEAD box protein polypeptide processes based on the combined list of upregulated (DDX1). In the WC-As, nucleobindin (NUCB2) and WC-As and SR-Ar genes. The upregulated genes in Ras p21 protein activator 2 (RASA2) were represent- WC-As were indicative of cytoskeletal remodeling and ed in the network. Another relevant network as scored cell adhesion (Table 4), whereas those genes in SR-Ar by GeneGo is shown in Figure 4. In this network, the were related to cytoskeletal remodeling and oxidative transcription factor SREBP2, which plays a regulatory phosphorylation (Table 4). role in lipid metabolism, was upregulated in the SR-Ar. Cytoskeletal remodeling was occurring in each breed, Vascular endothelial growth factor (VEGF-A) was el- but the specific activity was different. Genes for ACTB, evated in the WC-As. CAV1, DSTN, COL4A2, and COL5A2 were expressed IN VIVO PIGEON ATHEROSCLEROSIS GENES 2673 Table 3. Transcripts expressed in representational difference analysis from both atherosclerosis-susceptible White Carneau (WC-As) or atherosclerosis-resistant Show Racer (SR-Ar) 1-d-old aortic cells in vivo

Copy no.

Gene1 Gene name WC SR P-value Copy number differs significantly by chi-squared analysis ND1 NADH dehydrogenase subunit 1 5 13 0.0040 LUM Lumican, keratan sulfate proteoglycan 16 5 0.0167 Copy number does not differ significantly by chi-squared analysis 16S rRNA 16S ribosomal RNA 4 9 0.1319 RBP7 Retinol binding protein 7, cellular 1 4 0.1597 SLC25A6 Solute carrier family 25 (mitochondrial carrier; adenine nucleotide translocator), member 6 4 1 0.1924 LDHB Lactate dehydrogenase B 10 5 0.2153 ANXA1 Annexin-1; calcium-dependent membrane binding protein annexin I 1 3 0.2920 EIF4A2 Eukaryotic translation initiation factor 4A2 2 4 0.3775 ACTA2 Actin, α 2, smooth muscle, aorta 7 10 0.3957 CYTB Cytochrome b 3 5 0.4333

COII Cytochrome oxidase subunit II 10 13 0.4419 Downloaded from MYLK Smooth muscle kinase (61-kDa active fragment); telokin 2 1 0.5884 PRDX1 Peroxiredoxin 1 2 3 0.6150 FBLN5 Fibulin 5 4 3 0.7473 ATP6 ATP synthase F0 subunit 6 5 4 0.7873 COI Cytochrome oxidase subunit I 7 6 0.8441 ND4 NADH dehydrogenase subunit 4 9 9 0.9204

MYH11 Myosin, heavy chain 11, smooth muscle 2 2 0.9635 http://ps.oxfordjournals.org/ EIF1 Eukaryotic translation initiation factor 1 1 1 0.9743 RPS3A Ribosomal protein S3A 1 1 0.9743 UQCRFSL1 Ubiquinol-cytochrome c reductase, Rieske iron-sulfur polypeptide 1 1 1 0.9743 VIM 1 1 0.9743 Total 98 104 1Total of 22 genes.

in the WC-As, whereas ACTA2 and EIF4A2 were up- Changes in actin expression are of potential impor- at University of New Hampshire Library on July 17, 2014 regulated in the SR-Ar. Figure 1 presents further evi- tance to atherogenesis because ACTB and ACTA2 are dence of cytoskeletal remodeling through a relationship associated with a synthetic and contractile phenotype, between the transcription factor AP-1, the actin iso- respectively (Yamin and Morgan, 2012). At 1 d of age, forms, and IGF-1. Figure 2 shows that ACTB is also the synthetic phenotype is expected to dominate be- a target of PPAR-gamma, whereas ACTA2 expression cause both breeds are still in the developmental stage. is modulated by CNOT2 via the transcription factor In the current experiment, the nearly equal expression NF-AT. of fibulin 5 (FIBLN5), a transcript involved in arterial

Table 4. Pathway maps and gene ontology (GO) processes based on all representational difference analysis (RDA) transcripts having significant expression differences in both atherosclerosis-susceptible White Carneau and atherosclerosis-resistant Show Racer 1-d-old aortic cells in vivo

Item P-value GeneGo pathway map Cytoskeleton remodeling_Cytoskeleton remodeling 2.07E-07 Cytoskeleton remodeling_TGF, WNT and cytoskeletal remodeling 3.7E-07 Oxidative phosphorylation 4.93E-06 Development_TGF-β-dependent induction of EMT via RhoA, PI3K and ILK 4.19E-05 Cell adhesion_Integrin-mediated cell adhesion and migration 4.96E-05 Cytoskeleton remodeling_Regulation of actin cytoskeleton by Rho 0.000123 Development_Slit-Robo signaling 0.000276 Cell adhesion_Chemokines and adhesion 0.000852 Neurophysiological process_Receptor-mediated axon growth repulsion 0.000924 Chemotaxis_Inhibitory action of lipoxins on IL-8- and leukotriene B4-induced neutrophil migration 0.001333 GO process System development 1.48E-10 Developmental process 4.1E-10 Blood circulation 5.33E-10 Circulatory system process 5.92E-10 Anatomical structure development 8.61E-10 Vasoconstriction 1.56E-09 Vascular smooth muscle contraction 1.9E-09 Regulation of anatomical structure size 1.94E-09 Organ development 2.1E-09 Regulation of blood pressure 4.3E-09 2674 Anderson et al. Downloaded from http://ps.oxfordjournals.org/

Figure 1. Most inclusive network (GeneGo, Carlsbad, CA) incorporating genes upregulated in representational difference analysis from atherosclerosis-susceptible White Carneau and atherosclerosis-resistant Show Racer 1-d-old aortic cells in vivo. Abbreviations listed in http:// www.ncbi.nlm.nih.gov/genbank/. Color version available in the online PDF. at University of New Hampshire Library on July 17, 2014 development, demonstrated the synthetic phenotype. are consistent with the in vitro transcript profile ana- However, subtle actin profile changes are already appar- lyzed by Anderson et al. (2012b) who found ACAT2 ent, with differing ACTA2 expression between breeds, was expressed exclusively in the SR-Ar, and ACTB was and the absence of ACTB in the SR-Ar. These findings significantly upregulated in the WC-As. Likewise, in

Figure 2. Significant peroxisome proliferator-activated receptor-γ network (GeneGo, Carlsbad, CA) incorporating genes upregulated in rep- resentational difference analysis from atherosclerosis-susceptible White Carneau and atherosclerosis-resistant Show Racer 1-d-old aortic cells in vivo. Abbreviations listed in http://www.ncbi.nlm.nih.gov/genbank/. Color version available in the online PDF. IN VIVO PIGEON ATHEROSCLEROSIS GENES 2675 Table 5. Microarray analysis of genes upregulated in atherosclerosis-susceptible White Carneau (WC-As) or atherosclerosis-resistant Show Racer (SR-Ar) 1-d-old aortic cells in vivo

Fold Gene Name change P-value Genes upregulated in WC-As aortic tissue by microarray GPR107 G-coupled protein receptor 107 (integral to membrane) 0.0647 0.0022 GAP1m aka RASA2 - RAS p21 protein activator 2 (GTPase activator) 0.0735 0.0031 DRB1 Major histocompatibility complex (MHC) II (DR β 1)/human leukocyte antigen (HLA)- 0.1030 0.0028 DRB MRPL28 Mitochondrial ribosomal protein L28 (melanoma antigen p15) TF 0.1563 0.0005 FBXL10 F box and leucine rich repeat protein 10 0.1623 0.0012 SQSTM1 Sequestosome (oxidative stressed induced, Zn ion binding) 0.1703 0.0023 PSMC3IP PSMC3 (26S proteasome ATPase subunit 3) interacting protein 0.2027 0.0029 NUCB2 Nucleobindin 2 0.2107 0.0019 STRAP Serine/threonine kinase receptor associated protein 0.2566 0.0010 TTLL7 tyrosine -like family 7 0.2831 0.0014 PET56 Mitochondrial ribosomal RNA methylase 0.4399 0.0000 VEGFA Vascular endothelial growth factor A 0.4691 0.0007 Downloaded from MARCKS Myristoylated alanine rich C kinase 0.5150 0.0026 GCAT Glycine C-acetyl 0.8180 0.0031 SERPINF2 Serpin peptidase inhibitor Clade F 0.8749 0.0027 mdm1 p53 binding protein 0.9629 0.0012 HBB Hemoglobin, β chain 1.1986 0.0017 Genes upregulated in SR-Ar aortic tissue

HSP90 Heat shock protein 90 (β?) −1.2563 0.0021 http://ps.oxfordjournals.org/ ACAD8 Acyl CoA dehydrogenase, member 8 −1.2273 0.0025 SP90 Heat shock protein 90 (α?) −1.1471 0.0027 MADH2 SMAD2 (TGFB signaling, TF) −1.0016 0.0008 ETNK1 Ethanolamine kinase −0.9802 0.0002 CYB5R3 Cytochrome b5 reductase/diphorase −0.8575 0.0015 HLA-C MHC I −0.6656 0.0029 AKR1 Palmitoyl transferase (palmitoyls ) −0.6047 0.0012 PNN Pinin, desmosome associated protein −0.6038 0.0023 SREBP2 Sterol regulatory binding protein 2 −0.5785 0.0027

SEPW1 Selenoprotein W, 1 (translation termination, pseudogene?) −0.4644 0.0005 at University of New Hampshire Library on July 17, 2014 BDH1 3-Hydroxy butarate dehydrogenase type I −0.4638 0.0002 P2RX7 Purinergic receptor P2X, ligand gated ion channel 7 −0.4110 0.0028 CMPK1 Cytidylate kinase (pyrimidine synthesis reaction) −0.3291 0.0011 DDX1 DEAD box protein polypeptide 1 (RNA helicase) −0.3203 0.0012 SLC25A6 aka ANT3 (adenine nucleotide transporter 3) −0.3178 0.0025 RhoGAPX Rho GTPase activating protein 6 −0.2982 0.0016 EDARADD Ectodysplasin A1 receptor associated death domain −0.2967 0.0027 AP180 Clathrin assembly protein −0.2892 0.0013 MMS19 Nucleotide excision repair homolog (aka MET18) −0.2718 0.0014 RNF219 Ring finger protein 219 (C13orf7) −0.2684 0.0018 CYP39A1 Cytochrome p450 family 39, subfamily a, peptide 1 −0.2565 0.0018 cyp1b1 Cytochrome p450 family 1, subfamily b, peptide 1 −0.2340 0.0027 TPP2 Tripeptidyl peptidase II −0.2039 0.0030 OSBPL2 Oxysterol binding protein like-2 −0.1912 0.0020 ARID1A AT Rich interactive domain 1a/actin dependent chromatin regulation −0.1776 0.0023 CSPP1 Centrosome and spindle pole association protein 1 −0.1591 0.0016 GRPEL1 Mitochondrial Grpe-like 1 (binds to HSP70 protein folding) −0.1555 0.0026 CTCF CCCTC binding factor (zinc finger protein) −0.1458 0.0018 ASPM Abnormal spindle homology, microcephaly associated −0.1309 0.0001 KIAA0368 Proteasome accessory protein −0.0559 0.0006 the present study ACTB was exclusive to the WC-As, lent to aortic cells in vivo at approximately 12 wk of and ACTA2 had CN variation at 10:7, although that age (Cooke and Smith, 1968). Therefore, cells analyzed value was not significant. These results suggest that in previous in vitro studies are actually older than in the SMC phenotypic reversion hypothesis (Gomez and the tissue. The proteoglycan profile may be different as Owens, 2012) of atherogenesis may be relevant in the the phenotype switches, with lumican being expressed pigeon model. early (1 d) in the WC-As followed by subsequent down- In terms of the cell adhesion pathway operating in regulation. Further experiments in vivo are required to the WC-As, CAV1, and LUM expression could relate test the lumican expression profile during WC-As de- to the response to retention hypothesis of human ath- velopment. erogenesis. In vitro, lumican was expressed primarily Previous aortic cell results have associated oxidative in the SR-Ar, although it was detected in the WC-As. phosphorylation with SR-Ar and glycolysis with the Direct age comparisons between aortic cells in vitro WC-As (Anderson et al., 2012b). The current experi- and in vivo are complex. Prior studies indicate that ment demonstrates mitochondrial differences between 1-d-old aortic cells cultured in vitro for 7 d are equiva- breeds in vivo at 1 d of age, although the anaerobic 2676 Anderson et al. Table 6. Pathway maps and gene ontology (GO) processes based on all microarray transcripts having significant expression differ- ences in both atherosclerosis-susceptible White Carneau (WC-As) and atherosclerosis-resistant Show Racer (SR-Ar) 1-d-old aortic cells in vivo

Item P-value GeneGo pathway map Immune response_Antigen presentation by MHC class I 3.714E-05 Regulation of lipid metabolism_Regulation of lipid metabolism via LXR, NF-Y, and SREBP 9.418E-05 Mechanisms of CFTR activation by S-nitrosoglutathione [normal and cystic fibrosis transmembrane receptor (CF)] 1.674E-04 Development_Role of IL-8 in angiogenesis 3.341E-04 CFTR folding and maturation (norm and CF) 4.882E-04 Regulation of lipid metabolism_Regulation of acetyl-CoA carboxylase 1 activity in keratinocytes 4.001E-02 wtCFTR and delta508 traffic/clathrin coated vesicles formation (norm and CF) 4.462E-02 Cytoskeleton remodeling_Role of Activin A in cytoskeleton remodeling 4.691E-02 Development_Glucocorticoid receptor signaling 5.604E-02 Apoptosis and survival_Apoptotic Activin A signaling 5.831E-02 GO process Positive regulation of tolerance induction to nonself antigen 1.8553E-08

Regulation of tolerance induction to nonself antigen 1.8553E-08 Downloaded from Positive regulation of tolerance induction dependent upon immune response 4.6328E-08 Regulation of tolerance induction dependent upon immune response 4.6328E-08 T cell tolerance induction 4.6328E-08 Antigen processing and presentation of endogenous peptide antigen via MHC class I via endoplasmic reticulum (ER) pathway 3.8732E-07 Antigen processing and presentation of endogenous peptide antigen via MHC class I via ER pathway, 3.8732E-07 TAP-dependent

Positive regulation of T cell mediated cytotoxicity 4.1521E-07 http://ps.oxfordjournals.org/ Antigen processing and presentation of exogenous peptide antigen via MHC class I 5.5265E-07 Regulation of T cell mediated cytotoxicity 7.8319E-07 at University of New Hampshire Library on July 17, 2014

Figure 3. Most inclusive network (GeneGo, Carlsbad, CA) incorporating genes upregulated in microarray analysis from atherosclerosis- susceptible White Carneau and atherosclerosis-resistant Show Racer 1-d-old aortic cells in vivo. Abbreviations listed in http://www.ncbi.nlm.nih. gov/genbank/. Color version available in the online PDF. IN VIVO PIGEON ATHEROSCLEROSIS GENES 2677 Downloaded from http://ps.oxfordjournals.org/

Figure 4. Significant sterol regulatory element binding protein 2 network (GeneGo, Carlsbad, CA) incorporating genes upregulated in microar- at University of New Hampshire Library on July 17, 2014 ray analysis from atherosclerosis-susceptible White Carneau and atherosclerosis-resistant Show Racer 1-d-old aortic cells in vivo. Abbreviations listed in http://www.ncbi.nlm.nih.gov/genbank/. Color version available in the online PDF. pathway is not yet apparent. In addition to energy pro- higher glycosylated hemoglobin is a known human ath- duction differences, each breed expressed diverse pro- erosclerosis risk factor, particularly in diabetics (Bower teasome subunits. This proteasome variation was high- et al., 2012), a strong connection has not been reported lighted by SR-Ar upregulation of POMP, a maturation between free hemoglobin levels and atherosclerosis. In- protein, in the current study as well as previous in vitro creased hemoglobin has been related to oxidative stress detection of regulatory subunit (PSMD1; Anderson et (Alayash et al., 2001), which itself is a contributing fac- al., 2012b). The WC-As microarray samples expressed tor to atherosclerosis. Studies in transgenic mice have PSMC3 26S proteasome ATPase subunit 3 interacting shown that the α and β hemoglobin chain contribute protein (PSMC3IP), whereas PSMC3, the 26S prote- to the proinflammatory nature of high-density lipo- asome ATPase subunit 3, was found in vitro (Anderson protein and may be considered as a novel biomarker et al., 2012b). for atherosclerosis in humans (Watanabe et al., 2007), Representational difference analysis is a sensitive which warrants further study. In contrast, the finding technique with the capability of detecting gene expres- of mdm1, a p53 binding protein in the WC-As, has sion differences at less than one copy per cell (Sung et potential interest to the atherogenic phenotype. The al., 2005). The technique does not lend itself to high- p53 tumor suppressor gene promotes cell cycle arrest throughput profiling (Lucito et al., 2003). In contrast, and apoptosis. Previous experiments have found that the main power of microarray analysis is the simulta- mdm2 and mdm4 inhibit p53 expression and are there- neous quantification of thousands of genetic sequences fore oncogenic (Perry, 2010). The same has not been (Ding and Cantor, 2004). Therefore, microarray anal- found for mdm1, but the influence of its binding needs ysis was conducted as a complementary approach to to be investigated in light of its presence in the suscep- maximize the identification of genetic differences be- tible breed. If mdm1 also inhibits p53, it could promote tween the 2 pigeon breeds. the survival of an aberrant cell. Anti-apoptosis activity in the WC-As was reported in vitro (Anderson et al., Microarray Experiment 2012b) and may be an important factor early in ath- erogenesis. In the microarray analysis, HBB and mdm1 demon- In the SR-Ar, HSP90, acyl CoA dehydrogenase strated the greatest fold change in the WC-As. The dif- (ACAD8), and a transcription factor for TGFB, ferential HBB expression is surprising because although SMAD2, demonstrated the greatest differential expres- 2678 Anderson et al. sion (Table 5). Pathways were identified based on all of the current experiments, but also in vitro, suggesting significant genes from both breeds (Table 6). The data its importance in pigeon atherogenesis. Reorganization set contained almost twice the number of significant of the cytoskeleton occurs before foam cell development genes in SR-Ar compared with the WC-As (Table 5), and cellular proliferation (Hansen, 1977; Owens, 1996), which gave the former group greater influence but like- and thus may be one of the earliest events (Gomez and ly masked any defective WC-As pathways. Biological Owens, 2012) that denote a susceptible WC-As pheno- pathways operating in the SR-Ar included an immune type. response, regulation of lipid metabolism, and cytoskel- Cytoskeletal remodeling is a broad category, en- etal remodeling (specifically the role of activin A). The compassing multiple pathways, many of which include immune response was the most significant pathway, TGFB activity. The same topic was previously catego- represented by the major histocompatibility complex rized (Anderson et al., 2012b) as cytoskeletal reorgani- I (HLA-C) and HSP90. Figure 3 shows that HLA-C zation and regulation and placed in the SMC pheno- directly influences the expression of the STAT1 tran- type theme. The cell signaling pathways theme included scription factor. Disruptions in the JAK/STAT path- TGFB. Examining the reported genetic and proteomic way have been observed in atherosclerosis models (Guo differences between breeds (Anderson et al., 2012b), cy- Downloaded from et al., 2006), and it is interesting to note that ACTB is toskeletal remodeling is frequently in the most relevant also active in this pathway. The immune response was biological pathways by our analysis methods. The com- also identified in vitro, but the stimulus for this path- bined data sets have 4 subcategories, each found in at way remains unclear. least 2 separate experiments (Table 7). The TGFB is Regulation of lipid metabolism in the SR-Ar is in- related to cytoskeletal development, which our analy- http://ps.oxfordjournals.org/ dicated by the expression of SREBP2, a transcription sis divides into 2 subcategories (Table 7). The TGFB factor. Cholesterol levels are not different between signaling was differentially expressed in vitro, both at breeds (Wagner, 1978), but an increase in nonesterified the gene level (Activin A; Anderson et al., 2012b) and fatty acids (NEFA) has been observed in the WC-As protein level (Activin A Binding protein; Smith et al., as early as 1 d of age (Nicolosi et al., 1972; Hajjar et 2008). In the microarray, the TGFB pathway is repre- al., 1980). The fact that SREBP2 is upregulated in the sented in the SR-Ar by MADH2, a TGFB transcription SR-Ar suggests a regulatory function inhibiting NEFA factor. accumulation in aortic tissue. The ACAD8 expression In addition to cytoskeletal remodeling and TGFB at University of New Hampshire Library on July 17, 2014 would also prevent NEFA accumulation. As shown in signaling, the proteasome and heat shock proteins dem- Figure 4, SREBP2 appears to inhibit c-Src, a focal onstrate consistent differential expression in multiple point in this network, whereas VEGF (WC-As) ap- experiments. In the SR-Ar, KIAA0368, a proteasome pears to enhance the expression of c-Src. The potential accessory protein was upregulated on the microar- relationship of the proto-oncogene c-Src expression to ray, and POMP was up in the RDA. In the WC-As, atherogenesis has not been studied, but could lead to PSMC3IP8 was found on the microarray. This is consis- smooth muscle cell proliferation observed before lipid tent with in vitro experiments, where both PSMC3 and accumulation. PSMC2 were upregulated that breed. Finally, HSP70 (WC-As) was upregulated in the current RDA experi- Comparison of Experiments ment, and in the proteomics experiment (Smith et al., 2008). This HSP is expressed in human atherosclerotic Because the gene lists from the RDA and the micro- lesions and is known to recruit cytokines (Xu et al., array experiment are so dissimilar, they are best com- 2012). In the SR-Ar, HSP90 was upregulated on the pared at the pathway level. Cytoskeletal remodeling microarray, and HSP40 was found in vitro. The HSP90 was a recurring process, as genes associated with this has been shown to decrease proliferation and migration pathway were differentially expressed not only in each of various cell types and “could have a protective role in

Table 7. Cytoskeletal remodeling and transforming growth factor-β (TGFB) signaling pathways identified in multiple experiments using representational difference analysis (RDA) or microarray of atherosclerosis-susceptible White Carneau or atherosclerosis-resis- tant Show Racer pigeon aortic cells

RDA Microarray in vivo 1 d in vivo 1 d RDA in vitro Item (this study) (this study) (Anderson et al., 2012b) Cytoskeletal remodeling pathways TGF, WNT, and cytoskeletal remodeling X X Cytoskeletal remodeling X X Regulation of actin cytoskeleton by Rho GTPase X X Role of activin A in cytoskeleton remodeling X X TGFB signaling pathways TGFB dependent-induction of EMT via RHOA X X TGFB dependent-induction of SMADS X X IN VIVO PIGEON ATHEROSCLEROSIS GENES 2679 atherosclerosis” (Xu et al., 2012). Although HSP par- pothesis. In addition, the recent publication of the pi- ticipate in a variety of cellular functions, the consistent geon genome (Shapiro et al., 2013) provides a reference differential expression across the experiments may be sequence. This genomic resource will allow application relevant to pigeon atherogenesis. of RNA-Seq (Marioni et al., 2008) to investigate the Although RDA and microarray analyses revealed transcriptome of each breed. The RNA-Seq combines similar metabolic pathways, the question of the inde- the open discovery advantage of RDA with the high- pendent sets of differentially expressed genes arises. As throughput capacity of a microarray. an open system (Tyson et al., 2002), RDA allows for gene discovery because it is not limited to sequences printed on a microarray. The most convenient explana- ACKNOWLEDGMENTS tion for finding a gene in RDA and not by microarray is that the gene is not on the chip. Another explanation This research was supported by National Institutes of may be that the microarray-imprinted EST may have Health (NIH) #1R15HL072786-01 Candidate Gene(s) a different sequence possibly through splice variation for Pigeon Atherosclerosis. Partial funding was pro- such that the RDA transcript would not hybridize. vided by the New Hampshire Agricultural Experiment Downloaded from Andersson et al. (2001) compared data derived from Station (Durham, NH). This is Scientific Contribution RDA and microarray experiments and found that “45% Number 2513. of genes isolated by RDA more than once in random sequencing were confirmed in the microarray analy- REFERENCES sis while 44% were below the signal threshold.” They http://ps.oxfordjournals.org/ concluded that low abundant transcripts isolated us- Alayash, A. I., R. P. Patel, and R. E. Cashon. 2001. Redox reactions ing RDA are not detected by microarray. Linder et al. of hemoglobin and myoglobin: Biological and toxicological impli- (2004) reached the same conclusion, stating that the cations. Antioxid. Redox Signal. 3:313–327. Anderson, J. L., S. C. Smith, and R. L. Taylor Jr. 2012a. Spontane- “lower sensitivity of microarray may fail to verify differ- ous atherosclerosis in pigeons: A good model of human disease. ential transcript expression.” Therefore the 2 methods Pages 25–48 in Atherogenesis. S. Parthasarathy, ed. InTech Inc., should not be used to validate, but to complement each Rijeka, Croatia. Accessed Jan. 11, 2012. http://www.intechopen. com/articles/show/title/spontaneous-atherosclerosis-in-pigeons- other as investigative tools. The results of the present a-good-model-of-human-disease. study appear to support this view. Indeed, although Anderson, J. L., S. C. Smith, and R. L. Taylor Jr. 2013. Athero- at University of New Hampshire Library on July 17, 2014 RDA may be superior for “identifying rare transcripts sclerosis-susceptible and atherosclerosis-resistant pigeon aortic with large differential expression, it may fail to detect smooth muscle cells express different genes and proteins in vitro. Pages 165–186 in Current Trends in Atherogenesis. R. Rezzani, those abundant transcripts that have small differential ed. InTech Inc., Rijeka, Croatia. Accessed Feb. 27, 2013. http:// expression” (Kim et al., 2001). This may explain genes www.intechopen.com/articles/show/title/atherosclerosis-suscep- found in the microarray that were not found in RDA. tible-and-atherosclerosis-resistant-pigeon-aortic-smooth-muscle- Divergent genetics between pigeon and chicken may cells-express-. Anderson, J. L., R. L. Taylor Jr., E. C. Smith, W. K. Thomas, further explain lack of detection on a cross-species ar- and S. C. Smith. 2012b. Differentially expressed genes in aortic ray. This is likely the case with genes associated with smooth muscle cells from atherosclerosis-susceptible and athero- oxidative phosphorylation, as mitochondrial genes are sclerosis-resistant pigeons. Poult. Sci. 91:1315–1325. used to study evolutionary distance between avian spe- Andersson, T., S. Boräng, M. Larsson, V. Wirta, A. Wennborg, J. Lundeberg, and J. Odeberg. 2001. Novel candidate genes for ath- cies (Pacheco et al., 2011). Finally, genes may have not erosclerosis are identified by representational difference analy- been found by microarray because they were not dif- sis-based transcript profiling of cholesterol-loaded macrophages. ferentially expressed in the experiment. Lack of cross- Pathobiology 69:304–314. hybridization or the absence of a transcript from the Bassiouny, H., C. Zarins, M. Kadowaki, and S. Glagov. 1994. Hemo- dynamic stress and experimental aortoiliac atherosclerosis. J. microarray are limitations to identifying expression dif- Vasc. Surg. 19:426–434. ferences. The advantage of both the array and RDA is Bower, J. K., L. J. Appel, K. Matsushita, J. H. Young, A. Alonso, F. identifying true breed differences. L. Brancati, and E. Selving. 2012. Glycated hemoglobin and risk of hypertension in the atherosclerotic risk in communities study. In conclusion, genes involved in cytoskeletal remodel- Diabetes Care 35:1031–1037. ing, proteasome activity, cellular respiration, and the Burnside, J., P. Neiman, J. Tang, R. Basom, R. Talbot, M. Aron- immune response are distinctly different between pi- szajn, D. Burt, and J. Delrow. 2005. Development of a cDNA geon breeds at 1 d of age. These findings are consistent array for chicken gene expression analysis. BMC Genomics 6:13. Clarkson, T. B., R. W. Prichard, M. G. Netsky, and H. B. Lofland. with genetic, proteomic, and metabolic data previously 1959. Atherosclerosis in pigeons: Its spontaneous occurrence reported. The RDA and microarray are complementa- and resemblance to human atherosclerosis. AMA Arch. Pathol. ry methods to investigate differential gene expression. 68:143–147. One should not be used to validate the other because Cooke, P., and S. C. Smith. 1968. Smooth muscle cells: Source of foam cells in atherosclerotic White Carneau pigeons. Exp. Mol. they each have their unique strengths and limitations. Pathol. 8:171–189. Work is in progress to monitor actin and proteoglycan Ding, C., and C. Cantor. 2004. Quantitative analysis of nucleic gene expression during WC-As development as actin acids—The last few years of progress. J. Biochem. Mol. Biol. cytoskeleton changes affect focal adhesion (Yamin and 37:1–10. Eisen, M. B., P. T. Spellman, P. O. Brown, and D. Botstein. 1998. Morgan, 2012), suggesting a relationship between SMC Cluster analysis and display of genome-wide expression patterns. phenotypic switching and the response to retention hy- Proc. Natl. Acad. Sci. USA 95:14863–14868. 2680 Anderson et al.

Garosi, P., C. De Filippo, M. van Erk, P. Rocca-Serra, S. A. San- Ridker, P. 2000. New clue to an old killer: Inflammation and heart sone, and R. Elliott. 2005. Defining best practice for microarray disease. Nutrition Action 27:3–5. analyses in nutrigenomic studies. Br. J. Nutr. 93:425–432. Santerre, R., T. N. Wight, S. C. Smith, and D. Brannigan. 1972. Gomez, D., and G. K. Owens. 2012. Smooth muscle cell phenotype Spontaneous atherosclerosis in pigeons. A model system for switching in atherosclerosis. Cardiovasc. Res. 95:156–164. studying metabolic parameters associated with atherogenesis. Guo, F., C. Zarella, and W. D. Wagner. 2006. STAT4 and the pro- Am. J. Pathol. 67:1–22. liferation of artery smooth muscle cells in atherosclerosis. Exp. SAS Institute. 2007. JMP Genomics Version 3.0 User’s Guide. SAS Mol. Pathol. 81:15–22. Inst. Inc., Cary, NC. Hajjar, D. P., T. F. Wight, and S. C. Smith. 1980. Lipid accumu- Scheckhuber, C. 2005. Mitochondrial dynamics in cell life and death. lation and ultrastructural change within the aortic wall during Sci. SAGE KE 2005:pe36. early spontaneous atherogenesis. Am. J. Pathol. 100:683–705. Shapiro, M. D., Z. Kronenberg, C. Li, E. T. Domyan, H. Pan, M. Hansen, J. 1977. Spontaneous atherosclerosis: An ultrastructural Campbell, H. Tan, C. D. Huff, H. Hu, A. I. Vickrey, S. C. Nielsen, study in the White Carneau pigeon. Virchows Arch. A Pathol. S. A. Stringham, H. Hu, E. Willerslev, M. T. Gilbert, M. Yandell, Anat. Histol. 375:147–157. G. Zhang, and J. Wang. 2013. Genomic diversity and evolution in Hansson, G., and P. Libby. 2006. The immune response in athero- the head crest of the rock pigeon. Science 339:1063–1067. sclerosis: A double-edged sword. Nat. Rev. Immunol. 6:508–519. Smith, S. C., E. C. Smith, M. L. Gilman, J. L. Anderson, and R. Heron, M. 2012. Deaths: Leading causes for 2008. Natl. Vital Stat. L. Taylor Jr.. 2008. Differentially expressed soluble proteins in Rep. 60:1–94. aortic cells from atherosclerosis-susceptible and resistant pigeons.

Kerr, M. K., M. Martin, and G. A. Churchill. 2000. Analysis of Poult. Sci. 87:1328–1334. Downloaded from variance for gene expression microarray data. J. Comput. Biol. Smith, S. C., E. C. Smith, and R. L. Taylor Jr.. 2001. Susceptibility 7:819–837. to spontaneous atherosclerosis in pigeons: An autosomal recessive Kim, S., K. Zeller, C. V. Dang, E. P. Sandgren, and L. A. Lee. 2001. trait. J. Hered. 92:439–442. A strategy to identify differentially expressed genes using repre- St Clair, R. 1983. Metabolic changes in the arterial wall associated sentational difference analysis and cDNA arrays. Anal. Biochem. with atherosclerosis in the pigeon. Fed. Proc. 42:2480–2485. 288:141–148. St. Clair, R. 1998. The contribution of avian models to our under- Kjaernes, M., A. Svindland, L. Walloe, and S. Wille. 1981. Localiza- standing of atherosclerosis and their promise for the future. Lab. http://ps.oxfordjournals.org/ tion of early atherosclerotic lesions in an arterial bifurcation in Anim. Sci. 48:565–568. humans. Acta Pathol. Microbiol. Scand. [A] 89:35–40. Sung, B., K. J. Jung, H. S. Song, M. J. Son, B. P. Yu, and H. Y. Lichter, P. 2000. New tools in molecular pathology. J. Mol. Diagn. Chung. 2005. cDNA representational difference analysis used in 2:171–173. the identification of genes related to the aging process in rat kid- Linder, K., P. Arner, A. Flores-Morales, P. Tollet-Egnell, and G. ney. Mech. Ageing Dev. 126:882–891. Norstedt. 2004. Differentially expressed genes in visceral or sub- Tyson, K. L., and C. M. Shanahan. 2000. Use of cDNA represen- cutaneous adipose tissue of obese men and women. J. Lipid Res. tational difference analysis to identify disease specific genes in 45:148–154. human atherosclerosis. Pages 83–98 in Methods in Molecular

Lucito, R., J. Healy, J. Alexander, A. Reiner, D. Esposito, M. Chi, Medicine. A. H. Baker, ed. Humana Press, Clifton, NJ. at University of New Hampshire Library on July 17, 2014 L. Rodgers, A. Brady, J. Sebat, J. Troge, J. A. West, S. Rostan, Tyson, K. L., P. L. Weissberg, and C. M. Shanahan. 2002. Hetero- K. C. Nguyen, S. Powers, K. Q. Ye, A. Olshen, E. Venkatraman, geneity of gene expression in human atheroma unmasked using L. Norton, and M. Wigler. 2003. Representational oligonucleotide cDNA representational difference analysis. Physiol. Genomics microarray analysis: A high-resolution method to detect genome 9:121–130. copy number variation. Genome Res. 13:2291–2305. VanderLaan, P. A., and C. A. Reardon. 2005. Thematic review se- Marioni, J. C., C. E. Mason, S. M. Mane, M. Stephens, and Y. ries: The immune system and atherogenesis. The unusual sus- Gilad. 2008. RNA-seq: An assessment of technical reproducibil- pects: An overview of the minor leukocyte populations in athero- ity and comparison with gene expression arrays. Genome Res. sclerosis. J. Lipid Res. 46:829–838. 18:1509–1517. Wagner, W. 1978. Risk factors in pigeons genetically selected for Moghadasian, M. H., J. J. Frohlich, and B. M. McManus. 2001. increased atherosclerosis susceptibility. Atherosclerosis 31:453– Advances in experimental dyslipidemia and atherosclerosis. Lab. 463. Invest. 81:1173–1183. Watanabe, J., K. J. Chou, J. C. Liao, Y. Miao, H. H. Meng, H. Ge, Nicolosi, R. J., R. F. Santerre, and S. C. Smith. 1972. Lipid accumu- V. Grijalva, S. Hama, K. Kozak, G. Buga, J. P. Whitelegge, T. lation in muscular foci in White Carneau and Show Racer pigeon D. Lee, R. Farias-Eisner, M. Navab, A. M. Fogelman, and S. T. aortas. Exp. Mol. Pathol. 17:29–37. Reddy. 2007. Differential association of hemoglobin with proin- Nikolsky, Y., S. Ekins, T. Nikolskaya, and A. Bugrim. 2005. A novel flammatory high density lipoproteins in atherogenic/hyperlipid- method for generation of signature networks as biomarkers from emic mice. A novel biomarker of atherosclerosis. J. Biol. Chem. complex high throughput data. Toxicol. Lett. 158:20–29. 282:23698–23707. Ordovas, J., and A. Shen. 2002. Genetics, the environment, and lipid Wolfinger, R. D., G. Gibson, E. D. Wolfinger, L. Bennett, H. Hama- abnormalities. Curr. Cardiol. Rep. 4:508–513. deh, P. Bushel, C. Afshari, and R. S. Paules. 2001. Assessing gene Owens, G. K. 1996. Role of alterations in the differentiated state of significance from cDNA microarray expression data via mixed smooth muscle cell in atherogenesis. Pages 401–420 in Athero- models. J. Comput. Biol. 8:625–637. sclerosis and Coronary Artery Disease. V. Fuster, R. Ross, and Xu, Q., B. Metzler, M. Jahangiri, and K. Mandal. 2012. Molecular E. Topol, ed. Lippincott-Raven, Philadelphia, PA. chaperones and heat shock proteins in atherosclerosis. Am. J. Pacheco, M. A., F. U. Battistuzzi, M. Lentino, R. F. Aguilar, S. Physiol. Heart Circ. Physiol. 302:H506–H514. Kumar, and A. A. Escalante. 2011. Evolution of modern birds Yamin, R., and K. G. Morgan. 2012. Deciphering actin cytoskel- revealed by mitogenomics: Timing the radiation and origin of etal function in the contractile vascular smooth muscle cell. J. major orders. Mol. Biol. Evol. 28:1927–1942. Physiol. 590:4145–4154. Pastorian, K., L. Hawel, and C. Byus. 2000. Optimization of cDNA Yu, E., J. Mercer, and M. Bennett. 2012. Mitochondria in vascular representational difference analysis for the identification of dif- disease. Cardiovasc. Res. 95:173–182. ferentially expressed mRNAs. Anal. Biochem. 283:89–98. Perry, M. E. 2010. The regulation of the p53-mediated stress re- sponse by mdm2 and mdm4. Cold Spring Harb. Perspect. Biol. 2:a000968.