Rosiglitazone modulates pigeon atherosclerotic lipid accumulation and gene expression in vitro

Abstract

Atherosclerosis is a major contributor to the overall United States mortality rate, primarily in the form of heart attacks and stroke. Unlike the human disease, which is believed to be multifactorial, pigeon atherosclerosis is due to a single gene autosomal recessive trait. The White Carneau (WC-As) strain develops atherosclerotic plaques without the presence of known environmental risk factors such as diet and classic predictors such as blood pressure or blood cholesterol levels. With similar parameters, the Show Racer (SR-Ar) is resistant to plaque development. Thiazolidinediones, including rosiglitazone, activate the peroxisome proliferator-activated receptor gamma (PPARγ) raising cellular sensitivity to insulin. The effect of rosiglitazone was evaluated in aortic smooth muscle cells (SMC) from these 2 pigeon breeds. Primary SMC cultures were prepared from WC-As and SR-Ar squabs. Cell monolayers, which achieved confluence in 7 d, were treated with 0 or 4 µM rosiglitazone for 24 h. Cellular lipid accumulation was evaluated by oil red O staining. Control WC-As cells had significantly higher vacuole scores and lipid content than did the SR-Ar control cells. Rosiglitazone treatment decreased WC-As lipid vacuoles significantly compared with the control cells. On the other hand, lipid vacuoles in the treated and untreated SR-Ar cells did not differ significantly. The effect of rosiglitazone on WC-As SMC gene expression was compared with control SMC using representational difference analysis. Significant transcript increases were found for caveolin and RNA binding motif in the control cells compared with the rosiglitazone-treated cells as well as cytochrome p450 family 17 subfamily A polypeptide 1 (CYP171A) in the rosiglitazone-treated cells compared with the control cells. Although rosiglitazone was selected for these experiments because of its role as a PPARγ agonist, it appears that the drug also tempers c-myc expression, as genes related to this second transcription factor were differentially expressed. Both PPARγ and c-myc appear to affect WC-As SMC gene expression, which may relate to disease development, progression, or both.

INTRODUCTION

Atherosclerosis is one of the leading causes of death in the United States. As a prevalent complication of type II diabetes, the disease is affecting an increasing population in economically developed countries. Animal models for human disease are critical to understanding the mechanisms underlying specific pathologies, and the pigeon is one of several good animal models for atherosclerosis (Moghadasian et al., 2001; Anderson et al., 2011). White Carneau (WC-As) pigeons are susceptible to spontaneous atherosclerosis at the celiac bifurcation of the aorta, whereas Show Racers (SR-Ar) are resistant (Clarkson et al., 1959). No experimental manipulation is required to observe atherosclerotic progression in the WC-As, and lesions in the WC-As are quite similar to human lesions in their structure as well as anatomic location. Atherosclerosis in the WC-As is independent from known environmental risk factors and occurs in the absence of elevated blood cholesterol levels (St. Clair, 1983). Susceptibility resides at the level of the arterial wall (St. Clair, 1983) and is inherited in a pattern consistent with a single gene autosomal recessive trait (Smith et al., 2001). The spontaneous disease development, pathology, and aortic position make the WC-As a good model to study the earliest events in atherogenesis, as well as the overall disease process (Moghadasian et al., 2001; Anderson et al., 2011).

In comparison with corresponding aortic sites in SR-Ar, the earliest biochemical difference observed in WC-As is an increase in nonesterified fatty acids (NEFA) in the celiac focus at 1 d of age (Hajjar et al., 1980). This appears to reflect lesser oxidation of lipid during embryonic development in WC-As (Cramer and Smith, 1976). By 6 mo of age, cholesterol esters accumulate in susceptible WC-As foci. Although disease progression is spontaneous in the WC-As, addition of exogenous NEFA enhances lipid accumulation in aortic cells from both breeds (S. C. Smith and E. C. Smith, unpublished data). This indicates an accelerating role of these fatty acids in pigeon atherosclerosis. In humans, higher NEFA levels (Moinet et al., 1991; Phillips et al., 2005) appear before the development of both atherosclerosis and diabetes (Olsson et al., 1999) causing cells to become resistant to insulin (Rodriguéz-Lee et al., 2007). In addition, NEFA are endogenous ligands for the nuclear hormone transcription factor, peroxisome proliferator-activated receptor gamma (PPARγ; Lee et al., 2003), and as such may have a direct effect on gene expression (Sampath and Ntambi, 2005).

The PPARγ regulates overall energy metabolism and is an important determinant of vascular structure and function. All cells involved in atherosclerotic lesion formation, including smooth muscle cells (SMC), express PPARγ (Lin et al., 2004). When the genetic expression of WC-As and SR-Ar SMC is compared in vitro, genes involved in energy metabolism and cellular contractility are the most significant differences (Anderson et al., 2012). Oxidative phosphorylation is clearly upregulated in the resistant SR-Ar, whereas glycolysis predominates in the susceptible WC-As. In vivo PPARγ network gene expression differs between WC-As and SR-Ar aortic cells (Anderson et al., 2013). Therefore, it appears that differences in PPARγ structure, activation status, or both, contribute to the atherosclerotic phenotype in pigeons. To test this hypothesis, agents that regulate PPARγ activity should be useful.

Thiazolidinediones (TZD) are a class of drugs that serve as preferential ligands for PPARγ (Martens et al. 2002), competing with NEFA for the active binding sites on this receptor (Hsueh and Law, 2001). Pharmacological PPARγ activation with TZD reduces vascular lesion formation in animal models of atherosclerosis (Chen et al., 2001). The TZD increase insulin sensitivity in diabetic humans, decrease intimal thickening in carotid arteries, and in clinical studies minimize the aortic lumen narrowing that occurs during restenosis in those patients with cardiovascular complications (Marx et al., 2004). Thiazolidinediones have pro-apoptotic and anti-proliferative effects in vascular SMC, but the mechanism(s) for these effects are ill-defined because the target genes are not well characterized (Lin et al., 2004).

Rosiglitazone is one member of the TZD family that has been shown to reduce atherosclerotic progression in nondiabetic patients with coronary artery disease (van Wijk and Rabelink, 2004). In an apparent contradiction, use of rosiglitazone, marketed under the brand name Avandia, was restricted in 2010 because of potential heart attack and stroke risks. Recently, the FDA has considered lifting these restrictions after subsequent data analysis supported the original research (Morgan, 2013). The objective of the current study was to examine the effect of rosiglitazone on atherogenic events in pigeon SMC.

MATERIALS AND METHODS

Pigeons

Atherosclerosis-susceptible White Carneau and SR-Ar pigeons were housed in fly coops at ambient temperatures and allowed free access to water, Purina Pigeon Chow Checkers (St. Louis, MO), and Kaytee Red Grit (Chilton, WI). The University of New Hampshire colonies were established in 1962 with birds obtained from the Palmetto Pigeon Plant in Sumter, South Carolina, and have remained closed flocks. The birds were maintained under the supervision of the University of New Hampshire Animal Care and Use Committee (approval #050601).

Primary Cell Culture

Four 1-d-old WC-As and SR-Ar squabs were used to prepare primary cultures of aortic smooth muscle cells (SMC; Smith et al., 1965). Explants for each breed were grown for 7 d on coverslips in flasks until the cells formed a confluent monolayer. At that time, the flasks were divided into 2 groups. Four flasks received 4 µM of rosiglitazone (Pfizer, Groton, CT) and 3 flasks remained as the control.

Lipid Analysis

After 24 h explants grown on coverslips were stained with oil red O (ORO) and examined under phase contrast with microscopy at 20× (Smith et al., 2001). For each explant, 10 cells were examined in each of 4 different areas of growth around the explant and graded according to the criteria presented in Table 1. The grades were averaged and the mean vacuole grade was reported for each explant. Vacuole grades were converted to lipid content from a standard curve (Smith et al., 2001).

Table 1.

Vacuole grade criteria of individual aortic smooth muscle cells stained with oil red O (ORO; Smith et al., 2001)

Vacuole grade		Cell appearance
1		No ORO-stained droplets
2		One-third of cytoplasm occupied by ORO-positive droplets
3		Two-thirds of cytoplasm occupied by ORO-positive droplets
4		Foam cells. Entire cell cytoplasm occupied by ORO-positive droplets

Representational Difference Analysis

Parallel control or rosiglitazone-treated WC-As cultures grown in flasks without coverslips, were serially harvested in Trizol (Invitrogen, Carlsbad, CA) for RNA extraction. Total RNA was extracted and converted to double-stranded cDNA using the BD SMART PCR cDNA Synthesis kit (#K1052–1, Clontech, Mountain View, CA). The representational difference analysis (RDA) protocol followed Pastorian et al. (2000) including modifications of Tyson and Shanahan (2000) and Anderson et al. (2012). Three rounds of PCR-coupled subtractive hybridizations were performed in reciprocal experiments. Final difference products were cloned into the BamHI restriction site of the pBluescript SK+ phagemid vector (#212205, Agilent Technologies, Santa Clara, CA) as previously described (Anderson et al., 2012). Clones were sequenced by DNA Core at the Hubbard Center for Genome Studies (University of New Hampshire) on an ABI 3130 Genetic Analyzer (Life Technologies, Grand Island, NY). Raw sequence data were trimmed and used to query highly similar sequences in GenBank using megablast. If no significant match (e-10) was found, the sequences were subjected to blastx (e-10) and then tblastx (e-05 plus 20/20 amino acids).

RESULTS

Lipid Analysis

White Carneau and SR-Ar aortic smooth muscle cells were treated with rosiglitazone, and the amount of lipid present in each vacuole was determined. The vacuole grades are presented in Table 2a, and the overall lipid content in Table 2b. The WC-As control cells had a significantly higher vacuole grade (2.19) than the SR-Ar control cells (1.73), the SR-Ar treated cells (1.66), and the WC-As treated cells (1.10). Rosiglitazone had a significant (P < 0.0001) effect in the WC-As cells, decreasing the amount of cytoplasm occupied by ORO positive droplets from one-third to below the value of the resistant SR-Ar cells. The lipid content (1,152 µg of lipid/mg of protein) of WC-As control cells differed statistically from the SR-Ar control cells (880 µg of lipid/mg of protein), the SR-Ar treated cells (860 µg of lipid/mg of protein), and the WC-As treated cells (806 µg of lipid/mg of protein).

Table 2.

Mean vacuole grade and lipid content of control or rosiglitazone-treated aortic smooth muscle cells from White Carneau (WC-As) and Show Racer (SR-Ar) pigeons

Item	Explants (n)	White Carneau (WC-As)	Explants (n)	Show Racer (SR-Ar)
Vacuole grade
Control	45	2.19 ± 0.022a	55	1.73 ± 0.032b
Rosiglitazone	5	1.10 ± 0.045c	13	1.66 ± 0.073b
Lipid content (µg of lipid/mg of protein)
Control	45	1,152 ± 11.6a	55	880 ± 18.6b
Rosiglitazone	5	806 ± 33.0b	13	860 ± 43.3b

^a–cMeans within corresponding treatments having no common letter differ significantly (P < 0.001).

RDA

The difference products from the RDA experiment were cloned into 2 libraries. The genetic transcripts that were upregulated in the rosiglitazone-treated cells comprise the WC_ROSI library, and those genes expressed by the control cells are in the WC_CONT library. One hundred ninety-two clones were sequenced, revealing 120 difference products (DP) per treatment. GenBank requests that 18S and 28S rRNA sequences not be submitted to the expressed sequence tag database, so these transcripts were considered in the analysis but not submitted to GenBank. The library statistics and GenBank accession numbers of the differentially expressed transcripts are presented in Table 3.

Table 3.

Total number of expressed sequence tags (EST) from representational difference analysis (RDA) of rosiglitazone-treated and untreated atherosclerosis-susceptible White Carneau (WC-As) aortic smooth muscle cells submitted to the National Center for Biotechnology Information Basic Local Alignment Search Tool databases

Library (GenBank no.)		EST		GenBank accession no.	Identified, no. (%)	Unidentified, no. (%)
WC_ROSI (028248)		71		JZ476940–JZ477010	59 (83)	12 (17)
WC_CONT (028241)		120		JZ476749–JZ476868	110 (91)	10 (9)

On average, 87% of the sequences had an orthologous match in the National Center for Biotechnology Information databases. These sequences were considered to be identified and examined for redundancy. Twenty-four unique DP were determined to be exclusive to WC_ROSI (Table 4) including cytochrome p450 17 α hydroxylase (CYP17A1) and enolase (ENO1). Forty DP were exclusively expressed by the WC_CONT (Table 5) including RNA binding motif, single-stranded interacting protein (RBMS1), activin A receptor, type 1 (ACVR1), and caveolin (CAV1). Two DP were expressed in each treatment. One had equivalent expression in the control and rosiglitazone-treated cells and one DP that showed nonsignificant copy number variation (data not shown).

Table 4.

Copy number (CN) of transcripts expressed exclusively in rosiglitazone-treated (Rosi) atherosclerosis-susceptible White Carneau (WC-As) aortic smooth muscle cells

Gene		Gene name	Rosi	Control	P-value
CN differs significantly (P < 0.05)  by chi square analysis
CYP17A1		Cytochrome p450 17 α hydroxylase	13	0	0.0002
ENO1		Enolase	9	0	0.0022
CN does not differ significantly  by chi square analysis
EEF2		Eukaryotic translation elongation factor 2	2	0	0.1556
RPL27		L27 mt	2	0	0.1556
RPS4		RPS4-X-linked	2	0	0.1556
SULT1B1		Sulfotransferase family (cytosolic 1B member 1)	2	0	0.1556
UCHL5		Ubiquitin c variant 5	2	0	0.1556
AARS2		Alanyl-tRNA synthetase 2 (mt)	1	0	0.3163
ACTB		β actin	1	0	0.3163
ADH3		Alcohol dehydrogenase class 3	1	0	0.3163
ATP5A1W		ATP synthase α, mitochondrial like	1	0	0.3163
COL6A2		Collagen type VI α 2	1	0	0.3163
FTH1		Ferratin heavy chain	1	0	0.3163
GAPDH		Glyceraldehyde 3 phosphate dehydrogenase	1	0	0.3163
GDI2		GDP dissociation inhibitor 2	1	0	0.3163
IGFBP2		Insulin growth factor binding protein 2	1	0	0.3163
KDELR2		ER protein retention receptor 2	1	0	0.3163
LAMR		Laminin receptor 1	1	0	0.3163
PDLIM7		PD2 and LIM domain 7	1	0	0.3163
RPL3		RP L3	1	0	0.3163
RPS14		RP 40S S14	1	0	0.3163
SLC5A6		Sodium dependent multivitamin transporter like	1	0	0.3163
TAF9		TATA-box binding protein associated factor 9 aka TFID9	1	0	0.3163
TKT1		Transketolase	1	0	0.3163
Total			49	0

Table 5.

Copy number (CN) of transcripts expressed exclusively in control atherosclerosis-susceptible White Carneau (WC-As) aortic smooth muscle cells

Gene		Gene product	Rosi	Control	P-value
Copy number differs significantly (P < 0.05)  by  chi square analysis
RPS8		40S ribosomal protein S8	0	16	0.0000
RBMS1		RNA binding motif, ss interacting protein 1	0	14	0.0001
ACVR1		Activin A receptor, type 1	0	7	0.0072
LUM		Keratan sulfate proteoglycan lumican	0	7	0.0072
HSP90AA		Heat shock protein 90 kDa α class A	0	5	0.0238
PRKAR1A		Protein kinase, cAMP-dependent, regulatory type 1, α	0	5	0.0238
CAV1		Caveolin-1	0	4	0.0437
ORF		C18 ORF21	0	4	0.0437
Copy number does not differ significantly  by chi square analysis
PRDX1		Peroxiredoxin 1-like (variant 2)	0	3	0.0813
EEF1A1		Elongation factor 1-α 1	0	2	0.1556
EIF3E		Eukaryotic translation initiation factor 3, subunit E	0	2	0.1556
RPL30		Ribosomal protein L30	0	2	0.1556
RPL7		Ribosomal protein L7	0	2	0.1556
SAT1		Spermidine/spermine N1-acetyltransferase 1	0	2	0.1556
STT3B		STT3 subunit oligosaccharyltransferase complex homolog b	0	2	0.1556
TIMM22		Translocase of inner MT membrane 22 homolog	0	2	0.1556
ATP5A1		ATP synthase, H+ transporting, mt F1 complex, α 1	0	1	0.3163
CCDC80		Coiled-coil domain containing 80	0	1	0.3163
CCND2		Cyclin D2	0	1	0.3163
COL3A1		Collagen type III α 1	0	1	0.3163
COTL1		Coactosin	0	1	0.3163
CWC15		Splicesome associated protein	0	1	0.3163
DSTN		Destrin/actin depolymerization factor	0	1	0.3163
EIF2D		Eukaryotic translation initiation factor 2D	0	1	0.3163
EIF4A3		Eukaryotic translation initiation factor 4A, isoform 3	0	1	0.3163
EST		ChEST972124	0	1	0.3163
HNRNAB		Heterogenous nuclear ribonucleoprotein A/B	0	1	0.3163
ID2		Inhibitor of DNA binding 2, dominant negative helix-loop-helix protein	0	1	0.3163
ITA		Inhibitor of apoptosis protein-like (T cell in chicken)	0	1	0.3163
LY6E		Lymphocyte antigen 6E-like	0	1	0.3163
MRPL33		Mitochondrial ribosomal protein L33	0	1	0.3163
MTA1		Metastasis associated 1	0	1	0.3163
MYH9		Myosin, heavy chain 9, nonmuscle	0	1	0.3163
RPL26		Ribosomal protein L26	0	1	0.3163
SEC61G		Sec61, gamma subunit	0	1	0.3163
SLC25A6		Solute carrier family 25, member 6	0	1	0.3163
STAT1		Signal transducer and activator of transcription 1	0	1	0.3163
TIMP3		TIMP Metallopeptidase inhibitor 3	0	1	0.3163
XPO1		Exportin 1	0	1	0.3163
XRN2		5′-3′ exoribonuclease 2	0	1	0.3163
Total			0	103

DISCUSSION

Rosiglitazone decreased lipid in the WC-As cells to a level similar to the SR-Ar, unlike the rosiglitazone treatment vacuole grade in the WC-As that was reduced below the resistant breed. The treatment did not have a significant impact in the SR-Ar, possibly because there were fewer lipid inclusions present to reduce. Both lipid methods, vacuole grade and lipid concentration, indicate that the WC-As is susceptible to the rosiglitazone effects, whereas the SR-Ar is resistant. For this reason, RDA was only performed on the WC-As cells.

Significant examples of transcripts with higher expression in the control cells include RBMS1, ACVR1, LUM, and CAV1. Nonsignificant transcripts include PRDX1, SAT1, CCND2, COL3A1, and multiple eukaryotic translation initiation factors. Expression of direct downstream targets of PPARγ (Burgermeister et al., 2003), ACVR1 and CAV1, was not found in rosiglitazone-treated cells. This is consistent with drug-induced PPARγ antagonism.

The CAV1 regulates the flow of cholesterol in and out of the cell (Cohen et al., 2004), meaning its differential expression in the current study likely contributes to the observed differences in cellular lipid. Caveolin has previously been shown to be differentially expressed between susceptible and resistant pigeon breeds, both in vitro (Anderson et al., 2012) and in vivo (Anderson et al., 2013). In a crossbreeding study of CAV1 and ApoE knockout mice, the lack of CAV1 decreased the lesion area and actually provided protection against the known atherogenic effects of APOE deficiency (Frank et al., 2004). Caveolin expression has been associated with the early events of many human pathologies, including type II diabetes (Cohen et al., 2004), the disease targeted by the pharmaceutical form of rosiglitazone, Avandia.

The RBMS1 was also exclusively expressed in the WC-As control SMC (P < 0.0001). This gene is one direct activator of c-myc, a known oncogene in cervical cancer (National Center for Biotechnology Information, GenBank). The addition of rosiglitazone in this experiment appears to block RBMS1 expression. Although c-myc was not detected in the RDA, altered regulation of that oncogene suggested by differential RBMS1 expression in the current study is interesting for 2 reasons. First, previous in vitro studies have demonstrated that RBMS1 is expressed in the susceptible WC-As and not in the resistant SR-Ar (Anderson et al., 2012). Second, c-myc has been shown to regulate CAV1 (Carver et al., 2003). The transcriptional status of c-myc may be a key factor in pigeon atherogenesis because it provides a mechanistic link to the lipid accretion and cellular proliferation observed in susceptible pigeon aortic cells.

Although c-myc has been primarily recognized for its role with carcinogenesis (Dang et al., 1999; Nasi et al., 2001), several studies have found an association between c-myc expression and atherosclerosis. The c-myc oncogene, which regulates the transcription of multiple genes affecting cell cycle and transformation (Niki et al., 2000) is typically not expressed in quiescent cells (Nasi et al., 2001). Human SMC proliferation, an early event in atherogenesis, has been shown to be under c-myc influence (Shi et al., 1993). Quail that were fed an atherogenic diet showed increased c-myc transcriptional activity relative to the control diet, which was directly correlated to the degree of atherogenic lesions (Inafuku et al., 2007).

The c-myc is inhibited by a portion of the glycolytic enzyme ENO1 (Subramanian and Miller, 2000), which was exclusively expressed (P < 0.0022) by the rosiglitazone treated WC-As cells. Enolase expression in those cells is consistent with previous studies, where the enzyme transcript was upregulated relative to the SR-Ar (Anderson et al., 2012) and thought to be related to the susceptible breed’s reliance on glycolysis in the absence of efficient oxidative phosphorylation. However, the current experiment suggests WC-As enolase expression may be an attempt to dampen c-myc transcription, which is abnormally high in the WC-As.

Rosiglitazone appears to block c-myc expression in WC-As aortic cells by altering the transcription profile of ENO1 and RBMS1. Because rosiglitazone is a PPARγ-specific antagonist, the results do not reveal if the ENO1 decrease or RBMS1 increase is entirely a c-myc effect. However, previous studies have demonstrated that TZD18, an alternative TZD, inhibits the expression of c-myc (Liu et al., 2006) via a PPARγ independent pathway. In addition, rosiglitazone has been shown to confer anti-atherogenic properties independent of insulin sensitivity and glycemic control (Calkin et al., 2005). Perhaps this effect is mediated through c-myc rather than PPARγ.

Differences in c-myc activation could affect pigeon atherosclerotic susceptibility and explain why so many genes as well as proteins have differential expression between the 2 pigeon breeds. In addition to the proliferative effect of c-myc via CNCD2, there is the concurrent upregulation of CAV1, associated both with cellular cholesterol metabolism and tumor metastasis (Carver et al. 2003). The c-myc also inhibits the transcription of α-actin (Kimura et al., 1998), a key marker in healthy SMC differentiation. Previous in vitro (Anderson et al., 2012) and in vivo (Anderson et al., 2013) pigeon studies have shown an absence of α-actin expression in WC-As SMC relative to SR-Ar, which occurred before intracellular lipid differences.

Because of c-myc pleiotropic effects, further studies are warranted to investigate that gene’s regulation and structure in pigeons. It is possible that rosiglitazone does not affect SR-Ar because c-myc is not typically expressed in these resistant cells. Therefore, transcription does not need to be lowered. In contrast, c-myc overexpression in the WC-As cells makes them responsive to the TZD inhibitory effect. A consequence of inhibiting c-myc is to block CAV1 activity, eliminating lipid accumulation in the rosiglitazone-treated WC-As that was not observed in the SR-Ar.

The CYP17 was significantly (P < 0.0002) upregulated in the rosiglitazone-treated WC-As cells. Its expression is not surprising because CYP17 is directly involved in drug metabolism (GenBank). However, CYP17 is also involved in cholesterol metabolism, a result substantiated by lipid accretion differences between treatments. The lipid that accumulates in the WC-As SMC consists primarily of cholesterol esters (Nicolosi et al., 1972; Hajjar et al., 1980). Therefore, it is reasonable to suggest that the increased CYP17 expression is part of the rosiglitazone lipid-lowering effect. However, more studies are required to demonstrate this relationship.

Finally, we found thirty-one 18S and eighteen 28S transcripts in the rosiglitazone-treated cells relative to the controls. This number is not included in the total transcripts because GenBank does not want to confound its database with these commonly expressed genes. Although rRNA is often considered a false positive in RDA experiments, the exclusive nature of this finding might suggest increased protein synthesis in the rosiglitazone-treated cells. Skov et al., (2008) showed that pioglitazone, an alternate TZD, does enhance ribosomal biogenesis in skeletal muscle.

This study met the objective of examining possible anti-atherogenic effects of rosiglitazone on pigeon aortic SMC. We found a significant decrease in lipid accumulation in rosiglitazone-treated WC-As cells compared with the controls. By the assessments we used, the SR-Ar seems rather resistant to any effect of rosiglitazone, whereas the WC-As are sensitive and appear to be modulated to the level of the SR-Ar. Genes expressed by the atherosclerosis-susceptible WC-As control SMC were downregulated in the rosiglitazone-treated cells. Significant differences include RBMS1 and CAV1. Caveolin is a direct target of PPARγ, but also appears to be regulated by c-myc. Rosiglitazone was selected as a PPARγ agonist, but it appears to be a ligand for both transcription factors. The putative role of c-myc in pigeon atherogenesis warrants further investigation.