Retinoid-Induced Expression and Activity of an Immediate Early Tumor Suppressor Gene in Vascular Smooth Muscle Cells

Retinoids are used clinically to treat a number of hyper-proliferative disorders and have been shown in experimental animals to attenuate vascular occlusive diseases, presumably through nuclear receptors bound to retinoic acid response elements (RARE) located in target genes. Here, we show that natural or synthetic retinoids rapidly induce mRNA and protein expression of a specific isoform of A-Kinase Anchoring Protein 12 (AKAP12β) in cultured smooth muscle cells (SMC) as well as the intact vessel wall. Expression kinetics and actinomycin D studies indicate Akap12β is a retinoid-induced, immediate-early gene. Akap12β promoter analyses reveal a conserved RARE mildly induced with atRA in a region that exhibits hyper-acetylation. Immunofluorescence microscopy and protein kinase A (PKA) regulatory subunit overlay assays in SMC suggest a physical association between AKAP12β and PKA following retinoid treatment. Consistent with its designation as a tumor suppressor, inducible expression of AKAP12β attenuates SMC growth in vitro. Further, immunohistochemistry studies establish marked decreases in AKAP12 expression in experimentally-injured vessels of mice as well as atheromatous lesions in humans. Collectively, these results demonstrate a novel role for retinoids in the induction of an AKAP tumor suppressor that blocks vascular SMC growth thus providing new molecular insight into how retiniods may exert their anti-proliferative effects in the injured vessel wall.


Introduction
Vascular SMC are normally quiescent and express a repertoire of cytoskeletal and contractile proteins that subserve functions related to contractile tone and the maintenance of vascular integrity. A variety of vasculopathies shift the phenotype of SMC from one of quiescence and contractile competence to proliferation, migration, matrix production, and attenuated expression of contractile proteins [1]. A variety of therapeutic molecules have been shown to attenuate such phenotypic switching including a class of compounds known as retinoids [2]. Retinoids encompass synthetic and natural derivatives of retinol (vitamin A) that have found clinical utility in the management of several human hyperproliferative disorders [3,4]. Cultured SMC treated with the natural retinoid, all-trans retinoic acid (atRA) or its isoform (e.g., 9-cis RA), consistently show reduced growth potential during growth factor stimulation [5][6][7][8][9] and in some cases restoration of the contractile phenotype [10]. Moreover, a variety of animal models of vascular disease have been used to demonstrate retinoidmediated decreases in neointimal burden and increases in vessel patency [11][12][13][14][15][16][17][18][19]. Thus, retinoids represent a viable class of therapeutic molecules for the potential management of vascular occlusive disorders.
Retinoids exert their pleiotropic actions by binding ligandactivated nuclear receptors that control gene expression [20]. One logical approach to begin elucidating the mechanisms underlying retinoid action in the vessel wall is to define retinoid-responsive target genes. We previously performed a modified suppression subtractive hybridization screen in cultured SMC for the identification of atRA-responsive genes [21]. One of the genes reported to be induced by atRA was Src-Suppressed C Kinase Substrate (SSeCKS), the rodent ortholog [22] of human gravin [23] encoding for an A-kinase anchoring protein (AKAP). SSeCKS (official gene symbol, AKAP12) binds and localizes a number of signaling proteins including PKA, protein kinase C (PKC), calmodulin, and the b 2 -adrenergic receptor [24,25]. The assembly of such signaling complexes is linked to AKAP12 activities involved in growth suppression, actin cytoskeletal remodeling, and adrenergic signal transduction [24]. The growth suppressive activities of AKAP12, together with its attenuated expression in both transformed cell lines and a variety of human neoplasms, have led to the concept of AKAP12 being a tumor suppressor gene [24]. Further evidence for this hypothesis was recently demonstrated in AKAP12 knockout mice, which exhibit prostatic hyperplasia and focal dysplasia [26].
Previous studies have documented AKAP12 expression in SMC [21,27,28], but the regulatory control of its induction and function in SMC have not been well characterized. Moreover, since the AKAP12 locus comprises three independent transcription units, each under control of a unique promoter [29], the AKAP12 isoform responsive to the action of retinoids is unknown. In this report, we show that the AKAP12b isoform is rapidly and highly induced by both natural and synthetic retinoids. We further show that AKAP12b associates with PKA and mediates increases in activity of at least one downstream target of PKA. Acute or inducible over-expression of AKAP12b attenuates SMC growth in human and rodent SMC model systems. Finally, we demonstrate decreases in AKAP12 expression in vascular lesions where hyperproliferative activity exists. Our results establish AKAP12b as a novel retinoid-responsive tumor suppressor gene, making it an attractive target for therapy in a variety of disease contexts, including vascular occlusive diseases.

atRA Induces a Specific Isoform of AKAP12 in Multiple Species of Vascular SMC
In the course of defining a novel retinoid-response gene set, we identified a tumor suppressor gene called AKAP12 that was induced with atRA in RASMC [21]. To extend these results, we performed Northern blotting on multiple sources of SMC treated for varying times with atRA. Results in Figure 1A reveal a rapid induction of Akap12 mRNA in the rat PAC1 SMC line, RASMC, and HCASMC. The increase in Akap12 mRNA was also seen with 13 cis-RA stimulation, but not with agonists to PPAR gamma (data not shown). On the other hand, agonists to retinoic acid receptor (RAR) and retinoid X receptor (RXR) each elicited increases in Akap12 mRNA ( Figure 1B). The increase in Akap12 mRNA with atRA was dose-dependent ( Figure 1C) and RNA polymerase IIdependent as evidenced by complete suppression with actinomycin D treatment ( Figure 1D). atRA-stimulated Akap12 mRNA was not universally seen as some cell types (L6 and BC 3 H1 myoblasts) failed to show increases with retinoid treatment (data not shown). Akap12's rapid expression kinetics in SMC following retinoid stimulation, its dependence on de novo mRNA synthesis, and its independence for de novo protein synthesis [21] indicate that this tumor suppressor is a retinoid-induced, immediate-early gene.
We recently defined the Akap12 genomic landscape and discovered that three independent promoters direct expression of three Akap12 isoforms (Figure 2A and [29]). To ascertain which of the Akap12 isoforms is targeted for induction with retinoids, exonspecific probes to each Akap12 isoform were designed and applied to samples of RNA from atRA-stimulated PAC1 SMC. Results indicate that atRA specifically targets the Akap12b isoform with expression kinetics nearly identical to those observed with a probe common to all Akap12 isoforms (compare Figure 1A with Figure 2B). Similar kinetics of Akap12b induction was seen with the synthetic retinoid AM80 ( Figure S1A). mRNA kinetic studies suggest that the half-life of retinoid-induced Akap12b is on the order of 3 hr ( Figure S1B).
To determine whether retinoids elicit Akap12b-specific induction in the intact vessel wall, we administered atRA or corn oil to adult mice by oral gavage and measured serum retinoid levels as well as Akap12 isoforms in vascular tissue (enriched for SMC only) using PCR primers specific for Akap12a or Akap12b. No detectable levels of retinoids were seen in corn oil treated mice. However, consistent with a previous report in the rat [11], atRA-treated mice exhibited therapeutic levels of atRA (5144.86701 ng/ml), 13-cis RA (410.9686 ng/ml) and 9-cis RA (29.769 ng/ml) 6 hr following atRA administration. As with in vitro cell culture data, SMC-enriched vascular tissue exposed to atRA showed little change in Akap12a mRNA but clear increases in the Akap12b isoform ( Figure 2C). A similar induction of AKAP12 protein was observed in vascular SMC-enriched aortic tissue ( Figure 2D). Taken together, results indicate rapid and specific induction of the Akap12b isoform upon treatment with retinoids in various SMC culture models as well as mouse SMC of the vessel wall.

Akap12b Promoter Harbors an Atypical RARE
We previously characterized the Akap12b promoter and showed basal activity in a variety of cell types, including SMC [29]. To determine whether any conserved RARE is present in the Akap12b promoter, we compared the rat, mouse, and human Akap12b promoters for conserved RAREs based on a base frequency table of 67 experimentally-validated RAREs [4]. We found an RARE located -2,534 bp upstream of the annotated start site of transcription (see GenBank Accession number AY695060) and the sequence of this RARE indicates that it is a direct repeat (DR)-2 RARE, where each half site is spaced by two nucleotides ( Figure 3A). Extensive transient and stable transfections in SMC treated with atRA revealed weak activation of the native Akap12b promoter constructs carrying the RARE (data not shown). To determine whether the Akap12b RARE is more responsive to atRA in isolation, we multimerized the sequence and placed it upstream of a thymidine kinase minimal promoter ( Figure 3A). Robust activation of a DR-5 RARE from the Rarb gene is evident with atRA stimulation (data not shown). However, the Akap12b RARE is only weakly (,2-fold) activated with atRA stimulation ( Figure 3B). This level of activation is comparable to that seen with carbonic anhydrase (Ca2), a known retinoid responsive target gene containing a similar DR-2 RARE as that seen in the Akap12b RARE [30] ( Figure 3B). Mutagenesis experiments failed to reveal a clear-cut dependence on the DR-2 RARE (data not shown), likely because of the weak level of induction. Nevertheless, ChIP assays consistently showed enrichment for acetylated histone H3 within the region encompassing the DR-2 RARE, suggesting that the chromatin landscape is modified to a transcriptionally competent state with retinoid stimulation ( Figure 3C). Interestingly, we saw little evidence of retinoic acid receptor alpha enrichment in this region consistent with the weak activation of the RARE. These results reveal a mild, though consistent, activation of the Akap12b RARE and a transcriptionally competent chromatin landscape following atRA stimulation.

atRA-Induced AKAP12 Protein Expression and Association with PKA
To assess expression kinetics of AKAP12 protein, we performed Western blotting on extracts of PAC1 SMC treated with atRA. Results demonstrated increases in AKAP12b protein as early as 12 hr post-atRA treatment with levels gradually decreasing at 72 hr ( Figure 4A). Similar induction of AKAP12 protein was seen with AM80 treatment ( Figure S2). Virtually no change in expression of smooth muscle calponin (CNN1) was seen with atRA treatment ( Figure 4A). Immunofluorescence microscopy of PAC1 SMC showed atRA increases AKAP12 around the nucleus and at the periphery of the cell ( Figure 4B, panels a versus d). The expression of the regulatory II (RII) alpha subunit of PKA, to which AKAPs bind [31], displayed a perinuclear localization of expression with little to no changes in level or cellular distribution following atRA stimulation ( Figure 4B, panels b versus e, and data not shown). Importantly, the perinuclear expression of AKAP12 and RII alpha appeared to overlap in the perinuclear region only ( Figure 4B, panels c versus f). To extend these data further, we performed an RII alpha overlay assay in PAC1 SMC treated with either DMSO vehicle or atRA for 12 and 24 hr. These results demonstrated a dramatic induction of an AKAP12b-radiolabeled RII alpha complex with atRA treatment ( Figure 4C). Several lower molecular weight AKAPs showed little, if any, induction suggesting that the stimulatory effect of atRA is specific to the AKAP12 locus in SMC. Interestingly, AKAP12b over-expression studies revealed increases in downstream targets of PKA, namely CREB activation as well as elevated phosphorylation of vasodilator-stimulated phosphoprotein ( Figure S3). These results extend our Akap12bmRNA expression studies to the protein level and indicate a close association between AKAP12b and PKAmediated signaling in SMC.

AKAP12b Overexpression Attenuates SMC Growth
Studies in cancer cell lines suggest that both AKAP12a and b attenuate proliferation [24,32]. To evaluate the potential of AKAP12b over-expression to elicit growth suppressive properties in the context of SMC, we generated PAC1 SMC clones stablytransfected with doxycycline-inducible AKAP12b. Treatment of control (empty vector alone) stable cell lines with doxycycline showed no change in cell growth indicating there was no intrinsic growth inhibitory effects of doxycycline in these cells at a concentration of 1 mg/ml (data not shown). Cells carrying a Myc-tagged AKAP12b transgene showed robust expression of AKAP12b after 1 day of doxycycline treatment with levels AKAP12b persisting over the entire time course of study ( Figure 5A). Importantly, we showed doxycycline-dependent cell growth inhibition in three independent AKAP12b expressing cell lines as compared to the same cells where AKAP12b was not overexpressed ( Figure 5B). To evaluate the effects of AKAP12b in human SMC, we transduced HCASMC with adenovirus carrying AKAP12b under control of the CMV promoter (Ad-AKAP12b). Efficient over-expression of AKAP12b was shown over a 5 day time course by Western blot analysis ( Figure 5C). Similar to results seen in stably-transfected PAC1 SMC, HCASMC over-expressing AKAP12b showed a significant reduction in cell number ( Figure 5D). These results are consistent with the known growth suppressive effects of AKAP12 and suggest that retinoids may inhibit SMC growth, at least in part, through the induction of AKAP12b. Attempts to knock down retinoid-induced levels of AKAP12bmRNA were unsuccessful thus precluding rescue studies relating to SMC proliferation (data not shown).

Expression of AKAP12 in Vascular Lesions
A hallmark of a tumor suppressor gene is reduced expression within tissues associated with accelerated growth. To examine expression of AKAP12 in vascular lesions associated with SMC growth, we used a mouse model of neointimal formation [33] combined with Ki-67 staining. IgG control stained vessels revealed no background staining ( Figure S4). Uninjured carotid arteries exhibit abundant AKAP12 and no cell proliferation consistent with the contractile phenotype ( Figure 6A, 6D and Figure S4). When such vessels are subject to a partial ligation injury [33], Ki-67 positive cells increase in the media and neointima of 1 week (20.87% Ki-67 positive cells, Figure 6E) and 3 week (4.5% Ki-67 positive cells, Figure 6F) vessels. In general, and consistent with data in the cancer field [34], Ki-67 positive cells showed weak AKAP12 staining, especially in 3 week injured vessels where a prominent neointima is manifest ( Figure 6E, 6F). We also noted dramatic reductions in AKAP12 staining in the vessel wall following complete ligation of the carotid artery ( Figure S4). Importantly, human atherosclerotic lesions showed virtually no AKAP12 expression within the neointima ( Figure 7A, 7D). Such low AKAP12 expression correlates with reductions in CNN1 ( Figure 7B, 7E), a SMC differentiation marker known to be reduced in atherosclerosis [35]. Similar findings have been seen in multiple independent atheromas of varying severity (data not shown). To rule out an intrinsic loss in immunoreactivity within the neointima of these vessels as an explanation for the loss in AKAP12 and CNN1 expression, adjacent sections were analyzed for macrophage content using the Ham56 antibody ( Figure 7C, 7F). These results showed immunoreactive macrophages within the neointima indicating that loss in AKAP12 and CNN1 within the neointima is not a consequence of some intrinsic defect in immunolocalization. Thus, AKAP12 expression is attenuated in both experimental and human conditions of neointimal formation and such decreases appear to correlate with elevated SMC proliferation.

Discussion
Over 50 AKAPs have been defined in the human genome. AKAP functions include the ability to compartmentalize multiprotein complexes in order to specify unique spatio-temporal signaling events involving PKA and other signaling moieties [31]. Here, we show that the natural retinoid, atRA, and several synthetic retinoids elicit rapid and robust induction of a specific isoform of AKAP12 (AKAP12b) in vascular SMC. The increase in AKAP12b is also seen in intact vascular tissue exposed to therapeutic levels of atRA. Although we identified a conserved DR2-RARE located in the proximal Akap12b promoter, transient and stable transfection studies revealed only weak (,2-fold) activation of the Akap12b RARE with retinoid stimulation. Nevertheless, actinomycin D and ChIP assays suggest that retinoid signaling converges at the Akap12b promoter to effect gene transcription. A sensitive assay that detects pan-AKAP interactions with the PKA regulatory subunit II alpha indicates that AKAP12 is the only AKAP in SMC exhibiting dramatic increases in expression with retinoid treatment. Directed AKAP12b expression resulted in enhanced PKA activities and attenuated SMC growth. Finally, mouse and human vascular occlusive diseases were associated with reduced AKAP12 expression. Together, these studies establish AKAP12b as a strategic target of retinoid signaling thus providing a framework for further evaluation of this retinoid-AKAP12 axis in vascular SMC growth control and perhaps other cell types responsive to retinoid signaling.
AKAP12 (aka SSeCKS and Gravin) has demonstrable growth suppressive properties and is down-regulated in a variety of human neoplasms supporting the idea that AKAP12 is a bonafide tumor suppressor gene [32,34,[36][37][38][39][40]. However, there is little information on the role of AKAP12 in non-neoplastic cell growth inhibition. Here, we demonstrate that the AKAP12b isoform reduces human and rodent vascular SMC growth and that such growth inhibition correlates with elevated PKA activity. Previous studies have demonstrated PKA-mediated signaling in the inhibition of vascular SMC growth as well as neointimal formation [41][42][43]. The precise signaling mechanisms underlying PKA-mediated SMC growth inhibition are unknown but likely relate to PKA redistribution within cells via AKAPs such as AKAP12. In this regard, we and others have demonstrated context-dependent and isoform-specific localization of AKAP12 within cells [29,44,45]. Interestingly, results shown in this report suggest that retinoid-induced AKAP12 concentrates in the peri-nuclear region in close proximity to PKA ( Figure 4B). It is possible that such peri-nuclear localization of PKA may direct nuclear events such as the phosphorylation of CREB, which is known in some contexts to mediate SMC growth inhibition [46]. Further work is necessary to determine whether retinoid signaling via AKAP12 exerts effects on transcription factors, such as CREB, that may mediate SMC growth suppression. In this context, it will be informative to interrogate CREB factor binding, genome wide, following atRA stimulation of SMC to identify potentially important downstream target genes that may mediate AKAP12dependent SMC growth inhibition.   Tumor suppressors are thought to normally effect growth inhibition. Although several tumor suppressor genes have been shown to be retinoid-inducible [47], none exhibit the rapid onset and high level induction observed with AKAP12b. It will therefore be informative to test whether treating neoplastic cells with natural or synthetic retinoids induces AKAP12b to counter-balance the documented silencing of the AKAP12a isoform [39,40].
The Akap12 locus comprises three independent promoters controlling unique isoforms with distinct intracellular localizations [29]. Such diversity in Akap12 isoform expression exemplifies the complexity of the human genome with a total gene count only mildly greater than that in simpler vertebrate species. Limited information exists with respect to differential control of each Akap12 promoter with virtually all work reporting effects on the proximal, Akap12a promoter. Initial studies demonstrated the importance of serum response factor-binding CArG boxes in the regulation of Akap12a promoter activity [48]. Adjacent GC-rich sequences in the Akap12a promoter were shown to undergo hypermethylation and gene silencing in various cancer cell lines; in contrast, the Akap12b promoter exhibited less gene silencing in cancer cells [34,39,40]. A subsequent study demonstrated a methylation-independent mechanism for Akap12a promoter silencing through the recruitment of HDAC1 [49]. More recently, the anti-proliferative agent dexamethasone was shown to weakly (2-fold) activate the Akap12a promoter without influencing Akap12b promoter activity, though no glucocorticoid responsive elements were reported [37]. We provide evidence here for a conserved RARE in the upstream Akap12b promoter region that is weakly responsive to retinoids. The DR-2 RARE in the Akap12b promoter is very similar to that of the known retinoid-response gene Ca2 [30], which similarly exhibits weak activation with atRA stimulation ( Figure 3B). Although we could not show consistent binding of a retinoic acid receptor to the RARE containing region, several lines of evidence suggest that retinoid receptors target this or perhaps a more distal Akap12b promoter region. First, actinomycin D experiments showed that the induction of Akap12b by atRA proceeds in a RNA polymerase II-dependent manner. Second, three synthetic retinoids that directly bind and activate retinoid receptors each induced Akap12bmRNA expression. Third, a ChIP assay revealed retinoid-mediated enrichment of acetylated histones in the region encompassing the Akap12b RARE. The weak activation of Akap12b promoter with retinoid treatment may imply the absence of key cofactor modifications needed to fully activate the promoter. Alternatively, there may exist remote regulatory elements, as reviewed elsewhere [50], controlling retinoid-mediated transcription of Akap12b from a distance. Methods in bacterial artificial chromosome transgenic mouse technology and related recombineering would be ideal approaches to address the latter possibility. Whatever the full extent of the mechanism(s) may be, this study clearly demonstrates that the increase in Akap12 with retinoid treatment proceeds through Akap12b.
AKAP12 is not the only AKAP shown to suppress SMC growth responses. AKAP5 (aka AKAP79/AKAP75/AKAP150 in human, bovine, rat respectively) was previously demonstrated to inhibit SMC growth in vitro and, similar to our findings here, AKAP5 stimulated CREB-dependent transcriptional activity [51]. Moreover, local delivery of AKAP5 to the balloon-injured vessel wall reduced the extent of neointimal burden [51]. Since, as shown in this report, AKAP12 is reduced in human and rodent vascular lesions, one would not expect a compensatory, confounding influence of AKAP12 on AKAP5-mediated effects. In this context, there is similarity in key amino acid sequences between AKAP12 and AKAP5 [52]. Moreover, we have observed AKAP12 and AKAP5 share similar flanking genes suggesting that these two AKAPs are paralogs related via a segmental chromosomal duplication ( Figure S5). Despite the functional and genomic similarities between AKAP12 and AKAP5, only AKAP12b is induced with retinoids as we were unable to show similar induction of AKAP5 (data not shown).
In summary, the results of this report document retinoidinduced stimulation of a specific AKAP12 isoform which exhibits growth suppressive properties, most likely via PKA-mediated signaling. Future studies should evaluate the extent of neointimal formation and the effects of retinoids in suppressing such growth in mice where the Akap12 gene is genetically deleted [26]. Finally, effects of retinoids on AKAP12b expression should be evaluated in other disease contexts where cell proliferation is manifest (e.g., cancer).

Treatment of cells or animals with retinoids
Rat pulmonary artery SMC (PAC1) were grown and maintained as described previously [53]. Primary-derived rat aortic SMC (RASMC) were obtained from adult thoracic aorta of male Sprague-Dawley rats as described [54], grown in Dulbecco's modified Eagles Medium supplemented with 10% fetal bovine serum (FBS), and used between passage number 10-20. Human coronary artery SMC (hCASMC) were obtained from Cascade Biologics (Portland, OR) and grown in commercially-supplied growth medium per manufacturer's specifications. In all SMC cultures, we routinely validate their phenotype with a panel of SMC-restricted markers, including the SMC-restricted myocardin transcription factor [55,56]. For retinoid stimulation, cells were synchronized for 24 hr in 0.25% FBS and then stimulated in fresh medium containing 0.25% FBS for the indicated times with 2610 26 M of atRA, 13-cis RA, or 1 mM of one of three synthetic retinoids (AM80, BMY-46561, RAR agonists and BMS-188649, an RXR agonist). Control cells received 0.1% dimethylsulfoxide (diluent for retinoids) in medium containing 0.25% FBS. For in vivo experiments, we introduced either 10 mg/kg atRA in corn oil or an equivalent volume (,50 ml) of corn oil alone by oral gavage to adult male FVB/N mice (n = 4 per treatment) and collected blood by intra-cardiac puncture 6 hr later for the determination of serum levels of atRA and its two major stereoisomers (9-cis RA and 13-cis RA) by HPLC as described previously [11]. The same mice were euthanized 18 hr later and aortas (stripped of endothelium and adventitia) harvested for total RNA isolation (as above) and RT-PCR with isoform-specific primers as described [29].

Adenoviral construction
Full-length Akap12b cDNA with a carboxy-terminal FLAG epitope and LacZ were each cloned into pENTR/D-TOPO (Invitrogen) and then recombined into pAd/plDEST (Invitrogen) using LR clonase to create the Ad-AKAP12b or Ad-LacZ adenoviral constructs as described previously [48]. Viral production and titering were also done as described [48].
Western blotting and R2 overlay assays PAC1 SMC were synchronized with 0.25% FBS for 24 hr before stimulation with either 2610 26 M atRA or 0.1% DMSO and protein extracted for Western blotting as described [48]. Primary antibodies to AKAP12 [22] (1:2000 dilution), smooth muscle calponin (hCP, Sigma; 1:10,000 dilution) or alpha tubulin (T5168, Sigma; 1:2000) were applied to membranes followed by species-appropriate HRP-conjugated secondary antibodies. To evaluate the interaction of AKAP12 with the regulatory subunit of PKA, PAC1 SMC were stimulated with 2610 26 M atRA or 0.1% DMSO for 12 and 24 hr and 75 mg total protein resolved in a large format 5% SDS-PAGE gel. The gel was then electro-blotted to a PVDF (Immobilon-P) membrane and dried for a radiolabeled RII overlay as described [57].

Immunofluorescence microscopy
Cells were grown on glass chamber slides (Labtek), treated for retinoid stimulation as above, and at the indicated times, fixed, processed, and visualized with a fluorescence microscope as previously described [58]. Primary antibodies used were polyclonal rabbit anti-AKAP12 [22] and mouse anti-PKA RII alpha (BD Transduction Laboratories, Cat # 612242).

Bioinformatics
Identification of a conserved RARE around the promoter of Akap12b isoform was performed using VISTA (http://www-gsd. lbl.gov/vista/index.shtml)(citation) by searching for pairs of consensus half sites (RGGTCA) and characterized variants situated within several nucleotides of each other that are conserved in sequence and relative position within a 20 kb window centered on the human, rat, and mouse Akap12b promoters [4].

Cell transfections and reporter assay
To test and compare the Akap12b RARE for functionality, oligonucleotide primers comprising 3x multimerized RAREs from Akap12b, Rarb, or the retinoid-response gene, carbonic anhydrase (Ca2) were inserted upstream of the minimal thymidine kinase promoter of tk-luciferase. PAC-1 SMC were transfected with the indicated construct, made quiescent for 24 hours using 0.25% FBS-containing DMEM, and subsequently stimulated with 2610 26 M atRA for 24 hours. Dual luciferase assays were performed as previously described [29]. Data shown are representative of at least three independent experiments. Error bars represent the standard deviation from the mean. Data were analyzed by one-way ANOVA and Tukey's post-hoc test using GraphPad Prism software. A probability value less than 0.05 was considered statistically significant. In some experiments, similar luciferase assays were done in PAC1 SMC using a CREB reporter plasmid (PathDetect System, Stratagene) transfected with either an empty control vector or one carrying the AKAP12b open reading frame.

Cell transduction and growth assay
Stable cell lines expressing doxycycline-inducible, Myc-tagged AKAP12b were generated using the T-Rex System (Invitrogen) in PAC1 SMC according to the manufacturer's specifications. To induce expression of AKAP12b, cells were treated every other day with 1 mg/ml doxycycline (Sigma) and levels of AKAP12b measured with a Myc antibody. HCASMC were grown to subconfluence in 6-well dishes and transduced with 300 infectious units per cell (ifu/cell) of either CMV-driven LacZ (Ad-LacZ) or CMV-driven AKAP12b (Ad-AKAP12b) in 2% FBS as described [48]. Following overnight culture in 2% FBS, culture medium was changed and replaced with 0.25% FBS for 24 hr to synchronize the cells. Cells were then stimulated with full growth medium and the number of trypan blue negative cells manually counted with a hemocytometer. At least three independent measures per time point were made in two independent experiments. Results are presented as the average of three replicates from one experiment 6 the standard error of the mean.

Ligation injury model
Male C57BL/B6 mice (30 g) were subject to partial ligation of the left carotid artery [33] and FVB/N mice were injured by complete ligation of the common carotid artery [59]. One and three weeks after injury, animals were perfusion fixed with neutral buffered formalin, vessels removed and processed, and sections (5 mm) of injured arteries distal to the ligature stained by immunohistochemistry with a polyclonal antibody to AKAP12 [22] or an antibody to Ki-67 to detect proliferating cells of the vessel wall. All animal studies were approved by the University of Rochester's Institutional Animal Care and Use Committee.

Immunohistochemistry
Mouse vessels were fixed in 4% buffered paraformaldehyde and paraffin embedded. Samples of human coronary vessels with variable degrees of atherosclerosis were obtained from archived tissues in the University of Rochester Medical Center's Pathology Department. All tissues were sectioned at 5 micron thickness and slides were deparaffinized and rehydrated to PBS (pH, 7.4). Endogenous peroxidase activity was quenched using 3% aqueous hydrogen peroxide for 10 minutes and antigen retrieval was performed (for CNN1) utilizing heat induced epitope retrieval in 0.05% citraconic anhydride as described [60]. Primary antibodies (and their dilution) were as follows: rabbit polyclonal anti-AKAP12 (1:500), anti-CNN1 (DAKO; 1:1000), Ham56 (DAKO; 1:1000), and Ki-67 (DAKO, 1:100). Appropriate secondary biotinylated antibodies (Vector BA-2000, BA-1000, or DAKO rabbit anti-rat for Ki-67) were applied for 30 minutes at room temperature following washes in TBST. Immunoreactive signals were revealed by a 30 minute dark exposure to either alkaline phosphatase (Vector AK-5000) followed by Vector Red (Vector SK-5100) chromagen or horseradish peroxidase (Vector PK-6100) followed by diaminobenzidine chromagen. Specific immunoreactive prod-uct was indicated by inclusion of control isotype-matched IgG on adjacent sections that were handled in exactly the same manner as those stained with primary antibodies. Figure S1 A, PAC1 SMC were treated with 1 mM AM80 for the indicated times or DMSO diluent and Akap12b mRNA measured by qRT-PCR (n = 3). B, PAC1 SMC were treated with 1 mM AM80 for 6 hr and immediately thereafter were exposed to 1 mg/ ml actinomycin D or equal amount of water for the indicated times and Akap12b mRNA measured by qRT-PCR as in panel A. Akap12b mRNA was normalized to internal control Gapdh with the control (DMSO or 0 h) ratio set to a value of 1. (TIF) Figure S2 Extracts of PAC1 SMC were treated with DMSO or the synthetic retinoid, AM80 (1 mM), for 24 hrs and total protein analyzed for AKAP12, ACTA2, CNN1 and TUBA1 (control) proteins. (TIF) Figure S3 A, PAC1 SMC were co-transfected with a CREB reporter and either an empty vector or an AKAP12b expression plasmid to assess CREB activity in a luciferase assay. Results are expressed as normalized luciferase (see Methods). B, PAC1 SMC were co-transfected with EGFP (control) 6 AKAP12b and phosphorylation of VASP assessed by immunoblotting. (TIF) Figure S4 Non-immunogenic IgG control antisera was applied to uninjured right carotid (A), 7 day injured left carotid (B), and 21 day injured left carotid (C). Panels D-F represent AKAP12 staining of uninjured carotid (D) and femoral (E) artery or a 7 day complete ligation injured carotid artery (F). Note loss of AKAP12 staining in the media of the injured vessel (F) as compared to normal vessels (D,E). (TIF) Figure S5 Schematic shows evidence of segmental chromosomal duplication with percent amino acid homologies between AKAP5 and AKAP12 and their paralogous flanking genes. (TIF)