Retinoids Inhibit the Actions of Angiotensin II on Vascular Smooth Muscle Cells

Apart from causing vasoconstriction, angiotensin II (Ang II) affects the response of vascular smooth muscle cells (VSMCs) to hypertension, arteriosclerosis, or vascular damage of chronic transplant rejection.¹²³ Ang II may act as a comitogen for the proliferation or hypertrophy of VSMCs. It also promotes vascular inflammation and extracellular matrix deposition.³⁴ Ang II stimulates cell proliferation via the activator protein-1 (AP-1) complex by induction of proto-oncogene expression, such as c-fos and c-jun, which constitute the AP-1.⁵ This type of activation is mediated via the Ang II receptor subtype AT₁ but not AT₂.⁶ Ang II also activates growth and differentiation factors, such as transforming growth factor (TGF)-β, platelet-derived growth factor (PDGF), fibroblast growth factor (FGF), and others that in turn modulate the cellular response to Ang II.⁷ The mechanisms by which Ang II stimulates growth and vasoactive factors have not been clarified in detail. AP-1 may again play a role,⁸ although additional Ang II pathways have been described, ie, c-myc, serum response element, and cAMP response element.²⁹¹⁰

Retinoids, natural and synthetic derivatives of vitamin A, influence cellular differentiation in organ development.¹¹¹² They potently inhibit proliferation of a variety of cell types.¹³¹⁴¹⁵ They act, not exclusively but in part, via inhibition of AP-1–induced activation.¹⁶ Inhibition of AP-1 activity by retinoids is mediated via different mechanisms, which are cell- and context-dependent. These include direct inhibition of c-jun, c-fos downregulation, interruption of c-fos/c-jun dimerization, or inhibition of binding of AP-1 to DNA.¹⁵¹⁶

Retinoid signals are transduced by 2 families of receptors, ie, retinoid acid receptors (RARs) and retinoid X receptors (RXRs). RAR and RXR heterodimers bind to polymorphic cis-acting retinoid acid (RA)–responsive and retinoid X–responsive elements.¹⁷ RXRs also form heterodimers with other receptors of the steroid receptor family, such as the peroxisome proliferator–activated receptors and the vitamin D and thyroid hormone receptors, thus enabling crosstalk between different regulatory pathways.¹⁸ Although the vascular system is not considered a traditional target for the retinoids, VSMCs are sensitive to their action.¹³ Expression of RARs by VSMCs has been demonstrated on the RNA level. Functional studies were performed after RA transduction using reporter gene assays.¹³ The antiangiogenetic properties of retinoids have long been known. Retinoids effectively block neointima proliferation induced by balloon arterioplasty injury in the rat.¹⁴¹⁹²⁰ The consideration that Ang II has important nonvasoconstrictive effects on VSMCs and that VSMCs constitute a target for retinoids prompted us to examine whether retinoids interfere with the action of Ang II on VSMCs.

Materials and Methods

Culture of VSMCs

Primary rat aortic SMCs were prepared according to Owens et al²¹ and maintained as previously reported in our laboratory.²² Cells were used between passages 3 and 10 out of 3 different isolates. Cells were rendered quiescent after reaching the appropriate degree of confluence by placing them in DMEM/F12 containing 0.5% FCS for 24 hours unless otherwise stated.

Cell Counts

VSMCs (2×10⁴) were plated per well in 24-well dishes, stimulated every 48 hours, and counted in hemocytometers (Neubauer type; Brand).

Measurement of DNA and Protein Synthesis

Cells were plated in 96-well dishes (10⁴ per well) and rendered quiescent for 24 hours. The cells were treated with agonists for another 24 hours and labeled with tritiated thymidine or proline within the final 4 hours of the experiment, as described earlier.²²

Transient Transfection and Luciferase Assay

Confluent VSMCs in 6-well dishes were transfected with 1 μg/well of the reporter gene construct PathDetect AP-1 cis-reporting system (pAP-1-Luc; Stratagene) with FuGENE6 transfection reagent according to the manufacturer’s protocol (Roche Diagnostics). Transfections were normalized against β-galactosidase activity from a cotransfected (1 μg/well) pSV-β-galactosidase control vector (Promega). In cotransfection experiments, plasmid cytomegalovirus (pCMV)-Rα and pCMV-RXRα constructs were used (kindly provided by Professor Pierre Chambon, Institut de Génétique et de Biologie Moléculaire et Cellulaire [IGBMC], Illkirch, France). A human AT₁-promoter/exon-1 fragment (AT₁-luc) was amplified via polymerase chain reaction (PCR), sequenced, and cloned into a luciferase vector essentially as in Holzmeister et al.²³ The AP-1 consensus site of AT₁-luc (ΔAP-1) was point-mutated according to Holzmeister et al.²³ After 1-hour pretreatment with 1 μmol/L all-trans RA (Sigma), 1 μmol/L Ang II (Sigma) was added for another 4 (AP-1) or 21 (AT₁) hours, and the luciferase activity was assayed according to the manufacturer’s protocol (Promega). All experiments were repeated 3 to 5 times; assays were performed at least in duplicate or triplicate on 3 different cell isolates.

AT₁ receptor binding was determined in confluent monolayers of VSMCs treated with 1 μmol/L all-trans RA for 24 hours, as described,²⁴ using unlabeled AT₁-specific ligand DuP753 (Losartan) in concentrations ranging from 10 fmol/L to 100 μmol/L and 50 nmol/L of tritiated DuP753 (DuPont LS).

RNA Isolation and Reverse Transcriptase– PCR Analysis

Total RNA was isolated using RNeasy extraction columns (Qiagen). After reverse transcription, quantitative PCR was carried out in triplicates using strand-specific primers for AT₁ receptor (33 cycles), TGF-β₁ (29 cycles),²² and c-fos (31 cycles).²⁵ For quantification, we used deletion mutants of the respective genes sharing the same primer sequences.²² The results were expressed as ratio of the optical densities of wild-type versus mutant cDNA signal.²²

Immunoblotting and Immunoprecipitation

Whole-cell extracts of VSMCs were prepared from confluent transfected cultures and immunoprecipitated with monoclonal anti-mouse RXRα [4RX3A2] or RARβ [Ab8β(F)2] antibodies (IgG1 subclass) covalently linked to protein A beads according to Gaub et al.²⁶ Proteins with or without previous immunoprecipitation were fractionated by SDS-PAGE, electrotransferred onto nitrocellulose filters²⁶ (Schleicher & Schuell), and immunoprobed with polyclonal antibodies raised against RARα [RPα(F)], RARβ[RPβ(F)2], or RXRα[RPRXα(A)], as described,²⁷ and detected via chemoluminescence.²⁷ The specificity of the reactions was checked by depleting the specific antibodies from the antiserum. COS-1 cell extracts transfected with full-length RARα, RARβ, or RXRα expression vectors served as positive controls, as described.¹¹

Statistics

Results are presented as mean±SEM. Data were evaluated using the Mann-Whitney test. The zero hypothesis was rejected at P<0.05.

Results

VSMCs Express RAR and RXR Protein

To examine whether VSMCs contain RAR protein, we analyzed lysates of VSMCs by immunoblotting with polyclonal anti-RAR and anti-RXR antibodies. The presence of RARα protein is documented in Figure 1A. RARα protein was consistently demonstrated by immunoblotting in 10 and 20 μg of lysate. The 51-kDa band migrated at the same position as the RARα protein overexpressed in control COS cell lysates, which had been transiently transfected with a total RARα construct (pCMV-RARα).

RARβ could be demonstrated in VSMCs on the mRNA level only (data not shown). The polyclonal anti-RARβ antibody failed to detect RARβ protein in VSMC lysates. RARβ could also not be detected after immunoprecipitation with specific RARβ antibodies or after stimulation of VSMCs with 1 μmol/L all-trans RA (data not shown). In contrast, RARβ could be demonstrated in transiently transfected RARβ-COS control cell lysates.

RXRα protein concentration in VSMC lysates was below the detection limit of immunoblotting when a polyclonal anti-RXRα antibody was used. Immunoprecipitation yielded a clear signal for RXRα in lysates of VSMCs. Controls were lysates of transiently transfected RXRα COS cells (Figure 1B).

All-trans RA Inhibits Ang II–Induced Proliferation of VSMCs

Figure 2A shows the effect of 10 nmol/L or 1 μmol/L all-trans RA on VSMCs treated with 1 μmol/L Ang II in the presence of 0.5% FCS over 10 days. Dose-dependently lower rates of cell proliferation were observed in the cultures treated with all-trans RA plus Ang II compared with the cultures treated with Ang II alone.

Figure 2B indicates that all-trans RA completely abrogated the Ang II–induced DNA synthesis, as assessed by thymidine incorporation, but no effect of all-trans RA on basal DNA synthesis was noted. Similarly, protein synthesis in response to 1 μmol/L Ang II was abrogated by 1 μmol/L all-trans RA (Figure 2C), but basal protein synthesis was not affected. In an additional 48-hour experiment, the cumulative proline incorporation into the precipitable fraction was determined in the presence of 1 μmol/L all-trans RA, 1 μmol/L Ang II alone, or both compounds. The results showed significantly less proline incorporation in the presence of all-trans RA (data not shown).

RAR- and RXR-Specific Agonists Block Ang II–Induced Proliferation of VSMCs

To test whether inhibition of cell proliferation by all-trans RA is mediated via RAR- or RXR-dependent pathways, cells were treated with the RAR-specific agonists AM580 and all-trans RA as well with the RXR-specific agonists BMS649 or 9-cis RA (all reagents 1 μmol/L), respectively. In a 7-day experiment (Figure 3A), 1 μmol/L Ang II stimulated VSMC proliferation 3-fold over controls. In incubations without Ang II, retinoids (all-trans RA, AM580, BMS649, and 9-cis RA) induced a small increase in cell number over control, which was not significant. However, Ang II–induced proliferation was significantly inhibited by all retinoids. There was no difference between RAR- and RXR-specific agonists. To exclude loss of receptor specificity at high concentrations (1 μmol/L), we tested AM580 and BMS649 at 3 different concentrations (1 μmol/L, 100 nmol/L, and 1 nmol/L). The inhibitory efficiency was equivalent (Figure 3B).

All-trans RA Inhibits Ang II–Dependent Activation of TGF-β₁ Expression

After 4 hours of exposure to 1 μmol/L Ang II, the steady-state expression of TGF-β₁ mRNA was significantly increased compared with controls. All-trans RA alone (1 μmol/L) reduced basal TGF-β₁ mRNA concentration and completely abrogated the Ang II–induced stimulation of TGF-β₁ mRNA (Figure 4).

Inhibition of Ang II by All-trans RA Is Mediated Via AP-1

Ang II induced a 24-fold stimulation of luciferase activity of the AP-1 plasmid within 4 hours (Figure 5A). Pretreatment (1 hour) with 1 μmol/L all-trans RA reduced Ang II–dependent AP-1 induction by ≈30%. Costimulation of Ang II–treated cells with 1 μmol/L Losartan abrogated luciferase activity to control levels (data not shown). Dose-dependent effects of all-trans RA on AP-1-induction could not be detected, presumably because the overall effect was small. The anti–AP-1 activity of all-trans RA in response to Ang II, however, was markedly increased after cotransfection of plasmids encoding CMV promoter–driven retinoid receptors RARα and RXRα with the AP-1 reporter gene plasmid (Figure 5B). Cotransfection of both retinoid receptors reduced AP-1 luciferase activity by ≈50%. When different amounts of RARα were cotransfected, a clear dose-dependency could be demonstrated (Figure 5C).

All-trans RA Abrogates the Induction of c-fos Expression by Ang II

Treatment with 1 μmol/L Ang II for both 30 minutes and 1 hour stimulated c-fos RNA expression significantly. Pretreatment with 1 μmol/L all-trans RA completely abrogated the induction of c-fos mRNA by Ang II (data shown after 1-hour Ang II treatment only; Figure 6).

All-trans RA Reduces the Basal Expression of the AT₁ Receptor

To examine whether all-trans RA changes AT₁ receptor gene expression and thus alters Ang II sensitivity, we determined steady-state levels of AT₁ receptor mRNA. Reduction in AT₁ receptor gene expression reached statistical significance after 24 hours of incubation with 1 μmol/L all-trans RA (P<0.05, Figure 7A).

All-trans RA Downregulates AT₁ Receptor

After 24 hours of treatment with 1 μmol/L all-trans RA (Figure 7B), B_max decreased by 45%, from 995±51 to 564±31 cpm. EC₅₀ decreased from 0.31 to 1.1 nmol/L, indicating a negligible rise in affinity of AT₁ receptors to DuP753 when treated with all-trans RA.

Inhibitory Action of All-trans RA on Ang II–Induced AT₁ Receptor Promoter Activity Is Dependent on AP-1

The effects of 1 μmol/L all-trans RA on the Ang II–induced AT₁ receptor promoter activity were assessed in transiently transfected VSMCs. All-trans RA had no significant effect on the low basal promoter activity of the wild-type AT₁ receptor promoter under quiescent conditions. VSMCs were transfected with an AT₁ receptor promoter luciferase reporter gene construct either with (AT₁-luc) or without (AT₁-luc ΔAP-1) a functionally active AP-1 consensus site. All-trans RA abrogated the Ang II–induced (1 μmol/L) increase of luciferase activity in the wild-type (AT₁-luc) AT₁ promoter construct (Figure 7C). In transfected VSMCs carrying the construct without a functional AP-1 site (AT₁-luc ΔAP-1), all-trans RA was no longer able to inhibit the Ang II–dependent stimulation of AT₁ promoter activity (Figure 7D).

Discussion

The data document that all-trans RA inhibits the effects of Ang II in VSMCs. This inhibition may involve AP-1 repression and downregulation of the AT₁ receptor. VSMCs express both RAR and RXR subtypes on the RNA level¹³ and, as shown here, on the protein level. These observations suggest that VSMCs may respond not only to RAR ligands, such as all-trans RA, but also to RXR ligands, such as 9-cis RA or BMS649. Expression of RXRs allows for crosstalk with other receptors of the steroid superfamily, such as thyroid hormone, vitamin D, and peroxisome proliferator activator receptors. These receptors also act on cardiovascular tissues and cell growth.¹⁸²⁸

The induction of growth in VSMCs by Ang II has been studied intensively.²⁹³⁰ Ang II may evoke hypertrophic or hyperplastic responses, depending on the balance of antiproliferative and proproliferative autocrine factors, ie, TGF-β, PDGF, FGF, and others.³⁰³¹³² Our data confirm that Ang II stimulated c-fos expression²³¹ time-dependently and AP-1 activity by a factor of about 20 to 30 using a luciferase vector containing AP-1 consensus sites.⁸

Repression of c-fos gene expression has been identified as a mechanism of retinoid inhibition of AP-1 activity.¹⁵³³ All-trans RA has been shown to inhibit the serum-induced stimulation of c-fos and c-jun by serum in mesangial cells¹⁵³³³⁴ and in our VSMCs.

The anti–AP-1 activity of retinoids has been intensively studied and may be cell-type specific.¹⁶³⁵³⁶ All-trans RA reduced Ang II–induced activation of AP-1 modestly but significantly in reporter gene assays. This suggested low natural abundance of the receptors. Cotransfection with an RARα expression vector dose-dependently inhibited Ang II–induced luciferase activity up to 50%, confirming blockade of Ang II–dependent AP-1 activation by all-trans RA.

All-trans RA dose-dependently blocked Ang II–stimulated cell proliferation, as indicated by proliferation markers, such as cell counts or thymidine and proline incorporation. All-trans RA alone generally had no effect on cell proliferation, although, depending on cell preparation and culture conditions, occasionally a minor increase was observed, which was also described by Chen and Gardner.³⁷ In all doses of all-trans RA used in experiments, no sign of toxicity was present; signs of toxicity were only present at ≥10 μmol/L.

To clarify whether RARα inhibits Ang II–dependent AP-1 activation alone or with RARα pathways, we assessed retinoid receptor–specific ligands and cotransfection of either RARα or RXRα expression plasmids. Receptor-specific ligands, such as AM580 (which is specific for RAR) or BMS649 and 9-cis RA (which are specific for RXRs [Figures 3A and 3B]), showed that both receptor ligands were able to block Ang II–induced cell proliferation. RAR-dependent inhibition of VSMC proliferation is well known.¹³¹⁴ Our data support these findings. In our cells, RXR agonists, 9-cis RA (Figure 3A), and BMS649 (Figures 3A and 3B) were as efficient at inhibiting Ang II–induced cell proliferation as RAR agonists. We used a highly specific RXR agonist, BMS649, which cannot isomerize into other retinoids as 9-cis RA. Its K_d for RXRα is 4.9 nmol/L, for RXRβ is 2.9 nmol/L, and for RXRγ is 98 nmol/L. In contrast, the K_d of BMS649 for RARα is 42 500 nmol/L, for RARβ is 17 500 nmol/L, and for RXRγ is 14 000 nmol/L, which is about 10 000-fold higher than for RXRs. Furthermore, in COS-1 cell test systems comparable to Brand et al,³⁸ transient transfections of specific retinoid receptor subtype CAT constructs had shown no stimulation of CAT activity of RAR by BMS649 (C. Zusi, unpublished data, 1991). These data are different from those provided by Neuville et al,¹⁴ who found no RXR-dependent growth inhibition in SMCs. These differences may be explained by the use of 10% serum instead of Ang II and specific SMC populations derived from injured vasculature.¹⁴

Besides retinoid receptor–specific ligands, cotransfection with either RARα or RXRα caused inhibition of AP-1 activation by Ang II. This is in good agreement with studies where both RAR and RXR expression vectors were able to confer anti–AP-1 activity.³⁶ The data do not allow conclusions on differences in potency and efficacy between RAR- and RXR-specific compounds, because the effects varied slightly between different cell preparations, and differences were subtle at best.

Ang II induction of cell growth is modulated by autocrine factors, particularly TGF-β₁, which significantly contributes to the hypertrophic response of VSMCs to Ang II.⁷ Our data of a modest but consistent stimulation of TGF-β₁ gene expression by Ang II are in accordance with the literature, where an increase of 1.5- to 8-fold, depending on the cellular context, was reported.²²³¹³² All-trans RA completely abrogated this induction and even reduced basal TGF-β₁ mRNA levels in these cells. The effects of retinoids on TGF-β are cell-specific: all-trans RA stimulated TGF-β₁ mRNA in chondrocytes, where basal TGF-β is low, and inhibited TGF-β in myocytes, where basal expression is high.³⁹ The TGF-β₁ promoter contains 3 AP-1 sites, which are able to confer AP-1–mediated stimulation of gene transcription.³⁶ Salbert et al³⁶ demonstrated that both RARs and RXRs inhibit the TGF-β₁ gene promoter via inhibition of AP-1. Morishita et al⁴⁰ demonstrated that Ang II–stimulated TGF-β induction in VSMCs is mediated via AP-1 using an anti–AP-1 oligodeoxynucleotide decoy approach. These findings strongly suggest, but do not provide final proof, that abrogation of Ang II–dependent TGF-β induction by retinoids may be mediated via AP-1.

The proliferative response to Ang II is mediated via the AT₁ receptor involving protein kinase C and induction of proto-oncogenes.⁸ Our data of a 50% reduction of AT₁ receptor binding confirm the findings of Takeda et al,⁴¹ who also found a reduction in B_max but not K_d values. Holzmeister et al²³ previously demonstrated that the AT₁ receptor promoter contains a functional AP-1 binding site. It is thus conceivable that downregulation of this receptor by retinoids might be mediated via AP-1. To test this hypothesis, we point-mutated the AP-1 binding site on the human AT₁ receptor promoter. All-trans RA had no significant effect on the low basal AT₁ receptor promoter activity of the AT₁ receptor under quiescent conditions. This is in contrast to findings of Takeda et al,⁴¹ where at 10% FCS, all-trans RA reduced AT₁ receptor promoter activity, but inhibition was not AP-1 dependent. In our study, all-trans RA did not inhibit serum-induced AP-1 activity in the presence of 5% or 10% FCS (data not shown). Additionally, under serum conditions of >2%, no Ang II effects could be observed. This suggests that the AP-1 dependency of retinoid effects may be overrun under high-serum stimulation and prompted us to examine the effects of retinoids under low-serum conditions.

In contrast to the wild-type AT₁ receptor construct, the inhibition of Ang II–dependent promoter activity by all-trans RA was now abolished. These findings support the notion that the AP-1 binding site of the AT₁ receptor promoter mediates Ang II stimulation, as has previously been shown for tissue plasminogen activator stimulation by Holzmeister et al.²³

We have shown that retinoids inhibit the proproliferative action of Ang II on different levels, first by inhibiting AP-1–dependent gene transcription and second by reducing responsiveness to Ang II via reduced expression of the AT₁ receptor. Although retinoids belong to the steroid superfamily, they clearly differ in this respect from glucocorticoids. The latter sensitize VSMCs to vasopressors and upregulate the AT₁ receptor.²⁴⁴²

The effects of retinoids on Ang II are an additional example of their influence on the expression action and of vasoactive factors. Inhibition of endothelin-1, PDGF, inducible nitric oxide synthase, and others by retinoids has been reported in myocytes or mesangial cells.³⁵⁴³⁴⁴ Retinoids also inhibit neointima formation in response to vascular balloon injury in the rat.¹⁴¹⁹²⁰ Retinoids, therefore, are potentially interesting candidate drugs to modulate agonist-induced transcription in the cardiovascular system.

Original received January 26, 2000; resubmission received September 12, 2000; revised resubmission received February 20, 2001; accepted February 20, 2001.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Footnotes