INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

RETINOIC ACID (RA) exerts profound effects on embryonic development (8, 17), cellular growth (6, 20, 21, 29), and differentiation (6, 8). Extracellular matrix (ECM) proteins and their receptors, which often control tissue remodeling (11, 12, 14, 15, 19, 31), are an emerging class of products that are modified by RA treatment (2, 5, 7, 23, 25, 30). By regulating these molecules, the retinoids control both cell differentiation and growth. More recently, emphasis on the vascular remodeling capabilities of RA has been growing in importance as chronic doses of RA are being used for remission of the cancer, acute promyelocytic leukemia (APL; see Ref. 6). Retinoids are also being tested to limit neointimal formation associated with angioplasty, artherosclerosis, cardiovascular transplantation, and hypertension (20, 21).

Antiproliferative actions of RA have been observed in a number of cells (6, 21, 29), including vascular smooth muscle cells (VSMCs) in culture (21). One mechanism for this inhibition could be the triggering of intracellular signals initiated by specific environmental stimuli that are transduced by cell-matrix interactions. Integrins (11, 27), a class of ECM receptors that transduce signals for cell growth and migration, are controlled by RA (5, 7, 25). In addition, their substrates, such as collagen, matrix proteases, adhesion proteins, and other such molecules that control remodeling, have also been reported to be induced by RA (2, 23, 30).

Integrins have been found to be vital for growth and angiogenesis (3, 9) as well as cellular apoptosis (22). They are heterodimers made up of an - and -subunit that combine in different permutations to recognize a range of matrix proteins like laminin, vitronectin, fibronectin, fibrinogen, collagen, von Willebrand's factor, etc. (27). Thus integrin receptors can sense the extracellular environment and trigger specific intracellular pathways to initiate distinct cellular responses.

Treatment of rats with all-trans retinoic acid (ATRA; one of the active isomers of RA that is biologically synthesized from the inactive precursor, vitamin A) prevented neointimal hyperplasia after balloon injury of carotid arteries (20). The effects of ATRA included geometric remodeling of the perimeter of the injured vessel and specific attenuation of VSMC proliferation in the media (20). The mechanisms for this remain largely unexplored. To establish the effect of RA therapy in vascular disease, the action of RA on integrins expressed on VSMC was examined. This study describes an increase in the mRNA and protein for the ₁-integrin subunit in two types of cultured rat VSMC lines after exposure to ATRA. These results were followed up by looking for an increase in integrin ₁- protein in vivo in the aortas of adult rats that were fed therapeutic doses of ATRA. In addition, changes in the adhesion profile of cultured VSMCs to different matrix proteins after treatment with RA were also studied. The results showed a significant increase in attachment and spreading of the cells on fibronectin after exposure of the cells to growth inhibitory doses of ATRA.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell culture. Rat aortic smooth muscle cells (RASMCs; see Ref. 21) and a pulmonary arterial smooth muscle cell line (PAC; see Ref. 26) were obtained from Dr. Joe Miano and were cultured as mentioned in each protocol below. ATRA (Sigma Chemical, St. Louis, MO), and cells being treated with it, were always handled in the dark to avoid photoinactivation as described previously (20, 21).

Northern analysis. RASMC and PAC were grown to 80% confluency in DMEM containing 10% FBS and antibiotics, washed one time with DMEM, and treated overnight with 1 µM ATRA or the same volume of vehicle (DMSO) in DMEM-F-12 media (Sigma) containing 0.1% BSA. Total RNA was extracted from the cells using the Trizol reagent (3 ml/100-mm dish; GIBCO-BRL, Gaithersburg, MD), as specified by the manufacturer. Equal amounts of RNA from each sample (20 µg) were loaded on a formaldehyde agarose gel and were electrophoresed, blotted, and probed with ₁-cDNA as described (16). The ₁-probe was isolated after RT-PCR from the PAC total RNA (28) using the RT-PCR Ready-To-Go-Beads (Pharmacia, Piscataway, NJ) with ₁-specific oligomers 5'-cccagcaagtcccaagtgccatga-3' and 5'-tccacctgcacaggctggggcaac-3'. The product was electrophoresed on a 1% agarose gel, and a single band of expected size (644 bp) was purified with Eluquick beads (Schleicher and Schuell, Keene, NH). It was confirmed to be a ₁-integrin product by Southern analysis (28) using the labeled ₁-specific internal oligomer 5'-agagaacagctcagagatctgca-3' as a probe. The ₁-specific PCR product hybridized to the probe, whereas other DNA fragments on the same membrane did not.

Western analysis. RASMCs and PACs were cultured for 10 days in 100-mm dishes with vehicle (DMSO) or ATRA (1 µM). The serum was reduced to 0.5% after the cells reached 50% confluency. Hormones and media were replaced every 48 h. The cells were washed three times with PBS, and the proteins were solubilized and extracted with 500 µl RIPA buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 0.5% SDS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 1 mM EDTA, and 1× protease inhibitor cocktail obtained from Pharmingen, San Diego, CA). The lysates were used to estimate their protein content with the Bio-Rad DC Protein Assay Reagent (Bio-Rad, Hercules, CA). Equal amounts of protein (25 µg) from each sample were electrophoresed on a 7.5% nonreducing, SDS-polyacrylamide gel with running buffer (1). Care was taken not to reduce the samples with -mercaptoethanol, as the ₁-antibodies do not recognize the reduced antigen. Two types of color-tagged molecular weight markers (Kaleidoscope prestained protein markers from Bio-Rad and Benchmark prestained markers from GIBCO-BRL) as well as native protein markers were included to assess the molecular weights of the products. The gels were transferred to nitrocellulose as described (1). The membrane was treated overnight with primary antibody [anti-CD29 Hamster IgM (Ha2/5, 18) 1 µg/ml; Pharmingen], washed three times, and incubated with mouse anti-hamster antibody (Pharmingen) diluted 1:1,000 for 45 min followed by a treatment with a tertiary mouse anti-hamster horseradish peroxidase. The ₁-integrin bands were developed as described (1). The experiment was repeated with a second ₁-integrin antibody (1:30,000; Chemicon, Temecula, CA), and the membrane was treated with only one secondary antibody (horseradish peroxidase-conjugated goat anti-rabbit, 1:1,000). All gels were washed thoroughly with Tris-buffered saline-Tween 20 (1) and were reexposed to the monoclonal smooth muscle specific -actin antibody (1:5,000; Sigma Chemicals). X-ray films were scanned in a densitometer, and each ₁- as well as the corresponding smooth muscle -actin band were quantitated to normalize for protein loading.

Cell adhesion assay. The adhesion assays were done in 96-well polystyrene microtiter dishes as described (16). Wells were coated with 5 µg/ml human vitronectin (GIBCO-BRL), 25 µg/ml type I rat tail collagen, 5 µg/ml laminin, 5 µg/ml fibronectin, 5 µg/ml fibrinogen, or PBS (10 mm sodium phosphate, pH 7.0, 138 mM NaCl) overnight at 4°C. The PBS-treated wells were used to obtain values for nonspecific binding of the cells to the dishes. RASMCs and PACs were grown with ATRA (1 µM) or DMSO for 5 days in DMEM plus 5% FBS. The cells were lifted and aliquoted on the different matrixes as described (16). There were 10 wells of each matrix used for the four sets of cells [RASMC treated with DMSO, RASMC treated with ATRA (1 µM), PAC treated with DMSO, and PAC treated with ATRA (1 µM)]. Three additional wells from each set were incubated with blocking anti-integrin ₁-antibody Ha2/5, anti-_v3 (F-11; Chemicon; see Ref. 13), or their matching nonspecific immunoglobulins, each at a concentration of 10 µg/ml. Adhesion and spreading were allowed to proceed for 45 min at 37°C in the tissue culture incubator. The wells were then emptied and washed three times with PBS containing CaCl₂ and MgCl₂ (Sigma Chemical). At this time, six different dilutions of unplated cells were added to unused wells in the same dishes. This was used to obtain a standard curve for determining the cell number vs. optical density. The plate was developed to colorimetrically determine the cell number in each well (16). The adherent cell count was estimated using known values in the standard wells (16).

RA administration to rats by oral gavage. ATRA was mixed with corn oil at 1 mg/1.25 ml in the dark as described (20). The suspension was administered to 8-wk-old male Sprague-Dawley rats orally using an 18-gauge gavage needle. Four rats were administered 1 mg/kg RA every other day for 3 wk. At the same time, a similar number of animals (controls) was given corn oil alone. The rats were weighed throughout the experimental period, and the amount of ATRA was adjusted for changes in body weight.

Statistical analysis. All statistical analyses were performed between ATRA-treated samples vs. vehicle-treated control groups using Sigmastat 2.0 software. For antibody blocking in the adhesion assays, the tests were also run between samples treated with blocking antibodies or controls treated with class-matched immunoglobulins. Analyses were presented as unpaired mean values ± SE in two-tailed, one-way ANOVA tests. Tukey tests were run after the ANOVA to determine the significance of the difference between the variability of the mean values between the two groups. All data reported were statistically significant (P < 0.05 or as specifically mentioned in each result) and satisfied the Tukey test requirements.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effects of ATRA on expression of ₁-mRNA. Integrin profiles on vascular smooth muscle cells are very important to determine their fate. It has been shown that ATRA substantially attenuated the proliferative action of platelet-derived growth factor (PDGF-BB) and serum stimulation in VSMCs in culture (21). The present study attempts to correlate integrin expression with growth inhibitory doses of ATRA in the same RASMCs. The status of the RA receptors in these cells has already been characterized (21). To further verify the results, a second line (PACs; see Ref. 26) was also studied. Both lines were grown in culture and treated for 24 h with vehicle (DMSO) or ATRA (1 µM) dissolved in DMSO. The total RNA was extracted, electrophoresed as described in METHODS, and probed with labeled ₁-integrin cDNA. The samples showed only one band hybridizing to the probe at ~4.6 kb (Fig. 1A), which is the same size already described for the rat ₁-integrin (4). The amount of RNA loaded in each lane was normalized by quantitating the ethidium bromide-stained 28S or 18S bands in a fluorimager, and this value was used to express the amount of specific ₁-mRNA in each lane. There was a significant increase in the ₁-message in both cell lines after treatment with 1 µM ATRA (P = 0.02 for n = 3 in RASMC, P = 0.0021 for n = 4 in the PACs), as shown in Fig. 1B. The steady-state message was increased >60% after only 24 h of treatment with ATRA at concentrations that affected PDGF-BB or serum-stimulated growth of these cells.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Northern hybridization of vascular smooth muscle cell (VSMC) RNA after treatment with all-trans retinoic acid (ATRA). A: blot showing the 1-specific mRNA (~4.6 kb, top) in control cells (C, treated with vehicle) and cells treated with ATRA (R, 1 µM). Bottom: ethidium bromide-stained 28S rRNA. RASMC, rat aortic VSMC; PAC, pulmonary artery VSMC. B: graph showing the quantitative increase in 1-integrin message after treatment with 1 µM ATRA for 24 h. Each value for 1-mRNA was normalized for amount of RNA loaded on the gel using the corresponding densitometric reading for the 28S rRNA. n, No. of preparations.

Total ₁-protein expression in VSMC. It was more difficult to see a specific increase in ₁-protein after treatment with ATRA, as integrin messages are often long-lived and abundant (16). Also, regulation of these messages at the translational level remain poorly understood. In an effort to document the change in ₁- protein, the cells were treated with ATRA for varying times. An increase in expression was seen after 3 days of exposure to ATRA and was very significantly upregulated after 10 days (Fig. 2A). The delay to observe early changes in protein may also be due to technical factors such as affinity of the antibody and sensitivity of chemiluminescent assay used for this protocol. This is in contrast to Northern blots where hybridization of mRNA to long (>100 bp) [³²P]DNA is very sensitive. The nonreduced ₁-protein was around 121 kDa as has been published (18). We confirmed the integrity of this band with a second anti-CD29 polyclonal antibody (AB 1937; Chemicon; results not shown). The ₁-bands from independent experiments were quantitated and normalized for protein loading from the corresponding densitometer reading of the smooth muscle -actin bands in each sample. Treatment with ATRA increased the expression of the ₁-integrin subunit in both cell lines, as seen in Fig. 2, A and B, with P = 0.01 for n = 4 in RASMCs and P = 0.001 for n = 3 in PACs. The upregulation in the ₁-protein was most likely more pronounced than the increase in the total ₁ message seen in Effect of ATRA on expression of ₁ mRNA due to the longer treatment with ATRA.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Western blots showing increase in total 1 expressed in RASMCs and PACs after treatment with retinoic acid (RA). A: proteins from 10-day ATRA (1 µM)-treated cells (R) and controls (C) were used for Western analysis. Each gel was probed sequentially with 2 antibodies; one antibody was for the 1-integrin protein (~121 kDa when nonreduced; top), and the second was a smooth muscle specific -actin monoclonal antibody (bottom; demonstrating equal loading in each lane). B: quantitative changes of total 1-integrin protein in RASMCs and PACs after treatment with ATRA, after normalization with -actin.

Effect of RA on ₁-integrin in vivo. To see if this increase in ₁-integrin expression would occur in vivo, 8-wk-old adult male rats were given oral doses of ATRA for a prolonged period (3 wk). Previous studies (20) have reported that 2 wk and 4 days of treatment with ATRA had significant effects on vascular remodeling after balloon withdrawal injury of the common carotid artery in rats. The 3-wk time point chosen for this study was to ensure that uninjured aortas were given ample chance to express changes in protein induced by the treatment. The dose selected was close to that administered for therapy to patients with APL and was less than the dose fatal to 50% of test animals for ATRA. After 3 wk of treatment, the rats were euthanized, and their aortas were examined for ₁-protein expression. As shown in Fig. 3A and as seen in the case of the cultured cell lines, there was a consistent increase in ₁-expression in the aortas of the ATRA-treated animals (P = 0.002, n = 4). We were not able to quantitatively determine the class of cells that contributed to this result, although the VSMCs are the best candidates for this since they are the most abundant cell type in the aortas and have shown growth regulation by ATRA (20). Examination of another vessel, the basilar artery, did not give such reproducible results. Closer analysis of this result showed that the ratio of smooth muscle -actin to total protein in the basilar arteries decreased after treatment with ATRA (unpublished observation), implying that RA could initiate other remodeling events in this vessel.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Western analysis of rat aortic 1-integrin protein after prolonged feeding of ATRA (1 mg · kg1 · 48 h1) for 3 wk. A: proteins of corresponding segments of the aorta from 4 control (C1-C4) rats and 4 ATRA-fed littermates (R1-R4) were probed sequentially with 1 and then smooth muscle (SM) -actin monoclonal antibodies. B: values for 1-integrin normalized with the smooth muscle -actin.

Effects of ATRA on cell adhesion. The cellular consequences of induction of ₁-integrin by ATRA were analyzed using cultured VSMCs. Alteration of the adhesive profile of the cells that change interaction with the matrix would be signaled to the interior of the cell. RA may act by inducing adhesive changes that trigger antiproliferative pathways and stimulate cellular differentiation. To test this, the cultured RASMCs and PACs were treated with ATRA (1 µM) for 5 days and then were used to assay changes in cellular adhesion. Treatment beyond this time (10 days, as used to assay for upregulation of ₁-protein) was not possible as it was difficult to lift the cells off the dish nonenzymatically due to the prolonged accumulation of secreted matrix. Harsh enzymatic digestion would likely destroy the extracellular integrin receptors. The cells were plated on a number of substrates, including vitronectin, laminin, collagen, fibrinogen, and fibronectin. The only significant change in adhesion in cells treated with ATRA was seen in the case of fibronectin. This is demonstrated in Fig. 4 in which both cell types (RASMC and PAC) showed increased adhesion and spreading on fibronectin after treatment with ATRA (P = 0.0001 for n = 10 for RASMCS and P < 0.0001 for n = 10 in PACs). The untreated RASMCs adhered better to fibronectin than the PACs, but the increase in adhesion after exposure to ATRA was greater for the PACs. This could be due to a different array of -subunits that heterodimerize with the ₁-integrins being induced by ATRA within the two cell lines. Particularly interesting was the fact that the change of adhesion to fibronectin after exposing the RA-induced cells was completely blocked by the functional ₁-antibody Ha2/5 (18 and Fig. 4; P = 0.008 and n = 3 for RASMCs and P = 0.0011 and n = 3 for PACs). This was not seen when nonspecific hamster IgM (negative control) or anti-rat ₃-blocking antibody (F-11; see Ref. 13) and matching purified ascites fluid were used in PACs in a separate experiment (see Fig. 4). The blocking anti-₃- monoclonal and ascites fluid (control) did not block adhesion, but both treatments increased nonspecific binding of RASMC to fibronectin and were not included in Fig. 4. The experiment thus clearly demonstrates that the upregulation in expression of the ₁-integrin subunit by ATRA was responsible for the altered adhesion of the two cell lines to fibronectin.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Effect of ATRA on VSMC adhesion to fibronectin (5 µg/ml). The two cultured cell lines RASMC and PAC were treated for 5 days with RA and used for analyses for adhesion to fibronectin. Values for nonspecific adhesion to uncoated wells were subtracted to obtain results for specific binding only. There was a >3-fold increase in adhesion of the RASMCs, which was not significantly altered by treating the cells with nonspecific hamster IgM. However, pretreating the cells with 1-specific antibody blocked this ATRA-induced increase in adhesion, showing that it was brought about by 1-integrins. This increased adhesion was not mediated by 3-integrins on PACs, as shown by using ascites fluid (IgAF) or anti-3-monoclonal F-11.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The aim of this study was to test the hypothesis that RA induced ₁-integrin expression and altered VSMC adhesion, events that could result in inhibition of growth and initiation of tissue remodeling. RA has been shown to antagonize the growth-promoting actions of PDGF-BB and serum in cultured RASMCs (21). The results obtained in this study clearly show that similar concentrations of ATRA induce the expression of the ₁-integrin subunit in these cells. The increase was observed in the steady-state message and protein levels in both RASMC and PACs. There are a number of examples where ₁-integrins have been associated with decreased cell proliferation (10, 19, 24). Furthermore, there is evidence that decreased proliferation of chronic myelogenous leukemia progenitors and K-562 cells was mediated by a ₁-integrin, which resulted in increased adhesion to fibronectin (15).

Treatment of RASMCs and PACs with ATRA also showed an increased adhesion to fibronectin. Fibronectin can bind to a number of integrins, e.g., ₃₁, ₄₁, ₅₁, _v1, _v3, and _v6 (27). The results from the present study showed that the RA-increased adhesion to fibronectin was completely blocked by a specific ₁-integrin functional antibody demonstrating it is mediated by ₁-integrins only. There was no change in RA-induced adhesion to fibronectin when an antibody to rat ₃-integrin, F-11, (unpublished observation) was used, confirming that ATRA increased cellular interaction with fibronectin in VSMC via a ₁- and not a ₃-integrin. The ₃-integrins have been seen to promote vascular proliferation (3, 9). It is therefore entirely reasonable to expect the antigrowth effect of the retinoids to be mediated by an increase in the ₁-integrin expression, which also enhances adhesion to fibronectin.

There were a number of reasons to check if the increase in ₁-integrin observed in cultured cells also occurred in vivo. First was the preliminary step to verify the observations recorded in cell lines as being relevant in living animals. Second, ATRA has been seen to induce remodeling of the carotid artery after balloon injury, reducing intimal hyperplasia and increasing vessel wall perimeter (20). Both events increase the lumen size, favoring the use of retinoids as therapeutic agents for prevention of restenosis. Third, chronic doses of the retinoids are being administered to patients with APL that could result in remodeling of their vasculature. For the last two reasons, the findings in this study that ₁-integrins are upregulated by ATRA merit closer and more detailed investigations.

The ₁-integrins have been described as key molecules in remodeling. In VSMCs, dynamic conformational changes in ₁-integrins were necessary for collagen matrix reorganization (14). The ₁₁ is a critical receptor in rat artery smooth muscle cells involved in matrix remodeling after injury (12). With the use of a mouse cell line lacking ₁-integrin, _51A was found to be a prime function for fibronectin matrix assembly (31). Therapeutic doses of ATRA were therefore used in this study to quantitate any changes in ₁-integrins in the blood vessels of adult rats. The results demonstrated a clear increase in the total ₁-integrin protein in the aorta, validating the observations made in VSMCs in culture. It was surprising to see variable amounts of ₁-integrin in a smaller vessel, the basilar artery, where the drug seemed to induce more dramatic remodeling. The specific VSMC marker smooth muscle -actin seemed to diminish with the treatment, implying that these cells were dying out. This complicated efforts to record reproducible changes in ₁-integrin (especially an increase) in these vessels after treatment with ATRA. This event did not seem to occur in uninjured carotid arteries (20). Thus there is a need to assess the effects of ATRA on remodeling of different vascular beds and vessel sizes in adults, to assess current treatment for patients with APL, and to benefit pathological procedures such as intimal disease associated with angioplasty and hypertension.