INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE RESPONSE OF blood vessel wall morphology to processes such as atherosclerosis and hypertension and after angioplasty involves changes in vessel wall mass as well as controlled alterations in total vascular circumference. Although much work has focused on understanding the regulation of vascular smooth muscle proliferation and synthetic activity, relatively little is known concerning the mechanisms that control the three-dimensional arrangement of smooth muscle (SM) cells (SMC) in the arterial wall. The modulation of vascular wall geometry during normal embryonic development and in response to physiological stimuli is referred to as vascular remodeling. During this dynamic process, coordinated changes in the mass and composition of the vessel wall occur that define a critical lumen area for normal vessel function.

Several types of geometric responses can be distinguished in the vessel wall during the remodeling process. Outward or adaptive remodeling may occur at the site of atherosclerotic plaques to compensate for the luminal decrease caused by an abnormal tissue mass (15, 18, 39). Outward remodeling can also be found after vessel trauma in association with an aneurysmal dilation, implying an increase of luminal area achieved by a thinning of the vessel wall. Inward remodeling may occur as an adaptation to a diminished flow requirement or as a maladaptive response after vessel trauma such as angioplasty (21). This latter effect results in the reduction of lumen area due to shrinkage of the vessel circumference, which is superimposed on vessel wall thickening. Despite the importance of vessel remodeling in determining clinical outcomes, relatively little is known about its key underlying regulatory mechanisms, which comprise mechanical forces as well as inflammatory and apoptotic responses.

Tissue-targeted expression of the SV40 T antigen (TAg) in transgenic mice has been used to create specific models of increased cell proliferation (26) by taking advantage of the specific effect of TAg to inactivate cell cycle negative regulators such as p53 (14, 28) and proteins of the pRb family, including p130 (6, 11, 38). These cell cycle negative regulators are of particular interest, as highlighted by recent observations that the loss of p53 results in acceleration of neointimal lesions of vein bypass grafts in mice (27) and that the loss of p130 accelerates vessel wall mass by neointimal formation and favors adaptive remodeling in injured carotid arteries (34). Other studies (13) showed that pRb and p130 are downregulated during response to injury in a rat carotid model. Accordingly, cytostatic gene therapy with a constitutively active form of the retinoblastoma gene product and p130 gene transfer revealed a significant decrease in SMC proliferation and neointima formation in a rat carotid and porcine femoral artery model of restenosis (4, 5). This points out the pathophysiological relevance for studying the effects of cell cycle negative regulators for a variety of vascular diseases, such as restenosis, atherosclerosis, and bypass graft stenosis. Here we directed expression of TAg to SMC using a 2.3-kb portion of the SM-myosin heavy chain (SM-MHC) gene promoter to study the effect of directly induced SMC proliferation (by TAg-induced inhibition of cell cycle negative regulators) in the vascular wall in the absence of endothelial injury or inflammation. The results show that enhanced SMC proliferation in the vessel wall is associated with adaptive remodeling which conserves a strong positive correlation (r = 0.93) between vessel wall thickness and the vessel circumference defined by the external elastic lamina (EEL). This establishes a new model of adaptive arterial remodeling induced by a single proliferative input in the absence of a mechanical injury.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Molecular constructs. Standard techniques were used to construct a minigene comprised of the 5' region of the rabbit SM-MHC gene and the wild-type SV40 TAg early region (MHC-TAg construct). Briefly, the 2.3-kb fragment of the 5' SM-MHC regulatory sequence from 2,305 to 4 bp relative to the transcriptional start site was used, containing a 107 bp enhancer region (1,332 to 1,225 bp) thought to function in tissue-directed expression, but no intron (17). The TAg DNA sequence was 2.7-kb long (BglI/BamHI fragment) starting at the origin of DNA replication 80 nucleotides upstream from the start codon and including the endogenous TAg splice site. The exact DNA sequence and amino acid sequence of TAg early region have been published (12). This TAg fragment includes the coding sequences for the binding sites of p53 (residues 350-450 and 532-625) and the proteins of the pRb family (residues 105-114) (19, 24). Orientations of fragments were confirmed by restriction digestion and DNA sequencing. Transcripts originating from the SM-MHC promoter will thus target expression of TAg in the mice.

Generation of transgenic mice and SMC lines. Transgenic mice were generated in a C3HeB/FeJ (The Jackson Laboratory; Bar Harbor, ME) background using standard techniques for microinjecting purified insert DNA into zygotes. Microinjected embryos were cultured in vitro to the two-cell stage and then reimplanted into pseudopregnant Swiss Webster/Taconic (Taconic Farms; Germantown, NY) female mice. All manipulations were performed according to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and Institutional Animal Care and Use Guidelines. Pups derived from the microinjected embryos were screened for the presence of the transgene by PCR amplification using primers located respectively in the 3' region of the SM-MHC promoter and the 5' region of TAg gene. Internal control primers included in each reaction were sense and antisense to two regions of the murine p53 gene selected so as to amplify a readily distinguishable DNA fragment. Reaction products were analyzed by electrophoresis through a 1% agarose gel and detected by ethidium bromide staining. Transgene expression was tested in every transgenic mouse by immunohistochemistry.

Histology, immunohistochemistry, and morphometry. Mice were euthanized with an overdose of halothane. Two MHC-TAg mice at 8 and 13 wk of age were perfusion fixed for optimized tissue morphometry and histology. All other MHC-TAg mice were not perfusion fixed to permit other molecular analyses. For perfusion fixation, the chest was opened immediately after death, and the heart was punctured with a 25-gauge cannula for infusion of PBS under physiological pressure. The blood was drained through an incision of the inferior vena cava. After 5 min, PBS was substituted by 10% buffered formaldehyde, and the mice were perfused for an additional 5 min before dissection and overnight postfixation with 10% buffered formaldehyde. Tissues were dehydrated through various concentrations of ethanol and embedded in paraffin by using standard methods. Sections were rehydrated and visualized with Verhoeff-van Gieson stain, hematoxylin and eosin stain, or Goldner stain, respectively. For immunohistochemistry, tissues were blocked with PBS containing 5% goat serum and 20 mg/ml bovine serum albumin, and treated with 2% H₂O₂ for 20 min to inactivate endogenous peroxidases. Staining for TAg employed a polyclonal rabbit antibody diluted 1:700 in blocking solution (kindly provided by Dr. Loren J. Field, Krannert Institute of Cardiology, Indianapolis, IN) (35) or a monoclonal antibody specifically reacting with the COOH-terminal end of TAg (clone PAb 101, Pharmingen; San Diego, CA). The latter antibody was used at a dilution of 1:200 in blocking solution. -SM actin was stained using a mouse monoclonal antibody diluted to 1:1,000 (clone 1A4, Sigma; St. Louis, MO). Proliferative cell nuclear antigen (PCNA) was analyzed with the use of a mouse monoclonal antibody diluted to 1:300 (clone PC 10, DAKO; Carpinteria, CA). Endothelial cells were detected with anti-von Willebrand factor diluted 1:200 (A082, DAKO). Primary antibodies were labeled using Vectastain ABC kits (Vector; Burlingame, CA) and color development was performed by diaminobenzidine (Sigma). Sections were briefly counterstained with hematoxylin for the visualization of all nuclei. For immunofluorescence primary antibodies (polyclonal rabbit antibody recognizing TAg and monoclonal mouse antibody recognizing PCNA as listed above) were labeled with Alexa Fluor 488 goat anti-rabbit diluted 1:500 and Texas Red goat anti-mouse diluted 1:200, respectively (Molecular Probes; Eugene, OR). For all immunohistochemical staining, a negative control was included which was treated identically but received no primary antibody.

Gel electrophoresis and immunoblotting. Dissected tissues were immediately frozen in liquid nitrogen after being ground in a mortar and then dissolved in a gel dissociation buffer containing 62.5 nmol/l Tris, 3% sodium dodecyl sulfate, 20% glycerol, and 20 nmol/l dithiothreitol. Comparable amounts of protein were loaded onto 7.5% and 5% polyacrylamide gels and blotted onto nitrocellulose membranes (0.45 µm, Schleicher and Schuell; Dassel, Germany). Membranes were blocked with PBS containing 5% milk and 0.1% Tween 20 for 4 h after incubation with primary antibodies for 2 h. Polyclonal anti-TAg (kind gift from Dr. Loren J. Field) (35) was used at a dilution of 1:5,000; -SM actin was stained using a mouse monoclonal antibody diluted 1:1,000 (clone 1A4, Sigma); and anti-SM-MHC immunoglobulin (clone G4, Santa Cruz Biochemical; Santa Cruz, CA) was diluted 1:250. After the washing steps, secondary antibodies were incubated for 1 h before signal development was performed by enhanced chemiluminescence (ECL, Amersham; Arlington Heights, IL).

Preparation of genomic DNA, Southern blotting, and Northern blotting. Southern blotting was performed by standard methods. Briefly, snap-frozen liver tissues of MHC-TAg mice were ground in a mortar before the addition of genomic DNA cell lysis solution containing 0.05 mol/l Tris (pH 8), 0.02 mol/l EDTA, 0.1 mol/l NaCl, and 1% SDS for 10 min. The resulting solution was then treated with 50 µg/ml RNase (Roche Diagnostics; Indianapolis, IN), followed by digestion with 100 µg/ml proteinase K (Sigma) overnight. Genomic DNA was extracted with repeated phenol-chloroform-isoamyl alcohol steps and ethanol precipitation. The genomic DNA was digested with BamHI and XbaI and extracted by phenol-chloroform-isoamyl alcohol. Genomic DNA (20 µg per lane) was run on a 0.8% agarose gel before the gel was soaked in 0.25 N HCl for 15 min to partially hydrolyze the DNA. It was then soaked for 15 min in alkali (0.5 mol/l NaOH and 1.5 mol/l NaCl) for denaturing and 2 × 15 min in neutralizing solution (1 mol/l Tris and 1.5 mol/l NaCl). The DNA was blotted onto a nitrocellulose membrane (Schleicher and Schuell) by capillary transfer overnight and fixed by UV light.

Blood pressure measurements. Measurements were taken with the tail cuff method using a pneumatic pulse transducer and programmed electrosphygmomanometer (model PE-300, Narco Biosystems; Houston, TX). Mice were sedated with 2.5% Avertin (0.015 ml/g body wt ip) and kept on a heating plate at body temperature. Measurements were taken immediately before the mice were awakened completely and before the initial onset of mobility. Blood pressures were determined for nine control mice and four MHC-TAg mice between 6 and 8 wk of age. Data are given as means from at least three measurements per animal, and analyzed by t-test for statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

General phenotypic characteristics of transgenic mice. MHC-TAg mice expressing TAg in SMC exhibited enlargement of arterial blood vessels, enlargement of the bowel and gall bladder, and development of tumors of the uterus, seminal vesicles, and lacrimal gland. Of nine independent founders generated, six founders showed vascular and visceral transgene expression, whereas the other three founders featured visceral transgene expression only. Those with vascular expression featured a consistent and striking vascular phenotype as well as visceral phenotypes (see below). These animals, four females and two males, were euthanized between the ages of 3 and 13 wk due to abdominal distention associated with the development of generalized bowel enlargement, a uterine tumor, and a tumor of the seminal vesicles, respectively. The other three founders lacked the development of vascular alterations. One of these latter founders allowed the propagation of a mouse line, which was characterized by normal vessel development but a high susceptibility to visceral SM tumor development. Female animals developed uterine tumors in 70% of the cases by the age of 4-5 mo. The uterine hyperplasia was diffuse as well as nodular, of solid consistency, and comprised up to 50% of total body weight. The uterine tumors were characterized by -actin-positive hyperplastic stroma without glandular elements. Male animals developed a seminal vesicle tumor in 75% of cases by the age of 3-6 mo. The bowel enlargements included a megacolon associated with hyperplastic thickened muscular layers of the bowel wall without evident strictures or obstructing tumors.

3

3

Tissue-restricted expression of TAg in transgenic mice. Immunohistochemical analysis was performed on vascular and nonvascular tissues from all MHC-TAg mice using a polyclonal antibody against TAg. For the aorta, data obtained from a polyclonal antibody were compared with data obtained from a monoclonal antibody recognizing the TAg COOH terminus to confirm correct processing and translation of the transgene. Nuclear staining for TAg was consistently found in all SM-containing vessels as well as the muscular layers of the entire gastrointestinal tract comprising the esophagus, stomach, small intestine, colon, and the gall bladder, but was not detected in skeletal muscle, airway SM, or glandular elements of the gastrointestinal tract. The spleen exhibited TAg expression in the SMCs of the capsule and stroma. Furthermore, TAg expression was found in the muscular layers of the genitourinary tract comprising the bladder, uterus, epididymis, and seminal vesicle. The fraction of SMC nuclei that were TAg positive was evaluated for several positive tissues of transgenic mice, as depicted in Fig. 1. The extent of TAg expression varied between organs and ranged from 19 to 52% of SMC nuclei of arterial as well as visceral tissues. The number of TAg-positive nuclei (29%, polyclonal antibody) in the aorta was almost identical with the number of nuclei that stained positively with the COOH-terminal TAg (32%), thus confirming the expression of a complete transgene. Figure 2 shows typical TAg staining and -actin staining in vascular and visceral transgenic tissues. Figure 2, A and B, shows a comparison of adjacent sections in one aorta (3-wk-old transgenic mouse) stained with both TAg antibodies. Tissues of control mice did not show TAg staining (Fig. 2C). TAg expression was also found in visceral SM as shown for the esophagus (Fig. 2G) and colon (Fig. 2J). To further correlate TAg expression with SM tissue morphology, Fig. 2 reveals -actin staining of a transgenic aorta (Fig. 2D) and a control aorta (Fig. 2E), as well as a transgenic esophagus (Fig. 2H) and control esophagus (Fig. 2I) and a transgenic colon (Fig. 2K) and control colon (Fig. 2L), respectively. Expression of TAg was also seen in the kidney tubular epithelium and tracheal epithelium in one founder each, suggesting inappropriate expression due to transgene integration effects. All other TAg expression was restricted to SM tissues as described above.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Quantitative analysis of smooth muscle (SM) cell (SMC)-T antigen (TAg) expression in transgenic mice [myosin heavy chain (MHC)-TAg] mice. TAg positive smooth muscle nuclei of various tissues were quantitated on immunohistochemical stainings, as described in MATERIALS AND METHODS. Immunohistochemistry was performed with a polyclonal antibody against TAg except for the aortic sections, which were stained with a monoclonal antibody specifically reacting with the COOH-terminal end of TAg. Adjacent aortic cross sections were used for immunohistochemistry with the polyclonal and monoclonal antibody. Data are given as the percentage of SM nuclei revealing positive TAg signals. The bars indicate the means ± SE of six MHC-TAg animals (coronary arteries and bladder, n = 4 mice).

View larger version (120K): [in this window] [in a new window]   Fig. 2.   Immunohistochemical analysis of SMC-TAg expression and -actin expression; comparative displays of various MHC-TAg tissues and controls. A: aorta of a 3-wk-old MHC-TAg mouse with beginning neointima formation was stained for TAg with a polyclonal antibody. It was also compared with an adjacent section stained with a monoclonal antibody, which specifically recognized the COOH-terminal end of TAg (B), whereas C shows the absence of TAg stain in a control aorta. A nearby section of the same MHC-TAg aorta was stained for -actin (D) and compared with the aorta of a control mouse also stained for -actin (E). F shows negative staining control. G and J: TAg staining (polyclonal) of the esophagus and colon of a 13-wk-old MHC-TAg mouse and comparison with -actin stainings of the same tissues (H and K) and -actin stainings of controls (I and L). TAg and -actin were visualized by peroxidase and diaminobenzidine (DAB) (brown), whereas hematoxylin (blue) was used as a nuclear counterstain. Original magnifications: A-E, ×1,000; F-L, ×400.

View larger version (45K): [in this window] [in a new window]   Fig. 3.   Immunoblot revealing MHC, TAg, and -actin levels in MHC-TAg tissues. Tissue lysates were obtained from various organs of a 5-wk-old MHC-TAg mouse (MHC-TAg). Organs were snap-frozen in liquid nitrogen before the tissue was pulverized in a mortar and dissolved in gel dissociation buffer. Comparable amounts of protein (60 µg/lane) were loaded onto a 5% polyacrylamide gel and blotted onto a nitrocellulose membrane. Signals for SM-MHC, TAg, and -actin were obtained from the same membrane for optimal comparative analysis. Signal assessment was performed by chemiluminescence.

View larger version (37K): [in this window] [in a new window]   Fig. 4.   Vascular vs. nonvascular phenotype in MHC-TAg mice. A: immunoblotting: tissue lysates were prepared as described (Fig. 3) and run on a 7.5% polyacrylamide gel. Protein (22.5 µg) was loaded per lane. Signals for TAg and -actin were obtained from the same membrane. The signals shown for the uterine tissue were obtained from a separate blot, which was run under identical conditions. B: Southern blotting. Genomic DNA was digested with BamHI and XbaI and run on a 0.8% agarose gel (20 µg/lane). DNA was blotted by capillary transfer overnight onto a nylon membrane and finally hybridized with a 32P-labeled probe, which was generated by random prime labeling by using a 1.6-kb EcoNI fragment of SV40 TAg DNA as a template.

Analysis of arterial morphometric parameters in MHC-TAg mice and age-matched controls. To evaluate the phenotype of arterial vessels in MHC-TAg and control mice, we measured the circumference of the EEL, internal elastic lamina, and the lumen and determined areas for the intervening vessel wall layers of the thoracic aorta and the common carotid, femoral, and coronary arteries. For MHC-TAg mice, spontaneous formation of neointima was present and progressed in an age-dependent manner but was also dependent on the phenotype. Circumferential neointima was found in the aorta and the common carotid arteries of MHC-TAg mice, whereas the femoral arteries displayed patchy neointima formation. Even in vessels without circumferential neointima formation, vessel enlargement was characterized by medial thickening accompanied by lumen expansion to values distinctively larger than those of control mice. For the thoracic aorta, common carotid and femoral arteries all MHC-TAg mice at age 3-13 wks showed significant two- to threefold increases in both vessel wall area and lumen area compared with controls (each P < 0.05), as depicted in Fig. 5. Figure 6, A-J, shows representative Verhoeff-van Gieson stains of MHC-TAg and control arteries.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Area ratios of the vessel wall and the lumen of the thoracic aorta, common carotid, and femoral arteries. The vessel wall areas and lumen areas of MHC-TAg mice (n = 6) were evaluated and compared with age-matched control mice (n = 18). The area ratios (MHC-TAg/control) were analyzed for statistical significance by Wilcoxon's signed-rank test. Control values were set equal to 1. * Significant increase in areas of MHC-TAg mice over corresponding controls. Data are given as means ± SE.

View larger version (83K): [in this window] [in a new window]   Fig. 6.   Vessel morphology in MHC-TAg mice. Verhoeff-van Gieson stains (A-J) of the thoracic aorta, common carotid artery, femoral artery, and left coronary artery of a 13-wk-old MHC-TAg mouse and an age-matched control. A-D: comparison of the thoracic aorta in terms of size and wall thickness, including tunica media and neointima of a MHC-TAg animal and control. All vessels were obtained from animals perfusion fixed as described in MATERIAL AND METHODS (original magnifications: A and B, ×400; C and D, ×25; E-J, ×100).

View larger version (26K): [in this window] [in a new window]   Fig. 7.   Conserved correlation of vessel wall thickness and vessel circumference in MHC-TAg mice. MHC-TAg mice at ages of 3-13 wk (n = 6) and age-matched controls (n = 18) were analyzed for vessel circumference and vessel wall thickness of various arteries (aortic segments and common carotid, femoral, and coronary arteries). Vessel preparation and morphometry were performed as described in MATERIALS AND METHODS. Controls were prepared identically to each corresponding MHC-TAg mouse. Data are given as means of multiple sections analyzed per vessel in each animal. In case of controls data reflect means of three age-matched animals. EEL, external elastic lamina.

Proliferative characteristics of vessel wall and matrix expression. To evaluate the role of cellular proliferation in determining the growth of vessels of MHC-TAg mice compared with controls, we performed cell counts on the thoracic aorta and the common carotid, femoral, and coronary arteries (Table 1). The average cellular density of transgenic arteries was increased 3.2-fold compared with controls (19.02 ± 1.29 vs. 5.86 ± 0.30 cells/10³ µm², respectively, P < 0.001). The typical number of cellular layers lying between elastic laminas was increased from one to three in the transgenic aortic media, whereas the total number of laminar units was conserved (5.97 ± 0.12 for transgenic mice vs. 5.81 ± 0.06 for controls, NS).

View larger version (72K): [in this window] [in a new window]   Fig. 8.   Localization of proliferative cell nuclear antigen (PCNA) and TAg expression, endothelial integrity and matrix expression. A: immunofluorescent staining for coexpression of PCNA with TAg in a thoracic aorta of a 13-wk-old MHC-TAg mouse. Staining for PCNA is depicted by red fluorescence and TAg is revealed by green fluorescence, whereas direct overlay of both appears yellow. Pictures were obtained by confocal microscopy. B: negative staining control. C and D: von Willebrand factor stainings of the thoracic aortas of a 13-wk-old MHC-TAg animal and a control mouse to test for endothelial integrity. Goldner stains indicate matrix expression (green) for the aorta of a MHC-TAg mouse (E) and a control (F) (original magnifications: A and B, ×2,500; C and D, ×1,000; E and F, ×400).

View larger version (24K): [in this window] [in a new window]   Fig. 9.   Comparative displays of temperature-sensitive TAg expression, type VIII and type I collagen RNA expression in aortic SMC (MHC-tsA58). Aortic SMC were derived from MHC-TAg mice containing the 5' region of the rabbit SM-MHC gene connected to the temperature-sensitive TAg (tsA58) mutant SV40 TAg early region. A: immunoblot with 5 µg protein per lane testing for TAg expression and -actin expression at the permissive temperature of 33°C and at the restrictive temperature of 39.5°C. B and C: Northern blotting experiments with 2.0 µg RNA per lane (B) and 4.0 µg RNA per lane (C) obtained from SMC grown at 33°C and 39.5°C, respectively. The blots were hybridized at 72°C with either 1(VIII)procollagen (B) or 1(I)procollagen (C) antisense riboprobes. Both types of collagens are typically detected in a multiple-band pattern. Blots were hybridized for 18S RNA to test for equal loading of RNA. Detection was performed by enhanced chemiluminescence.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

This study presents a novel MHC-TAg mouse model demonstrating adaptive arterial remodeling associated with supranormal SMC proliferation and spontaneous neointima formation induced by the tissue-specific inactivation of cell cycle negative regulators. This SMC proliferation and hyperplasia occurs in the absence of mechanical vascular injury with its attendant variability and inflammatory responses. The data presented here provide evidence that direct induction of SMC proliferation is associated with adaptive remodeling as well as tissue growth. Accordingly, MHC-TAg animals displayed conservation and even enlargement of vessel lumen despite remarkable vessel wall thickening. The correlation between vessel wall thickness and vessel circumference (both EEL and lumen) noted in controls is maintained precisely in the MHC-TAg mice. This relationship implies the conservation of circumferential vessel wall stress () as a constant among both control and MHC-TAg animals, defined as
where P is the mean transmural pressure, r_i is the internal radius, and t is the vessel wall thickness (3). The blood pressures of MHC-TAg mice were found to be equivalent to controls, both by direct measurement; and by inference from morphometric evaluations of the MHC-TAg and control hearts with respect to weight ratio, gross dimensions, and muscle fiber diameters (9, 10). These dimensions were comparable for MHC-TAgs and controls, consistent with the absence of any hypertrophic response that could indicate an intermittently increased blood pressure. Thus the parallel increase in internal vessel radius and vessel wall thickness in the MHC-TAg animals provides evidence for conservation of circumferential wall stress as an important constant across multiple vessel types as well as in the context of hyperproliferation as found in the transgenic mice.

To further evaluate the conditions for adaptive remodeling in the context of SMC hyperproliferation we performed studies on matrix expression. Goldner stains did not reveal overtly differing quantitative expression for MHC-TAgs and controls, however, this does not rule out qualitative alterations in extracellular matrix. To further illuminate the matrix-specific effects of SMC hyperproliferation due to TAg expression, we have generated MHC-TAg mice that express a temperature-sensitive mutant of TAg under the same SM-MHC promoter (MHC-tsA58). These mice were established for the generation of SMC lines with inducible TAg expression. Previous studies (26) and in vitro experiments showed that the temperature-sensitive TAg is active to a large extent at body temperature (37°C). This was consistent with our observation that these mice also showed the characteristic vascular proliferative and adaptive arterial remodeling response as described for MHC-TAg mice. Northern blotting experiments with SMC obtained from the aorta of a MHC-tsA58 mouse revealed that SMC hyperproliferation due to TAg expression caused alterations in the RNA levels of type VIII and type I collagens. This establishes the hypothesis that alterations in fibrillar versus network forming collagens may support the conditions for conserving vessel wall stress during adaptive arterial remodeling in a hyperproliferative vessel wall.

The 2.3-kb fragment of the SM-MHC promoter employed appears sufficient to direct TAg expression to vascular and visceral SM tissues in vivo. All MHC-TAg mice were studied with regard to the integrity of the expressed transgene by immunohistochemical staining from an antibody recognizing the COOH-terminal end of TAg in addition to a polyclonal antibody for TAg. In addition, immunoblotting was performed on SM tissues of four founders (not all shown) revealing TAg signals of the expected size, thereby establishing expression of intact TAg.

Previous studies (26) using an -actin promoter failed to yield viable mice expressing TAg in the major vessels in vivo, presumably as a consequence of additionally significant extravascular TAg expression during embryonic development under regulation of the more promiscuous -actin promoter. The present results found with the SM-MHC promoter compare closely with the results from another study in which a longer, 16 kb SM-MHC promoter-intronic construct derived from rat directed SM gene expression in both vascular and visceral SMC (25). However, the prior study found that the presence of the intronic segments was necessary for significant expression. This discrepancy might be explained either by species-specific differences in the two promoter sequences or by an effect of TAg as the expressed gene to amplify the observable level of promoter function, at least in a subpopulation of SMC clones. In either case, the present data showing that the 2.3-kb rabbit SM-MHC fragment effectively directs SMC-specific transcription of TAg leaves remaining uncertainty as to whether such intronless fragments of the SM-MHC promoter can be expected to confer vascular-specific targeting under more general circumstances. This issue is of particular interest in designing gene-therapeutic strategies, but also is significant for transgenic model development. In the case of the mice described here, elimination of the visceral expression would be helpful to avoid the genitourinary tumor formation and bowel enlargement (the latter already found for TAg expression under a telokin promoter) (16), which generally precluded study of our animals beyond 4 mo of age. This nonvascular phenotype was a drawback of the study. Nevertheless, data were successfully obtained from six independent founders with vascular transgene expression. This circumstance actually strengthens the conclusion that the common vascular characteristics found in all six founders were attributable to the transgene rather than any secondary factors, e.g., the disruption of a naturally existing gene or other integration site-specific effects. Interestingly, one founder that began a line maintained visceral TAg expression in the absence of evident vascular expression. This likely represents an integration site-dependent effect (1, 37) of the transgene because Southern blotting revealed that vascular transgene expression was not associated with increased transgene incorporation, thereby making unlikely a gene dosage effect as the basis for the variable phenotypic expression.

The interplay of vascular remodeling with SMC proliferation is now recognized to be of utmost clinical relevance in determining the sequelas of atherosclerosis and restenosis after angioplasty. Many recent studies (2, 33, 36) have pointed out a dominant role of large vessel remodeling in determining luminal loss or gain, as expressed by the relationship between lumen area and vessel wall area or thickness. Adaptive remodeling appears to be an early physiological reaction to compensate for lumen loss in atherosclerosis (20, 36). After angioplasty, the impact of remodeling increases while the relative contribution of neointima to late lumen loss decreases over time (7). Our data clearly indicate that SMC proliferation and neointima formation is not necessarily accompanied by progressive vascular occlusion. In fact, increased proliferative activity of SMC is capable of creating an enlarged vessel to preserve wall stress. Similar conservation of wall stress was reported for transgenic mice overexpressing human growth hormone resulting in increased vessel growth and increased body weight (8). In contrast to the present study, however, an increase in vessel diameter was achieved in the context of a widespread effect with overall increases in body or organ weights, and was not related specifically to increased mural SMC proliferation, which is of relevance for atherosclerosis and restenosis. Mice have also been described in which the vascular remodeling response is disrupted. Studies of mice with homologous deletion of the elastin gene revealed that the loss of one allele is associated with an increase in elastic lamellas and SM, whereas mice completely lacking elastin died of SMC proliferation with eventual vascular occlusion (22, 23), suggesting that the absence of elastin rendered the vessel unable to maintain an adaptive relationship between wall mass increase and vessel circumference.

In summary, our results, taken together with previous studies, suggest that SMC hyperproliferation in the vessel wall is not a sufficient cause of vascular occlusion or constrictive vascular remodeling. Our study further suggests the hypothesis that the conservation of wall stress functions as a primary determinant of vascular remodeling by modulating the extent of radial versus circumferential directionality of vascular SM tissue expansion and migration, and this is supported by alterations in collagen expression. This insight into vascular remodeling may help to improve the understanding of pathophysiological mechanisms during early atherosclerosis. Our results may also pave the way to the development of new therapeutic paradigms based on the directional rather than absolute control of SMC proliferation in vascular diseases.