INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

VASCULAR HYPERTROPHY and remodeling are adaptive structural changes in response to sustained increases in arterial pressure or altered shear stress that favor restoration of normal physiological regulation. On the other hand, excessive wall growth, fibrosis, and inward or inadequate outward remodeling cause failure of surgical procedures (e.g., restenosis after angioplasty/ stent, atherectomy, and bypass grafting) and underlie diseases such as atherosclerosis, pulmonary hypertension, and accelerated arteriosclerosis (31, 35). Thus the mechanisms regulating growth of vascular wall cells are under intense investigation. Besides the vasoactive actions of norepinephrine (NE), there is growing evidence that NE may be a trophic mediator for vascular smooth muscle cells (SMCs) and adventitial fibroblasts (AFBs). In vivo studies using surgical or systemic sympathetic denervation (16), systemic infusion of catecholamines (7, 21), or -adrenoceptor (AR) antagonists (20), as well as positive correlation of plasma catecholamines with wall hypertrophy and stiffness (8) and severity of atherosclerosis (22) in humans, suggest that NE may have direct trophic effects on the normal and diseased vascular wall. Moreover, in the balloon-injured rat and rabbit carotid, chronic systemic ₁-AR antagonists reduced cell proliferation, neointimal growth, and restenosis by at least 50% (14, 18, 30, 34). ₁-AR antagonists also attenuated angiotensin II-induced DNA synthesis (33) and atherogenesis (26, 29). However, interpretation of these past in vivo studies is complicated by concomitant hemodynamic disturbances that, themselves, have trophic effects. For example, chemical or immunological systemic denervation and systemic -AR antagonists cause significant hypotension and humoral changes. As well, local denervation of an artery leads to dilation of its dependent branches, which increases flow and causes outward remodeling. Hypotension directly inhibits vascular wall hypertrophy. Thus these effects complicate interpretation of past studies regarding the concept that NE may exert a direct trophic action on the vascular wall.

Recent in vitro studies do support, however, the concept that catecholamines may be trophic for SMCs and AFBs. Blood vessels express multiple AR types that do not all mediate changes in SMC contraction (9, 13). For example, medial SMCs and, surprisingly, AFBs of the intact rat thoracic aorta and carotid artery express all three ₁-ARs (_1A, _1B, and _1D), as well as _2D/A (hereafter designated _2D) and -ARs (13); moreover, total ₁-AR abundance is the same on both the medial SMCs and adventitial AFBs (13). Whereas _1D- and _2D-ARs signal constriction and -ARs dilation of the rat aorta (see 9 and 13 for Refs.), the function of _1A- and _1B-ARs on medial SMCs as well as the multiple subtypes on noncontracting AFBs has been unclear. In cells cultured from the rat aorta, NE stimulated proliferation of subconfluent SMCs (and hypertrophy of growth-arrested SMCs) through activation of ₁-ARs, but not ₂- or -ARs (37 and references therein), and stimulated AFBs to proliferate (13). In the uninjured rat aorta maintained in organ culture under radial wall tension to simulate normal mean arterial pressure, NE induced hypertrophy of SMCs and proliferation of AFBs and reduced expression of marker proteins that characterize the differentiated SMC phenotype (38). Moreover, NE proliferation of SMCs and AFBs was strongly augmented (as were the changes in marker protein expression) in aortas that had received balloon injury either 4 or 12 days earlier in vivo (38). In that study, proliferation and phenotypic changes of intima-media cells were blocked by an _1A-AR antagonist (KMD-3213), whereas proliferation of adventitial cells was inhibited by an _1B-AR antagonist (AH11110A); _1D-, ₂-, and -AR antagonists (BMY-7378, RX-821002, and propranolol, respectively) had minimal to no effect (38). Because these organ culture studies were only studied for 48 h in culture, whether these effects, when prolonged, translate into a contribution of NE to inward remodeling, neointimal growth and lumen loss ("restenosis") could not be determined; moreover, a trophic role for endogenous NE could not be determined in that study. We have recently found that NE induces dose-dependent chemotaxis of cultured rat aorta SMCs and AFBs, where _1D- and _2D-ARs mediate migration of SMCs, whereas _1A-ARs stimulate migration of AFBs (39). These migration results further underscore the possibility that NE may directly contribute to vascular wall growth and remodeling.

Despite the aforementioned cell and organ culture studies demonstrating that NE exerts direct trophic actions and that these actions are strongly augmented in an injured artery, SMCs and AFBs are significantly altered by in vitro conditions. In particular, SMCs and AFBs maintained in primary cell culture undergo significant changes in the expression of ARs and cytoskeletal and contractile proteins (13, 38). As such, past in vitro studies, as well as in vivo studies using systemic AR agonists and antagonists with their attendant hemodynamic disturbances that themselves have trophic effects, have not addressed whether NE is directly trophic for the vascular wall in vivo. Thus it is important to test this hypothesis by using a method that avoids systemic hemodynamic and/or humoral disturbances. In the present study, NE and adrenoceptor antagonists were delivered by chronic local perivascular suffusion to the uninjured and balloon-injured rat carotid artery, and effects on intimal, medial, and adventitial growth were measured. The results provide the first in vivo evidence that NE can act directly to cause wall restructuring in the uninjured artery. Moreover, they suggest that basal as well as elevated wall NE may contribute significantly to intimal expansion and lumen loss after injury.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Balloon injury and chronic perivascular suffusion. Male 500-g Sprague-Dawley rats (n = 161 for reported data plus ~50 for preliminary experiments) were anesthetized with ketamine (125 mg/kg im) plus acepromazine (1.25 mg/kg im) and received atropine (54 µg/kg sc) and cephazolin (50 mg/kg im). The left thyroid, occipital, and external carotid arteries proximal to the lingual artery were isolated sterilely and ligated. After heparin administration (60 U/kg im and 110 U/kg sc), a 2-Fr embolectomy catheter (Baxter Healthcare; Irvine, CA) was advanced via an external carotid arteriotomy to the origin of the left common carotid. The balloon was inflated with 15 µl of saline and rotated while withdrawing it the length of the common carotid; the procedure was repeated twice and the external carotid artery was ligated. A sterile catheter for perivascular suffusion (modified from Ref. 6) was constructed from thin-walled vinyl tubing (0.58/0.96 mm, ID/OD), which had four 27-gauge needle slits placed radially 6 mm from its closed end. The catheter was anchored with 5-0 silk to the longus colli so as to pass under the omohyoideus and lie parallel with and ~0.5 mm lateral to the common carotid sheath, which was not disturbed.

Morphometry. Fourteen days after surgery (28 days for the prazosin experiment in which empty pumps were replaced at 14 days under ether anesthesia), pumps were verified as empty and catheters intact, and the vasculature was perfusion fixed at 100 mmHg with transcardial 4% paraformaldehyde in phosphate-buffered saline. The position of the left carotid adjacent to the catheter outflow slits was marked with India ink, and the left and right carotids were removed en bloc with their surrounding sheaths and skeletal muscle intact. Vessels were postfixed for 24 h in 4% phosphate-buffered saline at 4°C and embedded in paraffin, and the centermost 5-mm section was blocked. Three adjacent 8-µm sections were cut every 200 µm, extending 1.5 mm above and below the India ink-identified position of the catheter outflow slits, and were stained with Masson's trichrome, hematoxylin and eosin (H+E), or picrosirius red (10 sections for each stain per vessel). Three trichrome sections approximately equally separated over the length of the blocked carotid were selected for digital planimetry (Scion Image, NIH) by an observer blinded for the treatment group; adjacent sections to these were stained with and analyzed for nuclear density (cell number; H+E) and collagen I/III (picrosirius red). Within individual groups, there were no consistent differences in vessel area dimensions among the sections spanning the 3-mm length, indicating that the injury and drug effects were relatively uniform extending along 1.5 mm above and below the outflow slits of the suffusion catheter. Areas were determined as follows: lumen area = (lumen circumference)²/4; neointimal area = area between the lumen and internal elastic lamina (IEL); medial area = area between the IEL and external elastic lamina (EEL); circumference = length of EEL; and adventitial area = area of the dense collagen containing layer between the EEL and loose perivascular connective tissue. Neointima thickness was calculated as [(circumference of IEL  2)  (circumference of lumen  2)]; thickness of media and adventitia was similarly calculated using circumferences of the EEL and IEL, and the outer edge of the dense adventitia and EEL, respectively. Cell nuclei were counted by threshold masking and digital image-element recognition; automated counts were confirmed manually in a preliminary analysis. Relative collagen content was determined by color threshold masking of serius red-stained sections (Scion Image), with zero defined by the "red" intensity inside of a skeletal muscle fascicle (23). The same threshold value was used for all vessel sections.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

NE caused inward remodeling of uninjured carotid. Two-week local suffusion of NE, compared with vehicle, caused lumen area, adventitia area, and circumference to decrease by 10%, 14%, and 5%, respectively, indicating a modest inward remodeling (Fig. 1). At the time of vessel harvest, no histological evidence of catheter implantation was detectable at the carotid sheath or the structures contained therein (i.e., carotid, vagus, sympathetic chain, internal jugular vein, and intervening connective tissue, Fig. 2). The catheter, which was fixed ~0.5 mm lateral to the sheath, was covered with a thin connective tissue layer. Vehicle (0.2% ascorbated Ringer solution) had no effect on vessel morphometric measures compared with naive right carotid (Fig. 1 vehicle group vs. uninjured group in Fig. 6). Also, vehicle and catheter implantation had no effect on dimensions in preliminary experiments of balloon-injured animals that received vehicle suffusion for 14 or 28 days compared with naive rats without catheters; for example, in vehicle-suffused, balloon-injured animals (14 days, n = 12) versus injured animals without catheters (n = 9), lumen area was 0.33 ± 0.03 vs. 0.26 ± 0.03 mm², neointimal area was 0.15 ± 0.01 vs. 0.19 ± 0.02 mm², medial area was 0.13 ± 0.01 vs. 0.11 ± 0.01 mm², adventitial area was 0.20 ± 0.01 vs. 0.20 ± 0.02 mm², and circumference (of the EEL here and elsewhere) was 2.82 ± 0.06 vs. 2.73 ± 0.11 mm (all P > 0.05). In all experiments in this study, body weight gain over the duration of the experiments was not different among any of the groups.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Effect of 2-wk perivascular suffusion of norepinephrine (NE, 100 µM; concentration is for pump reservoir and infusion rate is 5 µl/h, here and elsewhere) or vehicle (0.2% ascorbated Ringer solution, here and elsewhere) on the uninjured rat common carotid artery. Circumference, here and elsewhere, is measured at the external elastic lamina. n, Number of vessels (one per animal), here and elsewhere.

View larger version (117K): [in this window] [in a new window]   Fig. 2.   Histological examples from animals from four different treatment groups: uninjured rat carotid (A) and carotids 28 days (B) and 14 days (C and D) after balloon injury, with perivascular suffusion of vehicle (A-C) or NE (D, 100 µmol/l). A and B: Masson's trichrome; C and D: picrosirius red. Bar in D = 100 µm.

NE increased neointimal formation after balloon injury. Balloon injury caused neointimal formation, lumen loss, and thickening of the adventitia into a "neoadventitia" (Fig. 2, A vs. B). This expected effect of injury alone is quantified in Figs. 6-8 and discussed below. Two-week suffusion of NE significantly increased neointimal area and lumen loss and reduced circumference and adventitial area (Fig. 2, C vs. D, Fig. 3). Neointimal cell (nuclei) number was increased (Fig. 4A), but in proportion to the increase in neointimal area (Fig. 3); thus cell density (Fig. 4B) was not increased. As also proposed for other growth factors in this injury model (31, 35), this suggests that NE may have augmented proliferation early on, but by 2 wk the neointimal area had increased and cell size and density had "returned" to normal for neointima (possibly from a combination of apoptosis, matrix accumulation, and maturation of cell size after mitosis). Total area of picrosirius red staining increased secondary to NE-induced expansion of the neointima area (Figs. 2 and 4D). Thus, when factored with intensity of collagen staining, total collagen content in the neointima increased (Fig. 4E). As a result, the volume fraction of collagen (collagen density) was increased in the neointima (Figs. 2 and 4F).

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Perivascular suffusion of NE (100 µmol/l) for 2-wk augmented neointimal growth, lumen loss, and reduction in circumference of the external elastic lamina (inward remodeling). See Fig. 1 for additional details.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   NE (2 wk, 100 µmol/l) increased cell number in neointima (A) in proportion to increased neointimal area (see Fig. 3), thus no effect on cell density (B). Cell number represents nuclear number per high power field (×200). NE increased neointimal collagen content (volume fraction): mean pixel density (C) equalled average red intensity of pixels. Total area of collagen (D) equalled area of all pixels exhibiting any red intensity above zero [zero defined as "red" intensity inside surrounding skeletal muscle fascicles (Fig. 2, C and D, yellow stain)]. Total collagen content (E) equalled product of C and D data. Volume fraction of collagen (F) equalled quotient of total collagen content per layer and area of layers as shown in Fig. 3. See Fig. 1 for additional details.

Blockade of ₁-adrenoceptors-attenuated response to balloon injury. Two-week suffusion of benoxathian attenuated neointimal area and tended (not significant by 2-tail) to lessen lumen loss and medial growth (Fig. 5). Similar inhibitory effects were obtained with 4-wk delivery of prazosin; when compared with vehicle (n = 8), prazosin (n = 7) reduced the neointimal area by 32 ± 3% (P < 0.01) and tended to reduce lumen loss (by 31 ± 11%, P = 0.051; 2-tail) but had no effect on the medial area or circumference (adventitial area was not measured in this experiment). Lumen reduction and increases in neointimal, medial, and adventitial areas were not different at 14 versus 28 days after balloon injury, confirming many reports that changes in rat carotid wall areas achieve their maximum within 2 wk after the balloon injury protocol used herein. Two-week suffusion of benoxathian to uninjured vessels (n = 8) had no effect compared with vehicle (n = 8). Lumen, media and adventitia areas, media and adventitia thickness, and circumference for vehicle was 0.64 ± 0.023 mm², 0.09 ± 0.004 mm², 0.15 ± 0.007 mm², 39.3 ± 5.2 µm, 49.4 ± 4.8 µm, and 3.08 ± 0.05 mm, respectively; and for benoxathian, 0.61 ± 0.029 mm², 0.09 ± 0.003 mm², 0.14 ± 0.006 mm², 33.0 ± 1.9 µm, 51.9 ± 7.1 µm, and 2.98 ± 0.06 mm (0.20 < P < 0.78 for all comparisons), respectively. Thus endogenous "basal" NE levels (i.e., present in the absence of behavioral or physiological stress) do not exert a tonic ₁-AR-dependent trophic action in the normal uninjured vessel.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Perivascular suffusion of 1-adrenoceptor (1-AR) antagonist benoxathian for 2 wk (10 µmol/l) attenuated restenosis induced by balloon injury. Prazosin (10 µmol/l, 4 wk) had similar effects (see RESULTS). See Fig. 1 for additional details.

_1A-AR antagonist attenuated restenosis. To examine possible involvement of ₂- or -ARs and to identify which ₁-AR subtypes are responsible for the effect of basal endogenous NE to contribute to restenosis after balloon injury (as indicated by the benoxathian and prazosin experiments), antagonists for these ARs (all at 10 µmol/l minipump concentration, 5 µl/h) were suffused for 2 wk beginning immediately after balloon injury (Figs. 6-8). "Uninjured" vessels in Figs. 6-8 consisted of the sham-injured right carotid artery (exposed but without balloon injury). Because of changes in vessel size (i.e., circumference of the EEL), vessel layer thicknesses were also determined. Propranolol and the ₂-AR antagonist RX-821992 had no significant effect nor did the _1D-AR antagonist BMY-7378 or the _1B-AR antagonist AH11110A (Figs. 6-8). In contrast, the _1A-AR antagonist KMD-3213 attenuated the decrease in lumen area by 70%, reduced neointimal area and thickness by 54% and 55%, abolished the increases in adventitial area and thickness, and reduced inward remodeling by 84% (Figs. 6-8).

View larger version (50K): [in this window] [in a new window]   Fig. 6.   Perivascular suffusion of 1-AR antagonist KMD-3213 (KMD) attenuated lumen loss, decrease in circumference, and increase in adventitial area and thickness. 2-AR (RX821002; RX) and -AR (propranolol; PROP) antagonists had no effect on response to balloon injury [vehicle (Veh) vs. uninjured right carotid (UI)]. In all experiments, antagonists were suffused for 2 wk at 10 µmol/l osmotic pump concentration. See Fig. 1 for additional details. Dotted line, aid in comparison of drug treatment groups to vehicle group.

View larger version (47K): [in this window] [in a new window]   Fig. 7.   Perivascular suffusion of 1B-AR (AH11110A; AH) and 1D-AR (BMY-7378; BMY) antagonists had no effect on response to balloon injury [Veh vs. UI]. Veh and UI groups are different from those in Fig. 6; this Veh group had stronger injury than Veh group in Fig. 6, which accounted for the differences in media between the two vehicle groups. In all experiments, antagonists were suffused for 2 wk at 10 µmol/l osmotic pump concentration. See Fig. 1 for additional details. Dotted line, aid in comparison of drug treatment groups to vehicle group.

View larger version (44K): [in this window] [in a new window]   Fig. 8.   Effect of AR antagonists, for groups shown in Figs. 6 and 7, on growth of neointima after balloon injury. In all experiments, antagonists were suffused for 2 wk at 10 µmol/l osmotic pump concentration. See Fig. 1 for additional details and Figs. 6 and 7 for abbreviations. Dotted line, aid in comparison of drug treatment groups to vehicle group.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The first major finding of this study was that chronic local elevation of wall NE in the balloon-injured rat carotid artery further increased neointimal area by 64%, reduced lumen area by an additional 45%, reduced vessel circumference, and increased cell number and collagen content in the neointima (Figs. 3 and 4). Expansion of the adventitial area by balloon injury, known to be associated with AFB proliferation (35), was not enhanced by elevated NE, even though NE stimulates proliferation of AFBs in cell culture (13) and in the uninjured and balloon-injured aorta in organ culture (38). Instead, NE caused adventitial area of the injured carotid to decrease (Fig. 3). However, this was a geometric consequence of the reduction of vessel circumference by NE, because in the uninjured vessel, NE caused virtually the same decrease in EEL and adventitial area, resulting, as well, in no change in adventitial thickness (Fig. 1). There is evidence that balloon injury may stimulate AFBs to migrate to the intima (35) and that NE itself stimulates migration of AFBs in vitro (39). Thus the absence of induction of adventitial hypertrophy by NE, despite its potential proliferative action on AFBs, could arise from enhanced centripetal migration of AFBs or could reflect the complex remodeling mechanisms activated by balloon injury.

The second perhaps more important finding of this study was that chronic 2-wk administration of nonspecific ₁-AR and selective _1A-AR subtype antagonists reduced neointimal growth and lumen loss, whereas _1B-, _1D-, ₂-, and -AR antagonists were all without effect. The critical time for ₁-AR blockade was during the initial 2 wk, because blockade for 4 wk with prazosin had no greater effect than 2 wk with benoxathian. These data suggest that basal endogenous NE levels augment experimental restenosis. In previous studies of the balloon-injured rat aorta maintained in organ culture, the threshold for NE-induced growth (48 h exposure) was between 10 and 100 nmol/l (38). This is 5- to 50-fold higher than the resting arterial blood level that, itself, increases with aging and can increase 10-fold with stress (19, 25). Plasma NE represents mostly spillover from nerve-derived concentrations in the wall of innervated vessels, where estimates range from 1 to 10,000 nmol/l over 1-10 Hz stimulation, depending on the proximity to varicosities and nerve density (2), which is less dense in carotid than in heavily innervated vessels such as superior mesenteric artery (4, 5).

These effects cannot be attributed to systemic hemodynamic actions of NE or the antagonists because the doses administered (in mg · kg¹ · day¹) are at least several hundred times below doses required for a detectable effect on arterial pressure. For example, the 3.6 ng · kg¹ · min¹ NE dose we used is 70 times below the intravenous dose (0.25 µg · kg¹ · min¹) that raises plasma NE approximately fivefold and that is just subthreshold for producing a detectable increase in mean arterial pressure in conscious rats (24, 32). Because NE was suffused extravascularly rather than intravenously, the plasma concentration was at least 150-fold times below (compare data in Refs. 7, 24, and 32) and more likely 700 times or more below this threshold, given the metabolic barrier to the perivascular route of delivery. Like NE, the actions of the effective antagonists (prazosin, benoxathian, and KMD-3213) also cannot be attributed to a systemic hemodynamic effect, because the total dose of each antagonist given outside of the carotid sheath was 10 µg · kg¹ · day¹, which is at least 1,000 times below the dose required to be given extravascularly to produce a just-threshold effect on arterial pressure.

Importantly, our results suggest that earlier findings of attenuated neointimal growth after balloon injury by systemic infusion of ₁-AR antagonists (14, 18, 29, 34) likely resulted from blockade of the direct hypertrophic effect of NE identified herein and possibly also from reduction in arterial pressure or in downstream resistance and thus flow-induced outward remodeling. The present antagonist results suggest that even basal endogenous NE levels in the vascular wall may increase wall growth induced by injury. Moreover, the NE results suggest that prolonged sympathoexcitation, secondary to physiological or behavioral stress, may augment growth of injury-induced vascular wall lesions. These findings also validate past in vitro studies and provide the rationale required for future studies to identify the signaling mechanisms responsible for the effect of injury to so strongly augment the trophic actions of NE as identified by us herein, previously (38), and in preliminary reports (10, 13) examining two different models of vascular injury in mice genetically devoid of catecholamine synthesis and specific ₁-AR subtypes.

In the uninjured carotid, 2-wk suffusion of NE caused a modest inward remodeling, without hypertrophy of media or adventitia (Fig. 1); the greater effect on lumen area is a consequence of its nonlinear relationship to circumference. In contrast, a small hypertrophic effect was evident in organ culture studies of uninjured rat aorta, where treatment with 1 µmol/l NE (a much higher concentration than endogenous wall levels, albeit only presented for 48 h) caused media protein content to increase by 10% (no increase in cell number) and caused AFB cell number and protein to increase by 10% (38). This is consistent with the action of NE in the cell culture to cause dose-dependent (0.005-1 µmol/l) hypertrophy of quiescent SMCs (37) and proliferation of quiescent AFBs (13) seen after as little as 12 h of NE exposure. It is possible that the different responses reflect differences in the in vitro and in vivo models or to duration of exposure and concentration of NE. Vascular wall hypertrophy is commonly observed with chronic systemic infusion of NE, although past studies have only examined doses that raise arterial pressure (7, 20, 21, 24, 32). However, physiological sympathoexcitation simultaneously increases both wall NE and arterial pressure. Because pressure-dependent increase in wall stress, per se, is believed to promote medial and adventitial growth, it is possible that, under physiological circumstances, the intracellular signals activated by the direct trophic effect of NE combine with signals concomitantly activated by increased wall stress to achieve the final amount of wall hypertrophy and remodeling observed. Concomitant increases in sympathetic activity and arterial pressure, combined with local pharmacological or genetic blockade of ₁-ARs, will be required to test this hypothesis.

It is not possible to determine the concentration of NE across the carotid wall that was achieved by the suffusion (5 µl/h) of the NE concentration present in the osmotic pump (100 µmol/l) at a point several hundred micrometers away from the intact carotid sheath. However, KMD-3213 is selective for blockade of _1A-ARs at 0.1 µmol/l in vitro (1, 13, 38, 39) and appeared to be selective in the present study when present in the pump at 10 µmol/l and suffused at 5 µl/h. This, together with the much greater biological half-life of KMD-3213 (12 h; Ref. 1) than NE (several minutes), suggests that the NE concentration achieved in the injured carotid wall was likely to be <1 µmol/l, which would be within the physiological range for dose-dependent, NE-mediated SMC and AFB trophic and chemotactic activity as determined in previous cell and organ culture studies (13, 37-39). In the present study, blockade of ₁-ARs with benoxathian, while significantly attenuating restenosis, had no effect on the uninjured carotid. These results suggest that basal endogenous NE levels may exert little, if any, tonic trophic effect on the vascular wall of healthy arteries of unstressed normotensive individuals; however, more prolonged elevation of NE alone, or normal or elevated NE levels in the presence of mediators activated by injury or increased arterial pressure, may do so.

A limitation of this study is the specificity of the antagonists employed, although we used antagonists with the highest selectivity available. Antagonists were delivered at the same concentration and rate, based on their affinity and selectivity (given in MATERIALS AND METHODS) and the following considerations. Benoxathian, prazosin, propranolol, and RX-821002 have similar high affinities (~0.5 nmol/l) and excellent selectivities (1,000-fold) for their respective AR types. In the present study, benoxathian and prazosin had almost identical inhibitory effects, whereas propranolol and RX-821002 were without effect. Higher concentrations of propranolol and RX-821002 were not tested because of the difficulty of these in vivo experiments and the large numbers of animals required, because the similar affinities of propranolol and RX-821002 to those of benoxathian and prazosin would predict that they were present in inhibitory concentrations, and because we previously found no contribution of ₂- and -ARs in studies of NE-induced growth of media or adventitia of injured arteries in organ culture (38) or of cultured SMCs (37) where maximal blocking concentrations were achieved. Nevertheless, a greater contribution to restenosis of _1A-AR stimulation, together with possible participation of _1B-, _1D-, _2D-, and/or -AR subtypes, cannot be completely ruled out by the current studies where full AR blockade cannot be assured. However, our previous in vitro studies do not support a significant trophic contribution of _1D-, _2D-, or -ARs (37, 38) nor do in vivo studies employing mice made genetically devoid of _1D-ARs (29).

The ₁-AR subtype antagonists do not have the very high selectivities of the above nonsubtype-selective ₁-, ₂-, and -AR antagonists. KMD-3213 has a higher affinity (~0.3 nmol/l for the _1A-AR) and selectivity than do BMY-7378 and particularly AH11110A for each of their targeted subtypes (see MATERIALS AND METHODS). However, at 0.1 µmol/l, each antagonist exhibited good selectivity for its respective subtype in previous studies of SMC and AFB growth and migration in cell and organ culture (37-39). Thus in the present study they were suffused at a 100-fold higher concentration to account for diffusional barriers during perivascular suffusion. The difficulty of using more animals (than the >200 already required) with this in vivo model, plus the lower affinity and selectivity of the available ₁-AR subtype antagonists, prevented the study of multiple concentrations. Thus the full contribution of _1A-ARs, and possible effects of other ₁-AR subtypes, may not have been identified. However in recent organ culture studies, _1D-ARs, which mediate constriction of the rat carotid and aorta (9), had no trophic effect, whereas _1A-ARs mediated SMC proliferation (38). It is noteworthy that AH11110A is the least satisfactory antagonist used herein, having an ~70 times lower affinity for its targeted subtype (_1B-AR) than KMD-3213 and BMY-7378 have for theirs, together with a low selectivity (~20-fold) over the _1A- and _1D-ARs (see 38 for Refs.). Thus in vivo confirmation of these results awaits development of more selective ₁-AR subtype antagonists or genetic methods for conditional manipulation of local NE concentration and adrenoceptor subtype signaling. In this regard, we have recently found that injury-induced hypertrophic outward remodeling of the carotid is abolished in gene knockout mice devoid of dopamine -hydroxylase or _1B-ARs, but unaffected in _1D-knockout mice (11). Similar results were obtained for an additional type of injury, i.e., hypoxic pulmonary vascular remodeling (11). In addition, chronic systemic administration of KMD-3213 at levels without effect on systolic or diastolic pressure, heart rate, or regional resistances, dose dependently reduced neointimal growth after balloon injury of the rat carotid (10).

How injury augments the vascular trophic effects of NE, or vice versa, is an important question for future studies. Contraction of SMCs may not be involved, because BMY-7378 at a concentration (0.1 µmol/l) that inhibits rat aorta contraction whose constriction is well known to be mediated by the _1D-AR (9) had no affect against NE growth of the same vessel in vitro (38); neither was it effective in the present in vivo study. Also, NE causes growth of AFBs that do not constrict (13, 38). As well, SMC contractility and NE constriction are reduced when measured several days and 2 wk after injury (3) at a time when the trophic effect of NE is augmented (38). Expression of mRNAs for all three ₁-AR subtypes and the _2D-AR are decreased by similar amounts (~50%) at day 4 after injury in neointima, media, and adventitia, whereas _1A- and _2D-AR transcript levels return to normal by 21 days, although at this time total ₁-AR density and _1B and _1D mRNA levels remain reduced by ~50% (12). Thus an increase in ₁-AR receptor density does not underlie the increased NE trophic response after injury. Increased emigration of inflammatory cells by a direct effect of NE is not supported because monocytes, macrophages, T cells, and neutrophils do not possess ₁-ARs, and because injury also strongly augments the trophic actions of NE in vitro in vessel organ culture (38). On the other hand, augmentation of the trophic activity of NE by injury may involve an increase in wall NE content after injury (3), which may arise from neurite outgrowth. Additional possibilities are interaction of NE with peptide growth factor mechanisms activated by injury (31, 35) or increased sensitivity of SMCs and AFBs to the trophic activity of NE, secondary to their already being displaced toward proliferation by other mediators induced by injury itself.

The current study suggests that NE may be capable of increasing collagen content in the neointima, presumably through actions on SMCs and/or AFBs that have migrated to and proliferated therein. There are no previous reports examining the effect of NE on collagen in cell or organ culture of SMCs or AFBs. There is also evidence that the adrenergic cotransmitter neuropeptide Y induces proliferation of cultured SMCs (40). Moreover, it is interesting to note the parallels between the current findings and evidence suggesting a direct _1A-AR involvement in hypertrophy of cardiac myocytes (15, 17) and smooth muscle-like stromal cells of the prostate (1, 15, 28).

In conclusion, this is the first in vivo study to identify that NE is directly trophic for SMCs and AFBs of the vascular wall. This action may have adaptive structural significance in the uninjured wall by inducing wall thickening, and thus normalizing wall stress, during elevated arterial pressure in chronic sympathoexcitatory states. Whether catecholamine trophic actions contribute to arteriosclerosis or hypertensive hypertrophy, which are known to be associated with increased sympathetic activity and aging in humans (8), or limit compensatory outward remodeling (35), merits additional study. This same trophic effect may worsen intimal lesion growth induced by surgical injuries and by atherogenic processes.