Neointimal growth has been a subject for intensive research during several decades. A collar placed around a target vessel is one of the models that have been used to induce concentric intimal thickening in different animal species under macroscopically intact endothelium [1]. However, good models for asymmetric human-like lesion formation and neointimal growth are still lacking especially if endothelial denudation should be avoided. The role of adventitia in the neointimal formation has not been studied very much and interest has been merely focused on medial smooth muscle cells (SMC) and endothelium. However, new interest in the role of adventitia has emerged due to several findings concerning the role of adventitia in the production of reactive oxygen species (ROS) [2], expression of inducible nitric oxide synthase (iNOS) [3]), extracellular matrix production and degradation [4], neovascularization [5] and synthesis of various growth factors [5].

Changes in shear stress are important determinants of intimal hyperplasia [6, 7]. Disturbed flow in curved regions and branches is proatherogenic, whereas laminar flow is atheroprotective. Laminar shear stress upregulates several genes involved in cell survival, cell cycle arrest, morphological remodelling and NO production [8]. Most models used for these studies have utilized isolated vessels in perfusion chambers, which is far from in vivo situation. Marano et al. used ligation of the right carotid artery thus causing high shear stress on the contralateral collared artery [9].

Collar models used previously have been based on symmetrical hollow collars [1]. In contrast to human atherosclerotic lesions this collar constantly produces a symmetric concentric intimal thickening (Figure 1). Here we describe a new collar model that reproducibly causes an asymmetrical eccentric neointimal thickening in rabbit carotid and femoral arteries.

After eight weeks, neointimal formation was detected in the collared arteries. Contrary to the previously used rabbit carotid collar models [1, 11, 12] predominantly asymmetric neointimal formation was observed, which in several cases resembled human lesions (Figure 3). Another new feature in this model was the gradual increase in the neointimal thickening along the flow direction in carotid segments. Intima/media ratio increased towards the distal end of the collar being 0.34 ± 0.09 in caudal segments, 0.48 ± 0.12 in caudomedial segments, 0.92 ± 0.38 in craniomedial segments (p = 0.002 vs. caudal) and 0.59 ± 0.17 in cranial segments. However, such findings were not seen in femoral arteries. (Figures 4, 5.)

Effect of the dotted collar on the velocity of the carotid artery blood flow was measured with Doppler ultrasound. Attenuation of the peak flow velocity in the carotid arteries was seen after the placement of the collar. However, the attenuation was modest and returned to normal levels after eight weeks follow-up despite of the development of neointima. In native, non-operated carotids blood flow proximal to the site of the collar was 1.27±0.20 m/s, blood flow immediately after the collar placement was 0.77±0.06 m/s. The corresponding value eight weeks after the collar implantation was 1.17±0.21 m/s. (Figure 6).

Immunohistochemical analysis showed macrophage accumulation in areas of the thickest neointimal formation and in adventitia (Figures 3 m-p; 5). In some collared vessels the density of macrophage accumulation also increased with the increasing neointimal area and according to the blood flow (Figure 5 a-d). The number of macrophages in the lesions varied between 5 to 20 cells/100 μm². Intact endothelium was present at eight weeks in almost all segments (Figure 3 e-h). Endothelial layer was always present in areas of large, plaque-like lesions. In a few sections, some local detachment of endothelial cells was noted, but this did not correlate with the increase in the neointimal thickening, and was mainly considered as a sporadic finding. The bulk of the lesions consisted of SMC as shown by α-actin immunostaining (Figure 3 i-l). In some lesions, accumulation of macrophages was also seen in the deep layers of the lesions (Figure 5 c,d,f) resembling early stages in atheroma formation. Lesion neovascularization could be seen increasing towards cranial (distal) segments in a similar fashion as plaque thickness, as detected by CD31 immunostaining (Figure 3 e-h).

Molecular and cellular factors leading to neointimal growth include proliferation and migration of vascular SMC, matrix remodelling, lipid accumulation, LDL oxidation, endothelial dysfunction and inflammatory reactions [13]. Physical determinants proposed to explain neointimal growth in the collar model are changes in shear stress, occlusion of vasa vasorum thus affecting oxygen balance and metabolic status, loss of perivascular innervation and twisting of the collared artery [11]. Marano et al. showed that increased shear stress was protective against neointimal thickening in rabbit carotid collar model [9]. However, they used hollow collars with an even inner surface as originally developed by Booth et al. [1]. In the new model our intention was to alter the normal laminar flow to a turbulent one by twisting the vessel thus giving differential pressure and flow signals on the endothelium.

Vascular adventitia is known to be activated in a variety of cardiovascular diseases. It has been postulated that adventitial fibroblasts proliferate and migrate into the neointima in response to injury [2, 14]. It has also been shown that direct adventitial injury in the absence of endothelial damage is sufficient to cause neointimal lesions [15]. Barker et al. [16] removed adventitia in hypercholesterolemic rabbits, which induced intimal hyperplasia and macrophage accumulation. Our dotted collar imposed an altered pressure on adventitia and we hypothesized that this is sufficient to disturb normal homeostasis in the vessel wall and cause lesion formation. It was found that dotted collar induced the formation of asymmetric SMC-rich lesions that also accumulated macrophages and started to resemble human-like lesions.

The effect of the collar placement in femoral versus carotid arteries has not been studied before. In this series a clear difference in the reactivity to the dotted collar was seen between carotid and femoral arteries: collar-induced neointimal thickening in the femoral arteries was only mildly increased along the direction of the blood flow and did not reach statistical significance. We consider this difference in the neointimal development probably as a result of the differential adventitial pressure. However, differences in the anatomical locations of the arteries may also have contributed to these findings. Limitations of this study include limited technical details in ultrasound acquisition: in future with intra-arterial flow probe the turbulence inside the dotted collar can be studied and correlated to the neointimal changes. Also, collar device can be used for local drug, protein and gene delivery [12, 17, 18], which should allow further evaluation of the pathogenetic mechanisms under macroscopically intact endothelium.

In conclusion, the dotted collar model causes asymmetrical neointimal lesion formation with the accumulation of macrophages. This model should be useful for studies addressing the pathogenesis of neointimal SMC accumulation and the formation of eccentric atherosclerotic lesions in large arteries.