ALLO- AND AUTOTRANSPLANTATION OF CAROTID ARTERY-A NEW MODEL OF CHRONIC GRAFT VESSEL DISEASE

Evaluation by Magnetic Resonance Imaging and Histology

Abstract

Background. 

Graft vessel disease is a special form of accelerated arteriosclerosis. Because immunological and nonimmunological factors can contribute to graft vessel disease, we developed a model that enables the study of both factors simultaneously.

Methods. 

A carotid artery was allografted from DA to Lewis rats, with the excised native artery autografted on the contralateral side. Five groups of six to seven rats were treated for 8 weeks with vehicle (placebo) or cyclosporine (CsA) (0.3, 1, 3, and 10 mg·kg^-1·day^-1), which was administered using subcutaneous osmotic minipumps. The carotid lumen area was estimated in vivo at 2, 4, and 8 weeks by magnetic resonance imaging (MRI); CsA blood levels were determined twice. Carotid neointimal thickening and medial and luminal area were measured with histological techniques.

Results. 

MRI showed bulging of the allografts but not autografts. Bulging disappeared over time with narrowing of the allograft lumina estimated by both MRI and histology. Histologically, vehicle-treated animals developed a massive neointima, which was inhibited in a dose-dependent manner by CsA. Autografts remained normal except for minimal subintimal thickening of two of four arteries in the group given the highest dose of CsA. Cellular rejection was detected in the allografts of all but the highest-dose group. The CsA blood levels were similar to those used in man at the two lower doses and about 10-fold higher at the highest dose.

Conclusions. 

Subintimal thickening did not correlate with in vivo lumen size, a phenomenon that we have previously described for balloon catheter-induced lesions. CsA blood concentrations similar to those used in patients suppressed neointima formation in part, and 10-fold higher concentrations almost completely suppressed neointima formation.

Transplanted organs frequently develop graft vessel disease (GVD^*), a form of accelerated arteriosclerosis that results in chronic rejection ⁽¹⁾, which is responsible for over 50% of organ failure. Vascular changes are seen in 40-60% of grafts within the first 5 years of transplantation ⁽²⁾, presenting as typically, persistent perivascular inflammation and concentric neointima formation ^{(3, 4)} often combined with medial thinning and necrosis ⁽⁵⁾.

In man, development of GVD is slow, and therefore impractical to study. Animal models may mimic only certain aspects, making extrapolation to man difficult. To address some of these problems, we developed an animal model that allows investigation of immunological and nonimmunological processes driving GVD ^{(6, 7)} concurrently in one individual. This should avoid many confounding factors, resulting in observations that are relatively easy to interpret, and should lower the number of animals required. Our new model consists of simultaneous allogeneic and autologous transplantation of rat carotid arteries. Allografts and autografts were subjected to exactly the same procedures.

We hoped to resolve the old controversy of whether cyclosporine (CsA) is able to suppress GVD ^{(8, 9)}, or even enhance such processes ^(10-13). We pursued this aim by using a wide range of CsA doses and a methodology that permitted a constant exposure with defined blood levels. Most previous studies have neglected this point, administering the compound usually once a day to rats even though in man b.i.d. administration is the rule. We used osmotic minipumps to achieve constant CsA blood levels and monitored the blood levels twice, in the middle of the pumping cycle and at the end, when the pumps were supposed to be nearly empty.

We hypothesized that the autotransplant, which does not benefit from immunosuppression, might show potential ill effects of high doses of CsA.

Lumen size has not previously been assessed in animal studies of GVD, although it is the degree of luminal obstruction rather than the thickness of the neointima that eventually causes perfusion problems. A further major purpose of this study was to assess the effects of CsA on lumen size in vivo and to correlate this with histological morphometric measurements. We based this approach on our previous experience with comparative studies on lumen size of rat carotid arteries after balloon-catheter injury ^{(14, 15)}. We had observed that there was no correlation between histological neointimal thickness and lumen size measured in vivo with magnetic resonance imaging (MRI), but that there was a correlation in histological specimens obtained after careful fixation in situ under physiological pressure. It appeared relevant for clinical studies to investigate how the immunological lesions of GVD would manifest in vivo with respect to lumen size.

MATERIALS AND METHODS

Experimental animals. Rats from inbred strains DA (RT1^a, donors) and Lewis (RT1¹, recipients), weighing between 320 and 340 g, were used. All animals were purchased from Harlan, Zeist, The Netherlands. They were allowed unrestricted access to food and water before and after the operation. Handling and care corresponded to the Swiss federal law for animal protection.

Carotid artery transplantation. The rats were anesthetized with isoflurane (Abbott, Cham, Switzerland; 4-5% for induction, 1.5-2% for maintenance), and 300 μg of atropine sulfate was injected subcutaneously after the induction. The left carotid artery was carefully dissected free without damaging the accompanying nerve and vein. The artery was clamped proximally and distally, and a segment of about 7-10 mm was removed and stored in ice-cold saline for 45 min for autologous grafting on the contralateral side. An allograft that also had been subjected to 45 min of cold ischemia was then inserted into the gap. Ethilon 10/0 sutures were used for the end-to-end anastomoses. Finally, the skin was closed with 4/0 Suturamid sutures. An Alzet osmotic minipump (Alza Corp., Palo Alto, CA) was then implanted subcutaneously on the back.

Study design. The rats were randomly subjected to one of the following treatments: vehicle (placebo) or CsA 0.3, 1, 3, or 10 mg·kg^-1·day^-1. Except for the last group (n=6), all groups consisted of seven animals. Vehicle and drugs were infused for 8 weeks using Alzet osmotic minipumps implanted subcutaneously: 2-week minipumps (mean pumping rate: 5 μl·hr^-1) for the first 4 weeks and 4-week minipumps (mean pumping rate: 2.5 μl·hr^-1) for the second half of the treatment period. In the group receiving the highest dose, 2-week pumps were used throughout.

The body weight of the recipients was recorded on the day of transplantation and after 1, 2, 4, 6, and 8 weeks. CsA blood levels were assayed at 1 and 8 weeks. The carotid lumen area was determined in vivo at 2, 4, and 8 weeks by MRI. The Alzet minipumps were exchanged after each measurement. The rats were killed at 8 weeks, and in situ perfusion-fixation was carried out at physiological pressure as described below.

Histological and morphometric analyses of the neointimal thickening, cross-sectional area of the media, and luminal diameter of the allo- and autografted carotid arteries were performed postmortem in a blind randomized fashion with 5, 4, 5, 6, and 5 allografted carotid arteries and 6, 6, 6, 6, and 4 autografted carotid arteries for the placebo and CsA 0.3, 1, 3, and 10 mg·kg^-1·day^-1 groups, respectively.

Statistics. Statistical analysis was done using Student's t test, Dunnett's test, and Pearson's correlation together with probability testing using SYSTAT version 5.03 (Systat, Inc., Evanston, IL). If the variances differed by more than the power of 10, the exact Wilcoxon-Mann-Whitney test was performed using StatXact-3. Values of P<0.05 were considered statistically significant.

Determination of CsA blood levels. The CsA whole blood levels were determined by a competitive enzyme-linked immunosorbent assay with a limit of quantification of 2 ng·ml^-1. The sample preparation involved extraction with methanol, centrifugation, decantation of the supernatant, evaporation, and reconstitution in Trizma-buffered saline/Tween 20 (0.05%)/bovine serum albumin (1%) buffer. Biotinylated CsA ((O-(2-biotinoyloxi-ethyl)Thr-2)cyclosporin) was added as a tracer, followed by the addition of the specific monoclonal antibody 78-299 to CsA. An aliquot of this mixture was transferred to a goat anti-mouse/IgG(Fc)-coated microtiter plate and incubated at 4°C overnight. A colorimetric reaction was produced by addition of peroxidase-labeled streptavidin and o-phenylenediamine dihydrochloride-substrate. The optical density was measured with a microtiter plate reader.

Magnetic resonance imaging. MRI experiments were performed on a Biospec 47/15 spectrometer (Bruker, Karlsruhe, FRG) equipped with a self-shielded gradient system with an inner diameter of 85 mm. The rats were anesthetized with isoflurane 1.5% (Abbott) in nitrous oxide:oxygen (2:1) administered via a face mask and positioned in supine position on a support made from glass fiber-enforced epoxy resin. Their necks were positioned above a single-turn surface coil of 3.5-cm diameter tuned to 200 MHz, which was used as radio-frequency transmitter/receiver. Images were acquired using a rapid gradient-recalled echo pulse sequence (snapshot FLASH; ¹⁶) with radiofrequency pulse angles of 45°. With this choice of pulse angles, signals due to stationary tissue were effectively suppressed while flowing blood yielded high-intensity signals, resulting in angiographic images. The field of view was 25.6 mm, with dimensions of the imaging matrix of 128×128 pixels and slice thickness of 1 mm. For image reconstruction, zero filling to 512×512 pixels was applied. The spectral width was set to 25,000 Hz, and the repetition and echo delay was set to 8.5 and 3.6 msec, respectively. Eight averages were accumulated, which resulted in a total imaging time of 8.2 sec/image.

Images were analyzed using the UXNMR software package supplied by Bruker. The lumen of the two carotid arteries was defined by interactively drawing a border and measuring within the region of interest. Twenty slices covering a distance of 20 mm were analyzed using the carotid bifurcation as the reference point (corresponding to 0 mm). This covered the full length of the transplant as well as approximately 5 mm of the native vessel at both ends of the transplant.

Perfusion fixation and histological examination. We have previously reported our method of perfusion fixation at physiological pressure ⁽¹⁴⁾. In brief, the rats were anesthetized with sodium pentobarbital (60 mg·kg^-1 i.p. Vetanarcol, Veterinaria AG, Zurich, Switzerland), a perfusion catheter was inserted through the left ventricle into the aortic arch, and an aspiration cannula was inserted into the right ventricle. The animals were perfused for 1 min with 0.1 M phosphate-buffered saline solution (pH 7.4) and then for 15 min with 2.5% glutaraldehyde in phosphate buffer (pH 7.4) at a perfusion pressure of 150 mmHg at the tip of the cannula ⁽¹⁴⁾. The carotid arteries were excised and incubated overnight at 4°C in 0.1 M cacodylate buffer (pH 7.4) containing 7% sucrose, then shaken for 1 hr at room temperature in 0.05% KMnO₄ in 0.1 M cacodylate buffer, rinsed with water, and dehydrated in a graded ethanol series. The carotid arteries were then infiltrated in RWL embedding medium (based on methacrylate; RWL Histotechnologie, Vagen, FRG) according to the manufacturer's recommendation and embedded in silicon molds. After overnight polymerization under argon, 1- to 2-μm-thick sections were cut from the central area of each carotid artery and stained with Giemsa stain.

Morphometry. The morphometric measurements were carried out as described previously ⁽¹⁴⁾ at magnification ×50 by using the microcomputer imaging M4 system version 2.0 (Imaging Research Inc., St. Catharines, Ontario, Canada). Three measurements were taken and analyzed from each section, and the means of these measures were used as final values. We measured the lumen area and the thickness of the media and intima separated by the external and internal elastic laminas, respectively.

Qualitative analysis. Morphological changes were evaluated from histological sections of the central areas of the transplanted carotid arteries using an optical microscope (Zeiss). The allo- and autografts were scored on an 0 to 3 scale for adventitial infiltration of mononuclear cells and necrosis (vacuolar degeneration, hypertrophy of cells), the number of smooth muscle cell (SMC) nuclei in the media (0-10, 11-100, 101-200, and >>200 nuclei for scores of 0, 1, 2, and 3, respectively), medial SMC necrosis (vacuolar degeneration and hypertrophy of SMC), and the intimal infiltration of mononuclear cells ⁽¹³⁾.

RESULTS

Fate of the animals, body weight, and mortality. During the first postoperative week, most animals lost weight (maximum of 3% in the CsA 0.3 mg·kg^-1·day^-1 group, not dose related); thereafter, the body weight in all groups increased linearly (mean 19%) during the follow-up period.

In the early phase of the experiment, we lost one animal from each group, except the 10 mg·kg^-1·day^-1 CsA group, during the MRI measurements. Since respiratory complications were the cause of the problem, we subsequently administered 100 μg·kg^-1 atropine s.c. 15 min before anesthesia. In addition, respiration was monitored using a strain gauge system based on a rubber belt.

Two animals from the 10 mg·kg^-1·day^-1 CsA group were excluded from the analysis due to malfunction of the osmotic minipumps (see also CsA blood levels, below).

CsA blood levels. The CsA blood levels were determined after 1 week of treatment, which corresponded to the midlife of the 2-week minipumps. The following mean values were obtained: 76±9, 188±17, 636±51, and 2186±88 ng·ml^-1 for the 0.3, 1, 3, and 10 mg·kg^-1·day^-1 CsA dose groups (n=7 for all except the last group, where n=6). At 8 weeks, the end of the 4-week infusion period of the 4-week minipumps, the CsA blood levels were measured for the 0.3, 1, 3, and 10 mg·kg^-1·day^-1 CsA groups and were as follows (mean ± SEM, expressed as ng·ml^-1): 61±11 (n=6), 141±28 (6), 632±174 (n=6), and 1460±235 (n=3).

As stated previously, the highest CsA dose had to be delivered in 2-week pumps, since the concentration needed for the 4-week pumps could not be achieved. Despite this, we lost two animals because of pump dysfunction (pump not empty at time of change). These animals were excluded. Thus, the minipumps produced comparable blood levels throughout their cycles for all doses except the highest, where levels were 33% lower (and where two out of six pumps failed completely).

Magnetic resonance examination of carotid lumen areas. The lumen areas of both allo- and autografted carotid arteries were measured in vivo by MRI at 2, 4, and 8 weeks. Figures 1 and 2 give an overview of all measurements obtained at 2 weeks for the allografts and the autografts, respectively. The lumen of the autografts (Fig. 2) showed an even longitudinal profile similar to that of normal carotid arteries. The absolute lumen areas of the untreated autografts were 0.68±0.12 mm² and 0.64±0.10 mm² at the proximal and distal anastomoses, respectively, and 0.62±0.07 mm² at the center of the graft. In contrast, the untreated allografts showed absolute lumen areas of 0.54±0.08 mm² and 0.59±0.11 mm² at the proximal and distal anastomoses and 0.77±0.15 mm² at the center of the transplant (Fig. 1). Compared with the autografts, the anastomotic region thus tended to be narrower and the center tended to be dilated. CsA treatment had no relevant influence on the lumen diameter of the allo- or autografts at 2 weeks. This pattern of bulging is further elaborated in Figure 3 for all MRI measurements obtained throughout the time course of the experiments. It is obvious that treatment had no relevant influence on the bulging, which, however, diminished with time.

At 2 weeks the absolute lumen areas at the center of the allografts tended to be larger than those of the autografts in all but the 0.3 and 10 mg·kg^-1·day^-1 CsA dose groups (Fig. 1 vs. Fig. 2). However, 4 weeks after transplantation, despite the bulging, the allografted carotid arteries tended to have smaller absolute lumen areas than the autografts, except for the 10 mg·kg^-1·day^-1 CsA dose group (data not shown). At 8 weeks the findings were similar, except that now the allografts treated with 3 mg·kg^-1·day^-1 CsA had significantly (Student's t test) larger lumina than the autografts (Fig. 4, top).

Postmortem histological evaluation of the rat carotid arteries. Histological specimens from the midpoint of the allo- and autografted carotid arteries were evaluated morphometrically.

Lumen: Except for the 10 mg·kg^-1·day^-1 CsA dose group, the lumen of the allografts tended to be smaller than the lumen of the autografts (Fig. 4, bottom). Within the CsA-treated groups, the allografts showed a significant increase in the luminal area with higher CsA doses, as shown in Figure 5 (linear regression, P=0.011, correlation coefficient r=0.554).

Neointima:Figure 6 shows a massive concentric neointima formation in the allografts in the placebo and 0.3 mg·kg^-1·day^-1 CsA groups (71,760±14,523 and 71,993±16,234 μm², respectively). Higher doses of CsA elicited significant, dose-dependent inhibition of neointima formation (linear regression, P<0.001, correlation coefficient r=-0.76; data not shown). It is noteworthy that CsA blood levels within the human therapeutic range (76±9 and 188±17 ng·ml^-1 at the 0.3 and 1 mg·kg^-1·day^-1 CsA doses, respectively) caused only a modest inhibition of the neointima formation. In the CsA-treated groups, neointima formation in the allografts correlated significantly with luminal narrowing, as shown in Figure 7 (linear regression, P=0.0004, correlation coefficient r=-0.72).

In the autografted carotid arteries, two out of four arteries in the highest CsA dose group showed focal intimal thickening. Only minimal foci were detected in two out of six autografted carotid arteries in the 1 mg·kg^-1·day^-1 CsA dose group, and no foci were found in the placebo, 0.3, and 3 mg·kg^-1·day^-1 CsA dose groups.

Media: The size of the medial area in the allo- and autografted carotid arteries was not altered by CsA treatment (data not shown), but in the placebo group, the medial area of the autografts was significantly higher than that in the allografts (Student's t test).

Comparison of histological and MRI measurements. The luminal areas of the allo- and autografted carotid arteries, measured in vivo by MRI, were always larger than those estimated histologically postmortem (Fig. 4, top vs. bottom). Therefore, the data are also expressed as ratio of allograft to autograft lumen area for both the in vivo MRI data and the postmortem histological data (Fig. 8). The histological measurements show a clear trend to larger ratios with higher CsA doses. The MRI measurements are more difficult to interpret, since the allograft to autograft ratio declined at the highest dose and also because untreated animals had a surprisingly high ratio.

No correlation was found between measurement of carotid lumen area for auto- and allografts by the two techniques (data not shown).

Qualitative histological analysis of the carotid arteries. To examine the degree of GVD, midpoint histological sections of allo- and autografted carotid arteries were scored for neointimal and adventitial infiltration of mononuclear cells, the necrosis of the media and adventitia, and the number of SMC nuclei in the media. Representative examples are shown in Figures 9 and 10.

The allografts of the placebo and 0.3 mg·kg^-1·day^-1 CsA groups formed a massive neointima. Evidence of cellular rejection was detected in the allografts of all but the 10 mg·kg^-1·day^-1 CsA dose group, as shown in Figures 9 and 11. The neointima and adventitia of these carotid arteries showed an infiltration of macrophages, lymphocytes, and neutrophil granulocytes, which gradually decreased with an increase of the CsA dose (Fig. 11). In all allografts, an intact endothelial layer was visible. The media of the allografts of the placebo group and the 0.3 mg·kg^-1·day^-1 CsA dose group showed a loss of cellularity which is generally attributed to the migration of SMC from the media to the intima (Fig. 11). CsA doses of 1 mg·kg^-1·day^-1 or more prevented the loss of SMC nuclei and reduced the degree of necrosis in the media and adventitia. The autografts showed no signs of cellular rejection (Fig. 10).

DISCUSSION

Allogeneic immune mechanisms ⁽²⁾ and nonimmunological responses ^{(6, 7, 17)} drive the process that leads to chronic GVD. In the rat, aortic transplantation is often used to model this process ^{(13, 18)}, and in the mouse, carotid artery transplantation is used ⁽¹⁹⁾. Transplanted arteries develop concentric and generalized intimal thickening ^{(3, 18)}, in contrast to classical arteriosclerosis, where the intimal hyperplasia is mostly focal and asymmetric ^{(3, 4)}.

The mechanism leading to intimal thickening is characterized by an early acute cellular rejection process in the adventitia, followed by a less intense chronic inflammation ^{(1, 20)}. Concurrently, SMC in the media become necrotic and/or start to migrate to the intima, where they also replicate ^{(13, 21, 22)}. Thus, the media loses cells and a massive neointima is formed ⁽⁵⁾. Necrosis of the media, a characteristic finding during rejection of untreated vascular allografts, is due to a toxic effect of activated lymphoid cells on medial myocytes ⁽²⁰⁾. We found a relatively good preservation of media volume, but a decrease of cellularity, which has also been observed by other investigators ⁽²³⁾.

CsA dose dependently inhibited all these phenomena to the point of almost complete suppression at the highest dose, which, however, resulted in blood levels about 10 times higher than those used therapeutically in man (as far as comparisons are possible considering the pharmacokinetic differences: infusion with minipumps yields constant levels without peaks and troughs). Preservation of the medial area and nuclear density by CsA was also observed by other investigators ^{(8, 11, 13, 24)}.

Syngeneic aorta grafts showed no intimal or medial alterations, although low-level inflammation in the adventitia was observed ⁽¹⁸⁾. Similarly, our autografts showed a normal intima except for a minimal focal intimal thickening detected in two out of four autografted carotid arteries that we could examine in the highest-dose group. This is possibly attributable to a high-dose CsA effect, suggesting that the autotransplants that do not benefit from immunosuppression manifest ill effects of CsA. The blood levels achieved at this dose were about 10-fold higher than those used in man, and this could lead to injury of the vessels. CsA has previously been reported to be toxic for vascular endothelium ^(25-27) and to lead to intimal SMC vacuolation (necrosis) ⁽¹²⁾. We observed no frank toxic effect of CsA on endothelial cells. However, the minimal neointimal thickening of the autografts, which cannot be the result of an immunological process, appears to represent stimulation by the highest dose of CsA.

This comparison of the histological data of the simultaneously allo- and autotransplanted carotid arteries permits the conclusion that the massive neointima formation in the allografts was almost completely driven by immunological processes. Surgical trauma or reperfusion injury ⁽²⁸⁾ appear to play very minor roles under the conditions of the present experiment with a short ischemia time (45 min). Our observations suggest that the highest dose of CsA might enhance neointima formation. However, the plasma concentrations were much higher than those used therapeutically in man.

Lumen size measurements of a vessel by histological techniques are difficult, and artifacts are unavoidable during perfusion fixation, dehydration, embedding, and staining. With in vivo MRI, we were able to estimate the lumen areas of the carotid arteries repeatedly throughout the treatment period. Irrespective of treatment and despite fixation at physiological pressure, the in vivo lumen areas were consistently larger than those measured histologically. However, these measurements are also difficult and prone to artifacts induced, for example, by differing levels of anesthesia and alterations in blood pressure, among many other possible influences. At 2 weeks the absolute lumen size at the center of the allografts tended to be larger than those of the autografts. At 4 and 8 weeks the trend reversed. The luminal narrowing of the allografts could be functional (endothelial dysfunction with endothelin release, lack of nitric oxide) or structural (neointima).

The structural hypothesis is supported by the finding that 8 weeks after transplantation the lumen size of the allografted vessels, as estimated by histology, showed a significant inverse correlation to neointimal area in the CsA-treated animals. Therefore, it appears that the thickness of the neointima indeed had an influence on the lumen size as measured by histological techniques. In contrast, the lumen size in vivo does not correlate with the neointimal thickness. This is similar to observations in the balloon angioplasty model ⁽¹⁴⁾, where we found no correlation between the lumen diameter of the injured carotid arteries and neointimal thickening. These, as much as the present experiments, indicate that the information gained from histology differs from that gained by techniques used to estimate lumen size in vivo. Neointimal thickening may be less important than other factors, such as arterial wall shrinkage ^{(29, 30)}. Finally, compensatory enlargement may minimize the impact of subendothelial thickening, at least for a time ⁽³¹⁾. Thus, histological changes in the neointima may reveal progressive allograft rejection but bear no relation to lumen size, at least in the early stages. At least in the early phase of the process, neointimal thickness does not correlate with lumen size in vivo and is of little use in assessing a functional impairment of blood flow. Therefore, qualitative aspects related to the rejection process appear to be more useful than morphometric measurements in the histological assessment.

The allografts showed bulging at midgraft and a stenosis at the anastomosis sites in vivo 2 weeks after transplantation, independent of treatment, although with time this effect disappeared. It would be logical to expect that bulging of the allografts occurs dose dependently as a result of the medial loss of cellularity. This was not the case; the bulging was observed in all treatment groups. At 8 weeks the phenomenon had disappeared. In accordance with our observations, Isik et al. ⁽²³⁾ also described bulging in allografts but not in isografts. Aneurysm formation was also observed after prolonged cold ischemia time ⁽³²⁾, which damaged the medial smooth muscle layer. The dilatation progressed over a follow-up time of 60 days and appeared to correlate with loss of medial SMC.

In conclusion, the present study demonstrated that CsA suppressed, in a dose-dependent manner, the neointima formation, loss of medial SMC, and the cellular rejection of allografted carotid arteries. However, the neointimal thickening was only partially inhibited with the 1 mg·kg^-1·day^-1 CsA dose, which produced blood levels comparable to those used in man. Irrespective of treatment and despite the loss of SMC nuclei, the medial area did not change during the observation period. The lumen area of the allografts, determined histologically, was indirectly proportional to the intimal thickening, which suggests a structural component to luminal narrowing. The appearance and regression of bulging can be only partially explained at present, especially since it is not influenced by treatment and further investigation is required. Finally, the simultaneous autograft serves as a “built-in control” useful in detecting nonimmunological influences and possibly ill effects of drugs on transplanted vessels.

Acknowledgments. The authors thank Dr. B. Dorobek for the determination of CsA in very small rat blood samples and D. Baumann for skillful assistance with the MRI measurements. They are also grateful to R. Bergmann for statistical advice and support.

Figure 1: Carotid lumen area of the allografts, determined in vivo by MRI 2 weeks after transplantation. The mean proximal and distal measurements of the grafts at the points indicated for all groups are shown relative to their anatomic position. The error bars (=SEM, shown for the control group) were similar for all groups (n=6, 6, 7, 6, and 6 for the placebo and CsA 0.3, 1, 3, and 10 mg·kg^-1·day^-1 groups).

Figure 2: Carotid lumen area of the autografts determined in vivo by MRI 2 weeks after transplantation. For details, see Legend to Figure 1.

Figure 3: Effect of CsA treatment in vivo on the allo- and autografted carotid lumen areas as a function of time after transplantation; MRI measurements are expressed as the ratio of center lumen area (middle of transplant) to anastomotic lumen area (distal anastomosis site). Columns are means, and bars indicate SEM; n=6, 6, 7, 6, and 6 for the placebo and CsA 0.3, 1, 3, and 10 mg·kg^-1·day^-1 groups. Asterisks indicate significant differences between the allografts and autografts (Student's t test; P<0.05).

Figure 4: Lumen area of the allo- and autografted carotid arteries in control rats (placebo treatment) and animals treated with CsA as indicated, determined in vivo by MRI (top) and by histological examination (bottom) 8 weeks after transplantation. Columns are means, and bars indicate SEM; n=6, 6, 7, 6, and 6 for MRI measurements of the placebo and CsA 0.3, 1, 3, and 10 mg·kg^-1·day^-1 groups; n=5, 4, 5, 6, and 5 (allografts) and n=6, 6, 6, 6, and 4 (autografts) for the histological measurements of the same groups. The plus sign indicates a significant difference between the allografts and autografts (Student's t test; P<0.05).

Figure 5: Semilogarithmic plot of the carotid lumen area of the CsA-treated allografts, estimated postmortem by histology. The individual values of each dose group are shown. The linear regression curve and 0.95 confidence band curves (dashed lines) are shown. P=0.011, correlation coefficient r=0.554, n=20.

Figure 6: Neointimal thickening of the allo- and autografted carotid arteries in control and treated animals, determined from histological sections 8 weeks after transplantation. Columns are means, and bars indicate SEM; n=5, 4, 5, 6, and 5 (allografts) and n=6, 6, 6, 6, and 4 (autografts) for the histological measurements of the placebo and CsA 0.3, 1, 3, and 10 mg·kg^-1·day^-1 groups. The plus signs indicate significant differences between the allografts and autografts for each group (Student's t test; P<0.05), and asterisks mark differences of the allografts from the placebo group (Dunnett's test; P<0.05).

Figure 7: Correlation between the carotid artery lumen size and neointimal area of the CsA-treated allografts as determined from histological sections 8 weeks after transplantation. The individual values for each dose are shown. The linear regression curve and 0.95 confidence band curves (dashed lines) are shown. P=0.0004, correlation coefficient r=-0.72, n=20.

Figure 8: Carotid artery lumen areas, expressed as ratio of allografted to autografted carotid lumen area, determined 8 weeks after transplantation. The left and right bars of each pair relate to the lumen area estimated by histological morphometry and by MRI, respectively. Columns are means, and bars indicate SEM; n=6, 6, 7, 6, and 6 for MRI measurements of the placebo and CsA 0.3, 1, 3, and 10 mg·kg^-1·day^-1 groups; n=5, 4, 5, 6, and 5 (allografts) and n=6, 6, 6, 6, and 4 (autografts) for the histological measurements of the same groups.

Figure 9: Photomicrographs of representative histological cross-sections of allografted carotid arteries of the (A) placebo group, (B) CsA 1 mg·kg^-1·day^-1 group, and (C) CsA 10 mg·kg^-1·day^-1 group determined 8 weeks after transplantation. Giemsa stain, magnification ×400.

Figure 10: Photomicrographs of representative histological cross-sections of autografted carotid arteries of the (A) placebo group, (B) CsA 1 mg·kg^-1·day^-1 group, and (C) CsA 10 mg·kg^-1·day^-1 group determined 8 weeks after transplantation. Giemsa stain, magnification ×400.

Figure 11: Morphological analysis of the allografted rat carotid arteries, determined from histological sections 8 weeks after transplantation. Columns present the mean of the morphologic parameter grading per dose group, and bars indicate SEM. The grafts were scored for adventitial infiltration of mononuclear cells (Infil. ADV) and necrosis (Nec. ADV), the number of SMC nuclei in the media (SMC Med) and necrosis in the media (Nec. Med), and the intimal infiltration of mononuclear cells (Infil. INT). The specimens were scored on a 0 to 3 scale (see Materials and Methods); n=5, 4, 5, 6, and 5 (allografts) and n=6, 6, 6, 6, and 4 (autografts) for the histological measurements of the placebo and CsA 0.3, 1, 3, and 10 mg·kg^-1·day^-1 groups. The solid square (▪) indicates a single allograft that showed a focal infiltration of mononuclear cells into a small area of the intima. Asterisks indicate significant changes when comparing treated groups with their respective placebo group (Dunnett's test; P<0.05).