INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ANITSCHKOW AND COLLEAGUES (1) showed early this century that aortic lesions in the cholesterol-fed rabbit occur particularly frequently downstream of branch sites. They suggested on the basis of experiments with intravital dyes that this resulted from an enhanced uptake of plasma constituents in such regions and further suggested that the enhanced uptake was caused by mechanical stresses. The patterns of lesions (15-17, 29, 37) and transport (5, 27, 41, 49) have been confirmed many times and, following the suggestion of Fry (22), are widely thought to result from effects of elevated hemodynamic shear stresses. However, there is a conspicuous lack of direct evidence that the variations in transport are flow dependent. The issue is further complicated by the discrepancy with the lesion pattern subsequently observed in adult human arteries. In these vessels, lesions occur on the upstream lip of branches and the lateral walls of bifurcations (11, 14, 24, 46), which as several model and numerical studies indicate are regions of low shear (2, 11, 21, 36, 53). This pattern is difficult to reconcile with the roles of transport and flow postulated for rabbit lesions.

It has recently been proposed that the occurrence of two different lesion distributions is related to age, not species, and that this can resolve the uncertainty concerning the role of transport (6, 7, 39, 40). Human fetuses, neonates, and infants have been known for some time to show the Anitschkow pattern of lipid deposition (43). These lesions must regress with increasing age, and new ones must then develop with the upstream distribution. It is now known that the spontaneous lesions occasionally seen in rabbits also switch from a downstream to a more upstream distribution with age (6). Furthermore, although lesions occur downstream of branches in young cholesterol-fed rabbits, the more upstream pattern can be induced in adult rabbits (7). Thus rabbit and human aortas seem to develop parallel but age-related distributions of lipid deposition. Data obtained in the rabbit aorta show that both distributions can be explained by transport patterns. The extensively investigated downstream pattern of transport is in fact a property only of young rabbits. In adult animals, short-term (40) and quasi steady (39) transport are greater upstream than downstream of branches.

It is possible that effects of age are also the key to understanding the relation between flow and transport. In a study of short-term transport near branches by Sebkhi and Weinberg (40), mature rabbit aortas inadvertently exposed to tracer after death, and hence in the absence of blood pressure and flow, gave the young pattern of uptake, leading to the suggestion that it is the adult, not the juvenile, pattern of transport that is flow dependent. The study described here (44) was designed to test this hypothesis. Patterns of uptake by the aortic wall near intercostal branch ostia were determined in young and old rabbits after flow around some ostia had been modified by occluding the side branches.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals. All animal procedures complied with Home Office and local regulations. Five young (68-96 days, 1.84-3.4 kg) and seven old (613-738 days, 4.3-5.6 kg) male New Zealand White rabbits (Murex Biotech) were used to assess the effects of flow on aortic tracer uptake. They were maintained on a standard diet (9603 TRB, Harlan Teklad).

Surgical procedures. Each animal was anesthetized with fentanyl fluanisone (Hypnorm, Janssen) and midazolam (Hynovel, Roche). Initial doses were 0.3 ml/kg im and 2 mg/kg iv, respectively; boosters of one-third these amounts were administered as required. Lidocaine (Xylocaine, 1-2 ml, Astra) was applied subcutaneously around the throat and ventral midline. The trachea was intubated and the animal ventilated with air (50 breaths/min, Harvard Small Animal Ventilator).

Tracer administration and fixation. Permeability was assessed using bovine serum albumin (fatty acid free, First Link) labeled with the fluorescent dye sulforhodamine B (Lissamine rhodamine B, CI 45100, Sigma). This tracer was prepared and purified of free dye by standard techniques, as previously described (20, 51). It was administered via the marginal ear vein (700 mg/kg) ~20 min after the intercostal arteries were tied off; preliminary experiments showed such a delay was necessary to observe the effects described below.

Measurement of tracer uptake. Tracer uptake was assessed by measuring fluorescence from longitudinal sections of aortic wall cut through the center of intercostal branch ostia, using the histological procedures and carefully calibrated techniques of digital imaging fluorescence microscopy previously described (45, 51). Average values were calculated for regions of the aortic wall up to 345 µm (~1 side branch diameter) from the ostium, in the intima and media combined. Equivalent autofluorescence intensities, obtained in 18 branches from two additional rabbits not administered tracer, were subtracted from these values. Measurements of tracer and autofluorescence were normally made on six sections from each branch.

Statistics. Data are expressed as means ± SE, and n indicates the number of rabbits unless otherwise stated. Statistical comparisons were made between pairs of means by unpaired Student's t-test and between >2 means by ANOVA.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mean arterial blood pressure, measured while tracer was circulating, averaged 71 ± 5 mmHg (range 49-81 mmHg), which is ~80% of the normal in vivo value. Blood pressure was depressed in both young and old rabbits. Heart rate was approximately equal to the normal value in both young and old rabbits, averaging 242 ± 7 beats/min (range 221-270 beats/min) while tracer circulated.

Patterns of transport are shown in Fig. 1. To express the pattern quantitatively, uptake in the upstream region was subtracted from that in the downstream region, and the difference was then expressed as a percentage of the mean in both regions. Numbers greater than zero therefore signify greater uptake in the downstream region, whereas numbers less than zero signify greater uptake in the upstream region. In the immature animals (Fig. 1A), there was no significant difference between the patterns of transport around control, sham-operated, or occluded branches (F = 0.008, P > 0.05). On average, a downstream pattern was observed in all three treatment groups, but the differences between regions were small and not significant (t = 0.04, P > 0.1; t = 0.02, P > 0.1; t = 0.04, P > 0.1 for control, sham-operated, and occluded branches, respectively).

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Patterns of tracer uptake by the aortic wall upstream and downstream of the origins of intercostal arteries that had been occluded, sham operated, or left alone. Differences in uptake between upstream and downstream regions are shown as a percentage of the mean uptake; values >0 signify greater uptake downstream of the branch, values <0 signify greater uptake upstream. A: in immature rabbits (n = 5), tracer uptake was not significantly different between regions upstream and downstream of all three types of branch. B: in mature rabbits (n = 6), tracer uptake was greater upstream of control branches and this pattern was unaffected by sham operation; however, data for occluded branches were significantly different. Bars show means ± SE.

This pattern was markedly different from those previously observed in immature conscious rabbits (39, 40) or in immature perfused aortas (18), where uptake has been substantially greater in the downstream region. To ensure that this discrepancy was not caused by a change in rabbit arterial properties over time or by some problem with the measuring equipment, but was instead due to the surgical protocol, the pattern of transport was determined in one conscious, untreated rabbit of the same age. This animal gave a value of 86 ± 20% (n = 8 branches), exactly as expected from the earlier studies.

In mature animals (Fig. 1B), control branches gave the upstream pattern of transport seen in all previous studies (18, 39, 40). The finding was consistent, each rabbit showing this pattern on average, and significant (t = 3.84, P < 0.02). There was no significant difference between the patterns of transport in control and sham-operated branches (t = 0.37, P > 0.7). However, values for the occluded branches were significantly different from both the control (t = 3.81, P < 0.01) and sham-operated (t = 2.38, P < 0.05) branches; on average, these branches showed the reverse pattern of transport, uptake being greatest in the downstream region.

The data from one mature rabbit were omitted from these analyses, because the results for its sham-operated branches (130 ± 7.2%, n = 3 branches) were much higher than those obtained in the other six animals; the occluded branches also gave an unusually high value. We speculate that the vessels in this animal were hyperreactive and went into spasm during the sham operation (or occlusion). However, even if data from this animal are included, the difference in values between occluded and control branches remains significant (t = 3.64, P < 0.01), and the lack of difference between control and sham-operated branches remains insignificant (t = 1.13, P > 0.2).

Figure 2 shows typical images of the fluorescence emitted by sections through occluded and control branches from a mature rabbit. Tracer concentrations are indicated by the brightness at each location. In the control branch, concentrations in the aortic intima media are visibly greater upstream of the branch ostium than downstream. In the occluded branch, the opposite trend is apparent. Tracer in the adventitia, where concentrations are higher than in the intima media, was not included in the measurements.

View larger version (116K): [in this window] [in a new window]   Fig. 2.   Digital images of the fluorescence emitted by sections through a control branch (A, upstream of ostium; B, downstream) and an occluded branch (C, upstream; D, downstream) from a mature animal. Aortic endothelium is at the top and the adventitia (with high levels of fluorescence) is at the bottom of all images. The lumen of the intercostal artery lies between each pair of images. Fluorescence intensity, indicating tracer concentration, is visibly greater in the intima-media upstream of the control branch but downstream of the tied branch. Bar = 100 µm.

The mean of uptake in the upstream and downstream regions combined was significantly greater in control branches from immature animals than in those from mature animals (Fig. 3A) (t = 4.10, P < 0.01). Again, this contrasts with previous studies in conscious animals, where mean uptake in immature rabbits (other than weanlings) was similar to that in mature rabbits (39, 40). There was no significant difference in mean uptake between treatment groups for either the immature (Fig. 3B) or mature (Fig. 3C) rabbits (F = 0.64, P > 0.2 and F = 0.77, P > 0.2, respectively).

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Mean tracer concentration in regions of aortic wall upstream and downstream of intercostal branches combined, expressed as a percentage of the tracer concentration in plasma. A: tracer concentration was significantly higher near control branches from immature rabbits (n = 5) than near those from mature rabbits (n = 6). In immature (B) and mature (C) rabbits, there was no significant effect of occlusion or sham operation on mean uptake. Bars show means + SE.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study is part of continuing attempts to elucidate the mechanisms underlying the two transport patterns seen near rabbit aortic branches; these patterns may be relevant to the distribution of human as well as rabbit lesions. Short-term uptake studies (40) suggest that both transport patterns are determined by variations in the permeability of the wall to the diffusive or convective entry of macromolecules. Because both patterns are seen in vessels perfused with a constant flow of physiological buffer (18), neither can depend on interactions between blood cells and the wall, on the precise composition of plasma, or on the pulsatility of flow, at least in the short term. Effects of adding nitric oxide synthase inhibitors to perfused vessels indicate that the upstream transport pattern, seen in mature rabbits, is nitric oxide dependent but the downstream one, seen in immature rabbits, is not (18). Here we tested the proposition that the upstream but not the downstream pattern is flow dependent.

Aortic hemodynamics were modified by occluding intercostal arteries. This procedure does not completely stop flow into the side branch because blood is still able to pass out of it into the vasa vasorum and any other small vessels originating proximal to the occlusion site. We did not attempt to prevent such flow because blocking the adventitial blood supply can injure the aortic wall (48) and increase uptake (13). Preliminary experiments showed that transport was not increased if the intercostal artery was occluded at least a few millimeters from the aorta. Flow into the branch will still have been very substantially reduced by such occlusion, and any changes in transport can be attributed to this change in flow provided that there are no effects of sham operation.

Control branches in the mature rabbits showed the expected upstream pattern of transport. Sham-operated branches gave the same result, showing that the pattern was not modified by surgical manipulation alone. As predicted, the pattern was modified at occluded branches; in fact, the downstream pattern was seen. The observed dependence of the mature transport pattern on flow fits well with its dependence on nitric oxide because nitric oxide synthesis is influenced by blood flow, increases in steady flow rate or the introduction of pulsatile flow both increasing its release (35, 38). In the short term, such increases are mediated by an enhanced activity of nitric oxide synthase (23), whereas in the longer term increased expression of this enzyme may also be important (47). The dependence of the mature pattern on flow probably also accounts for the inverse spatial correlation between shear rate and transport seen at the rabbit aortic bifurcation by Berceli et al. (9); we have speculated elsewhere (40) that the distribution of tracer they saw is equivalent to the mature pattern.

The average uptake in upstream and downstream regions combined did not differ between control, sham-operated, and occluded branches in mature rabbits. Hence, the change in pattern observed around occluded branches must have resulted from a decrease in uptake upstream of the branch and an increase downstream. Such a change is hard to explain but is consistent with previous data: reversals in pattern without changes in mean level also occur with age (39, 40) and, in mature rabbits, with cholesterol feeding (Sebkhi and Weinberg, unpublished observations) and administration of the nitric oxide synthase inhibitor N-monomethyl-L-arginine (L-NMMA) (18). We speculate that flow, age, and hypercholesterolemia all have the same striking effect as L-NMMA because they all act via influences on nitric oxide synthesis or activity. In studies by other groups (3, 32, 33, 52), nitric oxide has increased or decreased transport across blood vessel walls. Although the reasons for this discrepancy are a matter of controversy (26, 31), the occurrence of both trends is consistent with the concept that in the mature aorta nitric oxide increases transport upstream of branches but decreases it downstream.

The fact that occlusion had the same effect as L-NMMA further suggests that it reduced shear, and hence nitric oxide synthesis, in both upstream and downstream regions. That would also explain why it had the same effect as the inadvertent exposure to zero aortic flow (and consequent absence of shear in both regions) in the experiments of Sebkhi and Weinberg (40). This inference is supported by two recent numerical studies of flow at the rabbit aortocoeliac branch, which show that shear stresses are elevated both upstream and downstream of the branch (10, 12). As flow into the side branch is reduced, so are the elevations in shear stress (12).

The downstream pattern of transport normally seen in immature rabbits was abolished by some aspect of the protocol used in the present study. Thus we could not determine the effect of flow on this pattern. However, recent experiments by Lever and Murphy (34) have answered this point: a downstream pattern of uptake was observed around the aortorenal branch of young rabbits, and the pattern was not changed 1.5 h after the renal artery was occluded. The only studies appearing to show a causal relation between flow and the downstream pattern of transport of which we are aware are those in which aortas lost the downstream pattern of Evans blue dye (EBD) uptake when exposed to the dye in vitro (8). However, these studies predated the discoveries that depressurization of arteries causes endothelial damage (4) and that EBD inhibits the release or action of nitric oxide (19). Furthermore, there was an absence of pressure as well as flow in vitro.

It is not currently possible to say what does cause the downstream pattern in immature rabbits, but its disappearance during our experiments suggests that it cannot depend on structural or other long-term influences alone, contrary to previous speculation (40). Differences in conditions between the present experiment and those in which the downstream pattern was detected seem quite subtle. In particular, the downstream pattern was seen by Lever and Murphy (34), who employed the same tracer and similar techniques of in situ fixation, tracer detection, and side branch occlusion, although they studied a branch of the abdominal rather than the thoracic aorta. The small difference in protocols between the two studies reduces the number of factors that could be of importance. An effect of blood pressure is one possibility because pressures are lower when the thorax and abdomen are opened, as in the present study, than when the abdomen alone is opened, as in the study by Lever and Murphy. The reduction could affect the release of autocoids and the compaction or strain of the wall. It would also diminish convective transport. Such transport is likely to dominate at foci of enhanced endothelial permeability, and these occur particularly frequently downstream of intercostal branch ostia in young rabbits (5, 27). Another possibility is an influence of anesthetics on the immature but not the mature pattern of transport; although the two studies used the same anesthetics, the present one employed higher doses. These anesthetics are known to have diverse direct effects on arteries. For example, they are either vasoactive or can alter the response of endothelial and smooth muscle cells to vasoactive agents (e.g., 25, 28, 30, 42). Effects of anesthetics on transport in young but not mature rabbits might also explain why mean uptake was greater in the former in the present study but not in studies using conscious animals.

In conclusion, we have demonstrated for the first time a flow dependence of transport around arterial branches. It was the upstream pattern, seen in mature animals, that was found to be flow dependent. The results provide a novel explanation for the spatial correlation between the pattern of flow and the distribution of disease in adult human arteries. The downstream pattern of transport, seen in immature arteries, appears not to be flow dependent (34); preliminary evidence was obtained that it is abolished by low blood pressure or by pharmacological effects of anesthetics. Systematic investigation of the conditions where the downstream pattern is and is not detectable will allow the mechanisms underlying it to be more precisely established.