Effect of Age on the Pattern of Short-term Albumin Uptake by the Rabbit Aortic Wall Near Intercostal Branch Ostia
Abstract Lipid deposition occurs more frequently downstream than upstream of branches in immature human aorta but the opposite pattern is seen in mature vessels. These distributions may reflect variation in the uptake of plasma macromolecules by the aortic wall. We have recently shown that the quasi–steady state uptake of albumin is greater downstream than upstream of branches in immature rabbit aortas and that the opposite pattern occurs in mature animals. Additionally, there is a sharp drop in the mean uptake shortly after weaning. In the present study, the mechanisms underlying these phenomena were investigated by examining the short-term uptake of albumin and its distribution across the wall. Albumin was labeled with a fluorescent dye and introduced into the circulation of conscious New Zealand White rabbits. Thoracic aortas were fixed in situ 10 minutes later and were sectioned through the center of intercostal ostia. Fluorescence from sections was measured by using digital imaging fluorescence microscopy and was converted to tracer concentrations after appropriate autofluorescence levels had been subtracted. In animals aged 45 days, more tracer was detected in the wall downstream than upstream of branches; the difference between regions was >100% of the mean value. This percentage halved and the mean uptake decreased almost threefold by 75 days. In mature animals, the mean value remained at the 75-day level but the converse distribution was seen; 22% more tracer was detected upstream than downstream. These trends were insensitive to the depth of the intimal-medial layer examined. In each region, the maximum tracer concentration occurred close to the luminal surface but not always within the first 2.9-μm-thick layer of the wall. Maxima were similar in magnitude to those observed at quasi–steady state, but the fall with increasing distance into the wall was much sharper. In many cases concentrations remained constant over most of the media, and rises toward the adventitial boundary were rarely seen. Uptake after 10 minutes predominantly reflects the rate at which tracer enters the wall. The concentration profiles were consistent with most of the tracer having entered from the luminal surface and with the involvement of convective transport. The trends observed with age closely paralleled those occurring at quasi–steady state. Consequently, the latter are also likely to be determined by changes in the resistance of the wall to macromolecule influx.
- Received June 14, 1995.
- Accepted September 29, 1995.
Atherosclerotic lesions are nonuniformly distributed in adult human arteries. The outer walls of bifurcations and the upstream lip of side branches are among the most frequently affected sites, while flow dividers and regions downstream of them tend to be spared.1 2 3 4 5 The lipid and other components of these lesions appear to derive from plasma constituents,6 7 8 making it plausible that disease develops in areas of the wall where the uptake of circulating macromolecules is high.
Early studies investigated this hypothesis by examining the uptake of intravital dyes, but such methods may be unreliable.9 10 More recent studies have employed a range of novel techniques to quantify short-term and quasi–steady state patterns of macromolecule uptake, mainly in the rabbit aorta. Almost all have shown that uptake is enhanced in areas downstream of branch ostia.11 12 13 14 15 These sites are prone to diet-induced lipid accumulation in rabbits16 but are particularly resistant to disease in adult human arteries. Consequently, the results could be construed as evidence for or against the possibility that elevated uptake leads to the development of human lesions. Furthermore, an apparently contradictory trend has been identified in a study of the rabbit aorta by Berceli et al;17 elevated uptake was seen in the outer walls of the bifurcation, a site that is susceptible to atherosclerosis in human vessels.18
In an attempt to clarify these issues, we investigated the influence of age on transport. The potential importance of age was inferred from the observation by Sinzinger et al19 that the distribution of lipid deposition in immature human aortas is the opposite of the adult pattern. Sudanophilia occurs most frequently downstream of branch ostia in these arteries. We found that although quasi–steady state uptake of albumin is higher downstream than upstream of branches in immature rabbits, this difference decreases and then reverses with age.20 Thus, when age is taken into consideration, the pattern of transport in rabbit aortas does parallel the distribution of human lesions and is consistent with uptake being important in atherogenesis. The findings may also explain the apparently anomalous result of Berceli et al17 since this study, alone of those cited above, used mature rabbits.
The normal adult pattern of transport does not correlate well with the distribution of lesions in hyperlipidemic rabbits. Disease in this model seems to occur downstream of branches irrespective of age. However, we have subsequently shown that albumin transport reverts to its juvenile pattern, at least transiently, in older rabbits fed a cholesterol-enhanced diet.21
In the present study, mechanisms underlying the effects of age on quasi–steady state transport were investigated. Albumin was labeled with a fluorescent dye and introduced into the circulation of normal, conscious rabbits. After 10 minutes, animals were administered an overdose of anesthetic, and their aortas were fixed in situ. Tracer uptake was subsequently assessed by applying digital imaging fluorescence microscopy to sections through branches. Such short-term transport predominantly reflects rates of influx into the wall. Variations were found that appear to account for the patterns of uptake described above and also for a decrease seen in the mean uptake shortly after weaning.20 Profiles of tracer concentration across the wall and the patterns seen in animals in which fixation was delayed give some indications about the mechanisms involved. Part of this work is described in a preliminary report.22
Preparation of Tracer
Sulforhodamine B (Lissamine rhodamine, CI 45100, Sigma) was purified by extraction in 75% vol/vol ethanol and converted to its sulfonyl chloride by grinding with PCl5. The sulfonyl chloride was extracted in acetone and added to 20 times its own weight of bovine serum albumin (fraction V, Sigma) at 2% wt/vol in 0.33 mol/L carbonate buffer, pH 9, at 4°C.25 The resulting conjugate was purified of free dye and buffer salts by gel filtration (Sephadex G-25, Pharmacia), frozen dropwise in liquid nitrogen, freeze-dried, and stored at −20°C.
Just prior to use, the conjugate was reconstituted and then repurified of free dye by stirring with neutralized, activated charcoal (0.35 g/g protein) for 1 hour.25 Charcoal was removed by centrifugation (twice at 1000g for 30 minutes) and filtration (0.2-μm Minisart-plus, Sartorius). More than 99.8% of the dye, as assessed by absorbance at 575 nm, could be precipitated by trichloroacetic acid after this purification. The tracer is stable in vitro and in vivo and has similar physical, chemical, and biological properties to the unlabeled protein.11 26 27 28 29 Free dye does not bind to plasma proteins or tissue components.26
All animal procedures complied with Home Office and local regulations. Nineteen male New Zealand White rabbits aged 38 to 1463 days were used to assess tracer uptake or autofluorescence. All animals had been weaned and were maintained on a standard lab diet (200 g/d, Special Diet Services) with tap water ad libitum.
Tracer (700 mg/kg in 20 mL saline) was administered via the marginal ear vein to conscious, unrestrained animals over a period of approximately 1 minute. Unlabeled bovine serum albumin was used in control experiments. The vein was flushed with a small volume of saline before and after the introduction of tracer or albumin. Eight minutes after the end of this procedure, heparin (1000 IU, Sigma) was introduced in the same way. Two minutes later, the animals were killed with an overdose of pentobarbital (300 mg IV, Sagatal, Rhone Merieux). In rapid succession, the abdomen was opened along the midline, the viscera were deflected, and the diaphragm was partially removed. A blood sample was taken by cardiac puncture, and a retrograde cannula was tied into the aorta at the level of the diaphragm. The thoracic aorta was flushed for approximately 30 seconds with 50 mL isotonic saline at a pressure of 120 cm water. It was then fixed at the same pressure with formal sublimate (6% wt/vol HgCl2 plus 1% wt/vol formalin in water). This fixative rapidly immobilizes proteins but does not cause substantial tissue autofluorescence.30 The period from anesthesia until fixation averaged 3.6 minutes (SD=0.94, n=13) for the animals administered tracer. These figures exclude two animals for which the period was inadvertently extended to more than double the mean value. The latter experiments gave anomalous results and are discussed separately.
The proximal descending aorta was clamped after 2 minutes to reduce the volume of fixative required. After 20 minutes of fixation, the descending thoracic aorta was excised, and regions of the wall containing intercostal ostia were dehydrated in ascending concentrations of ethanol, transferred to propylene oxide, and embedded in epoxy resin31 (Polarbed 812, Fisons).
Plasma prepared from the terminal blood sample was diluted 1:100 with 12.5% wt/vol gelatin (type B, Sigma) in phosphate-buffered saline (0.15 mol/L, pH 7.4). Gels were set at 4°C, fixed with formal sublimate, and processed in the same way as arterial tissue.
Tracer Detection and Image Processing
Blocks of embedded tissue were trimmed under a dissecting microscope so that 2-μm-thick longitudinal sections could be cut through the center of intercostal ostia. Sections were cut by using an ultramicrotome and glass knives, and were mounted on coverslips. For every animal, three to eight branches were examined; seven or more sections were cut from each branch. Fluorescence from sections was stimulated and examined by using an epifluorescence microscope (Zeiss Axioplan) with oil immersion lenses, custom filters (577±7-nm excitation filter, 590-nm dichroic mirror, and 610±10-nm barrier filter; Omega Optical), reducing transfer optics (Microcam), and an intensified CCD camera (Darkstar, Photonic Science).
Output from the camera was digitized by using a Hawk V12 digitizer (Wild Vision) in an Archimedes 540 microcomputer (Acorn). The circuitry of the digitizer was modified to improve the accuracy of black-level clamping.24 Custom software was written for image acquisition, processing, and analysis in ARM assembler.
One image was obtained for the aortic wall upstream of the ostial lip and one for the equivalent downstream region in each section. The maximum length of each aortic segment was 370 μm. Eight frames were grabbed and averaged for each image to improve signal-to-noise ratios. Images were processed to ensure a uniform, linear, and constant relationship between tracer concentration and gray level. Offsets were subtracted from all frames and a flat-field correction was used to remove the effects of spatial biases in the system. A correction factor was calculated to allow for changes in the exciting illumination, intensifier voltage, video amplifier gain, and extent of photobleaching. By using these techniques, spatial biases in the corrected image are <5%, and measured intensity is proportional to tracer concentration over at least a thousandfold range.24
The upstream or downstream image was displayed on a monitor, and a line was drawn along the outer surface of the aortic endothelium by using a tracker ball. This line was smoothed by replacing the x and y coordinates of each pixel with the mean coordinates of itself and the six nearest neighbor pixels along the line. A series of lines parallel to this endothelial boundary was then generated through the intima and media. The spacing between lines was equivalent to 2.88 μm before magnification, and the number of lines was near to the maximum that could be fitted without impinging on the adventitia at any point. The edges of the grid were defined by lines perpendicular to each end of the endothelial boundary. An example of such a grid, superimposed on an image, is shown in Fig 1a⇓.
Corrected gray values of all the pixels within each depthwise slice of the intima and media were summed, and the pixels were counted. A mean concentration and a mean quantity of tracer for each slice were derived from these parameters. Concentrations were expressed as a percentage of the terminal plasma concentration of tracer as assessed by analysis of the images from calibration gels. Quantities of tracer were summed over various depths as described below. They were expressed as a normalized mass transfer coefficient32 that was defined as
where Qw is the quantity of tracer in the wall, Cp is the tracer concentration in plasma, Ae is the area of endothelium, and t is the duration of exposure to the tracer. The term “permeability” has been avoided since it is variously used in connection with diffusion, convection, or both phenomena. The expression reduced to
where ΣCGLw is the sum of corrected gray levels in the wall, Ap is the area of the section imaged by a single pixel, C̅G̅L̅g is the mean corrected gray level for gel, and L is the length of the endothelium.
Sections from control animals were analyzed to determine autofluorescence intensities. These values were subtracted from all experimental data prior to averaging or calculating ratios between regions.
Statistical methods were obtained from Armitage and Berry.33
Age, Weight, and Arterial Geometry
Transport was studied by using rabbits in five age groups: group 1, 38 to 51 days; group 2, 66 to 86 days; group 3, 272 to 389 days; group 4, 805 to 858 days; and group 5, 1295 to 1463 days. Four control rabbits had ages of 38, 85, 307, and 1085 days. The relation between age and weight for all the rabbits is shown in Fig 2⇓. The data are in reasonable agreement with previous results.20 In both samples, there was a nearly linear increase in weight up to an age of approximately 3 months, after which the rate of growth gradually declined to zero. The maximum weight of almost 5 kg was attained between 1 and 3 years of age. The data have been used to classify as immature or mature those rabbits used in earlier investigations for which weights but not ages have been recorded. Full sexual maturity occurs at around 5 months34 and hence at a weight of approximately 3.4 kg.
The wall had an irregular shape in upstream and downstream regions, making it difficult to define a representative intimal-medial thickness. However, an accurate indication of the fractional changes in thickness that occur with location and age can be obtained from the depth of the image-analysis grid, since the number of grid lines was adjusted for each branch to include most of the intima-media without impinging on the adventitia at any point. Grids penetrated to at least 90% of the minimum thickness of the intima-media in each region, and this proportion would not have varied systematically with location or age.
The grid depths (Table 1⇓) were on average 25% greater downstream than upstream. They increased by <50% between weaning and maturity; most of the change occurred before the age of 3 months. Systemic blood pressure in rabbits shows a similar pattern, increasing linearly from birth until 2.5 to 3 months and then remaining constant for several years.35
Naturally occurring autofluorescence from tissue components imposes a limit on the minimum concentration of tracer that can be reliably detected. Fig 1b⇑ shows a corrected image of autofluorescence from the region and age with the highest emission. The intensity is clearly negligible compared with fluorescence from the experimental sections shown in the same figure. Quantitative comparison of all control sections with corresponding experimental data showed this to be true for most locations and ages. The exceptions occurred in the mid media of animals older than 1 year, in which tracer concentrations sometimes fell substantially below 0.05% of the concentration in plasma. This resulted in autofluorescence and tracer fluorescence having intensities of comparable magnitude.
The variability of autofluorescence, as well as its mean intensity, determines the sensitivity of the technique. If autofluorescence were constant, it could be accurately subtracted from tracer fluorescence irrespective of their relative intensities. In the present study, variations in autofluorescence were found to be small. At each age and for each region of the wall, the SE of autofluorescence intensity never exceeded 7% of the mean value and was never equivalent to more than 0.0022% of the mean concentration of tracer in plasma. Consequently, the use of an autofluorescence subtraction technique was considered a reliable method for determining whether tracer was present across the entire wall of the artery and for measuring differences in tracer influx between upstream and downstream regions. However, as discussed below, the small fractional differences detected between mid medial areas of older rabbits were similar in size to variations in autofluorescence and cannot be regarded as reliable.
There was a significant increase in mean autofluorescence with age (ANOVA applied to regression yielded F=102.8, P<.005, n=59), and small, nonsignificant differences were also detected between upstream and downstream regions (for all ages: t>−1.90, P>.07, n=24) (Fig 3⇓). Hence, when subtracting autofluorescence, a mean value from the equivalent region of the control animal closest in age was used.
Differences Between Upstream and Downstream Regions
Fig 1c⇑ through 1f shows corrected images of the fluorescence from upstream and downstream regions of an immature rabbit and a mature rabbit administered tracer. Uptake is visibly greater downstream of the branch than upstream in the young rabbit, but the opposite pattern is apparent in the older animal.
For quantitative comparisons, differences in uptake between regions were scaled by the mean of both regions so that changes in the pattern of transport and changes in overall transport could be distinguished. The difference was calculated as
where U and D are the normalized mass transfer coefficients for the upstream and downstream regions, respectively. Differences were calculated for every section and were averaged to obtain a representative value for each branch. The mean and SEM of the branch averages were then calculated for each age group.
A problem in interpreting mass transfer coefficients is determining which parts of the intima-media receive tracer from the luminal surface and which from the adventitia. In the present study, differences between regions were calculated after integrating tracer quantities over a range of distances from the endothelium to allow this question to be addressed. Fig 4a⇓ through 4c shows the effect of age on the difference for depths of 10, 20, and 30 grid lines, corresponding to 29, 58, and 86 μm, respectively.
An alternative to examining the same depth for both regions and all ages is to scale the depth according to the thickness of the intima-media. This procedure has been used in several studies of nonbranch regions of wall.32 36 37 In the present study, in which thickness varied systematically with age and location, the procedure might have introduced biases by enhancing the relative uptake in downstream regions and in older animals. However, the method was used so that comparison could be made with previous studies; tracer quantities were integrated over approximately the full intimal-medial thickness of each region (Fig 4d⇑).
Considering first the depth of 29 μm (Fig 4a⇑), the normalized difference between mass transfer coefficients was significantly greater than zero in immature animals (groups 1 and 2: t=10.9, P<.001, n=34) but less than zero in mature animals (groups 3 through 5: t=−3.3, P<.005, n=43). There was a significant fall in the difference between groups 1 and 2 (t=−3.92, P<.001, n=34). There was no simple trend with age and no significant deviation from linearity between the mature groups (ANOVA applied to regression yielded F=0.19, P>.05, and F=0.80, P>.05, n=43). A large variation between branches was observed, particularly in the older animals. This variation was not found in a previous study of quasi–steady state transport.20
As the depth of the grid was increased from 10 to 30 lines, small changes occurred in the differences between regions but the main trends were not altered. The same general patterns were also apparent in the data for the full intimal-medial depth. Uptake in the downstream region was accentuated for groups 2 through 5, reflecting the greater depth and hence greater quantity of tracer that was sampled in this region. However, the difference between regions in mature animals still averaged −15% and remained significantly less than zero (t=−2.75, P<.005, n=43).
It is not clear whether the same distance from ostia should have been examined in all rabbits or whether this distance should have been scaled according to the length of the animal. To determine whether the latter procedure would have affected the trends described in Fig 4⇑, 10 branches from two animals in group 1 were reanalyzed over approximately half the distance employed in the main study. This scaling factor reflects the ratio of minimum and maximum animal lengths, estimated as the cube root of weight. For the full intimal-medial depth, this procedure altered the difference between regions from 119.8±10.43% to 125.5±10.25% (t=−0.038, P>.7, n=20). A similar lack of effect was found in the earlier study of quasi–steady state transport.20
Transmural Concentration Profiles
Profiles of tracer concentration across the intima-media are shown for upstream and downstream regions of the different age groups in Fig 5a⇓ through 5e. The profiles summarize data from a subset of specimens comprising three branches per animal and at least three sections from each branch. For the downstream regions of groups 1 and 2, the profiles show a nearly smooth decrease in concentration from the luminal surface to the outer media. For all other profiles the concentration decreased initially but then remained essentially constant throughout the central region of the wall. The discontinuities shown in the deepest layers of the wall are artifactual, reflecting changes in sample size; for the greatest distances into the media, data could be obtained from only a small number of sections. A slight rise toward the adventitial border was seen in a few profiles.
In groups 1 and 2, the maximum concentration and the depth required for the decline from this level to the minimum value were both substantially greater downstream of the branch. Comparable trends in the opposite direction were not readily discernible for groups 3 through 5, reflecting the smaller differences between regions in mature animals. Differences between regions were apparent at depths where the profiles had reached nearly steady levels but in some cases tracer concentrations were so low that these differences approached the limit of detection. Although they seem to be consistent, a small error in the assessment of autofluorescence intensities would have lead to a significant and systematic bias. Consequently, they must be interpreted with caution.
To ensure that such biases could not have introduced a significant error into the age-related patterns shown in Fig 4⇑, mean differences between regions in mature animals were reexamined over the first five grid lines (14 μm), since tracer concentrations are high enough for the reliable detection of small differences up to this depth. No evidence for the presence of biases was obtained. The mean difference of −22% was identical to the value for 10 lines and significantly below zero (t=3.26, P<.005, n=43). The variability remained high and therefore cannot be attributed to the difficulty of measuring low levels of tracer.
The distribution of tracer in the innermost layers of the wall sometimes showed unexpected patterns. In several regions, the concentration in the first slice was lower than would be predicted by extrapolation of the profile and in some cases was lower in absolute terms than the concentration in the next layer. This was particularly apparent in the downstream regions of younger animals and the upstream regions of older animals, where uptake was relatively high.
High-Resolution Profiles Across the Inner Wall
To obtain further information about the differences between regions in mature rabbits and the unexpected patterns seen in the innermost layers of the wall, the same images were reanalyzed by using a finer grid, in which lines were separated by 0.72 rather than 2.88 μm. This analysis was applied to the first 10 μm (14 lines) of the wall. The grid was superimposed over only a short segment of aorta in an attempt to reduce the blurring of profiles that results from averaging long regions of the wall. Data are shown in Fig 6a⇓ through 6e. Because only a short length of aorta was reanalyzed, there are minor discrepancies in the heights, but not the overall pattern, of the corresponding high- and low-resolution profiles.
Differences between upstream and downstream regions in mature animals were more clearly visible in the high- than the low-resolution profiles. Peak heights were greater and concentrations declined more slowly with depth in the upstream profiles. In addition, all the high-resolution profiles revealed an apparent rise in tracer concentration that occurred over at least the initial 1.7 μm of the wall. The rise is likely to be partly artifactual, reflecting the response of the measuring system to a sharp edge; the theoretical resolution of the system under the conditions used in the present study is approximately 1.5 μm. However, this limit is unlikely to account for the differences seen between profiles.
In groups 1 and 2, apparent maximum concentrations occurred deeper in the wall downstream than upstream of branches. This pattern was not seen in mature rabbits. In groups 3 and 5, the peaks occurred at the same depth in upstream and downstream regions, while in group 4 the peak was deeper in the upstream region. The depths at which the maxima occurred were not a simple correlate of the absolute fluorescence intensities, which would have suggested a measurement artifact. The trends were consistent with the patterns seen in the low-resolution profiles. In the downstream region of immature animals, and occasionally in the upstream region of older animals, the highest concentrations appeared not to occur in the innermost layer of the wall. Further investigations employing higher magnifications, higher-resolution equipment, and thinner sections are required to clarify which anatomic structures are involved and to allow comparison with previous studies that have examined tracer concentrations between the endothelium and internal elastic lamina.38
Comparison With Transmural Profiles at Quasi– Steady State
A sample of images obtained in a previous study of quasi–steady state transport was reanalyzed by using the grid technique so that the transmural profiles could be compared with those for short-term uptake. The methods used to obtain the images have been reported in detail.20 They were similar to the methods described above except that the tracer was allowed to circulate for 3 hours instead of 10 minutes; cannulation was performed under anesthesia rather than immediately after death; and, because reducing optics were not used, regions up to only 270 μm (not 370 μm) from the ostia were examined and the outermost layers of the media could not always be included.
Only single animals were examined at each age in the previous study,20 and no animal was as old as those in group 5. Quasi–steady state profiles were therefore derived for the four animals closest in age to the mean ages of groups 1 through 4. Animals were included only if terminal plasma concentrations had been obtained. At least three sections from each of four branches per rabbit were analyzed.
For the majority of regions and ages, peak tracer concentrations detected in the inner wall after 10 minutes approached or even exceeded those seen after 3 hours. This was invariably the case in mature age groups. The difference in the protocols used just prior to cannulation may have influenced this result but is unlikely to be the sole factor. A lower-resolution study of albumin transport in which all animals were cannulated rapidly after death also found that uptake in the first layer of the wall reached quasi–steady state levels within 10 minutes.32
The fall in tracer concentration from the inner to the outer wall was much greater at 10 minutes than 3 hours. The ratio of minimum and maximum concentrations, averaged across regions and ages, was 10% at 10 minutes but 35% at 3 hours. Although maximum tracer concentrations were comparable at the two times, the minima, expressed as a fraction of terminal plasma concentrations, were on average 3.4-fold lower at 10 minutes.
The quasi–steady state profiles for several ages showed a pronounced rise toward the medial-adventitial boundary that was absent or greatly attenuated at 10 minutes. A rise with increasing depth was also apparent for the innermost layers in several of the 3-hour profiles. Fig 7a⇓ through 7d shows, as an example of these trends, the low- and high-resolution profiles for group 1 compared with equivalent quasi–steady state data from an animal aged 37 days.
Mean Mass Transfer Rates Near Branches
The mean of the mass transfer coefficients for upstream and downstream regions also changed but at an earlier age than the reversal in the difference between regions. Averages were again calculated for 10, 20, and 30 grid lines into the wall and for approximately the full intimal-medial thickness. Values were obtained for every section and averaged to give a representative coefficient for each branch. The mean and SEM of the branch averages for each age group are shown in Fig 8a⇓ through 8d. The trends were very similar for all depth criteria. Taking 10 lines (Fig 8a⇓) as an example, mass transfer rates were significantly higher in group 1 than group 2 (t=−4.81, P<.001, n=34). Between groups 2 through 5 there was no simple trend with age (ANOVA applied to regression yielded F=1.33, P>.05, n=62), although there was a significant deviation from linearity (F=3.60, P<.05) as group 4 had a lower value than the other three groups. The value for group 1 was 2.5×10−8 cm/s, and equivalent values for groups 2 through 5 averaged 0.91×10−8 cm/s. As depth was increased to 20 lines, 30 lines, or the full intimal-medial thickness, these values increased by 31%, 56%, and 69% for group 1 and 40%, 77%, and 117% for groups 2 through 5.
To determine whether the increase in animal size affected the measurement of changes in the mean coefficient, 10 branches from two animals in group 1 were again reexamined over approximately half the distance along the endothelium that had been employed for mature animals. As with the difference between regions, this scaling procedure did not significantly alter the data. For the full intimal-medial thickness, the coefficient changed from 3.22±0.32 to 3.34±0.32×10−8 cm/s (t=0.28, P>.7, n=20).
Effects of Delayed Cannulation
Two animals were excluded from the analyses described above because, following the normal 10-minute exposure to tracer in the conscious state, there were inadvertent delays between administration of the overdose of anesthetic and fixation. These delays caused interesting deviations from expected values. In one animal from group 3, fixation commenced 9 minutes after anesthesia. The difference between regions was 61% and the mean mass transfer coefficient was 4.0×10−8 cm/s, both calculated for the full intimal-medial thickness. The second animal, from group 4, had a delay of 7.5 minutes and gave corresponding values of 20.5% and 1.7×10−8 cm/s. Thus, in both these mature animals the juvenile pattern of transport was seen, and overall uptake was approximately double the mean value for the age group.
These results make it plausible that relative levels of tracer in the downstream region and the mean value of the mass transfer coefficients might both be overestimated for the other rabbits used in this study, since fixation commenced at least 2 and occasionally 5 minutes after anesthesia. However, further analysis showed no correlation between the length of the delay and either the difference between regions or the mean coefficient (data not shown). It appears that significant effects occur only if the period between anesthesia and cannulation is prolonged beyond 5 minutes.
Transport was investigated by applying digital imaging fluorescence microscopy to sections of arteries that had been fixed after exposure to sulforhodamine B–labeled albumin in vivo. The properties of the labeled albumin, the accuracy and precision of the detection techniques, and the comparability of quasi–steady state data with results obtained by using other methods have been extensively discussed.11 20 23 24 Several of the observations made in the present study concerning short-term transport are also consistent with earlier findings. The mean mass transfer coefficients obtained for the intima-media varied between 1.00 and 4.24×10−8 cm/s. These figures are compatible with coefficients determined for a number of tracers,15 32 37 38 although a wide range of values has been reported. The sharp decrease seen in the coefficient shortly after weaning has not been quantified before. However, it may be related to the finding that Evans blue dye uptake can be detected in the aortas of rabbits aged 6 weeks but not 12 weeks.39 Short-term uptake of albumin by the rat aorta increases with age,40 but no comparable trend was detected in the present study.
The observation that mass transfer coefficients are greater downstream than upstream of branches in immature animals is consistent with reports that foci of high macromolecule uptake are more prevalent downstream than upstream of branches in young rabbits,14 15 although this pattern is not seen in rats.41 The same properties may also underlie an earlier finding that the influx of LDL is greater downstream of large aortic branches than in nonbranch regions.13 No comparable elevation was seen downstream of intercostal branches, but the technique employed may have had insufficient spatial resolution to detect variations around such small ostia. The present finding that the difference between regions decreases and then reverses with age and the measurement of transmural profiles around branches appear to be novel.
There is good agreement between those investigations in which animals of the same species and developmental stage have been studied, despite the diversity of techniques and tracers employed. This is evidence not only for the reliability of the methods used but also for the view that uptake near branches varies in a similar way for macromolecules that have a range of sizes and markedly different interactions with cells and extracellular matrix components. The phenomena observed for albumin may therefore have relevance to the transport of other molecules, including the much larger and more metabolically active lipoproteins that have been implicated in atherogenesis.
The main aim of this study was to investigate whether differences in the rate at which macromolecules enter the arterial wall can account for the age-related variations seen in quasi–steady state albumin transport near branches. Quantities of tracer in the wall may depend not only on rates of influx but also efflux, metabolism, and the space available for the tracer within tissue. The relative importance of these additional factors increases with the duration of exposure to tracer. Transport over a 10-minute period was assessed in the present study. This is the shortest duration for which uptake in the conscious state substantially outweighs that occurring between anesthesia and fixation.
Only an insignificant proportion of the albumin entering the wall during a 10-minute period will be metabolized.42 Efflux from the intima-media is also likely to be negligible. After 10 minutes, tracer concentrations in the mid media were much lower than those in the inner wall. Because this fall was several times sharper than at quasi–steady state, when influx and efflux are approximately equal, only a small proportion of tracer entering via the luminal surface could have been lost through the adventitia. If efflux was negligible, almost all the tracer that entered the intima-media will have remained there and will have been included in the measurements, irrespective of how its distribution was affected by variations in the available space. Furthermore, since tracer concentrations within the wall after 10 minutes are much lower than those seen at equilibrium,36 they can have little influence on the rate of influx itself. On balance, therefore, it appears reasonable to assume that the trends seen in 10-minute uptake predominantly reflect changes in the resistance of the wall to tracer influx.
Differences between upstream and downstream regions were compared for the 10-minute and 3-hour experiments as a function of age (see Table 2⇓). Each group of animals from the present study was matched with the animal closest in age from the study of quasi–steady state transport.20 There was an excellent quantitative agreement between the two sets of data. Furthermore, tracer concentrations in the wall fell 2.8-fold between the ages of 38 and 56 days in the quasi–steady state experiments, while mass transfer coefficients fell 2.8-fold between 45 and 75 days in the present study. Again, the agreement is excellent. The transport gradients present in short-term experiments are not the same as those occurring at quasi–steady state or for native macromolecules in vivo, since tracer concentrations are relatively low in the media. Consequently, transport processes could also differ. The good agreement seen between 10-minute and 3-hour patterns is consistent with the same processes being critical in both cases and with the quasi–steady state phenomena being determined by changes in the resistance to macromolecule influx.
The rate of entry of macromolecules into the wall is often assumed to depend primarily on endothelial permeability. However, the possible role of convection and the resistances of other barriers to transport also need to be considered. Some indications of the transport processes taking place may be derived from transmural concentration profiles, although no definite conclusions can be drawn since profiles reflect the space that is available to tracer across the wall as well as the concentration of tracer within that space. The wide variety of reported profiles may reflect the range of vessels and techniques that has been employed.32 36 43 Age and location may also be important determinants. In the present study, profiles showed a nearly continuous decrease in concentration from the inner wall to the outer media when areas downstream of branches in young animals were examined but were flat over large parts of the wall upstream of branches and in older animals. The space available for tracer would have to be highly nonuniform for the latter type of profile to be compatible with a predominantly diffusive transport. The increases in concentration occasionally seen in the first few layers of the wall are consistent with the occurrence of concentration polarization.
Entry of tracer from the adventitia is possible in theory, and some evidence exists for the importance of this route.32 43 The present results, however, make it unlikely that adventitial influx is significant near intercostal branches. No substantial rises in tracer concentration were observed in the outer media. Although the geometry of the image-analysis grid excludes some parts of the outer media, rises were observed when using the same technique to study tissue exposed to tracer for 3 hours. Furthermore, only small changes were seen in the difference between regions as increasing depths of wall were examined, an unlikely occurrence if significant entry had occurred through both boundaries.
Some further inferences concerning the factors that determine macromolecule influx can be drawn from the patterns of transport obtained in this study. First, differences in uptake between upstream and downstream regions were more variable after 10 minutes than after 3 hours, the coefficient of variance being particularly high in older animals. This indicates that the rate of influx varies over short distances or fluctuates rapidly, since the quasi–steady state uptake at any point reflects entry through channels distributed over a wider area and integrated over a longer period. Second, the observation of a juvenile distribution of tracer in those mature arteries fixed more than 5 minutes after death is consistent with blood pressure, flow, or some other systemic property being required to maintain the normal adult pattern and, again, with the existence of mechanisms capable of rapid change. A causal link between variations in hemodynamic shear stress and variations in endothelial transport properties near branches has been postulated in several theories of atherogenesis but not conclusively established. The effects of delayed cannulation are compatible with such an interaction being a characteristic of mature vessels and with the immature pattern of transport being independent of flow or related only to its long-term influence on more stable properties of the wall. If it is additionally correct to assume that the highest stresses occur downstream of branches,1 shear would appear to be reducing the uptake of macromolecules in mature vessels. The mechanisms involved may be inhibited by hyperlipidemia, since short-term administration of a cholesterol-enhanced diet can also induce the juvenile pattern in older animals.21
To investigate the role of hemodynamic stresses in more detail, an attempt was made to map the boundaries between endothelial cells near branches. Despite the use of standard silver-staining techniques, boundaries could not be visualized in areas upstream of ostia in young rabbits. Similar difficulties in observing cell boundaries near branches have been reported.44 45 It is possible that the low albumin influx and the poor staining in upstream regions of immature vessels are both related to unusual properties of the intercellular junctions.
The physiological and pathological significance of the variations in transport is uncertain. A previous suggestion11 that the elevated transport observed downstream of branches in young animals might serve a physiological role by helping to maintain the thicker wall of this region is contradicted by the present finding that transport variations reverse with age while the pattern of wall thickness does not. A direct connection with pathological processes in rabbits is also unlikely. Normocholesterolemic rabbits rarely develop spontaneous sudanophilia,16 whereas hyperlipidemic animals show different transport properties.21 However, the spatial correlation with the location of human lesions makes it plausible that similar transport variations occur in human arteries and influence the development of disease.
This study was funded by the British Heart Foundation.
Kjaernes M, Svindland A, Walloe L, Wille SO. Localization of early atherosclerotic lesions in an arterial bifurcation in humans. Acta Pathol Microbiol Scand Sect A. 1981;89:35-40.
Asakura T, Karino T. Flow patterns and spatial distribution of atherosclerotic lesions in human coronary arteries. Circ Res. 1990;66:1045-1066.
Cornhill JF, Herderick EE, Stary HC. Topography of human aortic sudanophilic lesions. In: Leipsch DW, ed. Blood Flow in Large Arteries: Application to Atherogenesis and Clinical Medicine. Basel, Switzerland: Karger; 1990:13-19. Atherosclerosis monograph vol 15.
Smith EB. Fibrinogen, fibrin and lipoproteins. In: Woodford FP, Davignon J, Sniderman A, eds. Atherosclerosis X. Amsterdam, Netherlands: Excerpta Medica, Elsevier; 1995:334-337.
Schwenke DC, Carew TE. Quantification in vivo of increased LDL content and rate of LDL degradation in normal rabbit aorta occurring at sites susceptible to early atherosclerotic lesions. Circ Res. 1988;62:699-710.
Schwenke DC, Carew TE. Initiation of atherosclerotic lesions in cholesterol-fed rabbits, II: selective retention of LDL versus selective increases in LDL permeability in susceptible sites of arteries. Arteriosclerosis.. 1989;9:908-918.
Herrmann RA, Malinauskas RA, Truskey GA. Characterization of sites with elevated LDL permeability at intercostal, celiac, and iliac branches of the normal rabbit aorta. Arterioscler Thromb. 1994;14:313-323.
Berceli SA, Warty VS, Sheppeck RA, Mandarino WA, Tanksale SK, Borovetz HS. Hemodynamics and low-density lipoprotein metabolism: rates of low-density lipoprotein incorporation and degradation along medial and lateral walls of the rabbit aorto-iliac bifurcation. Arteriosclerosis. 1990;10:688-694.
Sebkhi A, Weinberg PD. Effect of dietary cholesterol on the uptake of albumin by rabbit aortic wall in vivo. J Physiol. 1994;479P:77P. Communication.
Fothergill JE. Fluorochromes and their conjugation with proteins. In: Nairn RC, ed. Fluorescent Protein Tracing. 2nd ed. Edinburgh, Scotland: E & S Livingstone; 1964:4-33.
Mancini RE, Vilar O, Dellacha JM, Davidson OW, Gomez CJ, Alvarez B. Extravascular distribution of fluorescent albumin, globulin and fibrinogen in connective tissue structures. J Histochem Cytochem. 1962;10:194-203.
Fothergill JE. Properties of conjugated serum proteins. In: Nairn RC, ed. Fluorescent Protein Tracing. 2nd ed. Edinburgh, UK: E & S Livingstone; 1964:34-59.
Nairn RC. Direct protein tracing in experimental physiology and pathology. In: Nairn RC, ed. Fluorescent Protein Tracing. 2nd ed. Edinburgh: E. & S. Livingstone; 1964:88-102.
Glauert AM. Epoxy resins: an update on their selection and use. Microsc Anal. September 1991:15-20.
Bratzler RL, Chisolm GM, Colton CK, Smith KA, Zilversmit DB, Lees RS. The distribution of labelled albumin across the rabbit thoracic aorta in vivo. Circ Res. 1977;40:182-190.
Armitage P, Berry G. Statistical Methods in Medical Research. 2nd ed. Oxford, England: Blackwell Scientific Publications; 1987.
Berger M, Jean-Faucher Ch, de Turkheim M, Veyssiere G, Blanc MR, Poirier JC, Jean Cl. Testosterone, luteinizing hormone (LH) and follicle stimulating hormone (FSH) in plasma of rabbit from birth to adulthood: correlation with sexual and behavioural development. Acta Endocrinol. 1982;99:459-465.
McCloskey DI, Cleary EG. Chemical composition of the rabbit aorta during development. Circ Res. 1974;34:828-835.
Penn MS, Saidel GM, Chisolm GM. Relative significance of endothelium and internal elastic lamina in regulating the entry of macromolecules into arteries in vivo. Circ Res. 1994;74:74-82.
Belmin J, Corman B, Merval R, Tedgui A. Effect of aging on albumin endothelial permeability and distribution volume in the rat thoracic aorta. FASEB J. 1991;5:A1474. Abstract.
Chuang P-T, Cheng H-J, Lin S-J, Jan K-M, Lee MML, Chien S. Macromolecule transport across arterial and venous endothelium in rats: studies with Evans blue–albumin and horseradish peroxidase. Arteriosclerosis.. 1990;10:188-197.
Yedgar S, Carew TE, Pittman RC, Beltz WF, Steinberg D. Tissue sites of catabolism of albumin in rabbits. Am J Physiol. 1983;224:E101-E107.
Tompkins RG. Quantitative analysis of blood vessel permeability of squirrel monkeys. Am J Physiol. 1991;260:H1194-H1204.