| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Biomedical Engineering Program, Arizona Health Sciences Center, and 2Department of Physiology, College of Medicine, University of Arizona, Tucson, Arizona
Submitted 30 July 2004 ; accepted in final form 8 April 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODS AND MATERIALSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
interstitial hydration; deendothelialized vessels; convective transport
Studies have shown that the endothelium plays a significant role in controlling fluid filtration through the artery wall under steady-state pressure. These experiments (15, 19) demonstrated that the endothelium accounts for about half of the hydraulic resistance of the whole vessel wall. Therefore, one goal of the present study was to evaluate the role of the endothelium in the initial increase and subsequent decrease in filtration following the onset of pulsatility. It is important to examine arteries when they are exposed to pulsatile pressure with variations in frequency, because in vivo, pulsatile pumping of the heart causes cyclic changes in arterial transmural pressure, creating rapid shifts in arterial wall stress with possible changes in wall transport properties. In addition, the pulsatile frequency, although fairly constant when the body is at rest, increases abruptly with sympathetic stimulation. Therefore, a second goal of the present study was to determine whether the transient increase in filtration observed in the previous study (1) still occurred after an increase in frequency of an already established pulsatile pressure regime.
|
METHODS AND MATERIALS |
|---|
|
TOPABSTRACTMETHODS AND MATERIALSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Protocol for measuring fluid flux in deendothelialized vessels. To measure fluid filtration in the deendothelialized arteries under pulsatile pressure, carotid arteries were cannulated and excised without depressurization or change in length. These carotid arteries were the pairs of those used in the previous study (1), and so the intact arteries from the previous study served as a control group in the current experiment. After excision, the arteries used in the present study were placed in the experimental apparatus (Fig. 1), and each was deendothelialized by passing an air bubble along its length.
|
Protocol for measuring fluid flux in vessels receiving change in pulsatile frequency. In a second set of anesthetized animals, each pair of carotid arteries was surgically prepared (cannulated and then excised) such that the endothelium remained intact. Lp at a steady pressure of 60 mmHg was measured in each vessel. An initial pulsatile pressure regime (10–14 mmHg pulsatile pressure) was then superimposed on the baseline pressure (60 mmHg), after which the frequency of the pressure pulses was doubled (from 1 to 2 Hz) for a given set of pulses and the resultant fluid filtration was measured. A pressure transducer was used to measure and record the pressure changes over time, and OCT was used to measure arterial distension before, during, and after the frequency variation.
The observed experimental fluid filtration into the artery wall (corrected for residual distension) at the 2-Hz pulsatile frequency, subsequent to a "baseline" of 1-Hz frequency, was compared with that predicted for steady-state pressure and with the values previously measured in arteries that experienced steady pressure followed by the onset of 1-Hz frequency pressure pulses (1). The predicted steady-state values for filtration were calculated from the measured steady pressure Lp and the time-averaged pressure of the pulse train.
Surgery. During a previous study (1), both carotid arteries were surgically removed from each rabbit, and one vessel was used to determine the effects of pulsatile pressure on fluid filtration. The second carotid artery from each pair was used in the present study to determine the role of the endothelium in response to pulsatile pressure. Both vessels were cannulated, pressurized to 60 mmHg with PBS (4.0% bovine serum albumin, pH 7.4), excised, and then stored briefly in PBS before their use (1). Before excision, any branches from the main artery had been ligated. A stainless steel holder clipped to each cannula maintained the excised arteries at their original physiological lengths. The cannulated segment was placed into the apparatus shown in Fig. 1, and the segment was perfused with PBS containing 0.03% Trypan blue. Inclusion of the Trypan blue in the perfusate allowed for detection of any unligated branches when the vessels were pressurized and the outlet closed. The arterial segment was then deendothelialized by passing a small air bubble through its length. Previous studies (3) that used electron microscopy have shown that this method removes the endothelium without damaging the underlying internal elastic lamella. The segment was then preconditioned with repeated pressurization and depressurization between 10 and 100 mmHg, following a previously established procedure (4).
Steady-state Lp measurement.
An air bubble introduced into a length of transparent plastic tubing attached to the arterial inlet cannula was used to measure the artery Lp under steady pressure, following a previously established procedure (4). With the use of this technique, a step change in pressure was imposed within the arterial lumen, with the arterial outflow occluded. The air bubble moved because of a combination of pressure-induced arterial distension and transmural filtration. After a few minutes the bubble moved with a constant velocity that corresponded to the transmural fluid filtration rate. This value was then used to obtain Lp according to the following equation (4, 14):
 | (1) |
is the time-averaged transmural pressure, r is radius of plastic tube, and 
is bubble velocity. This calculation assumes that there is no osmotic pressure difference across the vessel wall. Earlier calculations (18) demonstrated that the colloid osmotic pressure difference in intact vessels was <10 mmHg for a vessel immersed in a 2% albumin solution with a 4% albumin solution perfusing the vessel. With a 4% albumin solution used for both immersion and as a perfusate, as used in the present study, the colloid osmotic pressure difference would be even smaller, probably an order of magnitude less than the hydrostatic pressure difference. It might be argued that pressurization of the artery for the 10- to 20-min time period required for measurement of Lp might hydrate the arterial wall. However, when Tedgui and Lever (18) tested for changes in extracellular space and water content in intact and deendothelialized vessels pressurized at 70 mmHg for 90 min, no significant increases were observed. Therefore, it would not be expected that the steady-state calculation would be affected by fluid accumulation.
Creation of pressure regimes. Pulsatile pressure was created in the arterial segments with a Harvard apparatus pump (model 1421, Pulsatile Blood Pump). The same apparatus described in the previous study (1) was used to create a high-resistance outlet with oscillatory pressure, but without net fluid flow, in series with the cannulated end of the artery (see Fig. 1). A Viggo-Spectramed pressure transducer was attached to the cannula outlet at the other end of the artery. As in previous experiments, baseline pressure was established in a pressure reservoir connected to the arterial inlet by using a sphygmomanometer bulb with attached mercury manometer, and the pressure tranducer was calibrated at the start of each experiment. The Harvard Apparatus pump was activated, and six sets of 5-pulse trains, followed by six sets of 20-pulse trains, at 60- and 80-mmHg baseline pressures at a frequency of 1 pulse/s were applied to the carotid arteries. In between the pulse trains there were 10- to 20-s periods of steady pressure to allow the apparatus to be reset. Data on the fluid flux into the artery during these interim periods were not obtained.
Synchronized data outputs were recorded from the transducer and imaging system throughout each pulse train. As was seen previously (1), the bubble oscillated back and forth with each pressure pulse and was often shifted toward the vessel after passage of the pulse train. The motion of the bubble during each pulse train was videotaped to determine the initial and final positions of the bubble and the total time of oscillatory motion. The net movement of the bubble toward the artery arose from the sum of any residual arterial distension and the fluid filtered from the artery lumen. From these measurements, and with knowledge of the geometry of the tube containing the bubble, the total volume of fluid filtering through the artery wall for a given pulse train could be calculated. This value had to be corrected for any residual distension that may have occurred, as described below. The corresponding fluid loss predicted to occur had the pressure been steady, rather than pulsatile, was found using the following equation:
 | (2) |
Thus the filtration measured under pulsatile pressure could be compared with that predicted for the same time-averaged steady pressure.
Measurement of residual distension.
To determine the fluid loss from the arterial lumen, it was necessary to measure the residual distention by using OCT as performed in a prior study (1). Details concerning the OCT system have been described previously (5, 13). The current system consists of a short-coherence length light source (superluminescent diode, 1,290-nm center wavelength and 49-nm bandwidth) allowing for a 16-µm coherence length (and thus axial resolution) in air. When a tissue index of refraction of 1.4 (6) is assumed, the tissue axial resolution is 
11 µm. The sample arm light is focused to a 15-µm spot on the tissue, and an image through the artery wall is obtained by collecting and analyzing the scattered light from the artery. The diameter measurements made using OCT have, at most, one-ninth of the error of previous diameter measurements made using a vernier caliper accurate to 0.1 mm (4). Furthermore, the OCT system measures the actual inner artery diameter rather than the external diameter (3, 4, 18). In these experiments the system was used to create two-dimensional OCT images [m-scans; i.e., functions of depth and time with time resolution of 100 ms (10 a-scans/s)] to determine variations in internal diameter caused by pulsatile pressure.
Volume displaced by residual distension (Vres) is given by the following equation:
 | (3) |
Changing frequency of the pulsatile pressure. To establish an initial pulsatile frequency in the artery, a baseline pressure of 60 mmHg was imposed, and the pump was then used to create a pulse train consisting of 150 continuous pressure pulses at an amplitude of 10–14 mmHg (added to the baseline pressure) and a frequency of 1 Hz. At 60-mmHg baseline pressure, after 150 total pulses, the filtered volume rate settled to a "steady" value (estimated 3.9 x 10–5 cm3/s) after the initial increase, and no further spurts occurred (1). A set of higher-frequency (2-Hz) 5-pulse trains was then generated in the artery. The whole procedure was repeated but with 20-pulse trains. Therefore, as an example, the artery began at a steady-state of 60-mmHg pressure, received an initial 150-pulse pressure train [(5 pulses x 6) + (20 pulses x 6)] at a set amplitude and frequency (10–14 mmHg and 1 Hz), and then was allowed to stabilize at steady pressure temporarily to measure the initial bubble position. Next, the artery received a 5-pulse pressure train at 2-Hz frequency, the experimental filtration volume was measured, and the process was then repeated for five more 5-pulse trains. After another 150-pulse train set, at 1 Hz, the artery was exposed to a set of six 20-pulse double-frequency pulse trains.
During the higher-frequency pulse trains, the pump plunger was forward in its cylinder when the baseline 60-mmHg pressure was set. This was necessary to discern that the plunger had started and stopped in the same position, especially when using a higher pulse frequency. The effect of having the plunger forward is that when the pump is engaged and the plunger retracts, the pulsatile component is subtracted from the baseline pressure. Therefore, the pulse train appears inverted. Physiologically, however, this is akin to measuring the diastolic pressure in an artery previous to the systolic pressure.
Statistics. Error bars represent means ± SE. The experimental volume corrected for residual distension, averaged over the six 5-pulse values between 5 and 30 pulses, was compared with the corresponding steady-state value using the Student’s t-test (P < 0.05), after assuring a normal distribution of the values using the Kolmorogov-Smirnov test. The 5- to 30-pulse mean experimental value was also compared with the average of the experimental values obtained between 50 and 150 pulses (six 20-pulse trains), the latter parameter being divided by 4 to account for the fact that these were 20-pulse trials rather than 5-pulse trials. The various plots were tested for normality and checked for equal variances, and then linear regression was performed where possible. When linear regression lines could be found between two regions, F tests were then performed to test whether the regions yielded significantly different linear regression lines. Any values that did not satisfy normality were tested using Wilcoxon’s rank sum test.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODS AND MATERIALSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Filtration in deendothelialized vessels. Figures 2 and 3 show the cumulative experimental volumes for the deendothelialized pulsatile vessels (experimental volume is defined hereinafter as the volume entering the artery, as measured from the bubble shift forward after a pulse train minus the volume accounted for by the residual distension) compared with 1) the predicted cumulative steady pressure filtration volume for deendothelialized vessels (hereinafter referred to as deendothelialized steady-state vessels) and 2) the cumulative experimental volumes for the intact pulsatile vessels (intact vessels data from Ref. 1). As in the prior study (1), "cumulative volume" referenced in Figs. 2–4 is the summed volume, either experimental or predicted (steady-state) volumes, over time (i.e., experimental cumulative volume, 20 s, is the sum of the experimental volume values at 5, 10, 15, and 20 s). Timescales in Figs. 2–4 represent the integrated duration of the pulse trains themselves (10 pulses are the sum of two 5-pulse trains, so the artery experiences approximately a total of 10 s of pulsatile pressure).
|
|
|
Linear regressions, where possible, have been summarized in Table 1 for the 5- to 30-pulse region and the 50- to 150-pulse region of Fig. 2. Because these are cumulative volumes, it is difficult to predict even the steady-state slope value without linear regression. The steady-state plots were calculated for each trial and then averaged and accumulated on a point-by-point basis using the time-averaged pressure measurements made during each experimental trial, the Lp value found before instigation of the pulsatile pressure, and the actual time of oscillation for each trial run. Linear regression of the deendothelialized steady pressure and the deendothelialized pulsatile plots showed that from 5 to 30 pulses, the slope of the deendothelialized pulsatile values exceeded the slope of the deendothelialized steady-state values by a factor of 5.3 (15.6 x 10–5 cm3/s versus 2.9 x 10–5 cm3/s, respectively).
|
Figure 3 shows cumulative volumes versus time scatterplots for the same vessels as in Fig. 2, except that these vessels were tested at 80-mmHg baseline pressure. Table 2 summarizes the linear regression slope values at 5–30 pulses and 50–150 pulses. The scatterplots for the intact pulsatile and the deendothelialized pulsatile volumes followed a similar trend as for the 60-mmHg plot; however, F tests of the deendothelialized pulsatile and the intact pulsatile vessels from 50 to 150 pulses did not show a significant difference. However, the point-by-point comparison over 50–150 pulses, using Wilcoxon’s rank sum test, indicated significant differences between the intact and deendothelialized pulsatile vessels at 50, 90, 110, and 130 total pulses. There was no significant difference between the slopes of the deendothelialized pulsatile vessels and the intact pulsatile vessels at 5–30 pulses. These results support the finding that the endothelium does not appear to regulate the transiently increased fluid filtration into the artery wall until about 1 min after the onset of pulsatile pressure.
|
15 s, followed by a region of lower filtration from 25 to 75 s. Because of the nonlinearity of the plots in Fig. 4, F tests were not performed. The values of cumulative volume at 5–30 pulses for intact pulsatile vessels (0- to 1-Hz change) were significantly smaller than those for the (1- to 2-Hz change) frequency vessels (Student’s t-test). The values at 50–150 pulses for the intact pulsatile vessels (0- to 1-Hz change) were also significantly smaller than the corresponding values for the 1- to 2-Hz frequency change vessels.
Calculation to determine whether increased fluid influx remains in interstitium.
One question posed by this study is whether the excess fluid entering the arterial intima, subsequent to onset of pulsatile pressure, filters all the way through the artery wall or just remains within the interstitium. To answer this question, a calculation was performed to determine the increase in medial wall thickness that would result if the accumulated volumes of the intact pulsatile vessels shown in Figs. 2 and 3 remained in the interstitium. At 60 mmHg, the accumulated volume eventually reaches a maximum value of 0.0105 cm3. The formula for the change in volume (
Vfill) corresponding to a change in wall thickness (w) is given in the following equation:
 | (4) |
is average artery length, and w is average artery thickness. By using the values of Ro = 1.0 mm, = 1.8 cm, and w = 0.14 mm, with a Vfill of 0.0105 cm3, we obtain a thickness increase of 92 µm, or a 66% increase in thickness, which should be visible under OCT (1). At 80 mmHg, the accumulated volume increases to 0.208 cm3, which would correspond to a thickness change of 180 µm, or a 131% thickness increase. Such increases in vessel thickness were not observed, indicating that some fluid must have filtered all the way through the artery walls. In addition, these estimated volumes of fluid influx equal or exceed the measured medial water content of the rabbit carotid artery (2), so fluid must, at least at 80 mmHg, be exiting from the adventitial side of the artery. |
DISCUSSION |
|---|
|
TOPABSTRACTMETHODS AND MATERIALSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
A way to adjust the resistance of the endothelium within the time frame seen in Fig. 3 may be linked to the ability of endothelial cells to either maintain or disrupt junctional integrity in the face of changes in circumferential stress or active contraction (10). Typically, in a contracted state, endothelial cells retract from each other, causing the junctional gaps between individual cells to become wider and the endothelium to become more permeable (9). This retraction is mediated by increased cross-bridge cycling of the actin-myosin complex. The retraction is counterbalanced by the tethering forces provided by adhesive proteins in the adherens junctions. These proteins are linked to the actin-myosin complexes, and their function is to maintain a basal level of endothelial barrier function. Endothelial barrier function is directly related to the amount of myosin light chain dephosphorylation, accomplished, for example, by molecules such as cAMP. The greater the dephosphorylation, the more effective is the endothelial barrier function. It is possible that the endothelium in the intact arteries experiences changes in myosin light chain dephosphorylation and hence alterations in the junctional adhesion and intimal permeability, following onset of pulsatile pressure, which could explain the initial increase, and subsequent decrease, in arterial permeability. Evidence that endothelial cell permeability may be altered by pulsatile pressure is provided by a study on cultured endothelial cells in which mechanical deformation of the cells by cyclic strain was found to increase oxidative stress (12). Future experiments to investigate the junctional gap size (i.e., by silver nitrate staining), before and after changes in the pulsatile pressure, could resolve whether modification of intimal resistance accounts for the lower filtration over time in the intact pulsatile vessels at 60 mmHg baseline pressure.
A transient increase in filtration, similar to that seen after the onset of pulsatile pressure following a steady pressure, was observed when the pulsatile frequency was changed from 1 to 2 Hz. Because an increase in pulse frequency often happens in vivo, the transient burst in filtration is therefore probably a normal physiological event as opposed to an artifact caused by a sudden onset of pulsatile pressure. When the frequency of pressure pulsatility is changed, the initial transient burst of filtration, seen with the 60-mmHg baseline pressure, is compressed into 2.5 to 15 s (compared with 5–30 s in the 0- to 1-Hz frequency change arteries). It is not surprising that in the 1- to 2-Hz vessels, the increased fluid flux occurred over a shorter time period compared with the 0- to 1-Hz vessels at 60 mmHg. Pulsing at twice the frequency but the same amplitude probably serves to move more fluid into the gel matrix at a faster rate.
In both the intact and the deendothelialized vessels there was increased filtration compared with that predicted from steady-state measurements. This would indicate that with the onset of a pulsatile pressure, internal changes occur in the arterial media, in addition to possible alterations in endothelial permeability, allowing greater filtration to take place. The internal wall modifications could involve either reorganization of the medial structures to allow more passage of fluid through the wall or a local variation in the internal wall pressure gradient. The lack of any observable difference in fluid influx between the intact and deendothelialized vessels soon after the onset of pulsatile pressure suggests that either the endothelial resistance is reduced during this time or it is counteracted by altered pressure gradients throughout the artery wall. If the phenomenon seen is due to altered pressure gradients, then the pressure gradients involved would have to be greater in the intact vessels than in the deendothelialized vessels to cause an equal amount of filtration. Alternatively, the pulsatile pressure could induce the endothelium to produce substances such as nitric oxide that may alter the configuration and hence the permeability of the arterial media, possibly by relaxing the smooth muscle cells.
In contemplation of the role of the endothelium in large blood vessels, the differences in the filtration responses at longer periods after the onset of pulsatile pressure, between the deendothelialized and the intact vessels, might provide insight into the disease processes associated with inflammation and endothelial dysfunction. For example, if an artery with endothelial damage was exposed to a change in pulsatile frequency, increased fluid flux could drag higher numbers of macromolecules into the artery wall and allow them to percolate interstitial regions formerly closed to fluid transfer. For example, during the initial period of high filtration, more spaces may open up in the medial "ground substance," (11) or mucopolysaccharide gels, allowing access of water and other molecules. Among the transported species could be types of molecules that damage connective tissues and extracellular matrix, such as oxidized LDL, thereby increasing and continuing the inflammatory process. In addition, if the wall ground substance saturates and dampens the excessive convection, then the macromolecules could accumulate beneath the damaged endothelium. Thus a damaged or leaking endothelial barrier might allow large fluxes of lipoproteins and inflammatory species to be convectively driven by pulsatile pressure into the wall to overwhelm wall clearance mechanisms. On the other hand, in an intact vessel, although the initial onset of a pulsatile frequency change would cause increased fluid flux, and hence increased macromolecular flux into the artery wall, the endothelium would then act as a dynamic barrier adjusting to different frequencies and pressures, controlling any subsequent increases in convective fluid and macromolecular fluxes into the artery wall.
A previous study (7) supports the idea that changes in arterial pressure waveforms may affect macromolecular influx as well as fluid influx. Chesler and Enyinna (7) investigated particle deposition in ex vivo arteries under steady, pulsatile, and oscillatory pressure regimes, with and without flow. Microspheres deposited in porcine carotid arteries under these regimes were imaged by using immunofluorescence and confocal microscopy. Under no-flow conditions, pressure waveforms were found to significantly affect the number of 200-nm particles accumulating in the arterial intima. In particular, it was found that a time-varying pressure profile enhanced the amount of particle deposition compared with a steady pressure at the same mean physiological pressure. The transient filtration spurts observed after changes in pulsatile pressure in the present study are consistent with the increased microsphere accumulation in the subintimal spaces found by Chesler and Enyinna (7). Further experiments testing the role of the endothelium in regulating macromolecular intimal deposition following changes in the arterial pressure waveform could provide insight into the major problem of arterial restenosis following stent placement, because this procedure always damages the endothelium.
|
GRANTS |
|---|
|
TOPABSTRACTMETHODS AND MATERIALSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODS AND MATERIALSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J.-B. Michel, O. Thaunat, X. Houard, O. Meilhac, G. Caligiuri, and A. Nicoletti Topological Determinants and Consequences of Adventitial Responses to Arterial Wall Injury Arterioscler. Thromb. Vasc. Biol., June 1, 2007; 27(6): 1259 - 1268. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
