In Vivo/In Vitro Comparison of Rat Abdominal Aorta Wall Viscosity
Influence of Endothelial Function
Abstract Arterial wall viscosity (AWV) is a potential source of energy dissipation in circulation. That arteries, which are known to be markedly viscous in vitro, have lower viscosity in vivo has been suggested but not demonstrated under similar pressure conditions. Endothelium, which may modulate AWV through smooth muscle tone, could contribute to the low level of viscosity in vivo. Our objectives were first to compare AWV of the rat abdominal aorta, in vivo and in vitro, with similar pulse-pressure waves, and second, to determine whether endothelial function influences AWV in vivo and in vitro. The diameter of the abdominal aorta and distending pressure were measured in vivo and in vitro with a high-resolution echotracking system and a micromanometer, respectively. AWV was calculated as the area of the pressure-volume curve hysteresis. After in vivo examination, the arterial segments were isolated in vitro and submitted to resynthesized pressure waves identical to those recorded in vivo. Deendothelialization was performed in vivo by balloon rubbing; then arteries were examined either in vivo or in vitro. AWV was markedly lower in vivo than in vitro (6.6±0.7 versus 22.7±3.7 J · m−1 · 10−5, respectively; P<.001). After deendothelialization, a sustained 40% increased AWV was observed during a 15-minute follow-up (P<.01). In vitro, deendothelialized arteries have a 64% higher AWV than segments with endothelium (P<.01). Our results indicate that the physiological effective viscosity, measured in vivo in intact animals, is threefold lower than the intrinsic viscosity of the arterial wall, measured in vitro. Endothelium removal determines a sustained increase in AWV, either in vivo or in vitro. These results suggest that active mechanisms compensate for intrinsic viscosity under physiological conditions. One of these energy-saving mechanisms might be dependent on normal endothelial function.
Reprint requests to Stéphane Laurent, MD, PhD, Service de Pharmacologie, Hôpital Broussais, 96, Rue Didot, 75674 Paris Cedex 14, France.
- Received March 5, 1996.
- Accepted November 1, 1996.
Viscoelasticity has long fascinated physiologists, because of the belief that the underlying mechanisms are the key to the fundamental mechanisms for generating smooth muscle force.1 Although the viscoelastic behavior of large arteries has been extensively investigated, few studies have focused on viscosity itself. The ejection of the heart is pulsatile, while peripheral circulation is continuous. Large artery function is not only to conduct the blood from the heart to the periphery but also to store the potential energy generated by the heart during systole2 3 4 and restore it during diastole. Thus, large arteries can preserve the potential energy (ie, pressure) from attenuation along the arterial tree so that driving pressure remains high enough at the periphery to drive the blood in resistive networks. During systole, the elastic behavior of the arterial wall allows the internal diameter to increase proportionally to blood pressure. The stored energy is proportional to the product of strain and stress.5 Purely elastic materials allow the whole stored energy to be restored during the unloading phase. However, biological tissues in general, and arteries in particular, are not purely elastic but exhibit a markedly viscous behavior. Vascular tissues exhibit greater dimension at any given pressure during unloading than during loading phase, which determines a hysteresis loop on the pressure-volume relationship. The area of the hysteresis loop has a dimension of energy. Thus, while part of the energy stored by the arterial wall during elastic distension is fully restored, the remaining part of the energy corresponding to the viscous deformation (area of the hysteresis loop) is dissipated within the arterial wall.5 6 7 Thus, AWV is considered as the main source of pressure-wave attenuation along the arterial tree in theoretical models.8 In these respects, AWV, an energy-dissipating mechanism, can reduce the efficiency of heart-vessel coupling.5
AWV has mainly been determined in vitro1 9 10 and is easily measurable. The determinants of AWV are the amplitude and frequency of pulsatile stress, together with arterial smooth muscle content and its degree of contraction.2 11 12 AWV was very small in the few published in vivo studies,4 6 7 13 14 despite the presence of physiological stimuli reported to increase AWV. It has not been clearly determined whether AWV is smaller in vivo than in vitro at the site of conduit arteries. Furthermore, AWV may have been underestimated in vitro. Indeed, as viscosity depends on the frequency and velocity of stretching,2 the low-amplitude sinusoidal pressure waves used in in vitro experiments may not affect arterial wall behavior in the same way as the complex pressure waves observed in vivo.1 To our knowledge, AWV values have not been compared in vivo and in vitro with similar pulse-pressure waves.
Various interdependent factors may contribute to the apparent in vivo/in vitro difference in AWV, including the pulse-pressure wave pattern, innervation and vascularization of the arterial wall, pulsatile flow and shear stress, and smooth muscle tone. In particular, viscosity correlates with the activity of smooth muscle in a given artery.1 2 This activity results from complex interactions between the endothelium, smooth muscle cells, and humoral and mechanical factors.15 16 The release of endothelium-derived relaxing factors is modulated by the amplitude and frequency of flow, pulse pressure,17 18 and mechanical deformation,19 while the endothelium modulates stretch-induced sustained and phasic myogenic responses.16 20 21 22 In this study, we focused on the influence of the endothelium on AWV.
We first compared AWV at the site of a large muscular artery (the rat abdominal aorta), in vivo and in vitro, with similar pulse-pressure waves, and then investigated the potential influence of endothelial function on AWV in vivo and in vitro.
We used 27 male Wistar rats (Iffa-Credo, France) aged 12 weeks and weighing 320±20 g (mean±SD). The animals were managed according to French Ministry of Agriculture guidelines. Experiments were conducted under disodium thiopental anesthesia (50 mg/kg IP). Animals were killed by exsanguination.
The comparison between in vivo and in vitro viscoelastic properties of rat abdominal aorta was performed in six rats during a first experiment, in which AWV (Fig 1⇓) was first determined in vivo and then in vitro on the same arterial segments. The role of endothelium was studied, under in vivo and in vitro conditions, in a second and third experiment. In the second experiment, nine rats were studied in vivo, before and after abdominal aorta DE, and compared with five sham rats. In the third experiment, DE of the abdominal aorta was performed in vivo in seven animals; the aortas were then studied under in vitro conditions and results compared with those of in vitro controls with intact endothelium, matched for mean and pulse pressures.
Experiment 1: In Vivo and In Vitro Determinations of AWV
In Vivo Determination
The 2F microtransducer was introduced through the left femoral artery of anesthetized animals for aortic blood pressure measurements. A median laparotomy was then performed and the AA was exposed but not dissected. A 10-MHz focalized transducer was stereotactically positioned 1 cm above the aortic trifurcation, using warm (37°C) isotonic saline as coupling medium. The position of the pressure probe was adjusted visually and assumed to be correct when the tip of the catheter was detected on the radiofrequency signal of the aorta. Pulsatile changes in diameter and pressure were then measured during three periods of 4 seconds each.
In Vitro Determination
The experimental setup developed for in vitro determination of AWV is represented in Figs 2 through 4⇓⇓⇓. Arterial diameter and pressure were measured and computed in the same way as in the in vivo determination.
Abdominal aortic segments were collected immediately after in vivo investigations. The collateral side branches were ligated carefully under a binocular microscope and the AA was gently separated from the inferior vena cava. Two 3-0 silk laces were then placed around the AA; the first at the level of the left renal artery, and the other at the level of the aortic termination. The length of the AA was measured in situ between the laces with a two-point compass. The segment of artery was quickly removed and placed in ice-cold Krebs buffer (118 mol/L · 10−3 NaCl, 4.6 mol/L · 10−3 KCl, 2.5 mol/L · 10−3 CaCl2 2H2O, 1.16 mol/L · 10−3 MgSO4, 1.16 mol/L · 10−3 KH2PO4, 11.1 mol/L · 10−3 glucose, 25 mol/L · 10−3 NaHCO3, 95% O2, 5% CO2). The explant length was measured and shrinkage (35±5%) calculated. The remaining side branches were ligated. The proximal end of the AA was cannulated with a 2F cannula and connected to a perfusion line. The distal end was cannulated with a 2F high-fidelity microtransducer to measure intraluminal pressure. The arterial segment was then mounted in a specially designed organ chamber containing oxygenated Krebs buffer at 37°C. The AA segment was then pressurized with Krebs, extended to its in vivo length, submitted to a constant pressure of 100 mm Hg, and allowed to equilibrate for 30 minutes. The leakage was checked by temporarily emptying the chamber and estimating the speed of refilling. When leakage exceeded 1 mL/min and could not be corrected, the experiment was discarded. Pressure was then set at the mean arterial pressure in vivo for 15 minutes, and the artery was connected to the pressure-wave synthesizer (described above). The perfusion line was composed of low-compliance polyethylene tubing, a pressure reservoir containing oxygenated Krebs buffer at 37°C, and a proximal adjustable Windkessel segment. A rectangular negative ramp signal was found to provide the best fit for the morphology of the pressure wave observed during the corresponding in vivo experiment (Fig 5⇓). The pressure-signal morphology was fine tuned by careful adjustment of the Windkessel chamber size.
The amplitude and frequency of pulse pressure were adjusted to the pulse pressure measured in vivo for each rat. The artery was submitted to this pulsatile pressure regimen for 5 minutes, and pressure and diameter were then measured for 4 seconds. To ensure that the frequency content of the distending pressure in vitro was identical to that in vivo, we performed a Fourier analysis (8192 samples) of the pressure waves under both conditions and compared them afterward.
Experiment 2: Effects of DE on AWV In Vivo
DE was performed by means of the balloon catheter procedure originally described by Clowes et al.23 As our objective was to determine the maximal DE not altering arterial wall function, we slightly modified the classical method to avoid overstretching the aortic wall and checked that aortic ring contraction and relaxation were preserved (see below).
The animals were prepared as described in experiment 1, and the pressure-diameter relationship was measured at baseline. A 2F Forgarty catheter (Baxter) was then introduced through the left common carotid artery and gently pushed down through the thoracic aorta toward the abdominal aorta. When the tip of the catheter was located near the area of ultrasonic diameter measurement, it was pushed 1 cm further. The balloon was then inflated with the minimal volume of air (usually 50 μL) required to obtain occlusion of the artery (confirmed by the disappearance of the downstream pressure signal). The catheter was then slowly withdrawn until resistance was felt at the level of the diaphragm. The procedure was repeated once, and the catheter was then deflated and withdrawn. In sham animals, the catheter was pushed until it reached the site of diameter measurement but was not inflated. The pressure-diameter relationship was measured 1, 5, and 15 minutes after the procedure.
To ensure that the frequency content of the blood pressure after DE was identical to that before DE, we performed a Fourier analysis (8192 samples) of the pressure waves under both conditions and compared them afterward.
To check that our DE procedure damaged the endothelium without altering smooth muscle function, aortic rings were studied in isometric conditions. The abdominal aorta was removed from five DE and five sham animals and placed in ice-cold Krebs buffer. The arteries were then dissected and cut into rings 3 to 4 mm long. Two rings from each animal were placed in an isometric organ chamber (Marty Technology) and allowed to equilibrate at 37°C in oxygenated Krebs buffer. Segments were exposed to 60 mol/L · 10−3 KCl for 15 minutes and then to cumulative concentrations of norepinephrine (10−9 to 10−5 mol/L). When the maximal contractile response was obtained, endothelium-independent relaxation was tested by adding 10−5 mol/L papaverine. The maximal force developed under KCl stimulation and the degree of endothelium-independent relaxation were taken as indices of smooth muscle function. The aortic rings were then contracted with 3×10−7 mol/L norepinephrine and exposed to cumulative doses of carbamylcholine (10−9 to 10−5 mol/L). The capacity of aortic segments to relax in a carbamylcholine concentration–dependent manner was used as an index of endothelial integrity.15 Data for each concentration are expressed as the percentage of maximum baseline contraction.
Experiment 3: Effects of DE on AWV In Vitro
DE was performed in vivo by means of a balloon catheter, following the procedure described in experiment 2. Then, abdominal aorta segments were collected in the same manner as in experiment 1 and studied in vitro as described above. The mean and pulse pressures applied in vitro to these arterial segments were chosen identical to those measured in vivo previously from the DE procedure.
Determination of the Dynamic Diameter-Pressure Relationship
As viscosity is measured by the area of the hysteresis of the diameter-pressure relationship (Fig 1⇑), we first simultaneously measured pulsatile changes in diameter and pressure to determine their relationship. In vivo and in vitro measurements were made in the same way. Pulsatile changes in diameter were measured with the NIUS 1 echotracking device (described below). Pulsatile changes in distending pressure were measured using a 2F high-fidelity microtransducer (Millar Instruments, Nycomed; Fig 4⇑).
Aortic diameter was measured by the NIUS 1 echotracking device (ASULAB Research Laboratory), which has been fully described elsewhere.24 25 26 Briefly, the device measures internal diameter and its systolic-diastolic variations with a precision close to 50 μm and 1 μm, respectively. This high precision is made possible by the quadratic fit of a specific window of the radiofrequency signal, the high frequency (75 MHz) of signal tracking, and the digitization of this signal on 8 bits. The software was slightly modified for viscosity measurements. Diameter and pressure were simultaneously digitized and stored at a frequency of 2.5 kHz, giving a 4×10−4-second time domain definition (0.0125 radian for a frequency of 5 Hz).
The Doppler technique was used to place the probe perpendicularly to the arterial axis, in its largest cross-sectional dimension. After the transducer was switched to radiofrequency mode, the backscattered echoes from both the anterior and posterior walls were visualized on an oscilloscope. The radiofrequency signals of both walls exhibiting a high signal-to-noise ratio were easily tagged by two electronic trackers, so that their movements could be monitored continuously. Recordings of the tracker movements were used to derive a digitized signal of arterial diameter variations. Distending pressure was measured simultaneously at the same position. The computerized acquisition system derived the lumen cross-sectional area-pressure curve from the two continuous signals of arterial diameter and pressure and fitted it using an arctangent model with three independent parameters (α, β, and γ), as described by Langewouters et al27 : where ATN is the arctangent, P is instantaneous distending pressure, and D is internal diameter. The three parameters α, β, and γ have a physiological meaning.27 α represents the maximal LCSA divided by π/2, β is the pressure at which the compliance is maximum, and γ is the pressure at which compliance is half the maximum.
Various definitions of compliance are used in the literature.28 In the case of a cylindrical vessel, arterial cross-sectional compliance (C) is defined by the change in cross-sectional area (ΔLCSA) for a given change in intravascular pressure (ΔP). Owing to the nonlinearity of the cross-sectional area–pressure curve, compliance falls as pressure increases. It is therefore important to establish the compliance-pressure curve in the systolic-diastolic range (dynamic cross-sectional compliance). This value was obtained by deriving equation 1 with respect to pressure as follows: Thus, equation 2 refers to dynamic cross-sectional compliance. Arterial cross-sectional distensibility (Dist) is defined by the relative change in cross-sectional area (ΔLCSA/LCSA) for a given change in distending pressure and represents compliance normalized to the corresponding cross-section of the artery: These equations allowed us to calculate geometrical and functional parameters at any pressure value in the systolic-diastolic range. As AWV determines a hysteresis of the pressure-diameter relationship, the rising limb is distinct from the descending limb. It is thus important to specify the part of the curve on which the fitting is done. For practical reasons, the software begins with a least-squares minimization of the hysteresis loop by iteratively shifting the time series of pressure and diameter. The fitting process is performed on the resulting curve, which corresponds to the diastolic part of the curve.24 Since the pressure-diameter relationship is curvilinear, the distensibility and compliance values depend on the pressure. We chose to normalize the geometrical and functional parameters at an arterial pressure of 125 mm Hg, which was close to mean blood pressure in the entire experimental group of animals.
Evaluation of AWV
AWV can be estimated by the hysteresis of the pressure-volume relationship or by the ratio of the dynamic-to-static elastic moduli.5 12 We focused on the hysteresis loop of the pressure-LCSA relationship, as the ratio of the dynamic-to-static elastic moduli cannot be easily determined in vivo for obvious reasons.
The energy per unit of arterial length exchanged from the blood to the arterial wall during distention can be written as: where dW is dissipated energy, ∂L is an arbitrary length, P is pressure, and dLCSA is the elementary variation in lumen area.
AWV determines the emergence of a hysteresis loop (Fig 1⇑), in which it is possible to distinguish the “distention” limb (A→B) during systole from the “retraction” limb during diastole (B→A). Arterial diameter is smaller during distention than retraction at a given distending force. The hysteresis loop corresponds to a classical adiabatic cycle in thermodynamics. Indeed, the energy stored during the distention phase (WE) is not fully restored during retraction. This energy dissipated by the arterial wall (dW) per unit length during one cycle (dL) represents Wv and can be written as: The quantity given by equation 5 corresponds graphically to the area of the hysteresis loop of the pressure-LCSA relationship (Fig 1⇑). We developed dedicated software to measure this area directly from the original pressure and diameter recordings, using an integrative algorithm. Energies are expressed in joules per meter during one cycle. Wv is expressed either in absolute values or as a percentage of WT, with WT=Wv+WE. This software permitted us to detect a phase lag as small as 10−5 radian on simulated sinusoidal time series. The practical precision of this method is given by the precision of pressure and diameter measurements (0.5 mm Hg and 1 μm, respectively) and can be estimated as >1×10−10 J · m−1. Viscosity was measured on all cardiac cycles and was averaged for one cycle.
In Vitro Validation of Viscosity Measurements
A crucial point in the measurement of the hysteresis loop of the pressure-LCSA relationship is the synchronization of the pressure and diameter signals. We therefore developed an experimental setup (Fig 2⇑) to estimate the potential time delay between these two signals. A two-entrance low-compliance chamber was closed with a purely elastic plastic membrane and filled with distilled water. A high-fidelity microtransducer was positioned inside the chamber immediately beneath the membrane, and its motion was measured with the echotracking NIUS system. The whole system was submitted to various pressure waveforms generated by a homemade pressure wave synthesizer. This device was composed of three elements. A function generator (Hewlett Packard) was programmed to generate appropriate electrical signals. These signals were amplified by a solid state power amplifier (Classé Audio 70) and fed into a powerful low-frequency loudspeaker (Focal) coupled mechanically to a low-friction microsyringe (Hamilton). This pressure synthesizer could generate pressure signals adjustable in amplitude, frequency, and shape. We submitted this elastic membrane to various pressure waveforms (square, sinusoidal) of various frequencies (1 to 50 Hz). Any desynchronization in the acquisition or processing of the pressure and diameter signals was detected by the appearance of a hysteresis loop on the pressure-displacement curve. Both transducers were found to be linear in the frequency and amplitude range observed in animals. The two signals were simultaneously acquired, and no time delay was observed. As a result, the hysteresis loop area was negligible (<10−7 J/m). To illustrate the synchronization and linearity of our system, a complex waveform associating a carrying wave of low frequency and high amplitude (2 Hz, 40 mm Hg, respectively), modulated by a high frequency and small amplitude (30 Hz, 10 mm Hg, respectively). The resulting pressure and displacement waves and the pressure-displacement relationship are represented in Fig 3⇑. Despite the wide range of frequency and amplitude present in the pressure signal, no hysteresis was discernible.
Data are expressed as mean±SEM. In vitro measurements were compared with in vivo measurements by using a paired nonparametric Wilcoxon’s test. The effects of DE on arterial wall elasticity and viscosity were analyzed by using repeated-measures ANOVA, with time as the within-subject factor and the group (DE, sham) as the between-subject factor. The C matrix post hoc test was used to test changes in the parameters versus baseline values. All tests were bilateral and values of P<.05 were considered to denote significant differences. All statistical tests were done on SYSTAT software.29
Experiment 1: In Vivo/In Vitro Comparison of Viscoelastic Properties
The frequency content of the in vitro pressure waves was not significantly different from that of in vivo waves (harmonic 1: 6.12±0.45 versus 5.54±0.49 mm Hg; harmonic 2: 3.60±0.32 versus 2.89±0.32 mm Hg; harmonic 3: 1.54±0.20 versus 1.42±0.25 mm Hg; and harmonic 4: 0.66±0.15 versus 0.69±0.10 mm Hg).
LCSA was lower in vivo than in vitro, while distensibility and compliance were markedly higher in vivo (Table 1⇓). AWV was markedly lower in vivo than in vitro, both when expressed in absolute values and when normalized to the total work exchanged. Typical pressure-LCSA relationships obtained in vivo and in vitro in the same artery are represented in Fig 6⇓. The frequency content of the distending pressure in vitro did not differ from that observed in vivo, since the first four Fourier components were not significantly different between conditions.
Experiment 2: Effect of DE on Arterial Parameters In Vivo
Arterial parameters did not differ between sham and DE animals at baseline (Table 2⇓). In sham rats, as no significant time-dependent change in arterial parameters was observed, the four (baseline, 1, 5, and 15 minutes) LCSA-pressure and distensibility-pressure curves were pooled (Figs 7⇓ and 8⇓).
A significant decrease in systolic, mean, and diastolic blood pressures was observed in DE animals 1 minute after DE, but not at later times. No change in pulse pressure or heart rate was observed (Table 2⇑). The frequency content of the distending pressure after DE did not differ from that observed before DE, since the first four Fourier components were not significantly different between conditions during follow-up. The pressure-diameter curve of DE animals was shifted upward by the procedure (Fig 7⇑). At a blood pressure of 125 mm Hg, the increase in aortic LCSA was significant 1 and 5 minutes after DE, but not at 15 minutes. Distensibility was marginally altered by DE, and the distensibility-pressure curve was slightly shifted downward: at a blood pressure of 125 mm Hg, there was a slight but significant decrease in distensibility 1 minute after DE. Aortic compliance, calculated at 125 mm Hg, was not modified 1 minute after DE, but was significantly increased thereafter.
The total work exchanged during the cardiac cycle (WT) did not change during the procedure in DE or sham rats. An important observation was that DE induced a significant increase in AWV (Wv, +41%; WP, +43%), which occurred at the first minute and persisted thereafter (Table 2⇑). A representative pressure-LCSA relationship trace obtained before and after DE is shown in Fig 9⇓.
Experiment 3: Effect of DE on Arterial Parameters In Vitro
DE arteries showed lower diameter, distensibility, and compliance than controls with intact endothelium when examined in vitro (Table 3⇓). The markedly reduced distensibility of DE arteries was accompanied by a decreased ability to store energy during distention: the absolute energy dissipated in viscosity was smaller than in controls. In addition, the percentage of exchanged energy dissipated in viscosity was 64% higher than in controls.
Endothelial and Smooth Muscle Cell Function
Balloon rubbing of the abdominal aorta led to complete inhibition of carbamylcholine-induced endothelium-dependent relaxation (Fig 10⇓). This inhibition was obtained with no detectable alteration of maximal endothelium-independent contraction or relaxation. Indeed, maximal contraction in response to 60 mol/L · 10−3 KCl did not differ between DE and sham animals (2.94±0.23 g versus 3.17±0.18 g, respectively). In addition, endothelium-independent relaxation induced by papaverine was not significantly different in DE and sham animals (−124±6% versus −106±4%, respectively).
Large arteries not only conduct blood from the heart to the periphery but also store energy during systole, restore it during diastole, and transmit it along the arterial tree. This ability to store and transmit energy, which is the main determinant of cardiovascular efficiency, is directly influenced by AWV.
The two main findings in this study are that (1) AWV is lower in vivo than in vitro under similar pulsatile conditions and (2) DE leads to a marked increase in AWV either in vivo or in vitro.
In Vivo/In Vitro Comparison of Viscoelastic Properties
Consideration of Methods
Two prerequisites for this study were accurate measurement of AWV in situ, under physiological pressure and flow conditions, and the comparability of these measurements with those obtained on the same arterial segments in vitro. We had to make theoretical choices as to the characterization of AWV. Indeed, although many models of viscoelasticity have been proposed, none has proved fully able to describe the behavior of the vascular wall.30 Thus, instead of using a model, we took advantage of the ability of the NIUS system to accurately determine the pressure-diameter relationship and chose to measure AWV directly, through its thermodynamic expression, ie, the hysteresis loop area of the pressure-diameter curve. This approach has the advantage of stressing on the physical consequence of viscosity, which is to dissipate energy. This method was initially applied by Brodley6 and Bertram7 to arterial wall mechanics. As discussed in “Methods,” a prerequisite for measuring the hysteresis loop of the pressure-diameter relationship was synchronization of the two signals. In a dedicated experimental setup including a highly elastic membrane, we showed that this was effectively the case, with a negligible hysteresis loop area.
The second challenge was to generate pressure waveforms in vitro as close as possible to those recorded in vivo. Previous works on AWV have involved low-amplitude sinusoidal pressure waveforms because of fundamental and technical problems. In these conditions of low-amplitude deformation, AWV is fairly insensitive to frequency,30 although the amplitude of stress oscillations was not controlled.3 Moreover, viscous phenomena depend not only on pulse-pressure frequency but also on its amplitude and velocity of change.12 The natural pressure waveform exhibits both high-amplitude and high-frequency components, which predominate during the distention phase. We thus developed a pressure wave synthesizer to generate a pressure waveform similar to the natural counterpart. A rectangular negative ramp signal provided the best fit to the morphology of the pressure wave observed during the corresponding in vivo experiment (Fig 5⇑). The pressure signal morphology was fine tuned by careful adjustment of the Windkessel chamber size. The amplitude and frequency of pulse pressure were adjusted to the pulse pressure measured in vivo for each rat. Because of the time constraints, this adjustment was done by “visual fit” during the experiment and checked by harmonic decomposition afterward. As presented in the “Results” section, the first four Fourier components of the pressure waves were not significantly different between in vivo and in vitro baseline conditions, and no correlation between the harmonic content of pressure and viscous parameters was found.
Consideration of Findings
One of the main findings is that AWV was lower in vivo than in vitro. Such a difference had only been suspected, on the basis of comparisons of experiments on different species, at different arterial sites reporting in vivo or in vitro data.3 5 7 9 10 13 In contrast, we observed this in vivo/in vitro difference in AWV in the same animals and at the same arterial site.
Although the experimentally generated pulsatile waveforms were similar to their natural counterparts, flow and smooth muscle tone were absent. The lack of smooth muscle tone may explain the larger in vitro aortic diameter but is unlikely to have led to an overestimation of AWV, as smooth muscle contraction is generally reported to increase AWV.1 2 12 The impact of the absence of flow on viscosity in vitro is more difficult to estimate. Arterial blood flow and smooth muscle tone may have opposite influences on AWV, the former decreasing it through flow-dependent vasorelaxation and the latter increasing it.2 Other mechanisms potentially explaining the in vivo/in vitro difference in AWV include an interruption of blood flow to the vasa vasorum,31 leading to ischemic changes. Although our experiments were conducted in oxygenated medium, we could not rule out transient ischemia and subsequent functional changes. The replacement of blood by Krebs buffer may also have had an influence. In addition, diffusion of electrolytes, water, and macromolecules into the arterial wall may have changed its viscoelastic properties.32 In contrast, an alteration in endothelial function in vitro can be ruled out, as the segments of abdominal aorta submitted to pulsatile pressure were prepared in the same way as in the isometric experiments (which indicated normal relaxation in response to carbamylcholine).
Whatever the mechanisms underlying the in vivo/in vitro difference in AWV, the most important point is that AWV is very low in vivo. AWV is generally described as a passive phenomenon due to the physical properties of arterial wall components.30 This passive phenomenon was best expressed in our in vitro conditions, as the arterial segment was isolated from neurohumoral and flow-dependent control. The fact that AWV is very low in vivo suggests that active mechanisms compensate for intrinsic viscosity in physiological conditions. Such active mechanisms would imply dynamic behavior of arterial wall components during the cardiac cycle, minimizing AWV. As viscosity dissipates energy, these mechanisms (including preserved endothelial function) may improve the efficiency of the cardiovascular coupling energy balance.
Arterial Parameters After DE
Smooth muscle tone results from complex interactions between the endothelium, smooth muscle cells, and mechanical factors. In particular, endothelium-derived relaxing factor release is modulated by the amplitude and frequency of flow, pulse pressure, and mechanical deformation. On the other hand, endothelium modulates stretch-induced sustained and phasic myogenic responses.16 19 20 22 This suggests that endothelial function could interfere with the dynamic behavior of the vessel wall and specifically with AWV.
Consideration of Methods
We used the classical balloon method to suppress the endothelial layer locally at the site of pressure-diameter measurements. We took special care to minimize any damage to the arterial wall caused by balloon overinflation, and no visible arterial dilation occurred. However, a degree of smooth muscle dysfunction cannot be ruled out, even if gross contraction and relaxation functions were preserved when studied isometrically in vitro.
DE may modify AWV through changes in the pressure waveform. Indeed, during the denudation procedure, the aortic pressure proximal to the occlusion likely increased and that distal to the occlusion decreased. These changes in hemodynamic conditions could alter the aortic pressure waveform, even during the follow-up period. To ensure that the frequency content of the distending pressure after DE was identical to that before, we performed a Fourier analysis of the pressure waves under both conditions. We checked that the first four Fourier components of the pressure waves were not significantly different between baseline and post-DE, under in vivo conditions. Therefore, the time-dependent change in AWV after the denudation may well be a true change in viscosity and not a reflection of changes in the waveform of aortic pressure. DE arteries, studied in vitro, were submitted to a pulsatile pressure regimen, the amplitude and frequency of which were adjusted to the pulse pressure measured in vivo after DE for each rat.
Consideration of Findings
The main finding in the DE experiments is that DE causes a large and sustained increase in AWV in vivo and that this increase in AWV is further increased when the DE arteries are studied in vitro. In vivo, the hemodynamic modifications that accompanied the DE procedure were transient and minimal and were thus unlikely to explain the increase in AWV, by contrast with the work of Pagani et al,14 in which the increased hysteresis after nitrix oxide donor administration in sheep was accompanied by large hemodynamic changes. Other mechanisms potentially explaining this increase in AWV include a rise in arterial smooth muscle tone after DE, but this is very difficult to assess in situ.
The relationship between changes in LCSA and distensibility after DE is less obvious for in vitro experiments than for in vivo ones. For instance, transient and opposite changes in LCSA and distensibility occurred in vivo after DE, the former increasing and the latter decreasing. The influence of blood pressure was minimized by the expression of distensibility as an isobaric parameter (Dist125). It is well accepted that a higher LCSA redistributes the loads from the more distensible to the less distensible components of the arterial wall, resulting in a lower distensibility. In vitro, the relationship between the low LCSA and the low distensibility is more difficult to explain, since a low LCSA is supposed to redistribute the loads from the less distensible to the more distensible components of the arterial wall, resulting in a higher distensibility, in a classical model of the arterial wall in which the elastin and collagen fibers are in parallel.33 A more complex model, suggested by O’Rourke et al,34 is that at normal distending pressure, smooth muscle in the arterial wall is in series with some of the stiffer collagenous components but in parallel with the elastic lamellae. Contraction of smooth muscle tenses the collagenous components, thus decreasing distensibility, whereas dilation transfers stresses to the elastic lamellae, thus increasing distensibility. Other mechanisms may be involved.
The high level of viscosity of in vitro DE arteries is accompanied by a very low distensibility compared with in vitro controls. As stated above, an increased AWV may not only increase hysteresis but also reduce distensibility. By nature, a viscous material is stiffer under dynamic pressure regimen, while it dissipates more energy (ie, through increased hysteresis).
In the present work, AWV was increased in vitro, with and without endothelium, while changes in lumen size were opposite. It is therefore unlikely that the increase in viscosity relates only to change in lumen size, shifting the load bearing from distensible to stiffer components of the arterial wall. In addition, although changes in LCSA and distensibility may result from changes in arterial smooth muscle tone, changes in AWV may result from other factors. Indeed, in a preliminary in vitro experiment using the same methodology, AWV was not significantly modified by either phenylephrine or sodium nitroprusside in arterial segments having functional endothelium, despite significant vasoconstriction and vasodilation, respectively.
Two other mechanisms besides smooth muscle contraction may possibly contribute to the increase in AWV after DE: overstretching in response to an abrupt and durable rise in distending pressure, which induces a rearrangement of wall components35 and arterial inflation and suppression of the endothelial barrier, which may promote electrolyte, water, and macromolecule diffusion into the arterial wall. On the contrary, endothelium could actively promote a fast adaptation of the vessel wall to rapid changes in stress (within the cardiac cycle), possibly through the release of NO in response to pulsatile pressure and/or pulsatile flow.
Although the precise mechanism underlying the increase in AWV after DE is unclear, our results indicate that an intact endothelial layer is necessary to keep AWV low. However, despite the large increase after DE, AWV remained lower than in vitro. This finding suggests that factors other than the endothelium are involved in the in vivo/in vitro AWV difference. In addition, it suggests that the endothelial layer, although functional in vitro, cannot bring AWV down to its in vivo level.
The importance of energy dissipation through viscosity for the whole cardiovascular system is difficult to evaluate. The viscous loss of energy, measured at a cross-sectional level, although appearing small in absolute value, may be much higher if it is integrated all along the arterial tree.
In conclusion, active mechanisms, including normal endothelial function, appear to compensate for the intrinsic viscosity of the arterial wall in physiological conditions, thereby improving the efficiency of the cardiovascular coupling energy balance.
Selected Abbreviations and Acronyms
|AWV||=||arterial wall viscosity|
|LCSA||=||luminal cross-sectional area|
|WP||=||viscous energy (as % of total)|
This study was supported by a grant from the Société Française d’Hypertension Artérielle, which is part of the Société Française de Cardiologie, and from Institut National de la Santé et de la Recherche Médicale (INSERM grant No. 494014). We thank Brigitte Laloux for her excellent technical assistance.
Dobrin P. Vascular mechanics. In: Shepherd JT, Abboud FM, eds. Handbook of Physiology: The Cardiovascular System, III. Bethesda, Md: American Physiological Society; 1983:80-83.
Murphy RA. Mechanics of vascular smooth muscle. In: Shepherd JT, Abboud FM, eds. Handbook of Physiology: The Cardiovascular System, III. Bethesda, Md: American Physiological Society; 1983:325-351.
Gow BS, Taylor MG. Measurement of viscoelastic properties of arteries in living dogs. Circ Res. 1968;23:111-122.
McDonald DA, Gessner U. Wave attenuation in viscoelastic arteries. Proceedings of the First International Conference on Hemorheology. Elmsford, NY: Pergamon Press Inc; 1968:113-125.
Learoyd BM, Taylor MG. Alteration with age in the viscoelastic properties of human arterial walls. Circ Res. 1966;18:278-292.
Milnor W. Principle of hemodynamics. In: Cardiovascular Physiology. New York, NY: Oxford University Press; 1990:78-80.
Patel DJ, Janicki JS, Vaishnav RN, Young JT. Dynamic anisotropic viscoelastic properties of the aorta in living dogs. Circ Res. 1973;32:93-107.
Pagani M, Mirski I, Baig H, Manders T, Kerkhof P, Vatner S. Effect of age on aortic pressure-diameter and stiffness-stress relationships in unanesthetized sheep. Circ Res. 1979;44:420-429.
Furchgott R. Role of endothelium in responses of vascular smooth muscle. Circ Res. 1983;53:557-573.
Rubanyi GM. Mechanoreception by the Vascular Wall. Mount Kisco, NY: Futura Publishing Company Inc; 1993:89-121.
Hutcheson IR, Griffith TM. Release of endothelium derived relaxing factor is modulated both by frequency and amplitude of pulsatile flow. Am J Physiol. 1991;261:H359-H363.
Lamontagne D, Pohl U, Busse R. Mechanical deformation of vessel wall and shear stress determine the basal release of endothelium-derived relaxing factor in the intact rabbit coronary vascular bed. Circ Res. 1992;70:123-130.
Katusic Z, Shepherd JT, Vanhoutte PM. Endothelium-dependent contraction to stretch in canine basilar arteries. Am J Physiol. 1986;252:H671-H673.
Lacolley P, Glazer E, Challande P, Boutouyrie P, Mignot JP, Duriez M, Levy B, Safar M, Laurent S. Structural changes and in situ aortic pressure-diameter relationship in long-term chemical-sympathectomized rats. Am J Physiol. 1995;296:H407-H416.
O’Rourke M. Mechanical principles in arterial disease. Hypertension.. 1995;26:2-9.
SYSTAT, Inc. SYSTAT, Version 5.01. Evanston, Ill: SYSTAT Inc.
Fung YC. Biomechanics: Mechanical Properties of Living Tissues. 2nd ed. New York, NY: Springer-Verlag; 1992:281-320.
Stefanadis C, Vlachopoulos C, Karayannacos P, Bouloudas H, Stratos C, Fillipides T, Agapidos M, Toutouzas P. Effect of vasa vasorum flow on structure and function of the aorta in experimental animals. Circulation. 1995;91:2669-2678.
Tedgui A, Merval R, Esposito B. Albumin transport characteristics of rat aorta in early phase of hypertension. Circ Res. 1992;71:932-942.
O’Rourke M, Kelly R, Avolio A, Hayward C. Potential for reversing the ill effects of aging and arterial hypertension on central systolic pressure and left ventricular hydraulic load by arterial dilator agents. Am J Cardiol. 1989;63:381-441.