INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE SPLANCHNIC VASCULATURE, and especially the liver, potentially contributes a large fraction of the blood mobilized under conditions of stress (p. 1539 of Ref. 13). Vasoactive hormones may induce this blood redistribution, but the mechanisms are not well defined. By measurement of total hepatic blood flow and the pressure in the hepatic venules, which collect the blood from the hepatic sinusoids, it is possible to localize changes in microvascular resistance. Also, by measurement of changes in liver lobe thickness, to estimate changes in liver volume, it is possible to partition, under some conditions, active changes in vascular capacitance properties from passive changes in stressed blood volume. A passive change in vascular volume is a change in stressed volume related to a change in distending pressure of the vessel. The stressed volume of a vascular segment is defined as the product of the compliance of the segment and the transmural distending pressure (31-33, 36, 37). An active change of vascular volume connotes a change in activity of the smooth muscle in the blood vessel walls (16, 32, 33, 37) or in the activity of contractile elements in the stellate cells of the hepatic sinusoids (3, 18, 20, 46). Thus vasoactive agents may induce active changes in the unstressed volume or compliance of the sinusoids or other blood vessels. The unstressed volume of a segment is defined as the volume at a transmural pressure of zero. The unstressed volume is calculated by extrapolating the relatively linear segment of the pressure-volume relationship over the normal operating range to zero transmural pressure. Experimentally, the unstressed volume is the difference between the total volume and the stressed volume (31-33).

Shoukas and Sagawa (36, 37) developed a technique to measure systemic vascular compliance and explicitly noted that the carotid sinus baroreceptor reflex could vary the vascular compliance or the unstressed vascular volume or both. In their early studies, they reported that the reflex caused no significant change in systemic vascular compliance (37), but their later studies and those of others have reported small changes in compliance (4, 7, 14). Greenway and co-workers (12, 16) have concluded that changes in splanchnic bed unstressed volume, and not compliance, provide the major blood volume reserve. In the studies reported here, we did not attempt to measure changes in compliance, which would have required measuring the change in distending pressure in response to a change in volume without any change in contractile element activity. Active venoconstriction may be assumed if the contained volume decreased in conjunction with an increase in distending pressures.

In a preceding study, we examined the mechanisms of capacitance and resistance responses to norepinephrine (NE) and the vasodilatory drugs isoproterenol, adenosine, histamine, and acetylcholine in rabbit livers (34). In that study, as in this one, norepinephrine significantly increased hepatic vein resistance and microvascular and portal vein pressure while actively decreasing the vascular volume. Isoproterenol, at closely similar molar infusion rates, had little effect in the rabbit livers. Histamine passively induced engorgement by increasing the outflow resistance. Adenosine also passively induced engorgement by reducing upstream resistance and so increasing flow to increase distending pressure, thereby passively increasing blood volume.

These hepatic microvascular capacitance and resistance responses to various vasoactive agents are not well know because of the difficulty of simultaneously measuring the blood flow, volume (or diameter), and pressure within the hepatic venules, sinusoids, and portal venules. We used the complex servo-null micropipette technique to measure hepatic venular pressures because we have reported (25) that a catheter placed in a hepatic vein to measure intrahepatic venular pressure gives erroneously high values. Lautt et al. (24), using retrograde hepatic venous catheters retracted ~5 mm from the wedge position, reported that hepatic intralobar pressures were insignificantly different from portal venous pressure, thereby placing almost all of the resistance between the portal vein and vena cava in the hepatic outflow veins. Our direct measurements suggest that the portal venule and hepatic venous resistances are similar (26, 34).

The methods used in this study were very similar to those we used to determine the normal pressures within the microscopic portal venules, sinusoids, and hepatic venules of individual lobules (26). A brief review of the anatomic and physiological characteristics of the hepatic circulation and the mechanisms related to redistribution of blood to or from the liver was recently published (34) as was a brief characterization of vascular capacitance (33).

The four vasoactive drugs studied have different systemic actions and were expected to selectively stimulate different mechanisms influencing liver capacitance. NE is released in response to stress, excitement, or vigorous physical activity. It actively reduces the blood volume of the liver but does not change the compliance (13, 22). Angiotensin (ANG) is an important participant in both the regulation of salt and water excretion by the kidneys and the regulation of arterial blood pressure via changes in total peripheral resistance. It is involved in some forms of hypertension. Endothelin (ET) is one of the most potent vasoconstrictors known (45). However, the physiological role of ET continues to be unclear. Vane (40) hypothesized that it might be contributing to disease states such as hypertension. Its primary function may be related only to serious pathology and cell death in which local necrotic areas are vascularly isolated from the remainder of the body to reduce the spread of toxins. Vasopressin (VP), the antidiuretic hormone, influences water excretion but at high concentrations is also a vasoconstrictor. Severe hemorrhage stimulates VP secretion to increase its blood concentration >50-fold (8). This helps sustain the systemic arterial blood pressure (8) and has been considered to participate in arterial pressure stabilization (6).

Our hypotheses were as follows: 1) the four vasoactive agents, ANG, ET, NE, and VP, cause active constriction of the hepatic capacitance vessels; and 2) the hepatic venous resistance responds in the same direction and proportion to each of the four vasoactive agents as the splanchnic resistance.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study, approved by the Indiana University Animal Care and Use Committee (study MD285b), closely followed the procedures used in our recent study (34). New Zealand White rabbits [3.6 ± 0.3 kg (means ± SD), n = 11 rabbits, 2 of which were males] were sedated with 4 mg/kg chlorpromazine. Under local anesthesia, a central ear artery and vein were cannulated to measure systemic arterial pressure (P_sa) and to infuse the anesthetic. A solution of 10 mg/ml -chloralose and 100 mg/ml urethan was infused intravenously for 30 min to a dose of 50 mg/kg body wt chloralose and 500 mg/kg body wt urethan. We infused additional anesthetic if the rabbit exhibited any unexpected movement or a response to a stimulus such as touching the eye or pinching a toe. Rabbits were chosen to provide a preparation large enough for insertion of catheters for the various measurements without interference of normal flow, yet small enough to be able to position the animal under the microscope.

Lactated Ringer (70%) and 6% dextran-70 with 5% dextrose (30%) were infused at a rate of 11 ± 1 ml · h¹ · kg body wt¹ to replace fluid losses. The animals were allowed to breathe spontaneously unless the arterial pressure or respiratory rate decreased (6 of 11 animals), at which time positive-pressure artificial respiration was used (34).

To measure the portal venous pressure (P_pv), a 1.0-mm-outer diameter (OD) polyethylene catheter was advanced from a cecal vein into the portal vein to within 3 cm of the liver hilum. To measure the abdominal vena caval pressure (P_avc), a 1.3-mm-OD catheter was advanced through a branch of the right jugular vein and positioned ~2 cm upstream from the diaphragm at the level of the hepatic venous outflow.

The liver lobe was exposed and supported as previously described (26, 34). Microvascular pressure was measured with the use of a servo-null micropipette pressure-measuring system (model 5A; Instruments for Physiology and Medicine, San Diego, CA) in conjunction with a microscope, video camera, and recorder (34). The field was epi-illuminated. The diameters of the hepatic venules were measured from the calibrated video image (1 mm = 1.55 µm). Hepatic venules were identified as large microvessels fed from sinusoids in a convergent flow pattern.

The pressure transducers were calibrated and zero-referenced to the middle of the right atrium and continuously flushed with saline (34). The servo-null micropressure system was calibrated by placement of the pipette in a chamber partially filled with 0.9% NaCl and pressurized at 0, 10.0, and 20.0 mmHg with the use of a water manometer. The pressure data resolution was 0.08 mmHg; the relative uncertainty was estimated to be <0.2 mmHg (34).

The total hepatic blood flow (F_hv) was measured with a 6-mm-wide ultrasound transit-time flow probe (Transonics, Ithaca, NY) placed around both the portal vein and the hepatic artery within 2 cm of the hepatic hilum. The surgical trauma required to adequately set up the preparation for microvascular pressure measurements limited our options for accurately measuring hepatic blood flow. Reduction of movement at the puncture site to less than ~10 µm during each breath required positioning of a lobe on a rigidly supported stationary surface. The liver was further isolated from the movement of the diaphragm with a barrier, with some risk of restricting hepatic blood flow. Because the rabbit hepatic arterial anatomy is complex and up to 1.5 h can be required to isolate the artery (2), we chose not to risk possible catastrophic failure and additional trauma but rather placed both vessels in the flow probe. Even so, once the lobe was adequately positioned for the microvessel pressure measurements, the probe was inaccessible and so could not be repositioned. This reduced the reliability and accuracy of our total hepatic flow data but did provide evidence of partial vascular occlusion as we manipulated the isolation barrier. We often saw small flow pulsations in synchrony with the heart rate in the F_hv recording, thus substantiating the inclusion of the hepatic artery.

Liver lobe thickness changes were obtained by measurement of the distance between the microscope focus point at the pipette tip in the vessel and the stationary support table with the use of a variable differential transformer (resolution 8 µm) attached to the microscope head to sense the position of the micropipette tip. To keep the ~5-µm pipette centered in the 50-µm-diameter vessels required continuous refocusing during the responses to the vasoactive agents. Otherwise, movements of the vessel, related to respiratory activity of the animal, could place the pipette tip next to the vessel wall, thereby distorting the conductivity field and leading to intolerable noise and loss of signal. During NE infusion, the thickness decreased ~350 µm, and lateral movement often occurred. The focusing for thickness estimates was not subjective, because during the responses to drugs, the operator was required to control the pipette position by use of a three-dimensional micromanipulator, move (with the other hand) the table holding the animal to keep the field containing the puncture site visible, and repetitively refocus the microscope (by use of a foot control) to keep the tip in view. The data were recorded for later objective analysis. A percent change in thickness underestimates the corresponding change in volume, because the thickness measurement was only in one dimension, whereas a volume change is likely to be in three dimensions (15).

Data acquisition. The raw signals from all transducers were averaged with a low-pass four-pole Bessel filter (f_c = 0.18 Hz) and digitized every 0.1 min with the use of a 12-bit converter (Keithley series 500). The last 0.1-min interval in each minute of data was used as the value for that minute. The hepatic venous resistance (R_hv) was computed as follows: R_hv = (P_µhv  P_avc)/F_hv, where P_µhv is hepatic venular pressure (by servo-null micropipette). Splanchnic bed resistance (R_spl) was computed as follows: R_spl = (P_sa  P_avc)/F_hv.

Protocol. The micropipette was immersed in the fluid pool above the vessel to establish the transmural pressure baseline and was then inserted into the vessel. We used a 2-min control period, a 5-min period of drug infusion, and a 3-min recovery period. The drugs were infused via the P_pv catheter from a microsyringe at up to 193 µl/min with an uncertainty of ~1 µl/min (34). The pressure flushing lines were clamped periodically for 0.2 min to measure and verify the constancy of the flow-related pressure offset (0.5-2.5 mmHg).

1

1

Statistics. The average of the effects of each drug during the last 3 min of infusion was compared with the preceding 2-min control periods and expressed as differences or as percent change from control. A Student's t-test of the change was computed. Differences were considered significant if P < 0.05. Data variability is described by the standard deviation (SD).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The livers at the end of the experiments, which were drained of blood and without the gall bladder, were 2.53 ± 0.36% of the body weight. The F_hv averaged 954 ± 167 ml · min¹ · kg liver wt¹. The thickness of the liver lobe at the site of measurement averaged 5.6 ± 2.0 mm. The hepatic venule diameters averaged 54 ± 17 µm.

Control data from the 2-min period before the drug infusions were obtained for 63 runs from 11 rabbits (Table 1). Mean P_sa and P_pv were similar to those reported earlier (34). The combined hepatic inflow (assumed to equal outflow, F_hv) averaged 24.4 ± 8.7 ml · min¹ · kg body wt¹ and was lower than in the earlier study. The P_µhv averaged 1.9 mmHg less than the P_pv and 2.0 mmHg higher than the P_avc. The venous pressure gradient across the liver (P_pv  P_avc) averaged only 3.9 mmHg, a value similar to that reported before (26, 34). The high coefficient of variability for flow (Table 1) accounts for some, but not all, of the uncertainty of the computed resistances. The portal venous concentration of the drugs was estimated as the ratio of the rate of drug infusion to the rate of blood flow.

NE was infused via the portal vein at 4.7 ± 0.6 µg · min¹ · kg body wt¹ to administer a dose of 138 nmol/kg body wt over the 5-min infusion period. The P_sa, P_pv, and P_µhv were increased significantly, but the liver lobe thickness (THK) decreased 7% (Fig. 1 and Table 1). The P_avc, heart rate (HR), and ratio of the P_µhv-to-P_avc gradient to the total P_pv-to-P_avc gradient (Grad) were not significantly changed by this infusion rate. The F_hv significantly increased after ~3 min of infusion. After the first minute of infusion, there were marked increases in both R_spl and R_hv (between the 54-µm hepatic venules and the abdominal vena cava; Fig. 1). Within 3 min after the infusion was stopped, all variables, except P_pv, F_hv, and THK, had returned to control. The increase in P_sa and R_spl strongly suggests that the liver extracted only part of the infused vasopressor agent. The responses were less than those reported earlier (Table 1 of Ref. 34) because the infusion rates were less.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Cardiovascular responses to norepinephrine (NE) infused via the portal vein at 4.6 ± 0.6 µg · min1 · kg body wt1. Psa, systemic arterial pressure (as percentage of control); Fhv, hepatic venous outflow (as percentage of control); Ppv, portal venous pressure; Pµhv, hepatic venular pressure; Pavc, abdominal vena cava pressure (change from control in mmHg); R, resistance; Rhv, hepatic venous outflow resistance [(Ppv  Pavc)/Fhv]; Rspl, splanchnic vascular resistance [(Psa  Pavc)/Fhv]; HR, heart rate; Thick, liver lobe thickness (percentage of control) at site of Pµhv measurement. During the 3- to 5-min periods of infusion, all variables significantly changed from control except Pavc and HR (see Table 1).

ANG II was infused at 0.38 ± 0.11 µg · min¹ · kg body wt¹ to administer a total dose of 1.8 nmol/kg body wt. This caused P_sa, P_pv, and P_µhv to increase (Fig. 2 and Table 1). The large increases in R_hv and R_spl lead to a small decrease in F_hv even though the perfusion pressure increased. The decrease in liver thickness, in conjunction with an increase in distending pressures (P_µhv and P_pv), suggests that most of the reduction in vascular capacitance was active. The significant decrease in HR was associated with the increased P_sa. Only P_avc and Grad were not changed significantly. By 3 min after the end of the infusions, the variable values returned almost to control.

View larger version (33K): [in this window] [in a new window]   Fig. 2.   Cardiovascular responses to angiotensin (ANG) II infusion via the portal vein at 0.38 ± 0.11 µg · min1 · kg body wt1. For explanation of abbreviations, see legend to Fig. 1. During the 3- to 5-min periods of infusion, all variables significantly changed from control except Pavc (see Table 1).

VP was infused at 0.047 ± 0.026 µg · min¹ · kg body wt¹ (equal to 0.043 nmol · min¹ · kg body wt¹) to administer a total dose of 0.22 nmol/kg body wt. This caused R_hv and R_spl to increase so much that F_hv decreased to 60% of control even though P_sa increased 28% (Fig. 3 and Table 1). As a consequence of the decrease in F_hv, P_pv and P_µhv also significantly decreased. Lobe THK was not changed even though the distending pressure was decreased just downstream from the sinusoids and in the portal vein. Most variables had not returned to control by 3 min after the VP infusion was stopped.

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Cardiovascular responses to vasopressin (VP) infused via the portal vein at 0.047 ± 0.026 µg · min1 · kg body wt1. Abbreviations as in Fig. 1. (Note change in scale for resistances.) During the 3- to 5-min periods of infusion, all variables significantly changed from control except Pavc and lobe thickness (see Table 1).

ET-1, when infused at a low rate (0.0012 ± 0.0006 µg · min¹ · kg body wt¹ rate), caused R_spl to significantly decrease to 91.6 ± 9.6% of control and flow to increase 6.8 ± 10.0%. Even at these low infusion rates, P_pv increased 0.38 ± 0.26 mmHg, suggesting a portal venule increase in resistance. No other variables were significantly changed. Because of the cumulative effects of ET-1 and because ET-1 was infused at the end of the experiment, the data in Table 1 and Figs. 4 and 5 are expressed as the change from the initial 2-min control period before any ET-1 administration.

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Cardiovascular responses to moderate infusion rate of endothelin (ET)-1 infused via the portal vein at 0.011 ± 0.007 µg · min1 · kg body wt1. (Cumulative dose of 0.112 ± 0.010 µg/kg body wt = 0.046 ± 0.004 nmol/kg body wt.) Abbreviations as in Fig. 1. [Most variables do not start at a control value of zero because of the cumulative effects of ET-1. The control value for the first (lowest infusion rate) run was used for all subsequent data.] During the 3- to 5-min periods of infusion, all variables significantly changed from control except Pavc (see Table 1).

View larger version (34K): [in this window] [in a new window]   Fig. 5.   Cardiovascular responses to high infusion rate of ET-1 infused at 0.044 ± 0.004 µg · min1 · kg body wt1. (Cumulative dose of 0.326 ± 0.030 µg/kg body wt = 0.136 ± 0.012 nmol/kg body wt.) Abbreviations as in Fig. 1. (Note change in scale for Rhv, Rspl, Thick, Pµhv and Ppv.) During the 3- to 5-min periods of infusion, all variables significantly changed from control except Pavc (see Table 1).

When ET-1 was infused at a moderate rate of 0.0109 ± 0.0067 µg · min¹ · kg body wt¹ (equal to 0.0044 nmol · min¹ · kg body wt¹) to administer a cumulative dose of 0.046 ± 0.004 nmol/kg body wt, R_spl increased to 36 ± 47% above the control value (recorded before the low infusion rate started), and F_hv decreased 23 ± 22% below control (Table 1). Furthermore, this dose caused a marked increase in R_hv, and thus P_µhv increased 0.84 ± 1.32 mmHg (Fig. 4). P_pv also increased 2.0 ± 1.1 mmHg, suggesting portal venule constriction. Liver lobe THK was not significantly changed compared with the preinfusion control period, even though P_pv and R_hv were increased, but the lobe THK was significantly less than the initial control average. At these moderate infusion rates, ET-1 is a potent vaso- and venoconstrictor, increasing R_hv and R_spl. However, the P_sa decreased and the HR increased slightly (Table 1). The changes in all variables, except P_avc and Grad, were statistically significant (P < 0.001) compared with the initial control.

During the ~25 min between the initial control observation, the low and moderate infusions of ET-1, and the control period before the high infusion rate (see below), a total of 0.112 ± 0.010 µg/kg body wt of ET-1 was infused. The splanchnic vascular resistances had increased: P_sa to P_avc (R_spl), 78%; and P_µhv to P_avc (R_hv), 321%. This caused hepatic distending pressures to increase: P_pv to 2.7 mmHg and P_µhv to 1.7 mmHg, but the P_sa did not change. The F_hv decreased 36%. All changes, except P_sa, were significantly different from the initial control. By the end of the moderate infusion, the lobe THK had decreased to a significant 4.9% of initial control average but only an insignificant 2.2% compared with the control period for the moderate infusion.

When ET-1 was infused at 0.0435 ± 0.0042 µg · min¹ · kg body wt¹ (equal to 4.4 pmol · min¹ · kg body wt¹) to administer a relatively high cumulative dose averaging 0.136 ± 0.012 nmol/kg body wt, R_hv, R_spl, P_µhv, and P_pv increased markedly (Fig. 5). (Please note the changes in scale for the resistances, lobe THK, P_µhv, and P_pv and also the residual changes in most variables at the start of the high ET-1 run, which had carried over from the previous infusions of ET-1.) During the 5 min of ET infusion, P_sa progressively increased, whereas F_hv decreased even more from the value at preinfusion control. The hepatic lobe THK did not significantly change compared with the preinfusion control period, but by 3 min of ET-1 infusion, the lobe THK started a progressive increase in parallel with the increase in P_pv and P_µhv (Fig. 5).

By 3 min after the end of the high infusion rate (total dose 0.329 ± 0.030 µg/kg body wt), the resistances were much higher compared with the initial control: R_spl, 158%; R_µpv, 514%; and R_hv, 678%. This caused the hepatic pressures to increase: P_pv to 6.9 mmHg, P_µhv to 3.2 mmHg, and P_sa to 8.2 mmHg. The F_hv was only 45% of the initial control. The lobe THK was then only 0.5% less than the initial control (Fig. 5). The small decrease in P_avc, although not significant at 3-5 min, progressively and significantly decreased after the infusion was stopped, suggesting blood sequestration. From these data, we conclude that the net influence on the vascular capacitance vessels is minor, but with the large increases in distending pressure, the contractile elements must have been activated by the infusion of ET-1 to resist a volume change. The resistances both upstream (R_µpv) and downstream (R_hv) from the hepatic venules markedly increased, thus confirming that ET-1 acts in a similar manner on venous as well as arterial and presinusoidal vessels. The onset of response to the infusion of ET-1 was slower than that of the other vasoactive agents, and the full response was often not evident until several minutes after the infusion was stopped. The recovery was slower. The variability between experiments was large (Table 1).

A striking observation in some experiments using ET-1 was the stopping of erythrocyte movement in the sinusoids even at low drug-infusion rates. We made a crude visual estimation of erythrocyte velocity compared with the pre-ET velocities by assigning one of six categories: 100% = no change; 90% = just-detectable decrease; 75% = clear decrease; 50% = about one-half; 20% = very sluggish; and 5% = no movement, or back and forth. These data, at the corresponding dose, were fitted with an exponential curve starting at 100% of control and decreasing to an asymptote of zero (Fig. 6). The 50% velocity reduction (estimated by interpolation) occurred at a cumulative dose of 0.16 µg/kg body wt. (The moderate cumulative dose averaged 0.112 ± 0.010 µg/kg body wt, and the high dose averaged 0.328 ± 0.012 µg/kg body wt.) The same function was fitted to the F_hv data. A 50% reduction in F_hv occurred at 0.37 µg/kg body wt. In the last three experiments, even higher doses were administered, but these three rabbits were relatively resistant to further reductions in hepatic blood flow.

View larger version (21K): [in this window] [in a new window]   Fig. 6.   Red blood cell (RBC) velocity (A) and Fhv (B) in the microvessels at progressively increasing doses of ET-1 expressed as µg/kg body wt of cumulative dose (the high dose averaged 0.33 ± 0.03 µg/kg body wt). Semilogarithmic curves were least-squares fitted with an initial value of 100 and asymptote of zero. For RBC velocity: 31 observations, 50% response at 0.16 µg/kg body wt, rate constant = 4.32, and R2 = 0.721. For Fhv: 36 observations, 50% response at 0.37 µg/kg body wt, rate constant = 2.01, and R2 = 0.867. Linear regression (lines not shown) for RBC velocity: zero intercept = 55.8%, and slope = 33.4% · µg1 · kg (P = 0.02, r2 = 0.159). Linear regression for Fhv: zero intercept = 74.6%, and slope = 45.6% · µg1 · kg (P = 0.1, r2 = 0.097).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All four vasoactive agents used cause a significant increase in R_hv as well as R_spl (Table 1). NE and ANG II caused an active reduction in vascular capacitance by reducing liver volume (lobe THK) in conjunction with increased distending pressures (P_µhv and P_pv). VP caused the R_spl to more than double, thereby decreasing blood flow and, consequently, P_µhv and P_pv, but lobe THK was not significantly changed (Table 1). ET-1 also increased all resistances and decreased flow, but P_µhv and P_pv increased. The lack of change in lobe THK could either be from a lack of response by the contractile elements of the venous vessels or from a balance between a reduction in volume by wall contraction and an increase in distending pressure by a downstream increase in resistance to flow. [A 10% change in volume of a segment of the vasculature with a relatively large cross-sectional area (e.g., 20-µm venules with a cross section of 37,000 cm² and composing 25% of the total volume; see Table 2 of Ref. 30) will have a much smaller net effect on resistance than a 10% change in volume of larger vessels.]

There were no significant changes in the fraction of the P_µhv-to-P_avc gradient with respect to the total P_pv-to-P_avc gradient (Grad; Table 1), suggesting that the pre- and postsinusoidal vessels were similarly sensitive to each of the agents. As a corollary, the R_spl increased, as did the R_hv. The resistance across the portal venules and sinusoids [(P_pv  P_µhv)/F_hv] must also have increased, because P_pv increased more than P_µhv, with the assumption that the hepatic arterial flow was negligible or changed in proportion to the change in F_hv.

Baroreceptor control of HR was functioning in our rabbits. When the mean P_sa increased, suggesting incomplete extraction from the blood by the liver of the vasoactive agents infused, the HR consequently decreased. However, the highly significant increase in P_sa by NE did not lead to a significant change in HR. During the moderate infusion rate of ET, P_sa decreased and HR increased.

NE. As expected (10, 11, 34, 35), NE caused an active reduction in liver size and thus can activate the liver to be a rapid and effective blood volume reservoir. The hepatic vascular capacitance response to NE was clearly active, because the volume was decreased even though the distending pressures (P_µhv and P_pv) and F_hv increased. The infusion rate via the portal vein of NE (4.7 g · min¹ · kg body wt¹) was high, but our goal was to define clearly the characteristics of the response. In both our earlier study (34) and this one, the F_hv during the first minute of infusion of NE decreased but then increased above control as the P_sa increased. Using an isolated, autologous blood-perfused dog liver preparation, Shibamoto et al. (35) reported that a bolus injection of NE into the circulating perfusate decreased the liver weight by >20%, even though the P_pv and sinusoidal pressure increased. (The concentration of NE in the perfusate was calculated to be 4.3 µM/l, an even higher concentration than we used; see Table 1.) Sinusoidal pressures were estimated via a triple-occlusion method and at control averaged 5.2 mmHg, with the postsinusoidal resistance making up 54% of the total P_pv-to-P_avc gradient. That value is similar to the 53% that we found for the faction that P_µhv composed of the total gradient at control (Table 1). We note that Greenway et al. (10) concluded on p. H992 that the "sympathetic control of the hepatic venous bed is mediated through the hepatic innervation, and circulating catecholamines play at most a minor role."

ANG II. The decrease in lobe THK in conjunction with significant increases in P_µhv and P_pv (Fig. 2) suggests that ANG II can induce an active constriction of the vasculature of the liver. Greenway and Lautt (11) reported an ~42% reduction in liver volume during infusion of ANG II at 0.5 µg · min¹ · kg body wt¹ into femoral veins of cats. The intestinal vascular conductance was reduced to ~25% of control (11). The P_pv was increased only ~2 mmHg. Bulkley et al. (5) have concluded that ANG II acts proportionally more on the R_spl than on other arteries and therefore is the agent leading to serious gastric ischemia in cardiogenic shock. Pang and Tabrizchi (29), using conscious rats, reported that ANG II increases the mean circulatory filling pressure. Our data suggest that the liver participates in the process.

VP. VP caused such a large increase in splanchnic and other arterial resistances that even though P_sa increased, the F_hv, P_pv, and P_µhv decreased. We assume that the hepatic arteries were stimulated to a similar degree as the intestinal and stomach arteries. Greenway and Lautt (11) reported only an ~10% reduction in liver volume during infusion of VP at 10 mU · mg¹ · kg¹. (With the assumption of a potency of 367 U/mg, their infusion rate was about one-half of what we used.) In our study, there was no significant change in lobe THK even though P_hv decreased 0.6 mmHg. The intestinal vascular conductance reduction to ~25% of control is similar to our 350% of control R_spl.

ET. ET influenced the presinusoidal portal venules of the liver, in that both the moderate and high infusion rates increased P_pv and the portal venular and splanchnic resistances and decreased F_hv. In addition, however, the P_µhv and the R_hv were also increased, and so some of the increase in P_pv was from the increase in P_µhv. Although the trend during the high ET infusion rates was a decrease in flow and erythrocyte velocity (Fig. 6), the between-animal variability was high for both variables, in part because of the uncertainty of our measurement technique. ET can apparently influence all contractile elements of the liver and not just the presinusoidal resistance.

Limitations of methods used. Limitations of our experimental approach to measure hepatic microvascular pressure have been, in part, described in previous reports (26, 34). Because the sinusoids contain most of the blood in the liver, we would have preferred to measure changes in sinusoidal pressure, but these ~10-µm-diameter vessels are too small for long-term, reliable pressure measurements. Use of an isolated in vitro liver preparation would have simplified the problem of controlling movement and interpreting the results, which occurred with recirculation of drug. However, isolated, artificially perfused preparations are much less representative of the live animal because of inadequate oxygenation, missing nutrients, absence of innervation, release of toxic materials or accumulation of waste materials, and possible abnormal osmotic and oncotic environment (2, 13).

Conclusion. In summary, we conclude that ANG II and NE induced active constriction of the hepatic capacitance vessels, because the liver lobe THK decreased significantly even though P_µhv and P_pv increased in these anesthetized rabbits with autoperfused livers. All four agents increased R_spl and R_hv in similar proportions. During the infusion of VP and ET-1, the R_spl increased so much that total hepatic flow decreased. With VP, P_µhv and P_pv decreased, but with ET-1, P_µhv and P_pv increased. However, the lobe THK was not significantly changed by either drug during the infusion compared with the 2-min control period. We thus suggest that VP and ET-1 have only a minor net effect on hepatic capacitance vessels; however, with ET-1, some contractile element activation must have occurred to counter the influence of the marked increase in hepatic microvessel distending pressures. ET-1, at 0.04 µg · min¹ · kg body wt¹, caused a 13% increase in P_sa, but the increase in hepatic portal venule resistance was so great that the erythrocyte movement through the sinusoids in some animals stopped. The hepatic microvascular responses to ET are highly heterogeneous.