INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

HEMORRHAGIC SHOCK IS FREQUENTLY associated with a poor prognosis, even if adequate volume resuscitation results in a restoration of systemic hemodynamics. For example, a recently published prospective study (17) in patients with hemorrhagic shock revealed an overall mortality of >50%. This high mortality rate seems primarily to be due to the subsequent development of multiple organ failure (39). Among those patients that develop multiple organ failure, the incidence of liver failure has been reported to exceed 60% (39). Moreover, based on the central role the liver plays in the organism's metabolic and immunological response to injury (37), the occurrence of liver failure is frequently associated with a further deterioration of the prognosis (19).

This raises the question of which pathophysiological mechanisms may be responsible for the development of liver injury after hemorrhagic shock and resuscitation (HSR). Although various factors have been implicated, accumulating evidence suggests that the occurrence of hepatic microcirculatory failure may be one major determinant. For example, Wang et al. (47) described a progressive impairment of hepatic microvascular blood flow after hemorrhagic shock despite adequate fluid resuscitation. In addition, physical prevention of microvascular shutdown using a flow-controlled reperfusion mode largely prevents parenchymal cell necrosis after ischemia-reperfusion of the liver, i.e., the degree of microcirculatory failure determines the extent of lethal hepatocyte injury (10). Moreover, the occurrence of sinusoidal perfusion failure after HSR is associated with deteriorations of the hepatic mitochondrial redox state and bile flow, enzyme release from the liver, and hepatocyte necrosis (2, 36).

It is of particular interest to note that recent evidence from our and other laboratories suggests that endothelins (ETs), a family of isopeptides that can exert potent and long-acting vasoconstriction (49), may play a crucial role in the dysregulation of hepatic microvascular perfusion (12). Exogenous ETs can cause a sustained increase in total portal resistance and a decrease in portal flow in the normal liver (14, 51). A video microscopic study (8) of the hepatic microcirculation has revealed that these vasoconstrictive effects of ETs involve both extrasinusoidal and sinusoidal sites. The pattern of microcirculatory disturbances caused by exogenously administered ETs closely resemble the pattern of changes observed in the liver under pathological conditions such as endotoxemia (30, 35) or ischemia followed by reperfusion (21). Moreover, under these conditions, ET levels are increased, and administration of ET antiserum or pharmacological blockade of ET receptors reduces microvascular injury in the liver (16, 32, 33, 48).

However, it remains unclear how an initial insult such as compensated and reversible HSR may lead to the subsequent development of progressive hepatic microcirculatory and organ failure. Evidence suggests that subtle changes occur in cells or even organs after exposure to a primary minor stimulus that produces hardly detectable injury in itself, e.g., short periods of ischemia, but that may aggravate the response to a secondary insult (13, 22). Consequently, in addition to increased generation, an enhanced response to ETs could contribute to this phenomenon, as previously shown in other models of hepatic stress such as chronic ethanol consumption or endotoxemia (6, 35). Therefore, we hypothesized that HSR, even if it is not primarily associated with profound impairments of sinusoidal perfusion, may alter the vascular responsiveness of the portohepatic circulation and prime the liver for the deleterious vasoconstrictive effects of ETs.

To test this hypothesis, we used a rat model of in vivo HSR and studied the effect of hemorrhagic shock on the intrinsic portal contractile response to ETs in the isolated perfused liver ex vivo. This methodological approach allows the combined analysis of 1) total portal resistance, 2) flow-dependent and -independent components of portal resistance using multiple-point pressure-flow (P-Q) relationships, and 3) hepatic microcirculation via direct in situ video microscopy without any confounding systemic hemodynamic effects of ETs.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals. Male Sprague-Dawley rats (Charles River; Sulzfeld, Germany), weighing between 300 and 350 g, were used for all experiments. The experimental protocol was approved by the local animal care and use committee, and all animals received humane care according to the criteria outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication 86-23, Revised 1985). Rats were fasted for 6 h before the induction of anesthesia but allowed free access to water. After anesthesia was induced with an intraperitoneal injection of pentobarbital sodium (50 mg/kg body wt), animals were placed supine on a heating pad (38-40°C) to maintain body temperature. After a tail vein was cannulated, an infusion of Ringer solution (10 ml · kg¹ · h¹) was started to compensate for evaporative losses during the surgical preparation. Anesthesia was maintained with supplemental intravenous bolus injections of pentobarbital sodium (5 mg/kg body wt) when indicated by any evidence of spontaneous muscle activity or lessening of the anesthetic plane. A tracheotomy was performed, and animals were allowed to breathe spontaneously. A fluid-filled polyethylene-50 catheter was inserted into the left carotid artery and connected to a pressure transducer (MX 860, Medex Medical; Lancashire, UK), and arterial blood pressure was continuously measured.

Hemorrhagic shock protocol. A nonlethal nonheparinized pressure-controlled model of compensated HSR was used. After surgical instrumentation and a stabilization period of 10 min, animals in shock groups were bled to a mean arterial pressure of 40 mmHg in <5 min. The rate of blood withdrawal was 2 ml/min. Shed blood was collected in syringes containing citrate-phosphate-dextrose solution (0.14 ml/ml of blood) (Sigma; St. Louis, MO). Mean arterial pressure was maintained at 40 ± 4 mmHg by intermittent withdrawal of 0.3- to 0.4-ml aliquots of blood or infusion of respective aliquots of Ringer solution. After 60 min of hemorrhagic hypotension, resuscitation of the animals was performed with 60% of the shed blood withdrawn (reinfused during the first 20 min of resuscitation), and twice the maximal bleedout volume was given as Ringer solution during the first hour of resuscitation, i.e., 200% of the shed blood volume. During the second hour of resuscitation, the infusion rate of the Ringer solution was lowered to a volume equaling the maximal bleedout volume, i.e., 100% of the shed blood volume. Thus the total volume of Ringer solution administered during the first 2 h of resuscitation equals 300% of the shed blood withdrawn and 500% of the amount of shed blood reinfused. For the remaining experimental period, the infusion rate of Ringer solution was kept constant at the basal rate of 10 ml · kg¹ · h¹. This resuscitation regimen ensures complete recovery and maintenance of mean arterial pressure until the end of the experiment. Time-matched sham shock animals were anesthetized, completely instrumented as described above, and received a constant infusion of Ringer solution of 10 ml · kg¹ · h¹ during the whole observation period but did not undergo hemorrhage (sham shock groups).

Isolated perfused liver system. A pressure-limited recirculating isolated liver perfusion system was employed as previously described (10, 34). Briefly, the perfusate [5% rat red blood cells (RBC) in Krebs-Henseleit bicarbonate buffer (KHB); pH 7.4] was pumped from an outflow reservoir through a Silastic tubing oxygenator (gassed with a mixture of 95% O₂-5% CO₂) into an overflow chamber that served as the inflow reservoir. A heat exchanger was used to warm the perfusate to 37°C. An ultrasonic in-line flow probe (Transonic; Ithaca, NY) was placed ahead of the inlet cannula and connected to a flowmeter (T-206, Transonic) to measure the total flow rate (Q_t). Portal inlet pressure (P_inlet) was measured via a pressure transducer (Transpac Transducer, Abbott; Wiesbaden, Germany) connected to a T-fitting in the inlet cannula. The outflow cannula drained into the outflow reservoir. Another pressure transducer was connected to a second T-fitting in the vena caval outlet cannula to measure outlet pressure (P_outlet). Both pressure transducers were calibrated simultaneously and zeroed against a column of fluid open to the atmosphere at the level of the cannula tips. The two cannulas were positioned at the level of the abdominal posterior vena cava. Signals were recorded using the Windaq 200 personal computer-based data-acquisition system (Dataq Instruments; Akron, OH). Total portal resistance (R_t) was calculated from (P_inlet  P_outlet)/Q_t.

Experimental protocol. On the basis of the time course of shock-induced hepatic expression of vasoactive mediators (3, 5), baseline measurements and all subsequent pharmacological interventions were performed at 6 h after shock induction (i.e., 1-h shock, 5-h resuscitation; shock groups) or in time-matched sham shock experiments (1-h sham shock, 5-h sham resuscitation; sham shock groups). At 30 min before baseline, a laparotomy was performed, and livers were isolated and perfused in situ via the portal vein essentially as previously described (10, 34). The initial flow rate was set to 30 ml/min, and the perfusate hematocrit was adjusted to ~5% by addition of rat RBC prepared as described above. Fluorescein isothiocyanate-labeled bovine serum albumin was added to the perfusate for assessment of sinusoidal boundaries and visualization of RBC in the sinusoids via negative contrast (26). Subsequently, the flow rate was reduced to 20 ml/min, and, after a 15-min stabilization period, all baseline measurements were obtained.

9

9

9

Epifluorescence liver microscopy. Videomicroscopy of the liver was performed essentially as reported in detail previously (10, 34). Briefly, the liver preparation was positioned on the stage of a Zeiss Axiotech fluorescence microscope (Axiotech Vario 100 HD; Carl Zeiss; Jena, Germany) and viewed with a ×40 water-immersion objective (Zeiss Achroplan, Carl Zeiss). The surface of the left liver lobe was epi-illuminated with a fiber-optic illuminator (KL 1500, Schott Glaswerke; Wiesbaden, Germany) similar to the oblique transillumination procedure described by MacPhee et al. (24), and a randomly chosen acinus was brought into focus so that sinusoids of zone 3 could be observed. This region of the acinus was chosen because the parallel arrangement of sinusoids permits unambiguous quantitative assessment of the sinusoidal microcirculation (11). With the same area in focus, the liver was epi-illuminated with a 100-W mercury lamp with 450-490 nm of excitation and 520-nm emission band-pass filters, which allows visualization of fluorescein isothiocyanate-conjugated albumin within hepatic sinusoids. All microscopic images were projected onto a charge-coupled device video camera (FK 6990 IQ-S, Pieper; Schwerte, Germany). On-line digital contrast enhancement was performed on all images for improved clarity using an Argus-20 image processor (Hamamatsu Photonics; Hamamatsu, Japan). Processed images were recorded for off-line analysis using a S-VHS video recorder (Panasonic AG 7350E, Matsushita Electrical Industrial; Osaka, Japan).

Hepatic oxygen delivery and consumption. To obtain estimates of total hepatic oxygen delivery (DO₂) and consumption (VO₂), 200 µl of perfusate were simultaneously withdrawn from the inflow and outflow tubing of the isolated perfused liver system into gas-tight syringes. Subsequently, hemoglobin concentration, oxygen saturation, and oxygen partial pressure were determined within each sample using an automated blood gas analysis system (ABL 625, Radiometer Medical A/S; Copenhagen, Denmark), which had been calibrated for rat blood using an OSM 3 hemoximeter (Radiometer Medical A/S). Estimates of hepatic DO₂ and VO₂ were derived using standard and previously described equations (31).

Data analysis. Data are presented as means ± SE. Raw data at baseline are provided in Table 1. Statistical differences from baseline were determined using a paired t-test. Differences between shock and sham shock groups were tested using an unpaired t-test within each experimental series. When criteria for parametric tests were not met (Kolomogorov-Smirnow test for normality and/or Levene-Mediane test for equal variance failed), the respective nonparametric tests (i.e., Wilcoxon signed-rank test or Mann-Whitney rank sum test) were used. All individual P-Q relationships were analyzed by linear regression to obtain the regression slope. A P value <0.05 was considered to indicate a significant difference. All statistical tests were performed using the SigmaStat software package (Jandel Scientific; San Rafael, CA).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Hepatic macrohemodynamics, microhemodynamics, and oxygenation after HSR. The model of compensated hemorrhagic shock used in the present study did not cause major disturbances of portal venous hemodynamics in the isolated perfused liver after 5 h of resuscitation. This is illustrated by the experimental data summarized in Table 1. At this time point, all macro- and microhemodynamic parameters measured, as well as hepatic DO₂ and VO₂, did not significantly differ between the shock and sham shock groups.

Effect of HSR on the change in total portal vein resistance in response to ET-1 and S6c. Whereas vehicle did not cause any major changes in the macrohemodynamic parameters measured, administration of ET-1 increased P_inlet  P_outlet (Fig. 1), caused a decrease in Q_t (Fig. 2), and caused a respective increase in R_t (Fig. 3) in all experimental groups. However, the resistive response to ET-1 was more pronounced in livers from animals that underwent HSR compared with livers from sham shock animals (Figs. 1-3). Addition of the preferential ET_B receptor agonist S6c to the perfusate also resulted in increases in P_inlet  P_outlet and R_t (Figs. 1 and 3) and decreases in Q_t compared with baseline (Fig. 2). Despite this similarity, livers from shock animals showed different behaviors in response to the two vasoconstrictive agents. In contrast with the enhancing effect of HSR on the portal contractile response to ET-1, the macrohemodynamic changes induced by S6c were either unaffected (Q_t and R_t; Figs. 2 and 3) or even attenuated (P_inlet  P_outlet; Fig. 1) after HSR compared with the respective sham control group.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Effect of hemorrhagic shock on changes in portal driving pressure (inlet-outlet pressure) in isolated perfused rat livers in response to either vehicle (1 ml H2O), endothelin-1 (ET-1; 0.5 × 109 M), a preferential ETA receptor agonist, or sarafotoxin 6c (S6c; 0.25 × 109 M), a selective ETB receptor agonist. Posttreatment measurements were performed 10 min after pharmacological intervention. Livers were obtained from rats subjected to hemorrhagic shock (1 h; mean arterial pressure: 40 ± 4 mmHg) and volume resuscitation (5 h; shock groups) or from time-matched sham shock animals. Data represent means ± SE; n = 5-9 experiments/group. *P < 0.05 vs. respective baseline value (within each group); #P < 0.05 vs. respective sham shock group (within each series).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Effect of hemorrhagic shock on changes in total portal flow rate in isolated perfused rat livers in response to either vehicle (1 ml H2O), ET-1 (0.5 × 109 M), or S6c (0.25 × 109 M). Posttreatment measurements were performed 10 min after pharmacological intervention. Livers were obtained from rats subjected to hemorrhagic shock and volume resuscitation (shock groups) or from time-matched sham shock animals. Data represent means ± SE; n = 5-9 experiments/group. *P < 0.05 vs. respective baseline value (within each group); #P < 0.05 vs. respective sham shock group (within each series).

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Effect of hemorrhagic shock on changes in total portal vascular resistance in isolated perfused rat livers in response to either vehicle (1 ml H2O), ET-1 (0.5 × 109 M), or S6c (0.25 × 109 M). Posttreatment measurements were performed 10 min after pharmacological intervention. Livers were obtained from rats subjected to hemorrhagic shock and volume resuscitation (shock groups) or from time-matched sham shock animals. Data represent means ± SE; n = 5-9 experiments/group. *P < 0.05 vs. respective baseline value (within each group); #P < 0.05 vs. respective sham shock group (within each series).

Effect of HSR on ET-1- and S6c-induced changes in portal vein P-Q relationships. The administration of vehicle to the perfusate did not have a significant effect on P_Q=0 (Fig. 4) or the slope_PQR in livers from sham shock controls or shock animals (Fig. 5). In contrast, both vasoactive agents caused increases in the flow-independent (Fig. 4) and flow-dependent components (Fig. 5) of R_t in sham control and shock groups. HSR did not alter the effect of ET-1 or S6c on the slope_PQR (Fig. 5). However, the changes in P_Q=0 were differentially affected after shock (Fig. 4). Whereas HSR did not influence the magnitude of the increase in P_Q=0 after S6c, the increase in P_Q=0 in response to ET-1 was much more pronounced after shock compared with sham shock.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Effect of hemorrhagic shock on changes in zero-flow portal inlet pressure (PQ=0) in isolated perfused rat livers in response to either vehicle (1 ml H2O), ET-1 (0.5 × 109 M), or S6c (0.25 × 109 M). Posttreatment measurements were performed 10 min after pharmacological intervention. The inlet pressure was measured under steady-state conditions while the outlet pressure was kept constant at a value of 1 cmH2O. Livers were obtained from rats subjected to hemorrhagic shock and volume resuscitation (shock groups) or from time-matched sham shock animals. Data represent means ± SE; n = 5-9 experiments/group. *P < 0.05 vs. respective baseline value (within each group); #P < 0.05 vs. respective sham shock group (within each series).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Effect of hemorrhagic shock on changes in the calculated regression slope of multiple-point pressure-flow relationships (slopePQR) in isolated perfused rat livers in response to either vehicle (1 ml H2O), ET-1 (0.5 × 109 M), or S6c (0.25 × 109 M). Posttreatment measurements were performed 10 min after pharmacological intervention. The inlet pressure was measured under steady-state conditions at multiple levels of portal flow (0, 2.5, 5, 10, 15, 20, and 25 ml/min) while the outlet pressure was kept constant at a value of 1 cmH2O. Slopes of individual pressure-flow relationships of each organ at each pre- and posttreatment time point were calculated over the entire range of flow. Livers were obtained from rats subjected to hemorrhagic shock and volume resuscitation (shock groups) or from time-matched sham shock animals. Data represent means ± SE; n = 5-9 experiments/group. *P < 0.05 vs. respective baseline value (within each group).

Effect of HSR on the hepatic microcirculatory response to ET-1 and S6c. Direct microscopic observation of the hepatic microcirculation within the intact organ revealed a significant narrowing of sinusoids in response to ET-1 that could not be observed in the vehicle-treated groups (Table 2 and Fig. 6). The magnitude of sinusoidal narrowing after ET-1 was much more pronounced in livers from animals that underwent HSR compared with those that had been obtained from sham shock animals (Fig. 6). Representative high-power video micrographs illustrating the enhanced ET-1-induced decrease in sinusoidal dimensions that could be observed after HSR are depicted in Fig. 7. Whereas S6c did also cause a reduction in D_s, its microcirculatory effects were not altered by HSR, i.e., the decrease in D_s was similar in the sham and the shock group (Table 2 and Fig. 6). Whereas V_RBC remained unchanged in the two vehicle groups, administration of ET-1 caused a decrease in V_RBC in the sham control group that could not be observed in the shock group (106 ± 37 vs. 6 ± 40 µm/s, P < 0.05). In contrast, addition of S6c to the perfusate resulted in an increase in V_RBC in livers from sham shock animals (169 ± 48 µm/s, P < 0.05) but did not cause any significant changes in V_RBC in livers from shock animals.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Effect of hemorrhagic shock on changes in sinusoid diameter in isolated perfused rat livers in response to either vehicle (1 ml H2O), ET-1 (0.5 × 109 M), or S6c (0.25 × 109 M). Posttreatment measurements were performed 10 min after pharmacological intervention. Sinusoid diameters were determined using in situ epifluorescence videomicroscopy. Pre- and posttreatment measurements were performed within the same segments of the same sinusoids. Livers were obtained from rats subjected to hemorrhagic shock and volume resuscitation (shock groups) or from time-matched sham shock animals. Data represent means ± SE; n = 5-9 experiments/group. *P < 0.05 vs. respective baseline value (within each group); #P < 0.05 vs. respective sham shock group (within each series).

View larger version (145K): [in this window] [in a new window]   Fig. 7.   Representative high-power in situ video micrographs of sinusoids as viewed by the oblique transillumination technique at baseline (A and C) and 10 min after the administration of ET-1 (0.5 × 109 M; B and D). The liver depicted in A and B was obtained from a sham shock animal, and the liver shown in C and D was from an animal that underwent hemorrhagic shock. Note that ET-1 causes only a moderate decrease in sinusoid diameters after sham shock (B), whereas the sinusoidal narrowing observed in response to ET-1 is much more pronounced after HSR (D).

Influence of HSR on the changes in hepatic oxygenation after ET-1 and S6c. Estimates of total hepatic DO₂ and VO₂ remained unchanged in the two vehicle groups (Fig. 8, A and B). In contrast, both ET-1 and S6c caused substantial reductions in DO₂ (Fig. 8A). Whereas the negative effect of S6c on DO₂ tended to be attenuated after HSR, the impairment of the hepatic oxygen supply in response to ET-1 was more pronounced under these conditions compared with respective sham shock controls (Fig. 8A). Although the changes in VO₂ did not reach statistical significance, there was a tendency toward a lower VO₂ in particular after ET-1 administration in the shock group, whereas VO₂ tended to increase on S6c administration after shock (Fig. 8B).

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Effect of hemorrhagic shock on changes in estimates of total hepatic oxygen delivery (A) and consumption (B) in isolated perfused rat livers in response to either vehicle (1 ml H2O), ET-1 (0.5 × 109 M), or S6c (0.25 × 109 M). Hemoglobin concentration, oxygen saturation, and oxygen partial pressure were determined within perfusate samples withdrawn from the inflow and outflow tubing to calculate estimates of oxygen delivery and consumption. Posttreatment measurements were performed 10 min after pharmacological intervention. Livers were obtained from rats subjected to hemorrhagic shock and volume resuscitation (shock groups) or from time-matched sham shock animals. Data represent means ± SE; n = 5-9 experiments/group. *P < 0.05 vs. respective baseline value (within each group); #P < 0.05 vs. respective sham shock group (within each series).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Accumulating evidence suggests that even a subtle initial insult that does not cause overt disturbances in organ perfusion, function, or integrity may nevertheless significantly alter the responsiveness to a secondary noxious stimulus, thus leading to the subsequent development of organ dysfunction or failure (13, 23). However, the pathophysiological mechanisms responsible for this "two-hit" phenomenon remain to be identified. Maintenance of the integrity of the microcirculation is crucial to prevent liver injury (10). Therefore, in addition to changes in hepatic metabolic activity, energy status, or immune function, alterations in the control of nutritive microvascular blood flow could provide a mechanistical basis for the negative priming effect exerted by hemorrhagic shock (28). In this regard, it is of particular interest to note, as a recently published study by Garrison et al. (15) showed, that HSR can substantially enhance the microvascular responsiveness of the small intestinal circulation to bacterial endotoxin. Thus it was the aim of the present study to test the hypothesis that hemorrhagic shock, even if it is not primarily associated with major impairments of sinusoidal perfusion, may prime the portohepatic circulation for the vasoconstrictive effects of ETs. The combined analysis of portal macro- and microhemodynamics revealed that HSR enhances the portal vasoconstrictive response to ET-1 and that this increase in responsiveness occurs predominantly within the sinusoidal compartment. These conclusions are derived from the following observations.

Administration of ET-1 to livers from sham shock animals increased R_t and caused a respective decrease in portal flow. This is in agreement with previous results from our and other laboratories (8, 45) and confirms that ET-1 acts as a potent vasoconstrictor within the portal venous circulation of the liver. Although HSR in itself did not cause major disturbances of hepatic perfusion under these experimental conditions, the responsiveness to ET-1 was substantially enhanced, i.e., the increase in P_inlet  P_outlet and R_t as well as the decrease in Q_t after ET-1 was much more pronounced in livers from animals that underwent HSR compared with livers from sham shock animals. A similar increase in the portohepatic contractile response to ETs has also been described after chronic ethanol consumption (6) or endotoxemia (35). It is important to note that these are very different pathological entities. For example, whereas endotoxemia causes a generalized inflammatory response, inflammation is a delayed and much less pronounced event after HSR. Therefore, the similar increase in the contractile response to ET-1 observed under these different conditions may be considered to reflect a rather uniform general remodeling of the hepatic hemodynamic responsiveness that occurs after diverse pathological stimuli.

Identification of the sites of altered resistance regulation could contribute to the characterization of the mechanisms responsible for the HSR-mediated increase in the portohepatic contractile response to ET-1. Thus we combined the analysis of portal venous P-Q relationships with the direct observation of the sinusoidal microcirculation using epifluorescence video microscopy. Previous studies (1, 27) have provided substantial indirect evidence that changes in the flow-independent component of portal venous resistance, i.e., P_Q=0, primarily reflect changes in resistance that occur in a downstream sinusoidal or postsinusoidal compartment. On the other hand, changes in the flow-dependent component of portal resistance, i.e., of the slope_PQR, reflect changes in resistance that occur upstream from the site of the P_Q=0 in a vessel where a positive P_Q=0 higher than the actual P_outlet exists (9, 27). Therefore, the increase in P_Q=0 and slope_PQR, associated with a respective decrease in D_s and V_RBC that could be observed in livers from sham shock animals in response to ET-1, strongly supports the results of previous studies (8, 50) in the normal liver: that ET-1 acts at both sinusoidal and extrasinusoidal sites. It is of great interest to note that the changes in the components of R_t after ET-1 were differentially affected by HSR. Whereas the ET-1-induced increase in P_Q=0 was much greater after HSR compared with sham, the increase in slope_PQR was comparable in both groups. On the basis of the assumptions summarized above, these data would suggest that the enhancement of the contractile response to ET-1 after HSR occurred primarily within the sinusoidal compartment of the liver. This is further supported by the fact that in situ microscopic measurements of D_s did indeed reveal that the sinusoidal narrowing caused by ET-1 was much more pronounced after HSR compared with sham while, at the same time, the reduction in V_RBC was attenuated. This raises the question of which factors could be responsible for the fact that manifestation of the enhanced responsiveness to ET-1 after HSR occurred primarily within the sinusoidal compartment.

On one hand, this could be the result of a HSR-induced increase in the contractility of hepatic stellate cells (HSC) (46). These nonparenchymal liver cells are localized within the space of Dissé and have been suggested to act as liver-specific pericytes with vasoregulatory properties at the level of the sinusoids (18, 38). For example, HSC can actively contract in response to ET-1 in vitro as well as in the intact liver in situ (20, 50). Moreover, activation of HSC has been shown to be associated with a specific increase in the sensitivity to ET-1 (40, 43). Therefore, the results of the present study would be internally consistent with the notion that the enhancement of the ET-1-mediated sinusoidal narrowing after HSR may result from a hypercontractile state of HSC. On the other hand, McCuskey (25) recently suggested that diameters of sinusoids might also decrease due to passive recoil when inflow is reduced via an increase in presinusoidal resistance and intrasinusoidal distending pressure falls. Thus the enhanced sinusoidal narrowing in response to ET-1 after HSR could also reflect a hypercontractility of smooth muscle cells located within the terminal portal venules. However, portal venular constriction, e.g., in response to the ₁-adrenoreceptor agonist phenylephrine, has been shown to be associated with an isolated decrease in slope_PQR in the absence of any significant changes in P_Q=0 (35). Because this was not the case under the experimental conditions of the present study, the sum of these data favor a predominant role of the sinusoidal compartment in the HSR-induced enhancement of the vasoconstrictive response to ET-1.

The biological effects of ETs are mediated through different receptors that are heterogeneously expressed among liver cells. This raises the question of which receptors could be involved in the sensitizing process after HSR. Whereas HSC and hepatocytes express ET_A and ET_B receptors, only the ET_B receptor can be found on Kupffer and endothelial cells (18). Both ET_A and ET_B receptor agonists may cause contraction of the respective hepatic target cells (42, 50). However, the relative contribution of the two receptors may largely vary depending on the actual experimental or pathological conditions. Thus we also studied the effect of HSR on the contractile response to the selective ET_B receptor agonist S6c. In sharp contrast with the enhancing effect of HSR on the response to ET-1, the portohepatic hemodynamic effects of S6c were unchanged or even slightly attenuated after shock. In this regard, it is of particular interest to note that Reinher et al. (40) recently proposed that activation of HSC may induce the appearance of a high-affinity ET_A receptor subtype, whereas in quiescent HSC a low-affinity ET_A receptor subtype prevails (40). However, these putative receptor subtypes have not been dissected at the molecular level, and neither ET_A nor ET_B receptor antagonists were used in the present study. Therefore, the sensitizing effect of HSR on the contractile response to ETs cannot be ascribed to a specific ET receptor. In addition, it is important to note that only a single dose of ET-1 and S6c was tested. On the basis of the results of previous dose-response determinations of ET-1 in the portal circulation (7), an ET-1 dose was chosen that caused a half-maximal increase in portal resistance to allow the detection of both increases as well as decreases in the contractile response. Consequently, a dose of S6c was chosen for comparison, which caused an increase in R_t in the normal liver of a similar magnitude as that of ET-1 did. However, based on this experimental design, we cannot exclude the possibility that the differential effects of HSR on the portohepatic contractile response may vary depending on the actual concentration of ET-1 or S6c.

The results of the present study were obtained using an isolated perfused liver system. This experimental approach was chosen to allow the analysis of the intrinsic hepatic hemodynamic response to ETs in the absence of confounding factors such as varying levels of anesthesia, changes in sympathetic outflow, differences in hematocrit, or systemic hemodynamics. Despite these advantages, the use of such a system may be associated with significant alterations of hepatic perfusion. For example, in agreement with previous results, mean D_s were ~0.5-2 µm wider, and D_s showed a larger variability in the isolated perfused liver in vitro compared with the normally perfused liver in vivo (4, 8, 51). Therefore, factors that could contribute to these phenomena, such as an inhomogeneous washout of vasoactive mediators, denervation, or altered intra- and extravascular pressure gradients, may in turn have affected the response to the agonists. Consequently, further in vitro and in vivo studies will be needed to more precisely define the mechanisms of the sensitization of the liver to ET-1 after HSR.

Various pathological conditions that frequently follow upon hemorrhagic shock, such as endotoxemia, hypoxemia, or recurrent periods of hemorrhagic hypotension, have been shown to result in increases in systemic or hepatic levels of ET-1 (29, 41, 44, 48, 52). On the basis of the results of the present study, exposure of the liver to increased amounts of ET-1 after HSR-mediated priming of the portohepatic circulation may cause critical perturbations of nutritive hepatic blood flow. The potential functional significance of these findings is further supported by the fact that the increased responsiveness to ET-1 was associated with respective impairments of hepatic oxygenation.

In conclusion, our results demonstrate that HSR primes the portohepatic circulation for the vasoconstrictive effects of ET-1 via a mechanism that predominantly involves the sinusoidal compartment. These findings could, at least in part, explain why compensated and reversible HSR frequently leads to the subsequent development of progressive hepatic microcirculatory and organ failure in response to a secondary insult.