Endothelin 1 Impairs Oxygen Delivery in Livers From LPS-Primed Animals

Abstract

Endothelin 1 (ET-1) is a potent vasoactive peptide that acts at sinusoidal and extrasinusoidal sites in the liver. Sensitivity to ET-1 increases in LPS-primed animals and is associated with impaired liver microcirculation in these animals. We hypothesized that LPS priming leads to an exacerbation in the impaired oxygen delivery in response to intraportal infusion of ET-1. Rats were studied 24 h after LPS injection (1 mg/kg, i.p.). Surface PO₂ was determined using a recently developed technology of O₂ mapping. The baseline portal pressure was higher in LPS-primed animals (P < 0.05), and increased to similar magnitude as sham animals after a 10-min infusion of ET-1. The resultant portal pressure remained elevated in LPS compared to sham animals. There was no significant difference in baseline mean arterial pressure, and no significant systemic response to ET-1 in either group. In contrast to the macrohemodynamic, the decrease in tissue surface PO₂ in response to ET-1 infusion was potentiated by LPS treatment (increased from baseline levels 33.8 ± 9 to 46.8 ± 8.3 in sham; 42.3 ± 9.1 to 69 ± 6.5 gray scale units in LPS;P < 0.01, sham vs. LPS) at end of infusion of ET-1 for 10 min. This indicates tissue hypoxia in response to ET-1, which is exacerbated in livers from LPS-primed animals compared to sham. Frequency distribution analysis showed a shift in mode from lower intensity (higher PO₂) to areas with higher fluorescent intensity ranges (lower PO₂), indicating areas with shut down in perfusion in LPS-treated animals. In the whole liver, ET-1 suppressed oxygen consumption, and this response was potentiated by LPS pretreatment. We propose that ET-1 impairs oxygen delivery in the liver during endotoxemia, resulting in areas of focal hypoxia. This response is possibly due to potentiated action of ET-1 at both sinusoidal and extrasinusoidal sites in the liver during endotoxemia.

INTRODUCTION

Endotoxemia causes microcirculatory dysfunction that can lead to a dramatic decrease in the sinusoidal blood flow (^1,2). Since a reduction in the sinusoidal flow can be a major determinant of hepatocellular injury (³), the development of microcirculatory dysfunction is likely to contribute to liver failure in septic shock. However, the underlying mechanisms behind microcirculatory dysfunction leading to the hepatic injury remain incompletely understood. Endothelin (ET) is a potent vasoactive peptide that acts at both sinusoidal and extrasinusoidal sites in the liver (^4–10). Addition of exogenous ET-1 in normal animals decreased sinusoidal diameter and sinusoidal flow, and increased total portal vascular resistance in the liver (^4–7). These studies provided evidence for the sinusoidal site of action of ET-1. Previous studies in our laboratory have demonstrated an increase in the levels of ET-1 mRNA in response to intraperitoneal injection of LPS at a dose concentration of 1 mg/kg body wt. (¹¹). Elevated endothelin plasma levels have been reported in patients with septic shock, and increases in systemic and portal ET-1 levels have been reported during experimental endotoxemia (^12,13). Furthermore, endotoxemia increased the sinusoidal response to ET-1 (¹⁴). The sinusoids constricted with a greater magnitude in response to ET-1, which contributed to a dramatic decrease in the sinusoidal flow in LPS-primed animals compared to sham (¹⁴). These experiments provided evidence for endothelin-mediated dysregulation of the hepatic microcirculation during endotoxemia; however, the mechanisms leading to the hepatic injury are still unclear.

Accompanying the changes in the sinusoidal hemodynamics, recent experiments in our laboratory showed heterogeneity of oxygen distribution in response to ET-1 (¹⁵). The results provided evidence for an altered oxygen distribution, possibly due an impaired sinusoidal perfusion as observed previously. A method was developed based on previous attempts (^16,17) to make a direct assessment of PO₂ on the liver surface. This technology utilizes an oxygen-quenched fluorescent dye incorporated in an oxygen-permeable silicone elastomer membrane. The heterogeneity was measured as changes in coefficient of variation of fluorescent intensity, which showed heterogeneity of oxygen distribution at the acinar level in response to ET-1. This suggested an inadequacy in the microcirculatory perfusion that could lead to a mismatch between oxygen supply and metabolic demand, leading to hepatic injury. Since there is an increase in the portal contractile response to ET-1 during endotoxemia, it is unknown if the higher microvascular response leads to an exacerbation of impaired oxygen supply.

Sepsis produces tissue injury via a combination of altered metabolic status and impaired oxygen delivery (^18–20). A diminished oxygen delivery resulting in a mismatch between oxygen supply and metabolic demand (^21,22) would potentiate any direct metabolic defects. Since there is an increase in the microvascular reactivity during endotoxemia, we would predict an exacerbation in tissue hypoxia in the liver in response to ET-1. In the present study we addressed the question whether or not the hypoxia on the liver surface in response to ET-1 is exacerbated during endotoxemia.

MATERIAL AND METHODS

Materials

Human ET-1 was purchased from American Peptide (Sunnyvale, CA). All other chemicals were purchased from Sigma Chemical (St. Louis, MO).

Oxygen-sensitive probe membrane

An oxygen-quenched fluorescent dye, Tris (1,10-phenanthroline) ruthenium (II) chloral hydrate (PRCH; Aldrich Chemical, Milkwaukee, WI) was used for the detection of the PO₂, as described previously (¹⁵). Briefly, the dye was directly dissolved in acetonitrile, 99% (Aldrich) and was then incorporated in an oxygen-permeable silicone elastomer dissolved in ether, 99+% anhydrous A.C.S reagent (Aldrich). Approximately 70 μL of the mixture was placed on a microscope cover slip 22 mm in diameter. The coverslip was centrifuged in order to spread the silicone in a uniform layer. The membrane was fixed to a window made in the center of a plastic 60-mm diameter petri dish, which fit onto the microscope stage. The exteriorized under-surface of the liver was directly placed on the probe membrane with no intervening space in between.

Experimental protocol

The animals were fasted overnight, but were allowed free access to water. Under metafane anesthesia, animals received an intraperitoneal injection of either 0.9% saline (1 mL/kg body wt; sham control group, 4 rats), or Escherichia coli LPS (Difco Laboratories, Detroit, MI), 1 mL/kg body wt (LPS group, 4 rats). The experiments were performed 24 h after injection. A single dose of ET-1 (American Peptide Company) was infused through a syringe infusion pump (Harvard syringe infusion 22, Harvard Instruments, South Natick, MA) at rate of 50 μL/min and a final dose of 60 pmol/min. This dose of ET-1 was chosen since it produced alteration in the oxygen distribution on the liver surface, along with an increase in the portal vascular resistance (¹⁵).

In a separate series of experiments, changes in the total oxygen consumption of the liver in response to ET-1 were determined in sham and LPS-treated animals. The livers were isolated and perfused from sham (n = 8) and LPS-treated animals (n = 7). ET-1 was infused at a rate that produced a final dose of 2.6 nM at the liver through a syringe infusion pump for 10 min.

Animal preparation for in vivo microscopy

The protocol was approved by the University of North Carolina at Charlotte Institutional Animal Care and Use Committee. The experiments were performed in adherence to the National Institutes of Health Guidelines on the use of Laboratory Animals. Male Sprague-Dawley rats (280–340 g, Charles River Laboratories, Wilmington, MA) were fasted for 24 h, but were allowed free access to water. After induction of anesthesia (pentobarbital sodium, 50 mg/kg body wt, i.p.), a tracheotomy was performed, but animals were allowed to breathe room air spontaneously. Rats were placed on a heating pad to maintain body temperature at 36°C to 37°C. The carotid artery was cannulated and the cannula was connected to a pressure transducer to monitor mean arterial pressure. A midline laparotomy was performed, the splenic vein was identified, and the splenic arteries were ligated to prevent splenic congestion. The splenic vein was cannulated with PE 50 catheter and advanced to the portal vein. The catheter was connected to a T tube connector, one end of which was connected to a Digi-Med Low Pressure analyzer for measurement of portal pressure, and the other end was kept for drug infusion. The transducers were connected to a Digi-Med Signal Analyzer (ASA 200) and to a Digi-med Blood Pressure Analyzer. 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 bowel was repositioned and covered with saline-wetted cotton gauze to minimize evaporative loss during the procedure. Animals were turned on their side, and the left liver lobe was gently exteriorized with lower surface uppermost and positioned on the oxygen sensitive membrane. The liver was covered by a piece of plastic wrap to prevent mechanical damage and drying of the surface.

Intravital videomicroscopy

The preparation was then moved to the stage of a Zeiss IM 35 fluorescence microscope (Carl Zeiss, Inc., Thornwood, NY). The surface was epi-illuminated with a 100 W mercury lamp with 490 nm excitation and 600 nm emission band pass filters to view the fluorescent images of oxygen distribution using a 6.3× objective. The emission wavelength used in this study is longer than that used previously (515–565 nm) (¹⁵). This longer wavelength allows more selective visualization of the PRCH dye and decreases the autofluorescence from the tissue. Although this improves the quantitation of tissue oxygen, it does not allow visualization of tissue structure.

Gain and black level settings on the camera were controlled manually to allow recording of changes in fluorescence intensity. At the beginning of each experiment, background subtraction was performed to minimize the effects of uneven illumination. After setting a field of interest involving three to four lobules, a background subtraction and image average (16 frames) was performed to normalize to baseline. This procedure enhances the quality of the images and avoids any artifact because of differences in the baseline fluorescence from one membrane to another. It does, however, artificially minimize the differences in baseline fluorescence. Thus, the procedure most accurately reflects changes from baseline. The images were recorded using a QC-200 intensified charge coupled device camera (Quantex, Sunnyvale, CA) connected to an S-VHS video recorder (Panasonic AG7300) and were analyzed during video playback. The images were recorded at 0, 10, 20, 30, and 40 min.

Off-line computer-assisted video analysis

The recorded images were digitized and analyzed frame by frame with a Image-Pro Plus image analysis system (Media Cybernetics, Silver Spring, MD). The average intensity for a field was measured for each time point. The intensity of each gray scale pixel was assigned a value of 0 to 255, with 0 corresponding to minimum and 255 to maximum fluorescent intensity. The range of 0 to 255 was divided into 25 divisions of 10 units for analysis of frequency distribution of intensity ranges.

Isolated perfused liver system

The nonrecirculating liver perfusion used in these experiments is essentially the same as described previously (²³). Under metafane anesthesia (Schering-Plough Animal Health Corp., Union, NJ), a laprotomy was performed, and the portal vein was cannulated with polyethylene tubing (PE-240, Becton Dickinson, Parsippany, NJ). The liver was constantly perfused at a rate of 10 mL/100 g bw/min with a Krebs-Henseleit bicarbonate buffer (KHB, pH 7.4), which was constantly bubbled with 95% O₂ 5% CO₂, and the temperature was maintained at ∼35°C by water jacketing. The livers were allowed to stabilize for a 20-min period, during which they were perfused with KHB without substrate. Following stabilization, the perfusate was changed to substrate buffer (containing 5 mM of lactate and 1 mM of pyruvate). ET-1 was infused through a syringe infusion pump at rates calculated to produce designated final concentrations in the liver.

Hepatic oxygen consumption

The hepatic VO₂ was calculated from the difference between inflow and outflow oxygen tension, as measured from platinum Clark type O₂ electrodes (Yellow Springs Instruments) situated in perfusion line before and after the liver. The livers were stabilized for a period of 20 min with substrate-free KHB. After altering the contents of the perfusate, values in the calculations were obtained at fixed time intervals for a period of 40 min.

Statistical analysis

Data are means ± SE. Differences between the two groups were tested using an unpaired t test or analysis of variance (ANOVA). Statistical differences from baseline within each group were determined by repeated measures ANOVA followed by post hoc Dunnett's test. When criteria for parametric testing were violated, the appropriate nonparametric (i.e., Friedman repeated measures ANOVA on ranks and Mann-Whitney U test) were used, and P < 0.05 was considered significant.

RESULTS

Systemic and portal hemodynamics: sham vs. LPS-primed animals

We previously demonstrated heterogeneity in oxygen distribution in the liver in response to ET-1 in normal rats (¹⁵). In the current study we determined the changes in the tissue surface oxygen distribution the livers of LPS-primed animals in response to ET-1. The systemic and the portal pressure were measured during and after the end of ET-1 infusion through the portal vein catheter. After 24 h of LPS injection at a dose of 1 mg/kg, i.p., there was no significant difference in the baseline systemic pressure between sham and LPS-treated animals (111.5 mmHg in sham vs. 108.9 mmHg in LPS;Fig. 1). Furthermore, there was no measurable mean arterial pressure response either in sham or LPS-treated animals in response to ET-1. This is in similar to our previous observations suggesting a high first-pass metabolism of ET-1 by the liver when ET-1 is directly infused into the portal vein. In contrast to the systemic pressure, the baseline portal pressure was higher in LPS-primed animals compared to sham animals (4.9 mmHg in sham vs. 7.4 mmHg in LPS-primed animals, P < 0.05), which is consistent with our previous reports of LPS-mediated increase in the portal vascular resistance. In response to ET-1, there was an increase in the portal pressure in both sham and LPS-treated animals. The portal pressure increased progressively, with a peak response at the end of the infusion followed by a gradual decline. However, the magnitude of the response to ET-1 was similar in the two groups. This response is different compared to our previous observations in an isolated, perfused livers, which showed a potentiated response in the LPS-primed animals. The results reflect a more complicated regulation of the hepatic microcirculation in vivo. This observation raises the question of whether the response to ET-1 is potentiated at the microvascular level following LPS treatment in vivo.

Fig. 1: Time course of changes in the mean arterial pressure and portal pressure in response to ET-1 infusion in sham and LPS-primed animals. Rats received either 0.9% saline (1 mL/kg body wt, i.p.; sham group) or LPS (E. coli, 1 mL/kg body wt, i.p.). The experiments were performed 24 h post-injection. ET-1 was injected at a dose concentration of 60 pmol/min for a period of 10 min. (A) Changes in the systemic pressure with ET-1 infusion in either sham or LPS-primed animals. (B) Changes in the portal pressure with ET-1 infusion in either sham or LPS-primed animals. Data are mean ± SE. * P < 0.05, between the groups, and ^# P < 0.05 vs. respective baseline value (within each group for both sham and LPS groups).

Oxygen distribution in sham and LPS-primed animals

In the present experiments the background subtraction procedure obscured any differences in tissue PO₂ that may have existed between sham and LPS animals. Nevertheless, in response to ET-1, the fluorescence intensity increased on the liver surface of both sham and LPS-treated animals, suggesting a decrease in the tissue PO₂ (Fig. 2). In the livers from LPS-treated animals, the magnitude of response was higher in the livers from LPS-treated animals compared to sham animals (Fig. 3). The mean fluorescence intensity (higher intensity = less O₂) increased from baseline levels of 33.9 ± 9 to 46.8 ± 8.3 fluorescence units in sham, whereas the intensity increased from 42.3 ± 9.1 to 69 ± 6.5 fluorescence units in LPS (P < 0.01, sham vs. LPS) at the end of infusion of ET-1 for 10 min (Fig. 3). The pattern of the fluorescence increase observed in the LPS-treated animals was indicative of generalized hypoxia rather than the patchy pattern that we previously reported in normal animals. This most likely reflects the increased responsiveness to exogenous ET-1 in the LPS-treated group. This increase in the fluorescence intensity was sustained for the whole period of observation. To further analyze the degree of hypoxia and to get an estimate of any heterogeneity, the range of 0 to 255 of gray scale pixels was divided into 25 divisions of 10 fluorescence intensity units for analysis of frequency distribution of intensity ranges, as described in Materials and Methods. Frequency distribution analysis showed a shift in mode from lower intensity (higher PO₂) to areas with higher fluorescent intensity ranges (lower PO₂) at the end of infusion with ET-1 (Fig. 4). The shift towards higher intensity ranges was more in LPS-primed animals as compared to sham animals, suggesting areas of higher level of hypoxia in response to ET-1. Most notably, only the LPS-treated animals showed any areas with an intensity level above 80 units, indicating most severe hypoxia.

Fig. 2: Images showing fluorescent intensity changes due to changes in PO₂ in sham and LPS-primed animals. Shown are the representative images of one experiment in the sham and LPS groups. (A) Sham. (B) LPS. Rats received either 0.9% saline (1 mL/kg body wt, i.p.; sham group) or LPS (E. coli, 1 mL/kg body wt, i.p.). The experiments were performed 24 h post-injection. ET-1 was injected at a dose concentration of 60 pmol/min for a period of 10 min. The captured video images were converted from a continuous gray scale (0–255) to one divided into 25 division of 10 intensity units range. For sake of graphic clarity, the image was reversed so that hypoxia shows as dark areas, while well-oxygenated areas are bright. All images were treated identically.

Fig. 3: Changes in the mean intensity in response to ET-1 in sham and LPS-primed animals plotted as a function of time. Rats received either 0.9% saline (1 mL/kg body wt, i.p.; sham group) or LPS (E. coli, 1 mL/kg body wt, i.p.). The experiments were performed 24 h post-injection. ET-1 was injected at a dose concentration of 60 pmol/min for a period of 10 min. The average intensity for a field was measured for each time point. The intensity of each gray scale pixel was assigned a value of 0 to 255, with 0 corresponding to minimum and 255 to maximum fluorescent intensity using the image analysis. The higher intensity corresponds to a lower PO₂ level and vice versa. Data are mean ± SE. * P < 0.05, between the groups, and ^# P < 0.05 vs. respective baseline value (within LPS group).

Fig. 4: Frequency distribution of percentage area with different gray scale intensity ranges. Rats received either 0.9% saline (1 mL/kg body wt, i.p.; sham group) or LPS (E. coli, 1 mL/kg body wt, i.p.). The experiments were performed 24 h post-injection. ET-1 was injected at a dose concentration of 60 pmol/min for a period of 10 min. The average intensity for a field was measured for each time point. The intensity of each gray scale pixel was assigned a value of 0 to 255, with 0 corresponding to minimum and 255 to maximum fluorescent intensity using the image analysis. The range of 0–255 was divided into 25 divisions of 10 units for analysis of frequency distribution of intensity ranges. Shown are the time points at baseline 0 min (A) and at the end of infusion (B) 10 min. Data are mean ± SE. * P < 0.05, between the groups.

Oxygen consumption in isolated perfused liver

In order to further test whether liver surface hypoxia as indicated by fluorescence microscopy reflected impairment of total liver oxygen extraction, we studied oxygen consumption in isolated perfused liver. The dose of ET-1 used in this study caused a marked increase in portal pressure, and this response was potentiated in endotoxemia (Fig. 5). This is consistent with our previous report (¹⁴). ET-1 also caused a transient decrease in oxygen consumption with a time course similar to that of the increase in portal pressure (Fig. 6). This decrease in oxygen consumption was potentiated by pretreatment with LPS. Taken together with the increased level of hypoxia on the liver surface in endotoxic rat livers during ET-1 infusion, these results indicate that the hypoxia is a result of decreased nutrient flow rather than increased consumption.

Fig. 5: Portal pressure response to endothelin in isolated perfused liver. Rats received either 0.9% saline (1 mL/kg body wt, i.p.; sham group) or LPS (E. coli, 1 mL/kg body wt, i.p.). The experiments were performed 24 h post-injection. ET-1 was infused to achieve a final concentration of 2.6 nM at the liver for 10 min. ANOVA showed a significant difference between sham and LPS. ^# P < 0.05 compared to baseline for the same group. * P < 0.05 vs. sham at the same time point.

Fig. 6: Oxygen consumption response to endothelin in isolated perfused liver. In the same experiments as in Figure 5, oxygen consumption was determined by measuring inflow and outflow PO₂. ET-1 infusion produced a significant decrease in oxygen consumption during the infusion. This decrease in oxygen consumption was potentiated in the livers from LPS-pretreated rats by ANOVA. ^# P < 0.05 compared to baseline for the same group. * P < 0.05 vs. sham at the same time point.

DISCUSSION

In the present study we determined the changes in the oxygen distribution on the liver surface in response to ET-1 in LPS-primed animals in order to test the hypothesis that a higher microvascular response to ET-1 contributes to the impaired oxygen delivery in the liver during endotoxemia. The results demonstrated the development of a greater hypoxia in response to ET-1 in livers from LPS-primed animals compared to sham animals. This suggests that ET-1 potentiates the mismatch between oxygen delivery and metabolic demands in such a way that it might lead to exacerbation of hepatic injury. This conclusion was further substantiated by the observation that total oxygen consumption decreased during ET-1 infusion and that this decrease in oxygen consumption was potentiated by LPS pretreatment.

In earlier reports we showed an increase in the hepatic microvascular response to ET-1 in LPS-primed animals (¹⁴). In situ microscopy revealed that sinusoidal constriction in response to ET-1 was profound and sustained in livers from animals pretreated with LPS compared with control livers, suggesting enhanced responsiveness of the sinusoids to ET-1 after LPS pretreatment. A significant sinusoidal narrowing contributed to a significant decrease in the sinusoidal flow. This observation led us to hypothesize that elevated levels of ET-1 could impair the delivery of nutrients and oxygen to the hepatic tissue during endotoxemia.

Previously, we determined the functional significance of hepatic microvascular response to ET-1 in normal animals (¹⁵). Since ET-1 affects hepatic microcirculation, we determined the tissue surface PO₂ change in response to ET-1 to examine if ET-1 altered hepatic oxygen delivery. The results demonstrated heterogeneity in distribution of oxygen on the liver surface in response to ET-1 that was suggested because of sinusoidal site of action of ET-1. In our current study we measured oxygen distribution on the liver surface using the same oxygen mapping. These results provide evidence for an exacerbation of the hypoxia in LPS, pretreated animals, suggesting that hyperresponse of the microcirculation to ET-1 decreases the availability of oxygen in the livers of LPS-pretreated animals.

Administration of LPS at a dose of 1 mg/kg body wt for a period of 24 h led to an increase in the portal pressure. A similar increase in the portal resistance was observed in a previous study (¹⁴). A hyperresponsive hepatic microcirculation to ET-1, along with higher levels of ET-1 during endotoxemia was also observed. Addition of combined ET_A and ET_B receptor antagonist (Bosentan) reverses the endotoxin induced portal hypertension (²⁴). Taken together, the results suggest a role for the endothelins in the development of portal hypertension in LPS-treated animals (^11,14). In our current experiments, the in vivo response of the portal pressure to ET-1 was similar in magnitude in both sham and LPS-primed animals. These results appear to contradict our previous experiments using isolated, perfused livers from LPS-primed animals. This response in isolated perfused liver was also confirmed in the present studies (Fig. 5). In the previous study, addition of ET-1 in an isolated and perfused liver increased the portal resistance; however, unlike the present in vivo experiments, the response was significantly higher in LPS-primed animals compared to sham animals. This discrepancy between isolated and in vivo experiments may result from a more complicated regulation of the hepatic circulation in vivo. Even though we saw no systemic pressure response, it is possible that ET-1 entering the systemic circulation would preferentially cause splanchnic vasoconstriction. Bauer et al. (⁴) demonstrated that a systemic infusion of ET-1 increased the portal pressure despite no measurable effect on the mean arterial pressure. This result suggested that even if changes occur systemically, the response to ET-1 is not reflected in changes in mean arterial pressure. A greater constriction at the upstream splanchnic vessels would attenuate the increase in portal resistance in LPS-primed animals as otherwise observed in a controlled condition of isolated and perfused livers. Further experiments are required to determine the effects of ET-1 in the splanchnic microcirculation and its contribution in the overall portal resistance in the intact perfused livers. A similar response of the portal resistance in vivo in LPS-treated and sham animals raised the question whether or not there is potentiated vascular response at the microcirculatory level in spite of similar results at the macrohemodynamic level.

Infusion of ET-1 in both sham and LPS-primed animals altered the oxygen distribution on the liver surface. In sham animals, there were areas with both high and low oxygen tension. However, the net result was an increase in the average intensity of the fluorescence, indicating decreased PO₂. In the LPS group, the livers showed significantly greater mean fluorescent intensity compared to sham, which suggests an exacerbation in the level of hypoxia. The frequency distribution demonstrated areas with very high fluorescence intensity, indicating severe hypoxia or even anoxia possibly due to sinusoidal shut down. The results indicated an enhanced sinusoidal response leading to potentiation of impaired oxygen delivery in presence of ET-1. Since our experiments were short term, it is not clear if the transient hypoxia would exacerbate overt hepatic injury in LPS-primed animals. Our results using isolated perfused liver did demonstrate that total oxygen consumption of the liver is decreased by endothelin-1 and that this effect is significantly potentiated by LPS pretreatment. Thus, it is likely that a sustained response to elevated endothelin levels in vivo would lead to exacerbation of any direct injury to hepatocytes that results from the endotoxin.

A potentiated response of the hepatic microcirculation to ET-1 during endotoxemia raises questions behind their mechanism of action. Endothelins are a family of 22-amino acid vasoactive peptides have been proposed to contribute to the failure of the microcirculation (^14,25–31). ET-1 is the most widely studied in the group and has been shown to act by binding to both ET_A and ET_B receptors (^4,7). ET-1 levels increase under the influence of cytokines such as transforming growth factor-β released by Kupffer cells during endotoxemia (³²). The higher levels of ET-1 act in a paracrine and autocrine manner in the hepatic microcirculation. ET-1 is cleared primarily in lungs and liver, and the clearance occurs via ET_B receptors. ET-1 acts on ET_A receptors (⁶) to cause vasoconstriction, whereas ET_B receptor stimulation leads to both vasoconstriction (ET_B2) and vasodilation (ET_B1) (^33,34). Studies determining the expression of vasoactive genes showed an increase in the level of ET_B receptor subtype during stresses such as endotoxemia, suggesting a role in the enhanced sensitivity to ET-1 (^11,35). Unlike ET-1, a selective ET_B agonist (IRL-1620) caused an increase in the portal resistance, however, it was less sustained in the LPS group compared to sham group (³³), suggesting a vasodilatory component balancing vasoconstriction during endotoxemia. The diminished portal resistance response was accompanied by amelioration in the hepatic injury. Blocking of nitric oxide (NO) synthesis by L-NAME (NO synthase inhibitor) led to a greater vascular response in LPS-pretreated animals (³³), providing evidence that NO balances vasoconstriction response of ET_B receptors. The study indicated that the higher ET_B proportion during endotoxemia may be linked to the vasodilation component that counterbalances vasoconstriction in the endothelin-mediated vascular response. The study raised a dilemma as to whether the enhanced ET_B receptor serves a protective or a detrimental role in the hepatic microcirculation, and it questioned the role of altered receptor expression in the microvascular hyperresponsiveness.

It is difficult to totally implicate the relative increase in the ET_B receptor proportion in the increase in vascular reactivity to ET-1. The enhanced responsiveness to ET-1 could be related to the differential expression of the receptor subtypes. We would speculate that the response to endotoxemia includes a selective upregulation of ET_B1 receptors in excess of ET_B2. Our previous competitive receptor binding study demonstrated a selective increase in ET_B receptor with no change in the ET_A receptor expression. However, while nearly all of the ET-1-specific binding could be displaced by a combination of ET_A and ET_B antagonists, there was a significant increase in the specific binding that was not displaced by these competitors, suggesting the existence of a novel endothelin receptor subtype (³⁶). The presence of such a receptor subtype with different kinetics from either ET_A or ET_B receptors has been reported in various studies (^37,38). Furthermore, previous experiments have shown that the contraction of hepatic stellate cells (responsible for sinusoidal constriction) in response to endothelins is not completely inhibited by bosentan, a mixed ET_A and ET_B receptor antagonist (³⁹). These results raise a possibility of this novel receptors being upregulated and responsible for the increased microvascular response in endotoxemia. An upregulated level of cytokines, growth factors, and other vasoactive mediators such as NO may alter the level of endothelin receptors during endotoxemia (^40–42). Since Kupffer cells are activated and release cytokines, further experiments are required to elucidate the direct role of Kupffer cells in the altered endothelin receptor expression in the liver during endotoxemia.

In conclusion, hyperresponse of the portal vascular system leads to functional impairment in the oxygen delivery. This causes focal areas of hypoxia in the liver, leading to local mismatch between the oxygen delivery and metabolic demands. A major component of hepatic injury in sepsis is metabolic dysfunction (^43–46). Deficiencies in the oxygen distribution via endothelin-mediated microcirculatory dysfunction would worsen the metabolic dysfunction during sepsis, leading to hepatocellular injury.