INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In critically ill patients, the biphasic oxygen uptake-oxygen delivery model (O₂-DO₂ model) has been used to assess oxygen deficiency, which has been suspected to cause organ dysfunction (1). The pathologic dependence of O₂ on DO₂ at low flows suggests that DO₂ should be increased to supranormal levels in critically ill patients to decrease organ failure (2, 3). In the O₂-DO₂ model, O₂ remains constant as DO₂ varies over a wide range, extracting only as much O₂ from the blood as needed to maintain vital metabolism: this is known as O₂ supply independence. Stable O₂ during O₂ supply independence is thought to signify tissue wellness and maintained oxidative metabolism despite variation in DO₂. When DO₂ declines to a critical threshold value (critical DO₂ or DO₂crit), O₂ can no longer be maintained constant and O₂ declines in direct proportion to DO₂: a state known as O₂ supply dependence. During this O₂ supply dependence, the decrease in O₂ can represent an adaptive reduction in O₂ demand or a manifestation of tissue hypoxia, with an O₂ supply that is inadequate to support O₂ demand. The most important assumption in this model is that O₂ demand is constant at all DO₂ values, so that the covariation of O₂ and DO₂ represents inadequate DO₂ to support metabolism and not O₂ demand dependence. Indeed, when O₂ demand is permitted to vary, the increased O₂ is normally not supported by an increase of O₂ extraction but rather by an increase in DO₂, and this covariation of O₂ and DO₂ is not indicative of tissue hypoxia (4).

In the liver, the O₂ supply-uptake relationship has already been studied in various conditions. Schlichtig and colleagues (5) found that hepatic O₂ remains relatively constant as DO₂ progressively decreases, until a DO₂crit is reached below which hepatic O₂ also decreases. By evaluating hepatic mitochondrial NAD redox state during O₂ supply independence and dependence, they found that the decreased O₂ during oxygen supply dependence represents hepatic dysoxia. Furthermore, by lowering the hepatic blood flow in a model of stagnant hypoxia. Samsel and colleagues (6) also correlated the hepatic oxygen supply-uptake dependence with a switch from lactate consumption to lactate production.

Few studies have sought to determine the relationship between hepatic flow and hepatic functions. Studying galactose elimination as a metabolic function of the liver, Keiding and colleagues (7) showed that galactose elimination and O₂ are independent of DO₂ at high DO₂, whereas at low DO₂ both parameters decrease in parallel, but they did not determine and compare the DO₂crit in the O₂-DO₂ and the galactose elimination-DO₂ relationships. In the present study, we sought to determine whether hepatic urea production is limited at low hepatic DO₂, and if so, whether the deficit in O₂ or the deficit in NH₃ donor induced by low flow rates limits the urea production. To answer this question, we isolated and perfused livers from normal rats and measured urea production (an important and well known function of the liver) at various hepatic flow rates (or hepatic DO₂).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Thirty-five male Sprague-Dawley rats weighing 240 to 410 g were fasted for 24 h prior to the experiment, with free access to water, and anesthetized intraperitoneally with sodium pentobarbital (Nembutal, 50 mg/kg) before the liver perfusion.

Liver Perfusion

Livers were perfused in situ following the method of Hems and colleagues (8). Briefly, the abdominal cavity was opened and the portal vein was cannulated and secured. The ligature was placed around the inferior vena cava above the left renal vein. The infrarenal vena cava was transected and the perfusate was pumped into the portal vein. The flow rate was slowly increased over 1 min (> 3 ml/min/g wet weight). The chest was opened and a second cannula was inserted through the right atrium into the inferior vena cava and secured with a ligature. Finally, the ligature around the lower part of the inferior vena cava was tied, thus closing the circuit. The animal preparation lasted less than 10 min, and the liver was placed in a Plexiglas^® box maintained at steady temperature (37° C). The temperature close to the liver was measured using a temperature probe (Yellow Springs Instrument Co., Yellow Springs, OH). Perfusion pressure was monitored manometrically from tubing attached proximally to the inflow cannula. The entire perfusion system consisted of reservoir, pump (Cole Palmer Instrument Co., Chicago, IL), bubble trap, filter, and oxygenator (9). The livers were perfused with a Krebs-Henseleit- bicarbonate (KHB) buffer (118 mM NaCl, 1.2 mM MgSO₄, 1.2 mM KH₂PO₄, 4.7 mM KCl, 26 mM NaHCO₃, and 2.5 mM CaCl₂). The perfusate was oxygenated with a mixture of 95% O₂-5% CO₂, and the pH was maintained at 7.40. After the surgical procedure, the livers were allowed to recover over 30 min using a nonrecirculation perfusion. Livers were perfused with KHB buffer, KHB + 5 mM glutamine (Gln), or KHB + 5 mM NH₄Cl.

Experimental Protocol

After the recovery period, the pump flow was lowered in stages, reducing gradually the hepatic DO₂ or the hepatic substrate delivery. Ischemic injury was avoided by limiting the stages (six or seven stages for each liver). After the 5-min wait at each new flow, we measured portal pressure, inflow and outflow blood gases, and lactate and urea levels in the outflow perfusate.

Because hepatic Gln delivery can be modified by changing either the hepatic flow rate or the Gln concentrations in the perfusate while keeping hepatic DO₂ constant, eight livers were perfused with a constant flow rate during each experiment while reducing Gln concentrations in the perfusate (5, 2.5, 1, and 0.5 mM).

Additionally, because the hepatic DO₂ can be modified by changing the hepatic flow rate at a constant PO₂ or by changing PO₂ at a constant flow rate, we perfused eight livers with a constant flow rate (and a constant Gln concentration; 5 mM) while gradually reducing PO₂ in the perfusate. In the oxygenator, N₂ was progressively increased while decreasing O₂. By varying the percentage of CO₂ in the oxygenator, the PCO₂ in the inflow perfusate was maintained constant.

Assays

Samples were obtained from the inflow and outflow tubings kept in ice, and assayed for PO₂ within 30 min, using an ABL₂ blood gas analyzer (Radiometer, Copenhagen, Denmark) that was calibrated hourly. Urea and lactate were determined by Sigma kits (Sigma Chemicals, St. Louis, MO).

Viability of Perfused Livers

The viability of perfused livers was assessed by measuring potassium and lactate dehydrogenase (LDH) release in the perfusate at the end of each experiment. Potassium was measured using an electrolyte analyzer (Nova Biomedical, Waltham, MA), and LDH was determined with a Technicon RA500 automatic analyzer (Technicon Instruments, Tarrytown, NY). To determine the degree of swelling during the perfusion, the livers were weighed and freeze-dried over 48 h at the end of each experiment, and the wet/dry weight ratio of each liver was calculated.

Data Analysis

Hepatic O₂ (µl/min/dry g) was calculated from the difference in O₂ contents between the inflow and outflow perfusates. The same solubility coefficient (24 µl of O₂ per ml buffer) was used in the various buffers. Urea and lactate release were determined as (urea or lactate) × flow/liver dry weight. Results were expressed in nmol/min/dry g.

Hepatic DO₂crit was calculated in each liver from a plot of DO₂ (x-axis) and O₂ (y-axis) as the point of intersection of the two best-fit linear regression lines (10). The data were sorted in each liver as DO₂- O₂ pairs with decreasing DO₂. A linear regression line was then calculated for the lowest DO₂ points (O₂ supply dependent), and another regression line was determined for the remaining points (O₂ supply independent). A point from the supply-independent portion was then moved to the supply-dependent portion, and new regression lines were calculated. This process was repeated until the two regression lines with the lowest sum of squared residuals were calculated. The DO₂crit was determined as the DO₂ at the point of intersection of these two lines. O₂ERcrit was calculated at the ratio of O₂ to DO₂ at the point of intersection. The maximal oxygen extraction ratio (O₂ERmax) was chosen among the calculated ratios during the final stage of each experiment (lower DO₂). DO₂crit and critical urea release (UreaRcrit) were also determined from a plot of urea release versus DO₂, using the same computing method as described above.

Data are given as means ± SD. F values were computed by analysis of variance with a Scheffe test to identify differences between groups; p < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

KHB Perfusion (n = 6)

Hepatic O₂ (Figure 1A) was independent of hepatic DO₂ at high DO₂ and became dependent on hepatic DO₂ below DO₂crit, as previously described in vivo (5, 6). For each liver, data adequately fitted the dual-line model. The DO₂crit for this group was 116.0 ± 42.7 µl/min/dry g (Table 1). O₂max was 79.1 ± 16.3 µl/ml/dry g with a slope in the high DO₂ region averaging 0.0015 ± 0.0032. The hepatic O₂ER (Figure 1B) increased in a nearly hyperbolic manner and continued to increase even after DO₂ fell below 116.0 µl/min/dry g (Table ). Lactate release (Figure 1C) appeared when hepatic DO₂ fell below this critical point.

View larger version (17K): [in this window] [in a new window]   Figure 1.   Relationship between oxygen delivery and oxygen consumption (A), oxygen extraction ratio (B), and lactate release (C ) in KHB-perfused livers (n = 6). KHB = Krebs-Henseleit-bicarbonate buffer.

KHB + 5 mM Gln Perfusion (n = 7)

As shown in Figure 2A, the shape of the relationship between DO₂ and O₂ in this group was similar to the previous description. However, perfusion with 5 mM Gln significantly increased DO₂crit, O₂crit, and O₂ERcrit compared with KHB perfusion (Table ). O₂max was 161.4 ± 26.7 µl/ml/dry g, with a slope in the high DO₂ region averaging 0.0088 ± 0.0051. Particularly noteworthy in this group is the high O₂crit and O₂max resulting from Gln addition to the perfusate.

View larger version (19K): [in this window] [in a new window]   Figure 2.   Relationship between oxygen delivery and oxygen consumption (A), oxygen extraction ratio (B), and lactate release (C ) in KHB + 5 mM Gln-perfused livers (n = 7). KHB = Krebs-Henseleit-bicarbonate buffer; Gln = glutamine.

KHB + 5 mM NH₄Cl Perfusion (n = 6)

In order to further investigate the high O₂ observed with 5 mM Gln perfusion, additional livers were perfused with 5 mM NH₄Cl, as a direct NH₃ donor (Figure 3A-C). Unexpectedly, all calculated parameters were similar to the KHB-perfused livers, as shown in Table . 

View larger version (17K): [in this window] [in a new window]   Figure 3.   Relationship between oxygen delivery and oxygen consumption (A), oxygen extraction ratio (B), and lactate release (C ) in KHB + 5 mM NH4Cl-perfused livers (n = 6). KHB = Krebs-Henseleit-bicarbonate buffer.

Urea Release with Gln and NH₄Cl Perfusion

Both relationships between urea release and DO₂ and between urea release and Gln delivery were similar to the previous descriptions (Figure 4). Urea release was independent of DO₂ or Gln delivery in the high DO₂ region, and became O₂ supply- dependent in the low DO₂ region. The DO₂crit (146.5 ± 32.0 µl/min/dry g) calculated from the DO₂-urea release relationship was similar to the value obtained from the DO₂-O₂ relationship (157.6 ± 26.4 µl/min/dry g). At this point, urea release or UreaRcrit was 1,173 ± 263 nmol/min/dry g. UreaRcrit determined from the relationship between Gln delivery and urea release (Figure 4B) was similar (1,166 ± 280 nmol/min/ dry g). Urea release in livers perfused with NH₄Cl was only detectable at low DO₂ values, and consequently the DO₂-urea release relationship could not be obtained in this group as in the KHB-perfused livers.

View larger version (19K): [in this window] [in a new window]   Figure 4.   Relationship between urea release and hepatic oxygen delivery (A) and urea release and glutamine (Gln) delivery (B) in livers perfused with Krebs-Henseleit-bicarbonate (KHB) + 5 mM Gln (n = 7). In each experiment, Gln concentration in the perfusate was kept constant and hepatic flow rate was gradually reduced.

Urea Release when Changing Gln Concentrations in the Perfusate at Constant Flow Rate (n = 8)

Because hepatic Gln delivery can be modified by changing either the hepatic flow rate (and the hepatic DO₂) or the Gln concentration in the perfusate while keeping hepatic DO₂ constant, we performed additional experiments, keeping the same flow rate during each experiment but reducing gradually Gln concentrations in the perfusate (5, 2.5, 1, and 0.5 mM). In this group, hepatic O₂ remained steady at various ranges of Gln delivery (Figure 5A), but urea release decreased with the decrease of hepatic Gln delivery (or Gln concentration in the perfusate) (Figure 5B).

View larger version (17K): [in this window] [in a new window]   Figure 5.   Relationship between hepatic oxygen consumption and Gln delivery (A) and urea release and Gln delivery (B) in livers perfused with Krebs-Henseleit-bicarbonate (KHB) and various Gln concentrations (5, 2.5, 1, and 0.5 mM) (n = 8). In each experiment, hepatic flow was kept constant.

Urea Release when Changing PO₂ in the Perfusate at a Constant Flow Rate and a Constant Gln Concentration (5 mM) (n = 8)

Because the hepatic DO₂ can be modified by changing the hepatic flow rate at a constant PO₂ or by changing PO₂ at a constant flow rate, we perfused eight livers with a constant flow rate (and a constant Gln concentration; 5 mM) while gradually reducing PO₂ in the perfusate (Figure 6). The shape of the relationship between DO₂ and O₂ and between DO₂ and urea release was similar to the one described in Figure 2A and in Figure 4A. The DO₂crit calculated from the two relationships was similar: 117.4 ± 7.8 µl/min/dry g (DO₂-O₂ relationship) and 98.3 ± 7.7 µl/min/dry g (DO₂-urea release relationship).

View larger version (19K): [in this window] [in a new window]   Figure 6.   Relationship between hepatic oxygen consumption and hepatic oxygen delivery (A) and urea release and hepatic oxygen delivery (B) in livers perfused with Krebs-Henseleit-bicarbonate (KHB) + 5 mM Gln at constant flow rate (n = 8). Hepatic oxygen delivery was modified by decreasing PO2 in the inflow perfusate.

Viability of Perfused Livers

As shown in Table 2, wet/dry weight ratio was higher in livers perfused with KHB + Gln than in livers perfused with KHB alone or KHB + NH₄Cl. Liver weight/body weight ratio was similar to the ratio observed in the literature. No LDH release was detected in the perfusate at the end of the experiment, and potassium release remained steady over the experimental periods.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

As observed in the biphasic O₂-DO₂ relationship, when livers are perfused with Gln, urea production remains constant at high DO₂ and declines in direct proportion to DO₂ below the DO₂crit. Furthermore, DO₂crit is similar when calculated from both the O₂-DO₂ model and the urea release-DO₂ relationship. Therefore, hepatic urea production becomes O₂ supply-dependent at the same DO₂crit as determined by the O₂-DO₂ relationship. When Gln concentration in the perfusate was gradually reduced while keeping hepatic DO₂ constant, urea production decreased proportionally with Gln concentrations in the perfusate. Furthermore, when Gln concentration and flow rate were maintained constant while decreasing PO₂ in the inflow perfusate (as well as hepatic DO₂), urea production declined below the DO₂crit. Consequently, urea production is dependent on Gln and O₂ availability and becomes limited at the same DO₂crit determined by the O₂-DO₂ relationship.

Numerous studies have emphasized the importance of determining the systemic DO₂-O₂ relationship to improve the mortality rate of critically ill patients. Besides the systemic DO₂- O₂ relationship, several investigators also investigated the relationship in various organs (11) and found that O₂ supply dependence in organs commences at values slightly different from whole-body DO₂crit values (15). Nelson and colleagues (17) demonstrated that the gut DO₂crit was reached at a stage when nongut O₂ was still independent of O₂ supply. Similarly, supply limitation of O₂ occurs at a higher systemic DO₂ in the contracting diaphragm, suggesting that in diseases associated with increased work of breathing and decreased DO₂, the diaphragm may become metabolically impaired before limitation of systemic O₂ is observed (15). These studies have demonstrated that undetectable hypoxia may occur in organs without changes in whole-body parameters of oxygenation, emphasizing the importance of studying the relationship in individual organs. However, besides parameters of oxygenation, few data exist about the relationship between functions and flow in these models, and the isolated perfused livers appear to be a useful tool in determining such relationships in steady experimental conditions.

In the liver, the DO₂-O₂ relationship has been studied previously. In these studies, the decrease in DO₂ was induced by progressive hemorrhage (5), selective decrease of hepatic blood flow (6), or partial occlusion of hepatic vessels (13). All these in vivo studies found a biphasic relationship between DO₂ and O₂. In our model, we reached similar conclusions, O₂ was independent of DO₂ at high DO₂ and became O₂ supply-dependent at low DO₂. Above the DO₂crit, O₂ continued to increase with a slope in the high DO₂ region, averaging 0.0015 ± 0.0032, which is similar to results reported by Samsel and colleagues (6).

Whether hepatic function becomes O₂ supply-dependent at the DO₂crit determined by the O₂-DO₂ model remains unclear. Schlichtig and colleagues (5) showed that below the DO₂crit, -hydroxybutyrate-to-acetoacetate ratio increased with the fall in DO₂, whereas Samsel and colleagues (6) demonstrated that below the DO₂crit, livers change from lactate consumption to lactate production. These findings further support the hypothesis that a decrease in O₂ accompanying O₂ supply dependence represents hepatic hypoxia. In our study, the shape of the relationship between urea release and DO₂ was similar to the O₂-DO₂ relationship. DO₂crit was similar when calculated from these two relationships. When Gln and flow rate were kept constant while hepatic DO₂ decreased by lowering PO₂ in the perfusate (Figure 6), urea release also decreased below the DO₂crit. Thus, urea production becomes O₂ supply-dependent below the DO₂crit. When hepatic Gln delivery was modified by changing Gln concentrations in the perfusate at constant flow rate (and constant DO₂), we found a direct relationship between urea release and Gln delivery (or Gln concentrations in the perfusate), demonstrating that urea production is also dependent on Gln availability.

In KHB- and KHB + 5 mM NH₄Cl-perfused livers, lactate release appeared below the DO₂crit, supporting the hypothesis that DO₂-O₂ dependence represents hepatic hypoxia. O₂ supply dependence of urea production also supports this hypothesis. However, absence of cell damage in our model did not exclude an adaptative modification of hepatic metabolism away from urea production toward preservation of cell morphology. The duration of low DO₂ needed to produce cell damage is highly variable among organs, and the liver is thought to be well preserved after several hours of ischemia. This high tissue resistance might explain the absence of cell damage despite hepatic hypoxia in our model.

This study also emphasized the high potency of Gln as a NH₃ donor for urea synthesis, whereas NH₄Cl was less efficient. In NH₄Cl-perfused livers, urea production was undetectable at high DO₂ and we were unable to describe the urea-DO₂ relationship in this group. The high potency of Gln as a NH₃ donor may be explained by the high intracellular concentrations of Gln (30 to 35 mM) obtained when livers from normal rats are perfused with 5 mM concentrations (18). When hepatic O₂ demand was increased by adding Gln to the perfusate, DO₂crit also increased, suggesting that DO₂crit depends on O₂. Similarly, Nelson and colleagues (19) showed that endotoxemia increases both systemic DO₂crit and O₂max.

The fact that hepatic O₂ doubled when Gln-perfused livers produced high amounts of urea is also surprising. However, when livers were perfused with low concentrations of Gln in the perfusate, urea release significantly decreased, whereas O₂ remained constant. Thus, urea production did not account for the high O₂ demand observed in Gln-perfused livers. Gln may be used in other pathways in hepatocytes and other hepatic cells such as Kupffer and endothelial cells (20). In a similar experimental model, when Gln was added to the perfusate, liver mass rapidly increased, with a net uptake of potassium followed by a marked release of potassium when the liver mass reached its new steady state (18). Additionally, transport of Gln in hepatocytes is associated with sodium entry, subsequent depolarization of the cell membrane, and activation of the Na⁺/K⁺-ATPase pump. Changes in the ion conductance through the membrane may explain the high O₂ observed in these livers.

Limitations in our model must be pointed out. Livers are denervated and perfused only through the portal vein with a hemoglobin-free buffer. Thus, the high PO₂ might exaggerate the O₂ gradient between vessels and hepatic cells. Hepatic O₂ was in the range previously reported in livers perfused with KHB buffer in the absence of energy substrate (21). The liver weight/body weight ratio was similar to previous studies and the increased wet/dry weight ratio in Gln-perfused livers confirms previous reports.

In summary, our results demonstrate that hepatic urea production is dependent on Gln and O₂ availability and becomes O₂ supply-dependent at the same DO₂crit as determined by the O₂-DO₂ relationship. Besides parameters of oxygenation, this finding points out the importance of studying the relationship existing between organ functions and flow. Whether the decrease in urea production at low DO₂ values represents an adaptative modification of hepatic metabolism away from urea production toward preservation of cell morphology remains to be determined. Study of additional hepatic functions should help to further understand the relationship between functions and flow, especially at low DO₂ values.