| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TRANSLATIONAL PHYSIOLOGY
1Department of Digestive Surgery, University Hospital Northern Norway, Tromsø, Norway; 2Liver Failure Group, The University College London Institute of Hepatology, University College London, London, United Kingdom; 3Department of Cellular Neuroscience, Max-Delbrück Center for Molecular Medicine, Berlin, Germany; 4Department of Surgery, Maastricht University, Maastricht, The Netherlands; and 5Hepatology Unit, Hospital Univeritario de Valme, Sevilla, Spain
Submitted 19 September 2005 ; accepted in final form 27 February 2006
ABSTRACT
Ammonia reduction is the target for therapy of hepatic encephalopathy, but lack of quantitative data about how the individual organs handle ammonia limits our ability to develop novel therapeutic strategies. The study aims were to evaluate interorgan ammonia metabolism quantitatively in a devascularized pig model of acute liver failure (ALF). Ammonia and amino acid fluxes were measured across the portal drained viscera (PDV), kidneys, hind leg, and lungs in ALF pigs. ALF pigs developed hyperammonemia and increased glutamine levels, whereas glutamate levels were decreased. PDV contributed to the hyperammonemic state mainly through increased shunting and not as a result of increased glutamine breakdown. The kidneys were quantitatively as important as PDV in systemic ammonia release, whereas muscle took up ammonia. Data suggest that the lungs are able to remove ammonia from the circulation during the initial stage of ALF. Our study provides new data supporting the concept of glutamate deficiency in a pig model of ALF. Furthermore, the kidneys are quantitatively as important as PDV in ammonia production, and the muscles play an important role in ammonia removal.
amino acids; hyperammonemia; hepatic failure; urea cycle
Previous studies in animal models and patients with liver disease have pointed to an important role for the gut and kidneys in the production of ammonia and the muscles in the removal of ammonia (13). In addition, because the lungs contain both PAG and glutamine synthase (GS) (27), they could also play a role in interorgan ammonia metabolism. The present study was therefore designed to determine the quantitative dynamics of ammonia metabolism and its relationship to interorgan glutamine and glutamate metabolism. Furthermore, the role of the lungs in interorgan ammonia trafficking was studied. Because interorgan ammonia and related amino acid trafficking are technically difficult to study in humans, we used a previously well characterized large (porcine) animal model of ALF (25, 35–38).
MATERIALS AND METHODS
Study outline.
The Norwegian Experimental Animal Board approved the present study. Sixteen female Landrace pigs (23–30 kg) were randomly allocated into sham-operated controls or ALF. Study outline is shown in Fig. 1. Blood and urine sampling was performed 
30–45 min after creation of the portacaval shunt (PCS) or completion of sham surgery (t = 0 h). ALF was induced shortly after completion of the sampling procedures. The experiments were terminated with an overdose of pentobarbital sodium and potassium chloride at t = 6 h.
|
All animals received 500 ml 0.9% NaCl containing 625 mg glucose as a preoperative load to prevent any preoperative dehydration. During the experiment, 0.9% NaCl was infused at a rate of 3 ml·kg–1·h–1. Anesthesia was stopped after the liver was devascularized. If the degree of sedation became insufficient, small doses of fentanyl and midazolam were given as a bolus. Sham-operated animals received continuous anesthesia during the experimental period and received equal amounts of intravenous fluids.
ALF was induced with an end-to-side portacaval shunt followed by ligation of the hepatic arteries. Details of the surgery, including the sham-operation procedure, have been described elsewhere (35, 36). After ALF induction, 50% glucose and 20% human albumin (Octapharm, Hurdal, Norway) were continuously infused at rates of 0.6048 and 0.66 ml·kg–1·h–1, respectively. However, sham-operated animals received only half the amount of glucose (0.3024 ml·kg–1·h–1) to make the glucose levels comparable between the groups. Heparin (2,500 IU) was given intravenously to all pigs at the start of the experiment.
Positioning of catheters, flow probes, sampling, and analytical procedures. Catheters developed by Hallameesch et al. (6) and Ten Have et al. (28) were inserted in the abdominal aorta, renal vein, portal vein, and femoral vein for arterial and venous blood sampling. A 16-G central venous catheter (Secalon T, Ohmeda, Swindon, UK) was introduced into the left external jugular vein for administration of drugs and fluids. p-Amminohippuric acid (PAH; 25 mM; A1422; Sigma) was infused at a rate of 30 ml/h through this catheter after an initial bolus of 6 ml (28). Portal and femoral blood flow was measured by the use of perivascular ultrasonic transit time flow probes (CardioMed Systems; Medistim A/S, Oslo, Norway). A 5-Fr Edwards Swan-Ganz catheter (Baxter Healthcare, Irvine, CA) was floated into the pulmonary artery via the right external jugular vein. The urine bladder was drained via a cystotomy.
Blood and urine samples were collected on ice at the times for measurement of blood flow and processed as described previously (5). Tissue samples were freeze clamped with Wollenberger tongs cooled in liquid nitrogen (34) and frozen at –80°C. Ammonia, urea, and PAH were determined spectrophotometrically (5). Amino acids were determined using HPLC (29).
Calculations. Plasma flow rate (ml·kg body wt–1·h–1) of the kidneys was calculated using the formulae based on the method of indicator dilution and Fick’s principle (5, 17). The PAH-determined blood flow and data from the perivascular blood flow probes were converted to plasma flow using the hematocrit. Substrate fluxes across organs were calculated as the venous-arterial concentration difference times the plasma flow. Positive values reflect substrate release, and negative fluxes reflect substrate uptake. Kidney and hind leg data are multiplied by two to reflect both organs.
A primed-constant infusion of the stable glutamine isotope tracer [L-5-15N]glutamine (Cambridge Isotope Laboratories, Woburn, MA) was used to calculate whole body rate of appearance of glutamine (RaGLN). [L-5-15N]glutamine (1.50 mg/kg body wt) was given as an intravenous bolus at t = –2 h, after which a continuously infusion of 1.06 mg·kg body wt·h L-[L-5-15N]glutamine was started. RaGLN was derived from the equation Ra = I/TTRA, where I is the tracer infused (µmol·kg body wt–1·min–1) and TTRA is the tracer/tracee ratio in arterial plasma (33). TTR values were corrected for background values.
Statistics.
Statistical analysis was performed using the Statistical Package for the Social Sciences, version 11.0 for Windows (SPSS, Chicago, IL). Data are expressed as means ± SE. The Wilcoxon’s signed-rank test was applied to test whether organ flux was different from zero. Two-way ANOVA was applied to test for differences within and between groups over time. An overall significance in analyses of variance for repeated measurements (F-test, P 
0.05) may be attributable to either the effect of group (PG) or the interaction for group and time (PGT). The overall significance for the effect of group meant that the groups were different when all of the repeated measurements were taken together (independent of time), whereas a significant interaction denotes a different time course in the two groups. PT 0.05 denotes a significant change during the experimental period but without any difference between the groups. Huynh-Feldt epsilon factor adjusted probability levels were used when Mauchly’s test of sphericity was significant.
The Mann-Whitney U-test was used to test for differences between the groups at preexperimental defined time points. Probability values 0.05 were considered significant for all tests applied.
RESULTS
Effects of PCS and ALF on arterial concentrations. Ammonia (Table 1) was significantly higher after the creation of PCS (P < 0.01). ALF induction, in addition to PCS, increased arterial ammonia levels further, and this difference increased during the experimental period (PGT = 0.001).
|
Glutamine was not different from sham-operated animals after initiation of PCS but significantly increased after ALF induction (P < 0.01). Glutamine increased further in ALF, whereas sham-operated animals remained stable for the rest of the experiment (PGT < 0.001). Glutamate was lower after PCS induction (P < 0.01) and decreased further after ALF induction, whereas controls remained stable during the experimental period (PGT < 0.001).
Alanine was not different between the groups after PCS induction but significantly higher after ALF induction (P < 0.01). Moreover, alanine remained higher throughout the experimental period compared with sham-operated animals (PGT = 0.035).
Whole body RaGLN in ALF. Figure 2 shows the RaGLN. There were no differences between sham and ALF at t = 0 h. RaGLN increased in ALF, whereas it slightly decreased in shams (PGT = 0.06). Accordingly, RaGLN for glutamine was overall higher in ALF compared with sham-operated animals during the time period studied (PG = 0.002).
|
|
|
Glutamate was taken up by the PDV in sham-operated animals at t = 0 h (P < 0.05), whereas flux was not different from zero after PCS induction and significantly different from sham animals (P < 0.01). ALF induction did not have an effect. Glutamate was taken up by PDV in shams, whereas it was released or not different from zero during the rest of the experimental period (PG < 0.001). Alanine flux was not different from zero at any time point in any of the groups.
Effects of PCS and ALF on renal metabolism in ALF. Venous-arterial differences are shown in Table 4. The kidneys released ammonia in both groups at all time points. There was no significant difference between the groups after PCS (Tables 4 and 5), but the kidneys released more ammonia into the systemic circulation after ALF induction (P < 0.05). However, ammonia flux was not different between the groups during the rest of the experimental period.
|
|
|
Glutamine flux was not different from zero in either group at any time point. Although a significant time effect was found (P = 0.04) indicating a trend toward glutamine uptake, this trend was similar in both groups causing no significant difference between the groups. Accordingly, no differences in glutamine flux across the kidneys were found (Table 6).
The kidneys took up glutamate in sham-operated animals (P < 0.05). PCS induction changed this pattern toward less uptake (P < 0.01), whereas ALF induction, on top of PCS, switched net glutamate uptake to net glutamate release (P < 0.01). Because ALF pigs released glutamate into the systemic circulation during the remaining experimental period, a highly significant group effect was detected (PG < 0.001).
Alanine flux across the kidneys was not different between or within the groups at any time point.
Effects of PCS and ALF on hind leg metabolism. Venous-arterial differences are shown in Table 7. Ammonia was taken up at all time points in both groups (Tables 7 and 8), but uptake of ammonia across the hind leg increased after PCS induction (P < 0.01). This increased uptake was sustained after ALF induction and during the rest of the experimental period (P < 0.01).
|
|
Glutamate was taken up at all time points in sham-operated controls. PCS induction did not change this. However, glutamate uptake was significantly decreased after ALF induction (P < 0.01), and this difference increased because glutamate was not taken up across the hind leg at t = 6 h in the ALF group, causing a significant interaction for glutamate flux (PGT = 0.035).
Alanine was released from the hind leg at all time points. PCS induction did not change this. However, alanine flux was decreased after ALF induction, and alanine was not significantly released during the rest of the experimental period, an effect that was confirmed by the ANOVA analysis (PG = 0.001).
Effects of PCS and ALF on lung metabolism. Venous-arterial differences are shown in Table 9. The lungs took up ammonia after the creation of PCS, whereas ammonia flux was not different from zero in shams. ALF induction did not change this difference at t = 2 h, because the lungs continued to take up ammonia and the shams remained not different from zero (Table 10). However, this difference vanished, and ammonia flux was not different from zero in pigs with ALF at t = 4 and 6 h. Glutamine, glutamate, and alanine fluxes were not significantly different from zero.
|
|
|
|
|
Reduction in ammonia concentration remains one of the main strategies to treat hepatic encephalopathy (HE), but incomplete understanding and the lack of sequential quantitative data regarding interorgan metabolism of ammonia have limited our ability to do so. Studies in patients with cirrhosis have identified important roles for the kidneys and muscle in addition to the gut to be involved in ammonia homeostasis (13, 14). Because interorgan ammonia trafficking is relatively difficult to study in human ALF, we decided to use a well-characterized porcine model of ALF for this purpose (25, 35–38). PCS and subsequent ligation of the hepatic arteries in this model induce hyperbilirubinemia, hyperammonemia, hyperdynamic circulation, intracranial hypertension, renal dysfunction, and an abnormal hemostasis (38). This model was thought to be suitable for our study, because our main focus was to evaluate interorgan ammonia metabolism in an animal model of ALF that could be repeatedly sampled and that depicted the characteristic end-organ dysfunction typically seen in human ALF (11). Additionally, during the time period of the study, this model does not show any activation of inflammatory cytokines allowing an assessment of the metabolic processes involved without the confounding effects of an inflammatory response (24).
Hyperammonemia developed shortly after PCS induction and increased further after ALF induction to arterial levels previously reported in experimental and human ALF (1, 2, 5, 8). Arterial ammonia levels remained within the normal range for sham-operated pigs (38). Ammonia release from the PDV was nearly significantly increased compared with sham-operated pigs (PG = 0.062). In the postabsorptive state, ammonia was produced in equal amounts in the small and large intestines. No difference in glutamine flux across the PDV was observed. This is different from the situation in the humans with cirrhosis and rats with portacaval shunting where an increase in the uptake of glutamine by the PDV was observed (13). This indicates that the PDV in ALF contributes to the hyperammonemic state mainly through increased portasystemic shunting and not as a result of increased ammonia production from glutamine.
The kidneys were a major contributor to the hyperammonemic state because ammonia was continuously released from the kidneys in the ALF group. As opposed to a previous study in rats (4), we did not observe any adaptation to the hyperammonemic state either by increased urinary ammonia excretion or by decreased ammonia release into the general circulation. In contrast, the kidneys excreted significantly less ammonia over time compared with sham-operated controls. This lack of adaptation may be related to the relatively short duration over which the study was conducted or to the mechanism of ammonia generation in the kidneys. In this model, ammonia genesis was not associated with increased glutamine uptake.
Hind leg muscle removed significantly more ammonia during the initial phase after PCS and during the first 2 h after ALF induction, compared with sham-operated controls. However, we were unable to find an overall increase in muscle ammonia uptake in ALF. Our observation supports earlier experimental data in rats, where the investigators were unable to show a significant net uptake of ammonia across the hind quarter (5). Also, our results are in agreement with a human study (1), which quantified ammonia consumption by leg muscle to be 100 nmol·100 g–1·min–1. Although glutamine was released from the hind leg muscle, we could not detect any differences compared with sham-operated controls. Net detoxification by muscle occurs if ammonia is taken up by muscle and glutamine is produced. The discussion of whether ammonia uptake by skeletal muscle will always lead to glutamine release and whether this glutamine release is stoichiometric to ammonia uptake has been confounded by contrasting literature (5). The data with respect to ammonia metabolism in the lung are difficult to interpret due to large variations, but they do indicate that the lungs are metabolically active. The reason for these large variations is likely to reflect small venoarterial differences and high blood flow. ALF induced an increase in total tissue ammonia in jejunum and lung tissue homogenates, but this was not significantly increased in the kidneys and hind leg muscle. The significance of high tissue ammonia in relationship to organ failure is likely to be important, because the concentrations were between 5 and 10 times higher than the circulating plasma levels. At these millimolar concentrations, ammonia has several other effects such as inhibition of essential enzyme function that may be important in maintaining cellular bioenergetics, metabolism, and integrity (16, 39).
Arterial glutamine levels increased rapidly and were more than twice the arterial glutamine levels observed in sham-operated controls after 6 h. However, the glutamine levels are lower than those previously reported in the literature and may represent substrate deficiency (glutamate) (1). This increase in arterial glutamine was associated with a significant increase in RaGLN.
Glutamate depletion after ALF induction was a particular feature of this study, which may explain some of the differences from previous studies in which glutamine has been shown to be the fuel for the generation of ammonia in the PDV and also in the kidneys. Accordingly, glutamate was reduced eightfold, 2 h after induction of ALF, and this low level was sustained throughout the course of the experiment. This observation has been previously explored in humans with liver failure, but the results of glutamate levels have been variable. In the older studies (18, 21), a high glutamate value was reported. However, with improvement in the techniques of analysis, the more recent studies have shown a low value of glutamate in patients with liver failure (1, 23, 26). These latter observations support the concept that glutamate depletion is indeed a feature of liver failure. The observed glutamate depletion does provide support for further development of the currently available agents and ideas for newer therapies. The administration of a large-quantity glutamate may improve outcome in liver failure was reported in a preliminary communication more than 50 years ago (31). L-Ornithine L-aspartate (LOLA), which is a mixture of two amino acids, provides intermediates that increase glutamate availability for synthesis of glutamine and illustrates that absolute or relative glutamate deficiency may underlie the pathogenesis of hyperammonemia. Administration of LOLA into animals with ALF resulted in reduced brain water (20) and into patients with HE resulted in an improvement in HE compared with placebo-treated controls (9). These data, if confirmed in future studies, will allow the development of newer strategies.
This glutamate depletion may explain the observed lack of uptake of ammonia and glutamate and production of glutamine by the hind leg after 2 h. The mechanism of this glutamate depletion is not clear but may reflect loss of hepatic glutamate synthesis capacity. Alternatively, or in addition, this depletion may be contributed to by consumption in the production of glutamine in the first 2 h. At the interorgan level, glutamate tended to be taken up in the PDV in the sham-operated animals, whereas glutamate was produced in the PDV in the ALF group (PG < 0.001), which may indicate an increased glutamine breakdown through the action of gut PAG. Glutamate was also released into the systemic circulation from the kidneys, which may be consequent on the activity of renal PAG (10).
Before the data can be generalized to humans, one must consider that our study describes a rapidly evolving liver failure in a matter of hours rather than days, is an irreversible surgical model, and we are describing events occurring in the early phase of the disease. At this stage, the animals do not show any evidence of an inflammatory response, which is often seen in patients with liver failure that proceed to develop severe encephalopathy (7, 19, 30). However, our data and those in the literature provide a compelling argument to study new therapeutic strategies based on the knowledge from the studies of interorgan metabolism (1, 12, 14, 15, 26)
In conclusion, the result of our study provides sequential data in the early phase of liver failure in a noninflammatory model. We have shown that the rate of glutamine production is increased in ALF, but its arterial levels are only modestly elevated. These observations support the concept of glutamate deficiency in ALF and the basis of possible new therapies. In addition, we provide further evidence that the kidneys are quantitatively as important as PDV in relationship to ammonia production and an important role for the muscles in ammonia metabolism. For development of future therapies to reduce ammonia in liver failure, this critical interplay between the organs involving ammonia, glutamate, and glutamine would have to be taken into account.
GRANTS
This work was supported by the Norwegian Research Council, Sir Siegmund Warburg Voluntary Settlement, the Liver Research Foundation, and Maastricht University.
ACKNOWLEDGMENTS
H. Hagerup, E. Hareide, H. Mæhre, and T. Solvang are acknowledged for excellent technical assistance during the experiments. We are grateful to Dr. T. Wilsgaard for support on statistical methods. Teraklin provided travel costs for the investigators.
Address for reprint requests and other correspondence: R. Jalan, Liver Failure Group, Institute of Hepatology, Univ. College London, 69–75 Chenies Mews, London WC1E 6HX, UK (e-mail: r.jalan{at}ucl.ac.uk)
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
REFERENCES
: a biochemical marker for increased intracranial pressure in pigs with acute hepatic failure. Scand J Gastroenterol 35: 546–551, 2000.[CrossRef][ISI][Medline] in pigs with acute encephalopathy. Scand J Clin Lab Invest 61: 217–225, 2001.[CrossRef][ISI][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
