Isoleucine infusion during “simulated” upper gastrointestinal bleeding improves liver and muscle protein synthesis in cirrhotic patients§

Authors

  • Steven W. M. Olde Damink,

    1. Department of Surgery, Academic Hospital, Maastricht University, Maastricht, The Netherlands
    2. Liver Unit, Royal Infirmary of Edinburgh, Edinburgh, Scotland, UK
    Search for more papers by this author
  • Rajiv Jalan,

    Corresponding author
    1. Liver Unit, Royal Infirmary of Edinburgh, Edinburgh, Scotland, UK
    2. Liver Failure Group, The UCL Institute of Hepatology, Division of Medicine, University College London, London, UK
    • Liver Failure Group, The UCL Institute of Hepatology, Division of Medicine, University College London, 69-75 Chenies Mews, London WC1 6HX, United Kingdom
    Search for more papers by this author
    • Potential conflict of interest: Dr. Jalan received grants from Wellcome Trust and Plasma Protein Therapeutic Association.

    • fax: (44) 2073800405

  • Nicolaas E. P. Deutz,

    1. Department of Surgery, Academic Hospital, Maastricht University, Maastricht, The Netherlands
    Search for more papers by this author
  • Cornelis H. C. Dejong,

    1. Department of Surgery, Academic Hospital, Maastricht University, Maastricht, The Netherlands
    2. Department of Surgery, Royal Infirmary of Edinburgh, Edinburgh, Scotland, UK
    Search for more papers by this author
  • Doris N. Redhead,

    1. Department of Radiology, Royal Infirmary of Edinburgh, Edinburgh, Scotland, UK
    Search for more papers by this author
  • Paula Hynd,

    1. Department of Dietetics, Royal Infirmary of Edinburgh, Edinburgh, Scotland, UK
    Search for more papers by this author
  • Peter C. Hayes,

    1. Liver Unit, Royal Infirmary of Edinburgh, Edinburgh, Scotland, UK
    Search for more papers by this author
  • Peter B. Soeters

    1. Department of Surgery, Academic Hospital, Maastricht University, Maastricht, The Netherlands
    Search for more papers by this author

  • Presented in part as an abstract at BASL September 2001 and at the 52th Annual Meeting of the American Association for the Study of Liver Diseases in Dallas, TX, November 2001.

  • Data on basal ammonia, urea, and amino acid metabolism (Olde Damink et al., HEPATOLOGY 2002;36:1163–1171) and on the effects of a “simulated bleed” without isoleucine infusion on ammonia, urea, and amino acid metabolism (Olde Damink et al., HEPATOLOGY 2003;37:1277–1285) are presented in detail elsewhere.

Abstract

Upper gastrointestinal (GI) bleeding in cirrhotic patients has a high incidence of mortality and morbidity. Postbleeding catabolism has been hypothesized to be partly due to the low biological value of hemoglobin, which lacks the essential amino acid isoleucine. The aims were to study the metabolic consequences of a “simulated” upper GI bleed in patients with cirrhosis of the liver and the effects of intravenous infusion of isoleucine. Portal drained viscera, liver, muscle, and kidney protein kinetics were quantified using a multicatheterization technique during routine portography. Sixteen overnight-fasted, metabolically stable patients who received an intragastric infusion of an amino acid solution mimicking hemoglobin every 4 hours were randomized to saline or isoleucine infusion and received a mixture of stable isotopes (L-[ring-2H5]phenylalanine, L-[ring-2H4]tyrosine, and L-[ring-2H2]tyrosine) to determine organ protein kinetics. This simulated bleed resulted in hypoisoleucinemia that was attenuated by isoleucine infusion. Isoleucine infusion during the bleed resulted in a positive net balance of phenylalanine across liver and muscle, whereas renal and portal drained viscera protein kinetics were unaffected. In the control group, no significant effect was shown. Conclusion: The present study investigated hepatic and portal drained viscera protein metabolism selectively in humans. The data show that hepatic and muscle protein synthesis is stimulated by improving the amino acid composition of the upper GI bleed by simultaneous intravenous isoleucine administration. (HEPATOLOGY 2007;45:560–568.)

Upper gastrointestinal (GI) bleeding in patients with cirrhosis is associated with life-threatening complications such as spontaneous bacterial peritonitis, sepsis, renal failure, and encephalopathy,1–4 resulting in the death of approximately 20% to 30% of patients despite adequate control of bleeding.5 These complications are thought to result in part from the metabolic consequences of the blood protein load in the digestive tract.6

Seventy-five percent of whole blood protein is hemoglobin, which comprises 98% of the erythrocyte protein. Hemoglobin is unique because it is totally devoid of the essential branched-chain amino acid (BCAA) isoleucine.6 Moreover, albumin and globulin also have an extremely low isoleucine content.6 Therefore, an upper GI bleed presents the gut with protein of very low biological value, and consequently the absorbed amino acids from blood protein cannot be used for protein synthesis.

An upper GI bleed results in extremely low isoleucine concentrations in experimental animals, healthy volunteers, and patients with normal and impaired liver function.7–9 We hypothesized that decreased plasma and tissue isoleucine concentrations could impair protein synthesis6 and therefore influence the function of rapidly dividing cells (e.g., immune cells) and short half-life proteins (e.g., clotting factors). This may increase the risk of further bleeding, infection, and catabolism in postbleeding patients with cirrhosis. Theoretically, supplying isoleucine in this situation would be beneficial.

The purpose of this study was to test the hypothesis that simulating an upper GI bleed in patients with cirrhosis impairs organ protein synthesis that can be restored by intravenous infusion of isoleucine.

Abbreviations

BCAA, branched chain amino acids; GI, gastrointestinal; NB, net balance; PDV, portal drained viscera; TIPSS, transjugular intrahepatic portosystemic stent-shunt.

Patients and Methods

Patients.

Sixteen metabolically stable patients with biopsy-proven cirrhosis who underwent portography were studied (Table 1). Exclusion criteria included severe ascites, pitting edema, hepatic encephalopathy, active alcohol abuse, diabetes, cardiovascular disease, serum creatinine >100 μM, malignancy, pregnancy, and shunt dysfunction (portal pressure gradient >12 mm Hg). Patients did not use lactulose, diuretics, or antibiotics. Patients maintained a food chart and had continued their usual diet during the week preceding the study. The study was approved by the Lothian Research Ethics Committee, and written informed consent was obtained from each subject. CTP score defined severity of liver disease.

Table 1. Patient Characteristics
CharacteristicsSB-SalineSB-Isoleucine
  1. NOTE. Data are expressed as the mean (SEM). Literature reference values19 for healthy control subjects are: body mass index, 20–25 kg/m2; fat-free body mass (% body weight), 86%; body cell mass (% fat-free body mass), 67%.

Age, years45 (4)51 (4)
Sex  
 Female24
 Male64
Weight (kg)83.7 (6.7)75.8 (6.6)
Body mass index (kg/m2)30.9 (2.9)28.8 (2.1)
Total body water (l)48.6 (4.5)43.7 (3.4)
Fat-free body mass (kg)66.4 (6.1)59.7 (4.6)
Fat-free body mass (% body weight)83.1 (3.7)79.7 (3.4)
Extracellular water (% total body water)45.2 (1.3)40.9 (2.7)
Body cell mass (kg)36.7 (4.0)35.5 (3.5)
Body cell mass (% fat-free body mass)54.9 (1.3)59.1 (2.7)
Etiology7 Alcoholic liver disease; 1 primary sclerosing cholangitis6 Alcoholic liver disease; 1 primary biliary cirrhosis; 1 cryptogenic
CTP score6.4 (0.5)7.0 (0.3)
Portal pressure gradient (mm Hg)5.4 (1.2)5.4 (1.4)
Months after TIPSS placement30.7 (6.7)25.9 (6.2)
Albumin (g/l)36.1 (1.6)36.0 (1.2)
Alanine aminotransferase (IU/l)32.0 (3.1)38.8 (6.0)
International normalized ratio1.4 (0.1)1.3 (0.1)
Creatinine (μM)81.4 (2.8)84.9 (4)
60% Daily protein intake (g)46 (4)42 (4)

Simulation of an Upper GI Bleed.

An upper GI bleed was “simulated” by administration of a solution that was identical to the amino acid composition of the hemoglobin molecule.2 To ensure that patients received a comparable nitrogen load that would lead to a moderate elevation of plasma ammonia, we administered an amino acid quantity that equaled 60% of daily nitrogen intake.2 The amino acid solution was infused nasogastrically continuously over the 4-hour study period at a mean rate of 113 ± 8 ml/h.

Infusion of Isoleucine or Placebo.

At the same time as the amino acid infusion was started (t = 0 h), a continuous intravenous infusion of an iso-osmotic solution containing 40 g/l of isoleucine (Clinical Pharmacy, azM, Maastricht, the Netherlands) or saline was commenced using an IVAC pump. The amount of isoleucine infused was calculated to equal half the amount of leucine administered in the amino acid solution to reflect the normal leucine/isoleucine ratio in average protein.10 Patients were randomly assigned to one of two study groups: SB-saline and SB-isoleucine.

Administration of Stable Isotopes.

Studies were started after an overnight fast 3 hours before the portography (t = −180 min) (Supplementary Table 1; Available at the HEPATOLOGY website: http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html). Twenty grams of deuterium oxide dilution (99.9 atom%, D-4501; Sigma, Boston, MA) and 30 ml of 150 mM sodium bromide were administered orally (Clinical Pharmacy) to determine total body water and extracellular water.6 A primed, continuous infusion of a mixture of the following stable isotopes (Cambridge Isotope Laboratories, Woburn, MA) was administered until the end of the experiment (t = 240 min): L-[ring-2H5]phenylalanine ([D5]-Phe, single priming dose 0.5 mg/kg body weight, infusion 0.5 mg/kg body weight/h), L-[ring-2H4]tyrosine ([D4]-Tyr, only prime 0.08 mg/kg body weight), and L-[ring-2H2] tyrosine ([D2]-Tyr, single priming dose 0.1 mg/kg, infusion 0.1 mg/kg body weight/h). Arterial blood samples were taken until t = 0 minutes and starting at t = 90 minutes with 30-minute intervals during the infusion of the simulated bleed to assess the achievement of steady state in isotope enrichments.

Insertion of Catheters.

Catheters were inserted after portography as described previously.10 In short, a nasogastric tube was inserted, the right femoral artery was cannulated, and a sheath with a sampling port was inserted in the right femoral vein. A sampling catheter was introduced through this sheath into the right renal vein. A Swan-Ganz catheter with a distal (infusion) and proximal (sampling) port was inserted through the right internal jugular vein sheath, and the proximal port was positioned in the trunk of the portal vein distal to the shunt.

Sampling and Measurement of Blood Flow.

Organ blood flow was determined as detailed recently.10, 11 Kidney, leg, and portal flow were determined using a primed, continuous infusion of para-aminohippuric acid (MSD, Haarlem, the Netherlands), and liver plasma flow was determined using a primed, continuous infusion of indocyanine green (Cardiogreen; BDMS, Aalst, Belgium).10–12 Portal, renal, and leg blood flow were measured immediately before simulation of the bleed (0 min) and 120, 180, and 240 minutes thereafter. Hepatic blood flow was measured prior to and 240 minutes after the simulated bleed when the position of the catheters was checked using fluoroscopy (Supplementary Table 1).10, 11

Analytical Procedures.

Blood samples were collected from the femoral artery and the portal, renal, femoral, and hepatic veins at the times of measurement of blood flow. A blood sample for deuterium and bromide analysis was taken 4 hours after ingestion. Blood electrolytes (CO-oximeter, IL 282; Instrumentation Laboratories), plasma ammonia, urea, PAH and ICG (spectrophotometry),10 bromide (fully automated HPLC),13 and deuterium (isotope ratio mass spectroscope)6 were measured.13 Plasma amino acid concentrations and tracer amino acid enrichments were measured using an LCQ chromatography mass spectrometry system (Thermoquest LCQ, Veenendaal, the Netherlands).14

Calculations.

Plasma flow was calculated according to Fick's principle.10 Substrate fluxes across organs were calculated as venous − arterial concentration difference × plasma flow. Liver fluxes were calculated using our recently described model11 in which we assume that all the portal venous blood is shunted through the transjugular intrahepatic portosystemic stent-shunt (TIPSS). To calculate substrate fluxes across the liver, we had to estimate the amount of portosystemic shunting through the TIPSS. Walser et al.15 showed that the shunted fraction of portal blood via a TIPSS is on average 93%. In the calculations we used the maximum fraction that can be shunted (100%), because all our patients had patent TIPSS.

Assuming 100% portal-systemic shunting:

Liver flux = hepatic venous − arterial concentration difference × liver plasma flow.

Renal and leg plasma flow and substrate fluxes are expressed per two kidneys and for one leg. Amino acid enrichments were calculated as tracer/tracee ratios. BCAAs were calculated as the sum of valine, leucine, and isoleucine. Total body water, fat-free body mass, and body cell mass were calculated from bromide and deuterium dilution.13

Organ Tracer Kinetics.

Net organ protein balance (protein synthesis minus breakdown) was determined by dilution of stable isotopes across the organs using a two-compartment model.16 In this model, disposal (nmol/kg body cell mass/min) is the total rate of metabolism of the amino acid across an organ (incorporation into protein plus degradation), whereas production (nmol/kg body cell mass/min) represents the rate of amino acid release from protein breakdown across that organ. In the present study, disposal and production of phenylalanine is calculated because phenylalanine is only incorporated in protein and is not produced de novo. The phenylalanine method has been proven to be valid in cirrhotic patients,17–19 whereas it has been shown that leucine metabolism is altered in patients with cirrhosis.20 (A detailed description of the method and its calculations is provided as Supplementary material at the HEPATOLOGY website [http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html].)

Statistics.

Results are presented as the mean ± SEM. Results are presented as baseline values (t = 0 h) and study end values (the mean values of the final hour of the amino acid infusion, when isotopic plateau conditions were obtained). Wilcoxon's signed rank test was used to test for differences from zero and to test for significant changes in arterial values. Because of missing data in the organ flux and protein synthesis data, the Mann-Whitney U test was used to test for differences between the basal and study-end data within the groups.

Results

General.

Patient characteristics were similar (Table 1), and all patients had patent TIPSS. A mean of 46 ± 4 g and 42 ± 4 g of amino acids were administered in the SB-saline and the SB-isoleucine group, respectively, representing the approximate hemoglobin content of 300 ml whole blood. One patient did not tolerate the nasogastric tube, and the study was stopped after 2.5 hours. Due to technical difficulties, we could not obtain an arterial sample at t = 2 hours in 1 patient, the portal and hepatic sample at t = 4 hours in a second patient, the arterial and portal sample at t = 3 hours in a third patient, and the renal samples at t = 3 and 4 hours in a fourth patient. Stable isotope data could not be obtained in 2 patients due to malfunction of the infusion pumps.

No complications were encountered either during catheter insertion or the infusions. Mean hepatic fractional extraction of ICG was 0.22 ± 0.06 and 0.30 ± 0.07 in the SB-saline and the SB-isoleucine group, respectively (P = 0.4).5 The body composition of the patients did not differ between groups19, 20 (Table 1).

Arterial Data.

The administration of the amino acid solution resulted in a significant decrease in arterial isoleucine concentration to 39% of initial value in the saline group, with a plateau at t = 180 minutes (Supplementary Table 2). The ammonia concentration increased significantly, with a plateau between t = 180 and 240 minutes that was not affected by isoleucine infusion (Supplementary Table 2). After the simulated bleed in the SB-saline group, all other amino acids increased. The most significant increase was in leucine (400%), valine (335%), and alanine (200%), resulting in an elevation of 170% of initial value in the sum of all amino acids measured (Table 2). Supplementation of isoleucine prevented the fall in isoleucine levels, resulting in a significantly higher isoleucine concentration that was equivalent to 37% of valine and 52% of leucine concentrations at the end of the study. The change in isoleucine concentration was significantly different between the two groups (P < 0.001).

Table 2. Arterial Substrate Concentrations
 TimeSB-SalineP ValueSB-IsoleucineP Value
  1. NOTE. Data are expressed as the mean ± SEM in μM. End values represent the mean values of the final hour of amino acid infusion. P = 0.001 only in isoleucine.

  2. Abbreviations: TAA, the sum of all amino acids; TEA, the sum of all essential amino acids; TNEA, the sum of nonessential amino acids.

Ammonia075 ± 6 82 ± 6 
 End125 ± 90.018140 ± 120.012
Urea03,691 ± 256 3,386 ± 276 
 End3,981 ± 2780.0463,883 ± 3210.012
Isoleucine038 ± 3 37 ± 4 
 End15 ± 10.018178 ± 100.012
Leucine079 ± 9 67 ± 6 
 End319 ± 290.018341 ± 270.012
Valine0144 ± 19 128 ± 10 
 End483 ± 390.018487 ± 320.012
Glutamine0631 ± 40 637 ± 33 
 End719 ± 340.018745 ± 490.025
Alanine0225 ± 35 196 ± 18 
 End435 ± 370.018490 ± 440.012
TNEA01,461 ± 99 1,457 ± 77 
 End2,064 ± 460.0012,433 ± 1690.001
TEA0719 ± 63 731 ± 53 
 End1,800 ± 1320.0012,393 ± 1870.001
TAA02,622 ± 159 2,638 ± 135 
 End4,325 ± 1600.0185,298 ± 3470.012

Tracer Steady State.

Isotopic steady state of all stable isotopes was achieved at baseline and also at 180 minutes after the amino acid solution administration was started. The tracer/tracee ratio of all infused isotopes was below 6% (data not shown).

Portal Drained Viscera Metabolism.

Portal drained viscera (PDV) plasma flow did not change significantly during the experiment (SB-saline 19.6 ± 4.8 to 25.5 ± 5.5 ml/kg body cell mass/min and SB-isoleucine 26.8 ± 5.5 to 32.4 ± 3.7 ml/kg body cell mass/min). The baseline amino acid fluxes across the PDV were comparable between the two groups (Table 3). The PDV produced ammonia in the postabsorptive state, and production increased after administration of the amino acid solution; however, this did not reach statistical significance because of the large SEMs. Isoleucine infusion during the simulated bleed resulted in a significantly enhanced uptake of isoleucine by the PDV. Consequently, the change in isoleucine uptake by the PDV was significantly greater in the SB-isoleucine group (P < 0.05). There was significant release of leucine and valine in the SB-saline group (both P = 0.029). This could not be demonstrated in the SB-isoleucine group, probably because of the large SEMs.

Table 3. Amino Acid Fluxes Across Organs
 TimePDVLiverMuscleRenal
SB-SalineSB-IsoleucineSB-SalineSB-IsoleucineSB-SalineSB-IsoleucineSB-SalineSB-Isoleucine
  • NOTE. Data are expressed as the mean (SEM) in nanomoles per kilogram body cell mass per minute. End values represent the mean values of the final hour of amino acid infusion.

  • Abbreviations: TAA, sum of all amino acids measured; TEAA, sum of essential amino acids measured; TNEAA, sum of nonessential amino acids measured.

  • *

    Fluxes significantly different from zero.

  • P < 0.05.

  • P < 0.01.

  • §

    P < 0.001.

Ammonia01,217 (412)*1,454 (278)*−439 (186)*−138 (101)−458 (109)*−843 (146)*234 (104)251 (125)
 End2,274 (783)*2142 (223)*−431 (191)−140 (128)−1259 (106)*§−950 (56)*1,351 (233)*§1,127 (329)*
Isoleucine046 (31)55 (23)*−11 (16)−41 (14)*5 (12)32 (25)5 (19)−14 (21)
 End40 (33)−397 (75)*§−51 (9)*−398 (62)*§2 (29)−257 (55)*−34 (29)−213 (99)*
Leucine0169 (58)*174 (42)*−111 (58)−112 (35)*35 (24)77 (42)68 (32)*59 (32)
 End2,739 (1,307)*2,368 (1,208)−311 (421)−1070 (336)*§−535 (327)−497 (127)*§−282 (302)−65 (302)
Valine095 (79)122 (86)−89 (41)*−126 (44)*−24 (60)92 (71)20 (75)14 (64)
 End2,765 (1,445)*1,638 (1,248)−253 (605)−1299 (479)*−553 (373)−537 (264)−205 (411)−182 (296)
TNEAA0−304 (520)−865 (937)−4,549 (1,882)*−2,114 (914)*1,526 (448)*2,354 (1,093)*−604 (457)−894 (885)
 End7,568 (3,768)6,506 (2,481)*−5,519 (2,675)*−7,774 (1,270)*1,757 (1,437)1,569 (667)*−3,627 (1,011)*−2,029 (1172)
TEAA0739 (407)207 (340)−1,325 (632)*−988 (377)*205 (198)1,120 (681)279 (278)−108 (200)
 End13,281 (6,234)9,749 (4,614)−3,511 (2,584)−7,577 (1,433)*§−1,694 (1,589)−1,990 (718)*−826 (1,660)−509 (1,176)
TAA0614 (911)*−805 (1,378)−5,777 (2,504)*−3374 (1,431)*1,781 (616)*3,285 (1,361)*−179 (702)−1,097 (1,181)
 End20,781 (9,954)*16,173 (7,049)−9,218 (5,109)−15,451 (2,401)*§30 (2,892)−205 (1,341)−4,483 (2,483)−2,399 (2,401)

PDV Protein Kinetics.

The PDV were in a net catabolic state evidenced by negative net balances (NBs) of phenylalanine at t = 0 hours and protein breakdown estimates that were higher than protein synthesis estimates (SB-saline: synthesis 80 ± 59, breakdown 131 ± 53, NB −51 ± 33 nmol phenylalanine/kg body cell mass/min; SB-isoleucine: synthesis −63 ± 205, breakdown 182 ± 67, NB −199 ± 153 nmol phenylalanine/kg body cell mass/min) (Fig. 1). The NB of phenylalanine did not change significantly during the protocol in either group, but the rate of protein turnover was increased, as evidenced by elevated protein synthesis and protein breakdown in both groups (SB-saline: synthesis 450 ± 205, breakdown 1,397 ± 412, NB −992 ± 567 nmol phenylalanine/kg body cell mass/min; SB-isoleucine: synthesis 861 ± 442, breakdown 954 ± 294, NB −166 ± 507 nmol phenylalanine/kg body cell mass/min) (Fig. 1).

Figure 1.

Data of protein synthesis, breakdown, and the resulting net balance, expressed as nanomoles of phenylalanine per kg body cell mass per minute, of the portal drained viscera (left panels) and the liver (right panels) for both the isoleucine study group (lower panels) and saline controls (upper panels). Protein synthesis data of the liver are corrected for hydroxylation (see Patients and Methods). Post values represent the mean values of the final hour of amino acid infusion. #P < 0.05 (Mann-Whitney U test for differences within groups).

Liver Metabolism.

Liver plasma flow did not change significantly during the experiment (SB-saline: 14.1 ± 2.9 to 16.9 ± 4.9; SB-isoleucine: 9.9 ± 3.0 to 10.8 ± 2.0 ml/kg body cell mass/). Basal ammonia and amino acid metabolism was comparable between groups and did not differ significantly (Table 3). In the SB-saline group, the simulated bleed resulted in an increased uptake of isoleucine by the liver but did not result in a change in uptake of ammonia or amino acids. Isoleucine infusion during the simulated bleed increased the uptake of most amino acids by the liver, including isoleucine. The change in uptake of isoleucine was significantly greater in the SB-isoleucine group. Isoleucine infusion did not alter liver removal of ammonia.

Liver Protein Kinetics.

Administration of the amino acid solution did not change hepatic protein kinetics in the SB-saline group; however, protein synthesis was stimulated in the SB-isoleucine group, resulting in an increased NB of phenylalanine at study end (Fig. 1). No differences were observed in protein breakdown estimates. (t = 0 h: SB-saline: synthesis 415 ± 120, breakdown 263 ± 50, NB 152 ± 76 nmol phenylalanine/kg body cell mass/min; SB-isoleucine: synthesis 218 ± 37, breakdown 109 ± 25, NB 98 ± 33 nmol phenylalanine/kg body cell mass/min. Study end: SB-saline: synthesis 274 ± 250, breakdown 108 ± 162, NB 166 ± 231 nmol phenylalanine/kg body cell mass/min; SB-isoleucine: synthesis 839 ± 221, breakdown 157 ± 204, NB 682 ± 165 nmol phenylalanine/kg body cell mass/min) (Fig. 1).

Muscle Metabolism.

Plasma flows did not change significantly during the experiment (leg flow 15.3 ± 5.1 to 21.3 ± 5.1 and 22.4 ± 3.9 to 13.3 ± 3.2 ml/kg body cell mass/min for the SB-saline and the SB-isoleucine group, respectively). Baseline ammonia and amino acid fluxes across the muscle were comparable between groups (Table 3). The administration of the amino acid solution resulted in increased ammonia uptake in the SB-saline group; this is in contrast with the SB-isoleucine group, in which no differences could be observed. The SB resulted in increased uptake of valine, leucine, isoleucine, and total essential amino acids in the SB-isoleucine group, but not in the SB-saline group.

Muscle Protein Kinetics.

At t = 0 h, muscle protein kinetics did not differ between the two groups (SB-saline: synthesis 117 ± 52, breakdown 137 ± 51, NB −20 ± 19 nmol phenylalanine/kg body cell mass/min; SB-isoleucine: synthesis −31 ± 201, breakdown 196 ± 61, NB −185 ± 152, nmol phenylalanine/kg body cell mass/min) (Fig. 2). Administration of the amino acid solution did not change the NB in the SB-saline group, whereas it increased significantly in the SB-isoleucine group, resulting in a net anabolic state (SB-saline: synthesis 372 ± 211, breakdown 288 ± 175, NB 87 ± 140 nmol phenylalanine/kg body cell mass/min; SB-isoleucine: synthesis 377 ± 135, breakdown 159 ± 100, NB 261 ± 153 nmol phenylalanine/kg body cell mass/min) (Fig. 2).

Figure 2.

Data of protein synthesis, breakdown, and the resulting net balance, expressed as nanomoles of phenylalanine per kilogram body cell mass per minute, of the muscle (left panels) and the kidney (right panels) for both the isoleucine study group (lower panels) and saline controls (upper panels). Protein synthesis data of the kidney are corrected for hydroxylation (see Patients and Methods). Post values represent the mean values of the final hour of amino acid infusion. #P < 0.05 (Mann-Whitney U test for differences within groups).

Renal Metabolism.

Renal plasma flows across organs did not change significantly during the experiment (18.0 ± 4.8 to 21.1 ± 6.7 and 21.1 ± 5.4 to 21.9 ± 5.5 ml/kg body cell mass/min for the SB-saline and the SB-isoleucine group, respectively). Baseline flux data were similar in both groups (Table 3). Administration of the amino acid solution resulted in elevated renal ammoniagenesis in both groups (P = 0.06 in the SB-isoleucine group) (Table 3).

Renal Protein Kinetics.

The NB of phenylalanine did not differ significantly between groups (SB-saline: synthesis −17 ± 65, breakdown 77 ± 66, NB −95 ± 23 nmol phenylalanine/kg body cell mass/min; SB-isoleucine: synthesis 88 ± 67, breakdown 89 ± 35, NB −3 ± 42 nmol phenylalanine/kg body cell mass/min) (Fig. 2) and did not change during the experiment in either group, although the SB-isoleucine group showed an increase in protein synthesis by the kidney (P = 0.014) (SB-saline: synthesis 67 ± 78, breakdown −106 ± 98, NB 173 ± 153 nmol phenylalanine/kg body cell mass/min; SB-isoleucine: synthesis 550 &plusmn, 148; breakdown 275 ± 201, NB 275 ± 233 nmol phenylalanine/kg body cell mass/min) (Fig. 2).

Discussion

We were able to investigate multiple organ protein metabolism selectively in cirrhotic patients. The results of this study demonstrate that isoleucine infusion during intragastric administration of an amino acid solution that resembles the amino acid pattern of hemoglobin stimulates liver and muscle protein synthesis. In the control group, this normal response to administration of an amino acid mixture18, 19 was impaired.

The simulated bleed resulted in increased arterial concentrations of all amino acids except isoleucine, which was decreased markedly. This amino acid pattern resembled the changes observed in patients admitted with an acute upper GI bleed21 and experimental animals receiving intragastric erythrocytes.8, 9, 22 The common finding of severe hypoisoleucinemia can be explained by the fact that hemoglobin does not contain isoleucine and by BCAA-antagonism,6 for which the present study supplies circumstantial evidence. The uptake of isoleucine by the liver was enhanced after the simulated bleed in the SB-saline group, although arterial isoleucine concentrations dropped. Hemoglobin comprises 98% of erythrocyte protein, is totally devoid of the BCAA isoleucine, and contains large amounts of the BCAAs valine and leucine. Consequently, the simulated upper GI bleed resulted in a 3-4-fold increase in arterial valine and leucine concentration. As we have detailed previously,6 the imbalance between the BCAAs can lead to BCAA antagonism, which presumably results from their common pathway of transport and degradation. First, increased plasma concentrations of leucine and valine could compete with the decreased isoleucine concentrations for transport across the cell membrane via their shared transport carrier.23, 24

Second, the three BCAAs have a common degradation pathway.23 Within the cell, the initial step is a reversible, concentration-dependent, transamination reaction by BCAA aminotransferase. The rate of transamination is concentration-dependent, and BCAA aminotransferase is widely distributed among tissues, with high activity in skeletal muscle and low activity in the liver and intestine.25 After this transamination reaction the resulting 2-oxoacids undergo an irreversible oxidative decarboxylation reaction by the BCOD-complex, the rate-limiting step in BCAA oxidation.25 The activity of this complex exhibits large differences in various organs and is activated most potently by the oxo-acid of leucine: alpha-ketoisocaproate.26, 27 Therefore, the high plasma leucine concentration that results from an upper GI bleed may stimulate BCAA oxidation, irrespective of the level of isoleucine, resulting in a further depression of plasma and tissue isoleucine concentrations.6, 28

The administration of the simulated bleed did not result in a positive NB in any of the organs studied. This supports our hypothesis that the limited availability of isoleucine prevents the normal stimulatory effect of hyperaminoacidemia on protein synthesis that follows an amino acid meal.29, 30 It is still not known whether anabolism after amino acid administration in healthy subjects and patients with cirrhosis is achieved via inhibition of proteolysis or stimulation of protein synthesis or both (see Charlton29 for review). In a recent study, McCullough et al.20 showed that suppression of proteolysis in response to amino acid administration was impaired in cirrhotic patients and that anabolism was achieved by increased protein synthesis.

Short-term limited availability of a single essential amino acid during insulin and amino acid infusion impairs the synthesis of any protein requiring that amino acid.31 Impaired protein synthesis after an upper GI bleed has been hypothesized to occur6 based on literature reports in which hypoisoleucinemia resulted in impaired protein synthesis, DNA synthesis, cell proliferation, and cell function.32–34 Recently, it has been shown that cells undergo apoptosis when deprived of isoleucine.35 Lecavalier et al.36 showed that even a short period (8 h) of hypoisoleucinemia impaired whole body protein synthesis in humans. Although the present study did not show a net catabolic state in any of the organs measured after amino acid ingestion, one has to bear in mind that only a bleeding of 300 ml was simulated. Further research is necessary to determine the metabolic consequences of an actual upper GI bleed that can result in several liters of blood in the gut.5 At present, clotting abnormalities, inadequate response to injury, and high incidence of infections after an upper GI bleed in patients with liver failure are thought to result from loss or use of clotting factors and acute phase proteins. We also hypothesize6 that impaired protein synthesis following an upper GI bleed may play an important role in interfering with the production of these proteins, which may further increase the risk of bleeding, infection, and catabolism.

The present study supports our hypothesis that parenteral infusion of isoleucine may be an effective and simple treatment for the adverse metabolic effects of upper GI bleeding.6 Isoleucine infusion balanced the BCAA pattern of arterial plasma and probably transformed the imbalanced amino acid composition of the hemoglobin molecule into a balanced protein, which can be used for protein synthesis in liver and muscle. In agreement with the recent report of McCullough et al.,20 a net anabolic state was reached via increased incorporation of amino acids for protein synthesis and not via inhibited proteolysis. Isoleucine infusion stimulated the uptake of most amino acids by the liver. We suggest that improving liver protein synthesis during an acute upper GI bleed in patients with cirrhosis has major clinical effects, because liver protein synthesis is of pivotal importance in generating adequate cellular immunity, an acute phase response, and normal clotting function.

In this study, we could not demonstrate an increase in net protein synthesis across the PDV with infusion of isoleucine. In the saline control group, there was a tendency toward a decrease in NB after the simulated bleed; however, this decrease was not statistically significant, probably because of the large variations observed in the fluxes measured across the PDV. Furthermore, these large SEMs prevented us from detecting a significant difference between the two groups. However, the present study provides circumstantial evidence that PDV protein kinetics during a simulated bleed may be positively influenced by intravenous isoleucine infusion. First, the NB of phenylalanine remained the same in the SB-isoleucine group, whereas there was a tendency toward a decrease in the SB-saline group. Second, in the SB-saline group there was a greater appearance of essential amino acids in the portal vein, suggesting that there was greater retention or use of essential amino acids by the gut in the SB-isoleucine group during the simulated bleed, because both groups received similar amounts of essential amino acids via the simulated bleed. Such increased retention in the gut wall would be in keeping with previous observations by our group in pigs37 and would fit in the concept of a labile protein pool.37 Third, in the SB-isoleucine group there was enhanced uptake of isoleucine by the PDV, suggesting enhanced use of this essential amino acid. Further research is necessary to put these observations into perspective.37

The kidneys have recently been recognized as organs with a very high protein turnover.38 In the present study, we showed an increase in renal protein synthesis by administration of isoleucine, but this did not result in significantly increased NB. The latter may be the consequence of a lack of power of the present study. As detailed recently,11 the kidneys have a key role in the hyperammonemia that follows an upper GI bleed in patients with cirrhosis. Renal ammoniagenesis after the administration of a simulated bleed is probably related to alanine uptake,11 which was not influenced by isoleucine infusion in the present study. This may explain the lack of influence of isoleucine infusion on postbleeding hyperammonemia.

In summary, we have shown that infusion of isoleucine during a simulated bleed in patients with cirrhosis restores impaired protein synthesis of liver and muscle leading to a net anabolic state in these organs. Further research is needed to evaluate the therapeutic benefit of intravenous isoleucine infusion during actual acute upper GI bleeding in patients with cirrhosis, with particular focus on the effects on clotting function, infection rate, and mortality.

Acknowledgements

We gratefully acknowledge the nursing staff of the Liver Unit and the Department of Radiology Royal Infirmary of Edinburgh for support and help. We thank Hans M. H. van Eijk, Ph.D., Jean L. J. M. Scheyen, B.Sc., and Gabrie A. M. Ten Have, B.Sc., for analytical help.

Ancillary