Evidence of Inter-organ Amino-Acid Transport by Blood Cells in Humans (erythrocytes/alanine/liver/gluconeogenesis/glucose-alanine cycle)

Transcription

1 Proc. Nat. Acad. Sci. USA Vol. 70, No. 6, pp , June 1973 Evidence of Inter-organ Amino-Acid Transport by Blood Cells in Humans (erythrocytes/alanine/liver/gluconeogenesis/glucose-alanine cycle) PHILIP FELIG*, JOHN WAHRENt, AND LARS RAFt * Department of Internal Medicine, Yale University School of Medicine, New Haven, Connecticut 06510; and t Department of Clinical Physiology and I Department of Surgery of the Serafimer Hospital and Karolinska Institute, Stockholm, Sweden Communicated by Lewis Thomas, April 6, 1973 ABSTRACT To evaluate the contribution of blood cellular elements to inter-organ transport of amino acids, net exchange across the leg and splanchnic bed of 17 amino acids was determined in seven healthy postabsorptive subjects by use of both whole blood and plasma for analysis. Arterial-portal venous differences were measured in five additional subjects undergoing elective cholecystectomy. By use of whole blood, significant net release of amino acids was noted from the leg and gut, while a consistent uptake was observed by the splanchnic bed. The output of alanine from the leg and gut and the uptake of this amino acid by the splanchnic bed exceeded that of all other amino acids and accounted for 35-40% of total amino-acid exchange. Transport by way of plasma could not account for total tissue release or uptake of alanine, threonine, serine, glutamine, methionine, leucine, isoleucine, tyrosine, and citrulline. For each of these amino acids, significant tissue exchange was calculated to occur by way of the blood cellular elements, the direction of which generally paralleled the net shifts occurring in plasma. For alanine, 30% of its output from the leg and gut and 22% of its uptake by the splanchnic area occurred by way of blood cells. We conclude that the blood cellular elements, presumably erythrocytes, contribute substantially to the net flux of amino acids from muscle and gut to liver in normal postabsorptive humans. Alanine predominates in the inter-organ transfer of amino acids occurring by way of blood cells as well as plasma. It is generally believed that plasma rather than erythrocytes is the vehicle of amino-acid exchange between tissues (1). The slow equilibration time of amino-acid transport across erythrocyte membranes, as indicated by in vitro studies (2), has led to the widely held notion that erythrocytes are of little, if any, significance in the inter-organ transfer of amino acids. As a consequence, studies of amino-acid metabolism in physiologic as well as pathologic circumstances have generally been restricted to measurements of amino-acid concentrations in plasma rather than whole blood (1, 3). However, recent observations on glutamate flux in intact humans suggest a dynamic role for erythrocytes in the transport of this amino acid across muscle tissue (4). In addition, studies in dogs indicate that erythrocytes and plasma may play independent and occasionally opposing roles in the exchange of several amino acids across the liver and gut (5, 6). The present investigation was consequently undertaken to determine the contribution of blood cells, primarily erythrocytes, to inter-organ amino-acid transport in normal humans. This was done by examination of peripheral, splanchnic, and portal exchange of 17 individual amino acids in the postabsorptive Abbreviations: A-FV, arterial-femoral venous; A-HV, arterialhepatic venous; A-PV, arterial-portal venous. To whom reprints should be addressed state; whole blood as well as plasma was used for the determination of arterio-venous differences. In particular, we were interested in the contribution of the blood cellular elements to the tissue exchange of alanine. Plasma analyses have emphasized the relative importance of alanine as the key gluconeogenic precursor on the basis of its predominance in the flux of amino acids from muscle to liver in postabsorptive (7, 8), prolonged fasted (7, 8), and exercising humans (9). METHODS Subjects. Two groups were studied. One group consisted of seven healthy adult male volunteers in whom amino-acid exchange across the splanchnic bed and leg was determined by means of hepatic venous and femoral venous catheterization. The second group consisted of five subjects in whom aminoacid exchange across the gut was investigated by means of portal venous blood sampling at the time of elective cholecystectomy for uncomplicated cholelithiasis. All subjects were informed of the nature, purpose, and possible complications of the study before obtaining their consent to participate. Procedures. All studies were performed with the patients in the recumbent position after an overnight (10-14 hr) fast. To examine leg and splanchnic amino-acid exchange, blood samples were obtained simultaneously from catheters placed in the brachial artery, hepatic vein, and femoral vein (9). For the portal vein study, a brachial artery catheter was inserted percutaneously before surgery. During the surgical procedure, at which time the patients were anesthetized with short-acting barbiturates and halothane, blood samples were obtained simultaneously from the arterial catheter and by direct needle puncture from the portal vein. The catheters in all studies were kept patent by intermittent flushing with saline. Neither glucose nor heparin was infused to the subjects during the study. Analyses and Calculations. Blood for amino-acid analysis was collected in heparinized tubes. Whole blood and plasma (obtained by centrifugation at 5600 X g and 50) were deproteinized with equal volumes of 10% sulfosalicyclic acid (10), and the filtrates were stored at -20. To remove glutathione, which interferes with the chromatographic determination of amino-acid concentration in whole blood in that it emerges as a broad peak underlying several amino acids, the wholeblood filtrates were treated with sodium sulfite before analysis (10). The whole-blood filtrates (4.0 ml) were neutralized to ph 7 with NaOH, 0.2 ml of 0.5 M sodium sulfite was then added, and the solution was allowed to stand at room temperature (250) for 3-4 hr, after which the filtrates were acidi-

2 1776 Physiology: Felig et al. TABLE 1. Arterial concentrations of free amino acids in whole blood and plasma Whole blood Plasma (#M) (MM) P* Taurine i ± 3.1 <0.001 Aspartate t Threonine i i 9.1 <0.01 Serine i i 7.9 <0.001 Glutaminet i ± 21.1 NS Proline i ± 8.4 NS Citrulline 50.7 i i 3.0 <0.005 Glycine i ± 15.1 <0.001 Alanine ± i 17.8 <0.005 a-aminobutyrate 25.9 ± ± 4.1 NS Valine ±t i 15.1 NS Cystine ± Methionine 15.9 i i 1.9 NS Isoleucine 62.4 i ± 3.0 NS Leucine ± i 6.0 NS Tyrosine 61.1 ±t ± 3.6 NS Phenylalanine 54.3 ± NS * P = significance of difference between whole blood and plasma concentration (paired t-test). Proc. Nat. Acad. Sci. USA 70 (1973) fied again with HCl to ph 2.2 (10). Individual acidic and neutral amino acids were determined by the automated ionexchange chromatographic technique (10) on a Beckman model 120C amino-acid analyzer (Beckman Instrument Co., Palo Alto, Calif.). By this technique, glutamine and asparagine emerge as a single peak, and are referred to collectively as glutamine, since inter-organ exchange of asparagine is negligible (11). Glutamate values on the other hand are often artifactually elevated with the chromatographic technique (12) and consequently are not reported. Aspartic acid, readily measurable in whole blood, was present in such small concentrations in all plasma samples (<10 MAmol/liter) as to preclude accurate integration of its peak. Cystine was not detectable in the whole-blood samples since the sodium sulfite treatment used for removal of glutathione results in the conversion of cystine to cysteine-s-sulfonate and its emergence as part of the undifferentiated peak appearing at column volume (10). Analysis of replicate samples of plasma and of whole blood revealed coefficients of variation ranging between 3 and 6% for each of the amino acids analyzed. From the hematocrit determination in each patient, the contribution of plasma alone to the amino-acid concentration in a liter of whole blood was calculated: (100-hematocrit) X (0.01) X Mtmol/liter of plasma plasma contribution. Arterial-venous differences for plasma amino acids (Tables 2-4) = are consequently reported in Mmol/liter of whole blood. The contribution of the blood cellular components was then calculated by the formula: whole blood AA plasma AA - = blood cell AA In this equation AA refers to amino-acid concentration in,4mol/liter of whole blood. The amount of plasma trapped with cells in determination of the hematocrit is quite small (13) and may be ignored. Hepatic blood flow was determined by the continuous infusion technique (14), with indocyanine green dye (15). Statistical analyses were determined by the paired t-test. Data are expressed as means 1 standard error. RESULTS Arterial concentrations The arterial concentrations of individual amino acids as determined by analyses of whole blood and plasma are shown in Table 1. The concentrations of taurine, aspartate, threonine, serine, citrulline, glycine, and alanine were significantly higher in whole blood than in plasma. This was particularly true in the case of aspartate, the concentration of which in plasma was too small (<10 Mumol/liter) to permit accurate integration of its peak on the chromatogram..for the remaining TABLE 2. Arterial-femoral venous differences of free amino acids in whole blood, plasma, and blood cells (Amol/liter of whole blood) Taurine 2.4 ± 7.5 NS 0.4 ± 0.9 NS 2.1 ± 7.5 NS Aspartate -2.0 ± 12.1 NS Threonine -7.7 ±t 3.9 < ± 1.7 < ± 4.2 NS Serine 8.6 ± 2.9 < ± < P < ± 4.4 NS Glutaminet ± 11.8 < ± 86 < ± < P < 0.1 Proline ± 5.4 < ± 5.0 NS 5.0 ± 15.1 NS Citrulline -6.0 ± 6.5 NS 1.1 ± 0.9 NS -7.1 ± 6.3 NS Glycine ±t 5.9 < ± 3.3 < ± 5.0 NS Alanine ± 7.8 < ±t 5.1 < ± 7.6 <0.025 a-aminobutyrate -2.6 ±t < P < ± 0.5 NS -2.6 ±E 1.8 NS Valine ±t 4.3 < ± 1.4 < ± 4.5 NS Cystine ± 2.0 <0.02 Methionine -1.6 ± '< P < ±t 0.5 < ± 1.0 NS Isoleucine -6.7 ± 1.7 < ± < P < ± 1.9 <0.025 Leucine -9.3 ±E 2.5 < ± < P < ± 2.9 <0.05 Tyrosine -5.6 ± 1.2 < ± 0.6 < ± 1.2 <0.02 Phenylalanine -4.0 ±t 1.3 < ± 0.9 < ± 1.9 NS * P = probability that arterial-femoral venous difference does not differ from zero (paired t-test).

3 Proc. Nat. Acad. Sci USA 70 (1973) Amino Acid Transport by Blood Cells 1777 TABLE 3. Arterial-hepatic venous differences of free amino acids in whole blood, plasma, and blood cells (pumol/liter of whole blood) Taurine 1.7 ± 5.4 NS 2.6 -± 1.4 NS -0.5 ± 4.1 NS Aspartate t 6.3 NS Threonine 21.5 i 4.0 < ± 1.7 < i 1.7 <0.05 Serine ± 4.7 < ± 2.6 < ± <P < 0.1 Glutaminet 65.3 ± 8.4 < ± 7.1 < ±t 8.0 NS Proline 9.2 ±t 7.2 NS 11.3 ± 3.5 < ± 8.9 NS Citrulline ±t 8.3 < ± 1.2 < ± < P < 0.1 Glycine 22.2 ± 5.6 < ± 3. 1 < ±t NS Alanine 92.2 ±t 5.9 < ±t 4.8 < ±t 4.3 < a-aminobutyrate -0.7 ± 0.9 NS 0.9 ± 0.5 NS -1.3 ± 0.6 <0. 05 Valine 9.5 ± 4.0 < L 2.8 NS 3.7 ±6 3.7 NS Cystine ± 1.7 < Methionine 4.8 ± 0.7 < ± 0.4 < ± 0.7 <0.01 Isoleucine -0.2 ±4 1.1 NS NS -0.3 ±- 1.7 NS Leucine 4.7 ± 2.3 < NS.3.8 ± 1.4 <0.025 Tyrosine 9.3 ±t 1.2 < ±t 0.6 < ± < P < 0.1 Phenylalanine 6.3 ± 1.4 < ± 0.8 < ±4 1.6 NS * P = probability that arterial-hepatic venous difference does not differ from zero (paired t-test). nine amino acids the concentrations in whole blood were not significantly different from those in plasma. Peripheral amino-acid exchange In Table 2 the arterial-femoral venous (A-FV) differences as determined in whole blood and plasma and the calculated contribution from blood cells to the whole blood A-V differences are shown for each of the amino acids. With whole blood, negative A-FV differences indicating net release from the leg were observed for 12 amino acids. The magnitude of this release was greatest for alanine, which accounted for 35% of the measured amino-acid output. In the case of alanine, methionine, leucine, tyrosine, and glutamine, the A-FV differences in whole blood could not be entirely accounted for by changes in plasma concentrations across the leg. For these amino acids, significant negative A-FV differences were calculated for the cellular compartment of the blood. As in the case of whole blood and plasma, the release of alanine from the leg by way of blood cells exceeded that of all other amino acids. Splanchnic amino-acid exchange Arterial-hepatic venous (A-HV) differences for individual amino acids in whole blood, plasma, and blood cells are shown in Table 3. Analyses of whole blood revealed positive A-V differences for 10 amino acids, indicating a net uptake by the splanchnic bed. The A-HV difference was largest for alanine, which was responsible for 35-40% of the total net amino-acid extraction by the splanchnic tissues. In the case of alanine, threonine, serine, methionine, leucine, and tyrosine, the positive A-HV differences observed with whole blood could not be accounted for solely on the basis of plasma exchange. For these amino acids, the calculated uptakes from blood cells on passage through the splanchnic bed were significant. As with whole blood and plasma, splanchnic extraction of alanine from blood cells was greater than that of all other amino acids. Noteworthy is the consistent splanchnic output of citrulline by way of plasma as well as blood cells. Estimated hepatic blood flow was 1124 ± 48 ml/min, and hepatic plasma flow was 660 ±t 25 ml/min. Portal amino-acid exchange Arterial-portal venous (A-PV) differences are given in Table 4. In whole blood consistently negative A-PV differences, indicating net release from the gut, were observed for nine amino acids. Output of alanine exceeded that of all other amino acids and was responsible for 35% of total amino-acid release from the gut. Exchange via plasma could not account for the total release of alanine, taurine, threonine, methionine, and leucine. For these amino acids, significant negative A-PV differences were calculated for blood cells. A consistently positive A-PV difference was observed only in the case of glutamine. However, the net uptake determined by analysis of whole blood was considerably lower than that observed with plasma, indicating a consistent output of glutamine into blood cells despite a concomitant tissue uptake from plasma on passage through the gut. DISCUSSION The present data provide evidence that amino-acid flux from peripheral tissues and gut to the liver in postabsorptive humans cannot be accounted for solely on the basis of transport by way of plasma. For alanine, threonine, serine, methionine, leucine, isoleucine, tyrosine, and citrulline, inter-organ transport as determined by whole-blood analysis was greater than that observed with plasma determination. For each of these amino acids a significant tissue exchange was calculated to occur via the blood cellular elements, the direction of which paralleled the net shifts observed to occur via plasma. Glutamine, on the other hand, was unique in that its movement on passage of whole blood through the gut was in opposite directions with respect to plasma and blood cells. Thus, while arterial-portal venous differences indicated a tissue uptake of glutamine from plasma, a simultaneous shift of glutamine was calculated to occur into the blood cellular components. Accordingly, the net uptake of glutamine by the gut, though

4 1778 Physiology: Felig et al. Proc. Nat. Acad. Sci. USA 70 (1973) TABLE 4. Arterial-portal venous differences offree amino acids in whole blood, plasma, and blood cells (umol/liter of whole blood) Taurine ± 4.4 < t 3.0 NS ± 4. 8 <0. 05 Aspartate -7.5 ± 5.8 NS - - Threonine ± 1.4 < ± 1.8 NS -8.4 ± 3.8 <0.05 Serine NS -4.0 ± 3.1 NS 7.6 ±t 5.2 NS Glutaminet 38.0 ±15.9 < < ± 9.9 <0.005 Proline -3.0 i 6.9 NS ±: 3.8 < ± 8.0 NS Citrulline ±t 3.9 < ± 2.1 < i 2.8 NS Glycine i 10.7 < ±4.4 < i 9.3 NS Alanine ± 13.0 < ±t 13.2 < ± 7.5 <0.05 a-aminobutyrate -0.8 i 0.9 NS -1.2 ± 0.4 < ± 1.0 NS Valine -9.2 ± 2.7 < ± < P < ± 3.9 NS Cystine NS - Methionine -2.2 ±t 1.0 < ± 0.8 NS -2.0 ± < P < 0.1 Isoleucine NS -1.0 ± 0.9 NS -2.6 ± 2.4 NS Leucine ± 1.8 < ± 1.5 < ± 2.1 <0.025 Tyrosine -2.0 ±2.0 NS -0.2 ± 0.5 NS -1.8 ±t 2.2 NS Phenylalanine -3.2 ± 1.4 < ± < P < ± 2.3 NS * P = probability that arterial-portal venous difference does not differ from zero (paired t-test). significant, was considerably less than that indicated by plasma determinations. Of particular interest is the predominance of alanine in the inter-organ movement of amino acids via plasma as well as blood cells. Alanine exchange accounted for 35-40% of the total amino-acid flux from peripheral tissues and gut to the liver. In the case of each of these tissues, the actual rate of output (from the leg and gut) or uptake (by the splanchnic bed) of alanine was underestimated on the basis of plasma measurements. As indicated by the data in Tables 2-4, transport via blood cells was responsible for 22-32% of the total net movement of this amino acid into or out of whole blood in each of the tissues examined. Prior studies with plasma have emphasized the quantitative importance of alanine in amino-acid flux from muscle to liver (7-9). Since alanine accounts for no more than 7-10% of the amino-acid residues in muscle proteins (16), peripheral synthesis of alanine by means of transamination of glucose-derived pyruvate has been suggested (8). Furthermore, on the basis of the large contribution of alanine to total hepatic amino-acid uptake (7) and its rapid conversion to glucose by perfused liver (17), a glucose-alanine cycle analogous to the Cori cycle for lactate has been proposed (8, 9, 17). The current findings indicate that measurements with plasma underestimate by 25-30% the actual rate of alanine movement from peripheral tissues to the splanchnic bed. Furthermore, these data underscore the unique role of alanine in amino-acid balance between tissues since its calculated transport by way of blood cells exceeded that of all other amino acids. The observations with whole blood also shed light on the relative availability of alanine and glutamine as substrates for hepatic gluconeogenesis. Splanchnic uptake of glutamine from plasma has recently been shown to be comparable to that of alanine (11). However, whereas blood cells add substantially to the total splanchnic uptake of alanine, they do not have a similar role with regard to splanchnic extraction of glutamine (Table 3). In addition, the splanchnic balance data (Table 3) underestimate the actual hepatic uptake of alanine since additional alanine is made available to the liver by the gut (Table 4). In marked contrast, the splanchnic balance data overestimate the net uptake of glutamine by the liver since extrahepatic tissues are responsible for at least a portion of the total splanchnic extraction of this amino acid (Table 4). In fact, comparison of the arterial hepatic venous and arterial-portal venous differences (Tables 3 and 4) indicates that more than 50% of splanchnic glutamine uptake occurs in tissues drained by the portal vein rather than in the liver. The present evidence of an important role for the cellular elements of the blood, presumably erythrocytes, in aminoacid transport would appear to be at variance with in vitro studies demonstrating a slow time constant for amino-acid diffusion across erythrocyte membranes (2). The amino acids whose in vivo transport into and out of tissues involves participation by blood cellular elements, as indicated by the current observations, have previously been demonstrated to require minutes (in the case of leucine, isoleucine, methionine, and tyrosine) to hours (in the case of alanine and serine) to reach equilibrium across the membrane of human erythrocytes when studied in vitro (2). Nevertheless, several recent observations suggest a more rapid transfer rate on passage of blood cells through the capillary bed. Thus Aoki et al., using whole blood and plasma analyses, demonstrated that erythrocytes are important in the transport of glutamate into muscle (4) despite in vitro evidence suggesting virtual impermeability of human erythrocytes to this amino acid (2, 18). Similarly Elwyn et al. observed that erythrocytes contribute to amino-acid exchange across the gut and liver in dogs (5, 6). Several explanations have been advanced to account for the apparent discrepancies between in vivo balance data across tissues and in vitro studies with erythrocytes separated from venous blood. Direct transfer of amino acids between erythrocytes and tissue cells (6), rapid increases in erythrocyte transport velocities upon entry of blood into the capillary bed (6), and rapid in vitro transfer that has been obscured by pre-existing optimal equilibration of erythrocytes separated

5 Proc. Nat. Acad. Sci. USA 70 (1973) from plasma (4) have been postulated. The present results do not permit precise conclusions as to which, if any, of these mechanisms is operative. Nevertheless the data clearly implicate the blood cellular elements as important carriers in the net flux of various amino acids between peripheral tissues, gut, and liver in normal humans. This work was supported by Grant AM from the National Institutes of Health, Grant 19X-3108 from the Swedish Medical Research Council, and a grant from the Tore Nilson Foundation for Medical Research. P.F. is recipient of a Research Career Development Award (AM 70219) from the National Institutes of Health. 1. Munro, H. N. (1970) in Mammalian Protein Metabolism, ed. Munro, H. N. (Academic Press, New York), pp Winter, C. G. & Christensen, H. N. (1964) J. Biol. Chem. 239, Scriver, C. (1971) Amer. J. Clin. Nutr. 24, Aoki, T. T., Brennan, M. F., Muller, W. A., Moore, F. D. & Cahill, G. F., Jr. (1972) J. Clin. Invest. 51, Elwyn, D. H. (1966) Fed. Proc. 25, Amino Acid Transport by Blood Cells 1779 (1. Elwyn, D. H., Launder, W. J., Parikh, H. C. & Wise, E. M., Jr. (1972) Amer. J. Physiol. 222, Felig, P., Owen, 0. E., Wahren, J. & Cahill, G. F., Jr. (1969) J. Clin. Invest. 48, Felig, P., Pozefsky, T., Marliss, E. & Cahill, G. F., Jr. (1970) Science 167, Felig, P. & Wahren, J. (1971) J. Clin. Invest. 50, Spackman, D. H., Stein, W. H. & Moore, S. (1958) Anal. Chem. 30, Marliss, E. B., Aoki, T. T., Pozefsky, T., Most, A. S. & Cahill, G. F., Jr. (1971) J. Clin. Invest. 50, Pagliara, A. S. & Goodman, A. D. (1969) N. Engl. J. Med. 281, Garby, L. & Vuille, J. C. (1961) Scand. J. Clin. Lab. Invest. 13, Bradley, S. E. (1946) in Transactions of the Fifth Conference on Liver Injury (Josiah Macy, Jr. Foundatioh, New York), pp Caesar, J., Shaldon, S., Chiandussi, L., Guevara, L. & Sherlock, S. (1961) Clin. Sci. 21, Kominz, D. R., Hough, A., Symonds, P. & Laki, K. (1954) Arch. Biochem. Biophys. 50, Mallete, L. E., Exton, J. H. & Park, C. R. (1969) J. Biol. Chem. 244, Ussing, H. H. (1943) Acta Physiol. Scand. 5,