0021-972X/91/7304-1106$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright © 1991 by The Endocrine Society
Vol. 73, No. 5 Printed in U.S.A.
Interactions between Oleic Acid and Drug Competitors Influence Specific Binding of Thyroxine in Serum* CHEN-FEE LIM, ANDREA J. CURTIS, JOHN W. BARLOW, DUNCAN J. TOPLISS, AND JAN R. STOCKIGT Ewen Downie Metabolic Unit, and Monash University Department of Medicine, Alfred Hospital, Melbourne, Victoria 3181, Australia
ABSTRACT. Long chain nonesterified fatty acids and various drugs may share albumin-binding sites in common. We questioned whether serum binding of T4 could be indirectly influenced by displacement of drug competitors from these sites by nonesterified fatty acids. The influence of oleic acid on druginduced inhibition of [125I]T4 binding was measured by equilibrium dialysis, using undiluted serum in order to avoid dilutionrelated artefacts. Oleic acid (1 mmol/L) alone did not inhibit serum protein binding of T4, but this concentration augmented the inhibitory effects on T4 binding of diflunisal, mefenamic acid, meclofenamic acid, and aspirin. This effect increased with increasing concentrations of mefenamic acid, meclofenamic acid, and furosemide. The T4-displacing effect of fenclofenac was not augmented by oleic acid. The mechanism of these interactions was studied by examining 1) oleic acid effects on drug binding, and 2) drug effects on oleic acid binding in undiluted serum. Increments in added oleic acid (0.5-2.0 mmol/L) progressively
T
HE FREE fraction of T 4 may be abnormally high in critically ill patients, consistent with the presence of circulating inhibitors of T4 binding (1-5), but despite detailed study, no single competitor has been convincingly identified. Various drugs (6, 7) and nonesterified fatty acids (NEFA) (8, 9) that can inhibit T4 binding have usually been considered alone, but since they may share common binding sites on albumin (10), we postulated that interactions between them could influence hormone binding synergistically. With respect to the inhibition of T4 binding by NEFA in normal serum, recent studies (6, 11) suggest that the total NEFA concentrations required to displace T4 are outside the pathophysiological range. The direct affinities of important drug competitors for T4-binding globulin (TBG) range
Received February 11,1991. Address all correspondence and requests for reprints to: J. R. Stockigt, M.D., Ewen Downie Metabolic Unit, Alfred Hospital, Commercial Road, Melbourne, Victoria 3181, Australia. * Presented in part at the 33rd Annual Scientific Meeting of the Endocrine Society of Australia, Perth, September 1990, and at the 10th International Thyroid Congress, The Hague, The Netherlands, February 1991. This work was supported in part by Project Grants 850366 and 880859 from the National Health and Medical Research Council of Australia and an equipment grant from the Alfred Hospital WholeTime Medical Specialists' Private Practice Scheme.
increased the mean unbound fractions of [14C]aspirin, [14C] diflunisal, and [14C]furosemide, but did not displace ["CJfenclofenac. At the relevant total and free drug concentrations, the inhibitory effect of oleic acid on drug binding and its influence on drug-induced displacement of T4 were concordant in the order: meclofenamic acid > aspirin > mefenamic acid > diflunisal > furosemide > fenclofenac. In contrast, drug-induced increases in the unbound fraction of [uC]oleic acid did not correlate with augmentation of T4 displacement. We conclude that synergistic effects of oleic acid and drugs on T4 binding result from drug displacement by oleic acid, rather than the reverse effect. Hence, substances that increase the unbound concentration of a competitor by displacing it from albumin can increase its T4-displacing potency. Interactions between various ligands may exert a greater hormone-displacing effect than the sum of each alone. (J Clin Endocrinol Metab 73: 1106-1110,1991)
from 3 orders of magnitude less (furosemide) to about 7 orders of magnitude less than T4 itself (aspirin) (7). However, such data do not reflect competition in vivo, which depends on the relative free circulating concentrations of hormone and competitor. The free concentrations of acidic drugs are determined by their binding to serum proteins, especially the high capacity binding sites of albumin. Occupancy of these sites by any ligand, endogenous or exogenous, may influence the free concentration of competitor. As serum is progressively diluted in vitro it becomes virtually impossible to maintain the concentrations of free hormone, free competitor(s), and unoccupied binding sites in the relationship that applies in vivo (12). Hence, we have studied binding of T4, drugs, and oleic acid in the mixed system of multiple sites and interacting ligands that is present in undiluted serum, using an equilibrium dialysis system that allows drug competitors to be added so as to achieve predetermined serum concentrations at equilibrium (6) based on the known unbound fraction of each drug (6,13). We also studied drug effects on serum binding of [14C] oleic acid by heptane partition (14) after preliminary purification of tracer (15). The six drugs chosen for study were those found to
1106
DRUG AND NEFA EFFECTS ON T4 BINDING be the most potent competitors for serum T 4 binding in undiluted serum at their relevant therapeutic concentrations (6).
Materials and Methods Drugs and reagents Aspirin was purchased from Ajax Chemicals (Melbourne, Australia), diflunisal from Merck, Sharp, and Dohme (Sydney, Australia), fenclofenac from Reckitt and Colman (Hull, England), and furosemide from Hoechst (Melbourne, Australia). Labeled preparations of drugs were gifts from the following sources: [14C]furosemide (9.8 mCi/mmol), Hoechst (Frankfurt, Germany); [UC]fenclofenac (13.6 mCi/mmol), Reckitt and Colman; and [14C]diflunisal (13.1 mCi/mmol), Merck, Sharp, and Dohme. [14C]Aspirin (40 mCi/mmol) was purchased from New England Corp. (Boston, MA.) [125I]T4 (40-60 /tCi/jig) and [14C] oleic acid (54-59 mCi/mmol) were obtained from Amersham International (Aylesbury, Buckinghamshire, United Kingdom). Nonradioactive oleic acid (99% pure) was obtained from Aidrich (Milwaukee, WI). n-Heptane of spectroscopic grade was purchased from E. Merck (Darmstadt, Germany). Antiinflammatory drugs were dissolved in 100% ethanol; the final ethanol concentration in the dialysate was less than 0.05%, which had no effect on [125I]T4 binding (6). Furosemide was dissolved in 0.04 mmol/L Tris-Cl buffer, pH 7.4 (at 37 C), containing physiological concentrations of K+, Ca2+, Mg2+, and Cl (16); this buffer was used for all studies. The test serum used was a pool from untreated euthyroid normal subjects. The total T4, binding protein, and total NEFA concentrations of this serum pool were: T4, 95 nmol/L; TBG, 20 mg/L; transthyretin, 280 mg/L; albumin, 41 g/L; and total NEFA, 0.7 mmol/L, giving a total NEFA/albumin molar ratio of about 1.2:1. Hormone and protein concentrations were measured as previously described (17), and total NEFA was measured by an enzymatic colorimetric method (Wako Industries, Osaka, Japan).
Measurement of free T4 fraction We measured the ability of each drug to displace [125I]T4 by equilibrium dialysis of undiluted serum, as previously described (6), using 20 mL buffer and 0.25 mL serum containing about 300,000 cpm [125I]T4 (40-60 >iCi/ng) in double knotted 10-mm dialysis tubing (no. 453103, Union Carbide, New York, NY). Serum was dialysed for 16-18 h at 37 C in quadruplicate in 20 mL glass scintillation vials; mixing was performed by intermittent rotation. Drugs were added to the buffer compartment in amounts calculated to achieve predetermined total and free drug concentrations in the serum compartment at equilibrium (6,18,19) based on the measured free fraction of each drug (6, 13). Oleic acid was added directly to serum as 2-5 jiL/mL of a 0.1-1.0 mol/L solution in pure ethanol, as previously described (6). During dialysis to equilibrium, less than 1% of the added oleic acid entered the dialysate, consistent with its very high degree of albumin binding (14,15). The free fraction of T4 was calculated from dialysate counts after MgCl2 precipitation (20). All dialyzable compounds were shown to reach equilibrium in this system within 12 h at 37 C.
1107
Measurement of the free drug fraction u
C-Labeled drug preparations
Poorly bound tracer contaminants were removed by adding the labeled compound to serum, followed by preliminary dialysis against buffer (21). Briefly, 1-2 /*Ci labeled drug in 0.5 mL pooled normal serum was dialyzed overnight at 4 C against 0.5-2 L Tris-Cr buffer, pH 7.4. The sac contents containing protein-bound tracer were diluted for binding measurements. To assess its effect on drug binding in undiluted serum, oleic acid (0.5-2.0 mmol/L) and 14C-labeled drugs (
Vol. 73, No. 5 Printed in U.S.A.
Interactions between Oleic Acid and Drug Competitors Influence Specific Binding of Thyroxine in Serum* CHEN-FEE LIM, ANDREA J. CURTIS, JOHN W. BARLOW, DUNCAN J. TOPLISS, AND JAN R. STOCKIGT Ewen Downie Metabolic Unit, and Monash University Department of Medicine, Alfred Hospital, Melbourne, Victoria 3181, Australia
ABSTRACT. Long chain nonesterified fatty acids and various drugs may share albumin-binding sites in common. We questioned whether serum binding of T4 could be indirectly influenced by displacement of drug competitors from these sites by nonesterified fatty acids. The influence of oleic acid on druginduced inhibition of [125I]T4 binding was measured by equilibrium dialysis, using undiluted serum in order to avoid dilutionrelated artefacts. Oleic acid (1 mmol/L) alone did not inhibit serum protein binding of T4, but this concentration augmented the inhibitory effects on T4 binding of diflunisal, mefenamic acid, meclofenamic acid, and aspirin. This effect increased with increasing concentrations of mefenamic acid, meclofenamic acid, and furosemide. The T4-displacing effect of fenclofenac was not augmented by oleic acid. The mechanism of these interactions was studied by examining 1) oleic acid effects on drug binding, and 2) drug effects on oleic acid binding in undiluted serum. Increments in added oleic acid (0.5-2.0 mmol/L) progressively
T
HE FREE fraction of T 4 may be abnormally high in critically ill patients, consistent with the presence of circulating inhibitors of T4 binding (1-5), but despite detailed study, no single competitor has been convincingly identified. Various drugs (6, 7) and nonesterified fatty acids (NEFA) (8, 9) that can inhibit T4 binding have usually been considered alone, but since they may share common binding sites on albumin (10), we postulated that interactions between them could influence hormone binding synergistically. With respect to the inhibition of T4 binding by NEFA in normal serum, recent studies (6, 11) suggest that the total NEFA concentrations required to displace T4 are outside the pathophysiological range. The direct affinities of important drug competitors for T4-binding globulin (TBG) range
Received February 11,1991. Address all correspondence and requests for reprints to: J. R. Stockigt, M.D., Ewen Downie Metabolic Unit, Alfred Hospital, Commercial Road, Melbourne, Victoria 3181, Australia. * Presented in part at the 33rd Annual Scientific Meeting of the Endocrine Society of Australia, Perth, September 1990, and at the 10th International Thyroid Congress, The Hague, The Netherlands, February 1991. This work was supported in part by Project Grants 850366 and 880859 from the National Health and Medical Research Council of Australia and an equipment grant from the Alfred Hospital WholeTime Medical Specialists' Private Practice Scheme.
increased the mean unbound fractions of [14C]aspirin, [14C] diflunisal, and [14C]furosemide, but did not displace ["CJfenclofenac. At the relevant total and free drug concentrations, the inhibitory effect of oleic acid on drug binding and its influence on drug-induced displacement of T4 were concordant in the order: meclofenamic acid > aspirin > mefenamic acid > diflunisal > furosemide > fenclofenac. In contrast, drug-induced increases in the unbound fraction of [uC]oleic acid did not correlate with augmentation of T4 displacement. We conclude that synergistic effects of oleic acid and drugs on T4 binding result from drug displacement by oleic acid, rather than the reverse effect. Hence, substances that increase the unbound concentration of a competitor by displacing it from albumin can increase its T4-displacing potency. Interactions between various ligands may exert a greater hormone-displacing effect than the sum of each alone. (J Clin Endocrinol Metab 73: 1106-1110,1991)
from 3 orders of magnitude less (furosemide) to about 7 orders of magnitude less than T4 itself (aspirin) (7). However, such data do not reflect competition in vivo, which depends on the relative free circulating concentrations of hormone and competitor. The free concentrations of acidic drugs are determined by their binding to serum proteins, especially the high capacity binding sites of albumin. Occupancy of these sites by any ligand, endogenous or exogenous, may influence the free concentration of competitor. As serum is progressively diluted in vitro it becomes virtually impossible to maintain the concentrations of free hormone, free competitor(s), and unoccupied binding sites in the relationship that applies in vivo (12). Hence, we have studied binding of T4, drugs, and oleic acid in the mixed system of multiple sites and interacting ligands that is present in undiluted serum, using an equilibrium dialysis system that allows drug competitors to be added so as to achieve predetermined serum concentrations at equilibrium (6) based on the known unbound fraction of each drug (6,13). We also studied drug effects on serum binding of [14C] oleic acid by heptane partition (14) after preliminary purification of tracer (15). The six drugs chosen for study were those found to
1106
DRUG AND NEFA EFFECTS ON T4 BINDING be the most potent competitors for serum T 4 binding in undiluted serum at their relevant therapeutic concentrations (6).
Materials and Methods Drugs and reagents Aspirin was purchased from Ajax Chemicals (Melbourne, Australia), diflunisal from Merck, Sharp, and Dohme (Sydney, Australia), fenclofenac from Reckitt and Colman (Hull, England), and furosemide from Hoechst (Melbourne, Australia). Labeled preparations of drugs were gifts from the following sources: [14C]furosemide (9.8 mCi/mmol), Hoechst (Frankfurt, Germany); [UC]fenclofenac (13.6 mCi/mmol), Reckitt and Colman; and [14C]diflunisal (13.1 mCi/mmol), Merck, Sharp, and Dohme. [14C]Aspirin (40 mCi/mmol) was purchased from New England Corp. (Boston, MA.) [125I]T4 (40-60 /tCi/jig) and [14C] oleic acid (54-59 mCi/mmol) were obtained from Amersham International (Aylesbury, Buckinghamshire, United Kingdom). Nonradioactive oleic acid (99% pure) was obtained from Aidrich (Milwaukee, WI). n-Heptane of spectroscopic grade was purchased from E. Merck (Darmstadt, Germany). Antiinflammatory drugs were dissolved in 100% ethanol; the final ethanol concentration in the dialysate was less than 0.05%, which had no effect on [125I]T4 binding (6). Furosemide was dissolved in 0.04 mmol/L Tris-Cl buffer, pH 7.4 (at 37 C), containing physiological concentrations of K+, Ca2+, Mg2+, and Cl (16); this buffer was used for all studies. The test serum used was a pool from untreated euthyroid normal subjects. The total T4, binding protein, and total NEFA concentrations of this serum pool were: T4, 95 nmol/L; TBG, 20 mg/L; transthyretin, 280 mg/L; albumin, 41 g/L; and total NEFA, 0.7 mmol/L, giving a total NEFA/albumin molar ratio of about 1.2:1. Hormone and protein concentrations were measured as previously described (17), and total NEFA was measured by an enzymatic colorimetric method (Wako Industries, Osaka, Japan).
Measurement of free T4 fraction We measured the ability of each drug to displace [125I]T4 by equilibrium dialysis of undiluted serum, as previously described (6), using 20 mL buffer and 0.25 mL serum containing about 300,000 cpm [125I]T4 (40-60 >iCi/ng) in double knotted 10-mm dialysis tubing (no. 453103, Union Carbide, New York, NY). Serum was dialysed for 16-18 h at 37 C in quadruplicate in 20 mL glass scintillation vials; mixing was performed by intermittent rotation. Drugs were added to the buffer compartment in amounts calculated to achieve predetermined total and free drug concentrations in the serum compartment at equilibrium (6,18,19) based on the measured free fraction of each drug (6, 13). Oleic acid was added directly to serum as 2-5 jiL/mL of a 0.1-1.0 mol/L solution in pure ethanol, as previously described (6). During dialysis to equilibrium, less than 1% of the added oleic acid entered the dialysate, consistent with its very high degree of albumin binding (14,15). The free fraction of T4 was calculated from dialysate counts after MgCl2 precipitation (20). All dialyzable compounds were shown to reach equilibrium in this system within 12 h at 37 C.
1107
Measurement of the free drug fraction u
C-Labeled drug preparations
Poorly bound tracer contaminants were removed by adding the labeled compound to serum, followed by preliminary dialysis against buffer (21). Briefly, 1-2 /*Ci labeled drug in 0.5 mL pooled normal serum was dialyzed overnight at 4 C against 0.5-2 L Tris-Cr buffer, pH 7.4. The sac contents containing protein-bound tracer were diluted for binding measurements. To assess its effect on drug binding in undiluted serum, oleic acid (0.5-2.0 mmol/L) and 14C-labeled drugs (
