Synthesis and preliminary evaluation of 18F-labeled 4-thia palmitate as a PET tracer of myocardial fatty acid oxidation

Nuclear Medicine & Biology, Vol. 27, pp. 221–231, 2000 Copyright © 2000 Elsevier Science Inc. All rights reserved.

ISSN 0969-8051/00/$–see front matter PII S0969-8051(99)00101-8

Synthesis and Preliminary Evaluation of 18F-labeled 4-Thia Palmitate as a PET Tracer of Myocardial Fatty Acid Oxidation Timothy R. DeGrado,1 Shuyan Wang,1 James E. Holden,2 R. Jerome Nickles,2 Michael Taylor2 and Charles K. Stone2 1

DEPARTMENT OF RADIOLOGY, DUKE UNIVERSITY MEDICAL CENTER, DURHAM, NORTH CAROLINA USA; AND 2DEPARTMENTS OF MEDICAL PHYSICS AND CARDIOLOGY, UNIVERSITY OF WISCONSIN-MADISON, MADISON, WISCONSIN, USA

ABSTRACT. Interest remains strong for the development of a noninvasive technique for assessment of regional fatty acid oxidation rate in the myocardium. 18F-labeled 4-thia palmitate (FTP, 16-[18F]fluoro-4thia-hexadecanoic acid) has been synthesized and preliminarily evaluated as a metabolically trapped probe of myocardial fatty acid oxidation for positron emission tomography (PET). The radiotracer is synthesized by Kryptofix 2.2.2/K2CO3 assisted nucleophilic radiofluorination of an iodo-ester precursor, followed by alkaline hydrolysis and by purification by reverse phase high performance liquid chromatography. Biodistribution studies in rats showed high uptake and long retention of FTP in heart, liver, and kidneys consistent with relatively high fatty acid oxidation rates in these tissues. Inhibition of carnitine palmitoyl-transferase-I caused an 80% reduction in myocardial uptake, suggesting the dependence of trapping on the transport of tracer into the mitochondrion. Experiments with perfused rat hearts showed that the estimates of the fractional metabolic trapping rate (FR) of FTP tracked inhibition of oxidation rate of palmitate with hypoxia, whereas the FR of the 6-thia analog 17-[18F]fluoro-6-thia-heptadecanoic acid was insensitive to hypoxia. In vivo defluorination of FTP in the rat was evidenced by bone uptake of radioactivity. A PET imaging study with FTP in normal swine showed excellent myocardial images, prolonged myocardial retention, and no bone uptake of radioactivity up to 3 h, the last finding suggesting a species dependence for defluorination of the omega-labeled fatty acid. The results support further investigation of FTP as a potential PET tracer for assessing regional fatty acid oxidation rate in the human myocardium. NUCL MED BIOL 27;3:221–231, 2000. © 2000 Elsevier Science Inc. All rights reserved. KEY WORDS. Fatty acids, ␤-Oxidation, PET, Thia fatty acids

INTRODUCTION Interest has remained strong for the development of a noninvasive technique for regional assessment of ␤-oxidation of long-chain fatty acids. ␤-Oxidation is depended upon for the majority of adenosine triphosphate production in the heart in normal conditons. Noninvasive assessment of regional ␤-oxidation rate may allow early detection of metabolic abnormalities in the myocardium that may presage irreversible tissue injury. However, the development of a measurement technique using a gamma-emitter labeled fatty acid analog has proved problematic in spite of extensive efforts with numerous tracers. [1-11C]Palmitate (CPA) has been studied with positron emission tomography (PET) in animals and humans as a tracer of myocardial fatty acid metabolism. In nonischemic conditions, washout of the oxidative product 11CO2 predominates the early clearance phase, whereas turnover of CPA incorporated into complex lipids determines the slower late clearance phase. The clearance rate and the relative size of the rapid phase correlate with cardiac work (11) and myocardial oxygen consumption rate (27). Noninvasive delineation of regional fatty acid oxidation rates from dynamic PET imaging Address correspondence to: Timothy R. DeGrado, Ph.D., Duke University Medical Center, Dept. of Radiology, Box 3949, Durham, NC 27710, USA; e-mail: [email protected]. Received 17 July 1999. Accepted 15 October 1999.

studies with CPA were proposed on the basis of these findings. A relatively complex tracer kinetic model is necessary to model CPA kinetics for quantitative measurement of fatty acid utilization that includes catabolite and lipid storage compartments (13). Bergmann et al. (2) developed a seven-parameter compartmental modeling technique for quantification of CPA kinetics that was in good agreement with measured arteriovenous differences of fatty acids across the canine heart in nonischemic conditions. The modeling technique may not be applicable in ischemic regions due to low signal-to-noise ratio in the images. Furthermore, the enhanced back-diffusion of extracted nonmetabolized tracer from ischemic myocardium is kinetically indistinct from clearance of 11CO2 during the early clearance phase (10). In this situation, the kinetics of the early clearance phase do not quantitatively reflect the oxidative rate of fatty acids by the myocardium. Structurally modified fatty acid tracers have been developed to prolong the retention of oxidative metabolites in tissues, allowing the application of more simplified imaging protocols and data analysis strategies. 11C-labeled ␤-methyl heptadecanoic acid (BMHA) was developed as a metabolically trapped tracer (9, 19). BMHA showed prolonged retention in myocardium due to incorporation mainly into triglycerides (1). Likewise, myocardial retention of a radioiodinated ␤-methylated fatty acid was found to be insensitive to decreases of ␤-oxidation rate secondary to pharmacologic inhibition of carnitine palmitoyltransferase-I (CPT-I) (5). The 6-thia fatty acid analog 14(R,S)-[18F]fluoro-6-thia-heptade-

T. R. DeGrado et al.

222

FIG. 1. Chemical structures of 18 F-labeled thia fatty acid analogs. The (␻-3) labeled, 6-thia analog 14(R,S)-[18F]fluoro-6-thiaheptadecanoic acid (FTHA) has been previously developed (6). FTP, 18F-labeled 4-thia palmitate; 17F6THA, 17-[18F]fluoro6-thia-heptadecanoic acid.

canoic acid (FTHA; Fig. 1) was developed as a metabolically trapped tracer of ␤-oxidation in myocardium-based (4, 7) pharmacology studies with 4-thia fatty acids (31). 4-Thia fatty acids are ␤-oxidized in the mitochondria to 4-thia-enoyl-CoAs, which apparently accumulate in the mitochondrion. Initial studies with FTHA have shown it to be metabolically retained in the myocardium and have excellent imaging properties in normal human subjects (8, 21) and in patients with coronary artery disease (28). Experimental validation studies have shown good correspondence of trapping of FTHA in the myocardium with fatty acid oxidation in various conditions; however, FTHA accumulation in the myocardium failed to track decreases in ␤-oxidation in hypoxic conditions (35). Evidently, ␤-oxidation is not solely responsible for intracellular retention of the radiolabel and the hypoxic condition presumably exposes retention of radiolabel in acyl intermediates or complex lipids. The suboptimal specificity of FTHA for ␤-oxidation motivated further radiotracer development, specifically 4-thia fatty acid analogs. The present study reports the synthesis and preliminary evaluation of 18F-labeled 4-thia palmitate (FTP) as a myocardial ␤-oxidation probe for PET. METHODS AND MATERIALS

ice-diluted HCl and the solution was extracted twice with ether (20 mL). The combined ether fractions were washed successively with dilute NaHCO3, water, and brine; dried (MgSO4); and evaporated under reduced pressure. The product 1 (0.84 g, 32% yield) was isolated by column chromatography (hexane/ethyl acetate 9:1) and recrystallized in MeOH. TLC (hexane/ethyl acetate 9:1) Rf ⫽ 0.4. Melting point ⫽ 33°C. 1H-NMR 1.4 (m, 20H, CH2), 2.52 (t, 2H, C(2)H2), 2.61 (t, 2H, C(5)H2), 2.79 (t, 2H, C(3)H2), 3.20 (t, 2H, C(16)H2), 3.71 (s, 3H, -COO-CH3).

16-Fluoro-4-thia-hexadecanoic Acid (2) Fluorination of 1 (0.2 g, 0.5 mmol) was performed by addition of tetrabutylammonium fluoride (TBAF; 4 mmol as 1 M tetrahydrofuran solution). The mixture was allowed to react at room temperature for 20 h. The resulting flouro-ester was hydrolyzed by addition of 1 mL 1 N potassium hydroxide (KOH) and 1 mL ethanol for 2 h at room temperature. The product 2 (0.05 g, 34% yield) was isolated by column chromatography (hexane:ether:acetic acid, 3:1:0.1) and recrystallized in hexane. TLC (hexane/ether/acetic acid 3:1:0.1) Rf ⫽ 0.25. Melting point ⫽ 59°C. 1H-NMR 1.4 (m, 20H, CH2), 2.55 (t, 2H, C(2)H2), 2.65 (t, 2H, C(5)H2), 2.80 (t, 2H, C(3)H2), 4.44 (dt, 2H, C(16)H2, JHF ⫽ 47.5, JHH ⫽ 6.2 Hz).

Synthesis of Radiotracers Chemical structures are shown in Figure 1 for the 18F-labeled thia fatty acids that were synthesized and evaluated: FTP and 17[18F]fluoro-6-thia-heptadecanoic acid (17F6THA). Chemicals were of analytical grade. Dry acetonitrile was obtained commercially (Pierce, Rockford, IL USA). 1H-Nuclear magnetic resonance (NMR) spectra were recorded with a Varian (Palo Alto, CA USA) Unity 500 MHz spectrometer using CDCl3 as a solvent (Me4Si, 0.00 ppm). Rf values refer to thin layer chromatography (TLC) performed on silica gel with the solvent system noted. Routine column chromatography was performed under normal pressure with silica gel (100 –200 mesh) and the solvent system noted.

Methyl 16-iodo-4-thia-hexadecanoate (1) 1,12-Diiodododecane (2.66 g, 6.3 mmol) was dissolved in 20 mL acetonitrile. Methyl 3-mercaptopropionate (0.76 g, 6.3 mmol) and K2CO3 (1.1 g, 8 mmol) were added and the mixture was allowed to react at room temperature for 20 h. The mixture was acidified with

Methyl 17-tosyloxy-6-thia-heptadecanoate (3) To a solution of 11-bromo-1-undecanol (10 g, 40 mmol) in dimethyl sulfoxide (DMSO; 80 mL) was added thiourea (3.64 g, 48 mmol), and the mixture was allowed to react at room temperature for 21 h. The mixture was extracted twice with hexane (20 mL) to remove the unreacted bromo alcohol. To the DMSO fraction was added 2 N KOH (50 mL) and the mixture was heated at 80°C for 5 min, releasing the thiol 11-mercapto-undecanol. The mixture was acidified (HCl) and extracted twice with ether (40 mL). The combined ether phases were washed successively with water and brine, dried over MgSO4, and evaporated under reduced pressure. The resulting thiol (approximately 7.2 g) was made to react with methyl 5-bromo-pentanoate (7 g, 37 mmol) according to the procedure for compound 1. Crystallization of the product in hexane yielded 5.5 g (18 mmol, 48% yield from the thiol) of methyl 17-hydroxy-6-thia-heptadecanoate. TLC (hexane:ether, 1:1) showed a single product at Rf ⫽ 0.3. Without further characterization, the hydroxy-ester (4 g, 13 mmol) was caused to react with Ts-Cl (2.7 g, 14 mmol) and pyridine (14 mmol) in CH2Cl2 (40 mL)

18

F-labeled 4-thia Palmitate in Heart

223

TABLE 1. Semi-preparative Reverse Phase HPLC Capacity Factors (kⴕ) of Fatty Acid Analogs (Nucleosil, C-18 (10␮) 250 ⴛ 10 mm, Flow ⴝ 4.3 ml/min, Mobile Phase MeOH/H2O/AcOH 90:9.5:0.5) Compound

kⴕ

16-Iodo-4-thia-hexadecanoic acid 16-Fluoro-4-thia-hexadecanoic acid (2, FTP) 17-Tosyloxy-6-thia-heptadecanoic acid 17-Fluoro-6-thia-heptadecanoic acid (4,17F6THA)

5.1 3.2 3.0 3.5

at 5°C for 4 h. The mixture was acidified with ice-diluted HCl, and the organic layer separated, dried with MgSO4, and evaporated under reduced pressure. The product 3 (2.4 g, 34% yield) was isolated by column chromatography (hexane:ethyl acetate, 7:3) and recrystallized in hexane. TLC (hexane:ethyl acetate, 7:3) Rf ⫽ 0.6. Melting point ⫽ 54°C. 1H-NMR 1.3 (m, 22H, CH2), 2.38 (t, 2H, C(2)H2), 2.5 (m, 7H, C(5)H2, C(7)H2, (tosyl) CH3), 3.7 (s, 3H, -COO-CH3), 4.0 (t, 2H, C(17)H2), 7.6 (m, 4H, aryl).

17-Fluoro-6-thia-heptadecanoic Acid (4) Compound 3 (1 g, 1.8 mmol) was caused to react with TBAF (2.5 mmol) according to the procedure for compound 2. The resultant fluoro-ester was hydrolyzed in KOH/ethanol as previously described and crystallized twice in ethanol/water to yield 0.4 g (70% yield) of product 4. Melting point ⫽ 54°C. 1H-NMR 1.3 (m, 22H, CH2), 2.38 (t, 2H, C(2)H2), 2.5 (m, 4H, C(5)H2, C(7)H2), 4.44 (dt, 2H, C(17)H2, JHF ⫽ 47.3, JHH ⫽ 6.2 Hz).

(75 mg/kg) before injection of radiotracer, remaining anesthetized throughout the study. The 18F-labeled radiotracer (20 – 40 ␮Ci) was injected into the femoral vein. A prescribed duration of time was allowed before procurement of heart, liver, lung, blood, kidney, bone (rib), brain (whole), and skeletal muscle. The tissues were counted and weighed. Radiotracer uptake was calculated as:

Uptake (%dose kg/g) ⫽

CPM(tissue) ⫻ Body Wt.(kg) ⫻ 100 Tissue Wt.(g) ⫻ CPM (dose)

(1)

Nonesterified fatty acid (NEFA) concentrations in plasma at time of euthanasia were measured colorimetrically (WACO Bioproducts, Richmond,VA USA). In one group of fasted rats, the CPT-I inhibitor etomoxir (40 mg/kg, Byk Gulden, Konstanz, Germany) was administered intrperitoneally 2 h prior to FTP injection. Crude analysis of the nature of the metabolites in the heart was performed by a Folch-type extraction procedure as previously described (7). The rat heart was excised 30 min postinjection and thoroughly homogenized and sonicated (20 s) in 7 mL chloroform/ methanol (2:1) at 0°C. Urea (40%, 1.75 mL) and 5% sulfuric acid (1.75 mL) were added and the mixture was sonicated for an additional 20 s. After centrifugation for 10 min, aqueous, organic, and protein interphase fractions were separated and counted. The organic phase was further analyzed by silica-gel TLC for radiolabeled diglycerides, fatty acids, triglycerides, and cholesterol ester as previously described (7).

18

Isolation and Perfusion of Rat Hearts

The precursors for FTP and 17F6THA were compounds 1 and 3, respectively. To a 2 mL glass vial was added Kryptofix 2.2.2 (10 mg), acetonitrile (0.5 mL), and 20 ␮L of a 9% K2CO3 solution in water. [18F]Fluoride, which was produced via proton bombardment of H18 2 O (⬎95 atom %), was then added, the vessel placed in an aluminum heating block at 85°C, and the solvent evaporated under a stream of helium or nitrogen. The residue was further dried by azeotropic distillation with acetonitrile (2 ⫻ 0.3 mL). A solution of the precursor (approximately 2 mg) in acetonitrile (0.5 mL) was added, and the vial was sealed and returned to the heating block. Reaction time was 15 min. The vial was briefly cooled by placing in ice water. The incorporation of [18F]fluoride was monitored by radio-TLC (hexane/ethyl acetate 3:1). Rf values were 0.0 and ⬎0.4 for [18F]fluoride and [18F]fluoro-ester, respectively. Subsequent hydrolysis of the resulting [18F]fluoro-ester was performed in the same vessel by the addition of 0.15 mL 0.2N KOH and continued heating at 90°C for 4 min. The mixture was cooled, acidified with concentrated acetic acid (25 ␮L), filtered, and applied to the preparative high performance liquid chromatography (HPLC) column (Table 1). An in-line ultraviolet detector (210 nm) was used to monitor the elution of unlabeled materials. The [18F]fluoro-fatty acid fraction was collected, evaporated to dryness, formulated in 1–2% albumin in isotonic NaCl solution, and filtered through a 0.22 ␮m filter (Millex-GS, Millipore, Bedford, MA USA).

Hearts were excised from pentobarbital anesthetized female Sprague-Dawley rats (225–250 g) that had been fed ad libitum. Following cannulation of the aorta, retrograde (Langendorff) perfusion of the hearts was commenced using a peristaltic pump. The perfusion medium was Krebs-Henseleit (K-H) bicarbonate buffer of the following composition (mmol/L): Na⫹ 143; K⫹ 5.9; Ca2⫹ 1.85; ⫺ Mg2⫹ 1.0; Cl⫺ 125.6; SO2⫺ 1.18; H2PO⫺ 4 4 1.2; HCO3 25, glucose 10; palmitate 0.15; and albumin 0.15. The perfusate was prefiltered using 5 ␮m in-line filters (Millipore Corp., Bedford, MA USA). Under standard conditions, the perfusate was gassed with a 95%/5% oxygen (O2)/carbon dioxide CO2 mixture. To assess the tracer kinetics in conditions of hypoxia, the fraction of oxygen was lowered to 35% while maintaining the CO2 fraction at 5%. The makeup gas was nitrogen. The apparatus utilized water-jacketed vessels and a heater/circulator to deliver the medium to the heart at a temperature of 37°C. Aortic pressure was monitored by a pressure transducer and allowed measurement of heart rate and mean aortic pressure. Hearts were not externally paced. Coronary flow was measured manually.

F-Labeling Procedure

Biodistribution Studies in Rats Female Sprague-Dawley rats (180 –225 g) were fasted overnight or fed ad libitum. The rats were anesthetized with pentobarbital

Bolus Injection of Radiotracer For experiments using bolus administration of 18F-labeled fatty acids, a catheter (PE-30 tubing) was inserted into the aortic cannula. The opening of the catheter was situated near the root of the coronary artery of the perfused rat heart. The standard protocol for bolus injection studies consisted of a 20-min stabilization period, after which the radiotracer (approximately 5 ␮Ci in 50 ␮L) was manually injected through the catheter in less than 0.5 s. A 20-min washout period followed.

T. R. DeGrado et al.

224

Pulse-perfusion of Radiotracer For pulse-perfusion of radiotracer, two reservoirs of perfusate were employed. Before each study, one reservoir was filled with 500 mL of perfusion medium containing 100 –200 ␮Ci/L 18F-labeled fatty acid and approximately 50 ␮Ci/L [9,10-3H]palmitate. A three-way valve was employed for switching the perfusate between the two sources. The standard pulse-perfusion protocol consisted of a 20-min stabilization period without tracers, a 20-min perfusion with radiotracers, and a 30 – 40-min washout period (without tracers) in succession. The 18F radioactivity concentration of the wash-in medium Cp (CPM/mL) was measured in a NaI well-counter. At the end of perfusion, ventricles were opened and gently blotted on paper toweling. Hearts were then placed in preweighed tubes, weighed, and counted in the NaI well-counter. As a consequence of the prolonged washout period, no radioactivity remained in the left heart chambers or perfusion cannulae at the end of perfusion. The apparent distribution volume (ADV) of radioactivity in the whole heart was calculated from the sum of radioactivity in the heart and the paper toweling (Ah), the dry mass of the heart (Mh), and Cp: ADV (tend) ⫽ Ah/(Cp ⫻ Mh)

(2)

where tend was the time at the end of perfusion.

Estimation of Palmitate Oxidation Rate Hearts were perfused with [9,10-3H]palmitate at a concentration of 50 ␮Ci/L. Arterial samples were collected from the reservoir used to perfuse the heart. Coronary effluent samples were collected after a 10 min equilibration period. Tritiated water was separated from nonmetabolized 3H-palmitate in the coronary samples by an organic solvent extraction procedure described by Saddik and Lopaschuk (25). Fractional palmitate oxidation rate (mL/min/g dry) was calculated as the product of mass-specific coronary flow (mL/min/g dry) and the fraction of coronary effluent 3H2O concentration relative to perfusate 3H-palmitate concentration at steady-state. Palmitate oxidation rate (nmol/min/g dry) was calculated by multiplying the fractional rate by the concentration of palmitate in the perfusate (nmol/mL).

Acquisition and Normalization of Curves in Isolated Rat Hearts

18

F Time-activity

18

F Radioactivity was externally monitored using two lead-collimated 2 ⫻ 2 inch NaI(Tl) scintillation probes configured electronically for ␥,␥-coincidence detection. The true coincidence count rate in the heart (Yi [counts/s]) was calculated as the difference of prompt coincidence rate and random coincidence rate, the latter of which was estimated by the delayed signal technique. The true coincidence count rate was corrected for radioactive decay. For bolus injection studies, the time-activity curves were normalized by setting the peak count rate equal to 1.0. For studies in which radiotracers were administered by pulse-perfusion, the resultant time-activity curve was converted to units of ADV. This was accomplished by multiplication of each data point Yi by a calibration factor (fc), ADV i共ml/g dry兲 ⫽ Y i f c where fc was given by the equation:

(3)

FIG. 2. Two-compartment model for describing the kinetics of 18F-labeled thia fatty acid analogs in isolated rat heart tissue. In addition to these compartments representing radioactivity in tissue compartments, a blood volume term is included in the modeling technique (Eq. 8) to account for radiolabel in the vascular spaces within the field of view of the radiation detectors. fc

冉

冊

ml/g dry ADV共t end兲 ⫽ counts/s Ye

(4)

where ADV(tend) is given by Eq. 3, and Ye is the average of Y over the last 10 s of perfusion.

Tracer Kinetic Modeling For bolus injection studies, the time-activity curves were leastsquares fit to the three-exponent model: y ⫽ A1 exp(⫺E1 t) ⫹ A2 exp(⫺E2 t) ⫹ A3 exp(⫺E3 t)

(5)

For pulse-infusion studies, a two-compartment model (Fig. 2) was fit to the time-activity curves. The first compartment (C1) represents radiolabel in reversible pools in the tissue and is in exchange with the blood (perfusate) through rate constants k1 and k2. The second compartment (C2) represents “trapped” metabolites in the tissue. The production rate of trapped metabolites is represented by rate constant k3. The slow clearance of radiolabel from the heart is due to loss of radioactivity from the trapped compartment. This clearance process is represented by the rate constant k4. A blood volume term was included in the model to account for radiotracer present in the vascular spaces. The differential equations describing the kinetics of the radiolabel are: dC1(t) ⫽ k 1C p共t兲 ⫺ 共k 2 ⫹ k 3兲C 1共t兲 dt

(6)

dC2(t) ⫽ k 3C 1共t兲 ⫺ k 4 C 2共t兲 dt

(7)

where Cp(t) is the concentration of radiotracer in the perfusate. The total apparent distribution volume of radiolabel in the heart is modeled as:

18

F-labeled 4-thia Palmitate in Heart

225

TABLE 2. Uptake (% dose kg/g) of FTP in Sprague-Dawley Rats Dietary Status Intervention Time p.i. (min) n Heart Blood Lung Liver Kidney Bone Brain Skeletal muscle

Fed Control 30 7

Fasted Control 30 5

Fasted Control 120 3

Fasted Etomoxir 30 5

0.423 ⫾ 0.203 0.053 ⫾ 0.013** 0.167 ⫾ 0.051** 1.24 ⫾ 0.44 0.39 ⫾ 0.04** 0.152 ⫾ 0.048 0.021 ⫾ 0.002 0.037 ⫾ 0.006**

0.320 ⫾ 0.170 0.032 ⫾ 0.008 0.078 ⫾ 0.018 1.15 ⫾ 0.16 0.248 ⫾ 0.041 0.238 ⫾ 0.073 0.022 ⫾ 0.003 0.020 ⫾ 0.005

0.172 ⫾ 0.046 0.026 ⫾ 0.003 0.060 ⫾ 0.007 1.11 ⫾ 0.15 0.183 ⫾ 0.016* 0.577 ⫾ 0.147* 0.018 ⫾ 0.003 0.017 ⫾ 0.002

0.060 ⫾ 0.012* 0.054 ⫾ 0.012** 0.101 ⫾ 0.037 1.12 ⫾ 0.07 0.156 ⫾ 0.023** 0.165 ⫾ 0.094 0.048 ⫾ 0.016* 0.035 ⫾ 0.006**

Note: Rats were female, 180 –200 g, fasted overnight. Etomoxir-treated rats were given 40 mg/kg etomoxir 2 hr prior to tracer injection. Uptake is normalized to body mass (kg). *p ⬍ 0.05 versus 30 min fasted controls. ** p ⬍ 0.01 versus FTP uptake at 30 min fasted controls in same organ.

ADV(t) ⫽ BV C p共t兲 ⫹ 共1 ⫺ BV兲共C 1共t兲 ⫹ C 2共t兲兲

(8)

The compartmental model solutions were derived numerically using a fourth order Runge-Kutta integrator. Parameter estimation utilized a nonlinear least-squares fitting routine (22). An operational equation is derived for the fractional metabolic rate of tracer (FR) as the rate of entry of radiolabel into compartment C2 at quasi-steady-state (dC1/dt ⫽ 0). In this case, the distribution volume of nonmetabolized tracer in tissue is given as: V1 ⫽

C1 Cp

冏

⫽ ss

k1 k2 ⫹ k3

(9)

The fractional tracer metabolic rate (FRFTP) (mL/min/g dry) in tissue is estimated as: FR ⫽

dC2(t)/dt Cp

冏

⫽ V 1k 3 ⫽ ss

k 1k 3 k2 ⫹ k3

to 60. Reverse phase HPLC purification successfully separated the F fluoro fatty acids from their respective hydrolyzed precursor compounds, although the tosyloxy fatty acid (hydrolyzed product of precursor 3) eluted nearest to the 18F product 17F6THA, necessitating the discarding of a small fraction of 17F6THA peak to avoid contamination. Use of the iodo-ester precursor for FTP was advantageous because its corresponding fatty acid eluted long after the elution of the 18F fluoro fatty acid product. The corresponding bromo-ester was also evaluated (data not shown) but found less preferable to the iodo-ester because it had a lower melting point, leading to less practical storage and handling. 18

(10)

PET Imaging of FTP in Normal Swine Dynamic PET imaging of two normal swine (45–52 kg) was performed using the Advance scanner (General Electric Medical Systems, Milwaukee, WI USA). Emission scanning was preceeded by a whole-body transmission scan. Emission scanning commenced with a dynamic study of the heart until 20 min postinjection, followed by three consecutive whole-body scans for a total scan time of 3 h. The dynamic heart scan sequence was 12 ⫻ 10 s, 3 ⫻ 30 s, and 1 ⫻ 5 min. Image analysis was performed by manual drawing of regions on the PET images. Regions were analyzed to create time-activity curves for left ventricular blood pool, myocardium (septum), lung, and liver. Results were expressed as %dose/mL of tissue.

Statistical Analysis Data are expressed as mean ⫾ standard deviation. The t-test (two-tailed) for unpaired samples was used to compare means of two groups of hearts. RESULTS

Radiotracer Syntheses The omega-labeled 18F-thia fatty acids FTP and 17F6THA were synthesized in decay-corrected radiochemical yields ranging from 25

Biodistribution Studies in Rats Table 2 show biodistribution of FTP in rats fed ad libitum or fasted overnight. Arterial nonesterified fatty acid concentrations were 0.19 ⫾ 0.06 mM and 0.43 ⫾ 0.07 mM (p ⬍ 0.001) in fed and fasted rats, respectively. The highest uptake for all the tracers was observed in heart, liver, and kidney, regardless of dietary status. This biodistribution is consistent with the relatively high ␤-oxidation activity in these three organs. Radioactivity levels in blood, lung, kidneys, and skeletal muscle were higher in the fed state than in the fasted state. Considering the twofold higher concentration of nonesterified fatty acids in the blood with fasting, the nearly equivalent fractional uptake of FTP in the heart and liver agrees with the expected increase of fatty acid utilization by these organs during fasting. Myocardial and hepatic uptake was not significantly changed for FTP in fasted rats from 30 to 120 min, although kidney uptake was somewhat decreased over time. The ␻-labeled analog FTP showed significantly higher (p ⬍ 0.01) bone uptake than previously observed using a (␻-3)-labeled thia-substituted fatty acid (6), consistent with more extensive defluorination of ␻-18F-labeled non-␤-oxidizable fatty acid analogs in rodents. Pretreatment of fasted rats with the CPT-I inhibitor etomoxir caused an 80% reduction of myocardial uptake of FTP at 30 min postinjection (Table 2) despite higher levels of radioactivity remaining in blood. Although the heart:blood ratio was 11.0 in controls at 30 min, the same ratio was 1.1 in etomoxir-treated rats. Kidney uptake was 37% lower in etomoxir-treated rats, but liver uptake was unchanged. Higher uptakes observed in lung, brain, and skeletal muscle of etomoxir-treated rats may be related to the higher level of radioactivity in the blood. Table 3 shows the distribution of 18F radioactivity in hearts excised from control and etomoxir-treated rats at 30 min postinjection of FTP. The results show the predominant fate of the

T. R. DeGrado et al.

226

TABLE 3. Distribution of 18F in Rat Hearts at 30 Min after Injection of FTP. Rats were Fasted Overnight. Values are Expressed as % Dose kg (Body wt)/g (Heart) for Direct Comparison with Total Uptake in Same Units as Given in Table 2. Values are Corrected for Incomplete Recovery of Radioactivity from the Homogenization Procedure. The Organic Soluble Metabolites were Analyzed by Silica-gel TLC using Petroleum Ether/ether/acetic Acid (70:30:1). The Rf Values and Designations (DG ⴝ diglycerides, FA ⴝ fatty acids, TG ⴝ triglycerides) of the TLC Fractions are Shown. Organic soluble Group (n)

Aqueous soluble

Protein interphase

Total

Polar Rf ⴝ 0–0.1

‘‘DG’’ ‘‘FA’’ “TG” Rf ⴝ 0.2–0.3 Rf ⴝ 0.35–0.5 Rc ⴝ 0.79–0.9

Control (5) 0.018 ⫾ 0.004 0.255 ⫾ 0.011 0.047 ⫾ 0.009 0.027 ⫾ 0.011 0.004 ⫾ 0.002 0.005 ⫾ 0.005 Etomoxir (3) 0.003 ⫾ 0.001* 0.027 ⫾ 0.006* 0.030 ⫾ 0.006* 0.014 ⫾ 0.006* 0.002 ⫾ 0.001 0.010 ⫾ 0.001*

0.006 ⫾ 0.005 0.002 ⫾ 0.002

* p ⬍ 0.05 versus control.

radiolabel in control hearts to be distribution into the protein interphase layer containing precipitated protein and to a lesser extent undisturbed membrane fragments. This fraction is decreased by 89% in etomoxir-treated hearts, which is consistent with the hypothesis that protein-bound metabolites of FTP are produced as a result of its oxidation in the mitochondrion. Aqueous soluble radiolabeled metabolites of FTP are relatively minor. Organic soluble metabolites are predominantly polar species that remained unidentified by this assay. They could be phospholipids or fatty acyl intermediates of FTP that did not remain protein-bound. Incorporation of FTP into triglycerides was found to be very small in both control and CPT-I inhibited conditions.

Studies in Isolated Langendorff Perfused Rat Hearts Studies were performed in Langendorff perfused rat hearts to delineate the kinetics of thia fatty acid analogs and determine the relationship of tracer retention to fatty acid oxidation rate in normoxic and hypoxic myocardium. The perfusate was K-H buffer with 5 mM glucose, 0.15 mM albumin, and 0.15 mM palmitate. The perfusate was gassed with mixtures of 95% O2/5% CO2 (normoxic condition) or 35% O2/60% N2/5% CO2 (hypoxic condition). Flow was maintained at 7 mL/min. Figure 3 shows a representative time-activity curve in a Langendorff perfused heart given a rapid bolus injection of FTP at the aortic root. The time-activity curves were fit to a three-exponent model (Table 4). The most rapid phase represents clearance of vascular spaces. The intermediate phase represents extracted but unretained radiolabel in tissue. The clearance half-time of this phase is approximately 11 s. The slowest component evidences turnover of the tracer in one or more slowly clearing compartments. On average, approximately 20% of initially extracted tracer is taken up into the slow turnover compartment(s). The mean clearance half-time of the slowest clearance phase is approximately 90 min. This “trapped” component is clearly delineated from the more rapid clearance components by 4 min after injection. The relationship of the steady-state accumulation rate of FTP and 17F6THA to palmitate oxidation rate was determined in isolated rat hearts using simultaneous pulse-infusion of a 18F-labeled tracer and [9,10-3H]palmitate. Experiments were performed in

normoxic (95% O2) and hypoxic (35% O2) conditions, the latter representing a condition in which fatty acid oxidation rate is limited by mitochondrial ␤-oxidation rate. After perfusion for 20 –30 min with radiotracers at constant concentration in the perfusate, perfusion without radiotracers (washout) followed for 40 min. The heart rate of the unpaced heart was 25% lower in the hypoxic condition (205 ⫾ 33 beats/min [hypoxic] vs. 276 ⫾ 33 beats/min [control]; p ⬍ 0.01). Mean aortic pressure was not significantly different in the hypoxic condition (47 ⫾ 8 mmHg [hypoxic] vs. 43 ⫾ 4 mmHg [control]). Therefore, changes in fatty acid oxidation may be partially due to the moderately lower workload in addition to intracellular hypoxia. Figure 4 shows representative time-activity curves for FTP during the uptake phase, illustrating the markedly lower accumulation rate of tracer in the hypoxic condition. The time-activity curves were fit to a twocompartment model (Fig. 5) with compartments representing reversible precursor (C1) and “trapped” metabolite (C2) compartments. Fitting of data from the last 25–30 min of the washout period gave estimates of the slow clearance rate of “trapped” radiolabel (k4). This value of k4 was subsequently fixed in the least-squares fitting of k1, k2, and k3 in the two-compartment model. From the fitted rate constants, FR was calculated from Eq. 10. Figure 6 shows the correlation of fractional metabolic rate for FTP and 17F6THA with palmitate oxidation rates. Table 5 summarizes the FR estimates and palmitate oxidation rates. Metabolic trapping rates of the 4-thia analog FTP were found to correlate well with rates of palmitate oxidation in normal and hypoxic conditions (r ⫽ 0.85). In contrast, 17F6THA metabolic trapping rate was insensitive to decreases in palmitate oxidation rate observed in hypoxia. The FR of FTP (FRFTP) was approximately 60% the fractional palmitate oxidation rate. Mean FRFTP was decreased 50% in the hypoxic group relative to the normoxic group, tracking a decrease of 40% in the oxidation rate of palmitate with hypoxia (Table 5).

PET Imaging of FTP Uptake in Normal Swine Whole-body dynamic PET imaging studies of FTP (5.5–7.6 mCi bolus) in two anesthetized normal swine showed rapid accumulation of tracer in heart, liver, and kidneys. No bone uptake of 18F was

TABLE 4. Parameter Estimates of 3-exponent Fits to FTP Kinetics in Langendorff Perfused Rat Heart A1 0.63 ⫾ 0.14

E1 (minⴚ1)

A2

E2 (minⴚ1)

A3

E3 (minⴚ1)

35.4 ⫾ 9.0

0.28 ⫾ 0.15

3.67 ⫾ 0.79

0.078 ⫾ 0.023

0.0076 ⫾ 0.0016

Note: Fitted model is y ⫽ A1 exp (⫺E1t) ⫹ A2 exp (⫺E2t) ⫹ A3 exp (⫺E3t). Three hearts were perfused.

18

F-labeled 4-thia Palmitate in Heart

227

FIG. 3. Representative time-activity curve with bolus injection of 18F-labeled 4-thia palmitate in Langendorff perfused rat heart. Rapid (“reversible”) and slow (“trapped”) clearance phases are readily observed. Table 4 gives parameter estimates from threeexponent fits to the curves.

observed over the 3 h imaging period, suggesting negligible metabolic defluorination of the omega-[18F]fluoro fatty acid in the swine. Left ventricular myocardial uptake of FTP was high (Fig. 7). Time-activity curves for myocardium, left ventricular blood pool, and lung regions are shown in Figure 8. Blood clearance was rapid: Myocardium-to-blood ratio exceeded 7:1 at 20 min postinjection. Myocardium-to-lung ratio was ⬎8:1 at 20 min postinjection. Myocardial clearance half-time of 18F radioactivity was approximately 5 h after 20 min postinjection, evidencing effective trapping of the radiolabel in tissue. Liver:heart ratios were 0.4 – 0.8. Clearance rate of radioactivity from liver (T1/2 approximately 50 min) was faster than that from heart. DISCUSSION Mitochondrial ␤-oxidation of long-chain fatty acids is an important process in tissues of the body because it supplies major proportions of the energy requirement in heart, liver, and kidney. Noninvasive assessment of ␤-oxidation may be a useful diagnostic tool because impairment of ␤-oxidation may be interpreted as a feature of disease progression in patients with coronary artery disease (18, 26, 29, 36) and various inherited (12 , 23) and acquired (15, 33) cardiomyopathies. Pharmacologic inhibition of fatty acid oxidation has been suggested as a potential therapy for enhancing recovery of reper-

fused ischemic myocardium (20), whereas gene therapy is a potential aid for patients with inherited deficiencies of fatty acid oxidation enzymes. Noninvasive assessment of fatty acid oxidation with PET could provide critical information for developing therapy protocols and monitoring the effects of therapy in individual patients. The advantages of an 18F-labeled fatty acid analog that is specifically retained as a consequence of ␤-oxidation would include 1) the favorable 110-min half-life of 18F for PET imaging, synthesis of multiple doses, and possibility for commercial production and distribution; 2) excellent imaging characteristics of 18F (short positron range in tissue); 3) simplified, “metabolic-trapping” kinetics in tissue such that ␤-oxidation rate is indicated by tracer accumulation, allowing quantitative estimates of ␤-oxidation rate in analogy to the estimation of glucose phosphorylation rate with 2-[18F]fluoro-2-deoxy-glucose (24); and 4) prolonged retention of radioactivity in tissue that allows imaging by low temporal resolution collimated single photon emission computed tomography or coincidence-capable gamma cameras, as well as higher signal-tonoise in dedicated PET scanners. Importantly, quantitative assessment of ␤-oxidation rate in ischemic tissue is feasible. Although the advantages of a metabolically trapped tracer are compelling, heavy demands are place on radiotracer design, characterization, and validation to establish the utility of a proposed tracer. A metabolically

TABLE 5. Comparison of Estimated Fractional 18F-labeled Thia Fatty Acid Trapping Rates and Palmitate Oxidation Rates in Langendorff Perfused Rat Hearts. All Values have Units ml/min/g Dry 18

F Tracer metabolic trapping rates [FR ⴝ k1k3/(k2 ⴙ k3)]

Fractional palmitate oxidation rate

Tracer

Normoxic (n)

Hypoxic (n)

% Decrease

Normoxic

Hypoxic

% decrease

FTP 17F6THA

1.45 ⫾ 0.39 (5) 1.28 ⫾ 0.18 (4)

0.73 ⫾ 0.16 (5) 1.17 ⫾ 0.35 (5)

50* 9

2.73 ⫾ 0.23 2.37 ⫾ 0.13

1.64 ⫾ 0.30 1.47 ⫾ 0.39

40* 39*

* p ⬍ 0.01. Oxygenation of perfusate: normoxic, 95% O2; hypoxic, 35% O2.

228

T. R. DeGrado et al.

FIG. 4. Time-activity curves of 18 F-labeled 4-thia palmitate in representative isolated rat hearts in normoxic (95% oxygen [O2]) and hypoxic (35% O2) perfusion conditions. The tracer pulse-infusion period (0 < t < 20 min) is shown. Coronary flow was 7 mL/min in both experiments. ADV, apparent distribution volume.

trapped tracer of ␤-oxidation should be retained in tissue only in the form of ␤-oxidation metabolites in a manner such that there is a close (and well-behaved) relationship between tracer trapping rate and fatty acid oxidation rates under a broad range of physiologic conditions. The 6-thia fatty acid analog FTHA was developed as a metabolically trapped tracer of ␤-oxidation in myocardium (4, 7). 4-Thia fatty acids are ␤-oxidized in the mitochondria to 4-thia-enoylCoAs, which apparently accumulate in the mitochondrion because they are very poor substrates for CPT-II (32), yet bind to acyl-CoA dehydrogenase enzymes in the mitochondrion (17). 4-Thia-enoylCoA is very slowly converted to ␤-hydroxy-4-thia acyl-CoA by the next step of the ␤-oxidation pathway (17). The resultant ␤-hydroxy-4-thia acyl-CoA splits spontaneously to an alkyl-thiol and

FIG. 5. Fit of two-compartment model to 18F-labeled 4-thia palmitate kinetics in normoxic Langendorff heart. Fitted parameter estimates are: blood volume ⴝ 0.78 mL/g dry; k1 ⴝ 5.01 mL/ min/g dry; k2 ⴝ 0.56/min; k3 ⴝ 0.36/ min; k4 ⴝ 0.0050/min. ADV, apparent distribution volume.

malonic acid semialdehyde-CoA ester. 4-Thia fatty acids inhibit fatty acid oxidation and induce fatty liver in vivo (30). The inhibition is presumably explained by accumulation of 4-thia-enoyl-CoAs in the mitochondrion leading to decreased mitochondrial CoASH, as well as inactivation of long-chain acyl-CoA dehydrogenase by binding of the accumulated 4-thia-enoyl-CoA to this enzyme (17). ␻-Oxidation and sulfur oxidation of thia fatty acids in liver give rise to dicarboxylic sulfoxy acids, which are efficiently cleared by kidneys and excreted into the urine (14). Incorporation of 4-thia fatty acids into complex lipids of hepatocytes is relatively low (31). Long-chain fatty acids labeled with 18F at or near the terminus carbon are taken up and metabolically sequestered in heart and liver in a manner that is similar to that of their unlabeled counterparts

18

F-labeled 4-thia Palmitate in Heart

FIG. 6. Correlation of estimated fractional tracer metabolic trapping rates of 18F-labeled thia fatty acids and fractional palmitate oxidation rate in perfused rat heart. Squares, normoxic (95% oxygen [O2]); triangles, hypoxic (35% O2). FTP, 18 F-labeled 4-thia palmitate; 17F6THA, 17-[18F]fluoro-6-thiaheptadecanoic acid. (16). Therefore, it was expected that a terminally 18F-labeled even-substituted thia fatty acid would trace fatty acid uptake and undergo ␤-oxidation to a potentially trapped metabolite in the mitochondrion. Placement of the 18F atom at the (␻-3) carbon was found to minimize defluorination of the tracer in mice as a consequence of cytochrome P-450 mediated ␻-oxidation mainly in liver (6).

229

Evaluation studies have shown the myocardial accumulation of FTHA to follow expected responses of ␤-oxidation in the following conditions: 1) reduced with pharmacologic inhibition of CPT-I in mouse (7); 2) unchanged by dipyridamole- or adenosine-induced hyperemia (8, 34); 3) increased by exercise in normal human volunteers correlating with increases in heart rate-pressure product (8); 4) decreased by lactate in pig heart with parallel decreases in oxidation of palmitate (34); 5) decreased by ischemia in pig heart with parallel decreases in oxidation of palmitate (35); 6) decreased in humans with coronary artery disease (28); and 7) decreased with glucose/insulin clamp in normal human myocardium (21). However, accumulation of FTHA was not sensitive to inhibition of fatty acid oxidation with hypoxia in extracorporeally perfused swine myocardium (35). Our present results with the corresponding ␻-labeled 6-thia analog 17F6THA in perfused rat hearts confirm the insensitivity of 6-thia fatty acid kinetics to hypoxia in the myocardium, suggesting the existence of a retention mechanism in the heart for 18F-labeled 6-thia fatty acid analogs that is not dependent on fatty acid oxidation. The placement of the sulfur atom at the sixth position of FTHA rather than the fourth was motivated by concerns that interference by the sulfur atom with initial metabolic and transport steps would be lower with the larger distance between the sulfur atom and the carboxylic group. The present results with 4-thia fatty acid analogs show this not to be the case. Accumulation of the 6-thia (17F6THA) and 4-thia (FTP) long-chain fatty acids were equivalent in normoxic perfused rat hearts, suggesting that both radiotracers were equally accepted by the fatty acid binding proteins, carriers, and enzymes involved in the transport and metabolic trapping of the tracers in the myocyte. However, both 4-thia and 6-thia analogs were more slowly oxidized by isolated heart than natural palmitate. Their fractional oxidation rate was approximately 60% that of palmitate in normal isolated rat hearts. It remains to be determined which steps of transport or metabolism are slower for the thia fatty acid analogs relative to natural fatty acids. In isolated rat heart, reduction in fatty acid oxidation rate with hypoxia was closely tracked by reduction in the trapping rate of 4-thia FTP. In contrast, the accumulation of the 6-thia analog 17F6THA was unchanged in hypoxia, confirming observations in hypoxic swine myocardium with FTHA (35). Evidently, the con-

FIG. 7. Positron emission tomography image of mid-ventricular short-axis slice of normal swine heart acquired 10 –20 min postinjection with 18F-labeled 4-thia palmitate.

T. R. DeGrado et al.

230

FIG. 8. Time-activity curves for 18Flabeled 4-thia palmitate in normal swine myocardium, lung, liver, and left ventricular blood pool.

dition of hypoxia exposes mechanism(s) of retention in myocytes for the 6-thia fatty acids that the 4-thia fatty acids do not share. The chemical nature of the retained fraction of radioactivity with 17F6THA in hypoxic myocardium was not identified in this study, but previous investigations in living mouse hearts showed incorporation of FTHA into polar lipids and, to a lesser extent, triglycerides (7). In addition, radiolabeled acyl intermediates of 17F6THA could possibly accumulate in hypoxic myocardium to a greater degree than those of FTP, depending on rates of formation and hydrolysis. The metabolic trapping rates of the long-chain analogs 17F6THA and FTP (1.3–1.5 mL/min/g dry) in relation to coronary flows of approximately 40 mL/min/g dry in normal conditions imply net (steady-state) extractions of approximately 3%. The discrepancy in this figure with the 8% “trapped” fraction in bolus injection experiments with FTP (Table 4) is likely due to a systematic underestimation of the total dose by the peak count rate in the bolus technique. At physiologic flows of 1–2 mL/min/g wet, the net extraction of FTP in vivo should be in the 10 –25% range. At these extraction levels, the tracer uptake is not likely to be influenced significantly by changes in myocardial perfusion per se. The slow clearance rate of FTP in the normal swine heart is consistent with the avid myocardial retention of FTP in living rats at 2 h postinjection (Table 2). Analytical studies in extracts of hearts from living rats showed that protein-bound metabolites comprise ⬎80% of the radioactivity in the heart in normal conditions. The protein-bound radioactivity is ⬎90% reduced in CPT-I-inhibited rat hearts. This result is consistent with decrease in fatty acid oxidation rates with CPT-I inhibition (3) and suggests that the accumulation of protein-bound metabolites in the myocardium is dependent on transport and oxidation of FTP in the mitochondrion. A putative end-metabolite of FTP in the mitochondrion that could bind to proteins is the product of the first step of ␤-oxidation, 16-[18F]fluoro-4-thia-hexadec-2-enoyl-CoA. The next step of ␤-oxidation (i.e., hydration of the 4-thia-enoyl-CoA ester) has been shown to be very slow with a nonfluorinated 4-thia fatty

acid (17). Further analytical work is necessary to demonstrate that the chemical form of “trapped” species of FTP is a product of ␤-oxidation and to validate the putative two-compartment model used to fit the kinetics of FTP in isolated rat hearts and estimate fatty acid oxidation rate over a broad range of conditions. Until such work is done, it cannot be firmly concluded that the trapping of FTP in the myocardium strictly measures ␤-oxidation rate. The lack of evidence of in vivo defluorination of FTP in the living swine was a surprising finding in light of the evidence for defluorination of FTP in rodents (Table 2). Initial studies with FTP in humans also show neglible bone uptake of radioactivity (to be published elsewhere), suggesting that the metabolic handling of FTP in swine and humans is different than in the rat. A species dependence of ␻-oxidative defluorination of FTP is suggested. Rodent species are known to have particularly strong hepatic cytochrome P-450 – dependent ␻-oxidative metabolism of fatty acids (6). The high liver uptake of FTP is undesirable for myocardial imaging, especially near the inferior wall of the left ventricle. Nevertheless, the spatial resolution of current state of the art PET scanners and image reconstruction algorithms may allow measurement of myocardial radioactivity concentration with minimal spillover problems from liver. Whether dietary status could be manipulated to optimize the heart:liver ratio of FTP uptake remains to be seen.

CONCLUSIONS Our results show that the 4-thia palmitate analog FTP has potential as a metabolically trapped tracer of fatty acid oxidation in heart. FTP shares the excellent imaging characteristics as the previously developed 6-thia analog FTHA but may possess higher specificity for indication of fatty acid oxidation. Further studies are needed to define the mechanism of retention and understand the relationship of FTP trapping rate in the myocardium to oxidation rate of long-chain fatty acids in various conditions and disease states.

18

F-labeled 4-thia Palmitate in Heart

This work was performed with the support of National Institutes of Health Grant HL-54882. The authors thank Larry F. Whitesell and Khristen J. Carlson for their excellent technical support. Etomoxir was generously provided by Dr. H. Wolf of Byk Gulden, Konstanz, Germany.

References 1. Abendschein D. R., Fox K. A. A., Ambos H. D., Sobel B. E. and Bergmann S. R. (1987) Metabolism of beta-methyl[1-11C]heptadecanoic acid in canine myocardium. Nucl. Med. Biol. 14, 579 –585. 2. Bergmann S. R., Weinheimer C. J., Markham J. and Herrero P. (1996) Quantitation of myocardial fatty acid metabolism using PET. J. Nucl. Med. 37, 1723–1730. 3. Bielefeld D. R., Vary T. C. and Neely J. R. (1985) Inhibition of carnitine palmitoyl-CoA transferase activity and fatty acid oxidation by lactate and oxfenicine in cardiac muscle. J. Mol. Cell. Cardiol. 17, 619 – 625. 4. DeGrado T. R. (1991) Synthesis of 14(R,S)-[18F]fluoro-6-thia-heptadecanoic acid (FTHA). J. Label. Comp. Radiopharm. 29, 989 –995. 5. DeGrado T. R., Holden J. E., Ng C. K., Raffel D. M. and Gatley S. J. (1989) ␤-Methyl-15-p-iodophenylpentadecanoic acid metabolism and kinetics in the isolated rat heart. Eur. J. Nucl. Med. 15, 78 – 80. 6. DeGrado T. R. and Moka D. C. (1992) Non-␤-oxidizable T-[18F]fluoro long-chain fatty acid analogs show cytochrome P-450 mediated defluorination: Implications for the design of PET tracers of myocardial fatty acid utilization. Nucl. Med. Biol. 19, 389 –397. 7. DeGrado T. R., Sto¨cklin G. and Coenen H. H. (1991) 14(R,S)[18F]Fluoro-6-thia-heptadecanoic acid (FTHA): Evaluation in mouse of a new in vivo probe of myocardial utilization of long-chain fatty acids. J. Nucl. Med. 32, 1888 –1896. 8. Ebert A., Herzog H., Stoecklin G., Henrich M., DeGrado T. R., Coenen H. H. and Feinendegen L. E. (1994) Kinetics of 14(R,S)-[F-18]fluoro6-thia-heptadecanoic acid (FTHA) in normal human heart at rest, during exercise and after dipyridamole. J. Nucl. Med. 35, 51–56. 9. Elmaleh D. R., Livni E., Levy S., Varnum D., Strauss H. W. and Brownell G. L. (1983) Comparison of 11C and 14C-labeled fatty acids and their b-methyl analogs. Nucl. Med. Biol. 10, 181–187. 10. Fox K. A. A., Abendschein D. R., Ambos H. D., Sobel B. E. and Bergmann S. R. (1985) Efflux of metabolized and nonmetabolized fatty acid from myocardium. Implications for quantifying myocardial metabolism tomographically. Circ. Res. 57, 232–243. 11. Goldstein R. A., Klein M. S., Welch M. J. and Sobel B. E. (1980) External assessment of myocardial metabolism with C-11 palmitate in vivo. J. Nucl. Med. 21, 342– 438. 12. Hale D. E. and Bennett M. J. (1992) Fatty acid oxidation disorders: A new class of metabolic diseases. J. Pediatr. 121, 1–11. 13. Huang S. C. and Phelps M. E. (1986) Principles of tracer kinetic modeling in positron emission tomography and autoradiography. In Positron Emission Tomography and Autoradiography. Principles and Applications for the Brain and Heart (Edited by Phelps M. E., Mazziotta J. C. and Schelbert H. R.), pp. 287–346. Raven Press, New York. 14. Hvattum E., Skrede S., Bremer J. and Solbakken M. (1992) The metabolism of tetradecylthiopropionic acid, a 4-thia stearic acid, in the rat. In vivo and in vitro studies. Biochem. J. 286, 879 – 887. 15. Kelly D. P., Mendelsohn N. J., Sobel B. E. and Bergmann S. R. (1993) Detection of impaired myocardial fatty acid utilization by positronemission tomography indicative of genetic deficiency of long-chain acyl-CoA dehydrogenase. Am. J. Cardiol. 71, 738 –744. 16. Knust E. J., Kuperfernagel C. and Sto¨cklin G. (1979) Long-chain F-18 fatty acids for the study of regional metabolism in heart and liver; odd-even effects in mice. J. Nucl. Med. 20, 1170 –1175. 17. Lau S. M., Brantley R. K. and Thorpe C. (1989) 4-Thia-trans-2alkenoyl-CoA derivatives: Properties and enzymatic reactions. Biochemistry 28, 8255– 8262. 18. Liedtke A. J. (1981) Alterations of carbohydrate and lipid metabolism in the acutely ischemic heart. Prog. Cardiovasc. Dis. 23, 321–336. 19. Livni E., Elmaleh D. R., Levy S., Brownell G. L. and Strauss W. H. (1982) Beta-methyl[1-11C]heptadecanoic acid: A new myocardial metabolic tracer for positron emission tomography. J. Nucl. Med. 23, 169 –175.

231

20. Lopaschuk G. D. (1997) Alterations in fatty acid oxidation during reperfusion of the heart after myocardial ischemia. Am. J. Cardiol. 80, 11A–16A. 21. Ma¨ki M. T., Haaparanta M., Nuutila P., Oikonen V., Luotolahti M., Eskola O. and Knuuti J. M. (1998) Free fatty acid uptake in the myocardium and skeletal muscle using fluorine-18-fluoro-6-thia-heptadecanoic acid. J. Nucl. Med. 39, 1320 –1327. 22. Marquardt D. W. (1963) An algorithm for least-squares estimation of nonlinear parameters. J. Soc. Ind. Appl. Math. 11, 431– 441. 23. North K. N., Hoppel C. L., De Girolami U., Kozakewich H. P. and Korson M. S. (1995) Lethal neonatal deficiency of carnitine palmitoyltransferase II associated with dysgenesis of the brain and kidneys. J. Pediatr. 127, 414 – 420. 24. Phelps M. E., Hoffman E. J., Selin C., Sokoloff L. and Kuhl D. E. (1979) Tomographic measurement of local cerebral metabolic rate in humans with (F-18) 2-fluoro-2-deoxy-D-glucose: Validation of method. Ann. Neurol. 6, 371–388. 25. Saddik M. and Lopaschuk G. (1992) Myocardial triglyceride turnover during reperfusion of isolated rat hearts subjected to a transient period of global ischemia. J. Biol. Chem. 267, 3825–3831. 26. Scho¨n H. R., Schelbert H. R., Najafi A., Hansen H., Huang S. C., Barrio J. and Phelps M. E. (1982) C-11 labeled palmitic acid for the noninvasive evaluation of regional myocardial fatty acid metabolism with positron-computed tomography. II. Kinetics of C-11 palmitic acid in acutely ischemic myocardium. Am. Heart J. 103, 532–547. 27. Scho¨n H. R., Schelbert H. R., Robinson G., Najafi A., Huang S. C., Hansen H., Barrio J., Kuhl D. E. and Phelps M. E. (1982) C-11 labeled palmitic acid for the noninvasive evaluation of regional myocardial fatty acid metabolism with positron-computed tomography. I. Kinetics of C-11 palmitic acid in normal myocardium. Am. Heart J. 103, 532–547. 28. Schultz G., Vom Dahl J., Kaiser H. J., Koch K. C., Sabri O., Banneitz I., Cremerius U. and Buell U. (1996) Imaging of ␤-oxidation by static PET with 14(R,S)-[18F]-fluoro-6-thiaheptadecanoic acid (FTHA) in patients with advanced coronary heart disease: A comparison with 18 FDG-PET and 99mTc-MIBI SPECT. Nucl. Med. Commun. 17, 1057– 1064. 29. Schwaiger M., Fishbein M. C., Block M., Wijns W., Selin C., Phelps M. E. and Schelbert H. R. (1987) Metabolic and ultrastructural abnormalities during ischemia in canine myocardium: Noninvasive assessment by positron emission tomography. J. Mol. Cell. Cardiol. 19, 259 –269. 30. Skorve J., Asiedu D., Solbakken M., Gjestdal J., Songstad J. and Berge R. K. (1992) Comparative effects of oxygen and sulfur-substituted fatty acids on serum lipids and mitochondrial and peroxisomal fatty acid oxidation in rat. Biochem. Pharmacol. 43, 815– 822. 31. Skrede S., Sorensen H. N., Larsen L. N., Steineger H. H., Hovik K., Spydevold O. S., Horn R. and Bremer J. (1997) Thia fatty acids, metabolism and metabolic effects. Biochim. Biophys. Acta 1344, 115–131. 32. Skrede S., Wu P. and Osmundsen H. (1995) Effects of tetradecylthiopropionic acid and tetradecylthioacrylic acid on rat liver lipid metabolism. Biochem. J. 305, 591–597. 33. Sochor H., Schelbert H. R., Schwaiger M., Henze E. and Phelps M. E. (1986) Studies of fatty acid metabolism with positron emission tomography in patients with cardiomyopathy. Eur. J. Nucl. Med. 12, S66 – S69. 34. Stone C. K., Pooley R. A., DeGrado T. R., Renstrom B., Nickles R. J., Nellis S. H., Liedtke A. J. and Holden J. E. (1998) Myocardial uptake of the fatty acid analog 14-18F-fluoro-6-thia-heptadecanoic acid (FTHA) in comparison to ␤-oxidation rates by tritiated palmitate. J. Nucl. Med. 39, 1690 –1696. 35. Stone C. K., Pooley R. A., DeGrado T. R., Renstrom B., Nickles R. J., Nellis S. H., Liedtke A. J. and Holden J. E. (1998) Comparison of fatty acid tracers FTHA and BMIPP retention during myocardial ischemia and hypoxia. J. Nucl. Med. 39, 1684 –1689. 36. Tamaki N., Kawamoto M., Takahashi N., Yonekura Y., Magata Y., Torizuka T., Nohara R., Kambara H. and Konishi J. (1993) Assessment of myocardial fatty acid metabolism with positron emission tomography at rest and during dobutamine infusion in patients with coronary artery disease. Am. Heart J. 125, 702–710.