Kinetic Analysis of Technetium-99m-Labeled Nitroimidazole (BMS-181321) as a Tracer of Myocardial Hypoxia
Background Experimental data have indicated that [99mTc]- nitroimidazole (BMS-181321) is preferentially taken up in hypoxic tissue; its kinetics, however, has not been fully investigated. The purpose of this study was to address the relation between perfusate oxygen level and myocardial retention of [99mTc]nitroimidazole.
Methods and Results Bolus injection and constant infusion experiments were performed in Langendorff buffer–perfused rat hearts in normoxic and hypoxic conditions. Data were acquired with a pair of NaI detectors. The initial clearance rate of [99mTc]nitroimidazole was approximately 20 seconds and independent of perfusate oxygen level. The slow clearance rate was greater than 3 hours in all perfusion conditions. The tissue retention of [99mTc]nitroimidazole varied from 0.61±0.14% in normoxic conditions to 5.94±1.16% in the most severe hypoxic conditions. In addition, tissue retention was inversely proportional to perfusate oxygen level in a sigmoidal manner. The constant infusion experiments established that the binding rate at 25% oxygen level (1.94±0.38 mL of perfusate/min−g dry wt) was twofold of that at 40% and sevenfold at 100%. The binding rate of [99mTc]nitroimidazole was independent of the perfusion sequence, suggesting irreversible binding.
Conclusions These data indicate that [99mTc]nitroimidazole may be a useful tracer for the identification of myocardial hypoxia. A sigmoidal relation was demonstrated for the uptake of the tracer, which suggests that a threshold level of hypoxia is necessary for the uptake of the tracer.
Currently available noninvasive approaches for the detection of ischemia in patients with coronary artery disease are based on either the identification of flow heterogeneity or regional alterations of myocardial metabolism.1 Studies are commonly performed with [99mTc]sestamibi and thallium-201 in single photon emission computed tomography (SPECT) and [18F]fluorodeoxyglucose, [13N]ammonia, and rubidium-82 in positron emission tomography (PET). These studies using either imaging modality, however, provide only an indirect assessment of regional myocardial ischemia. Since ischemia is tightly coupled to hypoxia, an assessment of tissue oxygen content by tracers extracted readily by the myocardium and retained significantly in direct response to tissue oxygen level may provide a means of positive imaging for the hypoxic myocardium. A class of compounds with high electron affinity, the radiolabeled nitroimidazoles, is therefore under active evaluation for this purpose. These lipophilic compounds diffuse across the cell membrane and undergo reduction in the cytoplasm to form a radical.2 3 4 When oxygen is abundant in the cell, it reacts with the radical anion formed to yield superoxide and noncharged nitroimidazole that then diffuse out of the cell. When intracellular hypoxia is present, the nitroimidazole radical anion is reduced further to form nitrous compounds that combine covalently with cytosolic macromolecules and then are trapped intracellularly. Enhanced retention of nitroimidazoles in hypoxic cells has been demonstrated in dogs5 6 and perfused hearts7 using [18F]fluoromisonidazole and in isolated adult rat myocytes8 and dogs9 using [3H]fluoromisonidazole. For wider clinical applications, [99mTc]nitroimidazole (BMS-181321) has been recently developed to be used as a potential hypoxic marker with the use of SPECT. Preliminary studies in dogs,10 11 perfused hearts,12 and isolated myocytes12 have indicated that [99mTc]nitroimidazole (BMS-181321) may serve as a sensitive marker for hypoxic myocardium. The kinetics of [99mTc]nitroimidazole, however, have not been fully evaluated in the myocardium.
The specific aims of this study were (1) to evaluate the kinetics of [99mTc]nitroimidazole in an isolated perfused heart preparation under a range of hypoxic conditions and (2) to determine the relation between tissue retention and perfusate oxygen level.
The structure of [99mTc]nitroimidazole (BMS-181321, Bristol-Myers Squibb) is shown in Fig 1⇓. The radiochemical synthesis of BMS-181321 has been described recently in detail by Linder et al.13 Briefly, a ligand vial was prepared by dissolving 2.0 mg of propylene amine oxime (PnAO)-1-(2-nitroimidazole) in 2 mL of saline containing 36 MBq of [99mTc]TcO4−. Four milliliters of saline was added to reconstitute a Squibb Techneplex kit, which consists of 10 mg pentetate calcium trisodium and 0.5 mg SnCl2 · 2H2O. The formation of [99mTc]nitroimidazole was initiated by adding 0.15 mL of the reconstituted Techneplex into the ligand vial enclosed in a lead container at room temperature. Radiochemical purity of [99mTc]nitroimidazole was checked by thin-layer liquid chromatography (solvent, diethyl ether) to be >90% after 10 minutes of reaction. [99mTc]pentetate was prepared with the same procedure except without the addition of the PnAO-1-(2-nitroimidazole) ligand.
Isolated Perfused Rat Heart
Male Sprague-Dawley rats (Harlan) (weight, 250 to 300 g) were fed ad libitum. The procedures followed for the perfused heart studies were in accordance with institutional guidelines. The schematic for the perfusion apparatus and the detection system are shown in Fig 2⇓. The perfusate buffer was delivered to a heat and gas exchanger from a reservoir by a roller pump. Hearts were perfused in Langendorff mode with a constant flow of 10 mL/min without recirculation in one of the following conditions: (1) normoxia with buffer saturated by 95% O2 and 5% CO2 (perfusate oxygen level of 100%) and (2) hypoxia with perfusate oxygen level reduced systemically from 100% to 75%, 50%, 40%, 25%, and 0% by mixing with 95% N2 and 5% CO2. Heart rate and systolic and diastolic pressures were also recorded on-line to determine whether the perfused heart was suitable for the experiment. The perfusate was Krebs-Henseleit buffer with the following composition (in mmol/L): NaCl, 119; KCl, 5.9; MgSO4, 7; H2O, 0.53; NaH2PO4, 1.4; NaHCO3, 25; CaCl2, 2; H2O, 1.25; glucose, 5; and insulin, 10 mU/mL.
External Detection of Tissue Radioactivity
Tissue residue curves of 99mTc-labeled compounds were measured externally with a pair of NaI detectors as shown in Fig 2⇑. The proximity of the detectors to the perfused heart was important for increasing detection sensitivity and minimizing motion effects. During acquisition of tissue residue curves, a cylindrical thin film surrounded the heart to keep it warm. Each NaI detector consisted of a 64-mm crystal (Bicron) encased in a 20-mm-thick cylindrical lead shield with a 38-mm-diameter collimator hole. Since the whole heart was in the field of view of the two detectors, total counts over the heart, which were always less than 40 000 counts/s, were recorded as a function of time in all studies. The percent dead-time loss was less than 1% at a count rate of 40 000 counts/s. Tissue residue curves were acquired on-line with a digital computer (Leading Edge) at a temporal sampling rate of 1 second per data point for the bolus injection experiments and 10 seconds per data point for the constant infusion experiments.
The experimental protocols for both the bolus injection and the constant infusion studies are shown in Fig 3⇓. All studies involved a baseline period of 15 minutes in which hearts were perfused at the normal oxygen level of 100%.
Protocol I: Bolus Injection Studies of [99mTc]TcO4−, [99mTc]Pentetate, and [99mTc]Nitroimidazole
In protocol I (Fig 3A⇑), the gas mixture was switched to either 0%, 25%, 40%, 50%, 75%, or 100% of perfusate oxygen level at 15 minutes before the injection of the radiopharmaceuticals. This additional 15-minute period allowed the perfused hearts to equilibrate with the gas mixture in the perfusate. After 30 minutes of perfusion, 1.5 MBq of [99mTc]TcO4−, [99mTc]- pentetate, or [99mTc]nitroimidazole was injected over 5 seconds via the tubing (Fig 2⇑) into the aortic root of the perfused hearts. Tissue residue curves then were acquired for 60 minutes, which was required for the determination of the slow clearance rate. Five hearts each were performed in experiments studied with [99mTc]nitroimidazole and three hearts each for other experiments. The percent yield of the [99mTc]nitroimidazole synthesis was about 92%, and the other 8% could be either [99mTc]TcO4− or [99mTc]pentetate; thus, the confounding effect of these two radiolabeled compounds on the kinetics of [99mTc]nitroimidazole was examined in additional experiments.
Protocols IIA and IIB: Constant Infusion Studies of [99mTc]Nitroimidazole
The accuracy of determining percentage of tissue retention using the bolus technique relies heavily on the detection of a true peak; therefore, a second set of experiments was performed to examine the applicability of using the constant infusion technique to estimate the binding rate of [99mTc]nitroimidazole under various oxygen levels. In addition, the constant infusion experiments allow the evaluation of [99mTc]nitroimidazole kinetics in response to changes in oxygen level within the same heart. Two protocols were used as outlined in Fig 3⇑. Protocol IIA (n=3 hearts) was first performed with normoxia followed by hypoxia, and protocol IIB (n=3 hearts) was performed in the reverse order. In both protocols, hearts were initially perfused with normal oxygen level for 15 minutes. Depending on the protocol, hearts then were perfused with a different oxygen level for an additional period of 15 minutes. After this equilibration period, the infusion of [99mTc]nitroimidazole (2.2 MBq/mL) into the aortic root was initiated at 0.1 mL/min. In protocol IIA (Fig 3B⇑), hearts were infused with [99mTc]nitroimidazole for an additional 30 minutes at 100% O2 followed by two successive intervals of 30 minutes and 20 minutes at 40% and 25% oxygen levels, respectively. In protocol IIB (Fig 3B⇑), hearts were infused with [99mTc]nitroimidazole for 30 minutes and 40 minutes at 40% and 100% oxygen levels, respectively. To ensure that the perfused heart would regain most of its contractility after a hypoxic insult, the experiment with 25% oxygen level was not performed in protocol IIB. Furthermore, the objective of protocol IIB was to examine the effect of 100% perfusate oxygen level on the binding kinetics of [99mTc]nitroimidazole after a period of hypoxic insult. Since the objective of the constant infusion protocols was to provide an alternative method to analyze the kinetics of [99mTc]nitroimidazole, no constant infusion experiments were performed with [99mTc]TcO4− and [99mTc]pentetate. In addition, since [99mTc]TcO4− was rapidly cleared from the vasculature and the right heart chambers, as seen in “Results,” constant infusion experiments with [99mTc]TcO4− were not appropriate.
Tissue Residue Curves of 99mTc-Labeled Compounds
Tissue time-activity curves obtained after a bolus of [99mTc]- nitroimidazole and [99mTc]pentetate were fitted triexponentially. The fast and slow clearance rates were determined by the first and third components, respectively. Given the objective of the current study, the information for the second component was not used. Tissue residue curves, however, still required three exponential terms for good fit. Clearance half-time then was estimated by dividing 0.693 by the clearance rate.14 All tissue time-activity curves were normalized to their corresponding peak counts to facilitate comparisons among different experimental groups. Tissue retention was calculated by averaging the normalized counts between 14 minutes and 16 minutes (retention 1) or between 58 minutes and 60 minutes (retention 2). The effect of perfusate oxygen level on tissue retention of [99mTc]nitroimidazole was described by the modified Hill equation15 :
where TR is the fractional tissue retention of [99mTc]nitroimidazole, [O] is the fractional perfusate oxygen level, and a, b, c, and n are constants. To include the observation that tissue retention was not zero at 100% perfusate oxygen level, the Hill equation was modified by the addition of the constant a. Furthermore, tissue retention reached almost a plateau as oxygen level increased beyond 75%; thus, constant a was set at a value of 0.0061, which was the average tissue retention fraction at 100%. Constants b, c, and n were estimated by nonlinear regression using kaleidagraph 3.0 (Synergy Software). Since the mean values of retention 1 were not statistically different from those of retention 2 for the [99mTc]nitroimidazole experiments, the mean values of retention 1 were used to estimate the parameters.
Estimation of Tissue Binding Rate of [99mTc]Nitroimidazole
Since the bolus data indicated that [99mTc]nitroimidazole exhibited very slow tissue clearance in all hypoxic conditions, the tracer is effectively trapped in the myocardium after the initial clearance through the vasculature. For tracers with irreversible binding kinetics, the linear slope indicates the binding rate when the tissue residue curve is normalized to the delivery of the tracer over time.16 The Patlak approach has been used extensively for the estimation of myocardial glucose metabolic rate using [18F]fluorodeoxyglucose and PET.17 Each tissue residue curve in the current study was scaled to the infusion dose by counting 1 mL of perfusate in the same detection system. The curve then represented the distribution volume of [99mTc]nitroimidazole over time. In other words, distribution volume describes the equivalent amount of perfusate containing [99mTc]nitroimidazole accumulated in the perfused heart at a particular time. The binding rate (mL of perfusate/min−g dry wt) of [99mTc]nitroimidazole for each perfusion condition was estimated by the fitted slope obtained with linear regression. In protocol IIA (normoxia-hypoxia), the slope was estimated from 45 minutes to 55 minutes, 70 minutes to 80 minutes, and 90 minutes to 100 minutes for perfusate oxygen level of 100%, 40%, and 25%, respectively. In protocol IIB (hypoxia-normoxia), the slope was estimated from 45 minutes to 55 minutes and from 90 minutes to 100 minutes for perfusate oxygen level of 40% and 100%, respectively.
All values are expressed as mean±SD. Dry weight was determined by placing the whole heart in an oven overnight. Statistical analysis was performed with use of statview 4.1 (Abacus Concepts, Inc). P<.05 was considered significant. The statistical significance of the changes induced by different study conditions was assessed by one-way ANOVA and Fisher’s test for Tables 1, 2, and 3. The paired, two-tailed t test was used to determine the statistical significance between individual values of retention 1 and retention 2 before being averaged in Table 2⇓, but the mean values were tested by unpaired, two-tailed t-test. Since Table 4⇓ was compiled based on Tables 2⇓ and 3⇓, the unpaired, two-tailed t test was used to evaluate the statistical significance between any two groups in this table. The Bonferroni procedure was used to adjust the critical P values for multiple tests performed in Tables 2⇓ and 4⇓.18
Tissue Residue Curves of [99mTc]Nitroimidazole
Characteristic tissue residue curves after bolus injection of [99mTc]nitroimidazole for various degrees of hypoxia are shown in Fig 4⇓. Curves were scaled to the peak count rate and plotted on a semi–log scale for easy visualization. Residue curves then were fitted by a triexponential function, which appeared to sufficiently describe the kinetics in all perfusion conditions (R>.99). The first exponential term mainly describes the tracer clearance through the vasculature and the right heart chambers. As shown in Fig 4⇓, the initial clearance was independent of perfusate oxygen level. The clearance half-time of [99mTc]nitroimidazole for this initial phase was about 20 seconds, with no statistical significance among all perfusion conditions (Table 1⇓). The second term was required to describe the kinetics at the knee portion of the residue curves. The kinetics during this phase is indicative of nonbound [99mTc]nitroimidazole equilibrating between the interstitial space and the intracellular space. This immediate phase was not part of the objectives in the current study; thus, its results were not reported. The triexponential function, nonetheless, was required for the determination of the slow clearance rate. The tail portion, which reflects the binding kinetics of cellular [99mTc]nitroimidazole anion after the nitroreduction process, reached a plateau in less than 10 minutes in all hypoxic conditions. As shown in Table 1⇓, the clearance half-time for this slow phase was about 5 hours independent of the perfusate oxygen level. The individual values of retention 1 were statistically different from those of retention 2 before being averaged, but the mean values were not (Table 2⇓). This indicates that a shorter study time can be used to evaluate myocardial hypoxia with the tracer retention because of the slow washout rate of the reduced forms of [99mTc]nitroimidazole.
Kinetic Disparity Between [99mTc]Nitroimidazole, [99mTc]TcO4−, and [99mTc]Pentetate
To further evaluate the effect of perfusate oxygen level on the kinetics of [99mTc]TcO4− and [99mTc]pentetate, experiments were performed with these radiolabeled compounds under two perfusion conditions: perfusate oxygen levels of 25% and 100%. The fast clearance half-time for [99mTc]TcO4− was about 5 seconds, which compared with [99mTc]nitroimidazole was 4 times faster in both perfusion conditions (Table 1⇑). As expected, [99mTc]TcO4− was not retained in the myocardium to any significant degree after 2 minutes (Fig 5⇓) because its negative charge does not permit the molecule to diffuse across the cell membrane. Subsequently, data for the slow clearance half-time of [99mTc]TcO4− was not available and therefore was not reported in Table 1⇑. The clearance half-time for [99mTc]pentetate was about 10 seconds, which was about half that for [99mTc]nitroimidazole. Although the tissue retention fraction of [99mTc]pentetate was almost twice that of [99mTc]nitroimidazole in normoxic conditions (Table 2⇑), the retention fraction of [99mTc]pentetate remained at about 0.9 with perfusate oxygen level reduced to 25%, while that of [99mTc]nitroimidazole increased to about 5.0. The higher retention of [99mTc]pentetate is probably a result of its ability to bind to the capillary wall and the interstitial space, since it has the same fast clearance rate as [99mTc]TcO4−.
Dependence of Tissue Retention Fraction on Perfusate Oxygen Level
Since the retention fraction of the tracer potentially can be estimated by in vivo imaging, it is imperative that a dependence of tissue retention fraction on perfusate oxygen level be established for the determination of cellular oxygen level. The reductive bioactivation of the nitro group on the imidazole ring requires a cellular enzyme called nitroreductase. Thus, it is reasonable to use the Hill equation developed from the work of enzyme kinetics to describe the relation between tissue retention fraction and fractional perfusate oxygen level. The Hill equation was modified to include a constant of 0.61 to better explain the data that showed an initial plateau. As shown in Fig 6⇓, the relation is well described, with an R value of .99. Constants b, c, and n were determined to be 0.053±0.004, 0.018±0.021, and 7.84±1.94, respectively. Large parameter errors were expected because only six groups with five hearts each were used for the estimations. It is important to note that cellular oxygen level must be reduced to at least 60% for the enzyme to be activated.
Binding Kinetics of [99mTc]Nitroimidazole in Tissue
Experiments with constant infusion were used to verify the percentage of tissue retention and the kinetic sensitivity of [99mTc]nitroimidazole to acute changes in perfusate oxygen level. Fig 7A⇓ demonstrates that the trapping of the [99mTc]nitroimidazole anion radical, although small, existed in the normoxic conditions. More importantly, the degree of trapping increased as perfusate oxygen level was reduced to 40% and 25%, respectively. Fig 7B⇓ further substantiates the results obtained with the bolus experiments in that the binding of [99mTc]- nitroimidazole anion radical is essentially irreversible. The binding remained irreversible even when the perfusate oxygen level was returned to 100%. The initial portion of the tissue residue curves in both Fig 7A⇓ and 7B⇓ probably is indicative of the filling process of [99mTc]- nitroimidazole in the ventricular chambers. When the perfusate oxygen level was returned to normal in protocol IIB (Fig 7B⇓), the distribution volume showed a slight decrease initially because the binding rate at normoxic conditions was much smaller than that at hypoxic conditions. This change was further confounded by smaller ventricular chambers in the normoxic condition. The binding rates for various perfusion conditions are summarized in Table 3⇓. For perfusate oxygen level of 40%, the binding rates were not significantly different between the two protocols, but the rates were different for 100%. The mechanism for this difference, however, is not clear. It is important to note that the ratio of binding rates in any two perfusion conditions was in the same order as the ratio of retention fractions obtained from the bolus injection experiments (Table 4⇓), thus confirming that the tissue residue curves obtained by bolus injection indeed consisted of the true peaks. Therefore, measurements of tissue retention fractions were valid in the bolus injection experiments.
The current study shows that [99mTc]nitroimidazole (BMS-181321), a new lipophilic 99mTc-labeled compound, is preferentially retained in hypoxic myocardium. The nitro group of the compound is responsible for retention of the tracer in the myocardium, while the remaining portion is responsible for the overall lipid solubility of the molecule. The perfused rat heart model with constant flow was chosen for studying the kinetics of [99mTc]nitroimidazole in myocardium because of the ease of inducing global hypoxia in a well-controlled environment. In this model, the effect of hypoxia on [99mTc]nitroimidazole kinetics could be directly evaluated independent of flow. A period of 15 minutes was used to allow adequate oxygen equilibration in the myocardium before the 99mTc-labeled compounds were administered into the aortic root of the perfused heart. The retention of [99mTc]nitroimidazole in myocardium was found to be inversely proportional to perfusate oxygen level in both bolus injection and constant infusion experiments. Furthermore, results from both sets of experiments demonstrated that the cellular binding of [99mTc]nitroimidazole anion radical was essentially irreversible, although a small amount of nonspecific binding was present in perfusion conditions with normal oxygen level. A threshold for cellular binding was established to be 60% perfusate oxygen level in the bolus injection experiments; ie, perfusate oxygen level must be less than 60% in order for the nitro group of [99mTc]nitroimidazole to be reduced significantly and trapped.
Tracer Kinetics of [99mTc]Nitroimidazole in Myocardium
In rat heart perfused with O2- or N2-equilibrated cell medium, Rumsey et al12 reported that the washout of [99mTc]nitroimidazole displayed biphasic kinetics. Our current data obtained from the bolus injection experiments appeared to display biphasic kinetics, but a triexponential function was clearly required to describe the kinetic behavior of [99mTc]nitroimidazole adequately as determined by the goodness of fit. The difference in kinetics between Rumsey et al12 and the current study may be due to differences in experimental protocols. The slow clearance rate of [99mTc]nitroimidazole was estimated from the bolus experiments to be about 5 hours regardless of the perfusate oxygen level. The optimum duration for the determination of mean life for any counting process with minimum error is of the order of 3 to 5 mean lives.19 Therefore, the accurate determination of the slow clearance rate in the current study will require experiments that last more than 20 hours, an impracticality for isolated heart studies. Since we chose an experimental duration less than 2 hours, estimation of the slow clearance rate is expected to have large variability, as evidenced by results shown in Table 1⇑. On the contrary, the constant infusion experiments with [99mTc]- nitroimidazole can statistically distinguish the tissue binding rate between a perfusate oxygen level of 25% and 40% (Table 3⇑). It appears that the estimation of binding rate with the constant infusion approach is much more sensitive than the estimation of the clearance half-time for the slow phase in the bolus injection experiments. As reported by Smith et al,20 exponential factors and coefficients cannot always be equated respectively with compartmental transport rate constants or compartmental sizes. Compartmental analysis of the tissue residue curves obtained in the bolus injection experiments may provide accurate estimation on the binding rate constant and the dissociation rate constant of [99mTc]nitroimidazole in tissue.
In contrast to [99mTc]TcO4−, tissue kinetics of [99mTc]- nitroimidazole showed a small degree of binding, which was 0.6% of the injected dose, even at normoxic conditions. It is unclear whether or not the binding is related to nonspecific binding of [99mTc]nitroimidazole similar to [99mTc]pentetate, the avidity of the nitroimidazole reductase enzymes for the reduced compound, or a nitro reduction reaction occurring with the intact technetium nitroimidazole. It is conceivable that nitroimidazole is also a substrate for an enzyme that catalyzes the nitroreduction process independent of the oxygen concentration.21 Nonetheless, the level of binding in normoxic conditions was minimal compared with the hypoxic conditions.
By using the multiple indicator–dilution method and the blood-perfused heart preparation, Dahlberg et al22 showed that [99mTc]nitroimidazole had a high initial extraction during both ischemic and hypoxic conditions, followed by a rapid clearance. However, they found that low-flow ischemia significantly increased the net extraction (retention) of [99mTc]nitroimidazole at 15 minutes, whereas hypoxia at normal flow did not alter the net extraction of [99mTc]nitroimidazole. As demonstrated by the current study, tissue retention in hypoxia, which was significantly different from normoxia, could be obtained only when a threshold of perfusate oxygen level was reached. Accordingly, they may not have chosen a level beyond this threshold.
The sigmoidal response provides a much more sensitive control of the retention rate in response to variations in perfusate oxygen level. In other words, once the threshold of 60% oxygen level is reached, any subsequent reduction of oxygen will cause a substantial amount of the [99mTc]nitroimidazole anion to be trapped intracellularly. The sigmoidal response suggests that the cellular nitroreduction process for [99mTc]nitroimidazole probably requires an enzyme that is dependent on oxygen concentration. A second possible mechanism may be the decrease in oxygen concentration is important for the increase in the amount of reduced nitroimidazole and the concentration of reducing equivalents being critical for the rate of nitroreductase reactions. If the sigmoidal relation holds true in the in vivo situation, myocardial oxygen level then can be estimated from the accumulation of tissue radioactivity. Myocardial studies with 18F-labeled fluoromisonidazole in intact dogs indicate that perfusion of less than 40% of normal appeared necessary to increase extraction of tracer, suggesting that this level of ischemia is necessary to create tissue hypoxia.6 This finding further supports the sigmoidal response obtained in the current study.
Tissue retention of [99mTc]nitroimidazole obtained in the bolus injection experiments can be affected easily by the detection of true peaks. Since the clearance of [99mTc]nitroimidazole was found to be very slow, we chose to perform a second set of experiments with the constant infusion technique, which is free from the problem of peak detection. In addition, the constant infusion experiments provide data that reflect the accumulation of [99mTc]nitroimidazole in tissue rather than its clearance from the tissue. The binding rates observed in normoxic conditions were significantly different from each other in the protocols. In protocol IIB, hearts were first perfused with 40% of oxygen level, then with 100%. The data appear to suggest that the initial perfusion with 40% oxygen level caused the [99mTc]nitroimidazole radical anion to occupy some binding sites that were not available for binding when the perfusate oxygen level was changed to 100%. The constant infusion experiments can be used to compare different hypoxic agents in terms of their binding ability.
Considerations in the Experimental Model
There is considerable disagreement about the adequacy of oxygenation in isolated hearts perfused with Krebs-Henseleit buffer. Issues concerning the validity of using the buffer-perfused hearts for physiological and biochemical measurements have been addressed extensively by Opie.23 Furthermore, Leiris and coworkers24 outlined the advantages of using isolated heart preparations in studies designed to investigate the effect of hypoxia or ischemia on myocardial cells. Because our major objective for the current study was to investigate the effect of hypoxia on the tissue kinetics of [99mTc]nitroimidazole, we therefore chose the isolated heart preparation perfused with buffer.
Slight hypoxia in the isolated heart preparation will cause additional tissue binding of [99mTc]nitroimidazole in the current study. If the preparation was becoming more hypoxic with time, the binding rate of [99mTc]nitroimidazole in the normoxic phase obtained in protocol IIB would have been larger than that obtained in protocol IIA. At the end of each 2-hour experiment, the perfused heart plus its water content rarely weighed more than 1.5 g. As shown in Fig 7A⇑, the distribution volume already increased as much as 1.2 mL over 30 minutes at 100% perfusate oxygen level. Because edema merely increases the water space for tracer accumulation, it is very unlikely that edema is the major cause for the temporal change in tissue binding kinetics of [99mTc]nitroimidazole in the constant infusion experiments.
Although we did not measure intracellular oxygen concentration, we allowed a period of 15 minutes for the perfused hearts to equilibrate with the perfusate oxygen concentration. Sarcoplasmic oxygen pressure was found to be nearly identical to the extracellular oxygen pressure in resting cells and only about 2 mm Hg less in rapidly respiring myocytes.25 Since differences of oxygen pressure within the cardiac myocyte are very small, the perfusate oxygen level in the current study should be an adequate indicator of the intracellular oxygen level. Furthermore, measured Mvo2 was previously found to be inversely proportional to the perfusate oxygen level in the same perfused rat heart preparation.26 The perfusate oxygen level was probably a suitable index for the indication of the oxygen state in the perfused heart.
Implications for In Vivo Imaging
The current study provides evidence that [99mTc]nitroimidazole may potentially serve as an imaging marker for hypoxic myocardium. Selective retention of [99mTc]nitroimidazole in hypoxic myocardium could be enhanced by more than ninefold (Table 2⇑, 5.94/0.61), depending on the severity of cellular oxygen deprivation. The sigmoidal curve in Fig 6⇑, however, shows that profound changes in perfusate oxygen levels of 20% to 50% may be associated with little changes in tracer retention. This steep gradient may pose a sensitivity limitation in the in vivo studies with [99mTc]nitroimidazole and SPECT. Preliminary data studied with [99mTc]nitroimidazole in a canine model with a severe left anterior descending coronary artery stenosis demonstrated preferential uptake of the tracer in the ischemic zone.11 The high uptake and slow clearance of [99mTc]nitroimidazole in the liver,10 11 however, will further complicate the feasibility of using this 99mTc-labeled compound clinically for the positive identification of ischemic myocardium. Because the current data suggest that a threshold exists for the bioactivation of [99mTc]nitroimidazole in the myocardium, the limitation of using this radiotracer in low-flow ischemia is yet to be determined in patients.
It is important to note that the [99mTc]nitroimidazole compound evaluated in the current study is just one member in a family of nitroimidazole compounds. Results from the current study support further investigation to identify which nitroimidazole compound has the appropriate characteristics for in vivo imaging. Many nitroimidazole compounds are being investigated as a radiosensitizer for tumor therapy.27 In addition, preliminary results indicate that [99mTc]nitroimidazole can be used to identify ischemic tissue at risk for infarction during stroke in a cat model.28 Thus, in vivo imaging with radiolabeled nitroimidazoles potentially has a wide range of applications not only in cardiology but also in neurology and oncology.
The initial rapid clearance of [99mTc]nitroimidazole is independent of perfusate oxygen level. Approximately 0.6% of the injected [99mTc]nitroimidazole was retained in the normoxic myocardium; thus, the nonspecific binding is minimal. The retention of [99mTc]nitroimidazole is inversely proportional to the perfusate oxygen level in a sigmoidal manner. Perfusate oxygen level must be reduced to below a threshold of 60% to have significant trapping of [99mTc]nitroimidazole. The binding of [99mTc]- nitroimidazole in the hypoxic myocardium is essentially irreversible. Furthermore, binding rate appears to be a more sensitive parameter for the measurement of change in perfusate oxygen level than the slow clearance rate.
This work was supported in part by a Grant-in-Aid from the Connecticut Affiliate of the American Heart Association to Dr Ng. The authors would like to thank Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton, NJ, for providing the materials required to prepare the hypoxia localizing agent 99mTc-labeled TcO(PnAO)-1-(2-nitroimidazole) (BMS-181321).
Presented in part at the American College of Cardiology, Atlanta, Ga, 1994, and at the Society of Nuclear Medicine, Orlando, Fla, 1994.
- Received October 12, 1994.
- Revision received February 16, 1995.
- Accepted February 28, 1995.
- Copyright © 1995 by American Heart Association
McGhie AI, Gould KL, Schelbert HR, Willerson JT. Coronary artery disease: nuclear cardiology. In: Willerson JT, Cohn JN, eds. Cardiovascular Medicine. New York, NY: Churchill-Livingstone; 1995:475-527.
Martin GV, Caldwell JH, Graham MM, Grierson JR, Kroll K, Cowan MJ, Lewellen TK, Rasey JS, Casciari JJ, Krohn KA. Noninvasive detection of hypoxic myocardium using fluorine-18-fluoromisonidazole and positron emission tomography. J Nucl Med. 1992;33:2202-2208.
Shelton ME, Dence CS, Hwang DR, Herrero P, Welch MJ, Bergmann SR. In vivo delineation of myocardial hypoxia during coronary occlusion using fluorine-18 fluoromisonidazole and positron emission tomography: a potential approach for identification of jeopardized myocardium. J Am Coll Cardiol. 1990;16:477-485.
Shelton ME, Dence CS, Hwang DR, Welch MJ, Bergmann SR. Myocardial kinetics of fluorine-18 misonidazole: a marker for hypoxic myocardium. J Nucl Med. 1989;30:351-358.
Martin GV, Cerqueira MD, Caldwell JH, Rasey JS, Embree L, Krohn KA. Fluoromisonidazole a metabolic marker for myocyte hypoxia. Circ Res. 1990;67:240-244.
Martin GV, Caldwell JH, Rasey JS, Grunbaum Z, Cerqueira M, Krohn KA. Enhanced binding of the hypoxic cell marker [3H]fluoromisonidazole in ischemic myocardium. J Nucl Med. 1989;30:194-201.
Rumsey WL, Patel B, Kuczynski B, Bauer A, DiRocco R, Pirro J, Nunn AD, Strauss HW, Narra RK. Planar and SPECT imaging of ischemic canine myocardium using a novel [99m]Tc-nitroimidazole. J Nucl Med. 1993;34:15P. Abstract.
Shi Q, Dione DP, Singer MJ, Heller E, Young L, Wackers F, Zaret BL, Sinusas AJ. Technetium-99m nitroimidazole: a positive imaging agent for the detection of myocardial ischemia. Circulation. 1993;88(suppl I):I-249. Abstract.
Sorenson JA, Phelps ME. Physics in Nuclear Medicine. Orlando, Fla: Grune & Stratton, Inc; 1987:475.
Segel IH. Biochemical Calculations. 2nd ed. New York, NY: John Wiley & Sons; 1976:309-312.
Gambhir SS, Schwaiger M, Huang SC, Krivokapich J, Schelbert HR, Nienaber CA, Phelps ME. Simple noninvasive quantification method for measuring myocardial glucose utilization in humans employing positron emission tomography and fluorine-18 deoxyglucose. J Nucl Med. 1989;30:359-366.
Wallenstein S, Zucker CL, Fleiss JL. Some statistical methods useful in circulation research. Circ Res. 1980;47:1-9.
Evans RD. The Atomic Nucleus. New York, NY: McGraw-Hill; 1955:813-815.
Dahlberg ST, Gilmore MP, Flood M, Leppo JA. Effect of hypoxia and low-flow ischemia on the myocardial extraction of technetium-99m-nitroimidazole. Circulation. 1993;88(suppl I): I-250. Abstract.
Wittenberg BA, Wittenberg JB. Oxygen pressure gradients in isolated cardiac myocytes. J Biol Chem. 1985;260:6548-6554.
Ng CK, Huang SC, Schelbert HR, Buxton DB. Validation of a model for [1-11C]acetate as a tracer of cardiac oxidative metabolism. Am J Physiol. 1994;266:H1304-H1315.
Rockwell S. Use of hypoxia-directed drugs in the therapy of solid tumors. Semin Oncol. 1992;19:29-40.