We appreciate the interest and comments of Alkire et al. regarding our human GABAA-receptor ligand binding data. 1We agree that anesthetics, such as isoflurane, have the potential to confound 11C-flumazenil ligand binding data by various mechanisms, such as alteration of regional cerebral blood flow (rCBF), ligand metabolism, or unmetabolized ligand binding in the plasma.
Although Alkire et al. raise only one of these confounders, it is worth discussing all of them to illustrate the systematic strategy employed to preserve the validity of the data in our article. 1The untoward effect of rCBF changes is adequately addressed in our study by ruling out any correlation between blood flow and ligand binding changes in agreement with previous studies that demonstrated the stability of the binding parameter, distribution volume ratio (DVRATIO), under such circumstances. 2The second potential confounder is related to a change in ligand concentration in the plasma that in turn could result in an error in the calculation of DVRATIOif it remains unaccounted for in the computation process. Specifically, as it is raised by Alkire et al. , the measure of brain tissue drug exposure, expressed as the area under the arterial plasma concentration-versus -time curve (AUC), could be changed by the combined administration of isoflurane and phenylephrine. 3,4This would result from a decrease in the early redistribution of the radioligand increasing its plasma concentration and brain tissue exposure, and that in turn would lead to an overestimation of receptor-specific ligand binding. To prevent this, as described in detail in our article, 1the time course of total 11C-flumazenil concentration was determined in the plasma in each experimental condition obtaining 20 arterial blood samples during the initial 2 min of the study and 20 more samples over the remainder of the 90-min time frame of the study. Furthermore, to account for potentially altered ligand metabolism during the isoflurane conditions, 11C-flumazenil metabolism was monitored in the plasma in each experimental condition by quantifying the time course of unmetabolized parent compound concentration in arterial blood samples obtained at 5, 10, 15, 45, and 75 min after injection. Receptor-specific 11C-flumazenil binding in turn, expressed as DVRATIO, was computed using these measured values. Therefore, any change in available ligand concentration, or brain tissue exposure to the ligand, in the presence of isoflurane and phenylephrine was eliminated as a confounder of receptor-specific 11C-flumazenil binding.
Although the aforementioned approach adequately addresses the concerns of Alkire et al. , to eliminate the third potential confounder, however, i.e. , altered unmetabolized ligand binding in the plasma, we also needed to measure nonreceptor binding in our experiments. This was necessary based on the well-established experimental evidence showing that nonreceptor ligand binding sites represent a nonsaturable compartment where the binding of the radioligand is linearly increasing with its plasma concentration. 5It follows that nonreceptor binding serves as a cumulative index of brain tissue ligand exposure. As shown in table 2 of our article, 1however, no significant differences were observed in nonreceptor ligand binding during the isoflurane conditions. This clearly indicates that the detected changes in binding were not due to altered unmetabolized ligand binding or altered ligand concentration in the plasma due to any mechanism, including the one proposed by Alkire et al.
Taken together, it appears that the employed strategies of measuring total and unmetabolized plasma concentrations of 11C-flumazenil as well as nonreceptor ligand binding in each experimental condition provide adequate protection against confounders, such as anesthesia-induced alterations in plasma tracer concentrations, in the measurement of 11C-flumazenil binding, eliminating the need for further validation experiments.