A proof-of-concept study using [11C]flumazenil PET to demonstrate that pagoclone is a partial agonist
The GABA-ergic system is central to many anxiety disorders and has therefore received much attention. Benzodiazepines are highly effective anxiolytics, leading to their widespread use clinically. However, benzodiazepines often have unwanted side effects, such as sedation, amnesia and ataxia, in the short term. With longer use, tolerance and dependence associated with withdrawal have led to their use being limited or discouraged (NICE 2004).
The benzodiazepine receptor site lies within the GABA-A receptor complex, the activation of which increases chloride channel opening, thus increasing inhibitory neurotransmission. Compounds that bind to the benzodiazepine receptor site such as diazepam or alprazolam are full agonists and increase inhibitory activity in the brain. It has been known for many years that certain compounds are partial agonists at the benzodiazepine receptor site. Such compounds do not fully activate the GABA-A system to the same extent as a full agonist, even when occupying all the receptors. Studies in animals suggest the anxiolytic actions of benzodiazepines are seen at relatively low levels of occupation, whereas higher occupation and/or greater efficacy leads to adverse effects such as sedation and ataxia (Facklam et al. 1992). It is possible therefore that partial agonists might exhibit anxiolytic activity with fewer such adverse effects than full agonists. For these theoretical advantages, a number of partial agonists were developed and some (abecarnil, bretazenil) were studied for their efficacy as clinical anxiolytics. They displayed some efficacy and were somewhat freer of adverse effects than full agonists though bretazenil was more sedating than expected.
Pagoclone is a more recently studied compound. A member of the cyclopyrrolone family, it is known to bind to the benzodiazepine receptor site of the GABA-A receptor (Doble et al. 1993). In animal models it is three to 30 times more anxiolytic, whilst being 15–200 times less sedating than full agonists (Doble et al. 1993). In man, pagoclone in phase 1 trials was well tolerated and was anxiolytic without being sedating or increasing EEG beta power, which is consistent with partial agonism (see Bateson 2003). We have also shown anxiolysis in the absence of typical benzodiazepine side effects in a randomised, double-blind, cross-over study of pagoclone in panic disorder (Sandford et al. 2001).
Despite the body of preclinical evidence and limited clinical trial data on partial agonists, as yet there has been no direct demonstration of the partial agonist concept in humans. Perhaps the best way of achieving this is to show a double dissociation between the pharmacodynamic actions and brain receptor occupation of a putative partial agonist compared with a full agonist; a partial agonist would be proven if it had less pharmacodynamic action despite equal or greater brain receptor occupation than a full agonist.
We have conducted such an experiment with pagoclone using an established pharmacokinetic/pharmacodynamic modelling paradigm validated for full agonists with concurrent PET measures of brain receptor occupation. We used [11C]flumazenil PET to measure benzodiazepine receptor occupancy by pagoclone, together with quantifying saccadic eye movements, to assess its pharmacodynamic effects. We chose this measure as the effects of benzodiazepines on saccadic eye movements are well characterized (Ball et al. 1991). For comparison with pagoclone, lorazepam was chosen as a full agonist at a dose (1 mg) used clinically to reduce anxiety and inducing only minimal drowsiness, so saccadic eye movements could be measured. We chose 0.4 mg pagoclone since this dose had been given to healthy volunteers without causing significant problems and was comparable to that used in our previous study (Sandford et al. 2001). If pagoclone is a partial agonist, we expected to show less effect on saccadic eye movements with greater levels of benzodiazepine receptor occupancy than lorazepam. The study was conducted in accordance with the provisions of the Declaration of Helsinki with Good Clinical Practice, and was approved by the local research ethics committee and the UK Administration of Radioactive Substances Advisory Committee.
Six medically and psychiatrically healthy male volunteers were recruited. Each individual underwent two [11C]flumazenil PET scans, a baseline and a ‘drug’ scan. The drug (pagoclone 0.4 mg, lorazepam 1 mg) was taken orally about 90 min before the injection of [11C]flumazenil in a single blind manner with two subjects in each group undergoing the baseline scan first. A bolus injection (370 MBq, 10 mCi) was administered and dynamic images were acquired over 90 min with a brain-dedicated ECAT-953B PET camera as described by Malizia et al. (1998). Spectral analysis of the images with a metabolite corrected plasma input function was used to calculate the volume of distribution (VD) of available GABA-benzodiazepine receptors in the presence and absence of pagoclone or lorazepam. Regions of interest (ROI; frontal cortex and cerebellum and thalamus) were drawn on the VD maps.
To assess the pharmacodynamic effects of lorazepam and pagoclone, saccadic eye movements were measured by using electro-oculography, with the subject required to follow a target displayed in a screen mounted within sight outside the PET camera as described elsewhere (Ball et al. 1991). Data were collected for 48 eye movements of 15–40° at eight time points throughout the PET scan. Peak deceleration, acceleration, velocity, errors and acceleration/deceleration were calculated for each subject. During the scanning procedure, the subject was asked to rate their level of anxiety and sedation using a visual analogue scale (not at all sedated/anxious=0 to the most sedated/anxious ever=100).
Lorazepam levels were measured by Dr. M. Franklin (Oxford University). Pagoclone levels and that of its major metabolite, hydroxy pagoclone, were measured by Interneuron.
The area under the curve (AUC) for time versus drug levels, sedation and anxiety ratings, and each saccadic eye movement parameter, was calculated for each scan and the difference between the baseline and drug scans was compared to the % occupancy in the ROIs.
After taking pagoclone, plasma levels were similar in the three subjects and remained steady throughout the scanning procedure (AUC: 420.9, 337.1, 376 ng/ml). Similarly for two of the subjects who had lorazepam, plasma levels remained constant during the scan (AUC: 3,790 and 5,795 ng/ml); however, one subject had a much higher lorazepam level at the start of the scan that declined during the scan to a level comparable to the other two subjects (AUC: 17,760 ng/ml).
Both lorazepam and pagoclone resulted in reduced [11C]flumazenil uptake, which is consistent with them occupying the benzodiazepine receptor. In all three ROIs, pagoclone occupied more benzodiazepine receptors than lorazepam. In the frontal cortex, the mean occupancy for pagoclone was 14.7±5.7% (range 17, 18.8, 8.2%) compared with 5.6±5.2% (range 1.7, 3.6, 11.5%) for lorazepam. A similar pattern was seen in the thalamus with the mean occupancy for pagoclone, 13.0±2.7% (range 15.6, 13.2, 10.2%), compared to 8.9±5.5% (range 6.5, 15.2, 5.0%) for lorazepam. In the cerebellum, one ‘pagoclone’ scan had insufficient tissue available for analysis but the % occupancy in the remaining two subjects (16.6 and 6.17%) was higher than that for three subjects who took lorazepam (3.9, 4.7 and 13.9%). Notably, the high plasma lorazepam levels seen in one individual were associated with the highest % receptor occupancy.
There were no significant differences in the subjective ratings of anxiety or sedation ratings to lorazepam and pagoclone or between the two drugs. However, lorazepam resulted in the impairment of all the saccadic eye movement (SEM) parameters calculated. For the most sensitive measure (Ball et al. 1991), the mean difference for peak deceleration between the AUC for the ‘drug’ and baseline scans was 803.3±530 deg s−2 (range 1398, 381, 631 deg s−2) for the three subjects who received lorazepam. Although pagoclone also impaired the SEM parameters, it did so to a lesser extent than lorazepam. For example, lower peak deceleration was seen in two of the subjects who received pagoclone (difference between the AUC for the ‘drug’ and baseline scans: 425 and 433 deg s−2) and in one individual, the converse was seen with faster eye movements evident on pagoclone (difference −271 deg s−2; mean of three individuals 176.7±389.1 deg s−2). Notably, this subject had the ‘drug’ scan first. The findings for the other SEM parameters, peak acceleration and peak velocity, were similar in all subjects.
We have shown in this proof-of-concept [11C]flumazenil PET study in a small number of subjects that pagoclone (0.4 mg) resulted in greater occupancy of the benzodiazepine receptor than lorazepam (1 mg), with less impairment of saccadic eye movements. This greater level of occupancy but with a smaller behavioural response is consistent with pagoclone acting as a partial agonist at the GABA-benzodiazepine receptor.
The low % occupancy shown by lorazepam is consistent with previous studies which have estimated that 95% of benzodiazepine receptors are unoccupied when a full agonist shows full behavioural potency. In humans, significant drowsiness and sleep with benzodiazepine agonists is associated with up to 30% occupancy (Malizia et al. 1996; Sybirska et al. 1993). In another study, sleepiness from an oral dose of zolpidem (20 mg) was associated with 21% receptor occupancy (Abadie et al. 1996). Concerning effects on eye movements, in our previous study of midazolam, individuals could not perform the saccadic eye movement task with benzodiazepine receptor occupancy in the same range as we have measured here with pagoclone (Malizia et al. 1996). It is therefore remarkable that after pagoclone, saccadic eye movements could still be measured and were less impaired than with lorazepam.
In conclusion, we have shown in this proof-of-concept study that pagoclone, at a dose that has been shown in studies to be anxiolytic, occupies more benzodiazepine receptors compared with an analogous dose of lorazepam and that this is associated with less effect on saccadic eye movements. This is compatible with pagoclone acting as a partial agonist at the benzodiazepine receptor. Larger studies are needed to confirm this finding to account for confounders such as an order effect and variability in plasma levels.
The authors acknowledge the technical assistance of Mr. N. Uhl and the support of Dr. K D'Orlando, previously of Interneuron. The study was funded by Interneuron.
- Bateson A (2003) Pagoclone indevus. Curr Opin Investig Drugs 1:91–95Google Scholar
- Doble A, Canton T, Dreisler S, Piot O, Boireau A, Stutzmann JM, Bardone MC, Rataud J, Roux M, Roussel G et al (1993) RP 59037 and RP 60503: anxiolytic cyclopyrrolone derivatives with low sedative potential. interaction with the gamma-aminobutyric acidA/benzodiazepine receptor complex and behavioral effects in the rodent. J Pharmacol Exp Ther 266(3):1213–1226PubMedGoogle Scholar
- Malizia AL, Gunn RN, Wilson SJ, Waters SH, Bloomfield PM, Cunningham VJ, Nutt DJ (1996) Benzodiazepine site pharmacokinetic/pharmacodynamic quantification in man: direct measurement of drug occupancy and effects on the human brain in vivo. Neuropharmacology 35(9–10):1483–1491CrossRefPubMedGoogle Scholar
- National Institute for Clinical Excellence (2004) Management of anxiety (panic disorder, with or without agoraphobia, and generalised anxiety disorder) in adults in primary, secondary and community care. Clinical Guideline No 22Google Scholar
- Sybirska E, Seibyl JP, Bremner JD, Baldwin RM, al-Tikriti MS, Bradberry C, Malison RT, Zea-Ponce Y, Zoghbi S, During M et al (1993) [123I]iomazenil SPECT imaging demonstrates significant benzodiazepine receptor reserve in human and nonhuman primate brain. Neuropharmacology 32(7):671–680CrossRefPubMedGoogle Scholar