The mixed β-adrenergic/serotonin type 1A (5-HT_1A) receptor ligand pindolol has been used as a novel strategy to augment selective serotonin reuptake inhibitors (SSRIs) in pharmacological treatment of major depressive disorder. However, the 11 controlled clinical trials of pindolol augmentation to date have produced inconsistent antidepressant effects (1).

In animals, when SSRIs are administered, the consequent short-term elevation of 5-HT in the midbrain raphe nuclei, the site of the 5-HT cell bodies, stimulates the 5-HT_1A autoreceptor and thus attenuates further 5-HT release from terminal projections. With prolonged exposure to SSRIs the raphe nuclei 5-HT_1A autoreceptors become desensitized and normal cell firing resumes, with concomitant increase of extracellular 5-HT in terminal regions (reviewed by Pineyro and Blier [2]). The time course of 5-HT_1A autoreceptor desensitization has been linked to the 2–3-week delay before the onset of clinical improvement with antidepressant therapy. The addition of pindolol (acting as a 5-HT_1A antagonist at the raphe nuclei autoreceptors) to SSRI treatment would be predicted to block the short-term effect of raised 5-HT level on the autoreceptor, thus causing increased and sustained elevations of 5-HT in terminal regions. Clinically, this effect may speed the onset of SSRI antidepressant action and possibly augment it in resistant cases.

A critical factor in the evaluation of clinical trials is whether the dose of pindolol (2.5 mg t.i.d.) used in all but one trial (3) has been sufficient for consistent occupancy of the 5-HT_1A autoreceptor in the human brain in vivo. This dose was chosen to avoid β-adrenergic side effects, as knowledge of the dose needed to effect consistent occupancy of the 5-HT_1A autoreceptor in humans was not available at the time.

Using positron emission tomography (PET) and [¹¹C]WAY-100635 (a radioligand specific for the 5-HT_1A receptor), we previously measured occupancy of the 5-HT_1A autoreceptor and postsynaptic receptor by single doses of pindolol (5.0 mg, 10.0 mg, and 20.0 mg) in healthy volunteers (4, 5). On the basis of these results and the supporting evidence of Martinez et al. (6), we hypothesized that the typical pindolol dose used in clinical trials was unlikely to achieve consistent 5-HT_1A autoreceptor occupancy in patients. In this study we tested this hypothesis by measuring pindolol occupancy directly in depressed patients receiving SSRI treatment, after one of two pindolol augmentation regimes. A pindolol dose of 2.5 mg t.i.d. was predicted to give inconsistent occupancy, while 5.0 mg t.i.d. was predicted to achieve more consistent effects. It was important to examine medicated depressed patients as concomitant antidepressant medication may influence pindolol plasma levels achieved after oral dosing, and the illness itself may alter 5-HT_1A autoreceptor function.

We examined eight patients (seven male and one female, aged 24–61 years) who met the DSM-IV criteria for major depressive disorder, were receiving therapeutic doses of antidepressant medication only (SSRIs, N=7; venlafaxine, N=1), and were not fully recovered. After giving informed written consent, the subjects underwent two [¹¹C]WAY-100635 PET scans, with an interval of 7–14 days between scans. The first was a baseline scan (during antidepressant-only therapy), while the second followed 1–2 weeks of pindolol augmentation of the antidepressant, at a dose of either 2.5 mg t.i.d. (N=4) or 5.0 mg t.i.d. (N=4). The postdrug scan was conducted 2 hours after the last dose of pindolol, to coincide with the peak plasma level (t_max=1–3 hours). The PET scan acquisition and subsequent analysis generated regional values of binding potential (BP) for [¹¹C]WAY-100635 (5). The percentage occupancy of the 5-HT_1A autoreceptor by pindolol was compared to test-retest data from a control group of 13 healthy volunteers, who were scanned on two occasions without drug administration. Occupancy was calculated as [(binding potential at baseline – binding potential with pindolol)/binding potential at baseline]×100.

One-tailed Student’s t test was used, as we predicted the direction of the change in binding of the radioligand from that in our healthy volunteers. Plasma pindolol levels were measured at 0, 16, 33, 50, and 66 minutes after the start of the postdrug PET scan.

Mean occupancy values are shown in F1. There was considerable variability in the occupancy achieved within both groups. For the four subjects in the 2.5-mg group the raphe nuclei occupancy values (and binding potential at baseline and during pindolol augmentation) were as follows: 32% for subject 1 (baseline, 3.1; pindolol, 2.1), –17% for subject 2 (baseline, 3.6; pindolol, 4.2), 20% for subject 3: (baseline, 2.9; pindolol, 3.5), and 2% for subject 4 (baseline, 2.6; pindolol, 2.5). For the four subjects in the 5.0-mg group the occupancy values were more consistent: 41% for subject 5 (baseline, 5.7; pindolol, 3.3), 18% for subject 6 (baseline, 2.4; pindolol, 2.0), –2% for subject 7 (baseline, 3.4; pindolol, 3.5), and 20% for subject 8 (baseline, 3.2; pindolol, 2.5). Compared with the test-retest data for the control subjects, 2.5 mg t.i.d. of pindolol did not produce consistent occupancy at the 5-HT_1A autoreceptor (mean=–1%, SD=24%) (t=–0.46, df=15, p=0.33). In comparison, the 5.0-mg t.i.d. dose led to a modest but significant occupancy of the 5-HT_1A autoreceptor (mean=19%, SD=18%) (t=–2.77, df=15, p=0.007). After pindolol treatment, scores on the Hamilton Depression Rating Scale and Beck Depression Inventory showed inconsistent changes: for the group receiving 2.5 mg, the mean Hamilton score decreased from 13 to 11 and the mean score on the Beck Depression Inventory decreased from 19 to 13, and for the group receiving 5.0 mg, the mean Hamilton score fell from 21 to 12 and the mean score on the Beck Depression Inventory decreased from 24 to 23. There was no significant correlation between change in score on the Hamilton depression scale or Beck Depression Inventory and either the autoreceptor or postsynaptic receptor occupancy (data not shown). Plasma levels of pindolol were higher in the group receiving 5.0 mg t.i.d. (mean=35.1 ng/ml, range=16–61) than in the group receiving 2.5 mg t.i.d (mean=11.8 ng/ml, range=6–17); however, there was no correlation of plasma level with the occupancy achieved (r=–0.23, df=6, p=0.63).

The occupancy of the autoreceptor achieved by 5.0 mg t.i.d. of pindolol, while significant, was modest (19%) and well short of the occupancy reported for comparable single doses (based on pharmacokinetic considerations) or long-term oral doses of pindolol taken by healthy volunteers. For example, 37% occupancy was achieved with a 10.0-mg single dose of pindolol in healthy volunteers (5) (F1), and 40% occupancy was achieved with a 7.5-mg long-term daily dose (6).

The findings of this in vivo study, of depressed patients receiving SSRIs, confirms predictions from studies of healthy volunteers (5, 6) that the oral dose of pindolol being used to augment SSRI therapy in clinical trials will not lead to a consistent occupancy of the 5-HT_1A autoreceptor. Of note, the plasma pindolol levels achieved by 2.5 mg t.i.d. were comparable to those in the two clinical studies for which plasma pindolol concentrations were reported: 7.0 ng/ml (7) and 9.9 ng/ml (8). Clinical trials have therefore used a suboptimal dose of pindolol, possibly accounting for their inconsistent results. While the magnitude of autoreceptor occupancy needed for the augmentation of SSRI-induced 5-HT release is unknown, a speculation on the basis of rat studies gives a minimum required occupancy of approximately 75%, considerably higher than that observed in our study. This calculation is based on an EC₅₀ (concentration to achieve a 50% effect) of about 0.3 mg/kg i.v. for pindolol occupancy of the 5-HT_1A(9), approximately 10 mg/kg i.p. of pindolol for augmentation of SSRI-induced release of 5-HT (10), and a 10:1 ratio of intraperitoneal-intravenous pindolol dosing. However, increasing the dose of pindolol to achieve effective occupancy will increase the likelihood of β-adrenergic side effects. A complete evaluation of pindolol augmentation of SSRIs will require further clinical work guided by PET occupancy studies.

Received Feb. 2, 2001; revision received May 8, 2001; accepted May 21, 2001. From the MRC Cyclotron Unit, Imperial College School of Medicine, Hammersmith Hospital. Address reprint requests to Dr. Rabiner, MRC Cyclotron Unit, Imperial College School of Medicine, Hammersmith Hospital, Du Cane Road, London W12 0NN, U.K.; eugenii.rabiner@ic.ac.uk (e-mail). The authors thank Andy Blythe, Joanne Holmes, Dave Turton, Ray Khan, Safiye Osman, Keith Poole, Abel Haida, and Len Schnorr for technical assistance in conducting the PET studies and Michael Clement and Michael Franklin for analysis of pindolol plasma levels.