Although extensive evidence has accumulated to support the role of serotonin (5-hydroxytryptamine, 5-HT) in major depression, Reference Yatham, Liddle, Shiah, Scarrow, Lam and Adam1–Reference Bhagwagar, Whale and Cowen3 the precise nature of alterations in the 5-HT system that underlie manic symptoms still remains elusive. For instance, the studies of cerebrospinal fluid 5-hydroxyindoleacetic acid (5-HIAA) levels, platelet 5-HT uptake or tritiated imipramine binding sites in people with mania have yielded inconsistent results (see Shiah & Yatham Reference Shiah and Yatham4 for a review). Neuroendocrine studies have reported unaltered prolactin responses to buspirone and dl-fenfluramine, growth hormone responses to sumatriptan, blunted prolactin responses to d-fenfluramine and enhanced cortisol responses to ipsapirone and 5-hydroxytryptophan in people with mania compared with healthy controls. Reference Shiah and Yatham4 Taken together, these studies suggest a reduction in presynaptic 5-HT activity, enhanced post-synaptic 5-HT1A receptor and likely 5-HT2 receptor sensitivity, and unaltered 5-HT1D receptor sensitivity in mania. Reference Shiah and Yatham4 These studies, however, were limited by the fact that they provide information about 5-HT activity in the hypothalamic region only and not in other brain regions. Although numerous positron emission tomography (PET) studies have assessed brain 5-HT2 receptors in people with major depression, Reference Yatham, Liddle, Shiah, Scarrow, Lam and Adam1,Reference Meyer2,Reference Bhagwagar, Hinz, Taylor, Fancy, Cowen and Grasby5 no study to date measured brain 5-HT2 receptors in people with acute mania. The purpose of the present study, therefore, was to assess brain 5-HT2 receptor density in drug-free or drug-naive people with acute mania in comparison with healthy individuals using [18F]setoperone with PET.
Method
Participants
Individuals with current mania (n = 10) between the ages of 18 and 65, who fulfilled DSM–IV criteria for bipolar I disorder – manic episode and were able to give informed consent were recruited for the study. 6 All study participants were assessed by a structured clinical interview for DSM–IV diagnosis (SCID). Reference First, Spitzer, Gibbon and Williams7 The DSM–IV diagnosis of mania was arrived at by the consensus of a research team based on information from an unstructured clinical interview as well as SCID. The severity of manic symptoms in participants was assessed by the Young Mania Rating Scale (YMRS). Reference Young, Biggs, Ziegler and Meyer8
Participants were either medication naive or psychotropic medication free for at least 2 weeks (6 weeks free for fluoxetine) with the exception of lorazepam prior to the PET scan. Patients with an Axis I comorbidity were excluded as were the individuals with a history of alcohol or drug misuse within the previous 6 months. Those with current major medical illnesses, and women of childbearing potential who were pregnant or not taking contraceptive measures were also excluded.
We also recruited healthy controls (n = 10) matched for age (within 6 years) and gender with the individuals with mania. The controls were assessed by a structured clinical interview for DSM–IV diagnosis non-patient version (SCID–NP) Reference First, Spitzer, Gibbon and Williams9 to determine lifetime history of a psychiatric diagnosis. The controls had no lifetime history of psychiatric diagnosis or family history of a mood disorder. All participants in the control group were medication free for at least 6 weeks and had no history of any major medical illnesses. Women of childbearing age who were pregnant or not taking contraceptive measures were excluded.
PET procedure
The procedure for scanning at our centre has been described previously. Reference Yatham, Liddle, Shiah, Scarrow, Lam and Adam1,Reference Yatham, Liddle, Dennie, Shiah, Adam and Lane10 Briefly, individuals were escorted to the PET suite by a research coordinator. Each participant had a transmission scan for 10 min for attenuation correction for PET images. The [18F]setoperone was synthesised as previously described and each participant had 4 to 7 mCi of [18F]setoperone injected intravenously. Individuals were then scanned with PET camera ECAT/953B (Siemens, Knoxville, TN, USA) for 110 min as previously described.
All study procedures were approved by the clinical research ethics board of the University of British Columbia and a written informed consent was obtained from all study participants.
Data analysis
Demographic and clinical variables
Student's t-tests were used to examine the differences in age between individuals with mania and controls. Relationships between 5-HT2 receptor binding potential and YMRS scores were assessed using Pearson's correlation coefficient. All tests were two-tailed, with significance set at P<0.05.
5-HT2 receptor binding potential
The rationale and methods for determining 5-HT2 receptor binding potential have been described previously in detail elsewhere. Reference Yatham, Liddle, Shiah, Scarrow, Lam and Adam1 Briefly, cerebellum was used as a reference region and the region/cerebellum ratios from scans during the pseudo-equilibrium period (70 to 110 min post-injection) were used to compute 5-HT2 receptor binding potential images for each participant. Previous studies have shown that the binding potential estimated using the ratio method correlates well with the other invasive and non-invasive methods, Reference Kapur, Jones, DaSilva, Wilson and Houle11 and this approach has been used in a number of other studies to estimate 5-HT2 receptor binding potential. Reference Yatham, Liddle, Shiah, Scarrow, Lam and Adam1,Reference Meyer, Kapur, Houle, DaSilva, Owczarek and Brown12,Reference Kapur, Zipursky, Remington, Jones, McKay and Houle13 Statistical parametric mapping (SPM5; Wellcome Department of Cognitive Neurology, Institute of Neurology, London; www.fil.ion.ucl.ac.uk/spm/software) software run on Matlab 7 (Mathworks Inc, Natick, MA, USA) was used for aligning frames from 70 to 110 min to create a mean image for each participant. The mean image was then coregistered to each person's magnetic resonance (MR) image. Each participant's PET and MR images were normalised to the standard coordinate frame (Montreal Neurological Institute (MNI) template) used in SPM5. Then a 5-HT2 receptor binding potential image was created for each participant by dividing each pixel in the normalised mean image with that image's average cerebellar value. A mean activity value from two large regions of interest (one on the right and one on the left) drawn on three contiguous cerebellar slices was used as that image's average cerebellar value.
In order to assess the relationship between 5-HT2 receptor binding potential and age, we extracted the 5-HT2 receptor binding potential values for frontal, temporal and parietal grey matter regions. The Pick_Atlas program (Wake Forest University School of Medicine, Winston-Salem, NC, USA; www.ansir.wfubmc.edu) run on Matlab 7 was used to create a grey matter mask for frontal, temporal, and parietal cortical regions. The mask for each cortical region was applied to the 5-HT2 receptor binding potential images of each participant and the mean for each cortical grey matter region was extracted.
SPM analysis
Since 5-HT2 receptor binding potential significantly declined with age in both individuals with mania and controls, the differences in 5-HT2 receptor binding potential between participants with mania and healthy individuals was computed using age as a covariate with the SPM5. A 12 mm Gaussian filter was used to smooth the binding images before SPM analysis was performed. The grey matter threshold was set at 130% of the mean image intensity as this threshold eliminated most white matter voxels and included most grey matter voxels. In addition to examining the significance of difference in 5-HT2 receptor binding potential for each voxel between the two groups, we also computed the significance of clusters of contiguous voxels in which the difference in 5-HT2 receptor binding potential exceeded a threshold of z = 2.57 corresponding to P = 0.01. The method implemented in SPM5 calculates the significance of clusters using cluster criteria that takes into account the number of contiguous voxels. The corrected cluster significance was set at P<0.05.
Results
Ten people with mania (three males and seven females) and ten controls (three males and seven females) participated in the study. The details of clinical characteristics and demographics for the ten people with mania are presented in Table 1.
Sociodemographic and clinical characteristics of the ten participants with acute mania
| Participant number | Age, years | Young Mania Rating Scale score on scan day | Drug status | Duration of current episode, weeks | Previous depressive episodes, n | Previous manic episodes, n |
|---|---|---|---|---|---|---|
| Female | ||||||
| 1 | 59 | 23 | Naive | 2 | 0 | 0 |
| 2 | 25 | 24 | 2 weeks free | 3 | 10 | 10 |
| 3 | 21 | 28 | >3 years free | 4 | 0 | 1 |
| 4 | 25 | 17 | Naive | 8 | 0 | 2 |
| 5 | 32 | 32 | Naive | 2 | 0 | 0 |
| 6 | 52 | 25 | Naive | 1.5 | 0 | 0 |
| 7 | 39 | 48 | Naive | 3 | 0 | 0 |
| Male | ||||||
| 8 | 29 | 24 | >2 weeks free | 5 | 0 | 0 |
| 9 | 30 | 27 | Naive | 4 | 0 | 0 |
| 10 | 28 | 23 | Naive | 2 | 1 | 0 |
| Mean 34 (s.d. = 12.4) | Mean 27.1 (s.d. = 8.3) |
Of the ten people with mania, seven were drug naive and the other three were drug free (one for more than 3 years, one for more than 2 weeks, and one had sertraline discontinued about 2 weeks prior to the baseline PET scan). Seven participants were in their first manic episode and the other three had one to ten previous manic episodes. Eight of the ten had no previous depressive episodes but one had one episode and the other had about ten previous depressive episodes. The duration of the current manic episode for individuals ranged from 1 to 8 weeks. The mean YMRS was 27.1 (s.d. = 8.3). There was no significant difference in age between people with mania (mean 34, s.d. = 12.4) and controls (mean 32.5, s.d. = 13.71) (P = 0.8).
As expected, there was a significant negative correlation between age and the 5-HT2 receptor binding potential (frontal r = −0.81; temporal r = −0.85; parietal r = −0.71; P<0.001 for all regions examined) and this correlation was similar in both groups.
Given that age had a significant negative correlation with 5-HT2 receptor binding potential, we assessed difference in 5-HT2 receptor binding potential between people with mania and controls using age as a covariate. The SPM analysis using age as a covariate revealed that binding potential was significantly lower in various cortical regions in people with mania compared with controls. The reductions in binding potential were observed in an extensive cluster of voxels (corrected P = 0.016) bilaterally in cortical regions but changes were more prominent in the right frontal, temporal, parietal and occipital cortical regions (Fig. 1). This cluster had 23 974 grey matter voxels. The mean decrease in binding potential for the entire cluster was 19.9%. There were a number of individual voxels in which the reductions in binding potential met significance criteria for false discovery rate (P<0.025 after correction for multiple comparisons). The voxels that showed the most significant decrease (z≥3.1) in binding potential were located in the right fusiform gyrus, right insula, right inferior temporal gyrus, right middle occipital gyrus, right medial frontal gyrus and right middle frontal gyrus. The 5-HT2 receptor binding potential remained significantly lower in individuals with mania when the analysis was repeated without controlling for age as a covariate. Further, to exclude the possibility that reductions in brain 5-HT2 receptor binding potential in our study participants are a result of reductions in brain grey matter volumes, we compared brain volumes and total grey matter volumes between participants with mania and controls in this study. The results showed no differences in either measure between the two groups.
Statistical parametric maps of t-values displayed as projections on the sagittal (a), coronal (b) and transverse (c) renderings of the brain.
These projections illustrate regions of significantly decreased [18F]setoperone binding potential in participants with acute mania compared with matched healthy controls.
We found no increase in 5-HT2 receptor binding potential in people with mania in any of the brain areas. There was no significant correlation between 5-HT2 receptor binding potential and YMRS scores in participants with mania. There was no significant correlation between the duration of manic episode and the 5-HT2 receptor binding potential.
Discussion
Given the difficulty recruiting drug-naive or drug-free people with mania, it is not surprising that this is the first study to examine brain 5-HT2 receptors in this population. It took over 4 years to recruit 10 people with acute mania who were drug naive or free and able to provide informed consent. The results showed that 5-HT2 receptor binding potential is decreased in frontal, temporal, parietal and occipital cortical regions, suggesting a reduction in brain 5-HT2 receptor density in these regions in individuals with mania. Furthermore, consistent with previous studies, the results of this study indicate that 5-HT2 receptor density decreases with age.
5-HT2 receptors in platelets and brain
Previous studies in mood disorders have used platelets as models for brain 5-HT neurons to assess various aspects of 5-HT function. Two previous studies that assessed 5-HT2 receptors in mania in platelets reported conflicting findings, with one reporting no change Reference Velayudhan, Sunitha, Balachander, Reddy and Khanna14 whereas the other reported an increase in 5-HT2 receptor density Reference Pandey, Pandey, Ren, Dwivedi and Janicak15 in drug-free people with mania compared with healthy controls. However, it must be remembered that although platelet 5-HT2 receptor protein is very similar to brain 5-HT2 receptor protein, the platelets' 5-HT2 receptors are not subjected to the same sort of regulatory mechanisms as brain 5-HT2 receptors, which are extensively modulated because of inter-neuronal connections. Reference Cho, Kapur, Du and Hrdina16 The present study and several previous studies have shown that brain 5-HT2 receptor density decreases with age but no such correlation was found between platelet 5-HT2 receptors and age. Reference Pandey, Pandey, Ren, Dwivedi and Janicak15 Furthermore, PET studies by our group Reference Yatham, Steiner, Liddle, Shiah, Lam and Zis17 and other groups Reference Cho, Kapur, Du and Hrdina16 have shown that the 5-HT2 receptor density in platelets does not correlate with brain 5-HT2 receptors, thus raising questions about the utility of gleaning information on brain 5-HT status by studying 5-HT binding in platelets.
Limitations
Some limitations of this study must be considered. First, the number of individuals with mania studied was small. However, we have applied a statistical test that provides a strict control against type I error and hence it is unlikely that the reduction in 5-HT2 receptors observed in the participants with mania in this study arose purely by chance.
Second, the study sample was somewhat unusual in that only two of the ten individuals in the mania sample had previous depressive episodes and that two had their first manic episode after age 50. However, it must be remembered that mania is not uncommon after age 50 and that all medical causes of manic symptoms, including substance misuse, were excluded in our participants. Nonetheless, given the smaller sample size and the nature of our study sample, a high degree of caution is warranted in the interpretation of findings as they might not be generalisable to the entire population with mania.
Third, the 5-HT2 binding potential in our study was estimated using cortex/cerebellum ratios and not by measuring maximum binding potential (B max) and the dissociation constant (K d) using arterial input function. Given this, we cannot exclude the possibility that reduced 5-HT2 binding potential was a result of changes in K d and not a result of changes in B max. This, however, is unlikely because previous studies of alteration in receptor binding in other psychiatric conditions have shown changes in B max and not K d, and thus the reduction in 5-HT2 binding potential in our study likely indicates reduced 5-HT2 receptors in people with mania. Furthermore, although the two previous platelet studies in mania yielded conflicting findings with regard to changes in 5-HT2 receptor density, both studies reported no alteration in K d for 5-HT2 receptors. Reference Velayudhan, Sunitha, Balachander, Reddy and Khanna14,Reference Pandey, Pandey, Ren, Dwivedi and Janicak15
Fourth, it is possible that the reduction in 5-HT2 binding potential is because of an increase in 5-HT in the synaptic space in people with acute mania that might be expected to occupy 5-HT2 receptors, thus leaving a fewer receptors for the setoperone to bind. This, however, is unlikely as most studies of 5-HT metabolite 5-HIAA levels in people with mania have not reported any consistent increases but rather lower levels compared with healthy controls. Further, PET studies have shown that increases in 5-HT levels in synaptic space with fenfluramine do not result in any changes in 5-HT2 Reference Hirani, Sharp, Sprakes, Grasby and Hume18 or 5-HT1A receptor binding Reference Hume, Hirani, Opacka-Juffry, Myers, Townsend and Pike19,Reference De Haes, Harada, Elsinga, Maguire and Tsukada20 suggesting that endogenous 5-HT levels do not affect estimates of 5-HT2 receptor density with PET.
Fifth, no studies to date assessed the effects of lorazepam on brain 5-HT2 receptors and hence it is not possible to exclude the effects of lorazepam on 5-HT2 receptors as some of our sample with acute mania but not controls received lorazepam prior to scanning.
Reduction in brain 5-HT2 receptor density: a state or trait marker?
Since some Reference Yatham, Liddle, Shiah, Scarrow, Lam and Adam1,Reference Biver, Wikler, Lotstra, Damhaut, Goldman and Mendlewicz21–Reference Messa, Colombo, Moresco, Gobbo, Galli and Lucignani23 but not all Reference Meyer, Kapur, Houle, DaSilva, Owczarek and Brown12 previous studies that examined brain 5-HT2 receptors in people with major depression have reported a reduction in these receptors compared with controls, and given the fact that we found reduced 5-HT2 receptors in people with mania, one could argue that a reduction in brain 5-HT2 receptors predisposes individuals to both depression as well as mania. However, decreased brain 5-HT2 receptor density is unlikely to predispose individuals to depression because several effective antidepressant treatments also down-regulate brain 5-HT2 receptors in people with depression, including the tricyclic antidepressant desipramine, Reference Yatham, Liddle, Dennie, Shiah, Adam and Lane10 the selective serotonin reuptake inhibitor (SSRI) paroxetine, Reference Meyer, Kapur, Eisfeld, Brown, Houle and DaSilva24 the SSRI and 5-HT2 antagonist nefazadone, Reference Mischoulon, Dougherty, Bottonari, Gresham, Sonawalla and Fischman25 as well as the somatic treatment electroconvulsive therapy (ECT) (details available from author on request). Given this, we have previously argued that a reduction in brain 5-HT2 receptors observed in people with depression is not a cause of depression but rather a compensatory mechanism of the brain to cope with the state of depression. Reference Yatham, Liddle, Shiah, Scarrow, Lam and Adam1 Such down-regulation of 5-HT2 receptors is expected to lead to spontaneous remission of depressive symptoms in some individuals but those people that are not able to mount effective compensatory down-regulation of brain 5-HT2 receptors may require treatment with antidepressants or ECT to further down-regulate these receptors to improve from depression. In contrast, antidepressant treatments either induce or worsen manic symptoms in people with bipolar disorder. Reference Leverich, Altshuler, Frye, Suppes, McElroy and Keck26–Reference Yatham, Calabrese and Kusumakar28 Given that antidepressant treatments down-regulate brain 5-HT2 receptors, the propensity of antidepressants to worsen/induce mania is consistent with the hypothesis that a reduction in brain 5-HT2 receptors either predisposes individuals to mania or is a cause of mania. The fact that YMRS scores did not correlate with 5-HT2 binding and that treatment of acute mania with valproate does not alter brain 5-HT2 receptors Reference Yatham, Liddle, Lam, Adam, Solomons and Chinnapalli29 lends support to the hypothesis that decreased 5-HT2 receptor density is a trait marker for bipolar disorder. However, studies that assessed the association between 5-HT2 receptor polymorphisms and bipolar disorder have yielded no consistent findings. Reference Ranade, Mansour, Wood, Chowdari, Brar and Kupfer30–Reference Mahieu, Souery, Lipp, Mendelbaum, Verheyen and DeMaertelaer33
Reduction in brain 5-HT2 receptors and increased dopamine in mania
Regardless of whether a reduction in brain 5-HT2 receptors is a state or a trait marker for mania, it is important to reconcile this observation with the fact that drugs that reduce dopamine transmission are effective anti-manic agents whereas drugs that increase dopamine transmission induce or worsen mania. Reference Silverstone34 Interestingly, there is evidence that 5-HT2 receptors are located on dopaminergic neurons. Reference EberleWang, Mikeladze, Uryu and Chesselet35–Reference Doherty and Pickel37 Within the 5-HT2 receptor family, 5-HT2A receptors facilitate stimulated but not basal dopamine release in the nucleus accumbens and striatum Reference Schmidt, Fadayel, Sullivan and Taylor38,Reference Schmidt and Fadayel39 whereas 5-HT2C receptors inhibit both basal and stimulated-impulse flow-dependent mesocortical dopamine function. Reference Di Giovanni, De Deurwaerdere, Di Mascio, Di Matteo, Esposito and Spampinato40–Reference Di Matteo, Di Giovanni, Di Mascio and Esposito42 Since the ligand in this study binds to both 5-HT2A and 5-HT2C receptors (although it has higher affinity for 5-HT2A receptors), it is likely that reduced binding observed in people with mania represents a reduction in both these receptor subtypes. Accordingly, a reduction in 5-HT2A receptors will have no effect on basal dopamine release whereas a reduction in 5-HT2C receptors would be expected to be associated with enhanced dopamine release and transmission. This may explain the efficacy of antidopaminergic drugs for mania.
Brain 5-HT2 receptors and grey matter volume
Previous studies of people experiencing a first manic episode have not shown any consistent reductions in grey matter volumes compared with age- and gender-matched controls. Reference Yatham, Lyoo, Liddle, Renshaw, Wan and Lam43 Further, a preliminary analysis of our study sample has indicated no changes in either total brain volume or grey matter volume compared with matched controls. Thus, the reduction in 5-HT2 receptor density observed in our sample cannot be attributable to changes in grey matter volume. Interestingly, the reduction in 5-HT2 receptors observed in our sample was diffuse and included frontal, temporal, parietal and occipital cortical regions although the changes were more prominent in the right limbic regions. A reduction in 5-HT2C receptors in these regions is expected to be associated with enhanced dopamine release. Further studies should assess both 5-HT2C receptors and the magnitude of dopamine release in acute mania to verify this hypothesis.
eLetters
No eLetters have been published for this article.