Molecular Psychiatry (1997) 2, 495–500  1997 Stockton Press All rights reserved 1359–4184/97 $12.00 ORIGINAL RESEARCH ARTICLE consistent with their specific distribution in the human .11,12 Within the range studied, postmortem mRNA interval (PMI) was not found to significantly influence the levels of prodynorphin, , or D1 and D2 receptor mRNAs expressed throughout the striatum expression is increased in (eg Spearman correlation values between PMI and pro- the patch vs matrix mRNA levels measured in the caudate, putamen, and , were −0.0099, compartment of the −0.0474, and 0.0442, respectively). No significant cor- relations were found between proenkephalin, D1,orD2 caudate nucleus in suicide mRNA expression and age, sex, race, blood ethanol level, or freezer storage time of the brain tissue subjects samples. The same applied for the expression of levels of prodynorphin mRNA and these variables, except for YL Hurd1, MM Herman2, TM Hyde2,LB 2 2 2 age, where there was a positive correlation in the Bigelow , DR Weinberger and JE Kleinman nucleus accumbens (Spearman correlation: 0.7592; = 1Karolinska Institute, Dept of Clinical Neuroscience, P 0.0004; with Bonferroni inequality the significance Psychiatry Section, S-171 76 Stockholm, Sweden; 2Clinical level was 0.0017). To consider any possible contri- Brain Disorders Branch, NIMH, Neuroscience Center at St bution of age, ANOVA statistical evaluations were also carried out with covariance for age. Elizabeths Hospital, Washington DC 20032, USA As presented in Table 1, no significant group differ- ences were seen in the mRNA expression levels of pro- Keywords: ; limbic; striatum; or D1 and D2 receptors measured depression; human; in situ hybridization; enkephalin; dopa- within the caudate, putamen, or nucleus accumbens. mine receptors Also, no significant group differences were apparent Experimental and clinical studies suggest an involve- for the prodynorphin hybridization signals measured ment of the opioid system in psychiatric within the total caudate, putamen, and nucleus accum- disorders. Notably, opioid immunoreactivity is bens. Of the markers tested, however, the mRNA altered in the cerebrospinal fluid of chronic schizo- expression of prodynorphin in the human striatum phrenics and manic-depressive subjects.1–3 Despite showed a pronounced heterogeneous distribution pat- these clinical findings, few postmortem investigations4,5 tern, with cell clusters of high mRNA expression sur- have examined the association of endogenous opioid rounded by regions of low expression (Figure 1), con- neuropeptides with schizophrenia and suicide. Anatom- sistent with its patch/matrix striatal organization.9 ically, a tight interaction exists within the neostriatum ANOVA with age as a covariate revealed a significant between the (dynorphin and enkephalin) group difference (P = 0.027) of the prodynorphin system and classical such as dopam- mRNA expression in the caudate nucleus patch com- ine6 which has been implicated in both the psychotic partment: suicide subjects expressed higher levels of symptoms and the cognitive deficits that characterize 7 prodynorphin mRNA than that measured in schizo- schizophrenia (see review). The neostriatum is differen- Յ tially organized into patch and matrix phrenics (P 0.05; Tukey-Kramer), and a similar trend mosaic compartments anatomically connected to limb- (P Յ 0.1; Tukey-Kramer) when compared to normal ic- and sensorimotor-related brain regions, respect- controls (Table 1 and Figure 1). Prodynorphin mRNA ively.6,8 Moreover, the human neostriatum is charac- expression in the caudate nucleus patch compartment terized by a heterogenous expression of the of schizophrenic subjects was not significantly differ- prodynorphin opioid gene: high in the patch, but low in ent from normal controls. No significant group differ- the matrix compartment.9,10 The present results show ences were found for prodynorphin mRNA expression for the first time a differential alteration of prodynorphin levels measured in the matrix compartment of the cau- within distinct striatal compartments in postmortem date nucleus. Thus overall, the same group difference tissue from nonschizophrenic suicide subjects. The pro- (P = 0.018) was apparent for the prodynorphin dynorphin patch/matrix mRNA expression was elevated in the caudate nucleus of suicide subjects as compared patch/matrix ratio in the caudate nucleus: high to normal controls and schizophrenics in which no alter- patch/matrix prodynorphin mRNA expression in the suicide subjects vs the control (P Յ 0.05) and schizo- ations in opioid or D1 and D2 mRNA expression were apparent. Altogether the findings suggest that dis- phrenic (P Յ 0.03) groups, and no change from control crete dysfunction of the endogenous opioid dynorphin for the schizophrenics. It was noted that although all system might contribute to depression and the risk of suicide victims showed elevated patch/matrix prody- suicide in nonschizophrenic subjects. norphin mRNA expression, one subject was found to The mRNA expression of prodynorphin, proenke- express very low prodynorphin mRNA throughout the phalin, and D1 and D2 receptors was assessed using in entire striatum, which could not be explained by any situ hybridization histochemistry. The hybridization documented clinical information. In the putamen, no signals obtained using selective oligonucleotides or statistically significant group difference was evident in riboprobes complementary to these mRNA transcripts either the patch or matrix compartment. showed unique hybridization patterns and intensities The failure to detect any significant differences in Prodynorphin mRNA expression in the striatum YL Hurd et al 496 Table 1 DPM mg−1 values (mean ± s.e.m.) in the striatum of schizophrenic, suicide, and normal control subjects

Caudate Putamen Nucleus accumbens

PDYN (total area) Normal 88.9 ± 3.0 84.6 ± 2.6 93.2 ± 6.0 Suicide 93.8 ± 2.8 90.9 ± 4.1 108.8 ± 10.7 Schizophrenic 84.6 ± 5.3 80.3 ± 5.5 101.3 ± 9.9 PDYN patch Normal 105.6 ± 2.7 97.2 ± 5.5 n.d. Suicide 120.5 ± 6.5b 103.9 ± 9.8 n.d. Schizophrenic 97.6 ± 6.1 80.9 ± 5.4 n.d. PDYN matrix Normal 54.1 ± 2.2 48.3 ± 7.2 n.d. Suicide 48.1 ± 5.4 43.9 ± 5.6 n.d. Schizophrenic 52.2 ± 4.9 39.8 ± 6.4 n.d. PDYN patch/matrix Normal 1.9 ± 0.10 2.2 ± 0.20 n.d. Suicide 2.6 ± 0.19a,b 2.4 ± 0.15 n.d. Schizophrenic 1.9 ± 0.17 2.2 ± 0.24 n.d. PENK Normal 125.8 ± 9.2 133.0 ± 8.8 129.1 ± 8.7 Suicide 133.3 ± 7.0 136.4 ± 6.0 139.7 ± 4.1 Schizophrenic 138.7 ± 6.4 141.9 ± 6.7 141.1 ± 6.5

D1 Normal 65.3 ± 2.9 65.7 ± 3.1 65.2 ± 2.9 Suicide 66.7 ± 2.2 65.9 ± 2.3 67.2 ± 2.3 Schizophrenic 65.2 ± 0.7 65.2 ± 1.1 65.6 ± 0.8

D2 dopamine receptor Normal 85.4 ± 2.2 85.6 ± 1.6 80.5 ± 2.5 Suicide 83.9 ± 1.2 85.6 ± 1.6 84.5 ± 2.6 Schizophrenic 87.1 ± 3.3 86.8 ± 4.4 82.0 ± 2.5

PDYN, prodynorphin; PENK, proenkephalin; n.d., not determined. aP Ͻ 0.05, vs control; bP Ͻ 0.05 vs schizophrenic; Tukey-Kramer.

17 the mRNA levels of dopamine D1 and D2 receptors typically inducing dysphoric experiences in humans. expressed in the caudate, putamen, or nucleus accum- While the present study cannot directly address the bens of drug-free schizophrenics is also matched at the state of depression, if any, at the time of death, the fact protein level; negative results were also obtained for remains that depression is the single most important 18 D1 and D2 receptor sites using radioligand binding on risk factor for suicide. Moreover, the prodynorphin the same postmortem brain specimen currently exam- elevations were predominantly localized in the patch 13 ined. The lack of D2 receptor alterations found at compartment of the caudate nucleus which has sig- postmortem fits with in vivo imaging studies that have nificant implications for affect, since it is the patch stri-

failed to detect any alterations of dopamine D2 receptor atal compartment that receives a greater input from density in drug-free schizophrenic patients.14 How- limbic brain regions such as the amygdala8 and allo- 6 ever, D1 and D2 receptor and proenkephalin mRNA cortices, in comparison to the matrix which is more expression were also found to be normal in schizo- closely linked to sensorimotor brain areas. phrenic subjects who had chronic neuroleptic treat- Although striatal prodynorphin preferen- 15 19 ment prior to death. It is feasible that although the D1 tially express D1 dopamine receptors, no changes and D2 receptor sites are major targets of antipsychotic were detected in mRNA expression for this dopamine drugs, other dopamine receptor subtypes, eg D3 and D4, receptor subtype or for D2 mRNA expression in the that are more tightly linked to the limbic system, are suicide group. Again, the lack of any alteration of the 16 affected to a greater extent in schizophrenic subjects. mRNA expression of D1 and D2 receptors is consistent In contrast to the failure to detect any mRNA abnor- with the negative results observed for D1 and D2 recep- malities at the level of striatal organization currently tor-binding sites in these subjects13 and in other sui- studied in schizophrenics, the increased expression of cide victims free of antidepressants prior to death.20 prodynorphin mRNA in the patch compartment of the In addition to expressing high levels of prodynor- caudate nucleus provides the first postmortem evi- phin mRNA, the human patch compartment is charac- dence of mRNA abnormalities in the striatal opioid terized by high ␮ receptor-binding sites.12 Recently, ␮ peptide system in suicide victims. Administration of opioid receptors were found to be elevated in the stria- dynorphin-like agents has a potent influence on affect, tum of suicide subjects.21 However, the latter was a bio- Prodynorphin mRNA expression in the striatum YL Hurd et al 497

Figure 1 Autoradiograms showing mRNA expression of prodynorphin in a normal control, suicide, and schizophrenic subject within the caudate nucleus (C), putamen (P), and nucleus accumbens (A). Bar scale is 5 mm.

chemical-binding study in which anatomical localiz- cide abnormalities, serotonin (5HT) activities have ation of the alterations could not be been the most often documented observations in suic- verified, so it remains to be established if the ␮ receptor ide victims and those attempting suicide, as well as in changes are in fact most associated with the patch com- patients with depression. Decreased 5HT function in partment. Although the ␮ receptor and prodynorphin suicide and depression has long been hypothesized, results provide growing evidence for significant alter- based on reduced CSF 5-hydroxyindolacetic acid in ations of the endogenous opioid system in suicide and suicidal patients,24 and reduced 5HT reuptake sites25 ␮ perhaps affective disorders, receptors are generally and increased 5-HT2 receptors in postmortem tissue of most linked with the proenkephalin system because suicide subjects.26 Consistent with the literature, an

opioid peptides derived from the proenkephalin gene increase in 5HT2 receptor sites was recently confirmed are specific ligands for the ␮ receptor sites.22 In the cur- in the prefrontal cortex of the same suicide postmortem rent study, we were unable to find any significant alter- brain specimens as those examined in this study ations in striatal proenkephalin mRNA expression (Ohuoha et al, submitted). Although the exact presyn- levels. Prodynorphin-derived peptides instead pre- aptic regulation of 5HT on opioid peptide transmission ferentially interact at ␬ opioid receptors;23 however, to has not been elucidated, it is apparent that a dysfunc- our knowledge, no evidence currently exists as to ␬ tion of 5HT could have a significant receptor changes in suicide subjects. influence on the prodynorphin system. While lesions While our results implicate opioid function in sui- of mesencephalic dopamine neurons result in a marked Prodynorphin mRNA expression in the striatum YL Hurd et al 498 upregulation of the striatal proenkephalin gene serum toxicology for neuroleptic drugs. Suicide sub- expression, these manipulations have weak or variable jects had negative toxicology for antidepressant and effects on the expression of prodynorphin mRNA.27 illicit drugs. Normal subjects had no documented his- Instead, lesions of the dorsal raphe, the major source tory of psychiatric or drug abuse disorders and toxi- of 5HT fibers, appear to alter selectively the gene cology was also negative, except for ethanol in the expression of striatal prodynorphin as compared to blood of some subjects, which was also detected in proenkephalin.28 In addition, depletion of 5HT by par- other groups (range up to 0.25%). Demographic infor- achloroamphetamine treatment elevates dynorphin- mation of the groups are listed in Table 2. like immunoreactivity in the ,29 a major The brains were held on wet ice until sectioning into target of opioid peptide striatal efferent neurons. It is 1.5 cm-thick coronal sections; rapidly frozen in a slurry thus feasible that the prodynorphin mRNA expression of dry-ice cold isopentane; and stored frozen at −70°C changes observed in suicide subjects are secondary to until the time of sectioning and in situ hybridization a dysfunction of the 5HT system. histochemistry. Fourteen ␮m-thick sections were taken In summary, the apparent abnormal expression of from blocked striatal tissue and dried onto gelatin- prodynorphin mRNA in the limbic-related striatal subbed glass slides. Right and left sides were ran- compartment in suicide subjects illustrates the appli- domly used. cation of neurochemical markers to detect subtle, but significant, neurobiological alterations associated with In situ hybridization histochemistry neuropsychiatric disorders. Although the patch/matrix In preparation for in situ hybridization histochemistry, striatal organization provides a unique possibility for tissue sections from the rostral striatum were pre- characterizing limbic- vs sensorimotor-related systems treated as previously described.12 Briefly, the sections in the striatum, it is doubtful that neurochemical were fixed with 4% paraformaldehyde in phosphate- events within these compartments will serve to dis- buffered saline, rinsed in phosphate-buffered saline, sociate neurochemical imbalances in all psychiatric incubated with 0.25% acetic anhydride in 0.1 M disorders. As evident in the present study, no prody- triethanolamine/0.9% saline, rinsed and dehydrated norphin patch/matrix alterations were observed in through a series of graded ethanols. Synthetic 48-base schizophrenic subjects; however, these subjects had oligodeoxynucleotides were used for the human proen- not been on neuroleptic medication at the time of their kephalin (bases encoding amino acid sequence 130–

death, and any previous neuroleptic treatment might 145), dopamine D1 (bases encoding amino acid have had long-term effects in regard to returning opioid sequence 706–754) and D2 (bases encoding amino acid peptide gene expression to ‘normal’ levels. Altogether, sequence 242–257) receptor cDNAs; and for prodynor- the current prodynorphin mRNA results suggest that phin, RNA probes were used which were complemen- the opioid system may contribute to the pathophysiol- tary to exon 4 of the prodynorphin gene. The oligonu- ogy of suicide and perhaps affective disorders. Further cleotide9,12 and RNA probes9,10 have been previously studies testing this hypothesis will be necessary to rep- characterized showing specific hybridization patterns licate this finding and/or understand its relevance to when carried out under the present in situ hybridiz- suicide and affective disorders. ation conditions. In addition, different probes which hybridize to mRNA transcripts with distinct regional patterns of expression were simultaneously studied in Materials and methods the in situ hybridization experiments to further vali- Postmortem tissue date the specificity of the hybridization signal. In brief, Brain samples used in this study were obtained at auto- the oligodeoxynucleotide probes were end-labeled psy from the Medical Examiner’s Office of the District with 35S-dATP with terminal transferase and tailing of Columbia (Washington, DC, USA) by the Section of buffer (25 units ␮l−1; Boehringer-Mannheim, Indiana- Neuropathology, Clinical Brain Disorders Branch, polis, IN, USA). RNA probes were generated from a National Institute of Mental Health. All brains were human cDNA fragment of the prodynorphin gene examined by a neuropathologist (MMH) and any sub- inserted into SP65 vector and labeled with 35S-uridine ject with pathological abnormalities as determined 5′-[␣-thio]triphosphate (1500 Ci mmol−1; New England after gross and microscopic histological examination, Nuclear, Boston, MA, USA).9 In situ hybridization was including Bielschowsky’s silver stain, were excluded. performed overnight at 37°C (oligonucleotides) or 55°C Blood and urine specimens were collected at autopsy (riboprobes) with 8 × 106 cpm-labeled probe. The sec- for toxicological examination for psychotropic agents tions were subsequently carried through high strin- such as psychostimulants, neuroleptics, and antide- gency washes in a buffer containing standard saline cit- pressant drugs on all subjects. Brain specimens were rate optimized for hybridizations carried out with chosen from three groups: schizophrenia, nonschizo- oligonucleotide or RNA probes,9 rinsed, and dried. phrenic suicide, or normal control, matched for age Tissue sections were apposed to hyperfilm for 10–28 and gender. Psychiatric diagnosis was determined by days along with 14C standards (previously correlated at least two independent psychiatrists according to against 35S standards) and developed (Kodak). Light DSM III-R criteria. Schizophrenic subjects had been off transmittance values from different striatal regions neuroleptic medication prior to death (exact length of were taken from the film images by use of a light box, time unknown), which was verified by a negative MTI CCD72 high-resolution video camera with a Canon Prodynorphin mRNA expression in the striatum YL Hurd et al 499 Table 2 Demographics of subjects examined in this study

Diagnosis Number Mean age Sex Race Mean PMI (years ± s.e.m.) W:B:A (h ± s.e.m.)

Normal 6 51 ± 12 2 F/4 M 1 : 4 : 1 30.7 ± 5.6 Suicide 6 47 ± 12 1 F/5 M 4 : 1 : 1 27.4 ± 6.0 Schizophrenia 6 57 ± 16 4 F/2 M 3 : 3 : 0 21.8 ± 3.7

F, female; M, male; W, white; B, black; A, Asian.

Macro FD/50 mm lens, and Macintosh-based image tides in human brain—postmortem studies. Prog Neuro-Psycho- analysis software system (IMAGE, Wayne Rasband, pharmacol Biol Psych 1985; 9: 91–95. 5 Toru M, Watanabe S, Shibuya H, Nishikawa T, Noda K, Mitsushio NIMH). Gain and black level settings were stan- H et al. Neurotransmitters, receptors and neuropeptides in post- dardized with the MTI CCD72 control unit for the cam- mortem brains of chronic schizophrenic patients. Acta Psychiatr era. The gray levels measured in the caudate, putamen, Scand 1988; 78: 121–137. and nucleus accumbens regions were converted to 6 Gerfen CR. The neostriatal mosaic: multiple levels of compart- DPM mg−1 values using the standard curve generated mental organization in the . Ann Rev Neurosci 1992; 14 15: 285–320. by the C standard and averaged. Clusters of high pro- 7 Jaskiw G, Weinberger D. Dopamine and schizophrenia—a cortically dynorphin mRNA expression in the caudate and corrective perspective. Sem Neurosci 1992; 4: 179–188. putamen have been demonstrated to be in register with 8 Graybiel AM. Neurotransmitters and neuromodulators in the basal the distribution of classic patch compartment mark- ganglia. Trends Neurosci 1990; 13: 244–254. ers.9 Thus, hybridization labeling within and outside 9 Hurd YL, Herkenham M. Neuropeptide compartmentalization within the human neostriatum: in situ hybridization and in vitro 4–10 randomly distributed patches of high prodynor- autoradiography characterization. Neuroscience 1995; 64: 571–586. phin mRNA expression were averaged and used for 10 Hurd YL. Differential messenger RNA expression of prodynorphin determining DPM mg−1 values within the patch and and proenkephalin in the human brain. Neuroscience 1996; 72: matrix compartments, respectively. 767–783. 11 Mengod G, Vilaro MT, Niznik HB, Sunahara RK, Seeman P, Statistical analyses were performed using the JMP O’Dowd BF et al. Visualization of a dopamine D1 receptor mRNA statistical program (version 3.1; SAS Institute Inc Stat- in human and rat brain. Molec Brain Res 1991; 10: 185–191. istical software). Significant group differences within 12 Hurd YL, Herkenham M. Molecular alterations in the neostriatum specific striatal regions were assessed using ANOVA of human addicts. Synapse 1993; 13: 357–369. with Tukey-Kramer post-hoc analysis for each mRNA 13 Knable MB, Hyde TM, Herman MM, Carter JM, Bigelow L, Klein- man JE. Quantitative autoradiography of dopamine-D1 receptors, species. Correlation analyses were carried out using the D2 receptors, and dopamine uptake sites in postmortem striatal Spearman’s coefficient of correlation with P-value set specimens from schizophrenic patients. Biol Psychiatr 1994; 36: after applying the Bonferroni procedure to correct for 827–835. multiple correlations. Results were regarded as sig- 14 Farde L, Wiesel F-A, Stone-Elander S, Halldin C, Nordstro¨ m AL, nificant for P Յ 0.05 and considered as a trend for Hall H et al. D2 dopamine receptors in neuroleptic-naive schizo- Յ phrenic patients: a positron emission tomography study with P 0.10. [11C]raclopride. Arch Gen Psychiat 1990; 47: 213–219. 15 Harrington KA, Augood SJ, Faull RLM, McKenna PJ, Emson PC. Dopamine D1 receptor, D2 receptor, proenkephalin A and subst- Acknowledgements ance P gene expression in the caudate nucleus of control and We thank Dr James Douglas (Amgen Inc, Thousand schizophrenic tissue: a quantitative cellular in situ hybridization study. Molec Brain Res 1995; 33: 333–342. Oaks, CA) for the human prodynorphin cDNA clone, 16 Seeman P, Guan H-C, Van Tol HHM. Dopamine D4 receptors elev- Jennifer Weber and Christopher Spurney for their ated in schizophrenia. Nature 1993; 365: 441–445. excellent technical assistance, and Jonas Nilson for his 17 Pfeiffer A, Brandt V, Herz A. Psychotomimesis mediated by kappa help with film analysis. YLH was supported in part by receptors. Science 1986; 233: 774–776. the Swedish Medical Research Council (11252). 18 Fawcett J, Scheftner W, Clark D, Hedeker D, Bibbons R, Coryell W. Clinical predictors of suicide in patients with major affective disorders: a controlled prospective study. Am J Psychiat 1987; 144: 35–40. References 19 Le Moine C, Normand E, Bloch B. Phenotypical characterization of 1 Lindstrom LH, Widerlov E, Gunne LM, Wahlstrom A, Terenius L. the rat striatal neurons expressing the D1 dopamine receptor gene. in human cerebrospinal fluid: clinical correlations to Proc Natl Acad Sci USA 1991; 88: 4205–4209. some psychotic states. Acta Psychiatr Scand 1978; 57: 153–164. 20 Bowden C, Theodorou AE, Cheetham SC, Lowther S, Katona CLE, 2 Naber D, Pickar D, Post RM, Van Kammen DP, Waters RN, Bal- Crompton MR et al. Dopamine D-1 and D-2 receptor binding sites lenger JC et al. Endogenous opioid activity and beta-endorphin in brain samples from depressed suicides and controls. Brain Res immunoreactivity in CSF of psychiatric patients and normal volun- 1997; 752: 227–233. teers. Am J Psychiatry 1981; 138: 1457–1462. 21 Gabilondo AM, Meana JJ, Garcia-Sevilla JA. Increased density of 3 Lindstrom LH, Besev G, Gunne LM, Terenius L. CSF levels of ␮-opioid receptors in the postmortem brain of suicide victims. receptor-active endorphins in schizophrenic patients: correlations Brain Res 1995; 682: 245–250. with symptomatology and monoamine metabolites. Psychiatr Res 22 Lord JAH, Waterfield AA, Hughes J, Kosterlitz HW. Endogenous 1986; 19: 93–100. opioid peptides: multiple agonists and receptors. Nature 1977; 267: 4 Kleinman JE, Hong J, Iadarola M, Govoni S, Gillin CJ. Neuropep- 495–499. Prodynorphin mRNA expression in the striatum YL Hurd et al 500 23 Chavkin C, James IF, Goldstein A. Dynorphin is a specific endogen- neurons regulate the expression of neuropeptide mRNAs in the rat ous ligand for the ␬-opioid receptor. Science 1982; 215: 413–415. forebrain. Proc Natl Acad Sci USA 1986; 83: 9827–9831. 24 Åsberg M, Tra¨skman L, Thore´n P. 5-HIAA in the cerebrospinal 28 Morris BJ, Reimer S, Ho¨ llt V, Herz A. Regulation of striatal prody- fluid: a biochemical suicide predictor? Arch Gen Psychiat 1976; 33: norphin mRNA levels by the raphe-striatal pathway. Molec Brain 1193–1197. Res 1988; 4: 15–22. 25 Stanley M, Virgilio J, Gershon S. Tritiated binding sites 29 Smiley PL, Johnson M, Bush L, Gibb JW, Hanson GR. Effects of are decreased in the frontal cortex of suicides. Science 1982; 216: cocaine on extrapyramidal and limbic dynorphin systems. J Phar- 1337–1339. mac Exper Ther 1990; 253: 938–943. 26 Arango V, Ernsberger P, Marzuk PM, Shen J-S, Tierney H, Stanley M et al. Autoradiographic demonstration of increased serotonin 5- Correspondence: Dr Yasmin Hurd, Karolinska Institute, Dept of Clini-

HT2 and ␤-adrenergic receptor binding sites in the brain of suicide cal Neuroscience, Psychiatry Section, Karolinska Hospital, S-171 76 victims. Arch Gen Psychiat 1990; 47: 1038–1047. Stockholm, Sweden. E-mail: yasmin.hurdȰneuro.ks.se 27 Young III WS, Bonner TI, Brann MR. Mesencephalic dopamine Received 28 April 1997; revised and accepted 30 June 1997