- Email: [email protected]
Expression of serotonin transporter mRNA in rat brain: Presence in neuronal and non-neuronal cells and effect of paroxetine
Journal of Chemical Neuroanatomy 13 (1997) 71 – 76
Expression of serotonin transporter mRNA in rat brain: Presence in neuronal and non-neuronal cells and effect of paroxetine Marc C. Swan a, A. Rahman Najlerahim b,1, Jonathan P. Bennett a,* a
Department of Anatomy and Cell Biology, Imperial College School of Medicine at St Mary’s, Norfolk Place, London, W2 1PG, UK b Institute of Gerontology, King’s College London, Waterloo Road, London, SE1 8TX, UK Received 22 November 1996; received in revised form 14 April 1997; accepted 14 April 1997
Abstract The expression of serotonin transporter mRNA in rat brain was examined by in situ hybridisation. Hybridisation was observed in cells of the known serotonergic nuclei and no other neuronal populations. It was also associated with ependymal cells of the aqueduct which may indicate a specialisation of this part of the ventricular system in anatomical and neurophysiological terms. The effect of the selective serotonin reuptake inhibitor paroxetine on neuronal expression of the serotonin tranporter mRNA was examined. Quantitation at the cellular level in the dorsal and median raphe nuclei was carried out by analysis of the mean number of silver grains per cell in autoradiographed sections. No significant change (P\ 0.1) in serotonin transporter mRNA expression was observed following 21 day administration of paroxetine (5 mg/kg per day). © 1997 Elsevier Science B.V. Keywords: Antidepressive agents; 5-Hydroxytryptamine; Neurotransmitter uptake; In situ hybridization; Raphe nuclei; Ependyma
1. Introduction Within the central nervous system (CNS), the cell bodies of serotonergic neurones are almost exclusively associated with the midsagittal raphe of the brain stem (Steinbusch, 1981; Tork, 1985) and have extensive anatomical projections appropriate to the diverse physiological and psychological functions that serotonin (5hydroxytryptamine, 5-HT) mediates. One of the features of serotonergic transmission is a specific 5-HT reuptake system and the Na + -dependent substrate-specific transporter molecule responsible has been identified and its gene sequenced by several groups (Blakely et al., 1991; Hoffman et al., 1991; Lesch et al., 1993a). The mRNA for the 5-HT transporter has been shown to be present in neurons of the raphe nuclei as expected if uptake is presynaptic (Uhl, 1992), though the ques* Corresponding author. Tel.: +44 171 5943764; fax: + 44 171 4026861; e-mail: [email protected] 1 Present address: Department of Neurology, Shohada Medical Centre, Shahid Behshti University, Tehran, Iran. 0891-0618/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 8 9 1 - 0 6 1 8 ( 9 7 ) 0 0 0 2 6 - 4
tion remains as to whether it may be present in additional locations (Lesch et al., 1993b). Selective serotonin reuptake inhibitors (SSRIs) such as paroxetine are a type of antidepressant which selectively inhibits the 5-HT transporter, thereby increasing synaptic cleft 5-HT concentration and prolonging postsynaptic receptor activation (Wolf and Kuhn, 1990). Because it has been reported that tricyclic antidepressants such as imipramine and amitriptyline cause an increase in 5-HT transporter mRNA levels (Lopez et al., 1994), we were interested to see whether the SSRI paroxetine has a similar effect in terms of mRNA expression. It is not difficult to imagine how a feedback mechanism might exist whereby the rate of serotonin reuptake could control 5-HT transporter synthesis and if such a mechanism were to exist it might have implications for the optimal dose regimen in clinical practice. This study used the technique of in situ hybridisation and quantitative image analysis to determine levels of expression of 5-HT transporter mRNA in serotonergic cells of the raphe nuclei.
72
M.C. Swan et al. / Journal of Chemical Neuroanatomy 13 (1997) 71–76
2. Materials and methods
2.1. Animals and drug treatment All animal procedures were conducted under UK Home Office licence. Adult female Sprague-Dawley rats of approximately 200 g weight were treated with paroxetine (5 mg/kg per day in saline solution by intraperitoneal injection) for 21 days. Paroxetine was obtained from Dr I. Tullock (SmithKline Beecham Pharmaceuticals). The rats were sacrificed 24 h after last treatment by lethal intraperitoneal injection of phenobarbitone and the brain immediately removed and frozen. Rats receiving saline were used as a control group. Cryostat sections of 10 mm were taken and processed as previously described (Najlerahim et al., 1990).
2.2. Probe selection and labelling A 30 base oligonucleotide probe (synthesised by Oswel DNA Service, Southampton) sequence 5%-CTG GGG TGG GCT CAT CAG AAA ACT GCA AAT-3% was used. This is complementary to bases 1635 – 1664 of the 5-HT transporter gene sequence (Blakely et al., 1991) and was not homologous to any other gene in the EMBL sequence database (for up to six mismatches). The probe was labelled with [35S]-deoxyadenosine 5%-[athio] triphosphate using 3% terminal deoxynucleotidyl transferase as previously described (Najlerahim et al., 1990).
2.3. Hybridisation The in situ hybridisation procedure used has been described in detail previously (Najlerahim et al., 1990; Swan et al., 1994). Section hybridisation was carried out using 16.7 kBq probe per section (specific radioactivity 0.14 MBq/pmol) in a buffer containing 0.78 M sodium ions and 50% formamide for 18 h at 27°C (15°C below the calculated melt temperature for the probe). Subsequent washes in 0.15 M sodium chloride and 15 mM sodium citrate included 1 h at 52°C (15°C below the recalculated melt temperature). Sections were autoradiographed either by direct apposition to film for 14 days, or by dipping in photographic emulsion and exposing for 6 weeks. In control experiments, hybridisation was abolished by preincubating sections with ribonuclease or by cold displacement using a 50-fold excess of unlabelled probe; in addition no specific hybridisation was observed with a 30 base probe of random degenerate sequence.
2.4. Quantification The number of silver grains overlying cell bodies of raphe neurons was determined by an image analyser
(Seescan, Cambridge) with a three stage thresholding routine. In a calibration experiment the amount of [35S]-labelled probe hybridised to cells on serial sections was varied by adjusting the specific radioactivity in the hybridisation solution (Gerfen, 1989) and there was found to be a linear relationship with grain number measured with the image analyser (r=0.988, PB0.02). The number of silver grains overlying ependymal cells was counted manually from enlarged micrographs.
3. Results Hybridisation for 5-HT transporter mRNA was carried out throughout the rat brain; a representative selection of coronal sections is shown in Fig. 1. It was highly specific relating almost entirely to the midsagittal raphe nuclei. The dorsal raphe nucleus is the largest labelled region, with individual composite cell clusters visible (eg. Fig. 1C). Labelling was also found in the median raphe, raphe pontis, raphe magnus, raphe obscurus and raphe pallidus nuclei. The only labelled extra-raphe neurons were the bilateral B9 groups located in the ventrolateral tegmentum of the midbrain, close to the medial lemniscus (Fig. 1B). No labelling was found in any other brain region including all areas of the cortex, hippocampus, thalamus, hypothalamus, cerebellum, brainstem and spinal cord (e.g. Fig. 1A,F). The emulsion dipped sections (Fig. 2) demonstrated specific labelling of serotonergic cell bodies (approximately half the neurons in the dorsal raphe) with negligible signal over others. Specific labelling was also exhibited around the caudal part of the aqueduct (Fig. 1C,D,E) where it enlarges to form the recess of the superior colliculus, sometimes termed the ‘mesencephalic ventricle’ (Mitro and Palkovits, 1981). Labelling was rapidly discontinued as the aqueduct opened into the fourth ventricle and neither was any associated with the lateral or third ventricles. Examination of emulsion dipped sections showed hybridisation to the ependymal layer (Fig. 3). However there did not appear to be homogenous labelling over all cells in the layer and the distribution of hybridisation was further studied. Quantitation of cellular labelling indicated that the number of silver grains overlying individual cells of the ependymal layer varied from zero to over 20 with a mean of 4.2 (Fig. 4A). This differed from background labelling which was determined by placing tracings of the ependymal cell outlines over unlabelled regions of the same micrographs. The background showed a tighter distribution (Fig. 4B) averaging 1.6 silver grains per cell and it resembled a Poisson distribution in shape as would be predicted for a random distribution of grains. If the assumption is made that the grains from hybridisation to ependymal cells should also show a Poisson distribu-
M.C. Swan et al. / Journal of Chemical Neuroanatomy 13 (1997) 71–76
73
Fig. 1. Distribution of 5-HT transporter mRNA in selected coronal sections of rat brain. (A – F) are photographs printed directly from autoradiographs with schematic line drawings below showing major visible anatomical features. Scale bar: 5.0 mm. Abbreviations: LV, lateral ventricle; 3V, third ventricle; Aq, aqueduct; 4V, fourth ventricle; DR, dorsal raphe nucleus; MnR, median raphe nucleus; RPn, raphe pontis nucleus; RMg, raphe magnus nucleus; RPa, raphe pallidus nucleus; ROb, raphe obscurus nucleus; B9, group of serotonergic cells within the medial lemniscus.
tion then modelling the observed data requires a minimum of three cell populations. For example, Fig. 4C models a mixed population of cells of which 45% have background labelling at an average 1.6 grains per cell, another 40% have specific labelling averaging five grains per cell and the final 15% average ten grains per cell; this predicted distribution is not significantly different from the experimental data (P \ 0.2, Kolmogorov-Smirnov test). The levels of neuronal hybridisation were studied in rats which had received paroxetine (5 mg/kg bodyweight per day for 21 days). Quantitation was based on the mean number of silver grains overlying labelled cells in the dorsal and median raphe nuclei, determined by image analysis. For each experimental animal 20 randomly chosen neuronal cell bodies which included a
nuclear profile were analysed for both the dorsal and median raphes and the mean number of grains determined. Following 21 days treatment with paroxetine, hybridisation was found not to differ significantly from the control group for either the dorsal raphe (P =0.15, Student’s 2-tailed t-test) or the median raphe (P =0.89) nuclei (Table 1).
4. Discussion These results confirm the anatomical distribution of 5-HT transporter mRNA to neuronal cell bodies within the raphe nuclei and the B9 region of the medial lemniscus (Fujita et al., 1993; Austin et al., 1994). This corresponds to the distribution of serotonergic cell bod-
74
M.C. Swan et al. / Journal of Chemical Neuroanatomy 13 (1997) 71–76
ies previously established using other techniques (Steinbusch, 1981; Tork, 1985). No further neuronal labelling was seen beyond the regions described. We conclude that neuronal expression of the transporter is entirely presynaptic, which is not necessarily the case for all neurotransmitter transporters (Swan et al., 1994). A finding of particular interest is the existence of a distinct loop of 5-HT transporter mRNA labelling surrounding the caudal part of the aqueduct, the circumference of which steadily increases as the aqueduct enlarges, until the fourth ventricle is reached upon which it disappears. The hybridisation appeared specific in that it was not observed in other parts of the ventricular system, it was displaced by an excess of unlabelled probe, and it was not observed when using a control probe (or indeed, probes to other mRNAs). It was observed in all 11 animals investigated. In principle we cannot rule out the existence in this part of the aqueduct of a different unknown mRNA containing a
Fig. 3. Detection of 5-HT transporter mRNA in the aqueductal ependyma. Dark-field (A,B) and the corresponding bright-field (C,D) photomicrographs of the aqueduct wall showing labelling of the ependyma; scale bar, 100 mm. Panels A and C illustrate hybridisation with the probe for 5-HT transporter mRNA; panels B and D illustrate a control experiment in which hybridisation was carried out in the presence of a 50-fold excess of unlabelled probe. Panel E shows a higher magnification photomicrograph of the labelled ependymal layer; scale bar, 20 mm.
Fig. 2. Cellular distribution of 5-HT transporter mRNA. Bright-field photomicrograph of (A) the dorsal raphe nucleus, (B) the median raphe nucleus, and (C) the B9 group of cells. Note the presence of both labelled and unlabelled cells. Scale bar, 20 mm.
sequence complementary to the probe, but the exquisite specificity of labelling elsewhere makes it most likely that we are detecting 5-HT transporter mRNA; even if it were a different mRNA it would still imply a functional specialisation of this part of the ventricular system.
M.C. Swan et al. / Journal of Chemical Neuroanatomy 13 (1997) 71–76
Histological investigation localises this labelling to the ependymal layer and quantitative analysis showed that the labelling pattern was consistent with just over one half of the ependymal cells undergoing hybridisation with the remainder negative which may reflect a functional heterogeneity among ependymal cells; this could be due to known ependymal cell types such as the ciliated epithelial cells and the tanycytes, or further
Fig. 4. Analysis of distribution of autoradiographic silver grains overlying ependymal cells hybridised for 5-HT transporter mRNA. (A) Distribution for 1499 cells (from five animals) analysed. (B) Distribution of background labelling, determined using outline tracings of the ependymal cells analysed in panel A in regions negative for hybridisation. (C) Model for the distribution of silver grains found in panel A, assuming a mixture of three cell populations each having a Poisson distribution of grain number with averages of 1.6 grains per cell (45% of cells), 5 grains per cell (40%) and 10 grains per cell (15%).
75
Table 1 Quantitation of 5-HT transporter mRNA in paroxetine-treated and control rats
Dorsal raphe Median raphe
Control (n =4)
Paroxetine-treated (n = 6)
85 95 68 98
739 15 67 910
Expression is determined as the number of silver grains (mean 9 S.D.) overlying neuronal cell bodies following in situ hybridisation.
degrees of specialisation. However the simplest explanation of the apparent heterogeneity is probably that of these polarised epithelial cells a proportion will not have their rough endoplasmic reticulum included in any particular histological section, and thus would not be labelled for membrane protein mRNAs associated with the endoplasmic reticulum including the 5-HT transporter. The fact that two cell populations with differing levels of hybridisation were required to model the experimental data (Fig. 4C) may not be meaningful since the assumption that distributions are of the Poisson type is not necessarily appropriate for a mixed cell population where the grain scatter due to the decay energy of sulphur-35 is of the order of a cell diameter. In theory there is the alternative explanation that the mRNA detected is not in the ependymal cells but the intercalated nerves. An extensive serotonergic plexus originating in the raphe nuclei has been shown to exist in the supra- and subependymal structures (Aghajanian and Gallager, 1975; Chan-Palay, 1976), which may act to modulate cerebrospinal fluid (CSF) composition in terms of 5-HT concentration. If the 5-HT transporter mRNA were present in these nerve terminals the small size of the nerves adjacent to close packed ependymal cells, coupled with the dispersion of silver grains resulting from the decay energy of sulphur-35, would mean that on hybridisation the label would appear to overlie a subpopulation of ependymal cells. However only in a few special cases has mRNA been found in the axonal domain of neurons (Steward and Banker, 1992), and the cellular machinery for synthesis of membrane proteins is absent. Instead we prefer to propose that the 5-HT transporter is synthesised by cells of the aqueductal ependyma. It indicates a specialisation of this region in terms of cellular function. Published images obtained by inhibitor binding have previously appeared to indicate the presence of the 5-HT transporter protein in the vicinity of the aqueduct but not other regions of the ventricular system (Hrdina et al., 1990) though in that study the identity of the cells where it was situated was not investigated. The presence of serotonin within CSF is well established and it has been hypothesised to control CSF volume and composition via the serotonin receptors on the apical surface of the choroid plexus
76
M.C. Swan et al. / Journal of Chemical Neuroanatomy 13 (1997) 71–76
epithelium (Nilsson et al., 1992); it may also have an effect on neural function judging by the correlation between the concentration of serotonin and its metabolites in the CSF and certain psychiatric conditions (Lester, 1995). An uptake mechanism specifically removing serotonin from the CSF would ensure sensitivity of the system to changes in the rate of secretion of serotonin into CSF via supraependymal nerve terminals. It is also conceivable that the location of the proposed serotonin uptake mechanism at the aqueduct would allow differential control of CSF composition between the forebrain and the hindbrain, spinal cord and subarachnoid space. No further novel sites of 5-HT transporter mRNA localisation were revealed in our study. Lesch et al. (1993b) used a reverse transcriptase-polymerase chain reaction approach to investigate areas of rat brain associated with ascending 5-HT pathways. They reported that 5-HT transporter mRNA may be present in low concentrations in the frontal cortex, hippocampus and neostriatum and attributed this to specific transport of mRNA to the terminal regions of the 5-HT pathways in order to institute local transporter synthesis. However we were unable to confirm their distribution using in situ hybridisation (see Fig. 1A,B). If such a transport mechanism exists the amount of mRNA involved must be so small that it can only be detected by sensitive amplification methods, which implies that its contribution to total 5-HT transporter synthesis will be marginal. It could easily be envisaged that expression of the 5-HT transporter might be under feedback control which would result in it being altered by pharmacological manipulation of extracellular 5-HT levels by SSRIs, similar to the change in expression reported following exposure to tricyclic antidepressants (Lesch et al., 1993b; Lopez et al., 1994). We quantified 5-HT transporter mRNA at the cellular level and though we found a marginal decrease in both the dorsal and median raphe nuclei after 21 days paroxetine treatment, the change was not statistically significant. Lesch et al. (1993b) estimated from Northern blot analysis that 21-day administration of another SSRI, fluoxetine (2.5 mg/kg per day via an osmotic pump), caused a 30% decrease in 5-HT transporter mRNA expression, but no statistical analysis was presented. We are unable to say whether our inability to confirm their finding is due to the use of a different SSRI, or the difference in methodology for determining mRNA expression levels.
Acknowledgements This work was supported in part by a grant from the Medical Research Council.
References Aghajanian, G.K., Gallager, D.W., 1975. Raphe origin of serotonergic nerves terminating in the cerebral ventricles. Brain Res. 88, 221–231. Austin, M.C., Bradley, C.C., Mann, J.J., Blakely, R.D., 1994. Expression of serotonin transporter messenger RNA in the human brain. J. Neurochem. 62, 2362 – 2367. Blakely, R.D., Berson, H.E., Fremau, R.T., Caron, M.G., Peek, M.M., Prince, H.K., Bradley, C.C., 1991. Cloning and expression of a functional serotonin transporter from rat brain. Nature 354, 66–70. Chan-Palay, V., 1976. Serotonin axons in the supra- and subependymal plexuses and in the leptomeninges; their roles in local alterations of cerebrospinal fluid and vasomotor activity. Brain Res. 102, 103 – 130. Fujita, M., Shimada, S., Maeno, H., Nishimura, T., Tohyama, M., 1993. Cellular localization of serotonin tranporter mRNA in the rat brain. Neurosci. Lett. 162, 59 – 62. Gerfen, C.R., 1989. Quantification of in situ hybridization histochemistry for analysis of brain function. Meth. Neurosci. 1, 79–97. Hoffman, B.J., Mezey, E., Brownstein, M.J., 1991. Cloning of a serotonin transporter affected by antidepressants. Science 254, 579 – 580. Hrdina, P.D., Foy, B., Hepner, A., Summers, R.J., 1990. Antidepressant binding sites in brain: Autoradiographic comparison of [3H]paroxetine and [3H]imipramine localization and relationship to serotonin transporter. J. Pharmacol. Exp. Ther. 252, 410–418. Lesch, K.P., Wolozin, B.L., Estler, H.C., Murphy, D.L., Riederer, P., 1993a. Isolation of a cDNA encoding the human brain serotonin transporter. J. Neural Transm. Gen. Sect. 91, 67 – 72. Lesch, K.P., Aulakh, C.S., Wolozin, B.L., Tolliver, T.J., Hill, J.L., Murphy, D.L., 1993b. Regional brain expression of serotonin transporter mRNA and its regulation by reuptake inhibiting antidepressants. Brain Res. Mol. Brain Res. 17, 31 – 35. Lester, D., 1995. The concentration of neurotransmitter metabolites in the cerebrospinal fluid of suicidal individuals: A meta-analysis. Pharmacopsychiatry 28, 45 – 50. Lopez, J.F., Chalmers, D.T., Vazquez, D.M., Watson, S.J., Akil, H., 1994. Serotonin transporter mRNA in rat brain is regulated by classical antidepressants. Biol. Psychiatry 35, 287 – 290. Mitro, A., Palkovits, M., 1981. Morphology of the Rat Brain Ventricles, Ependyma and Periventricular Structures. Karger, Basel. Najlerahim, A.R., Harrison, P.J., Barton, A.J.L., Heffernan, J., Pearson, R.C.A., 1990. Distribution of messenger mRNAs encoding the enzymes glutaminase, aspartate aminotransferase and glutamic acid decarboxylase in rat brain. Brain. Res. Mol. Brain Res. 7, 317–333. Nilsson, C., Lindvall-Axelsson, M., Owman, C., 1992. Neuroendocrine regulatory mechanisms in the choroid plexus-cerebrospinal fluid system. Brain Res. Rev. 17, 109 – 138. Steinbusch, H.W.M., 1981. Distribution of serotonin-immunoreactivity in the CNS of rat brain-cell bodies and terminals. Neuroscience 6, 557 – 618. Steward, O., Banker, G.A., 1992. Getting the message from the gene to the synapse: Sorting and intracellular transport of RNA in neurones. Trends Neurosci. 15, 180 – 186. Swan, M.C., Najlerahim, A.R., Watson, R.B., Bennett, J.P., 1994. Distribution of mRNA for the GABA transporter GAT-1 in the rat brain: Evidence that GABA uptake is not limited to presynaptic neurons. J. Anat. 185, 315 – 323. Tork, I., 1985. Raphe nuclei and serotonin containing systems. In: Paxinos, G., (Ed.), The Rat Nervous System, vol. 2. Academic Press, Sydney, pp. 43 – 69. Uhl, G.R., 1992. Neurotransmitter transporters (plus): A promising new gene family. Trends Neurosci. 15, 265 – 268. Wolf, W.A., Kuhn, D.A., 1990. Modulation of serotonin release: Interactions between the serotonin transporter and autoreceptors. Ann. New York Acad. Sci. 604, 505 – 513.
Expression of serotonin transporter mRNA in rat brain: Presence in neuronal and non-neuronal cells and effect of paroxetine Marc C. Swan a, A. Rahman Najlerahim b,1, Jonathan P. Bennett a,* a
Department of Anatomy and Cell Biology, Imperial College School of Medicine at St Mary’s, Norfolk Place, London, W2 1PG, UK b Institute of Gerontology, King’s College London, Waterloo Road, London, SE1 8TX, UK Received 22 November 1996; received in revised form 14 April 1997; accepted 14 April 1997
Abstract The expression of serotonin transporter mRNA in rat brain was examined by in situ hybridisation. Hybridisation was observed in cells of the known serotonergic nuclei and no other neuronal populations. It was also associated with ependymal cells of the aqueduct which may indicate a specialisation of this part of the ventricular system in anatomical and neurophysiological terms. The effect of the selective serotonin reuptake inhibitor paroxetine on neuronal expression of the serotonin tranporter mRNA was examined. Quantitation at the cellular level in the dorsal and median raphe nuclei was carried out by analysis of the mean number of silver grains per cell in autoradiographed sections. No significant change (P\ 0.1) in serotonin transporter mRNA expression was observed following 21 day administration of paroxetine (5 mg/kg per day). © 1997 Elsevier Science B.V. Keywords: Antidepressive agents; 5-Hydroxytryptamine; Neurotransmitter uptake; In situ hybridization; Raphe nuclei; Ependyma
1. Introduction Within the central nervous system (CNS), the cell bodies of serotonergic neurones are almost exclusively associated with the midsagittal raphe of the brain stem (Steinbusch, 1981; Tork, 1985) and have extensive anatomical projections appropriate to the diverse physiological and psychological functions that serotonin (5hydroxytryptamine, 5-HT) mediates. One of the features of serotonergic transmission is a specific 5-HT reuptake system and the Na + -dependent substrate-specific transporter molecule responsible has been identified and its gene sequenced by several groups (Blakely et al., 1991; Hoffman et al., 1991; Lesch et al., 1993a). The mRNA for the 5-HT transporter has been shown to be present in neurons of the raphe nuclei as expected if uptake is presynaptic (Uhl, 1992), though the ques* Corresponding author. Tel.: +44 171 5943764; fax: + 44 171 4026861; e-mail: [email protected] 1 Present address: Department of Neurology, Shohada Medical Centre, Shahid Behshti University, Tehran, Iran. 0891-0618/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 8 9 1 - 0 6 1 8 ( 9 7 ) 0 0 0 2 6 - 4
tion remains as to whether it may be present in additional locations (Lesch et al., 1993b). Selective serotonin reuptake inhibitors (SSRIs) such as paroxetine are a type of antidepressant which selectively inhibits the 5-HT transporter, thereby increasing synaptic cleft 5-HT concentration and prolonging postsynaptic receptor activation (Wolf and Kuhn, 1990). Because it has been reported that tricyclic antidepressants such as imipramine and amitriptyline cause an increase in 5-HT transporter mRNA levels (Lopez et al., 1994), we were interested to see whether the SSRI paroxetine has a similar effect in terms of mRNA expression. It is not difficult to imagine how a feedback mechanism might exist whereby the rate of serotonin reuptake could control 5-HT transporter synthesis and if such a mechanism were to exist it might have implications for the optimal dose regimen in clinical practice. This study used the technique of in situ hybridisation and quantitative image analysis to determine levels of expression of 5-HT transporter mRNA in serotonergic cells of the raphe nuclei.
72
M.C. Swan et al. / Journal of Chemical Neuroanatomy 13 (1997) 71–76
2. Materials and methods
2.1. Animals and drug treatment All animal procedures were conducted under UK Home Office licence. Adult female Sprague-Dawley rats of approximately 200 g weight were treated with paroxetine (5 mg/kg per day in saline solution by intraperitoneal injection) for 21 days. Paroxetine was obtained from Dr I. Tullock (SmithKline Beecham Pharmaceuticals). The rats were sacrificed 24 h after last treatment by lethal intraperitoneal injection of phenobarbitone and the brain immediately removed and frozen. Rats receiving saline were used as a control group. Cryostat sections of 10 mm were taken and processed as previously described (Najlerahim et al., 1990).
2.2. Probe selection and labelling A 30 base oligonucleotide probe (synthesised by Oswel DNA Service, Southampton) sequence 5%-CTG GGG TGG GCT CAT CAG AAA ACT GCA AAT-3% was used. This is complementary to bases 1635 – 1664 of the 5-HT transporter gene sequence (Blakely et al., 1991) and was not homologous to any other gene in the EMBL sequence database (for up to six mismatches). The probe was labelled with [35S]-deoxyadenosine 5%-[athio] triphosphate using 3% terminal deoxynucleotidyl transferase as previously described (Najlerahim et al., 1990).
2.3. Hybridisation The in situ hybridisation procedure used has been described in detail previously (Najlerahim et al., 1990; Swan et al., 1994). Section hybridisation was carried out using 16.7 kBq probe per section (specific radioactivity 0.14 MBq/pmol) in a buffer containing 0.78 M sodium ions and 50% formamide for 18 h at 27°C (15°C below the calculated melt temperature for the probe). Subsequent washes in 0.15 M sodium chloride and 15 mM sodium citrate included 1 h at 52°C (15°C below the recalculated melt temperature). Sections were autoradiographed either by direct apposition to film for 14 days, or by dipping in photographic emulsion and exposing for 6 weeks. In control experiments, hybridisation was abolished by preincubating sections with ribonuclease or by cold displacement using a 50-fold excess of unlabelled probe; in addition no specific hybridisation was observed with a 30 base probe of random degenerate sequence.
2.4. Quantification The number of silver grains overlying cell bodies of raphe neurons was determined by an image analyser
(Seescan, Cambridge) with a three stage thresholding routine. In a calibration experiment the amount of [35S]-labelled probe hybridised to cells on serial sections was varied by adjusting the specific radioactivity in the hybridisation solution (Gerfen, 1989) and there was found to be a linear relationship with grain number measured with the image analyser (r=0.988, PB0.02). The number of silver grains overlying ependymal cells was counted manually from enlarged micrographs.
3. Results Hybridisation for 5-HT transporter mRNA was carried out throughout the rat brain; a representative selection of coronal sections is shown in Fig. 1. It was highly specific relating almost entirely to the midsagittal raphe nuclei. The dorsal raphe nucleus is the largest labelled region, with individual composite cell clusters visible (eg. Fig. 1C). Labelling was also found in the median raphe, raphe pontis, raphe magnus, raphe obscurus and raphe pallidus nuclei. The only labelled extra-raphe neurons were the bilateral B9 groups located in the ventrolateral tegmentum of the midbrain, close to the medial lemniscus (Fig. 1B). No labelling was found in any other brain region including all areas of the cortex, hippocampus, thalamus, hypothalamus, cerebellum, brainstem and spinal cord (e.g. Fig. 1A,F). The emulsion dipped sections (Fig. 2) demonstrated specific labelling of serotonergic cell bodies (approximately half the neurons in the dorsal raphe) with negligible signal over others. Specific labelling was also exhibited around the caudal part of the aqueduct (Fig. 1C,D,E) where it enlarges to form the recess of the superior colliculus, sometimes termed the ‘mesencephalic ventricle’ (Mitro and Palkovits, 1981). Labelling was rapidly discontinued as the aqueduct opened into the fourth ventricle and neither was any associated with the lateral or third ventricles. Examination of emulsion dipped sections showed hybridisation to the ependymal layer (Fig. 3). However there did not appear to be homogenous labelling over all cells in the layer and the distribution of hybridisation was further studied. Quantitation of cellular labelling indicated that the number of silver grains overlying individual cells of the ependymal layer varied from zero to over 20 with a mean of 4.2 (Fig. 4A). This differed from background labelling which was determined by placing tracings of the ependymal cell outlines over unlabelled regions of the same micrographs. The background showed a tighter distribution (Fig. 4B) averaging 1.6 silver grains per cell and it resembled a Poisson distribution in shape as would be predicted for a random distribution of grains. If the assumption is made that the grains from hybridisation to ependymal cells should also show a Poisson distribu-
M.C. Swan et al. / Journal of Chemical Neuroanatomy 13 (1997) 71–76
73
Fig. 1. Distribution of 5-HT transporter mRNA in selected coronal sections of rat brain. (A – F) are photographs printed directly from autoradiographs with schematic line drawings below showing major visible anatomical features. Scale bar: 5.0 mm. Abbreviations: LV, lateral ventricle; 3V, third ventricle; Aq, aqueduct; 4V, fourth ventricle; DR, dorsal raphe nucleus; MnR, median raphe nucleus; RPn, raphe pontis nucleus; RMg, raphe magnus nucleus; RPa, raphe pallidus nucleus; ROb, raphe obscurus nucleus; B9, group of serotonergic cells within the medial lemniscus.
tion then modelling the observed data requires a minimum of three cell populations. For example, Fig. 4C models a mixed population of cells of which 45% have background labelling at an average 1.6 grains per cell, another 40% have specific labelling averaging five grains per cell and the final 15% average ten grains per cell; this predicted distribution is not significantly different from the experimental data (P \ 0.2, Kolmogorov-Smirnov test). The levels of neuronal hybridisation were studied in rats which had received paroxetine (5 mg/kg bodyweight per day for 21 days). Quantitation was based on the mean number of silver grains overlying labelled cells in the dorsal and median raphe nuclei, determined by image analysis. For each experimental animal 20 randomly chosen neuronal cell bodies which included a
nuclear profile were analysed for both the dorsal and median raphes and the mean number of grains determined. Following 21 days treatment with paroxetine, hybridisation was found not to differ significantly from the control group for either the dorsal raphe (P =0.15, Student’s 2-tailed t-test) or the median raphe (P =0.89) nuclei (Table 1).
4. Discussion These results confirm the anatomical distribution of 5-HT transporter mRNA to neuronal cell bodies within the raphe nuclei and the B9 region of the medial lemniscus (Fujita et al., 1993; Austin et al., 1994). This corresponds to the distribution of serotonergic cell bod-
74
M.C. Swan et al. / Journal of Chemical Neuroanatomy 13 (1997) 71–76
ies previously established using other techniques (Steinbusch, 1981; Tork, 1985). No further neuronal labelling was seen beyond the regions described. We conclude that neuronal expression of the transporter is entirely presynaptic, which is not necessarily the case for all neurotransmitter transporters (Swan et al., 1994). A finding of particular interest is the existence of a distinct loop of 5-HT transporter mRNA labelling surrounding the caudal part of the aqueduct, the circumference of which steadily increases as the aqueduct enlarges, until the fourth ventricle is reached upon which it disappears. The hybridisation appeared specific in that it was not observed in other parts of the ventricular system, it was displaced by an excess of unlabelled probe, and it was not observed when using a control probe (or indeed, probes to other mRNAs). It was observed in all 11 animals investigated. In principle we cannot rule out the existence in this part of the aqueduct of a different unknown mRNA containing a
Fig. 3. Detection of 5-HT transporter mRNA in the aqueductal ependyma. Dark-field (A,B) and the corresponding bright-field (C,D) photomicrographs of the aqueduct wall showing labelling of the ependyma; scale bar, 100 mm. Panels A and C illustrate hybridisation with the probe for 5-HT transporter mRNA; panels B and D illustrate a control experiment in which hybridisation was carried out in the presence of a 50-fold excess of unlabelled probe. Panel E shows a higher magnification photomicrograph of the labelled ependymal layer; scale bar, 20 mm.
Fig. 2. Cellular distribution of 5-HT transporter mRNA. Bright-field photomicrograph of (A) the dorsal raphe nucleus, (B) the median raphe nucleus, and (C) the B9 group of cells. Note the presence of both labelled and unlabelled cells. Scale bar, 20 mm.
sequence complementary to the probe, but the exquisite specificity of labelling elsewhere makes it most likely that we are detecting 5-HT transporter mRNA; even if it were a different mRNA it would still imply a functional specialisation of this part of the ventricular system.
M.C. Swan et al. / Journal of Chemical Neuroanatomy 13 (1997) 71–76
Histological investigation localises this labelling to the ependymal layer and quantitative analysis showed that the labelling pattern was consistent with just over one half of the ependymal cells undergoing hybridisation with the remainder negative which may reflect a functional heterogeneity among ependymal cells; this could be due to known ependymal cell types such as the ciliated epithelial cells and the tanycytes, or further
Fig. 4. Analysis of distribution of autoradiographic silver grains overlying ependymal cells hybridised for 5-HT transporter mRNA. (A) Distribution for 1499 cells (from five animals) analysed. (B) Distribution of background labelling, determined using outline tracings of the ependymal cells analysed in panel A in regions negative for hybridisation. (C) Model for the distribution of silver grains found in panel A, assuming a mixture of three cell populations each having a Poisson distribution of grain number with averages of 1.6 grains per cell (45% of cells), 5 grains per cell (40%) and 10 grains per cell (15%).
75
Table 1 Quantitation of 5-HT transporter mRNA in paroxetine-treated and control rats
Dorsal raphe Median raphe
Control (n =4)
Paroxetine-treated (n = 6)
85 95 68 98
739 15 67 910
Expression is determined as the number of silver grains (mean 9 S.D.) overlying neuronal cell bodies following in situ hybridisation.
degrees of specialisation. However the simplest explanation of the apparent heterogeneity is probably that of these polarised epithelial cells a proportion will not have their rough endoplasmic reticulum included in any particular histological section, and thus would not be labelled for membrane protein mRNAs associated with the endoplasmic reticulum including the 5-HT transporter. The fact that two cell populations with differing levels of hybridisation were required to model the experimental data (Fig. 4C) may not be meaningful since the assumption that distributions are of the Poisson type is not necessarily appropriate for a mixed cell population where the grain scatter due to the decay energy of sulphur-35 is of the order of a cell diameter. In theory there is the alternative explanation that the mRNA detected is not in the ependymal cells but the intercalated nerves. An extensive serotonergic plexus originating in the raphe nuclei has been shown to exist in the supra- and subependymal structures (Aghajanian and Gallager, 1975; Chan-Palay, 1976), which may act to modulate cerebrospinal fluid (CSF) composition in terms of 5-HT concentration. If the 5-HT transporter mRNA were present in these nerve terminals the small size of the nerves adjacent to close packed ependymal cells, coupled with the dispersion of silver grains resulting from the decay energy of sulphur-35, would mean that on hybridisation the label would appear to overlie a subpopulation of ependymal cells. However only in a few special cases has mRNA been found in the axonal domain of neurons (Steward and Banker, 1992), and the cellular machinery for synthesis of membrane proteins is absent. Instead we prefer to propose that the 5-HT transporter is synthesised by cells of the aqueductal ependyma. It indicates a specialisation of this region in terms of cellular function. Published images obtained by inhibitor binding have previously appeared to indicate the presence of the 5-HT transporter protein in the vicinity of the aqueduct but not other regions of the ventricular system (Hrdina et al., 1990) though in that study the identity of the cells where it was situated was not investigated. The presence of serotonin within CSF is well established and it has been hypothesised to control CSF volume and composition via the serotonin receptors on the apical surface of the choroid plexus
76
M.C. Swan et al. / Journal of Chemical Neuroanatomy 13 (1997) 71–76
epithelium (Nilsson et al., 1992); it may also have an effect on neural function judging by the correlation between the concentration of serotonin and its metabolites in the CSF and certain psychiatric conditions (Lester, 1995). An uptake mechanism specifically removing serotonin from the CSF would ensure sensitivity of the system to changes in the rate of secretion of serotonin into CSF via supraependymal nerve terminals. It is also conceivable that the location of the proposed serotonin uptake mechanism at the aqueduct would allow differential control of CSF composition between the forebrain and the hindbrain, spinal cord and subarachnoid space. No further novel sites of 5-HT transporter mRNA localisation were revealed in our study. Lesch et al. (1993b) used a reverse transcriptase-polymerase chain reaction approach to investigate areas of rat brain associated with ascending 5-HT pathways. They reported that 5-HT transporter mRNA may be present in low concentrations in the frontal cortex, hippocampus and neostriatum and attributed this to specific transport of mRNA to the terminal regions of the 5-HT pathways in order to institute local transporter synthesis. However we were unable to confirm their distribution using in situ hybridisation (see Fig. 1A,B). If such a transport mechanism exists the amount of mRNA involved must be so small that it can only be detected by sensitive amplification methods, which implies that its contribution to total 5-HT transporter synthesis will be marginal. It could easily be envisaged that expression of the 5-HT transporter might be under feedback control which would result in it being altered by pharmacological manipulation of extracellular 5-HT levels by SSRIs, similar to the change in expression reported following exposure to tricyclic antidepressants (Lesch et al., 1993b; Lopez et al., 1994). We quantified 5-HT transporter mRNA at the cellular level and though we found a marginal decrease in both the dorsal and median raphe nuclei after 21 days paroxetine treatment, the change was not statistically significant. Lesch et al. (1993b) estimated from Northern blot analysis that 21-day administration of another SSRI, fluoxetine (2.5 mg/kg per day via an osmotic pump), caused a 30% decrease in 5-HT transporter mRNA expression, but no statistical analysis was presented. We are unable to say whether our inability to confirm their finding is due to the use of a different SSRI, or the difference in methodology for determining mRNA expression levels.
Acknowledgements This work was supported in part by a grant from the Medical Research Council.
References Aghajanian, G.K., Gallager, D.W., 1975. Raphe origin of serotonergic nerves terminating in the cerebral ventricles. Brain Res. 88, 221–231. Austin, M.C., Bradley, C.C., Mann, J.J., Blakely, R.D., 1994. Expression of serotonin transporter messenger RNA in the human brain. J. Neurochem. 62, 2362 – 2367. Blakely, R.D., Berson, H.E., Fremau, R.T., Caron, M.G., Peek, M.M., Prince, H.K., Bradley, C.C., 1991. Cloning and expression of a functional serotonin transporter from rat brain. Nature 354, 66–70. Chan-Palay, V., 1976. Serotonin axons in the supra- and subependymal plexuses and in the leptomeninges; their roles in local alterations of cerebrospinal fluid and vasomotor activity. Brain Res. 102, 103 – 130. Fujita, M., Shimada, S., Maeno, H., Nishimura, T., Tohyama, M., 1993. Cellular localization of serotonin tranporter mRNA in the rat brain. Neurosci. Lett. 162, 59 – 62. Gerfen, C.R., 1989. Quantification of in situ hybridization histochemistry for analysis of brain function. Meth. Neurosci. 1, 79–97. Hoffman, B.J., Mezey, E., Brownstein, M.J., 1991. Cloning of a serotonin transporter affected by antidepressants. Science 254, 579 – 580. Hrdina, P.D., Foy, B., Hepner, A., Summers, R.J., 1990. Antidepressant binding sites in brain: Autoradiographic comparison of [3H]paroxetine and [3H]imipramine localization and relationship to serotonin transporter. J. Pharmacol. Exp. Ther. 252, 410–418. Lesch, K.P., Wolozin, B.L., Estler, H.C., Murphy, D.L., Riederer, P., 1993a. Isolation of a cDNA encoding the human brain serotonin transporter. J. Neural Transm. Gen. Sect. 91, 67 – 72. Lesch, K.P., Aulakh, C.S., Wolozin, B.L., Tolliver, T.J., Hill, J.L., Murphy, D.L., 1993b. Regional brain expression of serotonin transporter mRNA and its regulation by reuptake inhibiting antidepressants. Brain Res. Mol. Brain Res. 17, 31 – 35. Lester, D., 1995. The concentration of neurotransmitter metabolites in the cerebrospinal fluid of suicidal individuals: A meta-analysis. Pharmacopsychiatry 28, 45 – 50. Lopez, J.F., Chalmers, D.T., Vazquez, D.M., Watson, S.J., Akil, H., 1994. Serotonin transporter mRNA in rat brain is regulated by classical antidepressants. Biol. Psychiatry 35, 287 – 290. Mitro, A., Palkovits, M., 1981. Morphology of the Rat Brain Ventricles, Ependyma and Periventricular Structures. Karger, Basel. Najlerahim, A.R., Harrison, P.J., Barton, A.J.L., Heffernan, J., Pearson, R.C.A., 1990. Distribution of messenger mRNAs encoding the enzymes glutaminase, aspartate aminotransferase and glutamic acid decarboxylase in rat brain. Brain. Res. Mol. Brain Res. 7, 317–333. Nilsson, C., Lindvall-Axelsson, M., Owman, C., 1992. Neuroendocrine regulatory mechanisms in the choroid plexus-cerebrospinal fluid system. Brain Res. Rev. 17, 109 – 138. Steinbusch, H.W.M., 1981. Distribution of serotonin-immunoreactivity in the CNS of rat brain-cell bodies and terminals. Neuroscience 6, 557 – 618. Steward, O., Banker, G.A., 1992. Getting the message from the gene to the synapse: Sorting and intracellular transport of RNA in neurones. Trends Neurosci. 15, 180 – 186. Swan, M.C., Najlerahim, A.R., Watson, R.B., Bennett, J.P., 1994. Distribution of mRNA for the GABA transporter GAT-1 in the rat brain: Evidence that GABA uptake is not limited to presynaptic neurons. J. Anat. 185, 315 – 323. Tork, I., 1985. Raphe nuclei and serotonin containing systems. In: Paxinos, G., (Ed.), The Rat Nervous System, vol. 2. Academic Press, Sydney, pp. 43 – 69. Uhl, G.R., 1992. Neurotransmitter transporters (plus): A promising new gene family. Trends Neurosci. 15, 265 – 268. Wolf, W.A., Kuhn, D.A., 1990. Modulation of serotonin release: Interactions between the serotonin transporter and autoreceptors. Ann. New York Acad. Sci. 604, 505 – 513.