Significant outcomes
Even drastic alterations of extracellular serotonin levels translate into comparatively minor changes in serotonergic gene expression.
While transcript levels of serotonergic genes in the raphe region and areas such as amygdala and striatum were clearly affected by manipulations of serotonin levels; small or no changes could be observed in the hypothalamus and prefrontal cortex.
Short-term SSRI treatment is associated with downregulation of genes encoding enzymes regulating serotonin synthesis in the raphe nuclei (Tph2, Ddc).
Limitations
The dissection techniques employed unfortunately preclude analysis of potentially important effects in sub-areas, such as differences in expression patterns between the dorsal and median raphe nuclei.
Although analysis of behavioural correlates of observed effects on gene expression could have been potentially informing, this is not possible as no such tests were performed.
As with all gene expression studies, the functional relevance (e.g. the impact on protein levels and function) is somewhat unclear.
Introduction
The assumption that brain serotonin plays a key role in the regulation of various aspects of human behaviour, and is the target for important psychoactive drugs, has inspired numerous attempts to measure brain serotonergic activity in experimental animals. To this end, a variety of techniques has been applied, among which assessment of the expression of serotonin-related genes, not least because of its convenience, has become one of the most commonly used. Thus, analysis of the expression of genes encoding serotonin-related proteins has, for example, been applied for addressing how serotonergic transmission may be influenced by factors such as drugs (Barbon et al., Reference Barbon, Orlandi, La Via, Caracciolo, Tardito, Musazzi, Mallei, Gennarelli, Racagni, Popoli and Barlati2011; Yamamura et al., Reference Yamamura, Abe, Nakagawa, Ochi, Ueno and Okada2011; McQuade et al., Reference McQuade, Leitch, Gartside and Young2004), electroconvulsive treatment (Shen et al., Reference Shen, Numachi, Yoshida, Fujiyama, Toda, Awata, Matsuoka and Sato2003), hormones (McQueen et al., Reference McQueen, Wilson, Sumner and Fink1999; Donner & Handa, Reference Donner and Handa2009), maternal separation (Gardner et al., Reference Gardner, Hale, Lightman, Plotsky and Lowry2009), stress (Bethea et al., Reference Bethea, Phu, Reddy and Cameron2013) and social defeat (Boyarskikh et al., Reference Boyarskikh, Bondar, Filipenko and Kudryavtseva2013).
While the underlying assumption for this strategy has been that messenger RNA (mRNA) levels to some extent may reflect the levels of the corresponding protein, and that regulation of gene expression may be one important route for factors exerting long- or short-term influences on brain serotonergic activity, it remains to be clarified to what extent up- and downregulation of gene expression are indeed important mechanisms for inducing transient fluctuations and/or stable alterations in the status of a certain transmitter, for example, for the purpose of maintaining homeostasis. For example, while studies such as those cited above indicate that the expression of serotonin-related genes is indeed influenced by various external influences, it has been shown that knockout of the gene encoding the serotonin-synthesising enzyme tryptophan hydroxylase 2 (Tph2), leading to marked depletion of brain serotonin levels, did not influence the expression of any of a large number of serotonin-related genes that were assessed (Kriegebaum et al., Reference Kriegebaum, Song, Gutknecht, Huang, Schmitt, Reif, Ding and Lesch2010). Thus, to facilitate the interpretation of studies using mRNA assessment to reflect the status of brain serotonergic transmission, it is important to increase the insight into how responsive the expression of various serotonin-related genes is to interventions known to influence extracellular serotonin levels.
In this vein, the present study was undertaken to evaluate the possible influence of short-term alterations in synaptic levels of serotonin on a large number of serotonin-related genes, both in the raphe nuclei, where the serotonergic cell bodies reside, and in various terminal regions; the hypothesis being that such changes would induce rapid adaptive responses in some but not all of the studied genes. To this end, one group of animals was injected for 3 days with the selective serotonin reuptake inhibitor (SSRI) paroxetine, that is, a drug that is likely to cause at least a moderate increase in extrasynaptic levels of serotonin in at least some of the studied terminal regions (see refs in Fuller, Reference Fuller1994) and, at the same time, to dampen the electrical activity of the serotonergic cell bodies in the raphe region (as the result of an autoreceptor-mediated feedback) (Hajós et al., Reference Hajós, Gartside and Sharp1995). Another group of rats was subjected to 3 days of treatment with the serotonin synthesis inhibitor, para-chlorophenylalanine (p-CPA), in a regimen known to cause a depletion of tissue serotonin exceeding 95% (Miczek et al., Reference Miczek, Altman, Appel and Boggan1975; Näslund et al., Reference Näslund, Studer, Nilsson, Westberg and Eriksson2013). Supporting that the applied treatments do in fact produce a significant alteration of serotonergic output, they are known to elicit reciprocal behavioural effects that are likely to be the result of such an influence; thus, while p-CPA enhances aggression (Miczek et al., Reference Miczek, Altman, Appel and Boggan1975) and sexual activity (Tagliamonte et al., Reference Tagliamonte, Tagliamonte, Gessa and Brodie1969) and reduces anxiety-like behaviour (as assessed, e.g. using the elevated plus maze; Treit et al., Reference Treit, Robinson, Rotzinger and Pesold1993; Näslund et al., Reference Näslund, Studer, Petterson, Hagsater, Nilsson, Nissbrandt and Eriksson2015), acute administration of SSRIs leads to the opposite effects (Griebel et al., Reference Griebel, Moreau, Jenck, Misslin and Martin1994; Vega Matuszcyk et al., Reference Vega Matuszcyk, Larsson and Eriksson1998; Ho et al., Reference Ho, Olsson, Westberg, Melke and Eriksson2001).
Aims of the study
We aimed to investigate, by way of measurement of transcript levels of a comprehensive set of serotonergic genes, the responsiveness of the serotonin system to short-term manipulation of synaptic serotonin levels. We also desired to characterise the effects of such interventions per se and to compare these results with earlier work by other groups.
Experimental procedures
Study design and protocol
Thirty-three male Wistar rats, aged 9–10 weeks at arrival, were kept in an animal facility for 5 weeks, whereupon they were randomly divided into three equal groups and administered paroxetine, p-CPA or saline for 3 days. On the morning of the 4th day, they were given a final injection and sacrificed 2 h later. Brains were immediately extracted and dissected on dry ice.
In five regions receiving substantial serotonergic innervation, that is, amygdala, hippocampus, striatum, hypothalamus and prefrontal cortex, we assessed the genes encoding i) nine different serotonergic receptor subtypes, 5HT1A (Htr1a), 5HT1B (Htr1b), 5-HT1D (Htr1d), 5-HT2A (Htr2a), 5-HT2C (Htr2c), 5-HT3A (Htr3a), 5-HT4 (Htr4), 5-HT6 (Htr6) and 5-HT7 (Htr7); ii) the two subtypes of the monoamine metabolising enzyme, monoamine oxidases A and B (Maoa and Maob); iii) brain-derived neurotrophic factor (Bdnf), which is a protein known to exert an important impact on serotonergic transmission (Rumajogee et al., Reference Rumajogee, Vergé, Darmon, Brisorgueil, Hamon and Miquel2005); iv) the BDNF receptor (Ntrk1); and v) P11 (S100a10), which is another protein attributed to important interactions with brain serotonergic neurons (Anisman et al., Reference Anisman, Du, Palkovits, Faludi, Kovacs, Szontagh-Kishazi, Merali and Poulter2008; gene names within parentheses). In the raphe nuclei, where the serotonergic cell bodies reside, we studied a number of genes known to be expressed by serotonergic neurons and/or suggested to interact with these, such as those encoding i) the three serotonergic autoreceptors, that is, Htr1a, Htr1b and Htr1d; ii) the enzymes involved in the synthesis of serotonin, that is, one of the two isoforms of tryptophan hydroxylase (Tph2) and amino acid-decarboxylase (Ddc); iii) the serotonin transporter (Slc6a4); iv) the monoamine vesicular transporter (Slc18a2); v) Maoa and Maob; vi) three transcription factors expressed by serotonergic neurons and of importance for the development (and possibly maintenance) of serotonergic neurons and expressed by these: GATA-2 (Gata2; Craven et al., Reference Craven, Lim, Ye, Engel, de Sauvage and Rosenthal2004), MASH-1 (Ascl1; Pattyn et al., Reference Pattyn, Simplicio, van Doorninck, Goridis, Guillemot and Brunet2004) and PET-1 (Fev) (Liu et al., Reference Liu, Maejima, Wyler, Casadesus, Herlitze and Deneris2010); vii) Bdnf; viii) Ntrk2; and ix) S100a10. Finally, in order to shed further light on the possible differences between different regions with respect to how strongly they were influenced by the two interventions, the expression of the immediate-early gene c-Fos was assessed both in raphe and in the different terminal regions.
Animals
Animals were obtained from Taconic (Ejby, Denmark) and housed with a 12-h light/dark cycle (lights on at 06:00 a.m.) and with standard chow and water available ad libitum. All procedures were carried out in accordance with national and European legislation and with the approval of the local ethics committee and in accordance with the institutional guidelines.
Drug treatment
The p-CPA (Sigma-Aldrich, St Louis, MO, USA) was dissolved in 0.9% saline and administered intraperitoneally as one injection of 300 mg/kg per day for 3 days, with the last injection being given roughly 24 h before sacrifice. Paroxetine hydrochloride (Jai Radhe Chemicals, Ahmedabad, India) was dissolved in 0.9% saline and administered subcutaneously at a dose of 10 mg/kg 2 times per day. All animals were given two injections daily 10–12 h apart; the paroxetine group was given paroxetine 10 mg/kg at both occasions, the p-CPA group was given p-CPA (300 mg/kg) in the first injection and saline in the next and the saline group was given saline at both occasions. On the 4th day, the paroxetine group was administered a last dose of paroxetine 2 h before sacrifice, while the other two groups received saline.
Dissection
Brains were extracted immediately after decapitation. The extended amygdala, hippocampus, striatum, hypothalamus, prefrontal cortex and raphe nuclei were dissected out for gene expression analysis. Brain tissue samples were immediately frozen on dry ice and stored at −80°C.
Gene expression
Individual samples of brain tissue were homogenised in Qiazol (Qiagen, Hilden, Germany) using a TissueLyzer (Qiagen). Total RNA was extracted with an RNeasy Lipid Tissue Mini Kit (Qiagen) using a QiaCube (Qiagen). RNA quantities were determined, and quality assessed, by means of spectrophotometric measurements (Nanodrop 1000; Thermo Scientific, Wilmington, DE, USA). For complementary DNA (cDNA) synthesis, 4000 ng of total RNA was reversely transcribed using random hexamers (Applied Biosystems, Sundbyberg, Sweden) and Superscript III reverse transcriptase (Invitrogen Life Technologies, Paisley, UK) according to the manufacturer’s description. Recombinant RNaseout® Ribonuclease Inhibitor (Invitrogen) was added to prevent RNase-mediated degradation. All cDNA reactions were run in duplicate and the products pooled for further analysis.
Real-time qPCR was performed by means of TaqMan® Custom Arrays using TaqMan probe and primer sets for target genes and reference genes chosen from an online catalogue (Applied Biosystems). Two separate cards were used: one to investigate the raphe region and one to investigate the various target areas. Names and assay numbers for genes investigated are shown in Supplementary Table 1. The sets were factory loaded into the 384 wells of the TaqMan® Array, each port being loaded with cDNA corresponding to 500 ng total RNA combined with nuclease-free water and 50 µl TaqMan® Gene Expression Master Mix (Applied Biosystems) to a final volume of 100 µl. The TaqMan® Arrays were analysed using the 7900 HT system with a TaqMan Array Upgrade (Applied Biosystems). Thermal cycling conditions were 50°C for 2 min and 94.5°C for 10 min, followed by 40 cycles of 97°C for 30 s and 59.7°C for 1 min.
All reactions were run in duplicate, and reactions were excluded if the Cycle threshold (CT) values of the duplicates differed by more than 5%. By means of the NormFinder algorithm (http://moma.dk/normfinder-software), a combination of Hmbs and Ppia was found to display the highest stability among the four reference genes in all areas examined and was therefore used to normalise the expression levels. Gene expression values were calculated based on the ΔΔC t method (Livak & Schmittgen, Reference Livak and Schmittgen2001). Table 1 provides the relative expression of each gene; ΔΔCt values for each gene are given in Supplementary Table 2.
Gene expression effects of short-term manipulation of synaptic 5-HT levels
Treatment effects of short-term treatment with p-CPA or paroxetine. Numbers indicate expression levels relative to the saline group, with asterisks (*) indicating level of significance for the comparison.
n = 10–11 (Slc6a4 in amygdala = 9)
*p < 0.05, **p < 0.01, ***p < 0.001.
Statistical analyses
Student’s t-tests were used to test for treatment effects as compared to placebo; we chose not to use an analysis of variance-based approach as we did not aim to investigate the possible differences between SSRI- and p-CPA-treated animals. As this was a hypothesis-driven study, where some of the observed genes were expected to change in a certain direction, whereas for others there was no a priori hypothesis, the results are presented without any correction for multiple testing; nevertheless, permutation analyses were performed and are included in the supplementary online information (Supplementary Table 3). SPSS for Mac version 21 (IBM, Chicago, IL, USA) was used for all statistical procedures, except for the permutation analysis, where R (R Core team, Vienna, Austria) was employed.
Results
para-Chlorophenylalanine
Htr1b, Htr2a and Htr3a were upregulated in the amygdala, while Htr2c was upregulated in the striatum. Htr4 was upregulated in the amygdala, while Htr6 was downregulated in hippocampus and prefrontal cortex. Maob was upregulated in the striatum and raphe. Bdnf was upregulated in the amygdala, hippocampus and raphe, while Ntrk2 was upregulated in the striatum. Slc6a4 was downregulated in the hypothalamus, prefrontal cortex and raphe. S100a10 was upregulated in the hippocampus. Ddc was downregulated in the raphe. Fos was upregulated in the hippocampus and raphe (Table 1). Nine of the observed effects survived correction for multiple comparisons by means of permutation analyses with subsequent area-by-area Holm-Bonferroni correction: Htr1b and Htr2a in the amygdala; Bdnf in the hippocampus; Htr2c in the striatum; Htr4 and Slc6a4 in the prefrontal cortex; and Bdnf, Slc6a4 and Ddc in the raphe (Supplementary Table 3).
Paroxetine
Htr1a and Htr1d were upregulated in the hippocampus. Htr2a was downregulated in the hypothalamus. Htr2c and Htr3a were upregulated in the striatum, and the latter gene was also upregulated in the amygdala. Htr4 was upregulated in the amygdala. Maoa was upregulated in the amygdala, while Maob was upregulated in amygdala and hippocampus. Ntrk2 and Slc6a4 were upregulated in the striatum. Three genes were significantly downregulated by paroxetine in the raphe region: Tph2, Ddc and Fev. Fos was upregulated in hypothalamus (Table 1). None of the effects except for the downregulation of Tph2 in the raphe survived correction for multiple comparisons (Supplementary Table 3).
Discussion
One major conclusion of this study is that both paroxetine and p-CPA left the expression of most of the studied genes unaffected, the major exceptions being discussed below. It may hence be concluded either that short-term changes in extracellular serotonin levels and/or serotonergic cell firing, elicit an adaptive modulation of the formation of only a minority of the proteins in question, or that such adaptations are not easily captured using conventional assessment of mRNA expression. Of note is, for example, that the expression of the gene encoding the 5HT1A receptor (Popova & Naumenko, Reference Popova and Naumenko2013), which is regarded as one of the most important mediators of the effects of serotonin on postsynaptic neurons, and which also serves as an autoreceptor at serotonergic cell bodies, was not influenced by any of the two treatments.
The observation that serotonin depletion induced a marked upregulation of Bdnf both in the raphe region and in the hippocampus is in agreement with an earlier study (Zetterström et al., Reference Zetterström, Pei, Madhav, Coppell, Lewis and Grahame-Smith1999) as well as with a report on Tph2 knock-out mice (Migliarini et al., Reference Migliarini, Pacini, Pelosi, Lunardi and Pasqualetti2012). The lack of effect of short-term administration of paroxetine on Bdnf, both in raphe and in the terminal regions, is also in agreement with the literature, where acute and sub-acute SSRI treatment has been reported to exert either modest down-regulating or no effects on the expression of this gene (Nibuya et al., Reference Nibuya, Morinobu and Duman1995; Zetterström et al., Reference Zetterström, Pei, Madhav, Coppell, Lewis and Grahame-Smith1999; Mannari et al., Reference Mannari, Origlia, Scatena, del Debbio, Catena, dell’Agnello, Barraco, Giovannini, Dell’osso, Domenici and Piccinni2008). Long-term administration of SSRIs, on the other hand, is reported to increase Bdnf expression (Mannari et al., Reference Mannari, Origlia, Scatena, del Debbio, Catena, dell’Agnello, Barraco, Giovannini, Dell’osso, Domenici and Piccinni2008), that is, to exert an effect similar to that of acute serotonin depletion.
Apart from the effect on Bdnf, there were very minor effects of 72 h of serotonin depletion on the expression of serotonin-related genes in the raphe. This finding is indirectly in line with previous reports according to which p-CPA does not influence the firing rate of raphe serotonergic neurons, that is, that these neurons are not under tonic feedback inhibition by extracellular levels of serotonin (Aghajanian et al., Reference Aghajanian, Graham and Sheard1970; Chaput et al., Reference Chaput, Lesieur and de Montigny1990). Likewise, blockade of 5-HT1A autoreceptors does not exert any effect on raphe firing when administered per se (Mundey et al., Reference Mundey, Fletcher and Marsden1994).
According to the traditional concept of denervation supersensitivity, an expected effect of serotonin depletion on serotonergic receptors would be one of compensatory upregulation. In line with this, we observed an increase in Htr1b and Htr2a expression in the amygdala and of Htr2c in the striatum, in p-CPA-treated animals, the latter finding being in line with a previous study (Compan et al., Reference Compan, Segu, Buhot and Daszuta1998). In contrast, none of the other receptors were significantly upregulated by this treatment. While it cannot be excluded that also other receptors would have been upregulated had the depletion lasted for longer than 72 h, our results suggest that 5-HT1B and 5-HT2A receptors in the amygdala, and 5-HT2C receptors in the striatum, are particularly sensitive to changes in extracellular levels of serotonin.
With respect to the 5-HT6 receptor, the opposite effect of serotonin depletion was found; p-CPA thus induced a marked reduction in the expression of the gene encoding this receptor in prefrontal cortex and a somewhat less pronounced effect in the same direction in the hippocampus. A previous study found support for 5-HT6 receptors in the prefrontal cortex to exert an inhibitory influence on the local release of serotonin (Schechter et al., Reference Schechter, Lin, Smith, Zhang, Shan, Platt, Brandt, Dawson, Cole, Bernotas, Robichaud, Rosenzweig-Lipson and Beyer2008); so from the perspective of adaptive feedback, a drug-induced shortage of serotonin leading to a downregulation of a receptor inhibiting serotonin release bears some logic.
Previous studies have suggested that serotonin depletion, as obtained using reserpine for 3 days (Xiao et al., Reference Xiao, Pawlyk and Tejani-Butt1999) or p-CPA for 1–2 (Rattray et al., Reference Rattray, Baldessari, Gobbi, Mennini, Samanin and Bendotti1996) or 10 days (Linnet et al., Reference Linnet, Koed, Wiborg and Gregersen1995), leads to downregulation of the expression of the gene encoding the serotonin transporter, that is, to a change in the expression of this gene, which would make sense from an adaptive point of view. Our observation of reduced levels of Slc6a4 expression in several brain regions of serotonin-depleted animals is hence in line with earlier work.
It has since long been known that inhibition of serotonin reuptake, as the result of an autoreceptor-mediated feedback, elicits an immediate reduction in serotonergic cell firing (Gallager & Aghajanian, Reference Gallager and Aghajanian1975; Hajós et al., Reference Hajós, Gartside and Sharp1995) as well as a decrease in serotonin turnover (Carlsson et al., Reference Carlsson, Jonason and Lindquist1969; Fuller & Wong, Reference Fuller and Wong1977). In line with these observations, raphe Tph2 expression was markedly reduced in rats exposed to paroxetine for 3 days. The TPH2 enzyme being downregulated by SSRIs has been shown before, both by means of histochemical methods (MacGillivray et al., Reference MacGillivray, Lagrou, Reynolds, Rosebush and Mazurek2010) and by means of assessment of mRNA expression (Dygalo et al., Reference Dygalo, Shishkina, Kalinina, Yudina and Ovchinnikova2006; Abumaria et al., Reference Abumaria, Rygula, Hiemke, Fuchs, Havemann-Reinecke, Rüther and Flügge2007; Shishkina et al., Reference Shishkina, Kalinina and Dygalo2007; Klomp et al., Reference Klomp, Hamelink, Feenstra, Denys and Reneman2014); however, a common view has been that such an effect requires weeks of treatment to be at hand (Dygalo et al., Reference Dygalo, Shishkina, Kalinina, Yudina and Ovchinnikova2006). In contrast, our results suggest that increased extracellular serotonin levels in the raphe region leads to a downregulation of Tph2 that takes no more than 3 days to be manifest. In addition, it is of note that two other raphe genes that are (more or less) exclusively expressed by serotonergic neurons, and the expression of which hence might be expected to be influenced by the activity of the neurons, that is, Ddc and Fev, also displayed reduced expression in SSRI-treated animals, and that there was a tendency for Slc6a4 in the same direction. A previous study showed reduced expression of Slc6a4 after 7, but not 4, days of fluoxetine treatment (Neumaier et al., Reference Neumaier, Root and Hamblin1996). With respect to the influence of paroxetine on serotonergic receptors, it is noteworthy that this was always one of up- rather than downregulation, with Htr2a in the hypothalamus being the sole exception.
It is noteworthy that the striatum, the hippocampus and the amygdala were the areas that appeared most influenced by sub-acute SSRI treatment and that this observation is also supported by the assessment of c-Fos activation (Beck, Reference Beck1995; Torres et al., Reference Torres, Horowitz, Laflamme and Rivest1998; Morelli et al., Reference Morelli, Pinna, Ruiu and del Zompo1999). Partly in line with this, previous studies using in vivo microdialysis have shown short-term administration of an SSRI to cause a marked increase in the extracellular levels of serotonin in the striatum (Kalén et al., Reference Kalén, Strecker, Rosengren and Björklund1989), hippocampus (Sabol et al., Reference Sabol, Richards and Seiden1992; Bosker et al., Reference Bosker, Klompmakers and Westenberg1995) and the amygdala (Sundblad & Eriksson, Reference Sundblad and Eriksson1997; Bosker et al., Reference Bosker, Cremers, Jongsma, Westerink, Wikström and den Boer2001); in contrast, the possible effect of the same treatment in prefrontal cortex, where we observed no effects on gene expression, appears less clear-cut and has remained a matter of controversy (Sarkissian et al., Reference Sarkissian, Wurtman, Morse and Gleason1990; Adell & Artigas, Reference Adell and Artigas1991; Gartside et al., Reference Gartside, Umbers, Hajós and Sharp1995).
To summarise, in order to facilitate the interpretation of studies assessing the expression of serotonin-related genes to gain insight into the status of brain serotonergic neurotransmission after various experimental interventions, we have explored to what extent gene expression is influenced by short-term changes in extracellular levels of serotonin and/or in serotonergic nerve activity. The results show some of the studied genes to be markedly influenced while most were not. Follow-up studies assessing the possible influence of similar manipulations of the extracellular levels of serotonin being in place for a more prolonged time period are warranted.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/neu.2020.9.
Acknowledgements
The authors gratefully acknowledge the expert technical assistance of Ms. Gunilla Bourghardt, Ms. Inger Oscarsson and Ms. Ann-Christine Reinhold. We thank the Genomics Core Facility at Sahlgrenska Academy, University of Gothenburg, Sweden, for participating in the gene expression experiments.
Authors’ contributions
JN and EE designed the study. JN and ES performed the experiments. SN provided statistical expertise. JN wrote the first draft of the manuscript. All authors contributed to, and have approved, the final version of the manuscript.
Financial support
Financial support was obtained from the Swedish Science Council (grant number K2010-61x-4961-01-3), Söderberg’s Foundation (MT30/09), Hållsten’s Foundation (N/A) and the Brain Foundation (FO2011-0293). Apart from providing grants, none of the funding sources took any active part in this work.
Conflicts of interest
None of the authors report any conflict of interest of relevance to the work here presented.
Open access
