International Journal of Neuropsychopharmacology (2012), 15, 617–629. f CINP 2011 doi:10.1017/S1461145711000733
ARTICLE
Molecular adaptation to chronic antidepressant treatment: evidence for a more rapid response to the novel a2-adrenoceptor antagonist/ 5-HT-noradrenaline reuptake inhibitor (SNRI), S35966, compared to the SNRI, venlafaxine Florence Serres1, Mark J. Millan2 and Trevor Sharp1 1 2
University Department of Pharmacology, Oxford, UK Institut de Recherches Servier, Croissy/Seine, France
Abstract Evidence of early changes in neural plasticity may aid the prediction of rapid-onset antidepressant drugs. Here we compared the dual a2-adrenoceptor antagonist/5-HT-noradrenaline reuptake inhibitor (SNRI), S35966, to the SNRI, venlafaxine, with regards to their effect on rat brain expression of a panel of neural plasticity-related genes : Arc, BDNF, and VGLUT1, as well as Homer1a and Shank1B (not studied previously). Abundance of mRNA was determined by in-situ hybridization in cortical and hippocampal regions 2 h and 16 h following drug administration for 14, 7 and 1 d. After 14 d, both S35966 and venlafaxine increased mRNA of all genes, including Homer1a and Shank1B, and effects were similarly time- and region-dependent. After 7 d, S35966 elevated Arc, Shank1B and BDNF mRNA, whereas venlafaxine increased Shank1B mRNA only. Finally, after 1 d (acute administration), S35966 increased Arc and Homer1a mRNA whereas venlafaxine had no effect on any gene examined. In summary, a 14-d course of treatment with S35966 or venlafaxine induced similar region- and time-dependent increases in expression of neural plasticity-related genes including Shank1B and Homer1a. Some genes responded earlier to S35966, suggesting that drugs with combined a2-adrenoceptor antagonist/SNRI properties may elicit more rapid changes in markers of neural plasticity than a SNRI alone. Received 17 November 2010 ; Reviewed 14 December 2010 ; Revised 11 April 2011 ; Accepted 12 April 2011 ; First published online 17 May 2011 Key words : Antidepressant, a2-adrenoceptor, neural plasticity, S35966, venlafaxine.
Introduction Currently available antidepressant drugs require administration over several weeks to produce a full therapeutic effect, implicating neuroadaptive changes in their mechanism of action (Katz et al. 2004 ; Millan, 2006 ; Papakostas et al. 2006). The search for antidepressant drugs with an early onset of therapeutic effect is hampered by a limited number of suitable experimental models with predictive and construct validity (Cryan & Slattery, 2007 ; Frazer & Morilak, 2005 ; Krishnan et al. 2008) although there are promising reports to the contrary (e.g. Santarelli et al. 2003 ; Stone Address for correspondence : Dr F. Serres, University Department of Pharmacology, Mansfield Road, Oxford OX1 3QT, UK. Tel. : +441865 271851 Fax : +44 1865 271853 Email : fl
[email protected]
et al. 2008). Recent attempts to model the neuroadaptive changes induced by treatment with antidepressants have included a genome-wide molecular approach. To date, genome-wide array analysis have produced interesting data but no precise and consistent molecular ‘fingerprint ’ of antidepressant action has yet emerged (Conti et al. 2007 ; Drigues et al. 2003 ; Sillaber et al. 2008). An alternative to the genome-wide approach is to investigate the effect of antidepressant treatment on candidate genes known to influence neural plasticity. Indeed, recent data suggest that a course of several weeks’ administration of standard antidepressants induces changes in specific molecular markers of neural plasticity, of potentially high relevance to the pathophysiology of depression and the actions of antidepressant drugs (Castren, 2005 ; Krishnan & Nestler,
618
F. Serres et al.
2008 ; Mathew et al. 2008). Thus, mRNA and protein expression of brain-derived neurotrophic factor (BDNF), activity-regulated cytoskeletal associated protein (Arc) and vesicular glutamate transporter1 (VGLUT1) increased in response to repeated but not acute administration of clinically used antidepressant drugs including monoamine uptake inhibitors (De Foubert et al. 2004 ; Nibuya et al. 1995 ; Pei et al. 2003 ; Tordera et al. 2005). These data suggest that several weeks’ administration of conventional antidepressants is required to induce a cascade of multiple genes linked to neural plasticity. Moreover, an antidepressant with a faster onset of therapeutic effect would produce such changes but with an earlier onset. Drugs that block the reuptake of both 5-hydroxytryptamine (5-HT ; serotonin) and noradrenaline, such as venlafaxine, markedly increase extracellular levels of monoamines in the brain and exert robust antidepressant actions in experimental models as well as patients (Koch et al. 2003 ; Millan et al. 2001a, b ; Thase, 2008 ; van den Broek et al. 2009). However, the actions of these drugs are limited by feedback inhibitory actions of noradrenaline at a2-autoreceptors. To overcome this feedback inhibition, one putative rapidonset antidepressant strategy under investigation involves the combination of a2-adrenoceptor antagonism and monoamine reuptake blockade (Andres et al. 2005 ; Blier et al. 2010 ; Invernizzi & Garattini, 2004 ; Millan, 2006). This strategy is supported by findings that a2adrenoceptor antagonists potentiate the increase in extracellular corticolimbic levels of noradrenaline and 5-HT induced by monoamine reuptake inhibitors (Gobert et al. 1997 ; Millan et al. 2000 ; Ortega et al. 2010). Further, a2-adrenoceptor antagonists augment antidepressant effects of monoamine reuptake inhibitors in certain animal models of depression (Dhir & Kulkarni, 2007 ; Dwyer et al. 2010) and depressed patients (Cappiello et al. 1995 ; Sanacora et al. 2004). More recently, an a2-adrenoceptor antagonist was found to accelerate the antidepressant effect of chronic imipramine in behavioural models, and enhance the expression of Arc and BDNF, as well as neurogenesis (Yanpallewar et al. 2010). Preliminary data on drugs with combined a2adrenoceptor antagonist and selective serotonin and noradrenaline reuptake inhibitor (SNRI) actions, including S35966 [4(5)-(5-fluoroindan-2-ylmethyl)-4,5dihydro-imidazole], have been described (Andres et al. 2005 ; Cordi et al. 2001 ; Millan et al. 2001c). S35966 is a potent antagonist at native, rat, cerebral and cloned, human a2-adrenoceptors (A, B, C subtypes), and rat and human 5-HT and noradrenaline (but not
dopamine) transporters (Gobert et al. 2002). In microdialysis experiments S35966 increased cortical extracellular noradrenaline and 5-HT to levels greater than obtained with a SNRI alone (Millan et al. 2001c). In the hippocampus, S35966 enhanced neurogenesis (Soumier et al. 2006). Moreover, over doses of 0.63–10.0 mg/kg, S35966 expresses antidepressant properties in a range of behavioural models in rodents including forced swim test, chronic mild stress, marble-burying and social aggression (Millan et al. 2001c ; Servier, unpublished data). The present study investigated the effect of S35966, compared to the SNRI venlafaxine, on the expression in corticolimbic brain regions of a panel of genes involved in mechanisms of neural plasticity. This panel comprised BDNF, Arc, VGLUT1 as well as the immediate early gene Homer1a and the post-synaptic density scaffolding protein Shank1B. Like BDNF, Arc and VGLUT1, the modulation of Homer1a by antidepressants (Dell’aversano et al. 2009) and the depressivelike phenotype induced by the genetic deletion of Homer (Jaubert et al. 2007), points to a key role of this gene in the pathophysiology of depression. Moreover, Shank1B expression was decreased in an animal model of depression (Serres et al. 2010), although no studies have investigated its modulation by antidepressant treatment. Methods and materials Animals Male Sprague–Dawley rats (220–250 g, Harlan Olac, UK) were housed at constant temperature (21¡1 xC) and humidity under a 24-h light/dark cycle (lights on 07:00 hours). All experiments were performed in accordance with the UK Scientific Procedures Act (1986) and Home Office guidelines. Drug treatment protocols Drugs were injected at a dose of 10 mg/kg i.p. and compared to saline vehicle. Experiments comprised three groups of rats (n=6 per group) as follows : (i) a single treatment with S35966, venlafaxine or saline x2 h or 16 h post-drug interval, (ii) 7-d once-daily treatment with S35966, venlafaxine or saline – 2 h or 16 h post-drug interval. (iii) 14-d once-daily treatment with S35966, venlafaxine or saline – 2 h or 16 h post-drug interval. These different time-points were selected as it was expected that the expression of certain genes (Arc and Homer1a) would be more dynamic than others (full-length BDNF, Shank1B and VGLUT1).
Antidepressants and gene expression In-situ hybridization procedure Brains were removed and snap-frozen in isopentane (x40 xC) prior to processing for in-situ hybridization as described previously (Pei et al. 2003). In brief, cryostatcut tissue sections (12 mm) were thaw-mounted onto gelatine-coated slides and pretreated using standard methods. Oligonucleotides complementary to mRNA encoding for Arc (5k-CTTGGTTGCCCATCCTCACCTGGCACCCAAGACTGGTAT-TGCTGA-3k), BDNF (5k-GGTCTCGTAGAAATATTGCTTCAGTTGGCCTTTTGA), Homer1a (5k-TGCCTTTGAGCCGTCTAGACTGATTATCCTATACACATTCCT), Shank1B (5k-GTACCACATCCTGTTCCCGATGGTTACGAATCAGTT-3k or VGLUT1 (5k-GCACTGGGCACAAGGGAAGACTTGCATGTT-3k) were 3k-tail-labelled with [35S]dATP and applied to each section in hybridization buffer (2.4r106 cpm/section for all except VGLUT1 when 1.2r106 cpm/section was applied). After overnight incubation at 35 xC, slides were washed in 1rSSC at 55 xC for 3r20 min followed by 2r60 min at room temperature. Sections were then air-dried and exposed to Biomax films (Amersham, UK) for 5–7 d at room temperature. Controls included hybridization of sections using oligonucleotides in the sense orientation and displacement with unlabelled probes. Image and data analysis The relative abundance of Arc, BDNF, VGLUT1, Homer1a and Shank1B mRNA in selected areas was determined by densitometric quantification of autoradiograms (MCID, Canada). Optical density values were calibrated to 35S tissue equivalents using 14C microscales (Amersham, UK). Densitometric values were measured from the three sections of each animal and averaged. Data are expressed as % of controls and presented as means¡S.E.M. for each treatment group. The effects of drugs were compared statistically, region by region against saline-treated controls, using one-way ANOVA followed by Dunnett’s test. Drugs S35966 [4(5)-(5-fluoroindan-2-ylmethyl)-4,5-dihydroimidazole] and venlafaxine were synthesized by Servier. All drugs were dissolved in saline. Results Regional distribution of gene expression The abundance of mRNA for each of Arc, BDNF, VGLUT1, Homer1a and Shank1B was detectable and
619
displaceable by sense controls. Each mRNA had the expected gene- and region-specific distribution pattern (Fig. 1). Effect of treatment with S35966 and venlafaxine on gene expression Autoradiograms illustrating the effect of 14-d oncedaily injection of S35966 (10 mg/kg i.p.), venlafaxine (10 mg/kg i.p.) or saline on the mRNA abundance in rat cortical regions are presented for Arc, BDNF and VGLUT1 (Fig. 1) as well as Homer1a and Shank1B (Fig. 2). Autoradiograms illustrating the effect of a single daily injection of S35966, venlafaxine or saline on the mRNA of Arc and Homer1a are also presented (Fig. 3). Quantified changes in regional brain mRNA abundance 2 h and 16 h after the last injection of 14-, 7- and 1-d treatment with S35966, venlafaxine or saline are shown for each of Arc (Fig. 4), Homer1a (Fig. 5), Shank1B (Fig. 6), BDNF (Fig. 7) and VGLUT1 (Fig. 8). Fourteen days’ treatment with S35966 and venlafaxine Compared to saline-injected controls both S35966 and venlafaxine increased the expression of each gene examined, and these effects were statistically significant (Figs 4–8). Generally speaking, abundance of each mRNA increased in more than one cortical or hippocampal region, although the increases in VGLUT1 were only statistically significant in parietal cortex (Fig. 8). For example, Arc, Homer1a and Shank1B mRNA each increased in the orbital and cingulate cortices after S35966 and venlafaxine (Figs 4–6). BDNF mRNA increased in dentate gyrus after venlafaxine (F2,17=5.484, p=0.007) as well as in CA3 of hippocampus after S35966 and venlafaxine (F2,17=9.19, p=0.02, and p=0.001, respectively) (Fig. 7). In the case of both S35966 and venlafaxine, the effects on gene expression were clearly regiondependent. Thus, significant increases in Arc, Shank1B and Homer1a mRNA were detected in cingulate and orbital cortices post-S35966 or venlafaxine (Arc 2 h, cingulate : F2,15=7.266, p=0.026 and p=0.005, respectively ; orbital : F2,15=3.892, p=0.042 and p=0.115, respectively ; Shank1B 16 h, cingulate : F2,17=2.703, p= 0.492 and p=0.043, respectively ; orbital : F2,17=3.068, p=0.344 and p=0.047, respectively ; Homer1a 16 h, cingulate : F2,15=7.129, p=0.022 and p=0.08, respectively ; orbital : F2,15=6.966, p=0.027 and p=0.008, respectively) (Figs 4–6) while there was no change in their expression in CA3 (F2,15=0.641, p=0.256) (Table 1). On the other hand, both drugs evoked a significant increase in VGLUT1 mRNA in the parietal
F. Serres et al.
620
Saline
S35966
Venlafaxine
(a) Cg Orb
Arc
(b)
Homer1a (c)
Shank1B Fig. 1. Representative autoradiograms showing distribution of Arc, Homer1a and Shank1B in anterior cortex of rats treated once daily for 14 d with saline, S35966 (10 mg/kg i.p.) or venlafaxine (10 mg/kg i.p.). (a, b) Arc and Homer1a mRNA 2 h after the last injection ; (c) Shank mRNA 16 h after the last injection. Cg, Cingulate cortex ; Orb, orbital cortex.
Saline
S35966
Venlafaxine
(a)
Ptx
DG
CA3
VGLUT1
(b)
BDNF Fig. 2. Representative autoradiograms showing distribution of VGLUT1 and BDNF mRNA in posterior cortex including hippocampus of rats treated once daily for 14 d with saline, S35966 (10 mg/kg i.p.) or venlafaxine (10 mg/kg i.p.). (a) VGLUT1 mRNA 2 h after the last injection ; (b) BDNF mRNA 16 h after the last injection. Ptx, Parietal cortex ; DG, dentate gyrus.
cortex (F32,16=4.174, p=0.048 and p=0.05, respectively) but not in any other area (Fig. 8). It is noteworthy that in the case of both S35966 and venlafaxine, the effects on gene expression were timedependent. For example, in most cortical regions examined both S35966 and venlafaxine caused a greater increase in Arc mRNA after 2 h (cingulate : F2,15=7.266, p=0.026 and p=0.005, respectively ; orbital : F2,15=3.892, p=0.042 and p=0.115, respectively ; parietal : F2,11=8.691, p=0.026 and p=0.007, respectively) but not 16 h post-drug (frontal : F2,17=2.752 ;
orbital : F2,17=1.607) (Fig. 4). On the other hand, Shank1B mRNA was significantly increased after 16 h (cingulate : F2,17=2.703, p=0.043 ; orbital : F2,17=3.068, p=0.047) after venlafaxine but not at 2 h post-drug (cingulate : F2,16=0.194, p=0.826 ; orbital : F2,17=1.070, p=0.368) (Fig. 6). Seven days’ treatment with S35966 and venlafaxine After 7 d treatment, S35966 and venlafaxine (10 mg/ kg i.p. once daily) increased the expression of Arc,
Antidepressants and gene expression Saline
S35966
621
Venlafaxine
Cg Orb Arc
Homer1a Fig. 3. Representative autoradiograms showing the distribution of Arc and Homer1a mRNA in anterior cortex of rats treated with a single injection of saline, S35966 (10 mg/kg i.p.) or venlafaxine (10 mg/kg i.p.). Measurements were made 2 h post-drug. Cg, Cingulate cortex ; Orb, orbital cortex.
Arc mRNA (% control)
2 h post-drug 300 200
S35966 VLX
*
*
* *
*
* *
*
*
7d
14 d
*
100 0 Acute
7d
14 d
Acute
Acute
7d
14 d
16 h post-drug
*
Arc mRNA (% control)
150
100
50
0 Acute
7d
14 d
Cingulate cortex
Acute
7d
14 d
Orbital cortex
Acute
7d
14 d
Parietal cortex
Fig. 4. Time-course changes of Arc mRNA abundance in the rat anterior cortex (orbital, cingulate and parietal cortices) after acute, 7-d and 14-d treatment with saline, S35966 (10 mg/kg i.p., once daily) or venlafaxine (10 mg/kg i.p., once daily). Abundance of mRNA was measured 2 h (upper panels) and 16 h (lower panels) after the last injection (* p
