Depression is a common and serious mental health problem affecting approximately 121 million people, and is estimated to be the fourth major cause of disability worldwide (WFMH 2010). It is generally accepted that the etiology of depression involves changes in the CNS monoamine system (Prange 1964; Schildkraut 1965; Coppen 1967). Accordingly, commonly used antidepressants – monoamine reuptake inhibitors – act by enhancing central biogenic amine function by targeting the monoamine transporters. However, the resultant immediate neurochemical enhancement of monoamine transmission does not match with the onset of clinical therapeutic effect, which may start even several weeks after continuous administration of the antidepressant. Therefore, the observed clinical response is often attributed to long-term compensatory effects, resulting from antidepressant treatment. These compensatory effects may include antidepressant-induced modulation of second messenger processes, transcription factors and growth factors (Morilak and Frazer 2004).
Two important suggestions can be derived from this. First, such research has led to fundamentally new hypotheses regarding the underlying pathophysiology of depression, which allow focusing on other aspects than monoamine transmission levels. For example, depression pathophysiology may be related to neuronal death or to deficiencies in specific brain regions, such as the hippocampus, and antidepressants may induce neuroprotective mechanisms or even neurogenesis to resolve such deficits (McEwen 2000; Duman et al. 2001). Second, even when considering the secondary compensatory effects derived from the prolonged changes in monoamine signaling, it does not fully explain the fact that the clinical response of antidepressants is achieved in only about two thirds of the patients. Both of these two notations imply that novel therapeutic approaches to mood disorders are needed.
Association of neuropeptides, including galanin, and classical neurotransmitters with the mechanisms of depression have been proposed already 20 years ago (Grenhoff et al. 1993; Weiss et al. 1998). The potential involvement of galanin in depression was first hypothesized (Fuxe et al. 1991) because of its co-localization with noradrenaline (NA) in the locus coeruleus (LC) and serotonin (5-HT) in dorsal raphe nucleus (DR) in the rat (Skofitsch and Jacobowitz 1985; Melander et al. 1986; Xu et al. 1998a, b). The first pharmacological proof for involvement of brain galanin in depression-like behavior was described when galanin infusion into the rat ventral tegmental area (VTA) resulted in increased immobility time in the rodent depression screen test, the forced swim test (FST) (Weiss et al. 1998). This effect was attributed to galanin mediated inhibition of dopamine signaling in nucleus accumbens, because of ‘burst’ activation of NA-ergic fibers from LC to the VTA (Weiss et al. 1998). The increased immobility time in FST test translates as ‘potentially inducing depression-like state’, and galanin-induced pro-depressive effects have since been confirmed by several groups (e.g., see ref Kuteeva et al. 2007). Similarly, it has been shown that galanin inhibits the firing rate of 5-HT neurons in the rat DR (Xu et al. 1998b) and galanin in DR inhibits 5-HT release in the rat hippocampus (Kehr et al. 2002).
The biological activity of galanin is mediated trough interactions with three different G-protein coupled receptors, galanin receptor type 1–galanin receptor type 3 (GalR1–GalR3), which are all associated with different signaling pathways. Previously, it has been shown that activation of GalR2, in contrast to GalR1 and GalR3, significantly increases 5-HT release in DR (Mazarati et al. 2005) and leads to attenuation of depression-like behavior, while activation of GalR1 and GalR3 increases it (Lu et al. 2005; Swanson et al. 2005; Barr et al. 2006; Kuteeva et al. 2008b). Taken together, all this suggests that GalR2 subtype specific ligands could potentially be used as drug candidates for treating mood disorders.
This study has been initiated by the fact that peptides, because of high receptor specificity, potency, and low toxicity, are known to have great therapeutic potential, but at the same time, their physicochemical properties strongly restrict their use as psychopharmacological tools. In addition, there is a lack of systemically active GalR2 subtype specific ligands in the galanin field. Here, we have designed an approach to increase in vivo usability of peptide ligands and we hereby demonstrate a series of systemically active novel GalR ligands with modest preferential binding toward GalR2, including more in-depth characterization of one of these peptide ligands. Furthermore, we show in side by side comparison with a common clinically used antidepressant medication imipramine their ability to attenuate depression-like behavior.
- Top of page
- Materials and methods
- Supporting Information
In this study, we present a series of novel systemically active GalR2 subtype preferring ligands J17-J21 (Table 1). The design of the ligands is based on two previously published GalR2 agonists M1145 and M1153 (Runesson et al. 2009; Saar et al. 2011). The N-terminal part of the novel ligands was elongated with the same four amino acid residues (Arg-Gly-Arg-Gly) from the galanin-like peptide (GALP) as was done in M1145 and M1153 (Ohtaki et al. 1999; Runesson et al. 2009; Saar et al. 2011). To disturb the ligand binding to GalR1 and GalR3, the N-terminal Gly in position 5 was replaced with Asn (Sollenberg et al. 2006; Runesson et al. 2009; Saar et al. 2011).
Blood–brain barrier (BBB) is known to act as a major obstacle to deliver macromolecular entities, such as peptides, to CNS. In addition, short half-lives and large volumes of distribution in the blood of the peptides are a concern. Therefore, two commonly known strategies were used (Kokaia et al. 2001) to improve the pharmacokinetic properties (i.e. stability to degradation) and the CNS activity of the ligands. First, the ligands were elongated with additional positively charged amino acids (cationization). Second, a stearic acid was coupled to the side chain of Lys in the C-terminal part of ligands (lipidization). Both cationization and lipidization have been shown to enhance the bioavailability of the peptide in the brain (Bulaj et al. 2008). In addition, to reduce enzymatic degradation and further increase the bioavailability, three C-terminally positioned l-lysines were substituted to d-lysines in the peptide J18.
All these modifications resulted in a slight loss of subtype specificity for GalR2 and in a decrease in binding affinity when compared to previously published GalR2 agonists M1145 and M1153 (Table 1). However, in the displacement studies, all the tested peptides still exhibited a reasonable preferential binding toward GalR2. The best subtype specificity was demonstrated by the peptide J18, with approximately six to seven-fold selectivity toward GalR2, compared to GalR1 and GalR3. Importantly, all peptides showed nanomolar binding affinities toward GalR2 ranging from 4.9 nM to 83 nM. The lowest Ki values to GalR2 were exhibited by J19 and J21, which could be because of substitution of the C-terminally positioned l-lysines with ornithine. In addition, two l-arginines were substituted with ornithines in the position 2 and 4 of the peptide J21. This modification led to a two fold increase in Ki values when peptides J19 and J20 were compared, probably because of the smaller molecular size of ornithine that might have disrupted, to a lesser degree, the binding of the ligand to GalR1-GalR3. At the same time, in the peptide J20 the substitution of two l-lysines with aminohexanoyls in the positions 18 and 19 increased the binding toward GalR2. This suggests that moving the fatty acid further away from the original peptide backbone decreases its ability to disturb the binding toward GalRs.
Nevertheless, and as discussed subsequently, somewhat decreased binding affinity and lower receptor subtype selectivity of the peptides may not indicate anything about its usefulness in vivo, because increased stability may compensate for lower affinity. The main focus of this study was to develop in vivo usable, systemically active peptide-based ligands, and to confirm the utility of these in a field which notably lacks usable compounds, but has profound clinical potential. The rationale for concentrating on galanin system was based on previous studies, suggesting that GalR2 agonists should exhibit antidepressive effect (Kuteeva et al. 2008b). Therefore, we tested the behavior of the novel galanin receptor ligands in vivo.
First, we compared the biodistribution of J18 with a previously published GalR2 subtype specific ligand M1145, which does not contain stearoyl and d-amino acids (Fig. 8). As hypothesized, the relative distribution of the peptides was notably different. The differences arise from two aspects. First, J18 clearly showed much higher half-life, which reflects its higher stability to degradation. Figure 8 shows that the levels of J18 were higher in all organs 1 h and 3 h post-injection. At 24 h post-injection, J18 had been cleared out from the organism, while M1145 had been largely eliminated already at 3 h. Second, the levels of J18 in the brain were 2.5 fold higher than that of M1145. These results suggest that the combination of increased stability and lipofilicity increased the ability of J18 to penetrate the BBB.
We also demonstrate the antidepressant activity of the peptides by performing FST. The first notable thing is that we are able to see a CNS-derived behavioral effect upon systemic administration of a (modified) peptide, which, according to classical pharmacology, would not be able to reach the brain. We showed that i.v. administration of several peptides reduced the immobility time in FST. These results are in concordance with a previous publication demonstrating that GalR2(R3) agonist AR-M1896 decreased immobility, while GalR2 antagonist M871 increased immobility time in the FST (Kuteeva et al. 2008b) – the effects that have been observed upon i.c.v. administration of the drugs. The highest effect in this study was obtained by the peptide J18 at the concentration of 0.5 mg/kg, comparable to the effect presented by the clinically relevant antidepressant medication imipramine. Remarkably, the effective doses of J18 of 0.5 and 0.25 mg/kg are equal to 192 and 96 nmol/kg, which reflect strikingly high potency of the peptides.
The U-shaped dose–response relationship peaking at 0.5 mg/kg shown by J18 could possibly be explained by increased binding to GalR1 and GalR3 at higher concentrations, an assumption, which is supported by the binding profile (see Table 1) of the J-peptides, mostly showing modest specificity toward GalR2. Previous research has shown that the activation of GalR1 and GalR3 can increase the pro-depressive effect of galanin (Kuteeva et al. 2008b). Accordingly, the activation of GalR1 and GalR3 caused by the higher concentrations of the ligand could possibly counterattack the effect mediated through GalR2, the one that is known to lead to the attenuation of depression-like behavior (Kuteeva et al. 2008a). Further studies with lipidated peptides with higher receptor subtype selectivity might confirm this hypothesis; if this is the case, higher selectivity toward GalR2 should exhibit more linear dose–response curve in depression animal models.
Nevertheless, it should be noted that high receptor subtype selectivity in itself may not always yield the best-behaving pharmaceutical compound. For example, the same antidepressant drug used in this article, imipramine, has different binding sites and directly modulates several neurotransmitter levels in the brain (including serotonin, noradrenalin, and dopamine). Therefore, simultaneous binding to several receptor subtypes may contribute to the unique behavioral effect (or clinical effect) of the compound (in this case, antidepressive effect).
We also used another depression model (tail suspension test) to further confirm systemic activity of the novel peptides (Fig. 3). As expected, similar to the effect in FST, J18 exhibited potent systemic activity, with slight shift in the dose curve, the dose of 0.25 mg/kg being most efficient in this test paradigm. In addition, we show that the antidepressive effect of J18 is retained also via i.p. administration route, which is important for animal studies, as it permits easy, stress-free handling, and experimental setup with chronic administration. Antidepressive drug screening animal models utilize acute administration of a drug (two injections for rats and one injection for mice in TST or FST). However, it is generally known that chronic administration for 2–4 weeks is required for SSRIs to see antidepressive effect in clinics. Therefore, we investigated if it is possible to use J18 in other depression models which use repeated administration of a drug, as theoretically, different feedback mechanisms or receptor desensitization may decrease the pharmacological potency of the compound. Chronic administration of J18 (0.25 mg/kg i.p. for 17 days) still retained its antidepressant-like effect in TST (Fig. 5), suggesting that further characterization of J18 and other currently presented peptides in different models is possible.
Considering that the behaviors observed and quantified in mouse depression paradigms are purely motor (differentiation between locomotion vs. immobility), and not necessarily because of central effects, we confirmed, using pharmacological tools, that the effect of the novel peptides is really CNS-derived, and mediated by GalR2. Accordingly, i.c.v. infusion of a known galanin receptor antagonist blocked the antidepressive effect of systemically administered J18 (Fig. 6), suggesting that indeed, brain galanin receptors are activated by J18. Moreover, as J18 failed to elicit antidepressant-like effect in GalR2KO animals supports the hypothesis that J18 acts through GalR2. The mechanism by which GalR2 stimulation leads to antidepressant-like effect may be mediated by interactions between galanin and 5-HT/DR and NA/LC – the same CNS areas that mediate pro-depressive effects through GalR1 activation (Weiss et al. 1998). It is hypothesized that galanin is released under stressful conditions (Lundberg and Hokfelt 1986), creating a tonic galanin transmission. Accordingly, it has been shown that activation of GalR1 and R2 mediate opposing effects, where i.c.v. galanin, GalR1 agonist, or GalR2 antagonist induced pro-depressive behavior, and GalR2(R3) agonist induced antidepressive behavior (Kuteeva et al. 2008b). However, antidepressant treatment may augment the GalR2 tone (Kuteeva et al. 2008b), by increasing galanin levels and GalR2 binding sites (Lu et al. 2005), leading to relative shift from normally major pro-depressive GalR1 to now dominating antidepressive GalR2. It is known that GalR1 mainly activates Gi/o types of G-proteins, mediating inhibition via adenylate cyclase (Habertortoli et al. 1994) and inhibition of monoamine system in LC and DR (Kuteeva et al. 2008a). In contrast, the GalR2 subtype may transmit stimulatory effects of galanin, activating Gq/11 types of G-proteins (Wang et al. 1998) and thus may mediate antidepressive effects through LC (Kuteeva et al. 2008b) and DR (Lu et al. 2005).
In our study, i.c.v infusion of non-specific galanin antagonist M35 blocked the antidepressive activity of GalR2 agonist J18, but did not, by itself, reduce immobility in a rodent depression test, as has been shown previously for M35 (Kuteeva et al. 2007), or for other non-specific antagonists (for example, see ref Weiss et al. 1998). However, there are other works that have, similar to the current one, not observed this effect (for example Lu et al. 2005). One possible explanation that may explain this is species differences between mouse and rat (Larm et al. 2003).
Taken together, we demonstrate a set of chemical modifications that alter peptide pharmacokinetics and biodistribution, increasing its stability to degradation and penetration to brain. Hence, we characterize a series of novel systemically active GalR2 agonists and demonstrate their potent antidepressant effect at notably low doses in several common rodent screening models of depression-like behavior.