|Home | About | Journals | Submit | Contact Us | Français|
Neurotrophins, the cognate ligands for the Trk receptors, are homodimers and induce Trk dimerization through a symmetric bivalent mechanism. We report here that amitriptyline, an antidepressant drug, directly binds TrkA and TrkB and triggers their dimerization and activation. Amitriptyline, but not any other tricyclic or SSRI antidepressants, promotes TrkA autophosphorylation in primary neurons and induces neurite outgrowth in PC12 cells. Amitriptyline binds the extracellular domain of both TrkA and TrkB and promotes TrkA-TrkB receptor heterodimerization. Truncation of amitriptyline binding motif on TrkA abrogates the receptor dimerization by amitriptyline. Administration of amitriptyline to mice activates both receptors and significantly reduces kainic acid-triggered neuronal cell death. Inhibition of TrkA, but not TrkB, abolishes amitriptyline's neuroprotective effect without impairing its antidepressant activity. Thus, amitriptyline acts as a TrkA and TrkB agonist, and possesses marked neurotrophic activity.
Neurotrophins, which include NGF (nerve growth factor), BDNF (brain-derived neurotrophic factor), NT-3 and NT- 4/5, are critical for the development and maintenance of the peripheral and the central nervous systems. Neurotrophins exert their physiological actions through two classes of receptor: Trk tyrosine kinase receptors and p75NTR. The Trk receptors are members of a transmembrane tyrosine kinases family (TrkA, TrkB and TrkC). NGF primarily binds TrkA, BDNF and NT-4/5 mainly interact with TrkB, and NT-3 predominantly associates with TrkC, while having weaker interactions with TrkA and TrkB receptors. Docking of neurotrophic factors on Trk receptors initiates receptor homodimerization, autophosphorylation of cytoplasmic tyrosine residues on the receptors, and a cascade of cell signaling events including Ras/Raf/MAPK, PI 3-kinase/Akt and PLC-γ1 (Kaplan and Stephens, 1994). These signals prevent apoptotic cell death, promote cellular differentiation and axon elongation, and up-regulate choline acetyl transferase (ChAT). Several neuronal cell types that are lost in certain pathologies express TrkA and respond to NGF, and NGF has been reported to reverse basal forebrain cholinergic atrophy and reduce cognitive decline and stimulate cholinergic fiber growth in humans with mild Alzheimer's disease (Mufson et al., 2008) and ameliorate peripheral diabetic neuropathies (Apfel, 2002). Other applications proposed for NGF include treatment of neuronal damage (Hughes et al., 1997) and targeting of neuroectoderm-derived tumors (Cortazzo et al., 1996; LeSauteur et al., 1995). In addition, neurotrophins such as BDNF act as key regulators of neurite outgrowth and synaptic plasticity, and it has been proposed that a deficit of these molecules may underlie the psychopathology of stress and depression, while antidepressants may act via neurotrophins to produce the molecular and behavioral changes associated with their efficacy (Altar et al., 1997; Duman et al., 1997). Thus, both NGF and BDNF might contribute to antidepressant-induced formation and stabilization of synaptic connectivity.
Nonetheless, the use of NGF for the treatment of these diseases has been limited by its poor pharmacological properties, such as the low blood–brain barrier permeability (Poduslo and Curran, 1996) and relevant side effects, like hyperalgesia (Apfel, 2002). To search for NGF and BDNF mimetics with better pharmacokinetic properties, tremendous efforts have been made to design small, proteolytically stable molecules with neurotrophic activity and specificity for TrkA or TrkB. For example, ligand mimicry and antibody mimicry strategies have been intensively explored to generate peptide analogs of two agonists directed to the extracellular domain (ECD) of TrkA or TrkB (Beglova et al., 2000; O'Leary and Hughes, 2003; Saragovi et al., 1991; Xie et al., 2000). Additionally, an NGF-mimetic peptide has recently been reported to partially mimic NGF and reduce neuropathic pain in rat (Colangelo et al., 2008). However, none of these antibodies and peptidomimetics can fully mimic NGF or BDNF function in animals.
In this report, we show that tricyclic antidepressant amitriptyline, which is traditionally thought to exert its therapeutic effects via blockade of the serotonin and noradrenaline transporters, interacts directly with both TrkA and TrkB receptors on the ECD (extracellular domain). Further, our data reveal that amitriptyline induces TrkA and TrkB homo- and heterodimerization and activation in mouse brain, but that the heterodimerization is not required for Trk receptor activation. Truncation of amitriptyline binding motif on TrkA but not the corresponding region on TrkB abolishes the receptor homo- and hetero-dimerization. Moreover, amitriptyline suppresses neuronal apoptosis elicited by kainic acid in a TrkA-dependent manner. Hence, amitriptyline acts as a TrkA and TrkB receptor agonist and possesses marked neurotrophic activity.
Recently, we developed a cell-based screening approach for novel TrkA agonist discovery, and reported that gambogic amide is a selective agonist for TrkA that possesses robust neurotrophic activity and prevents neuronal cell death (Jang et al., 2007). During the screening, we also surprisingly identified numerous tricyclic antidepressant compounds that selectively protected TrkA expressing T17 cells but not parental SN56 cells lacking TrkA from apoptosis (Figure 1A & 1B). Titration assays with T17 cells for apoptosis inhibitory activity revealed that EC50 values were 8, 30 and 50 nM for clomipramine, imipramine and amitriptyline, respectively (Figure 1C). TrkA and p75NTR are up-regulated in hippocampal and cortical neurons under pathophysiological conditions (Kokaia et al., 1998; Lee et al., 1998). Moreover, neuroprotective effects of NGF in hippocampal and cortical neurons have been demonstrated in vitro and in vivo (Culmsee et al., 1999; Zhang et al., 1993). Therefore, to test whether the tricyclic compounds can also protect primary hippocampal neurons from apoptosis, we pretreated primary cultures with test compounds (0.5 μM each) for 30 min, followed by glutamate treatment. NGF, gambogic amide, or amitriptyline pretreatment significantly protected hippocampal neurons from apoptosis, while other tricyclic drugs tested had no effect (Figure 1D and data not shown).
NGF reduces cortical infarction and apoptosis in transgenic mice and protects PC12 cells from apoptosis in OGD (Oxygene-Glucose-Deprivation) (Beck et al., 1992; Guegan et al., 1998). To explore whether amitriptyline and/or other tricyclics could protect hippocampal neurons from OGD-provoked apoptosis, we pretreated primary cultures with various tricyclic drugs, followed by OGD stimulation for 3 h. Amitriptyline significantly suppressed apoptosis, whereas neither imipramine nor clomipramine exhibited any protective activity (Figure 1E, left panel). Titration assays showed that amitriptyline repressed neuronal apoptosis in a dose-dependent manner (Figure 1E, right panel). Thus, amitriptyline but not any other tricyclic anti-depressant drugs selectively protects hippocampal neurons from apoptosis.
NGF binds TrkA and elicits its autophosphorylation and downstream MAP kinase and PI 3-kinase/Akt pathways activation in primary hippocampal and cortical cultures that express demonstrable TrkA (Culmsee et al., 2002; Kume et al., 2000). To explore whether amitriptyline could stimulate TrkA, we treated hippocampal neurons with 0.5 μM amitriptyline or other tricyclic drugs for 30 min. Immunofluorescent staining showed that amitriptyline, like NGF, triggered TrkA tyrosine phosphorylation, whereas other tricyclic compounds did not (Figure 2A). Both Akt and Erk 1/2 were markedly activated in NGF- or amitriptyline-treated hippocampal neurons. In contrast, none of the other tricyclic drugs was capable of simultaneously activating Akt and Erk 1/2 (Figure 2B). It was worth noting that amitriptyline induced TrkA phosphorylation on both tyrosine Y751 and Y794. Surprisingly, Y490 was not phosphorylated at all. In contrast, NGF and gambogic amide activated all three tyrosine residues on TrkA receptor. Although trimipramine induced TrkA phosphorylation on Y794, it failed to induce phosphorylation on either Y490 or Y751 residue (Figure 2B). K252a is an inhibitor of the Trk receptors. K252a potently blocked amitriptyline-triggered TrkA tyrosine phosphorylation, indicating that the stimulatory effect by amitriptyline represents Trk receptor-dependent autophosphorylation. Strikingly, amitriptyline, but not NGF, also induced TrkB tyrosine phosphorylation, which was also blocked by K252a (Figure 2C). However, amitriptyline failed to provoke TrkC activation (Supplemental Figure 1). Amitriptyline swiftly activated both MAPK and Akt signaling cascades in hippocampal neurons in a manner temporally similar to NGF (Figure 2D, left panels). Titration assays demonstrated that 250 nM amitriptyline stimulated both Erk 1/2 and Akt signalings activation and the signal became stronger at 500 nM (Figure 2D, right panels). Pretreatment with anti-NGF or anti-BDNF failed to block the stimulatory effect of TrkA or TrkB by amitriptyline in cortical neurons, suggesting that amitriptyline provokes TrkA and TrkB activation independent of neurotrophins (Supplemental Figure 2). Together, these results demonstrate that amitriptyline strongly induces TrkA and TrkB receptor phosphorylation and activation in a dose-dependent manner.
One of the most prominent neurotrophic effects of NGF is to trigger neurite outgrowth in neuronal cells and incur differentiation. To assess whether amitriptyline and/or other antidepressant drugs possess this activity, we incubated PC12 cells with NGF or various antidepressant drugs (0.5 μM) for 5 days. As expected, NGF induced pronounced neurite sprouting in PC12 cells after 5 days of treatment. Among all tested tricyclic antidepressant drugs, only amitriptyline markedly triggered demonstrable neurite outgrowth in PC12 cells, and the neurite network generated was comparable to that initiated by NGF (Figure 3A). Titration assays revealed that 100 nM of amitriptyline was sufficient to provoke substantial neurite sprouting in PC12 cells (Figure 3B). Because PI 3-kinase and MAPK signaling pathways are required for NGF's neurite outgrowth effect, we performed experiments using pharmacological agents, including K252a, PI 3-kinase inhibitors and PD98058 (MEK1 inhibitor), which all substantially blocked the neurite outgrowth effect by NGF (Figure 3C). These inhibitors also inhibited amitriptyline's neurotrophic activity, indicating that amitriptyline's neurite outgrowth effect is Trk receptor dependent. Thus, amitriptyline possesses notable neurotrophic activity and robustly induces neurite outgrowth.
To determine whether Trk receptors directly bind amitriptyline, we conducted an in vitro binding assay with purified Trk ECD and ICD proteins and [3H]-amitriptyline. Remarkably, the ECD but not ICD from TrkA and TrkB receptors bound to amitriptyline, with TrkA exhibiting stronger binding (TrkA binding constant, 3 μM, TrkB, 14 μM) (Figure 4A). Gradual increase of cold amitriptyline concentration progressively diminished [3H] amitriptyline binding to the ECD of TrkA and TrkB receptors, indicating specific binding (Figure 4B). By contrast, amitriptyline was unable to compete with NGF or BDNF for binding to TrkA or TrkB ECD, respectively (data not shown). This might be due to its low affinity to the receptors. To map which region is involved in TrkA binding to amitriptyline, we systematically deleted Ig1, Ig2, Ig1+2 domains in TrkA receptor and found that none of the immunoglobulin-like domains were required for the interaction with amitriptyline, while removal of the entire ECD domain significantly diminished the binding by amitriptyline (Figure 4C). Truncation assays with a variety of TrkA ECD proteins showed that the 1st leucine-rich motif (a.a. 72−97) was essential for full binding of amitriptyline to TrkA (Figure 4D, lower panel). Sequence alignment for the 1st LRM (a.a. 72−97) of Trk receptors is shown (Figure 4D, upper panel).
We next sought to determine whether the binding by amitriptyline to Trk receptors incurs their dimerization. Coimmunoprecipitation demonstrated that amitriptyline, but not imipramine, elicited TrkA homodimerization to as NGF. Strikingly, amitriptyline uniquely caused TrkA to heterodimerize with TrkB (Figure 4E, top panel). Amitriptyline elicited marked tyrosine phosphorylation in TrkA and TrkB, but not in TrkC receptor when HEK293 cells were individually transfected by TrkA, TrkB or TrkC (Figure 4F). By contrast, TrkA-KD (kinase-dead) was not tyrosine phosphorylated, indicating that tyrosine phosphorylation of Trk receptors provoked by amitriptyline is exerted by the receptors themselves but not by any other cytoplasmic tyrosine kinases. Hence, amitriptyline mimics NGF and induces TrkA dimerization and autophosphorylation, and also possesses novel TrkA-TrkB heterodimerazation activity.
To assess whether amitriptyline can induce Trk receptor activation in mouse brain, we injected amitriptyline (15 mg/kg, i.p.) into mice. Amitriptyline induced both TrkA (Y794) and TrkB (Y816) activation in mouse brain after 4 h and the effect persisted at 8 h (Figure 5A, left top and 3rd panels). By contrast, imipramine failed to activate any of the receptors (Figure 5A, right panels). Consequently, we also observed robust Akt and ERK1/2 activation with the same time course, supporting that amitriptyline can provoke both TrkA and TrkB receptor activation in mice. Although imipramine failed to induce phosphorylation of the Trk receptors, it did promote phosphorylation of Akt and ERK1/2 (Figure 5A, right panels). The onset of Trk receptor activation fits with the peak plasma concentration of amitriptyline that is reached within 6 h. RTPCR analysis showed that neither TrkA nor TrkB mRNA levels were altered after treatment by amitriptyline or imipramine (Figure 5B), indicating that acute treatment with amitriptyline or imipramine might not significantly regulate Trk receptor transcription.
To explore whether amitriptyline can block the excitotoxicity initiated by Kainic acid (KA), we injected the mice with saline or amitriptyline (15 mg/kg, i.p.), followed by saline or KA (25 mg/kg, i.p.). Five days following treatment, TUNEL staining revealed significant apoptosis in the hippocampus of KA-treated mice compared to vehicle control or amitriptyline alone. KA-provoked apoptosis was substantially diminished by pretreatment with amitriptyline, while administration of amitriptyline after KA was not as effective (Figure 5C). Quantitative analysis revealed that amitriptyline suppressed KA-induced cell death by 70% when injected prior to KA, compared to 45% when injected after (Figure 5C, right panel). Immunohistochemical analysis showed that TrkA and TrkB were significantly activated in the hippocampus by amitriptyline. In addition, amitriptyline evidently augmented the expression of TrkA but not TrkB (Figure 5D). Hence, amitriptyline strongly provokes TrkA and TrkB activation in mouse brain and protects hippocampal neurons from KA-triggered cell death.
To test whether amitriptyline regulates Trk receptors expression, we fed the mice with various anti-depressant drugs sub-chronically and chronically. After 5 days or 28 days, we monitored TrkA and TrkB protein expression levels and activation by immunoblotting. In 5 days, amitriptyline prominently provoked TrkA tyrosine phosphorylation, whereas imipramine, fluoxetine and control vehicle had no effect (Figure 6A, left top panel). TrkA expression levels in amitriptyline-treated mice were higher than other drugs or saline-treated mice (left 2nd panel). Remarkably, TrkB was activated by all three drugs compared to saline, and total TrkB protein levels were higher in imipramine and fluoxetine-treated mice than amitriptyline and saline-treated mice (Figure 6A, 3rd and 4th panels). Compared to vehicle control and other antidepressant drugs, RT-PCR analysis revealed that TrkA transcription was substantially augmented after amitriptyline. Nevertheless, TrkB mRNA levels remained similar in all samples (Figure 6A, right panels), indicating that sub-chronic treatment with antidepressant drugs does not significantly affect TrkB receptor transcription. We made the similar observation in 28 days-treated samples (data not shown). The comparable effects were recapitulated in 5-HT1a-null mice, suggesting that serotonin receptor 5-HT1a is not implicated in these events (supplemental Figure 3). Thus, these findings support the idea that amitriptyline activates TrkA receptor and upregulates its expression in both sub-chronic and chronic treatment conditions. However, imipramine and fluoxetine are more potent than amitriptyline in provoking TrkB activation and elevating TrkB protein levels.
Coimmunoprecipitation revealed that amitriptyline strongly provoked endogenous TrkA to bind endogenous TrkB receptor in the brain after sub-chronic treatment. By contrast, imipramine or fluoxetine failed (Figure 6B). Cotransfection and pull-down assays demonstrated that amitriptyline induced TrkA and TrkB homodimerization like NGF and BDNF, but also triggered heterodimerization between TrkA and TrkB receptors in HEK293 cells. In contrast, imipramine and fluoxetine did not induce either of these effects (Figure 6C). To explore the role of amitriptyline binding motifs in mediating TrkA and TrkB receptors hetero-dimerization, we deleted the binding motif and transfected the truncated receptors into HEK293 cells. Deletion of amitriptyline binding motif on TrkA completely abolished the homo- and hetero- dimerization activity (Figure 6D, left panels). By contrast, truncation of the similar region on TrkB failed to abolish the stimulatory activity by amitriptyline (Figure 6D, right panels), indicating that the deleted fragment on TrkB might not be the binding spot for amitriptyline. Notably, the cotransfected wild-type TrkA and TrkB were potently phosphorylated by amitriptyline. Interestingly, truncated TrkB but not deleted TrkA was still capable to be activated by amitriptyline. By contrast, truncated TrkA and TrkB were robustly activated by NGF or BDNF, respectively (Figure 6D), suggesting that amitriptyline binding motif is not required for neurotrophins to trigger Trk receptor activation.
Previous studies show that TrkA immunoglobulin-like ligand binding domains inhibit spontaneous dimerization and activation of the receptor (Arevalo et al., 2000). To explore whether amitriptyline elicits receptor dimerization via blocking the autoinhibitory effect by the Trk receptors, we transfected the truncated receptors with different tags into HEK293 cells. Cotransfection of truncated TrkA failed to form a homodimer regardless of amitriptyline (Supplemental Figure 4A), supporting that amitriptyline induces TrkA dimerization not through blocking its autoinhibitory effect by the binding motif on TrkA. In contrast, amitriptyline elicited potent homodimer by the truncated TrkB receptors as BDNF, though TrkB activation by amitriptyline was evidently reduced as compared to BDNF. Ligand binding assays demonstrated that deletion of 72−97 residues abolished the binding activity by amitriptyline on TrkA but on TrkB (Supplemental Figure 4B & C). This finding might explain why the truncated TrkB is still able to form a homodimer by amitriptyline. Hence, Trk receptor binding by amitriptyline is essential for the receptor homo- and hetero- dimerization.
To assess whether amitriptyline's neurotrophic activity is solely mediated through TrkA receptor, we explored its ability to promote neuronal survival in wild-type neurons and those that lack TrkA. Heterozygous (TrkA +/−) males and females were crossed, and hippocampal neurons were harvested from newborn (P0) pups. Amitriptyline triggered TrkA tyrosine phosphorylation in wild-type but not TrkA knockout neurons (Figure 7A, left 2nd panel). Fluoxetine also elicited weak tyrosine phosphorylation in both TrkA and TrkB receptor in a TrkA-dependent manner (Figure 7A, right 2nd and 4th panels). Interestingly, these experiments additionally revealed a different effect of amitriptyline on TrkB in the absence of TrkA. In contrast to TrkA, amitriptyline activated TrkB in both wild-type and TrkA −/− neurons, indicating that TrkA is not required for amitriptyline-induced phosphorylation of TrkB, and amitriptyline activates TrkB in a TrkA-independent way.
We next assessed the apoptotic pathway by examining caspase-3 activation at baseline and following glutamate exposure. In the absence of glutamate treatment, a low amount of active caspase-3 was demonstrable in TrkA-null neurons but absent in wild-type neurons, consistent with the substantial neuronal cell death that has been observed in TrkA-null mice (Smeyne et al., 1994). Amitriptyline completely inhibited the basal caspase-3 activation in TrkA −/− neurons (Figure 7A, left 5th panel). Caspase-3 activation by glutamate was moderately enhanced in TrkA −/− neurons, suggesting that TrkA normally inhibits the apoptotic response to excitotoxicity. Amitriptyline pretreatment partially blocked glutamate-induced caspase-3 activation in both wild-type and TrkA −/− neurons, suggesting that activation of TrkB by amitriptyline in TrkA −/−neurons contributes to this protective action. Neither fluoxetine nor imipramine inhibited caspase-3 activation in wild-type or TrkA −/− neurons (right 5th panel). To determine whether activation of TrkA by amitriptyline requires TrkB, we extended our analysis to TrkB −/− neurons. Amitriptyline, but not imipramine, selectively activated TrkA in both wild-type and TrkB −/−neurons (Figure 7B, 3rd panel). Depletion of TrkB incurred spontaneous caspase-3 activation, which was completely blocked by amitriptyline but not imipramine. Moreover, amitriptyline also markedly suppressed glutamate-provoked caspase-3 activation in both wild-type and TrkB −/−neurons, whereas imipramine had no protective effect (Figure 7B, bottom panel). These results indicate that the activation of TrkA by amitriptyline is independent of TrkB, and TrkA activation in TrkB −/− neurons might account for the protective effect of amitriptyline pretreatment.
TrkA F592A knock-in mice can be selectively blocked by the inhibitor 1NMPP1, which results in Trk-null phenotypes (Chen et al., 2005). Consistent with previous reports, NGF-provoked TrkA, Akt, and Erk 1/2 phosphorylation was selectively blocked by 1NMPP1 but not K252a, a pattern that was mimicked by amitriptyline treatment (Figure 7C). Imipramine and fluoxetine failed to activate TrkA but did stimulate Akt and Erk 1/2, although these activities were not blocked by 1NMPP1 or K252a, indicating that these antidepressants were acting via a TrkA-independent pathway (Figure 7C, 3rd and 5th panels).
We next tested whether our in vitro results could be recapitulated in vivo by assessing KA-induced caspase-3 activation in TrkA F592A mice. As expected, 1NMPP1, amitriptyline alone or 1NMPP1 + amitriptyline combined treatment had no effect on caspase-3 in TrkA F592A mice. KA provoked significant caspase-3 activation that was suppressed by amitriptyline, and this protective effect was abolished by 1NMPP1 pretreatment (Figure 7D, top panel). 1NMPP1 also blocked TrkA F592A phosphorylation by amitriptyline, and imipramine did not activate TrkA under any condition (Figure 7D, middle panel). KA-provoked caspase-3 activation in 1NMPP1-sensitive TrkB F616A knock-in mice was also inhibited by amitriptyline. However, in contrast to the TrkA F592A mice, this protective effect persisted in the presence of 1NMPP1 (Figure 7E). Although imipramine weakly activated TrkB F616A, it failed to suppress caspase-3 activation by KA (Figure 7E). These results demonstrate that amitriptyline selectively activates both TrkA and TrkB receptors in mice, but TrkA is more important than TrkB in mediating the neuronal survival functions by amitriptyline.
In this study, we show that amitriptyline binds to a motif in the 1st LRM in the TrkA receptor ECD with a Kd of ~ 3 μM. Interestingly, amitriptyline also interacts with the ECD from TrkB but not TrkC with a decreased binding affinity (Kd ~ 14 μM). Previous study suggests that brain [amitriptyline] in treating depression is around 5−7 μm (Glotzbach and Preskorn, 1982). To treat pain associated peripheral neuropathy, [amitriptyline] is 2-fold lower (Esser and Sawynok, 1999). Hence, amitriptyline affinity to TrkA might be sufficient to exert its biological actions. The known Kd for NGF to Trk A and BDNF to TrkB is ~1 nM in the absence of p75NTR. These differences in binding affinities are consistent with the quantitative dose-response studies that amitriptyline is some 100-fold less potent than the neurotrophin ligands (Figure 3).
NGF is a homodimer and it binds to both the Ig2 domain and LRM domain on TrkA and brings two TrkA receptors together, which subsequently autophosphorylate each other. NGF selectively binds the 24 amino acids (a.a 97−120) of the 2nd LRM with Kd ~ 1.3 nM (O'Connell et al., 2000). Here we show that amitriptyline can only bind the 1st LRM domain (a.a. 72−97) and promotes TrkA receptor dimerization. In addition, we also show that amitriptyline binds both TrkA and TrkB receptors and provokes TrkA to interact with TrkB, leading to a heterodimer. However, NGF fails to do so (Figure 4). Previous studies show that BDNF, NT-3, and NT-4 bind to the LRM3 cassette of TrkB, whereas NGF does not. These binding characteristics clearly reflect in vivo specificities. A more precise mapping of the region(s) responsible for binding BDNF, NT-3, and NT-4 identified the 2nd LRM of TrkB as a functional unit capable of binding all three neurotrophins (Windisch et al., 1995). Why does amitriptyline but not NGF or BDNF provoke the heterodimer formation by TrkA and TrkB? NGF and BDNF bind p75NTR, while amitriptyline does not (data not shown), suggesting that these two categories of molecules bind neurotrophin receptors in fundamentally different ways.
Antidepressant drugs and electroconvulsive stimuli significantly influence brain neurotrophins concentrations. It has been proposed that BDNF and NGF may play a role in depression (Angelucci et al., 2000). Subchronic treatment with lithium increases NGF content in brain of adult rat, supporting that NGF may be implicated in the mechanism of antibipolar treatments (Hellweg et al., 2002). To test whether amitriptyline exerts its antidepressant action through the TrkA receptor, we conducted forced swim tests with TrkA F592A mice. Blocking TrkA signaling with 1NMPP1 did not significantly alter the immobility. Sub-chronic treatment of mice with amitriptyline or imipramine substantially decreased the immobility no matter whether TrkA receptor was blocked or not. Moreover, both drugs exhibited demonstrable antidepressant effect in NGF +/+ and +/− mice (Supplemental Figure 5). Thus, TrkA signaling might be dispensable for at least some of the therapeutic actions of amitriptyline. This finding fits with previous reports that amitriptyline increased BDNF but not NGF concentration in serum of depressed patients (Hellweg et al., 2008). Moreover, normal TrkB signaling is required for the behavioral effects typically produced by antidepressants (Saarelainen et al., 2003). Conceivably, amitriptyline exerts its antidepressant action through TrkB but not TrkA receptor.
While tricyclics are generally regarded as comparable antidepressants, why is amitriptyline uniquely effective versus the other tricyclics? Imipramine differs chemically from amitriptyline only in the presence of an exocyclic double bond in amitriptyline, which is absent in imipramine (Figure 1A). The exocyclic double bond inhibits the free rotation of the side chain of amitriptyline, rendering it slightly more “rigid”. Thus, this chemical feature may account for the selective and potent Trk agonistic activity of amitriptyline. Consistently, it has been shown before that amitriptyline but not imipramine possesses robust anticholinergic activity (Snyder and Yamamura, 1977). Tricyclic antidepressants like amitriptyline inhibit the monoamine reuptake inactivation of norepinephrine and serotonin neurotransmitters. They are potent blockers of muscarinic cholinergic (Snyder and Yamamura, 1977), alpha adrenergic and 5-HT receptors (Peroutka and Snyder, 1980). To explore whether serotonin is implicated in amitriptyline's action, we depleted serotonin with pCPA and found that amitriptyline's stimulatory effect on TrkA was not significantly affected. Moreover, serotonin itself was unable to provoke TrkA or TrkA activation in primary neurons (data not shown), suggesting that serotonin might not be implicated in amitriptyline's agonistic effect. Amitriptyline has anticholinoceptor actions on both the CNS and peripheral organs. The best correlation to the biologic activity of amitriptyline is the anticholinergic actions. For instance, inhibition of nAChRs by mecamylamine had antidepressant-like effects and potentiated the antidepressant activity of amitriptyline when the two drugs were used in combination. Mice lacking high-affinity nAChRs showed no behavioral response to amitriptyline. Hence, the antagonism of nAChRs might be an essential component of the therapeutic action of antidepressants like amitriptyline (Caldarone et al., 2004). In alignment with this finding, we show that mecamylamine also strongly elicited neurite sprouting processes in PC12 cells (supplemental figure 6). Our data demonstrate that amitriptyline exhibits potent neurotrophic activities including neurite outgrowth in PC12 cells, neuronal survival in primary neurons and neuroprotection in mice, which mimic the primary neurotrophic effects of NGF. It is also noteworthy that amitriptyline effectively treats chronic neuropathic pain (Ho et al., 2008; Watson, 2000). Unlike NGF, which causes intense pain, conceivably, amitriptyline achieve its pain reducing effects might entirely distinct from NGF/TrkA mechanisms involving peripheral nerves.
Neurotrophins exert the physiological functions through provoking neurotrophin receptors (TrkA, B and C) homo-dimerization. However, it remains elusive whether the ligand has to be a homodimer in order to trigger the receptor dimerization. Moreover, it is unknown whether the Trk receptors can form a heterodimer or not. In current study, we show that a small tricyclic antidepressant drug amitriptyline provokes both TrkA and TrkB homo- and heterodimerization and activation in transfected HEK293 cells and primary neurons. Amitriptyline elicits potent TrkA and TrkB activation in mice with a similar temporal pattern. Thus, we demonstrate that amitriptyline acts as a novel agonist for both TrkA and TrkB. Strikingly, this small molecule but not any other antidepressant drugs selectively stimulates endogenous TrkA/TrkB heterodimer formation in mouse brain. This finding provides a novel molecular mechanism in dimerization and activation of transmembrane receptor tyrosine kinase (RTK). Endogenous cognate ligands for Trk receptors are homodimers, which can only selectively bind to one of the Trk receptors, whereas amitriptyline binds to both TrkA and TrkB receptors, leading to a heterodimer formation. Hence, this discovery establishes a proof-of-concept model for identifying small molecular agonists and antagonists for RTK. Conceivably, using the established screening assay, numerous small compounds can be identified for mimicking growth factors including EGF, insulin, netrins etc.
Mouse septal neuron × neuroblastoma hybrids SN56 cells were created by fusing N18TG2 neuroblastoma cells with murine (strain C57BL/6) neurons from postnatal 21 days septa. SN56 cells were maintained at 37°C with 5% CO2 atmosphere in DMEM medium containing 1 mM pyruvate and 10% FBS. T17 cells, stably transfected with rat TrkA were cultured in the same medium containing 300 μg/ml G418. NGF and BDNF were from Roche. The specificity of anti-p-TrkA 794 has been previously described (Jeanneteau et al., 2008; Rajagopal et al., 2004). The specificity of anti-p-TrkB 816 has been described before (Arevalo et al., 2006). Anti-TrkB antibody was from Biovision. Anti-TrkA was from Cell Signaling. The chemical library containing 2000 biologically active compounds was from The Spectrum Collection (MicroSource Discovery System, Inc. Gaylordsville, CT 06755). TrkA F592A and TrkB F616A mice have been described previously (Chen et al., 2005). TrkA F592A and TrkB F616A mice, TrkA +/−, TrkB +/− and NGF +/− C57BL/6 mice were bred in a pathogen-free environment in accordance with Emory Medical School guidelines. All chemicals not included above were purchased from Sigma.
T17 cells were seeded in a 96-well plate at 10,000 cells/well in 100 μl complete medium. Cells were incubated overnight, followed by 30 min pretreatment with 10 μM compounds in DMSO (10 mM stock concentration from The Spectrum Collection library). The cells were then treated with 1 μM staurosporine for 9 h. One h before the termination of the experiment, 10 μM MR(DEVD)2, a cell permeable caspase-3-activated fluorescent dye was introduced. Cells were fixed with 4% paraformaldehyde for 15 min. Cells were washed with PBS and incubated with 1 μg/ml of Hoechst 33342 for 10 min. Cover slides were washed with PBS, mounted, and examined using a fluorescence microscope.
Purified TrkA ECD or ICD proteins were incubated with different [3H-amitriptyline] at 4 °C for 10 min in 1 ml of binding buffer (50 mM Na-K phosphate pH 7.1, 200 mM NaCl and 2 nM 3H-amitriptyline (68000 cpm)). After the incubation, the reaction mixture was loaded on filter paper. The mixture was washed with 3 × 5 ml washing buffer (100 mM Tris, pH 7.1). The dried filter paper was put into a small vial and subjected to liquid scintillation counter analysis. The value of the dissociate constant and the number of sites were obtained from Scatchard plots by using the equation r/[L]free = n/Kd - r/Kd, where r is the ratio of the concentration of bound ligand to the total protein concentration and n is the number of binding sites.
Male C57BL/6 mice aged of 60 days were injected intraperitoneally (i.p.) with a single dose of either 30% ethanol in saline or KA (25 mg/kg) (Sigma, MO) or 7,8-dihydroxyflavone (5 mg/kg) followed by KA. Animals were continually monitored for 2 h for the onset of seizure activity. At 5 days following treatment, animals were anesthetized and perfused with 4% paraformaldehyde in 0.1 M phosphate buffered saline. Brains were removed, post-fixed overnight and processed for paraffin embedding. Serial sections were cut at 5 μm and mounted on slides (Superfrost-plus, Fisher). The slides were processed for TUNEL staining in order to assess the degree of DNA fragmentation.
Brain tissues were fixed in 4% paraformaldehyde overnight followed by paraffin embedding. Sections of 6 μm were cut. For immunohistochemical staining, brain sections were deparaffinized in xylene and rehydrated in graded alcohols. Endogenous peroxidase activity was blocked by 3% hydrogen peroxide for 5 minutes and all slides were boiled in 10 mM sodium citrate buffer (pH 6.0) for 10 minutes. Phosphorylated Trk A, Trk A, phosphorylated Trk B, and Trk B were detected using specific antibodies and Zymed Histo-SP AEC kit (Invitrogen, USA). Slides were then counterstained with hematoxylin.
2−4 month-old TrkAF592A and TrkBF616A mice were pretreated with 1NMPP1 in drinking water (25 μM) 1 day before amitriptyline (15 mg/kg) treatment. After 4 h, KA (25 mg/kg) was intraperitoneally injected into the mice. The mice were housed for 4 more days with 1NMPP1 in the drinking water. At day 5, animals were anesthetized and perfused and treated as described above. The brain slides were processed for TUNEL staining in order to assess the degree of DNA fragmentation.
This work is supported by grants from NIH (RO1, NS045627) to K. Ye. The authors are thankful to Dr. David D. Ginty at Johns Hopkins University for the TrkA F592A and TrkB F616A knock-in mice. We thank Dr. Lino Tessarollo at NCI, NIH for Trk heterozygous mice and Dr. Moses Chao at New York University for anti-p-TrkB Y816 antibody.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.