Animal models suggest the TRIP8b-HCN interaction is a therapeutic target for major depressive disorder

KEYWORDS:

• HCN
• TRIP8b
• depression
• Pex5
• ion channel

1. Introduction

Major depressive disorder (MDD) remains a substantial public health problem despite the existence of many available treatments. One issue with existing pharmacological options is that nearly all of them act to increase monoaminergic neurotransmission. Given the mechanistic similarity, it is not surprising that many patients fail to respond adequately to current treatments. Recent work using animal models of MDD has suggested that the excitability of the hippocampus may play an important role in mediating the symptoms of MDD [1]. In line with this hypothesis, human patients diagnosed with MDD show less hippocampal functional connectivity [1]. This potential MDD mechanism suggests that therapeutic interventions that increase hippocampal excitability may function as antidepressants.

2. Hyperpolarization-activated cyclic nucleotide-gated channels as a target for the treatment of MDD

Hyperpolarization-activated cyclic nucleotide-gated (HCN) channels regulate the excitability of neurons in the central nervous system [2]. These channels are encoded by four pore-forming subunits (HCN1–4), wherein HCN1 and HCN2 are the predominant isoforms in the brain. Because these channels are open at resting membrane potentials, they lower the membrane resistance [2], and in CA1 pyramidal neurons, they have been shown to reduce excitability [3]. Given that HCN channels are important for limiting hippocampal excitability, several labs have begun to investigate if these channels might also be a target for the treatment of MDD. Interestingly, shRNA knockdown of HCN1 and knockout (KO) of HCN1 or HCN2 have all been shown to increase antidepressant-like behavior in the tail suspension task (TST) and forced swim task (FST) [3,4]. These behavioral tasks are commonly used to screen for antidepressants, and suggests that interventions that limit hippocampal HCN channel function may be sufficient to function as antidepressants [5]. However, because HCN channels play a critical role in regulating heart rate, systemically administered HCN channel antagonists are expected to produce cardiac effects that could limit therapeutic utility in MDD [2].

To circumvent cardiac effects and take advantage of the potential for inhibiting HCN channel function as a novel antidepressant strategy, our lab recently proposed interfering with the function of a brain-specific auxiliary subunit of HCN channels named tetratricopeptide repeat (TPR)-containing Rab8b-interacting protein (TRIP8b) [6,7]. As in HCN1 KO mice, TRIP8b KO mice also exhibit an antidepressant-like phenotype with less time immobile in the TST and FST tasks [3,8] (Table 1). To date, the most salient role identified for TRIP8b in vivo is trafficking HCN channels to the neuronal cell membrane and into the distal dendrites of CA1 pyramidal neurons [3]. As such, our recent work has focused on whether TRIP8b-mediated HCN channel trafficking in CA1 is sufficient to cause changes in antidepressant-like behavior [8]. We found that viral rescue of TRIP8b in CA1 restored HCN channel trafficking and reversed the increase in antidepressant-like behavior that is normally seen in TRIP8b KO mice [8]. Combined with earlier work showing that loss of HCN1 in the same region leads to decreased immobility in the TST and FST tasks [4], these results indicate that antidepressant-like behavior is highly sensitive to manipulations of HCN channel trafficking in CA1.

Table 1. Summary of studies linking HCN channel function in the hippocampus to antidepressant-like behavior.

Model	Distal dendritic enrichment of HCN channels in CA1	Antidepressant-like behavior	Reference
Hcn1^-/- mice	Decreased	Less immobile time in FST and TST	Lewis et al. 2011 [3]
Hcn2^ap/ap* mice
Trip8b^-/- mice
shRNA knockdown of HCN1 in the dorsal hippocampi of rats	Decreased	Less immobile time in FST	Kim et al. 2012 [4]
AAV-TRIP8b in the dorsal hippocampi of Trip8b^-/- mice	Increased	More immobile time in FST and TST, compared with Trip8b^-/- mice	Han et al. 2016 [8]
AAV-TRIP8b(Δ58)** in the dorsal hippocampi of Trip8b^-/- mice	No change	No change in immobile time in FST or TST
AAV-TRIP8b(N13A)*** in the dorsal hippocampi of Trip8b^-/- mice	Decreased	Less immobile time in FST and TST, compared with Trip8b^-/- mice

*The apathetic allele of Hcn2 contains a nonsense mutation that prevents any HCN2 protein from being produced [3]. **AAV-TRIP8b(Δ58) is a TRIP8b mutant that is unable to bind with the CNBD of HCN channels [8]. ***AAV-TRIP8b(N13A) is a point mutation in a key asparagine residue in one TPR domain of TRIP8b that disrupts the interaction with the C terminus of HCN channels [8]. AAV: Adeno-associated virus.

3. Structure of the TRIP8b–HCN interaction

TRIP8b binds to HCN pore-forming subunits in a 1:1 ratio at two distinct locations [9] (Figure 1). First, there is an interaction that occurs between a conserved 80 amino acid stretch of TRIP8b (referred to as TRIP8b_core [10]) and the cyclic-nucleotide binding domain (CNBD) of HCN channels. This interaction is exclusively responsible for the effect of TRIP8b on HCN channel gating and cAMP dependence [10]. In addition, there is a second interaction between TRIP8b and HCN that occurs between the TPR domains of TRIP8b and the C terminal tripeptide tail of HCN subunits (‘SNL’ in HCN1, HCN2, and HCN4 but ‘ANM’ in HCN3). Although each of the two interactions plays a role in trafficking HCN channels in vitro, the C terminal tail interaction is the stronger of the two interactions [10]. In vivo experiments have revealed findings similar to those from in vitro studies to establish the importance of the C terminal tail interaction for TRIP8b-mediated HCN channel trafficking [8]. Viral expression of an HCN1 construct lacking the substrate for this interaction [11] or expression of a TRIP8b construct incapable of binding the C terminal tail of HCN both resulted in impaired HCN channel trafficking [8]. As an important note for therapeutic purposes, loss of only the C terminal tail interaction actually increased antidepressant-like behavior of TRIP8b KO animals and reduced the amount of HCN protein [8]. This result suggests that binding of TRIP8b_core to the CNBD of HCN channels in the absence of the C terminal tail interaction facilitates degradation of the channel and in turn promotes antidepressant-like behavior.

Figure 1. Schematic of the interaction between HCN and TRIP8b, reproduced from our previous report [8]. A single subunit of HCN is represented in black, with the cyclic-nucleotide binding domain (CNBD) and C terminal tail (SNL) highlighted. TRIP8b contains a variable N terminus (labeled) as well as a domain that interacts with the CNBD (represented by a dark gray shape) and a series of TPR domains (light gray shapes) that interact with the C terminus of HCN.

Targeting TRIP8b circumvents cardiac issues associated with directly antagonizing HCN channels [7], but raises the question of off-target effects associated with proteins structurally related to TRIP8b. TRIP8b was initially identified as Pex5l for Peroxin 5-like protein, based on its homology to Peroxin 5 (Pex5). Although the two proteins differ substantially at their N terminus, the C terminus of each contains a series of TPR domains that mediate protein–protein interactions [9]. Pex5 is the peroxisomal import receptor responsible for the transport of peroxisomal proteins from the cytosol into the peroxisome [12]. This is accomplished by binding Peroxisomal Targeting Signal 1 (PTS1) motifs [13], which typically occur as variations on a C terminal ‘SKL’ consensus sequence. Importantly, there is a wide range of PTS1 motifs that are bound by Pex5, with some sequences varying substantially in composition and affinity from the canonical ‘SKL’ sequence [12]. Despite this similarity in terms of function and substrate, knockout of TRIP8b does not affect peroxisomal function [3] and HCN channels are not targeted to peroxisomes. The crystal structure of the TPR domains of both TRIP8b [9] and Pex5 [13] have been solved, and differences in the residues involved in binding their respective cargo proteins are thought to give rise to this specificity. As such, despite their considerable structural similarity, differences in their endogenous substrates indicate that the TPR domains of the two proteins are sufficiently distinct to allow TRIP8b to be molecularly targeted to the exclusion of Pex5 [9].

4. The CNBD binding site as a therapeutic target

The TPR domains of TRIP8b form a deep pocket to accommodate the C terminal tail of HCN [9] and this interaction appears to be susceptible to disruption by small molecules [6]. In contrast to this well-defined interaction, the TRIP8b–HCN interaction that occurs between the CNBD of HCN and TRIP8b_core involves broad surfaces on both proteins [14], suggesting a target that is less amenable to small molecule interference. Although the CNBD binding site may not be easily disrupted, evidence from in vivo viral expression experiments suggests that disrupting this interaction could still increase antidepressant-like behavior [8]. While overexpressing wild-type TRIP8b reversed the antidepressant-like behavior of TRIP8b KO mice, overexpressing a TRIP8b mutant protein in which only the CNBD binding is impaired did not [8]. Of note, although eliminating the CNBD interaction prevented behavioral effects of restoring TRIP8b expression in the knockout, it did not augment antidepressant-like behavior as was seen following expression of the TRIP8b mutant protein in which only the C terminal tail interaction was ablated [8]. Overall, these data indicate that the CNBD binding site is a more difficult target and that it may be less effective than blocking the C terminal binding site for regulating behavioral symptoms.

5. Expert opinion

Targeting the TRIP8b–HCN interaction is emerging as an attractive therapeutic target for MDD as it offers the possibility of limiting HCN channel function only in the brain [6]. Recent work on the structure–function relationship between TRIP8b binding to HCN and subsequent HCN channel trafficking into the distal dendrites of CA1 pyramidal neurons indicates that loss of either of the two TRIP8b–HCN binding sites impairs channel function [8]. However, we reason that the TPR domains of TRIP8b represent the best candidate for small molecule targeting because of the deep pocket formed and the high specificity of this interaction in vivo [9]. As an alternative to the development of small molecules [6], it may also be possible to use a gene therapy-based approach [4,8] in order to limit hippocampal HCN channels with an shRNA strategy.

Although evidence exists that impairing HCN channel function in the hippocampus increases antidepressant-like behavior, it remains unclear what role HCN channels play in the etiology of MDD. Recent work in other brain structures has identified changes in the current mediated by HCN channels in rodent models of depression [15], but these studies did not examine the hippocampus. Using human data, genome-wide association studies have not shown a direct link between HCN channel function and MDD, but these studies are not typically designed to evaluate depression resistance, the behavioral phenotype expected with mutations in genes encoding TRIP8b, HCN1, or HCN2. Regardless whether or not HCN channels are involved in the pathogenesis of MDD, converging lines of evidence from studies in animal models suggest that limiting HCN channel function holds promise as a novel approach to treating symptoms of depression.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Additional information

Funding

D. M. Chetkovich is supported by National Institutes of Health Grant 2R01NS059934, R01MH106511 and R21MH104471, Brain Research Foundation SG 2012-01 and the Chicago Biomedical Consortium HTS-004. Y. Han is also supported by the Chicago Biomedical Consortium HTS-004.