A De Novo Peptide That Induces Axonal Growth Through TrkB Activation

Abstract

Brain‐derived neurotrophic factor (BDNF) and its receptor, tropomyosin‐related kinase B (TrkB), are key regulators of neuronal survival and synaptic plasticity. Impaired BDNF–TrkB signaling is strongly implicated in neurodegenerative and neuropsychiatric disorders, including Alzheimer's disease and major depressive disorder. However, the clinical use of BDNF is limited by poor pharmacokinetics and low tissue stability, highlighting the need for alternative TrkB agonists. Many previously reported agonists have shown inconsistent or indirect TrkB activation, raising concerns about their mechanism of action. In this study, we employed the random nonstandard peptides integrated discovery (RaPID) system to identify de novo macrocyclic peptide ligands that bind the extracellular domain of TrkB with high affinity and selectivity. One of the peptides was turned into a dimeric peptide, diTrbL3, which induced TrkB autophosphorylation and activated canonical downstream pathways, including ERK1/2 and AKT, without detectable cross‐activation of other receptor tyrosine kinases. Functionally, diTrbL3 promoted axonal elongation and upregulated activity‐dependent immediate early genes in primary hippocampal neurons. These findings establish diTrbL3 as a potent de novo TrkB agonist and highlight RaPID‐derived macrocyclic peptides as a promising class of BDNF mimetics for targeting TrkB‐dependent signaling in neurological disease.

A homodimeric anti‐TrkB macrocyclic peptide, DiTrbL3, originally discovered by the RaPID system as monomeric form, dimerizes the TrkB receptor and activate the downstream signaling. DiTrbL3 acts as a mimic of BDNF (the natural protein ligand) and stimulates hippocampal neurons to extend axons.

1. Introduction

Neurotrophins are endogenous polypeptides responsible for regulating the development, function, and survival of neuronal cells and networks across the entire brain. Among the neurotrophin family members, brain‐derived neurotrophic factor (BDNF) has emerged as the leading candidate in potential treatments for neurological and psychiatric disorders [1, 2]. BDNF acts as a natural ligand to two receptors of distinct classes: tropomyosin‐related kinase receptor type B (TrkB) and neurotrophin receptor p75, where BDNF binds with at least 10‐fold higher affinity to TrkB over p75 [3]. TrkB and its ligand BDNF are widely expressed in the adult mammalian brain, including the cerebral cortex, hippocampus, brainstem, and spinal cord nuclei [4]. TrkB is a transmembrane tyrosine kinase receptor protein that belongs to the tropomyosin‐related kinase (Trk) family and upon binding to BDNF undergoes autophosphorylation of tyrosine residues near the cytoplasmic C‐terminal domain (Figure 1a) [2, 5]. BDNF‐induced phosphorylation of TrkB in turn activates three major downstream pathways: PLCγ, PI3K, and MEK‐ERK pathways [2]. Phosphorylation of phospholipase C‐gamma 1 (PLCγ1) results in the production of inositol triphosphate (InsP3) and diacylglycerol (DAG) that releases intracellular calcium stores and thus influences synaptic plasticity via activation of calcium‐dependent protein kinases. Activation of phosphoinositide 3‐kinase (PI3K) leads to the stimulation of protein kinase B (AKT), which results in the suppression of proapoptotic proteins and ultimately promotes cell survival [6]. BDNF‐associated neurite outgrowth, cellular differentiation, and neuronal survival during development are implicated by the downstream stimulation of mitogen‐activated kinase kinases 1/2 (MEK 1/2) and extracellular signal‐regulated kinases 1/2 (ERK 1/2). In addition, the MEK‐ERK pathway is associated with maintaining neuronal structure and function in the adult brain [2, 6].

FIGURE 1.

An artificial dimeric peptide induces potent phosphorylation of endogenous TrkB. (a) BDNF‐induced phosphorylation of TrkB, which in turn activates three major downstream pathways: phospholipase C‐gamma (PLCγ), phosphoinositide 3‐kinase (PI3K), and mitogen‐activated kinase kinases and extracellular signal‐regulated kinases (MEK‐ERK) pathways. In this study, we mainly focused on MEK‐ERK pathway, which is implicated in neurite outgrowth, cellular differentiation, and neuronal survival during development. (b) Anti‐TrkB macrocyclic peptides discovered by RaPID system were converted into homodimeric macrocyclic entities using a polyethylene glycol (PEG) linker of suitable length. Resulting homodimeric peptides were further used for the discovery of a BDNF mimic. (c) Representative result of screening of six types of dimeric peptide by western blotting for phosphorylated Trk (P‐Trk). Primary mouse hippocampal neurons (DIV6) were treated with DMSO, BDNF (0.1 nM), or dimeric peptides (1 nM) for 15 min. Anti‐total TrkB antibody was used for internal control (TrkB). (d) Target specificity of diTrbL3 was evaluated by using ANA‐12, a selective inhibitor for the kinase domain of TrkB. ANA‐12 (100 μM) was added to culture medium 30 min before treatment of peptides. (e) Results of Trk phosphorylation. Phosphorylation levels of TrkB were normalized by total TrkB and, then, normalized by mean of DMSO‐treated group. Error bars indicate standard error of the mean (s.e.m.), n = 3 for each group. ** p < 0.01 (compared with DMSO), ## p < 0.01 (compared between diTrbL3 1 nM and diTrbL3 1 nM with ANA‐12 at 100 μM). One‐way ANOVA followed by Tukey/Kramer's multiple comparison test was used for statistical analysis.

There is growing evidence that BDNF's neurotrophic actions have positive implications in various neurological disorders, including amyotrophic lateral sclerosis (ALS), Parkinson's disease, and Alzheimer's disease [7, 8, 9, 10, 11]. In rodent ALS models, administration of BDNF led to the survival of degenerating motor neurons and rescued axotomized motor neurons [7]. In the nonhuman primates, BDNF protein infusion displayed beneficial anatomical and behavioral effects, and in rodents, BDNF treatment prevented the loss of dopaminergic neurons in the substantia nigra, where in both the models parkinsonism was chemically induced [8, 11]. In transgenic mouse models expressing a mutant form of the amyloid precursor protein, BDNF improved learning and memory, while it prevented neuronal cell death in adult rats following a perforant pathway lesion [9]. It is clear that BDNF therapy has a huge potential in a variety of neurological disorders; however, due to short in vivo half‐life, the clinical trial outcomes of using recombinant BDNF have been disappointing [12, 13]. As BDNF binding to p75 is known to activate cell death pathways, dosage control to alleviate off‐target effects will be challenging [14].

To overcome the limitations of BDNF, small molecules, monoclonal antibodies, and peptide mimetics have been devised as TrkB agonists [1, 15, 16, 17]. Among the list of small‐molecule TrkB agonists, 7,8‐dihyroxyflavone (7,8‐DHF) has been extensively studied in cellular and animal models, showing a promise in managing BDNF‐related neurological disorders [1]. 7,8‐DHF was reported to directly bind to TrkB and induce TrkB autophosphorylation [18, 19]; however, subsequent studies done by independent research groups could not reproduce the autophosphorylation of TrkB by 7,8‐DHF as well as several previously reported small‐molecule TrkB agonists questioning their ability to act as true TrkB agonists [20, 21]. Therefore, for 7,8‐DHF, a path to clinic will be difficult unless the exact molecular mechanism of 7,8‐DHF induction of downstream TrkB pathway is made clear. Besides small‐molecule TrkB agonist discovery and development, limited yet promising efforts have been made toward the discovery of TrkB agonist antibodies that have been shown to activate TrkB in vitro [1, 21, 22, 23, 24]. One of the exciting works on agonist antibodies was recently reported by Merkouris et al., where they selected human antibodies against TrkB using a function‐based target receptor reporter assay from a combinatorial human antibody library. One of the candidate antibodies, ZEB85, was shown to be a full agonist of human TrkB while activating downstream phosphorylation of PLCγ, AKT, and ERK at levels similar to that for BDNF. However, ZEB85 elicited a very weak response in mouse neurons toward TrkB phosphorylation and dendritic outgrowth. For further development of a human antibody as a replacement of BDNF therapy, mouse models will play an important role, for which a human–mouse cross‐reacting antibody would be required. Finally, a cyclic peptide containing a Lys‐Lys‐Arg (KKR) motif, derived from BDNF loop 4, has been reported as a peptide mimetic of BDNF and was shown to promote peripheral myelination through TrkB activation [25]. The peptide exhibited potent agonistic activity; however, it was also found to inhibit the BDNF–TrkB interaction. In addition, its specificity toward TrkB remains unclear because the KKR motif represents a relatively simple binding determinant that may not confer high receptor selectivity.

The random nonstandard peptides integrated discovery (RaPID) system developed in our laboratory has emerged as a powerful platform for the discovery of macrocyclic peptide ligands exhibiting exceptionally high target affinity, remarkable specificity, and improved pharmacokinetic properties compared with their linear counterparts (Figure S1) [26, 27, 28, 29]. By combining the Flexible In vitro Translation (FIT) system with genetic code reprogramming mediated by flexizymes, the RaPID platform enables the generation of extraordinarily diverse libraries of thioether‐macrocyclic peptides. Each peptide is covalently linked to its encoding mRNA via puromycin, allowing direct genotype–phenotype coupling and efficient selection of ligands against a target of interest. Previous studies have demonstrated that macrocyclic peptides discovered through the RaPID system can bind their targets with low‐nM to sub‐nM dissociation constants [26, 29]. Moreover, dimerization of such macrocycles can induce dimerization of their protein targets, thereby activating downstream signaling paths. For instance, dimerization of monomeric macrocycles that bind the hepatocyte growth factor (HGF) receptor resulted in receptor dimerization and activation, leading to HGF‐like cellular responses, including branching morphogenesis, whereas the corresponding monomeric peptides were unable to stimulate HGF receptor signaling [26]. Stimulated by these HGF‐like peptides, we applied a similar strategy to identify macrocyclic peptides that bind TrkB with high affinity and selectivity, leading to the development of a de novo dimeric peptide that selectively activates downstream TrkB signaling (Figure 1b).

2. Results and Discussion

2.1. RaPID Selection Against hTrkB Discovers Potent Peptide Binders

Using a set of two thioether‐macrocyclic peptide libraries (D‐ and L‐library, where the respective ClAc‐D‐tyrosine and ClAc‐L‐tyrosine charged onto tRNAfMetCAU by a flexizyme (eFx) were used to initiate the translation and macrocyclized with a downstream cysteine residue) in the RaPID system (Figure S1), we performed five rounds of selection against a His‐tagged extracellular domain of human TrkB immobilized on Ni‐NTA magnetic beads. After five rounds, the recovery of peptide–mRNA fusion molecules from the immobilized protein beads was significantly increased for both the L‐ and D‐libraries (as compared to the magnetic beads alone; Figure S2a), suggesting that the peptide libraries are selectively enriched with the TrkB‐binding macrocycles. The recovered DNA after the fifth round of screening from each library was sequenced via next‐generation sequencing (Figure S2b). Three most abundant peptides, TrbL1–3 and TrbD1–3, from the respective libraries (Table S1), were synthesized chemically on an automated solid support system without the (GS)₃ linker.

We first assessed the strength and selectivity of hTrkB over hTrkA and hTrkC by quantitative measurements of binding kinetics (binding affinity KD) by means of surface plasmon resonance, SPR, against the extracellular domain of hTrkA–C proteins immobilized on the sensor chip via their His6‐tag (Table 1 and Figure S3). All TrbL/D‐peptides showed potent affinities with KD values of 1.5–70 nM for hTrkB. Most importantly, three peptides, TrbL1, TrbL3, and TrbD2, exhibited high selectivity to hTrkB, i.e., no sign of binding to hTrkA and hTrkC by SPR. Particularly, TrbL3 possesses a 1.5 nM KD and high selectivity, making it the most attractive molecule among them. However, because our ultimate goal was to activate the receptor via dimerization, potent monomeric binding affinity might not directly predict a peptide's ability to induce TrkB dimerization, which could instead depend on where on the receptor each peptide binds. We therefore decided to forward all peptides for further studies.

TABLE 1.

Chemically synthesized monomeric and homodimeric peptides show potent binding to TrkB by SPR. Three most abundant peptides from each library were synthesized as monomers by SPPS such that an N‐terminal acetyl moiety formed a thioether macrocycle through the side chain of a downstream Cys residue (in bold). Each monomeric peptide was linked with a suitable PEG linker to prepare the homodimeric peptide (see also Figure S4). Dissociation constant (KD) with respect to TrkA, TrkB, and TrkC was determined by SPR.

Monomer	Peptide sequencea	K D TrkB, nM	K D TrkA/C, nM	Homodimer	K D TrkB, nM	K D TrkA/C, nM
TrbL1		26	—/—b	DiTrbL1	60	—/—
TrbL2		70	100/—	DiTrbL2	71	7/—
TrbL3		1.5	—/—	DiTrbL3	2	—/—
TrbD1		6	150/—	DiTrbD1	10	29/120
TrbD2		18	—/—	DiTrbD2	36	4/—
TrbD3		15	123/—	DiTrbD3	15	7/—

^a Peptides are cyclized via a thioether bond between N‐term acetyl group and C‐term Cys residue (Figure S1a), where ^AcY = N‐acetyl‐L‐tyrosine and ^Acy = N‐acetyl‐D‐tyrosine.

^b — indicates that reliable KD value could not be determined due to weak affinity (i.e., the measurement profile did not fit to an appropriate 1:1 fitting model and likely KD values are above 1 µM).

2.2. Dimerization of Peptides Retains the Potency of Their Binding to hTrkB

BDNF binds to the extracellular domain of hTrkB and causes TrkB homodimerization, which activates downstream signaling via TrkB autophosphorylation. We reasoned that if two anti‐TrkB macrocyclic peptides are connected with a suitable flexible linker, the resulting homodimeric macrocyclic peptide entity should mimic BDNF's agonistic activity, a concept known as chemically induced dimerization [30]. This mechanism is also consistent with our previous findings in the HGF/MET receptor system, where high‐affinity monomeric macrocyclic peptides bound the receptor but failed to activate signaling, whereas dimerization converted the binders into potent agonists [26]. Therefore, we selected a PEG5–DAP–PEG5 linker (58 Å in length) to synthesize six homodimeric TrbL/D‐peptides, diTrbL1–3 and diTrbD1–3, following a strategy described previously by Bashiruddin et al. (Figure S4) [31]. After the synthesis and purification, binding kinetics of the respective diTrbL1–3 and diTrbD1–3 to the extracellular domain of hTrkA–C were assessed by SPR in the same manner as the study of their monomeric form (Table 1 and Figure S3). The binding affinities of diTrbL/D‐peptides remained the same as those of their monomeric form, indicating that their ability to bind hTrkB was unchanged. In principle, dimerization should show enhanced binding affinity when compared to that of their monomeric counterparts. We did not observe this phenomenon due to the limitation of the SPR assay. In our SPR experiments, both monomeric and dimeric peptides were evaluated over a concentration range of 1–1000 nM. As 1 nM represents the practical lower limit of the assay, any avidity‐driven improvement that reduced the KD below this value would be difficult to resolve accurately. As a result, the measured affinity is expected to approach the assay floor and may not show a discernible increase despite the presence of avidity effects. In contrast, avidity effects lead to loss of selectivity of dimeric peptides to hTrkB, where four out of six peptides lost the hTrkB selectivity and bind hTrkA strongly but did not bind or only weakly bind (diTrbD1) to hTrkC. The most potent binder, diTrbL3, retains high selectivity to hTrkB; i.e., it shows binding to neither hTrkA nor hTrkC on the SPR sensor chip. It is known that TrkA, TrkB, and TrkC share high sequence homology; nevertheless, there is no BDNF cross‐reactivity to hTrkA and hTrkC, which is attributed to the Ig‐like domain known as Trk‐d5 [32, 33]. As RaPID selection allows peptides to bind to any part of hTrkB, it is expected that some peptides may bind to a region that shares homology with hTrkA/C. Therefore, it is not surprising that some peptides show cross‐reactivity to TrkA/C in SPR.

2.3. An Artificial Dimeric Peptide Potently Induced Phosphorylation of TrkB

To define which peptides could exhibit agonist activity, we performed the induction of phosphorylation of mouse TrkB (mTrkB) by diTrbL1–3 and diTrbD1–3 in mouse primary hippocampal neurons. Within the mature domain, human BDNF is 100% identical in amino acid sequence to mature mouse BDNF, making the two proteins functionally equivalent [33]. Consistent with this, BDNF has been shown to act as a TrkB agonist in both human and mouse cell lines [34]. We therefore reasoned that the peptides selected against hTrkB could be tested directly in mouse cell lines. We set a positive control experiment where the exogenous application of the natural protein agonist, BDNF (0.1 nM, for 15 min), induced Trk phosphorylation clearly detected by Western blotting using an anti‐phospho‐Trk antibody (Figure 1c, lane 2), whereas the negative DMSO control did not induce it at all (lane 1). We selected 0.1 nM BDNF as a literature‐supported, subsaturating benchmark that produces robust native TrkB activation while maintaining sensitivity for comparison with diTrbL3 [35]. The exogenous application of diTrbL1–3 and diTrbD1–3 at a concentration of 1 nM defined only a single peptide, diTrbL3, capable of inducing the phosphorylation of mTrkB (lane 5), whereas other peptides had no effect (lanes 3, 4, and 6–8).

Since the anti‐phospho‐Trk antibody used in this analysis also recognizes TrkA and TrkC, though TrkA is not normally expressed in the hippocampal neurons [36, 37], we needed to determine the specificity of diTrbL3 for the induction of the TrkB phosphorylation. For this purpose, we utilized ANA‐12 [38], a specific kinase inhibitor for the intracellular catalytic domain of TrkB. The treatment of ANA‐12 to the hippocampal neurons diminished the phosphorylation of Trk induced by diTrbL3 (Figure 1d, lane 3 vs. lane 4), suggesting that diTrbL3 could trigger the phosphorylation of TrkB selectively. Quantitative analysis of TrkB phosphorylation in dose titration of diTrbL3 revealed that there was clear TrkB phosphorylation even in a subnanomolar concentration of diTrbL3 (0.2 nM, p < 0.01) (Figure 1e). Taken together with the in vitro assessment of the selective binding of diTrbL3 to hTrkB and the above cellular assays, diTrbL3 is a selective agonist for TrkB phosphorylation and is fortuitously cross‐reactive to both mouse and human TrkB proteins.

2.4. DiTrbL3 Peptide Is a Specific TrkB Agonist

The phosphorylation of TrkB Y515 by BDNF induces the association of SH2 (src homology‐type 2) linker proteins such as shc (src homology domain containing) and IRS1/2 (insulin receptor substrate 1/2) and leads to the activation of the PI3K and ERK signaling pathways [39]. Therefore, we examined whether the downstream signaling kinases of TrkB were activated by the exogenous treatment of diTrbL3. As expected, diTrbL3 drastically activated the phosphorylation of Akt and Erk1/2 in hippocampal primary neurons (lane 3, Figure 2a,b), whereas the phosphorylation activity was diminished in the presence of ANA‐12 (lane 4, Figure 2a,b), strongly suggesting that the downstream signaling is TrkB‐dependent. Quantitative results showed significant difference between DMSO‐ and diTrbL3‐induced activation of Akt and Erk1/2 (Figure 2c–e, p < 0.01). In addition, the phosphorylation of these kinases was suppressed by ANA‐12 treatment (Figure 2c–e, p < 0.05 for Akt and p < 0.01 for Erk1/2), which implied that the downstream signaling cascade was initiated by diTrbL3 and induced autophosphorylation of TrkB.

FIGURE 2.

DiTrbL3 peptide is a specific TrkB agonist. Primary mouse hippocampal neurons (DIV6) were treated with DMSO (a negative control of peptide), BDNF (0.1 nM), or diTrbL3 (1 nM) for 15 min. ANA12 (100 μM) was added to medium 30 min before treatment of peptides. (a) Representative western blots for phospho‐Akt (P‐Akt at S473). (b) Representative western blots for phospho‐Erk1/2 (P‐Erk1/2). (c–e) Quantified results of phosphorylation levels of (c) Akt, (d) Erk1, and (e) Erk2. Arrowheads and arrows indicate Erk1 (44 kDa) and Erk2 (42 kDa), respectively. Relative phosphorylation level of each kinase was normalized by mean of the DMSO‐treated group. Error bars indicate s.e.m., n = 3 for each group. ** p < 0.01 (compared with DMSO), ## p < 0.01 and # p < 0.05 (compared between diTrbL3 1 nM and diTrbL3 1 nM with ANA‐12). One‐way ANOVA followed by Tukey/Kramer's multiple comparison test was used for the statistical analysis. (f) Results of phospho‐RTK array. A commercial phospho‐RTK array was used for this study, where 40 kinds of RTKs were included (bottom panel). Hippocampal primary neurons (DIV6) were incubated with stimulants for 15 min. ANA‐12 (100 μM) was added 30 min before treatments of diTrbL3. Each spot indicates phosphorylation level of each RTK. Red rectangle areas indicate duplicated TrkB spots.

We also examined the off‐target activity of diTrbL3 with other receptor tyrosine kinases in an array‐based experiment containing 40 different RTKs (Figure 2f). One hour incubation in primary hippocampal neurons with BDNF caused the phosphorylation of TrkB (Figure 2f, bottom left panel). Very similar to this observation, diTrbL3 only phosphorylated TrkB (Figure 2f, upper right panel). Moreover, ANA‐12‐treated samples gave weaker signals of TrkB phosphorylation induced by diTrbL3 both in RTK‐array and densitometric analysis (Figures 2f and S5). All the above results, spanning in vitro binding selectivity by SPR and unbiased cellular activation profiling across 40 receptor tyrosine kinases, provide strong evidence that diTrbL3 is a specific TrkB agonist at the cellular level.

2.5. DiTrbL3 Treatment Induces Neuronal Gene Activation and Enhances Axonal Outgrowth

Since BDNF plays a critical role in the regulation of synaptic plasticity in response to immediate early genes such as c‐fos and Arc [40, 41, 42], we next quantified c‐fos and Arc mRNA expression levels under the treatment of diTrbL3. Primary cortical neurons were treated with BDNF (1 nM) or diTrbL3 (1 nM) for 1 h. As a result, both c‐fos and Arc expressions were significantly elevated by diTrbL3 (Figure S6, p < 0.01), indicating that diTrbL3 engages the activity‐dependent transcriptional program downstream of TrkB that underlies BDNF's role in synaptic plasticity. The activation of TrkB was implicated in various cellular responses such as neurite arborization, neuronal survival, and axonal elongation [39]. Therefore, we next verified the effect of diTrbL3 treatment on the cellular level in primary neurons. Three days treatment of diTrbL3 caused elongation of axonal length of primary neurons, similar to BDNF treatment (Figure 3a). Quantitative results showed that 10 and 100 nM of diTrbL3 significantly enhanced axonal outgrowth (Figure 3b, compared to DMSO treatment, p < 0.01). The neuronal images in Figure 3a are snapshot to represent a visual in support of neurite growth, and no significant qualitative differences beyond those captured by neurite length quantification were observed. This demonstrates that diTrbL3 can substitute for BDNF in promoting axonal growth, a process fundamental to how neurons form and strengthen the connections underlying learning, memory, and overall brain function.

FIGURE 3.

DiTrbL3 enhances axonal outgrowth of hippocampal neurons. (a) Representative images of primary hippocampal neurons treated with Di‐L3 peptides. Hippocampal neurons were treated with 1, 10, or 100 nM of diTrbL3 or 2 nM BDNF from DIV0 to 3, and then, immunocytochemistry was performed by using anti‐Tau (red) and MAP2 (green) antibodies with nuclear staining (DAPI; blue). Axons were identified as Tau‐positive and MAP2‐negative neurites. The length of each axon was quantified by Neuron J44. Scale bar indicates 50 mm. (b) A quantitative result of axonal outgrowth experiment (n = 60 for DMSO, 20 for BDNF, and 30 for each concentration of diTrbL3 drug). Error bars indicate s.e.m., ** p < 0.01 (compared with DMSO). One‐way ANOVA followed by Tukey/Kramer's multiple comparison test was used for statistical analysis.

Several efforts for the development of specific agonists for TrkB have been made and identified a variety of small compounds including 7,8‐dihydroxyflavone (7,8‐DHF) [18], LM22A‐4 [34], deoxygedunin [43], demethylasterriquinone B1 (DMAQ‐B1) [44], amitriptyline [45], and deprenyl [46].

7,8‐DHF, the most extensively studied of these, has shown protective effects in models of neurodegenerative, neurodevelopmental, and neuropsychiatric disorders. However, its neuroprotective effect has also been reported to be independent of TrkB activation as the Trk inhibitor K252a does not block it, suggesting that 7,8‐DHF may act via other protein targets [47, 48, 49]. Therefore, to discover a potent and TrkB‐selective mimic of BDNF and overcome the limitation of small molecules, we undertook this study. Using the RaPID system, we identified isoform‐specific novel monomeric and dimeric macrocyclic peptide binders for hTrkB, with low nanomolar affinities. One of the dimeric peptides (diTrbL3) induced potent phosphorylation of TrkB and its downstream effectors, including ERK or AKT, in neurons derived from mouse brain, where these effects were blocked by the TrkB‐selective inhibitor ANA‐12. Subsequent analysis using the RTK array showed that the effect of the peptides for TrkB was highly specific. Finally, the peptide enhanced neurite outgrowth in primary hippocampal neurons and induced the expression of immediate early genes. Although the monomeric TrbL3 originated from a selection against hTrkB, diTrbL3 shows potent activation of TrkB in mouse neuronal cells. With antibody‐like affinity and cross‐reactivity to mTrkB, diTrbL3 could prove to be a valuable tool in further investigating the BDNF–TrkB pathway in mouse models.

3. Conclusion

The present study describes the discovery of de novo macrocyclic peptide binders of human TrkB. These macrocyclic peptides were converted into their respective homodimeric entity linked by a flexible PEG linker to potentially act as a mimic of TrkB's natural agonist, BDNF. Among six homodimeric macrocyclic peptides exhibiting tight binding to the extracellular domain of hTrkB, one of the homodimeric peptides, diTrbL3, was able to activate downstream signaling via TrkB autophosphorylation in mouse neuronal cell lines. DiTrbL3 showed potent and specific mouse TrkB agonism as well as induced neuronal gene activation and enhanced axonal outgrowth. Therefore, diTrbL3 could provide a significant lead in the development of therapeutic candidate in treatments for neurological and psychiatric disorders. As with many small molecules being developed for central nervous system indications, however, an outstanding challenge for diTrbL3 is poor cell permeability and the need to cross the blood–brain barrier (BBB) [50]. Although TrkB is an extracellular target and diTrbL3 is potent even at subnanomolar concentrations, an effective delivery strategy to overcome the BBB will be required before diTrbL3 can be advanced as a therapeutic candidate for diseases of the central nervous system. Moreover, our results further validate that homodimeric peptides are great alternatives to natural protein ligands in inducing cell‐surface receptor dimerization, thereby extending the concept of chemically induced dimerization to impact other therapeutic areas.

Author Contributions

R.M. and H.S. conceived and designed the study. R.M. performed RaPID selection, peptide synthesis, and biochemical assays. K.T., K.F., and T.T. performed cellular assays. All authors analyzed the experimental data, discussed the results, and were involved in the preparation of the manuscript.

Funding

This work was supported by the Japan Agency for Medical Research and Development (JP19am0101072 and JP16am0301001h) and the Japan Society for the Promotion of Science (JP20H05618, 16H06463, and 16H06316).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplementary Material

Maini Rumit, Tamada Kota, Fukuda Kunihiro, Takumi Toru, Suga Hiroaki, A De Novo Peptide That Induces Axonal Growth Through TrkB Activation, ChemBioChem 2026, 0, e70517. 10.1002/cbic.70517

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information and on request from the corresponding author.