GV1001 reduces pathological 4R tau and functional deficits in models relevant to progressive supranuclear palsy

Abstract

GV1001, a peptide drug derived from human telomerase reverse transcriptase, has showed the therapeutic effect in the animal model of Alzheimer’s disease (AD), a representative chronic neurodegenerative disease having impaired learning and memory. In our previous studies, GV1001 has showed the multi-functions including anti-apoptosis, anti-oxidative stress, and anti-neuroinflammation in AD-related in vitro and in vivo systems. Here, in addition to these previously reported functions, GV1001 was discovered to reduce the protein level of 4R tau isoform in the pathological condition. There is no studies providing the potential of GV1001 as a therapeutic drug for neurodegenerative movement disorders. Progressive supranuclear palsy (PSP) is a rare atypical Parkinsonism in the midbrain region, leading to more severe motor symptoms and very rapid pathological progression. Increased 4R tau isoform in an affected brain region of the primary 4R tauopathy is a distinct pathological character in PSP patients. In this study, GV1001 down-regulated the protein level of 4R tau specifically in an annonacin-induced PSP in vitro neuronal model as well as in vivo study using 4R TauP301L-BiFC mouse model. These findings suggest a novel role of GV1001 in 4R tauopathy and support its disease-modifying potential within the context of 4R tau–driven neurodegenerative models, including PSP.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-42195-7.

Introduction

Progressive supranuclear palsy (PSP) is a neurodegenerative disease that progresses rapidly after the onset of symptoms and leads to death within on average of 5–8 years¹. Disease prevalence is approximately 5–7 per 10,000 people, and the proportion of patients diagnosed with PSP among those with onset Parkinsonism has been reported at approximately 5–6%^2,3. Although the pathological mechanism of PSP has not yet been completely elucidated, abnormal accumulation of 4R tau proteins in the midbrain region have been reported to lead to neuronal cell death, resulting in midbrain atrophy and motor dysfunctions^4,5.

Tau is a microtubule-associated protein (MAP) that is abundant in neuronal axons, where it plays a central role in the stabilization of microtubule (MT) bundles^6–8. MT, which plays a fundamental role in many pivotal processes such as neuronal activity, is a key structural and functional element in neuronal axons, supporting microtubule dynamics and transporting small organelles through the axon. The function of tau in stabilizing MT is primarily regulated by phosphorylation, and disruption of the MT network due to the loss of tau function, is observed in a heterogeneous group of neurodegenerative disorders called tauopathies, which include corticobasal degeneration, frontal temporal lobar dementia (FTLD), and PSP.

Human tau is encoded by the MAPT gene, located at the q21 locus on chromosome 17. The tau gene consists of 16 exons numbered 0–14¹¹. In the adult human central nervous system, MAPT expression produces six major tau isoforms, while two are also produced in the peripheral nervous system. Tauopathy can be classified using the isoforms that accumulate in NFTs. The brain-specific isoforms vary in the number of N-terminal inserts (0 N, 1 N, or 2 N) and C-terminal repeat domains 3-repeat (3R) or 4-repeat (4R) because of the alternative splicing of exons 2, 3, and 10, resulting in proteins of between 48 kDa (0N3R) and 67 kDa (2N4R)¹².

PSP is generally classified as a representative 4R tauopathy in which neuronal degeneration occurs mainly due to 4R tau overexpression and aggregation. PSP develops as a 4R-tau pathology in the basal ganglia, resulting in motor dysfunctions such as an ataxic gait, followed by clinical features such as dysarthria, agitation, visual impairment, and cognitive or executive dysfunction¹³.

Current therapeutic strategies for PSP target the three types of 4-repeat tau pathology and include (1) regulating MAPT expression with antisense oligonucleotides or splicing modulators, (2) modulating tau post-translational modifications, and (3) inhibiting tau propagation through trans-synaptic pathways or microglial involvement^1,14. Despite the extensive testing of various drugs, none have shown sufficient efficacy in alleviating PSP symptoms.

GV1001 is a synthetic peptide composed of 16 amino acids (EARPALLTSRLRFIPK), as derived from the active site of human telomerase reverse transcriptase (hTERT_611–626)¹⁵. GV1001 was originally developed as an anticancer treatment for solid tumors such as advanced pancreatic cancer, melanoma, non-small cell lung cancer, advanced hepatocellular carcinoma, cutaneous T-cell lymphoma, and B-cell chronic lymphocytic leukemia^16,18,18.

Studies investigating the use of GV1001 in benign prostatic hyperplasia (BPH) Phase 3 and an AD Phase 2 clinical trials demonstrated no serious adverse reactions^19,20.

It has been reported that GV1001 exhibits multi-several functions, such as neuroprotection, recovery from mitochondrial dysfunction, anti-oxidative stress, and anti-inflammation in cells of the central nervous system have been reported for GV1001, indicating its efficacy^21,23,23. Koh et al. showed improved learning and memory in a GV1001 injected AD Tg mouse model^24,25.

However, the potential of GV1001 as a therapeutic drug for neurodegenerative movement disorders has not been tested yet. PSP is a rare atypical Parkinsonism related to the neuropathology in the midbrain region, leading to more severe motor symptoms and very rapid pathological progression. In this study, GV1001 injection significantly improved the motor performance and cognitive function of 4R TauP301L-BiFC transgenic mouse, a 4R tauopathy mouse model showing motor dysfunction. In addition, using in vitro PSP neuronal model induced by annonacin treatment, GV1001 was revealed to alleviate the 4R tau level specifically. Increased 4R tau isoform in brain regions affected by primary 4R tauopathy is a specific pathological character in PSP patients. This novel efficacy of GV1001 was confirmed in 4R Tau P301L-BiFC mouse as well. This study suggests a disease-modifying therapeutic potential of GV1001 for PSP patients through the novel function of GV1001 in reducing the level of 4R tau protein.

Materials and methods

Animals

To monitor tau self-assembly in the brain, a novel tau transgenic mouse was generated, denoted 4R TauP301L-BiFC²⁶. Tau self-assembly was monitored and quantified by introducing a bimolecular fluorescence complementation technique to human tau with a P301L mutation, allowing fluorescence measurements in the brains of the transgenic 4R TauP301L-BiFC mice. Insoluble 4R TauP301L-BiFC aggregates were observed to accumulate with endogenous mouse tau in the brain, showing a subsequent increase in fluorescence after 9 months. Neuronal degeneration and cognitive deficits were observed from 12 months, for which the general symptoms were identified by visual inspection. The experiment was conducted in the Animal Room at the Korea Institute of Science and Technology (KIST), with conditions set at 22 ± 3 ℃, 55 ± 15% relative humidity, 12 h light and dark (08 : 00 light on ~ 20 : 00 light off), 150 to 300 lx lx illumination, and ventilation performed 10 to 20 times/h. Mice were maintained on IVC racks (MSRS-M70S; Orient Bio Inc.) during the acclimation, administration, and observation periods, with three mice in each cage. Solid feed for rats and mice (LabDiet 5053) was provided freely over the experimental period, and tap water was freely provided and replaced once per day. Animal protocols followed the principles and practices outlined in the approved guidelines by the Institutional Animal Care and Use Committee of the Korea Institute of Science and Technology and ARRIVE guidelines²⁷. All animal experiments were approved by the Korea Institute of Science and Technology.

Rota-rod test

The rotarod test was conducted to evaluate motor ability and motor learning. During the 3-day adaptation training period, mice were allowed to adapt to the experimental space for one hour before a behavioral experiment in which mice were forced to walk on a treadmill at speeds of 5, 15, and 20/25 rpm for 360 s. Mice that lost their balance and fell were placed back on the treadmill. After 3 days of adaptation training, the treadmill speed was gradually increased from 5 rpm to 40 rpm for 480 s and the behavioral experiment performed, with the time(s) until mice lost balance and fell to the floor recorded.

Open field test

General activity and locomotion were assessed in an open-field arena (40 × 40 cm) with black walls and a white floor. Mice were habituated to the testing room for 1 h prior to the experiment, then placed in the center and allowed to explore freely for 10 min. The total distance moved and pathways travelled were recorded automatically and analyzed using theEthoVision XT 11.5 video tracking system.

Novel object recognition test

Training (familiarization phase) for the recognition test was performed by allowing the mice to investigate two identical objects to which they were previously habituated for 10 min in an open field. Testing (recognition phase) was conducted 24 h later by exchanging one of the familiar objects with a novel object that differed in color and shape, and allowing mice to explore the objects for 10 min. The exploration times of the familiar (old) and new (novel) objects were recorded using an EthoVision XT 11.5 video tracking system. Memory was operationally defined using the recognition index (RI), which was calculated by dividing the time spent investigating the novel or familiar object by the total time spent investigating the objects during testing. Investigation time was measured when mice were facing the vicinity of an object.

Brain tissue preparation

After the behavioral test, mice were perfused with 0.9% saline. The brains were then rapidly extracted for analysis. For imaging analysis, one hemisphere is fixed in PBS containing 4% paraformaldehyde for 24 h, before dehydration in 20% sucrose for 24 h followed by 30% sucrose for 48 h. Brains were then embedded in O.C.T. and samples cut serially to obtain 30-µm thick coronal sections on a cryostat before transferring slices to PBS containing 0.05% sodium azide and preserving at 4 °C for analysis. For immunoblot analysis, the other hemisphere is suspended in RIPA lysis buffer containing a protease and phosphatase inhibitor cocktail and the tissues disrupted using homogenizers before incubation at 4 °C for 2 h. Homogenized mixtures were then centrifuged at 13,000 rpm at 4 °C for 20 min, and the supernatants collected and stored at − 80 °C.

Immunofluorescence image analysis

For immunofluorescence, brain tissue slices were stained with AT8 (pS202/T205, 1:200) and primary antibodies detected using Alexa Fluor 633-conjugated anti-mouse secondary antibodies (1:500). Nuclei were stained with Hoechst 33,342 in PBS (1:2000) and fluorescence images (BiFC; λex = 460–490 nm, λem = 500–550 nm, AT8; λex = 620–640 nm, λem = 650–700 nm, Hoechst; λex = 330–375 nm, λem = 430–470 nm) acquired using a Zeiss Axio Scan Z1.

Immunoblot analysis

Protein lysates were quantified using BCA assay (Invitrogen, USA) and then 20 µg of samples were boiled for 5 min, separated in 4–12% Bis-Tris Nu-PAGE gels (Invitrogen, USA), transferred onto polyvinylidene difluoride (PVDF) or nitrocellulose (NC) membranes (Bio-rad). Membrane were blocked with 10% skim milk (Bioworld, USA) for 1 h and then probed with antibodies against the indicated proteins (Table 1).

Table 1.

Reagents.

Name	Cas no.	Manufacturer	Catalog no.
Anti-β-Actin antibody	N/A	Sigma-Aldrich	A1978
Anti-Tau antibody [Tau-5]	N/A	Abcam	ab80579
Anti-Tau (Phospho S199)	N/A	Abcam	ab81268
Anti-Tau (Phospho S396)	N/A	Abcam	ab109390
Anti-Phospho-Tau (Ser202,Thr205) Monoclonal Antibody (AT8)	N/A	Invitrogen	MN1020
Anti-Tau (4-repeat isoform RD4) Antibody	N/A	Merck	05–804
Tau Monoclonal Antibody (HT7)	N/A	Invitrogen	MN1000
GAPDH (D16H11) XP® Rabbit mAb	N/A	Cell Signaling	5174s
Anti-PSD95 antibody [7E3-1B8]	N/A	Abcam	Ab13552
Synaptophysin (D8F6H) XP® Rabbit mAb	N/A	Cell Signaling	36,406
Anti-Glial Fibrillary Acidic Protein Antibody	N/A	Merck	MAB360
Anti-Iba1 antibody (EPR16588)	N/A	Abcam	ab178846
FKBP51/FKBP5 Antibody	N/A	Novus Biologicals	NBP1-84676
Complement C3 Polyclonal Antibody	N/A	Invitrogen	PA5-21349
COX-2 (H-3)	N/A	Santa Cruz	sc-376,861
TSPO Polyclonal Antibody	N/A	Invitrogen	PA5-75544
Cleaved Caspase-3 (Asp175) Antibody	N/A	Cell Signaling	9661s
Caspase-3 Antibody	N/A	Cell Signaling	9662s
Bovine serum albumin	9048-46-8	Cellconic	A3294
DMSO	67–68-5	Sigma-Aldrich	D2650
DNaseⅠ	N/A	Sigma-Aldrich	10,104,159,001
DPBS	N/A	Welgene	LB 001–02
Frozen section media	N/A	Leica	3,801,480
Goat anti-Mouse IgG (H+L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 633	N/A	Invitrogen	A21050
Goat Anti-Mouse IgG H&L (HRP)	N/A	Abcam	ab6789
Goat Anti-Rabbit IgG H&L (HRP)	N/A	Abcam	ab6721
Hoechst 33,342	23491-52-3	Invitrogen	H3570
Paraformaldehyde	30525-89-4	Sigma-Aldrich	158,127
Phosphatase inhibitor cocktail 3	N/A	Sigma-Aldrich	P0044
Protease inhibitor cocktail	N/A	Sigma-Aldrich	P8340
RIPA buffer	N/A	Sigma-Aldrich	R0278
Sterile saline	7647-14-5	JW Pharmaceutical	N/A
Sucrose	57-50-1	Sigma-Aldrich	84,097
Trypsin-EDTA	N/A	Cellgro	25–053-CI
Tween 20	9005-64-5	Sigma-Aldrich	P9416
10% SDS	151-21-3	Welgene	ML 009-01
10X TBS	N/A	Bio-rad	1,706,435
10X Tris/Glycine Buffer	N/A	Bio-rad	1,610,771
10X Tris/Glycine/SDS Buffer	N/A	Bio-rad	1,610,772
2,2,2-Tribromoethanol	75–80-9	Sigma-Aldrich	T48402
Retinoic acid 98% (HPLC), powder	N/A	Sigma-Aldrich	R2625
Recombinant Human BDNF Protein, CF	N/A	R&D systems	11,166-BD
Annonacin	N/A	MedChem express	HY-N2877
Lipopolysaccharides from escherichia coli O55:B5	N/A	Sigma-Aldrich	L2880
Human IFN-gamma Recombinant Protein	N/A	PeproTech	300-02
Hydrogen peroxide solution	7722-84-1	Sigma-Aldrich	H1009

Statistical analysis

Two-tailed t tests were used to compare two groups, and one-way or two-way analysis of variance (ANOVA) used to compare multiple groups depending on the number of independent variables. Statistical analyses were performed using Excel (Microsoft Corp.) or Prism (GraphPad Software Inc.).

Results

Administration of GV1001 in TauP301L-BiFC mice

To evaluate the therapeutic effect of GV1001 on motor dysfunction and cognitive impairment, 7-month-old TauP301L-BiFC mice were treated with GV1001 (0.1, 1, and 2 mg/kg) or leuco-methylthioninium-bis(hydromethanesulfonate) (LMTM) (15 mg/kg) three times per week by subcutaneous (SC) or oral (PO) administration for 21 weeks (n = 10 per group) (Fig. 1A; Table 2). The body weight of the mice was measured once every 2 weeks to check any weight change induced by administration of the drug; however, no significant difference was observed in any of the groups over the 21 weeks (Supplementary Data 1 A, 1 B), indicating that the administered drugs did not cause stress.

Fig. 1.

Behavioral tests for motor ability and locomotor activity of TauP301L-BiFC mice subjected to GV1001 administration. A Experimental design for GV1001 administration. TauP301L-BiFC mice were subcutaneously injected with GV1001 (0.1, 1, and 2 mg/kg) three times a week over a total of 21 weeks (n = 10 per group), from 7 to 12.5 months of age. B, C TauP301L-BiFC mice injected with GV1001 (1 and 2 mg/kg) remained on the Rota-rod treadmill for 480 s in the motor learning ability test. D Measurement of total distance travelled (E) and speed (F) of TauP301L-BiFC mice in each group in an open field arena. Data represent the mean ± S.D. and were analyzed using one-way analysis of variance with Dunnett’s multiple-comparisons test; *p < 0.05; **p < 0.01; n.s., not significant compared with vehicle group.

Table 2.

Rota-rod test.

Article	Dose (mg/kg)	Cumulative Duration (s)
Vehicle	N/A	226.4±151.5
LMTM	15	354.7±114.8
GV1001	0.1	364.3±108.3
GV1001	1	409.5±89.3**
GV1001	2	391.1±102.9*

The data represent the means ± S.D. and analyzed using two-way ANOVA with Tukey's multiple-comparisonstest; *p<0.05; **p<0.01 compared with vehicle group

This experiment does not include a blind test. Sex: Male and Female. Age, number, and body weight range: 7 months, 50 (30 males, 20 females), Body weight range (Female = 23.4±2.7 g; Males = 30.4±2.7 g). Quarantine and acclimation: The general symptoms were observed by visual inspection of the appearance when animals were obtained and purified it for more than 7 days. Identification method: Experimental animals were identified by ear-punching.

Recovery of impaired motor ability in TauP301L-BiFC mice due to GV1001 injection

A Rota-rod test was performed to evaluate the motor performance of the TauP301L-BiFC mice following 21 weeks of GV1001 administration (Fig. 1B). The vehicle group remained on the treadmill for an average of 226 s and the LMTM-administered group for an average of 354 s; however, the observed differences were not statistically significantly different (Fig. 1B). The GV1001-injected group that received 0.1 mg/kg remained on the treadmill for an average of 364 s, while the 1 mg/kg group stayed for 409 s, and the 2 mg/kg remained for 391 s (Table 3), indicating increased retention time for the 1 mg/kg and 2 mg/kg GV1001-injected groups (p < 0.01, p < 0.05) (Fig. 1C). This result suggests that the GV1001 treatment improved the motor-learning ability of the TauP301L-BiFC mice.

Table 3.

Open field test.

Article	Dose (mg/kg)	Total distance moved (cm)	Velocity (cm/s)
Vehicle	N/A	1878.3 ± 246.2	3.1 ± 0.4
LMTM	15	2607.7 ± 595.6*	4.4 ± 1.0*
GV1001	0.1	2123.1 ± 785.6	3.5 ± 1.3
GV1001	1	2573.3 ± 462.5*	4.3 ± 0.8*
GV1001	2	2607.3 ± 528.3*	4.4 ± 0.9*

The data represent the means ± S.D. and analyzed using one-way ANOVA with Dunnett's multiplecomparisons test; *p<0.05 compared with vehicle group.

Recovery of locomotor activity in GV1001-injected TauP301L-BiFC mice

An open field test was also performed following the 21 weeks of GV1001 administration to evaluate the locomotor activity of the TauP301L-BiFC mice (Fig. 1D). The total distance covered and the speed at which the TauP301L-BiFC mice moved when placed in an open-field arena were analyzed. The results showed that the vehicle group moved an average distance of 1878 cm at an average speed of 3.1 cm/s, whereas the GV1001-injected groups covered average distances of: 2123 cm (0.1 mg/kg), 2573 cm (1 mg/kg), and 2607 cm (2 mg/kg), at average speeds of: 3.5 cm/s (0.1 mg/kg), 4.2 cm/s (1 mg/kg), and 4.3 cm/s (2 mg/kg)) (Table 4). The results showed that both the total distance covered (Fig. 1E) and the speed (Fig. 1F) increased significantly in the GV1001 (1 or 2 mg/kg)-injected groups as compared to the vehicle group, suggesting that GV1001 improved the locomotion and exploratory behavior of the TauP301L-BiFC mice.

Table 4.

Novel object recognition test.

Article	Dose (mg/kg)	Cumulative duration (old)	Cumulative duration (new)
Vehicle	N/A	0.443 ± 0.220	0.557 ± 0.220
LMTM	15	0.357 ± 0.140	0.643 ± 0.140*
GV1001	0.1	0.465 ± 0.228	0.535 ± 0.228
GV1001	1	0.321 ± 0.163	0.679 ± 0.163***
GV1001	2	0.390 ± 0.210	0.610 ± 0.210

The data represent the means ± S.D. and analyzed using two-way ANOVA with Sidak's multiple comparisonstest; *p<0.05, ***p<0.001 compared with vehicle group.

Restoration of cognitive dysfunction in GV1001-injected TauP301L-BiFC mice

A novel object recognition test was performed the day after the open field test to evaluate the effect of GV1001 on the cognitive function of the TauP301L-BiFC mice (Fig. 2). In the recognition phase, the search time for a familiar (old) or new object (new) was recorded for 10 min (Fig. 2A), and the cognitive memory function defined as the Recognition Index (RI), which was calculated by dividing the time spent investigating new or old objects by the total time spent exploring all objects during the test period (Table 5). No significant difference was observed in the vehicle group in terms of the time spent investigating the new object as compared to the old object, indicating that the cognitive memory function was impaired in these mice. However, a significant increase in the RI was observed in the GV1001 (1 mg/kg)-treated groups (Fig. 2B), with significantly longer time spent investigating the new object as compared to the old object (p < 0.001), suggesting that the cognitive function of the TauP301L-BiFC mice was improved by GV1001.

Fig. 2.

Behavioral test investigating cognitive function of GV1001-injected TauP301L-BiFC mice. A Schematic diagram showing novel object recognition test. Mice were initially exposed to two identical objects and one object was replaced 24 h later. Mice were then allowed to freely investigate the two different objects for a fixed time period. B The recognition index (RI), in which the novel object investigation was divided by total exploration time, was used to analyze the results. Data represent the mean and S.D. and were analyzed using two-way analysis of variance with Sidak’s multiple comparisons test; *p < 0.05, ***p < 0.001; n.s., not significant, compared with vehicle group.

Table 5.

Quantified Tau-BiFC fluorescence intensity for tau aggregation.

Article	Dose (mg/kg)	Tau-BiFC fluorescence intensity
		Somatosensory cortex	Motor cortex	Hippocampus (CA1)	Substantia nigra
Vehicle	N/A	65.8 ± 13.8	60.1 ± 14.2	85.7 ± 16.4	24.8 ± 10.4
LMTM	15	56.1 ± 13.5	44.0 ± 10.3*	67.7 ± 15.5*	18.4 ± 6.3
GV1001	0.1	47.1 ± 6.4**	46.7 ± 18.2	80.9 ± 14.3	14.1 ± 6.5*
GV1001	1	45.0 ± 12.3**	34.4 ± 10.8***	66.9 ± 12.9*	21.1 ± 8.1
GV1001	2	41.9 ± 11.7***	43.7 ± 12.6*	47.5 ± 16.5***	16.9 ± 6.4*

Quantifi ed fl uorescence intensity was statistically analyzed by one-way ANOVA with Dunnett's multiplecomparisons test; *p<0.05, **p<0.01, ***p<0.001 compared with vehicle group.

Reduced tau oligomerization and phosphorylation in GV1001-injected TauP301L-BiFC mice

Brain tissue sections of TauP301L-BiFC mice were prepared when the behavioral experiments were complete, and the results used to evaluate tau pathology (Supplementary data 2). TauP301L-BiFC mice showed the BiFC fluorescence, with a significantly increased in intensity in amygdala, thalamus and striatum at 12 months²⁶. Based on this result, substantia nigra was also observed to check the tau pathology in this study. Tau-BiFC fluorescence images can be used to show the degree of tau assembly, including tau oligomers and aggregates, while immunostaining with the AT8 antibody, which is the most widely used antibody for detecting phosphorylated tau protein at serine 202 and threonine 205, can be utilized to show the degree of tau pathology²⁸. Both tau aggregation (Fig. 3) and phosphorylation (Fig. 4) were measured and analyzed in four regions of the brain tissue: the somatosensory cortex, motor cortex, hippocampal CA1, and substantia nigra, with results indicating that the degree of Tau-BiFC fluorescence was decreased in the GV1001-injected groups as compared to the vehicle, with 28% (0.1 mg/kg), 31% (1 mg/kg), and 36% (2 mg/kg) reductions observed in the somatosensory cortex (Fig. 3A, B); 22% (0.1 mg/kg), 42% (1 mg/kg), and 27% (2 mg/kg) in the motor cortex (Fig. 3C, D); 5% (0.1 mg/kg), 21% (1 mg/kg), and 44% (2 mg/kg) in the hippocampus (CA1) (Fig. 3E, F); and 43% (0.1 mg/kg), 14% (1 mg/kg), and 31% (2 mg/kg) in the substantia nigra (Fig. 3G, H) (Table 6). The degree of AT8 antibody-positive phosphorylated tau was also reduced in the GV1001-injected group as compared to the vehicle, with results indicating reductions of 14% and 31% in the somatosensory cortex (Fig. 4A, B); 26% and 25% in the motor cortex (Fig. 4C, D); 21% and 27% in the hippocampus (CA1) (Fig. 4E, F); and 25% and 15% in the substantia nigra (Fig. 4G, H) as compared to the vehicle (Table 7). These results suggest reduced tau aggregation and phosphorylation in the brains of TauP301L-BiFC mice as a result of the GV1001 administration, indicating that GV1001 improved the tau pathology in the TauP301L-BiFC mice.

Fig. 3.

Decreased tau oligomerization and aggregation in the brains of GV1001-injected TauP301L-BiFC mice. BiFC images from each brain region, including the A somatosensory cortex, C motor cortex, E hippocampus, and G substantia nigra of TauP301L-BiFC mice in the vehicle-, LMTM-, and GV1001-injected groups. Scale bars: 100 μm and 50 μm. B, D, F, H Tau-BiFC fluorescence intensity was quantified using ImageJ and statistical analysis performed using one-way analysis of variance with Dunnett’s multiple-comparisons tests; *p < 0.05, **p < 0.01, ***p < 0.001, n.s., not significant compared with the vehicle group.

Fig. 4.

Decreased tau phosphorylation in the brains of GV1001-injected TauP301L-BiFC mice. AT8-immunostaining images from each brain region, including the A somatosensory cortex, C motor cortex, E hippocampus, and G substantia nigra of TauP301L-BiFC mice in the vehicle-, LMTM-, and GV1001-injected groups. Scale bars; 100 μm, 50 μm. B, D, F, H Quantification of AT8-immunofluorescence intensity. Fluorescence intensity was analyzed using ImageJ and statistically analyzed by one-way analysis of variance with Dunnett’s multiple-comparisons test; *p < 0.05, **p < 0.01, ***p < 0.001, n.s., not significant compared with the vehicle group.

Table 6.

Quantified AT8-immunofluorescence intensity for tau phosphorylation.

Article	Dose (mg/kg)	AT8-immunofluorescence intensity
		Somatosensory cortex	Motor cortex	Hippocampus (CA1)	Substantia nigra
Vehicle	N/A	48.8 ± 8.9	49.8 ± 9.9	52.7 ± 11.0	35.2 ± 11.0
LMTM	15	33.5 ± 11.2***	36.9 ± 12.3*	38.0 ± 9.2*	29.7 ± 10.7
GV1001	0.1	37.2 ± 6.9*	37.2 ± 6.9*	42.8 ± 11.8	24.7 ± 9.7
GV1001	1	31.1 ± 5.0***	36.8 ± 10.3*	34.0 ± 12.3**	24.0 ± 7.6
GV1001	2	30.4 ± 9.2***	25.0 ± 7.8***	34.5 ± 8.6**	21.6 ± 11.9*

Quantifi ed fl uorescence intensity was statistically analyzed by one-way ANOVA with Dunnett's multiplecomparisons test; *p<0.05; **p<0.01, ***p<0.001 compared with vehicle group.

Decreased tau phosphorylation and aggregation in brain extracts from GV1001-injected TauP301L-BiFC mice

To further confirm the effects of GV1001 on tau phosphorylation, protein lysates of whole brain region from the TauP301L-BiFC mice were analyzed by western blotting (Fig. 5A, F, Supplementary data 3), with results showing phosphorylated human tau (hTau) levels of 26% (0.1 mg/kg), 22% (1 mg/kg), and 43% (2 mg/kg) at residue S199 (Fig. 5B), and phosphorylated hTau levels of 32% (0.1 mg/kg), 32% (1 mg/kg), and 48% (2 mg/kg) at residue S396 (Fig. 5C) as compared to the vehicle. The levels of phosphorylated mouse tau (mTau) were also measured, with results showing 41% (0.1 mg/kg), 43% (1 mg/kg), and 54% (2 mg/kg) (Fig. 5D) at residue S199 in the brain lysates and 38% (0.1 mg/kg), 38% (1 mg/kg), and 49% (2 mg/kg) at residue S396 (Fig. 5E) as compared to the vehicle. The amount of RIPA-insoluble tau was also reduced, by 81% (0.1 mg/kg), 66% (1 mg/kg), and 70% (2 mg/kg) in the GV1001-administered group as compared to the vehicle group (Fig. 5F, G). These results indicate that GV1001 decreased tau phosphorylation and aggregation in the TauP301L-BiFC mice.

Fig. 5.

Decreased level of tau phosphorylation and aggregation in the brain extracts of GV1001-injected TauP301L-BiFC mice. A Immunoblot analysis using anti-phosphorylated tau (pS199 and pS396) in RIPA-soluble lysates from TauP301L-BiFC mice. Black arrows; human tau labeled with the BIFC compartments (VN173 or VC155), red arrows; endogenous mouse tau. Original blots are presented in Supplementary data 7. B–E Relative levels of phosphorylated human or mouse tau were quantified. F Immunoblot analysis using anti-Tau (Tau5) in RIPA-insoluble brain lysates from TauP301L-BiFC mice. G Relative level of aggregated tau was quantified using Image J and statistically analyzed by one-way analysis of variance with Dunnett’s multiple-comparisons test; *p < 0.05, **p < 0.01, ***p < 0.001, n.s., not significant, compared with vehicle group.

Synaptic loss reduction in GV1001-injected TauP301L-BiFC mice

Previous reports have indicated that 4R tauopathies, including PSP and corticobasal degeneration, are associated with increased 4R tau levels, the accumulation of hyperphosphorylated tau, neuroinflammation, and severe synaptic loss, which ultimately lead to neuronal death²⁹, and GV1001 has been reported to exhibit protective effects against oxidative stress and toxic amyloid beta in neural stem cells (NSCs) isolated from rodents^22,23. To confirm the neuroprotective function of GV1001 in human neuronal cells, SH-SY5Y cells were treated with hydrogen peroxide (H2O2), upregulating the expression of cleaved caspase-3, a marker representative of apoptotic neuronal death. The upregulation of cleaved caspase-3 was reduced when cells were pre-treated with GV1001 30 min before H2O2 treatment (Supplementary data 4). Similar to the results in primary neural stem cells and primary cultured neurons isolated from rodents, GV1001 showed neuroprotective functioning against oxidative stress in human neuronal cells, which has been observed in the brains of patients with PSP, indicating a crucial role in the pathogenesis of neurodegenerative diseases³⁰.

Postsynaptic density protein-95 (PSD-95) is a major postsynaptic scaffolding protein that plays a key role in bidirectional synaptic plasticity, and synaptophysin is an abundant presynaptic vesicle membrane protein involved in the localized retrieval and recycling of synaptic vesicles. Thus, to identify whether synaptic loss was relieved by GV1001 injection in the TauP301L-BiFC mice, the levels of both proteins were analyzed in whole brain extracts from the mice (Fig. 6A). Synaptic dysfunction is a key pathogenic event in brain network physiology, with reports indicating that pathological oligomeric tau leads to reduced synaptic plasticity and density^31,32. The obtained PSD-95 and synaptophysin levels showed patterns of decrease in the brain lysates of the TauP301L-BiFC mice as compared to wild-type mice, and both the GV1001-injected and LMTM-administered groups exhibited PSD-95 (Fig. 6B) and synaptophysin (Fig. 6C) recovery, with GV1001 injection more significant as compared to LMTM administration. These results indicate that GV1001 reduced the progressive synaptic loss related to the declining cognitive and motor functions in the TauP301L-BiFC mice.

Fig. 6.

Recovered synaptic loss in the brain lysates from GV1001-injected TauP301L-BiFC mice. A Representative images showing immunoblot analysis using anti-PSD95 and anti-synaptophysin antibodies as markers for post-synapse and pre-synapse, respectively, in the whole brain lysates of wild-type (WT) mice and TauP301L-BiFC mice administered with vehicle, LMTM, or GV1001. Original blots are presented in Supplementary data 8. Quantified graphs showing B PSD-95 and C synaptophysin levels. Data were analyzed using one-way analysis of variance with Tukey’s multiple comparisons test; *p < 0.05 compared with wild-type group; ^#p < 0.05, ^##p < 0.01 compared with vehicle group.

Reduced activated astrocyte and microglia levels as a result of GV1001 injection

GV1001 has been found to decrease neuroinflammation by modulating the activation of astrocytes and microglia in a 3xTg AD mouse model with accumulate amyloid plaques and tauopathy^24,25. In our previous studies, the subcutaneous injection of GV1001 successfully decreased the levels of amyloid plaques and tau aggregation, resulting in the recovery of recognition, learning, and memory functions in 3xTg mice. In addition, the enhanced activation of astrocytes and microglia, which play a key role in neuroinflammation, in the hippocampus and cortex was reduced in GV1001-injected 3xTg mice. Based on these studies, to determine whether neuroinflammation is decreased in GV1001-injected TauP301L-BiFC mice, the degree of activation in astrocytes and microglia was analyzed based on the GFAP and Iba1 expression levels, respectively (Fig. 7A), with results showing increased levels of GFAP and Iba1 in the whole brain lysates of TauP301L-BiFC mice as compared to wild-type mice, as expected. The TauP301L-BiFC group injected with GV1001 exhibited a statistically significant decrease in GFAP (Fig. 7B) and Iba1 (Fig. 7C) levels as compared to the vehicle group. The group administered LMTM, a positive control substance, also showed a decrease in the levels of GFAP and Iba1; however, this effect was less pronounced and not statistically significant as compared to the GV1001 groups (Fig. 7B, C).

Fig. 7.

Reduced levels of activated astrocyte and microglia as a result of GV1001 treatment. A Immunoblot analysis using anti-GFAP and anti-Iba1 antibodies to detect the activated astrocyte and microglia, respectively, in whole brain lysates of wild-type (WT) mice and TauP301L-BiFC mice administered with vehicle, LMTM, or GV1001. Original blots are presented in Supplementary data 9. Quantified graphs showing B GFAP and C Iba1 levels. Data were analyzed using one-way ANOVA with Tukey’s multiple comparisons test; *p < 0.05 compared with wild-type group; ^#p < 0.05, ^##p < 0.01, n.s., not significant compared with vehicle group. D Immunoblot analysis using total lysates of U-87MG cells stimulated with conditioned medium from LPS (1 µg/ml for 24 h)-treated HCM3 cells for 18 h and then GV1001 was post-treated for 6 h Original blots are presented in Supplementary data 10. Quantified graphs: E GFAP, F FKBP5, G C3 levels. Relative levels of GFAP, FKBP5, and C3 were quantified using Image J and statistically analyzed by one-way ANOVA with Tukey’s multiple comparisons test; *p < 0.05, ***p < 0.001 compared with non-stimulated group; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001 compared with MCM-treated group. (H) Immunoblot analysis of total lysates of HMC3 cells stimulated with IFN-γ + LPS. GV1001 was added 6 h later and post-treatment continued for 18 h. Original blots are presented in Supplementary data 11. Relative levels of I Iba1 and J COX2 were quantified using Image J and statistically analyzed by one-way analysis of variance with Tukey’s multiple comparisons test; *p < 0.05, **p < 0.01 compared with non-stimulated group; ^##p < 0.01 compared with IFN-γ + LPS-treated group.

Next, to determine whether the decrease in the activation of astrocytes and microglia observed in TauP301L-BiFC mice injected with GV1001 was due to direct regulation by GV1001 or a result of pathological recovery, in vitro study was performed using U87-MG cells, a human astrocyte cell line, and HMC3 cells, a human microglial cell line (Fig. 7D, H). To activate the U87-MG cells, microglia conditioned medium (MCM) was added to U87-MG cells over 24 h^33,34, and to enhance the activation of the cells, CM from 1 µg/mL LPS-stimulated HMC3 cells (MCM) was added, leading to the successful up-regulation of the GFAP level (Fig. 7E). Post-treatment with GV1001 was then performed 6 h after MCM stimulation, and analysis performed 18 h later. The results showed downregulation of the increased GFAP levels as a result of GV1001 treatment (Fig. 7E). In addition, the S100B, FKBP5, and C3 levels, which are representative markers for A1 reactive astrocytes in humans, were regulated by the post-treated GV1001 (Fig. 7F, G). Next, HMC3 cells were activated via IFN-γ (20 ng/mL) + LPS (1 µg/mL) treatment over 24 h (Fig. 7H), and to test the therapeutic effect of GV1001, GV1001 was post-treated 6 h after HMC3 cells were stimulated with IFN-γ + LPS, with analysis performed 18 h later. The results showed that the increased Iba1 levels were diminished by GV1001 treatment (Fig. 7I). In addition, the level of COX2, an M1-type microglial marker, was downregulated by GV1001 (Fig. 7J). These data indicate that GV1001 decreases neuroinflammation by directly modulating the abnormal activation of astrocytes and microglia. Furthermore, the results also suggest that GV1001 does not simply reduce the astrocyte and microglia activation, but is rather involved in more a subtle regulation of the polarization of the two cell types from the pro-inflammatory/neurotoxic type to the anti-inflammatory/neuroprotective type.

GV1001 down-regulates the protein level of 4R tau in an annonacin-induced in vitro PSP model

In addition to general functions such as neuroprotection and anti-neuroinflammation, which can be commonly applied to other neurodegenerative diseases, an in vitro PSP neuronal model was established using a differentiated SH-SY5Y cell line to determine whether GV1001 has functions that can be more specifically applied to PSP (Fig. 8A). For this investigation, SH-SY5Y cells were incubated with 50 ng/mL human BDNF (hBDNF) and 10 µM retinoic acid (RA) for six days to differentiate into neurons with generated neurites and then annonacin added to the differentiated SH-SY5Y cells over 48 h to induce 4R tauopathy. Annonacin, a lipophilic inhibitor of complex I in the mitochondrial respiratory chain (nicotinamide adenine dinucleotide [NADH]-quinone oxidoreductase), has been reported as an environmental model for PSP^35,36. Cultured neurons incubated with low doses of annonacin show increased 4R tau protein levels, hyperphosphorylation of tau, and somatodendritic distribution of tau in cells³⁷. To establish an annonacin-induced in vitro PSP model using SH-SY5Y cells, differentiated SH-SY5Y cells were treated with different doses of annonacin (25, 50, and 100 nM) for 48 h (Fig. 8B). As reported, the lowest dose (25 nM) of annonacin showed the highest increase in 4R tau levels as compared to those treated with higher doses (50 and 100 nM) of annonacin (Fig. 8C). The immunoblot for total tau protein as detected by the HT7 antibody showed no significant changes following annonacin treatment (Fig. 8D); however, a significant increase in the ratio of 4R tau to total tau was observed in the 25 nM annonacin-treated cells (Fig. 8E). To test the therapeutic effects of GV1001 on neuronal cells showing 4R tauopathy, GV1001 was post-treated for 24 h to the cells incubated with annonacin for 24 h (Fig. 8A, F). Interestingly, the upregulated 4R tau protein levels in the annonacin-treated cells were abolished by the GV1001 treatment (Fig. 8G). In addition, the increased ratio of 4R tau to total tau was completely diminished by GV1001 (Fig. 8I), whereas GV1001 had no significant effect on total tau levels (Fig. 8H). These data indicate that GV1001 has a fundamental therapeutic function for PSP by directly modulating 4R tau levels.

Fig. 8.

Regulation of 4R tau level by GV1001 in vitro and in vivo. A Schematic diagram for the in vitro PSP neuronal model. SH-SY5Y cells were incubated in 1% FBS media containing retinoic acid (RA) and BDNF to optimize neuronal differentiation for six days. To induce 4R tauopathy, annonacin (0, 25, 50, and 100 nM) treatment was applied for 48 h. B 4R tau and total tau protein levels were analyzed using anti-4R tau antibody (RD4) and anti-total tau antibody (HT7), respectively. Original blot for 4R tau is presented in Supplementary data 12. Quantified graphs showing C 4R tau, D total tau, and E the ratio of 4R tau, with total tau obtained from three biological repeats. Data represent the mean and S.D. and analyzed using one-way ANOVA with Tukey’s multiple comparisons test; *p < 0.05 vs. 0 nM annonacin-treated cells. F GV1001 (1 µM) was post-treated for 24 h to the neuronal cells incubated with 25nM annonacin for 24 h. Quantified graphs for G 4R tau, H total tau, and I the ratio of 4R tau with total tau are presented with three biological repeats. Data represent the mean and S.D. and analysis was performed using one-way ANOVA with Tukey’s multiple comparisons test; *p < 0.05 vs. Ctrl group and #p < 0.05 vs. +Ann group. J Immunoblot analysis of whole brain lysates from TauP301L-BiFC mice administered with the vehicle, LMTM, or GV1001. Original blot for 4R tau is presented in Supplementary data 12. Quantified graphs for K mouse 4R tau, L mouse total tau, and M the ratio of 4R tau with total tau are presented. Data were analyzed using one-way analysis of variance with Tukey’s multiple comparisons test; *p < 0.05 compared with vehicle group.

Endogenous level of 4R tau is down-regulated in TauP301L-BiFC mice injected with GV1001

Next, to confirm the regulatory function of GV1001 on 4R tau levels in vivo, brain extracts of TauP301L-BiFC mice were analyzed using an anti-4R tau antibody^38,40,40. Endogenous 4R tau levels in the brains of TauP301L-BiFC mice were downregulated in the GV1001-injected group (Fig. 8K), while no significant changes were observed in the level of total tau under treatment with the HT7 antibody (Fig. 8L). In addition, the 4R tau to total tau ratio was significantly reduced in the GV1001-injected TauP301L-BiFC mice (Fig. 8M). Mice that were given LMTM showed a tendency for decreased 4R tau protein levels and a lower ratio of 4R tau to total tau; however, this effect was not statistically significant (Fig. 8K, M). These in vivo data support the effect of GV1001 on the regulation of 4R tau level in 4R tauopathy.

Regulation of translocator protein (TSPO) level by GV1001

Many studies of patients with PSP have explored the association between clinical severity and neuroimaging markers of tau pathology, brain atrophy, and neuroinflammation^41,43,43. Subcortical neuroinflammation is associated with the clinical severity of PSP; in particular, molecular imaging with PET for microglial activation and tau pathology is expected to predict the clinical severity and progression of PSP⁴⁴. Several groups have reported PET imaging with radio ligands that bind to the 18 kDa translocator protein (TSPO) and TSPO PET has been used to visualize changes in neuroinflammation in patients with neurodegenerative diseases^45,47,47. To test whether TSPO PET can be used to monitor the pathological severity and progression in clinical trials of GV1001 for PSP treatment, changes in the TSPO levels as a result of GV1001 administration were analyzed through in vitro and in vivo studies, with results showing that HMC3 cells stimulated with IFN-γ + LPS for 24 h exhibited increased TSPO levels, which were then successfully abolished by GV1001 post-treatment, indicating that GV1001 directly regulates the TSPO levels in microglia (Fig. 9A, B). Furthermore, the amount of TSPO protein, which was increased compared to that of wild-type mice, was significantly reduced in TauP301L-BiFC mice injected with GV1001 (Fig. 9C, D, Supplementary data 6). The group administered with LMTM, a tau aggregation inhibitor that does not directly regulate neuroinflammation, showed a decreasing trend in TSPO levels but it was not statistically significant. These data support the hypothesis that GV1001 regulates neuroinflammation by modulating microglial activation, and that TSPO PET could be used in GV1001 clinical studies.

Fig. 9.

Regulation of TSPO levels by GV1001. A Immunoblot analysis using total lysates of HMC3 cells stimulated with IFN-γ + LPS. GV1001 was added 6 h later and post-treatment continued for 18 h. B Relative TSPO level was quantified using Image J and statistically analyzed by one-way ANOVA with Tukey’s multiple comparisons test; *p < 0.05 compared with non-simulated cells; ^##p < 0.01 compared with IFN-γ + LPS-treated cells. C Representative images showing immunoblot analysis of TSPO protein levels in the whole brain lysates of wild-type (WT) mice and TauP301L-BiFC mice administered with the vehicle, LMTM, or GV1001. Original blots are presented in Supplementary data 13. D Quantified graph for TSPO. Data were analyzed using one-way analysis of variance with Tukey’s multiple comparisons test; *p < 0.05 compared with wild type (WT) group; ^#p < 0.05, n.s., not significant compared with the vehicle.

Discussion

Our previous studies investigating the therapeutic effects of GV1001 in AD indicated that AD model mice that were injected with GV1001 showed decreased amyloid beta plaque accumulation and tau aggregation in the cerebral cortex and hippocampus of the brain²⁴. As an AD model, a triple-transgenic mouse model engineered by introducing the APPSwe, MAPTP301L, and PS1M146V transgenes, named 3xTg, was used. This AD model exhibited both amyloid plaques and neurofibrillary tau tangles, leading to impaired learning and memory⁴⁸. However, the tau pathology in PSP differs from that in AD at the clinical, neuroanatomical, and biochemical levels. Pathological tau forms are straight, unbranched filaments in PSP and paired helical filaments and tangles in AD^49,50. The conformational and filamentous forms of the tau aggregates in patients are related to specific pathologies distinctive of PSP. In addition, the tau inclusions in PSP show distinct pathological substrates such as globose NFT in the neurons, astrocytes (tufted astrocytes), and oligodendrocytes (coiled bodies)^51,53,53. In addition, amyloid beta and tau are reported to interact with each other rather than acting independently, accelerating the progression of neurodegeneration⁵⁴ and related studies were performed, which support the hypothesis that the synergistic interplay between amyloid beta and tau in Alzheimer’s disease⁵⁵. These studies indicate that animal model having only 4R tauopathy should be used to test the therapeutic effect of GV1001 although GV1001 improved the behavioral defect and the neuropathology in 3xTg mouse model. However, presently no animal model replicates the key anatomical and neuropathological progression that characterize PSP. Although the translational value of 4R TauP301L-BiFC mouse model is somewhat limited, this model is reported to provide advantages compared to other tau transgenic mouse models²⁶. TauP301L-BiFC transgenic mice with cDNA constructs of full-length human tau (2N4R) containing the P301L mutation were thus fused to the BiFC compartment (hTauP301L-VN173 and hTauP301L-VC155). The TauP301L-BiFC model showed tau oligomerization after three months, with significantly enriched BiFC fluorescence in the brain²⁶, and the TauP301L-BiFC mice exhibited motor, locomotor, and cognitive dysfunctions. In comparison to P301S tau transgenic mouse model, our model shows improved visualization of tau dynamics by conjugating BiFC to tau while it does not promote or delay insoluble tau aggregation²⁶.

LMTM, a tau aggregation inhibitor that blocks tau polymerization by trapping tau monomers, has completed one Phase II and two Phase III clinical trials for AD treatment. However, no significant reduction in functional decline has been observed in subjects with mild to moderate AD in the trials^56,57. Although LMTM failed to ameliorate cognitive decline in patients with AD, it is still undergoing clinical trials for the treatment of tauopathy, due to evidence suggesting that LMTM administration reduced tau pathology and improved the cognitive function in a TauP301L-BiFC mouse model⁵⁸. In this study, groups that were given LMTM, a positive control substance, also exhibited improved behavioral functions and decreased brain pathologies; however, the efficacy was less pronounced or not statistically significant that observed with GV1001 administration. Differences in the MoA may explain the distinct effects observed for the two substances. Unlike LMTM, which solely inhibits tau aggregation, GV1001 provides therapeutic benefits through multiple mechanisms including neuroprotection, restored mitochondrial function, and reduced oxidative stress and neuroinflammation, indicating has a fundamentally different MoA in PSP pathology as compared to LMTM. Lee et al. recently reported a fluorescent phenotyping method that can be used to distinguish the effects of drug treatment and understand the MoA of various drugs⁵⁹. Analysis of blood samples collected from LMTM-administered and GV1001-injected TauP301L-BiFC mice using fluorescent probes showed that samples from the LMTM- or GV1001-treated groups formed separate clusters, supporting the notion that GV1001 has an MoA distinct from that of LMTM.

Gunter et al. observed that annonacin induces neuropathological abnormalities in the basal ganglia and brainstem nuclei⁶⁰, with significant loss of neurons observed in the substantia nigra and striatum, accompanied by increased astrocyte and microglia activation. In addition, it was found that annonacin induced the retrograde transport of mitochondria and redistribution of tau protein from axons to the cell body. They demonstrated that annonacin-induced ATP depletion led to mitochondrial retrograde transport and altered the intracellular distribution of tau protein³⁷. In 2014, Bruch et al. also found that annonacin and MPP+, two prototypical mitochondrial complex I inhibitors, increase 4R tau protein in human neurons, whereas no significant change is observed in the relative levels of 3R tau and total tau detected using an HT-7 antibody⁶¹. SRSF2, a splicing factor, has been identified to be upregulated in annonacin-treated neurons, resulting in increased 4R tau mRNA levels. However, the mechanism by which annonacin-induced ATP depletion increases the mRNA level of 4R tau through SRSF2 upregulation has not been elucidated.

Tau mutations related to splicing efficiency are located in either exon 10 or intron 10, and result in a shift in the ratio of mRNA levels between 4R and 3R tau^62,63. Interestingly, it has been reported that increased levels of intracellular tau protein can affect microtubular functioning and lead to tau aggregation even in the absence of mutations in the tau gene^64,66,66. Significantly increased 4R tau levels in the frontal cortex and caudate have been reported in association with 4R-tau-dominant pathology in patients with PSP. However, no differences in 3R tau levels have been observed between PSP and control brains^67,68. In terms of PSP therapy strategies, antisense oligonucleotides and splicing modulators have been developed to block 4R tau protein expression^1,69.

Retrograde transport in the axons is essential for mitochondrial repair and clearance. Damaged mitochondria move in a retrograde direction in axons that are to be degraded by lysosomes in the soma of neurons^70,71. The transport of cargo within cells occurs in two directions: anterograde transport, mediated by kinesin, and retrograde transport, mediated by dynein. Both kinesin and dynein motors interact with MT, and this interaction is potentially affected by the relative content of 3R and 4R tau isoforms because tau modulates axonal transport by direct interaction with molecular motors^72,74,75. An imbalance toward the 4R tau isoform promotes the retrograde movement of cargo, while the predominance of 3R tau favors anterograde transport⁷⁵. It is possible that the expression level of 4R tau is enhanced for transporting dysfunctional mitochondria in a retrograde direction to clear and repair annonacin-treated neurons although this should be proved by in-depth studies. GV1001 restores impaired mitochondrial membrane potential (MMP) and ATP production in damaged mitochondria in neuronal cells treated with hydrogen peroxide or toxic beta-amyloid^21,23,23. Based on previous studies and the present findings, we hypothesize that GV1001 may reduce pathological 4R tau levels in association with restoration of mitochondrial dysfunction under neuropathological conditions. However, the precise molecular mechanisms linking mitochondrial recovery to selective regulation of 4R tau were not directly addressed in this study.

Recently, we reported the results of GV1001 in a Phase 2a clinical trial evaluating a drug for PSP (NCT05819658), with the results showing sufficient promise for proceeding to Phase 3 trials. Participants in all groups (0.56 mg or 1.2 mg dose of GV1001, and the placebo) experienced disease progression; however, a 48% reduction was observed in the progression of the 0.56 mg group. Furthermore, when reviewing the differences in the scores for the different PSP subtypes, patients with PSP-Richardson syndrome (PSP-RS) had significantly lower PSP rating scale scores than the control group. Notably, the PSP-RS participants experienced stabilization or even improvement in their symptoms during the trial. Additionally, the safety data for GV1001 were consistent with previously gathered data, with PSP researchers noting no serious adverse effects and GV1001 generally well tolerated by participants.

Several limitations of the present study should be acknowledged. First, although the TauP301L-BiFC mouse model provides a robust platform to investigate 4R tau aggregation and related functional impairments relevant to 4R tauopathies, it may not fully capture all anatomical and pathological features of human progressive supranuclear palsy (PSP). Accordingly, the present findings should be interpreted as providing mechanistic insights into 4R tauopathy models with relevance to PSP, rather than as a direct representation of the full spectrum of PSP pathology. Second, while tau pathology was observed in the substantia nigra region, dopaminergic neuron–specific validation such as tyrosine hydroxylase (TH) co-staining was not performed due to limited tissue availability. As a result, the cell-type specificity of tau pathology within this region could not be definitively determined. Third, synaptic alterations were assessed using synaptic marker expression in whole-brain lysates, which limits region-specific and structural interpretation at the cellular level. Finally, behavioral and biochemical analyses were not conducted under blinded conditions, which should be considered when interpreting the results.

In summary, we demonstrated the therapeutic efficacy of GV1001 in 4R tauopathy using in vitro and in vivo models. Subcutaneous injection of GV1001 ameliorated motor dysfunction and tau-related pathologies in 4R TauP301L-BiFC mouse model. Furthermore, we discovered a novel and critical efficacy of GV1001 to reduce 4R tau level induced by annonacin, an environmental model for PSP.

This study provides important insights into the therapeutic effects of GV1001 in 4R tauopathy models and supports its potential relevance to 4R tau–driven neurodegenerative disorders, such as PSP.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

KBJ, MHL, DMK, N. Lukianenko: conducted the experimentsYKK, SSL: provided the experimental materials, supervised the data collection from in vivo experiment and reviewed the manuscriptSJK: reviewed the manuscript, provided oversight.HJC: conceived the study, designed the experiments, and wrote the manuscript. Provided oversight and management of the project.

Funding

No funding.

Data availability

The data from this study is available upon a reasonable request from Dr. Cho [hjcho@gemvax.com](mailto: hjcho@gemvax.com).

Declarations

Competing interests

Declaration of Competing Interest. The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: SJK holds equity in GemVax&KAEL Co., Ltd. and KBJ, MHL, and HJC are employee of GemVax. Other authors declare no competing interests.