Bi-directional regulation of AIMP2 and its splice variant on PARP-1-dependent neuronal cell death; Therapeutic implication for Parkinson's disease

SK-SH5Y, SK-N-SH, and N2A cells were obtained from the Korean Cell Line Bank (KCLB, Seoul, Korea). Cells were grown in RPMI-1640, and supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin (HyClone, PA, USA). The transient transfection of myc-tagged or GFP-tagged parkin, AIMP2, DX2, HA-tagged ubiquitin, and empty vector was performed using lipofectamine 2000 (Invitrogen, CA, USA). Propidium iodide solution (PI), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetra-zolium bromide (MTT), rotenone, and 6-OHDA were obtained from Sigma-Aldrich (St. MO, USA). siRNA targeting DX2 (CACGUGCAGGAUUACGGGGC) was purchased from Bioneer (Seoul, Korea).

To prepare primary neurons, embryonic brain cortices at E18-E19 were removed and transferred to new conical tubes and gently washed three times with PBS. The washed cortices were incubated in papain solution containing 3.5 mg/ml papain (0.5 units/g; Wako, VA, USA), 0.5 mg/ml EDTA-disodium salt (Wako, VA, USA), and deoxyribonuclease I (DNase I, 5 units/mL; Takara, Shiga, Japan) for 15 min. The digested cortices were dissociated by pipetting 12 times with a glass Pasteur pipette and filtered with a cell strainer (40-mm mesh; BD Biosciences, NJ, USA). The harvested neuronal cells were cultured in MEM containing 20% FBS and N2 supplement (Thermo Fisher Scientific, MA, USA).

Cells were detached using Trypsin/EDTA solution and subjected to centrifugation at 300 g for 5 min. A portion of the collected cells were then sampled and suspended in a 1:1 mixture of media and trypan blue solution (Cat no. 15250061, Thermo Fisher, CA, USA). The cell suspension was subsequently loaded into a cell counting device (Cat no. AMQAF2001, Countess 3 Autometated cell counter, Thermo Fisher, CA, USA) following the device process. All cell counting experiments were conducted in triplicate for quantification.

HEK293 cells transfected with myc-tagged AIMP2 or DX2 were then incubated in a methionine-free medium for 1 h. Then, [³⁵S] methionine (50 μCi/ml) was added and incubated for 1 h. After the radioactive methionine was washed off with fresh medium, targeted proteins AIMP2 and DX2 were immunoprecipitated with the myc antibody (SC, CA; USA Cat. #sc-40), separated by 12% SDS-PAGE, and subjected to autoradiography using a BAS scanner (FLA-3000; Tokyo, Japan).

Hippocampal CA3-CA1 regions from (0–1 day-old) Sprague Dawley rats (DBL; Strain code: NTac:SD) were dissected, dissociated, and plated onto poly-ornithine-coated coverslips inside a 6-mm-diameter cylinder. For live imaging of synaptic physiology, the plasmids (vGlut1-pHluorin with strep-DX2 or strep-AIMP2) were transfected 8 days after plating. Experiments were conducted 14–21 days after plating. All animal experiments were performed with the approval of the Institutional Animal Care and Use Committee of Kyung Hee University. The plasmids were incubated with 2 mM Ca²⁺, 2 × HeBS (273 mM NaCl, 9.5 mM KCl, 1.4 mM Na₂HPO₄.7H₂O, 15 mM D-glucose, 42 mM HEPES), and the mixture was transfected to hippocampal neurons at DIV 8 using the Ca²⁺ phosphate precipitation method.

For synaptic function imaging, we utilized a pHluorin-based assay. Coverslips with neurons were mounted in a laminar-flow-perfusion system of a metal chamber on the stage of a custom-built, laser-illuminated epifluorescence microscope. Live-cell images were acquired using an Andor iXon Ultra 897 (Model #DU-897U-CS0-#BV) back-illuminated EM CCD camera with a diode-pumped OBIS 488 laser (Coherent), shuttered by TTL on/off signal from the EMCCD camera during imaging. Fluorescence excitation/emission and collection were achieved using a 40× Fluor Zeiss objective lens (1.3 NA) with 500–550 nm emission and 498 nm dichroic filters (Chroma) for pHluorin. Action potentials were caused by passing a 1 ms current pulse through platinum-iridium electrodes using an isolated current stimulator (World Precision Instruments). Neurons were perfused in saline-based Tyrode’s buffer consisting of 119 mM NaCl, 2.5 mM KCl, 2 mM CaCl₂, 2 mM MgCl₂, 25 mM HEPES, 30 mM glucose, 10 μM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), and 50 μM D,L-2-amino-5-phosphonovaleric acid (AP5), adjusted to pH 7.4. All live presynaptic terminal imaging was conducted at 30 °C; all images were acquired at 2 Hz with a 50-ms exposure.

After washing the prepared cells with cold PBS, cell lysate was collected with 1% Trition × 100 lysis buffer diluted in PBS after incubation at 4 °C for 20 min. The 500 μg of protein in the supernatant was incubated with 1 μg antibody for pull down for 2 h and then incubated with 25 μl protein A/G plus-agarose beads (sc-2003, SC; CA, USA) at 4 °C overnight. After washing the sinked bead with PBS-T (0.05% tween in PBS), 2 × sample buffer was mixed with the bead. For denaturation, the samples were boiled for 10 min. Pull down was performed with PARP-1 antibody, c-Myc antibody for Parkin, and Flag antibody for AIMP2 and DX2.

The prepared protein supernatant was loaded into a 10% acrylamide gel for Western blotting electrophoresis, before transferring the protein from the gel to a PVDF membrane; a 5% skimmed milk solution was then used for blocking. The membrane was incubated in primary antibody diluted with Tris-buffered saline (TBS) with 0.05% Tween (TBS-T) for 2 h. After three washes with TBS-T, the diluted secondary antibody was incubated for 1 h. Detection was performed after luminol (SC; CA, USA). Antibodies against c-myc (sc-40), Ha-tag (sc-7392), PARP-1 (sc-8007), GFP (sc-101525), PCNA (sc-56), and α-tubulin (sc-5286) were purchased from Santa Cruz (SC, CA, USA) and tyrosine hydroxylase antibody (p40101-150) was purchased from Pel-Freez (AR, USA). Poly-(ADP-ribose) (83732s), lamin A/C (2032s), AIF (4642s) and YY1 (46365s) antibodies were purchased from Cell Signalling (MA, USA). For detecting basal AIMP2 and DX2 expression, NMS-01-0011 (curebio, Seoul) was used. All experiments were 3 times repeated independently.

All mice experiments were performed under the Kyung Hee University Institutional Animal Care & Use Committee guidelines (KHUASP(SE)-18-101).

To confirm the anti-neurodegenerative role of DX2 in vivo, chicken-β-actin promoter induced whole body DX2 transgenic animals were provided by Dr. Sunghoon Kim (Seoul National University, South Korea). Male C57Bl/6n-based DX2 transgenic mice and wild type mice were used. Since 8 weeks of age, fresh rotenone solution was orally administrated at a concentration of 30 mg/kg by gavage once a day for 4 weeks to wild type and DX2-TG mice. Rotenone (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in 4% carboxymethylcellulose (CMC, Sigma-Aldrich) with 1.25% chloroform (Beijing Chemical Works, Beijing, China). During rotenone treatment, the rotarod behaviour test was performed once a week, and the latency to fall was measured. At 4 weeks after chemical administration, pole test was performed to measure the turning time above top pole and descending time from top to bottom. To confirm this dopaminergic neuron survival in the substantial nigra, immunohistochemistry staining was performed using tyrosine hydroxylase antibody with cryo-sectioned brain.

To produce AAV-GFP and AAV-DX2, AAV-GFP and AAV-DX2 were generated by ligating oligonucleotides encoding the GFP or DX2 sequence into the pSF-AAV-ITR-CMV-ITR-KanR (OXGENE, Oxford, UK) vector. AAV viral vectors were prepared and transfected with AAV2 helper (OXGENE, Oxford, UK), AAV2 Rep-Cap (OXGENE, Oxford, UK), and pSF AAV-ITR-DX2 or pSF AAV-ITR-GFP in HEK293 cells. Media and cells were then collected and lysed.

To produce recombinant AAV serotype 2 (rAAV2) encoding DX2 or GFP at 40 L (20 L × 2) scales, one day before transfection of plasmid vectors, HEK 293 cells were counted (8 × 10⁵ cells/ml) and centrifuged at 300 × g for 7 min. The cell pellet was resuspended in HEK293 medium supplemented with 4 mM ultraglutamine and incubated for 24 h. Then, the DNA-PEIproHQ (transfection reagent) complexes (transfection mix) were generated according to the manufacturer’s recommendations. Briefly, two mixtures were generated in culture medium: (1) the plasmid DNA mix (scAAV2 DX2 or GFP–4 mg, rep/cap plasmid–7.2 mg, helper plasmid–8.8 mg, total 20 mg of DNA per 1 L of HEK293 medium with final 4 mM ultraglutamine) and (2) the PEIproHQ mix (60 mg of PEI per 1 L of HEK293 medium with final 4 mM ultraglutamine). These were combined and incubated for 10–15 min at room temperature. Finally, a volume equating to 10% of the production volume was added to each of the wave bags to be transfected (2 L of DNA-PEIproHQ complexes was added to 18 L of culture). Three days post-transfection, the cells were harvested, lysed with lysis buffer containing 300 mM NaCl, 50 mM Tris–Cl, pH8.0, and 0.1% PF68. To remove the residual DNA, the lysate was incubated for 1 h at 37 °C after adding a nuclease. Then, we purified the AAV product using tangential flow filtration (TFF) and affinity chromatography, and the neutralized affinity chromatography eluate material was solubilized in Tris-based buffer solution. In consideration of the effect of the AAV formulation on safety, the Tris-based buffer was changed to PBS-based buffer (PBS with 300 mM NaCl and 0.001% PF-68) by the 2nd TFF diafiltration to produce high-titer virus vector stock. Part of the formulated (diafiltrated) viral particles were concentrated using 100 kDa Amicon Centricon™ and the concentrated products were sterilized by filtration using a 0.22 μm syringe filter (Satorius). A final product (post purification) viral genome (VG) titer of 4.05 × 10¹³ VG (1.62 × 10¹³ VG/mL, by qPCR) was obtained with this process.

AAV-GFP or -DX2 was administrated to N2A and SK-N-SH cells. After 3 days, RNA was isolated using TRIzol Reagent (Thermo Fisher Scientific Inc., MA, USA). For control and test RNAs, the construction of the library was performed using QuantSeq 3′ mRNA-Seq Library Prep Kit (Lexogen Inc., VIE, Austria) according to the manufacturer’s instructions. QuantSeq 3′ mRNA-Seq reads were aligned using Bowtie2 [37]. Bowtie2 indices were either generated from genome assembly sequence or the representative transcript sequences for aligning to the genome and transcriptome. The alignment file was used for assembling transcripts, estimating their abundances, and detecting differential expression of genes. Differentially expressed genes were determined based on counts from unique and multiple alignments using coverage in Bedtools [44]. The RC (Read Count) data were processed based on the quantile normalization method using EdgeR within R using Bioconductor [21]. Gene category was based on searches done in DAVID (http://​david.​abcc.​ncifcrf.​gov/​) and Medline databases.

All experiments in mice were performed under the Kyung Hee University Institutional Animal Care & Use Committee guidelines (KHUASP(SE)-18-101). In the 6-OHDA group, male c57bl/6n mice aged 8 weeks were injected with 4 μg/μl of 6-OHDA in the right striatum. Each mouse was ethically anesthetized using ketamine and a muscle relaxant (Rompun). The anesthetized mice were then placed on the stereotaxic device with tooth and ear bars. The total volume of 6-OHDA per mouse was 3 μl. The injection was performed using a Hamilton syringe at the following coordinates: AP: + 0.5 mm, ML: + 1.8 mm, DV: − 3.7 mm. To evaluate the efficacy of DX2 in the 6-OHDA mice model, AAV-DX2 and its control virus AAV-GFP were injected on both sides of the lesion. The injection was performed using a Hamilton syringe at the following coordinates: coordinates 1: AP: + 1.2 mm, ML: + 1.5 mm, DV: − 3.5 mm; coordinates 2: AP: − 0.1 mm, ML: + 2.1 mm, DV: − 3.2 mm. All stereotaxic injections were performed at a rate of 0.5 μl/min and the needle was removed out of place for another 10 min after the injection before slowly being removed.

Every mouse in the 6-OHDA model group was behaviourally tested by using a cylinder with a total diameter of 11 cm and height of 20 cm. The blind test method was used through all experiments and analysis. Behaviour tests were conducted at different points after the injection of AAV.

The mouse was placed on top of a pole, and the duration for which the mouse stayed on the pole was measured. Since mice with deficits in motor activity will likely fall off the pole and mice with improved motor activity will descend the pole, an increase in the time in pole descent indicates an improvement in motor activity

To examine the motility of the rotenone mouse models, a rotarod test was conducted at 3 and 15 days post lesion. All mice were trained on the same machine with equal rotating speed to show a stable performance. Training procedures consisted of three sessions on 2 days in which each session included three individual trials, lasting at least 200 s each. All mice were trained at 5, 10, and 15 rpm (revolutions per minute). The final test was conducted at 15 rpm (three sessions, 300 s each) on the third day. To reduce fatigue, each mouse was given at least 5 min of rest between each measurement.

At the end of the behavioural assessment, the total duration was 12 weeks for the 6-OHDA model. Mice were sacrificed after being anesthetized with isoflurane and perfused with 0.9% saline mixed with 4% paraformaldehyde (PFA; Yakuri Pure Chemicals) in phosphate buffer (PBS) at a pH of 7.4. Organs including the lungs, heart, liver, kidney, ovary, spleen, bladder, spinal cord, and brain were collected. The brains were fixed overnight in 4% PFA followed by a dehydration step with 30% sucrose in PBS. Brain samples were then frozen with optimum cutting temperature compound (O.C.T. compound; Tissue-Tek) and cut on a cryotome (Leica CM1850; Leica Biosystems) into 30 μm thick coronal sections through the entire substantia nigra and striatum. All sections were made from the posterior end of the SN to the anterior end of the striatum. The sections were stored at − 20 °C in a cryoprotectant solution made with ethylene glycol, glycerol, and 0.2 M phosphate buffer.

For the staining, the floating brain tissues in cryoprotectant solution or attached cells on a coverslip were moved to a slide glass. After washing with PBS three times for 5 min, blocking with 5% BSA in PBS was performed for 30 min followed by permeabilization with 0.2% triton × 100 in PBS. The primary tyrosine hydroxylase antibody was 1:500 diluted with 1% BSA in PBS, applied to the section/cells, and incubated at room temperature for 1 h. After three washes with PBS-T (0.1% tween in PBS) for 5 min, the secondary antibody diluted with 1% BSA in PBS was incubated in a dark room for 1 h. Mounting with VECTASHILD antifade mounting medium with DAPI was used for nucleus counterstaining.

Images were analyzed with Image J (http://​rsb.​info.​nih.​gov/​ij) using the Time Series Analyzer plugin, accessible at https://​imagej.​nih.​gov/​ij/​plugins/​time-series.​html. Synaptic boutons of neurons were appointed as oval regions of interest (diameter, 10 pixels), and the amplitude of pHluorin-based fluorescence at synapses was counted using Origin Pro 2020. The kinetics of endocytosis was measured using a single exponential decay function.

The p values are represented for mean ± SEM and data were statistically analyzed using Student's t-test or ANOVA, where appropriate. A p-value of less than 0.05 was considered to be statistically significant.

In our previous study, we demonstrated that while AIMP2 binds to TRAF2 and p53, inducing apoptosis, DX2 counteracts the apoptotic function of AIMP2 by obstructing its binding to TRAF2 and p53, thereby masking the interaction site of AIMP2 [13]. Moreover, it has been reported that wild-type Parkin1 interacts with AIMP2, resulting in its degradation. However, mutant Parkin1 fails to bind AIMP2, which is one of the molecular mechanisms underlying mutant Parkin1-induced parkinsonism. Interestingly, DX2 displayed a comparable binding affinity to AIMP2 for Parkin1 interaction (Fig. 1a), and both AIMP2 and DX2, as Parkin substrates, were degraded at similar levels (Fig. 1b). Additionally, under conditions treated with MG132 proteasome inhibitor, parkin not only ubiquitinated AIMP2 but also DX2 similarly (Fig. 1c and Additional file 1: Fig. S1). Considering that protein stability may affect binding opportunities to a target molecule, we compared the stabilities of AIMP2 and DX2. The pulse-chase assay revealed that DX2 maintained a similar stability to full-length AIMP2 for up to 4 h after halting de novo protein synthesis with cycloheximide (Fig. 1d). AIMP2 and DX2 are known to translocate into the nucleus in response to cellular stress [39]. Under hydrogen peroxide (H₂O₂)-treated conditions, DX2 appeared to localize in the nucleus more rapidly than AIMP2 (Fig. 1e). It has been previously demonstrated that AIMP2 interacts with PARP-1, leading to aberrant neuronal cell death in PD [38]. Consequently, we investigated whether DX2 also binds to PARP-1. In a pull-down assay with PARP-1, the binding affinity of DX2 to PARP-1 was found to be stronger than that of AIMP2 (Fig. 1f), suggesting that DX2 has the potential to modulate neuronal cell fate in contrast to AIMP2 during PARP-1 activation.

Taken together, our results show that both AIMP2 and DX2 can be ubiquitinated and degraded in the cytoplasm as common substrates of Parkin (upper panel of Fig. 1g). However, DX2 responds to reactive oxygen species (ROS) signals and predominantly localizes in the nucleus, potentially having more opportunities to bind to PARP-1 (lower panel of Fig. 1g).

To determine whether DX2 acts as a competitive inhibitor of AIMP2, suppressing AIMP2-induced PARP-1 activation and neuronal cell death, we transfected neuroblastoma cells with DX2-plasmid vectors and -siRNA, and subsequently measured the binding of AIMP2 to PARP-1. We found that AIMP2-PARP-1 binding was reduced in DX2-overexpressing cells (Fig. 2a) and increased in DX2-knockdown cells (Fig. 2b; Additional file 1: Fig. S2), indicating that DX2 expression levels play a critical role in AIMP2-PARP-1 interactions. Given that PARP-1 activation occurs upon oxidative stress and its cleavage serves as a marker for apoptosis, we monitored PARP-1 cleavage in cells transfected with empty vector (EV), AIMP2, or DX2, both in the absence and presence of hydrogen peroxide (H₂O₂). AIMP2-transfected cells exhibited increased PARP-1 cleavage under oxidative stress conditions compared to other transfected cells (Fig. 2c, fourth lane). Notably, PARP-1 cleavage was not observed in DX2-transfected cells (Fig. 2c). Since PARylation is a post-translational process indicative of PARP-1 activation, we investigated the effects of AIMP2 or DX2 on PARylation. AIMP2 enhanced PARylation in the presence of H₂O₂, whereas DX2 had no effect (Fig. 2d). Next, to determine whether the PARylation are actually involved in parthanatos, we examined the nucleus translocation of AIF, one of the markers of parthanatos under H₂O₂-treatmed condition. Expectedly AIMP2 overexpression induced more nucleus localization of AIF and reversely DX2 seemed to suppress the translocation of AIF (Fig. 2e). Considering that AIMP2 tends to form dimers [42], we sought to determine if AIMP2-DX2 heterodimers exist and whether homo- or heterodimer formation impacts PARP-1 activation. To assess the binding affinities of AIMP2 and DX2 and the likelihood of heterodimer formation, we employed two different tags, green fluorescent protein (GFP) and myc, on each gene. Pull-down assays with myc-tag on transfected cell lysates revealed the formation of AIMP2 homodimers and AIMP2-DX2 heterodimers (Fig. 2f, first and second lanes of left upper panel). However, the amount of DX2 homodimers was considerably lower (Fig. 2f, third lane of upper left panel) compared to AIMP2-AIMP2 homodimers or AIMP2-DX2 heterodimers. Immunoprecipitation with GFP tag yielded a similar binding pattern between AIMP2 and DX2 as observed in Fig. 1K when immunoprecipitated by myc tag in the opposite direction (Fig. 2g).

Collectively, our findings suggest that when DX2 is insufficiently expressed, AIMP2 forms homodimers, interacts with PARP-1, activates PARP-1, and induces neuronal cell death. However, as DX2 exhibits a significantly higher binding affinity for PARP-1 than AIMP2, DX2-expressing cells display reduced AIMP2-PARP-1 binding, leading to the inhibition of PARP-1 activity and neuronal survival. These results indicate that DX2 is a crucial factor for neuronal cell viability under PARP-1 activating conditions, and that DX2 overexpression may ameliorate neuronal cell death in neurodegenerative diseases (Fig. 2h).

Drawing upon these findings, we propose that DX2 acts as an inhibitory molecule in oxidative stress-induced PARP-1 activation. We investigated the influence of AIMP2 and DX2 on cell death in neuronal cells. Plasmids expressing either AIMP2 or DX2 were introduced into SH-SY5Y neuroblastoma cells. Subsequently, apoptosis was induced through ROS stress by treating the cells with H₂O₂. Under normal conditions, overexpression of AIMP2 or DX2 had little impact on cell viability; however, in H₂O₂-treated conditions, AIMP2 overexpression significantly decreased cell viability, while DX2 overexpression mitigated this effect (Fig. 3a). When cells were treated with 6-OHDA, a selective neurotoxin targeting dopaminergic neurons, AIMP2 administration induced cell death in a statistically significant manner compared to untreated cells. Intriguingly, DX2 more potently suppressed cell death relative to 6-OHDA-treated conditions (Fig. 3b).

Subsequently, we examined the role of AIMP2 and DX2 in regulating cell viability under H₂O₂-induced cell death conditions. Even with increased AIMP2 expression (1–2 μg), a 1 μg transfection of DX2 diminished AIMP2/ H₂O₂-induced cell death to a level comparable to that achieved with 2 μg of DX2 (Fig. 3c). Conversely, AIMP2 expression levels had no significant effect on neuronal cell death in DX2-expressing cellular conditions (1–2 μg) (Fig. 3d). Consequently, DX2-expressing cells exhibited substantially reduced cell death compared to control cells under oxidative stress conditions. Collectively, these findings indicate that the anti-apoptotic effect of DX2 is considerably more potent than the pro-apoptotic activity of AIMP2 when both proteins are co-expressed.

We hypothesized that DX2 expression confers a protective effect on synaptic functionality in a neurodegenerative model system. To validate this notion, we employed a pHluorin-based assay, which is widely used for monitoring synaptic functionality, including synaptic transmission and synaptic vesicle endocytosis [2, 34]. Primary hippocampal neurons were transfected with vGlut-pHluorin in combination with either AIMP2 or DX2, and activity-driven synaptic transmission was monitored in the presence of 6-OHDA. Treatment with 6-OHDA resulted in a significant decrease in synaptic transmission to 80% at 60 min compared to control neurons, with AIMP2-overexpressing neurons exhibiting an even greater reduction in synaptic transmission, as evidenced by a loss of bouton signal (Fig. 3e–g). In contrast, DX2-overexpressing neurons demonstrated no deficits in synaptic transmission and, in fact, enhanced synaptic function. To evaluate endocytic activity, which reflects neuronal damage, we indirectly measured endocytosis by monitoring changes in the exocytic signal using the declining curve from the peak. In synaptic vesicle retrieval, AIMP2-overexpressing conditions displayed lower activity than control conditions, while DX2-overexpressing neurons exhibited a clear protective effect in the presence of 6-OHDA (Fig. 3h and i; for example, compare with the y-axis value of the red curve at 30 s), suggesting that DX2 promotes neuronal survival in neurotoxic environments.

To assess whether AAV-DX2 treatment exhibits the same anti-apoptotic properties as DX2 overexpression, we exposed a variety of cell types, including primary neurons, mouse embryonic fibroblasts (MEF), hepatocytes, mesenchymal stem cells (MSC), and neuroblastoma cells to AAV-DX2. Notably, AAV-DX2-transduced cells consistently demonstrated a marked reduction in cell death compared to AAV-GFP-transduced cells following H₂O₂ treatment (Fig. 4a), suggesting that AAV-DX2 serves as an effective anti-apoptotic intervention. Intriguingly, under basal conditions, cell viability remained unaffected by the presence or absence of AAV-DX2, indicating that DX2 operates exclusively in response to stress-induced apoptotic stimuli (Fig. 4b).To elucidate the therapeutic mode of action and anticipate potential unexpected responses, we treated SK-N-SH human neuroblastoma cells with AAV-DX2 and performed RNA-seq analysis. Differentially expressed genes (DEGs) were classified into signaling pathways or ontology databases, revealing a common involvement in suppressing p53-associated cell death pathways and TNF-α or interleukin-related signaling (Fig. 4c). These findings corroborate previous knowledge of AIMP2 and DX2-related pathways and display a predominance of downregulated genes (Fig. 4d). By characterizing each gene, we aimed to understand its contribution to cell death or survival (Fig. 4e, Additional file 1: Table S1). Statistical significance was visualized through a p-value distribution graph for all pathways following DX2 overexpression (Fig. 4f).

In summary, AAV-DX2-based gene therapy is anticipated to downregulate inflammation or oxidative stress-induced neuronal apoptosis-related genes while upregulating neurosurvival-associated genes, highlighting its potential as a promising therapeutic approach.

To determine the importance of DX2 expression in Parkinson's disease (PD) treatment, we generated whole-body DX2-overexpressing transgenic mice (DX2-TG). Additionally, there is are no different life span between WT and DX2-TG mice (Additional file 1: Fig. S3) We administered rotenone to both control and DX2-TG mice to examine its effects on parkinsonism. Four weeks after rotenone treatment, we conducted rotarod and pole tests and analyzed the mice's behavior (Fig. 5a). We observed a reduction in motor coordination in rotenone-treated wild-type mice; however, the rotenone-treated DX2-TG mice demonstrated significant resistance to the decline in motor activity (Fig. 5b). Additionally, we assessed the expression of tyrosine hydroxylase (TH), a marker for dopaminergic neurons. Under normal conditions, there was no difference in TH expression between wild-type and DX2-TG mice, implying normal viability of dopaminergic cells in the substantia nigra pars compacta (SNpc). In contrast, under rotenone-treated conditions, TH expression was significantly reduced in wild-type mice, while TH-positive cells in DX2-TG mice displayed notable recovery compared to cells from wild-type mice (Fig. 5c–e). We also confirmed improvements in motor symptoms in DX2-overexpressing mice following rotenone treatment. In the pole test, DX2-TG mice turned around faster than wild-type mice (Fig. 5f) and exhibited a longer descending time, indicative of a stronger grip on the pole (Fig. 5g). In conclusion, our results reveal that DX2 overexpression in transgenic mice confers resistance to Parkinsonism induced by rotenone treatment, suggesting a potential therapeutic role for DX2 in PD.

To investigate the effects of DX2 in the 6-hydroxydopamine (6-OHDA)-induced Parkinson's disease (PD) model [27], mice were first intracranially injected with either AAV-GFP (control) or AAV-DX2 and subsequently received a 12 μg injection of 6-OHDA in the right striatum. Two weeks later, mouse behavior was assessed twice at 4-week intervals (Fig. 6a). According to the apomorphine test, motor symptoms were notably ameliorated in DX2-treated mice compared to saline or GFP-treated mice (Fig. 6b).

Furthermore, to compare the neuroprotective effects between AAV-GFP and AAV-DX2-treated mice, we conducted the contralateral forepaw contact test and the elevated body swing test. In the cylinder test, both saline and GFP mice exhibited unequal distribution of touch. However, the touch rate of contralateral and ipsilateral paws increased in the AAV-DX2-injected group (Fig. 6c). In the elevated body swing (EBS) test, head swing to the right was diminished in the AAV-DX2-injected group (Fig. 6d). Collectively, mouse behavior was indistinguishable between saline and GFP-treated mice, yet recovery of 6-OHDA-induced PD symptoms was observed in AAV-DX2-treated mice.

To further analyze tyrosine hydroxylase (TH) expression in the AAV-transduced mouse brain, we performed immunostaining. The 6-OHDA-injected right hemisphere displayed a decrease in TH intensity due to the loss of dopaminergic neurons in the substantia nigra (Fig. 6e and Additional file 1: Fig. S4). However, the number of dopaminergic neuronal cells recovered in DX2-injected mice compared to GFP-transduced ones (Fig. 6e and f).

We assessed the cleavage level of poly (ADP-ribose) polymerase 1 (PARP-1), a marker of apoptosis, in the striatum region of the tested animals. Cleavage of PARP-1 was little detected in the striatum of AAV-DX2-injected mice even after 6-OHDA injection (Fig. 6g). Also, PARP-1 was deactivated in the striatum region of AAV-DX2-injected mice compared with 0.9% saline or AAV-GFP-injected mice (Fig. 6h). To confirm the anti-apoptotic effect on dopaminergic neurons in DX2-injected mice, we analyzed the mRNA expression levels of DX2 and the apoptotic marker gene, Bax, using quantitative RT-PCR. As anticipated, DX2 expression levels significantly increased in DX2-injected mice (Fig. 6i), and Bax expression, a neuronal cell death marker, diminished in these mice (Fig. 6j).

In conclusion, our findings demonstrate that AAV-DX2 administration improves motor function and neuroprotection in a 6-OHDA-induced PD mouse model. This study underscores the therapeutic potential of DX2 for PD treatment and emphasizes the need for further research to optimize and develop this approach for clinical applications.