The animal experiment was performed in the CACU (Center of Animal Care and Use) of LCDI (Lee Gil Ya Cancer and Diabetes Institute) of GACHON University. Eight weeks old male C57BL/6N mice (20~25 gm) were randomly divided and housed into 5 mice per cage under a temperature-controlled environment with a 12 hours light/dark cycle, free access to food and water. All the animal protocols described in this study were approved by the CACU animal center ethical board. To establish suitable PD model, 30 mg/kg of rotenone (Sigma-Aldrich) which is suspended in 0.5% CMC was treated orally once in a day for 2 months.

UCB-MSCs were grown in α-MEM medium (DMEM, Gibco^® Life Technologies Corp.) supplemented with 10% fetal bovine serum (FBS, Gibco^® Life Technologies Corp.) and 1% Penicillin and streptomycin (Sigma-Aldrich). These cells were maintained at 37°C in a humidified air under 5% CO₂. Cells were subcultured at 80% of confluence.

To generate sRAGE secreting UCB-MSC, transfection was done with CRISPR/Cas9 ribonucleoprotein system (ToolGen, Inc.) which is targeting safe harbor site known as adeno associated virus site 1 (AAVS1). Donor vector was purchased from Sigma Aldrich (pZDonor-AAVS1 Puromycin vector, PZD0020). Inside multiple cloning sites of donor vector, the sRAGE construct was placed by restriction enzyme digestion. EcoRI and NotI restriction sites were successfully used to generate complete donor vector for knock in. This insert consists of human elongation factor 1-alpha (EF1-α) promoter, sRAGE coding sequence and poly A tail. The nucleofection was performed with 4D nucleofector (Lonza). 15 μg of Cas9 protein, 20 μg of gRNA and 1 μg of donor DNA were prepared for 2×10⁵ cells. Cas9 protein and gRNA were incubated for 15 minutes at room temperature before transfection was performed. P3 kit (Lonza, V4XP-3024) was used to perform efficient transfection. After 15 minutes, 4D nucleofector was set in conditions of FF-104 code (Lonza provided). Cells were seeded after transfection in 6 well plates and each contains 8×10⁵ cells.

After one month of rotenone oral injection, animals were randomly divided into 2 groups: PD and sRAGE-MSC. Animals were anesthetized with 3:1 ratio of Zoletil 50 (Virbac) and Rompun (Bayer Korea) by 1 ml/kg intraperitoneally prior to surgical procedures. The injection was performed with unilateral administration of drugs into the right CS (anterior and posterior 0.4, medial and lateral 1.8, dorsal and ventral from Bregma 3.5 mm). 1×10⁶ Cells were prepared in 10μl of α-MEM medium without fetal bovine serum (FBS) and antibiotics. Perfusion was performed under the anesthesia through the heart with 50 ml 1XPBS followed by 50 ml cold fixative consisting 4% paraformaldehyde (PFA). After perfusion, the brain was removed and post-fixed for 5 hours in 4% PFA then stored overnight in 20% sucrose solution. Cryoprotected brain blocks were cut into 10 μm slices on a cryostat.

Frozen sections of mouse brain were washed with 1XPBS 5 times and non-specific binding of antibodies was blocked by normal horse serum (Vector Laboratories). Tissues were incubated with protein-specific antibodies (Table 1) for overnight at 4°C, and then samples were washed with 1XPBS, followed by a secondary antibody incubation step at room temperature for 1 hour. For counterstaining of the nucleus, samples were stained with DAPI (4′6-diamino-2-phenylindole; 1 μg/ml, Sigma-Aldrich) for 20 seconds. After washing with 1XPBS, coverslips were mounted on a glass slide using Vectashield mounting medium (Vector Laboratories) and analyzed by using an LSM710 confocal microscope (Carl Zeiss).

Frozen sections of mouse brain were dried at room temperature for 5 minutes and washed with 1XPBS for 5 times, then incubated in graded ethanol (95% ethanol 15 minutes, 70% ethanol 1 minute, 50% ethanol 1 minute). After washing with distilled water, the brain tissues were stained with 0.5% cresyl violet acetate (Sigma-Aldrich) solution for 12 minutes and passed through distilled water (1 minute), 50% ethanol (1 minute), 70% ethanol (2 minutes), 95% ethanol (2 times 2 minutes), 100% ethanol (1 minute) and finally cleaned with xylene (5 minutes). Stained slides were mounted with DPX mounting medium for histological analysis (Sigma-Aldrich).

Brain tissues were prepared with RIPA lysis buffer (AMRESCO) and added 1x protease inhibitor (ROCHE) followed by sonication. The tissues were centrifuged at 14000 X g for 20 minutes at 4°C. Total protein concentration was measured by BCA (Life Technologies) according to the manufacturer’s method. Equal amounts (20 μg) of proteins were separated on 10% polyacrylamide gels (Life Technologies) and transferred to PVDF membrane (Millipore Corp.). Proteins were detected with protein-specific antibodies (Table 1). ECL (Animal Genetics Corp.) detection reagent was used to visualize the immunoreactivity proteins on the membrane.

Cell activity was measured by (3-(4,5-dimethylthiazol-2-yr)-5-(3-carboxylmethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt) (MTS) assay (Ez-Cytox cell viability assay). 2×10³ of SHSY5Y cells are seeded in each 96 well plate. After seeding, cells were treated with AGE-Albumin (50 nM) during 12 hours. Co-treated cells were incubated with an sRAGE secreting UCB-MSC conditioned medium for 1 hour before AGE-Albumin treatment and incubated for 12 hours. MTS was performed according to the manufacturer’s protocol. The intensity of staining in the culture medium was measured by a spectrophotometer (450 nm) and presents the proportion of living cells.

Paraffin-embedded heart tissues were deparaffinized in xylene for 10 minutes at room temperature and washed in 1X PBS for 5 times then incubated with permeabilization solution (0.1% Triton X-100 (Amresco), 0.1% Sodium citrate (Sigma-Aldrich), freshly prepared 2 minutes on ice. After permeabilization, tissues were rinsed 3 times with 1XPBS. TUNEL reaction mixture (Roche) was added to tissues and incubated in a humidified atmosphere for 1 hour at 37°C in the dark. Tissues were washed again in 1XPBS 3 times and coverslips were mounted onto glass slides using the Vectashield mounting medium (Vector Laboratories).

Human soluble receptor for advanced glycosylation end products (sRAGE) ELISA kit (Aviscera Bioscience) was used to quantify the amount of secreted sRAGE. The samples and standards were loaded in a pre-coated 96-well microplate and incubated for 2 hours at room temperature. After washing the wells four times with washing buffer, 100 μl of detection antibody was added to each well and incubated at room temperature. After 2 hours of incubation, the wells were rinsed for four times with washing buffer and then 100 μl of ARIG-HRP conjugated working solution was added to each well incubated at room temperature for 1hour at room temperature. After washing the plate with washing buffer, 100 μl of Substrate solution to each well followed by incubating for 10 minutes at room temperature. Finally, 100 μl of Stop solution added each well and optical density was determined at 450 nm.

All experimental data were presented as the mean±standard deviation (SD). Statistical significance was evaluated using Student’s t-test and p≤0.05 was considered as significant.

For investigating the involvement of activated microglial cells in PD, we performed triple immunostaining against AGE, albumin, and Iba-1 (activated microglial cell marker). AGE-albumin was localized in activated microglial cells after triple staining in Corpus striatum in mouse PD model (Fig. 1A). However, no labeling was observed in control. Triple labeling was observed 14.4 folds significantly higher in striatum area of PD model compared to control mouse model. Co-localization coefficient was around 5% in control and 72% in PD model as shown in the bar graph (Fig. 1B).

For the generation of sRAGE secreting UCB-MSC, pzDonor vector, CRISPR gRNA and Cas9 proteins were delivered together by Neon transfection system (Fig. 2A~C). This insert consists of human EF1-α promoter, sRAGE coding sequence and poly A tail (Fig. 2B). Finally, the gene of sRAGE successfully integrated into AAVS1 (Fig. 2C) and the cells start generating sRAGE proteins.

sRAGE secretion was measured by western blotting with Flag-antibody which can detect only protein from the successfully transfected vector. Only sRAGE secreting UCBMSC with pzDonor transfection has an expression of Flag (Fig. 2D). To determine the secretion level of the whole sRAGE in the medium, ELISA was performed. In conditioned medium from sRAGE secreting UCB-MSC, 17870.9 pg/ml of sRAGE was detected. However, medium from mock-MSC has 389.37 pg/ml of sRAGE expression (Fig. 2E).

To check the protective effect of sRAGE secreting UCBMSC, immunohistochemistry and TUNEL staining were employed. The expression of RAGE increased after AGE-albumin treatment in neuronal cells compared to the control group. However, the expression was decreased after treating with conditioned medium (Fig. 3A, Supplementary Fig. S1). TUNEL (apoptotic cell marker) positive cells were also increased after AGE-albumin treatment compared to control group. However, the expression was decreased after treating with conditioned medium (Fig. 3A, Supplementary Fig. S2).

To prove the protective effect of sRAGE on cell death, MTS assay with neuronal cells was performed. As a result, MTS assay showed that AGE-albumin treatment induced cell death and the viability of cell was significantly decreased from 96% to 82% (Fig. 3B). However, the viability of co-treated group with AGE-albumin and conditioned medium remained similar to a control group.

Human neuronal cells were treated with AGE-albumin or AGE-albumin co-treated with conditioned medium from sRAGE secreting UCB-MSC to detect the mechanism of cell death trough RAGE-related mitogen-activated protein kinases (MAPK) pathway. The result shows that the expressions of pp38 and pERK were upregulated after treatment of AGE-albumin and decreased after co-treatment with AGE-albumin and conditioned medium. The expression of pJNK was increased by AGE-albumin treatment. However, it remained the same even after co-treatment with AGE-albumin and conditioned medium. Also, the expression of Bax was increased after AGE-albumin treatment. However, if the conditioned medium was co-treated with AGE-albumin, the expression level of Bax was decreased (Fig. 3C~G).

To check the protective effect of sRAGE secreting UCBMSC, cresyl violet staining was employed. The number of neuronal cells in CS was decreased from around 791±47 cells in the control group to 654±39.4 in the PD mice. After sRAGE secreting UCB-MSC treatment, cells were dramatically increased to 717±21.5, which was similar to control (Fig. 4A). There were significant differences in survival cell numbers between with or without sRAGE secreting UCB-MSC treatment in PD mice (Fig. 4B). To check the protective effect of sRAGE secreting UCB-MSC, immunohistochemistry and TUNEL staining were also employed in PD mice with or without sRAGE secreting UCB-MSC treatment. The expression of RAGE increased PD mice in CS compared to the control group. However, the expression was decreased after treatment with sRAGE secreting UCB-MSC (Fig. 4C). TUNEL positive cells were also decreased after sRAGE secreting UCB-MSC treatment compared to no-treatment in PD mice (Fig. 4D). To examine the anti-inflammatory effects, activated microglial cells were stained by immunohistochemistry. The number of activated microglial cells were increased in PD mice brain, however, decreased dramatically after sRAGE secreting UCB-MSC treatment in PD mice (Supplementary Fig. S3).

To examine changes in the movement ability, Rotarod and Pole tests were performed. Firstly, mice were placed on the rotarod machine and they were tested in the continuous speed. The average maintaining times were recorded as 65.54±10.73, 29.30±13.48, and 58.19±18.70 seconds for Control, PD, and sRAGE secreting UCB-MSC treated PD respectively (Supplementary Fig. S4A). Control group has the highest average maintaining time and PD group was recorded as the lowest. The movement ability was increased by treating sRAGE and sRAGE secreting UCB-MSC to mice. sRAGE secreting UCB-MSC treated group showed higher behavioral recovery than the PD condition.

Recovery of animal behavior was measured also by Pole test. The time costs were measured for four different groups of animals; Control, PD, sRAGE treated PD and sRAGE secreting UCB-MSC treated PD groups. From each group, 10 mice were tested and the average maintaining time was measured in seconds. The result was shown as 5.00±1.20, 6.06±1.40, and 3.56±0.44 for Control, PD, and sRAGE secreting UCB-MSC PD groups respectively. The average time was increased in PD compared with control groups, which states that the behavior ability was dropped. However, when the sRAGE and sRAGE secreting UCB-MSC was treated, the behavioral ability was improved (Supplementary Fig. S2B).