The care of animals and the experiments were conducted in accordance with the guidelines approved by the Committee on Animal Research of Kyung Hee University (KHU-ASP-20-235; 30 June 2020). Female Sprague–Dawley rats (10 weeks of age, weighing 240–270 g), obtained from Daehan Biolink (originally from Taconic Co., Albany, NY, United States), were housed in temperature-controlled (21–23 °C) and humidity-controlled conditions with a 12:12-h light/dark cycle. The rats for the control or experimental groups were separated and housed until they were sacrificed. During the animal housing and experimental periods, all animals had unrestricted access to food and water. All experiments were conducted with the aim of minimizing animal suffering and utilizing the minimum number of animals required to generate meaningful scientific data.

Stereotaxic surgery was performed according to previously described procedures (Park et al., 2012). Rats were anesthetized through intraperitoneal injection of chloral hydrate (360 mg/kg; Millipore Sigma, St. Louis, MO, United States) and then secured in a stereotaxic apparatus (David Kopf Instrument, Tujunga, CA, United States).

Dextroamphetamine (2.5 mg/kg, intraperitoneally; US Pharmacopeia, Bethesda, MD, United States) was used to induce ipsilateral rotation behavior in rats that were unilaterally lesioned in the nigrostriatal pathway by MPP⁺, as described previously (Nam et al., 2015; Kim et al., 2021). Ipsilateral rotations were counted for 1 h at 1 and 2 weeks after MPP⁺ by Ethovision video analysis software (Noldus, Wageningen, Netherlands). The absolute value of the number of rotations can indicate disease severity; however, it is highly variable across samples. Therefore, we assessed statistical significance within groups by comparing the before and after values for poly I: C or Vehicle treatment.

Human brain tissue was acquired from the Victoria Brain Bank Network (VBBN) located in Parkville, Victoria, Australia. Details on control subjects and PD patients are provided in Supplementary Table S1. The human tissues were deparaffinized in xylene and subjected to citrate antigen retrieval prior to immunohistochemistry. Briefly, brain tissues were boiled in an antigen retrieval buffer containing 10 mM sodium citrate for 20 min. And tissues were chilled and washed with PBS containing 0.5% BSA. After that, tissues were blocked for 1 h in PBS containing 1% BSA, 0.1% cold fish gelatin, 0.5% Triton X-100, 0.05% Tween-20, and 0.05% sodium azide. The procedures and materials used are the same as the rat brain immunostaining protocol.

PD rats were transcardially perfused with physiological saline containing 0.5% sodium nitrate and heparin (10 U/mL) and fixed with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). After dehydration in 30% sucrose solution, rat brains were coronally cut into 40-μm-thick slices using a sliding microtome. Serial coronal sections were processed for immunohistochemical staining using procedures previously described. Briefly, sections were rinsed in PBS and then incubated with the following primary antibodies: mouse GFAP (Millipore sigma; G3893, 1:500), mouse anti-OX-42 (CD11b; Bio-rad, Hercules, CA, United States; MCA275G, 1:300), mouse anti-TH (Abcam, Cambridge, UK; ab5423, 1:1000), rabbit anti-TH (Pel-freez, Rogers, AR, United States; P40101-150, 1:2000), rabbit anti-TLR3 (Novus Biologicals, Centennial, CO, United States; NB100-56571, 1:1000), rabbit anti-α-synuclein (phospho-S129) (abcam; ab51253, 1:2000), rabbit anti-CNTF (Santa-Cruz Biotechnology, Dallas, TX, United States; sc-13996, 1:200), rabbit anti-VEGF-B (Santa-Cruz Biotechnology; sc-13083, 1:200), and rabbit anti-GDNF (Santa-Cruz Biotechnology; sc-328, 1:400). After overnight incubation, in preparation for light microscopy, the brain tissues were incubated with the following biotinylated secondary antibodies: biotin-conjugated anti-mouse (Seracare Life Sciences Inc., Milford, MA, United States; 5260-0051, 1:400), or biotin-conjugated anti-rabbit (Vector Laboratories, Burlingame, CA, United States; BA-1000, 1:400), followed by avidin-biotin-peroxidase complex (Vector Laboratories; PK-6100, 1:1:100). The signal was detected by incubating the sections with 3,3′-diaminobenzidine tetrahydrochloride hydrate (DAB; Milliore Sigma; D5637) and viewed under a bright-field microscope (Olympus Optical, Tokyo, Japan). For fluorescent microscopy, sections were incubated with the following fluorescent-conjugated secondary antibodies: FITC-conjugated-anti-mouse (Vector Laboratories; FI-2000, 1:400), Cy3-conjugated-anti rabbit (Millipore Sigma; AP132C, 1:500), Alexa Fluor 488-conjugated-anti-mouse (Invitrogen, Waltham, MA, United States; A21200, 1:400), Alexa Fluor 594-conjugated anti-mouse (Invitrogen; A11037, 1:500) and CF405M-conjugated-anti-mouse (Biotium, Fremont, CA, United States; 20180, 1:400). The stained sections were covered with Vectashield mounting medium (Vector Laboratories), and signal was viewed under fluorescence microscopy (Carl Zeiss, Jean, Germany; LSM 700).

To count the number of TH + neurons in SNpc, we used the optical fractionator method on a bright-field microscope (BX51, Olympus Optical) with Stereo Investigator software (MBF Bioscience, Williston, VT, United States). This unbiased stereological method of cell counting is not affected by either the reference volume (SNpc) or the size of the counted elements (neurons). In brief, draw the reason of interest (ROI), including the entire SNpc on each stained tissue section placed on the slide glass. In addition, set the grid options in the software for segmentation and draw them in the ROI. Then, count cells with positive signals on randomly selected grids, integrate the counts, and multiply by the grid number.

Imaging data obtained from the bright-field and confocal microscopes were analyzed as pixel intensity using the ImageJ version 1.54f (National Institutes of Health, Bethesda, MD, United States). ImageJ was used to quantify chromogenic signal intensity by thresholding the images, thereby eliminating non-specific background signal. For single-channel measurements, images were converted to 8-bit grayscale and adjusted to the threshold histogram endpoint. Then, the pixel intensity was quantified and normalized by an unstained area. The pixel intensity is the numerical value that represents the brightness of fluorescence-stained signals captured in each image. For the measurement of merged channels, the adjusted images of each channel are colocalized using a colocalization plugin. The pixel intensity of the overlaid signal is quantified. The variable, which ranged from 0 to 255 in pixel intensity, was simplified to the numerical value of the Y axis, as shown in the calculation that lowered numerical value by 1/100.

Statistical analyses were performed using GraphPad Prism 10.3.0 (GraphPad Software, San Diego, CA, United States). Results are expressed as Mean ± Standard Deviation (SD). To determine statistical significance, Student’s unpaired t-test was performed between the two groups. One-way analysis of variance (ANOVA) with Turkey’s multiple comparisons test or Bonferroni multiple comparisons was performed in multiple groups. p < 0.05 was considered statistically significant.

We first examined TLR3 expression in SN sections obtained from human post-mortem PD patients compared to controls (Supplementary Table S1). We detected robust TLR3 immunoreactivity in GFAP immunopositive (+) astrocytes in the SNpc of PD patients compared to controls (Figures 1A,B). The expression of TLR3 was also tested in an MPP⁺-induced experimental PD rat model. Consistent with prior reports (Park et al., 2012; Nam et al., 2015; Kim et al., 2021), the unilateral injection of MPP⁺ into the median forebrain bundle (MFB) resulted in the destruction of 52% of tyrosine hydroxylase (TH) + DA neurons in the SNpc and 56% of TH + fibers in the striatum at one-week post-MPP⁺ injection compared to the PBS (as a control) (Figures 1C–E). We next determined the expression of TLR3 in glial fibrillary acidic protein (GFAP) + astrocytes, TH + DA neurons, and OX-42 (recognizes CD11b) + microglia/macrophages using immunofluorescence staining. We observed a high expression of TLR3 in GFAP+ astrocytes in the SNpc in MPP⁺-lesioned PD rats compared to the control (PBS)-injected rats (Figures 1F,I). Conversely, TLR3 expression was decreased in TH + DA neurons, while OX-42 + microglia/macrophages remained relatively unchanged (Figures 1G–I). The above data indicate that higher TLR3 expression in astrocytes in the MPP⁺-induced PD rat model is potentially relevant to human PD pathology.

Next, to assess the effect of endogenous TLR3 activation on MPP⁺-lesioned PD rats, the rats were treated with poly I: C, a potent TLR3 agonist. We administered 4 μg of poly I: C or vehicle into SN 1 week post MPP⁺ injection. The poly I: C dose showed no neurotoxicity, as evidenced by no significant loss of TH + DA neurons (Supplementary Figure S1). At 2 weeks, the brains were harvested, and we analyzed TLR3 activation. Interestingly, intra-nigral injection of poly I: C significantly enhanced the TLR3 immunoreactivity in astrocytes in MPP⁺-lesioned rats compared to PBS as a control for MPP⁺ and vehicle as a control for poly I: C (Figure 2).

The literature showed that activated astrocytes produce neuronal survival factors, including NTF (Patel et al., 2019; Poyhonen et al., 2019; Duarte Azevedo et al., 2020). Thus, we next investigated whether endogenous TLR3 activation induces the production of neurotrophic factors in astrocytes. Immunohistochemical analysis revealed a significant increase in expression of CNTF and VEGF-B in GFAP+ astrocytes in the SNpc of poly I: C-treated MPP⁺-lesioned rat brain compared to vehicle-treated control at 2 weeks post-MPP⁺ injection (Figure 3). The data suggest that activation of endogenous TLR3 on astrocytes produces neurotrophic and neuronal survival factors.

In parallel with showing neurotrophic astrocytes in the SNpc (Figure 3), we next tested whether MPP⁺-lesioned rats treated with poly I: C attenuate the neuronal death in the SNpc. Before investigating the pathology of the PD rat brain, we performed behavioral tests. We randomly and blindly divided the MPP⁺-lesioned rats that exhibited ipsilateral rotation behavior, as observed in the amphetamine-induced rotation test conducted 1 week after MPP⁺ administration (Figure 4A) (Nam et al., 2015; Kim et al., 2021). We subsequently injected poly I: C or vehicle. One week later (at 2 weeks), we performed the amphetamine-induced rotation test. Interestingly, the rats administered with poly I: C had a reduced number of amphetamine-induced ipsilateral rotations compared to the before poly I: C treatment (by 24.1%; 1 week vs. 2 weeks, p < 0.05) or vehicle treatment (by 47.1%; M + Vehicle vs. M + poly I: C, compared at week 2, p < 0.01) (Figure 4B). The data suggest that poly I: C attenuated progressive asymmetrical motor behavior. Next, the brains were harvested, and we analyzed whether endogenous TLR3 activation promotes the survival of DA neurons against progressive neurotoxicity caused by MPP⁺. The treatment with poly I: C significantly attenuated the MPP⁺-induced loss of nigrostriatal DA neurons, as measured by stereological counts of TH⁺ cells in the SNpc (Figures 4C,D) and density of TH⁺ fibers in the striatum (STR) (Figures 4C,E).

The expression of TLR3 in each cell type in the MPP⁺-injected PD rat model is replicated in the AAV2-α-syn-A53T-induced PD model, which overexpresses human A53T mutant α-syn (A53T) (Sato et al., 2011). At 3 weeks post-AAV2-α-syn-A53T injection, we observed an enhanced phosphorylation of α-syn (serine 129) as a marker of the pathological form of α-Syn in PD (Sato et al., 2011; Anderson et al., 2006), which indicates efficient transduction of α-syn into the DA neurons in the SNpc (Supplementary Figures S2A,B). Notably, in AAV2-α-Syn-A53T-injected rats, we observed increased TLR3 expression on GFAP+ astrocytes or OX-42 + microglia/macrophages compared to eGFP, while the expression of TLR3 on TH + DA neurons was decreased (Supplementary Figures S2C–F). In the treatment with poly I: C or a vehicle at 3 weeks post-AAV2-α-synuclein-A53T injection, poly I: C treatment significantly preserved the number of TH + cells in the SNpc at 4 weeks post-AAV2-α-synuclein-A53T injection compared to vehicle-treated rats (Supplementary Figures S2G,H). In accordance with the MPP⁺-induced PD rats, the findings suggest that TLR3, activated by poly I: C, alleviated the progressive neuronal death in PD rats.

Taken together, pharmacological activation of endogenous TLR3 on astrocytes induces NTF production, thereby attenuating DA neuron death and providing a disease-modifying strategy.