Microglial cGAS drives neuroinflammation in the MPTP mouse models of Parkinson's disease

Abstract Background Neuroinflammation has been widely accepted as a cause of the degenerative process. Increasing interest has been devoted to developing intervening therapeutics for preventing neuroinflammation in Parkinson's disease (PD). It is well known that virus infections, including DNA viruses, are associated with an increased risk of PD. In addition, damaged or dying dopaminergic neurons can release dsDNA during PD progression. However, the role of cGAS, a cytosolic dsDNA sensor, in PD progression remains unclear. Methods Adult male wild‐type mice and age‐matched male cGAS knockout (cGas −/−) mice were treated with MPTP to induce neurotoxic PD model, and then behavioral tests, immunohistochemistry, and ELISA were conducted to compare disease phenotype. Chimeric mice were reconstituted to explore the effects of cGAS deficiency in peripheral immune cells or CNS resident cells on MPTP‐induced toxicity. RNA sequencing was used to dissect the mechanistic role of microglial cGAS in MPTP‐induced toxicity. cGAS inhibitor administration was conducted to study whether GAS may serve as a therapeutic target. Results We observed that the cGAS‐STING pathway was activated during neuroinflammation in MPTP mouse models of PD. cGAS deficiency in microglia, but not peripheral immune cells, controlled neuroinflammation and neurotoxicity induced by MPTP. Mechanistically, microglial cGAS ablation alleviated the neuronal dysfunction and inflammatory response in astrocytes and microglia by inhibiting antiviral inflammatory signaling. Additionally, the administration of cGAS inhibitors conferred the mice neuroprotection during MPTP exposure. Conclusions Collectively, these findings demonstrate microglial cGAS promote neuroinflammation and neurodegeneration during the progression of MPTP‐induced PD mouse models and suggest cGAS may serve as a therapeutic target for PD patients. Limitations of the Study Although we demonstrated that cGAS promotes the progression of MPTP‐induced PD, this study has limitations. We identified that cGAS in microglia accelerate disease progression of PD by using bone marrow chimeric experiments and analyzing cGAS expression in CNS cells, but evidence would be more straightforward if conditional knockout mice were used. This study contributed to the knowledge of the role of the cGAS pathway in PD pathogenesis; nevertheless, trying more PD animal models in the future will help us to understand the disease progression deeper and explore possible treatments.


| BACKG ROU N D
Parkinson's disease (PD), a multifactorial and age-related neurodegenerative disease, is characterized by the dramatic loss of dopaminergic neurons in the substantia nigra (SN). PD causes motor symptoms with bradykinesia, resting tremors, rigidity, and many non-motor disturbances. 1 It is well known that inflammatory attacks can induce disability and loss of dopaminergic neurons, thereby contributing to the pathogenesis of PD. 2 Notably, virus infections, including Epstein-Barr virus, herpes simplex virus, hepatitis C virus, and Coronavirus disease-19 were associated with an increased risk of PD, [3][4][5][6] suggesting virus infection-associated inflammation may be a probable contributor to the pathogenesis of PD. Since several epidemiological studies reported that non-steroidal anti-inflammatory drugs (NSAID) reduce the risk of PD, 7-9 much attention has been devoted to understanding the benefits of targeting the inflammatory response to functional recovery after PD onset.
More precise anti-inflammatory targets and therapeutic strategies must be further explored. During PD, apoptotic loss of neurons and dsDNA breaks occur in the brain. However, the role of broken genomic dsDNA and the mechanisms through which these dsDNA regulate PD neuroinflammation and pathogenesis remains largely unknown. Therefore, exploring DNA sensor-mediated antiviral inflammation may provide a new potential target for the treatment of PD.
Several reports have shown the role of innate immune cells and inflammation in PD pathogenesis. [10][11][12] Innate immune cells, including macrophages, monocytes, dendritic cells, and microglia (the specialized population of macrophages-like cells in the brain), are essential in triggering an inflammatory response to infection or sterile injury by their pattern recognition receptors (PRRs). 13,14 Among the PRRs, the cyclic GMP-AMP synthase (cGAS) is a primary cytosolic double-stranded DNA sensor that initiates inflammation in response to infection or sterile tissue damage. 15 Upon DNA binding, cGAS is activated to catalyze the synthesis of 2′3′-cGAMP, which binds and activates the adaptor stimulator of interferon genes (STING). 16 STING, in turn, activates TBK1 and IKK, leading to the activation of the transcription factors IRF3 and NF-κB, respectively, which induce the expression of type-I IFNs and other immune regulatory molecules. 17 cGAS has been reported to play essential roles in host defense, tumor growth, and inflammatory diseases. 15 It is worth noting that the cGAS pathway's effect on the central nervous system (CNS) has begun to unravel. Recently, cGAS has been reported to be involved in the pathogenesis of some neurodegenerative progress, such as Huntington's disease (HD) and experimental autoimmune encephalomyelitis (EAE). 18,19 However, the exact function of cGAS in the context of PD is largely unknown.
In the present study, to explore the exact role of cGAS in the progression of PD, we utilized MPTP-induce neurotoxic PD model Laboratory of Reproductive Medicine of Nanjing Medical University, Grant/Award Number: SKLRM-2022BP3 Conclusions: Collectively, these findings demonstrate microglial cGAS promote neuroinflammation and neurodegeneration during the progression of MPTP-induced PD mouse models and suggest cGAS may serve as a therapeutic target for PD patients.

Limitations of the Study:
Although we demonstrated that cGAS promotes the progression of MPTP-induced PD, this study has limitations. We identified that cGAS in microglia accelerate disease progression of PD by using bone marrow chimeric experiments and analyzing cGAS expression in CNS cells, but evidence would be more straightforward if conditional knockout mice were used. This study contributed to the knowledge of the role of the cGAS pathway in PD pathogenesis; nevertheless, trying more PD animal models in the future will help us to understand the disease progression deeper and explore possible treatments.

K E Y W O R D S
antiviral inflammatory signaling, cGAS, MPTP, neuroinflammation, Parkinson's disease F I G U R E 1 cGAS-STING pathway is activated during the pathogenesis of PD. (A) Comparison of cGAS, STING and IFNα1 transcript in SN tissue between healthy controls and PD patients from public datasets in GEO (healthy controls = 7/group, PD patients = 9/group). (B) Western blot analysis of cGAS, STING, IRF3, TBK1, and β-Actin (loading control) in the brain SN tissue of the indicated mice (n = 3/group). (C) Quantitative PCR analysis of indicated genes in the brain SN tissue of PBS or MPTP-induced mice (n = 4/group). Data were normalized to the reference gene, Hprt. (D) ELISA assay of IFNβ, CXCL10 and TNFα in the brain SN tissue of PBS or MPTP-induced mice (n = 4/group). (E) Representative immunofluorescence staining for p-TBK1 (green) in microglia (CD11b, red) of midbrain tissues from PBS or MPTP-induced mice (n = 4/group). The merge of p-TBK1with CD11b is indicated by white arrowheads. Scale bar, 50 μm. (F) Flow cytometry analysis of p-TBK1 in the brain SN tissue of PBS or MPTP-induced mice (n = 4/group). Mean fluorescence intensity (MFI) are detailed with representative overlaid histograms (left) and quantified summary graph (right). Data are representative of two independent experiments for (B). Data are pooled from three independent experiments for (C-F). Error bars show means ± SEM. Unpaired t-test for (A) and (F). Multiple t-test for (C) and (D). and bone marrow chimeric experiment. We found that microglial cGAS exacerbates pathologic neuroinflammation and neurotoxicity by positively regulating antiviral inflammatory signaling during MPTP exposure, establishing a novel potential target for PD therapeutics.

| Animals
Male mice with the C57BL/6 background were used in this study.
The band size of wild type and mutant is 188 and 298 bp, respectively. The knockout mice bred, grew and developed similarly to WT mice, same as previously reported. 20

| In vivo experimental treatments
The mice received intraperitoneal injection of MPTP at a dose of 20 mg/kg four times with 2 h intervals between injections. Seven days later, the mice underwent behavioral testing, and then brain samples were collected for histological analyses, quantitative polymerase chain reaction (qPCR), and western blotting. In some case, mice received MPTP/p administration as described previously. 21 In brief, 10 doses of MPTP HCL in saline plus probenecid in dimethyl sulfoxide were given twice a week for 5 weeks.
Each time, the mice were subcutaneously injected with MPTP-HCl (20 mg/kg), and 1 h later, the mice were intraperitoneally injected with probenecid (250 mg/kg, Macklin, P822732). Animals underwent behavioral testing or were sacrificed for further analysis 7 days after the last treatment. For cGAS inhibitor RU.521 (MCE, HY-114180) treatment, the mice were intraperitoneally injected with RU.521 (10 mg/kg) daily starting the same day as MPTP administration.

| Pole test
As previously described, a pole test was used to evaluate mouse balance and coordination. 23 Briefly, the mice were placed on a vertical pole with their head up (the pole was 0.5 m long and 1 cm in diameter). The total time from the top to descend back into the cage was recorded. Before the start of the test, the training experiment was repeated five times continuously, with 30 min intervals between each training trial. The following day after the training experiment, each mouse was tested thrice with 3 min intervals between tests. The average value of the three tests was recorded.
F I G U R E 2 cGAS deficiency protects mice from MPTP toxicity. (A) Time to descend during pole test analysis of MPTP-treated WT or cGas −/− mice (n = 10/group). (B) Latency to fall during the accelerating rotarod test analysis of MPTP-induced WT or cGas −/− mice (n = 10/ group). (C and D) Open field test analysis of MPTP-induced WT or cGas −/− mice (n = 10/group). Data are presented as representative movement tracks in (C), quantified total traveled distance, movement speed and percentage of center area duration in (D) (n = 10/group). (E) Immunohistochemistry and immunofluorescence analysis of TH neurons and microglia (IBA1, green) in the brain SN tissue from indicated mice (n = 6/group). Data are presented as representative pictures (left), quantified number of TH neurons and relative IOD of IBA1 (right). Scale bar, 200 μm. (F) Quantitative PCR analysis of indicated genes in the brain SN tissue of PBS or MPTP-induced mice (n = 4/group). Data were normalized to the reference gene, Hprt. (G) ELISA analysis of IFNβ, CXCL10 and TNFα in the brain SN tissue of PBS or MPTP-induced mice (n = 4/group). Data are pooled from three independent experiments. Error bars show means ± SEM. Multiple t-test.

| Accelerating rotarod test
The accelerating rotarod test was conducted as described previously. 22 The mice were acclimated to the rotarod apparatus for training, which accelerated from 5 to 30 rpm in 300 s. Each mouse was trained for 3 consecutive days with five trials per day with 30 min intervals between each trial. On the day of the experiment, three trials were conducted with 3 min intervals between each test. Latency to fall was recorded, and the average value of the three trials of mice was finally calculated. After each test, the excrements were cleaned, and the arena was decontaminated with 75% alcohol.

| Immunohistochemistry and immunofluorescence staining
After all behavioral tests, mice were anesthetized and perfused with PBS and 4% paraformaldehyde (PFA) to collect brain samples. For

| Western blot analysis
Brain SN tissue was prepared as single-cell suspensions previously described. 25 In brief, brain SN tissue was digested with DNase I

| Quantitative polymerase chain reaction (qPCR) analysis
Total RNA was extracted using TRIzol reagent (Invitrogen) following the manufacturer's instructions. The cDNAs were synthesized using a cDNA synthesis kit (Vazyme) according to the manufacturer's F I G U R E 3 cGAS deficiency protects mice from MPTP/p-induced toxicity. (A) Schematic representation of the experiments in (B-G). (B) Time to descend during pole test analysis of MPTP-induced WT or cGas −/− mice (n = 10/group). (C) Latency to fall during the accelerating rotarod test analysis of MPTP-induced WT or cGas −/− mice (n = 10/group). (D and E) Open field test analysis of MPTP-induced WT or cGas −/− mice (n = 10/group). Data are presented as representative movement tracks in (D), quantified total traveled distance, movement speed and percentage of center area duration in (E) (n = 10/group). (F) Immunohistochemistry and immunofluorescence analysis of TH neurons and microglia (IBA1, green) in the brain SN tissue from indicated mice (n = 6/group). Data are presented as representative pictures (left), quantified number of TH neurons and relative IOD of IBA1 (right). Scale bar, 200 μm. (G) ELISA analysis of IFNβ, CXCL10 and TNFα in the brain SN tissue of PBS or MPTP-induced mice (n = 4/group). Data are pooled from three independent experiments. Error bars show means ± SEM. Multiple t-test.
protocols. Quantitative PCR was performed using SYBR Green Supermix (Vazyme). The following primers were used:

| Flow cytometry
Brain SN tissue was prepared as single-cell suspensions previously described. 25 In brief, brain SN tissue was digested with DNase I (

| Primary microglial culture and conditioned medium collection
Meninges and blood vessels were carefully stripped from the cerebral cortices of newborn mice at 1-3 days old. After dissection, the

| Primary astrocytes culture and treatment
Primary cultures of hippocampus astrocytes were prepared from neonatal mice, as described previously. 26 After 7 days, astrocytes were purified using more than four repetitions of trypsinization and re-plating. Astrocytes were seeded in plates pre-coated with PDL overnight, then incubated with CM (CM:DMEM/F12 = 1:2) for 24 h.
Then cells were collected to measure cytokine gene expression by qPCR and immunofluorescence staining.

| Primary neurons culture and treatment
Primary neurons were prepared from the ventral mesencephalon of fetuses (E15-16) by treatment with 0.125% trypsin EDTA, as described previously. 27 The neurons were cultured in a neurobasal medium supplemented with 2% B27 and 0.5 mM glutamine for 6 days and treated with CM (CM:neurobasal = 1:2) for 24 h.

| Hoechst staining
Neurons were fixed with 4% PFA for 30 min and stained with Hoechst 33324 (1:1000 dilution) for 10 min. Apoptotic neurons were quantified by imaging in a fluorescence microscope (Olympus BX 60).

| The cGAS-STING pathway is involved in the pathogenesis of PD
Public datasets in GEO showed that cGAS, STING, and IFNα1 expression increased in the SN tissues of PD patients compared with healthy donors ( Figure 1A), suggesting that the cGAS-STING pathway may be involved in regulating the progression of PD. 28 To further explore the cGAS-STING activation, we constructed the MPTP-induced neurotoxic PD model. In brief, MPTP was intraperitoneally injected into mice at a dose of 20 mg/kg four times at 2 h intervals. Then, behavioral testing and brain tissue isolation were

| cGAS deficiency protects against MPTP toxicity
To directly elucidate whether cGAS is involved in limiting or preventing MPTP-induced toxicity, we treated age-matched cGAS knockout (cGas −/− ) and wild-type (WT) mice with MPTP to compare their phenotype ( Figure S1A), and the ablation of cGas was demonstrated using western blot ( Figure S1B). The pole and accelerating rotarod test showed cGas −/− mice descend faster and fall less easily than WT mice (Figure 2A,B). The open test showed that cGas −/− mice traveled more distance and spent more time in the center ( Figure 2C,D).

| cGAS deficiency in CNS resident cells controls MPTP-induced neurotoxicity and neuroinflammation
To further determine whether cGAS deficiency in peripheral im-  Figure S3D). 29 These results indicate that cGAS deficiency in the CNS resident cells but not peripheral immune cells protect against MPTP-induced neurotoxicity and neuroinflammation.

| cGAS deficiency inhibits microglial antiviralrelated inflammatory gene expression during MPTP treatment
A public database showed that cGAS was highly expressed in microglia compared to other CNS cell types in mice and humans, such as astrocytes, oligodendrocytes (OPC), neurons, and endothelial cells ( Figure 5A). Consistently, cGAS has been reported to be expressed mainly in microglia in CNS. 30 Therefore, we next focused on exploring Axl, was significantly decreased in cGas −/− microglia ( Figure 5F,G). These results suggest that cGAS deficiency restricts the microglial antiviral pathway, facilitating microglial inflammation during MPTP treatment.

| cGAS deficiency inhibits neuronal death and controls the phenotypic switch of microglia and astrocytes from a homeostatic to an inflammatory state
The dramatic loss of dopaminergic neurons occurs in the SN during PD. Indeed, we observed a significant increase of γH2AX, the marker of DNA damage, in the SN of mice treated with MPTP ( Figure S5A).
We next determined whether cGAS-dependent neuroinflammation and neurotoxicity in the MPTP-induced PD mice were caused by microglial activation. We firstly detected decreased levels of IFNβ in con- in the primary microglia compared with that of WT ( Figure 6D,E).
Consistently, the CM from cGas −/− microglia also significantly inhibited the expression of neurotoxic A1-related factors, such as Tnfα, Cxcl10, and C3, but elevated the expression of neuroprotective A2-related genes, such as S100β, in the primary astrocytes compared with that of WT ( Figure 6F, Figure S5C). These data suggest that microglial cGAS mediates the phenotypic switch of microglia and astrocytes from a homeostatic to an inflammatory state and neuronal death in vitro.

| The administration of cGAS inhibitors attenuates the neurotoxicity and neuroinflammation
To confirm that cGAS restricts MPTP-induced neurotoxicity and neuroinflammation, we treated WT mice intraperitoneally with the cGAS inhibitor RU.521 ( Figure 7A). We found that RU.

| DISCUSS ION
It has been reported that DNA damage affects mechanisms central to PD pathogenesis, including protein homeostasis, mitochondrial function, and redox homeostasis. 31 progression by promoting neuroinflammation in microglia. [33][34][35] Consistently, we found that cGAS enhances the disease severity of PD, by aggregating PD-related behavior, and decreasing the loss of TH-positive neurons and the number of activated microglial cells in cGas −/− mice during PD. While it has been widely considered that peripheral macrophages are implicated in the pathogenesis of PD, [36][37][38] we identified that cGAS in CNS resident cells but not peripheral immune cells accelerate disease progression of PD by bone marrow chimeric experiments. We further showed that microglial cGAS contribute to PD development by analyzing cGAS expression in CNS cells but without further evidence from conditional knockout mice.
Notably, consistent with our findings, a recent study reported that STING-deficient mice resist PD induction using αSyn-preformed fibril. 39 Still, the role of cGAS in PD progression and the efficacy of  Historically, neuroinflammation has been widely considered a cause of neurodegeneration and not an effect 43 ; recent efforts in therapeutic intervention seek to prevent neuroinflammation in PD.
Most proposed inflammation-modulating therapies target nonspecifically, thus resulting in a broad range of side effects. 44 Recently, exploring TLRs in PD provides a specific route for therapeutic development. Notably, NF-κB, the major pro-inflammatory transcription factor of TLRs pathways, can be targeted for an effective treatment for PD. 45 It is well established that Mucuna pruriens, ursolic acid and chlorogenic acid exhibit anti-inflammatory and neuroprotective activity by inhibiting NF-κB in the MPTP-induced PD mouse model. 45,46 However, targeting TLRs attenuates neuroinflammation mainly triggered by aggregated α-synuclein and gut dysbiosis. 33 In our study, cGAS inhibitor RU.521 significantly attenuates the toxicity of MPTP by suppressing antiviral-associated neuroinflammation.
Therefore, targeting the cGAS-STING signaling pathway shows potential as a therapeutic strategy for treating virus or damaged DNA fragment-induced PD pathogenesis.
The two main experimental animal models for PD are the genetic and toxin models. The transgenic models only simulate the familial form of PD, which accounts for only 10% of PD subjects. 47 The standard toxin models, to some extent, phenocopy the salient feature of PD, especially in sporadic PD, which accounts for an overwhelming majority of PD subjects. 47,48 Although the MPTP-induced mouse toxin model does not fully recapitulate motor deficits of PD patients, it is considered the gold standard in PD research, due to its low cost and significant clinical correlation over other models. [48][49][50] Moreover, 1-methyl-4-phenylpyridinum (MPP + ), the final toxic metabolite of MPTP, has been reported to trigger an inflammatory process characterized by microglia activation in the SN and striatum 51 ; thus, the MPTP-induced mouse model is suitable for studying the mechanism of neuroinflammation during PD development.
Overall, these findings demonstrate cGAS in microglia driving neuroinflammation in mouse models of PD, suggesting that cGAS is involved in PD progression. Moreover, the neuroinflammatory and neurodegenerative effects of the cGAS-STING pathway may not be limited to PD, as evidenced by cGAS-STING activation in other neurodegenerative diseases such as Huntington's disease, Niemann-Pick disease type C, ALS, and AD. 18,[52][53][54] Therefore, targeting the cGAS-STING signaling pathway highlights the potential for broad-therapeutic treatment of antiviralinduced neuroinflammation in PD and other neurodegenerative diseases.

| CON CLUS IONS
In summary, our study demonstrates that cGAS-STING is activated

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All are included in the article and the supplementary data. Raw data and processed files of RNA-seq have been deposited in the NCBI Sequence Read Archive (SRA) database under the accession code PRJNA904829.