Introduction

Parkinson’s disease (PD), the second most common neurodegenerative disorder, tends to affect aging population [1]. Accumulating evidence demonstrates that advancing age confers the major risk for PD progression. Old-age onset PD patients manifest faster progression of motor symptoms, decreased levodopa responsiveness, and increased cognitive impairment [2,3,4]. During aging process in the brain, senescence is observed in both glia cells and post-mitotic neurons, which plays a detrimental role in age-associated neurodegenerative disorders progression [5, 6]. Recently, astrocyte senescence has been proved to participate in PD progression and initiation of cognitive decline with aging [7, 8]. Senescent astrocytes exhibit large and flat morphology with enhanced senescence-associated β-galactosidase (SA-β-Gal) activity, accumulation of p16, and elevated senescence-associated secretory phenotype (SASP) secretion [9]. Importantly, clearance of senescent astrocytes prevents tau pathology and cognitive decline [10, 11], providing insights into therapeutic strategy targeting astrocyte senescence for neurodegenerative diseases. Although emerging evidence has demonstrated a correlation between astrocyte senescence and PD progression, further investigation is warranted to reveal the molecular mechanism underlying astrocyte senescence in the context of PD.

As a cytosolic DNA sensor, cyclic GMP-AMP synthase (cGAS) and its downstream signaling protein stimulator of interferon genes (STING) are components of the innate immune system to clear pathogenic entities [12,13,14,15]. But, excessive activation of the cGAS-STING pathway in the brain contributes to neuroinflammation [16]. Besides the canonical role of the cGAS-STING signal in immune response and inflammation, recent studies uncovered the control of cGAS-STING in cellular senescence. For instance, cGAS-STING signal was required for senescence of mouse embryonic fibroblasts [17, 18] and chondrocytes [19]. However, the effect of cGAS-STING on astrocyte senescence is poorly documented. Depicting astrocytic cGAS-STING signaling cascade may provide therapeutic targets for PD.

In present study, we found that cGAS-STING signal was activated in senescent astrocytes. More importantly, silencing astrocytic cGAS-STING signal delayed astrocyte senescence and PD progression in both MPTP-treated PD mice and aged mice. Further, we identified LCN2 as the downstream effector of the cGAS-STING pathway in the control of astrocyte senescence. Mechanistically, YY1 was discovered to be the transcription factor that negatively regulates LCN2 expression. Nuclear translocation of YY1 was blocked by binding of STING to YY1. Thus, we have demonstrated that the cGAS-STING-YY1 axis promotes astrocyte senescence through upregulation of LCN2 expression which contributes to PD progression. Our study deepens the understanding of the molecular mechanism for astrocyte senescence, and sheds light on the development of disease modifying therapy for PD.

Materials and methods

Experimental animals

Mice (C57Bl/6 background) were maintained at the Animal Resource Centre of the Faculty of Medicine (Nanjing Medical University) and housed under standard laboratory conditions (22 ± 1 °C, 12 h light-dark cycle, food and water ad libitum). All procedures concerning animal care and treatment were performed in accordance with the protocols approved by the Institutional Animal Care and Use Committee of Nanjing Medical University.

MPTP/p-induced PD mouse model

The MPTP protocol was performed as described previously [20]. Mice were randomly divided into four groups (15/group). The randomly mice (male, 16–18 week) were injected subcutaneously with MPTP hydrochloride (25 mg/kg, Sigma-Aldrich, St. Louis, MO, USA) followed by intraperitoneal injection of probenecid (250 mg/kg, Sigma, St. Louis, MO, USA) at 1 h interval for four consecutive days and left for 3 days.

Stereotaxic injection

The surgical procedure was carried out as previously described [21]. Under anesthesia, the 1 μL of adeno-associated virus (AAV-GFAP promoter-cGAS-shRNA-EGFP, 5.5 × 1012 vg/mL, GOSV0290175 or AAV-m-LCN2-GFAP promoter-MCS 3FLAG hGH PolyA, 7.1 × 1013 vg/mL, GOSV0334717, GeneChem, Shanghai, China) was bilaterally delivered into the substantia nigra pars compacta (SNpc) at a rate of 0.2 μl/min using the following coordinates: −3.0 mm A/P, ±1.3 mm M/L, and −4.5 mm D/V from bregma.

Behavioral analysis

Behavioral analyses were performed as described previously [22]. The pole test, rotarod test, and locomotor activity test were performed at 7 days after the final injection of MPTP. For the pole test, the mice were placed head upward on the top of a vertical wooden rough-surfaced pole (diameter 1 cm, height 50 cm). The total time until the mouse reached the floor with its four paws was recorded (T-total). The time needed for the mouse to turn completely head downward was recorded (T-turn). For the rotarod test, mice were accustomed to the apparatus before testing. The mice were then placed on the rod and tested at 20 rpm for 300 s. The latency time that each mouse stayed on the rod at rotarod speed was recorded. Locomotor activity of mice was detected in an activity monitor under bright illumination. The mouse was placed into activity monitor chambers (20 cm × 20 cm × 15 cm) for 30 min, and the activities were recorded at 5-min intervals. The tester was blinded to all genotypes and treatment groups for each behavioral testing.

Immunohistochemistry and immunofluorescence

To detect tyrosine hydroxylase (TH), as described previously [23], brain tissue encompassing each midbrain was cut into 30-μm slices using a freezing microtome (Leica M1950, Nussloch, Germany). Brain slices were incubated with mouse antibody against TH (T1299, 1:1000, Sigma, St Louis, MO, USA) overnight and then for 1 h with secondary antibodies. Immunoreactivity was visualized by incubation in substrate-chromogen solution (DAB). Control staining was performed without primary antibodies. The total number of TH-positive neurons in the SNpc were counted stereologically using the Optical Fractionator (Stereo Investigator 7, MBF bioscience, Williston, VT, USA).

For immunofluorescence staining, the slices or the astrocytes were incubated with anti-GFAP (MAB360, 1:1000, Millipore, Billerica, MA, USA), anti-lamin B1 (Abcam, ab16048, 1:400 dilution), anti-p16 (Santa Cruz Biotechnology, sc-56330, 1:200 dilution), anti-cGAS (31659S, 1:500, Cell Signaling Technology, USA), anti-TH (AB152, 1:2000, Sigma-Aldrich, USA), anti-MAP2 (sc-74421, 1:1000, Santa Cruz Biotechnology), anti-IBA-1 (ab5076, 1:1000, Abcam), anti-p-stat3 (9145S, 1:1000, Cell Signaling Technology, USA), anti-p-STING (72971S, Cell Signaling Technology, USA) or anti-YY1 (MA5-42708, 1:1000, Invitrogen, USA) overnight at 4 °C and then incubated with Alexa Fluor 555-conjugated antibody (Invitrogen, A21432; 1:1000) or Alexa Fluor 488-conjugated antibody (Invitrogen, A21202; 1:1000) for 1 h at 20 °C. DAPI (P36931, Life Technologies) visualizes nuclei. Images were acquired by a confocal microscope (Axiovert LSM510, Carl Zeiss Co., Germany) and then processed by Image J.

Quantitative RT-PCR (qPCR) and ELISA

As previously described [8], total RNA was extracted from these cultured astrocytes and SNpc tissue with Trizol reagent (Invitrogen, USA). Reverse transcription PCR was carried out using a TAKARA PrimeScript RT reagent kit and qPCR was performed in duplicate for each sample using a QuantiTect SYBR Green PCR kit (Qiagen, Germany) with an ABI 7300 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). GAPDH and β-actin were used as an internal control for the real-time PCR amplification. The sequences of primers used are as follows: p16Ink4a forward: CGCTTCTCACCTCGCTTGT, reverse: TGACCAAGAACCTGCGACC. IL-1α forward: AGTCAACTCATTGGCGCTTG, reverse: GAGAGAGATGGTCAATGGCAGA. IL-6 forward: TCCTTCCTACCCCAATTTCCA, reverse: GTCTTGGTCCTTAGCCACTCC. MMP-3 forward: GTTCTGGGCTATACGAGGGC, reverse: TTCTTCACGGTTGCAGGGAG. MMP-9 forward: CGACTTTTGTGGTCTTCCCC, reverse: AGCGGTACAAGTATGCCTCTGATTTCCA. Nd1 forward: CAAACACTTATTACAACCCAAGAACA, reverse: TCATATTATGGCTATGGGTCAGG. Nd2 forward: CCATCAACTCAATCTCACTTCTATG, reverse: GAATCCTGTTAGTGGTGGAAGG. L1 gDNA forward: TAGGAAATTAGTTTGAATAGGTGAGAGG, reverse: TCCAGAAGCTGTCAGGTTCTCTGGC. IL-1β forward: TCATTGTGGCTGTGGAGAAG, reverse: AGGCCACAGGTATTTTGTCG. GAPDH forward: CAAAAGGGTCATCTCC, reverse: CCCCAGCATCAAAGGTG. β-actin forward: CACTGTGCCCATCTACGA, reverse: TGATGTCACGCA CGATTT.

IL-1β (MLB00C), IL-1α (MLA00), IL-6 (M6000B), MMP3 (MMP300) and MMP9 (MMPT90) levels in cell culture supernatants and SNpc tissues were determined using a mouse ELISA kits from R&D Systems according to the manufacturer’s instructions.

Culture and treatment of mouse primary astrocytes

Astrocyte primary culture was described as previously [24]. The neonatal midbrain (P0–3) was trypsinized and dissociated. Tissue was centrifuged for 5 min at 1000 rpm centrifugation, triturated, and resuspended in Dulbecco’s modified Eagle’s medium/Ham’s F12 medium containing 10% fetal bovine serum, GIBCO, Gaithersburg, MD, USA). Cells were plated onto poly-D-lysine-coated T-75 flasks at 50,000 cells/cm2 to generate mixed glial cultures. Confluent mixed glial cultures were shaken at 220 rpm for 6 h at 37 °C to remove unwanted cell types (microglia, oligodendrocytes, neurons, and fibroblasts). The purity of astrocytes was >95% as determined with GFAP immunocytochemistry. To induce premature senescence model, astrocytes were treated with MPP+ (200 μM, Sigma, St. Louis, MO, USA) for 24 h or α-synuclein aggregate (α-Syn PFF, 1 μg/ml, ab218819, abcam, USA) for 48 h. To induce naturally senescence model, astrocytes were cultured for 40 days (passages 8–10) in vitro.

Cell transfection

For LCN2 plasmid transfection, astrocytes were transfected with plasmids expressing Flag-LCN2 (pcDNA3.1-LCN2, Hanbio Biotechnology Co., Ltd., Shanghai, China) in OPTI -MEM-reduced serum medium (Gibco, USA) using lipofectamine 3000 reagent (Invitrogen, Life Technologies) for 48 h. For knockdown of YY1, STING, cGAS, LCN2, the siRNA targeting YY1 (sense: GACGGUUGUAAUAAGAAGUTT; antisense: CAAAGUCUACUCCACAAGCTT), STING (sense: GGAGCCGAAGACUGUACAUTT; antisense: AUGUACAGUCUUCGGCUCCTT), cGAS (sense: GGAUUGAGCUACAAGAAUATT; antisense: UAUUCUUGUAGCUCAAUCCTT), or LCN2 (sense: CCAGUUCACUCUGGGAAAUTT; antisense: AUUUCCCAGAGUGAACUGGTT) (Genepharma, Shanghai, China) was transfected into astrocytes using lipofectamine 3000 reagent for 48 h before stimulation according to the instructions provided.

SA-β-gal staining

SA-β-gal staining was performed with a β-galactosidase-based Senescence Cells Staining Kit (CS0030-1KT, Sigma-Aldrich, USA) according to the manufacturer’s instructions. Astrocytes were fixed in 4% paraformaldehyde for 30 min at 20 °C and then incubated with the SA-β-gal staining solution overnight at 37 °C. The senescent astrocytes stained blue at pH 6 were counted. Positive cells were expressed as a percentage of total cell number.

Flow cytometry analysis of cell cycle

To determine the cell cycle, astrocytes were collected and fixed with ice-cold 70% ethanol overnight at 4 °C. The fixed cells were stained with 50 μg/ml propidium iodide in the dark at 4 °C for 30 min in the presence of RNase A. The stained samples were then analyzed by flow cytometry (BD Bioscience, Franklin Lakes, USA). All the experiments were repeated three times.

Quantification of neuron count and neuronal processes

As described previously [24], mesencephalic primary neurons were treated with astrocytic conditioned medium (ACM) for 24 h and then incubated with anti-MAP2 antibody (ab32454, Abcam, USA) at 4 °C overnight followed by incubation with Alexa Fluor 488-conjugated antibody for 1 h at 20 °C. Neurons were counted in 8 randomly selected fields and the total length of cell processes was quantitated using Image Pro Plus 5.1.

Liquid chromatograph—tandem mass spectrometry (LC-MS/MS)

Immunoprecipitation of protein complexes followed by LC-MS/MS has been used to identify targets that bind to a protein of interest. For detecting proteins binding to p-STING in astrocytes, cells were lysed, centrifuged and the supernatants were incubated with anti-p-STING (72971S, Cell Signaling Technology, USA) antibodies at 4 °C overnight with gentle shaking, and precipitated with protein A/G-agarose beads. Immunocomplex of anti-p-STING antibodies was resolved in 1D-PAGE gel followed by silver staining of the gel. Protein bands were excised and digested with trypsin. Peptides were extracted and analyzed using the automated LC-MS/MS method according to reported procedure [25]. LC-MS/MS was performed by Applied Protein Technology Corporation (APTBIO, Shanghai, China).

RNA sequencing

Total messenger RNA (mRNA) was extracted from the SNpc tissues of AAV-control and AAV-cGAS shRNA PD mice. RNA libraries were sequenced on the Illumina sequencing platform by LC-BIO Co., Ltd (HangZhou, China). Differentially expressed genes were defined as fold changes cutoff with |log2 ratio| ≥ 1 and p < 0.05.

Chromatin immunoprecipitation (ChIP)

The ChIP assay was carried out by the Pierce Magnetic ChIP Kit (Thermo Scientific, 26,157). The crosslinked complexes were then incubated with anti-YY1 or anti-IgG antibodies and isolated using Pierce Protein A/G Magnetic Beads. After decrosslinking, the enrichment of specific fragments was assessed by qPCR. The results were presented as a percentage of input.

Dual luciferase reporter assay

pGL3 LCN2 promoter vector (Hanbio, Shanghai, China) containing different fragments of YY1 binding sites, pcDNA3.1 vector, and YY1 pcDNA3.1 vector (Hanbio, Shanghai, China) were co-transfected into the 239 T cells using Lipofectamine 3000 for 48 h. Dual-Luciferase Reporter Gene Assay Kit (E1910, Promega, USA) was used to detect the luciferase activities and then normalized with Renilla luciferase activity according to the protocol provided.

WB analysis and co-immunoprecipitation (co-IP)

Brain tissues and cells were homogenized in RIPA lysis buffer. The protein was separated and transferred onto PVDF membranes (IPVH00010, Millipore, Billerica, MA, USA). Membranes were incubated with following primary antibodies overnight at 4 °C. Immuno-reactive bands were detected by ImageQuant™ LAS 4000 imaging system (GE Healthcare, Pittsburgh, PA, USA) and quantified using ImageJ software. The following primary antibodies were used: anti-p16 (ab211542, abcam, USA), anti-LCN2 (ab63929, abcam, USA), anti-laminB1 (ab16048 abcam, USA), anti-p-sting (72971S, Cell Signaling Technology, USA), anti-sting (13647S, Cell Signaling Technology, USA), anti-p-p65 (3039S, Cell Signaling Technology, USA), anti-p21 (37543, Cell Signaling Technology, USA), anti-cGAS (31659S, Cell Signaling Technology, USA), anti-TH (MAB318, 1:1000, Millipore, Billerica, MA, USA), anti-YY1 (MA5-42708, 1:1000, Invitrogen, USA), anti-β-actin (BM0627, Boster, Pleasanton, CA, USA).

For co-IP, cells were lysed, centrifuged and the supernatants were incubated with anti-YY1 (MA5-42708, 1:1000, Invitrogen, USA) or anti-p-STING (72971S, Cell Signaling Technology, USA) antibodies at 4 °C overnight, and precipitated with protein A/G-agarose beads (sc-2003, Santa Cruz, CA, USA) for 4 h at 4 °C. The immunoprecipitated proteins were analyzed by WB analysis.

Statistical analysis

All data were analyzed using Prism7 software. All results were shown as means ± SEM. One way or two-way analysis of variance with the Tukey’s post hoc test was used for comparison among different treatments and genotypes and Student’s t test was used to assess the differences between two groups. The results were considered significant at p < 0.05.

Results

cGAS-STING signal is activated in senescent astrocytes of MPTP-treated mice and aged mice

Cell senescence has suggested to be involved in neurodegenerative diseases and aging progression. To investigate whether senescent astrocytes are observed in PD and aging, we established MPTP-induced mouse model of PD and employed 18–24 month aged mice. Compared to control mice (3 month old), MPTP-treated mice and aged mice showed a significant behavior impairment (Fig. 1A–D) and elevated expression of the senescence marker p16INK4a in astrocytes (Fig. 1E, F), which confirmed senescent astrocytes in MPTP-treated and aged mice. To identify critical factors regulating astrocyte senescence, we repeated the in vivo model in primary culture astrocytes. Interestingly, we found that mitochondrial DNA (mtDNA), but not nuclear DNA (nDNA), was obviously increased in both MPP+ and α-Syn PFF-induced premature senescent astrocytes and long-term culture-induced naturally senescent astrocytes by qPCR analysis (Fig. 1G–I). Strikingly, cGAS-STING signal, a cytosolic DNA sensor, was markedly activated in those senescent astrocytes (Fig. 1J–L, Fig. S1A–H). Meanwhile, we also found the activation of cGAS-STING signal was significantly enhanced in the SNpc of MPTP-treated mice and aged mice, detected by western blot analysis (Fig. 1M, N, Fig. S1I–L). Importantly, the cGAS-STING level was positively correlated with the p16INK4a expression in aged mice (Fig. 1O, P). Subsequently, double immunostaining further confirmed that expression of cGAS was evidently increased in MPTP-treated mice and aged mice, mainly in GFAP+ astrocytes from the SNpc (Fig. 1Q–T, Fig. S1M, N). Collectively, the above findings suggest that cGAS-STING signal is associated with astrocyte senescence and is potentially involved in PD progression.

Fig. 1: cGAS-STING signal is activated in senescent astrocytes of MPTP treated mice and aged mice.
figure 1

Time was recorded in the rotarod test and pole test in aged mice (A, B), n = 13 animals for each group) and MPTP treated mice (C, D, n = 12 animals for each group). Representative double-immunostaining for p16Ink4a (red) and astrocytic marker GFAP (green) in the SNpc of aged mice (E) and MPTP treated mice (F). qPCR measurement of mitochondrial DNA (mtDNA-ND1 and mtDNA-ND2) and nuclear DNA (nDNA) in astrocytes cultured for 40 days (G, six independent experiments), treated with α-Syn PFF (H, six independent experiments), or treated MPP+ (I, six independent experiments). Representative immunoblots of relative expression of cGAS and p-STING in astrocytes cultured for 40 days (J), treated with α-Syn PFF (K), or treated MPP+ (L). Representative immunoblots of relative expression of cGAS and p-STING in the SNpc of aged mice (M) and MPTP treated mice (N). O, P The cGAS-STING protein is positively correlated with p16 protein in the SNpc in aged mice. Quantification of the cGAS level in GFAP/TH/IBA positive cells in MPTP treated mice (Q, n = 4-6 animals for each group) and in aged mice (R, n = 4-5 animals for each group). Representative double-immunostaining for cGAS (red) and astrocytic marker GFAP (green), DA neuron marker TH (green) or microglia marker IBA (green) in the SNpc of MPTP treated mice (S) and aged mice (T). DAPI stains nucleus (blue). The data shown are the mean ± SEM. Unpaired t test was used (A–D, G–I, Q, R) and correlation was analyzed by Pearson’s correlation coefficient (O, P).

Full size image

cGAS-STING signal is required for astrocyte senescence

To study the potential role of cGAS-STING in astrocyte senescence, we knocked down cGAS-STING in astrocytes using siRNA-mediated gene silencing. qPCR and ELISA analysis showed that the SASP factors secretion, such as IL-6, IL-1α, IL-1β, MMP-3 and MMP9, were significantly reduced in cGAS siRNA astrocytes compared to control siRNA astrocytes (Fig. 2A, B, Fig. S2A–F). Flow cytometry analysis showed that the percentage of G0/G1 phase was significantly increased in MPP+ treated astrocytes and this cell-cycle arrest was abrogated by cGAS deletion. (Fig. 2C). In addition, cGAS deficiency markedly augmented the nuclear level of lamin B1 and decreased the level of p16INK4a in astrocytes, detectable by immunostaining (Fig. 2D, E, Fig. S2G, H). Moreover, cGAS deletion significantly reduced the expression of p16INK4a and p21 in MPP+-induced senescent astrocytes (Fig. 2F, Fig. S2I–K). SA-β-gal staining showed that the percentage of β-galactosidase positive cells was noticeably lower in cGAS ablation astrocytes than in control astrocytes (Fig. 2G, Fig. S2L). Meanwhile, STING deletion also obviously inhibited the MPP+-induced astrocyte senescence phenotype, including reduced SA-β-gal activity, downregulation of p16INK4a, and upregulation of laminB1, and reduction of SASP factors secretion including IL-1α, IL-1β, and MMP-3 (Fig. 2H–J, Fig. S3).

Fig. 2: cGAS-STING knockdown prevents senescence of astrocytes in vitro.
figure 2

A Heatmap of relative indicated mRNA levels of SASP in astrocytes transfected with cGAS siRNA and then treated with MPP+. B The levels of IL-6, IL-1α, IL-1β, MMP3, and MMP9 in astrocytes measured by ELISA (Six independent experiments). C Astrocytes were analyzed by flow cytometry to evaluate cell cycle distribution (Three independent experiments). D, E Immunofluorescence and quantification of lamin B1 in GFAP+ astrocytes (Four independent experiments). DAPI stains nucleus (blue). F Representative immunoblots of relative expression of p-STING, p16 and p21 in astrocytes. G Quantification of the percentage of SA-β-gal+ astrocytes over total astrocytes (Six independent experiments). H Representative immunoblots of relative expression of p16 in astrocytes transfected with STING siRNA and then treated with MPP+ (Four independent experiments). I, J Immunofluorescence and quantification of lamin B1 in GFAP+ astrocytes (Four independent experiments). K Representative immunoblots of relative expression of cGAS, p-STING and p16 in astrocytes transfected with cGAS siRNA and then treated with α-Syn PFF (Three independent experiments). L Representative immunoblots of relative expression of p16 in astrocytes transfected with STING siRNA and then treated with α-Syn PFF (Three independent experiments). M Representative immunoblots of relative expression of cGAS, p-STING and p16 in astrocytes transfected with cGAS siRNA and cultured for 7 days or 40 days in vitro (Four independent experiments). N The levels of IL-6, IL-1α, IL-1β, MMP3, and MMP9 in astrocytes measured by ELISA (Six independent experiments). O Astrocytes were analyzed by flow cytometry to evaluate cell cycle distribution (Three independent experiments). P Quantification of the percentage of SA-β-gal+ astrocytes over total astrocytes (Six independent experiments). Q Representative immunoblots of relative expression of p16 in astrocytes transfected with STING siRNA and cultured for 7 days or 40 days in vitro (Three independent experiments). The data shown are the mean ± SEM. Two-way ANOVA with Tukey’s post-hoc tests were used.

Full size image

To confirm the effect of cGAS-STING on astrocyte senescence, we treated the astrocytes with α-Syn PFF. α-Syn PFF induced astrocyte senescence phenotype, including upregulation of senescence markers p16INK4a and increase in the secretion of SASP factors such as IL-1α, IL-1β, IL-6, MMP-9 and MMP-3. cGAS deletion significantly reduced the α-Syn PFF-induced p16INK4a expression and the SASP factors secretion (Fig. 2K, Fig. S4A–F). The inhibitory effect of STING deficiency on astrocyte senescence was similar to that of cGAS deficiency (Fig. 2L, Fig. S4G–L).

Long-term culture-induced naturally senescence of astrocytes were further verify the role of cGAS-STING in astrocyte senescence. We found that cGAS-STING deletion markedly inhibited the astrocyte senescence phenotype, such as downregulation of senescence markers p16INK4a, decreased SA-β-gal activity, increase in the percentage of S phase and reduction of IL-1α, IL-1β, IL-6, MMP-9 and MMP-3 secretion (Fig. 2M–Q, Fig. S5). These findings indicate that cGAS-STING promotes astrocyte senescence in vitro.

Astrocytic cGAS ablation alleviates PD-like pathology via delay of astrocytes senescence in MPTP-treated mice

Emerging evidence suggests that senescent astrocytes is involved in PD [11]. To explore the role of astrocytic cGAS in PD, we stereotaxically injected AAV carrying GFAP-promoter-cGAS-shRNA-EGFP to mouse SNpc regions to specifically down-regulate astrocytic cGAS expression. As expected, cGAS-shRNA-EGFP was strongly expressed in GFAP+ astrocytes of SNpc on day 28 after injection (Fig. 3A, B), and the efficiency of cGAS silencing was confirmed by double immunostaining, which showed 62% reduction of protein level of cGAS in virus-infected cells (Fig. S6A, B). We next determined whether deletion of cGAS in astrocytes had any effect on Parkinsonism in mice. First, we found that astrocytic cGAS deletion significantly improved the behavior of MPTP-treated mice, including shorter time in the pole test, better performance in the rotarod test and longer distance in the open field test (Fig. 3C–G). In addition, astrocytes-specific cGAS deletion obviously reversed the loss of dopamine (DA) neurons in PD model mice, as defined by TH immunostaining in the SNpc (Fig. 3H, I) and in the striatum (Fig. S6C). Moreover, the expression of TH protein was markedly reduced in MPTP-treated PD mice, but this decrease was noticeably rescued by astrocytic cGAS deletion (Fig. 3O, P). These data demonstrate that astrocytic cGAS ablation alleviates PD like pathology in the MPTP/p -induced mouse model of PD.

Fig. 3: Astrocytic cGAS ablation alleviates PD like pathology via delay of the senescence of astrocytes in MPTP treated mice.
figure 3

A Diagram of the experimental design. B IHC results confirm that AAV-cGAS shRNA (EGFP) is expressed, mainly in GFAP+ astrocytes (red). C Time on the rod was measured by the rotarod test (n = 9 animals for each group). D, E the time taken to turn around (Time-turn) and descend a pole (Time-total) was recorded in pole test (n = 9 animals for each group). F, G Movement distance within 5 min was recorded by open field test (n = 9 animals for each group). H Microphotographs of TH-positive neurons in the SNpc. I Stereological counts of TH-positive neurons in the SNpc (n = 6 animals for each group). J Heatmap of relative mRNA levels of SASP as indicated. K The levels of IL-6, IL-1α, IL-1β, MMP3, and MMP9 in astrocytes measured by ELISA (n = 6 animals for each group). L Representative double-immunostaining for lamin B1 or p16 (red) and EGFP (green) in astrocytes in the SNpc. Quantification of lamin B1 immunofluorescence intensity (M) and p16 immunofluorescence intensity (N) in GFAP+ astrocytes in the SNpc (n = 6–8 animals for each group). DAPI stains nucleus (blue). Representative immunoblots (O) and quantification of relative expression of TH (P), p-STING (Q), p-p65 (R) and p16 (S) in the SNpc (n = 4 animals for each group). The data shown are the mean ± SEM. Two-way ANOVA with Tukey’s post-hoc tests were used.

Full size image

Meanwhile, we observed a significant decrease in secretion of SASP factors, such as IL-6, IL-1α, IL-1β, MMP-3 and MMP9 in astrocytic cGAS-shRNA infected mice compared to control mice (Fig. 3J, K, Fig. S6D–I). In addition, astrocytic cGAS deletion markedly enhanced nuclear level of lamin B1 and diminished level of p16INK4a in astrocytes of PD mice, detectable by immunostaining (Fig. 3L–N). Moreover, astrocyte-specific deletion of cGAS notably downregulate the expression of p16INK4a, p-STING, and p-p65 in SNpc of PD mice (Fig. 3Q–S). These findings indicate that astrocytic cGAS ablation alleviates DA neurodegeneration through inhibition of astrocytes senescence in MPTP/p-induced PD mice.

Lipocalin-2 (LCN2) acts as an effector target for cGAS-STING-mediated astrocyte senescence

To uncover the downstream targets of cGAS-STING in mediating astrocyte senescence, RNA-seq was performed with total RNAs extracted from the SNpc of astrocytic cGAS deletion and control mice. A cutoff with p value < 0.05 and |log2 ratio| ≥ 1 was used to define differentially expressed genes (Fig. 4A). 198 genes were downregulated and 35 genes were upregulated in the SNpc of astrocytic cGAS-deficient mice (Fig. S7A). The RNA-seq results were confirmed by qPCR with selected the TOP nine downregulated genes and eight upregulated genes in the SNpc and in the astrocytes (Fig. 4B–D). Among the above RNA-seq-identified differentially expressed genes, LCN2 was the most significantly upregulated gene in PD mice and in astrocytes and this increase was completely blocked by cGAS deletion (Fig. 4C, D). Consistent with the mRNA assay, the protein expression of LCN2 was markedly augmented in the SNpc of MPTP-treated mice and in senescent astrocytes, and this upregulation was also abolished by cGAS ablation (Fig. 4E–G, Fig. S7B–D). These findings imply that LCN2 may be a downstream target of cGAS during astrocyte senescence.

Fig. 4: Identification of LCN2 as an effector target for cGAS-STING-mediated astrocyte senescence.
figure 4

A Volcano plot of downregulated (down) or upregulated (up) genes in the SNpc of AAV-cGAS shRNA PD mice compared to AAV-control PD mice by RNA-seq analysis (n = 3 animals for each group). B Heatmap of top 17 up- or downregulated genes indicated between the SNpc of AAV-control and AAV-cGAS shRNA PD mice. qPCR analysis measuring the mRNA levels of indicated genes in the SNpc (C, n = 3 animals for each group) and in astrocytes (D, four independent experiments). Representative immunoblots of relative protein level of LCN2 in the SNpc (E) and in astrocytes treated with MPP+ (F) and cultured for 7 days or 40 days in vitro (G). H Representative immunoblots of relative expression of p16 in astrocytes transfected with LCN2 siRNA (si-LCN2) and then treated with MPP+ (Three independent experiments). I Quantification of the percentage of SA-β-gal+ astrocytes over total astrocytes (Four independent experiments). J, K Immunofluorescence and quantification of lamin B1 in GFAP+ astrocytes (Four independent experiments). DAPI stains nucleus (blue). L Representative immunoblots of p16 expression in astrocytes transfected with LCN2 siRNA and cultured for 7 days or 40 days in vitro (Three independent experiments). M Quantification of the percentage of SA-β-gal+ astrocytes over total astrocytes (Four independent experiments). The astrocytes were co-transfected with LCN2 plasmid and cGAS siRNA (STING siRNA) for 48 h. N, O Immunofluorescence and quantification of lamin B1 in GFAP+ astrocytes (Four independent experiments). DAPI stains nucleus (blue). P, Q Representative images of SA-β-gal staining and quantification of the percentage of SA-β-gal+ astrocytes over total astrocytes in astrocytes treated with MPP+ and in astrocytes cultured for 7 days or 40 days in vitro (Four independent experiments). Representative immunoblots of p16 expression in astrocytes treated with MPP+ (R) and in astrocytes cultured for 7 days or 40 days in vitro (S). The data shown are the mean ± SEM. Two-way ANOVA with Tukey’s post-hoc tests were used.

Full size image

We next explored whether the LCN2 mimics the effect of cGAS-STING on astrocyte senescence. Indeed, siRNA-based LCN2 suppression significantly reduced the expression of p16INK4a, the secretion of SASP factors, such as IL-6, IL-1α, IL-1β, MMP-3 and MMP9, and the activity of SA-β-gal, and markedly increased the nuclear level of lamin B1 in both MPP+-induced premature senescence of astrocytes (Fig. 4H–K, Fig. S8A–G) and in long-term culture-induced naturally senescence of astrocytes (Fig. 4L–M, Fig. S8H–N). In addition, astrocytic LCN2 overexpression aggravated the behavioral deficits and promoted the loss of DA neurons in the SNpc in young and aged mice (Fig. S9). These findings demonstrate that LCN2 plays a similar role in promoting astrocyte senescence and neurodegeneration as cGAS -STING.

To determine whether the LCN2 is necessary for the cGAS-STING-induced astrocyte senescence, we re-expressed LCN2 in cGAS-STING deficient astrocytes by transfecting the LCN2 gene. The enforced expression of LCN2 restored astrocyte senescence phenotype in cGAS-STING deficient cells, including increased secretion of SASP factors, upregulation of senescence markers p16INK4a, decreased laminB1 expression and enhanced SA-β-gal activity (Fig. 4N–S, Fig. S10). Thus, our findings demonstrate that cGAS-STING promotes astrocyte senescence through inducing the expression of LCN2.

STING interacts with transcriptional factor Yin Yang 1 (YY1) to prevent YY1 nuclear translocation to increase LCN2 transcription

To identify critical p-STING interaction proteins, LC-MS/MS was performed and recognized YY1 as the major p-STING binding protein (Fig. 5A). Next, we predicted transcription factors regulating LCN2 transcription using promoter analysis tools PROMO, UCSC, and Animal TFDB. A total number of 7 transcription factors that appeared in the three prediction databases were obtained (Fig. 5B). We selected YY1 as the interested transcription factor due to its involvement in cell senescence and its interaction with p-STING. Then, ChIP assays were further performed to validate whether YY1 directly regulates LCN2 transcription as transcription factor. As expected, the results showed that YY1 was recruited to the promoter region of LCN2 (Fig. 5C). YY1 overexpression reduced LCN2 luciferase activity in HEK293 cells by transfecting LCN2 reporter system (Fig. 5D). Immunoblotting assays showed that YY1 knockdown significantly increased the expression of LCN2 in astrocytes (Fig. 5E, Fig. S11A, B). These findings indicate that YY1 directly binds to the LCN2 promoter and inhibits its transcription.

Fig. 5: STING directly binds to YY1 and prevents YY1 nuclear translocation to increase LCN2 transcription.
figure 5

A p-STING-bound protein YY1 in astrocytes is identified by LC-MS/MS. B The Venn diagram presents the overlap of the predicted transcription factors of LCN2 between the PROMO, UCSC, and Animal TFDB datasets. C Binding of YY1 to LCN2 promoter was analyzed using ChIP-qPCR in HEK293 cells (Five independent experiments). IgG, immunoglobulin G. D Luciferase reporter activity of LCN2 in HEK293 cells transfected with YY1-expressing plasmid or vector (Five independent experiments). E Representative immunoblots of LCN2 expression and YY1 in astrocytes transfected with YY1 siRNA (si-YY1) or control siRNA (si-control) (Four independent experiments). F Immunoprecipitation and immunoblot analysis of the interaction of p-STING with YY1 in astrocytes treated with MPP+ using anti-p-STING and anti-YY1 antibodies (Three independent experiments). G Immunoprecipitation and immunoblot analysis of the interaction of p-STING with YY1 in astrocytes cultured for 7 days or 40 days (Three independent experiments). H Representative images showing the localization of p-STING (red) and YY1 (green) in astrocytes. I Representative Immunoblot of YY1 expression in the nucleus and the cytoplasm from astrocytes treated with MPP+. J Representative Immunoblot of YY1 expression in the nucleus and the cytoplasm from astrocytes cultured for 7 days or 40 days. K Representative images of immunofluorescence staining on YY1 (green) and DAPI (blue) in astrocytes transfected with cGAS/STING siRNA and then treated with MPP+. The data shown are the mean ± SEM. Unpaired t test was used (C, D).

Full size image

The binding between p-STING and YY1 was significantly increased in MPP+-induced premature senescence of astrocytes and in long-term culture-induced naturally senescence of astrocytes; this effect was abolished by knockdown of cGAS (Fig. 5F, G, Fig. S11C, D). The interaction of p-STING and YY1 was further confirmed by co-staining of p-STING and YY1 in astrocytes (Fig. 5H). More importantly, we found that the cytoplasmic abundance of YY1 was significantly enhanced, whereas the nuclear content was markedly reduced in MPP+-induced premature senescent astrocytes and in long-term culture-induced naturally senescent astrocytes; this effect was reversed by knockdown of cGAS or STING (Fig. 5I, J, Fig. S11E–L). Subsequently, confocal imaging further confirmed that cGAS-STING deletion promoted YY1 nuclear translocation in astrocytes (Fig. 5K). These data collectively demonstrate that p-STING interacts with YY1 to prevent YY1 from the cytoplasm to the nucleus to increase LCN2 transcription in astrocytes.

LCN2 is required for cGAS-STING-mediated astrocyte senescence and neurodegeneration in MPTP treated mice

Senescent astrocytes produce SASP factors that cause neurotoxicity. We first determined whether LCN2 functionally affects cGAS-STING-triggered neurodegeneration in vitro. Murine mesencephalic primary neurons were then treated with ACM from cGAS-STING-deficient astrocytes. MAP2 immunostaining revealed that cGAS-STING ablation in astrocytes significantly increased both the number of neurons and the length of neuronal processes, but this effect was almost rescued by overexpression of LCN2 in astrocytes (Fig. 6A–C). Subsequently, to investigate whether LCN2 is required for cGAS-STING-mediated astrocyte senescence and neurodegeneration in vivo, we then injected AAV viruses carrying GFAP promoter-LCN2 to the SNpc region in astrocytic cGAS deficient mice to specifically enforce the LCN2 expression in astrocytes (Fig. S12A, B). Immunoblot analysis confirmed that the protein expression of LCN2 was significantly increased in the SNpc region of astrocytic cGAS deficient mice (Fig. S12C). Astrocytic cGAS deletion mice showed improved behavior, including shorter time in the pole test, longer time in the rotarod test and greater distance in open field test, however, overexpression of LCN2 in astrocytes largely blunted this effect (Fig. 6D–G). Furthermore, astrocytic cGAS deletion markedly alleviated DA neuron loss in MPTP-induced PD mice, as defined by TH immunostaining and western blot in the SNpc. Overexpression of LCN2 in astrocytes also rescued this effect (Fig. 6H–K). Collectively, these results demonstrate that LCN2 is required for cGAS-STING-mediated neurodegeneration in vitro and in vivo.

Fig. 6: LCN2 is required for cGAS-STING-mediated astrocyte senescence and neurodegeneration in MPTP treated mice.
figure 6

A Representative pictures of MAP2 (green) immunostaining. DAPI stains nucleus (blue). Quantification of relative the number of neurons (B, four independent experiments) and mean total neuritis length (C, four independent experiments). D The time taken to descend a pole (Time-total) was recorded in pole test (n = 14 animals for each group). E Time on the rod was measured in the rotarod test (n = 10 animals for each group). F, G Movement distance within 5 min was recorded in open field test (n = 13 animals for each group). H Microphotographs of TH-positive neurons in the SNpc. I Stereological counts of TH-positive neurons in the SNpc (n = 6 animals for each group). Representative immunoblots (J) and quantification of TH (K) and p16 expression (L) in the SNpc (n = 5 animals for each group). qPCR measurement of IL-1α (M), IL-1β (N), IL-6 (O), and MMP9 (P) mRNA expression in the SNpc (n = 6 animals for each group). Q Representative double-immunostaining for lamin B1 (red) and EGFP (green) in astrocytes in the SNpc. R Quantification of lamin B1 immunofluorescence intensity in GFAP+ astrocytes in the SNpc (n = 6 animals for each group). DAPI stains nucleus (blue).The data shown are the mean ± SEM. Two-way ANOVA with Tukey’s post-hoc tests were used.

Full size image

More importantly, while knockdown of cGAS in astrocytes led to a significant downregulation of senescence markers p16INK4a, a significant decrease in secretion of SASP factors and a significant increase in nuclear level of lamin B1, AAV-mediated LCN2 gain of function in astrocytes almost abolished this effect (Fig. 6L–R). Together, these results indicate that cGAS-STING-LCN2 axis is important for astrocytes senescence in MPTP treated mice.

cGAS-STING-LCN2 pathway mediates age-related neurodegeneration in aged mice

Our preceding data have strongly indicated that cGAS-STING-LCN2 pathway is activated in aged mice. To further validate the role of cGAS-STING-LCN2 signal in age-related neurodegeneration, 16 month old mice were injected with AAV carrying GFAP promoter-cGAS-shRNA for 4 month in the SNpc regions to reduce astrocytic cGAS expression. Remarkably, we found that behavioral deficits in aged mice were significantly reversed by astrocytes-specific cGAS deletion, as shown by improvement in performance in the pole test, the rotarod test and open field test (Fig. 7A–D). In addition, astrocytic cGAS ablation significantly rescued the DA neuron injury and increased TH protein levels in the SNpc and in the striatum of aged mice (Fig. 7E–I). Moreover, AAV-mediated cGAS loss of function in astrocytes notably delayed the senescence of astrocytes in aged mice, as evidenced by downregulation of senescence markers p16INK4a, upregulation of laminB1 in astrocytes, and reduction of the secretion of SASP factors including IL-6, IL-1α and MMP9 (Fig. 7J–O). Furthermore, the STING activation and LCN2 expression were significantly enhanced in aged mice, and this effect was completely abolished by astrocytic cGAS deletion (Fig. 7P, Q). These results indicate that cGAS-STING-LCN2 signal contributes to aged-related neurodegeneration in mice.

Fig. 7: cGAS-STING-LCN2 pathway mediates age-related neurodegeneration in aged mice.
figure 7

A The time taken to descend a pole (Time-total) was recorded in pole test (n = 14 animals for each group). B Time on the rod was measured in the rotarod test (n = 10 animals for each group). C, D Movement distance within 5 min was recorded in open field test (n = 10 animals for each group). E Microphotographs of TH-positive neurons in the SNpc. F Stereological counts of TH-positive neurons in the SNpc (n = 6 animals for each group). G Microphotographs of TH staining of striatum (n = 6 animals for each group). Representative immunoblots (H) and quantification of relative expression of TH (I) and p16 (J) in the SNpc (n = 4 animals for each group). qPCR measurement of IL-6 (K), IL-1α (L), and MMP9 (M) mRNA expression in the SNpc (n = 6-8 animals for each group). N, O Representative double-immunostaining and quantification of lamin B1 immunofluorescence intensity in GFAP+ astrocytes in the SNpc (n = 6 animals for each group). DAPI stains nucleus (blue). Quantification of relative expression of p-sting (P) and LCN2 (Q) in the SNpc (n = 4 animals for each group). R Proposed model depicting the crucial role of cGAS-STING-YY1-LCN2 in mediating astrocyte senescence, consequently, contributing to neurodegeneration in brain aging and PD. The data shown are the mean ± SEM. One-way ANOVA with Tukey’s post-hoc tests were used.

Full size image

Discussion

The cGAS-STING pathway senses cytosolic DNA to induce innate immune response to clear pathogens and is found to be activated in most central nervous system diseases [26,27,28]. Excessive activation of this pathway contributes to microglia-mediated neuro-inflammation and neuronal cell death [29,30,31]. But, the involvement of cGAS-STING in astrocyte senescence remains unclear. In this work, we showed that cGAS-STING was activated in senescent astrocytes in vitro and in vivo, suggesting that cGAS-STING may be involved in astrocyte senescence. To clearly investigate the role of cGAS-STING specifically in astrocyte senescence, we used multiple cell models including long-term culture-induced naturally senescence model and MPP+/α-Syn PFF -induced premature senescence model. Our results showed that the knockdown of cGAS-STING in astrocytes decreased SA-β-gal activity, downregulated p16Ink4a expression, reduced several SASP factors production, and increased nuclear level of lamin B1. In addition, to further explore the effect of cGAS-STING on astrocyte senescence in vivo, we injected mice with AAV-GFAP promoter-cGAS shRNA to exclusively delete cGAS in astrocytes. We also demonstrated that conditional deletion of cGAS in astrocytes reduced the p16Ink4a expression, increased the nuclear lamin B1 level and decreased the secretion of several SASP factors in astrocytes in both the MPTP-induced PD model mice and the aged mice. These results provide direct evidence that cGAS-STING signal is essential for astrocyte senescence.

To elucidate the molecular mechanism underlying cGAS-STING regulating astrocyte senescence, by using multiple techniques, we identified LCN2 as an effector molecular target for cGAS-STING-mediated astrocyte senescence. LCN2 was recently reported as a potential clinical biomarker in several diseases such as multiple sclerosis and aging-related cognitive decline [32,33,34,35,36]. Astrocytes are considered as the major source of LCN2 [37,38,39]. However, the function of LCN2 and the related molecular mechanisms involved in disease progression remain unclear. Here, we provide novel insights into the biological functions of LCN2 in astrocyte senescence. LCN2 gene expression is regulated mostly at the transcription level. Interestingly, we unveiled that YY1 can directly bind to LCN2 promoter and negatively regulate LCN2 transcription. Furthermore, we revealed that p-STING interacted with YY1 to prevent YY1 from the cytoplasm to the nucleus, thus decreased the abundance of nuclear YY1. This led to the increased transcription of LCN2, further promoting astrocyte senescence (Fig. 7R). More importantly, knockdown of LCN2 was similar to that of cGAS–STING deletion in regulation of astrocyte senescence. Re-expression of LCN2 in cGAS-STING deletion astrocytes restored senescent phenotypes in primary cultured astrocytes and in PD model mice. Collectively, the current work reveals that the effects of cGAS-STING on astrocyte senescence are mediated through YY1-LCN2.

Recent studies have shown that astrocyte senescence participates in the PD pathogenesis [40, 41]. Given that cGAS-STING-LCN2 axis is necessary for astrocyte senescence, we next explored the role of cGAS-STING-LCN2 axis in the pathogenesis of PD. Here, our results showed that the expression of cGAS-STING-LCN2 signal was significantly increased in MPTP-induced PD model mice, while conditional deletion of cGAS in astrocytes improved motor function, reduced DA neuron loss, decreased the accumulation of senescent astrocytes and downregulated the LCN2 expression. More importantly, AAV-mediated LCN2 gain of function in astrocytes reversed cGAS deletion-induced neuroprotection in PD model mice. Additionally, the ACM from cGAS-STING deletion astrocytes results in more slight DA neuron injury, which further supported the crucial role of astrocytic cGAS-STING-LCN2 axis in PD pathogenesis. These results provide direct evidence that cGAS-STING-LCN2 signal in astrocytes indeed plays a critical role in the pathogenesis of PD, and have established a causality between the activation of cGAS-STING-LCN2 axis in astrocytes and neurodegeneration.

As the biggest risk factor, aging accelerates several neurodegenerative diseases progression such as PD [42]. Next, we confirm the role of cGAS-STING-LCN2 signal in aged mice. We showed that the cGAS-STING-LCN2 signal was activated in aged mice and that astrocytic cGAS deletion delayed astrocyte senescence, improved the motor function and alleviated DA neuron injury in the aged mice. These evidences indicate that inhibition of the cGAS-STING-LCN2 pathway may delay brain aging and the progress of aging-related neurodegenerative diseases.

In conclusion, our study identified unrecognized functions of cGAS-STING-LCN2 axis in promoting astrocytic senescence and DA neurodegeneration. This work ascertains that targeting astrocytic cGAS-STING-LCN2 axis represents a novel therapeutic strategy for delaying PD progression.