Peli1 controls the survival of dopaminergic neurons through modulating microglia-mediated neuroinflammation

Chronic neuroinflammation is known to contributes to the toxicity of neurodegeneration of Parkinson’s disease (PD). However, the molecular and cellular mechanisms controlling inflammatory responses in the central nervous system remain poorly understood. Here we found that a E3 ubiquitin ligase Peli1 is dramatically induced only in the substantia nigra (SN) of the human and mouse PD brains. The ablation of Peli1 significantly suppressed LPS-induced production of neurotoxic mediators and proinflammatory cytokines in SN and in primary microglia, whereas Peli1 is dispensable for the inflammatory responses in astrocyte. Accordingly, Peli1 deficiency markedly inhibited neuron death induced by the conditioned medium from LPS-stimulated microglia. Mechanistical study suggested that Peli1 acts as a positive regulator of inflammatory response in microglia through activation of NF-κB and MAP kinase. Our results established Peli1 as a critical mediator in the regulation of microglial activation and neuroinflammation-induced death of dopaminergic neurons during PD pathogenesis, suggesting that targeting Peli1 may have therapeutic effect in neuroinflammation.

reported to play essential roles in the modulation of TLR, IL-1 receptor (IL-1R) and TNR receptor (TNRR) signaling in innate immune cells [16][17][18][19] , and in the regulation of adaptive immune cell activation 20,21 . Our recent studies have indicated that Peli1 protein is highly expressed in the brain and spinal cord, in which Peli1 is predominantly expressed in microglia, but not in other CNS-resident cells like neuron, astrocytes and oligodendrocytes. In such case, Peli1 functions as an important mediator for microglia activation and contribute to the pathogenesis of multiple sclerosis and viral encephalitis 19,22 .
Based on ours and others' previous reports, the current study hypothesized that Peli1 could regulate the PD pathogenesis through modulating microglial inflammatory responses. Our experiments were first aimed to characterize Peli1 function in PD and secondly to decipher the mechanism how Peli1 regulate the survival of dopaminergic neurons during neuroinflammation.

Results
Peli1 is induced in the SN of human and mouse PD brain. Recent evidences have suggested that not only CNS-resident cells, but also peripheral immune system, critically regulate the pathology of PD 23,24 . Since we previously demonstrated that Peli1 plays important roles in regulating the function of both CNS-resident microglia and peripheral immune cells 19,20,22 , so in order to examine the biological function of PELI1 during PD pathogenesis, we initially searched the public NCBI GEO database (GDS2519, GDS2821, GDS3128, GDS3129 and GDS4145) and examined the relative expression of PELI1 in SN, medullary regions, and peripheral blood mononuclear cells (PBMCs) that isolated from health donors (HD) and PD patients. We found that not only one dataset (GDS2821, GDS3128 and GDS3129) suggested that PELI1 expression is dramatically increased in the SN of PD patients as compared to that of HD (P = 0.0196, P = 0.0008, and P = 0.0451 Fig. 1a). Although there is no statistical difference, the data from GDS4145 suggested that the PELI1 expression levels have an increased tendency in the medullary regions of PD patients as compared to that from HD (Fig. 1a). However, there is no obvious difference of PELI1 mRNA expression in PBMCs between HD and PD patients (Fig. 1b). To confirm these observations, we stereotaxically injected LPS into the mouse SN to generate the inflammation-induced PD model, and examine the Peli1 induction. The result revealed that Peli1 mRNA and its protein expression in SN is dramatically increased upon LPS priming as compared to that in PBS-injected control SN (Fig. 1c,d). These data collectively suggested that Peli1 may function directly in CNS-resident cells but not in the peripheral immune system to regulate PD pathogenesis.

Peli1 deficiency protects against inflammation-induced TH + neurons death. To figure out
whether Peli1 indeed modulate PD pathogenesis, we generated the PD model by injecting LPS into SN of WT and Peli1-KO mice, and then analyzed the survival condition of TH + dopaminergic neurons by using the Figure 1. Peli1 is induced in the substantia nigra of human and mouse PD brains. (a,b) Normalized PELI1 expression signals in peripheral blood mononuclear cells (PBMCs) (a) and substantia nigra (SN) (b) from health donors (HD) and Parkinson's disease (PD) patients, the data were obtained from NCBI GEO dataset (GDS2821, GDS3128, GDS3129, GDS4145, GDS2519). Each dot in the graphs represent a value from one person. (c,d) QPCR and immunoblot analysis of relative mRNA (c) and protein (d) expression for Peli1 in the SN of WT and Peli1-KO mice that stereotaxically injected with 2 μl PBS or LPS (5 μg/ml). The immunoblot results are presented as Peli1 and Actin blot panels (d, left panel) and summary bar graph quantifying the relative Peli1 protein levels to Actin (d, right panel). Data with error bars represent mean ± SD. Each panel is representative of three independent experiment. *P < 0.05, *** P < 0.001 as determined by unpaired Student's t test.
immunohistochemical staining of SN 7 days after LPS treatment. The data suggested that LPS injection induced a remarkable death of TH + neurons in SN as reported. In contrast, Peli1 deficiency markedly suppressed the loss and enhanced the survival rate of TH + neurons in SN as compared to that in WT mice (Fig. 2a,b). Accordingly, we observed there are more accumulation of Iba1 + microglia and GFAP + astrocytes in the SN of WT PD mice than that in KO PD mice (Fig. 2c,d), suggesting the suppressed activation of the glial cells in Peli1-KO mice during PD pathogenesis. Peli1 contributes to the enhanced inflammation in SN of PD mice. To confirm that Peli1 mediates the enhanced inflammation in SN during PD pathogenesis, we examined the expression of proinflammatory genes in SN (Fig. 3a). In consistent with the immunofluorescence data, Peli1 deficiency markedly inhibited LPS-induced expression of Nos2 and proinflammatory genes (Il1b, Il6, Tnf) in SN as compared to that injected with PBS (Fig. 3b), suggesting that Peli1 served as a positive regulator of neuroinflammation in PD pathology.

Peli1 mediates LPS-induce production of neurotoxic factors in microglia.
It is reported that microglia are initially activated and responsible for the neurotoxicity of LPS-mediated neuroinflammation in SN 25 , so we tested the LPS-mediated gene induction in mouse primary microglia. Expectedly, we found that LPS-induced expression of Nos2 and proinflammatory genes that encoding IL-1β, IL-6 and TNF, were significantly inhibited in Peli1-deficient microglia (Fig. 4a). Accordingly, the loss of Peli1 in microglia significantly suppressed the secretion of these proinflammatory cytokines and neurotoxic nitrites in the microglial cell culture supernatant (Fig. 4b). Therefore, these results suggested that Peli1 mediates the activation and production of neurotoxic factors in microglia during PD pathogenesis.

Peli1 does not affect the inflammatory responses in astrocytes.
Published studies showed that astrocytes have also been well-characterized to function as a key modulator of neuroinflammation during neurodegeneration 26,27 . In this model, astrocyte can be activated by the IL-1β and/or TNF that produced by activated microglia, and together with microglia, synergistically promote the neurotoxic effect 25 . Although Peli1 is known as a positive regulator of IL-1R signaling in innate immune cells [16][17][18] , we found that IL-1β-induced inflammatory responses are not affected in Peli1-deficient microglia and astrocyte as compared to WT cells (Fig. 4c,d). In addition, we have previously demonstrated Peli1 is dispensable for the inflammatory responses induced by TNF in www.nature.com/scientificreports www.nature.com/scientificreports/ astrocytes 19 . Together, these results demonstrated that the suppression of LPS-induced neuroinflammation in SN of Peli1-deficient mice is due to the defective activation of microglia, but not astrocyte.

Peli1 deficiency inhibits microglia-mediated neuron death.
To confirm that Peli1-mediated microglial activation is responsible for the neuron death during LPS-induced SN inflammation, we examined the neurotoxicity to Neuro2A cells by using different conditioned medium (CM) that isolated from LPS-stimulated primary cultured WT or Peli1-deficient microglia or astrocyte (Fig. 5a). We also used TNF plus cycloheximide (CHX) stimulation as a positive control, which can induce Neuro2A cell death. As expected, WT microglia-derived CM significantly promoted the apoptosis of Neuro2A cells, whereas CM from Peli1-deficient microglia dramatically inhibited Neuro2A cell death. However, CM of astrocytes from both WT and Peli1-deficient mice didn't affect the survival of Neuro2A cells, and there is no significant difference of these two groups. Interestingly, WT microglia-derived CM significantly promoted the production of neurotoxic factors of both WT and Peli1-deficient astrocytes. This effect was markedly inhibited when using the CM from Peli1-deficient microglia (Fig. 5b). More interestingly, CM from WT primary microglia also induced significant cell death of primary neurons, but Peli1 deficiency in microglia almost abolished the toxic effect on primary neurons that induced by microglia-derived inflammatory mediators (Fig. 5c).
To exclude the possibility that Peli1 in neurons may directly regulate the inflammation-induced cell death, we knocked down Peli1 gene expression in Neuro2A cells and stimulated with TNF plus CHX, a positive inducer of cell death. The results showed in Fig. 5d indicated that Peli1 knockdown in Neuro2A cells does not affect neuron cell death induced by TNF plus CHX. These results further confirmed that Peli1-mediated activation and production of neurotoxic mediators in microglia contribute to the neuron death during LPS-induced SN inflammation.

Peli1 mediates NF-κB and MAP kinase activation in microglia.
To dissect the molecular mechanisms of Peli1-mediated microglial activation, we reduced Peli1 expression in murine BV2 microglia cells (Fig. 6a), and Peli1 knockdown does not affect the growth and survival of BV2 cells (Fig. 6b). We next investigated the function of Peli1 in modulating LPS-induced gene expression in Peli1-knockdown BV2 cells. The results revealed that the expression of Nos2 and proinflammatory cytokine genes were significantly inhibited in Peli1-knockdown BV2 cells in response to LPS stimulation (Fig. 6c), which confirmed the phenotype we discovered in primary microglia. Since NF-κB and MAP kniase activation play important role in LPS-mediated www.nature.com/scientificreports www.nature.com/scientificreports/ inflammation, we next tested the activation status of these signal pathways in BV2 cells. The results indicated Peli1 knockdown in BV2 cells significantly inhibited NF-κB activation by EMSA analysis, and MAP kinase activation by immunoblot analysis (Fig. 6d-f), which were consistent with our previously published data obtained from primary microglia. Taken together, Peli1 mediated LPS-induced activation of NF-κB and MAP kniase, and thus elevated the activation neurotoxic gene expression in microglia.

Discussion
The specific loss of dopaminergic neurons in the SN is the hallmark of PD 10,28 . Until now, the detailed mechanism controlling dopaminergic neuronal cell death during PD pathogenesis are not fully understood. However, published studies have suggested that the local inflammation in SN may be function as an active contributor to promote the pathology of both human and mouse PD [29][30][31] . Therefore, an improved understanding of the mediators that regulate CNS inflammatory responses during PD pathogenesis will promote the understanding of the pathogenic mechanism of PD and thus help develop new therapeutic strategies. Here, we found that the E3 ubiquitin ligase Peli1 function as an important mediator to regulates the pathogenesis of PD. Loss of Peli1 significantly protected neurons from inflammation-induced neurotoxicity.
Microglial activation has been recognized as a critical event that initiates inflammatory responses in the pathogenesis of PD 14,25 . In the present study, we found that deficiency of Peli1 in primary cultured microglia dramatically suppressed LPS-induced secretion of neurotoxic mediators, leading to promoted neuron cell survival. However, Peli1 deficiency does not affect astrocyte activation by IL-1β stimulation, which is consistent with previous report that Peli1 is dispensable for the IL-1R signaling pathway in mouse embryo fibroblasts (MEF) 32 . In addition, LPS-mediated microglial activation and production of neurotoxic factor are sufficient to induce neuron death. Peli1 deficiency in microglia obviously inhibited this inflammation-induced cell death. More interestingly, LPS-induced pro-inflammatory cytokines by microglia further exhibit a paracrine function to activate astrocytes, which increase the production of neurotoxic factors, leading to exaggerated neurotoxic effect. Consistent with the gene expression results, Peli1 deficiency in astrocyte does not affect neuron death by astrocyte conditioned www.nature.com/scientificreports www.nature.com/scientificreports/ medium that stimulated with the same microglial conditioned medium. Collectively, our data further confirmed that microglial activation is a critical initiator of inflammation in PD pathogenesis.
The microglia-specific function of Peli1 may be attribute to its expression pattern in different glial cells. Our previous data have shown that astrocyte express three Peli family members (Peli1/2/3) at comparable levels 19 . Given the extremely high levels of sequence identity of the different Peli proteins, it is very likely that they have functional redundancies in IL-1R signaling functions in astrocyte. Due to the predominant expression of Peli1 in microglia, knockdown of Peli1 in BV2 cells not only affected the TRIF-mediated TLR signaling in NF-κB pathway, but also impaired the MyD88-dependent TLR signaling in the activation of MAPKs. These results are consistent with our previously findings in Peli1-deficient innate immune cells 32 and primary microglia 19 .
Although no evidences indicated that LPS or bacterial infection directly involve in the human PD pathogenesis, very little LPS has been found in the brains under normal physiological condition 33 , suggesting a possibility that the entered LPS may initiate the neuroinflammation. Accumulating evidences suggested that aging increase the generation of endogenous TLR ligands, which may promote the disease initiation or progression 34,35 . In fact, repeated intraperitoneal injection of LPS exacerbates motor axon degeneration in the mouse system 36,37 , suggesting LPS or bacteria infection may be a risk factor to promote the neuroinflammation and thus induce the death of dopamine neurons. The present findings further confirmed that modulation of LPS-induced TLR inflammatory signaling by Peli1 in microglia indeed controlled the pathogenesis of PD.
In conclusion, we provided the evidences that Peli1 is induced in the SN of human and mouse PD brain, in which Peli1 facilitates microglial over-activation through NF-κB and MAP kinases. The activated microglia that produced amounts of neurotoxic factors, together with the inflammatory mediators produced by activated astrocytes, contributed to inflammation-induced death of dopaminergic neurons. Therefore, targeting Peli1 may suppress expression of neurotoxins, and thus have therapeutic effect in neuroinflammation and related neurological diseases.

Methods
Mice. Peli1-deficient mice (on the C57BL/6 background) were obtained as described previously 21   LPS-induced PD mouse model. The induction of PD mouse model was as described previously 25 . Briefly, WT and Peli1-defiicient 8-10-week-old male mice (n = 4-5 mice/group) were anesthetized and then immobilized in a stereotaxic apparatus. Two microliters of LPS (5 μg/ml) or equal volume of PBS were stereotaxically injected with into the SN (AP, −3.3 mm; ML, ±1.2 mm; and DV, −4.6 mm) over a 5-minute period by using a stainless-steel syringe (Hamilton). Mice were euthanized 6 h after the stereotaxic injection, and the injected SN was dissected under a dissection microscope and the tissue was processed for quantitative PCR. Seven days after the injection, the mice were sacrificed, and the brain slices were collected for immunohistochemistry analysis. www.nature.com/scientificreports www.nature.com/scientificreports/ Immunohistochemistry (IHC) and immunofluorescence (IF). The IHC or IF staining was as described previously 25,26 . Experimental mice were anesthetized and perfused transcardially with PBS followed by 4% paraformaldehyde. Brain samples were postfixed with 4% paraformaldehyde overnight and equilibrated in 30% sucrose. Coronal sections of 10 μm were prepared with a sliding microtome, and were then incubated with primary antibodies: rabbit anti-tyrosine hydroxylase pAb (1:500; Chemicon); rabbit anti-Iba1 pAb (1:500; WAKO); rabbit anti-GFAP pAb (1:800; DAKO); mouse anti-GFAP mAb (1:1,000, Sigma-Aldrich). The brain slices were then incubated with the horseradish peroxidase (HRP)-or fluorescence-conjugated secondary antibodies. The peroxidase activity of immune complexes was revealed with a DAB kit according to the manufacturer's instruction (Beyotime, P0203). Sections were imaged using either a cooled CCD (DP72, Olympus) on a microscope (BX51; Olympus).

Cell quantification. The number of tyrosine-hydroxylase (TH)-positive cells was quantified in adult
Peli1-KO and their littermates in brain cryosections with typical morphology of the substantia nigra as previously described 26 . In brief, four series of cryosections were collected and every fourth section (10 μm) was used for quantification of TH-positive neurons. The average intensities of GFAP or Iba1 were calculated using ImageJ in the substantia nigra.
Primary glial cell culture. Mixed glial cultures were prepared from neonatal mice that age around 1-2 days as described previously 26 . In brief, after removing the meninges, the neonatal brains were dissociated by 0.25% trypsin, filtered with a 40-μm mesh, and the dissociated cells were plated in 10 cm dish in DMEM/ F12 medium containing 10% FBS, penicillin and streptomycin at 37 °C in humidified 5% CO 2 /95% air. Culture media were changed twice a week. The mixed glial cells were reached to 90% confluence at around day 9, were then re-plated after trypsinization. At day 20 in vitro, cultures were mildly trypsinized with trypsin solution (0.07% trypsin in DMEM/Ham's F12) at 37 °C for 15-20 min. Floating cells (astrocytes and dead cells) were removed by rinsing cultures with D-Hanks' solution. The resulting enriched microglial cultures were maintained in DMEM/Ham's F12 complete medium containing 10% FBS, penicillin and streptomycin until use. The purity of the microglia was >97% as determined by flow cytometry to measure the percentage of CD11b + cells.
Primary neuron culture. The primary mouse neurons were prepared by following our previous protocol 19 .
Briefly, we dissociated the cortex of newborn mouse brains in 0.25% trypsin, filtered the cell suspension with a 40 μm mesh and cultured the cells in Neurobasal medium (Invitrogen) supplemented with B-27 (Invitrogen), 100 U/ml penicillin and 100 mg/ml streptomycin. The culture medium was changed every 3 d for 2 weeks. The collected primary neurons were then applied for the survival assay by using the CM from LPS-stimulated primary microglia.
Quantitative RT-PCR. Brain tissues or cell samples were homogenized in Trizol reagent 26 . cDNA was synthesized from 1 μg of extracted total RNA using M-MLV Reverse Transcriptase kit (Takara) according to the manufacturer's instructions. Quantitative PCR was performed with SYBR-Green premix ExTaq (Roche) and detected by a Real Time PCR System by using gene-specific primer sets ( Table 1). The relative genes' expression was assessed in triplicate, normalized to a reference gene Actb (encoding β-actin) and determined based on 2 −∆∆Ct method 38 .
Immunoblot (IB) and electrophoresis mobility shift assays (EMSA). IB analysis were performed as previously described 19 . Primary microglia were stimulated with 100 ng/ml LPS for the indicated time period and lysed in RIPA buffer. The whole cell extracts were separated via 8.25% SDS-PAGE, transferred to PVDF membranes, blocked and subject to IB analysis. Nuclear extract preparation and EMSAs were carried out as described previously 19 .