caspase-3 pathway in the cerebellum of mice

Journal Pre-proof FGF9 knockout in GABAergic neurons induces apoptosis and inflammation via the Fas/caspase-3 pathway in the cerebellum of mice Moran Guo, Huifang Chen, Weisong Duan, Zhongyao Li, Yuanyuan Li, Yanqin Ma, Xiangyang Xu, Le Yi, Yue Bi, Yakun Liu, Jie Zhang, Chunyan Li

PII:

S0361-9230(19)30486-1

DOI:

https://doi.org/10.1016/j.brainresbull.2019.10.012

Reference:

BRB 9794

To appear in:

Brain Research Bulletin

Received Date:

21 June 2019

Revised Date:

18 October 2019

Accepted Date:

26 October 2019

Please cite this article as: Guo M, Chen H, Duan W, Li Z, Li Y, Ma Y, Xu X, Yi L, Bi Y, Liu Y, Zhang J, Li C, FGF9 knockout in GABAergic neurons induces apoptosis and inflammation via the Fas/caspase-3 pathway in the cerebellum of mice, Brain Research Bulletin (2019), doi: https://doi.org/10.1016/j.brainresbull.2019.10.012

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier.

FGF9 knockout in GABAergic neurons induces apoptosis and inflammation via the Fas/caspase-3 pathway in the cerebellum of mice

Moran Guo1, 2,* , Huifang Chen3,*, Weisong Duan1, 2, Zhongyao Li1, 2, Yuanyuan Li1, 2, Yanqin Ma4, Xiangyang Xu4, Le Yi1, 2, Yue Bi1, 2, Yakun Liu1, 2, Jie Zhang1, 2, Chunyan Li1, 2,+ 1

ro of

Department of Neurology, The Second Hospital of Hebei Medical University, Shijiazhuang, Hebei 050000, China 2 Neurological Laboratory of Hebei Province, Shijiazhuang, Hebei 050000, China 3 Hebei General Hospital, Shijiazhuang, Hebei 050000, China 4 Jiangsu Nhwa Pharm. Co. Ltd, Nantong, Jiangsu 210000, China

authors contributed equally to this work. +Corresponding author. E-Mail: [email protected]

re

-p

*These

lP

Highlights

 The deletion of Fgf9 in GABAergic neurons leads to severe ataxia in mice.

na

 Fgf9 ablation in GABAergic neuron induce Purkinje cells death and disrupt dendritic arborization.

 Fgf9 ablation disrupts Bergmann fiber scaffold formation, impairs granule

ur

Jo



neuron migration. Fas-caspase-3-ERK -induced apoptosis and inflammation are involved in those processes.

Abstract

Fibroblast growth factor 9 (FGF9) is a member of the fibroblast growth factor family and is widely expressed in the central nervous system (CNS). However, it is not clear how the working mechanism of FGF9 is involved in cerebellar development. To address this question, we deleted the Fgf9 gene specifically in GABAergic neurons or

glutamatergic neurons, and demonstrated that Fgf9 ablation in GABAergic neurons rather than the glutamatergic neurons caused severe ataxia. We showed that FGF9 played a key role in the survival and development of Purkinje cells. GABAergic neuron-specific knockout of FGF9 (Fgf9VGAT) significantly affected the survival and development of Purkinje cells, disrupting Bergmann fiber scaffold formation and granule neuron migration in mice. RNA sequencing revealed that 976 differentially expressed genes (DEGs) were identified between Fgf9VGAT and control mice. The DEGs with significantly upregulated expression were found to be involved in apoptotic and inflammatory signaling. Selected genes including Fas, Bid, Caspase3, Cxcl10, CCl2, Bik and Fos, were validated by qRT-PCR and exhibited increases in expression in Fgf9VGAT mice compared to control mice similar to those seen in the

ro of

RNA-sequencing data. The expression levels of apoptosis- and inflammation-related proteins were also increased, especially those of Fas and caspase-3 pathway related proteins. Interestingly, activated ERK signaling has been observed in apoptosis

-p

and inflammatory responses induced by deleting Fgf9 in GABAergic neurons.

Introduction

re

The cerebellum plays an important role in the precise coordination of body, balance and motor learning [13]. In the developing cerebellum, the ventricular

Jo

ur

na

lP

zone generates GABAergic neurons (inhibitory Purkinje cells and cerebellar interneurons), and glutamatergic neurons (granule cells and others) are generated in the anterior rhombic lip [8, 21]. In early postnatal stages, the granule cells are migrated from the external granule layer (EGL) to the internal granule layer (IGL) along the Bergmann glial radial fibers, which plays a dominant role in cerebellar development [19]. Bergmann glia are unipolar cerebellar astrocytes that align next to the Purkinje cell layer during late embryogenesis. The radial fibers of Bergmann glia have been shown to provide guidance for granule neuron migration and to form synapses with Purkinje cell dendrites [2]. Purkinje neurons can receive the excitatory inputs from granule cells and the inferior olive, integrate and process information, and then sent to the neurons of the deep cerebellar nuclei via inhibitory GABAergic synapses [5]. Thus, the symptoms of heritable ataxia are closely associated with pathological changes in Purkinje neurons. Fibroblast growth factors (FGFs) are a large family of growth factors that participate in the proliferation, differentiation and migration of a variety of cell types [12, 14]. The different levels of FGF signaling also affect the brain and cerebellum development in embryonic stages [1]. FGF7, FGF8, FGF9, FGF10, FGF14, FGF17, FGF18, and FGF22 are expressed in the cerebellum and play important roles in

cerebellum development [4, 30, 31]. FGF9 is predominantly expressed in neurons and

ro of

functions as a glia-activating factor (GAF) to promote the differentiation and survival of neurons and glia in the central nervous system (CNS) [32]. Lin Y. et al. found that neural-specific Fgf9-knockout mice (Nestin-Cre based) develop severe ataxia and demonstrated that granule neurons secrete FGF9 to control the formation of the Bergmann fiber scaffold [19]. However, the expression of Nestin, a neural stem cell/progenitor cell marker, is downregulated when CNS stem cell/progenitor cells differentiate into neurons or glial cells [23]. In addition, Nestin also identifies many other embryonic tissues, so it is not a specific marker for the CNS. Therefore, the molecular mechanism underlying the role of FGF9 in cerebellum development is still obscure. To study the role of Fgf9 in the developing cerebellum, we generated GABAergic neuron- specific or glutamatergic neuron-specific Fgf9-knockout mice. In this study, we demonstrated that Fgf9 deletion in GABAergic neurons was directly

re

-p

related to ataxia in mice. Fgf9 ablation in GABAergic neurons affected Purkinje cells survival and development, disrupted Bergmann fiber scaffold formation and impaired granule neuron migration. RNA sequencing identified that Fgf9 deletion in GABAergic neurons induced apoptosis and inflammation in the cerebellum through the Fas/Caspase3 and ERK signaling pathway. This work establishes that FGF9 performs protective functions in the survival and development of Purkinje cells during

lP

cerebellum development.

Results

na

Deleting the Fgf9 gene in GABAergic neurons causes ataxia

Jo

ur

To investigate the role of FGF9 in cerebellum development, the loxP-Cre recombination system was employed to specifically delete the Fgf9 allele by crossing Fgf9flox mice with VGATCre and VGLUT1Cre transgenic mice. The deletion of the floxed sequence resulted in a defective Fgf9 allele lacking the translational initiation site and exon 1 coding sequences. Conditional Fgf9 null GABAergic neurons (Fgf9VGAT) and glutamatergic neurons (Fgf9VGLUT1) were verified by genotyping (Fig. 1A).

Fgf9VGAT pups and Fgf9VGLUT1 pups appeared normal in the first week after birth. Beginning in the second week, Fgf9VGAT mice began to show ataxic symptoms, such as clumsy movements, abnormal postures and impaired balance (Sup Video.1). Images of inked foot prints showed that Fgf9VGAT mice had a diminished stride and typical ataxic gaits, which lacked forefoot-hindfoot correspondence (Fig. 1B). In addition, Fgf9VGAT mice also showed growth retardation (Fig. 1C). On average, adult

Fgf9VGAT mice had 50% lower body weight than control littermates (Fig. 1D). In contrast, Fgf9VGLUT1 mice, WT mice, Fgf9 hemizygous knockout mice and Fgf9loxp/loxp (F/F) mice did not display any abnormalities (Sup Video.2 and Video.3). F/F mice were used as the control in this study. Together, these results indicate that the deletion of Fgf9 in GABAergic neurons leads to severe ataxia. Fgf9 ablation in GABAergic neurons rather than glutamatergic neurons disrupts granule neuron migration The ataxic phenotype suggests that Fgf9VGAT mice might have defects in the cerebellum. The cerebellum was examined at 21 days after birth to test this possibility. All the granule cells had migrated into the IGL in F/F mice and Fgf9VGLUT1 mice.

-p

ro of

However, in the Fgf9VGAT mice, the granular cells were not arranged neatly, and many granular neurons remained in the EGL and molecular layer (ML), a localization pattern that persisted in 12-month-old or older mice (Fig. 2A and 2B). These results show that knocking out Fgf9 in GABAergic neurons causes defects in the migration of cerebellar granule neurons. Fgf9 ablation damages the survival and development of Purkinje cells

re

As the major GABAergic neuron type, Purkinje cells are the most affected cell type in cerebellar ataxia, so we evaluated changes in Purkinje cells after knocking out Fgf9. In the control cerebellum, Purkinje cells were always positioned in a single

Jo

ur

na

lP

layer between the IGL and ML. However, the Purkinje cells in Fgf9VGAT mice were lost and exhibited a disrupted alignment by immunostaining for HE and calbindin, a Purkinje cell marker (Fig. 3A). In addition, the protein expression of calbindin in the cerebellum of Fgf9VGAT mice was markedly decreased (Fig. 3B). Golgi stains showed that dendritic arborization of Purkinje cells was disrupted in Fgf9VGAT mice; compared to those in control mice, the Purkinje cells in Fgf9VGAT mice exhibited elongated primary dendrites and reduced secondary dendrites (Fig. 3C). In order to determine if Fgf9 ablation induce directly Purkinje cells death, calbindin and caspase-3 (a marker of apoptosis activation) double immunostaining was performed. The number of caspase-3 positive apoptotic cells was increased significantly, and the labeling was colocalized with Purkinje cells in the cerebellum of Fgf9VGAT mice (Fig. 3D). These results demonstrate that FGF9 plays important roles in Purkinje cell survival, development and morphology, and Fgf9 ablation in GABAergic neuron induce directly Purkinje cells death. Loss of Fgf9 disrupts the formation of the radial fiber scaffold of Bergmann glia in the cerebellum of Fgf9VGAT mice The cytodifferentiation of Bergmann glial cells proceeds in association with the

migration, dendritogenesis, synaptogenesis and maturation of Purkinje cells, and the migration of granule neurons is guided by Bergmann glial fibers [26]. Therefore, we determined whether the formation of the radial fiber scaffold comprising Bergmann glia is affected in Fgf9VGAT mice. Anti-GFAP immunostaining was used to assess Bergmann glial fiber development in Fgf9VGAT mice. In the Fgf9VGAT cerebellum, many Bergmann glial fibers did not develop the uni/bipolar radial processes characteristic of the radial glia and instead developed a stellate morphology that caused them to lose contact with the pial surface (Fig. 4A). Bipolar radial glial cells were clearly visible in the direct neuronal migration and laminar patterning of the cerebellum in F/F mice. These results indicate that FGF9 in Purkinje cells is required for the formation of radial fibers of Bergmann glia during the expansion of the

ro of

cerebellum.

Many microglia are activated in the cerebellum of Fgf9VGAT mice

re

-p

Microglia are distributed throughout the CNS and can sense and react to many types of damage. To test whether damage to Purkinje cells induces inflammation, we first examined the expression pattern of Iba-1, a microglia marker. Compared to F/F mice, Fgf9VGAT mice exhibited an obviously increased number of microglia, and the microglia changed their morphology from a ramified form to an amoeboid form (Fig. 4B). Western blot results also showed that the expression levels of CD11b (a microglia marker protein) were significantly upregulated in the cerebellum of

lP

Fgf9VGAT compared with that of F/F mice (Fig. 6C).

na

Numerous genes associated with apoptosis and inflammation exhibit upregulated expression in Fgf9VGAT mice

Jo

ur

To further investigate the molecular mechanism involving in the development of the Fgf9VGAT cerebellum, we compared the gene expression profiles of the cerebellum of Fgf9VGAT and F/F mice by RNA sequencing. The RNA-sequencing data revealed significant changes in gene expression: 600 (61.48%) transcripts exhibited upregulated expression, and 376 (38.52%) transcripts exhibited downregulated expression (Fig. 5A). Approximately 21,300 genes were commonly expressed between the Fgf9VGAT and F/F mice (Sup Fig. 1B). The heatmap shown in Fig. 5B provides a visual representation of the expression levels of all differentially expressed genes (DEGs) between the Fgf9VGAT and F/F mice. We evaluated three Gene Ontology (GO) categories: biological process, cell component and molecular function. Most of the DEGs were involved in biological processes such as biological regulation, cellular process, response to stimulus, signaling, cell proliferation and killing. In terms of cellular components, the DEGs were involved in cell, membrane and organelle. In terms of molecular functions, the DEGs were involved in binding,

catalytic activity, signal transducer activity and transporter activity (Sup Fig. 1A). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis showed that the DEGs were involved in cellular community, transport and catabolism, cell growth and death and cell motility (Fig. 5C). Most notably, the DEGs associated with apoptosis and inflammation exhibited significantly upregulated expression in the Fgf9VGAT mice (Table 1). Confirmation of the RNA sequencing data by qRT-PCR and Western blot analysis We selected 8 genes from the DEGs that played important roles in apoptosis and inflammation (indicated in red in Table 1). The genes Fas, Bik and Bid are related to apoptosis. The genes Fas, Fosb, Mapk11, Cxcl10 and Ccl2 are related with

ro of

neuroinflammation. To test the reliability of RNA sequencing data, RT-PCR analysis was performed with specific primers for these eight genes (Fig. 5D). The results showed that the selected genes had expression trends and fold changes that matched those in the RNA sequencing data.

re

-p

We also detected changes in protein expression in the cerebellum of Fgf9VGAT mice. In the current study, Fgf9 ablation promoted apoptosis via the upregulation of Fas, Bid, Bik and Caspase-3 expression (Fig. 6A). LC3-II and p62 expression was markedly increased in the cerebellum of Fgf9VGAT mice compared to that of F/F mice (Fig. 6B). in addition, the mRNA expression of proinflammatory cytokines, including

lP

TNF-a, IL-12b and IL-1b, was increased in the cerebellum of the Fgf9VGAT mice (Fig. 6D), and the expression of inflammation-related proteins, such as CD11b and p-P65, was also increased (Fig. 6C).

na

MAPK pathways are activated in the cerebellum of Fgf9VGAT mice

ur

The mitogen-activated protein kinases (MAPKs) (including ERK1/2, P38MAPK and JNK) and PI3-K/Akt are crucial signaling pathway involved in survival, proliferation, inflammation and apoptosis. Next, to characterize the Fgf9 ablation in GABAergic neurons induced changes in kinase signaling pathways, we analyzed the phosphorylation statuses of MAPKs and Akt. The phosphor-ERK1/2, JNK and

Jo

p38 were all shown to be activated in the cerebellum of Fgf9VGAT mice via Western blotting, with maximal increases occurring in the phosphorylation status of ERK1/2 (Fig. 6E). Western blot analysis showed no change in Akt pathways in the Fgf9VGAT mice compared with F/F mice. The results demonstrate that MAPKs are activated in the cerebellum of Fgf9VGAT mice and ERK1/2 signaling pathway plays an important role in process.

Discussion

FGF9, one of the predominantly expressed FGFs in the CNS, exhibits a broad spectrum of biological activities [35]. A previous study showed that FGF9 is secreted by granule neurons and controls scaffold formation of radial fibers and inward migration of granule neurons in Fgf9Nestin mice [19]. However, in this study, we observed that mice with Fgf9 specifically knocked out in glutamatergic neurons (granule neurons in the cerebellum) displayed a normal phenotype; in contrast, the deletion of the Fgf9 gene in GABAergic neurons caused ataxia (Fig. 1). In addition, obvious and identical pathological changes were found in the cerebellum of Fgf9VGAT mice but not in that of Fgf9VGLUT1 mice. Based on these results, it is implied that GABAergic neuron-derived FGF9 is essential for preventing ataxia in mice.

ro of

Then, we observed the number and morphological changes of Purkinje cells (the mainly type of GABAergic neurons) in the cerebellum of Fgf9VGAT mice. The Purkinje cells in Fgf9VGAT mice exhibited a dramatic numerical loss and an irregular arrangement (Fig. 3A). Staining for the apoptosis marker caspase-3 indicated that

re

-p

there was substantial apoptotic cell death occurring in the Purkinje cells in Fgf9VGAT mice (Fig. 3D). In addition, the primary dendrites of the Purkinje cells in Fgf9VGAT mice were elongated, and the number of secondary dendrites was reduced (Fig. 3C). These results indicate that FGF9 plays a key role in Purkinje cell survival and dendrite differentiation. Purkinje cells express DNER, a Notch ligand that can interact with a Notch

ur

na

lP

receptor on Bergmann glia to promote radial glial identity in the Bergmann glia [7, 9], which collaboratively guides the inward migration of granule cells. Yoon K et al. found that fibroblast growth factor receptor signaling promotes radial glial identity and interacts with Notch1 signaling in telencephalic progenitors [34]. In our study, many Bergmann glial fibers completely lost their uni/bipolar morphology, and the number of GFAP-positive cells with a stellate morphology was increased in the regions where Purkinje cell damage was observed (Fig. 4A). These results demonstrate that FGF9 in Purkinje cells is required for the formation of radial fibers, and for the alignment of Bergmann glia cell bodies. Since it has been reported that

Jo

Bergmann glia radial fibers also provide structural support for the directional growth of Purkinje cell dendrites [27], it is interesting to clarify glial-neuronal functions. The analysis of RNA-sequencing profiling may provide a comprehensive understanding of the mechanism underlying the pathology driven by Fgf9 ablation in GABAergic neurons. In this study, we identified 600 genes with upregulated expression and 376 with downregulated expression in the cerebellum of Fgf9VGAT mice compared to that of F/F mice (Fig. 5A). GO terms and KEGG pathways were evaluated to obtain more insight into the mechanism. The GO and KEGG data

indicate the involvement of these genes in the regulation of apoptosis- and inflammation-related signaling or activities (Fig. 5, Table 1). Fas (also named CD95) is a death receptor belonging to the tumor necrosis factor receptor superfamily [16]. After binding to Fas ligand, Fas initiates an intracellular cascade that leads to the induction of apoptosis in target cells. Bid is a BH3-only protein that links the extrinsic and mitochondrial pathways of apoptosis by transducing signals from death receptors on the cell surface [15]. Bcl-2 interacting killer (Bik), a protein anchored in the endoplasmic reticulum (ER), initiates Bakdependent release of ER Ca2+ stores, resulting in DRP1-regulated mitochondrial fission and cytochrome c release to initiate apoptosis [22]. Caspases play important

ro of

roles in the apoptotic process. Caspase-3 is a major executioner protein in apoptosis and is used as a marker of apoptosis activation. In our study, the expressions of Fas, Bid, Bik and Caspase-3 were increased at both the mRNA and protein levels (Fig. 5D and 6A), suggesting that Fas-caspase-3 induced apoptosis is activated in Fgf9VGAT

re

-p

mice. Interestingly, LC3-II and p62 were increased in the cerebellum of Fgf9VGAT mice (Fig. 6B), which indicated that autophagy was overwhelmed [29]. These results are consistent with those of most studies, considering that apoptotic signaling serves to inhibit autophagy [10]. In response to the pathological changes driven by apoptosis in Purkinje cells, there is a rapid increase in the expression of proinflammatory cytokines and

ur

na

lP

chemokines. Cxcl10 is a proinflammatory cytokine produced by a variety of cell types including glia, dendritic cells, leukocytes and endothelial cells [24], that has been shown to be expressed in numerous inflammatory diseases [28]. The chemokine (CeC motif) ligand 2 (CCL2) is a member of the CC subtype chemokine family and signals through its cognate receptor chemokine receptor type 2 (CCR2). The expression of CCL2 is elevated in both acute disease and chronic inflammation [33]. In addition, some proinflammatory cytokines (e.g., TNF-α, IL-1b, IL-12b) can initiate a cascade of inflammatory mediators by targeting the endothelium. In this study, we identified greatly increased expression levels of Cxcl10, CCL2, TNF-α, IL-1b and IL-12b in

Jo

Fgf9VGAT mice (Fig. 5D and 6D), which supports the notion that microglial activation in the CNS activates inflammation in the cerebellum of Fgf9VGAT mice. In our results, we showed that MAPK signaling pathways, especially the ERK1/2 pathway, are activated in the cerebellum of Fgf9VGAT mice (Fig. 6E). In neuronal apoptosis, the ERK pathway has been shown to have dual roles [17]. Although the majority of studies have demonstrated that the ERK pathway has an anti-apoptotic role in neurons, pro-apoptosis effects-mediated by ERK signaling have also been observed in neurons in the hippocampus or cerebellum [3, 25]. In addition, caspase-3

was also found to be cleaved in the cerebellum of Fgf9VGAT mice. These results are consistent with the finding that LC3/ERK/Caspase-3 signaling is at least partly activated in autophagy and apoptosis [20]. In addition, our results demonstrate that the absence of Fgf9 is not related to PI3‐ K/Akt signaling.

ro of

In summary, this study suggests that FGF9 plays important roles in survival, development and maturation of Purkinje cells. The absence of Fgf9 in Purkinje cells disrupts Bergmann fiber scaffold formation, impairs granule neuron migration and activates microglia. Fas-caspase-3-induced apoptosis may play an important role in those processes. In addition, MAPKs, especially ERK, are involved in apoptosis and inflammation in the cerebellum of Fgf9VGAT mice.

Materials and methods

-p

Mice

Fgf9fl/fl mice were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. The

re

Fgf9fl/fl mice were on the C57BL/6 background and had floxp sites inserted across exon 1 of the target gene by a conventional gene-targeting strategy [6]. To generate Fgf9 CKO mice, Fgf9fl/fl

lP

mice were crossed with VGAT-Cre and VGLUT1-Cre mice (The Jackson Laboratory). Mice were genotyped by PCR analysis (Table S1) of DNA obtained from tail tissue. Mice were maintained in

na

the specific pathogen-free facility of the Animal Center of the Second Hospital of Hebei Medical University. All animal experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee of the Second Hospital of Hebei Medical

ur

University.

Antibodies and chemicals

Jo

The following primary antibodies were used for Western blotting and staining tissue sections: antiFGF9 (Abcam, EPR19937), anti-CD11b (Proteintech, 20991-1-AP), anti-Iba-1 (Abcam, ab5076), anti-GFAP (Millipore, MAB360), anti-Calbindin (Sigma, C2724), anti-Fas (Proteintech, 13098-1AP), anti-Bid (Proteintech, 10988-1-AP), anti-Bik (Cell Signaling, #4592), anti-Caspase-3 (Bioworld Technology. Inc., BS1518), anti-LC3 (Cell Signaling, #12741), anti-p-p65 (Cell Signaling, #3033), anti-p65 (Cell Signaling, #8242), anti-p-ERK (Cell Signaling, #4370), antiERK (Cell Signaling, #4695), anti-p-JNK (Beyotime, AF762), anti-JNK (Beyotime, AF1048),

anti-p-p38 (Cell Signaling, #4511), anti-p38 (Cell Signaling, #8690), anti-p-Akt (Immunoway, YP006), anti-Akt (Immunoway, YT0176) and anti-β-actin (Proteintech, 60008-1-Ig). Secondary antibodies for Western blotting were obtained from Rockland Immunochemicals (USA), and fluorescein isothiocyanate-conjugated secondary antibodies were purchased from Jackson ImmunoResearch (West Grove, PA). A Hoechst stain (1:100, Invitrogen) and RNeasy Lipid Tissue Mini kit (QIAGEN, 74804) were obtained. The Hito Golgi-Cox OptimStainTM PreKit was purchased from Hitobiotec Corp (HTKNS1125NH). Immunohistochemistry and immunofluorescence

ro of

Mice were sacrificed and perfused with 4% paraformaldehyde. Immunofluorescence and immunohistochemical (IHC) staining were performed as described previously with some

modifications [18]. For immunofluorescence staining, cerebellum sections were probed with a

primary antibody at 4 °C overnight, and then incubated with a secondary antibody and Hoechst

-p

stain for 1 h at room temperature. Images were obtained using a confocal fluorescence microscope (Olympus FV1000). For IHC staining, sections were incubated with a primary antibody at 4 °C

re

overnight and then separately incubated with a biotinylated secondary antibody and horseradish peroxidase (HRP)-conjugated streptavidin at room temperature for 30 min, followed by staining

DP72 digital camera.

na

Golgi staining

lP

with 3,3’-diaminobenzidine. Images were captured using an Olympus BX51 microscope and a

Golgi-Cox staining was conducted according to the manufacturer’s instructions (Hitobiotec Corp., USA). Briefly, whole brain tissues samples were removed rapidly, immersed in a mixture of

ur

solution 1 and solution 2 for 14 days in the dark, and then transferred to solution 3 for an incubation of 24-72 h at 4 °C in the dark. Sections (200-μm-thickness) were cut with a vibratome

Jo

(Leica CM3050S, Germany). For staining, the sections were mounted on 3% gelatin-coated glass slides, air-dried, stained with solutions 4 and 5, incubated in graded ethanol solutions, dehydrated and cleared with xylene. Purkinje cells were identified with a 20× objective and captured under a 40× objective. A total of 20–25 neurons were measured for each group. The length and distribution of dendrites were evaluated by Sholl analysis [36]. RNA sequencing and qRT-PCR Total RNA was extracted from the cerebellum using TriPure Isolation Reagent (Roche) according

to the manufacturer’s protocol. Then, the RNA samples were subjected to 50-bp single-end sequencing with a BGISEQ-500 sequencer as described previously [11]. At least 20 million clean reads of sequencing depth were obtained for each sample. The DEGs were defined as genes with a false discovery rate (FDR) less than 0.01 and a log2 fold change greater than 1 (upregulation) or smaller than −1 (downregulation). The DEGs were screened after gene annotation, normalization and statistical analyzed by an advanced method. We performed real-time qRT-PCR (qPCR) using gene-specific primer sets (Table S2), and we assessed relative gene expression (in triplicate) after normalization using a reference gene (Gapdh).

ro of

Western blot analysis Tissue samples were harvested from sacrificed mice and transferred to tubes. Proteins were

extracted using a protein extraction kit (Applygen Technologies Inc., P1250), separated by SDSPAGE and then transferred to PVDF membranes. The membranes were incubated with primary

-p

antibodies for 12 h at 4 °C, washed 3 times for 15 min, and incubated with secondary antibodies

for 1 h at 37 °C. Finally, the membranes were scanned with the Odyssey Infrared Imaging System

re

(LI-COR, Lincoln, NE). Statistical analyses

lP

Data were analyzed using GraphPad Prism 5 and Adobe Illustrator CS6. qPCR and Western blot data were compared by Student’s t-test, while the results of other experiments were compared by

na

the nonparametric Mann-Whitney test. Data are presented as the mean ± s.e.m. (error bars). P<0.05 was considered to indicate statistical significance.

ur

Competing interests. The authors declare no competing financial interests.

Jo

Animal ethical approval. All animal experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee of the Second Hospital of Hebei Medical University. Acknowlegements. This work was supported by the National Natural Science Foundation of China grants (31371089, 81171210).

Reference [1] M.A. Basson, D. Echevarria, C.P. Ahn, A. Sudarov, A.L. Joyner, I.J. Mason, S. Martinez, G.R. Martin, Specific regions within the embryonic midbrain and cerebellum require different levels of FGF signaling during development, Development, 135 (2008) 889-898. [2] A. Buffo, F. Rossi, Origin, lineage and function of cerebellar glia, Prog Neurobiol, 109 (2013) 42-63.

ro of

[3] X. Chen, X. Lan, S. Mo, J. Qin, W. Li, P. Liu, Y. Han, R. Pi, p38 and ERK, but not JNK, are involved in copper-induced apoptosis in cultured cerebellar granule neurons, Biochem Biophys Res Commun, 379 (2009) 944-948.

-p

[4] C.L. Chi, S. Martinez, W. Wurst, G.R. Martin, The isthmic organizer signal FGF8 is

re

required for cell survival in the prospective midbrain and cerebellum, Development, 130 (2003) 2633-2644.

lP

[5] R. Chopra, V.G. Shakkottai, Translating cerebellar Purkinje neuron physiology to progress in dominantly inherited ataxia, Future Neurol, 9 (2014) 187-196.

na

[6] B. Deng, W. Lv, W. Duan, Y. Liu, Z. Li, X. Song, C. Cui, X. Qi, X. Wang, C. Li, FGF9

ur

modulates Schwann cell myelination in developing nerves and induces a pro-inflammatory environment during injury, J Cell Biochem, 119 (2018) 8643-8658.

Jo

[7] M. Eiraku, A. Tohgo, K. Ono, M. Kaneko, K. Fujishima, T. Hirano, M. Kengaku, DNER acts as a neuron-specific Notch ligand during Bergmann glial development, Nat Neurosci, 8 (2005) 873-880.

[8] X. Feng, A.H. Juan, H.A. Wang, K.D. Ko, H. Zare, V. Sartorelli, Polycomb Ezh2 controls the fate of GABAergic neurons in the embryonic cerebellum, Development, 143 (2016) 1971-

1980. [9] N. Gaiano, J.S. Nye, G. Fishell, Radial glial identity is promoted by Notch1 signaling in the murine forebrain, Neuron, 26 (2000) 395-404. [10] C. Gordy, Y.W. He, The crosstalk between autophagy and apoptosis: where does this lead?, Protein Cell, 3 (2012) 17-27. [11] Y. Huang, Z. Xu, S. Xiong, F. Sun, G. Qin, G. Hu, J. Wang, L. Zhao, Y.X. Liang, T. Wu, Z.

ro of

Lu, M.S. Humayun, K.F. So, Y. Pan, N. Li, T.F. Yuan, Y. Rao, B. Peng, Repopulated microglia are solely derived from the proliferation of residual microglia after acute depletion, Nature Neuroscience, 21 (2018) 530-540.

-p

[12] Q. Hui, Z. Jin, X. Li, C. Liu, X. Wang, FGF Family: From Drug Development to Clinical

re

Application, Int J Mol Sci, 19 (2018).

[13] M. Ito, Cerebellar circuitry as a neuronal machine, Prog Neurobiol, 78 (2006) 272-303.

lP

[14] N. Itoh, H. Ohta, Y. Nakayama, M. Konishi, Roles of FGF Signals in Heart Development, Health, and Disease, Front Cell Dev Biol, 4 (2016) 110.

na

[15] M.V. Joglekar, P.M. Trivedi, T.W. Kay, W.J. Hawthorne, P.J. O'Connell, A.J. Jenkins,

ur

A.A. Hardikar, H.E. Thomas, Human islet cells are killed by BID-independent mechanisms in response to FAS ligand, Apoptosis, 21 (2016) 379-389.

Jo

[16] M. Le Gallo, A. Poissonnier, P. Blanco, P. Legembre, CD95/Fas, Non-Apoptotic Signaling Pathways, and Kinases, Front Immunol, 8 (2017) 1216. [17] Q. Li, M. Chen, H. Liu, L. Yang, T. Yang, G. He, The dual role of ERK signaling in the apoptosis of neurons, Front Biosci (Landmark Ed), 19 (2014) 1411-1417. [18] H. Lin, H. Hu, W. Duan, Y. Liu, G. Tan, Z. Li, B. Deng, X. Song, W. Wang, D. Wen, Y.

Wang, C. Li, Intramuscular Delivery of scAAV9-hIGF1 Prolongs Survival in the hSOD1(G93A) ALS Mouse Model via Upregulation of D-Amino Acid Oxidase, Mol Neurobiol, 55 (2018) 682695. [19] Y.S. Lin, L.J. Chen, C.H. Lin, Y.D. Luo, R.Y.L. Tsai, F. Wang, Neuron-derived FGF9 is essential for scaffold formation of Bergmann radial fibers and migration of granule neurons in the cerebellum, Dev Biol, 329 (2009) 44-54.

ro of

[20] M. Lou, L.N. Zhang, P.G. Ji, F.Q. Feng, J.H. Liu, C. Yang, B.F. Li, L. Wang, Quercetin nanoparticles induced autophagy and apoptosis through AKT/ERK/Caspase-3 signaling

pathway in human neuroglioma cells: In vitro and in vivo, Biomed Pharmacother, 84 (2016) 1-

-p

9.

re

[21] H. Marzban, M.R. Del Bigio, J. Alizadeh, S. Ghavami, R.M. Zachariah, M. Rastegar, Cellular commitment in the developing cerebellum, Front Cell Neurosci, 8 (2014) 450.

lP

[22] Y.A. Mebratu, I. Leyva-Baca, M.G. Wathelet, N. Lacey, H.S. Chand, A.M.K. Choi, Y. Tesfaigzi, Bik reduces hyperplastic cells by increasing Bak and activating DAPk1 to juxtapose

na

ER and mitochondria, Nat Commun, 8 (2017) 803.

ur

[23] J. Namiki, S. Suzuki, T. Masuda, Y. Ishihama, H. Okano, Nestin protein is phosphorylated in adult neural stem/progenitor cells and not endothelial progenitor cells,

Jo

Stem Cells Int, 2012 (2012) 430138. [24] F. Niu, K. Liao, G. Hu, S. Sil, S. Callen, M.L. Guo, L. Yang, S. Buch, Cocaine-induced release of CXCL10 from pericytes regulates monocyte transmigration into the CNS, J Cell Biol, 218 (2019) 700-721. [25] W. Pang, H. Lu, Y.D. Hu, H.P. Yang, X. Leng, Y.G. Jiang, Depletion of intracellular zinc

induced apoptosis in cultured hippocampal neurons through Raf/MEK/ERK pathways, Nutr Neurosci, 15 (2012) 18-24. [26] S. Sakaguchi, S. Katamine, N. Nishida, R. Moriuchi, K. Shigematsu, T. Sugimoto, A. Nakatani, Y. Kataoka, T. Houtani, S. Shirabe, H. Okada, S. Hasegawa, T. Miyamoto, T. Noda, Loss of cerebellar Purkinje cells in aged mice homozygous for a disrupted PrP gene, Nature, 380 (1996) 528-531.

ro of

[27] H.G. Shim, S.S. Jang, S.H. Kim, E.M. Hwang, J.O. Min, H.Y. Kim, Y.S. Kim, C. Ryu, G. Chung, Y. Kim, B.E. Yoon, S.J. Kim, TNF-alpha increases the intrinsic excitability of

cerebellar Purkinje cells through elevating glutamate release in Bergmann Glia, Sci Rep, 8

-p

(2018) 11589.

re

[28] D. Skinner, B.S. Marro, T.E. Lane, Chemokine CXCL10 and Coronavirus-Induced Neurologic Disease, Viral Immunol, 32 (2019) 25-37.

lP

[29] C.Y. Song, J.F. Guo, Y. Liu, B.S. Tang, Autophagy and Its Comprehensive Impact on ALS, Int J Neurosci, 122 (2012) 695-703.

na

[30] C. Sotelo, Cellular and genetic regulation of the development of the cerebellar system,

ur

Prog Neurobiol, 72 (2004) 295-339.

[31] H. Umemori, M.W. Linhoff, D.M. Ornitz, J.R. Sanes, FGF22 and its close relatives are

Jo

presynaptic organizing molecules in the mammalian brain, Cell, 118 (2004) 257-270. [32] S. Wang, Y. Li, C. Jiang, H. Tian, Fibroblast growth factor 9 subfamily and the heart, Appl Microbiol Biotechnol, 102 (2018) 605-613. [33] J. Xu, H. Dong, Q. Qian, X. Zhang, Y. Wang, W. Jin, Y. Qian, Astrocyte-derived CCL2 participates in surgery-induced cognitive dysfunction and neuroinflammation via evoking

microglia activation, Behav Brain Res, 332 (2017) 145-153. [34] K. Yoon, S. Nery, M.L. Rutlin, F. Radtke, G. Fishell, N. Gaiano, Fibroblast growth factor receptor signaling promotes radial glial identity and interacts with Notch1 signaling in telencephalic progenitors, J Neurosci, 24 (2004) 9497-9506. [35] I.O. Yusuf, H.M. Chen, P.H. Cheng, C.Y. Chang, S.J. Tsai, J.I. Chuang, C.C. Wu, B.M. Huang, H.S. Sun, S.H. Yang, Fibroblast growth factor 9 activates anti-oxidative functions of

ro of

Nrf2 through ERK signalling in striatal cell models of Huntington's disease, Free Radic Biol Med, 130 (2019) 256-266.

[36] K. Zhang, Q. Zhang, J. Deng, J. Li, J. Li, L. Wen, J. Ma, C. Li, ALK5 signaling pathway

-p

mediates neurogenesis and functional recovery after cerebral ischemia/reperfusion in rats via

Jo

ur

na

lP

re

Gadd45b, Cell Death Dis, 10 (2019) 360.

ro of -p re lP

na

Fig. 1 Conditional knockout of Fgf9 in GABAergic neurons in mice causes ataxia. (A) PCR analysis shows the genotyping of Fgf9loxP/loxP VGAT-cre (Fgf9VGAT), Fgf9loxP/loxP VGLUT1-cre (Fgf9VGlUT1) and control mice. (B) Fgf9VGAT mice showed ataxic gaits that lacked the forefoot–

ur

hindfoot correspondence observed in control mice. The front paws were labeled with red ink, and the rear paws with blue ink. The arrows indicate the direction of movement. (C) Fgf9VGAT mice

Jo

exhibited growth retardation beginning on postnatal day 10. Data are presented as the mean  s.e.m. for n=6 mice per genotype. (D) Adult Fgf9VGAT mice had 40-50% lower body weight than control littermates at 6 weeks. Data are presented as the mean  s.e.m. for n=20 mice per genotype.

ro of -p re lP na ur Jo

Fig. 2 Fgf9 ablation in GABAergic neurons disrupts the migration and alignment of cerebellar granule neurons. (A and B) Hematoxylin and eosin stain (HE staining) of midsagittal sections of F/F, Fgf9VGlUT1and Fgf9VGAT cerebellums collected on postnatal day 21. Defects in the migration of granule cells were found in Fgf9VGAT mice. Higher magnification views of the framed areas in the low-magnification images in rows 1 and 3 (Scale bar, 200 μm) are shown in rows 2 and 4 (Scale bar, 100 μm).

ro of -p re lP na ur

Fig. 3 Ablation of Fgf9 in GABAergic neurons affects the survival and development of

Jo

Purkinje cells. (A) Purkinje cells were lost and their alignment was disrupted in Fgf9VGAT cerebellum at 21 days after born, as observed by immunostaining for HE and Calbindin. Purkinje cells indicated by green arrows, and the loss of Purkinje cells indicated by black arrows. Scale bar, 50 μm. (B) Western blot analyses demonstrated an increase in calbindin expression in Fgf9VGAT mice at P21 compared to control (F/F) mice. β-actin was used as the internal loading control. (C) Golgi stains showed that Fgf9VGAT Purkinje cells had elongated primary dendritic bundles and reduced dendritic arborization compared to F/F Purkinje cells at 21 days after born. Scale bar, 20

μm. (D) Double immunostaining with Calbindin (red) and Cleaved caspase-3 (green) in cerebellum of F/F and Fgf9VGATmice at P21. Scale bar, 50 μm. Images are representative of n = 3 mice. Data

ur

na

lP

re

-p

ro of

with error bars represent mean ± SD.**P < 0.01 as determined by unpaired Student’s t test (B).

Jo

Fig. 4 Loss of Fgf9 in GABAergic neurons perturbs the formation of the radial fiber scaffold of Bergmann glia and activates microglia in the cerebellum. (A) Many GFAP-positive cells with a stellate morphology were found in the ML of the Fgf9VGAT cerebellum at 21 days after born. Scale bar, 50 μm. (B) An increase in the number of cells expressing Iba-1, a microglial marker, was found in the Fgf9VGAT cerebellum at 21 days after born. Higher magnification views of the framed areas in the low-magnification images in rows 1 (Scale bar, 200 μm) are shown in rows 2 (Scale bar, 50 μm). Images are representative of n = 3 mice.

ro of -p re lP na

5

RNA-sequencing analysis

of

differentially

expressed genes

(DEGs) and

ur

Fig.

signaling pathways in the cerebellum of Fgf9VGAT and F/F mice. (A) Volcano plot to visually

Jo

compare DEGs between Fgf9VGAT and F/F mice at 21 days. (B) Heatmap of DEGs between Fgf9VGAT mice and F/F mice. (C) KEGG pathway analysis showing changes in DEGs associated with cell growth and death in Fgf9VGAT mice and F/F mice. (D) qRT-PCR validation of some selected differentially expressed genes. Each panel is representative of at least three independent experiment. Significant differences versus the control: *P<0.05 and **P<0.01, unpaired Student’s t-test.

ro of -p re lP na ur Jo

Fig. 6 Western blot analyses demonstrate increased expression of apoptosis- and inflammation-related proteins in Fgf9VGAT mice compared to F/F mice. (A) The quantification in western blot analysis showed that Fas, Bid, Bik and cleaved caspase-3 expression increased in Fgf9VGAT mice at P21. (B) Western blotting analysis showed the levels of LC3, p62 and ubiquitin in

Fgf9VGAT and F/F mice at P21. (C and D) The expression of proinflammatory cytokines and inflammation-related proteins was increased at the mRNA and protein levels in Fgf9VGAT mice at P21. (E) MAPK pathways, especially the ERK pathway, were activated in Fgf9VGAT mice at P21. Images are representative of n = 3 mice. β-actin was used as the internal loading control. Data are

Jo

ur

na

lP

re

-p

ro of

presented as the mean ± SEM; *P<0.05 and **P<0.01, unpaired Student’s t-test.

log2(CKO/ Gene ID

P value(WT-

Gene

Kegg Orthology WT)

vs-CKO)

vs-CKO)

14282

Fosb

6.6775384

0

0

IL-17 signaling pathway

2

226896

Tfap2d

5.7419117

1.20E-138

2.00E-140

transcription factor AP-2 delta

3

12124

Bik

4.3964441

1.42E-05

3.97E-06

Bcl-2-interacting killer

4

14186

Fgfr4

3.1740516

0.000407

0.000139

fibroblast growth factor receptor 4

5

12047

Bcl2a1d

2.8297794

0.000337

0.000113

hematopoietic Bcl-2-related protein A1

6

13489

Drd2

2.7590141

4.27E-46

2.22E-47

dopamine receptor D2

7

214230

Pak6

2.677653

4.70E-76

1.50E-77

p21-activated kinase 6

8

100040048

Ccl27b

2.4714192

6.98E-10

1.28E-10

Chemokine signaling pathway

9

20296

Ccl2

2.4635583

2.08E-05

10

15945

Cxcl10

2.3964441

1.96E-06

11

20295

Ccl17

2.3666967

12

545192

Baiap3

2.3092109

13

14281

Fos

14

60367

15

54612

16

-p

re

C-C motif chemokine 2

4.94E-07

IL-17 signaling pathway

1.36E-05

3.81E-06

C-C motif chemokine 17

9.42E-73

3.21E-74

BAI1-associated protein 3

0

0

IL-17 signaling pathway

lP

5.92E-06

na 2.2771019

ro of

1

ur

SN

Q value(WT-

2.1198394

1.64E-21

1.66E-22

interleukin 1 receptor accessory protein-like

Sfrp5

2.0915895

4.86E-14

6.77E-15

Wnt signaling pathway

58208

Bcl11b

1.9432697

1.20E-45

6.27E-47

B-cell lymphoma/leukemia 11B

17

192199

Rspo1

1.8674142

7.47E-06

2.02E-06

R-spondin-1

18

235505

Cd109

1.8602172

6.36E-28

5.21E-29

CD109 antigen

19

14284

Fosl2

1.7317301

#######

#######

fos-like antigen 2

Jo

Il1rapl2

12484

Cd24a

1.6796917

#######

#######

CD24 antigen

21

269389

Tox2

1.6775271

2.55E-46

1.32E-47

toll-like receptor 13

22

12062

Bdkrb2

1.676552

3.81E-05

1.12E-05

Inflammatory mediator regulation of TRP channels

23

252838

Tox

1.6406194

1.76E-14

2.40E-15

toll-like receptor 13

24

225471

Ticam2

1.5516144

0.000927

0.000337

NF-kappa B signaling pathway

25

12575

Cdkn1a

1.5438541

0

0

cyclin-dependent kinase inhibitor 1A

26

14573

Gdnf

1.4667536

1.48E-08

3.01E-09

glial cell derived neurotrophic factor

27

107995

Cdc20

1.4471497

7.45E-19

8.39E-20

cell division cycle 20

28

12064

Bdnf

1.2817998

#######

#######

brain-derived neurotrophic factorase

29

12122

Bid

1.2520541

5.60E-10

1.02E-10

BH3 interacting domain death agonist

30

19094

Mapk11

1.1268641

1.48E-94

3.70E-96

p38 MAP kinase

Ripk4

1.0762764

1.0426008

-p

re

8.31E-05

2.57E-05

0.000223

0.000568

7.28E-05

6 transcriptional activator Myb receptor-interacting serine/threonine-protein kinase

0.000198 4

ur

72388

Myb

1.0856565

tumor necrosis factor receptor superfamily member

-1.030866

2.16E-27

1.80E-28

fibroblast growth factor

Jo

33

17863

Fas

lP

32

14102

na

31

ro of

20

Cd80

-1.176561

0.000371

0.000125

CD80 antigen

116701

Fgfrl1

-1.297936

4.72E-33

3.33E-34

fibroblast growth factor receptor 4

37

21948

Cd70

-2.803228

2.71E-05

7.83E-06

tumor necrosis factor ligand superfamily member 7

38

24117

Wif1

-2.893662

7.68E-64

2.95E-65

WNT inhibitory factor 1

39

13142

Dao

-2.9901

0

0

D-amino-acid oxidase

34

14178

35

12519

36

Fgf7

100043324

Bnip3l-ps

-8.220454

3E-31

2.2E-32

BCL2/adenovirus E1B interacting protein 3-like

ur

na

lP

re

-p

ro of

Table. 1 Some genes were up- and down- regulated in the FGF9VGAT compared to F/F.

Jo

40