Defective Astrocyte Maturation Drives Cerebellar Neuroinflammation and Degeneration

ABSTRACT

While persistent neuroinflammation and neurodegeneration are hallmarks of many diseases, the exact mechanisms triggering neurodegeneration are not fully established. Neurodegeneration is accompanied by activation of astrocytes that can have both neuroprotective and neurotoxic functions. Much less is known about how intrinsic dysfunction of astrocytes can lead to neuroinflammation and neurodegeneration. To study astrocyte‐driven neurodegeneration, we examined aging cerebella of adult astrocyte‐specific Yin Yang1 (Yy1) conditional knockout mice that contain improperly matured dysfunctional astrocytes. We found that deletion of Yy1 from astrocytes during development results in subsequent cerebellar neurodegeneration in adult mice. The neurodegeneration was accompanied by profound changes in astrocyte morphologies and expression of astrocyte‐specific genes, and development of severe neuroinflammation that preceded cerebellar neurodegeneration and Purkinje cell (PC) loss. Mechanistically, we found that sustained β‐catenin expression by Bergmann glia (BG) correlated with their decreased adenomatous polyposis coli (APC) expression and diminished expression of synaptic proteins by glutamatergic neurons, suggesting that Yy1 supports astrocytic APC expression needed for β‐catenin degradation and proper BG morphology. Our findings highlight the critical role of YY1 in sustaining cerebellar astrocyte functions and suggest that dysfunction of astrocytes has widespread consequences for cerebellar integrity, function, and leads to neurodegeneration.

In the Bergmann glia (BG) of the cerebellum, YY1 supports expression of glutamate transporters, GLT1 and GLAST, which uptake glutamate. YY1 also enhances expression of adenomatous polyposis coli (APC) needed for β‐catenin degradation and the acquisition of proper BG morphology. In the absence of YY1, decreased APC expression in BG leads to the accumulation of β‐catenin and morphological abnormalities. In addition, diminished GLT1 and GLAST expression leads to the accumulation of excitotoxic glutamate, microglia activation, and cerebellar degeneration.

Abbreviations

APC

adenomatous polyposis coli

BG

Bergmann glia

BLBP

brain lipid‐binding protein

CNS

central nervous system

FA

fibrous astrocytes

FC

Fañanas cells

GCL

granular cell layer

GFAP

glial fibrillary acidic protein

MOL

molecular layer

PC

Purkinje cells

PCL

Purkinje cell layer

VA

velate astrocytes

VGLUT1

vesicular glutamate transporter 1

VGLUT2

vesicular glutamate transporter 2

WM

white matter

YY1

Yin Yang 1

1. Introduction

Neurodegeneration characterized by a progressive loss of neurons is a hallmark of many diseases, including Alzheimer's disease, Huntington's disease, Parkinson's disease, amyotrophic lateral sclerosis, multiple sclerosis, and many other rarer neurodegenerative diseases [1]. While genetic, environmental, and endogenous factors have been implicated in these neurodegenerative diseases, the exact mechanisms triggering neurodegeneration remain to be fully established [2, 3]. The implicated processes include oxidative stress, mitochondrial dysfunction, abnormal accumulation of proteins, and neuroinflammation [1, 2]. Neuroinflammation is a protective mechanism that develops in response to injury and infection to prevent damage and to restore homeostasis [4, 5, 6, 7]. This response involves the resident innate immune and nonimmune cells of the brain [7, 8, 9] as well as cells of the adaptive immune system [7, 10]. Although neuroinflammation promotes tissue repair, unresolved chronic neuroinflammation, characterized by persistent activation of microglia and sustained presence of proinflammatory cytokines and chemokines, leads to neurodegeneration [11, 12, 13]. While the critical role of microglia in neurodegenerative disorders has been well established, astrocytes have recently attracted increased attention [13, 14, 15, 16]. Astrocytes, the most abundant glial cells in the brain, play critical physiological functions by maintaining energy and ion balance, regulating the blood–brain barrier, synapses, and concentrations of neurotransmitters [17, 18, 19]. Neurodegenerative disorders are associated with a state of reactive astrogliosis characterized by changes to the morphological, metabolic, transcriptional, and physiological properties of astrocytes [13, 14, 15, 16]. However, reactive astrocytes can be both neuroprotective and neurotoxic [20, 21], and they may lose normal astrocytic functions but also gain new neurotoxic functions, rapidly killing neurons [21]. The responses of astrocytes are also very heterogeneous, driven by their diversity, specific location, and contexts of their responses to different stimuli [22, 23, 24, 25]. Nevertheless, much less is known of how intrinsic dysfunction of astrocytes can lead to neuroinflammation and neurodegeneration.

The transcription factor Yin Yang 1 (YY1) is ubiquitously expressed in mammalian cells and regulates multiple processes, including embryogenesis [26], cell cycle [27, 28], X chromosome inactivation [29], oncogenesis [30], and differentiation [31, 32, 33, 34, 35]. YY1 interacts with other regulatory proteins [27, 36, 37, 38, 39, 40, 41, 42, 43] and also creates active loops of chromatin [35, 42, 43]. In the central nervous system, YY1 binding reorganizes chromatin architecture during early neural lineage commitment [35, 42], is essential for the development of the neuroepithelium [44], and supports neural progenitor cell proliferation and survival [45]. YY1 also affects neuroinflammation by promoting M2 microglia polarization [46]. YY1 is also highly expressed in astrocytes [47, 48], can mitigate manganese‐induced neurotoxicity [49] and provides protection against apoptosis, oxidative stress, and inflammation [50]. Furthermore, YY1 is required for proper maturation of cerebellar astrocytes during development and the maintenance of the mature astrocyte phenotype in the adult cerebellum [51]. Despite these well‐documented roles of astrocytic YY1, little is known about the consequences of the presence of dysfunctional YY1‐deficient astrocytes that did not mature properly. In this study, we found that the presence of the cerebellar astrocytes that did not mature properly initially leads to the development of neuroinflammation, characterized by activation of microglia and secretion of proinflammatory cytokines. Subsequently, persistent neuroinflammation promotes cerebellar neurodegeneration that manifests with the loss of synapses and Purkinje cells (PCs) degeneration. Our findings highlight the critical role of YY1 in sustaining cerebellar health and suggest that astrocyte dysfunction can have widespread consequences for cerebellar integrity and function.

2. Materials and Methods

2.1. Mice

The generation of Yy1 ^GFAP‐CRE (Yy1 ^ΔAST ) mice has been previously described [51]. These mice were obtained by crossing mice carrying the Yy1 allele flanked by loxP sites (Jackson Laboratory; strain B6;129S4‐Yy1tm2Yshi/J) with Gfap‐Cre mice (Jackson Laboratory; strain 77.6mGFAPcre). Subsequently, heterozygous Yy1 ^{WT/LoxP;GFAP‐CRE} mice were crossed with Aldh1l1‐EGFP mice (provided by Dr. Cagla Eroglu, Duke University) to generate Yy1 ^{ΔAST;Aldh1L1‐EGFP} mice expressing EGFP in astrocytes. All mice were accommodated at Virginia Commonwealth University following the guidelines set by the Institutional Animal Care and Use Committee. They were subjected to a 12‐h light/dark cycle, fed standard laboratory chow, and had unrestricted access to water. Yy1 ^ΔAST and Yy1 ^{ΔAST;Aldh1L1‐EGFP} mice exhibiting severe symptoms were supplemented with DietGel 76A (ClearH20). Both male and female littermates were chosen randomly for experimental purposes, with group sizes outlined in the figure legends.

2.2. Hematoxylin and Eosin (H&E) Staining

Histological analysis was performed as previously described [52, 53]. Animals were first perfused with 0.9% Sodium Chloride and then 4% paraformaldehyde. Subsequently, brains were postfixed in 4% paraformaldehyde for 48 h. Tissue samples were then embedded in paraffin, sectioned to a thickness of 5 μm, stained with H&E, and imaged using the Vectra Polaris Imaging system at the VCU's Cancer Mouse Models Core Facility.

2.3. Immunofluorescence

Animals were perfused with 0.9% Sodium Chloride followed by 4% paraformaldehyde. Brains were postfixed in 4% paraformaldehyde overnight and then cryopreserved in 30% sucrose in PBS for 48 h at 4°C. Subsequently, brains were embedded in optimal cutting medium (Tissue‐Tek, VWR), and 50 μm frozen sections were prepared on probe plus microscopy slides (Fisher Sci). For permeabilization, a solution of 1% Triton X‐100in PBS was applied for 1 h, followed by blocking in PBS containing 0.5% Triton X‐100 and fish skin gelatin (Electron Microscopy) for 1 h. Primary antibodies Chicken Anti‐GFAP (Aves Lab., GFAP, 1:500), Rabbit Anti‐IBA1 (FUJIFILM, 016‐20001, 1:200), Rabbit Anti‐β‐Catenin (Millipore, ABE208, 1:200), Rabbit Anti‐NeuN (Cell Signaling Tech., D4G40, 1:200), Rabbit Anti‐Calbindin (Cell Signaling Tech., D1I4Q, 1:200), Mouse Anti‐APC (CC1) (Millipore, OP80, 1:200), Rabbit Anti‐GLAST (Cell Signaling Tech., D44E2, 1:200), Guinea Pig Anti‐GLT1 (Millipore, AB1783, 1:200), Rabbit Anti‐Aquaporin (CiteAB, AQP‐004, 1:200), Rabbit Anti‐S100β (Cell Signaling Tech., E7C3A, 1:200), Guinea Pig Anti‐VGLUT1 (Millipore, AB5605, 1:100), Guinea Pig Anti‐VGLUT2 (Millipore, AB2251‐1, 1:100) and Rabbit Anti‐BLBP (Abcam, ab32423, 1:200) were diluted in the blocking buffer, applied onto slides, and incubated overnight at 4°C. Subsequently, sections underwent three washes before being exposed to the Alexa Fluor‐488, −594, or −633 labeled secondary antibodies (diluted at 1:1000, Invitrogen) for 45 min at 37°C. After another three washes, slides were mounted using Vectashield mounting medium (Vector Laboratories) and imaged using the Zeiss LSM 880 confocal microscope. Maximum projection images were generated from z‐stacks, ensuring no fluorescence crossover between channels. Subsequent image analysis was carried out using either ImageJ or Imaris 9.7 software. Three to four animals per group were examined, and representative images are shown.

2.4. Cell Morphology Analysis

Image analysis was conducted using Imaris 9.7 software (Bitplane) or ImageJ. The Imaris software was employed for the quantification of cell processes (filament tracer), volume assessment (surface tool), determination of BG angle and orientation (filament tracer), cell count, and analysis of VGLUT2 puncta (spots tool). ImageJ was used to examine mean fluorescence intensity and thickness analysis.

2.5. Quantitative RT‐PCR

Gene expression was analyzed by qRT‐PCR as previously described [54]. Cerebella were flash‐frozen, tissue was powdered in mortar, and total RNA was extracted using Trizol (Life Technologies). cDNA was prepared using the high‐capacity cDNA kit (Applied Biosystems). The resulting cDNA was then amplified utilizing the BioRad CFX Connect Real‐Time System. SYBR Green intron‐spanning predesigned qPCR primers from BioRad were utilized for the amplification process. Expression levels of genes were normalized to GAPDH and presented as fold change relative to the control.

2.6. Western Blotting

Western blotting was performed as described [54]. Briefly, powdered tissues underwent lysis in a buffer composed of 10 mM Tris (pH 7.4), 150 mM sodium chloride, 1 mM EDTA, 0.5% Nonidet P‐40, 1% Triton X‐100, 1 mM sodium orthovanadate, 0.2 mM PMSF, and a protease inhibitor mixture (Roche Applied Science). Subsequently, samples were subjected to SDS–PAGE separation and transferred onto nitrocellulose membranes. Membranes were incubated with primary anti‐GFAP (Santa Cruz Biotech., sc‐33673, 1:1000), anti‐GLAST (Santa Cruz Biotech., sc‐365634, 1:1000), anti‐GLT1 (Invitrogen, PA5‐17099, 1:1000), or anti‐β tubulin (Santa Cruz Biotech., sc‐9104, 1:1000) antibodies diluted in TTBS buffer containing 5% dry milk overnight at room temperature. Membranes were extensively washed in TTBS and incubated with HRP‐labeled secondary antibodies for 1 h at room temperature. Visualization of antigen–antibody complexes was achieved using ECL (Immobilon Western blotting kit, Millipore).

2.7. Statistical Analysis

Statistical analysis was performed using GraphPad Prism 7 software. Quantitative data are expressed as mean ± SEM or mean ± SD (as specified). Sample sizes are indicated in figure legends. Statistical analysis was conducted using two‐tailed t‐tests, and the significance levels are indicated in the figure legends.

3. Results

3.1. Deletion of Yy1 From Astrocytes Results in Cerebellar Neurodegeneration

YY1 is ubiquitously expressed during brain development and regulates neuroepithelium [44] and proliferation and survival of neural progenitor cells [45]. It also functions as either a transcriptional activator or repressor in the adult brain in neurons, astrocytes, microglia, and oligodendrocytes [31, 44, 45, 55]. YY1 protects astrocytes against apoptosis, oxidative stress, and inflammation [50]. We have recently shown that YY1 regulates cerebellar astrocyte maturation during development and the maintenance of their mature phenotype in the adult cerebellum [51]. In the current study, we aimed to understand the long‐term consequences of the presence of Yy1‐deficient dysfunctional astrocytes in the aging cerebellum. We examined brains of 5–7‐month‐old Yy1 ^ΔAST mice in which the Yy1 gene is specifically deleted from cells expressing GFAP [51]. We found a significant decrease in the weights of the cerebella in Yy1 ^ΔAST mice compared to their control littermates (Figure 1A,B). Furthermore, the presence of degenerative morphological changes was evident in Yy1 ^ΔAST cerebella visualized by both H&E (Figure 1C) and Calbindin (Figure S1) staining. The most pronounced changes were observed in the anterior and central domains of the Yy1 ^ΔAST cerebella, particularly in lobes I–VII (Figure S1C,D). Detailed analysis of the length of the rostrocaudal axis and dorsoventral axis of the cerebella revealed a significant reduction in the cerebellar sizes in Yy1 ^ΔAST mice (Figure 1D,E). Furthermore, the thickness of the cerebellar molecular layer (MOL), granular cell layer (GCL), and white matter (WM) was significantly decreased in Yy1 ^ΔAST mice (Figure 1F), suggesting cerebellar neurodegeneration.

FIGURE 1.

Cerebellar neurodegeneration in 5–7‐month‐old the Yy1 ^ΔAST mice. (A) Comparison of the Yy1 ^loxP/loxP and Yy1 ^∆AST brain morphology (n = 3 mice in each group). (B) Weights of the cerebella (n = 8, 7, two‐tailed t‐test; **p < 0.01). (C) H&E staining of the Yy1 ^loxP/loxP and Yy1 ^∆AST brains. (D) The comparison of rostrocaudal axis length of the Yy1 ^loxP/loxP and Yy1 ^∆AST cerebella (n = 3 for each group, two‐tailed t‐test; ****p < 0.0001). (E) The length of Dorso‐Ventral axis length of the Yy1 ^loxP/loxP and Yy1 ^∆AST cerebella (n = 3 for each group, two‐tailed paired t‐test. *p < 0.05). (F) Thickness of three different layers of the Yy1 ^loxP/loxP and Yy1 ^∆AST cerebella, (n = 3 for each group, two‐tailed t‐test. *p < 0.05; ****p < 0.0001). GCL, granular cell layer; MOL, molecular layer; WM, white matter. (D–F) Lobes IV–V were analyzed.

3.2. Profound Changes in Cerebellar Astrocyte Morphologies and Protein Expression Induced by Deletion of Yy1

To evaluate the impact of Yy1 deletion on astrocytes in aging cerebellum, we subsequently conducted comprehensive morphological analyses of GFAP‐positive Bergmann Glia (BG), velate astrocytes (VA), and fibrous astrocytes (FA) (Figure 2A). This analysis showed that all three astrocyte subpopulations had significantly decreased cell volumes, filament lengths, and numbers of dendrite segments in Yy1 ^ΔAST cerebella (Figure 2B–D). In addition, the dendrite diameters of BG were significantly increased (Figure 2B), while cell areas of VA and FA were reduced in Yy1 ^ΔAST cerebella (Figure 2C,D). Interestingly, while BG were GFAP/BLBP positive in control littermates, most of them were GFAP positive but BLBP negative in Yy1 ^ΔAST cerebella (Figure 2E). In agreement with the above observations, Sholl analysis demonstrated significant changes in the number of intersections for BG, VA, and FA, with the most dramatic changes in BG (Figure 2F). These morphological changes were accompanied by increased GFAP intensity in the MOL, but no changes in GCL and WM of Yy1 ^ΔAST cerebella were observed (Figure 2. G). Moreover, Yy1 deficient astrocytes showed changes in dendritic branching (Figure 2H) and dendritic orientation (Figure 2I) angles. We also observed drastically diminished numbers of BG in the MOL of Yy1 ^ΔAST cerebella (Figure 2J). Since the most pronounced changes in the cerebellar structure were observed in lobes I–VII (Figure S1), we subsequently examined astrocyte morphology specifically in lobes IV/V, which exhibited the most pronounced structural changes. For comparison, we also examined lobe IX, which showed limited changes. While structural changes between lobes IV/V and IX were obvious, BG, VA, and FA displayed relatively comparable changes in lobes IV/V (Figure S2) and lobe IX (Figure S3) of Yy1 ^ΔAST cerebella. In addition to the three major subpopulations of cerebellar astrocytes, Fañanas cells (FCs) characterized by unipolar morphology, stubby processes adorned with numerous short protrusions, and small cell bodies situated in the MOL have been described [56, 57, 58]. The majority of FCs are predominantly found in Lobe IX/X [58]. We examined FC morphology in lobes IV/V and IX. We found only small differences in the morphologies of these cells in Yy1 ^ΔAST cerebella, such as larger dendritic diameter and fewer segments in lobe IV/V (Figure S4A–D). Nevertheless, these cells are morphologically very different from BG (Figure S4E,F). Overall, our data indicate that Yy1 is needed to sustain normal morphologies of various astrocyte subtypes in aging mouse cerebellum.

FIGURE 2.

Dramatic differences in morphology, astrocyte processes, and GFAP expression in the three subpopulations of the cerebellar astrocytes in 5–7‐month‐old Yy1 ^ΔAST mice. (A) GFAP IF staining of 5–7‐month‐old cerebella. DAPI was used to visualize nuclei. (B) Analysis of Bergmann glia morphology in the molecular layer. Mean diameter of branches, average volume, filament length, and filament number of dendrite segments were examined. (C) Assessment of velate astrocyte morphology in the granular cell layer. Cell area, cell volume, filament length, and filament number per dendritic segment were examined. (D) Analysis of fibrous astrocyte morphology in white matter. Cell area, cell volume, filament length, and filament number were examined. (E) GFAP/BLBP IF staining of 5–7‐month‐old cerebella. DAPI was used to visualize nuclei. MOL, molecular layer; GCL, granular cell layer; and WM, white matter. (F) Sholl analysis for Bergmann glia, velate astrocytes, and fibrous astrocytes. (G) Quantification of GFAP fluorescence intensity in MOL, molecular layer; GCL, granular cell layer; and WM, white matter. (H) Analysis of angles of Bergmann glia processes. (I) The orientation analysis of Bergmann glia processes. (J) Quantification of the number of BG. n = 3 mice for each group; two‐tailed t‐test; ****p < 0.0001. (A–J) n = 3 mice for each group; two‐tailed t‐test; *p < 0.05; **p < 0.01; ***p < 0.0005; ****p < 0.0001, ns, not significant.

To assess whether morphological differences of astrocytes correlate with altered expression of the key astrocyte proteins, we examined the expression levels of GLAST, GLT1, S100β, and AQP4 (Figure 3A,B). Expression of GLAST and S100β, which are mostly expressed in the MOL, was significantly reduced in Yy1 ^ΔAST cerebella. Expression of GLT1, which is expressed in both the MOL and GCL, was also significantly reduced in both of these layers in Yy1 ^ΔAST cerebella, while expression of AQP4 remained unchanged. Additionally, qPCR data indicated that while Gfap and Nestin mRNA expression was significantly increased in Yy1 ^ΔAST cerebella indicating astrocyte activation, the expression of mRNAs for glutamate transporters, Slc1a3 (GLAST) and Slc1a2 (GLT1), and mRNAs for glutamate receptors, Gria1 and Gria4, was either significantly decreased or trending lower in Yy1 ^ΔAST cerebella (Figure 3C). The expression of Aqp4 mRNA remained unchanged in Yy1 ^ΔAST cerebella (Figure 3C). These data indicate that Yy1 is essential for the accurate expression of critical astrocyte proteins by cerebellar astrocytes.

FIGURE 3.

Diminished expression of critical astrocytic proteins in 5–7‐month‐old Yy1 ^ΔAST mice. (A) IF costaining for GFAP with either GLAST, GLT1, S100β, or AQP4 depicted on low‐magnification (white box) and high‐magnification in 5–7‐month‐old cerebella. DAPI was used to visualize nuclei. (B) Quantification of GLAST, GLT1, S100β, or AQP4 IF (n = 3 mice for each group; two‐tailed t‐test; *p < 0.05; ***p < 0.0005; ****p < 0.0001; ns, not significant). (C) Expression analysis (qPCR) of the indicated mRNAs expressed by astrocytes. n = 8 and 6 mice, two‐tailed t‐test; **p < 0.01; ***p < 0.0005; ****p < 0.0001; ns, not significant.

3.3. Severe Neuroinflammation in the Cerebella of 5–7‐Month‐Old Yy1^ΔAST Mice

Since astrocytes and microglia continuously cross‐communicate [59, 60], we examined the morphology and numbers of IBA1‐positive cells in the MOL, GCL, and WM in 5–7‐month‐old cerebella of Yy1 ^loxP/loxP and Yy1 ^∆AST mice (Figure 4A–G). Microglia of the Yy1 ^∆AST MOL were characterized by significantly larger areas, volumes, dendritic diameters, and shorter filament lengths (Figure 4A,B), indicating their activation. Similarly, activation of microglia was found in the Yy1 ^∆AST GCL with increased areas and dendritic diameters but reduced filament lengths (Figure 4C,D), while microglia in the Yy1 ^∆AST WM did not exhibit significant morphological changes (Figure 4C,E). Nevertheless, increased numbers of IBA1‐positive cells were found in all three layers of the Yy1 ^∆AST cerebellum, and the increase was most prominent in the MOL (Figure 4F). Subsequently, Sholl analysis indicated marked differences in microglia's intersections across radii from the soma in all three cerebellar layers (Figure 4G). Consistent with the observed morphological changes indicating microglia activation, we found increased mRNA expression of the major proinflammatory cytokines (IL‐1β, TNFα, and IL‐6), chemokines (CCL5 and CXCL10), as well as the microglia marker (SiglecH) in Yy1 ^∆AST cerebella (Figure 4H). Overall, we concluded that severe neuroinflammation associated with activated microglia and induced expression of proinflammatory markers develops during the aging of cerebella containing dysfunctional Yy1‐deficient astrocytes.

FIGURE 4.

Development of severe neuroinflammation in the cerebella of 5–7‐month‐old Yy1 ^ΔAST mice. (A) Iba1 IF staining of the molecular layer (MOL) of cerebella. DAPI was used to visualize nuclei. (B) Morphological analysis of Iba1 positive cells in the MOL. Soma area, cell volume, dendrite diameter, and filament length were examined. (n = 3 mice in each group; two‐tailed t‐test; *p < 0.05; **p < 0.01; ***p < 0.0005; ****p < 0.0001). (C) Iba1 IF staining of the granular cell layer (GCL) and white matter (WM). DAPI was used to visualize nuclei. (D and E) Morphological analysis of Iba1 positive cells in the GCL (D) and WM (E). Soma area, cell volume, dendrite diameter, and filament length were examined. n = 3 mice in each group; two‐tailed t‐test; *p < 0.05; ****p < 0.0001, ns, not significant. (F) Number of Iba1 positive cells in the three layers of cerebellum (n = 3 mice in each group; two‐tailed t‐test; *p < 0.05; ****p < 0.0001). (G) Sholl analysis of Iba1 positive cell in the three cerebellar layers. n = 3 mice for each group. (H) Expression analysis (qPCR) of the indicated neuroinflammatory markers. (n = 8 and 6 mice; two‐tailed t‐test; *p < 0.05; **p < 0.01; ***p < 0.0005; ****p < 0.0001).

3.4. Purkinje‐Cell Degeneration in the Cerebella of 5–7‐Month‐Old Yy1^ΔAST Mice

Since aging Yy1 ^ΔAST mice had significantly smaller cerebella containing reactive astrocytes and activated microglia, we subsequently examined whether this inflammatory cerebellar environment affects PCs forming synapses in the MOL. We stained PCs for Calbindin, which is their well‐established marker (Figure 5A). We found significantly decreased numbers of Calbindin‐positive PC cells in Yy1 ^ΔAST cerebella (Figure 5A,B). Furthermore, the location of PC somas was significantly altered in Yy1 ^∆AST cerebella with PC bodies often aberrantly located in the MOL or GCL (Figure 5C), indicating disruption of the normal architecture. Indeed, morphological properties of PC in Yy1 ^∆AST cerebella were significantly altered with reduced cell volumes, filament lengths, and dendritic segments compared to Yy1 ^loxP/loxP cerebella (Figure 5D). Sholl analysis further confirmed differences in the number of intersections around the soma of between PC of Yy1 ^loxP/loxP and Yy1 ^∆AST mice (Figure 5E). While the numbers of PC and their morphology were clearly affected in Yy1 ^∆AST cerebella, expression of neuronal‐specific Rb‐foxP3 mRNA (Figure 5F) and its corresponding protein NeuN (Figure 5G,H) was not altered in Yy1 ^∆AST cerebella. However, we noticed that NRG1 expression was significantly increased (Figure 5F). These results collectively suggest that Yy1 deletion in cerebellar astrocytes indeed induces neurodegeneration impacting neurons, particularly PC, in aging mice.

FIGURE 5.

Degeneration of Purkinje cells in the cerebella of 5–7‐month‐old Yy1 ^ΔAST mice. (A) Calbindin IF staining of the cerebella. DAPI was used to visualize nuclei. (B) Quantification of the number of Calbindin‐positive cells. n = 3 mice for each group; two‐tailed t‐test; ****p < 0.0001. (C) Distance of Calbindin‐positive cells from the Purkinje cell layer (PCL). n = 3 mice for each group; two‐tailed t‐test; **p < 0.01. (D) Analysis of Purkinje cell morphologies, including cell volume, filament length, and the number of filament segments. n = 3 mice for each group; two‐tailed t‐test; *p < 0.05. (E) Sholl analysis of Calbindin‐positive cells. (F) Expression analysis (qPCR) of the indicated markers. n = 8 and 6 mice; two‐tailed t‐test; ***p < 0.0005, ns, not significant. (G and H) Lobes IV–V, (G) NeuN IF staining, (H) Quantification of NeuN‐positive cells. n = 3 mice for each group, two‐tailed t‐test, ns, not significant.

3.5. Neuroinflammation and Aberrant Gene Expression in 7–8‐Week‐Old Yy1^ΔAST Mice Precede Cerebellar Neurodegeneration

Although cerebellar neurodegeneration was evident in 5–7‐month‐old Yy1 ^ΔAST mice, these mice do show obvious signs of neurodegeneration at 8–12 weeks [51]. We decided to use younger animals (7–8 weeks old) and examine whether altered gene expression and inflammation precede subsequent neurodegeneration. In contrast to what we found in 5–7‐month‐old animals, the length of the rostrocaudal axis and dorsoventral axis of the cerebella was not significantly changed in 7–8‐week‐old Yy1 ^ΔAST mice (Figure 6A,B). However, expression of the proinflammatory cytokine mRNAs encoding IL‐1β and TNFα was drastically increased, suggesting an ongoing inflammatory response (Figure 6C). This ongoing inflammatory response seems not to strongly correlate with microglia polarization since expression of mRNAs encoding iNOS, ARG1, IL‐10, and FIZZ1 was only mildly affected (Figure 6C). Significantly, the expression of Gfap mRNA was drastically enhanced (Figure 6D), indicating astrocyte activation. Since glutamate excitotoxicity is associated with neurodegeneration [61], we examined expression of glutamate transporters in the cerebella of 7–8‐week‐old mice. Indeed, expression of GLAST (Slc1A3) and GLT‐1 (Slc1A2) protein and mRNAs were significantly diminished in the cerebella of Yy1 ^ΔAST mice (Figure 6D,E). This decrease correlated with a simultaneous decrease in the expression of mRNAs encoding AMPA receptors, Gria1 and Gria4, in the cerebella of Yy1 ^ΔAST mice (Figure 6F). Thus, the deletion of Yy1 from astrocytes during development manifests as altered gene expression and development of neuroinflammation in young animals that subsequently leads to cerebellar neurodegeneration in older animals.

FIGURE 6.

Profound changes in the cerebellar gene expression already at 7–8‐week‐old Yy1 ^ΔAST mice. (A) The Rostro‐Caudal axis length of the Yy1 ^loxP/loxP and Yy1 ^∆AST cerebella (n = 3 for each group, two‐tailed t‐test, ns, not significant). (B) The Dorso‐Ventral axis length of the Yy1 ^loxP/loxP and Yy1 ^∆AST cerebella (n = 3 for each group, two‐tailed paired t‐test. ns, not significant). (C) Expression analysis (qPCR) of the indicated inflammation‐associated mRNAs. n = 10 and 11 mice, two‐tailed t‐test; *p < 0.05; **p < 0.01; ****p < 0.0001; ns, not specific. (D) Western blot analysis of cerebellar lysates depicting changes in protein expression (GFAP, GLAST, and GLT1). B‐tubulin is a loading control. (E and F) Expression analysis (qPCR) of the indicated mRNAs expressed by astrocytes. n = 10 and 11 mice, two‐tailed t‐test; **p < 0.01; ****p < 0.0001.

3.6. Profound Changes in Cerebellar Astrocyte Morphologies in 7–8‐Week‐Old Yy1^{ΔAST;Aldh1L1‐EGFP} Mice

Since GFAP staining does not visualize fine astrocyte processes [62], we generated Yy1 ^{ΔAST;Aldh1L1‐EGFP} mice by crossing Yy1 ^ΔAST mice with mice expressing EGFP under the control of the Aldh1L1 promoter [63]. We examined astrocyte morphologies in 7–8‐week‐old cerebella of control and Yy1 ^{ΔAST;Aldh1L1‐EGFP} mice simultaneously costaining for GFAP (Figure 7A). Focusing on EGFP‐labeled cells, we found that the morphologies of all three major astrocyte subpopulations, BG, VA, and FA, were dramatically affected, including cell areas, volumes, filament lengths, and the number of dendrite segments (Figure 7B–D). Additionally, while EGFP fluorescence intensity was decreased in the MOL of Yy1 ^{ΔAST;Aldh1L1‐EGFP} mice (Figure 7E), the GFAP fluorescence signal was significantly increased (Figure 7F) suggesting strong activation of BG. This was further supported by Sholl analysis of astrocytes that showed dramatic changes in BG (Figure 7G).

FIGURE 7.

Changes in morphologies of the three subpopulations of the cerebellar astrocytes in 7–8‐week‐old Yy1 ^{ΔAST;Aldh1L1‐EGFP} mice. (A–G) Lobes IV–V. (A) GFAP IF staining of 7–8‐week‐old cerebella. DAPI was used to visualize nuclei. MOL, molecular layer; GCL, granular cell layer; and WM, white matter. Cells expressing EGFP (green). (B) Analysis of Bergmann glia morphology in the molecular layer. Cell area, cell volume, filament length, and filament number of dendrite segments were examined. (C) Assessment of velate astrocyte morphology in the granular cell layer. Cell area, cell volume, filament length, and filament number per dendritic segment were examined. (D) Analysis of fibrous astrocyte morphology in the white matter. Cell area, cell volume, filament length, and filament number were examined. (E and F) Quantification of EGFP (E) and GFAP (F) IF. (G) Sholl analysis for Bergmann glia, velate astrocytes, and fibrous astrocytes. (A–G) n = 3 mice in each group; two‐tailed t‐test; *p < 0.05; ***p < 0.0005; ****p < 0.0001, ns, not significant).

3.7. Sustained β‐Catenin Expression by BG Correlates With Their Decreased CC1 Expression and Diminished Expression of Synaptic Proteins by Glutamatergic Neurons

The pivotal role of adenomatous polyposis coli (APC) in the establishment of the unique BG morphology has been reported, and the loss of APC from BG induces nonautonomous neurodegeneration of Purkinje neurons [64]. Since the morphology of BG was dramatically affected by the deletion of Yy1, we examined APC (anti‐CC1) expression by astrocytes in both 5–7‐month‐old Yy1 ^ΔAST mice (Figure 8A,B) and 7–8‐week‐old Yy1 ^{ΔAST;Aldh1L1‐EGFP} mice (Figure 8C,D). The expression of APC by both GFAP‐positive cells (Figure 8B) and EGFP‐positive cells (Figure 8D) was drastically diminished in astrocytes lacking Yy1. It is well established that APC is a critical regulator of the canonical Wnt signaling pathway [64, 65]. APC promotes the degradation of β‐catenin, a downstream effector of the Wnt pathway, effectively inhibiting the nuclear translocation of β‐catenin and thus preventing the activation of its target genes [65]. Activation of β‐catenin signaling in astrocytes initiates their activation [66]. Furthermore, aberrant accumulation of β‐catenin in APC‐deficient astrocytes promotes cerebellar neurodegeneration [64]. We assessed β‐catenin accumulation in GFAP‐positive cells of 5–7‐month‐old Yy1 ^ΔAST mice (Figure 8E,F) and EGFP‐positive cells of 7–8‐week‐old Yy1 ^{ΔAST;Aldh1L1‐EGFP} mice (Figure 8G,H). We found that Yy1 deletion induces accumulation of β‐catenin in astrocytes in the MOL, GCL, and WM (Figure 8E–H). These results suggest that Yy1 is needed to support astrocytic APC expression, subsequent β‐catenin degradation, and proper BG morphology.

FIGURE 8.

Increased β‐catenin and decreased CC1 expression by Bergmann glia correlates with diminished expression of synaptic proteins by glutamatergic neurons in Yy1 ^∆AST and Yy1 ^{ΔAST;Aldh1L1‐EGFP} mice. Yy1 ^∆AST (A, B, E, and F) and Yy1 ^{ΔAST;Aldh1L1‐EGFP} (C, D, and G–L) mice. (A) costaining for CC1 and GFAP in 5–7‐month‐old cerebella. (B) Quantification of CC1 in GFAP‐positive cells of the molecular layer (MOL). (C) Immunofluorescence costaining for CC1 and GFAP in 7–8‐week‐old cerebella. Cells expressing EGFP (green). (D) Quantification of CC1 in EGFP‐positive cells in the MOL. (E) Immunofluorescence costaining for β‐catenin and GFAP in 5–7‐month‐old cerebella. (F) Quantification of β‐catenin in GFAP‐positive cells in the MOL, granular cell layer (GCL), and white matter (WM). (G) Immunofluorescence costaining for β‐catenin and GFAP in 7–8‐week‐old cerebella. Cells expressing EGFP (green). (H) Quantification of β‐catenin in EGFP‐positive cells in the MOL, GCL, and WM. (I and K) Immunofluorescence costaining for VGLUT1 (I) or VGLUT2 (K) and Calbindin in 7–8‐week‐old cerebella of Yy1 ^{ΔAST;Aldh1L1‐EGFP} mice. Cells expressing EGFP (green). (J and L) Quantification of VGLUT1 (J) or VGLUT2 (L) in EGFP‐positive cells in the MOL and GCL. (A–L) Lobes IV–V. n = 3 mice in each group; two‐tailed t‐test; *p < 0.05; **p < 0.01; ***p < 0.0005; ****p < 0.0001.

To evaluate whether deletion of Yy1 from astrocytes affects synaptic contacts, we stained for vesicular glutamate transporter 1 (VGLUT1) that marks synapses between the Parallel Fibers of excitatory granule cells and PC dendrites. Indeed, the intensity of VGLUT1 was dramatically diminished in 7–8‐week‐old Yy1 ^{ΔAST;Aldh1L1‐EGFP} mice (Figure 8I,J). We also stained for VGLUT2 that marks synapses between PCs and the Climbing Fibers and also found significantly decreased numbers of VGLUT2 puncta in 7–8‐week‐old Yy1^{ΔAST;Aldh1L1‐EGFP} mice (Figure 8K,L). These data suggest a potential disruption in synaptic contacts between Parallel Fibers and Climbing Fibers with PC dendrites, which are typically enwrapped and supported by BG microdomains.

4. Discussion

Neurodegeneration is not regarded as a cell‐autonomous disorder of neurons since it affects multiple cell types, including astrocytes, microglia, oligodendrocytes, and other cells within the affected area of the neural tissue. While astrocytes have long been known to become reactive during neurodegeneration, their responses are very heterogeneous and either neuroprotective or neurotoxic. Furthermore, astrocytes may either become reactive in response to external stimuli or become intrinsically dysfunctional and drive neurodegeneration. Although astrocyte reactivity is associated with many devastating diseases, including Alzheimer's, Huntington's, and Parkinson's diseases, and amyotrophic lateral sclerosis, defining whether astrocyte activation is the cause or an effect of the ongoing disease is not straightforward for many neurodegenerative diseases. One of the most recognized examples of astrocyte‐driven neurodegeneration is Alexander disease caused by mutations in the GFAP gene [67, 68]. Astrocyte dysfunction also triggers neurodegeneration in a multiple sulfatase deficiency [69] and hereditary spastic paraplegia [70]. In addition, acquired autoimmune responses against astrocyte targets, such as GFAP [71] and AQP4 [72] result in encephalopathy, tremor, ataxia, and oligodendrocyte death, and demyelination, respectively. In this study, we found that improper maturation of astrocytes during their development in Yy1 ^ΔAST mice [51] leads to subsequent cerebellar neuroinflammation, which is followed by cerebellar neurodegeneration with the loss of synapses and PC degeneration. The effects of the presence of dysfunctional astrocytes were particularly evident in the cerebellar MOL, suggesting that dysfunctional BG are particularly strongly affected. Activation of BG, termed Bergmann gliosis, has been described in both humans [73, 74] and in animal models [75, 76, 77, 78, 79] and is associated with cerebellar ataxias and severe motor and behavioral outcomes. Indeed, adult Yy1 ^ΔAST mice display severe motor deficits [51]. Reactive Bergmann gliosis has recently been shown to play a central role in driving spinocerebellar ataxia inflammation via unique activation of the JNK pathway [80]. In addition, chronic activation of BG leads to the dysfunction of GLAST and PC death [81]. In this study, we found diminished expression of GLT1 throughout the cerebellum and drastically decreased expression of GLAST in the cerebellar MOL in adult Yy1 ^ΔAST mice. Importantly, diminished expression of both GLT1 and GLAST manifested already in 7–8‐week‐old Yy1 ^ΔAST mice, preceding PC loss and suggesting that accumulated extracellular glutamate may cause excitotoxicity and drive subsequent neurodegeneration. The unchanged levels of Aqp4 mRNA and protein suggest that not all astrocytic proteins are equally dependent on YY1 for their expression, pointing to a selective regulatory role. The upregulation of Gfap and Nestin mRNAs further indicates a reactive astrocyte phenotype. Such activation could exacerbate the dysregulation of glutamate transport and amplify neuronal stress, aligning with prior studies linking reactive astrocytes to neurodegenerative processes [82]. Furthermore, we found diminished expression of AMPA receptors and decreased numbers of VGLUT1 and VGLUT2 synapses between PCs and the parallel fibers of excitatory granule cells, as well as PCs and the climbing fibers. The AMPA receptors are pivotal for maintaining PC synapses and regulate the fine‐tuning of motor skills [83]. Importantly, the loss of synaptic VGLUT1 and VGLUT2 in Yy1 ^ΔAST mice suggests a breakdown in astrocyte‐mediated synaptic maintenance. Astrocytes, including BG, play a key role in enwrapping synapses and modulating their function. Reactive transformation impairs this supportive role, disrupting the balance of excitatory and inhibitory neurotransmission. These deficits are especially critical in the cerebellum, where BG microdomains are essential for the proper function of PCs and granule neurons [20, 84]. The decreased expression of both GLT1 and GLAST in 7–8‐week‐old Yy1 ^ΔAST mice correlated with the development of neuroinflammation characterized by increased proinflammatory cytokine expression and microglia activation. This neuroinflammation persisted in older mice and likely contributed to the loss of synapses and cerebellar neurodegeneration. It is also important to point out that the Yy1 gene was specifically and exclusively deleted in cerebellar astrocytes but not from other cell types [51] precluding the possibility that the observed neurodegeneration is driven by YY1 deletion in other cell types, including neurons or oligodendrocytes.

It has been previously reported that APC is pivotal for the acquisition of unique BG morphology and its loss induces nonautonomous neurodegeneration of PCs [64]. We found that APC expression was drastically diminished in astrocytes lacking Yy1. The loss of APC expression correlated with the accumulation of β‐catenin, which is also known to initiate astrocyte activation [66]. This parallels previous findings that aberrant accumulation of β‐catenin in APC‐deficient astrocytes promotes cerebellar neurodegeneration [64]. Moreover, the activation of β‐catenin signaling in the Yy1 ^ΔAST mice aligns with the broader literature on astrocytic reactivity and its deleterious effects. β‐catenin activation has been implicated in the pathological transformation of astrocytes, driving proinflammatory and neurodegenerative cascades. Sustained β‐catenin expression can further enhance the release of proinflammatory cytokines and reactive oxygen species, which compound neuronal damage [85]. Our data support the notion that Yy1 is needed to support astrocytic APC expression, subsequent β‐catenin degradation, and the acquisition of proper BG morphology and gene expression. However, other mechanisms cannot be excluded, including cross‐talk between β‐catenin signaling and YY1 via cis‐acting circular RNA [86] or other yet to be identified mechanisms.

Although we reported that YY1 is required for the maintenance of the mature astrocyte phenotype in the adult cerebellum [51], we now demonstrate that YY1 plays a critical role in sustaining cerebellar health, and Yy1 deficient astrocytes drive subsequent neuroinflammation and neurodegeneration. It remains to be established whether astrocytes are dysfunctional in the human Gabriele‐de Vries YY1 haploinsufficiency syndrome [87] and contribute to the observed symptoms such as intellectual disability and developmental delay. Overall, our data demonstrate that the presence of the cerebellar astrocytes that did not mature properly initially leads to the development of cerebellar neuroinflammation, its persistence, and subsequent cerebellar neurodegeneration. These findings show that YY1 is pivotal for sustaining cerebellar astrocyte functions and cerebellar health.

Author Contributions

K.M. and M.Z.‐K. planned and performed most experiments, with assistance from A.K.G., A.H., L.D., J.V., and S.K.S. T.K. conceived the study and contributed to the planning of the experiments. T.K. and M.Z.‐K. drafted the manuscript. All authors read and approved the final manuscript.

Ethics Statement

Mice were housed at Virginia Commonwealth University according to the guidelines of the Institutional Animal Care and Use Committee (IACUC). The mouse protocols were approved by the IACUC.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1. Effect of astrocytic Yy1 deletion on cerebellar lobes in 5‐7‐month‐old Yy1 ^ΔAST mice (A‐D) Calbindin IF staining. (A) The lobes of the Yy1 ^loxP/loxP cerebellum (indicated by the Roman numbers). (B) The lobes of the Yy1 ^ΔAST cerebellum (three different mice). (C, D) The four main cerebellar domains (denoted by different colors) in Yy1 ^loxP/loxP (C) and Yy1 ^ΔAST mice (D).

Figure S2. Differences in morphology of astrocytes in the cerebellar lobes IV/V of 5–7‐month‐old Yy1 ^ΔAST mice. (A) GFAP IF staining of 5–7‐month‐old cerebella. DAPI was used to visualize nuclei. MOL, molecular layer; GCL, granular cell layer; and WM, white matter. (B) Analysis of Bergmann glia morphology in the molecular layer of lobes IV/V. Mean diameter of branches, average volume, filament length, and filament number of dendrite segments were examined. (C) Assessment of velate astrocyte morphology in the granular cell layer of lobes IV/V. Cell area, cell volume, filament length, and filament number per dendritic segment were examined. (D) Analysis of fibrous astrocyte morphology in the white matter of lobes IV/V. Cell area, cell volume, filament length, and filament number were examined. (E) Sholl analysis for Bergmann glia, velate astrocytes, and fibrous astrocytes. (A–D) n = 3 mice for each group; two‐tailed t‐test; *p < 0 0.05, **p ⟨ 0.01, ns; not significant.

Figure S3. Differences in morphology of astrocytes in the cerebellar lobe IX of 5–7‐month‐old Yy1 ^ΔAST mice. (A) GFAP IF staining of 5–7‐month‐old cerebella. DAPI was used to visualize nuclei. MOL, molecular layer; GCL, granular cell layer; and WM, white matter. (B) Analysis of Bergmann glia morphology in the molecular layer of lobes IX. Mean diameter of branches, average volume, filament length, and filament number of dendrite segments were examined. (C) Assessment of velate astrocyte morphology in the granular cell layer of lobes IX. Cell area, cell volume, filament length, and filament number per dendritic segment were examined. (D) Analysis of fibrous astrocyte morphology in the white matter of lobes IX. Cell area, cell volume, filament length, and filament number were examined. (E) Sholl analysis for Bergmann glia, velate astrocytes, and fibrous astrocytes. (A–D) n = 3 mice for each group; two‐tailed t‐test; *p < 0.05, **p < 0.01, ***p < 0.0005, ns; not significant.

Figure S4. Differences in the morphology of Fañanas cells in 5–7‐month‐old Yy1 ^ΔAST mice. (A) GFAP IF staining of cerebellar lobes IV/V and IX. DAPI was used to visualize nuclei. MOL, molecular layer. (B, C) Morphological quantification of GFAP‐positive Fañanas cells (FC) in lobes IV/V (B) and lobe IX (C). Dendrite diameter, cell volume, filament length, and filament number per dendritic segment were examined. (D) Sholl analysis for FC in both lobes IV/V and lobe IX. (E) Filament length analysis of FC and BG in lobes IV/V and lobe IX. (F) and Sholl analysis of FC and BG in lobes IV/V and lobe IX.

Mockenhaupt K., Zarei‐Kheirabadi M., Gonsiewski A. K., et al., “Defective Astrocyte Maturation Drives Cerebellar Neuroinflammation and Degeneration,” The FASEB Journal (2025): e70824, 10.1096/fj.202501225RR.

Data Availability Statement

Information regarding the experimental methods used, and the data in this paper are available to scientific communities upon direct contact to the authors. Individual requests for shipment of mice to AAALAC accredited institutions will be honored. An appropriately signed MTA will be required, as well as permission from the original source of the Aldh1l1‐EGFP mice (Dr. Cagla Eroglu).