Phenotypical changes of satellite glial cells in a murine model of GM1‐gangliosidosis

Abstract Satellite glial cells (SGCs) of dorsal root ganglia (DRG) react in response to various injuries in the nervous system. This study investigates reactive changes within SGCs in a murine model for GM1‐gangliosidosis (GM1). DRG of homozygous β‐galactosidase‐knockout mice and homozygous C57BL/6 wild‐type mice were investigated performing immunostaining on formalin‐fixed, paraffin‐embedded tissue. A marked upregulation of glial fibrillary acidic protein (GFAP), the progenitor marker nestin and Ki67 within SGCs of diseased mice, starting after 4 months at the earliest GFAP, along with intracytoplasmic accumulation of ganglioside within neurons and deterioration of clinical signs was identified. Interestingly, nestin‐positive SGCs were detected after 8 months only. No changes regarding inwardly rectifying potassium channel 4.1, 2, 3‐cyclic nucleotide 3‐phosphodiesterase, Sox2, doublecortin, periaxin and caspase3 were observed in SGCs. Iba1 was only detected in close vicinity of SGCs indicating infiltrating or tissue‐resident macrophages. These results indicate that SGCs of DRG show phenotypical changes during the course of GM1, characterized by GFAP upregulation, proliferation and expression of a neural progenitor marker at a late time point. This points towards an important role of SGCs during neurodegenerative disorders and supports that SGCs represent a multipotent glial precursor cell line with high plasticity and functionality.

very similar to the role of astrocytes in the central nervous system (CNS). 5,6 Interestingly, SGCs also share certain molecular markers with astrocytes, namely glutamine synthetase (GS) 7 and the inwardly rectifying potassium channel (Kir) 4.1. 8 Since these two markers neither are expressed in neurons nor in Schwann cells (SCs), they can be used as SGC-specific markers for reliable identification of this cell population within murine DRG. 9 Another shared feature of astrocytes and SGCs is represented by their ability to react to various noxious stimuli. [10][11][12] Comparable to astrocytes within the CNS, SGCs are thought to proliferate and to convert into an activated state in response to injury, concomitant with an increased expression of glial fibrillary acidic protein (GFAP). 13,14 Therefore, GFAP is commonly used as a marker of murine SGC activation. [15][16][17] Over the last years, increasing evidence suggests that SGCs not only respond to pathological conditions within the peripheral nerves (PN), 18,19 but also act as key modulators in chronic pain conditions. [20][21][22][23] This was further investigated using animal models of traumatic PN injury, 20,23-25 diabetic neuropathic pain, 5,26,27 inflammatory pain as, for example in response to injection of the hindpaw with Complete Freund's Adjuvant 11,[28][29][30] as well as herpes simplex infection. 31,32 Moreover, it is suggested that SGCs might possess a multipotent character with the ability to differentiate into sensory neurons as a consequence of nerve injury. 23,33 These results indicate that SGCs are not only functionally involved but might also represent a potential source of regenerative capacity in various pathological conditions. 23,24 These special features make SGCs extremely interesting research objects.
The knowledge of their exact role and behaviour in the context of storage diseases is scarce. Therefore, the aim of the present study was to characterize and investigate potential phenotypical changes in the expression pattern of SGCs in a lysosomal storage disease mouse model of G M1 -gangliosidosis. G M1 -gangliosidosis in humans is associated with degenerative changes within CNS and PNS. [34][35][36] The clinical disease is divided into three types, according to the age of onset including an infantile type with early onset and rapid clinical deterioration, a late infantile/juvenile type with later onset and prolonged progression as well as an adult type. 37 It is caused by an accumulation of G M1 -ganglioside due to a deficiency of β-galactosidase (GLB1). [37][38][39] Consequently, G M1 -gangliosides and related glycoconjugates are deposited in several tissues but especially within neurons. 37 This leads to distension with subsequent death of neurons within both, CNS and PNS. 37,40,41 Studying murine models for G M1gangliosidosis, it was observed that Glb1-deficient mice, despite increasing accumulations of G M1 -ganglioside, did not show clinical abnormalities up to the age of 4-5 months. 35,42 The present study investigates the phenotypical changes of SGCs within the DRG of C57BL/6 wild-type (Glb1 +/+ ; WT) and homozygous β-galactosidase-knockout (Glb1 −/− ) mice. Because of the close interrelationship between neuronal somata and SGCs as well as the intense communication between both cell types, it was hypothesized that SGCs of affected mice will show phenotypical changes during the course of disease. A better understanding of the reaction pattern and the potential involvement of SGCs during the course of G M1 -gangliosidosis will increase our knowledge about the nature and the potential of this unique cell population.

| Animals and clinical investigation
Dorsal root ganglia of homozygous Glb1 knockout and WT mice were obtained from previous experiments. 40 In brief, Glb1 −/− mice were generated via insertion of a lacZ gene fragment of 636 base pairs into exon 15 of the Glb1 gene of murine C57BL/6 oocytes.
Genotyping of mice was achieved using conventional polymerase chain reaction and gel electrophoresis. 40 Both Glb1 −/− and WT mice were bred and housed in parallel, as described previously. 40 Mice were examined regularly assessing clinical parameters like appearance/posture, behaviour/activity and gait. Furthermore, animals were screened for neurological deficiencies using the parachute reflex test and grid-walking test (horizontal wire netting). 40 At 2, 4, 6 and 8 months of age, 6 WT and Glb1 −/− mice were euthanized, and DRG were removed at the height of the cervical vertebral column.
Tissue was routinely fixed in 10% formalin and embedded in paraffin wax.

| Tissue processing and evaluation
Formalin-fixed and paraffin-embedded tissue samples were used for immunofluorescence (IF) analysis. All tissue samples were cut into approximately 4 µm thick sections on a microtome and subsequently mounted on SuperFrost-Plus ® slides (Thermo Fisher Scientific Inc., Fisher Scientific GmbH). IF staining was performed as previously described. 9 Briefly, deparaffinization and rehydration were performed following standard procedures using xylene and graded alcohols. Sections were blocked in 20% goat serum in phosphate buffered saline (PBS) containing 1% bovine serum albumin (BSA) and 0.1% Triton-X (Triton ® X-100, Merck millipore, Merck KGaA) after antigen retrieval in citrate buffer (pH = 6) for 20 min in a microwave. Primary antibodies (for details see Table 1  For transmission electron microscopy, selected DRG were fixed in 5% glutaraldehyde in cacodylate buffer and further processed as described previously. 40 Accumulated storage material within neurons was visualized using a transmission electron microscope (Zeiss EM 10C electron microscope; Zeiss).

| Statistical analysis
Statistical analysis was performed using SPSS for Windows (version 27; IBM ® SPSS ® Statistics, SPSS Inc.). Differences between groups at different time points were analysed via Mann-Whitney U tests.
Statistical significance was accepted at a p-value of <0.05. For statistical evaluation of immunostainings, initially only 2-and 8-month-old animals were investigated for significant alterations between affected and non-affected animals (CNPase, doublecortin, GFAP, Ki67, nestin, periaxin, Sox2) and/or in between time points (Kir 4.1).
Where significant changes were detected, 4-and 6-month-old animals were evaluated accordingly to further investigate the chronological sequence of changes. For evaluation of clinical data matching the investigated tissue specimens, the number of mice included in the statistical analysis was n = 18 (2 months of age), n = 17 (4 months of age), n = 12 (6 months of age) and n = 6 (8 months of age) for Glb1 −/− mice, and n = 16 (2 months of age), n = 12 (4 months of age), n = 12 (6 months of age) and n = 5 (8 months of age) for WT mice.

| Ethics
All animal experiments were conducted in accordance with the German Animal Welfare Law and were approved by local authori-

| Kir 4.1 expression in Glb1 −/− mice remains constant during G M1 -gangliosidosis
Kir 4.1 is a suitable marker for the detection of murine SGCs and stains a high percentage of these cells per DRG. 9 Kir 4.1 is also reported to be downregulated in various pathological conditions. 43,44 However, within the present study, no significant changes of Kir 4.1 expression could be detected in Glb1 −/− mice between 2 and 8 months of age ( Figure 4). These data indicate that SGCs increasingly upregulate GFAP.

| Satellite glial cells show proliferation but no increased apoptosis in Glb1 −/− mice
It has also been demonstrated that SGCs show proliferation after injury. 23,47,48 In order to further characterize the response of mu-

| Satellite glial cells show expression of nestin in Glb1 −/− mice at the end of the investigation period
Nestin is an intermediate filament that is commonly expressed in neural progenitor cells and gets downregulated during cellular differentiation. 49 Interestingly, it has been reported that SGCs show nestin expression during embryonic stages and that nestin expression might be re-activated following injury. 23,50 Within the present study, the occurrence of nestin-positive SGCs was detected in Glb1 −/− mice at the end of the investigation period (Figure 7; Figure S3). Nestinpositive SGCs surround sensory neurons, which show a positive immunoreaction for GM1 ( Figure S3).

| Murine satellite glial cells show no expression of Sox2 and doublecortin
In search for the potential upregulation and/or expression of further progenitor markers, DRG were investigated using anti-Sox2, anti-NG2 and anti-doublecortin antibodies. No immunoreactivity for either Sox2 or doublecortin ( Figure S4) was found in SGCs of Glb1 −/− and WT mice. Moreover, DRG neurons did not show any immunoreactivity for doublecortin. As experienced previously, 9 evaluation of NG2 staining on FFPE tissue did not reveal a distinct and evaluable staining pattern.

| Murine satellite glial cells show no expression of markers for myelination
To further determine a potential re-myelinating capacity of SGCs, DRG were investigated using anti-CNPase and anti-periaxin antibodies. These two markers represent myelin markers of the CNS (CNPase) and PNS (periaxin) respectively. SGCs of neither affected nor unaffected animals showed immunopositivity for CNPase and/ or periaxin.

| DISCUSS ION
G M1 -gangliosidosis is a lysosomal storage disease caused by a predominantly intraneuronal accumulation of G M1 -gangliosides, leading to a progressive neurodegenerative disorder. 38 Although CNS disease represents the most prominent feature of most lysosomal storage

| Satellite glial cells upregulate GFAP concomitant to progression of G M1 -gangliosidosis
Glial fibrillary acidic protein is an intermediate filament that is mainly expressed in astrocytes and is upregulated following brain injury with increasing GFAP levels frequently being used as a marker for CNS astrogliosis. 52,53 It is believed that the increased expression of GFAP in astrocytes is involved in the formation of elongated and thickened processes and represents a prominent feature following traumatic and degenerative events within the CNS. 54 A reactive astrogliosis of the CNS is also present in murine and canine models of G M1 -gangliosidosis as well as during the course of the human disease, especially within areas exhibiting neuronal vacuolation, neuronal death and demyelination. 37,55,56 Although separated within distinct compartments of the ner-

| Murine satellite glial cells exhibit an increased proliferation rate at the age of 6 months
As stated before, SGCs tightly envelop neurons and provide support. They perform similar functions in the periphery as astrocytes in the CNS. Neuronal injury within the CNS is frequently associated with the development of a reactive astrogliosis, which is not only characterized by upregulation of GFAP but also by proliferation of astrocytes. 54 In line with this, previous studies have shown that activation of SGCs also includes both upregulation of GFAP and proliferation. 11,19,20,31,45,46,60 In general, SGCs retain the ability of cell Proliferation might therefore represent an attempt of SGCs to maintain homeostasis and structural integrity within the DRG as well as to secure neuronal survival, comparable to glial cell proliferation within the injured CNS. 62 Interestingly, previous studies indicate that GFAP expression not only promotes proliferation within CNS astrocytes 52 but also relates to proliferation of SCs of the PNS and subsequent regeneration. 60 This may point to a possible link between GFAP expression in SGCs and the following increased proliferation rate of SGCs.

| Glb1 −/− mice upregulate nestin in satellite glial cells at the end of the investigation period
The intermediate filament nestin has been widely accepted as a marker for multipotent stem cells and progenitor cells in various tissue, for example muscle, 63 hair follicle sheath, 64 pancreas 65 and teeth. 66 Importantly, nestin is also a marker for neuronal and glial cells, together with their shared progenitor cells. 67 However, nestin expression is temporary in most of these cells and downregulated during cell differentiation. 68 In adults, the expression of nestin is mainly restricted to stem cell niches like the subventricular zone as well as the hippocampus in the CNS. 69 However, an increased number of nestin-positive cells can be seen in response to several pathologic conditions affecting the CNS, as for example inflammation, ischaemia and epilepsy. 67 Several studies suggest that nestin is re-expressed and upregulated in activated astrocytes, 70,71 as well as in SGCs following nerve injury. 72,73 The present study shows that adult murine SGCs in Glb1

| Murine satellite glial cells do not upregulate expression of precursor markers Sox2 and doublecortin
In this study, murine SGCs of neither Glb1 −/− nor WT mice showed any immunoreaction for Sox2 or doublecortin. Sox2 is a transcription factor expressed by neural/glial precursors, 51 and doublecortin is expressed by neuronal precursors. 52 The lack of Sox2 expression in adult murine SGCs correlates with previous studies. 9 Doublecortin as a neuronal precursor marker in adult neurogenesis 75 can also be expressed in sensory neurons of adult murine DRG. 76 However, in this study, no immunoreaction for doublecortin was observed in sensory neurons of DRG ( Figure S4).
Furthermore, NG2, a well-established marker for oligodendrocyte precursor cells which is expressed in a subgroup of SGCs in C57BL/6 wild-type mice, 9 was tested. As experienced previously, 9 NG2 is very sensitive towards formalin fixation with lack of a distinct and evaluable staining pattern on FFPE tissue. Therefore, this marker was determined as non-suitable for evaluation of the available tissue.

| Murine satellite glial cells do not express markers of myelination
Myelination of nerve fibres is crucial for nerve conduction.
Oligodendrocytes and SCs within the CNS and PNS, respectively, are specialized glial cells that are in charge of myelin production. 77 CNPase is a myelin-associated enzyme mainly found in oligodendrocytes that is mandatory for physiological function of the axonmyelin unit. 78,79 Within the PNS, periaxin represents a SC-specific protein of non-compact myelin sheaths. 80 Previous studies have shown that canine SGCs of healthy, adult dogs express CNPase; however, this does not apply to murine SGCs. 9,51 Interestingly, CNPase expression is reported to increase in rat SGCs following injury. 86 Moreover, it was found that embryonic rat SGCs are able to divide and differentiate into other glial cell populations like oligodendrocytes, SCs and astrocytes in vitro. 87 Another study reported that rat SGCs resemble SC precursors with the ability to myelinate embryonic axons in co-cultures. 88 Regarding murine SGCs, overexpression of Sox10 during embryogenesis was shown to transform SGCs towards an oligodendrocyte-like phenotype, supporting the theory of SGCs being multipotent glial precursor cells with high plasticity. 89 Within the present study, neither

| Clinical deterioration in Glb1 −/− mice correlates with first signs of activation in dorsal root ganglia
Increased clinical scores, indicating a neurological impairment of Glb1 −/− mice, were noticed at 4 months of age (Figures 1 and 2).
An increased expression of GFAP within SGCs was observed at the same time point ( Figure 5). Therefore, it can be assumed that clinical deterioration in affected mice is accompanied by a reactive change in SGCs. Significant clinical differences between healthy WT and diseased Glb1 −/− mice were detected as early as 6 months of age in the grid-walking and in the parachute test, accompanied by increased proliferation (6 months) and upregulation of nestin (8 months) in SGCs. In

CO N FLI C T O F I NTE R E S T S
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data were available on request from the authors.