GM2 ganglioside accumulation causes neuroinflammation and behavioral alterations in a mouse model of early onset Tay-Sachs disease

Background Tay-Sachs disease (TSD), a type of GM2-gangliosidosis, is a progressive neurodegenerative lysosomal storage disorder caused by mutations in the α subunit of the lysosomal β-hexosaminidase enzyme. This disease is characterized by excessive accumulation of GM2 ganglioside, predominantly in the central nervous system. Although Tay-Sachs patients appear normal at birth, the progressive accumulation of undegraded GM2 gangliosides in neurons leads to death. Recently, an early onset Tay-Sachs disease mouse model, with genotype Hexa−/−Neu3−/−, was generated. Progressive accumulation of GM2 led to premature death of the double KO mice. Importantly, this double-deficient mouse model displays typical features of Tay-Sachs patients, such as cytoplasmic vacuolization of nerve cells, deterioration of Purkinje cells, neuronal death, deceleration in movement, ataxia, and tremors. GM2-gangliosidosis is characterized by acute neurodegeneration preceded by activated microglia expansion, macrophage, and astrocyte activation, along with the production of inflammatory mediators. However, the mechanism of disease progression in Hexa−/−Neu3−/− mice, relevant to neuroinflammation is poorly understood. Method In this study, we investigated the onset and progression of neuroinflammatory changes in the cortex, cerebellum, and retina of Hexa−/−Neu3−/− mice and control littermates by using a combination of molecular genetics and immunochemical procedures. Results We found elevated levels of pro-inflammatory cytokine and chemokine transcripts, such as Ccl2, Ccl3, Ccl4, and Cxcl10 and also extensive microglial and astrocyte activation and proliferation, accompanied by peripheral blood mononuclear cell infiltration in the vicinity of neurons and oligodendrocytes. Behavioral tests demonstrated a high level of anxiety, and age-dependent loss in both spatial learning and fear memory in Hexa−/−Neu3−/− mice compared with that in the controls. Conclusion Altogether, our data suggest that Hexa−/−Neu3−/− mice display a phenotype similar to Tay-Sachs patients suffering from chronic neuroinflammation triggered by GM2 accumulation. Furthermore, our work contributes to better understanding of the neuropathology in a mouse model of early onset Tay-Sachs disease.

instance, the levels of TNF-α pro-inflammatory cytokine were significantly increased in the cerebrospinal fluid of TSD patients [9]. In addition, five inflammation biomarkers, ENA-78, MCP-1, MIP-1α, MIP-1β, and TNFR2 were also detected in the cerebrospinal fluid of patients with infantile and juvenile gangliosidosis [10]. Various lysosomal storage disorders including Niemann-Pick type C disease [11], Gaucher disease [12], mucopolysaccharidosis type I, IIIA and III [13] and neuronal ceroid-lipofuscinoses [14] also exhibit neuropathological alterations such as prominent microglial and astrocyte activation. A previous study showed that Hexa-/-Neu3-/-mice exhibited an inflammatory response, with astrogliosis, in the hippocampus, cortex, and cerebellum [5]. In this work, we aimed to further investigate the neuroinflammatory response to GM2 accumulation in the brain and retina of Hexa-/-Neu3-/-mice, and its effect on behavior during the progression of disease.
Food and water were available ad libitum. The protocols used to perform the experiments described in this work were approved by the Institutional Animal Care and Use Committee of the Izmir Institute of Technology.

RNA array for cytokine-and chemokine-encoding genes
For a broad analysis of inflammatory gene expression in the cortex and cerebellum of Hexa-/-and Hexa-/-Neu3-/-mice, the mRNA profile of cellular mediators of inflammation, specifically cytokine and 5 chemokines, was assessed using RT² Profiler PCR Array (Qiagen, The Netherlands), according to manufacturer's instructions. Total RNA was transcribed using the RT 2 First Strand Kit (Qiagen, The Netherlands). Real-Time SYBR green PCR master mix (Qiagen, The Netherlands) was used and PCR cycles were programmed in accordance with the manual. Briefly, (10 min at 95 °C, followed by 40 cycles of 15 s at 95 °C for and 1 min at 60 °C) followed by routine melting curve analysis. All data were normalized to an average of five housekeeping genes, i.e. Gusb (beta-glucuronidase), Hprt (Hypoxanthine-guanine phosphoribosyltransferase), Hsp90ab1 (Heat Shock Protein 90 Alpha Family Class B Member 1), GADPH (D-Glyceraldehyde-3-Phosphate Dehydrogenase) and Actb (beta-actin).
Each sample had a single peak in each reaction at a temperature greater than 80°C which showed that there is no non-specific product. The threshold cycle changes (△Ct) denote the difference in Ct for the gene of interest based on the Ct level of housekeeping genes. The expression of each gene was normalized using the mean expression of five different housekeeping genes in the array. Fold changes in the expression level of genes were obtained according to the 2 −ΔΔCT method. Genes that had positive or negative fold changes of more than 1.5 fold with a p-value smaller than 0.05, were considered to be statistically significant.

Retina analysis
Eyes were removed and fixed for 24 hours with 4% PFA, at room temperature. Retinas were prepared as flattened whole-mounts, by making four radial cuts, and labeled with anti-lectin (20μg/mL) (Vector Laboratories, USA) and anti-phalloidin (1:100) (Biotium, USA) antibodies. Retinas were permeabilized by washing in PBST for 30 min, then, washed them in PBS 3 times for 10 min and, finally, anti-lectin was applied at 4 o C overnight. The following day, they were washed in PBS 3 times for 5 min and incubated in anti-phalloidin at 4 o C overnight. Retinas were mounted with an anti-fade reagent containing Hoechst-33258 to label nuclei (Molecular Probes, USA). Twenty µm-thick sagittal sections were assessed for potential anatomical alterations. Finally, whole-mount retinas and sagittal sections were imaged using a light microscope (400x, Eclipse E600, Nikon Instruments Inc., Japan) equipped with a manually controlled specimen stage for X, Y, and Z-axis; a color camera (model DXM1200, Nikon Instruments Inc., Japan); a fluorescent light source (HB-10104AF, Nikon Instruments Inc., Japan); and image analysis software (NIS-Elements, Version 3.22, Nikon Instruments Inc., Japan).

Passive avoidance test
This test was conducted as previously described [16]. Briefly, the apparatus, which is divided has 7 bothinto a lit compartment and a dark compartment, with a vertical sliding door between the two.
Animals were allowed to explore both compartments on the first day. Each mouse was placed into the bright compartment, and after 30 s the central door opened allowing it to migrate into the dark compartment, an environment which they generally prefer. The following day, upon entry into the dark compartment, the door was closed and the mouse received a mild electric shock (0.2 mA for 2 s).
On the final day, animals were given the same choice to enter the dark compartment (300 s maximum time). Latency to cross through the door between compartments was scored with ShutAvoid v1.8 (Harvard Apparatus, USA).

Morris water maze test
Spatial learning and memory were assessed as previously described [16]. The water maze consists of a large circular pool (diameter 140 cm; depth, 45 cm) filled with an opaque solution of dry milk in water (22 0 C), that is located in a room with numerous visual cues. Spatial acquisition was organized in 3 training sessions (days 1-3) with a visible platform (1 cm above the platform). Each mouse has 60 s to reach the platform. The location of cues and platform were changed every day, and the mice were released into the water at 1 of 3 different locations. On days 4-8 animals received reversal training with an invisible platform (1 cm below the water surface). For this training, the platform location was kept constant, and mice were placed into the water at 1 of 3 different locations and given 90 s to find the platform. Measurements were acquired with a Sony camera (model SSC-G18), centrally positioned above the water tank. Behavioral differences were analyzed using the Panlab SMART Video Tracking System v0.3 (Harvard Apparatus, USA). Animals were allowed to dry under a heat lamp after each trial to avoid hypothermia, and all experiments were started at the same time each day.

Grip strength
Mice were subjected to forelimb grip strength test, with a Grip Strength Meter (IITC Life Science, USA).
The gauge was reset to 0 g after stabilization, and each mouse was encouraged to grab a T-shaped bar. Connection of the bar to a digital force transducer allowed quantification of strength as pull force in grams. The order of mice tested was randomized, and the experimenter was blinded to the 8 genotypes of mice.

Open field test
The apparatus for this assay consisted of a 40X40 cm surface area and was surrounded from all sides by a 40 cm transparent wall. A digital camera was mounted directly above the apparatus. Mice were placed in one of the corners of the open field and allowed to explore undisturbed for 5 min. Behavioral differences were analyzed using the Panlab SMART Video Tracking System v0.3 (Harvard Apparatus, USA).

Statistical analysis
GraphPad QuickCalcs (GraphPad Software, USA) software was used for statistical analysis. All values are expressed as the mean ± S.E.M. Differences were tested using one-way-ANOVA for behavioral analysis and immunofluorescence analysis. qPCR data were analyzed by unpaired t-test. A p-value of less than 0.05 was considered to represent statistical significance.

Altered levels of the inflammatory cytokines and chemokines
Neuroinflammation involves the release of both inflammatory cytokines and chemokines [17].
Therefore, we first evaluated the transcript levels of pro-and anti-inflammatory cytokines and chemokines in the cortex and cerebellum of 4.5-month-old Hexa-/-and Hexa-/-Neu3-/-mice. We used a preformatted gene pathway array to compare the expression of 84 inflammatory chemokine, cytokine and interleukin receptor genes relative to that in Hexa-/-mice. After disregarding unregulated, non-detectable gene products and evaluating the relative levels of gene expression across all samples, we identified 42 inflammatory genes, which we categorized into 3 distinct groups.
Pro-inflammatory cytokines and chemokines, secreted mostly by activated macrophages in microglial cells, are involved in the up-regulation of inflammatory reactions [18]. Our data showed that Hexa-/-Neu3-/-mice displayed an altered expression profile for the pro-and anti-inflammatory cytokines, as well as chemokines in both of the cortical ( predominantly expressed in the cerebellum. It has been established that Ccl2 cytokine displays chemotactic activity for monocytes, lymphocytes, and neutrophils, while Ccl3 is involved in the migration of monocytes, lymphocytes, and neutrophils together with Ccl2 [18] and activates T cells and macrophages with Ccl4 [19]. Increased expression of Ccl3 and Ccl4 in brain and cerebellum were also observed in different lysosomal storage diseases such as MPSIIIA [20] and MPSIIIB [21], and Gaucher Disease [22,23]. In addition, Cxcl10, an inflammatory chemokine produced by astrocytes which recruits activated T lymphocytes by increasing their migration to the site of tissue damage, is increased in both cortex and cerebellum of Hexa-/-Neu3-/-mice [24,25]. Anti-inflammatory cytokines and chemokines regulate the expression of pro-inflammatory cytokines to arrest or slow down the immune response [26]. We found not only an up-regulation of pro- was increased compared with that in Hexa-/-mice. Csf1, a cytokine that controls the production, differentiation, and function of macrophages [29], was increased in the cortex (1.6-fold) and cerebellum (1.3-fold) of Hexa-/-Neu3-/-mice, which may account for the higher number of macrophages seen in these regions (Fig. 2H, P).

Microglial activation
Microglial cells are the immune cells of the CNS responsible for sensing stress signals released by damaged or dying neurons [19]. They migrate to sites of injury where they release cytokines to promote the removal of dead and dying cells by phagocytosis [30]. Once stimulated, they can act to activate neighboring microglia, astrocytes, neurons, or oligodendrocytes. Several studies have shown that the accumulation of undegraded macromolecules in lysosomal storage disorders, including gangliosidosis, result in activation of a neuroinflammatory response in the CNS [17,19]. To visualize the location of active microglia, brain sections from WT, Hexa-/-, Neu3-/-, and Hexa-/-Neu3-/-mice were immunostained with anti-Moma2 and anti-Lamp1 antibodies, which are specific to the activated microglial/macrophage system [31]. The number of Moma2-positive cells was significantly increased in the hippocampus, cortex, thalamus, and pons of 4.5-month-old Hexa-/-Neu3-/-compared with that in WT, Hexa-/-and Neu3-/-mice at ( Fig. 2 and Supp. Fig. 1). We found significantly increased microgliosis in the brains of 4.5-month-old Hexa-/-Neu3-/-mice compared with that in age-matched Hexa-/-mice, with the following ratios: approximately 60-fold in hippocampus (Fig. 2V), 9-fold in cortex (Fig. 2W), 13-fold in thalamus (Fig. 2X), and 19-fold in pons (Fig. 2Z). In the cerebellum, microgliosis was observed only in Hexa-/-Neu3-/-mice (Fig. 2P, Y). Significantly increased microgliosis was also observed in 2.5-month-old Hexa-/-Neu3-/-mice compared with that in age-matched Hexa-/mice, with the following ratios: approximately 37-fold in the hippocampus (Supp. Fig. 1V), 4-fold in the cortex (Supp. Fig. 1W) and 5-fold in the thalamus (Supp. Fig. 1X). Only Hexa-/-Neu3-/-mice displayed microgliosis in the cerebellum (Supp. Fig. 1P, Y). There was no microglial activation in the hippocampus, cortex, thalamus and cerebellum of WT and Neu3-/-mice, with the exception of in the pons area. The presence of neuroinflammatory conditions such as astrogliosis in the hippocampus, cortex and cerebellum of 4.5-month-old Hexa-/-Neu3-/-mice has been previously reported [5]. In this study we showed that activated astrocytes are present in the hippocampus, cortex, and cerebellum of 11 not only 4.5-month-old but also 2.5-month-old Hexa-/-Neu3-/-mice (Supp Fig. 2).

Increased infiltration of peripheral blood mononuclear cells (PBMCs) in the brain
Neuroinflammation contributes to disease progression in several brain disorders. Specifically, activated immune cells induce the infiltration of T and B cells into the CNS, promoting the release of chemokines and stimulating the migration of more immune cells to the CNS via a disrupted bloodbrain-barrier (BBB) [35]. Previous studies showed that neuroinflammation leads to significant infiltration of PBMC in the brain of the Sandhoff mouse model [36].

Decreased number of oligodendrocyte and neurons
To characterize the degree to which neuroinflammation affected neuronal cells and oligodendrocytes, 10 µm coronal brain sections from 2.5 and 4.5-month-old WT, Hexa-/-, Neu3-/-and Hexa-/-Neu3-/mice were immunostained with NeuN and CNPase antibodies, which recognize nonspecific neurons and oligodendrocytes, respectively. As shown in Fig. 4, 4.5-month-old Hexa-/-Neu3-/-mice showed a loss in neuronal density in different regions of CNS compared to age-matched control groups. The analysis demonstrated that the number of NeuN-positive neurons was significantly reduced by approximately 50% in cortex, 40% in thalamus, and 50% in pons of Hexa-/-Neu3-/-mice (Fig. 4D, H, P, respectively) compared with that in WT mice (Fig. 4A, E, M, respectively). There was no significant change in neuronal density, in the granular layer of the cerebellum (Fig. 4T). Furthermore, there was no significant change in the number of neurons in 2.5-month-old Hexa-/-Neu3-/-compared with that in age-matched control groups in different regions of the CNS (Supp. Fig. 4). Loss of NeuN immunoreactivity could be explained by increased neuronal death in damaged areas of the brain [37].
Thus, a decrease in NeuN immunostaining correlates with increased numbers of apoptotic cells. In a previous study, a TUNEL assay showed that there was apoptotic cell death in the brain of Hexa-/-Neu3-/-mice compared with that in WT, Hexa-/-and Neu3-/-mice [5]. Furthermore, NeuN-positive neurons in the cortex (Fig. 4D), thalamus (Fig. 4H), and hippocampus (Fig. 4P) appeared swollen, with accumulated storage material as previously shown in the Sandhoff mice brain [38].
Oligodendrocytes are types of glial cells that function in the formation of myelin, which then supplies support and insulation to axons [30]. Oligodendrocytes provide axons with myelin sheaths, and have the ability to renew their myelin sheaths three times within a day. Myelination is essential for optimal signal transduction in the CNS. Activated microglia produces various pro-inflammatory mediators, chemokines, and cytokines. In such environments, oligodendrocytes are particularly susceptible to microglia-derived factors, resulting in the production of poor-quality myelin sheaths and oligodendrocyte death [39]. As shown in Fig. 5, Hexa-/-Neu3-/-mice showed lower numbers of 2',3'cyclic-nucleotide 3'-phosphodiesterase-positive cells compared with age-matched WT, Hexa-/-and Neu3-/-mice. The number of oligodendrocytes was reduced by approximately 45% in the cortex 55% in thalamus and cerebellum and 35% in pons area of Hexa-/-Neu3-/-mice (Fig. 5D, H, L, P, respectively) compared with that in WT mice (Fig. 5A, E, I, M, respectively). The number of oligodendrocytes was significantly decreased in the cortex of Hexa-/-mice, by approximately 30% compared with that in WT mice. We observed no significant changes in the number of oligodendrocytes in the thalamus, cerebellum, and pons of Hexa-/- (Fig. 5F, J, N, respectively) and Neu3-/- (Fig. 5G, K, O, respectively) mice, compared with that in WT mice. These results are consistent with previous studies. Gene expression profile studies in the cerebral cortex of normal and GM2 gangliosidosis (Tay-Sachs and Sandhoff) patients revealed that the myelin basic protein gene, expressed by oligodendrocytes, was also significantly depressed [8]. Sialiated gangliosides, especially GD1a and GT1b, are present on the axonal membrane and interact with the Myelin-associated glycoprotein (MAG) on the periaxonal surface to promote myelin sheath stability [40]. GM2 is also a sialic acid containing ganglioside; thus, accumulation of GM2 could lead to instability of myelin sheath. We found no significant changes in the number of oligodendrocytes in 2.5-month-old Hexa-/-Neu3-/-mice compared with that in age-matched control groups (Supp. Fig. 3).

Increased number of microglia in the retina
Neuroinflammation or neurodegeneration leads to reactive gliosis in the retina through hypertrophy, causing thickening and enlargement of processes via Müller cells and astrocytes [42]. To characterize the neuroinflammatory conditions affecting the retina, coronal eye sections from 4.5-month-old WT and Hexa-/-Neu3-/-mice were immunostained with anti-lectin (vessels and glia) and anti-phalloidin antibodies (vessels and actin filaments). We found significantly higher numbers of microglial cells in the retinas of Hexa-/-Neu3-/-, mice compared with that in WT mice (Fig. 6A, D). Microglial staining was prominent around the vessels. However, sagittal sections of Hexa-/-Neu3-/-mice showed no anatomical alterations in their retinal layers (Fig. 6G).

Impairments in spatial learning and memory
Up-regulation of cytokines and their receptors within the CNS during inflammation, and concomitant effects on brain function, have been reported [6]. The Morris water maze task was used to detect spatial learning and memory deficits [43][44][45][46]. We used a two-way ANOVA to analyze the data. Here we report that both 2.5-and 4.5-month-old Hexa-/-Neu3-/-mice displayed deficits in spatial learning and memory. Also, we found that in the first 3 days of training WT (p<0.001), Hexa-/-(p<0.025) and Neu3-/-(p<0.05) mice learned to use the visual clues to quickly reach the visible platform, whereas Hexa-/-Neu3-/-mice took a longer time to swim toward it (Fig. 7A, C). All groups initially had difficulty in finding the exact location of the hidden platform on day 4. While WT, Hexa-/-and Neu3-/-mice groups quickly improved their ability to find the platform, both (p<0.01) 2.5-and (p<0.025) 4.5month-old Hexa-/-Neu3-/-mice were not able to learn the location of the hidden platform compared to WT. Also, 2.5-month-old WT (p<0.025), Hexa-/-(p<0.05) and Neu3-/-(p=ns) mice and 4.5-month-old WT (p<0.025) and Neu3-/-(p=ns) mice spent continuously less time, from day one to five, to find the hidden platform (Fig. 7B, D). Our data showed that WT, Hexa-/-and Neu3-/-mice used distal clues more efficiently than Hexa-/-Neu3-/-mice.
Typical swim patterns revealed that both 2.5-and 4.5-month-old Hexa-/-Neu3-/-mice were unable to find the escape platform by the final day of the test, whereas age-matched WT, Hexa-/-and Neu3-/mice were successful at using distal clues to find it. In addition, 4.5-month-old Hexa-/-Neu3-/-mice showed more anxiety-related behaviors compared with that in WT and single deficient Hexa-/-and Neu3-/-mice (Fig. 7I, J).

Deficits in cognitive learning and memory
The passive avoidance test was used to detect cognitive learning and memory deficits [44]. We found that 2.5-month-old WT, Hexa-/-, Neu3-/-and Hexa-/-Neu3-/-mice re-entered the dark compartment at the average times: 219 ± 41.3 s, 203 ± 31.8 s, 132 ± 21.8 s, and 118 ± 60 s, respectively, after having previously received the electric shock through the feet. These results may indicate that the younger age group could not tolerate light, and prefer the dark compartment in spite of the 0.2 mA electric shock, for 2 s (Fig. 8A). Nearly all the 4.5-month-old WT mice avoided entrance into the dark compartment during the 300 s observation period, whereas the Hexa-/-Neu3-/-mice re-entered the chamber at an average time of 160 s (292.1 ± 7.9 s in WT and 159 ± 33.5 s in Hexa-/-Neu3-/-; p<0.001). In addition, Hexa-/-and Neu3-/-mice re-entered the chamber in slightly less time than WT mice (Fig. 8B). These results may show that Hexa-/-Neu3-/-mice display a significant deterioration in memory function compared with that in other groups.

Discussion
Neuroinflammation is a common hallmark in the pathogenesis of several lysosomal storage disorders including GM2 gangliosidosis, and may negatively impact neuronal survival [17,19]. In a previous study, we showed that severe neuronal apoptosis and astrogliosis in the hippocampus, cortex, and thalamus of Hexa-/-Neu3-/-mice correlated with undegraded GM2 ganglioside accumulation [5].
However, the mechanism of disease progression relevant to neuroinflammation was not yet elucidated. In the present work, we focused on investigating the age-dependent mediators of neuroinflammation in Hexa-/-Neu3-/-mice for future therapeutic studies.
It has been shown that the accumulation of undegraded gangliosides is linked to activation of microglial cells [30,48]. Here, we demostrated that an accumulation of GM2 ganglioside led to the activation of the microglial/macrophage system in the brain and retina of Hexa-/-Neu3-/-mice, compared with that in age-matched WT, Hexa-/-and Neu3-/-mice. Activation of this cellular system led to an altered expression profile of both pro-and anti-inflammatory cytokines, as well as chemokines, in the cortex and cerebellum. We found that while pro-inflammatory cytokines and chemokines, such as Ccl2, Ccl3, Ccl4, and Cxcl10 were significantly up-regulated in the brain of Hexa-/-Neu3-/-compared with that in Hexa-/-mice, anti-inflammatory cytokines and chemokines such as IL10, IL13, IL11, and IL24 were significantly down-regulated (Fig. 1). It has been shown that simultaneous increase of Ccl2, Ccl3, and Ccl5 in mice brain leads to epileptic seizures [49].
Importantly, their elevated levels may account for the neuroinflammation and the seizure activity observed in the neuropathology of the early-onset Tay-Sachs disease mouse model. Furthermore, these cytokines play a role in the attraction, migration, and activation of monocytes, lymphocytes, and neutrophils in the CNS [18,24,25]. Their increased levels in the CNS of Hexa-/-Neu3-/-mice triggered PBMC infiltration similar to the Sandhoff mouse model. This may suggest that inhibition of PBMC infiltration may delay disease progression and neurodegeneration in Hexa-/-Neu3-/-mice, similar to the Hexb-/-Ccl2-/-mice [36]. Moreover, it was reported that the deletion of the Ccl3 gene caused a delay of neuronal loss in Sandhoff mice (Hexb-/-Mip1α-/-) mice) by inhibiting macrophage infiltration to the inflammatory sites, which resulted in an improved neurologic status and a longer lifespan [37]. Additionally, Hexb-/-Tnfα-/-mice displayed a delayed neurodegenerative cascade by an extending lifespan, improving sensorimotor coordination, decreasing levels of astrogliosis and neuronal cell death [51]. We suggest that antagonists of pro-inflammatory cytokines and chemokines such as Ccl3 or Ccl4 may dampen neuroinflammation and improve disease outcome in Hexa-/-Neu3-/mice.
Activated microglia produce various pro-inflammatory mediators, chemokines, and cytokines which negatively impact neurons and oligodendrocytes [40]. Here, we showed that Hexa-/-Neu3-/-mice exhibited significant loss in neuronal density and oligodendrocytes. These results are consistent with the previous research indicating increased apoptotic cell death in the brain of Hexa-/-Neu3-/-mice [5]. Increased level of Lif under inflammatory conditions has been shown to inhibit cell proliferation [52], a similar increase was observed in the cortex of Hexa-/-Neu3-/-mice. Furthermore, transcription levels of IL 11, whose overexpression has been shown to inhibit demyelinization by protecting oligodendrocytes [53], was significantly reduced in both the cortex and cerebellum of Hexa-/-Neu3-/mice compared with that in Hexa-/-mice.
The Bmp protein family is responsible for the regulation of bone formation, maintenance, and repair [54]. CNS as well as reduction in neuronal density and oligodendrocytes [30,40]. Additionally, deficits in memory and learning might be related to impaired hippocampal function, which is involved in spatial and/or temporal processing of memory [55], deterioration of Purkinje cells, which play a role in spatial navigation [5,56], and/or damaged retrosplenial and secondary motor corteces, which are involved in memory, navigation, and control of movement [57]. Memory impairments in Hexa-/-Neu3-/-mice, observed in the passive avoidance test, might also be accounted for by a decreased sensitivity to electric shock, due to damage in primary motor and somatosensory cortices, which are regions that process pain control [58].
The swimming speeds of 2.5-and 4.5-month-old Hexa-/-Neu3-/-mice were significantly lower than those of age-matched control groups, although they move equal distances to find the target platform ( Fig. 7). These data suggest that Hexa-/-Neu3-/-mice may not be able to learn the location of the hidden platform due to the memory impairment and damaged signaling pathways in the CNS.

Conclusion
In conclusion, our data suggest that abnormal GM2 ganglioside accumulation in the CNS of early-   The passive-avoidance test. Latencies to enter the dark compartment on the 3rd day were shown for 2.5-(A) and 4.5-month-old (B) mice group. If mice did not enter the dark compartment within 300 s, retention latency time was recorded as 300 s. The data are presented as the mean ± SEM. One-way ANOVA was used for statistical analysis (****p<0.001).