Characterization of a phenotypically severe animal model for human AB-Variant GM2 gangliosidosis

AB-Variant GM2 gangliosidosis (ABGM2) is a rare and lethal genetic disorder caused by mutations in the GM2A gene that lead to fatal accumulation of GM2 gangliosides (GM2) in neurons of the central nervous system (CNS). GM2A encodes a transport protein known as GM2 activator (GM2A) protein, which is essential for degrading GM2 into their GM3 form. ABGM2 presents in infantile-, juvenile-, and adult-onset forms; of the three, the infantile-onset is the most prominent, and by far the most severe, as evidenced by high levels of GM2 accumulation, widespread neurodegeneration, and death by the age of 4. Gm2a−/− mice are commonly used as a model of ABGM2. These mice are characterized by phenotypes most representative of predicted adult-onset form of ABGM2, which include moderate GM2 accumulation and mild neurological defects. This mild phenotype has been attributed to compensation by alternative GM2 degradation pathways mediated by sialidase, neuraminidase 3 (NEU3), a pathway that is more prominent in mice than humans. To assess the extent to which NEU3 contributes to GM2 degradation, we generated double knock-out (Gm2a−/−Neu3−/−) mice. Compellingly, these mice present with a clinical phenotype resembling that of a more severe ABGM2, including ataxia, reduced mobility and coordination, weight loss, poor body scores, and lethality by 6–7 months. Furthermore, these phenotypes correlate with a dramatic increase in GM2 accumulation in the CNS compared to levels observed in either Gm2a−/− or Neu3−/− mice. Taken together, these studies, for the first-time, confirm that the mild neurological phenotype of Gm2a−/− mice is due to compensatory activity on GM2 catabolism through an alternate breakdown pathway involving NEU3. These studies support the use of double knockout mice as a novel and highly relevant model for pre-clinical drug studies in a more severe form of ABGM2.


Introduction
Gangliosides are important components of the plasma membrane of vertebrate cells that modulate cell-cell and cell-pathogen signaling by interacting with membrane receptors, adhesion molecules, and ion channels (Sipione et al., 2020;Quinville et al., 2021).These amphipathic molecules are highly enriched in the central nervous system (CNS), and their synthesis and catabolism are tightly regulated.Gangliosides are transported via endocytosis and sorted to lysosomes where they are degraded by hydrolytic enzymes.A negative surface charge and a narrow pH range (3.8-4.6)stimulates the catabolic activity of these enzymes and their activator proteins.Catabolism of gangliosides involves stepwise hydrolysis of their glycan chains by lysosomal exoglycosidases (reviewed by Quinville et al., 2021).
Gangliosides are the main carriers of sialic acid in the CNS and are essential lipids for the maintenance and maturation of the brain and spinal cord.GM1 and GM2 gangliosides (hereafter referred to as GM1 and GM2, respectively) are a type of ganglioside that have a single sialic acid carbohydrate as a head group.Aberrant regulation or function of these gangliosides has been implicated in a number of diseases; these include Parkinson's disease, Alzheimer's disease, and a class of disorders known as GM2 gangliosidosis (Quinville et al., 2021).
TSD, SD and ABGM2 are a group of severe, autosomal recessive, lysosomal storage disorders characterized by widespread accumulation of GM2 with resultant neuronal apoptosis and early death (Bley et al., 2011).There are three distinct forms of these diseases that are distinguished by their onset, with the most severe being the infantile form, followed by the rarer and milder juvenile-and adult-onset forms.The severity and onset is inversely proportional to the amount of residual β-HEXA or GM2A activity present in affected patients (Conzelmann and Sandhoff, 1991;Leinekugel et al., 1992;Bley et al., 2011).Infantile-onset of these diseases typically manifest between the ages of 3-6 months of age and exhibit the lowest levels of functional β-HEXA or GM2A activity (>2%), resulting in death by 4 years of age (Bley et al., 2011;Sheth et al., 2016).In contrast, the juvenile and adult forms have comparatively higher residual protein activity (5-15%) with time of onset as early as 2 and 21 years of age, respectively.Premature death is evident in juveniles, occurring around 10-15 years of age, while patients with the adult form can live for over 60 years (Neudorfer et al., 2005;Maegawa et al., 2006;Toro et al., 2021).Regardless of the form each patient present with, there is a substantial reduction in quality of life and there is currently no curative treatment available.
The animal model commonly used to study ABGM2 is the Gm2a −/− mouse model (Liu et al., 1997).These mice are homozygous for a Gm2a m1R1p targeted mutation and thus do not express GM2A protein.As such, they were predicted to phenocopy the severity of pathologies that arise in human infantile ABGM2; however, unlike comparable SD mouse models (Phaneuf et al., 1996), the GM2 gangliosidosis-like pathologies in these mice were unexpectedly less severe.They exhibit moderate levels of GM2 accumulation in the brain (at 20 weeks), mild behavioral and motor dysfunction (e.g., decreased coordination), and a normal life span (Liu et al., 1997).Hence, Gm2a −/− mice fail to accurately model the severity of neurological decline and lethality of the infantile-or juvenile-onset forms of human ABGM2.This shortcoming has been postulated to be due to the presence of an alternative, sialidase-mediated GM2 catabolic pathway more prominent in mice than humans (Figure 1).Such a pathway provides an alternative GM2 breakdown route that bypasses or compensates for absent (or weakened) β-HEXA/GM2Amediated breakdown in Gm2a −/− mice (Kolter and Sandhoff, 1998).
Sialidase(s) are capable of hydrolyzing terminal sialic acid from the glycan chains of gangliosides and could potentially convert GM2 to its asialo-GA2 form, which can be further hydrolysed by β-HEXB (an isozyme of β-HEXA) which is present at normal levels in ABGM2 mice (Sango et al., 1995).Three types of sialidases are enriched in mouse neurons: neuraminidase 1 (NEU1), neuraminidase 3 (NEU3) and neuraminidase 4 (NEU4).NEU1 forms a multi-enzyme complex with β-galactosidase and cathepsin A in the lysosome to mediate the breakdown of sialylated glycopeptides and oligosaccharides (reviewed in Monti and Miyagi, 2015).It has strong hydrolyzing activity for GM3 and GD1A, with low activity toward GM2; suggesting that NEU1 is unlikely to Schematic diagram of the proposed pathway for GM2 catabolism.GM1 breakdown is mediated by GM1-β-galactosidase, cleaving the terminal β-D-galactose with the assistance of GM2A or saposin B, resulting in GM2.GM2 is reduced to GM3 by β-HEXA and GM2A to remove N-acetylgalactosamine.Next, GM3 breakdown is mediated by sialidase and saposin B to remove the sialic acid residue, producing LacCer.Alternatively, GM2 can be hydrolyzed by a murine-specific pathway involving the sialidase NEU3.In this pathway, NEU3 removes the terminal sialic acid residue, resulting in GA2.GA2 is then reduced to LacCer by isoforms β-HEXA and β-HEXB by removing the terminal N-acetylgalactosamine.The galactose is then hydrolysed from LacCer by GM1-B-galactosidase with saposin B or galactosylceramide-β-galactosidase with saposin C into GlcCer.Finally, GlcCer is reduced to ceramide by glucosylceramide-ß-glucosidase with saposin C. (reviewed in Quinville et al., 2021).GM1, GM1; GM2, GM2; GM3, GM3; LacCer, lactosylceramide; GlcCer, glucosylceramide; Cer, ceramide; GA2, glycolipid GA2; Glc, glucose; Gal, galactose; GalNAc, remove N-acetylgalactosamine; SA, sialic acid.mediate an alternate pathway in GM2 catabolism (Timur et al., 2015).NEU3 and NEU4 have high substrate specificity to gangliosides in the presence of an activator protein (Akyıldız Demir, 2019).When Neu4 was knocked out in the context of Hexa −/− mice (which are commonly used as a TSD mouse model), significantly higher GM2 levels, epileptic seizures, and neuronal death was observed (Seyrantepe et al., 2010).However, the extent of pathological abnormalities was not as severe as age-matched SD mouse model (Hexb −/− ) (Sango et al., 1995) and mice were still surviving around 12 months of age (Seyrantepe et al., 2010).This data suggests that Neu4 is not the only contributor potentiating the pathological severity of GM2 gangliosidosis in mice.Moreover, NEU3 is responsible for about 30% of neuraminidase activity in the brain and for about 50% of activity against gangliosides (Pan et al., 2017).Mice doubly-deficient in Hexa and Neu3 display severe neurological abnormalities, ataxia, tremors and significantly higher GM2 accumulation in the brain (Pan et al., 2017;Seyrantepe et al., 2018).Neu3 −/− mice also displayed inclusions and vacuoles that are typically observed in lysosomal sphingolipid disorders, and which were not observed in Hexa −/-Neu4 −/− mice (Pan et al., 2017).These studies suggested that NEU3 may mediate an alternative degradative pathway that could compensate for disruption of β-HEXA/GM2Adependent catabolism of GM2.This has led us to hypothesize that ablating Neu3 in a Gm2a −/− mouse background may increase neurological and motor impairment severity and be a more representative model of a severe form of ABGM2.To explore this hypothesis, Neu3 −/− and Gm2a −/− mice were crossbred to generate a Gm2a −/-Neu3 −/− model.Behavioral and motor phenotypes of Gm2a −/− Neu3 −/− mice were then assessed to evaluate their suitability as a more severe animal model to represent human ABGM2.
2 Materials and methods

Generation of a double knockout mouse
Gm2a −/− mice (Liu et al., 1997) were purchased from Jackson Laboratories (Maine, United States).Neu3 −/− (Yamaguchi et al., 2012) mice obtained from the Centre Hospitalier Universitaire Sainte-Justine Research Centre (Montreal, Quebec, Canada) with a kind permission from Professor T. Miyagi, who generated this strain (Yamaguchi et al., 2012).Mice were bred, housed, and maintained at the Animal Care Facility at Queen's University (Kingston, Ontario, Canada) as previously described (Liu et al., 1997;Yamaguchi et al., 2012).All experimental procedures were approved by the institutional Animal Care Committee in accordance with the Canadian Council for Animal Care.

Behavioral testing
A series of behavioral tests were conducted to characterize potential physical impairments in double knockout mice.Testing began at 8 weeks of age and continued every 4 weeks, until 16 weeks of age.After this period, testing was conducted bi-weekly until endpoint.Frequency of behavioral testing was increased to weekly, starting at 16 weeks of age, to determine the rate of behavioral decline.Two behavioral parameters were tested.These included an Open Field Test (OFT, ActiMot, TSE systems, Berlin, Germany) to evaluate gross locomotor activity (Osmon et al., 2018), and a RotaRod test (RR; IITC Life Sciences, California, United States) to evaluate motor coordination and balance (Osmon et al., 2018).
For the OFT, mice were placed in a bordered arena (40 cm × 40 cm) for 5 min and were allowed to roam freely.Resting time, distance traveled, and speed during this time was recorded (Osmon et al., 2018).
For the RR, mice were placed on elevated rotating cylinders that accelerated from 4 to 40 rotations per minute (rpm) over 5 min.When the mouse fell, it off balanced a beam below them, which signaled the end of the test.Each mouse was tested on the RR apparatus three times per time point, with a minimum of 10 min of rest between each trial.Latency to fall, end rotations per minute (rpm) and distance traveled were recorded (Osmon et al., 2018).

Euthanization
Gm2a −/-Neu3 −/− mice were euthanized at their humane endpoint and all other cohorts were euthanized between 27 and 30 weeks of age.A humane endpoint is defined as the point at which a mouse loses 15% of its peak weight, or when a mouse is unable to right itself (which indicates a reduced quality of life).Mice are sacrificed by CO 2 asphyxiation according to the Animal Care Committee guidelines.Organs were perfused with phosphate-buffered saline (PBS) and collected thereafter, for future biochemical and histological analysis.
Cardiac puncture on euthanized mice was used to collect blood to assess serum protein levels.Organ perfusion with PBS was carried out following cardiac puncture.Gross organs and CNS tissue were collected for Western blot analysis, ganglioside accumulation assays and histological imaging.Regions of the brain and spinal cord dissected for tissue collection and sectioning are shown in Supplementary Figure 1.

Ganglioside storage assay
A modified Folch extraction was used as described in Osmon et al. (2018).Briefly, mid-sections of the brain were taken to be representative of whole murine brain.Each mid-section was sonicated and transferred to a glass tube (UltiDent Scientific Inc., Quebec, Canada).Next, a series of methanol and chloroform dilutions were added to the sample and then flowed through a column (Phenomenex, California, United States) to collect gangliosides.Samples were eluted and added to a thin layer chromatography plate (TLC; MilliporeSigma, Massachusetts, United States) for development.The plate was placed in a tank containing chloroform, methanol, and calcium chloride.The plate was developed in the tank for several hours to allow individual gangliosides to separate, and then sprayed with a drying solution made up of orcinol and sulfuric acid.Individual gangliosides were quantified by densitometry by comparing the intensity of the GM2 bands against an internal control (GD1A) using IMAGE J software (National Institute of Health, Maryland, United States).A TLC Monosialoganglioside mixture (MJS Biolynx Inc., Ontario, Canada) and Gm2a −/− mid-section samples were used as controls.GD1A was used as an internal control as it is a constant presence in mice, throughout adulthood (Yu et al., 2012).
MUG and MUGal activity were assessed by incubating serum samples at 37°C in a 3.2 mM MUG or MUGal substrate solution.After 1 h, enzyme activity was stopped with 0.1 M 2-amino-2-methyl-1propanol (AMP).MUG and MUGal activity were estimated based on a 4-methylumbelliferone (4-MU) standard curve, which ranged from 0.73 nM to 130 μM.Activity was detected by colorimetric intensity at an excitation wavelength of 365 nm and an emission wavelength of 450 nm on a Varioskan™ microplate reader (Thermo Fisher Scientific, Massachusetts, United States).HexA activity was calculated by subtracting HI sample activity (representative of HexB activity) from total MUG activity (all hexosaminidases), after being normalized to MUGal concentrations.

Histology
Portions of the mid-and caudal-sections of the brain were fixed in 4% paraformaldehyde (PFA) for 24-48 h and then transferred to 70% ethanol for another 24-48 h.Samples were then transferred to PBS and shipped to The Center for Phenogenomics (Toronto, Ontario, Canada) for slide preparation and staining with hematoxylin and eosin, or with an anti-GM2 antibody (a gift from Kyowa Hakko Kirin Co. Ltd., Tokyo, Japan).Ganglioside accumulation was quantified using QuPath Software (Belfast, Ireland).

Statistical analysis
All statistical analyses were performed with GraphPad PRISM 9.3.1.Statistical testing for all behavioral modalities were performed using two-way repeated measures analysis of variance (ANOVA) with Tukey post hoc test to compare cohorts across all timepoints.Ganglioside accumulation and hexosaminidase activity were analyzed by a one-way ANOVA with a Tukey post hoc test for multiple pairwise comparisons.A Kaplan Meier curve was used to present the survival patterns of each cohort, which was analyzed with a Log-Rank (Mantel-Cox) test to assess significant survival differences.A nonlinear regression (curve fit) was used to determine if there was a correlation between symptom onset and age of death.

Protein and enzyme activity altered in
Gm2a −/− Neu3 −/− mice A Western blot was performed to confirm the absence of GM2A protein and to determine the effect of knocking out Neu3 on the expression of the αand β-subunits in the brain (Figure 2A).As expected, GM2A protein was absent in Gm2a −/− and Gm2a −/-Neu3 −/− mice and unaffected in wild-type and Neu3 −/− mice (Figures 2A,B).Expression of the α-subunit was not impacted by knocking out Neu3 in Gm2a −/− mice, and there were relatively equal levels of the α-subunit between all cohorts (Figure 2C).Expression of the β-subunit in Gm2a −/-Neu3 −/− mice was significantly lower than wild-type mouse brains (Figure 2D).A similar trend was seen in Gm2a −/− mice, which had reduced expression of the β-subunit compared to wild-type and Neu3 −/− mice (Figure 2D).To determine the correlation between expression of each subunit and β-HEXA activity, a hexosaminidase activity assay was conducted on mouse serum (Figure 2E).Here, we saw no significant differences in β-HEXA activity in any of the cohorts, regardless of the slightly altered expression levels in the subunits observed in some of the cohorts.
The newly developed mouse strain exhibited a significantly reduced average lifespan of 27 ± 1.5 weeks (range from 24.43 to 30.0 weeks), compared to Gm2a −/− mice (91.9 ± 9.8 weeks) (Figure 3A).Gm2a −/-Neu3 −/− mice exhibited a range of physical impairments indicative of disease manifestation (Supplementary Figure 3B).A Protein expression and enzyme activity in Gm2a −/− Neu3 −/− mice.(A) Representative Western blot analysis of GM2A protein and the α-and β-subunits of β-HEXA expression in murine brains.Mid-sections of the brain were dissected from Gm2a −/− Neu3 −/− mice at humane endpoint and compared to brain sections extracted from controls (Gm2a −/− Neu3 −/− and wild-type) that were euthanized between 27 and 30 weeks of age.Bands migrating at ~20 kDa or 22 kDa depict the mature and precursor forms of GM2A, respectively.Bands migrating at 61 kDa depict the α-subunit of β-HEXA and those migrating at 63 kDa depict the β-subunit of β-HEXA.β-actin was utilized as an internal control and migrates at 42 kDa.(B) Quantification of GM2A protein signal in the Western blot.The intensity of the mature and precursor forms of GM2A were taken as a representation of total GM2A protein signal and were normalized to β-actin intensity.Intensities were quantified by densitometry analysis.Gm2a   3A).Depending on symptom severity, Gm2a −/− Neu3 −/− mice lived, on average, for 1.5-3 weeks after the appearance of the first signs of the disease.Moreover, the earlier the symptoms appeared, the earlier the humane endpoint was reached (see Methods: Euthanizations).The predominant symptoms indicative of disease included tremulous movement and ataxic gait (Supplementary Figure 3B); these began as early as 21 weeks of age in some mice, while others were relatively asymptomatic until 28 weeks of age.In addition, Gm2a −/− Neu3 −/− mice began to lose weight around 20-22 weeks of age with at least 66.7% of them losing over 15% of their peak body weight at humane endpoint (Figures 3B,C).Other symptoms included seizures (4.2%), eye infections (16.7%), and death (8.3%) although these were of substantially less predominance (Supplementary Figure 3B).Four Gm2a −/− Neu3 −/− mice experienced hindlimb paralysis around 9-12 weeks of age; however, they have not been included in any of the analyses as we were unable to confirm the cause.Although, this is a classic symptom observed in the SD mouse model (Phaneuf et al., 1996).

Coordination and balance are severely impaired in Gm2a
Coordination and balance were assessed on a RR apparatus which measures the time that elapses before a mouse falls off an elevated rotating cylinder.Three RR parameters were assessed, including latency to fall, distance traveled, and end RPM.All cohorts tested on this apparatus performed comparably at approximately 8 weeks of age but started to deviate at around 12 weeks of age (Figures 4A,E,I).After 12 weeks of age, Gm2a −/− Neu3 −/− mice exhibited a downward trend in performance for all three RR parameters assessed.These trends became significant for all 3 parameters by 20-24 weeks of age and remained significant at later time points (Figures 4B-D,F).By 24 weeks of age, most of the Gm2a −/− Neu3 −/− mice were incapable of staying on the rotating cylinder at all.This trend was consistent across all other RR parameters.This progressive decrease in performance on RR with the Gm2a −/-Neu3 −/− mice was not observed in other cohorts, suggesting that knocking out Neu3 in Gm2a −/− mice imparts coordination and balance defects that are absent in Gm2a −/− .In addition, Gm2a −/− mice appear to perform similarly to the wild-type cohort, which opposes a previous paper that characterizes this model (Liu et al., 1997).

Locomotion is severely impaired in
Disparities in motor function were much more apparent between cohorts when evaluated by the OFT, with differences in the Gm2a −/− Neu3 −/− mice beginning as early as 8 weeks of age (Figures 5A,E,I).For the OFT parameters evaluated-mean speed, Remarkably however, Gm2a −/− Neu3 −/− mice exhibited significantly reduced scores for the mean speed and distance traveled parameters of the OFT when compared to all the cohorts (Figures 5B-D,F).Correspondingly, Gm2a −/− Neu3 −/− mice resting time was significantly longer than the other cohorts (5B-5D).These differences became significant at 8-12 weeks of age and persisted through all time points.Furthermore, motor impairments in Gm2a −/− Neu3 −/− mice grew progressively worse as indicated by their OFT motor scores, which increasingly diverged from the scores of other cohorts over time (Figures 5A,E,I).These data clearly demonstrate that double knockout mice have impaired gross locomotor activity beginning very early in life.

GM2 accumulates in Gm2a
mice to levels representative of a more severe form of ABGM2 GM2 is a breakdown product of GM1 in the ganglioside catabolic pathway, which then continues to be metabolized in a stepwise fashion into glucosylceramide (GlcCer).This process involves the intermediary products GM3 and LacCer (Figure 1).Importantly, a branching pathway from GM2 to GA2 and then LacCer has also been identified as an alternative catabolic route for GM2 breakdown (Sango et al., 1995; Figure 1).Hence, GlcCer is a key convergence point for two GM2 breakdown pathways and its levels are an important marker of total GM2 breakdown activity.
GM1, GM2, GM3 and GlcCer were extracted and quantified using TLC.Quantitative analysis of glycolipids was carried out on the mid-section of brains, which are taken as representative of the whole murine brain.The brains were collected at humane endpoints for Gm2a −/− Neu3 −/− mice and at 26-30 weeks of age for all other cohorts.Interestingly, we were unable to detect GM2 accumulation in Neu3 −/− brain tissue indicating that GM2 catabolic efficiency is unimpacted by knocking out the NEU3-mediated route of GM2 breakdown (Figures 1,  6A), likely because GM2 is more efficiently catabolized by β-HEXA/ GM2A-mediated pathway.These results were consistent with a previous study (Seyrantepe et al., 2018).In contrast, appreciable GM2 accumulation was detected in the brains of Gm2a −/− mice, suggesting that GM2 breakdown is impaired, and confirms that GM2A co-factor activity for the β-HEXA enzyme is involved in GM2 degradation.However, knocking out Neu3 in Gm2a −/− backgrounds resulted in a striking increase (approximately 3-to 4-fold) in GM2 levels relative to Gm2a −/− mice.This demonstrates that significantly less GM2 is degraded in Gm2a −/-Neu3 −/− mice than in Gm2a −/− mice (Figure 6A).This also indicates that the NEU3-mediated pathway may be compensating for impaired GM2 breakdown in Gm2a −/− mice, and that in the absence of this compensation, GM2 accumulation is significantly higher.The levels of other gangliosides quantified remain unchanged in all cohorts (Figures 6B,C).Importantly, the levels of GM2 accumulation in the double knockouts parallel the severity of GM2 accumulation observed in a bona fide model of infantile SD cohort of age-matched Hexb −/− mice (Figure 6A).In a separate experiment, the progression of GM2 accumulation in double knockout mice was also assessed over five different time points: 8, 12, 16, 20 and 26 weeks of age (Figure 6D).There were no significant differences seen between any of the cohorts at any of the timepoints.It is possible that this experiment was not sufficiently powered to see differences in ganglioside levels throughout their lifetime.
Histological analysis of murine brain sections supported our GM2 accumulation findings (Figure 7).The cortex (Figure 7A), cerebellum (Figure 7B) and pons (Figure 7C) in Gm2a −/− Neu3 −/− mice display large accumulations of GM2 with no buildup visible in other cohorts.Accumulation in the cortex and pons is uniform across the entire section, whereas GM2 is almost exclusively built up in the white matter of the cerebellum (Figure 7).While GM2 is more abundant in the gray matter, it also has a major role in the development of myelin (Kraĉun et al., 1984;Sipione et al., 2020).Vacuolization is also visible in all histological brain sections of Gm2a −/− Neu3 −/− mice but is most notable in the cerebrum and pons.These vacuoles appear to be larger and more abundant in all knockout cohorts compared to wild types.

Discussion
ABGM2 is the rarest of the GM2 gangliosidosis, with an estimated incidence rate of <1/1,000,000 and less than 30 reported cases, most of which are infantile onset (Sheth et al., 2016).The disease, with no current curative treatment, is believed to be underdiagnosed as there is not a consistent protein assay established for GM2A quantification in most clinical laboratories and the disease was less suspected/examined until recently when molecular testing with gene panels became available (Turro et al., 2020).To develop effective therapies, a representative animal model was essential.In this study, we knocked out Neu3 in a Gm2a −/− mouse, which is involved in the alternate catabolic pathway for GM2 ganglioside metabolism (Kolter and Sandhoff, 1998).The resultant double knockout mouse model (Gm2a −/− Neu3 −/− ) represents neuropathological and clinical abnormalities of a more severe form of ABGM2, compared to Gm2a −/− mice, and provides a valuable resource for future treatment development.
GM2 accumulation in Gm2a −/− Neu3 −/− mice was comparable from 8 weeks of age to the animal's humane endpoint (Figure 6D).In SD mice, GM2 accumulation is detected early in life; however, behavioral symptoms are not apparent until 12-16 weeks of age (Phaneuf et al., 1996;Hooper and Igdoura, 2016), suggesting that it is likely the downstream effects that trigger the progression of neurological decline in these animals.There is a sharp increase in microgliosis around 8-12 weeks of with subsequent astrogliosis around 12-14 weeks of age that coincides with increased caspase 9, a biomarker of apoptosis, along with other pro-inflammatory pathway (Hooper and Igdoura, 2016).Concurrent with astrocyte activation is a large decrease in motor function, suggesting that neuroinflammation is what drives disease pathology in GM2 gangliosidosis (Hooper and Igdoura, 2016).These timelines coincide with the decline in the behavioral parameters observed in Gm2a −/− Neu3 −/− mice, which becomes most pronounced around 12 weeks of age and continues to deteriorate progressively (Figures 4, 5).
In the current investigation, a widespread accumulation of GM2 is observed in the white matter of murine brains, particularly in the cerebellar cortex (Figure 7).This accumulation is believed to also contribute to the phenotypic manifestation observed in behavioral testing, as disruption in the normal maturation of white matter can affect motor pathways, such as the pyramidal tract (Maegawa et al., 2009).Gangliosides play a major role in the development and stability of myelin (Schnaar, 2010;Vajn et al., 2013;Sipione et al., 2020), which is decreased in all GM2 gangliosidosis animal models (Maguire and Martin, 2021).Additionally, striking vacuolization is also visible in all histological brain sections of each cohort but is most notable in the cerebrum and pons of knockout mice.An abundance of vacuoles is characteristic of lysosomal storage disorders and has been previously observed (Pan et al., 2017).
The present study provides evidence that corroborates previous research regarding the involvement of NEU3 in ganglioside catabolism (Kolter and Sandhoff, 1998;Pan et al., 2017;Seyrantepe et al., 2018).The disruption of NEU3 protein activity resulted in an accelerated disease pathology, which is most likely attributed to the increased accumulation of GM2.However, if NEU3 were the sole sialidase involved in the metabolism of GM2, the phenotypic severity would match that of Hexb −/− mice (Phaneuf et al., 1996).When Neu3 was knocked out in Hexa −/− mice, minor accumulation of GA2 was observed, indicating a possible involvement of another sialidase in this pathway (Seyrantepe et al., 2018).
The alternate sialidase involved is likely NEU4, given that a loss-offunction mutation to this protein in Hexa −/− mice resulted in a more severe phenotype than that observed in single knockout Hexa −/− mice (Seyrantepe et al., 2010).Additionally, it is possible that in the absence of NEU3, both NEU4 and NEU1 may partially compensate for this sialidase loss, as all neuraminidase activity has been shown to be increased in mouse models of GM2 gangliosidosis (Seyrantepe et al., 2018).
We also found that knocking out Gm2a impacted expression of the β-subunits in the brain (Figures 2C,D).This is the first study to note these differences in expression as hexosaminidase isozyme activity is typically unaltered in ABGM2 models (Liu et al., 1997).An increase in expression of the β-subunit is observed in wild-type and Neu3 −/− mice, but not Gm2a −/− Neu3 −/− and Gm2a −/− mice.It's possible that when the interaction between β-HEXA (αβ) and β-HEXB (ββ) and GM2A is altered, due to its absence, β-HEXA (αβ) and β-HEXB (ββ) are downregulated.β-HEXA enzyme activity was equal amongst all cohorts (Figure 2E), suggesting that the discrepancy in β-subunit expression is likely solely due to β-HEXB.However, there is a wide range of intra-cohort variability in the hexosaminidase assay, thus, further experimentation with a larger sample size is required to draw more concrete conclusions.
Our study presents compelling evidence regarding the involvement of NEU3 in GM2 metabolism and ABGM2 pathogenesis in mice.The newly developed Gm2a −/− Neu3 −/− mouse model provides a more accurate representation of a severe form of the disease, both phenotypically and biochemically.These mice offer a unique opportunity to assess treatment effects on significant symptoms of the human disease, such as lifespan, coordination, gross locomotor activity and pathogenesis.As such, our findings hold great promise for future research on ABGM2 and potential treatments.