The Multifaceted Roles of Brain Glycogen

Abstract

Glycogen is a biologically essential macromolecule that is directly involved in multiple human diseases. While its primary role in carbohydrate storage and energy metabolism in the liver and muscle is well characterized, recent research has highlighted critical metabolic and non-metabolic roles for glycogen in the brain. In this review, the emerging roles of glycogen homeostasis in the healthy and diseased brain are discussed with a focus on advancing our understanding of the role of glycogen in the brain. Innovative technologies that have led to novel insights into glycogen functions are detailed. Key insights into how the cellular localization impacts neuronal and glial function are discussed. Perturbed glycogen functions are observed in multiple disorders of the brain, including where it serves as a disease driver in the emerging category of neurological glycogen storage diseases (n-GSDs). n-GSDs include Lafora disease (LD), adult polyglucosan body disease (APBD), Cori disease, Glucose transporter type 1 deficiency syndrome (G1D), GSD0b, and late onset Pompe disease (PD). They are neurogenetic disorders characterized by aberrant glycogen which results in devastating neurological and systemic symptoms. In the most severe cases, rapid neurodegeneration coupled with dementia results in death soon after diagnosis. Finally, we discuss current treatment strategies that are currently being developed and have the potential to be of great benefit to n-GSD patients. Taken together, novel technologies and biological insights have resulted in a renaissance in the brain glycogen that dramatically advanced our understanding of both biology and disease. Future studies are needed to expand our understanding and the multifaceted roles of glycogen and effectively apply these insights in human disease.

Graphical Abstract

Glycogen is an essential carbohydrate storage molecule involved in multiple tissue-specific processes. This review provides a comprehensive summary of our understanding of brain glycogen and the multifaceted roles of glycogen proposed by recent articles. The roles are described in health and disease with a focus on current evidence linking glycogen and glycosylation. Finally, we highlight the therapeutic potential of targeting glycogen metabolism and potential avenues for future research.

Glycogen is one of the fundamental biological molecules and is critical for primary carbohydrate storage, metabolism, and emerging novel roles. Its central role in energy homeostasis has long been studied (Bernard 1857; Foster 1899; Carpenter, Smith, and Meneses 1876). Glycogen is found in high levels in the liver and muscle, where it serves as systemic and local carbohydrate reserves. However, the machinery for glycogen synthesis and degradation is widely expressed throughout the body, and other tissues and organs have been increasingly recognized to critically depend on cellular glycogen (Brewer and Gentry 2019a; Dienel and Carlson 2019; Obel et al. 2012). Through the lenses of newly developed technologies, spatially unique glycogen distribution in the brain, lung, testis, kidney, and bone marrow has been increasingly studied (Oe et al. 2016; Young et al. 2020; 2022; Sun et al. 2019; 2021).

Interestingly, glycogen displays heterogeneous patterns of distribution within the brain (Oe, Akther, and Hirase 2019) and is observed in diverse cell types within the central nervous system (CNS), including neurons, astrocytes, and microglia (A. M. Brown and Ransom 2007; Hertz, Peng, and Dienel 2007). This has significance for key aspects of brain biology as well as implications for diseases that affect the brain. Currently, our understanding of brain glycogen includes its pivotal roles in learning and memory, signaling events, neurotransmitter metabolism, and protein glycosylation. Here, we review multiple roles of glycogen in the brain. We also highlight key new technologies and insights into the biology of glycogen that may have profound implications for the brain. Finally, we discuss the role of glycogen in neurological disorders and address therapeutic strategies targeting glycogen metabolism to alleviate major symptoms of neurological disorders where glycogen is the primary disease driver.

Brain Glycogen – architecture

Glycogen consists of glucose molecules linked via glycoside bonds with α-1,4 chains and α-1,6 branches. This design allows the storage of up to ~50,000 glucose units (Shearer and Graham 2002; Meléndez, Meléndez-Hevia, and Cascante 1997; Marchand et al. 2002). Glycogen synthase is responsible for elongation of the α-1,4 chains and glycogen branching enzyme catalyzes the α-1,6 branches. Conversely, glycogen degradation in the cytosol is carried out by glycogen debranching enzyme and glycogen phosphorylase, while lysosomal glycogen is degraded by glucosidases.

In addition to glucose moieties, recent work revealed other key covalent components in brain glycogen. Brain glycogen contains covalently attached phosphate. Glycogen phosphate content is regulated by laforin, a glycogen phosphatase (Tagliabracci et al. 2007; Gentry, Worby, and Dixon 2005; Raththagala et al. 2015). More recently, work revealed the incorporation of monosaccharides other than glucose in brain glycogen, which is discussed in more detail below. Glycogen structure is defined by branching pattern, glucose chain length, and composition including covalent phosphate (Brewer and Gentry 2019b). Differences in these key factors modulate glycogen solubility, accessibility to interacting enzymes, and granular size (Brewer and Gentry 2019a; Gentry et al. 2018). Multiple glycogen storage disorders (GSDs) present with increased phosphate content and/or abnormal branching and chain elongation (Sullivan et al. 2019; Nitschke et al. 2017; B. I. Brown and Brown 1966; Mercier and Whelan 1973). Tissue- and context-specific glycogen architecture with respect to biology and disease remains a critical knowledge gap in the field.

Brain Glycogen – distribution

Determination of glycogen distribution in the brain was previously challenging due to technical limitations and the detection limits of methods available at the time. However, the development of specific glycogen antibodies and improved rapid fixation methods have allowed visualization of brain glycogen in two dimensional space with cellular details (Oe et al. 2016; Oe, Akther, and Hirase 2019; Baba 1993). Brain glycogen distribution exhibits high spatial and cellular heterogeneity, being present in the hippocampus, cerebellum, layers of the neocortex, striatum, substantia nigra, and pons; regions surrounding the vasculature, and white matter regions such as the hypothalamus and thalamus (Oe, Akther, and Hirase 2019; A. Brown, Baltan, and Ransom 2009; A. M. Brown, Rich, and Ransom 2019). Interestingly, glycogen distribution exhibits spatial heterogeneity within white and gray matter regions, which may be coupled to functional differences. This spatial heterogeneity raises interesting questions regarding the integrative roles of glycogen in these respective brain regions.

Glycogen is observed in the cytosol of multiple cell types in the brain. Astrocytes store the largest quantities of glycogen, with neurons and microglia also observed to have significant glycogen stores (Saez et al. 2014; Oe et al. 2016). Of note, neurons possess the enzymatic machinery for glycogen metabolism and have been shown to actively engage in glycogen turnover. Indeed, neuronal glycogen is essential for neuronal survival under hypoxic conditions in vitro, and it plays important roles for memory formation and hypoxia tolerance in vivo (Duran et al. 2013; Saez et al. 2014). However, the relative fraction of glycogen and glycogen turnover in neurons and glia cells remains to be further investigated.

Interestingly, brain glycogen can be further compartmentalized on a subcellular level and the concentration and turnover may vary according to specific compartments. For instance, elevated glycogen has been observed in the brain in cellular processes including synaptic endfeet (Dienel and Carlson 2019; Richter, Hamprecht, and Scheich 1996; Pfeiffer-Guglielmi et al. 2003; Bak and Walls 2018). In peripheral organs, glycogen is also compartmentalized at the subcellular level, with glycogen being identified in the nucleus, mitochondria, and in close proximity to the endoplasmic reticulum (Baird and Fisher 1957; De Man, Blok, and Beens 1966; Cardell 1977; Nielsen et al. 2010; Ishikawa and Pel 1965). The first description of nuclear glycogen was reported in hepatocytes in the 1940s and subsequently confirmed by multiple laboratories (Chipps and Duff 1942; Baird and Fisher 1957; Bogoch et al. 1955). The recently identified role of nuclear glycogen in histone acetylation suggests that subcellular localization of glycogen plays key roles in fundamental biology which have not been previously appreciated (Sun et al. 2019). To this end, the subcellular localization of glycogen is an exciting area for future scientific advances and remains to be further elucidated within brain cells.

Brain Glycogen – detection and localization

Multiple methods are widely utilized to analyze and quantify glycogen, including periodic acid-Schiff (PAS) staining, alcohol or trichloroacetic acid (TCA) precipitation, followed by biochemical enzymatic assay, and electron microscopy (EM) (Figure 1) (Brewer and Gentry 2019a). Glycogen was first purified by ethanol and TCA precipitation from 1909 to the 1930s, but this method required a substantial amount of tissue (Pflüger 1909; Somogyi 1934; Willstätter and Rohdewald 1934). In the 1930s-1940s, direct visualization of glycogen was achieved through the implementation of electron microscopy (EM). EM provided key insights into glycogen structure/architecture (scanning EM) and sub-cellular locations within a cell (transmission EM).

Figure 1: Glycogen and glycan detection methods.

Overview of common methods to detect glycogen and glycans, including overlapping methods. MALDI-MSI can be utilized to detect both simultaneously. Created with BioRender.com. Abbreviations: EM – electron microscopy, PAS – periodic acid-Schiff, HPAEC-PAD – high performance anion-exchange chromatography with pulsed amperometric detection, GCMS – Gas chromatography mass spectrometry, MALDI-IMS – Matrix-assisted laser desorption/ionization imaging mass spectrometry, MS – mass spectrometry, SDS-PAGE – sodium dodecyl sulfate–polyacrylamide gel electrophoresis, NMR – nuclear magnetic resonance.

One of the first techniques to provide histological annotation of glycogen came with the development of PAS staining by McManus in 1946 (McManus 1946; 1948; Aterman and Norkin 1963). PAS is easy to perform, and coupled with a rapid turnaround time, it is an ideal method for clinical pathology. Even today, PAS is utilized in clinical settings for the diagnosis of diverse diseases, including Ewing’s sarcoma (Mizuguchi et al. 2016; Hartman et al. 1991). As a research tool, PAS staining has several limitations due to low specificity caused by reactivity with a wide range of polysaccharides, including glycoproteins, glycolipids, glycogen, and mucins. Additionally, while PAS in combination with diastase treatment (amylase) can improve selectivity, the technique is not sensitive enough to detect physiological glycogen levels in many tissues and cell types.

The development of glycogen antibodies dramatically improved our knowledge of regional glycogen distribution (Baba 1993; Nakamura-Tsuruta et al. 2012). Two mouse monoclonal IgM glycogen antibodies, IV58B6 and ESG1A9, have been verified to detect glycogen by immunohistochemical methods. While both antibodies detect glycogen, ESG1A9 possesses higher affinity for larger glycogen molecules, and IV58B6 displays a more uniform affinity for glycogen of all sizes (Oe, Akther, and Hirase 2019; Oe et al. 2016). Additionally, brain glycogen is difficult to quantify and study due to its rapid post-mortem degradation (Wu, Wong, and Swanson 2019; Pontén, Ratcheson, and Siesjö 1973). In fact, research on glycogen extraction methods and degradation time suggest that up to 95% of glycogen may be degraded during improper experimental- and extraction conditions (Cruz and Dienel 2002; Dienel 2020; Hsu et al. 2021). The implementation of focused microwave fixation in situ has provided a significant advantage for the accurate detection of physiological glycogen compared to other approaches (Oe et al. 2016; Juras et al. 2022).

Importantly, recent technological advances are allowing multi-scale definition of the cellular, spatial, and systemic metabolism of the central nervous system. Multiple techniques for accurate and sensitive glycogen detection have been established and widely implemented to measure glycogen from mouse to man. These techniques include gas chromatography mass spectrometry (GCMS), matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI), magnetic resonance spectroscopy (MRS), and magnetic resonance imaging (MRI) (Figure 1) (Soares, Gruetter, and Lei 2017). The development of GCMS-based methods to detect glycogen dramatically improved accuracy and sensitivity (Young et al. 2020; Andres et al. 2020). MALDI-MSI combines the accuracy and sensitivity of mass spectrometry with added spatial information and glycogen architecture (Conroy, Allison, et al. 2021; Young et al. 2022; Hawkinson and Sun 2022). This technique offers crucial quantification of regional- and architectural information and is approaching single-cell resolution. While MS-based assays focus on endpoint analysis, glycogen quantitation in living organisms is made possible through recent development in MRI (Y. Zhou et al. 2020). While having reduced sensitivity compared to MS-based methods, these methods have the significant advantage of being non-invasive and applicable to imaging the entire human brain, allowing live imaging of glycogen. One such MRI method relies on magnetic coupling between glycogen and water through nuclear Overhauser enhancement (i.e., glycoNOE), and represents a promising technique for glycogen detection and treatment evaluation for GSDs (Y. Zhou et al. 2020). Collectively, utilizing these newer techniques will be crucial for deepening our understanding of brain glycogen with respect to spatial resolution and imaging glycogen in live organisms.

The pleiotropic roles of glycogen

The brain is an energy-demanding organ relying on a continuous supply of glucose (Bordone et al. 2019). Interestingly, the brain is also uniquely susceptible to damage caused by fluctuating glucose and oxygen levels. Brain glycogen promptly responds to changing cerebral energy demands and is degraded for energy production as an emergency fuel in response to low glucose levels. However, brain glycogen cannot sustain metabolic demands for long due to the high energy demands of the brain and low glycogen levels. Illustrating this, glycogen is very rapidly degraded post-mortem (Dienel and Carlson 2019; Dienel 2019b; Pontén, Ratcheson, and Siesjö 1973; Wu, Wong, and Swanson 2019). Furthermore, brain glycogen quickly reacts to changing energy demands and is rapidly degraded during a hypoglycemic episode or ischemia in both astrocytes and neurons (Guo et al. 2021; Bastian et al. 2019; Dienel and Rothman 2019). Importantly, neurons possess active glycogen metabolism proteins, and these enzymes are needed to protect neurons from hypoxia (Saez et al. 2014). Indeed, cultured neurons both de-novo synthesize and degrade glycogen in response to hypoxia by activation of glycogen synthase and glycogen phosphorylase (Vilchez et al. 2007). Additionally, it has been increasingly appreciated in recent years that glycogen has pleiotropic roles that support normal brain function, and dysregulation of glycogen metabolism contributes to multiple diseases (Waitt et al. 2017; Obel et al. 2012; Bak and Walls 2018; Dienel 2019b; DiNuzzo and Schousboe 2019) (Figure 2).

Figure 2: Integrated and proposed novel roles of Glycogen.

Glycogen is a diverse molecule rapidly responding to changing energy levels and serve multiple vital functions in several bodily organs. Created with BioRender.com.

Further, glycogen can be utilized in a coupled multi-cellular context. Work by Magistretti and colleagues demonstrated a relationship between glycogen and lactate, suggesting that astrocytes degrade glycogen to produce lactate for neuronal energy demands known as the Astrocyte-Neuron-Lactate-Shuttle (ANLS) (Magistretti and Allaman 2015) (Figure 3). However, the ANLS remains a controversial hypothesis due to stoichiometric disagreements and the ability of neurons to utilize lactate. Thus, other hypotheses explaining the relationship between glucose/glycogen metabolism and neurotransmission have been proposed, including the Glucose Sparing by Glycogenolysis (GSG) hypothesis (Rothman et al. 2022; Dinuzzo et al. 2012; Swanson 1992). The GSG hypothesis suggests that astrocytes utilize glycogen to spare glucose and thereby increase glucose availability for neurons during high energy demands. This hypothesis also provides an explanation for the stoichiometry and energy demands of the astrocytic glutamate/GABA oxidation that the ANLS does not fully explain. Yet, both hypotheses include limitations that must be further elucidated. Collectively, the complexity of metabolic pathways and interrelationship between neurons and astrocytes awaits further investigation, but the central importance of glycogen in neurons and astrocyte remains clear (Dienel 2019b; Bélanger, Allaman, and Magistretti 2011).

Figure 3. Glycogen is a critical component of cerebral metabolism.

Glycogen metabolism directly supports neuronal and astrocytic function. Glycogen is also found in microglia. Glycogen supports brain energy metabolism, but controversies exist regarding the underlying mechanisms and energetic coupling between neurons and astrocytes. Created with BioRender.com. Abbreviations: Glu – glutamate, gln – glutamine, ATP – adenosine triphosphate.

Cognitive function and memory formation

The earliest descriptions of glycogen in memory and learning were demonstrated in chickens and based on impaired learning after injections of a glycogen phosphorylase inhibitor 1,4-dideoxy-1,4-imino-d-arabinitol (DAB) (Gibbs, Anderson, and Hertz 2006). These findings have been supported by other studies demonstrating that impaired memory and learning in rat astrocytes and neurons, utilizing DAB or lactate transporter inhibition/deletion to perturb mobilization of astrocytic glycogen-derived lactate (Alberini et al. 2018). Furthermore, multiple other studies suggest that astrocytes contribute to memory formation as well and are crucial for reward location (Doron et al. 2022). While physiologically important, mechanistic insights were limited, thus leaving knowledge gaps regarding the corroboration of cell types. Emerging evidence supports a role for both astrocytic and neuronal glycogen in memory formation that calls for further investigation (Duran et al. 2020; Duran, Gruart, Varea, et al. 2019).

Astrocytic glycogen can support multiple functions, including lactate release, glutamate uptake, and astrocytic glutamine synthesis (Suzuki et al. 2011; Sickmann et al. 2009; Andersen and Schousboe 2022). Glutamine is a precursor for glutamate, and biosynthesis of glutamine occurs primarily in astrocytes and can be synthesized from glycogen. Interestingly, pharmacological inhibition of glutamine synthesis resulted in glycogen accumulation in astrocytes and perturbed memory consolidation in chickens and mice (Folbergrová et al. 1969; Swanson et al. 1989; Gibbs, Anderson, and Hertz 2006). Additionally, glutamine supplementation can improve impaired memory and learning caused by perturbed glycogen metabolism in chickens (Gibbs et al. 2007; Gibbs, Anderson, and Hertz 2006). Thus, glutamine is intimately linked to memory formation and glutamatergic neurotransmission awaiting further elucidation.

Glutamate is the major excitatory neurotransmitter in the brain, and glutamatergic neurotransmission is closely coupled to energy metabolism and tightly regulated to maintain homeostasis and prevent seizure generation (Bordone et al. 2019). Multiple laboratories demonstrated that glutamatergic neurotransmission is coupled to glycogen metabolism by employing inhibitors of glycogen metabolism enzymes in cultures, co-cultures, and slices (Dienel 2019b; Sickmann et al. 2009; Drulis-Fajdasz et al. 2015; Mozrzymas, Szczęsny, and Rakus 2011). Indeed, it is well-established that inhibition of glycogenolysis results in disrupted excitatory signaling and impaired long-term potentiation (Suzuki et al. 2011; Mozrzymas, Szczęsny, and Rakus 2011; Duran, Gruart, Varea, et al. 2019). However, the mechanism for synaptic perturbations caused by inhibition of glycogenolysis is not fully elucidated and has been linked to both neuronal and astrocytic processes.

Astrocytes provide another link to glycogen and glutamatergic neurotransmission through their clearance of extracellular K+. Strikingly, astrocytic clearance of extracellular K+ directly causes glycogen degradation. The current model for this degradation is the energy demands of the Na+/K+-ATPase that stimulates glycogenolysis and respiration for ATP production (Hof, Pascale, and Magistretti 1988; Xu et al. 2013; Choi et al. 2012). Thus, clearance of K+ depends on glycogen, and inability in clearance of K+ could lead to seizure generation and hyperexcitability (Hof, Pascale, and Magistretti 1988; Choi et al. 2012). Taken together, glycogen directly impacts glutamatergic neurotransmission by multiple processes, including providing substrate for glutamate production and K+ clearance.

More recent, studies with transgenic models have demonstrated that genetically altered mice devoid of CNS glycogen, i.e., Nestin-Cre-Gys1 knockout, survive and have a normal lifespan, but present with impaired memory formation and long-term memory consolidation (Duran et al. 2013; Gibbs, Anderson, and Hertz 2006). Follow-up studies based on genetic manipulation of neuronal glycogen metabolism demonstrated that animals devoid of glycogen in hippocampal neurons have impaired synaptic plasticity and memory formation, but did not exhibit increased susceptibility to hippocampal seizures (Duran, Gruart, Varea, et al. 2019). Additionally, studies utilizing genetic strategies to specifically target and ablate neuronal and astrocytic glycogen suggested that astrocytes are more linked to neurodegeneration and inflammation, while neurons contribute more to epileptic seizure susceptibility (Duran et al. 2021; Duran, Gruart, López-Ramos, et al. 2019). Thus, data from these genetic models suggest neuronal glycogen as a primary contributor to glycogen-derived memory perturbations but the potential adaptation occurring in transgenic mice should be kept in mind.

Thermoregulation, viscosity, and phase separation

Glycogen as a macromolecule has also been shown to contribute to diverse biological processes directly affecting intermolecular processes and homeostasis, including osmolarity, crowding, temperature regulation, and viscosity (N A Chebotareva 2007; Natalia A Chebotareva et al. 2004; Persson, Ambati, and Brandman 2020; Kerly 1929). Early studies determined that the viscosity of a liquid decreases in response to increasing temperatures (Longsworth 1954). This decreased viscosity at higher temperatures consequently leads to increased diffusion rates and molecular interaction frequency, affecting signal transduction and kinetics of complex assembly (Longsworth 1954; M. E. Johnson 2018; Stojanovski et al. 2017). Interestingly, glycogen increases heat stress tolerance, analogous to homeoviscous adaptation in the cell membrane (Wilson, Wang, and Roach 2002; Persson et al. 2020).

Work in the 1970s, pioneered by Sinensky and colleagues, elegantly demonstrated that cells adapt membrane lipid content in response to temperature changes and thereby allow the cell to control intermolecular interactions (Budin et al. 2018; Sinensky 1974). However, the mechanism for thermoregulation and viscosity in the cytosol remained unclear until it was recently demonstrated that glycogen and trehalose, a disaccharide of glucose, contribute to these processes (Persson, Ambati, and Brandman 2020). Through a combination of in vitro and in vivo techniques, it was demonstrated that yeast cells adjust glycogen and trehalose synthesis in response to temperature changes to prevent viscosity perturbations and maintain constant diffusion rates. Furthermore, data suggest that this viscoadaption impacts phase separation and the solubility of biomolecules in the cell. Indeed, it has been reported that glycogen undergoes liquid-liquid phase separation, creating large dynamic droplets in vitro that require the glycogen phosphatase laforin for activation of cell survival pathways (Liu et al. 2021). Collectively, glycogen appears to be crucial for thermoregulation, viscosity, and phase separation which may have specific impacts in the brain.

Glycogen in gene expression via histone acetylation

The nucleus is a highly specialized cellular organelle in eukaryotes that serves as a key cellular processing and DNA regulatory center. Histone acetylation in the nucleus impacts cell proliferation and growth, which are well-established hallmarks of aberrant cancer metabolism (Gujral et al. 2020; Unnikrishnan, Gafken, and Tsukiyama 2010; Z. Zhou et al. 2019; Zois, Favaro, and Harris 2014). Strikingly, increased glycogen in the nucleus is also a hallmark of some cancers, as first reported nearly 80 years ago by microscopy studies (Chipps and Duff 1942). However, the physiological function of nuclear glycogen remained largely unknown until recent work demonstrated that nuclear glycogen is linked to histone acetylation (Sun et al. 2019; Dong 2019; Donohue, Gentry, and Sun 2020). It was demonstrated that de novo nuclear glycogen synthesis and the metabolic fate of the glycogenolysis product, pyruvate, directly controls epigenetic regulation through histone acetylation. Furthermore, decreased levels of the E3 ubiquitin ligase malin impaired nuclear glycogenolysis by directly decreasing glycogen phosphorylase levels in the nucleus, causing nuclear glycogen accumulation (Sun et al. 2019; Donohue, Gentry, and Sun 2020). This novel molecular mechanism and role for nuclear glycogen has fascinating implications for brain function and is the subject of ongoing research.

Glycogen and protein glycosylation

Protein glycosylation is a highly regulated posttranslational modification that is critical for multiple cellular functions including protein folding, cell-cell interactions, cell adhesion, proliferation, and inflammation (Schwarz and Aebi 2011; Ng and Freeze 2018; Sun et al. 2021). Its importance is emphasized by the fact that >2% of the genome encodes genes involved in glycan metabolism, and >100 human diseases arise from glycosylation defects (Sun et al. 2021; Reily et al. 2019; Conroy, Hawkinson, et al. 2021; Ng and Freeze 2018). Glycogen and N-linked glycans share significant biochemical connections. Both are branched polysaccharides required for life which share the common substrate UDP-glucose. Additionally, glycogenolysis has been shown to preferably supply the hexosamine pathway for the biosynthesis of UDP-GlcNAc. Finally, brain glycogen has been shown to be a rich source of glucosamine, that can be directly converted to UDP-GlcNAc upon release from glycogen (Sun et al. 2021).

Asparagine- or N-linked glycosylation is the most common form of protein glycosylation (Eichler 2019). The most abundant extracellular monosaccharides and building blocks for glycosylation in the brain are glucose, glucosamine, fructose, mannose, fucose, and galactose. These monosaccharides enter cells by facilitated diffusion through the glucose transporter (GLUT) family of membrane transporters encoded in the SLC2 genes. Historically, the chemical heterogeneity of glycosylation has proven to be a barrier to study. However, emerging methods utilizing mass spectrometry and in situ enzymatic degradation of glycans known as MALDI-MSI provides improved sensitivity and spatial resolution (McDowell et al. 2021; Blaschke et al. 2020; Powers et al. 2013; Drake et al. 2018; Stanback et al. 2021; Hawkinson and Sun 2022).

Early studies suggested that muscle and liver glycogen contain small amounts of covalently bound glucosamine and provided an early connection to protein glycosylation. These studies utilized radiolabeled glucosamine and galactosamine to prove that glycogen synthase catalyzes the in vitro incorporation of UDP-glucosamine into glycogen (Tarentino and Maley 1976; Maley, McGarrahan, and DelGiacco 1966). Furthermore, Kirkman and Whelan demonstrated that glucosamine is a normal, albeit <1%, component of pig and rabbit liver glycogen, where it replaces α-1,4-linked glucose residues (Kirkman and Whelan 1986). More recently, a key pathway in brain carbohydrate metabolism that directly connects brain glycogen and N-linked protein glycosylation has been reported (Sun et al. 2021). By a multidisciplinary approach, it was demonstrated that N-linked protein glycosylation in the brain is critically and directly connected to glycogenolysis via the metabolite glucosamine. Strikingly, brain glycogen is composed of 25% glucosamine, while liver and muscle glycogen contain <1% glucosamine. Indeed, cells in the CNS leverage glycogen-associated glucosamine as a substantial substrate pool to produce UDP-acetylglucosamine (UDP-GlcNAc) and modulating glycogenolysis impacts N-linked glycan biosynthesis. Inhibiting glycogen utilization restricted UDP-GlcNAc pools and thereby total N-linked glycan levels in the brain. This mechanism was confirmed both in vitro and in vivo using two separate genetic mouse models lacking different glycogen metabolism enzymes. This study biochemically demonstrated that glucosamine is incorporated into glycogen by glycogen synthase and released from glycogen as glucosamine-1-phopshate by glycogen phosphorylase and glycogen debranching enzyme. The development of innovative multiplexed MALDI-MSI techniques allowed simultaneous visualization of glycogen and N-linked glycans in multiple disease models and revealed increases in glycogen and decreases in N-linked glycans (Conroy, Stanback, et al. 2021; Young et al. 2022). This technology provides an exciting new avenue to study spatial brain metabolism in the context of health and diseases.

It is well-established that both glycogen metabolism and N-linked protein glycosylation are crucial for synaptic modeling and memory formation. Perturbations in either pathway delay neuronal outgrowth and initiate neuro-inflammation that leads to cognitive delay, memory loss, and epilepsy (Dienel 2019a; Scott and Panin 2014; Gibbs, Anderson, and Hertz 2006). In fact, multiple disorders are either directly caused by or have perturbed glycosylation as a hallmark of the disease, including congenital disorders of glycosylation (CDG), some cancers, Alzheimer’s disease, Pompe disease, and Lafora disease (Gaunitz, Tjernberg, and Schedin-weiss 2020; Hawkinson et al. 2021; Sun et al. 2021; de-Souza-Ferreira, Ferreira, and de-Freitas-Junior 2023; Ng and Freeze 2018; Conroy, Stanback, et al. 2021; Wens et al. 2014). Strikingly, there are clinical studies reporting N-linked hypo-glycosylation in GSD patients; and aberrant glycogen aggregates in CDG patients (Papi et al. 2023; Sun et al. 2021; Tegtmeyer et al. 2014; McMahon and Frost 1996). Indeed, aortic stiffness in Pompe disease is caused by advanced glycan formation in aortic smooth muscle (Wens et al. 2014). Furthermore, multiple GSD and CDG patients share many neurological symptoms, including epilepsy and cognitive impairment (Piedade et al. 2022; Nitschke et al. 2018b; Korlimarla et al. 2019). While the novel route of hexosamine metabolism in the brain provides a direct metabolic connection between glycogen and N-glycans, key insights remain to be determined. Excitingly, these novel avenues of research have the potential to be therapeutically actionable for key glycogen-driven diseases of the CNS.

Glycogen in disease

GSDs are a group of inherited metabolic disorders and the first subtype was described in 1929 by von Gierke (Moses 2002; von Gierke 1929). This initial description of two patients with increased renal glycogen was later named GSD type I (Gierke’s disease), and more than 16 types of GSDs have subsequently been identified (Yeo, Moawad, and Grunewald 2023; Ellingwood and Cheng 2018). GSDs are a heterogeneous group of diseases driven by pathogenic glycogen with a wide range of symptoms, clinical onset, and life expectancy.

GSDs are caused by mutations in a wide range of genes encoding glycogen metabolism enzymes, leading to glycogen accumulation in most of the diseases (Figure 4A). However, there are at least two known GSDs that do not present with increased glycogen levels, including GSD type 0 (GSD0) and Glucose transporter 1 (Glut1) deficiency syndrome (G1D). These diseases are caused by mutations in genes encoding glycogen synthase and Glut1, preventing normal production of glycogen from glucose (Rajasekaran et al. 2023; Yeo, Moawad, and Grunewald 2023; Ellingwood and Cheng 2018). GSD0 and G1D mouse brains exhibit decreased glycogen and have similar symptoms to other GSDs (Rajasekaran et al. 2023; Chen and Weinstein 2016).

Figure 4. Neurological glycogen storage diseases and emerging treatments.

Overview of neurological glycogen storage diseases (n-GSDs) and key treatment strategies under development. A) Schematic representation of glycogen metabolism with affected enzymes in n-GSD and disease names. B) Current clinical status of treatment strategies for Lafora and Pompe disease and administration route for each treatment with green check signifying published and/or presented work and yellow check signifying ongoing investigation. ASOs targeting GYS1 have been tested via IT delivery for Lafora disease and siRNAs targeting GYS1 have been tested via IV delivery for Pompe disease. Preclinical work using AEFs has been performed for both diseases with multiple administration routes tested for Lafora disease. Created with BioRender.com. Abbreviations: ASO – anti-sense oligonucleotides, AEF – antibody-enzyme fusion, IT – Intrathecal, ICV – Intracerebroventricular, IV – Intravenous, siRNA – Small interfering ribonucleic acid, AAV – Adeno-associated virus, LD – Lafora disease, G1D – glucose transporter type 1 deficiency syndrome, Glc– glucose, UDP-Glc – uridine diphosphate glucose, P – phosphate, G1P – glucose-1-phosphate, G6P – glucose-6-phosphate, PGM1 – phosphoglucomutase 1, G6PASE - glucose-6-phosphatase, GSD0 – glycogen storage disease type 0, GP – glycogen phosphorylase, GYS – glycogen synthase, LA – laforin, GDE – glycogen debranching enzyme, GBE – glycogen branching enzyme, APBD – adult polyglucosan body disease.

Neurological glycogen storage diseases (n-GSDs)

GSDs can be classified by the primary affected organ. n-GSDs primarily affect the brain and CNS. Further, a number of disorders caused by inborn errors of metabolism present with a wide range of neurological symptoms, with disease severity ranging from mild delay in intellectual development to severe seizures, loss of cognition, and death. Additionally, it is increasingly recognized that a subset of inborn errors of metabolism share the pathogenic feature of aberrant glycogen-like aggregates and overlap with known n-GSDs. These diseases are particularly deadly and provide a unique window into metabolic perturbations in the nervous system.

Collectively n-GSDs, including Lafora disease (LD) (Gentry et al. 2018), adult polyglucosan body disease (APBD) (Zebhauser et al. 2022), Cori disease (Hobson-Webb et al. 2010), Glucose transporter type 1 deficiency syndrome (G1D) (Wang, Pascual, and De Vivo 1993), GSD0b, and late onset Pompe disease (PD) (Korlimarla et al. 2019), are neurogenetic disorders that are clinically characterized by aberrant glycogen most commonly manifested as pathogenic glycogen-like accumulations termed polyglucosan bodies (PGB) or dramatically decreased levels of glycogen in the CNS (Cafferty et al. 1991; Nitschke et al. 2018a). n-GSD patients can display progressive paraparesis, sensory deficit, peripheral neuropathy, myoclonic seizure and in the most severe cases, rapid neurodegeneration coupled with dementia and death can occur within 5 years of disease onset (Nitschke et al. 2018a; Zebhauser et al. 2022; Boulan-Predseil et al. 1995). Many n-GSD patients are adolescents and young adults and thus far there is no effective cure or targeted therapy. These diseases highlight the importance of brain glycogen, and research to date has provided a unique window into studying how neurological glycogen impacts the CNS in health and disease (Figure 4).

Glycogen as a therapeutic target

Given the severity and consequences of perturbed glycogen and glycan metabolism, several treatment strategies are being developed (Liu et al. 2021). Multiple strategies are being pursued that directly target enzymes involved in glycogen metabolism to prevent glycogen accumulation or promote degradation of aberrant glycogen aggregates. Further, a wide range of therapeutics platforms and techniques are being explored, including small molecule, biologic, and genetic approaches (Brewer et al. 2019; Ahonen et al. 2021; Gumusgoz et al. 2022; Markussen et al. 2021; Tang et al. 2020; Austin et al. 2019a). While some strategies remain in early preclinical drug development stages, a few treatments have already provided encouraging data in clinical trials (Figure 4B).

Small molecule inhibitors of glycogen synthase are one of the very promising therapies being translated into the clinical. These substrate reducing therapies (SRTs) inhibit GYS1 and decrease glycogen levels. The small molecule MZE001 from Maze Therapeutics recently completed Phase I clinical trials (NCT05249621). The selective inhibition of muscle glycogen synthase by MZE001 renders it a promising drug for Pompe disease and multiple other GSDs. MZE001 was granted Orphan Drug Designation by the FDA, and initial reports from a large Phase I are promising (Therapeutics 2022). However, any SRT must cross the blood brain barrier for therapeutic efficacy in n-GSD therapeutics.

Anti-sense oligonucleotides (ASOs) targeting GYS1 developed in collaboration with Ionis Pharmaceuticals have also displayed promising results in pre-clinical Lafora disease models (Ahonen et al. 2021). The GYS1 ASO reduced brain glycogen synthase levels in both young and old LD mouse models, halted aberrant glycogen formation, and reduced neuroinflammation. In this study, the earlier treatment in younger mice provided the most robust response. Similarly, the GYS1 siRNA ABX1100 from Aro Biotherapeutics received FDA orphan drug designation for treatment of Pompe disease (Beyer 2022).

Antibody Enzyme Fusions (AEFs) are another promising class of drugs under development for GSDs from Parasail, LLC (Brewer et al. 2019; Austin et al. 2019a). VAL-0417 is a promising AEF comprised of a cell-penetrating antibody fragment (Fab) fused with pancreatic α-amylase, allowing in vivo degradation of glycogen aggregates and normalization of both brain metabolism and N-linked glycosylation (Sun et al. 2021). Importantly, the utilized VAL-0417 antibody fragment efficiently delivers the drug by entering cells via the equilibrative nucleoside transporter 2 (ENT2, SLC29A2) and localizes to the cytosol and nucleus (Hansen et al. 2013; Weisbart et al. 2012; Hansen, Weisbart, and Nishimura 2005; Weisbart et al. 2015; Hansen et al. 2007). Since ENT2 is ubiquitously expressed on cell membranes in mice and humans, VAL-0417 has potential applications in multiple diseases (Crawford et al. 1998; Lu, Chen, and Klaassen 2004; Hansen et al. 2007). As with other modalities, key aspects of brain delivery remain to be resolved for clinical translation, with initial data demonstrating that both intrathecal and intracerebroventricular administration of VAL-0417 is effective (Austin et al. 2019b). AEF therapy may be particularly beneficial for late-stage symptomatic patients who require degradation of already formed aggregates (Brewer et al. 2019; Austin et al. 2019a; Sun et al. 2021).

Adeno-associated virus (AAV)-mediated gene therapy is a promising approach for treating both the systemic and the neurological symptoms of Pompe disease and potentially multiple other GSDs. Key components to gene therapy strategies include the choice of vector, promoter, and route of administration. To target the neurological symptoms of the disease, intrathecal (IT), intracisternal magna (ICM) (Hordeaux et al. 2017), and intracerebroventricular (ICV) (Lee et al. 2018) routes have been considered (Salabarria et al. 2020). One group treated adult GAA −/− mice with a single ICM injection of either AAV9 or AAVrh10-CAG-hGAA (Hordeaux et al. 2017). The authors demonstrated that the glycogen buildup was completely cleared four months post-ICM throughout the CNS, including brain cerebellum, brainstem, glia cells, myelin sheaths, and the proximal and distal spinal cords. These changes corresponded to improved motor function and reduction of hypertrophic cardiomyopathy one-year post-treatment (Hordeaux et al. 2017). Similarly, ICM administration of AAV5-GAA in Gaa^−/− mice decreased intraneuronal glycogen content and improved ventilation (Qiu 2012). Another group (Lee et al. 2018) treated Pompe mice using rAAVN-GAA driven by a synapsis promoter. Treated animals exhibited increased GAA activity, decreased glycogen content in the CNS but not in skeletal muscle or the liver, improvement in motor coordination and balance with reduced astrogliosis and improved myelination. Two ongoing clinical trials focus on liver-directed gene therapy approaches ((Smith et al. 2023) and NCT04093349)) by creating a liver depot for GAA production. The trial by SPARK Therapeutics using an AAVRh74 derived capsid is currently on clinical hold due to adverse events (NCT04093349). Recently, the preliminary results were reported in three subjects receiving AAV8-LSPhGAA at a dose of 1.6 x 10¹²vg/Kg (Smith et al. 2023). The authors demonstrated safety; however, the lack of glycogen lowering suggested that despite the presence of increased GAA activity in skeletal muscle, the efficacy of gene therapy at this dose was not sufficient to replace ERT. While there is some potential overlap between different modalities, there are also significant differences in terms of cellular targeting, suitability for use in the CNS, and potentially synergistic effects that make the parallel development and testing of different modalities particularly critical for n-GSD patients.

Future directions and concluding remarks

Recent research into glycogen biology has provided key insights into brain biology and human disease. These include insights into the localization and direct utilization of glycogen in the brain and key metabolic and non-metabolic roles for brain glycogen with direct relevance to human health. Further, recent general insights into glycogen biology in other organs have revealed novel ongoing research avenues to broaden and deepen our understanding of glycogen function in the brain.

Intriguingly, multiple groups have established that glycogen is a key molecular component in many aspects of brain health. Multiple laboratories have demonstrated that brain glycogen increases in response to disease or injury, including interictal stages of epilepsy (Seo et al. 2022), stroke (Cai et al. 2020), and anesthesia (A. M. Brown, Rich, and Ransom 2019). Additionally, brain glycogen likely impacts neuronal plasticity during aging and dementia (Riba et al. 2021; Mann et al. 1987; E. C. B. Johnson et al. 2020). Thus, glycogen may have direct causal effects on diseases of the brain well beyond what is currently considered.

Continued progress on n-GSDs makes it evident that there are likely connections with more common neurological diseases. As the role of glycogen metabolism is defined in these other diseases, it is possible that treatments for n-GSDs could be more broadly utilized. Thus, the overlap and potential common therapeutic approaches between these groups of diseases is of great interest.

Abbreviations:

AAV

Adeno-associated virus

AEF

Antibody-enzyme fusion

ASO

Antisense oligonucleotide

ANLS

Astrocyte-Neuron-Lactate-Shuttle

APBD

adult polyglucosan body disease

ATP

adenosine triphosphate

CDG

congenital disorders of glycosylation

CNS

Central nervous system

DAB

1,4-dideoxy-1,4-imino-d-arabinitol

DNA

Deoxyribonucleic acid

EM

Electron microscopy

G1D

Glucose transporter type 1 deficiency syndrome

G1P

glucose-1-phosphate

G6P

glucose-6-phosphate

G6PASE

glucose-6-phosphatase

GBE

glycogen branching enzyme

GCMS

Gas chromatography mass spectrometry

GDE

glycogen debranching enzyme

Glc

glucose

Gln

Glutamine

Glu

Glutamate

GLUT1

Glucose transporter type 1

GP

glycogen phosphorylase

GS

glycogen synthase

GSD

Glycogen storage disorder

GSG

Glucose Sparing by Glycogenolysis

HPAEC-PAD

high performance anion-exchange chromatography with pulsed amperometric detection

ICM

intracisternal magna

ICV

intracerebroventricular

IT

intrathecal

LA

laforin

LD

Lafora disease

MALDI-IMS

Matrix-assisted laser desorption/ionization imaging mass spectrometry

MRS

magnetic resonance spectroscopy

MRI

magnetic resonance imaging

MS

mass spectrometry

n-GSDs

neurological glycogen storage diseases

NMR

nuclear magnetic resonance

P

phosphate

PAS

periodic acid-Schiff

PD

Pompe disease

PGB

Polyglucosan

PGM1

phosphoglucomutase 1

SDS-PAGE

sodium dodecyl sulfate–polyacrylamide gel electrophoresis

siRNA

Small interfering ribonucleic acid

SRT

Substrate reducing therapy

TCA

trichloroacetic acid

UDP

Uridine diphosphate

UDP-Glc

uridine diphosphate glucose

UDP-GlcNac

Uridine diphosphate N-acetylglucosamine