Therapeutic potential of rescuing protein O-GlcNAcylation in tau-related pathologies

KEYWORDS:

• O-GlcNAcylation
• phosphorylation
• tau protein
• neurofibrillary tangles
• glucose metabolism

1. Introduction

O-GlcNAcylation is the non-canonical glycosylation of nucleocytoplasmic proteins with a single O-linked N-acetylglucosamine (O-GlcNAc) moiety. The dynamic cycling of O-GlcNAc on proteins is regulated by the concerted actions of two enzymes: the O-GlcNAc transferase (OGT) and a neutral β-hexosaminidase known as O-GlcNAcase (OGA). The O-GlcNAcylation of proteins occurs on serine and threonine residues, which can be commonly occupied by phosphate groups. Therefore, O-GlcNAcylation and phosphorylation are mutually related and by this cells can modulate a variety of signaling pathways and transcription factor in response to nutrients or stress [1]. O-GlcNAcylation is the product of nutrient flux through the hexosamine biosynthetic pathway (HBP), which integrates glucose, amino acid, fatty acid, and nucleotide metabolism to generate the donor substrate for O-GlcNAcylation, uridine diphosphate GlcNAc (UDP-GlcNAc). Approximately 2–5% of all glucose entering the cell is channeled into the HBP to generate UDP- GlcNAc. Glutamine–fructose-6-phosphate amide transferase (GFAT), the rate-limiting enzyme of the HBP that catalyzes the formation of glucosamine 6-phosphate, is shown to be subject to feedback inhibition by UDP-GlcNAc. Because O-GlcNAcylation depends on the availability of UDP-GlcNAc, and in turn intracellular UDP-GlcNAc level determines OGT activity, O-GlcNAcylation is considered a valuable intracellular sensor of glucose metabolism that can be directly regulated in a glucose-responsive manner.

2. Protein O-GlcNAcylation and tauopathies

Despite the O-GlcNAcylation process was discovered in the 80s by Torres and Hart [2], the O-GlcNAcylation-related research has developed rapidly only after the finding that human brain tau is O-GlcNAc modified representing a potential link with neurodegeneration [3]. Noticeably, since tau hyperphosphorylation represent the causative event leading to neurofibrillary tangles (NFT) formation, protein O-GlcNAcylation and its relationship with phosphorylation hold an essential significance in the investigation of the molecular mechanism leading to brain damage and to the development of the so-called tauopathies that include Alzheimer disease (AD), as well as, Down syndrome (DS) [4]. Initial studies, in cell culture and in ex vivo tissue slices, have supported that human tau O-GlcNAcylation can be reciprocal to tau phosphorylation suggesting its role in the build-up of NFT [5]. The analysis of human brain tissue established that hyperphosphorylated tau contains up to four-fold less of O-GlcNAc levels [6]. Further studies, by Gong and co-workers, demonstrated that short-term fasting of mice leads to decreased O-GlcNAc levels and increased tau phosphorylation, supporting the potential metabolic nature of such alterations [7]. Subsequently, it was reported that increasing protein phosphorylation levels, by the use of phosphatase 2A inhibitors or neuronal OGT gene deletion, lead to decreased O-GlcNAc levels [8]. Studies on human AD brain showed a marked decrease in tau-specific and global levels of O-GlcNAcylation, which are associated with the decline of (1) cerebral glucose metabolism, (2) brain insulin signaling and (3) neuronal GLUTs expression [5,6,9], suggesting that the O-GlcNAc modification of protein could represent the link between glucose hypometabolism and NFT formation. Furthermore, aberrant O-GlcNAc levels have been found on several proteins involved in glucose metabolism, strengthening this link [10,11]. Besides, aberrant levels of O-GlcNAc has been reported also on APP implying that this modification could be involved in the induction of its non-amyloidogenic processing by γ-secretase and to the formation of senile plaques [12]. Noteworthy, recent studies reported that the reduction of protein O- GlcNAcylation in AD and tau-driven pathologies do not occurs indiscriminately in all the brain proteome but it appears to vary between brain regions and a specific subcellular fraction of the neurons [13].

3. Therapeutic modulation of protein O-GlcNAcylation

Over the last decade several compounds, able to regulate OGT/OGA cycling, have been used to control the relative activities of OGT and OGA in both in vitro and in vivo models of neurodegenerative diseases [4]. Among these compounds, Thiamet G demonstrated, in several reports, to have a high selectivity for OGA together with a good oral bioavailability, achieving as expected the increase of brain O-GlcNAcylation levels. The first preclinical study using Thiamet G in drinking water aimed to treat hemizygous JNPL3 tau transgenic mice [14]. Authors found that the treatment increased tau O-GlcNAc, impeded the formation of tau aggregates, and decreased the number of NFT. However, OGA inhibition was shown to increase O-GlcNAcylation of human tau in a short period, but not after sustained OGA inhibition. In a different study, the injection of Thiamet-G into the lateral ventricle of the mouse brain led to the decrease of tau phosphorylation levels in several residues controlling its neurotoxicity [15]. Subsequently, Thiamet-G, was given to rTg4510 mouse model of tauopathy for 4 months obtaining results similar to those of JNPL3 mice. Short-term OGA inhibition reduced phosphorylation of soluble tau, but the long-term administration had no influence on its phosphorylation [16]. Borghgraef and colleagues used Thiamet-G in the Tau.P301L mouse model, demonstrating that the treatment was able to decrease the number of neurons showing tau pathology, decrease behavioral defects and reduce mice mortality [17]. Recently, Hastings and collaborators observed that the chronic inhibition of OGA with Thiamet-G reduces pathological tau in the brain and total tau in the cerebrospinal fluid of rTg4510 mice by directly increasing tau O- GlcNAcylation and thereby maintaining tau in the soluble and non- toxic form [18]. Up to now, treatment using Thiamet G provided valuable information to design future feasible therapeutic strategies in humans. Generally, all the data reported positive outcomes in restoring O-GlcNAcylation in mouse models of tau-related pathologies supporting the use of this compound in rescuing cognitive functions. However, few limitations on the use of Thiamet G or of OGA inhibitors in general should be taken under consideration before translating from mouse to human therapy. Few studies reported evident differences between the effects of short- and long-term administration of Thiamet G on tau phosphorylation, raising questions on its potential feasibility in the clinical setting [15,19]. It was proposed that these differential effects could be related to the fact that long-term treatment might lead to cells adaptation restoring the aberrant activity of kinases and phosphatases, or alternatively, that the sustained action of kinases could overcome the barrier of O-GlcNAc blockade recovering the aberrant phosphorylation levels. In this scenario, the length of the treatment seems to represent a critical point for the use of OGA inhibitors in patients, despite subsequent work tried to clarify this controversy [18]. Therefore, an in-depth analysis in this direction is needed to fully elucidate the effects of long-term treatments in the brain. A foster crucial restrain, in considering the use of OGA inhibitors as a valuable therapeutic strategy for tau-related pathologies, is represented by the variability of O-GlcNAc levels alteration in pathological brains, as recently observed. Data collected so far, on AD brains from human and mouse models of the disease, suggest that the cytoplasmic and mitochondrial proteome of the hippocampal region represents one of the primary target of intervention [6,9,10], indicating that a brain region-specific administration strategy focusing on these areas should be preferred to a systemic delivery (by oral or intraperitoneal injection distribution). Hence, a targeted administration could allow to maximize the effects, rescue cognition, and avoid the modification of unaffected brain regions or peripheral organs. Therefore, the experimenter, before beginning any pharmacological treatment using OGA inhibitors, should take into account the administration routes, as well as, the compound half-life and distribution at CNS.

The regulation of brain nutrients could represent an alternative approach to restore protein O-GlcNAcylation by increasing glucose uptake and/or induce the HPB pathway. The so-called western diet represents a well–known risk factor for the development of cognitive decline-pathologies related to overnutrition. As a nutrient sensor, O-GlcNAcylation is one of the key homeostatic mechanisms in response to diet changes. Hyperglycemia has been shown to raise cellular O-GlcNAcylation levels in various tissues in vivo [20]. However prolonged hyperglycaemia can lead, at CNS, to brain insulin resistance resulting in reduced glucose uptake [21]. Numerous groups have demonstrated the increase of protein O-GlcNAcylation in response to the modulation of extracellular glucose concentrations in vitro [22] . Data concerning the effect of caloric restriction (CR) on a mouse model of obesity-induced diabetes demonstrated, that CR was able to regulate brain dysmetabolism, increase the levels of hippocampal O-linked-N-acetylglucosamine and OGT, decrease the levels of phosphorylated tau and relieve learning impairment [23]. However, the relationship between nutrient availability and O-GlcNAcylation is not necessary a direct positive correlation since, beside the abundance of the UDP-GlcNAc substrate, it might depend on by the expression levels and activity of OGT and OGA and of their respective adaptor proteins. A recent work by Dai et al. [24] demonstrated that the CaMKIIα promoter-dependent neuronal KO of OGT in adult mice led to short-term overeating, body weight gain, and peripheral insulin resistance, accompanied by marked elevation of serum insulin and leptin levels and neuronal cell death, revealing an important role of neuronal OGT in controlling these aspects. Further, the neuronal OGT KO exacerbated obesity and insulin resistance induced by high-fat diet.

Intriguingly, the recovery of brain proteins from aberrant protein O-GlcNAcylation could be theoretically achieved using pharmacological approaches able to target and reduce insulin resistance. Several reports highlighted that the use of compounds designed to reduce brain insulin resistance (e.g. insulin, rapamycin) rescue glucose metabolism alterations, decrease oxidative stress and ameliorate cognitive decline [25]. Despite no data, concerning the involvement of O-GlcNAcylation have been published yet, we suggest that the recovery of O-GlcNAcylation levels in the brain could have a prominent role in mediating the beneficial outcomes observed using these compounds, our efforts are currently focused in this direction.

At final, recent studies demonstrate that physical exercise could represent a valid approach to modulate metabolism in mice and rats therefore re-establishing physiological protein O-GlcNAcylation in different tissues, comprising the brain [26].

4. Conclusions

Overall, despite studies of O-GlcNAcylation in neurodegeneration are still at an early stage, pharmacological interventions aimed to rescue O-GlcNAcylation in mice have produced satisfying results in tau-driven pathologies, supporting that targeting O-GlcNAcylation represent one of the most promising therapeutic opportunity. Indeed, this approach by 1) decreasing tau hyperphosphorylation and 2) recovering brain dys-metabolism, lead to reduced brain damage and improved cognitive performance. In this scenario, despite further studies to uncover the role of protein O-GlcNAcylation in the brain is ongoing, we are confident that the introduction of a clinically translatable therapy is imaginable in the next few years.

Declaration of interest

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This work was supported by the Ministry of Instruction, Universities and Research (MIUR) under the SIR program n° RBSI144MT by Sapienza University of Rome “Fondi di Ateneo Progetti Grandi” [# RG116154C9214D1A] to FDD.