Targeting Impaired Nutrient Sensing via the Glycogen Synthase Kinase-3 Pathway With Therapeutic Compounds to Prevent or Treat Dementia: A Systematic Review

Background: Dementia is a global challenge with 10 million individuals being diagnosed every year. Currently, there are no established disease-modifying treatments for dementia. Impaired nutrient sensing has been implicated in the pathogenesis of dementia. Compounds that inhibit the glycogen synthase kinase-3 (GSK3) pathway have been investigated as a possible treatment to attenuate the progression of the disease, particularly the suppression of the hyper-phosphorylation process of the tau protein. Aims: Systematically summarizing compounds which have been tested to inhibit the GSK3 pathway to treat cognitive impairment and dementia. Methods: PubMed, Embase and Web of Science databases were searched from inception until 28 July 2021 for articles published in English. Interventional animal studies inhibiting the GSK3 pathway in Alzheimer’s disease (AD), Parkinson’s dementia, Lewy body dementia, vascular dementia, mild cognitive impairment (MCI) and normal cognitive ageing investigating the change in cognition as the outcome were included. The Systematic Review Centre for Laboratory animal Experimentation’s risk of bias tool for animal studies was applied. Results: Out of 4,154 articles, 29 described compounds inhibiting the GSK3 pathway. All studies were based on animal models of MCI, AD or normal cognitive ageing. Thirteen out of 21 natural compounds and five out of nine synthetic compounds tested in MCI and dementia animal models showed an overall positive effect on cognition. No articles reported human studies. The risk of bias was largely unclear. Conclusion: Novel therapeutics involved in the modulation of the GSK3 nutrient sensing pathway have the potential to improve cognitive function. Overall, there is a clear lack of translation from animal models to humans.


INTRODUCTION
The number of people aged 60 years and older is increasing globally, with a similar trend being seen in those affected by neurodegenerative diseases. In 2020, globally over 50 million people were living with dementia, of which 60-70% with Alzheimer's disease (AD). Dementia has multiple causes and is characterized by impaired cognitive function. The neurodegenerative process of dementia begins a long time before the presentation of clinical symptoms (Katsuno et al., 2018). Factors contributing to the neurodegenerative process include metabolic disorders such as hyperglycaemia (Gale et al., 2018). As such it has been hypothesized that targeting ageing pathways may be a viable therapeutic option for treating dementia. One such pathway is the nutrient sensing pathway such as glycogen synthase kinase-3 (GSK3) which has been linked to the onset of dementias (Efeyan et al., 2015).
Impaired nutrient sensing is defined as dysregulation in processing nutrients for mammalian cells (de Lucia et al., 2020). Hyperglycemia such as in diabetes can lead to inflammation and apoptosis associated with high GSK3 activity. Hyperlipidemia and obesity can also increase GSK3 activity (Liu and Yao, 2016). The regulation of GSK3, one of the main molecules involved in insulin signalling (Jolivalt et al., 2008), plays an important role in this process. GSK3 is a protein kinase and present in many processes in organisms such as cell signalling and cellular transport (Souder and Anderson, 2019). It has two isoforms: alpha and beta. Both take part in glycogen metabolism and the phosphorylation process of over hundred substrates (Beurel et al., 2015), which are important for memory formation (Takashima, 2012). Impaired GSK3, when phosphorylated, induces hyper-phosphorylation of the tau protein, and subsequently lead to the formation of neurofibrillary tangles (NFTs) (Sayas and Ávila, 2021). Impaired GSK3 also contributes to the formation of amyloid beta (Aβ) plaques through the amyloid precursor protein (APP) cleavage pathway (Giese, 2009). Higher levels of GSK3 are associated with neuronal loss in Huntington's disease and AD (L'Episcopo et al., 2016;Gale et al., 2018). Therefore, inhibition of GSK3 by therapeutic compounds could be a promising approach in managing dementia (Jope and Roh, 2006;Bhat et al., 2018).
This systematic review summarizes novel compounds inhibiting the GSK3 pathway in the context of preventing and treating cognitive impairment and dementia.

Eligibility Criteria
Articles included in this review met the following inclusion criteria: 1) study population: animals or humans; normal cognitive ageing or dementia (AD, VD, LBD, PD). Populations likely to have a faster pace of cognitive ageing such as those with type 2 diabetes mellitus, insulin-resistance and obesity, were also included. For animals, normal ageing was defined as a strain not at a greater propensity to develop dementia and not manipulated to mimic dementia. Dementia models were defined as strains at a greater propensity to develop dementia compared to normal ageing strains. 2) Interventional study design with comparators, including randomised controlled trial, quasirandomised controlled trials, and pre/post studies. 3) Intervention: compounds targeting cognition by inhibiting the GSK3 nutrient sensing pathway. 4) Outcome: cognitive function measured using standardized cognitive tests.
Articles were excluded if they met one of the following exclusion criteria: in vitro data only, conference abstract, review, editorial, or letter to the editor, or published in a language other than English.

Study Selection and Data Extraction
Two review authors (AM and SP) independently screened the titles and abstracts and subsequently the full text articles of potentially relevant studies against the inclusion and exclusion criteria. A third reviewer (ABM or LF) resolved any disagreements between the two review authors. All studies were divided into two groups: studies testing natural compounds (containing substances produced naturally by living organisms (Ouyang et al., 2014) and studies testing synthetic compounds.
The following variables were independently extracted for the included studies by two reviewers (AM and SP): author, year of publication, intervention, species/animal model, sample size (treatment group, control group), age, sex, duration of intervention, dose of therapeutic and type of model (dementia or normal ageing). For binary outcomes, the number of events, percentage of events or odds ratios with 95% confidence intervals were extracted. For continuous outcomes, the mean or median value, standard deviation, standard error, 95% confidence intervals or interquartile range, mean difference, and p-values were extracted.

Data Analysis
Interventions were considered preventative if the intervention was administered prior to the onset of dementia, with a delayed onset of dementia or reduced incidence rate after administration. Interventions were considered therapeutic if the intervention was administered after the dementia onset and subsequently slowed its progression or improved cognitive function.
An overall positive effect of the administered compound on cognitive performance was defined as positive finding on primary cognitive outcome, or >50% of the cognitive tests demonstrating a statistically significant improvement in the treatment group compared to the comparator group. A moderately-positive result was defined as ≥20% of the cognitive tests demonstrating a statistically significant improvement. A finding was considered negative where <20% of the cognitive outcomes were positive in the treatment group compared to the comparator group.
To investigate if the cognitive effect of compounds is dependent on study characteristics, studies were grouped based on study outcome (positive, moderately positive, negative), study population (MCI/dementia, normal ageing), the duration of intervention and the dosage of the compound.

Registered Clinical Trials in Humans
To identify ongoing and unpublished, completed human trials investigating GSK3 inhibitors, the ClinicalTrials.gov database (ClinicalTrials.gov, 2022) was screened for trials registered before 19 February 2022. Trials were included if GSK3 inhibition was mentioned, or compounds summarized in this review were utilized in combination with cognitive function, dementia, MCI, AD, VD, PD or LBD.

Risk of Bias
Two reviewers assessed the risk of bias (AM and SP). The Systematic Review Centre for Laboratory animal Experimentation's risk of bias tool (SYRCLE) was used for animal studies (Hooijmans et al., 2014), which includes eight sources of bias: sequence generation, baseline characteristics, allocation concealment, random housing, blinding of personnel, random outcome assessment, incomplete outcome data, and selective outcome reporting. Results were denoted green for low risk, yellow for unclear risk, and red for high risk.

Study Selection and Characteristics
The literature searches and selection process are illustrated in Figure 1. After exclusion of 4,154 duplicates, 3,517 articles were screened for title and abstracts of which 844 underwent full text screening. Thirty-two studies from 29 articles were included, investigating impaired nutrient sensing using compounds interfering with the GSK3 pathway on cognition in mice (19 studies), rats (12 studies) and zebrafish [one study (Koehler et al., 2019)] ( Table 1). The Wistar rat was used most often to study dementia (8 out of 32 studies), whereas the C57BL/6 mice were used to study normal ageing [two studies (Jiang et al., 2020;Zhou et al., 2020)]. The sample size per treatment group varied between six Lin et al., 2021) to 20 animals Ibrahim et al., 2020). The majority of used animal dementia models were transgenic, overexpressing or producing mutant products of human genes such as amyloid precursor protein (APP) [Jiang et al., 2020;Zhou et al., 2020;Lin et al., 2021) and tau ]. Dementia was also induced by administration of intracerebroventricular Aβ (Huang et al., 2019;Akhtar et al., 2020;Fan et al., 2020;Sun et al., 2020;Yan et al., 2021) or streptozotocin (Tang et al., 2018;Wang et al., 2018;Zhu et al., 2019;Akhtar et al., 2020). Cognitive tests used in the trials included Morris water maze, open field test, Barnes maze task, elevated plus maze, locomotor activity, Y-maze test, passive avoidance test, passive avoidance learning and novel object recognition test. Details of the tests are summarized in Supplementary Table 1. GSK3 inhibition was studied at different stages of the nutrient sensing pathway, which is illustrated in Figure 2.

Human Registered Trials
No trials were found mentioning the inhibition of the GSK3 pathway and cognitive outcomes. Based on the compounds being described in this review, three human trials were found testing curcumin to improve cognition. One active, not recruiting phase 2 trial investigates curcumin (with aerobic yoga exercise) on levels of blood biomarkers and cognitive function in individuals with MCI or subjective cognitive impairment (NCT01811381). A completed phase 2 trials investigate the effect of curcumin on physical function and cognition in older adults (NCT03085680), and another phase 2 trial the effect of curcumin on mental capacity and amyloid-beta blood concentrations in AD patients (NCT01001637). Table 2 shows the SYRCLE risk of bias ratings for animal studies. All studies showed an unclear risk of bias on allocation concealment, random housing, blinding of personnel and random outcome assessment. A low risk was found across the sequence generation, baseline characteristics, incomplete outcome data and selective outcome reporting. Eleven studies showed unclear risk in the sequence generation criterion Madhavadas and Subramanian, 2017;Tang et al., 2018;Wang et al., 2018;Ahmad Rather et al., 2019;Huang et al., 2019;Koehler et al., 2019;Liao et al., 2019;Akhtar et al., 2020;Thapak et al., 2020;Saleh et al., 2021). Unclear risk was mainly due to the lack of reported data.

DISCUSSION
Compounds inhibiting the GSK3 pathway may improve cognition in animals with cognitive impairment, but there is limited evidence for delaying cognitive decline or dementia onset in ageing animal models.
Frontiers in Aging | www.frontiersin.org July 2022 | Volume 3 | Article 898853 et al., Li et al., 2019;Yao et al., 2019;Akhtar et al., 2020;Ibrahim et al., 2020;Ren et al., 2021). The PI3K/Akt pathway is a key for regulating cell survival, growth and metabolism, which is important for cognitive function such as learning and memory (Shu et al., 2013). It plays an important role in apoptosis and autophagy in the nervous system, but also in reducing neuronal and nerve damage (Li et al., 2016). There is evidence that the insulin-PI3K/Akt pathway is reduced in AD, which makes compounds activating this pathway promising for the treatment of cognitive decline (Kong et al., 2013). Of the aforementioned compounds, osthole has been shown to inhibit the PI3K/Akt pathway via the activation of PTEN (Zhu et al., 2018). Brain-derived neurotrophic factor and diapocynin inhibits PI3K and AKT directly. Sodium orthovanadate inhibits PI3K and AKT downstream via the inhibition of PTP (Takada et al., 2004). The exact mechanism for sea cucumber cerebroside, and schisandra chinensis inhibition of this pathway remains uncertain. The compounds evodiamine, curcumin, cinnamon extract, bee pollen extract, Asiatic acid, seed of litchi chinensis fraction, SCR-1693, DL0410 and P.harmala activate Akt, which leads to GSK3 phosphorylation impairment (Madhavadas and Subramanian, 2017;SoukhakLari et al., 2018;Tang et al., 2018;Wang et al., 2018;Ahmad Rather et al., 2019;Liao et al., 2019;Bi et al., 2020;Zhou et al., 2020;Chou and Yang, 2021;Saleh et al., 2021). Akt is important in protecting injured neurons (Huang et al., 2017) and promoting neurite outgrowth (Park et al., 2012). In early stages of neurofibrillary pathology as well as in hippocampal neurons in type 2 diabetes animal models, the levels of phosphorylated Akt are downregulated (Griffin et al., 2005;Xiang et al., 2015). Asiatic acid reduces oxidant mediated neuronal apoptosis, which leads to the inhibition of GSK3β (Ahmad Rather et al., 2019).
Frontiers in Aging | www.frontiersin.org July 2022 | Volume 3 | Article 898853 (Koehler et al., 2019). 20(S)-protopanaxadiol and oleanolic acid inhibit GSK3β via Wnt/GSK3β/β-catenin pathway activation (Lin et al., 2021). The brain-derived neurotrophic factor activates protein phosphatase 2 (alpha isoform) (PP2A), which leads to a decrease in p-Tau. The result is a dephosphorylation of active GSK3β and an enhancement of inactive GSK3β Ser9 . Compound N. incisum extract modulates extracellular Aß production caused by unphosphorylated GSK3β. Specifically, it suppresses the expression of the cleaving enzyme beta-site APP cleaving enzyme 1 (BACE1), preventing the cleavage of GSK3 substrate APP. The result is a decrease in Aß production and subsequently the decrease of p-Tau and neuronal death (Jiang et al., 2020). Compound lychee seed extract reduces the deposit of Aß associated with insulin dysfunction and resistance (Tang et al., 2018). IRS-1 is activated by catechin, procyanidin A1 and procyanidin A2 derived from lychee seed extract. The result is the inhibition in the activity of GSK3β in the PI3K/Akt/GSK3 pathway (Xiong et al., 2020).
The GSK3 pathway is one of the multi-aetiological factors implicated in insulin resistance (Leng et al., 2010), which might be of importance as more than 80% of patients with AD have either type 2 diabetes mellitus or an impaired fasting glucose (Barbagallo, 2014). Disturbed glucose metabolism has also been shown to increase the risk of developing dementia by 50% (Biessels et al., 2014).
GSK-3β can phosphorylate tau protein at various sites in vitro and in cell culture models and the epitopes are consistent with those found to be hyperphosphorylated in AD brains. Consequently, activation of GSK-3β leads to aggregation of the tau protein (Lei et al., 2011). Activated GSK-3β is also involved in Aβ formation and accumulation in human AD brains because it modulates the cleavage of APP (Lauretti et al., 2020). These mechanisms could be potentially targeted by GSK3 inhibitors. There has consistently been a poor translation of successful therapeutics of pre-clinical animal dementia models to successful interventions in human dementia clinical trials (Franco and Cedazo-Minguez, 2014).Only two compounds included in this review, α-Lipoic acid and curcumin were studied in human clinical trials. Results from one of the trials concluded that α-Lipoic acid can slow cognitive decline (Shinto et al., 2013). Three trials concluded that the administration of curcumin can disaggregate the Aß deposit in the brain (Baum et al., 2008), prevent Aß aggregation (Begum et al., 2008) and improve memory in healthy adults (Small et al., 2018).
Compounds tested in normal ageing animal models did not show cognitive improvement (Jiang et al., 2020;Zhou et al., 2020). Most of the animal models rely on the utilization of genetic mutations associated with familial AD, while the most common form of AD is sporadic AD (LaFerla and Green, 2012). Furthermore, publication bias cannot be ruled out, particularly in animal studies which are unlikely to be registered and may be less likely to be published if results are negative (van der Naald et al., 2020). It was not possible to perform formal statistical analysis due to small sample sizes of studies examining a number of compounds and the conclusion was based on reported p-values. The significance of p-values is influenced by the sample size, which is often small in animal studies. The search Frontiers in Aging | www.frontiersin.org July 2022 | Volume 3 | Article 898853 8 included studies, which investigated the modulation of the GSK3 pathway. Therefore, compounds potentially inhibiting the GSK3 pathway without mentioning the pathway could have been missed. Finally, information on baseline characteristics, outcome assessment, blinding of personnel, housing, allocation concealment, and sequence generation, were often unclear. Future animal trials should consider following the SYRCLE guidelines (Hooijmans et al., 2014).

CONCLUSION
The results of this systematic review suggest that the investigated compounds can improve cognitive function in MCI and dementia animal models. Further studies are required to fully elucidate the potential of GSK3 inhibitors in MCI/dementia as well as initiation of clinical trials in humans with compounds inhibiting the GSK3 pathway.

AUTHOR CONTRIBUTIONS
ABM designed the article, AM, SK, LF and ABM analysed and interpreted the data, and AM and SK drafted the article and LF and ABM revised it critically for important intellectual content, and all authors approved the version to be published, and all authors agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

FUNDING
This work was funded by the University of Melbourne and the Medical Research Future Fund via the Melbourne Academic Centre for Health. The funder has no role in the design and conduct of the review, the analysis and interpretation of the data.