Lithium as a Treatment for Alzheimer’s Disease: The Systems Pharmacology Perspective

Abstract

Systems pharmacology is a novel framework for drug research that models traditional and innovative pharmacological parameters and provides the overall efficacy and safety profile of a drug across body systems and complex, non-linear, molecular interactions. Lithium chloride, a pharmacological compound approved for the therapy of psychiatric disorders, represents a poorly explored compound for the treatment of Alzheimer’s disease (AD). Lithium has been shown to reduce downstream effects associated with the aberrant overactivation of certain molecular pathways, such as glycogen synthase kinase 3 subunit β (GSK3-β)-related pathways, involved in AD-related pathophysiology. It seems that overactivation and overexpression of GSK3-β lead to an impairment of long-term potentiation and amyloid-β induced neurotoxicity that can be normalized using lithium. Moreover, a growing body of evidence has demonstrated that lithium’s GSK3-β inhibitory effect prevents tau phosphorylation in mouse models of tauopathies. Clinical data have been inconclusive, partly due to methodological limitations. The lack of studies exploring the dynamics of protein misfolding in AD and investigating the specific tau-isoforms appearing prior to the accumulation of neurofibrillary tangles calls for new and optimized clinical trials. Advanced computer modeling based on a formal implementation of quantitative parameters and basic enzymatic insights into a mechanism-based model would present a good start to tackle these non-linear interactions. This innovative approach will pave the way for developing “molecularly” biomarker-guided targeted therapies, i.e., treatments specifically adapted (“tailored”) to the individual, consistently with the primary objectives and key conceptual points of precision medicine and precision pharmacology.

Keywords

Alzheimer’s disease GSK3 lithium neurotoxicity precision medicine post-translational modification systems pharmacology tau

INTRODUCTION

Precision pharmacology for precision neurology and treatment of neurodegenerative diseases

Complex chronic diseases such as cancer, immune system, and neurodegenerative diseases, including Alzheimer’s (AD), are characterized: by 1) a multifactorial nature, because of the coexistence of poly(epi)genetic and environmental susceptibility and 2) dysfunction of multiple networks, at both the molecular and large-scale biological level. It is well established that spatial-temporal evolution of AD pathophysiology has non-linear fashion. For instance, the alteration of evolutionarily highly conserved molecular and intracellular pathways as well as of several body systems cross-talks underlie the core features of the disease, including brain proteinopathies and neurodegeneration [1 –4]. Continuous failure of late stage clinical drug trials, following the traditional drug development paradigm in AD, demonstrates that a conceptual shift in drug Research and Development (R&D) pipelines is required to attain therapeutic breakthroughs [5, 6].

Systems-based biology, physiology, and pharmacology are expected to transform R&D strategies considerably, paving the way for developing biomarker-guided targeted therapies, i.e., pathway-based treatments, specifically adapted (tailored) to the individual biology and genetics [5 , 7–11]. In particular, systems pharmacology is a novel conceptual framework that models traditional pharmacological parameters, derived from pharmacodynamics and pharmacokinetics, operating under the experimental and computational framework of systems-medicine [5 , 7–11]. Systems pharmacology represents the ultimate breakthrough in biomarker-based approaches to drug R&D pipelines enabling accurate pathway-based targets identification. Indeed, multimodal exploratory and predictive outcomes, throughout proof-of-pharmacology as well as all relevant decision-making steps, can be integrated into a single prediction model [5 , 7–11]. The consequence of a standardized systems-wide approach to develop drugs will be represented by precision pharmacology-oriented R&D programs which will ultimately facilitate precise future medical strategies [5, 12]. In the coming future, we expect all individuals to be accurately clustered by degree of vulnerability to complex diseases and treated with preventive therapeutic strategies tailored to the individual biological make-up in line with the precision medicine paradigm shift [5, 12].

Through implementation of R&D pipelines with multimodal biomarkers for different context-of-use, the next-generation of trials investigating putative disease-modifying compounds will rely on robust proof of pharmacology: 1) alteration of pathophysiological mechanism(s) due to active target engagement; 2) modifications of disease pathophysiology toward physiological conditions (from molecular pathways to higher complexity network re-modeling); 3) long-term improvement of clinical outcomes upon target engagement and related pathomechanistic changes. This biomarker-based approach has already been adopted in certain advanced medical areas (i.e., Oncology and Clinical Immunology) in which several phenotypes of cancer differing in histology, cytologic topography, and clinical manifestations but sharing pathomechanistic alterations and molecular commonalities are treated with the same drug [8, 13–15 , 8, 13–15].

The advanced holistic systems-level approach of systems biology and systems pharmacology is assumed to facilitate the drug repositioning process—also known as the drug re-profiling or drug repurposing process—indicating that a drug with a recognized biological effect could be utilized to treat a disease for which it has not been registered [8, 13–15 , 8, 13–15].

In order to advance the development of precision medicine and precision pharmacology in AD, the international Alzheimer Precision Medicine Initiative (APMI) and its associated Cohort Program (APMI-CP) have been recently launched by our consortium and conceptually linked to the U.S. Precision Medicine Initiative (PMI) (available at https://obamawhitehouse.archives.gov/precision-medicine) and the U.S. “All of Us” Research Program (available at https://allofus.nih.gov/) [5, 16]. The APMI is an innovative research platform promoting the following objectives: 1) predicting the individual’s susceptibility to developing AD in the asymptomatic preclinical phase, 2) accomplishing an accurate AD detection and diagnosis targeting early pathophysiological alterations over the preclinical course, 3) optimizing patients’ assignment to treatments more likely to be beneficial with the lowest risk of adverse events [5, 16]. The APMI acts at the level of generating theoretical and explorative biomarker-based frameworks aimed at supporting substantial progress in the discovery, development, and validation of multimodal biomarkers (body fluids, functional and structural neuroimaging, electrophysiological measures) serving at several context of use [5, 16]. The APMI research activity operates at the systems-wide level consistent with the systems biology and systems physiology approaches for which complex diseases, such as AD, originate from failure of network adaptation and compensatory mechanisms leading to loss of resilience and decompensation [12 , 17–19] (see Glossary). The final step of this detrimental cascade is represented by loss of systems homeostasis which underlies the widespread pathophysiological dynamics such as synaptic failure and loss of neural plasticity, cerebral overaccumulation of amyloid-β (Aβ) and neurofibrillary pathology, neuroinflammation, and endothelial dysfunction [20]. Thus, the APMI research activity, in line with the precision pharmacology and precision medicine paradigms, aims at developing effective biomarker- and disease stage-guided targeted therapies, ultimately serving as a preventive and individualized medicine [12 , 17–19].

With regards to the relevance of drug repositioning programs inherent to the precision pharmacology-oriented pipelines, lithium chloride, a therapeutic agent approved for major psychiatric disorders [21], deserves to be further explored and developed also in the field of neurodegenerative diseases, including AD [22 –26]. Although lithium chloride has been shown to effectively modulate downstream effects of aberrant molecular pathways involved in AD, it has been poorly investigated so far [22–25 , 27]

Among the multi pathomechanistic alterations occurring in AD, the misfolding of tau, its aberrant post-translational modification(s), and ultimately the accumulation into tau-composed neurofibrillary tangles (NFT) and neuropil threads, play a primary role and have a predictable evolution in space and time [28] that can be targeted by lithium. The in-vivo and in-human tracking of tau post-translational modifications (PTMs) will be soon feasible through high-throughput technologies and breakthroughs currently utilized within in-silico sciences. Thus, we firmly think that focusing on lithium-induced modulation of tau pathology is, at the moment, a therapeutic avenue to pursue for accomplishing preventive treatments of AD.

The preliminary phase of investigations on novel therapeutic targets, including neuroinflammation, alongside the increasing debate on the validity of the etiological hypothesis, i.e., the amyloid cascade [29, 30], represents a strong rationale for a comprehensive focus on tau-mediated pathways. Moreover, it is well established that the magnitude of tau-related pathophysiological alterations correlates better with the rate of cognitive decline than Aβ-induced lesions. Therefore, in the recent years, several drug R&D pipelines investigating candidate disease-modifying compounds have introduced tau targets [31].

LITHIUM AND THE TREATMENT OF AD: FROM IN VITRO TO ANIMAL MODELS

Lithium is the oldest and most widely used treatment for major psychiatric disorders, particularly bipolar disorder and treatment-resistant depression [21], yet the exact mechanism underlying its therapeutic action remains unclear. Pharmacological and genetic studies suggest that lithium affects multiple steps in cellular signaling, usually increasing basal and decreasing stimulated activities [21]. The main pharmacodynamic targets include inositol monophosphatase (IMP) within the phosphatidylinositol signaling pathway, glycogen synthase kinase 3 (GSK3), Cyclic adenosine monophosphate (cAMP), response element binding protein (CREB), and Na+-K+-ATPase [21 , 27]. Due to its pleiotropic effects, lithium influences neuronal homeostasis involved in the activation of neuroprotective and neurotrophic responses, in the inhibition of oxidative stress and inflammatory cascades, and in the upregulation of mitochondrial function [26, 27 , 32–34]. Such findings provided the basis for investigating its beneficial effects in animal models of neurodegenerative disease, including AD.

The widespread misfolding and accumulation of tau and Aβ, respectively in NFT and neocortical Aβ plaques are established neuropathological hallmarks of AD. Robust translational evidence has demonstrated that loss of axonal sprouting and remodeling of dendritic spines are primary features of AD-related pathophysiology. It is widely recognized that loss of synaptic plasticity associated with progressive decline of synaptic function, manifested both as long-term potentiation (LTP) impairment and long-term depression (LTD) amplification, always precede neuronal death (i.e., neurodegeneration) [35, 36].

It has been hypothesized that conditions that favor LTD may promote the loss of synapses via a process termed synaptosis (a GSK3-dependent mechanism) [32, 37].

The modulation of GSK3 to slow down AD pathophysiological evolution

In mammals, there are two isoforms of GSK3, the subunits α and β [38]. GSK3-β is highly expressed in neuronal tissues, especially in some regions such as the hippocampus where neurogenesis (the process of generating functional neurons from precursors), neurotransmission, and synaptic plasticity have been shown to be clearly linked to GSK3 activity. In neurons, the GSK3 deficiency leads to embryonic lethality [39].

GSK3, initially identified as a key regulator of glycogen synthase, to date is well known being involved in several cell processes related to both pro-survival and apoptotic signaling as well as pathways leading to cell development/differentiation, endocytosis, microtubule stability, and lipid membrane dynamics [40, 41]. While over 100 substrates GSK3 can bind are known [42], it is assumed that many others downstream effectors of GSK3-β have not been completely described yet.

A fine negative regulation of GSK3 activity, based on evolutionarily highly conserved feedback loops, has also been demonstrated. An example is represented by insulin that increases the activity of PKB, a kinase that phosphorylates serine 21 residue in GSK3-α or serine 9 residue in GSK3-β inhibiting GSK3 activity [43].

GSK3 dysfunction has been linked to a broad range of pathomechanistic alterations occurring in different human diseases, including neurodegenerative diseases such as AD.

Regarding AD, it has been demonstrated that amyloid-β protein precursor (AβPP) is a substrate for GSK3 [44] and that GSK3 may play a role as regulator of both gamma-secretase activity and beta secretase 1 (BACE1) transcription [45 –47]. There is a feedback loop regulation of GSK3 through Aβ since the latter interferes with insulin or the Wnt/beta-catenin signaling, thus, increasing GSK3 activity [38 , 43]. Furthermore, presenilin 1 (PS1) may inactivate GSK3 through PI3K/PKB signaling [42].

As a proof of concept, a selective GSK3 inhibitor was able to prevent the acute effect of Aβ (500 nM) to inhibit LTP [48, 49]. Interestingly, this GSK3 inhibitor was still fully effective if applied after the Aβ peptide had inhibited LTP, demonstrating reversibility in the deleterious effects of acute Aβ. The ability of a different GSK3 inhibitor to prevent the effects of Aβ to inhibit LTP was also reported [50]. A link with the involvement of GSK3-β in chronic AD mouse models is the finding that either lithium or kenpaullone, two structurally distinct GSK3 inhibitors, reverse LTP deficit in slices from the Tg2576 APPswe mouse model of AD [51]. Thus, it seems that overactivation of GSK3-β is a key factor in the mediation of the pathological effect of Aβ on synaptic plasticity [52 –54]. Consistent with this idea, overexpression of GSK3-β leads to an impairment of LTP that can be recovered by treatment with lithium [49].

Certainly, GSK3-β plays a major role in the phosphorylation of the microtubule-binding protein tau [55] and its deregulated activity can lead to the formation of NFT (see below) [56]. A role for tau in pathological plasticity is also supported by the fact that Aβ inhibition of LTP is prevented in tau knockout mice [50]. Similarly, pharmacological blockade of Cdk-5, another primary kinase involved in the phosphorylation of tau, with either butyrolactone or roscovitine, prevents Aβ-mediated inhibition of LTP [57]. Likewise, microtubule-stabilizing agents preserve synaptic markers in response to lysosomal stress [58]. Therefore, tau could be a point of convergence of multiple kinases underlying pathological plasticity in AD, some of which are direct targets of lithium (see below).

It has been also described that GSK3 hyperactivity may promote pro-inflammatory pathways [41, 47]. On the flip side, a few pro-inflammatory cytokines may trigger or enhance GSK3 activity [59]. GSK3 slows down the neuronal production of fractalkine whose decrease may result in microglia activation [60, 61].

Besides an effect at the synaptic level, GSK3 inhibitors also improve cognitive deterioration in non-clinical models of AD [62, 63]. Genetic and pharmacological manipulations aimed at stimulating GSK3 activity are associated with cognitive impairment in the Morris water maze, as well as with hyperphosphorylation of tau, reactive astrocytosis and microgliosis, and neuronal death [64, 65]. Conversely, GSK3 transgene shutdown in symptomatic AD mice leads to normal phospho-tau levels, diminished neuronal death, and rescue of the cognitive deficit, thus confirming the potential of GSK3 inhibitors in the treatment of AD [66, 67]. Similarly, deletion of tau protein in GSK3-β overexpressing mice significantly reduces impairment in the Morris water maze task, further supporting a role for GSK3–tau interplay underlying cognitive impairment [68].

A growing body of evidence has demonstrated that lithium can prevent tau phosphorylation in mouse models of tauopathies [54 , 69–73].

Mice overexpressing GSK3 in neurons [65] or in glia cells [74, 75] could reverse their GSK3 induced phenotype by inhibition, for example by lithium, of GSK3 [66]. As a proof of concept, transgene shutdown in conditional GSK-3β overexpressing mice leads to normal p-tau levels, and more interesting to recover of the cognitive deficit observed in this animal model [65]. More recently, it was demonstrated that lithium could have a therapeutic benefit in neurodegenerative tauopathies by preventing tau hyperphosphorylation and reversing hyperphosphorylated somatodendritic tau in a double transgenic model, overexpressing GSK3-β in a conditional manner, using the Tet-off system and tau protein carrying a triple frontotemporal dementia and parkinsonism linked to chromosome 17 (FTDP-17) mutation. On the contrary, lithium was not able to change back tau already aggregated associated into NFTs [66].

Previous works have shown that chronic microdose lithium treatment reverses cognitive decline and histopathological alterations in a transgenic model of AD [76].

Microdose formulations have shown better bioavailability and safety profiles than conventional formulations. Indeed, microdose can be administered through routes of administration bypassing the gastrointestinal tract and liver metabolism, thus leading to high bioavailability. It is well acknowledged that lithium’s central and peripheral toxicity is dose-dependent which may substantially narrow long-term administration for chronic diseases [33, 77, 78].

A similar effect was also obtained when combining microdose lithium with pyrroloquinoline quinone, a strong antioxidant compound [33]. Notably, a recent study conducted in the 3xTg-AD mice revealed that chronic lithium treatment was associated with longer telomeres in the hippocampus and in the parietal cortex, hence providing new insights in the mechanism of action of lithium [32].

The beneficial effect of lithium microdose is not limited to hitting GSK3-β [77]. Indeed, Wilson and colleagues recently reported that an innovative experimental formulation of lithium microdose release is associated with the lowering of BACE1 gene expression and overall cerebral Aβ accumulation. They also found that lithium is associated with a neurogenic effect suggesting a potential neuroprotective property of the compound [77].

Further experimental evidence obtained from worms, flies, and mice revealed that lithium’s neuroprotective effect is mediated by GSK-3 dependent and independent pathways [26]. To summarize, GSK-3 inhibition might lead to 1) activation of nuclear factor (erythroid-derived 2)-like 2 (Nrf2) to prevent oxidative damage, 2) inhibition of signal transducer and activator of transcription 3 (STAT3) to counteract neuroinflammation, and 3) increase of Wnt-dependent gene transcription to enhance adult neurogenesis. On the other hand, GSK-3-independent effects include: 1) inhibition of toll-like receptor 4 (TLR4) to exert anti-inflammatory action, 2) activation of CREB-dependent transcription of brain-derived neurotrophic factor (BDNF) to promote neurotrophic actions including neurogenesis, cell survival and synaptic plasticity, and 3) inhibition of IMP leading to activation of autophagy and protein turnover [26].

Moreover, it has been reported that inhibitors of GSK3 could increase the degradation of acetylcholinesterase, thus enhancing the cholinergic firing [79, 80].

As homeostatic regulator, lithium may prevent several pathomechanistic features related to a broad spectrum of disorders. Indeed, several studies have clearly shown that dysregulation of GSK-3 and slowdown of neurogenic pathways play a pivotal role in pathogenesis of the so-called synucleopathies such as Huntington’s disease (HD) and Parkinson’s diseases (PD) [81 –83] as well as in neuropsychiatric condition as schizophrenia and bipolar disorders [34].

Genetic and pharmacological manipulation of these molecular cascades are expected to enhance multilevel adaptation and compensation (resilience), from subcellular to synaptic, until large-scale brain networks, ultimately prolonging brain homeostasis [12 , 84–86].

Drug repositioning of lithium for AD: In-human clinical trials

In the last ten years, human studies conducted on individuals suffering from AD have reported that lithium may improve cognitive performance as well as impact on exploratory outcome measures such as AD-related biomarkers. Regarding the cognitive outcomes, in a meta-analysis of randomized placebo-controlled trials testing lithium as a treatment for subjects with mild cognitive impairment (MCI) and AD dementia has suggested that lithium may have beneficial effects on cognitive performance [22]. Indeed, the main finding indicated that lithium significantly inhibited the progression of cognitive decline compared to placebo and with an effect size almost equal to the approved cholinergic drugs [22]. The rate of cognitive improvement in active treatment harm was significantly higher in AD dementia compared with MCI [22].

In addition, more recent studies [22, 25] using microdose lithium indicates promising results in MCI individuals.

For what concerns AD core biomarkers, lithium has shown mixed effects. In this context, one of the earliest clinical lithium-trial in AD was conducted by Astra Zeneca [23]. This underfunded and too short duration (only 10-weeks) cerebrospinal fluid (CSF) biomarker outcome trial with an insufficient number of recruited patients could not achieve sufficient power to demonstrate hypothesized biological effects on hyperphosphorylation using low lithium concentrations among the other restrictive methodological and trial design issues.

As a consequence of the results of this trial, the NIH decided to discontinue their planned large-scale lithium trial in AD [24]. Two years later, Fortaleza and colleagues carried out a longer protocol, with a 12-month treatment period, finding a decrease in CSF p-tau181 in the active harm compared with the placebo one (besides reporting lower rate of conversion from MCI to dementia) [25].

Therefore, future clinical trials with appropriate duration of treatment and including biological markers of AD pathophysiology seem mandatory to better explore the putative long-term cognitive and biological effects of that lithium may have in subjects with MCI.

In line with the objective of preventing the development of a dementia syndrome and exploring whether lithium may slow down AD-related molecular pathways, in 2017, Ariel Gildengers in collaboration with the National Institute on Aging (NIA) has started the LATTICE (Lithium As a Treatment to Prevent Impairment of Cognition in Elders) trial to evaluate the capacity of lithium in delaying brain and cognitive changes in elderly individuals with MCI. Of note, besides cognitive measures, the primary outcome measures will be performed by biomarker assessment analyzing GSK-3β activity in blood and CSF (https://clinicaltrials.gov/ct2/show/NCT03185208).

SYSTEMS PHARMACOLOGY STUDIES WITH GSK3-β

Multiple aspects of tau post-translational modifications

The tau molecule is phosphorylated at multiple sites in a number of tauopathies. There seems to be a dynamic molecular signature of tau phosphorylation through time and progression of a certain tauopathy or in clusters of individuals affected by certain tauopathies, such as AD.

Augustinack and colleagues found that this appears to start early and preclinical with phosphorylation at Thr231, then at Thr181, and later at clinical Ser199, this has to be further worked out including more phosphorylation sites and in subjects longitudinally through all stages to identify clusters of dynamic tau phosphorylation molecular non-linear sequences which in turn will present targets for therapies at different timepoint and different clusters of people) [87]. In AD, increased phosphorylation at Thr181 measured in CSF is a diagnostic marker. Many tau kinases have been identified and the list is growing, but GSK3-β is one of the earliest discovered and most important tau kinases.

An updated list of tau phosphorylation sites and their associated kinases can be found online (https://docs.google.com/spreadsheets/d/1hGYs1ZcupmTnbB7n6qs1r_WVTXHt1O7NBLyKBN7EOUQ/edit#gid=0 and from the Diane Hanger laboratory: https://www.kcl.ac.uk/ioppn/depts/bcn/Our-research/Neurodegeneration/diane-hanger-dementia-tauopathies/About-us.aspx).

Although hyperphosphorylation is a major PTM of tau protein in neurodegenerative diseases, it is by no means the only one. Additional modifications include acetylation, methylation, glycation, truncation, and ubiquitination, and the molecules often interact either directly by competing for the same site or indirectly by changing the conformation of the protein, thereby affecting the total outcome. Depicting accurately the overall dynamic leading to loss of proteostasis in neurodegenerative diseases is challenging.

For instance, phosphorylation by protein kinase A (PKA) primes GSK3-β-mediated phosphorylation of Thr181, Ser199, Ser202, Thr205, Thr217, Thr231, Ser396, Ser413, and Ser422, but inhibits phosphorylation at Thr212 and Ser404 [88]. Phosphorylation at Ser214 and Thr217 (both substrates for GSK3-β) promote subsequent phosphorylation at Ser210 and Ser214 [89] likely through casein kinase 1. This illustrates the many non-linear interactions during PTM, and the challenges encountered in influencing these processes for therapeutic purposes.

Functional consequences of GSK3-β-mediated changes

Hyperphosphorylation of tau tends to enhance dissociation from microtubules and, therefore, induce destabilization of the neuronal cytoskeleton. Phosphorylation by GSK-3β of tau at Ser131, Thr153, Ser202, and especially Ser262 leads to substantial dissociation from microtubules. However, some phosphorylation sites such as Thr50 (longest 2N-4R 441 AA numbering) that are weakly affected by GSK3-β, lead to microtubule stabilization [90]. Increased phosphorylation of serine/threonine tau sites at Thr181, Ser202, Thr212, Ser214, and Thr231 have been documented in the postmortem brain of patients with Creutzfeldt-Jakob disease. Similarly, hyperphosphorylation at Ser198, Ser199, and Ser202 is transiently increased after traumatic brain injury [91]. The contribution of GSK3-β to these changes is unclear, but these observations suggest that tau hyperphosphorylation might be a general reaction to neuronal cell stress and death. Besides being truncated, presumably toxic tau oligomers are extensively phosphorylated [92]. Whether this is due to reduced phosphatase activity or because phosphatases cannot reach the tau sites due to the conformational changes is so far unknown.

Tau phosphorylation at sites that are substrates for GSK3-β reduces proteolysis of misfolded tau proteins, leading to increased intracellular stress, and enhances mislocalization of toxic tau species in dendritic spines in a Fyn-independent manner [93]. While this tau phosphorylation leads to a variety of PTM and subsequent impact on cellular functions, it also allows, in principle, to link the individual PTM ‘signature’ of patients to a unique environment of tau pathology.

It is of interest to note that other modifications at tau sites can interfere with the phosphorylation kinetics or have different effects on their own. Such modifications include, for example, acetylation, glycation, methylation, ubiquitination, and truncation. Recent technological advances such as FlexiTau approach make it possible to quantitatively describe the modifications at individual sites [94].

While they are not used yet in clinical practice, they have in principle the capability to report on the individual tau signature. Another biomarker used in clinical studies is Tau-PET imaging. However, it is unclear at this point how modulating GSK3-β might affect this biomarker and one of the next challenges is to develop computational models that link molecular processes such as hyperphosphorylation to these imaging biomarkers (see the paragraph below).

Therefore, in the case of lithium and the related putative neuroprotective role, it is expected in the next years to explore traditional and novel outcome measures and endpoints assessed through multimodal biomarker-based study design. Novel emerging systems pharmacology frameworks will enable to accomplish a multilevel (from molecular to synaptic until large-scale networks) proof of pharmacology for a disease-modifying effect.

Developing a quantitative systems pharmacology platform

The sheer complexity of these non-linear dynamic interactions calls for a novel approach to develop a more predictive platform, especially when therapeutic interventions combining, for instance, GSK-3β inhibitors, such as lithium, with modulators of other PTM are considered. To complicate the situation, there is some evidence that GSK3-β activity is highest in early Braak stages and decreases as the disease progresses [95].

Tau hyperphosphorylation is a crucial step for Aβ-induced neuronal loss and tau-mediated neurotoxicity even precedes accumulation of NFT, i.e., the Braak “pre-tangle” stage [28, 96]. A converging pathway initiating tau aggregation and propagation, and the distinctive biological signatures specific for each brain tauopathy, including AD, remain elusive. A deeper understanding of molecular mechanisms such as the GSK3-mediated upstream regulation of tau homeostasis is required to determine the pathophysiological commonalities and differences within the broad spectrum of brain proteinopathies with neurodegeneration.

To date, all clinical trials investigating anti-tau agents with putative disease-modifying effects have shown their relevant efficacy and safety in animal models of AD but weak or no benefits in human individuals. This discrepancy can be partially attributed to inappropriate study methodology and design with improper biomarker selection for target engagement.

In general, the most important characteristic of an AD biomarker in clinical trials is the mechanistic relationship with the target on trial, followed by the correlation with clinical outcome. Efforts to generate the biomarker data and cost (for instance the high cost of tau-imaging) are also quite important.

The vast number of tau strains generated from aberrant PTM in humans indicate the need for a paradigm shift when developing clinical trials. Indeed, CSF concentrations of tau phosphorylated at Thr231 (p-Tau231) correlate with brain NFT accumulation rates significantly better than the CSF concentrations of p-Tau181, a surrogate marker of neurofibrillary tau pathology for AD diagnosis. We have previously reported that p-Tau231, among others hyperphosphorylated tau epitopes, may be generated by GSK-3 overactivity [97]. In addition, Augustinack and colleagues showed that p-Tau231 seems to appear as the earliest and most prominent phosphorylated tau strain detectable in the preclinical “pre-tangle” stage of AD-related neurofibrillary pathology [87].

Different molecular signatures related to tau hyperphosphorylation patterns occur at different times during the cellular and molecular phases of AD progression. Such pathophysiological sequences may display interindividual differences related to the (epi)genetic background and its interaction with environmental stressors. The lack of studies exploring such dynamics and investigating the specific tau-isoforms occurring even before the accumulation NFT indicates the need for a change in approach in drug R&D programs. Besides the established p-Tau181, a more comprehensive panel of tau PTM-related biomarkers should be investigated. With regard to the epidemic burden of AD, such a panel of blood biomarkers should be developed by conducting large-scale longitudinal studies in asymptomatic individuals.

Advanced computer modeling based on a formal implementation of quantitative parameters and basic enzymatic insights into a mechanism-based model is a good start to tackle these dynamic non-linear interactions. For example, the interaction between GSK3-β and PKA has been modeled [98], based on in vitro studies.

We have recently started MAPTA (Modeling Alliance for Systems Pharmacologies in Tauopathies), where the knowledge about tau biology and pathology that is available in the public domain will be modeled, so that observations from preclinical animal models or human induced pluripotent stem (hIPSC) cells can be translated into a more relevant human tau pathology situation (see Fig. 1). This will allow estimation of the indirect impact of specific kinase inhibitors on the general signature of the tau molecule with possible implications for biomarkers related to target engagement in novel therapeutics.

Fig.1

Validation of Quantitative Systems Pharmacology (QSP) by clinical trials with lithium. Phosphorylation as a post-translational modification (PTM, light brown) has been studied most extensively, but other modifications such as acetylation (yellow), O-GlcNAcylation (pink), methylation (red), truncation (light grey), and glycation (black) affect neuronal function. Neuronal functions affected by tau PTM changes include loss of binding to microtubules (MT) and reduced MT stability, direct and indirect effects on synaptic function, aggregation into oligomers and neurofibrillary tangles (NFT) and increased resistance against autophagy or proteasome-mediated degradation. Some of these post-translational modifications can possibly be detected in biological fluids using antibodies or most recently and quantitatively using mass spectrometry (FlexiTau), providing a “fingerprint” tau modification of the individual patient. QSP intends to link the difference in tau PTM ‘fingerprint’ at the individual patient level with treatment to these changes and ultimately to the firing dynamics of more extensive neuronal circuits that have been shown to be correlated to clinical outcomes. Comparison between the predicted and actual changes in clinical trials allow to validate the QSP model or to identify differences that can be addressed with an improved QSP model. ADAS-cog, Alzheimer’s Disease Assessment Scale-Cognitive subscale.

As an example, the MAPTA project simulates the kinetics of GSK3-β mediated tau phosphorylation at different tau sites based on transition rates between the enzymatic reaction states and constrained from biochemical studies using either antibody-mediated readouts [99] or MS-based approaches such as FlexiTau [94, 100]. Similarly rate constants for the enzymatic reaction of phosphatase (both PP2A and PP1)-mediated tau dephosphorylation can be derived from experimental condition [101]. Finally, the interaction between these two enzymatic reactions will be modeled in an ’in vivo cellular environment’ and various scenarios will be explored to reproduce experimental observations in clinical samples [81]. This allows in principle the exploration of subsequent changes in ATP, tau, phosphatase or GSK3-β levels on the phosphorylation state. Expansion of this modeling approach to different sites and modifications to the degree of specific site phosphorylation (such as Thr181) and comparison with experimentally measured CSF levels can in principle identify the activation states of both classes of enzymes. Such an approach, however, depends heavily upon the availability of experimental clinical data on post-translational tau site modification and might not be easily generalizable to other conditions. In addition, the number of possible combinations of tau-site PTMs can rapidly become prohibitive for calculations, so prioritization of the biological processes will be mandatory.

A compartmental model will be developed to relate levels of interstitial tau (tau secreted in the extracellular space) to CSF tau, mostly based on rodent preclinical studies. Furthermore, a computer model is being developed for the slow axonal transport of tau along the microtubule; the phosphorylation status of tau as determined by the competition between GSK3-β and phosphatase can be implemented both as a decrease in the binding rate of free tau (both monomeric and seed-competent) to microtubule, affecting the transport in neuronal projections as well as an effect on microtubule stability.

A major challenge in all clinical trials is determining the level of target exposure at the site of action. The combination of a Quantitative Systems Pharmacology (QSP) [7, 102] model describing the balance between kinase and phosphatase activity combined for instance with the FlexiTau mediated quantification of CSF tau in individual patient during treatment might generate a better estimate of the inhibition level of neuronal GSK3-β.

Using FLEXITau, differences in these PTM fingerprints after lithium treatment can be detected in human sample [94] at the individual patient level. The QSP platform can then predict (see Fig. 1) the changes in downstream effects on the different biological readouts, such as loss of binding to microtubules (MT) and reduced MT stability, direct and indirect effects on synaptic function, aggregation into oligomers and NFT and increased resistance against autophagy or proteasome-mediated degradation and ultimately their effect on neuronal circuit firing dynamics that are closely associated with clinical readouts such as Alzheimer’s Disease Assessment Scale-Cognitive subscale [7 , 103].

The QSP model will use these individual patient fingerprints in combination with comedications and genotypes to predict clinical outcomes which can be compared to the actual observed clinical change after lithium treatment. This would allow to validate the platform, better understand the clinical trial results or identify areas for improvement. Performing this operation for multiple clinical trials with compounds that affect tau post-translational modification allows to develop an increasingly better and predictive computer model that may ultimately pave the way to pathway-based therapies tailored to the individual’s biological profile, in line with the primary objectives of precision pharmacology [5 , 104].

In conclusion, systems pharmacology, through innovative and unique platforms and technologies, will enable researchers to decipher early AD-related pathophysiological signatures even in asymptomatic individuals, offering preventive and individualized strategies in line with precision pharmacology and precision medicine.

CONTRIBUTORS TO THE ALZHEIMER PRECISION MEDICINE INITIATIVE–WORKING GROUP (APMI–WG)

Mohammad AFSHAR (Paris), Lisi Flores AGUILAR (Montréal), Leyla AKMAN-ANDERSON (Sacramento), Joaquín ARENAS (Madrid), Jesus AVILA (Madrid), Claudio BABILONI (Rome), Filippo BALDACCI (Pisa), Richard BATRLA (Rotkreuz), Norbert BENDA (Bonn), Keith L. BLACK (Los Angeles), Arun L.W. BOKDE (Dublin), Ubaldo BONUCCELLI (Pisa), Karl BROICH (Bonn), Francesco CACCIOLA (Siena), Filippo CARACI (Catania), Juan CASTRILLO(Derio), Enrica CAVEDO (Paris), Roberto CERAVOLO (Pisa), Patrizia A. CHIESA (Paris), Jean-Christophe CORVOL (Paris), Augusto Claudio CUELLO (Montréal), Jeffrey L. CUMMINGS (Las Vegas), Herman DEPYPERE (Gent), Bruno DUBOIS (Paris), Andrea DUGGENTO (Rome), Enzo EMANUELE (Robbio), Valentina ESCOTT-PRICE (Cardiff), Howard FEDEROFF (Irvine), Maria Teresa FERRETTI (Zürich), Massimo FIANDACA (Irvine), Richard A. FRANK (Malvern), Francesco GARACI (Rome), Hugo GEERTS (Berwyn), Filippo S. GIORGI (Pisa), Edward J. GOETZL (San Francisco), Manuela GRAZIANI (Roma), Marion HABERKAMP (Bonn), Marie-Odile HABERT (Paris), Harald HAMPEL (Paris), Karl HERHOLZ (Manchester), Felix HERNANDEZ (Madrid), Dimitrios KAPOGIANNIS (Baltimore), Eric KARRAN (Cambridge), Steven J. KIDDLE (Cambridge), Seung H. KIM (Seoul), Yosef KORONYO (Los Angeles), Maya KORONYO-HAMAOUI (Los Angeles), Todd LANGEVIN (Minneapolis-Saint Paul), Stéphane LEHÉRICY (Paris), Alejandro LUCÍA (Madrid), Simone LISTA (Paris), Jean LORENCEAU (Paris), Dalila MANGO (Rome), Mark MAPSTONE (Irvine), Christian NERI (Paris), Robert NISTICÓ (Rome), Sid E. O’BRYANT (Fort Worth), Giovanni PALERMO (Pisa), George PERRY (San Antonio), Craig RITCHIE (Edinburgh), Simone ROSSI (Siena), Amira SAIDI (Rome), Emiliano SANTARNECCHI (Siena), Lon S. SCHNEIDER (Los Angeles), Olaf SPORNS (Bloomington), Nicola TOSCHI (Rome), Steven R. VERDOONER (Sacramento), Andrea VERGALLO (Paris), Nicolas VILLAIN (Paris), Lindsay A. WELIKOVITCH (Montréal), Janet WOODCOCK (Silver Spring), Erfan YOUNESI (Esch-sur-Alzette).

Authors’ disclosures available online (https://www.j-alz.com/manuscript-disclosures/19-0197r1).

Footnotes

ACKNOWLEDGMENTS

This research benefited from the support of the Program “PHOENIX” led by the Sorbonne University Foundation and sponsored by la Fondation pour la Recherche sur Alzheimer.

HH is supported by the AXA Research Fund, the “Fondation partenariale Sorbonne Université” and the “Fondation pour la Recherche sur Alzheimer”, Paris, France. Ce travail a bénéficié d’une aide de l’Etat “Investissements d’avenir” ANR-10-IAIHU-06. The research leading to these results has received funding from the program “Investissements d’avenir” ANR-10-IAIHU-06 (Agence Nationale de la Recherche-10-IA Agence Institut Hospitalo-Universitaire-6).

The MAPTA project is funded by Cohen Veterans Biosciences (CVB), a non-profit patient foundation and the authors appreciate the discussions with Drs Kristophe Diaz and Andreas Jeromin.

Appendix

Glossary: Evolving lexicon and terminology of the Alzheimer Precision Medicine Initiative (APMI) paradigm

Concept	Abbreviation	Definition
Precision Medicine	PM	Translational science paradigm related to both health and disease. PM is a biomarker-guided targeted medicine on systems-levels taking into account methodological advancements and discoveries of the comprehensive pathophysiological profiles of complex polygenic, multi-factorial neurodegenerative diseases (proteinopathies of the brain). It aims at optimizing the effectiveness of disease prevention and therapy, by considering (customized) an individual’s specific “biological make-up” (e.g., genetic, biochemical, phenotypic, lifestyle, and psychosocial characteristics) for targeted interventions through P4M implementation.
Pathway-based therapy		A treatment developed following the systematic analysis of specific genes, their functions, and the related interactomes underlying a specific complex disease. By using reliable exploratory strategies (i.e., GWAS, proteomics, and microarrays), within the systems pharmacology approach, therapies can realistically be developed according to the molecular mechanism and large-scale network alterations regardless the clinical manifestation.
Systems Biology	SB	Evolving hypothesis-free, exploratory, holistic (non-reductionistic), global, integrative, and interdisciplinary paradigm using advances in multimodal high-throughput technological platforms. SB enables the examination of multi-level networks from molecular pathways, to cellular signals until large scale networks and across body systems.
Systems Pharmacology	SP	Science of advancing knowledge about drug action at the molecular, cellular, tissue, organ, organism, and population levels” (available at http://www.aaps.org/Systems_Pharmacology/).
		Systems pharmacology aims at exploring and predicting the whole effect and safety profile of a drug across body systems through the acquisition and integration of multi-modal biomarkers and operating at both experimental and computational level. Systems pharmacology provides the accurate detection of a drug effect also computing the interindividual differences in terms of (epi)genetic background and interactomes.
Precision Pharmacology	PP	Conceptual paradigm operating under the System biology and Systems pharmacology approach. The mission of precision pharmacology is to discover and develop pathway-based therapies to target individuals’ pathophysiological mechanism(s) with the most proper drug (i.e., best efficacy and safety profile) for the single patient at any specific disease stage.
Homeostasis		A spontaneous tendency towards a condition of a dynamic equilibrium based on a continuous counterbalance between regulatory-defense mechanisms and disrupting stress-induced signals. Homeostasis is common to any biological system. Homeostatic signaling is hierarchically organized from subcellular to cellular level, across organs, and, finally, systems. Homeostasis is essential for protecting all core biosynthetic processes necessary to optimal functioning and survival.
Adaptation		Biological output arising from multi-level anti-stress response, generating advantageous morpho-functional alterations in cells and higher levels inside a system. Adaptation is essential for re-allocate bio-energetic resources to face stressors, prevent systems damage and finally promote survival
Compensation (resilience)		Protective process in which a morpho-functional alteration is counterbalanced by another morpho-functional alteration without any change in biological output, thus preserving system homeostasis. Compensatory mechanisms are hierarchically organized through systems levels and aim at preserving the homeostasis under pathophysiological conditions.
Decompensation (Failure)		Breakdown or overexpression of one or more compensatory mechanisms finally resulting in detrimental alterations. This, in turn, reflects a homeostatic imbalance at different levels of complexity in body systems and may trigger irreversible systems failure.

Modified from Hampel et al. [5 , 16].

References

Shine

, Breakspear

, Bell

, Ehgoetz Martens

, Shine

, Koyejo

, Sporns

, Poldrack

(2019) Human cognition involves the dynamic integration of neural activity and neuromodulatory systems. Nat Neurosci 22, 289–296.

Palop

, Mucke

(2016) Network abnormalities and interneuron dysfunction in Alzheimer disease. Nat Rev Neurosci 17, 777.

Henstridge

, Hyman

, Spires-Jones

(2019) Beyond the neuron-cellular interactions early in Alzheimer disease pathogenesis. Nat Rev Neurosci 20, 94–108.

Hoeijmakers

JHJ

(2009) DNA damage, aging, and cancer. N Engl J Med 361, 1475–1485.

Hampel

, Toschi

, Babiloni

, Baldacci

, Black

, Bokde

ALW

, Bun

, Cacciola

, Cavedo

, Chiesa

, Colliot

, Coman

C-M

, Dubois

, Duggento

, Durrleman

, Ferretti

M-T

, George

, Genthon

, Habert

M-O

, Herholz

, Koronyo

, Koronyo-Hamaoui

, Lamari

, Langevin

, Lehericy

, Lorenceau

, Neri

, Nistico

, Nyasse-Messene

, Ritchie

, Rossi

, Santarnecchi

, Sporns

, Verdooner

, Vergallo

, Villain

, Younesi

, Garaci

, Lista

(2018) Revolution of Alzheimer Precision Neurology. Passageway of systems biology and neurophysiology. J Alzheimers Dis 64, S47–S105.

Cummings

, Tong

, Ballard

(2019) Treatment combinations for Alzheimer’s disease: Current and future pharmacotherapy options. J Alzheimers Dis 67, 779–794.

Roberts

, Spiros

, Geerts

(2012) Simulations of symptomatic treatments for Alzheimer’s disease: Computational analysis of pathology and mechanisms of drug action. Alzheimers Res Ther 4, 50.

Harrold

, Ramanathan

, Mager

(2013) Network-based approaches in drug discovery and early development. Clin Pharmacol Ther 94, 651–658.

Richiardi

, Altmann

, Milazzo

A-C

, Chang

, Chakravarty

, Banaschewski

, Barker

, Bokde

ALW

, Bromberg

, Buchel

, Conrod

, Fauth-Buhler

, Flor

, Frouin

, Gallinat

, Garavan

, Gowland

, Heinz

, Lemaitre

, Mann

, Martinot

J-L

, Nees

, Paus

, Pausova

, Rietschel

, Robbins

, Smolka

, Spanagel

, Strohle

, Schumann

, Hawrylycz

, Poline

J-B

, Greicius

(2015) BRAIN NETWORKS. Correlated gene expression supports synchronous activity in brain networks. Science 348, 1241–1244.

10.

Lyu

, Wang

, Balius

, Singh

, Levit

, Moroz

, O’Meara

, Che

, Algaa

, Tolmachova

, Tolmachev

, Shoichet

, Roth

, Irwin

(2019) Ultra-large library docking for discovering new chemotypes. Nature 566, 224–229.

11.

Zhao

, Iyengar

(2012) Systems pharmacology: Network analysis to identify multiscale mechanisms of drug action. Annu Rev Pharmacol Toxicol 52, 505–521.

12.

Hampel

, Vergallo

, Aguilar

, Benda

, Broich

, Cuello

, Cummings

, Dubois

, Federoff

, Fiandaca

, Genthon

, Haberkamp

, Karran

, Mapstone

, Perry

, Schneider

, Welikovitch

, Woodcock

, Baldacci

, Lista

(2018) Precision pharmacology for Alzheimer’s disease. Pharmacol Res 130, 331–365.

13.

Drilon

, Laetsch

, Kummar

, DuBois

, Lassen

, Demetri

, Nathenson

, Doebele

, Farago

, Pappo

, Turpin

, Dowlati

, Brose

, Mascarenhas

, Federman

, Berlin

, El-Deiry

, Baik

, Deeken

, Boni

, Nagasubramanian

, Taylor

, Rudzinski

, Meric-Bernstam

, Sohal

DPS

, Ma

, Raez

, Hechtman

, Benayed

, Ladanyi

, Tuch

, Ebata

, Cruickshank

, Ku

, Cox

, Hawkins

, Hong

, Hyman

(2018) Efficacy of Larotrectinib in TRK fusion-positive cancers in adults and children. N Engl J Med 378, 731–739.

14.

Deng

, Nakamura

(2017) Cancer precision medicine: From cancer screening to drug selection and personalized immunotherapy. Trends Pharmacol Sci 38, 15–24.

15.

Power

, Berger

, Ginsburg

(2014) Genomics-enabled drug repositioning and repurposing: Insights from an IOM Roundtable activity. JAMA 311, 2063–2064.

16.

Hampel

, O’Bryant

, Durrleman

, Younesi

, Rojkova

, Escott-Price

, Corvol

J-C

, Broich

, Dubois

, Lista

(2017) A Precision Medicine Initiative for Alzheimer’s disease: The road ahead to biomarker-guided integrative disease modeling. Climacteric 20, 107–118.

17.

Perez-Nievas

, Stein

, Tai

H-C

, Dols-Icardo

, Scotton

, Barroeta-Espar

, Fernandez-Carballo

, de Munain

, Perez

, Marquie

, Serrano-Pozo

, Frosch

, Lowe

, Parisi

, Petersen

, Ikonomovic

, Lopez

, Klunk

, Hyman

, Gomez-Isla

(2013) Dissecting phenotypic traits linked to human resilience to Alzheimer’s pathology. Brain 136, 2510–2526.

18.

Scheller

, Schumacher

, Peter

, Lahr

, Wehrle

, Kaller

, Gaser

, Kloppel

(2018) Brain aging and APOE epsilon4 interact to reveal potential neuronal compensation in healthy older adults. Front Aging Neurosci 10, 74.

19.

Seeley

(2011) Divergent network connectivity changes in healthy APOE epsilon4 carriers: Disinhibition or compensation? Arch Neurol 68, 1107–1108.

20.

Hampel

, O’Bryant

, Molinuevo

, Zetterberg

, Masters

, Lista

, Kiddle

, Batrla

, Blennow

(2018) Blood-based biomarkers for Alzheimer disease: Mapping the road to the clinic. Nat Rev Neurol 14, 639–652.

21.

Alda

(2015) Lithium in the treatment of bipolar disorder: Pharmacology and pharmacogenetics. Mol Psychiatry 20, 661–670.

22.

Matsunaga

, Kishi

, Annas

, Basun

, Hampel

, Iwata

(2015) Lithium as a treatment for Alzheimer’s disease: A systematic review and meta-analysis. J Alzheimers Dis 48, 403–410.

23.

Hampel

, Ewers

, Burger

, Annas

, Mortberg

, Bogstedt

, Frolich

, Schroder

, Schonknecht

, Riepe

, Kraft

, Gasser

, Leyhe

, Moller

H-J

, Kurz

, Basun

(2009) Lithium trial in Alzheimer’s disease: A randomized, single-blind, placebo-controlled, multicenter 10-week study. J Clin Psychiatry 70, 922–931.

24.

Tariot

, Aisen

(2009) Can lithium or valproate untie tangles in Alzheimer’s disease? J Clin Psychiatry 70, 919–921.

25.

Forlenza

, Aprahamian

, de Paula

, Hajek

(2016) Lithium, a therapy for AD: Current evidence from clinical trials of neurodegenerative disorders. Curr Alzheimer Res 13, 879–886.

26.

Kerr

, Bjedov

, Sofola-Adesakin

(2018) Molecular mechanisms of lithium action: Switching the light on multiple targets for dementia using animal models. Front Mol Neurosci 11, 297.

27.

Freland

, Beaulieu

J-M

(2012) Inhibition of GSK3 by lithium, from single molecules to signaling networks. Front Mol Neurosci 5, 14.

28.

Braak

, Zetterberg

, Del Tredici

, Blennow

(2013) Intraneuronal tau aggregation precedes diffuse plaque deposition, but amyloid-beta changes occur before increases of tau in cerebrospinal fluid. Acta Neuropathol 126, 631–641.

29.

Panza

, Lozupone

, Logroscino

, Imbimbo

(2019) A critical appraisal of amyloid-beta-targeting therapies for Alzheimer disease. Nat Rev Neurol 15, 73–88.

30.

Behl

(2017) Amyloid in Alzheimer’s disease: Guilty beyond reasonable doubt? Trends Pharmacol Sci 38, 849–851.

31.

Congdon

, Sigurdsson

(2018) Tau-targeting therapies for Alzheimer disease. Nat Rev Neurol 14, 399–415.

32.

Cardillo G de

, De-Paula V de

, Ikenaga

, Costa

, Catanozi

, Schaeffer

, Gattaz

, Kerr

, Forlenza

(2018) Chronic lithium treatment increases telomere length in parietal cortex and hippocampus of triple-transgenic Alzheimer’s disease mice. J Alzheimers Dis 63, 93–101.

33.

Zhao

, Gong

, Liu

, Pan

, Sang

, Sun

, Yu

, Fang

, Zhao

, Fei

, Jin

, Zhong

, Xu

(2014) Beneficial synergistic effects of microdose lithium with pyrroloquinoline quinone in an Alzheimer’s disease mouse model. Neurobiol Aging 35, 2736–2745.

34.

Machado-Vieira

(2018) Lithium, stress, and resilience in bipolar disorder: Deciphering this key homeostatic synaptic plasticity regulator. J Affect Disord 233, 92–99.

35.

Nistico

, Pignatelli

, Piccinin

, Mercuri

, Collingridge

(2012) Targeting synaptic dysfunction in Alzheimer’s disease therapy. Mol Neurobiol 46, 572–587.

36.

Selkoe

(2002) Alzheimer’s disease is a synaptic failure. Science 298, 789–791.

37.

Nisticò

, Collingridge

(2012) The synaptic basis of Alzheimer’s disease. Eur J Neurodegener Dis 1, 21–33.

38.

Xie

, Helmerhorst

, Taddei

, Plewright

, Van Bronswijk

, Martins

(2002) Alzheimer’s beta-amyloid peptides compete for insulin binding to the insulin receptor. J Neurosci 22, RC221.

39.

Townsend

, Mehta

, Selkoe

(2007) Soluble Abeta inhibits specific signal transduction cascades common to the insulin receptor pathway. J Biol Chem 282, 33305–33312.

40.

Cohen

, Frame

(2001) The renaissance of GSK3. Nat Rev Mol Cell Biol 2, 769–776.

41.

Hooper

, Killick

, Lovestone

(2008) The GSK3 hypothesis of Alzheimer’s disease. J Neurochem 104, 1433–1439.

42.

Baki

, Shioi

, Wen

, Shao

, Schwarzman

, Gama-Sosa

, Neve

, Robakis

(2004) PS1 activates PI3K thus inhibiting GSK-3 activity and tau overphosphorylation: Effects of FAD mutations. EMBO J 23, 2586–2596.

43.

Magdesian

, Carvalho

MMVF

, Mendes

, Saraiva

, Juliano

, Garcia-Abreu

, Ferreira

(2008) Amyloid-beta binds to the extracellular cysteine-rich domain of Frizzled and inhibits Wnt/beta-catenin signaling. J Biol Chem 283, 9359–9368.

44.

Hernandez

, Lucas

, Avila

(2013) GSK3 and tau: Two convergence points in Alzheimer’s disease. J Alzheimers Dis 33 Suppl 1, S141–144.

45.

Kimura

, Yamashita

, Nakao

, Park

J-M

, Murayama

, Mizoroki

, Yoshiike

, Sahara

, Takashima

(2008) GSK-3beta is required for memory reconsolidation in adult brain. PLoS One 3, e3540.

46.

Bhat R

, Budd Haeberlein

, Avila

(2004) Glycogen synthase kinase 3: A drug target for CNS therapies. J Neurochem 89, 1313–1317.

47.

Jope

, Yuskaitis

, Beurel

(2007) Glycogen synthase kinase-3 (GSK3): Inflammation, diseases, and therapeutics. Neurochem Res 32, 577–595.

48.

, Whitcomb

, Olsen

, Kerrigan

, Lo

S-C

, Bru-Mercier

, Dickinson

, Scullion

, Sheng

, Collingridge

, Cho

(2011) Abeta(1-42) inhibition of LTP is mediated by a signaling pathway involving caspase-3, Akt1 and GSK-3beta. Nat Neurosci 14, 545–547.

49.

Hooper

, Markevich

, Plattner

, Killick

, Schofield

, Engel

, Hernandez

, Anderton

, Rosenblum

, Bliss

, Cooke

, Avila

, Lucas

, Giese

, Stephenson

, Lovestone

(2007) Glycogen synthase kinase-3 inhibition is integral to long-term potentiation. Eur J Neurosci 25, 81–86.

50.

Shipton

, Leitz

, Dworzak

, Acton

CEJ

, Tunbridge

, Denk

, Dawson

, Vitek

, Wade-Martins

, Paulsen

, Vargas-Caballero

(2011) Tau protein is required for amyloid beta-induced impairment of hippocampal long-term potentiation. J Neurosci 31, 1688–1692.

51.

, Hoeffer

, Capetillo-Zarate

, Yu

, Wong

, Lin

, Tampellini

, Klann

, Blitzer

, Gouras

(2010) Dysregulation of the mTOR pathway mediates impairment of synaptic plasticity in a mouse model of Alzheimer’s disease. PLoS One 5, e12845.

52.

Schaeffer

, Catanozi

, West

, Gattaz

(2017) Stereological investigation of the CA1 pyramidal cell layer in untreated and lithium-treated 3xTg-AD and wild-type mice. Ann Anat 209, 51–60.

53.

Zhang

, Heng

, Li

, Yang

, Zhang

, Du

, Doody

, Le

(2011) Long-term treatment with lithium alleviates memory deficits and reduces amyloid-beta production in an aged Alzheimer’s disease transgenic mouse model. J Alzheimers Dis 24, 739–749.

54.

Habib

, Sawmiller

, Li

, Xiang

, Rongo

, Tian

, Hou

, Zeng

, Smith

, Fan

, Giunta

, Mori

, Currier

, Shytle

, Tan

(2017) LISPRO mitigates beta-amyloid and associated pathologies in Alzheimer’s mice. Cell Death Dis 8, e2880.

55.

Plattner

, Angelo

, Giese

(2006) The roles of cyclin-dependent kinase 5 and glycogen synthase kinase 3 in tau hyperphosphorylation. J Biol Chem 281, 25457–25465.

56.

Ishiguro

, Shiratsuchi

, Sato

, Omori

, Arioka

, Kobayashi

, Uchida

, Imahori

(1993) Glycogen synthase kinase 3 beta is identical to tau protein kinase I generating several epitopes of paired helical filaments. FEBS Lett 325, 167–172.

57.

Wang

, Walsh

, Rowan

, Selkoe

, Anwyl

(2004) Block of long-term potentiation by naturally secreted and synthetic amyloid beta-peptide in hippocampal slices is mediated via activation of the kinases c-Jun N-terminal kinase, cyclin-dependent kinase 5, and p38 mitogen-activated protein kinase as well as metabotropic glutamate receptor type 5. J Neurosci 24, 3370–3378.

58.

Butler

, Bendiske

, Michaelis

, Karanian

, Bahr

(2007) Microtubule-stabilizing agent prevents protein accumulation-induced loss of synaptic markers. Eur J Pharmacol 562, 20–27.

59.

Wyss-Coray

, Rogers

(2012) Inflammation in Alzheimer disease-a brief review of the basic science and clinical literature. Cold Spring Harb Perspect Med 2, a006346.

60.

Fuster-Matanzo

, Jurado-Arjona

, Benvegnu

, Garcia

, Martin-Maestro

, Gomez-Sintes

, Hernandez

, Avila

(2017) Glycogen synthase kinase-3beta regulates fractalkine production by altering its trafficking from Golgi to plasma membrane: Implications for Alzheimer’s disease. Cell Mol Life Sci 74, 1153–1163.

61.

Arranz

, De Strooper

(2019) The role of astroglia in Alzheimer’s disease: Pathophysiology and clinical implications. Lancet Neurol 18, 406–414.

62.

King

, Pardo

, Cheng

, Downey

, Jope

, Beurel

(2014) Glycogen synthase kinase-3 inhibitors: Rescuers of cognitive impairments. Pharmacol Ther 141, 1–12.

63.

Martinez

, Gil

, Perez

(2011) Glycogen synthase kinase 3 inhibitors in the next horizon for Alzheimer’s disease treatment. Int J Alzheimers Dis 2011, 280502.

64.

Hernandez

, Borrell

, Guaza

, Avila

, Lucas

(2002) Spatial learning deficit in transgenic mice that conditionally over-express GSK-3beta in the brain but do not form tau filaments. J Neurochem 83, 1529–1533.

65.

Lucas

, Hernandez

, Gomez-Ramos

, Moran

, Hen

, Avila

(2001) Decreased nuclear beta-catenin, tau hyperphosphorylation and neurodegeneration in GSK-3beta conditional transgenic mice. EMBO J 20, 27–39.

66.

Engel

, Goni-Oliver

, Lucas

, Avila

, Hernandez

(2006) Chronic lithium administration to FTDP-17 tau and GSK-3beta overexpressing mice prevents tau hyperphosphorylation and neurofibrillary tangle formation, but pre-formed neurofibrillary tangles do not revert. J Neurochem 99, 1445–1455.

67.

Engel

, Hernandez

, Avila

, Lucas

(2006) Full reversal of Alzheimer’s disease-like phenotype in a mouse model with conditional overexpression of glycogen synthase kinase-3. J Neurosci 26, 5083–5090.

68.

Gomez de Barreda

, Perez

, Gomez Ramos

, de Cristobal

, Martin-Maestro

, Moran

, Dawson

, Vitek

, Lucas

, Hernandez

, Avila

(2010) Tau-knockout mice show reduced GSK3-induced hippocampal degeneration and learning deficits. Neurobiol Dis 37, 622–629.

69.

Nakashima

, Ishihara

, Suguimoto

, Yokota

, Oshima

, Kugo

, Terada

, Hamamura

, Trojanowski

, Lee

VM-Y

, Kuroda

(2005) Chronic lithium treatment decreases tau lesions by promoting ubiquitination in a mouse model of tauopathies. Acta Neuropathol 110, 547–556.

70.

Noble

, Planel

, Zehr

, Olm

, Meyerson

, Suleman

, Gaynor

, Wang

, LaFrancois

, Feinstein

, Burns

, Krishnamurthy

, Wen

, Bhat

, Lewis

, Dickson

, Duff

(2005) Inhibition of glycogen synthase kinase-3 by lithium correlates with reduced tauopathy and degeneration in vivo . Proc Natl Acad Sci U S A 102, 6990–6995.

71.

Perez

, Hernandez

, Lim

, Diaz-Nido

, Avila

(2003) Chronic lithium treatment decreases mutant tau protein aggregation in a transgenic mouse model. J Alzheimers Dis 5, 301–308.

72.

Sudduth

, Wilson

, Everhart

, Colton

, Wilcock

(2012) Lithium treatment of APPSwDI/NOS2-/- mice leads to reduced hyperphosphorylated tau, increased amyloid deposition and altered inflammatory phenotype. PLoS One 7, e31993.

73.

Lei

, Ayton

, Appukuttan

, Moon

, Duce

, Volitakis

, Cherny

, Wood

, Greenough

, Berger

, Pantelis

, McGorry

, Yung

, Finkelstein

, Bush

(2017) Lithium suppression of tau induces brain iron accumulation and neurodegeneration. Mol Psychiatry 22, 396–406.

74.

Jurado-Arjona

, Llorens-Martin

, Avila

, Hernandez

(2016) GSK3beta overexpression in dentate gyrus neural precursor cells expands the progenitor pool and enhances memory skills. J Biol Chem 291, 8199–8213.

75.

Jurado-Arjona

, Rodriguez-Matellan

, Avila

, Hernandez

(2019) GSK3beta overexpression driven by GFAP promoter improves rotarod performance. Brain Res 1712, 47–54.

76.

Nunes

, Schowe

, Monteiro-Silva

, Baraldi-Tornisielo

, Souza

SIG

, Balthazar

, Albuquerque

, Caetano

, Viel

, Buck

(2015) Chronic microdose lithium treatment prevented memory loss and neurohistopathological changes in a transgenic mouse model of Alzheimer’s disease. PLoS One 10, e0142267.

77.

Wilson

, Do Carmo

, Iulita

, Hall

, Ducatenzeiler

, Marks

, Allard

, Jia

, Windheim

, Cuello

(2017) BACE1 inhibition by microdose lithium formulation NP03 rescues memory loss and early stage amyloid neuropathology. Transl Psychiatry 7, e1190.

78.

Grunfeld

J-P

, Rossier

(2009) Lithium nephrotoxicity revisited. Nat Rev Nephrol 5, 270–276.

79.

Jing

, Zhang

J-Y

, Ouyang

, Wu

, Zhang

X-J

(2013) Lithium treatment induces proteasomal degradation of over-expressed acetylcholinesterase (AChE-S) and inhibit GSK3beta. Chem Biol Interact 203, 309–313.

80.

Hampel

, Mesulam

M-M

, Cuello

, Farlow

, Giacobini

, Grossberg

, Khachaturian

, Vergallo

, Cavedo

, Snyder

, Khachaturian

(2018) The cholinergic system in the pathophysiology and treatment of Alzheimer’s disease. Brain 141, 1917–1933.

81.

Credle

, George

, Wills

, Duka

, Shah

, Lee

Y-C

, Rodriguez

, Simkins

, Winter

, Moechars

, Steckler

, Goudreau

, Finkelstein

, Sidhu

(2015) GSK-3beta dysregulation contributes to Parkinson’s-like pathophysiology with associated region-specific phosphorylation and accumulation of tau and alpha-synuclein. Cell Death Differ 22, 838–851.

82.

Gassowska

, Czapski

, Pajak

, Cieslik

, Lenkiewicz

, Adamczyk

(2014) Extracellular alpha-synuclein leads to microtubule destabilization via GSK-3beta-dependent Tau phosphorylation in PC12 cells. PLoS One 9, e94259.

83.

Nubling

, Bader

, Levin

, Hildebrandt

, Kretzschmar

, Giese

(2012) Synergistic influence of phosphorylation and metal ions on tau oligomer formation and coaggregation with alpha-synuclein at the single molecule level. Mol Neurodegener 7, 35.

84.

Fontana

, Partridge

, Longo

(2010) Extending healthy life span–from yeast to humans. Science 328, 321–326.

85.

Childs

, Durik

, Baker

, van Deursen

(2015) Cellular senescence in aging and age-related disease: From mechanisms to therapy. Nat Med 21, 1424–1435.

86.

Hampel

, Lista

, Khachaturian

(2012) Development of biomarkers to chart all Alzheimer’s disease stages: The royal road to cutting the therapeutic Gordian Knot. Alzheimers Dement 8, 312–336.

87.

Augustinack

, Schneider

, Mandelkow

E-M

, Hyman

(2002) Specific tau phosphorylation sites correlate with severity of neuronal cytopathology in Alzheimer’s disease. Acta Neuropathol 103, 26–35.

88.

Liu

, Liang

, Shi

, Yin

, El-Akkad

, Grundke-Iqbal

, Iqbal

, Gong

C-X

(2006) PKA modulates GSK-3beta- and cdk5-catalyzed phosphorylation of tau in site- and kinase-specific manners. FEBS Lett 580, 6269–6274.

89.

Hanger

, Byers

, Wray

, Leung

K-Y

, Saxton

, Seereeram

, Reynolds

, Ward

, Anderton

(2007) Novel phosphorylation sites in tau from Alzheimer brain support a role for casein kinase 1 in disease pathogenesis. J Biol Chem 282, 23645–23654.

90.

Feijoo

, Campbell

, Jakes

, Goedert

, Cuenda

(2005) Evidence that phosphorylation of the microtubule-associated protein Tau by SAPK4/p38delta at Thr50 promotes microtubule assembly. J Cell Sci 118, 397–408.

91.

Rubenstein

, Chang

, Davies

, Wagner

, Robertson

, Wang

KKW

(2015) A novel, ultrasensitive assay for tau: Potential for assessing traumatic brain injury in tissues and biofluids. J Neurotrauma 32, 342–352.

92.

Zhou

, Shi

, Chu

, Hu

, Guan

, Gong

C-X

, Iqbal

, Liu

(2018) Relevance of phosphorylation and truncation of tau to the etiopathogenesis of Alzheimer’s disease. Front Aging Neurosci 10, 27.

93.

Xia

, Li

, Gotz

(2015) Pseudophosphorylation of tau at distinct epitopes or the presence of the P301L mutation targets the microtubule-associated protein tau to dendritic spines. Biochim Biophys Acta 1852, 913–924.

94.

Mair

, Muntel

, Tepper

, Tang

, Biernat

, Seeley

, Kosik

, Mandelkow

, Steen

(2016) FLEXITau: Quantifying post-translational modifications of tau protein in vitro and in human disease. Anal Chem 88, 3704–3714.

95.

Hampel

, Buerger

, Kohnken

, Teipel

, Zinkowski

, Moeller

, Rapoport

, Davies

(2001) Tracking of Alzheimer’s disease progression with cerebrospinal fluid tau protein phosphorylated at threonine 231. Ann Neurol 49, 545–546.

96.

Morris

, Knudsen

, Maeda

, Trinidad

, Ioanoviciu

, Burlingame

, Mucke

(2015) Tau post-translational modifications in wild-type and human amyloid precursor protein transgenic mice. Nat Neurosci 18, 1183–1189.

97.

Buerger

, Ewers

, Pirttila

, Zinkowski

, Alafuzoff

, Teipel

, DeBernardis

, Kerkman

, McCulloch

, Soininen

, Hampel

(2006) CSF phosphorylated tau protein correlates with neocortical neurofibrillary pathology in Alzheimer’s disease. Brain 129, 3035–3041.

98.

Stepanov

, Karelina

, Markevich

, Demin

, Nicholas

(2018) A mathematical model of multisite phosphorylation of tau protein. PLoS One 13, e0192519.

99.

Liu

, Zhang

, Li

, Fang

, Wang

, Deng

, Gong

, Grundke-Iqbal

, Iqbal

, Wang

(2004) Tau becomes a more favorable substrate for GSK-3 when it is prephosphorylated by PKA in rat brain. J Biol Chem 279, 50078–50088.

100.

Singh

, Winter

, Bilimoria

, Bonni

, Steen

(2012) FLEXIQinase, a mass spectrometry-based assay, to unveil multikinase mechanisms. Nat Methods 9, 504–508.

101.

Liu

, Grundke-Iqbal

, Iqbal

, Gong

C-X

(2005) Contributions of protein phosphatases PP1, PP2A, PP2B and PP5 to the regulation of tau phosphorylation. Eur J Neurosci 22, 1942–1950.

102.

Geerts

, Spiros

, Roberts

, Carr

(2013) Quantitative systems pharmacology as an extension of PK/PD modeling in CNS research and development. J Pharmacokinet Pharmacodyn 40, 257–265.

103.

Geerts

, Spiros

, Roberts

(2018) Impact of amyloid-beta changes on cognitive outcomes in Alzheimer’s disease: Analysis of clinical trials using a quantitative systems pharmacology model. Alzheimers Res Ther 10, 14.

104.

Hampel

, Frank

, Broich

, Teipel

, Katz

, Hardy

, Herholz

, Bokde

ALW

, Jessen

, Hoessler

, Sanhai

, Zetterberg

, Woodcock

, Blennow

(2010) Biomarkers for Alzheimer’s disease: Academic, industry and regulatory perspectives. Nat Rev Drug Discov 9, 560–574.