Modulation of serine/threonine phosphatases by melatonin: therapeutic approaches in neurodegenerative diseases

Abstract

Melatonin is an endogenous hormone produced by the pineal gland as well as many other tissues and organs. The natural decline in melatonin levels with ageing contributes significantly to the development of neurodegenerative disorders. Neurodegenerative diseases share common mechanisms of toxicity such as proteinopathy, mitochondrial dysfunction, metal dyshomeostasis, oxidative stress, neuroinflammation and an imbalance in the phosphorylation/dephosphorylation ratio. Several reports have proved the usefulness of melatonin in counteracting the events that lead to a neurodegenerative scenario. In this review, we have focused on the fact that melatonin could rectify the altered phosphorylation/dephosphorylation rate found in some neurodegenerative diseases by influencing the activity of phosphoprotein phosphatases. We analyse whether melatonin offers any protective activity towards these enzymes through a direct interaction.

Linked Articles

This article is part of a themed section on Recent Developments in Research of Melatonin and its Potential Therapeutic Applications. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v175.16/issuetoc

Abbreviations

AD

Alzheimer's disease

CaM

calmodulin

HDAC4

histone deacetylase 4

HMGB1

high‐mobility group protein B1

MCAO

middle cerebral artery occlusion

MLCP

myosin light chain phosphatase

NDD

neurodegenerative diseases

NFAT

nuclear factor of activated T‐cells

OA

okadaic acid

PIP

PP1‐interacting proteins

PME‐1

protein phosphatase methylesterase‐1

PP1

phosphoprotein phosphatase 1

PP2A

phosphoprotein phosphatase 2A

PP3 or PP2B

calcineurin

PP5

phosphoprotein phosphatase 5

PPP

phosphoprotein phosphatases

RONS

reactive oxygen and nitrogen species

Introduction

Molecular physiopathology of neurodegenerative diseases

Neurodegenerative diseases (NDD) are a growing concern, especially in an ageing population, causing a corresponding increase in the associated socio‐economic impact. These multifactorial diseases are characterized by the progressive loss of cognitive and functional skills as a consequence of neuronal cell death. Common NDD include Alzheimer's disease (AD), Parkinson's disease (PD), amyotrophic lateral sclerosis (ALS), and Huntington's disease. Most of these pathologies are often associated with the aggregation of misfolded proteins, such as http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=4865 (Aβ), http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=9275, α‐synuclein or http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=5199, mitochondrial impairment, metal dyshomeostasis, high concentrations of reactive oxygen and nitrogen species (RONS) causing oxidative damage, neuroinflammatory processes and/or an alteration in the fine balance between protein phosphorylation and dephosphorylation processes. However, the precise mechanisms that lead to neuronal dysfunction and death in these diseases are still poorly understood.

Current treatment strategies for NDD

The current lack of knowledge about the aetiology of NDD hinders the discovery of efficient drugs and is exacerbated by their multifactorial nature. AD is a typical example. Three decades since both the amyloid and Tau protein theories were proposed, no drugs based on any of these hypotheses have been approved, and the current therapies are still limited to three anticholinesterase drugs, that is, http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=6599, http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=6602 and http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=6693 (Anand et al., 2014), and the antagonist of http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=4268 http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=75, http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=4253 (Alam et al., 2017). These four drugs only attack the symptoms of the disease with limited success. The case of ALS is even more dramatic, because the only two drugs prescribed, http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2326 (Wokke, 1996) and edaravone (Dorst et al., 2018), can only slightly prolong a patient's life by a few months. Moreover, despite the positive therapeutic outcomes using drugs which target the dopaminergic neurotransmission for PD, these medicines are actually only prescribed to improve symptoms, which means that this disease still does not have a cure. Therefore, NDD urgently require new innovative pharmacological approaches to replace the current and mostly unsuccessful therapies. The results of recent studies have suggested that Ser/Thr phosphoprotein phosphatases (PPP) are an interesting alternative in the treatment of NDD.

Phosphatases: Tyr and Ser/Thr phosphatases

In eukaryotes, the principal mechanism of cellular regulation is the reversible phosphorylation of proteins on amino acids with hydroxyl groups (Shi, 2009). Tyrosine is phosphorylated/dephosphorylated by enzymes (http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=980) that are different from those acting on serine and threonine (Ser/Thr phosphatases). This is due to the formation of a phosphate ester at the Tyr phenyl ring that requires a reactivity that is different to that of the aliphatic alcohols present in both Ser and Thr residues (Olsen et al., 2006). On phosphorylation of the hydroxyl group, the strong negative charge of the phosphate group distorts the structure and directly affects protein functions, triggering conformational changes that alter the catalytic function, the affinity for ligands, the subcellular localization or the protein stability (Olsen et al., 2006). The phosphorylation/dephosphorylation processes must be tightly controlled because their deregulation is associated with a wide range of diseases such as cancer, diabetes, cardiovascular and neurodegenerative diseases (Zhang et al., 2013). Although there is a considerable amount of evidence that kinases and phosphatases are altered under pathological conditions, most of the research into the design of new drugs has been based on the inhibition of kinases (Cohen, 2002). This discrimination is in part due to the simplified view that phosphatases are non‐specific enzymes. Nevertheless, recent research has demonstrated the opposite and introduced phosphatases as promising therapeutic targets.

There are three types of phosphatases with different substrate specificities: Tyr phosphatases, Ser/Thr phosphatases and dual phosphatases. The Ser/Thr phosphatases are classified as PPP, metal‐dependent protein phosphatases and aspartate‐based phosphatases (Shi, 2009). The PPP family consists of seven subfamilies, phosphoprotein phosphatase 1 (PP1), phosphoprotein phosphatase 2A (PP2A), calcineurin (PP2B or PP3), phosphoprotein phosphatase 4, phosphoprotein phosphatase 5 (PP5), phosphoprotein phosphatase 6 and phosphoprotein phosphatase 7. The majority of these enzymes work as holoenzymes, formed by a highly conserved catalytic subunit and a variety of regulatory subunits which modulate the enzymic activity of the former, providing them with substrate specificity and subcellular localization (Shi, 2009). Despite being a superfamily of enzymes, PP1 and PP2A are responsible for about 90% of the total Ser/Thr phosphatase activity within cells.

All the PPP enzymes share high homology at the catalytic site, the amino acid sequence and a reaction mechanism based on a metal‐catalysed phosphate ester hydrolysis. Six amino acid residues, three His, two Asp and one Asn, at the active site coordinate two metal ions that can be Mg²⁺, Mn²⁺, Fe²⁺ or Zn²⁺, which mediate the phosphate transfer without forming an intermediate (Zhang et al., 2013). These metals bind a water molecule, activating it into a hydroxyl ion that produces a nucleophilic attack on the phosphate group of the protein substrate.

Therapeutic potential of melatonin in NDD

In recent years, http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=224 (N‐acetyl‐5‐methoxytryptamine) has emerged as a promising therapeutic agent against NDD. Its pleiotropic actions and ubiquitous presence in various extra‐pineal organs, for example, the immune system cells, the brain, the airway epithelium, bone marrow, gut, ovaries, testes, skin and possibly other tissues (Acuna‐Castroviejo et al., 2014), have been described. Its natural decline with age causes melatonin deficiency that may contribute to numerous dysfunctions and pathophysiological changes (Acuna‐Castroviejo et al., 2014). Several reports have shown the versatility of melatonin to counteract oxidative stress. The CNS is highly vulnerable and sensitive to oxidative stress because it consumes a substantial amount of oxygen and generates RONS. The potent antioxidant ability of melatonin responds to a dual action: on the one hand at high concentrations, melatonin acts as a direct free‐radical scavenger, and on the other at low concentrations, melatonin could increase the activity and/or induce the expression of antioxidant enzymes (Rodriguez et al., 2004). The main target of the RONS is the mitochondria, which possess high concentrations of melatonin (Venegas et al., 2012). Recently, it has been suggested that melatonin is not only synthesized in the mitochondrial matrix and released into the cytosol but also undergoes an ‘automitocrine’ pathway, an action mechanism within the mitochondria based on intracellular receptor‐ligand interactions, whereby melatonin prevents neurodegeneration (Suofu et al., 2017).

The physiological actions of melatonin in the brain are mediated by activating two distinct receptor types: the membrane receptors, http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=287 and http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=288&familyId=39&familyType=GPCR and the nuclear receptors (Cutando et al., 2011). MT₁ and MT₂ membrane receptors mediate their functions through a G‐protein‐coupled second messenger pathway and participate in the regulation of circadian and seasonal rhythms. On the other hand, nuclear receptor signalling appears to be mediated via the transcription factor RZR/ROR and the immunomodulatory and antitumour effects of melatonin depend on http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=88 nuclear signalling (Winczyk et al., 2002). In humans, both mRNA and protein for the MT₁ receptor have been detected in cerebral areas such as the hypothalamus, cerebellum, frontal cortex, nucleus accumbens, amygdala and hippocampus (Ng et al., 2017). The clinical significance of melatonin receptors has been documented in post‐mortem studies of AD patients, where a significant decrease in MT₂ receptors was found (Savaskan et al., 2005), as well as an increase in MT₁ receptor immunoreactivity. Indeed, melatonin exerts receptor‐mediated actions in memory formation and neurogenesis (Liu et al., 2016), which represents an important line of research into the improvement of brain health.

During the last decade, the importance of inflammation in age‐related diseases has attracted increasing interest. Neuronal over‐excitation and inflammatory responses are tightly associated, although they are probably initiated at different sites. Pro‐inflammatory mediators are found in neurons, astrocytes and microglia, and melatonin can counteract the pro‐inflammatory signals in all cell types involved (Hardeland et al., 2015), serving as a potent inhibitor of NF‐κB expression (Korkmaz et al., 2012).

The contribution of Aβ‐metal complexes that undergo redox cycling and alterations to both Cu²⁺ and Zn²⁺ distribution are important events that occur in some NDD as a consequence of an imbalance in brain metal concentrations (Hung et al., 2010). Exploring the chelating properties of melatonin, our research group recently suggested the possibility that melatonin might reduce metal‐induced neurotoxicity by generating a stable organometallic complex (Romero et al., 2014). To date, there have not been any reports of any significant side effects from melatonin, even at high doses (Andersen et al., 2016). In this context, the approach known as ‘two are better than one’ leads to a more effective or synergic combination of drugs. Therefore, the co‐administration of melatonin with other pharmaceutical agents studied for NDD may not only enhance the protective effect of melatonin but also minimize the potential side effects of the accompanying drug.

Different research groups have reported the modulatory actions of melatonin on some PPP, which represent an important milestone in the study of new therapeutic alternatives for NDD. In the following sections, we discuss and analyse how melatonin may regulate these enzymes and modulate the physiopathological pathways affected by phosphorylation/dephosphorylation processes.

Rationale of the review: potential interaction of melatonin with PPP

In this review, we have briefly summarized the role played by melatonin in the down‐regulation of pathological events that lead to neurodegenerative situations and why we consider it to be an emerging therapeutic tool. To substantiate this view, the exogenous administration of melatonin has been shown to rescue cell cultures and cerebral preparations from AD‐related toxic damage, reducing functional and cognitive impairment in animal models of neurodegeneration (Buendia et al., 2016). Moreover, melatonin could be considered as a biomarker because its endogenous reduction closely correlates with the progression of age‐related diseases like AD. Conversely, the rising interest in the PPP enzymes lies mainly in their role as regulators of neuronal death/survival processes, particularly the microtubule‐stabilizing Tau protein dephosphorylation. Liu et al. (2005) reported that in physiological conditions, the PPP enzymes PP1, PP2A, Cn and PP5 dephosphorylate Tau in relative ratios of about 11, 71, 10, and 7% respectively. The brains of AD patients exhibited a substantial drop (about 50%) in Tau dephosphorylation, especially in the cortex and hippocampus regions (Gong and Iqbal, 2008), where PP2A is the enzyme most susceptible to decrease, although reductions of PP1 and PP5 enzymic activities were not marginal (Liu et al., 2005). Collecting these data, one might wonder whether the neuroprotective property of melatonin could reverse the altered phosphorylation rate observed in some NDD by modulating the activity of the PPP enzymes. This hypothesis would go beyond recognizing the wide‐spectrum antioxidant effects of melatonin, given that it is widely accepted that the control of the cellular redox state favours the correct enzymic activity of the PPP enzymes. Consequently, the rationale of this review is to analyse whether melatonin confers any protective activity towards the PPP enzymes, or at least towards those closely implicated in the Tau dephosphorylation, by mechanisms other than the conventional radical scavenging. This theory is based on the recent discovery that structurally related gramine derivatives have been shown to protect PPP enzyme activity against inhibition by the selective PP2A inhibitor http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=5349 (OA). The computational studies carried out by Lajarin‐Cuesta et al. (2016) lead to the hypothesis that these ligands would bind to PP2A near the catalytic site. We also studied the PP2A‐melatonin complex by computational docking and found that melatonin efficiently docks into several binding sites of the PP2A subunit C. The highest energy scores were reached when docking at a site located in a region close to the PP2A structural subunit A. This finding is consistent with the activity of melatonin in the formation of the trimer ABC (Koh, 2012). The strongest intermolecular bonds were generated with the amino acids Arg 49 and 110, Glu 47 and 109, Val 48 of subunit C and Asp 572 of subunit A (Figure 1). The arginines interact with the indole of melatonin in a pocket through weak Van der Walls interactions and H‐bonds, formed with the π cloud of melatonin and a strong H‐bond with the methoxy group of melatonin. Other H‐bonds, such as those formed by glutamates and the amide group of melatonin, help to insert melatonin into this binding pocket (Figure 1B). These interactions suggest that the effect of melatonin on the enzymic activity of Ser/Thr phosphatases could be ascribed at least in part to a direct interaction with the PPP enzymes.

Figure 1.

Docking of the melatonin−PP2A complex. (A) Pose of melatonin (yellow) in the PP2Aa‐c intersecting binding site (main residues of PP2Ac in green and main residues of PP2Aa in blue). (B) Principal hydrogen bonds between melatonin and PP2A. (C) Van der Waals contacts (Arg 49 and Arg 110) and hydrophobic interactions (Val 48) of melatonin.

Effect of melatonin on PP1

PP1 is expressed ubiquitously in the organism as the Ser/Thr phosphatase enzyme, with a wide range of biological substrates. It plays an essential role in many cellular mechanisms, such as gene transcription, protein expression, the cell cycle, the metabolism of glycogen and lipids and muscle contraction (Ceulemans and Bollen, 2004). In the CNS, PP1 participates in the maintenance of the synaptic plasticity (Ceulemans and Bollen, 2004). Holoenzymatic PP1 consists of PP1c with a catalytic function and the so‐called regulatory PP1‐interacting proteins (PIP), which show 50 possible polymorphisms (Garcia et al., 2004). Such interactions depend on the presence of a binding motif in PIP, the RVxF, which behaves as an anchor interacting with a hydrophobic groove that is only present in PP1c, far from its catalytic site. PP1c can have four different isoforms, PP1α, PP1β/δ, PP1γ1 and PP1γ2. The PIPs provide substrate specificity to PP1 as well as subcellular location, and PP1 has the ability to dephosphorylate PIP (Garcia et al., 2004). When the PIP/PP1 assembly is inhibited, for instance, by RVxF motif mimetics, PP1c exhibits a broad phosphatase activity. Peripherally, the myosin phosphatase‐targeting subunit 1 (MYPT1/PPP1R12A) that is widely expressed in smooth muscle binds PP1cβ to form the so‐called myosin light chain phosphatase (MLCP). This superstructure regulates muscle contraction (Xia et al., 2005), and its inhibition improved myosin motor activity and the stability of the filaments. In this context, Gomez‐Pinilla et al. (2008) described that melatonin restored contractility in an animal model of urinary bladder malfunction through a mechanism that involves MLCP inhibition. Endogenously, PP1 is inhibited by several peptides such as the dopamine and cAMP‐regulated neuronal phosphoprotein, the heat stable protein inhibitor‐1 (INH‐1 or I1^PP1) and the inhibitor‐2 (INH‐2 or I2^PP1).

Montilla‐Lopez et al. (2002) were the first to investigate the potential relationship between PPP enzymes and melatonin in neurodegeneration. They demonstrated that melatonin counteracted the oxidative stress elicited by the PP1/PP2A selective inhibitor OA in N1E‐115 neuroblastoma cells. Low levels of OA (50 nM) increased, approximately fivefold, the generation of lipid peroxidation products reducing the activity of the antioxidant enzymes catalase, GSH transferase and GSH reductase by 2.5‐fold or more, which was more effective than H₂O₂ (0.2 mM). Melatonin, which did not have any effect on the basal (OA‐untreated) group, prevented the damaging effects of OA, and the combination of OA and H₂O₂, when pre‐incubated at a concentration of 10⁻⁵ M, with N1E‐115 neuroblastoma cells. The neuroprotective profile of melatonin in OA‐treated N1E‐115 cells was shown to be MT1 and MT2‐independent of MT₁ and MT₂ receptors, as the selective antagonist http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=1363 did not inhibit the oxidative events, that is, the lipid peroxidation products or the down‐regulation of oxidative enzymes, triggered by OA (Montilla et al., 2003). Additionally, pre‐incubation of luzindole with melatonin did not affect the neuroprotective action of the latter in the OA‐exposed N1E‐115 cells. Similar results were observed in in vivo where, after i.c.v. injection of OA (200 ng·kg⁻¹) in rats (Tunez et al., 2003), 4.5 mg·kg⁻¹ of melatonin prevented inhibition of GSH‐peroxidase, reductase and transferase enzymes, as well as http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=6737 overconsumption. Although OA selectively inhibited PP1 and PP2A, none of these reports measured the phosphatase activity or the phosphorylation rate of the PPP substrates. By contrast, the experimental protocols were aimed at monitoring fluctuations in the cell redox state. Hence, any possible direct effect of melatonin on PPP remains unproven. Experiments carried out by other groups have revealed a closer interaction. Li et al. (2004) used calyculin A as pharmacological tool to inhibit the enzymatic activities of PP1 and PP2A in SH‐SY5Y neuroblastoma cells. Melatonin protected against the loss of cell viability due to calyculin A. In addition, Western blot analyses revealed that it markedly attenuated the hyperphosphorylation rate of neurofilaments at 50 μM. Neurofilaments are cytoskeletal proteins that form a complex network which play an essential role in axonal growth, as well as in anchoring cytoplasmic constituents to neurons (Kevenaar and Hoogenraad, 2015). The plasticity of these neurofilaments depends on phosphorylation/dephosphorylation processes, but their arrest in a hyperphosphorylated state leads to neurodegeneration. This has been demonstrated in AD brains, which showed an accumulation of hyperphosphorylated neurofilaments (Wang et al., 2001). This is due to an imbalance in the enzymic activities of kinases and phosphatases, in particular PP1 and PP2A, which are the most sensitive and is similar to that described for Tau‐pathies (Veeranna et al., 2011). Furthermore, hyperphosphorylated neurofilaments co‐localize with neurons and increase the probability of the formation of neurofibrillary tangles (NFT) (Vickers et al., 1990). Neurofilaments are made up of three subunits, namely, a heavy chain, a middle chain and a light chain. Interestingly, melatonin counteracted the calyculin A‐triggered reduction of middle chain neurofilament expression. Thus, melatonin would be restoring the abnormal transcription and phosphorylation of these neurofilaments (Li et al., 2004) through a mechanism independent of its antioxidant properties that, according to the authors, could be related to GPCR‐mediated http://www.guidetopharmacology.org/GRAC/FamilyDisplayForward?familyId=284 inhibition and PPP stimulation, which has been predicted by previous studies (Karlsson et al., 2000).

Ling et al. (2009) reported surprising results regarding the relationship between melatonin and PP1 using in vivo models. They were able to inhibit the endogenous synthesis of melatonin by constant illumination in rats, which showed pathological hallmarks and cognitive impairments that were compatible with AD‐like neurodegeneration such as memory deficit, oxidative stress, synaptic damage or multi‐site Tau hyperphosphorylation. The phosphorylation rate imbalance was due to the activation of kinase enzymes, such as http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=2030 and PKA, and interestingly to PP1 depression, as shown by testing the total phosphatase activity of rat hippocampal extracts with [³²P]‐phosphorylase a as a substrate. The rats subjected to constant light (400 Lux) showed a 70% decrease in PP1 phosphatase activity. However, the simultaneous administration of exogenous melatonin (1 mg·kg⁻¹) not only restored the PP1 activity but also enhanced it by more than threefold when compared with the control non‐light‐stimulated group. Overall, it seems that PP1 is one of the enzymes most sensitive to fluctuations in brain melatonin brain levels, as the reduction of endogenous melatonin noticeably reduced PP1 phosphatase activity, but supplementation with exogenous melatonin sharply increased it.

Effect of melatonin on PP2A

PP2A is an essential Ser/Thr phosphatase that plays an important role in cell development, cell proliferation and death, cell mobility, cytoskeleton dynamics and cell cycle control, as well as in the regulation of numerous signalling pathways (Janssens and Goris, 2001). PP2A has also been claimed to be an important tumour suppressor in some cancer types. Taking all these observations into account, it is not surprising that its impairment has been associated with many pathologies, including cancer, AD, PD and some metabolic disorders. PP2A is one of the most abundant enzymes in eukaryotes, ubiquitously expressed and highly conserved from yeasts to humans. It is the major phosphatase in the brain accounting for more than 70% of Tau dephosphorylation.

PP2A exists in the cells in two forms: as a heterodimeric core enzyme and as a heterotrimeric holoenzyme. However, complete enzymic activity is only achieved with the heterotrimeric holoenzyme. The PP2A core consists of a scaffold subunit (A subunit or PR65 subunit) and a catalytic subunit (subunit C). Both subunits have two isoforms, α and β. The individual subunits cannot be found in vivo even though AC dimers are abundant in tissues, the most prevalent forms of PP2A are the heterotrimers. The PP2A core interacts with a wide variety of regulatory subunits B that are assembled in the holoenzyme. The regulatory subunit comprises four families: B (B55 or PR55), B′ (B56 or PR61), B″ (PR48, PR72 or PR130), and B″ (PR93 or PR110). The B subunit regulates substrate specificity and complex formation (Sents et al., 2013). Such diversity in the PP2A subunits can lead to various combinations of more than 200 PP2A heterocomplexes, thereby increasing the possibilities of selectivity at various phosphorylation sites and phosphoprotein substrates.

Three processes regulate PP2A activity: phosphorylation, methylation and the binding of endogenous inhibitors such as inhibitor I1^PP2A, inhibitor I2^PP2A (also called SET) and the oncoprotein cancerous inhibitor of PP2A (CIP2A). All of these inhibit phosphatase activity by direct intervention with the catalytic subunit (PP2A_C) (Martin et al., 2013). Abnormal cleavage of I2^PP2A in its nuclear localization sequence leads to the overexpression of I1^PP2A and I2^PP2A/SET and consequently to a decrease in PP2A activity. Methylation of the PP2A_C terminal carboxy, catalysed by PP2A methyltransferase, is necessary for the in vivo assembly of the trimer that is compensated by the opposing action of protein phosphatase methylesterase‐1 (PME‐1). The other major post‐translational modification is the phosphorylation of PP2A_C at Tyr 307 carried out by PTPA that produces inhibition of PP2A. It is worthwhile mentioning that PP2A exhibits a partial ability to dephosphorylate itself at Tyr 307.

In post‐mortem AD brains, PP2A activity was significantly impaired as down‐regulated expression of PP2A subunits reduced PP2A_C carboxymethylation, increased expression of endogenous inhibitors and increased PP2A_C phosphorylation was observed (Sontag et al., 2004). As a consequence, Tau protein becomes hyperphosphorylated and in turn aggregated, leading to the formation of NFT, microtubule disruption and subsequent neurodegeneration. Mutations in the PP2A subunits disrupt PP2A holoenzymes and are an unequivocal biomarker of different human cancers such as endometrial and ovarian carcinomas, lung and colon cancers and liver carcinoma. Endogenous inhibitors can also be over expressed such as CIP2A in colorectal cancer, SET in acute myeloid leukaemia, α4 in liver carcinoma, lung carcinoma or breast cancer and PME‐1 in both glioma and endometrial cancer.

In summary, a more profound knowledge and better understanding of PPP activity, specifically of PP2A, could provide new pharmacological strategies in the future to treat a wide range of diseases, including cancer and NDD. Interestingly, melatonin has also been object of study as a drug to modulate PP2A. Karlsson's research group (Karlsson et al., 2000) focused on the study of organelle movement using melanophores from poikilothermic vertebrates as a model of cytoskeleton‐mediated organelle transport, which is essential for axonal transport, endocytosis, secretion and inter‐compartmental trafficking. This intracellular movement appears to be regulated by the reversible phosphorylation of proteins by Ser/Thr kinases and phosphatases. The group investigated the signal transduction pathways involved in the regulation of intracellular melanosome transport in Xenopus laevis melanophores and found that melatonin led to the aggregation of melanosomes towards the cell centre. As it has been shown that G_i/o‐protein‐coupled receptors can mediate cell proliferative signals, microtiter‐plate assays and Western blot experiments were performed to study the role of tyrosine kinases in melanosome movement. The results showed that melatonin induced aggregation of the melanosomes to the centre of the cell through a G_i/o‐protein‐coupled receptor, http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=107, which mediates the inhibition of PKA and the stimulation of PP2A.

As for the possible effect of melatonin, it protected the cell viability of SH‐SY5Y neuroblastoma cells exposed to OA, by 43% at a concentration of 1 μM, as measured by MTT reduction (Lajarin‐Cuesta et al., 2016). In this cell line model, PP2A was selectively inhibited by OA (20 nM).

The work presented by Montilla et al. (2003) reviewed in the previous section discussed the influence of OA on the generation of oxidative stress when tested at a concentration where both PP1 and PP2A were compromised, which reproduces the biochemical changes that resemble AD, as well as the important antioxidant and protective effect of melatonin under these conditions. Again, these effects were independent of MT₁ and MT₂ receptors.

Koh (2012) reported that melatonin exhibited a neuroprotective profile through the induction of the PP2A regulatory subunit B in rats with ischaemic brain injury, following middle cerebral artery occlusion (MCAO). The adult male rats were treated either with the vehicle or melatonin (5 mg·kg⁻¹), and the middle cerebral artery generated was collected from the cerebral cortex tissues 24 h later. Through a proteomic approach, significant lower levels of the PP2A subunit B were observed in the animals treated with the vehicle only and this decrease was prevented by treatment with melatonin. The same results were observed using Western blot analyses. Immunohistochemical staining assays showed a decrease in the number of cells positive for the PP2A subunit B in cerebral cortices of the MCAO rats, treated with the vehicle, compared with the control group. Again, a decrease in the PP2A_B subunit was attenuated by melatonin. Additionally, there was a decrease in the PP2A subunit B levels of http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=1369‐exposed cell cultures that was reversed by the presence of melatonin.

In a manner similar to PP1, light modulates PP2A levels and activity. Eye diseases, such as age‐related macular degeneration or retinitis pigmentosa, are linked to light as a risk factor due to apoptotic cell death provoked by continuous or intense light. Nonetheless, the mechanism responsible for the morphological alteration that initiates the death signal after light exposure remains elusive. Sripathi et al. (2012) suggest that appropriate levels of NO generated by retinal pigment epithelium under oxidative stress induce the phosphorylation of both vimentin and PP2A, facilitating the triggering of stress‐dependent apoptotic pathways. As mentioned previously, the phosphorylation of PP2A at Tyr 307 inactivates it. Zhang et al. (2010) observed the opposite result using an experimental protocol defined by 12 h light/12 h dark cyclic conditions of a retina in vivo. They identified five proteins which increased during constant light and which were down‐regulated upon the administration of melatonin. These five proteins, including vimentin and PP2A, play important roles in phototransduction, rhodopsin trafficking and 11‐cis retinal regeneration. The authors hypothesized that melatonin acts as a downstream regulator of light and reverses light‐induced protein changes. Modifications by vimentin are critical for the reorganization of the intermediate filament network and Müller cell function. Melatonin decreased the light‐induced modifications of vimentin and improved the normal structure of the intermediate filaments in Müller cells and therefore maintained the neuro‐retinal architecture and neuronal survival. PP2A directly binds to vimentin, dephosphorylating it. Proteomic results showed that PP2A is notably up‐modulated in constant light, presumably due to vimentin activation. Hence, in this experimental model, melatonin could be controlling the formation of more stable vimentin‐PP2A complexes and negatively regulating apoptosis, by PP2A down‐regulation. As melatonin modulates many processes in the eye, such as retinomotor movements, neurotransmitter release and retinal pigment epithelium phagocytosis (Baba et al., 2009), the identification of these melatonin‐targeted proteins could make them a candidate for achieving the normal maintenance of retinal health, and for them to become a potential therapy for oxidative retinal diseases, providing a new role for melatonin, which hitherto is better known an antioxidant.

Melatonin is also implicated in pain perception, and some studies have shown that melatonin produces anti‐allodynia of neuropathic pain in animal models that is blocked by MT₂ receptor antagonists. Melatonin caused a dephosphorylation of http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=2659 (HDAC4), prompting its nuclear import, and preventing neuropathic pain. Recently, Lin et al. (2016) found that HDAC4 phosphorylation induced its cytosolic accumulation in dorsal horn neurons promoting nociceptive hypersensitivity. Previously, Zhang et al. (2003) found that PP2A inhibition resulted in an increase in heat allodynia and heat hyperalgesia in http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2486‐treated rats, demonstrating the essential role of PP2A in nociceptive hypersensitivity. PP2Ac makes contact with the N‐terminus of HDAC4 leading to dephosphorylation of Ser residues and results in the movement of HDAC4 into the nucleus. The high‐mobility group protein B1 (HMGB1) is one of the most important chromatin proteins and this protein takes part in the neuropathic pain caused by tibia nerve injury and spinal nerve ligation. Lin et al. (2016) recently reported that neuropathic pain caused by spinal nerve ligation induces behavioural allodynia, which is related to a decrease in PP2A_C levels in dorsal horn neurons. In addition, they also observed that rats subjected to knockdown of the spinal PP2A_C expression developed behavioural allodynia. Furthermore, spinal PP2A_C down‐regulation is linked to HDAC4 phosphorylation and shuttling, as well as a decrease in PP2A_C‐HDAC4 coupling, which induces HDAC4 phosphorylation. The HDAC4 remained restricted to the cytoplasm of dorsal horn neurons and therefore unable to bind to the nuclear promoter region, developing behavioural allodynia. However, melatonin through its interaction with MT₂ receptors increased PP2Ac expression, PP2Ac‐HDAC4 coupling and the subsequent cascade ending in attenuated allodynia.

Effect of melatonin on PP3/Cn/PP2B

The PP3 holoenzyme, also called Cn or PP2B, is a ubiquitous heterodimer consisting of a 60 kDa catalytic ‘A’ subunit, which interacts with http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=2351 (CaM) in a Ca²⁺‐dependent reaction, and an intrinsic Ca²⁺‐binding 19 kDa regulatory ‘B’ subunit, which contains four Ca²⁺‐binding domains (Dodge and Scott, 2003). Cn is activated by elevated Ca²⁺ levels and the subsequent activation of CaM. This holoenzyme can have specific roles in several physiological processes such as neurotransmission, cardiac and insulin signalling and immune response modulation. It is abundant in the hippocampus, neocortex, striatum and cerebellum, and has an important role in both memory and synaptic plasticity. It is the most abundant Ca²⁺‐ and CaM‐dependent enzyme in adult brain (Klee et al., 1979). Thus, the homeostatic balance between the activities of http://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=1554 (CaMKII) and Cn is essential for the regulation of synaptic Ca²⁺ levels. The reduction in both the synaptic activity of CaMKII and the overactivation of Cn due to calpain‐mediated cleavage plays an important role in AD pathogenesis (Ghosh and Giese, 2015). These mechanisms suggest that excessive activation of Cn would lead to spine loss in AD. Therefore, it would be a potential therapeutic target for the prevention of memory loss. LTP induction requires CaMKII activation in the hippocampus. Studies carried out by Fukunaga et al. (2002) showed that melatonin might act on CaMKII‐dependent signals involved in LTP induction, whereas http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=1024, a Cn inhibitor, significantly potentiated LTP induction, and melatonin completely abolished it.

As described above, Cn regulates the immune response through diverse signalling pathways in mammals. The NFAT proteins are a family of transcription factors that are important mediators of the immune response. Thus, when continuous Cn hyperactivation leads to the translocation of the phosphorylated cytoplasmic NFAT protein in the nucleus, it combines with the inducible nuclear transcription factor component to form a protein complex, which triggers target gene expression. Continuous NFAT activation and nuclear signalling result in neurite dystrophy and dendritic spine loss. Because the first step of NFAT activation is the dephosphorylation by Cn, melatonin would bind to CaM in a Ca²⁺‐dependent and MT receptor–independent manner (Turjanski et al., 2004) and specifically prevent, in a dose‐dependent fashion, dephosphorylation of NFAT by Cn and its translocation to the nucleus, which plays an important role in the T cell‐mediated immune responses. More recently, it has been reported that CaMKII and Cn expression was increased by http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=5019 exposure in myocardial cells, while melatonin attenuated this action when administered prior to LPS stimulation (Lu et al., 2015).

Concluding remarks

An analysis of all these contributions indicates that melatonin has a significant influence on the regulation of phosphorylation processes. The research work presented in this review characterizes the pharmacological activity of melatonin beyond its antioxidant properties or its interaction with MT receptors. The latest findings point to a more direct relationship with Ser/Thr phosphastases in general and PPP in particular. We hypothesize the existence of direct binding by melatonin to these enzymes, which would be partly responsible for its modulatory activity in the phosphorylation/dephosphorylation processes (Figure 2).

Figure 2.

Modulation of the cerebral activity of Ser/Thr phosphatases by melatonin. Beyond its direct effect on melatonin receptors, melatonin exerts important antioxidant actions that prevent the impairment of the ROS‐sensitive PP2A and PP1. In addition, direct interactions with these enzymes, either activating or inhibiting, would lead to the maintenance of neuronal plasticity and health, by preventing NFT formation, heavy neurofilament phosphorylation or loss of dendritic spines.

The increasing interest by those members of the scientific community devoted to drug research and development in placing PPP, such as PP1, PP2A or Cn, as validated biological targets for NDD has stimulated the study of their regulation by one of the most promising drugs of today, melatonin. The therapeutic potential of melatonin, for diseases related to ageing, as well as brain damage or cancer, has been widely documented in the last decade. Although most of the research reviewed here has studied melatonin in cell cultures, there are also contributions using in vivo models where melatonin has been studied in terms of its pharmacokinetics, assessing its metabolism, binding to plasma proteins or cell absorption. In these experiments, melatonin also showed modulation of PPP enzymes. Consequently, this review fulfils two aims. Firstly, by demonstrating that many of the therapeutic actions described for melatonin can be explained by its regulatory effects on PPP, and secondly, the chemical optimizations conducted into the design of new melatonin derivatives with enhanced neuroprotective properties should focus on potentiating the pharmacological modulation of several PPP, namely, PP1, PP2A and Cn, which play an essential role in a number of physiological and pathological processes.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 (Alexander et al., 2017a,b,c).

Conflict of interest

The authors declare no conflicts of interest.

Arribas, R. L. , Romero, A. , Egea, J. , and de los Ríos, C. (2018) Modulation of serine/threonine phosphatases by melatonin: therapeutic approaches in neurodegenerative diseases. British Journal of Pharmacology, 175: 3220–3229. 10.1111/bph.14365.