Tau phosphorylation by glycogen synthase kinase 3β modulates enzyme acetylcholinesterase expression

Abstract In Alzheimer's disease (AD), the enzyme acetylcholinesterase (AChE) co‐localizes with hyperphosphorylated tau (P‐tau) within neurofibrillary tangles. Having demonstrated that AChE expression is increased in the transgenic mouse model of tau Tg‐VLW, here we examined whether modulating phosphorylated tau levels by over‐expressing wild‐type human tau and glycogen synthase kinase‐3β (GSK3β) influences AChE expression. In SH‐SY5Y neuroblastoma cells expressing higher levels of P‐tau, AChE activity and protein increased by (20% ± 2%) and (440% ± 150%), respectively. Western blots and qPCR assays showed that this increment mostly corresponded to the cholinergic ACHE‐T variant, for which the protein and transcript levels increased ~60% and ~23%, respectively. Moreover, in SH‐SY5Y cells differentiated into neurons by exposure to retinoic acid (10 µM), over‐expression of GSK3β and tau provokes an imbalance in cholinergic activity with a decrease in the neurotransmitter acetylcholine in the cell (45 ± 10%). Finally, we obtained cerebrospinal fluid (CSF) from AD patients enrolled on a clinical trial of tideglusib, an irreversible GSK3β inhibitor. In CSF of patients that received a placebo, there was an increase in AChE activity (35 ± 16%) respect to basal levels, probably because of their treatment with AChE inhibitors. However, this increase was not observed in tideglusib‐treated patients. Moreover, CSF levels of P‐tau at the beginning measured by commercially ELISA kits correlated with AChE activity. In conclusion, this study shows that P‐tau can modulate AChE expression and it suggests that AChE may possibly increase in the initial phases of AD.


| INTRODUC TI ON
During the progressive course of Alzheimer's disease (AD), the most common neurodegenerative dementia in the elderly, there is a loss of forebrain cholinergic neurons and a marked synaptic cholinergic deficit that most likely contributes to the cognitive, behavioral, and functional symptoms of AD (Bohnen et al., 2005;Mufson et al., 2003Mufson et al., , 2008. This cholinergic deficiency courses with a progressive decline in synaptic acetylcholine (ACh) levels (Davies & Maloney, 1976;Perry et al., 1977), affecting both choline acetyltransferase (ChAT), the rate-limiting enzyme that synthesizes ACh, and acetylcholinesterase (AChE), the ACh hydrolyzing enzyme, activities. However, AChE activity is increased around the two neuropathological hallmarks of the disease: the amyloid plaques, which are extracellular deposits of aggregated β-amyloid (Aβ) protein, and the neurofibrillary tangles (NFT) of microtubule-associated protein tau abnormally hyperphosphorylated (P-tau) (Mesulam et al., 1987).
The increase in AChE in the plaques may be triggered by the deposition of aggregated Aβ, since Aβ peptides influence AChE levels both in vitro (Hu et al., 2003;Sberna et al., 1997) and in vivo (Dumont et al., 2006;Sberna et al., 1998). The early increase in AChE around NTF, even in regions lacking amyloid plaques (Ulrich et al., 1990) has yet to be explained.
AChE is very polymorphic, with different molecular forms adopting distinct subcellular locations and performing a variety of physiological roles (Massoulié, 2002;Meshorer & Soreq, 2006). Alternative mRNA splicing generates three AChE transcripts that differ in their C-terminal region. The cholinergic AChE form is a tetramer formed by subunits encoded by the AChE-T (tailed) transcript, which co-exists in brain with minor amounts of the AChE-R ("readthrough") variant whose expression is increased in stress (Meshorer et al., 2002), and AD (Campanari et al., 2016). A third splicing variant AChE-H ("hydrophobic"), in most species is expressed in non-nervous cells . Moreover, an alternate upstream promoter usage produces another version of each of these AChE variants with an extended N terminus (N-AChE-T and N-AChE-R) (Meshorer & Soreq, 2006).
Our group has previously shown that over-expression of P-tau leads to an increase in AChE expression in the brain of Tg-VLW mice expressing 3 missense mutations of human tau (G272V-P301L-R406W) associated with the autosomal dominantly inherited frontotemporal dementia and parkinsonism linked to chromosome 17 (FTDP-17) . Moreover, AChE activity and mRNA of AChE-T transcript have been reported to be increased in the septum, but not in other brain areas, of Tg601 mice, a transgenic model of tauopathy that expresses P-tau-positive neurons, but does not develop NFTs (Hara et al., 2017). These previous results suggest that an altered tau phosphorylation may underlie some of the changes in AChE levels observed in patients with AD.
The major tau phosphorylating kinase is glycogen synthase kinase-3β (GSK3β) that may have a central role in the AD pathogenesis (Hooper et al., 2008), as well as in other neurodegenerative diseases (Yang et al., 2008). GSK3β over-expression and subsequent increased P-tau, have been correlated with neurodegeneration  and formation of NFT (Chu et al., 2017;Jaworski et al., 2011). Moreover, GSK3β inhibitors have been proposed as therapeutic agents for AD (Hooper et al., 2008;Medina & Castro, 2008) since it has been demonstrated that GSK3β inhibition reduces tau phosphorylation (Hooper et al., 2008;Medina & Castro, 2008) and even amyloid production, both in vitro and in vivo (Hooper et al., 2008;Su et al., 2004).
In this study, we have investigated the interaction between AChE and P-tau in human neuroblastoma SH-SY5Y cells by analyzing the effect of increased tau phosphorylation mediated by GSK3β on AChE mRNA and protein expression. We also have studied whether inhibition of GSK3β influences AChE expression in neuronal primary cultures. Finally, we have examined AChE activity levels in the cerebrospinal fluid (CSF) of AD patients treated with the GSK3β inhibitor tideglusib (Lovestone et al., 2015;del Ser et al., 2012).  (Cat. No. 17,504,044: Gibco BRL), 100 IU/mL penicillin, 100 μg/mL streptomycin, and 2 mM glutamine (Cat. No. 11,500,626: Thermo Fisher Scientific). A total number of three female pregnant mice were used for these experiments. After 7 days in culture, primary cortical neurons were transfected with GSK3β and tau cDNAs using Lipofectamine LTX (Cat. No.15338100: Thermo Scientific), according to the manufacturer's instructions. The efficacy of transfection was assayed using a GFP-pCI cDNA vector.

| Pharmacological treatments
After DNA transfection, SH-SY5Y cells and primary cortical neurons were treated for 24 hr with 20 μM of SB216763 (Cat. No. S3442: Sigma-Aldrich Co.), a GSK3β inhibitor (Wagman et al., 2004). After treatment, the conditioned medium was removed and the cells were washed twice with PBS, harvested, suspended in ice-cold extraction buffer and solubilized as described previously in Cell culture section.
Cell viability was assessed using the MTS assay, as described earlier.

| AChE assay and total protein determination
AChE activity was assessed with a microassay adapted from the colorimetric Ellman method (Sáez-Valero et al., 1993), adding 1 mM acetylthiocholine (Cat. No. A5626: Sigma-Aldrich Co.) in the presence of 50 µM tetraisopropyl pyrophosphoramide (Iso OMPA, Cat. No. T1505: Sigma-Aldrich Co.) to block any contamination by butyrylcholinesterase. One milliunit (mU) of AChE activity was defined as the number of nmoles of acetylthiocholine hydrolyzed per min at 22°C.
Protein concentrations were determined using the bicinchoninic acid method, with bovine serum albumin as standard (Cat. No. 23,225: Pierce, Rockford, IL).

| Sedimentation analysis
Molecular forms of AChE were separated according to their sedimentation coefficients by ultracentrifugation on 5%-20% (w/v) sucrose gradients containing 0.5% (w/v) Triton X-100 (Sáez-Valero et al., 1993). Ultracentrifugation was performed in a SW41Ti Beckman rotor at 150,000 g for 18 hr at 4 ºC. Approximately 40 fractions were collected from the bottom of each tube and assayed for AChE activity to identify individual AChE forms (G 4 : tetramers; G 1 : monomers) by comparison with the position of molecular weight markers, catalase (11.4S, Cat. No. C9322: Sigma-Aldrich Co.) and alkaline phosphatase (6.1S, Cat. No. P0114: Sigma-Aldrich Co.).

| Determination of acetylcholine levels
Cellular ACh levels were measured using the commercial fluorometric Choline/Acetylcholine Assay Kit (Cat. No. ab65345: Abcam, Cambridge, UK) following the manufacturer's guidelines. Briefly, SH-SY5Y cells transfected as previously described were harvested in PBS and re-suspended in assay buffer. After centrifugation, the levels of total choline and free choline were determined in the supernatant. The concentration of ACh in the samples was calculated as the difference between total choline and free choline levels.

| Human CSF samples
This study was approved by the ethics committee of the Hospital General Universitario de Elche (License no: PI 10/2011) and was carried out in accordance with the Declaration of Helsinki. This study was not pre-registered. The CSF samples used for this study were deidentified leftover aliquots from the ARGO study, a phase II clinical trial with the GSK3β inhibitor tideglusib (NP031112-010B04) performed by Noscira SA (Madrid, Spain) on AD patients (ClinicalTrials. gov number NCT01350362; see ( ;Lovestone et al., 2015) for details about the study). The total length of the trial was 26 weeks in which AD patients were orally administered 1,000 mg of tideglusib using two different regimes, either once a day (QD; seven cases) or every other day (QOD; seven cases) or a matching placebo (five cases). See Table 1 for details.
All AD patients (6 females/13 males; 71 ± 2 years old) fulfilled the NINCDS-ADRDA criteria for "probable" AD (McKhann et al., 1984), had a mean MMSE score of 19 ± 1, had been taking a stable and well-tolerated dose of an AChE inhibitor (AChE-I: rivastigmine, donepezil, or galantamine; see Table 1) for at least 4 months prior to the enrollment into the trial, and AChE-I treatment was continued during the trial.
CSF was obtained by lumbar puncture at baseline and week 26, it was centrifuged at 1,000 g for 15 min to eliminate cells and insoluble material and immediately stored at −80ºC. Assays were performed blind for the treatment assignment and no randomization was performed to allocate subjects in the study. The study was exploratory, and no exclusion criteria were pre-determined with the samples obtained from Noscira.

| Increased tau phosphorylation lead to increased AChE levels
SH-SY5Y cells express AChE  and have been used to study the effects of tau and GSK3β over-expression (Bijur et al., 2000;Katsinelos et al., 2018).
Thus, in this study, human wild-type tau and GSK3β were overexpressed in SH-SY5Y cells to increase cellular levels of P-tau as compared with cells transfected with a control pCI "empty" vector or only with either tau or GSK3β alone (Figure 1a). Viability assays showed there was no significant cell death of GSK3β + tau overexpressing cells (11 ± 5% reduction relative to pCI, p =.1).
When AChE enzymatic activity was determined later in the cellular extracts, the levels of AChE were increased (20 ± 2%; p <.001) in cells over-expressing both tau and GSK3β compared with cells over-expressing tau or GSK3β alone (Figure 1b). AChE protein levels were also determined by western blot using the N-19 antibody, raised against a peptide that recognizes the N-terminus of human AChE, which is common to all variants. A major immunoreactive band of ~66 kDa, consistent with the molecular mass of full-length AChE, and a faint band of 55 kDa that could not be quantified reproducibly, were detected in all samples. Immunoreactivity of the 66 kDa TA B L E 1 Demografic data and I-AChE treatment AD patients were administered with placebo or 1,000 mg of tideglusib orally using two different regimes: once a day (QD) and every other day (QOD). . Over-expression of VLW increased P-tau levels ( Figure S2a) and AChE activity (a 14% ± 3 increment, p < 0.001: Figure S2b), as compared with pCI-transfected cells. Moreover, the AChE protein also increased in these cells as a result of VLW transfection (385% ±43, p < 0.001: Figure S2c).

F I G U R E 1
Increase in acetylcholinesterase (AChE) levels in neuroblastoma SH-SY5Y cells with increased tau phosphorylation as a result of GSK3β and tau over-expression. SH-SY5Y cells were transfected with DNA vectors that encode GSK3β, wild-type tau (tau), or both proteins (GSK3β + tau) or with a pCI empty control vector. (a) Each lane contained 30 µg of protein from cell extracts. Proteins were resolved by electrophoresis and probed with specific primary antibodies to GSK3β, tau or P-tau (clone AT8). Representative blots are shown. Increased tau phosphorylation was observed in cells that over-express GSK3β + tau. (b) AChE-specific activity (mU/mg of total protein) was also determined in cellular extracts. Percentages (%) of AChE activity relative to pCI control cells are represented. (c) Fifty microgram protein extracts were then assayed by immunoblotting using the N-19 antibody that recognizes all the AChE variants. Representative western blot (left panel) and densitometric quantification (right panel) expressed as percentage (%) relative to immunoreactivity of the control group are shown. Immunoreactivity values obtained by densitometry were normalized relative to that of the housekeeping protein GAPDH. (d) SH-SY5Y cells transfected with a pCI empty vector or GSK3β + tau were treated with 20 μM of SB216763, a GSK3β inhibitor, or with vehicle (DMSO) for 24 hr. Inhibition of tau phosphorylation was quantified by a decrease in P-tau immunoreactivity. AChE protein expression was also analyzed by western blots probed with N-19 antibody. Representative blots are shown. (e) AChE enzymatic activity was assayed in cells treated with 20 μM of SB216763, and expressed as percentage (%) relative to control cells. Results were confirmed in n = 18 independent cell determinations (obtained from 3 independent cell sets of experiments). Represented values are means ± SEM. *Significantly different (p <.05) from the control group, as assessed by one-way ANOVA followed by Tukey test for pair-wise comparisons

| Cholinergic AChE-T is the splicing variant increased in cells over-expressing GSK3β and tau
The influence of P-tau on particular AChE variant/s was further ana- script, also expressed in SH-SHY5Y cells (Bi et al., 2014), were also measured and no changes were found in GSK3Β and tau over-expressing cells compared with controls ( Figure 3d). In summary, these results suggested that a specific increase in the expression of the AChE-T cholinergic species can be triggered by increased phosphorylation of tau by GSK3β.

Undifferentiated SH-SHY5Y cells express monomeric forms of
AChE-T, and only after differentiation to a neuronal phenotype SH-SHY5Y cells express a pattern AChE isoforms similar to that in cholinergic neurons, with monomers (G 1 ) and tetramers (G 4 ) of the AChE-T variants (Grisaru et al., 1999;Massoulié, 2002;Taylor & Radic, 1994).
Thus, AChE activity and protein levels in RA-differentiated SH-SY5Y cells were analyzed to further investigate whether P-tau influences AChE cholinergic species. However, no differences were found between RA-differentiated SH-SY5Y cells over-expressing GSK3β + tau and control cells (Figure 4e).
We further studied whether the increase in AChE might result in an alteration in the levels of the neurotransmitter ACh. Cellular ACh levels were measured using a fluorometric method (Figure 4f). A reduction in ACh levels (45 ± 10% decrease; p = 0.04) was observed in RA-differentiated SH-SY5Y cells that over-express GSK3β + tau compared to control cells.

| AChE co-localizes with P-tau in cytoplasmatic regions
We have also analyzed the cellular location of AChE and P-tau to compare their distribution by immunocytochemistry on CHO cells that stably over-expresses AChE-T, and P-tau levels were induced by transfection of GSK3β and tau. P-tau immunochemistry was performed using an anti p-T181-tau antibody instead of an anti-Ptau202/205AT8 antibody. Previous western blot assays probed with anti p-T181-tau antibody also showed the increment on P-tau levels in GSK3β and tau over-expressing cells (data not shown). Confocal microscopy analysis ( Figure 5) showed that

F I G U R E 3 Increased levels of "tailed" acetylcholinesterase (AChE-T) splice variant in SH-SY5Y cells with increased tau phosphorylation.
Immunodetection of AChE variants in cellular extracts from SH-SY5Y cells with elevated P-tau resulting of glycogen synthase kinase-3β (GSK3β) plus total tau transfection (GSK3β + tau) and control cells transfected with a pCI empty vector (control). Thirty microgram of protein from cell extracts was resolved by electrophoresis and probed with specific primary antibodies raised to: (a) the C-terminus of the AChE-T variant; (b) the C-terminus of "readthrough" acetylcholinesterase (AChE-R) variant; (c) the extended N-terminus of AChE (N-AChE) variants. Representative blots and densitometric quantification of the immunoreactive bands, expressed as percentage (%) relative to immunoreactivity of the control group are shown. For semiquantitative analysis, the levels were normalized to the housekeeping protein GAPDH. (d) Relative mRNA levels of the transcripts for AChE splice variants were analyzed by qRT-PCR. The specificity of the PCR products was confirmed by dissociation curve analysis. Transcript levels were calculated by the comparative 2 −ΔCt method with respect to GAPDH.
Values are means ± SEM from at least of 24 cell wells independent determinations from four independent cell culture experiments. *p <.05 significantly different from the control group (Student's t-test) P-tau co-localizes with AChE mainly in cytoplasmatic regions with a Mander's coefficients of 0.51 ± 0.09 for P-tau and 0.41 ± 0.09 for AChE co-localization with P-tau.

| GSK3β inhibitor treatment influences AChE levels in AD patients
The in vitro studies have shown that an increase in P-tau levels in a cellular model triggers an increment in AChE expression and that treatment with the GSK3β inhibitor SB216763 can block this effect.
To assess the role of tau phosphorylation on AChE in AD patients, CSF samples from patients enrolled in a clinical trial of the GSK3Β inhibitor tideglusib (NP031112-10B04)) were analyzed. CSF levels of core AD biomarkers, Aβ42, T-tau, and P-tau, as well AChE activity were measured in aliquots of CSF samples collected before and after 26 weeks of tideglusib treatment (see Table 1 for details). In agreement with previous results (Lovestone et al., 2015), a non-significant (p =.087) decreasing trend was observed in CSF P-tau levels after treatment with tideglusib, compared with placebo-treated patients, whereas no changes were found in the levels of T-tau (p =.84) and Aβ42 (p =.14: Figure S3). No differences were found for these core F I G U R E 4 Increased tau phosphorylation leads to a cholinergic imbalance with decreased acetylcholine levels in retinoic acid (RA)differentiated SH-SY5Y cells. RA-differentiated SH-SY5Y cells were transfected with pCI empty vector (control) or plasmid cDNAs that encode GSK3β and tau (GSK3β + tau). (a) Ultracentrifugation on sucrose gradient served to verify the expression of tetramers of acetylcholinesterase (AChE) in non-transfected RA-differentiated cells. Molecular forms of AChE (tetramers: G 4 ; and light monomers: G 1 ) were identified by comparison with the position of molecular weight markers catalase (11.4S) and alkaline phosphatase (6.1S). Representative profile is shown. (b) Immunoblots probed to total tau (T-tau), GSK3β and phosphorylated tau (P-tau) were then performed to verify transfection and the effect on tau phosphorylation. Equal amounts of protein were loaded in all lanes. Representative blots are shown. (c) AChE enzymatic activity was assayed in cellular extracts and expressed as percentage (%) of the activity respect to pCI controltransfected cells. (d) AChE protein levels were assayed by western blotting of 30 μg of protein from cellular extracts using the anti-AChE antibody N-19. Immunoreactivity of the AChE bands was quantified and normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Immunoreactive levels were expressed as % of control levels. Representative blot and densitometric quantification are presented.
(e) The levels of choline acetyltransferase (ChAT) were assayed by western blot in cellular extracts, with GAPDH as loading control. (f) Cellular levels of the neurotransmitter acetylcholine (ACh) were also measured using a commercial fluorometric method and expressed as % respect to control. Results were confirmed in at least of n = 18 cell wells from three independent cell cultures experiments. Mean value ± SEM are represented. *Significantly different (p < 0.05) from the control group, as assessed by the Student's t test AD biomarkers between QOD and QD treatment administration groups.
AChE levels before tideglusib treatment were similar in the patients that received the placebo or tideglusib treatment (27.23 ± 6.46 and 22.60 ± 3.30, respectively). Patients with rivastigmine treatment displayed significantly lower AChE activity levels (11.46 ± 1.54) than those treated with galantamine (33.01 ± 2.79) or donepezil (30.03 ± 11.85). Interestingly, AChE activity levels correlated with P-tau levels (n = 19 subjects; r = 0.525; p =.021: Figure 6a) and this correlation was higher when patients under rivastigmine treatment were excluded (n = 12 subjects; r = 0.709; p = 0.010). The baseline correlations of AChE activity with T-tau (r = 0.44; p =.06), and with Aβ42 (r = 0.027; p =.91) were non-significant. Interestingly, placebo-treated patients had a significant increase in AChE activity after treatment (35 ± 16%; p =.04: Figure 6b), whereas those treated with tideglusib did not display significant changes from baseline levels (5 ± 3% increase; p = .94: Figure 6b). This steady AChE activity during the trial was independent of the prescribed AChE-I (ANOVA test comparison of the three sub-groups: p =.24). A positive correlation was found between the change in P-tau levels and AChE F I G U R E 5 Acetylcholinesterase (AChE) colocalizes with phosphorylated tau (P-tau) in cytoplasmatic regions in Chinese hamster ovary (CHO) cells. CHO cells that stably over-express ACHE-T variant were transfected with plasmid cDNAs that encode glycogen synthase kinase-3β and tau to increase P-tau levels. Immunoassay was performed and confocal images were obtained with a x63 oil immersive objective lens. From left to right we observe: stained nuclei with Hoechst in blue; AChE probed with anti-N-terminus AChE antibody plus anti-rabbit IgG Alexa Fluor 488 in green; P-tau with anti-PHF-tau antibody combined with an anti-mouse IgG Cy5  r = 0.768; p <.001). The correlation was also observed for the placebo-treated patients (n = 5; r = 0.949; p =.014) although only showed a non-significant tendency in the tideglusib-treated subgroup alone (n = 14, r = 0.450; p =.10: Figure 6c). This tendency increased when the tideglusib-treated patients under rivastigmine medication were excluded from the analysis (n = 8 subjects, r = 0.661; p =.074: data not shown).

| D ISCUSS I ON
This study demonstrated that an increased phosphorylation of tau by the kinase GSK3β can modulate the levels of the cholinergic AChE. These changes may consequently compromise cholinergic neurotransmission. Nonetheless, we cannot discard that GSK3β also exerts a cholinergic regulatory effect not related with P-tau, a possibility that should be explored. Anyhow, our results are in accordance with a previous report where we showed that in a mouse model of FTDP17 tauopathy (Tg-VLW), there are higher levels of AChE-T protein, enzymatic activity and transcript levels than in wild-type background strain mice . However, other au- tation, as well as other tau mutations, make tau a more favorable substrate for brain protein kinases, favoring its hyperphosphorylation. This VLW mutation increases the hyperphosphorylation of tau at phosphothreonine 231 and phosphoserines 199/202 (Lim et al. 2001). We could not discard that tau phosphorylation by other kinases that are implicated in AD, such as CDK5 (Liu et al., 2016), AMPK (Tu et al., 2014), MAPKs (Zu et al. 2001), could also affect AChE expression, like GSK3β. However, further studies will be necessary to clarify this issue. The analysis of AChE levels in a small sample of AD patients undergoing a clinical trial with the GSK3β inhibitor tideglusib showed a similar but inconclusive tendency since only a minor not statically significant reduction was noticed for P-tau levels. However, our data suggest that there may be cross-talk between P-tau and AChE.
In this study, AChE protein levels were analyzed by SDS- The alteration in cellular ACh levels may have physiopathological consequences. In AD, decreased ACh is linked to impaired cognition, behavior, and daily living activities. A previous report from our group using a rat model of liver cirrhosis showed that impairment of the  (Jing et al., 2007). The regulation of AChE-T during apoptosis (Zhang et al., 2002) may be driven by specific mechanisms. Indeed, it was shown that GSK3β may stabilize AChE-T protein preventing its proteasomal degradation in HEK cells (Jing et al., 2013). GSK3β revealed multifaceted roles also related with transcriptional regulation (Lauretti et al., 2020). Thus, we cannot discard that GSK3β also modulates AChE expression by other mechanisms independent of changes in tau phosphorylation.
The increase in AChE-T may also be a consequence of the increase in intracellular calcium. AChE-T expression is modulated by AChE promoter activity as a result of perturbations in intracellular calcium homeostasis (Luo et al., 1994;Zhu et al., 2007) and extracellular tau can induce an increase in intracellular calcium in cultured neuronal cells, probably via M1 and M3 muscarinic receptors (Gómez-Ramos et al., 2008). Moreover, an increment of intracellular calcium could also trigger an enhanced AChE promoter activity as a result of the binding of CCAAT-binding factor (CBF/NF-Y) to the CCAAT motif within AChE promoter. Thus, the increase in AChE-T could also be related to the alterations of the binding to transcription-binding sites within the AChE promoter region such as Sp-1, cyclic AMP-responsive element, AP-1, and NFAT that could be regulated by tau phosphorylation and GSK3β. Indeed, the deposition of phosphorylated tau has been related to an increased c-Jun, c-Fos, and CREB-1 expression in neurons in Picks disease brains (Nieto-Bodelón et al., 2006). In this regard, binding of c-jun to the AP1 site of AChE promoter has been related to an increased AChE-T in PC12 cells (Zhang et al 2008).
AChE-T over-expression has been linked to an increase in neurodegeneration (Farchi et al., 2007) and programmed cell death (Greenberg et al., 2010;Toiber et al., 2009). Interestingly, it has been reported that AChE-T facilitates Aβ fibril formation and AD plaque formation (Berson et al., 2008). Indeed, transgenic mice over-expressing AChE-T and the APP Swedish mutation, show early deposition and more abundant β-amyloid plaques than mice over-expressing mutant APP Swedish mutation alone (Rees et al., 2005). We speculate that AChE-T is part of a vicious circle in which Aβ deposition may be potentiated as a result of an increase in AChE-T by P-tau.
Several GSK3β inhibitors are currently being tested for the treatment of AD. Tideglusib is an irreversible GSK3β inhibitor, which reduces tau phosphorylation and prevents apoptotic death in human neuroblastoma cells and murine primary neurons (Domínguez et al., 2012). In AD patients treated with tideglusib a trend toward a  (Lovestone et al., 2015), tideglusib-treated patients showed a trend toward decreased CSF P-tau levels as compared with placebo-treated patients that was not statistically significant.
This result might suggest that tideglusib does not act as a GSK3β inhibitor, yet the authors indicated that an inhibitory effect was present since there was a significant decrease in CSF BACE1 relative to the patients that received the placebo. This alteration to BACE1 is in accordance with the reduction observed in an AD transgenic model treated with lithium, suggesting an inhibitory effect of the drug. There were no changes in any of the other core AD biomarkers T-tau and Aβ1-42. Importantly, patients enrolled in this clinical trial were under AChE-I treatment for at least 4 months prior to baseline and maintained this AChE-I treatment during the trial. AChE activity in CSF is differently affected by AChE-Is; rivastigmine causes persistent inhibition of AChE (Darreh-Shori et al., 2002;Nordberg et al., 2009;Parnetti et al., 2011), whereas donepezil and galantamine cause a rebound increase in CSF AChE activity (García-Ayllón et al., 2007;Nordberg et al., 2009;Parnetti et al., 2011). In agreement with these data, AChE activity levels at baseline were lower in our rivastigmine-treated patients than in those under donepezil or galantamine treatment. The influence of AChE-I on CSF AChE activity may affect the interpretation of a potential effect of tideglusib on AChE levels. While AChE levels at baseline correlate with P-tau, suggesting that P-tau could influence this enzyme, a cause-effect relationship has not been demonstrated. A positive correlation was also observed for the placebo-treated patients between the change in P-tau levels and AChE activity during the trial. Interestingly, the levels of CSF AChE were still increased in the placebo group during the 26 weeks of trial (for four out of the five patients, two receiving donepezil, one galantamine and another one rivastigmine; one patient under donepezil treatment did not change). Tideglusib-treated patients displayed similar CSF AChE activity levels at the end of the trial, compared with baseline. This result suggests that even if the GSK3β inhibition by tideglusib fails to drive significant changes in CSF P-tau, it is able to act on CSF AChE activity. In a neurodegenerative mouse model over-expressing transgenic tau, the suppression of tau expression resulted in improved memory function although NFTs accumulation persisted (Santacruz et al., 2005). Thus, the pathological and therapeutic effect of changes in tau hyperphosphorylation should be evaluated not only in terms of tau pathology.
More studies are needed to clarify the potential effect of tideglusib on AChE activity and cholinergic function, if possible, in early diagnosed AD patients without previous treatment with AChE-I.
Taken together our findings point to a possible influence of tau hyperphosphorylation on cholinergic AChE activity that could be relevant in the physiopathology of AD. Therefore, the early increase in AChE expression that occurs around NFT (Mesulam et al., 1987) may be a consequence of disturbed tau phosphorylation. Indeed, we have shown in cellular models the colocalization of P-tau and AChE in cytoplasmatic regions like in neurons of the Tg-VLW mutant . Moreover, it has been reported that AChE activity may be preserved or even increased in some brain areas of patients with mild AD (Herholz et al., 2004). Additionally, the study supports the view that early tau hyperphosphorylation may cause an impairment of cholinergic activity with a decrease in ACh levels that may be a contributing factor for the degeneration of cholinergic neurons of the basal forebrain. Finally, our results indicated the relevance of measuring CSF AChE as a biomarker for trials with GSK3β inhibitors, and the need to analyze the cholinergic enzymes and neurotransmitter in future studies using animal models or clinical trials.