Memantine Attenuates Alzheimer's Disease-Like Pathology and Cognitive Impairment

Abstract

Deficiency of protein phosphatase-2A is a key event in Alzheimer's disease. An endogenous inhibitor of protein phosphatase-2A, inhibitor-1, I1PP2A, which inhibits the phosphatase activity by interacting with its catalytic subunit protein phosphatase-2Ac, is known to be upregulated in Alzheimer's disease brain. In the present study, we overexpressed I1PP2A by intracerebroventricular injection with adeno-associated virus vector-1-I1PP2A in Wistar rats. The I1PP2A rats showed a decrease in brain protein phosphatase-2A activity, abnormal hyperphosphorylation of tau, neurodegeneration, an increase in the level of activated glycogen synthase kinase-3beta, enhanced expression of intraneuronal amyloid-beta and spatial reference memory deficit; littermates treated identically but with vector only, i.e., adeno-associated virus vector-1-enhanced GFP, served as a control. Treatment with memantine, a noncompetitive NMDA receptor antagonist which is an approved drug for treatment of Alzheimer's disease, rescued protein phosphatase-2A activity by decreasing its demethylation at Leu309 selectively and attenuated Alzheimer's disease-like pathology and cognitive impairment in adeno-associated virus vector-1-I1PP2A rats. These findings provide new clues into the possible mechanism of the beneficial therapeutic effect of memantine in Alzheimer's disease patients.

Fig 1. Infection with AAV1-I₁ ^PP2A upregulates I₁ ^PP2A in rat brain.

A) Four and one-half months post-infection, Western blots developed with mouse monoclonal anti-I₁ ^PP2A (5G6) showed a significant increase in the level of the transgene (normalized with actin as loading control; actin blot the same as shown in Fig 4A) in the ventricular area, the cerebellum, the cerebral cortex and the hippocampus in AAV1-I₁ ^PP2A rats as compared with that in AAV1-GFP control rats. Memantine treatment didn’t alter the level of I₁ ^PP2A in either AAV1-I₁ ^PP2A or AAV1-GFP rats. *P<0.05, **P<0.01 vs. AAV1-GFP control rats. B) Immunofluorescence staining with anti-GFP showed transgene expression widely spread in the cerebral cortex(a and e), the hippocampus (b and c) and the ventricular area (d) in AAV infected rats. I₁ ^PP2A including the endogenous protein was observed with 5G6 staining both in AAV1-GFP rats (f) and AAV1-I₁ ^PP2A rats (g-j). The double staining showed that GFP colocalized with 5G6 only in AAV1-I₁ ^PP2A rats (l-o), not in AAV1-GFP rats (k). 3–4 rats were employed in each experiment. MMT, memantine; Vh, vehicle. Scale bar, 50 μm.

Fig 2. Memantine restores PP2A activity by decreasing the demethylation of PP2Ac at Leu 309.

A, B) Western blots developed with mouse anti-PP2A catalytic α (1:10000; BD Transduction Laboratories^™) showed no significant change in the level of PP2Ac in the AAV1-I₁ ^PP2A rat brain when compared with AAV1-GFP control group and treatment with memantine didn’t influence the level of PP2Ac. C) PP2A activity as determined by phosphatase ELISA was markedly decreased in all brain areas studied in AAV1-I₁ ^PP2A rats. Memantine only reversed pathologic change of PP2A activity in AAV1-I₁ ^PP2A rats, but didn’t influence the level of PP2A activity in AAV1-GFP rats. *P<0.05, **P<0.01 AAV1-I₁ ^PP2A vs. AAV1-GFP control rats; ^ΔP<0.05, ^ΔΔP<0.01, AAV1-I₁ ^PP2A with memantine vs. AAV1₁ ^PP2A vehicle. D) Western blots developed with anti-demethylated PP2Ac (4b7) showed that the demethylation of PP2Ac at Leu 309 in AAV1-I₁ ^PP2A rats with a chronic memantine treatment was markedly decreased as compared with vehicle-treated- I₁ ^PP2A rats. Memantine did not influence the demethylation of PP2Ac in AAV1-GFP rats. 3–4 rats were employed in each experiment. *P<0.05, **P<0.01 AAV1-I₁ ^PP2A vs. AAV1-GFP control rats; ^ΔP<0.05, ^ΔΔP<0.01 AAV1-I₁ ^PP2A-memantine vs. AAV1-I₁ ^PP2A-Vh.

Fig 3. In vitro inhibition of PP2A activity with I₁ ^PP2A is prevented by memantine.

A) The standard curve of PP2A activity assay showing linearity of the reaction up to 20 mu PP2A/assay of purified PP2A holoenzyme in dephosphorylation of tau phosphopeptide P17. B) Recombinant I₁ ^PP2A concentration-dependent inhibition of PP2A activity. C) 20 mu PP2A/assay purified PP2A and 4.8 μg/ml recombinant I₁ ^PP2A determined from A and B were employed to evaluate the effect of memantine on inhibition of PP2A activity by I₁ ^PP2A. Memantine, 2 μg/ml assay, completely prevented the inhibition of PP2A with I₁ ^PP2A. 3–4 rats were employed in each experiment.

Fig 4. Memantine inhibits I₁ ^PP2A-induced tau hyperphosphorylation.

A, B) Western blots showed increase in abnormal hyperphosphorylation of tau in AAV1-I₁ ^PP2A rats. Quantitative analysis of the blots of hyperphosphorylated tau normalized with total tau showed significant increase in tau hyperphosphorylation at several sites studied and memantine was found to reverse this change. *P<0.05, **P<0.01 AAV1-I₁ ^PP2A vs. AAV1-GFP rats; ^ΔP<0.05, ^ΔΔP<0.01 AAV1-I₁ ^PP2A-memantine vs. AAV1-I₁ ^PP2A-vehicle.

Fig 5. I₁ ^PP2A and hyperphosphorylated tau are colocalized and memantine reduces expression of hyperphosphorylated tau.

A,B) GFP served as a reporter protein for AAV1-induced expression. Strong immunofluorescence in the cerebral cortex (Aa-Ad and Am-Ap), the ventricular area (Ba-Bd and Bm-Bp) was observed. Tau was hyperphosphorylated at pT231/pS235 (M4), and pS262/pS356 (12e8) epitopes in the cerebral cortex (Ag, As), the ventricular area (Bg, Bs) of AAV1-I₁ ^PP2A, but not AAV1-GFP rats (Ae, Aq, Be and Bq). Double immunostaining showed that I₁ ^PP2A colocalized with hyperphosphorylated tau (Ak, Aw, Bk and Bw). Memantine inhibited tau hyperphosphorylation and phosphorylated tau distribution in the I₁ ^PP2A rats (Ah, At, Bh and Bt) while memantine didn’t influence the distribution and expression of I₁ ^PP2A (Ac, Ad, Ao, Ap and Bc, Bd, Bo, Bp). 3–4 rats were employed in each experiment. Scale bar in A and B, 50 μm.

Fig 6. Memantine inhibits I₁ ^PP2A-induced neurodegeneration.

A,B) As compared with AAV1-GFP rats, AAV1-I₁ ^PP2A rats showed a marked decrease in the expression of somatodendritic marker MAP2 and synaptic markers synapsin and synaptophysin in the hippocampus. Memantine treatment restored dendritic and synaptic loss. B) Western blots showed that β III tubulin staining was significantly decreased in the hippocampus of AAV1-I₁ ^PP2A rats as compared with AAV1-GFP control animals. Memantine restored the decrease to physiological level. C)Fluoro Jade staining showed a selective neurodegeneration in the AAV1-I₁ ^PP2A rat dentate gyrus of the hippocampus. Memantine prevented the neurodegeneration in the AAV1-I₁ ^PP2A rats. 3–4 rats were employed in each experiment. *P<0.05, **P<0.01 AAV1-I₁ ^PP2A vs. AAV1-GFP rats; ^ΔP<0.05, ^ΔΔP<0.01 AAV1-I₁ ^PP2A-memantine vs. AAV1-I₁ ^PP2A-vehicle. Scale bar in A and D, 100 μm.

Fig 7. Memantine decreases the level of activated GSK-3β and the expression of intraneuronal Aβ.

A) While the level of total GSK-3β was not changed, pSer-9 GSK-3β was markedly decreased in AAV1-I₁ ^PP2A as compared with the AAV1-GFP rats in hippocampus, cerebral cortex and ventricular areas. Memantine inhibited the phosphorylation of GSK-3β at Ser-9. *P<0.05 vs. AAV1-GFP control rats; ^ΔP<0.05, ^ΔΔP<0.01 vs. AAV1-I₁ ^PP2A rats. B) Immunohistochemical staining with rabbit polyclonal antibodies to Aβ_34–40 and to Aβ_36–42 showed increase in the intraneuronal expression of Aβ in AAV1-I₁ ^PP2A rats and memantine was found to reverse this change. 3–4 rats were employed in each experiment. Scale bar, 100 μm.

Fig 8. Memantine attenuates spatial learning and memory impairment in AAV1-I₁ ^PP2A rats.

A) Neurological examination did not reveal any significant difference among groups. B) There was no difference between groups to visit the center of the arena in open field, meaning that all animals displayed similar anxiety levels. C) During the twenty minutes of free exploration, all groups covered similar distance, suggesting that all animals exhibited similar level of exploration. D) In the water-maze task, all groups of animals swam at similar speeds. E) During the training of the water-maze task, AAV1-I₁ ^PP2A rats displayed delayed performance compared to other three groups and the treatment with memantine rescued this impairment. F, G) Western blots data showed that the level of activated phospho-CREB (Ser-133) in AAV1-I₁ ^PP2A rats was markedly decreased compared with AAV1-GFP control animals in all brain regions studied, while there was no significant change in the level of total CREB. Memantine treatment rescued the phosphorylation of CREB at Ser-133 in hippocampus in AAV1-I₁ ^PP2A rats. A significant increase of phosphorylation of pSer 133 CREB was found in the ventricular area, the cerebral cortex and the cerebellum in AAV1-I₁ ^PP2A rats after the memantine treatment. 8–10 rats were employed for behavioral tests. 3–4 rats were employed for Western blots. *P<0.05, **P<0.01 AAV1-I₁ ^PP2A vs. AAV1-GFP control rats; ^ΔP<0.05 AAV1-I₁ ^PP2A-memantine vs. AAV1-I₁ ^PP2A-vehicle rats.

Fig 9. Hypothetical scheme showing how I₁ ^PP2A inhibits PP2A activity and memantine rescues it.

A) I₁ ^PP2A on the one hand interacts with the catalytic subunit of PP2A (PP2Ac) and inhibits the PP2A activity [8]. On the other hand, I₁ ^PP2A probably upregulates the methylesterase PME, and/or downregulates the leucine carboxyl methyltransferase LCMT1 (broken line) and leads to an increase in demethylation of PP2Ac at Leu 309, resulting in the decrease of PP2A activity [22]. Demethylation of Leu 309 causes B subunit of PP2A to dissociate from C subunit of PP2A, further decreasing the PP2A activity. B) On the one hand, memantine may competitively bind to I₁ ^PP2A and rescue PP2A activity (broken line). On the other hand, as a non-competitive NMDA receptor antagonist with anti-oxidative property, memantine may inhibit the cytochrome P450, a demethylase and thereby decrease the demethylation of PP2A at Leu 309 and also lead to rescue of PP2A activity.