JNK: A Putative Link Between Insulin Signaling and VGLUT1 in Alzheimer’s Disease

The concept of dysfunctional insulin signaling in sporadic Alzheimer’s disease (AD) is now widely accepted and diabetes is recognized as one of the main risk factors for developing the disease. Early stages of AD are marked by metabolic dysfunction associated with altered brain insulin signaling and significant abnormalities in insulin-regulated gene expression and kinase activation. Postmortem AD brains showed decreased insulin and re-ceptor activities in proportion to the severity of disease [1–3].

The initial components of the insulin receptor (IR) signaling cascade in the brain are largely similar to those of the periphery, but the downstream targets of the cascade are quite different, probablyinvolving, among others, the glutamatergic system [4]. An essential step in glutamate neurotransmission is the concentration of glutamate into synaptic vesicles by vesicular glutamate transporters (VGLUTs) before release from the presynaptic terminal, and, in fact, VGLUT1 is considered as a marker of glutamatergic terminals. Some lines of evidence indicated that the decreased VGLUT1 expression in frontal regions of AD patients is strongly correlated with the progression of cognitive decline inAD [5].

c-Jun N-terminal kinases (JNKs), as member of the mitogen-activated protein kinase (MAPK) family, is a central stress signaling pathway implicated in gene expression, neuronal plasticity, regeneration, cell death, and regulation of cellular senescence. It has been shown that there is a strong activation of JNK pathway after exposure to different stressing factors including Aβ peptides [6–8]. JNK may also directly induce insulin resistance [9].

The hypothesis of the present work is that activation of JNK in AD might inhibit insulin signaling which in turn, might lead to a decreased expression of VGLUT1 and therefore, to the glutamatergic deficit describedin AD.

Brain tissues were obtained from the Oxford Project to Investigate Memory and Ageing (OPTIMA, see http://www.medsci.ox.ac.uk/optima). Subjects for this study constituted a randomly selected subset of the participants, now part of the Thomas Willis Oxford Brain Collection within the Brains for Dementia Research Initiative (BDR). Subjects had been assessed annually with the Mini-Mental State Examination (MMSE). Postmortem brain tissues used in this study consisted of frontal (Brodmann area 10, BA10) cortex. The inclusion/exclusion criteria and point-of-entry characteristics of the participants have previously been described in detail [10]. The study had Local Ethics Committees’ approval. All subjects fulfilled CERAD criteria for the neuropathological diagnosis of AD and were staged at Braak V/VI. Controls did not have dementia or other neurological diseases, did not meet CERAD criteria for AD diagnosis, and were staged at Braak 0-II. Brain pH has been used as an indication of tissue quality in postmortem research, with pH >6.1 considered acceptable.

All the experiments in animals were carried out in strict compliance with the recommendations of the EU (DOCE L 358/1 18/2/1986) for the care and use of laboratory animals. Male mice (C57BL/6J), 3 months old, were used. The following experimental groups were used (n = 9 mice per group): saline, corticosterone, JNK inhibitor D-JNKi, corticosterone+D-JNKi. In the corticosterone groups, mice received for 4 weeks drinking water with a solution containing 100μg/ml corticosterone (Sigma-Aldrich, St. Louis, MO). D-JNKi1 (0.3 mg/kg; Enzo Life Sciences, New York, USA) was dissolved in saline and administered (i.p.) once a day during the whole corticosterone treatment. All efforts were made to minimize animal suffering, to reduce the number of animals used, and to utilize alternatives to in vivo techniques, if available.

Total mRNA was extracted from human postmortem brain samples according to the instructions of NucleoSpin® RNA II kit (Macherey-Nagel, Germany). DNAase treatment was performed with DNA-free kit (Ambion, TX, USA), and purified total RNA used as a template to generate first-strand cDNA synthesis using M-MLV reverse transcriptase (Invitrogen, CA, USA) as described by the manufacturer.Quantitative real-time PCR was performed as described by the provider (Applied Biosystems, CA, USA) using an ABI PRISM 7000 HT Sequence Detection System. Taqman probes were supplied by Applied Biosystems. Gene expression levels IR were normalized using GAPDH as internal control. Fold change between different groups of brains were calculated using the 2^–ΔΔCt method.

Western blotting experiments: tissues were homogenized in 10 volumes of lysis buffer containing (mmol/L): 50 Tris–HCl (pH 8), 150 mM NaCl, 2 mM EDTA, 2 mM EGTA, 1% Nonidet P-40, 1:100 of phosphatases and proteases inhibitors cocktail set II (Calbiochem, Darmstadt, Germany). Proteins (10μg for VGLUT and 50μg for (p)Akt, (p)JNK and (p)ERK2) were separated by electrophoresis on a SDS–polyacrylamide gel (7.5% ) and detected using the following antibodies: anti-pAkt Ser 473 and total Akt (Cell Signalling Technology, Beverly, MA. USA), anti-p42/p44 MAPK (ERK1/2) and total ERK1/2 (Cell Signalling Technology) and anti-VGLUT1 (donated by Dr. S. El Mestikawy, Paris, France) and anti-pSAPK/JNK and total SAPK/JNK (Cell Signalling Technology). Immunopositive bands were visualized using an enhanced chemiluminescense western blotting-detection reagent (ECL; Amersham, Buckinghamsire, England). The optical density (O.D.) of reactive bands visible on X-ray film was determined densitometrically. β-actin was used as internal control. Results were expressed as percentage of O.D. values.

Data were analyzed by SPSS for Windows, release 15.0 and normality was checked by Shapiro–Wilks’s test (p <  0.05). Data was analyzed by two-way ANOVA or Student’s t-test. Correlation studies between variables were studied by Pearson’s correlation coefficient.

The total number of cases analyzed was 15 controls (7 males/8 females) and 16 AD (6 males/10 females). There were no significant differences between controls and AD regarding age at death was (73 ± 3 years in controls versus 81 ± 2 in AD cases), postmortem delay (40.1 ± 6.4 h in controls versus 49.8 ± 7.4 h in AD), or brain pH (6.6 ± 0.1 versus 6.4 ± 0.1 Student’s t-test p >  0.05, in all cases). In addition, there were no significant correlations between age, postmortem delay, or brain pH and any of the neurochemical variables studied in either control patients or those with dementia (p >  0.05). Average MMSE scores in AD was 5 ± 1.

In AD postmortem samples, levels of IR mRNA (Student’s t-test; p <  0.01; Supplementary Figure 1a), and markers of downstream signaling pathways (pAkt and pERK2, Student’s t-test p <  0.01 and p <  0.05 respectively, Fig. 1a and b) were significantly decreased. There was a significant positive correlation between pAkt or pERK2 levels and cognitive deficits in AD, as shown by MMSE score (Pearson’s, r = 0.596; p <  0.01; and Pearson’s, r = 0.875; p <  0.01, respectively; Supplementary Figure 1b and c). pJNK levels were significantly increased in AD (Student’s t-test; p <  0.001, Fig. 1c). There was a very strong trend toward significant negative correlations in AD between IR mRNA levels and pJNK expression (Pearson’s, r = –0.522; p = 0.06; n = 16). As depicted in Fig. 1d, VGLUT1 expression was reduced in AD (Student’s t-test; p <  0.01), and a significant positive correlation was observed between IR mRNA levels and VGLUT1 expression (Pearson’s, r = 0.729; p <  0.01; n = 16)

As previously reported [11], corticosterone treatment led to cognitive deficits, increased pJNK expression, and central insulin resistance, as decreased levels of insulin receptor phosphorylation, the phosphorylation status of the adapter protein insulin receptor substrate 1 (IRS1) at the active site (Ser636/639) and downstream signaling markers (pAkt and pERK) were observed, and all these effects were reversed by treatment with D-JNKi [11]. Using this same cohort of animals, it was presently found that corticosterone treatment induced a significant decrease in hippocampal VGLUT1 levels [significant interaction, F_1,35 = 4.105, p <  0.05] that was reversed by D-JNKi1 (Fig. 2).

In the present work, it has been corroborated that postmortem AD brain shows signs of altered insulin signaling, including reduced brain IR expression and hypoactivity of the IR downstream pathways. It is increasingly accepted that increased accumulation of Aβ oligomers and their binding to synapses lead to removal of IRs from the neuronal plasma membrane [12]. Altered IR downstream pathways could be related to cognitive deficits in AD, as activation of the insulin signal transduction cascade (Akt/ERK) is required for the induction of long-term potentiation, basic process underlying learning and memory [13]. Activation of JNK might be related to alterations in insulin pathways in AD insulin resistance through inhibition of IRS-1 [12], and suppression of the JNK pathway has been shown to improve insulin resistance and glucose tolerance [14]. Supporting this idea, it has been shown that the administration of a JNK inhibitor reverses insulin resistance and cognitive deficits in a model of chronic corticosterone administration [11].

Impairments in insulin signaling might lead to deficits in energy metabolism shifting brain metabolic profile from glucose-driven bioenergetics towards a compensatory, but less efficient, ketogenic pathway that seems to occur in early AD [1–15, 16]. Interestingly, it has been suggested that excessive ketone bodies levels regulates VGLUTs activity [17]. Therefore, it is possible to speculate that deficiencies in insulin signaling could switch neuronal metabolism to produce ketone bodies, which in turn, might lead to a decreased expression of VGLUT1, and therefore to a decreased release of glutamate and hence, to the glutamatergic deficit described in AD. Inhibition of JNK, by reversing insulin alterations, has presently proven to reverse also glutamatergic deficiencies, supporting our hypothesis that JNK may serve as a crucial link between metabolic diseases and glutamatergic deficits.

Nowadays Memantine is commonly prescribed to AD patients to reduce glutamate-induced excitotoxicity mediated by excessive activation of NMDA-receptors which leads to deficient synaptic functioning. It has been claimed that memantine, preferentially blocks excessive NMDA receptor activity without disrupting normal activity [18]. Therefore,it is possible to speculate that initially, alterations in insulin signaling could contribute to decrease glutamate release (present results) which in turn develops into deterioration of proper synaptic functioning.

Overall, the results of the present project could help not only to a better understanding of the common mechanisms that could link and underlie both to diabetes and AD, but also to consider new therapeutic options. Modulation of VGLUT1 activity through modulating pJNK may be considered as a possible therapeutic target for the treatment of metabolic disturbances in AD.