treated metformin intraperitoneally to another set of 8-weeks-old SAMP8/SAMR1. Results: The level of cortical p-tau 396 of 2- months-old SAMP8 was significantly lower than SAMR1, while p-tau 262 level of 2-months-old SAMP8 was comparable, and these levels increased upto 10 months-old SAMP8. The pattern of p- GSK3b expression showed a negative correlation with p-tau 396 . In the hippocampus, p-tau 396 and p-tau 262 were observed to in- crease as SAMP8 aged, as compared to the levels of age-matched SAMR1. The early activation of cortical AMPK of 2-months-old SAMP8 disappeared in 5-months- and 10-months-old mice, as compared to SAMR1; however, we could not observed the activa- tion of AMPK in the hippocampus of SAMP8 across all ages. The level of Sirt1 in the cortex of 2-months-old SAMP8 was signifi- cantly lower than SAMR1, which persisted until 5-months-old and disappeared at 10-months-old, compared to the levels of SAMR1. The level of insulin receptor substrate-1 (IRS-1) expres- sion in the cortex of SAMP8 was significantly lower than SAMR1 across all ages. In the hippocampus, the level of Sirt1, but not IRS-1, was lower in 2-months-old SAMP8 than SAMR1; however, the levels of Sirt1 in 10-months-old SAMP8 were compa- rable. However, when we treated metformin to activate AMPK, the AMPK-mediated regulation of p-tau 396 in SAMP8 was not clear. Conclusions: Our data showed that the differential and possibly in- direct regulation of tau phosphorylation by AMPK in cerebral cor- tex of SAMP8 AD model. However, further investigations to clarify precise mechanism of AMPK in tau phosphorylation and energy metabolism in the cerebral cortex are needed. P4-084 SUPPRESSION OF TAU INCREASE IN CEREBROSPINAL FLUID OF APP TRANSGENIC MICE PROVIDES EVIDENCE FOR DOWNSTREAM EFFECT OF BACE1 INHIBITION Stephan A. Kaeser 1 , Juliane Schelle 1,2 , Lisa Haesler 1 , Jens C. Goepfert 3 , Thomas O. Joos 3 , Hugo Marcel Vanderstichele 4 , Erik Stoops 4 , Ulf Neumann 5 , Derya Robert Shimshek 5 , Matthias Staufenbiel 1 , Mathias Jucker 1 , 1 German Center for Neurodegenerative Diseases, Hertie Institute for Clinical Brain Research, Tuebingen, Germany; 2 Graduate School of Cellular and Molecular Neuroscience, University of Tuebingen, Tuebingen, Germany; 3 Natural and Medical Sciences Institute at the University of Tuebingen, Reutlingen, Germany; 4 ADx NeuroSciences, Gent, Belgium; 5 Novartis Pharma AG, Basel, Switzerland. Contact e-mail: stephan.kaeser@uni-tuebingen.de Background: The best-validated and established fluid biomarkers for Alzheimer’s disease (AD) are reduced Abeta42 and increased tau levels in the cerebrospinal fluid (CSF). We have previously shown that this CSF signature is mirrored in aging Abeta precursor protein (APP) transgenic mice. Despite the lack of neurofibrillary tangles and global neuron loss, they exhibit higher CSF tau concen- trations with progressing plaque deposition, indicating that beta- amyloidosis is a major trigger for CSF tau increase. Moreover it suggests that APP transgenic mouse models can be instrumental to predict CSF biomarker changes in clinical trials directed against Abeta. One of the main therapeutic approaches tested for the treat- ment of AD is the inhibition of the beta-site amyloid precursor pro- tein-cleaving enzyme 1 (BACE1). Methods: We now used a potent BACE1 inhibitor to treat early-depositing APPPS1 and APP23 mice during six months and evaluated CSF tau changes. For the ac- curate quantification of murine tau we developed a novel high- sensitivity assay based on the single molecule array (Simoa) tech- nology. In addition, we measured Abeta42 and 40 in the CSF and brain tissue of the same mice and assessed their plaque load by ste- reological analysis on histological brain sections. Results: As ex- pected, long-term BACE1 inhibition attenuated amyloid deposition in both APP transgenic mouse models. Strikingly, the CSF tau increase, which is normally seen without treatment, was completely abolished. CSF Abeta42 and 40 were reduced compared to baseline levels, which was already evident after one week of BACE1 inhibition. CSF tau concentration however, was unaffected after short-term treatment, accentuating again its link to amyloid plaque load. Conclusions: The prevention of CSF tau increase in two different APP transgenic mouse models suggests that BACE1 inhibition also affects down-stream AD pathologies. The strong correlation between CSF tau and cerebral amyloid renders CSF tau a useful marker for the monitoring of the effectiveness of a BACE1 inhibitor in current clinical trials. S.A.K. and J.S. contrib- uted equally to this work. P4-086 TAU MODULATES BDNF EXPRESSION AND MEDIATES Ab-INDUCED BDNF DOWN- REGULATION IN ANIMAL AND CELLULAR MODELS OF ALZHEIMER’S DISEASE Elyse Rosa 1 , Sujeivan Mahendram 1 , Yazi D. Ke 2 , Lars M. Ittner 2 , Stephen D. Ginsberg 3 , Margaret Fahnestock 1 , 1 McMaster University, Hamilton, ON, Canada; 2 University of New South Wales, Sydney, Australia; 3 New York University Langone Medical Center, Orangeburg, NY, USA. Contact e-mail: rosae@mcmaster.ca Background: In Alzheimer’s disease (AD), soluble tau is hyper- phosphorylated, and some of this population aggregates and precip- itates as neurofibrillary tangles (NFTs). This pathological tau is believed to cause neurodegeneration, although a precise toxic mechanism is not well understood. Our previous work showed that brain-derived neurotrophic factor (BDNF), a protein required for neuronal survival and function, is lost in brain tissue from AD and non-AD tauopathies. Thus, we hypothesized that pathological tau down-regulates BDNF. Methods: We quantified BDNF mRNA from the cortex of transgenic mice (8c-het and hTau mice) over-ex- pressing wild-type human tau, as well as wild-type human tau (hTau40)-transfected human neuroblastoma (SH-SY5Y) cells, us- ing qRT-PCR. These models allowed us to examine the effect of excess tau on BDNF expression. Additionally, we quantified BDNF mRNA from amyloid-b over-expressing (APP23) mice and APP23xTauKO to determine if tau mediates amyloid-b induced BDNF down-regulation. Results: We found a significant reduction in BDNF mRNA in both 8c-het and hTau mice compared to non-transgenic control mice. The 8c-het mice exhibit increased tau phosphorylation but do not develop NFTs, while the hTau mice do exhibit NFT pathology, yet both models significantly down-regulate BDNF. We also found a significant reduction in BDNF and BDNF transcript IV in hTau40-transfected SH-SY5Y cells compared to controls. Lastly, we found that APP23 mice, which over-express Ab, have significantly reduced BDNF expres- sion. However, when crossed with Tau knockout (KO) mice, the re- sulting APP23xTauKO animals have cortical BDNF expression intermediate between APP23 and wild-type animals and not statis- tically different from wild-type animals. Conclusions: Our results demonstrate that excess wild-type tau can down-regulate BDNF, and that neither a mutation in tau nor NFTs are required for toxicity as measured by BDNF expression. Furthermore, the partial rescue of BDNF levels by tau knockout suggests that tau contributes to Ab-induced BDNF down-regulation. Thus, loss of BDNF may mediate tau neurotoxicity, which has profound implications for therapeutic intervention in AD and tauopathies. Poster Presentations: Wednesday, July 27, 2016 P1045