WO2011067420A1

WO2011067420A1 - Arf6 as a new target for treating alzheimer's disease

Info

Publication number: WO2011067420A1
Application number: PCT/EP2010/069000
Authority: WO
Inventors: Wim Annaert; Ragna Sannerud; Katrijn Coen; Bart De Strooper
Original assignee: Vib Vzw; Katholieke Universiteit Leuven, K.U.Leuven R&D
Priority date: 2009-12-04
Filing date: 2010-12-06
Publication date: 2011-06-09
Also published as: CA2782459A1; US8790887B2; EP2506869A1; US20120276076A1

Abstract

The present invention relates to the field of disorders of the peripheral or central nervous system, in particular Alzheimer's disease, and the prevention and/or treatment thereof. In particular, the present invention relates to ARF6 and/or ARF6 effector proteins as new targets in Alzheimer's disease, and based thereon screening methods for compounds that reduce amyloid beta peptide formation in mammalian cells by affecting ARF6-mediated endosomal sorting.

Description

ARF6 as a new target for treating Alzheimer's disease

FIELD OF THE INVENTION

The present invention relates to the field of disorders of the peripheral or central nervous system, in particular Alzheimer's disease, and the prevention and/or treatment thereof. In particular, the present invention relates to A F6 and/or ARF6 effector proteins as new targets in Alzheimer's disease, and based thereon screening methods for compounds that reduce amyloid beta peptide formation in mammalian cells. The present invention also relates to inhibiting and/or activating agents targeting ARF6 and pharmaceutical compositions thereof, and their use in therapeutic applications of said disorders. The present invention also relates to restoring dysfunction in endosomal sorting and degradation seen as a very early pathological feature in Alzheimer's disease. Finally, the present invention also relates to a cell line and the use thereof for identifying compounds blocking ARF6- mediated endosomal sorting.

BACKGROUND

Alzheimer's disease (AD) is the most common neurodegenerative disorder afflicting the elderly. AD is clinically characterized by progressive neuronal loss and inflammation, memory impairment, cognitive deficits, and behavioral changes. Neuropathologically, the AD brain is characterized by two proteinaceous aggregates, amyloid plaques, mainly composed of the amyloid β-protein (Αβ), and neurofibrillary tangles (NFT), comprised of hyperphosphorylated aggregates of the tau protein (Selkoe 2001). Two major hypotheses have driven pharmaceutical research in the search for a medication for AD: the amyloid hypothesis (Hardy and Selkoe 2002) and the cholinergic hypothesis (Bartus et al. 1982). Although significant progress has been made toward understanding the pathophysiology of AD, significant questions remain unanswered, e.g. the potential link between amyloid pathology and the cholinergic deficit observed in AD patients and the relationship between Αβ generation, neuronal cell death, and NFTs.

Αβ is derived from proteolysis of the β-amyloid precursor protein (APP), a type I integral membrane protein, following sequential cleavage by th e β- (BACE1) a n d γ-secretases. The γ-secretase is a tetrameric complex that cleaves APP within its transmembrane domain, thereby liberating the intact Αβ peptide, which ranges in length from 39-43 residues (De Strooper et al. 2003). The majority of Αβ produced is 40 amino acids in length (Αβ₄₀), whereas a small proportion (~10%) is the 42-residue variant (Αβ₄₂). Αβ₄₂ is more hydrophobic, aggregates much faster than Αβ₄₀, is more toxic than Αβ_40, and is the major Αβ species found in cerebral plaques (Selkoe 2001; Iwatsubo 1994).

Despite intensive research during the last 100 years, prognosis of AD patients now is still quite the same as that of patients a century ago, since there is still no real cure available. There are two types of drugs approved by the U.S. Food and Drug Administration and used in clinic today to treat AD: Acetylcholinesterase (AchE) inhibitors and Memantine. There is ample evidence in the art that the amyloid beta peptide, the main component of the amyloid plaques that are specific to the AD etiology, has a key role in the development of AD disease (Hardy et al. 2002, Golde et al. 2006). Therefore, one of the most common strategies to lower Αβ is to diminish its production by γ- and β-secretase inhibition. One strategy was the development of gamma-secretase inhibitors however such inhibitors often result in serious side effects since gamma-secretase is involved in the proteolytic processing of at least 30 proteins (De Strooper et al. 2003). Yet another attractive strategy is the development of β- secretase (BACE1) inhibitors, as BACE1 knock-out mice are viable and have no obvious pathological phenotype (e.g. oberds et al. 2001, Ohno et al. 2004, Ohno et al. 2006).

Yet, there still is a continuous need for alternative approaches to lower Αβ in view of treating patients with neurodegenerative disorders, such as Alzheimer's disease.

SUMMARY OF THE INVENTION

There is increasing evidence that endosomes constitute a major site where Αβ peptides are produced. Αβ peptide, the primary constituents of senile plaques which are a hall mark in the pathology of Alzheimer's disease, is generated by the sequential cleavage of the amyloid precursor protein (APP) by BACE1 and γ-secretase. U nderstand ing BACE 1 intracel l ular sorting a nd trafficking is of major importance for the development of therapeutic inhibition of Αβ production as it is the enzyme responsible for the generation of this peptide.

It was surprisingly found that BACE1 enters the cell via a distinct route than APP, more specifically, BACE1 enter the cells via a clathrin-independent ARF6-mediated pathway before reaching the Rab5- positive endosome, whereas APP is internalized via a clathrin-dependent pathway. ARF6 is a small GTPase that regulates the trafficking of endosomal membrane (D'Souza-Schorey and Schavrier 2006). It was found that ARF6 cycling activity influences the processing of APP. It was demonstrated that blocking BACE 1 in the ARF6 vacuoles prevents BACE1 reaching the Rab5-positive endosomes and thereby prevents processing of APP, supporting that shedding of APP occurs in the endosomal compartments. These findings highlight the importance of the (early) endosomal compartment as the major site of BACE1 processing and open novel avenues for interfering with Αβ production through selectively interfering with the distinct internalization and endosomal recycling routes.

So, the present invention relates to ARF6 and/or ARF6 effector proteins as new targets in Alzheimer's disease, since it was surprisingly found that by modulating the activity of ARF6, the formation of amyloid beta peptides in mammalian cells can be significantly reduced.

A first aspect of the present invention relates to a method for reducing amyloid beta peptide formation in a mammalian cell comprising modulating the ARF6 cycling activity and/or ARF6 effector protein activity. Accordingly, the invention relates to the use of ARF6 and/or an ARF6 effector to reduce amyloid beta peptide formation in a mammalian cell.

A second aspect of the present invention relates to a method for identifying compounds that reduce amyloid beta peptide formation in a mammalian cell comprising modulating the ARF6 cycling activity and/or ARF6 effector activity. Accordingly, the invention relates to the use of ARF6 and/or an ARF6 effector to identify compounds that reduce amyloid beta peptide formation in a mammalian cell.

In another aspect, the present invention relates to an agent such as an antisense polynucleotide, a ribozyme, or a small interfering RNA (siRNA), characterized in that it is reducing amyloid beta peptide formation, and wherein said agent comprises a nucleic acid sequence complementary to, or engineered from, an ARF6 encoding polynucleotide.

In still another aspect, the invention relates to a pharmaceutical composition comprising a therapeutically effective amount of any of the above described agents and at least one of a pharmaceutically acceptable carrier, adjuvant or diluents.

A further aspect of the present invention relates to a cell line characterized by lacking endogenous presenilin expression or function (PSENdKO cell line) and the same cell line stable expressing virally transduced ARF6

In particular, the PSENdKO cell line and the PSENdKO cell line expressing ARF6 can be used in an assay for screening compounds that are capable of modulating ARF6 protein activity in a mammalian cell. Preferably, said compounds are capable of reducing amyloid beta peptide formation in a mammalian cell. Even more preferably, said compounds are therapeutic candidates for the prevention and/or treatment of a disorder of the peripheral or central nervous system, in particular Alzheimer's disease. In particular, said cell lines are MEF cell lines.

A particular embodiment of the present invention relates to a method for identifying a compound that modulates the endosomal redistribution in a mammalian cell based on the imaging of one or more morphological parameters of cells from a cell suspension culture of the above described PSENdKO cell line in the presence of a compound and comparing, under the same conditions, to the same morphological parameter(s) of cells of the corresponding wild type and/or PSENdKO cell line expressing ARF6.

Objects of the present invention will be clear from the description that follows.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1: (a,b) BACE1 localize to ARF6-positive membrane. HeLa cells co-transfected with BACE1 and HA- ARF6-Q67L (a) or HA- ARF6-T27N (b) were fixed and stained for HA, BACE1 and MHCI. (c-e) BACE1 localized with the ARF6-TBC1D3-GGA3 complex at the cell surface, (c) HeLa cells transfected with either (c) FLAG-tagged GGA3 and BACE1, or (d) FLAG-tagged-GGA3, BACE1 and HA-TBC1D3, or (e) FLAG-GGA3, BACE1 and HA- ARF6-Q67L, were fixed and stained for the tags FLAG and HA, and BACE1 as required, (c) Arrow indicates colocalization of GGA3 and BACE1 at the plasma membrane, in the inset arrowheads indicate coloca lization in the endosoma l compartment, (d ) Arrows ind icate colocalization of BACE1 with GGA3 and TBC1D3. (a-e) Magnifications of selected areas, indicated by a square, are shown in insets, arrowheads highlight colocalization. To facilitate visualization, black and white picture for each channel and pseudo colours as well as two merges if three channels, were used. Bars= 10 μιτι.

Figure 2: BACE1, but not APP, internalize via a clathrin-independent pathway, (a) APP endocytosis is not mediated by ARF6. HeLa cells cotransfected with APP and HA-ARF6-Q67L were fixed after 24h and stained for the tag HA and APP. Not that APP is clearly absent from the vacuoles created by overexpression of ARF6-Q67L. Magnification of this region is shown in the inset. (b,c) Internalization of

BACE1 ( b) and APP (c) together with CD59 anti bodies were performed in HeLa cel ls transiently expressing BACE1 or APP accordingly (24h). After 10 min at 37°C, with EGF (200 ng/ml) stimulation, cells were fixed on ice. Before permea bilization, cel ls were incu bated with pacific blue-labelled secondary antibody to visualize the cell surface, then after permeabilization 488- and 568-labbeled secondary antibodies were used to stain internalized primary antibodies of BACE1 (b) or APP (c) and

CD59, respectively. Cells kept on ice (0 min) do not show internalization of CD59, BACE1 or APP. Right panels show magnifications of selected, (d) HeLa cells were cotransfected with BACE1 and empty vector or with BACE1 and a truncated form of AP180 (AP180-Cterm). 24 h later the cells were labelled with EZ-Link Sulfo-NHS-SS-Biotin at 4°C for 15 min. After washes the cells were incubated at 37°C for 10 min allowing endocytosis. Remaining biotin at the cell surface was reduced before lysing the cells and internalized biotin were pooled down and detected as indicated in materials and methods. Westernblot results of a representative experiment is shown in (d) left panel and quantification of the signal of three independent experiments was generated for BACEl, transferrin receptor (TfR) and APP (d, right panel) (* * *, p<0,l%). (e) BACEl is trapped in ARF6-Q67L vacuoles in neuronal cell. Rat hypocampal neurones were co-transfected with ARF6-Q67L and BACEl 6 days after plating. 24hrs later the cells were fixed and stained for tag HA, BACEl and APP. Two merges are shown on the right hand side to highlight BACEl (red) trapped in the ARF6-Q67L (green) vacuoles and that BACEl (green) and APP do not colocalized in these vacuoles.

Figure 3: (a, b) BACEl internalizes via an ARF6-mediated route before reaching the RAB5 early endosome. (a) HeLa cells were cotransfected with RAB5-Q67L and BACEl, fixed and stained for EEAl(red) and BACEl (green). Inset shows a magnification of a selected area indicated by a square in the merge image, to highlight that BACEl and EEA1 are labelling distinct domain in RAB5-enlarged endosome. Note that BACEl is present only in a subset of RAB5-enlarged endosomes labelled with EEA1. (b) HeLa cells were cotransfected with cerulean-RAB5-Q79L, HA-tagged ARF6-Q67L and BACEl, fixed after 24h and stained using antibodies against HA and BACEl. Pseudo-colour was used for cerulean- RAB5-Q79L to facilitate visualization and colocalization. RAB5-enlarged endosomes are devoid of BACEl as shown in the inset picture which represent only BACEl (red) and cerulean- RAB5- Q79L (green), (c, d) ARF6-Q67L overexpression prevents BACEl from reaching APP, while both colocalize in ARF6-T27N compartments. HeLa cells transfected with BACEl and APP, and with either HA- ARF6-Q67L (a) or HA- ARF6-T27N (b), were fixed and stained for BACEl, APP and HA. Magnification of a selected area is indicated by a square in the merge image. Bar= 10 μιτι.

Figure 4: APP processing is affected by overexpression of the ARF6 mutants, (a) HeLa cells cotransfected with APP and with ARF6-wt, ARF6-Q67L or ARF6-T27N were lysed and total protein (20μg per lane) were analysed by westernblot, shown on the left and quantification of APP processing was estimated by dividing APP-CTF fragment to APP full length for each ARF6 construct, right panel, (b, d) Modulating ARF6 cycle activity affect APP processing, (b, c) HeLa cells were transfected with swAPP together with either empty vector or ARF6- WT or ARF6-Q67L or ARF6-T27N or EFA6-WT or EFA6-mut or ACAPl-wt or ACAPl-mut and fu rther process for meta bol ic la bel l ing experiment 24hrs later as described in materials and methods. Means of ratio A0/APP-FL were calculated for each transfection and normalized to the control sample (swAPP with empty vector, * * *, p<0, l%). (d) 24hrs after ARF6 down regulation cells were transfected with swAPP and 24hrs later cells were processed for metabolic la bell ing experiment. ARF6 was ~80% down regulated after 48hrs (inset). NS= non specific sequence RNAi (*, p < 5%). Figure 5: Differential sorting of BACEl and APP. Schematic representation of BACEl and APP trafficking within the endosomal sytem in cells overexpressing RAB5-Q79L (a), ARF6-Q67L (b) and ARF6-T27N (c). (a) Cells overexpressing RAB5-Q79L develop enlarged endosomes where APP and BACEl accumulate promoting processing, (b) Overexpression of ARF6-Q67L blocks the fusion of ARF6 vesicles to endosome ending up with a massive accumulation of vacuoles which trap BACEl preventing its trafficking to the endosomes and therefore cleavage of APP. (c) Blocking ARF6-dependent recycling promote APP processing.

Figure 6: Adhesion-dependent raft endocytosis towards the perinuclear recycling compartment is not affected by presenilin deficiency. Stably adherent fibroblasts were plasma membrane-labeled with CTxB-AlexaFluor488 on ice, and brought in suspension for the indicated times. Cells were kept at 37°C during the suspension phase, and after fixation immunofluorescently labeled for caveolinl. The central region is enlarged in the left panel. In both, WT and PSENl&2dKO MEFs raft markers are localized in the perinuclear region of the cell after 30min in suspension.

Figure 7: Redistribution of raft components from the recycling compartment are blocked in PSENl&2dKO MEFs. After lh in suspension, fibroblasts were replated on a fibronectin-coated surface (2C^g/ml) for lh. In WT MEFs the raft markers CTxB and caveolinl redistribute back to the plasma membrane in a polarized manner, while in PSENl&2dKO cells both remained largely in the perinuclear region. Caveolinl distribution is much more fragmented in fine spots compared to WT cells and does not reach the plasmamembrane at all.

Figure 8: ARF6 activity is deregulated at the mRNA level in presenlin deficient fibroblasts, (a) WT and PSENKO MEFs were lysed, and active ARF6 was pulled down on GST-GGA3 beads. Representative western blots of bound samples (ARF6-GTP) and the respective whole cell lysates (total ARF6) show a clear decrease in ARF6 activity in all PSKO cell lines. Blots from at least 3 independent experiments were quanified, and normalized intensities were calculated relative to WT MEFs for total ARF6 (b), and the ratio of ARF6-GTP to total ARF6 (c). (d) RTqPCR results indicate that the reduction in ARF6 activity was partially caused by a decrease in mRNA levels in PSENKO cells. (N=3-7; meansiSEM; * P<0.05, ** P<0.01, *** P<0.005)

Figure 9: Confocal microscopy of MEF cells immunolabeled for endogenous caveolinl (Green), β-actin (Red), and a-tubulin (Blue) reveals in more detail the aberrant morphology of PSENl&2dKO MEFs. In WT cells, caveolinl shows a polarized plasma membrane localization at one edge of the cells. However, immunoreactivity of caveolinl in PSENl&2dKOs is clustered intracellular, with only very little plasma membrane staining left. PSENl&2dK0 revert to WT morphologies upon retroviral transduction with hARF6-WT, including a polarized redistribution of caveolinl to the plasma membrane

Figure 10: Expression of human ARF6 'rescues' increased Racl and EGFR levels in PSENl&2dKO MEFs. (A) Western blot analysis of PSEN l&2dKO MEFs retrovirally infected with hARF6 WT-HA or hARf6 T157A-HA shows that Racl and EGFR levels decrease dependent on the amount of ARF6 expression (quantified in B). No differences were seen for the stability of the different γ-secretase components, nor for APP processing. Therefore, we can postulate that the increase in Racl levels and EGFR levels in PSENl&2dKO MEFs is independent of γ-secretase activity.

Figure 11: Retroviral expression of ARF6 in PSENl&2dKO MEFs reduces wound healing significantly. (A) A scratch assay was performed on gelatin-coated surfaces under serum free conditions. Cells were stained with ΙμΜ calcein-AM for lh before pictures were taken, either immediately after scratching or 15h later. ( B) The percentage of the healed wounds was quantified for WT, PSENl&2dKO, and retrovirally infected PSENl&2dKO MEFs. Rescue was performed using retroviral transduction of human ARF6 WT-HA and ARF6 T157A-HA constructs. (N=3; meansiSEM; * P<0.05, ** P<0.01, *** P<0.005)

Figure 12: ARF6 is involved in the decreased turnover of degradative vacuoles in PSENl&2dKO MEFs. PSEN deficiency leads to the overall accumulation of lysotracker-positive vesicles representing late endosomes or lysosomes. These acidic compartments are accompanied by the accumulation of EGFR in PSENl&2dKO MEFs as shown by a co-distribution with lysotracker and anti-Lampl. PSENl&2dKO MEFs retrovirally transduced with human ARF6 WT-HA induce the clearance of these acidified organelles together with an increased degradation of EGFR, resulting in much lower anti-EGFR fluorescence intensity.

Figure 13: Morphology screen assay.

Figure 14: Endogenous ARF6 levels decrease with aging. Cortices from mice with ages ranging from prenatal E14 to postnatal 24 months were dissected out, homogenized and extracted in Tris-EDTA buffer (pH 7.3) containing 250mM sucrose, 1% TritonX-100 and protease inhibitors. Protein concentration of cleared extracts was measured and equal amounts of protein (2C^g) were loaded on precasted 4-12% M ES S DS-PAGEgels (Invitrogen) and following electrophoresis transferred on a nitrocellulose membrane (protran). These blots were blocked and subsequently probed for RAB5, ARF6 and presenilin 1 (PS1) using appropriate primary antibodies. Immunodetection was performed with HRP-conjugated secondary antibodies followed by ECL. Quantitave westernblotting was performed on a FujiLas Mini (acquisition) and data were processed using Lais software (Fuji). Protein concentration and GAPDH were used for normalization between the different experiments. Mean ± SEM of two to three independent experiments is shown.

Figure 15: ARF6 levels are downregulated in the brain of Alzheimer's disease (AD) patients. Samples of the frontal cortex of human control (18) and AD (47) brains were extracted in Tris-EDTA buffer (pH 7.3) containing 250m M sucrose, l%Triton X-100 a nd protease inh ibitors. After measuring protein concentration, equal protein amounts (3C^g) of cleared extracts were processed for SDS-PAGE and westernblotting. Blots were pro bed for RAB5, AR F6 a nd GGA3 using the a ppropriate primary antibodies and quantified. GAPDH was used for normalization. No statistically differences were found for RAB5 between control and AD brain. In the contrary, in AD brain, ARF6 and GGA3 levels dropped 42% and 28%, respectively, compared to control brain.

DETAILED DESCRIPTION OF THE INVENTION

The terms "a myl oid beta peptide" or "a mylo id beta protei n" or "Αβ pe ptide" o r "Αβ" a re interchangeably used further herein. Amyloid beta peptides are processed from the amyloid beta precursor protein (APP) and include the amyloid beta peptides 1-42, 1-40, 11-42, 11-40 which can ultimately be found in plaques and are often seen in cerebral spinal fluid.

The terms "modulating", "modulation", "modulated" means an up-regulation or down-regulation of the expression, or an increase or decrease in activity of a protein. Modulation of a protein includes the up-regulation, down-regulation, increase or decrease in activity of a protein or compound that regulates a protein. Modulation also includes the regulation of a gene, the mRNA, or any other step in the synthesis of the protein of interest.

The terms "protein", "polypeptide", "peptide" are interchangeably used further herein. The ADP-ribosylation factor (ARF) family of proteins belongs to the Ras superfamily of small GTPases that regulate vesicular trafficking and organelle structure by recruiting coat proteins, regulating phospholipid metabolism and modulating the structure of actin at membrane surfaces. ARF1 and ARF6, two of the best characterized ARF proteins, provide a molecular context for ARF protein function in fundamential biological processes, such as secretion, endocytosis, phagocytosis, cytokinesis, cell adhesion and tumour-cell invasion (D'Souza-Schorey and Schavrier 2006). Like other Ras-related GTP binding proteins, the ARF proteins cycle between their active-GTP-bound and inactive-GDP-bound conformations. Hydrolysis of bound GTP is mediated by GTPase-activating proteins (GAPs), whereas the exchange of GDP for tri-phosphate nucleotide is mediated by guanine nucleotide-exchange factors (GEFs). Several ARF-specific G EFs and GAPs that interact with one or more ARF proteins have been identified in vitro, however, distinct GAPs and GEFs regulate GTP-GDP cycle of individual ARF proteins in vivo (Jackson and Casanova, 2000; Randazzo et al. 2004). The ARF proteins are ubiquitously expressed and the amino-acid sequences seem to be well conserved in all eukaryotes, from yeast to humans, with remarkable fidelity.

ARF6 is the least conserved member of the ARF family of proteins and shares only 66% amino acid identity with Arfl. ARF6 regulates the trafficking of endosomal membrane and structural organization at the cell surface. The polynucleotide and amino acid sequence of the human ARF6 protein are accessible in public databases by accession numbers CR541939 and CAG46737, respectively. The ARF6 protein as referred to in the present invention also includes homologues as well as active fragments of the full length ARF6 polypeptide. "Active", with respect to the ARF6 polypeptide, refers to those forms, fragments or domains of a ARF6 polypeptide which retain the biological and/or antigenic activity of a ARF6 polypeptide.

The term "ARF6 cycling" as used herein means the ARF6 cycling between its active-GTP-bound and inactive-GDP-bound state. The term "ARF6 effector protein" in the context of the present invention means any protein that directly or indirectly interacts with ARF6 or which intervenes in the ARF6 endocytotic pathway, either upstream or downstream of the ARF6 signaling pathway.

In a particular embodiment of any of the above methods said ARF6 effector is chosen from a GAP, such as GIT-1, GIT-2, Centaurin a or a G EF, such as EFA6A, EFA6C, EFA6D, ARNO, msec7-l, GRP1, ARF- GEP100 (Jaworski 2007). Preferably, said GAPs and GEFs are predominantly expressed in the nervous system. Other effectors linked to the ARF6 pathway include, but are not limited to, phospholipase D2 (PLD2), PI(4)P5-kinase, other GTPases such as Ras or Rac, Jun N-terminal kinase interacting protein 3 and -4 (JIP3 &4), other Rabs such as Rabll, Rab22A, RablO, Rab35 (reviewed in D'Souza-Schorey and Chavrier 2006). ARF6 functioning at the cell surface can be mediated by its effect on phospholipid metabolism. In this regard, A F6 activates phospholipase D2 (PLD2) leading to the formation of lysophosphatic acid and diacylglycerol which are important for endosomal recycling via this route. ARF6 also activates PI(4)P5-kinase, generating PI(4,5)P2 with a similar function as above. ARF6 is also activated via activation of the ERK pathway or via activities of other GPTases such as Ras or Rac. ARF6 also binds to Jun N-terminal kinase interacting protein 3 and -4 (J IP3 &4) for targeting the recycling endosomal membrane to the cell surface. Other Rabs that interfere in this route are Rabll, Rab22A, RablO, Rab35.

The term "compound" is used herein in the context of a "test compound" or a "drug candidate compound" described in connection with the methods of the present invention. As such, these compounds comprise organic or inorganic compounds, derived synthetically or from natural resources. The compounds include polynucleotides, lipids or hormone analogs that are characterized by low molecular weights. Other biopolymeric organic test compounds include small peptides or peptide-like molecules (peptidomimetics) comprising from about 2 to about 40 amino acids and larger polypeptides comprising from about 40 to about 500 amino acids, such as antibodies or antibody conjugates.

Examples of assay methods for identifying compounds in the context of the present invention are described in the Example section, without the purpose of being limitative. It should be clear to the skilled artisan that the present screening methods might be based on a combination or a series of measurements, particularly when establishing the link with amyloid beta peptide generation. Also, it should be clear that there is no specific order in performing these measurements while practicing the present invention.

For high-throughput purposes, compound libraries may be used. Examples include, but are not limited to, natural compound libraries, allosteric compound libraries, peptide libraries, antibody fragment libraries, synthetic compound libraries, etc.

Determining the level of amyloid beta peptides produced can be done by using specific ELISAs using antibodies specifically recognizing the different amyloid beta peptides species or by identifying amyloid beta peptides and other APP processing products (APP-carboxyterminal fragments, ectodomains) following metabolic labeling. A reduction in level of amyloid beta peptides formed is preferably by at least 5%, more preferably by at least 10%, and most preferably by at least 25%, 50% or more.

Assays can be performed in eukaryotic cells, advantageously in mammalian cells, such as human cells. Appropriate assays can also be performed in prokaryotic cells, reconstituted membranes, and using purified proteins in vitro. Polypeptide therapeutics and in particular antibody-based therapeutics have significant potential as drugs because they have exquisite specificity to their target and a low inherent toxicity. In particular, the features of monoclonal antibodies such as high affinity, high selectivity, and distinct structure and function domains amenable to protein engineering for therapeutic delivery, make them potential drug candidates. Given the growing potential for the utilization of monoclonal antibodies as therapeutics, GP 3-specific monoclonal antibodies can be generated using techniques well-known by the skilled person as these form part of the current state of the art and the effectiveness of these antibodies as modulators of Αβ generation can also be determined in the context of the present invention.

Active fragments of the above described antibodies form also part of the invention. The term "active fragment" refers to a portion of a n anti body that by itself has high affinity for a n antigen ic determinant, or epitope, and contains one or more CDRs accounting for such specificity. Non-limiting examples include Fab, F(ab)'2, scFv, heavy-light chain dinners, nanobodies, domain antibodies, and single chain structures, such as a complete light chain or complete heavy chain.

The antibodies of the invention, or their active fragments, can be labeled by an appropriate label, said label can for instance be of the enzymatic, colorimetric, chemiluminescent, fluorescent, or radioactive type.

It is known by the skilled person that an antibody which has been obtained for a therapeutically useful target requires additional modification in order to prepare it for human therapy, so as to avoid an unwanted immunological reaction in a human individual upon administration. The modification process is commonly termed "humanization". It is known by the skilled artisan that antibodies raised in species, other than in humans, require humanization to render the antibody therapeutically useful in humans ( (1) CDR grafting : Protein Design Labs: US6180370, US5693761; Genentech US6054297; Cel ltech : EP626390, US5859205; (2) Veneering: Xoma : US5869619, US5766886, US5821123). Humanisation of antibodies entails recombinant DNA technology, and is departing from parts of rodent and/or human genomic DNA sequences coding for H and L chains or from cDNA clones coding for H and L chains. Techniques for humanization of non-human antibodies are known to the skilled person as these form part of the current state of the art. Non-human mammalian antibodies or animal antibodies can be humanized (see for instance Winter and Harris 1993). The antibodies or monoclonal antibodies according to the invention may be humanized versions of for instance rodent antibodies or rodent monoclonal antibodies.

The use of ARF6 and/or an ARF6 effector to reduce amyloid beta peptide generation in a mammalian cell is also envisioned in the present invention, as well as the use of ARF6 and/or an ARF6 effector to identify compounds that reduce amyloid beta peptide formation in a mammalian cell. In one embodiment said ARF6 and/or said ARF6 effector can be a mutant or can be overexpressed or can be downregulated. Examples of ARF6 mutants are known in the art, e.g. ARF6-Q67L, which is a dominant active mutant locking ARF6 in its GTP bound state or ARF6-T27N, which is a GDP-locked inactive ARF6 (Peters et al. 1995). Other non-limiting examples of how ARF6 and/or ARF6 effectors can be mutated, overexpressed or downregulated are described in the Example section.

In a further aspect, the invention also relates to the use of ARF6 and/or ARF6 effector to diagnose or prognose Alzheimer's disease. The use as meant here is any use of the nucleic acid or protein, and may be, as a non-limiting example, the genomic DNA, for the detection of mutation, the mRNA or derived cDNA for the analysis of the expression, or the protein, for the analysis of translated protein. Methods for mutation and SNP analysis, expression analysis and detection and quantification of protein (e.g. via antibodies recognizing the protein) are known to the person skilled in the art.

In still another aspect, the present invention relates to an agent such as an antisense polynucleotide, a ribozyme, or a small interfering RNA (siRNA), characterized in that it is reducing amyloid beta peptide formation, and wherein said agent comprises a nucleic acid sequence complementary to, or engineered from, an ARF6 encoding polynucleotide.

The terms "polynucleotide", "polynucleic acid", "nucleic acid" are interchangeably used further herein.

In a preferred embodiment, the agents of the present invention encompass short interfering RNA (siRNA) molecules that down regulate expression of ARF6 mRNA by RNA interference. RNA interference refers to the process of sequence-specific post transcriptional gene silencing in animals mediated by short interfering RNAs (siRNA) (Fire et al. 1998). siRNA molecules are short pieces of dsRNA obtained by processing of the dsRNA by Dicer, a ribonuclease III enzyme (Bernstein et al. 2001). Short interfering RNAs derived from Dicer activity are typically about 21-23 nucleotides in length and comprise about 19 base pair duplexes. siRNAs up to 26 nucleotides have proven to be effective at specifically silencing gene expression without causing any interferon response. The siRNA molecules of the present invention encompass human ARF6 siRNAs, which are useful for research to analyse the function of ARF6, and which may be used for therapy in humans, e.g. in the prevention and/or treatment of a disorder of the peripheral or central nervous system, in particular Alzheimer's disease. I n a specific embodiment, the small interfering RNAs (siRNA) of the present invention comprise a nucleic acid sequence as defined by SEQ ID NOs 1 and 2 (Table 1). Table 1. Specific siRNAs for ARF6

Based on the RNA sequence of human ARF6, siRNA molecules with the ability to knock-down ARF6 activity can be obtained by chemical synthesis or by hairpin siRNA expression vectors (as described by Yu et al. 2002). There are numerous companies that provide the supply of costumer-designed siRNAs on a given RNA sequence, e.g. Ambion, Imgenex, Dharmacon.

The ARF6 siRNAs of the invention may be chemically modified, e.g. as described in US20030143732, by ph osp ho roth ioate i ntern ucl eotide l i n kages, 2'-0-m et hyl ri bo n u c l eotid es, 2'-deoxy-2'fluoro ribonucleotides, "universal base" nucleotides, 5-C-methyl nucleotides, and inverted deoxyabasic residue incorporation. The sense strand of ARF6 siRNAs may also be conjugated to small molecules or peptides, such as membrane-permeant peptides or polyethylene glycol (PEG). Other siRNA conjugates which form part of the present invention include cholesterol and alternative lipid-like molecules, such as fatty acids or bile-salt derivatives.

In a further embodiment, the present invention relates to an expression vector comprising any of the above described polynucleotide sequences encoding at least one ARF6 siRNA molecule in a manner that allows expression of the nucleic acid molecule, and cells containing such vector. The polynucleic acid sequence is operably linked to regulatory signals (promoters, enhancers, suppressors etc.) enabling expression of the polynucleic acid sequence and is introduced into a cell utilizing, preferably, recombinant vector constructs. A variety of viral-based systems are available, including adenoviral, retroviral, adeno-associated viral, lentiviral, herpes simplex viral vector systems. Selection of the appropriate viral vector system, regulatory regions and host cell is common knowledge within the level of ordinary skill in the art.

As gene delivery and gene silencing techniques improve, the selective deletion of ARF6 in particular tissues or cellular populations may prove useful in order to limit the impact of protein deletion to a particular system under study. The ARF6 siRNA molecules of the invention may be delivered by known gene delivery methods, e.g. as descri bed in US 20030143732, incl uding the use of naked si RNA, synthetic nanoparticles composed of cationic lipid formulations, liposome formulations including pH sensitive liposomes and immunoliposomes, or bioconjugates including siRNAs conjugated to fusogenic peptides. Del ivery of si R NA expressi ng vectors ca n be system ic, such as by i ntraven ous or intramuscular administration or by any other means that would allow for introduction into the desired target cell (see US 20030143732).

In still another aspect, the invention relates to a pharmaceutical composition comprising a therapeutically effective amount of any of the above described agents and at least one of a pharmaceutically acceptable carrier, adjuvant or diluents. Any of the above agents or said pharmaceutical composition can be used for the manufacture of a medicament to prevent and/or treat a disorder of the peripheral or central nervous system, in particular Alzheimer's disease. One of ordinary skill in the art will recognize that the potency and, therefore, an "effective amount" can vary for the inhibitory agents of the present invention. One skilled in the art can readily assess the potency of the inhibitory agent.

A medicament to prevent and/or to treat a disorder of the peripheral or central nervous system, in particular Alzheimer's disease, relates to a composition comprising agents as described above and a pharmaceutically acceptable carrier or excipient (both terms can be used interchangeably) to treat or to prevent diseases as described herein.

The administration of pharmaceutical compositions may be by way of oral, inhaled or parenteral administration. In particular, pharmaceutical compositions can be delivered through intrathecal or intracerebroventricular administration. The active ingredient may be administered alone or preferably formulated as a pharmaceutical composition. An amount effective to treat Alzheimer's disease depends on the usual factors such as the nature and severity of the disorder being treated and the weight of the mammal. It is greatly preferred that the pharmaceutical composition is administered in the form of a unit-dose composition, such as a unit dose oral, parenteral, or inhaled composition. Such compositions are prepared by admixture and are suita bly adapted for oral, inhaled or parenteral administration, and as such may be in the form of tablets, capsules, oral liquid preparations, powders, granules, lozenges, reconstitutable powders, injectable and infusa ble solutions or suspensions or suppositories or aerosols. Tablets and capsules for oral administration are usually presented in a unit dose, and contain conventional excipients such as binding agents, fillers, diluents, tabletting agents, lubricants, disintegrants, colourants, flavourings, and wetting agents. The ta blets may be coated according to well known methods in the art. Suitable fillers for use include cellulose, mannitol, lactose and other similar agents. Suitable disintegrants include starch, polyvinylpyrrolidone and starch derivatives such as sodium starch glycollate. Suitable lu bricants include, for example, magnesium stearate. Suitable pharmaceutically acceptable wetting agents include sodium lauryl sulphate. These solid oral compositions may be prepared by conventional methods of blending, filling, tabletting or the like. Repeated blending operations may be used to distribute the active agent throughout those compositions employing large quantities of fillers. Such operations are, of course, conventional in the art. Oral liquid preparations may be in the form of, for example, aqueous or oily suspensions, solutions, emulsions, syrups, or elixirs, or may be presented as a dry product for reconstitution with water or other suitable vehicle before use. Such liquid preparations may contain conventional additives such as suspending agents, for example sorbitol, syrup, methyl cellulose, gelatin, hydroxyethylcellulose, carboxymethyl cellulose, aluminium stearate gel or hydrogenated edible fats, emulsifying agents, for example lecithin, sorbitan monooleate, or acacia; non-aqueous vehicles (which may include edible oils), for example, almond oil, fractionated coconut oil, oily esters such as esters of glycerine, propylene glycol, or ethyl alcohol; preservatives, for example methyl or propyl p-hydroxybenzoate or sorbic acid, and if desired conventional flavouring or colouring agents. Oral formulations also include conventional sustained release formulations, such as tablets or granules having an enteric coating. Preferably, compositions for inhalation are presented for administration to the respiratory tract as a snuff or an aerosol or solution for a nebulizer, or as a microfine powder for insufflation, alone or in combination with an inert carrier such as lactose. In such a case the particles of active compound suitably have diameters of less than 50 microns, preferably less than 10 microns, for example between 1 and 5 microns, such as between 2 and 5 microns. For parenteral administration, fluid unit dose forms are prepared containing a compound of the present invention and a sterile vehicle. The active compound, depending on the vehicle and the concentration, can be either suspended or dissolved. Parenteral solutions are normally prepared by dissolving the compound in a vehicle and filter sterilising before filling into a suitable vial or ampoule and sealing. Advantageously, adjuvants such as a local anaesthetic, preservatives and buffering agents are also dissolved in the vehicle. To enhance the stability, the composition can be frozen after filling into the vial and the water removed under vacuum. Parenteral suspensions are prepared in substantially the same manner except that the compound is suspended in the vehicle instead of being dissolved and sterilised by exposure to ethylene oxide before suspending in the sterile vehicle. Advantageously, a surfactant or wetting agent is included in the composition to facilitate uniform distribution of the active compound. Where appropriate, small amounts of bronchodilators for example sympathomimetic amines such as isoprenaline, isoetharine, salbutamol, phenylephrine and ephedrine; xanthine derivatives such as theophylline and aminophylline and corticosteroids such as prednisolone and adrenal stimulants such as ACTH may be included. As is common practice, the compositions will usually be accompanied by written or printed directions for use in the medical treatment concerned.

The terms "therapeutically effective amount", "therapeutically effective dose" and "effective amount" mean the amount needed to achieve the desired result or results (modulating A F6 activity; treating or preventing Alzheimer's disease). "Pharmaceutically acceptable" means a material that is not biologically or otherwise undesirable, i.e., the material may be administered to an individual along with the compound without causing any undesirable biological effects or interacting in a deleterious manner with any of the other components of the pharmaceutical composition in which it is contained.

A "carrier", or "adjuvant", in particular a "pharmaceutically acceptable carrier" or "pharmaceutically acceptable adjuvant" is any suitable excipient, diluent, carrier and/or adjuvant which, by themselves, do not induce the production of antibodies harmful to the individual receiving the composition nor do they elicit protection. Preferably, a pharmaceutically acceptable carrier or adjuvant enhances the immune response elicited by an antigen. Suitable carriers or adjuvantia typically comprise one or more of the compounds included in the following non- exhaustive list: large slowly metabolized macromolecules such as proteins, polysaccharides, polylactic acids, polyglycolic acids, polymeric amino acids, amino acid copolymers and inactive virus particles.

A "diluent", in particular a "pharmaceutically acceptable vehicle", includes vehicles such as water, saline, physiological salt solutions, glycerol, ethanol, etc. Auxiliary su bstances such as wetting or emulsifying agents, pH buffering substances, preservatives may be included in such vehicles.

It should be clear that the inhibitory agents of the present invention for Alzheimer's disease can also be used in combination with any other AD disease therapy known in the art such as gamma-secretase inhibitors, or beta-secretase inhibitors.

A further aspect of the present invention relates to a cell line characterized by lacking endogenous presenilin expression or function and the same cell line stably expressing virally transduced A F6 (see Example 3). Presenilins (PS, PSEN) have been shown to form the catalytic subunit of the γ-secretase complex that produces the Αβ peptide. So, a PSEN double knock out cell line ( PSENdKO), lacking endogenous presenilin expression (PSEN1 and PSEN2), lacks γ-secretase activity. However, despite the γ-secretase dependent (catalytical) function, presenilins also have other functions, such as a role in trafficking of membrane proteins. So, a PSENdKO cell line has an aberrant morphological phenotype due to typical endosomal trafficking defects, such as aberrant accumulation of membrane lipids and proteins. It was shown in the present invention that by "rescuing" the PSENdKO cell line with the ARF6 GTPase, the aberrant morphological phenotype observed in the PSENdKO cell line could be restored. Therefore, said "rescue" cell line allows to discriminate between the γ-secretase independent versus γ- secretase dependent (catalytical) function and as such is a val ua ble screening tool for new drug compounds. In particular, the PSENdKO celline and the PSENdKO cell line expressing ARF6 can be used in an assay for screening compounds that are capa ble of modulating ARF6 protein activity in a mammalian cell. Preferably, said compounds are capable of reducing amyloid beta peptide formation in a mammalian cell. Even more preferably, said compounds are therapeutic candidates for the prevention and/or treatment of a disorder of the peripheral or central nervous system, in particular Alzheimer's disease. In particular, said cell lines are MEF cell lines.

A particular embodiment of the present invention relates to a method for identifying compounds that modulate the endosomal redistribution in a mammalian cell comprising the steps of:

a) providing a cell culture characterized by lacking endogenous presinilin expression/function;

b) administering a test compound to said cell culture;

c) imaging at least one morphological parameter of the cells in said cell culture; wherein, under the same test conditions, a deviation or aberration in said at least one parameter compared to the same at least one parameter of cells of a corresponding wild type cell culture and/or of said cell culture characterized by lacking endogenous presinilin expression/function stably expressing A F6, identifies said test compound as a compound that modulates the endosomal redistribution in a mammalian cell.

The term "morphological parameters" in the context of the present invention includes, but is not limited to, cell area, cell perimeter, cell ratio area/perimeter, cell elongation, cell diameter, cell intensity, cell count, cell roundness. Differences in morphological phenotype between cell cultures in the presence of test compounds can be screened in a high-throughput imaging setup (e.g. InCell 2000, GE Healthcare; see also Example 4).

The phrase "modulate (or modulation of modulating) of the endosomal redistribution" as used herein refers to interfering with the activity of the endocytic pathway, wherein said activity includes but is not limited to (i) endocytic rates; (ii) endosomal fusion and recycling; (iii) degree of accumulation of lipid species such as cholesterol and sphingomyelins and/or proteins or enzymes; (iv) degree of accumulation of β-amyloidogenic fragments or aggregated proteins or protein fragments in endosomal compartments; (v) degree of accumulation of autophagic vacuoles. Usually, an abnormal endocytic pathway is one that exhibits an increase or decrease in one or more of the foregoing activities, in other words a change in endosomal balance that is aberrant from the normal situation. In a preferred embodiment, a compound will "decrease" or "reduce" the abnormal activity of the endocytic pathway. Said reduction is preferably by at least 5%, more preferably by at least 10%, and most preferably by at least 25%, 50% or more. Assays and methods for measuring the activity of the endocytic pathway are known in the art. As a non-limiting example, filipin can be administered to the above mentioned cell cultures to identify cholesterol accumulations as a readout for endosomal accumulations. In a more specific embodiment, said compound modulates A F6 cycling activity and/or ARF6 effector protein activity. Preferably, said compound is capable of reducing amyloid beta peptide formation in a mammal ian cell and as such is a therapeutic candidate for the prevention and/or treatment of Alzheimer's disease. Said reduction is preferably by at least 5%, more preferably by at least 10%, and most preferably by at least 25%, 50% or more.

The following examples are intended to promote a further understanding of the present invention. While the present invention is described herein with reference to illustrated embodiments, it should be understood that the invention is not limited hereto. Those having ordinary skill in the art and access to the teachings herein will recognize additional modifications and embodiments within the scope thereof. Therefore, the present invention is limited only by the claims attached herein.

EXAMPLES

Example 1. BACEl internalises via the clathrin-independent ARF6-mediated endocytotic pathway

The β-amyloid precursor protein (APP) internalises via a clathrin-dependent pathway (Schneider et al.

2008) however the molecular machinery involved in the regulation of BACEl internalisation remains unknown. Unlike APP, BACEl does not contain a sorting motif within its tails which could regulate its internalisation via the clathrin route (Traub 2009). However, BACEl's cytosolic tail (aa 496-500) contains an acid cluster-dileucine motif (ACDL, DISLL sequence) which has been shown to regulate its trafficking between the endosomal compartments as well as its internalisation (Capell et al. 2000; Huse and Pijak 2000; Pastorino et al. 2002). The ACDL motif of BACEl binds to the Golgi-localised gamma-ear containing ADP ribosylation factor-binding (GGA) family (He et al. 2002; Shiba et al. 2004; von Arnim et al. 2004; He et al . 2005; Wahle et al. 2005). These are small monomeric adaptors involved in the transport of protein mostly between the TGN and the endosome and in the recycling pathway from the endosomes to the TGN (Bonifacino 2004; He et al. 2005). All three GGA ( 1, 2 and 3) appear to be involved in the trafficking of BACEl as depletion of any of them causes a significant change in the distribution of BACEl (He et al. 2005). Recently, GGA3 was shown to be recruited by ARF6, a small GTPase involved in membrane trafficking (D'Souza-Schorey and Chavrier 2006; Grant and Donaldson

2009) , and TBC1 D3, a TBC-containing protein lacking GAP activity, at the plasma membrane and involved in a macropinocytosis route (Frittoli et al. 2008). GGA3 has been also implicated in trafficking and down regulation of BACEl (Tesco et al. 2007).

It was investigated whether BACEl could be a cargo of this ARF6-mediated macropinocytotic pathway by first looking at BACEl localisation in cells overexpressing ARF6 cycle mutants (Peters et al. 1995). We tested this by first co-expressing BACEl with ARF6-Q67L, a dominant active mutant locking ARF6 in its GTP bound state (Peters et al. 1995). As shown in Fig. la, BACEl becomes clearly trapped together with ARF6 cargo proteins like MHCI in characteristic grape-like vacuoles (Brown et al. 2001; Naslavsky et al. 2003; Naslavsky et al. 2004). On the other hand blocking ARF6 activation, by expressing the GDP- locked inactive ARF6, ARF6-T27N mutant, inhibits recycling to the cell surface (Peters et al. 1995). In this condition, BACEl also co-localized with MHCI trapped in ARF6-T27N positive recycling endosomal structures (Fig. lb, inset). In addition, in the case of both mutants, MHCI and BACEl often colocalized to ARF6-positive protrusions or ruffles at the cell surface (Fig. la-b). Alternatively, ARF6 activation through overexpression of its specific guanine nucleotide exchange factor (GEF) EFA6 equally induces protrusions and macropinosomes (Franco et al. 1999) which turned out to be positive for BACEl when co-overexpressed (data not shown). Also both BACEl and GGA3 colocalized at the cell surface as well as intracellular^ (Fig. lc). It has been shown that TBC1D3 overexpression induce extensive formation of ruffle-like structures at the dorsal surface of the cells thereby recruiting GGA3 (Frittoli et al. 2008). Indeed, the co-localization of GGA3 with BACEl was clearly enhanced at TBClD3-induced ruffles (Fig. Id, arrows and arrowheads in the inset).

Also when co-expressed with ARF6-Q67L, BACEl and GGA3 became trapped in ARF6-positive vacuoles (Fig. le)(Frittoli et al. 2008). These results clearly show that GGA3 likely recruits BACEl for internalization through the ARF6 dependent macropinocytic uptake route.

In contrast, APP has been shown to require the clathrin-dependent pathway for internalization (Carey et al. 2005; Schneider et al. 2008). Indeed, when APP was co-expressed with ARF6-Q67L, we did not observe APP accumulating in ARF6-positive endocytic vacuoles (Fig. 2a) underscoring that the substrate APP and its sheddase follow distinct internalization routes.

To further corroborate these findings, we followed the fate of cell surface localized BACEl and APP using an antibody uptake assay. Hela cells were incubated at 4°C with antibodies directed against BACEl and CD59, a GPI-anchor protein following the ARF6 route, and subsequently chased at 37°C in the presence of high EGF concentrations (200 ng/ml) to stimulate macropinocytosis (Lanzetti et al. 2004). After 10 min, most of the internal ized antibody-BACEl conjugates colocalized with CD59- positive endosomal compartments (Fig. 2b) indicating that the major pool of surface localized BACEl entered the cell via the ARF6-dependent pathway. On the contrary, in a similar setup, no internalised APP was colocalizing with CD59 (Fig. 2c). Next, it was investigated whether a selective inhibition of the clathrin-dependent pathway would selectively affect APP but not BACEl internalization. Therefore, BACEl or APP were overexpressed together with the carboxy-terminal domain of AP180 (AP180-C). AP180 is a nonconventional neuronal adaptor protein of the clathrin pathway that when truncated blocks internalization (Zhao 2001; Frittoli et al. 2008). Next, cells were surface biotinylated and following a lOmin chase, the amount of internalized biotinylated BACEl, APP or endogenous transferrin receptor (TfR) was quantified (Fig. 2c-d). While endocytosis of APP was efficiently inhibited by 80%, as reported (Schneider et al. 2008), BACEl internalization was only slightly but not significantly affected. Also another typical clathrin-dependent cargo protein, TfR (Schneider et al. 2008) was inhibited by ~40% (p<5%, Figure 2d). The lower efficiency here might however be due to the fact that TfR is endogenously expressed. Again, our data confirm independently that internalization of BACEl and APP can be separately controlled. Interestingly, when we transfected primary hippocampal neurons with ARF6-Q67L, overexpressed BACEl but not APP accumulated in ARF6-positive vacuoles (Fig. 2e) indicating that distinct internalization routes for BACEl and APP also exist in the brain.

It has been described that early endosomes, the typical recipient of clathrin-mediated endocytic organelles, constitute a major site of BACEl activity (Rajendran et al. 2006; Schneider et al. 2008). However, in these studies, BACEl itself was not yet convincingly shown to reside in early endosomes. By using the dominant active RAB5-Q79L mutant, a GTP-locked RAB5 that blocks exit and maturation from early endosomes causing them to enlarge (Stenmark et al. 1994; Rink et al. 2005), we now clearly show that co-expressed BACEl readily accumulated in distinct domains of enlarged endosomes (Fig. 3a). This was further confirmed by antibody uptake experiments where internalised antibody-BACEl conjugates reached RAB5-Q79L-positive endosomes (data not shown). On the other hand, no significant uptake of BACEl antibodies was detected in cells co-expressing the GDP-bound RAB5-S34N mutant (data not shown). This is interesting since it accords with a proposed role of RAB5 in macropinocytosis. Indeed such a route from ARF6-dependent internalization to RAB5-positive early endosomes is described for certain ARF6 cargo molecules, such as MHCI and CD59 (Naslavsky et al. 2003; Grant and Donaldson 2009). To investigate whether BACEl follows a similar route we co- expressed BACEl together with both ARF6-Q67L and RAB5-Q79L. In this case, and as opposed to RAB5- Q79L alone, no significant immunolabeling for BACEl was detected in RAB5-Q79L enlarged endosomes clearly indicating that BACEl reaches the RAB5-positive endosomes via the ARF6-mediated route (Fig. 3b). Next, cargo is sorted to a endosomal recycling compartments from where proteins like MHCI are recycled to the cell surface again in an ARF6-dependent manner. This formation of recycling tubules can be blocked by overexpressing the GDP-locked ARF6-T27N mutant and when applied here, this indeed resulted in the perinuclear accumulation of BACEl (Fig. lb). Even more, in triple transfected cells, both BACEl and APP accumulated after ARF6-T27N mediated blockade of recycling as opposed to the effect of ARF6-Q67L (Fig. 3c and 3d). Altogether, the above data imply that BACEl and APP, enzyme and substrate respectively, can enter the cell via two distinct routes. BACEl internalises majorly via an A F6-dependent clathrin- independent pathway and that as previously described (Schneider et al. 2008), APP endocytosis is clathrin mediated.

Example 2. Modulation of the clathrin-independent ARF6-mediated endocytotic pathway affects the processing of APP.

As shown in Example 1, the ARF6 cycling mutants had opposite effect on the colocalisation of BACEl and APP, with the ARF6-Q67L clearly blocking BACEl transport prior to encountering APP. Here, it was investigated how these mutants affected the processing of APP.

Firstly, we looked at the effect of wild type and mutants ARF6 on APP processing using quantitative westernblotting. Overexpression of APP with ARF6-Q67L resulted in a nearly 50% drop in the ratio of APP-CTF over full-length APP while the ARF6-T27N caused a 2,5 fold increase (Fig.4a). We next moved to metabolic labelling allowing us to evaluate in more detail the effects of wild type versus mutant ARF6 on newly synthesized swedishAPP (sweAPP). This mutant was chosen as it increases β-secretase processing due two mutations at codons 670 and 671 (APP₇₇₀ transcript) at the N-terminus of the Αβ sequence (Mullan et al. 1992). Like for APP-CTF, ARF6-Q67L caused a dramatic decrease in both the secretion of soluble ectodomain fragments of sweAPP (APPs as well as Αβ peptides indicating that overall processing of sweAPP was strongly inhibited in ARF6-Q67L expressing cells (Fig. 4b). On the other hand, ARF6-T27N expression had an adverse effect on sweAPP processing resulting in more secreted APPs and Αβ as compared to wild type ARF6 overexpression (Fig. 4b). However, it was suprising to find that these increases were statistically not significant. This could be explained by the assumption that despite the increased shedding, the ARF6-T27N mediated transport blockade also affected the accessibility of APP-CTFs to become processed by γ-secretase. This is at least supported by the stronger accumulation of APP-CTF in these cells (Fig.4a) although alternative explanations cannot be excluded. To further scrutinize the importance of ARF6 activity on BACEl trafficking and hence APP processing, we explored the effect of (hyper)activating the ARF6 cycle by overexpressing the ARF6 specific GEF, EFA6a (Franco et al. 1999) and GAP, ACAP1 (Jackson et al. 2000), that promote the GTP- and GDP-bound state of ARF6, respectively. In both cases, overexpression resulted in a significant decrease in secreted APPs (not shown) and Αβ Fig. 4c). On the other hand, overexpressing their respective mutant EFA6A and ACAP1 did not affect Αβ secretion as compared to mock transfection. This suggest that the observed effects are caused by the respective increased GEF and GAP activity on ARF6 function possible resulting both in a preference towards the GTP-bound state of ARF6. It is not surprising therefore that both effectors have a similar effect on APP processing as observed for ARF6- Q67L. Finally, we investigated the effect on APP processing when the ARF6 route overall was suppressed. As shown in Fig. 3d inset, siRNA-mediated downregulation of endogenous ARF6 resulted in nearly 80% reduced protein levels as compared to non-specific oligonucleotides. This significantly reduced Αβ secretion al beit not to the extent as observed by a more dramatic blockade of ARF6 mediated transport for instance using ARF6-Q67L.

Altogether our data demonstrate that BACEl and APP enter the cell/neuron via two distinct routes, a clathrin-independent/ ARF6 dependent and clathrin-dependent one, respectively (Fig. 5). Overexpression of ARF6-Q67L blocks the transport of BACEl at an early stage preventing it for reaching the early endosome and hence access to its major substrate APP (Fig. 5b). The net result is a prominent decrease in proteolysis. On the other hand, inhibiting BACEl recycling to the cell surface using ARF6- T27N, enhances residence time of both BACEl and APP resulting in increased processing (Fig. 5c). Hence, keeping BACEl and APP separated until they encounter each other in the early endosome provides a clear physiological means to control ectodomain shedding and hence Αβ production.

Additionally our data suggest that BACEl utilizes this selective ARF6 route for balancing its levels in the cell and neuron. Indeed, the ARF6-mediated endocytosis as degradative route has been previously described also for the epidermal growth factor (EGF) receptor (Sigismund et al. 2008). When cells are treated with high (200 ng/ml) EGF doses, a pool of the EGF-receptor becomes internalized via an ARF6- dependent route for degradation instead of the clathrin-mediated pathway which promotes recycling (Sigismund et al. 2008). Similarly, high EGF concentrations, a condition that stimulates macropinocytosis, also enhances BACEl internalization (this study; Lanzetti et al. 2004; Zou et al. 2007; Frittoli et al. 2008), which by this mechanism allow the cell to adapt and regulate the amount of protein at the cell surface. Our data also provide a cell biological explanation for the inverse correlation of BACEl and GGA3 expression levels described earlier (Tesco et al. 2007). Indeed, down regulation of GGA3 was shown to inhibit macropinocytosis (Frittoli et al. 2008) and hence decreased levels of GGA3 as for instance observed in AD brain, prevents BACEl from being sorted in the ARF6-dependent pathway and instead stabilized.

Although the major site of Αβ generation has been assigned to endosomal compartments we now prove for the first time that BACEl reaches this compartment independently from APP via a clathrin- independent/ ARF6-mediated route. Ou r study therefore not on ly increases sign ificantly ou r understanding of the intracellular transport regulation of BACEl versus APP, but, moreover, demonstrate that sorting of BACEl and APP can be independently obstructed by interfering with their respective endocytic machineries, as shown with ARF6 mutants but also with AP180-C. This implies that factors or compounds selectively affecting internalization of BACE1 or APP, may provide a unexplored avenue for therapeutic inhibition of Αβ production in AD.

Example 3. Retroviral expression of human ARF6 in a PSEN double knockout cell line and its use in screening assays

Presenilins (PSEN1&2) can act in a γ-secretase independent manner regulating cell migration and endocytosis. The migration and adhesion defects of PSENKO cells are likely caused by problems in caveolinl transport and redistribution from intracellular caveosomes. Therefore the role of PSENs in adhesion-dependent endocytosis and recycling pathways of raft proteins in wild-type (WT) versus PSENl&2dKO MEFs was tested. Adherent fibroblasts were pre-labeled with CTxB, specifically labeling the raft marker GMl. CTxB and caveolinl localize sharply at the plasma membrane immediately after detachment, and accumulate in the perinuclear region after 30 min in suspension. Despite significant differences in the steady state localization of caveolinl in PSENKO MEFs (Fig. 9), there was no delay in the uptake kinetics towards the intracellular recycling compartments during the suspension phase (Fig. 6). After lh in suspension, fibroblasts were replated on a fibronectin-coated surface for lh. In WT MEFs this resulted in the polarized redistribution of caveolinl and CTxB to the plasma membrane. In PSENl&2dKO cells the dynamic relocalization of CTxB and caveolinl was not apparent, and both molecules remained largely in the perinuclear region. Caveolinl distribution is much more fragmented in fine spots in PSENl&2dKO cells compared to controls and does not reach the plasma membrane at all (Fig. 7).

ARF6 controls the adhesion-regulated recycling of rafts from the recycling endosomes to the plasma membrane (Balasubramanian e aL 2007). Looking at the ARF6 levels in PSENKO MEFs, it was found that total ARF6 levels as well as ARF6-GTP levels were decreased in all PSENKO cells. Via RTqPCR it was shown that presenilins regulate the amount of ARF6 mainly at the level of mRNA synthesis (Fig. 8). Thus ARF6 activation, regulated upstream by PSENs, is coupled to the efflux of recycling endosomes and redistribution of raft components. In addition, it was found that retroviral expression of human ARF6 in PSENl&2dKO cells can rescue the intracellular accumulations of caveolinl, resulting in its polarized plasma membrane distribution at one edge of the cell (Fig. 9) and generating PSENDKO MEFs lacking γ-secretase activity but with a morphology and balanced endosomal trafficking reminiscent of wild-type cells.

Since the internalization and redistribution of GMl-containing membrane domains regulate Racl signaling, we speculated that Racl does not get internalized upon adhesion, resulting in problems in cell spreading, polarity, and cell migration, as shown for caveolinl KO MEFs (del Pozo et al. 2004; Grande-Garcia et al. 2007). It is therefore not surprising that the normalized caveolinl distribution in PSENl&2dKO MEFs expressing human ARF6 results in decreased Racl levels (Fig. 10) and a normal fibroblast-like phenotype. These 'rescue' cell lines underwent dramatic morphological changes from a round morphology to an elongated cell shape, with reduced lamellipodia formation compared to presenilin deficient cells. ARF6 expression in PSENl&2dKO MEFs also compensates for the increased cell migration speed in wound healing assays (Fig. 11).

Earlier reports showed that PSEN1 is involved in degradative organelle turnover. Deficiency of PSEN1 in hippocampal neurons leads to the accumulation of ICAM5 (Esselens et al. 2004), a n d a- and β- synuclein in autophagic vacuoles (Wilson et al. 2004). Similarly to the neuron data, a dramatic increase in the amount of acidified compartments was seen in PSENl&2dKO MEFs, concomitant with increased anti-EGFR staining compared to control cells (Fig. 12). Retroviral expression of human ARF6 WT in these cells resulted in a dramatic decrease of EGFR expression levels dependent on the amount of ARF6 rescue (Fig. 10). ARF6 is involved in the regulation of protein degradation, resulting in largely diminished acidic accumulations in PSl&2dKO cells. Anti-EGFR staining is dispersed over the complete cell due to increased ARF6 expression levels.

Altogether, by "rescuing" the PSENdKO cell line with the ARF6 GTPase, the aberrant morphological phenotype observed in the PSENdKO cell line could be restored. Therefore, said "rescue" cell line allows to discriminate between the γ-secretase independent versus γ-secretase dependent (catalytical) function and as such is a valuable screening tool for new drug compounds.

Example 4. Screening assay

A morphological screening assay has been set up making use of wild-type (WT) and Presenilinl&2 (PSEN1&2) knockout Mouse Embryonic Fibroblasts (MEFs). PSEN1&2 knockout cells stably rescued with ARF6 are used as a positive control. The screening assay can be dissected in different steps. First cells are brought into culture and several cell lines are maintained. These cells are treated with compounds, incubated and afterwards fixed and stained. These stained cells are imaged and the images are further analyzed. The screening assay was optimized and is described into further detail in the Materials and Methods to the Examples section.

A medium to high throughput screen was performed using a chemical compound library. The chemical genetic screen took advantage of a library of 10,000 compounds available at the Plant Systems Biology department of VIB (Ghent, Belgium). As an initial screen for morphology, in each well 50 to 100 cells were randomly analyzed with respect to the surface area they occupy as well as the perimeter. The ratio of surface area to perimeter, defined as 'circularity' was read-out to quantify changes in morphology, i.e. changes from more circular, adhesive cells to elongated migratory cells. An example and proof-of-principle is given (Fig. 13) in which the circularity is measured between WT, PSEN 1&2 knockout and PSEN 1&2 knockout cells stably rescued with A F6 as a positive control.

After optimization, five parameters are calculated to study changes in the cell morphology, namely the cell intensity, cell count, cell roundness, cell area and cell elongation. The ratio of the short over the long axis of the cells is called the cell elongation. A not elongated or symmetric cell has a value of one. All the other values situate between zero and one. The cell roundness is also called the cell 1/form factor. It is calculated via parameter (perimeter) over area. These values lie between one and + infinity. A cell with a roundness of one, forms a perfect circle.

Using this morphology assay we screened about 2,500 chemical compounds out of the available 10,000 compound library (hence 25%). Each compound was analyzed simultaneously in three MEF lines, namely wild-type, PSEN-dKO and PSEN-dKO stably rescued with ARF6, using the InCell 2000 analyzer (GE Healthcare). At the end of the experiment, the morphology of the cells (i.e. along the parameter of 'roundness') was evaluated and scored. From the 2,500 chemical compounds we could identify 23 positive hits. With positive hits we mean those compounds that affect the morphological parameter of PSEN-dKO M EFs but not that of wild-type and ARF6-rescued MEFs. Such compounds likely affect the morphology of PSEN-dKO MEFs through acting on up- or downstream effectors of the ARF6 pathway or signaling.

Positive hits are further analyzed on the basis of the following parameters:

Localization of endogenous caveolin 1 in these MEF cell lines

Accumulation (and its rescue) of lysotracker in acidic endosomal compartments

Uptake of fluorescently-tagged Epidermal Growth Factor (EGF)

Measurement of APP processing after metabolic labeling

Example 5. Endogeneous ARF6 levels decrease with aging

Given the findings in Example 2, that interfering with ARF6 activity affects APP processing, and in example 3, that ARF6 levels are downregulated in PSEN deficient cells, it follows that we hypothesized that endogenous ARF6 expression levels may undergo changes during aging of the brain. We analyzed extracts of brain cortices obtained from mice at different stages of development (prenatal and early postnatal days and weeks) up to adult mice, between 6 months and 24 months of age, the latter referred to as old or aged mice. Using quantitative westernblotting of endogenous ARF6 protein expression levels, normalized to endogenous GAPDH protein expression levels, we observe an increase of ARF6 protein expression during late embryonal development and postnatally. However these levels dropped from 3 months onwards to over 50% between 6 and 24 months (Fig. 14). In contrast, endogenous RAB5, a major RAB GTPase involved in early endosomal sorting remains, after an increase during embryonale stages, very stable up to 24 months underscoring a selective vulnerability for ARF6 and the ARF6 pathway. Interestingly, also endogenous presenilin 1 protein levels decrease to a similar extent from 1 month onwards to 24 months. This agrees well with the observed lower ARF6 levels in presenilinl deficient cells and neurons and highlights a genetic interaction between both.

Example 6. ARF6 levels are downregulated in the brain of Alzheimer's disease patients

As ARF6 levels decrease during aging in mice, we challenged the idea that a similar event occurs in brains for Alzheimer's disease (AD) patients. To investigate this, we obtained brain samples of the frontal cortex of human control and AD brain (Brainbank of Lille, in collaboration with Prof. Em. A Delacourte). Protein extracts were analyzed by quantitative western blotting and normalized to endogenous GAPDH. As a control, we compared the levels of endogenous RAB5 (see example 5) as well as the adaptor molecule GGA3. The latter has been demonstrated to be downregulated in AD brain and is a n interactor of BACE 1 (Tesco et a l. 2007) . We o bserve that both AR F6 a nd GGA3 are downregulated in AD brain, as compared to control brain samples. Again, endogenous RAB5 levels are not altered not only underscoring the reliability of quantitative westernblotting but also the selectivity in affected endosomal regulators. Of importance is also the recent observation that GGA3 interacts with TBC1D3 and that overexpression of TBC1D3 activates the ARF6-mediated macropinocytotic route. It is therefore not surprising that both ARF6 and GGA3 levels are lowered as they operate in the same internalization/endosomal route.

Materials and methods to the examples

Cell culture

HeLa (CCL2 clone) cells were routinely grown in Dulbecco's modified Eagle's medium (DMEM/F12, Invitrogen) supplemented with 10% fetal calf serum and maintained in a humidified chamber with 5% C02 at 37°C. The culture of primary hippocampal neurons has been described previously [1]. Antibodies

The following monoclonal antibodies (mabs) were commercially obtained: Rat mab against theHA tag (clone 3F10,Roche Diagnostics),mab to ARF6 (SC-7971, Santa Cruz);, mab to MHCI (clone W6/32, Abeam) , mab to CD59 (Chemicon); mab to FLAG (clone M2, Sigma), and mab to GGA3 (clone 8, BD transduction Laboratories); mab to transferrin receptor (TFR,clone H68.4, Invitrogen) and mabs to Αβ (6E10, 4G8, Signet Laboratories). Rabbit polyclonal antibodies (pabs) were obtained from: anti-APP (A8717, Sigma); and anti-myc (clone A-14, Santa Cruze) . Pa b to APP ( B63) has been described previously (Esselens et al. JCB 2004) and generated using a synthetic peptide corresponding to the final 16 amino acids of APP coupled to KLH (Pierce Chemical Co.) as an antigen.

A mab, 10B8, against BACEl, was produced in house by immunizing BACEl-/- mice with recombinant hBACEl ectodomain followed by generation of a hybridoma cell line according to established procedures (Esselens et al. 2004).

Plasmids, Transfection, RNAi

Plasmids encoding ARF6 proteins (pXS-HA-ARF6wt, pXS-HA-ARF6Q67L and pXS-HA-ARF6T26N) were generously provided by J. Donaldon (NIH, Bethesda, Maryland, USA); myc-tagged AP180C was from H. McMahon (MRC, Cambridge, UK), GFP-EFA6 from P. Chavrier (Institut Curie, Paris, France), HA-tagged TBC1D3 and FLAG-tagged GGA3 were from S. Confalonieri (IFOM, Milan, Italy); Rab5Q79L from M. Zerial (MPI, Dresden, Germany), ACAPl-wt and ACAP1-R448Q from V. Hsu (Harvard Medical School, Boston, USA) and finally pcDNA-hBACEl from C. Haass (Ludwig-Maximilians-Universitat Adolf- Butenandt-lnstitute, Germany). cDNAs of wild-type APP695 and the Swedish mutant, APPsw, were cloned into pcDNA-3.1 (Zeo+) vector (Invitrogen). ARF6 constructs were recloned into pcDNA as follows: HA-tagged ARF6T27N using EcoRI and Xbal, and ARF6Q67L using EcoRI and EcoRV restriction sites. All cDNAs were verified by sequencing.

HeLa cells were transfected with FugeneHD (Roche Diagnostics) and primary neurones with Lipofectamine 2000 (Invitrogen) as described by the manufacturers.

The siRNA sequence targeting ARF6 are oligo#l: 5 ' -G C ACCG C ATTATC AATG ACCG -3 ' and oligo#2: 5'- G GTCTC ATCTTCGTAGTG G -3' [2]. Oligo#l was used for meta bol ic la bell ing experiments. The GL2 luciferase RNAi duplex was used as a nonspecific control. RNAi duplexes were manufactured by Dharmacon, Lafayette, CO and transfection was done using Oligofectamine (Invitrogen) as described by the manufacturer.Cells were analyzed 48 hres after downregulation. Confocal laser scanning microscopy

Cells were routinely plated on glass coverslips, transfected 24hrs later and processed for indirect immunolabeling the next day. After fixation (4% paraformaldehyde/ 4% sucrose in PBS, 20min RT)) and washing, cells are permeabilized (0,1% Triton in PBS, 5min RT) and blocked ( 2% bovine serum albumin (BSA), 2% foetal bovine serum (FBS), 1% gelatin, 2% goat serum in PBS, lhr, RT). Primary anti bodies were dil uted in the same blocking buffer and applied to fixed cel ls (4°C, overnight). Following washes in PBS, cells were incubated with the appropriate secondary antibodies conjugated to either Alexa-488, -568, -647 or Pacific Blue (Invitrogen) diluted in blocking buffer (lhr, RT). Finally, cel ls were washed and mounted with Mowiol . Images were captured on a confocal microscope (Radiance 2100; Carl Zeiss Microimaging, Inc.) connected to an upright microscope (Eclipse E800; Nikon) and using an oil-immersion plan Apo 60x A/1.40 NA objective lens. Image acquisition and final processing was done with Lasersharp (v..., Zeiss) and Adobe Photoshop 8.0 (Adobe, CA).

Primary antibody uptake hBACEl- or APPwt-transfected cells were serum-starved for 4hrs, rinced twice with ice-cold serum- free medium and incubated with the appropriate primary antibodies (mab 10B8 for BACE1 and 6E10 for APP) diluted in serum-free medium on ice for 30 min. Then, cells were rinsed twice with ice-cold serum free medium to remove unbound antibodies and placed back in the incubator at 37°C for 10 min in pre-warmed medium containing 200 ng/ml EGF (Sigma), to stimulate macropinocytosis [3] or kept on ice for the controls. Internalization was stopped by putting cells on ice, and replacing the medium with washes with ice-cold PBS. Following fixation (4°C, 20min), but prior to permeabilization remaining cell surface bound antibodies were first immunolabelled by incubating cells with Pacific- Blue-conjugated secondary antibodies (lhr). Then cells were rinsed and permeabilized (0,1% Triton X- 100 in PBS, 5min) and blocked (see above). Internalized primary antibodies were immunolabelled by incubating cells with appropriate Alexa-tagged secondary antibodies. Imaging and processing was done as described above.

Cell surface biotinylation and internalisation assay

All reagents, except otherwise stated, were kept on ice. Following 24hrs transfection with the appropriate constructs, HeLa cells were placed on ice, washed in PBS (pH 8) and next incubated in PBS (pH 8) supplemented with 0,25 mg/ml sulfo-NHS-SS-Biotin (Pierce) (15 min at 4°C). Excess of biotin was washed outand cells were incubated with 1% BSA in PBS (10 min at 4°C). After removing BSA, cells were incubated at 37°C for the appropriate time or kept at 4°C as a control. Endocytosis was quickly stopped by placing the dishes back on ice and washing the cells with ice-cold PBS. Remaining biotin at the cell surface was cleaved off by incubating the cells twice in PBS containing 100 mM 2-sodium-2- mercaptoethanesulfonate (Sigma) as a reducing agent (15 min at 4°C). To determine total surface biotinylation cells were incubated in PBS lacking the reducing agent. After this, cells were washed with ice-cold 5 mg/ml iodoacetamide for 5 min, then twice with ice-cold PBS, and finally extracted in in lysis buffer (50 mM HEPES, pH 7.2, 100 mM NaCI, 1% Triton X-100, + proteases inhibitor cocktail (Sigma)). Total protein was measured and biotinylated proteins were pulled down from equal amounts of extracts usingstreptavidin Sepharose beads (Pierce) (4°C, overnight on a rotation wheel). After washing, the bound material was eluted from the beads using 2x loading buffer (Invitrogen) containing 2% 0-mercaptoethanol (70°C for 10 min), separated on precasted 4-12% Bis-Tris NuPAGE gels in MES running buffer (Invitrogen) and processed for Western blotting and immunodetection. For each data point, three samples were prepared, one being non-reduced (N R) but kept at 4°C during the whole procedure (hence representing the total pool of surface biotinylated proteins), one reduced (R0) but kept at 4°C (to monitor the efficiency of reduction of biotin) and finally one reduced (R37) following internalization at 37°C (representing the internalized pool of proteins). To compare the internalization efficiency for each protein, the (R37-R0)/NR ratio for each experiment was measured and normalized to the control (control = 100). Each experiment was performed at least three times and statistical significance was measured using two-sided Student t-test. Values are presented as mean ± SEM.

Protein determination and Western blotting

Protein concentrations were determined by the Bio-Rad DC protein assay (Bio-

Rad) as described by the manufacturers. Samples were separated by SDS-PAGE (4-12% Bis-Tris NuPAGE gels in MES running buffer (Invitrogen) and transferred onto nitrocellulose membranes (Invitrogen). After blocking in 5% non-fat milk, membranes were incubated with primary antibody (4°C, overnight) fol lowed by washing a nd incu bation with horserad ish peroxidase ( H RP)-conjugated secondary antibodies (lh, RT). After final washing, immunodetection was done using a chemiluminescence reaction (Western Lightning-Plus ECL reagent (PerkinElmer)), and immunoreactive protein bands were digitally processed and quantified on a Fuji MiniLAS 3000 imager (Fuji) and using Aida software raytest (Isotopenmessgerate GmbH, Germany).

Metabolic labelling

24 hrs after transfecting HeLa cells with pcDNA-APPsw, cells were washed twice with serum-free and once with methionine- and cysteine-free medium (Sigma) supplemented with 2 mM L-glutamine, 0.5 m M sodiumpyruvate, and Pen/Strep (100 units/ml) (Invitrogen) and incu bated (10 min at 37°C). Medium was replaced with the same fresh methionine- and cysteine-free medium (Sigma) but additionally supplemented with 0,07 mCi [S³⁵] translabel (EasyTag Express Protein labelling Mix, Perkin Elmer). After 3 hrs at 37°C, the conditioned medium was collected and the cell layer was washed once in PBS, and then scraped in extraction buffer (lx Tris-buffered saline (TBS: 50 mM Tris.HCI (pH 7.4) and 150 mM NaCI), 1% Triton X-100, Complete protease inhibitor cocktail (Roche Diagnostics)). Both the conditioned media and cell extracts were centrifuged (14,000 rpm, 20 min) to remove detached cells and unsolubilized material). Cleared conditioned media were first immunoprecipitated with mab 4G8, then the unbound fraction was split in two and used to immunoisolate secreted ΑΡΡβ ectodomain and Αβ peptides using pab ANJJ (4μΙ) [4] and 6E10 (3μΙ), respectively, together with protein G-Sepharose (Pharmacia) (overnight,at 4°C, on a rotating wheel). Full-length APP and APP-CTFs were isolated likewise from extracts using pab B63. Immunoprecipitates were washed five times in extraction buffer, once in TBS. and bound material was eluted with 2x sample buffer (Invitrogen) containing 2% 0- mercaptoethanol (10 min at 70°C). Immunoisolates from cell extracts were electrophoresed on 10% M ES gel s ( I nvitrogen) wh ile for conditioned med ia imm u noprecipitates 7% Tris-Acetate gels (Invitrogen) were used. After SDS-PAGE, gels were dried and radiolabeled bands were detected using a Phosphorlmager ( Molecu la r Dynamics, I nc. ) a nd ana lysed usinglmageQua NT 5.1. Levels of APP fragments (secreted Αβ ,≤ΑΡΡβ and APP-CTF) were normalized to the expression level of full-length APP. Data are presented as mean ± SEM.

Retroviral infection

Constructs containing the human ARF6 WT and T157A proteins preceded by the HA-tag were kindly provided by J. Donaldson (Laboratory of Cell Biology, NHLBI, National Institutes of Health, Bethesda, Maryland 20892, USA). Human ARF6 constructs were cloned in the retroviral vector pMSCV*- puromycin (Clontech Laboratories; containing an extended multiple cloning site). Cloned pMSCV*- constructs, verified by sequencing, were used for the generation of retroviral particles via co- transfection with the helper plasmid ρΐκ (Ecopac) in HEK293T cells for packaging of the retroviruses. Viruses were harvested, and particles were either used directly or snap-frozen and aliquots were stored at -80°C until use. PSEN l&2dKO M EFs were used for transduction with retrovirus for 24h followed by puromycin selection in Dulbecco's modified Eagle's medium-F12 supplemented by 10% FCS and 5μg/ml puromycin, and later su bcloned to obtain single colonies with the same genetic background.

Cloning of human ARF6 WT-HA in pMSCV*

Cloning of human ARF6 WT-HA in pGEMt- Strategy: Digestion of pCDNA3.1+ ARF6 WT-HA with ECORI and Xbal

Klenow of ARF6 WT-HA fragment

Adding an A-overhang using TAQ to ARF6 WT-HA fragment

Ligation of insert in pGEMt vector

Cloning of human ARF6 WT-HA in pMSCV*

Strategy:

Digestion of vector pMSCV* and pGEMt + ARF6 WT-HA with Ncol and Sail Ligation of ARF6 WT-HA into pMSCV*

Oligo's

pGEM-T Forward = 2261= TAA TAC GAC TCA CTA TAG GGC GA (SEQ ID NO: 3) pGEM-T Reverse = 2262= AAG CTA TTT AGG TGA CAC TAT AGA A (SEQ ID NO: 4) pMSCV Forward = 2259= CCCTTGAACCTCCTCGTTCGACC (SEQ ID NO: 5) pMSCV Reverse = 2260= GAGACGTGCTACTTCCATTTGTC (SEQ ID NO: 6)

Cloning of human ARF6 T157A-HA in pMSCV*

Normal site directed mutagenese on template pGEM-t+ARF6 WT-HA

Colony PCR with Forward and Reverse pGemt primers

Digestion of vector pMSCV* and pGEMt +ARF6 T157A-HA with Ncol and Sail

Ligation

Oligo's: wa 2322_ARF6wt_T157A_F: CCTCCTGTGCCGCCTCAGGGGACG (SEQ ID NO: 7) wa 2323_ARF6wt_T157A_R: CGTCCCCTGAGGCGGCACAGGAGGG (SEQ ID NO: 8) pGEM-T Forward = 2261= TAA TAC GAC TCA CTA TAG GGC GA (SEQ ID NO: 9) pGEM-T Reverse = 2262= AAG CTA TTT AGG TGA CAC TAT AGA A (SEQ ID NO: 10)

Lentiviral infection

SMARTvector Lentiviral shRNA particles (thermo-scientific) were used for the stable knockdown of ARF6 in WT MEFs via lentiviral transduction. Different MOIs were used to obtain the best possible knockdown, following the manufacturing protocol. Stable cell lines were obtained after puromycin selection in Dul becco's modified Eagle's mediu m-F 12 su ppl emented by 10% FCS a nd 5μg/m l puromycin, and later subcloned to obtain single colonies with the same genetic background. Biological valorization assays

MEFs will be plated out in 96-wells and treated with DSMO or DSMO+active compound (concentration range as above) for 24 or 48hrs and next processed for the listed read-outs.

To localize caveolin 1 (Del Pozo et al., 2005), after treatment, cells will be fixed (4% paraformaldehyde/ 4% sucrose in PBS, 20min RT), permeabilized (0,1% Triton in PBS, 5min RT) and blocked ( 2% bovine serum albumin (BSA), 2% foetal bovine serum (FBS), 1% gelatin, 2% goat serum in PBS, lhr, RT). Primary antibody to caveolin 1 and tubulin (as a control) will be diluted in the same blocking buffer and applied to fixed cells (4°C overnight). Following washes in PBS, immunolabel will be visualized using appropriate secondary antibodies conjugated to either Alexa-488, -568, -647 or Pacific Blue (Invitrogen) diluted in blocking buffer (lhr, RT). Finally, cells will be analyzed on the InCell 2000, or alternatively on the Zeiss Radiance2100 confocal microscope. Data acquisition and processing will be with Lasersharp, ImageJ and Photoshop.

To identify acidic compartments, live cells will be, after treatment with the lead compounds (same concentration range and incubation time), incubated with Lysotracker Red (Invitrogen) for lh at 37°C in normal growth medium containing 0.2μΜ LysoTracker (Molecular Probes; Invitrogen). Cells are washed with PBS-/- and fixed with 4% paraformaldehyde for 20min at room temperature and processed for imaging.

To monitor fluorescently tagged EGF(lnvitrogen), pulse-chase experiments are performed. Cells are labeled at 4°C with xng/ml labeled EGF for lOmin, and briefly washed with PBS-/- before the chase period starts. EGF-488 is chased for 10, 30 and 60 min at 37°C, later fixed with 4% paraformaldehyde for 20min at room temperature and processed for imaging accordingly.

To quantify APP processing products (secreted Αβ, solube APP ectodomain fragments and APP- Carboxyterminal fragments (CTF)(Annaert et al., 1999), cells are transfected 24 hrs after compound treatment with pcDNA-APPsw. 24hrs later cells are washed twice with serum-free and once with methionine- and cysteine-free medium (Sigma) supplemented with 2 m M L-glutamine, 0.5 mM sodiumpyruvate, and Pen/Strep (100 units/ml) (Invitrogen) and incubated (10 min at 37°C). Medium are replaced with the same fresh methionine- and cysteine-free medium (Sigma) but additionally supplemented with 0,07 mCi [S³⁵] translabel (EasyTag Express Protein labelling Mix, Perkin Elmer). After 3 hrs at 37°C, the conditioned medium is collected and the cell layer washed once in PBS, and then scraped in extraction buffer (lx Tris-buffered saline (TBS: 50 mM Tris.HCI (pH 7.4) and 150 mM NaCI), 1% Triton X-100, Complete protease inhibitor cocktail (Roche Diagnostics)). Both the conditioned media and cell extracts are centrifuged (14,000 rpm, 20 min) to remove detached cells and unsolubilized material. Cleared conditioned media are first immunoprecipitated with mab 4G8, then the unbound fraction is split in two and used to immunoisolate secreted ΑΡΡβ ectodomain and Αβ peptides using pab ANJJ (4μΙ) ( ajendran et al., 2006) and 6E10 (3μΙ), respectively, together with protein G-Sepharose (Pharmacia) (overnight,at 4°C, on a rotating wheel). Full-length APP and APP-CTFs are isolated likewise from extracts using pab B63. Immunoprecipitates are washed five times in extraction buffer, once in TBS. and bound material was eluted with 2x sample buffer (Invitrogen) conta i n i ng 2% b-mercaptoethanol (10 min at 70°C). Immunoisolates from cell extracts are electrophoresed on 10% MES gels (Invitrogen) while for conditioned media immunoprecipitates 7% Tris-Acetate gels (Invitrogen) are used. After SDS-PAGE, gels are dried and radiolabeled bands detected using a Phosphorlmager (Molecular Dynamics, Inc.) and analysed using ImageQuaNT 5.1. Levels of APP fragments (secreted Αβ ,≤ΑΡΡβ and APP-CTF) are normalized to the expression level of full-length APP.

Assay development

A screening assay is set up that can be dissected in different steps. First cells are brought into culture and several cell lines are maintained. These cells are treated with compounds, incubated and afterwards fixed and stained. These stained cells are imaged and the images are further analyzed.

Maintaining of MEF cell cultures. MEF (mouse embryonic fibroblasts) cell lines are passaged (every 3 days at confluency of 90%) under sterile circumstances in T75 bottles appropriate for adhering cells. Two T75 bottles were used per cell line to provide enough cells for the experiments. A sterile environment is created by usage of a LAF-cabinet (laminar airflow cabinet) and a solution of 70% ethanol. First the cells are washed with phosphate buffered saline (PBS [-] CaCI₂ [-] MgCI₂) to remove dead cells and remaining medium. After trypsinizing the cells two or three minutes fresh DMEM + L- glutamine + HEPES + 10% FBS + PENSTREP medium is added to inactivate the trypsin activity. Then a part of the cell suspension is transferred to a new bottle with fresh medium. In MEF PSdKO + hArf6 cl.10 cells 5μg/μL puromycin is added to keep them selected for ARF6 expression. All cell suspensions are maintained at 37°C and passed through a cell strainer to obtain single cells. The concentration of this cel l suspension is ca lculated by usage of a Neu bauer improved hemacytometer. This cell suspension is d il uted to 200.000cel ls/ml . At least 3m l cel l suspension of each cel l l ine at th is concentration is needed to manually plate

into a 384-well-plate.

Liquid handling. To handle the 384-well plates, the Freedom EVOware 150 pipetting robot (TECAN) is used according to Manufacturer's instructions. In these experiments only the multichannel head and the washing station are used to dilute the compounds of the masterplate, to treat the cells with compounds, to fix the cells and to stain them afterwards. Treat cells with compounds. First the masterplate with compounds (96-well plate) is diluted so the concentration of the compounds is lowered from 25mM to ΙΟΟΟμΜ. Two 384-well plates are used to perform the experiment in duplicate, the cells are manually added with a multichannel pipette (8 channels). After 12 hours in culture, the cells are attached to the bottom of the 384-well plate and ready to be treated. Every cell line in one plate is treated once with a final concentration of 50μΜ of every compound. Again the cells are incubated for 24 hours in a C0₂-incubator. All 3 different cell lines and a negative control, consisting of DMSO, are plated in a 384-well plate. The wells only filled with DMSO serve as negative controls for the fluorescence of the compounds. The first (1 & 2) and last (23 & 24) two columns contain cells but are not treated with compound so they serve as negative controls for the cell morphology. This plate is made twice to provide duplicate experiments.

Fixing and staining. After the incubation of the cells with compounds they are fixed for 15 minutes with a 4% PFH (paraformaldehyde) solution. To stain the cells, the liquid is first removed from the fixed cells, followed by three washes with PBS [+] CaCI₂ [+] MgCI₂. A 0,1% Triton X-100 solution (in PBS [+] CaCI₂ [+] MgCI₂) is added to permeabilize the cells and to allow phalloidin-TRITC to enter the cells. After an incubation time of three minutes, the cells are washed twice with PBS [+] CaCI₂ [+] MgCI₂. Before adding the dye, aspecific interactions are avoided by incubating the cells for 15 minutes with blocking buffer (PBS [-] CaCI₂ [-] MgCI₂ containing + 2% BSA + 2% FBS + 0,2% gelatin). Now the phalloidin-TRITC dye diluted to 1μί/500μί from a 2,5mg/mL stock solution (with PBS [+] CaCI₂ [+] MgCI₂), is added to the cells and incubated for 25 minutes. The final concentration of the dye is 0,005mg/mL. After staining the cells with phalloidin, a DAPI staining was performed as a control. First the phalloidin dye was removed and the cells were washed again twice with PBS [+] CaCI₂ [+] MgCI₂. A DAPI stock solution of lmg/mL is diluted in a rate of 1μί/700μί (in PBS [+] CaCI₂ [+] MgCI₂). This staining solution is added to the cells. Because the used dyes are fluorescent, they lose some of their strength when exposed to light; so they are covered with aluminum foil and stored in the refrigerator. Every time a liquid is used, a volume of 50μί is added.

Imaging. After fixing and staining, images were taken with the InCell Analyzer 2000 (GE Healthcare). These images were obtained from 384-well plates with a lOx objective lens (Nikon lOx - NA 0,45 - Plan Apo - CFI/60), a large chip CCD camera (CoolSNAP K4 - 2048 x 2048 pixel array - 7,40μιτι square pixel) and the QUAD1 polychroic mirror. To save time two fields per well were imaged and for focusing, the hardware autofocus (laser) was used. For the phalloidin-TRITC staining three wavelengths were used. The DAPI excitation and emission filters were used with an exposure time of 0,050sec. The Cy3 excitation and emission filters were used with an exposure time of 2,000sec and the brightfield excitation (transmitted light) and the DAPI emission filter were used with an exposure time of 0,050sec to obtain a brightfield image.

Analysis. The acquired images were further automatically analyzed with the InCell Workstation 3.5 software (GE Healthcare, Amersham, United Kingdom). For the analysis, five parameters are calculated to study changes in the cell morphology, namely the cell intensity, cell count, cell roundness, cell area (size) and cell elongation (length). The ratio of the short over the long axis of the cells is called the cell elongation. A not elongated or symmetric cell has a value of one. All the other values situate between zero and one. The cell roundness is also called the cell 1/form factor. It is calculated via parameter (perimeter) over area. These values lie between one and + infinity. A cell with a roundness of one, forms a perfect circle.

Compounds. The compound screening collection at the Compound Screening Facility of VIB (Gent) amounts to a total of 42,000 compounds. The collection comprises three different chemical libraries acquired through ChemBridge Corporation (http://chembridge.com/chembridge/); DIVE Set™ (22,000 compounds), CNS-Set™ (10,000 compounds) and NOVACore (10,000 compounds). For selection of the DIVERSet™ compounds, a range of filtering methods are applied to ensure maximal diversity with a minimal num ber of compound, and to remove unstable, toxic and non-drug-like compounds. The DIVERSet™ library can be used in initial screening programs that require high diversity and qualitative lead-like compounds. For the CNS-Set™ library, additional computational methods are applied to select compounds with increased proba bility of oral bioavailability and blood-brain-barrier penetration. N OVACo re is a d iverse a nd d rug-like library comprising compounds that are synthesized via combinatorial chemistry. The main focus of NOVACore is 'novelty'; all compounds are recently synthesized by ChemBridge and contain mainly proprietary ChemBridge building blocks. All NOVACore compounds have a low molecular weight, which allows more margin for further lead optimization.

REFERENCES

- Annaert W.G., et a!. Presenilin 1 controls γ-secretase processing of the amyloid precursor protein in pre- Golgi compartments of hippocampa neurons. J. Ce l BioL, 147(2), 277-294, 1999.

- Baiasubrarnaman, N., et al., ARF6 and microtubules in adhesion-dependent trafficking of lipid rafts.

Nat Cell Biol, 9, 1381-91 (2007).

- Bartus RT, Dean RL, 3rd, Beer B, Lippa AS, The cholinergic hypothesis of geriatric memory dysfunction. Science 217, 408 (1982).

- Bonifacino, J. S. (2004). "The GGA proteins: adaptors on the move." Nat Rev Mol Cell Biol 5(1): 23- 32.

- Brown, F. D., A. L. Rozelle, et al. (2001). "Phosphatidylinositol 4,5-bisphosphate and ARF6-regulated membrane traffic." J Cell Biol 154(5): 1007-17.

- Capell, A., H. Steiner, et al. (2000). "Maturation and pro-peptide cleavage of beta-secretase." J Biol Chem 275(40): 30849-54.

- Carey, R. M., B. A. Balcz, et al. (2005). "Inhibition of dynamin-dependent endocytosis increases shedding of the amyloid precursor protein ectodomain and reduces generation of amyloid beta protein." BMC Cell Biol 6: 30.

- D'Souza-Schorey, C. and P. Chavrier (2006). "ARF proteins: roles in membrane traffic and beyond." Nat Rev Mol Cell Biol 7(5): 347-58.

- Del Pozo, .A. et al., Integrins regulate Rac targeting by internalization of membrane domains.

Science, 303, p. 839-42 (2004).

- del Pozo, .A., et al., Phospho-caveolin-1 mediates integrin-regulated membrane domain internalization. Nat Cell Biol, 2005, 7(9): p. 901-8.

- De Strooper, Aph-1, Pen-2, and Nicastrin with Presenilin generate an active gamma-Secretase complex. Neuron 38, 9-12 (2003).

- Esselens et al. J Cell Biology; 2004 September; 166(7): 1041-1054

- Franco, M., P. J. Peters, et al. (1999). "EFA6, a sec7 domain-containing exchange factor for ARF6, coordinates membrane recycling and actin cytoskeleton organization." Embo J 18(6): 1480-91.

- Frittoli, E., A. Palamidessi, et al. (2008). "The primate-specific protein TBC1D3 is required for optimal macropinocytosis in a novel ARF6-dependent pathway." Mol Biol Cell 19(4): 1304-16.

- Golde TE, Dickson D, Hutton M. Filling the gaps in the abeta cascade hypothesis of Alzheimer's disease. Curr Alzheimer Res. 3:421-30 (2006).

- Grande Garcia, A. et al. , Caveolin-1 regulates cell polarization and directional migration through

Src kinase and Rho GTPases. J Cell Biol, 177, 683-94 (2007). - Grant, B. D. and J. G. Donaldson (2009). "Pathways and mechanisms of endocytic recycling." Nat Rev Mol Cell Biol 10(9): 597-608.

- Hardy J, Selkoe DJ, The amyloid hypothesis of Alzheimer's disease: progress and problems on the road to therapeutics. Science 297, 353-6 (2002).

- Hashimoto, S., et al., Requirement for ARF6 in breast cancer invasive activities. Proc Natl Acad Sci U S A, 2004. 101(17): p. 6647-52.

- He, X., W. P. Chang, et al. (2002). "Memapsin 2 (beta-secretase) cytosolic domain binds to the VHS domains of GGA1 and GGA2: implications on the endocytosis mechanism of memapsin 2." FEBS Lett 524(1-3): 183-7.

- He, X., F. Li, et al. (2005). "GGA proteins mediate the recycling pathway of memapsin 2 (BACE)." J Biol Chem 280(12): 11696-703.

- Huse, J. T., D. S. Pijak, et al. (2000). "Maturation and endosomal targeting of beta-site amyloid precursor protein-cleaving enzyme. The Alzheimer's disease beta-secretase." J Biol Chem 275(43): 33729-37.

- Iwatsubo T, Odaka A, Suzuki N, Mizusawa H, Nukina N, lhara Y, Visualization of A beta 42(43) and A beta 40 in senile plaques with end-specific A beta monoclonals: evidence that an initially deposited species is A beta 42(43). Neuron 13, 45-53 (1994).

- Jackson, T. R., F. D. Brown, et al. (2000). "ACAPs are arf6 GTPase-activating proteins that function in the cell periphery." J Cell Biol 151(3): 627-38.

- Jaworski, J. (2007). ARF6 in the nervous system. Eur J Cell Biol 86, 513-524

- Kabat EA, Wu TT, Identical V region amino acid sequences and segments of sequences in antibodies of different specificities. Relative contributions of VH and VL genes, min igenes, and complementarity-determining regions to binding of antibody-combining sites. The Journal of Immunology, 147(5):1709-19 (1991).

- Lanzetti, L., et al., Rab5 is a signalling GTPase involved in actin remodelling by receptor tyrosine kinases. Nature, 2004. 429(6989): p. 309-14.

- Mullan, M., F. Crawford, et al. (1992). "A pathogenic mutation for probable Alzheimer's disease in the APP gene at the N-terminus of beta-amyloid." Nat Genet 1(5): 345-347.

- Naslavsky, N., . Weigert, et al. (2003). ^"Convergence of non-ciathrin- and clathrin-derived endosornes involves ARF6 mactivation and changes in phosphoinositides." Mol Biol Cell 14(2): 417- 31.

- Naslavsky, hi,, R. Weigert, et al. (2004). "Characterization of a nonc!athrin endocytic pathway: membrane cargo and lipid requirements." Mol Biol Ceil 15(8): 3542-52. - Ohno M, Sametsky EA, Younkin LH, Oakley H, Younkin SG, Citron M, Vassar R, Disterhoft J F, BACEl Deficiency Rescues Memory Deficits and Cholinergic Dysfunction in a Mouse Model of Alzheimer's Disease. Neuron, 41:27-33 (2004).

- Ohno M, Chang L, Tseng W, Oakley H, Citron M, Klein WL, Vassar R, Disterhoft J F Temporal memory deficits in Alzheimer's mouse models: rescue by genetic deletion of BACEl. Eur J Neurosci, 23:251-260 (2006).

- Pastorino, L., A. F. Ikin, et al. (2002). "The carboxyl-terminus of BACE contains a sorting signal that regulates BACE trafficking but not the formation of total A(beta)." Mol Cell Neurosci 19(2): 175-85.

- Peters, P. J., V. W. Hsu, et al. (1995). "Overexpression of wild-type and mutant ARF1 and ARF6: distinct perturbations of nonoverlapping membrane compartments." J Cell Biol 128(6): 1003-17.

- Rajendran, L., M. Honsho, et al. (2006). "Alzheimer's disease beta-amyloid peptides are released in association with exosomes." Proc Natl Acad Sci U S A 103(30): 11172-7.

- Rink, J., E. Ghigo, et al. (2005). "Rab conversion as a mechanism of progression from early to late endosomes." Cell 122(5): 735-749.

- Roberds SL, Anderson J, Basi G, Bienkowski MJ, Branstetter DG, Chen KS, Freedman SB, Frigon N L, Games D, Hu K, et al BACE knockout mice are healthy despite lacking the primary β- secretase activity in brain : implications for Alzheimer's disease therapeutics. Hu m Mol Genet, 10: 1317-1324 (2001).

- Schneider, A., L. Rajendran, et al. (2008). "Flotillin-dependent clustering of the amyloid precursor protein regulates its endocytosis and amyloidogenic processing in neurons." J Neurosci 28(11): 2874-82.

- Selkoe DJ, Alzheimer's disease: genes, proteins, and therapy. Physiol Rev 81, 741-66 (2001).

- Shiba, T., S. Kametaka, et al. (2004). "Insights into the phosphoregulation of beta-secretase sorting signal by the VHS domain of GGA1." Traffic 5(6): 437-48.

- Stenmark, H., R. G. Parton, et al. (1994). "Inhibition of rab5 GTPase activity stimulates membrane fusion in endocytosis." Embo J 13(6): 1287-1296.

- Tesco, G., Y. H. Koh, et al. (2007). "Depletion of GGA3 stabilizes BACE and enhances beta-secretase activity." Neuron 54(5): 721-37.

- Traub, L. M. (2009). "Tickets to ride: selecting cargo for clathrin-regulated internalization." Nat Rev Mol Cell Biol 10(9): 583-96.

- von Arnim, C. A., M. M. Tangredi, et al. (2004). "Demonstration of BACE (beta-secretase) phosphorylation and its interaction with GGA1 in cells by fluorescence-lifetime imaging microscopy." J Cell Sci 117(Pt 22): 5437-45. - Wahle, T., K. Prager, et al. (2005). "GGA proteins regulate retrograde transport of BACEl from endosomes to the trans-Golgi network." Mol Cell Neurosci 29(3): 453-61.

- Wilson et al. J Cell Biology; 2004 May; 165(3): 335-346).

- Winter G, Harris WJ, Humanized antibodies. Trends Pharmacol Sci. 14{5):139-43 (1993).

- Zhao, X,, T. Greener, et al. (2001). "Expression of auxiSin or AP180 inhibits endocytosis by misiocalizing clathrin: evidence for formation of nascent pits containing API or AP2 but not c!athrin." J Ceil Sci 114(Pt 2): 353-65.

Claims

1. Use of ARF6 and/or an ARF6 effector to reduce amyloid beta peptide formation in a mammalian cell.

2. Use of ARF6 and/or an ARF6 effector to identify compounds that reduce amyloid beta peptide formation in a mammalian cell.

3. Use according to any of claims 1 or 2 wherein said ARF6 effector is chosen from a GAP, such as GIT-1, or a GEF, such as EFA6A.

4. Use according to any of claims 1 to 3 to prevent or treat Alzheimer's disease.

5. Use of ARF6 and/or ARF6 effector to diagnose or prognose Alzheimer's disease.

6. An agent characterized in that it is reducing amyloid beta peptide formation, wherein said agent is a small interfering RNA (siRNA) or an antisense polynucleotide or a ribozyme, and wherein said agent comprises a nucleic acid sequence engineered from an ARF6 encoding polynucleotide.

7. A pharmaceutical composition comprising a therapeutically effective amount of the agent according to claim 6 and at least one of a pharmaceutically acceptable carrier, adjuvant or diluent.

8. The agent according to claim 6 or the pharmaceutical composition according to claim 7 for use in the prevention and/or treatment of Alzheimer's disease.

9. An isolated cell line lacking endogenous presenilin expression/function and further characterized by stably expressing ARF6.

10. A method for identifying compounds that modulate the endosomal redistribution in a mammalian cell comprising the steps of:

b) administering a test compound to said cell culture;

c) imaging at least one morphological parameter of the cells in said cell culture; wherein, under the same test conditions, a deviation in said at least one parameter compared to the same at least one parameter of cells of a corresponding wild type cell culture and/or of a cell culture from the cell line of claim 9, identifies said test compound as a compound that modulates the endosomal redistribution in a mammalian cell.

11. The method according to claim 10 wherein said compound modulates ARF6 cycling activity and/or ARF6 effector protein activity.

12. The method according to any of claims 9 to 11 wherein said compound is capable of reducing amyloid beta peptide formation in a mammalian cell.

13. The method according to any of claims 9 to 12 wherein said compound is a therapeutic candidate for the prevention and/or treatment of Alzheimer's disease.