Final Report Summary - NAD (NANOPARTICLES FOR THERAPY AND DIAGNOSIS OF ALZHEIMER DISEASE)
Recent statistics indicate that over three million people in the EU have Alzheimer disease (AD). This figure is destined to rise as the population ages, doubling by 2040 in Western Europe and trebling in Eastern Europe. Therefore, the search for effective therapies and early diagnosis is imperative. A hallmark of the disease in the AD brain is the abnormal accumulation of the peptide β-amyloid (Aβ) forming extracellular aggregates: oligomers, considered the most neurotoxic form of aggregation and plaques, detectable in the brain of AD patients. This project intended to synthesize and to use nanoparticles (NPs) for AD therapy and diagnosis, singly or combined (theranostics), focusing on brain Aβ as the target. Brain and blood Aβ are in equilibrium across the blood-brain barrier (BBB), so the project also considers blood Aβ as a target.
Different NPs (liposomes, solid lipid NPs, polymeric NPs) have been multiple-functionalized with: i) molecules interacting with Aβ, ii) molecules stimulating BBB crossing, ii) PET or MRI contrast agents. Artificial and cellular models have been used to improve and fine-tune NP binding to Aβ, biocompatibility, BBB crossing and physical stability, keeping to a minimum the use of animals, although the efficacy of NPs has eventually been evaluated in rodent models of AD. Different routes of administration (i.v. oral, nasal) and different protocols have been employed to boost NP brain delivery. The Consortium has realized NP that showed the ability, once injected in transgenic mouse models of AD, to disaggregate and remove brain Aβ, either plaques or, noticeably, oligomeric aggregates which are considered the most toxic for of the peptide. Besides this action, the treatment with NPs has been able to restore animal impaired memory.
Project Context and Objectives:
Recent statistics estimate that 24.3 million people have dementia today, with 4.6 million new cases of dementia every year (one new case every 7 seconds). In the EU, about 5 million people have dementia, with Alzheimer’s disease (AD) accounting for over 3 million. This is likely to increase as the population of old people grows, double this figure predicted by 2040 in Western Europe and treble in Eastern Europe.
AD causes continuous deterioration of higher nervous functions, due to progressive and irreversible neuro-degeneration of the limbic and association cortices, leading to the total loss of autonomy and eventually to death from cachexia or opportunistic infections. One in 20 people over 65, and one in five over 85, have AD. Given the constant increase of the elderly population in Europe, more healthcare, social and economic resources will be required, with huge social and financial costs.
Although substantial progress has been made in the scientific understanding of AD, there remains an urgent need to identify effective therapies and early detection strategies, in order to avert a financially overwhelming public health problem. Genetic, pathological and biochemical clues suggest that the progressive production and subsequent accumulation of β-amyloid (Aβ), a proteolytic fragment of the membrane-associated amyloid precursor protein (APP), play a central role. This peptide is released from cells in a soluble form that progressively forms oligomeric, multimeric and fibrillar aggregates, ending with extracellular plaques, one of the morphological hallmarks of the disease, detectable post-mortem in AD brains. Oligomers are considered the most toxic form of Aβ aggregation. Accumulation of abnormally hyper-phosphorylated protein tau in neurons in the form of neurofibrillary tangles, another hallmark of AD, is postulated to be secondary to Aβ deposit (Fig. 1).
Finally, brain and blood Aβ are in equilibrium through the blood brain barrier (BBB), and sequestration of Aβ in the blood may shift this equilibrium, drawing out the excess from the brain (“sink” effect) (Matsuoka Y et al., J Neurosci. 23,29,2003 ; Sagare A., Nature Med. 13, 1029, 2007).
All these premises strongly suggest an approach to both the therapy and diagnosis of AD, based on Aβ as the target. Nanoparticles (NPs) offer an attractive mean of achieving this task, with their high potential for surface multi-functionalization. Multi-functionalization may confer on NPs a wide array of properties, such as stealth characteristics once they enter the bloodstream (ability to avoid the reticulo-endothelial system, RES), ability to cross the blood-brain barrier (BBB) and, of course, ability to interact with Aβ. In addition, since it is possible to confer on them features of biocompatibility, lack of toxicity and immunogenicity, biodegradability, simple preparation and high physical stability, the NP-based system is an intelligent tool for diagnostics, prognostics, and controlled and sustained delivery of therapeutic agents to specific targets.
This project intended to use NPs specifically engineered for targeting of Aβ, for the diagnosis and therapy of AD.
This has been pursued: 1) By seeking an interaction with the different pools and forms of Aβ (monomers, oligomers, plaques). The interaction with soluble extracellular Aβ could inhibit its oligomerization, thus preventing amyloidogenesis and plaque formation. The interaction with aggregates, plaques included, could achieve their disaggregation in vitro and reduce toxicity in vitro. 2) By functionalization of NPs to cross the BBB.
3) By functionalization with MRI and PET contrast agents of NPs crossing the BBB and carrying Aβ ligands. 4) By seeking an interaction with Aβ circulating in the blood. NPs could interact with Aβ in the bloodstream, withdrawing the peptide from the brain through the “sink effect” (Fig. 2).
It is important to stress that none of the above approaches can reverse the neuronal cell loss. However, these interventions should limit further damage and/or delay the progression of the disease, or could help restore the activity of partially damaged neurons.
Ligands for Aβ peptide
Selection of most promising or most effective Aß ligands was an important issue in the first phase. The possibility to improve the efficiency of ligands already existing has also been taken into account, for instance by chemical modification. Moreover, the chemical nature of selected molecules has been modified, to accomplish the subsequent functionalization of NPs. The ligands chosen have been selected among small molecules or synthesized within the Consortium. The ligands utilized in the project were amphipatic lipids, anti-Aβ antibodies and naturally occurring molecules.
1. Amphipatic Lipids
Evidences from literature show that Aβ peptide, in particular the monomeric and oligomeric forms, strongly interacts with the plasma membrane (either raft or non-raft regions); these information allowed us to study Aβ-lipids interaction. We have performed a preliminary screening of the candidate lipids able to bind the Aβ peptide using TLC-immunostaining assay (Fig. 3). These results gave us a qualitative information about the preferential binding of Aβ to lipids. The data suggested that acidic phospholipids (phosphatidic acid (PA), cardiolipin (CL) and ganglioside GM1 have the strongest ability to bind Abeta in an in vitro system.
Antibodies are by definition one of the most specific ligand for Abeta. For this reason clones of cells producing antibodies against Abeta have been produced within the Consortium (Fig. 4).
3. Other naturally occurring ligands
Different classes of putative ligands have been selected after an exhaustive survey of literature data among the small molecules reported to be able to inhibit Aβ aggregation. We selected among ligand for Abeta to be used in the project and concentrated our efforts on curcumin as the most promising candidate (Fig. 5).
On the base of these information we have used:
Phosphatidic acid/Cardiolipin; Anti-Abeta antibodies, curcumin as ligands in NPs to perform experiments to check if they retain the ability to bind Abeta
NPs with stealth character have been prepared: The NPs chosen were SLN, liposomes or PEGylated polymeric NPs (Fig. 6). Phospholipid analogues have been synthesized for the generation of functionalisable lipid-based NPs. Monomers bearing the same group have been synthesized and used for the generation of surface-functionalisable polymeric NPs.
First of all, a protocol for NPs preparation has been set up, allowing a possible scale-up when a pharmacological prototype for the treatment of AD will be hopefully made available. Therefore, the necessity of using simple, cheap and fast methods for preparation of NPs has been taken into account as much as possible from the very beginning of the project. The fine tuning of NPs characteristics was the result of a careful programming that took into account different aspects: the major one to mention is that all molecules utilized for successive functionalization of NPs (conferring stealth features, ligands for Abeta, ligands for BBB) did not interfere, either because of size or chemical nature, with each other; related with this, were the chemical problems connected with the realization and the sequence of steps to perform such multiple functionalization.
NPs surface-decorated with ligands for Aβ peptide have been synthesized taking advantage of their amphiphilic nature or after chemical modification.
Antibodies - It was decided to use the biotin/streptavidin system for antibody conjugation on the surface of liposomes, after preparing liposomes that incorporate the appropriate molar amount of lipid-PEG-biotin conjugate. This methodology has been identified as a good method to use in in vitro studies. For in vivo studies, an alternative anchoring technique, as the formation of a covalent bond between a maleimide group on the NP surface and SH groups on the antibody, after their chemical thiolation, was exploited.
Curcumin - In order to generate NP able to easily link other curcumin, we designed and generated a monomer with a lipid moiety, a PEG spacer ending with an azido function to be used for the covalent linkage to alkyne bearing entities. For this purpose the native molecule of curcumin was modified and an alkyne moiety was introduced and linked to NP by click chemistry reaction.
Phosphatidic acid/Cardiolipin - Liposomes and SLN functionalized with acidic phospholipids have been prepared, exploiting their amphiphilic nature, by addition to the lipid mixture during the NP preparation
NPs decorated with Aβ ligands showed high affinity for Aβ peptide in all aggregation forms.
Surface Plasmon Resonance (SPR) and Ultracentrifugation have been utilized for these experiments. Liposomes or SLN containing phosphatidic acid (PA), cardiolipin or mixtures of gangliosides showed a tendency to bind Aβ 1-42 in different aggregation forms (oligomers and fibrils) immobilized on the sensor chip. Binding to Abeta monomers was generally less intense than the binding to Abeta oligomers and fibrils. The analysis of the curves suggested an high affinity of PA for Abeta fibrils, with a KD ≈ 6-60 nM. These results have been confirmed by experiments carried out by ultracentrifugation on a discontinuous sucrose density gradient, followed by dot-blot procedure.
Liposomes functionalized with curcumin have been also analyzed by SPR. The affinity values of NP with the curcumin-derivative, for Aβ1-42 fibrils was in the low nM range (2-10 nM) and this very high affinity likely involve avidity effects.
Surprisingly, also Pegylated Poly (Alkyl Cyanoacrylate) Nanoparticles were found to bind Aβ1-42 fibrils. Although the interaction between these NPs and Aβ species was clear, the corresponding sensorgrams could not be fitted by using the classic binding equations, suggesting complex interactions.
NP surface has been further decorated, in combination with Aβ ligands, with different molecules to promote the crossing of the Blood Brain Barrier (BBB).
Even if the final proof of principle of the ability of crossing the BBB has been performed in vivo, the screening of the molecules and of candidate NPs has been first carried out in vitro; for this purpose, suitable models have been set up and tested.
The ability of NPs to cross a BBB barrier cellular model has been evaluated using isotopically radiolabeled or fluorescently labeled NPs. The BBB model was constituted by a transwell system in which upper and lower chambers were separated by a layer of endothelial BBB cells. NPs to be tested were added to the upper chamber (mimicking the blood compartment) and their passage across the endothelial cell layer was evaluated by the amount recovered in the lower chamber (mimicking the brain compartment.) For these experiments in vitro either rodent capillary endothelial cells or human immortalized brain capillary endothelial cells have been utilized.
The molecules selected to enhance BBB crossing were: i) TAT, a cell-penetrating peptide from a protein sequence of human immunodeficiency virus (HIV) exploiting the adsorptive pathway, ii) anti-Transferrin receptor antibody (anti- TfR) exploiting receptor-mediated transcytosis, and iii) a modified peptide taken from the receptor-binding domain of apolipoprotein E (ApoE) exploiting in part the adsorptive and in part the receptor-mediated pathway.
The BBB ligands were covalently linked to a maleimide group inserted into NPs and exploiting the reaction with an SH- group present on the BBB ligand. For this purpose tat and ApoE peptides were added with a terminal Cysteine moiety, while anti-Tfr antibodies were submitted to chemical thiolation.
Functionalization with any of the ligand mediated an efficient NLs uptake by cultured endothelial cells that increased with ligand density. Furthermore we established that the chemical strategy chosen for the ligation (via a bioti- streptavidin bridge, or by covalent bond) affects the transport across the BBB model. NPs covalently bound and NLs carrying monomeric peptide from apolipoprotein E (mApoE) (residues 141-150) performed the best. mApoE-SLN achieved comparable results. Moreover, the permeability of a tritiated derivative of Curcumin was enhanced after its entrapment into mApoE-NLs.
Abbreviations: Chol, cholesterol; LIP, liposomes; mAb, monoclonal antibody; [14C]-PA, 14C-radiolabelled phosphatidic acid; [3H]-Sm, tritiated sphingomyelin; PA, phosphatidic acid; Sm, sphingomyelin.
The ability of NPs to bind Aβ, even after functionalization for BBB, has been demonstrated by different techniques (SPR, ultracentrifugation, ITC).
It has been established that the further functionalization for BBB crossing could affect the binding with Abeta of NPs designed for this scope. Therefore, the conditions were established ( maximal BBB ligand density ) in order not to affect these features
NPs functionalized for imaging
It was necessary to set up the techniques for incorporation of contrast agents into NPs and to test the instrumentation and the experimental conditions to follow their fate in vivo. These issues were studied with the NPs already available, in the absence of any functionalization. The use of non-functionalized NPs allowed the standardization of protocols Nps loaded with Iron oxides, or Gd or Fluorinated molecules, to be employed for MRI or PET have been prepared. Assessment of their applicability has been checked in vitro and in vivo.
Characterization of NPs their physicochemical features.
The precise knowledge of NPs features is necessary to assess the influence of any single operation, chemical or physical, on their structure and stability. Also the Abeta binding affinity can be affected by their multiple functionalization and for this reason this feature had to be followed continuously.
Since the final goal of the project was the preparation of NPs for therapy and diagnosis of AD, their use in vivo was intrinsic with their preparation. This opened the theoretical possibility that, beside a therapeutic action on diseased tissues, NPs may exert a toxic effect on healthy cells. Most of the components utilized for preparation of the NPs utilized in the present project were naturally-occurring molecules, and available literature data indicate a lack of toxicity. Even if these premises are positive, the assessment of biological impact of NPs was mandatory and has been followed for the duration of the project in different aspects: cellular uptake and storage of NPs; influence of NPs on biochemical functions of cell membranes; cellular metabolism of NPs; NPs haemocompatibility; NP complement activation.
Different techniques (Viability, MTT, hemolysis, NO production, ER stress markers, apoptosis..) and different cell lines or tissues have been utilized: fibroblasts, neuronal cell lines, endothelial cells (human umbilical, human brain capillary), macrophages, blood. In particular, the possible influence on complement activation has been evaluated; these experiments established the proportion of ligands to be administered in vivo.
All the studies showed that NPs functionalized to bind Aβ and to cross the BBB were biocompatible and stable in biological fluids.
Complement system studies clarified that liposomes functionalized with Aβ –binding amphiphilic lipids did not induce activation in AD sera when below a given size and below a given proportion of ligand.
Proof of principle in vitro
The objective was to establish the conditions to follow in vivo the performance of NPs for the therapy and diagnosis of AD. This objective included the development of analytical methods for the determination of NPs and/or their components, for instance the Abeta ligands, in biological fluids and in cellular systems. Moreover, the objective included the assessment of the conditions for MRI imaging of NPs functionalized with probes and prepared, and the detection ex vivo of amyloid deposits in diseased brain tissue. The final aim was to obtain an in vivo proof of principle for diagnosis and therapy of AD. The results showed that functionalized nanoparticles are able to rescue Aβ toxicity in cultured cells, increasing cell viability and reducing Tau phosphorylation. Moreover, other experiments showed that bi-functionalized NP are able to prevent Aβ peptide aggregation and, most importantly, are able to disaggregate Aβ supramolecular assemblies in vitro.
Molecular modeling experiments carried out in order to investigate the mechanisms of interaction between Aβ and NPs have clarified which chemical groups either of the peptide or on the NP contribute to the interaction.
Proof of principle in vivo has been carried out in wild type and in transgenic mouse models of AD.
NPs functionalized to bind Aβ
Liposomes functionalized with Phosphatidic acid or Cardiolipin
We found that liposomes functionalized to bind Abeta with Phosphatidic acid or Cardiolipin, i.p. administered to APP/PS1 transgenic mice for 3 weeks showed a decrease beta-amyloid in blood, more evident for Aβ 1-40 than for Aβ 1-42. However no statistically significant differences were observed in brain samples from treated mice. Some control protein as such as GAPDH, APP or BACE were not modified along treatments, however the analysis of neuronal elements such as Tau and the phosphorylation status of tau indicated some clear differences. For instance, in the phospho-epitope of AT100 is clearly reduced by the CL-Liposome treatment. In addition, when we analysed some putative Tau kinases we detected that the stress kinase JNK was significantly inhibited, as inferred from its “activation” phosphorylation levels.
Liposomes functionalized with anti Aβ antibody.
We found that liposomes functionalized to bind Abeta to APP/PS1 mice exerted a clear reduction of circulating beta amyloid after i.p. treatment for 3 weeks, for both Aβ 1-40 and Aβ 1-42. However the data obtained from brain samples showed a non statistically significant decrease of Aβ.
Liposomes functionalized to bind Abeta and to cross the BBB
As first, experiments carried out with radiolabeled NPs injected i.v. in healthy mice, showed that the functionalization to cross the BBB in vitro enhances the amount of radioactivity reaching the brain, suggesting the efficacy of the ligand chosen.
We found that liposomes bi-functionalized to cross the BBB and to bind Abeta, i.p. administered to APP/PS1 transgenic mice for 3 weeks, were able to restore impaired memory to normal, evaluated by Novel Object Recognition test. At the same time, histological analysis of brain sections showed a very important plaque disruption, with strong reduction of their number and area. This results has been obtained either histologically, by examination of brain sections of treated mice, or by Positron Emisssion Tomography (PET) using [11C-PIB] as a probe. Importantly, the amount of soluble oligomeric forms, which are recognized as the best correlate of synaptic dysfunction and disease severity, were decreased after treatment.
We also found that polymeric nanoparticles, either functionalized with curcumin derivative or linked to an anti-A antibody, significantly restored memory, although without affecting plaque load.
On the other end, other types of nanoparticles, such as mApoE-PA-SLN functionalized nanoparticles, were no as efficient in restoring memory in Tg mouse models as the aforementioned. The inefficacy of these last preparations may be linked to an insufficient dosage or to an inappropriate administration route or treatment schedule.
PET imaging. Therapeutic effect of nanoliposomes double functionalized was studied with single transgenic APP23 mice. The effectiveness of the nanoliposome treatment in these mice was followed with longitudinal [11C]PIB PET scans. Previously, [11C]PIB has been shown to be the most suitable tracer to longitudinally follow the changes in the Aβ deposition in the mouse brain
Longitudinal study (before, post and three months after the treatment) of LIPO-PA-mApoE in APP23Tg mice showed bound-to-free ratios similar to WT mice, in parallel with a reduced Aβ deposition in the brain.
MRI Imaging. Three different intrinsic MR-parameters were explored ( diffusion kurtosis imaging (DKI), resting state functional MRI (rsfMRI) and magnetic transfer contrast (MTC)) in differentiating characteristics of the brain of APP/PS1 mice from control mice. Histological evidence suggests that all 3 MR-parameters are sensitive to the severity of the amyloid pathology with associated neuroinflammation
All together these data, especially considering the evidence that some of the preparations tested did succeed in inducing significant positive effects, promote the nanoparticle-based therapy for AD as an innovative and suitable approach.
The present project addressed the necessity to develop new therapeutic and diagnostic tools to cope with Alzheimer’s Disease. The results of the project, carried out on a large set of cellular and animal models, will provide new chances for the treatment and the diagnosis of AD in humans; the combined use of therapy and diagnosis will allow the efficacy of therapy to be followed. The use of new diagnostic and therapeutic methods based on nanotechnology is one of the potential future answers to the immense societal and economic problems linked to AD. Future applications of the nanotechnology systems for AD treatment will facilitate progress towards a more effective and socially sensitive medicine and improve the quality of life for elderly. The results are also expected to have a positive long-term impact on a broad spectrum of the medical research landscape in Europe, by generating NP-based diagnostics and CNS-therapy systems for further basic and applied research in this complex medical area, thus reinforcing competitiveness of European industry.
51 articles have been published on the main Journals of Nanomedicine and Biochemistry
4 International Patents have been filed
List of Websites:
Project WEB site: www.nadproject.eu
Relevant contact details
Prof. Massimo Masserini
Università degli Studi di Milano Bicocca
Via Cadore 48 Monza, IT 20052 Italy
Mrs. Fiorenza Coviello
Università degli Studi di Milano-Bicocca
Via Cadore, 48 Monza – IT 20052 – Italy
Prof. András Dinnyés
BioTalentum Ltd. [BIO], Gödöllő, Hungary