Final Report Summary - NEURONANO (Do nanoparticles induce neurodegenerative diseases? Understanding the origin of reactive oxidative species and protein aggregation and mis-folding phenomena in the pr
Executive summary:
The EU FP7-funded project NEURONANO conducted an investigation into how nanoparticles interact with the human brain. The overall aim was to establish a risk assessment framework to categorise nanomaterials in terms of their potential for access to the brain, and potential for harm once there.
The NEURONANO project resulted in a number of key findings:
-The existing body of research into the impacts of nanomaterials was not sufficient for the project partners' purposes. As a result, they had to produce their own evidence base from scratch, using both cellular systems and carefully selected animal studies.
-Despite the brain being a 'sacrosanct' organ which most particles cannot enter, very small quantities of nanoparticles can pass through the blood-brain barrier (BBB). However, there is no evidence that the amounts of engineered nanomaterials that pass the barrier are sufficient to affect human health.
-Several new modes of interactions of nanoparticles with the BBB were observed, using the most advanced microscopy every applied to such systems. For example, a propensity for nanoparticles to accumulate in the lysosomes of the BBB endothelial monolayer was observed. Additional work is required to understand if similar effects are also observed in animals, as current approaches to quantify uptake in vivo do not distinguish between the nanoparticle load in the brain versus in the endothelium separating the brain from the blood (i.e. within the BBB).
References:
[1] Devasagayam, TPA et al., J. Association of Physicians of India (JAPI) 2004, 52: 796.
[2] Walsh DM, Teplow DB. Prog Mol Biol Transl Sci. 2012, 107:101-24.
[3] Calderon-Garcidueñas L, et al., Toxicol Pathol. 2007, 35:154-62.
Project Context and Objectives:
The overall science and technology objective of the NEURONANO project was to determine if engineered nanoparticles could constitute a significant neuro-toxicological risk to humans for two diseases endpoints, namely Alzheimer's and Parkinson's diseases. Importantly, the NEURONANO consortium did not presume a neurotoxic hazard from engineered nanoparticles / nanomaterials. Instead, the project sought to understand if exposure to nanoparticles resulted in them reaching the brain, and if so under what circumstances. Additionally, the project aimed to understand whether the presence of nanoparticles in the brain could exacerbate or accelerate existing cellular mechanisms such as the rate of generation of Reactive Oxygen Species (ROS) or the rate of fibrillation of proteins involved in the neurodegerative diseases of interest, amyloid-beta and alpha-synuclein, respectively.
The project was designed to be compliant with the European Commission's policy on reduction, replacement and refinement (the 3Rs) of the use of animals in scientific testing and experimentation. As such, much of the work was performed using in vitro methods, many developed specifically within the project, and which are now being further refined for onward development towards standardised alternative tests, both within the QualityNano research infrastructure project (see http://www.qnano-ri.eu online), and in other EU projects such as MARINA (see http://www.marina-fp7.eu/ online). Where animal studies were necessary (such as for assessment of the distribution of nanoparticles following exposure via different routes, and to assess the impacts of nanomaterials in the brain on animal behavior and amlyloid plaque load), there were designed to correlate and potentially confirm in vitro findings, and were designed to minimise the numbers of animals needed whilst still being scientifically robust. The tiered approach to the three stands of work relating to the assessment of neurodegenerative disease (solution studies, cell studies, and finally animal studies) represents a scientifically and ethically balanced approach to the work, balancing the necessary level of scientific excellence with the need to reduce the numbers of animal experiments.
At the time of development of the NEURONANO project, much of the knowledge of nanoparticle-induced oxidative stress has been derived (or deduced) from environmental or occupational exposure to small particulates including asbestos, coalmine dust, quartz, and PM10.[12-14] Oxidative stress caused by particles and other sources is capable of graded (dose-response-related), effects in cells.[15, 16] At low level there may be only very mild effects that are dealt with by the cells constitutive anti-oxidants and are not 'sensed'. At higher levels there can be activation of anti-oxidant defences, like catalase, to deal with mild oxidative stress. As the extent of oxidative stress increases, there is activation of stress-responsive signalling that leads to changes in gene expression for inflammation in epithelial cells, for example. At very high doses there can be cell death by necrosis or apoptosis.
Thus, there were several pathways by which nanoparticles might generate (and respond to) harmful oxidative stress in the brain:
1) Direct effects of nanoparticle-derived free radicals on neurones after contact or entry into them;
2) Contact or entry into neurones causing proteasomal or mitochondrial dysfunction;
3) Uptake by microglial cells causing activation of NADPH oxidase and an oxidative burst.
If indeed oxidative stress is an important mechanism in the brain, then it was anticipated that those particles most active in the generation of free radicals production would be more harmful than others.[17] However, given the complexity even in cell systems, it was possible that oxidative stress might be correlated not just with the degree of necrosis/apoptosis, but also with mitochondrial dysfunction, proteasomal dysfunction, induction of antioxidant defence, respiratory burst and inflammatory gene expression by microglial cells. To form a coherent picture the NEURONANO team proposed the application of gene arrays, proteomics and redox proteomics to determine the nano-particle-induced perturbation of cellular metabolism and oxidative stress pathways.
Whilst the precise mechanisms leading to neurodegenerative diseases are still not fully clarified (and were even less so at the time of initiation of the NEURONANO project), it is broadly agreed that the key effects involve the presence of early pre-fibrillar structures, neuroinflammation and ROS-related processes. All of the links in the causal chain are now present for a credible expectation that nanoparticles could have impacts on the onset and progression of neurodegenerative diseases, and thus present a neurotoxicity risk of greater or lesser significance depending on their physico-chemical composition and potential to access the brain.
Alzheimer's disease is characterized by loss of short-term memory and visual-spatial skills. It is diagnosed clinically based on characteristic neurological and neuropsychological phenomena. There is as yet no commonly accepted marker for the disease, and certain diagnosis is made only post-mortem. Recent research has generally considered the disease is caused by misfolding of either amyloid ß or tau proteins.[18] Evidence that amyloid ß is the key is provided by the fact that the Down syndrome patients (who almost invariably progress to Alzheimer's at an early age[19]) have an additional copy of the gene for amyloid ß precursor. There is also growing support for the hypothesis which suggests that the cytotoxic species is an intermediate oligomeric misfolded species of the protein, rather than the mature fibrils. It may be that Alzheimer's is primarily a disease driven by extracellular events where by Amyloid Precursor Protein (APP) is cleaved leading to release of the amyloid ß fragment containing residues 1-40/42, and that the resulting extracellular amyloid ß then forms the toxic oligomers (pre-fibrils) leading to the disease. Indeed the delivery of biological extracts containing these oligomers to mice leads directly to Alzheimer's-like symptoms.[6] Exposure to environmental toxins is believed to increase the likelihood of developing Alzheimer's disease, and there have been persistent claims that the ROS mechanism is also involved, but the manner is ill-understood as yet. Studies in cities with high pollution loads, where there are significant loads of nanoparticles of the size that can access the brain have suggested that Alzheimer's could be associated to this form of pollution, and it is now known that nanoparticles can dramatically affect the rate of protein fibrillation (including amyloid ß).[3]
Parkinson's disease is characterized by muscle rigidity, tremor, a slowing of physical movement, followed by (in extreme cases) a loss of physical movement. Neurons in diseased tissue generally contain Lewy bodies (dense aggregates involving proteins a-synuclein and Parkin). Parkin, a protein responsible for targeting damaged proteins for destruction in the proteasome, is believed to be involved in the onset of the disease, leading to protein accumulation, ER stress and production of ROS, resulting in an increased misfolding of a-synuclein. There is sufficiently broad consensus that protein fibrillation is at the root of the disease that it is a major target for drug therapy. In general Parkinson's disease is believed not to have been common until the beginning of the Industrial Revolution. There has thus long been a suspicion that toxic substances might lead to generation of ROS and possibly represent some sort of co-factor for the disease.[20, 21] In summary then, besides the physical presence of any foreign body in the neurons, the additional elements of a-synuclein fibrillation and ROS are quite clearly issues of central concern.
The key conclusion was that oligomeric protein fibril clusters and oxidative stress are both believed to be the cause or (in the cases of ROS) significant contributing factors in neurodegenerative diseases.[22, 23] Thus, according to our understanding, all of the key warnings signals for neurodegenerative hazard were present at the time of formulating the project: nanoparticles can reach the brain, and when they get there, evidence (in vitro and in some cases in vivo) suggests that they are able to trigger the two key processes (oxidative stress and accelerated fibrillation) associated with neurodegenerative diseases.
To this end, the overall objective of the NEURONANO project was divided into eight scientific objectives, as follows:
1. Develop nanoparticles that have appropriate labels of sufficient activity and duration for neurotoxicological studies. In particular, NEURONANO aimed to develop nanoparticles that retain their label activity (radio, fluorescent etc.) over time periods of up to two or more years.(WP1)
2. Identify classes of nanoparticles that lead to either (a) oxidative stress in relevant cell models (WP2); (b) increased rates of fibrillation (WP3); (c) up-regulation of the relevant pathway proteins involved in the diseases (at the cellular level) (WP4). These explicit measurables of oxidative stress and amyloid load (in solution, in cells), validated by animal models (WP4) will provide the basis for identification of hazard for those particles (WP5).
3. Identify the physiochemical properties, and surface expression properties that lead to given levels of oxidative stress and amyloid load. NEURONANO sought to correlate the physiochemical properties with (a) increased rates of fibrillation; (b) up-regulation of the relevant pathway proteins involved in the diseases (at the cellular level); or (c) oxidative stress in relevant cell models. NEURONANO also sought to test the hypothesis that the composition, organization, and reorganization time scales of the biomolecule corona (the proteins that adsorb to nanoparticles under physiological conditions) can be more naturally correlated to these key biological impacts, rather than any specific set of material properties. (WPs 1- 4)
4. Understand what constitutes a lead nanoparticle candidate for passing the Blood Brain Barrier (BBB). The original supposition was that not all nanoparticles that could be toxic will in fact pass through the BBB. Indeed, preliminary data suggest that there is a very great dependence on particle type and size, and these effects must be investigated and understood. Part of our strategy here was to carry out early screening of the particles using the in vitro BBB model in parallel with limited animal trials. NEURONANO also sought to clarify if the nanoparticle corona is more naturally correlated with the ease of crossing the BBB that the underlying nanoparticle physico-chemical properties. (WP4)
5. Quantify transport efficiency to the brain. NEURONANO sought to quantify the amount of nanoparticles that arrive in the brain (of small animals) from each of the potential exposure routes (inhalation, instillation, ingestion, intravenous). Furthermore, NEURONANO sought to show how this depends on the material, again both in the conventional sense (size, zeta potential etc.), and on the surface expression, or evolving protein corona. (WPs 1, 4)
6. Understand the detailed pathways that nanoparticles take to reach the brain. The animal studies were also intended to provide as much information as possible about how the nanoparticles reached their destination, including the detailed nature of what is expressed on their surface as they pass through different intermediate tissue types on going to the brain. Thus, even if no neurotoxicity was observed, the investment of resources, and animal experimentation would result in significant new data and a deeper understanding of the in vivo biodistribution of nanoparticles and the role of the protein corona. (WPs 1, 4)
7. Determine if there are indeed neurodegenerative disease endpoints for a range of animal models, and relate the extent of those endpoints to the amounts and properties of nanoparticles that have reached the brain. If, according to the usual measures of diagnosis (cognitive, behavioral and pathological), nanoparticles were found to affect the time of onset, or extent of the disease, then several examples of arrival dose versus response (by these conventional methods of diagnosis) would subsequently be constructed. The NEURONANO project thus sought to test the hypothesis that the more natural correlation is between the elevation of oxidative stress/amyloid mass and the disease outcome, i.e. to test the concept of the FIBROS property hazard map. (WPs 1, 2, 3, 4)
8. Provide input for a risk assessment framework (and screening protocol) for engineered nanoparticles for use by regulatory authorities and industry. The ultimate aim of the NEURONANO project was to begin development of a simple rational screening methodology for NEURONANOtoxic hazard, the so-called FIBROS map. The basis of the FIBROS map was the hypothesis to be tested within NEURONANO that screening for hazard in the future could be based on the three (presumed) contributing factors of engineered nanoparticles to neurotoxicity: access to the brain, oxidative stress generation potential, and amyloid fibril generation potential, possibly encapsulated in the arrival efficiency of the nanoparticles to the brain, and the location on the FIBROS hazard map. (WPs 1, 2, 3, 4, 5)
The NEURONANO project was thus a very focussed and detailed study of the potential for engineered nanoparticles to pose a neurotoxicity risk, and the development of a framework within which to predict the level of nerutoxic risk posed by specific nanoparticles.
Project Results:
The NEURONANO project was organised into 5 scientific and 2 management Work Packages. The inter-dependencies of the Work Packages are also illustrated. The flow of work in the 3 central experimental Work-Packages was based on a tiered approach, where experiments were conducted in order of increasing system complexity - experiments in solution, experiments in vitro, and finally, experiments in vivo. The purpose of this approach was to establish (where possible) in vitro methodologies to assess the potential neurotoxicity of engineered nanoparticles, and to reduce the numbers of animals needed in the course of the project, in line with the European Commission 3Rs policy regarding animal studies (reduction, replacement and refinement - Directive 86/609/EEC).
Progress towards the objectives and beyond the State of the Art
NEURONANO was an extremely ambitious project that set itself objectives far beyond the state of the art in the field. NEURONANO set itself eight broad objectives against which its success could be measured, spanning the activities of the five research work packages, and the dissemination workpackage. A summary of the NEURONANO consortium's progress towards achieving each of its objectives, during the three years of research activity, is given in the paragraphs below:
Objective 1: Develop nanoparticles that have appropriate labels of sufficient activity and duration for neurotoxicological studies
Radiolabelled nanoparticles: The main task of the JRC in the project was to develop cyclotron-based methods for radiolabelling of nanoparticles. Initially, extensive effort went into investigating the radiolabelling of TiO2 nanoparticles (selected as a priority nanoparticle at the project kick-off meeting for in vivo biokinetics studies at HELMUC). The result of theoretical considerations and many experimental tests was a successful protocol for ion-beam activation and subsequent dispersion and size selection of ST-01 TiO2 NPs, achieving samples suitable for the in vivo studies, whilst also ensuring that the structure of the nanoparticles was undamaged by the selected activation procedure. Several test irradiations on different types of TiO2 were carried out, and XRD structural damage assessment was performed on P25 and ST-01 TiO2 nanoparticles after proton irradiation. The XRD results indicated that no significant damage was caused, even at activity levels as high as 1MBq/mg, achieved after 20 hours irradiation. Much time was spent in developing a dispersion and size-selection protocol in order to achieve a 'bioavailable' suspension of activated ST-01. This included several steps of dispersion, leaching, centrifugation, 'washing' and filtration. The results are described elsewhere,[24] but it can be noted that the suspension consisted of nanoparticles of approximately 100nm hydrodynamic size (DLS performed at HELMUC), that were stably labelled to an activity level of more than 1MBq/mg. The in vivo studies indicated only minimal release of 48V from the activated ST-01 TiO2 nanoparticles.
Throughout the project further experiments have been carried out on TiO2, as it proved to be an extremely useful material (see below) for validation of calculations of the thermal effects of ion-beam irradiation on nanoparticle samples in general. Following this, experiments were also carried out on Fe3O4, CeO2, nanodiamonds, carbon black, and carbon nanotubes. Gold nanoparticles were also successfully activated and were used for in vitro work at IHCP/JRC.
In addition to direct ion-beam activation, a highly novel method for nanoparticle radiolabelling, based on recoil of 7Be from a Li containing powder, was developed. This method was shown to also be suitable for radiolabelling of SiO2 and Al2O3 nanoparticles and nanotubes.
In addition, the JRC activated Au and Ag electrodes for use in the HELMUC spark ignition aerosol inhalation facility, an additional task not originally foreseen in the NEURONANO workplan that is highly relevant for the project results. Finally, radiochemical synthesis of 56Co labelled SiO2 was achieved, and progress was made regarding synthesis of 89Zr labelled Al2O3.
Another approach of synthesis of labelled nanomaterials from activated precursors was also investigated, and radiochemical synthesis of SiO2 labelled with 56Co was achieved. A foil of iron was irradiated to produce the radiotracer, and then dissolved and the 56Co separated. This was added to nanoparticle synthesis precursor solutions and SiO2 nanoparticles were then created. Practically all the 56Co was incorporated in a stable way in the SiO2 nanoparticle matrix. Preliminary experiments showed that 89Zr Al2O3 could be synthesised in a similar fashion and radiochemical synthesis of several other NP types is now being pursued.
Fluorescently labelled nanoparticles: UCD were responsible for development of fluorescently-labelled particles that fulfilled the dual goals of being suitable for use in a biological context, such that they did not elute the fluorescent label inappropriately, and being sufficiently bright and stable that they can be tracked over time in complex milieu. Indeed, within the previous NanoInteract project, UCD had observed that many laboratory synthesized and commercially available fluorescent nanoparticles eluted free dye which confounded the uptake experiments, and had developed approaches to assess the elution of free dye from nanoparticles.[25]
Within the NEURONANO project, UCD focussed on development of optimised fluorescent silica nanoparticles that were resistant to dissolution in vitro over 48 hours. During the course of the NEURONANO project, the UCD team observed that the commonly employed Stöber type silica nanoparticles degrade by hydrolytic dissolution, which is accelerated in the presence of fluorescent moieties, which increase the porous nature of the particles, and under biological media conditions as compared to water alone. UCD have thus developed a method to render such nanoparticles' biologically stable for at least 48hrs, using an arginine-based approach to construct nanoparticles with a core-shell architecture, in which the highly fluorescent silica core is made by the Stöber method and a silica shell is added using the arginine approach in a two-step process. In making such particles, no effort need be made to limit the fluorescence intensity of the core, as this is no longer a significant element of the structural integrity of the nanoparticle.[26]
These particles showed good dispersibility in relevant media (as is usual for silica nanoparticles). They were incubated under biological type conditions with resultant dye leakage/dissolution monitored by time-resolved 1-D SDS-Page Gel and molybdate assay. The modified particles show vastly improved stability against dye leakage. The absence of any visible dye release over 48 hours is quite striking. When combined with the freedom to load a large amount of dye (or other internal) label these structures represent an ideal combination for biological application. The epifluorescence microscopy imaging after 22 hours incubation, in agreement with the gel result, show the increased background fluorescence for the Stober and Stober Shell particles as compared the Arg-Sil Shell ones (rightmost panel). Indeed, under different biological conditions, the contrast can be more striking.
Note: These particles are being further developed within QNano (the Research Infrastructure for nanosafety assessment) in WP6 on the optimisation of labelled nanoparticles for biological assays, and are being considered as for onward development as a QNano quality assured test nanomaterial.
Objective 2a: Identify classes of nanoparticles that lead to oxidative stress in relevant cell models
The original hypothesis was that the potential of the nanoparticles to cause oxidative stress (the Free Radical Generation Potential, FRGP) would differ for different particles depending on their chemical composition and potentially surface curvature and other properties, and that this could be connected to the physic-chemical properties of the nanoparticles as a hazard classification approach.
Thus, UEdin determined the intrinsic free radical activity of the NEURONANO panel of using the using DMPO spin trap approach to measure hydroxyl radicals. This well established approach uses Electron paramagnetic resonance (EPR) or electron spin resonance (ESR) spectroscopy, which is a technique for studying chemical species that have one or more unpaired electrons, such as organic and inorganic free radicals or inorganic complexes possessing a transition metal ion. Spin traps, such as Tempone H (superoxide anions) or DMPO (hydroxyl radicals) serve to 'trap' the fugitive free radicals so that their concentration can be assessed.
From the NEURONANO panel of nanoparticles, only one particle, TiV+ was cytotoxic and haemolytic. However, this particle (TiV+) did not have an excess of free radical generation as measured by EPR compared to the non-toxic NPs in the panel. Thus, the source of its toxicity is not related to ROS production. Similarly, amine-modified polystyrene were found to induce platelet aggregation, but not via the classical up-regulation of adhesion receptors. Rather the amine-PS nanoparticles appeared to act by perturbation of the platelet membrane, revealing anionic phospholipids. Neither oxidative stress generation by particles nor metal contamination was responsible for these effects, which were a result of differential surface derivatization.[27] The study reveals that NP composed of insoluble low-toxicity material are significantly altered in their potency in causing platelet aggregation by altering the surface chemistry.
In a previous study with a large panel of metal oxide nanoparticles, UEdin found a clear relationship between high positive zeta potential and haemolytic activity and between high positive zeta potential and ability to cause inflammation in rat lungs.[28] Thus, the zeta potential of the full NEURONANO panel of particles was determined as a possible factor in toxicity in the light of research in other projects. The TiV+ were found, along with the aminated polystyrene beads (one of the positive control nanoparticles), to have a high positive zeta potential under mild acid conditions.
Objective 2b: Identify classes of nanoparticles that lead to increased rates of fibrillation
At its conception, NEURONANO envisaged that, in analogy to the Reactive Oxygen Generation Potential (FRGP) concept described under objective 2a above, a set of solution studies to determine the ease with which fibrillation occurs for the relevant proteins (i.e. amyloid-ß and a-synuclein for Alzheimer's and Parkinson's diseases, respectively) in the presence of different nanoparticles, producing a 'fibrillation score', termed Amyloid Fibril Generation Potential (AFGP).
Objective 2c: Identify classes of nanoparticles that lead to up-regulation of the relevant pathway proteins involved in neurodegenerative diseases (at the cellular level)
The control of cell number plays a critical role in organ development and in maintaining the integrity of structural elements. Balance between cell proliferation and cell death represents the central aspect of this regulation, as well as in maintaining homeostasis.[31] Failure to regulate apoptosis is a common feature in several diseases, including autoimmune disorders, neurodegenerative diseases, cancer and AIDS.[31] In contrast to rapid turnover of cells in proliferative tissues, neurons commonly survive for the entire life of the organism. This enduring nature of neurons is necessary for maintaining the function of those cells within neuronal circuits. Excessive death of one or more populations of neurons results in disease or injury. For example, death of hippocampal and cortical neurons results in Alzheimer's disease (AD), death of mid brain neurons results in Parkinson's disease (PD), death of neurons in the stratium results in Huntington's disease (HD) and finally, death of lower motor neurons results in amyotrophic lateral sclerosis (ALS).[31]
Objective 3: Identify the physiochemical properties, and surface expression properties that lead to given levels of oxidative stress and amyloid load.
NEURONANO sought to test the hypothesis that the composition, organization, and reorganization time scales of the biomolecule corona (the proteins that adsorb to nanoparticles under physiological conditions) can be more naturally correlated to the key biological impacts associated with neurodegenerative diseases (i.e. oxidative stress and protein fibrillation), rather than any specific set of material properties. Thus, for the test nanomaterials, the nature and impact of the protein corona was assessed as the basis for such a correlation.
Nanoparticle protein corona studies
HMGU performed an in vitro incubation of triphenylphosphine surface-modified gold nanoparticle (5nm, 15nm and 80nm, identical to those used in the biodistribution studies) with diluted serum of mice. The nanoparticles were incubated for 24h at 37°C. After incubation the nanoparticles were washed three times and an SDS gel electrophoresis was made. After Coomassie blue staining the gel spots were cut and prepared for subsequent digestion with trypsin. The digestion was performed according to an established standard protocol of the core facility Proteomics of the Helmholtz Center Munich. After digestion the Mass Spec analysis was made with the help of the Proteomics Analyzer 4700, which is a tandem time of flight matrix assisted laser desorption and ionization apparatus (MALDI-TOF-MS). From this analysis, a number of different proteins which conjugated significantly to the gold nanoparticles.
Objective 4: Understand what constitutes a lead nanoparticle candidate for passing the Blood Brain Barrier (BBB)
Again, this initial conception while valuable in framing the project, was somewhat limited by our collective lack of knowledge at the time, as it missed two key factors that have since emerged as being of very considerable importance, namely:
(1)that nanoparticles can (potentially) signal across barriers via paracrine signalling, as has been demonstrated for the wear-induced particles across the foetal barrier,[40] and with the NEURONANO project for non-toxic carboxyl-modified polystyrene nanoparticles.[41]
(2)The low amount of particles reaching the brain could be a result of significant accumulation of nanoparticles in the brain endothelium itself. Indeed, using the in vitro barrier model, we observed significant accumulation of nanoparticles in lysosomes of the barrier cells, from as early as 4 hours after exposure, and reaching a plateau level at account 24 hours, suggesting that the barrier itself could be a target for nanoparticle accumulation.
Objective 5: Quantify transport efficiency to the brain. NEURONANO sought to quantify the amount of nanoparticles that arrive in the brain (of small animals) from each of the potential exposure routes (inhalation, instillation, ingestion, intravenous)
In vivo Exposure by Peripheral Injection to nanoparticles in rats
UU investigated the amount of nanoparticles reaching the brain of rats using a combination of TEM imaging and ICP-MS quantification of particle load.[45, 46] Peripheral injections of ceria (64 nm), titania (3-9 nm), silica (50 nm), gold (BBI 30 nm), gold-ApoE-PEG 2000 (30nm), gold-ApoE-PEG 5000 (30nm) and gold-ApoE-PEG 10000 (30nm) were carried out using intravenous and intra-tracheal methods. Particle suspensions were intratracheally instilled synchronous with the animal's inspiration followed by 300 μl of air. Animals were culled after 24 hours and blood was collected immediately by cardiac puncture, aliquoted into EDTA Vacutainer tubes and chilled. An aliquot of CSF was removed from the cisterna magna and frozen immediately. Animals were then perfused through the heart with PBS solution and all major organs (brain, heart, lungs, liver, kidneys, spleen, pancreas, stomach, intestines and muscle) removed for ICP-MS analysis. Ultrathin sections (100nm) were microtomed and mounted on copper TEM grids. Samples were imaged with a FEI Titan transmission electron microscope equipped with an energy dispersive X-ray (EDX) elemental analysis system.
Objective 6: Understand the detailed pathways that nanoparticles take to reach the brain
The animal studies described under Objective 5 above were also intended to provide as much information as possible about how the nanoparticles reached their destination, including the detailed nature of what is expressed on their surface as they pass through different intermediate tissue types on going to the brain. Thus, even if no neurotoxicity was observed, the investment of resources, and animal experimentation would result in significant new data and a deeper understanding of the in vivo biodistribution of nanoparticles and the role of the protein corona.
Mechanisms of uptake and intracellular transport of nanoparticles
To date, most studies have focused on identifying the point of entry and final destination of nanoparticles. However, to truly understand and predict the fate and impacts of nanoparticles, and specifically to identify rare events, a deeper knowledge and understanding of the entire intracellular itinerary and event sequence will be essential. It is, in principle, possible that the original corona, combined with new molecules picked up inside the cell, could, even if rarely, provide the nanoparticles with signals to allow them to access different subcellular destinations rather always following the endo-lysosomal pathway.
Role of cell cycle in nanoparticle uptake (including in polarised / barrier cells)
UCD have recently demonstrated the strong relation between nanoparticle uptake and cell cycle in proliferating cells and developed methods to measure if nanoparticle export is present[54]. In proliferating cell populations, during cell division, nanoparticles are split among daughter cells, thus the internalized nanoparticle dose is diluted. These findings can have important implications for slow- or non-dividing cells, such as for instance the blood barrier itself, where such diluting effect may be missing or less active, and if export is not present, bioaccumulation could arise over longer time scales. Thus, together with the question of transcytosis, it is important to study how nanoparticle uptake and cell cycle relates in such specialized barriers and if nanoparticles can impact the barrier itself.
In vitro transport studies in the model BBB - evidence of nanoparticle transcytosis
In recent years, significant effort has been directed towards mimicking the in vivo BBB using immortalised cells. The hCMEC/D3 cell line was chosen as it is the closest available correlate to the human in vivo situation, and does not require co-culture with primary astrocyte cells, making it easier to work with for quantitative studies. In addition, the porous transwell system upon which the hCMEC/D3 monolayer was grown has previously been utilised in the presence of monolayers from alternate cell lines by other groups (19, 20).[55] As this cell line can differentiate into a monolayer in culture after 7-10 days, seeding the cells onto a permeable filter, the so called 'transwell' system enables an in vitro BBB model to be established. The transwell is composed of an insert filter on which the cell monolayer is grown, and an acceptor well, with the apical chamber mimicking the blood and the basolateral chamber mimicking the brain. Thus, molecules loaded into the apical chamber are presented to the cellular barrier formed on the filter, which can then to be taken up by the cells by a receptor-mediated (transcytosis) process.
Role of proteins in corona in vivo on nanoparticle distribution
The pulmonary route is considered to be one of the primary exposure routes for inadvertent exposure to nanomaterials, as well as being very promising for drug delivery by inhalation. In this regard, understanding the role of lunch surfactant proteins on the biodistribution of nanoparticles is vital. Directly after their deposition, inhaled Au NP come into contact with pulmonary surfactant protein D (SP-D). SP-D can agglomerate Au NP in vitro, and this may influence the clearance as well as the systemic translocation in vivo.
Mytilus edulis (mussel)
The mussel genus Mytilus is one of the most widely-used groups of sentinel species in aquatic toxicology, having been used for many years in environmental monitoring programs such as Musselwatch.[58] Advantages of this species include the mussels' wide geographic range, abundance, ease of identification, filter-feeding habit and sedentary nature. The animal is also robust enough to withstand moderate levels of pollution. This allows a broad range of loadings, with a resultant range oxidative stresses. It is also believed that the overall range of oxidative stress processes in Mytilus is comparable to that of animals, and therefore its whole-organism proteome is an attractive analytical tool to screen for tissue level oxidative stress. Indeed, in recent years, proteins expressed in Mytilus have come to be widely used as biomarkers of environmental stress, which is now considered a form of validation of the species for environmental (including oxidative) stress.[59]
Nereis virens (ragworm)
The marine polychaete, Nereis virens, a common species in estuarine sediment and infaunal deposit feeder, was used as a further animal model. This species is known to select and ingest large amounts of fine sediment particles that tend to be enriched in organic material and metals. The Worms used in this exposure trial were bought at Halfway angling centre, Cork. Clean sediment was bought in from a local supplier (artificial play sand).
Neuropathological and histological studies of rat brains
The effect of titania nanoparticles (ST01 titania anatase) on plaque formation was analysed in the 5XFAD Alzheimer's mouse model. These transgenic mice express 5 forms of human familial Alzheimer's Disease mutations (Co-overexpression of the human APP695 with the Swedish (K670N, M671L), Florida (I716V), and London (V717I) FAD mutations in the same APP molecule and human PS1 harbouring two FAD mutations, (M146L and L286V). These mice develop amyloid plaques, show impairments in neuronal transmission and synaptic plasticity as they age, and develop memory impairments similar to patients of Alzheimer's disease. Amyloid deposition and gliosis is accelerated in these mice compared to other models, with significant levels observed at 2 months of age.
Behavioral studies in rats and mice
Behavioral and memory changes are the first subtle changes of brain functionality that are observed well before any anatomical changes occur in neurodegenerative diseases. Thus, in addition to looking for physiological markers of neurodegeneration in response to exposure to nanoparticles, NEURONANO partners also performed some behavioral studies on nanoparticle-exposed and control animals.
The conclusion from these studies is that very little behavioural effects were observed in the rats or mice. Further data analysis is still underway and will be published over the coming months. These findings are also consistent with very small amounts of the ST-01 titania particles reaching the animal brains following IV or IT exposure.
Objective 8: Provide input for a risk assessment framework (and screening protocol) for engineered nanoparticles for use by regulatory authorities and industry
The ultimate aim of the NEURONANO project was to begin development of a simple rational screening methodology for NEURONANOtoxic hazard, the so-called FIBROS map. The basis of the FIBROS map was the hypothesis to be tested within NEURONANO that screening for hazard in the future could be based on the three (presumed) contributing factors of engineered nanoparticles to neurotoxicity: access to the brain, oxidative stress generation potential, and amyloid fibril generation potential, possibly encapsulated in the arrival efficiency of the nanoparticles to the brain, and the location on the FIBROS hazard map.
We can summarise the findings from the NEURONANO project as they relate to the concepts of the FIBROS map as follows:
Fibrillation potential
-Induction of fibrillation is a surface area effect - all nanoparticles induce or modulate the rate of protein fibrillation, and there appears to be a detailed interplay between monomer, oligomer and fibril for residence at the nanoparticle surface.
-The presence of a protein corona on nanoparticles, such as would inevitable occur during nanoparticle uptake in vivo, slowed the fibrillation of Aβ, regardless of size, surface chemistry, charge or aspect ratio when compared with the same pristine (bare) nanomaterial.
-In order to be effective as an inhibitor of Aß fibril formation the proteins in the nanoparticle-protein corona must be in their native form: corona proteins that have been denatured by heat were found to actually accelerate the formation of protein fibrils, mostly likely due to the exposure of hydrophobic amino acid residues from proteins that do not refold after heating.
-Further work is required to understand more fully the specific mechanism by which fibril formation is inhibited by the presence of a protein corona around nanomaterials, and also to determine whether slow protein degradation in vivo in the presence of nanoparticles could contribute to protein fibrillation and neurodegeneration over the long term.
-The fact that particles bring their corona with them all the way to their final location in cells suggests that there may also be subtle effects from the presence of serum proteins in cells when they are normally excluded from cells. Whether this would also have a role in protein fibrillation in cells, or indeed induce other impacts requires further investigation.
-The fact that the positively charged polystyrene particles, which induced apoptosis in astrocytes, were also found to up-regulate signatures related to protein aggregation and fibrillation is significant, and suggests that this could be an alternative measure against which to assess fibrillation potential of nanoparticles. A wider panel of particles would need to be assessed, along with an in-depth biomarker validation process in order to demonstrate this link robustly.
Reactive oxygen stress generation potential
-No link between cellular oxidative stress and nanoparticle effects could be confirmed. Indeed, the particle that was found to induce the most effects in cellular assays, the TiV+ which was cytotoxic, haemolytic and induced mitochondrial damage, did not have an excess of free radical generation as measured by EPR compared to the non-toxic nanoparticles in the panel.
-A link between particles with a positive surface charge and toxicity was also investigated, and while the connection held in some assays, the TiV+ particles were found to damage mitochondrial function, whereas the positively charged control nanoparticle, amine-modified polystyrene nanoparticles were not, suggesting that more than simple charge was involved in the induction of this particular effect.
-Oxidation of protein thiols and carbonyls is a generally-applicable, sensitive marker of oxidative stress induced by nanoparticles in both cells and more complex organisms such as mussels and worms (much more robust than measuring cellular oxidative stress in the presence of nanoparticles). However, as with the biomarkers for protein fibrillation, significant additional screening, plus validation work is required.
Access to the brain
-Route of exposure matters in terms of how much of the applied particle dose reaches the brain, although even with the most efficient transfer to the brain, less than 1% of the applied dose reaches the brain.
-Using ST-01 titania nanoparticles, brain accumulation was found to peak at around day 7, but particles could still be detected by day 28 following a two-hour inhalation (mean and SD, n=4) in rats. On the other hand, following peripheral injection of the same nanoparticles in rats, titania particles were found only in liver and lung and no particles were found in brain tissues.
-A significant increase in cerium content against control (6-fold; p<0.05) was observed in brain tissue of rats injected intravenously with ceria nanoparticles. This data may suggest that ceria passes the blood brain barrier readily.
-Nanoparticle size plays a role in determining the amount of access to the brain, although to a smaller degree that expected for exposure via intertrachael instillation and intravenous administration: brain accumulation of 5 nm - 200 nm gold NPs is similar for all particle sizes tested. Only the 1.8 nm particles showed significantly enhance brain uptake via these routes.
-The exposure route dependence is related to the nature of the biofluid milieu that the particles initially encounter and thus to the nature and composition of the protein corona that forms: exposure via the oesophagus led to the highest accumulation in the brain for the 18 nm gold NPs, indicating that something happens with these NP in the GIT which alters their biokinetic fate. The most obvious explanation is a different specific protein coating, related to the exposure route. In the stomach, it is conceivable that the acidic environment, as well as the presence of gastric enzymes, are likely to degrade the triphenylphosphine surface coating of the applied NP. Thereby, protein coating of the blank NP surface may occur.
-Although uptake is rather low compared to the initially injected dose, conjugation of gold nanoparticles with either ApoE or albumin significantly increased the amount of particles detected in all investigated brain compartments compared to the bare citrate stabilized gold NP. Retention after 24h was very similar in the total brain for both apoE conjugates.
-The distribution pattern for single brain compartments was very similar for both apoE conjugates. Therefore we conclude that both conjugation methods led to an increased translocation into the brain across the blood-brain barrier. Compared to the apoE conjugates the albumin conjugates seem to have an even better access to the brain.
Potential for indirect signaling
-The low amount of particles reaching the brain could be a result of significant accumulation of nanoparticles in the brain endothelium itself. Indeed, using the in vitro barrier model, we observed significant accumulation of nanoparticles in lysosomes of the barrier cells, from as early as 4 hours after exposure, and reaching a plateau level at account 24 hours.
-While such significant accumulation of non-toxic nanoparticles, such as the Ps-COOH shown here, and also silica nanoparticles or gold nanoparticles which we have also studied using this model, do not seem to perturb the cells, there are indications that signalling molecules are secreted by the endothelium in response to the presence of the PS COOH NPs, which are subsequently amplified by the astrocytes when using a co-culture model.
-Growth of human astrocytes below the BBB monolayer allows the communication of supporting signals in vitro which aids in mimicking a more realistic in vivo situation, allowing these signalling pathways to occur. Assessing impacts of less benign particles accumulating in endothelium is necessary, although as the particles themselves begin to induce effects on the endothelium barrier separating out the sequences of signalling events will become more difficult, and isolating paracrine signalling from inter-cellular signalling will require detailed molecular biology studies.
Need to connect access to the brain with impact
-The neuropathological review of animals exposed intra-tracheally to ST01 titania anatase (3-9nm primary particles, 55 nm in dispersion) indicated that these particles did not cause any detectable neuropathology effects: No significant differences in the volume density of plaque, the total plaque volume or the total activated astrocyte number were demonstrated between control and treatment groups following 6 months exposure to titania (ST-01) nanoparticles administered intra-tracheally.
-Thus, from a risk assessment viewpoint, simply observing that a small percentage of particles can reach the brain may not actually represent a risk, if there is no associated hazard. However, given that these studies were performed on limited numbers of animals, a very limited panel of particles, significantly more research is required in order to make a real connection between access and impact. Additionally, the particle tested were themselves very benign, and thus more inherently toxic particles should be tested in the future. However, from the viewpoint of the FIBROS model for risk assessment, it is clear that access and impact need to both be assessed, and dose-response elements included also.
All NEURONANO publications are available at http://www.ucd.ie/cbni
Potential Impact:
Dissemination activities and the exploitation of results
The NEURONANO project has already resulted in over 50 publications , with several more in press or under review. The knowledge and experience gained is now forming the basis for other projects, both in the EU and across the world, including incorporation into several other EU projects, such as QualityNano, the EU infrastructure for nanosafety assessment which is disseminating many of the NEURONANO findings to the community via training schools and the development of positive and negative control nanomaterials; and NanoTransKinetics, a modelling project that aims to develop a predictive tool for determining uptake, localisation and impact of nanomaterials based on their physic-chemical characteristics and their subsequent biomolecular corona.
From the project conception stage, emphasis was placed on ensuring impact for, and contribution to standardisation efforts, including:
-ensuring that the NEURONANO partner's efforts to establish best practices within the group should lead to protocols of value to the international community of scientists working in the nanotoxicology arena. The project was designed so that members of the Advisory Board who liaise with bodies responsible for standards such as CEN WG 166, ISO newly created TC on nanotechnology will be aware of the best practices that we are developing.
-building on the technical and focused nature of the project, which was heavily biased towards quantitative and verifiable methods, to ensure that the outputs are framed in such a way that they contribute key scientific data to the development of standards, test and measurement methods in the field of nanoparticle interactions with biological systems.
Enduring impact in a range of scientific and quality assurance areas
The NEURONANO project has already had, and will continue to have, impact far beyond its size and budget, in part due to the formative stage of the research field at the outset of the project, and in part as a result of the fact that it has facilitated the development of several insights that will have a durable impact on the research field. Among the key scientific developments resulting from the NEURONANO project were the fact that the cell cycle may play a significant role in determining the accumulation or otherwise of nanoparticles in key cells, the fact that the nature of the protein corona can be significantly different under in vitro versus in vivo conditions due to the significant differences in the amounts of proteins present relative to the available particle surface area.
Radiolabelled nanoparticles synthesised by JRC
A key output from NEURONANO is the availability of a range of radio-labeled nanoparticles for use in dosimetric / biokinetics or environmental fate and behaviour studies, or other studies where tracking of nanoparticles over specific time periods is required.
Pre-synthesised Au, CeO2, Fe3O4, and TiO2 nanoparticles can be radiolabeled using ion-beam methods, and Au and Ag electrodes can be activated for subsequent nanoparticle synthesis.
Positive and negative control Nanoparticles for apoptosis and cell cycle regulation
Having studied in detail in several cell lines and several species the impacts of amine-modified polystyrene nanoparticles, and found a consistent and dose-dependent onset of apoptosis (linked with autophagy under some circumstances), as well as a low-dose impact on cell cycle, these nanoparticles were determined to be an ideal candidate for onward development, and round-robin validation as positive control nanomaterials for apoptosis and cell cycle regulation. The fact that they are easily labelled for tracking and surface functionalised with a variety of functional groups adds to their attractiveness as a positive control candidate. Since the carboxyl-modified equivalent particles show almost no effects on tested cells or organisms at equivalent exposure doses and times they are an ideal candidate for onward development as negative control nanoparticles for the same end-points.
Protocols covering a range of aspects of nanoparticle interactions with cells, barriers and organisms
Protocols covering a variety of aspects, from nanomaterials synthesis and characterisation, surface functionalisation and dispersion, to all aspects of nanoparticle uptake and localisation, apoptosis, assessment of changes in gene expression, oxidative stress and redox proteomics, dosimetrics and biokinetics and assessment of behavioral impacts of nanomaterials, have been developed within NEURONANO. Many have already been published as part of scientific manuscripts, and many others are currently in final stages of preparation for publication.
New assays and/or experimental approaches developed within NEURONANO
Effect of ratio of available protein to nanoparticle surface area on resultant corona (in vitro versus in vivo coronas)
A key advance developed within NEURONANO has been the understanding that the corona is context dependent: i.e. the corona formed under low protein conditions (i.e. in vitro conditions) may not be identical to that formed under high protein conditions (i.e. in vivo conditions).[65] The protein adsorption for two compositionally different NPs, namely sulphonated-polystyrene (PSOSO3) and silica (SiO2) NPs was studied at different protein concentrations keeping the particle surface area constant. NP-protein complexes were characterized by Differential Centrifugal Sedimentation (DCS), Dynamic Light Scattering (DLS) and Z-potential both in situ and once isolated from plasma as a function of the protein/NP ratio. A semi-quantitative determination of their hard corona using one-dimensional sodium dodecyl sulphate poly-acrylamide gel electrophoresis (1D PAGE) and electrospray Liquid Chromatography Mass Spectrometry (LC MS/MS) was introduced, which allowed us to follow the total binding isotherms for the particles, identifying simultaneously the nature and amount of the most relevant proteins as a function of the plasma concentration. We find that the hard corona can evolve quite significantly as one passes from protein concentrations appropriate to in vitro cell studies to those present in in vivo studies, which has deep implications for in vitro-in vivo extrapolations and will require some consideration in future.
Cell cycle effects
Work at UCD regarding the role of cell cycle in determining nanoparticle uptake rates and intracellular dose[66] have resulted in some important protocols and experimental design considerations. In summary, the nanoparticles studied here do not disrupt cell cycle progression and, once internalized, localize in lysosomes, where they remain. Nanoparticle export is negligible and dilution of the intracellular nanoparticle load occurs by cell division. Uptake rates during the different cell cycle phases are comparable.
TEM protocol for assessing integrity of the in vitro BBB
In light of the increasing drive to reduce reliance on animal testing there is a concerted effort to develop and validate alternative in vitro methods for research, screening and regulatory assessment of the impacts of chemicals, including nanoparticles. However, to date there remain important concerns about the reliability and relevance of such in vitro methods, and whether they represent a realistic alternative to animal studies. Methods to ensure the quality and reproducibility of in vitro methods and the comparability of the data generated to the in vivo models they are intended to replace are required to begin to address these issues.
The neutral red retention time (NRRT) assay
The NRRT assay is useful for detecting decreased lysosomal membrane stability in haemocytes sampled from bivalves, a phenomenon often associated with exposure to environmental pollutants including nanomaterials. Bivalves are popular sentinel species in ecotoxicology and use of NRRT in study of species in the genus Mytilus is widespread in environmental monitoring. The NEURONANO partner UCC has shown that this assay can be applied to a panel of metal and metal oxide nanoparticles (2ppm) and distinguish between those that are toxic and those that are not, and connect this with markers of oxidative stress or other toxicity pathways, and they suggest that this could support use of NRRT as a generally applicable in vitro assay in nanosafety assessment.
Impact for standardization
Several of the approaches developed or optimised in NEURONANO are suitable for standardisation, and indeed NEURONANO was presented at two of the European Commission workshops on standardisation activities from EU NMP projects, as follows:
Dr. Iseult Lynch presented data specifically related to standardisation from NEURONANO at the NMP Seminar on Standards & Standardisation on December 14th 2010 in Brussels. Talk title: 'NEURONANO (EU FP7 214547-2) Contribution to standardisation processes'. Significant interest was expressed by CEN in the aspects of the protocols for Differential Centrifugal Sedimentation characterisation of nanoparticle dispersions in situ in biological fluids and in the validation of the in vitro model Blood brain barrier, and follow-up is underway.
Why do we need to describe the dose of nanoparticles?
There is a need to know the in situ dose in nanoform (< 100 nm). To achieve this, NEURONANO proposed:
- Characterisation in situ in biofluids
- Combination of DCS and TEM describes dose presented to cells
- Time-resolved data (evolution of sample)
- Applicable to any biofluid, including environmental (e.g. natural organic matter).
Impact for regulatory agencies and policy groups
Neurodegenerative diseases currently affect over 1.6% of the European population, with dramatically rising incidence that is likely (in part) due to the increase in the average age of the population. There are persistent claims, based on the epidemiology that pollution may be a cofactor in Alzheimer's disease, although the evidence is controversial. The risk that engineered nanoparticles could introduce unforeseen hazards to human health is now a matter of deep and growing concern in regulatory bodies, governments and industry. However, at present there is only circumstantial evidence that nanoparticles could impact on such diseases.
Recommendations for further research (from NEURONANO Final meeting and workshop)
As part of the NEURONANO final meeting (5th December 2011), and in collaboration with the ESF EpitopeMap Research Networking Programme , a joint workshop was held in Dublin on 6th - 7th December 2011 that brought together leading European experts in Engineered nanoparticle transport across biological barriers to identify the central questions that require scientific and policy attention as a matter of priority. Among the outcomes from the workshop of particular relevance for NEURONANO and its stakeholders were:
Accidental 'targeting' of ENP to the brain - need for implementation of 'protein corona' screening
Many of the 'success' stories to date in terms of using engineered nanoparticles to deliver drugs to the brain have been the result of the ENP surface offering an appropriate scaffold for known blood-brain-barrier receptors, such as the LDL receptor. In fact, one key study shows that having a specific surfactant coating (polysorbate 80 - Tween 80) that spontaneously binds apolipoprotein E and apolipoprotein A-I from the dispersion medium was as effective at delivering the drug to the brain as the specifically surface-engineered ENP where these apolipoproteins were chemically grafted to the ENP.[70-72] This raises a very important question as to whether other ENP, which were never intended to reach the brain (or indeed other key organs, or the foetus), could unintentionally bind such targeting protein coatings and then reach delicate organs with very high efficacy.
Funding recommendation: Research is needed to improve existing and create new barrier models and develop clear in vitro-in vivo correlations, including appropriate culture conditions (serum free) to optimise barrier functioning and cell-cell (paracrine) signalling. Research should be directed towards the use of co-culture and 3D cell culture approaches, and towards comparative in vitro and in vivo studies under similar conditions, and where limitations are documented and understood.
Infrastructure needs for nanosafety and nanomedicine
A key concern relates to large scale facilities and specialised research centres for supporting the safety and efficacy assessment of ENP (for example facilities for radiolabelling ENP and assessment of the biodistribution and biokinetics of radiolabelled-ENP). Funding to support such facilities, and the mobility of EU researchers / post-docs for their wider exploitation, over the coming decade is vital, as many questions regarding safety and efficacy of ENPs can only be answered by the use of not yet available radioactive nanomaterials. In a parallel fashion, support to maintain the intellectual infrastructure (appropriately trained personnel) is required to respond adequately to EU needs in relation to ENP.
Project website: http://www.neuronano.eu
Coordinator contact details:
Prof. Kenneth Dawson, Director Centre for BioNano Interactions, University College Dublin, Belfield, Dublin 4, Ireland.
Phone: 00353 1 716 2459
Email: Kenneth.A.Dawson@cbni.ucd.ie
The EU FP7-funded project NEURONANO conducted an investigation into how nanoparticles interact with the human brain. The overall aim was to establish a risk assessment framework to categorise nanomaterials in terms of their potential for access to the brain, and potential for harm once there.
The NEURONANO project resulted in a number of key findings:
-The existing body of research into the impacts of nanomaterials was not sufficient for the project partners' purposes. As a result, they had to produce their own evidence base from scratch, using both cellular systems and carefully selected animal studies.
-Despite the brain being a 'sacrosanct' organ which most particles cannot enter, very small quantities of nanoparticles can pass through the blood-brain barrier (BBB). However, there is no evidence that the amounts of engineered nanomaterials that pass the barrier are sufficient to affect human health.
-Several new modes of interactions of nanoparticles with the BBB were observed, using the most advanced microscopy every applied to such systems. For example, a propensity for nanoparticles to accumulate in the lysosomes of the BBB endothelial monolayer was observed. Additional work is required to understand if similar effects are also observed in animals, as current approaches to quantify uptake in vivo do not distinguish between the nanoparticle load in the brain versus in the endothelium separating the brain from the blood (i.e. within the BBB).
References:
[1] Devasagayam, TPA et al., J. Association of Physicians of India (JAPI) 2004, 52: 796.
[2] Walsh DM, Teplow DB. Prog Mol Biol Transl Sci. 2012, 107:101-24.
[3] Calderon-Garcidueñas L, et al., Toxicol Pathol. 2007, 35:154-62.
Project Context and Objectives:
The overall science and technology objective of the NEURONANO project was to determine if engineered nanoparticles could constitute a significant neuro-toxicological risk to humans for two diseases endpoints, namely Alzheimer's and Parkinson's diseases. Importantly, the NEURONANO consortium did not presume a neurotoxic hazard from engineered nanoparticles / nanomaterials. Instead, the project sought to understand if exposure to nanoparticles resulted in them reaching the brain, and if so under what circumstances. Additionally, the project aimed to understand whether the presence of nanoparticles in the brain could exacerbate or accelerate existing cellular mechanisms such as the rate of generation of Reactive Oxygen Species (ROS) or the rate of fibrillation of proteins involved in the neurodegerative diseases of interest, amyloid-beta and alpha-synuclein, respectively.
The project was designed to be compliant with the European Commission's policy on reduction, replacement and refinement (the 3Rs) of the use of animals in scientific testing and experimentation. As such, much of the work was performed using in vitro methods, many developed specifically within the project, and which are now being further refined for onward development towards standardised alternative tests, both within the QualityNano research infrastructure project (see http://www.qnano-ri.eu online), and in other EU projects such as MARINA (see http://www.marina-fp7.eu/ online). Where animal studies were necessary (such as for assessment of the distribution of nanoparticles following exposure via different routes, and to assess the impacts of nanomaterials in the brain on animal behavior and amlyloid plaque load), there were designed to correlate and potentially confirm in vitro findings, and were designed to minimise the numbers of animals needed whilst still being scientifically robust. The tiered approach to the three stands of work relating to the assessment of neurodegenerative disease (solution studies, cell studies, and finally animal studies) represents a scientifically and ethically balanced approach to the work, balancing the necessary level of scientific excellence with the need to reduce the numbers of animal experiments.
At the time of development of the NEURONANO project, much of the knowledge of nanoparticle-induced oxidative stress has been derived (or deduced) from environmental or occupational exposure to small particulates including asbestos, coalmine dust, quartz, and PM10.[12-14] Oxidative stress caused by particles and other sources is capable of graded (dose-response-related), effects in cells.[15, 16] At low level there may be only very mild effects that are dealt with by the cells constitutive anti-oxidants and are not 'sensed'. At higher levels there can be activation of anti-oxidant defences, like catalase, to deal with mild oxidative stress. As the extent of oxidative stress increases, there is activation of stress-responsive signalling that leads to changes in gene expression for inflammation in epithelial cells, for example. At very high doses there can be cell death by necrosis or apoptosis.
Thus, there were several pathways by which nanoparticles might generate (and respond to) harmful oxidative stress in the brain:
1) Direct effects of nanoparticle-derived free radicals on neurones after contact or entry into them;
2) Contact or entry into neurones causing proteasomal or mitochondrial dysfunction;
3) Uptake by microglial cells causing activation of NADPH oxidase and an oxidative burst.
If indeed oxidative stress is an important mechanism in the brain, then it was anticipated that those particles most active in the generation of free radicals production would be more harmful than others.[17] However, given the complexity even in cell systems, it was possible that oxidative stress might be correlated not just with the degree of necrosis/apoptosis, but also with mitochondrial dysfunction, proteasomal dysfunction, induction of antioxidant defence, respiratory burst and inflammatory gene expression by microglial cells. To form a coherent picture the NEURONANO team proposed the application of gene arrays, proteomics and redox proteomics to determine the nano-particle-induced perturbation of cellular metabolism and oxidative stress pathways.
Whilst the precise mechanisms leading to neurodegenerative diseases are still not fully clarified (and were even less so at the time of initiation of the NEURONANO project), it is broadly agreed that the key effects involve the presence of early pre-fibrillar structures, neuroinflammation and ROS-related processes. All of the links in the causal chain are now present for a credible expectation that nanoparticles could have impacts on the onset and progression of neurodegenerative diseases, and thus present a neurotoxicity risk of greater or lesser significance depending on their physico-chemical composition and potential to access the brain.
Alzheimer's disease is characterized by loss of short-term memory and visual-spatial skills. It is diagnosed clinically based on characteristic neurological and neuropsychological phenomena. There is as yet no commonly accepted marker for the disease, and certain diagnosis is made only post-mortem. Recent research has generally considered the disease is caused by misfolding of either amyloid ß or tau proteins.[18] Evidence that amyloid ß is the key is provided by the fact that the Down syndrome patients (who almost invariably progress to Alzheimer's at an early age[19]) have an additional copy of the gene for amyloid ß precursor. There is also growing support for the hypothesis which suggests that the cytotoxic species is an intermediate oligomeric misfolded species of the protein, rather than the mature fibrils. It may be that Alzheimer's is primarily a disease driven by extracellular events where by Amyloid Precursor Protein (APP) is cleaved leading to release of the amyloid ß fragment containing residues 1-40/42, and that the resulting extracellular amyloid ß then forms the toxic oligomers (pre-fibrils) leading to the disease. Indeed the delivery of biological extracts containing these oligomers to mice leads directly to Alzheimer's-like symptoms.[6] Exposure to environmental toxins is believed to increase the likelihood of developing Alzheimer's disease, and there have been persistent claims that the ROS mechanism is also involved, but the manner is ill-understood as yet. Studies in cities with high pollution loads, where there are significant loads of nanoparticles of the size that can access the brain have suggested that Alzheimer's could be associated to this form of pollution, and it is now known that nanoparticles can dramatically affect the rate of protein fibrillation (including amyloid ß).[3]
Parkinson's disease is characterized by muscle rigidity, tremor, a slowing of physical movement, followed by (in extreme cases) a loss of physical movement. Neurons in diseased tissue generally contain Lewy bodies (dense aggregates involving proteins a-synuclein and Parkin). Parkin, a protein responsible for targeting damaged proteins for destruction in the proteasome, is believed to be involved in the onset of the disease, leading to protein accumulation, ER stress and production of ROS, resulting in an increased misfolding of a-synuclein. There is sufficiently broad consensus that protein fibrillation is at the root of the disease that it is a major target for drug therapy. In general Parkinson's disease is believed not to have been common until the beginning of the Industrial Revolution. There has thus long been a suspicion that toxic substances might lead to generation of ROS and possibly represent some sort of co-factor for the disease.[20, 21] In summary then, besides the physical presence of any foreign body in the neurons, the additional elements of a-synuclein fibrillation and ROS are quite clearly issues of central concern.
The key conclusion was that oligomeric protein fibril clusters and oxidative stress are both believed to be the cause or (in the cases of ROS) significant contributing factors in neurodegenerative diseases.[22, 23] Thus, according to our understanding, all of the key warnings signals for neurodegenerative hazard were present at the time of formulating the project: nanoparticles can reach the brain, and when they get there, evidence (in vitro and in some cases in vivo) suggests that they are able to trigger the two key processes (oxidative stress and accelerated fibrillation) associated with neurodegenerative diseases.
To this end, the overall objective of the NEURONANO project was divided into eight scientific objectives, as follows:
1. Develop nanoparticles that have appropriate labels of sufficient activity and duration for neurotoxicological studies. In particular, NEURONANO aimed to develop nanoparticles that retain their label activity (radio, fluorescent etc.) over time periods of up to two or more years.(WP1)
2. Identify classes of nanoparticles that lead to either (a) oxidative stress in relevant cell models (WP2); (b) increased rates of fibrillation (WP3); (c) up-regulation of the relevant pathway proteins involved in the diseases (at the cellular level) (WP4). These explicit measurables of oxidative stress and amyloid load (in solution, in cells), validated by animal models (WP4) will provide the basis for identification of hazard for those particles (WP5).
3. Identify the physiochemical properties, and surface expression properties that lead to given levels of oxidative stress and amyloid load. NEURONANO sought to correlate the physiochemical properties with (a) increased rates of fibrillation; (b) up-regulation of the relevant pathway proteins involved in the diseases (at the cellular level); or (c) oxidative stress in relevant cell models. NEURONANO also sought to test the hypothesis that the composition, organization, and reorganization time scales of the biomolecule corona (the proteins that adsorb to nanoparticles under physiological conditions) can be more naturally correlated to these key biological impacts, rather than any specific set of material properties. (WPs 1- 4)
4. Understand what constitutes a lead nanoparticle candidate for passing the Blood Brain Barrier (BBB). The original supposition was that not all nanoparticles that could be toxic will in fact pass through the BBB. Indeed, preliminary data suggest that there is a very great dependence on particle type and size, and these effects must be investigated and understood. Part of our strategy here was to carry out early screening of the particles using the in vitro BBB model in parallel with limited animal trials. NEURONANO also sought to clarify if the nanoparticle corona is more naturally correlated with the ease of crossing the BBB that the underlying nanoparticle physico-chemical properties. (WP4)
5. Quantify transport efficiency to the brain. NEURONANO sought to quantify the amount of nanoparticles that arrive in the brain (of small animals) from each of the potential exposure routes (inhalation, instillation, ingestion, intravenous). Furthermore, NEURONANO sought to show how this depends on the material, again both in the conventional sense (size, zeta potential etc.), and on the surface expression, or evolving protein corona. (WPs 1, 4)
6. Understand the detailed pathways that nanoparticles take to reach the brain. The animal studies were also intended to provide as much information as possible about how the nanoparticles reached their destination, including the detailed nature of what is expressed on their surface as they pass through different intermediate tissue types on going to the brain. Thus, even if no neurotoxicity was observed, the investment of resources, and animal experimentation would result in significant new data and a deeper understanding of the in vivo biodistribution of nanoparticles and the role of the protein corona. (WPs 1, 4)
7. Determine if there are indeed neurodegenerative disease endpoints for a range of animal models, and relate the extent of those endpoints to the amounts and properties of nanoparticles that have reached the brain. If, according to the usual measures of diagnosis (cognitive, behavioral and pathological), nanoparticles were found to affect the time of onset, or extent of the disease, then several examples of arrival dose versus response (by these conventional methods of diagnosis) would subsequently be constructed. The NEURONANO project thus sought to test the hypothesis that the more natural correlation is between the elevation of oxidative stress/amyloid mass and the disease outcome, i.e. to test the concept of the FIBROS property hazard map. (WPs 1, 2, 3, 4)
8. Provide input for a risk assessment framework (and screening protocol) for engineered nanoparticles for use by regulatory authorities and industry. The ultimate aim of the NEURONANO project was to begin development of a simple rational screening methodology for NEURONANOtoxic hazard, the so-called FIBROS map. The basis of the FIBROS map was the hypothesis to be tested within NEURONANO that screening for hazard in the future could be based on the three (presumed) contributing factors of engineered nanoparticles to neurotoxicity: access to the brain, oxidative stress generation potential, and amyloid fibril generation potential, possibly encapsulated in the arrival efficiency of the nanoparticles to the brain, and the location on the FIBROS hazard map. (WPs 1, 2, 3, 4, 5)
The NEURONANO project was thus a very focussed and detailed study of the potential for engineered nanoparticles to pose a neurotoxicity risk, and the development of a framework within which to predict the level of nerutoxic risk posed by specific nanoparticles.
Project Results:
The NEURONANO project was organised into 5 scientific and 2 management Work Packages. The inter-dependencies of the Work Packages are also illustrated. The flow of work in the 3 central experimental Work-Packages was based on a tiered approach, where experiments were conducted in order of increasing system complexity - experiments in solution, experiments in vitro, and finally, experiments in vivo. The purpose of this approach was to establish (where possible) in vitro methodologies to assess the potential neurotoxicity of engineered nanoparticles, and to reduce the numbers of animals needed in the course of the project, in line with the European Commission 3Rs policy regarding animal studies (reduction, replacement and refinement - Directive 86/609/EEC).
Progress towards the objectives and beyond the State of the Art
NEURONANO was an extremely ambitious project that set itself objectives far beyond the state of the art in the field. NEURONANO set itself eight broad objectives against which its success could be measured, spanning the activities of the five research work packages, and the dissemination workpackage. A summary of the NEURONANO consortium's progress towards achieving each of its objectives, during the three years of research activity, is given in the paragraphs below:
Objective 1: Develop nanoparticles that have appropriate labels of sufficient activity and duration for neurotoxicological studies
Radiolabelled nanoparticles: The main task of the JRC in the project was to develop cyclotron-based methods for radiolabelling of nanoparticles. Initially, extensive effort went into investigating the radiolabelling of TiO2 nanoparticles (selected as a priority nanoparticle at the project kick-off meeting for in vivo biokinetics studies at HELMUC). The result of theoretical considerations and many experimental tests was a successful protocol for ion-beam activation and subsequent dispersion and size selection of ST-01 TiO2 NPs, achieving samples suitable for the in vivo studies, whilst also ensuring that the structure of the nanoparticles was undamaged by the selected activation procedure. Several test irradiations on different types of TiO2 were carried out, and XRD structural damage assessment was performed on P25 and ST-01 TiO2 nanoparticles after proton irradiation. The XRD results indicated that no significant damage was caused, even at activity levels as high as 1MBq/mg, achieved after 20 hours irradiation. Much time was spent in developing a dispersion and size-selection protocol in order to achieve a 'bioavailable' suspension of activated ST-01. This included several steps of dispersion, leaching, centrifugation, 'washing' and filtration. The results are described elsewhere,[24] but it can be noted that the suspension consisted of nanoparticles of approximately 100nm hydrodynamic size (DLS performed at HELMUC), that were stably labelled to an activity level of more than 1MBq/mg. The in vivo studies indicated only minimal release of 48V from the activated ST-01 TiO2 nanoparticles.
Throughout the project further experiments have been carried out on TiO2, as it proved to be an extremely useful material (see below) for validation of calculations of the thermal effects of ion-beam irradiation on nanoparticle samples in general. Following this, experiments were also carried out on Fe3O4, CeO2, nanodiamonds, carbon black, and carbon nanotubes. Gold nanoparticles were also successfully activated and were used for in vitro work at IHCP/JRC.
In addition to direct ion-beam activation, a highly novel method for nanoparticle radiolabelling, based on recoil of 7Be from a Li containing powder, was developed. This method was shown to also be suitable for radiolabelling of SiO2 and Al2O3 nanoparticles and nanotubes.
In addition, the JRC activated Au and Ag electrodes for use in the HELMUC spark ignition aerosol inhalation facility, an additional task not originally foreseen in the NEURONANO workplan that is highly relevant for the project results. Finally, radiochemical synthesis of 56Co labelled SiO2 was achieved, and progress was made regarding synthesis of 89Zr labelled Al2O3.
Another approach of synthesis of labelled nanomaterials from activated precursors was also investigated, and radiochemical synthesis of SiO2 labelled with 56Co was achieved. A foil of iron was irradiated to produce the radiotracer, and then dissolved and the 56Co separated. This was added to nanoparticle synthesis precursor solutions and SiO2 nanoparticles were then created. Practically all the 56Co was incorporated in a stable way in the SiO2 nanoparticle matrix. Preliminary experiments showed that 89Zr Al2O3 could be synthesised in a similar fashion and radiochemical synthesis of several other NP types is now being pursued.
Fluorescently labelled nanoparticles: UCD were responsible for development of fluorescently-labelled particles that fulfilled the dual goals of being suitable for use in a biological context, such that they did not elute the fluorescent label inappropriately, and being sufficiently bright and stable that they can be tracked over time in complex milieu. Indeed, within the previous NanoInteract project, UCD had observed that many laboratory synthesized and commercially available fluorescent nanoparticles eluted free dye which confounded the uptake experiments, and had developed approaches to assess the elution of free dye from nanoparticles.[25]
Within the NEURONANO project, UCD focussed on development of optimised fluorescent silica nanoparticles that were resistant to dissolution in vitro over 48 hours. During the course of the NEURONANO project, the UCD team observed that the commonly employed Stöber type silica nanoparticles degrade by hydrolytic dissolution, which is accelerated in the presence of fluorescent moieties, which increase the porous nature of the particles, and under biological media conditions as compared to water alone. UCD have thus developed a method to render such nanoparticles' biologically stable for at least 48hrs, using an arginine-based approach to construct nanoparticles with a core-shell architecture, in which the highly fluorescent silica core is made by the Stöber method and a silica shell is added using the arginine approach in a two-step process. In making such particles, no effort need be made to limit the fluorescence intensity of the core, as this is no longer a significant element of the structural integrity of the nanoparticle.[26]
These particles showed good dispersibility in relevant media (as is usual for silica nanoparticles). They were incubated under biological type conditions with resultant dye leakage/dissolution monitored by time-resolved 1-D SDS-Page Gel and molybdate assay. The modified particles show vastly improved stability against dye leakage. The absence of any visible dye release over 48 hours is quite striking. When combined with the freedom to load a large amount of dye (or other internal) label these structures represent an ideal combination for biological application. The epifluorescence microscopy imaging after 22 hours incubation, in agreement with the gel result, show the increased background fluorescence for the Stober and Stober Shell particles as compared the Arg-Sil Shell ones (rightmost panel). Indeed, under different biological conditions, the contrast can be more striking.
Note: These particles are being further developed within QNano (the Research Infrastructure for nanosafety assessment) in WP6 on the optimisation of labelled nanoparticles for biological assays, and are being considered as for onward development as a QNano quality assured test nanomaterial.
Objective 2a: Identify classes of nanoparticles that lead to oxidative stress in relevant cell models
The original hypothesis was that the potential of the nanoparticles to cause oxidative stress (the Free Radical Generation Potential, FRGP) would differ for different particles depending on their chemical composition and potentially surface curvature and other properties, and that this could be connected to the physic-chemical properties of the nanoparticles as a hazard classification approach.
Thus, UEdin determined the intrinsic free radical activity of the NEURONANO panel of using the using DMPO spin trap approach to measure hydroxyl radicals. This well established approach uses Electron paramagnetic resonance (EPR) or electron spin resonance (ESR) spectroscopy, which is a technique for studying chemical species that have one or more unpaired electrons, such as organic and inorganic free radicals or inorganic complexes possessing a transition metal ion. Spin traps, such as Tempone H (superoxide anions) or DMPO (hydroxyl radicals) serve to 'trap' the fugitive free radicals so that their concentration can be assessed.
From the NEURONANO panel of nanoparticles, only one particle, TiV+ was cytotoxic and haemolytic. However, this particle (TiV+) did not have an excess of free radical generation as measured by EPR compared to the non-toxic NPs in the panel. Thus, the source of its toxicity is not related to ROS production. Similarly, amine-modified polystyrene were found to induce platelet aggregation, but not via the classical up-regulation of adhesion receptors. Rather the amine-PS nanoparticles appeared to act by perturbation of the platelet membrane, revealing anionic phospholipids. Neither oxidative stress generation by particles nor metal contamination was responsible for these effects, which were a result of differential surface derivatization.[27] The study reveals that NP composed of insoluble low-toxicity material are significantly altered in their potency in causing platelet aggregation by altering the surface chemistry.
In a previous study with a large panel of metal oxide nanoparticles, UEdin found a clear relationship between high positive zeta potential and haemolytic activity and between high positive zeta potential and ability to cause inflammation in rat lungs.[28] Thus, the zeta potential of the full NEURONANO panel of particles was determined as a possible factor in toxicity in the light of research in other projects. The TiV+ were found, along with the aminated polystyrene beads (one of the positive control nanoparticles), to have a high positive zeta potential under mild acid conditions.
Objective 2b: Identify classes of nanoparticles that lead to increased rates of fibrillation
At its conception, NEURONANO envisaged that, in analogy to the Reactive Oxygen Generation Potential (FRGP) concept described under objective 2a above, a set of solution studies to determine the ease with which fibrillation occurs for the relevant proteins (i.e. amyloid-ß and a-synuclein for Alzheimer's and Parkinson's diseases, respectively) in the presence of different nanoparticles, producing a 'fibrillation score', termed Amyloid Fibril Generation Potential (AFGP).
Objective 2c: Identify classes of nanoparticles that lead to up-regulation of the relevant pathway proteins involved in neurodegenerative diseases (at the cellular level)
The control of cell number plays a critical role in organ development and in maintaining the integrity of structural elements. Balance between cell proliferation and cell death represents the central aspect of this regulation, as well as in maintaining homeostasis.[31] Failure to regulate apoptosis is a common feature in several diseases, including autoimmune disorders, neurodegenerative diseases, cancer and AIDS.[31] In contrast to rapid turnover of cells in proliferative tissues, neurons commonly survive for the entire life of the organism. This enduring nature of neurons is necessary for maintaining the function of those cells within neuronal circuits. Excessive death of one or more populations of neurons results in disease or injury. For example, death of hippocampal and cortical neurons results in Alzheimer's disease (AD), death of mid brain neurons results in Parkinson's disease (PD), death of neurons in the stratium results in Huntington's disease (HD) and finally, death of lower motor neurons results in amyotrophic lateral sclerosis (ALS).[31]
Objective 3: Identify the physiochemical properties, and surface expression properties that lead to given levels of oxidative stress and amyloid load.
NEURONANO sought to test the hypothesis that the composition, organization, and reorganization time scales of the biomolecule corona (the proteins that adsorb to nanoparticles under physiological conditions) can be more naturally correlated to the key biological impacts associated with neurodegenerative diseases (i.e. oxidative stress and protein fibrillation), rather than any specific set of material properties. Thus, for the test nanomaterials, the nature and impact of the protein corona was assessed as the basis for such a correlation.
Nanoparticle protein corona studies
HMGU performed an in vitro incubation of triphenylphosphine surface-modified gold nanoparticle (5nm, 15nm and 80nm, identical to those used in the biodistribution studies) with diluted serum of mice. The nanoparticles were incubated for 24h at 37°C. After incubation the nanoparticles were washed three times and an SDS gel electrophoresis was made. After Coomassie blue staining the gel spots were cut and prepared for subsequent digestion with trypsin. The digestion was performed according to an established standard protocol of the core facility Proteomics of the Helmholtz Center Munich. After digestion the Mass Spec analysis was made with the help of the Proteomics Analyzer 4700, which is a tandem time of flight matrix assisted laser desorption and ionization apparatus (MALDI-TOF-MS). From this analysis, a number of different proteins which conjugated significantly to the gold nanoparticles.
Objective 4: Understand what constitutes a lead nanoparticle candidate for passing the Blood Brain Barrier (BBB)
Again, this initial conception while valuable in framing the project, was somewhat limited by our collective lack of knowledge at the time, as it missed two key factors that have since emerged as being of very considerable importance, namely:
(1)that nanoparticles can (potentially) signal across barriers via paracrine signalling, as has been demonstrated for the wear-induced particles across the foetal barrier,[40] and with the NEURONANO project for non-toxic carboxyl-modified polystyrene nanoparticles.[41]
(2)The low amount of particles reaching the brain could be a result of significant accumulation of nanoparticles in the brain endothelium itself. Indeed, using the in vitro barrier model, we observed significant accumulation of nanoparticles in lysosomes of the barrier cells, from as early as 4 hours after exposure, and reaching a plateau level at account 24 hours, suggesting that the barrier itself could be a target for nanoparticle accumulation.
Objective 5: Quantify transport efficiency to the brain. NEURONANO sought to quantify the amount of nanoparticles that arrive in the brain (of small animals) from each of the potential exposure routes (inhalation, instillation, ingestion, intravenous)
In vivo Exposure by Peripheral Injection to nanoparticles in rats
UU investigated the amount of nanoparticles reaching the brain of rats using a combination of TEM imaging and ICP-MS quantification of particle load.[45, 46] Peripheral injections of ceria (64 nm), titania (3-9 nm), silica (50 nm), gold (BBI 30 nm), gold-ApoE-PEG 2000 (30nm), gold-ApoE-PEG 5000 (30nm) and gold-ApoE-PEG 10000 (30nm) were carried out using intravenous and intra-tracheal methods. Particle suspensions were intratracheally instilled synchronous with the animal's inspiration followed by 300 μl of air. Animals were culled after 24 hours and blood was collected immediately by cardiac puncture, aliquoted into EDTA Vacutainer tubes and chilled. An aliquot of CSF was removed from the cisterna magna and frozen immediately. Animals were then perfused through the heart with PBS solution and all major organs (brain, heart, lungs, liver, kidneys, spleen, pancreas, stomach, intestines and muscle) removed for ICP-MS analysis. Ultrathin sections (100nm) were microtomed and mounted on copper TEM grids. Samples were imaged with a FEI Titan transmission electron microscope equipped with an energy dispersive X-ray (EDX) elemental analysis system.
Objective 6: Understand the detailed pathways that nanoparticles take to reach the brain
The animal studies described under Objective 5 above were also intended to provide as much information as possible about how the nanoparticles reached their destination, including the detailed nature of what is expressed on their surface as they pass through different intermediate tissue types on going to the brain. Thus, even if no neurotoxicity was observed, the investment of resources, and animal experimentation would result in significant new data and a deeper understanding of the in vivo biodistribution of nanoparticles and the role of the protein corona.
Mechanisms of uptake and intracellular transport of nanoparticles
To date, most studies have focused on identifying the point of entry and final destination of nanoparticles. However, to truly understand and predict the fate and impacts of nanoparticles, and specifically to identify rare events, a deeper knowledge and understanding of the entire intracellular itinerary and event sequence will be essential. It is, in principle, possible that the original corona, combined with new molecules picked up inside the cell, could, even if rarely, provide the nanoparticles with signals to allow them to access different subcellular destinations rather always following the endo-lysosomal pathway.
Role of cell cycle in nanoparticle uptake (including in polarised / barrier cells)
UCD have recently demonstrated the strong relation between nanoparticle uptake and cell cycle in proliferating cells and developed methods to measure if nanoparticle export is present[54]. In proliferating cell populations, during cell division, nanoparticles are split among daughter cells, thus the internalized nanoparticle dose is diluted. These findings can have important implications for slow- or non-dividing cells, such as for instance the blood barrier itself, where such diluting effect may be missing or less active, and if export is not present, bioaccumulation could arise over longer time scales. Thus, together with the question of transcytosis, it is important to study how nanoparticle uptake and cell cycle relates in such specialized barriers and if nanoparticles can impact the barrier itself.
In vitro transport studies in the model BBB - evidence of nanoparticle transcytosis
In recent years, significant effort has been directed towards mimicking the in vivo BBB using immortalised cells. The hCMEC/D3 cell line was chosen as it is the closest available correlate to the human in vivo situation, and does not require co-culture with primary astrocyte cells, making it easier to work with for quantitative studies. In addition, the porous transwell system upon which the hCMEC/D3 monolayer was grown has previously been utilised in the presence of monolayers from alternate cell lines by other groups (19, 20).[55] As this cell line can differentiate into a monolayer in culture after 7-10 days, seeding the cells onto a permeable filter, the so called 'transwell' system enables an in vitro BBB model to be established. The transwell is composed of an insert filter on which the cell monolayer is grown, and an acceptor well, with the apical chamber mimicking the blood and the basolateral chamber mimicking the brain. Thus, molecules loaded into the apical chamber are presented to the cellular barrier formed on the filter, which can then to be taken up by the cells by a receptor-mediated (transcytosis) process.
Role of proteins in corona in vivo on nanoparticle distribution
The pulmonary route is considered to be one of the primary exposure routes for inadvertent exposure to nanomaterials, as well as being very promising for drug delivery by inhalation. In this regard, understanding the role of lunch surfactant proteins on the biodistribution of nanoparticles is vital. Directly after their deposition, inhaled Au NP come into contact with pulmonary surfactant protein D (SP-D). SP-D can agglomerate Au NP in vitro, and this may influence the clearance as well as the systemic translocation in vivo.
Mytilus edulis (mussel)
The mussel genus Mytilus is one of the most widely-used groups of sentinel species in aquatic toxicology, having been used for many years in environmental monitoring programs such as Musselwatch.[58] Advantages of this species include the mussels' wide geographic range, abundance, ease of identification, filter-feeding habit and sedentary nature. The animal is also robust enough to withstand moderate levels of pollution. This allows a broad range of loadings, with a resultant range oxidative stresses. It is also believed that the overall range of oxidative stress processes in Mytilus is comparable to that of animals, and therefore its whole-organism proteome is an attractive analytical tool to screen for tissue level oxidative stress. Indeed, in recent years, proteins expressed in Mytilus have come to be widely used as biomarkers of environmental stress, which is now considered a form of validation of the species for environmental (including oxidative) stress.[59]
Nereis virens (ragworm)
The marine polychaete, Nereis virens, a common species in estuarine sediment and infaunal deposit feeder, was used as a further animal model. This species is known to select and ingest large amounts of fine sediment particles that tend to be enriched in organic material and metals. The Worms used in this exposure trial were bought at Halfway angling centre, Cork. Clean sediment was bought in from a local supplier (artificial play sand).
Neuropathological and histological studies of rat brains
The effect of titania nanoparticles (ST01 titania anatase) on plaque formation was analysed in the 5XFAD Alzheimer's mouse model. These transgenic mice express 5 forms of human familial Alzheimer's Disease mutations (Co-overexpression of the human APP695 with the Swedish (K670N, M671L), Florida (I716V), and London (V717I) FAD mutations in the same APP molecule and human PS1 harbouring two FAD mutations, (M146L and L286V). These mice develop amyloid plaques, show impairments in neuronal transmission and synaptic plasticity as they age, and develop memory impairments similar to patients of Alzheimer's disease. Amyloid deposition and gliosis is accelerated in these mice compared to other models, with significant levels observed at 2 months of age.
Behavioral studies in rats and mice
Behavioral and memory changes are the first subtle changes of brain functionality that are observed well before any anatomical changes occur in neurodegenerative diseases. Thus, in addition to looking for physiological markers of neurodegeneration in response to exposure to nanoparticles, NEURONANO partners also performed some behavioral studies on nanoparticle-exposed and control animals.
The conclusion from these studies is that very little behavioural effects were observed in the rats or mice. Further data analysis is still underway and will be published over the coming months. These findings are also consistent with very small amounts of the ST-01 titania particles reaching the animal brains following IV or IT exposure.
Objective 8: Provide input for a risk assessment framework (and screening protocol) for engineered nanoparticles for use by regulatory authorities and industry
The ultimate aim of the NEURONANO project was to begin development of a simple rational screening methodology for NEURONANOtoxic hazard, the so-called FIBROS map. The basis of the FIBROS map was the hypothesis to be tested within NEURONANO that screening for hazard in the future could be based on the three (presumed) contributing factors of engineered nanoparticles to neurotoxicity: access to the brain, oxidative stress generation potential, and amyloid fibril generation potential, possibly encapsulated in the arrival efficiency of the nanoparticles to the brain, and the location on the FIBROS hazard map.
We can summarise the findings from the NEURONANO project as they relate to the concepts of the FIBROS map as follows:
Fibrillation potential
-Induction of fibrillation is a surface area effect - all nanoparticles induce or modulate the rate of protein fibrillation, and there appears to be a detailed interplay between monomer, oligomer and fibril for residence at the nanoparticle surface.
-The presence of a protein corona on nanoparticles, such as would inevitable occur during nanoparticle uptake in vivo, slowed the fibrillation of Aβ, regardless of size, surface chemistry, charge or aspect ratio when compared with the same pristine (bare) nanomaterial.
-In order to be effective as an inhibitor of Aß fibril formation the proteins in the nanoparticle-protein corona must be in their native form: corona proteins that have been denatured by heat were found to actually accelerate the formation of protein fibrils, mostly likely due to the exposure of hydrophobic amino acid residues from proteins that do not refold after heating.
-Further work is required to understand more fully the specific mechanism by which fibril formation is inhibited by the presence of a protein corona around nanomaterials, and also to determine whether slow protein degradation in vivo in the presence of nanoparticles could contribute to protein fibrillation and neurodegeneration over the long term.
-The fact that particles bring their corona with them all the way to their final location in cells suggests that there may also be subtle effects from the presence of serum proteins in cells when they are normally excluded from cells. Whether this would also have a role in protein fibrillation in cells, or indeed induce other impacts requires further investigation.
-The fact that the positively charged polystyrene particles, which induced apoptosis in astrocytes, were also found to up-regulate signatures related to protein aggregation and fibrillation is significant, and suggests that this could be an alternative measure against which to assess fibrillation potential of nanoparticles. A wider panel of particles would need to be assessed, along with an in-depth biomarker validation process in order to demonstrate this link robustly.
Reactive oxygen stress generation potential
-No link between cellular oxidative stress and nanoparticle effects could be confirmed. Indeed, the particle that was found to induce the most effects in cellular assays, the TiV+ which was cytotoxic, haemolytic and induced mitochondrial damage, did not have an excess of free radical generation as measured by EPR compared to the non-toxic nanoparticles in the panel.
-A link between particles with a positive surface charge and toxicity was also investigated, and while the connection held in some assays, the TiV+ particles were found to damage mitochondrial function, whereas the positively charged control nanoparticle, amine-modified polystyrene nanoparticles were not, suggesting that more than simple charge was involved in the induction of this particular effect.
-Oxidation of protein thiols and carbonyls is a generally-applicable, sensitive marker of oxidative stress induced by nanoparticles in both cells and more complex organisms such as mussels and worms (much more robust than measuring cellular oxidative stress in the presence of nanoparticles). However, as with the biomarkers for protein fibrillation, significant additional screening, plus validation work is required.
Access to the brain
-Route of exposure matters in terms of how much of the applied particle dose reaches the brain, although even with the most efficient transfer to the brain, less than 1% of the applied dose reaches the brain.
-Using ST-01 titania nanoparticles, brain accumulation was found to peak at around day 7, but particles could still be detected by day 28 following a two-hour inhalation (mean and SD, n=4) in rats. On the other hand, following peripheral injection of the same nanoparticles in rats, titania particles were found only in liver and lung and no particles were found in brain tissues.
-A significant increase in cerium content against control (6-fold; p<0.05) was observed in brain tissue of rats injected intravenously with ceria nanoparticles. This data may suggest that ceria passes the blood brain barrier readily.
-Nanoparticle size plays a role in determining the amount of access to the brain, although to a smaller degree that expected for exposure via intertrachael instillation and intravenous administration: brain accumulation of 5 nm - 200 nm gold NPs is similar for all particle sizes tested. Only the 1.8 nm particles showed significantly enhance brain uptake via these routes.
-The exposure route dependence is related to the nature of the biofluid milieu that the particles initially encounter and thus to the nature and composition of the protein corona that forms: exposure via the oesophagus led to the highest accumulation in the brain for the 18 nm gold NPs, indicating that something happens with these NP in the GIT which alters their biokinetic fate. The most obvious explanation is a different specific protein coating, related to the exposure route. In the stomach, it is conceivable that the acidic environment, as well as the presence of gastric enzymes, are likely to degrade the triphenylphosphine surface coating of the applied NP. Thereby, protein coating of the blank NP surface may occur.
-Although uptake is rather low compared to the initially injected dose, conjugation of gold nanoparticles with either ApoE or albumin significantly increased the amount of particles detected in all investigated brain compartments compared to the bare citrate stabilized gold NP. Retention after 24h was very similar in the total brain for both apoE conjugates.
-The distribution pattern for single brain compartments was very similar for both apoE conjugates. Therefore we conclude that both conjugation methods led to an increased translocation into the brain across the blood-brain barrier. Compared to the apoE conjugates the albumin conjugates seem to have an even better access to the brain.
Potential for indirect signaling
-The low amount of particles reaching the brain could be a result of significant accumulation of nanoparticles in the brain endothelium itself. Indeed, using the in vitro barrier model, we observed significant accumulation of nanoparticles in lysosomes of the barrier cells, from as early as 4 hours after exposure, and reaching a plateau level at account 24 hours.
-While such significant accumulation of non-toxic nanoparticles, such as the Ps-COOH shown here, and also silica nanoparticles or gold nanoparticles which we have also studied using this model, do not seem to perturb the cells, there are indications that signalling molecules are secreted by the endothelium in response to the presence of the PS COOH NPs, which are subsequently amplified by the astrocytes when using a co-culture model.
-Growth of human astrocytes below the BBB monolayer allows the communication of supporting signals in vitro which aids in mimicking a more realistic in vivo situation, allowing these signalling pathways to occur. Assessing impacts of less benign particles accumulating in endothelium is necessary, although as the particles themselves begin to induce effects on the endothelium barrier separating out the sequences of signalling events will become more difficult, and isolating paracrine signalling from inter-cellular signalling will require detailed molecular biology studies.
Need to connect access to the brain with impact
-The neuropathological review of animals exposed intra-tracheally to ST01 titania anatase (3-9nm primary particles, 55 nm in dispersion) indicated that these particles did not cause any detectable neuropathology effects: No significant differences in the volume density of plaque, the total plaque volume or the total activated astrocyte number were demonstrated between control and treatment groups following 6 months exposure to titania (ST-01) nanoparticles administered intra-tracheally.
-Thus, from a risk assessment viewpoint, simply observing that a small percentage of particles can reach the brain may not actually represent a risk, if there is no associated hazard. However, given that these studies were performed on limited numbers of animals, a very limited panel of particles, significantly more research is required in order to make a real connection between access and impact. Additionally, the particle tested were themselves very benign, and thus more inherently toxic particles should be tested in the future. However, from the viewpoint of the FIBROS model for risk assessment, it is clear that access and impact need to both be assessed, and dose-response elements included also.
All NEURONANO publications are available at http://www.ucd.ie/cbni
Potential Impact:
Dissemination activities and the exploitation of results
The NEURONANO project has already resulted in over 50 publications , with several more in press or under review. The knowledge and experience gained is now forming the basis for other projects, both in the EU and across the world, including incorporation into several other EU projects, such as QualityNano, the EU infrastructure for nanosafety assessment which is disseminating many of the NEURONANO findings to the community via training schools and the development of positive and negative control nanomaterials; and NanoTransKinetics, a modelling project that aims to develop a predictive tool for determining uptake, localisation and impact of nanomaterials based on their physic-chemical characteristics and their subsequent biomolecular corona.
From the project conception stage, emphasis was placed on ensuring impact for, and contribution to standardisation efforts, including:
-ensuring that the NEURONANO partner's efforts to establish best practices within the group should lead to protocols of value to the international community of scientists working in the nanotoxicology arena. The project was designed so that members of the Advisory Board who liaise with bodies responsible for standards such as CEN WG 166, ISO newly created TC on nanotechnology will be aware of the best practices that we are developing.
-building on the technical and focused nature of the project, which was heavily biased towards quantitative and verifiable methods, to ensure that the outputs are framed in such a way that they contribute key scientific data to the development of standards, test and measurement methods in the field of nanoparticle interactions with biological systems.
Enduring impact in a range of scientific and quality assurance areas
The NEURONANO project has already had, and will continue to have, impact far beyond its size and budget, in part due to the formative stage of the research field at the outset of the project, and in part as a result of the fact that it has facilitated the development of several insights that will have a durable impact on the research field. Among the key scientific developments resulting from the NEURONANO project were the fact that the cell cycle may play a significant role in determining the accumulation or otherwise of nanoparticles in key cells, the fact that the nature of the protein corona can be significantly different under in vitro versus in vivo conditions due to the significant differences in the amounts of proteins present relative to the available particle surface area.
Radiolabelled nanoparticles synthesised by JRC
A key output from NEURONANO is the availability of a range of radio-labeled nanoparticles for use in dosimetric / biokinetics or environmental fate and behaviour studies, or other studies where tracking of nanoparticles over specific time periods is required.
Pre-synthesised Au, CeO2, Fe3O4, and TiO2 nanoparticles can be radiolabeled using ion-beam methods, and Au and Ag electrodes can be activated for subsequent nanoparticle synthesis.
Positive and negative control Nanoparticles for apoptosis and cell cycle regulation
Having studied in detail in several cell lines and several species the impacts of amine-modified polystyrene nanoparticles, and found a consistent and dose-dependent onset of apoptosis (linked with autophagy under some circumstances), as well as a low-dose impact on cell cycle, these nanoparticles were determined to be an ideal candidate for onward development, and round-robin validation as positive control nanomaterials for apoptosis and cell cycle regulation. The fact that they are easily labelled for tracking and surface functionalised with a variety of functional groups adds to their attractiveness as a positive control candidate. Since the carboxyl-modified equivalent particles show almost no effects on tested cells or organisms at equivalent exposure doses and times they are an ideal candidate for onward development as negative control nanoparticles for the same end-points.
Protocols covering a range of aspects of nanoparticle interactions with cells, barriers and organisms
Protocols covering a variety of aspects, from nanomaterials synthesis and characterisation, surface functionalisation and dispersion, to all aspects of nanoparticle uptake and localisation, apoptosis, assessment of changes in gene expression, oxidative stress and redox proteomics, dosimetrics and biokinetics and assessment of behavioral impacts of nanomaterials, have been developed within NEURONANO. Many have already been published as part of scientific manuscripts, and many others are currently in final stages of preparation for publication.
New assays and/or experimental approaches developed within NEURONANO
Effect of ratio of available protein to nanoparticle surface area on resultant corona (in vitro versus in vivo coronas)
A key advance developed within NEURONANO has been the understanding that the corona is context dependent: i.e. the corona formed under low protein conditions (i.e. in vitro conditions) may not be identical to that formed under high protein conditions (i.e. in vivo conditions).[65] The protein adsorption for two compositionally different NPs, namely sulphonated-polystyrene (PSOSO3) and silica (SiO2) NPs was studied at different protein concentrations keeping the particle surface area constant. NP-protein complexes were characterized by Differential Centrifugal Sedimentation (DCS), Dynamic Light Scattering (DLS) and Z-potential both in situ and once isolated from plasma as a function of the protein/NP ratio. A semi-quantitative determination of their hard corona using one-dimensional sodium dodecyl sulphate poly-acrylamide gel electrophoresis (1D PAGE) and electrospray Liquid Chromatography Mass Spectrometry (LC MS/MS) was introduced, which allowed us to follow the total binding isotherms for the particles, identifying simultaneously the nature and amount of the most relevant proteins as a function of the plasma concentration. We find that the hard corona can evolve quite significantly as one passes from protein concentrations appropriate to in vitro cell studies to those present in in vivo studies, which has deep implications for in vitro-in vivo extrapolations and will require some consideration in future.
Cell cycle effects
Work at UCD regarding the role of cell cycle in determining nanoparticle uptake rates and intracellular dose[66] have resulted in some important protocols and experimental design considerations. In summary, the nanoparticles studied here do not disrupt cell cycle progression and, once internalized, localize in lysosomes, where they remain. Nanoparticle export is negligible and dilution of the intracellular nanoparticle load occurs by cell division. Uptake rates during the different cell cycle phases are comparable.
TEM protocol for assessing integrity of the in vitro BBB
In light of the increasing drive to reduce reliance on animal testing there is a concerted effort to develop and validate alternative in vitro methods for research, screening and regulatory assessment of the impacts of chemicals, including nanoparticles. However, to date there remain important concerns about the reliability and relevance of such in vitro methods, and whether they represent a realistic alternative to animal studies. Methods to ensure the quality and reproducibility of in vitro methods and the comparability of the data generated to the in vivo models they are intended to replace are required to begin to address these issues.
The neutral red retention time (NRRT) assay
The NRRT assay is useful for detecting decreased lysosomal membrane stability in haemocytes sampled from bivalves, a phenomenon often associated with exposure to environmental pollutants including nanomaterials. Bivalves are popular sentinel species in ecotoxicology and use of NRRT in study of species in the genus Mytilus is widespread in environmental monitoring. The NEURONANO partner UCC has shown that this assay can be applied to a panel of metal and metal oxide nanoparticles (2ppm) and distinguish between those that are toxic and those that are not, and connect this with markers of oxidative stress or other toxicity pathways, and they suggest that this could support use of NRRT as a generally applicable in vitro assay in nanosafety assessment.
Impact for standardization
Several of the approaches developed or optimised in NEURONANO are suitable for standardisation, and indeed NEURONANO was presented at two of the European Commission workshops on standardisation activities from EU NMP projects, as follows:
Dr. Iseult Lynch presented data specifically related to standardisation from NEURONANO at the NMP Seminar on Standards & Standardisation on December 14th 2010 in Brussels. Talk title: 'NEURONANO (EU FP7 214547-2) Contribution to standardisation processes'. Significant interest was expressed by CEN in the aspects of the protocols for Differential Centrifugal Sedimentation characterisation of nanoparticle dispersions in situ in biological fluids and in the validation of the in vitro model Blood brain barrier, and follow-up is underway.
Why do we need to describe the dose of nanoparticles?
There is a need to know the in situ dose in nanoform (< 100 nm). To achieve this, NEURONANO proposed:
- Characterisation in situ in biofluids
- Combination of DCS and TEM describes dose presented to cells
- Time-resolved data (evolution of sample)
- Applicable to any biofluid, including environmental (e.g. natural organic matter).
Impact for regulatory agencies and policy groups
Neurodegenerative diseases currently affect over 1.6% of the European population, with dramatically rising incidence that is likely (in part) due to the increase in the average age of the population. There are persistent claims, based on the epidemiology that pollution may be a cofactor in Alzheimer's disease, although the evidence is controversial. The risk that engineered nanoparticles could introduce unforeseen hazards to human health is now a matter of deep and growing concern in regulatory bodies, governments and industry. However, at present there is only circumstantial evidence that nanoparticles could impact on such diseases.
Recommendations for further research (from NEURONANO Final meeting and workshop)
As part of the NEURONANO final meeting (5th December 2011), and in collaboration with the ESF EpitopeMap Research Networking Programme , a joint workshop was held in Dublin on 6th - 7th December 2011 that brought together leading European experts in Engineered nanoparticle transport across biological barriers to identify the central questions that require scientific and policy attention as a matter of priority. Among the outcomes from the workshop of particular relevance for NEURONANO and its stakeholders were:
Accidental 'targeting' of ENP to the brain - need for implementation of 'protein corona' screening
Many of the 'success' stories to date in terms of using engineered nanoparticles to deliver drugs to the brain have been the result of the ENP surface offering an appropriate scaffold for known blood-brain-barrier receptors, such as the LDL receptor. In fact, one key study shows that having a specific surfactant coating (polysorbate 80 - Tween 80) that spontaneously binds apolipoprotein E and apolipoprotein A-I from the dispersion medium was as effective at delivering the drug to the brain as the specifically surface-engineered ENP where these apolipoproteins were chemically grafted to the ENP.[70-72] This raises a very important question as to whether other ENP, which were never intended to reach the brain (or indeed other key organs, or the foetus), could unintentionally bind such targeting protein coatings and then reach delicate organs with very high efficacy.
Funding recommendation: Research is needed to improve existing and create new barrier models and develop clear in vitro-in vivo correlations, including appropriate culture conditions (serum free) to optimise barrier functioning and cell-cell (paracrine) signalling. Research should be directed towards the use of co-culture and 3D cell culture approaches, and towards comparative in vitro and in vivo studies under similar conditions, and where limitations are documented and understood.
Infrastructure needs for nanosafety and nanomedicine
A key concern relates to large scale facilities and specialised research centres for supporting the safety and efficacy assessment of ENP (for example facilities for radiolabelling ENP and assessment of the biodistribution and biokinetics of radiolabelled-ENP). Funding to support such facilities, and the mobility of EU researchers / post-docs for their wider exploitation, over the coming decade is vital, as many questions regarding safety and efficacy of ENPs can only be answered by the use of not yet available radioactive nanomaterials. In a parallel fashion, support to maintain the intellectual infrastructure (appropriately trained personnel) is required to respond adequately to EU needs in relation to ENP.
Project website: http://www.neuronano.eu
Coordinator contact details:
Prof. Kenneth Dawson, Director Centre for BioNano Interactions, University College Dublin, Belfield, Dublin 4, Ireland.
Phone: 00353 1 716 2459
Email: Kenneth.A.Dawson@cbni.ucd.ie
