Final Report Summary - INLIVETOX (Intestinal, Liver and Endothelial Nanoparticle Toxicity Development and evaluation of a novel tool for high-throughput data generation)
Increasing concerns about the safety of new active chemical ingredients used in pharmaceuticals, cosmetics and the food industry have been further highlighted by the emergence of materials with dramatic new properties based on nanoparticles (NPs). NPs are particles with at least one dimension on the nano-scale of 100 nm or less. In parallel, ethical concerns about the use of animals to test the safety or efficacy of new compounds are growing. In vitro testing offers a potential solution to the challenge of how to ensure that as NPs are developed and used, any unintended consequences of exposure to humans are minimised. The INLIVETOX project focused on the impact of NP exposure via ingestion, on the vascular endothelium, liver and gastrointestinal (GI) tract. Exposure via ingestion is particularly relevant due to the inclusion of NPs in food, food packaging and in oral medicines.
During the project, a novel in vitro model of the GI tract, vascular endothelium and liver was developed (the ILT system) for in vitro studies of nanotoxicology by ingestion. CaCo-2 cells were used to model the intestinal epithelium, HUVECs the vascular endothelium and C3As the liver. Each model tissue was maintained in a separate bioreactor. The model included two fluidics circuits: one circuit modeled the GI tract and the other - the model blood stream - connected all three tissue models. A new bioreactor (ILT2) was developed during the project to house the intestinal epithelium. Both fluidics circuits flowed through ILT2, separated by the model intestinal epithelium.
A major effort in the project was focused on reference measurements and the establishment of protocols for the ILT system. The physicochemical properties of the NPs to be used - gold (Au), silver (Ag) titanium dioxide (TiO2) and polystyrene (PS) - were characterised. Common culture conditions were established for the three cell types in the ILT system. Protocols for assays for cell viability and function were defined. Assays were also established for markers of oxidative stress, apoptosis and inflammation. Finally, baseline values for toxicity for the different NPs and individual cell lines under static conditions were determined.
A model of the inflamed intestinal epithelium based on cell lines (CaCo-2, THP-1 and MUTZ-3) was also established. The model was not suitable for the ILT system. However, comparative studies of NP toxicity using this model and simple CaCo-2 layers demonstrated a protective effect of the immune cells in triple co-culture. This protective effect was impaired when the co-culture model was inflamed. These models represent an important step towards in vitro modelling of, not only the healthy, but also the diseased gastrointestinal tract in nanotoxicology.
In vivo experiments carried out using rats, provided comparative data on NP exposure by both ingestion and injection. Biokinetics experiments provided data on the distribution and fate of Au and TiO2 NPs. Not on Ag NPs, however, due to severe problems with their labeling. Selected tissues - the intestine, Peyers patches (PP) aorta and liver - were also analysed for NP uptake, oxidative stress and changes in gene expression for a panel of genes involved in inflammation, oxidative stress and apoptosis.
Comparison of the data obtained in vivo on exposure by injection and ingestion with data obtained from static, single cell in vitro assays and the ILT system showed a remarkable pattern of differences and similarities. In vivo exposure to Ag and TiO2 NPs by intravenous injection gave strikingly similar gene expression responses to those obtained using static in vitro cultures of C3A, suggesting that simple in vitro models using human cell lines can compare very well with intravenous in vivo rodent studies. This result is particularly noteworthy since hepatocyte cell lines are often regarded as sub-optimal for toxicology testing due to their lack of expression of metabolising enzymes.
Modelling nanotoxicity by ingestion is much more complex than by injection, as interactions of particles with the food, translocation through the GIT and distribution, as well as interactions with cells along the way and the communication of these cells with each other and on a systemic level have to be taken into account. Poor correlation was obtained between in vivo and static single-cell in vitro experiments. In vitro data on nanotoxicity by ingestion was also obtained for Ag NPs using the ILT system. In the ILT model of toxicity by ingestion, Ag NPs caused a dose dependent decrease in viability of the three cell types. Dose-dependent losses of cell function (epithelial barrier function, production of albumin) and increases in inflammatory markers were also observed. These results compared poorly with the in vivo data on Ag NP toxicity by ingestion, correlating much better with Ag NP toxicity by injection.
It would be useful to test the ILT response to the lower toxicity Au NPs and compare it with the in vivo responses to ingestion. Although Au was lower in toxicity, it induced a wider array of changes in gene expression than Ag allowing for a more in depth comparison. It would also be useful to compare the results of the ILT system with injection responses in vivo for a wider array of NP and to compare these with static cultures to assess whether the additional complexity of the ILT system provides any advantages over the cheaper and easier static models for predicting responses to intravenous injection.
The INLIVETOX system developed in this project has potentially much wider application than just the testing of responses to NPs. Throughout the project there has been extensive dialogue with members of an industrial advisory group from the pharmaceutical, cosmetics, food and household products manufacturing industries. Outputs from the project include tested demonstrators of a cell culture bioreactor that will be commercialised by partners from the consortium.
Project context and objectives:
Engineered NPs (also called manufactured nanomaterials) can be considered a new class of substance, since chemical and physical properties at the nano-scale can differ hugely from those of bulk materials. Typical examples include silver particles used as an antibacterial and antifungal agent in water bottles, zinc oxide used to block UVA and UVB sun rays in sunscreen, and the various materials used as pigments in paints and cosmetics. The novel properties of NPs - that make them so attractive for many applications - means that they can also interact with the body in different and unexpected ways, and their small size allows them to potentially cross barriers in the body that hold back larger materials. For this reason, a rapid and thorough evaluation of NP toxicity is essential.
Given the large numbers of NPs that exist in the lab and sometimes in commercial products, researchers and industry are faced with the almost impossible task of analysing the safety of thousands of new materials in a very short time frame. It has become clear that, from both ethical and financial viewpoints, new animal-free testing methods are necessary. Unfortunately, the animal-free cell culture or 'in vitro' models currently in use are often not representative of the human body. They simplify enough to allow the production of data, but often so much that the data is not representative and cannot replace most testing on animals. Most in vitro tests work with one type of cell in culture, but this is, of course, not the way cells work in our body, where the surrounding environment, including other cell types, has a strong influence on how cells behave. The interactions between different organs, or the transport of a material through barriers, for example from the intestine into the bloodstream, cannot yet be modelled in a realistic, or 'physiological'. way.
The EU-sponsored project INLIVETOX (please see http://www.inlivetox.eu(öffnet in neuem Fenster) online) has brought together engineers and biologists from universities and research institutes across Europe in order to develop an improved cell culture system. This model will help us gain insight into how NPs taken up in food or otherwise ingested can cross the intestinal wall and how they can affect the different tissues in the body.
The global objectives of INLIVETOX were:
- to improve understanding of NP toxicity by ingestion;
- to develop improved in vitro toxicity testing methods.
In order to achieve these objectives, the project aimed to develop a modular microfluidics-based in vitro test system, the ILT system (WP 1); a system in which different cell types - each modelling a target tissue - could be maintained and interact via cytokines. The target tissues chosen for this project were the GI tract, the liver and the vascular endothelium so that NP toxicity by ingestion could be modelled with the system.
Subsidiary objectives in the project were the establishment of common culture conditions for the cells and test protocols for the analyses to be carried out with the ILT system (WP 2). Once established, these were then to be implemented in the ILT system (WP 3).
Validation data were to be obtained via an in vivo study of NP toxicity by ingestion in rats carried out in parallel. Data on both biokinetics and toxic response from the target tissues were the goal of WP 4.
The final validation, bringing together in vivo data with new data generated by the ILT system, was the objective of WP 5.
The long-term objectives of the project are:
- to ensure the safe development and use of NMs for commercial applications;
- to commercialise a test system to screen NMs for their toxicity.
These two objectives are related to the foreseen impacts of the project, with future commercialisation of the INLIVETOX system making it available to the whole toxicology community.
Project results:
WP 1: Development of microfluidic systems
WP 1 was the engineering and design WP of the project, and also involved mathematical modelling. The underlying concept of the project is that of connected cultures in which two or more tissues are connected together by a fluid flowing between them, much as in the human body in which distant tissues and organs are connected by the bloodstream. Our aim was to engineer a connected culture fluidic system in which nanoparticles could pass through an intestinal barrier and then influence downstream tissues to simulate the route of ingested nanoparticles in the body. It was decided early on in the project to build a modular fluidic system so that different tissues could be added or removed from the fluidic circuit as required. We also choose to use cell culture systems which were easily translatable to biological experimental standards.
The first part of the WP was dedicated to the scaling of the INLIVETOX system using a method known as allometry. This method is based on the well-known fact that the different features of animals are correlated to their body mass. Using allometry we calculated the liver, vascular tissue and intestinal cell numbers necessary to properly represent the human body in the reduced scale of our system. At the same time, we also considered other parameters such as flow, oxygen and fluid forces. Having identified the cell numbers and other parameters, we proceeded with the design of the system. During this period, the engineering partners trained the other partners in the consortium in fluidic culture methods using a bioreactor chamber known as ILT0. ILT0 is an interconnectable commercial cell culture chamber designed and commercialised respectively by two partners in the team.
The main challenge in the WP was the design of the intestinal chamber which requires a semi-permeable membrane to separate the intestinal side from the blood side and two different fluid flows on either side. Two versions of the chamber were developed, ILT1 and the much improved second version, ILT2.
Challenges identified were:
i) the need to continuously monitor the function of the intestinal cells using a technique known as TEER (transepithelial electrical resistance). TEER is a standard and sensitive method of analysing the integrity of epithelial barriers such as the intestinal wall.
ii) the need to use ultrathin microporous membranes to allow NPs to pass from the intestinal side to the blood side. Most commercially available microporous membranes rapidly become clogged when exposed to NPs.
To meet these challenges an ultrathin silicon nitride membrane was fabricated with built in electrodes for TEER measurements. The first bioreactor chamber ILT1 was designed to hold the membrane and had electrical contacts for TEER measurements. After the first few prototypes were fabricated and tested, a training course was held to introduce partners to the new system. The cell culture partners then started using the ILT1 to perform baseline measurements on intestinal epithelia.
A few months later enough feedback had been collected to begin the design of ILT2. The most fundamental problem was a difficulty in seeding cells on the microfabricated membranes and then placing the membranes in ILT1. Other problems concerned leakage, air bubbles and difficulty with stable TEER measurements. ILT2 was designed using innovative solutions (some were patented) to overcome these problems. It was also more robust and the TEER measurements more stable due to new watertight contacts and more robust electronics.
The final part of the WP was dedicated to training of other partners, modelling NP passage through the membrane and NP sedimentation in the fluidic system and troubleshooting when technical hitches arose during cell culture.
WP 2: In vitro biological models
Laying the ground work for all further in vitro studies in the INLIVETOX project, WP 2 established protocols to ensure reproducible and stable size distributions in the NP suspensions as well as defining assays and endpoints to quantify cell viability and toxic responses in the INLIVETOX system. Baseline data under static conditions was generated for comparison in later stages of the project. Furthermore, the in vitro models used in the project were optimised with regards to culture conditions, selection of functional markers and suitability for higher throughput applications (e.g. an existing three-dimensional (3D) model of the inflamed intestinal mucosa was modified to improve reproducibility).
As the INLIVETOX system is based on studying the cross talk between three different tissue models connected to each other via a fluidics systems, common culture conditions had to be identified for the three cell lines Caco-2 (enterocytes), C3A (hepatocytes), and HUVEC (endothelial cells). A common medium was developed based on EMEM. This medium, referred to as EMEM-GF, affected neither cell adhesion nor morphology either in static or dynamic conditions. Under the chosen conditions, all cells maintained their viability as well as their tissue-specific phenotype and were able to withstand the shear stress of dynamic conditions. Albumin expression was used as the phenotypic endpoint for C3A hepatocytes while expression of von Willebrand factor was the marker for endothelial cells and barrier function (quantified via transepithelial electrical resistance) characterised the functionality of Caco-2 cells.
In order to interpret the toxicity of nanoparticles and understand the underlying mechanism of potential toxic effects, the physicochemical characteristics of the NPs need to be known. Previous studies have correlated toxicity of various nanomaterials to parameters such as surface area, surface reactivity or phase composition. Furthermore, it was shown that the dispersion medium used can greatly influence the physiochemical properties of NPs. Therefore, we characterised the NPs to be used in the INLIVETOX system with regards to their size, polydispersity, aggregation behaviour and surface charge in various media to be used for in vivo and in vitro investigations. Most nanomaterials were found to form stable dispersions in EMEM-GF with particle sizes close to the nominal size. Only NM101 TiO2 NP showed significant aggregation behaviour, forming 700 nm sized clusters (primary particle size 7-11 nm). Serum proteins sterically enhanced the stability of the NP dispersions as aggregation behaviour was more pronounced in serum free conditions. Dispersions were stable for at least 48 h, allowing their use in the INLIVETOX system.
Different assays evaluating various aspects of NP toxicity such as membrane damage, cell death, apoptosis, inflammation and oxidative stress were established in the partner labs and adapted for use in the interconnected fluidic ILT setup. Furthermore, baseline values for toxicity for the different ILT NPs and individual cell lines under static conditions were determined. The different NPs were tested up to concentrations of 625 µg/cm2 except for the gold NPs which were tested up to concentrations of 40 µg/cm2.
For all three cell lines, NM300 Ag NP were found to be highly toxic at very low concentrations. NM101 TiO2 NP demonstrated strong pro-inflammatory activity in HUVEC cells and to a limited degree also in C3A cells. There is some indication that very high concentrations of 55 nm and 211 nm fluoresbrite polystyrene NP induce toxicity in C3A and HUVEC cells as shown in the Alamar blue and LDH assay. For Caco-2 cells there seems to be slight oxidative stress reaction induced by Au 15 nm NP.
In general, C3A cells and HUVEC were more sensitive to the potential toxic effects of NP than Caco-2 cells, as they showed lower LC50 values and gave slightly positive readouts for the polystyrene NP at high concentrations and for inflammation induced by TiO2.
The LDH assay and measurement of apoptosis via FAS ligand and of inflammation via IL-8 release seem most suitable for routine application in the ILT setup as all three biomarkers can be measured from the cell supernatant with no need to remove the cells from the system. This allows online monitoring. Furthermore, the measured LC50 values were about 5-10 fold lower in the LDH assay compared to Alamar blue assay, indicating higher sensitivity. IL-8 release mirrored the cytokine expression at mRNA level.
For the healthy intestine, there are several more or less complex multi-cellular in vitro models of the intestinal mucosa described in literature and patents. In most cases, these are based on permanent intestinal epithelial cell lines such as Caco-2 or HT-29 However, only one of these models (combining Caco-2 cells with M-cell like Raji cells) has been used to study NP interactions focusing on uptake of MPs across healthy intestinal mucosa without assessment of local inflammatory or cytotoxic reactions. As in INLIVETOX we also wanted to address engineered nanomaterial interactions with the susceptible (i.e. inflamed) intestinal barrier, we introduced an inflamed intestinal model for toxicity testing. Previously, a 3D co-culture model based on Caco-2 enterocytes and primary immune cells had been established by one of the partners. It was decided to adapt the model to for higher throughput screening and also for easier lab-to-lab transfer by replacing, primary blood-derived macrophages and dendritic cells by permanent cell lines.
The cell lines THP-1 and U937 were evaluated to replace the primary blood-derived macrophages while MUTZ-3 cells were targeted to replace the dendritic cells. Studies on activation of THP-1 and U937 cells identified THP-1 cells as the only suitable candidate. As a first step a Caco-2/THP-1 double co-culture was established. In a second step, MUTZ-3 cells were integrated to form a triple culture of enterocytes and dendritic and macrophage-like cells. The modified model had a comparable performance to the original triple culture both in the non-inflamed state and upon stimulation to an inflamed state with pro-inflammatory cytokine IL-1. In the inflamed state, barrier properties were reduced and an increased release of pro-inflammatory marker IL-8 was observed
WP 3: Implementation of biological models in the ILT system
This WP focussed on taking the cell culture models from WP 2 and establishing them in the fluidics system(s) of WP 1 to provide the starting point for the validation work of WP 5.
The first task in the work package was to establish the ILT1 system at the different partners' institutes. After the training course for ILT1 had been carried out, USAAR and UNIPI were provided with complete ILT1 systems and started working with them in their own laboratories. After some weeks of work by the consortium using ILT1, feedback from the different users was analysed and was used to define the requirements for the second version bioreactor, ILT2 that was developed in WP 1.
It was originally planned to establish the 3-tissue culture model - with all three model tissue maintained in the same fluidics system - using ILT1 and ILT0 at this point. However, this was not feasible due to time constraints. Instead, the individual tissue models were set up in ILT0 (for the C3A and HUVECs) or ILT1 bioreactors and the viability and function of the models under the common culture conditions was tested.
However, once the ILT2 bioreactor had been completed, the three-tissue culture model was established. As described in previous WPs, the models established were: for the intestinal epithelium CaCo-2; for the vascular endothelium, HUVEC, for the liver, C3A.
One tissue model that was not established in the ILT system was the triple culture model of the inflamed intestine. This model (described in WP 2) was based on a collagen layer deposed on the surface of the microporous cell culture support. Unfortunately, the collagen layer effectively blocked the transport of NPs from apical to basolateral compartments, making the model unsuitable for study of NP transport. In addition, the collagen layer adhered poorly to the silicon nitride microporous cell culture support used in the ILT system. This did not pose a problem under static culture conditions. However, under flow conditions, the model intestinal barrier peeled off the support so that the separation of the two (apical and basolateral) flow circuits was not maintained. For these reasons, it was decided to abandon the establishment of this model and to focus efforts on the remaining tasks in the project.
Finally, at the end of WP 3 a system baseline was established in the absence of NPs. The baseline provided a negative and a positive (0.1 %Triton-X 100) control for the subsequent experiments in WP 5. The baseline included assays for viability, inflammation, apoptosis and function.
WP4: In vivo assays
The animal studies of INLIVETOX used rodent models to determine the quantitative biokinetics of intravenously injected and orally administered radio-labelled NM (TiO2, Au and Ag) with a special emphasis on cardiovascular and liver uptake since these two targets relate to the targets included in the INLIVETOX model. Tissues from these animals were used to investigate toxicological responses, again focusing on the gut, cardiovascular system and liver. This data was then compared with other biokinetic studies using similar particles, but other routes of exposure (e.g. respiratory route).
Oral gavage biokinetic studies:
- After oral administration to rats, only a small fraction of all of the NM administered were absorbed through the gastro-intestinal-tract (GIT) wall into blood circulation for subsequent accumulation in organs and tissues. The major fraction of the ingested dose for all NM administered were eliminated from GIT into the faeces.
- The 24-hours biokinetics study of the orally delivered monodisperse, different sized (1.4 2.8 5, 18, 80, 200 nm) gold nanoparticles (AuNP) showed a clear, inverse size-dependent absorption through the gut wall into blood circulation (0.4 % of administered 1.4 nm AuNP to 0.02 % of 80 nm AuNP). This led to low but detectable accumulations in organs in the following decreasing order: kidneys, liver, spleen, lungs, heart, brain. Interestingly, AuNP also accumulated in the skeleton. This pattern differed strongly from the pattern after intravenous injection of the same AuNP indicating different surface modifications according to the route of entry which may relate to the formation of the protein corona.
- Similar to AuNP most of the orally administered, agglomerated titanium dioxide NP (TiO2 NP) of 70 nm size were also excreted in faeces. Compared to 80 nm AuNP the GIT absorption of TiO2 NP was five times higher indicating NP material differences. Accordingly, the quantities of TiO2 NP in organs and skeleton were higher but ranked similarly as for 80 nm AuNP. Absorption through the gut into blood circulation increased during the first 24 hours but decreased during the next six days indicating NP clearance from the body.
- Silver NPs (AgNP) were not used for biokinetics studies due to enhanced physico-chemical reactions during the process of radiolabelling resulting in dangerous difficulties in their handling.
Intravenous injection biokinetic studies
- All NM administered intravenously were almost completely retained in the body and accumulated in organs and tissues.
- Intravenous injection of NP suspensions predominantly resulted in accumulation in the liver, justifying the choice of hepatocytes in the INLIVETOX model.
- These studies allowed for kinetic measurements of the hepato-biliary clearance of NP into the small intestine and further into faeces. The monodisperse, different sized (1.4 2.8 5, 18, 80, 200 nm) AuNP showed a clear inverse size dependency of this hepatic clearance pathway. Clearance of the 70 nm TiO2 NP was similar to that of 80 nm AuNP but increased linearly with time up to 2 % of the initial NP dose over one month indicating a persistent but moderate NP clearance pathway out of the liver.
- Given the physico-chemical reaction during the process of radiolabelling the silver NP were not used for biokinetics studies. Instead intravenous injected silver ions were used. This is relevant for Ag NP toxicity due to the potential for AgNP to dissolve. Less than 5 % of the silver ions were found to accumulate in organs and the skeleton. However, they were not excreted in urine as expected but predominantly in faeces indicating effective translocation from blood into the gut lumen. This was not seen for the metal ions of the other NP.
In addition to biokinetics studies, in vivo toxicology studies were carried out. The data from these studies was compared to data from the static in vitro studies of WP2 and to the in vitro studies with the INLIVETOX system in WP 5.
Oral gavage toxicology studies
- A depletion of reduced glutathione (GSH) levels in the liver was used as a marker of oxidative stress. No significant effects were observed for intestinal or hepatic oxidative stress for any NP administered by gavage.
- Similarly, no significant changes in GSH levels were observed in Peyer's patches after gavage of Au NPs of 15 and 80 nm.
- However, there was a significant increase in GSH in Peyer's patches after gavage of 250 µ g TiO2 and Ag per animal indicating oxidative stress in this target tissue for these two particle types.
Intravenous injection toxicology studies
- Of the injected particles (Ag, 15 nm Au and TiO2), only the TiO2 particles caused a decrease in reduced GSH content of livers. No significant changes in GSH levels were observed in Peyer's patches or intestines after injection of any of the particles used.
Inflammation
Exposure of rats to NPs by both injection and ingestion resulted in changes in the expression of genes involved in inflammation, oxidative stress and apoptosis in the target organs examined (liver, Peyer's patches, intestine and aorta). Some of the most interesting results were the following:
- Changes in gene expression were not limited to the primary target organs (i.e. livers / aorta for injection and intestine / Peyer's patches for ingestion), but systemic changes were also observed.
- Changes in gene expression were dose-dependent, and a particularly interesting difference was observed in Peyer's patches, where small doses of 15 nm Au particles (50 µg per animal) caused an increase in the expression of a number of genes, but higher doses (250 µg per animal) caused a strong decrease in gene expression.
Overall, 80 nm Au particles caused the fewest changes in gene expression, and when compared to 15 nm Au particles, there was a pronounced size effect for the Au NPs. When examining H&E-stained sections of livers, intestine, Peyer's patches and aorta, no obvious effects such as strong influx of inflammatory cells or tissue damage could be observed.
No NPs were detected in the Peyer's patches of rats exposed to the range of NPs used in the in vivo exposures (gavages) using TEM.
Ag, TiO2 and 15 nm Au particles were detected by TEM in livers of animals exposed via the tail vein. Ag and TiO2 NPs were detected in both hepatocytes and Kupffer cells, and some had reached the nuclei of hepatocytes. The particles appeared in small clusters and free within the cytoplasm / nuclei. In contrast, 15 nm Au particles were only detected in the cytoplasm of Kupffer cells, and were without exception enclosed within membrane-bound vesicles.
WP 5: Validation of the system and analysis of NP toxicity by ingestion
The first task in WP 5 was to produce and distribute a protocol handbook for the INLIVETOX system, in particular, for assessing NP toxicity. This handbook was produced and has been shared among the ILT partners. It includes:
- an introduction;
- instructions for setup, use and cleaning;
- cell culture, common media and transfer of cells into the ILT system;
- particle suspension preparation and exposure of cells;
- protocols for a number of toxicological endpoints;
- protocols for cell-specific functional tests.
The main goal of the WP was to characterise the fate and behaviour of NPs within the completed INLIVETOX model and the impact of the particles on the cells in the system and to compare this with in vivo data. This was delayed because of the additional time needed to finalise the ILT system and deal with some problems as highlighted in WP 3. Experiments were carried out using Ag NPs, which were selected because they had the highest toxicity to all three cell types used (see WP 2).
A range of concentrations of NM300 NPs (from 0.05 to 50 µg/ml) was applied to the ILT system. This range included a dose which was non-toxic in static experiments, a dose with low toxicity and one with high toxicity above the LC50. Effects of the NPs on cell-specific markers such as TEER and albumin release could be measured, as well as cytotoxicity and inflammatory mediators in the medium.
Ag NP induced a dose dependent decrease in viability of the three cell types in the ILT system with the lowest concentration (0.05 µg/ml) inducing no significant effect. A slightly higher concentration of 0.5 µg/ml induced a moderate decrease in TEER (by about 15 %), but significant LDH and reduced HUVEC viability. This toxicity was also associated with an increase in the production of the pro-inflammatory mediator IL8. At the higher Ag NP concentrations of 5 and 50 µg/ml epithelial barrier function was greatly affected as shown by TEER measurements. The loss of barrier integrity of the Caco2 cells was associated with greater downstream tissues effects such as cell death in HUVEC and reduced albumin production in C3A cells. High levels of LDH and inflammatory proteins could be measured in both circuits.
Comparisons were made between in vivo responses to NP exposure by injection and the responses of static single cell in vitro models.
Remarkable similarities were observed between the static in vitro models and the in vivo data. These similarities were observed especially for Ag and TiO2 NPs for:
- intracellular particle uptake;
- oxidative stress (measured by depletion of reduced glutathione);
- gene expression of inflammatory markers, particularly IL-8/MIP-2 (the rat homologue of IL-8), TNF and IL-1RI.
Since particle accumulation mostly occurs in the liver, studies focussed on this organ. Both high toxicity Ag and low-toxicity TiO2 NPs were detected in cells in the liver after injection and were taken up on exposure in vitro. However, the responses observed for the two NPs were very different. No oxidative stress was observed for Ag NP either in vitro or in vivo. However, IL-8/MIP-2, TNF and IL-1RI expression were all increased both in vitro and in vivo. In contrast, TiO2 NP did cause oxidative stress in vivo and at high doses in vitro. TNF and IL-8 expression were increased after 24 hours both in vivo and in vitro, while IL-1RI was relatively unchanged.
A calculation of doses delivered to the liver in vivo and hepatocytes in vitro indicated that they are comparable. The similarities observed suggest that simple in vitro models using human cell lines compare very well with intravenous in vivo rat studies.
In vivo data on NP exposure by ingestion was compared with the ILT model. The biokinetics studies showed that in the ingestion model, only a small number of particles (maximum 0.5 % for 1.4 nm Au particles) translocated from the gastrointestinal tract to the blood and secondary target organs. This is in strong contrast to the injection model, where rapid and near-quantitative accumulation of particles in the liver occurred. This makes in vitro modelling of ingestion with simple models very difficult, because interactions of particles with the food, translocation through the GIT and distribution, as well as interactions with cells along the way and the communication of these cells with each other and on a systemic level have to be taken into account.
This situation was clearly reflected in the much more complex in vivo tissue response to particle exposure via intraoesophagial gavage as seen for gene expression. This type of integrated and complex response was observed in the ILT model in which:
i) very low translocation of PS-FITC to the basal compartment was nonetheless accompanied by substantial NP uptake in the target C3A and HUVEC cells; and
ii) Ag-induced toxicity and inflammation was observed at very low concentrations and translocation.
In the in vivo ingestion studies, Ag had no significant impact on cell death (as measured by gene expression markers of apoptosis) or pro-inflammatory gene expression (including IL8). In contrast, the Peyer's patches were very sensitive to the Ag NP with increases in some pro-inflammatory genes (e.g. TNF-alpha and MCP-1), decreases in others (e.g. IL1-beta) but no significant changes in markers of apoptosis. When compared with the results of the ILT system (see above), this suggests that the ILT system did not reflect well the downstream responses to Ag NP of tissues such as the endothelium and liver following exposure via ingestion. The ILT system response to Ag NPs therefore reflects the static in vitro model responses rather that the in vivo ingestion model responses. The ILT response also reflects the in vivo liver responses following injection of Ag NPs.
Data from the single cell studies, inflamed cultures and three-tissue experiments under flow conditions were analysed by three-way principle components analysis assessing a combination of variables (subjects), objects (response variables) and conditions. This study is the first application of three-way PCA to toxicity data.
In summary, regarding viability, it was concluded that C3A cells are the most sensitive to NP exposure, followed closely by Huvec and then Caco-2 and the inflamed model. Inflammation was higher for TiO2 than for Ag NPs. The inflamed model gave the highest response, and C3A were the most sensitive cells, followed by Huvec and Caco-2. Viability of cells in the static and dynamic 3 tissue model was also examined. Under static conditions, there was a clear difference in susceptibility of the different cell types. However, under flow conditions with crosstalk between cells, there was overlap between all three cell types, highlighting the fact that the three tissue model under flow conditions is more than just the sum of the three single cell types, and represents interactions between the cells which cannot be modelled in a single cell system or without medium flow.
Overall, based on silver data, cells in the system are more susceptible to NP exposure when under flow than as single cells in static conditions. This indicates once more the importance of a co-culture system which enables communication between the different tissue models.
Finally, nanoparticle toxicity in the inflamed intestinal model was studied. The results were compared to the non-inflamed triple culture model and the Caco-2 monoculture.
Testing in the non-inflamed and inflamed co-culture model of the intestinal mucosa changed the toxicity profile of the engineered nanoparticles investigated in the INLIVETOX project. TiO2 nanoparticles showed no toxic or inflammatory effect in any of the models. In contrast, the Au NPs had no significant toxicity but induced strong inflammation in both triple culture setups, which couldn't be observed in the Caco-2 monoculture. Differences between the three cell culture test systems were most pronounced for Ag NP. Toxicity was observed in both mono- and co-cultures, but the LC 50 values were markedly lower in the triple culture setups compared to Caco-2 monoculture. Furthermore, significant inflammation was observed in the triple culture models while being much weaker in the monoculture.
In general the co-culture model more easily picked up pro-inflammatory potential of NP. The innate immune cells have a protective effect, preferentially taking up NPs, reducing epithelial exposure and inducing an immune response to a perceived threat. In the inflamed model, this protective function is impaired. In summary, the triple culture setups provided valuable data for interpretation and correlation of in vivo results and mechanistic understanding of NP interaction with the gut mucosa. This is an important step to in vitro modelling of not only the healthy, but also the diseased gastrointestinal tract.
In conclusion, the developmental and experimental work has resulted in a working system for NP exposures with a set of instructions which ensures reproducibility of experimental procedures among ILT users. The baseline created in WP 3 has been expanded upon with further experiments including a range of doses of high-toxicity Ag NPs, which show effects on all three types of cells. These effects vary between cell types and are dose-dependent, and differ from those observed in single cell or static cultures.
Comparisons of in vivo data with data generated by simple static in vitro models and by the ILT system showed that simple in vitro models can give surprisingly good correlations with in vivo studies in some cases. Here, simple static cultures of C3A cells showed striking similarities to the in vivo data for studies based on injection. In contrast, no such similarities were found between the equivalent static cultures, the ILT system and the in vivo studies of Ag NP induced nanotoxicity by ingestion. Ag was chosen because of its relatively high toxicity observed in the static in vitro models used in the project. It would be useful to test the ILT response to the lower toxicity material Au to see whether the results induced are comparable to the in vivo responses observed following ingestion. Although Au was lower in toxicity, it induced a wider array of changes in gene expression of pro-inflammatory, anti-inflammatory, antioxidant defence and apoptosis genes than Ag allowing for a more in depth comparison. It would also be useful to compare the results of the ILT system with injection responses in vivo for a wider array of NP and to compare these with static cultures to assess whether the additional complexity of the ILT system provides any advantages over the cheaper and easier static models for predicting responses to intravenous injection.
Potential impact:
Socio-economic impact can occur at the micro level (companies or partners using the project results) at the meso level (specific industry sectors or geographic regions benefitting) or at the macro level, where benefits are experienced across countries or the whole EU. The INLIVETOX project will have impact at all these levels.
At a micro level, the project has delivered an exciting and innovative technology that has the potential to underpin new product developments in the field of in vitro testing. This has created (and will continue to create) commercial opportunities for some project partners and for those organisations that licence the technology from the partners - as demonstrated by the very high level of interest observed from potential end users in both industry and research organisations. The very high quality of the project work and the results obtained will further enhance the reputations of the research organisations that participated and strengthen their capability to take a leading role in future internationally leading research in the field of in vitro testing and toxicity.
At a meso level, the results of the INLIVETOX project have the potential to change the way that the pharmaceutical, chemical, cosmetic and food sectors of industry are testing the safety and efficiency of new materials. The improved methods could deliver significant economic benefits both through reduction of testing costs compared to the use of animals, but also through the opportunity to bring safer products to market faster than existing methods. The technology developed in the project could provide a significant competitive advantage to the early adopters.
At a macro level, the project confirms the internationally competitive position that Europe's research organisations hold in the fast developing field of in vitro testing. A report by the United States (US) National Research Council in 2007 entitled 'Toxicity testing in the 21st century' outlined the scale of the challenges and suggested a 15 year time scale for the replacement of many animal testing methods by in-vitro techniques.
The WYSS Institute in Boston, US, has recently been granted USD 26 million for a project to develop a 'lab-on-a-chip' solution to toxicity testing. The capability demonstrated by the INLIVETOX project shows that EU researchers could be leaders in this field, if they are able to secure similar levels of funding.
The motivation for such an investment is the size of the existing (GBP 1.8 billion) market for drug toxicity testing and also the emerging (GBP 2 billion) market for the testing of chemicals to comply with REACH legislation.
Any change in methodology for the testing of new drugs will require regulatory approval. The regulatory bodies (ECVAM, FDA etc.) are justifiably cautious in approving new methods. However, the cornerstone to any change is sound science and the work of the INLIVETOX team has demonstrated the capability of European research organisations to make significant progress.
To summarise the potential impact of the INLIVETOX project:
- Direct commercial benefit to the partners or their licensees, opening up new commercial product opportunities that could reach revenues of (EUR 7 million) within 5 years.
- Potential annual savings for the pharmaceutical and chemical sector industries using the technology measured in 10's of millions in the short-term (5 years) and 100's of millions in the longer term (10 years).
- Potential safer drugs and reduced risk in chemical trials through the reduction in false positives or false negatives in toxicity testing.
- Significant reductions in the number of animals used for testing of new drugs and chemicals.
Main dissemination activities
A project identity set consisting of project logo, leaflet and all associated templates was created at the start of the project and distributed at many conferences and exhibitions throughout the duration of the project.
INLIVETOX website (please see http://www.inlivetox.eu/(öffnet in neuem Fenster) online) was created by ALMA and updated with data from all the partners, and will be maintained by Alma Consulting for a period of five years following the completion of the project.
Three press releases were made between June 2010 and the end of the project and articles or interviews appeared in Edinburgh Napier News, Radio Suisse Romande, l'Express / l'Impartial, Le Temps, La Gruyère, Migros magazine, The Scotsman, BBC news and Lepoint.fr.
Two workshops and seminars were delivered in Ispra and Saarbrücken in 2010 and 2011.
Specific training courses are planned by Pisa University and Kirkstall to disseminate the methods developed during the project. The first of these was held on 17 and18 September 2012 in Rome. It is expected that these courses can be self financing and so can continue at regular intervals during 2012 and 2013.
Papers were presented by the project partners at 13 conferences in Rome, London, Lausanne (2), Saarbrücken (2), Montpellier, Edinburgh, Grenoble, Krakov, York, Essen and Utrecht.
The project results were exhibited at three major trade shows in Paris, Manchester and Munich as well as several smaller events.
The following scientific papers have been published already or will be published in the near future:
- Quality control in in vitro nanoparticle testing: Collagen coating can optimise particle distribution on adherent cell lines and improve data quality
B. Gaiser
'Modern Polymeric Materials for Environmental Applications', Vol. 4, Issue 1, 2010, Ed. K. Pielichowski, 978-83-930641-1-3
- A complementary definition of nanomaterial
W. Kreyling
Nanotoday, Vol. 5, Issue 3, June 2010, pp. 165 -168, Elsevier, 10.1016/j.nantod.2010.03.004
- Generation and characterisation of stable, highly concentrated titanium dioxide nanoparticle aerosols for rodent inhalation studies
W. Kreyling
Journal of Nanoparticle Research, 2011, Vol. 13 No. 2, pp. 511-524, Springer, 10.1007/s11051-010-0081-5
- Engineering quasi-vivo in vitro organ models
A. Ahluwalia
Adv Exp Med Biol. 2012; 745:138-53, Springer, 10.1007/978-1-4614-3055-1_9
- Size and surface charge of gold nanoparticles determine absorption across intestinal barriers and accumulation in secondary target organs after oral administration
Nanotoxicology, February 2012, Vol. 6, No. 1, pp. 36-46, doi:10.3109/17435390.2011.552811)
http://informahealthcare.com/doi/abs/10.3109/17435390.2011.552811(öffnet in neuem Fenster)
- Particle size-dependent and surface charge-dependent biodistribution of gold nanoparticles after intravenous administration
Hirn S.
Eur J Pharm Biopharm. 2011 Apr;Vol. 77 No. 3, pp. 407-16, Elsevie
http://www.sciencedirect.com/science/article/pii/S093964111000370X(öffnet in neuem Fenster)
- Using in vitro models to assess the toxicity of nanomaterials
Stone V.
In Vitro Toxicology Society
Three chapters in a book entitled 'New developments in cell-based in-vitro testing - methods and protocols' have been written by INLIVETOX partners. The book is available online in June 2012 by Pan Stanford Publishing (please see http://www.panstanford.com(öffnet in neuem Fenster) online).
Direct mailshots of a newsletter were made to 24 industrial companies who were potential users and 2 organisations involved in regulatory matters (ECVAM and LGC).
Exploitation of results and commercialisation strategy
The overall objective of the strategy is to maximise the uptake of the project results. The plan accommodates both those beneficiaries who wish to be involved directly in commercial activities and those who wish to licence others to do this on their behalf. The plan takes account of regulatory aspects and ethical issues.
There are two main sectors that are being targeted in the commercialisation plan: the academic research community working on advanced in-vitro models of cell culture and the commercial companies involved in the testing of safety or efficacy of compounds for the pharmaceutical, cosmetic household products or food industry.
In the commercialisation plan careful consideration was given as to whether to publish, licence to others or develop commercially ourselves. The latter two options are viable because steps have been taken to protect the IP: one patent has been applied for on the final ILT2 holder. Copyright designs and secrecy are further options still available to the consortium.
Products and services
The following products and services that build on the IPR created during the project will be commercialised by the partners:
- protocols to set up cell cultures to create viable biological tissue models of the intestine and other barriers within the body (service);
- protocols to determine Cytotox and inflammatory responses (service);
- data on biokinetics from in vivo studies (service);
- modular microfluidics based in vitro test system 'INLIVETOX system' including necessary protocols of how to use it in various applications (product);
- improved microfabricated membranes for cell culture (product).
Routes to market
The service offers described above will be marketed direct by the organisations involved. For the products a different approach is being taken by Kirkstall, i.e. initial market creation in the local market (UK) by direct sales. This is followed by selection of distributors to serve regional markets.
Competition
The market for products and services in the field of in-vitro testing of safety and efficacy for drugs, cosmetics and chemicals is very large and hence there are many competitors already and many more emerging. The leading players have invested heavily in development of their offers and then followed by perhaps even greater investment in getting their tests validated and approved by the regulatory bodies such as ECVAM.
The INLIVETOX consortium members will be seeking further investment and partnership with major global companies to ensure the maximum take up of the technologies developed during the project.
The service offers could be available before the end of 2012. The product offers are expected to take one to two years further development subsequent to the end of the project.
Website: http://www.inlivetox.eu(öffnet in neuem Fenster)
Coordinator:
Centre Suisse d'Electronique et de Microtechnique SA (CSEM), Switzerland
Dr Martha Liley
E-mail: Martha.LILEY@csem.ch
During the project, a novel in vitro model of the GI tract, vascular endothelium and liver was developed (the ILT system) for in vitro studies of nanotoxicology by ingestion. CaCo-2 cells were used to model the intestinal epithelium, HUVECs the vascular endothelium and C3As the liver. Each model tissue was maintained in a separate bioreactor. The model included two fluidics circuits: one circuit modeled the GI tract and the other - the model blood stream - connected all three tissue models. A new bioreactor (ILT2) was developed during the project to house the intestinal epithelium. Both fluidics circuits flowed through ILT2, separated by the model intestinal epithelium.
A major effort in the project was focused on reference measurements and the establishment of protocols for the ILT system. The physicochemical properties of the NPs to be used - gold (Au), silver (Ag) titanium dioxide (TiO2) and polystyrene (PS) - were characterised. Common culture conditions were established for the three cell types in the ILT system. Protocols for assays for cell viability and function were defined. Assays were also established for markers of oxidative stress, apoptosis and inflammation. Finally, baseline values for toxicity for the different NPs and individual cell lines under static conditions were determined.
A model of the inflamed intestinal epithelium based on cell lines (CaCo-2, THP-1 and MUTZ-3) was also established. The model was not suitable for the ILT system. However, comparative studies of NP toxicity using this model and simple CaCo-2 layers demonstrated a protective effect of the immune cells in triple co-culture. This protective effect was impaired when the co-culture model was inflamed. These models represent an important step towards in vitro modelling of, not only the healthy, but also the diseased gastrointestinal tract in nanotoxicology.
In vivo experiments carried out using rats, provided comparative data on NP exposure by both ingestion and injection. Biokinetics experiments provided data on the distribution and fate of Au and TiO2 NPs. Not on Ag NPs, however, due to severe problems with their labeling. Selected tissues - the intestine, Peyers patches (PP) aorta and liver - were also analysed for NP uptake, oxidative stress and changes in gene expression for a panel of genes involved in inflammation, oxidative stress and apoptosis.
Comparison of the data obtained in vivo on exposure by injection and ingestion with data obtained from static, single cell in vitro assays and the ILT system showed a remarkable pattern of differences and similarities. In vivo exposure to Ag and TiO2 NPs by intravenous injection gave strikingly similar gene expression responses to those obtained using static in vitro cultures of C3A, suggesting that simple in vitro models using human cell lines can compare very well with intravenous in vivo rodent studies. This result is particularly noteworthy since hepatocyte cell lines are often regarded as sub-optimal for toxicology testing due to their lack of expression of metabolising enzymes.
Modelling nanotoxicity by ingestion is much more complex than by injection, as interactions of particles with the food, translocation through the GIT and distribution, as well as interactions with cells along the way and the communication of these cells with each other and on a systemic level have to be taken into account. Poor correlation was obtained between in vivo and static single-cell in vitro experiments. In vitro data on nanotoxicity by ingestion was also obtained for Ag NPs using the ILT system. In the ILT model of toxicity by ingestion, Ag NPs caused a dose dependent decrease in viability of the three cell types. Dose-dependent losses of cell function (epithelial barrier function, production of albumin) and increases in inflammatory markers were also observed. These results compared poorly with the in vivo data on Ag NP toxicity by ingestion, correlating much better with Ag NP toxicity by injection.
It would be useful to test the ILT response to the lower toxicity Au NPs and compare it with the in vivo responses to ingestion. Although Au was lower in toxicity, it induced a wider array of changes in gene expression than Ag allowing for a more in depth comparison. It would also be useful to compare the results of the ILT system with injection responses in vivo for a wider array of NP and to compare these with static cultures to assess whether the additional complexity of the ILT system provides any advantages over the cheaper and easier static models for predicting responses to intravenous injection.
The INLIVETOX system developed in this project has potentially much wider application than just the testing of responses to NPs. Throughout the project there has been extensive dialogue with members of an industrial advisory group from the pharmaceutical, cosmetics, food and household products manufacturing industries. Outputs from the project include tested demonstrators of a cell culture bioreactor that will be commercialised by partners from the consortium.
Project context and objectives:
Engineered NPs (also called manufactured nanomaterials) can be considered a new class of substance, since chemical and physical properties at the nano-scale can differ hugely from those of bulk materials. Typical examples include silver particles used as an antibacterial and antifungal agent in water bottles, zinc oxide used to block UVA and UVB sun rays in sunscreen, and the various materials used as pigments in paints and cosmetics. The novel properties of NPs - that make them so attractive for many applications - means that they can also interact with the body in different and unexpected ways, and their small size allows them to potentially cross barriers in the body that hold back larger materials. For this reason, a rapid and thorough evaluation of NP toxicity is essential.
Given the large numbers of NPs that exist in the lab and sometimes in commercial products, researchers and industry are faced with the almost impossible task of analysing the safety of thousands of new materials in a very short time frame. It has become clear that, from both ethical and financial viewpoints, new animal-free testing methods are necessary. Unfortunately, the animal-free cell culture or 'in vitro' models currently in use are often not representative of the human body. They simplify enough to allow the production of data, but often so much that the data is not representative and cannot replace most testing on animals. Most in vitro tests work with one type of cell in culture, but this is, of course, not the way cells work in our body, where the surrounding environment, including other cell types, has a strong influence on how cells behave. The interactions between different organs, or the transport of a material through barriers, for example from the intestine into the bloodstream, cannot yet be modelled in a realistic, or 'physiological'. way.
The EU-sponsored project INLIVETOX (please see http://www.inlivetox.eu(öffnet in neuem Fenster) online) has brought together engineers and biologists from universities and research institutes across Europe in order to develop an improved cell culture system. This model will help us gain insight into how NPs taken up in food or otherwise ingested can cross the intestinal wall and how they can affect the different tissues in the body.
The global objectives of INLIVETOX were:
- to improve understanding of NP toxicity by ingestion;
- to develop improved in vitro toxicity testing methods.
In order to achieve these objectives, the project aimed to develop a modular microfluidics-based in vitro test system, the ILT system (WP 1); a system in which different cell types - each modelling a target tissue - could be maintained and interact via cytokines. The target tissues chosen for this project were the GI tract, the liver and the vascular endothelium so that NP toxicity by ingestion could be modelled with the system.
Subsidiary objectives in the project were the establishment of common culture conditions for the cells and test protocols for the analyses to be carried out with the ILT system (WP 2). Once established, these were then to be implemented in the ILT system (WP 3).
Validation data were to be obtained via an in vivo study of NP toxicity by ingestion in rats carried out in parallel. Data on both biokinetics and toxic response from the target tissues were the goal of WP 4.
The final validation, bringing together in vivo data with new data generated by the ILT system, was the objective of WP 5.
The long-term objectives of the project are:
- to ensure the safe development and use of NMs for commercial applications;
- to commercialise a test system to screen NMs for their toxicity.
These two objectives are related to the foreseen impacts of the project, with future commercialisation of the INLIVETOX system making it available to the whole toxicology community.
Project results:
WP 1: Development of microfluidic systems
WP 1 was the engineering and design WP of the project, and also involved mathematical modelling. The underlying concept of the project is that of connected cultures in which two or more tissues are connected together by a fluid flowing between them, much as in the human body in which distant tissues and organs are connected by the bloodstream. Our aim was to engineer a connected culture fluidic system in which nanoparticles could pass through an intestinal barrier and then influence downstream tissues to simulate the route of ingested nanoparticles in the body. It was decided early on in the project to build a modular fluidic system so that different tissues could be added or removed from the fluidic circuit as required. We also choose to use cell culture systems which were easily translatable to biological experimental standards.
The first part of the WP was dedicated to the scaling of the INLIVETOX system using a method known as allometry. This method is based on the well-known fact that the different features of animals are correlated to their body mass. Using allometry we calculated the liver, vascular tissue and intestinal cell numbers necessary to properly represent the human body in the reduced scale of our system. At the same time, we also considered other parameters such as flow, oxygen and fluid forces. Having identified the cell numbers and other parameters, we proceeded with the design of the system. During this period, the engineering partners trained the other partners in the consortium in fluidic culture methods using a bioreactor chamber known as ILT0. ILT0 is an interconnectable commercial cell culture chamber designed and commercialised respectively by two partners in the team.
The main challenge in the WP was the design of the intestinal chamber which requires a semi-permeable membrane to separate the intestinal side from the blood side and two different fluid flows on either side. Two versions of the chamber were developed, ILT1 and the much improved second version, ILT2.
Challenges identified were:
i) the need to continuously monitor the function of the intestinal cells using a technique known as TEER (transepithelial electrical resistance). TEER is a standard and sensitive method of analysing the integrity of epithelial barriers such as the intestinal wall.
ii) the need to use ultrathin microporous membranes to allow NPs to pass from the intestinal side to the blood side. Most commercially available microporous membranes rapidly become clogged when exposed to NPs.
To meet these challenges an ultrathin silicon nitride membrane was fabricated with built in electrodes for TEER measurements. The first bioreactor chamber ILT1 was designed to hold the membrane and had electrical contacts for TEER measurements. After the first few prototypes were fabricated and tested, a training course was held to introduce partners to the new system. The cell culture partners then started using the ILT1 to perform baseline measurements on intestinal epithelia.
A few months later enough feedback had been collected to begin the design of ILT2. The most fundamental problem was a difficulty in seeding cells on the microfabricated membranes and then placing the membranes in ILT1. Other problems concerned leakage, air bubbles and difficulty with stable TEER measurements. ILT2 was designed using innovative solutions (some were patented) to overcome these problems. It was also more robust and the TEER measurements more stable due to new watertight contacts and more robust electronics.
The final part of the WP was dedicated to training of other partners, modelling NP passage through the membrane and NP sedimentation in the fluidic system and troubleshooting when technical hitches arose during cell culture.
WP 2: In vitro biological models
Laying the ground work for all further in vitro studies in the INLIVETOX project, WP 2 established protocols to ensure reproducible and stable size distributions in the NP suspensions as well as defining assays and endpoints to quantify cell viability and toxic responses in the INLIVETOX system. Baseline data under static conditions was generated for comparison in later stages of the project. Furthermore, the in vitro models used in the project were optimised with regards to culture conditions, selection of functional markers and suitability for higher throughput applications (e.g. an existing three-dimensional (3D) model of the inflamed intestinal mucosa was modified to improve reproducibility).
As the INLIVETOX system is based on studying the cross talk between three different tissue models connected to each other via a fluidics systems, common culture conditions had to be identified for the three cell lines Caco-2 (enterocytes), C3A (hepatocytes), and HUVEC (endothelial cells). A common medium was developed based on EMEM. This medium, referred to as EMEM-GF, affected neither cell adhesion nor morphology either in static or dynamic conditions. Under the chosen conditions, all cells maintained their viability as well as their tissue-specific phenotype and were able to withstand the shear stress of dynamic conditions. Albumin expression was used as the phenotypic endpoint for C3A hepatocytes while expression of von Willebrand factor was the marker for endothelial cells and barrier function (quantified via transepithelial electrical resistance) characterised the functionality of Caco-2 cells.
In order to interpret the toxicity of nanoparticles and understand the underlying mechanism of potential toxic effects, the physicochemical characteristics of the NPs need to be known. Previous studies have correlated toxicity of various nanomaterials to parameters such as surface area, surface reactivity or phase composition. Furthermore, it was shown that the dispersion medium used can greatly influence the physiochemical properties of NPs. Therefore, we characterised the NPs to be used in the INLIVETOX system with regards to their size, polydispersity, aggregation behaviour and surface charge in various media to be used for in vivo and in vitro investigations. Most nanomaterials were found to form stable dispersions in EMEM-GF with particle sizes close to the nominal size. Only NM101 TiO2 NP showed significant aggregation behaviour, forming 700 nm sized clusters (primary particle size 7-11 nm). Serum proteins sterically enhanced the stability of the NP dispersions as aggregation behaviour was more pronounced in serum free conditions. Dispersions were stable for at least 48 h, allowing their use in the INLIVETOX system.
Different assays evaluating various aspects of NP toxicity such as membrane damage, cell death, apoptosis, inflammation and oxidative stress were established in the partner labs and adapted for use in the interconnected fluidic ILT setup. Furthermore, baseline values for toxicity for the different ILT NPs and individual cell lines under static conditions were determined. The different NPs were tested up to concentrations of 625 µg/cm2 except for the gold NPs which were tested up to concentrations of 40 µg/cm2.
For all three cell lines, NM300 Ag NP were found to be highly toxic at very low concentrations. NM101 TiO2 NP demonstrated strong pro-inflammatory activity in HUVEC cells and to a limited degree also in C3A cells. There is some indication that very high concentrations of 55 nm and 211 nm fluoresbrite polystyrene NP induce toxicity in C3A and HUVEC cells as shown in the Alamar blue and LDH assay. For Caco-2 cells there seems to be slight oxidative stress reaction induced by Au 15 nm NP.
In general, C3A cells and HUVEC were more sensitive to the potential toxic effects of NP than Caco-2 cells, as they showed lower LC50 values and gave slightly positive readouts for the polystyrene NP at high concentrations and for inflammation induced by TiO2.
The LDH assay and measurement of apoptosis via FAS ligand and of inflammation via IL-8 release seem most suitable for routine application in the ILT setup as all three biomarkers can be measured from the cell supernatant with no need to remove the cells from the system. This allows online monitoring. Furthermore, the measured LC50 values were about 5-10 fold lower in the LDH assay compared to Alamar blue assay, indicating higher sensitivity. IL-8 release mirrored the cytokine expression at mRNA level.
For the healthy intestine, there are several more or less complex multi-cellular in vitro models of the intestinal mucosa described in literature and patents. In most cases, these are based on permanent intestinal epithelial cell lines such as Caco-2 or HT-29 However, only one of these models (combining Caco-2 cells with M-cell like Raji cells) has been used to study NP interactions focusing on uptake of MPs across healthy intestinal mucosa without assessment of local inflammatory or cytotoxic reactions. As in INLIVETOX we also wanted to address engineered nanomaterial interactions with the susceptible (i.e. inflamed) intestinal barrier, we introduced an inflamed intestinal model for toxicity testing. Previously, a 3D co-culture model based on Caco-2 enterocytes and primary immune cells had been established by one of the partners. It was decided to adapt the model to for higher throughput screening and also for easier lab-to-lab transfer by replacing, primary blood-derived macrophages and dendritic cells by permanent cell lines.
The cell lines THP-1 and U937 were evaluated to replace the primary blood-derived macrophages while MUTZ-3 cells were targeted to replace the dendritic cells. Studies on activation of THP-1 and U937 cells identified THP-1 cells as the only suitable candidate. As a first step a Caco-2/THP-1 double co-culture was established. In a second step, MUTZ-3 cells were integrated to form a triple culture of enterocytes and dendritic and macrophage-like cells. The modified model had a comparable performance to the original triple culture both in the non-inflamed state and upon stimulation to an inflamed state with pro-inflammatory cytokine IL-1. In the inflamed state, barrier properties were reduced and an increased release of pro-inflammatory marker IL-8 was observed
WP 3: Implementation of biological models in the ILT system
This WP focussed on taking the cell culture models from WP 2 and establishing them in the fluidics system(s) of WP 1 to provide the starting point for the validation work of WP 5.
The first task in the work package was to establish the ILT1 system at the different partners' institutes. After the training course for ILT1 had been carried out, USAAR and UNIPI were provided with complete ILT1 systems and started working with them in their own laboratories. After some weeks of work by the consortium using ILT1, feedback from the different users was analysed and was used to define the requirements for the second version bioreactor, ILT2 that was developed in WP 1.
It was originally planned to establish the 3-tissue culture model - with all three model tissue maintained in the same fluidics system - using ILT1 and ILT0 at this point. However, this was not feasible due to time constraints. Instead, the individual tissue models were set up in ILT0 (for the C3A and HUVECs) or ILT1 bioreactors and the viability and function of the models under the common culture conditions was tested.
However, once the ILT2 bioreactor had been completed, the three-tissue culture model was established. As described in previous WPs, the models established were: for the intestinal epithelium CaCo-2; for the vascular endothelium, HUVEC, for the liver, C3A.
One tissue model that was not established in the ILT system was the triple culture model of the inflamed intestine. This model (described in WP 2) was based on a collagen layer deposed on the surface of the microporous cell culture support. Unfortunately, the collagen layer effectively blocked the transport of NPs from apical to basolateral compartments, making the model unsuitable for study of NP transport. In addition, the collagen layer adhered poorly to the silicon nitride microporous cell culture support used in the ILT system. This did not pose a problem under static culture conditions. However, under flow conditions, the model intestinal barrier peeled off the support so that the separation of the two (apical and basolateral) flow circuits was not maintained. For these reasons, it was decided to abandon the establishment of this model and to focus efforts on the remaining tasks in the project.
Finally, at the end of WP 3 a system baseline was established in the absence of NPs. The baseline provided a negative and a positive (0.1 %Triton-X 100) control for the subsequent experiments in WP 5. The baseline included assays for viability, inflammation, apoptosis and function.
WP4: In vivo assays
The animal studies of INLIVETOX used rodent models to determine the quantitative biokinetics of intravenously injected and orally administered radio-labelled NM (TiO2, Au and Ag) with a special emphasis on cardiovascular and liver uptake since these two targets relate to the targets included in the INLIVETOX model. Tissues from these animals were used to investigate toxicological responses, again focusing on the gut, cardiovascular system and liver. This data was then compared with other biokinetic studies using similar particles, but other routes of exposure (e.g. respiratory route).
Oral gavage biokinetic studies:
- After oral administration to rats, only a small fraction of all of the NM administered were absorbed through the gastro-intestinal-tract (GIT) wall into blood circulation for subsequent accumulation in organs and tissues. The major fraction of the ingested dose for all NM administered were eliminated from GIT into the faeces.
- The 24-hours biokinetics study of the orally delivered monodisperse, different sized (1.4 2.8 5, 18, 80, 200 nm) gold nanoparticles (AuNP) showed a clear, inverse size-dependent absorption through the gut wall into blood circulation (0.4 % of administered 1.4 nm AuNP to 0.02 % of 80 nm AuNP). This led to low but detectable accumulations in organs in the following decreasing order: kidneys, liver, spleen, lungs, heart, brain. Interestingly, AuNP also accumulated in the skeleton. This pattern differed strongly from the pattern after intravenous injection of the same AuNP indicating different surface modifications according to the route of entry which may relate to the formation of the protein corona.
- Similar to AuNP most of the orally administered, agglomerated titanium dioxide NP (TiO2 NP) of 70 nm size were also excreted in faeces. Compared to 80 nm AuNP the GIT absorption of TiO2 NP was five times higher indicating NP material differences. Accordingly, the quantities of TiO2 NP in organs and skeleton were higher but ranked similarly as for 80 nm AuNP. Absorption through the gut into blood circulation increased during the first 24 hours but decreased during the next six days indicating NP clearance from the body.
- Silver NPs (AgNP) were not used for biokinetics studies due to enhanced physico-chemical reactions during the process of radiolabelling resulting in dangerous difficulties in their handling.
Intravenous injection biokinetic studies
- All NM administered intravenously were almost completely retained in the body and accumulated in organs and tissues.
- Intravenous injection of NP suspensions predominantly resulted in accumulation in the liver, justifying the choice of hepatocytes in the INLIVETOX model.
- These studies allowed for kinetic measurements of the hepato-biliary clearance of NP into the small intestine and further into faeces. The monodisperse, different sized (1.4 2.8 5, 18, 80, 200 nm) AuNP showed a clear inverse size dependency of this hepatic clearance pathway. Clearance of the 70 nm TiO2 NP was similar to that of 80 nm AuNP but increased linearly with time up to 2 % of the initial NP dose over one month indicating a persistent but moderate NP clearance pathway out of the liver.
- Given the physico-chemical reaction during the process of radiolabelling the silver NP were not used for biokinetics studies. Instead intravenous injected silver ions were used. This is relevant for Ag NP toxicity due to the potential for AgNP to dissolve. Less than 5 % of the silver ions were found to accumulate in organs and the skeleton. However, they were not excreted in urine as expected but predominantly in faeces indicating effective translocation from blood into the gut lumen. This was not seen for the metal ions of the other NP.
In addition to biokinetics studies, in vivo toxicology studies were carried out. The data from these studies was compared to data from the static in vitro studies of WP2 and to the in vitro studies with the INLIVETOX system in WP 5.
Oral gavage toxicology studies
- A depletion of reduced glutathione (GSH) levels in the liver was used as a marker of oxidative stress. No significant effects were observed for intestinal or hepatic oxidative stress for any NP administered by gavage.
- Similarly, no significant changes in GSH levels were observed in Peyer's patches after gavage of Au NPs of 15 and 80 nm.
- However, there was a significant increase in GSH in Peyer's patches after gavage of 250 µ g TiO2 and Ag per animal indicating oxidative stress in this target tissue for these two particle types.
Intravenous injection toxicology studies
- Of the injected particles (Ag, 15 nm Au and TiO2), only the TiO2 particles caused a decrease in reduced GSH content of livers. No significant changes in GSH levels were observed in Peyer's patches or intestines after injection of any of the particles used.
Inflammation
Exposure of rats to NPs by both injection and ingestion resulted in changes in the expression of genes involved in inflammation, oxidative stress and apoptosis in the target organs examined (liver, Peyer's patches, intestine and aorta). Some of the most interesting results were the following:
- Changes in gene expression were not limited to the primary target organs (i.e. livers / aorta for injection and intestine / Peyer's patches for ingestion), but systemic changes were also observed.
- Changes in gene expression were dose-dependent, and a particularly interesting difference was observed in Peyer's patches, where small doses of 15 nm Au particles (50 µg per animal) caused an increase in the expression of a number of genes, but higher doses (250 µg per animal) caused a strong decrease in gene expression.
Overall, 80 nm Au particles caused the fewest changes in gene expression, and when compared to 15 nm Au particles, there was a pronounced size effect for the Au NPs. When examining H&E-stained sections of livers, intestine, Peyer's patches and aorta, no obvious effects such as strong influx of inflammatory cells or tissue damage could be observed.
No NPs were detected in the Peyer's patches of rats exposed to the range of NPs used in the in vivo exposures (gavages) using TEM.
Ag, TiO2 and 15 nm Au particles were detected by TEM in livers of animals exposed via the tail vein. Ag and TiO2 NPs were detected in both hepatocytes and Kupffer cells, and some had reached the nuclei of hepatocytes. The particles appeared in small clusters and free within the cytoplasm / nuclei. In contrast, 15 nm Au particles were only detected in the cytoplasm of Kupffer cells, and were without exception enclosed within membrane-bound vesicles.
WP 5: Validation of the system and analysis of NP toxicity by ingestion
The first task in WP 5 was to produce and distribute a protocol handbook for the INLIVETOX system, in particular, for assessing NP toxicity. This handbook was produced and has been shared among the ILT partners. It includes:
- an introduction;
- instructions for setup, use and cleaning;
- cell culture, common media and transfer of cells into the ILT system;
- particle suspension preparation and exposure of cells;
- protocols for a number of toxicological endpoints;
- protocols for cell-specific functional tests.
The main goal of the WP was to characterise the fate and behaviour of NPs within the completed INLIVETOX model and the impact of the particles on the cells in the system and to compare this with in vivo data. This was delayed because of the additional time needed to finalise the ILT system and deal with some problems as highlighted in WP 3. Experiments were carried out using Ag NPs, which were selected because they had the highest toxicity to all three cell types used (see WP 2).
A range of concentrations of NM300 NPs (from 0.05 to 50 µg/ml) was applied to the ILT system. This range included a dose which was non-toxic in static experiments, a dose with low toxicity and one with high toxicity above the LC50. Effects of the NPs on cell-specific markers such as TEER and albumin release could be measured, as well as cytotoxicity and inflammatory mediators in the medium.
Ag NP induced a dose dependent decrease in viability of the three cell types in the ILT system with the lowest concentration (0.05 µg/ml) inducing no significant effect. A slightly higher concentration of 0.5 µg/ml induced a moderate decrease in TEER (by about 15 %), but significant LDH and reduced HUVEC viability. This toxicity was also associated with an increase in the production of the pro-inflammatory mediator IL8. At the higher Ag NP concentrations of 5 and 50 µg/ml epithelial barrier function was greatly affected as shown by TEER measurements. The loss of barrier integrity of the Caco2 cells was associated with greater downstream tissues effects such as cell death in HUVEC and reduced albumin production in C3A cells. High levels of LDH and inflammatory proteins could be measured in both circuits.
Comparisons were made between in vivo responses to NP exposure by injection and the responses of static single cell in vitro models.
Remarkable similarities were observed between the static in vitro models and the in vivo data. These similarities were observed especially for Ag and TiO2 NPs for:
- intracellular particle uptake;
- oxidative stress (measured by depletion of reduced glutathione);
- gene expression of inflammatory markers, particularly IL-8/MIP-2 (the rat homologue of IL-8), TNF and IL-1RI.
Since particle accumulation mostly occurs in the liver, studies focussed on this organ. Both high toxicity Ag and low-toxicity TiO2 NPs were detected in cells in the liver after injection and were taken up on exposure in vitro. However, the responses observed for the two NPs were very different. No oxidative stress was observed for Ag NP either in vitro or in vivo. However, IL-8/MIP-2, TNF and IL-1RI expression were all increased both in vitro and in vivo. In contrast, TiO2 NP did cause oxidative stress in vivo and at high doses in vitro. TNF and IL-8 expression were increased after 24 hours both in vivo and in vitro, while IL-1RI was relatively unchanged.
A calculation of doses delivered to the liver in vivo and hepatocytes in vitro indicated that they are comparable. The similarities observed suggest that simple in vitro models using human cell lines compare very well with intravenous in vivo rat studies.
In vivo data on NP exposure by ingestion was compared with the ILT model. The biokinetics studies showed that in the ingestion model, only a small number of particles (maximum 0.5 % for 1.4 nm Au particles) translocated from the gastrointestinal tract to the blood and secondary target organs. This is in strong contrast to the injection model, where rapid and near-quantitative accumulation of particles in the liver occurred. This makes in vitro modelling of ingestion with simple models very difficult, because interactions of particles with the food, translocation through the GIT and distribution, as well as interactions with cells along the way and the communication of these cells with each other and on a systemic level have to be taken into account.
This situation was clearly reflected in the much more complex in vivo tissue response to particle exposure via intraoesophagial gavage as seen for gene expression. This type of integrated and complex response was observed in the ILT model in which:
i) very low translocation of PS-FITC to the basal compartment was nonetheless accompanied by substantial NP uptake in the target C3A and HUVEC cells; and
ii) Ag-induced toxicity and inflammation was observed at very low concentrations and translocation.
In the in vivo ingestion studies, Ag had no significant impact on cell death (as measured by gene expression markers of apoptosis) or pro-inflammatory gene expression (including IL8). In contrast, the Peyer's patches were very sensitive to the Ag NP with increases in some pro-inflammatory genes (e.g. TNF-alpha and MCP-1), decreases in others (e.g. IL1-beta) but no significant changes in markers of apoptosis. When compared with the results of the ILT system (see above), this suggests that the ILT system did not reflect well the downstream responses to Ag NP of tissues such as the endothelium and liver following exposure via ingestion. The ILT system response to Ag NPs therefore reflects the static in vitro model responses rather that the in vivo ingestion model responses. The ILT response also reflects the in vivo liver responses following injection of Ag NPs.
Data from the single cell studies, inflamed cultures and three-tissue experiments under flow conditions were analysed by three-way principle components analysis assessing a combination of variables (subjects), objects (response variables) and conditions. This study is the first application of three-way PCA to toxicity data.
In summary, regarding viability, it was concluded that C3A cells are the most sensitive to NP exposure, followed closely by Huvec and then Caco-2 and the inflamed model. Inflammation was higher for TiO2 than for Ag NPs. The inflamed model gave the highest response, and C3A were the most sensitive cells, followed by Huvec and Caco-2. Viability of cells in the static and dynamic 3 tissue model was also examined. Under static conditions, there was a clear difference in susceptibility of the different cell types. However, under flow conditions with crosstalk between cells, there was overlap between all three cell types, highlighting the fact that the three tissue model under flow conditions is more than just the sum of the three single cell types, and represents interactions between the cells which cannot be modelled in a single cell system or without medium flow.
Overall, based on silver data, cells in the system are more susceptible to NP exposure when under flow than as single cells in static conditions. This indicates once more the importance of a co-culture system which enables communication between the different tissue models.
Finally, nanoparticle toxicity in the inflamed intestinal model was studied. The results were compared to the non-inflamed triple culture model and the Caco-2 monoculture.
Testing in the non-inflamed and inflamed co-culture model of the intestinal mucosa changed the toxicity profile of the engineered nanoparticles investigated in the INLIVETOX project. TiO2 nanoparticles showed no toxic or inflammatory effect in any of the models. In contrast, the Au NPs had no significant toxicity but induced strong inflammation in both triple culture setups, which couldn't be observed in the Caco-2 monoculture. Differences between the three cell culture test systems were most pronounced for Ag NP. Toxicity was observed in both mono- and co-cultures, but the LC 50 values were markedly lower in the triple culture setups compared to Caco-2 monoculture. Furthermore, significant inflammation was observed in the triple culture models while being much weaker in the monoculture.
In general the co-culture model more easily picked up pro-inflammatory potential of NP. The innate immune cells have a protective effect, preferentially taking up NPs, reducing epithelial exposure and inducing an immune response to a perceived threat. In the inflamed model, this protective function is impaired. In summary, the triple culture setups provided valuable data for interpretation and correlation of in vivo results and mechanistic understanding of NP interaction with the gut mucosa. This is an important step to in vitro modelling of not only the healthy, but also the diseased gastrointestinal tract.
In conclusion, the developmental and experimental work has resulted in a working system for NP exposures with a set of instructions which ensures reproducibility of experimental procedures among ILT users. The baseline created in WP 3 has been expanded upon with further experiments including a range of doses of high-toxicity Ag NPs, which show effects on all three types of cells. These effects vary between cell types and are dose-dependent, and differ from those observed in single cell or static cultures.
Comparisons of in vivo data with data generated by simple static in vitro models and by the ILT system showed that simple in vitro models can give surprisingly good correlations with in vivo studies in some cases. Here, simple static cultures of C3A cells showed striking similarities to the in vivo data for studies based on injection. In contrast, no such similarities were found between the equivalent static cultures, the ILT system and the in vivo studies of Ag NP induced nanotoxicity by ingestion. Ag was chosen because of its relatively high toxicity observed in the static in vitro models used in the project. It would be useful to test the ILT response to the lower toxicity material Au to see whether the results induced are comparable to the in vivo responses observed following ingestion. Although Au was lower in toxicity, it induced a wider array of changes in gene expression of pro-inflammatory, anti-inflammatory, antioxidant defence and apoptosis genes than Ag allowing for a more in depth comparison. It would also be useful to compare the results of the ILT system with injection responses in vivo for a wider array of NP and to compare these with static cultures to assess whether the additional complexity of the ILT system provides any advantages over the cheaper and easier static models for predicting responses to intravenous injection.
Potential impact:
Socio-economic impact can occur at the micro level (companies or partners using the project results) at the meso level (specific industry sectors or geographic regions benefitting) or at the macro level, where benefits are experienced across countries or the whole EU. The INLIVETOX project will have impact at all these levels.
At a micro level, the project has delivered an exciting and innovative technology that has the potential to underpin new product developments in the field of in vitro testing. This has created (and will continue to create) commercial opportunities for some project partners and for those organisations that licence the technology from the partners - as demonstrated by the very high level of interest observed from potential end users in both industry and research organisations. The very high quality of the project work and the results obtained will further enhance the reputations of the research organisations that participated and strengthen their capability to take a leading role in future internationally leading research in the field of in vitro testing and toxicity.
At a meso level, the results of the INLIVETOX project have the potential to change the way that the pharmaceutical, chemical, cosmetic and food sectors of industry are testing the safety and efficiency of new materials. The improved methods could deliver significant economic benefits both through reduction of testing costs compared to the use of animals, but also through the opportunity to bring safer products to market faster than existing methods. The technology developed in the project could provide a significant competitive advantage to the early adopters.
At a macro level, the project confirms the internationally competitive position that Europe's research organisations hold in the fast developing field of in vitro testing. A report by the United States (US) National Research Council in 2007 entitled 'Toxicity testing in the 21st century' outlined the scale of the challenges and suggested a 15 year time scale for the replacement of many animal testing methods by in-vitro techniques.
The WYSS Institute in Boston, US, has recently been granted USD 26 million for a project to develop a 'lab-on-a-chip' solution to toxicity testing. The capability demonstrated by the INLIVETOX project shows that EU researchers could be leaders in this field, if they are able to secure similar levels of funding.
The motivation for such an investment is the size of the existing (GBP 1.8 billion) market for drug toxicity testing and also the emerging (GBP 2 billion) market for the testing of chemicals to comply with REACH legislation.
Any change in methodology for the testing of new drugs will require regulatory approval. The regulatory bodies (ECVAM, FDA etc.) are justifiably cautious in approving new methods. However, the cornerstone to any change is sound science and the work of the INLIVETOX team has demonstrated the capability of European research organisations to make significant progress.
To summarise the potential impact of the INLIVETOX project:
- Direct commercial benefit to the partners or their licensees, opening up new commercial product opportunities that could reach revenues of (EUR 7 million) within 5 years.
- Potential annual savings for the pharmaceutical and chemical sector industries using the technology measured in 10's of millions in the short-term (5 years) and 100's of millions in the longer term (10 years).
- Potential safer drugs and reduced risk in chemical trials through the reduction in false positives or false negatives in toxicity testing.
- Significant reductions in the number of animals used for testing of new drugs and chemicals.
Main dissemination activities
A project identity set consisting of project logo, leaflet and all associated templates was created at the start of the project and distributed at many conferences and exhibitions throughout the duration of the project.
INLIVETOX website (please see http://www.inlivetox.eu/(öffnet in neuem Fenster) online) was created by ALMA and updated with data from all the partners, and will be maintained by Alma Consulting for a period of five years following the completion of the project.
Three press releases were made between June 2010 and the end of the project and articles or interviews appeared in Edinburgh Napier News, Radio Suisse Romande, l'Express / l'Impartial, Le Temps, La Gruyère, Migros magazine, The Scotsman, BBC news and Lepoint.fr.
Two workshops and seminars were delivered in Ispra and Saarbrücken in 2010 and 2011.
Specific training courses are planned by Pisa University and Kirkstall to disseminate the methods developed during the project. The first of these was held on 17 and18 September 2012 in Rome. It is expected that these courses can be self financing and so can continue at regular intervals during 2012 and 2013.
Papers were presented by the project partners at 13 conferences in Rome, London, Lausanne (2), Saarbrücken (2), Montpellier, Edinburgh, Grenoble, Krakov, York, Essen and Utrecht.
The project results were exhibited at three major trade shows in Paris, Manchester and Munich as well as several smaller events.
The following scientific papers have been published already or will be published in the near future:
- Quality control in in vitro nanoparticle testing: Collagen coating can optimise particle distribution on adherent cell lines and improve data quality
B. Gaiser
'Modern Polymeric Materials for Environmental Applications', Vol. 4, Issue 1, 2010, Ed. K. Pielichowski, 978-83-930641-1-3
- A complementary definition of nanomaterial
W. Kreyling
Nanotoday, Vol. 5, Issue 3, June 2010, pp. 165 -168, Elsevier, 10.1016/j.nantod.2010.03.004
- Generation and characterisation of stable, highly concentrated titanium dioxide nanoparticle aerosols for rodent inhalation studies
W. Kreyling
Journal of Nanoparticle Research, 2011, Vol. 13 No. 2, pp. 511-524, Springer, 10.1007/s11051-010-0081-5
- Engineering quasi-vivo in vitro organ models
A. Ahluwalia
Adv Exp Med Biol. 2012; 745:138-53, Springer, 10.1007/978-1-4614-3055-1_9
- Size and surface charge of gold nanoparticles determine absorption across intestinal barriers and accumulation in secondary target organs after oral administration
Nanotoxicology, February 2012, Vol. 6, No. 1, pp. 36-46, doi:10.3109/17435390.2011.552811)
http://informahealthcare.com/doi/abs/10.3109/17435390.2011.552811(öffnet in neuem Fenster)
- Particle size-dependent and surface charge-dependent biodistribution of gold nanoparticles after intravenous administration
Hirn S.
Eur J Pharm Biopharm. 2011 Apr;Vol. 77 No. 3, pp. 407-16, Elsevie
http://www.sciencedirect.com/science/article/pii/S093964111000370X(öffnet in neuem Fenster)
- Using in vitro models to assess the toxicity of nanomaterials
Stone V.
In Vitro Toxicology Society
Three chapters in a book entitled 'New developments in cell-based in-vitro testing - methods and protocols' have been written by INLIVETOX partners. The book is available online in June 2012 by Pan Stanford Publishing (please see http://www.panstanford.com(öffnet in neuem Fenster) online).
Direct mailshots of a newsletter were made to 24 industrial companies who were potential users and 2 organisations involved in regulatory matters (ECVAM and LGC).
Exploitation of results and commercialisation strategy
The overall objective of the strategy is to maximise the uptake of the project results. The plan accommodates both those beneficiaries who wish to be involved directly in commercial activities and those who wish to licence others to do this on their behalf. The plan takes account of regulatory aspects and ethical issues.
There are two main sectors that are being targeted in the commercialisation plan: the academic research community working on advanced in-vitro models of cell culture and the commercial companies involved in the testing of safety or efficacy of compounds for the pharmaceutical, cosmetic household products or food industry.
In the commercialisation plan careful consideration was given as to whether to publish, licence to others or develop commercially ourselves. The latter two options are viable because steps have been taken to protect the IP: one patent has been applied for on the final ILT2 holder. Copyright designs and secrecy are further options still available to the consortium.
Products and services
The following products and services that build on the IPR created during the project will be commercialised by the partners:
- protocols to set up cell cultures to create viable biological tissue models of the intestine and other barriers within the body (service);
- protocols to determine Cytotox and inflammatory responses (service);
- data on biokinetics from in vivo studies (service);
- modular microfluidics based in vitro test system 'INLIVETOX system' including necessary protocols of how to use it in various applications (product);
- improved microfabricated membranes for cell culture (product).
Routes to market
The service offers described above will be marketed direct by the organisations involved. For the products a different approach is being taken by Kirkstall, i.e. initial market creation in the local market (UK) by direct sales. This is followed by selection of distributors to serve regional markets.
Competition
The market for products and services in the field of in-vitro testing of safety and efficacy for drugs, cosmetics and chemicals is very large and hence there are many competitors already and many more emerging. The leading players have invested heavily in development of their offers and then followed by perhaps even greater investment in getting their tests validated and approved by the regulatory bodies such as ECVAM.
The INLIVETOX consortium members will be seeking further investment and partnership with major global companies to ensure the maximum take up of the technologies developed during the project.
The service offers could be available before the end of 2012. The product offers are expected to take one to two years further development subsequent to the end of the project.
Website: http://www.inlivetox.eu(öffnet in neuem Fenster)
Coordinator:
Centre Suisse d'Electronique et de Microtechnique SA (CSEM), Switzerland
Dr Martha Liley
E-mail: Martha.LILEY@csem.ch
