Final Report Summary - NANOLYSE (Nanoparticles in Food: Analytical methods for detection and characterisation)
Executive Summary:
The NanoLyse project was set up to develop methods for the reliable determination of nanoparticles in food matrices. The need for such methods resulted from the increasing interest into the presence of nanomaterials in food, both from risk assessors and regulators, but also from industry in expectation of upcoming regulations.
NanoLyse established a dual, tiered approach that is consistent with current procedures for food safety monitoring. Tier 1 involves rapid screening and imaging methods for a cost-efficient yes/no decision on the presence of nanoparticles in the sample. Tier 2 provides sophisticated coupled separation and detection methods for the unambiguous identification and reliable quantification of nanoparticles in food and beverage matrices. Validation is based on suitable guidelines and makes use of stable and homogeneous reference materials. This concept was successfully translated into practice by the development and validation of methods for selected nanoparticles and complemented by the preparation of respective reference materials. The project accomplished major achievements in the following areas:
- Screening: Rapid, easy to use and cost-efficient screening methods were developed, based on two different platforms. This involved ready to use ELISA kits for the detection of cross-linked gelatine nanoparticles and a ß-lactoglobulin based nanocarrier for food additives in beverages, furthermore surface plasmon resonance (SPR) biosensor assays for these organic nanoparticles as well as an SPR assay for silver nanoparticles.
- Imaging: Electron microscopy (EM) methods for inorganic nanoparticles in food (i.a. silver particles in meat and silica particles in soup) were developed with emphasis on an optimised sample preparation for these complex matrices. An object-based image analysis system for the (semi) automated smart analysis of large numbers of EM images was delivered, enabling rapid and cost-effective EM data evaluation.
- Methods for inorganic particles: Field flow fractionation coupled online with light scattering and ICP-MS detectors was used for quantitative, element specific determination of silver nanoparticles in meat and silica in soup. Additionally, a single particle ICP-MS method for the determination of number-based size distribution of Ag particles was developed and subjected to an interlaboratory method performance study.
- Methods for organic particles: Organic nanoparticles differ significantly from inorganic ones, in particular, they are much more fragile. Care has been taken to develop adequate methods and yielded a universal method for protein-, carbohydrate- and lipid-based NP that is based on separation by hydrodynamic chromatography, quantification by UV absorbance detection and identification by MALDI-ToF-MS fingerprinting. This was complemented by methods for a fast screening for nanoparticle components based on DART-MS and UPLC-Orbitrap MS as well as a parallel GEMMA method for particle number concentrations.
- Reference materials: Stable, homogeneous and well characterised reference materials (pairs of pure suspensions and matrix materials) were prepared. Labelled Ag and SiO2 particles were synthesised for use as internal standards. All developed methods were validated according to a generic validation approach drafted in the project.
Dissemination of the project results to the scientific community as well as to regulatory and industrial stakeholders was realised via presentations at conferences, scientific journals, NanoLyse Open Days as well as policy support workshops. Technology transfer to potential end-user was achieved by hands-on training workshops.
Project Context and Objectives:
Nanotechnology applications for the food sector are intensively investigated and developed at the moment. A number of nanomaterials are already in use as food additives or in food contact materials. Furthermore, approved (bulk) food additives may have a size distribution which extends down to the sub 100 nm range, e.g. fumed silica (E551). At the same time, limited knowledge is available on the potential impact of engineered nanoparticles (ENP) on consumers’ health. A prerequisite for toxicological, toxicokinetic, migration and exposure assessment studies is the availability of analytical tools for the detection and characterisation of ENP in complex matrices such as food.
At the starting point of the project, the availability of methods for the detection and characterisation on engineered nanoparticles in complex matrices, and in particular in food, was very limited. Some approaches existed for the measurement of inorganic particles in environmental matrices and electron microscopic imaging in e.g. tissues while no methods were available for screening of any ENP and for the measurement of organic ENP in food.
The NanoLyse project focussed on the development of validated methods and reference materials for the analysis of engineered nano-particles (ENP) in food and beverages. The developed methods were to cover relevant classes of ENP with reported or expected food and food contact material applications, i.e. metal, metal oxide/silicate, and organic ENP. Priority ENPs have been selected out of each class as model particles to demonstrate the applicability of the developed approaches, e.g. nano-silver, nano-silica, a protein-based nano-carrier system as well as fullerenes (potential food contaminants). Priority was given to methods which can be implemented in existing food analysis laboratories. Rapid imaging and screening should be achieved by the application of automated smart electron microscopy imaging as well as screening techniques in sensor and immunochemical formats. More sophisticated, hyphenated methods were intended to allow the unambiguous characterisation and quantification of ENP. These include elaborate sample preparation, separation by flow field fractionation and chromatographic techniques as well as mass spectrometric and electron microscopic characterisation techniques. The developed methods were validated using the well characterised food matrix reference materials that were produced within the project. In summary, this includes the following objectives:
• Reference materials for the analysis of nanoparticles in food and beverages
• Sample preparation methods for the detection of nanoparticles in food
• Rapid imaging and screening methods for nanoparticles in food
• Analytical methods for the identification, characterisation and quantification of inorganic and organic nanoparticles from the food matrix
• Dissemination and training of the new methods to relevant stakeholders
The European Commission published a recommendation for a definition of a nanomaterial in October 2011, halfway of the duration of the project. This number-based definition shifts the analytical requirements to particle counting methods. NanoLyse reacted to the resulting changed needs of stakeholders by replacing a mass based method for inorganic ENP by an easy implementable particle counting method (single particle ICP-MS), in addition to the more laborious electron microscopic techniques foreseen in the original work plan.
Project Results:
Methods for the measurement of nanoparticles (NPs) in food were virtually not available at the start of the project and there were strong doubts if reliable approaches can be developed for such complex matrices as food. NanoLyse succeeded, in 3,5 years’ time, not only to demonstrate that nanoparticles can be detected, identified and even quantified in food, but also delivered a number of fully validated methods for relevant NP in realistic food and beverage matrices. These results were achieved within 4 distinct RTD work packages (WP) that closely collaborated. WP1 (Reference Materials) supplied well characterised NP suspensions and NP spiked matrix materials, synthesised labelled analogues and provided a general validation approach for NP analysis in food. WP2 (Imaging and Screening methods) developed electron microscopy imaging methods and an automated image analysis system as well as screening assays in biosensor and ELISA format. WP3 was dedicated to the development of sophisticated coupled separation and detection methods for inorganic NP, while WP4 had the same task for organic NP. The main achievements are presented per WP below.
REFERENCE MATERIALS (WP1)
- NP suspensions
The initially foreseen target NP were thoroughly tested for their suitability as future reference materials, involving criteria such as stability and well defined material characteristics (e.g. size distribution), but also the possibility to measure them with the foreseen analytical techniques. As a result of this revision, some target particles had to be replaced by others and the final target NP to be used in the project were: silver NP (example for a metal NP), silica NP (oxidic NP), cross-linked gelatine NP (organic NP) and C60 fullerenes (carbon NP).
Silica nanoparticle (SiO2 NP) suspensions and silver nanoparticle (Ag NP) suspensions were prepared from commercially available materials. They were tested for homogeneity and stability and characterised for particle size by dynamic light scattering (DLS), electron microscopy (EM), field-flow-fractionation-inductively coupled plasma mass spectrometry (FFF-ICPMS), gas-phase electrophoretic molecular mobility analysis (GEMMA) as well as for total silica or silver content by ICPMS or k0-neutron activation analysis (k0NAA). Gelatine suspensions and C60-fullerene solutions were tested for homogeneity and stability by DLS (gelatine) and liquid-chromatography-mass spectrometry (LCMS) (fullerenes). The gelatine particles were characterised for size by GEMMA, DLS, TEM and AFM.
- Labelled analogues
Reliable quantification of the target analytes is a major issue in the statutory monitoring for food safety. (Isotopically) labelled internal standards play a major role for this purpose in the analysis of conventional analytes (molecules, ions). NanoLyse established the same approach for NP analysis by the synthesis of labelled analogues of the target analytes. Au labelled Ag NPs were produced from a mixture of AgNO3 and HAuCl4 by reduction with NaBH4. The suspension was purified by centrifugation and washing with distilled water. The UV-VIS spectrum showed only one peak, confirming incorporation of the Au in the Ag matrix. Transmission electron microscope (TEM) images showed roughly spherical particles with particle sizes ranging from 4 to 24 nm. X-ray photoelectron spectroscopy was used to quantify the Au/Ag ratio and particles were found to contain 2 % (Mol/Mol) Au. The particles are readily suspensible in water.
GeO2 labelled SiO2 was produced from a mixture of tetraethoxysilane and germanium ethoxide. The material was purified by centrifugation and repeated extraction with ammonium nitrate in ethanol. Scanning electron microscope (SEM) pictures show spherical particles of approximately 100 nm. X-ray photoelectron spectroscopy was used to quantify the Ge/Si ratio and particles were found to contain 0.8 % (m/m) GeO2. The dried powder was readily dispersible in ethanol, but dispersion in water required sonication for more than 20 minutes. Therefore, a second batch of material was produced which was not dried to a powder, but was distributed as paste instead.
Fluorescent labelled gelatine was produced using a commercial protein labelling kit and fluorescein hydrochloride (DTAF). The outcome was investigated by confocal microscopy. Localised fluorescent points show that the labelling with the fluorescent dye was successful.
13C labelled C60 fullerenes are commercially available.
- Food matrix reference materials
Four homogeneous and stable matrix reference materials (RM) for Ag NP in chicken meat and SiO2 NPs in tomato soup were produced. The chicken meat materials were characterised for total Ag content by k0NAA and for Ag NP size by single particle-ICPMS (sp-ICPMS) and transmission electron microscopy (TEM). The soup material was characterised for total silica content by ashing and for SiO2 NP size by FFF-ICPMS. The development of a homogeneous and stable suspension of gelatine NPs in beverage was not successful, presumably due to interactions of the gelatine NPs with the gum Arabic in the beverage. An approach for the production of a homogeneous solution of C60-fullerenes in oil was developed. These materials are complemented by the corresponding neat NP dispersions (see above), which gives users the unique opportunity to compare results on matrix materials after sample preparation with the results on the NPs before adding to the matrix. The reference materials are available to other interested parties and projects after the end of NanoLyse as long as stocks last.
- Validation approach
A generally applicable validation approach for the analysis of NP in food (and other complex matrices) was developed. It was based on generally accepted principles in analytical chemistry as laid down in international guidelines, taking into consideration potential differences arising from the particulate matter of the measurands. This is the first published systematic approach for the validation of methods detecting and quantifying NP. The validation approach was applied in the validation of methods developed in the project and refined based upon the resulting experiences. Furthermore, the validation approach was posted on the public project website and comments from external experts invited to further develop the proposed approach.
RAPID SCREENING AND IMAGING METHODS (WP2)
- Sample preparation for electron microscopy (EM)
Appropriate sample preparation is an essential first element of any analytical method. It is important for nanoparticle detection/characterisation methods to ensure that there is no significant change in the state of nanoparticles due to dissolution or agglomeration/ aggregation in order to avoid erroneous analytical results. As water is the main component in the majority of food products, its removal as required for EM analysis can also trigger changes to the nanoparticles. In NanoLyse different sample preparation methods were compared for suitability for unperturbed preparation of nanoparticle containing dispersions and food matrices. The methods included air drying, freeze drying, sedimentation by ultracentrifugation, removal of water by blotting, drying by chemical means, resin embedding, snap freezing, etc. The evaluation showed that each method has certain advantages and disadvantages for specific purposes. Results of validation showed ultracentrifugation and blotting to be the most suitable methods for sample preparation due to the least variation in nanoparticle sizes when measured by EM. A respective guidance document was issued and the experiences used to optimise the methods developed in the project.
- EM imaging methods
Imaging approaches based on electron microscopy (EM) have been thoroughly tested and validated against other analytical methods studied under NanoLyse for detection/characterisation of nanoparticles in dispersions and selected food matrices. The data and information generated provides a comprehensive insight to the capabilities and limitations of the EM methods. Based on these results dedicated methods for the NanoLyse analyte/matrix combinations were developed, in particular a scanning EM (SEM) method for silica NP in tomato soup and a transmission EM (TEM) method for Ag NP in chicken meat.
- Automated image analysis system
A major bottleneck in the use of imaging methods for high throughput analyses is the time and effort needed to manually analyse each individual EM image. To address this problem, NanoLyse developed a software-based platform capable of processing large numbers of EM images to derive data on particle numbers, size distribution, and certain morphological (shape) characteristics. The software solution was developed in the form of an object-based image analysis system based on eCognition® Architect using the Cognition Network Technology®. Rule sets can be applied and parameterised using the developed “Architect-Solution” which in this case would reflect a typical workflow of object based image analysis (OBIA) to detect and count nanoparticles. The system is easy to use and guides the user step by step through the workflow. It also allows for parameterisation wherever necessary and to test and exclude intermediate results. This way, the user can focus on relevant parameters for object detection and can generate report statistics in simple steps. In case the results are not precise enough, the user can either re-parameterise single steps of the solution or correct them manually in the image at any time.
Although the system was originally intended to be developed as a fully automated platform, following testing on nanoparticle images in dispersions and food matrices of increasing complexity, as well as variable quality of images (brightness, contrast, focus) it was deemed more appropriate to retain a degree of expert oversight and thus make the system semi-automated. The developed solution modules were validated for analysis of EM images of increasing complexity in food matrices. The developed OBIA system was tested against currently available software based on pixel based imaging analysis (PBIA).
- Enzyme-linked immunosorbent assay (ELISA) for organic NP
ELISA is a robust and cost-efficient technique for the rapid analysis of i.a. food samples. Within NanoLyse, it was demonstrated that this technique can also be used for the detection of NP in food and beverage samples. It is in particular suited for organic NP which are often more difficult to analyse by other techniques due to their fragile nature. Prerequisite is that they possess immunogenic properties that can be used to raise specific antibodies. This fact excluded some of the originally foreseen target particles (organic nano-encapsulates of which the shell compounds did not trigger an immune response). Antibodies were successfully raised against cross-linked gelatine NP (a potential carrier for food supplements or flavours). These antibodies are highly specific to the cross-linked gelatine NP and do not cross-react with different native forms of gelatine. They were used to develop a screening assay for the detection of cross-linked gelatine NP in beverages (juices, soft drinks) with detection limits of 250 – 500 ug/L. The method was validated for limits of detection (LoD), repeatability, precision, and robustness according to 2002/657/EC and the validation guidelines developed in WP1. A prototype ELISA kit for gelatine NP was manufactured. Furthermore, an existing antibody against native ß-lacto globulin was shown to recognise also nanoparticulate ß-lacto globulin complexes with epigallocatechin gallate (EGCG, an antioxidant food supplement) and an assay based on this antibody was developed and validated.
- Surface plasmon resonance (SPR) biosensor assays for silver and organic NP
SPR biosensors allow a label-free high throughput screening and are used for the detection of residues and contaminants in food safety monitoring. Assays for silver NP and organic NP were developed and validated in the project. The assay for silver NP is based on a metal-binding protein (metallothionein MT1) that was immobilised on the sensor chip surface. Although this metallothionein also binds silver ions and other metals the selectivity towards Ag NP was increased by an optimised composition of the buffer and a size specific filtration step in the sample preparation. The assays for organic NP make use of the antibodies that were also used for the ELISAs. The developed assays showed adequate sensitivity in the low ppm range and were tested in different fruit juices and carbonated and non-carbonated flavoured mineral water. For ß-lacto globulin NPs a selective rapid, 2 minutes, sample-pre-treatment step was developed to further improve the assay’s selectivity towards the nanoform of ß-lacto globulin, using disposable spin-through size exclusion cartridges. Total assay time was about 12 minutes and the sensor chips showed high robustness and stability over a period of several months. The utilization of the sensor described here for detection of protein-based NPs in food, can facilitate rapid and automated screening needed for efficient food safety monitoring
SEPARATION AND CHARACTERISATION METHODS FOR INORGANIC PARTICLES (WP3)
- Sampling for NP analysis
Sampling protocols were developed for inorganic engineered nanoparticles in at least three food matrices. The issue of preparing a representative sub-sample was addressed by referring to different ISO standards and to a NMKL guideline on the subject for sampling in the context of the NanoLyse project of olive oil, soup and meat. Different types of distributions were considered, which were assumed to be relevant for count data from experimental methods detecting nanoparticles. The Poisson distribution was chosen as one of the most common discreet distributions, which are frequently used to describe count data. Calculations have been made for silver and silica nanoparticles regarding the minimum sampling volumes necessary for having a representative sub-sample. The calculations showed that even at low concentrations (i.e. in the 10 ng/ml range) the number of sampled nanoparticles in very low volumes (a few microliters. i.e. lower than typically taken by for example pipetting) was sufficient to allow for representative sub-sampling when working with nanoparticles in the size range of 1 to 100 nm.
- Quantitative multidetector FFF method for Ag NP in meat
A method has been developed that is capable of extracting Ag NP from chicken meat samples, determine their size and quantify their mass amount. After comparison of different approaches enzymatic digestion was selected and optimised for sample preparation. Recovery rates were > 80%. The analytical separation of the Ag NP was achieved by asymmetric flow field flow fractionation (AF4). A systematic optimisation approach for a large set of parameters in FFF (carrier composition and concentration, cross flow, injection flow and time, injected mass, membrane type and spacer height) was developed and applied. The resulting method provides optimal recovery, separation power, retention time, peak shape, and run time for the target analyte. Separation has been coupled on-line with a suite of three different detectors (UV/Vis absorbance at a wavelength of 400 nm, one angle (90°) of the MALS detector and the 107Ag trace of the ICPMS. ICPMS was used for the mass based quantification of the Ag NP in samples. The overall method was validated for precision, trueness (recovery), limit of quantification and ruggedness. A small scale inter-laboratory comparison of the method demonstrated that it is transferable to well-trained laboratories. A standard operating procedure (SOP) for the isolation, detection, characterization and quantification silver nanoparticles (Ag-ENPs) in lean chicken meat by AF4 was provided.
- Quantitative multidetector FFF method for silica NP in soup
A method has been developed that is capable of extracting silica NP from tomato soup samples, determine their size and quantify their mass amount. After comparison of different approaches acid digestion was chosen with the aim to completely oxidize the organic matrix with only minor effects on the SiO2-NPs. Applying a sample pre-treatment with heating and homogenization treatment followed by acid digestion proved to be suitable for extraction of SiO2-NPs from tomato soup. Si mass recoveries were in an acceptable range (> 90%). The analytical separation of the silica NP was achieved by asymmetric flow field flow fractionation (AF4). The same systematic optimisation approach as described for Ag NP has been applied and resulted in a method with an optimal compromise between recovery, separation power, retention time, peak shape, and run time for silica NP. The FFF separation was coupled on-line with light scattering (MALS) and ICP-MS detectors. MALS data was used to determine the rms-radius applying the Debye fit (3rd order) while ICP-MS was applied for the mass based quantification of the silica NP. The overall method was validated for precision, trueness (recovery), limit of quantification and ruggedness. A standard operating procedure (SOP) for the isolation, detection, characterization and quantification silica particles (SiO2-ENPs) in tomato soup by AF4 was provided.
- Particle counting single particle ICP-MS method for Ag NP
A particle counting method for Ag Np was developed in response to the new EC recommendation for a definition of nanomaterial. The method is capable of extracting Ag NP from a meat matrix and to measure the mass of individual particles in the resulting suspensions. Subsequently, number-based size distributions can be determined. Enzymatic digestion and dilution in water was used for sample preparation. Detection was achieved by ICP-MS in time resolved mode. The method was validated first in a single laboratory approach and then subjected to a full interlaboratory method performance study with 23 participants. The results indicate that single particle ICP-MS is a suitable (screening) technique for silver NP. A standard operating procedure (SOP) as well as a data evaluation macro were provided.
SEPARATION AND CHARACTERISATION METHODS FOR ORGANIC PARTICLES (WP4)
Organic NP are usually more fragile and complex in nature as compared to most inorganic NP. Typical structures for such NP are micelles, liposomes, encapsulates and complexes build from large or macromolecules (proteins, carbohydrates, lipids) carrying small molecules (e.g. vitamins, antioxidants, flavours, preservatives). This requires from analytical methods that they leave the NP structures intact and allow a complex chemical characterisation (in contrast to inorganic NP where elemental composition often is sufficient). This was achieved in WP4 by a number of different approaches.
- Universal method for organic NP based on HDC-UV-MALDI-TOFMS
This method was developed for the size determination, quantification and characterisation of organic NP of different nature (protein-, carbohydrate-, lipid-based). Sample preparation from beverage matrices was realised by the application of filtration and size exclusion chromatography. The analytical separation and size determination (after calibration with respective standards) is achieved by hydrodynamic chromatography (HDC). The on-line coupled UV detector allows the reliable quantification of the NP amount. The unambiguous chemical identification is carried out by means of an off-line coupled MALDI-ToF-MS (fingerprinting of the NP). The method has been developed and tested for different types of organic NP. A full validation according to EC/2002/657 was exemplarily carried out for the determination of a commercial lipid-based nanoparticle in beverages and a standard operating procedure (SOP) was provided.
- GEMMA
Differential mobility analysis was combined with an electrospray unit for particle ionisation (GEMMA). This combination is very successful in particle size separation and the production of number-based size distributions according to the requirement of the EC definition of a nanomaterial. A SOP was produced for the combination of GEMMA with a condensation particle counter (CPC). Validation of the method shows that GEMMA is useful as a qualitative method for ENP characterisation.
- Mass spectrometry based screening for organic NP
Two rapid methods for a decision on the presence of target NP in food/beverage matrices have been developed as an alternative to the more laborious HDC and GEMMA methods. Components of the organic NP (e.g. building blocks of micelle shells or active ingredients encapsulated inside) can be detected and quantified by the developed methods. The first option uses UHPLC-TOFMS. The second option uses DART (Direct analysis in real time) ionisation coupled to high resolution (Orbitrap) mass spectrometry. This is a high resolution ambient MS technique used for rapid fingerprinting/profiling directly from the sample without sample preparation or chromatographic separation. Nanoparticle components in the sample can be detected on the basis of characteristic fragment ions. Both methods have been validated for the direct determination of commercial micelle-type polysorbate based ENPs in beverages.
- Method for fullerenes in edible oils
A method was developed for the determination of C60 fullerenes in vegetable oils (as potential environmental contaminants). The isotope dilution method employs acidic destruction of the sample matrix and analyses with LC/MS. A SOP has been produced and the complete method was validated using C60 fullerenes in olive oil. Analysis of ~20 commercial olive oils did not show the presence of C60 fullerenes in these oils.
Potential Impact:
As a result of the efforts described above, analytical tools are available which allow the robust detection and characterisation of man-made NPs. The technology transfer via dissemination activities such as presentations and publications, the public website, e-newsletters, NanoLyse Open Days and policy support workshops, but in particular also via stakeholder directed hands-on training workshops allowed the direct implementation of the developed methods in practice. On the one hand this facilitates studies into the oral toxicity of ENP, on the other hand it empowers food safety authorities to monitor food products (as well as other commodities such as cosmetics) for the presence of ENP. In this way, NanoLyse provides building blocks for a proper assessment of the risk of man-made NPs. Consumer confidence in the safety of novel foods is crucial for the further exploitation of this market, a key focus of the creation of added value in the knowledge based bio-economy.
Furthermore, the experience and results obtained in NanoLyse provide impulses to the analytical industries. On the short term the commercialisation of screening assays such as ELISA and SPR sensor test kits for food safety applications is likely once respective regulations come into force. In a mid-term view, the automated electron microscopy smart imaging tool may also be developed into food laboratory routine instrumentation, upon a massive increase in the need for ENP screening, e.g. after regulatory introduction of maximum residue limits for certain ENP. Field flow fractionation combined with ICP-MS will probably become a workhorse for the determination of inorganic ENP. Single particle ICP-MS was shown to be a suitable routine tool. The experiences from the project have already resulted in the improvement of existing instrumentation by manufacturers of FFF and ICP-MS machinery. Some of the NanoLyse results are picked up and developed further in recently started FP7 projects (e.g. NanoDefine).
List of Websites:
Project website: www.NanoLyse.eu
Contact details Coordinator:
Dr. Stefan Weigel
RIKILT - Institute of Food Safety, Wageningen UR
PO Box 230
6700 AE Wageningen
The Netherlands
Tel: 0031 317 485518
Fax: 0031 317 417717
email: stefan.weigel@wur.nl
The NanoLyse project was set up to develop methods for the reliable determination of nanoparticles in food matrices. The need for such methods resulted from the increasing interest into the presence of nanomaterials in food, both from risk assessors and regulators, but also from industry in expectation of upcoming regulations.
NanoLyse established a dual, tiered approach that is consistent with current procedures for food safety monitoring. Tier 1 involves rapid screening and imaging methods for a cost-efficient yes/no decision on the presence of nanoparticles in the sample. Tier 2 provides sophisticated coupled separation and detection methods for the unambiguous identification and reliable quantification of nanoparticles in food and beverage matrices. Validation is based on suitable guidelines and makes use of stable and homogeneous reference materials. This concept was successfully translated into practice by the development and validation of methods for selected nanoparticles and complemented by the preparation of respective reference materials. The project accomplished major achievements in the following areas:
- Screening: Rapid, easy to use and cost-efficient screening methods were developed, based on two different platforms. This involved ready to use ELISA kits for the detection of cross-linked gelatine nanoparticles and a ß-lactoglobulin based nanocarrier for food additives in beverages, furthermore surface plasmon resonance (SPR) biosensor assays for these organic nanoparticles as well as an SPR assay for silver nanoparticles.
- Imaging: Electron microscopy (EM) methods for inorganic nanoparticles in food (i.a. silver particles in meat and silica particles in soup) were developed with emphasis on an optimised sample preparation for these complex matrices. An object-based image analysis system for the (semi) automated smart analysis of large numbers of EM images was delivered, enabling rapid and cost-effective EM data evaluation.
- Methods for inorganic particles: Field flow fractionation coupled online with light scattering and ICP-MS detectors was used for quantitative, element specific determination of silver nanoparticles in meat and silica in soup. Additionally, a single particle ICP-MS method for the determination of number-based size distribution of Ag particles was developed and subjected to an interlaboratory method performance study.
- Methods for organic particles: Organic nanoparticles differ significantly from inorganic ones, in particular, they are much more fragile. Care has been taken to develop adequate methods and yielded a universal method for protein-, carbohydrate- and lipid-based NP that is based on separation by hydrodynamic chromatography, quantification by UV absorbance detection and identification by MALDI-ToF-MS fingerprinting. This was complemented by methods for a fast screening for nanoparticle components based on DART-MS and UPLC-Orbitrap MS as well as a parallel GEMMA method for particle number concentrations.
- Reference materials: Stable, homogeneous and well characterised reference materials (pairs of pure suspensions and matrix materials) were prepared. Labelled Ag and SiO2 particles were synthesised for use as internal standards. All developed methods were validated according to a generic validation approach drafted in the project.
Dissemination of the project results to the scientific community as well as to regulatory and industrial stakeholders was realised via presentations at conferences, scientific journals, NanoLyse Open Days as well as policy support workshops. Technology transfer to potential end-user was achieved by hands-on training workshops.
Project Context and Objectives:
Nanotechnology applications for the food sector are intensively investigated and developed at the moment. A number of nanomaterials are already in use as food additives or in food contact materials. Furthermore, approved (bulk) food additives may have a size distribution which extends down to the sub 100 nm range, e.g. fumed silica (E551). At the same time, limited knowledge is available on the potential impact of engineered nanoparticles (ENP) on consumers’ health. A prerequisite for toxicological, toxicokinetic, migration and exposure assessment studies is the availability of analytical tools for the detection and characterisation of ENP in complex matrices such as food.
At the starting point of the project, the availability of methods for the detection and characterisation on engineered nanoparticles in complex matrices, and in particular in food, was very limited. Some approaches existed for the measurement of inorganic particles in environmental matrices and electron microscopic imaging in e.g. tissues while no methods were available for screening of any ENP and for the measurement of organic ENP in food.
The NanoLyse project focussed on the development of validated methods and reference materials for the analysis of engineered nano-particles (ENP) in food and beverages. The developed methods were to cover relevant classes of ENP with reported or expected food and food contact material applications, i.e. metal, metal oxide/silicate, and organic ENP. Priority ENPs have been selected out of each class as model particles to demonstrate the applicability of the developed approaches, e.g. nano-silver, nano-silica, a protein-based nano-carrier system as well as fullerenes (potential food contaminants). Priority was given to methods which can be implemented in existing food analysis laboratories. Rapid imaging and screening should be achieved by the application of automated smart electron microscopy imaging as well as screening techniques in sensor and immunochemical formats. More sophisticated, hyphenated methods were intended to allow the unambiguous characterisation and quantification of ENP. These include elaborate sample preparation, separation by flow field fractionation and chromatographic techniques as well as mass spectrometric and electron microscopic characterisation techniques. The developed methods were validated using the well characterised food matrix reference materials that were produced within the project. In summary, this includes the following objectives:
• Reference materials for the analysis of nanoparticles in food and beverages
• Sample preparation methods for the detection of nanoparticles in food
• Rapid imaging and screening methods for nanoparticles in food
• Analytical methods for the identification, characterisation and quantification of inorganic and organic nanoparticles from the food matrix
• Dissemination and training of the new methods to relevant stakeholders
The European Commission published a recommendation for a definition of a nanomaterial in October 2011, halfway of the duration of the project. This number-based definition shifts the analytical requirements to particle counting methods. NanoLyse reacted to the resulting changed needs of stakeholders by replacing a mass based method for inorganic ENP by an easy implementable particle counting method (single particle ICP-MS), in addition to the more laborious electron microscopic techniques foreseen in the original work plan.
Project Results:
Methods for the measurement of nanoparticles (NPs) in food were virtually not available at the start of the project and there were strong doubts if reliable approaches can be developed for such complex matrices as food. NanoLyse succeeded, in 3,5 years’ time, not only to demonstrate that nanoparticles can be detected, identified and even quantified in food, but also delivered a number of fully validated methods for relevant NP in realistic food and beverage matrices. These results were achieved within 4 distinct RTD work packages (WP) that closely collaborated. WP1 (Reference Materials) supplied well characterised NP suspensions and NP spiked matrix materials, synthesised labelled analogues and provided a general validation approach for NP analysis in food. WP2 (Imaging and Screening methods) developed electron microscopy imaging methods and an automated image analysis system as well as screening assays in biosensor and ELISA format. WP3 was dedicated to the development of sophisticated coupled separation and detection methods for inorganic NP, while WP4 had the same task for organic NP. The main achievements are presented per WP below.
REFERENCE MATERIALS (WP1)
- NP suspensions
The initially foreseen target NP were thoroughly tested for their suitability as future reference materials, involving criteria such as stability and well defined material characteristics (e.g. size distribution), but also the possibility to measure them with the foreseen analytical techniques. As a result of this revision, some target particles had to be replaced by others and the final target NP to be used in the project were: silver NP (example for a metal NP), silica NP (oxidic NP), cross-linked gelatine NP (organic NP) and C60 fullerenes (carbon NP).
Silica nanoparticle (SiO2 NP) suspensions and silver nanoparticle (Ag NP) suspensions were prepared from commercially available materials. They were tested for homogeneity and stability and characterised for particle size by dynamic light scattering (DLS), electron microscopy (EM), field-flow-fractionation-inductively coupled plasma mass spectrometry (FFF-ICPMS), gas-phase electrophoretic molecular mobility analysis (GEMMA) as well as for total silica or silver content by ICPMS or k0-neutron activation analysis (k0NAA). Gelatine suspensions and C60-fullerene solutions were tested for homogeneity and stability by DLS (gelatine) and liquid-chromatography-mass spectrometry (LCMS) (fullerenes). The gelatine particles were characterised for size by GEMMA, DLS, TEM and AFM.
- Labelled analogues
Reliable quantification of the target analytes is a major issue in the statutory monitoring for food safety. (Isotopically) labelled internal standards play a major role for this purpose in the analysis of conventional analytes (molecules, ions). NanoLyse established the same approach for NP analysis by the synthesis of labelled analogues of the target analytes. Au labelled Ag NPs were produced from a mixture of AgNO3 and HAuCl4 by reduction with NaBH4. The suspension was purified by centrifugation and washing with distilled water. The UV-VIS spectrum showed only one peak, confirming incorporation of the Au in the Ag matrix. Transmission electron microscope (TEM) images showed roughly spherical particles with particle sizes ranging from 4 to 24 nm. X-ray photoelectron spectroscopy was used to quantify the Au/Ag ratio and particles were found to contain 2 % (Mol/Mol) Au. The particles are readily suspensible in water.
GeO2 labelled SiO2 was produced from a mixture of tetraethoxysilane and germanium ethoxide. The material was purified by centrifugation and repeated extraction with ammonium nitrate in ethanol. Scanning electron microscope (SEM) pictures show spherical particles of approximately 100 nm. X-ray photoelectron spectroscopy was used to quantify the Ge/Si ratio and particles were found to contain 0.8 % (m/m) GeO2. The dried powder was readily dispersible in ethanol, but dispersion in water required sonication for more than 20 minutes. Therefore, a second batch of material was produced which was not dried to a powder, but was distributed as paste instead.
Fluorescent labelled gelatine was produced using a commercial protein labelling kit and fluorescein hydrochloride (DTAF). The outcome was investigated by confocal microscopy. Localised fluorescent points show that the labelling with the fluorescent dye was successful.
13C labelled C60 fullerenes are commercially available.
- Food matrix reference materials
Four homogeneous and stable matrix reference materials (RM) for Ag NP in chicken meat and SiO2 NPs in tomato soup were produced. The chicken meat materials were characterised for total Ag content by k0NAA and for Ag NP size by single particle-ICPMS (sp-ICPMS) and transmission electron microscopy (TEM). The soup material was characterised for total silica content by ashing and for SiO2 NP size by FFF-ICPMS. The development of a homogeneous and stable suspension of gelatine NPs in beverage was not successful, presumably due to interactions of the gelatine NPs with the gum Arabic in the beverage. An approach for the production of a homogeneous solution of C60-fullerenes in oil was developed. These materials are complemented by the corresponding neat NP dispersions (see above), which gives users the unique opportunity to compare results on matrix materials after sample preparation with the results on the NPs before adding to the matrix. The reference materials are available to other interested parties and projects after the end of NanoLyse as long as stocks last.
- Validation approach
A generally applicable validation approach for the analysis of NP in food (and other complex matrices) was developed. It was based on generally accepted principles in analytical chemistry as laid down in international guidelines, taking into consideration potential differences arising from the particulate matter of the measurands. This is the first published systematic approach for the validation of methods detecting and quantifying NP. The validation approach was applied in the validation of methods developed in the project and refined based upon the resulting experiences. Furthermore, the validation approach was posted on the public project website and comments from external experts invited to further develop the proposed approach.
RAPID SCREENING AND IMAGING METHODS (WP2)
- Sample preparation for electron microscopy (EM)
Appropriate sample preparation is an essential first element of any analytical method. It is important for nanoparticle detection/characterisation methods to ensure that there is no significant change in the state of nanoparticles due to dissolution or agglomeration/ aggregation in order to avoid erroneous analytical results. As water is the main component in the majority of food products, its removal as required for EM analysis can also trigger changes to the nanoparticles. In NanoLyse different sample preparation methods were compared for suitability for unperturbed preparation of nanoparticle containing dispersions and food matrices. The methods included air drying, freeze drying, sedimentation by ultracentrifugation, removal of water by blotting, drying by chemical means, resin embedding, snap freezing, etc. The evaluation showed that each method has certain advantages and disadvantages for specific purposes. Results of validation showed ultracentrifugation and blotting to be the most suitable methods for sample preparation due to the least variation in nanoparticle sizes when measured by EM. A respective guidance document was issued and the experiences used to optimise the methods developed in the project.
- EM imaging methods
Imaging approaches based on electron microscopy (EM) have been thoroughly tested and validated against other analytical methods studied under NanoLyse for detection/characterisation of nanoparticles in dispersions and selected food matrices. The data and information generated provides a comprehensive insight to the capabilities and limitations of the EM methods. Based on these results dedicated methods for the NanoLyse analyte/matrix combinations were developed, in particular a scanning EM (SEM) method for silica NP in tomato soup and a transmission EM (TEM) method for Ag NP in chicken meat.
- Automated image analysis system
A major bottleneck in the use of imaging methods for high throughput analyses is the time and effort needed to manually analyse each individual EM image. To address this problem, NanoLyse developed a software-based platform capable of processing large numbers of EM images to derive data on particle numbers, size distribution, and certain morphological (shape) characteristics. The software solution was developed in the form of an object-based image analysis system based on eCognition® Architect using the Cognition Network Technology®. Rule sets can be applied and parameterised using the developed “Architect-Solution” which in this case would reflect a typical workflow of object based image analysis (OBIA) to detect and count nanoparticles. The system is easy to use and guides the user step by step through the workflow. It also allows for parameterisation wherever necessary and to test and exclude intermediate results. This way, the user can focus on relevant parameters for object detection and can generate report statistics in simple steps. In case the results are not precise enough, the user can either re-parameterise single steps of the solution or correct them manually in the image at any time.
Although the system was originally intended to be developed as a fully automated platform, following testing on nanoparticle images in dispersions and food matrices of increasing complexity, as well as variable quality of images (brightness, contrast, focus) it was deemed more appropriate to retain a degree of expert oversight and thus make the system semi-automated. The developed solution modules were validated for analysis of EM images of increasing complexity in food matrices. The developed OBIA system was tested against currently available software based on pixel based imaging analysis (PBIA).
- Enzyme-linked immunosorbent assay (ELISA) for organic NP
ELISA is a robust and cost-efficient technique for the rapid analysis of i.a. food samples. Within NanoLyse, it was demonstrated that this technique can also be used for the detection of NP in food and beverage samples. It is in particular suited for organic NP which are often more difficult to analyse by other techniques due to their fragile nature. Prerequisite is that they possess immunogenic properties that can be used to raise specific antibodies. This fact excluded some of the originally foreseen target particles (organic nano-encapsulates of which the shell compounds did not trigger an immune response). Antibodies were successfully raised against cross-linked gelatine NP (a potential carrier for food supplements or flavours). These antibodies are highly specific to the cross-linked gelatine NP and do not cross-react with different native forms of gelatine. They were used to develop a screening assay for the detection of cross-linked gelatine NP in beverages (juices, soft drinks) with detection limits of 250 – 500 ug/L. The method was validated for limits of detection (LoD), repeatability, precision, and robustness according to 2002/657/EC and the validation guidelines developed in WP1. A prototype ELISA kit for gelatine NP was manufactured. Furthermore, an existing antibody against native ß-lacto globulin was shown to recognise also nanoparticulate ß-lacto globulin complexes with epigallocatechin gallate (EGCG, an antioxidant food supplement) and an assay based on this antibody was developed and validated.
- Surface plasmon resonance (SPR) biosensor assays for silver and organic NP
SPR biosensors allow a label-free high throughput screening and are used for the detection of residues and contaminants in food safety monitoring. Assays for silver NP and organic NP were developed and validated in the project. The assay for silver NP is based on a metal-binding protein (metallothionein MT1) that was immobilised on the sensor chip surface. Although this metallothionein also binds silver ions and other metals the selectivity towards Ag NP was increased by an optimised composition of the buffer and a size specific filtration step in the sample preparation. The assays for organic NP make use of the antibodies that were also used for the ELISAs. The developed assays showed adequate sensitivity in the low ppm range and were tested in different fruit juices and carbonated and non-carbonated flavoured mineral water. For ß-lacto globulin NPs a selective rapid, 2 minutes, sample-pre-treatment step was developed to further improve the assay’s selectivity towards the nanoform of ß-lacto globulin, using disposable spin-through size exclusion cartridges. Total assay time was about 12 minutes and the sensor chips showed high robustness and stability over a period of several months. The utilization of the sensor described here for detection of protein-based NPs in food, can facilitate rapid and automated screening needed for efficient food safety monitoring
SEPARATION AND CHARACTERISATION METHODS FOR INORGANIC PARTICLES (WP3)
- Sampling for NP analysis
Sampling protocols were developed for inorganic engineered nanoparticles in at least three food matrices. The issue of preparing a representative sub-sample was addressed by referring to different ISO standards and to a NMKL guideline on the subject for sampling in the context of the NanoLyse project of olive oil, soup and meat. Different types of distributions were considered, which were assumed to be relevant for count data from experimental methods detecting nanoparticles. The Poisson distribution was chosen as one of the most common discreet distributions, which are frequently used to describe count data. Calculations have been made for silver and silica nanoparticles regarding the minimum sampling volumes necessary for having a representative sub-sample. The calculations showed that even at low concentrations (i.e. in the 10 ng/ml range) the number of sampled nanoparticles in very low volumes (a few microliters. i.e. lower than typically taken by for example pipetting) was sufficient to allow for representative sub-sampling when working with nanoparticles in the size range of 1 to 100 nm.
- Quantitative multidetector FFF method for Ag NP in meat
A method has been developed that is capable of extracting Ag NP from chicken meat samples, determine their size and quantify their mass amount. After comparison of different approaches enzymatic digestion was selected and optimised for sample preparation. Recovery rates were > 80%. The analytical separation of the Ag NP was achieved by asymmetric flow field flow fractionation (AF4). A systematic optimisation approach for a large set of parameters in FFF (carrier composition and concentration, cross flow, injection flow and time, injected mass, membrane type and spacer height) was developed and applied. The resulting method provides optimal recovery, separation power, retention time, peak shape, and run time for the target analyte. Separation has been coupled on-line with a suite of three different detectors (UV/Vis absorbance at a wavelength of 400 nm, one angle (90°) of the MALS detector and the 107Ag trace of the ICPMS. ICPMS was used for the mass based quantification of the Ag NP in samples. The overall method was validated for precision, trueness (recovery), limit of quantification and ruggedness. A small scale inter-laboratory comparison of the method demonstrated that it is transferable to well-trained laboratories. A standard operating procedure (SOP) for the isolation, detection, characterization and quantification silver nanoparticles (Ag-ENPs) in lean chicken meat by AF4 was provided.
- Quantitative multidetector FFF method for silica NP in soup
A method has been developed that is capable of extracting silica NP from tomato soup samples, determine their size and quantify their mass amount. After comparison of different approaches acid digestion was chosen with the aim to completely oxidize the organic matrix with only minor effects on the SiO2-NPs. Applying a sample pre-treatment with heating and homogenization treatment followed by acid digestion proved to be suitable for extraction of SiO2-NPs from tomato soup. Si mass recoveries were in an acceptable range (> 90%). The analytical separation of the silica NP was achieved by asymmetric flow field flow fractionation (AF4). The same systematic optimisation approach as described for Ag NP has been applied and resulted in a method with an optimal compromise between recovery, separation power, retention time, peak shape, and run time for silica NP. The FFF separation was coupled on-line with light scattering (MALS) and ICP-MS detectors. MALS data was used to determine the rms-radius applying the Debye fit (3rd order) while ICP-MS was applied for the mass based quantification of the silica NP. The overall method was validated for precision, trueness (recovery), limit of quantification and ruggedness. A standard operating procedure (SOP) for the isolation, detection, characterization and quantification silica particles (SiO2-ENPs) in tomato soup by AF4 was provided.
- Particle counting single particle ICP-MS method for Ag NP
A particle counting method for Ag Np was developed in response to the new EC recommendation for a definition of nanomaterial. The method is capable of extracting Ag NP from a meat matrix and to measure the mass of individual particles in the resulting suspensions. Subsequently, number-based size distributions can be determined. Enzymatic digestion and dilution in water was used for sample preparation. Detection was achieved by ICP-MS in time resolved mode. The method was validated first in a single laboratory approach and then subjected to a full interlaboratory method performance study with 23 participants. The results indicate that single particle ICP-MS is a suitable (screening) technique for silver NP. A standard operating procedure (SOP) as well as a data evaluation macro were provided.
SEPARATION AND CHARACTERISATION METHODS FOR ORGANIC PARTICLES (WP4)
Organic NP are usually more fragile and complex in nature as compared to most inorganic NP. Typical structures for such NP are micelles, liposomes, encapsulates and complexes build from large or macromolecules (proteins, carbohydrates, lipids) carrying small molecules (e.g. vitamins, antioxidants, flavours, preservatives). This requires from analytical methods that they leave the NP structures intact and allow a complex chemical characterisation (in contrast to inorganic NP where elemental composition often is sufficient). This was achieved in WP4 by a number of different approaches.
- Universal method for organic NP based on HDC-UV-MALDI-TOFMS
This method was developed for the size determination, quantification and characterisation of organic NP of different nature (protein-, carbohydrate-, lipid-based). Sample preparation from beverage matrices was realised by the application of filtration and size exclusion chromatography. The analytical separation and size determination (after calibration with respective standards) is achieved by hydrodynamic chromatography (HDC). The on-line coupled UV detector allows the reliable quantification of the NP amount. The unambiguous chemical identification is carried out by means of an off-line coupled MALDI-ToF-MS (fingerprinting of the NP). The method has been developed and tested for different types of organic NP. A full validation according to EC/2002/657 was exemplarily carried out for the determination of a commercial lipid-based nanoparticle in beverages and a standard operating procedure (SOP) was provided.
- GEMMA
Differential mobility analysis was combined with an electrospray unit for particle ionisation (GEMMA). This combination is very successful in particle size separation and the production of number-based size distributions according to the requirement of the EC definition of a nanomaterial. A SOP was produced for the combination of GEMMA with a condensation particle counter (CPC). Validation of the method shows that GEMMA is useful as a qualitative method for ENP characterisation.
- Mass spectrometry based screening for organic NP
Two rapid methods for a decision on the presence of target NP in food/beverage matrices have been developed as an alternative to the more laborious HDC and GEMMA methods. Components of the organic NP (e.g. building blocks of micelle shells or active ingredients encapsulated inside) can be detected and quantified by the developed methods. The first option uses UHPLC-TOFMS. The second option uses DART (Direct analysis in real time) ionisation coupled to high resolution (Orbitrap) mass spectrometry. This is a high resolution ambient MS technique used for rapid fingerprinting/profiling directly from the sample without sample preparation or chromatographic separation. Nanoparticle components in the sample can be detected on the basis of characteristic fragment ions. Both methods have been validated for the direct determination of commercial micelle-type polysorbate based ENPs in beverages.
- Method for fullerenes in edible oils
A method was developed for the determination of C60 fullerenes in vegetable oils (as potential environmental contaminants). The isotope dilution method employs acidic destruction of the sample matrix and analyses with LC/MS. A SOP has been produced and the complete method was validated using C60 fullerenes in olive oil. Analysis of ~20 commercial olive oils did not show the presence of C60 fullerenes in these oils.
Potential Impact:
As a result of the efforts described above, analytical tools are available which allow the robust detection and characterisation of man-made NPs. The technology transfer via dissemination activities such as presentations and publications, the public website, e-newsletters, NanoLyse Open Days and policy support workshops, but in particular also via stakeholder directed hands-on training workshops allowed the direct implementation of the developed methods in practice. On the one hand this facilitates studies into the oral toxicity of ENP, on the other hand it empowers food safety authorities to monitor food products (as well as other commodities such as cosmetics) for the presence of ENP. In this way, NanoLyse provides building blocks for a proper assessment of the risk of man-made NPs. Consumer confidence in the safety of novel foods is crucial for the further exploitation of this market, a key focus of the creation of added value in the knowledge based bio-economy.
Furthermore, the experience and results obtained in NanoLyse provide impulses to the analytical industries. On the short term the commercialisation of screening assays such as ELISA and SPR sensor test kits for food safety applications is likely once respective regulations come into force. In a mid-term view, the automated electron microscopy smart imaging tool may also be developed into food laboratory routine instrumentation, upon a massive increase in the need for ENP screening, e.g. after regulatory introduction of maximum residue limits for certain ENP. Field flow fractionation combined with ICP-MS will probably become a workhorse for the determination of inorganic ENP. Single particle ICP-MS was shown to be a suitable routine tool. The experiences from the project have already resulted in the improvement of existing instrumentation by manufacturers of FFF and ICP-MS machinery. Some of the NanoLyse results are picked up and developed further in recently started FP7 projects (e.g. NanoDefine).
List of Websites:
Project website: www.NanoLyse.eu
Contact details Coordinator:
Dr. Stefan Weigel
RIKILT - Institute of Food Safety, Wageningen UR
PO Box 230
6700 AE Wageningen
The Netherlands
Tel: 0031 317 485518
Fax: 0031 317 417717
email: stefan.weigel@wur.nl