Service Communautaire d'Information sur la Recherche et le Développement - CORDIS

FP7

NANOGNOSTICS Résumé de rapport

Project ID: 242264
Financé au titre de: FP7-HEALTH
Pays: Germany

Final Report Summary - NANOGNOSTICS (Quantum Dot-Based Highly Sensitive Immunoassays for Multiplexed Diagnostics of Alzheimer's Disease)

Executive Summary:

6.1 million people currently live with a form of dementia in the European Union with an addition of 1.4 million new cases every year. Because the diagnosis of Alzheimer’s disease (AD) is complicated and time consuming, there’s a great demand for a rapid, sensitive and specific immunoassay for protein markers inside blood to shorten the timeframe of diagnosis and support the treatment of the disease. The NANOGNOSTICS project, taking place from October 2009 until March 2013, focused on developing a multiplexed homogeneous immunoassay based on FRET from one dye labelled specific antibody (AB1) to another (AB2) within an “AB1-biomarker-AB2” assay architecture or an aptamer assay utilizing lanthanide complexes (LC) and quantum dots (QD) as FRET donors and acceptors, respectively. During the project significant progress could be achieved in a variety of fields. New water soluble QDs were obtained via a coating procedure with an amphiphilic polymer, which wrappes around the QDs. The entirely present carboxylic groups at the polymer backbone allowed for a further functionalization of the QDs with biomolecules or LCs. New highly luminescent LCs with Terbium and Europium as central ion and a variety of different structures were synthesized, functionalized with biomolecules and established in different fluorescence immunoassays, e.g. for the AD biomarker EGF (epidermal growth factor).
Besides developing new FRET donors and acceptors research also concentrated on new assay architectures with antibodies and aptamers. Here we could develop different new assays, e.g. a QD FRET-based immunosensor for casein kinase, a general new sensing platform based on QDs and Chemiluminescence Resonance Energyx Transfer (CRET) or an aptamer-based sensing platform for the Vascular Endothelial Growth Factor (VEGF) utilizing FRET or CRET. Additionally, we used commercial available components to develop a multiplexed immunoassay, which allows for simultaneous detection of up to 5 biomarkers while achieving very low limits of detection (LODs), which are comparable to commercially available single marker assays. Besides utilizing FRET for the detection of biomarkers, we also contributed to a profound understanding of the FRET between LCs and QDs in general. Extensive structural and photophysical characterization of the LCs and QDs was then followed by characterizing the FRET for various systems and distances between LCs and QDs. Thus we could determine FRET parameters that were unknown before and where assumptions had to be made in order to theoretically describe the FRET. In order to implement the developed FRET-based immunoassays, a multi-modality, high throughput immunoassay plate reader meeting the design goals and specifications requested by the NANOGNOSTICS partners was developed. The plate reader provides a plethora of research quality data which could prove invaluable for many research and assay development applications.
During the project the research in the field of AD specific biomarkers was continuously monitored. Patient samples (plasma and if possible cerebrospinal fluid) were collected and analyzed. We successfully established several ELISA based methods for potential relevant AD biomarkers from peripheral blood. Unfortunately we were not able to confirm literature data (Ray S. et al., Nature Medicine, 2007). In agreement with Björkqvist et al. (PLoS ONE, 2012) we conclude that the panel of 18 proteins suggested by Ray et al. is not suitable for AD detection. Nevertheless, highly sensitive and specific detection of VEGF demonstrated the feasibility of the NANOGNOSTIC spectroscopic platform technology for biomarker detection. First results from ongoing research indicate that plasma levels of complement proteins might be indicative for AD development.
Project Context and Objectives:
Combination of psychological testing, brain-imaging and exclusion of other neurological disorders makes the diagnosis of Alzheimer’s disease (AD) complicated and time consuming (taking up to 20 months). A rapid, sensitive and specific immunoassay for protein markers in whole blood or plasma would therefore largely improve early diagnosis as well as therapy- and disease progression-monitoring for the benefit of clinicians (fast, easy and inexpensive) and, more importantly, of patients (uncomplicated diagnosis, better treatment). Homogeneous assays based on FRET (Förster or Fluorescence Resonance Energy Transfer) from one dye labelled specific antibody (AB1) to another (AB2) within an “AB1-biomarker-AB2” immune complex are an ideal basis to meet these challenging requirements of in vitro diagnostics. They do not require any washing or separation steps, fast solution-phase kinetics allow short incubation times and time-resolved detection permits nearly background-free measurements. Moreover, the ratiometric format (luminescence detection of FRET donor and acceptor) offers an instantaneous suppression of sample or measurement fluctuations resulting in an extremely good reproducibility (very low coefficient of variation). Thus, homogeneous assays possess unrivalled properties for high throughput screening (HTS) as well as point-of-care (POC) or small laboratory testing.
Instead of antibodies for biomarker binding, aptamers can also be used. Compared to antibodies, aptamers have similar affinity for their targets and possess many advantages such as cost-effective and highly reproducible chemical synthesis in vitro, easy labelling, less nonspecific adsorption and more flexibility for binding sites of the target. Thus aptamers are extremely interesting for homogeneous immunoassays and by choosing a parallel investigation with anti-bodies as well as aptamers we are increasing flexibility and scientific impact by a concomitant decrease of risk.
As the detection of several protein markers is obligatory for a highly sensitive and specific diagnosis an optical multiplexing approach with dyes of different colours is an excellent solution. Semiconductor quantum dots (QDs) are the ideal candidates due to their size-dependent absorption and emission. Moreover, they possess unique photophysical properties (e.g. high absorbance and photostability) that overcome conventional fluorescence dyes. In combination with lanthanide complexes (LCs) QDs render a powerful multiplexing tool for highly sensitive diagnostics even for large immune complexes. QD luminescence lifetimes can reach up to microseconds, which is short compared to the LC lifetimes (up to milliseconds) but long compared to organic fluorophores (up to few nanoseconds). When the luminescence lifetime of the FRET donor is significantly longer than the decay time of the background luminescence (directly excited acceptor and autofluorescence of the sample) time-resolved detection (or time gating) allows for efficient background suppression. Thus, QDs can be used both as donors (with organic fluorophores) and acceptors (with LCs) for highly sensitive time-resolved immunoassays. One big advantage of LCs is their broad emission wavelength spectrum with well separated and narrow bands enabling the possibility of using one single LC type as donor for different QD acceptors. Figure 1 shows the multiplexing immunoassay concept, where FRET from one LC donor (D) to several different QD acceptors (A1-AX) can only occur when the labelled antibodies are in close proximity due to antibody binding on the specific protein markers (1-X). The FRET process results in new long-lived QD luminescence signals at different wavelengths, which allow for direct quantification of the various marker proteins inside the diagnostic test. This immunoassay provides not only multiple specificity (due to antibody binding and multiplexing) and high sensitivity (luminescence measurement), but is also very fast.
The time from blood withdrawal to diagnosis can become shorter than 10 min (Figure 2) if detection in whole blood (ca. 1 – 2 orders of magnitude higher detection limits compared to detection in serum or plasma) and fast binding kinetics are possible. This makes our technology very suitable for POC testing with 1 or few samples. For highly sensitive assays measurement in plasma (or serum) longer incubation times might be required. In this case HTS (hundreds to thousands of samples) is the method of choice.
Our immunoassay system will bring many advantages to both POC as well as HTS. By offering a modular spectroscopic platform we meet the demands of many in vitro diagnostic applications. Moreover we can adapt the system to our optimal requirements in order to achieve full control over the multiplexing assays with high sensitive-ties for AD. Once the technology is working as expected an integration e.g. into POC tests and the adaption to other markers (e.g. cancer or cardio-vascular markers) can be achieved relatively easily. Although QDs offer all the important properties for a successful use in diagnostics this is to date restricted to academic research. Diagnostic QD-based FRET applications are very rare in particular when using QDs as FRET acceptors and a commercial QD-based FRET immunoassay is not available. A complete understanding of QD-based FRET will be a big step towards the successful integration into commercial applications. For a comprehensive analysis the use of lanthanide complexes is mandatory, because they are the only known donors for efficient FRET to QD acceptors.
NANOGNOSTICS strived for a profound understanding of QD-based FRET, the synthetic creation of highly efficient QD-based FRET immune sensors for detection of several Alzheimer-specific biomarkers, and the development of a modular high-throughput-screening immunoanalyser for the integration of QD-based multiplexing immunoassays into early diagnosis for improved patient outcome in dementia therapy. In order to fulfil these goals we established a multidisciplinary consortium with 9 partners from Germany, Finland, UK, France and Israel. For the understanding of QD-based FRET and the multiplexing immunoassay development the partners provide an internationally recognized expertise in nanoparticles/QDs and their application to bioanalysis (PUM and HUJI), antibody and aptamer based FRET bioassays (HUJI), lanthanide complex synthesis and their bioanalytical applications (CNRS and UTU) and diagnostic and bioanalytical FRET spectroscopy and microscopy (Fraunhofer, UPS, UP and UTU). Strengthening the competitiveness of European industry by focusing on applied research for the benefit of industrial partners (Fraunhofer) is another important aspect of the project. One SME partner contributed highly advanced skills in technical development of ultra-sensitive time-resolved spectroscopy and novel light sources (EI). The development of a modular immune analyzer for commercialization by the SME partner was planned. In order to successfully integrate the novel multiplexing assays into clinical applications for the benefit of patients an end user of the technology was integrated in the consortium (Charité), which is one of the largest university clinics in Europe having access to a large amount of patient samples and clinical as well as neurological expertise for a profound “real life” study of Alzheimer diagnostics with the QD-based immunoassays.
The immunoassay system we wanted to develop within NANOGNOSTICS should be an innovative tool to identify alterations of protein expression patterns caused by AD. As we chose an extremely sensitive (quantity) and reproducible (quality) luminescence detection approach with colour multiplexing and the measurement is performed in patient whole-blood or plasma samples the technique will be able to provide an early, specific and parallel detection of various protein markers for diagnostic, prognostic and monitoring purposes without the common inconveniences the patients have to suffer from current Alzheimer analysis. We will quantitatively identify the most relevant proteins expressed due to AD and a distinction between close variants will be possible due to identification through highly specific antibodies and aptamers.

Project Results:
Synthesis and functionalization of QDs (WP1)

To transfer inorganic QDs, which are carrying hydrophobic surfactants, to an aqueous solvent, an amphiphilic polymer was prepared. The polymer consisted of poly(isobutylene-alt-maleic anhydride) (MW ≈ 6000 g/mol) which was reacted with dodecylamine to an extent that leaves 25% of the anhydride rings unreacted. Thus a polymer, which possesses hydrophobic side chains, which function as counterpart to the QD surfactants, as well as carboxylic groups, which enable water solubility, was formed (Figure 3).
A mixture of QDs and polymer (both diluted in chloroform) was stirred for about 10 minutes followed by a complete evaporation of the solvent under reduced pressure. The obtained solid film was diluted in 50mM sodium borate buffer at pH = 12 (SBB pH 12), so that the QDs were finally transferred to an aqueous solvent. The basic pH leads to deprotonated carboxylic groups so that the QDs are colloidally stable. To purify the QDs from excess polymer strands gel electrophoresis was the medium of choice (Figure 4). Herein, the QDs were loaded into a 2% agarose gel and an electric field of 10 V/cm was applied for a certain time (usually 60 to 120 min). Thus the polymer strands, which usually are migrating faster through the gel compared to the QDs, can be efficiently separated from the QDs. To extract the purified QDs from the gel matrix they were cut out from the gel and transferred into a dialysis membrane and again the same voltage was applied.
To link functional molecules, e.g. organic dyes which can function as FRET acceptors, to QDs, the mentioned polymer was pre-modified before the coating process. The anhydride rings serve as anchor points and primary amines can be attached forming an amide bond once a ring was opened. This way of functionalization is usually performed during the attachment of hydrophobic side chains. In the present case ATTO-590 (absorption maximum: 597 nm, emission maximum: 615 nm, ATTO-TEC GmbH, Germany) or DY700 (absorption maximum: 707 nm, emission maximum: 730 nm, DYOMICS, Germany), both derivatized with an amino group, were attached to the polymer. After the mentioned coating process (QDs emitting at 565 nm with ATTO-590 functionalized polymer and QDs emitting at 655 nm with DY700 functionalized polymer) the hydrophilic QDs can be separated from empty polymer micelles via gel electrophoresis (Figure 5).
In contrast to the presented pre-modification method also a post-modification after the polymer coating can be carried out. The choice of the modification method depends mainly on the solubility of the functional molecules. In case the molecules to be attached are not soluble in organic solvents a pre-modification becomes difficult. This behaviour was observed for the lanthanide complexes (LCs) which were delivered by CNRS and UTU (WP 2). The amino-modified LCs, namely Tb(III) and Eu(III) complexes were attached to the readily coated and purified QDs via N-(3-Dimethylaminopropyl)-N’-ethylcarbodiimide (EDC-) coupling. The LCs are potential donors regarding FRET applications. After attachment the QDs were again purified from free/unbound molecules via gel electrophoresis (Figure 6).
With the same strategy spacer molecules, here PEG, can be attached to the surface of QDs. In first order PEG has two important impacts on the QD system: 1) The colloidal stability is increased, especially in biological environments involving proteins, due to steric reasons. 2) The PEG allows for an attachement of functional molecules in a certain distance to the QD surface, which reduces surface triggered effects. Therefore the used PEG molecules should be heterofunctional in a manner that one end group allows for an efficient attachment to the QD surface (primary amine) and another end group allows for an attachment of the desired molecule. Here the strategy was to have a biotinyl derivative which allows for an attachment of streptavidin modified LCs. To ensure the desired PEG coverage of the QD surface (saturated or with a certain number of PEG molecules) the amount of EDC during the coupling reaction was varied, what resulted in more retarded bands, when analyzed via gel electrophoresis, in case more EDC was added (Figure 7).
In case the functional molecules of interest, like antibodies, possess a biotin group, streptavidin can also function as crosslinker between the two biotin fractions.
The colloidal characterization of NPs was mainly based on three methods: 1) Gel electrophoresis: In the previous paragraphs gel electrophoresis was mentioned as an efficient purification and characterization method. According to the net charge and the diameter of a certain species of NPs the migration distance in the gel can serve as a first indication of colloidal stability and/or homogeneity of the sample. Moreover NPs can be broadly purified from smaller molecules which are not attached to the NP surface. 2) Size exclusion chromatography served as further purification and colloidal characterization method. Therefore polymer coated QDs were loaded into a high-performance liquid chromatography (HPLC) system (Agilent 1200 series) for purification and analytical proof of colloidal homogeneity. 3) As a third opportunity to characterize water soluble NPs strategies based on pH titration combined with dynamic light scattering (DLS) and zeta potential measurements were developed. As the NPs are coated with an amphiphilic polymer which enables water solubility by the presence of carboxylic groups at the entire polymer backbone the pKa of the latter gives information about colloidal properties and functionalization abilities. An example for a titration curve for polymer coated 7 nm Au NPs coated is given in Figure 8. The entire profile could be fitted with a pKa of 7.2 what is 2 to 3 pH units higher compared to actual pKa values of free carboxylic acids. It was shown that the resulting higher pKa has its origin in structural conformations that cause hydrogen bonds between specific sites of the polymer backbone.
DLS and zetapotential measurements confirmed these findings as in the same pH region agglomeration as well as neutralization occurred when Au NPs were titrated against an acid. However, these methods give information about the colloidal stability at a certain pH, what would be a crucial parameter for the use of such NPs in biological systems. On the other hand they give information about the feasibility of functionalization strategies like EDC activated peptide coupling. Additionally, the amount of carboxylic groups can be estimated from the width of the buffering plateau.
To test if PEG is suitable as spacer molecule to minimize the influence of the surface on functional molecules (here in first order regarding the electrostatic influence on a sensor dye) the chloride sensitive dye Amino-MQAE was mounted to water soluble Au NPs. The free dye responds to a linearly increasing chloride concentration with exponentially decreasing fluorescence intensity (quenching). In the present system PEG spacers of variable length were mounted in-between the NP surface and the dye. With this approach a NP-dye hybrid with a short PEG spacer should map the close proximity of the NP surface regarding the local chloride concentration and a NP-dye hybrid with a long PEG spacer should map the local chloride concentration in a certain distance to the NP surface: Thus the dye response is more similar to the bulk response of the free dye the longer the PEG spacer. These expectations were proven and fulfilled experimentally (Figure 9). As the NP surface is negatively charged due to entirely present deprotonated carboxylic groups the local chloride concentration in close proximity to the NP surface is lower as in bulk due to electrostatic repulsion. With increasing distance to the NP surface the local chloride concentration increases until it reaches bulk concentration. Exactly this behavior was mapped with the presented Au NP system, what demonstrated the suitability of PEG as spacer molecule.
Another approach for an efficient linkage of functional molecules to NPs with tunable distance and orientation was verified within a so called SNAP-tag strategy. Herein guanine premodified NPs (guanine attached via EDC chemistry and a PEG spacer) are further functionalized with a model protein, namely the antibody fragment scFv consisting of variable VH and VL regions of a monoclonal antibody. First scFv was linked to O6-alkylguanine-DNA alkyltransferase (SNAP protein). The natural object of the SNAP protein used herein is to irreversibly transfer the alkyl group from its substrate O6-alkylguanine-DNA to one of its cysteine residues. However, in the present study the SNAP-scFv conjugate was reacted with the functionalized NPs under release of guanine to form SNAP-scFv-functionalized NPs (Figure 10). The resulting model object illustrates several advantages of this linkage procedure: The protein (or dye) is grafted in an oriented, irreversible and defined way (no unspecific binding), the binding occurs quickly and under physiological conditions, it excludes cross linking effects due to the monovalent recognition partners.

Synthesis and functionalization of LCs (WP2)

The synthesis and characterization of established LCs have been performed by CNRS according to the established plan within the 6 first months. All the details concerning the technical description are given in the technical annex of the activity report n°1.
The list of established LCs sent by CNRS and UTU to other partners are:
- 5 mg of NHS actived Tb LCs (Compound reference : AL030) sent by CNRS to HUJI on January the 1st 2010.
- 7 mg of amino functionalized Tb LC (Compound reference : AL044) sent by CNRS to PUM on February the 1st 2010.
- 10 mg of NHS activated Tb LCs (Compound reference AL063) sent by CNRS to Fraunhofer on April the 20th 2010.
Starting from M6, the synthesis of optimized LCs was done by CNRS. In a first step we developed a series of coordination sites containing carboxylate binding arms and an heptinoic spacer that allowed for the introduction of a labelling function (L1, L2 and L8). But it soon appeared that the corresponding complexes of Eu and Tb were not stable enough in biological media such as sera targeted for the detection of biomarkers in the blood. We then turned our attention towards the possibility to improve the stability of the complexes by the introduction of phosponate functions. A new series of prototype complexes, containing no activated functions was developed (Ligand L3, L5, L11 and L12). The introduction of phosphonated functions as coordinating arms was shown to be very efficient to improve the stability of the complexes. For the pyridylbispyrazolyl plateform L3, they were even demonstrated to be stable in bovin serum (Inorg. Chem. 2011, 50, 1689). The stability was similar in the case of the terpyridyl group (L12), weaker with the bulky indazolyl groups (L5), but extremelly high with the small pyridyl derivative (L11). Nevertheless, regarding the photophysical properties of the complexes, only those derivatives of L3 and L12 should be envisaged for the formation of highly luminescent Tb complexes.
The next logical step was then to introduce an activated function on the ligand to allow for the grafting of the complexes on biological compounds. For the bispyrazolylpyridine platform three strategies were followed for this purpose, with the formation of C-C bond in the para-position of the central pyridyl ring with an ethylene (L4), a para-carboxy-benzene ring (L7 and L10) or an amide link (L18).
The examination of the photophysical properties of the Tb complexes obtained with these ligands showed L4 to be the best candidate for further experiments, with a very long luminescence lifetime of 2.60 ms in water, a luminescence quantum yield of 24% and an excitation maximum at 328 nm (far from the absorption bands of proteins at ca. 280 nm).
In the case of the terpyridyl plateform, the introduction of the activated function was similarly envisaged according to three chemical pathways with modifications at the para-position of the central pyridyl ring: the introduction of an ethylene linker (L13), of an amide bond (L18) or of an electron donating ethoxy group (L19). Here again, the comparison of the photophysical properties of the Tb complexes showed L13 to be highly more efficient than the two other candidates (quantum yield ~ 19% in water). A global comparison of all the newly developed ligands and their complexes unambiguously showed L4 to be the best candidate for biolabeling. Considering the next step of the project, the implementation of the labels in fluoroimmunoassays, derivatives of compounds L4 were synthesized in which the activated function was introduced: L9 bearing an NHS ester for bioconjugation on lysine aminoacids of proteins and antibodies, L16 with a maleimide function for labelling of cysteine aminoacids, L15 functionalized by a terminal biotin function for fixation on streptavidin and L17 having a terminal lipoic acid function for a potential fixation of the complex directly at the surface of quantum dots. While the Tb complex of L15 has been directly engaged in FRET experiments with QDs (collaboration with UP) the biolabelling has been perform with L4 and its Tb complex.
In addition, UTU also provided different LCs using conventional (2 - 5) and novel structures (1, 6, 7).
The last task of WP2 was the implementation of the newly developed lanthanide labels into fluoroimmuno assays. For this purpose, the activated form of L9 was used for the coupling to the different biological compounds. All the labelled biomolecules were then characterized by MALDI mass spectrometry to unravel the number of chelates per biomolecules and the photophysical properties of the Tb chelates labelled onto the biomolecules were determined, with emphasis on the Tb luminescence lifetime. In a preliminary set of experiments, streptavidin was used as a reference biomolecule to assess the optimal conditions (buffer, pH, number of equivalent). As L-selectin was identified as one of the biomarkers to target, it has soon be decided to label DREG55 and DREG200 (supplied by Charité), two anti-Lselectin antibodies, which can be used within the frame of a sandwich immunoassay. Labelling ratio of 9 and 10 were obtained respectively, with luminescence lifetimes of 1.6 ms and a maximum of excitation at 328 nm. The two antibodies were provided to Partner 10 (UPS) for FRET experiments. Similarly, it was decided to label L-selectin itself in order to use it in possible competitive FRET immunoassays. Following the negative return from the first FRET experiments, it was then decided to label a pair of antibodies known to be efficient within the frame of sandwich assays. For that purpose, PSS233 and PSR222, two antibodies targeted against prostate specific antigen (PSA) have also been labelled and the labelled antibodies have been fully characterized and sent to UPS for FRET analysis. Finally, considering that troubles in the FRET efficiency should arise from the very large sizes of antibodies, the labelling of fragment of antibodies was envisaged. EgB4 and EgA1, two single domain antibodies targeted against EGF, one of the biomarker of AD, were labeled. Unfortunately in the case of these fragments of antibodies, the labelling resulted in a precipitation of the fragments which could not be redissolved. In conclusion, we have been able to develop a new family of luminescent Tb labels that were optimized for their spectroscopic properties and for their stability in biological media. The best label was then successfully used two label different antibodies, among which those targeted towards L-selectin and EGF, two biomarkers of AD.
Prepared LCs 1 to 7 were purified with HPLC and flash chromatography and characterized using techniques such as: mass spectrometry, fluorescence spectroscopy and NMR. The prepared LCs were conjugated to antibodies, EGF, aptamers and biotin. The novel structures were fully characterized to find photophysical differences between the similar chelate structures. The novel 9- and 11-dentate chelates (6b,7b) proved to possess higher absorptivity and quantum yield than the conventional 7-dentate chelate (1b). This potentially leads to improved detection schemes. Tb-chelate (4) was conjugated to EGF antibody and EGF successfully and heterogeneous and homogeneous assays were investigated. The 9-dentate Eu-chelate (3) was coupled to aptamers and purified with HPLC. The chelate was successfully run in homogeneous VEGF and bFGF assays. Structures 1c, 2, 5, 6c and 7c were conjugated to biotin and studied for streptavidin binding.
Within this WP also commercial Tb-complexes (Lumi4-Tb, see Figure 15) were successfully labelled to different antibodies (IgG, F(ab)2, Fab and nanobodies), peptides, oligonucleotides and streptavidin and photophysically characterized. The bioconjugates were then used in WP3 and WP5.

Profound understanding of QD-based FRET (WP3)

Within this WP we have been studying different FRET configurations using QDs both as FRET donors and acceptors in order to understand the different possibilities of QD-based FRET. We could show for different QDs (commercial ones and QDs from academic collaborators) and for different biomolecules (streptavidin, biotin, oligonucleotides, peptides) that QDs can be used as acceptors (with Tb), as donors (with dyes) and as simultaneous acceptors and donors in so-called FRET relays, where the Tb-to-QD FRET is the 1st FRET step followed by a 2nd FRET step from QD to dye. This work has lead to several publications in high impact scientific journals as well as the co-application to a US-patent in collaboration with the Naval Research Laboratories, group of Dr. Igor Medintz, in Washington, D.C, USA (see dissemination section for publications and patent application). Within another collaboration (Dr. Thomas Pons, ESPCI Paris, France) we could also show FRET from Tb-labeled streptavidin to biotinylated QDs using two different commercial QDs (Invitrogen by Life Technologies and eBioscience) as well as one self-synthesized QD for a detailed study of the influence on surface coatings, functionalization and luminescence quantum yields of QDs on the sensitivity of FRET bioassays (see dissemination section for publication).
These results are exemplified in Figures 16 (for the FRET relays using Tb-to-QD-to-dye FRET) and Figure 17 (for the Tb-to-dye FRET using different QDs). Figure 16 describes the principle of the FRET relay system and the different applied biomolecules. Part (A) (i) shows time-gated FRET sensitization of QD photoluminescence (PL) via FRET1. Both the Tb and QD are initially excited by a flash of UV light; however, the QD relaxes to its ground state after a suitable microsecond delay (time gate) and becomes a good FRET acceptor for a proximal long-lifetime Tb donor. Part (A) (ii) shows time-gated sensitization of A647 PL via FRET1 and FRET2. The coassembly of a fluorescent dye, A647, with the Tb around a QD permits a two-step energy transfer relay with the QD as an intermediary. Part (B) shows that the QD is able to serve as a nanoscaffold for the controlled assembly of biomolecules labeled with Tb and A647. Three configurations were used in this work: (i) peptide assembly, (ii) oligonucleotide assembly and hybridization, and (iii) both peptide and oligonucleotide assembly/hybridization.
With such a configuration we could demonstrate that multiplexed (duplexed) biosensing is possible with only one QD color by using spectro-temporal multiplexing. This system was applied to detect duplexed DNA hybridization and duplexed enzyme kinetics.
Figure 17 describes the streptavidin-biotin binding-enabled FRET between Tb and different QDs. Once the binding is established, the close proximity between the Tb-complexes (Lumi4-Tb) and the QD enables efficient FRET. Three different biotinylated QDs with emission maxima at ca. 605 nm were investigated: Commercial Qdot605 (Invitrogen/Life Technologies – left TEM image), eFluorNC605 (eBioscience – center TEM image), and self-made CdSe/CdS/ZnS core/shell/shell QDs (right TEM image). The cross shape of streptavidin (sAv) represents the four binding sites for biotin. The four Lumi4-Tb are bound to free lysine groups of sAv and are therefore randomly distributed over the sAv. This distribution as well as the non-spherical shapes of the QDs lead to a Lumi4-Tb to QD-center distance distribution. Out of this distribution, we could identify three main distances that were further averaged to one average Tb-to-QD distance.
Within these systems we could use time-resolved FRET analysis for shape and distance determination of the different QD. Moreover we could determine biotin/QD ratios of the different biotinylated QDs. We showed the suitability of all studied QDs for highly sensitive clinical FRET-bioassays. We have performed a very precise and detailed analysis of biocompatible QDs with TR-FRET from Tb-based donor complexes. On the contrary to most of the other analytical technologies (e.g., dynamic light scattering, HPLC and TEM) our FRET method can analyze the biocompatible QDs under physiological conditions at subnanomolar concentrations and is therefore highly suited (ideally in combination with the other techniques) to give a more accurate picture of the QD properties at concentrations and conditions in which they are usually applied within fluorescence sensing applications.
Understanding FRET in general and biomolecule interaction, folding and unfolding of proteins it is necessary to go down to the single molecule level. This can be achieved with confocal microscopy coupled with lasers or a super continuum white light source. To minimize background and autofluorescence of the proteins and buffer media, two-photon excitation fluorescence can be also used as excitation source. With this technique it is possible to detect FRET from QD to dye in human plasma, serum and whole blood.
Single pair FRET experiments were executed with commercial quantum dots, QD525Biot, (donors) coated with 4-6 biotin molecules and labeled AlexaFluor organic dyes, AF546-Strep (acceptors). In order to find the most efficient donor acceptor pair AF555, AF568 and AF546 were tested in common FRET assays. Among this selection, the AF546 dye showed the best results and it has been chosen for further single pair FRET experiments. Different donor acceptor ratios were mixed and incubated in different media (PBS, PBS with BSA, serum, plasma) and the FRET signal was detected with the experimental setup.
The experimental data show the decrease in donor fluorescence intensity (ChD) associated with the increase in dye concentration and corresponding increase in acceptor fluorescence (ChA).
Two-photon fluorescence microscopy (TPM) possesses several unique advantages enabling noninvasive study of living specimens in three dimensions with low photodamage and high signal to background contrast. The combination of QDs, with their high theoretical two-photon excitation (TPE) cross-sections, and TPM will open a new field of for nonlinear spectroscopy.
We investigated the TPE cross-sections of 5 different water soluble, biocompatible commercial QDs (525, 565, 605, 655 and 705, Qdot(R) Nanocrystals – Invitrogen(TM)) and the resulting TPE spectra over a wavelength range from 700 to 960 nm. The commercial QDs covered by our study display ultra-high two-photon excitation (TPE) cross-sections over a wide wavelength range without impairing their excellent photophysical properties (Figure 18).
Typical TPE cross-sections of organic molecules are in the range of 1 – 1000 GM (Göppert-Mayer). The highest presently known TPE cross-sections of organic molecules are owned by meso, β-linked porphyrin oligomeres. But in the wavelength range covered by our study the best known molecular fluorophores just reach TPE cross-sections of mere 10000 GM. Therefore the TPE cross-sections reached by the examined QDs are at least two orders of magnitude higher than those of common molecular fluorophores.
In addition to the experimental determination of TPE cross-sections, a homogeneous TPE FRET assay was performed, using a QD as donor and an Alexa Fluor Dye(R) (Invitrogen(TM)) as acceptor.

Immunoassay geometry (WP4)

Within this WP commercial quantum dots (eBioscience) were successfully labelled with different antibodies (IgG, F(ab)2, Fab and nanobodies), peptides, oligonucleotides and streptavidin and photophysically characterized. The bioconjugates were then used in WP3 and WP5.
In addition, synthetic methodologies for the conjugation of antibodies to CdSe/ZnS-QDs were developed. These included the modification of the QDs with glutathione or mercapto hexanoic acid and the subsequent coupling of the antibodies to the protecting layers. The synthesis of highly luminescent stable QDs was demonstrated, and the application of the QDs for the analysis of Caseine Kinase was demonstrated.
Synthetic methodologies for the modification of nucleic acids (aptamers) or aptamer subunits with QDs, lanthanide complexes (LCs), dyes, and fluorescence quenchers were developed. The methodologies are based on the coupling of amino-functionalized aptamers, or their subunits, to carboxylic acid-modified QDs, LCs, dyes or fluorescence quenchers.
A variety of generic optical QDs-based aptasensors configurations were developed. These included the development of FRET-based optical sensors through the self-assembly of QDs-dye-modified aptamer subunits, in the presence of the analyte substrate, and the activation of the FRET process.
A new aptasensor platform using chemiluminescence resonance energy transfer (CRET) as readout signal was also investigated. In this context, the chemiluminescence generated by a hemin/G-quadruplex, formed upon the self-assembly of the aptamer-substrate complex, stimulated the CRET process to the QDs. The luminescence of the QDs provided the readout signal for the sensing process (Figure 19).
We could also demonstrate a novel amplification route to detect aptamer-substrate complexes through the Exonuclease III (Exo III) regeneration of the substrate-analyte. This method was implemented to amplify the detection of ATP, cocaine and thrombin as model systems (Figure 20).
The anti-Vascular Endothelial Growth Factor (VEGF) aptamer was implemented to develop a set of fluorescence or CRET QDs-based aptasensors for the detection of the VEGF, known as a biomarker for the Alzheimer’s disease. These efforts led to a set of sensor configurations that addressed the following issues:
a) Fluorescence or CRET QDs-based sensors were developed, and their analytical performances were compared.
b) The Exo III amplified regeneration of VEGF as a means to enhance the sensitivity of sensor devices was demonstrated, and an analytical platform that meets the sensitivity to detect VEGF in Alzheimer’s disease patients, was demonstrated.
c) The optimized analytical platform was successfully applied to detect VEGF in serum.
d) A further biomarker that was explored was L-selectin. An aptasensor configuration for the successful sensing of L-selectin was demonstrated. The aptamer against L-selectin revealed, however, a low affinity constant for the biomarker and thus the sensitivity required to monitor L-selectin in biological fluid was not accomplished.
In addition, a competitive L-selectin-targeted aptamer-based FRET fluoroassay has been developed within UP and Fraunhofer activities in the NANOGNOSTICS project. The principle of this assay is presented in Figure 21. In the absence of L-selectin, a Tb-complex-labeled L-selectin and Cy5-labeled aptamer bind to each other and the FRET signal can be observed. When a native L-selectin is present in the solution, it works as a competitor to the Tb-complex-labeled L-selectin, therefore, the FRET signal decreases. The decrease in the FRET signal can be considered as a measure of L-selectin concentration. In our assay, 90% decrease in the FRET signal can be achieved at L-selectin concentration equal to 100 ng/ml. We demonstrated L-selectin detection in the range from 1 to 100 ng/ml, i.e. at concentrations from five hundred times to five times lower than usually observed in healthy patients. Currently, we are optimizing the L-selectin fluoroassay to achieve a change in FRET signal over the whole dynamic range for relevant L-selectin concentrations.
Immunoassay spectroscopy (WP5)

Broadband light absorption allows for efficient excitation QDs at any wavelength below their emission wavelength. Their intense and narrow emission makes QDs excellent FRET donors that can transfer energy more specifically. In our study the FRET reference system was composed of commercially available, biotinylated quantum dots (Biot-QDs) as donors and streptavidin conjugated with various Alexa Fluor organic dyes as acceptors. Particularly, donors Biot-QD525, Biot-QD565, Biot-QD605, Biot-QD655, and Biot-QD705 and acceptors AF568-Strep, AF647-Strep and AF680-Strep were chosen to facilitate multiplexing. To study a particular donor-acceptor pair we used a well-established interaction between biotin and streptavidin. This model system provides fast and strong binding characterised by a high formation constant yielding 10^12 1/M.
In order to properly choose a QD-Dye FRET pair, two issues must be addressed:
(1) An overlap between donor emission and acceptor absorption should be as large as possible to provide high FRET efficiency.
(2) An overlap between donor and acceptor emissions should be reduced to allow for the crosstalk-free multiplexed detection.
Steady-state fluorescence spectra of FRET pairs in which to Biot-QD565 donor Strep AF568 and Strep AF647 were titrated are shown in Figure 22. In Figure 22 (left) both donor and acceptor emissions are well separated but the FRET efficiency is reduced. In contrast, Figure 22 (right) shows a FRET pair in which the process efficiency is high but both donor and acceptor emissions overlap largely. Then the crosstalk-free wavelength separation is difficult and unambiguous detection is unavailable.
Relative fluorescence intensity in a donor (Biot-QD605) and acceptor (AF647-Strep) channel is presented in Figure 23 left and right, respectively. The donor intensity (red curve) decreases as an acceptor concentration increases. Simultaneously, in an acceptor channel a significant increase in fluorescence was detected, which indicates FRET. A titration with a plain buffer was used as a reference in this case (Figure 23 left, black curve). The donor signal was essentially unaffected when titrated only with buffer. A control experiment was executed to confirm the origin of the acceptor signal change (here AF647). The acceptor was also titrated in the absence of a donor (Figure 23 right, black curve). Only a slight fluorescence increase can be observed. Analogous results were obtained for all other studied pairs.
Our study revealed that following systems: Biot-QD565 / AF568-Strep, Biot-QD605 / AF647-Strep, Biot-QD655 / AF680-Strep are best choices to become FRET pairs in further experiments.
We also determined the best time-gated detection range that could be used to study FRET between QD donors and dye acceptors. Unambiguous temporal differentiation between background fluorescence and donor and acceptor emissions make time-gated detection an excellent method for efficient multiplexing. Importantly, QD luminescence decays noticeably longer than luminescence of organic dyes. Usually QD luminescence decay time can be found in the range from 10 to even 100 ns (depending on QD size and material) while organic dye fluorescence usually decays after 3-5 ns. In order to choose the most suitable time window for the time-gated detection of QD-to-dye FRET, the Biot-QD605/QF647-Strep assay was performed. The fluorescence signal was collected in a time-resolved mode and the fluorescence decays were recorded in the timeframe of 0-200 ns. Several time windows were investigated to obtain respective time-gated intensities and corresponding limits of detection (LODs). The detection limits (LODs) were calculated by dividing 3 times the standard deviation of the ratio at zero acceptor concentration by the slope of the linear rising part of the ratio curve. The LODs obtained from this method are displayed in Figure 24.
As presented in Figure 24, the lowest detection limits were found in following time windows: 5-15 ns, 5-20 ns 10-20 ns. The LODs calculated from those data are lower than the average LOD calculated from the overall time window 0-200 ns (which corresponds to pseudo steady-state luminescence). As the time-window 10-20 ns gives the best sensitivity and detection limits (LODs), this time window is recommended for time-gated detection in biotin-streptavidin QD-to-dye FRET assay.
The most commonly used luminescence detectors are photomultipliers (PMT), micro channel plate photomultiplier tubes (MCP-PMT) and avalanche photodiodes (APD). The choice of appropriate detector is a key factor in FRET spectroscopy and it can determine the general outcome. That is why it is important to know the performance of a detector in a spectral range of interest. Technical specifications give a general overview on detector capabilities but an experimental evaluation can clearly show advantages and disadvantages of a detector in regard to specific samples and experimental setup. The FluoroMax 4 has a Hamamatsu R928 PMT as detector, whereas the FLS920 has a Hamamatsu R1527 PMT and a MCP-PMT as detector. According to the data sheets of the PMT, the FluoroMax 4 PMT can be characterised as a red sensitive detector, because till 800 nm the quantum efficiency is high. But above 800 nm the quantum efficiency decreases strongly. From here, the red sensitive PMT of the FluoroMax 4 is abbreviated as red PMT. In contrast to this red PMT, the quantum efficiency of the FLS920 PMT strongly decreases at 600 nm, so this PMT can be characterised as a blue sensitive detector. Because of that, this PMT is abbreviated as blue PMT. The excitation sources of the two instruments are also different. The FluoroMax 4 has a Xe-lamp as excitation source. The excitation source of the FLS920 is a supercontinuum source SC450-AOTF (Fianium - Southampton, UK), because this instrument is normally used for time-resolved measurements. For a better comparability of the detectors it is useful to excite the sample with the same excitation energy independently of the used instrument. Therefore the excitation energy was measured at different excitation slit widths, wavelengths and in case of the supercontinuum source with different repetition rates. For the choice of the excitation energy it is important to take into account the limitation of the different detectors, because if too many photons are detected, the detector could be saturated or even damaged in the worst case. The choice of the excitation wavelength depends on the used FRET pair. For this report as FRET donor the biotinylated quantum dot biot-QD605 (Invitrogen Corporation, Carlsbad, USA) and as FRET acceptor the AlexaFluor dye AF647 labeled to streptavidin was used. The strong streptavidin-biotin binding leads to a close proximity between donor and acceptor and FRET becomes possible. In previous measurements this FRET pair showed very good FRET, because of the large spectral overlap between donor emission and acceptor absorption. Another point is the good spectral separation between the emission bands of the donor and acceptor, which avoid spectroscopic cross-talk. Generally it can be said, that the integrated intensities of AF647-strep increase with increasing acceptor concentration. The control measurements showed that adding AF647-strep to pure buffer also results in a slight increase of acceptor emission. This is attributed to direct excitation of the acceptor. In contrast to the integrated donor intensities there are large differences between the integrated acceptor intensities which were measured with the different detectors. Especially for the blue PMT and MCP-PMT there are large differences, because the integrated acceptor intensities measured with the blue PMT show a significantly stronger increase. It can be concluded that the blue PMT is more sensitive than the MCP-PMT within the measured wavelengths range. Comparing the red PMT and the blue PMT, the increase of the acceptor intensities of the red PMT is not so strong but in the same range of the blue PMT values.
In summary among three tested detectors the red-light-sensitive-photomultiplier shows the best performance in QD-to-dye FRET. This can be associated with the used of most of tested FRET acceptors emit in the red (600-700 nm), where the blue-PMT and MCP-PMT have reduced sensitivity. This feature is particularly relevant for reaching very low limits of detection (less than 1pM), as envisaged in WP5. In order to realize ultrasensitive spectroscopy investigations on QD-to-dye FRET successfully, optimal accuracy is required. However it may decrease at high intensities due to detector saturation. Optimal intensity can be achieved via adjusting sample concentrations, excitation intensity as well as slits of monochromators both in excitation and emission path.
A new functional luminescent lanthanide complex (LLC) has been synthesized with terbium as a central lanthanide ion and biotin as a functional moiety. The complex synthesis strategy and photophysical properties were studied as well as the performance in time-resolved Förster Resonance Energy Transfer (FRET) assays. In those assays, this biotin-Tb-LLC transferred energy either to acceptor organic dyes (Cy5 or AF680) labelled on streptavidin or to quantum dots (QD655 or QD705) surface-functionalised with streptavidins. The permanent spatial donor-acceptor proximity is assured through strong and stable biotin-streptavidin binding. The energy transfer has been concluded from the quenching observed in donor emission and from a decrease in donor luminescence decay both associated with simultaneous increase in acceptor intensity and in the decay time. The dye-based assays were realised in TRIS and in PBS, whereas QD-based systems are studied in borate buffer. The delayed emission analysis allows for quantifying the recognition process and for autofluorescence-free detection, which is particularly relevant for application in bioanalysis. In accordance with Förster theory, Förster radii (R0) were found to be around 60 Å for organic dyes and around 105 Å for QDs. The FRET efficiency reached 80% and 25% for dye and QD acceptors, respectively. Physical donor-acceptor distances (r) have been determined in the range 45-60 Å for organic dye acceptors, while for acceptor QDs between 120 Å and 145 Å. This newly synthesised biotin-LLC extends the class of highly sensitive analytical tools to be applied in the bioanalytical methods such as time-resolved fluoroimmunoassays (TR-FIA) or luminescent imaging.
A homogeneous time-resolved luminescence resonance energy transfer (TR-LRET) assay has been developed to quantify proteins. The competitive assay is based on resonance energy transfer (RET) between two luminescent nanosized particles. Polystyrene nanoparticles loaded with Eu(III) chelates (EuNPs) act as donors, while protein-coated quantum dots (QDs), either CdSe/ZnS emitting at 655 nm (QD655-strep) or CdSeTe/ZnS with emission wavelength at 705 nm (QD705-strep), are acceptors. In the absence of analyte protein, in our case bovine serum albumin (BSA), the protein-coated QDs bind non-specifically to the EuNPs leading to RET. In the presence of analyte proteins, the binding of the QDs to the EuNPs is prevented and the RET signal decreases. RET from the EuNPs to the QDs was confirmed and characterized with steady-state and time-resolved luminescence spectroscopy. According to Förster theory, the approximate average donor-acceptor distance is around 15 nm at RET efficiencies equal to 15% for QD655 and 13% for QD705 acceptor, respectively. The limits of detection are below 10 ng of BSA with less than 10% average coefficient of variation. The assay sensitivity is improved, when compared to the most sensitive commercial methods. The presented mix-and-measure method has potential to be implemented into routine protein quantification in biological laboratories.
The synthesized LCs (1-7) were applied to a number of heterogeneous and homogeneous assays. LCs were utilized as such and in combination with QDs and acceptor dyes. Moreover, lanthanide nanoparticles were studied in combination with QDs and acceptor dyes:
a) A functional highly sensitive assay was developed to measure protein concentration using time-resolved fluorescence energy transfer (TR-FRET) from carboxylated europium(III) nanoparticles to the commercial streptavidin-coated QDs (emission at 665 [SA-655] and 705 [SA-705] nm).
b) Aggregation processes are prominently involved in the Alzheimer’s disease as several proteins are prone to aggregation. A sensitive detection method for the aggregation of amyloid- peptide and microtubulin with the Eu-nanoparticles and acceptor dyes was constructed.
c) Three prepared europium chelates (1c,6c,7c) were preliminary characterized in a streptavidin binding assay. As the assay has no real relevance in bio-community, the biotinylated LCs were further applied to competitive heterogeneous c-reactive protein assays successfully. A clinically relevant assay was developed as the concentration range was from 3 to 300 μg/L. Although the range is lower than desired for the CRP assay this allows a high serum dilution further reducing any adverse effect of serum.
d) Homogeneous assays were investigated with the novel biotinylated LCs (1c,6c,7c) using two approach: 1) fluorescence resonance energy transfer between LCs and streptavidin labelled acceptor dye, and 2) quenching resonance energy transfer (developed by UTU) where the luminescence of LC labelled biotin was quenched in solution with soluble quenchers through RET process and protected when bound to streptavidin. The simple streptavidin-biotin approach was taken to compare different detection schemes using the same or very similar assay components. 7-dentate chelate appears to be the most promising candidate for FRET assay and 9-dentate chelate for the QRET assay. No differences in the assay sensitivity were found as the affinity drives this property.
e) Aptamers were labelled with Eu- and Tb-chelates and aptamer binding and signal formation in growth factor binding was characterized and investigated. Changes at the growth factors levels have been associated with AD. The assay principle is extremely simple as the labelled aptamer binds to the target protein and luminescence signal is detected in time-resolved mode. The non-bound labelled aptamer is quenched in solution with soluble quenchers. The TR-signal is directly proportional to the concentration of the growth factor. The sensitivity of the assays for VEGF and bFGF was below 1 nM. Although the assay sensitivity was high, the methods are not highly suitable to serum-based concepts as the sensitivity is compromised. The method matches well for the drug discovery assays as the matrix effects are not problematic.
f) EGF antibody and EGF protein were labelled with the Tb-chelate. A weak TR-FRET was found in the non-competitive sandwich assay with the labelled antibody and this approach was not further investigated. The low assay performance was due to large distances between antibodies bound to both ends of the EGF protein. The work was continued with the labelled EGF and the competitive TR-FRET approach. An extremely high sensitivity was observed (LOD=20 pM). Further work is dedicated to understand the assay performance.
Within this WP we have been also developing multiplexed immunoassays for serum-biomarkers. For these assays we used the bioconjugates prepared in WP2 (Tb) and WP4 (QD). For the multiplexed assays the following biomarkers were used: PSA (prostate specific antigen), NSE (neuron specific enolase), CEA (arcinoembryonic antigen), the cytokeratin-19 fragment Cyfra21-1, SCC (squamous cell carcinoma antigen), the carbohydrate antigen CA15.3, EGFR (epidermal growth factor receptor) and L-selectin. A first assay system consisted of a Tb-donor and dye-acceptor FRET pair. The aim was to develop an optically multiplexed assay, which applies a spectral crosstalk correction of the different dyes and to use this crosstalk correction for achieving a 5-fold multiplexed immunoassay, which provides clinically relevant detection limits for biomarkers in serum samples. This work was published in The Journal of the American Chemical Society (see dissemination section). The results are summarized in Table 1.
It becomes apparent that the detection limits achieved with this 5-fold multiplexed serum-biomarker immunoassay are in the same concentration range than the commercially and clinically used single-biomarker immunoassay. The unique combination of high-sensitivity fluorescence detection with straightforward spectral crosstalk correction allowed an efficient distinction between healthy and elevated biomarker levels as well as a prediction of the type of lung cancer (SCLC or NSCLS) from a single 50 µL human serum sample. We expect our method to provide significant benefits for real-time investigations of simultaneous biomolecular interactions and ultrasensitive detection of multiple biomarkers for early disease diagnostics, in particular for AD detection once several serum-biomarkers have been identified.
In order to demonstrate multiplexable immunoassays for the detection of biomarkers in small-volume serum samples using Tb-to-QD FRET we have conducted a detailed study using different kinds of antibodies. The first approach consists of the application of full IgG as well as reduced F(ab)2 and Fab antibodies (as detailed in deliverables 5.9 and 5.10). The assay principle is explained in Figure 25. Tb-antibody conjugates (top) and QD-antibody conjugates (bottom) contain two different primary antibodies (which are further reduced to F(ab)2 and Fab fragments) against PSA. All possible combinations of the six different conjugates were used to specifically recognize PSA (P) in FRET sandwich immunoassays (center). The second approach used single domain antibodies, which are also called nanobodies or VHH. The principle of this assay is explained in Figure 26. The combination of Tb-labeled VHH and VHH-labeled QDs allows the formation of (Tb-VHH)-EGFR-(QD-VHH) sandwich complexes, which results in efficient FRET from several Tb to the central QD. Time-gated detection of the FRET quenched Tb photoluminescence (PL) and the FRET-sensitized QD PL is used for a sensitive quantification of EGFR.
The results of the immunoassays are presented in Figure 27 and 28 (for more details see Deliverables 5.9 and 5.10). The limits of detection of all homogeneous FRET assays are in the sub-nanomolar (few ng/mL) total PSA range and the dynamic range spans approximately three orders of magnitude. The difference of ABs for the Tb conjugates is negligible, which we attribute to the relatively large distances in our sandwich immunoassay (Tb-donors labeled to the Fc region of the IgG are too far away from the QD-acceptor to participate in efficient FRET). Another remarkable aspect concerns the very low coefficients of variation for the QD-Fab containing FRET systems. This important advantage is most probably caused by the higher labeling ratio of the small Fab fragments per QD, which leads to a much lower background signal of directly excited QDs per AB-PSA-AB binding. Therefore the fluorescence reader system can be used at a higher sensitivity without detector saturation. The best immunoassay system combining maximum sensitivity, minimum antibody modification (no IgG reduction for the Tb conjugates) and maximum separation efficiency is therefore the (Tb-IgG)+(QD-Fab) system. The VHH immunoassay is the first demonstration of a homogeneous FRET-immunoassay combining quantum dot nanocrystals and VHH nanobodies. The immunoassay requires relatively short incubation times (ca. 1h), does not need any washing and separation steps (simple mix-and-measure procedure), is measured quickly (5 s per sample) and ratiometrically (FRET-ratio), and time-gated detection (void of autofluorescence background and directly excited QDs) allows sub-nanomolar detection limits of biomarkers in both buffer and serum samples, which is in a relevant range of EGFR serum levels for cancer diagnostics. We believe that our proof-of-concept in combination with the fast, easy and inexpensive production of ultra-small VHH antibodies (compared to monoclonal IgG antibodies) and the unrivaled optical properties of QDs for sensitive and multiplexed detection will also make this immunoassay a valuable tool for future in-vitro diagnostic applications.
Both results present major milestones for the integration of QD-based immunoassays into clinical diagnostics and two manuscripts have been submitted for publication.

Immunoanalyzer and development of modular components (WP6)

Work at Edinburgh Instruments has focused on the development of modular components used in demonstrators of a multi-modality, high throughput immunoassay plate reader meeting the design goals and specifications requested by NANOGNOSTICS partners.
The dual channel plate (96 or 384 well) reader combines a confocal optical geometry with ‘plug and play’ exchangeable light sources (steady state and pulsed xenon lamp or diode laser, supercontinuum laser, nitrogen laser etc), wavelength selection excitation and emission filters and single photon counting sensitive cooled and gated photomultiplier detectors with state of the art data acquisition electronics and sophisticated data analysis software.
Two of the multi-modality plate reader systems have been made to date, one installed at the Université Paris Sud and the other in house at Edinburgh Instruments as a test facility and available for measurements by any other NANOGNOSTICS partner (see Figure 29).
The plate reader as designed is capable of the following modalities: fluorescence intensity (steady state and gated), fluorescence polarisation, absorbance, time resolved (TR) fluorescence, time resolved anisotropy, (TR-)FRET, fluorescence lifetime, luminescence
The plate reader can be easily configured to measure fluorescence in every form rapidly and quantitatively from standard 96 and 384 well plates with sample volumes down to <20µl. The features of the multimodality plate reader system are summarized in Table 2.
The area of maximum interest for the NANOGNOSTICS project is time resolved measurements of TR-FRET and fluorescence lifetime. Demonstration of system performance and sensitivity has been proven using commercially available assays. For example, the TR-FRET based TPSA (total prostate specific antigen) assay was used to test the plate reader sensitivity and to show that the instrument accuracy is sufficient to detect antigen concentrations below the disease diagnostic threshold (2.5 ng/ml), i.e. for cancer diagnostics. Fluorescence lifetime assays using long lifetime 9-aminoacridine dye have been measured for several interesting target classes including protein kinases and protease using TCSPC.
In all cases the plate reader provides a plethora of research quality data which could prove invaluable for many research and assay development applications.
This work entailed the development of system sub-assemblies including excitation light sources (150 W xenon arc lamp with ellipsoidal mirror for efficient light collection, extension of sub-ns pulsed LEDs down to 250 nm wavelength, novel linear variable filter based wavelength selector for supercontinuum lasers and other broadband light sources), integrated detector housing for gated, cooled single photon counting photomultipliers, hardware controller for 20 stepper motors, data acquisition electronics for fast photon counting, multi-channel scaling and time-correlated single photon counting and a sophisticated library of analytical software (FAST fluorescence analysis software), many of which have been rolled out commercially or been integrated in the latest FLS980 series fluorescence spectrometers launched by Edinburgh Instruments in Autumn 2012.

Clinical development and evaluation (WP7)

Charité performed continuously literature search to update and discuss with the partners ongoing research in the field of biomarker evaluation specific for Alzheimer´s disease (AD). Collection of patient samples (plasma and if possible cerebrospinal fluid) was conducted throughout the whole period after receiving ethical approval.
We successfully recruited blood and cerebrospinal fluid (CSF) samples from AD/MCI patients and non-demented controls. Patients were scored by MMSE testing and inflammatory influences analysed by partial blood counts and determination of the proinflammatory marker IL-6. In the course of this project we collected together with our colleagues from the Department of Neurology 21 samples from none demented controls, 7 samples from MCI patients, 3 samples from patients with a diagnosis of being in a status between MCI and AD (MCI/AD) and 46 samples of AD patients (Table 3). From each proband 5 ml plasma inactivated with EDTA was collected and from some probands additionally 5 ml CSF. In the course to diagnose AD the CSF samples were tested for the clinically established markers total tau protein, hyperphosphorylated tau protein, Aβ42-protein, Aβ40-protein and the ratio of Aβ42-protein and Aβ40-protein (Table 4). In total we have collected sample material from 77 probands.
Using matched pairs of antibodies we were able to detect concentration dependant all marker proteins (EGF, VEGF, L-selectin) in plasma inactivated with EDTA. It was possible to detect EGF and L-selectin at physiological concentrations. It was not possible to detect VEGF at physiological concentration in plasma inactivated with EDTA. But we were able to detect VEGF at elevated concentrations that reflect the published concentrations of VEGF in plasma inactivated with EDTA of AD patients (Corsi M.M. et al., Biogerentology, 2011).
The difficulties in identifying a reliable panel of AD markers is also reflected in the inconsistency of different study results on AD markers. On a single parameter level this is also true for VEGF and EGF. Corsi et al. and Chiappelli et al. identified elevated concentrations of VEGF in AD sample material using plasma inactivated with EDTA in contrast to sample material of none demented controls (NDC) (Corsi M.M. et al., Biogerontology, 2011; Chiappelli et al., Rejuvenation Research, 2006). Del Bo described no differences in VEGF concentrations between sample material of none demented controls and Alzheimer patients by using serum (Del Bo R. et al. Ann Neurol., 2005). In contrast Mateo found reduced concentrations of VEGF in AD sample material in comparison to NDC sample material by analyzing serum (Mateo I., Acta Neurol Scand., 2007).
Ray et al. identified reduced concentrations of EGF in AD sample material compared to none demented control samples using plasma inactivated with EDTA (Ray S. et al., Nature Medicine, 2007). In contrast, Marksteiner et al. described elevated concentrations of EGF in AD sample material using Biocoll(R) isolated plasma (Marksteiner J. et al., Neurobiol Aging, 2011).
We successfully established several ELISA based methods for potential relevant AD biomarkers from peripheral blood, but were not able to confirm data from the literature (Ray S. et al., Nature Medicine, 2007, Classification and prediction of clinical Alzheimer´s diagnosis based on plasma signaling proteins). In agreement with Björkqvist et al. (PLoS ONE, 2012, Evaluation of a previous suggested plasma biomarker panel to identify Alzheimer´s disease) we conclude that the panel of 18 proteins suggested by Ray et al. is not suitable for AD detection. Exemplarily performed analyses on different sample treatments, demonstrated for the marker proteins EGF and VEGF, resulted in varying results and pointed to the importance of standardization in biomarker research. Highly sensitive and specific detection of VEGF demonstrated the feasibility of the NANOGNOSTIC spectroscopic platform technology for biomarker detection. First results from ongoing research indicate that plasma levels of complement proteins might be indicative for AD development.
The inconsistency of VEGF and EGF concentrations reported by different studies brought our focus on studying the influence of preanalytical sample handling on VEGF and EGF concentrations. First we analyzed the concentrations of VEGF and EGF of matched samples of plasma inactivated with ETDA and serum of none demented control probands (n=8). The mean VEGF and EGF concentrations in serum samples were 7 and 123 times higher than in plasma samples inactivated with EDTA, respectively (Figure 30). These results indicate that VEGF and EGF are released during the process of clotting into serum.
We analyzed this process in more detail and found that a delay in blood sample processing led to an increase in VEGF and EGF concentrations only in samples in which clotting was possible (serum vials). Most interesting in one out of three EDTA plasma samples VEGF and EGF concentration did also change over time, leading to a 4-fold and 7.2-fold increase, respectively.
Additionally we looked at the influence of storage conditions on VEGF and EGF concentrations (Figure 31). To investigate the impact of platelet derived VEGF and EGF during storage we stored fresh platelet rich plasma (PRP) and plasma samples (n=3) for two hours at 22°C, 4°C and -80°C or analyzed samples after three cycles of freezing and thawing (positive control). Storage of PRP at -80°C did increase VEGF and EGF concentrations 1.4-fold and 6.4–fold, respectively. In plasma samples no influence of storage at -80°C on VEGF and EGF concentrations was observed.
Those results clearly show the impact of preanalytical sample handling on the concentrations of VEGF and EGF in blood derived sample material. It further indicates that serum is not the suitable sample material for analyzing VEGF and EGF concentrations in none demented controls and Alzheimer patients. Additionally we could show that even by analyzing plasma inactivated with EDTA the sample material should be analyzed or stored (-80°C) within one hour, because even in individual EDTA plasma VEGF and EGF concentrations did increase.
Taken together, the influence of preanalytical sample processing on the concentrations of AD markers as shown for VEGF and EGF has to be taken into account for any AD marker from blood sample material. Among other reasons this may explain the inconsistency of published data in this field of research. It also may contribute to the fact that no reliable set of biomarkers from blood samples could be identified for the diagnosis of mild cognitive impairment (MCI) or Alzheimer´s disease until now. Due to the lack of such a reliable set of biomarkers for the diagnosis of AD it was not possible to establish a multiplexed detection system during this project. Nevertheless, we could show during the project that we are able to perform the envisioned assay technology with other markers, which is a great step forward in the diagnosis and monitoring of diseases for which a set of biomarkers, that have to be tracked, is known.
Potential Impact:
Most of the initial plans of dissemination of foreground have been fulfilled within the 42 months of the NANOGNOSTICS project. A large amount of peer reviewed publications (48, cf. Template A1) as well as presentations on workshops, conferences, public seminars etc. (138, cf. Template A2) have been used for dissemination to the scientific community and to the public.
Socioeconomic impact has been mainly created by the development of innovative technologies, from which some have been submitted as patent applications (cf. Template B1) and the development, production and integration in the product portfolio of the participating SME Edinburgh Instruments of several spectroscopy-related devices (cf. Template B2).
Apart from the tables in templates A1, A2, B1 and B2, the following paragraphs will give more details about dissemination and impact.

2.1 Communication and dissemination within the consortium

- Daily contact between the different partners of NANOGNOSTICS has been an important issue in order to organize the workflow and to exchange about results and necessary interactions between all partners. This has been realized by regular exchange via telephone, e-mail and data exchange between the NANOGNOSTICS ftp-server. As this communication has been very efficient a newsletter (as originally planned to assure an efficient communication between all partners) has not been established.
- A dedicated NANOGNOSTICS website was set up at the beginning of the project. Apart from a public space an ftp-site was also created to allow an easy and efficient exchange of data, reports etc.
- Regular reports have been submitted by every partner to the coordinator in order to assure an efficient communication and the submission of periodic reports, deliverables etc. to the European Commission.
- Internal NANOGNOSTICS meetings with all partners have been organized twice a year in order to exchange results and ideas, for discussion of technical problems and possible solutions, and for planning the exchange of research staff, equipment, and materials. These meetings were also used to get a frequent update on deliverables and milestones and to adjust the planning of the project in the cases of unfulfilled deliverables or milestones. Some meetings were held with the participation of external experts in order to get an independent outside view on the results.
- The close interaction between the partners was enforced by the exchange of researchers (mainly PhD students) who visited the laboratories of the other partners for collaborative experiments. Also several partners have visited the facilities of the SME partner Edinburgh instrument in order to stimulate the academia-SME interaction.

2.2 Training and education of junior scientists and students

- All academic partners have recruited one or more junior investigators on the graduate student or post-doctoral researcher basis. Since much of the research has been performed in direct collaboration with other consortium laboratories, the junior researchers have learnt to work in different environments and facilities. During these exchanges, they also learnt to communicate with researchers from other disciplines, acquire new skills and technologies, and most importantly, develop novel interdisciplinary approaches. For the post-docs the participation in the very interdisciplinary and collaborative NANOGNOSTICS project gave them the opportunity to further strengthen their research, mobility and collaboration profile toward becoming an own independent group leader. For the graduate students the NANOGNOSTICS project offered a very interdisciplinary and innovative educational program and several NANOGNOSTICS students have received their PhD degree within the project time or will receive their PhD degrees within the following one to two years.
- The senior investigators of the NANOGNOSTICS consortium who were teaching in their universities could make use of the project results for lectures and seminars in order to give students an inside look into actual research programs and state-of-the-art science on an international level. During the NaNaX conference in Spain in 2012, which was co-organized by the NANOGNOSTICS consortium, many PhD students and post-docs from inside and outside of the project have participated and the NANOGNOSTICS satellite conference was given a summer school character, where the invited speakers (from inside and outside of the consortium) dedicated a large part of their presentations on educational aspects for explaining their technologies to the students from different research disciplines.

2.3 Protection and exploitation of intellectual property (IP)

As the commercial exploitation of the components and/or IP rights was organized in the consortium agreement no conflicts of interest have been appeared during the NANOGNOSTICS project. Apart from the publication of innovative technologies, which has led to contacts with the industry and the creation of new research projects, some of the results have been submitted as patents. Moreover, the SME partner Edinburgh Instruments could bring several devices and components developed during the NANOGNOSTICS project into production and these products have significantly increased and improved the product portfolio of Edinburgh Instruments.

2.4 Dissemination of the scientific results to the scientific community

- The consortium of NANOGNOSTICS co-organized an international conference NaNaX 2012 in Fuengirola, Spain, with an own satellite conference purely dedicated to NANOGNOSTICS. Many internationally renowned scientists and researchers, researchers and managers from the industry and students participated in this conference, which was open to everyone. Members of the External Advisory Board as well as invited speakers of several non-participating research groups and biotechnological industrial enterprises, both European and overseas were taking part in this very interesting conference. Several members of the NANOGNOSTICS consortium also participated in organizing external related activities such as workshops or conferences.
- In addition to the meetings organized by the NANOGNOSTICS consortium, all the consortium members actively participated in different scientific seminars and meetings, international conferences and congresses, and industrial events by presenting the NANOGNOSTICS project in oral and poster presentations. This channel of dissemination to the scientific community has also lead to new scientific / industrial collaborations and the creation of new research projects on the national and international level. All participations in such dissemination activities are summarized in the Template A2.
- In addition to the presentations at conferences, the primary public means of dissemination of the research conducted within the NANOGNOSTICS project was individual and joint publications of scientific papers in peer-reviewed high-impact journals (e.g. Nanoletters, Journal of the American Chemical Society, Angewandte Chemie, Trends in Biotechnology etc.). Due to the reviewing system used by scientific journals, this way of dissemination provided an excellent way of external quality control. The importance of the scientific results obtained within the NANOGNOSTICS consortium can be judged by the number of publications (48, cf. Template A1) and the impact factors (several above 10) of the journals where the results are published.
- A website with the goals and strategies of NANOGNOSTICS has been presented to the scientific community. This website also includes information about the participating research groups and information for patients or participants of the clinical study.

2.5 Raising participation and public awareness

- Public dissemination with the goal of raising public awareness has been performed in seminars, scientific events and discussions for the public, publications and teaching in the universities. The main goals of public awareness actions were focused on enhancing the acceptance of nanotechnologies and their industrial development, in particular related to clinical approaches. Our activities assisted in demonstrating the technological possibilities and future perspectives of nanotechnology in diagnosis (for AD and other diseases) on the one side, and in identifying problems of AD diagnostics such as the lacking availability of a sufficiently specific biomarker panel. Collaboration with other projects on the national and international level (e.g. the active participation in the EU-Cluster Targeted Nanopharmaceuticals and Diagnostics) also strengthened the dissemination of results, problems and risks of nanotechnology and diagnostics.
- The NANOGNOSTICS partners also organized open-house days to present their work locally in an informal way to all those who share a scientific interest. This type of (educational) outreach was much appreciated by the public in almost all age and societal levels and we believe that such actions were quite beneficial for the public towards understanding bio- and nanotechnology, and thus for their impact and development.
- Several members of the NANOGNOSTICS project also invited school classes or groups from schools to their laboratories and took school interns in order to realize a direct practical experience for the pupils with biotechnology and nanotechnology. Many of the students were very interested in the medical applications of scientific technologies.

2.6 Socioeconomic Impact

Within the NANOGNOSTICS project novel and innovative results have been created in the areas of nanotechnology, supramolecular chemistry, biochemistry, physical chemistry and in-vitro diagnostics. The collaborative research between different academic groups and a technology-based SME has led to many high quality publications, a significant advancement of European research in the aforementioned areas, the development of several highly innovative products by the participating SME Edinburgh Instruments as well as a deeper understanding of AD-related biomarkers in AD-patients and non-AD control persons. The mixture of advances in understanding AD diagnostics, the development of novel diagnostic technologies and the creation and production of novel materials, instruments and devices has led to a a significant (still small on the international level but we believe that the results of our small-to-medium-sized project have brought several social, societal and economic aspects a step forward) impact for the European society and economy. The main impacts concerning the different areas can be summarized as follows:
Nanotechnology: Novel strategies for quantum dot bioconjugation and biocompatibility have been developed, which facilitate the use of QDs as efficient biosensors.
Supramolecular chemistry: Several novel lanthanide complexes were developed, which allow the attachment to biomolecules and flexible photophysical properties. Such complexes can be applied for various diagnostic applications for high-sensitivity biomarker detection.
Biochemistry: Several novel homogeneous aptamer-based bio- and chemosensors have been developed, which allow the replacement of antibody-based detection scheme and improve multiplexing capability. Such aptamer-based senors will allow an easier and more reproducible synthetic production of biomolecular detection molecules.
Physical chemistry: Many quantum dot-based FRET and CRET interactions (QDs as donors and acceptors) in biological binding systems have been developed, which provides a deeper understanding of QD-based FRET. This deeper understanding will be the basis for the development of even better and more efficient QD-FRET-based biosensors.
Innovative products: Several different important components (e.g. light sources, electronics) and a stand-alone multiflexible time-resolved fluorescence plate reader have been developed, which is of great interest for time-resolved biomarker assay development and significantly increased the product portfolio and competiveness of the SME partner Edinburgh Instruments. These products will have a direct impact on strengthening European technology and economics.
In-vitro diagnostics: Homogeneous immunoassays and aptamer assays for several different biomarkers have been developed, which will have a significant impact on the improvement of multiplexed diagnostics for fast and easy-to-use test, e.g. in point-of-care diagnostics or high throughput screening toward companion diagnostics. Such assays are of high importance for disease screening of a large part of the (aging) population.
Alzheimer’s disease: 77 probands have been recruited and plasma as well as CSF from MCI, AD and non-demented control patients have been collected and analysed, which provided new insights into possible AD biomarkers and their diagnostic value as well as in the importance of standardization in biomarker analysis. These results provided important information for the future of AD diagnostics.
Regarding this summary-list of some of the most important results from the NANOGNOSTICS project there will be a significant impact on multiple disciplines combining chemistry, physics and biology for a successful application in the important health sector of multiplexed diagnostics. Moreover, there will be a possible impact on the society, as the developed technologies might help to provide a more efficient and faster detection of biomarkers for different diseases (AD, cancer and others). A direct economic impact has been created by the development of different devices and components that are already integrated in the product portfolio of Edinburgh Instruments. The scientific interdisciplinarity, the SME-academia interaction as well as the direct access to patient samples within the NANOGNOSTICS project has been therefore been a big success to provide scientific, economic as well a societal impact.

List of Websites:

Website: www.nanognostics.eu, www.nanognostics.org

Contact:
Dr. André Geßner
Fraunhofer Institute for Applied Polymer Research
Geiselbergstr. 69
D-14476 Potsdam-Golm
Tel: +49 331 568 3331
Fax: +49 331 568 3000
E-mail: andre.gessner@iap.fraunhofer.de

Informations connexes

Reported by

FRAUNHOFER-GESELLSCHAFT ZUR FOERDERUNG DER ANGEWANDTEN FORSCHUNG E.V
Germany
See on map