Skip to main content
European Commission logo print header

In Vivo Imaging of Beta cell Receptors by Applied Nano Technology

Final Report Summary - VIBRANT (In Vivo Imaging of Beta cell Receptors by Applied Nano Technology)

Executive Summary:


In Vivo Imaging of Beta-cell Receptors by Applied Nano Technology

Currently, around 30 million people in the enlarged Europe suffer from diabetes, with a prevalence of 7.5% in the member states. By 2025, the number of people with diabetes is expected to rise to around 50 million in Europe, thus increasing prevalence to 10.9%. This devastating disease is ranked among the leading causes of fatal cardiovascular diseases, kidney failure, neuropathy, lower limb amputation and blindness. Estimates of annual direct costs of diabetes care in Europe are currently EUR 50 billion. The indirect costs of diabetes i.e. the costs of lost production are as high as the direct costs or even higher, not to speak of the intangible suffering of the patients and their families.

Diabetes results from an absolute or relative decline in pancreatic ß-cell function and/or mass. Studies suggest that gradual loss of ß-cell mass (BCM) starts many years before the disease becomes overt. Therefore, non-invasive in-vivo imaging and quantification of BCM is of ultimate importance for diabetes management and the development of novel therapies. Unfortunately, such diagnostic methods hitherto do not exist.

The principal goal of VIBRANT is the development of nano-technology-based magnetic resonance imaging (MRI) agents for diabetes diagnosis. The two major objectives of VIBRANT are

1. to develop a method suitable for the non-invasive imaging and quantification of insulin-producing ß-cells in the pancreas of diabetic human subjects by means of target specific nanoparticular probes, and

2. to study the beneficial or adverse effects of ligand functionalized nanoparticular probes for potential therapeutic applications, e.g. targeted drug delivery systems.

In VIBRANT, a biocompatible, non-toxic polymeric nanocontainer containing a sensitive -optionally multimodal- reporter system or therapeutic agents was accomplished via highly reproducible microfluidic processes. The nanocapsules showed low unspecific binding, superior sensitivity and longitudinal signal stability in MRI and optical readouts (T1, T2*, 19F, fluorescent). The outer surface of this nanocapsule was decorated with various recognition motifs for the in vivo targeting of the pancreatic ß-cell. Substantial progress has been accomplished in in vivo imaging of functional β-cell mass and a novel ex vivo mesoscopic optical imaging (OPT/SPIM) provided exact cross-validation of prelabeled islet transplants in the pancreas. However, all recognition motifs tested for ß-cell specific binding showed insufficient target affinity; hence determination of ß-cell mass after systemic administration of labeled nanocontainers needs future improvements.

Project Context and Objectives:
2. VIBRANT Summary Description

The increasing global incidence of diabetes and the tremendous socio-economical impact of the disease and its sequels urgently demands for major improvements in preventive medicine, early diagnosis and therapy. Studies suggest that one of the main causations of diabetes, namely the gradual deterioration of pancreatic β-cell mass (BCM), precedes overt diabetes for more than a decade. In order to find a remedy against these deficiencies, the principal goal of VIBRANT was the development of nano-technology-based magnetic resonance imaging (MRI) agents for early diabetes diagnosis by determination of BCM.
To this end, VIBRANT was constituted around a consortium of nine renowned European research institutions and one SME in the fields of diabetology, β-cell physiology, imaging and nanotechnology.

Name Short name Country
Center for Applied Nanotechnology (CAN) GmbH (Coordinator) CAN Germany
Dept. of Chemistry, Univ. Hamburg UH Germany
Rolf Luft Research Center for Diabetes and Endocrinology, Karolinska Insitutet, Stockholm KI Sweden
Laboratory of Experimental Hormonology, Brussels Free University ULB Belgium
Centre for Molecular Medicine, Umeå University UMU Sweden
Centre for Genomic Regulation, Barcelona CRG Spain
Katholieke Universiteit Leuven K.U.Leuven Belgium
Centre for Biomolecular Interactions Bremen UniHB Germany
Universität Bayreuth UBT Germany
University of Copenhagen UCPH Denmark

The two major objectives of VIBRANT were:

1. to develop a method suitable for the non-invasive imaging and quantification of insulin-producing ß-cells in the pancreas of diabetic human subjects by means of target specific nanoparticular probes,


2. to study the beneficial or adverse effects of ligand functionalized nanoparticular probes for potential therapeutic applications, e.g. targeted drug delivery systems.

Two anatomical features of the pancreas demand for technical solutions, far more advanced than those required for other imaging applications. First, β-cells exist in tiny (100–300 µm in diameter) micro-organs, the islets of Langerhans, which are distributed throughout the exocrine pancreas. Each is too small to be spatially resolved by current non-invasive in vivo imaging methods. Second, the total β-cell mass, even in healthy humans, is so small (about 1 g) as to constitute only about 1–2% of the total pancreatic mass.
This required

i. a contrast agent providing a stable and high signal intensity,
ii. a contrast agent that is highly specific to the cell type, but devoid of toxicity, and
iii. an imaging technology with the potential of paramount spatial resolution.
For a more comprehensive knowledge about β-cell functionality with regard to proliferation and apoptosis, in order to eventually stall β-cell deterioration and induce increase of functional BCM, an agent would be desirable, which

iv. allows monitoring of functional aspects (viability status of the β-cell)
v. holds the potential for a novel β-cell specific drug delivery system

In order to address these objectives, VIBRANT aimed for novel MRI contrast agents, which provide sufficient sensitivity and spatial resolution power, as well as the potential for quantification. Magnetic resonance imaging (MRI) as a diagnostic tool, offers numerous advantages, e.g. high spatial resolution, potential of functional imaging, and lack of radiation exposure (suitable for screening). Superparamagnetic iron oxide nanoparticles (SPIONs) offer high sensitivity and spatial resolution, whereas novel 19F-containing imaging agents hold the potential of quantification. Furthermore, the hyperintense contrast of the 19F signal holds the potential to avoid the common problem of hypointense artifacts and to image the contrasted structures in absence of anatomical background. For this purpose, VIBRANT envisioned micellar nanocontainers comprised of amphiphilic polymers micelles, enclosing optionally superparamagnetic nanoparticles or fluorinated agents and –optionally - fluorescent quantum dots (QD) suitable for a cross-validation by multiplex read-out. The magneto-fluorescent interdependencies of the constituents were studied, adverse interferences were minimized and the potential of this multi-component system optimized. In summary, a first result of VIBRANT are Nano-Containers (NCs), which stably enclose superparamagnetic nanoparticles (Fe2O3, Fe3O4) or fluorous phases for high sensitivity/high resolution MRI, optionally together with fluorescent quantum dots.
The outer shell of the nanocontainers is comprised of biocompatible polymer chains, amenable to functionalization with biomolecules (antibodies, carbohydrates or peptides). The latter were used in order to achieve affinity and specifity to insulin secreting cells (see Fig. 1).
In summary, the second result of VIBRANT is the fully fledged technology for the preparation of diverse ligand functionalized NCs.
A plethora of variations in size, shape, payload, linker chemistries and geometries, as well as ligands, ligand density or surface coverage of the nanocontainer were prepared and tested, which pivotally determined the biological properties. In VIBRANT a cascade of cell assays was used in order to optimize the ligand functionalized FPNC with regard to biocompatibility, lack of toxicity, affinity and specificity. In particular, the effects on β-cell function, survival and regeneration was investigated in vitro by analyzing various functional parameters of insulin secreting cells, like glucose utilization, glucose oxidation, mitochondrial membrane potential, insulin release, and apoptosis. In summary, the third result of VIBRANT are biocompatible, non-toxic nanoconstructs, which did not affect cellular viability and metabolism and show moderate to high affinity as well as selectivity for insulin secreting cells in vitro, suitable for further in vivo testing.
In VIBRANT, a globally unique in vivo anterior chamber model (ACE) of the eye was utilized, thus directly exploring the effects of the nanoconstructs on fully vascularized and innervated islet transplants. The optimized ligand functionalized NCs finally were tested in vivo on islet transplants, as well as in type 1 and type 2 diabetes models, in order to differentiate and quantify β-cell mass of healthy vs. impaired pancreata using confocal microscopy and MRI. Unfortunately –despite of a careful selection of the most state-of-the-art assays and models- VIBRANT encountered a common issue in drug research, i.e. that the in vitro results only partially translated into in vivo models. In particular, the NC constructs, identified in cellular tests showed insufficient labeling of islets in vivo. Nevertheless, a novel small molecule carbohydrate derivative, AJ070 -not in the primary scope of VIBRANT- showed promising results in vitro and in vivo and will be pursued further.
Modified NCs comprised of a biodegradable polymer shell were studied for their potential as drug delivery systems in order to restore, maintain or improve in vivo β-cell function. First evidence for the feasibility of this approach was gained. In summary, a fourth result of VIBRANT is the demonstration of the potential of this concept to diagnostic and therapeutic applications in vivo.
It was planned to cross-validate and calibrate the results by an independent method, - optical projection tomography (OPT). This novel ex vivo imaging technology offers unprecedented power for 3D whole-organ imaging, but with a resolution approaching the cellular scale. Within VIBRANT, this unique technology was further improved and optimized. Because of insufficient effectivity of ligand functionalized NCs in vivo, surrogate studies were performed demonstrating the feasibility of the concept for in vivo BCM imaging and quantification in type 1 and type 2 diabetes disease models. In summary, a fifth result of VIBRANT is the independent cross-validation and confirmation of MRI results by OPT in a proof-of-principle study.
In summary, a sixth result of VIBRANT is the consolidation of the developed nanoparticular agents, processes and associated methodology in view of comprehensive IP protection, the dissemination to the scientific community and the societal stakeholders, and the commercialization of nanoparticular components (e.g. STREM Chemicals Inc.). Three patent applications were filed, besides 56 peer-reviewed research publications, about 80 conference contributions, and 7 articles and interviews to the general public. An Industrial Advisory Board (IAB) associated to VIBRANT (Servier, Lilly, BMS/Amylin and JDRF) will facilitate uptake after the end of VIBRANT for follow-up cooperations with partners from the pharmaceutical industry.

Overall strategy

The VIBRANT workplan was structured in the form of four major strategic segments (see Figure 2), which combine in a comprehensive framework and corresponded to the major development phases of the value chain in the project:
(see Fig. 2: VIBRANT PERT Chart)


As a basis and starting point of the project work, the first segment (I) consisting of WPs 1 to 3, addressed all aspects in context with the synthesis of novel ligand functionalized nanocontainers. As a result, ligand functionalized NCs, containing 19F or superparamagnetic MRI contrast agents, optionally combined with fluorescent QDs, as well as the associated technology to produce, characterize and iteratively optimize these constructs were provided.
The second segment (II), WP 4 to 6, built on segment I, pursuing the testing and optimization of ligand functionalized NCs with regard to β-cell specific targeting, sensitivity, spatial resolution, toxicity, signal stability, and potential quantification for diagnostic non-invasive in vivo imaging of the pancreas. As a result, stable, nontoxic NCs with high sensitivity and high spatial resolution power for imaging were provided, demonstrating β-cell specificity in vitro. However, translation of these findings into in vivo was not fully conclusive.
The third segment (III), consisting of WP 7, leveraged on the observations made in part (II) in order to explore ligand functionalized NCs as vehicles for the targeted delivery of therapeutics to the pancreatic β-cell. A robust 3D MR imaging of the mouse abdomen was developed for the analysis of ß-cell function/ mass and measurement techniques of the mouse abdomen with manganese in two animal models of diabetes. Targeting of ß-cell survival and function in vivo was studied. As a result, first evidence that drug-loaded nanocontainers specifically support pro-survival in ß-cells was gained.
The fourth segment (IV) incorporated in WP 8, concluded the development work by providing an independent ex vivo cross validation of the in vivo experiments of part (II) and (III) using innovative OPT and SPIM technology. Due to the lack of functional NCs in vivo, surrogate studies were performed in MRI in such a way that labeled islets were grafted into pancreata and cross validated with OPT. As a result, precise superimposition of islet grafts and 3D quantification of BCM was demonstrated by optimized correlation and calibration algorithms. Additionally, novel correlative OPT/SPIM and near infrared OPT (NIR-OPT) instruments with increased multichannel capacities were engineered.

In summary, VIBRANT provided comprehensive technology for the manufacturing of nanoparticular probes, bearing biological recognition motifs, for multi-modal imaging and drug-delivery. Superparamagnetic iron oxide NCs were highly sensitive for the detection of single islets and superior to commercial materials. The fluorous phase MRI agents developed, offer new opportunities for improved imaging modalities.
Knowledge on ß-cell function, survival and regeneration was substantially expanded. During VIBRANT, highly sophisticated in vitro assays and in vivo models were developed and adapted to MRI. An independent cross-validation of the MRI results was accomplished by OPT. Mn2+-enhanced MRI can detect small changes in functional ß-cells in vivo and is a potential method for early noninvasive detection of changes in functional ß-cell mass.
Unfortunately, -at the time- in vivo β-cell imaging by means of nanoparticular probes was not achievable, possibly due to insufficient ß-cell affinity of the selected ligands. However, search for novel high-affinity ligands was not in the scope of VIBRANT. Once as such will become available, studies should be resumed, because all technical success factors have been developed. In particular, surrogate proof-of-principle studies clearly indicate feasibility of anatomical and functional pancreatic β-cell imaging and quantification of BCM. Promisingly, a small molecule fluorescent carbohydrate derivative AJ 070 showed specific labeling of islets in vitro and in vivo and will be further pursued.

Project Results:
3. VIBRANT main S&T results/foregrounds

WP 1 Component Synthesis

In workpackage 1 (WP 1) all necessary components, like superparamagnetic or fluorescent nanoparticles, a range of biocompatible polymers for encapsulation, various fluorous phases for fluorine (19F)-MRI, carbohydrate and proteinogenic bioligands for affinity labeling have been provided.

WP 1.1 Particle synthesis
Quantum dots: The highly fluorescent quantum dots used in VIBRANT were made using an innovative continuous flow reactor .

Compared to the common “batch” synthesis, this process leads to superior reproducibility of the particle properties with significance lower than 0.1%. Figure 1.1 shows absorption and emission measurements of three different batches of quantum dots.

Figure 1.1: Absorption and emission spectra of three different batches of quantum dots

Fluorescence and stability were improved by growing CdS and ZnS shells around the CdSe cores. Absorption and emission spectra of CdSe/CdS/ZnS particles can be tuned between 550 and 625 nm. The quantum yields are usually above 40%.

Figure 1.2: absorption and emission spectra of CdSe/CdS/ZnS particles

Recently, these materials were commercialized under the trademark CANdots® Series A and distributed by STREM Chemicals .

Iron oxide:

The superparamagnetic iron oxide nanoparticles (SPIOs) used within VIBRANT were produced with diameters between 4 and 20 nm and a narrow distribution smaller than 10%. Alternatively, continuous flow synthesis of SPIOs of different sizes was successful, albeit size distribution still leaves room for improvements.

Fig. 1.3 SPIOs

WP 1.2 Polymers
For biological applications, the hydrophobic inorganic nanoparticles synthesized as described above need to be encapsulated into a biocompatible polymer shell in order to render the particles water soluble. To this end, so-called amphiphilic diblock copolymers, comprised of a threadlike hydrophobic and a hydrophilic block each were used. In water, these polymers self- assemble to spherical micelles enclosing the nanoparticle in the center (see WP2). The hydrophobic blocks adhere to the nanoparticle surface, wherein the hydrophilic blocks point outwards to the aqueous environment, hence rendering this nanocontainer water soluble. The synthesis and properties of these amphiphilic polymers are described herein.

The hydrophilic block of all block copolymers consisted of polyethylene glycol (PEG), in combination with different hydrophobic blocks, namely poly lactic acid (PLA), poly caprolactone (PCL), poly caprolactone-co-glycolide PCLGL, or poly isoprene (PI). All block copolymers were synthesized by sequential living anionic polymerization and/or coordination-insertion polymerization, hence providing narrow dispersity and good control over molecular weight. Typical examples of poly(lactid acid-b-ethylene glycol) (PLA-PEG), poly(caprolactone-b-ethylene glycol) (PCL-PEG) and poly(isoprene-b-ethylene glycol) (PI-PEG) are shown in Fig. 1.4

Fig. 1.4 Diblockpolymers

The PEG-terminal OH-group was converted into terminal amino, isocyanate, alkynyl and carboxylic acid groups by standard functional group transformations for subsequent conjugation with biotin, carbohydrates, peptides, antibodies, or other putative ß-cell specific ligands.
Additionally, for 19F-MRI the standard diblock polymers were perfluorinated at the hydrophobic terminus. However, signal intensity was insufficient. In contrast, encapsulation of monomeric F-phases, e.g. perfluorooctyl bromide (PFOB), perfluoro-15-crown-5–ether, or perfluorohexyl octane (F6H8) provided satisfying signals (see WP3). Advantageously, most monomeric F-phases (PFOB, F6H8) are clinically approved.

WP 1.3 Carbohydrate Ligands

Specific labeling of pancreatic ß-cells is extremely challenging, because of the scarcity of ß-cell specific targets, e.g. receptors etc. In VIBRANT numerous potential targets from the literature were studied for their suitability. In particular, it was reported that streptozotocin (STZ) as well as mannoheptulose and their derivatives show selectivity for the GLUT2 transporter of ß-cells. Accordingly, various series of optionally fluorinated gluco- or manno-heptulose derivatives were synthesized, with linkers for the straightforward conjugation to the nanocontainers.
These derivatives either resemble the (desnitroso) urea pharmacophore of the STZ, or a glycosidic bond (Fig.1.5).

Fig. 1.5: STZ and heptulose derivatives

Additionally, regioisomeric F-heptuloses were synthesized and successfully tested in MRI (in vitro and in vivo). Notably, a non-toxic fluorescent mannoheptuloside AJ070 showed significant binding to islets (in vitro and in vivo). A novel SPR-based assay for surface analysis of glycosylated NPs revealed strong cooperative binding effects (KD 200.000 higher than for monomeric saccharides).

WP 1.4 Proteinogenic Affinity Ligands

Within VIBRANT, two antibodies, namely GLUT2ab and the recently described ß-cell specific single chain antibody SCA B5 , as well as various GLP-1 peptide derivatives were procured for conjugation to nanocontainers and subsequent biological testing. In a first optimization cycle a uniform conjugation protocol was chosen based on the well-known biotin/NeutrAvidin dyad. Commercially available biotinylated GLUT2 antibody was selected for proof of principle experiments, because within the pancreas, GLUT2 is exclusively expressed in ß-cells. The full length antibody was chosen for benchmarking, temporarily ignoring potential immunogenicity and putative co-expression of GLUT2 in adjacent tissues like liver, kidney and the small intestine. Direct covalent, as well as biotin/NeutrAvidin conjugation was accomplished. As an additional target, a recently published single-chain antibody SCA B5, claiming high ß-cell specificity was chosen. To this end, the construction of an 8x-histidine, thrombin cleavage site and biotin AviTag labeled version of the published SCA B5 amino acid sequence was achieved.

WP 2: Nanocontainer Assembly

According to the VIBRANT workplan, in WP2 all components procured in WP1 were assembled to a functional nanocontainer. This requires the phase transfer of the hydrophobic nanoparticle, e.g. fluorescent quantum dots (QD) or super-paramagnetic iron oxide nanoparticles (SPIO) or liquid fluorous phase from the organic phase into water, as well as bio-conjugation to the affinity ligand. A standard operation procedure (SOP) was developed, using a continuous flow phase transfer process, which provides highly reproducible nanocontainers with hydrodynamic diameter in the range of 30 to 120 nm at comparably high production volumes. A subsequent purification step removed by-products or empty micelles. Numerous series of fluorinated and non-fluorinated micellar NCs have been prepared using different methodologies, offering optimized solutions for either diagnostic (contrast agent) or therapeutic (controlled release) applications.
Nanocontainers comprised of polymers containing carbon-carbon double bonds, like PI-b-PEG, offer the option of radical induced crosslinking, thus providing exceptionally stable nanocontainers. Evidently, crosslinked NCs offer benefits e.g. increased circulation half live and longitudinal signal stability for imaging, whereas crosslinking is rather detrimental for controlled release. For the latter, nanocontainers comprised of biodegradable polyester-b-PEG polymers, mainly poly(lactic acid-b-ethylene glycol) (PLA-PEG) and poly(caprolactone-b-ethylene glycol) (PCL-PEG), were developed (Fig. 2.1).

Fig. 2.1: Optimized polymer shells for contrast agents or drug delivery

All polymers in use were biocompatible and functionalization was achievable by various methods (see below). For diagnostic applications, highly fluorescent quantum dots (QDs) and superparamagnetic iron oxide nanoparticles (SPIONs), as well as monomeric F-phases were optionally co-encapsulated into FPNCs or non-fluorinated NCs. With regard to fluorinated nanocontainers, only monomeric F-phases provided sufficient NMR/MRI signal (see WP3). However, polymeric F-phases, e.g. RF-PCL-PEO-X, RF-PLA-PEO-X or RF-PEO-X, strongly facilitated solubilization of monomeric F-phases.

Three different methodologies for the assembly and functionalization of the nanocontainers were developed (Fig 2.2 and Fig. 2.3):

1. Post-Assembly
A nanocontainer is formed in a first step and functionalized with a bio-ligand thereafter
2. Pre-Assembly
A polymer chain is functionalized with a bio-ligand and subsequently assembled to a nanocontainer
3. Seeded emulsion polymerisation (SEP) with “in situ” functionalization
A mixture of matrix monomers and tethered (bio)-ligands were co-polymerized around nanoparticles, hence forming (bio)-functionalized nanocontainers in a single step

The “post-assembly” approach was preferred for delicate affinity ligands, e.g. antibodies, which were conjugated to the full-fledged nanocontainer via biotin/Neutravidin, or via amino- or carboxy groups using conventional bioconjugation methodologies. However, the conjugation yield and ligand density on the surface of the NC was difficult to control and to analyze.
The “pre-assembly” strategy offered better control of the surface coverage of affinity ligands on the NC by employing defined mixtures of functionalized and unfunctionalized polymers. This approach was well suited for robust small molecules, e.g. pyranoses or biotin, which are sufficiently soluble in the required organic solvents and compatible with reaction conditions, like radical induced cross-linking and elevated temperatures. Furthermore, structural elucidation of the affinity molecule polymer conjugate by standard spectroscopy (e.g. NMR) prior to the assembly was usually possible, whereas these methods, when applied to a full-fledged NC, generally failed.

Fig. 2.2: Assembly strategies

Alternatively, using seeded emulsion polymerisation (SEP) with “in situ” functionalization by transferring hydrophobic nanoparticles into water, using surfmers (i.e. polymerizable surfactants) and adding variable mixtures of matrix forming monomers, e.g. styrene or divinyl benzene in combination with bioligands attached to a polymerizable tether, provided excellent results in a single polymerization step (vide infra).

Fig. 2.3: Encapsulation of QDs by SEP: I) PI-N3 coated QD; II) QD in PI-b-PEO micelle after phase transfer; III) QD encapsulated in polystyrene: A) without additional functionality; B) with copolymerized functional monomer; C) with functional surfmer.

The continuous flow phase transfer process developed within VIBRANT, not only allows precise tuning of the NCs’ size distribution and surface coverage, but also provides control over the numbers of nanoparticles enclosed. Hence, by tuning the process parameters, e.g. absolute and relative concentrations of the components or the geometries of the mixing chamber, all kinds of NCs, either bearing a singular nanocrystal, or enclosing a large number of nanocrystals were deliberately accomplished. Clustering of nanocrystals offers advantages, either by clustering different e.g. fluorescent and superparamagnetic nanocrystals providing hybrids for multichannel read-outs, or by clustering solely superparamagnetic nanocrystals for optimal MRI contrast (Fig. 2.4 see WP3)

Fig. 2.4: TEM images of single and clustered 9.8 nm SPIONs. The size bar is equivalent with 100 nm. Determined dhyd by DLS measurements were 59 nm (1.), 73 nm (2.), 126 nm (3.), and 132 nm (4.).

However, the increased hydrodynamic diameter of the clusters has to be considered with respect to the biological properties, e.g. passing biological barriers (see WP5).
Overall, within VIBRANT, more than 100 diverse nanocontainers were prepared, physico-chemically characterized and subsequently tested in a staggered panel of biological assays.

WP 3: Physico-chemical Characterization

In WP 3 the diverse series of nanocontainers (NCs) prepared in WP 2, were characterized by spectroscopic, electron microscopic and magnetic analytical methods. Furthermore, NCs were evaluated and optimized with regard to their suitability as contrast agents for cellular imaging.

19F-MRI contrast media

Initially, two series of PLA/PCL-PEG block copolymers with terminal perfluorinated appendages or perfluorinated blocks (e.g. RF-PCL-PEO-X and RF-PLA-PEO-X, see WP1) were successfully synthesized and used for the preparation of stable micelles with a hydrodynamic diameter of about 50 nm (WP 2). However, both series of micellar FPNCs gave no usable 19F signals, assumedly to the restricted motility of polymer-bound F-atoms.

In contrast, encapsulation of monomeric F-phases, e.g. perfluoro octyl bromide (Perflubron®, PFOB) or perfluoro-15-crown-5-ether (CellSense ®, PF-15-c5) succeeded in stable micelles providing high sensitivity in 19F-MRI. The detectability limit of at least 5 x 1015 equivalent 19F atoms per voxel was established with the help of radio-frequency coils especially developed within VIBRANT (see also WP6). In particular, the PFOB containing PI-PEO micelles showed good stability within a tuneable hydrodynamic diameter of 30 to 110 nm, depending on the PFOB content. However, PFOB displays a multi-signal 19F NMR spectrum, due to non-equivalent F-atoms. Thus “ghosts”, i.e. several images of the same object, were detected in 19F MRI, provided that the usual large bandwidth of 8000Hz was used. Elimination of these ghosting artefacts succeeded, by using selective excitation pulses with a reduced bandwidth of 1000Hz.
Favourably, PF-15-c5 contains 20 equivalent F-atoms, thus providing a singlet, most suitable for MRI. Routine encapsulation with PI-b-PEG initially failed, but eventually was successful by using the partially fluorinated polymers RF-PCL-PEO-X and RF-PLA-PEO-X, or by means of a semi fluorinated solubilizer perfluorohexyl octane (F6H8). Quantification of the 19F-MRI signal was achieved by comparison with an external standard (e.g. calibration with a 5-fluoro cytosine or NaF solution).

Additionally, libraries of fluorinated mannoheptulose and glucoheptulose derivatives were synthesized (WP 1) and their cellular uptake in hepatocytes, INS-1 cells and pancreatic islets was studied (WP 4) . Cell labeling with intracellular concentrations of up to 20mM (up to 1012 19F per cell) of fluoro sugars was sufficient for MR imaging. The detectability of labeled cells was considerably improved by using custom build radio-frequency coils (detectability limit 20,000 cell µl-1). 19F-labelled mannoheptuloses showed high specificity to the GLUT2-transporter, thus suggesting their utility as novel imaging agents.

Iron oxide contrast media (SPIOs)

Iron oxide containing contrast agents were assessed to confirm their size, size distribution (monodispersity), magnetic properties and potential clustering using TEM, dynamic light scattering, determination of the blocking temperature, squid, MRI and EPR (see Figure 3.1). For better uptake/ binding, quantification, sensitivity and biocompatibility, it was our aim to obtain monodisperse, superparamagnetic particles with high T2 relaxivity. Iron oxide based particles developed in VIBRANT were superior to commercially available particles in all aspects.
In order to further optimize the already near perfect MR contrast, we assessed the dependency of MR contrast on the size of nanoparticles. Hereby, we used PEGylated iron oxide based particles and magnetoliposomes synthesized via the seed-mediated growth method, providing mono-disperse particles of different sizes. As can be seen, the larger particles obtained after two seed-mediated growth steps showed superior MRI properties compared to the original seed particles (Fig. 3.1 right column).

Fig. 3.1: (A),(B) Bright field TEM images and (C),(D) corresponding histogramps of Seed-mPEGSi (A),(C), and SMG²-mPEGSi (B),(D). Scale bar of the TEM images is 200 nm. The histograms, plotted as diameter versus number frequency, were fitted with a lognormal curve. (E),(F) Normalized magnetic hysteresis curves of Seed-mPEGSi (E) and SMG²-mPEGSi (F) measured at 300 K. The grey curves are the corresponding Langevin curve fits.

Due to the intracellular degradation of the outer phospholipid bilayer of magnetoliposomes (ML), MLs are prone to intracellular clustering. This was confirmed by experiments that compared PEG-silane coated particles with MLs of the same core. Comparison of MR contrast of cells labelled initially with the same amount of iron oxide particles showed that MLs resulted in stronger contrast and contained particles for a longer time.

Fig. 3.2: TEM images of PEG-silanes and MLs taken up by MSCs confirm the intra-cellular clustering of MLs, which contributes to their better MR contrast and longitudinal in vivo monitoring. (B) Clustered MLs provide a better MR contrast as confirmed by re-laxivity measure-ments.

In order to transfer the benefical effects of clustering to biofunctionalized NCs, seeded emulsion polymerization (SEP), as described in WP2 was used, providing biofunctionalizable SPIO clusters with precisely defined properties (Fig 3.3).

Fig. 3.3: Clustering of SPIOs by seeded emulsion polymerisation.

Fluorescent contrast media (QDs)

Application of seeded emulsion polymerization (SEP) to the encapsulation of fluorescent quantum dots (QDs) provided nanocontainers (NCs) with remarkably improved stability against various reagents, e.g. metal ions, EDTA, aldehydes etc. which may be used in the preparation of biological specimens . This was a major progress, inevitable for the crossvalidation studies (WP8).

Hybrid contrast media

For multimodal read-out, as required for the crossvalidation studies of WP8, binary combinations of the foregoing described contrast media were tested. Again, the newly developed SEP process provided excellent results.

Magneto-fluorescent nanocontainers: QD/SPIO hybrids

The main challenge of hybridisation is to minimize mutual adverse effects of different nanoparticles, e.g. to prevent fluorescence quenching of QDs by the release of iron ions from SPIOs.
As can be seen in Fig. 3.5 SEP of spherical SPIOs and longitudinal quantum dot rods provided hybrids (A) preserving all properties of the disparate particles (B,C,D)

Fig. 3.5: A: TEM image of the QDQR iron oxide hybrids (inlay: hybrids under room light (left) and under UV light (right)); B. DLS histograms of the hybrids (volume distribution) of the encapsulated iron oxide particles before (dashed line) and after hybrid formation (solid line); C: black: absorption spectra of the hybrids (solid line), polystyrene encapsulated iron oxide nanocrystals (dashed line), QDQRs (dot dashed line); red: normalized emission spectra of the hybrids (solid line) and QDQRs solely (dashed line); D: confocal fluorescence microscopy images of the hybrids.

Fluorous phase-fluorescent nanocontainers: 19F/QD hybrids

Co-encapsulation of fluorous phase and QDs was successful. Fluorescence and 19F-signals were fully maintained without mutual impairment. The size of the NCs was tuneable via the fluorous phase/polymer ratio, e.g. at a ratio of 1 µL fluorous phase per mg polymer NC diameter increased to 100nm or higher.

19F/SPIO hybrids

Co-encapsulation of fluorous phase and SPIOs was technically feasible, but the 19F/SPIO hybrids did not provide any 19F-signal.
Two series of PI-PEG micelles loaded with PFOB and SPIOs of 10 nm or 6 nm diameters respectively were prepared. In both cases no 19F MRI-signal could be detected with standard NMR and MRI methods. In further experiments the PFOB/SPIO (10 nm) hybrids were subjected to "pulse-acquire" MR spectroscopy. Excitation was performed with a 10µs pulse and data were acquired after 2.5 µs. However, by 3 µs complete dephasing of the F-signal occurred. Consequently, further efforts towards 19F/SPIO hybrids were abandoned.

WP 4: In vitro screening of NCs on cell lines and primary cells

In WP 4 the nanocontainers (NCs) prepared in WP 2 and subsequently characterized in WP 3 were screened in cellular assays, with regard to toxicity, unspecific adhesion, ß-cell selectivity and cellular response.

Standard toxicity screening of NCs on human lung carcinoma cell line A549

The initial screen of all NCs was a standard toxicity assay on the human lung carcinoma A549 cell line. Cell viability and integrity were measured via WST-1 (water soluble tetrazolium) and LDH (lactate dehydrogenase) assay, respectively. This screening was well suited for the identification of putatively toxic by-products from NC manufacturing. After optimization and standardization of the preparation protocols, all NCs were nontoxic in the relevant concentration range. In particular, concentrations of the NCs up to 0.5 µM did not
• reduce cell proliferation or vitality
• cause cell lyses
• show optical/microscopic evidence for cellular changes
Preparations of NCs, which did not pass these criteria, either were purified again or excluded from further testing.

Testing NCs for toxicity on INS1 cell line using the xCELLigence platform

Further development of the NCs for possible in vivo applications required in-depth understanding of their effects on ß-cells. In particular, in later clinical applications the residual ß-cell mass of type 1 or type 2 diabetes patients may be impaired by metabolic and inflammatory stress. Consequently, a high content functional model was required, which preferably maps ß-cell biology and allowed screening of NCs on both unchallenged and challenged ß-cells and in ß-cells exposed to factors that affect ß-cell regeneration and replication. This was accomplished by adapting a novel real-time, on-line technology, the xCELLigence cell-analyser platform, offered by Roche. Different insulin producing ß-cell lines were tested, from which the INS1 was chosen as mostly suitable, thus establishing a powerful high content screening tool.
Routine screening of NCs typically showed no toxicity up to 250 nM, whereas at 1000 nM all NC were toxic to INS1 cells.
Furthermore, the INS1-xCelligence platform was remarkably sensitive for the detection of toxic byproducts. Even minute amounts of sodium azide, conventionally used for the preservation of NCs against bacterial contamination were detected.
Most notably, a unique profile was screened out for a NC conjugate to GLP1 (EU-0533), which solely showed a proliferative effect at concentrations of up to 100nM. Binding to INS1 cells was proven by fluorescence microscopy (Fig. 4.1).

Fig. 4.1: Viability of INS1 cells, when exposed to positive and negative controls, as well as EU-0533. 20000 INS1 cells were exposed to medium, cytokines (150 pg/ml IL-1ß and 5 ng/ml IFNγ), and the NPs were tested at 50 nM to 1000nM.

Toxicity screening of NCs on mesenchymal stem cells, MSC and multipotent adult progenitor cells, MAPC

Deriving insulin-expressing cells from Multipotent Adult Progenitor Cells (MAPC), induced pluripotent stem cells (iPSC) and other progenitors would be a major progress for the treatment of diabetes. Unfortunately so far, heterogeneous cell populations were obtained, from which insulin-expressing cells/ß-cells are difficult to discriminate. Therefore, NCs were tested for toxicity on these cells, in order to estimate their potential for specific labeling of insulin expressing cells derived from MAPCs. After careful purification, NCs were devoid of toxicity in the relevant concentration range.

Functional effects of NCs on pancreatic islets

The effects of NCs on glucose-induced insulin secretion were tested on freshly isolated rat islets. The non-targeted nanoparticles only affected adversely insulin release evoked by D-glucose or other nutrient secretagogues when used in high concentrations, i.e. in the 0.5 to 1.0 μM range in the case of NC (+ FeOx) and in the 0.3 to 1.0 μM range in the case of NC (+ QD). The NC (+ QD) and NC (+ FeOx), even when tested at an 1.0 μM concentration, only provoked minor changes in the metabolism in D-glucose (16.7 mM). Likewise, the functionalized nanoparticles all tested at a 50 nM concentration did not alter insulin output and D-glucose metabolism significantly, irrespective of the preparative route (post- vs. pre-assembly) and nanoparticle load (QD or SPIOs).

Cell Adhesion Tests

For reliably imaging of a small cell mass, like pancreatic ß-cells, unspecific adhesion or uptake of NCs needs to be minimized.
The PEG-coated unfunctionalized NCs were devoid of unspecific binding to cells in protein supplemented culture media. Unspecific binding was evoked, for NCs providing positive surface charge by protein depletion of the medium. The biotin/NeutrAvidin conjugation strategy for the attachment of the affinity ligands reduced unspecific background to less than 1%, even if incubation was performed in protein free medium. Conjugates obtained directly via heterobifunctional linkers were preferably incubated in full medium to reduce background.
In summary, a conjugation system was established that provided nontoxic nanoparticle-affinity molecule conjugates, virtually free of unspecific adhesion.

Specific binding of ligand functionalized NCs

NC conjugates providing GLUT2 antibody and a GLP-1 derivative proved specific binding on cell culture strains and tissue slides but not on isolated islets.
NC conjugates presenting the anti-GLUT2 antibody exhibited the strongest binding to INS1 cells. However, for the envisioned in vivo application putative immunogenicity, as well as co-expression of GLUT2 in adjacent tissues has to be considered. Nevertheless the anti-GLUT2 antibody conjugate was chosen for benchmarking.

The recombinant SCAB5 derivate proved binding to human and rat pancreatic islets in immunohistology experiments and may serve as a backup candidate. Unfortunately, after conjugation to NCs, the construct was dysfunctional.

Exendin-4-NC conjugates showed moderate binding. Notably, the K19-modified GLP-1-NC conjugate EU 0533 showed intense binding in INS1 cell culture even with short incubation time (10 min), but adversely affected the cells. Surprisingly, in the xCELLigence assay EU 0533 showed no toxicity, but even improved cell viability compared to the control. Unfortunately, these promising findings at the end of VIBRANT could not studied in detail anymore (Fig.4.2).

Fig. 4.2: Staining of pancreatic islets is clearly visible for the GLUT-2 (A) and KGLP1 (B) probes. Exendin-4 probe did not stain the islets on the slides (C).

Additionally, a fluorescent mannoheptulose conjugate AJ070, not in the scope of a nanoparticular probe, showed accumulation in ß-cells in vitro and in vivo and holds the potential for a novel low molecular weight imaging agent (see WP 5).

In summary, two lead candidates were identified for extensive in vitro and in vivo evaluation.

WP 5: In vivo Biological Testing in the Anterior Chamber of the Eye (ACE)

One of the main goals of VIBRANT was to develop strategies for the in vivo quantification of ß-cell mass (BCM), in order to monitor loss or gain of BCM as well as to assess the efficiency of specific therapeutic interventions on BCM. This quantification depends on the development of specific ß-cell labeling and detection techniques. This is a challenging task due to the fact that ß-cells are deeply embedded in the exocrine pancreas and therefore are difficult to access optically for labeling and functional studies.

When transplanted into the anterior chamber of the eye, pancreatic islets become fully vascularized and innervated. These engrafted islets reflect exactly the status of endogenous islets, thus key aspects of ß-cell morphology and function can be studied noninvasively using in vivo imaging techniques, due to the ideal optical and structural properties of the eye. The ACE model also offers a unique opportunity to noninvasively detect and visualize fluorescent NCs in islet transplants. In addition to the development of techniques allowing in vitro screening of functionalized fluorescent NCs for their specific binding to ß-cells, the ACE model permits to assess their potential binding properties under in vivo conditions.

Monitoring loss of BCM

The potential for non-invasively monitoring loss of BCM in the ACE model was clearly demonstrated on islet grafts expressing green fluorescent protein (GFP) in ß- cells. Cell death can be visualized by staining with annexin V conjugated to allophycocyanin (APC). If cell death was induced by i.v application of alloxan, the fluorescent staining was strongly increased. Backscatter imaging, which provides structural information about the morphology of endocrine cells, also clearly indicated a diminished BCM.
Loss of BCM was also observed in the BB rat, a model for human type 1 diabetes (T1DM). Genetically, in diabetes-prone (DP) BB strains, ß-cell decay and thus onset of diabetes occurs within a narrow, defined time-period, i.e. between 50-70 days of age (Holmberg R. et al., 2011, PNAS 108:10685-9). Diabetes-prone BB rat islets were transplanted into the ACE of DP-BB rats and allowed to vascularize and innervate. Similarly diabetes-resistant BB rat islets were transplanted into the ACE of diabetes-resistant BB rats. FITC-labeled dextran was injected i.v. for the visualization of islet vascularization, backscatter imaging was used to indicate morphological changes of islet grafts. At 100 days of age, islets transplanted into the diabetic BB rats vanished, while islets transplanted into non-diabetic BB controls were still intact and vascularized. Notably, the loss of BCM in the ACE coincided with the onset of diabetes, thus strongly suggesting that the dynamics and changes in BCM imaged in the eye mirror the changes observed in the native islets in the pancreas.
Importantly, the loss of BCM can be prevented by treatment of DP-BB rats with antisense-oligonucleotides targeting ApoCIII, which has been identified as one of the T1DM serum factors promoting ß-cell death (Holmberg R. et al., 2011, PNAS 108:10685-9). The lowered ApoCIII serum levels prolonged survival of islet grafts and thus delayed onset of T1DM after 93 days vs. 56 days in the non-treated group.

Monitoring gain of BCM

To monitor changes in BCM we developed a novel image analysis protocol permitting precise quantification of individual islet graft volumes in the anterior chamber of the eye over time. This protocol has been cross-validated by optical projection tomography (OPT), proving that imaging of islet grafts in the eye can successfully report on individual islet volumes and, importantly, on the possible increase or decrease in BCM.

Using this image analysis protocol we successfully reported on the BCM increase occurring in the leptin-deficient ob/ob mouse. This obese mouse is characterized by, among other things, its strong appetite. The increased demand for insulin leads to a quick increase in the number of ß-cells in the pancreas, thus to an increase in BCM. We were able to determine that the increase in ß-cell mass in the pancreas is reflected by the increase of ß-cell mass in the islets engrafted in the anterior chamber of the eye.
Therefore, because islets transplanted in the anterior chamber of the eye reflect endogenous pancreatic islet properties, the ACE model allows for monitoring changes in pancreatic BCM.

Validating fluorescent molecules for their specific binding to ß-cells using the ACE

The ACE model serves as a perfect imaging platform allowing for a precise assessment of pancreatic ß-cell mass regulation and future validation of functionalized NCs. After i.v. administration of fluorescent unfunctionalized NCs (EU0503) we could successfully image the ACE islet graft vasculature in vivo via confocal microscopy. This proves that fluorescent compounds can reach the islet graft for assessing their potential binding to ß-cells.

We then established an in vitro procedure to evaluate binding of functionalized NCs to insulin releasing cells. For validation, cells were incubated with exendin-4 conjugated to a small fluorescent dye. The fluorescence could be detected on mouse islet cells after 2 min incubation at a concentration of 100 nM. Displacement of the fluorescent compound in presence of non-labeled exendin-4 was successful, proving specificity of the ligand binding to islet cells (Fig. 5.1).

This targeting molecule was then assayed in vivo (Figure 5.2) and proved to be suitable both for in vitro and in vivo specific labeling of islet cells.

Fig. 5.2: In vivo imaging of islet transplant, 1 minute after i.v. injection of a fluorescently-labeled exendin-4

We assayed various labeling compounds based on anti-GLUT2, SCAB5 and GLP-1 derivatives, and conjugated either to FITC or to a nanocontainer (NC). Unfortunately none of them proved to exhibit specific binding properties to ß-cells (Fig. 5.3). Hence, with the proteinogenic affinity ligands available, no suitable NCs could be identified.

Diverse carbohydrate-modified NCs, in particular desnitroso-STZ conjugates, were tested but none was found to show specific binding to ß-cells either. We could however see an accumulation of signal in islet cells using a fluorescent mannoheptulose derivative (AJ070) both in vitro and in vivo (Fig. 5.4).

These promising results were achieved at the very end of VIBRANT and will be pursued further in a continued collaboration. In conclusion, we were able to develop and provide a reliable in vivo imaging platform for both the assessment of changes in BCM and for screening of compounds targeting specifically ß-cells in vivo.

WP 6 In vivo imaging technology

In WP6 imaging techniques were developed and optimized that allow non-invasive, in vivo imaging of pancreatic islets. Hereby, we have focused mainly on magnetic resonance imaging (MRI), which is readily available in the clinic. In addition, in vivo optical imaging techniques (mainly fluorescence imaging) were utilized for cross-validation in pre-clinical research.
In order to visualize the FPNCs developed and tested in WP1 to WP4, robust 1H and 19F MRI protocols for imaging the pancreas in experimental models of diabetes and controls were established and exchanged between partners. This includes on one hand the establishment of sensitive and possibly quantitative acquisition protocols but also purpose build hardware components like radio frequency coils for detection of 1H and 19F MR images of the rodent pancreas or the eye for the detection of FPNCs in the anterior chamber model (see WP5).

Detection of highly sensitive FPNCs by MRI and fluorescence imaging

To overcome limitations of individual imaging methods like the low spatial resolution of fluorescence imaging or the lack of absolute quantification of iron oxide based nanoparticles by MRI, bimodal FPNCs were used that target beta cells and contain imaging reporters for fluorescence (for example quantum dots) or MR Imaging (iron oxide based nanoparticles or fluorine containing compounds). For in vivo imaging, iron oxide based particles have proven to be highly sensitive, resulting in the detection of single islets. In combination with quantum dots, unspecific background signal in the MRI can clearly be distinguished from the signals originating from the FPNCs (see also Figure 6.1).

Figure 6.1: MR images and in vivo fluorescence images of FPNCs developed in the Vibrant project. The particles were functionalized to target GLUT2 (IR-anti GLUT2) or GLP1 (IR-K-GLP1) of the ß-cells. Those FPNCs contained iron oxide based nanoparticles and CANdot Series M quantum dots carrying a DyLight 680 labeled neutravidin linker, which was coupled to the respective targeting molecule or just biotin (IR-Biotin) as a control. Particles were systemically injected to test their specificity for ß-cells. Imaging was performed repeatedly between 2 to 24 hours after FPNC administration. As seen in the left panel, MR images resulted in a hypointense contrast in the liver, spleen as well as the pancreas region, indicating high sensitivity of the FPNCs. Fluorescence images (right panel) indicates fluorescence signal from the liver, spleen and pancreas region as also confirmed by ex vivo fluorescence images. Lack of specificity indicates that particles were also cleared via the spleen and liver.

Quantitative in vivo imaging using fluorine containing compounds and 19F MRI

FNPCs that contained fluorinated compounds (perfluoro carbons, PFC) or small molecules like deoxy-fluoro-mannoheptuloses were tested for their suitability for quantitative in vivo assessment of beta cell mass. 19F MRI proofed to be a suitable tool to assess the biodistribution of administered particles in vivo with high specificity (no background signal, see Figure 6.2). While PFC containing FPNC proofed to be highly sensitive, resulting in the detectability of less than ten islets (see Figure 6.3) deoxy-fluoro-mannoheptuloses showed high specificity to ß-cells and were able to reach them under in vivo conditions due to the small size of those molecules. However, the sensitivity of deoxy-fluoro-mannoheptuloses based in vivo islet detection was insufficient, due to the fact that those molecules contained only one fluorine atom and where cleared from the body relatively quickly.
It was possible to quantify the amount of fluorinated compounds and indirectly the amount of cells with in vivo 19F MRI when internal or external standards were applied.

Figure 6.2: Biodistribution of PFC containing FNPCs after systemic administration. The 19F MR image indicates clearance of the FPNCs via the liver. The 19F MR image indicates high specificity as no background signal was observed. Utilization of purpose build double tuned 1H/ 19F detectors allows overlay of those 19F MRI with 1H MRI to superimpose anatomical detail.

Figure 6.3: In vivo detection of engrafted pancreatic islets prelabelled with PFCs. The left panel shows an 1H MRI, indicating anatomical details in the abdominal cavity. The right panel shows an overlay of 1H and 19F MRI, indicating the location of the engrafted islets (color coded).

The application of 1H and 19F MRI to the anterior chamber model resulted in quantitative visualization of PFC labeled implanted islets, hereby providing an independent method for fluorescence methods (see Figure 6.4).

Figure 6.4: Top row from left to right: Anatomical image of the eye (1H MRI); visualization of engrafted islets using 19F MRI; overlay of the 1H and 19F MRI of the anterior chamber model and 19F NMR spectrum for indirect islet robust quantification. The fluorescent microscopy image below confirms successful labeling of the implanted islets.

Quantitative in vivo imaging using manganese-based MRI contrast

Manganese-based MRI protocols were developed for the visualization of tissue-specific accumulation of Mn2+ with the aim to visualize the rodent pancreas and the assess its function non-invasively. An optimized protocol of T1-weighted MR imaging of the abdomen fulfilled the requirements of least-invasive MR imaging of tissue specific Mn2+-accumulation (see also Figure 6.5). As presented under WP7, this method is sensitive to visualize non-invasively pathological changes during the progression of diabetes.

Figure 6.5: Typical time course experiment after Mn2+-injection.
A: Time course of MR images. Numbers indicate time relative to Mn2+-injection at t=0min.
B: Reference MR-image without inversion preparation.
C: Graph of relative signal intensities (reference: “B”) in tissue specific ROIs. Error bars represent the scatter of signal intensities in the selected ROIs, not the measurement error! Yellow asterisks indicate times of MR-images in “A”.

WP 7 Therapeutic Applications: in vitro and in vivo testing

In WP 7 we provided:
• Instrumentation for monitoring of morphological/ biochemical/ pharmacological changes, e.g. a robust method for 3D MR imaging of the mouse abdomen (in collaboration with WP 6)
• A thorough understanding of the molecular drug target for the therapeutic intervention

Putative drug candidates for targeted delivery to the ß-cell

Specific targeting of the ß-cell to improve ß-cell survival and function is currently one of the most challenging aims in diabetes therapy. Thus, delivering anti-apoptotic agents directly to the pancreatic islets via ß-cell specific drug-loaded nanocontainers, as proposed in VIBRANT, is a promising strategy. We have shown that inflammation processes play a pivotal role in diabetes. If islets are exposed, either directly to pro-inflammatory cytokines or are damaged through gluco- and lipotoxicity, cytokine release within the islets leads to ß-cell apoptosis and impaired function. According to our studies, three classes of drugs are able to prevent the activation of NFkB, being one of the major pro-apoptotic pathways in the ß-cell. Hence, the interleukin-antagonist IL-1Ra, the histone deacetylase inhibitor ITF2357 and c-Jun N-terminal kinase inhibitors restore ß-cell function and survival under diabetogenic conditions in animal models in vivo, and in human islets in vitro . Special polyester block copolymers such as poly(caprolactone-b-ethylene glycol) (PCL-PEG) and its copolymers with polyglycolide (PCL/PGL-b-PEG) were used for the encapsulation. It was shown that these nanocontainers collapse e.g. upon endosomal acidification and release their payload.
Robust 3D MR imaging of the mouse abdomen

In order to gain insight about the morphological and functional changes induced by external stimulation, a robust 3D MRI measurement technique, as well as an improved method for the analysis of ß-cell function/ mass and measurement techniques of the mouse abdomen were established . ß-cell function rather than ß-cell mass can be estimated by means of divalent manganese ions (Mn2+). Mn2+ (like Ca2+) enters metabolically active ß-cells through voltage-gated calcium channels, which open in response to a rise in extracellular glucose (Fig.1) . Therefore, Mn2+ uptake is glucose dependent and can be used in vivo to assess the functionality of both grafted and endogenous pancreatic islets. MRI signal intensities of paramagnetic Mn2+ and ß-cell function were well correlated, as demonstrated in diabetic and non-diabetic mice (Fig.7.2).

Fig.7.1: Mn2+ enters the ß-cell in a glucose dependent manner through Ca2+ channels

Representative 3D-MR images of a normal diet fed wild type (WT) C57Bl/6J mouse are shown in Figure 7.2 below. Circles mark our regions of interest in the pancreas.

No long-term effect of Mn2+ on glucose tolerance

To detect the effect of Mn2+ on glucose tolerance, MnCl2 or vehicle control was injected once weekly in 3 consecutive and ipGTT measured before, immediately after injection and at days 1 and 3. Despite the expected changes in ipGTT after MnCl2 injection, the effect was not severe and was washed out 2 days after each injection. The results show that also multiple Mn-injections had no long-term effects on glucose tolerance and can therefore be considered as safe tool, which does not affect glucose tolerance.
Measuring functional ß-cell mass by non-invasive MRI

To analyze ß-cell function and mass during diabetes progression, C57Bl/6 mice were fed a normal chow diet (ND) or a high fat high sucrose diet (HFD, Surwit) and were compared by analyzing in vivo-ipGTT, insulin secretion, Mn2+ signals in the MRI and ß-cell mass in fixed tissue.
We have shown before13b) that the HFD results in impaired glucose tolerance after 4 weeks impaired fasting glucose and hyperinsulinemia after 8 weeks and hyperglycaemia after 12 weeks of diet. Accordingly, ß-cells compensate and increase their mass from 4 to 12 weeks of diet, but ß-cell mass is decreased after 16 weeks of diet13i), MRI signals showed an initial induction after 1 week of HFD feeding, but from 4 weeks on, there was a decline in MRI signal intensity, which highly correlated with the impaired ß-cell function after 4 weeks of HFD feeding (Fig.7.3).

Fig.7.3: Relative pancreas signal normalized to liver signal in MRT during 30 min of analysis from age-matched normal diet (ND) and high fat diet (HFD) fed mice over 12 weeks. Each data point represents 3 mice, area under the curve (AUC) was analyzed from the 30 minutes of measurement.

Progressive impairment of ß-cell function and mass during 12 weeks of HFD correlates with the MRI signal

After 4 weeks of diet, glucose tolerance and insulin secretion were impaired, but mice compensate by increasing ß-cell mass. In line with the impaired function but in contrast to the increased ß-cell mass, Mn2+ uptake in the pancreas is lower in the HFD fed mice.
Data from the MRI analysis highly correlate with the functional measures of ß-cell function. Finally, ß-cell mass was analyzed in the same mice over 12 weeks. Again, the compensatory increase in ß-cell mass already after 1-3 weeks correlates with (1) the improved function and (2) the increased MRI signal. In contrast, ß-cell mass increases up to 8 weeks of diet, and after 12 weeks, a drop in ß-cell mass is observed. Such correlation was confirmed in another mouse model of diabetes.
Our results provide for a 1st time a non-invasive measurement of functional ß-cell mass during diabetes progression. We further provide here a method to reduce the number of mouse in an experiment, since ß-cell mass can be monitored in one mice over a period of time, which highly complies to the 3R (refinement, replacement, reduction) rules of experiments with laboratory animals.

In summary, we achieved:

• Robust 3D MR imaging of the mouse abdomen
• Targeting of ß-cell survival and function in vivo
• Analysis of ß-cell function/ mass and measurement techniques of the mouse abdomen with manganese in 2 animal models of diabetes
• Analysis of the first distributed FPNCs delivered by the project partners
o Treatment of ß-cells with functionalized particles in vitro
o Injection, recognition and quantification of functionalized particles in mice were accomplished.
• First evidence that drug-loaded NCs specifically support pro-survival in ß-cells.

CONCLUSION: Mn2+-enhanced MRI can detect small changes in functional ß-cells in vivo and is a potential method for early noninvasive detection of changes in functional ß-cell mass.

WP 8: Cross validation

For the development of systemically administered contrast agents intended for non-invasive imaging by MRI (as pursued within VIBRANT) the possibility to cross evaluate the contrast agent up-take specificity throughout the volume of the gland and to calibrate the non-invasive read out is of outermost importance. Optical Projection Tomography (OPT) imaging currently provides the highest resolution available for whole organ ex vivo imaging of islets in intact pancreata. In WP 8, this technique has been further improved to generally increase the quality of the generated data and to enable cross-validation studies of MRI imaging of pancreatic islets. Further, the cell-level resolution imaging technique Selective Plane Illumination Microscopy (SPIM) has been integrated into the OPT setup to enable complementary high-resolution cross validation studies. Proof of principle for the successful adaptation/ development of the two technologies for these undertakings has been obtained within WP8. However, the lack of functional nano-containers has impeded their full implementation in the project.

Fig. 8.1: Near-infrared OPT images of rodent pancreata A, OPT image of a Non Obese Diabetic (NOD) pancreas labeled for ASMA (blood vessels, red, Insulin (β-cells, blue) and CD3 (autoimmune T-cells, green) illustrating the multichannel capacity of the technique. B, OPT generated image of a mouse pancreas (left) and a rat pancreas (right). The rat pancreas is around 5 times larger than its mouse counterpart and could not be imaged intact by previous OPT setups. The islets are reconstructed based on the signal from insulin specific antibodies (red).

Development/optimization of protocols for OPT based cross evaluation.

New image processing assisted algorithms for OPT were developed that generally contribute to increase acquisition speed and quality of pancreatic OPT data. More specifically this included 1) algorithms for semi-automatic and precise positioning of a sample at the axis of rotation, 2) a fast and robust algorithm for determination of post alignment values throughout the specimen as compared to existing methods , and 3) a computational statistical approach implementing the concept of contrast limited adaptive histogram equalisation (CLAHE). The latter approach significantly increases both, the sensitivity of OPT for islet imaging and the conservation of islet morphology . Further, by adapting the OPT technology to enable imaging in the near infrared range the imaging depth and multichannel capacity of the technique could be significantly increased. The latter feature is a pre-requisite for conducting cross-evaluation experiments of MRI contrast agents (having a fluorescent label) . Finally, an approach was developed to improve structural data obtained from Optical Projection Tomography (OPT) using Image Fusion (IF) and contrast normalization. This enables the visualization of molecular expression patterns in biological specimens with highly variable contrast values.As such the technique may e.g. be used to study islet grafts in relation to vascular networks in experimental research.

Co-registration of OPT and MRI

The capacity for simultaneous detection of fluorochrome labelled MRI contrast agents and antibody labelled islets in “pancreas phantoms” by OPT was evaluated. Hereby islets were labelled ex vivo using fluorescent magnetic particles and grafted into excised pancreata to generate the correct anatomical context for imaging assessments. Using such embedded ex vivo pancreases, it was

Fig. 8.2: 3D Co-registration of OPT and MR images in a pancreas phantom harbouring ex vivo labelled islets. Panel A shows the co-registration of the OPT (purple) and the MR image (grey) of a pancreas. In panel B, the signal of prelabeled islets from OPT imaging is shown in red and the MRI signal is shown in green. The tissue outline (in grey) is reconstructed based on tissue autofluorescence using OPT. Artefacts from air bubbles trapped in the agarose have been manually removed. By the implementation of a set of computational techniques the absolute majority of islets detected by OPT can also be recognized by MRI showing that the technique may be used to directly evaluate up-take specificity of in vivo labelled islets and for calibration of non-invasive read outs. Note: Endogenous islets are not displayed in the image.

established that islets can be imaged by both methods. A pipeline (software) was developed to co-register OPT and MR images, segment the pancreas and quantify the number of pancreatic islets. As expected, MRI showed some false-positive signals due to intrinsic hypointense contrast. Those results confirmed that OPT is a valuable tool for future cross-validation of in vivo MRI of labelled islets. In the process of developing the cross-validation procedure, a range of tissue processing protocols were evaluated and a technique was identified that allows clearing of the sample for OPT and SPIM imaging as well as for MRI using “negative” contrast (manuscript in preparation).

Integration of SPIM into the OPT setup and high-resolution mesoscopic analyses.

A main objective for WP8 was the proof-of-principle for the utility of the 3D tiling approach. The tiling allows for gathering high resolution data of mesoscopic samples in a whole intact organ like the pancreas. Two mesoscopic imaging techniques (OPT and SPIM) were combined in one single setup. Hereby, to our knowledge, a worldwide unique setup was generated. Imaging with the new setup showed a significant higher resolution and detail in its sub-tiles compared to data that was acquired previously by OPT. Further improvement to the optical coupling of the resonant scan mirror and increased precision of sample positioning in the hybrid imaging setup allowed imaging an entire pancreatic lobe in high resolution with multispectral capabilities.

Fig. 8.3: SPIM scan of a gastric lobe of murine pancreas, showing a stitched maximum projection. Insets reveal the high resolution achieved. Blood vessels labelled in blue, alpha cells in green and beta-cells in red. The islets of Langerhans show an almost complementary composition of alpha- and beta-cells, as they are the major components in the islets.


Speier et al., Nature Medicine, 2008, 14, 574 – 578.
2 Niehaus JS et al., patent application WO2009101091 A1
5Ran C, et al., Angew. Chem. Int. Ed. 2007, 46, 8998-9001.
6 Leclercq-Meyer V, Kadiata MM, Malaisse WJ; Int J Mol Med. 2000,6(2),143-52.
7 Ueberberg S et al., Diabetes October 2009, 58 (10), 2324-2334
8 a) Poeselt E, Kloust H et al., patent application EP2585210, b) Kloust H et al. Langmuir 2012, 28, 7276-7281, c) ibid, 2013, 29 (15), 4915–4921).
9 Malaisse WJ, et al. Arch Biochem Biophys. 2012 Jan 15;517(2):138-43
10 Thiem J et al., patent application WO2012016935A1.
1 Kloust H, et al., J. Phys. Chem. (manuscript submitted)
12 Malaisse WJ et al,, Met Funct Res Diab, 2013, (Online) Vol 6, 1-5
13 a) Lundh M, Christensen DP, Rasmussen DN, Mascagni P, Dinarello CA, et al. Diabetologia (2010) 53, 2569-2578. b) Mandrup-Poulsen T, Pickersgill L, Donath MY; Nat Rev Endocrinol (2010) 6, 158-166. c) Lewis EC, Blaabjerg L, Storling J, Ronn SG, Mascagni P, et al. Mol Med (2011) 17, 369-377. d) Ablamunits V, Henegariu O, Hansen JB, Opare-Addo L, Preston-Hurlburt P, et al. Diabetes (2012) 61, 145-154.
e) Cavelti-Weder C, Babians-Brunner A, Keller C, Stahel MA, Kurz-Levin M, et al. Diabetes Care (2012) 35, 1654-1662. f) Lundh M, Christensen DP, Damgaard Nielsen M, Richardson SJ, Dahllof MS, et al. Diabetologia (2012) 55, 2421-2431. g) Maedler K, Dharmadhikari G, Schumann DM, Storling J Handb Exp Pharmacol (2011), 257-278. h) Ardestani A, Sauter NS, Paroni F, Dharmadhikari G, Cho JH, et al. J Biol Chem. 2011; 286(19), 17144–17155. i) . Owyang AM, Maedler K, Gross L, Yin J, Esposito L, et al. Endocrinology. 2010, 151(6), 2515-27. j) Sauter NS, Schulthess FT, Galasso R, Castellani LW, Maedler K; Endocrinology 2008, 149, 2208-2218.
14 a) Haase A, Matthaei D, Bartkowski R, Duhmke E, Leibfritz D; J Comput Assist Tomogr (1989) 13, 1036-1040. b) Kustermann E, Meyer A, Godbole A, Dreher W, Maedler K; 19th Annual Meeting & Exhibition International Society for Magnetic Resonance in Medicine (2010). c) Kustermann E, Meyer A, Godbole A, Dreher W, Maedler K; Proc Intl Soc Mag Reson Med 2011, 19, 810.
15 Malaisse WJ, Maedler K; Diabetes Res Clin Pract. 2012, 98 (1), 11–18.
16 Glas R, Sauter NS, Schulthess FT, Shu L, Oberholzer J, et al. Diabetologia 2009, 52, 1579-1588.
17 Cheddad et al., IEEE Transactions on medical imaging, 2012
18 Hörnblad A, Cheddad A, Ahlgren U; Islets 2011, 3, 204–208

Potential Impact:
In summary, VIBRANT provided comprehensive technology for the manufacturing of nanoparticular probes, bearing biological recognition motifs, for multi-modal imaging and drug-delivery. Superparamagnetic iron oxide NCs were highly sensitive for the detection of single islets and superior to commercial materials. The fluorous phase MRI agents developed, offer new opportunities for improved imaging modalities.
Knowledge on ß-cell function, survival and regeneration was substantially expanded. During VIBRANT, highly sophisticated in vitro assays and in vivo models were developed and adapted to MRI. An independent cross-validation of the MRI results was accomplished by OPT. Mn2+-enhanced MRI can detect small changes in functional ß-cells in vivo and is a potential method for early noninvasive detection of changes in functional ß-cell mass.
Unfortunately, -at the time- in vivo β-cell imaging by means of nanoparticular probes was not achievable, possibly due to insufficient ß-cell affinity of the selected ligands. However, search for novel high-affinity ligands was not in the scope of VIBRANT. Once as such will become available, studies should be resumed, because all technical success factors have been developed. In particular, surrogate proof-of-principle studies clearly indicate feasibility of anatomical and functional pancreatic β-cell imaging and quantification of BCM. Promisingly, a small molecule fluorescent carbohydrate derivative AJ 070 showed specific labeling of islets in vitro and in vivo and will be further pursued


Within VIBRANT, different kinds of nanoparticles have been developed and optimized for putative labeling of ß-cells, but eventually will find broad application in preclinical biomedical applications in general.

Fluorescent Quantum Dots

During the last decade, considerable efforts have been made in order to exploit the unique spectral properties of fluorescent nanocrystals -so-called semiconductor quantum dots (QDs) - in biological and preclinical applications. In particular, absorption and emission wavelengths of QDs are size-tuneable, providing extremely broad and intense absorption bands, very narrow emission bands and high fluorescence quantum yields. Conventional synthetic routes via batch processes suffer from poor size control, hence resulting in undesirable broadening of emission bands. Partner 1 (CAN GmbH) developed and optimized a proprietary continuous flow synthesis, providing superior control and reproducibility of the semiconductor core materials. As a result, these QDs are now commercialized under the trademark CANdots Series A ® and distributed by STREM Chemicals Inc.

Since QDs are hydrophobic after synthesis in high boiling organic solvents, a surface modification has to be done to render them water-soluble and biocompatible, without compromising the spectral properties. To this end, several different, often polymeric ligand systems have been developed. An optimized ligand system will provide superior colloidal stability in biological media, as well as high fluorescence quantum yields for extended periods. Additionally, particles will be straightforwardly functionalized. Both aims – facile functionalization and protection against quenching – are still major challenges. Within VIBRANT these goals have been fully achieved by partner 2 (UH, Weller group) by polymer encapsulation, providing highly fluorescent, functionalized nanocontainers with excellent stability and very low toxicity in biological media in vitro and in vivo. Furthermore, the probes show unrivalled stability against chemicals, frequently used in histology for tissue preparations, making them highly suitable for specific histological stainings. Finally, these probes will spur progress in life science and preclinical applications.

Superparamagnetic Iron Oxide Nanoparticles (SPIOs)

Inorganic nanocrystals are well-established in biological and medical applications.1 Especially the superparamagnetic iron oxide nanocrystals are of great interest and find application as contrast agents for T2- or T2*-, and recently even T1-weighted magnetic resonance imaging (MRI). The SPIOs developed during VIBRANT are highly sensitive for the detection of even single islets and superior to commercial materials in all aspects, as demonstrated by partner 9 (KULeuven). Additionally, protocols for the controlled clustering have been developed for optimized magnetic relaxivities. The SPIO-nanocontainers, as well as the associated technology will be further exploited in ongoing third party collaborations.

Plasmonic Gold nanoparticles

Albeit not in the original focus of VIBRANT, plasmonic gold nanoparticles were used as gauges for the endothelial fenestration of islets. During VIBRANT, gold NPs were developed marketable and are currently commercialized in conjugation kits, distributed by AppliChem.

Fluorous Phase Nanocontainers

19F-MRI is a quite recent concept for preclinical imaging. The main advantages is a hyperintense contrast, which is less prone to artifacts, can be quantified, and can be superimposed with anatomical images from 1H-MRI. Promising results have been shown for islet labeling. Hitherto, microemulsions of fluorine compounds suffered from low stability and high background due to unspecific cellular binding. Functionalization was unsatisfactory. Within VIBRANT, novel fluorous phase nanocontainers have been accomplished, which resolve these issues. A novel encapsulation process was developed and used in collaboration with partner 11 (Univ. Bayreuth). Further development will be needed from partners 1 and 9. The results will be confidential until patentability is evaluated. The fluorous phase nanocontainers and associated 19F imaging technology, developed by partner 9 (KULeuven) bear high potential in preclinical research.

Hybrid Nanocontainers

The concept of inorganic hybrid nanostructures is to combine the disparate physico chemical properties of diverse nanocrystals towards higher integrated functionalities, affording diagnostic modalities, otherwise not obtainable with the discrete nanomaterials. Undeniably, different kinds of nanoparticles should fully maintain, rather than mutually compromise their original properties, resulting in e.g. undesirable fluorescence quenching, energy transfer, or concurring light absorption processes, due to the proximity of fluorescent components to the other materials. Within VIBRANT, partner 2 (Weller group) succeeded in providing optimized magneto-fluorescent, as well as fluorescent F-phase nanocontainers for multimodal imaging. Furthermore, partner 1 accomplished superparamagnetic NIR probes for independent cross validation of biological specimens via MRI and NIR scanning, as demonstrated by partners 6 (Univ. Umea), 7 (CRG) and 9 (KULeuven). This concept requires additional research until commercial exploitation.

Biodegradable Polyester nanocontainers for Drug Delivery

Within VIBRANT, partner 11 (Univ. Bayreuth) developed bio-compatible and –degradable nanocontainers, which could be functionalized with affinity molecules and enclose a lipophilic payload, e.g. JNK-inhibitor SP600125 (1,9-Pyrazoloanthrone), as demonstrated in a proof-of-principle study together with partner 10 (Univ. Bremen). However, albeit very promising, considerable research will be required to further apply the concept in practice.


Partner 9 (KULeuven) succeeded in the encapsulation of premium quality SPIOs provided by partner 1 (CAN) into phospholipid liposomes, which were highly suitable for cell labeling, because of their low toxicity, hence superior to conventional agents. A product development for the scientific life science community appears feasible.


Testing NCs for toxicity on INS1 cell line using the xCELLigence platform

By adapting a commercial cell analyzer xCelligence (Roche) to insulin secreting cell lines (INS1) Partner 12 (Univ. Copenhagen) developed a high content/high capacity screen for toxicity and biological response of effectors and nanoparticular probes. This highly predictive screening could be of considerable interest for the pharmaceutical industry for the development of anti-diabetic drugs. A further development is under evaluation. Remarkably, a NC conjugate to GLP1 (EU-0533) showed a unique proliferative effect on INS1 cells at concentrations of up to 100nM. Binding to INS1 cells was proven by fluorescence microscopy. Further research is required for validation of these findings.

Islet Transplants in the Anterior Chamber of the Eye

When transplanted into the anterior chamber of the eye, pancreatic islets become fully vascularized and innervated. These engrafted islets reflect exactly the status of endogenous islets, thus key aspects of ß-cell morphology and function can be studied noninvasively using in vivo imaging techniques, due to the ideal optical and structural properties of the eye. The ACE model also offers a unique opportunity to noninvasively detect and visualize fluorescent NCs in islet transplants. In addition to the development of techniques allowing in vitro screening of functionalized fluorescent NCs for their specific binding to ß-cells, the ACE model permits to assess their potential binding properties under in vivo conditions. This unique method developed by partner 4 (Karolinska Institute) is of great potential for diagnostic and therapeutic applications in diabetes management. The further development and exploitation will be pursued by partner 4 and his associated spin-offs. Partner 9 accomplished an adaptation of this sophisticated model to MRI monitoring in combination with fluorous phase nanocontainers. This method holds high potential for the scientific community and will be further pursued in research collaborations.

19F MRI for the detection, targeting and quantification of β-cells in vivo

A methodology for in vivo 19F MRI for small animals was developed by Partner 9 (KU Leuven, U. Himmelreich). By using a series of regioisomeric fluorinated ketoheptuloses as a ß-cell specific targets in vivo provided by Partner 2 (Univ. Hamburg, J.Thimm/J.Thiem) functional imaging, concentration dependent sensitivity and detection limits were evaluated, providing structural/mechanistic insights to specificity as well as organ specificity.

Mn2+-enhanced MRI for the detection of small changes in functional ß-cells in vivo

In order to monitor ß-cell survival and function in vivo, a robust 3D MR imaging of the mouse abdomen by using Mn2+ ions was established by partner 10 (Uni Bremen). Manganese ions display similar properties as calcium, but in contrast are paramagnetic, thus allowing MRI. Consequently, analysis of ß-cell function/ mass and measurement techniques of the mouse abdomen with manganese in 2 animal models of diabetes was accomplished. Mn2+-enhanced MRI can detect small changes in functional ß-cells in vivo and is a potential method for early noninvasive detection of changes in functional ß-cell mass.

Integration of SPIM into the OPT setup and high-resolution mesoscopic analyses and co-registration of OPT and MRI

Optical Projection Tomography (OPT) imaging currently provides the highest resolution available for whole organ ex vivo imaging of islets in intact pancreata. Further, the cell-level resolution imaging technique Selective Plane Illumination Microscopy (SPIM) has been integrated into the OPT setup. By combination of these two mesoscopic imaging techniques (OPT and SPIM) in one single setup a worldwide unique instrument was generated by partners 6 (Univ. Umea) and 7 (Centre of Genomic Regulation, Barcelona). Imaging with the new setup showed a significant higher resolution and detail in its sub-tiles compared to data that was acquired previously by OPT. The instrument, together with the associated technology will be exploited by the inventor’s spin-off (Bioptonics). The method was further adapted for the cross-validation of MRI studies, demonstrated by pre-labeling of pancreata with magneto-fluorescent nanoparticles, provided by partner 1. This method, enabled by the tissue clearing protocols developed by partner 6, holds very high potential to set new standards for the independent cross-validation of imaging studies in general.

Synthesis of seven carbon sugars and derivatives (AJ070)

Partner 2 (Univ. Hamburg, J.Thimm/J.Thiem) synthesized manno-and glucoheptulose derivatives, optionally conjugated to a fluorescent label. Remarkably, such a conjugate, namely AJ070 showed fluorescent labeling of islets in vitro and in vivo in models provided by partner 4 (Karolinska Institute). Because of their potential, these experiments will be pursued further after the end of VIBRANT. After positive validation, such a reagent will be of major impact for functional and anatomical ß-cell imaging. Currently, the results will be kept confidential. After validation, protection of IPR in a patent application will be pursued. Further development will be done in collaboration with the pharmaceutical industry.

List of Websites:

VIBRANT public website:

Principal Investigators


Theo Schotten (Coordinator),
Katja Werner,
Jan Niehaus,

Dept. of Chemistry, Univ. Hamburg

Horst Weller,
Joachim Thiem,

Rolf Luft Research Center for Diabetes and Endocrinology, Karolinska Insitutet Stockholm

Per-Olof Berggren,
Ingo Leibiger,

Laboratory of Experimental Hormonology, Brussels Free University

Abdullah Sener,
Willy Malaisse,

Centre for Molecular Medicine, Umeå University

Ulf Ahlgren,

Centre for Genomic Regulation, Barcelona

James Sharpe,
Katholieke Universiteit Leuven

Centre for Biomolecular Interactions Bremen

Kathrin Maedler,
Ekkehard Küstermann,

Universität Bayreuth

Stefan Foerster,
Købnhavns Universitet

Thomas Mandrup-Poulsen,
Dan Ingemar Holmberg,