Skip to main content

Light-based functional in vivo monitoring of diseases related enzymes

Final Report Summary - LIVIMODE (Light-based functional in vivo monitoring of diseases related enzymes)


Executive Summary:

A major gap in molecular imaging approaches is the availability of target-specific and tissue-specific imaging probes. Molecular information of relevance for a multitude of diseases are often associated with an up-regulation or activation of specific disease-related factors. Among such disease-related factors, proteases are a class of enzymes, which is frequently highly up-regulated. The combination of their high disease-specific expression levels with their ability to cleave substrates at high turnover rates renders proteases attractive targets for imaging probes. Cleavage of a smart probe through the enzymatic reaction generates a disease specific signal (activatable probe), whereas the high protease expression level and high turnover rates provide significant signal amplification thereby increasing the sensitivity of the assay. Therefore, such selective smart imaging probes offer a great potential for monitoring disease progression and efficacy of therapeutic agents in vivo.

LIVIMODE has tackled those goals by developing a toolbox of novel, highly selective probes for 11 proteases, including cathepsins S, B, and K, MMP-9, MMP-12 and MMP-13, caspase-1, neutrophil elastase, FAP and metallocarboxypeptidases, with several of them being tested in vivo in animal models of cancer and osteoarthritis (OA). Most of the probes are based on the reverse-design principle, resulting in high turnover of the probes. To overcome one of the major problems of such probes, i.e. their rapid clearance from the disease site, where the cleavage occurs, we have applied lipidation and conjugation of the probes to polymer carriers, in particular linear polyglutamates (PGA) and dendritic polyglicerols (dPG).. Using these approaches, we have significantly improved the homing properties and/or retention time of the smart probes targeted the above proteases, thereby paving the path for further preclinical and clinical studies. Using one such probe, PGA-MA063 which detects the MMP-13 (a key collagenase) activity associated with OA, we have validated the proof-of-concept for the utilisation of such probe for diagnosing early stage of the disease and to monitoring the efficacy of an oral available MMP-13 inhibitor in the animal models. Moreover, the novel probes developed for cathepsins S and B and for FAP utilized in the same principle showed very promising results in imaging of non-invasive cancer.

The probes developed in our collaborative studies have proven to be excellent tools not only for in vivo imaging, but also for unravelling of many key cellular and biological functions of the target proteases in the complex systems such as in normal physiology. For instance, we have developed the first metallocarboxypeptidase probes, as well as neutrophil elastase probe that proved to be excellent tools for imaging their activities in situ. We have developed the first cell permeable and highly selective caspase-1 probe that should enable a much better understanding of the inflammatory processes such as inflammasome activation. We have further advanced the repertoire of imaging agents by developing highly specific tissue-targeting proteins DARPins (designed ankyrin repeat proteins) against FAP and conjugated them to imaging moieties, thereby making another class of imaging agents for cancer. Finally, we were able to make a step further developing a strategy for the synthesis of a dual reporter polymer conjugate bearing two probes that allow the combination of optical imaging and magnetic resonance imaging (MRI) strategies opening the field to further applications.

LIVIMODE developed an interdisciplinary and conceptually novel approach for the design, optimization and testing of smart activatable imaging agents for functional imaging of disease related molecular events, which can be easily adapted to a number of other diseases, including most of the inflammation-associated diseases such as atherosclerosis, chronic obstructive pulmonary disease and rheumatoid arthritis. Once fully developed, such selective agents will provide significant benefits both for preclinical and clinical drug discovery and diagnostic imaging of disease related events, and LIVIMODE has contributed a great part to those goals.

Project Context and Objectives:

Understanding the molecular processes in living cells during the formation and progression of diseases such as cancer, rheumatoid arthritis and osteoarthrosis is crucial for the development of new diagnostic tools and treatments. LIVIMODE (Light-based Functional In Vivo Monitoring of Disease-related Enzymes) is a European research consortium with scientists from universities, research organizations, and a global healthcare leader, Sanofi. LIVIMODE was developing novel powerful imaging agents which are intended for the direct visualization and investigation of disease-related molecular processes in living cells and in diagnostic imaging. The LIVIMODE consortium focused particularly on highly sensitive optical imaging methods, which allow such direct visualization of specific molecular events in the living cell. Typical examples are disease-related enzyme activities, among which proteases were the major focus of the project. Moreover, protease are now gaining importance also as important class of drug targets1. Such molecular tools as developed within the LIVIMODE project would therefore in a long run allow earlier diagnosis and improved treatment of diseases not limited only to cancer and osteoarthritis, but also amenable to other inflammation-associated diseases.

Design of highly selective smart probes for proteases

The major aim of the project was the preparation of highly selective probes for the determination of specific activities of proteases under pathological disease conditions. In contrast to the alternative probes described by Bogyo and others in the MMP field, these probes are turnover-based, making them particularly useful for the detection of activities of low abundant proteins because the signal accumulates over time.

In principle, there are two ways to obtain turnover-based probes, either by following a more conventional approach, redesigning identified substrates or applying the so-called reverse design strategy, using optimized inhibitors as starting point for protease substrate design. Both approaches were successfully applied to the synthesis of novel, highly selective protease probes for a wide variety of proteases, among them caspase-1, neutrophil elastase, fibroblast- activated protein (FAP or seprase), cathepsins S, B and K, MMP 9, 12 and 13 and different subtypes of metallo-carboxypeptidases, accounting for 11 proteases in total.

Probes were designed specifically for different biological tasks like serving the purpose of cellular or in vivo examinations and therefore the fluorophore pairs were carefully optimized. The probe design was mostly based on fluorescence resonance energy transfer (FRET) and probes were equipped with several different sets of fluorophores and quenchers for applications in in vitro systems and in intact organisms, respectively. The key criterion for optimization was excellent selectivity against similar proteases, as specificity of probes employed in pharmacological experiments is a key requirement to assess pathophysiological significance of a given target in a particular disease state. Positions tolerating attachment of flexible linker systems to allow attachment to various polymer supports by different chemical reactions were incorporated. In this respect the probes developed for MMP-9, MMP-12 and MMP-13 possess an unprecedented selectivity profile when tested on a set of 10 MMPs, while maintaining an extremely high turn-over number, even when grafted on biopolymers. In addition, we also probed the concept of lipidation for improved homing of probes to the respective sites of interest by successful synthesis of lipidated probes for five proteases, including cathepsins B and S and FAP, allowing pharmacological characterization and direct comparison to the alternatively applied conjugates (WP3 and 4). Particular attention was paid to the point that sufficient quantities of the respective probes could be provided to the partners performing conjugation to polymeric scaffolds or performing the biochemical and pharmacological characterization including the commercial partners in a timely fashion.

In total, more than 60 novel highly selective probes targeting 11 proteases of interest have been obtained and distributed to the partner institutions.

Polymer conjugation as a step towards improved in vivo properties of the probes

The main aim was to enhance the tissue and molecular selectivity of the low molecular weight (MW) FRET smart activatable probes synthesised as the availability of target-specific and tissue-specific imaging probes it was considered one of the most critical gaps in molecular imaging approaches. In LIVIMODE this goal has been achieved by means of a rational design of polymer therapeutics: (i) seeking selective enrichment in the tumor/joint environment through the enhanced permeability and retention (EPR) effect and by specific tissue-targeting residues (designed ankyrin repeat proteins (DARPins) (cancer), alendronate biphosphonates for bone targeting (OA), cartilage targeting oligopeptides (OA) or cyclic RGD peptides (cancer); (ii) improving pharmacokinetics of probes (optimised half-life, tissue distribution) and by (iii) optimising signal intensity by targeting retention of cleaved probe.

We have found dendritic polyglycerols (dPGs) and linear polyglutamates (PGAs) as delivery enhancers. For this reason, we have developed synthetic pathways to a plethora of functional PGA and dPGs with well-defined structure, adjustable MW and low dispersion. Additionally, various functionalities such as alkyne, azides, reactive disulphides and protected amines were introduced by “post-polymerization modification” yielding a set orthogonal reactive attachment sites.

Once the functionalities were introduced into the polymer backbones, they were used for the bioconjugation of highly specific FRET smart probes (substrates for cysteine cathepsins, metalloproteases (MMPs) and FAP, targeting moieties (including DARPins) and/or magnetic-resonance imaging (MRI) probes enabling the synthesis of tissue-specific polymeric based imaging agents.

In most cases the probes showed to be superior compared to the non-conjugated substrates in terms of enzyme and tissue specificity, showing the same or even improved selectivity for i.e. MMP-12, cathepsin B, or cathepsin S. One of the best results obtained within this consortium was indeed coming from a PGA-MMP13 substrate conjugate that was tested in an osteoarthritis (OA) mouse model, demonstrating the proof-of-concept that such probes can be used to monitor the therapeutic efficacy of an MMP-13 inhibitor in vivo

Finally, we were able to make a step further developing a strategy for the synthesis of a dual reporter polymer conjugate bearing two probes that allow the combination of optical imaging and magnetic resonance imaging (MRI) strategies opening the field to further applications.

Probe selectivity as a key point for in vivo applications

Biochemical profiling of the probes is a required step as it gives basic information about the probes specificity. Although it may seem that only the very selective probes are useful for further work, this is not really true as also the more broad spectrum probes may serve as excellent tools for elucidating protease function(s) in cellular or in vivo environment. Another important criterion is to evaluate cell permeability of the probes developed, although it is really important only for the probes that target intracellular proteases, such as caspases and cathepsins, but not for metalloproteases. However, even cathepsins can be often found extracellularly in various diseases, including various forms of arthritis and cancer. All the probes were therefore screened against panels of related enzymes and only the probes exhibiting high selectivity were selected for further work. Although modification of the probes through lipidation or conjugation to polymers sometimes changed the selectivity of the probes, we were still able to obtain a high number of selective probes for our 11 target enzymes. Moreover, we were able to obtain the first cell permeable caspase-1 probe that has a major potential in studies of inflammasome activation. However, also a number of other smart probes were cell permeable and retained their selectivity also in the more complex environment of a cell, thereby providing excellent tools for studying the physiological roles of the target proteases. Moreover, none of the probes tested exhibited any cytotoxicity in cellular assays, thereby diminishing the risk in for the in vivo studies. Finally, we were able to produce several selective protease-sensitive FRET probes in a lipidated form that permitted to compare enzyme activities on cell surfaces versus the open extracellular space. Neutrophil elastase, for instance, was shown to only be active on the surface of cells and not in the supernatant of native cells isolated from the mouse lung.

In vivo imaging of cancer and osteoarthritis

The main objective of the in vivo studies was to evaluate the imaging probes targeting cathepsins (B and S), MMPs, and FAP in animal models of cancer and osteoarthritis. Similarly, integrin-targeting probes have been developed to visualize activated endothelium in association with the tumor neo-angiogenesis. Cancer models used comprised colon (C51), mammary (4T1) and brain tumors /GL261) implanted either subcutaneously or orthotopically in mice. Biodistribution and pharmacokinetic studies were carried out for a number of cathepsin B and S specific probes to identify those providing the highest tumor-to-background signal ratio suitable for pharmacological studies. In the second generation of probes, where the probes were either lipidated or polymer conjugated, the in vivo kinetics and biodistribution was evaluated and compared with that of the nonmodified probes. Lipidation strongly enhanced the probe retention in the tumor both in case of the cathepsin S probes as well as the FAP probes. However, in the case of the cathepsin B probes, lipidation was of no advantage. Furthermore, it even reduced the probe retention capacity in mouse tumor models and also showed high hepatic clearance. Conjugation with dendrimer-based polymers further enhanced the performance in case of Cathepsin B specific probes. PGA/dPG conjugation did not enhance the probe performance and tissue retention substantially, though it delayed its clearance, thereby prolonging it’s half-life in the system. Among all the cathepsin probes evaluated, the cathepsin S lipidated probe HHY-297 was the most promising, followed by the non-lipidated cathepsin S probe HHY-276 and the cathepsin B probes (non-lipidated HHY-223 and lipidated HHY275). These proof-of-concept studies accomplished the major goals of the project in the cancer area. In addition, we were able to identify several probes (Cat B nonlipidated probe HHY223, Cat S lipidated probe HHY297, and functionalized and lipidated FAP probe MS19a) as potential candidates for evaluating the tumor microenvironment for their respective targets with good efficacy.

One of the phenotypic hallmarks of tumors is neoangiogenesis, which is triggered by hypoxia and requires remodelling of the tumor extracellular matrix mediated by proteases. We therefore complemented the protease project by developing imaging tools to monitor tumor hypoxia, hypoxia signaling and angiogenesis.

In the OA model we initially found that the probe following MMP-13 (collagenase 3) activity developed in the previous FP6 consortium CAMP was capable of detecting the disease progression in living mice. We also established that a significant amount of the polyglutamic acid (PGA) polymer was retained in the knee articular joints of mice following systemic delivery. The consortium subsequently developed an improved MMP-13 smart probe MA063 with higher sensitivity and specificity. Conjugation of this probe to PGA yielding PGA-MA063 that enabled the successful detection of in vivo signals at early stages of OA. Furthermore, we were able to use the PGA-MA063 probe to monitor the therapeutic efficacy of a collagenase 3 inhibitor in vivo.

The main achievements of the LIVIMODE project are summarized below:

- Consequent application of modern design strategies to obtain specific protease probes for eleven pathologically relevant proteases
- Systematic and thorough exploration of beneficial delivery effects of conjugation to various polymeric supports (size, polarity, loading, conjugation type etc.)
- Employment of a battery of highly sophisticated targeting principles to alter tissue distribution (DARPINS, a novel cartilage targeting concept, demonstration of lipidation as an extremely valuable technology: lipidated protease reporters have been shown to have some surprisingly favourable features. First, they reside easily in the outer leaflet of the plasma membrane and report with high spatial resolution. Second, they serve as strong homing tools to sequester fluorophores in tumours and hence visualize them.

- Thorough characterization of kinetic and specificity properties of the obtained probes to allow optimal and objective comparison of the different targeting principles, this wealth of data is rarely found in other approaches, though being key to assess overall probe utility
- State-of-the art in-vivo imaging in healthy and diseased animals to allow head-to head comparison in the most relevant condition
- Preparation of a rich toolbox of highly selective protease probes for eleven target proteases, among them highlights like:

- the first cell-permeable caspase-1 probe
- novel FAP and NE probes
- a highly specific MMP-13 probe
- a novel bimodal cartilage targeting agent, combining MRI and NIRF imaging
- outstanding tumor targeting properties for a lipidated cathepsin-S and a dendrimer conjugated cathepsin-B derivative
- some of the first metallocarboxypeptidase probes

- Proof of concept for the utility of the developed probes as efficacy markers in an MMP-13 study, non-invasively demonstrating efficacy of an oral MMP inhibitor in a chronic OA study

Finally the array of protease probes equipped with similar fluorophores and linkers can now be truly called a ‘protease reporter platform’, and is probably the first of its kind.

Project Results:

Comment:

Figures and Table for Description of the main S&T results/foregrounds are in attached file Description of MAIN ST Results Figures Table.pdf

Text, Figures and Table for Description of the main S&T results/foregrounds are in attached file Desc MAIN ST Results Text Figures Table.pdf

The LIVIMODE project aimed at developing specific tools for non-invasive optical in vivo imaging of disease related molecular events. In order to generate smart (activatable) imaging agents that change their fluorescence properties upon disease-related events, i.e. upregulation of specific proteases, LIVIMODE combined expertises in medicinal chemistry, pharmacology and imaging in a conceptually new approach in designing smart agents. The rationale for the selection of proteases was based on their relevance in disease processes and on the expertise of the partners acquired especially during the last decade. Specific focus was given to tumor hypoxia, angiogenesis and to the development of tumor microenvironment as critical factors in oncogenesis, and to inflammatory and degenerative processes in osteoarthritis, i.e. diseases that cannot yet be monitored with non-invasive technology.

For optimal efficacy, the work was divided into five Work packages (WP):

WP1: Enzyme Selective smart activatable imaging agents

WP2: Tissue-targeted selective smart activatable imaging agents

WP3: Biochemistry and cell biology

WP4A: Disease pharmacology: Monitoring tumor microenvironments, angiogenesis and hypoxia

WP4B: Disease pharmacology: Disease progression and cartilage protection in osteoarthritis

WP5: Project management and dissemination of results

The first two WP aimed at developing novel smart agents, WP3 aimed at producing all recombinant proteases required, enzymatic and cellular screening and profiling of the smart probes developed in the previous two WP and on preliminary cytotoxicity studies. In WP4 selected Smart probes were evaluated for their potential as non-invasive early diagnostic markers and as tools for monitoring disease progression in vivo focusing on cancer and on osteoarthritis. Finally, special emphasis was given also to identification of possible disease targets and to assessment of the therapeutic efficacy of compounds.

We will here describe the results generated in the WP1 to WP4. The dissemination of results related to WP5 is presented in the next section.

WP1: Enzyme Selective smart activatable imaging agents

The major goal of this workpackage was the preparation of highly selective probes for the determination of specific activities of proteases under pathological disease conditions. In contrast to alternative probes described by Bogyo, these probes are turnover-based, making them particularly useful for the detection of activities of low abundant proteins.

In principle, there are two ways to obtain turnover-based probes. The first is the conventional approach, starting from a substrate mapping effort, followed by identification of a promising starting sequence and characterization of the potential cleavage by other, related proteases. Careful optimization e.g. by determination of fragments resulting from cleavage by MS and subsequent alterations of the cleavage site e.g. by incorporation of non-natural amino acids should ultimately lead to specific substrates, only recognized by the protease of interest.

This strategy was successfully applied for the synthesis of highly selective FAP and cathepsin K probes; the latter one resulted from improvement of selectivity of an early FAP probe, for which the original activity on FAP was eliminated (EMBL and Sanofi). The CEA group focused on the development of additional smart probes for the zinc endopeptidases including MMP-9, MMP-12 and MMP-13, and was able to advance this area significantly. In specific, a series of substrates sequences with unnatural amino acids have been synthesized using both natural and unnatural amino acids. Exploiting electrostatic interactions for controlling the probe selectivity, a library of 16 peptides having the generic chemical structure:

Mca-Gly-Xaa1-Pro-Xaa2-Gly-Leu-Xaa3-Xaa4-Dpa-NH2

Where Xaa = Arg or Glu

has been prepared, all peptides were purified and were systematically evaluated on a set of 8 MMPs. These substrates were purified, fully characterised analytically and tested on a panel of MMPs. This led to 3 specific smart probes for MMP-9, MMP-12 and MMP-13.

The CEA and UAB groups optimized a series of phosphinic tri- or tetra-peptides as carboxypeptidase inhibitors of metallo-carboxypeptidases (CPs), to be used as basic elements for building activity (inhibitory)- based probes carrying a PEG linker and a fluorophor head. More than 15 specific inhibitors for metallo-carboxypeptidases (CPs) as well as 2 complete derived fluorescent probes for them have been obtained within this effort. Most of the constructs showed high affinity for the A-type CPs (A4, A1, A2...) with low nanomolar Ki, and some of them full discrimination for the B-type CPs (CPB, TAFI, CPD...). Interestingly some of these derivatives show cell-penetrating properties, a behaviour that could be useful for localization/targeting/diagnostics purposes.

The second approach is making use of the previously reported protease inhibitors derived from industry-based optimization efforts. Usually these inhibitors display excellent selectivity and good cellular permeability. By following principles of the reverse design2, the warhead is replaced by a scissile bond. Ideally selectivity is maintained while turning the former inhibitor molecule into a substrate for the protease of interest. Careful introduction of two fluorophores will result in a quenched substrate. Upon cleavage of the sequence by the target protease, the quenching residue will be removed and fluorescence will be detectable.

In addition, positions were identified tolerating attachment of flexible linkers allowing attachment of the probe to various polymers.

Using this approach, we obtained a wide variety of cathepsin B and S probes with optimized fluorophores for cellular and in vivo experiments and attachment points for conjugation. An example of application of the reverse design principles for the cathepsin S and B probes is shown in figures 3 and 4.

A similar approach, using pralnacasan as the starting compound, was successfully applied to the synthesis of caspase-1 probes, yielding novel cell-permeable and highly specific reporters. Cell-permeable smart probes for this enzyme will provide new means to monitor inflammasome assembly in cells and possibly in vivo.

As a further approach we explored the effect of probe lipidation to induce homing effects and increase residence time of the probe at the membranes. This new concept was evaluated successfully to visualize proteases that are selectively activated on the cell surface resulting in 5 lipidated smart probes (FAP, NE, cathepsins S, B and K). Lipidation thus proved to be an effective means to improve homing of selective probes to cell surfaces.

Incorporation of additional charges to prevent internalization of the probe before cleavage by the target protease was another strategy employed. Upon cleavage, the probes get internalized, leading to excellent signal in the target cell. Another very successful example of lipidation was demonstrated for cathepsin B in a tumor model (See WP4)

Collectively, we can conclude that we managed to develop a toolbox for measuring protease activity, which has now reached a highly sophisticated level and permits measuring activities in vitro, in intact cells as well as in living animals (mice).

WP2: Tissue-targeted selective smart activatable imaging agents

The main aim of WP2 was to enhance the tissue and molecular selectivity of the low molecular weight (Mw) smart activatable probes synthesised in WP1 as it was considered that one of the most critical gap in molecular imaging approaches is the availability of target-specific and tissue-specific imaging probes. In Livimode this goal has been achieved by means of a rational design of polymer therapeutics, in particular polymer conjugates3: (i) seeking selective enrichment in the tumor/joint environment through enhanced permeability and retention (EPR) effect and by specific tissue-targeting residues (i.e. highly specific designed ankyrin repeat proteins (DARPins)4, RGD peptides5, alendronates); (ii) improving pharmacokinetics of probes (optimised half-life, tissue distribution) and by, (iii) optimising signal intensity by targeting retention of cleaved probe.

The whole workpackage required extensive collaboration and good logistics between the groups involved in probe synthesis (WP1) those involved in enzymatic testing and in characterization in cells (WP3) and in in vivo models (WP4).

WP2 proposed dendritic polyglycerols (dPGs)6 and linear polyglutamates (PGAs)7 as delivery enhancers for many SMART probes, which could substantially increase tissue specificity and the internalization of active components specifically into targeted cells enhancing the specific activity of the whole drug delivery system and thereby decreasing the adverse side effects in vitro and in vivo.

Dendritic polyglycerols (dPGs), as characterized by tunable end group functionalities, defined topological 3D architecture, and inertness to non-specific interactions with biological systems, present a novel platform for next generation biomaterials6. Along with the synthesis of hyperbranched PG, fabrication routes to perfect dendrimers, dendrons, microgels, and hydrogels have been reported over the last decade (dendron and hyperbranched based polyglycerol in Figure 6). With the goal of preparing different polymeric architectures for coupling to SMART-probes, dendrons and hyperbranched polymers based in polyglycerol were synthesized.

Hyaluronic acid (HA). HA polymers are highly attractive as carrier for delivery of dyes to cancer cells. Recent reports showed excellent internalization of drug-loaded pegylated HA nanoparticles into various cancer cell types via receptor-mediated endocytosis, displaying little uptake in normal fibroblasts10. Furthermore, intra-articularly hyaluronic acid polymers display a significantly prolonged residence time in osteoarthritic knees, thereby representing an appealing option to be exploited for imaging purposes

Polyglutamates as linear polymeric platform. PGA-based polymers are considered good candidates for their use in drug delivery or imaging probes for several reasons7: (i) PGA is a biocompatible, water-soluble and a multivalent polymer; (ii) PGA has been already used for the design of anticancer conjugates (as single agents or in combination therapy) as well as conjugates with application in tissue regeneration; (iii) it is a polyanion with preferential accumulation in macrophages; (iii) It is biodegradable in presence of cathepsin B which allows use of high Mw in order to enhance passive targeting by the EPR effect and finally, (iv) its FDA approval is expected after approval of the PGA-paclitaxel conjugate, OpaxioTM for the treatment of various cancer in particular ovarian in Phase III clinical trials and glioblastoma in combination with radiotherapy.

We have developed synthetic pathways to a plethora functional polyglutamates with well-defined structure, adjustable molecular weight and low dispersity (D = Mw/Mn <1.2) applying the ring opening polymerization (ROP) of N-carboxyanhydrides (NCA)8. Additionally, the acid moieties of the polyglutamates can be activated with 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium (DMTMM) and various functionalities such as alkyne, azides, reactive disulphides or protected amines can be easily introduced by “post-polymerization modification” yielding a set orthogonal reactive attachment sites9 (Figure 7a). As an example, the incorporation of amino-PEG-azide was performed (Fig 7). The introduction of PEGylated units into the polymer backbone not only will allow us the inclusion of a spacer between the polymer and the corresponding bioactive compound, but also can modify the in vivo behaviour, biodistribution and therapeutic application of the selected smart probe. On the one hand, the azide and alkyne moieties were chosen to have suitable moieties for a click chemistry attachment of the corresponding bioactive compound with all the benefits that this kind of reaction provides. The Copper Catalyzed Azide/Alkyne Cycloaddition (CuAAC) has been extensively used for post-polymerization modification, because it provides high yields under mild conditions in aqueous as well as organic media; it presents high tolerance of functional groups (Fig 7b). On the other hand, the introduction of activated disulphides for thiol/maleimide chemistry, and protected amines, for acid/base bioconjugations was also successfully accomplished.

Once the functionalities were introduced into the polymer backbones, they were used for the bioconjugation of highly specific FRET smart probes (substrates for cysteine cathepsins, metalloproteases (MMPs), and Fibroblast activation protein (FAP or seprase)), targeting moieties (including DARPins) an/or Magnetic-resonance imaging (MRI) probes enabling the synthesis of tissue-specific polymeric based imaging agents. The conjugation reactions have been optimized in terms of choice of catalyst, reaction conditions (e.g. temperature, time, and solvent), spacer and probe payload. The final products have been fully characterized by several physic-chemical techniques including UV/Vis, GPC, HPLC, FPLC, NMR or aminoacid analysis. The conjugated probes were sent to WP3 for enzymatic testing, the ones with the best profiles in vitro in terms of selectivity and specificity against the specific protease target were selected and scaled up in order to understand the in vivo behavior in comparison with the free substrates. Therefore the conjugates were send back to WP3 for cellulr and toxicity studies and to WP4 for in vivo evaluation.

In most cases the probes showed to be superior compared to the non-conjugated substrate and commercially available small molecule probes in terms of enzyme specificity showing the same or even improved selectivity for i.e. MMP-12, cathepsin B or cathepsin S compared with the non-conjugated probes. One of the best results obtained within this consortium is indeed coming from a PGA-MMP13 substrate conjugate that was tested in an osteoarthritis (OA) mouse model (see details in WP4).

Active-Targeted Probe Conjugates. Following the same line as above-described, several technologies for the coupling of targeting moieties to the polymeric carriers were assayed. The targeting moieties used on frame of Livimode range from cartilage targeting oligopeptides (OA), alendronates biphosphonates for bone targeting (OA), cyclic RGD peptides for integrin targeting (Cancer) targeting and DARPins for FAP targeting (Cancer) (Figure 8).

For this purpose, pure recombinant DARPins for surface marker protein such as FAP were obtained. For the in vivo experiments, these DARPins were re-cloned as fusion proteins, containing a C-terminal mCherry as fluorescence marker. The binding of these fusion proteins against human and murine FAP was confirmed by SEC. Additional SPR measurements showed that the affinity was not affected by the C-terminal mCherry. The best mCherry DARPins were prepared for in vivo imaging (WP 4) and also conjugated to dPGs and PGA (Figure 8).

After conjugation, the targeted macromolecular probes were evaluated regarding their binding to FAP. SEC experiments with human FAP and purified DARPin-conjugates (DARPin-PGA-cy5.5 DARPin-PG-cy5.5 DARPin-mCherry-PGA and DARPin-mCheryy-PG) revealed that all DARPin-conjugates formed complexes with FAP. These conjugated DARPins were also sent to WP4 for in vivo analysis.

Finally, we were able to combine in one PGA conjugate the specificity provided by a CatS specific substrate, and RGD moieties for cancer targeting, achieving one of the main aims of this project. Importantly, we have even gone a step further developing a strategy for the synthesis of a polymeric platform of conjugates bearing two probes for the combination of optical imaging and magnetic resonance imaging (MRI) (Figure 9).

In conclusion, we can successfully report the use of dendritic polyglycerols and linear polyglutamates as carriers of highly specific smart probes for specific proteases MMPs (MMP-13, MMP-12, MMP-9), cysteine cathepsins (Cat B, Cat S), and FAP to enhance site-specific accumulation and even probe specificity in selected cases.

It is clear with these results that the combination of two research fields for a specific target has been proven to be successful.

WP3: Biochemistry and cell biology

Context and objective

This Workpackage focused on the production of recombinant proteases that were needed for the biochemical probe characterization, establishment of standard procedures for biochemical screening assays assuring that all the groups were using the same methodology, as well as on biochemical and cellular screening of all the probes generated in WP1 and WP2.

Biochemical profiling of the probes, which was the core activity of this workpackage, is a required step, which gives basic information about the probe selectivity. This allows to further select the selective and non-toxic probes, which can be then used in in vivo studies. However, it is not true that only the highly selective probes are useful for further work, as often the broad spectrum probes may serve as excellent tools for elucidating protease function in cellular or in vivo environment. An important goal was also to evaluate the cellular permeability of the probes, which are targeting the intracellular proteases, such as caspase-1, the cathepsins, and for a few carboxypeptidases, but is not really so important for most of metalloproteases such as matrix metalloproteases, which are secreted enzymes. However, even the cathepsins can be often found extracellularly in various diseases, in particular in inflammation-associated diseases including cancer and at least partially also osteoarthritis, which were both the major topic of Livimode. On the other hand, evaluating cellular toxicity is a necessary step for all the probes used in vivo.

The whole workpackage required extensive collaboration and good logistics between the groups involved in probe synthesis (WP1) and conjugation (WP2), and those involved in characterization (WP3), as the newly synthesized probes had initially to be profiled for selectivity in WP3, before sending them to WP2 for suitable conjugation. Finally, the conjugated probes were again profiled for selectivity and cell permeability in WP3. All the probes that were further used in in vivo studies were also found to be nontoxic on cell lines at the highest concentration tested.

In vitro testing of SMART probes

Initially, we established standard assays for selectivity profiling of probes against a panel of recombinant target enzymes, including the cathepsins, caspases, MMPs, ADAM-TS's, carboxypeptidases, FAP and legumain. The majority of probes generated were targeting the cathepsins and MMPs, whereas some work was also performed for other proteases, including carboxypeptidases, caspase-1, FAP and neutrophil elastase.

SMART probes against cysteine cathepsins

During this project several probes evaluated exhibited excellent selectivity for cathepsin B, cathepsin K or cathepsin S, which are the major cathepsins involved in cancer (cathepsins B and S) or osteoarthritis (cathepsin K). We were dealing with three different types of the probes based on the small molceule reverse-design principle, which were either lipidated or conjugated to polymeric scaffolds, thereby exploring a completely novel area within the design of activity-based probes. The major idea behind and a major goal of the project was to improve the in vivo properties of these probes, primarily their retention time in the tissue. After the initial in vitro screening for selectivity, a more detailed kinetic validation was performed for several of the probes (see below, Fig. 10 ).

In general, we have found that conjugation increased the Km value (usually for an order of magnitude), which is in good agreement with the enzyme selectivity profiling, whereas the turnover efficiency was usually decreased after conjugation. The main reason is likely the decreased accessibility of the probe as it is coupled to the polymer. Lipidation was found not to have a strong effect on the Km value, but drastically decreased the efficiency as seen from the kcat/KM values. The likely reason is changed hydrophobicity as the lipid side chain can also hinder the enzyme-probe interactions. This was a good starting point for further studies as a decreased turnover suggested a potentially increased retention time due to reduced elimination from the tissue. Next, several of the novel activity-based probes for cathepsins S and B were further investigated for in vitro selectivity and sensitivity in complex samples, such as cell lysates and conditioned media. Several of the probes for cathepsins B and S exhibited good results in the conditioned media from cocultures of cancer cells and macrophages, and their selectivity was confirmed by inhibition with E-64 a broad spectrum cysteine cathepsin inhibitor. Moreover, microscopy studies demonstrated that a number of probes were successfully internalized, largely by endocytosis (see below Fig. 11):

Finally, using fluorescent microscopy we have observed a vesicular pattern of all cell-permeable cleaved probes in all tested cell lines, which suggested that the cleavage occurred in the endolysosomal system. The most selective cathepsin S probe and its conjugate were found to be completely selective also in cellular experiment and thus suitable for further in vivo characterization. A significant amount of the Cy5.5 cleaved product was observed within 40 minutes, gradually increasing to cca 60 minutes, after which it remained constant. Such delayed and slow processing of the probes is highly promising for the in vivo work. Moreover, it also offers an excellent tool for cellular studies.

It can be concluded that the cathepsin probes tested represent an ideal platform for the development of novel SMART tools for monitoring of their activity ranging from in vitro to in vivo applications.

Caspase-1 probes

Using pralnacasan scaffold, a very selective and sensitive bioluminescent probe for caspase-1 was initially generated. However, the probe was not cell permeable, which lead to the design of the second generation of caspase-1 probes, which were fluorescent and based on the FRET principle. The new probes retained the excellent selectivity, but were also cell permeable, which is a major achievement, as these are the first cell-permeable probes ever made for caspase-1. Although the differences in signal increase in cellular experiments are rather small, they are statistically significant and in excellent agreement with the generally low expression levels of caspase-1 in immune cells.

MMP probes

Part of the project was devoted to the development of novel probes against matrix metalloproteases, which play important roles in the development of cancer and osteoarthritis. The major focus was on the development of smart probes for MMP-9, MMP-12 and MMP-13, which were all screened against a panel of 9 different MMPs. Finally, the work has focused on three probes, which exhibited high selectivity against the three MMPs, one for each. The probes were also conjugated to polymers and were found to retains the good selectivity against their target enzymes. Moreover, the probes were found to be stable in plasma and retained their selectivity even under these conditions (see below), thereby being suitable for further in vivo testing.

Carboxypeptidases probes

The synthesis of the carboxypeptidase probes was particulatly challenging because they require to have a free C-terminus, equivalent to the C-terminus in peptide-protein structures or mimetics, a point that limited the placement of the fluorophore and its desirable pairing with a quencher. Finally, a phosphinic tri- or tetra-peptide was adopted as the core structure to which a PEG linker and a fluorophore was attached (at the N-terminus). The designed precursors and probes (15+2 constructs) showed high inhibitory/binding capabilities (nanomolar levels of Ki), and alowed discrimination between the A- and B-types of the enzymes (CPs). The kinetic analyses indicated that the incorporation of the PEG linker decreased slightly the affinity for the enzymes, whilst the further bonding of the fluorophor clearly recovered it, probably because of its hydrophobic-sticky character. Although it was not intended initially, some of the probes showed unexpected property to internalize in cell cultures, something that could be taken as an advantage for localization and diagnostic purposes in future.

Cytotoxicity of the probes

Finally, all the probes that were taken into in vivo studies were also tested for their potential toxicity on cells. None of the smart probes exhibited any cytotoxicity at the concentration used in in vivo experiments.

WP 4 Disease pharmacology

The major aim of this workpackage was to evaluate the efficacy of the smart probes generated and tested in previous workpackages in vivo in animal models of disease with the major focus on cancer and osteoarthritis. A further goal of the project was to evaluate contribution of several proteases to the disease progression, and, if possible, to evaluate the probes in combination with a selective inhibitor. Due to the complexity of the topic, the work was split in two parts, one dealing with cancer and the second with osteoarthritis.

a) In vivo studies using cancer models

Initially, a several tumor models have been established for this project, including subcutaneously grown colon carcinoma (C51), glioma (GL261) and mammary carcinoma (4T1) as well as orthotopically grown glioma (GL261) and mammary carcinoma (4T1). In the next stage all the models have been evaluated for their suitability as tumor models for assessing protease activity using the commercially available protease probes, which served as a reference for the performance of the probes developed by the consortium, as well as with probes for assessing the vascular status (AngioSense and IntegriSense). Based on these results, the 4T1 mammary cancer model was selected for further work. In the course of the project, we have evaluated 24 probes developed within LIVIMODE with regard to their pharmacokinetic properties. Among the 17 smart probes, at least one for each enzyme was nonmodified, one lipidated and one conjugated to dPG and one to PGA (Table 1).

For Cathepsins B and S and Fibroblast activation protein (FAP), the novel smart probes, either nonfunctionalized or lipidated/conjugated were used. In addition, a series of FAP probes based on Darpins (Designed Ankyrin Repeat Proteins) has also been evaluated. PGA-based integrin v3 targeting probes have been applied for neo-angiogenesis measurements. All the probes have been evaluated for their biodistribution and pharmacokinetic properties. A set of commercial probes developed by Perkin Elmer targeting various aspects of tumor microenvironment were also evaluated as control.

Dendrimer conjugated cathepsin B probe enhances blood circulation and tumor retension in animal models of breast cancer:

The consortia probes for cathepsin B were in vivo imaged for enzyme kinetics and probe biodistribution, and compared with the commercial probes targeting cathepsin B activity. For all the probes the activity peaked 24h after probe injection, when the signal to noise ratios were maximal (Fig. 15a). The Cathepsin targeted probes fluorescent probes were conjugate to macromolecular scaffold in order to increase blood circulation time and therefore target exposure. Two types of scaffolds have been used, dendrimers (PG) and linear polymers (PGA). Conjugation with PG enhanced the performance in case of Cathepsin B-specific probes. PGA-PEG conjugation did not enhance the probe performance and tissue retention substantially, though it delayed its clearance thereby making the probe available in the system for a longer duration (Fig. 15b). Besides, the conjugation significantly affected the biodistribution of the tracer; for example the probe 223 was effectively cleared in the kidney while conjugation to PGA strongly enhanced renal elimination, while conjugation to PG led to predominant hepatic clearance. A representative image for probe injection (HHY223) over time in breast tumor (4T1) bearing animals is given in Figure 1c.

Lipidation improves the substrate targeting and tumor retension of cathepsin S probe: Polymer conjugation of Both Cat S and B probes changed the probe kinetic in vivo. Lipidating Cat S and B probes also changed the probe targeting to tumors effectively, with Cathepsin S-lipidated probe HHY-297 showing the best performance compared to lipidated Cathepsin B probe (Fig. 16a). Interestingly on histological analysis, for all the probes the signal localized to macrophages to a larger extent as compared to tumor cells (Fig. 16b).

MS19a, a funcitonalized and lipidated FAP probe is an ideal candidate for in vivo imaging of tumor-associated FAP:

The specificity of the FAP probes was evaluated in vivo by testing the MS19a (functional and lipidated) versus the MS17a (non functional) probes. followed by probe kinetics mice bearing 4T1 tumors. Inspite of our in vitro data showing a specificity of the MS19a probe toward FAP, the in vivo cleavage of the MS17a probe suggest an unspecific cleavage of the probe by some proteases like for instance cathepsin K. The FAP probes were then coupled to macromolecular carriers (polyglycerol and polyglutamate) and their biodistribution was by comparing with MS27 non conjugated probe. No significant difference were found in the probe distribution and cleavage on injection of SP16 (PolyGlutamate 152G.A -MS-27 conjugate) and SP17 (PGA184G.A -MS27 conjugate) in mice bearing 4T1 tumors. In parallel, PG10-MS27 (PolyGlycerol 10G -MS-27 conjugate) and PG130 (PG130 -MS27 conjugate) probes were evaluated in the same tumoral model and we found an increase in the non specific cleavage in the peritoneal cavity without improvement of the cleavage within the tumors. The fluorescence was observed in kinetic and the illustration was done at 24h. Thus in our in vivo studies the coupling of both polymers (polyglycerol and polyglutamate) on the probes did not improve the tumoral distribution.

In summary, lipidation strongly enhanced the probe retension for both cathepsin S and FAP probes. Although, lipidation was of no advantage in the case of cathepsin B probes, it reduced the probe retension capacity in mouse tumor models and also had a high hepatic clearance. Further on, conjugation with dendrimer-based polymers further enhanced the performance in case of cathepsin B specific probes. PGA/dPG conjugation did not enhance the probe performance and tissue retention substantially, though it delayed its clearance, thereby prolonging it’s half-life in the system. Among all the cathepsin probes evaluated, the cathepsin S lipidated probe HHY-297 had the best performance, followed by the non-lipidated cathepsin S probe HHY-276 and cathepsin B probes (non-lipidated HHY-223 and Lipidated HHY275).

b) In vivo studies in models of osteoarthritis:

MMP-13 is a valid target for OA probe development.

Initially we used the probes developed in the previous FP6 consortium, CAMP, to follow the activities of the matrix metalloproteinases (MMPs) in the mouse surgical model of OA. The probes were specific for MMP-12 (metalloelastase) and MMP-13 (collagenase 3). Using these probes, we determined that MMP-12 was not involved specifically during the development of disease. MMP-13, however, was specifically upregulated in OA disease joints. We therefore identified MMP-13 as the lead metalloprotease to develop novel smart activity probes towards.

Development of PGA-MA063 as the novel MMP-13 smart activity probe.

The second generation probes of MMP-13 were obtained by screening a small library of coumarin substrates containing amino acid with unnatural side chains. The peptide probe obtained in WP1 (MA063, partner 8) was highly selective for MMP-13. However, when we tested MA063 for its ability to detect MMP-13 actvity in the OA joints of live mice the signals were weak. The probe MA063 was then conjugated to polyglutamic acid (PGA) polymer developed in WP2 (PGA-MA063, partner 10). Signals were detected in OA joints after systemic injection of PGA-MA063 (Figure 18A). The activation of PGA-MA063 was inhibited in the presence of a specific MMP-13 inhibitor (Figure 18B).

Proof of concept that PGA-MA063 can be used to determine the efficacy of inhibitor treatment.

SA (partner 5) provided an orally available MMP-13 chemical inhibitor to determine if the smart probe PGA-MA063 could evaluate the inhibitor’s long term efficacy. We tested the inhibitor at two doses, an optimal daily dose of 30 mg/kg and a sub-optimal daily dose of 0.3 mg/kg on the surgical model of OA. The amount of PGA-MA063 activated at the optimal dose of inhibitor was comparable to that detected in the knee that has not undergone surgery. At the sub-optimal dose, there was progressively increasing amount PGA-MA063 activated (Figure 19). PGA-MA063 was successful in monitoring the efficacy MMP-13 inhibitor treatment. Based on these results, partners OX, CEA and CIF are in the process of a patent application for PGA-MA063.

References

1. B. Turk, Nat Rev Drug Disc. 2006, 5(9): 785-99.
2. A Watzke, G Kosec, M Kindermann, V Jeske, HP Nestler, V Turk, B Turk, KU Wendt Angew Chem Int Ed Engl. 2008, 47(2):406-9.
3. R Duncan and MJ Vicent. Adv. Drug Deliv. Rev. 2013, 65(1), 60-70; b) R. Haag and F Kratz, Angew.Chem. Int. Ed, 2006, 45, 1198-1215 c) F Canal, J Sanchis, MJ Vicent Curr Opin Biotechnol 22(6), 894-900, 2011.
4. HK Binz, et al., J Mol Biol, 2003. 332(2): p. 489-503; b) J Hanes and A. Plückthun, Proc Natl Acad Sci U S A, 1997. 94(10): p. 4937-4.
5. A Eldar-Boock, K Miller, J Sanchis, R Lupu, MJ Vicent, R Satchi-Fainaro. Biomaterials 2011, 32, 3862-3874.
6. M Calderón, MA Quadir, S.K. Sharma and R. Haag, Adv. Mater., 2010, 22, 190–218. B)A. Sunder, R. Hanselmann, H. Frey, and R. Mülhaupt, Macromolecules, 1999, 32, 4240-4246.
7. C. Li and S. Wallace Adv Drug Deliv Rev 2008, 60, 886-898; b) MJ Vicent and E Perez-Paya. J Med Chem. 2006 49, 3763-5; c) SD Chipman et al. Int J Nanomedicine 2006, 1, 375-383.
8. MJ Vicent, M Barz, F Canal, I Conejos-Sanchez, A Duro-Castano, R England. P201131713. b) I. Conejos-Sánchez, A. Duro-Castano, A. Birke, M. Barz, M.J. Vicent. Polym. Chem., 2013, 4 (11), 3182 – 3186.
9. M. Barz, A. Duro-Castano, M.J. Vicent. Polym. Chem., 2013, 4 (10), 2989 – 2994.
10. Choi Choi KY, Yoon HY, Kim JH, Bae SM, Park RW, Kang YM, Kim IS, Kwon IC, Choi K, Jeong SY, Kim K, Park JH. ACS Nano 2011 (11) 8251-8159.

Potential Impact:

Non-invasive quantitative imaging of disease-specific molecular markers has the potential to overcome the limitations of the currently available imaging technologies such as X-ray-CT, MRI or ultrasound that cannot be used to monitor molecular processes and to provide highly specific and potentially early indicators of ongoing disease processes. In particular, these approaches lack the capability to correctly assess the specific activity of a given target which can differ significantly between particular disease states, albeit expressed protein levels may be similar. Establishing such technologies is of critical relevance both for preclinical drug discovery and human diagnostics. Non-invasive monitoring of molecular events in vivo ultimately enables us to combine different imaging modalities into an integrated multimodular view on the tissue anatomy, physiology, metabolism and cellular and molecular status of the intact organism. The most critical gap in achieving these goals lies in the lack of target-specific and tissue-specific imaging probes.

The major goal of LIVIMODE was therefore to overcome this gap with a novel approach aiming towards highly specific smart (activatable) imaging agents. In LIVIMODE we exemplified this approach on a selected set of disease- relevant protease targets, with consequences for improved monitoring of life-threatening diseases such as cancer and chronic diseases such as degenerative joint disease. A long term goal, which is, however, beyond the reach of the LIVIMODE project, is to impact disease management at multiple levels, including early diagnosis and disease staging, development and characterization of novel disease-relevant animal models that are critical for therapy development, and finally the evaluation of novel therapeutic drug candidates. These goals could have only been achieved by gathering a highly interdisciplinary team of experts in medicinal chemistry, polymer sciences, molecular and cellular biology, biochemistry and structural biology, in vivo pharmacology and in vivo imaging. Moreover, this could only have been implemented at a pan-European level, since the broad panel of expertises required cannot be assembled in any national scientific community. The excellent collaboration between the academic and industry researchers, which is clearly visible from the project results has finally lead to the establishement of a highly versatile imaging platform that could reach well beyond the initial goals of the project.

Through validation of the imaging strategies several novel selective smart imaging probes revealed novel potential protease targets either as disease biomarkers for demonstrating proof-of-concept of mechanism, including cathepsins B and S, and FAP for cancer, as well as MMP-13 for osteoarthritis, MMP and also proof-of-efficacy for novel drug candidates, such as exemplified for the selective orally delivered MMP-13 inhibitor, developed by the Sanofi partner. Although LIVIMODE was focused on experimental studies in animal models of human disease, the fact that it was based on non-invasive imaging readouts has a potential to facilitate the translation of the concepts into the clinics, as demonstrated both for cancer and osteoarthritis.

One should not forget that after cardiovascular diseases, cancer is the main cause of death in developed countries. Moreover, cancer is one of the diseases, which leads to a significant reduction of the quality of life. Although the anti-cancer therapies have significantly advanced, there is still no universal cure of cancer. In addition to the classical chemotherapeutic agents targeting cancer cells, recent approaches for cancer therapy target the tumor microenvironment, i.e. the complex interplay between cancer cells and tumor-associated non-cancerous cells in the hypoxic tumor stroma. The majority of communications within the tumor microenvironment are mediated by proteases, working in a concerted manner. During cancer progression a number of proteases have been found to be overexpressed, thereby disturbing the normal balance between the proteases and their endogenous inhibitors and triggering activation of downstream effector molecules (e.g. cytokines, chemokines, growth factors, and cell and matrix adhesion molecules). Another major problem with cancer is its early detection and diagnosis. Currently there are only a few reliable biomarkers available, one of them being a protease (PSA for prostate cancer). Furthermore, current diagnostic imaging methods including MRI are of limited usage. Therefore, the design of novel anti-cancer treatment strategies depends on improved knowledge of molecular mechanisms underlying tumour cell scattering and spreading in ectopic tissue microenvironments in patients with cancer. Proteases, specifically overexpressed in the tumor microenvironment such as FAP, cathepsins B and S, and some metalloproteases, are therefore highly attractive targets for both diagnosis and, possibly, treatment. However, the new protease-targeting approaches need to be much more selective to avoid the problems of the first generation of the broad spectrum metalloprotease inhibitors (marimistat, batimastat,...) which all failed in advanced clinical trials. Several of the probes developed during the project, in particular those targeting fibroblast-associated protease (FAP) and cathepsins B and S, which were found to be highly selective, therefore represent a significant potential for such translational efforts of pharmaceutical industry in the future.

In contrast, OA is the most prevalent age-related degenerative joint disease. With the expanding ageing population, it imposes a major socio-economic burden on society. A key feature of OA is a gradual loss of articular cartilage and deformation of bone at the margin, which eventually results in the impairment of joint function and disability. Currently, there is no effective disease-modifying treatment except joint replacement surgery. The primary cause of cartilage destruction is elevated levels of active proteases, and those proteases are primarily produced by cartilage itself. It is therefore attractive to consider protease inhibitors as potential therapeutics. While on-going research continues to improve the design of selective protease inhibitors, one of the major limitations is the lack of sufficient methods to measure an improvement of the joint structure, and the current techniques only capture the late-stage disease. Thus, the development of diagnostic tools to identify early-stages of the disease and to monitor the efficacy of disease-modifying therapeutics of OA is a major challenge. For this purpose we have developed a non-invasive method to detect molecular changes (i.e. up-regulation of MMP-13 which degrades cartilage collagen) in the joint tissues before structural changes occurs. The smart probe PGA-MA063 proves a novel imaging agent to detect early OA and to monitor the efficacy of MMP-13 inhibitor in the mouse model of OA, thereby demonstrating a successful examination in the early preclinical stage. Based on these encouraging results PGA-MA063 offers the potential to be used in advanced preclinical models in larger animal models of OA (e.g. rabbit and dog) and eventually to further develop it for the diagnostic and disease monitoring tools in man. Moreover, a similar approach can be applied to develop similar in vivo imaging probes for other key metalloproteinases.

A further breakthrough of our project was achieved in smart probe development. There is a major debate about the design of the smart activity-based probes targeting proteases and their potential usage in the clinics. The two major design principles are (i) the covalent design, based on an inhibitory warhead, which sticks to the active site and remain bound to the protease, thereby achieving a stable and localized signal, but has a drawback of potential accumulation and covalent modification of the target which could potentially lead to adverse immune responses upon degradation of the modified enzyme, and (ii) the turnover design, which was also used in LIVIMODE. Whereas the main advantage of the turnover probes is their rapid turnover and signal amplification, their major drawback is their rapid clearance from the disease site, where the cleavage occurs. Nevertheless, Perkin-Elmer after their acquisition of VisEn Medical Inc., is advancing the first broad spectrum protease probe for monitoring cathepsin B activity, based on the turnover principle, into the clinical studies for cancer. Within LIVIMODE, we have overcome this problem by means of a rational design of polymer therapeutics through: (i) seeking selective enrichment in the tumor/joint environment through the enhanced permeability and retention (EPR) effect and by specific tissue-targeting residues (designed ankyrin repeat proteins (DARPins) (cancer), alendronate biphosphonates for bone targeting (OA), cartilage targeting oligopeptides (OA) or cyclic RGD peptides (cancer)); (ii) improving pharmacokinetics of probes (optimised half-life, tissue distribution) and by (iii) optimising signal intensity by targeting retention of cleaved probe. Using lipidation and conjugation to PGA or dPG polymers, we have significantly improved the homing properties and/or retention time of the smart turnover protease targeting probes, thereby overcoming the major problem and paving the path for further preclinical and clinical studies.

The probes developed do not offer only excellent tools for in vivo imaging of the proteases in the context of disease. They were demonstrated to be also extremely valuable tools for cellular and in vitro experiment that would help to unravel the biology of the target proteases in complex systems, including normal physiological role. For instance, we have developed the first metallocarboxypeptidase probes, as well as neutrophil elastase probes that proved to be excellent tools for imaging those proteases. Moreover, we have developed the first cell permeable and highly selective caspase-1 probe that should enable a much better understanding of the inflammatory processes such as inflammasome activation. Furthermore, when these probes would be combined with the cathepsin S and B probes that can be used for general monitoring of inflammation, a much better picture of the inflammation can be established, which has a potential to be further developed into preclinical and, possibly, clinical tools for diagnosis and treatment of other inflammation-associated disease. In addition to the small molecule smart probes, we have advanced the repertoire of imaging agents by developing highly specific tissue-targeting proteins DARPins (designed ankyrin repeat proteins) against FAP and conjugated them to imaging moieties, thereby making another class of imaging agents for cancer. Finally, we were able to make a step further developing a strategy for the synthesis of a dual reporter polymer conjugate bearing two probes that allow the combination of optical imaging and magnetic resonance imaging (MRI) strategies opening the field to further applications.

Reinforcing European competitivness

In addition to the scientific achievements, the Livimode project has also a great impact on the competitiveness of Europe in the field of chemical biology, in particular smart protease probe development, and preclinical diagnostic imaging. The complexity of the topic research required a high degree of networking between different experts in the field. The LIVIMODE consortium therefore represents a highly elaborated network with a critical mass of complementary expertise, which has created a coherent and highly competitive European scientific network with unique expertise on functional and pre-clinical validation of smart probes in vitro and in vivo in tumour and OA models. Therefore, the LIVIMODE project represents an European added value because no single group could have performed such a study alone. Moreover, the integration of all the research efforts of the LIVIMODE project has been an important stimulator of increased competitiveness.

Dissemination and exploitation of the results

To ensure dissemination and exploitation of results emanating from the LIVIMODE consortium, the LIVIMODE web site (http://livimode.eu/) has been set up at the beginning of the project and maintained during the course of the project. The web page allowed constant and up-to-date communication within the consortium (internal communications) and with the scientific community (external communications). Therefore, the Internet portal was split into a private and a public domain. This web site was extremely helpful to provide a platform for dissemination of knowledge generated throughout the project, as well as to maintain an updated basis for the data generated.

Internal dissemination of project results

Through the LIVIMODE web site, all partners have a restricted access to the following relevant information:

1. Scientific information:

- project information: objectives, summary of WP, list of deliverables, list of milestones.
- publications
- reports
- minutes of meetings

2. Technical information:

- Standard Operating Procedures
- models and probes info

3. Relevant contacts:

- links to partners and partner Institutions

External dissemination of project results and activities

The continuous dissemination activities are listed in the annexed tables and concerned the following issues:

- Publications in peer-reviewed journal
- Presentations of results at national and international meetings (posters, lectures).
- A satellite conference organized

Public awareness

- Project information (summary) is posted at the LIVIMODE web site.
- Several Press releases have been performed by different partners when the project started.

Exploitation of project results

In order to ensure that the possibilities for Intellectual Property Protection are not compromised, a Consortium Agreement, encompassing all the procedures relative to Intellectual Property Rights (IPR) management, knowledge dissemination, and clarifying the points relative to pre-existing knowledge and to the management of new discoveries (exploitation, patenting and publications), has been established and signed at the beginning of the project. This document provided the foundation for a comprehensive knowledge management process, with the aim to transform knowledge into economic value. In this aspect LIVIMODE has taken the advantage of the expertise of the Legal Department at SAD that providde support for patent applications and IP resources such as Standard Confidential Agreements, Material Transfer Agreements and Technical Transfer Agreements.

A policy of wide dissemination of project findings and results has been pursued throughout the course of the project, but care has been taken in implementing activities of dissemination not to compromise the Intellectual Property Protection possibilities, finally resulting in 3 patent applications in preparation.

During the regular meetings in the second half of the project, we have assembled a list of potentially exploitable results which the partners expect to be achieved during the course of the LIVIMODE project. Individual participants were then identified as responsible for each of these expected results, to monitor its progress during the course of the project and to raise awareness in the consortium in case the result seemed to get out of the focus of the project. This list was updated regularly, during our meetings, where items were removed from the list either due to too many problems materialising (competition results such as in the case of caspase-3 and -7 probes and partially legumain; the latter also because of non-supporting animal studies, and the problems of availability of some proteases, such as ADAM-TS) or when new opportunities arose which lead to the inclusion of new results in the list (e.g. development of novel neutrophil elastase probes as proof-of-principle). This process also helped us to identify and focus our resources to the most promising exploitable results. In parallel, we have also maintained a similar list of not (yet) exploitable results, including new basic knowledge relevant to be presented to the scientific community but not directly useful for development of applications.

List of Websites:

Website addresshttp://livimode.eu

Relevant contact details:
Prof. Boris Turk
Jožef Stefan Institute (JSI)
Position: Head and Professor in the Department of Biochemistry and Molecular and Structural Biology
Jamova 39
SL-1000 Ljubljana
Slovenia
Tel. +386 1 477 3772 / (Handy: +386 51 348 925)
Fax: +386 1 477 3984
E-mail: boris.turk@ijs.si