Final Report Summary - VISION (Prolonged inhibition of semaphorine3a pathway via a bio-degradable implant towards a better therapy for visual sensory impairments)
Executive Summary:
Glaucoma is a term which covers different pathological conditions leading to neuropathy, degeneration of the vulnerable optic nerve axons and cell bodies (Retinal Ganglion Cells, RGC). The most common glaucoma is associated with high intraocular pressure (IOP). The treatment today is based on lowering the IOP but is inefficient in preventing RGC loss. Thus, there is an urgent need for novel treatment.
Semaphorin 3A (Sema3A) is a cell secreted protein that participates in the axonal guidance pathways. Partner TAU was antibodies. Efficacy assay showed 3H4 to have the upper hand, consequently making it the best lead coming out of the antibody part of the VISION project. In addition, injury to the optic nerve (leading to sema3A elevation in the eye) was shown to specifically change the pharmacokinetics of Fab 3H4 but not that of a non-specific isotype-control antibody.
Small molecules as inhibitors of Sema3A to serve as glaucoma therapeutic agent:
The focus was in two main directions: a) increase the repertoire of small molecules potentially active to inhibit the target protein, and b) explore the mechanism of action of such inhibition in order to design more potent and safe inhibitors. The first goal has been achieved by constructing chemical libraries of SICHI (a peptoid named SICHI, discovered few years ago by partner CSIC) derivatives from which it has been identified an analogue, named CSIC-002, that has shown activity properties in vitro and in vivo to be considered a lead compound.
The active implant manufactured by electrospinning
A single electrospinning process for manufacturing thin poly(lactic-co-glycolic acid) (PLGA) cylindrical implants of 0.5 mm diameter loaded with small MW inhibitors was developed. Proof of concept for core and shell structure was demonstrated
The active implant manufactured by extrusion
An implant model for each Sema3a inhibitor candidate has been designed. The drug release of the implant prototypes has been studied in a laboratory model that simulates the vitreous cavity with a constant flow simulating the aqueous flow. The integrity of the Sema3a inhibitor after the manufacture of the implant has been proof. The developed implants provided a controlled release of the Sema3a inhibitor and are made by biodegradable polymers accepted by the regulatory authorities EMA and FDA for its use in medical devices for human being.
To complement the achievements of the implant development using the electrospinning technic, and to overcome the barrier limiting the amount of substance capable of being loaded in the implant manufactured by electrospinning, an alternative manufacturing process was proposed using extrusion.
In vivo studies
Acute glaucoma in rabbits models were used, Intra ocular implants loaded with small molecule CSIC002 and and Fab 3H4 were tested counting live Retinal Ganglion Cells (RGC).
zIN All tested models of assault to RGC, the intraocular injection of small molecule CSIC002 and Fab 3H4 presented a clear salvation activity of the RGC following the assault to the optic nerve and retina. Both therapeutic agents have potential to be a preventing degeneration drug in degenerative processes of CNS. No toxic or inflammatory reaction was observed during the whole period of the experiments
Regulatory Aspects
Upon the identification of the regulatory framework that governs the drug delivery system consisting of a polymeric implant + a Sema 3A inhibitor, the consortium has defined a non-clinical program that should support a 3-month phase IIb clinical trial in Europe. In agreement with Directive 2001/83/EC and Directive 2003/63/EC amending the former, non-clinical studies have to be performed under Good Laboratory Practice (GLP) as instructed by EMA and ICH guidelines. Hence, partner AX has identified which specific tests have to be conducted under GLP. 41 protocols outlines have been prepared.
Project Context and Objectives:
Glaucoma is a term which covers different pathological conditions leading to neuropathy, degeneration of the vulnerable optic nerve axons and cell bodies (Retinal Ganglion Cells = RGC) . The most common glaucoma is associated with high intraocular pressure (IOP). The treatment today is based on lowering the IOP but is inefficient in preventing RGC loss. Thus there is an urgent need for treatment that protects the RGCs . From other side the worldwide prevalence of glaucoma is increasing. This is in part due to the rapidly aging population. In 2010, 8.4M people worldwide were blind from primary open-angle glaucoma with incident growing to an estimated 11 M by 2020 .
Semaphorin 3A (Sema3A) is a cell secreted protein that participates in the axonal guidance pathways. Partner TAU was the first to show that Sema3A is also capable of inducing neuronal cell death . It was further shown that Sema3A is mediating the vast RGC apoptosis following optic nerve injury . Importantly, marked inhibition of RGC loss was achieved when eyes with axotomized optic nerves were co-treated by intravitreous injection of antibodies against Sema3A providing the proof of concept for the therapeutic approach for neuroprotection via inhibiting the Sema3A pathway. This concept was further validated by partner CSIC who developed a small molecular weight peptoid inhibitor (SICHI) of Sema3A and showed that this inhibitor promotes neural regeneration of damaged axons .
Many of the pathological mechanisms in glaucoma are apparent in acute optic nerve neuropathy which is characterized by neuronal death following stroke (an age related disease).
Thus, this project goal is to develop a therapy for glaucoma and acute optic nerve neuropathy using the same therapeutic approach, i.e. inhibiting further death of vision related neural cells by prolonged inhibition of Sema3A apoptotic pathway. The therapy will be based on a minimally invasive implant for controlled release of novel therapeutic moieties. Thus, the project will develop two Sema3A inhibitors: a low MW compound and a Sema3A targeted antibody. These inhibitors will be loaded into a novel controlled and prolonged release minimal invasive, injectable implant. Sema3A is a cell secreted protein thus its putative interaction with inhibitors will occur outside the cells. This is a substantial advantage of the proposed therapeutic strategy, due to the fact that it eliminates the need for structural optimisation and delivery vehicles required to facilitate cell penetration.
The following are the main project challenges (see also figure 1):
• Develop a novel and highly potent low molecular weight inhibitor of Sema3A signalling via construction and screening of a dynamic combinatorial library based on SICHI peptoid.
• Develop function blocking Sema3A human antibodies using a phage library.
• Develop a minimal invasive, injectable implant based on PLGA (poly-lactic-co-glycolic acid) and PEO (polyethylene oxide) polymers as “core and shell” nanofibres via electro-spinning which will be designed for controlled and prolonged release, of the Sema3A inhibitors (low MW compound or Sema3A targeted antibody).
• Produce animal models with optic nerve injury and optic nerve ischemia to mimic glaucoma and ischemic optic neuropathy. Rabbits with inherited glaucoma , (which appears only in 2% - 4% of the normal rabbit population) will be also produced for the final tests.
• Study in vivo efficacy, drug release kinetics, biocompatibility, biodegradability and safety of the implant using the developed models.
The scientific program is designed to generate a data pack to support an IMPD submission for clinical trial of the project product
Scientific and technological objectives
Objective 1 Implant requirements and specifications (R&S)
1.1 Requirements and specification of the polymeric implant (dimension, physicochemical properties, resorption rate, etc.).
1.2 Define the Sema3A inhibitors (target, dose, frequency of application, size, application process etc.) and the formulation for optimal implant release.
1.3 R&S for therapeutic affinity, solubility, permeability, pharmacokinetics, and efficacy based of the drainage dynamics modelling of the target anatomy and the proposed injection position.
1.4 Regulatory requirements of the polymeric implant as per EMA.
1.5
Objective 2 Develop Sema3A human antibody as therapeutic agent
2.1 Affinity selection of Sema3A antibody leads from the “Ronit 1” human antibody phage display library by the following baits:
• Full length Sema3A protein.
• Peptides sequences derived from Sema3A region that binds the receptor
2.2 Produce small-size single-chain antibodies in a soluble manner using E.coli and evaluate in vitro
2.3 Produce full size human IgG1 antibodies using mammalian cultures
2.4 Purification of the full size human IgG1 antibodies from stable cells
2.5 In vitro evaluation of the Sema3A inhibitors (chemo-repulsive and receptor binding assays)
2.6 Fluorescent labelling of the antibody for further in vivo studies
Objective 3 Develop a low molecular weight inhibitor of Sema3A by further modulating the structure of SICHI peptoid
3.1 Cloning and purification of the Sema3A
3.2 Define the interaction between SICHI and Sema3A by NMR
3.3 Generate small molecules derivatives for Sema3A inhibition by synthesis of a peptoid trimers library and preparation of dynamic combinatorial libraries (DCLs).
3.4 Screening the peptoid trimers and the DCL for identification of Sema3A inhibitors
3.5 In vitro evaluation of the Sema3A inhibitors (chemo-repulsive and receptor binding assays)
3.6 Scale-up synthesis of the Sema3A inhibitors and a back-up candidate under GMP-like conditions
Objective 4 Develop an injectable polymeric implant (loaded with the Sema3A modulator).
4. 1. Develop the formulation for therapeutic agent encapsulation in the injectable implant
4. 2. Develop and produce the PLGA based and controlled release implant, via electro-spinning
4. 3. Develop an intravitreal injecting device designed for inserting the implant
4. 4. In silico modelling of the Pharmacokinetic/Pharmacodynamic (PK/PD) of the Sema3A inhibitors for optic nerve injury and optic nerve ischemic model.
4. 5. Implant biodegradation (release kinetics) optimization via in vitro studies
4. 6. In vitro efficacy studies of the neuroprotective activated (loaded with theSema3A inhibitors) implant effect
Objective 5 In vivo studies of the active implant as per EMA guileless for clinical studies.
5.1 Statistical design of in vivo (rabbit models) study of the drug delivery system in glaucoma and optic nerve neuropathy
5.2 Define in vitro, in vivo toxicity, biocompatibility and biodegradability of the active implant
5.3 Define PK/PD of the activated implant
5.4 In vivo efficacy evaluation of the activated implant for chronic and acute insult:
• Acute insult: in two models to mimic glaucoma and optic nerve neuropathy (rats and rabbits)
• Chronic insult: in rabbits with inherited glaucoma
Objective 6 Draft IMPD using the data generated
6.1 Develop GLP protocols for preclinical studies according to Commission Directive 2003/63/EC
6.2 Summarize the preclinical study in vitro and in vivo. PK/PD, ADME, toxicity
6.3 Develop protocols for GMP synthesis of the Sema3A inhibitors and of the implant
Project Results:
Implant Requirements and Specifications (R&S) (WP1)
Two implant manufacturing lines were developed for obtaining suitable implants; electrospinning and extrusion methods. In both cases, the manufacturing processes carried for the small molecule and for the antibody are different, specialized for the characteristics and requirements of both substances. In both cases they were well tolerated by the animals.
Due to the degradation of the polymers, two different hypotheses have been studied. On one hand, if the total of the active ingredient is released prior to initiation of degradation of the product, a simple controlled release is achieved. On the other hand, if the polymer degrades while realizing the active ingredient, the changes in the exposed surface would affect the drug release kinetics. Based on this second hypothesis, we studied the possibility of coating the implants to provide an extra layer that would act as a protection layer and would reduce the exposed surface to the ends of the cylinder. Coating the implants presented some benefits but it was discarded from the current candidate when optimal drug release kinetics was achieved.
Lead candidate activated implants developed within the project:
• ES#202PR0202.
Manufacturing method: electrospinning.
Ingredients: PLGA and sema3a inhibitor small molecule.
1 cm implant releases 1-3 µg/day for the first 4 days and slower release from then. On day 14, ca. 40% of the total loading is already released.
• SYN-102-30.
Manufacturing method: extrusion.
Ingredients: PCL and sema3a inhibitor small molecule.
1 cm implant releases >11µg/day for the first 9 days and slower release for the next 6 days. Maximum dose of 52µg/day on day 3.
• SYN-103-33.
Manufacturing method: extrusion.
Ingredients: PCL and sema3a inhibitor antibody.
Sema3A human antibody as therapeutic agent
- Isolation of peptides binding antibodies
At the beginning of the study TAU ordered synthetic peptides corresponding to different parts of sema3a that are suspected to have functional significance. 4 such peptides were designed and ordered. Following 4 cycles of affinity selection of the “Ronit 1” human antibody phage display library, only a single high affinity clone was obtained. Other leads failed affinity or specificity validations. This clone binds a peptide that corresponds to the putative dimerization interface of sema3a. When it was tested for binding to full-size sema3a, it showed a very low binding signal, both as a phage displayed antibody and later as a soluble Fab. This is probably due to the fact the, indeed (as reported and as we found ourselves. Sema3a is a homodimer in solution, hence, the binding sequence to which this antibody binds is hidden in the mature protein.
Isolation of sema3a protein binding antibodies
For isolating phage antibodies that bind sema3a protein, we used a partially-purified sema3a preparation that was prepared at TAU from condition medium of HEK293 cells that were transfected to secrete human sema3a to the medium. While we could not purify sufficient quantities of full-size sema3a protein from the material that did bind to the column, the 65 kDa fragment, corresponding to Furin-clipped seam3a (which corresponds to the extracellular domain we wish to target) was found in the unbound fraction, after concentration was sufficient if quantity and purify for antibody isolation.
Selection of phage antibodies was carried out essentially as described in Azriel-Rosenfeld et al, 2004. However, since the antigen was only partially purified, the protocol modified as follows: A single well of a 24-well tissue culture plate was coated with 5 µg/ml mouse monoclonal anti sema3A in PBS. Library phages were mixed with sema3a (of the Q-sepharose unbound fraction) and the phage-sema3a complexes were captured on the immobilized monoclonal antibody. This way, phages that bind contaminating proteins in the sema3a fraction are not captures and washed away. 3-4 affinity selection cycles were carried out where in the odd numbered cycles such a “sandwich capture was applied while in the even numbered cycles phages were captured by sema3a coated directly onto wells. Phage ELISA was used to identify sema3a binders. Such affinity selection of the library was carried out 3 times independently by different investigators. The sequence of the antibody genes was determined to identify intact antibody clones that are unique. Combined, our efforts yielded 16 unique phage antibodies that were further analysed as soluble antibodies.
Evaluation of sema3a binding by soluble Fabs and IgGs
The purified Fabs were tested by ELISA for binding to sema3a and to several un-related proteins as specificity controls. Apparent binding affinity as estimated from the half maximal binding signal of the ELISA was between 10-7-10-8 M. A graph showing binding evaluation of 4 IgGs (including the “winning” lean 3H4 is shown in Fig. 1. The binding of the “winning” lead antibody, 3H4 to additional proteins from the sema3 family and to neuropilin was evaluated by us at TAU as well as by out collaborators at CSIC. 3H4 binds to sema3a, sema3e and neuropilin but much less so (negligible) to sema3a .
In vitro evaluation of Sema3A inhibiting antibodies
The ability of specific anti-Sema3A antibodies to inhibit Sema3a was tested in two cell culture assays (DRGs and a scratch assay using U87 cells. These experiments were carried out by the TAU. Even though the cell migration inhibition of Sema3A was rather low, as shown by a DRG assay (both antibodies 3H4 and 3E12 could significantly reduce Sema3A activity both as Fabs and as IgGs.
In both cases the Fabs displayed better results compared to the corresponding IgGs. The (DRG) repulsion assay we found that only 3H4 IgG could provide protection against Sema3A. No inhibitory effects were found in 3E12 antibody. In addition to 3H4 and 3E12, which were found to be function blocking antibodies (mainly the Fab form) we also tested two additional antibodies L4B8 and L4E6. Whereas B8 displayed inhibitory effect, E6 did not show any Sema3A inhibitory effect. Thus we did not continue to work with the E6 antibody. Using the scratch assay, we analyzed the inhibitory effects of 3H4 and L4B8 Fabs as a function of their dose response. We found that that both Fab 3H4 is a more efficient inhibitor than Fab L4B8.
In vivo evaluation of Sema3A inhibiting antibodies
Using a rat axotomy model (experiments carried out within the project also by TAU) we found increased levels of Sema3a in the retina following optic nerve axotomy. We found that several antibodies, but in particular antibody 3H4 could partially protect retinal ganglion cells (RGCs) from death following axotomy. To examine whether axotomy will change the pharmacokinetics of 3H4 Fab, we labelled antibodies with fluorescent dyes, we intraocularly injected fluorescent antibody and at different time point measured the level of the antibody. As shown in Fig. 4, We found that IgGs are cleared more slowly from the rat eye that Fabs and also found that that clearance kinetics was markedly slower in the axotomized eyes, suggesting the formation of immunocomplexes in the eye. As a control experiment we asked whether the reduced rate of antibody clearance following axotomy was specific to anti- Sema3A antibodies. For that purpose we repeated the experiment. Whereas axotomy significantly reduced the clearance of 3H4 Fab, no such effect was detected for non-specific (isotype control) antibody demonstrating the higher levels of Sema3A (as a result of axotomy) are responsible for reducing the clearance kinetics of 3H4 Fab.
Antibody-containing implants and additional in vivo studies
In order to determine whether the delayed clearance of Fab 3H4 was due to sema3A elevation rather than the injury itself, the experiment was repeated with an unrelated Fab against streptavidin (SA). Two groups of rats (3 rats in each group) were injected with either Alexa 594-conjugated Fab 3H4 or Alexa 594-conjugated Fab SA following optic nerve axotomy. Two additional groups served as control - they were injected with either Fab 3H4 or Fab SA without injury to the optic nerve. No difference in the clearance rate was detected between the axotomized and control groups injected with Fab SA. On the other hand, we again observed a significant difference between the clearance rates of the axotomized and control groups treated with Fab 3H4 both 24 and 48 hours following injury. These results indicate that the slow clearance rate in the axotomized group treated with Fab 3H4 did not arise from the axotomy of the optic nerve itself but rather from the formation of immuno-complexes between 3H4 Fab and sema3A.
A low molecular weight inhibitor of Sema3A by modulating the structure of SICHI peptoid
Clone, express and purify Sema3A.
In addition to cloning, expression and purification of Sema3A, we have developed a method for the efficient production of unlabelled and isotopically labelled (for NMR studies) constructs of the Sema3A C-terminal domain. Moreover, we have worked on the development of an efficient method of production and purification for the Sema3A C-terminal domain in sufficient amounts for biophysical studies. In the case of the Sema3A C-t, predicted to be unfolded, E.Coli was the method of choice for the expression of the protein. E.Coli allows the production of isotopically labelled proteins as it can grow in minimal medium (M9). Initial purification was performed in denaturing conditions (8M Urea or 6M guanidinium chloride) to avoid protein degradation. Given that the Sema3A C-terminal domain has a highly basic isoelectric point (pI > 11), the protein can be produced without any affinity tag (such as poly-His or GST), and purified by cationic exchange at high pH (9-10), which removes most of the proteins that fall in a more physiological range of pls. The amount of pure protein following this protocol afforded modest yields, (0,75-1,5 mg/L). Whereas these amounts might be sufficient for most of the experiments planned, problems arose when attempting to produce isotopically labelled protein (15N or 15N/13C) for NMR studies. Another strategy was the use of synthetic DNA templates with optimized codon usage, in our particular case, for E.Coli. This approach was useful and we were able to increase the levels of Sema3A C t domain expression from 0-1,5 mg/L to 2-12 mg/L (depending on the specific construct) for the codon optimised constructs. For the best expressing constructs (all construct boundaries and purification yields were listed in previous Reports), we were also able to obtain 15N-labeled protein, at around 2 mg/L. Although these amounts could be enough to perform some NMR experiments, we tried to further increase the levels of expression for 15N-labelled constructs by using the expression system developed by the BioSilta company, which essentially is a method that optimizes the growing and protein expression conditions for E.Coli (detailed protocols at http://biosilta.com/recombinant-proteins/e-coli/protein-for-nmr-analysis/). In this way, we were able to get yields for 15N-labelled Sema3A C-t constructs between 5-15 mg/L, even higher than the yields for the corresponding unlabelled constructs prepared in LB medium. Another strategy that we tested for partial Sema3A C-t labelling consisted in the methylation (using 13C – formaldehyde) of purified unlabelled protein. The strategy worked out well and methyl groups were introduced in all (6) lysine side chains of the domain, as confirmed by MALDI-TOF.
The interaction between SICHI and Sema3A by NMR
The molecular details of the interactions between the Sema3A C-t and GAGs have not been yet described. A deeper knowledge of this association can add valuable information regarding the proposed mechanism for the action of SICHI (our initial small molecule hit, and derived analogues in the inhibition of the Sema3A signaling. The goal of this part was to study the Sema3A C-terminal - GAGs interactions by the use of a combination of biophysical techniques, namely, Nuclear Magnetic Resonance (NMR), Surface Plasmon Resonance (SPR), and Fluorescence spectroscopy. In addition, the hypothetical perturbations that SICHI and analogues exert on the Sema3A C-t - GAGs association were investigated by binding competition experiments (using NMR or fluorescence).
Our data support the hypothesis that the interaction between Sema3A and GAGs observed at cellular level and described in the literature takes place through the involvement of the C-terminal polybasic region. Due to the importance of Sema3A-GAG interaction in the Sema3A signalling pathway, this knowledge is valuable for future therapeutic strategies for Sema3A modulation. The prime candidate for SICHI action was the protein-protein interaction between secreted Sema3A and Nrp1 receptor, but our results showed that SICHI does not significantly bind to the Sema3A sema domain nor to the Nrp1 extracellular domain. Therefore, we sought for an alternative mechanism for SICHI inhibition of the Sema3A pathway. Accordingly, we studied the interaction of SICHI and positively charged model peptides (FS2 and FS3, for structures cf. previous Reports) with GAGs by means of biophysical techniques and molecular dynamics simulations. First, the biophysical studies show that the two peptides FS2 and FS3 bind to GAGs and therefore, can be used as models to study Sema3A C-terminal domain-GAGs interaction. Next, we confirmed that SICHI binds to GAGs. Finally, we demonstrated that SICHI efficiently interferes in the Sema3A-GAGs interaction. Both NMR and fluorescence spectroscopy experiments showed a dose-dependent SICHI competition effect. Also, the acquisition of NMR spectra in the presence of salt (150 mM NaCl) showed that SICHI could compete with cationic peptides for the binding to GAGs in physiological salt conditions. By comparing the binding constants obtained by isothermal titration calorimetry (ITC) for the interaction between peptides or SICHI with heparin, a GAG model, we observed that the affinity of FS2 and FS3 peptides to heparin was 5 to 10-fold stronger than the affinity observed for SICHI, and these results are in agreement with the partial displacement of peptides from heparin exerted by SICHI and observed by NMR and fluorescence. These results support our hypothesis that the in vitro biological activity observed for SICHI could be related to SICHI exerting its function by binding to the GAG side in the C-terminal Sema3A-GAG interaction.
The medicinal chemistry strategy mostly used until now to inhibit protein-GAGs interactions has consisted of generating a target protein-specific synthetic glycan. In the present case, the suspected protein target for GAG interaction was the 55-amino acid long Sema3A C-terminal basic region (with 6 Lys and 10 Arg). Our results provide confirmation that a small peptoid molecule, such as SICHI or, more interestingly, the CSIC-002 analogue, could inhibit the interaction between polycationic peptides and GAGs, and provide a set of information that could be exploited for the development of protein-GAGs interaction inhibitors of therapeutic interest. At this point stands the studies carried out along the second half of the Project for optimising the activity of SICHI by designing new analogues. Among them, CSIC-002 has resulted to be the most active compound in vivo. However, in front of the uncertainties about the possible intellectual protection of this compound, we have worked on the design and preparation of a second library of SICHI analogues in addition to the testing of other polycationic amino derivatives (i.e. spermine). At the present time, a compound more active than CSIC-002 has not been yet identified. Taken overall, our findings have opened a new way for developing potent inhibitors of Semaphorin 3A.
Generation of small molecules for Sema3A inhibition by synthesis of a library of peptoid trimers
and of dynamic combinatorial libraries (DCLs)
This objective was mainly carried out along the first 18 months of the Project and it is fully described in the 1-18M Report. Briefly, a library of 14 SICHI derivatives was designed, synthesised, purified and structurally characterised. From the in vitro evaluation of these library components it was identified CSIC-002 as the most promising hit inhibitor of Sema-3A. In fact, this compound has been converted in the lead compound of the Project regarding small organic molecules. Consequently, in this second part of the Project, our group has been working in improving the preparation protocol for this compound (cf. Section 3.6) and has provided to Consortium members (in particular TAU, Synovo and Nicast) of CSIC-002 batches as demanded.
The following conclusions were reached:
▪ The first procedure of the preparation of SICHI was optimised. However, the methodology developed will be hardly applicable to the synthesis of the large amounts of this compound required eventually for preclinical trials.
▪ The synthesis of an analogue of SICHI, CSIC-02, exhibiting promising in vitro and in vivo activities as Sema-3A inhibitor, has been also studied. We developed a methodology that permits the scale-up of this compound and then fulfil the demands for further preclinical trials.
▪ AS a conclusion of the above methodologies, the preparation of a collection of SICHI analogues bearing different peptoid and heterocyclic structures was perpared. These compounds were delivered to for biological testing.
▪ From the computational studies, it is evident that SICHI and CSIC-02 could bind to the GAGs models selected by establishing several polar interactions, i.e. hydrogen bonds and salt bridges, mainly between the cationic moieties of SICHI/CSIC-002 and the anionic sulfate/carboxylate groups of the GAGs.
In addition, a virtual screening of a library of commercial lead-like compounds (650.000 from the MOE lead-like library, Chemical Computing Group Inc., Montreal, Canada) against the Neuropilin-1 binding domain of Sema 3A rendered interesting possibilities. Specifically, hits consisting of two aromatic rings bound through a cationic linker chain were identified. With this information, the possibility of constructing a dynamic combinatorial library (DCL) with a diamine and different aldehydes was highlighted. The imine formation/exchange is a suitable covalent bond to perform DCLs, because the reaction is reversible and the equilibrium can be shifted to adapt to the presence of the target. Moreover, further reduction of the imine bond would freeze the equilibrium, allowing the isolation of the amplified species. Accordingly, we designed a library of potential modulators of protein-protein interactions generated from two diamines and five aromatic aldehydes (see below), and used DCC to identify the best binder(s) to control the Semaphorin 3A-Neuropilin 1 (Sema3A-Nrp1) interaction.
A blind docking screening process was carried out with those compounds from the combinatorial dynamic library (CDL) that were amplified in the presence of Sema3A (1A, 1AB, 1AC, 1AD, 1AE, 1B, 1BC, 1CD, 1CE, 1D, 2AC, 2BC and 2C), to determine their potential binding site(s) on the protein. Thus, among 14 potential interaction sites identified on the surface of Sema3A, two of them were determined to be preferential sites for interaction, based on the higher docking scores obtained for most of the compounds bound at these sites. From these two sites, one of them involving residues 240-242, is located on a region of the Sema 3A surface that has been crystallographically identified as the interface that interacts with Neuropilin-1, suggesting that if the DCL compounds bind at this site they could interfere with the interaction between these two protein partners (Sema 3A – Nrp1). Two compounds that showed the best docking scores at this site, i.e. 1CE and 1AD, were also among the most amplified of the DCL. The second site identified on Sema 3A, involving residues 309, 310 and 490, is located in a funnel shaped cavity located in the middle of the protein. However, the biological relevance of the interaction of the DCL compounds at this site cannot be predicted at this moment.
From the 30 possible compounds, 7 potential binders were identified by the DCC. These compounds and two inactive ones were synthesized separately to evaluate their biological activities. We performed some STD-NMR studies of these compounds with Sema3A and Neuropilin-1, both together and separately. These compounds were also delivered to TAU to assess their biological activities. Summarizing, ligand-protein binding studies done by NMR showed that the most amplified compounds obtained by DCL bind Sema3A sema domain, but they exhibit poor selectivity (these compounds bind also Nrp1). Also, our results indicate that these compounds do not bind the Sema3A-Nrp1 binding interface (PPI region), which was our initial target. Consequently, at the present time this approach has not rendered results as satisfactory as those obtained from the initial library of peptoid derivatives.
Scale-up synthesis of lead candidate for Sema3A inhibition
The scale-up synthesis of CSIC-002, the small organic molecule that has emerged as lead candidate for Sema3A inhibition is described with full details in Deliverable 3.5. The scale-up synthesis of two lead candidates for Sema-3A inhibition (SICHI and CSIC-002) has been successfully achieved. We have developed a new convergent 2+1 strategy to overcome some problems found during the optimisation of the previously proposed procedures. Moreover, we have discarded the use of solid-phase synthesis steps because a procedure carried out in homogeneous media is more suitable for this scale-up. Consequently, we are capable of producing batches of hundreds of milligrams of either CSIC-001 (SICHI) or CSIC-002 (the current lead candidate) and it is not expected that a multigram production would add any additional problem to those encountered and solved in this study. Preparation of CSIC-002, the selected lead compound identified in this Project, involves 8 steps following a convergent strategy with different purifications steps during the process. The most important drawbacks in the scale-up of these syntheses are the distillation processes. But we think that when trying to scale up the synthesis at the industrial level, these problems will be overcome due to the equipment facilities of the Fine Chemicals companies.
Significant results
1. The production of selected protein fragments related to Sema3A and Neuropilin-1 has been optimised. In particular, different residues from the terminal amino basic tail of Sema3A have been produced in enough amounts to permit biological evaluation as well as the performance of biophysical Studies (NMR, SPR, ITC, fluorescence and UV-VIS spectroscopy). This production has involved also the construction of residues labelled with 13C and 15N addressed to specific NMR experiments.
2. The library of SICHI analogues designed and constructed led, after the corresponding activity assays, to the identification of compound CSIC-002 as lead candidate for the inhibition of Sema3A.
3. Concerning the inhibitory mechanism of action of SICHI and CSIC-002 on Sema3A, our biophysical and computational studies studies support the hypothesis that the interaction between Sema3A and glycoaminoglycans (GAGs) observed at cellular level and described in the literature takes place through the involvement of the Sema3A C-terminal polybasic region. In fact, we have confirmed that SICHI binds to GAGs, and that SICHI efficiently interferes in the Sema3A-GAGs interaction. The suspected protein target for GAG interaction is the 55-amino acid long Sema3A C-terminal basic region (with 6 Lys and 10 Arg). Our results provide confirmation that a small peptoid molecule could inhibit the interaction between polycationic peptides and GAGs and provide a set of information that could be exploited for the development of novel protein-GAGs interaction inhibitors of therapeutic interest.
4. The scale-up synthesis of our lead candidate, CSIC-002, has been optimised for the preparation of tens of milligram batches of this compound.
5. The first dynamic combinatorial library of potential Sema3A inhibitors has been designed, constructed and characterised. From the structural features learned, an improved version of this library is in due course.
An active polymeric implant (loaded with Sema3A inhibitor
Development of polymeric formulations by electrospinning loaded with small molecular weight sema3a inhibitors
Several small molecular weight sema3a inhibitors were synthesized and tested for stability and potency by CSIC and TAU, leading to the selection of the CSIC-002 inhibitor as the leading candidate for further studies. Throughout the process, NIC has been developing polymeric formulations containing the inhibitors to enable successful electrospinning forming a nano-fibrous membrane. Several hundred formulations were evaluated, each of these tested in the electrospinning setup by tuning various process parameters, followed by wide characterization of the manufactured material.
Formulation parameters that were evaluated include polymer type, composition and concentration (PCL, PLGA 50:50, PLGA 75:25), solvents type and composition (N,N-dimethylformamide (DMF), dichloromethane (DCM) and mixtures of thereof) and inhibitor concentration. Some of the electrospinning process parameters that were evaluated include the electrical potential, solution flow rate, distance, time, mandrel type, number and type of needles, drying and post-treatment conditions.
Fine tuning of the electrospun prototypes was conducted using low and high loaded CSIC-002 samples: 13 and 14% PLGA 50:50 loaded with 3-4% w/w CSIC-002, 22% PLGA 75:25 loaded with 3% w/w CSIC-002, 14% PLGA 50:50 loaded with 7 and 10% w/w CSIC-002 and 22%PLGA 75:25 loaded with 6 and 8% w/w CSIC-002.
Viscosity, density and conductivity were measured for each formulation prototype.
Development of polymeric formulations loaded with sema3a antibody inhibitors
Parallel to the development of a formulation containing small molecular weight sema3a inhibitors, sema3a antibody inhibitors were also considered. However, due to the bio-chemical nature of the antibodies, they could not be used in organic solvents formulation. Therefore, a core-shell structure, consisting of an organic soluble polymeric shell, and a water soluble polymeric core containing the antibody, were designed. Such a structure could be achieved using the co-axial electrospinning method. The development of the co-axial electrospinning setup was highly challenging due to delicate design and tuning required of the various solution and process parameters, which is essential to achieve a stable and reliable electrospinning process. The co-axial setup was established for PCL based shell and water soluble polymeric based core (PEG35K, POLYOx300, PVA). PCL as shell and PEG35K as core have been selected as most relevant candidates following preliminary experiments IgG antibody (Avastin) and human gamma globulin (HGG) were selected as antibody model to be incorporated in the core. Erbitux and 3H4-Fab manufactured by Prof. Itai Benhar at TAU were also incorporated to obtain a prototype loaded with antibody sema3a inhibitors. The development of the co-axial electrospinning process required measurements of surface tension and conductivity, in order to better understand the mutual influence of the core and shell jets in the formation of the Taylor’s cone in presence of the antibody inhibitors. The core-shell nano-fibrous structure was evaluated using fluorescent confocal microscopy, by incorporating fluorescent dyes in the core and shell solutions.
A sugar (trehalose) based method was developed in parallel to obtain a simpler electrospinning process, enabling to incorporate sema3a antibody inhibitors in the formulation. This technique requires a single electrospinning process, whereas the sugar acts as a stabilizer to protect the antibody against organic solvent.
Development of implant by electrospinning loaded with small molecular weight sema3a and antibody inhibitors
Following the development of the various formulations containing the CSIC-002 inhibitor, an implant that could be easily inserted into the eye was designed. Different membranes consisting of submicronic fibers of PLGA loaded with Sema3A inhibitor (CSIC-002) were electrospun (with a 15 mm mandrel) from polymeric solutions. Scanning electron microscopy (SEM) imaging was carried out to study the morphology of CSIC-002 loaded PLGA electrospun fibers, whereas recovery of the inhibitor was determined to be 100%. Samples of 13 and 14% PLGA (50:50) – prototype I and II – as well as 22% PLGA (75:25) – prototype III – loaded with different %wt of CSIC-002 resulted in submicronic fibers with width ranging from 0.2 to 1.4 µm for PLGA (50:50) and 0.4-4.2 µm for PLGA (75:25). Thicker fibers were obtained from PLGA (75:25) and present a much more curved structure, compared to those based on PLGA (50:50)
The release kinetics of the CSIC-002 from the electrospun fibers were analysed by SYN using a model of the vitreous humor with constant aqueous flow washing. The obtained release profiles were used to study the release mechanism and prioritize the formulations most suitable for in vivo testing. It was observed that prototype I samples, 3% CSIC-002 @ 14% PLGA (50:50) demonstrated an ideal and linear drug sustained system with a constant delivery rate; 55% of CSIC-002 was released after 60 days. In the relatively high load samples, i.e. 7% CSIC-002 @ 14% PLGA (50:50) and 8% CSIC-002 @ 22% PLGA (75:25), all the CSIC-002 was released within 7 days. The experimental results of CSIC-002 release as well as its proposed kinetic model demonstrated that an erosion mechanism of the fibers controls the drug release. The ideal prototype of 3% CSIC-002 loaded PLGA (50:50) has been chosen to be the leading candidate for future in vivo study due to the fact that degradation of the nano-fibers leads to a constant and linear release rate over a long period up to 60 days with a zero-order-kinetic.
The human gamma globulin (HGG) release, however, was very slow resulting in less than 1% release after 30 days. (Increasing the concentration of the HGG could not be achieved, as it led to an unstable electrospinning process. The slow release, and the requirement to dilute the antibody during the solution preparation, could not achieve the therapeutic demands of the implant, and therefore, at this point, the antibody inhibitor process can not present feasibility.
Based on the chosen formulation of 3% CSIC-002 @ 14% PLGA (50:50), a cylindrical implant was designed in order to be tailored for the use of OZURDEX® applicator for intravitreal injections.However due to shifting from membrane to hollow cylindrical geometry, the formulation had to be slightly altered. The cylindrical implant formulation that was found to mimic the membrane release profile was of 2.7% CSIC-002 @ 13% PLGA (50:50).
The cylindrical implant was later re-evaluated and a new design of 0.24mm wide membrane stripes was found more suitable for insertion into the eye, using a standard 23G cannula.This design, again, required an update of the formulation. In addition, the drug release profile was modified to reflect a quick, up to 7 days release profile. This modification follows a literature search and evaluation of the negative systemic effect, advocated in the literature, of the sema3a inhibitor.
The final formulation which was achieved after optimization of the electrospinning process was 5.5% CSIC-002 @13% PLGA (50:50), exhibiting an initial quick release of 37% of the active compound within 7 days, followed by a slower rate additional release.
Implant by extrusion loaded with sema3a inhibitor small molecule (CSIC002) and sema3a antibody (3H4)
To complement the achievements of the implant development using the electrospinning technic, and to overcome the barrier limiting the amount of substance capable of being loaded in the implant manufactured by electrospinning, an alternative manufacturing process was proposed.
A laboratory scale extruder was designed and created by the partner Synovo. Several implants were manufactured using biodegradable polymers like PCL, PLGA, PLA, and PEO, and testing as well different MWs of the previously mention polymers or different PLA and PGA ratios in PLGA.
Prior to extrusion, the polymer is combined with the substance, for the polymer encapsulate the substance. The process is different for each substance. Once the mix is done, the extrusion can be performed. The extrusion device is made of stainless steel and consists on the loading cavity, the extrusion channel, and the pressure mechanism with a rotating handle. When turning the handle, the material in the loading cavity is forced into the extrusion channel leaving the minimum left material in the loading cavity. The system has a temperature regulator incorporated. The required temperature depends on the properties of the implant ingredients. As a final product, a long cylinder exists through the extrusion channel to the exit hole.
The small molecule and the antibody were successfully integrated into the implants up to 30-33 % w/w loading. Several implants were created and the drug release studied until two lead extrusion based implants were selected as leading products. The lead extrusion based implants are loaded with 30% sema3a inhibitor small molecule (CSIC002) and 33 % sema3a antibody (3H4) respectively. They have the shape of a long cylinder of 0.3 mm diameter. The length can be adjusted after manufacturing for lower or higher dose.
To study the drug release profile of the implants loaded with the small molecule, a laboratory model was created within the project, which mimics the intravitreal environment. It consists in a device with cavities including synthetic vitreous mimicking the eye, washed with a constant flow mimicking the aqueous flow through the eye. The samples of the synthetic aqueous flow exiting the vitreous cavity were analysed by HPLC/MS/MS, Quantification of the remaining substance in the implants after the release study was as well done to confirm previous results.
Both extrusion based implant candidates were tested in vitro and in vivo with positive outcome.
• SYN-102-30.
Manufacturing method: extrusion.
Ingredients: PCL and sema3a inhibitor small molecule.
1 cm implant releases >11µg/day for the first 9 days and slower release for the next 6 days. Maximum dose of 52µg/day on day 3.
• SYN-103-33.
Manufacturing method: extrusion.
Ingredients: PCL and sema3a inhibitor antibody.
In vivo studies of the active implant (Sema3A inhibitors), as per EMA guileless for clinical studies
Pharmacokinetics
At the beginning of the project, ex vivo studies were performed with SICHI and analogs provided with an idea of the distribution of the substances in the eye. CSIC-002 was injected in ex-vivo eyes from rabbits (taken post mortem from cadavers used for other purposes) for a concentration in the eyes of 1 or 10 µM (2 groups, 5 eyes per group). The eyes were incuvated 3h at room temperature under slight shaking and overnight at 4°C. Vitreous, lent, retina, pigmented epithelium and optic nerve were extracted and analyzed.
i.v. pharmacokinetics and toxicokinetics murine studies in mice
BL/6 male mice were subjected to single i.v. administration of CSIC001, CSIC002, CSIC004, CSIC008 or CSIC010 (1.8µmol/Kg of each. Formulation: serum). Samples from peripheral plasma at 5min, 15min, 30min, 1h, 2h and 3h after injection were anlysed. The animals were sacrificed 3 h after dose and different organs were sampled for analytics. Following i.v. administration, the half-life in plasma of the CSIC compounds was ca. 30 min and no compound was found in plasma 2 h after dose. Every CSIC compound was found in the eyes 3h after i.v. administration. This confirms the affinity of the Sema3a inhibitor for the eye, especially CSIC-004 and CSIC-002. No signs of intolerance or rejection was observed. No toxicity effects were assigned to the treatments.
Intravitreal pharmacokinetics and toxicokinetics in mice
16 mice, albino Balb/C and pigmented BL/6 mice were subjected to intravitreal administration of 0.15 µg/eye of CSIC001 or CSIC002 (10µL/eye). Two hours after injection the animals were sacrificed. The concentration of the CSIC compounds in the eye was analysed. Following intravitreal administration, no systemic exposure was presented. CSIC002 presented high concentration in the back of the eye, compared to other eye parts. No signs of intolerance or rejection were observed. No toxicity effects were assigned to the treatments.
Subcutaneous toxicokinetics in mice
Following previous tolerability studies with the sema3a inhibitor active substance, the activated implants, loadeded with sema3a inhibitor, were tested in several in vitro and in vivo studies, including toxicity and pharmaokinamic studies. No signs of rejection or inflammation were presented. The implants were well tolerated.
The lead PCL implants manufactured by extrusion were tested in s.c. murine model. The s.c. implants were well tolerated during a 39 days study in 18 C57BL/6 mice. Each implant injected was 0.7 cm long. The implants injected had 0.3 mm, 0.5 mm or 1 mm diameter. The mice were observed during 39 days and the implants were monitored by palpation regularly along the study. On day 39, some of the implants had been broken due to the palpation, however, those which were still s.c. has the original length and appearance. Pictures of the implants after sacrifice of the animals were recorded.
Intravitreal toxicokinetics in rabbits
Four New Zealand Rabbits were subjected to intravitreal injection of extrusion based PCL implant in the left eye. The animals were observed 3 times a week during the first 16 days and weekly after. Half of the animals were sacrificed after 2 weeks and the other half 4 weeks after injection. No signs of rejection and/or inflammation were observed in the animals. Both control eye and implant carry eye presented same appearance during the whole study. The cornea, lens, retina and vitreous presented healthy a status.
Complete cut of optic nerve axons within its covers created in laboratory rats
Adult male rats, 12-15 –weeks old, were deeply anaesthetized. The right optic nerve(ON) was exposed by separation of the bulbar conjunctiva and entering the orbital space retro bulbar. While separating the surrounding extra ocular muscles and surrounding tissue the optic nerve was exposed. Using a special designed glass dissector with a blunt tip of 50 microns diameter and opening was created in the surrounding meninges and with a slow movement the axons of the nerve were separated. At the end of the procedure a clear separation of two optic nerve stumps can be observed. The injury was unilateral in all animals done in the right eye. The surgery was immediately followed by an injection of Sema3A inhibitor or saline (as control) into the vitreous space, close to the retina. After 12 days the retrograde neurotracer 4-di-10-Asp was inserted into the optic nerve between and near the sight of injury and the eye globe. This neurotraces stains only live retinal ganglion cells (RGC).Four days following the staining , the animals were sacrificed. The retinas were separated from the eye and mounted on slide to be viewed under a fluorescence microscope. Random fields were selected and the number of live RGC were counted.(Ref 1)
Retinal detachment in rats
The same type and age of rats were used for creating retinal detachment in the right eye. Under deeply anesthetized rat (Xylasine 50mg/kg and Ketamine 35 mg/kg) and following dilatation of the pupil with Tropicamide drops 0,5% the retinal detachment was created. The procedure was done using a syringe with a 32G needle and containing a volume of 5 microliter of saline. The needle was in traduced at the cornel border in the anterior cahmber and introduce under the iris between the lens and the ora serata ( the dead end of the retina.). The needle was introduced under the retina and the saline injected. This procedure created approximately half of the retinal area detached.The rats were devide in two treated groups:Group A was immediately injected with 2 microliters of Fab 3H4 and Group B 3 days later.The third group was injected with saline and served as control. The same protocol of staining the RGC and flat mounting of the retina for RGC live counting as described in previous chapter was done also in this experiment.
Acute glaucoma in rabbits
Male, 12 weeks old, New Zeeland albino rabbits were used in creating this model. Under deeply anesthetized rabbits, (Xylasine and Ketamine), High intraocular pressure (IOP) was created by introducing in the anterior chamber a maintainer cannula (used in cataract surgery and connecting it to an intravenous infusion pack of 500 cc volume containing saline. The pack was fixed at 80 cm height and created an IOP of 45 mmHg to 50 mmHg. This IOP was maintained for 1,5 hour. Twenty four hours following the assault, intra ocular implants (PLGA – created by Synovo) loaded with small molecule CSIC02 or blank implants, serving as control, were introduced into the eye globe using vitrectomy technique. Following 14 days from the assault the same procedure of staining the RGC, flat mounting and counting live RGC was done.
The intra ocular injection of CSIC02 in the right eye of rats following the creation of complete cut of the optic nerve presented a significant number of live RGC following the assault. The retinas were stained at the 14th day following the injury. This is the climax time of the apoptotic death program created by Sema3A.
There can be observed that about 28% of RGC are live when compared to normal retina and the saline treated retinae present only 1,5% of live RGC.
The results of intra ocular injection of Fab 3H4, to inhibit the apoptotic death of the RGC following complete cut of optic nerve axons in the right eye of rats: Retrograde staining of the RGC reveals that over 40% of RGC are live in the treated eyes while compared with normal retinae. Only 2% of RGC remained live in the saline treated group.
Intraocular injection of Fab 3H4 saved RGC following the creation of retinal detachment in the right eye of rats. In contrary, the right eye of injured optic nerve of rats with retinal detachment and treated with unrelated Fab against streptavidin (SA) presented only few live RGC.
Conclusion
In all three models of assault to RGC, the intraocular injection of small molecule CSIC002 and Fab 3H4 presented a clear salvation activity of the RGC following the assault to the optic nerve and retina.
Both approaches have potential to be a preventing degeneration drug in degenerative processes of CNS.
No toxic or inflammatory reaction was observed during the whole period of the experiments.
Regulatory aspects - Draft IMPD using the data generated
To support and optimize the development of an injectable polymeric implant (loaded with the Sema3A modulator) the VISION project has included a follow up of the project’s developments and a design of the future regulatory aspects that will need to be covered to advance the final product to the clinical trials phase.
On the grounds of the European regulation, the relevant regulatory guidelines for VISION implant development into clinical phases have been identified. The drug product developed in the VISION project is a combination product consisting of a Sema 3A inhibitor and a polymeric implant marketed. The Sema 3A inhibitor falls in the category of medicinal products, whereas the polymeric implant is a medical device, thus two different types of regulations apply to the final product development. In addition, the regulatory aspects that will guide the preclinical development are partially dependent on the type of clinical design that will be used in the future. The consortium came to an agreement on which would be, a priory, the main principles of an initial trial based on the initial data and the clinical experience of TAU:
• The indication foreseen will be ischemic optic neuropathy
• The most appropriate clinical study would be a phase IIa (proof-of-concept) clinical trial in patients with ischemic optic neuropathy. A phase I clinical trial will be probably not feasible for ethical reasons since the implant should be introduced into the eye
• The duration foreseen for the first clinical trial would be at least 3 months with a time point at 6 months
• Single or multiple administrations, but not a chronic treatment, is envisaged
• The proposed therapy will be administrated together with current treatments mainly intended to lower IOP (intraocular presure)
• Two scenarios will be considered for the preclinical development proposal: (1) polymeric implant +NCE and (2) polymeric implant + MoAb
Taking into account international standards and guidelines, a toxicological program enabling clinical phase IIa with Implant + NCE or Implant + MoAb has been designed. A list of go/no-go decisions based on regulatory needs have guided the non-clinical development of the VISION implant within the project and projected beyond.
A market evaluation reviewing available implants for ocular administration and other polymeric dosage forms has been done in order to gather relevant information for the design, development and fabrication of the VISION implant. An epidemiological analysis of the selected indications (retinal vein occlusion, ischemic optic neuropathy, retinal arterial occlusion, traumatic neuropathy and acute demyelinating optic neuropathy) to select the best optimal target for further exploitation has also been included.
An evaluation of the potential cross-reactivity risk between the compounds developed in VISION (NCE / MoAb) and semaphorins other than 3A has been done to assess the need of secondary pharmacodynamics studies. The possibility of such undesired interactions taking place in the eye environment was estimated on the basis of available expression and structural data. The subsequent modulation of the expected therapeutic outcome was assessed taking into account functional data.
Commonly accepted inflammatory markers that could be used in the evaluation of the implant based on regulatory guidelines and standards have been identified to ensure that the experiments done in the course of the project can support as much as possible the requirements of the regulatory requirements that the product will face later on.
The regulatory pharmacodynamic studies are better accepted with at least some knowledge about the mechanism of action of the compound under study. During the course of the project the SICHI compound has been suggested to achieve neuroprotection through the interaction with glycosaminoglycans (GAGs) in the eye, which have been reported to have a regulatory action in the apoptotic pathway. With the aim of properly guiding the development of new compounds for glaucoma through any of the following strategies: maintaining Sema3A-Nrp1 as a target or target the interaction between Sema3A and GAGs; a market and literature survey on other compounds targeting GAGs has been included. The aspects evaluated included: a) the commercial use of GAG modulators, b) the prior art regarding GAG modulation in the treatment of neurodegenerative diseases, c) the suitability of these compounds for treating glaucoma and d) the use of these compounds as a positive control for future studies with SICHI.
Chemical development guidelines along with current protocols for production of CSIC-002 and analytical measurements have been compiled setting the bases for further developments under GMP (Good Manufacturing Practices) by an appropriate CMO (Contract Manufacturing Organization). A pre-selection of GMP-compliant facilities has been included
The relevant non-clinical pharmacology, pharmacokinetics and toxicology studies performed within the project have been compiled in the format of the Investigational Medicine Product Dossier (IMPD) for the small molecule (CSIC-002) and a monoclonal antibody (Fab 3H4). The regulatory studies needed for NCE-implant and the MoAb-implant are coincident in many points but there are some relevant points that need to be addressed differently, thus two different documents have been generated in order to be able to use them independently in case of independent further development. The relevant results from in vitro and in vivo studies conducted within VISION project have been used to address the efficacy and safety aspects of the document. While efficacy studies do not have to be conducted under GLP, thus the data collected may be sufficient to support an application, the safety studies need to be complemented with studies performed under GLP terms. The data collected within VISION has been included as supporting data for the safety aspects of the document. The further compulsory safety studies that need to be performed under good laboratory practices (GLP) prior to moving to a clinical phase have been identified. The protocol outlines of GLP-Compliant Studies have been set-up and optimized based on regulatory guidelines and the data obtained in the course of the project (41 protocol outlines) and linked to the corresponding sections of the IMPD.
An implementation calendar of GLP-Compliant Studies in a 15-month period that guarantees GLP compliance at minimal cost, based upon European Medicines Agency requirements and practical experience has been outlined. A pre-selection of GLP-compliant facilities able to provide services that meet the needs and expectations of the VISION project has also been included.
A basic introduction and guideline for a request of Scientific Advice to the EMA has been included along with some possible questions for the agency.
Potential Impact:
Impact
Enhancing Europe competitiveness
Furthermore the European Commission has identified active and healthy ageing as a major societal challenge common to all European countries, and an area which presents considerable potential for Europe to lead the world in providing innovative responses to this challenge.
The pilot European Innovation Partnership on Active and Healthy Ageing is aiming to:
1. Enabling EU citizens to lead healthy, active and independent lives while ageing;
2. Improving the sustainability and efficiency of social and health care systems;
3. Boosting and improving the competitiveness of the markets for innovative products and services, responding to the ageing challenge at both EU and global level, thus creating new opportunities for businesses.
The overarching target of this pilot partnership is to increase the average healthy lifespan by two years by 2020.
The VISION project supports the above pilot thus enhancing the European competitiveness
Socio-Economic impact
After cancer, Americans fear blindness most . Increasing aging of populations throughout the world, paired with other factors, including an increase in the prevalence of age-related diseases and sedentary lifestyles, will ultimately lead to a far greater worldwide prevalence of many diverse ophthalmic diseases, in addition to those related to only aging and obesity (Figure 3.1)
Blindness and irreversible sight impairment cost an estimated $50 billion each year in the United States and, thus, improving on the treatments available today could substantially reduce the financial burden on society.
The project will lead to a real difference to the lives of elderly Europeans by providing treatment for neurodegenerative diseases (e.g. glaucoma and acute optic nerve neuropathy) to stop further death of neural cells, never achieved before.
List of Websites:
http://fp7-vision.eu/
Coordinator Professor Arieh Solomon
asolomon@post.tau.ac.il
Glaucoma is a term which covers different pathological conditions leading to neuropathy, degeneration of the vulnerable optic nerve axons and cell bodies (Retinal Ganglion Cells, RGC). The most common glaucoma is associated with high intraocular pressure (IOP). The treatment today is based on lowering the IOP but is inefficient in preventing RGC loss. Thus, there is an urgent need for novel treatment.
Semaphorin 3A (Sema3A) is a cell secreted protein that participates in the axonal guidance pathways. Partner TAU was antibodies. Efficacy assay showed 3H4 to have the upper hand, consequently making it the best lead coming out of the antibody part of the VISION project. In addition, injury to the optic nerve (leading to sema3A elevation in the eye) was shown to specifically change the pharmacokinetics of Fab 3H4 but not that of a non-specific isotype-control antibody.
Small molecules as inhibitors of Sema3A to serve as glaucoma therapeutic agent:
The focus was in two main directions: a) increase the repertoire of small molecules potentially active to inhibit the target protein, and b) explore the mechanism of action of such inhibition in order to design more potent and safe inhibitors. The first goal has been achieved by constructing chemical libraries of SICHI (a peptoid named SICHI, discovered few years ago by partner CSIC) derivatives from which it has been identified an analogue, named CSIC-002, that has shown activity properties in vitro and in vivo to be considered a lead compound.
The active implant manufactured by electrospinning
A single electrospinning process for manufacturing thin poly(lactic-co-glycolic acid) (PLGA) cylindrical implants of 0.5 mm diameter loaded with small MW inhibitors was developed. Proof of concept for core and shell structure was demonstrated
The active implant manufactured by extrusion
An implant model for each Sema3a inhibitor candidate has been designed. The drug release of the implant prototypes has been studied in a laboratory model that simulates the vitreous cavity with a constant flow simulating the aqueous flow. The integrity of the Sema3a inhibitor after the manufacture of the implant has been proof. The developed implants provided a controlled release of the Sema3a inhibitor and are made by biodegradable polymers accepted by the regulatory authorities EMA and FDA for its use in medical devices for human being.
To complement the achievements of the implant development using the electrospinning technic, and to overcome the barrier limiting the amount of substance capable of being loaded in the implant manufactured by electrospinning, an alternative manufacturing process was proposed using extrusion.
In vivo studies
Acute glaucoma in rabbits models were used, Intra ocular implants loaded with small molecule CSIC002 and and Fab 3H4 were tested counting live Retinal Ganglion Cells (RGC).
zIN All tested models of assault to RGC, the intraocular injection of small molecule CSIC002 and Fab 3H4 presented a clear salvation activity of the RGC following the assault to the optic nerve and retina. Both therapeutic agents have potential to be a preventing degeneration drug in degenerative processes of CNS. No toxic or inflammatory reaction was observed during the whole period of the experiments
Regulatory Aspects
Upon the identification of the regulatory framework that governs the drug delivery system consisting of a polymeric implant + a Sema 3A inhibitor, the consortium has defined a non-clinical program that should support a 3-month phase IIb clinical trial in Europe. In agreement with Directive 2001/83/EC and Directive 2003/63/EC amending the former, non-clinical studies have to be performed under Good Laboratory Practice (GLP) as instructed by EMA and ICH guidelines. Hence, partner AX has identified which specific tests have to be conducted under GLP. 41 protocols outlines have been prepared.
Project Context and Objectives:
Glaucoma is a term which covers different pathological conditions leading to neuropathy, degeneration of the vulnerable optic nerve axons and cell bodies (Retinal Ganglion Cells = RGC) . The most common glaucoma is associated with high intraocular pressure (IOP). The treatment today is based on lowering the IOP but is inefficient in preventing RGC loss. Thus there is an urgent need for treatment that protects the RGCs . From other side the worldwide prevalence of glaucoma is increasing. This is in part due to the rapidly aging population. In 2010, 8.4M people worldwide were blind from primary open-angle glaucoma with incident growing to an estimated 11 M by 2020 .
Semaphorin 3A (Sema3A) is a cell secreted protein that participates in the axonal guidance pathways. Partner TAU was the first to show that Sema3A is also capable of inducing neuronal cell death . It was further shown that Sema3A is mediating the vast RGC apoptosis following optic nerve injury . Importantly, marked inhibition of RGC loss was achieved when eyes with axotomized optic nerves were co-treated by intravitreous injection of antibodies against Sema3A providing the proof of concept for the therapeutic approach for neuroprotection via inhibiting the Sema3A pathway. This concept was further validated by partner CSIC who developed a small molecular weight peptoid inhibitor (SICHI) of Sema3A and showed that this inhibitor promotes neural regeneration of damaged axons .
Many of the pathological mechanisms in glaucoma are apparent in acute optic nerve neuropathy which is characterized by neuronal death following stroke (an age related disease).
Thus, this project goal is to develop a therapy for glaucoma and acute optic nerve neuropathy using the same therapeutic approach, i.e. inhibiting further death of vision related neural cells by prolonged inhibition of Sema3A apoptotic pathway. The therapy will be based on a minimally invasive implant for controlled release of novel therapeutic moieties. Thus, the project will develop two Sema3A inhibitors: a low MW compound and a Sema3A targeted antibody. These inhibitors will be loaded into a novel controlled and prolonged release minimal invasive, injectable implant. Sema3A is a cell secreted protein thus its putative interaction with inhibitors will occur outside the cells. This is a substantial advantage of the proposed therapeutic strategy, due to the fact that it eliminates the need for structural optimisation and delivery vehicles required to facilitate cell penetration.
The following are the main project challenges (see also figure 1):
• Develop a novel and highly potent low molecular weight inhibitor of Sema3A signalling via construction and screening of a dynamic combinatorial library based on SICHI peptoid.
• Develop function blocking Sema3A human antibodies using a phage library.
• Develop a minimal invasive, injectable implant based on PLGA (poly-lactic-co-glycolic acid) and PEO (polyethylene oxide) polymers as “core and shell” nanofibres via electro-spinning which will be designed for controlled and prolonged release, of the Sema3A inhibitors (low MW compound or Sema3A targeted antibody).
• Produce animal models with optic nerve injury and optic nerve ischemia to mimic glaucoma and ischemic optic neuropathy. Rabbits with inherited glaucoma , (which appears only in 2% - 4% of the normal rabbit population) will be also produced for the final tests.
• Study in vivo efficacy, drug release kinetics, biocompatibility, biodegradability and safety of the implant using the developed models.
The scientific program is designed to generate a data pack to support an IMPD submission for clinical trial of the project product
Scientific and technological objectives
Objective 1 Implant requirements and specifications (R&S)
1.1 Requirements and specification of the polymeric implant (dimension, physicochemical properties, resorption rate, etc.).
1.2 Define the Sema3A inhibitors (target, dose, frequency of application, size, application process etc.) and the formulation for optimal implant release.
1.3 R&S for therapeutic affinity, solubility, permeability, pharmacokinetics, and efficacy based of the drainage dynamics modelling of the target anatomy and the proposed injection position.
1.4 Regulatory requirements of the polymeric implant as per EMA.
1.5
Objective 2 Develop Sema3A human antibody as therapeutic agent
2.1 Affinity selection of Sema3A antibody leads from the “Ronit 1” human antibody phage display library by the following baits:
• Full length Sema3A protein.
• Peptides sequences derived from Sema3A region that binds the receptor
2.2 Produce small-size single-chain antibodies in a soluble manner using E.coli and evaluate in vitro
2.3 Produce full size human IgG1 antibodies using mammalian cultures
2.4 Purification of the full size human IgG1 antibodies from stable cells
2.5 In vitro evaluation of the Sema3A inhibitors (chemo-repulsive and receptor binding assays)
2.6 Fluorescent labelling of the antibody for further in vivo studies
Objective 3 Develop a low molecular weight inhibitor of Sema3A by further modulating the structure of SICHI peptoid
3.1 Cloning and purification of the Sema3A
3.2 Define the interaction between SICHI and Sema3A by NMR
3.3 Generate small molecules derivatives for Sema3A inhibition by synthesis of a peptoid trimers library and preparation of dynamic combinatorial libraries (DCLs).
3.4 Screening the peptoid trimers and the DCL for identification of Sema3A inhibitors
3.5 In vitro evaluation of the Sema3A inhibitors (chemo-repulsive and receptor binding assays)
3.6 Scale-up synthesis of the Sema3A inhibitors and a back-up candidate under GMP-like conditions
Objective 4 Develop an injectable polymeric implant (loaded with the Sema3A modulator).
4. 1. Develop the formulation for therapeutic agent encapsulation in the injectable implant
4. 2. Develop and produce the PLGA based and controlled release implant, via electro-spinning
4. 3. Develop an intravitreal injecting device designed for inserting the implant
4. 4. In silico modelling of the Pharmacokinetic/Pharmacodynamic (PK/PD) of the Sema3A inhibitors for optic nerve injury and optic nerve ischemic model.
4. 5. Implant biodegradation (release kinetics) optimization via in vitro studies
4. 6. In vitro efficacy studies of the neuroprotective activated (loaded with theSema3A inhibitors) implant effect
Objective 5 In vivo studies of the active implant as per EMA guileless for clinical studies.
5.1 Statistical design of in vivo (rabbit models) study of the drug delivery system in glaucoma and optic nerve neuropathy
5.2 Define in vitro, in vivo toxicity, biocompatibility and biodegradability of the active implant
5.3 Define PK/PD of the activated implant
5.4 In vivo efficacy evaluation of the activated implant for chronic and acute insult:
• Acute insult: in two models to mimic glaucoma and optic nerve neuropathy (rats and rabbits)
• Chronic insult: in rabbits with inherited glaucoma
Objective 6 Draft IMPD using the data generated
6.1 Develop GLP protocols for preclinical studies according to Commission Directive 2003/63/EC
6.2 Summarize the preclinical study in vitro and in vivo. PK/PD, ADME, toxicity
6.3 Develop protocols for GMP synthesis of the Sema3A inhibitors and of the implant
Project Results:
Implant Requirements and Specifications (R&S) (WP1)
Two implant manufacturing lines were developed for obtaining suitable implants; electrospinning and extrusion methods. In both cases, the manufacturing processes carried for the small molecule and for the antibody are different, specialized for the characteristics and requirements of both substances. In both cases they were well tolerated by the animals.
Due to the degradation of the polymers, two different hypotheses have been studied. On one hand, if the total of the active ingredient is released prior to initiation of degradation of the product, a simple controlled release is achieved. On the other hand, if the polymer degrades while realizing the active ingredient, the changes in the exposed surface would affect the drug release kinetics. Based on this second hypothesis, we studied the possibility of coating the implants to provide an extra layer that would act as a protection layer and would reduce the exposed surface to the ends of the cylinder. Coating the implants presented some benefits but it was discarded from the current candidate when optimal drug release kinetics was achieved.
Lead candidate activated implants developed within the project:
• ES#202PR0202.
Manufacturing method: electrospinning.
Ingredients: PLGA and sema3a inhibitor small molecule.
1 cm implant releases 1-3 µg/day for the first 4 days and slower release from then. On day 14, ca. 40% of the total loading is already released.
• SYN-102-30.
Manufacturing method: extrusion.
Ingredients: PCL and sema3a inhibitor small molecule.
1 cm implant releases >11µg/day for the first 9 days and slower release for the next 6 days. Maximum dose of 52µg/day on day 3.
• SYN-103-33.
Manufacturing method: extrusion.
Ingredients: PCL and sema3a inhibitor antibody.
Sema3A human antibody as therapeutic agent
- Isolation of peptides binding antibodies
At the beginning of the study TAU ordered synthetic peptides corresponding to different parts of sema3a that are suspected to have functional significance. 4 such peptides were designed and ordered. Following 4 cycles of affinity selection of the “Ronit 1” human antibody phage display library, only a single high affinity clone was obtained. Other leads failed affinity or specificity validations. This clone binds a peptide that corresponds to the putative dimerization interface of sema3a. When it was tested for binding to full-size sema3a, it showed a very low binding signal, both as a phage displayed antibody and later as a soluble Fab. This is probably due to the fact the, indeed (as reported and as we found ourselves. Sema3a is a homodimer in solution, hence, the binding sequence to which this antibody binds is hidden in the mature protein.
Isolation of sema3a protein binding antibodies
For isolating phage antibodies that bind sema3a protein, we used a partially-purified sema3a preparation that was prepared at TAU from condition medium of HEK293 cells that were transfected to secrete human sema3a to the medium. While we could not purify sufficient quantities of full-size sema3a protein from the material that did bind to the column, the 65 kDa fragment, corresponding to Furin-clipped seam3a (which corresponds to the extracellular domain we wish to target) was found in the unbound fraction, after concentration was sufficient if quantity and purify for antibody isolation.
Selection of phage antibodies was carried out essentially as described in Azriel-Rosenfeld et al, 2004. However, since the antigen was only partially purified, the protocol modified as follows: A single well of a 24-well tissue culture plate was coated with 5 µg/ml mouse monoclonal anti sema3A in PBS. Library phages were mixed with sema3a (of the Q-sepharose unbound fraction) and the phage-sema3a complexes were captured on the immobilized monoclonal antibody. This way, phages that bind contaminating proteins in the sema3a fraction are not captures and washed away. 3-4 affinity selection cycles were carried out where in the odd numbered cycles such a “sandwich capture was applied while in the even numbered cycles phages were captured by sema3a coated directly onto wells. Phage ELISA was used to identify sema3a binders. Such affinity selection of the library was carried out 3 times independently by different investigators. The sequence of the antibody genes was determined to identify intact antibody clones that are unique. Combined, our efforts yielded 16 unique phage antibodies that were further analysed as soluble antibodies.
Evaluation of sema3a binding by soluble Fabs and IgGs
The purified Fabs were tested by ELISA for binding to sema3a and to several un-related proteins as specificity controls. Apparent binding affinity as estimated from the half maximal binding signal of the ELISA was between 10-7-10-8 M. A graph showing binding evaluation of 4 IgGs (including the “winning” lean 3H4 is shown in Fig. 1. The binding of the “winning” lead antibody, 3H4 to additional proteins from the sema3 family and to neuropilin was evaluated by us at TAU as well as by out collaborators at CSIC. 3H4 binds to sema3a, sema3e and neuropilin but much less so (negligible) to sema3a .
In vitro evaluation of Sema3A inhibiting antibodies
The ability of specific anti-Sema3A antibodies to inhibit Sema3a was tested in two cell culture assays (DRGs and a scratch assay using U87 cells. These experiments were carried out by the TAU. Even though the cell migration inhibition of Sema3A was rather low, as shown by a DRG assay (both antibodies 3H4 and 3E12 could significantly reduce Sema3A activity both as Fabs and as IgGs.
In both cases the Fabs displayed better results compared to the corresponding IgGs. The (DRG) repulsion assay we found that only 3H4 IgG could provide protection against Sema3A. No inhibitory effects were found in 3E12 antibody. In addition to 3H4 and 3E12, which were found to be function blocking antibodies (mainly the Fab form) we also tested two additional antibodies L4B8 and L4E6. Whereas B8 displayed inhibitory effect, E6 did not show any Sema3A inhibitory effect. Thus we did not continue to work with the E6 antibody. Using the scratch assay, we analyzed the inhibitory effects of 3H4 and L4B8 Fabs as a function of their dose response. We found that that both Fab 3H4 is a more efficient inhibitor than Fab L4B8.
In vivo evaluation of Sema3A inhibiting antibodies
Using a rat axotomy model (experiments carried out within the project also by TAU) we found increased levels of Sema3a in the retina following optic nerve axotomy. We found that several antibodies, but in particular antibody 3H4 could partially protect retinal ganglion cells (RGCs) from death following axotomy. To examine whether axotomy will change the pharmacokinetics of 3H4 Fab, we labelled antibodies with fluorescent dyes, we intraocularly injected fluorescent antibody and at different time point measured the level of the antibody. As shown in Fig. 4, We found that IgGs are cleared more slowly from the rat eye that Fabs and also found that that clearance kinetics was markedly slower in the axotomized eyes, suggesting the formation of immunocomplexes in the eye. As a control experiment we asked whether the reduced rate of antibody clearance following axotomy was specific to anti- Sema3A antibodies. For that purpose we repeated the experiment. Whereas axotomy significantly reduced the clearance of 3H4 Fab, no such effect was detected for non-specific (isotype control) antibody demonstrating the higher levels of Sema3A (as a result of axotomy) are responsible for reducing the clearance kinetics of 3H4 Fab.
Antibody-containing implants and additional in vivo studies
In order to determine whether the delayed clearance of Fab 3H4 was due to sema3A elevation rather than the injury itself, the experiment was repeated with an unrelated Fab against streptavidin (SA). Two groups of rats (3 rats in each group) were injected with either Alexa 594-conjugated Fab 3H4 or Alexa 594-conjugated Fab SA following optic nerve axotomy. Two additional groups served as control - they were injected with either Fab 3H4 or Fab SA without injury to the optic nerve. No difference in the clearance rate was detected between the axotomized and control groups injected with Fab SA. On the other hand, we again observed a significant difference between the clearance rates of the axotomized and control groups treated with Fab 3H4 both 24 and 48 hours following injury. These results indicate that the slow clearance rate in the axotomized group treated with Fab 3H4 did not arise from the axotomy of the optic nerve itself but rather from the formation of immuno-complexes between 3H4 Fab and sema3A.
A low molecular weight inhibitor of Sema3A by modulating the structure of SICHI peptoid
Clone, express and purify Sema3A.
In addition to cloning, expression and purification of Sema3A, we have developed a method for the efficient production of unlabelled and isotopically labelled (for NMR studies) constructs of the Sema3A C-terminal domain. Moreover, we have worked on the development of an efficient method of production and purification for the Sema3A C-terminal domain in sufficient amounts for biophysical studies. In the case of the Sema3A C-t, predicted to be unfolded, E.Coli was the method of choice for the expression of the protein. E.Coli allows the production of isotopically labelled proteins as it can grow in minimal medium (M9). Initial purification was performed in denaturing conditions (8M Urea or 6M guanidinium chloride) to avoid protein degradation. Given that the Sema3A C-terminal domain has a highly basic isoelectric point (pI > 11), the protein can be produced without any affinity tag (such as poly-His or GST), and purified by cationic exchange at high pH (9-10), which removes most of the proteins that fall in a more physiological range of pls. The amount of pure protein following this protocol afforded modest yields, (0,75-1,5 mg/L). Whereas these amounts might be sufficient for most of the experiments planned, problems arose when attempting to produce isotopically labelled protein (15N or 15N/13C) for NMR studies. Another strategy was the use of synthetic DNA templates with optimized codon usage, in our particular case, for E.Coli. This approach was useful and we were able to increase the levels of Sema3A C t domain expression from 0-1,5 mg/L to 2-12 mg/L (depending on the specific construct) for the codon optimised constructs. For the best expressing constructs (all construct boundaries and purification yields were listed in previous Reports), we were also able to obtain 15N-labeled protein, at around 2 mg/L. Although these amounts could be enough to perform some NMR experiments, we tried to further increase the levels of expression for 15N-labelled constructs by using the expression system developed by the BioSilta company, which essentially is a method that optimizes the growing and protein expression conditions for E.Coli (detailed protocols at http://biosilta.com/recombinant-proteins/e-coli/protein-for-nmr-analysis/). In this way, we were able to get yields for 15N-labelled Sema3A C-t constructs between 5-15 mg/L, even higher than the yields for the corresponding unlabelled constructs prepared in LB medium. Another strategy that we tested for partial Sema3A C-t labelling consisted in the methylation (using 13C – formaldehyde) of purified unlabelled protein. The strategy worked out well and methyl groups were introduced in all (6) lysine side chains of the domain, as confirmed by MALDI-TOF.
The interaction between SICHI and Sema3A by NMR
The molecular details of the interactions between the Sema3A C-t and GAGs have not been yet described. A deeper knowledge of this association can add valuable information regarding the proposed mechanism for the action of SICHI (our initial small molecule hit, and derived analogues in the inhibition of the Sema3A signaling. The goal of this part was to study the Sema3A C-terminal - GAGs interactions by the use of a combination of biophysical techniques, namely, Nuclear Magnetic Resonance (NMR), Surface Plasmon Resonance (SPR), and Fluorescence spectroscopy. In addition, the hypothetical perturbations that SICHI and analogues exert on the Sema3A C-t - GAGs association were investigated by binding competition experiments (using NMR or fluorescence).
Our data support the hypothesis that the interaction between Sema3A and GAGs observed at cellular level and described in the literature takes place through the involvement of the C-terminal polybasic region. Due to the importance of Sema3A-GAG interaction in the Sema3A signalling pathway, this knowledge is valuable for future therapeutic strategies for Sema3A modulation. The prime candidate for SICHI action was the protein-protein interaction between secreted Sema3A and Nrp1 receptor, but our results showed that SICHI does not significantly bind to the Sema3A sema domain nor to the Nrp1 extracellular domain. Therefore, we sought for an alternative mechanism for SICHI inhibition of the Sema3A pathway. Accordingly, we studied the interaction of SICHI and positively charged model peptides (FS2 and FS3, for structures cf. previous Reports) with GAGs by means of biophysical techniques and molecular dynamics simulations. First, the biophysical studies show that the two peptides FS2 and FS3 bind to GAGs and therefore, can be used as models to study Sema3A C-terminal domain-GAGs interaction. Next, we confirmed that SICHI binds to GAGs. Finally, we demonstrated that SICHI efficiently interferes in the Sema3A-GAGs interaction. Both NMR and fluorescence spectroscopy experiments showed a dose-dependent SICHI competition effect. Also, the acquisition of NMR spectra in the presence of salt (150 mM NaCl) showed that SICHI could compete with cationic peptides for the binding to GAGs in physiological salt conditions. By comparing the binding constants obtained by isothermal titration calorimetry (ITC) for the interaction between peptides or SICHI with heparin, a GAG model, we observed that the affinity of FS2 and FS3 peptides to heparin was 5 to 10-fold stronger than the affinity observed for SICHI, and these results are in agreement with the partial displacement of peptides from heparin exerted by SICHI and observed by NMR and fluorescence. These results support our hypothesis that the in vitro biological activity observed for SICHI could be related to SICHI exerting its function by binding to the GAG side in the C-terminal Sema3A-GAG interaction.
The medicinal chemistry strategy mostly used until now to inhibit protein-GAGs interactions has consisted of generating a target protein-specific synthetic glycan. In the present case, the suspected protein target for GAG interaction was the 55-amino acid long Sema3A C-terminal basic region (with 6 Lys and 10 Arg). Our results provide confirmation that a small peptoid molecule, such as SICHI or, more interestingly, the CSIC-002 analogue, could inhibit the interaction between polycationic peptides and GAGs, and provide a set of information that could be exploited for the development of protein-GAGs interaction inhibitors of therapeutic interest. At this point stands the studies carried out along the second half of the Project for optimising the activity of SICHI by designing new analogues. Among them, CSIC-002 has resulted to be the most active compound in vivo. However, in front of the uncertainties about the possible intellectual protection of this compound, we have worked on the design and preparation of a second library of SICHI analogues in addition to the testing of other polycationic amino derivatives (i.e. spermine). At the present time, a compound more active than CSIC-002 has not been yet identified. Taken overall, our findings have opened a new way for developing potent inhibitors of Semaphorin 3A.
Generation of small molecules for Sema3A inhibition by synthesis of a library of peptoid trimers
and of dynamic combinatorial libraries (DCLs)
This objective was mainly carried out along the first 18 months of the Project and it is fully described in the 1-18M Report. Briefly, a library of 14 SICHI derivatives was designed, synthesised, purified and structurally characterised. From the in vitro evaluation of these library components it was identified CSIC-002 as the most promising hit inhibitor of Sema-3A. In fact, this compound has been converted in the lead compound of the Project regarding small organic molecules. Consequently, in this second part of the Project, our group has been working in improving the preparation protocol for this compound (cf. Section 3.6) and has provided to Consortium members (in particular TAU, Synovo and Nicast) of CSIC-002 batches as demanded.
The following conclusions were reached:
▪ The first procedure of the preparation of SICHI was optimised. However, the methodology developed will be hardly applicable to the synthesis of the large amounts of this compound required eventually for preclinical trials.
▪ The synthesis of an analogue of SICHI, CSIC-02, exhibiting promising in vitro and in vivo activities as Sema-3A inhibitor, has been also studied. We developed a methodology that permits the scale-up of this compound and then fulfil the demands for further preclinical trials.
▪ AS a conclusion of the above methodologies, the preparation of a collection of SICHI analogues bearing different peptoid and heterocyclic structures was perpared. These compounds were delivered to for biological testing.
▪ From the computational studies, it is evident that SICHI and CSIC-02 could bind to the GAGs models selected by establishing several polar interactions, i.e. hydrogen bonds and salt bridges, mainly between the cationic moieties of SICHI/CSIC-002 and the anionic sulfate/carboxylate groups of the GAGs.
In addition, a virtual screening of a library of commercial lead-like compounds (650.000 from the MOE lead-like library, Chemical Computing Group Inc., Montreal, Canada) against the Neuropilin-1 binding domain of Sema 3A rendered interesting possibilities. Specifically, hits consisting of two aromatic rings bound through a cationic linker chain were identified. With this information, the possibility of constructing a dynamic combinatorial library (DCL) with a diamine and different aldehydes was highlighted. The imine formation/exchange is a suitable covalent bond to perform DCLs, because the reaction is reversible and the equilibrium can be shifted to adapt to the presence of the target. Moreover, further reduction of the imine bond would freeze the equilibrium, allowing the isolation of the amplified species. Accordingly, we designed a library of potential modulators of protein-protein interactions generated from two diamines and five aromatic aldehydes (see below), and used DCC to identify the best binder(s) to control the Semaphorin 3A-Neuropilin 1 (Sema3A-Nrp1) interaction.
A blind docking screening process was carried out with those compounds from the combinatorial dynamic library (CDL) that were amplified in the presence of Sema3A (1A, 1AB, 1AC, 1AD, 1AE, 1B, 1BC, 1CD, 1CE, 1D, 2AC, 2BC and 2C), to determine their potential binding site(s) on the protein. Thus, among 14 potential interaction sites identified on the surface of Sema3A, two of them were determined to be preferential sites for interaction, based on the higher docking scores obtained for most of the compounds bound at these sites. From these two sites, one of them involving residues 240-242, is located on a region of the Sema 3A surface that has been crystallographically identified as the interface that interacts with Neuropilin-1, suggesting that if the DCL compounds bind at this site they could interfere with the interaction between these two protein partners (Sema 3A – Nrp1). Two compounds that showed the best docking scores at this site, i.e. 1CE and 1AD, were also among the most amplified of the DCL. The second site identified on Sema 3A, involving residues 309, 310 and 490, is located in a funnel shaped cavity located in the middle of the protein. However, the biological relevance of the interaction of the DCL compounds at this site cannot be predicted at this moment.
From the 30 possible compounds, 7 potential binders were identified by the DCC. These compounds and two inactive ones were synthesized separately to evaluate their biological activities. We performed some STD-NMR studies of these compounds with Sema3A and Neuropilin-1, both together and separately. These compounds were also delivered to TAU to assess their biological activities. Summarizing, ligand-protein binding studies done by NMR showed that the most amplified compounds obtained by DCL bind Sema3A sema domain, but they exhibit poor selectivity (these compounds bind also Nrp1). Also, our results indicate that these compounds do not bind the Sema3A-Nrp1 binding interface (PPI region), which was our initial target. Consequently, at the present time this approach has not rendered results as satisfactory as those obtained from the initial library of peptoid derivatives.
Scale-up synthesis of lead candidate for Sema3A inhibition
The scale-up synthesis of CSIC-002, the small organic molecule that has emerged as lead candidate for Sema3A inhibition is described with full details in Deliverable 3.5. The scale-up synthesis of two lead candidates for Sema-3A inhibition (SICHI and CSIC-002) has been successfully achieved. We have developed a new convergent 2+1 strategy to overcome some problems found during the optimisation of the previously proposed procedures. Moreover, we have discarded the use of solid-phase synthesis steps because a procedure carried out in homogeneous media is more suitable for this scale-up. Consequently, we are capable of producing batches of hundreds of milligrams of either CSIC-001 (SICHI) or CSIC-002 (the current lead candidate) and it is not expected that a multigram production would add any additional problem to those encountered and solved in this study. Preparation of CSIC-002, the selected lead compound identified in this Project, involves 8 steps following a convergent strategy with different purifications steps during the process. The most important drawbacks in the scale-up of these syntheses are the distillation processes. But we think that when trying to scale up the synthesis at the industrial level, these problems will be overcome due to the equipment facilities of the Fine Chemicals companies.
Significant results
1. The production of selected protein fragments related to Sema3A and Neuropilin-1 has been optimised. In particular, different residues from the terminal amino basic tail of Sema3A have been produced in enough amounts to permit biological evaluation as well as the performance of biophysical Studies (NMR, SPR, ITC, fluorescence and UV-VIS spectroscopy). This production has involved also the construction of residues labelled with 13C and 15N addressed to specific NMR experiments.
2. The library of SICHI analogues designed and constructed led, after the corresponding activity assays, to the identification of compound CSIC-002 as lead candidate for the inhibition of Sema3A.
3. Concerning the inhibitory mechanism of action of SICHI and CSIC-002 on Sema3A, our biophysical and computational studies studies support the hypothesis that the interaction between Sema3A and glycoaminoglycans (GAGs) observed at cellular level and described in the literature takes place through the involvement of the Sema3A C-terminal polybasic region. In fact, we have confirmed that SICHI binds to GAGs, and that SICHI efficiently interferes in the Sema3A-GAGs interaction. The suspected protein target for GAG interaction is the 55-amino acid long Sema3A C-terminal basic region (with 6 Lys and 10 Arg). Our results provide confirmation that a small peptoid molecule could inhibit the interaction between polycationic peptides and GAGs and provide a set of information that could be exploited for the development of novel protein-GAGs interaction inhibitors of therapeutic interest.
4. The scale-up synthesis of our lead candidate, CSIC-002, has been optimised for the preparation of tens of milligram batches of this compound.
5. The first dynamic combinatorial library of potential Sema3A inhibitors has been designed, constructed and characterised. From the structural features learned, an improved version of this library is in due course.
An active polymeric implant (loaded with Sema3A inhibitor
Development of polymeric formulations by electrospinning loaded with small molecular weight sema3a inhibitors
Several small molecular weight sema3a inhibitors were synthesized and tested for stability and potency by CSIC and TAU, leading to the selection of the CSIC-002 inhibitor as the leading candidate for further studies. Throughout the process, NIC has been developing polymeric formulations containing the inhibitors to enable successful electrospinning forming a nano-fibrous membrane. Several hundred formulations were evaluated, each of these tested in the electrospinning setup by tuning various process parameters, followed by wide characterization of the manufactured material.
Formulation parameters that were evaluated include polymer type, composition and concentration (PCL, PLGA 50:50, PLGA 75:25), solvents type and composition (N,N-dimethylformamide (DMF), dichloromethane (DCM) and mixtures of thereof) and inhibitor concentration. Some of the electrospinning process parameters that were evaluated include the electrical potential, solution flow rate, distance, time, mandrel type, number and type of needles, drying and post-treatment conditions.
Fine tuning of the electrospun prototypes was conducted using low and high loaded CSIC-002 samples: 13 and 14% PLGA 50:50 loaded with 3-4% w/w CSIC-002, 22% PLGA 75:25 loaded with 3% w/w CSIC-002, 14% PLGA 50:50 loaded with 7 and 10% w/w CSIC-002 and 22%PLGA 75:25 loaded with 6 and 8% w/w CSIC-002.
Viscosity, density and conductivity were measured for each formulation prototype.
Development of polymeric formulations loaded with sema3a antibody inhibitors
Parallel to the development of a formulation containing small molecular weight sema3a inhibitors, sema3a antibody inhibitors were also considered. However, due to the bio-chemical nature of the antibodies, they could not be used in organic solvents formulation. Therefore, a core-shell structure, consisting of an organic soluble polymeric shell, and a water soluble polymeric core containing the antibody, were designed. Such a structure could be achieved using the co-axial electrospinning method. The development of the co-axial electrospinning setup was highly challenging due to delicate design and tuning required of the various solution and process parameters, which is essential to achieve a stable and reliable electrospinning process. The co-axial setup was established for PCL based shell and water soluble polymeric based core (PEG35K, POLYOx300, PVA). PCL as shell and PEG35K as core have been selected as most relevant candidates following preliminary experiments IgG antibody (Avastin) and human gamma globulin (HGG) were selected as antibody model to be incorporated in the core. Erbitux and 3H4-Fab manufactured by Prof. Itai Benhar at TAU were also incorporated to obtain a prototype loaded with antibody sema3a inhibitors. The development of the co-axial electrospinning process required measurements of surface tension and conductivity, in order to better understand the mutual influence of the core and shell jets in the formation of the Taylor’s cone in presence of the antibody inhibitors. The core-shell nano-fibrous structure was evaluated using fluorescent confocal microscopy, by incorporating fluorescent dyes in the core and shell solutions.
A sugar (trehalose) based method was developed in parallel to obtain a simpler electrospinning process, enabling to incorporate sema3a antibody inhibitors in the formulation. This technique requires a single electrospinning process, whereas the sugar acts as a stabilizer to protect the antibody against organic solvent.
Development of implant by electrospinning loaded with small molecular weight sema3a and antibody inhibitors
Following the development of the various formulations containing the CSIC-002 inhibitor, an implant that could be easily inserted into the eye was designed. Different membranes consisting of submicronic fibers of PLGA loaded with Sema3A inhibitor (CSIC-002) were electrospun (with a 15 mm mandrel) from polymeric solutions. Scanning electron microscopy (SEM) imaging was carried out to study the morphology of CSIC-002 loaded PLGA electrospun fibers, whereas recovery of the inhibitor was determined to be 100%. Samples of 13 and 14% PLGA (50:50) – prototype I and II – as well as 22% PLGA (75:25) – prototype III – loaded with different %wt of CSIC-002 resulted in submicronic fibers with width ranging from 0.2 to 1.4 µm for PLGA (50:50) and 0.4-4.2 µm for PLGA (75:25). Thicker fibers were obtained from PLGA (75:25) and present a much more curved structure, compared to those based on PLGA (50:50)
The release kinetics of the CSIC-002 from the electrospun fibers were analysed by SYN using a model of the vitreous humor with constant aqueous flow washing. The obtained release profiles were used to study the release mechanism and prioritize the formulations most suitable for in vivo testing. It was observed that prototype I samples, 3% CSIC-002 @ 14% PLGA (50:50) demonstrated an ideal and linear drug sustained system with a constant delivery rate; 55% of CSIC-002 was released after 60 days. In the relatively high load samples, i.e. 7% CSIC-002 @ 14% PLGA (50:50) and 8% CSIC-002 @ 22% PLGA (75:25), all the CSIC-002 was released within 7 days. The experimental results of CSIC-002 release as well as its proposed kinetic model demonstrated that an erosion mechanism of the fibers controls the drug release. The ideal prototype of 3% CSIC-002 loaded PLGA (50:50) has been chosen to be the leading candidate for future in vivo study due to the fact that degradation of the nano-fibers leads to a constant and linear release rate over a long period up to 60 days with a zero-order-kinetic.
The human gamma globulin (HGG) release, however, was very slow resulting in less than 1% release after 30 days. (Increasing the concentration of the HGG could not be achieved, as it led to an unstable electrospinning process. The slow release, and the requirement to dilute the antibody during the solution preparation, could not achieve the therapeutic demands of the implant, and therefore, at this point, the antibody inhibitor process can not present feasibility.
Based on the chosen formulation of 3% CSIC-002 @ 14% PLGA (50:50), a cylindrical implant was designed in order to be tailored for the use of OZURDEX® applicator for intravitreal injections.However due to shifting from membrane to hollow cylindrical geometry, the formulation had to be slightly altered. The cylindrical implant formulation that was found to mimic the membrane release profile was of 2.7% CSIC-002 @ 13% PLGA (50:50).
The cylindrical implant was later re-evaluated and a new design of 0.24mm wide membrane stripes was found more suitable for insertion into the eye, using a standard 23G cannula.This design, again, required an update of the formulation. In addition, the drug release profile was modified to reflect a quick, up to 7 days release profile. This modification follows a literature search and evaluation of the negative systemic effect, advocated in the literature, of the sema3a inhibitor.
The final formulation which was achieved after optimization of the electrospinning process was 5.5% CSIC-002 @13% PLGA (50:50), exhibiting an initial quick release of 37% of the active compound within 7 days, followed by a slower rate additional release.
Implant by extrusion loaded with sema3a inhibitor small molecule (CSIC002) and sema3a antibody (3H4)
To complement the achievements of the implant development using the electrospinning technic, and to overcome the barrier limiting the amount of substance capable of being loaded in the implant manufactured by electrospinning, an alternative manufacturing process was proposed.
A laboratory scale extruder was designed and created by the partner Synovo. Several implants were manufactured using biodegradable polymers like PCL, PLGA, PLA, and PEO, and testing as well different MWs of the previously mention polymers or different PLA and PGA ratios in PLGA.
Prior to extrusion, the polymer is combined with the substance, for the polymer encapsulate the substance. The process is different for each substance. Once the mix is done, the extrusion can be performed. The extrusion device is made of stainless steel and consists on the loading cavity, the extrusion channel, and the pressure mechanism with a rotating handle. When turning the handle, the material in the loading cavity is forced into the extrusion channel leaving the minimum left material in the loading cavity. The system has a temperature regulator incorporated. The required temperature depends on the properties of the implant ingredients. As a final product, a long cylinder exists through the extrusion channel to the exit hole.
The small molecule and the antibody were successfully integrated into the implants up to 30-33 % w/w loading. Several implants were created and the drug release studied until two lead extrusion based implants were selected as leading products. The lead extrusion based implants are loaded with 30% sema3a inhibitor small molecule (CSIC002) and 33 % sema3a antibody (3H4) respectively. They have the shape of a long cylinder of 0.3 mm diameter. The length can be adjusted after manufacturing for lower or higher dose.
To study the drug release profile of the implants loaded with the small molecule, a laboratory model was created within the project, which mimics the intravitreal environment. It consists in a device with cavities including synthetic vitreous mimicking the eye, washed with a constant flow mimicking the aqueous flow through the eye. The samples of the synthetic aqueous flow exiting the vitreous cavity were analysed by HPLC/MS/MS, Quantification of the remaining substance in the implants after the release study was as well done to confirm previous results.
Both extrusion based implant candidates were tested in vitro and in vivo with positive outcome.
• SYN-102-30.
Manufacturing method: extrusion.
Ingredients: PCL and sema3a inhibitor small molecule.
1 cm implant releases >11µg/day for the first 9 days and slower release for the next 6 days. Maximum dose of 52µg/day on day 3.
• SYN-103-33.
Manufacturing method: extrusion.
Ingredients: PCL and sema3a inhibitor antibody.
In vivo studies of the active implant (Sema3A inhibitors), as per EMA guileless for clinical studies
Pharmacokinetics
At the beginning of the project, ex vivo studies were performed with SICHI and analogs provided with an idea of the distribution of the substances in the eye. CSIC-002 was injected in ex-vivo eyes from rabbits (taken post mortem from cadavers used for other purposes) for a concentration in the eyes of 1 or 10 µM (2 groups, 5 eyes per group). The eyes were incuvated 3h at room temperature under slight shaking and overnight at 4°C. Vitreous, lent, retina, pigmented epithelium and optic nerve were extracted and analyzed.
i.v. pharmacokinetics and toxicokinetics murine studies in mice
BL/6 male mice were subjected to single i.v. administration of CSIC001, CSIC002, CSIC004, CSIC008 or CSIC010 (1.8µmol/Kg of each. Formulation: serum). Samples from peripheral plasma at 5min, 15min, 30min, 1h, 2h and 3h after injection were anlysed. The animals were sacrificed 3 h after dose and different organs were sampled for analytics. Following i.v. administration, the half-life in plasma of the CSIC compounds was ca. 30 min and no compound was found in plasma 2 h after dose. Every CSIC compound was found in the eyes 3h after i.v. administration. This confirms the affinity of the Sema3a inhibitor for the eye, especially CSIC-004 and CSIC-002. No signs of intolerance or rejection was observed. No toxicity effects were assigned to the treatments.
Intravitreal pharmacokinetics and toxicokinetics in mice
16 mice, albino Balb/C and pigmented BL/6 mice were subjected to intravitreal administration of 0.15 µg/eye of CSIC001 or CSIC002 (10µL/eye). Two hours after injection the animals were sacrificed. The concentration of the CSIC compounds in the eye was analysed. Following intravitreal administration, no systemic exposure was presented. CSIC002 presented high concentration in the back of the eye, compared to other eye parts. No signs of intolerance or rejection were observed. No toxicity effects were assigned to the treatments.
Subcutaneous toxicokinetics in mice
Following previous tolerability studies with the sema3a inhibitor active substance, the activated implants, loadeded with sema3a inhibitor, were tested in several in vitro and in vivo studies, including toxicity and pharmaokinamic studies. No signs of rejection or inflammation were presented. The implants were well tolerated.
The lead PCL implants manufactured by extrusion were tested in s.c. murine model. The s.c. implants were well tolerated during a 39 days study in 18 C57BL/6 mice. Each implant injected was 0.7 cm long. The implants injected had 0.3 mm, 0.5 mm or 1 mm diameter. The mice were observed during 39 days and the implants were monitored by palpation regularly along the study. On day 39, some of the implants had been broken due to the palpation, however, those which were still s.c. has the original length and appearance. Pictures of the implants after sacrifice of the animals were recorded.
Intravitreal toxicokinetics in rabbits
Four New Zealand Rabbits were subjected to intravitreal injection of extrusion based PCL implant in the left eye. The animals were observed 3 times a week during the first 16 days and weekly after. Half of the animals were sacrificed after 2 weeks and the other half 4 weeks after injection. No signs of rejection and/or inflammation were observed in the animals. Both control eye and implant carry eye presented same appearance during the whole study. The cornea, lens, retina and vitreous presented healthy a status.
Complete cut of optic nerve axons within its covers created in laboratory rats
Adult male rats, 12-15 –weeks old, were deeply anaesthetized. The right optic nerve(ON) was exposed by separation of the bulbar conjunctiva and entering the orbital space retro bulbar. While separating the surrounding extra ocular muscles and surrounding tissue the optic nerve was exposed. Using a special designed glass dissector with a blunt tip of 50 microns diameter and opening was created in the surrounding meninges and with a slow movement the axons of the nerve were separated. At the end of the procedure a clear separation of two optic nerve stumps can be observed. The injury was unilateral in all animals done in the right eye. The surgery was immediately followed by an injection of Sema3A inhibitor or saline (as control) into the vitreous space, close to the retina. After 12 days the retrograde neurotracer 4-di-10-Asp was inserted into the optic nerve between and near the sight of injury and the eye globe. This neurotraces stains only live retinal ganglion cells (RGC).Four days following the staining , the animals were sacrificed. The retinas were separated from the eye and mounted on slide to be viewed under a fluorescence microscope. Random fields were selected and the number of live RGC were counted.(Ref 1)
Retinal detachment in rats
The same type and age of rats were used for creating retinal detachment in the right eye. Under deeply anesthetized rat (Xylasine 50mg/kg and Ketamine 35 mg/kg) and following dilatation of the pupil with Tropicamide drops 0,5% the retinal detachment was created. The procedure was done using a syringe with a 32G needle and containing a volume of 5 microliter of saline. The needle was in traduced at the cornel border in the anterior cahmber and introduce under the iris between the lens and the ora serata ( the dead end of the retina.). The needle was introduced under the retina and the saline injected. This procedure created approximately half of the retinal area detached.The rats were devide in two treated groups:Group A was immediately injected with 2 microliters of Fab 3H4 and Group B 3 days later.The third group was injected with saline and served as control. The same protocol of staining the RGC and flat mounting of the retina for RGC live counting as described in previous chapter was done also in this experiment.
Acute glaucoma in rabbits
Male, 12 weeks old, New Zeeland albino rabbits were used in creating this model. Under deeply anesthetized rabbits, (Xylasine and Ketamine), High intraocular pressure (IOP) was created by introducing in the anterior chamber a maintainer cannula (used in cataract surgery and connecting it to an intravenous infusion pack of 500 cc volume containing saline. The pack was fixed at 80 cm height and created an IOP of 45 mmHg to 50 mmHg. This IOP was maintained for 1,5 hour. Twenty four hours following the assault, intra ocular implants (PLGA – created by Synovo) loaded with small molecule CSIC02 or blank implants, serving as control, were introduced into the eye globe using vitrectomy technique. Following 14 days from the assault the same procedure of staining the RGC, flat mounting and counting live RGC was done.
The intra ocular injection of CSIC02 in the right eye of rats following the creation of complete cut of the optic nerve presented a significant number of live RGC following the assault. The retinas were stained at the 14th day following the injury. This is the climax time of the apoptotic death program created by Sema3A.
There can be observed that about 28% of RGC are live when compared to normal retina and the saline treated retinae present only 1,5% of live RGC.
The results of intra ocular injection of Fab 3H4, to inhibit the apoptotic death of the RGC following complete cut of optic nerve axons in the right eye of rats: Retrograde staining of the RGC reveals that over 40% of RGC are live in the treated eyes while compared with normal retinae. Only 2% of RGC remained live in the saline treated group.
Intraocular injection of Fab 3H4 saved RGC following the creation of retinal detachment in the right eye of rats. In contrary, the right eye of injured optic nerve of rats with retinal detachment and treated with unrelated Fab against streptavidin (SA) presented only few live RGC.
Conclusion
In all three models of assault to RGC, the intraocular injection of small molecule CSIC002 and Fab 3H4 presented a clear salvation activity of the RGC following the assault to the optic nerve and retina.
Both approaches have potential to be a preventing degeneration drug in degenerative processes of CNS.
No toxic or inflammatory reaction was observed during the whole period of the experiments.
Regulatory aspects - Draft IMPD using the data generated
To support and optimize the development of an injectable polymeric implant (loaded with the Sema3A modulator) the VISION project has included a follow up of the project’s developments and a design of the future regulatory aspects that will need to be covered to advance the final product to the clinical trials phase.
On the grounds of the European regulation, the relevant regulatory guidelines for VISION implant development into clinical phases have been identified. The drug product developed in the VISION project is a combination product consisting of a Sema 3A inhibitor and a polymeric implant marketed. The Sema 3A inhibitor falls in the category of medicinal products, whereas the polymeric implant is a medical device, thus two different types of regulations apply to the final product development. In addition, the regulatory aspects that will guide the preclinical development are partially dependent on the type of clinical design that will be used in the future. The consortium came to an agreement on which would be, a priory, the main principles of an initial trial based on the initial data and the clinical experience of TAU:
• The indication foreseen will be ischemic optic neuropathy
• The most appropriate clinical study would be a phase IIa (proof-of-concept) clinical trial in patients with ischemic optic neuropathy. A phase I clinical trial will be probably not feasible for ethical reasons since the implant should be introduced into the eye
• The duration foreseen for the first clinical trial would be at least 3 months with a time point at 6 months
• Single or multiple administrations, but not a chronic treatment, is envisaged
• The proposed therapy will be administrated together with current treatments mainly intended to lower IOP (intraocular presure)
• Two scenarios will be considered for the preclinical development proposal: (1) polymeric implant +NCE and (2) polymeric implant + MoAb
Taking into account international standards and guidelines, a toxicological program enabling clinical phase IIa with Implant + NCE or Implant + MoAb has been designed. A list of go/no-go decisions based on regulatory needs have guided the non-clinical development of the VISION implant within the project and projected beyond.
A market evaluation reviewing available implants for ocular administration and other polymeric dosage forms has been done in order to gather relevant information for the design, development and fabrication of the VISION implant. An epidemiological analysis of the selected indications (retinal vein occlusion, ischemic optic neuropathy, retinal arterial occlusion, traumatic neuropathy and acute demyelinating optic neuropathy) to select the best optimal target for further exploitation has also been included.
An evaluation of the potential cross-reactivity risk between the compounds developed in VISION (NCE / MoAb) and semaphorins other than 3A has been done to assess the need of secondary pharmacodynamics studies. The possibility of such undesired interactions taking place in the eye environment was estimated on the basis of available expression and structural data. The subsequent modulation of the expected therapeutic outcome was assessed taking into account functional data.
Commonly accepted inflammatory markers that could be used in the evaluation of the implant based on regulatory guidelines and standards have been identified to ensure that the experiments done in the course of the project can support as much as possible the requirements of the regulatory requirements that the product will face later on.
The regulatory pharmacodynamic studies are better accepted with at least some knowledge about the mechanism of action of the compound under study. During the course of the project the SICHI compound has been suggested to achieve neuroprotection through the interaction with glycosaminoglycans (GAGs) in the eye, which have been reported to have a regulatory action in the apoptotic pathway. With the aim of properly guiding the development of new compounds for glaucoma through any of the following strategies: maintaining Sema3A-Nrp1 as a target or target the interaction between Sema3A and GAGs; a market and literature survey on other compounds targeting GAGs has been included. The aspects evaluated included: a) the commercial use of GAG modulators, b) the prior art regarding GAG modulation in the treatment of neurodegenerative diseases, c) the suitability of these compounds for treating glaucoma and d) the use of these compounds as a positive control for future studies with SICHI.
Chemical development guidelines along with current protocols for production of CSIC-002 and analytical measurements have been compiled setting the bases for further developments under GMP (Good Manufacturing Practices) by an appropriate CMO (Contract Manufacturing Organization). A pre-selection of GMP-compliant facilities has been included
The relevant non-clinical pharmacology, pharmacokinetics and toxicology studies performed within the project have been compiled in the format of the Investigational Medicine Product Dossier (IMPD) for the small molecule (CSIC-002) and a monoclonal antibody (Fab 3H4). The regulatory studies needed for NCE-implant and the MoAb-implant are coincident in many points but there are some relevant points that need to be addressed differently, thus two different documents have been generated in order to be able to use them independently in case of independent further development. The relevant results from in vitro and in vivo studies conducted within VISION project have been used to address the efficacy and safety aspects of the document. While efficacy studies do not have to be conducted under GLP, thus the data collected may be sufficient to support an application, the safety studies need to be complemented with studies performed under GLP terms. The data collected within VISION has been included as supporting data for the safety aspects of the document. The further compulsory safety studies that need to be performed under good laboratory practices (GLP) prior to moving to a clinical phase have been identified. The protocol outlines of GLP-Compliant Studies have been set-up and optimized based on regulatory guidelines and the data obtained in the course of the project (41 protocol outlines) and linked to the corresponding sections of the IMPD.
An implementation calendar of GLP-Compliant Studies in a 15-month period that guarantees GLP compliance at minimal cost, based upon European Medicines Agency requirements and practical experience has been outlined. A pre-selection of GLP-compliant facilities able to provide services that meet the needs and expectations of the VISION project has also been included.
A basic introduction and guideline for a request of Scientific Advice to the EMA has been included along with some possible questions for the agency.
Potential Impact:
Impact
Enhancing Europe competitiveness
Furthermore the European Commission has identified active and healthy ageing as a major societal challenge common to all European countries, and an area which presents considerable potential for Europe to lead the world in providing innovative responses to this challenge.
The pilot European Innovation Partnership on Active and Healthy Ageing is aiming to:
1. Enabling EU citizens to lead healthy, active and independent lives while ageing;
2. Improving the sustainability and efficiency of social and health care systems;
3. Boosting and improving the competitiveness of the markets for innovative products and services, responding to the ageing challenge at both EU and global level, thus creating new opportunities for businesses.
The overarching target of this pilot partnership is to increase the average healthy lifespan by two years by 2020.
The VISION project supports the above pilot thus enhancing the European competitiveness
Socio-Economic impact
After cancer, Americans fear blindness most . Increasing aging of populations throughout the world, paired with other factors, including an increase in the prevalence of age-related diseases and sedentary lifestyles, will ultimately lead to a far greater worldwide prevalence of many diverse ophthalmic diseases, in addition to those related to only aging and obesity (Figure 3.1)
Blindness and irreversible sight impairment cost an estimated $50 billion each year in the United States and, thus, improving on the treatments available today could substantially reduce the financial burden on society.
The project will lead to a real difference to the lives of elderly Europeans by providing treatment for neurodegenerative diseases (e.g. glaucoma and acute optic nerve neuropathy) to stop further death of neural cells, never achieved before.
List of Websites:
http://fp7-vision.eu/
Coordinator Professor Arieh Solomon
asolomon@post.tau.ac.il