Beta cell preservation via antigen-specific immunotherapy in Type 1 Diabetes: Enhanced Epidermal Antigen Delivery Systems

Final Report Summary - EE-ASI (Beta cell preservation via antigen-specific immunotherapy in Type 1 Diabetes: Enhanced Epidermal Antigen Delivery Systems.)

Executive Summary:
The EE-ASI programme aimed to develop a novel therapeutic approach to the immunotherapy of type 1 diabetes which could slow further loss of insulin production in newly-diagnosed patients and may also have value in preventing the onset of diabetes in at-risk individuals. The programme planned to combine a beta cell epitope recognised by T lymphocytes with tolerance enhancing elements targeted to antigen presenting cells using a gold nanoparticle (GNP) platform and delivery of the combination into the very superficial layers of the skin using microneedles to optimise access to the immune system. The Consortium combined the expertise of 8 partners, including 5 academic partners (Cardiff University (CU), UK; Leiden University Medical Centre (LUMC), the Netherlands; INSERM-Marseille, France; King’s College London (KCL), UK; Linkoping University (LINK), Sweden) 2 SMEs (Midatech, UK/Spain – GNP production; Nanopass, Israel – microneedles) and a management partner (Inserm Transfert, France).
The management of the consortium (WP1) was successfully coordinated by Inserm Transfert, and all reports were provided on time. In WP2, an optimal method for preparation of gold nanoparticles (GNPs) coupled to proinsulin peptide antigen and cytokine was developed that could be transferred to our GMP production facility. It was also defined that when injected intradermally into ex-vivo human skin, these GNPs are not disrupted, retain their negative zeta potential and reflux into the epidermis as well as being distributed in the dermis. Furthermore, human dendritic cells can present peptide antigen bound to GNPs efficiently without inducing maturation and with reduced proinflammatory cytokine production. In WP3 a microneedle technique suitable for use in mice, along with the synthesis of GNPs carrying peptide relevant to the murine system, allowed system 1 to be trialled for tolerance in in vivo mouse models in WP4. The successful transfection of human epidermis with plasmid alone (system 2) when injected via microneedles – although unexpectedly not when attached to GNP - will allow this approach to be used to express beta cell antigens in skin in a future project. In WP4, the peptide conjugated GNPs developed for mice were tested for safety and efficacy in a murine model of type 1 diabetes (NOD mice). The results suggested that peptides conjugated to nanoparticles are safe following single and repeated injections may have benefit in distribution of antigenic peptide for induction of tolerance. However, disease onset was not delayed in this model with GNP-peptide alone suggesting that addition of a tolerogenic adjuvant is likely to be required. In WP5, proinsulin peptide conjugated GNPs were administered for the first time in man. Following regulatory advice, a toxicology programme was completed, an IB and IMPD prepared, GMP production of peptide-GMPs was achieved satisfying stability and quality criteria for clinical trial release and an ethical and regulatory approval obtained from the appropriate authorities in UK and Sweden. Delays were introduced following amendments required by the regulatory authorities necessitating a no-cost extension of 9 months which was approved by the EC. Complexities in contracting resulted in additional delays to opening the Swedish site for recruitment. GNPs were stable for more than 18 months after manufacture were administered in man with no systemic safety concerns after 3 doses. The GNPs were noted to persist in the skin for more than 6 months. Immunological monitoring suggested possible expansion of regulatory cells and skin biopsy showed an infiltrated with a mixture of CD4 and CD8 cells. In WP6, a patent was filed and published for a platform technology based on GNPs. Dissemination activities included 3 international workshops, 24 publications and information sent to over 35,000 stakeholders.
In summary, we have developed a method to generate proinsulin peptide conjugated GNPs that have been tested for safety in a preclinical model and in a phase 1A first-in-man clinical trial and protected this platform technology with a published patent. Data from the clinical trial imply that GNPs persist for over six months in the skin, providing opportunities for enduring tolerance induction if inflammation can be minimized. We have confirmed that the product can be made to GMP and has no major adverse reactions. In addition, we have demonstrated that a second cargo can be attached to the GNPs (eg the tolerogenic cytokine IL-10) and that intradermal injection of naked DNA via microneedles is an efficient method of inducing expression in human skin. Taken together, these data provide the basis of a platform technology based on GNPs for the development of antigen specific immunotherapy for type 1 diabetes and other organ-specific autoimmune diseases ultimately leading to a novel non-immunosuppressive approach to immunotherapy. For type 1 diabetes, a condition in which over 1 million adults and children in Europe are dependent on insulin injections, such immunotherapy should slow the loss of insulin production from the body after diagnosis (and potentially prior to diagnosis) resulting in improved metabolic control, quality of life and long-term outcomes.
Project Context and Objectives:
The overall aims of the EE-ASI programme were to develop a novel therapeutic approach to the immunotherapy of type 1 diabetes which could slow further loss of insulin production in newly-diagnosed patients and may also have value in preventing the onset of diabetes in at-risk individuals. The programme aimed to combine a beta cell epitope recognised by T lymphocytes with tolerance enhancing elements targeted to antigen presenting cells using a gold nanoparticle (GNP) platform and delivery of the combination into the very superficial layers of the skin using microneedles to optimise access to the immune system.
The objectives of the project were the following: i) to couple the T cell epitope, proinsulin C19-A3 with interleukin-10 (IL-10) onto gold nanoparticles (System 1) and show that these elements can be efficiently targeted to human skin dendritic cells (DC) using a microneedle delivery system; ii) to adapt System 1 for use in a well-established mouse model of Type 1 diabetes (NOD mouse) for preclinical testing ; iii) to develop an alternative system (System 2) which uses DNA constructs linked to nanoparticles to express similar elements in DCs ; iv) to generate System 1 to GMP standards and perform a phase 1 study of safety including biomarkers of potential efficacy in subjects with Type 1 Diabetes.

The consortium comprised 8 partners, including 5 academic partners (Cardiff University (CU), UK; Leiden University Medical Centre (LUMC), the Netherlands; INSERM-Marseille, France; King’s College London (KCL), UK; Linkoping University (LiU), Sweden) 2 SMEs (Midatech, UK/Spain – GNP production; Nanopass, Israel – microneedles) and a management partner (Inserm Transfert, France).

The programme was divided into six work packages (WPs) with the specific objectives detailed below.

WP1: Governance, management and coordination
WP1 was dedicated to management activities and dealt with all usual and contractual administrative tasks (financial, reporting, organisation of meetings (annual consortium, ExCom), management of Intellectual Property, ethical issues, gender equality). These included the following activities:
- Coordinating the project’s administrative and financial issues as defined by the European Commission (by acting as a permanent helpdesk for all partners to answer their questions).
- Assisting the interface between the coordinator and the European Commission.
- Organizing and facilitating the consortium and ExCom meetings during the course of the project (agenda, date, place, logistic, minutes).
- Developing and updating effective communication channels within and outside the consortium (address book, public and private websites).
- Ensuring correctly and timely scientific and administrative reporting to the European Commission and reporting to all partners.
- Ensuring that the periodic reports are prepared in the most efficient and practical way according to European Commission guidelines by quality-checking all the partners’ contributions.
- Monitoring complex aspects related to the networks’ specificity, such as information hub and database management, safety issues, legal, ethical and intellectual issues and gender equality promotion (examples: by establishing the public and private websites or training sessions on IP and gender issues).

WP2: Generate System 1 (protein based) for humans and ex vivo testing
WP2 was aimed at defining the core methodology for optimal preparation of nanoparticles linking a pancreatic beta cell antigen (proinsulin C19-A3) to a tolerance enhancing element (IL-10). GNP with either a glucose or mannose C2 corona, and an escalating number (1-10) of proinsulin peptide C19-A3 molecules covalently attached to the gold core, were prepared for in vitro and ex vivo experiments at LUMC and CU, respectively. One of these constructs was selected as a clinical candidate based on the in vitro and ex vivo work performed in this work package. Some of the GNP constructs contained fluorescent markers for imaging studies to support the clinical candidate selection process. This methodology was then used in WP5 for GMP manufacture of the optimised C19-A3-GNP construct and its application in early phase human clinical studies.
Specific tasks/objectives were:
- Generation of peptide-IL10 coupled GNPs.
- Testing of peptide-IL10 GNPs for uptake by human DCs.
- Modification of GNPs to further enhance uptake.
- Formulation of GNPs for microneedle delivery.
- Delivery of GNPs to human skin via microneedles.
- Functionality of microneedle-delivered GNPs.

WP3: Generate System 1 and System 2 (DNA based) for use in preclinical model
The general objective of WP3 is to generate a peptide-IL-10 GNP (System 1) and a DNA-linked GNP (System 2) together with a microneedle-based administration system for in vivo testing in murine disease models of Type 1 diabetes.
The specific objectives of WP3 were:
- To conduct studies on the effects of microneedle injection on skin DC subsets in the mouse.
- To generate a DNA construct for expression of beta cell antigen, IL-10 and other tolerance enhancing elements in murine DCs under the control of an appropriate promoter (System 2).
- To demonstrate uptake and expression of System 1 in murine epidermal DCs.
- To prepare system 2-GNPs (an expression construct to express antigen in a tolerogenic environment targeted to murine DCs via NPs).

WP4: Preclinical safety, PK, PD and efficacy of System 1 and System 2
The specific objectives of WP4 were:
- To provide adequate numbers of NOD mice for the studies in the WP.
- To measure safety, pharmacokinetics and pharmacodynamics of single peptide-IL10-NP. Three different peptides were tested for distribution and safety before and after the addition of IL-10.
- To measure safety, pharmacokinetics and pharmacodynamics of repeated peptide-IL10-NP in vivo and optimisation of treatment regime.
- To measure safety, pharmacokinetics and pharmacodynamics of system 2 (DNA NP).
- Disease prevention with system 1 NPs in preclinical model. Using the protocol established, a total of 60 mice were treated with the optimized protocol to test for the efficacy of system 1 with diabetes observed over 30 weeks.
- Testing system 2 NPs for safety and efficacy.

WP5: Phase 1A study of System 1 in subjects with diabetes
This work package aimed to prepared and conduct a first-in-man clinical trial. The specific objectives were as follows:
- To prepare C19-A3-IL-10 NPs and an appropriate microneedle system to GMP standard.
- To obtain required preclinical toxicology data on the GMP C19-A3-IL-10 NP product and microneedles system.
- To obtain necessary regulatory approvals for the use of C19-A3-IL-10 NP in early phase studies in man.
- To perform a phase 1 study of C19-A3-IL-10 NPs delivered by microneedles in man and obtain safety data.
- To obtain preliminary data on changes in the autoimmune response following C19-A3-IL-10 NP administration in humans with type 1 diabetes.

WP6: Dissemination and exploitation plans
The specific objectives of this work package were as follows:
- Producing as early as possible a Marketing Requirements Document (MRD) that will define all specifications required to reach the target market at the level of our expectations.
- Making sure all participants are aware of and understand the Intellectual Property aspects raised by the project via training activities, common rules and monitoring system.
- Establishing as early as possible a network with future end-users, regulatory bodies and the community for (1) optimizing our dissemination effort (2), preventing bottlenecks and (3) collecting useful feedback from the community.
- Adopting a corporate dissemination strategy, with consortium rules aimed at protecting the IP potential of the generated foreground.
- Reaching in an early phase of the project a consensus on our exploitation strategy that will satisfy all parties involved in the EE-ASI project, avoiding any conflict between the partners.
Project Results:
The scientific and technical results of the programme (WP2-6) are summarised by work packages below.

WP2: Generate System 1(protein based) for humans and ex-vivo testing
WP leader: MIDATECH
Duration from Month 1 to Month 24
Partners involved: Partners 1 CU, 3 MID, 4 NPSS, 5 LUMC, 6 KCL

Summary: This work package aimed to define the core methodology for optimal preparation of nanoparticles linking a beta cell antigen (proinsulin C19-A3) to a tolerance enhancing element (IL-10). The tasks, aims and timelines of WP2 did not change from the original plans. This work package was completed in Period 2 and reported in Reporting Period 2. Some additional GNP constructs were synthesised in Period 3 to support in vitro dendritic cell-T cell experiment elaborated in LUMC. These are reported below in Task 2.a. and Task 2.c. The methodology developed in WP2 generated and validated a method for GMP production of GNPs carrying proinsulin C19A3+/- IL-10 for application in the clinical trial reported in WP5.

Task 2.a: (M1-12) Generation of peptide-IL10 coupled NPs
Partner responsible: Partner 3 MIDATECH
Status: Completed during RP1
Work description and progress:
The task aims to generate GNPs carrying human proinsulin C19-A3 peptide and a tolerance enhancing element (human recombinant IL-10). In parallel to the C19-A3 GNP preparations, GNPs linked to IL-10 or myoglobin and non-GMP indicator peptides to which we have T cell clones e.g. GAD (339 – 352) and EBV peptides were also generated.
Initially, two types of C19-A3 conjugated GNPs were generated: NP51 (10% Glucose C2, 10% C19-A3) and NP52 (10% Mannose C2, 10% C19-A3) using a 75% methanol preparation method. HPLC analysis of peptide released from the C19-A3-GNPs following treatment with KCN solution (100 mM) found that the % of GNPs carrying peptide was 24.6% (NP51) and 20% (NP52, with 2.5 and 1.3 peptides bound per GNPs for NP51 and NP52, respectively. However, incorporation efficiency of 10% C19-A3 peptide into these GNPs was low (5% for NP52 and 3% for NP51). See Task 2c for discovery of a method for significant improvement of incorporation efficiency. TEM size distribution analysis confirmed that the GNPs were 1-5 nm in diameter.

Binding studies, using myoglobin (from equine skeletal muscle) as a surrogate for human recombinant IL-10, as it has similar pI and molecular weight (Mol. Wt. = 17.6 kDa; pI: 7.3) showed that pH 4.6 0.05 M sodium acetate-acetic acid buffer was optimal for electrostatic binding of myoglobin to Midatech formulated GNPs. Microdialysis of human recombinant IL-10 into sodium acetate-acetic acid buffer (pH 4.6 0.05 M) enabled electrostatic binding of human rIL-10 to C19-A3-AuNPs, with 60% human rIL-10 bound to NP51 (10% Glucose C2 (NP55) and 90% bound to NP52 (10% Mannose C2 (NP56)). These GNPs were sent to partners CU and LUMC. Human rIL-10+AuNPs without any peptide (90% Glutathione, 10% Glucose C2) (NP57) were also generated, with 83% human rIL-10 bound. The NP57 GNPs were sent to partner CU. Electrostatic binding of human rIL-10 to C19-A3 GNPs was performed using C19-A3 peptide conjugated GNPs with 5% glucose C2 or 5% mannose C2. rIL-10-C19-A3 GNPs were suspended and stored in Na borate buffer (pH 8.5 0.05 M). Human rIL-10 incorporation into 1% C19-A3 glucose (NP231) and mannose C2 GNPs (NP232) was 90-93% (140 and 135 μg rIL-10/ ml GNP respectively), with 2 and 1.8 rIL-10 bound per GNP, respectively. Human rIL-10 incorporation into 3% C19-A3 conjugated glucose and mannose C2 GNPs was 100% (133 and 140 μg rIL-10/ml for NP259 and NP260, respectively), with 1.6 rIL-10 bound per GNP. The zeta potential of all the rIL-10-peptide-AuNPs were less negatively charged compared to peptide-GNPs, without IL-10. The rIL-10-C19-A3 peptide GNPs and control rIL-10 glucose GNPs were sent to partner 1 CU for zeta potential stability studies and to partner LUMC for in vitro studies.

In parallel to the C19-A3 GNP preparations, GNPs linked to IL-10 or myoglobin and non-GMP indicator peptides to which we have T cell clones e.g. GAD and EBV peptides were also generated using this refined water-based preparation method. Dimerization of the GAD (339 – 352) peptide [TVYGAFDPLLAVAD] was undertaken but this proved to be challenging due to very low yield from the dimerization process. Therefore, monomeric GAD (339–352) peptide (dissolved in 10 mg/ml DMSO) was applied for conjugation to glucose C2 and mannose C2 GNPs. HPLC analysis confirmed that 1% GAD GNPs conjugated to 5% glucose C2 (NP233) or 5% mannose C2 (NP234) carried 1 peptide per GNP, with 84-86% of these GNPs carrying GAD peptide. For 5% glucose C2 or 5% mannose C2 GNPs conjugated with 3% GAD peptide, 88-92% of the GNPs carried GAD peptide, with 2 and 3 peptides per GNP, for NP235 and NP236, respectively. The zeta potential of the GAD-GNPs ranged between -48 mV to -59 mV, indicating good dispersion and stability in water. GAD monomeric GNPs were sent to partner LUMC for in vitro testing. Dialysed human rIL-10 (pH 4.6) was electrostatically bound to 1% or 3% GAD peptide GNPs. The rIL-10-GAD-GNPs were suspended in 0.05 M Na borate buffer. HPLC analysis confirmed that there were ~2 rIL-10 bound per GNP for 1% GAD+rIL-10 (NP318), with 86% rIL-10 incorporation. For 3% GAD+rIL-10 GNP (NP319), there were 2 rIL-10 bound per GNP, with 100% rIL-10 incorporation. The zeta potential of NP318 and NP319 was -21.1 mV and -27.1 mV, respectively compared to -59.8 mV and – 54.6 mV, respectively for GAD-GNPs (NP233 and NP235) without rIL-10, while the supernatants of NP318 and NP319, containing any remaining free rIL-10 and unbound GNP were positively charged (+8.5 mV and +4.1 mV, respectively). These nanoparticles were sent to partner LUMC for in vitro studies.

To determine whether the buffer strength influenced the stability of the rIL-10-GAD GNPs, human rIL-10-GAD monomer GNPs were synthesised and suspended in either 0.05 M or 0.01 M Na borate buffer (pH 8.5). HPLC analysis found 100% release of rIL-10 from 1% GAD monomer-rIL-10 glucose C2 GNPs (suspended in 0.05 M Na borate buffer, pH 8.5) compared to 67% and 58% for 1% GAD monomer conjugated glucose and mannose C2 GNPs, suspended in 0.01 M Na borate buffer, pH 8.5. While it is not clear whether the lower release of rIL-10 may be due to some covalent binding, it was decided to store future IL-10 GNP preparations in 0.05 M Na borate buffer (pH 8.5) since this buffer concentration enabled full release of the rIL-10 from the GNP.

Human EBV peptide [GLCTLVAML] GNPs were generated using EBV peptide dissolved in DMSO (10 mg/ml) as the peptide was insoluble in water. HPLC analysis of peptide released from the EBV-GNPs following treatment with KCN solution (100 mM) found that the % of GNPs carrying peptide was only 1-3% for 1% EBV-GNPs and 6-13% EBV peptide incorporation for 2% EBV-GNPs, with 0.1 and 0.4-0.5 peptides/GNP, for NP and NP respectively. When the EBV peptide was dissolved in a lower concentration of DMSO (1-2 mg/ml), peptide incorporation into 5% glucose C2 and 5% mannose C2 GNPs increased to 15-29% using a methanol-based preparation method. A concentration of 4% EBV peptide was found to be optimal, with 89-92% of GNPs bound to peptide, and 2.6 and 3.2 EBV peptides/GNP, for 5% mannose C (NP185) and glucose C2 GNPs (NP186), respectively. These GNPs were sent to partner 1 CU.

Fluorescently labelled C19-A3 GNPs using 1% C19-A3 peptide conjugated GNPs with 5% glucose C2 or mannose C2 ligand were also generated for partner LUMC for uptake experiments by blood-derived DCs. By fluorimetry, 3% FITC-C19-A3 GNP had greater signal intensity (827 - 963 a.u.) compared to 1% FITC-C19-A3 glucose and mannose C2 GNPs (279 - 384 a.u.). HPLC confirmed that all free FITC was removed by Vivaspin from all the GNP preparations. 1% FITC-labelled C19-A3 peptide glucose C2 and mannose C2 GNP (NP205, NP206), had 1.4 -1.7 FITC-C19-A3 bound per AuNP while those conjugated with 3% FITC-labelled C19-A3 (NP207, NP208) had 5 FITC-C19-A3 bound/GNP. Incorporation efficiency of FITC-labelled C19-A3 peptide was 31-48% and the proportion of GNPs carrying at least 1 peptide/GNP was 95% - 98%. The GNPs all had a negative zeta potential (-55 to -58 mV). These GNPs were sent to partner 1 CU for testing both in vitro and using human excised skin.

In Period 3, a medium-scale reactor batch of 3% C19-A3 peptide conjugated GNP with 5% glucose C2 (67/109/1) was electrostatically bound to human recombinant IL-10. rIL-10-C19-A3 GNPs were suspended and stored in Na borate buffer (pH 8.5 0.05 M). Human rIL-10 was incorporated with 100% efficiency into 3% C19-A3-GNP (213 µg rIL-10/ml) and 3% GAD-GNP (255 µg rIL-10/ml), with no rIL-10 GNP remaining in the supernatants. DLS size analysis found that rIl-10-C19A3-GNP (NP374) were 4.4 nm and rIL-10-GAD-GNP (NP377) were 8.7 nm. The zeta potential of rIL-10-C19A3-GNP (NP374) and rIL-10-GAD-GNP (NP377) were both less negatively charged (-9.6 and -18.2 mV, respectively) compared to peptide-GNP, without IL-10. The rIL-10-peptide GNPs, C19A3 peptide-GNP (67/109/1), GAD-GNP(NP235) and control GNPs (NP302) were sent to partner LUMC for further in vitro studies.

Respective contribution of the partners:
Partner 3 MID has designed, manufactured and quality controlled over 400 GNP constructs to support the experiments in the labs of CU, LUMC and INSERM.

Task 2.b: (M6-24) Testing of peptide-IL10 NPs for uptake by human DCs
Partner responsible: Partner 6 LUMC
Status: Completed
Work description and progress:
1. Effect of Gold Nanoparticles (GNP) on human dendritic cell phenotype and maturation
Uptake of C19-A3 GNP have been shown not to induce spontaneous pro-inflammatory differentiation of immature dendritic cells in vitro, nor did they prevent subsequent maturation of DCs with LPS treatment. LPS induced up-regulation of MHC-class II and co-stimulatory molecules to a same level in C19-A3-GNP treated as in non-GNP treated mature DCs. C19-A3-GNP-treated DCs showed reduced production of cytokines, indicating that treatment with C19-A3-GNPs changed the maturation capacity of iDCs.

2. Toxicity of GNPs to human DCs
The uptake of GNPs did not cause toxicity of DCs up to a concentration equivalent to 50 μM of gold. Microscopic evaluation indicated GNP uptake by DCs, which was confirmed by experiments using fluorescently-labelled GNPs. With regards to cytotoxicity of GNPs, we performed an experiment in which GNPs were added to the DC culture starting from day 0. This experiment revealed that exposure to C19-A3-GNPs from the monocyte stage did not reduce DC yield, nor change DC phenotype. Interestingly, like in the case of GNP-treatment of immature DCs, the capacity to produce cytokines after LPS was reduced if DCs were treated with GNPs during the culture period.

3. Presentation by human DCs of GNP coupled proinsulin C19-A3 peptide to T cells
Next, we have addressed the presentation of the GNP-bound peptide C19-A3 taken up by DCs. Data showed antigen-specific stimulation of C19-A3-specific T regulator cells (Tregs) upon incubation with C19-A3 GNP-treated DCs. In period 3, we further investigated the consequences of this presentation for the T cell differentiation. The GNP-C19-A3-treated DCs were incubated with naïve T cells after which, the phenotype and functional properties of activated T cells were investigated. Compared to T cells stimulated with control DCs (not treated with GNP-C19-A3), T cells stimulated with GNP-treated DCs showed enlarged IL-10 production and ability to inhibit proliferation of CFSE-labelled responder T cells in a suppression assay.

4. Effect of GNP-coupled IL-10 on human DC phenotype and function
Incubation of iDCs with IL-10 coupled on the GNPs did not yield significant effect on the immature DC phenotype, compared with peptide-only-carrying GNPs. However, IL-10-GNPs did change the ability of DCs to up regulate HLA-DR and activating molecules CD80 and CD25, upon stimulation with LPS. Resulting LPS-treated DCs incubated with soluble IL-10 or IL-10 coupled to GNPs had an immature phenotype, and also reduced pro-inflammatory cytokine production (IL-12 and TNF). These data suggest significant change of antigen-presenting capacity of DCs treated with peptide-GNPs coupled to IL-10. Of note, IL-10 effect was similar when added as soluble protein or coupled to GNPs. In addition, DCs treated with soluble IL-10 or peptide-IL-10-GNPs (final conc. IL-10 was 100¬g/ml), showed significantly higher release of IL-10 in the culture. To discriminate whether moDCs retrieve this IL-10 from GNPs or de novo produce this cytokine, we isolated RNA from DCs treated with IL-10 alone, peptide-GNPs alone, IL-10-GNPs or medium control, synthesized cDNA and tested cytokine mRNA using RT-PCR (IL-6, TNFa and IL-10). Relative expression of IL-10 mRNA increases upon stimulation with IL-10, independent of whether it was added as free cytokine or coupled to the nanoparticle. This finding is confirmed in two independent DC cultures. This increase in local IL-10 concentration is expected to enhance regulatory T cell induction on subsequent interaction between the DCs and T cells.

5. Comparison of properties of glucose and mannose linked GNPs
An additional point addressed as part of Task 2.b and 2.c was the comparison of glucose-containing GNPs with mannose-containing GNPs. The reason for this comparison was to investigate a possible preferential uptake of mannose-containing GNP by DCs, because dendritic cells use macro-pinocytosis and the mannose receptor to concentrate macromolecules in the intracellular MHC compartment. However, we did not detect significant differences in several parameters such as DC differentiation, maturation, survival, GNP uptake and peptide presentation, or cytokine production.

6. Presentation by human DCs of peptide+IL-10 coupled GNPs to proinflammatory T cells
The efficacy of peptide presentation by GNPs to pro-inflammatory T cells was tested using human glutamic acid decarboxylase (GAD) peptide (339-352) specific T cell clones (generated by LUMC). These cells were used to test the possible effects of IL-10 on specific immune stimulation of peptide specific regulator T cells. We do have available proinsulin peptide C19-A3 specific proinflammatory T cell clones but these are already of regulatory phenotype. Hence in this experiment, Th1 – type T cells specific for a different diabetes-linked peptide epitope – GAD 339-352 were used. To that end, 1 % GAD peptide - GNPs with either 5% glucose or 5% mannose were prepared. Human rIL-10 was also incorporated into a portion of the GAD-GNP preparations reaching 100% efficiency of incorporation of IL-10 at an approximate ratio of 1:2, IL-10 to GAD peptide. Monocyte derived dendritic cells (moDCs) of HLA type DR 17,4 were used as stimulator cells and a CD4+ T-cell clone specific for GAD339-252 in DR17 were the effector cells. After incubation of moDCs with the GNP constructs and subsequent maturation induced by LPS, DCs were collected, thoroughly washed and incubated with CFSE-labelled GAD-specific T cells. Cell proliferation was determined after 3-day co-culture, demonstrating strongly decreased proliferation of GAD-specific T cells stimulated with DCs incubated with IL-10-coupled GNPs. In period 3, we addressed the consequences of IL-10 coupling to GNPs for the naïve T cell differentiation after two weeks of co-culture. Resulting T cells were far more potent at IL-10 production and inhibition in the suppression assay as compared to T cells stimulated with control DCs (without GNP) or DCs treated with GNP-C19-A3 (without IL-10). The IL-10 effect was dose dependent and T cells generated using a high GNP-C19-A3-IL-10 dose (100ug/ml) showed a strong antigen-independent bystander inhibition.

Respective contribution of the partners:
Partner 3 MID: supplied the conjugated nanoparticles for the studies and prepared the respective nanoparticle constructs for the studies performed at LUMC.
Partner 5 LUMC: tested peptide-IL10 AuNPs for uptake by human DCs and effects on DC phenotype.

Task 2.c: (M6-18) Modification of NPs to further enhance uptake
Partner responsible: Partner 3 MIDATECH
Status: Completed
Work description and progress:
The working hypothesis underlying this specific task was that dendritic cells and other antigen-presenting cells are using macro-pinocytosis and the mannose receptor to bring and concentrate preferentially foreign macromolecules into the cellular compartment where proteins are processed and peptides associated with the MHC class proteins for subsequent presentation on the cell surface (Sallusto, F. et al., J. Exp. Med., 182: 389-400; 1995). Hence, the task aimed to modify GNPs to further enhance uptake. The RP1 document summarises the method development undertaken to optimize the incorporation of C19-A3 peptide into GNPs using a water-based GNP synthesis method and the co-coupling of rIL-10 to C19-A3 conjugated GNP with 5% glucose C2 or 5% mannose C2. Fluorescently labelled C19-A3 and GAD peptide conjugated GNPs with 5% glucose C2 or mannose C2 ligand were also synthesized and sent to partner LUMC for in vitro uptake experiments using human monocyte-derived DCs. No or almost no differences in bio-potency (cell uptake, antigen presentation, induction of T cell proliferation) were observed between glucose C2 and mannose C2-containing peptide-GNP. It is possible that the mannose residues on a short C2 linker are not sufficiently exposed on the gold nanoparticle surface to be picked up by the mannose receptors on dendritic cells. During Period 3, glucose C11 and mannose C11 were custom synthesised by Selvita to investigate whether mannose residues on a longer C11 linker could improve binding of peptide-GNP to mannose receptors on dendritic cells. GAD monomeric peptide (3%) was covalently conjugated to glucose C11 (5% or 10%) (NP397, NP398) or mannose C11 linker (5% or 10%) (NP399, NP400) using the water-based synthesis method. HPLC analysis confirmed that the GAD-peptide GNP with glucose C11 or mannose C11 had ~2-3 peptides per GNP. The DLS nanoparticle size ranged from ~3-6 nm while the zeta potential ranged from -34 to -48 mV, confirming that the nanoparticles were stable in water and well dispersed. These particles were shipped to partner LUMC for in vitro experiments to determine the ability of the mannose C11 long linker on peptide-GNP to enhance monocyte derived dendritic cell uptake through targeting of the mannose receptor. However, no significant difference in bio-potency (cell uptake, antigen presentation, induction of T cell proliferation) was observed between the 5% and 10% glucose C11 and mannose C11-containing peptide-GNP.

CONCLUSION: We have generated peptide and IL-10 co-coupled GNPs and shown that these are not toxic or proinflammatory to human DCs, are taken up readily by DCs and result in efficient presentation of antigen to T cells. In addition, these dual cargo GNPs result in the release of biologically active IL-10 into the milieu around the DCs at concentrations able to suppress the proliferation of proinflammatory T cells and reduce maturation in response to LPS. Hence, these GNPs have many characteristics that would make them appropriate for use in tolerance induction in vivo.
Because partner Midatech (MID) has quite a lot of working experience with and safety data for glucose-coated GNP, but much less data for mannose-coated GNP, and the fact that the glucose constructs showed the anticipated features of potentially immunotherapeutic peptide delivery nanoparticles, the consortium decided to select as clinical candidate C19-A3 peptide GNPs with 5% glucose C2.

Respective contribution of the partners:
Partner 6 LUMC: performed cell culture tests on functionality of C19-A3 or GAD peptide conjugated GNPs with either glucose or mannose C2 or C11.
Partner 3 MID: prepared the peptide-conjugated glucose and mannose C11 nanoparticle constructs for further evaluation of nanoparticle targeting of the mannose receptor on dendritic cells.

Task 2.d: (M1-24) Formulation of NPs for microneedle delivery
Partners responsible: Partner 4 NPSS, Partner 1 CU, Partner 6 LUMC
Status: Completed
Work description and progress:
This task aims to confirm that the NPs and NP formulations are appropriate for microneedle delivery and remain stable and functional following microneedle injection. The formulations tested to date comprise NPs in sterile water. Transmission electron microscopy (TEM) and elemental analysis has confirmed that the GNPs are spherical and 1-5 nm in diameter. The zeta potentials (approximation of surface charge in water) of NP21, NP22 and NP23 are -27.7mV -50.7mV and -45.4mV respectively. Partner NPSS has supplied microneedle arrays of 450 µm and 600 µm microneedle length for testing. Single microneedles will not be used.
The diameter and zeta potential of NP formulation of various C19-A3 peptide concentrations and including either glucose (G) or mannose (M) sugars (to potentially confer efficient/preferential uptake by relevant cells) were determined. Transmission electron microscopy (TEM) indicated that suspensions are relatively homogenous and particle size is 2-4 nm for all the GNP preparations (1%, 3% and 10%). The identity of the sugar moiety in the formulation (glucose or mannose C2) had little observable impact. TEM analysis also indicates that the concentration of the storage buffer (0.01M versus 0.05M borate buffer) did not have a notable effect on the particle size. The particle size was also evaluated using dynamic laser scattering with results supporting TEM observations. The diameter of NPs was determined before and after injection through the MicronJet 600 µm microneedles, with the results showing comparable particle sizes. Accompanying studies were also performed to measure the absorbance of the gold core, as a correlate of GNP concentration, following microneedle injection. The results indicate that gold concentration does not change following injection of the formulation through microneedles and therefore there is no loss of the formulation during the injection process.
Changes to the surface properties of C19-A3 GNP following MicronJet 600 µm microneedles injection was determined be retesting zeta potential (ZP) of formulations after injection. The ZP of complexes containing 3% and 5% C19-A3 peptide was between -40mV and -60mV and remained in this range following injection through the microneedle device. This suggests that these complexes are physically stable to the injection shear forces. The zeta potential of particles containing the highest concentration of peptide (10%) however appeared to be altered following injection via microneedles. Further analysis of zeta potential distribution profiles show that microneedle injection has a notable impact on the previously unimodal surface charge properties of the formulation containing 10% of peptide.
Particle size and zeta potential measurements of GNPs containing C19-A3 peptide and IL-10 following microneedle injection has also been determined.
Distribution and cellular uptake of GNPs in human skin following microneedle injection in vitro and ex vivo has also been investigated under other tasks.
The efficacy of peptide presentation by GNPs to pro-inflammatory T cells following microneedle injection was tested in vitro using human moDCs and GAD-specific T cell clone (as described in 2.b.6). GAD peptide-GNPs were added into the moDC culture through the microneedle and after an overnight incubation and subsequent maturation, DCs were collected, thoroughly washed and incubated with CFSE-labelled GAD-specific T cells. Cell proliferation was similar in cultures with DCs irrespective of whether the GNPs were added by pipetting or injected through a microneedle.

Respective contribution of the partners:
Partner 3 MID: synthesised C19-A3-GNPs, GAD peptide-GNP and C19-A3-IL-10-GNPs.
Partner 1 CU: performed TEM on model GNPs and glucose and mannose GNP conjugated to C19-A3 peptide, particle size and zeta potential measurements on C19-A3 GNPs and C19-A3-IL-10-GNPs pre- and post-microneedle injection.
Partner 4 NPSS: supplied microneedles of 450 µm and 600 µm length.
Partner 6 LUMC: performed cell culture tests on functionality.

Task 2.e: (M6-24) Delivery of NPs to human skin via microneedles
Partners responsible: Partner 4 NPSS, Partner 1 CU
Status: Completed
Work description and progress:
This task evaluates the capability of the microneedle device to inject NP formulations into human skin, determines the distribution of NPs in human skin following microneedle injection and investigates whether NPs are taken into human skin cells.
600 μm and 450 μm MicronJet needles have been evaluated as a means to deposit nanoparticles in human skin, at both the tissue and cellular level. A range of model NP formulations and gold nanoparticle formulations have been used in this studies, the identity of which is dictated by the timing of the studies with respect to the development of the formulation.
Initial studies were performed using suspensions of model particles to determine the immediate deposition of a colloidal formulation in the skin following injection by a MicronJet device. Optical coherence tomography (OCT) images of iron oxide particles that have been injected into ex vivo human skin, show gross deposition of a MicronJet injected particulate formulation, which was distributed primarily in the reticular dermis.This observation was further confirmed by injection of 30nm diameter fluorescent nanoparticles into human skin using both 450 μm and 600 μm microneedles, sectioning the skin and observing under fluorescence microscopy. Increased injection volume (10-100 μl) resulted in a more extensive area of immediate deposition, but in all cases nanoparticle diffusion was primarily restricted to the reticular dermis. There was limited evidence of model nanoparticles in the papillary dermis and viable epidermis. These model particulate formulation experiments used nanoparticles of 30nm diameter, whereas the GNPs used in this project are 2-4nm in diameter (more akin to a solution) and therefore may diffuse very differently in the skin tissue following injection. For visualisation of Midatech gold nanoparticles a silver staining technique was used. Brightfield microscopy images initially indicated that the principal location of MicronJet injected gold nanoparticles (NP23: gold NPs with 90%GSH and 10% Glc), immediately after injection, is the papillary and reticular dermis, with the basement membrane of the epidermis acting as a potential barrier to epidermal distribution. Analysis of skin sections 4 hours after injection, show reduced signal intensity from the dermis, indicating that the NPs can diffuse freely in the dermal tissue during this time. Closer examination of further samples revealed that gold NPs (NP23) were however detectable in the epidermal layer of the skin. Light microscopy suggests that particles distribute rapidly into some epidermal cells and accumulate both within and in close proximity to certain cells. This provides evidence that MicronJet 450 μm and 600 μm microneedles can facilitate delivery of gold NPs to both the epidermal and dermal regions of human skin. Further, tissue analysis by transmission electron microscopy (TEM) showed the presence of gold nanoparticles in epidermal cells that had morphologies indicative of keratinocytes.
These studies were repeated using Au NPs complexes that are covalently linked to the C19A3 peptide. It was found that 4 h after 50 μl injection the GNPs were distributed extensively in the papillary and reticular dermis. However, there was also evidence of gold NPs in the viable epidermis, both surrounding the microneedle injection site and in other areas of the skin. TEM analysis supported these observations, with gold NPs accumulating at the dermal-epidermal junction. However, it was also possible to visualise NPs within dermal cells. Although a greater proportion of delivered dose, which correlates to 10 μg of C19A3 peptide in 50 μl of injection volume (338 μg/μl with respect to the gold), is deposited in the dermis, examination of the epidermal cells indicated that gold NPs were also present in this highly cellular layer. Importantly, intracellular examination of epidermal cells by TEM identified GNPs within cytoplasm of cells containing features indicative of Birbeck granules; a Langerhans cell histological marker. This provides evidence that C19A3 Au NPs can be delivered to epidermal Langerhans cells using the 600 μm MicronJet device.
Larger commercial NPs (50 nm and 100nm in diameter) were injected to confirm that the size of the NPs plays an important role in their distribution and cellular uptake. Analysis of skin sections showed that both 50 nm and 100nm Au NPs were distributed mainly in dermis near collagen and elastic fibers.
Based on the light microscopy data showing epidermal delivery of the GNPs, we also performed flow cytometric studies looking at the nanoparticle uptake in human keratinocyte cell lines (both HaCaT and primary cells) and human skin-derived cell suspensions.
Initial in vitro experiments using the fluorescently-labelled-GNPs [Aminolinker-FITC (NP19), Aminolinker-Dylight 649 (NP20) and GSH-FITC (NP25)] showed that the GNPs are internalised by immortalized keratinocyte cells in culture and the fluorophore is released, as measured by flow cytometry and fluorescent microscopy. The GNPs also remain capable of internalisation into cells after microneedle injection. Exposure of primary skin keratinocytes, taken from freshly excised human skin, to NP20/NP25 demonstrated cellular internalization of GNPs as assessed by high-resolution fluorescent microscopy. Cellular internalization occurs very rapidly post exposure with FITC signal intensification 10 min and then 30 min post-exposure. Cellular uptake at 4°C suggests that GNP uptake is a passive process.
In later studies the NPs were conjugated to C19A3 peptide, which carried a fluorescent marker. In these studies NPs loaded with different amount of peptide (either 1% or 3% peptide) were exposed to both HaCaT cells and primary keratinocytes. Flow cytometry analysis showed detectable levels of fluorescence present in the majority of cells treated with the 3% peptide NPs (90% for HaCaT, 75% for primary keratinocytes) after 4 hours of exposure; as opposed to around 30% of cells exposed to 1% peptide. All the cells (both types) had detectable levels of the 3% peptide NPs after 24 hours exposure as opposed to 90% (HaCaT) and 75% (primary keratinocytes) of cells with the 1% of peptide NPs. A comparison of various types of NPs showed that there was no difference in cellular uptake of NPs containing glucose or mannose. Cellular uptake of NPs was also explored at the intracellular level using confocal microscopy, which showed the NPs were efficiently internalised into primary keratinocytes and present throughout the cell cytoplasm.
The uptake of fluorescently labelled C19-A3 GNPs in relevant skin cell types was further explored by incubating the NPs with suspensions of cells isolated from the epidermis and dermis of freshly excised ex vivo human skin and determining cell uptake by flow cytometry and confocal microscopy. Epidermal cells were isolated using either a trypsinisation or walkout protocol; dermal cells were isolated by walkout alone. NP uptake in epidermal cell suspensions was highly efficient with over 70% of cells displaying FITC labelling within 30 minutes of exposure (over 90% after 24 hours exposure) irrespective of the method used to isolate the cells. NP uptake by dermal cells was more variable and less efficient; with 30% of cells taken up the C19-A3 GNPs after 30 min and 60% after 24 hours of incubation. Confocal microscopy data confirmed that the NPs were internalised within the cells.
The next study was performed to determine whether the C19-A3 loaded NPs were taken into cells following microneedle injection into ex vivo human skin. In these studies the MicronJet 600 μm devices were used to inject FITC-labelled C19-A3 NPs into freshly excised human skin explants prior to removal of the cells and flow cytometry analysis. The cells were isolated from the skin 4 hours after injection. At an injection concentration of 100 μg/ml (with respect to gold) 1% peptide loaded NPs were taken up into 6% of epidermal cells and 3% peptide loaded NPs were taken up into 15% of epidermal cells. Further studies were performed using the higher concentration of C19-A3 (3% peptide) GNPs injected into excised human skin using 600 μm MicronJet microneedles. Initially the concentration of NPs administered was increased to 300 μg/ml (with respect to gold). This led to an increase in epidermal cell uptake to a maximum of 60% of total cells. At this concentration, approximately 60% of dermal cells also took up the C19-A3 GNPs. Uptake could be further enhanced by increasing the GNP dose to 10 μg of C19-A3 peptide in 50 μl of injection volume (equating to approximately 440 μg/ml with respect to gold). At this dose (equating to a therapeutically relevant dose of peptide for antigen specific immunotherapy) uptake was observed in up to 80% of epidermal skin cells, although dermal cell uptake did not increase beyond 60%.
Further experiments were carried out to determine the internalisation of fluorescently labelled GNPs in different epidermal and dermal cell subsets. The FITC-labelled GSH GNPs (NP25) were initially injected into skin using microneedles and at 24h after injection, the epidermis was separated from the dermis and trypsinised to obtain a single cell suspension. Cells were obtained from the dermis following a further 48 h of tissue migration. The cells were stained and fixed for FACS, to analyse cellular uptake. Although the phenotype of the DCs was well defined, the cells exhibited minimal levels of FITC fluorescence. Similar experiments were performed with fluorescently labelled C19-A3 GNPs (as described above). Staining of the skin cells with HLADR/CD207 (for epidermal cells) and HLADR/CD11c (for dermal cells) was used to confirm the relative efficiency of uptake of injected GNPs by Langerhans cells and dermal dendritic cells when compared against other cells in the epidermis and dermis. It was shown that 85% of Langerhans cells have taken up the C19-A3 GNPs when only 27% of keratinocytes were positive for FITC after microneedle injection. Further experiments showed that increasing the amount of C19-A3 GNPs led to an increase in keratinocyte uptake up to 34% and a higher percentage of FITC-positive Langerhans cells, which reached 94%.
Similarly, against a background of approx. 45% of total dermal cells being FITC positive after fluorescent C19-A3 GNP injection, approx. 30% of dermal dendritic cells had taken up the NPs. It is important to note, that in the dermal skin samples we often observe 20-30% of the cells as staining positive for dendritic cell phenotype – this is expected as dermal dendritic cells will be preferentially collected from a walkout protocol.
We have also investigated the presentation of NP-bound peptide taken up by skin-resident dendritic cells and presented to antigen-specific T-cells. In these studies a flu peptide was used as a surrogate. Hemagglutinin (306-318) Influenza A virus peptide (FP1) was injected by MicronJet 600 µm needles to viable, freshly excised, human breast skin. After incubation, a single cell suspension of epidermal cells was obtained and co-cultured with cloned T-cells (HLADR1 restricted, FP1 specific). The cloned T-cells responded with a higher IFNγ stimulation index when cell suspensions from HLADR1 positive skin donors were used (5.2±3.1) in comparison to skin cells from HLADR1 negative donors (1.0±0.1). This demonstrates that the NPs can deliver the peptide to the appropriate immune-competent cells in the skin to mediate antigen-specific presentation.

We have preliminary evidence that IL-10 bound to GNPs is functional on injection into human skin as seen by a reduction in the ability of extracted DCs from the skin to induce a mixed lymphocyte reaction.

Respective contribution of the partners:
Partner 3 MID: has supplied model GNPs.
Partner 1 CU: has performed in vitro cell delivery and ex vivo skin delivery experiments and analysis.
Partner 4 NPSS: have supplied microneedles of 450 µm and 600 µm length and training on use.

Task 2.f: (M6-24) Functionality of microneedle-delivered NPs
Partners responsible: Partner 5 LUMC, Partner 1 CU
Status: Completed
Work description and progress:
IL-10 conjugated GNPs injected via microneedles into ex vivo human and skin were shown to reduce the mixed epidermal cell-lymphocyte response using cells extracted after 12-16 hours from the treated skin. Reduction in the interferon gamma (IFNγ) response with preservation of the IL-10 response was seen, at the higher dose of IL-10-GNPs (1000 ng/ml equivalent), comparable to IL-10 injected alone.
GNPs can promote a pro-tolerogenic environment. In Period 3, further experiments were conducted to detect antigen presentation following skin injection with T cell clones (LUMC, CU) and with the later GNP production technique, which more efficiently incorporates IL-10.
GAD-GNP solution (cfr. Task 2.b/5) was applied to the culture of human dendritic cells (DCs) directly to the culture (through a pipet-tip) or applied through microneedles. Upon 3 hours or overnight culture with GNP, incubated DCs demonstrated equal capability to take up, process and present GNP-derived peptides to T cells, irrespective of the way they were applied to the culture. Also, DCs treated with GNP solution delivered through a pipet-tip or through a microneedle had the same phenotype and cytokine production pattern.
The results showed that the moDCs take up the GAD-GNP, process and present the GAD peptide as efficiently as was shown earlier for the C19-A3 peptide. The GAD peptide was presented equally well upon delivery with either glucose- or mannose-containing GNP, and both constructs induced T cell proliferation equally well and to a similar magnitude as free peptide. Passage of the GNP solution through microneedles had no influence on the bio-potency of the GNP constructs, which demonstrated that in all likelihood the choice of GNP administration would not interfere with the anticipated immune stimulation in the patients.

Respective contribution of the partners:
Partner 3 MID: supplied IL-10 conjugated GNPs.
Partner 1 CU: performed skin organ bath studies.

WP3: Generate System 1 and System 2 (DNA based) for use in preclinical model
WP leader: NanoPass
Duration from Month 1 to Month 24
Partners involved: Partners 1 CU, 3 MID, 4 NPSS, 5 LUMC, 7 INSERM

Summary: All the experiments were completed in Periods 1 and 2. After 36 months of activities NPSS, MID, CU, and INSERM have achieved the following.
- Confirmation that the shorter (0.45 mm) microneedles are suitable for use in mice and can deliver volumes of up to 50 µl.
- Demonstration that this technique delivers fluorescent GNP into the dermis.
- Distribution of gold nanoparticles in murine skin was tested, showing that the majority of injected particles are distributed within the dermis region of the skin.
- Peptide GNP suitable for use in murine studies have been developed and shown to deliver peptide to murine T cells effectively in vitro and in vivo.
- Peptide coupled to GNP has been shown to disseminate to lymphoid tissues more rapidly than peptide alone when injected intradermally.
- Test plasmid DNA has been coupled to GNP.
- Transfection of NPs: DNA complexes in human cell lines (HaCaT and primary keratinocyte) show that GNP do not support pDNA expression in vitro.
- Injection of NP: DNA complexes into ex vivo human skin resulted in accumulation of aggregates in the tissue.
- Plasmid injected using the MicronJet device without NPs resulted in successful expression, which occurred mostly in the epidermis region of the skin, however reduced, but significant levels of gene expression were observed also in dermis site of the skin.
- Successful transfection was observed for different plasmids tested (pCMVB and pGFP-N1).
- MicronJet (600 μm) was shown to be an effective delivery system for gene delivery to human skin.
- Delivery of plasmid alone to in vivo mouse skin results in reduced and more varied plasmid expression compared with human skin.
- Whilst delivery of plasmid alone to in vivo mouse skin varies greatly between different injections and different experiments, successful transfection of plasmid delivered by MicronJet 600 μm was observed in both epidermal and dermal sites of murine skin.
- Successful transfection of plasmid alone was also observed in few cases of plasmid delivered intradermally in mice, however it varied between injections and experiments.

Task 3.a: (M1-18) Development of a shortened microneedle delivery system for murine use
Partners responsible: Partner 4 NPSS, Partner 1 CU
Status: Completed in RP1
The 450nm and 600nm Micronject microneedles systems developed by Nanopass were tested for intradermal injection in mice. True intradermal injection is challenging in mice because the skin is much thinner (250um) than in humans (1000nm). Angled injection was found to be possible in mice with either length of needle. Although not ideal, it was concluded that the construction of a shorter needle would not represent a significant advantage over angled injection.

Task 3.b: (M1-18) Evaluation of an accompanying slow release system for microneedle delivery of NPs
Partners responsible: Partner 4 NPSS, Partner 1 CU
Status: Deferred
Work description and progress:
The development of a slow release delivery system was deferred. Following review of the delivery data in human and mouse skin, it was not clear that this would be required. In addition, at the General Assembly held in Leiden in October 2013, one of the options raised was that the target product profile would be a treatment injected periodically (several times per year) to achieve the required objectives. NPSS is testing the concept of developing a system for self-administration based on a pen injector (which the patients are familiar with) and this may ultimately replace the need for a slow release system.

In order to improve usability and reduce user errors Nanopass developed a new prototype aimed at improving on the current MicronJet 600 device for intradermal injection (called MicronJet G). The MicronJet G uses the extensive database and user input from the MicronJet 600 to examine the required learning curve and failure modes during intradermal injection with the device. The new device prototype aims to reduce the learning curve by introducing a new design with a completely new microneedle-skin interface with integrated intuitive mechanical guides for exact orientation and location.

Respective contribution of the partners:
Partner 4 NPSS: development and evaluation of MicronJet G prototype.

Task 3.c: (M1-12) Preparation of a peptide-IL-10 NP for murine studies
Partners responsible: Partner 3 MID
Status: Completed in RP1
Peptides suitable for murine studies including the BDC 2.5 mimotope, the WE14 chromogranin peptide and the insulin B9-23 have been successfully coupled to GNP at 1%, 3% and 5% peptide concentrations by Partner 3 (MID) using the water-based synthesis method (for WE14 and insulin B9-23) or 75% methanol based method (BDC2.5 mimotope) and a tripeptide linker. GNP carrying ovalbumin (OVA) derived peptide to which partner 7 (Inserm) have transgenic T cells for ovalbumin detection have also been prepared using the water-based method. Efficient conjugation of peptide was confirmed by HPLC (High pressure Liquid Chromatography) and stability in water was confirmed by zeta potential measurements, which were all >30mV. However, GNP conjugated with higher ovalbumin peptide concentration >3% were limited by difficulties in removing unbound peptide from the AuNP preparations. Therefore, 3% peptide-AuNP was chosen as the optimal peptide concentration for electrostatic binding to human rIL-10. Efficient non-covalent coupling of peptide-GNP and GNP without peptide to human recombinant IL-10 (suitable for murine studies) or control equine skeletal muscle has also been achieved. Peptide coupling to mannose and glucose GNP has been performed, to allow experiments in which AuNP uptake by these two sugar moieties can be compared.

Respective contribution of the partners:
Partner 3 MID: Preparation of peptide coupled GNP, coupling to IL-10.

Task 3.d: (M6-M24) Effects of epidermal delivery of NPs on skin DCs
Partner responsible: Partner 7 INSERM
Status: Completed
Work description and progress:
The aim of this task was to study in vivo the efficiency of antigen delivery within the skin using nanoparticles and microneedles (MN). Indeed, we assessed if the antigen bound to the nanoparticles was targeting skin dendritic cells, resulting in antigen presentation by the dendritic cells. To address this question, the model antigen chosen was the ovalbumin (OVA). OVA peptide-coated GNPs were injected with microneedles in the mouse skin and the efficiency of DC targeting was followed through the proliferation of CFSE-labelled OVA specific CD4+ T cells. We showed significant proliferation of naïve T cells suggesting that DC were able to capture the antigen bound to the NPs, to process it and to present it to T cells. To confirm that skin dendritic cells were involved in this process, we compared WT mice with CCR7 KO mice in which DC migration is impaired. The significant reduction of T cell proliferation in the draining lymph nodes of CCR7 KO mice confirmed that after MN injection of peptide-GNPs within the skin, skin dendritic cells were capturing and processing the antigen, migrating to the LN and subsequently activating naïve T cells. Furthermore, we compared mannose and glucose based nanoparticles and showed that both nanoparticles were taken up by skin dendritic cells and permitted antigen presentation to naïve T cells by migratory dendritic cells within the skin draining lymph nodes.
We also compared different doses of peptide coupled to the NPs (1; 3; 5%) and showed that similar T cell proliferation was obtained with the 3 different NPs suggesting that minimal amount of peptide could be used limiting the presence of free peptide.

Respective contribution of the partners:
Partner 3 MID: Supply of peptide coupled GNP.
Partner 4 NPSS: Provision of shortened microneedles.
Partner 7 INSERM: Injection of test antigen into murine skin and detection with CFSE labelled T cells; development of protocols to identify different skin DCs subsets.

Task 3.e: (M1-18) Generation of a DNA construct for expression of beta cell antigen linked to TEEs in DCs (System 2)
Partners responsible: Partner 7 INSERM, Partner 1 CU
Status: Completed
Work description and progress:
The version 1 of System 2 that was originally designed consisted of a minimal Ea promoter, a pre-proinsulin ORF, a GSG2A sequence, an IL-10 ORF and a bGH polyA signal. Following a recent publication showing that the numbers of CpG motifs present in synthesized DNA pieces have to be minimized to avoid their immunogenicity, an Ea-mpreproinsulin-GSG-2A-mIL10 cassette with reduced numbers of CpG motifs has been thus redesigned.
The cassette has been constructed and archived. In view of the need to use cumbersome T cell-based assay to visualize and track the expression of the Ea-mpreproinsulin-GSG-2A-mIL10 cassette in skin cells much of the efforts of Partner 7 have focused on the engineering and validation of two cassettes the expression of which are far more easily tractable – that is via fluorescence instead of T cell-based assay – permitting to define the best delivery protocol in mouse and human skin and thus preparing the ground to the use of the Ea-mpreproinsulin-GSG-2A-mIL10 cassette.

Among several plasmids tested, plasmid pEGFP-N1 turned out to give the highest expression of the green reporter GFP in several cell lines, including DC1, a line of mouse dendritic cells. pEGFP-N1 expresses the GFP under the control of the early immediate CMV promoter and SV40 polyadenylation sequences are appended at the 3’ end to permit the proper termination of the transcription. Plasmid pEGFP1-N1 has been sent to Partner 3 and used to express GFP in human skin explants. Based on successful GFP expression reached by Partner 3, the same plasmid backbone was used to construct a plasmid called pCMC-OVA-2A-EGFP (pCG434-15). Accordingly, the full-length chicken ovalbumin (OVA) sequence (http://www.uniprot.org/uniprot/P01012) was synthesized and appended after the CMV promoter. The OVA sequence was also connected to the GFP sequence via a sequence coding for a self-cleaving 2A peptide sequence. In the process of translation of the OVA-2A-GFP open reading frame, the self-cleaving 2A peptide sequence permits high cleavage efficiency (1 to 1 ratio) between the ovalbumin protein and the GFP protein. Accordingly, using pCMC-OVA-2A-EGFP is a able to express EGFP to testify for the presence of equimolar amounts of OVA. The whole sequence of pCG434-15 is shown below.

pCG434-15 has been transiently transfected into several cell lines and gave the expected green fluorescence, although at levels lower than pEGFP-N1.

Respective contribution of the partners:
Partner 7 INSERM: Developing of system 2 construct.

Task 3.f: (M12-18) Coupling of the System 2 DNA construct to NPs
Partners responsible: Partner 3 MID
Status: Completed
Work description and progress:
Further transfection studies with various plasmids (pGFP and TdTomato) were performed in vitro in human cell lines (HaCaT and Primary keratinocytes). Various types of NPs (containing low or high Zn) were synthesised by partner MID and used at different molar ratios of NP:pDNA. Lower transfection was generally obtained for TdTomato plasmid compared to pGFP. Transfection was effective only when chloroquine was used, suggesting that addition of nanoparticles do not support pDNA expression in human cell lines.
Further experiments with plasmid alone were performed to examine gene expression in skin tissue using ex vivo human skin – see task 3g.

Respective contribution of the partners:
Partner 3 MID: preparation of cDNA-AuNP complexes.
Partner 7 INSERM: preparation of cDNA constructs.

Task 3.g: (M6-M24) Uptake and expression of peptide-IL-10 NPs and System 2 NPs in dendritic cells
Partners responsible: Partner 7 INSERM, Partner 1 CU, Partner 5 LUMC
Status: Completed
Work description and progress:
Peptide-IL-10 NPs
LUMC have confirmed that immature DCs generated from different individuals take up and process NPs. To test antigen presentation, DCs were pre-incubated with NPs, stimulated with LPS and cultured with autoreactive T-cell clone recognizing a peptide coupled to NPs. Co-culture of the T cell clone with DCs pre-incubated with peptide-coupled NPs induced proliferation of effector T cells, in a dose dependent manner. This indicated that iDCs retrieve intact peptides from NPs and present to T cells.
To test the role of IL-10 on the presentation of peptides from NPs, iDCs were incubated with IL-10 coupled on NPs, which did not change the phenotype of immature DC, compared to peptide-only-carrying NPs. However, IL-10-NPs prevented DCs to upregulate HLA-DR and activating co-stimulatory molecules upon stimulation with LPS. Resulting DCs retained immature phenotype and showed reduced pro-inflammatory cytokine production (IL-12 and TNF). These data suggest significant change of antigen-presenting capacity of DCs treated with peptide-NPs coupled to IL-10.
Of note, IL-10 effect was similar when added as soluble protein.

System 2
Initial transfection studies were performed in human keratinocyte cell lines (HaCaT and primary keratinocytes) using various types of Au NPs (low or high Zn) supplied by Midatech and plasmids: pEGFP-N1 and TdTomato. Obtained results indicate that coupling of Au NPs to pDNA does not enhance pDNA expression in cell culture systems.

Experiments in viable human skin explants were performed to determine ex vivo gene expression following delivery of various concentrations of NP:pDNA complex or plasmid alone via MicronJet 600 μm microneedles. Following 24 h incubation, skin was stained with X-gal solution and examined for β-galactosidase expression. The results show that NPs do not support expression of DNA in human skin. Further examination of transverse sections indicates accumulation/agregation of complexes in dermal regions of the skin. Successful expression was however obtained when plasmid was injected alone, i.e. naked pDNA. Interestingly, it was noted that the site of transfection often occurred distant from the injection site. Transverse sections indicate that transfection occurs predominantly in viable epidermis although reduced levels of gene expression have also been observed in the dermis. Successful transfection was evident in at least three independent experiments and occurred for various concentrations of pCMVB (0.5 1 and 5 μg/μl ; injection volume was 50 μl).
Examination of β-galactosidase expression injected via MicronJet 450 μm or 600 μm microneedles into ex vivo human skin resulted in similar level of expression. As a result subsequent studies were performed using the CE-marked clinical ready MicronJet 600 μm needles.
To further confirm successful expression of naked plasmid delivered alone, pEGFP-N1 plasmid was injected using Micronjet 600 μm microneedles. Intravital imaging of human skin 24 hours after MicronJet injection of naked pEGFP-N1 plasmid (50 μl ; 6μg/μl) confirmed observations seen with pCMVB, with increased and widely dispersed fluorescence intensity seen in all four treated skin samples, compared to the negative control. Fluorescence imaging of epidermal sheets isolated from skin samples post-transfection support these observations. Analyses of transverse sections indicate that exogenous gene expression occurs within the viable epidermis. In a further experiment pEGFP-N1 plasmid was delivered into ex vivo human skin via MicronJet microneedles prior to isolation of the epidermal layer, suspension of the cells and analysis by flow cytometry. The results confirmed that, it was possible to observe gene expression in a small number of cells (approximately 1% of cells expressing GFP) enzymatically and mechanically removed from the skin after MicronJet injection.

To assess the potential of MicronJet for successful, efficient and reproducible gene delivery the results were compared with direct intradermal (ID) injection of naked pDNA as an established delivery method. Following ID delivery of pCMVB plasmid positive expression of the exogenous gene was detected in the viable epidermis, however it was not as extensive as that observed following MicronJet injection and was restricted to the site of injection. Further histological evaluation of skin showed that the transfection areas were either associated with the epidermal region that is directly disrupted upon needle entry into the skin, or, if the tip of the needle was positioned in a cellular region, at the tip of the needle. The lower plasmid expression occurred demonstrates an apparent advantage of MicronJet for successful gene delivery in human skin. To explore the mechanism for this improved delivery we have performed optical coherence tomography (OCT) imaging of ex vivo human skin explants during MicronJet injection. The imaging shows, in real time, how the formation of a bleb in the skin results in pockets of micro-disruption in the dermis, and possibly the epidermal-dermal junction. It is possible that this micro-disruption facilitates access of the plasmid formulation to the viable cells in the epidermis. It may even cause transient damage to the cells themselves which would promote plasmid uptake. In further experiments we have shown that MicronJet injection can increase DNA uptake in cells in culture as a result of the force of the injection stream. We currently believe therefore that MicronJet microneedles facilitate gene expression in ex vivo human skin through a combination of tissue and/or cell micro-disruption and hydrodynamic pressure effects.

For results in mouse skin, see WP4.

Respective contribution of the partners:
Partner 1 CU: In vitro and in vivo antigen presentation studies.
Partner 3 MID: preparation of plasmid-GNP complexes.
Partner 5 LUMC: performed the in vitro experiments with human iDC.
Partner 7 INSERM: Contributed to the in vivo experiments.

WP4: Preclinical safety, PK, PD and efficacy of System 1 and System 2
WP leader: CU
Duration from Month 6 to Month 57
Partners involved: Partners 1 CU, 3 MID, 7 INSERM

Summary:
This WP tested the in vivo preclinical safety, pharmacokinetics and pharmacodynamics of:
a) System 1 which consists of single and potentially multiple peptides which are coupled to nanoparticles with or without IL-10.
b) System 2 which consists of plasmid DNA coupled to nanoparticles.

1) For system 1
We used the non-obese diabetic (NOD) mouse model which develops spontaneous diabetes and which is a good model for type 1 diabetes, as well as transgenic mouse systems on the NOD genetic background which facilitated short-term in vitro and in vivo readouts of efficacy of the systems.
The animal model colonies were expanded and we initially showed that the readouts were useful and effective for the purpose and we tested single and multiple system 1 peptides with the addition of nanoparticles, with and without IL-10, using microneedles.
We found that:
a) Single and multiple peptide nanoparticle injections in vivo in the NOD mouse were safe, and no adverse events were seen,
b) Several peptides were investigated for the BDC2.5 TCR transgenic cells - a highly diabetogenic CD4 T cell in the NOD mouse which recognizes a hybrid peptide of chromogranin A and insulin. The identity of the endogenous peptide was not known until 2015, and the earlier experiments were conducted firstly by using a high affinity mimotope stimulatory peptide, and secondly by a peptide initially reported to be the peptide, but in fact was only a portion of the hybrid peptide and of very low affinity.
c) When used in a range of doses and injected intradermally, the peptide nanoparticles, particularly at the higher doses had much wider distribution, and stimulated more proliferative responses from the indicator autoantigen specific CD4 T cells, and this was seen with both higher and lower affinity peptides,
d) At lower peptide doses, there was an increase in potentially regulatory responses as seen by increase in regulatory markers on the CD4 T cells as well as increase in the regulatory cytokine TGF-beta,
e) Short term in vivo experiments were carried out to try to ascertain an appropriate dose with which to perform an in vivo experiment to test for reduction in diabetes, an experiment that takes place over 35 weeks. The dose selected was based on in vitro experiments that indicated some efficacy when 2 doses were administered 3 weeks apart, and showed reduction in proliferation of BDC2.5 T cells injected in vivo and analysed for their proliferation in the draining lymph nodes and pancreatic lymph nodes.
f) Although the microneedle used was trialled in the earlier experiments, the variability seen was too great and therefore for the in vivo experiments, the microneedles, which show good efficacy in humans were not used for later in vivo experiments,
g) The addition of IL-10 on the nanoparticle on the experimental system that was used, was not found to enhance any potentially regulatory effects of the injection of the peptide nanoparticle complexes,
h) The in vivo experiment that was carried out in NOD female mice, 20 mice per group, where the mice at 6 weeks old were injected intradermally with 6mcg of hybrid peptide conjugated to nanoparticle, or 6mcg of free hybrid peptide or left untreated. The incidence of diabetes was however no different between the 3 groups of mice.
i) In testing the mice in the long term in vivo experiment, it was seen that there were residual gold particles in the skin of the mouse. The implications of this are currently not clear and histology is awaited.
j) No clear changes were noted in the endogenous cells in the draining or pancreatic lymph nodes as a result of the injections.
k) We conclude that there are differences in the distribution and pharmacokinetics of peptide-nanoparticles, which have some effects on pathogenic CD4 T cells, but at the dose and dosing regime used, that this is not sufficient to protect against autoimmune diabetes in the NOD mouse.

2) For system 2
System 2 was developed and initially tested in the skin of mice ex-vivo. We determined, as shown in WP3, GNP coupled DNA was not effective at resulting in gene expression and that while there was expression from naked plasmid constructs in human skin, there was very variable expression of the constructs in murine skin. Expression following gene gun delivery of an OVA expressing construct resulted in proliferation of OVA specific T cell clones, however this form of gene delivery was not considered practical for experiments to mimic human delivery. The variable expression with naked DNA in mouse skin would have made it very difficult to obtain a reliable expression in vivo and the approach will be modified in future projects, beyond the scope of the current project.

Task 4.a: (M6-12) Expansion of NOD mouse colony
Partner responsible: Partner 1 CU
Status: Completed in RP1
The colony was expanded to provide sufficient mice at appropriate ages for the duration of the project.

Task 4.b: (M12-24) Safety, pharmacokinetics and pharmacodynamics of single peptide-IL-10/BDC2.5 peptide-IL-10 NP administration in vivo
Partners responsible: Partner 1 CU, partner 3 MID
Status: Completed in RP2
Work description and progress:
BDC2.5/NP preparations were generated and tested in vitro - with 3 different peptides. These were a) a mimotope peptide stimulation BDC2.5 CD4 diabetogenic T cells, b) a peptide WE14, that was previously reported to be a native peptide of BDC2.5 T cells that requires post-translational modification and c) a more recently reported insulin-chromogranin A hybrid peptide. These peptides in single injections were safe in vivo with no adverse reactions, immediately or over a 3 day time course.
The peptide-NPs were compared with peptides and injected concurrently with BDC2.5 indicator cells and tested over 3 days. 2 linkers - mannose and glucose were also tested, and gave comparable results. Proliferation was seen in the draining lymph nodes, non draining lymph nodes and spleen for the peptide-NP complexes but only in the draining lymph node for the free peptide. These results were seen particularly at higher doses of peptide. Similar pharmacokinetics were seen when this was delivered both intradermally as well as through the skin with microneedles.

Similar experiments were carried out with an alternative system using ovalbumin as antigen. In addition, the use of a CCR7 knockout model was used which reduced considerably the wide distribution when peptide-NP was used. This indicated that some of the distribution related to transport by dendritic cells, which are dependent on CCR7.

Respective contribution of the partners:
Partner 1 CU carried out all experiments with the NOD system.
Partner 7 INSERM carried out the experiments with the ovalbumin and CCR7 knockout models.
Partner 3 MID manufactured the peptide nanoparticle complexes for the experiments.

Task 4.c: (M12-30) Safety pharmacokinetics and pharmacodynamics of repeated peptide-IL-10 NP administration in vivo and optimisation of treatment regime
Partners responsible: Partner 1 CU, Partner 3 MID
Status: Completed
Work description and progress:
A number of protocols to administer 2 doses of peptide separated by 7 days, 14 days and 21 days and 7 days later indicator cells were administered and then indicator cells were tested for proliferative responses and cytokine production over 3 days. When the two injections were administered 7 or 14 days apart, there was no inhibitory effect, but some inhibition was seen after 21 days and this was the time interval selected. 3 different peptides were tested – each of which has an effect to stimulate BDC2.5 T cells. The peptides were a) BDC2.5 mimotope – high affinity synthetic peptide b) WE14 – a peptide of chromogranin A that requires post-translational modification for optimal effect and high doses are required to stimulate BDC2.5 cells c) WE14HIP peptide – a hybrid peptide of insulin and chromogranin A recently reported to stimulate BDC2.5 cells, as for the single peptide injections. No adverse effects were seen and the regime was considered to be safe. However, there was no reduction in the proliferation when the strong BDC2.5 mimotope peptide was used and concluded that this stimulation is likely to be too strong. The proliferation with the native WE-14 peptide had a much weaker direct effect on the BDC2.5 T cells. Some initial reduction in the proliferation was seen with both peptide and peptide NP, but the most effect was seen with WE14HIP peptide and this was ultimately selected for further experiments.

Respective contribution of the partners:
Partner 1 CU carried out all experiments with the NOD system.
Partner 3 MID manufactured the peptide nanoparticle complexes for the experiments.

Task 4.d: (M18-30) Safety pharmacokinetics and pharmacodynamics of administration of system 2 (DNA NPs) in vivo
Partners responsible: Partner 1 CU, Partner 3 MID, Partner 7 INSERM
Status: Completed
Work description and progress:
Different NPs that included C2 glucose, pegamine +/- zinc, glutathione and mercapto-succinic acid were proved for binding DNA to the NPs and it was seen that pegamine was the most efficient system. The transfection efficiency was tested in HEK cell line and this was successful although the efficiency could not be improved beyond 20% by increasing DNA concentration. However, bone marrow derived dendritic cells could not be transfected in the same way in vitro.
In vivo, it could be seen that in contast to naked DNA, the DNA+/- NP pegamine was injected into mouse skin, either intradermally using microneedles, with choroquine, that there was no signal, either in skin draining lymph nodes or the skin APCs. Thus, NP-DNA complexes are not efficient at transfecting antigen presenting cells in vitro or in vivo and had no advantages over the non-NP conjugated DNA complexes.
Experiments had been compared with transfer of DNA plasmid into human skin which could be successfully delivered. However, this was not seen with the mouse skin. Accordingly, the NP-plasmid was not been used to deliver beta cell antigen to induce tolerance in mouse models. Further strategies using mRNA delivery will be tested in the future.
Expression following gene gun delivery in mice of an OVA expressing construct resulted in proliferation of OVA specific T cell clones (OT-I, OT-II), however this form of gene delivery was not considered practical or ethically acceptable for experiments to mimic human delivery in mice.

Respective contribution of the partners:
Partner 1 CU provided the mice and also testing in the mouse and human skin.
Partner 3 MID provided the NP complexes.
Partner 7 INSERM tested the DNA complexes in mice.

Task 4.e: (M12-57) Disease prevention with system 1 NPs
Partners responsible: Partner 1 CU, Partner 3 MID
Status: Completed
Work description and progress:
To generate a protocol that would be used for long term in vivo experiment to test for reduction of diabetes, we had tested different doses of peptide NP had been assessed in short term experiments in vivo and the dose which had given short term reduction in proliferation of indicator BDC2.5 T cells was 6μcg. We used the WE14HIP peptide that had given the best efficacy, injected intradermally with a hypodermic syringe, rather than microneedles as these had been shown on multiple injections to be less reliable. For the disease prevention experiment, we used female mice at the pre-diabetic age of 6 weeks and used 20 mice in each group. The experimental group was given 6μcg of WE14HIP peptide NP intradermally and then monthly thereafter for up to 6 injections. There were 2 control groups, one that had free WE14 HIP peptide and the other that was untreated. There was no difference in the incidence of diabetes in any of the groups of mice. Thus, we conclude that at this dose, with this dosing regimen, that WE14 HIP peptide administration was safe but did not induce any delay or prevention of diabetes. Different doses and in particular combining with a second tolerogenic signal on the GNP is likely to be required for tolerance induction.
We noted that there was a residue of gold NP in the skin, but no apparent remaining peptide in the pancreatic lymph nodes when the non-diabetic mice were examined after 35 weeks of age, 7 weeks after the last dose of treatment. Histological results have indicated non-specific immune cell collections.

Respective contribution of the partners:
Partner 1 CU carried out all experiments with the NOD system.
Partner 3 MID manufactured the peptide nanoparticle complexes for the experiments.

Task 4.f: (M24-57) Disease prevention with system 2 NPs
Partners responsible: Partner 1 CU, Partner 3 MID, Partner 7 INSERM
Status: Completed
Work description and progress: Although the use of NP-plasmid for delivery of beta cell antigens to mouse skin was assessed, and plasmid DNA could be delivered and expressed in human skin using microneedles, the expression in mouse skin was highly variable and it had not been possible to reliably test in vitro. Thus, we had not been able to use NP-plasmid for delivery of beta cell antigen to induce tolerance for testing in mouse models. This strategy will be further explored, pursuing delivery by use of mRNA.

WP5: Phase 1A study of System 1 in subjects with diabetes
WP leader: CU
Duration from Month 6 to Month 57
Partners involved: Partners 1 CU, 3 MID, 4 NPSS, 5 LUMC, 6 KCL, 8 LIU

Summary:
Major progress has been made in WP5 despite several challenges, leading to the testing of GNP proinsulin C19-A3 particles for the first time in man. The clinical material was prepared to GMP (MID, KCL), clinical grade microneedles have been provided (NPSS), toxicology in two species has been completed (MID, CU), the clinical protocol, IB and IMPD have been prepared (MID, CU, LIU) and ethical and regulatory approval obtained in the UK and Sweden (CU, LIU). Eleven patients have been screened and one has completed three doses of the GNPs. Five substantial amendments have been approved.
As anticipated by the Independent Advisory Board and to be expected for an early phase study, the translation to a first-in-man clinical trial presented hurdles and challenges. The initial plan was to test both GNP C19-A3 and a dual cargo particle with a tolerogenic cytokine, GNP C19-A3 IL-10.
The extension to follow-up requested by the UK regulatory agency, the revised IMP stability monitoring required, the extensive delays in contracting especially in Sweden and the high screen failure rate at the UK site resulted in delays to completion of the clinical study, although valuable progress has been made. As a result, some trial and immune monitoring data has already been obtained, but the partners CU, MID, LIU, KCL and NPSS have agreed to continue to support and extension to the trial to May 2018, beyond the end date of the EC funded project, to ensure that the maximum data can be obtained.
No significant side-effects were observed or safety signals. However, a local reaction and visible gold particles were present for at least 6 months after dosing. Immunological monitoring of draining lymph nodes and peripheral blood showed an increase in proliferation of Treg and CD8 cells in the blood and an increase in number of memory Treg cells. Skin biopsy from the site of injection after 6 months showed gold particles still present in the dermis.
Taken together this data confirms that we have conducted a first-in-man study of GNPs carrying the pro-insulin peptide C19-A3 administered intradermally via microneedles. The GNPs were generally well tolerated with no safety concerns. The GNPs persist in the skin for 6 months or more after intradermal administration. Further studies are required to determine whether the response to the GNPs promotes immune regulation.

Task 5.a: (M12-30) GMP formulation of system 1 NPs and microneedle delivery system performing relevant toxicology studies
Partner responsible: Partner 1 CU
Status: Complete
Work description and progress:
GMP formulation of GNPs with covalently bound proinsulin C19-A3 peptide and bulk manufacture has been successfully completed. Although a method to make GNP that carried IL-10 (non-covalently) as well as C19-A3 was also developed, but GMP material could not be developed due to the lack of available GMP IL-10 for in vivo use.

GMP GNP C19-A3 was successful aliquoted into single use vials (“fill and finish”) and Qualified Person (QP) released by PharmaKorell (Germany) for use in the clinical trial.

Extensive toxicology studies were performed in two species (rats and dogs) with daily administration up to dose 1000 times those for clinical use with no safety concerns.

Respective contribution of the partners:
Partner 1 CU: Subcontracted to PharmaKorell (Germany) for QP release of GMP material.
Partner 3 MID: Performed the formulation and bulk manufacture of GMP GNP C19-A3 with GMP peptide supplied by Bachem (Switzerland); “Fill and Finish” was performed under subcontract from MID by Baccinex (Switzerland). Toxicology was performed under subcontract from MID by MediTox SRO (Czech Republic).

Task 5.b: (M24-45) Obtain regulatory approvals for phase 1 study
Partner responsible: Partner 1 CU
Status: Complete
Work description and progress:
Advice from a regulatory consultant David Fairlamb (Protherax) was obtained and a meeting held with the UK regulatory agency (MHRA) in December 2014. The initial plan was to test both GNP C19-A3 and a dual cargo particle with a tolerogenic cytokine, GNP C19-A3 IL-10. The regulatory consultant advised that the due to regulatory approval issues, the protocols for the two entities should be separated. This proved wise, as although MID successfully manufactured non-GMP GNP C19-A3 IL-10, it became apparent that the available source of GMP IL-10 was “not for clinical use”. Hence this would require us to perform our own toxicology on this material. Toxicology is more complex for genetically engineered material, we therefore performed an extensive search for other sources. This identified Provepep (France) who were conducted a chemical synthesis of IL-10 for another customer and agreed to provide are part of this material to us for toxicology and clinical use. However, their synthesis process generated less material than they had anticipated and it contained two chemical entities (including a likely truncated product) with incomplete dimerization. A repeat synthesis for our purposes only would have been very costly and would not have been completed within the timescale of the project. It was therefore agreed to proceed with planned GNP C19-A3 studies only.
The toxicology plan and clinical protocol was completed using information derived from this meeting. Following completion of toxicology (July 2015) an Investigative Medicinal Product Dossier (IMPD) and Investigators Brochure (IB) were prepared and regulatory submission made to the UK NHS Research Ethics Committee, the UK MHRA and Swedish Ethics Committee and Swedish MPA in November 2015. All bodies recommended amendments which were made and approval from all 4 bodies was obtained in June 2016.

Respective contribution of the partners:
Partner 1 CU: Contracted the regulatory consultant, prepared the clinical protocol and participant information and consent material, contributed to the IB, IMPD and lab testing protocol and made the regulatory submissions in the UK.
Partner 3 MID: Prepared the IB and IMPD, attended the UK MHRA advisory meeting.
Partner 4 NPSS: Advised and provided information on the microneedles system used.
Partner 6 KCL: Prepared the lab testing protocol for immune monitoring.
Partner 8 LIU: Advised on the clinical protocol and lab testing, created the Swedish participant information and consent material, and made the regulatory submissions in Sweden.

Task 5.c: (M36-57) Administration of system 1 NPs in man
Partners responsible: Partner 1 CU, Partner 3 MID, Partner 4 NPSS, Partner 8 LiU
Status: Ongoing
Work description and progress:
Cardiff University agreed to be the Sponsor of the first-in-man clinical trial. Following regulatory submission, the UK MHRA requested further laboratory characterization of the IMP for ongoing monitoring and that an additional 8 weeks be added to the follow-up period for each participant to ensure gold excretion was completed (guided by the toxicology data). These delays necessitated a no-cost extension to the project from August 2016 to May 2017 which was requested and approved.
Since the EE-ASI project was devised and submitted to the European Commission in 2011/2012, the UK MHRA has changed its guidance and recommended that all IMP studies in the UK are performed with the support of a UK registered Clinical Trials Unit (CTU). A CTU therefore had to be identified and subcontracted (Swansea Trials Unit) necessitating an increased cost and change to the subcontracting agreement.
Contracting for the clinical trial also proved more complex than anticipated. Although regulatory approvals were obtained in June 2016, it took until September 2016 to complete contracting, sample transfer and processing and with the Trials Unit and local UK R&D. In Sweden, it became apparent that the Pharmacy (Apoteket) was separately operated to the hospital, and that there was a separate monitoring company (Forum Ostergotland). Furthermore, the monitoring company were not registered to submit perform serious adverse event reporting the EC and hence it was required to subcontract to another organization in Sweden, Karolinska Trial Alliance Support (Stockholm) to provide these services. Negotiating these contracts, and the agreement of which legal jurisdiction applied (English or Swedish) was very protracted taking a year, especially with Apotoket, and the Swedish contracts were eventually all signed in June 2017. The delays in study start, required a review of the stability of the IMP, to provide shelf-life approval to 24 months. This was arranged between MID and the QP, PharmaKorell.

Swansea Trials Unit (STU, UK registered) was contracted as the clinical trials unit. An electronic Case Report Form was prepared tested and signed off by STU. A monitoring plan was prepared for the UK site and agreed by the sponsor. UK contracts were established between the Sponsor (Cardiff University) and the Hospital (Cardiff and Vale Health Board), the QP (PharamaKorell), Midatech, Swansea Trials Unit, Swansea Laboratories, King’s College London and Leiden University. A Data and Safety Monitoring Board (DSMB) was appointed and a Trial Management Group. Sponsor approval was given and local Research and Development Approval in Cardiff and Vale Health Board was issued. Following this, a site initiation visit was conducted and the UK site was opened for recruitment in Sept 2016. A no-cost extension to the project was agreed with the EC from Aug 31st 2017 to May 31st 2017. Contractual issues were significantly more complicated with the Swedish site, requiring separate contracts with the private pharmacy company (Apoteket), the Hospital (University Hospital, Linkoping), the independent Monitoring Company (Forum Ostergotland) and a separate company to submit Serious Adverse Events (Karolinska Trial Alliance Support, Stockholm), these contracts were all completed in June 2017. To extend the shelf-life of the IMP for the ongoing trial the QP (PharmaKorell) required an extended (24 month) stability monitoring programme. This was submitted in detail in July 2017. The Sponsor has requested the Site Initiation Visit for the Swedish site to be updated due to the contractual delays. This will occur in August 2017 and the Swedish site can be opened.
The UK site has screened 11 subjects. Cardiff planned to recruit 4 subjects and has screened 11 with a predicted screen failure rate of around 65%. However 8 of 11 failed on genotype (expected to by around 60% failure) – this is assumed to have occurred by chance. 1 subject failed on c-peptide threshold and hence after discussion within the Trial Management Group and the DSMB, it was agreed to reduce the threshold for c-peptide entry by 50% and a substantial amendment was submitted. 1 subject failed to attend for follow-up screening tests. Screening continues at the Cardiff site and will commence at the Swedish site in August 2017.

5 substantial amendments have been submitted. These covered updated stability and impurity testing, a change to the entry criteria to allow extended safety follow-up, lower urinary c-peptide levels and serum c-peptide testing, a change to the participant information sheet to allow dosing on the inner arm, the use of photograph monitoring of the injection site, permission for an (optional) biopsy and blister study of the injection site. More than 50 trial site teleconferences including 10 Trial Management Group meetings and 2 DSMB meetings have been conducted.

Respective contribution of the partners:
Partner 1 CU: Provided trial sponsorship and trial management, contracted the trials unit, prepared the monitoring plan, drafted all contracts, appointed the DSMB, obtained UK local R&D approval, submitted the substantial amendments, convened the trial site and management meetings, and screened and recruited subjects.
Partner 3 MID: Prepared the stability plan and performed the stability studies, contributed to the trial site and trial Management meetings.
Partner 4 NPSS: Provided the microneedles and training on their use, contributed to the Trial Management Group.
Partner 8 LIU: Took part in the Swedish contract negotiations and all the Trial Site and Trial Management meetings, and submitted amendments within Sweden.

Task 5.d: (M36-57) Immune studies on treated subjects
Partner responsible: Partner 1 CU
Status: Ongoing
Work description and progress:
Data obtained from the first participant who has completed involvement in the study included serial blood and ultrasound guided LN sampling. Lymph node Elispots performed before and 5 days after the first dose showed no detectable IL-10 to the C19-A3 peptide. Lymph node IL-10 and interferon gamma elispots also showed no detectable response to peptide 30 days after the second injection. These results may reflect a lack of sensitivity of this test methodology in this setting. In future participants flow cytometry based studies of T cell activation will be used in place of Elispots. Peripheral blood flow cytometry data has been analysed from 3 times points examining over 50 white cell sub populations. Initial indications suggest an increase in eosinophils and neutrophils, an increase in proliferation of Treg and CD8 cells, and an increase in the number of memory Treg cells. Further data will be obtained from future participants. Stored samples will be sent to LUMC for antigen specific CD8 T cell analysis.
Following dosing, it was noted that the gold material was visible at the injection sites for 6 months or more and that there was a local erythematous reaction (not disturbing to the subject) at each site. This was not anticipated and hence after dermatological advice, a substantial amendment to allow optional biopsy of the injections site was obtained. A biopsy was performed which showed gold particles in the dermis and a local inflammatory response but not a “foreign body reaction”. Skin biopsy from the site of injection after 6 months showed gold particles still present in the dermis. There were also areas of lymphocyte and macrophage (histiocyte) aggregation around blood vessels. Initially immunohistochemistry showed a 1:1 CD4:CD8 ratio of T cells with some B cells. Studies of further patients and more detailed analysis of the skin response are ongoing.

Respective contribution of the partners:
Partner 1 CU: Obtained the blood and ultrasound guided LN samples as well as the skin biopsy and performed the LN Elispots. Performed the skin biopsy analysis.
Partner 5 LUMC: Will analyse stored samples from KCL for CD8 antigen specific T cell responses.
Partner 6 KCL: Performed peripheral blood analyses and provided training for CU and LIU.
Partner 8 LIU: Trained in performing LN and peripheral blood analysis and will analyse local samples.

WP6: Dissemination and exploitation plans
WP leader: IT
Duration from Month 1 to Month 57
Partners involved: Partners 1 CU, 2 IT, 3 MID, 4 NPSS, 5 LUMC, 6 KCL, 7 INSERM, 8 LIU

Summary:
WP6 has successfully guided the exploitation and dissemination strategies of the EE-ASI project.
The exploitation strategy began with the Outline Business Plan which defined the Target Product Profile and the Market Requirements. It was agreed that of the 6 possible markets, the initial focus would be newly-diagnosed adults and children with Type 1 Diabetes (market #1) and a Business Plan Workflow was developed comprising three workstreams: regulatory strategy, IP strategy and second generation products. A regulatory consultant was hired then from the company Protherax in 2014 who advised on dividing the IMP into two separate entities (as sourcing the IL-10 component to clinical grade encountered a bottleneck) leading to scientific review with the UK regulators (December 2014), a successful preclinical toxicology programme (July 2015) and approval of the GNP-peptide product for clinical trial use in June 2016. In parallel, a patent protecting the GNP-peptide entity was prepared and filed in April 2015 and published in October 2016. Further discussions through regular BPC meetings lead to the Exploitation Plan (deliverable 6.5 Sept 2015). The licensing of the C19-A3 peptide from partner KCL to the company UCB required further complex negotiation, and repeated discussions regarding pathways to commercialization were held with UCB in 2015-2016. A final patent funding and revenue sharing arrangement was agreed in July 2017 in which CU will lead on patent prosecution and there will be further discussion on commercialization strategy once all the results of the clinical trial are available in 2018.
The dissemination strategy was discussed by the BPC early in the project. It was agreed that the work was too early phase to recruit a Marketing Consultant or engage publicly with target stakeholders and that detailed discussion and publication in the academic community had to await filing of the patent. However, a series of satellite workshops with key sectors of the academic community were held in association with major international research meetings: Langerhans Cells (LC-2013), Amsterdam; Dendritic Cells (DC-2014), Tours; European Association for the Study of Diabetes (EASD – 2015), Stockholm. All were well attended. Publication of many of the research papers (24 so far, three more to follow) followed patent filing in April 2015. The company Sci-Impact was engaged for “general” wider dissemination to over 35,000 policymakers and other stakeholders via two articles/brochure in November 2016 and June 2017 according to our Dissemination Plan (deliverables 6.2 6.3).

Task 6.a: (M1-12) Set up of a Business Plan Committee (BPC): Marketing requirements document (MRD), IP and business perspectives
Partner responsible: Partner 1 CU
Status: Completed
Work description and progress:
A BPC was established and has had 28 meetings during the period of the project. Work in the BPC to protect IP has led to the filing of the patent ref WO2016162495A1 “Nanoparticle based antigen-specific immunotherapy” published on Oct 13th 2016. In addition, the BPC has produced an Outline Business Plan including a Market Requirements review (Deliverable 6.1) and an advanced Exploitation Plan (deliverable 6.5) – see Task 6d.

Respective contribution of the partners:
Partner 1 CU and 3 MID lead on patent development, provided inventors and subcontracted to patent lawyers (Mewburn Ellis).
Partner 4 NPSS drafted Outline Business Plan.
Partner 1 CU drafted Exploitation Plan.
All partners reviewed all documents for final versions.

Task 6.b: (M1-12) Communication network with stakeholders and patients’ associations established
Partner responsible: Partner 1 CU
Status: Completed
Work description and progress:
It was determined in Period 1 that the project represented too early a stage in product development for detailed contact with patients’ associations. However, academic and policy holder networks were established through our Workshop Series (deliverable 6.4) at relevant international meetings and through the network of stakeholders established by Sci-Impact. The Juvenile Diabetes Research Foundation (international patients’ association) was represented on our Scientific Advisory Board (Olivier Arnaud) and the project coordinator presented each year to the national congress of Diabetes UK (UK patients’ association).

Respective contribution of the partners:
Partner 2 CU and 2 IT developed the communication network and commissioned Sci-Impact.

Task 6.c: (M12-57) Deployment of our dissemination plan
Partners responsible: Partner 1 CU, Partner 2 IT
Status: Completed
Work description and progress:
Three international workshops were held as satellites to major, relevant international meetings: with Langerhans Cells (LC-2013), Amsterdam; Dendritic Cells (DC-2014), Tours; European Association for the Study of Diabetes (EASD – 2015), Stockholm. External participants were invited via the Congress Secretariat although IT managed the registration. Key external speakers were invited to make a highly attractive programme and all workshops were well attended (see deliverable 6.4 for detail). Publication of research papers (5 so far, three more to follow) followed patent filing in April 201. The company Sci-Impact was engaged for “general” wider dissemination to over 35,000 policymakers and other stakeholders via two articles/brochure in November 2016 and June 2017 according to our Dissemination Plan (deliverables 6.2 6.3).

Respective contribution of the partners:
Partner 1 CU wrote the material for Sci-Impact.
Partner 1 CU and Partner 2 IT arranged workshops.
Partner 8 LIU hosted the Stockholm workshop in 2015.

Task 6.d: (M24-36) Exploitation plan ready
Partner responsible: Partner 1 CU
Status: Completed
Work description and progress:
An exploitation plan was drafted and revised within the BPC meetings (deliverable 6.5). After extensive discussion it was agreed that CU and MID will share the patent costs, which will be recouped from any revenue and the residual distributed as follows: CU – 37.5%; MID – 37.5%; NPSS – 5%; LUMC - 5%; INSERM-Marseille – 5%; KCL – 5%; LIU – 5%. More advanced plans for commercialization will be agreed between CU and MID once all the results of the clinical trial complete. The patent will be pursued in Europe and North America.

Respective contribution of the partners:
Partner 1 CU drafted the exploitation plan and drafted the patent and revenue agreement.
All partners reviewed the exploitation plan.

WP1: Governance, management and coordination
WP leader: IT
Duration from Month 1 to Month 57
Partners involved: All

The WP1 is not strictly speaking a scientific and technical WP, but it has been closely related to the other ones, facilitating the whole process of the EE-ASI project.

Work description for whole WP1 at the end of Period 3:
During the last months of the project, CU, IT, the WP leaders and the other EE-ASI partners have achieved the following:
- Maintenance of an effective management framework for the EE-ASI consortium, which ensures the progress of the project towards its planned objectives and adequate reporting;
- Organisation and compilation of the second periodic reports for the Commission using partner-customized templates and simplified guidelines to help all partners to report their individual technical contribution, Form Cs and use of resources into the project’s programme of work; quality control of all financial reports of all partners; training on using the electronic signature on the Participant Portal;
- Maintenance of an efficient interface between the consortium and the European Commission (regular updates and invitation of the Project Officer to the project meetings);
- Support for the organisation of meetings, especially of the 3rd General meeting in Stockholm, the 4th General meeting in Cardiff and the final meeting in Paris; six additional Executive Committee meetings via Conference Calls in December 2015, February, May and October 2016, as well as in January and April 2017;
- Creation and update of the corporate EE-ASI communication tools (website with updates regarding the publications, meetings and newsletters);
- All actions were performed correctly and within the FP7 rules and regulations established by the European Commission and in the Consortium Agreement including financial and legal management and correct distribution and accounting of received funds;
- Each partner and the European Commission were continuously informed about the project status, scientific issues, the work planning (adjustments) and all other issues which were important and relevant to partners in order to obtain maximum transparency and achieve synergy of the cooperation;
- All partners were informed of important and impacting information (regulatory, ethical, business issues) that could influence the outcome of the project, notably via meetings and minutes thereof sent by emails to all partners and maintain on the private section of the EE-ASI website 24/7.
Potential Impact:
The final result for WP2 is the development of an optimal method for preparation of gold nanoparticles (GNPs) coupled to proinsulin peptide antigen and cytokine that can be transferred to a GMP production facility. In addition, WP2 has defined that when injected intradermally into ex-vivo human skin, these GNPs are not disrupted, retain their negative zeta potential and reflux into the epidermis as well as being distributed in the dermis Furthermore, human dendritic cells can present peptide antigen bound to GNPs efficiently without inducing maturation and with reduced proinflammatory cytokine production. The development, in WP3, of a microneedle technique suitable for use in mice, along with the synthesis of GNPs carrying peptide relevant to the murine system, allowed system 1 to be trialled for tolerance in in vivo mouse models in WP4. The successful transfection of human epidermis with plasmid alone (system 2) when injected via microneedles – although unexpectedly not when attached to GNP - will allow this approach to be used to express beta cell antigens in skin in future studies. In WP4, the final results suggest that peptides conjugated to nanoparticles may have benefit in distribution of antigenic peptide for induction of tolerance but that a further tolerogenic adjuvant is likely to be required, with further optimization of dose to delay disease in vivo. We will be carrying out further tests to understand the local effects of the nanoparticles and these results are awaited. In WP5, proinsulin peptide conjugated GNPs have been administered for the first time in man. All immunological monitoring (including ultrasound guided lymph nodes sampling) and dosing of further subjects will be completed by May 2018. We are also currently arranging further immunological studies on the skin biopsy specimens. We expect the data to confirm that the GNPs are stable for up to two years after manufacture and can be administered in man with no systemic safety concerns. Immunological monitoring will generate potential biomarkers for future studies and the need for a “second cargo” on the GNPs of a pro-tolerogenic partner molecule. Pharmacokinetic studies will show the time course of retention of the particles and excretion of the gold in man. In WP6, a patent has been filed and published for a platform technology based on GNPs for the development of antigen specific immunotherapy for type 1 diabetes and other organ-specific autoimmune diseases. Dissemination activities included 3 international workshops, 24 publications and information sent to over 35,000 stakeholders.

In summary, we have developed a method to generate proinsulin peptide conjugated GNPs that have been tested for safety in a preclinical model and in a phase 1A first-in-man clinical trial and protected this platform technology with a published patent. Data from the clinical trial imply that GNPs persist for over six months in the skin, providing opportunities for enduring tolerance induction if inflammation can be minimized. We have confirmed that the product can be made to GMP and has no major adverse reactions. In addition, we have demonstrated that a second cargo can be attached to the GNPs (eg the tolerogenic cytokine IL-10) and that intradermal injection of naked DNA via microneedles is an efficient method of inducing expression in human skin. Taken together, these data will provide the basis of a platform technology based on GNPs for the development of antigen specific immunotherapy for type 1 diabetes and other organ-specific autoimmune diseases. This opens the door to further phase 1 studies and later phase 2 (efficacy studies) with the optimal product, ultimately leading to a novel non-immunosuppressive approach to immunotherapy. For type 1 diabetes, a condition in which over 1 million adults and children in Europe are dependent on insulin injections, the aim is to slow the loss of insulin production from the body after diagnosis (and potentially prior to diagnosis) to improve metabolic control, quality of life and long-term outcomes.
List of Websites:
www.ee-asi.eu

Scientific Coordinator
Dr. Colin Dayan
Professor of Clinical Diabetes and Metabolism and Director of the Institute of Molecular and Experimental Medicine
Diabetes Research Group,
C2 Link,
Institute of Molecular and Experimental Medicine,
Cardiff University School of Medicine,
Heath Park,
Cardiff CF14 4XN,
Wales,
UK
Tel: (+44) 029 20 742182
dayancm@cardiff.ac.uk

Project Manager
Anna Boitard
Inserm Transfert
Biopark
7 rue Watt
75013 Paris
France
Tel: (+33) 1 55 03 01 55
anna.boitard@inserm-transfert.fr

Final Report Summary - EE-ASI (Beta cell preservation via antigen-specific immunotherapy in Type 1 Diabetes: Enhanced Epidermal Antigen Delivery Systems.)

Compartir esta página

Descargar