Skip to main content

Development & Evaluation of a Viable Stent Device for the Treatment of BronchoTracheal Cancer

Final Report Summary - PULMOSTENT (Development & Evaluation of a Viable Stent Device for the Treatment of BronchoTracheal Cancer)

Executive Summary:
Within the project “Pulmostent - Development & Evaluation of a Viable Stent Device for the Treatment of BronchoTracheal Cancer” the consortium has made great strides towards a fully viable endobronchial stent, addressing current and future clinical needs. In order to achieve a fully viable stent the principles of tissue-engineering and stent technologies have been brought together and their synergies have been fully exploited.
A computational finite element model for both laser-cut and hand-braided stent design, in both uncovered and covered configurations, has been developed. This model has been demonstrated to provide predictive capability during the design process and has the potential to guide towards faster prototyping to get to the required mechanical performance. Furthermore, a finite element model of the airway with realistic configuration based on CT, data has been developed. This model can provide an output of stresses and strains in the bronchial tissue during simulation of stent deployment as well as stresses, strains and final geometry of the deployed stents. The model outputs compare well with actual experimental observations from the initial animal studies.
Extensive efforts have been put into developing and testing of different stent platforms that could support the PulmoStent concept. The testing ultimately led to the selection of wire-based stents manufactured using hand braiding process and FE-guided design, manufactured from tube using laser-cutting. Both stent types were made of nitinol. Furthermore, an applicator device was developed in the process for the delivery of both stent types.
In-vitro tests have proven that PLGA-microspheres with a diameter of 50 µm - 100 µm offer the most suitable Gefitinib-controlled release system for the PulmoStent application. Further mechanical and biological testing of PU-based foam and fleece, with and without drug loaded PLGA, have led to an optimised manufacturing process and ultimately to PU-fleece coating from UKAachen-CVE as the final cover for the animal study. The fleece properties, determined by various manufacturing parameters, should allow nutrient transport through the fleece and at the same time avoid cancer re-ingrowth. Therefore, the porosity and thickness was adapted iteratively.”
Instead of seeding cells in-vitro, a device was developed to seed the cells in situ after stent placement. This innovative approach facilitates handling, as the critical steps for cell proliferation, such as stent crimping and loading into the applicator device, are avoided. Furthermore, the cell device will be filed as patent, as it has great potential for other applications such as cell therapy.
One of the most crucial success criteria, the assembly of the parts to a complete PulmoStent has been fully achieved. The proof of principle, which is the second crucial success criteria, is currently evaluated in a large animal study. The initial results look very promising.

Industrial partners have been:
3T TextilTechnologieTransfer GmbH, EPITHELIX SÀRL, Vysera Biomedical Limited

Research partners have been:
Institut für Textiltechnik der RWTH Aachen University, National University of Ireland GALWAY, Uniklinik RWTH AACHEN, University Utrecht
Project Context and Objectives:
Lung cancer is the most common cancer in terms of both incidence and mortality, worldwide. With a median age at diagnosis of 71, lung cancer is mainly affecting the aging population. Airway stenosis is a key problem with significant morbidity and premature death.
At the moment, stenting is the palliative method of choice for patients with bronchotracheal cancer. There are different stent types available: metal stents with an open strut structure, silicone stent which completely cover the airway mucosa and hybrid stent which combine a metal stent backbone with a (silicone) cover. However, none of these has optimal properties. While metal stents provide good mechanical forces and do not cover the mucosa, there is a high risk of tumour ingrowth. Silicone stents prevent tumour ingrowth but have a higher risk of migration and while covering the mucosa, mucus plugging distal to the stent is a major short coming.
The PulmoStent concept can overcome these disadvantages of commonly used stents, by combining a metal stent backbone with a cover which prevents tumour ingrowth. The decisive component here is an additional layer of respiratory epithelial cells on the inner surface of the stent cover which propel the mucus out of the respiratory system and therefore prevent mucus plugging. Hence, the concept can have a major impact on developing biohybrid implants and bringing them closer to a clinical use.
The PulmoStent is a step change beyond the state-of-the-art from a passive to a viable and functional active implant tailored to the patient. The combination of different kinds of biomaterials to a co-scaffold system for the biofunctionalisation of the stent will lead to an improved performance of endobronchial stents and thereby to longer durability. The novel PulmoStent will improve the quality of life and increase the life expectancy of lung cancer patients, because of the local tumour suppression in combination with the reduced mucus retention in the stented area, and herewith the reduced risk of life-threatening pneumonia.
The interdisciplinary consortium of seven partners bring their skills from four European countries in the areas of tissue-engineering, textile-engineering, computational modelling and drug-development in the project. The PulmoStent Project is divided into 9 different work packages (WP), which mirror the production process from the stent development (WP2), the PU coating and biofunctionalisation (WP3), cultivation of the respiratory epithelium as well as the cell lining (WP4) to the PulmoStent Production itself (WP5), followed by the proof of principle in the large animal study (WP6). By an alternative cell lining with mucosa cells the potential of multiple use will be evaluated with regard to a GI-stent (WP7). The objectives of each work package are briefly described in the following sections.

WP1 Modelling and Model Validation
This work package has the objective of developing finite element (FE) models of the stents being developed.
These FE models are specifically to help guide the stent design process, allowing prediction of mechanical performance characteristics, when variables are being changed. This allows progression of the iterative design process, without the time and expense involved in prototyping each variant. In particular the model evaluates forces and deformations during the crimping and deployment process enabling comparisons against design targets and benchmark commercial devices. Chronic outward force during deployment is a key mechanical criterion from a clinical perspective. For laser-cut stents (from NUIG) the FE model readily integrates with the 3D CAD models, while models are also developed for the braided stents (from ITA). In addition to the ‘bare’ stent structures, models are also developed of the PU coating configurations from the different partners (VYS, UKAachen-CVE). Validation of the core models is achieved by comparing experimental data against the output from the models. This validation of the models allows continued use of the models to predict behaviour when design parameters are being varied.
In addition to models to predict fundamental mechanical behaviour of stents and covers, this work package aims to develop a model of in-vivo performance of the devices. This requires novel development of anatomical FE models representing the airway and lung structures. This is built using geometry collected from computational tomography and tissue properties collect from biomechanical testing. Using this model to identify stress and strains at locations in the anatomy could allow for prediction of tissue responses in terms of inflammation and granulation.

WP2 Stent Development
The PulmoStent concept depends on having a suitable underlying stent structure to act as the scaffold to support the epithelialized cover. The objective of this work package is to develop such a suitable stent structure.
Given the many variants of stent design, geometry and manufacturing routes, this work package evaluates two fundamental alternatives; laser-cut (from tube) and wire/filament based structures. Laser-cutting allow the option of many different design patterns and therefore can give very specific tuning of mechanical properties. Wire/filament techniques such as warp-knitting and braiding offer very efficient and less expensive processes but may not provide the same extent of options on mechanical performance. Within this work package, ITA are utilizing their expertise in warp-knitted and braided stents while NUIG are providing their experience with design and modelling of laser-cut stents. All stents are being made from nitinol material as the shape memory and superelasticity characteristics are essential to the crimping, deployment and in-vivo flexibility needed. Design specifications are developed early within this work package to allow development of proposed solution – these design specifications are based on clinician input, animal cadaver testing and mechanical evaluations of commercially available tracheobronchial devices. Extensive prototyping of stents is performed within this work package, with the designs evolving from prototypical samples through to those suitable for implantation in the final animal study. The effect of incorporating different PU cover materials and configurations is thoroughly evaluated.
This work package also includes development and evaluation of a delivery system (applicator) that is needed in order to deploy the stents into the bronchial locations, through a bronchoscope.

WP3 Coating & Biofunctionalisation
Objective of this work package was to design and develop a biofunctional stent coating containing an anti-cancer drug.
An ideal coating has to allow the passage of glucose through the coating material enabling the feed, and therefore the survival, of the cells lying in the luminal part of the stent. The coating has also to provide a barrier for tumour cells in order to contrast tumour suppression and infiltration thus avoiding stent displacement and restenosis. The barrier is created by mechanical and pharmacological means, in fact the coated stent has to exercise sufficient radial force against the wall to remain well anchored and it releases an anti-cancer drug to kill/weaken tumour cells. In addition, the coating has to adhere well during the crimping and deploying of the stent.
Two partners, Vysera and UKAachen-CVE developed respectively a PU-foam and PU-fleece coating for PulmoStent. UU was responsible of designing and producing polymeric microspheres loaded with the chosen anti-cancer drug (Gefitinib). The resulting PU coating were investigated in cooperation with Epithelix, NUIG and UU by means of glucose diffusion experiments, tensile strength tests, cytotoxicity testing, laserscanning analysis and drug releasing measurements. Results of those tests were used to guide and optimise the coating manufacturing techniques and ultimately to select the most suitable coating.

WP4 Respiratory Epithelium
In the first part of this project, Epithelix (WP4) was in charge to develop and standardize the cell culture technology, which allows the cultivation of airway cells (epithelial cells and fibroblasts) isolated from pig and sheep. For practical and scientific reasons, sheep was selected for animal experiment. However, standard operating procedures have been developed for the cultivation airway cells from pigs and sheeps.
The second objective was to evaluate the capability of airway cells to adhere, grow and differentiate onto the material selected for the coating of the stent. Numerous material samples from various companies and partners were tested. Most of the work was initially done with a material (sprayed fibres of polyurethane) developed from NonWoTecc partner. However, this partner finally left the consortium in the middle of the project and all works were restarted with the newly introduced partner, UKAachen-CVE. Important variations in results were observed between the different coatings tested. Finally, for technical reasons, the PU fleece coating from UKAachen-CVE was selected for stent production. The results suggest that the PU fleece coating is not toxic, even if the cells were not able to attach to the UKAachen-CVE material. As a consequence of the fact that cells were not able to adhere on the PU from UKAachen-CVE, it was decided to add a small layer of fibrin on the coating to allow the cells to adhere and proliferate onto it. Additionally a sophisticated device was developed to seed the cells in situ after stent placement.
The third part of the work was to evaluate the acute and long term toxicity of the drugs candidates (Erlotinib and Gefitinib) selected to be incorporated into the coating in order to reduce or stop the cellular invasion of tumour cells into the bronchus or the trachea. Our results demonstrated that there is no visible toxicity of those compounds on normal human airway epithelial cells indicating that Erlotinib or Gefitinib can be used.
The last part of the work was to evaluate the potential toxicity of all the components used for making the stent (metal, coating, coating adhesion components, sterilisation process, etc.) as well as evaluate the potential toxicity of the final assembled PulmoStent. The results suggest that all the components used to reconstitute the stent are safe. Moreover, the same results were observed with the final assembled PulmoStent suggesting that the PulmoStent could be safely used for animal studies.

WP5 PulmoStent Production
In this work package, the expertise from all partners is brought together. The results from the first four work packages are used to develop the PulmoStent design and prepare its implantation. Developing this final design involves several parts:
Firstly, bioreactors for in-vitro culture of cell seeded stents are developed. A 2D bioreactor is built for testing influences of airflow on respiratory epithelium. This is important to find out if and which mechanical stimulation respiratory cells need for pre-culture and if they can withstand the shear stresses of physiological breathing. A consequent development of this bioreactor is a version which allows culture of 3D constructs to evaluate the cell-seeded stent behaviour.
Part of this work package includes extensive in-vitro testing of stents and applicator to allow for a simple and easy implantation of the stents. The stent applicator is tested in a cadaver study to gain experience in an operation room setting.
As a last step of this work package, a final stent design is selected for implantation. Here, all results from the previous work packages are brought together: Inputs from modelling and stent development are needed to decide for the stents itself; a decision for a specific coating is made based on results of cell growth on and glucose transfer through the coating, mechanical properties of coating and crimping behaviour of coated stents. The respiratory epithelium is used autologously and culture carried out as defined before.
The decision for a final stent design involves setting up standard operation procedures for all steps of the PulmoStent manufacturing process: Stent manufacture, coating, cleaning and loading the applicators, sterilisation, cell harvest, expansion and seeding as well as for the implantation itself.

WP6 Proof of Concept
In order to evaluate all the developments from the previous work packages, in this work package an in-vivo study is carried out. The aim is to prove the concept of the PulmoStent in a large animal model.
As the airway diameters are similar to human, sheep were chosen as the experimental animal. Two stent designs are tested, one has a laser-cut backbone (from NUIG), and the other one is braided (from ITA). Both stents are covered with a PU-non-woven (made by UKAachen-CVE) as it allows nutrition of the seeded epithelial cells.
In a first pre-trial, in each main bronchus of a sheep, a bare and a covered stent are implanted for testing feasibility of implantation and possible stent migration. After completing this trial, the proof of the PulmoStent concept is started. Here, two covered stents are implanted in the main bronchi of sheep; one of which is covered with respiratory epithelial cells. This allows direct comparison of a cell seeded versus the unseeded stent by means of mucociliary clearance.
To assess the stents’ behaviour, regular bronchoscopic controls are conducted as well as blood is drawn to measure a possible inflammation. After having the stent implanted for 1, 3 or 6 months, the sheep are euthanized and dissected. This allows getting an insight in tissue reaction in the stent’s surrounding. Further, the tissue is analysed histologically and electron microscopically.

WP7 Potential for multiple use
The PulmoStent Project allows multiple technologies to be developed. An important part of the project is to evaluate possible applications for one or more of those developed technologies in the gastrointestinal (GI) area. Vysera, one of the partners of the consortium, has an extensive experience in developing devices for GI applications using its approved PU-foam material which is the same material used to coat PulmoStent. The objectives of the Work Package 7 “Potential for multiple use” changed slightly during the project to comply with the best outcomes of PulmoStent while they were achieved. In the end, two concepts were considered of immediate interest, i.e. a tissue engineered oesophageal stent and a drug eluting device for Crohn’s applications. In particular, a preliminary proof of concept was realised for the tissue engineered oesophageal stent with the aim to outline possible bioreactor set-ups and pre-conditioning protocols. The eluting device for Crohn’s disease applications has been deeply investigated. Protocols and prototypes were produced and underwent through in-vitro characterisation tests.
Project Results:
The S&T results are described work package per work package starting from work package 1. Work package 8 and 9 are excluded, as they focus on standardization, dissemination and exploitation tasks.

WP1 Modelling and Model Validation
T1.1 Development of computational models of filament-based and laser cut coated stents:
The objective of this task was to develop computational structural mechanics models and simulation of stent deformation.
NUIG created finite element models of both their own laser-cut designs and the ITA braided designs. Initial models were based on a 10 mm x 30 mm stent however as more animal cadaver data became available, these changed after a number of iterations to 15 mm x 30 mm designs. Laser-cut designs were modelled using Abaqus with a constitutive model to describe the superelastic properties. The models were developed to provide outputs of device radial forces during crimping (radial resistive force) and during deployment (chronic outward force). Based on these models, the designs were optimized to give chronic outward forces in a range similar to benchmark devices (WP2). Braided stents were modelled both analytically (to compare with published methods) and using finite element methods in both stent extension and radial compression modes. Each of these models allow for assessment of the influence of wire diameters, braid angles, number of wires and the effect of open or closed end cells. The stent cover is modelled; membrane elements provide appropriate representation of stent radial forces, while shell elements are best for detecting buckling.

T1.2 Validation of structural mechanics models:
The objective was to compare the initial structural mechanics models against the in-vitro/bench mechanical testing.
A radial crimping and release experimental procedure was developed in order to obtain test data for comparison with the model predictions. Samples are crimped at 37 °C from 15 mm down to 5 mm and released, with a continuous plot of radial force vesus diameter obtained. Test data for uncoated laser-cut stents compared well with model data for chronic outward force – which is clinically the most important. The model data did not however pick up an axial folding/buckling that occurred experimentally during crimping. Significant refinements of the models were therefore undertaken in order to be able to predict this folding behaviour. The coated models are predicting increased radial forces with increasing coating thickness – there is good correlation between computational and experimental data. Test data for uncoated braided stents also compares well with the models.

T1.3 Model enhancement:
The objective of this task is to provide stent model enhancements to better reflect actual evolving design configurations and coatings likely for in-vivo use.
Enhancement to the basic stent model includes the ability to incorporate manufacturing related defects and variability in the case of the laser-cut stents. In the case of braided stents, this has included more accurate representation of actual wire properties. In relation to the coating, the main enhancement has been the testing and characterization of coatings from VYS and NWT, so as to accurately build the coating into the model – as the coating material properties and configuration on the stent continue to evolve. SEM and tensile data has been collected to input to these enhancements. Refinements of the mesh technique used for the coating itself have been undertaken. These enhanced models can be used to predict the mechanical effects of coating changes. Drug delivery models are also included in this task. This work has focused on exploration of models for drug release from the polymeric microspheres. A number of suitable analytical and finite element approaches have been identified in order to predict release from the PLGA microspheres into a perfect sink.
As stent designs and coating formulations continued to evolve, the models were continuously updated to reflect changes and refine accuracy. Samples of ovine trachea and bronchi were mechanically tested to provide property data for in-vivo models. Mechanical property testing was supported by histology evaluations to validate anisotropy observations in mechanical properties. Initial bronchial models showed that computation techniques can be used to assess deformations in the tissue during stent deployment.

T1.4 Validation of enhanced models:
The objective here was to compare mechanical performance testing of coated stents against model predictions. This provides an understanding of how useful the models are for both stent design purposes and for predicting in-vivo performance.
Evolving design changes for both the laser-cut and hand-braided stent designs were incorporated into both the computational modelling activities and the experimental mechanical testing. In addition, updates and enhancements to the PU coating formulation were also built into the models. In particular, the introduction of the PU fleece coating from UKAachen-CVE was assessed – the experimental data provided a good fit to a hyperelastic Neo-Hookean material model inbuilt to the finite element code used. The adhesion mediator used to attach the UKAachen-CVE cover to the stent was also modelled to capture the contribution this makes to stent radial force. The models allowed for evaluation of stent radial force at different points during the crimp/deploy cycle and allowed for identification of stresses, strains and pinching loads on the cover.
The enhanced FE model of the bronchus now incorporated realistic ovine airway geometries – these were built using computational tomography data collected from the animal study work in WP6. Mimics software was used to convert the CT data to the FE model. The models allow for the investigation of the stresses and strains in the bronchus after stent expansion and during physiological loading. Stress contour plots can be used to predict areas of inflammation caused by high stresses in the tissue. Data from the initial animal studies demonstrate that these in-vivo models have good predictive capabilities in relation to device geometry during deployment and for inflammatory granulation response.

T1.5 Modelling of stent performance in-vivo:
The objective here was to demonstrate final refinements to the stent models and in-vivo models for the implanted designs.
In relation to the braided stents, the final models of the implanted dog-boned devices incorporated substantial refinements. In particular, the wire cross-over points were modelled to allow contact, with cross slip, rather than joints rotating about a fixed point. The models showed an excellent correlation between the actual and computation geometric configuration of the braided stents during crimping and deployment. In addition, radial force experimental testing and computational models showed that the models are accurate for prediction of the clinically-relevant chronic outward force of the device.
For the laser-cut stent models, in addition to full representation of the heat treated nitinol, accurate properties and configuration for both the PU cover and the adhesion mediator prove to be critical to ensuring accurate predictive capability. Full details on both the laser-cut models and hand-braided models are reported in D1.5 and D1.6.
The final in-vivo model captures both the airways and the lungs. Material properties for these tissues were obtained experimentally and from the literature. The model captures deformations and loading initially due to the delivery catheter positioning and then due to the deployment and release of the stents. The models clearly show how the location (dimensions) and geometry of the bronchi influence the final deployed configuration and the stresses in the devices and in the tissues. Full details of the in-vivo model for both hand-braided and laser-cut stents are described in D1.6.

WP 2 Stent Development
T2.1 Definition of design parameters:
The objective was to identify the clinical/anatomical requirements with physician input and input from sheep cadaver studies. In addition, to identify mechanical performance requirements using input from benchmark testing of commercial devices.
Key design requirements were documented in a ‘user inputs’ document – this included aspects such as crimped profile, foreshortening, radiopacity and delivery mechanism. To define exact dimensions, a cadaver study for sheep and pig was performed in collaboration with the animal facilities at RWTH Aachen University. According to the results, a stent with a diameter of 15 mm and a length of 30 mm was proposed with appropriate oversizing for right and left caudal lobar bronchi. Additional mechanical testing and histology has been performed at NUIG to support design and model activities.
The design and clincial performance of commercially available tracheobronchial stents were reviewed. This provided inputs to help select appropriate benchmark devices for fundamental mechanical performance such as radial force and crimped profiles. Samples of commercial stents (UltraFlex, Boston Scientific and AerStent, Leufen Medical) were obtained for benchmark radial force testing.

T2.2 Design and development of filament-based stent:
Filament based stent structures were produced both with braiding and warp-knitting technology. Warp knitting technology was initially identified to be unsuitable for the PulmoStent project due to the fact that the coating partners would have difficulties to coat this structure and to load it into the stent applicator.
The braided stents are manufactured in both a manual and an automated process. For the first stent version Nitinol with a diameter of 100 μm was used. To achieve higher radial forces and to increase the stent’s ability for coating, the Nitinol diameter was changed to 200 μm. Because of this modification the stent structure is much stiffer and the handling for the coating procedure was highly improved. Beside, single filament based stents were produced by hand. With this process, two different configurations, one with 24 and one with 44 crossing Nitinol wires were produced. Because of the closed braided structure of the single wire stents, a significant higher radial force compared to the machine braided stent was achieved.
Finally, to get a tighter position in the lung or bronchi a stent structure with funnel ends and 24 crossing wires were produced. The expanding ends have a length of 5 mm and a diameter of 17 mm. The coating part in the middle has a length of 20 mm and a diameter of 15 mm. In the animal trail this stent structure was used.

T2.3 Design and development of laser-cut stent:
The objective of this task was to use 3D solid model systems combined with finite element computational techniques to develop bronchial stent designs that can be laser-cut from tube and provide tuneable performance when material properties, design and shape-setting heat treatments are all combined
Laser-cut concepts were prepared based on the design inputs. Initial prototypes (version U3) were prepared for mechanical evaluations and were also supplied to the coating partners for coating trials. Radial forces and in particular chronic outward force were in the desired range; however the crimping behaviour of the devices was not ideal; some asymmetric folding/axial buckling was observed during diameter reduction. Taking inputs from the model development and validation activities in WP1, a number of design iterations (U4) were therefore progressed to eliminate this buckling. In particular, the designs are made more robust by accounting for process variations and true material properties. A further U5 design was extensively modelled and experimentally tested. The performance during crimping and deployment was optimized to provide controlled expansion and a predictable radial force. In particular, the radial force was tuned to be within the range of that recorded for commercial stents such as the AerStent (Leufen Medical) and the UltraFlex stent (Boston Scientific). The impact of PU cover performance was also assessed; in general the PU covers added to the stiffness and radial strength. The moulded PU from VYS had a more pronounced effect in this context. A variation of the U5 design was also developed during this period – this involved adding a larger diameter flare to the proximal end of the device, to improve the resistance against migration. This U5F version was also extensively modelled and tested – and ultimately included in the animal study.
Finally, with a view to optimization for human studies in the future, a U6F version of the device was developed. This U6F version uses alternate locations for the ring connectors – this should ultimately allow the device to be crimped to smaller profiles that may be expected for the human anatomy.

T2.4 Manufacture of laboratory samples stents:
The objective here was to manufacture lab samples initially (and ultimately functional samples) of the selected braided and laser-cut stents.
Hand-braided and machine-braided laboratory samples were produced by ITA, as well as warp-knitted stents originally. The laser-cut stents designed by NUIG were converted to 2D CAD patterns, with laser-cutting subcontracted to Admedes. Working with Admedes to optimize the material and heat treatment processes, several iterations were made and tested. Ultimately the laser-cut U5F was selected and progressed to functional status

T2.5 Selection of stent applicator:
The objective here was to identify a suitable applicator mechanism and to subsequently prototype and test this.
It was initially decided to use an outer sheath withdrawal mechanism that would have capability to track over a guidewire and could be placed into the bronchi using a flexible bronchoscope if necessary. Both NUIG and ITA initially worked on practical concepts for this prototype – respectively evaluating the ability to load both the braided and laser-cut stents into the applicator.. Ultimately the same stent applicator design was used successfully for both stent types – and used in the animal study. This outer sheath withdrawl device was designed and prototyped at NUIG.

T2.6 Testing and evaluation of laboratory sample stents and stent applicator:
The objective here was to evaluate the mechanical performance of all bare and covered stents – in particular interactions with the stent delivery system (applicator) were to be evaluated.
Primarily radial force testing was used to evaluate the mechanical performance of stent samples. This was mainly performed at NUIG using an iris mechanism radial force tester at 37 °C – as the properties of nitinol stents are highly temperature-sensitive. However additional testing was also performed by ITA using a V-Block system. The computational models were also developed to correlate with both approaches. This stent testing provided guidance to optimize all designs and aided the selection of designs to progress to functional status and ultimately for use in the animal study.
Related to WP3, much of the applicator testing focused on assessing the ability to load the coated devices into the applicator and to deploy without buckling or permanent damage or deformation. In general, the addition of the coating made it more difficult to load both types of stents. The high thickness and stiffness of the VYS foam/process was particularly challenging. Ultimately, this testing activity resulted in selection of the PU fleece coating – it had less of an impact on radial forces, loading forces and was easier to crimp and deploy.
In addition to mechanical evaluations, further sheep cadaver work was also performed in this period. Studies at NUIG using tracheal and bronchial samples helped identify which devices gave the best apposition and lowest migration risk when in contact with the tissue. Further work with UKAachen and NUIG showed that the stent applicator worked well (for both stent types) in the full airway anatomy of a sheep cadaver.

WP3 Coating & Biofunctionalisation
T3.1 Structural design of the PU coating
The objective of this task was the optimisation of the structural design of the PU-based stent coating.
In particular, two different coating were developed by Vysera and UKAachen-CVE.
Vysera developed a triblock segmented polyurethane (SPU) produced though a 2 steps methods. During the first step a prepolymer is obtained by reacting diisocyanates with polyols groups and creating the characteristic urethane linkage. In the second step, the low molecular weight chain extender is used to link the pre-polymer segments yielding a high molecular weight polymer. The final material consists of alternating sections of hard segments, composed of diisocyanate and low molecular weight diol chain extender and soft segments, composed of polyols. By varying the ratio of soft and hard segments, final foam material properties can be customised and tuned according to the desired application.
UKAachen-CVE used a different technique based on a spraying technology to generate a polymer fleece. In order to produce sprayed PU fleece coating with the properties as named above, many spraying iterations were carried out. During those iterations, the different process parameters spraying distance, pressure, material flow rate and spraying time were varied consequently. As a result it can be seen, that larger spraying distance results in a more porous and fleece-like structure on contrast to dense and film-like structures at low distances. The porosity can be further ensured by a combination of high pressure and low flow rate respectively low pressure and high flow rate. This combination contributes to a more uniform surface structure with less PU drops between the fleece fibres. This results not only in a uniform surface but also in a more equal property distribution along the whole fleece. Additionally, information about glucose transport (UU) and tensile strength (NUIG) were gathered.

T3.2 Stent coating strategies:
The objective of this task was to define and optimise the used coating technique.
The coating technique developed by Vysera is a type of compression moulding technique in which the PU material is dispensed inside the mould and a system of clamp applies a certain pressure. Parameters such as the amount of dispensed material, pressure applied, curing time and temperature are responsible of the final material structure. The moulding process assures a perfect adhesion between the PU and the nitinol stent and further adhesion mediator are not required.
The final parameters were chosen based on the coating performance to support cell growth, allow glucose passage, load and release anti-cancer drug and comply with mechanical requirements. The polyurethane foam developed by Vysera is biocompatible and meets the requirement of the ISO 10993-1:2009. Further tests run within the project by Epithelix, confirmed the ability of the material to support adhesion and proliferation of primary human epithelial cells. Even though the polyurethane foam is not glucose permeable in its standard form, a new technology was developed to overcome this issue, resulting in a laser drilled holes of 30-50µm diameter through the PU foam. This treatment allows glucose transport across the coating within 72 hours.
PU-foam coatings underwent to several mechanical tests. Tensile tests were run on flat material samples, while radial force tests were done on coated stents. Simple PU-foam coating was used as a reference and many variations (laser-drilled, with blank-microspheres, with Gefitinib, etc) were also tested. In general, presence of microspheres, drugs or holes does not have a major impact on the strength and elongation of the material. On the contrary, the density of the final foam has a major influence on the coating behaviour. For examples, low density samples have more elongation and therefore, facilitate the stent crimping during the loading into the delivery catheter. One of the major problems for the PU-foam coating is the presence of material between the struts, resulting in a very bulky coating. This only feature prevented the Vysera coating to be selected as final coating for the PumoStent. Within the project, Vysera optimised the PU-foam coating (density and formulation) and obtained a coating that enables the hand-braided stent load/unload into the delivery catheter in a similar way to the PU fleece coating, used in the animal trial.
UKAachen-CVE uses a different technique based on a spraying technology to generate a polymer fleece. In order to guarantee the adhesion of the fleece to the nitinol stent, UKAachen-CVE developed a chemical-based method of binding a polymer to a metal surface. The stent surface has to be pre-treated with several layers of adhesion mediator (see Deliverable 5.8) before the PU fleece can be sprayed onto the stent. Tensile strength tests proved the coating a sufficient stability and adhesion strength during crimping and deploying. During several iterations, the pre-treatment was improved with regard to its thickness and the spared and thus uncoated ends of the stent.

T3.3 Development and optimization of drug-loaded microspheres:
Objective of this task was to develop and optimise drug loaded microspheres.
At the first meeting, UU had presented the general strategy for developing drug loaded microspheres. Research into the feasibility of preparing kinase inhibitor loaded microspheres using the model kinase inhibitor sunitinib led to the current concept of making Gefitinib loaded microspheres. These microspheres can be loaded in a fibrin hydrogel on the outer layer of the stent coating.
In parallel, the use of poly-urethane (PU) type polymers for making Gefitinib-loaded microspheres has been investigated. For this purpose, UU evaluated two different PUs provided by Vysera. But the studies were not successful because the polymers have a high molecular weight and, consequently, dissolving them in organic solvents proved difficult. The existing process of making drug-loaded microspheres is based on the emulsification of the dissolved polymer together with the drug in another hydrophilic solvent. Hence the dissolution of the polymer is an essential step in the preparation of drug-loaded microspheres.
Finally, it was decided to focus the efforts on making larger PLGA microspheres which will have a more extended/prolonged release window due to the longer diffusion distances within the microspheres. Standardized mesh sieves with three different sizes of 100 µm, 50 µm and 20 µm, respectively, and single emulsion solvent evaporation technique for making bigger size microspheres. Four different size fractions of Gefitinib-loaded PLGA microspheres were prepared: >100 µm, 50 µm - 100 µm, 20 µm - 50 µm and <20 µm and the release behaviour of the Gefitinib from the different fraction microspheres and from unfractionated microspheres were compared. It turned out that microspheres with smaller size had a faster release of Gefitinib compared with bigger size microspheres. Finally, the microspheres with size range of 50 µm - 100 µm which gave a sigmoidal drug release profile and the drug release last for 3 - 4 months were chosen for embedding into PU coating.

T3.4 Embedding of microspheres into the stent coating:
The objective was to establish successful methods for embedding microspheres into the PU coating.
PLGA microspheres (diameter ranging between 50 - 100 µm) containing Gefitinib, chemotherapy drug of choice for this project, were embedded inside the PU-foam developed by Vysera and the PU-fleece coating developed by UKAachen-CVE. Vysera developed also a protocol to embed directly Gefitinib inside the foam without microspheres.
The embedding protocol developed by Vysera involves dispersing the particles (microspheres or drug) inside the pre-polymer and then the addition of the chain extender/cross-linker. A maximum of 15 % w/w and 10 % w/w respectively for drug-loaded microspheres and Gefitinib were embedded inside the PU-foam. The reaction kinetics between the cross-linker and the prepolymer is fast therefore the reaction of Gefitinib and/or the PLGA with the other components of the foam is minimised. The drug (or the drug-loaded microspheres) remains entrapped in the matrix and can elute out slowly.
UU investigated how microspheres can be embedded in fibrin hydrogels without disrupting the integrity of the fibrin network. In general, fibrin hydrogels are relatively weak and investigations are currently exploring whether the strength of this type of microsphere/fibrin material is adequate for application in the PulmoStent device. Alternatively, stronger hydrogels can be applied for the embedding of microspheres
UKAachen-CVE and UU have explored how microspheres can be embedded in PU fleece in a sandwich like structure consisting of PU-fleece/fibrin hydrogel and a top layer of PU fleece. PLGA microspheres can be embedded in the fibrin gel layer between the two layers of PU fleece. For this purpose, the microspheres were mixed with a fibrinogen solution and applied on the first already sprayed and dried layer of fleece. The fibrinogen was dried after which a second PU layer was sprayed on top of the composite coating layers.

T3.5 Drug release kinetics from embedded microspheres:
Objective of this task was to investigate the release of Gefitinib from the different PU coatings.
Gefitinib is the 1st choice of treatment in pulmonary cancer and therefore it is also a good candidate drug for the PulmoStent project. UU successfully created a Gefitinib loaded poly (lactide-co-glycolide) (PLGA) microspheres with 8 % w/w drug loading. These initial batches of Gefitinib-loaded microspheres had a relative fast drug release profile, allowing for about 2 weeks of drug release. The optimal microspheres would have a 1-2 month drug release profile.
Afterwards data was collected regarding larger Gefitinib PLGA microspheres that were sieved into different size fractions between 20 µm - 100 µm. While small microspheres (<20 µm) showed rapid release, the microspheres with sizes between 50 µm - 100 µm showed the preferred release profile, i.e. sustained Gefitinib release over 2-3 month.
In vitro release studies with Gefitinib and Gefitinib-loaded microspheres embedded in PU foam are shown in. The in vitro release was studied at 37 °C for over 3 months and is still ongoing. These experiments showed that the drug is released in a sustained manner at a zero-order release profile reaching 20 % - 450 % cumulative release after 3 months.
Gefitinib-loaded PLGA microspheres that had been embedded in PU fleece via the above described sandwich layer technique also showed a sustained release profile when incubated at 37 °C, reaching 20 %-40 % cumulative release after three months. The material UKAachen-CVE_a sample had an incomplete second layer, which explains its faster release profile.
The diffusion of Gefitinib and model compounds across PU coatings have also been studied. In order to evaluate the relation between material properties such as coating thickness and porosity and their behaviour in the PulmoStent device, a set of PU membranes from both UKAachen-CVE and Vysera were provided to UU. UU studied drug and nutrient transport across these membranes by a tailor-made diffusion apparatus (modified Franz cell).

T3.6 Evaluation of different sterilisation methods:
The objective of this task was to evaluate the impact of different sterilisation methods on the non-viable components of PulmoStent.
The preferred method of sterilisation for both foam and fleece-based PU materials is the ethylene oxide (EO) sterilisation, therefore this method was used for the PU coated PulmoStent as well.
Ten samples were prepared and were exposed to EO sterilization process. Following completion of the sterilization cycle the samples were tested for material integrity using SEM, FTIR and tensile strength. All samples remained functional within parameters of the material.

WP4 Respiratory Epithelium
T4.1 Isolation and cultivation techniques of the respiratory cells:
Epithelix has profound know-how in the cultivation of human cells and just little knowledge about cultivation of animal cells. Even though many aspects seem to be similar, the reality reveals numerous differences. In fact, the cell culture medium needs to be adapted and reformulated in order to sustain the growth and the differentiation of the isolated cells. In course of the project a standard operating procedure (SOP) for the isolation and cultivation of pig and sheep airway epithelial cells as well as fibroblasts has been developed. At least 4 different conditions for this process have been tested and an optimized process and culture medium has been found. The work done revealed that pig cells were more difficult to handle and cultivate than cells isolated from sheep. However, the results also suggest that sheep cells are difficult to dissociate when cultured on conventional flask or plastic treated for cell culture. The dissociation and expansion of sheep cells remains a difficult task. Most of the cells do not survive with a classical trypsin treatment even if it is extensively diluted.

T4.2-1 Influence of Polyurethane coating on growth and differentiation of respiratory epithelium:
In order to grow and differentiate, cells need first to adhere onto the material. When the material is biocompatible, cells can grow and proliferate onto the material and cells rapidly cover all the surface of the substrate. To induce the differentiation of the airway epithelial cells it is generally necessary to polarize the cells. This can be obtained in-vitro by exposing one side of the cell monolayer to the air. The condition is also that cells continue to be fed with nutrients from the basal side. This gradient induces the differentiation of the cells into a fully functional epithelia composed of basal, mucus and ciliated cells. The gradient is possible because cells are generally cultured onto microporous membrane allowing fluid movements. The challenge was to reproduce all those conditions allowing the cells to adhere, grow and differentiate. In this context different PU coatings developed by NonWoTecc, Vysera and UKAachen-CVE have been tested. Commercially available microporous membrans from Millipore, Costar, Oxyphene have also been tested as alternatives
The results revealed that the PU coating from NonWoTecc allows the adhesion as well as the proliferation of the cells. However, the material was not porous enough to induce the differentiation of the cells. The cells can’t be switched to air-liquid interface because nutrients can’t go through the PU from the basal side. When switched to the air, the cells die. Due to the termination of NonWoTecc the optimal conditions could not be identified. On the samples from Vysera, cells were able to adhere and grow, however the porosity characteristics were not able to promote cell differentiation. The PU fleece coating from UKAachen-CVE was also tested and the results indicated that the cells were not able to adhere onto the material.
The results on the microporous membranes show that cells are able to adhere, proliferate and differentiate onto polycarbonate and polyethylene. One major challenge was how to perform a combination of these coating for the PulmoStent.
For physical and technical reasons, the consortium decided to select the PU fleece coating. As a consequence of the fact that cells were not able to adhere on the PU fleece coating from UKAachen-CVE, it was decided to add a small layer of fibrin on the coating to allow the cells to adhere and proliferate onto it. Additionally a sophisticated device was developed to seed the cells in situ after stent placement.

T4.2-2 Toxicity testing of Erlotinib and Gefitinib on reconstituted human respiratory epithelium:
The objective of this task was to evaluate the potential acute and chronic toxicity of Erlotinib and Gefitinib onto reconstituted airway epithelia.
To reach that goal it was decided to use the commercially available product MucilAirTM composed of human airway epithelial cells fully differentiated onto microporous membranes. The results suggest that Erlotinib and Gefitinib are well tolerated and no cytotoxic effect was detected. In order to determine if Erlotinib and Gefitinib can be used in the PulmoStent system the potential toxic effects of these molecules during a chronic exposure was tested (5 times a week during 2 weeks). The trans-epithelial electrical resistance (TEER), Il-8 and lactate deshydrogenase (LDH) release, as well as the mucociliary clearance was monitored to reveal potential adverse effects. No effect were observed on TEER, the LDH release, the secretion of IL-8 or mucociliary clearance, suggesting that Erlotinib and Gefitinib could be used in the coating of the Stent in order to block the invasion of tumour cells.

T4.2-3 Toxicity testing of all used materials:
The objective of this task was to evaluate the potential toxicity of all the materials used for making the stent.
In that part the potential toxicity and biocompatibility of the laser-cut and braided stent on human airway epithelia reconstituted in-vitro (MucilAirTM) was tested. The biocompatibility of the coating made from Vysera, NonWoTeccc and UKAachen-CVE, were also tested. Many approaches and analyses were performed to reveal the potential toxicity and biocompatibility of the stent components. Briefly the trans-epithelial electrical resistance (TEER) has been monitored. This technique allows the monitoring of the physical integrity of the tissue. The physical integrity of epithelia is a crucial indicator of the barrier function of the airway tissue. The secretion of the lactate dehydrogenase (indicator of cell death) has also been measured. Final tests were conducted to measure the capability of human airway epithelial cells to adhere and proliferate onto the various PU coating used. The results reveal that stent materials are safe and do not alter the tissue integrity of the human epithelia. The stent materials do not induce any cell death.
The results also reveal that some toxic elements were secreted from the PU coating after the sterilisation process. This is probably due to a remaining EO in the PU coating. In view of those observations it is recommended to wait enough time between the sterilizing procedure and the implantation due to a potential remaining EO into pore of the PU coating.

T4.2-4 Indirect toxicity testing of the assemble PulmoStent:
The objective of that part was to evaluate the potential indirect toxicity of the assembled PulmoStent.
The results showed that there was no visible indirect toxicity of the full assembled stent on the human airway epithelial cells. It is concluded that the washing process used for the full assemble stent is sufficient to remove any potential toxic element on the surface of the coating. However, it cannot be excluded that the toxic elements trapped inside of the coating could be released with time at a very low rate into the environment around the stent. The methodology used is not adapted to answer this question.
In view of these results, it is important to take into account that the tear or stretching of the coating during crimping or implantation could have an impact on the release of toxic element around the stent after implantation. However, the indirect toxicity testing of the final assembled PulmoStent (crimp and release) revealed that the stretching of the coating during crimping had apparently no effect on the release of toxic element into the supernatant.

T4.2.5 Direct toxicity testing of the assembled PulmoStent:
The objective of that part was to evaluate the potential direct toxicity of the final assembled PulmoStent with fibrin and cells.
Tests have been conducted to coat the stent with cells and fibrin. Due to technical difficulties, which couldn’t be solved with the available means, the tests couldn’t be fully completed. Even though the direct toxicity of the final assembled PulmoStent couldn’t be properly assessed, WP5 has developed a methodology and a sophisticated device for the application of cells and fibrin after stent placement

WP 5: PulmoStent Production
Production of the PulmoStent required the expertise of all partners. The results of the different components (stent, PU-cover, Gefitinib-releasing microspheres and respiratory epithelium) were combined to prepare the in-vivo study.
The objectives here were to firstly develop a bioreactor which allows the culture of a cell seeded stent under air-liquid interface and with biomechanical stimulation of the cells; secondly define the fabrication steps of the final PulmoStent and cell seeding process, additionally the stent applicator needs to be tested and the final design of all components has to be selected for the animal trial.

T5.1 Development of a PulmoStent bioreactor:
The start of WP5 involved the development of two novel bioreactor systems (2D and 3D) for in-vitro culture of a flat construct and the viable bronchotracheal stent. This is important to guarantee optimal culture conditions by means of nutrition and stimulation of the cells when pre-culturing the stent before implantation. Both bioreactors provide air-liquid-interface conditions (thus, medium supply at the basal, air contact of the apical epithelial surface) to induce epithelial cell differentiation and the formation of a multi-layered mucosal structure. In order to provide a functional respiratory epithelium with directed mucociliary function, the systems are able to apply defined shear stress conditions to the surface of the tissue construct. Next to the bioreactors we developed a membrane pump which simulates physiological breathing by means of shear stresses. It is based on an actuator connected to a membrane. As for the respiratory system the pump generates negative pressure as process of inspiration and a positive pressure as process of expiration. The pump volume can be adjusted such that the shear stresses are as high as in the trachea, allowing direct comparison of the in-vitro system to in-vivo conditions.

T5.2 Fabrication of the PulmoStent:
Due to different reasons (i.e. difficulties when crimping a cell-seeded stent), instead of seeding cells on the stent in-vitro, a device was developed to seed cells in situ after stent placement. This simplifies the in-vitro cell handling (no pre-culture of the seeded stent necessary), stent crimping and applicator loading. Such a system was not described before and will hence be filed as patent as it might have interesting potential for other applications as well. Thus, fabrication of the PulmoStent was accomplished by combining the competences of all partners: two different stents as backbones (as provided by NUIG and ITA), two different covers (produced by Vysera and UKAachen-CVE), combining the over with the tumour-specific drug Gefitinib (UU) and seeding epithelial cells with a newly developed and highly innovative device (Epithelix and UKAachen).

T5.3 In-vitro culturing of the assembled PulmoStent :
Due to the changes for implantation and cell seeding, the focus of some of the tasks define before, did change as well. With the bioreactor, we could show in-vitro that simulated physiological breathing does not change the cells’ morphology in-vitro. Hence, seeding the cells in situ on the stent is a reasonable option. Additionally, the seeding device was evaluated extensively in-vitro to show that it has no adverse effects on cells behavior and differentiation: After in-vitro culture of the epithelial cells, no changes in morphology could be seen.

T5.4 Evaluation of the stent applicator:
As the cells are seeded on the stent after placement, there will not be an influence on the cell surface by crimping. However, the crimping and release of bare and covered stents and loading into the applicator was evaluated. Different applicators and crimpers were tested to find an optimal solution. As applicator, a custom-made system was developed in WP 2 as it suited the requirements most. It is made from different tubing and works without a guide wire; as the outer tubing has a small bend, navigation in the respiratory system is possible without. Inside, there is a pusher tube which allows easy stent deployment. To prove functionality of the applicator system, a cadaver study was performed in which stents were placed in sheep cadaver under surgical conditions.

T5.5 Selection of PulmoStent designs to be implanted:
As a necessary step towards WP 6, the proof of principle, two PulmoStent designs were chosen for implantation in sheep. As the characteristics of a braided and a laser-cut stent vary, it was decided to test both stent types. Both stents have a diameter of 15 mm and a length of 30 mm as defined by a cadaver study on sheep before. For the braided stent, it was chosen to use a hand-braided dogbone stent with flared ends, made from a single nitinol wire. The laser-cut stent used in the animal study is called U5F. It has a flare on the proximal end which prevents stent migration and allows better tissue anchorage. Both stents were chosen to be covered with non-woven PU as it showed better properties for epithelial cell growth.

WP6: Proof of Concept
In work package 6 (Proof of concept), the assembled PulmoStent is advanced from in-vitro testing to in-vivo evaluation in sheep. As mentioned before, two stents will be tested: the laser-cut U5F-version (NUIG) and the hand-braided-dogbone stent (ITA). The procedures used conform to the “Guide for the care and use of laboratory animals” published by the US National Institutes of Health (The National Academies Press, 2011) and were approved by the local ethical committee (84-02.04.2013.A452).
As experimental animal sheep was chosen after consulting the responsible veterinarians and conducting a cadaver study to determine airway diameters. In total, stents were implanted in 14 animals. There was a pre-trial in which a bare and a covered stent were implanted as feasibility study and to evaluate the migration risk of the stents. To test the actual PulmoStent concept, each stent type was implanted in six sheep with groups of one, three and six months duration. This allowed evaluating short-term stent migration, epithelial development and tissue reaction, and testing the typical clinical stent duration. Tissue for cell isolation was harvested with biopsy forceps three weeks prior to the implantation. In work package 3, a lot effort was put in developing suitable methods for incorporating Gefitinib-loaded microspheres into the non-woven from UKAachen-CVE. In the animal study we did not include drug-loaded stents in this first animal study as at present, no animal models exist for bronchial cancer in large animals and the development of such an animal model is beyond the scope of the PulmoStent project. On the other hand, the stent developed in PulmoStent are designed for bronchi that have an inner lumen diameter that is in the same dimension as the human bronchi and they do not fit in the airways of small animals like rats and mice for which bronchial cancer models have been described.

First objective of this work package was the production of all PulmoStents for implantation. Further objectives were the implantation of the stents, in-vivo follow up and macroscopic and microscopic analysis of the stents and the surrounding tissue after euthanasia and dissection.
In the pre-trial the implantations went smoothly and no stent migrated. Thus, no further improvements of the stent designs were necessary.
As part of this work package, all PulmoStents to be implanted and spare ones were produced by the respective partners in time and according to the implantation schedule such that all implantations could be accomplished as planned. To evaluate the in-vivo behaviour of the stents in detail, bronchoscopic controls were accomplished with intervals of two to four weeks. During these interventions, blood was drawn to measure inflammation levels and bacterial colonisation of the respiratory system was analysed microbiologically.
After euthanasia, all animals were dissected to gain a macroscopic picture of the stents and the whole lungs. Tissue samples of the lung are collected as well as the stent including the surrounding tissue for further histological analysis. As the last step of the in-vivo study, the stents and the surrounding tissue are microscopically analysed. For histology and immunohistochemistry, the stents are embedded to get sections which can be stained. To evaluate the tissue reactions, H&E staining are done, to prove epithelial differentiation we stain for pan-cytokeratin. Furthermore, we analyse the epithelium on the stent cover electron microscopically to prove ciliary function.
As the experiments are still running, we cannot present results here as they shall be published in a peer-reviewed journal. This will be done when the experiments are finalised. However, so far no stent migrated during implantation periods of up to six months.

WP7 Potential for multiple use
T7.1 Stent seeding with epithelial cells from gastrointestinal origin and dynamic preconditioning:
Objective of this task was to design a tissue engineered stent similar to PulmoStent for GI application.
Oesophageal stents are currently used for the treatment of a variety of benign and malignant oesophageal conditions. Even though the use of self-expandable oesophageal stents has grown immensely over the past decade, improving the patient quality of life, these prostheses have been associated with morbidity rates and high complication, including pain, bleeding, perforation, but also tumour ingrowth and overgrowth. A tissue engineered stent, similar to PulmoStent, might provide mechanical strength pushing and circumscribing the tumour, while the layer of cells can mimic the surrounding environment reproducing the physiological functions of the oesophagus. Even though the presence of a living layer of cells is very interesting, new and provides the most physiological-like solution, the role of the cells in this tract is not as critical as for the respiratory epithelium. For this reason, we have limited the work to a preliminary proof of concept. In particular, a plan to modify an existing bioreactor, used for vascular tissue engineered stents, was addressed. The chamber containing the stent holder can be easily changed according to the oesophageal stent dimensions. A special connector can be used to spray a fibrin-cell layer inside and/or outside the stent, using the same technology developed within PulmoStent. The rest of the set-up remains almost unvaried and it consists of a circuit in which medium can be pumped to feed the living cells layer. Accordingly to the pressure felt by the oesophagus, in resting condition and during swallowing, the medium can be pumped at different pressure stretching and preconditioning the stent.

T7.2 Embedding of Crohn’s disease drug into the PU foam:
Objective of this task was to load drugs used for Crohn's disease treatment inside the Vysera PU foam and characterised the drug kinetics release.
Crohn's disease (CD) is a chronic inflammatory condition of the gastrointestinal tract, characterized by transmural granulomatous inflammation, a discontinuous pattern of distribution, and fistulae. It is a disabling life-long condition that afflicts 300,000-500,000 people in Europe.
There is no known cure for CD. The healing of the ulcerated and inflamed gut mucosa seems to produce the best long-term outcomes. Medical therapy is the most popular approach used to heal the ulcerated area in mild cases or, as a post-surgery therapy, in severe cases, to prevent recurrence of the ulcer. Two of the most common drugs used to manage CD patients are Rapamycin (anti-suppressant) and Metronidazole (antibiotics).
Those drugs have been loaded inside the Vysera PU-foam, a GI stable material. The same protocol developed to load Gefitinib inside the PU-foam was followed. Samples with different geometries (flat and bulky samples) and densities (low and high) were produced and then, analysed by UU.
The tests showed a high loading efficiency and reproducibility (drug recovery rate: 81 %-96 %). The releasing kinetics for the first month showed a slow zero-order release profile for Rapamycin-loaded samples. On the contrary, a fast release was measured for the Metronidazole-loaded samples. The in-vitro drug release behaviour shows that hydrophilicity of the drug greatly influences the release rate of the drug from the PU foams. The metronidazole, which has much higher hydrophilicity compared to Rapamycin release 100 % after 1 week, while Rapamycin released less than 5 % after 1 week. Those preliminary results show the importance of the drug physiochemical properties while choosing the desired kinetics release. More data is needed to understand the role of the sample geometries. So far it is possible to speculate that samples with higher surface to volume ratio have faster release profile but because bigger volume results in more porosity, the prediction of the releasing rate, based only on the surface to volume ratio, is very difficult and requires further experiments.
In conclusion, preliminary studies for a tissue engineered oesophageal stent with similar characteristics to PulmoStent were completed and a possible bioreactor set-up was established. Moreover, a drug-delivery polyurethane matrix has been developed for controlled release of Rapamycin and Metronidazole, respectively an immuno-suppressant and an antibiotic, drugs commonly used to manage patients with Crohn’s disease. Slow kinetics release was measured for Rapamycin loaded samples and a faster one for Metronidazole loaded samples. The findings on Rapamycin controlled elution from the PU-foam are very promising results because the released dose and kinetics comply with clinical requirements.

Potential Impact:
The stent design and modelling activities can have a significant impact to the future of tracheobronchial stenting including in potential human clinical application of the PulmoStent concept. Specifically, the laser-cut design can be readily adapted to suit human dimensions and anatomy and the testing during this project has shown that performance is comparable to conventional devices. It remains to be seen if indeed hand-braided stents would be better for the PulmoStent epithelial cell concept but importantly the design process and FE models mean that modifications can be easily introduced to optimize if needed. Furthermore, NUIG has substantially enhanced its core design and modelling expertise in this project and this may have broader institutional benefits outside the scope of PulmoStent. Applications in other areas, such as gastro-intestinal or vascular could possibly be exploited.
However perhaps more significant than the specific designs themselves has been the development of in-vivo models that simulate the airway and lung anatomy and then simulate the deployment of stents into the bronchi. These simulation models have been developed for both laser-cut and hand-braided stents; upon simulated deployment, the stresses and strains in both the device and the biological tissue can be identified. The regions of high stresses or strain in the bronchi can be used to predict the potential for inflammation and granulation formation. Initial animal study data indicates this is feasible – stent geometry and deformations, as well as tissue stresses, correspond with actual animal study observations. Much work would be needed to validate this in a human clinical setting but it does point the way to a personalized medicine approach i.e. where the stent design can tuned using a pre-placement CT scan of the specific patients anatomy. In addition, the in-vivo models may have substantial application in supporting regulatory approvals for such medical devices.
There has been extensive dissemination of results of these activities. Stent design and model work has been presented at a number of national conferences in Ireland and also at a major biomechanics conference in the US in 2014. The design work has also been the subject of a peer-reviewed publication in the Journal of Mechanical Behaviour of Biomedical Materials. Furthermore, the in-vivo model work has been accepted for presentation this year at the European Society of Biomaterials conference (Prague) and at an FDA-sponsored conference on medical device modelling in Washington DC. Further peer-reviewed papers are also expected.
In the course of the PulmoStent project the Institut für Textiltechnik of the RWTH Aachen University significantly increased its competency in the conceptual design and production of a braided single filament based stents. Here the processing of different nitinol diameters (100 µm up to 200 µm) and variation of stent geometries were analysed. The findings from the respiratory tract can be transferred to stent in the area of oesophagus, bile duct and urinary tract.
Additionally, two new projects have arisen out of the project PulmoStent, which were formulated to proposals. Both project ideas were submitted by national funding and are still under evaluation. The first project idea deals with an automation of the production for the currently manual hand-braiding process. It is the clear goal to increase the reproducibility of the single filament based stent. During the course of PulmoStent a concept development was started due to the experience of the manual hand-braiding process. In September 2014 the Institut für Textiltechnik of the RWTH Aachen University has been honoured with the Walter-Reiners-Stiftung award for the concept of “An automation of the current manual hand-braiding process”. With the second project idea, which is uncoupled from the braiding process, a single filament based stents will be developed. With this both project ideas the Institut für Textiltechnik succeeded to build up the competences in the manual braiding process and to place new research issues on this field.
The main breakthrough for Vysera during the PulmoStent project concerns the ability to load anti-cancer and immuno-suppressant drugs directly into the polyurethane foam and release them in a controlled way without the help of any other intermediary like the polymeric microspheres.
This technology can be used in a very broad range of applications. In particular, because of the proven stability of the material in the gastro-intestinal tract, immediate applications can be researched in the GI area.
In the WP7, the ability to load and release drugs administrated to Crohn’s disease patients has been already demonstrated. This might have a very big impact in the future of the medical therapies in CD. More than 60,000 persons in EU develop a perianal fistula as a CD outcome. Current mechanical solutions (drainage seton, perianal plug,etc. ) aim only to drain the fistula and combined with medical therapies achieve a slow healing with low rates of success and high rates of fistula recurrence. A new device, able to deliver in situ the medical therapy and with the required mechanical design, can dramatically improve patient’s quality of life. Beside the evident clinical need, a drug-delivery device for CD can reduce costs associated with the failure of current devices and costs related to the reoccurrence of the fistula. Another application in the GI area is a coating for oesophageal stent to release chemotherapic drugs in case of oesophageal cancer.
However, the applications of this technology do not limit only to the GI area but regard all clinical conditions where a controlled drug release is needed. The ability to mould the PU-foam in almost any kind of shape and tune its mechanical properties thanks to the chemical freedom provided by polyurethanes, makes this technology incredibly powerful.
The ability to coat stents for tracheobronchial stents by means of spraying has a significant impact on the future of tracheobronchial cancer treatment. As learned during this project, the properties of the fleece coating can be specifically adapted by changing one or more of the process parameters. This allows the production of a coating which is as dense as necessary and at the same time as porous as required. This knowledge can be use and transferred to other applications of sprayed polymer and especially PU fleece as for example vascular grafts or artificial heart valves.
Furthermore, the development of a pre-treatment procedure resulting in a strong and stable adhesion between PU and nitinol offers great possibilities in the field of cardiovascular engineering, as nitinol is the state of the art material used for any kind of vascular stents. Polymers in general and specifically PUs are strongly investigated materials for the use in blood contact. A good biocompatibility is the base for considering the material as a native tissue replacement. But the connection between a polymer and a metal is a complex chemical process which is not easy to handle, especially with the requirement of biocompatibility. Thus, this newly developed method is notified in an invention report and will be turned into a patent in the future.
UKAachen has developed bioreactors to cultivate tissue engineered 2D and 3D constructs. With the 2D bioreactor, not only different pre-culture protocols can be tested on respiratory epithelial cells, but also it can be used to test the influence of different breathing patterns (i.e. from artificial breathing machines) on epithelium und also ciliogenesis. Up to now, there was not much research in this area which makes it highly interesting. Projects resulting from this achievement can be from two types: basic research on ciliary development and applied research to develop tissue engineering of respiratory replacements.
Also, a new device for coating implants with single cell layers was developed at UKAachen and research in this area will be continued. Not only coating of implants, but also other possible applications as cell therapy of different organs are highly feasible with this device. It will be explored in a project which is already funded by another funding source together with researchers from Karolinska Institute. Furthermore, UKAachen is currently exploring the possibility of patenting this device to be able to exclusively use it for further trials and be able to work in close cooperation with interested companies. Further developing this device and standardizing it will allow advancing it to clinical use.
Especially the results of evaluation of a pre-version of the coating device were already presented at different European conferences and a peer reviewed manuscript was accepted for publication. After filing the patent, a manuscript to describe this new device will also be published in a peer-reviewed journal.
With the proof of concept study, we made a huge progress of advancing the PulmoStent concept or even parts of it into the clinic. Clinical approval is based on pre-clinical trials which were accomplished here. This applies not only to the complete PulmoStents but also parts thereof. With a successful animal trial also the newly developed stents or the cover (used for the first time in a respiratory application) can be brought to clinical application. Here for the first time, a trial was accomplished combining a stent with tissue engineered respiratory epithelium and is hence a huge step forward. Also we were able to locally establish sheep as experimental animals for respiratory stents, which has not been tried before. With the results after finalising the experiments within the next time, we are the first to report on such experiments and will publish these in high-ranked peer-reviewed journals.

Univ.-Prof. Dr. med. Stefan Jockenhövel
jockenhoevel@ame.rwth-aachen.de

Uniklinik RWTH Aachen
Institute of Applied Medical Engineering
Tissue Engineering & Textile Implants
Pauwelsstr. 20
52062 Aachen
Germany