Type-1 diabetes and influenza cause significant illness and unnecessary medical costs despite the existence of insulin for maintenance of diabetes and a vaccine for prevention of influenza. This dissertation describes three studies on the development and clinical translation of microneedles to improve the administration of these biopharmaceuticals. The first study reports on a sharp-tipped hollow metal microneedle designed to reduce manufacturing costs, improve insertion into skin, and improve fluid flow compared to other hollow microneedles used for drug delivery. The results showed sharp-tipped metal microneedles could be fabricated using an inexpensive electroplating and sacrificial micromolding process. Single-microneedle devices made by this method achieved high flow rates and delivered model drugs into tissue. The second study reports on insulin delivery using microneedles in children with type-1 diabetes. The results showed microneedle insertion was less painful, which is a promising result for improving injection compliance in children. Additionally, microneedle delivery showed rapid onset of insulin action compared to subcutaneous catheter delivery, which may enable automatic closed-loop insulin therapy. This was the first study of drug delivery to children using microneedles. The last study reports on microneedle patches for self-vaccination against influenza. Human subjects were recruited from greater Atlanta, were asked to self-administer placebo microneedle patches, and were then given a dynamic questionnaire to determine their views and preferences regarding influenza vaccination using microneedles compared to conventional intramuscular injection. The results showed that microneedles were usable by the participants, the introduction of microneedles may improve vaccination coverage by approximately 20%, and self-administration of vaccines may significantly reduce vaccination costs for a healthcare payer. This was the first study to assess the ability of human subjects to self- administer a microneedle patch and the first study to determine the potential impact of self-vaccination against influenza using a microneedle patch on vaccination coverage and vaccination cost. Overall, the fabrication advances and positive findings from human subjects research support additional translation of microneedles for insulin delivery and self-vaccination toward clinical use.

Therapeutic efficacy of drugs depends on their ability to reach the treatment target. Drugs that exert their effect within cells are constrained by an inability to cross the cell membrane. Methods are being developed to overcome this barrier including biochemical and biophysical strategies. The application of ultrasound with microbubbles increases the permeability of cell membranes allowing molecules, which otherwise would be excluded, to enter the intracellular space of cells; a phenomenon known as sonoporation. This thesis describes studies aimed at improving our understanding of the mechanism underpinning sonoporation and of the exposure parameters affecting sonoporation efficiency. Cancer cells (KHT-C) in suspension were exposed to ultrasound and microbubbles – total of 97 exposure conditions. The effects on cells were assessed through uptake of cell-impermeable molecules (10 kDa to 2 MDa FITC-dextran), cell viability and microscopic observations of the plasma membrane using flow cytometry, colony assay and electron microscopy techniques. Sonoporation was a result of the interaction of ultrasound and microbubbles with the cell membrane. Disruptions (30-100 nm) were generated on the cell membrane allowing cell impermeable molecules to cross the membrane. Molecules up to 2 MDa in size were delivered at high efficiency (~70% permeabilisation). Sonoporation was short lived; cells re-established their barrier function within one minute, which allowed compounds to remain inside the cell. Following uptake, cells remained viable; ~50% of sonoporated cells proliferated. Sonoporation efficiency depended on ultrasound and microbubble exposure conditions. Microbubble disruption was a necessary but insufficient indicator of ultrasound-induced permeabilisation. The exposure conditions can be tailored to achieve a desired effect; cell permeability of ~70% with ~25% cell death versus permeability of ~35% with ~2% cell death. In addition, sonoporation depended on position in the cell cycle. Cells in later stages were more prone to being permeabilised and killed by ultrasound and microbubbles. This study indicated that sonoporation can be controlled through exposure parameters and that molecular size may not be a limiting factor. However, the transient nature may necessitate that the drug be in close vicinity to target cells in sonoporation-mediated therapies. Future work will extend the investigation into in vivo models.

Water-soluble fullerene containing polymers comprising of poly(2-(dimethylamino) ethyl methacrylate)-fullerene (PDMAEMA-C60) with targeting moieties, poly(oligo(ethylene glycol) methyl ether methacrylate)-C60 (POEGMA-C60), nanocrystalline cellulose-fullerene (NCC-C60) and NCC-C60-POEGMA were synthesized and their solution properties were investigated. PDMAEMA-C60 with galactose targeting moiety was prepared by atom transfer radical polymerization (ATRP) and atom transfer radical addition (ATRA) processes. The self-assembly of galactose functionalized PDMAEMA-C60 structure in aqueous solutions was investigated using dynamic light scattering (DLS) at different pHs. A smaller hydrodynamic radius (Rh) was observed at pH 10 than at pH 3 due to electrostatic repulsion at low pH values. In addition, free PDMAEMA chains induced the demicellization of self-assembled nanostructures caused by the formation of charge transfer complex between PDMAEMA and C60. A well-defined poly(di(ethylene glycol) methyl ether methacrylate–stat-oligo(ethylene glycol) methyl ether methacrylate)-block-poly(di(ethylene glycol) methyl ether methacrylate ((PMEO2MA-stat-POEGMA300)-b-PMEO2) was successfully synthesized at room temperature via a two-step ATRP process. The block copolymer exhibited two thermal transitions at ~ 30 and 45 oC, which was believed to be associated with the formation of micelles and larger aggregates. The Rh of the aggregates increased from 47 to 90 nm, the aggregation number increased from 76 to ~9800 and Rg/Rh increased from 0.75 to 1.2 within the temperature range of 34 to 45oC. Well-defined statistical (PMEO2MA-stat-POEGMA300)-C60 was synthesized via ATRP and ATRA. The lower critical solution temperature (LCST) of (PMEO2MA-stat-POEGMA300)-C60 increased with methanol content in water, exhibiting lower LCSTs than PMEO2MA-stat-POEGMA300 for all methanol/water compositions. Higher critical micelle concentration (CMC) and larger spherical micelles were observed for (PMEO2MA-stat-POEGMA300)-C60 with increasing methanol content. The Rh of the micelles remained constant at temperature below the LCST and increased dramatically at temperature greater than the LCST, and (Rg/Rh) increased from ~ 0.75 to ~ 1.0. Nanocrystalline cellulose (NCC) was modified with water-soluble C60-(β-cyclodextrin) and (PMEO2MA-stat-POEGMA300)-C60) through a radical coupling reaction. NCC-C60-(PMEO2MA-stat-POEGMA300) possessed thermal responsive behavior in water and ~3.5 oC hysteresis associated with the heating/cooling cycles. No observable damage to NCC occurred during the radical coupling reaction as determined by TEM. NCC-C60-(β-cyclodextrin) possessed a similar thermal degradation behavior as NCC except it possessed a broader temperature range. Both NCC-fullerene systems demonstrated a radical scavenging activity when screened with the 2,2-diphenyl-1-picrylhydrazyl (DPPH). In addition, the drug loading and delivery using PDMAEMA-C60 with targeting moieties was explored. Two model drugs, namely fluorescein and pyrene were employed to evaluate the location of drug in the self-assembled structure of PDMAEMA-C60. It was found that the hydrophobic drugs were partitioned between the PDMAEMA shells and the hydrophobic fullerene cores. The drug delivery profiles indicated that PDMAEMA-C60 is an efficient drug carrier, however, it was cytotoxic to cells. The gene transfection efficacy of PDMAEMA-C60 to different cell lines was investigated and the results demonstrated that PDMAEMA-C60 exhibited good gene transfection performance. However, the targeting selectivity to liver cells cannot be determined in both cases. This study demonstrates that nanostructures of stimuli-responsive fullerene polymers can be controlled and manipulated by changing the external environments. Several potential applications, such as in drug and gene delivery, and free radical scavenging can be further explored.

The peak and valley problems caused by oral administration, injection or other conventional methods, call for developing systems that can deliver therapeutics more effectively. As one of the techniques, diffusion-controlled drug release membranes have significant interest due to great ease with which they can be designed to achieve near-zeroth-order release kinetics. Since diffusion is the rate-limiting step in these systems, determining the permeability and diffusivity of drug molecules in the membrane is therefore important in evaluating drug release performance. This study focuses on the Membrane Permeation Controlled Release (MPC) system, which involves a non-porous (dense) membrane, comprising of two biopolymers, sericin and chitosan. Ciprofloxacin hydrochloride and (+)-cis-diltiazem hydrochloride were used as hydrophilic model drugs, and nitro-2-furaldehyde semicarbazone (Nitrofurazon) was used as a hydrophobic model drug. Permeation experiments were carried out in a semi-infinite reservoir/receptor system to simulate in-vitro drug release. The intrinsic permeability and diffusivity (P, D) of the drugs through the membranes were determined using a modified time-lag method based on short time permeation and mass balance method based on long time permeation. The partition coefficients Kd of the drugs in the membranes and the swelling degree of the membranes were determined by sorption/desorption experiments. The diffusivities of the drugs were also determined from the sorption/desorption kinetics. Over the experimental ranges tested, the drug concentration and membrane cross-linking did not have significant effects on these parameters presumably due to the relatively low drug concentrations and mild crosslinkings of the membranes. The diffusivity coefficients of ciprofloxacin hydrochloride, (+)-cis-diltiazem hydrochloride and nitrofurazon in the membranes were found to be in the range of (2.0-2.6)×〖10〗^(-9)±2.6×〖10〗^(-10) cm2/s, (2.5-2.6) ×〖10〗^(-9)±1.1×〖10〗^(-10) and (38-134) ×〖10〗^(-9)±33.1×〖10〗^(-9) (cm2/s), respectively, and their permeability coefficients were in the range of (24-29)×〖10〗^(-8),(51-52) ×〖10〗^(-8) and (131-169) ×〖10〗^(-8) (cm2/s), respectively. The partition coefficients were determined to be around 0.91±0.21, 25±0.12 and 26±0.31, respectively. The diffusivity coefficients determined from sorption experiments for ciprofloxacin hydrochloride, diltiazem hydrochloride and nitrofurazon were found to be in the range of (3.2-7.6) ×〖10〗^(-9)±6.3×〖10〗^(-8), (6-10) ×〖10〗^(-9)±2.6×〖10〗^(-8) and (15-18) ×〖10〗^(-9)±2.7×〖10〗^(-7) (cm2/s), respectively. Also the diffusivity coefficients determined from sorption experiments for ciprofloxacin hydrochloride, diltiazem hydrochloride and nitrofurazon were in the range of (20-47) ×〖10〗^(-9), (12-24) ×〖10〗^(-9) and (11-20) ×〖10〗^(-9) (cm2/s), respectively. Nonetheless the differences in the diffusivities calculated from permeation and sorption/desorption experiments are considered to be acceptable, in view of the different experimental techniques used in this work, for the purpose of comparison of the membrane diffusivity and permeability.

Modeling interstitial fluid flow involves processes such as fluid diffusion, convective transport in the extracellular matrix, and extravasation from blood vessels. In all of these processes, computational fluid dynamics can play a crucial role in elucidating the mechanisms of fluid flow in solid tumors and surrounding tissues. To date, microvasculature flow modeling has been most extensively studied with simple tumor shapes and their capillaries at different levels and scales. With our proposed numerical model, however, more complex and realistic tumor shapes and capillary networks can be studied. First, a mathematical model of interstitial fluid flow is developed, based on the application of the governing equations for fluid flow, i.e., the conservation laws for mass and momentum, to physiological systems containing solid tumors. Simulations of interstitial fluid transport in a homogeneous solid tumor demonstrate that, in a uniformly perfused tumor, i.e., one with no necrotic region, the interstitial pressure distribution results in a non-uniform distribution of drug particles. Pressure distribution for different values of necrotic radii is examined, and two new parameters, the critical tumor radius and critical necrotic radius, are defined. In specific ranges of these critical dimensions the interstitial fluid pressure is relatively lower, which in turn leads to a diminished opposing force against drug movement and a subsequently higher drug concentration and potentially enhanced therapeutic effects. In this work, the numerical model of fluid flow in solid tumors is further developed to incorporate and investigate non-spherical tumor shapes such as prolate and oblate ones. Using this enhanced model, tumor shape and size effects on drug delivery to solid tumors are then studied. Based on the assumption that drug particles flow with the interstitial fluid, the pressure and velocity maps of the latter are used to illustrate the drug delivery pattern in a solid tumor. Additionally, the effects of the surface area per unit volume of the tissue, as well as vascular and interstitial hydraulic conductivity on drug delivery efficiency, are investigated. Using a tumor-induced microvasculature architecture instead of a uniform distribution of vessels provides a more realistic model of solid tumors. To this end, continuous and discrete mathematical models of angiogenesis were utilized to observe the effect of matrix density and matrix degrading enzymes on capillary network formation in solid tumors. Additionally, the interactions between matrix-degrading enzymes, the extracellular matrix and endothelial cells are mathematically modeled. Existing continuous and discrete models of angiogenesis were modified to impose the effect of matrix density on the solution. The imposition has been performed by a specific function in movement potential. Implementing realistic boundary and initial conditions showed that, unlike in previous models, the endothelial cells accelerate as they migrate toward the tumor. Now, the tumor-induced microvasculature network can be applied to the model developed in Chapters 2 and 3. Once the capillary network was set up, fluid flow in normal and cancerous tissues was numerically simulated under three conditions: constant and uniform distribution of intravascular pressure in the whole domain, a rigid vascular network, and an adaptable vascular network. First, governing equations of sprouting angiogenesis were implemented to specify the different domains for the network and interstitium. Governing equations for flow modeling were introduced for different domains. The conservation laws for mass and momentum, Darcy’s equation for tissue, and a simplified Navier Stokes equation for blood flow through capillaries were then used for simulating interstitial and intravascular flows. Finally, Starling’s law was used to close this system of equations and to couple the intravascular and extravascular flows. The non-continuous behavior of blood and the adaptability of capillary diameter to hemodynamics and metabolic stimuli were considered in blood flow simulations through a capillary network. This approach provided a more realistic capillary distribution network, very similar to that of the human body. This work describes the first study of flow modeling in solid tumors to realistically couple intravascular and extravascular flow through a network generated by sprouting angiogenesis, consisting of one parent vessel connected to the network. Other key factors incorporated in the model for the first time include capillary adaptation, non-continuous viscosity blood, and phase separation of blood flow in capillary bifurcation. Contrary to earlier studies which arbitrarily assumed veins and arteries to operate on opposite sides of a tumor network, the present approach requires the same vessel to run and from the network. Expanding the earlier models by introducing the outlined components was performed in order to achieve a more-realistic picture of blood flow through solid tumors. Results predict an almost doubled interstitial pressure and are in better agreement with human biology compared to the more simplified models generally in use today.

Chronic inflammatory diseases such as rheumatoid arthritis, type 2 diabetes and cardiovascular diseases, are associated with the homeostatic imbalance of one of several physiological systems combined with the lack of spontaneous remission, which causes the disease to persevere throughout patients’ lives. The inflammatory response relies mainly on tissue-resident, pro-inflammatory M1 type macrophages and, consequently, a chance for therapeutic intervention lies in driving macrophage polarization towards the anti-inflammatory M2 phenotype. Therefore, anti-inflammatory cytokines that promote M2 polarization, including interleukin-4 (IL4), have promising therapeutic potential. Unfortunately, their systemic use is hampered by a short serum half-life and dose-limiting toxicity. On the way towards cytokine therapies with superior safety and efficacy, this thesis is focused on designing bioresponsive delivery systems for the anti-inflammatory cytokine IL4. Chapter 1 describes how anti-inflammatory cytokines are tightly regulated in chronic, systemic inflammation as in rheumatoid arthritis but also in acute, local inflammation as in myocardial infarction. Both diseases show a characteristic progression during which anti-inflammatory cytokine delivery is of variable benefit. A conventional, passive drug delivery system is unlikely to release the cytokines such that the delivery matches the dynamic course of the (patho-)physiological progress. This chapter presents a blueprint for active drug delivery systems equipped with a 24/7 inflammation detector that continuously senses for matrix metalloproteinases (MMP) as surrogate markers of the disease progress and responds by releasing cytokines into the affected tissues at the right time and place. Because they are silent during phases of low disease activity, bioresponsive depots could be used to treat patients in asymptomatic states, as a preventive measure. The drug delivery system only gets activated during flares of inflammation, which are then immediately suppressed by the released cytokine drug and could prevent the steady damage of subclinical chronic inflammation, and therefore reduce hospitalization rates. In a first proof of concept study on controlled cytokine delivery (chapter 2), we developed IL4-decorated particles aiming at sustained and localized cytokine activity. Genetic code expansion was deployed to generate muteins with the IL4’s lysine 42 replaced by two different unnatural amino acids bearing a side chain suitable for click chemistry modification. The new IL4 muteins were thoroughly characterized to ensure proper folding and full bioactivity. Both muteins showed cell-stimulating ability and binding affinity to IL4 receptor alpha similar to those of wild type IL4. Copper-catalyzed (CuAAC) and strain-promoted (SPAAC) azide–alkyne cycloadditions were used to site-selectively anchor IL4 to agarose particles. These particles had sustained IL4 activity, as demonstrated by the induction of TF-1 cell proliferation and anti-inflammatory M2 polarization of M-CSF-generated human macrophages. This approach of site-directed IL4 anchoring on particles demonstrates that cytokine-functionalized particles can provide sustained and spatially controlled immune-modulating stimuli. The idea of a 24/7 sensing, MMP driven cytokine delivery system, as described in the introductory chapter, was applied in chapter 3. There, we simulated the natural process of cytokine storage in the extracellular matrix (ECM) by using an injectable solution of IL4 for depot formation by enzyme-catalyzed covalent attachment to ECM components such as fibronectin. The immobilized construct is meant to be cleaved from the ECM by matrix-metalloproteinases (MMPs) which are upregulated during flares of inflammation. These two functionalities are facilitated by a peptide containing two sequences: a protease-sensitive peptide linker (PSL) for MMP cleavage and a sequence for covalent attachment by activated human transglutaminase FXIIIa (TGase) included in the injection mix for co-administration. This peptide was site-selectively conjugated to the unnatural amino acid at IL4 position 42 allowing to preserve wild type bioactivity of IL4. In vitro experiments confirmed the anticipated MMP response towards the PSL and TGase-mediated construct attachment to fibronectin of the ECM. Furthermore, the IL4-peptide conjugates were able to reduce inflammation and protect non-load bearing cartilage along with the anterior cruciate ligament from degradation in an osteoarthritis model in rabbits. This represents the first step towards a minimally invasive treatment option using bioresponsive cytokine depots with potential clinical value for inflammatory conditions. One of the challenges with this approach was the production of the cytokine conjugate, with incorporation of the unnatural amino acid into IL4 being the main bottleneck. Therefore, in chapter 4, we designed a simplified version of this depot system by genetically fusing the bifunctional peptide via a flexible peptide spacer to murine IL4. While human IL4 loses its activity upon C-terminal elongation, murine IL4 is not affected by this modification. The produced murine IL4 fusion protein could be effectively bound to in vitro grown extracellular matrix in presence of TGase. Moreover, the protease-sensitive linker was selectively recognized and cleaved by MMPs, liberating intact and active IL4, although at a slower rate than expected. Murine IL4 offers the advantage to evaluate the bioresponsive cytokine depot in many available mouse models, which was so far not possible with human IL4 due to species selectivity. For murine IL4, the approach was further extended to systemic delivery in chapter 5. To increase the half-life and specifically target disease sites, we engineered a murine IL4 variant conjugated with a folate-bearing PEG chain for targeting of activated macrophages. The bioactive IL4 conjugate had a high serum stability and the PEGylation increased the half-life to 4 h in vivo. Surprisingly, the folate moiety did not improve targeting in an antigen-induced arthritis (AIA) mouse model. IL4-PEG performed better in targeting the inflamed joint, while IL4-PEG-folate showed stronger accumulation in the liver. Fortunately, the modular nature of the IL4 conjugate facilitates convenient adaption of PEG chain length and the targeting moiety to further improve the half-life and localization of the cytokine. In summary, this thesis describes a platform technology for the controlled release of cytokines in response to inflammation. By restricting the release of the therapeutic to the site of inflammation, the benefit-risk ratio of this potent class of biologics can be positively influenced. Future research will help to deepen our understanding of how to perfectly combine cytokine, protease-sensitive linker and immobilization tag or targeting moiety to tackle different diseases. / Chronische Entzündungskrankheiten wie rheumatoide Arthritis, Typ-2-Diabetes oder Herz-Kreislauf-Erkrankungen werden durch das Ungleichgewicht eines von mehreren physiologischen Systemen in Verbindung mit fehlender spontaner Remission verursacht, wodurch die Krankheiten lebenslang bestehen bleiben. Die zugrunde liegenden Entzündungsreaktionen beruhen hauptsächlich auf im Gewebe vorhandenen Makrophagen und deren Polarisation in Richtung des entzündlichen M1-Phänotyps, was gleichzeitig die Möglichkeit einer therapeutischen Intervention bietet. Entzündungshemmende Zytokine, einschließlich Interleukin-4 (IL4), haben ein großes therapeutisches Potenzial, da sie Makrophagen in Richtung des entzündungshemmenden M2-Phänotyps zu polarisieren vermögen. Leider ist ihre systemische Anwendung durch eine kurze Serumhalbwertszeit und dosislimitierende Toxizität eingeschränkt. Auf dem Weg zu Zytokintherapeutika mit verbesserter Sicherheit und Wirksamkeit konzentriert sich diese Arbeit auf die Entwicklung von bioresponsiven Freisetzungssystemen für das entzündungshemmende Zytokin IL4. Kapitel 1 beschreibt, wie entzündungshemmende Zytokine bei chronischen systemischen Entzündungen wie rheumatoider Arthritis im Vergleich zu akuten lokalen Entzündungen wie dem Myokardinfarkt reguliert werden. Beide Erkrankungen zeigen einen charakteristischen Verlauf, währenddessen die Freisetzung von entzündungshemmenden Zytokinen von unterschiedlich großem Nutzen ist. Gewöhnliche, passive Arzneimittelfreisetzungssysteme sind nicht in der Lage, Zytokine in idealer Menge zur optimalen Unterdrückung des dynamischen, (patho-)physiologischen Verlaufs der Krankheit freizusetzen. In diesem Kapitel werden deshalb aktive Arzneimittelfreisetzungssysteme vorgestellt, die mit einer Sensorik für die Entzündung ausgestattet sind, mit der sie kontinuierlich die Konzentration von Matrix-Metalloproteinasen (MMP) als Indikatoren für den Krankheitsverlauf erfassen können. Somit kann das aktive Arzneimittelfreisetzungssystem krankes Gewebe zum richtigen Zeitpunkt und am richtigen Ort mit Zytokinen behandeln. Solche bioresponsiven Depots können zur vorbeugenden Behandlung von asymptomatischen Patienten eingesetzt werden, da sie während Phasen geringer Krankheitsaktivität inaktiv sind. Das Freisetzungssystem wird erst durch Entzündungsschübe aktiviert, die dann sofort durch die freigesetzten Zytokine unterdrückt werden. Dadurch könnte die dauerhafte Schädigung durch subklinische, chronische Entzündung verhindert und als Konsequenz die Hospitalisierungsrate gesenkt werden. In einer ersten Machbarkeitsstudie wurden in Kapitel 2 IL4-dekorierte Partikel mit dem Ziel entwickelt, eine langanhaltende und lokalisierte Zytokinaktivität zu gewährleisten. Dazu wurden IL4-Muteine erzeugt, bei denen das Lysin 42 mittels Erweiterung des genetischen Codes durch zwei verschiedene unnatürliche Aminosäuren ersetzt wurde, die jeweils eine für Klick-Chemie geeignete Seitenkette tragen. Die IL4-Muteine wurden ausführlich charakterisiert, um eine korrekte Faltung und volle Bioaktivität sicherzustellen. Beide Muteine zeigten zellstimulierende Fähigkeit und Bindungsaffinität an IL4-Rezeptor-alpha, die mit der von Wildtyp-IL4 vergleichbar ist. Anschließend wurde kupferkatalysierte (CuAAC) und kupferfreie (SPAAC) Azid-Alkin-Cycloaddition verwendet, um IL4 ortsspezifisch auf Agarosepartikeln zu verankern. Die Partikel waren in der Lage, die IL4-Aktivität über längere Zeit aufrecht zu erhalten, was durch TF-1-Zellproliferation und M2-Polarisation von M-CSF-generierten, humanen Makrophagen gezeigt werden konnte. Dieser Ansatz der ortsspezifischen Verankerung von IL4 auf Agarosepartikeln zeigt, dass zytokinfunktionalisierte Partikel anhaltende und räumlich kontrollierte, immunmodulierende Stimuli liefern können. Die Idee eines MMP-gesteuerten Zytokinfreisetzungssystems mit 24/7-Sensorik, das im Einleitungskapitel vorgestellt wurde, wurde in Kapitel 3 umgesetzt. Der natürliche Prozess der Zytokinspeicherung in der extrazellulären Matrix (EZM) wurde mithilfe einer injizierbaren IL4-Lösung zur enzymatischen Depotbildung durch kovalente Bindung an EZM-Komponenten, z. B. Fibronektin, simuliert. Nach der Bindung soll das Konstrukt durch Matrix-Metalloproteinasen (MMPs), die während Entzündungsschüben hochreguliert werden, aus der EZM freigesetzt werden können. Eine Peptidsequenz, die ein Protease-sensitives Verbindungsstück und eine Sequenz, mit der das Zytokin bei gleichzeitiger Injektion von aktivierter menschlicher Transglutaminase FXIIIa (TGase) kovalent auf der EZM immobilisiert wird enthält, wurde ortsspezifisch über eine unnatürliche Aminosäure an Position 42 von IL4 gekoppelt. Dadurch wurde die Bioaktivität von IL4 vollständig erhalten, während das Protease-sensitive Verbindungsstück auf MMPs reagierte und das Konstrukt durch TGase an das Fibronektin der EZM gebunden werden konnte. Die IL4-Peptid-Konjugate waren in einem Osteoarthritis-Modell bei Kaninchen in der Lage, die Entzündung des Kniegelenks zu verringern und den nicht-tragenden Knorpel sowie das vordere Kreuzband vor Degradation zu schützen. Dies ist der erste Schritt in Richtung einer minimalinvasiven Behandlung durch Verwendung von bioresponsiven Zytokindepots mit potenziellem klinischem Nutzen bei Entzündungserkrankungen. Eine der Herausforderungen bei diesem Vorgehen war die Herstellung der Zytokinkonjugate, wobei der Einbau der unnatürlichen Aminosäure in IL4 den größten Engpass darstellte. Deshalb wurde in Kapitel 4 eine vereinfachte Version dieses Depotsystems entworfen, indem das bifunktionelle Peptid über eine flexible Verbindungssequenz mit murinem IL4 genetisch fusioniert wurde. Während humanes IL4 bei C-terminaler Verlängerung an Aktivität verliert, ist murines IL4 durch die Modifikation nicht beeinflusst. Die murinen IL4-Fusionsproteine konnten in Gegenwart von TGase wirksam an in vitro generierte extrazelluläre Matrix gebunden werden. Darüber hinaus wurde das Protease-sensitive Verbindungsstück selektiv von MMPs erkannt und gespalten, wobei intaktes und aktives IL4 freigesetzt wurde, wenn auch mit einer langsameren Rate als erwartet. Murines IL4 bietet die Möglichkeit das bioresponsive Zytokindepot in den vielen verfügbaren Mausmodellen zu testen, was mit humanem IL4 aufgrund der Speziesselektivität nicht möglich ist. Für murines IL4 wurde die Entwicklung in Kapitel 5 auf die systemische Applikation ausgeweitet. Um die Serumhalbwertszeit zu erhöhen und eine Wirkstofflokalisation im entzündeten Gewebe zu erreichen, wurde eine murine IL4-Variante entwickelt, die mit einer Folat-tragenden PEG-Kette konjugiert wurde, um aktivierte M1 Makrophagen zu adressieren. Das bioaktive IL4-Konjugat wies eine hohe Serumstabilität auf und die PEGylierung erhöhte die Halbwertszeit in vivo auf 4 h. Allerdings konnte durch die Konjugation der Folatgruppe an IL4 die Wirkstofflokalisation in einem Mausmodell mit Antigen-induzierter Arthritis (AIA) nicht verbessert werden. IL4-PEG akkumulierte sich stärker im entzündeten Gelenk, während IL4-PEG-Folat eine stärkere Anreicherung in der Leber zeigte. Erfreulicherweise erleichtert der modulare Aufbau des IL4-Konjugats die bequeme Anpassung der PEG-Kettenlänge und der zielorientierten Einheit, um die Halbwertszeit und Lokalisierung des Zytokins weiter zu verbessern. Zusammenfassend beschreibt diese Arbeit eine Plattformtechnologie zur kontrollierten Freisetzung von Zytokinen als Reaktion auf Entzündungen. Durch die Beschränkung der Freisetzung des Therapeutikums auf den Ort der Entzündung kann das Nutzen-Risiko-Verhältnis dieser potenten Klasse von Biologika positiv beeinflusst werden. Zukünftige Forschungen werden dazu beitragen zu verstehen, wie Zytokin, Protease-sensitives Verbindungsstück und Immobilisierungsanhängsel oder etwaige zielorientierte Einheiten zur Bekämpfung verschiedener Krankheiten perfekt kombiniert werden können.

Since its discovery 15 years ago there has been an explosion of research in the field of DNA immunisation. Unfortunately despite early promises that DNA immunisation had the potential to cure almost any infectious disease, autoimmune disease or even cancer, progress towards clinical trials has been slow. This has been due in part to the huge range of permutations possible in delivering the DNA. One approach is to deliver the DNA by gene gun. Gene gun delivery is a very efficient way of transfecting cells however also has a number of possible disadvantages. These drawbacks include a weak immunogenicity in larger animals as well as the tendency to bias towards the development of a strong type 2 response. In an effort to enhance antigen-specific immune responses and counter the type 2 polarisation of gene gun delivery, a series of DNA vaccines were created where the extracellular portion of the hemagglutinin (HA) gene from influenza A/PR8/34 virus was genetically fused the type 1 cytokines IFNγ, IL-12 and IL-23. Interleukin-23 has been recently discovered and even though both IL-12 and IL-23 contain the p40 subunit they seem to have dissimilar functions. The vaccine constructs were first tested in cellular assays in vitro to ensure correct production and biological activity of the attached cytokines. They were then delivered in various combinations to groups of BALB/c mice to test development of immune responses and the effect of different delivery regimes. Finally mice were immunised then challenged with live influenza virus to determine the different DNA vaccines� protective efficacy. DNA vaccines containing the HA gene alone (pHA) or fused to IFNγ (pIFNγHA), IL-12 (pIL-12HA) or IL-23 (pIL-23HA) were successfully constructed. The fusion of the HA gene to the genes for IFNγ, IL-12 or IL-23 did not significantly disturb the structure of the antigen or prevent the biological actions of the cytokines. Mice immunised three times with pHA had high titres of serum IgG1 antibody and their splenocytes produced approximately equal amounts of IFNγ and IL-5. Co-delivery of IFNγ was unable to alter immune responses regardless of whether it was delivered at the first, last or during all immunisations. Surprisingly co-delivery of IL-12 acted to suppress both antibody and cellular immune responses, possibly through an IFNγ/nitric oxide feedback loop. On the other hand co-delivery of IL-23 tended to enhanced immune responses and, while it did not significantly alter the type 1 to type 2 balance, it was able to increase the ability of mice to clear live influenza virus from their lungs when they were challenged 26 weeks after immunisation. This protection was associated with increased levels of neutralising antibody in the serum of pIL-23HA immunised mice. This research has illuminated several of the pitfalls in the development of DNA vaccines and the use of cytokine as adjuvants. However it has also broadened our understanding of IL-23 and implies that IL-23 could be effectively used to increase the development of longterm immunity after immunisation.

Purpose: This thesis aimed to gain a better understanding of the effects of salts in modifying in vitro phase behaviour of lecithin and cholesterol solid implants and to obtain further information on in vitro protein release and stability. Methods: Raman spectroscopy and partial least squares regression (PLSR) were used to investigate lecithin-cholesterol molecular interactions as a function of method of preparation. Lipid-salt interactions were studied by attenuated total reflectance Fourier transform infrared (ATR-FTIR) and Raman spectroscopy using principal component analysis (PCA). In vitro release of bovine serum albumin (BSA), a model protein, from lecithin and lecithin:cholesterol implants comprising 10 and 30% NaCl and CaCl₂ were performed. Size exclusion (SE) HPLC was used for quantitative and qualitative analysis of the released BSA. On hydration, changes in phase behaviour and implant morphology were studied by ATR spectroscopy and light microscopy. SE-HPLC, ATR and fluorescence spectroscopy were used to evaluate the structure of unreleased BSA. Protein adsorption on lipid films was studied by flow through ATR spectroscopy. Increased amide II peak area upon recirculation of BSA in salt solutions over hydrated lecithin and lecithin:cholesterol films cast on ZnSe prisms was used to quantify the deposition of BSA onto the lipid surfaces. Results: Shifts in the Raman spectra suggested the lecithin headgroup may be involved in lecithin-cholesterol interactions. Greater R� and root mean square error of cross validation in the calibration curves of physical mixing and heating (120�C) methods reflected poor mixing in these preparations. The mean absolute residue and mean Mahalanobis distance values from the physical mixing and granulation methods indicated their spectral similarity and comparable level of lecithin-cholesterol interactions. Calcium exhibited stronger affinity for phospholipids than sodium and it induced headgroup hydration and reorganisation upon binding. PCA of ATR spectra was sensitive to cholesterol addition, calcium binding and method of preparation whilst PCA of Raman spectra only differentiated the presence of cholesterol. In vitro release of BSA from implants produced from wet granulation mixtures of lecithin and lecithin:cholesterol in the absence of salt showed retention of a high monomer content and the release profiles were similar to the literature. Cholesterol increased the swelling, induced phase transformation of lecithin and, subsequently, reduced the BSA release. Salts only slightly modified the BSA release from the lecithin implants. In contrast, for lecithin:cholesterol matrices salts greatly enhanced implant swelling, induced the formation of hydrated lecithin of heterogeneous size and inhibited the in vitro BSA release. Analyses of the protein showed increased aggregation of BSA with a high retention of native structure while retained within the swollen matrices. ATR spectra suggested that salts promoted protein adsorption onto hydrated lecithin surfaces and the effects depend on salt types (NaCl > CaCl₂) and concentration (0.1 M > 1.0 M) but not on lecithin:cholesterol surfaces. Conclusion: PLSR and PCA can be used to investigate molecular interactions in the solid lipid matrices. In lecithin:cholesterol implants, salts modified the phase behaviour of lecithin which resulted in enhanced swelling, formation of hydrated lecithin of altered morphology and inhibition of in vitro BSA release.

Purpose: The aim of this thesis was to investigate a suitable formulation for the topical delivery of growth factors to chronic wounds, and then to determine the concentrations reached within an animal wound model. A secondary aim was to determine if the chosen growth factor was present at levels able to stimulate the production of other cytokines, specifically IL-1β and MCP-1. Methods: An in vitro testing apparatus was designed and made and the release of model actives [bromophenol blue (BPB) and horseradish peroxidase (HRP)] from gels and films of hydroxylpropylmethylcellulose (HPMC) (E4M CR, K4M CR and E10M CR) was determined. In this study, Fibroblast Growth Factor -2 (FGF-2) (0.3 [mu]g) was incorporated into three formulations (solution, gel and dried gel film on Melolin[TM] backing) and together with a control formulation were administered to punch biopsy wounds in rats. The in vivo release was followed over three time periods (two, five and eight hours) and the amount of FGF-2 at various wound depths was quantified by ELISA. Two biological markers IL-1β and MCP-1 were quantified using ELISA. The FGF-2 was additionally tagged with a fluorescent dye so that visualisation of the penetration could be obtained via confocal microsopy. Results: For the HPMC gels, the more viscous gel (E10M) provided a greater diffusional barrier and slowed the release of BPB (12 � 3.5 [mu]g/min compared with 16 � 1.7 [mu]g/min and 18 � 1.4 [mu]g/min for K4M and E4M respectively). However, when HPMC was formulated as a dried film a burst release was seen and release of BPB was slowest from the more rapidly hydrating K4M. With the larger model active HRP, there was a slower diffusion through the gel barrier formed upon film hydration, such that time of 100% release was up to 300 minutes compared to approximately 60 minutes for BPB. When the film was dried onto a supportive backing, the initial burst release was minimised as the film did not break apart on contact with the wound, and hence film integrity was maintained and release prolonged. The in vivo studies showed that, two hours after application, the highest concentration of FGF-2 was seen in the surface granulation tissue of rats that received the solution formulation (2280 � 790 pg/g). The concentration decreased with increasing tissue depth but was significantly greater than the saline control in the surface granulation and subcutaneous fat layers (p<0.05). Tissue concentrations following application of the gel and film formulations were only marginally greater than control in the surface granulation layer. After eight hours, rats that received the solution retained elevated surface tissue concentrations (surface granulation and subcutaneous fat) of FGF-2. Repeated measures ANOVA using a general linear model showed statistically significant differences in the mean FGF-2 level with respect to formulation and length of time of application of the formulation (p<0.05). In terms of other cytokines, there was a release of both IL-1β and MCP-1 in all groups, immediately post-wounding, probably in response to cellular damage. After eight hours, the film formulation appeared to elevate IL-1β levels which may be useful to drive wound healing. Confocal microscopy images showed diffuse distribution of FGF-2 eight hours after application of the solution formulation after eight hours and that with the gel formulation FGF-2 initially aggregated at the wound surface. Conclusion: In vitro experiments investigating the effect of hydration rate and viscosity of HPLC polymers allowed a formulation to be chosen for further in vivo study. Elevated FGF-2 could be measured in superficial wound tissues up to eight hours after application of a solution. However, application of a comparable amount of FGF-2 in HPMC gels or films did not produce appreciable elevations in FGF-2 in wound tissues, although confocal microscopy indicated the penetration of FGF-2 into the wound for up to eight hours.

Pulmonary and oral drug administrations are usually the preferred methods of delivery of active pharmaceutical ingredients.Generally,pulmonary drug formulations are more attractive compared to oral formulations since they consist of micron-sized powders with high surface area thus having faster onset of action,as well as minimizing the drug dosage and side effects.Oral insulin formulations,if achievable,would provide an alternative to injectable insulin,as the common drawbacks of injectable insulin are the multiple daily injections and the possibility of skin infections at the injection site. In this study,the feasibility of using dense gas particle processing techniques known as the Aerosol Solvent Extraction System (ASES),Gas Anti-Solvent (GAS)and High-Pressure Media Milling (HPMM)for pharmaceutical processing was assessed.The ASEStechnique,utilizing dense ethane,was employed to prepare insulin-lactose formulations for pulmonary administration whilst the GAS and ASES techniques,utilizing dense CO2,were employed to prepare microencapsulated formulations containing insulin and Eudragit?? S100 for oral administration.Furthermore,the HPMM technique,utilizing dense hydrofluocarbon (HFC)134a/227ea,was employed to prepare suspension Metered Dose Inhaler (MDI)formulations containing budesonide and various surfactants. The Fine Particle Fraction (FPF)of processed insulin without the presence of lactose was found to be 44%.In other words,44% of processed insulin delivered to the impactor stages (excluding the throat and neck)has aerodynamic diameter of less than 5??m.With the addition of lactose as carrier,the FPFof the insulin-lactose (1:1w/w)formulation increased to 64%.The increase in FPFwas attributed to the lower density of lactose particles compared to that of insulin particles to produce an intimate mixture with enhanced powder flowability and aerodynamic performance. Proteins for oral delivery should ideally be formulated with acid-resistant polymer as a protective coating to protect against enzymatic degradation in the stomach.Eudragit?? S100,which is insoluble or almost impermeable at pH 1-4and soluble at pH 5-7,was used to prepare oral insulin formulations.The insulin release at pH 3was sustained by the Eudragit?? S100coating and the encapsulation efficiency of insulin??Eudragit?? S100formulations varied between 6% and 24% depending on the initial drug to polymer ratio. One of the major therapies utilizing metered dose inhaler formulations in the treatment of asthma has been studied using the HPMM process.The HPMM process has been demonstrated to be an efficient milling process for the enhancement of the physical stability and aerodynamic performance of budesonide in HFC-134a/227ea propellant formulations.No significant change in physical stability was observed in the formulations for 2 weeks.