Lipid implants for controlled release of proteins

In recent years, we have seen significant progress towards understanding the role of proteins in both physiological and pathological processes. Especially neurotrophic factors have increasingly come into focus for the treatment of neurodegenerative diseases, however still no satisfying delivery strategies for long-term applications have been established. Most proteins possess a hydrophilic character and can therefore be retarded by encapsulation into more hydrophobic matrix materials. It was the aim of this thesis to establish physiological triglyceride matrices as a suitable alternative to commonly used polymeric carriers for the long-term controlled release of protein drugs. The success in comprehensively achieving this goal was to a great extent based on the parallel investigation into both mechanistic and application based aspects. Visualization of internal processes inside the matrices at all stages by confocal microscopy played a major role in advancing both fields of work and thoroughly understanding the system. First, in a comparison with the so far used manufacturing methods, PEG co lyophilization was developed as a new strategy with superior performance in finely and homogenously distributing lysozyme as a model protein in lipid matrix material with complete retention of activity. The method was designed as a one-pot procedure, where protein was micronized in a first step due to freezing�induced aqueous phase separation in the presence of PEG, and then dispersed in the solid state in a solution of the lipid in an organic solvent. Confocal microscopy pointed to the fine and homogenous distribution being the crucial prerequisite for long-term delivery. With this, a reduction in matrix size to dimensions of 1mm diameter and 1mm height was feasible, for the first time enabling their investigation in the animal brain, where they showed excellent biocompatibility. BDNF and IL-18 were incorporated by this method into glyceryl tripalmitate under complete retention of activity as demonstrated by BDNF-ELISA and a cell based IL 18 bioassay. A thorough understanding of the mechanisms underlying release from a drug delivery device is crucial for developing strategies to tailor the release profile and being able to react to drug stability problems. It could be demonstrated by diffusion studies of fluorescently labelled BSA, excipient PEG and release buffer visualized with the help of confocal microscopy that buffer could only penetrate into the matrices, where excipient or protein starts to dissolve and diffuses out. A linear concentration gradient of diffusing protein was formed between a buffer penetration front and the matrix surface until the buffer reached the centre of the device. The power law was an adequate mathematical model for the description of release profiles of FITC-BSA and lysozyme at loadings between 1 and 8%, with data being linear to time0.45 for up to 60% cumulative release. At 3-5% protein loading, a percolation threshold could be identified, predicting the loading, above which complete release of the matrices is possible due to the formation of a continuous network of pores. Buffer penetration and drug diffusion, the two crucial mechanisms for release identified by confocal microscopy, were further investigated for their dependence on matrix and drug properties. It could be demonstrated, that wettability of the matrix governed the buffer penetration velocity. This could be employed to tailor the release profile in a time range of more than 14months by varying matrix lipophilicity through changing the chain length of the fatty acid in the triglyceride. However, this was only possible when the incorporated drug was a protein, being able to reduce buffer contact angles via its own surface active properties. Non-surface active model substances, such as FITC-dextrans remained trapped within the triglyceride, due to their inability to induce wetting and buffer penetration. This, however, could be triggered by the addition of the surfactants Tween®20 and 81 at time points up to 57days after initial incubation, leading to a complete release of the matrix loading. Apart from drug surface activity, another important characteristic was its molecular weight, being decisive for the diffusion step during release. An impact could be proven not only for seven different model proteins (lysozyme, trypsin, ovalbumin, BSA, ADH, catalase, thyroglobulin) but also for FITC-dextrans of different molecular weights in the presence of surfactant in the release buffer. As a conclusion, with the work described in this thesis, lipid matrices could be established as a controlled release systems for long-term delivery of protein drugs. Concomitantly, it presents a fundamental insight into both release mechanisms and potential applications. It thus can provide a basis for a better understanding also for other kinds of matrix geometries and lipophilic materials, for which diffusion is the major release controlling factor