DNA vaccines offer unique promise as a means of generating immunity against infectious and malignant disease. Unfortunately a number of obstacles, including rapid degradation of naked plasmid DNA (pDNA), poor cellular uptake by antigen presenting cells (APCs) and subsequent low levels of gene expression have limited the ability of DNA vaccines to raise sufficient immune responses towards the target antigen. This thesis is focused on investigating the immunostimulatory potential of liposomal nanoparticulate (LN) formulations of pDNA (stabilized plasmid lipid particles; SPLP) and cytosine-guanine oligodeoxynucleotides (CpG-ODN; LN CpG-ODN), and examining their ability to act together as a non-viral DNA vaccine in attempt to address the shortcomings of current DNA vaccine approaches. One focus of this thesis concerns investigating the immunostimulatory activity of LN formulations of CpG-ODN and pDNA. It is shown that despite dramatic differences in pharmacokinetics and biodistribution of LN CpG-ODN following intravenous (i.v.) and subcutaneous (s.c.) administration the resultant immune response is very similar, which is concluded to be due to the intrinsic ability of APCs to sequester LN CpG- ODN. In addition, it is demonstrated that lipid encapsulation dramatically enhances the immunostimulatory potential of pDNA and it is observed that SPLP maintains immunostimulatory activity in Toll-like receptor 9 (TLR9) knock-out mice. Together theses findings highlight the need for DNA-based therapies to consider both TLR9-dependent and -independent immunostimulatory activities of pDNA when constructing non-viral vectors. Furthermore, a new role for SPLP as a non-viral gene delivery vehicle for the generation of a systemically administered genetic vaccine in the presence of LN CpG-ODN is introduced. The ability of vaccination with SPLP to act prophylactically, to protect mice from tumour challenge, and therapeutically, in a novel vaccination strategy where the antigen is expressed at the tumour site as a result of SPLP-mediated transfection, is explored, demonstrating that in the presence of LN CpG-ODN SPLP possesses potential as a non-viral delivery system for DNA-based cancer vaccines. In summary, this work represents a substantial advance in the understanding of the immunostimulatory potential of both SPLP and LN CpG-ODN and provides insight into their ability to work together as a non-viral DNA vaccine. / Medicine, Faculty of / Biochemistry and Molecular Biology, Department of / Graduate

One of the intriguing features of biological systems is the prevalence of highly selective and often very strong interactions among different cellular components. Such interactions play a variety of organizational, mechanical, and physiological roles at the cellular and organism levels. Antigen-antibody complexes are representative examples of highly selective and potent interactions involving proteins. The marked specificity of protein-antibody complexes have led to a wide range of applications in cellular and molecular biology related research. They have become an integral research tool in the present genomic and proteomic era. Unfortunately, the production of selective tools based on antigen-antibody interactions requires cumbersome protocols. The long term goal of this project explores the possibility of manipulating liposomes to serve as the chemical receptors ("artificial antibodies") against selected proteins. Cellular lipids (e.g., lipid rafts) are known to facilitate highly selective binding of proteins on cell membranes. The binding of proteins to cell membranes can be envisaged to be modulated via interactions between polar (charged) and non-polar head groups of lipids and the complementary amino acid residues of proteins. Their interaction is facilitated by a combination of van der Waals, electrostatic, hydrogen bonding and hydrophobic forces. A further interesting aspect of the above interaction is the "fluidity" of the membrane resident lipids, which can migrate from other regions to further enhance the complementary interactions of proteins on the initially "docked" membrane surface. With these features in mind, the end goal of this project is expected to deliver lipid-based chemical receptors "synthetically" designed against proteins to function as "artificial antibodies". Protein sensing will be accomplished with lipid receptors assembled in templated polymerized liposomes. The research presented here specifically focus on the analytical aspects of protein sensing via polymerized liposome vesicles. Lanthanide ions (Eu(III) and Tb(III)) are incorporated into polymerized liposome with the expectation to "report" quantitative and qualitative information on the interacting protein. Our proposition is to extract quantitative and qualitative information from the luminescence intensity and the luminescence lifetime of the lanthanide ion, respectively. A thorough investigation is presented regarding the analytical potential of these two parameters for protein sensing. Two chemometic approaches - namely partial least squares (PLS-1) and artificial neural networks (ANN) - are compared towards quantitative and qualitative analysis of proteins in binary mixtures.

One of the intriguing features of biological systems is the prevalence of highly selective and often very strong interactions among different cellular components. Such interactions play a variety of organizational, mechanical, and physiological roles at the cellular and organism levels. Antigen-antibody complexes are representative examples of highly selective and potent interactions involving proteins. The marked specificity of protein-antibody complexes have led to a wide range of applications in cellular and molecular biology related research. They have become an integral research tool in the present genomic and proteomic era. Unfortunately, the production of selective tools based on antigen-antibody interactions requires cumbersome protocols. The long term goal of this project explores the possibility of manipulating liposomes to serve as the chemical receptors ("artificial antibodies") against selected proteins. Cellular lipids (e.g., lipid rafts) are known to facilitate highly selective binding of proteins on cell membranes. The binding of proteins to cell membranes can be envisaged to be modulated via interactions between polar (charged) and non-polar head groups of lipids and the complementary amino acid residues of proteins. Their interaction is facilitated by a combination of van der Waals, electrostatic, hydrogen bonding and hydrophobic forces. A further interesting aspect of the above interaction is the "fluidity" of the membrane resident lipids, which can migrate from other regions to further enhance the complementary interactions of proteins on the initially "docked" membrane surface. With these features in mind, the end goal of this project is expected to deliver lipid-based chemical receptors "synthetically" designed against proteins to function as "artificial antibodies". Protein sensing will be accomplished with lipid receptors assembled in templated polymerized liposomes. The research presented here specifically focus on the analytical aspects of protein sensing via polymerized liposome vesicles. Lanthanide ions (Eu3+ and Tb3+) are incorporated into polymerized liposome with the expectation to "report" quantitative and qualitative information on the interacting protein. Our proposition is to extract quantitative and qualitative information from the luminescence intensity and the luminescence lifetime of the lanthanide ion, respectively. A thorough investigation is presented regarding the analytical potential of these two parameters for protein sensing. Two chemometic approaches - namely partial least squares (PLS-1) and artificial neural networks (ANN) - are compared towards quantitative and qualitative analysis of proteins in binary mixtures.

Ce travail, qui fait partie d’un projet européen, « NANOTHER », est focalisé sur la fonctionnalisation de nanoparticules polymériques et lipidiques fonctionnalisées par des anticorps Herceptine® pour cibler des cellules du cancer du sein HER2+. Deux stratégies de fonctionnalisation ont été étudiées : une a reposé sur l’utilisation de protéines de fusion, l’Anx5-ZZ, composée d’Annexine A5 et deux domaines Z homologues de la protéine A de Staphylococcus aureus qui peuvent se fixer des anticorps d’une manière orientée par leur fragment cristallisable ; l’autre a porté sur le couplage direct d’anticorps modifiés pour exposer des groupes sulfhydryles aux nanoparticules exposant des groupes maléimides.La première partie concerne le développement d’un agent de ciblage simplifié du complexe l’Anx5-ZZ-anticorps, à savoir l’Anx5-scFv (single-chain variable fragment). Puisque la cible n’avait pas été décidée au début de ce travail, deux scFvs ont été utilisé comme système modèle. L’expression de protéines de fusion a été essayée chez Escherichia Coli avec différentes constructions de protéines de fusion, différentes conditions d’expression et différentes souches bactériennes. Toutes les protéines sont soient agrégées soient non surexprimées.La deuxième partie consiste à fonctionnaliser les polymersomes par l’Herceptine® via l’Anx5-ZZ. D’abord, nous avons validé une méthode de modification de la surface de polymersome pour présenter des groupes maléimides. Ensuite, le couplage covalent de l’Anx5(SH)-ZZ aux polymersomes-maléimide a été réalisé et quantifié. Nous avons obtenu maximum 30 Anx5-ZZ par polymersome. Puis, la liaison d’affinité d’anticorps aux polymersomes-Anx5-ZZ a été caractérisée, réalisée et quantifiée. Pour 30 Anx5-ZZ par polymersome, nous avons 60 Herceptine® par polymersome. Cependant, l’efficacité de ciblage de ces systèmes est très faible.La troisième partie consiste à fonctionnaliser les liposomes par l’Herceptine® via couplage direct. Tout d’abord, la modification de l’Herceptine® pour présenter des groupes SH a été caractérisée et contrôlée. Ensuite, le couplage covalent d’Herceptine®-SH aux liposomes-maléimides a été réalisé et quantifié. L’étude de ciblage montre que les liposomes fonctionnalisés par une molécule d’Herceptine® sont capable de cibler les cellules HER2+. / This work, which is part of a European project "NANOTHER", focus on the functionalization of polymeric and lipidic nanoparticles by Herceptin® to target HER2+ cancer cells. Two functionalization strategies were studied: one was based on the use of a fusion protein, Anx5-ZZ, composed of Annexin A5 and two Z domains which are homologous with the protein A of Staphylococcus aureus that can bind antibodies by their crystallizable fragment in a oriented way; the other focused on the direct coupling of modified antibodies exposing sulfhydryl groups to nanoparticles exposing maleimid groups.The first part concerns the development of a targeting agent simplified from the Anx5-ZZ-antibody complex, namely Anx5-scFv (single-chain variable fragment). Since the target had not been decided at the beginning of this work, two scFvs were used as model system. The expression of fusion proteins was tested in Escherichia coli with different fusion protein constructions, different expression conditions and different bacterial strains. All proteins are either aggregated or non-overexpressed.The second part is to functionalize the polymersomes by Herceptin® via Anx5-ZZ. First, we validated a method for modifying the surface of polymersome to expose maleimid groups. Then, the covalent coupling of Anx5(SH)-ZZ to polymersomes-maleimid was performed and quantified. We obtained maximum 30 Anx5-ZZ per polymersome. Then, the affinity binding of antibodies to polymersomes-Anx5-ZZ was characterized, performed and quantified. For 30 Anx5-ZZ per polymersome, we have 60 Herceptin® per polymersome. However, the targeting efficiency of this system is very low.The third part consists in functionalizing the liposomes by Herceptin® via direct coupling. Firstly, the modification of Herceptin® to expose SH groups was characterized and controlled. Then, the covalent coupling of Herceptin®-SH to liposomes exposing maleimid groups was performed and quantified. The targeting study shows that liposomes functionalized with one Herceptin® are able to target HER2+ cells.

Les liposomes d’ATP incluant des ligands hépatiques pourraient contribuer à améliorer le statut énergétique du greffon hépatique. Une première phase de développement a mis en évidence la nécessité de stabiliser la forme (i) et de valider un modèle cellulaire à statut énergétique altéré (ii) afin de pouvoir tester différentes formulations liposomales d’intérêt. Afin de résoudre la première problématique, différentes stratégies ont été mises en place lors de la lyophilisation de liposomes blancs ou chargés en ATP. Le saccharose et le tréhalose ont conduit à une stabilisation de la forme avant et après lyophilisation. Néanmoins, quel que soit le procédé, la lyophilisation s’est accompagnée d’une fuite en ATP. L’ajout d’un agent inhibant la cartinine palmitoyl transferase, l’Etomoxir, s’est révélé un modèle intéressant pour moduler le niveau énergétique des HepG2 en fonction de la température et de la concentration utilisée. En conclusion, ce travail contribue à mieux comprendre les facteurs critiques liés à la lyophilisation des liposomes (i) et décrit des modèles cellulaires hépatiques à statut énergétique altéré qui pourraient être mis à profit pour tester différents principes actifs ou formes galéniques innovantes. / ATP liposome incorporating hepatic ligands may contribute to improve the energetic status of the liver graft. In a first phase of development, it has been emphasized the great need of stabilizing the liposome (i) and of validating a cellular model with an altered energetic status in order to test the formulations of interest. To provide a stable liposomal preparation, different strategies have been carried out to freeze-dry liposome with or without ATP. Sucrose and trehalose better stabilize the liposome preparation during the freeze-drying process. Nevertheless, in all conditions, a significant ATP leakage has been observed. An inhibitor of the cartinine palmitoyl transferase (i.e. Etomoxir) has shown great interest to modulate the energetic level in HepG2 cells by varying Etomoxir concentration and culture temperature. In conclusion, this study contributes to a better understanding of the critical factors related to liposome freeze-drying and describes hepatic cell models with altered energetic status that may have great interest to test specific agent or innovating formulations.

Les systèmes théranostiques associant des propriétés thérapeutiques et des propriétés d'imagerie sont développés pour permettre le suivi de traitement in vivo. La stratégie que nous proposons dans cette thèse est de formuler des liposomes magnétiques thermosensibles chargés en principe actif pour traiter des tumeurs murines superficielles du côlon CT26. Ces nanovecteurs peuvent être accumulés dans les tumeurs par ciblage magnétique et le relargage du principe actif peut être déclenché par HIFU (Ultrasons Focalisés de Haute Intensité). Ces liposomes sont développés par co-encapsulation de nanoparticules de maghémite (?-Fe2O3) et de Combrétastatine A4 Phosphate (CA4P) dans des liposomes thermosensibles. La forte encapsulation des nanoparticules magnétiques dans les Liposomes Ultra-Magnétiques (LUM) permet à la fois le ciblage magnétique et leur suivi in vivo par IRM. Le chauffage par HIFU in vitro à la température de transition des membranes des LUM a permis une amélioration du relargage de la CA4P. La biodistribution des LUM in vivo a été évaluée par IRM dynamique de contraste de susceptibilité avec une résolution temporelle adaptée à la cinétique de capture des nanovecteurs, notamment par le foie. L'efficacité du ciblage magnétique a été démontrée grâce à une nouvelle méthode de traitement de l'histogramme des intensités IRM. Enfin, l'ajustement in vivo des séquences d'HIFU permet le relargage du principe actif. L'efficacité du traitement est ensuite suivie par IRM multiparamétrique anatomique, pondérée T2* de diffusion et de perfusion pour évaluer l'impact sur la fonctionnalité vasculaire et l'évolution tumorale, ainsi que par histologie. / Theranostic systems with imaging and therapeutic properties are developed to monitor treatments in vivo. The strategy we propose here is to design thermosensitive drug-loaded magnetic liposomes to treat superficial colon tumors CT26 on mice. These nanocarriers can be accumulated in the tumor by using a magnetic field gradient and the drug release can be triggered by a local heating induced by HIFU (High Intensity Focused Ultrasounds). They have been developed by co-encapsulation of magnetic nanoparticles and the antitumoral drug CA4P (Combretastatin A-4 Phosphate) in thermosensitive liposomes. The high loading of maghemite (γ-Fe2O3) magnetic nanoparticles enables both magnetic targeting and in vivo monitoring by MRI of this Ultra-Magnetic Liposomes (UML). In vitro HIFU heating at the UML membrane transition temperature improved the drug release. In vivo UML biodistribution was evaluated with dynamic susceptibility contrast imaging adjusted in time acquisition in MRI and the magnetic targeting efficiency was shown with a new MRI imaging processing. Adjustments of in vivo HIFU sequences to locally heat the tumor at the UML transition temperature allowed the triggering of drug release. Treatment efficiency was monitored by multiparametric diffusion, T2* weighted, anatomical and perfusion MRI and histology.

Le psoriasis est une maladie de peau auto-immune et chronique. Le rhumatisme psoriasique est une de ses principales complications qui est très invalidante pour les patients. Cette pathologie reste encore incurable à ce jour. L'usage des médicaments disponibles actuellement dans le psoriasis est limité par leurs effets secondaires dépendant de la dose et de la durée d'utilisation. Le but de ce travail était de développer des nanovecteurs médicamenteux à base de lipides pour un usage topique, en particulier ciblant l'épiderme viable qui est le site principal de la physiopathologie du psoriasis, mais aussi le derme et au-delà pour atteindre les articulations endommagées. Grâce à une nouvelle technique que nous avons développé et optimisé, le double déplacement de solvants, basée sur une organisation des phospholipides en deux temps, nous avons préparé des vésicules lipidiques encapsulant du diclofénac d'une part et de la ciclosporine A 'autre part. Ensuite, nous avons évalué leur aptitude à traverser la peau et cibler les régions d'intérêt. Après une étude systématique permettant d'optimiser les paramètres de préparation, les vésicules lipidiques encapsulant le diclofénac et la cyclosporine A ont montré une efficacité d'encapsulation (EE%) comprise entre 50% et 90% respectivement, selon la concentration en phospholipides. Après réalisation des études in vitro sur peau de cochon, nous avons observé que la formulation contenant une concentration basse en phospholipides (8,5 mg / mL) permettait d'encapsuler plus de 80% du diclofénac et de cibler le derme et au-delà. La formulation de vésicules lipidiques chargées de cyclosporine A qui encapsule la quantité la plus élevée (environ 80%) était également celle contenant la concentration basse de phospholipides. Contrairement au diclofénac, cette formulation n'était pas la meilleure pour cibler une couche profonde de la peau comme l'épiderme viable, alors que c'était le cas pour la formulation avec une concentration élevée de phospholipides (15 mg / mL), bien que l'EE% était d'environ 55%. Le double déplacement de solvant est une technique très prometteuse de préparation de vésicules lipidiques, capable de produire une population monodisperse d'échelle nanométrique. Cette méthode n'est que légèrement impactée lors d'une transposition d'échelle et serait donc facile à mettre en oeuvre à l'échelle industrielle. Cette méthode a été conçue dès le début pour utiliser des solvants favorisant la pénétration cutanée mais l'étendue de ces applications reste à explorer / Psoriasis is an auto-immune and chronic skin disease. Psoriatic arthritis is the main complication which is very disabling for patients. This pathology still remains incurable to date. The currently psoriasis indicated medicines use is limited by their side effects which are dose and use duration dependent. The aim of this work was to develop lipid based nanocarriers for skin targeting, especially the viable epidermis which is the main site of psoriasis physiopathology but also the dermis and beyond in order to reach the damaged articulations. Thanks to a new technique we developed and optimized called the double solvent displacement, based on a two-step phospholipid organization, we prepared diclofenac and cyclosporine A loaded lipid vesicles. Then, we evaluated their potential to cross the skin and target the skin layers of interest. After a systematic study to optimize preparation parameters, diclofenac and cyclosporine A loaded lipid vesicles displayed an encapsulation efficiency (EE %) between 50% and 90% respectively, according to the phospholipid concentration. After in vitro skin studies, we observed that the formulation containing the lower phospholipid concentration (8.5 mg/mL) allowed to encapsulate more than 80% of diclofenac and also to target the dermis and beyond. The formulation of cyclosporine A loaded lipid vesicles which encapsulates the higher amount (around 80%) is also the one containing the lower phospholipid concentration. Unlike to diclofenac, this formulation was not the better to target the viable epidermis whereas the formulation with the higher phospholipid concentration (15 mg/mL) was even though the EE% was of around 55%. The double solvent displacement is a very promising technique of lipid vesicle preparation, capable to produce monodisperse population of nanoscale carriers. This method is hardly impacted during scale-up and would be easy to implement at an industrial scale. This method was designed from the beginning to use skin penetration enhancer solvents but the scope of its applications still remains to be explored

L’objectif des travaux de cette thèse était de développer un système de délivrance stimulus-sensible innovant. Basé sur des vésicules lipidiques, il permet la libération d’une substance anti-cancéreuse hydrophile encapsulée dans leur cœur aqueux, sous l’effet de la lumière. Des porphyrines, incorporées dans leur bicouche, permettent, une fois illuminées, de générer de l’oxygène singulet qui oxyde les chaînes acyl insaturées des phospholipides. Cela induit une augmentation de la perméabilité des liposomes et permet la libération de leur cargo. Nous avons, dans un premier temps, effectué une sélection de phospholipides et de porphyrines permettant de construire le système. Les résultats expérimentaux ont pu être corrélés à une étude de simulation de dynamique moléculaire. L’ensemble a mis en exergue l’importance de la profondeur d’insertion de la porphyrine dans la bicouche lipidique et de sa proximité avec la double-liaison des phospholipides. Mais il a aussi montré les limites de ce système. Nous avons alors développé deux nouvelles molécules, dérivées de phospholipides naturels auxquels a été couplée la pheophorbide a. Malgré leur possible autoassemblage sous la forme de vésicules, ces derniers n’étaient pas stables et s’agrégeaient rapidement. Nous avons donc associé ces conjugués à des lipides classiques (DSPC, cholestérol) et analysé les propriétés des mélanges obtenus. Les propriétés photothermiques des systèmes conçus ont été confirmées, capables d’induire une élévation en température de 14°C. La chaleur générée, responsable d’une plus grande fluidité de la bicouche lipidique, a permis de favoriser la libération du cargo. Enfin, les deux conjugués synthétisés ont montré eux-mêmes une activité phototoxique (PDT), additionnée d’une sélectivité vis-à-vis de cellules du cancer de l’œsophage. Ces nouvelles molécules offrent donc de nombreuses opportunités pour le développement de systèmes multimodaux, bio-inspirés et biodégradables, pour la délivrance d’un médicament sous l’effet de la lumière. / The aim of this work was to develop an innovative stimulus-responsive delivery system. Based on lipid vesicles, it allows the controlled release, by light, of a hydrophilic anti-cancer substance encapsulated in their aqueous core. Once illuminated, porphyrin molecules inserted into the lipidic bilayer, generate singlet oxygen which oxidizes the unsaturated acyl chains of the phospholipids. This induces an increase in the permeability of the liposomes and the release of their cargo. We first made a selection of phospholipids and porphyrins to build the system. Our experimental study could be correlated with results of molecular dynamics simulations. The whole work highlighted the importance of the depth of insertion of porphyrin into the lipid bilayer and its proximity to the double bond of phospholipids. But it also showed the limits of this system. We then developed two new molecules, derived from natural phospholipids, to which pheophorbide a was coupled. The conjugates were able to form self-assembled vesicles but were unstable and quickly aggregated. We therefore associated these conjugates with classical lipids (DSPC, cholesterol) and analyzed the properties of these mixtures. We highlighted photothermal properties of the designed systems, capable of inducing a temperature rise of 14 °C. The generation of heat, responsible for a greater fluidity of the lipid bilayer, subsequently promoted the encapsulated cargo release. Finally, the two synthesized conjugates showed a phototoxic activity (PDT), with selectivity towards esophageal cancer cells. These new molecules therefore offer many opportunities for the development of multimodal, bio-inspired and biodegradable systems, for the delivery of a drug under the effect of light.

<p>When drug molecules are passively absorbed through the cell membrane in the small intestine, the first key step is partitioning of the drug into the membrane. Partition data can therefore be used to predict drug absorption. The partitioning of a solute can be analyzed by drug partition chromatography on immobilized model membranes, where the chromatographic retention of the solute reflects the partitioning. The aims of this thesis were to develop the model membranes used in drug partition chromatography and to study the effects of different membrane components and membrane structures on drug partitioning, in order to characterize drug–membrane interactions.</p><p>Electrostatic effects were observed on the partitioning of charged drugs into liposomes containing charged detergent, lipid or phospholipid; bilayer disks; proteoliposomes and porcine intestinal brush border membrane vesicles (BBMVs), and on the retention of an oligonucleotide on positive liposomes. Biological membranes are naturally charged, which will affect drug partitioning in the human body.</p><p>Proteoliposomes containing transmembrane proteins and cholesterol, BBMVs and bilayer disks were used as novel model membranes in drug partition chromatography. Partition data obtained on proteoliposomes and BBMVs demonstrated how cholesterol and transmembrane proteins interact with drug molecules. Such interactions will occur between drugs and natural cell membranes. In the use of immobilized BBMVs for drug partition chromatography, yet unsolved problems with the stability of the membrane were encountered. A comparison of partition data obtained on bilayer disks with data on multi- and unilamellar liposomes indicated that the structure of the membrane affect the partitioning. The most accurate partition values might be obtained on bilayer disks.</p><p>Drug partition data obtained on immobilized model membranes include both hydrophobic and electrostatic interactions. Such partition data should preferably be used when deriving algorithms or computer programs for prediction of drug absorption.</p>

When drug molecules are passively absorbed through the cell membrane in the small intestine, the first key step is partitioning of the drug into the membrane. Partition data can therefore be used to predict drug absorption. The partitioning of a solute can be analyzed by drug partition chromatography on immobilized model membranes, where the chromatographic retention of the solute reflects the partitioning. The aims of this thesis were to develop the model membranes used in drug partition chromatography and to study the effects of different membrane components and membrane structures on drug partitioning, in order to characterize drug–membrane interactions. Electrostatic effects were observed on the partitioning of charged drugs into liposomes containing charged detergent, lipid or phospholipid; bilayer disks; proteoliposomes and porcine intestinal brush border membrane vesicles (BBMVs), and on the retention of an oligonucleotide on positive liposomes. Biological membranes are naturally charged, which will affect drug partitioning in the human body. Proteoliposomes containing transmembrane proteins and cholesterol, BBMVs and bilayer disks were used as novel model membranes in drug partition chromatography. Partition data obtained on proteoliposomes and BBMVs demonstrated how cholesterol and transmembrane proteins interact with drug molecules. Such interactions will occur between drugs and natural cell membranes. In the use of immobilized BBMVs for drug partition chromatography, yet unsolved problems with the stability of the membrane were encountered. A comparison of partition data obtained on bilayer disks with data on multi- and unilamellar liposomes indicated that the structure of the membrane affect the partitioning. The most accurate partition values might be obtained on bilayer disks. Drug partition data obtained on immobilized model membranes include both hydrophobic and electrostatic interactions. Such partition data should preferably be used when deriving algorithms or computer programs for prediction of drug absorption.