Formation mechanism of insulin fibrils and structural aspects of the insulin fibrillation process.

The therapeutic importance of gaining a thorough knowledge on insulin fibrillation in relation to type I diabetes has lead to six decades of studies focusing on its formation kinetics and structural characteristics. Insulin fibrils feature characteristics common to amyloid fibrils such as an elongated morphology, characteristic cross-beta diffraction pattern, Thioflavin T fluorescence, and Congo Red birefringence. A full understanding of the fibrillation process requires structural elucidation of every species and determination of the kinetics of interconversion between species on the reaction pathway. Therefore, describing the underlying mechanism is complicated and different mechanisms have been proposed. In the recent years increased knowledge has been obtained on the importance of prefibrillar oligomeric species present during the process. A solution structure of such a species and also low-resolution structures of mature insulin fibrils have been obtained as well as high-resolution structures of two insulin hexa-peptide segments forming cross-beta sheet structures. However, it still remains to be elucidated whether these cross-beta structures corresponds to the structure in the fibril formed from the full-length protein. Different morphologies of insulin fibrils are observed depending on the arrangement of the protofilament, and even circular amyloids and spherulites composed of a core with many fibrils extending from it have been observed. This review will mainly focus on the structure of the species present during the insulin fibrillation process such as the partially unfolded monomeric intermediate, prefibrillar oligomeric species, and the insulin fibrils. Furthermore, it will address the formation mechanism and potential inhibition of the fibrillation process.

Characterisation of salmon calcitonin in spray-dried powder for inhalation. Effect of chitosan.

Salmon calcitonin (sCT) powders suitable for inhalation, containing chitosan and mannitol as absorption enhancer and protection agent, respectively, were prepared using a spray-drying process. The effect of chitosan on physicochemical stability of sCT in the dry powder was investigated by different analytical techniques. High-performance liquid chromatography (HPLC) analysis indicated that sCT was chemically stable upon spray-drying. With the proportion of chitosan in spray-drying formulation being increased, dissolution of sCT from the dry powders was decreased both in phosphate buffer and acetate buffer. The thioflavine T fluorescence assay showed that no fibrils were present in the spray-dried powder. However, sCT partly fibrillated in the phosphate buffer, but not in acetate buffer. Fourier transform infrared (FTIR) spectra showed that the secondary structure of sCT was slightly changed in the dry powder, yet no aggregate signal was observed. Circular dichroism analysis indicated that the structure of sCT in an aqueous formulation was slightly altered by addition of chitosan. Nevertheless, recovery of sCT was not influenced by chitosan in the aqueous formulation as indicated by HPLC analysis. This study suggested that sCT, in absence of any additives, was stable during the spray-drying process under certain conditions. Addition of chitosan affects recovery of sCT from spray-dried powders, which may be due to formation of a partially irreversible complex between the protein and chitosan during the spray-drying process.

The stability of insulin in solid formulations containing melezitose and starch. Effects of processing and excipients.

Solid insulin formulations obtained by different methods of preparation were compared with respect to chemical stability and morphology. Spray- and freeze-drying, solution enhanced dispersion by supercritical fluids (SEDS) and precipitation into starch microspheres were the methods used for preparation of solid powders. The excipients applied were melezitose, starch, and sodium taurocholate. The stability of the samples was evaluated after storage in open containers at 25 degrees C and 30% RH for 6 months. All samples were amorphous after processing and storage as detected by XRD, except for the starch microspheres which were semi-crystalline. The spray- and freeze-dried samples containing melezitose and sodium taurocholate experienced a significant water uptake during storage, resulting in changes in morphology and disappearance of Tg. However, the chemical stability of these samples did not seem to be affected by the water uptake. Changes in morphology were not observed for the SEDS powders and the starch microspheres. The chemical stability of the samples was assessed by HPLC. In general, conventional spray- and freeze drying resulted in samples with higher chemical stability compared to SEDS powders and starch microspheres. Nevertheless, the excipients applied were observed to be of major importance, and further optimization of the formulation as well as processing conditions may lead to slightly different conclusions.

Preparing and evaluating delivery systems for proteins.

From a formulation perspective proteins are complex and therefore challenging molecules to develop drug delivery systems for. The success of a formulation depends on the ability of the protein to maintain the native structure and activity during preparation and delivery as well as during shipping and long-term storage of the formulation. Therefore, the development and evaluation of successful and promising drug delivery systems is essential. In the present review, some of the particulate drug delivery systems for parenteral delivery of protein are presented and discussed. The challenge for incorporation of protein in particulate delivery systems is exemplified by water-in-oil emulsions.

Lung surfactant as a drug delivery system.

Lung surfactant is a complex mixture of mainly phospholipids and proteins. The composition leads to a unique spreading effect of the surfactant as well as spontaneous vesicle formation, which may be favourable characteristics of a drug delivery system for pulmonary delivery. The aim of study was to investigate the potential use of the surfactant extract, HL10 (LeoPharma, DK) as a drug delivery system. Studies involved incorporation of hydrophilic- and amphipathic model drugs (sucrose and acylated peptides) into HL10 and elucidation of the influence of surfactant proteins on the HL10 behaviour. Results showed that HL10 vesicles did not retain sucrose indicating formation of leaky vesicles. Studying the influence of surfactant proteins on release from DPPC-liposomes showed tendencies toward a protein-induced release. Hence, the surfactant proteins may influence the membrane lipid packing and characteristics resulting in leakiness of the membranes. Incorporation of acylated peptides into HL10 depended on the chain length rendering a successful incorporation of the peptide acylated with C14-acyl chains. This study suggests that HL10 may be a promising drug delivery system for the pulmonary delivery of amphipathic drug substances, e.g. therapeutically active acylated peptides (e.g. acylated insulin).

Displacement of adsorbed insulin by Tween 80 monitored using total internal reflection fluorescence and ellipsometry.

PURPOSE: This study was conducted to investigate the mechanism of action in the displacement of adsorbed insulin from a hydrophobic surface by Tween 80 and of the competitive adsorption of the two species. METHODS: Total internal reflection fluorescence (TIRF) and ellipsometry were used as in situ methods to examine the processes taking place at hydrophobic model surfaces in the presence of insulin and Tween 80. RESULTS: TIRF studies showed that the displacement of insulin by Tween 80 could be fitted to a sigmoidal function, indicating a nucleation-dependent process. Furthermore, a linear dependence between the apparent rate constant and the logarithm of the Tween 80 concentration was found. Competitive adsorption from solution mixtures of insulin and Tween 80 indicated that insulin was adsorbed first, but subsequently displaced by the surfactant. This displacement proved also to be dependent on the concentration of Tween 80 in the mixture. CONCLUSIONS: The results indicate that Tween 80 at concentrations above critical micelle concentration can be used to protect insulin against surface adsorption. The presence of a lag phase in the displacement at low surfactant concentration indicates that the mechanism of action for Tween 80 to reduce adsorption of insulin may be by competing for sites at the surface.

Adsorption of human insulin and AspB28 insulin on a PTFE-like surface.

The interactions of human insulin, Zn-free human insulin, and AspB28 insulin with a hydrophobic surface were studied by ellipsometry. All three insulin types investigated adsorbed with high affinity onto the hydrophobic surface, as the plateau of the adsorption isotherm, represented by the irreversible bound fraction, was reached at concentrations >10(-3) mg/ml. The plateau values for human insulin and Zn-free human insulin could not be distinguished with statistical significance, whereas the plateau value for AspB28 insulin was lower than those for the two others, with an adsorbed amount corresponding to a monolayer of insulin monomers. The results observed may be explained by differences in self-association patterns of the insulin types or by enhanced charge repulsion between the AspB28 analog and the negatively charged surface.

Epidermal growth factor and insulin short-term increase hPepT1-mediated glycylsarcosine uptake in Caco-2 cells.

AIMS: Little is known about the physiological regulation of the human intestinal di/tri-peptide transporter, hPepT1. In the present study we evaluated the effects of epidermal growth factor (EGF) and insulin on hPepT1-mediated dipeptide uptake in the intestinal cell line Caco-2. METHODS: Caco-2 cells were grown on filters for 23-27 days. Apical dipeptide uptake was measured using [14C]glycylsarcosine([14C]Gly-Sar). HPepT1 mRNA levels were investigated using RT-PCR, cytosolic pH was determined using the pH-sensitive fluorescent probe BCECF. RESULTS: Basolateral application of EGF increased [14C]Gly-Sar uptake with an ED50 value of 0.77 +/- 0.25 ng mL-1 (n = 3-6) and a maximal stimulation of 33 +/- 2% (n = 3-6). Insulin stimulated [14C]Gly-Sar uptake with an ED50 value of 3.5 +/- 2.0 ng mL-1 (n = 3-6) and a maximal stimulation of approximately 18% (n = 3-6). Gly-Sar uptake followed simple Michaelis-Menten kinetics. Km in control cells was 0.98 +/- 0.11 mM (n = 8) and Vmax was 1.86 +/- 0.07 nmol cm-2 min-1 (n = 8). In monolayers treated with 200 ng mL-1 of EGF, Km was 1.11 +/- 0.05 mM (n = 5) and Vmax was 2.79 +/- 0.05 nmol cm-2 min-1 (n = 5). In monolayers treated with 50 ng mL-1 insulin, Km was 1.03 +/- 0.08 mM and Vmax was 2.19 +/- 0.06 nmol cm-2 min-1 (n = 5). Kinetic data thus indicates an increase in the number of active transporters, following stimulation. The incrased Gly-Sar uptake was not accompanied by changes in hPepT1 mRNA, nor by measurable changes in cytosolic pH. CONCLUSIONS: Short-term stimulation with EGF and insulin caused an increase in hPepT1-mediated uptake of Gly-Sar in Caco-2 cell monolayers, which could not be accounted for by changes in hPepT1 mRNA or proton-motive driving force.

Association of acylated cationic decapeptides with dipalmitoylphosphatidylserine-dipalmitoylphosphatidylcholine lipid membranes.

The interaction of three acylated and cationic decapeptides with lipid membranes composed of dipalmitoylphosphatidylcholine (DPPC) and dipalmitoylphosphatidylserine (DPPS) has been studied by means of fluorescence spectroscopy and differential scanning calorimetry (DSC). The synthetic model decapeptides that are N-terminally linked with C(2), C(8), and C(14) acyl chains contain four basic histidine residues in their identical amino acid sequence. A binding model, based on changes in the intrinsic fluorescent properties of the peptides upon association with the DPPC-DPPS membranes, is used to estimate the peptide-membrane dissociation constants. The results clearly show that all three peptides have a higher affinity to liposomes containing DPPS lipids due to non-specific electrostatic interactions between the cationic peptides and the anionic DPPS lipids. Furthermore, it is found that the acyl chain length of the peptides plays a crucial role for the binding. A preference for fluid phase membranes as compared to gel phase membranes is generally observed for all three peptides. DSC is used to characterise the influence of the three peptides on the thermodynamic phase behaviour of the binary DPPC-DPPS lipid mixture. The extent of peptide association deduced from the heat capacity measurements suggests a strong binding and membrane insertion of the C(14) acylated peptide in accordance with the fluorescence measurements.

Dipeptide model prodrugs for the intestinal oligopeptide transporter. Affinity for and transport via hPepT1 in the human intestinal Caco-2 cell line.

The human intestinal di/tri-peptide carrier, hPepT1, has been suggested as a drug delivery target via increasing the intestinal transport of low permeability compounds by designing peptidomimetic prodrugs. Model ester prodrugs using the stabilized dipeptides D-Glu-Ala and D-Asp-Ala as pro-moieties for benzyl alcohol have been shown to maintain affinity for hPepT1. The primary aim of the present study was to investigate if modifications of the benzyl alcohol model drug influence the corresponding D-Glu-Ala and D-Asp-Ala model prodrugs' affinity for hPepT1 in Caco-2 cells. A second aim was to investigate the transepithelial transport and hydrolysis parameters for D-Asp(BnO)-Ala and D-Glu(BnO)-Ala across Caco-2 cell monolayers. In the present study, all investigated D-Asp-Ala and D-Glu-Ala model prodrugs retained various degrees of affinity for hPepT1 in Caco-2 cells. These affinities are used to establish a QSAR of our benzyl alcohol modified model prodrugs, aided at elucidating the observed differences in model prodrug affinity for hPepT1; additionally, these data suggest that the hydrophobicity of the side-chain model drug is the major determinant in the compounds affinity for hPepT1. Transepithelial transport studies performed using Caco-2 cells of D-Asp(BnO)-Ala and D-Glu(BnO)-Ala showed that the K(m) for transepithelial transport was not significantly different for the two compounds. The maximal transport rate of the carrier-mediated flux component does not differ between the two model prodrugs either. The transepithelial transport of D-Asp(BnO)-Ala and D-Glu(BnO)-Ala follows simple kinetics, and the release of benzyl alcohol is pH-dependent, but unaffected by 1 mM of the esterase inhibitor Paraoxon in 80% human plasma and Caco-2 cell homogenate.

Probing the mechanism of insulin fibril formation with insulin mutants.

The molecular basis of insulin fibril formation was investigated by studying the structural properties and kinetics of fibril formation of 20 different human insulin mutants at both low pH (conditions favoring monomer/dimer) and at pH 7.4 (conditions favoring tetramer/hexamer). Small-angle X-ray scattering showed insulin to be monomeric in 20% acetic acid, 0.1 M NaCl, pH 2. The secondary structure of the mutants was assessed using far-UV circular dichroism, and the tertiary structure was determined using near-UV circular dichroism, quenching of intrinsic fluorescence by acrylamide and interactions with the hydrophobic probe 1-anilino-8-naphthalene-sulfonic acid (ANS). The kinetics of fibril formation were monitored with the fluorescent dye, Thioflavin T. The results indicate that the monomer is the state from which fibrils arise, thus under some conditions dissociation of hexamers may be rate limiting or partially rate limiting. The insulin mutants were found to retain substantial nativelike secondary and tertiary structure under all conditions studied. The results suggest that fibril formation of the insulin mutants is controlled by specific molecular interactions that are sensitive to variations in the primary structure. The observed effects of several mutations on the rate of fibril formation are inconsistent with a previously suggested model for fibrillation [Brange, J., Whittingham, J., Edwards, D., Youshang, Z., Wollmer, A., Brandenburg, D., Dodson, G., and Finch, J. (1997) Curr. Sci. 72, 470-476]. Two surfaces on the insulin monomer are identified as potential interacting sites in insulin fibrils, one consisting of the residues B10, B16, and B17 and the other consisting of at least the residues A8 and B25. The marked increase in the lag time for fibril formation with mutations to more polar residues, as well as mutations to charged residues, demonstrates the importance of both hydrophobic and electrostatic interactions in the initial stages of fibrillation. A model for insulin fibril formation is proposed in which the formation of a partially folded intermediate is the precursor for associated species on the pathway to fibril formation.

Epidermal growth factor inhibits glycylsarcosine transport and hPepT1 expression in a human intestinal cell line.

The human intestinal cell line Caco-2 was used as a model system to study the effects of epidermal growth factor (EGF) on peptide transport. EGF decreased apical-to-basolateral fluxes of [(14)C]glycylsarcosine ([(14)C]Gly-Sar) up to 50.2 +/- 3.6% (n = 6) of control values. Kinetic analysis of the fluxes showed that maximal flux (V(max)) of transepithelial transport decreased from 3.00 +/- 0.17 nmol x cm(-2) x min(-1) in control cells to 0.50 +/- 0.07 nmol x cm(-2) x min(-1) in cells treated with 5 ng/ml EGF (n = 6, P < 0.01). The apparent Michaelis-Menten constant (K(m)) was 2.71 +/- 0.31 mM (n = 6) in control cells and 1.89 +/- 0.28 mM (n = 6, not significantly different from control) in EGF-treated cells. Similarly, apical uptake of [(14)C]Gly-Sar decreased in cells treated with EGF, with an ED(50) value of 0.36 +/- 0.06 ng/ml (n = 6) EGF and a maximal inhibition of 80 +/- 0.02% (n = 6). V(max) decreased from 2.61 +/- 0.4 to 1.06 +/- 0.1 nmol x cm(-2) x min(-1) (n = 3, P < 0.05), whereas K(m) remained constant. Basolateral Gly-Sar uptake showed no changes in V(max) or K(m) after EGF treatment (n = 3). RT-PCR showed a decrease in hPepT1 mRNA (using glucose-6-phosphate dehydrogenase mRNA as control) in cells treated with EGF. Western blotting indicated a decrease in hPepT1 protein in cell lysates. We conclude that EGF treatment decreases Gly-Sar transport in Caco-2 cells by decreasing the number of peptide transporter molecules in the apical membrane.

Model prodrugs for the intestinal oligopeptide transporter: model drug release in aqueous solution and in various biological media.

The human intestinal di/tri-peptide carrier, hPepT1, has been suggested as a target for increasing intestinal transport of low permeability compounds by creating prodrugs designed for the transporter. Model ester prodrugs using the stabilized dipeptides D-Glu-Ala and D-Asp-Ala as pro-moieties for benzyl alcohol have been shown to have affinity for hPepT1. Furthermore, in aqueous solution at pH 5.5 to 10, the release of the model drug seems to be controlled by a specific base-catalyzed hydrolysis, indicating that the compounds may remain relatively stable in the upper small intestinal lumen with a pH of approximately 6.0, but still release the model drug at the intercellular and blood pH of approximately 7.4. Even though benzyl alcohol is not a low molecular weight drug molecule, these results indicate that the dipeptide prodrug principle is a promising drug delivery concept. However, the physico-chemical properties such as electronegativity, solubility, and log P of the drug molecule may also have an influence on the potential of these kinds of prodrugs. The purpose of the present study is to investigate whether the model drug electronegativity, estimated as Taft substitution parameter (sigma*) may influence the acid, water or base catalyzed model drug release rates, when released from series of D-Glu-Ala and D-Asp-Ala pro-moieties. Release rates were investigated in both aqueous solutions with varying pH, ionic strength, and buffer concentrations as well as in in vitro biological media. The release rates of all the investigated model drug molecules followed first-order kinetics and were dependent on buffer concentration, pH, ionic strength, and model drug electronegativity. The electronegativity of the model drug influenced acid, water and base catalyzed release from D-Asp-Ala and D-Glu-Ala pro-moieties. The model drug was generally released faster from D-Asp-Ala- than from the D-Glu-Ala pro-moieties. In biological media the release rate was also dependent on the electronegativity of the model drug. These results demonstrate that the model drug electronegativity, estimated as Taft (sigma*) values, has a significant influence on the release rate of the model drug.

Effect of environmental factors on the kinetics of insulin fibril formation: elucidation of the molecular mechanism.

In the search for the molecular mechanism of insulin fibrillation, the kinetics of insulin fibril formation were studied under different conditions using the fluorescent dye thioflavin T (ThT). The effect of insulin concentration, agitation, pH, ionic strength, anions, seeding, and addition of 1-anilinonaphthalene-8-sulfonic acid (ANS), urea, TMAO, sucrose, and ThT on the kinetics of fibrillation was investigated. The kinetics of the fibrillation process could be described by the lag time for formation of stable nuclei (nucleation) and the apparent rate constant for the growth of fibrils (elongation). The addition of seeds eliminated the lag phase. An increase in insulin concentration resulted in shorter lag times and faster growth of fibrils. Shorter lag times and faster growth of fibrils were seen at acidic pH versus neutral pH, whereas an increase in ionic strength resulted in shorter lag times and slower growth of fibrils. There was no clear correlation between the rate of fibril elongation and ionic strength. Agitation during fibril formation attenuated the effects of insulin concentration and ionic strength on both lag times and fibril growth. The addition of ANS increased the lag time and decreased the apparent growth rate for insulin fibril formation. The ANS-induced inhibition appears to reflect the formation of amorphous aggregates. The denaturant, urea, decreased the lag time, whereas the stabilizers, trimethylamine N-oxide dihydrate (TMAO) and sucrose, increased the lag times. The results indicated that both nucleation and fibril growth were controlled by hydrophobic and electrostatic interactions. A kinetic model, involving the association of monomeric partially folded intermediates, whose concentration is stimulated by the air-water interface, leading to formation of the critical nucleus and thence fibrils, is proposed.

Drug delivery by phospholipase A(2) degradable liposomes.

The effect of poly(ethylene glycol)-phospholipid (PE-PEG) lipopolymers on phospholipase A(2) (PLA(2)) hydrolysis of liposomes composed of stearoyl-oleoylphosphatidylcholine (SOPC) was investigated. The PLA(2) lag-time, which is inversely related to the enzymatic activity, was determined by fluorescence, and the zeta-potentials of the liposomes were measured as a function of PE-PEG lipopolymer concentration. A significant decrease in the lag-time, and hence an increase in enzymatic activity, was observed with increasing amounts of the negatively charged PE-PEG lipopolymers incorporated into the SOPC liposomes. The enhancement of the PLA(2) enzymatic activity might involve a stronger PLA(2) binding affinity towards the negatively charged and polymer covered PEG liposomes.

Association of an acylated model peptide with DPPC-DPPS lipid membranes.

The interaction between a small positively charged peptide with a N-terminally linked acyl chain and dipalmitoylphosphatidylcholine-dipalmitoylphosphatidylserine (DPPC-DPPS) lipid membranes has been studied by means of fluorescence resonance energy transfer. Two different lipid compositions were used: a neutral membrane (100 mol% DPPC), and a negatively charged membrane (30 mol% DPPS in DPPC). The fluorescence resonance energy transfer results reveal that the peptide associates with both types of membranes. Furthermore, it is found that the slope of the titration curve for the negatively charged membranes is much steeper than that for the neutral membranes. This indicates a higher binding affinity of the acylated peptide towards negatively charged lipid membranes as compared with neutral lipid membranes.

Activity of mammalian secreted phospholipase A(2) from inflammatory peritoneal fluid towards PEG-liposomes. Early indications.

Due to an increase in the activity of phospholipase A(2) (PLA(2)) in various inflammatory diseases, this enzyme may play a key role in the degradation of liposomes and the subsequent release of drug when PEG-liposomes passively target inflammatory tissue. The activity of mammalian secreted phospholipase A(2) (sPLA(2)) in casein stimulated peritoneal fluid was tested toward liposomes of different compositions. Early results indicate only a slight degradation of conventional dipalmitoylphosphatidylcholine (DPPC) liposomes as well as DPPC liposomes incorporated with different concentrations of PEG(2000). However, the DPPC degradation increased to 7% when inclusion of 30 mol% phosphatidylethanolamine (PE) in the lipid bilayer. The increase in degradation may be due to an improvement of the substrate - as it is well known, that PE is a better substrate for the mammalian sPLA(2) than PC. Incorporation of PE into the bilayer may increase the binding properties of the bilayer resulting in improved conditions for the enzymatic attack by sPLA(2). In addition, inhibitory zones of Staphylococcus aureus in an agar diffusion test showed that PLA(2) from Crotalus atrox venom was able to catalyze the release of gentamicin from PEG-liposomes. In conclusion, this study suggest that degradation of the lipid bilayer of PEG-liposomes by PLA(2) result in release of incapsulated drug, e.g. gentamicin and inclusion of PE in the liposomal bilayer, may enhance the activity of the mammalian sPLA(2) toward liposomes composed of DPPC.

Parenteral water/oil emulsions containing hydrophilic compounds with enhanced in vivo retention: formulation, rheological characterisation and study of in vivo fate using whole body gamma-scintigraphy.

The preparation and characterization of parenteral water-in-oil (w/o) emulsions with a potential for sustained release of hydrophilic drugs was described with emphasis on rheological behaviour and spreading phenomenon after intramuscular (i.m.) injection in rabbit thigh muscle. Both steady state and dynamic rheological parameters were investigated showing Newtonian behaviour at low fraction of disperse phase ratio as opposed to viscoelastic and pseudoplastic behaviour at high fraction of disperse phase. Disappearance and spreading behaviour of hydrophilic radioactive markers, aprotinin (6512 g/mol) and pertechnetate (193 g/mol) entrapped in w/o emulsions from an i.m. injection site was studied by whole body gamma-scintigraphy. The retention of entrapped aprotinin 24 h postinjection was 83 +/- 5% for a low spreading emulsion and 76 +/- 6% for a high spreading emulsion. The corresponding values for pertechnetate were 50 +/- 11 and 23 +/- 2%, respectively. The relatively long retention times were suggested to be related to the good physical stability properties of the present emulsions. It was concluded that the presented w/o emulsions are promising vehicles for sustained release of hydrophilic drugs from an i.m. injection site.

Sustained elevated plasma aprotinin concentration in mice following intraperitoneal injections of w/o emulsions incorporating aprotinin.

This study was initiated to test the feasibility of w/o emulsions as a sustained release system for aprotinin following intraperitoneal injection in mice. The emulsion was well tolerated in mice and sustained release was observed over a period of 96 h. The time for maximum plasma concentration of aprotinin was 10 min and 12 h after injection of a control solution and the emulsion dosage form, respectively. Furthermore, the hemolytic activity of the emulsion constituents was low indicating a low acute toxicological potential of the emulsion. The present study also showed that the lipolytic activity in peritoneal exudate from mice is important for the clearance of oily vehicles from the peritoneal cavity with lipolytic rate constants ranging from 50 to 130 nmol free fatty acid released/min/mg exudate protein at 37 degrees C, pH 8.5. It was concluded that the w/o emulsion was well suited to provide sustained elevated plasma aprotinin concentrations in mice.