In vivo efficiency of targeted norfloxacin against persistent, isoniazid-insensitive, Mycobacterium bovis BCG present in the physiologically hypoxic mouse liver.

Persistence of Mycobacterium tuberculosis is a hypoxia-inducible state in which the bacteria are phenotypically insensitive to currently available antituberculous drugs. In humans, persistent M. tuberculosis is found in granulomatous lesions, either inside macrophages or in necrotic tissue, where the partial oxygen pressure (pO(2)) is very low. Persistent bacteria can remain silent for decades before overt tuberculosis develops. Due to insensitivity to classical drugs, M. tuberculosis persistence prevents rapid and definitive clearance of bacteria. Consequently, therapeutic molecules are required that are both active against persistent bacilli and able to reach their intramacrophagic location. In contrast to its native form, norfloxacin is active in vivo against Mycobacterium bovis BCG present in the lungs when temporarily linked to a macromolecular carrier targeted to macrophages. To study the efficiency of this macromolecular prodrug targeted to persistent mycobacteria confined inside macrophages, we established a short-term in vivo model based on the physiological pO(2) differences between lungs, spleen and liver. Whereas lungs and spleen are well oxygenated, the liver has a low pO(2) due to its portal irrigation. Therefore, studying mycobacteria in the liver yields information about in vivo persistent bacilli exposed to low pO(2). To our knowledge, no similar short-term in vivo model has been published to date. Using this model, we demonstrated the insensitivity to isoniazid of M. bovis BCG present in hypoxic sites, and showed that norfloxacin given as a mannosylated macrophage-targeted prodrug was able to kill these isoniazid-insensitive mycobacteria. This demonstrates that intracellular persistent mycobacteria are amenable to antibiotic treatment.

Interactions between red blood cells and a lethal, partly quaternized tertiary polyamine.

Partially quaternized poly[thio-1-(N,N-diethyl-aminomethyl) ethylene]s, Q-P(TDAE)(x) with x indicating the percentage of quaternized subunits, have been proposed as potential carriers for drugs insoluble in water. However these cationic polyelectrolytes form emboli upon intravenous administration. In order to study the mechanism, Q-P(TDAE)(11) was incubated in vitro with red blood cells (RBCs) suspended in various aqueous media such as autologous plasma, autologous serum, albumin dissolved in phosphate buffer, plasma-serum mixtures and Tris buffer. The deformability of the RBC membrane studied by viscometry was not affected by the polycation. Q-P(TDAE)(11)-induced hemagglutination was studied by optical microscopy. It depended on the polymer concentration and on the presence of plasma proteins. As ghosts were formed in some cases, hemolysis was investigated by measuring potassium and hemoglobin released from RBCs. Fibrinogen and serum proteins, except albumin, protected RBCs from hemolysis. Moreover the order of addition of the suspension components modulated dramatically the Q-P(TDAE)(11)-induced hemolysis. Addition of Q-P(TDAE)(11) to whole blood caused hemolysis whereas addition of the polymer to plasma prior to contact with RBCs did not affect the cell integrity. In contrast, addition of the polymer to RBCs suspended in albumin solution caused greater hemolysis than the addition to whole blood, and the contact between Q-P(TDAE)(11) and albumin prior to RBC addition still enhanced cell lysis. Two conclusions can be drawn from these observations: (i) Q-P(TDAE)(11) induces both hemagglutination, probably through electrostatic interaction, and hemolysis, because Q-P(TDAE)(11) disrupted the RBC lipid bilayer; (ii) proteins can decrease or increase the deleterious effects of Q-P(TDAE)(11) on RBCs.

Poly(ethylene glycol): protein-repulsive or albumin-compatible?

In the literature, many papers deal with the behavior of proteins in aqueous media in the presence of poly(ethylene glycol) (PEG) molecules or poly(ethylene oxide) (PEO) segments, physically adsorbed onto, or covalently attached to, macromolecules or to solid surfaces. In particular, it is well known that PEO segments make foreign materials stealthy, i.e. they are much less detected by the immune system either through humoral reactions or, at the cell level, through opsonins. Revisiting the literature led us to challenge the largely accepted opinion that the decreased recognition of PEO segment-bearing foreign macromolecules and particles by the mononuclear phagocyte system is primarily the consequence of the repulsion of all blood proteins by PEG segments through the excluded volume effect. This challenge is based on the finding that albumin and PEG are compatible in phosphate-buffered saline at room temperature and at concentrations comparable to those measured by others on the surface of PEO segment-bearing species, whereas fibrinogen and PEG phase-separated and were incompatible despite the much lower concentration of the latter protein. According to literature and to these observations, it is proposed that the stealth effect induced by PEO segments is primarily due to the compatibility between PEO segments of intermediate molar mass and albumin, thus rendering PEO-bearing macromolecules or surfaces to look like native albumin. Under such conditions, the hospitality offered by PEG macromolecules or PEO segments to albumin, the dominant plasma protein, results in a 'chameleon' effect that prevents the activation of other PEG-compatible or -incompatible plasma proteins or cells involved in foreign body recognition and elimination. PEG with molar masses > or = 8000 did not accommodate albumin in agreement with the excluded volume phenomenon.

Polymeric prodrugs of antibiotics with improved efficiency.

Macromolecular prodrugs of the antibiotic norfloxacin were prepared by coupling the drug via a peptide spacer onto a mannosylated dextran. The tetrapeptide gly-phe-gly-gly-gly-OMe was selected as substrate for lysosomal enzymes. The drug was coupled on the alpha-C of the terminal glycine. In vitro degradation studies demonstrated the release of the parent drug in the presence of cathepsin B. In vivo experiments on mice showed a promising therapeutic effect.

Monoacylation of ribonuclease A enables its transport across an in vitro model of the blood-brain barrier.

A major challenge in correcting disorders affecting the central nervous system is to induce blood-brain barrier (BBB) crossing of exogenous biological compounds such as proteins or specific nucleic acid sequences. Fatty acids, due to their high membrane affinity and low toxicity, are good potential candidates to promote this barrier crossing when covalently bound to proteins. In this paper, we report that regiospecific monoacylation of ribonuclease A (RNase A) enables its transport across an in vitro model of the BBB. Myristoylated, palmitoylated and stearoylated RNases A were prepared using reversed micelles as microreactors. All the purified acylated RNases A kept their original enzymatic activity. A single fatty acid moiety was linked to RNase A through the alpha-amino group of its N-terminal lysine as shown by powerful analytical techniques. The ability of monoacylated RNases A to cross an in vitro model of the BBB is strictly dependent on the acyl chain length, which must be at least 16 carbon atoms long.

Effect of protein chemical hydrophobization on antiglucose oxidase immunoglobulin production in mouse.

One problem resulting from the therapeutic use of enzymes is the adverse immunological reactions. In order to study the immunoglobulin production elicited into mice by different derivatives of an enzyme, glucose oxidase was chosen as a model. The immunoglobulin productions induced by apoglucose oxidase, prepared by removing flavine adenine dinucleotide from the native enzyme through an acidic treatment and devoid of enzymatic activity, by metaperiodate-oxidized glucose oxidase that lost about 50% of its carbohydrate moiety, and by propyl aliphatic chains-coupled glucose oxidase were as intense as that induced by native glucose oxidase. On the other hand, coupling hexyl aliphatic chains to the enzyme did change its ability to stimulate antibody production. This hydrophobized preparation induced a low titer of antibody after repeated intravenous or subcutaneous injections. This result suggests a simple strategy for reducing the immunogenicity of foreign proteins and for decreasing the risk of immunological complications in enzyme therapy.

Immunoassay for native enzyme quantification in biological samples.

In order to detect low levels of enzyme activity, specifically glucose oxidase, in biological samples, an immunoenzymatic assay was developed since currently available methods could not be used because of either their lack of sensitivity or the conditions prevailing in our samples: turbidity of the medium, presence of redox systems other than glucose oxidase, and high concentration of proteins. The principle of the method is to coat a polystyrene surface with a fragment Fc-specific anti-IgG, then with an antibody directed against the looked-for enzyme, which is simultaneously the antigen and the enzyme activity required for immunoenzymatic detection. We applied this concept to biological samples after glucose oxidase administration to mice. This method achieves specificity and sensitivity (20 ng/mL or 1 ng) with samples of biological origin. No marker is needed since the antigen itself possesses an enzyme activity. This method, which requires a small sample volume (50 microL, 20 microL, if necessary), can be extended easily to the many enzymes currently used as markers. It could also be applied to the native enzymes of medical interest for which antibodies and a colorimetric reaction are available.

Fatty acid acylation of RNase A using reversed micelles as microreactors.

A water soluble protein, RNAse A, was fatty-acylated using AOT reversed micelles in 2,2,4-trimethyl pentane as microreactors and myristoyl chloride as reagent. Artificial attachment of lipid molecules to this protein was performed for different hydration degrees by changing Wo = [water]/ [AOT], the parameter which controls the microreactor size. The chemically modified protein was monitored using reverse phase HPLC and characterized by HPLC, free amino groups titration, and electrophoresis. An RNase A/myristoyl chloride ratio of 1:4 (mol/mol) at Wo = 7 was found to give 60% of modified protein.

Influence of barrier-crossing limitations on the amount of macromolecular drug taken up by its target.

Macromolecules (substitutive enzymes, polymeric prodrugs, immunotoxins, radiolabeled antibodies, or peptide hormones) are of interest in the treatment of several diseases. To reach the tissues, these macromolecular drugs have to cross the capillary wall, which represents an important transfer limitation. While pharmacokinetics usually studies the changes in drug concentration in different body compartments, analyzing the amount of drug gaining access to its target may be more relevant for assessing the efficiency of macromolecules than for low molecular mass drugs. To determine the influence of different parameters on the fraction of the injected dose gaining access to the pharmacologic target, we constructed pharmacokinetic models where two uptakes, both linear or nonlinear, work either in the same compartment (no transport limitation), or in compartments separated by a transport barrier. Numerical applications were carried out with parameters obtained either experimentally or from the literature. We conclude that it is of little use to increase the affinity (K(uptake)) of a macromolecular drug for its target when a transport limitation and an undesired elimination from the plasma space are both present. Likewise, an increase of the uptake (rate of uptake or maximal velocity) by the target is not very productive because permeability of the capillary wall is the factor limiting access of macromolecules to tissues. Maximal efficiency of therapeutic macromolecules could be achieved by increasing, where feasible, the transport across the barrier between the plasma and the target, and by preventing the undesired eliminations as much as possible.

Chemically processed bovine heterografts of the second generation as arterial substitutes: a comparative evaluation of three commercial prostheses.

The use of chemically processed bovine heterografts is primarily confined to the construction of arterio-venous blood accesses in those patients requiring hemodialysis, plasmapheresis or chemotherapy. The grafts of the first generation i.e. Artegraft and Solcograft are now being supplanted by those of the second generation i.e. Reinforced Artegraft, Solco P and NCGT. We have investigated these three types of arterial prostheses as a biomaterial in terms of sterility, inflammatory response and cytocompatibility and as a blood conduit in dogs in terms of patency and healing. For each type of graft, two implantations were carried out for durations of 4 hours, 24 hours, 48 hours, one week, two weeks, one month, three months, and six months. Therefore a total of 48 grafts were implanted. All grafts but five were patent at sacrifice: thromboses were observed in two Reinforced Artegraft (after two weeks and after one month) and in three NCGT (after 24 hours, after 48 hours, and after one month). Therefore the following patencies were observed: Reinforced Artegraft 14/16, Solco P 16/16 and NCGT 13/16. In all the patent grafts, the healing was reduced to the formation of a pannus along both anastomoses; thrombotic accumulations were observed on the surface defects of the grafts, particularly the NCGT graft. The Reinforced Artegraft presents only minor advantages over the previous Artegraft; the Solco P, somewhat more acceptable is no longer commercially available since the manufacturer withdrew it after early clinical failures. The improvements noted in the bovine heterografts of the second generation appear to be marginal as compared to those of the first generation.

Elimination of artifacts due to glutaraldehyde fixation in the histochemical detection of glucose oxidase with tetrazolium salts.

High background staining due to glutaraldehyde fixation prevents phenazine methosulphate and a tetrazolium salt being used to visualize glucose oxidase activity in tissue slices prepared from mice injected with the enzyme. Experiments in solution showed that products formed during the reaction between amino groups and glutaraldehyde are, at least in part, responsible for the non-enzymatic reduction of tetrazolium salts. Experiments performed with artificial membranes chemically akin to glutaraldehyde-fixed sections and prepared by cross-linking albumin by glutaraldehyde, showed that double bonds in amino-glutaraldehyde products are mainly responsible for the background staining development, whereas thiol groups play only a minor role. A sequential treatment with sodium borohydride and N-ethylmaleimide greatly reduced the background staining, thus permitting the detection of glucose oxidase activity. Optimal conditions for glucose oxidase activity demonstration (maximum enzyme velocity for minimum 'nothing dehydrogenase' phenomenon) were studied: choice of the tetrazolium salt, nature, pH and molarity of the buffer used for the staining mixture. A procedure similar to that developed with artificial membranes was applied to tissue sections of mice in which glucose oxidase had been injected intravenously. It allowed detection of glucose oxidase activity without artifactual staining in control slices.

In vivo evaluation of polyester arterial grafts coated with albumin: the role and importance of cross-linking agents.

Previous in vitro studies have predicted that the type of chemical used to cross-link albumin-coated polyester arterial prostheses may influence the rate of bioerosion of the albumin layer in vivo. This study has confirmed that the healing process of this type of compound prosthesis does indeed depend on the nature and concentration of the cross-linking agent used. Four series of implantations in the thoracic aorta of dogs for scheduled periods for 4 h up to 6 months were conducted using 1.6% glutaraldehyde, 2.5% glutaraldehyde and 0.2 M carbodiimide as the alternative cross-linking agents plus a nonalbuminated preclotted polyester prosthesis which served as the control. The pathology of the explanted grafts revealed that in the short and medium term the rate of healing and the extent of tissue ingrowth was dependent initially on the presence of and later on the rate of bioerosion of the albumin layer. After 3 months in situ, the prostheses coated with albumin cross-linked with 1.6% glutaraldehyde and carbodiimide had healed more rapidly and were invaded by more extensive tissue ingrowth than the one cross-linked with 2.5% glutaraldehyde or the preclotted control. Moreover, the migration of cells over the carbodiimide-treated surface was the most fully developed and most regularly organized of all four series. Immunostaining revealed that the presence of glutaraldehyde induced an inflammatory response which failed to support the growth of normal luminal cells with the endothelial phenotype.

Blood hemolysis by PTFE and polyurethane vascular prostheses in an in vitro circuit.

In order to improve understanding of the appearance of bright yellow stains in vivo (consecutive to the absorption of bilirubin) on a novel microporous, hydrophilic polyetherurethaneurea vascular prosthesis, the in vitro hemolytic activity of the material was compared with expanded polytetrafluoroethylene and silicone rubber. The results show that the tendency of the polyetherurethaneurea to produce free hemoglobin is low, so that the yellow staining observed is likely to be a result of the contact between the polymer and thrombi: Bilirubin is produced because of hemoglobin degradation in the thrombi rather than an active hemolysis on the surface of the prosthesis itself.

Effect of prosthetic sugar groups on the pharmacokinetics of glucose-oxidase.

The administration of enzymes is of potential therapeutic value in many disease states, e.g. lysosomal storage diseases, provided problems in the metabolism and targeting of large proteins can be overcome. We have addressed ourselves to these problems by studying the pharmacokinetics and distribution of glucose-oxidase (GO) and some of its derivatives in mice. A saturable mechanism was responsible for GO uptake by mononuclear phagocytes. After construction of a pharmacokinetic model, the Kuptake (850 nmol/l) and the number of capturing cells were determined; uptake was half the initial plasma concentration in about 10 min. Deglycosylated GO's had half-lives of about 100 min and were taken up by the same organs that took up native GO. Galactosylated GO had a half-life of 4 min and a different distribution; it was taken up preferentially by the liver in hepatocytes. Our results illustrate the role sugars might play in the targeting of foreign proteins to different cell types, and the feasibility of determining in vivo microscopic constants such as the affinity between molecules and certain cells.

Degradability of crosslinked albumin as an arterial polyester prosthesis coating in in vitro and in vivo rat studies.

In order to avoid the preclotting procedure in knitted polyester arterial prostheses and in woven models, compound polyester grafts have been proposed, containing preadsorbed collagen or albumin. Since we are currently investigating grafts impregnated with crosslinked albumin, it was decided to establish the degradation rate of this coating after stabilization with either glutaraldehyde (GA) or carbodiimide (CDI). Tests were performed in vitro by incubation in either PBS, plasma or pancreatin and in vivo by implantation in the abdominal cavity of rats. In PBS or plasma in vitro, the coatings were very stable (2% degradation after 144 h incubation), however, in pancreatin the CDI crosslinked albumin degraded much faster than the GA crosslinked albumin (more than 50% degradation in 12 h compared to less than 30% in 48 h). In vivo the degradation rates of the two types of crosslinked albumin were similar (almost all of the albumin having been lost after 4 weeks) but the cellular response was very different: a mild tissue reaction was observed with the CDI crosslinked coating whereas many foreign body giant cells were present on the GA crosslinked material.