AU703034B2 - Microporous pathogen-killing composition - Google Patents

Description

WO 97/15661 PCT/US96/17136 -1- MICROPOROUS PATHOGEN-KILLING COMPOSITION Field of the Invention Although the threat of hepatitis and AIDS are decreased by screening blood and blood products for the presence of viral antigens, antibodies and surrogate markers, the safety of blood and blood product transfusion is still a problem. The present invention provides a microporous substance for sequestering viruses and other pathogenic organisms in lethal proximity to a singlet oxygen generating system bound to the microporous substance.

Background of the Invention Aspects of this invention derive from several disparate arts that are discussed below.

Photodynamic Oxygenating Activity: The microbicidal effects of photodynamic action have been known for a century (Raab, 1900). Recently, photochemical-based methodologies have been developed to inactivate viruses in blood (Matthews et al., 1988; Sieber et al., 1989; Neyndorffet al., 1990; O'Brien et al., 1990; Horowitz et al., 1991) and blood products (Lin et al., 1989; Prodouz et al., 1991; Dodd et al., 1991: Margolis-Nunno et al., 1992). The combination of a photosensitizing molecule and light (hv) at an appropriate wavelength (color) can generate molecular oxygen (02) that has potent oxygenating activity capable of destroying pathogenic microbes (see Mohr et al., 1995, for general review), although few have succeeded in devising practical applications for this killing method. In other examples, psoralen activated by ultraviolet A light is reported to WO 97/15661 PCTIUS96/17136 -2inactivate vesicular stomatitis virus (Margolis-Nunno et al., 1992), and hypericin activated by fluorescent light is reported to inactivate HIV-1 (Lavie et al., 1995).

Photosensitizers can use either a type I or a type II mechanism (Schenck and Koch, 1960). The equations below explain the two types of mechanisms, and use the following terminology. The "spin multiplicity" of an atom or molecule is the spectroscopic expression of its total spin quantum number as defined by the expression I 2S 1. The spin multiplicities of the reactants, photosensitizer and 02, are described by an antecedent superscript; and an asterisk is used to indicate electronically excited molecules.

A photosensitizer in the S 0 state, has a singlet multiplicity; 12(0)1 1 1 (singlet, 1 photosensitizer molecule). The excited product remains singlet if an electron becomes excited but still has the same spin number. If the photosensitizer has the proper electronic structure and symmetry, intersystem crossing can occur; i.e., and the excited singlet photosensitizer molecule can relax to a lower energy electronically excited state by changing the spin quantum number of its excited electron. In doing so S is changed from 0 to 1, and the multiplicity, 12(1) 1 1 3, is changed to an electronically excited triplet 3 photosensitizer*): lPhotosensitizer* -Intersystem Crossing 3 Photosensitizer*. (1) For either type I or type II mechanisms, the initial event is the same. For this initial event, the ground state singlet multiplicity dye (1Photosensitizer) absorbs a photon of light and is transformed to its electronically excited singlet state (1Photosensitizer*): 1 Photosensitizer Light (hv) 1Photosensitizer* (2) In type I reactions, an electron is transferred from a reducing substrate to the triplet exited dye yielding two radicals, the reduced doublet photosensitizer anion and the oxidized doublet substrate cation: 3 Photosensitizer* 1Substrate 2 Photosensitizer 2 Substrate (3) or an electron is transferred from the excited triplet photosensitizer molecule to the substrate: 3 Photosensitizer* 1 Substrate 2 Photosensitizer 2 Substrate' (4) The radicals generated can react with other radicals in annihilation reactions or can react with higher multiplicity molecules in radical propagation reactions.

I

WO 97/15661 PCT/US96/17136 -3- For example, the photosensitizer radical can be harnessed via a type I reaction to the production of oxygen radicals that are capable of destroying pathogenic organisms. The reaction of oxygen with biomolecules is highly exergonic, but such reactions do not occur spontaneously. Biomolecules are typically of singlet multiplicity, but the oxygen we breathe is a paramagnetic molecule of triplet multiplicity. In accordance with the Wigner spin conversation rules, reactions of this form of oxygen with biomolecules are of low probability (Allen, 1994). However, ground state triplet oxygen (302) can be converted to singlet oxygen in a type II reaction with the triplet electronically excited photosensitizer molecule: 3 Photosensitizer* 302 1 Photosensitizer 102* This triplet-triplet annihilation proceeds through a singlet reaction surface, and yields the ground state singlet photosensitizer molecule and singlet molecular oxygen (102*) as products. The 1 Ag electronically excited form of 02 is 22.5 kcal/mol more energetic than the 3 yg ground state 02. In its singlet state, 02 can directly participate in spin allowed reactions with singlet multiplicity biomolecules. As such, the electrophilic oxygenating activity of 102* can be directed to the killing of bacteria or other pathogens, or to viral inactivation: 102* Microbe Microbe Killing or Inactivation. (6) Photosensitizer-mediated inactivation of viruses in blood occurs principally through such a type I singlet oxygen mechanism, whereas both type I and II mechanisms contribute to erythrocyte cytotoxicity (Rywkin et al., 1992).

Most singlet electronically excited molecules have relatively short lifetimes, 10-8 second. Their ephemeral nature limits their participation in chemical reactions. Their fate is to return to ground state by photon emission. Fluorescence is the energy product of singlet excited molecules relaxing to their singlet ground state.

Triplet electronically excited molecules have relatively long lifetimes, 10- 2 to 102 seconds. Phosphorescence is the energy product of triplet excited molecules relaxing to their singlet ground state. The Equation 2 transition of an excited singlet state to an excited triplet state by intersystem crossing is a low probability event. Therefore, the excited triplet is in a meta-stable state and can participate in chemical reactions.

Unlike the fluorescence of singlet*-4singlet transitions, the phosphorescence of triplet*-+singlet transitions is highly susceptible to chemical quenching, especially to 02 quenching. The relatively longer lifetime of the triplet excited state is a consequence of the low probability of the spin transition.

WO 97/15661 PCT/US96/17136 -4- Oxygen is unusual in that its ground state is triplet. Relaxation from its singlet excited state to ground state triplet also requires intersystem crossing but in the opposite direction. The low probability of this transition is responsible for the relative meta-stability of 102*. The lifetime of 102* in aqueous solution is reported to be in the microsecond range (Merkel and Kearns, 1972). As such, 102* can serve as a chemical reactant, but because of its brief lifespan, its radius of reactivity is confined to within 0.1 to 0.2 micrometer of its point of generation (Lindig and Rodgers, 1981).

Haloperoxidase Oxygenating Activity: Haloperoxidases, such as myeloperoxidase (MPO) and eosinophil peroxidase (EPO), can catalyze the hydrogen peroxide oxidization of chloride or bromide to hypochlorous or hypobromous acid (Allen, 1975a,b): Cf H 2 0 2 MPO-4 HOCI H 2 0. (7) The reaction of singlet multiplicity hypohalous acid with an additional singlet multiplicity hydrogen peroxide proceeds through a singlet reaction surface: HOCI H 2 0 2 Cl H 2 0 2 102, (8) and as such, all of the reaction products including oxygen will be of singlet multiplicity. Thus, haloperoxidases provide an enzymatic system for generating 102*, the same reactive product that can be generated by photosensitizers.

Carbohydrate-Lectin Mechanisms in Microbial Infection and Defense: The mechanisms by which viruses infect cells generally employ lectins, which are a family of proteins that bind tightly to specific sugars on glycoproteins or glycolipids. The example of influenza virus illustrates how a viral lectin participates in the infection process. To access the cytoplasmic compartment of a victim cell, the hemagglutinin protein on the surface of an influenza virus particle will first attach to the cell by binding to sialic acid on the cell's membrane (Weis et al., 1988). Thus, influenza hemagglutinin is a lectin that binds specifically to sialic acid.

Similarly, HIV-1 and HIV-2 also possess lectin-like activity that may be required for cell binding and infection. Two envelope glycoproteins of the human immunodeficiency virus (HIV-1), appear to have lectin-like activity. The HIV-1 outer -membrane protein, gpl20, and the transmembrane protein, gp41 are both derived from cleavage of a gpl60 precursor, and both behave as mannosyl and N-acetylglucosaminyl (GlcNAc) binding lectins (Haihar et al., 1992a and The a-D-mannosyl and a-D-glucosyl residues of glycoproteins on host cell membranes may serve as anchoring sites for these HIV-1 proteins, thus facilitating viral entry into WO 97/15661 PCTIUS96/17136 the cell. This mechanism may account for the infection of non-CD4 cells and/or may assist in the CD4-dependent infectious mechanism.

Other lectins are capable of binding to bacteria and viruses. The mannosylspecific serum lectin, mannan-binding protein (MBP) is an acute phase protein linked to broad spectrum bacterial recognition and complement activation (Holmskov et al., 1994). Salmonella montevideo containing mannose-rich lipopolysaccharide shows a high level of MBP binding (Van Emmerick et al., 1994). MBP also binds to HIV-1 envelope glycoprotein and inhibits the in vitro infection of cells by HIV-1 (Lifson et al., 1986; Robinson et al., 1987; Ezekowitz et al., 1989: Hansen et al., 1989). Concanavalin A (ConA), a plant lectin with affinity for o-D-mannosyl and ac -D-glucosyl residues, also attaches to HIV-1, although it does not interfere with viral envelope glycoprotein binding to CD4 (Gattegno et al., 1992). However, the infectivity of Con-A bound HIV for lymphoid cells is reduced in a carbohydrate specific, dose-dependent manner (Gattegno et al., 1992).

Many bacteria are known to bind and agglutinate erythrocytes and this agglutination is inhibited in some cases by specific simple sugars, indicating that bacterial lectins mediate this binding. Mannose-specific surface lectins are present on Escherichia coli, Klebsiella pneumonia, Salmonella spp., Serratia marcescens, Shigella spp., Enterobacter spp., and Erwinia spp. (Duguid and Old, 1980; Speert et al., 1984). An E. coli lectin has also been reported with binding specificity for N-acetylglucosamine (Vaisanen-Rhen et al., 1983). These and other lectins with a variety of carbohydrate specificities are found on filamentous bacterial appendages called fimbriae or pili that mediate binding of bacteria to mucosal surfaces and provide the anchoring mechanism required for initiation of infection. The relative specificity ofE. coli Type 1 fimbriae lectin binding to various mannose-derivatives and complex mannans has been reported (Firon et al., 1983). Lectin binding was directed to short oligomannose chains of N-linked glycoproteins. Such structures are common to mucosal cell surfaces (Sharon and Lis, 1982). The same pattern of lectin specificity was observed from several members of the same bacterial genus (Firon et al., 1984), suggesting that mannose may play a major role in infection by a wide variety of bacteria.

Binding and Microbicidal Properties of Haloperoxidases: The leukocyte haloperoxidases myeloperoxidase (MPO) and eosinophil peroxidase (EPO) are cationic glycoproteins that selectively bind to and kill bacteria (Allen, 1992, U.S.

patent application Serial No. 07/660,994; Allen, 1995, U.S. No. 5,389,369). MPO and EPO binding and killing specificity for certain gram negative microbes involves a WO 97/15661 PCTIUS96/17136 -6lectin-carbohydrate binding mechanism. Both MPO and EPO are rich in mannose and N-acetylglocosamine (Yamada et al., 1981; Miyasaki et al., 1986; Olsen et al., 1985; Allen, unpublished observations of lectin binding specificities. As such, MPO and EPO can be bound by the mannose-specific pili lectins of many gram-negative microbes including Escherichia coli, Kelbsiella pneumonia, Salmonella spp., Serratia marcescens, Shigella spp., Enterobacter spp., and Erwinia spp. (Duguid and Old, 1980; Speert et al., 1984). An E. coli lectin with N-acetylglucosamine binding specificity has also been reported (Vaisenen-Rhen et al., 1983). The binding of MPO and EPO to these lectin bearing bacteria is essentially of relatively high affinity and is essentially independent of ionic strength.

As discussed above, HIV-1 and HIV-2 bind through their envelope proteins with a.-D-mannosyl and N-acetylglucosaminyl carbohydrate structures. Accordingly, both viruses can selectively bind with MPO and EPO. MPO and EPO are known to inactivate viruses, including HIV-1 (Klebanoff and Belding, 1974; Klebanoff and Coombs, 1991). Furthermore, NIH-sponsored contract studies using ExOxEmis prepared MPO and EPO demonstrate high leval inactivation of AZT-resistant HIV-1.

Molecular and Particle Sieving (Gel Filtration): Microporous substances exist that are interlaced with regularly spaced channels of subcellular dimensions. The network of pore channels can comprise a substantial proportion of the total volume of the microporous substance. Porous substances include crystalline aluminosilicates or aluminophosphates, and gels composed of cross-linked dextran.

Porath and Flodin (1959) introduced the concept of using cross-linked dextran beads, Sephadex (Pharmacia), as a chromatography method for rapid separation of macromolecules based on size. Although commonly referred to as gel filtration, the procedure does not require the use of gel and is not a true filtration. The threedimensional molecular network of the bead consists of many open pores of a certain size range. If the pore size is too small to allow entry of a large molecule, the molecule will be excluded, whereas molecules smaller than the pore size can enter the pore matrix. Molecules too large to enter the pore matrix are thus excluded from this compartment of the gel, and will travel flow through the gel at the same speed as an aqueous solution. Much larger amounts of solution are required to elute the smaller molecules that have accessed the labyrinth within the pores. The smaller porepenetrating molecules will be retained by an equilibrium of molecules entering and leaving the pores. The size and shape of the small molecules and the size and shape of the pores will determine the equilibrium condition, and as such, the degree of physical association determines the delay in elution time.

WO 97/15661 PCT/US96/17136 -7- The gel filtration technique has been used to isolate and purify viruses. For example, bovine papilloma virus has been purified from crude extracts of bovine warts by gel filtration using Sephacryl® S-1000 Superfine (Pharmacia Fine Chemicals, Uppsala, Sweden). Hjorth and Moreno-Lopez, J. Virol. Methods, 5:151-158 (1982).

Similarly, barley yellow dwarf virus has been purified from infected oats using gel filtration on Sephacryl® S-1000 Superfine. Hewish and Shukula, J. Virol. Methods, 7:223-228 (1983). In both instances, the purity (as judged by SDS-PAGE) and yield of the isolated viruses was high, thus illustrating the selectivity and capacity of gel filtration solid supports for sequestering virus particles in solution. These results demonstrate that a virus particle may be readily separated from larger particles by the gel filtration technique.

A variety of molecules can be covalently bound to sieving materials that have been chemically activated. A number of well-known techniques are available for these purposes, as provided in Affinity Chromatography (Pharmacia Fine Chemicals), which is hereby incorporated by reference.

Summary of the Invention The invention provides a microporous substance having pores sized to permit entry of pathogenic particles but exclude blood cells, wherein a singlet oxygen generating system is bound to the microporous substance. Representative microporous substances for this purpose include controlled pore glass and carbohydrate polymers. In a representative embodiment, the pores are sized in the range of from about 0.1 gm to about 1.0 pm to permit entry of virus particles.

Alternatively, the pores may be sized up to about 3 lm to permit entry of bacteria.

The singlet oxygen generating system may be a light-activated photosensitizer or an enzymatic system. Suitable photosensitizers include phtholocyanines, porphyrins, methylene blue, hypericin, fluorescein derivatives, and psoralen. A preferred enzymatic system for generating singlet oxygen includes an oxidase and a haloperoxidase. The oxidase is preferably capable of oxidizing a substrate present in blood. Suitable oxidases for this purpose include lactate oxidase, oxalate oxidase, glucose oxidase, and cholesterol oxidase. The haloperoxidase may be selected from among myeloperoxidase, eosinophil peroxidase, and fungal chloroperoxidase. The singlet oxygen generating system may be bound to the microporous substance by covalent bonding or non-covalent binding. The microporous substance may constitute or be configured into a device such as a bead, wafer, gel filtration matrix, filter, bag, or tubing.

I

WO 97/15661 PCT/US96/17136 -8- In a particularly preferred embodiment, a ligand that binds to a pathogenic particle is also bound to the microporous substance. The ligand may be selected from among high mannose glycans, N-glucosamine, the N-acetylglucosaminyl core of oligosaccharides, the monosyl core of complex-type N-linked glycans, mannan, amethylmannoside, haloperoxidases, sulfated polysaccharides, low molecular weight dextran sulfate, and lectins. Preferred ligands for this purpose include the haloperoxidase myeloperoxidase and the lectin conconavalin A. In a representative embodiment, the microporous substance takes the form of a carbohydrate polymer bearing conconavalin A to which lactate oxidase and myeloperoxidase are respectively bound.

Brief Description of the Drawings FIGURE 1 depicts a microporous substance in contact with blood, and illustrates how a pore diameter of 1 pm effectively excludes red blood cells (RBC) but permits viral particles to enter the pores.

FIGURE 2 illustrates a pore channel (lumen) of a microporous substance (sieve material) to which myeloperoxidase (MPO) and oxidase have been covalently bound. Schematically illustrated are oxidase and MPO-catalyzed reactions producing singlet oxygen in close proximity to virus particles that have been captured by a ligand on the MPO or that are simply present within the pore lumen.

FIGURE 3 illustrates the effective radius (overlapping red circles) within which the virus-killing activity of singlet oxygen is confined due to its brief lifetime.

The lifetime of singlet oxygen in aqueous solution is reportedly in the microsecond range, hence its radius of reactivity is limited to within 0.1 to 0.2 pm of its generation.

FIGURE 4 illustrates sieve material bearing conconavalin A to which MPO is non-covalently bound. An oxidase is bound to the sieve material. Virus are selectively retained within the pores by ligand-receptor interaction with the conconavalin A, the MPO, or both. Singlet oxygen generated by the immobilized enzymes kill or inactivate the virus.

Detailed Description of the Preferred Embodiment This invention provides a microporous substance for purifying a biological fluid such as blood of pathogenic particles like virus and bacteria. The microporous substance has pores that are sized to permit the entry of pathogenic particles but exclude desirable cells in the biological fluid, such as the cellular elements of blood.

Pursuant to the invention, a singlet oxygen generating system is bound to the microporous substance. The microporous substance forms a sieving matrix in which pathogens viruses, bacteria, prions, or other similarly-sized pathogenic particles) WO 97/15661 PCT/US96/17136 -9are removed and inactivated from blood and other biological fluids. In an illustrative embodiment, the enzyme components of a singlet oxygen generative system are bound to the channels of a sieving gel. Pathogen-binding ligand molecules may also be bound to the matrix channels.

The nature of the sieving gel, and particularly its pore size, defines the size of the molecules or particles retained by the material. Generally, gel particles are polymeric materials including natural and synthetic polymers. A polymer is a high molecular weight compound consisting of smaller subunits that are covalently linked together to form long chains that are straight or branched and that also often have interchain linkages among the subunits. Polymeric materials used in gel filtration include natural polysaccharides such as agarose, dextran, and cellulose, and synthetic polymers such as polyacrylamide, polytrisacrylamide, and hydroxylated vinyl polymers. These polymers may be formed into particles (or beads) that have varying pore size. The pore size of the matrix may be controlled by the extent of cross-linking between the polymer chains of the gel particle. Generally, highly cross-linked polymers have smaller pore sizes than those polymers that are lightly cross-linked.

For the subject invention, the porous sieving matrix can consist of cross-linked carbohydrate, polydextran, controlled pore glass, or other appropriate material.

A number of these materials are commercially available, including cross-linked agarose and bisacrylamide cross-linked allyl dextran (Sepharose® and Sephacryl® gels, respectively; Pharmacia Fine Chemicals, Uppsala, Sweden). In these gels, the polymeric material is a polysaccharide cross-linked to varying degrees with bisacrylamide.

A variety of devices can be readily constructed from the sieving material, for example, beads, tubes, flat porous surfaces, lining plastic blood bags, filters, hollow cellulose tube filters, and gel filtration matrices. The sieving material will contain pores, a pore being a channel or void or intersticial space within a solid material that communicates with the outside liquid and permits the passage of liquids or gases through the material. The diameter of the pore channel will be large enough to allow the entry of pathogenic viruses, bacteria, or other similarly-sized pathogens. As illustrated in FIGURE 1, the pore diameter will be small enough to exclude desirable cells, particularly the cellular components of blood.

As described in Table 1, virus particles that contaminate blood range in diameter from 20 (parvoviruses) to 150 nm (0.02 to 0.15 gm). Platelets, the smallest cellular elements of blood have a free-floating diameter of approximately 3,000 nm (3 lm). Erythrocytes are next in size with a diameter of 7.2 to 7.9 upm. As such, any WO 97/15661 PCT/US96/17136 material with pore sizes in the range of 155 nm (about 0.15 rm) to about 2,500 nm (or less than 3 gim) will physically accommodate virus particles while excluding platelets and other cellular elements of blood.

Pore sizes of about 2.5 gim are particularly useful for sequestering non-viral pathogens having diameters less than the diameter of platelets. Bacteria are typically less than 2 pm in diameter, and most or all bacteria will enter pore channels having diameters of about 2-3 gim. Thus, pore sizes just slightly smaller than the 3 gm diameter of blood platelets will accommodate most pathogens that are expected to contaminate blood supplies.

In a first embodiment of the subject invention, a photosensitizer, such as a fluorescein derivative or a phthalocyanine derivative or other appropriate molecule, is bound covalently or non-covalently to the surfaces of the channels present in the pores of the sieving material. A "photosensitizer," sometimes known also as a "photodynamic dye," is a compound or a molecule that is capable of transferring its excited state energy to another compound or molecule. A photosensitizer becomes electronically excited after absorbing a photon of light, and subsequently transfers its excited state energy to another compound or molecule. A photosensitizer that transfers its excited energy to an oxygen molecule ground state oxygen, 302) thereby generating an excited oxygen molecule excited state oxygen, 102) a is a singlet oxygen photosensitizer.

Photosensitizers are characterized by the property of being converted to the triplet excited state when illuminated with light of an appropriate wavelength. The excited triplet photosensitizer molecule then participates in a type II reaction with ground state triplet oxygen present in the blood to generate highly reactive singlet oxygen, 102*, which is capable of killing pathogenic organisms.

Photosensitizers contemplated by the subject invention are those that absorb visible light (from about 350 nm to about 700 nm) and near-infrared light (from about 700 nm to about 1000 nm). Photosensitizers preferred for purifying blood include compounds that absorb visible light in the red region (from about 600 nm to about 700 nm) and in the near-infrared region, as the light required to activate these photosensitizers will not be absorbed by the red blood cells. Exemplary families of photosensitizers suitable for the subject invention include acridines, fluoresceins, xanthines, rhodamines, porphyrins, cyanines, and phthalocyanines. Specific photosensitizers useful for the subject compositions include methylene blue, hypericin, hematoporphyrin, fluorescein, and fluorescein derivatives such as rose bengal, eosin, and erythrosin.

WO 97/15661 PCT/US96/17136 -11- In a related embodiment, a virus or bacteria-binding composition is also bound covalently or non-covalently to the surface of the pore channels to aid in the capture and retention of pathogen particles. The enhanced capture of viruses can be accomplished by linking substances that can serve as ligands, such as concanavalin A or other lectins, to the surfaces of the pore channels. The capture of additional pathogens can be accomplished by coating the pore channels with substances such as sialic acid, mannose, N-acetylglucosamine, haloperoxidases such as myeloperoxidase or eosinophil peroxidase, or any other substance known to have the capacity of binding with molecules present on the surfaces of viruses and bacteria. A significant advantage of this embodiment is that the ligand-mediated binding of pathogens in and of itself partially serves the purpose of removing pathogenic particles from the blood.

The major advantages of using photosensitizers attached to microporous sieve material is that this approach obviates the problem of the photosensitizers binding to essential blood components, eliminates toxicity associated with introducing photosensitizers into the circulation, and also minimizes damage to cellular components. The relative disadvantage of this approach is that light is required to drive the photodynamic action. However, the 3 1g-+ 1 Ag electronic excitation of 02 is a relatively low energy transition requiring about 23 kcal/mol. For purification of blood, red light, which is not absorbed by hemoglobin, and even near infrared light will drive this process with an adequate transducer molecule.

In a second and preferred embodiment of the invention, the sieve material is bound covalently or non-covalently with an enzymatic system capable of chemically generating HOCI and 102. One such system is comprised of the previously described haloperoxidases, which require halide and H202 as substrates generating singlet oxygen. Haloperoxidases catalyze the hydrogen peroxidase oxidation of chloride or bromide to hypochlorous or hypobromous acid. The amount of chloride normally present in blood provides a sufficient amount of halide for reactions catalyzed by myeloperoxidase (Allen, U.S. Patent No. 5,389,369). The hypohalous acid thus produced reacts with an additional hydrogen peroxide molecule to yield singlet molecular oxygen. Virus or bacteria that are in the immediate vicinity of the reactive singlet oxygen molecules will be inactivated or killed. Blood components that are too large to enter the pores will not be significantly exposed to either H202 or singlet oxygen. Haloperoxidases suitable for the subject invention include myeloperoxidase, eosinophil peroxidase, and fungal chloroperoxidase.

Oxidases capable of acting on substrates normally present in the blood can be conveniently employed to provide the hydrogen peroxide substrate required by the WO 97/15661 PCTIUS96/17136 -12haloperoxidase. Such oxidases include lactate oxidase, oxalate oxidase, cholesterol oxidase, glucose oxidase, or other oxidases capable of using substrates in the blood.

If insufficient amounts of, lactate or glucose are present, these substrates could be added to the blood or to the blood-collecting bags to ensure sufficient hydrogen peroxide generation by the chosen oxidase. Lactate oxidase produces H 2 0 2 in the following reaction: Lactate +02 Pyruvate H 2 0 2 (9) In Equation peroxide is generated from lactate by L-lactate:0 2 oxidoreductase EC 1.1.3.x (Naka, 1993). Lactate is an excellent substrate in that it is typically present in a relatively high concentration in blood. The venous blood of normal healthy adults has a lactate concentration of 1 mm (9 mg/dL). Lactate is the end product of glycolytic metabolism and is the major product of erythrocyte metabolism of glucose. Every molecule of glucose metabolized yields two molecules of lactate. Lactate continues to be generated by erythrocytes during blood bank storage. The peroxide generated by lactate oxidase thus provides a substrate for the sieve-bound haloperoxidase.

Suitable substrates for other oxidases are also present in the blood, e.g., oxalate oxidase (oxalate:0 2 oxidoreductase, EC cholesterol oxidase (cholesterol:0 2 oxidoreductase, EC glucose oxidase, etc., can be employed in a similar manner. These and other oxidases are well-known and several of them are described in some detail in U.S. Patent No. 5,389,369.

FIGURE 2 schematically depicts a section through a virus-entrapping pore with surface bound lactate oxidase and myeloperoxidase. This pore is approximately 0.2 to 0.3 p.m in diameter and contains two viral particles the size of HIV, i.e., approximately 0.1 pm in diameter. Larger pores can be used to entrap bacteria and even cellular pathogens. Aside from the advantage of capturing pathogen particles, the isolated milieu of the pore also serves to concentrate and link the substrates and products required for effective enzymatic interaction. FIGURE 2 further illustrates how the oxidation of lactate to pyruvate drives the generation of 102* in close proximity to the captured virus particles. FIGURE 3 depicts the range (overlapping circles) within which the activity of 102* is limited because of its brief lifespan.

For the subject enzyme systems, conconavalin A can be used to noncovalently bind the haloperoxidase to the microporous substratum. Conconavalin A is a protein lectin derived from the jack bean. The conconavalin A molecule has four sites capable of binding non-covalently to carbohydrates, and thus has the capacity to WO 97/15661 PCT/US96/17136 -13agglutinate a variety of animal cells. As illustrated in FIGURE 4, the tetravalent molecule concanavalin A can serve as a "platform" for bringing several components of the composition into proximity within the pores of the microporous matrix. To accomplish this multi-component proximity, one of the reactive sites of the conconavalin A is first covalently bound to the microporous substrate itself, and then the lectin's reactive sites are allowed to bind with haloperoxidase and with mannosylated oxidase. Under empirically adjusted stoichiometric conditions, the lectin molecules will still have reactive sites available to bind with a receptor present on a virus or bacterium. This strategy is illustrated in FIGURE 4, which depicts covalently bound conconavalin A to which MPO is non-covalently bound. In this example, the oxidase is bound directly to the sieve material, but in other applications it can be mannosylated and bound non-covalently to the conconavalin A.

Confining H202 generation to within the pore environment has multiple advantages. It maximizes contact between pathogen particles and pore-bound myeloperoxidase, limits H202 dilution, and insures that H202 will not contact with erythrocytes (and thus avoids the conversion ofH202 by erythrocyte catalase). Blood cellular components are also protected by limiting the destructive effect of H 2 0 2 to the pore environment. Most importantly, this confines peroxide generation by the oxidase to the location of MPO. Confinement to the pore space also insures that HOCI, the product of H 2 0 2 -dependent MPO-catalyzed chloride oxidation, will react with an additional H202 to yield 102*. Enzymatic generation of 102* within the pore space ensures a high level of focused anti-pathogen activity without requiring an exogenous source of energy. Unlike systems reliant on photosensitizers, the enzymatic generation of 102* has the advantage of not requiring an external light source. Spatial confinement optimizes the reactive potential of 102* with its limited lifetime by localizing the pathogen to within the less than 0.2 gm radius of reactivity required for maximum viricidal or bacteriocidal action.

The invention thereby provides compositions useful for removing pathogenic particles, particularly viruses, bacteria, and prions, from blood or other fluids. For example, the subject microporous compositions can be used to remove viruses from pharmaceutical preparations that are produced by cultured eucaryotic or procaryotic cells. The compositions moreover can be applied to remove pathogens from virtually any fluid whose purity is desired.

The subject compositions are composed in whole or part of a microporous substance configured so that its pores open onto a surface that interfaces with the liquid to be purified. Representative microporous substances include controlled pore WO 97/15661 PCTIUS96/17136 -14glass and polymers such as cross-linked carbohydrates. These pores are of appropriate size to permit the entry of pathogenic particles such as viruses and bacteria. Methods for creating pores of the desired diameter are known in the art.

For example, the size of pores in cross-linked polymers can be controlled by manipulating the amount of cross-linkers, while the pores in glass beads can be established by controlled etching. Particularly useful are pores ranging in size from about 0.1 pm to about 1.0 im to permit entry of virus particles, and pores sized up to about 2.5 pm to permit entry of bacteria. Pores of less than about 2.5 Lm in diameter provide the advantage of excluding cellular components of the fluid, blood, from the interstices of the porous substance.

The subject compositions may also employ one or more ligand-receptor pairs.

Generally, a ligand is a molecule or group of molecules that binds to a receptor molecule or site on the surface of a cell or virus. Such a pair consists of two molecules with a sufficient affinity for one another such that if one member of the pair is bound to the microporous substance and the other member is present on the surface of a microbe, a virus or bacterium, the microbe will become at least temporarily attached to the microporous substance by virtue of the non-covalent binding between the two members of the pair. For the subject compositions, the member of the ligandreceptor pair that is bound to the microporous surface is referred to as the "ligand", and the ligand-binding molecules on the surface of a virus or bacterium is the "receptor". Conconavalin A is an example of a ligand capable of binding mannan oligosaccharide receptor(s) on viruses such as HIV-1.

A variety of materials may form components of the ligand-receptor pairs, including carbohydrate moieties, glycoproteins, lectins, and others. Suitable ligands include high mannose glycans, mannosylated surfaces, the manosyl core of complextype N-linked glycans, N-acetylglucosamine and its derivatives, mannan, sulfated polysaccharides such as heparin or low molecular weight (MW) dextran sulfate (DS), and lectins such as conconavalin A and other lectins that share conconavalin A's affinity for a-D-mannosyl, a-D-glucosyl residues, and N-acetylglucosamine.

Conconavalin A attaches to HIV-1 (Gattegno et al., 1992), and can also bind to EPO, MPO, and mannosylated oxidase.

The subject compositions require that the ligands be bound to the microporous substance, especially to the pores that permeate this substance. Methods are widely available for covalently linking these types of molecules to polymeric surfaces.

The subject microporous substances can be configured into a variety of devices. Such devices include beads, gel filtration matrices, filters, bags, tubing, and WO 97/15661 PCT[US96/17136 filters composed of hollow cellulose tubules. For example, to remove microbes from blood, the blood is passed ex vivo through a filter or a column containing the microporous substance. The blood can be filtered prior to being placed inside storage bags, or filtered upon removal from the bags, or both. The inner surface of bags in which blood is stored can be coated with a thin layer of the microporous substance, or tubing lined with the substance or beads or wafers composed of the substance can be simply placed inside the bag to inactivate particulate viruses and/or bacteria while the blood is stored in the bag. These and other devices satisfy the criterion that the microporous substance contacts the fluid to be purified so that an equilibrium can be established between pathogens within and without the pores.

EXAMPLES

1. Controlled pore microporous substances.

Any composition having pores sized to the required dimensions can be employed as the microporous substance, the sieve material. Several different types of sieve materials are commercially available. Sephacryl S-1000 (Pharmacia, Sigma S-1000) is a cross-linked co-polymer of allyl dextran and N,N'methylenebisacrylamide with a bead diameter of 40-105 pm and a pore-size exclusion limit of 1 x 10 9 daltons. Sepharose CL 2B (Pharmacia, Sigma CL-2B-300), a crosslinked agarose with a bead diameter of 60-200 pm and a pore-size exclusion of 0.2 x 10 9 daltons, may also be used with small viruses. Controlled pore glass is prepared by heating treating borosilicate glass and leaching out the boric oxide with acid. Controlled pore Glyceryl Glass (Sigma GG3000-200) is available in 120-200 mesh (75-125 pm in diameter) with a pore diameter of 0.3 lpm (300 nm).

Many techniques are available for chemically activating sieve materials and covalently binding proteins to the activated materials (Jakoby and Wilchek, 1974).

The following are examples of CNBr-activation of Sephacryl, Sepharose, and glyceryl glass, and the binding of MPO, EPO, lactate oxidase, and concanavalin A to the activated sieve material.

2. Preparation of Sieve Materials.

The following preparatory steps were seperately applied to: Sepharose 2B (Sigma CL-2B-300), 20 ml stock suspension in 20% EtOH; Sephacryl S-1000 (Sigma S-1000), 20 ml stock suspension in 20% EtOH; and Glyceryl-Glass, Controlled Pore (Sigma GG3000-200, 10 ml (10 cc) solid.

For the first step, the sieve materials were brought up to 50 ml with distilled

H

2 0, and placed on a rocker for 30-60 min. Next, the material was allowed to settle, and the aqueous supernatant was decanted. The material was washed as before with a WO 97/15661 PCTIUS96/17136 -16second 50 ml of distilled water. After the second water wash had been decanted, 2 M phosphate buffer pH 11.4, (536g Na 2

HPO

4 -7H 2 O/L H20 and adjusted with NaOH to pH 11.4) was added to a final volume of 50 ml. The solution was placed under vacuum for 1 hour to degas, then placed into an ice bath to cool.

3. Activating the Sieve Material with Cyanogen Bomide (CNBr) (Sigma C-6388).

A highly buffered alkaline cyanogen bromide (CNBr) method (Porath, 1974; Parikh et al., 1974) was used for bead activation and protein coupling. Two g of CNBr crystals were added to 20ml cold distilled H 2 0 and mixed to dissolve. (The crystals were only 50% dissolved after 30 minutes.) To each of the ice cold bead suspensions was added a total of 4ml of CNBr (approx. 50mg/ml for 200 mg total) in ml additions over a 10 min interval. The beads were washed over a 0.2 l filter with 4 volumes of H 2 0 at 4 0

C.

4. Covalently Linking Proteins to the Cyclic Imino Carbonates on the Beads.

MPO-Lactate Oxidase (MPO-LOx), EPO-LOx, and concanavalin A were prepared for bead coupling as follows: MPO-LOx (and lactate oxidase): Five ml of myeloperoxidase (10 mg, porcine, ExOxEmis, Lot 1899201, 70 nm total) was combined with 3.2mg lactate oxidase, Pediococcus species (3.2mg, 108 units, Sigma L-06380).

EPO-LOx (and lactate oxidase): Five ml of eosinophil peroxidase mg, porcine, ExOxEmis, Lot 1929201, 150 nmol total) was combined with 3.2mg lactate oxidase (Pediococcus species, 3.2mg, 108 units, Sigma L-06380).

Con-A: Ten mg of concanavalin A, Type IV Canavalin ensiformis (Sigma C-2010) was dissolved in 5ml distilled Similarly, an oxidase capable of using substrates present in blood could be concomitantly prepared and permitted to bind to the activated beads.

Beads, activated as described above, were resuspended in 0.25 M bicarbonate buffer pH 9.0 (21g NaHCO3/L H 2 0 adjusted to pH with NaOH) to a final volume of 30ml. Each 30ml bead suspension was divided equally into 3 tubes, for a total of nine tubes, including three each of Sepharose 2B, Sephacryl S-1000, and Glyceryl-Glass as described above. To each tube of bead preparations was added: 1.75 ml of MPO-LOx, 1.75 ml of EPO-LOx, and 1.75 ml of Con-A. The tubes were then placed on a rocker table for gentle mixing and allowed to react overnight (12 hr) at 22 0

C.

On completion of the coupling, a 1 M glycine solution was prepared (75 g glycine/) and added to each bead suspension to a final volume of 50ml. The glycine is provided to react with any unreacted iminocarbonates remaining on the beads.

WO 97/15661 PCT/US96/17136 -17- It was apparent from the color of the derivatized beads that the haloperoxidases had successfully bound to the beads. The bead preparations, once they had settled, displayed the pale green color characteristic of MPO and the orangebrown color ofEPO. UV-Visible spectroscopy was performed to quantify the protein remaining in the supernatant after the beads had settled. The amount of MPO bound to each type of bead was quantified by measuring the reduced (dithionite)-oxidized difference spectrum at 475 nm using a millimolar extinction coefficient of 78 mMlcm-1: Sepharose 2B wash 9.3 nmol unbound 60% bound Sephacryl S-1000 wash 0.3 nmol unbound 99% bound Glyceryl-Glass 2.9 nmol unbound 87% bound.

Results indicated that 23.3 nmol MPO had bound to each type of bead.

The amount of EPO bound to the beads was quantified as the reduced (dithionite)-oxidized difference spectrum at 449 nm using an extinction coefficient of 125 mM-lcm-1: Sepharose 2B wash 1.8 nmol unbound 96% bound Sephacryl S-1000 wash 1.4 nmol unbound 97% bound Glyceryl-Glass 0.1 nmol unbound 99% bound.

Results indicated that 50 nmol EPO had become bound to each type of bead.

Beads bound to either MPO or EPO were washed one additional time with Dulbecco's Phosphate Buffer, pH 7.4 and adjusted to a final volume of 10 ml in Dulbecco's.

Experiments were not done to determine the maximum amount of MPO and EPO capable of binding to each type of activated bead. However, results not shown here suggest that the binding capacity of the beads is much greater than the amounts that were bound in this experiment. The amount bound can be maximized by adding increasingly higher concentrations of MPO and EPO during the binding step until an amount is determined beyond which further increases do not result in increased amounts of protein bound to the beads.

5. A Bacteriophage Model of Pore:Lactate-Oxidase-Haloperoxidase Virus Inactivation.

Like eukaryocytes, bacteria are susceptible to virus infection and lysis. This type of virus is called a bacteriophage or phage. Bacteriophages are comparable to animal viruses in size, and their presence can be measured by their ability to form plaques, zones of lysis or clearing, produced on a "lawn" of susceptible bacteria.

WO 97/15661 PCT/US96/17136 -18- The number of bacteriophage can be measured as the number of plaque forming units per volume of sample.

Bacteriophage are not enveloped, and present a different surface than the HIV and hepatitis viruses of interest with regard to blood infectivity. However, their size and ease of detection allow testing of the size specific inactivating quality of the subject Pore:Oxidase-Haloperoxidase preparations.

Bacteriophage suspensions are contacted with the MPO- and EPO-bound beads prepared above. Bacteriophage enter the pores of the beads and interact noncovalently with the conconavalin A and/or the EPO or MPO present in the pore channels. After the beads are permitted to settle, the supernatant is collected and the titer of bacteriophage therein is determined using standard bacteriological techniques.

As controls, bacteriophage suspensions are contacted with beads that have been subjected to all of the above preparatory steps except that no MPO or EPO was added during the binding step. The effectiveness of phage killing by the MPO and EPO enzyme system bound to the beads is determined by comparing the resulting control and experimental titers.

6. A Bacterial Model of Pore:Lactate-Oxidase-Haloperoxidase Inactivation.

The functional capacities of untreated and MPO-LOx bound sieve material (porous beads) was tested using Escherichia coli, a short gram-negative bacillus with a cell width of approximately 0.5 Lm. The E. coli were cultured in trypticase soy broth for approximately 16 hours at 37 0 C. The bacteria were concentrated by centrifugation, and the final concentration adjusted to 106 to 107 bacteria per milliliter by turbidimetric measurement (titration).

The binding-killing capacities of the different sieve preparations described above were tested as follows. Into each test tube was placed: 100 ul (microliters) of the sieve material (beads) suspension; 100 gl of E. coli suspension (approximately 106 bacteria, total); 100 pl of 0.1 M lactic acid (10 mm lactate, final); and 700 .l of normal saline (0.85% NaCI).

The materials were gently mixed and allowed to incubate at 24 0 C for 1 hour.

The material were mixed to obtain a uniform suspension, diluted, and plated on trypticase soy agar. The number of E. coli colony forming units (CFU) were determined after 20 hours incubation at 37 0

C.

The results were as follows: WO 97/15661 PCTIUS96/17136 -19- Sieve Material E. coli (CFU/pl, final) NONE 4,520,000 Sepharose CL 2B untreated 3,000,000 MPO-LOx Bound 830,000 Sephacryl S1000 untreated 2,920,000 MPO-LOx Bound 200,000 Glyceryl-Glass untreated 3,160,000 MPO-LOx Bound 91,000 The small decrease in E. coli CFU observed with untreated sieve materials is thought to reflect binding and capture of the bacteria without killing. Notably, the removal and/or killing ofE. coli by the different MPO-LOx bound sieve materials was proportional to the pore size. Three different MPO-LOx-bound sieve materials were tested. Of these, the pore size of the glyceryl-glass was greater than that of the Sephacryl S1000, while the smallest pores of the group were those of Sephorose CL-2B. Accordingly, MPO-LOx-bound sephorose CL-2B was the least effective of the three sieve materials tested. This latter material also bound the least amount of MPO-LOx.

While the preferred embodiment of the invention has been illustrated and described, it will be appreciated that various changes can be made therein without departing from the spirit and scope of the invention. In particular, the compositions and methods described here can be applied to remove pathogenic microorganisms from any form of liquid.

WO 97/15661 PCT/US96/17136 Citations Raab, 1900 Matthews et al., 1988 Sieber et al., 1989, Blood 73:345-350.

Neyndorffet al., 1990, Transfusion 30:485-490.

O'Brien et al., 1990 Horowitz et al., 1991, Transfusion 31:102-108.

Lin et al., 1989 Prodouz et al., 1991 Dodd et al., 1991, Transfusion 31:483-490.

Margolis-Nunno et al., 1992 Transfusion 32:541-547.

Mohr et al., 1995, Immunological Investigations 24:73-85.

Lavie et al., 1995, Transfusion 35: 392-400.

Schenck and Koch, 1960 Allen, 1994 Merkel and Kearns, 1972, 7244-7253 Lindig and Rodgers, 1981 Allen, 1975 a,b Weis et al., 1988 Nature 333:426-431.

Haidar et al., 1992, Glycobiology 2:429-435.

Haidar et al., 1992, Glycoconjugate Journal 9:315-323.

Holmskoy et al., 1994 Van Emmerik et al., 1994, Clin. Exp. Immunol. 97:411-416.

Lifson et al., 1986, Journal of Experimental Medicine 164:2101-2106.

Robinson et al., 1987, Aids Research and Human Retroviruses 3: 265-282.

Ezekowitz et al., 1989, J. Exp. Med. 169:185-196.

Hansen et al., 1989 Gattegno et al., 1992, Aids Research and Human Retroviruses 8:27-37.

Duguid and Old, 1980 Speert et al., 1984 Vaisanen-Rhen et al., 1983, FEBS Letters 159:233-236.

Firon et al., 1983, Carbohydrate Research, 120:235-249.

Sharon and Lis, 1989, Science 246:227-234.

Firon et al., 1984 Allen, 1992, U.S.S.N. 07/660,994.

Allen, 1995, U.S.P.N. 5,389,369.

WO 97/15661 PCT/US96/17136 -21- Yamiada et al., 198 1, Biochemistry 20:766-77 1.

Miyasaki et al., 1986, Archives of Biochemistry and Biophysics 246:751-764.

Olsen et al., 1985, Biochem. J. 229:779-784.

Kiebanoff and Belding, 1974, The Journal of Infectious Diseases 129:345-348.

Klebanoff and Coombs, 1991, J. Exp. Med. 174:289-292.

Porath and Flodin, 1959, Nature 183:1657-1659.

Ifjorth and Moreno-Lopez, 1982, Journal of Virological Methods 5:151-158.

Hewish and Shukia, 1983, Journal of Virological Methods 7:223-228.

Naka, 1993 Jakoby and Wilchek, 1974 Porath, 1974 Parikh et al., 1974 The foregoing publications are incorporated by reference herein.

Claims (21)

1. A composition comprising a microporous substance having pores sized to permit entry of pathogenic particles but exclude blood cells, wherein a singlet oxygen generating system is bound to the microporous substance.

2. The composition of claim 1, wherein the microporous substance is selected from among the group consisting of controlled pore glass and carbohydrate polymers.

3. The composition of claim 1, wherein the pores are sized in the range of from about 0.1 rpm to about 1.0 gm to permit entry of virus particles.

4. The composition of claim 1, wherein the pores are sized up to about 3 tim to permit entry of bacteria.

The composition of claim 1, wherein the singlet oxygen generating system is a photosensitizer.

6. The composition of claim 5, wherein the photosensitizer is selected from among the group consisting of phtholocyanines, porphyrins, methylene blue, hypericin, fluorescein derivatives, and psoralen.

7. The composition of claim 1, wherein the singlet oxygen generating system is an enzymatic system.

8. The composition of claim 7, wherein the enzymatic system comprises an oxidase and a haloperoxidase.

9. The composition of claim 8, wherein the oxidase is capable of oxidizing a substrate in blood.

The composition of claim 9, wherein the oxidase is selected from among the group consisting of lactate oxidase, oxalate oxidase, glucose oxidase, and cholesterol oxidase.

11. The composition of claim 8, wherein the haloperoxidase is selected from among the group consisting of myeloperoxidase, eosinophil peroxidase, and fungal chloroperoxidase.

12. The composition of claim 8, wherein the haloperoxidase bears a ligand of a pathogenic particle.

13. The composition of claim 1, wherein the singlet oxygen generating system is bound to the microporous substance by covalent bonding.

14. The composition of claim 1, wherein the singlet oxygen generating system is bound to the microporous substance by non-covalent binding.

15. The composition of claim 1, wherein a ligand of a pathogenic particle is bound to the microporous surface.

16. The composition of claim 15, wherein the ligand is selected from among the group consisting of high mannose glycans, N-glucosamine, the N-acetylglucosaminyl core S of oligosaccharides, the monosyl core of complex-type N-linked glycans, mannan, a- S 15 methylmannoside, haloperoxidases, sulfated polysaccharides, low molecular weight dextran sulfate, and lectins.

17. The composition of claim 16, wherein the haloperoxidase is myeloperoxidase.

18. The composition of claim 16, wherein the lectin is conconavalin A.

19. The composition of claim 15, wherein the ligand is conconavalin A and the singlet oxygen generating system comprises myeloperoxidase that is non-covalently bound to the conconavalin A.

20. A composition comprising a microporous substance having pores sized to permit entry of pathogenic particles but exclude blood cells, substantially as hereinbefore described with reference to any one of the Examples.

21. A device comprising the composition of any one of claims 1 to 20, wherein the device is selected from among the group consisting of a bead, a wafer, a gel filtration matrix, a filter, a bag, and tubing. Dated 21 May, 1998 EOE, INC. Patent Attorneys for the Applicant/Nominated Person SPRUSON FERGUSON [n:\libc]03574:MEF