Method of Inactivation of Viral and Bacterial Blood Contaminants

This invention relates to the general field of biochemistry and medical sciences, and particularly to inactivation of viral contamination of compositions comprising peripheral blood cells (red blood cells, platelets, etc.), plasma protein fractions (albumin, clotting factors, etc.) from collected whole blood, the blood of virally infected persons, ex vivo, media used in the preparation of anti-viral vaccines, and cell culture media such as fetal bovine serum, bovine serum or products derived from these sources. The invention is also applicable to a method for the selective inactivation of cancerous cells.

A major concern in the transfusion of donated, stored whole human blood or the various blood cells or protein fractions isolated from whole blood is the possibility of viral contamination. Of particular concern are the blood-borne viruses that causes hepatitis and acquired immune deficiency syndrome (AIDS). While any number of cell washing protocols may reduce the viral contamination load for samples of blood cells, by physical elution of the much smaller virus particles, such washing alone is insufficient to reduce viral contamination to safe levels. In fact, some viruses are believed to be cell-associated, and unlikely to be removed by extensive washing and centrifugal pelleting of the cells. Current theory suggests that safe levels will ultimately require at least a 6 log (6 orders of magnitude) demonstrated reduction in infectious viral titer for cellular blood components. This 6 log threshold may be greater for plasma protein components, especially the clotting factors (Factor VIII, Factor IX) that are administered throughout the life of some hemophilia patients.

Viral inactivation by stringent sterilization is not acceptable since this could also destroy the functional components of the blood, particularly the erythrocytes (red blood cells) and thrombocytes (platelets) and the labile plasma proteins. Viable RBC's can be characterized by one or more of the following: capability of synthesizing ATP; cell morphology; P.sub.50 values; oxyhemoglobin, methemoglobin and hemochrome values; MCV, MCH, and MCHC values; cell enzyme activity; and in vivo survival. Thus, if lyophilized (freeze-dried) then reconstituted and virally inactivated cells are damaged to the extent that the cells are not capable of metabolizing or synthesizing ATP, or the cell circulation is compromised, then their utility in transfusion medicine is compromised.

Viral inactivation by stringent steam sterilization is not acceptable since this also destroys the functional components of the blood, particularly the blood cells and plasma proteins. Dry heat sterilization, like wet steam, is harmful to blood cells and blood proteins at the levels needed to reduce viral infectivity. Use of stabilizing agents such as carbohydrates does not provide sufficient protection to the delicate blood cells and proteins from the general effects of exposure to high temperature and pressure.

Methods that are currently employed with purified plasma protein fractions, often followed by lyophilization of the protein preparation, include treatment with organic solvents and heat or extraction with detergents to disrupt the lipid coat of membrane enveloped viruses. Lyophilization (freeze-drying) alone has not proven sufficient to inactivate viruses, or to render blood proteins sufficiently stable to the effects of heat sterilization. The organic solvent or detergent method employed with purified blood proteins cannot be used with blood cells as these chemicals destroy the lipid membrane that surrounds the cells.

Another vital inactivation approach for plasma proteins first demonstrated in 1958 has involved the use of a chemical compound, beta-propiolactone, with ultraviolet (U.V.) irradiation. This method has not found acceptance in the United States due to concern over the toxicity of beta-propiolactone in the amounts used to achieve some demonstrable viral inactivation and also due to unacceptable levels of damage to the proteins caused by the chemical agents. Concern has also been raised over the explosive potential for beta-propriolactone as well.

It is therefore a desideratum to devise an effective viral inactivation treatment for human blood components, which will not damage the valuable blood cells or proteins. The treatment must be nontoxic and selective for viruses, while allowing the intermingled blood cells or proteins to survive unharmed.

There is an immediate need to develop protocols for the deactivation of viruses that can be present in the human red blood cell supply. For example, only recently has a test been developed for Non A, Non B hepatitis, but such screening methods, while reducing the incidence of viral transmission, do not make the blood supply completely safe or virus free. Current statistics indicate that the transfusion risk per unit of transfused blood is as high as 1:3000 for Non A, Non B hepatitis, and ranges from 1:40,000 to 1:1,000,000 for HIV, depending on geographic location. Clearly, it is desirable to develop a method which inactivates or removes virus indiscriminately from the blood.

Contamination problems also exist for blood plasma protein fractions, such as plasma fractions containing immune globulins and clotting factors. For example, new cases of non A, non B hepatitis and hepatitis A have occurred in hemophilia patients receiving protein fractions containing Factor VIII which have been treated for viral inactivation according to approved methods. Therefore, there is a need for improved vital inactivation treatment of blood protein fractions.

The present invention thus provides a method for the inactivation of viral and bacterial contaminants present in blood and blood protein fractions.

The present invention also provides a method for viral or bacterial decontamination of frozen or lyophilized cells, recombinant protein preparations, blood components including red cells, platelets and leukocytes, stem cells, protein solutions or other frozen or lyophilized compositions intended for subsequent in-vivo use such as plasma derived factors. The present invention involves utilization of sensitizers which bind selectively to a viral nucleic acid, coat protein or membrane envelope. The sensitizer is also a moiety which can be activated upon exposure to radiation, which may be in the form of ultra-violet radiation or ionizing radiation, such as X-rays, which can penetrate the sample containing the contamination. Damage to cells in lyophilized preparations by hydroxy radicals is minimal due to the absence of water molecules. However, while not intending to be limited to a particular theory, in frozen cell or protein containing compositions, most of the water is present in the form of ice crystals but there is also noncrystalline water trapped in a highly viscous glassy state. Water molecules which are present in these glassy states have low mobility and may possibly form hydroxy radicals which can randomly damage cells. However, due to the low mobility in the glassy state, damage to cells from these hydroxy radicals is reduced. Therefore, by irradiating a frozen suspension of cells containing the sensitizers, random damage of the cells due to the hydroxy radicals may be avoided due to the inability of the sensitizer to migrate in the frozen suspension and the inability of the hydroxy radicals to form and migrate through the frozen suspension. In this manner, damage is localized on the targeted viral or bacterial particle.

The present invention is also applicable to inactivation of blood-borne bacterial contaminants, and to blood-borne parasitic contaminants, since such infectious organisms rely on nucleic acids for their growth and propagation. Since purified blood plasma protein factions are substantially free of human nucleic acids, and mature human peripheral blood cells, particularly red blood cells and platelets lack their own genomic DNA/RNA, the use of nucleic acid-binding sensitizers is especially useful for the problem of treating blood contaminants.

The present invention provides a method for viral/bacterial inactivation of dried (lyophilized or evaporatively dried), frozen, liquid or reconstituted cells (erythrocytes, platelets, hemosomes and other cellular or cell-like components) or blood protein fractions, which allows for the cells or protein fractions to be useful in a transfusable state, while still maintaining relatively high cell viability, ATP synthesis and oxygen transport, in the case of cellular components, and therapeutic efficacy, in the case of protein fractions. The present invention may also be applied to vital inactivation of tissues and organs used for transplantation, and used in topical creams or ointments for treatment of skin disorders or for topical decontamination.

The lyophilization and reconstitution media according to the present invention may be utilized to lyophilize and reconstitute proteins, particularly, blood plasma protein fractions. The protein fraction may be virally/bacterially deactivated by mixing with a chemical sensitizer, lyophilized (freeze-dried) or frozen, then irradiated. If the lyophilization media of the invention is used, it is contemplated that the constituents of the media also serve to provide some degree of protection of the dry proteins during irradiation.

The present invention utilizes a class of compounds based on 3-carboethoxy psoralens, psoralens, angelicins, khellins and coumarins which contain a halogen substituent and a water solubilization moiety, such as, quaternary ammonium ion or phosphonium ion. These materials comprise a relatively low toxicity class of compounds, which can selectively bind to the nucleic acid (single-stranded DNA, double-stranded DNA, or RNA) that comprise the genetic material of viruses. The bound compound can be activated by exposure to radiation, such as ultraviolet radiation (U.V. light of a defined wavelength), or ionizing radiation such as x-rays, after which the activated compound damages the bound viral nucleic acid or viral membranes rendering the virus sterile and non-infectious. Activation of the selectively bound chemical sensitizer focuses the photochemistry and radiation chemistry to the viral nucleic acid or vital membranes and limits exposure to nearby cellular components or plasma proteins.

According to the present invention, a radiation sensitizing chemical compound is added to a liquid suspension of infectious viruses, and the mixture is exposed to U.V. light or ionizing radiation. Assays of viral infectivity demonstrate the effectiveness of the compounds in inactivating the viruses, compared to radiation treatment alone. The present invention is also applicable to frozen or lyophilized and rehydrated blood components, in particular red blood cells, platelets, and plasma protein fractions.

An effective radiation sensitizer must bind specifically to nucleic acids and must not accumulate in significant amounts in lipid bilayers, which are common to viruses, erythrocytes, and platelets. Although there is evidence that psoralens bind to nucleic acids by intercalation, neutral psoralens such as 8-MOP are uncharged and thus also have a high affinity for the interior of lipid bilayers and cell membranes. The present invention thus provides an additional advantage in that the sensitizers do not bind to membranes. ##STR1##

The binding of 8-MOP (8-methoxypsoralen) to cell membranes, shown above, would be acceptable if the psoralen bound to lipid was photochemically inert. However, Midden (W. R. Midden, Psoralen DNA photobiology, Vol II (edi. F. P. Gaspalloco) CRC press, pp. 1. (1988) has presented evidence that psoralens photoreact with unsaturated lipids and photoreact with molecular oxygen to produce active oxygen species such as superoxide and singlet oxygen that cause lethal damage to membranes. Thus, it is believed that 8-MOP is an unacceptable sensitizer because it sensitizes indiscriminate damage to both cells and viruses.

Positively charged psoralens such as AMT (4'-aminomethyl-4,5',8-trimethylpsoralen) will not bind to the interior of phospholipid bilayers (membranes) because of the presence of the charge. However, AMT contains an acidic hydrogen which can bind to the phospholipid head group by hydrogen bonding, shown below. ##STR2##

Thus AMT is believed to be an unacceptable sensitizer because it will indiscriminately sensitize damage to viral membranes and to the membranes of erythrocytes and platelets.

The quaternary ammonium or phosphonium substituted halo-psoralens described herein do not accumulate in the interior of lipid bilayers (membranes) because of the presence of the charge, nor will they bind to the phospholipid head groups of the membrane because they lack acidic hydrogen for hydrogen bonding.

The psoralens (such as 8-MOP and AMT) must often be used in combination with a quencher (e.g. mannitol, dithiothreitol, vitamin E, etc.) to protect, repair or otherwise offset the deleterious effects of the sensitizer and light on cell membranes, and to quench the production of free oxygen radicals in solution that cause indiscriminate damage. The psoralen sensitizers herein do not accumulate in viral membranes and as a consequence do not require the presence of a quencher additive to the blood product. Preferably the psoralen sensitizers herein containing halogen also generate a minimal amount of free radicals in solution, thereby avoiding the need for quenchers.

In solid samples (frozen or lyophilized), a feature of the invention is that radical and radical ions are generated preferably by X-ray activation as bound radicals generated by the sensitizer for advantageous selective viral destruction. The exposure of water and organic molecules to ionizing radiation ordinarily produces undesirable hydroxyl radicals, hydrogen atoms, solvated electrons and organic radicals and radical ions. These are short lived species which react at nearly diffusion controlled rates with proteins and cell membranes. Ultimately these side reactions result in cell death or protein denaturation, but according to the invention these side chemical reactions are minimized in favor of the nucleic acid bound sensitizer radicals that are harnessed for viral inactivation.

When a liquid composition containing red cells, water, and virus is exposed to X-ray radiation, water absorbs essentially all of the incident X-rays due to its presence in large molar excess. X-rays of energy less than about 50 Kev interact with molecules primarily by the photoelectric effect. A water molecule is ionized by an X-ray photon to produce a radical cation and a photoelectron of high energy. The primary photoelectron has enough excess energy to ionize nearby water molecules to produce additional water radical cations and secondary electrons. The secondary photoelectrons have less energy than the primary photoelectrons, but still contain enough energy to ionize nearby water molecules and produce tertiary photoelectrons. The cycle of ionization continues until the energy of the photoelectrons is degraded below the ionization threshold of water. The thermalized photoelectrons are then solvated by water. Water radical cations react with water to form hydroxy radical and hydronium ion (Equation 5). Solvated electrons react with hydronium ions to form hydrogen atoms (Equation 6). This leads to the creation of hydroxyl radicals (see equation above) and solvated electrons randomly dispersed throughout the sample. These reactive intermediates are then free to diffuse throughout the sample and react indiscriminately with cells and virus. This protocol produces non-selective damage to the virus and cells that are suspended in the solution, ##STR3##

There are several strategic advantages which result from exposure of lyophilized or frozen solid samples to X-rays relative to the exposure of hydrated liquid samples to X-rays. As water is present in molar excess in liquid samples it will absorb the incident X-rays. This results in a random and unselective distribution of reactive intermediates in the sample. The water content in lyophilized samples is greatly reduced and thus the X-ray radiation can now be absorbed by a molecule other than water, in particular, a radiation sensitizer, which is in proximity to the virus and efficiently absorbs X-ray radiation, so that the reactive intermediates will be generated in a cluster close to the virus. As there is little diffusion in a lyophilized sample, the reactive intermediates cannot migrate and must undergo reactions in the immediate vicinity in which they are generated. The only reactions possible in the solid state involve the movement of small, light hydrogen atoms and electrons, most likely by quantum mechanical tunneling (QMT). This scheme works equally well in the frozen state. In this case, the frozen water molecules are restricted in their mobility. The most likely pathway for reaction for them involves the recombination of the radicals generated to form the neutral parent molecule. In the frozen system as well, the diffusion of reactive species is limited.

All atoms have a cross section for the absorption of X-rays. In general the larger the atomic number (Z) of the atom, the larger the cross section of the atom to high energy radiation. Secondly, the higher the energy of the X-ray photon, the smaller the cross section of the atom to radiation. Thus, a bromine atom absorbs X-rays more strongly than carbon, nitrogen, or oxygen atoms. X-rays of very high energy (>100 Kev) penetrate samples deeply because they have a small probability of absorption by a particular type of atom. For this reason high energy, deeply penetrating X-rays, are less selectively absorbed by a heavy atom such as bromine relative to a light atom such as carbon, nitrogen, or oxygen. Soft X-rays (<35 Kev) penetrate samples less deeply than hard X-rays (>100 Kev) but are more selectively absorbed by heavy atoms relative to light atoms.

The cross section (Mass Attenuation Coefficient) ratios of platinum to carbon, bromine to carbon, and iodine to carbon as a function of the X-ray source atom are shown in FIG. 1. These values suggest that Mo, Rb, Pd, and Ag X-ray target tubes could produce suitable wavelength radiation to provide a dramatic selectivity in X-ray cross section for bromine or platinum compared to carbon. The selective absorption of X-ray radiation for bromine and platinum peaks with a molybdenum source as opposed to the widely used tungsten source available in most industrial or medical X-ray units.

The preferred source of radiation for viral inactivation is the combination of the soft X-rays produced from a molybdenum X-ray tube with a polybrominated radiation sensitizer. The bromine atom K-shell absorption edge matches the energy of the characteristic X-ray emission spikes of the molybdenum tube thereby allowing selective deposition of the X-ray energy into the sensitizer. When the sensitizer is bound to the virus or in close proximity to it, this will lead to selective viral inactivation.

Exemplary molecules which meet these criteria include brominated psoralen derivatives as described in the appendix I and positively charged porphyrin compounds, appendix II.

Other compounds such as fluorescein derivatives, merocyanine 540, hematoporphyrin, benzoporphyrin, and phthalocyanines may be Utilized for this binding capacity upon derivatization to contain halogens or platinum or both (appendix III). These may be activated by UV, as well visible light or X-rays. This feature will permit their use in the solid state where the side reactions (such as generation of singlet oxygen) which hampers their use in the liquid state can be avoided or suppressed.

A preferred class of sensitizers is selected from the group consisting of compounds of the formula (A): ##STR4## wherein u is an integer from 1 to 6; X is an anionic counterion; Z is N or P; R.sub.1, R.sub.2, R.sub.3, R.sub.4, R.sub.5, and R.sub.6 are independently halo; H; linear or branched alkyl of 1-10 carbon atoms; linear or branched alkoxy of 1-10 carbon atoms; .paren open-st.CH.sub.2 .paren close-st..sub.m O.paren open-st.CH.sub.2 .paren close-st..sub.p Z.sup..sym. R'R"R'" or --O.paren open-st.CH.sub.2 .paren close-st..sub.n Z.sup..sym. R'R"R'" wherein n, m and p are independently integers from 1 to 10 and R', R" and R'" are independently H or linear or branched alkyl of 1 to 10 carbon atoms with the proviso that on each Z atom, not more than one of R', R" or R'" may be H; and at least one of R.sub.1, R.sub.2, R.sub.3, R.sub.4, R.sub.5, or R.sub.6 is .paren open-st.CH.sub.2 .paren close-st..sub.m O.paren open-st.CH.sub.2 .paren close-st..sub.p N.sup..sym. R'R"R'" or --O.paren open-st.CH.sub.2 .paren close-st..sub.n N.sup..sym. R'R"R'". Particularly preferred are compounds wherein R.sub.4 is --O.paren open-st.CH.sub.2 .paren close-st..sub.n N.sup..sym. R'R"R'", especially wherein R', R" and R'" are ethyl and n=3. Preferably, R.sup.6, R.sup.5, R.sup.2 and R.sup.1 are hydrogen and R.sup.3 is H or halo, preferably bromo.

In general, the above compounds may be made by halogenating psoralens and isolating the appropriately substituted isomers. For compounds wherein the ring substituent is a quaternary ammonium alkoxy or phosphonphonium alkoxy group, that group may be made from the corresponding hydroxy-substituted psoralen, as exemplified by the following scheme. ##STR5##

In general, derivatives possessing more than one free hydrogen atom on the preferred ammonium group show viral inactivation but unacceptable levels of cell damage. This is particularly the case for 4'-aminomethyl-4,5',8-trimethylpsoralen (AMT), which is an experimentally utilized psoralen derivative.

Therefore, according to the invention, sensitizing compounds for viral inactivation preferably do not contain substituents which possess free hydrogen groups capable of exhibiting hydrogen bonding to the cell membrane.

In one embodiment, the present invention is applicable to lyophilized cells. The term lyophilization is broadly defined as freezing a substance and then reducing the concentration of the solvent, namely water, by sublimation and desorption, to levels which will no longer support biological or chemical reactions. Usually, the drying step is accomplished in a high vacuum. However, with respect to the storage of cells and particularly erythrocytes, the extent of drying (the amount of residual moisture) is of critical importance in the ability of cells to withstand long-term storage at room temperature. Using the procedure described herein, cells may be lyophilized to a residual water content of less than 10 weight %, preferably less than 3%, and still be reconstituted to transfusable, therapeutically useful cells. Cells with about 3 weight % water content using this procedure may be stored for up to two weeks at room temperature, and at 4.degree. C. for longer than eight months, without decomposition. This far exceeds the current A.A.B.B. standard for refrigerated storage of red blood cells of six weeks at 4.degree. C. or less than one day at room temperature without decomposition. These dried cells may be deactivated using a chemical sensitizer described herein.

According to the embodiment of the present invention for treating dried or lyophilized cells, the washed, packed cells are mixed with a chemical sensitizer, then washed to remove excess sensitizer not bound to viral or bacterial nucleic acid, and the treated cells are then dried or lyophilized. The dry cell and sensitizer mixture will then be irradiated, for example, with gamma radiation, at an intensity of about 3K-50K rads, for a period of time sufficient to destroy viruses (in particular, the single-stranded or double-stranded RNA/DNA viruses), without any substantial adverse effect on the recovery and usefulness of the cells. Other wavelengths of electromagnetic radiation such as visible light or X-rays, may be used.

In another embodiment, the chemical sensitizers may be added to liquid protein preparations, then lyophilized and irradiated. Particularly preferred are blood protein preparations, including but not limited to, plasma proteins, blood protein extracts, clotting factors (such as Factors VIII and IX), immune globulins and serum albumin.

Dry or lyophilized cells or protein fractions may be directly mixed with the chemical sensitizer, then irradiated.

From the foregoing description, it will be realized that the invention can be used to selectively bind a chemical sensitizer to blood-transmitted viruses, bacteria, or parasites. Also monoclonal or polyclonal antibodies directed against specific viral antigens (either coat proteins or envelope proteins) may be covalently coupled with a sensitizer compound.

Since cell compositions also comprise a variety of proteins, the method of decontamination of cells described herein is also applicable to protein fractions, particularly blood plasma protein fractions, including, but not limited to, fractions containing clotting factors (such as Factor VIII and Factor IX), serum albumin and/or immune globulins. The viral and bacterial inactivation may be accomplished by treating a protein fraction with a sensitizer as described herein.

Although described in connection with viruses, it will be understood that the methods of the present invention are generally also useful to inactivate any biological contaminant found in stored blood or blood products, including bacteria and blood-transmitted parasites.

FIG. 1 is a plot of X-ray cross section for various sensitizers described in Example 3.

FIG. 2 is a plot of efficiency of sensitizers with Mo and W targets.

FIG. 3 is a plot of Phi 6 with inactivation according to the procedure of Example 5.

FIG. 4 is a plot viral reduction in plasma according to the procedure of Example 6.

FIG. 5 is a plot of viral reduction in samples described in Example 7.

FIG. 6 is a plot of viral reduction in samples described in Example 8.

FIG. 7 is a plot of residual Factor VIII in X-ray irradiated lyophilized plasma.

FIG. 8 is a plot of viral reduction of samples described in Example 9.

FIG. 9 is a plot of residual Factor VIII activity in UV irradiated plasma.

FIGS. 10, 11 and 12 are plots of viral reduction in samples described in Example 10.

FIGS. 13, 14, 15 and 16 are plots of viral reduction in samples described in Example 11.

FIG. 17 is a plot of lambda virus reduction with psoralen sensitizer (with and without bromine) in hydrated plasma in sensitizer concentrations of 0.05, 0.1 and 0.2 mg/ml. The sensitizers are psoralen (S#9) and Br-psoralen (S#8), activated with UV.

FIGS. 18 through 21 are graphs of platelet characteristics as a function of irradiation time according to the procedure of EXAMPLE 12.

FIGS. 22 through 25 are graphs of platelet characteristics of platelet concentrates seeded with phi 6, irradiated in presence of sensitizers 8B and 8C, according to EXAMPLE 12.

FIG. 26 shows the viral reduction results of the test described in Example 13.

FIG. 27 shows the results of bacterial reduction described in Example 19.

FIG. 28 shows the attenuation coefficients described in Example 21.

FIG. 29 shows results of the test described in Example 22.

The present invention is directed to methods for reducing viral, bacterial and other parasitic contamination in blood, blood components, cell cultures or cell culture components by irradiation in the presence of a chemical sensitizer. Sensitizers are disclosed which are particularly useful to decontaminate liquid compositions, such as blood, blood components, reconstituted lyophilized cells, and the like, using UV irradiation. Other sensitizers are disclosed which are particularly useful to decontaminate solid compositions, such as frozen, dried or lyophilized cells or proteins.

One embodiment involves lyophilized cells (or proteins).

The cells are preferably prepared by immersing a plurality of erythrocytes, platelets and/or hemosomes, etc. in a physiologic buffered aqueous solution containing a carbohydrate, and one or more biologically compatible polymers, preferably having amphipathic properties. By the term amphipathic it is meant there are hydrophobic and hydrophilic portions on a single molecule. This immersion is followed by freezing the solution, and drying the frozen solution to yield novel freeze-dried erythrocytes containing less than 10%, and preferably about 3% or less by weight of moisture, which, when reconstituted, produce a significant percentage of viable, transfusably useful red blood cells, platelets or hemosomes. Preferred methods of reconstitution of the lyophilized composition are described below. Although described in connection with red blood cells, it will be understood that the methods are generally also useful to lyophilize platelets, hemosomes, and blood protein fractions.

The carbohydrate utilized to prepare erythrocyte, platelet and/or hemosome compositions according to the invention is biologically compatible with the erythrocytes, platelets or hemosomes, that is, non-disruptive to the cells or hemosome membrane, and one which permeates, or is capable of permeating, the membrane of the erythrocytes, platelets or hemosomes. It is also advantageous to stabilize proteins, especially labile blood proteins, with the carbohydrates during lyophilization and irradiation according to the invention. The carbohydrate may be selected from the group consisting of monosaccharides, since disaccharides do not appear to permeate the membrane to any significant extent. Monosaccharide pentoses and hexoses are preferred as is a final concentration of from about 7.0 to 37.5 weight % in phosphate buffered saline (PBS) or a phosphate buffered solution, preferably about 26%. Xylose, glucose, ribose, mannose and fructose are employed to particular advantage.

It will be understood that the cells may be lyophilized using other protocols and irradiated as described below. Although vital inactivation will be attained, the advantage of retaining a significant percentage of viable useful red blood cells is lost if the described lyophilization procedure is not followed.

The following description is written in connection with erythrocytes (RBC's) but it will be understood it is also applicable to platelets, hemosomes or other blood cell types or biological cells, as well as protein fractions, particularly plasma protein fractions.

The erythrocytes will preferably be prepared from whole blood centrifugation, removal of the plasma supernatant and resuspending the cells in PBS or a phosphate buffered solution or a commercial dextrose-saline solution. This wash cycle may be repeated 2-3 times preferably using a commercial dextrose-saline solution, then the packed cells are diluted with the lyophilization buffer described above so that the final diluted concentration of carbohydrate and polymer are maintained in the necessary ranges.

Alternatively, commercially available packed blood cells may be used, which typically are prepared in CPDA (commercial solution containing citrate, phosphate, dextrose and adenine).

Preferably, but not necessarily, the cells will have been previously lyophilized using a lyophilization solution buffered in the range of pH of 7.0 to 7.4 preferably by a phosphate-buffered solution. A typical phosphate-buffered lyophilization solution will comprise mono- and di-basic potassium and sodium phosphate (usually in the range of 1-10 mM each) and 5-10 mM adenine. This solution maintains the pH at around 7.2.

A preferred phosphate-buffered solution to be used as the lyophilization buffer will comprise nicotinic acid, reduced glutathione, glutamine, inosine, adenine, monopotassium phosphate, magnesium chloride disodium phosphate all of which will serve as a basic salt buffer at a pH of about 7.2. In addition this lyophilization buffer will contain a final concentration of about 26% weight by volume of a monosaccharide, preferably 1.7M glucose, and a final concentration of about 3.0% weight by volume of polyvinylpyrrolidone (average molecular weight of 360K), and a final concentration of about 15% weight by volume of hydroxyethyl starch (average molecular weight of 500K).

Upon lyophilization to a moisture content of less than 10%, and preferably less than 3%, the lyophilized cells may be maintained under vacuum in vacuum-tight containers, or under nitrogen or other inert gas, at room temperatures for extended periods of time in absence of or without significant degradation of their desirable properties when reconstituted for use as transfusable cells. In using the preferred lyophilization method disclosed herein, a particular advantage of the present invention is that the lyophilized cells may be stored at room temperature for extended periods of time, thus obviating the need for low temperature refrigeration which is required for storing liquid CPDA preserved red blood cells prepared by methods of the prior art. The present invention also obviates the need for very low temperature (-80.degree. C.) frozen storage of red blood cells in glycerol.

By using the preferred reconstitution method disclosed herein it is a further advantage that the lyophilized red blood cells may be reconstituted at normal temperatures, i.e. greater than about 4.degree. C. up to about 37.degree. C., which corresponds to normal human body temperature, and preferably at room temperature (about 22.degree. C.). The reconstitution medium is preferably a solution comprising a polymer or mixture of polymers having a molecular weight of from about 2.5K to 360K, preferably 5K to about 360K, present in a concentration in the range of about 12 to 30% weight by volume. This polymer may be the same polymer utilized to lyophilize the red blood cells as described above. Hence the polymers polyvinylpyrrolidone, hydroxyethyl starch, and dextran are particularly preferred and most preferred is polyvinylpyrrolidone (preferably molecular weight about 10K) present in a concentration of about 19% weight by volume in the reconstitution solution. The reconstitution solution will be buffered again typically by phosphate-buffered solution comprising monopotassium phosphate and disodium phosphate as described hereinabove to maintain a pH within the range of about 7.0 to 7.4. The most particularly preferred polymer is polyvinylpyrrolidone of an average molecular weight of about 10K. The most preferred reconstitution buffer will also contain adenosine triphosphate (ATP) in a final concentration of about 5 mM.

The polymers may be present in the various solutions from a final concentration of about 3.6K weight % up to saturation, and have a molecular weight in the range of from about 2.5K to about 360K. Preferably, the polymers have molecular weights in the range of from about 2.5K to about 500K, most preferably from about 2.5K to 50K, and are present in a concentration of from about 3.6 weight % up to the limit of solubility of the polymer in the solution. Polymers selected from the group consisting of polyvinylpyrrolidone (PVP) and polyvinylpyrrolidone derivatives, and dextran and dextran derivatives provide significant advantages. Most preferred is the use of polyvinylpyrrolidone (an amphipathic polymer) of average molecular weight in the range of 2.5-360K in an amount in the range of 3-20% weight by volume in the solution prior to lyophilization. Amino acid based polymers (i.e., proteins), dextrans or hydroxyethyl starch may also be employed. In the lyophilization buffer hydroxyethyl starch (M-HES) with an average molecular weight of about 500K is employed in a 15% weight by volume final concentration. Other amphipathic polymers may be used, such as poloxamers in any of their various forms. The use of the carbohydrate-polymer solution in the lyophilization of red blood cells allows for the recovery of intact cells, a significant percentage of which contain biologically-active hemoglobin.

The most preferred reconstitution buffer will be a solution comprising monopotassium phosphate, disodium phosphate and ATP, all of which form a basic salt buffer at a pH of about 7.2, which also contains about 19% weight by volume of polyvinylpyrrolidone (average molecular weight about 10K).

The reconstitution solution may also optionally contain a monosaccharide, preferably present in the concentration range of about 7.0 to 37.5% weight by volume. The preferred monosaccharides are xylose, glucose, ribose, mannose and fructose.

In the most preferred embodiment, the lyophilized erythrocytes can be reconstituted by mixing with an equal volume of the reconstitution buffer at a temperature of about 37.degree. C. and mixed. By "equal" it is meant that the volume is the same as the starting volume prior to lyophilization. After initial reconstitution, the solution is preferably diluted 1:1 with 1-4 additional volumes of the reconstitution buffer at a temperature of about 37.degree. C. with added mixing until fully hydrated.

Then, it is preferred that the rehydrated cells be washed according to the following procedure. It is realized, however, that once the cells are reconstituted with reconstitution buffer they are in a hydrated and useful form, but the combination of washings described hereinafter are preferred, specifically for clinical purposes.

After separating the cells from the reconstitution buffer by centrifugation, the resulting packed red blood cells are preferably resuspended at room temperature in (approximately the volume used in the initial reconstitution) a wash buffer comprising nicotinic acid, inosine, adenine, glutamine, and magnesium chloride, all present at about 0.4-10 mM further comprising sodium chloride and potassium chloride each at about 30 mM, buffered by 10 mM disodium phosphate to pH 7.2. This wash buffer further comprises a monosaccharide, preferably glucose at a concentration of about 20 mM, and a polymer, preferably polyvinylpyrrolidone, of a molecular weight 40K and present at a concentration of about 16% weight by volume. Separation by centrifugation completes the first post-rehydration step, a washing step.

After the washing step the rehydrated cells may be suspended in a dextrose-saline transfusion buffer, if transfusion is the intended use of the cells, at room temperature which preferably contains polyvinylpyrrolidone at a 10% weight by volume final concentration, with an average 2.5K molecular weight. The cells can be used as is or be returned to autologous plasma. Additional wash steps in a phosphate-buffered diluent buffer can further remove viruses, but this step is optional for preparation of rehydrated, transfusible cells.

The reconstitution and washings described above will in most instances achieve about 4 log reduction of any viral and bacterial contamination, where 1 log reduction is achieved by drying and 3 log reduction is achieved by washing. Of course, different viruses may respond differently, potentially resulting in more than 4 log reduction of contamination.

The reconstituted cells have characteristics which render them transfusable and useful for therapeutic purposes in that their properties are similar to that of fresh (i.e. not previously lyophilized) red blood cells. Typically reconstituted red blood cells according to the present invention have an oxyhemoglobin content greater than about 90% of that in normal red blood cells. Hemoglobin recovery prior to any washing step is typically in the range of 80 to 85%. The overall cellular hemoglobin recovery including the post-hydration washing steps is about 20 to 30%. The morphology of the reconstituted cells according to the present invention (by scanning electron microscope) typically shows no holes or gaps, and primarily discocytic with some stomatocytic morphology. The oxygen carrying capacity of fresh red blood cells (as measured by P.sub.50, the oxygen partial pressure at which 50% of the oxygen molecules are bound) was measured to be in the range of about 26 to 28 (average 26.7); with an average Hill coefficient (a measure of the cooperative binding of oxygen molecules to native hemoglobin) of 1.95. The typical P.sub.50 for erythrocytes lyophilized and reconstituted according to the present invention is about 27.5 (average) with an average Hill coefficient of 2.08. Assays of ATP in the reconstituted cells indicate ATP levels suggesting normal ATP to ADP metabolism. Normal hemagglutination by readily available blood typing antisera of red blood cells made according to the present invention is also typically found.

This lyophilization and reconstitution procedure advantageously and significantly diminishes viral/bacterial contamination in cell-like material (such as hemosomes), and protein fractions. The contamination can be further reduced by the radiation sensitizing and treatment, particularly while the cells or protein fractions are in the dry state.

The starting packed red blood cells or proteins (which may initially be in a liquid or lyophilized state) are mixed with a sufficient amount (based on total wet weight of cells) of a chemical sensitizer. Preferably, in a composition of packed red blood cells (about 10% hematocrit) about 0.1 to 1 mg of the chemical sensitizer will be used per ml of packed cells. Preferably, the mixture will be irradiated with gamma radiation in the range of 3K-50K rads, typically about 3K rads. Preferred exposure is from 1-10 minutes, if using gamma radiation. Alternatively, UV light (320 nm) may be used, particularly for protein fractions. Preferred exposure is from 1-10 minutes, preferably 3 minutes, if using UV radiation. By this irradiation in presence of a sensitizer, there will be about a 6 log reduction of viral and bacterial contamination, based on contamination present prior to washing and irradiation.

The present invention provides a selective method of generating bound radicals derived from chemical sensitizers only in the vicinity of viral RNA or DNA. Indiscriminate radiolysis of blood containing virus in a hydrated state produces free hydroxyl radical in solution. However, the hydroxyl free radical will damage both the red blood cells and associated proteins as well as the viral target. Thus, viral inactivation would be achieved at the sacrifice of red cell viability. Therefore, sensitizers which bind to DNA, RNA, viral coat proteins, and/or viral membranes and which can be selected to generate bound radicals upon irradiation, are required. Since the radiolysis can be performed in the dry state (preferably less than 10% residual moisture), generation of free hydroxyl radicals from water is greatly reduced. In this manner indiscriminate radical damage is further prevented. Exemplary compounds include: ##STR6##

The preparations of these compounds are known. See Martin, R. F. and Kelly, D. P., Aust. J. Chem., 32, 2637-46 (1979); Firth, W., and Yielding, L. W., J. Org. Chem., 47, 3002 (1982). Other radical-generating reagents which generate radicals upon irradiation are disclosed by Platz et al., Proc. SPIE-Int. Soc. Opt. Eng. 847, 57-60 (1988) and Kanakarajah et al., JACS 110 6536-41 (1988).

The radiation-sensitizing compound (which may also be modified to bear a metal atom substituent) may also be selected from the class consisting of DNA-binding drugs, including, but not limited to, netropsin, BD peptide (a consensus peptide from HMG-1), S2 peptide, and the like. These and other DNA-binding drugs are disclosed in Pjura, P. E., Grzeskowiak, K. and Dickerson, R. E. (1987), J. Mol. Biol. 197, 267-271; and Tengi, M., Usman, N., Frederick, C. A. and Wang, A. H. J. (1988), Nucleic Acids Res. 16, 2671-2690.

The radiation sensitizing compound (which may also bear a metal atom) can also comprise a class of DNA-binding proteins and/or polypeptides and/or peptides. Examples of this class of DNA-binding proteins and/or polypeptides and/or peptides are disclosed in Churchill, M. E. A. and Travers, A. A. (1991) Trends in Biochemical Sciences 16, 92-97. Specific examples of DNA-binding peptides include the SE peptide and BD peptide disclosed in the reference herein.

Another class of sensitizers comprises the positively charged porphorins and pthalocyanates, which bind DNA and RNA. These sensitizers can be activated by irradiation with visible light (500-700 nm).

The DNA-binding specificity can be achieved by covalently coupling the radiation sensitizing compound and/or metal atom to either a DNA-binding drug or to a DNA-binding protein or polypeptide or peptide. ##STR7##

Other sensitizers include specially designed molecules which form triplex DNA, such as those disclosed by Youngquist and Dervan PNAS 82 2565 (1985); van Dyke and Dervan, Science 225 1122 (1984); Van Dyke and Dervan, Nuc. Acids Res. 11 5555 (1983); Barton and Raphael, PNAS 82 6460 (1985); Barton et al., JACS 106 2172 (1984); and Barton, PNAS 81 1961 (1984). These molecules bind to DNA and RNA, site specifically, if desired, and carry reactive moieties which can generate radicals in the proximity of the DNA or RNA.

While not intending to be bound by a theory, it is believed that the ejected electron will be captured by that site with the most favorable electron affinity, which is most likely a second molecule of sensitizer elsewhere in the sample. Electron capture by R--I (or R--Br) leads to dissociation of RX with the formation of a radical. The radical so generated will abstract a C--H hydrogen atom from a sugar moiety of a nearby nucleic acid which in turn will lead to DNA or RNA cleavage and viral inactivation.

The radical cation of the sensitizer (R--X.sup.+.cndot.) will eventually abstract an electron from that component of the sample with the most favorable oxidation potential. This is most likely guanine. The electron transfer reaction forms guanine radical cation. This substance will react with O.sub.2 upon reconstitution with aerated H.sub.2 O. This process also leads to DNA cleavage and viral inactivation. Unreacted material and reaction by-products will be removed during the washing steps involved in the reconstitution of the lyophilized cells. This process will also further remove any virus not inactivated by the treatment described above.

Compounds (1) and (2) bind tightly to DNA and RNA by either intercalation and/or by electrostatic interactions between positively charged ammonium ion groups and the negatively charged phosphate groups of the nucleic acid target. Red blood cells do not contain nucleic acids and accordingly will not bind to such compounds by intercalation.

Dried or lyophilized solid formulations are preferably exposed to 10,000 rads of x-ray or gamma radiation. It is known that the red cells will survive these doses of radiation in the absence of a sensitizer. Lyophilized blood can withstand higher dosage levels of radiation than hydrated blood.

The gamma radiation or x-ray will be absorbed primarily by the heavy atom of the sensitizer, which will be bound to viral DNA or RNA. The radiation will ionize the sensitizer as follows:

(X-Ray)

In some instances, particularly if the sensitizer and red blood cells are allowed to stand together for more than several minutes, sensitizers may diffuse into the red blood cells prior to lyophilization. Antioxidants such as glutathione (an excellent hydrogen atom donor) may be added to the preparation to augment the red cell defenses against free radical initiated damage. It will be understood that incorporation of the sensitizer into cells will also allow inactivation of intracellular viruses, especially viruses thought to reside inside white blood cells (most packed red blood cell units contain residual white cells), or intracellular blood parasites, such as malaria parasite which infects red blood cells.

The sensitizers are removed from the reconstituted blood serum or protein fraction by the washing protocol described above for lyophilized cells.

It is preferred that gamma or X-ray radiolysis take place in a dried lyophilized blood (or protein), virus, and sensitizer formulation rather than in a wet, fluid material for several reasons. Firstly, the dry material is less sensitive to radiation and can be exposed to larger doses of .gamma.-rays or other penetrating radiation without damage to red blood cells (Table 1). This increases the extent of radiolysis of the sensitizer. Secondly, sensitizer radicals bound to DNA or RNA in the dry state can not dissociate from the virus due to the lack of diffusion in the solid material. This will force the sensitizer radical to react with viral RNA or DNA. Thirdly, the solid state conditions will enhance hydrogen atom transfer reactions of the sensitizer radical with the viral nucleic acid, perhaps by quantum mechanical tunneling. Fourthly, the reconstitution and washing protocol used with lyophilized blood or protein fraction serves as a means to remove unreacted material or reaction by-products, and further removes any virus not affected by the treatment (Table 2).

Other types of radiation may be used including ionizing radiation in general, such as X-ray radiation. In one embodiment a metal and/or halogen atom may be a substituent on a chemical radiation sensitizer molecule which binds to nucleic acids, thereby targeting the embodiments such as bacteria, parasites and viruses. Metal and halogen atom substituents of chemical sensitizers for this purpose include Pt, Br, I, Zn, Cl, Ca and F. The X-ray source is preferably a tunable source, so that the radiation may be confined to a narrow wavelength and energy band, if so desired. The tunable feature allows for optimization of energy absorption by the metal atoms, thereby directing the absorbed penetrating radiation energy to the production of radicals by a chemical sensitizer bound to nucleic acid.

A preferred metal to be used as the radiation sensitizer is platinum. Another preferred group comprises the halogens, bromine, iodine, chlorine and fluorine. Based on their increasing ability to interact with impinging X-rays, the order of enhancement of radiation sensitizing is expected to be platinum, which is much greater than bromine, which is much greater than iodine, chlorine and fluorine, all of which are much greater than hydrogen.

Compounds containing these atoms, when exposed to X-rays or other forms of ionizing radiation, are capable of forming a reactive species which can interact with the viral nucleic acid, coat protein or lipid envelope, thus destroying it and rendering it non-infectious. This process may be most effective in a dry state where quenching and side reactions due to the presence of water are avoided.

If the composition suspected of being contaminated is to be treated in liquid form, i.e., as a suspension of cells or protein solution, it is preferred that the sensitizer be UV-activated. The most preferred sensitizers for this purpose are halogenated psoralens.

Psoralens are naturally occurring compounds which have been used therapeutically for millenia in Asia and Africa. The action of psoralens and light has been used to treat vitiligo and psoriasis (PUVA therapy; Psoralen Ultra Violet A) and more recently various forms of lymphoma.

Psoralen will bind to nucleic acid double helices by intercalation between base pairs; adenine, guanine, cytosine and thymine (DNA) or uracil (RNA). Upon absorption of a UVA photon the psoralen excited state will react with a thymine or uracil double bond and covalently attach to both strands of a nucleic acid helix. ##STR8##

The crosslinking reaction is specific for a thymine (DNA) or uracil (RNA) base and will proceed only if the psoralen is intercalated in a site containing thymine or uracil. The initial photoadduct can absorb a second UVA photon and react with a second thymine or uracil on the opposing strand of the double helix to crosslink the two strands of the double helix.

Lethal damage to a cell or virus occurs when a psoralen intercalated into a nucleic acid duplex in sites containing two thymines (or uracils) on opposing strands sequentially absorb 2 UVA photons. This is an inefficient process because two low probability events are required, the localization of the psoralen into sites with two thymines (or uracils) present and its sequential absorption of 2 UVA photons.

The halogenated psoralens according to the present invention are improved and more efficient sensitizers because they require only a single UVA photon for activation the ability of the halogen sensitizer to react with any base pair imposes no limitation for the site of intercalation. As shown in Scheme 1 absorption of a UVA photon by a bromopsoralen in the presence of guanine (or any nucleotide base) leads to electron transfer and the formation of bound radicals and ultimately nucleic acid cleavage and viral or cell death. This cleavage mechanism is more efficient than the conventional crosslinking reaction of non-halogenated psoralens. ##STR9##

The psoralen radical 2 can inflict damage on the nucleic acid double helix to which it is bonded by abstraction of a ribose (RNA) or deoxyribose (DNA) sugar carbon hydrogen bond. This can lead to DNA cleavage by known mechanisms. The guanine radical cation shown as an example is also known to react with molecular oxygen, initiating a series of reactions which cleave DNA. The byproduct of the bound radical photochemistry is debrominated psoralen 4, which can then sensitize DNA cross links as described above.

A disadvantage of PUVA therapy (see U.S. Pat. No. 4,683,889, for example) is that it increases the chances of DNA damage (cross linking of the double helix) to normal cells, if normal DNA-containing cells are present in the treated composition. The angelicins, khellins, and coumarins can not cross link DNA because they lack a second double bond or this bond is located in a position unsuitable for crosslinking. In addition, halogenated angelicins, khellins and coumarins bearing a quaternary ammonium or phosphonium side chain will reduce the side effects of cell damage by reducing interactions with the cell membranes.

The structures of exemplary chemical sensitizers utilized in accordance with the present invention are listed in Appendix I. The synthesis of the parent heterocycles to these sensitizers are well known.

The best mode for using the invention on liquid samples is to add a UV-sensitive sensitizer to potentially contaminated blood suspensions, and to expose to UV radiation.

In another embodiment of the present invention the sensitizers will be utilized in conjunction with solvent detergent systems used to treat blood plasma protein fractions. Such detergents are known to decrease the viral titre of plasma or separated plasma fractions, presumably by dissolution of the viruses. Such detergents include, Tween.RTM., sodium cholate, sodium deoxycholate, Triton.RTM. and common organic solvents such as ether. Reduction of viral titre by use of these solvent detergents is described for example by Horowitz, et al., Transfusion, 25, (6), 516-522 (1985), and 25, (6), 523-527 (1985); and U.S. Pat. No. 4,946,648. The level of reduction by such solvent detergents may vary as reported in literature to a reduction of one log to greater than five logs of viral titre for such viruses as VSV, Sindbis, and Sendai. The present invention may enhance the reduction of viral titre by these solvent detergents when used in conjunction with the sensitizers and exposure to radiation as set forth herein. While not intending to be bound by a theory, it is believed that the solvent detergents act on the viral proteins or lipid membranes to denature or alter them in a manner which makes them more susceptible to the actions of the sensitizers through the changes induced by the detergents.

A particularly preferred class of sensitizers comprise DNA intercalators, such as hydroxyl, amino methyl, or methyl substituted psoralens, which may be added to plasma or plasma fractions followed by UV radiation to reduce the viral contamination therein. The substituted psoralens are described in U.S. Pat. No. 4,727,027 wherein a reduction of about 4 to 7 logs of viral contamination was obtained with extended exposure to ultraviolet radiation. The proposed mechanism of action for these non-halogenated psoralens is to form a photoadduct between the sensitizer and the DNA or RNA of viral origin, which results in loss of infectivity of the virus. According to the present invention, the reduction of viral contamination can be unexpectedly reduced by utilizing brominated psoralens or other halogenated psoralens. For example, it was observed that the bromopsoralens are about 200,000 times more effective in reducing viral activity when compared to use of their non-brominated counterparts. While not intending to be bound by any theory, it is believed that the mechanism of action of the brominated psoralens may be a bound radical generation in the proximity of the DNA or RNA resulting in damage of vital nucleic acids of viruses, by direct cleavage of the nucleic acid strands.

The brominated psoralens are an improvement over the known psoralens and other substituted psoralens when used as sensitizers because only one photon of light is required to activate the brominated sensitizer whereas two photons are required to activate a non-brominated sensitizer. Secondly, a brominated psoralen is effective in virtually every intercalative site, whereas a non-brominated sensitizer is effective only in intercalation sites containing a uracil or thymine on different strands of the DNA or RNA. Thirdly, the brominated psoralens may be activated by X-rays as well as UV light.

The use of the brominated or halogenated psoralens is particularly useful for activation in hydrated systems such as plasma, immune sera, tissue culture media containing animal serum or serum components (such as fetal calf serum), or recombinant products isolated from tissue culture media.

Other types of intercalators may be utilized besides the psoralens and substituted psoralens such as those listed below. These intercalators may be used to target viruses or other blood contaminants, or cancer cells. Thus, halogenated or metal atom-substituted derivatives of the following compounds may be utilized as sensitizers:

dihematoporphyrin esters

hematoporphyrin derivatives

benzoporphyrin derivatives

hydrodibenzoporphyrin dimaleimade

hydrodibenzoporphyrin

dicyano disulfone

tetracarbethoxy hydrodibenzoporphyrin

tetracarbethoxy hydrodibenzoporphyrin dipropionamide

The above compounds in their non-halogenated or non-metal atom substituted forms are disclosed in U.S. Pat. Nos. 4,649,151, 4,866,168, 4,883,790, 5,053,423 and 5,059,619, incorporated by reference herein. When modified with halogen atoms or metal atoms, the above-identified classes of compounds may be sensitized with a range of electromagnetic radiation, including ultraviolet light and x-rays.

The present invention may be applied to treatment of liquid blood in ex vivo irradiation, such as by methods and apparatus described in U.S. Pat. Nos. 4,889,129 and 4,878,891 and 4,613,322.

The above compounds are included in a class named lipophilic dyes which include dyes such as merocyanine 540 and phthalocyanine derivatives. Merocyanine 540 has been disclosed as useful for the treatment of cancer and viral inactivation of blood cells and plasma proteins (Sieber, et al., Photo Chem. and Photo Biology, 46 707-711 (1987)). Phthalocyanine derivatives and other lipophilic dyes are known to bind to the membranes of cancer cells or enveloped viruses. When these compounds are activated with suitable wavelength of electromagnetic radiation, they produce free singlet molecular oxygen (Kalyanaraman, et al., PNAS, 84 2999-3003 (1987)), which damages the membrane resulting in the killing of the cancer cells or in viral inactivation. With the addition of radiation sensitizer atoms (metal atoms or halogens) according to the present invention, use of these compounds in combination with suitable radiation produces bound radicals in the proximity of the DNA/RNA/viral membranes, which then results in destruction of the vital membranes or nucleic acid to inactivate the virus.

Halogenated or metal atom substituted fatty acids also may be utilized according to the present invention as radiation sensitizers. Fatty acids per se have been used in viral inactivation by Horowitz, et al. as disclosed in U.S. Pat. No. 4,841,023. According to the present invention, these fatty acids may be utilized with sensitizer atoms to target viral membranes in plasma protein solutions, for example, and by subsequent activation with suitable radiation the chemical radicals are produced to inactivate the viral membrane.

Treatment of biological compositions with detergents are disclosed in U.S. Pat. Nos. 4,820,805 and 4,764,369. The clinical application of psoralens in conjunction with photodynamic treatment is discussed by Edelson, Scientific American 50-57 (August 1988), and U.S. Pat. Nos. 4,683,889, 4,428,744, 4,321,919, 4,398,906, 4,612,007, 4,464,166, 4,613,322, and 4,838,852, which are incorporated by reference herein.

The following compounds are illustrative of sensitizers which contain or which may be modified to contain metal substituents or halogen substituents in accordance with the present invention: ##STR10##

Other sensitizers include compounds such as doxorubicin and daunomycin, which may be modified with halogens or metal atoms in accordance with the present invention, which are known in chemotherapy to attack dividing cells.

The sensitizers also may be utilized in vivo and delivered in liposomes (artificial cells) or drug-loaded natural cells. After introduction of the liposome or drug-loaded cell, the patient may be treated by radiation to activate the sensitizer.

The present invention is applicable to contaminants which comprise single or double-stranded nucleic acid chains, including kNA and DNA, and viruses, bacteria or other parasites comprising RNA and/or DNA.

To illustrate the invention in lyophilized cells, red blood cells were lyophilized as described above, irradiated, and tested for erythrocyte characteristics measured. The results are shown in Table 1. The same procedure was then used, except that the bacteriophage T4 (in dextrose saline) was mixed with the cells and then washed successively with four different wash buffers. The results are shown in Table 2.

Packed human red blood cells purified from donated whole blood are washed free of the anticoagulant storage solution (commercially available CPDA, containing citrate/phosphate/dextrose/adenine), and suspended in dextrose-saline at a 10% hematocrit. Approximately 10 ml of washed packed red cells is placed in a quartz chamber and exposed to U.V. light, preferably at 320 nm, for 2 minute time intervals, up to a 10 minute total exposure. At each 2 minute interval the suspension is mixed and a small sample of red cells (10 microliters) is removed and diluted into 2 ml of water for spectrophotometric assay of hemoglobin. At each step the temperature of the irradiated red cell suspension is measured, to ensure that the suspension did not overheat. At no point did the suspension exceed 26 degrees C. (normal body temperature is 37 degrees C.). Untreated red cells contain a high proportion of functional oxyhemoglobin (oxyHb), usually in the range of 96% or higher. Oxidation damage can form a semi-stable methemoglobin species (metHb), which can normally be reduced back to oxyhemoglobin by a cellular repair enzyme. Hemichrome represents a more severely damaged form, and can be irreversible. Normal red cells can tolerate a moderate level of methemoglobin. Hemichrome degradation can produce free heme, the iron-porphyrin component of native hemoglobin, which is damaging to cell membranes. Thus it is desirable to minimize hemichrome levels. Each hemoglobin species can be detected at a specific wavelength, using a standard spectrophotometer.

The following data show the sensitivity of the hemoglobin to damage by the increased U.V. exposure. An exposure of 3 minutes was judged to be usable for viral inactivation using a radiation sensitizer, without inflicting excessive damage to red blood cells.

A suspension (0.1 ml) of bacteriophage lambda or bacteriophage phi-X174, of at least 10 EV PFU/ml, is separately added to 4 ml of dextrose-saline containing 1 mg/ml of compounds I or II or III. See Martin et al. and Firth et al., supra. Each suspension of bacteriophage with a radiation sensitizing compound is then exposed to U.V. radiation of the preferred wavelength (320 nm) in a quartz chamber for the preferred time (3 minutes). A control sample of each bacteriophage suspension, containing a sensitizer, is not exposed to U.V. light. Serial dilutions are performed to quantitate the level of infectious titer, and aliquots of the various bacteriophage samples are then mixed with host bacteria and spread on nutrient agar. Following a normal growth period, the plates are assayed for plaques. Other bacteriophage suspensions are separately irradiated as above, but without added sensitizer, to demonstrate the effect of this dose of U.V. alone.

From these data it can be seen that all three tested compounds significantly increase the sensitivity of double-stranded DNA virus (lambda) to U.V. of the preferred exposure. Compound I is also effective against a single-stranded DNA virus, phi-X174. Compound I is most preferred, showing a high (at least 6 log reduction) inactivation efficacy against both single-strand and double-strand DNA viruses.

Referring to FIG. 1, the X-ray cross-section ratios of bromine to carbon, iodine to carbon and platinum to carbon were calculated for their cross-sections in CM.sup.2 /G for various X-ray target tubes. Reference: W. H. McMaster, et al. UCRL-50174 Section II, Section III, Section IV, and CRC Handbook of Chemistry and Physics, pp. E147, 1979. The data suggest that Mo, Rh, Pd and Ag X-ray target tubes will produce suitable wave length radiation and selectivity in X-ray cross-sections for sensitizer atoms over the carbon element.

FIG. 2 shows the calculated mass attenuation coefficients (X-ray cross-section) of molybdenum and tungsten X-ray targets for cross-sections for iodine, bromine and platinum sensitizer compounds. The graph indicates that platinum and bromine sensitizer atoms absorb 100 to 175 times more radiation energy than carbon for low atomic number elements using a molybdenum target. However, the selectivity is reduced with a tungsten target. This suggests that the combination of platinum and bromine sensitizer atoms and a molybdenum target tube will allow activation of these sensitizers in the presence of a large excess of cellular and protein material. The viral inactivation obtained with molybdenum X-ray radiation and sensitizers with different heavy atoms are illustrated in FIG. 3.

Stock solution of .phi.6 phage was added to plasma to obtain the final virus concentrate of 1.2.times.10.sup.7 PFU/ml. Compounds 19 and 20 (formulas given above) and fluorescein were added to the mixture to give the sensitizer concentration of 0.5 mg/ml. After addition of the sensitizer, the solution was mixed on a mechanical shaker for 1 hour at room temperature. The sample was transferred to a plastic Petri dish (35.times.10 mm) and irradiated in a Pantak HP 160 X-ray unit equipped with a Mo target tube operating at 28 ma and 40 kv settings. Approximate radiation dose delivered was 353 k. After irradiation the residual viral titre was measured by the plaque method. A phage containing irradiated sample was mixed with suitable phage host bacteria and 3 ml of melted soft agar. The mixture was poured over hard nutrient agar plate. After one day of incubation the lysed area stood out as plaque against the dense background. The plaques were counted with a colony counter. All samples were treated with X-ray radiation unless otherwise stated. For dry lyophilized state irradiation, the samples were prepared as described above and transferred to 50 ml round bottom flasks, lyophilized on a bench freeze dryer for 16 hours. The dry powder was placed in Petri dishes and X-ray irradiated to 353 k. The residual viral titre was determined in reconstituted plasma by the plaque method as described above. The results (FIG. 3) are consistent with the hypothetical calculated selectivities shown in FIG. 2.

The initial titre of .phi.6 virus 1.2.times.10.sup.7 PFU/ml in compound 20 (structure given above) and 0.5 mg/ml were used in hydrated plasma. For R-17 virus the starting viral titre of 4.2.times.10.sup.8 PFU/ml in compound No. 20 at 0.5 mg/ml were used. The results (FIG. 4) indicate that enhanced viral reduction is obtained with the use of sensitizers in combination with the X-ray radiation treatment.

FIG. 5 shows the inactivation of enveloped and non-enveloped viruses obtained with molybdenum X-ray radiation treatment in lyophilized plasma preparations. The residual moisture content of Sample 1 was about 7.7% and of Sample 2 was about 2.4% as determined by a Karl/Fisher titrater. The starting concentrations of the viruses in the samples were .phi.6 (2.0.times.10.sup.7 PFU/ml and R-17, 4.4.times.10.sup.8 PFU/ml). Sample 1 and Sample 2 were lyophilized in a Petri dish and round bottom flask respectively. All samples were treated with 353 kr dose and the final virus concentration was determined by the plaque method. The results indicate that the amount of the residual moisture in lyophilized samples significantly influences the degree of inactivation obtained with X-ray radiation treatment in the presence of and absence of a sensitizer.

Samples have starting .phi.6 virus titre of 8.1.times.10.sup.8 PFU/ml and compound No. 20 (structure given above) at 0.5 mg/ml were used in lyophilized plasma. The radiation dose employed was 420 kr. For the R-17 virus the conditions were: initial titre 3.7.times.10.sup.8 PFU/ml and compound No. 7 (structure given above) at 0.2 mg/ml, 353 kr radiation dose. Samples not irradiated with X-ray radiation showed very small or no change in starting viral titre value. The results are shown in FIG. 6 where there is shown a 7 log reduction in viral titre of .phi.6 and a 4.6 log reduction in viral titre of R-17 using a sensitizer. FIG. 7 shows the residual factor H activity X-ray irradiated lyophilized plasma (AHF concentrate) from the same samples. The damage factor to Factor VIII activity in lyophilized plasma during viral inactivation treatment is negligible. Greater than 95% recovery of Factor VIII recovery is shown using the sensitizer.

A stock solution of bacterial phage .lambda. is added to plasma to obtain final titre of 1.2.times.10.sup.8 PFU/ml and compound No. 3 (structure given above) at 0.1 mg/ml concentration. The mixture was transferred to Pyrex glass photolysis cells and exposed to UV (300-360 nm blue lamp) for 5 minutes. After UV irradiation treatment the final viral titre and Factor VIII activity were measured as described earlier. FIG. 8 shows viral reductions obtained. About 3-6 logs of increased viral inactivation was obtained with protein coated viruses using a sensitizer. The recovery of Factor VIII in these samples after inactivation treatments is shown in FIG. 9. Only 10% of Factor VIII activity was recovered after viral inactivation treatment. However, addition of antioxidants, such as vitamin E, increase Factor VIII recovery under UV irradiation.

The starting titre of .phi.6 virus was 2.5.times.10.sup.8 PFU/ml in these samples. The plasma mixture was mixed with either 1% Tween 80 detergent or 1% tri-(n-butyl)phosphate (TNBP) for 2 hours at room temperature. The samples were treated with 353 kr radiation dose in hydrated and lyophilized states as described above. The results are shown in FIG. 10 (detergent) and FIG. 11 (organic solvent). The data indicates that the viral inactivation obtained with molybdenum X-ray radiation in hydrated and lyophilized plasma is substantially enhanced by addition of either organic solvent (TNBP) or detergent (Tween 80) individually or in combination. FIG. 12 shows the results using a stock solution of R-17 virus added to plasma to give a starting titre of 6.2.times.10.sup.8 PFU/ml and 1% Tween 80 with 1% TNBP for two hours at room temperature. The hydrated and lyophilized samples were treated with 353 k radiation dose and the final viral titre was determined as described above.

Fresh platelets (24 hours old) were spun to remove residual red cells. The platelets were diluted with plasma to obtain approximately 500.times.10.sup.6 cells ml. The starting titre of .phi.6 virus in liquid platelet preparation was 2.0.times.10.sup.5 PFU/ml. After addition of compound 20, 0.3 mg/ml of the sample was transferred to a polyolefin bag (2 inches by 2 inches dimensions) and irradiated to deliver approximately a radiation dose of 195.4 kr. Subsequently the sample was analyzed for morphological evaluation under microscope, cell recovery calculations were made, and aggregation response to collagen (200 .mu.g/ml) was determined by an aggregometer. The final virus titre was determined by the plaque method. The liquid state results were shown in FIGS. 13 and 14. FIG. 13 shows the X-ray radiation effect on inactivatin of viruses in platelets (liquid state). FIG. 14 shows the effect of sensitizer and radiation on platelet viability indices, i.e., morphology score, cell recovery and aggregation response. For the frozen samples, fresh platelets 500.times.10.sup.6 cells/ml were diluted with 10% DMSO in PBS buffer (1:1, v/v). The starting .phi.6 virus titre was 2.0.times.10.sup.6 PFU/ml. The compound No. 20 was added at a concentration of 0.3 mg/ml and the sample (2 ml) was transferred to a polyolefin bag and frozen. The frozen bag was treated with 26.3 kr radiation dose. The sample bag temperature was maintained at -80.degree. C. using liquid nitrogen during the irradiation treatment. After irradiation the sample was thawed at 37.degree. C. and diluted with PBS slowly to 10 ml. The washed platelets were resuspended in plasma and analyzed for morphological evaluation, cell recovery and aggregation response to collagen (200 mg/ml). The final virus titre after irradiation was determined by thawed samples by the plaque method. Results were shown in FIGS. 15 and 16. FIG. 15 shows the X-ray radiation effect on inactivation of viruses in frozen platelets. FIG. 16 shows the effect of sensitizer and radiation on frozen platelet viability indices. These results suggest that the addition of sensitizer increases the viral inactivation by 1 to 2 logs. The irradiation damage to platelets is not significant (70% to 90% control of viability assays) as determined by the morphological score, cell recovery and aggregation response to collagen agonist.

Human platelet concentrates are mixed with phi 6 bacteriophage, with or without sensitizer, then exposed to U.V. light. Cell quality assays after treatment include morphology score conducted by visual inspection using a light microscope (using a Kunicki scoring system), hypotonic shock response (HSR), aggregation response induced by addition of ADP measured in an aggregometer, and cell number count.

1. Soft spin platelet unit at 1000 rpm for 5 min to remove RBC.

2. Dilute concentrated platelets with spun autologous plasma to obtain a conc. of 500.times.10.sup.6 cells/ml.

3. Add 5.times.10.sup.6 of .phi.6 to the platelet suspension and mix gently.

4. Add sensitizer to the platelet suspension and incubate for 15 min while protecting from light. Control samples have PBS added.

5. Pipet 2 ml aliquots into small plastic petri dishes and irradiate with UV while rotating samples at 100 rpm on a shaker. Irradiate samples without cover.

6. Transfer irradiated samples to 15 ml plastic centrifuge tubes.

7. Determine morphology, HSR, ADP, Viral inactivation, cell count.

Sensitizer conc.: 0.050 mg/ml or 50 mg/ml

Sensitizer stock: 2.00 mg/ml in PBS

Irradiation times: 5', 7.5', 10', 12.5' (FIGS. 18-25)

Controls: 0', 12.5' no Sensitizer 0' Sensitizer 8, 8-QA.

FIGS. 18-25 are graphs of the experimental data on platelet quality as a function of U.V irradiation time. The data in FIGS. 22-25 were for platelet numbers 144, 145 and 146 (platelet rich plasma, n=3), with sensitizers S#8-B and #8-C (0.100 mg/ml).

To test the viral inactivation efficiency of brominated sensitizers, platelet preparations (2 mL, PRP) were spiked with .phi.6 phage and brominated psoralen (8B) and non-brominated psoralen(8C) were added to obtain final sensitizer concentration of 0.1 mg/mL. The samples were exposed to UVA light in a home made reactor consisting of 4 UVA lamps for the indicated time periods. The petri dishes containing samples were mixed on an orbital shaker during UVA exposure. The UVA light intensity at the sample position was measured at 3.3 mW/cm.sup.2. The residual viral titer after each time point was measured by plaque method The viral reduction values are illustrated in FIG. 26. These results clearly indicate the enhancement in the level of viral inactivation observed with brominated compound when compared with non-brominated derivative under identical conditions. These results demonstrate that as much as a 400 fold improvement in the level of viral inactivation observed when the psoralen molecule is brominated.

The vital inactivation values and in vitro properties of platelets subjected to viral inactivation treatment using various derivatives of psoralen are summarized in Table 1. The psoralen derivative with acidic hydrogens (AMT) produced substantial reduction in hypotonic shock response (HSR) and ADP aggregation response relative to the control sample. These results indicate that the presence of acidic hydrogens on the side chain nitrogen is detrimental to blood cells because of the hydrogen bonding with cell membrane phospholipid head groups, whereas quaternary ammonium derivatives produced significantly lower damage to platelets due to a charge screening effect

Single unit platelet concentrates were obtained 24 hours after collection from certified blood banks. The platelet were subjected to soft spin to remove residual RBC and about 6 logs of .phi.6 phage and sensitizer 8B were added in PBS solution to obtain 0.03 mg/mL final sensitizer concentration. The samples were transferred to Stericon.TM. polyolefin bag (which transmits 60% of UVA light) and irradiated from top (3.6 mW/cm.sup.2) and bottom (3.4 mW/cm.sup.2) in a home made reactor for 10 min. The bags were continuously mixed on extended glass platform attached to orbital shaker. The viral inactivation values and in vitro properties of platelet concentrates were monitored for 2 days after the treatment. The results are summarized in Table 2. The data indicates that >5.3 logs of viral reduction was obtained with excellent retention of platelet in vitro properties.

Units of packed red cells were transferred to Stericon.TM. red cell freezing bags and the hematocrits were adjusted to 60% with PBS. The samples were spiked with 5.times.10.sup.6 PFU/mL .phi.6 phage and 0.3 mg/mL sensitizer 8B. The contents of the bag was exposed to UVA light from top (3.6 mW/cm.sup.2) and bottom (3.4 mW/cm.sup.2) for 6 hours (151.2 J/cm.sup.2). The samples were mixed continuously during UVA exposure on an orbital shaker. The final viral titer in the sample was determined by the plaque method. In vitro biochemical and biophysical properties of treated red cells were monitored for 7 day following the treatment. The results are summarized in Table 3 below. The data indicates that >6.6 logs (complete inactivation of added virus) of viral reduction was obtained without any significant alterations in in vitro biochemical and biophysical properties.

The viral inactivation kinetics in red cells is dependent on the intensity of the UVA light source. The data in Table 4 below indicate that the time required to achieve 6 logs of viral reduction varies with the intensity of the UVA light source used. The time required to achieve 6 logs viral reduction in red cells is substantially reduced to 90 min. using a very high intensity UVA light source (40 mW/cm.sup.2 (top) and 40 mW/cm.sup.2 from bottom).

Viral inactivation in FBS was carried out in 100 mL serum samples in 225 cm.sup.2 tissue culture flask in the presence of sensitizer 8B (0.03 mg/mL). The sample was exposed to UVA light from top (3.6 mW/cm.sup.2) and bottom (3.4 mW/cm.sup.2) while continuously mixing on an orbital shaker for 30 min. (8.4 J/cm.sup.2). Human plasma was irradiated in a Pyrex glass photolysis cell in the presence of sensitizer 8 (0.1 mg/mL) for 5 min. Inactivation of various bacteriophages were carried out using brominated psoralens in combination with UVA light. The results in Table 5 below clearly indicate that 5-7 logs of viral reduction was obtained with various phage (enveloped and non-enveloped) in human plasma fractions and fetal bovine serum products.

FBS (5 mL) sample was spiked with E. coli bacteria and sensitizer 8 (0.03 mg/mL) and irradiated for 15 min. (3.3 mW/cm.sup.2). The residual bacterial titer was determined by colony forming assay using agar plates. The results (FIG. 27) indicate that brominated psoralens effectively inactivate bacteria as well as viral contaminants.

The standard Ames mutagensity test using salmonella mutants was performed on brominated sensitizer 8B at 0.6 mg(mL. The results are summarized in Table 6 below. The data indicates that the compound at 0.6 mg/mL concentration is non-mutagenic.

The X-ray cross-section ratios of bromine to carbon, iodine to carbon and platinum to carbon were calculated for their cross-sections in for various X-ray target tubes (ref.: W. H. McMaster, et al., UCRL-50174 section 11, Section III, Section W, and CRC Handbook of Chemistry and Physics, pp. E147, 1979. The calculated values (FIG. 28) suggest that Mo, Rh, Pd and Ag X-ray target tubes will produce suitable wave length radiation and selectivity in X-ray cross-sections for sensitizer atoms over carbon.

Stock solution of .phi.6 phage was added to plasma to obtain the final virus concentration of 1.2.times.10.sup.6 PFU/mL. Compounds RB (rose bengal), EY (eosin Y) FL (fluorescein) were added to the mixture to give the sensitizer concentration of 0.5 mg/mL. After addition of the sensitizer, the solution was mixed on a mechanical shaker for 1 hour at room temperature. The sample was transferred to a plastic petri dish (35.times.10 mm) and irradiated in a Pantak HP 160 X-ray unit equipped with a Mo target tube operating at 28 MA and 40 KV settings. Approximate radiation dose delivered was 353 Kr. Meter irradiation the residual viral titer was measured (FIG. 29) by the standard double agar plaque method.

For lyophilized state irradiation, the samples were prepared as described above and transferred to 50 mL round bottom flasks, lyophilized on a bench freeze dryer for 16 hours. The dry powder was placed in petri dishes and X-ray irradiated to 353 Kr. The residual viral titer was determined in reconstituted plasma by the plaque method. The results in FIG. 4 clearly demonstrate that sensitizer compound containing bromine substitutients (EY) produced significantly higher viral reduction in combination with a Mo X-ray source. Higher vital reduction was obtained in the lyophilized state where the interference of water in selective activation of brominated sensitizer was reduced to minimal.

The samples spiked with .phi.6 phage had starting titer of 8.1.times.10.sup.8 PFU/mL and compound 20 at 0.5 mg/mL was subjected to 420 Kr radiation treatment in lyophilized state. For R17 phage samples the initial titer, 3.7.times.10.sup.8 PFU, compound 8B.sub.5 at 0.2 mg/mL and 353 Kr radiation dose were employed. The samples not irradiated with X-ray radiation showed very small or no change in starting viral titer value. Table 7 below shows about 3.6 log viral reduction with .phi.6 phage and 4.6 log reduction in viral titer of R-17 using the pentabromo psoralen derivative. The damage to factor VIII activity in viral inactivation treated samples is negligible. Greater than 95% recovery of the factor VIII activity was retained after the treatment as measured by APTT times.

Packed red cells (2.5 mL) were spiked with .phi.6 phage to obtain final titer of 5.times.10.sup.6 PFU/mL and suspended in 7.5 mL of lyophilization solution containing radiation sensitizer (eosin, 0.6 mg/mL). The samples were frozen (-25.degree. C.) or lyophilized in T-75 tissue culture flasks. Except liquid samples frozen and lyophilized samples were x-ray irradiated while maintaining the sample temperature at -40.degree. C. using dry ice/ethanol bath. Liquid samples were irradiated in the presence of radiation sensitizer (eosin, 0.6 mg/mL) in saline solution. The residual viral titer in each samples was measured after processing the sample with appropriate reconstitution buffers. The viral reduction values and in vitro properties of treated red cells are summarized in Table 8 below. Respectable levels of viral reductions were obtained in frozen and lyophilized state without significant alterations in red cell properties. The liquid samples produced lower viral reductions and abnormal red cell properties were observed due to the treatment.

8-methoxypsoralen (8-MOP, Aldrich, Milwaukee, Wis., 0.75 g, 3.5 m mol) was dissolved in tetrahydrofuran (THF, 75 mL), Bromine (1.5 mL, 3.5 mmol) in THF (5 mL) was added dropwise. The resulting mixture was stirred at room temperature for overnight. 10% Sodium thiosulfate (5 mL) was added and the solution was basified using conc. aqueous NH.sub.4 OH, and then extracted twice with chloroform. The organic layer was recrystallized from chloroform to give the product 5-bromo-8-methoxypsoralen (0.75 g, 73.5% yield).

5-bromo-8-methoxypsoralen (0.75 g, 2.4 mmol) was dissolved in dry methylene chloride (50 mL) and added to a three-neck round bottom flask fitted with a reflux condenser under nitrogen. Boron tribromide in hexane (5.1 mL, 5 mmol) was then added in one portion to the flask via syringe. The resulting reaction mixture was stirred at room temperature for 7 hours after which some yellow solid precipitated. Water (125 mL) was added slowly and cautiously. The resulting heterogeneous mixture was stirred at room temperature for overnight. The crude product was collected by vacuum filtration and air dried. The crude product was recrystallized from acetonitrile (45 mL) to give the product 5-bromo-8-hydroxypsoralen (0.40 g, 58% yield).

5-bromo-8-hydroxypsoralen (0.35 g, 1.22 mmol) was dissolved in anhydrous acetone (100 mL). 1,3-dibromopropane (1.5 mL) and anhydrous K.sub.2 CO.sub.3 (3.0 g) were added. The resulting mixture was refluxed for 48 hours. After cooling, the acetone solution was filtered and the solid residue was washed twice with acetone solution was filtered and the solid residue was washed twice with acetone (950 mL). The combined acetone solutions were concentrated to yield an oil which was dissolved in benzene (150 mL). This was chromatographed through a silica plate column containing water (5%). The product eluted with benzene and was concentrated to give the product 5-bromo-8-(.gamma.-bromopropyloxy)psoralen (0.3 g, 60% yield).

5-bromo-8-(7-bromopropyloxy)psoralen (0.3 g, 0.73 mmol) was dissolved in anhydrous ethanol (10 mL). Triethylamine ((was added and heated at 60.degree. C. for 5 hours. After cooling, the brown solution was poured into water (100 mL) and extracted twice with chloroform (200 mL). The organic layer was dried with MgSO.sub.4 and concentrated to give an oil. The hydrochloride salt of the product was precipitated by adding anhydrous ethanol (2 mL), conc. aqueous HCl (0.5 mL) and ether (2 mL). The resulting mixture was collected in the freezer overnight and the precipitate was vacuum filtered to give a white solid which was recrystallized from anhydrous ethanol:hexane (1:5, v/v) to give the product 5-bromo-8(.gamma.-triethylaminopropyloxy)psoralen hydrochloride (0.15 g, 48% yield). Melting Point 196.degree.-198.degree. C. The product was characterized by .sup.1 H NMR, .sup.13 C NMR and FAB mass spectrometry.

The mixture of aluminum chloride (13.4 g, 100 mmol) and 40 mL of 1,2-dichloroethane was purged with nitrogen, cooled with ice-salt bath, and stirred as the solution of 4.32 g (20.0 mmol) of 8-methoxypsoralen in 60 mL of 1,2-dichloroethane was added dropwise over 30 min. Then followed was the addition of bromine (9.6 g, 60 mmol) over 1 h. After stirred at 0.degree. C. for 2 h, the reaction was then allowed to stand still overnight at room temperature. The reaction mixture was poured into the mixture of ice (200 g) and concentrated hydrochloric acid (30 mL0, and stirred over 3 h. The solid was collected by vacuum filtration, washed with acetone, and then transferred into 500 mL of acetone, which consequently was refluxed over 2 h. After cooled, the gray solid was collected, weighted 8.74 g (100%), mp 305.degree.-10.degree. C., .sup.1 HNMR(DMSO-d.sub.6, ppm): 11.40 (Br, 1H), 8.12 (d, J=9.2 Hz, 1H), 6.55 (d, J=9.2 Hz, 1H). .sup.13 CNMR: the spectrum was taken due to the lack of suitable solvent, HRMS mass cald for C.sub.11 H.sub.3.sup.79 Br.sub.3 O.sub.4 436.1441, found 435.7561, major fragments m/e (relative intensity %): 442(0.93, M.sup.+ +6), 440(2.98, M.sup.+ +4), 438(3.14, M.sup.+ +2), 436(0.95, M.sup.+), 396(20.76), 394(27.69), 317(11.70), 315(45.52), 313(34.56), 271(23.45), 269(35.28), 82(98.56), 80(100), 79(41.99).

To the mixture of 4.39 g (10.0 mmol) of 8-Hydroxyl-2,3,9-tribromopsoralen and 10.1 g (50 mmol) of 1,3-dibromopropane in 8 omL of DMSO, was added 7.0 g of K.sub.2 CO.sub.3. After stirred at 60.degree. C. for 2 days, the reaction mixture then was poured into 250 mL of ethyl acetate, which was followed by 3 times of washing with 150 mL of water, dried over Na.sub.2 SO.sub.4 and concentrated by rotatory evaporation to about 50 mL. About the same amount of petroleum ether was added to the residue to get precipitate, which was collected by filtration to give out 3.70 g (66%) of yellow solid. mp 116.degree.-8.degree. C. .sup.1 HNMR(DMSO-d.sub.6, ppm): 8.13 (d.d., J.sub.1 =10 Hz, J.sub.2 =2 Hz, 1H), 6.59 (d, J=10 Hz, 1H), 4.43 (t., J=5.9 Hz, 2H), 3.74 (t., J=6.5 Hz, 2H), 2.27 (m. 2H). .sup.13 CNMR: HRMS mass cald for C.sub.14 H.sub.8.sup.79 Br.sub.4 O.sub.4 556.2197, found 555.7178, major fragments m/e (relative intensity %): 564 (0.43, M.sup.+ +8), 562 (1.56, M.sup.+ +6), 560 (3.02, M.sup.+ +4), 558 (1.34, M.sup.+ +2), 556 (0.36, M+), 518 (14.90), 516 (23.75), 514 (18.28), 474 (11.00), 472 (34.13), 470 (36.94), 468 (13.86), 396 (44.12), 394 (60.20), 352 (47.04), 349 (100), 348 (60.91), 271 (33.81), 269 (54.70). ##STR12##

3-[8-(2,3,9-Tribromopsoralen)oxy]propyltriethyl ammonium bromide 2.0 mL of triethylamine was added to the mixture of 0.561 g (1.00 mmol) of 8-(3-Bromopropan)oxyl-2,3,9-tribromopsoralen and 8 mL of dimethyl sulfone. After stirred at 50.degree.-60.degree. C. for 2 days, the reaction mixture was poured into about 30 mL of water, the produced precipitate was then collected by filtration, recrystallized from ethanol and ethyl acetate to give 0.435 g (66%) yellow solid. mp 224.degree.-6.degree. C. (dec.). .sup.1 HNMR(DMSO.sub.d-6, ppm): 8.21 (d d, J.sub.1 =10.0 Hz, J.sub.2 =2.0 Hz, 1H), 6.64 (d, J=10.0 Hz, 1H), 4.46 (t, J=4.8 Hz, 2H), 3.47-3.26 (m, 8H), 2.15 (m, 2H), 1.23 (t, J=7.0 Hz, 2H). ##STR13##

3-[8-(2,3,9-Tribromopsoralen)oxy]propyldimethyl-2-dimethylamminoethyl ammonium bromide 2.0 mL of N,N,N',N'-tetramethylethylenediamine was added to the mixture of 1.12 g (2.00 mmol) of 8-(3-Bromopropan)oxyl-2,3,9-tribromopsoralen and 20 mL of dimethyl sulfone. After stirred at 50.degree.-60.degree. C. for 2 days, the reaction mixture was poured into about 100 mL of water, the produced precipitate was then collected by filtration, recrystallized from ethanol and ethyl acetate to give 0.700 g yellow solid. mp 189.degree.-92.degree. C. (dec.) .sup.1 HNMR(CDCl.sub.3, ppm): 8.18, 8.12 (dd, J=10.0 Hz, 1H), 6.45 (d, J=10.0 Hz, 1H), 4.49 (t, J=6.2 Hz, 2H), 4.12 (t, J=8.4 Hz, 2H), 3.88 (t, J=5.7 Hz, 2H), 3.55 (s, 6H), 2.89 (t, J=5.4 Hz, 2H), 2.53-2.41 (M, J.sub.1 =6.3 Hz, 2H), .sup.13 CNMR(CDCl.sub.3, ppm): 159.7, 147.1, 146.9, 144.1, 142.5, 131.5, 127.8, 115.7, 115.6, 107.4, 105.6, 72.7, 72.6, 71.3, 70.6, 70.3, 61.6, 49.1, 46.9, 42.7, 27.7, 11.6. HRMS mass cald for C.sub.24 H.sub.33.sup.35 ClNO.sub.7 556. 2197, found 555.7178, major fragments m/e (relative intensity %): 564(0.43, m.sup.+ +8), 562(1.56, M.sup.+ +6), 560(3.02, M.sup.+ +4), 558(1.34, M.sup.+ +2), 556(0.36, M+), 518(14.90), 516(23.75), 514(18.28), 474(11.00), 472(34.13), 470(36.94), 468(13.86), 396(44.12), 394(60.20), 352(47.04), 349(100), 348(60.91), 271(33.81), 269(54.70). ##STR14##

3-[8-(2,3,9-Tribromopsoralen)oxy]propyl-(2-hydroxyl)ethyldimethylammonium bromide. The mixture of 0.261 mg (0.500 mmol) of 8-(3-Bromopropan)oxyl-2,3,9-tribromopsoralen and 0.5 mL of the required amine in 5 mL of isopropanol was refluxed for 24 hrs. After cooled, the mixture was poured in 10 mL of water. The yellow solid was collected by filtration and recrystallized from ethanol-ethyl acetate. Yield 0.221 g (67%). mp 225.degree.-8.degree. C. (dec.) .sup.1 HNMR(DMSO.sub.d-6, ppm): 8.27 (d d, J.sub.1 =10.0 Hz, J.sub.2 =2.0 Hz, 1H), 6.67 (d, J.sub.1 =10.0 Hz, J.sub. 2=1.3 Hz, 1H), 4.44 (t, J=5.4 Hz, 2H), 3.62-3.35 (m, 6H), 3.10 (s, 6H), 2.28-2.19 (m, 2H), 1.92-1.83 (m, 2H). ##STR15##

(3-Bromo)propyltriethylphosphonium bromide The solution of 0.591 mL (0.473 g, 4.00 mmol) of triethlphosphine in 5 mL of acetonitrile was added to the refluxing solution of 4.04 g (20.0 mmol) of 1,3-dibromopropane in 15 mL of acetronitrile over 1 hr. under argon atmosphere. After the addition completed, the reaction mixture was refluxed for another 6 hrs. After cooled, acetonitrile was evaporated, 15 mL of ethyl acetate was then introduced into the residue. After stored in refrigerator for a couple of hours, the highly hygroscopic white solid was collected by filtration. Yield 1.09 g (85%). .sup.1 HNMR(DMSO.sub.d-6, ppm): 3.64 (t, J=6.7 Hz, 2H), 2.36-2.17 (m, 8H) 2.08-2.02 (m, 2H), 1.13 (d t, J.sub.1 =18.2 Hz, J.sub.2 =7.7 Hz, 1H).

3-[8-(2,3,9-Tribromopsoralen)oxy]propyltriethylphosphonium bromide The mixture of 0.439 g (1.00 mmol) of 8-Hydroxyl-2,3,9-tribromopsoralen, 0.351 g (1.10 mmol) of (3-Bromo)propyltriethylphosphonium bromide, and 0.4 g of potassium carbonate in 3 omL of acetone was refluxed under argon for 5 days. After acetone was evaporated, 30 mL of water was poured into the residue, and the water solution was then saturated with potassium bromide. The dark yellow crude product was collected by filtration and recrystallized with ethanol-ethyl acetate. Yield 0.439 g (65%). mp 225.degree.-8.degree. C. (dec.). .sup.1 HNMR(CDCl.sub.3, ppm): 8.20 (d d, J.sub.1 =10.0 Hz, J.sub.2 =2.6 Hz, 1H), 6.46 (d, J=10.0 Hz, 1H), 4.52 (t, J=4.9 Hz, 2H), 3.04-2.90 (m, 2H), 2.60 (d q, J.sub.1 =13.1 Hz, J.sub.2 =7.7 Hz, 6H), 2.33-2.27 (m, 2H), 1.36 (d t, J.sub.1 =18.1 Hz, J.sub.2 =7.7 Hz, 9H).