NO890097L - LIPOSOM / DOCSORUBICIN COMPOSITION AND PROCEDURES. - Google Patents

Description

The present invention relates to a new liposome/doxorubin composition and method, and in particular, a composition that can be stored for a longer time in freeze-dried form, and reconstituted into an injectable suspension with selected liposome sizes and between 85-95 percent liposome-encased doxorubicin.

2. References.

Aubel-Sadron, G., et al, Biochemie, 66:333 (1984).

Forssen, E.A. , et al, Proe Nat Acad Sei, USA, 78(3) 1873

(1981).

Gabizon, A., et al, Cancer Res, 42:4734 (1982).

Gabizon, A., et al, Cancer Res, 42:4734 (1983). Goormaghtigh, E., et al, Biochem Biophys Acta, 779:271 (1984).

Gutteridge, J.M.C. Biochem Pharm, 33(11):1725 (1984a). Juliano, R.L., et al, Biochem Pharmacol, 27:21 (1978). Poznansky, M.L., et al, Pharm Revs 36(4):277 (1984). Sunamoto, J., et al, Biochem Biophys Acta, 833:144 (1985). Szoka, F., et al, Proe Nat Acad Sci (USA) 75:4194 (1978). Szoka, F., et al, Ann Rev Biophys Bioeng, 9:467 (1980). Young, R.C., et al, N Eng J Med, 305:139 (1981).

Doxorubicin (DXR) is a potent chemotherapeutic agent effective against a broad spectrum of neoplasms (Aubel-Sadron et al and Young). Clinical use of the drug in free form is, however, limited by serious side effects and acute toxicity including malaise, nausea, emetic myelosuppression and severe hair loss. In addition, cumulative and irreversible heart damage occurs with repeated administration, which thus considerably limits the use of the drug in long-term treatment (Young).

One approach that has been used to combat DXR toxicity in vivo is to administer the drug in liposome-encapsulated form. Animal studies and, more recently, clinical testing have demonstrated that DXR in liposome-bound form retains therapeutic efficacy against animal tumors, but is considerably less toxic, due to reduced mortality and reduction in cardiotoxic effects (Forssen, Gabizon 1985). The drug-protective effect of liposomes is caused, at least in part, by a marked change in tissue disposition and drug release rate of the injected drug (Gabizon 1982; Gabizon 1983; Juliano).

Recent animal model studies point to three factors that are important in achieving increased therapeutic effect and reduced toxicity in DXR/liposomes. One of these is liposome size. Studies aimed at determining the biodistribution and drug clearance of DXR/liposomes after intravenous administration as a function of liposome size have been performed (Gabizon 1982, 1983 and U.S. Patent Application for "Liposome/Anthraquinone Drug Composition and Method, Serial No. 806,084, filed December 6, 1985).A brief summary of the results shows that DXR/liposomes with average sizes between approximately 0.1-0.2 have increased drug levels in the liver and spleen, and decreased drug levels in the heart, lung, intestine, kidney and skeletal muscle compared to the free drug.

Liposomes having this size are thus particularly advantageous for the treatment of liver and spleen-located tumors, and for reducing the toxicity related to drug levels in non-target tissues, especially the heart. Liposomes in this size range also exhibit slower drug clearance in liver and spleen tissue. The biodistribution and drug clearance rates in small unilamellar vesicles (in the approximately 0.03-0.08 size range) were in the middle between the values of the larger liposomes and free drug.

Due to many reasons, the optimal upper size limit of the liposomes is about 0.5 µm and, preferably, about 0.2-0.3 µm. First, desired target tissue areas such as liver sinusoids and parenchyma, spleen and bone marrow are more receptive to liposomes smaller than about 0.2-0.3 µm. Second, liposomes in the 0.2-0.3 pm size range can easily be rendered sterile by filtration through a depth filter. Vesicles of this size also have little tendency to aggregate upon storage, thus reducing a potentially serious problem when the composition is administered parenterally. Liposomes that have a size in the sub-micro range provide more uniform biodistribution and drug clearance characteristics, due to their more uniform sizes.

Another feature of DXR/liposomes that is important for therapeutic efficacy is the degree of chemical degradation of lipid and drug components that may occur during storage. Phospholipid degradation can take place as hydrolytic release of fatty acyl chain groups and lipid peroxidative damage, especially at unsaturated binding sites in the lipid acyl chain components (Gutterage, 1984; Sunamoto). In addition, anthaquinone-type drugs, such as DXR, can themselves initiate oxidative reactions (Goormaghtigh, 1984; Gutteridge, 1984a), and in the presence of lipids also participate in free radical/oxidative reactions, and also undergo rapid chemical changes on storage in liposomes, as reported in the above cited patent application for "Liposome/Anthraquinone Drug Composition and Method".

In ongoing studies supported by the above-cited application, it was found that the combined presence of a lipid-soluble free radical quencher, such as alpha-tocopheral, and a water-soluble trihydrosam-gelating agent, such as desferal, reduced the the peroxidative damage to the lipid and drug components substantially more than the sum of the individual protection provided by any one protective agent alone. A key feature of this protective effect appears to involve gelation of ferric iron in a form that does not catalyze peroxidation formation as readily, combined with the lipid-phase free-radical quenching of the lipophilic quencher.

The toxicity of the DXR/liposome formulations is also sensitive to the percentage of free drug in the formulation. Clinical trial calculations indicate that DXR/liposomes containing moderate levels of free drug (approximately 35%) produce substantially higher toxicity at a dose of 50 mg/m<2> than DXR/liposomes with low free drug content (10-15$) administered in a dose of 70 mg/m<2>. A number of different methods exist for the removal of free DXR from DXR/liposome preparations, including centrifugation, diafiltration, filtration by molecular sieve chromatography, and, as described in US Patent No. 4,460,577, by filtration through an ion-exchange resin. These methods are effective in reducing the free drug levels in freshly prepared DXR/liposome preparations by 10% or less of total DXR. When stored in solution, on the other hand, free drug can gradually be released when increased lipid and drug-oxidative damage occur. Conventional freeze-drying methods and reconstitution methods for storing DXR/liposomes in a dried state lead to a significant degree of release of free drug upon reconstitution. Experiments conducted pursuant to the present invention indicate that conventional freeze-drying procedures and reconstitution procedures result in substantial release of liposome-associated DXR, typically resulting in 20-30% free drug in the reconstituted liposome suspension.

It is therefore a general purpose of the present invention to provide a DXR/liposome formulation which has selected liposome size characteristics, and which is stable during long-term storage, and which contains at least approximately 85%-95% DXR in liposome-associated form.

Another purpose of the invention is to provide a freeze-dried liposome composition and liposome concentrate used to prepare such a formulation.

Yet another purpose of the invention is to provide a method for the production of DXR/liposomes which, after storage for a long time, have predetermined sizes, relatively little free DXR, and minimal oxidative damage.

The invention includes in one aspect a lyophilized doxorubicin/liposome composition which, upon reconstitution with a predetermined volume of aqueous medium, provides a liposome concentrate characterized by: a. a concentration of liposomes of greater than about 100 mM

liposome lipid,

b. llposome sizes which are mainly in a selected size range of between approximately 0.1-0.5 pm,

c. liposome-encapsulated DXR, at a concentration between about 2-10 mole percent liposome lipids, and between about 85-95%

total DXR, and

(d) between about 1$ to 10$ cryoprotectant.

In a preferred embodiment, the composition is reconstituted with low-osmolarity medium or distilled water to provide a concentrate close to physiological osmolarity. A preferred cryoprotectant is trehalose or lactose, at a concentration of about 5%.

In another aspect, the invention includes the DXR/liposome concentrate formed by reconstitution of the above composition. The concentrate preferably has a physiological osmolarity, and may be further formulated to contain a lipophilic free-radical quencher and a water-soluble iron-specific gelling agent, to minimize free radical and oxidative damage that may occur during the initial liposome preparation and storage and after reconstitution. The invention further includes a method for producing DXR/liposomes for intravenous administration in cancer chemotherapy, and with between 85-95$ DXR in liposome-associated form. The method first includes production of a liposome dispersion characterized by: a. liposome size mainly in a selected size range

between about 0.1 to 0.5 pm,

b. liposome-encapsulated DXR, in a concentration of between about 2-10 mole percent liposome lipids, and more than

about 90$ liposome-associated DXR; and

c. between about 1$ - 1$ cryoprotectant.

The dispersion, which is preferably prepared to a lipid concentration greater than 100 mM at physiological osmolarity, is lyophilized for storage, then reconstituted by the addition of an aqueous medium, and preferably distilled water, to a reconstituted concentrate having a lipid concentration of at least about 100 mM. The percentage of free drug in the concentrate is usually about 5$ higher than that in the original dispersion. The reconstituted DXR/liposomes are diluted with an injection medium, such as physiological saline, to form a DXR/liposome suspension suitable for IV injection. The percentage of free DXR in the diluted suspension is essentially the same as that in the reconstituted substance.

These and other purposes and features of the invention will be more obvious after the following detailed description of the invention.

1. DXR/1iposome preparation

The DXR/liposome concentrate and suspension in the invention is prepared from a concentrated, pre-freeze-dried liposome dispersion and the preparation is described in this section. Although the suspension can be prepared according to a number of procedures, two methods will be highlighted. These are (1) thin-film hydration, with subsequent size separation and treatment to remove free drug and to concentrate the suspension; and (2) solvent injection, according to a new method that produces high concentration dispersions.

A. Liposome dispersion components

DXR/liposomes in the dispersion are formed from standard vesicle-forming lipids, which usually include neutral phospholipids, such as phosphatidylcholine (PC); negatively charged phospholipids, such as phosphatidylglycerol (PG), phosphatidylserine (PS), phosphatidylinositol (PI) and phosphatidic acid (PA); negatively charged sterols, such as cholesterol sulfate, and cholesterol hemisuccinate; and sterols such as cholesterol. Selection of lipids is made by the following considerations (a) drug-release rate in serum, (b) drug-encapsulating efficiency, (c) liposome toxicity, and (d) biodistribution and targeting properties. The effect of phospholipid and cholesterol components on in vivo drug release rates, drug entrapment efficiency and liposome toxicity has been investigated in studies conducted under the above patent applications for "Liposome/Anthraquinone Drug Composition and Method". A brief summary of the results of these studies: 1. Negatively charged phospholipids, such as PG, PS and PI, enhance drug liposome stability, according to measurements of DXR release into 50$ plasma. Optimal concentrations are between about 10 to 30 mol-$ negatively charged lipid; 2. Cholesterol at a concentration between about 20 to 50 mol-$ produces a 2-3 fold enhancement in DXR retention in PC liposomes, according to measurements of DXR release into serum, but has less stabilizing effect when negatively charged phospholipids also is present; 3. Negatively charged phospholipids appear to increase drug-encapsulating efficiency, despite cholesterol having little effect; 4.

Liposomes mainly composed of phospholipids with saturated acyl chain components are more toxic than liposomes composed of corresponding unsaturated acyl chain components, despite the fact that saturated lipids are less susceptible to oxidative damage and a corresponding increase in toxicity upon storage.

A preferred liposome composition includes between about 30-70 mol% PC, 20-50 mol% cholesterol, and 10-15 mol% negatively charged phospholipid or sterol.

The interaction between lipid and DXR components in liposomes, and the effects of protective agents that inhibit free radical and lipid peroxidation reactions in DXR/liposomes have also been extensively studied, as reported in the above-cited application. The studies showed that DXR accelerates oxidative reactions in liposomes and is itself degraded by free-radical peroxidation reactions. According to an important finding in the studies, the presence of both a lipid-soluble free-radical quencher, such as alpha-tocopherol, and a water-soluble ferric-iron gelling agent, such as desferal, produced a significantly higher inhibition of oxidative damage to lipids and drug components than would be expected from the inhibition arising from each component alone. These results suggest that inhibition of free-radical and/or peroxidative events in both the aqueous and lipophilic phases of the dispersion is essential to achieve effective inhibition of oxidative damage in DXR/liposomes.

The lipophilic free radical scavenger used in the composition of the invention is preferably alpha-tocopherol, or a pharmacologically acceptable analogue or ester thereof, such as alpha-tocopherol succinate. Other suitable free radical scavengers include butylated hydroxytoluene (BHT), propyl gallate, and their pharmacological salts and analogs. The free radical quencher is usually included in the lipid components that are used to produce the liposomes, according to conventional methods. Preferred concentrations of the protective compound are between about 0.2 and 2 mol% of the total lipid components making up the liposomes.

The water-soluble protective agent is an iron-specific gelling agent selected from the class consisting of neutral and synthetic trihydroxanoic acids and characterized by a very high binding constant for ferric iron (of the order of 10^<0>) Dg and a relatively low binding constant for 2-valent cations, such as calcium and magnesium. A number of trihydroxanoic acids of natural origin have been described, including compounds of the ferrichrome class, such as ferrichrome, ferrichrome A and albomycin; compounds of the ferrioxamine class, including ferrioxamines and ferrimycins; and compounds in the fusaramine class. The structure and iron coordination properties of these compounds have been reviewed.

A preferred gelling agent is ferrioxamine B, also known as ferrioxamine, deferoxamine, desferrioxamine B and desferal. This compound exhibits exceptional iron binding affinity and has been shown to be safe for parenteral use in humans in the treatment of iron storage disease and iron poisoning.

B. Thin-film hydration method

As indicated above, the DXR/liposome dispersion can be prepared using a number of known methods for the preparation of liposomes (see, for example, Szoka 1980). A preferred method that leads to relatively high initial DXR encapsulation in liposomes (up to about 95$) is thin-film hydration illustrated in Example 1. In this method, vesicle-forming lipids are taken up in a suitable organic solvent or solvent system and dried physiological osmolarity (about 300 mO insoluble iron chelating agent can also 1 t preferred medium used in example 1 ictosis, 0.45$ NaCl, 50 mM desferal, pH 5-6Drosent lipid) DXR. The aqueous medium b Limen, and the lipids are hydrated without shaking) or slowly (without shaking) condition liposomes and produce a defined size range with optimal pharmacokinetic properties. A preferred method for achieving the desired size distribution of llposome sizes is by extrusion of the liposomes through a small pore polycarbonate membrane size at selected pore sizes, such as 0.1, 0.2 or 0.4 pm, approximately corresponding to the size distribution of liposomes after one or more passes through the membrane. Typically, the liposomes are extruded through the membrane several times until the size distribution stabilizes. This method of size separation is described in general by Szoka et al, 1978.

Alternatively, the liposomes can be extruded through a ceramic filter with pore sizes between about 0.2 and 1.2 µm, as described in co-owned patent application for "Liposome Extrusion Method", serial no. 829,710, filed February 28, 1986.

After liposome size separation, the dispersion is further processed to remove free DXR, i.e. DXR not immediately associated with liposomes. The purpose of removing the drugs is to reduce the amount of free drug to a level of less than about 10$ of total DXR in the dispersion, and preferably between about 3$ and 7$. The suspension can be pelleted by high-speed centrifugation after dilution, whereby free drug and very small liposomes remain in the supernatant. An advantage of this method is that it combines the removal of free drug with a concentration step which is also useful in the preparation of the final dispersion. Another method uses gel filtration by molecular sieve chromatography to separate liposomes from free DXR molecules. Alternatively, the dispersion may be placed in contact with a dialysis membrane or diafiltration/ultrafiltration membrane with a pore size sized to selectively retain liposomes but exclude free drug. The free-drug side of the membrane may contain ion-exchange or hydrophobic groups that actively bind free DXR, to promote separation of free drug from the liposomes. The diafiltration/ultrafiltration approach has the same advantage as ultracentrifugation in that the dispersion can be concentrated at the same time.

Another approach, described in U.S. Pat. patent no. 4,460,577, involves filtering a drug/liposome composition through an ion exchange or hydrophobic resin layer to remove free drug, such as DXR. Alternatively, the substance can be passed through a solid support material which, with surface-bound ion yields or hydrophobic groups, can selectively bind DXR when the dispersion is passed through the material. Such materials are available from Romicon (Woburn, MA).

In a preferred method for removing free drug, the dispersion is passed through a diafiltration apparatus designed to actively remove DXR which is passed through the diafiltration membrane. This can be done by placing an ion-exchange or hydrophobic resin substance on the filtrate side of the diafiltration membrane, or by attaching ion-exchange groups or hydrophobic groups to the support structure on this side of the membrane. The material is circulated through the membrane space until a final concentration greater than about 100 mM lipid is achieved.

In the final preparation step, the dispersion is preferably concentrated to a final lipid concentration of at least about 100 mM. The purpose of concentrating the dispersion is that after freeze-drying and storage, the freeze-dried substance is reconstituted originally to a concentration which is also at least approximately 100 mM. At the same time, the reconstituted concentrate should also have a final osmolarity close to physiological osmolarity, and in particular this condition includes intra- and extra-liposomal aqueous phase. This condition can best be achieved by (a) preparing the dispersion at a physiological osmolarity, and at a final concentration greater than about 100 mM, lyophilizing the concentrated dispersion, and reconstituting the dispersion to about the same volume with distilled water.

As indicated above, several methods for removing free drug, such as centrifugation, diafiltration/ultrafiltration, and diafiltration/ultrafiltration in the presence of drug-binding resins, act to concentrate the dispersion, and the methods can be easily performed to achieve the desired lipid concentration of at least about 100 mM. When other free-drug removal steps are used, the dispersion is suitably concentrated, e.g. by ultrafiltration or centrifugation.

The final DXR/dispersion has the following characteristics:

a. llposome sizes mainly in a selected size range

between about 0.1 to 0.5 pm, and

b. liposome-encapsulated DXR, at a concentration between about 2-10 mole percent liposome lipids, and at least about 90%

liposome-associated DXR; and

c. between about 1 and 10% cryoprotectant.

The dispersion also preferably has an lipid concentration of at least 100 mM, and close to physiological osmolarity.

The dispersion may be rendered sterile by filtration through a conventional 0.22 or 0.45 µm depth filter, and lyophilized for storage.

C. Solvent injection procedure

The solvent injection method used here is based on a novel liposome preparation method described in co-owned U.S. Pat. patent applications for "High-Encapsulation Liposome Processing Method, Serial No. 908,765; and "High-Concentration Liposome Processing Method; serial No. 909,122, both filed September 18, 1986. These methods are based on the fact that, with selected solvent injection conditions, liposomes with final concentrations up to 300-500 pmol/ml, and drug encapsulation efficiencies of up to 60-70$ for water-soluble compounds and up to 90$ or more for lipophilic compounds be produced in a simple batch process expensive. The system can further be readily adapted for size separation and/or removal of free drug, providing a final product with a desired concentration, liposome size distribution, and percentage of free drug.

In preparing the DXR/liposome dispersion by solvent injection, a solution of liposome lipids in a selected solvent system is injected, at a specific rate, into a mixing chamber containing an aqueous medium with DXR, cryoprotectant, salt, buffer and iron chelating agent. The lipid solubilizer preferably has a boiling point between about 0°-10°C, and the aqueous medium is preferably maintained between about 10e<->25<s>C. The chamber is maintained at a constant temperature and pressure during the solvent injection. The aqueous and lipid components are agitated during the infusion procedure by means of a motorized blade.

The liposome dispersion which is formed in the mixing space can be size-distributed by circulation through a size-distribution device designed to size-distribute the liposomes by extrusion. The device may contain a polycarbonate membrane or ceramic filter for extrusion and size distribution. The circulation shunt containing the extrusion device may also include a dialysis unit or ultrafiltration unit for removing free DXR from the liposomal dispersion. The device is designed to provide flow through of the liposomal dispersion, with diffusion of free DXR through a size-selective dialysis membrane or diafiltration membrane and encapsulation of ion-exchange groups or hydrophobic groups on the outer side of the membrane.

In operation, the lipid/solvent solution is injected into the mixing chamber, with extensive agitation, until the desired lipid concentration of greater than about 100 mM is achieved. At this stage, the dispersion is recirculated through a size distribution shunt/drug removal shunt until the desired size range and free drug concentration is achieved.

The liposome preparation can be carried out under sterile conditions, or the dispersion can be filter sterilized as above after the solvent/injection process. The dispersion is then freeze-dried for storage.

II. Llposomes reconstitution

To prepare the lyophilized DXR/liposome composition for injection, in diluted form, the composition is first reconstituted to form a liposome concentrate. The concentrate is then diluted with an injection medium, such as physiological saline, to form a diluted DXR/liposome suspension suitable for parenteral injection. The reconstitution steps/dilution steps used to prepare the lyophilized composition for injection are important for minimizing the amount of free drug in the diluted suspension, as indicated below, and form an important part of the present invention. This section examines the behavior of the freeze-dried DXR/liposomes in the reconstitution step.

A. Size character trait

In order to control liposome sizes in the reconstituted DXR/liposome substance, it is necessary to minimize liposome size changes during freeze-drying and reconstitution. Experiments conducted according to the present invention indicate that size-distributed liposomes, in the absence of cryoprotectants, increase substantially in size in a freeze-drying/reconstitution cycle.

This can be seen from the study reported in Example 3, which examines size changes in a liposome concentrate containing either 125 mM trehalose (approximately 5%), 5% lactose, or 5% mannitol. Initial llposome sizes, produced by extrusion through a 0.2 µm polycarbonate filter, had an average size of between approximately 280 to 335 nm, as Table 1 shows. Each of the three dispersions was slowly frozen in controlled freezers, or quickly frozen in liquid nitrogen, then freeze-dried to complete dryness. The preparations were then reconstituted with distilled water to the original dispersion volume, and analyzed for liposome size distribution. Average sizes are shown in the bottom rows of Table 1. This shows that both trehalose and lactose, but not mannitol, were effective in preventing liposome size changes after freeze-drying/reconstitution. The increase in llposome sizes observed in the mannitol formulation could be compared to that occurring in the absence of a cryoprotective sugar.

B. Lipid concentration

According to an important aspect of the invention, it has been discovered that the rate of release of liposome-bound DXR from freeze-dried DXR/liposomes can be greatly reduced by initial reconstitution to a relatively high lipid concentration, at least about 100 pmol lipid/ml, then further dilution with an isotonic injection medium. The effect of the lipid concentration upon DXR drug release on the reconstitution can be seen from the study reported in Example 4. Briefly, four DXR/llposome dispersions were prepared as above, and concentrated to increasing lipid concentrations between 40 and 211 mM, initially examined for content of percentage free DXR. This was carried out, as described in example 4, by molecular sieve chromatography in a salt solution which is iso-osmotic with the liposome formulation. As shown in Table 2, the formulations had between 11$ and 13$ free DXR.

After lyophilization, the dried formulation was reconstituted with either 250 µl (one-half of original volume) or 500 µl (original volume) of distilled water, and the reconstituted formulations were again analyzed for percent free DXR, and the results are shown in the bottom two rows of Table 2. As shown, dilution of the lyophilized material to 40 or 54 mM (third row, first two columns) produced a 7-8% increase in the amount of free DXR, while at higher concentrations, (80 to 422 mM), an increase of 4$ in free DXR was mostly observed.

Similar results were provided in the second study reported in Example 4. Here, DXR/liposome dispersions with lipid concentrations shown in Table 2, and free drug concentrations prior to lyophilization of approximately 5$, were lyophilized and reconstituted with distilled water to an original volume. The column to the right of the table shows the measured percentages of free drug after reconstitution. As shown at concentrations of reconstitution lipid below about 100 mM, the final free drug concentrations increased about 8-9%. At higher concentrations, the increase was approximately 3-4$.

After reconstitution to form an iposome concentrate, further dilution with isotonic medium has little or no effect on the percentage of free drug in the suspension. For example, a reconstituted liposome concentrate with a lipid concentration of approximately 130 mM and a measured free drug concentration of approximately 11% can be diluted 1:30 or 1:60 with isotonic medium without any significant change in percent free DXR, even after a 4 -week storage period at 4°C. It is also noted here that all of the percentages of free or bound drug reported above were determined by molecular sieve chromatography, which itself produces multiple dilutions of the DXR/liposome concentrate or suspension being analyzed.

However, further dilution of the concentrate may lead to the release of DXR from the liposomes if the dilution medium is hypotonic. This occurs even at a high lipid concentration, as demonstrated in the study in Example 5. Here, a DXR/liposome dispersion containing 138 mM lipid in 5% lactose and 0.45% saline was freeze-dried in 0.5 ml aliquots, then reconstituted with increasing amounts of distilled water, from 50 to 500 pl. After reconstitution and after standing at room temperature for 1 hour, the concentrate was diluted with distilled water to approximately the original concentration (138 mM). The first sample shown in Table 4 was thus, for example, reconstituted with 50 µl (10X original concentration), and then further diluted with 450 µl distilled water to approximately 1X concentrate, while the final sample shown in the table was reconstituted directly with 500 µl distilled water to IX concentration endpoint. Table 4 provides percent DXR associated with the resulting DXR/liposomes, as measured by molecular sieve chromatography, for each of two samples that were treated identically. As shown, dilution conditions that provide greater hypertonic liposome swelling also provide greater loss of liposome-bound DXR.

Due to reconstitution of DXR/liposomes under conditions that cause liposome swelling, the amount of free drug released increases considerably, and it is also of interest to determine whether hypotonic swelling upon reconstitution increased the release of free DXR. Here, DXR/liposomes were prepared in an isotonic dispersion (approximately 300 mOs), lyophilized and reconstituted to a final lipid concentration of approximately 150 mM with either distilled water or 0.9% NaCl. In the case of reconstitution with distilled water, the bulk of the phase medium was isotonic with the liposome inner regions, whereas in the case of reconstitution with isotonic saline, the main outer phase has approximately twice the osmolarity of the liposome inner regions, which causes some degree of liposome shrinkage. Percent of free drug in the pre-lyophilized dispersion was approximately 4$. In both reconstitution media, the percentage of free DXR after reconstitution is between approximately 6-7$. The results show that at least mild shrinkage during reconstitution can be tolerated without increased release of free DXR in the reconstitution medium.

Three main features of the above data stand out. First, reconstitution of freeze-dried liposomes necessarily causes some loss of liposome-bound DXR. (It is shown here that liposome concentrates show no increase in free DXR when stored at 4°C for several months, indicating that the DXR leakage observed upon freeze-drying is related to the freeze-drying process). This shows that reconstitution of the liposomes involves a certain membrane change in which DXR can be released from the liposomes. This loss can be minimized, according to one aspect of the invention, by reconstituting the liposomes into a relatively concentrated solution. A possible explanation for this phenomenon is that during the short period when the drug is released from the membrane, it is quickly distributed between the lipid and the main aqueous phases in the reconstitution medium in proportion to the relative concentrations of lipid and the main aqueous phases in the medlum. A higher lipid concentration would thereby promote greater distribution in the liposomes during reconstitution. Alternatively, the degree of membrane change and drug release during the reconstitution may be less at higher liposome concentration.

Second, a DXR/liposome concentrate can be diluted with an isotonic dilution medium without the desired release of free drug, even after extended storage periods. This behavior indicates that DXR binding to liposomes is not controlled by a dynamic equilibrium exchange, which would be concentration dependent. It appears that DXR release from (and perhaps binding to) liposomes only occurs during periods of membrane perturbation, such as during reconstitution or under conditions of osmotic swelling. This feature allows the liposome concentrate, which itself has a minimal free drug content, to be diluted for injection without a significant increase in the percentage of free drug.

Third, liposome swelling causes a considerable increase in the release of free DXR from reconstituted membranes, despite the fact that slight shrinkage upon reconstitution appears to have no effect on the release of free drug. This feature causes certain limitations on the type of dilution media suitable for use in the method according to the invention. Typically, the final dilution medium is physiological saline, in which the liposome concentrate should not be higher than approximately 2X physiological saline, to avoid significant liposome swelling. Hypotonic shrinkage is also preferably avoided by reconstitution of the lyophilized liposomes to a final osmolarity of no less than about 0.5 physiological osmolarity (about 300 mOs). In a preferred embodiment of the invention, the liposome dispersion is prepared at physiological osmolarity, reconstituted to the original concentrate volume with distilled water, and diluted with physiological saline. The final concentration of the diluted injectable suspension is preferably between about 2 to 20 mM lipid.

C. Temperature conditions.

The effect of storage temperature and material temperature on reconstitution of free drug release has also been investigated, and the results are shown in Example 6. Briefly explained, lyophilized DXR/liposomes were stored at either -20°C or 4°C for several days, and the samples were reconstituted to the original volume with distilled water at room temperature either at the storage temperature, or after equilibration of the lyophilized substance to room temperature. In all cases, the percentage of liposome-bound DXR was between approximately 90$ to 92$, as shown in the data in Table 5. The storage temperature or the temperature of the freeze-dried material thus has little or no effect on the degree of release of free drug from the liposomes upon reconstitution.

Ill. DXR/ Llposome administration

This section examines the preparation and use of the DXR/liposome formulation according to the invention in a clinical setting, including preparation and storage considerations, and use in tumor treatment.

A. Preparation and storage

The DXR/liposome formulation is supplied as a lyophilized composition. As explained in Section II, the composition, when reconstituted with a predetermined amount of aqueous medium, is characterized by: a. a concentration of liposomes greater than about 100 mM

liposome lipid,

b. llposome sizes mainly in a select size range

between about 0.1 to 0.5 pm,

c. liposome-encapsulated DXR, at a concentration between about 2-10 mole percent liposome lipids and between about 85-95%

liposome-associated DXR; and

d. between about 1$ and 10$ cryoprotectant.

The composition was preferably reconstituted with distilled water to an original volume of liposome dispersion.

An advantage of the freeze-dried composition is that it can be stored for a longer time in a cooler or at room temperature without significant lipid oxidative degradation or hydrolytic degradation. At the same time, the drug can be reconstituted into a concentrate with a desired selected liposome size distribution in the 0.1 to 0.5 pm size range, and with only a small loss of DXR into the main phase medium.

For appropriate preparation and also to minimize loss of drug, it is desirable to add the lyophilized substance in a multi-dose quantity which, when reconstituted, provides multiple aliquots for dilution and individual-dose drug administration. As an example, the lyophilized substance may be supplied in an amount containing 1 g (about 1.84 mmol) of DXR in a total of about 20 mmol of lipid. The lyophilized material is then reconstituted with approximately 100 mL of distilled water to a final concentration of approximately 200 mM lipid. The liposome concentrate can be stored at 4°C for several weeks without significant degradation, if it is protected with alpha-tocopherol and desferol, as explained.

Typically, an individual dose of between about 25-100 mg in a total IV drip volume of about 100-150 mL is optimal for chemotherapeutic administration. The individual diluted dose can easily be prepared from the concentrate above, by simple dilution with sterile saline. For example, to administer a 50 mg dose, a 5 ml aliquot of the concentrate is withdrawn and added to approximately 120 ml of sterile saline, and this suspension is then conventionally dispensed by IV drip or the like. Alternatively, for syringe injection, the concentrate may be diluted in a smaller volume, preferably 25 to 50 ml, for administration. The concentrate can thus be used to prepare between approximately 10-40 individual doses.

Alternatively, the lyophilized substance may be supplied in individual dosage form, such as in a multicompartment package containing, in individual sealed compartments, (a) the desired amount of lyophilized DXR/liposome composition, (b) an amount of reconstitution medium sufficient to forming the DXR/liposome contrast, when mixed with the lyophilized composition and (c) the dilution medium. The concentrate is individually formed by mixing the (a) and (b) compartments, and the final dilution medium by allowing mixing between the concentrate and compartment (c).

B. Therapeutic Applications in the Treatment of Tumors Previous studies described in the above-cited patent application for "Liposome/Anthaquinone Drug Composition and Method" have shown that small, drug-carrying liposomes can concentrate liposome-associated drug in tissue regions, such as liver sinusoids and parenchyma, spleen, and bone marrow accessible to relatively small liposomes. These findings are of enormous importance for the treatment of metastatic diseases of the liver, primary hepatoma, lymphoid proliferative diseases, and leukemias, all of which are major cancers, both in terms of number and geographic distribution.

Clinical trials using DXR/liposomes to treat metastatic tumors in the liver have been conducted. The treatment results provided to date indicate a significant improvement in overall patient comfort with liposome-encapsulated DXR, caused by substantially complete elimination of gastrointestinal toxicity (nausea, vomiting, diarrhea, and stomatitis), and no pain at the injection site. The results represent a significant improvement in the treatment of certain forces with DXR. In addition, there is a considerable reduction in alopecia in the treated patient, and the therapeutic efficacy appears to be as good or better than free DXR alone, according to changes in tumor size. The clinical trials with the DXR/liposome suspensions containing different percentages of free DXR indicate that a considerably greater reduction in side effects occurs when the percentage of free DXR in the suspension is reduced.

From what has been explained above, it becomes clear how different purposes and features according to the invention are countered. The freeze-dried composition according to the invention provides an appropriate and stable form of DXR/liposomes that can be stored for a long time (up to several months or more) in a cooler or at room temperature without degradation of either the lipid or the drug components.

The lyophilized composition is readily reconstituted into a concentrate with a desired liposome size distribution and at least about 85$, and typically 90-95$ liposome-associated DXR. The concentrate can be diluted for parenteral administration without significant change in liposome size or percentage of liposome-bound drug. The diluted medium has favorable biodistribution properties, caused by the size distribution of liposomes, and low toxicity, caused by both the integrity of the liposome and the DXR components, and the low percentage of free DXR.

The following examples illustrate the method of preparation and properties of the DXR/liposome concentrates and suspensions which are formed according to the invention. The examples shall in no way limit the scope of the invention.

Materials

Doxorubicin HC1 (DXR) was obtained from Laboratorie Roger Bellon, SA; EPC and EPG, from Avanti Polar Lipids, Inc., (Birmingham, AL); cholesterol, alpha-tocopherol succinate, lactose, and mannitol, from Sigma Chemical (St. Louis, MO); desferal, from CIBA (Summit, NJ); and trehalose dihydrate, from Pfanstiehl Laboratories, Inc., (Waukegan, IL).

Example 1

Production of freeze-dried DXR/ liposomes

A chloroform solution of lipids containing 518.6 mg EPC, 210.4 mg EPG, 140.0 mg cholesterol and 9.8 mg alpha-tocopherol succinate was prepared. The solvent was removed by rotary evaporation under reduced pressure followed by subjecting the sample to a vacuum of less than 50 µM Hg for 2 h to eliminate residual solvent. 20 ml of an aqueous solution containing 2.67 mg/ml DXR, 5% lactose, 0.45% NaCl and 50 µM desferal was added to the dried lipid (molar ratio PC:PGrcholesterol:alpha-tocopherol succinate:DXR 7:3: 4:0,2:1). About 20 glass beads of 6 mm diameter were also added to facilitate the lipid hydration process. The mixture was then shaken on an ankle shaker for 2-3 h at r.t. The MLVs thus formed were extruded through a 0.2 µm polycarbonate membrane five times in a 50 ml Ami con ultra filtration cell under pressure. The DXR/liposomes had the following characteristics at this stage of the manufacturing process: (a) total DXR incorporated into the liposomes was 85-90% of the original amount of drug added (determined by Sephadex gel permeation chromatography); (b) total lipid concentration was 55-68 pmol/ml (determined by organic phosphate present); (c) the average particle size of the preparation was 300-400 nm in diameter (determined by the dynamic laser light scattering technique); (d) 80$ of total DXR was still liposome-associated after the incubation of the preparation with 50$ human plasma in lt at a 37° at a final liposome concentration of 2.5 pmol/ml.

The DXR/liposome dispersion was treated with Dowex ion-exchange resin (19 mg Dowex 50X4-400 resin used/mg DXN) for 10 min at r.t., resulting in less than about 5% free DXR. After removal of the resin substances, the liposome concentration was brought up to approximately 150 pmol/ml by a centrifugal sedimentation process.

500 µl samples of this concentrated liposome preparation were aliquoted into individual containers and placed in the freeze-dryer where the shelf temperature was set to 5' C (Virtis

freeze dryer). When the sample temperature equilibrated to 5°C, the shelf temperature was reduced to -5°C. The vacuum was activated when the temperature of the samples reached -30° to -35°C. The samples were freeze-dried at -35°C for 48 h in a vacuum of less than 5 pm Hg, raising the shelf temperature to 25°C for 4 h. The samples were then sealed under vacuum and stored at 4°C until use .

Example 2

Size characteristics of reconstituted DXR/liposomes Liposomes were prepared in 125 mM trehalose-0.45% NaCl, 5% lactose-0.45% NaCl or 5% mannitol containing DXR in 10 mM Tris-HCl at pH 7.4 using the method described in Example 1, except that glass beads were not added to the lipid hydration mixture. After the polycarbonate membrane extrusion step, 6 ml of each of the three preparations was diluted sixfold in 10 mM TrIs-HCl-0.9% NaCl and pelleted at approximately 150,000 x g for 30 minutes. The pellets were then suspended in 10 ml of buffer containing the appropriate cryoprotectant. Approximately 30% of total DXR was not liposome-associated at this stage. 500 µl samples were frozen either by exposing them to liquid nitrogen or by placing them in a controlled freezer (Cryomed). Subsequent freeze-drying and reconstitution steps were performed as described in example 1. For particle size measurements, pre- or post-freeze-dried (reconstituted) samples were diluted 200 times in appropriate sugar solutions and measurements performed in a dynamic laser light-scattering instrument (Nicomp model 200, Maryland). As the particle size data shown in Table 1 below shows, both trehalose and lactose effectively prevented the liposome size weight on lyophilization and reconstitution.

Example 4

Effect of lipid concentration

Liposomes were prepared as described in Example 1 with the exception that the liposomes were not treated with Dowex resin. After polycarbonate membrane extrusion, aliquots of the preparations were diluted tenfold with 0.9% NaCl-50pM desferal and the liposomes were pelleted at 150,000 x g for 1 h. The liposome pellets were then resuspended in 5% lactose-0.45% NaCl-50pM desferal until different lipid concentrations (40, 54, 143, 211 mM). These preparations contained 11-13$ of free DXR at this stage. 500 µl aliquots of each of these samples were freeze-dried as described in example 1. The freeze-dried substances were reconstituted by either adding 250 µl or 500 µl of distilled water. Percent free DXN was determined by running the sample on a Sephadex G-50 column and analyzing the amount of DXN in the liposome peak eluted in the void volume of the column relative to what was eluted as free DXN in the bed volume of the column. Samples reconstituted with 500 µl distilled water were analyzed on a column equilibrated in 0.9% saline-50 µM desferal while those reconstituted with 250 µl water were analyzed on a column equilibrated in 1.8% saline-50 µM desferal . The percentage of free DXR in the analyzed samples is shown in table 2. The results are discussed in the text below.

In another study, DXR/liposomes were prepared and lyophilized as described in Table 2, except that the liposomes were treated with Dowe's resin as described in Example 1 to bring the free DXR level to 5-6$ prior to liposome concentration by pelleting at centrifugation. The lyophilized substances were reconstituted in 500 µl of distilled water and percent free DXN was determined using a standard Sephadex G-25 centrifugation technique. The percentage of liposome-bound drug is shown in Table 3.

Example 4

Rehydration conditions

Liposomes were prepared and freeze-dried as described in Example 1. Pre-freeze-dried sample had 4% free DXN and an lipid concentration of 200 pmol/ml. The freeze-dried substances were reconstituted by adding an original amount of distilled water (50, 100, 250 or 500 pl), by allowing the samples to stand at r.t. oxygen, by adding additional water to bring the total volume of water added to 500 pl. Percent liposome-bound DXR was determined as above and the results are shown in Table 4. The data, discussed above, show increasing release of DXR into the medium with greater hypertonic liposome swelling.

Example 5

Temperature conditions during reconstitution

Liposomes were prepared and freeze-dried as described in example 1. Pre-freeze-dried sample contained 6% free DXR and had an lipid concentration of 110 pmol/ml. The freeze-dried substances were stored at 4° or -20° for 2 weeks and reconstituted with 500 µl of water according to the procedure described

1 Table 5. Percent % DXR was analyzed by the above-mentioned Sephadex G-25 centrifugation technique. DXR liposome association data shown in Table 5 show that the release of free drug from liposomes was substantially independent of the storage temperature and the reconstitution temperature.

Example 6

Storage stability and reconstituted DXR/ liposomes

Liposomes were prepared and freeze-dried as described in example 5. The freeze-dried samples were reconstituted immediately after freeze-drying. Reconstituted samples were stored at 4° and percent free DXR was determined using the Sephadex G-25 centrifugation technique at various storage time points, as shown on the left in Table 6 below. The table also indicates the percentages of liposome-bound DXR measured after the indicated storage period.

Despite the fact that the invention has been described with regard to specific formulations, preparation conditions and applications, it is essential that various modifications or changes can be carried out without deviating from the invention.

Claims (15)

1. Freeze-dried adriamycin/liposome composition which, upon reconstitution with a predetermined volume of aqueous medium, provides a liposome concentrate characterized by: a. a concentration of liposomes greater than about 100 mM liposome lipid, b. llposome sizes mainly in a selected size range of between about 0.1-0.5 pm, c. liposome-encapsulated doxorubicin, at a concentration between about 2-10 mole percent lipids, and between about 85-95$ total doxorubicin, and (d) between about 1$ and 10$ cryoprotectant. 2. Composition according to claim 1, characterized in that, when reconstituted with distilled water, it has a significant physiological osmolarity. 3. Composition according to claim 1, characterized in that it further includes liposome-enclosed alpha-tocopherol or a pharmacologically acceptable analog, a derivative, or an ester thereof. 4. Composition according to claim 1, characterized in that it further includes, upon reconstitution with water, a trihydroxam-gelating agent in a molar excess in relation to ferric iron in the concentrate. 5. Composition according to claim 1, characterized in that the liposomes contain between approximately 30-70 mole percent phosphatidylcholine, 20-50 mole percent cholesterol, and between approximately 10-50 mole percent of a negatively charged phospholipid or sterol. 6. Reconstituted doxorubicin/liposome concentrate, characterized by: a. a concentration of liposomes greater than 100 mM, b. llposome sizes predominantly in a selected size range between about 0.1 to 0.5 pm; c. liposome-encapsulated doxorubicin, at a concentration between about 2-10 mole percent liposome lipids, and between about 85% to 95% of total doxorubicin; and d. between about 1 to 10% cryoprotectant. 7. Concentrate according to claim 6, characterized by a significant physiological osmolarity. 8. Concentrate according to claim 6, characterized in that it further includes liposome-enclosed alpha-tocopherol or a pharmacologically acceptable analogue, a derivative or an ester thereof. 9. Concentrate according to claim 6, characterized in that it further includes a trihydroxane chelating agent in a molar excess in relation to ferric iron in the suspension. 10. Concentrate according to claim 6, characterized in that the liposomes contain between approximately 30-70 mole percent phosphatidylcholine, 20-50 mole percent cholesterol, and between approximately 10-50 mole percent of a negatively charged phospholipid or sterol. 11. Process for the production of doxorubicin/liposomes for intravenous administration in cancer therapy, and which contains between approximately 85$ and 90$ of liposome-enclosed doxorubicin, characterized by preparing a liposome dispersion where: a. llposome sizes are generally in a selected size range between about 0.1 to 0.5 µm; b. liposome-encapsulated doxorubicin, at a concentration between about 2-10 mole percent liposome lipids, and at least about 90$ of total doxorubicin, and c. between about 1 to 10% cryoprotectant, and that the concentrate is freeze-dried to form a dehydrated doxorubicin/liposome composition, reconstituting the dehydrated composition with an aqueous medium to a final liposome concentration of greater than about 100 mM lipid, and diluting the reconstituted composition with a physiological aqueous medium to a final liposome concentration suitable for intravenous injection. 12. Method according to claim 11, characterized in that the liposome concentrate which is used further includes liposome-enclosed alpha-tocopherol or a pharmacologically acceptable analogue, a derivative or ester thereof. 13. Method according to claim 11, characterized in that the liposome concentrate which is used further includes a trihydroxam-gelating agent in molar excess in relation to the ferric iron in the suspension. 14. Composition according to claim 11, characterized in that the liposomes used contain between approximately 30-70 mole percent phosphatedylcholine, 20-50 mole percent cholesterol and between approximately 10-50 mole percent of a negatively charged phospholipid or sterol. 15. Method according to claim 11, characterized in that the liposome dispersion used has a lipid concentration of at least approximately 100 mM, and an osmolarity close to physiological osmolarity, and said reconstitution includes distilled water or diluted saline in an amount that comes close to the lipid concentration of the lipid dispersion.