CA2088975A1 - Stable doxorubicin/liposome composition - Google Patents

Abstract

A lyophilized doxorubicin/lipid composition which is stable against doxorubicin breakdown on storage. The composition is formed by lyophilizing a doxorubicin liposome suspension having a pH
between 3.0 and 4.4.

Description

, 208~75 STABLE DOXORUBICIN/LIPOSOME COMPOSITION

1. Field of the Invention The present invention relates to a stable, }yophilized ; doxorubicin/liposome composition.

2. References Anchordogu~, T., et al., Biochim Biophys Acta, 946(2):299 (1988).
Anchordoguy, T., et al., Cryobiology, 24(4):324 ~ (1987). Aubel-Sadron, G., et al., Biochemie, 66:333 ; ~1984).
Bearer, E.L., et al., Biochim Biophys Acta, 693(1):93 (1982)-Crowe, L.M., et al., et al., Biochim Biophys Acta, 769:141 (1983).
Forssen, E.A.,, et al., Proc Nat Acad Sci, USA, 20 78(3):1873 (1981).
Gabizon, A., et al., Cancer Research 42:4734 (1982).
~ Gabizon, A., et al., Cancer Research 43:4730 (1983).
< Gabizon, A., et al., Brit J Cancer, 51:681 (1985).
Higgins, J., et al., J Pharm Pharmacol, 39(8):577 25 (1987).
Higgins, J., et al., J Pharm Pharmacol, 38(4):259 (1986)--:, , . . .
., ~

. .

~,. ~ . ., .
;~ . . . . .. .
, .
., .
.

~ W092/02208 PCT/US91/~551~

o3 ~

Juliano, R.L., et al., Biochem Pharmacol, 27:21 (1978) .
Madden, T.D., Biochim Biophys Acta, 817: 67 (1~85) .
Strauss, G., et al., Biochim Biophys Acta, 858 (1): 169 5(1986).
Szoka, ~., Jr., et al., Proc Nat Acad Sci (USA) 75:4194 (1978).
Szoka, F., Jr., et al., Ann Rev Biophys Bioeng 9:467 (1980) -- 10Young, R.C., et al., N Eng ~ Med, 305: 139 (1981) .

3. Backqround of the Invention Doxorubicin is a potent chemotherapeutic agent effec-; tive against a broad spectrum of neoplasms ~Aubel~Sadron, Young). However, the use of the drug is limited by serio~s side effects. Its acute toxicity includes malaise, nausea, vomiting, myelosuppression, and severe alopecia. In addi-tion, cumulative and irreversible cardiac damage occurs with repeated administration, which seriously limits the use of the drug in protracted treatment (Young).
When administered in liposomal form, doxorubicin re-tains its therapeutic effectiveness against animal tumors, but is significantly less toxic (Forssen, Gabizon, 1985).
The drug-protectiVe effect of liposomes is due, at least in part, to a marked alteration in plasma pharmacokinetics and tissue disposition of the injected drug (Gabizon, 1982, 1983; Juliano).
Recently, it has been recognized that liposomal doxo-rubicin preparations are relatively unstable on storage in liquid form, as evidenced by rapid breakdown of both doxo-- rubicin and liposomal lipids. The instability of doxorubi-cin, when combined in liposomal form, appears to involve -` free radical and related oxidative mechanisms, since the rate and extent of lipid and drug damage is substantially reduced by free-radical quenchers, such as alpha-tocopher-. .
. .
.
; ' ' ' ` , ' ";, ~ , : r ol, and b~ an iron specific chela~or, sucn as desferriox-amine (~ s ?a~en, ~,~9~,73~) ~ he rate and ex ent of drug and lipid damage in a doxorubicin/liPcsome formulation can also be reduced in a lyophilized storage form However, experiments conducted in support of the present invention show tha; relatively high rates of doxorubicin breakdown can oc_ur in a lyophilized formula.ion under conditions OI accelerated s.orage, even in tne presence o~ free-~adicâl ~uenchers or i~on-specific 10 chela~ors 4 Summar~- of _he ~nven.ion ~ = is one general object of _ne invention _o p-o~-ide a lyophilized doxorubicin/liposome composi.ion which is stable agains- doxorubicin breakdown on long-term s.orage 1~ In one aspect .he invention includes a lyophilized doxorubicin/liposome (L-D0~) composi.ion which is charac;erized by less .han 1~%, and prefe-ably less than about '0%, d~xorubicin breakdown after ac-elerated storage in lyophi1ized form at 40C fo- ~ weeks The composition is 20 prepared by lyophilizing an aqueous liposome suspension having a pH be-ween 3 0 and 4 ~, and preferably between 3 5 and 4 0, and con~aining (i) liposomes whose do~inan. lipid componen~s are neutral phosphclipids, chcles~erol, and a negatively charged lipid, (ii) doxorubicin,'a~ a drug lipid 2~ ratio of between ~-10 mole percen~, and a doxorubicin concentration of less than 10 mg/ml, preferably between about 4-6 mg/ml, and (iii) a bulking agent These and other objects and features of the invention will become more fully apparen~ when the following 30 detailed description of the invention is read in conjunc-tion with the accompanying drawings Brief Descri~tion of the Drawinqs ., : :.
:
SVE~STITUTE SHEET

'''~
:. .
. :
';

:

W092/02208 PCT/US91/05~19 Figure l is an HPLc chromatogram showing doxorubicin and breakdown products in a lyophili~ed L-DOX composition of the invention after 2 weeks at 50C;
Figure 2 is an HPLC chromatogram showing doxorubicin and breakdown products in the same L-DOX composition, but after storage in liquid suspension form for 2 weeks at 40C;
and - Figure 3 is an HPLC chromatogram showing doxorubicin and breakdown products in a lyophilized L-DOX composition at pH 4.3 in the presence of succinate after 2 weeks at 50C.

Detailed DescriPtion of the Invention A. Preparation of Lyophilized L-DOX Composition 15Liposomes produced by the method of the present inven-tion are formed from standard vesicle-forming lipids, typi-cally including neutral phospholipids, such as phosphati-dylcholine (PC), negatively charged lipids, such as ~ phosphatidylglycerol (PG), phosphatidylserine (PS), : .20 phosphatidylinositol (PI), phosphatidic acid (PA) or negatively charged sterol lipids, such as cholesterol sulfate and cholesterol hemisuccinate, and cholesterol or .:neutral cholesterol analogs. One preferred formulation includes 40-60 mole percent PC, 10-30 mole percent of a .25 negatively charged lipid, such as PG or cholesterol ~sulfate, and remainder cholesterol.
.,The phospholipid components may contain either satura-ted acyl chains, such as dipalmitoyl acyl chains, or ;.unsaturated acyl chains, such as the mixture of unsaturated ::30 chains present in egg PC or egg PG. The effect of acyl ~chain composition, lipid purity, and negatively charged ~-lipid on storage properties, under accelerated storage conditions, is discussed in Section B below.

.~;~'' .
; ~
' .' .

The liposomal lipids may also include a lipophilic - free-radical quencher, such as alpha-~ocopherol (~-T), or an acià or sal ,hereof, e.g. alpha-toc~pherol s~ccinc_e (~-TS), or bu.yl2~ed hyd~o~:~toluene (3HT), at a preferred concentration of abou- 0-~-2 mole percent of to,al lipids.
One preferred lipid composition, described in Example 1, contains 47.1 mg/ml egg PC (EPC), 19.9 mg1ml egg PG (E?G), 13.4 mg/ml cholesterol, and 0.91 mg/ml alphz-tocopherol acid succinate (~TS), in prelyophilized liquid-10 suspension form- .~nother preferred composi.ion is given in ~ ~xample 2.
- ~he L-DOX compcsi~ion can be prepared by a varie.v of liposome-forming methods, such as have been re~iewec ; (Szok2, 19~0). 'n one lipid-hydration method suitable for ; 15 large-scale liposome production, the lipid components are initially dissolved in a volatile non-polar solvent or i solvent sys-em, such as chloroform, or a chlorofl~orocar30n solvent. Freon ~ 11 or a solven~ Sys~em con~aining FreonT~
~; and 2-5 ~/v percen_ ethanol are well suited for dissolving 20 lipid components for use in the present invention. The low 1 bolling point of this solven- permits rapid, and ;I substantially complete solvent removal under the solvent-removal conditions of the inven~ion~ The solven. poses minir,al health and safety-hazard risks and can be read~ly 25 reclaimed by condensation.
The lipid solution is dried, under vacuum, in a suitable drying vessel. For efficient large-scale liposome production, drying and rehydration steps are prefera~ly : carried out in a planetary mixer par~ially filled with 30 chemical-inert, preferably hydrophobic, spherical particles which become coated with the lipid during dehydration, as has been detailed in co-owned U.S. patent application for "Large-; Scale Liposome ~roduction," filed March 30, 1990, Serial No.
502,222. One suitable particle is a ~/16 inch ' SUB';TITUTE SHEET

,.
:~
. `:, , ` .
'.'` ~ :
. ~ . , ` , ~. `
.
:
, .`': .
, W092/02208 ~ ~ PCT/~S91/0~19 Teflon bead, preferably having a roughened surface, such as obtained from Clifton Plastics ~Clifton Heights, PA). The quantity of particles added to the vessels is preferably ~; such as to produce an area of the particles of between 5about 0.02 and 0.04, and preferably about Q.025 and 0.03 cm2/~mol total lipid added to the vessel.
Planetary mixers having mixer volumes between about 1-10 liters and greater are commercially available. One pre-ferred mixer is a 2- or 4- gallon mixer supplied by Charles Ross and Sons (Hauppauge, NY). Suitable mixing speeds are given in Example 1. Drying with mixing is typically car-~- ried out for 3-4 hours for large lipid-salution volumes.
After complete solvent removal, the particles in the mixer are coated with an irregular film of lipid, providing a high surface area o~ dried lipids.
After solvent removal, the lipids are hydrated with an aqueous volume to a ~inal concentration of lipids of between about 100-300 ~mol/ml. The aqueous medium contains ~ doxorubicin, at a concentration of between about 10-20 ;~ 20 mg/ml, and preferably about 15-16 mg/ml doxorubicin in a pyrogen-free aqueous medium. Typically, the mole concen-tration ratio of drug to phospholipid is between 1:10 and 1:50, depending on the drug being e~capsulated. For :~ example, common doxorubicin/phospholipid concentration ratios are on the order of 1:15.
Other components in the hydration medium may include desferioxamine, at a concentration of about 50 ~M to l mM
and preferably about 0.2 mM, and a physiological salt, such as NaCl, at a f inal concentration which gives a substan-tially isoosmolar solution. The hydration medium may . include a bulking agent ~Section B), at a weight concentra-... .
tion between about 1-10%, and preferably about 5%.
Alternatively, the bulking agent may be added to the re-.,':

:: :

~ .
:, , . . : .,. :, :. . ' W092/02208 P~T/US91~05519 $~

hydrated liposome suspension just prior to lyophilization, as described in Section B.
According to an important feature of the invention, the pH of the aqueous medium is between about 3.0 and 4.4, and preferably between about 3.5-4Ø As will be seen in -Section C below, a pH of 4.4 or below significantly enhances the stability of doxorubicin against breakdown under accelerated storage conditions in lyophilized L-DOX
form. At the same time, maintaining the pH above 3.0 reduces the extent of phospholipid hydrolysis, which is promoted by low pH.
One exemplary aqueous medium, used for producing doxo-rubicin liposomes, is prepared by dissolving desferal in pyrogen-free water, then adding doxorubic~- with stirring until the drug is dissolved. 'rO ~his solution is added the bulking agent and NaCl solution, to a final concentration of components of 200 ~M desferal, 5 mg/ml doxorubicin, 0.45 NaCl, and 5% (w/v) bulking agent. The solution is adjusted to pH 3.8 by addition of HCl (Example l).
20Hydration of the lipid-coated particles preferably occurs under mixing conditions similar to those used in preparing the lipid-coated particles. In particular, for large-scale preparation, the hydration procedure is prefer-ably carried out in a planetary mixer, under the mixing condition described above. The thorough mixing action pro-~ vided by the bi-axial motion of the mixing blades breaks up `~ particle-particle aggregates, and thus exposes more lipid-coated surface area for hydration. The greater degree of lip:d surface exposure to the aqueous medium enhances the rate and final yield of liposome formation.
The liposome suspension may be sized to achieve a selected size distribution of vesicles in a slze range less than about l micron and preferably between about 0.05 to 0.5 microns, and most preferably between about 0.05 and 0.2 . :

. '.'~' W092/0220~ ~ PCT/US91/055l9 ,. ,:

microns. The sizing serves to eliminate larger liposomes and to produce a defined size range having optimal pharma-cokinetic properties.
~everal techniques are available for reducing the sizes and size heterogeneity of liposomes. Extrusion of liposomes through a small-pore polycarbonate membrane is an ; effective method for reducing liposome sizes down to a well-defined size distribution whose average is in the range between about 0.08 and l micron, depending on the pore size of the membrane. Typically, the suspension is cycled through the membrane several times until the desired liposome size distribution is achieved. The liposomes may be extruded through successively smaller-pore membranes, to ~9 achieve a gradual reduction in liposome size.
}5 ~ree doxorubicin, i.e., doxorubicin present in the bulk a~ueous phase of the medium, is preferably removed to increase the ratio of liposome-entrapped to free drug. The drug removal is designed to reduce the final concentration of free doxorubicin to less than about 20% and preferably, less than about 10% of the total drug present in the composition.
Several methods are available for removing free druq from a liposome suspension. A sized liposome suspension can be pelleted by high-speed centrifugation, leaving free drug and very small liposomes in the supernatant. Another method involves concentrating the suspension by ultrafil-- tration, then resuspending the concentrated liposomes in a drug-free replacement medium. Alternatively, gel filtra-tion can be used to separate larger liposome particles from solute (free drug) molecules.
One preferred procedure for removing free doxorubicin, or analog thereof, utilizes an ion-exchange resin capable of bindinq drug in free, but not in liposome-entrapped, form. The preferred resin is a cation-exchanger, since the '' '~' - ' :;, W092/02208 PCT/US91/0~5]9 ~s 37~

. - .g drug is positively charged at neutral pH. One preferred ion exchange resin is Dowex SOW-X4 50-lO0 mesh resin.
After free drug removal, the liposomes may be diluted by addition of a physiological buffer, or concentrated, - 5 e.g., by ultrafiltration, to a desired drug concentration.
Bulking agent, if not already present, is added to the suspension at this stage. The liposomes are then lyophi-lized for storage, as will now be described.
: ;
B. LYoPhilization and Storaqe ; In general, storage of drug/liposomes in liquid suspension form over prolonged periods can lead to loss of entrapped drug from the liposomes, liposome size changes, and degradative changes in the lipid and/or drug molecules in the composition. In general, degradative changes in a liposome or cell suspension can be arrested, for long-term storage, by freezing or lyophilization, i.e., freezing followed by water removal by sublimation under vacuum.
As is known, the steps of freezing, dehydrating, and/or rehydrating dried liposomes can themselves cause undesired changes in drug/liposome properties. Studies conducted in support of the present invention, as well as studies report@d by others (e.g., Anchord~guy, 1988, 1987;
Higgins, 1987, 1986; Strauss), suggest that liposomes may -; 25 undergo two types of disruptive damage during freezing, prior to lyophilization. One type of damage is membrane ;;~ rupture caused by ice crystal formation inside and outside the vesicle spaces during freezing. This type of damage can lead to substantial loss of an encapsulated, water-~- 30 soluble drug molecule. Such drug solute loss due to `membrane rupture can be reduced by cryoprotectants such as ylycerol, DMSO, polyethylene glycol, polypropylene glycol, 1,3,-butanediol, 2,3, butanediol, 1,3-propanediol, a variety of mono- and disaccharides, such as lactose, sucrose, and trehalose, and polysaccharides, such as ~ , .
., .

.
:; .

W092/02208 ~ PCT/US9]/05~19 "' 10 dextran and hydroxyethyl starch (HES) which appear to interrupt or minimize ice crystal formation, when present : both inside and outside the liposomes.
In addition to membrane rupture, as evidenced by a - 5 loss of encapsulated solute, liposomes tend to show a pro-gressive size growth with freezetthaw cycles. The liposome size growth may be due to a solute-exclusion effect in . which the formation of ice crystals concentrates solutes and liposomes into microenvironments of very high lipid and salt concentrations, in effect, forcing liposome together.
- This effect can be reduced by including in the bulk ~extra-liposomal) phase of a liposome suspension, a bulking agent which is effective to interrupt formation of large ice crystals.
15The bulking agent should also have the property of forming a solid, somewhat porous, non-crystalline matrix on drying, to allow escape of water by sublimation from the frozen sample, and to reduce additional solvent exclusion effects due to crystal formation in the bulking agent.
Note that the requirement for a solid bulking agent -~` excludes a variety of small viscous-liquid cryoprotectants such as DMSO, glycerol, ethylene glycol, and propylene glycol as the bulking agent in the present-rinvention.
; one general class of compounds which are suitable as ; 25 bulking agents are non-crystalline carbohydrates, such as j trehalose and lactose, and polysaccharides, such as dextrans and hydxoxyethylcellulose. The use of these compounds as cryoprotectants or bulking agents for lipo-somes, microsomal membranes or blood cells has been reported (e.g., Crowe; Anchordoguy, 1988; Strauss; Madden).

Another general class of bulking agents incIudes amino acids, and amino acid analogs, such as 2-aminobutyric acid, 4-hydroxyproline, sarcosine, glycine betaine, and basic amino acids, such as lysine and histidine, which may inter-: i .

:;~

, . . .
i`4' :
.:

W O 92/02208 PC~r/US91/05519 2B~89 75 .. . 11 act with the negatively charged head groups of the lipo--~ somes (Anchordoguy, 1988).
A third general class of bulking agents includes a variety of non-sugar glycolytic pathway compounds, such as :i5 sodium or potassium salts of tartrate, oxaloacetate, fuma-rate, malate, ketoglutarate, and pyruvate. The bulking agent is preferably a mixture of two or more of these com-pounds, to minimiæe crystal formation effects on freezing and dehydration. As will be seen in Section C below, suc-cinate is not a suitable bulking agent since it enhances doxorubicin breakdown on storage in lyophilized form, nor are di- or tri-dicarboxylic acid compounds, such as ci-trate, since these may precipitate with doxorubicin, which is positively charged. A ~ourth class of bulking agents includes non-saccharide polymeric compounds, such as higher molecular weight polyethylene glycol (PEG), which are (a) solids at room temperature, (b) readily soluble in water, and ~b) pharmaceutically acceptable for parenteral adminis-tration. In particular, PEG polymers with molecular weights above about 1,500-2,000 daltons are contemplated.
In one embodiment, the polymer is included in the bulk phase of the suspension, at a weight concentration between about 1-10 percent.
In another embodiment, the liposomes themselves are ;25 derivatized with a PEG or other polyalkyl oxide, for pur-poses of enhanced lifetime in the bloodstream when admini-stered intravenously, as described in co-owned U.S. patent application for "Liposomes with Enhanced Circulation Times," filed October 10, 1989, Serial No. 425,224. Here the surface-derivatized molecules act to prevent ice cry-stal growth in the region of the liposomes. The liposome -suspension is preferably concentrated to insure a relative-ly high polymer concentration in the bulk phase. Ihe ,iposome suspension may also contain a solution-phase . ~
"..'' ,~, W092/0220~ PC~/~S91/05519 bulking agent selected from one of the classes above, in addition to the liposome-bound agent.
The bulking agents may be included in the aqueous hydration medium used in forming the liposomes, yielding a suspension with an equal concentration of agent in the bulk and encapsulated liposomal aqueous phases. Alternatively, the agent can be added to the final sized liposome suspen-sion, wherein the agent is present only in the bulk phase of the suspension. As indicated above, the final concen-tration of the bulking agent in the suspension is between about l-lO weight percent and preferably about 4-6 weight percent.
In still another embodiment, liposome fusion on ~- freezing and dehydration is minimized by forming the initial liposome dispersion in a low-ionic-strength medium in which the doxorubicin/liposomes have a partially ordered gel structure, by virtue of the surface charge repulsion effects. This type of liposome gel has been described in co-owned U.S. application ~or "Liposome Gel Composition and 20 Method," filed May 22, 1989, Serial No.356,262. The gel formulation re~uires an ionic strength comparable to that of NaCl, at a concentration of less than about 20 mM NaCl.
The medium may contain non-ionic species, such as mono or disaccharide protective agents.
In forming the lyophilized freeze-dried suspension, the liposome dispersion is frozen and lyophilized as described below. The lyophilized material can be recon-stituted, for parenteral injection, by addition of a '';'`~
`; :

.'' ' " ' ' ::; ~ . .
, - , , ~ ' ' ' ' ' :
.

W092/02208 PC~`/US9~/OS519 . ' ~.^:
20~975 . .~ ,;
.. ~

,. .
rehydration medium containing physiological salts or other ionic species.
In a preferred lyophilization method, the suspension is first cooled to 4-5~C, then frozen to -45C and main-tained at this temperature for 12 hours in a conventionallyophiliæer. The lyophilizer chamber is pumped down to - about 200 ~ pressure, after which the chamber temperature is raised, at about 10C/hr, to -20C. The chamber is maintained at this temperature until the lowest reading product thermocouple is within 3C of shelf temperature.
The chamber temperature is now allowed to rise, again at about 10C/hour, to room temperature, and the material is held under vacuum at this temperature for an additional 15 hours. The chamber is backfilled with nitrogen to a pressure of about 2 inches Hg, and the sample vessels stoppered ~or storage. One suitable lyophilizer is a ; Edwards Lyoflex 08 lyophilizer supplied commercially from Edwards High Vacuum, lnc. (New York).
The stability of the lyophilized preparation can be examined by accelerated studies conducted at elevated temperatures, according to known principles. The storage conditions which were used in the studies described below were 40C for 4 weeks, and 50C for two ~eeks. In each storage study, the lyophilized sample in a stoppered vessel was placed in an incubator at the selected temperature and analyæed after the two- or four-week incubation period for doxorubicin breakdown, with the results discussed in Section C.
C. Doxorubicin Stability on Storaqe : 30 Table I below shows composition and pH variables in . several lyophilized L-DOX preparations which were studied under accelerated storage conditions. Each of the 20 pre-parations were prepared according to the methods outlined ;~ above, and detailed for Preparation No. 11 in Example 1.

~ '' ' ', ' .. ..
.

$ W092/02208 PCT/VS~1/0~519 ,.~

` 14 TABLE l , Anionic Anti--~ 5 NO. Lipid Oxidant Desferal Succlnate PH
1 EPG oTS .2 ~ 4.8 2~ EPG oTS .2 + 4.8 3 EPG oTS .2 - 4.8 4 EPG - .2 ~ 4.8 EPG BHT .2 ~ 4.8 6b DPPG ~TS .2 ~ 4.8 7 EPG uTS - + 4.8 8 EPG oTS l.O + 4.8 ; gc EPG oTS .2 + 4.8 CS oTS .2 ~ 4.8 11 EPG oTS .2 ~ 3.8 12 EPG oTS .2 - 3.o , 13 EPG oTS .2 - 3.8 14 EPG ~TS .2 ~ 3.8 EPG B~T .2 - 4.0 16 EPG BHT .2 - 4 2 17 EPG oT .2 - 3 8 18 - _ _ 19 - - .2 + 4.8 - - .2 - 5.3 99~ pure EPC and EPG from Avanti Polar Lipids, Inc.
(Birmingham, AL) were used.
bA 99% pure dipalmitoylphosphatidylcholine (DPPC) and di-palmitoylphosphatidylglycerol (DPPG) (Avanti Polar - Lipids) were used to assess the effect of the degree of fatty acid saturation on the stability profile.
C99% pure EPG from Genzyme Corp. tBoston, MA) was uti-lized.
, dEPG was completely removed from the formulation; sodium ; cholesterol sulfate was incorporated in its place to , 40 provide the negative charge for the optimal incorporation of doxorubicin into liposomes.
, elOmM lactic acid substituted for l0 mM succinic acid in ; the aqueous phase.
:~ 45 After storage under the above accelerated storage condi-tions, each preparation was reconstituted by addition of dis-tilled water to a final doxorubicin concentration of about 5 mg/ml. An aliquot of the material was diluted with mobile phase to a final dilution of 1:50 at a final liquid concentra-, ''. '' :~ , .
"

. .

~, :

~ W092/02208 PCT/US91/0~519 ~ 2088975 .

tion of ab~ut 2.5 ~mol/ml (1.75 mg/ml). At this concentra-tion, the lipid and drug components of the L-DOX suspension ; are in solution. A typical example of this material, chroma-tographed by HPLC is detailed in Example 3.
Figure 1 shows an HPLC profile of doxorubicin and doxoru-bicin breakdown products observed in a lyophilized preparation corresponding to composition No. 11 in Table 1, after acceler-ated storage for 2 weeks at 50C. The peaks in the chromato-gram represent absorbance at the doxorubicin absorbance peak of at 480 nm. Three distinct peaks, with RT (retention time) values of 9.92, 12.75, and 14.4 ~doxorubicin) were observed, with the measured relative peak areas shown. The vertical markers on either side of each peak indicate the portion of the curve that was integraked.

Peak# Area% RT
1 0.695 9.92 - 20 2 1.227 12.75 3 98.077 14.4 , :
The concentration of doxorubicin in the reconstituted sample, both before and after storage at 50C for two weeks, was determined from a standard curve of known concentrations of doxorubicin as a function of peak areas on X~LC. The values calculated were 4.96~0.03 and 4.51~0.05 mg/ml doxorubi-cin before and a~`ter storage (Table 5). Thus, about 9% of the total doxorubicin present in the pre-stored sample was lost to breakdown products on storage.
The doxorubicin peak in the Figure 1 HPLC profile consti-tutes all but about 2% of the detected products, leaving approximately 7% of the breakdown products unaccounted for.
One possible source of this discrepancy is that one or more of the breakdown products have lower absorbance coefficients at - 480 nm, so that these product are either not seen or underes-':

.

~; :
W092/02208 PCT/US91/0~5]9 9 ~ ;i , S

timated in amount. Another possible source of discrepancy is that the breakdown products do not have well-defined peaks, and therefore are not included in the total peak area calcula-tions. Finally, it is possible that some of the breakdown products are not eluted from the column under the chromato-graphy conditions employed.
Figure 2 shows an HPLC profile of the same L-DOX formula-tion as in Figure 1, but after accelerated storage at 40~C for two weeks in liquid suspension form. The relative peak areas of the four peaks are given in Table 3 below. The total -- amount of doxorubicin in the main peak in the figure repre-sents about 67% of the doxorubicin present prior to storage.
- Thus, storage in liquid suspension form substantially increa-ses the breakdown of doxorubicin in the a low-pH L-DOX formu-lation.
. . .

Peak# Area~ RT

` 1 5.193 10.16 2 87.212 14.88 3 4.808 20.06 , . .
4 2.407 21.15 ~- 25 Figure 3 shows an HPLC profile of a lyophilized L-DOX
formulation corresponding to composition No. 2 in Table 1, ` after accelerated storage at 50C for 2 weeks. As seen in the ;' Figure, and in Table 4 below, at least 16 different peaks were identified, with the doxorubicin peak accounting for about 85%
of the total peak area. The total amount of doxorubicin in the main peak in the figure represents about 74% of the doxo-rubicin present prior to storage.

'~,' .
.

.`

W092/02~08 PCT/US91/05519 ~8897~

; ~
. TABLE 4 Peak~ Area~ RT
. 1 0.069 3.34 . 2 0.104 3.6 3 0.129 4.03 : 4 0.216 7.35 0.427 9.25 ~ 6 0.729 10.38 ~ 7 0.077 11.11 ~ .:
- 8 0.705 13.63 : 9 85.509 15.31 0.344 20.75 11 0.623 21.S8 ~ 12 0.631 22.31 . 13 0.124 25.45 . 14 0.302 30.48 0.756 31.55 16 1.255 69.61 . , , . Each of the 20 formulations in Table 1 were tested as . above, for loss of doxorubicin after storage at 40C or 50C
for 4 or 2 weeks, respectively. The results are given in Table 5 below, along with the pH for the suspension, prior to lyophilization.

.
''."
~' :., ..
.
, ............................... . .
..
.

. . .

W092/02208 PCT/VS91/05~l9 . . . ~,~Y

~, ,.
::~ Formulation Formulation ~DOX~3 mgJmL (~ of Original) No. Description p~
Original 2 weeks 4 weeks at 50c b at 40Cb .
1 5.2 4.92+0.072.44+0.043.g9+0.03 (50%) (81%) 2 hiyh purity 5.2 4.90+0.042.13+0.093.58+0.07 EPC & EPG (443) (73%) 3 No Succinate 6.0 4.86+0.043.41+0~07 4.08+0.01 15 4 No aTs 5.3 4.90+0.022.64+0.093.94_0.10 BHT, no aTs 5,3 5.05+0.232.48+0.253.94+0.08 (49%) (7~3%) 6 DPPC/DPPG' 5.0 4.32 1.69 3.72 (39~) (86~) 7 No Desferal 5.0 4.76+0.063.21+0.074.05+0.06 ;-' (67%) (85~) 8 1 mM Desferal 5.1 4.68+0.162.32~0.88 4.1010.02 25 9 high purity 5.1 4.69+0.06~50~) (88%) Chol. Sulphate, 5.2 4.92+0.113.51~0.02 4.28+0,oo 11 No Succinate, 3.8 4.96~0.03 t71%) (87~) , ~ 30 pH 3.8 (91%) (92%) 12 No Succinate, 3.2 4.75+0.034.50+0.03 4.51+0.02 pH 3.0 (95%) (95%) ,~- 13 Lactate, 3.8 4.73+0.023.88+0.124.31+0.03 , No Succinate (82%) (91%) ' 3514 Succinate, 3.9 5.08+0.003.61+0.094.26+0.02 pH 3.8 (71%) (84 No Succinate, 3.9 5.03i0.054.41+0.04 * d pH 4.0 (88~) ' 16 No succinate, 4.1 5.11~0.074.38+0.04 ~ d pH 4.2 (86~) , 17 oT, No ~Ts, 3.7 5.34+0.034.81+0.07 ~ d No Succinate, pH 3.8 (90%) 18 DOX --- 5.07 5.06+0.034.74+0.06 19 DOX in Suc- 4.8 5.25~0.22 3.64_0.og ~94~) cinate buffer (69%) (91~) DOX in Buffer, 5.1 5.49+0.055.15+0.04 5.28+0.14 No Succinate (94%) ~96%) ' ~DOX] was measured using HP~C. Triplicate vials were analyzed at each time point, with the exception that 4 weeks, 40C samples 2-5, 19 and 20 were done in duplicate.
. b 4.5 mL of pre-lyophilization bulk suspension ~nominal [DXN] = 5.0 mg/mL) was filled into 20cc clear type I glass vials for lyophiliza-tion. Reconstitution was accomplished by adding 4.0 mL of water to provide a target ~DXN] of 5.0 mg/mL. All vials were stored upright during the incubation period.
, ~
.~ .

`

.

, WO9~/02208 PCT/US9l/05519 , ~ .
, ~ 2~8~9 75 The most important factors effecting doxorubicin stabi-- lity were reduced pH (between pH 3-4) and the absence of suc-: cinate. Removal of succinic acid from the L-DOX formulation significantly increased drug stability (sample ~3 vs. sample #1). Less doxorubicin degradation was also seen when the pH
of the formulation was lowered to 3.8 (sample #14 vs. sample #1). Maximal stability was obtained with the simultaneous elimination of succinic acid and lowering of pH. For example, only 5% degradation was observed in a pH 3.0 formulation that did not contain succinic acid (sample #12) after two weeks incubation at 50c., whereas a 9-10% decrease in drug potency was exhibited in similar formations at pH 3.8 under the same : condition (sample #11 and sample ~17). Increasing the pH
further to 4.0 and 4.2 (sample #15 and sample #16) resulted in slightly more DOX degradation. Adding lactate resl~lted in significantly more DOX degradation (sample ~13 ~. sample #11), although lactate had a less detrimental effect than succinate in this regard (sample #13 vs. sample #14). A less dramatic effect was observed in samples incubated at 40OC than those incubated at 50OC. No lysophosphatidylcholine (lyso-PC) was detected in any of these samples after incubation at 40OC
or 50C (data not shown).
No systematic trend relating the stability profile to the purity, degree of saturation of phospholipid used, or type of negatively charged lipid used was seen, based on results from , samples incubated at 50 C. The use of EPG from Genzyme Corporation and sodium cholesterol sulfate ' DPPC - Dipal~itoylphosphatidy}choline DPPG - dipalmitoylphosphatidylglycerol 50 mL batch size was prepared for this formulation; only - one vial was analayzed for each time point.
~ Data are not yet available at this time.
pH of the prelyophilization bulk suspension except for samples 6-10 ~pH of reconstituted samples).
."

.. . . .

~ W092/0~208 P~T/US9l/05519 1.9~ ;

appears to result in some improvement in DOX potency preser-vation compared to the original formulation (sample #l) at the end of the incubation period. No dramatic effect was observed in samples incubated at 40C.
To assess the effect of the antioxidant on the stability profile, a-Ts was removed (sample #4) and replaced with BHT
(sample #5). No significant difference in DOX stability was observed in these samples at either 50C or 40.
Doxorubicin stability was compared in formations con-taining no Desferal (sample #7), 200 ~M (sample #l) or 1 mM
Desferal (sample #8). The samples without Desferal, incubated at 50C, showed a slight improvement in DOX stability.
The present invention is useful as an improved storagable form of doxorubicin/liposomes, for use in the treatment of a variety of tumor types. Reduced side effects have been observed in Phase I and Phase II clinical trials with doxoru-bicin liposome formulations prepared by reconstituting lyophilized preparations. The present ~-DOX invention provides the additional advantage of long-term stability on storage, without significant loss of active drug or generation of undesired breakdown products. The preparation is easily prepared by a method that is suitable for large-scale manufac-ture.
It will be appreciated that the advantages and features of the present invention will apply to a variety of related anthracene glycoside anti-neoplastic drugs, such as daunomy-cin, carcinomycin, N-acetyladriamycin, N-acetydaunomycin, rubidazone, 5-imidodaunomycin, and epirubicin.
The following example illustrates the method o~ the in-vention for preparing doxorubicin liposomes. The example il-lustrates, but in no way is intended to limit the scope o~ the invention.
:

:` ' ,''' ' ' .

W092/0220~ PCT/US91/05519 - 2`~

Example 1 Preparation of LYophilized L-DOX Composition (Pre~aration~ rom Table 1) :A 300 ml chloroform solution containing 20.1 g EPC
(Lipoid), 8.5 g EPG (Asahi), 5.7 g CH (Croda~, and 0.4 g ~-tocopherol succinate (Henkel~) was added to a 2-liter round-bottomed flask containing 180 g of 3 mm diameter glass beads and the whole dried in vacuo using a rotary evaporator. The ; lipid film was subsequently exposed to a vacuum of 50 mTorr overnight to complete drying.
The lipids were hydrated to a final total lipid concen-tration of about 240 ~mole/ml by addition of a doxorubicin-containing solution and mechanically agitating the mixture for 2 hours at room temperature. This solution contained about 15 mg/ml DOX, 5% (w/v) lact.ose monohydrate, 0.4% (w/v) sodium chloride, 200 ~M desferal in water and was prepared by dissol-ving the doxorubicin in water and adding the other excipients when complete dissolution was achieved. The pH of the mixture was adjusted to 3.8 with HCl. The resultant liposome suspen-sion was sized by passage five times through 0.4 ~m and fivetimes through 0.2 ~m pore-size polycarbonate membranes.
Unincorporated DOX was removed by passing the sized liposome suspension over a Dowex 50W-X4 cation exchange resin.
Finally, the suspension was diluted to a DOX concentration of about 5 mg/ml prior to lyophiliæation, using a 5% lactose, 0.4% sodium chloride, 200 ~M desferal solution, pH 3.8. The characteristics o~ this pre-lyophilization solution are given in Table 6.
.: .

.:
:
' ... . . . . . .

. . ' . .
' ' ' ' ' :
. , :
.

W092/022~8 PCT/US91/0~519 ~:: 9~

Theoretical.
. Com~osition .m~tml DOX HCl 5 0 -:- EPC 47.1 - EPG 19.9 CH 13.3 - ..................... a -TS 0. 91 .~ 10 Desferoxamine mesylate 0.132 ... Lactose monohydrate 52.6 NaCl 4.0 Water for injection qs to 1 ml HCl qs to pH 3.80 Assay values for pre-lyophilization solution:
Doxorubicin HCl (mg/ml) 4.91 % encapsulation 96 Liposome diameter ~nm) 290 . Example 2 Large Scale L-DOX Production :' 25 A lipid solution containing 1500 ml Freon llTM, 151.1 g EPC (Lipoid), 63.9 g EPG (Asahi), 42.9 g cholesterol (Croda), and 0.86 g o~ butylated hydroxytoluene (Penta) was added to a 2-gallon double planetary mixer (Ross mode~ 130 LD~) which . 30 contained about 4,100 PTFE Teflon beads (Clifton Plastics) having a mean diameter o~ 1/4 inch. The beads had been washed in an aqueous solution containing a detergent (Alconox) to : remove any oily residues from the sur~aces o~ the beads, rinsed with ethanol, and dried, before placement into the mixer.
The mixer was sealed and the pressure inside the mixer was reduced while the solution and beads were mixed at an orbital stirring rate of 20 rpm and an axial stirring rate of : 26 rpm for about two hours, to trans~orm the lipids to a solid state attached to the beads. A vacuum was maintained for an additional 3 hours without mixing to achieve additional sol-.~ vent removal.
. . .
,;

.. : . , ..:i : , ~.

: - 23 After solvent removal, an aqueous solution of doxorubicin HCl to a final lipid concentration of about 262 ~mol/ml was added to the lipid coated particles in the planetary mixer.
The drug solution was prepared to contain 0.132 mg/ml deferox-amine mesylate in pyrogen-free distilled water. Doxorubicin hydrochloride (Farmitalia Carlo Erba) was then added with stirring to a final concentration of about 12.5 mg/ml. After complete dissolution of the drug, lactose was added to a final concentration of 5 weight percent and NaCl 0.4 weight percent.
The pH of the solution was then adjusted to 3.7 with HCl solutlon .
Hydration of the lipid was carried out at a temperature of 25 C. for 2 hours, with mixer stirring at abou, 25 rpm.
The resulting liposome suspension was sized by passes through 0.4 and 0.2 micron polycarbonate filters. Free doxorubicln was reduced by treating the sized-liposome suspension with a Biorad AG 50w-X4 50-lOO ion exchange resin. The preparation was then diluted with a solu~ion containing deferoxamine mesylate, lactose, sodium chloride and hydrochloric acid in pyrogen-free distilled water.
The approximate final composition of the liposome suspen sion was:
4.3 mg/ml doxorubicin hydrochloride 40.0 mg/ml EPC
16.9 mg/ml EPG
11.3 mglml cholesterol 0.23 mg.ml butylated hydroxytoluene 0.132 mg.ml deferoxamine mesylate 50.0 mg/ml lactose 4.0 mg/ml sodium chloride The preparation had the following properties: pH of approxi-mately 3.8, mean particle size of approximately 232 nm;
approximately 96~ of the total drug was liposome associated.

Example 3 HPLC Chr~mato~ra~hV

',. - :
.

~ W092/02208 PCT/US91/05519 . 9~a C~

.: .-,,~
; A. HPLC Chromatography - Lyophilized L-DOX was reconstituted with water to a final doxorubicin concentration of about 5 mg/ml. A 0.25 ml aliquot of the rehydrated suspension was diluted to 5 ml with mobile phase.
The separation of the doxorubicin and its degradation products was carried out on a Waters HPLC using a Whatman Partisil ODS-3 column, 250 x 4.6 mm held at ambient tempera-ture. The mobile phase was 43~ aqueous buffer in methanol.
The aqueous buffer was 95 mM ammonium phosphate/ 5 mM tri-ethylamine, pH 4Ø 15 ~l of sample was injected in the system at a 1 ml/min flow rate. Effluent was monitored at 480 nm. An external standard, prepared fresh each day, consisted of doxorubicin powder dissolved in mobile phase to a final concentration of 0.35 mg/ml.

B. Analysis after Storage After an i~cubation period of two weeks at 50C, the L-- DOX was analysed by HPLC, as above. Doxorubicin potency was reduced 9~. The level of doxorubicin breakdown appeared to ; plateau at this level: a further two-week incubation at 40C
did not result in further reduction of doxorubicin concentra-tion.

Although preferred methods and formulations have been ,, . ,~
','`',~

,~
" '' " :~
., .
. , " . , . ' ...

e~
. :` W092/02208 PCI/US91/05S19 8~97S

. . .

; described, it will be apparent that various changes and modi-fications may be made without departing from the invention.

' ' , ' .' , . ' , .

.

,, -.. .. .

Claims (12)

1. A lyophilized doxorubicin/liposome composition which is:
(a) characterized by less than 15% doxorubicin breakdown after storage in lyophilized form at 40°C for 4 weeks, and (b) prepared by forming liposomes, in suspension form, at a pH of between about 3.2 and 4.1 and at a concentration of succinic acid which is substantially less than 10 mM, and lyophilizing the suspension.

2. The composition of claim 1, which is characte-rized by less than 10% doxorubicin breakdown after such storage, and the suspension pH is between 3.5 and 4Ø

3. The composition of claim 1, which is substan-tially free of succinate.

4. The composition of claim 1, wherein the lipo-somes contain between about 40-60 mole percent phospha-tidylcholine, 20-40 mole percent cholesterol, and 10-30 mole percent phosphatidylglycerol.

5. The composition of claim 1, wherein the lipo-somes contain a lipophilic free radical quencher selected from the group consisting of alpha-tocopherol or acid or salt thereof, and butylated hydroxytoluene.

6. The composition of the 1, wherein daxorubicin is present in the suspension at a concentration between about 4-6 mg/ml.

7. A method of storing doxorubicin in a stable, liposome-entrapped form, as evidenced by less than 15%
doxorubicin breakdown after storage at 40°C for 4 weeks, comprising lyophilizing an aqueous liposome suspension having a pH between 3.0 and 4.4 and containing (i) liposomes whose dominant lipid components are neutral phospho-lipids, cholesterol, and a negatively charged lipid, (ii) doxorubicin, at a drug:lipid ratio of between 5-10 mole percent, and a doxorubicin concentration of less than 10 mg/ml, and (iii) a cryoprotectant, and storing the lyophilized suspension.

8. The method of claim 7, which is effective to produce less than 10% conversion of doxorubicin to inactive degradation products after such accelerated storage, and the suspension pH is between 3.5 and 4Ø

9. The method of claim 7, wherein the liposomes contain between about 40-60 mole percent phosphatidyl-choline, 20-40 mole percent cholesterol, and 10-30 mole percent phosphatidylglycerol.

10. The method of claim 7, wherein the liposomes contain a lipophilic free radical quencher selected from the group consisting of alpha-tocopherol or acid or salt thereof, and butylated hydroxytoluene.

11. The composition of claim 7, wherein doxorubi-cin is present in the suspension at a concentration between about 4-6 mg/ml.

12. A method of reducing the extent of doxorubi-cin breakdown in a liposome composition containing doxorubicin, when the composition is stored in a lyophilized form over a period of two week or more, comprising preparing the composition, in liquid form, with a pH between about 3.2 and 4.1 and a concentration of succinic acid which is substantially less than 10 mM, and lyophilizing the composition.