Compositions for delivery of drug combinations (2025)

U.S. patent number 10,058,507 [Application Number 14/990,167] was granted by the patent office on 2018-08-28 for compositions for delivery of drug combinations. This patent grant is currently assigned to CELATOR PHARMACEUTICALS, INC.. The grantee listed for this patent is CELATOR PHARMACEUTICALS, INC.. Invention is credited to Marcel Bally, Troy Harasym, Andrew Janoff, Lawrence Mayer, Clifford Shew, Paul Tardi, Murray Webb.

Compositions for delivery of drug combinations

Abstract

Compositions which comprise delivery vehicles having stablyassociated therewith non-antagonistic combinations of two or moreagents, such as antineoplastic agents, are useful in achievingnon-antagonistic effects when combinations of drugs areadministered.

Prior Publication Data

Related U.S. Patent Documents

References Cited [Referenced By]

U.S. Patent Documents

Foreign Patent Documents

Other References

Primary Examiner: Kishore; Gollamudi
Attorney, Agent or Firm: Morrison & Foerster LLP

Parent Case Text

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is a continuation of U.S. Ser. No. 10/553,373having an international filing date of 16 Apr. 2004, which is thenational phase of PCT application PCT/US2004/011812 filed 16 Apr.2004, which is a continuation of U.S. application Ser. No.10/417,631 filed 16 Apr. 2003 (now U.S. Pat. No. 7,850,990), whichis a continuation-in-part of U.S. Ser. No. 10/264,538 filed 3 Oct.2002 (abandoned), which claims benefit under 35 U.S.C. .sctn.119(e) of provisional applications U.S. Ser. No. 60/326,671 filed 3Oct. 2001; Ser. No. 60/341,529 filed 17 Dec. 2001; Ser. No.60/356,759 filed 15 Feb. 2002; Canadian informal application SerialNo. CA 2,383,259 filed 23 Apr. 2002; provisional applications U.S.Ser. No. 60/401,984 filed 7 Aug. 2002 and U.S. Ser. No. 60/408,733filed 6 Sep. 2002. The contents of these applications areincorporated herein by reference.

Claims

The invention claimed is:

1. A pharmaceutical composition for parenteral administration,comprising particulate delivery vehicles having associatedtherewith at least a first antineoplastic agent and a secondantineoplastic agent, wherein said first and second agents are in amole ratio which exhibits a non-antagonistic cytotoxic orcytostatic effect in an in vitro assay, over at least 20% of theconcentration range over which the fraction of cells affected is0.2-0.8; and wherein said first and second agents are associatedwith the delivery vehicles to maintain said non-antagonistic ratioin the blood for at least one hour after administration, whereinsaid delivery vehicles comprise liposomes, and/or lipid micelles,and/or block copolymer micelles, and/or polymer microparticles,and/or polymer nanoparticles, and/or polymer lipid hybrid systems,and/or derivatized single chain polymers, and wherein thecomposition comprises one or more additional therapeuticagents.

2. The composition of claim 1 wherein said delivery vehicles are 4to 6,000 nm in diameter.

3. The composition of claim 1 wherein said delivery vehicles have amean diameter of between 4.5 and 500 nm.

4. The composition of claim 1 wherein said vehicles have a meandiameter of less than 250 nm.

5. The composition of claim 1 wherein said delivery vehicles arefrom 4 .mu.m to 50 .mu.m in diameter.

6. The composition of claim 1 wherein said delivery vehiclescomprise liposomes.

7. The composition of claim 1 wherein said first and secondantineoplastic agents are co-encapsulated.

8. The composition of claim 1 wherein at least one of theantineoplastic agents is selected from the group consisting of aDNA damaging agent, a DNA repair inhibitor, a topoisomerase Iinhibitor, a topoisomerase II inhibitor, a cell checkpointinhibitor, a CDK inhibitor, a receptor tyrosine kinase inhibitor, acytotoxic agent, an apoptosis inducing agent, an antimetabolite, acell cycle control inhibitor, a therapeutic lipid, a telomeraseinhibitor, an anti-angiogenic agent, a mitochondrial poison, asignal transduction inhibitor and an immunoagent.

9. The composition of claim 7 wherein the first antineoplasticagent is a cytoxic agent and the second antineoplastic agent is acell-cycle inhibitor, or wherein the first antineoplastic agent isa DNA damaging agent and the second antineoplastic agent is a DNArepair inhibitor, or wherein the first antineoplastic agent is atopoisomerase I inhibitor and the second antineoplastic agent is aS/G.sub.2- or a G.sub.2/M-checkpoint inhibitor, or wherein thefirst antineoplastic agent is a G.sub.1/S checkpoint inhibitor or acyclin-dependent kinase inhibitor and the second antineoplasticagent is a G.sub.2/M checkpoint inhibitor, or wherein the firstantineoplastic agent is a receptor kinase inhibitor and the secondantineoplastic agent is a cytotoxic agent, or wherein the firstantineoplastic agent is an apoptosis-inducing agent and the secondantineoplastic agent is a cytotoxic agent, or wherein the firstantineoplastic agent is an apoptosis-inducing agent and the secondantineoplastic agent is a cell-cycle control agent, or wherein thefirst antineoplastic agent is a telomerase inhibitor and the secondantineoplastic agent is a cell-cycle control inhibitor, or whereinthe first and second antineoplastic agents are antimetabolites, orwherein the first and second antineoplastic agents are cytotoxicagents, or wherein the first antineoplastic agent is a therapeuticlipid and the second antineoplastic agent is a cytotoxic agent, orwherein the first antineoplastic agent is a topoisomerase Iinhibitor and the second antineoplastic agent is a DNA repairinhibitor, or wherein the apoptosis-inducing antineoplastic agentis a serine-containing lipid.

10. The composition of claim 1 wherein the first antineoplasticagent is irinotecan and the second antineoplastic agent is 5 FU orFUDR, or wherein the first antineoplastic agent is cisplatin (orcarboplatin) and the second antineoplastic agent is 5 FU or FUDR,or wherein the first antineoplastic agent is idarubicin and thesecond antineoplastic agent is AraC or FUDR, or wherein the firstantineoplastic agent is oxaliplatin and the second antineoplasticagent is 5 FU or FUDR, or wherein the first antineoplastic agent isirinotecan and the second antineoplastic agent is cisplatin (orcarboplatin), or wherein the first antineoplastic agent isgemcitabine and the second antineoplastic agent is cisplatin (orcarboplatin), or wherein the first antineoplastic agent ismethotrexate and the second antineoplastic agent is 5 FU or FUDR,or wherein the first antineoplastic agent is paclitaxel and thesecond antineoplastic agent is cisplatin (or carboplatin), orwherein the first antineoplastic agent is etoposide and the secondantineoplastic agent is cisplatin (or carboplatin), or wherein thefirst antineoplastic agent is docetaxel or paclitaxel and thesecond antineoplastic agent is doxorubicin, or wherein the firstantineoplastic agent is doxorubicin and the second antineoplasticagent is vinorelbine, or wherein the first antineoplastic agent iscarboplatin and the second antineoplastic agent is vinorelbine, orwherein the first antineoplastic agent is 5-FU or FUDR and thesecond antineoplastic agent is gemcitabine.

11. The composition of claim 6 wherein the first antineoplasticagent is daunorubicin and the second antineoplastic agent isAraC.

12. The composition of claim 1 wherein the one or more additionaltherapeutic agents is a cytotoxic agent.

13. The composition of claim 1 where the one or more additionaltherapeutic agents is selected from an apoptosis-inducing agent, asignal transduction inhibitor or a receptor tyrosine kinaseinhibitor.

14. A method to prepare a composition of claim 1, which methodcomprises stably associating with said particulate deliveryvehicles a mole ratio of agents that has been determined to exhibita non-antagonistic cytotoxic or cytostatic effect in an in vitroassay over at least 20% of the concentration range over which thefraction of cells affected is 0.2-0.8; wherein said stableassociation is such that said ratio is maintained in the blood forat least one hour after administration, wherein said method furthercomprises adding said one or more additional therapeuticagents.

15. The method of claim 14, wherein said ratio has been determinedin an assay that employs testing at least one ratio of said agentsat a multiplicity of concentrations and applying an algorithm tocalculate a synergistic, additive, or antagonistic effect for saidratio over a range of concentrations.

16. The method of claim 15 which employs testing a multiplicity ofratios, and wherein said algorithm is the Chou-Talalay medianeffect method.

17. The method of claim 14 wherein the delivery vehicles areliposomes.

18. The method of claim 17 wherein the first antineoplastic agentis irinotecan and the second antineoplastic agent is 5-FU or FUDR,or wherein the first antineoplastic agent is cisplatin (orcarboplatin) and the second antineoplastic agent is 5-FU or FUDR,or wherein the first antineoplastic agent is idarubicin and thesecond antineoplastic agent is AraC or FUDR, or wherein the firstantineoplastic agent is oxaliplatin and the second antineoplasticagent is 5-FU or FUDR, or wherein the first antineoplastic agent isirinotecan and the second antineoplastic agent is cisplatin (orcarboplatin), or wherein the first antineoplastic agent isgemcitabine and the second antineoplastic agent is cisplatin (orcarboplatin), or wherein the first antineoplastic agent ismethotrexate and the second antineoplastic agent is 5-FU or FUDR,or wherein the first antineoplastic agent is paclitaxel and thesecond antineoplastic agent is cisplatin (or carboplatin), orwherein the first antineoplastic agent is etoposide and the secondantineoplastic agent is cisplatin (or carboplatin), or wherein thefirst antineoplastic agent is docetaxel or paclitaxel and thesecond antineoplastic agent is doxorubicin, or wherein the firstantineoplastic agent is doxorubicin and the second antineoplasticagent is vinorelbine, or wherein the first antineoplastic agent iscarboplatin and the second antineoplastic agent is vinorelbine, orwherein the first antineoplastic agent is 5-FU or FUDR and thesecond antineoplastic agent is gemcitabine.

19. The method of claim 14 wherein the first antineoplastic agentis daunorubicin and the second antineoplastic agent is AraC.

20. A method to treat a disease or condition in a subject, whichmethod comprises administering to a subject in need of suchtreatment the pharmaceutical composition of claim 1.

Description

TECHNICAL FIELD

The invention relates to compositions and methods for improveddelivery of synergistic or additive combinations of therapeuticagents. More particularly, the invention concerns delivery systemswhich ensure the maintenance of synergistic or additive ratios whenthe agents are delivered to an intended target by providingformulations comprising delivery vehicles.

BACKGROUND ART

The progression of many life-threatening diseases such as cancer,AIDS, infectious diseases, immune disorders and cardiovasculardisorders is influenced by multiple molecular mechanisms. Due tothis complexity, achieving cures with a single agent has been metwith limited success. Thus, combinations of agents have often beenused to combat disease, particularly in the treatment of cancers.It appears that there is a strong correlation between the number ofagents administered and cure rates for cancers such as acutelymphocytic leukemia. (Frei, et al., Clin. Cancer Res. (1998)4:2027-2037). Clinical trials utilizing combinations ofdoxorubicin, cyclophosphamide, vincristine, methotrexate withleucovorin rescue and cytarabine (ACOMLA) or cyclophosphamide,doxorubicin, vincristine, prednisone and bleomycin (CHOP-b) havebeen successfully used to treat histiocytic lymphoma (Todd, et al.,J. Clin. Oncol. (1984) 2:986-993).

The effects of combinations of drugs are enhanced when the ratio inwhich they are supplied provides a synergistic effect. Synergisticcombinations of agents have also been shown to reduce toxicity dueto lower dose requirements, to increase cancer cure rates(Barriere, et al., Pharmacotherapy (1992) 12:397-402; Schimpff,Support Care Cancer (1993) 1:5-18), and to reduce the spread ofmulti-resistant strains of microorganisms (Shlaes, et al., Clin.Infect. Dis. (1993) 17:S527-S536). By choosing agents withdifferent mechanisms of action, multiple sites in biochemicalpathways can be attacked thus resulting in synergy (Shah andSchwartz, Clin. Cancer Res. (2001) 7:2168-2181). Combinations suchas L-canavanine and 5-fluorouracil (5-FU) have been reported toexhibit greater antineoplastic activity in rat colon tumor modelsthan the combined effects of either drug alone (Swaffar, et al.,Anti-Cancer Drugs (1995) 6:586-593). Cisplatin and etoposidedisplay synergy in combating the growth of a human small-cell lungcancer cell line, SBC-3 (Kanzawa, et al., Int. J. Cancer (1997)71(3):311-319).

Additional reports of synergistic effects are found for:Vinblastine and recombinant interferon-.beta. (Kuebler, et al., J.Interferon Res. (1990) 10:281-291); Cisplatin and carboplatin(Kobayashi, et al., Nippon Chiryo Gakkai Shi (1990) 25:2684-2692);Ethyl deshydroxy-sparsomycin and cisplatin or cytosine arabinoside(AraC) or methotrexate or 5-FU or vincristine (Hofs, et al.,Anticancer Drugs (1994) 5:35-42); All trans retinoic acid andbutyric acid or tributyrin (Chen, et al., Chin. Med. Engl. (1999)112:352-355); and Cisplatin and paclitaxel (Engblom, et al., Br. J.Cancer (1999) 79:286-292).

In the foregoing studies, the importance of the ratio of thecomponents for synergy was recognized. For example, 5-fluorouraciland L-canavanine were found to be synergistic at a mole ratio of1:1, but antagonistic at a ratio of 5:1; cisplatin and carboplatinshowed a synergistic effect at an area under the curve (AUC) ratioof 13:1 but an antagonistic effect at 19:5.

Other drug combinations have been shown to display synergisticinteractions although the dependency of the interaction on thecombination ratio was not described. This list is quite extensiveand is composed mainly of reports of in vitro cultures, althoughoccasionally in vivo studies are included.

In addition to the multiplicity of reports, a number ofcombinations have been shown to be efficacious in the clinic. Theseare described in the table below.

TABLE-US-00001 REFERENCE DRUG 1 DRUG 2 DRUG 3 Langer, et al. (1999)Drugs 58 Suppl. Cisplatin or +UFT (Tegafur/ 3: 71-75 Vindesineuracil) FDA.sup.a (Colon or Rectal Cancer) Leucovorin +5-FU FDA(Colon or Rectal Cancer) Irinotecan +Leucovorin +5-FU FDA (BreastCancer) Herceptin +Paclitaxel FDA (Breast Cancer) Xeloda +Docetaxel(other names: Capecitabine) FDA (Ovarian and Lung Cancer)Paclitaxel +Cisplatin FDA (Lung Cancer) Etoposide +OtherFDA-approved Chemotherapeutic agents FDA (Lung Cancer) Gemcitabine+Cisplatin FDA (Prostate) Novantrone +Corticosteroids (mitoxantronehydrochloride) FDA (Acute Nonlymphocytic Novantrone +OtherFDA-approved drugs Leukemia) FDA (Acute Nonlymphocytic Daunorubicin+Other FDA-approved drugs Leukemia/Acute Lymphocytic Leukemia)(DNR, Cerubidine) FDA (Chronic Myelogenous Busulfex+Cyclophosphamide Leukemia) (Busulfan; (Cytoxan) 1,4-butanediol,dimethanesulfonate; BU, Myleran) .sup.aFDA: United States Food andDrug Administration

In addition, certain other combinations can be postulated fromvarious reports in the literature to have the potential forexhibiting non-antagonistic combination effects or clinicalefficacy or accepted as the standard of care by region studygroups. These are:

TABLE-US-00002 DISEASE DRUG 1 DRUG 2 DRUG 3 (Colon Cancer)Oxaloplatin +5-FU (or FUDR) +Leucovorin (Metastatic Breast Cancer)Taxol +Doxorubicin Adriamycin +Cytoxan (cyclophosphamide)(doxorubicin) Methotrexate +5-FU (or FUDR) +Cytoxan Vinblastine+Doxorubicin (Non-small Cell Lung Cancer) Carboplatin +TaxolCisplatin +Docetaxel (Taxotere .RTM.) Vinorelbine +CisplatinIrinotecan +Cisplatin (Small Cell Lung Cancer) Carboplatin +TaxolCisplatin +Etoposide (Prostate Cancer) Estramustine +TaxolEstramustine +Mitoxantrone Estramustine +Taxotere (Hodgkin'sLymphoma) Bleomycin +Vinblastine (as part of ABDV: Adriamycin,Bleomycin, DTIC, Vinblastine) (Non-Hodgkin's Lymphoma) Carboplatin+Etoposide (as part of ICE: Ifosfamide, Carboplatin, Etoposide)(Melanoma) IL-2 +Cisplatin (Acute Myeloid Leukemia) Daunorubicin+Cytosine Arabinoside Vincristine +Doxorubicin (Bladder Cancer)Carboplatin +Taxol Carboplatin +Gemcitabine Gemcitabine +TaxolVinblastine +Doxorubicin (as part of MVAC: Methotrexate,Vinblastine, Adriamycin, Cisplatin) (Head and Neck Cancer) 5-FU (orFUDR) +Cisplatin +Leucovorin (Pancreatic Cancer) Gemcitabine +5-FU(or FUDR) Additional Combinations: Carboplatin +5-FU (or FUDR)Carboplatin +Irinotecan Irinotecan +5-FU (or FUDR) Vinorelbine+Carboplatin Methotrexate +5-FU (or FUDR) Idarubicin +AraCAdriamycin +Vinorelbine Safingol +Fenretinide

Despite the aforementioned advantages associated with the use ofsynergistic drug combinations, there are various drawbacks thatlimit their therapeutic use. For instance, synergy often depends onvarious factors such as the duration of drug exposure and thesequence of administration (Bonner and Kozelsky, Cancer Chemother.Pharmacol. (1990) 39:109-112). Studies using ethyldeshydroxy-sparsomycin in combination with cisplatin show thatsynergy is influenced by the combination ratios, the duration oftreatment and the sequence of treatment (Hofs, et al., supra).

It is thus known that in order for synergy to be exhibited by acombination of agents, these agents must be present in amountswhich represent defined ratios. Indeed, the same combination ofdrugs may be antagonistic at some ratios, synergistic at others,and additive at still others. It is desirable to avoid antagonisticeffects, so that the drugs are at least additive. The presentinvention recognizes that the result obtained at an individualratio is also dependent on concentration. Some ratios areantagonistic at one concentration and non-antagonistic at another.The invention ensures ratios of components in the synergistic oradditive range by delivering these agents in formulations thatmaintain the desired or administered ratio when the target locationin the subject are reached and by selecting the ratios to bepredominantly non-antagonistic at a desired range ofconcentrations, since the concentration at the target may bedifferent from that administered.

PCT publication WO 00/51641 describes administering a combinationof antiviral agents which is said to be synergistic. In vitro testswere used to determine synergistic ratios. However, there is noteaching of any mode of administration which would maintain thisratio in vivo. Indeed, the publication states that the componentsmay be administered sequentially or simultaneously.

PCT publication WO 01/15733 describes putatively synergisticcompositions for treating autoimmune disease. Again, the method offormulation does not ensure maintenance of this ratio afterdelivery.

Daoud, et al., Cancer Chemother. Pharmacol. (1991) 28:370-376,describe synergistic cytotoxic actions of cisplatin and liposomalvalinomycin on human ovarian carcinoma cells. This paper describesan in vitro assay in which cisplatin which is free and valinomycinwhich is encapsulated in liposomes are used to treat cultures ofCaOV-3, a human ovarian tumor-derived cell line. The authorsdetermined the concentration ranges over which synergism andantagonism was exhibited. Liposome encapsulation was employed tosolubilize the valinomycin. As the experiments are performed invitro, in vivo delivery is irrelevant.

U.S. Pat. No. 6,214,821 issued 10 Apr. 2001 to Daoud, describespharmaceutical compositions containing topoisomerase I inhibitorsand a staurosporine. The claims appear to be based on the discoverythat staurosporines have the ability to abrogate topoisomerase Iinhibitor-induced S-phase arrest and to enhance its cytotoxicity tohuman breast cancer cells lacking normal p53 function. Noparticular pharmaceutical formulation is suggested.

U.S. Pat. No. 5,000,958 to Fountain, et al., describes mixtures ofantimicrobial agents encapsulated in liposomes which are said toexert an enhanced therapeutic effect in vivo. Suitable ratios ofantimicrobial agents are determined by a combination effect testwhich empirically tests for synergy in vitro. There is nodiscussion of assuring a synergistic ratio over a range ofconcentrations.

Schiffelers, et al., J. Pharmacol. Exp. Therapeutic (2001)298:369-375, describes the in vivo synergistic interaction ofliposome co-encapsulated gentamicin and ceftazidime. The desiredratios were determined using a similar combination effect test tothat of Fountain (supra), but there is no discussion ofdetermination of a ratio wherein synergism is maintained over arange of concentrations.

The present invention recognizes, first, that it is possible tomaintain a determined synergistic or additive ratio of therapeuticagents by controlling the pharmacokinetics of the formulation inwhich they are administered, and second, that the non-antagonisticratio must be exhibited over a range of concentrations, since theconcentration of components in a drug cocktail which reaches thetarget tissue may not be the same as that which is administered.The problem of maintaining synergy or additivity is solved by therecognition that when therapeutic agents are encapsulated in (i.e.,stably associated with) delivery vehicles, such as liposomes, thedelivery vehicles determine the pharmacokinetics and thus agentswhich are encapsulated will behave in a similar manner, and byselecting ratios which are predominantly synergistic/additive overa range of concentrations.

DISCLOSURE OF THE INVENTION

The invention relates to methods for administering non-antagonisticratios of therapeutic agents, preferably antitumor drugs, usingdelivery vehicle compositions that encapsulate two or more agents,wherein the agents are present in the vehicles at ratiossynergistic or additive (i.e. non-antagonistic) over a range ofconcentrations. Prior to encapsulation, the ratios of therapeuticagents in the combination are selected so that the combinationexhibits synergy or additivity over a desired concentration range.Encapsulation in delivery vehicles allows two or more agents to bedelivered to the disease site in a coordinated fashion, therebyassuring that the agents will be present at the disease site at anon-antagonistic ratio. This result will be achieved whether theagents are co-encapsulated in delivery vehicles, or are separatelyencapsulated in delivery vehicles administered such thatnon-antagonistic ratios are maintained at the disease site. Thepharmacokinetics (PK) of the composition are controlled by thedelivery vehicles themselves such that coordinated delivery isachieved (provided that the PK of the delivery systems arecomparable).

Thus, in one aspect, the invention provides a delivery vehiclecomposition for parenteral administration comprising two or moreagents encapsulated in the vehicle composition at a ratio that issynergistic or additive over a desired concentration range. Thedelivery vehicle composition is prepared by a process comprisingencapsulating the agents in the delivery vehicle composition atthese ratios. The non-antagonistic ratio of the agents isdetermined by assessing the biological activity or effects of theagents on relevant cell culture or cell-free systems over a rangeof concentrations and, in one embodiment, applying an algorithm todetermine a "combination index," (CI). As further described below,using recognized algorithms, a combination index can be calculatedat each concentration level. Ratios are selected where the CIrepresents synergy or additivity over a range of concentrations.Preferably the CI is synergistic over a wide concentration range.Preferred agents are antitumor agents. Any method which results indetermination of a ratio of agents which maintains anon-antagonistic effect over a desired range of concentrations maybe used.

More particularly, the invention relates to a composition whichcomprises delivery vehicles, said delivery vehicles havingencapsulated therein at least a first therapeutic agent and asecond therapeutic agent in a mole ratio of the first agent to thesecond agent which exhibits a non-antagonistic biologic effect torelevant cells in culture or cell-free system over at least 5% ofsuch concentration range where greater than 1% of the cells areaffected (Fraction affected (f.sub.a)>0.01) or to a compositionwhich comprises delivery vehicles, said delivery vehicles havingencapsulated therein at least a first therapeutic agent and asecond therapeutic agent in a mole ratio of the first agent to thesecond agent which exhibits a non-antagonistic cytotoxic effect orcytostatic effect to relevant cells wherein said agents areantineoplastic agents. By "relevant" cells, applicants refer to atleast one cell culture or cell line which is appropriate fortesting the desired biological effect. For example, if the agent isan antineoplastic agent, a "relevant" cell would be a cell lineidentified by the Developmental Therapeutics Program (DTP) of theNational Cancer Institute (NCI)/National Institutes of Health (NIH)as useful in their anticancer drug discovery program. Currently theDTP screen utilizes 60 different human tumor cell lines. Thedesired activity on at least one of such cell lines would need tobe demonstrated.

In another aspect, the invention is directed to a method to delivera synergistic or additive ratio of two or more therapeutic agentsto a desired target by administering the compositions of theinvention. The administration of such compositions need not be inthe form of a single composition, but may also include simultaneousor near simultaneous administration of separate compositionscomprising therapeutic agents in delivery vehicles such that thepharmacokinetics of the delivery vehicles is coordinated--i.e.,designed in such a way that the ratio of therapeutic agentsadministered is maintained when target tissues or organs arereached. Thus, separate compositions, each comprising deliveryvehicles stably associated with one or more therapeutic agents maybe delivered to the subject in a ratio of the therapeutic agentswhich has been determined to be non-antagonistic as describedherein.

In another aspect, the invention is directed to a method to preparea therapeutic composition comprising delivery vehicles, saiddelivery vehicles containing a ratio of at least two therapeuticagents which is non-antagonistic over a range of concentrationswhich method comprises providing a panel of at least twotherapeutic agents wherein the panel comprises at least one, butpreferably a multiplicity of ratios of said agents, testing theability of the members of the panel to exert a biological effect ona relevant cell culture or cell-free system over a range ofconcentrations, selecting a member of the panel wherein the ratioprovides a synergistic or additive effect on said cell culture orcell-free system over a suitable range of concentrations; andencapsulating (i.e., stably associating) the ratio of agentsrepresented by the successful member of the panel into drugdelivery vehicles. The ratio resulting from the determinationdescribed above, in addition to being used as a guide for preparinga single formulation, may also be used to determine the relativeamounts to be administered to a subject of separate compositions,each comprising delivery vehicles stably associated with at leastone therapeutic agent. Thus, the ratios of therapeutic agentsherein determined to be additive or synergistic may be supplied tothe subject in a single composition or in the correct proportion ofseparately prepared compositions.

In another aspect, the invention is directed to kits said kitscomprising, in separate containers, a first composition comprisinga first therapeutic agent stably associated with delivery vehiclesand a second composition comprising delivery vehicles stablyassociated with the second therapeutic agent. The two containersmay be calibrated so that the correct proportion of the twocompositions is administered; alternatively, or in addition the kitmay simply include instructions with regard to the correctratio.

As further described below, in a preferred embodiment, in designingan appropriate combination in accordance with the method describedabove, the non-antagonistic ratios are selected as those that havea combination index (CI) of .ltoreq.1.1 over a range of at least 5%of those doses or concentrations that affect greater than 1% ormore of the cells (f.sub.a>0.01), preferably between 20 and 80%of the cells (f.sub.a=0.2 to 0.8), as defined by relevant cellculture or cell-free assay systems.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a diagram outlining the method of the invention fordetermining an appropriate ratio of therapeutic agents to includein formulations.

FIGS. 2 (A-E) illustrates 5 methods for presenting combination andsynergy data.

FIG. 3A is a graph of combination index (CI) for irinotecan:5-FU atmole ratios of 1:10 (filled squares) and 1:1 (filled circles) as afunction of the fraction of HT29 cells affected (f.sub.a).

FIG. 3B is a graph of CI for etoposide:carboplatin at mole ratiosof 1:10 (filled diamonds) and 10:1 (filled squares) as a functionof the fraction of MCF-7 cells affected (f.sub.a).

FIG. 4 is a graph of the CI for cisplatin:edelfosine at mole ratiosof 10:1 (filled triangles) and 1:1 (filled circles) as a functionof the fraction of H460 cells affected (f.sub.a).

FIG. 5A is a graph of the CI maximum as a function ofcarboplatin:daunorubicin at 10:1, 1:1 and 1:10 mole ratios in H460cells. The inset is a histogram of the CI forcarboplatin:daunorubicin at mole ratios of 10:1 and 1:1 atEffective Dose (ED) values of 50, 75 and 90 in MCF-7 cells.

FIG. 5B is a graph of the CI for carboplatin:daunorubicin at moleratios of 1:10 (filled triangles), 1:1 (filled squares) and 10:1(filled circles) as a function of the fraction of H460 cellsaffected (f.sub.a). The inset is a histogram of the CI forcarboplatin:daunorubicin at mole ratios of 1:10, 1:1 and 10:1 at EDvalues of 50, 75 and 90 in H460 cells.

FIG. 6 is a graph of the carboplatin (open circles) anddaunorubicin (filled circles) concentrations in plasma (nmoles/mL)as a function of time after intravenous administration when thedrugs are formulated in a single liposome (DSPC/DSPG, 80:20 mol %)at a non-antagonistic ratio (10:1).

FIG. 7A is a graph of the carboplatin:daunorubicin mole ratio as afunction of time after intravenous administration at threedifferent ratios when the drugs are formulated in a single liposome(DSPC/DSPG, 80:20 mol %) at 10:1 (filled circles), 5:1 (opencircles) and 1:1 (filled triangles).

FIG. 7B is a graph of the 1:1 carboplatin:daunorubicin data in FIG.7A re-plotted as a function of time after intravenousadministration.

FIG. 8 is a graph of carboplatin (filled circles) and daunorubicin(open circles) concentrations in plasma (nmoles/mL) as a functionof time after intravenous administration when the drugs areformulated at a non-antagonistic mole ratio (10:1) in a singleliposome (DSPC/sphingomyelin/DSPE-PEG2000, 90:5:5 mol %).

FIG. 9 is a graph comparing the activity of a cocktail ofcarboplatin and daunorubicin (filled inverted triangles),carboplatin and daunorubicin formulated in a single liposome (openinverted triangles) or saline control (filled circles) given tomice bearing the human H460 non-small cell lung tumor. Carboplatinand daunorubicin were formulated in DSPC/DSPG (80:20 mol %)liposomes at a 1:1 mole ratio. The arrows indicate the days atwhich the doses were administered.

FIG. 10 is a graph comparing the activity of a cocktail ofcarboplatin and daunorubicin (filled triangles), carboplatin anddaunorubicin formulated in a single liposome (open triangles) orsaline control (filled circles) given to mice bearing the humanH460 non-small cell lung tumor. Carboplatin and daunorubicin wereformulated in DSPC/SM/DSPE-PEG2000 (90:5:5 mol %) liposomes at a10:1 mole ratio. The arrows along the x-axis indicate the dosingschedule.

FIG. 11A is a graph of the CI for cisplatin:daunorubicin at moleratios of 1:1 (filled squares) and 10:1 (filled circles) as afunction of the fraction of H460 cells affected (f.sub.a).

FIG. 11B is a graph of the CI maximum as a function of thecisplatin:daunorubicin at 10:1, 1:1 and 1:10 mole ratios againstH460 cells.

FIG. 12 is a graph of cisplatin (open circles) and daunorubicin(closed circles) concentrations in plasma (.mu.moles/mL) as afunction of time after intravenous administration when the drugsare formulated at a non-antagonistic mole ratio (10:1) in a singleliposome (DMPC/Chol, 55:45 mol %).

FIG. 13 is a graph of cisplatin (closed circles) and daunorubicin(open circles) concentrations in the plasma (.mu.moles/mL) as afunction of time after intravenous administration when the drugsare formulated at a non-antagonistic mole ratio (10:1) in twoseparate liposomes (DMPC/Chol, 55:45 mol % for cisplatin andDSPC/DSPE-PEG2000, 95:5 mol % for daunorubicin).

FIG. 14 is a graph comparing the activity of a cocktail ofcisplatin and daunorubicin (filled inverted triangles), cisplatinand daunorubicin formulated in separate liposomes (open invertedtriangles) or saline control (filled circles) given to mice bearingthe human H460 non-small cell lung tumor. Cisplatin was formulatedin DMPC/Chol (55:45 mol %) liposomes and daunorubicin wasformulated in DSPC/DSPE-PEG2000 (95:5 mol %) liposomes andadministered at a non-antagonistic mole ratio (10:1). Arrowsindicate the days on which the doses were administered.

FIG. 15 is a graph showing concentrations of cisplatin (closedcircles) and daunorubicin (open circles) remaining in the plasma(nmoles/mL) at various times after intravenous administration whenthe drugs were formulated in a single liposome (DMPC/Chol, 55:45mol %) at an antagonistic 1:1 mole ratio. The inset shows thecisplatin:daunorubicin mole ratio at various time points afteradministration.

FIG. 16 is a graph comparing the activity of a cocktail ofcisplatin and daunorubicin (filled triangles), cisplatin anddaunorubicin formulated in a single liposome (open triangles) orsaline control (filled circles) given to mice bearing the humanH460 non-small cell lung tumor. The drugs were formulated inDMPC/Chol (55:45 mol %) liposomes at an antagonistic mole ratio(1:1). Arrows indicate the days on which the doses wereadministered.

FIG. 17A is a graph of the CI for cisplatin:topotecan at moleratios of 1:1 (filled circles) and 10:1 (open circles) as afunction of the fraction of H460 cells affected (f.sub.a).

FIG. 17B is a graph of the CI maximum as a function of thecisplatin:topotecan mole ratio against H460 cells.

FIG. 18 is a graph showing concentrations of cisplatin (closedcircles) and topotecan (open circles) remaining in the plasma(.mu.moles/mL) at various times after intravenous administrationwhen the drugs are formulated in separate liposomes (DMPC/Chol,55:45 mol % for cisplatin and DSPC/Chol, 55:45 mol % fortopotecan). The inset shows the cisplatin to topotecan mole ratioat various time points after administration.

FIG. 19 is a graph comparing the activity of a cocktail ofcisplatin and topotecan (filled triangles), cisplatin and topotecanformulated in separate liposomes (open triangles) or saline control(filled circles) given to mice bearing the human H460 non-smallcell lung tumor. Cisplatin was formulated in DMPC/Chol (55:45 mol%) liposomes and topotecan was formulated in DSPC/Chol (55:45 mol%) liposomes and were administered at a non-antagonistic mole ratio(10:1). Arrows indicate the days on which the doses wereadministered.

FIG. 20A is a graph of the CI for cisplatin:irinotecan at moleratios of 1:1 (squares), 10:1 (circles), 1:5 (triangles) and 1:10(diamonds) as a function of the fraction of H460 cells affected(f.sub.a).

FIG. 20B is a graph of the CI maximum as a function of thecisplatin:irinotecan mole ratio against H460 cells.

FIG. 21 is a graph showing the concentrations of cisplatin (filledcircles) and irinotecan (open circles) remaining in the plasma(nmoles/mL) at various time points after intravenous administrationwhen the drugs were co-loaded into a single liposome (DSPC/DSPG,80:20 mol %).

FIG. 22 is a graph showing the concentrations of cisplatin (closedcircles) and irinotecan (open circles) remaining in the plasma(nmoles/mL) at various time points after intravenous administrationwhen the drugs are formulated in separate liposomes (DMPC/Chol,55:45 mol % for cisplatin and DSPC/DSPE-PEG2000, 95:5 mol % foririnotecan).

FIG. 23 is a graph comparing the activity of a cocktail ofcisplatin and irinotecan (filled squares), cisplatin and irinotecanformulated in separate liposomes and administered at differentdoses (open symbols) or saline control (filled circles) given tomice bearing the human H460 non-small cell lung tumor. Cisplatinformulated in DMPC/Chol (55:45 mol %) liposomes and irinotecanformulated in DSPC/DSPE-PEG2000 (95:5 mol %) liposomes wereadministered at a non-antagonistic mole ratio (1:5). Arrowsindicate the days on which the doses were administered.

FIG. 24 is a graph of CI for vinorelbine in combination with POPS(inverted triangles), DPPS (upward triangles), DLPS (circles), DSPS(diamonds) or DOPS (squares) as a function of the H460 cellsaffected (f.sub.a) at vinorelbine:PS mole ratios of 1:1.

FIG. 25A is a graph of the vinorelbine concentration in plasma as afunction of time after intravenous administration to SCID/rag2 miceof free vinorelbine (filled circles) or encapsulated inSM/Chol/DPPS/DSPE-PEG2000, 35:45:10:10 mol % liposomes (opencircles) at a vinorelbine:PS mole ratio of 1:1.

FIG. 25B is a histogram showing plasma concentration area under thecurve (AUC) for free vinorelbine (black bar) or encapsulated inSM/Chol/DPPS/DSPE-PEG2000, 35:45:10:10 mol % (grey bar) afterintravenous administration to SCID/rag2 mice, using the data ofFIG. 25A.

FIG. 26 is a graph comparing the activity of free vinorelbine (opencircles), vinorelbine encapsulated in DSPC/Chol/DPPS/DSPE-PEG2000,35:45:10:10 mol % liposomes (filled inverted triangles),vinorelbine encapsulated in SM/Chol/DPPS/DSPE-PEG2000, 35:45:10:10mol % liposomes (open triangles) or saline control (filled circles)given to mice bearing the H460 non-small cell lung tumor.Vinorelbine and phosphatidylserine (DPPS) were formulated at anon-antagonistic mole ratio (1:1). Arrows indicate the days onwhich the doses were administered.

FIG. 27 shows the effect of saline control (filled circles); freevinorelbine (open circles); vinorelbine encapsulated in:SM/Chol/DPPS/DSPE-PEG2000, 35:45:10:10 (filled inverted triangles),DAPC/Chol/DPPS/DSPE-PEG2000, 35:45:10:10 mol % (open triangles),and DSPC/Chol/DSPS/DSPE-PEG2000, 35:45:10:10 mol % (filled squares)liposomes given to mice bearing the H460 non-small cell lung tumor.Vinorelbine and phosphatidylserine (DPPS or DSPS) were formulatedat a non-antagonistic mole ratio (1:1). Arrows indicate the days onwhich the doses were administered.

FIG. 28 shows the effect of saline control (open triangles); freevinorelbine (filled circles); and vinorelbine encapsulated inSM/Chol/DPPS/DSPE-PEG2000, 35:45:10:10 mol % liposomes (filledinverted triangles) on percent survival of P388 murine leukemiabearing mice. Vinorelbine and phosphatidylserine were formulated ata non-antagonistic mole ratio (1:1). The arrow along the x-axisindicate the day on which the doses were administered.

FIG. 29 shows CI plotted as a function of the fraction of HT-29cells affected by combinations of FUDR:CPT-11 at various ratios:10:1 (solid squares); 5:1 (solid circles); 1:1 (solid triangles);1:5 (solid inverted triangles); and 1:10 (open circles).

FIG. 30 is a graph of plasma concentration levels of FUDR (solidcircles) and CPT-11 (open circles) as a function of time afterintravenous administration.

FIG. 31 is a graph of tumor volume versus time after tumor cellinoculation for saline controls (solid circles) injection of acocktail of CPT-11/FUDR (open inverted triangles) and the liposomalformulation of CPT-11/FUDR (solid inverted triangles).

MODES OF CARRYING OUT THE INVENTION.sup.1

The method of the invention involves determining a ratio oftherapeutic drugs which is non-antagonistic over a desiredconcentration range in vitro and supplying this non-antagonisticratio in a manner that will ensure that the ratio is maintained atthe site of desired activity. The synergistic or additive ratio isdetermined by applying standard analytical tools to the resultsobtained when at least one ratio of two or more therapeutic agentsis tested in vitro over a range of concentrations against relevantcell cultures or cell-free systems. By way of illustration,individual agents and various combinations thereof are tested fortheir biological effect on cell culture or a cell-free system, forexample causing cell death or inhibiting cell growth, at variousconcentration levels. The concentration levels of the preset ratiosare plotted against the percentage cell survival to obtain acorrelation which can be manipulated by known and establishedmathematical techniques to calculate a "combination index" (CI).The mathematics are such that a CI of 1 (i.e., 0.9-1.1) describesan additive effect of the drugs; a CI>1 (i.e., >1.1)represents an antagonist effect; and a CI of <1 (i.e., <0.9)represents a synergistic effect. .sup.1AbbreviationsThe followingabbreviations are used:PE: phosphatidylethanolamine; PS:phosphatidylserine; DPPS: dipalmitoylphosphatidylserine; DSPS:distearoylphosphatidylserine DLPS: dilauroylphosphatidylserine;DOPS: dioleoylphosphatidylserine; POPS:palmitoyloleoylphosphatidylserine; PC: phosphatidylcholine; SM:sphingomyelin; PG: phosphatidylglycerol; PI: phosphatidylinositol;PA: phosphatidic acid; DSPC: distearoylphosphatidylcholine; DMPC:dimyristoylphosphatidylcholine; DSPG:distearoylphosphatidylglycerol; DSPE:distearoylphosphatidylethanolamine; Chol: cholesterol; CH or CHE:cholesteryl hexadecyl ether;PEG: polyethylene glycol; DSPE-PEG:distearoylphosphatidylethanolamine-N-[polyethylene glycol]; whenPEG is followed by a number, the number is the molecular weight ofPEG in Daltons; DSPE-PEG2000:distearoylphosphatidylethanolamine-N-[polyethylene glycol2000];SUV: small unilamellar vesicle; LUV: large unilamellarvesicle; MLV: multilamellar vesicle;MTT:3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H tetrazolium bromide;DMSO: dimethylsulfoxide; OD: optical density; OGP: N-octylbeta-D-glucopyranoside; EDTA: ethylenediaminetetraacetic acid;HEPES: N-[2-hydroxylethyl]-piperazine-N-[2-ethanesulfonic acid];HBS: HEPES buffered saline (20 mM HEPES, 150 mM NaCl, pH 7.4); SHE:300 mM sucrose, 20 mM HEPES, 30 mM EDTA; ED50, ED75 and ED90:effective dose required to affect 50, 75 and 90% of the cells inculture; LD50: dose required to cause 50% lethality of the cells inculture; CI: combination index; CI max or CI maximum: CI valuetaken for a single L value (between 0.2 and 0.8) where the greatestdifference in CI values for the drugs at different ratios isobserved; f.sub.a: fraction affected; TEA: triethanolamine;FDA:United States Food and Drug Administration; NCI: National CancerInstitute.

One general approach is shown in FIG. 1. As shown, agents A and Bare tested individually and together at two different ratios fortheir ability to cause cell death or cell stasis as assessed by theMTT assay described below. Initially, correlations between theconcentration of drugs A, B, and the two different combinationratios (Y:Z and X:Y) are plotted against cytotoxicity, calculatedas a percentage based on the survival of untreated control cells.As expected, there is a dose-dependent effect on cell survival bothfor the individual drugs and for the combinations. Once thiscorrelation has been established, the cell survival or fractionaffected (f.sub.a) can be used as a surrogate for concentration incalculating the CI.

The results of the CI calculation are also shown in FIG. 1; thisindex is calculated as a function of the fraction of cells affectedaccording to the procedure of Chou and Talalay, Advance Enz. Regul.(1985) 22:27-55. In this hypothetical situation, the first ratio(X:Y) of drugs A plus B is non-antagonistic at all concentrationswhile the combination in the second ratio (Y:Z) is antagonistic.Thus, it is possible to provide a ratio of drugs A plus B (ratio 1)which will be non-antagonistic regardless of concentration over awide range. It is this ratio that is desirable to include in thecompositions of the invention.

The present inventors have also devised an alternative illustrationof the effect of ratio and concentration on synergy by calculatinga "CI maximum" for various ratios of combinations of agents. The"CI maximum" is defined as the CI value taken for a single f.sub.avalue (between 0.2 and 0.8) where the greatest difference in CIvalues for the drugs at different ratios was observed. This isillustrated in FIGS. 2A and 2B; as shown, when theirinotecan/carboplatin ratio is 1:10, its CI differs most from thatof the remaining ratios where the fraction affected value is 0.2.The CI value for this ratio at f.sub.a 0.2 is, as shown,approximately 2.0.

While the determination in vitro of non-antagonistic ratios hasbeen illustrated for a combination of only two drugs, applicationof the same techniques to combinations of three or more drugsprovides a CI value over the concentration range in a similarmanner

The ratio obtained in this way is maintained in the pharmaceuticalcomposition by encapsulating the agents in the predetermined ratioin liposomes or other particulate forms which assures that thenon-antagonistic ratio will be maintained. The compositions, thus,contain delivery vehicles which are particulate in nature andcontain the desired ratio of therapeutic agents.

While it is preferred to co-encapsulate the agents so that both arecontained in the same delivery vehicle, this is not necessary.Since particulate carriers can share similar pharmacokinetics, theactive substances experience coordinated delivery from theformulation even if encapsulated separately.

By "encapsulation", it is meant stable association with thedelivery vehicle. Thus, it is not necessary for the vehicle tosurround the agent or agents as long as the agent or agents is/arestably associated with the vehicles when administered in vivo.Thus, "stably associated with" and "encapsulated in" or"encapsulated with" or "co-encapsulated in or with" are intended tobe synonymous terms. They are used interchangeably in thisspecification. The stable association may be effected by a varietyof means, including covalent bonding to the delivery vehicle,preferably with a cleavable linkage, noncovalent bonding, andtrapping the agent in the interior of the delivery vehicle and thelike. The association must be sufficiently stable so that theagents remain associated with the delivery vehicle at anon-antagonistic ratio until it is delivered to the target site inthe treated subject.

Delivery vehicles may include lipid carriers, liposomes, lipidmicelles, lipoprotein micelles, lipid-stabilized emulsions,cyclodextrins, polymer nanoparticles, polymer microparticles, blockcopolymer micelles, polymer-lipid hybrid systems, derivatizedsingle chain polymers, and the like. Liposomes can be prepared asdescribed in Liposomes: Rational Design (A. S. Janoff ed., MarcelDekker, Inc., N.Y.), or by additional techniques known to thoseknowledgeable in the art. Liposomes for use in this invention maybe prepared to be of "low-cholesterol." Such liposomes are"cholesterol free," or contain "substantially no cholesterol," or"essentially no cholesterol." The term "cholesterol free" as usedherein with reference to a liposome means that a liposome isprepared in the absence of cholesterol. The term "substantially nocholesterol" allows for the presence of an amount of cholesterolthat is insufficient to significantly alter the phase transitioncharacteristics of the liposome (typically less than 20 mol %cholesterol). The incorporation of less than 20 mol % cholesterolin liposomes can allow for retention of drugs not optimallyretained when liposomes are prepared with greater than 20 mol %cholesterol. Additionally, liposomes prepared with less than 20 mol% cholesterol display narrow phase transition temperatures, aproperty that may be exploited for the preparation of liposomesthat release encapsulated agents due to the application of heat(thermosensitive liposomes). Liposomes of the invention may alsocontain therapeutic lipids, which include ether lipids,phosphatidic acid, phosphonates, ceramide and ceramide analogues,sphingosine and sphingosine analogues and serine-containing lipids.Liposomes may also be prepared with surface stabilizing hydrophilicpolymer-lipid conjugates such as polyethylene glycol-DSPE, toenhance circulation longevity. The incorporation of negativelycharged lipids such as phosphatidylglycerol (PG) andphosphatidylinositol (PI) may also be added to liposomeformulations to increase the circulation longevity of the carrier.These lipids may be employed to replace hydrophilic polymer-lipidconjugates as surface stabilizing agents. Embodiments of thisinvention may make use of cholesterol-free liposomes containing PGor PI to prevent aggregation thereby increasing the blood residencetime of the carrier.

Micelles are self-assembling particles composed of amphipathiclipids or polymeric components that are utilized for the deliveryof sparingly soluble agents present in the hydrophobic core.Various means for the preparation of micellar delivery vehicles areavailable and may be carried out with ease by one skilled in theart. For instance, lipid micelles may be prepared as described inPerkins, et al., Int. J. Pharm. (2000) 200(1):27-39 (incorporatedherein by reference). Lipoprotein micelles can be prepared fromnatural or artificial lipoproteins including low and high-densitylipoproteins and chylomicrons. Lipid-stabilized emulsions aremicelles prepared such that they comprise an oil filled corestabilized by an emulsifying component such as a monolayer orbilayer of lipids. The core may comprise fatty acid esters such astriacylglycerol (corn oil). The monolayer or bilayer may comprise ahydrophilic polymer lipid conjugate such as DSPE-PEG. Thesedelivery vehicles may be prepared by homogenization of the oil inthe presence of the polymer lipid conjugate. Agents that areincorporated into lipid-stabilized emulsions are generally poorlywater-soluble. Synthetic polymer analogues that display propertiessimilar to lipoproteins such as micelles of stearic acid esters orpoly(ethylene oxide) block-poly(hydroxyethyl-L-aspartamide) andpoly(ethylene oxide)-block-poly(hydroxyhexyl-L-aspartamide) mayalso be used in the practice of this invention (Lavasanifar, etal., J. Biomed. Mater. Res. (2000) 52:831-835).

Cyclodextrins comprise cavity-forming, water-soluble,oligosaccharides that can accommodate water-insoluble drugs intheir cavities. Agents can be encapsulated into cyclodextrins usingprocedures known to those skilled in the art. For example, seeAtwood, et al., Eds., "Inclusion Compounds," Vols. 2 & 3,Academic Press, NY (1984); Bender, et al., "CyclodextrinChemistry," Springer-Verlag, Berlin (1978); Szeitli, et al.,"Cyclodextrins and Their Inclusion Complexes," Akademiai Kiado,Budapest, Hungary (1982) and WO 00/40962.

Nanoparticles and microparticles may comprise a concentrated coreof drug that is surrounded by a polymeric shell (nanocapsules) oras a solid or a liquid dispersed throughout a polymer matrix(nanospheres). General methods of preparing nanoparticles andmicroparticles are described by Soppimath, et al. (J. ControlRelease (2001) 70(1-2):1-20) the reference of which is incorporatedherein. Other polymeric delivery vehicles that may be used includeblock copolymer micelles that comprise a drug containing ahydrophobic core surrounded by a hydrophilic shell; they aregenerally utilized as carriers for hydrophobic drugs and can beprepared as found in Allen, et al., Colloids and Surfaces B:Biointerfaces (1999) November 16(1-4):3-27. Polymer-lipid hybridsystems consist of a polymer nanoparticle surrounded by a lipidmonolayer. The polymer particle serves as a cargo space for theincorporation of hydrophobic drugs while the lipid monolayerprovides a stabilizing interference between the hydrophobic coreand the external aqueous environment. Polymers such aspolycaprolactone and poly(d,l-lactide) may be used while the lipidmonolayer is typically composed of a mixture of lipid. Suitablemethods of preparation are similar to those referenced above forpolymer nanoparticles. Derivatized single chain polymers arepolymers adapted for covalent linkage of a biologically activeagent to form a polymer-drug conjugate. Numerous polymers have beenproposed for synthesis of polymer-drug conjugates includingpolyaminoacids, polysaccharides such as dextrin or dextran, andsynthetic polymers such as N-(2-hydroxypropyl)methacrylamide (HPMA)copolymer. Suitable methods of preparation are detailed in Veroneseand Morpurgo, IL Farmaco (1999) 54(8):497-516 and are incorporatedby reference herein.

Delivery vehicles are thus provided such that consistent deliveryof the administered ratio of the therapeutic components isaccomplished. Thus, the ratio may be maintained by simpleco-encapsulation of the agents in the vehicles that comprise thecomposition or the agents can be encapsulated in separate vehiclesif the vehicles control the pharmacokinetics of the composition tomaintain non-antagonistic drug ratios in the same manner

Preferably, the compositions of the invention are used to delivercompositions of antitumor agents that are not antagonistic. Thefollowing detailed description sets forth the manner in which theratios of therapeutic agents are determined and methods forencapsulating the desired ratios into the delivery systems of theinvention.

Briefly, in one scenario, first, individual agents are screenedseparately in a variety of in vitro or in vivo assays to determinetheir individual activities. Then, pairs of agents are combined andassayed in the same screening method. In this initial screen, theratios of the agents are the mole ratios of the concentrationshaving 50% activity (IC.sub.50 value) identified previously.Alternatively, other fixed ratios (typically mole ratios of 1:10,1:1 and 10:1) are chosen based on considerations for formulationpurposes. The mean values, calculated based on agent effects oncell survival, and drug doses are entered into the CalcuSyncomputer program and the output data is evaluated to define aCombination Index (CI) value as a function of the fraction of cellsaffected (f.sub.a).

The CalcuSyn method has been successfully applied to test variousagents such as antitumor drugs, immunosuppressants for organtransplant, combined purging of leukemic cells for autologous bonemarrow transplantation, insecticides, biological responsemodifiers, multiple drug resistance inhibitors, anti-microbialagents, anti-HIV agents, anti-herpetic and other anti-viralagents.

Combinations of agents displaying interaction behavior similar tothat of cisplatin:daunorubicin at a mole ratio of 1:1 in FIG. 11A,i.e., are antagonistic, and are not pursued. Combinations ofcompounds having non-antagonistic interactions over substantialranges (preferably at least about 20%) of f.sub.a values greaterthan fa>0.01 (i.e., irinotecan:carboplatin at mole ratios of 1:1and 10:1; FIG. 2A) are re-evaluated in this in vitro screeningassay at a variety of different drug/drug ratios to define theoptimum ratio(s) to enhance both the strength of thenon-antagonistic interaction (i.e., lower CI values) and increasethe f.sub.a range over which synergy is observed.

Optimized non-antagonistic drug combinations thus identified definea composition for formulation in a delivery vehicle as a dual-agentcomposition and/or can be used as a single pharmaceutical unit todetermine synergistic or additive interactions with a thirdagent.

In Vitro Determination of Non-Antagonistic Ratios

In order to prepare the compositions of the invention, the desiredratio of agents contained in the delivery vehicles must first bedetermined. Desirably, the ratio will be that wherein synergy oradditivity is exhibited by the combination over a range ofconcentrations. Such ratios can be determined in vitro in cellcultures or cell-free systems using various mathematicalmodels.

Determination of ratios of agents that display synergistic oradditive combination effects over concentration ranges may becarried out using various algorithms, based on the types ofexperimental data described below. These methods includeisobologram methods (Loewe, et al., Arzneim-Forsch (1953)3:285-290; Steel, et al., Int. J. Radiol. Oncol. Biol. Phys. (1979)5:27-55), the fractional product method (Webb, Enzyme and MetabolicInhibitors (1963) Vol. 1, pp. 1-5. New York: Academic Press), theMonte Carlo simulation method, CombiTool, ComboStat and theChou-Talalay median-effect method based on an equation described inChou, J. Theor. Biol. (1976) 39:253-76; and Chou, Mol. Pharmacol.(1974) 10:235-247). Alternatives include surviving fraction (Zoli,et al., Int. J. Cancer (1999) 80:413-416), percentage response togranulocyte/macrophage-colony forming unit compared with controls(Pannacciulli, et al., Anticancer Res. (1999) 19:409-412) andothers (Berenbaum, Pharmacol. Rev. (1989) 41:93-141; Greco, et al.,Pharmacol Rev. (1995) 47:331-385).

The Chou-Talalay median-effect method is preferred. The analysisutilizes an equation wherein the dose that causes a particulareffect, f.sub.a, is given by:D=D.sub.m[f.sub.a/(1-f.sub.a)].sup.1/m in which D is the dose ofthe drug used, f.sub.a is the fraction of cells affected by thatdose, D.sub.m is the dose for median effect signifying the potencyand m is a coefficient representing the shape of the dose-effectcurve (m is 1 for first order reactions).

This equation can be further manipulated to calculate a combinationindex (CI) on the basis of the multiple drug effect equation asdescribed by Chou and Talalay, Adv. Enzyme Reg. (1984) 22:27-55;and by Chou, et al., in: Synergism and Antagonism in Chemotherapy,Chou and Rideout, eds., Academic Press: New York 1991:223-244. Acomputer program for this calculation (CalcuSyn) is found in Chou,Dose-effect analysis with microcomputers: quantitation of ED50,LD50, synergism, antagonism, low-dose risk, receptor ligand bindingand enzyme kinetics (CalcuSyn Manual and Software; Cambridge:Biosoft 1987).

The combination index equation is based on the multiple drug-effectequation of Chou-Talalay derived from enzyme kinetic models. Anequation determines only the additive effect rather than synergismand antagonism. However, according to the CalcuSyn program,synergism is defined as a more than expected additive effect, andantagonism as a less than expected additive effect. Chou andTalalay in 1983 proposed the designation of CI=1 as the additiveeffect, thus from the multiple drug effect equation of two drugs,we obtain: CI=(D).sub.1/(D.sub.x).sub.1+(D).sub.2/(D.sub.x).sub.2[Eq. 1] for mutually exclusive drugs that have the same or similarmodes of action, and it is further proposed thatCI=(D).sub.1/(D.sub.x).sub.1+(D).sub.2/(D.sub.x).sub.2+(D.sub.1)(D.sub.2)-/(D.sub.x).sub.1(D.sub.x).sub.2 [Eq. 2] for mutually non-exclusivedrugs that have totally independent modes of action. CI<1,=1,and >1 indicates synergism, additive effect, and antagonism,respectively. Equation 1 or equation 2 dictates that drug 1,(D).sub.1, and drug 2, (D).sub.2, (in the numerators) incombination inhibit x % in the actual experiment. Thus, theexperimentally observed x % inhibition may not be a round numberbut most frequently has a decimal fraction. (D.sub.x).sub.1 and(D.sub.x).sub.2 (in the denominators) of equations 1 and 2 are thedoses of drug 1 and drug 2 alone, respectively, inhibiting x %.

For simplicity, mutual exclusivity is usually assumed when morethan two drugs are involved in combinations (CalcuSyn Manual andSoftware; Cambridge: Biosoft 1987).

The underlying experimental data are generally determined in vitrousing cells in culture or cell-free systems. Preferably, thecombination index (CI) is plotted as a function of the fraction ofcells affected (f.sub.a) as shown in FIG. 1 which, as explainedabove, is a surrogate parameter for concentration range. Preferredcombinations of agents are those that display synergy or additivityover a substantial range of f.sub.a values. Combinations of agentsare selected that display synergy over at least 5% of theconcentration range wherein greater than 1% of the cells areaffected, i.e., an f.sub.a range greater than 0.01. Preferably, alarger portion of overall concentration exhibits a favorable CI;for example, 5% of an f.sub.a range of 0.2-0.8. More preferably 10%of this range exhibits a favorable CI. Even more preferably, 20% ofthe f.sub.a range, preferably over 50% and most preferably over atleast 70% of the f.sub.a range of 0.2 to 0.8 are utilized in thecompositions. Combinations that display synergy over a substantialrange of f.sub.a values may be re-evaluated at a variety of agentratios to define the optimal ratio to enhance the strength of thenon-antagonistic interaction and increase the f.sub.a range overwhich synergy is observed.

While it would be desirable to have synergy over the entire rangeof concentrations over which cells are affected, it has beenobserved that in many instances, the results are considerably morereliable in an f.sub.a range of 0.2-0.8. Thus, although the synergyexhibited by combinations of the invention is set forth to existwithin the broad range of 0.01 or greater, it is preferable thatthe synergy be established in the f.sub.a range of 0.2-0.8.

The optimal combination ratio may be further used as a singlepharmaceutical unit to determine synergistic or additiveinteractions with a third agent. In addition, a three-agentcombination may be used as a unit to determine non-antagonisticinteractions with a fourth agent, and so on.

As set forth above, the in vitro studies on cell cultures will beconducted with "relevant" cells. The choice of cells will depend onthe intended therapeutic use of the agent. Only one relevant cellline or cell culture type need exhibit the requirednon-antagonistic effect in order to provide a basis for thecompositions to come within the scope of the invention.

For example, in one preferred embodiment of the invention, thecombination of agents is intended for anticancer therapy.Appropriate choices will then be made of the cells to be tested andthe nature of the test. In particular, tumor cell lines aresuitable subjects and measurement of cell death or cell stasis isan appropriate end point. As will further be discussed below, inthe context of attempting to find suitable non-antagonisticcombinations for other indications, other target cells and criteriaother than cytotoxicity or cell stasis could be employed.

For determinations involving antitumor agents, cell lines may beobtained from standard cell line repositories (NCI or ATCC forexample), from academic institutions or other organizationsincluding commercial sources. Preferred cell lines would includeone or more selected from cell lines identified by theDevelopmental Therapeutics Program of the NCI/NIH. The tumor cellline screen used by this program currently identifies 60 differenttumor cell lines representing leukemia, melanoma, and cancers ofthe lung, colon, brain, ovary, breast, prostate and kidney. Therequired non-antagonistic effect over a desired concentration rangeneed be shown only on a single cell type; however, it is preferredthat at least two cell lines exhibit this effect, more preferablythree cell lines, more preferably five cell lines, and morepreferably 10 cell lines. The cell lines may be established tumorcell lines or primary cultures obtained from patient samples. Thecell lines may be from any species but the preferred source will bemammalian and in particular human. The cell lines may begenetically altered by selection under various laboratoryconditions, and/or by the addition or deletion of exogenous geneticmaterial. Cell lines may be transfected by any gene-transfertechnique, including but not limited to, viral or plasmid-basedtransfection methods. The modifications may include the transfer ofcDNA encoding the expression of a specific protein or peptide, aregulatory element such as a promoter or enhancer sequence orantisense DNA or RNA. Genetically engineered tissue culture celllines may include lines with and without tumor suppressor genes,that is, genes such as p53, pTEN and p16; and lines created throughthe use of dominant negative methods, gene insertion methods andother selection methods. Preferred tissue culture cell lines thatmay be used to quantify cell viability, e.g., to test antitumoragents, include, but are not limited to, H460, MCF-7, SF-268, HT29,HCT-116, LS180, B16-F10, A549, Capan pancreatic, CAOV-3, IGROV1,PC-3, MX-1 and MDA-MB-231.

In one preferred embodiment, the given effect (f.sub.a) refers tocell death or cell stasis after application of a cytotoxic agent toa cell culture. Cell death or viability may be measured, forexample, using the following methods:

TABLE-US-00003 CYTOTOXICITY ASSAY REFERENCE MTT assay Mosmann, J.Immunol. Methods (1983) 65(1-2): 55-63. Trypan blue dye exclusionBhuyan, et al., Experimental Cell Research (1976) 97: 275-280.Radioactive tritium (.sup.3H)-thymidine Senik, et al., Int. J.Cancer incorporation or DNA intercalating assay (1975) 16(6):946-959. Radioactive chromium-51 release assay Brunner, et al.,Immunology (1968) 14: 181-196. Glutamate pyruvate transaminase,creatine Mitchell, et al., J. of Tissue Culture Methodsphosphokinase and lactate dehydrogenase (1980) 6(3&4): 113-116.enzyme leakage Neutral red uptake Borenfreund and Puerner, Toxicol.Lett. (1985) 39: 119-124. Alkaline phosphatase activity Kyle, etal., J. Toxicol. Environ. Health (1983) 12: 99-117. Propidiumiodide staining Nieminen, et al., Toxicol. Appl. Pharmacol. (1992)115: 147-155. Bis-carboxyethyl-carboxyfluorescein (BCECF) Kolber,et al., J. Immunol. Methods retention (1988) 108: 255-264.Mitochondrial membrane potential Johnson, et al., Proc. Natl. Acad.Sci. USA (1980) 77: 990-994. Clonogenic Assays Puck, et al., J. ofExperimental Medicine (1956) 103: 273-283. LIVE/DEADViability/Cytotoxicity assay Morris, Biotechniques (1990) 8:296-308. Sulforhodamine B (SRB) assays Rubinstein, et al., J. Natl.Cancer Instit. (1990) 82: 1113-1118.

The "MTT assay" is preferred.

Non-antagonistic ratios of two or more agents can be determined fordisease indications other than cancer and this information can beused to prepare therapeutic formulations of two or more drugs forthe treatment of these diseases. With respect to in vitro assays,many measurable endpoints can be selected from which to define drugsynergy, provided those endpoints are therapeutically relevant forthe specific disease.

Thus, for example, one skilled in the art will be able to definenon-antagonistic ratios of two or more agents selected fortreatment of inflammatory disorders by measuring, in vitro,suppression of proinflammatory cytokines such as IL-1, IL-18,COX-2, TNF or interferon-gamma. Other inflammatory signals include,but are not limited to, inhibition of prostaglandin E2 andthromboxane B2. In particular, endotoxin-mediated macrophageactivation provides a suitable in vitro assay for measuring theanti-inflammatory effects of an added agent or combinations ofagents and is commonly used in the art. In such an assay,macrophages grown in large quantities are activated by the additionof an endotoxin, such as lipopolysaccharide. Upon activation,macrophage secretion of cytokines such as IL-1 and TNF ismeasurable as well as activation of COX-2. Candidateanti-inflammatory drugs are added and evaluated based on theirability to suppress IL-1, TNF and COX-2. Titration with1.times.10.sup.-7 M dexamethasone is typically used as a positivecontrol. It will be apparent to those skilled in the art thatassays involving macrophage activation are suitable for wide-spreadscreening of drug combinations and that suppression of IL-1, TNFand COX-2 are suitable endpoints for defining synergy. In additionto measuring inflammatory signals, investigators can consider theuse of in vitro models that measure the effect of two or moreagents on leukocyte functions. Functional tests can involve, butare not limited to, inhibition of degranulation, superoxidegeneration, and leukocyte migration.

Similar to cancer, proliferation is a key event in the developmentof arteriosclerosis, restenosis or other cardiovascular diseaseswith vasculoproliferative attributes. Thus, one skilled in the artcan find non-antagonistic ratios of two or more agents by assessingdrug synergy by the methods set forth herein, applied to relevantproliferating cell populations of blood vessels. In particular,restenosis, such as coronary and peripheral artery restenosis thattypically results following angioplasty, is attributable to smoothmuscle and endothelial cell proliferation (Fuster, Arch Mal CoeurVaiss (1997) 90 Spec No 6:41-47). Using standard methods, set forthherein, one skilled in the art can measure whether two or moreagents act non-antagonistically to inhibit endothelial cell orsmooth muscle cell proliferation. These assays can be undertakenusing immortalized cell lines or, preferably, using primary celllines. These cell lines can be obtained from commercial sources(e.g., Clonetics, California) or from fresh tissue (e.g., umbilicalveins, arteries, brain) and must be maintained in appropriategrowth factors that promote cell proliferation. Similar to assaysmeasuring synergy of two or more agents on cancer cells, suchassays can include, but are not limited to, endpoints of inhibitionof proliferation and migration. Proliferation endpoints can rely onlive/dead assays such as the MTT assay described in thisapplication, measurements of proliferation that rely on[.sup.3H]-thymidine incorporation, or other similar assays. Alsosimilar to dividing cancer cells, proliferation of endothelialcells and smooth muscle cells is regulated by checkpoints in thecell cycle and assays that measure cell cycle inhibition can beused to define non-antagonistic ratios of two or more agentsselected for treatment of vasculoproliferative disorders.

Non-antagonistic combinations of agents may also be identified fortheir activity against microbial or viral infections. As a firststep in identifying antimicrobial agents, the minimum inhibitoryconcentration (MIC) for an agent can be determined by the classicalmicrotitre broth dilution or agar dilution antimicrobial assaysknown to those skilled in the art. These assays are regulated bythe National Committee of Laboratory Safety and Standards (NCLSS).The standard broth dilution assays are published in Amsterdam(1996) Susceptibility testing of Antimicrobials in liquid media in"Antibiotics in Laboratory Medicine", Lorian, V. 4.sup.th Edition,pages 52-111, Williams and Wilkins, Baltimore. The MIC is definedas the lowest concentration of an antibiotic that will inhibit thein vitro growth of an infectious organism. In the above-mentionedassays, the MIC can be determined by plating an inoculum ofmicrobes in a small spot (at, for example, 10.sup.4 colony-formingunits [CFU] per spot) on growth medium (for example, agar) havingdifferent concentrations of the drug. Alternatively, microbes canbe inoculated into a suspension of growth media that containsdifferent concentrations of the drug. In addition, the microbes maybe either treated as above or may be resident as intracellularinfections in a specific cell population (i.e., a macrophage). Inthe latter instance, mammalian cells grown in culture by standardmethods are given intracellular microbial infections by briefexposure to a low concentration of microbes. After a period of timeto allow the intracellular replication of the microbes, the cellsand their intracellular microbes are treated with a drug in thesame manner as described for cytotoxicity tests with mammaliancells. After an appropriate period of time sufficient for the drugto inhibit microbial growth when given at effective concentrations,the bacterial growth can be determined by a variety of meansincluding: (i) determination of the absence or presence (and size,as appropriate) of the inoculum spot; (ii) plating and serialdilution of known volumes of the suspension of treated bacteriaonto agar growth plates to allow calculation of the number ofmicrobes that survived treatment; (iii) macroscopic (by eye)determination; (iv) time-kill curves in which microbes in thelogarithmic phase of growth are suspended into a growth mediacontaining a drug(s) and, at various times after inoculation, knownvolumes are removed and serial diluted onto growth agar forcounting of the surviving microbes; (v) other spectroscopic,analytic, in vitro or in vivo methods known by those skilled in theart to allow the counting of viable microbes. The efficacy of adrug, or combinations of drugs to kill intracellular-residentinfections are typically assessed after the host cells are lysedwith detergents (such as 1% Triton X-100 plus 0.1% sodium dodecylsulfate) to release the microbes, then the lysates are serialdiluted onto agar growth plates for counting of the numbers ofsurviving microbes.

Combinations of effective agents are assessed for theirantagonistic, additive or synergistic activity using the meansdescribed above. Specifically, pairs of compounds are applied tothe bacteria in fixed ratios that can be equimolar, or the ratio ofthe MIC values or other fixed ratios, and the bacteria treated at avariety of concentrations of the pair of compounds. Activity isdetermined as described above. Antagonism, additivity or synergyare determined from a variety of mathematical treatments forexample by isobolograms, CI, and the like.

Extensive screening of agents or combinations of agents withantiviral activity can be performed by a number of in vitro assays,typically plaque reduction and cytopathic effects (CPE) inhibitionassays, which are well known to those of skill in the art. Theseassays are able to directly measure the extent to which anantiviral drug or drugs inhibits the effects of viral infection intissue culture. The plaque reduction assay is preferred for virusand cell line combinations which produce a well-defined plaque.Michaelis, et al., demonstrated the use of plaque reduction assayscombined with the Chou-Talalay method for determiningnon-antagonistic antiviral effects of aphidicolin and itsderivatives on a number of viruses at various mole ratios(Michaelis, et al., Arzneimittelforschung (2002) 52(5):393-399). Ifa well-defined plaque is not producible by particular virus andcell line combinations, CPE inhibition assays are preferred.Additional methods for rapid and convenient identification ofnon-antagonistic combinations of antiviral agents include, but arenot limited to, cell viability, virus yield and HIV acute orchronic infection assays. Cell viability is used to measure anantiviral agent's or combination of agent's ability to increasecell viability and can be achieved using quantitative assays suchas the MTT assay previously described. Alternatively, the virusyield assay and the acute HIV infection assays evaluate an agent'sability to reduce virus yield allowing for direct measurements ofantiviral activity. It will be apparent to those knowledgeable inthe art that the aforementioned assays are suitable for screeningantiviral drug combinations for synergistic, additive orantagonistic effects in vitro and are therefore included within thescope of the invention.

Preferred Agent Combinations

Various combinations of therapeutic agents, having been found tosatisfy the criteria for non-antagonistic effects set forth above,are then provided in the form of formulations of drug deliveryvehicles. A "therapeutic agent" is a compound that alone, or incombination with other compounds, has a desirable effect on asubject affected by an unwanted condition or disease.

Certain therapeutic agents are favored for use in combination whenthe target disease or condition is cancer. Examples are: "Signaltransduction inhibitors" which interfere with or prevents signalsthat cause cancer cells to grow or divide; "Cytotoxic agents";"Cell cycle inhibitors" or "cell cycle control inhibitors" whichinterfere with the progress of a cell through its normal cellcycle, the life span of a cell, from the mitosis that gives itorigin to the events following mitosis that divides it intodaughter cells; "Checkpoint inhibitors" which interfere with thenormal function of cell cycle checkpoints, e.g., the S/G2checkpoint, G2/M checkpoint and G1/S checkpoint; "Topoisomeraseinhibitors", such as camptothecins, which interfere withtopoisomerase I or II activity, enzymes necessary for DNAreplication and transcription; "Receptor tyrosine kinaseinhibitors" which interfere with the activity of growth factorreceptors that possess tyrosine kinase activity; "Apoptosisinducing agents" which promote programmed cell death;"Antimetabolites," such as Gemcitabine or Hydroxyurea, whichclosely resemble an essential metabolite and therefore interferewith physiological reactions involving it; "Telomerase inhibitors"which interfere with the activity of a telomerase, an enzyme thatextends telomere length and extends the lifetime of the cell andits replicative capacity; "Cyclin-dependent kinase inhibitors"which interfere with cyclin-dependent kinases that control themajor steps between different phases of the cell cycle throughphosphorylation of cell proteins such as histones, cytoskeletalproteins, transcription factors, tumor suppresser genes and thelike; "DNA damaging agents"; "DNA repair inhibitors";"Anti-angiogenic agents" which interfere with the generation of newblood vessels or growth of existing blood vessels that occursduring tumor growth; and "Mitochondrial poisons" which directly orindirectly disrupt mitochondrial respiratory chain function.

Especially preferred combinations for treatment of tumors are theclinically approved combinations set forth hereinabove. As thesecombinations have already been approved for use in humans,reformulation to assure appropriate delivery is especiallyimportant.

Preferred agents that may be used in combination include DNAdamaging agents such as carboplatin, cisplatin, cyclophosphamide,doxorubicin, daunorubicin, epirubicin, mitomycin C, mitoxantrone;DNA repair inhibitors including 5-fluorouracil (5-FU) or FUDR,gemcitabine and methotrexate; topoisomerase I inhibitors such ascamptothecin, irinotecan and topotecan; S/G2 or G2/M checkpointinhibitors such as bleomycin, docetaxel, doxorubicin, etoposide,paclitaxel, vinblastine, vincristine, vindesine and vinorelbine;G1/early-S checkpoint inhibitors; G2/M checkpoint inhibitors;receptor tyrosine kinase inhibitors such as genistein, trastuzumab,ZD1839; cytotoxic agents; apoptosis-inducing agents and cell cyclecontrol inhibitors.

The mechanism of action of one or more of the agents may not beknown or may be incorrectly identified. All synergistic or additivecombinations of agents are within the scope of the presentinvention. Preferably, for the treatment of a neoplasm,combinations that inhibit more than one mechanism that leads touncontrolled cell proliferation are chosen for use in accordancewith this invention. For example, the present invention includesselecting combinations that effect specific points within the cellcycle thereby resulting in non-antagonistic effects. For instance,drugs that cause DNA damage can be paired with those that inhibitDNA repair, such as anti-metabolites. The present invention alsoincludes selecting combinations that block multiple pathways thatwould otherwise result in cell proliferation.

Particularly preferred combinations are DNA damaging agents incombination with DNA repair inhibitors, DNA damaging agents incombination with topoisomerase I or topoisomerase II inhibitors,topoisomerase I inhibitors in combination with S/G2 or G2/Mcheckpoint inhibitors, G1/S checkpoint inhibitors or CDK inhibitorsin combination with G2/M checkpoint inhibitors, receptor tyrosinekinase inhibitors in combination with cytotoxic agents,apoptosis-inducing agents in combination with cytotoxic agents,apoptosis-inducing agents in combination with cell-cycle controlinhibitors, G1/S or G2/M checkpoint inhibitors in combination withcytotoxic agents, topoisomerase I or II inhibitors in combinationwith DNA repair inhibitors, topoisomerase I or II inhibitors ortelomerase inhibitors in combination with cell cycle controlinhibitors, topoisomerase I inhibitors in combination withtopoisomerase II inhibitors, and two cytotoxic agents incombination.

Specific agents that may be used in combination include cisplatin(or carboplatin) and 5-FU (or FUDR), cisplatin (or carboplatin) andirinotecan, irinotecan and 5-FU (or FUDR), vinorelbine andcisplatin (or carboplatin), methotrexate and 5-FU (or FUDR),idarubicin and araC, cisplatin (or carboplatin) and taxol,cisplatin (or carboplatin) and etoposide, cisplatin (orcarboplatin) and topotecan, cisplatin (or carboplatin) anddaunorubicin, cisplatin (or carboplatin) and doxorubicin, cisplatin(or carboplatin) and gemcitabine, oxaliplatin and 5-FU (or FUDR),gemcitabine and 5-FU (or FUDR), adriamycin and vinorelbine, taxoland doxorubicin, flavopuridol and doxorubicin, UCN01 anddoxorubicin, bleomycin and trichlorperazine, vinorelbine andedelfosine, vinorelbine and sphingosine (and sphingosineanalogues), vinorelbine and phosphatidylserine, vinorelbine andcamptothecin, cisplatin (or carboplatin) and sphingosine (andsphingosine analogues), sphingosine (and sphingosine analogues) anddaunorubicin and sphingosine (and sphingosine analogues) anddoxorubicin.

Preferred combinations in general include those set forthhereinabove as already shown to be efficacious in the clinic asrecognized by the FDA and those further suggested based onliterature reports. While the candidate agents for use in themethod of the invention are not limited to these specificcombinations, those set forth hereinabove have been disclosed assuitable combination therapies, and are thus preferred for use inthe methods and compositions of the present invention.

Some lipids are "therapeutic lipids" that are able to exerttherapeutic effects such as inducing apoptosis. Included in thisdefinition are lipids such as ether lipids, phosphatidic acid,phosphonates, ceramide and ceramide analogues, dihydroxyceramide,phytoceramide, sphingosine, sphingosine analogues, sphingomyelin,serine-containing lipids and sphinganine The term"serine-containing phospholipid" or "serine-containing lipid" asdefined herein is a phospholipid in which the polar head groupcomprises a phosphate group covalently joined at one end to aserine and at the other end to a three-carbon backbone connected toa hydrophobic portion through an ether, ester or amide linkage.Included in this class are the phospholipids such asphosphatidylserine (PS) that have two hydrocarbon chains in thehydrophobic portion that are between 5-23 carbon atoms in lengthand have varying degrees of saturation. The term hydrophobicportion with reference to a serine-containing phospholipid orserine-containing lipid refers to apolar groups such as longsaturated or unsaturated aliphatic hydrocarbon chains, optionallysubstituted by one or more aromatic, alicyclic or heterocyclicgroup(s).

Combinations of therapeutic lipids and other agents can also beused to achieve synergistic or additive effects (see Examples17-21).

High Throughput Screening for Determining Ratios that DisplayNon-Antagonistic Combination Effects

Chemical libraries of agents may be screened against one another atdifferent ratios to identify novel non-antagonistic drugcombinations. Chemical libraries may comprise novel or conventionalagents. In addition to screening for two agent combinations, threeor four agent combinations may also be screened fornon-antagonistic combination effects. Preferably, the data analysismethodology employed to determine drug synergy is theaforementioned Median Effect Analysis. According to this method,libraries of agents are tested individually and in combination atdifferent ratios. Combination indexes are then calculated using theaforementioned method developed by Chou and Talalay. Drugcombinations that display non-antagonistic effects at specificratios are encapsulated in delivery vehicles at a non-antagonisticratio.

High throughput screening systems are commercially available (see,e.g., Zymark Corp., Hopkinton, Mass.; Air Technical Industries,Mentor, Ohio; Beckman Instruments, Inc. Fullerton, Calif.;Precision Systems, Inc., Natick, Mass., etc.). These systemstypically automate entire procedures including all sample andreagent pipetting, liquid dispensing, timed incubations, and finalreadings of the microplate in detector(s) appropriate for theassay. These configurable systems provide high throughput and rapidstart-up, as well as a high degree of flexibility andcustomization. The manufacturers of such systems provide detailedprotocols for the various high throughput screening methods.

Preparation of Non-Antagonistic Compositions

When the appropriate ratios of the agents have been determined asdescribed above, the agents at the appropriate ratio are placedinto one or more delivery vehicle compositions wherein one or moredelivery vehicles encapsulates two or more agents. Not all thedelivery vehicles in the composition need be identical. Thedelivery vehicles in the compositions are particles of sizes thatdepend on their route of administration, which can be suspended inan aqueous or other solvent and are able to encapsulate the agentsof the invention. Such vehicles include, for example, lipidcarriers, liposomes, cyclodextrins, polymer nanoparticles andpolymer microparticles, including nanocapsules and nanospheres,block copolymer micelles, lipid stabilized emulsions, derivatizedsingle-chain polymers, polymer lipid hybrid systems, lipidmicelles, lipoprotein micelles as mentioned previously. Forintravenous administration, delivery vehicles are typically about4-6,000 nm in diameter. Preferred diameters are about 5-500 nm indiameter, more preferably 5-200 nm in diameter. For inhalation,intra-thecal, intra-articular, intra-arterial, intra-peritoneal orsubcutaneous administration, delivery vehicles are typically from 4.mu.m to an excess of 50 .mu.m. Delivery vehicle compositionsdesigned for intra-ocular administration are generally smaller.

As explained above, the biologically active agents may beformulated into a single composition at the predetermined ratio, orseparate compositions comprising delivery vehicles with coordinatedpharmacokinetics can be employed along with instructions foradministering these compositions in a proportion consistent withthe predetermined ratio. Thus, the desired ratio may be achieved byadministering the agents in separate compositions simultaneously orsequentially in the proportion described.

The therapeutic agents are "encapsulated" in the delivery vehicles."Encapsulation," as previously described, includes covalent ornon-covalent association of an agent with the delivery vehicle. Forexample, this can be by interaction of the agent with the outerlayer or layers of the delivery vehicle or entrapment of an agentwithin the delivery vehicle, equilibrium being achieved betweendifferent portions of the delivery vehicle. For example, forliposomes, encapsulation of an agent can be by association of theagent by interaction with the bilayer of the liposomes throughcovalent or non-covalent interaction with the lipid components orentrapment in the aqueous interior of the liposome, or inequilibrium between the internal aqueous phase and the bilayer. Forpolymer-based delivery vehicles, encapsulation can refer tocovalent linkage of an agent to a linear or non-linear polymer.Further, non-limiting examples include the dispersion of agentthroughout a polymer matrix, or the concentration of drug in thecore of a nanocapsule, a block copolymer micelle or a polymer-lipidhybrid system. "Loading" refers to the act of encapsulating one ormore agents into a delivery vehicle.

Encapsulation of the desired combination can be achieved eitherthrough encapsulation in separate delivery vehicles or within thesame delivery vehicle. Where encapsulation into separate deliveryvehicles, such as liposomes, is desired, the lipid composition ofeach liposome may be quite different to allow for coordinatedpharmacokinetics. By altering the vehicle composition, releaserates of encapsulated drugs can be matched to allownon-antagonistic ratios of the drugs to be delivered to the tumorsite. Means of altering release rates include increasing theacyl-chain length of vesicle forming lipids to improve drugretention, controlling the exchange of surface grafted hydrophilicpolymers such as PEG out of the liposome membrane and incorporatingmembrane-rigidifying agents such as sterols or sphingomyelin intothe membrane. It should be apparent to those skilled in the artthat if a first and second drug are desired to be administered at aspecific drug ratio and if the second drug is retained poorlywithin the liposome composition of the first drug (e.g.,DMPC/Chol), that improved pharmacokinetics may be achieved byencapsulating the second drug in a liposome composition with lipidsof increased acyl chain length (e.g., DSPC/Chol). Alternatively,two or more agents may be encapsulated within the same deliveryvehicle.

Techniques for encapsulation are dependent on the nature of thedelivery vehicles. For example, therapeutic agents may be loadedinto liposomes using both passive and active loading methods.

Passive methods of encapsulating agents in liposomes involveencapsulating the agent during the preparation of the liposomes. Inthis method, the drug may be membrane associated or encapsulatedwithin an entrapped aqueous space. This includes a passiveentrapment method described by Bangham, et al., J. Mol. Biol.(1965) 12:238, where the aqueous phase containing the agent ofinterest is put into contact with a film of dried vesicle-forminglipids deposited on the walls of a reaction vessel. Upon agitationby mechanical means, swelling of the lipids will occur andmultilamellar vesicles (MLV) will form. Using extrusion, the MLVscan be converted to large unilamellar vesicles (LUV) or smallunilamellar vesicles (SUV). Another method of passive loading thatmay be used includes that described by Deamer and Bangham, Biochim.Biophys. Acta (1976) 443:629. This method involves dissolvingvesicle-forming lipids in ether and, instead of first evaporatingthe ether to form a thin film on a surface, this film beingthereafter put into contact with an aqueous phase to beencapsulated, the ether solution is directly injected into saidaqueous phase and the ether is evaporated afterwards, wherebyliposomes with encapsulated agents are obtained. A further methodthat may be employed is the Reverse Phase Evaporation (REV) methoddescribed by Szoka and Papahadjopoulos, P.N.A.S. (1978) 75:4194, inwhich a solution of lipids in a water insoluble organic solvent isemulsified in an aqueous carrier phase and the organic solvent issubsequently removed under reduced pressure.

Other methods of passive entrapment that may be used includesubjecting liposomes to successive dehydration and rehydrationtreatment, or freezing and thawing. Dehydration is carried out byevaporation or freeze-drying. This technique is disclosed by Kirby,et al., Biotechnology (1984) 979-984. Also, Shew and Deamer(Biochim. et Biophys. Acta (1985) 816:1-8) describe a methodwherein liposomes prepared by sonication are mixed in aqueoussolution with the solute to be encapsulated, and the mixture isdried under nitrogen in a rotating flask. Upon rehydration, largeliposomes are produced in which a significant fraction of thesolute has been encapsulated.

Passive encapsulation of two or more agents is possible for manydrug combinations. This approach is limited by the solubility ofthe drugs in aqueous buffer solutions and the large percentage ofdrug that is not trapped within the delivery system. The loadingmay be improved by co-lyophilizing the drugs with the lipid sampleand rehydrating in the minimal volume allowed to solubilize thedrugs. The solubility may be improved by varying the pH of thebuffer, increasing temperature or addition or removal of salts fromthe buffer.

Active methods of encapsulating may also be used. For example,liposomes may be loaded according to a metal-complexation or pHgradient loading technique. With pH gradient loading, liposomes areformed which encapsulate an aqueous phase of a selected pH.Hydrated liposomes are placed in an aqueous environment of adifferent pH selected to remove or minimize a charge on the drug orother agent to be encapsulated. Once the drug moves inside theliposome, the pH of the interior results in a charged drug state,which prevents the drug from permeating the lipid bilayer, therebyentrapping the drug in the liposome.

To create a pH gradient, the original external medium can bereplaced by a new external medium having a different concentrationof protons. The replacement of the external medium can beaccomplished by various techniques, such as, by passing the lipidvesicle preparation through a gel filtration column, e.g., aSephadex G-50 column, which has been equilibrated with the newmedium (as set forth in the examples below), or by centrifugation,dialysis, or related techniques. The internal medium may be eitheracidic or basic with respect to the external medium.

After establishment of a pH gradient, a pH gradient loadable agentis added to the mixture and encapsulation of the agent in theliposome occurs as described above.

Loading using a pH gradient may be carried out according to methodsdescribed in U.S. Pat. Nos. 5,616,341, 5,736,155 and 5,785,987incorporated herein by reference. A preferred method of pH gradientloading is the citrate-based loading method utilizing citrate asthe internal buffer at a pH of 2-6 and a neutral externalbuffer.

Various methods may be employed to establish and maintain a pHgradient across a liposome all of which are incorporated herein byreference. This may involve the use of ionophores that can insertinto the liposome membrane and transport ions across membranes inexchange for protons (see for example U.S. Pat. No. 5,837,282).Compounds encapsulated in the interior of the liposome that areable to shuttle protons across the liposomal membrane and thus setup a pH gradient (see for example U.S. Pat. No. 5,837,282) may alsobe utilized. These compounds comprise an ionizable moiety that isneutral when deprotonated and charged when protonated. The neutraldeprotonated form (which is in equilibrium with the protonatedform) is able to cross the liposome membrane and thus leave aproton behind in the interior of the liposome and thereby cause andecrease in the pH of the interior. Examples of such compoundsinclude methylammonium chloride, methylammonium sulfate,ethylenediammonium sulfate (see U.S. Pat. No. 5,785,987) andammonium sulfate. Internal loading buffers that are able toestablish a basic internal pH, can also be utilized. In this case,the neutral form is protonated such that protons are shuttled outof the liposome interior to establish a basic interior. An exampleof such a compound is calcium acetate (see U.S. Pat. No.5,939,096).

Two or more agents may be loaded into a liposome using the sameactive loading methods or may involve the use of different activeloading methods. For instance, metal complexation loading may beutilized to actively load multiple agents or may be coupled withanother active loading technique, such as pH gradient loading.Metal-based active loading typically uses liposomes with passivelyencapsulated metal ions (with or without passively loadedtherapeutic agents). Various salts of metal ions are used,presuming that the salt is pharmaceutically acceptable and solublein an aqueous solutions. Actively loaded agents are selected basedon being capable of forming a complex with a metal ion and thusbeing retained when so complexed within the liposome, yet capableof loading into a liposome when not complexed to metal ions. Agentsthat are capable of coordinating with a metal typically comprisecoordination sites such as amines, carbonyl groups, ethers,ketones, acyl groups, acetylenes, olefins, thiols, hydroxyl orhalide groups or other suitable groups capable of donatingelectrons to the metal ion thereby forming a complex with the metalion. Examples of active agents which bind metals include, but arenot limited to, quinolones such as fluoroquinolones; quinolonessuch as nalidixic acid; anthracyclines such as doxorubicin,daunorubicin and idarubicin; amino glycosides such as kanamycin;and other antibiotics such as bleomycin, mitomycin C andtetracycline; and nitrogen mustards such as cyclophosphamide,thiosemicarbazones, indomethacin and nitroprusside; camptothecinssuch as topotecan, irinotecan, lurtotecan, 9-aminocamptothecin,9-nitrocamptothecin and 10-hydroxycamptothecin; andpodophyllotoxins such as etoposide. Uptake of an agent may beestablished by incubation of the mixture at a suitable temperatureafter addition of the agent to the external medium. Depending onthe composition of the liposome, temperature and pH of the internalmedium, and chemical nature of the agent, uptake of the agent mayoccur over a time period of minutes or hours. Methods ofdetermining whether coordination occurs between an agent and ametal within a liposome include spectrophotometric analysis andother conventional techniques well known to those of skill in theart.

Furthermore, liposome loading efficiency and retention propertiesusing metal-based procedures carried out in the absence of anionophore in the liposome are dependent on the metal employed andthe lipid composition of the liposome. By selecting lipidcomposition and a metal, loading or retention properties can betailored to achieve a desired loading or release of a selectedagent from a liposome.

Passive and active loading methods may be combined sequentially inorder to load multiple drugs into a delivery vehicle. By way ofexample, liposomes containing a passively entrapped platinum drugsuch as cisplatin in the presence of MnCl.sub.2 may subsequently beused to actively encapsulate an anthracycline such as doxorubicininto the interior of the liposome. This method is likely to beapplicable to numerous drugs that are encapsulated in liposomesthrough passive encapsulation.

Kits

The therapeutic agents in the invention compositions may beformulated separately in individual compositions wherein eachtherapeutic agent is stably associated with appropriate deliveryvehicles. These compositions can be administered separately tosubjects as long as the pharmacokinetics of the delivery vehiclesare coordinated so that the ratio of therapeutic agentsadministered is maintained at the target for treatment. Thus, it isuseful to construct kits which include, in separate containers, afirst composition comprising delivery vehicles stably associatedwith at least a first therapeutic agent and, in a second container,a second composition comprising delivery vehicles stably associatedwith at least one second therapeutic agent. The containers can thenbe packaged into the kit.

The kit will also include instructions as to the mode ofadministration of the compositions to a subject, at least includinga description of the ratio of amounts of each composition to beadministered. Alternatively, or in addition, the kit is constructedso that the amounts of compositions in each container ispre-measured so that the contents of one container in combinationwith the contents of the other represent the correct ratio.Alternatively, or in addition, the containers may be marked with ameasuring scale permitting dispensation of appropriate amountsaccording to the scales visible. The containers may themselves beuseable in administration; for example, the kit might contain theappropriate amounts of each composition in separate syringes.Formulations which comprise the pre-formulated correct ratio oftherapeutic agents may also be packaged in this way so that theformulation is administered directly from a syringe prepackaged inthe kit.

Therapeutic Uses of Delivery Vehicle Compositions EncapsulatingMultiple Agents

These delivery vehicle compositions may be used to treat a varietyof diseases in warm-blooded animals and in avian species. Thus,suitable subjects for treatment according to the methods andcompositions of the invention include humans, mammals such aslivestock or domestic animals, domesticated avian subjects such aschickens and ducks, and laboratory animals for research use.Examples of medical uses of the compositions of the presentinvention include treating cancer, treating cardiovascular diseasessuch as hypertension, cardiac arrhythmia and restenosis, treatingbacterial, viral, fungal or parasitic infections, treating and/orpreventing diseases through the use of the compositions of thepresent inventions as vaccines, treating inflammation or treatingautoimmune diseases.

In one embodiment, delivery vehicle compositions in accordance withthis invention are preferably used to treat neoplasms. Delivery offormulated drug to a tumor site is achieved by administration ofliposomes or other particulate delivery systems. Preferablyliposomes have a diameter of less than 200 nm Tumor vasculature isgenerally leakier than normal vasculature due to fenestrations orgaps in the endothelia. This allows the delivery vehicles of 200 nmor less in diameter to penetrate the discontinuous endothelial celllayer and underlying basement membrane surrounding the vesselssupplying blood to a tumor. Selective accumulation of the deliveryvehicles into tumor sites following extravasation leads to enhanceddrug delivery and therapeutic effectiveness. Because carriersextravasate, it can be assumed that the carrier drug-to-drug ratiodetermined in the blood will be comparable to the carrierdrug-to-drug ratio in the extravascular space.

Administering Delivery Vehicle Compositions

As mentioned above, the delivery vehicle compositions of thepresent invention may be administered to warm-blooded animals,including humans as well as to domestic avian species. Fortreatment of human ailments, a qualified physician will determinehow the compositions of the present invention should be utilizedwith respect to dose, schedule and route of administration usingestablished protocols. Such applications may also utilize doseescalation should agents encapsulated in delivery vehiclecompositions of the present invention exhibit reduced toxicity tohealthy tissues of the subject.

Preferably, the pharmaceutical compositions of the presentinvention are administered parenterally, i.e., intraarterially,intravenously, intraperitoneally, subcutaneously, orintramuscularly. More preferably, the pharmaceutical compositionsare administered intravenously or intraperitoneally by a bolusinjection. For example, see Rahman, et al., U.S. Pat. No.3,993,754; Sears, U.S. Pat. No. 4,145,410; Papahadjopoulos, et al.,U.S. Pat. No. 4,235,871; Schneider, U.S. Pat. No. 4,224,179; Lenk,et al., U.S. Pat. No. 4,522,803; and Fountain, et al., U.S. Pat.No. 4,588,578.

In other methods, the pharmaceutical preparations of the presentinvention can be contacted with the target tissue by directapplication of the preparation to the tissue. The application maybe made by topical, "open" or "closed" procedures. By "topical", itis meant the direct application of the pharmaceutical preparationto a tissue exposed to the environment, such as the skin,oropharynx, external auditory canal, and the like. "Open"procedures are those procedures that include incising the skin of apatient and directly visualizing the underlying tissue to which thepharmaceutical preparations are applied. This is generallyaccomplished by a surgical procedure, such as a thoracotomy toaccess the lungs, abdominal laparotomy to access abdominal viscera,or other direct surgical approach to the target tissue. "Closed"procedures are invasive procedures in which the internal targettissues are not directly visualized, but accessed via insertinginstruments through small wounds in the skin. For example, thepreparations may be administered to the peritoneum by needlelavage. Likewise, the pharmaceutical preparations may beadministered to the meninges or spinal cord by infusion during alumbar puncture followed by appropriate positioning of the patientas commonly practiced for spinal anesthesia or metrazamide imagingof the spinal cord. Alternatively, the preparations may beadministered through endoscopic devices.

Pharmaceutical compositions comprising delivery vehicles of theinvention are prepared according to standard techniques and maycomprise water, buffered water, 0.9% saline, 0.3% glycine, 5%dextrose and the like, including glycoproteins for enhancedstability, such as albumin, lipoprotein, globulin, and the like.These compositions may be sterilized by conventional, well-knownsterilization techniques. The resulting aqueous solutions may bepackaged for use or filtered under aseptic conditions andlyophilized, the lyophilized preparation being combined with asterile aqueous solution prior to administration. The compositionsmay contain pharmaceutically acceptable auxiliary substances asrequired to approximate physiological conditions, such as pHadjusting and buffering agents, tonicity adjusting agents and thelike, for example, sodium acetate, sodium lactate, sodium chloride,potassium chloride, calcium chloride, and the like. Additionally,the delivery vehicle suspension may include lipid-protective agentswhich protect lipids against free-radical and lipid-peroxidativedamages on storage. Lipophilic free-radical quenchers, such asalpha-tocopherol and water-soluble iron-specific chelators, such asferrioxamine, are suitable.

The concentration of delivery vehicles in the pharmaceuticalformulations can vary widely, such as from less than about 0.05%,usually at or at least about 2-5% to as much as 10 to 30% by weightand will be selected primarily by fluid volumes, viscosities, andthe like, in accordance with the particular mode of administrationselected. For example, the concentration may be increased to lowerthe fluid load associated with treatment. Alternatively, deliveryvehicles composed of irritating lipids may be diluted to lowconcentrations to lessen inflammation at the site ofadministration. For diagnosis, the amount of delivery vehiclesadministered will depend upon the particular label used, thedisease state being diagnosed and the judgment of theclinician.

Preferably, the pharmaceutical compositions of the presentinvention are administered intravenously. Dosage for the deliveryvehicle formulations will depend on the ratio of drug to lipid andthe administrating physician's opinion based on age, weight, andcondition of the patient.

In addition to pharmaceutical compositions, suitable formulationsfor veterinary use may be prepared and administered in a mannersuitable to the subject. Preferred veterinary subjects includemammalian species, for example, non-human primates, dogs, cats,cattle, horses, sheep, and domesticated fowl. Subjects may alsoinclude laboratory animals, for example, in particular, rats,rabbits, mice, and guinea pigs.

In the instance where a single composition containing more than oneactive agent is included, the above procedures are followed per se.Where the agents are administered in separate delivery vehiclecompositions, the administration should be timed in such a mannerthat the desired ratio is maintained. Typically, this canaccomplished by simultaneously administering the compositions inthe calculated proportions.

Evaluation of Therapeutic Activity In Vivo

Therapeutic activity of delivery vehicle compositions comprisingtwo or more encapsulated agents may be measured afteradministration into an animal model. Preferably, the animal modelcomprises a tumor although delivery vehicle compositions may beadministered to animal models of other diseases. Rodent speciessuch as mice and rats of either inbred, outbred, or hybrid originincluding immunocompetent and immunocompromised, as well asknockout, or transgenic models may be used.

Models can consist of solid or non-solid tumors implanted as cellsuspensions, bries or tumor fragments in either subcutaneous,intravenous, intraperitoneal, intramuscular, intrathecal, ororthotopic regions. Tumors may also be established via theapplication or administration of tumorigenic/carcinogenic agents ormay be allowed to arise spontaneously in appropriate geneticallyengineered animal models. Tumor types can consist of tumors ofectodermal, mesodermal, or endodermal origin such as carcinomas,sarcomas, melanomas, gliomas, leukemias and lymphomas.

In a preferred embodiment, mouse models of tumors are employed.Human xenograft solid tumors grown in immune compromised mice maybe utilized and selected on the basis of defined genetics andgrowth attributes. Tumor cells utilized in these experiments can begenetically manipulated or selected to express preferableproperties and are injected into mice.

Once the tumors have grown to a palpable (measurable) size,delivery vehicle compositions can be administered, preferablyintravenously, and their effects on tumor growth are monitored.Intended therapeutic treatments can consist of single bolus or pushadministrations or multiple or continuous administrations overseveral days or weeks and by any appropriate route such as by theoral, nasal, subcutaneous, intravenous, intraperitoneal,intrathecal, intratumoral routes using syringes, tablets, liquids,and pumps (such as osmotic). Dose and schedule dependency may beevaluated in order to determine the maximum anti-tumor activitythat can be achieved.

Various methods of determining therapeutic activity in animalmodels comprising a tumor may be utilized. This includes solidtumor model evaluation methods and non-solid tumor model evaluationmethods.

Solid tumor model evaluation methods include measurement of tumorvolume (mass), tumor weight inhibition (TWI %), tumor growth delay(T-C), tumor regression, cell kill and clonogenic assays.

Tumor volume measurements are determined from vernier calipermeasurements of perpendicular length and width measurements (heightmeasurements can often be obtained as well). Tumor volume (mL) ormass (g) is calculated from: volume=(length.times.width.sup.2/2; orvolume=n/.pi./6.times.(length.times.width.times.height). Data isplotted with respect to time.

Tumor weight inhibition (TWI %) is determined by measuring the meantumor weight of a treated group divided by the mean tumor weight ofa control group, minus 1.times.100 at a defined time point.

Tumor growth delay (T-C) is measured as the median time in days fora treated group (T) to reach an arbitrarily determined tumor size(for example, 300 mg) minus median time in days for the controlgroup to reach the same tumor size.

Tumor regression as a result of treatment may also be used as ameans of evaluating a tumor model. Results are expressed asreductions in tumor size (mass) over time.

Cell kill methods of solid tumor model evaluation can involvemeasuring tumors repeatedly by calipers until all exceed apredetermined size (e.g., 200 mg). The tumor growth and tumordoubling time can then be evaluated. Log.sub.10 cell killparameters can be calculated by: log.sub.10 cellkill/dose=(T-C)/((3.32)(T.sub.d)(No. of doses)) log.sub.10 cellkill (total)=(T-C)/((3.32(T.sub.d)) log.sub.10 cell kill(net)=((T-C)-(duration of R.sub.x))/((3.3(T.sub.d)) Where:(T-C)=tumor growth delay T.sub.d=Tumor doubling time

Clonogenicity assays express the effectiveness of therapy. Theseassays include excision assays and characterization of cellsuspensions from solid tumors.

Excision assays, used to assess what fraction of cells, in asuspension prepared from tumors, have unlimited proliferativepotential (i.e., are clonogenic). Three types of excision assaysare: i) TD.sub.50, or endpoint dilution assays: determines thenumber of cells required to produce tumor takes from inocula invivo. ii) In vivo colony assay: assesses the ability of individualcells to form nodules (colonies) in, for example, the lung. iii) Invitro colony assay, tests the ability of individual cells to growinto colonies either in liquid media, when colonies form on theplastic or glass surface of culture dishes, or in semisolid mediasuch as agar, in which the colonies form in suspension.

Characterization of cell suspensions from solid tumors are requiredfor in vitro and in vivo clonogenic assays, flow-cytometricmeasurements, and for numerous biochemical and molecular analysesperformed on a per cell basis. Preparation is by a number ofmethods such as enzymatic, mechanical, chemical, combinationsthereof, and surface activity agents. Evaluations could include,cell yield, cell morphology, tumor cell clonogenicity, retention ofbiochemical or molecular characteristics.

Non-solid tumor model evaluation methods include measurement ofincrease in life span (ILS %), tumor growth delay (T-C), long-termsurvivors (cures).

Increase in life-span (ILS %) measures the percentage increase inlife-span of treated groups versus control or untreated groups.Tumor growth delay (T-C) measures median time in days for treated(T) group survival minus median time in days for control (C) groupsurvival. Long-term survivors (cures) measures treatment groupsthat survive up to and beyond 3-times the survival times ofuntreated or control groups.

Methods of determining therapeutic activity in humans afflictedwith cancer include measurements of survival and surrogateendpoints. The time at which survival is reasonably evaluateddepends on the tumor in question. By way of example, survival ratesfor patients with low-grade lymphomas may be examined at 5 or 10years post diagnosis, whereas the survival or patients havingaggressive diseases such as advanced non-small cell lung cancer maybe best evaluated at 6 or 12 months post diagnosis.

Methods of determining therapeutic activity using surrogateendpoints includes measuring complete response (CR), partialresponse (PR), progression-free survival (PFS), time-to-progression(TTP) or duration of response (DOR), plasma and urine markers,enzyme inhibition and/or receptor status, changes in geneexpression and quality of life (QOL).

A complete response means the disappearance of all known sites ofdisease without the development of any new disease for a period oftime appropriate for the tumor type being treated. Assessments arebased on a variety of examinations such as those stated above.

Partial response is at least a 50% decrease in the sum of theproducts of the bidimensional measurement of all lesions with nonew disease appearing for a period of time appropriate for thetumor type being treated. Assessments are based on a variety ofexaminations (CT scan, MRI, ultrasound, PET scan, bone scan,physical examination) of patients.

Progression-free Survival (PFS): Duration from treatment in which apatient survives and there is no growth of existing tumor norappearance of new tumor masses. PFS may be expressed as either theduration of time or as the proportion of patients who are survivingand progression-free at a given time after diagnosis.

Time-to-progression (TTP) or duration of response (DOR) refer tothe duration of time from treatment to a progression of tumorgrowth, measured either as an increase in size of existing tumormasses or the appearance of new tumor masses.

Plasma and urine markers include measuring markers such as, but notlimited to, the following markers: prostate specific antigen (PSA)and carcinoembryonic antigen (CEA).

Enzyme inhibition and/or receptor status. Growth factor receptorssuch as, but not limited to, tyrosine kinase receptors, EGFreceptor, PDGF receptor, Her-1 and Her-2 receptors. Enzymes suchas, but not limited to, integrin-linked kinases, protein kinasesand the like.

Changes in gene expression include serial analysis of geneexpression (genomics) and changes in protein expression(proteomics).

Quality of Life (QOL) include methods such as the EORTC QLQ-C30scoring method that evaluates yields scores for five functionalscales (physical, role, cognitive, social, and emotional), threesymptom scales (nausea, pain, and fatigue), and a global health andquality of life scale. The measure also yields single-item ratingsof additional symptoms commonly reported by cancer patients(dyspnea, appetite loss, sleep disturbance, constipation, anddiarrhea) as well as the perceived financial impact of the diseaseand its treatment.

The following examples are given for the purpose of illustrationand are not by way of limitation on the scope of the invention.

EXAMPLES

The examples below employ the following methods of determiningcytotoxicity and for evaluating non-antagonistic effects.

Cytotoxicity Assay

In the following examples the standard tetrazolium-basedcolorimetric MTT cytotoxicity assay protocol (Mosmann, et al., J.Immunol Methods (1983) 65(1-2):55-63) was utilized to determine thereadout for the fraction of cells affected. Briefly, viable cellsreduce the tetrazolium salt,3-(4,5-diethylthiazoyl-2-yl)-2,5-diphenyltetrazolium bromide (MTT)to a blue formazan which can be read spectrophotometrically. Cells,such as human H460 non-small-cell lung carcinoma (NSCLC) cellsgrown in 25 cm.sup.2 flasks are passaged (passage number<20),resuspended in fresh RPMI cell culture medium and added into96-well cell culture plates at a concentration of 1000 cells perwell in 100 .mu.L per well. The cells are then allowed to incubatefor 24 hours at 37.degree. C., 5% CO.sub.2. The following day,serial drug dilutions are prepared in 12-well cell culture plates.The agents, previously prepared in various solutions, are dilutedin fresh RPMI cell culture media. Agents are administered to theappropriate or specified wells for single agents (20 .mu.L) and atspecific fixed ratio dual agent combinations (increments of 20.mu.L) using a Latin square design or "checkerboard" dilutionmethod. The total well volumes are made up to 200 .mu.L with freshmedia. The drug exposure is for a duration of 72 hours.

Following drug exposure, MTT reagent (1 mg/mL in RPMI) is added toeach well at a volume of 50 .mu.L per well and incubated for 3-4hours. The well contents are then aspirated and 150 .mu.L ofdimethylsulfoxide (DMSO) is added to each well to disrupt the cellsand to solubilize the formazan precipitate within the cells. The96-well plates are shaken on a plate shaker, and read on amicroplate spectrophotometer set at a wavelength of 570 nm. Theoptical density (OD) readings are recorded and the OD values of theblank wells (containing media alone) are subtracted from all thewells containing cells. The cell survival following exposure toagents is based as a percentage of the control wells (cells notexposed to drug). All wells are performed in triplicate and meanvalues are calculated.

Median-Effect Analysis for Drug Combinations

For the drug combination analysis, the software program CalcuSyn,(Biosoft, Ferguson, Mo., USA) based on the median-effect principleby Chou and Talalay, was utilized. The fixed ratios for thedual-agent combinations are initially derived from theIC.sub.50:IC.sub.50 ratios from single agent cytotoxicity profiles.Subsequently, more relevant fixed ratios (e.g. ranging from 10:1 to1:10; mole ratios) are chosen based upon considerations forformulation purposes. From the mean values calculated based onagent effects on cell survival, doses and respective fractionaleffect values are entered into the CalcuSyn computer program. Thesoftware then determines whether the drug combinations aresynergistic, additive or antagonistic based on combination index(CI) values.

Example 1

Multiple Representation of Dose-Effect Analysis

Quantitative analysis of the relationship between an amount (doseor concentration) of drug and its biological effect as well as thejoint effect of drug combinations can be measured and reported in anumber of ways. FIG. 2 illustrates 5 such methods using, as anexample, a combination of irinotecan and carboplatin.

Based on Chou and Talalay's theory of dose-effect analysis, a"median-effect equation" has been used to calculate a number ofbiochemical equations that are extensively used in the art.Derivations of this equation have given rise to higher orderequations such as those used to calculate Combination Index (CI).As mentioned previously, CI can be used to determine ifcombinations of more than one drug and various ratios of eachcombination are antagonistic, additive or synergistic. CI plots aretypically illustrated with CI representing the y-axis versus theproportion of cells affected, or fraction affected (f.sub.a), onthe x-axis. FIG. 2A demonstrates that a 1:10 mole ratio ofirinotecan/carboplatin is antagonistic (CI>1.1), while 1:1 and10:1 have a synergistic effect (CI<0.9).

The present applicants have also designed an alternative method ofrepresenting the dependency of CI on the drug ratios used. Themaximum CI value is plotted against each ratio to better illustratetrends in ratio-specific effects for a particular combination asseen in FIG. 2B. The CI maximum is the CI value taken at a singlef.sub.a value (between 0.2 and 0.8) where the greatest differencein CI values for the drugs at different ratios was observed.

Because the concentrations of drugs used for an individual ratioplay a role in determining the effect (i.e., synergism orantagonism), it can also be important to measure the CI at variousconcentrations. These concentrations, also referred to as"Effective Doses" (ED) by Chou-Talalay, are the concentration ofdrug required to affect a designated percent of the cells in an invitro assay, i.e., ED.sub.50 is the concentration of drug requiredto affect 50% of the cells relative to a control or untreated cellpopulation. As shown in FIG. 2C, trends in concentration-effect arereadily distinguished between the various ratios. The error barsshown represent one standard deviation around the mean and isdetermined directly through the CalcuSyn program.

A synergistic interaction between two or more drugs has the benefitthat it can lower the amount of each drug required in order toresult in a positive effect, otherwise known as "dose-reduction."Chou and Talalay's "dose-reduction index" (DRI) is a measure of howmuch the dose of each drug in a synergistic combination may bereduced at a given effect level compared with the doses for eachdrug alone. DRI has been important in clinical situations, wheredose-reduction leads to reduced toxicity for the host whilemaintaining therapeutic efficacy. The plot in FIG. 2D shows thatthe concentrations of irinotecan and carboplatin required toachieve a 90% cell kill on their own is significantly higher thantheir individual concentrations required when they are combined ata non-antagonistic ratio.

Furthermore the aforementioned data can be represented in aclassical isobologram (FIG. 2E). Isobolograms have the benefit thatthey can be generated at different ED values; however, they becomemore difficult to read as more effect levels are selected forinterpretation. For this reason, the data in the examples below aregenerally presented in accordance with the types of plots shown inFIGS. 2A and 2B.

Example 2

CI is Dependent Upon Concentrations

Drug combinations of irinotecan and 5-Fluorouracil (5-FU) at moleratios of 1:1 and 1:10 and etoposide and carboplatin at mole ratiosof 10:1 and 1:10 were tested for additive, synergistic orantagonistic effects using the standard tetrazolium-basedcolorimetric MTT cytotoxicity assay and the median-effect analysisas described in the previous example sections. HT29 or MCF-7 cellswere exposed to single agents as well as agents in combination atdefined ratios. Eight drug concentrations were utilized for singleagents and combinations. Optical density values were obtained fromthe MTT assay, converted into a percentage of the control, averagedand then converted into fraction affected values. Dose and fractionaffected values were entered into CalcuSyn which yielded the CIversus f.sub.a graph, shown in FIG. 3.

FIG. 3A shows that irinotecan and 5-FU at a mole ratio of 1:1 werenon-antagonistic over the entire range of concentrations asmeasured by the fraction-affected dose. In contrast, at a moleratio of 1:10, the same two drugs were non-antagonistic at lowconcentrations, yet antagonistic at higher concentrations. As seenin FIG. 3B, etoposide and carboplatin were antagonistic at a moleratio of 10:1 over the entire concentration range. In contrast, ata 1:10 mole ratio, etoposide and carboplatin were antagonistic atlow concentrations while non-antagonistic at higherconcentrations.

Cisplatin and edelfosine at mole ratios of 10:1 and 1:1 were alsoshown to exhibit distinct combination effects in H460 cells assummarized by plotting CI versus f.sub.a. As shown in FIG. 4, thecombination at a 10:1 mole ratio was non-antagonistic forapproximately 50% of the fraction affected range at lowconcentrations and antagonistic at higher concentrations, while a1:1 mole ratio demonstrated synergy over the entire concentrationrange.

These results thus demonstrate that synergy is highly dependent onnot only the ratio of the agents to one another but also theirconcentrations.

Example 3

Determination of CI for Various Two-Drug Combinations

Various drug combinations presented in the table below were testedfor additive, synergistic or antagonistic effects using the MTTcytotoxicity assay protocol and the median-effect analysisprocedure described above. Results from the CI versus f.sub.agraphs are tabulated below. The approximate percentage of thef.sub.a range that exhibited a non-antagonistic effect is reportedin brackets following the ratio. Measurements were taken betweenf.sub.a values of 0.2 and 0.8 and the percent of that f.sub.a rangeexhibiting a synergistic or additive effect (non-antagonistic) wascalculated by determining the percentage of the curve falling belowa CI value of 1.1. Data is derived from at least one experimentperformed in triplicate.

TABLE-US-00004 DRUG COMBINATION CELL LINE MOLE RATIO [% Synergisticor Additive.sup.a] Irinotecan:5-FU H460 1:10 [83%], 1:1 [17%], 10:1[100%] Irinotecan:5-FU MCF-7 1:10 [48% additive.sup.b], 1:1 [58%],10:1 [90%] Irinotecan:5-FU HT29 1:10 [75%], 1:1 [100%]FUDR:Irinotecan HCT-116 1:10 [0%], 1:5 [92%], 1:1 [100%], 5:1[100%], 10:1 [100%] FUDR:Irinotecan HT29 1:10 [40%], 1:5 [73%], 1:1[100%], 5:1 [100%], 10:1 [95%] 5-FU:Carboplatin H460 1:10 [48%],1:1 [100%], 10:1 [100%] FUDR:Carboplatin H460 1:10 [37%], 1:5[100%], 1:1 [100%], 5:1 [100% additive.sup.b], 10:1 [100%additive.sup.b] Irinotecan:Carboplatin H460 1:10 [0%], 1:1 [13%],10:1 [100% additive.sup.b] Irinotecan:Carboplatin A549 1:10 [0%],1:1 [100%], 10:1 [100%] Cisplatin:Irinotecan H460 1:10 [100%], 1:1[56%], 10:1 [100% additive.sup.b] Cisplatin:Irinotecan MCF-7 1:10[100%], 1:1 [92%], 10:1 [50%] Etoposide:Carboplatin H460 1:10[55%], 1:1 [76% additive.sup.b], 10:1 [0%] Etoposide:CarboplatinMCF-7 1:10 [65%], 1:1 [30%], 10:1 [0%] Carboplatin:Taxol H460 1:10[100%], 1:1 [100%], 1:100 [0%] Carboplatin:Taxol MCF-7 1:10 [100%],1:1 [43%], 1:100 [0%] Taxol:Doxorubicin H460 1:5 [52%], 1:1 [37%additive.sup.b], 1:10 [22%] Taxol:Doxorubicin MCF-7 1:5 [70%], 1:1[100%], 1:10 [63%] Camptothecin:Taxol H460 1:1 [0%], 1:10 [100%]Doxorubicin:Vinorelbine H460 20:1 [0%], 1:1 [100%]Cisplatin:Etoposide H460 50:1 [0%], 1:1 [100%] Cisplatin:EtoposideMCF-7 25:1 [0%], 1:1 [100%] Suramin:Vinorelbine H460 10:1 [0%],20,000:1 [72%] Cisplatin:Edelfosine H460 10:1 [72%], 1:1 [100%]Cisplatin:Safingol H460 1:1 [0%], 0.1:1 [100%] Cisplatin:SafingolMCF-7 1:1 [58%], 0.1:1 [100%] Cisplatin:.beta.-sitosterol H460 10:1[0%], 0.1:1 [100%] Cisplatin:.beta.-sitosterol MCF-7 10:1 [34%],0.1:1 [100%] Cisplatin:Suramin H460 1:100 [37%], 1:40 [0%]Vinorelbine:Cisplatin H460 1:500 [0%], 1:200 [8% additive.sup.b]Vinorelbine:Edelfosine H460 1:10 [0%], 1:1 [0%]Doxorubicin:Cytosine H460 1:0.45 [0%] ArabinosideDoxorubicin:Methotrexate H460 1:0.36 [0%] .sup.a"% Synergistic orAdditive" is calculated as the percent of the f.sub.a range thatdoes not fall in the antagonistic range (CI values >1.1 areantagonistic) on a CI vs. fraction affected (f.sub.a) plot, basedon the Chou-Talalay Method, between f.sub.a values of 0.2 to 0.8.CI was measured by entering dose and f.sub.a values into CalcuSyn..sup.bThe data set for this ratio was in the "additive" range (CIbetween 0.9 and 1.1).

Example 4

Synergism of Carboplatin and Daunorubicin

The procedure set forth above for measuring additive, synergisticor antagonistic effects was repeated using carboplatin/daunorubicinat 10:1, 1:1 and 1:10 mole ratios in H460 cells and at 10:1 and 1:1ratios in MCF-7 cells. A combination index was determined for eachdose by producing CI versus f.sub.a curves as described above andthen determining the CI at f.sub.a values of 0.50, 0.75 and 0.90(to yield CI values at ED50, ED75 and ED90, respectively). Standarddeviations were calculated by the CalcuSyn program. As shown in theinset of FIG. 5A, carboplatin and daunorubicin at a mole ratio of10:1 displays a synergistic interaction at ED50, ED75 and ED90values in MCF-7 cells. As further shown in the inset of FIG. 5A,carboplatin and daunorubicin at a 1:1 mole ratio is synergistic, asjudged by the mean CI values at ED75 and ED90 while being additiveat ED50. In H460 cells, a plot of the CI maximum versus mole ratioof carboplatin/daunorubicin reveals that at a mole ratio of 10:1,the drugs are synergistic while at a mole ratio of 1:1, a slightlyantagonistic effect is observed. In contrast, a stronglyantagonistic effect is exhibited at a ratio of 1:10 (FIG. 5A). Datahave also been plotted in FIG. 5B as CI versus the fraction of H460cells affected to better illustrate the effect of concentration onsynergy. A 1:1 mole ratio of carboplatin/daunorubicin isnon-antagonistic at fraction affected values up to 0.42. At a ratioof 10:1, synergy is observed over a substantial range of f.sub.avalues (greater than 0.2) and a 1:10 ratio is antagonistic at allf.sub.a values. The inset of FIG. 5B shows that at a 10:1 ratio inH460 cells, synergy (as judged by the mean CI values) is observedat ED50, 75 and 90 and at a 1:1 ratio, additivity is indicated atthe ED50. At a 1:10 ratio, carboplatin/daunorubicin is stronglyantagonistic at ED50, 75 and 90 values. Based on these results,carboplatin and daunorubicin at a 1:10 mole ratio would thereforenot be selected for further formulation and in vivo studies, asantagonism is observed at all ED values measured and over the fullf.sub.a range in the CI versus f.sub.a plots. Mole ratios of 10:1and 1:1 carboplatin:daunorubicin are selected for formulation andefficacy studies as at each of these ratios, the drugs demonstratesynergistic effects over at least 5% of the f.sub.a range (wheregreater than 1% of the cells are affected).

Example 5

Maintaining Synergism of Carboplatin and Daunorubicin In Vivo

Carboplatin and daunorubicin were co-loaded into a singlecholesterol-free liposome at mole ratios of 10:1, 5:1 and 1:1(carboplatin/daunorubicin). DSPC was dissolved in chloroform andDSPG was dissolved in chloroform/methanol/water (50:10:1 vol/vol)with trace amounts of .sup.14C-CHE. The solutions were combined ata mole ratio of 80:20 (DSPC/DSPG). Solvent was removed under astream of N.sub.2 gas while maintaining the temperature at greaterthan 60.degree. C. The lipid film was then placed in a vacuum pumpfor 2 minutes and subsequently redissolved in chloroform only. Thechloroform was then removed as above. The resulting lipid filmswere left under vacuum overnight to remove any residual solventfollowed by rehydration in 150 mM CuSO.sub.4, pH 7.4 (pH adjustedwith triethanolamine) containing 80 mg/mL carboplatin with 4% (v/v)DMSO to increase carboplatin solubility. The resultingmultilamellar vesicles (MLVs) were extruded at 70.degree. C.through two stacked 80 and 100 nm pore size filters for a total often passes. The samples were exchanged into saline and then into300 mM sucrose, 20 mM HEPES, 30 mM EDTA, pH 7.4 (SHE) usingtangential flow dialysis. Daunorubicin (with trace amounts of.sup.3H-daunorubicin) was loaded into the liposomes by incubationat 60.degree. C. for 5 minutes at drug to lipid ratios to achievecarboplatin/daunorubicin mole ratios of 10:1, 5:1 and 1:1.Subsequently, each sample was buffer exchanged into saline bytangential flow. To determine the extent of drug loading at varioustimes, during preparation of the co-loaded formulation,daunorubicin and lipid levels were measured by liquid scintillationcounting. Carboplatin concentrations were measured by atomicabsorption spectrometry. Balb/c mice were intravenouslyadministered 8 mg/kg carboplatin and daunorubicin was dosed at 1.2mg/kg, 6 mg/kg and 12 mg/kg for mole ratios of 10:1, 5:1 and 1:1carboplatin/daunorubicin, respectively in the co-loadedformulation. At the indicated time points (3 mice per time point),blood was collected by cardiac puncture and placed into EDTA coatedmicrotainers. The samples were centrifuged and plasma was carefullytransferred to another tube. Liquid scintillation counting was usedto quantitate plasma daunorubicin and lipid levels; plasmacarboplatin levels were determined by atomic absorptionspectrometry. For quantitation by atomic absorption spectrometry,samples were diluted in 0.1% nitric acid to fall within the linearrange of a standard curve.

Results in FIG. 6, where the mean plasma drug concentration(+/-standard deviation, SD) is plotted at the specified times,indicate that the co-loaded liposomal formulations containingcarboplatin and daunorubicin at a 10:1 mole ratio maintained theratio of the drugs after intravenous administration as the moleplasma concentrations of carboplatin were present at ten times thatof daunorubicin. Results in FIGS. 7A and 7B demonstrate that 10:1,5:1 and 1:1 mole ratios of carboplatin to daunorubicin formulatedin DSPC/DSPG liposomes were maintained in the blood compartmentover the 24 hour time course (3 mice per time point) afterintravenous administration of formulations prepared at these ratios(FIG. 7B more clearly highlights the results obtained followingadministration of the 1:1 carboplatin/daunorubicin formulation).These results thus demonstrate that coordinated release kinetics oftwo drugs at a variety of mole ratios can be achieved.

Carboplatin and daunorubicin were also co-formulated intoDSPC/SM/DSPE-PEG2000 (90:5:5 mol %) liposomes in order to determinewhether coordinated release of the drugs in vivo could be achievedusing this formulation as well. A mole ratio of 10:1 was selectedthat was determined to be synergistic in Example 4.

Lipid films (with trace amounts of .sup.14C-CHE) were prepared asdescribed above by solubilizing the lipids in chloroform, removingthe chloroform under N.sub.2 gas and placing the samples in avacuum pump overnight. The resulting lipid films were hydrated in150 mM CuSO.sub.4, 20 mM histidine, pH 7.4 (pH adjusted withtriethanolamine) containing 40 mg/mL carboplatin. MLVs wereextruded at 70.degree. C. through two stacked filters of 100 nmpore sizes for a total of ten passes. Samples were then exchangedinto 300 mM sucrose, 20 mM HEPES, pH 7.4 by tangential flowdialysis to remove unencapsulated metal solution (or carboplatin).Daunorubicin loading (with trace levels of .sup.3H-daunorubicin)was carried out at 60.degree. C. for 5 minutes at a drugconcentration to achieve a 10:1 mole ratio ofcarboplatin/daunorubicin. To determine the extent of drug loading,daunorubicin and lipid levels were measured by liquid scintillationcounting; carboplatin levels were determined by atomic absorptionspectrometry. Male SCID/rag2 mice were administered 2.25 mg/kgdaunorubicin and 15 mg/kg carboplatin intravenously of thecombination co-loaded in DSPC/SM/DSPE-PEG2000 liposomes. At theindicated time points (3 mice per time point), blood was collectedby cardiac puncture and placed into EDTA coated microtainers. Thesamples were centrifuged and plasma was carefully transferred toanother tube. Plasma carboplatin and daunorubicin levels weredetermined by atomic absorption spectrometry and liquidscintillation counting, respectively.

The results set forth in FIG. 8, where the mean plasma drugconcentration (+/-standard deviation, SD) is plotted at theindicated times, reveal that carboplatin and daunorubicin wereeliminated from the plasma compartment at the same rate followingintravenous administration when formulated in DSPC/SM/DSPE-PEG2000liposomes. Carboplatin and daunorubicin were thus maintained at a10:1 mole ratio, as the plasma concentration of carboplatin(nmoles/mL) was present at roughly ten times that of daunorubicin(nmoles/mL) during the time course. These results illustrate that avariety of formulations can be utilized to coordinate thepharmacokinetics of two drugs co-encapsulated in a single liposomesuch that similar pharmacokinetic release profiles areachieved.

Example 6

Efficacy of Liposomal Carboplatin and Daunorubicin

DSPC/DSPG liposomes (80:20 mol %) co-encapsulated with daunorubicinand carboplatin at a mole ratio of 1:1 (that was selected forformulation in Example 4) were prepared as described in Example 5except lipid films were hydrated in a 150 mM CuSO.sub.4, pH 7.4 (pHadjusted with triethanolamine), solution containing 25 mg/mL ofcarboplatin. As well, the lipid films were re-dissolved after beingdried down in chloroform to remove methanol or water and thensolvent was removed as described previously.

As in the method of Example 26, efficacy studies were carried outby first inoculating H460 cells (1.times.10.sup.6 cells)subcutaneously into the flank of female SCID/rag2 mice. Tumors wereallowed to grow until about 50 mg (0.05 cm.sup.3) in size at whichtime (day 12) the formulations were injected via the tail vein.Animals (4 mice per group) were treated with three injections, withinjections being given every fourth day (q4d schedule; on days 12,16 and 20). Tumor growth was determined by direct calipermeasurements. Mice were treated with saline, free drug cocktail ata 1:1 mole ratio or a liposomal formulation ofcarboplatin/daunorubicin at a 1:1 mole ratio. For both the free andliposome-formulated treatments, the doses were 6.6 mg/kgcarboplatin and 10 mg/kg daunorubicin. Lipid doses were 260 mg/kglipid for liposome formulated samples.

Results presented in FIG. 9 (points represent mean tumorsize+/-standard error of the mean (SEM) determined on the specifiedday) show that administration of liposomal carboplatin anddaunorubicin at a 1:1 mole ratio increased efficacy in relation tofree drug cocktail and saline controls.

Efficacy was also examined in sphingomyelin containing liposomesco-loaded with carboplatin and daunorubicin at a 10:1 mole ratio(determined to be synergistic in Example 4) to examine if the largeimprovements in efficacy observed for DSPC/DSPG liposomes could beachieved using this formulation as well. Carboplatin anddaunorubicin were co-formulated into DSPC/SM/DSPE-PEG2000 (90:5:5mol %) liposomes according to the procedure outlined in Example 5except liposomes were extruded through an 80 nm and a 100 nm poresize filter ten times. As well, the samples were buffer exchangedinto SHE buffer prior to loading of daunorubicin by fixed volumedialysis rather than tangential flow dialysis. As detailed inExample 26, H460 tumor bearing female SCID/rag2 mice (4 mice pergroup) were administered 15 mg/kg carboplatin and 2.25 mg/kgdaunorubicin for liposome formulated drug and free drug cocktail ondays 14, 18 and 22. Liposomal drug was administered at a lipid doseof 375 mg/kg.

Results presented in FIG. 10 (points represent mean tumorsize+/-SEM determined on the specified day) show that liposomalcarboplatin and daunorubicin encapsulated at a 10:1non-antagonistic mole ratio in sphingomyelin-containing liposomesexhibit substantially increased efficacy in relation to controlsconsisting of free drug and saline

Example 7

Synergism of Cisplatin and Daunorubicin

Cisplatin/daunorubicin combinations were tested for additive,synergistic or antagonistic effects using the methods describedabove. The results are summarized in FIG. 11. As shown in FIG. 11A,synergy was observed at a cisplatin/daunorubicin mole ratio of 10:1over the entire f.sub.a range while the 1:1 mole ratios displayedantagonism over the complete f.sub.a range. FIG. 11B, a plot of CImaximum (CI max) vs. cisplatin-to-daunorubicin ratio, furtherillustrates the dependence of the combination ratio of two agentson the combination index. These results show that at a 10:1 moleratio, the CI max value is synergistic while at 1:1 and 1:10 moleratios the CI max value is antagonistic.

Example 8

Maintaining Synergism of Cisplatin and Daunorubicin In Vivo

Cisplatin and daunorubicin were co-loaded into DMPC/Chol (55:45 mol%) liposomes at a 10:1 mole ratio identified in Example 7 as beingnon-antagonistic.

Cisplatin was passively entrapped in liposomes by firstsolubilizing the drug (at 40 mg/mL) in a solution consisting of 150mM CuCl.sub.2, 20 mM histidine (pH 7.4, pH adjusted withtriethanolamine) plus 4% (v/v) DMSO and heating the resultingsolution to 80.degree. C. to enhance the solubility of cisplatin.The cisplatin solution was then added at 80.degree. C. to a lipidfilm composed of DMPC and cholesterol with trace levels of.sup.14C-CHE. The hydrated lipid films were extruded at 80.degree.C. through two 100 nm filters and the liposomes cooled to roomtemperature. Upon cooling, the samples were centrifuged in a benchtop centrifuge at 2000.times.g for 5 minutes to pellet anyunencapsulated cisplatin, and the supernatant collected. Removal ofexcess metal ions was carried out by passage through a SephadexG-50 gel filtration column and collection of the liposomefraction.

The cisplatin-loaded liposomes were further loaded withdaunorubicin (labeled with trace levels of .sup.3H-daunorubicin) ata 10:1 cisplatin/daunorubicin mole ratio by incubation of theliposomes with the drug at 60.degree. C. for 15 minutes. In orderto determine the extent of drug loading, cisplatin levels weremeasured by atomic absorption spectrometry and .sup.3H-daunorubicinand lipid levels were measured by liquid scintillationcounting.

In order to determine whether coordinated release was achieved bythis formulation, the loaded liposomes were injected into the tailvein of male SCID/rag2 mice at 5.0 mg/kg cisplatin and 1.0 mg/kgdaunorubicin per mouse. At the indicated time points (3 mice pertime point), blood was collected by cardiac puncture and placedinto EDTA coated microtainers. The samples were centrifuged andplasma was carefully transferred to another tube. Liposomal lipidand daunorubicin levels in the plasma were both determined byliquid scintillation counting and cisplatin levels were measured byatomic absorption spectrometry.

Results depicted in FIG. 12 (points represent mean drugconcentration in plasma +/-SD determined at the specified time)indicate that coordinated release of daunorubicin and cisplatin wasachieved as the concentrations in the plasma (.mu.moles/mL) weremaintained at a mole ratio of 10:1 at the time points measured.

Although liposomes may be co-loaded with cisplatin and daunorubicinby the method described above, other techniques may be employed toload the drugs into a single liposome. An alternative methodemploys the use of a pH gradient to load daunorubicin afterpassively entrapping cisplatin along with citrate, pH 4.0, andimposing a pH gradient across the membrane by buffer exchange. Thistechnique may be carried out as follows:

Lipid films consisting of DSPC/Chol (55:45 mol %) are prepared asdescribed above along with trace amounts of .sup.3H-CHE. Acisplatin solution is prepared by dissolving cisplatin powder into150 mM NaCl and 150 mM citrate (pH 4). To maximize the solubilityof cisplatin in the buffer, the solution is heated to 65.degree. C.and added to the lipid films. The resulting MLVs are extruded at65.degree. C. through two 100 nm pore size filters for a total often passes. Unencapsulated cisplatin is then removed from theformulation by centrifuging the solution at 2000.times.g for 10minutes. The resulting supernatant containing liposomal cisplatinis passed down a Sephadex G-50 column that is pre-equilibrated in150 mM NaCl and 20 mM HEPES (pH 7.4) to remove any residualunentrapped cisplatin and to establish a pH gradient across thebilayer.

Daunorubicin is subsequently loaded into the liposomes by firstincubating the liposomes at 60.degree. C. for 5 minutes to achievethermal equilibration and then adding daunorubicin to the lipidformulation at a 0.1:1 drug/lipid mole ratio while vortexing. Todetermine the extent of drug loading at various times, theconcentration of daunorubicin is determined by solubilizing theliposomes with OGP and measuring the absorbance of daunorubicin at480 nm. The cisplatin concentration of the formulation is measuredusing atomic absorption spectrometry. Lipid concentrations aremeasured by liquid scintillation counting.

An alternative means of coordinating the release kinetics of twodrugs can be achieved by formulating each drug in separatecarriers. This was demonstrated by formulating cisplatin inDMPC/cholesterol liposomes and daunorubicin in DSPC/DSPE-PEG2000liposomes and administering them intravenously to mice at a 10:1mole ratio.

Liposomal cisplatin was prepared by first dissolving cisplatin (8.5mg/mL) in 150 mM NaCl at 80.degree. C. The solution was next addedto a DMPC/cholesterol (55:45 mol %) lipid film containing traceamounts of .sup.3H-CHE and allowed to hydrate. The resulting MLVswere extruded at 80.degree. C. through two 100 nm pore size filtersand the liposomes were subsequently exchanged into 20 mM HEPES, 150mM NaCl (pH 7.4) (HBS) by tangential flow dialysis to remove excessmetal ions. The liposomes were centrifuged to pellet anyunencapsulated cisplatin after extrusion. The cisplatinconcentration was determined by atomic absorption spectrometry andlipid levels were determined by liquid scintillation counting.

Liposomal daunorubicin was prepared by hydration of a lipid filmcomposed of DSPC/ DSPE-PEG2000 (95:5 mol %) and trace amounts of.sup.14C-CHE with a solution of 300 mM CuSO.sub.4. The resultingMLVs were extruded by ten passes through two stacked 100 nm poresize filters at 70.degree. C. After extrusion, the liposomes wereexchanged into HBS (pH 7.4) by tangential flow dialysis. Loading ofdaunorubicin (with trace levels of .sup.3H-daunorubicin) wasinitiated by the addition of daunorubicin to a final drug/lipidweight ratio of 0.1 and holding the solution at 60.degree. C. for10 minutes. The extent of drug loading was measured by liquidscintillation counting to measure .sup.3H-daunorubicin and.sup.14C-CHE levels.

Male SCID/rag 2 mice were injected intravenously with liposomalcisplatin at a drug dose of 2 mg/kg and liposomal daunorubicin at adrug dose of 0.375 mg/kg. At the indicated time points (3 mice pertime point), blood was collected by cardiac puncture and placedinto EDTA coated microtainers. The samples were centrifuged andplasma was carefully transferred to another tube. Plasma cisplatinlevels were determined by atomic absorption spectrometry anddaunorubicin levels were determined by scintillation counting.

Results shown in FIG. 13 (points represent mean drug concentrationsdetermined in plasma+/-SD at the specified time points) reveal thatcisplatin and daunorubicin formulated in separate liposomes weremaintained at a 10:1 mole ratio at various time points afterintravenous administration.

Example 9

Efficacy of Liposomal Cisplatin and Daunorubicin

The efficacy of cisplatin and daunorubicin formulated in separateliposomes was determined in SCID/rag2 mice (H460 xenograft model)as detailed in Example 26. H460 tumor bearing mice (4 mice pergroup) were treated with saline or with cisplatin/daunorubicin at a10:1 mole ratio that was identified in vitro in Example 7 as beingnon-antagonistic. Cisplatin and daunorubicin were formulated inDMPC/Chol (55:45 mol %) and DSPC/DSPE-PEG2000 (95:5 mol %)liposomes respectively as set forth in Example 8, except DMPC/Cholliposomes were dialyzed against HBS after extrusion. Animalstreated with the drug combination received the agents as either acocktail of the free agents (cocktail; 10:1, mole ratio) or byco-administration of liposomal daunorubicin and liposomal cisplatin(liposome formulation; 10:1 mole ratio) on days 14, 17 and 21. Forboth the free and formulated treatments, the doses were 2.0 mg/kgof cisplatin and 0.375 mg/kg of daunorubicin. Lipid doses were 400mg/kg for liposomal cisplatin and 3.75 mg/kg for liposomaldaunorubicin.

FIG. 14 shows the results, where each data point represents meantumor size+/-SEM determined on the specified day. The salinecontrol (solid circles) did not inhibit tumor growth; similarly,the free cocktail (solid inverted triangles) showed only a slighteffect on tumor growth. In comparison, the liposomal formulation(open triangles) inhibited tumor growth over a period of at least32 days.

Example 10

Effect of Liposomal Administration of a Drug Combination at anAntagonistic Mole Ratio

Cisplatin and daunorubicin were co-loaded into DMPC/Chol (55:45 mol%) liposomes at a 1:1 mole ratio that was determined in Example 7to be antagonistic. Cisplatin was passively entrapped anddaunorubicin actively entrapped to achieve a cisplatin/daunorubicinmole ratio of 1:1. The procedure outlined in Example 8 was employedto load the drugs into a single liposome.

In order to determine whether coordinated release was achieved byformulation in DMPC/Chol liposomes, the loaded liposomes wereinjected into the tail vein of Balb/c mice at 2 mg/kg cisplatin and3.75 mg/kg daunorubicin. At the indicated time points (3 mice pertime point), blood was collected by cardiac puncture and placedinto EDTA coated microtainers. The samples were centrifuged andplasma was carefully transferred to another tube. Lipid anddaunorubicin plasma levels were both determined by liquidscintillation counting and cisplatin levels were measured by atomicabsorption spectrometry. Results summarized in FIG. 15 (data pointsrepresent mean drug concentrations determined in plasma+/-SD at thespecified time points) show that daunorubicin and cisplatin wereeliminated from the plasma at the same rate, thus theconcentrations in the plasma (nmoles/mL) were maintained at a moleratio of 1:1 (see insert to FIG. 15).

Efficacy studies were carried out as described in Example 26, whereH460 tumor bearing female SCID/rag2 mice were dosed at 2.5 mg/kgcisplatin, 4.7 mg/kg daunorubicin in either cocktail or liposomalformulation and 52.83 mg/kg lipid on days 11, 15 and 19.

Efficacy results in FIG. 16 (data points represent mean tumorsize+/-SEM determined on the specified day) show that treatmentwith daunorubicin and cisplatin at an antagonistic ratio isineffective at reducing tumor growth when compared to results at anon-antagonistic ratio (10:1 mole ratio) of the agents where tumorgrowth was substantially inhibited (see FIG. 14). These resultsthus highlight the importance of selecting drug combinations atratios that exhibit non-antagonistic effects over a range ofconcentrations in vitro. It should be noted that the drug dosesused in FIG. 16 (2.5 mg/kg cisplatin and 4.7 mg/kg daunorubicin)are actually higher than those used in FIG. 14 (2 mg/kg cisplatin,0.375 mg/kg daunorubicin).

Example 11

Synergism of Cisplatin and Topotecan

The procedure set forth above (see Example 1) for determiningsynergistic, additive or antagonistic effects was repeated usingcisplatin/topotecan, both at a 10:1 mole ratio and at a 1:1 moleratio. As shown in FIG. 17A, cisplatin/topotecan at al0:1 moleratio has a non-antagonistic interaction over a wide range of dosesthat affect 5% to 99% of cells (f.sub.a=0.05 to f.sub.a=0.99). Incontrast, cisplatin/topotecan at a 1:1 mole ratio was stronglyantagonistic over the same f.sub.a range (FIG. 17A).

This effect of concentration was also evidenced by calculating a CImaximum for various mole ratios of cisplatin/topotecan. As shown inFIG. 17B, an antagonist effect appears maximized at a 1:1 moleratio and non-antagonistic effects are apparent when either drug isin excess.

Example 12

Maintaining Synergism of Cisplatin and Topotecan In Vivo

Cisplatin and topotecan were formulated into DMPC/Chol andDSPC/Chol liposomes, respectively, and injected intravenously intomice at a 10:1 mole ratio identified in Example 11 to besynergistic.

Liposomal cisplatin was prepared by hydration of a lipid filmconsisting of DMPC and cholesterol (55:45 mol %) with a solutionconsisting of 150 mM NaCl and 8.5 mg/mL of cisplatin. The resultingMLVs were extruded at 80.degree. C. by ten passes through twostacked 100 nm pore size filters. After extrusion, the sample wascooled and precipitated cisplatin was removed by centrifugation.The remaining soluble cisplatin that was not encapsulated in theliposomes was removed by dialysis against HBS. After the removal ofnon-encapsulated cisplatin, the concentration of the drug wasmeasured by atomic absorption spectrometry.

Liposomal topotecan was prepared by hydration of a lipid filmcomposed of DSPC and cholesterol (55:45 mol %) with a solution of300 mM MnSO.sub.4. The resulting MLVs were extruded at 65.degree.C. by ten passes through two stacked 100 nm filters. Afterextrusion, the liposomes were exchanged into SHE buffer (300 mMsucrose, 20 mM HEPES and 30 mM EDTA, pH 7.4) by gel filtrationchromatography. Loading of topotecan was initiated by the additionof 1 .mu.g of A23187/.mu.mol lipid (A23187 is a cationic ionophorethat mediates the exchange of a divalent metal ion for two protonsacross a bilayer) and topotecan to a final topotecan/lipid ratio of0.08 (w/w), then holding the solution at 65.degree. C. for 15minutes. The extent of topotecan loading was measured by absorbanceat 380 nm after separation of encapsulated and non-encapsulateddrug using gel filtration chromatography and solubilization inTriton X-100.

The preparations were injected intravenously via the tail vein intoSCID/rag2 female mice. Doses of the liposomal formulations were 5mg/kg of cisplatin and 0.758 mg/kg of topotecan. At the indicatedtime points (3 mice per time point), blood was collected by cardiacpuncture and placed into EDTA coated microtainers. The samples werecentrifuged and plasma was carefully transferred to another tube.Liquid scintillation counting was used to quantitate radiolabeledlipid. Cisplatin was measured using atomic absorption spectrometrywhile topotecan was measured by fluorescence spectroscopy(excitation at 380 nm and emission at 518 nm) after disruption ofthe liposomes with excess detergent.

FIG. 18 (data points represent mean drug concentrations determinedin plasma +/-SD at the specified time points) shows that plasmalevels of cisplatin and topotecan were maintained at a 10:1 moleratio as plasma levels of cisplatin were roughly ten times that oftopotecan at various time points after intravenous administrationwhen they were delivered in the above-described liposomes. Theseresults demonstrate that the drug retention and liposomeelimination characteristics of two encapsulated agents in twodifferent liposomes can be coordinated such that coordinated drugelimination rates are realized. The inset of FIG. 18 shows that theplasma cisplatin-to-topotecan mole ratios (+/-SD) present in theplasma after intravenous administration vary little over time.

Cisplatin and topotecan can also be formulated in a single liposomein order to ensure non-antagonistic ratios are maintained in vivo.This may be carried out by passive entrapment of cisplatin followedby ionophore-mediated loading of topotecan. A cisplatin solution isfirst prepared by dissolving cisplatin powder into a solution of150 mM MnCl.sub.2. To maximize the solubility of cisplatin in theMnCl.sub.2 solution, the solution is heated to 65.degree. C. Alipid film composed of DSPC/Chol (55:45 mol %) along with traceamounts of .sup.3H-CHE is hydrated with the cisplatin/MnCl.sub.2solution. The resulting MLVs are extruded at 65.degree. C. throughtwo 100 nm filters for a total of ten passes. Insoluble cisplatinis then removed from the formulation by cooling the formulation toroom temperature and centrifuging the solution at 2000.times.g. Theresulting supernatant containing liposomal and soluble butunencapsulated cisplatin is dialyzed against SHE buffer, 300 mMsucrose, 20 mM HEPES, and 30 mM EDTA (pH 7.4) overnight at roomtemperature.

Topotecan is subsequently loaded into the liposomes using anionophore-mediated proton gradient. Drug uptake is performed at a0.08:1 drug to lipid weight ratio (w/w). The divalent cationionophore A23187 (1 .mu.g ionophore/.mu.mol lipid) is added to theliposomes, and then the mixture is incubated at 60.degree. C. for15 minutes to facilitate A23187 incorporation into the bilayer.Subsequently, topotecan is added, and the mixture is incubated at60.degree. C. for 60 minutes to facilitate drug uptake.Unencapsulated topotecan and A23187 are removed from thepreparation by dialyzing the sample against 300 mM sucrose. Theextent of topotecan loading is quantified by measuring absorbanceat 380 nm Cisplatin levels are measured by atomic absorptionspectrometry and lipid levels by liquid scintillation counting.

Example 13

Efficacy of Liposomal Cisplatin and Topotecan

The efficacy of cisplatin and topotecan loaded into separateliposomes was investigated by formulating the two drugs in separateliposomes and administering the formulation at a 10:1 mole ratioidentified in Example 11 as being non-antagonistic. Liposomalcisplatin was passively entrapped in DMPC/Chol (55:45 mol %)liposomes as described in the procedures of Example 12. Topotecanwas formulated in DSPC/Chol (55:45 mol %) as in Example 12 as well,except loading of topotecan was to a final topotecan/lipid weightratio of 0.1 (w/w). Following loading, the external buffer wasexchanged into HBS.

Efficacy studies were conducted as detailed in Example 26, whereH460 tumor bearing female SCID/rag2 mice (4 mice per group) weretreated intravenously (on days 13, 17, 21) with saline (control),free cocktail or a liposomal mixture of cisplatin/topotecan at a10:1 mole ratio identified as non-antagonistic in Example 11. Forboth the free and liposome-formulated treatments, the doses were1.6 mg/kg of cisplatin and 0.25 mg/kg of topotecan. Lipid doseswere 250 mg/kg arising from the cisplatin formulation plus 2.5mg/kg from the topotecan formulations.

FIG. 19 shows the results (data points represent mean tumorsize+/-SEM determined on the specified day). The saline control(solid circles) and the cocktail of cisplatin/topotecan 10:1 (solidtriangles) did not effectively arrest tumor volume. However, theliposomal preparation of cisplatin/topotecan 10:1 (open triangles)prevented the increase in tumor volume for a period of at least 35days.

Example 14

Synergism of Cisplatin and Irinotecan

Combinations of cisplatin and irinotecan at mole ratios of 1:1,10:1, 1:5 and 1:10 were tested for synergy, additivity orantagonism according to the methods described above (see Example1). Results summarized in FIG. 20A show that mole ratios of 10:1,1:5 and 1:10 were non-antagonistic over the complete range off.sub.a values whereas a 1:1 ratio was antagonistic over asubstantial range of f.sub.a values. FIG. 20B further illustratesthe dependency of the ratio on the nature of the combination effectas summarized by plotting the combination index maximum against thecisplatin to irinotecan mole ratio.

Example 15

Maintaining Synergism of Cisplatin and Irinotecan In Vivo

Cisplatin and irinotecan were co-loaded into DSPC/DSPG (80:20 mol%) liposomes, which were prepared as described in Example 5 exceptthat lipid films were rehydrated in 225 mM copper (75 mMCuCl.sub.2, 150 mM CuSO.sub.4, triethanolamine (TEA), pH 6.8)containing 6.0 mg/mL of cisplatin. The liposomal cisplatinconcentration after extrusion and removal of unencapsulated drugwas 0.025 mole cisplatin/mole lipid. The resulting liposomes weredialyzed against SHE, pH 6.8 overnight. Irinotecan was then addedto the preparation and the liposomes were incubated at 45.degree.C. for 1.5 hours. The liposomes loaded 60% of the added irinotecanas determined by HPLC. The liposomes were then buffer exchangedinto 0.9% saline by tangential flow. After tangential flow, theliposomes retained approximately 80% of the original cisplatin andirinotecan. Analysis of cisplatin and irinotecan, as determined byatomic absorption spectrometry and HPLC analysis, respectively,indicated that the final preparation had a cisplatin-to-irinotecanmole ratio of 1:3. SCID/rag2 mice were intravenously administered 2mg/kg cisplatin and 38.6 mg/kg irinotecan. At the indicated timepoints (3 mice per time point), blood was collected by cardiacpuncture and placed into EDTA coated microtainers. The samples werecentrifuged and plasma was carefully transferred to another tube.Plasma irinotecan and cisplatin levels were determined by HPLC andatomic absorption spectrometry, respectively.

Results in FIG. 21 (data points represent mean drug concentrationsdetermined in plasma+/-SD at the specified time points) show thatfollowing intravenous injection of formulations containingcisplatin and irinotecan, co-loaded into DSPC/DSPG liposomes, therates of drug elimination were comparable and non-antagonistic moledrug ratios could be maintained over the 24-hour time course afteradministration.

Coordinated release of liposomal cisplatin and irinotecan in vivowas also achieved by formulating the two drugs in separate deliveryvehicles and administering the drugs at a 1:5 mole ratio(cisplatin/irinotecan).

Liposomal cisplatin was prepared according to the passive loadingtechnique described above. Lipid films consisting of DMPC/Chol(55:45 mol %) were hydrated with a solution of 150 mM NaClcontaining 8.5 mg/mL cisplatin, then extruded as described above.The liposomes were collected in the supernatant aftercentrifugation as above then exchanged into HBS by tangential flowdialysis.

Liposomal irinotecan was prepared by hydrating lipid filmsconsisting of DSPC/DSPE-PEG2000 (95:5 mol %) with a solutionconsisting of 150 mM CuCl.sub.2, 20 mM histidine, pH 6.8 (pHadjusted with TEA). The resulting MLVs were extruded at 65.degree.C. through two stacked 100 nm pore size filters and bufferexchanged with HBS by tangential flow. The extruded liposomes wereloaded with irinotecan at 60.degree. C. for 1 minute at a 0.1:1drug to lipid weight ratio. The extent of loading of irinotecan wasdetermined by absorbance at 370 nm after solubilization in TritonX-100; lipid levels were measured by liquid scintillationcounting.

Liposomal cisplatin was administered to male SCID/rag2 mice at adrug dose of 2.0 mg/kg and liposomal irinotecan was administered tothe mice at 20 mg/kg. At the indicated time points (3 mice per timepoint), blood was collected by cardiac puncture and placed intoEDTA coated microtainers. The samples were centrifuged and plasmawas carefully transferred to another tube. Plasma irinotecan levelswere measured by HPLC and cisplatin was measured by atomicabsorption spectrometry.

Cisplatin and irinotecan administered together in these liposomalformulations at this synergistic ratio (1:5 mole ratio) maintainthis ratio at 1:5 following intravenous administration as evidencedby the plasma concentrations of irinotecan (nmoles/mL) beingroughly five times that of cisplatin (nmoles/mL) at various timepoints (FIG. 22).

Example 16

Efficacy of Liposomal Cisplatin and Irinotecan

Efficacy studies were carried out on liposomal cisplatin andirinotecan formulated into separate liposomes. Cisplatin waspassively entrapped in DMPC/Chol (55:45 mol %) liposomes andirinotecan was loaded into DSPC/DSPE-PEG2000 (95:5 mol %) liposomesas detailed in Example 15. Liposomal cisplatin and irinotecan wereco-administered to H460 tumor bearing SCID/rag2 mice according tothe methods described in Example 26 at a 1:5 mole ratio determinedto be non-antagonistic in Example 14. Liposomal cisplatin andirinotecan were administered (4 mice per group on days 14, 18 and22) at the non-antagonistic mole ratio of 1:5 with doses of 1 mg/kgcisplatin, 10 mg/kg irinotecan and 130 mg/kg lipid (open squares);2.5 mg/kg cisplatin, 25 mg/kg irinotecan and 175 mg/kg lipid (openupward triangles); or, 5 mg/kg cisplatin, 50 mg/kg irinotecan and250 mg/kg lipid (open inverted triangles). Freecisplatin/irinotecan was dosed at 1 mg/kg cisplatin and 10 mg/kgirinotecan which reflects a 1:5 mole ratio (solid squares).

FIG. 23 (data points represent mean tumor size+/-SEM determined onthe specified day) illustrates that tumor growth for the liposomalpreparations was substantially suppressed in relation to free drugcocktail and saline treated mice.

Example 17

Synergism of Drug and Lipid Combinations

Combinations comprising vinorelbine at a 1:1 mole ratio withvarious potentially therapeutic lipids incorporated into the lipidbilayer, such as POPS (inverted triangles), DPPS (upwardtriangles), DLPS (circles), DSPS (diamonds) or DOPS (squares), weretested for additive, synergistic or antagonistic effects using themethod described above (see Example 1).

Results in FIG. 24 show that all combinations of vinorelbine andlipids tested on H460 cells exhibit synergy over a substantialrange of f.sub.a values. In particular, the combinations ofvinorelbine with DLPS, DSPS and DOPS exhibit synergy at themajority of f.sub.a values, most notably between f.sub.a=0.2 tof.sub.a=0.8.

Example 18

Pharmacokinetics of Liposomal Vinorelbine andPhosphatidylserine

Liposomes consisting of SM/Chol/DPPS/DSPE-PEG2000 (35:45:10:10 mol%) were prepared and loaded with vinorelbine as follows:

Lipids were dissolved in chloroform at 100 mg/mL, and then combinedin the appropriate amounts. The exception to this is DPPS which wasdissolved at 25 mg/mL using CHCl.sub.3/methanol/H.sub.20/citratebuffer (20:10.5:1:1 v/v). Trace amounts of the radioactive lipid.sup.3H-CHE was added at this point to follow the lipid throughoutthe formulation process. The chloroform was removed under a streamof N.sub.2 gas until very little solvent remained. The resultinglipid films were left under vacuum overnight to remove any residualsolvent. The lipid films were rehydrated in citrate buffer (300 mM,pH 4.0) and the resulting MLVs were extruded at 65.degree. C.through two 100 nm pore size filters for a total of ten passes.

Vinorelbine was loaded into these formulations using the pHgradient loading method by titrating up the external buffer pH withthe use of 0.2 M Na.sub.2HPO.sub.4. A known amount of liposomeswere combined with the corresponding amount of vinorelbine (0.1drug/lipid weight ratio (w/w)) and incubated at 60.degree. C. for15 minutes. In order to establish a pH gradient, 0.2 MNa.sub.2HPO.sub.4 was added at ten times the volume of the citratebuffer. Vinorelbine was loaded into the liposomes to achieve avinorelbine/phosphatidylserine mole ratio that was identified asnon-antagonistic in Example 17.

The detergent OGP was used to solubilize the vinorelbine-loadedliposomes; drug levels were measured by absorbance at 270 nm andliquid scintillation counting was used to quantify lipid.

The resulting vinorelbine-loaded liposomes and free vinorelbinewere administered intravenously into SCID/rag2 mice at a drug doseof 10 mg/kg. At the indicated time points (3 mice per time point),blood was collected by cardiac puncture and placed into EDTA coatedmicrotainers. The samples were centrifuged and plasma was carefullytransferred to another tube. Blood was analyzed for remaining.sup.3H-CHE liposomal marker using scintillation counting. Plasmalevels of vinorelbine were assayed by HPLC.

FIGS. 25A and 25B show that SM/Chol/DPPS/DSPE-PEG2000 liposomesencapsulating vinorelbine exhibit substantially increased plasmadrug levels in relation to administration of free vinorelbine. Thefree vinorelbine mean area under the curve (AUC) of 0.112 .mu.gh/mL was increased to 125.3 .mu.g h/mL by formulation in theliposomes, representing a 1120 fold increase in mean AUC.

Example 19

Efficacy of Liposomal Phosphatidylserine and Vinorelbine in theH460 Human Lung Cancer Model

DSPC/Chol/DSPS/DSPE-PEG2000 (35:45:10:10 mol %),SM/Chol/DPPS/DSPE-PEG2000 (35:45:10:10 mol %) andDAPC/Chol/DPPS/DSPE-PEG2000 (35:45:10:10 mol %) liposomes wereprepared and loaded with vinorelbine as described in Example 18.Phosphatidylserine and vinorelbine were present in the liposomes ata non-antagonistic mole ratio (1:1). Efficacy studies were carriedout in the H460 human lung cancer model as described in Example26.

FIG. 26 shows for H460 tumor bearing mice (4 mice per group) givenintravenous administration of liposomes consisting ofDSPC/Chol/DPPS/DSPE-PEG2000 and SM/Chol/DPPS/DSPE-PEG2000 andencapsulated vinorelbine, that treatment engendered decreased tumorgrowth rates relative to those observed following treatment withfree vinorelbine and saline. Free vinorelbine was administered at 5mg/kg and liposomal vinorelbine was administered at a dose of 5mg/kg of the drug and 50 mg/kg lipid at 13, 17 and 21 days posttumor cell inoculation.

FIG. 27 (data points represent mean tumor size+/-SEM determined onthe specified day) shows that liposomes consisting ofSM/Chol/DPPS/DSPE-PEG2000; DAPC/Chol/DPPS/DSPE-PEG2000 andDSPC/Chol/DSPS/DSPE-PEG2000 and encapsulating vinorelbine displaydecreased tumor volume with time relative to free vinorelbine andsaline. Tumor-bearing mice (4 per group) were treated at avinorelbine dose of 5 mg/kg (free and liposomal) and a lipid doseof 50 mg/kg for the liposomal group. Mice were treatedintravenously on days 13, 17 and 21.

Example 20

Efficacy of Liposomal Phosphatidylserine and Vinorelbine in theMurine Leukemia Cancer Model

Liposomes consisting of SM/Chol/DPPS/DSPE-PEG2000 (35:45:10:10 mol%) were prepared and loaded with vinorelbine as described inExample 18, except that liposomes were extruded through a 100 nmpore filter stacked with an 80 nm filter.

P388/wt cells were inoculated intraperitonealy into BDF-1 mice asdescribed in Example 27. Subsequently, BDF1 female mice wereintraperitonealy administered one of the following: saline; freevinorelbine (10 mg/kg) and SM/Chol/DPPS/DSPE-PEG2000 liposomesloaded with vinorelbine (10 mg/kg vinorelbine and 100 mg/kg lipid).Intraperitoneal administration of free and liposomal vinorelbinewas carried out on day 1 with 4 mice per treatment group.

The survival curves shown in FIG. 28 demonstrate thatadministration of vinorelbine encapsulated in liposomes consistingof SM/Chol/DPPS/DSPE-PEG2000 results in substantially increasedsurvival rates in BDF-1 mice relative to free vinorelbine andsaline treatment.

Example 21

Co-Formulation of Sphingosine and Doxorubicin

Other therapeutic lipids besides phosphatidylserine may beincorporated into liposome membranes. For instance, sphingosine andsphingosine analogues are lipids that are amenable to formulationin liposomes and may be co-formulated with a therapeutic agent thatis encapsulated in the aqueous interior (for example, doxorubicin).The preparation of such a pharmaceutical composition (sphingosine)may be carried out as follows:

A typical liposomal formulation of sphingosine is composed ofDSPC/Chol/sphingosine (45:45:10 mol %). Lipid films are prepared asdetailed in the previous examples. The lipid films are rehydratedin citrate buffer (300 mM, pH 4) and the resulting MLVs areextruded at 65.degree. C. through two 100 nm filters for a total often passes. Doxorubicin is subsequently loaded into theseformulations using the pH gradient loading method by exchanging theexternal buffer of the liposomes by passage down a Sephadex G-50column that is equilibrated in HBS (pH 7.4) to establish a pHgradient.

The liposomes and doxorubicin solution are then incubated togetherat 60.degree. C. to allow loading to occur. To determine the extentof loading at various times, 100 uL of the sample is applied to a 1mL Sephadex G-50 spun column and then centrifuged. A drug to lipidratio for the spun column eluent is generated using liquidscintillation counting to quantitate lipid and absorbance at 480 nmto quantitate doxorubicin. To assay for drug, the liposomes aresolubilized by incubation in Triton X-100 before absorbancereadings are taken.

Example 22

Synergism of Floxuridine (FUDR) and Irinotecan (CPT-11)

The procedure set forth above for measuring additive, synergisticor antagonistic effects was repeated using FUDR/CPT-11 at 10:1,5:1, 1:1, 1:5 and 1:10 mole ratios in HT 29 cells. A combinationindex was determined for each dose by producing CI versus f.sub.acurves as described above. Data in FIG. 29, plotted as CI versusthe fraction of HT-29 cells affected, clearly illustrates theeffect of concentration on synergy. At a ratio of 5:1 or 1:1synergy is observed over the entire range of fraction affectedvalues (0.2 to 0.8) while a 10:1 ratio is non-antagonistic atf.sub.a values below 0.76 and a 1:5 mole ratio of FUDR/CPT-11 isnon-antagonistic at f.sub.a values less than 0.62. A 1:10 ratio isantagonistic over a substantial range of f.sub.a values (more than50%). Based on these results, a mole ratio of 1:1 FUDR:CPT-11 wasselected for formulation and efficacy studies as this ratiodemonstrated synergistic effects over a significant range off.sub.a values (at least 20% where greater than 1% of the cells areaffected). Formulations prepared at the 5:1 and 10:1 ratio wouldalso meet the requirements of a defined non-antagonistic ratio overa substantial range of f.sub.a values.

Example 23

Maintaining Synergism of FUDR and CPT-11 In Vivo

FUDR and CPT-11 were formulated into DSPC/DSPG/Chol (70:20:10 mol%) liposomes at a 1:1 mole ratio identified in Example A to besynergistic. Lipid films were prepared by dissolving DSPC andcholesterol in chloroform and DSPG in chloroform/methanol/water(16/18/1). The solutions were combined together such that thespecified mole ratio was achieved and trace quantities of.sup.14C-CHE were added as a liposomal lipid label. Followingsolvent removal the resulting lipid films were hydrated in asolution consisting of 250 mM CuSO.sub.4 and 25 mg/mL of FUDR (withtrace amounts of .sup.3H-FUDR) at 70.degree. C. The resulting MLVswere extruded at 70.degree. C. by ten passes through two stacked100 nm pore size filters. Subsequently, the liposomes were bufferexchanged into SHE, pH 7.4, by tangential flow dialysis, thusremoving any unencapsulated FUDR and CuSO.sub.4.

CPT-11 was added to these liposomes such that the FUDR to CPT-11mole ratio would be 1:1. Loading of CPT-11 into the liposomes wasfacilitated by incubating the samples at 50.degree. C. for 5minutes. After loading, the samples were exchanged into HBS, pH7.4, by tangential flow dialysis to remove EDTA or unencapsulateddrug. The extent of CPT-11 loading was measured using HPLC. FUDRand lipid levels were measured using liquid scintillation.

The preparations were injected intravenously via the tail vein intoBalb/c female mice. Doses of the liposomal formulations were 8.38mg/kg of FUDR and 20 mg/kg of CPT-11. At the indicated time points(3 mice per time point), blood was collected by cardiac punctureand placed into EDTA coated microtainers. The samples werecentrifuged and plasma was transferred to another tube. Liquidscintillation counting was used to quantitate radiolabeled lipidand FUDR in the plasma. CPT-11 plasma levels were quantified withHPLC.

FIG. 30 shows that plasma levels of FUDR and CPT-11 were maintainedat a 1:1 mole ratio as plasma levels of FUDR were roughly equal tothat of CPT-11 at various time points after intravenousadministration when they were delivered in the above-describedliposomes. Data points represent mean drug concentrations (nmolesdrug/mL plasma) determined in plasma+/-standard deviation at thespecified time points.

Example 24

Efficacy of Liposomal FUDR and CPT-11

DSPC/DSPG/Chol (70:20:10 mol %) liposomes co-encapsulated with FUDRand irinotecan at a mole ratio of 1:1 were prepared as described inExample B except that after drug loading the external liposomebuffer was exchanged to 0.9% NaCl.

Using the methods of Example 26, efficacy studies were carried outin female SCID/rag2 mice that had been inoculated subcutaneously inthe flank with 2.times.10.sup.6 HT-29 cells. Tumors were allowed togrow until they measured to be 180 mg (0.18 cm.sup.3) in size, atwhich time (day 21) the indicated formulations were injected. Tumorgrowth was determined by direct caliper measurements. Mice weretreated with a single dose (arrow) of saline, free drug cocktail ata 1:1 mole ratio or a liposomal formulation of FUDR/CPT-11 at a 1:1mole ratio. For both the cocktail and liposome-formulatedtreatments, the doses were 9.25 mg/kg FUDR and 25 mg/kg CPT-11.Lipid doses were 278 mg/kg lipid for liposome formulatedsamples.

Results presented in FIG. 31 show that administration of FUDR andCPT-11 encapsulated in a single liposome at a 1:1 mole ratioprovided significantly better therapeutic activity when compared toanimals injected with either the free drug cocktail or saline. Datapoints represent mean tumor size+/-standard error of the mean(SEM).

Example 25

Determination of CI for Various Three-Drug Combinations

Combinations comprising topotecan, cisplatin, HB5-5A (an analog ofedelfosine) and sphingosine were tested for additive, synergisticor antagonistic effects using the standard tetrazolium-basedcolorimetric MTT cytotoxicity assay (see Examples--CytotoxicityAssay). Combination effects were calculated using the median-effectanalysis described in the previous examples. CI versus f.sub.agraphs were created as described in the preceding examples and CIvalues corresponding to f.sub.a values at 0.50, 0.75 and 0.90(represented by ED50, 75 and 90) are reported in table below:

TABLE-US-00005 COMBINATION FIXED INDEX.sup.a AGENT 1 AGENT 2 AGENT3 RATIO ED.sub.50.sup.b ED.sub.75 ED.sub.90 Topotecan CisplatinHB5-5A 1:10:1 0.56 0.34 0.26 Topotecan Cisplatin HB5-5A 1:10:100.73 0.53 0.43 Topotecan Cisplatin HB5-5A 1:10:100 2.22 1.78 1.45Topotecan Cisplatin Sphingosine 1:10:1 0.23 0.12 0.07 TopotecanCisplatin Sphingosine 1:10:10 0.47 0.34 0.29 Topotecan CisplatinSphingosine 1:10:100 1.22 0.95 0.76 .sup.aCombination Index (CI) isused to determine synergy (CI < 0.9) or additivity (CI between0.9 and 1.1) based on the Chou Talalay theory of dose effectanalysis. Values are calculated using CalcuSyn Software..sup.bED.sub.50, ED.sub.75, ED.sub.90 refer to the dose of theagent(s) affecting 50, 75 or 90% of the measured response,respectively.

Example 26

Preparation of Tumor Models, Cell Preparation and Implantation fora Solid Subcutaneous Tumor Method

H460 human non-small cell lung carcinoma cells are obtained fromthe DCTC Tumor Repository of the NCI. The cells are maintained intissue culture for up to 20 passages. After 20 passages, new cellsare expanded from a frozen stock stored in liquid nitrogen. Whenthe cultured cells reached a confluence of 80-90% they are rinsedwith Hanks Balanced Salt Solution and the adherent cells areremoved with a 0.25% trypsin solution. Cells are counted on ahaemocytometer and diluted with media to a concentration of20.times.10.sup.6 cells/mL.

A patch of hair approximately 2 cm.times.2 cm is shaved usingelectric clippers in the lower back region of each mouse. Using a28 g needle, mice are inoculated subcutaneously with1.times.10.sup.6 tumor cells on day 0 (one inoculum/mouse) in avolume of 50 .mu.L.

When tumors reach a defined size of approximately 0.50-to-0.100cm.sup.3, either one-day prior to treatment or on the day oftreatment (.about.day 10-14), all tumors are measured. Afterselecting the appropriate tumor sizes, excluding tumors too smallor large, the tumors are randomly distributed (n=4) and the meantumor volume of the groups are determined.

Mice are organized into appropriate treatment groups and consist ofcontrol and treatment groups such as, saline control, vehiclecontrol, positive control and various dilutions of testarticles.

Treatment groups are as follows:

TABLE-US-00006 MICE/ DOSE VOLUME GROUP GROUP TREATMENT (MG/KG)SCHEDULE.sup.a INJECTION 1 4 Saline control N/A q4dx3 10 .mu.L/g 24 Vehicle control 20 q4dx3 10 .mu.L/g 3 4 Positive control 10 q4dx310 .mu.L/g 4 4 Test agent (low dose) 5 q4dx3 10 .mu.L/g 5 4 Testagent (medium dose) 10 q4dx3 10 .mu.L/g 6 4 Test agent (high dose)20 q4dx3 10 .mu.L/g .sup.aAlternative dosing schedules can beconsidered such as a single dose or 3 doses every 4-7 days

Mice are injected intravenously with the required volume of sampleto administer the prescribed dose (10 .mu.L/g as indicated) to theanimals based on individual mouse weights.

Tumor growth measurements are monitored using vernier calipersbeginning on the day of treatment. Tumor length measurements (mm)are made from the longest axis and width measurements (mm) will beperpendicular to this axis. From the length and width measurementstumor volumes (cm.sup.3) are calculated according to the equation(L.times.W.sup.2/2)/1000. Animal weights are collected at the timeof tumor measurement.

Individual mouse body weights are recorded at various days(generally two days apart such as Monday, Wednesday and Friday)during the efficacy study for a period of 14-days after the lastdosing.

All animals are observed at least once a day, more if deemednecessary, during the pre-treatment and treatment periods formortality and morbidity. In particular, signs of ill health arebased on body weight loss, change in appetite, rough coat, lack ofgrooming, behavioral changes such as altered gait, lethargy andgross manifestations of stress. Should signs of severe toxicity ortumor-related illness be seen, the animals are euthanized (CO.sub.2asphyxiation) and a necropsy is performed to assess other signs oftoxicity. Moribund animals must be terminated for humane reasonsand the decision to terminate will be at the discretion of theAnimal Care Technician and the Study Director/Manager. Any and allof these findings will be recorded as raw data and the time ofdeath will be logged as the following day.

Data are presented in either tabular or figure form as follows: 1.Plot of individual mouse tumor volumes with respect to each group,prior to treatment start and after grouping. 2. Mean body weightsfor each group as a function of time. 3. Mean tumor volumes foreach group as a function of time. 4. Raw data including figures andtables are generated and include tumor growth vs. time, tumorgrowth inhibition, and tumor growth delay. 5. Summary of abnormalor remarkable observations.

Example 27

Preparation of Tumor Models, Cell Preparation and Implantation foran Intraperitoneal Tumor Method

Mice are grouped according to body weight. Animals (n=4) areinoculated (Day=0) with 1.times.10.sup.6 P388 cells implanted inthe peritoneum cavity of BDF-1 mice in a volume of 500 .mu.L with a25 g needle. P388 cells from the ATCC tumor repository aremaintained as an ascitic fluid in the BDF-1 mouse, which arepassaged to new mice weekly. Mice are euthanized, and the asciticcells removed through the abdominal wall with a 20 g needle. Thecells used for experiment are used within passage 3-20. After 20passages in the mice, new cells are brought up from the frozenstock in liquid nitrogen, and mice are inoculated. For experiments,cells are rinsed with Hanks Balanced Salt Solution, counted on ahaemocytometer and diluted with HBSS to a concentration of2.times.10.sup.6 cells/mL.

Study groupings are performed randomly after all mice have beenadministered tumor cells. The required groupings are similar towhat is performed for solid tumor studies (see Example 26).

Mice are injected intravenously or intraperitonealy with therequired volume of sample to administer the prescribed dose (10.mu.L/g as indicated) to the animals based on individual mouseweights. With intraperitoneal tumors, administrations generallybegin 1-day post tumor cell inoculation.

Animal well-being is closely monitored daily. Signs of ill healthand progression of morbidity are closely monitored as described inExample 26. Animals are weighed at the time of examination.

Upon termination of any mice, gross necropsies are performed toevaluate the extent of tumor burden and/or physiologicallyobservable changes in organ appearances. Findings are recorded.

Group body weights are recorded Monday through Friday during theefficacy study for a period of 14 days after the last dosing.

All animals are observed at least once a day, more if deemednecessary, during the pre-treatment and treatment periods formortality and morbidity. In particular, signs of ill health arebased on body weight loss, change in appetite, behavioral changessuch as altered gait, lethargy and gross manifestations of stress.Should signs of severe toxicity or tumor-related illness be seen,the animals are terminated (CO.sub.2 asphyxiation) and a necropsyis performed to assess other signs of toxicity. Moribund animalsmust be terminated for humane reasons and the decision to terminatewill be at the discretion of the animal care technician and thestudy manager. These findings are recorded as raw data and the timeof death is logged on the following day.

Data is presented in tables or figures and includes mean bodyweights for each group as a function of time and increase inlife-span.

* * * * *