CA2264227A1 - Hydrolyzable prodrugs for delivery of anticancer drugs to metastatic cells - Google Patents

Abstract

Hydrolyzable prodrugs according to the present invention are activated by proteases located in the cell membranes of metastatic cells to yield active anticancer drugs that can be taken up by the metastatic cells. In general, a hydrolyzable prodrug according to the present invention comprises an aminoterminal capped peptide that is a substrate for a peptidohydrolase located on the surface of a metastatic cell covalently linked to a therapeutic drug through a self-immolating spacer of sufficient length to prevent the occurrence of steric hindrance. The therapeutic drug is typically an anticancer drug. The anticancer drug is typically doxorubicin, taxol, camptothecin, mitomycin C, or esperamycin. Typically, the peptidohydrolase that hydrolyses the substrate of the hydrolyzable prodrug is cathepsin B.

Description

1015202530W0 98/ 13059CA 02264227 1999-02-23PCTIUS97/17410HYDROLYZABLE PRODRUGS FOR DELIVERY OFANTICANCER DRUGS TO METASTATIC CELLS BACKGROUND OF THE INVENTIONThis invention is directed to hydrolyzable prodrugs for deliveryof therapeutic drugs to metastatic cells, particularly anticancer drugs.Metastasis is the hallmark of cancer. A tumor that does notmetastasize is termed "benign" because it poses a threat of survival that issmall compared to a "malignant" tumor that does metastasize (McGuire, Ne_wEng. J. Med., 320:525 (1989)).Metastasis involves a sequence of events that few cells cansuccessfully complete (Sanchez, Am. J. Med. Sci., 292:376 (1986); Poste,Nature, 283:139 (1980)). Metastatic cells must break away from the primarytumor, survive attack by the immune system during transit in the blood, lodgesomewhere while resisting the shear force of the moving blood stream,penetrate basement membrane to reach a safe haven where they canmultiply, and finally create a blood supply of their own when the demand fornourishment of the growing metastatic tumor exceeds what is availablelocally by diffusion. Metastatic cells are not a representative sample of thetumor (itself highly heterogeneous) (Poste, Ann. New York Acad. Sci.,397234 (1982); Heppner, Cancer Res., 4422259 (1984); Fidler, Science,2171998 (1982)) but rather constitute a small subpopulatlon that increaseswith the age of the primary tumor (Fidler, Cancer Res., 5016130 (1990);Science, 1972873 (1977); Kerbel, Int. J. Cancer, 487:118 (1991)). Eachmetastasis is clonal but rapidly diversities (Fidler, Cancer Treat. Flep., 682193(1984)).Although metastases are very small and thereforecomparatively accessible to chemotherapy, they are highly resistant topresent day drugs, for in spite of heavy medication, survival rates for e.g.,phases 2 and 3 breast carcinoma (lymph node involvement signalingspread) are very low compared to phase 1 (no spread). In the absence of- 1 -W0 98/130591015202530CA 02264227 1999-02-23PCT/US97/17410spread, the survival rate is 70% or greater; in the presence of spread, it isless than 10% (McGuire, New Eng. J. Med., 320:525 (1989)). Inhibiting ascontrasted with killing metastases can only prolong life a short whilebecause by the time cancer is typically diagnosed, metastasis has alreadytaken place. Therefore, there is a great need in cancer therapy for means ofspecifically eradicating metastatic foci, either after or before excision of theprimary tumor (Fisher, , 48321488 (1983); Jacquillat, ed., "Neo-Adjuvant Chemotherapy" (Libbey & Co., London, 1986); Bonadonna, J. Nat.Cancer |nst., 8221539 (1990); Fichtner, Anti‘-Cancer Res., 7:227 (1987)).Preferably, such anti-metastatic reagents should be usableagainst many types of metastases and not depend for their activity oncharacteristics of the primary tumor cells that might not be shared by themetastases. Preferably, such anti-metastatic reagents should be readilyabsorbed and lack toxicity, particularly in patients who are already subject toregimens consisting of multiple drug treatment.SUMMARYl have developed anti-metastatic reagents that meet theseneeds in the form of hydrolyzable prodrugs. In general, a hydrolyzableprodrug according to the present invention comprises an amino-terminalcapped peptide covalently linked to a therapeutic drug through a self-immolating spacer of sufficient length to prevent the occurrence of sterichindrance. The amino-terminal capped peptide is a substrate for apeptidohydrolase located on the surface of a metastatic cellTypically, the peptidohydrolase is cathepsin B or collagenaseIV. Preferably, the peptidohydrolase is cathepsin B.Typically, the amino-terminal capped peptide isbenzyloxycarbonylphenylalanyllysine, benzyloxycarbonylvalinyllysine, D-phenylalanylphenylalanyllysine, benzyloxycarbonylvalinylcitrulline,1-butyloxycarbonylphenylalanylysine, benzyloxycarbonyl-alanyllarginylarginine, benzyloxycarbonylphenylalanyl-N-tosylarginine,2--2-_ W0 98/ 1305910152025CA 02264227 1999-02-23PCT/US97Il7410aminoethylthio-succinimidopropionylvalinylcitrulline, 2—aminoethylthio-succinimidopropionyllysylphenylalanyllysine, acetylphenylalanyllysine, orbenzyloxycarbonylphenylalanyl-O-benzoylthreonine.Typically, the therapeutic drug is an anticancer drug.Preferably, the anticancer drug is doxorubicin, mitomycin C, taxol,esperamycin, or camptothecin. A particularly preferred anticancer drug isdoxorubicin.Typically, the spacer is Q-aminobenzyl carbonyl (“PABC”) or thebis-carbamate Q-NH-Ph-CH=(CH2OCO-)2 of bis(hydroxymethyl) styrene or“BHMS".The hydrolyzable prodrug can further comprise a peptidederived from a protein to which metastatic cells adhere in establishingcolonies covalently linked to the therapeutic drug. Typically, the peptide is aRGD—derived active peptide or a YIGSR-derived active peptide. Preferably,the peptide is YIGSR (SEQ ID N021) or GRGDS (SEQ ID NO:2).Preferred hydrolyzable prodrugs according to the presentinvention include benzyloxycarbonylphenylalanyllysyl-Q—aminobenzylcarbamoyldoxorubicin, acetylphenylalanyllysyl-Q-aminobenzylcarbamoyldoxorubicin, acetylphenylalanyllysyl-Q-aminobenzylcarbamoylmitomycin C, benzyloxycarbonylphenylalanyllysyl-Q—aminobenzylcarbonyl-7-paclitaxel, acetylphenylalanyllysyl-Q-aminobenzylcarbonylcamptothecin, 2-aminoethylthio-succinimidopropionyl—valinylcitrullinyl-BHMS-didoxorubicin, 2-aminoethylthio-succinimidopropionyl-lysylphenylalanyllysyl-BHMS-didoxorubicin,benzyloxycarbonylvalinyllysyl-Q-aminobenzyl carbamoyldoxorubicin, D-phenylalanylphenylalanyllysyl-Q-aminobenzyl carbamoyldoxorubicin, andbenzyloxycarbonylvalinylcitrullinyl-Q-aminobenzyl carbamoyldoxorubicin.Another aspect of the present invention is a method fordelivering a therapeutic drug to a metastatic cell comprising the steps of:W0 98/1305910152025CA 02264227 1999-02-23PCTIUS97/17410(1) contacting a hydrolyzable prodrug according to the presentinvention with a metastatic cell;(2) allowing the peptidohydrolase located on the surface of themetastatic cell to hydrolyze the hydrolyzable prodrug and release thetherapeutic drug from the prodrug; and(3) allowing the therapeutic drug to enter the metastatic cell.Yet another aspect of the present invention is a pharmaceuticalcomposition comprising:(1) a hydrolyzable prodrug according to the present invention;and(2) a pharmaceutically acceptable carrier.BRIEF DESCRIPTION OF THE DRAWINGSThese and other features, aspects, and advantages of thepresent invention will become better understood with reference to thefollowing description, appended claims and accompanying drawings where:Figure 1 is a depiction, showing structural formulas andreaction conditions, of the initial stages in the synthesis of the hydrolyzableprodrug Ac-Phe-Lys-PABC-Dox;Figure 2 is a similar depiction of the final stages in thesynthesis of Ac-Phe-Lys-PABC-Dox;Figure 3 is a similar depiction of the synthesis of thehydrolyzable prodrug Ac-Phe-PABC-MMC, beginning with an intermediatein the synthesis of Ac-Phe-Lys-PABC-Dox prior to the coupling of thedoxorubicin residue;Figure 4 is a similar depiction of the initial stages of thesynthesis of the hydrolyzable prodrug Z—Phe-Lys-PABC-Paclitaxel;-4-W0 98/1305910152025CA 02264227 1999-02-23PCT/US97l17410Figure 5 is a similar depiction of the final stages of thesynthesis of the hydrolyzable prodrug Z-Phe-Lys-PABC-Paclitaxel;Figure 6 is a similar depiction of the early stages of thesynthesis of the hydrolyzable prodrug CA—SP—Lys-Phe—Lys-BHMS-Doxz;Figure 7 is a similar depiction of the intermediate stages of thesynthesis of CA-SP-Lys-Phe-Lys-BHMS-Doxg;Figure 8 is a similar depiction of the final stages of thesynthesis of CA—SP-Lys-Phe-Lys-BHMS-Doxz;Figure 9 is a table showing the killing of BT-20 tumor cells,which are high—cathepsin B-secreting cells, and MCF-10A tumor cells, whichare low-cathepsin B-secreting cells, by several hydrolyzable prodrugs andcontrol compounds in the presence or absence of the cathepsin inhibitor CA-074;Figure 10 is a table showing the killing of BT-20 tumor cells atvarious times and prodrug concentrations with several hydrolyzableprodrugs and a control compound;Figure 11 is a table showing the killing of BT-20 and MCF-10Atumor cells at various times and prodrug concentrations with thehydrolyzable prodrug CA-SP-Lys-Phe-Lys-BHMS-Doxz; andFigure 12 is a table showing the stability of hydrolyzableprodrugs according to the present invention under physiological conditionsin the presence or absence of cathepsin B.DESCRIPTIONOne new approach to developing reagents specific formetastatic cells takes advantage of the properties of the metastatic cellsthemselves, particularly those properties possessed by the metastatic cellsthat allow them to spread through the body and adhere to specific tissues.-5-W0 98/ 130591015202530CA 02264227 1999-02-23PCT/US97/17410One such property is the ability of metastatic cells to penetratebasement membrane (Sanchez, Am. J. Med. Sci, 2922376 (1986); Poste, , 2831139 (1980)), shared only by the peripheral cells of the primarytumor (Poole, Nature, 2731545 (1978); Shamberger, Nature, 2132617 (1967);Grat, , 451587 (1981 ); Baici, , 4:13 (1984); Duffy, E_ur_._J_.Cancer Clin. Oncol., 23:583 (1987)) and by virtually no normal body cells.Metastatic cells do this by means of hydrolytic enzymes whichthey secrete into the medium (Duffy, Eur. J. Can. Clin. Oncol., 23:583 (1987);MacKay, Cancer Res., 50:5997 (1990); Goldfarb, Sem. Thromb. Hemostas.,12:294 (1986); Pietras, Gynec. Oncol., 7:1 (1979)) or in their plasmamembranes (Sylven, Virchows Arch. B., 17:97 (1974); Sloane, Biomed.Biochim. Acta, 50:549 (1991); Keren, Cancer Res., 48:1416 (1989); Pietras,J. Biol. Chem., 25628536 (1981); Weiss, Proc. Am. Assoc. Cancer Res., 31:73(1990); Sloane, Bio. Chem. Hoppe—Seyler, 371 Suppl.:193 (1990); Rozhln,Cancer Res., 47:6620 (1987); Sloane, Proc. Natl. Acad. Sci., 8312483(1986)). Tumors are proteolytic (Fischer, Arch. Entw. Mech. Arq., 1042210(1925); Duffy, Eur. J. Cancer Clin. Oncol., 23:583 (1987); Strauli et al., ed.,"Proteinases and Tumor lnvasion" (Raven Press, New York, 1980)), and thispower is correlated with metastatic propensity (Duffy, Eur. J. Cancer Clin.Oncol., 23:583 (1987); Sylven, Virchows Arch. B., 17:97 (1974); Sloane,Biomed. Biochim. Acta, 50:549 (1991); Keren, Cancer Res., 48:1416 (1989);Pietras, J. Biol. Chem., 25628536 (1981); Weiss, Proc. Am. Assoc. CancerBeg, 31:73 (1990); Sloane, Bio. Chem. Hoppe-Seyler, 371 Suppl.:193(1990); Rozhln, Cancer Res., 47:6620 (1987); Sloane, Proc. Natl. Acad. Sci.,83:2483 (1986); Sheahan, Cancer Res., 49:3809 (1989); Koppel, Exp. CellB_i_oL, 52:293 (1984); Sloane, Science, 21211151 (1981); Sloane, ’£4 4513636 (1985); Nakajima, Cell. Biochem., 362157 (1988); Nakajima,Science, 220:611 (1983); Qian, Cancer Res., 4914870 (1989); Hendrix,Molec. Cell Probes, 6:59 (1992)). Enzymes known to be secreted bymetastasizing cells include cathepsin B (Sloane, Biomed. Biochim. Acta,50:549 (1991); Keren, Cancer Res., 48:1416 (1989); Pietras, J. Biol. Chem.,25628536 (1981); Weiss, Proc. Am. Assoc. Cancer Res., 31 :73 (1990);Sloane, Bio. Chem. Hoppe—Seyler, 371 Suppl.:193 (1990); Rozhln, Cancer- 5 _W0 98/130591015202530CA 02264227 1999-02-23PCT/US97l17410fies, 4726620 (1987); Sloane, , 8322483 (1986);Sheahan, Cancer Res, 4923809 (1989); Koppel, Exp. Cell Biol., 522293(1984); Sloane, Science, 21221151 (1981); Sloane, Cancer Res., 4523636(1985); Nakajima, , 36:157 (1988); Nakajima, Science,220:611 (1983); Qian, , 4924870 (1989)), cathepsin D(Montcourrier, Cancer Res, 5026045 (1990); Vasishta, Brit. J. Surg., 722386(1985)), cathepsin L (Rozhin, Biochem. Biophys. Res. Comm., 1642556(1989); Vasishta, 721386 (1985)), cathepsin H (Vasishta, §_u_r_g_., 722386 (1985)), collagenase lV (Goldfarb, Sem. Thromb. Hemostas.,122294 (1986); Hendrix, , 6:59 (1992)), urokinase—typeplasminogen activator (Goldfarb, , 122294 (1986);Marutsuka, Inv. Metas. 112181 (1991)), B-glucuronidase (heparanase)(Rozhin, Cancer Res., 4726620 (1987); Nakajima, Cell. Biochem., 36:157(1988); Nakajima, §gi_e_n_c_e, 2202611 (1983)), gelatinase (Aoyama, _F?_rc&Natl. Acad. Sci., 8728996 (1990)), guanldinobenzoatase (Steven, Ami;Cancer Res, 112143 (1991)); and undefined tryptic enzymes (Goldfarb, Sem.Thromb. Hemostas., 122294 (1986)).increased enzyme secretion by malignant cells vis-a-vis nearby normal cellsfrom the same patient (Poole, Nature, 2732545 (1978); Sheahan, CancerR_e§,, 4923809 (1989); Flozhin, Biochem. Biophys. Res. Comm., 1642556(1989); Watanabe, Hepato-Gastro—Enterology, 342126 (1987); Murnane,Cancer Res., 5121137 (1991); Chauhan, Cancer Res., 5121478 (1991);Durdey, Brit. J. Surg., 722378 (1985); Sedo, J. Cancer Res. Clin. Oncol.,Many studies have reported greatly1172249 (1991); Lah, Int. J. Cancer, 50:36 (1992); Dengler, Biomed. Biochim.Acta, 502555 (1991)).A potential application of these enzymes is to unmask cytotoxicdrugs at the sites of metastatic foci. Preferred drugs are those that arereadily ingested by cells such as doxorubicin (Dox) (also known asadriamycin (ADM)), mitomycin C (MMC), taxol, camptothecin (CPT), andesperamycin, as well as derivatives of these drugs. Typically, suchanticancer drugs have substantial hydrophobic moieties so that they canpass through the plasma membrane of metastatic cells. Other anticancerdrugs can be derivatized with appropriate hydrophobic moieties to improve- 7 -WO 98/130591015202530CA 02264227 1999-02-23PCT/US97l174l0their permeability to the lipid bilayer of the plasma membrane of themetastatic cell.A particularly useful enzyme with prodrugs according to thepresent invention is cathepsin B, principally because it is an exopeptidase(Takahashi, J. Biol. Chem., 26119375 (1986); Koga, J. Biochem., 110:179(1991)) for which numerous peptide substrates are already known (Dingle,ed., "Lysosomes" (North-Holland, Amsterdam, 1977); Strauli et al., eds.,"Proteinases and Tumor lnvasion" (Raven Press, New York, 1980);Neuberger, ed., "Hydrolytic Enzymes" (Elsevier, Amsterdam, 1987)).Cathepsin B is a lysosomal enzyme that is ubiquitous within cells (Dingle,ed., "Lysosomes" (North-Holland, Amsterdam, 1977); Strauli et al., eds.,"Proteinases and Tumor invasion" (Raven Press, New York, 1980);Neuberger, ed., "Hydrolytic Enzymes" (Elsevier, Amsterdam, 1987)), butalmost never secreted normally. If small amounts of cathepsin B do escapeduring exocytosis or unprogrammed cell death, it loses all activity within 15minutes at neutral pH (Sheahan, Cancer Res., 4923809 (1989); Dingle, ed.,"Lysosomes" (North Holland Publishing Co., Amsterdam, 1977); Buck,Biochem. J., 2822273 (1992)). There is a strong association of cathepsin Bwith cancer as opposed to normal cells (Duffy, Eur. J. Cancer Clin. Oncol.,232583 (1987); Goldfarb, Sem. Thromb. Hemostas., 12:294 (1986); Pietras,Gynecol. Oncol., 7:1 (1979)), often within the same patient (Poole, ,2731545 (1978); Watanabe, Hegato-Gastro-Enterology, 342126 (1987);Murnane, Cancer Res., 51:1137 (1991); Durdey, Brit. J. Surg., 72:378(1985); Sedo, J. Cancer Res. Coin. Oncol., 117:249 (1991); Lah, |_m._J_.Cancer, 50:36 (1992); Dengler, Biomed. Biochim. Acta, 502555 (1991)), andits secretion correlates with the metastatic propensity of cells from patientsand animals, and with the degree of malignancy of their disease (Duffy, E_ur_.J. Cancer Clin. Oncol., 232583 (1987); Goldfarb, Sem. Thromb. Hemostas..121294 (1986); Pietras, Gynecol. Oncol., 7:1 (1979); Sloane, Biomed.Biochim. Acta, 502549 (1991); Keren, Cancer Res., 48:1416 (1989); Pietras,J. Biol. Chem., 256:8536 (1981); Weiss, Proc. Am. Assoc. Cancer Res., 31:73(1990); Sloane, Bio. Chem. Hogpe-Seyler, 371 SuppI.:193 (1990); Rozhin,Cancer Res., 47:6620 (1987); Sloane, Proc. Natl. Acad. Sci., 8322483_ 8 -WO 98/130591015202530CA 02264227 1999-02-23PCT/US97/ 17410(1986); Sheahan, , 49:3809 (1989); Koppel, Exp. Cell Biol,522293 (1984); Sloane, SE9; 21221151 (1981); Sloane, Cancer Res.,4523636 (1985); Nakajima, , 36:157 (1988); Nakajima,Science, 220:611 (1983); Qian, , 4914870 (1989) Lah, milCancer, 50:36 (1992)). At pH 7, cathepsin B degrades a number ofextracellular matrix components including type IV collagen (Buck, Biochem.¢, 2822273 (1992); Maciejewicz, , 502561 (1991);Maciejewicz, , 43:478 (1989)), laminin (Buck. Biochem. J.,2821273 (1992)), and fibronectin (Buck, Biochem. J., 2822273 (1992)).Although not normally stable at pH 7 as opposed to pH 5, cathepsin B fromtumors is stable at pH 7 (Sheahan, , 4913809 (1989); Buck,Biochem. J., 2822273 (1992); Sloane, Cancer Metastas. Rev., 3:249 (1984)),especially when kept within the cells‘ plasma membranes (Rozhin, Cancer3, 4726620 (1987)).their plasma membranes.In fact, malignant cells do secrete cathepsin B intoThe plasma membrane is obviously a highly desirable place toactivate a latent cytotoxic drug which should be delivered, not only asdirectly as possible to the target cells, but also to neighboring cancer cellsthat might not display as much cathepsin B as the majority, owing to the highgenetic instability (Fidler, Cancer Treat. Rep., 681193 (1984)) ofmetastasizing cells.I. HYDROLYZABLE PRODRUGSA preferred reagent for delivering therapeutic drugs tometastasizing cells according to the present invention is a hydrolyzableprodrug comprising an amino-terminal capped peptide covalently linked to atherapeutic drug through a self-immolating spacer of sufficient length toprevent the occurrence of steric hindrance. The amino-terminal cappedpeptide is a substrate for a peptidohydrolase located on the surface of ametastatic cell.Typically, the peptidohydrolase is cathepsin B or collagenaseIV; preferably, the peptidohydrolase is cathepsin B.-9-W0 98/ 130591015202530CA 02264227 1999-02-23PCTIUS97I17410If the peptidohydrolase is cathepsin B, the amino-terminalcapped peptide that acts as a substrate for the peptidohydrolase is typicallyone of benzyloxycarbonylphenylalanyllysine,benzyloxycarbonylvalinyllysine, D-phenylalanylphenylalanyllysine,benzyloxycarbonylvalinylcitrulline, 1-butyloxycarbonylphenylalanylysine,benzyloxycarbonylalanyllarginylarginine, benzyIoxycarbonylphenylalanyl-N-tosylarginine,2-aminoethylthio-succinimidopropionylvalinylcitrulline, 2-aminoethylthio-succinimidopropionyllysylphenylalanyllysine,acetylphenylalanyllysine, and benzyloxycarbonylphenylalanyl-O-benzoylthreonine. Derivatives of these peptides in which the derivatizinggroups do not interfere with the cleavage of the peptide by cathepsin B canalso be used. Alternatively, the amino-terminal capping groups of thesepeptides can be replaced by others known in the art.Preferably, the amino-terminal capped peptide isbenzyloxycarbonylphenylalanyllysine or acetylphenylalanyllysine.Other suitable substrates for cathepsin B are known. Forexample, substrates containing paired basic residues can be hydrolyzed bycathepsin B (J.K. McDonald & S. Ellis, Life Sci., 17212694276 (1975)).For hydrolyzable prodrugs according to the present invention,the amino-terminal residue must be “capped” or protected with a protectinggroup. Such protecting groups are well—known in peptide chemistry andinclude, for example, benzyloxycarbonyl (also known as carbobenzoxy andgenerally abbreviated as Z), acetyl, 2-aminoethylthio-succinimidopropionyl,1-butyloxycarbonyl, and other amino-terminal protecting groups such asthose disclosed in M. Bodanszky, “Principles of Peptide Synthesis” (2d Ed.,Springer-Verlag, Berlin, 1993). These groups include triphenylmethyl, Q-methoxybenzyloxycarbonyl, adamantyloxycarbonyl,biphenylylisopropyloxycarbonyl, formyl, isonicotinyloxycarbonyl, Q-nitrophenylsulfenyl, 9-fluorenylmethyloxycarbonyl, derivatives ofbenzyloxycarbonyl substituted on the aromatic ring of the benzyl group, or, insome cases, in which the phenyl moiety of the benzyl group is replaced with-10-W0 98/ 1305910152025CA 02264227 1999-02-23PCTIUS97/17410another fully aromatic moiety such as furan or pyridine, phthaloyl,dithiasuccinyl, Q-toluenesulfonyl (tosyl), and other groups.Alternatively, the amino-terminal protecting group can be a D-amino acid such as D-phenylalanine. When the amino-terminal protectinggroup is a D-amino acid, the carboxyl group of the D-amino acid forms apeptide bond with the amino-terminal residue of the amino-terminalprotected peptide.Preferably, when the peptide portion of the amino-terminalprotected peptide is phenylalanyllysine, the protecting group isbenzyloxycarbonyl or acetyl, so that the amino-terminal protected peptide isbenzyloxycarbonylphenylalanyllysine or acetylphenylalanyllysine.The hydrolyzable prodrug includes a spacer of sufficient lengthto prevent the occurrence of steric hindrance between the amino-terminalprotected peptide and the therapeutic drug. If the spacer is too short, thetherapeutic drug may prevent the binding of the substrate for thepeptidohydrolase to the active site of the peptidohydrolase by sterichindrance. A particularly suitable spacer is Q-aminobenzyl carbonyl(“PABC”). This has an approximate length of 10 angstroms.Derivatives of Q-aminobenzyl carbonyl can also be used, suchas compounds substituted on the aromatic moiety of the benzyl group.Alternatively, other spacer groups can be used. The length ofthe spacer should be greater than about 10 angstroms; spacers ofsignificantly greater length can be used, and can incorporate, for example,additional aliphatic or aromatic moieties. in general, such spacers should berelatively unbranched so as not to introduce steric hindrance of their own.The chemical functionality terminating the spacer can vary but one end isable to react with the carboxyl-terminal residue of the substrate for thepeptidohydrolase. Typically, this is an amino group. The other functionalityterminating the spacer is capable of reacting with the therapeutic drug. in-11-WO 98/130591015202530CA 02264227 1999-02-23PCT/US97/17410one preferred embodiment, this functionality reacts at an amino group ofthedrug to form a carbamate or urethane linkage as part of the spacer.Another preferred spacer is the bis-carbamate ofbis(hydroxymethyl)styrene (“BHMS”), which has the structure p-NH-Ph-CH=(CH2OCO-)2. This spacer also reacts at an amino group of the drug toform a carbamate or urethane linkage. This spacer is bifunctional and canbind two drug moieties such as doxorubicin.Preferably, such spacers have the property of self-immolation.A self-immolating spacer is one in which the residual portion of the spacerattached to the therapeutic drug subsequent to the hydrolysis of the peptidebond by the peptidohydrolase is then further cleaved by spontaneous,nonenzymatic hydrolysis in an aqueous medium to restore the originalunconjugated drug. Both PABC and BHMS are self-immolating. Forexample, if the spacer is Q-aminobenzyl carbonyl (“PABC”) and thetherapeutic drug is doxorubicin, the portion of the spacer remaining attachedto the drug after hydrolysis of the peptide bond by cathepsin B, thensubsequently undergoes spontaneous hydrolysis to Q-aminobenzyl alcohol,carbon dioxide, and doxorubicin.Typically, the therapeutic drug is an anticancer drug. However,other therapeutic drugs can be incorporated into hydrolyzable prodrugsaccording to the present invention and can be delivered to metastatic cells.Preferred anticancer drugs incorporated into hydrolyzable prodrugsaccording to the present invention include doxorubicin, taxol, camptothecin,mitomycin C, and esperamycin, as well as derivatives thereof. Otheranticancer drugs that have hydrophobic moieties that allow them to be takenup efficiently by metastatic cells or that can be derivatized with such moietiescan also be used.Therefore, particularly preferred hydrolyzable prodrugsaccording to the present invention includebenzyloxycarbonylphenylalanyllysyl-Q-aminobenzyl carbamoyldoxorubicin,acetylphenylalanyllysyl-Q-aminobenzyl carbamoyldoxorubicin,-12-W0 98/ 130591015202530CA 02264227 1999-02-23PCT/US97/17410acetylphenylalanyllysyl-Q-aminobenzyl carbamoylmitomycin C,benzyloxycarbonylphenylalanyllysyl-Q-aminobenzyl carbonyl-7-paclitaxel,acetylphenylalanyllysyl-Q-aminobenzyl carbonylcamptothecin, 2-aminoethylthio-succinimidopropionyl-valinylcitrullinyl-bis(hydroxymethyl)~styryl-bisdoxorubicin, 2-aminoethylthio-succinimidopropiony|-lysylphenylalanyllysyl-bis(hydroxymethyl)styryl-bisdoxorubicin,benzyloxycarbonylvalinyllysyl-Q-aminobenzyl carbamoyldoxorubicin, D-phenylalanylphenylalanyllysyl-Q-aminobenzyl carbamoyldoxorubicin, andbenzyloxycarbonylvalinylcitrullinyl-g-aminobenzyl carbamoyldoxorubicin.Synthesis of hydrolyzable prodrugs according to the presentinvention can be accomplished by condensation reactions that are wellknown in the art. Examples of syntheses are given below in Examples 1-6.in general, using the spacer PABC, the synthetic procedure comprises: (1)synthesizing the peptide that is the substrate of the peptidohydrolase byconventional peptide synthetic techniques, with the oc-amino group of theamino-terminal amino acid residue protected and appropriate protection forreactive side chains of the amino acids; (2) linking the Q-aminobenzyl moietyto the carboxyl group of the carboxyl-terminal amino acid; (3) activating thep_-aminobenzyl moiety for covalent linkage of the therapeutic drug; (4)covalently linking the therapeutic drug, which may have certain reactive sidechains protected as well; and (5) removing the remaining protecting groupson the peptide and the therapeutic drug.In general, using the spacer BHMS, the synthetic procedurecomprises: (1) synthesizing a peptide that is the substrate of thepeptidohydrolase by conventional peptide synthetic techniques, withappropriate protective groups as above; (2) linking the BHMS moiety to thecarboxyl group of the carboxyl-terminal amino acid of the peptide; (3)activating the BHMS moiety for coupling of the therapeutic drug; (4) couplingthe therapeutic drug to the activated BHMS moiety; (5) removing anyprotecting groups on amino acid side chains, such as the e-amino group oflysine; and (6) modifying the amino-terminal blocking group so that thedesired capping group is present.-13-W0 98/130591015202530CA 02264227 1999-02-23PCT/US97/17410As another alternative, if the substrate for cleavage by the ipeptidohydrolase is a tripeptide or peptide with more than three amino acids,the peptide can be extended at its amino-terminus after the coupling of thecarboxyl-terminus to the spacer. This involves removing the amino-terminalprotecting group of the peptide, activating the carboxyl group of the aminoacid to be added, and coupling it to the deblocked amino group to form apeptide bond. This step can be repeated if a longer peptide is desired.Then, the therapeutic drug is coupled to the completed peptide andsynthesis of the hydrolyzable prodrug is completed as above.In one embodiment according to the present invention, thehydrolyzable prodrug further comprises a peptide derived from a protein thatadheres to metastatic cells. Such peptides include peptides derived fromthe protein fibronectin (GFlGDS)(SEQ ID NO:1)(Humphries, Science2332467 (1986); Olden, Ann. New York Acad. Sci. 421 (1989); Dedhar, Q§s_._sa3Ls, 12:583 (1990)) and Iaminin (Y|GSFt)(SEQ ID NO:2)(lwamoto,Science, 23821132 (1987); Saiki, Brit. J. Cancer 59:194 (1989)).Typically, the peptide is covalently linked to the amino side ofthe amino-terminal capped peptide that is the substrate for thepeptidohydrolase, either directly as an amide, or indirectly via attachment tothe capping group,benzyloxycarbony|, acetyl, maleimidopropionyl or the like,which has been modified to accept the peptide. The linkage can be througheither the amino or carboxyl group of the peptide, or, in some cases, throughfunctional groups of the peptide such as the carboxyl of aspartic acid, thehydroxyl of serine or other functional groups of other residues. Dependingupon the peptide chosen and the therapeutic drug, various cross-linkingreagents can be used. For example, carbodiimides, such asdicyclohexylcarbodiimide, can form an amide linkage between a carboxylgroup and an amine. Other reactive groups are known and are described,for example, in G.T. Hermanson, “Bioconjugate Techniques” (AcademicPress, San Diego, 1996), in S.S. Wong, “Chemistry of Protein Conjugationand Crosslinking” (CRC Press, Boca Ftaton, Fla. 1991), and in T.E.Creighton, Ed., “Protein Function: A Practical Approach” (IRL Press, Oxford-14-WO 98/130591015202530CA 02264227 1999-02-23PCT/US97/174101989). The peptide can be linked, for example, to a maleimidopropionylcapvia a cysteine whose thiol group is added to the maleimido group, or to aglycine or other amino acid cap (replacing the acetyl cap) via acylation of theamino group of the glycine, or by attaching the peptide to the Q-position ofthe benzyloxycarbonyl cap, or by other methods known to the art. Thelinkage of these peptides such as YIGSR (SEQ ID NO: 2) to the prodrug maybe with or without intervening links which might or might not consist of otheramino acids.Other conjugation or crosslinking methods can also be used toattach the peptides such as the peptides from Iaminin or fibronectin to otherportions of the hydrolyzable prodrug,. In most cases, attaching the peptide tothe self-immolating spacer would either prevent hydrolysis of the substrateby the peptidohydrolase or result in steric hindrance.A number of peptides derived from the fibronectin and lamininpeptides can be linked to the hydrolyzable prodrugs. These peptides can beclassified in terms of their structure and homology to the fibronectin orIaminin sequence as follows:Fibronectin-derived peptides include: (1) the GRDGS (SEQ IDNO: 1) pentapeptide derived from the fibronectin sequence (I. Hardan et al.,“inhibition of Metastatic Cell Colonization in Murine Lungs and Tumor-Induced Morbidity by Non-Peptidic Arg-Gly-Asp Mimetics,” Int. J. Cancer 55:1023-1028 (1993)); (2) derivatives of the fibronectin pentapeptide sequencewith conservative amino acid substitutions, such as GRGES (SEQ ID NO: 3)(Ft.J. Tressler et al., “Correlation of Inhibition of Adhesion of Large CellLymphoma and Hepatic Sinusoidal Endothelial Cells by RGD-ContainingPeptide Polymers with Metastatic Potential: Role of lntegrin-Dependent and -Independent Adhesion Mechanisms,” Cancer Commun. 1: 55-63 (1989)); (3)truncated peptides derived from this sequence, such as RGD and RGDS(SEQ ID NO: 4) (I. Saiki et al., “Anti-Metastatic and Anti-Invasive Effects ofPolymeric Arg-Gly-Asp (RGD) Peptide, Poly(RGD), and its Analogues,”Japan J. Cancer Res. 81: 660-667 (1990)); (4) truncated peptides with amino-15-. ..2...... ......_.............................................,,....,s .. . . , . . , . ~....,...,......................,..m.........,...g,,.-,.c~.~....,....,W0 98/130591015202530CA 02264227 1999-02-23PCT/US97/17410acid substitutions, such as RGDT (SEQ ID NO: 5) (I. Saiki et al. (1990),§up_r_a_), as well as RGE and RGET (SEQ ID NO: 6); (5) extended peptidesderived from the fibronectin sequence, such as GRGDSP (SEQ ID NO: 7)(M.D. Pierschbacher & E. Ruoslahti, “lnfluence of Stereochemistry of theSequence Arg-Gly-Asp-Xaa on Binding Specificity in Cell Adhesion,” d_._BiolChem. 262: 17294-17298 (1987)) and GRGDSPA (SEQ ID NO: 8) (H.Kumagai et al., “Effect of Cyclic RGD Peptide on Cell Adhesion and TumorMetastasis,” Biochem. Biobhvs. Res. Commun. 177: 74-82 (1991); (5)extended peptides with amino acid substitutions, such as GRGDXPC, whereX is a naturally-occurring L-amino acid other than M, C, H, Y, G, or P (M.D.Pierschbacher & E. Ruoslahti (1987), ;s_u;&1_), GRGDNPC (SEQ ID NO: 9) (A.Hautanen et al., “Effects of Modification of the RGD Sequence and ItsContext on Recognition by the Fibronectin Receptor,” J. Biol. Chem. 264:1437-1442 (1989)), GRGDAPC (SEQ ID NO: 10) (M.D. Pierschbacher & E.Ruoslahti, “Variants of the Cell Recognition Site of Fibronectin that RetainAttachment-Promoting Activity,” Proc. Natl. Acad. Sci. USA 81: 5985-5988(1984)), GRGDXPA, where X is a naturally-occurring L-amino acid other thanM, C, H, Y, G, or P (by analogy to results reported in M.D. Pierschbacher & E.Ruoslahti (1987), sutga, and H. Kumagai et al. (1991), flog), GRGDSG(SEQ ID NO: 11) (by analogy to results with branched peptides reported inM. Nomizu et al., "Multimeric Forms of Tyr-Ile-Gly-Ser-Arg (YIGSR) PeptideEnhance the Inhibition of Tumor Growth and Metastasis,” Cancer Res. 53:3459-3461 (1993)), GRGDXG, where X is a naturally-occurring L-amino acidother than M, C, H, Y, G, or P (by analogy to results reported in M.D.Pierschbacher & E. Ruoslahti (1987), supg, and M. Nomizu et al. (1993),;up_na_), and GRDGXPA, where X is a naturally-occurring L-amino acid otherthan M, C, H, Y, G, or P (by analogy to results reported in M.D. Pierschbacher& E. Ruoslahti (1987), supra, and in H. Kumagai et al. (1991), §L_iQr_a), as wellas analogous peptides in which the D residue in the fourth position isreplaced by an E (by analogy to results of R.J. Tressler et al. (1989), supg);(6) extended substituted peptides with a D-amino acid replacing one of thenaturally-occurring L-amino acids, such as G(dR)GDSP and GRGD(dS)P(M.D. Pierschbacher & E. Ruoslahti (1987), _s_up_ga_); (7) cycllzed peptides,-16-W0 98/130591015202530CA 02264227 1999-02-23PCT/US97l174l0including c(GRGDSPA), c(GRGDSP), c(GRGDS), c(GRGD), and c(RGDS)(H. Kumagai et al. (1991), _s_up_ra_); (8) cyclized peptides with a D-amino acidreplacing one of the naturally-occurring L-amino acids, such as c(RGD(dF)V)and c(FlGDF(dV) (M. Aumailley et al., “Arg—Gly-Asp Constrained Within CyclicPentapeptides,” FEBS Lett. 291: 50-54 (1991)); (9) oligomers of peptides,including oligo (RGD) (1.5 kDa molecular weight) (J. Murata et al.,“Molecular Properties of Poly(RGD) and Its Binding Capacities to MetastaticMelanoma Cells,” Int J. Peptide Protein Res. 38: 212-217 (1991), (GFtGDS),,,(SEQ ID NO: 12) (GFlGES),,, (SEQ ID NO: 13) and(GRGDS)(GRGES)2(GFlGDS) (SEQ ID NO: 14)(R.J. Tressleret al. (1989),flgga); (10) polymers of peptides, including poly (RGD) (10 kDa molecularweight) (J. Murata et al. (1991), s_u;fl), poly (RGDT) (10 kDa molecularweight) (1. Saiki et al. (1990), flgfl), and copoly (RGD, YIGSR) (10 kDamolecular weight) (I. Saiki et al. (1989), §_up_i;r=_i); (11) branched peptides, suchas (AcGRGDSG),6K8K4K2KG-OH (M. Nomizu et al. (1993), s_up_r_a), and (12)cyclic peptides incorporating penicillamine, such as G(Pen)GFlGDSPC (M.D.Pierschbacher & E. Ruoslahti (1987), flgra). Derivatives of the RGDsequence that possess activity in blocking adhesion of metastatic cells,including derivatives in categories (1)-(12), are referred to generically hereinas “RGD-derived active peptides” and are within the scope of the invention.Laminin-derived peptides include: (1) the YIGSR (SEQ ID NO:2) pentapeptide derived from the laminin sequence (J. Murata et al. (1989),_sgma_); (2) derivatives of the laminin pentapeptide sequence withconservative amino acid substitutions such as YCGSR (SEQ ID NO: 15) (K.Kawasaki et al., “Amino Acids and Peptides. XXI. Laminin-Related PeptideAnalogs Including Po|y(Ethy|ene Glycol) Hybrids and Their Inhibitory Effecton Experimental Metastasis Chem. Pharm. Bull. 422917-921 (1994); (3)truncated peptides derived from this sequence, such as YIGS (SEQ ID NO:16) (K. Kawasaki et al. (1994), sgma); (4) extended sequences such asCDPGYIGSR (SEQ ID NO: 17) (K. Kawasaki et al. (1994), supra) (5)peptides, including substituted peptides and extended sequences withamino acid substitutions, in which a D-amino acid replaces one of thenaturally occurring L-amino acids, such as CDPGYl(dA)SR and YlG(dA)SR-17-WO 98/1305910152025CA 02264227 1999-02-23PCT/US97/17410(G.J. Ostheimer et ai., “NMR Constrained Solution Structures for LamininPeptide 11,” 4 267: 25120-25125 (1992)); (6) branchedpeptides such as (Ac-YIGSFtG),6K8K,,K2KG-OH, (YlGSRG),5KBK,,K2KG-OH,(Ac—YlGSRG)8K,,K2KG-OH, and (Ac-YlGSRG),,K2KG-OH (M. Nomizu et ai.(1992), _s_u_p_rg); and (7) polymers of peptides including poly (YIGSR) (10 kDamolecular weight) (I. Saiki et al. (1989), s_L;gr_a). Derivatives of the YIGSRsequence that possess activity in blocking adhesion of metastatic cells,including derivatives in categories (1)-(7), are referred to generically hereinas “YlGSR-derived active peptides” and are within the scope of theinvention.ll. METHODS FOR DELIVERING THERAPEUTIC DRUGS TOMETASTATIC CELLSAn additional aspect of the present invention is a method fordelivery of a therapeutic drug to a metastatic cell. Typically, such a methodcomprises the steps of:(1) contacting a hydrolyzable prodrug comprising an amino-terminal capped peptide covalently linked to a therapeutic drug through aself-immolating spacer of sufficient length to prevent the occurrence of sterichindrance with a metastatic cell, the amino-terminal capped peptide being asubstrate for a peptidohydrolase located on the surface of the metastatic cell;(2) allowing the peptidohydrolase located on the surface of themetastatic cell to hydrolyze the hydrolyzable prodrug and release thetherapeutic drug from the prodrug; and(3) allowing the therapeutic drug to enter the metastatic cell.Typically, as above, the therapeutic drug is an anticancer drug.Typically, the hydrolyzable prodrug is delivered to themetastatic cells under conditions under which the prodrug is stable in theabsence of enzymatic hydrolysis. Typically, such prodrugs are stable inplasma at pH 7.4 at 379 C. for at least 6 days in the absence of a-18-W0 98/130591015202530CA 02264227 1999-02-23PCT/US97/ 17410peptidohydrolase such as cathepsin B. In some cases, the prodrugs arestable for 16 days or more or 20 days or more in the absence ofcathepsin B. Thus, the prodrug can be delivered to the metastatic cellseither in vivo or in vitro. Typically, the hydrolyzable prodrugs of the presentinvention are administered in a quantity sufficient to kill at least a fraction ofthe metastatic cells.The hydrolyzable prodrugs of the present invention can beadministered in vivo using conventional modes of administration including,but not limited to, intravenous, intraperitoneal, oral or intralymphatic. Otherroutes of injection can alternatively be used. Oralor intraperitonealadministration is generally preferred. The composition can be administeredin a variety of dosage forms which include, but are not limited to, liquidsolutions or suspensions, tablets, pills, powders, suppositories, polymericmicrocapsules or microvesicles, liposomes, and injectable or infusiblesolutions. The preferred dosage form depends on the mode ofadministration and the quantity administered.Pharmaceutical compositions for administration according tothe present can include conventional pharmaceutically acceptable carriersand adjuvants known in the art such as human serum albumin, ionexchangers, alumina, lecithin, buffer substances such as phosphates,glycine, sorbic acid, potassium sorbate, and salts or electrolytes such asprotamine sulfate. The most effective modes of administration and dosageregimen for the hydrolyzable prodrugs as used in the methods of the presentinvention depend on the severity and course of the disease, the patient'shealth, the response to treatment, the particular type of metastatic cellscharacteristic of the particular primary tumor, the location of the metastases,pharmacokinetic considerations such as the condition of the patient‘s liverand/or kidneys that can affect the metabolism and/or excretion of theadministered hydrolyzable prodrugs, and the judgment of the treatingphysician. Accordingly, the dosages should be titrated to the individualpatient.-19-CA 02264227 1999-02-23W0 98/ 13059 PCT/US97/1 7410The invention is further exemplified by the following Examples.These examples are for illustrative purposes only and are not intended tolimit the scope of the invention..20-W0 98l130591015202530CA 02264227 1999-02-23PCT/US97I 17410Examgle 1Synthesis of Ac-Phe-Lys-PABC-CPTOne example of a hydrolyzable prodrug according to thepresent invention is acetylphenylalanyllysyl-Q-aminobenzylcarbonylcamptothecin (Ac-Phe-Lys-PABC-CPT).Synthesis of Fmoc-Phe-Lys-Boc-PABC-CPT. Camptothecin(534 mg, 1.53 mmol) was suspended in 14 ml of methylene chloride. Tothis was added 3.09 ml. of 1.93 M CIZCO (in toluene) and pyridine(0.14 ml, 1.68 mmol). The resulting slurry was stirred overnight after whichthe solvent and reagents were removed under vacuum. The resulting solidwas resuspended in methylene chloride and the solvent removed undervacuum to ensure complete removal of excess Cl2CO. This step wasrepeated two more times. The material was suspended a final time in 3.0 mlof methylene chloride and then added by pipet to a 5.0 ml suspension ofFmoc-Phe-Lys(Boc)—PABOH (1.61 g, 2.27 mmol). (In this intermediate,Fmoc refers to the 9—fluorenylmethyloxycarbonyl protecting group and Boorefers to the t-butyloxycarbonyl group.) The reaction was stirred for 4 hoursthen transferred to a separatory funnel and diluted with additional methylenechloride. The organic layer was washed with saturated aqueous sodiumbicarbonate, and saturated sodium chloride, then dried over sodium sulfate.The solution was filtered and the solvent removed under vacuum. Theproduct was isolated by column chromatography (2 x 12 in silica, 98:2methylene chloride/ethanol) and isolated as a solid.Synthesis of Ac-Phe-Lys (Boc)-PABC-CPT. The compoundisolated from the first stage of the reaction, Fmoc~Phe-Lys(Boc)—PABC-CPT,has, in place of the acetyl group, a 9-fluorenylmethyloxycarbonyl (“Fmoc”)group and has the e-amino group of the lysine protected with abutyloxycarbonyl (“Boc”) blocking group. The first step is the conversion ofthe 9-fluorenylmethyloxycarbonyl amino-terminal blocking group to theacetyl blocking group. To perform this conversion, Fmoc-Phe-Lys-(Boc)-PABC-CPT (200 mg, 0.18 mmol) was suspended in 6.0 ml of methylene-21-WO 98/130591015202530CA 02264227 1999-02-23PCT/US97l174l0chloride and 1.0 ml of diethylamine added. The suspension graduallydissolved as the solution and was stirred for 3 h. The solvent anddiethylamine were removed under vacuum. The resulting foamy solid wasredissolved in methylene chloride and then acetic anhydride (0.068 ml,0.72 mmol) and diisopropylethylamine (0.13 mol) added. The reactionmixture was stirred overnight, transferred to a separatory funnel, and washedwith pH 7 buffer. The organic layer was dried over sodium sulfate, filtered,and the solvent removed under vacuum. The product was purified bycolumn chromatography (2 x 3 in, silica, with a gradient of 713-9525methylene chloride/ethanol) to yield 110 mg (67 °/o yield) of the material asa yellow solid.Svnthesis of Ac-Phe-Lvs-PABC-CPT. The final step in thesynthesis was the removal of the protecting butyloxycarbonyl group on the 2:-amino group of the lysine in the hydrolyzable substrate. Ac—Phe-Lys-(Boo)-PABC-CPT (20 mg, 0.02 mmol) was suspended in 0.5 ml ofdichloromethane to which was added 0.5 ml of dichloroacetic acid. Theresulting bright yellow solution was stirred for 3 h after which it was added to50 ml of diethyl ether to precipitate the product. The solid was filtered togive 15 mg (83% yield) of a light yellow solid. The mass spectroscopy-electrospray ionization (MS-ESI) molecular weight calculated for MH*(C45H,,6N6O9) was 815.3. The measured value was 815.5.Example 2Svnthesis of CA-SP-Val-Cit-BHMS-Dox._,Another example of a hydrolyzable prodrug according to thepresent invention uses the bifunctional self-immolating spacer BHMS (p-NH-Ph—CH=(CH2OCO-)2). This bifunctional spacer can bind two doxorubicinmolecules. This compound also has a capping group of 2-aminoethylthio(CA) linked to the valine residue through a succinimidopropionyl (SP) group.Synthesis of MP-Val-Cit-BHMS(OTES)g; The first step in thesynthesis of this compound is the synthesis of MP-Val-Cit-BHMS-(OTES)2,-22-W0 98/ 130591015202530CA 02264227 1999-02-23PCT/US97/ 17410where MP is maleimidopropionyl and TES is triethylsilyl. The startingmaterial, Val-Cit-BHMS-(OTES)2 (500 mg, 0.76 mmol) was dissolved in2.0 ml of dimethylformamide followed by the addition ofdiisopropylethylamine (0.20 ml, 1.14 mmol) and N-succinimidyl-3-maleimidopropionate (303 mg, 1.14 mmol). After stirring for 2 h, thereaction mixture was transferred to a separatory funnel, diluted withmethylene chloride (200 ml) and washed with water (2 x 200 ml). Theorganic layer was dried over sodium sulfate with the addition of methanol tosolubilize the product, filtered, and the solvent removed under vacuum. Theproduct was purified by column chromatography with a 1.5 x 10 cm columnof silica, 95:5 methylene chloride/methanol to yield 250 mg (41% yield)product as a solid. The ESI-MS value calculated for M-H(C40H66N6O8Sl2)was 813.4. The value found was 813.4.Synthesis of MP—Val-Cit-BHMS-(OPNP)3. The next step in thesynthesis was the conversion of the triethylsilyl (TES) groups to p-nitrophenyl (PNP) groups. Forthis step, MP-Val-Cit-BHMS-(OTES)2(500 mg, 0.61 mmol) and di(Q—nitrophenyl) carbonate were placed in a10 ml flask and dissolved in 2.0 ml of dimethylformamide. Cesium fluoride(220 mg, 1.4 mmol) was added and the reaction mixture stirred for 1 h.The solution was then transferred to a separatory funnel, diluted with 100 mlof methylene chloride and washed with water (3 x 200 ml). The organiclayer was dried over sodium sulfate, filtered, and the solvent removedyielding 420 mg (75% yield) of product as a solid.Synthesis of MP-Val-Cit-BHMS-(Dox)2. The next step is thecoupling of the doxorubicin molecules. For this step, MP-Va|-Cit-BHMS-(OPNP)2 (420 mg, 0.46 mmol), was dissolved in 4.0 ml ofdimethylformamide to which was added diisopropylethylamine (0.40 ml,2.3 mmol) and doxorubicin hydrochloride (665 mg, 1.15 mmol). Thereaction was stirred 1 h followed by precipitation from methanol. Theresulting solid was filtered and purified by column chromatography on a4.8 x 10 cm column of silica with a gradient of 90:10 to 85:15trichloromethane/methanol to isolate 315 mg (32% yield) of product as a red-23-.........s4.....................r._..n..«..................................,».......W..... .. . . .......W................ ................................................n. .WO 98/130591015202530CA 02264227 1999-02-23PCT/US97/17410solid. The ESI-MS calculated for M-H (C8,,H92N8O32) was 1723.6 the valuefound was 1723.6.Svnthesis of CA—SP—Val—Cat-BHMS-(Dex): The final step inthe synthesis was the reaction of cysteamine with the maleimidopropionylmoiety to yield a 2-aminoethylthio-succinimidopropionyl capping group atthe amino-terminus of the peptide. MP-Val-BHMS-(Dox)2 (74 mg, 0.043mmol) was suspended in 1.0 ml methanol. Dimethyiformamide (15 drops)was added followed by cysteamine hydrochloride (8 mg, 0.07 mmol). Afterstirring for 0.5 h, the product was precipitated by the addition ofapproximately 10 ml of diethyl ether to yield 80 mg of red solid. The materialwas redissolved in dimethylformamide and precipitated fromdichloromethane to yield 56 mg (73% yield) of product as a red solid. TheESl-MS calculated for MH* (C86H,00N9O32S) was 1802.8. The value foundwas 1802.7.Example 3Svnthesis of Ac—Phe-Lvs-PABC-DoxSvnthesis of Fmoc-Lvs (MMT). The first step in the synthesis ofthe hydrolyzable prodrug Ac-Phe-Lys-PABC-Dox is the synthesis of Fmoc-Lys (MMT). This lysine derivative has its cc-amino group protected with theprotecting group 9-fluorenylmethoxycarbonyl and its e-amino groupprotected with the blocking group monomethoxytrityl. A stirred suspension ofFmoc-Lys hydrochloride (23.78 g, 56.42 mmol) and dry methylene chloride(250 ml) under argon at room temperature was treated with trimethylsilylchloride (15 ml, 2.1 equiv.) and diisopropylethylamine (10.3 ml, 1.05 equiv.).The mixture was heated at reflux for 1 h, during which time it becamehomogenous, and then cooled to 09C. Diisopropylethylamine (31 ml, 3.1equiv.) was added, followed by Q-anisyldiphenylmethyl chloride (18.29 g,1.05 equiv.). The reaction was stirred at room temperature for 14 h. Thesolvent was evaporated and the residue partitioned between ethyl acetateand pH 5 buffer (0.05 M biphthalate). The organic phase was washed withmore pH 5 buffer. water, and brine, dried over sodium sulfate, and-24..W0 98/130591015202530CA 02264227 1999-02-23PCT/US97/17410evaporated to give a pa|e—yellow foam (34.71 g, 96% yield). Proton nuclearmagnetic resonance (NMR) was performed in CDCI3 to yield: 5 1.26 and1.68 (m, 2H and 4H), 2.45 (rn, 2H), 3.71 (s, 3H), 4.05-4.40 (rn, 4H), 6.81 (d,2H), and 7.15-7.77 (m, 20H). MS-FAB yielded peaks at 641 (MH)*, 663(M+Na)*, and 679 (M+K)*. The HRMS calculated was 641.3015. The valuefound was 641.3001. The structure is shown in Figure 1.Svnthesis of Lvs(MMT). The next step is the removal of the 9-fluorenylmethoxycarbonyl group on the e-amino residue of the lysine so thatthis group is available for condensation. Fmoc-Lys (MMT) (5.25 g, 8.19mmol) in 1:1 methylene chloride/acetonitrile (80 ml) at room temperaturewas treated with diethylamine. After 1.5 h, the solvents were evaporated.The residue was flushed with acetonitrile (2x 50 ml) at 60‘-’C, and thentriturated with diethyl ether (80 ml). The resulting solid was collected byfiltration, washed with diethyl ether, and then dissolved as far as possible in1.1 methylene chloride/methanol. Some solid by-product was removed byfiltration and the filtrate was concentrated in vacuo. The resulting light-tansolid was dried in vacuo for 4 h (3.221 g, 94% yield). Proton NMR in DMSO-ds gave: 8 1.34, 1.57, and 1.72 (rn, 6H), 2.05 (rn, 2H), 3.38 (rn, 1H), 3.68 (s,3H), 3.71 (d, 2H), 7.03, 7.40) (m, 12H). MS-FAB yielded peaks at 419.2(MH)*, 441.4 (M+Na)*, and 457.4 (M+K)*. The analytical data calculated forC26H30N2O3-0.5H2O was C-73.04, H—7.31, N—6.55. The values found were N-73.62, H-7.59, N—6.56. The resulting structure is shown in Figure 1.Synthesis of Ac-Phe-Lys (MMT). The next step in the synthesisis the condensation of a phenylalanyl residue with the oi-amino-terminus ofthe lysine. The phenylalanyl residue has its own or-amino-terminusprotected with an acetyl group. A stirred solution of Lys (MMT) (4.0940 g,9.781 mmol) and lithium hydroxide monohydrate (410.4 mg, 1 equiv.) inwater (25 ml) and dimethoxyethane (70 ml) at room temperature was treatedwith a solution of Ac-Phe-OSu (2.9763 g, 1 equiv.) in dimethoxyethane(70 ml), where Ac refers to the acetyl group blocking the oz-amino group andSu refers to the succinimidyl group. The stirred mixture gradually becamehomogeneous within a few hours. After 16 h, as much dimethoxyethane was-25-W0 98/ 130591015202530CA 02264227 1999-02-23PCT/U S97/ 17410removed on the rotary evaporator as possible. The residue was partitionedbetween ethyl acetate and pH 4 buffer. The organic phase was washed withmore pH 4 buffer, water, and brine, dried over sodium sulfate, andevaporated to give a pale—ye|low solid (5.347 g, 90°/o yield). Proton NMR inCDCI3/CD3OD gave: 8 1.22 (m, 2H), 1.58 (m, 3H), 1.71 (m, 1H), 1.82 (s, 3H),2.49 (m, 2H), 3.00 (m, 2H), 3.75 (s, 3H), 4.26 (t, 1H), 4.63 (t, 1H), 6.82 (d,2H), 7.10-7.43 (m, 17H). MS-ESI yielded a peak at 608.5 (MH)*. Theanalytical values calculated for C37H,,,N3O5o2.5H2O was: C-68.08, H-7.10, N-6.44. The values found were: C-68.39, H—7.10, N-6.23. This structure isshown in Figure 1.Svnthesls of Ac-Phe-Lvs (MMT)-PABOH. The next step in thesynthesis is the addition of the Q-aminobenzyl moiety to introduce the self-immolating linker. A stirred mixture of Ac-Phe-Lys (MMT) (5.3096 g, 8.736mmol) and di-1-butylpyrocarbonate (2.8601 g, 1.5 equiv.) in methylenechloride (120 ml) at room temperature was treated with pyridine (0.741 ml,1.05 equiv.). After 15 min, Q-aminobenzyl alcohol (1.6140 g, 1.5 equiv.) wasadded. Stirring was continued for 16 h, and then the solvent wasevaporated. The residue was dried in vacuo for 1 h, and then triturated withdiethyl ether. The resulting solid was collected by filtration, washedrepeatedly with diethyl ether, and air-dried (5.2416 g, 84% yield). ProtonNMR (CDCl3/CD300) yielded: 5 1.31 (m, 1H), 1.50 (m, 1H), 1.89 (m and s,7H), 2.18 (m, 2H), 3.00 (m, 2H), 3.74 (s, 3H), 4.40 (t, 1H), 4.61 (s, 2H), 4.68(m, 1H), 6.67 (cl, 1H), 6.77 (d, 2H), 7.00-7.55 (m, 21H), 8.92 (br, 1H), MS-ESIyielded peaks at 713.6 (MH)“‘, 735.7 (M+Na)*. The analytical valuescalculated for C4,,H,,aN,,O5o0.5H2O were C-73.21, H-6.84, N-7.76. The valuesfound were: C-73.48, H-7.07,N-7.77. This compound is shown in Figure 1.Svnthesls of Ac-Phe-Lvs(MMT)—PABC-PNP. The next step isthe activation of the hydroxyl of the benzyl alcohol moiety of PABOH byconverting it into a Q-nitrophenyl ester. A mixture of Ac-Phe-Lys(MMT)-PABOH (5.861 g, 7.134 mmol), bis-p_-nitrophenyl carbonate (6.511 g, 3equiv.), and freshly activated powdered sieves (10 g) under argon at roomtemperature was treated with dry methylene chloride (120 ml) and then-26..WO 981130591015202530CA 02264227 1999-02-23PCT/US97/17410diisopropylethylamine (3.71 ml, 3 equiv.). The mixture was stirred at roomtemperature for 16 h and then filtered. The filtrate was evaporated and theresidue dried in vacuo for several hours and then dissolved in methylenechloride (20 ml). To this was added diethyl ether (40 ml) with moderatestirring. The resulting solid was collected by filtration, washed repeatedlywith 2:1 ether/methylene chloride, and air dried (4.1966 g, 67% yield).Proton NMR in DMF-d7 yielded: 8 1.43 (rn, 2H), 1.58 (m, 2H), 1.72 (rn, 1H),1.87 (m and s, 4H), 2.09 (rn, 2H), 2.4 (brt, 1H), 3.05 (rn, 2H), 3.78 (s, 3H), 4.52(rn, 1H), 4.72 (rn, 1H), 5.36 (s, 2H), 6.90 (d, 2H), 7.29 (rn, 16H), 7.41 (d, 2H),7.50 (d, 4H), 7.68 (d, 2H), 8.13 (d, 1H), 8.19 (d, 1H), 8.41 (d, 2H), 10.11(s, 1H). MS-ESI yielded peaks at 878.5 (MH)* and 900 (M+Na)*. Theanalytical values calculated for C5,H5,N5O9oH5O were C-68.36, H-5.96, N-7.82. The values found were: C-68.35, H-5.98, N-8.26. This compound isshown in Figure 1.Synthesis of Ac-Phe-Lys(MMT)-PABC-Dox. The next step inthe synthesis is the coupling of the doxorubicin moiety, which is theanticancer drug. A stirred mixture of Ac-Phe-Lys(MMT)-PABC-PNP (1.1006g, 1.253 mmol) and doxorubicin hydrochloride (7.634 mg, 1.05 equiv.) indimethylformamide (60 ml) at room temperature with treated withdiisopropylethylamine (0.23 ml, 1.05 equiv.). After 2 d, the mixture waspoured into ethyl acetate (400 ml). This solution was washed 4x with water,and then evaporated to give an orange solid that was chromatographed onsilica, eluting with (1) 20:1 and (2) 15:1 methylene chloride/methanol, to givethe product as an orange solid (959.8 mg, 60% yield). Proton NMR in DMF-d, gave: 8 1.25 (d, 3H), 1.41 (m, 2H), 1.56 (rn, 2H), 1.87 (m and s, 4H), 2.09(m, 4H), 2.34 (rn, 4H), 3.12 (m, 4H), 3.63 (b, r, s, 1H), 3.78 (s, 3H), 3.92 (rn,1H), 4.11 (s, 3H), 4.33 (rn, 1H), 4.51 (rn, 1H), 4.68 (m, 1H), 4.81 (s, 2H), 4.90(m, 1H), 5.00 (brs, 2H), 5.13 (brs, 1H), 5.40 (brs, 1H), 5.61 (s, 1H), 6.78 (d,1H), 6.89 (d, 2H), 7.28 (rn, 17H), 7.50 (d, 4H), 7.71 (rn, 3H), 8.05 (m, 3H),9.98 (s, 1H). MS-ESI yielded peaks at 1280.3 (M—H)‘ and 1282.4 (MH)*.. Analytical values calculated for C,2H,5N5O,7o2H2O were: C-65.59, H-6.04, N-5.31. The values found were: C-65.46, H-5.99, N-5.25. This compound isshown in Figure 2.-27-........._......_............................................... _ . ....WO 98/130591015202530CA 02264227 1999-02-23PCT/US97ll74l0Svnthesis of Ac—Phe—Q(s-PABC-DoxoHCl. The final step is theremoval of the monomethoxytrityl group protecting the a—amino group of thelysine residue to yield the final hydrolyzable prodrug according to thepresent invention. A stirred suspension of Ac-Phe-Lys(MMT)-PABC-Dox(1.8932 g, 1.476 mmol), and anisole (16 ml, 100 equiv.) in methylenechloride (50 ml) at room temperature was treated with dichloroacetic acid(1.22 ml, 10 equiv.). After 1.5 h, the mixture was poured in ethyl acetate (400ml) and the resulting suspension was stirred for 1 h. The orange solid wascollected by filtration, washed repeatedly with ethyl acetate, and thendissolved in methanol (80 ml). The solution was slowly eluted through acolumn of AG2-X8 ion exchange resin (50 g, chloride form). The orangefractions were collected and the solvent evaporated. The residue wastriturated with methylene chloride and the resulting solid collected byfiltration, washed with methylene chloride and dried in vacuo (1.5138 g, 98%yield). Proton NMFl in DMF-d7 gave: 6 1.25 (d, 3H), 1.41 (m, 2H), 1.56 (m,2H), 1.87 (m and s, 4H), 2.09 (m, 4H), 2.34 (m, 4H), 3.12 (m, 4H), 3.63 (brs,1H), 3.78 (s, 3H), 3.92 (m, 1H), 4.11 (s, 3H), 4.33 (m, 1H), 4.51 (m, 1H), 4.68(m, 1H), 4.81 (s, 2H), 4.90 (m, 1H), 5.00 (brs, 2H), 5.13 (brs, 1H), 5.40 (brs,1H), 5.61 (s, 1H), 6.78 (d, 1H), 6.89 (d, 2H), 7.28 (m, 17H), 7.50 (d, 4H), 7.71(m, 3H), 7.91 (m, 1H), 8.32 (d, 1H), 8.45 (br, 3H), 10.21 (brs, 1H), MS-ESIyielded a peak at 1010.5 (MH)*. Analytical values calculated forC52H5oN5O,5Cl-2.5H2O were: C~57.22, H-6.00, N-6.41. Values found were:C-57.16, H-6.03, N-6.34. The structure of this compound is shown in Figure2.Example 4Svnthesis of Ac—Phe-Lvs-PABC-MMCBy analogous methods, a hydrolyzable prodrug according tothe present invention with an acetyl capping group and mitomycin C (MMC)as the anticancer drug was synthesized. The peptide substrate for cathepsinB is Phe-Lys.-28-W0 98/ 130591015202530CA 02264227 1999-02-23PCT/US97/ 17410Synthesis of Ac-Phe-Lys(MMT)-PABC-MMC. A mixture of thepreviously prepared activated intermediate Ac-Phe-Lys(MMT)-PABC-PNP(614.1 mg, 0.6994 mmol), mitomycin C (245.5 mg, 1.05 equiv.),hydroxybenzotriazole (945.1 mg, 10 equiv.), and freshly activated powderedsieves (4 g) under argon at room temperature were treated with dimethylformamide (15 ml) and diisopropylethylamine (1.22 ml, 10 equiv.). After 2 d,the mixture was diluted with ethyl acetate (150 ml) and the solution washedwith water (4x) and brine, dried over sodium sulfate and evaporated. Theresidue was chromatographed on silica, eluting with 15:1 methylenechloride/methanol to give the product as a purple solid (501.0 mg, 67%yield). Proton magnetic resonance in DMF-d7 gave: 8 1.41 (m, 2H), 1.56 (m,2H), 1.69 (m, 1H), 1.77 (s, 3H), 1.84 (m and s, 4H), 2.08 (m, 2H), 2.39 (brt,1H), 3.04 (m, 2H), 3.25 (s, 3H), 3.67 (m, 3H), 3.78 (s, 3H), 4.12 (t, 1H), 4.39(d, 1H), 4.49 (m, 1H), 4.68 (m, 1H), 4.91 (ABq, 1H), 5.04 (ABq, 2H), 6.68 (br,2H), 6.89 (d, 2H), 7.05 (br, 1H), 7.29 (m, 15H), 7.49 (d, 4H), 7.70 (d, 2H), 8.13(t, 2H), 9.99 (s, 1H). MS-ESI yielded peaks at 1071.6 (M-H)‘, 1,073.5 (MH)*,1.095.6 (M+Na)*. Analytical data calculated for C60H64N8O,,-2H2O was: C-64.97, H-6.18, N—10.10. The values found were: C-65.01, H-6.11, N-10.12.This compound is shown in Figure 3.Synthesis of Ac-Phe-Lys-PABC-MMCoClCH2Q_Qgfl. The finalstep in the synthesis of the hydrolyzable prodrug is the removal of themonomethoxytrityl group blocking the e-amino group of the lysine residue. Astirred suspension of Ac-Phe-Lys(MMT)-PABC-MMC (269.5 mg, 0.2511mmol) and anisole (2.73 ml, 100 equiv.) in methylene chloride (18 ml) atroom temperature was treated with 1 M chloroacetic acid in methylenechloride (2.5 ml, 10 equiv.). After 3.5 h, diethyl ether (30 ml) was added. Theresulting suspension was stirred for 1 h, and then the purple solid wascollected by filtration, washed with diethyl ether and dried in vacuo (218.3mg, 97% yield). Proton nuclear magnetic resonance in DMF-d, gave: 6 1.51(m, 2H), 1.73 (s and m, 5H), 1.88 (s and m, 5H), 3.07 (m, 4H), 3.24 (s, 3H),4.09 (s, 2H), 4.12 (t, 1H), 4.39 (d, 1H), 4.51 (m, 1H), 4.68 (m, 1H), 4.92 (ABq,1H), 5.05 (ABq, 2H), 6.69 (br, 2H), 7.04 (brs, 2H), 7.30 (m, 7H), 7.78 (d, 2H),8.49 (d, 1H), 8.70 (cl, 1H), 10.30 (s, 1H). MS-ESI yielded peaks at: 801.6- -W0 98ll30591015202530CA 02264227 1999-02-23PCT/US97/17410(MH)*, 823.8 (M+Na)*. Analytical data calculated for C,,2H5,N,,O,2CIo2.5H2Owas: C-53.64, H-6.00, N-11.91. The values found were: C-53.58, H-5.95,N-11.51. This structure is shown in Figure 3.Example 5Svnthesis of Z-Phe-Lvs—PABC-7-PaclitaxelSvnthesis of Z-Phe—Lys(MMT). The first step in the synthesis isthe coupling of the phenylalanyl moiety with the lysine moiety, whose 5-amino group is protected with a monomethoxytrityl (MMT) residue. Couplingis accomplished with the use of a succinimidyl derivative of thephenylalanine, whose carboxyl group is thereby activated. A stirred mixtureof Lys(MMT) (7.4001 g, 17.68 mmol) and Z-Phe-OSu (7.0084 g, 1 equiv.)(“Su” refers to the succinimidyl group) in dimethylformamide (80 ml) at roomtemperature was treated with diisopropylethylamine (9.2 ml, 3 equiv.). After3 d, the reaction was diluted with ethyl acetate (400 ml) and the solution waswashed with pH 4 buffer (2x), water (2x), brine, dried over sodium sulfate,and evaporated to give a yellow foam. This was flushed several times withmethylene chloride until the foam was solid enough to be broken up with aspatula (11,259? g, 91% yield). Proton NMR in DMF-d7 gave: 5 1.35-1.95(m, 6H), 2.17 (m, 2H), 3.09 (m, 2H), 3.81 (s, 3H), 4.42 (m, 1H), 4.52 (m, 1H),4.99 (ABq, 2h), 6.91 (d, 2H), 7.41 (m, 22H), 8.27 (cl, 1H). MS-ESI yielded apeak of 700.5 (MH)*. Analytical values calculated for C43H,,5N3O6o1.5H2Owere: C-71.05, H-6.65, N-5.78. The values found were: C-71.21, H-6.43, N-5.57. The compound resulting from this synthetic step is shown in Figure 4.Svnthesis of Z—Phe—Lvs(MMT)-PABOH. The next step is theaddition of the Q-aminobenzyl moiety at the carboxyl group of the lysineresidue for the linkage to paclitaxel. A stirred mixture of Z-Phe—Lys(MMT)(7.8703 g, 11.24 mmol) and di-1-butylpyrocarbonate (3.681.6 g, 1.5 equiv.) inmethylene chloride (400 ml) at room temperature was treated with pyridine(0.955 ml, 1.05 equiv.). After 20 min, p—aminobenzyl alcohol (2.0775 g, 1.5equiv.) was added. The mixture was stirred overnight at room temperatureand then the solvent was evaporated. The residue was dried in vacuo and-30-WO 98/130591015202530CA 02264227 1999-02-23PCT/US97/17410then triturated with diethyl ether. The resulting solid was collected byfiltration and washed repeatedly with diethyl ether (6.2674 g, 69% yield).Proton NMR in DMF-d, gave: 8 1.35-1.95 (m, 6H), 2.09 (m, 2H), 2.37 (m, 1H),3.08 (m, 2H), 3.77 (s, 3H), 4.50 (m, 2H), 4.55 (d, 2H), 4.99 (ABq, 2H), 5.12 (t,1H), 6.88 (d, 2H), 7.30 (m, 24H), 7.69 (d, 2H), 8.22 (d, 1H), 10.04 (s, 1H).MS-ESI yielded peaks at 803.4 (M-H)‘, 805.7 (MH)*, 827.4 (M+Na)*.Analytical data calculated for C50H52N,,O6o0.5H2O were: C—73.78, H-6.56, N-6.88. The data found were: C-73.99, H-6.81, N-7.10. The resulting structureis shown in Figure 4.Synthesis of 2’-Monomethoxytrityl-Paclitaxel. The next step isthe synthesis of the protected paclitaxel derivative, 2’-monomethoxytrityl-paclitaxel. A stirred solution of paclitaxel (0.51 g, 0.597 mmol), and p_-anisyldiphenylmethyl chloride (4.63 g, 25 equiv.) in methylene chloride (14ml) under nitrogen at room temperature was treated with pyridine (1.23 ml,25 equiv.). After 16 h at room temperature, the solvent was evaporated, andthe residue dissolved in ethyl acetate. The solution was washed with coldpH 5 buffer (2x 100 ml), water, and brine, dried, and evaporated. Theresidue was chromatographed on silica, eluting with 3%methanol/methylene chloride to give the product as a white solid (482 mg,72% yield). Proton NMR in CDCl3 gave: 61.11 (s, 3H), 1.17 (s, 3H), 1.55 (s,3H), 1.67 (s, 3H), 1.90 and 2.54 (m, 2H), 2.26 (s, 3H), 2.51 (s, 3H), 2.54 (m,2H), 3.66 (d, 1H), 3.78 (s, 3H), 4.21 (ABq, 2H), 4.41 (m, 1H), 4.63 (d, 1H),4.92 (d, 1H), 5.62 (d, 1H), 5.70(m, 2H), 6.22(s, 1H), 6.74(d, 2H), 7.09-7.60(m,23H), 7.80(d, 2H), 8.09(d, 2H). MS-FAB yielded peaks at 1148(M+Na)*,1164(M+K)*. Analytical data calculated for C67H67NO,5oO.5H2O: was C-70.88, H-6.04, N-1.23. The values found were: C-70.58, H-6.20, N-1.25.The product is shown in Figure 5.Synthesis of Z-Phe—Lys(MMT)-PABC—7-Paclitaxel-2’-OMMT.The next step is the linkage of the paclitaxel moiety to the protected peptide.A stirred solution of 2’ monomethoxytrityl-paclitaxeI (1.5251 g, 1.354 mmol)in methylene chloride (10 ml) at 09 C. under argon was treated withdiisopropylethylamine (0.236 ml, 1 equiv.), pyridine, and diphosgene (0.09-31-W0 98/130591015202530CA 02264227 1999-02-23PCT/US97/17410ml, 0.55 equiv.). After 10 min. the ice bath was removed and the mixture wasstirred at room temperature for 4 h. The crude chloroformate solution wasthen added to a stirred solution of Z—Phe—Lys (MMT)-PABOH (1.1560 g, 1.06equiv.) and diisopropylethylamine (0.236 ml, 1 equiv.) in methylene chloride(20 ml). The mixture was stirred at room temperature for 2 d and then thesolvent was evaporated. The residue was partitioned between ethyl acetateand pH 5 buffer. The organic phase was washed with water and brine, driedover sodium sulfate, and evaporated. The residue was chromatographed onsilica. (the column was prepared in solvent containing 0.1% triethylamine),eluting with 4:1 methylene chloride/ethyl acetate, to give the product as acolorless glass (1.9307 g, 73% yield). Proton NMR in CDCl3/CD3OD) gave: 51.11(s, 6H), 1.23 (m, 2H), 1.47(m, 2H), 1.79(s, 3H), 1.96(m, 2H), 2.11(m, 2H),2.18(s, 3H), 2.29(s, 3H), 2.58(m, 2H), 3.05(m, 2H), 3.77(s, 3H), 8.82(s, 3H),4.22(ABq, 2H), 4.40(m, 2H), 4.70(d, 1H), 4.94(brd, 1H), 5.04(brs, 2H),5.18(ABq, 2H), 5.46(m, 2H), 5.67(brt, 1H), 5.73(m, 1H), 6.31(s, 1H), 6.79(m,4H), 7.00-7.60(m, 49H), 7.81(d, 2H), 8.08(d, 2H), 8.73(br, 1H). Analyticaldata calculated for C,,8H,,7N5O22-H20 is: C-71.75, H-6.07, N-3.54. Thevalues found were: C—71.77, H—6.14, N-3.45. The product is shown in Figure5.Svnthesis of Z-Phe-Lvs-PABC-7-PaclitaxeloCl2CHCOgtl_. Thefinal step is the removal of the protecting monomethoxytrityl residues on thee-amino group of the lysine and the paclitaxel moiety. A stirred solution of Z-Phe-Lys (MMT)-PABC-7-Paclitaxel-2’-OMMT (1.192 g, 0.6086 mmol) andanisole (13.2 ml, 100 equiv.) in methylene chloride (50 ml) at roomtemperature was treated with dichloroacetic acid (100 ml, 20 equiv.). After1.75 h diethyl ether (80 ml) was added and the resulting suspension wasstirred for 2 h and stored at 49 C. overnight. The white solid was thencollected by filtration, washed repeatedly with diethyl ether, and air-dried.(894.1 mg, 95% yield).Proton NMR in DMF-d7 gave: 5 1.15(s, 3H), 1.18(s, 3H), 1.47(m,2H), 1.98(m, 2H), 2.15(s, 3H), 2.22(m, 2H), 2.33(s, 3H), 2.54(m, 1H), 2.85(m,2H), 3.00(m, 2H), 3.86(m, 1H), 4.21(ABq, 2H), 4.47(m, 2H), 4.81(d, 1H),-32-WO 98/130591015202530CA 02264227 1999-02-23PCT/US97/174104.91(d, 1H), 4.98(d, 2H), 5.13(ABq, 2H) 5.42(m, 1H), 5.65(d, 1H), 5.72(m,1H), 5.87(s, 1H), 6.11(brt, 1H), 6.31(s, 1H), 7.05-7.70(m, 25H), 7.75(d, 2H),8.08(d, 2H), 9.30(br, 1H). MS-ESl yielded peaks at 1410.7(M-H)‘,1412.8(MH)*. Analytical data calculated for C80H8,N5O22Cl2oH2O was: C-60.91, H-5.81, N-4.44. The values found were: C-60.77, H~5.77, N-4.31.The product is shown in Figure 5.Example 6Synthesis of CA-SP-Lys-Phe-Lys-BHMS-Dox2Synthesis of Bis(hydroxymethyl)9-aminostyrene(BHMS). Thefirst step in the synthesis of the bifunctional hydrolyzable prodrug CA-SP-Lys-Phe-Lys-BHMS-Dox2 is a synthesis of the bifunctional linkerbis(hydroxymethyl)p-aminostyrene (BHMS). Raney nickel (5.28 ml, 50%slurry in H20) and hydrazine monohydrate (21 ml, 1.5 equiv.) were addedto a stirred solution of 2-(,c_>—nitrobenzylidene)-propane-1,3-diol (P. Vanelle etal., Eur. J. Med. Chem. 262709 (1991)) (60.29 g, 0.2885 mol) in a mixture oftetrahydrofuran (950 ml) and methanol (950 ml) at 30°C under a N2atmosphere. Vigorous gas evolution was observed and the reactiontemperature rose to 47°C. Additional quantities of hydrazine monohydrate(21 ml, 1.5 equiv.) were added after 30 min and 1.5 h while maintainingthe reaction temperature between 45-50°C. The reaction was then allowedto cool to room temperature and the catalyst was removed by filtrationthrough a celite pad. Removal of the solvent from the filtrate left the productas a yellow solid (51.43 g, 99.6% yield). The melting point was 119°C witha sample crystallized from ethyl acetate. infrared spectroscopy (KBr) yieldedpeaks at 3368, 1628, 1510, and 1022 cm". Proton nuclear magneticresonance (NMR) (300 Hz) in DMSO-d,, gave: 5 4.03 (s, 2H, CH2), 4.05 (s,2H, CH2), 4.65 (m, 2H, OH, exchanges with D20), 5.09 (s, 2H, NH2,exchanges with D20), 6.33 (s, 1H, vinylic H), 6.50 (d, 2H, J = 8.4Hz aromaticH), 6.99 (d, 2H, J =8.4 Hz aromatic H). “C NMR in DMSO-d6 yielded peaksat 57.5, 63.8, 113.6, 124.5, 126.2, 129.7, 136.0, and 147.6. Massspectroscopy gave a m/z of 179 (M*). High resolution mass spectroscopy(HRMS) calculated for C,2H,,,NO2 (MH*) was 180.1024. The value found was-33-.. . 2.22.... ......................................u...m...z . ,..,W0 98/ 130591015202530CA 02264227 1999-02-23PCT/US97/17410180.1017. The analytical values calculated for C,,,H,,,NO2 were: C-67.02, (H-7.31, N-7.82. The values found were: C-66.82, H-7.31, N-7.80. Thisstructure is shown in Figure 6.Svnthesis of Fmoc-Phe-Lvs(MM'DBHMS. The next step in thesynthesis is the coupling of the BHMS that is part of the linker moiety with apeptide, Phe-Lys, which has its oi-amino terminus blocked with a 9-fluorenylmethoxycarbonyl (Fmoc) group and which has the &:-amino group ofthe lysyl residue blocked with the monomethoxytrityl (MMT) group. A stirredsolution of Fmoc-Phe-Lys(MMT)-OH (62.0 g, 78.3 mmol),bis(hydroxymethyl)p-aminostyrene (14.4 g, 1.1 equiv.) and 1-ethoxycarbonyl-2-ethoxy-1,2-dihydroquinoline (EEDQ) (28.9 g, 1.5 equiv.)in a mixture of methylene chloride (400 ml) and methanol (100 ml) was leftstirring at room temperature overnight. The solvents were removed and theresidue was triturated with diethyl ether to leave the product as a light solid(47.3 g, 50% yield). IR (KBr) gave 3284, 1692, 1644, and 1510 cm“.Proton NMR (300 Hz) in DMSO-d6 gave: 5 1.20-1.75 (m, 6H), 1.90 (t, 2H),2.39 (t, 1H), 2.74-2.99 (m, 2H), 3.64 (s, 3H), 4.00-4.16 (m, 7H), 4.26-4.42 (m,2H), 4.79 (t, J = 5.0 Hz, 1H), 4.84 (t, J = 5.6 Hz, 1H), 6.48 (s, 1H), 6.77 (d, J=8.7 Hz), 7.06-7.57 (m, 27H), 7.84 (d, J = 7.6 Hz, 2H), 8.18 (d, J = 7.6 Hz,1H), 10.07 (s, 1H). Mass spectroscopy gave a m/z of 948 (M*). HRMScalculated for C6oH60N,,O, (M Na*) was 971.4360. The value found was971.4375. This structure is shown in Figure 6.Svnthesis of H-Phe—Lvs(MMT)—BHMS. The next step in thesynthesis was the removal of the blocking Fmoc group at the amino-terminusof the peptide to yield an amino group that is available for further coupling. Astirred solution of Fmoc-Phe-Lys(MMT)-BHMS (2.66 g, 2.81 mmol) anddiethylamine (7 ml) in dimethylformamide (15 ml) was stirred at roomtemperature for 5 min after which the solvent was removed. The residuewas chromatographed on silica gel (gradient elution with methylenechloride/methanol 98:2 to 92:8) to give the product as a tan solid (0.77 g,38% yield). Proton NMR in DMSO-d,, plus D20 gave:8 1.20-1.91 (m, 6H),2.59, 2.96 (m, 2H), 3.44 (m, 1H), 3.67 (s, 3H), 4.04 (s, 2H), 4.10 (s, 2H), 4.43-34-W0 98/ 130591015202530CA 02264227 1999-02-23PCT/US97l 17410(m, 1H), 6.47 (s, 1H), 6.79-7.56 (m, 23H). Mass spectroscopy byelectrospray ionization (MS-ESI) yielded a m/z of 725(M-H*). HRMScalculated for C,5H5,,N4O5 (MNa*) was 749.3679. The value found was749.3678. This structure is shown in Figure 6.Synthesis of H-Lys(MMT)-Phe-Lys(MMT)—BHMS. The next stepis the addition of the amino-terminal lysyl residue. This is accomplished byreacting a lysine with its on-amino terminus blocked with Fmoc, its s-aminogroup blocked with a monomethoxytrityl group, and its carboxyl groupactivated with a succinimidyl (Su) moiety. A solution of H-Phe-Lys(MMT)-BHMS (0.770 g, 1.06 mmol) and Fmoc-Lys(MMT)-OSu (0.860 g,1.1 equiv.) in dry dimethylformamide (5 ml) was left stirring at roomtemperature for 2h. This was then diluted with ethyl acetate and washedwith water. The organic phase was filtered through celite, washed with brineand then dried over sodium sulfate. Removal of the solvents left a solid thatwas dissolved in a mixture of methylene chloride (10 ml) and diethylamine(10 ml). After 3h at room temperature, the solvents were removed and theresidue was chromatographed on silica gel with gradient elution withmethylene chloride/methanol of 99:1 to 91 :9 to give it a product as a tan solid(0.661 g, 56% yield). Proton NMR in DMSO-d6 plus D20 gave: 6 1.04-1.90(m, 12H), 2.70-3.00 (m, 2H), 3.40 (m, 1H), 3.66 (s, 3H), 3.69 (s, 3H), 4.05 (s,2H), 4.11 (s, 1H), 4.35 (m, 1H), 4.60 (m, 1H), 6.48 (s, 1H), 6.77-7.58 (m, 37H).This structure is shown in Figure 6.Svnthesis of PNP Carbonate Derivative of MP—Lvs(MMT)-Phe-Lys(MMT)-BHMS. The next step is the synthesis of a gnitrophenyl (PNP)carbonate derivative for coupling of the doxorubicin moieties to the linker.Simultaneously, the amino-terminus of the peptide is activated by reactionwith N-succinimidyl 3-maleimidopropionate. A stirred mixture of H-Lys(MMT)-Phe-Lys(MMT)-BHMS (0.661 g, 0.59 mmol) and N-succinimidyl3-maleimidopropionate (0.172 g, 1.1 equiv.) in dry dimethylformamide(5 ml) was left stirring at room temperature for 2h. This was diluted withethyl acetate and washed with water, brine, and then dried over sodiumsulfate. The solvents were removed and the residue was dissolved in dry-35- .........~........,.. ...... ........_..............,,.....,._....M_.. .. . ,WO 98/130591015202530CA 02264227 1999-02-23PCT/US97/17410tetrahydrofuran (10 ml). This was cooled in an ice bath anddiisopropylethylamine (0.82 ml, 8 equiv.), p_-nitrophenyl chloroformate(0.713 g, 6 equiv.) and pyridine (0.024 ml, 0.5 equiv.) were added. Thereaction was left stirring for 1h after which it was quenched by the addition ofwater and then extracted with methylene chloride. The organic extracts werecombined, washed with water, saturated aqueous sodium bicarbonatesolution, water, and then dried over sodium sulfate. The solvent wasremoved and the residue was dissolved in methylene chloride (8 ml).Diethyl ether was added to precipitate the product as a tan solid (0.472 g,50% yield). Proton NMR in DMSO—d6 gave: 5 1.04-1.90 (m, 12H), 2.36 (m,4H), 2.65-3.10 (m, 2H), 3.52 (m, 2H), 3.65 (s, 3H), 3.67 (s, 3H), 4.03 (m, 1H),4.33 (m, 1H), 4.45 (m, 1H), 5.01 (s, 2H), 5.04 (s, 2H), 6.76-8.30 (m, 52H).This structure is shown in Figure 7.Svnthesis of MP-Lvs(MMT)-Phe-Lvs(MMT)-BHMS-Doxg. Thenext step in the synthesis is the coupling of the two doxorubicin moieties tothe activated linker. A stirred mixture of the Q-nitrophenyl carbonatederivative of MP-Lys(MMT)—Phe-Lys(MMT)-BHMS (0.472 g, 0.294 mmol)and doxorubicin hydrochloride (0.340 g, 2 equiv.) in dry dimethylformamide(9 ml) at room temperature was treated with diisopropylethylamine(0.153 ml, 3 equiv.). After 7 h, the solution was concentrated to about 3 ml.This was then added dropwise to a stirred solution of methanol (20 ml). Theresulting precipitate was collected and washed with methanol and diethylether to give the crude product as a red solid (527 mg). This was dissolvedin 5 ml of a 2% solution of methanol in methylene chloride andchromatographed on a silica gel column (gradient elution with methylenechloride/methyl alcohol of 99:1 to 93:7) to afford pure material as a red solid(0.308 g, 43% yield). Distinguishing peaks in the proton-NMR spectrum inDMSO-d6 were: 6 1.09 (brs, 6H, methyl group of Dox), 3.64 (s, 3H, methylgroup of monomethoxytrityl), 3.67 (s, 3H, methyl group of monomethoxytrityl),3.86 (s, 3H, methyl group of Dox), 3.87 (s, 3H, methyl group of Dox). Theresulting structure is shown in Figure 7.-35-WO 98/130591015202530CA 02264227 1999-02-23PCT/US97/17410S nthesis of MP-L s-Phe-L s-BHMS-Doxg. The next step inthe synthesis is the removal of the blocking monomethoxytrityl groups on thes-amino moieties of the lysine residues. A stirred suspension of MP-Lys(MMT)-Phe-Lys(MMT)-BHMS(Dox)2 (0.300 g, 0.124 mmol) and anisole(2.7 ml, 200 equiv.) in methylene chloride (20 ml) at room temperature wastreated with dichloroacetic acid (0.205 ml, 20 equiv.). After 1.5h, themixture was poured into ethyl acetate (20 ml). The product separated as afine red solid. This was collected using a centrifuge, washed with ethylacetate, and dried (0.252 g, 95% yield). Distinguishing peaks in the proton-NMR spectrum in DMSO-d6 were: 5 1.10 (brs, 6H, methyl group of Dox), 3.88(s, 3H, methyl group of Dox), 3.89 (s, 3H, methyl group of Dox), 6.99 (s, 2H,maleimide residue). MS-ESI gave a m/z of 1872 (MH*). This resultingstructure is shown in Figure 8.S nthesis of CA—SP—L s-Phe-L s-BHMS-Doxz. The final stepin the synthesis is the reaction of cysteamine with the maleimidopropionylgroup to yield the amino-terminal cap. MP-Lys-Phe-Lys-BHMS-Doxz(0.068 g, 0.032 mmol) was added to a solution of cysteamine hydrochloride(0.012 g, 3.3 equiv.) in methanol (1 ml). After 1 h, the mixture was filteredand concentrated to about 0.5 ml. Diethyl ether was added and theprecipitate was collected using a centrifuge. This was washed with diethylether and dried to give the product as a red solid (0.063 g, 88% yield).Distinguishing peaks in the proton NMR spectrum in DI‘/ISO—d6 were: 81.10 (brs, 6H, methyl group of Dox), 3.88 (s, 3H, methyl group of Dox), 3.89(s, 3H. methyl group of Dox). MS-ESI yielded m/z of 1949 (MH*) Thisstructure is shown in Figure 8.Example 7Cflotoxicity Assays of Hydrolyzable ProdrugsMaterials and MethodsCell Culture. The BT-20 cell line was maintained in E-MEM—10(minimal essential medium (Earle’s Salts) supplemented with 10% fetal-37-W0 98/ 130591015202530CA 02264227 1999-02-23PCT/US97/17410bovine serum, penicillin (100 U/ml) and streptomycin (100 pg/ml)) in 5% CO2at 37‘-’ C. MCF 10A cell line was maintained in DMEM/Ham’s F12supplemented with 5% horse serum, EGF (20 ng/ml), insulin (0.5 pg/ml),hydrocortisone (0.5 pg/ml), penicillin (100 U/ml), and streptomycin (100pg/ml) in 5% CO2 at 37‘-’ C. Corning tissue culture multiwell plates (24wells/plate) were seeded with 105 cells/16 mm well in 2 ml maintenancemedium, refed 48 h after seeding and were ready to use in cytotoxicityassays 4 8h later, with the cells just reaching confluency.Solubilization of Hvdrolvzable Prodruds and Anticancer Drucisin Agueous Medium. Concentrated stocks (200 mM) of the hydrolyzableprodrugs and the anticancer agents doxorubicin, taxol, mitomycin C, andcamptothecin were prepared in 100% dimethyl sulfoxide. Dilution withdimethyl sulfoxide to 33.33 mM followed by partial hydration with addition ofE-MEM-0% (serum free E-MEM) resulted in 25 mM drug or prodrug in asolvent of 75°/o dimethyl sulfoxide/25% EMEM-0%. A dilution of 121000 inEMEM-0% resulted in 25 uM concentrations with DMSO vehicleconcentration of 0.075%. Serial 1:10 dilutions of the 25 uM preparationswere made with E-MEM-O. This stepwise hydration of drugs and prodrugsfacilitated the delivery of otherwise aqueously insoluble compounds to cellsin culture.Cvtotoxicitv Assav and Inhibition with E-64 or CA-074. Thecytotoxicity assay protocol employing MCF 10A (low cathepsin B secreters)and BT-20 (high cathepsin B secreters) and inhibition of cytotoxicity with L-t_rar_1§-epoxysuccinyl-leucylamido (4-guanido) butane (cysteine proteaseinhibitor) or CA-074 (N-(L—3-tri-propylcarbamoyloxirane-2-carbonyl)-L-isoleucyl-L-proline)(specific cathepsin B inhibitor) was standardized with theonly variations occurring with compound, molar concentration, time ofexposure, and presence or absence of inhibitors. Maintenance medium wasaspirated from the cell walls after feeding, and the cells washed twice with2 ml/well Hanks balanced salt solution (HBSS). After the second wash,1.0 ml E-MEM-0 (medium without phenol red indicator) containingchemotherapeutic agents, hydrolyzable prodrug derivatives, medium alone,-33-WO 98/130591015202530CA 02264227 1999-02-23PCT/US97/ 17410or solubiiization vehicle controls (0.1% dimethyl sulfoxide). For assays inwhich inhibitor was employed, E-MEM-O containing E64, CA-074, or noinhibitor was added 30-60 min prior to addition of anticancer drugs orhydrolyzable prodrugs. Medium was aspirated from the wells and replacedwith fresh E—MEM-O containing anticancer agents, hydrolyzable prodrugs,inhibitors, medium alone, or vehicle controls in the molar concentration andcombinations indicated. incubation was continued in 5°/o CO2 at 37°C for theindicated times after drug/prodrug addition. To each well containing 1.0 ml,50 pl of 0.5% Trypan Blue in HBSS was added, the plate was gently swirledto mix the dye, the plate was allowed to sit for 5 minutes, and the percentageof dead cells (blue stained) was read. In this example, the followingcompounds were used: Compound (1) was Ac-Phe-Lys-PABC-Dox;Compound (2) was Ac-Phe-Lys—PABC—MMC; Compound (3) was Ac-Phe-Lys-PABC-CPT; Compound (4) was Z-Phe—Lys—PABC-7-Paclitaxel;Compound (5) was CA-SP-Val—Cit-BHMS-Doxz; Compound (6) was CA-SP—Lys-Phe-Lys-BHMS-Doxz; Compound (7) was Z-Phe-Lys-Dox; andCompound (8) was 2-Hydroxyethylthio-SP-D—Phe-Lys-PABC-Dox. In thesecompounds, MMC is mitomycin C, CPT is camptothecin, Z isbenzyloxycarbonyl, CA is 2-aminoethylthio, SP is succinimidopropionyl,PABC is Q—aminobenzyl carbonyl, and BHMS is Q-NH-Ph-CH=C(CH2OCO-)2.in the results shown in Figure 9, the prodrugs were used at100 uM, CA-074 (a cathepsin B-specific inhibitor) was used at 40 uM, BT-20 (high cathepsin B-secreting cells) and MCF-10 (low cathepsin B-secreting cells) were used, and the percent cell kill in 24 hours is shown forCompounds (1), (2), (7) and (8). In a separate experiment at 25 pM drug,Compounds (1) and (2) were 60% and 75% inhibited by 4 uM CA-074,respectively. These results show that high cathepsin B-secreting cells weremore readily killed than low cathepsin B-secreting cells. The killing of highcathepsin B-secreting cells was inhibited by CA-O74. The killing of lowcathepsin B-secreting cells was not inhibited by CA-074. This is attributed tokilling, albeit inefficiently, of the low secretors, not as the result of the activityof externally secreted cathepsin B, but rather by non-specific endocytosis ofthe drug into lysosomes where all cells possess cathepsin B. CA-074 does-39.WO 98/130591015202530CA 02264227 1999-02-23PCT/US97ll74l0not inhibit this killing because it does not enter lysosomes. Compounds(;7)and (8) are control compounds, weaker cytotoxic agents thanCompounds (1) and (2) because they are less susceptible to cathepsin B-mediated hydrolysis to an active drug because of their particular structures.Compound (7) lacks the PABC-self-immolating linker that facilitatesenzymatic cleavage, resulting in probable steric hindrance that places thebulky doxorubicin moiety in the active site of the cathepsin B.Compound (8) has the amino acids in the unnatural D instead of the naturalL configuration.Figure 10 depicts the percent cell kill at various times andprodrug concentrations with BT-20 cells. The results with CA-O74 present at40 pM are given in parentheses. The results indicate that all testcompounds show dose- and time-dependent killing of BT-20 (cathepsin B+)cells. The cathepsin B inhibitor CA-O74 strongly inhibits cytotoxicity.Compound (8), the compound containing the amino acids in the D-configuration, is very much less active.The results shown in Figure 11 are the percent cell kill forvarious times and prodrug concentrations using Compound (6), with bothBT-20 (high cathepsin B-secreting) cells and MCF-10A (low cathepsin B-secreting) cells. Results in parentheses are with added E-64 (10 uM), abroad-spectrum cysteine protease inhibitor. (Other experiments showed thatE-64 and medium alone caused 10% killing of BT-20 cells after 24 hours.)These results show that Compound (6), with two doxorubicinmoieties linked to the prodrug, causes dose- and time-dependent killing ofBT-20 (high cathepsin B-secreting) cells. The killing of MCF-10A (lowcathepsin B-secreting) cells is much less efficient. The cysteine proteaseinhibitor E-64 strongly inhibits cytotoxicity.In conclusion, tumor cells do indeed secrete enough cathepsinB to release enough cytotoxic drug to kill the cells efficiently. Tumor cellsthat secrete less cathepsin B are resistant to the hydrolyzable prodrugs.These cells are surrogates for normal cells that secrete no cathepsin B at all.-40-WO 98/1305910152025CA 02264227 1999-02-23PCT/U S97/ 17410Cathepsin B inhibitors strongly reduce cytotoxicity of the prodrugs, showingthat cathepsin B is the principal means of unmasking them. A prodruglacking the self—immolating linkers PABC or BHMS has much reducedcytotoxicity, showing that it is an enzyme, presumably cathepsin B, thatreleases active drug. This is because the absence of the self-immolatinglinker results in steric hindrance. Additionally, a prodrug with amino acids inthe unnatural D configuration has much reduced cytotoxicity, again showingthe role of cathepsin B.Example 8Stability of Hvdrolvzable ProdrudsFigure 12 shows the stability of a number of hydrolyzableprodrugs according to the present invention linked to doxorubicin using aPABC-self-immolating linker. These hydrolyzable prodrugs releaseanticancer drug under cathepsin B catalysis at 37°C, pH 7.4, at reasonablerates, and are stable for days or weeks in freshly drawn human plasmaunder the same conditions.ADVANTAGES OF THE PRESENT INVENTIONThe present invention provides an efficient way to treat cancercells secreting peptidohydrolases on their surface, particularly metastaticcells. The hydrolyzable prodrugs of the present invention are usable againstmany types of metastases and do not depend for their activity oncharacteristics of the primary tumor cells that might not be shared by themetastases. Hydrolyzable prodrugs of the present invention are readilyabsorbed and lack toxicity.Although the present invention has been described withconsiderable detail, with reference to certain preferred versions thereof,other versions are possible. Therefore, the spirit and scope of the appendedclaims should not be limited to the description of the preferred versionscontained herein.-41-

Claims (47)

What Is Claimed Is:

1. A hydrolyzable prodrug comprising an amino-terminal capped peptide covalently linked to a therapeutic drug through a self-immolating spacer of sufficient length to prevent the occurrence of steric hindrance, the amino-terminal capped peptide being a substrate for a peptidohydrolase located on the surface of a metastatic cell.

2. The hydrolyzable prodrug of claim 1 wherein the peptidohydrolase is selected from the group consisting of cathepsin B and collagenase IV.

3. The hydrolyzable prodrug of claim 2 wherein the peptidohydrolase is cathepsin B.

4. The hydrolyzable prodrug of claim 3 wherein the amino-terminal capped peptide is benzyloxycarbonylphenylalanyllysine, benzyloxycarbonylvalinyllysine, D-phenylalanylphenylalanyllysine, benzyloxycarbonylvalinylcitrulline, t-butyloxycarbonylphenylalanylysine, benzyloxycarbonylalanyllarginylarginine, benzyloxycarbonylphenylalanyl-N-tosylarginine,2-aminoethylthio-succinimidopropionylvalinylcitrulline, 2-aminoethylthio-succinimidopropionyllysylphenylalanyllysine, acetylphenylalanyllysine, or benzyloxycarbonylphenylalanyl-O-benzoylthreonine.

5. The hydrolyzable prodrug of claim 4 wherein the amino-terminal capped peptide is benzyloxycarbonylphenylalanyllysine.

6. The hydrolyzable prodrug of claim 4 wherein the amino-terminal capped peptide is acetylphenylalanyllysine.

7. The hydrolyzable prodrug of claim 4 wherein the amino-terminal capped peptide is 2-aminoethylthio-succinimidopropionylvalinylcitrulline.

8. The hydrolyzable prodrug of claim 4 wherein the amino-terminal capped peptide is 2-aminoethylthio-succinimidopropionyllysylphenylalanyllysine.

9. The hydrolyzable prodrug of claim 1 wherein the therapeutic drug is an anticancer drug.

10. The hydrolyzable prodrug of claim 9 wherein the anticancer drug is doxorubicin, mitomycin C, taxol, esperamycin, or camptothecin.

11. The hydrolyzable prodrug of claim 10 wherein the anticancer drug is doxorubicin.

12. The hydrolyzable prodrug of claim 1 wherein the self-immolating spacer is selected from the group consisting of p-aminobenzyl carbonyl and p-NH-Ph-CH=(CH2OCO-)2.

13. The hydrolyzable prodrug of claim 11 wherein the spacer is p-aminobenzyl carbonyl.

14. The hydrolyzable prodrug of claim 1 further comprising a peptide derived from a protein to which metastatic cells adhere in establishing colonies covalently linked to the therapeutic drug.

15. The hydrolyzable prodrug of claim 14 wherein the peptide is a RGD-derived active peptide or a YIGSR-derived active peptide.

16. The hydrolyzable prodrug of claim 14 wherein the peptide is YIGSR (SEQ ID NO:1) or GRGDS (SEQ ID NO:2).

17. Benzyloxycarbonylphenylalanyllysyl-p-aminobenzyl carbamoyldoxorubicin.

18. Acetylphenylalanyllysyl-p-aminobenzyl carbamoyldoxorubicin.

19. Acetylphenylalanyllysyl-p-aminobenzyl carbamoylmitomycin C.

20. Benzyloxycarbonylphenylalanyllysyl-p-aminobenzyl carbonyl-7-paclitaxel.

21. Acetylphenylalanyllysyl-p-aminobenzyl carbonylcamptothecin.

22. 2-aminoethylthio-succinimidopropionyl-valinylcitrullinyl-bis(hydroxymethyl)styryl-bis-doxorubicin.

23. 2-aminoethylthio-succinimidopropionyl-lysylphenylalanyllysyl-bis(hydroxymethyl)styryl-bis-doxorubicin.

24. Benzyloxycarbonylvalinyllysyl-p-aminobenzyl carbamoyldoxorubicin.

25. D-phenylalanylphenylalanyllysyl-p-aminobenzyl carbamoyldoxorubicin.

26. Benzyloxycarbonylvalinylcitrullinyl-p-aminobenzyl carbamoyldoxorubicin.

27. A method for delivering a therapeutic drug to a metastatic cell comprising the steps of:

(a) contacting a hydrolyzable prodrug comprising an amino-terminal capped peptide covalently linked to a therapeutic drug through a self-immolating spacer of sufficient length to prevent the occurrence of steric hindrance with a metastatic cell, the amino-terminal capped peptide being a substrate for a peptidohydrolase located on the surface of the metastatic cell;

(b) allowing the peptidohydrolase located on the surface of the metastatic cell to hydrolyze the hydrolyzable prodrug and release the therapeutic drug from the prodrug; and (c) allowing the therapeutic drug to enter the metastatic cell.

28. The method of claim 27 wherein the peptidohydrolase is selected from the group consisting of cathepsin B and collagenase IV.

29. The method of claim 28 wherein the peptidohydrolase is cathepsin B.

30. The method of claim 29 wherein the amino-terminal capped peptide is benzyloxycarbonylphenylalanyllysine, benzyloxycarbonylvalinyllysine, D-phenylalanylphenylalanyllysine, benzyloxycarbonylvalinylcitrulline, t-butyloxycarbonylphenylalanylysine, benzyloxycarbonylalanyllarginylarginine, benzyloxycarbonylphenylalanyl-N-tosylarginine,2-aminoethylthio-succinimidopropionylvalinylcitrulline, 2-aminoethylthio-succinimidopropionyllysylphenylalanyllysine, acetylphenylalanyllysine, or benzyloxycarbonylphenylalanyl-O-benzoylthreonine.

31. The method of claim 30 wherein the amino-terminal capped peptide is benzyloxycarbonylphenylalanyllysine.

32. The method of claim 30 wherein the amino-terminal capped peptide is acetylphenylalanyllysine.

33. The method of claim 30 wherein the amino-terminal capped peptide is 2-aminoethylthio-succinimidopropionylvalinylcitrulline.

34. The method of claim 30 wherein the amino-terminal capped peptide is 2-aminoethylthio-succinimidopropionyllysylphenylalanyllysine.

35. The method of claim 27 wherein the therapeutic drug is an anticancer drug.

36. The method of claim 35 wherein the anticancer drug is doxorubicin, mitomycin C, taxol, esperamycin, or camptothecin.

37. The method of claim 36 wherein the anticancer drug is doxorubicin.

38. The method of claim 27 wherein the spacer is selected from the group consisting of p-aminobenzyl carbonyl and p-NH-Ph-CH=(CH2OCO-)2,

39. The method of claim 38 wherein the spacer is p-aminobenzyl carbonyl.

40. The method of claim 27 wherein the hydrolyzable prodrug further comprises a peptide derived from a protein to which metastatic cells adhere and in establishing colonies covalently linked to the therapeutic drug.

41. The method of claim 40 wherein the peptide is a RGD-derived active peptide or a YIGSR-derived active peptide.

42. The method of claim 41 wherein the peptide is YIGSR
(SEQ ID NO:1) or GRGDS (SEQ ID NO:2).

43. The method of claim 27 wherein the hydrolyzable prodrug is benzyloxycarbonylphenylalanyllysyl-p-aminobenzyl carbamoyldoxorubicin, acetylphenylalanyllysyl-p-aminobenzyl carbamoyldoxorubicin, acetylphenylalanyllysyl-p-aminobenzyl carbamoylmitomycin C, benzyloxycarbonylphenylalanyllysyl-p-aminobenzyl carbonyl-7-paclitaxel, acetylphenylalanyllysyl-p-aminobenzyl carbonylcamptothecin, 2-aminoethylthio-succinimidopropionyl-valinylcitrullinyl-bis(hydroxymethyl)styryl-bis-doxorubicin, 2-aminoethylthio-succinimidopropionyl-lysylphenylalanyllysyl-bis(hydroxymethyl)styryl-bis-doxorubicin, benzyloxycarbonylvalinyllysyl-p-aminobenzyl carbamoyldoxorubicin, D-phenylalanylphenylalanyllysyl-p-aminobenzyl carbamoyldoxorubicin, or benzyloxycarbonylvalinylcitrullinyl-p-aminobenzyl carbamoyldoxorubicin.

44. A pharmaceutical composition comprising:

(a) the hydrolyzable prodrug of claim 1; and (b) a pharmaceutically acceptable carrier.

45. The pharmaceutical composition of claim 44 wherein the hydrolyzable prodrug further includes a peptide derived from a protein to which metastatic cells adhere in establishing colonies covalently linked to the therapeutic drug.

46. The pharmaceutical composition of claim 45 wherein the peptide is a RGD-derived active peptide or a YIGSR-derived active peptide.

47. The pharmaceutical composition of claim 46 wherein the peptide is