WO1996005865A1 - Combination immunotoxin/antineoplastic agent therapy for b-lineage cancer - Google Patents

Abstract

A method is provided which utilizes highly potent and cell-type specific immunotoxins (IT) to systemically treat therapy refractory high risk ALL patients. The method of the present invention decreases the post BMT relapse rate by eliminating the radiation resistant and/or drug resistant residual leukemia burden. With slight modifications the method of the present invention should be generally applicable to preparation of other PAP-MoAb conjugates for treatment of other types of cancer or AIDS.

Description

THERAPY FOR B-T JNFAGE CANCER Background of ttie Invention

Acute I. mphoMastic Leukemia.

Acute lymphoblastic leukemia (ALL) is the most common form of childhood malignancy. Champlin et al., Blood, 11, 2051 (1989). Each year about 1250 children less than 15 years of age are found to have acute lymphoblastic leukemia. Champlin et al., cited supra. Recently, dramatic improvements in the multiagent chemotherapy of children with ALL have resulted in cure rates of 70-75%. Poplack et al., Pediatric Clinics of North America, Vol. 35, No.4, pp 903-932, (1988). However, despite these recent improvements as many as 1 in 5 patients will eventually suffer leukemic relapse. Riehm et al.. Haemalol. Blood Transf. 33. 439 (199QY This occurrence of relapsed patients equates to 250 cases/year and is equivalent to the number of newly diagnosed cases of childhood acute nonlymphoblastic leukemia, medulloblastoma, and mabdomyosarcoma. Furthermore, this relapse rate surpasses the number of newly diagnosed cases of childhood Ewings sarcoma, osteogenic sarcoma, hepatoma, and germ cell tumors. The unsatisfactory outcome of this population makes a significant contribution to overall pediatric cancer mortality, despite the excellent outcome for the substantial majority of children with ALL.

Currently, the major challenge in the treatment of childhood ALL is to cure patients who have relapsed despite intensive multiagent chemotherapy. Champlin et al., cited supra. For patients who have relapsed while on therapy or shortly after elective cessation of therapy, the overall survival is very poor. Poplack et al., cited supra. Treatment of these relapsed children has generally employed either intensive chemotherapy to achieve a second remission, subsequent use of either nonablative chemotherapy or ablative radiochemotherapy and bone marrow transplantation (BMT). Kersey et al., N. Engl. J. Ivied.. 317. 461 (1987). However, recurrence of leukemia is the major obstacle to the success of either approach. Dicke et al., Clin. HemaloL 15, 86 (1986). Notably, recent studies demonstrated that primary clonogenic blasts from B-lineage ALL patients are among the most radiation resistant human tumors (Uckun et al., J. Gin. Invest. 21, 1044 (1993)) and they are capable of repairing sublethal radiation damage. Uckun et al., Cancer Res., 51, 1431 (1993). Therefore, it is not surprising that the vast majority of B- lineage ALL patients undergoing BMT relapse despite the use of total body irradiation (TBI) regimens. Kersey et al., N. Engl. J. Med.. 3JI, 461 (1987). Furthermore, treatment of these relapsed patients by the intensification of cytotoxic therapy using conventional drugs will likely cause overlapping toxicities and may result in delays which may erode the intensity of therapy. Consequently, the development of new potent anti-ALL drugs and the design of combinative treatment protocols utilizing these new agents, have emerged as focal points for research in the therapy of relapsed ALL. Inin unotoxins

Immunotoxins (antibody-toxin conjugates) are a relatively new class of immuTiopharrnacologic agents that are prepared by covalently linking cell type- specific polyclonal or monoclonal antibodies to a variety of cytotoxins either directly or via a linking agent. Immunotoxin therapy provides an alternative strategy that may enhance anti-leukemic effect with non- overlapping toxicities. 1. Monoclonal Antibodies

Monoclonal antibodies (MoAbs) are produced by the fusion of spleen lymphocytes with malignant cells (myelomas) of bone marrow primary tumors. Milstein, Sci. Am.. 243_, 66 (1980). The procedure yields a hybrid cell line, arising from a single fused cell hybrid, or clone, which possesses characteristics of both the lymphocytes and myeloma cell lines. Like the lymphocytes (taken from animals primed with sheep red blood cells as antigens), the fused hybrids or hybridomas secrete antibodies (immunoglobulins) reactive with the antigen. Moreover, like the myeloma cell lines, the hybrid cell lines are immortal. Specifically, whereas antisera derived from vaccinated animals are variable mixtures of antibodies which cannot be identically reproduced, the single-type of immunoglobulin secreted by a hybridoma is specific to one and only one determinant on the antigen, a complex molecule having a multiplicity of antigenic molecular substmrtures, or determinants (epitopes). Hence, monoclonal antibodies raised against a single antigen may be distinct from each other depending on the determinant that induced their formation. However, all of the antibodies produced by a given clone are identical. Furthermore, hybridoma cell lines can be reproduced indefinitely, are easily propagated in vitro and in vivo, and yield monoclonal antibodies in extremely high concentration.

B43 is a murine IgGl, K monoclonal antibody (MoAb) recognizing a 95 kDa target B lineage restricted phosphoglycoprotein, which is identified as the CD19 antigen according to the World Health Organization (WHO) established CD (cluster of differentiation) nomenclature. The chemical, immunological and biological features of B43 MoAb have been described in detail in previously published reports. Uckun et al., Blood 7_i, 13 (1988). 2. Toxins The variety of toxins that have been employed in immunotoxins by various investigators can be broadly categorized into two groups. The first group consists of intact toxins, such as intact ricin. Ricin consists of two subunits, the A chain which is capable of inactivating as many as 1,500 ribosomes per minute and the B chain which recognizes non-reducing terminal galactose residues on cell surfaces and facilitates A chain entry. Although intact ricin immunotoxins are highly effective destroyers for their target cells, they cannot be applied for in vivo treatment of leukemia because of the nonselectabity of their B chain moiety.

The second group of toxins are referred to as hemitoxins. Hemitoxins are single-chain ribosome inactivating proteins that act catalytically on eukaryotic ribosomes and inactivate the 60-S subunit, resulting in an irreversible shut-down of cellular protein synthesis at the level of peptide elongation. Such polypeptide toxins have been isolated from pokeweed (Phytolacca americana), bitter gourd (Momordica chattntia), wheat (Tritium vulgcris), soapwort (Scponaria of cindis), Gelonium multiflorum, and several other plants. Since these ribosome inactivating proteins, unlike intact ricin, do not have a B chain subunit with nonselective cell binding capacity, they cannot easily cross the cellular membrane. Therefore, hemitoxins are practically devoid of toxicity to intact eukaryotic cells. a. PAP

There are three subtypes of pokeweed antiviral protein (PAP) the expression of which are dependent upon the season. PAP is found in spring leaves, PAP II is found in late summer leaves, and PAP-S is found in the seeds. Irvin, Pharmacol. Ther.. 21, 371 (1983). Small differences exist in their sizes (all are approximately 29,000 MW) and there are only small differences, if any, between their ability to inhibit ribosomes caialytically. Houston et al., "Immunotoxins made with Toxins and Hemitoxins other than Ricin", in ϊmmunological Antibody Conjugates in Radioimaging and Therapy of Cancer. C.W. Vogel, ed., New York, Oxford University Press, P. 71 (1987).

PAP is a member of the hemitoxin group of toxins and thus inactivates ribosomes by the specific removal of a single adenine from the conserved loop sequence found near the 3' teπriinus of all larger rRNAs. Irvin et al., Pharmacology and Therapeutics. 5 i, 279, (1992). This specific depiirination greatly reduces the capability of elongation factors to interact with ribosomes and results in an irreversible shut-down of protein synthesis. Irvin et al., cited supra,

3. B43-PAP

B43-PAP is an anti-CD19 immunotoxin composed of anti-human CD19 monoclonal antibody B43 covalently coupled to the ribosome inhibitory plant toxin PAP (BB-IND-3864). U.S. Patent 4,831 , 117, issued to Uckun, discloses B43-PAP and that it may be useful for bone marrow purging in the treatment of ALL as well as for the treatment of patients with B-lineage leukemia/lymphoma. However, this patent contains only in vitro examples employing human B-lineage ALL progenitor cells, and the problems encountered in the clinical development of immunotoxins are largely due to the lack preclinical model systems which allow an accurate prediction of possible side effects or the efficacy of a given immunotoxin in humans. Limitations on Biotherapy

Although there is potential for immunotoxins to provide effective therapy for human diseases, many difficulties remain. Ideally, consistently locatable and reliable markers on target cells would permit the binding portion of immunoconjugates to completely avoid non-target tissue. In reality, cross- reactivity with antigens expressed by vital life-maώtaining organs has been sufficient to being an early close to some phase I trials. Patients in dose escalation studies have also demonstrated immune responses to antibody and toxin components of immunotoxins, even in cases where patients were immunosuppressed by the course of their disease. Moreover, while immunotoxins have shown remarkable cell-type specific cytotoxicity in vitro and in vivo, clinical results are limited by the lack of potency of tolerated doses. Furthermore, solid tumors are difficult to penetrate thoroughly and, in hematologic malignancies, residual disease can cause relapse despite easier access to target cells in leukemias and lymphomas.

Also, toxicity studies using immunotoxins in mice and monkeys have not been predictive of the toxicity of the immunotoxins in clinical trials. For example, no neurotoxicity has been observed in monkeys treated with ricin A chain immunotoxins directed to B-cell surface antigens CD19 or CD22. However, when these immunotoxins were used in patients with lymphoma, a significant fraction showed peripheral neuropathy as well asaphsia (loss of speech).

Similarly, no neurotoxicity was observed in preclinical animal studies using a recombinant ricin A chain immunotoxin of 454A12 mouse antitrramferrin receptor monoclonal antibody or a natural pseudomonas exotoxin immunotoxins of OVB3 mouse anti-adrenocarcinoma monoclonal antibody. Both immunotoxin caused lethal neurotoxicity with severe encephalopathy and brainstem .nflammation when used in patients with cancer. Immunotoxins have been used in a number of phase I-II clinical trials in patients with lymphoid malignancies. Grossbard et al., BlΩΩit £Ω, 863 (1992); Hertler et al., J. Clin. Oncol.. 1, 1932 (1989). However, the objective response rates have been in the order of only 1-5%. Uckun, Brit. J. Haematology. £5. 435 (1993).

For instance, although the efficacy of an anti-CD19 (B4) blocked ricin immunotoxin was demonstrated by the depletion of leukemia cells in vitro as well as in surrogate mouse models of human leukemia (Uckun et al., Leukemia. 2, 341 (1993)), these results were not reproducible in human clinical trials. In a phase I clinical trial with B4-blocked ricin, 25 patients with refractory B-lineage leukemia or lymphoma were treated with bolus infusions with the immunotoxin. (Nadler et al., Proceedings of the 2nd Intemational Symposium on Immunotoxins. p. 58.) Only one lymphoma patient with B-lineage lymphoma had a complete response and significant hepatotoxicity was observed in most patients. Other toxicities included thrombocytopenia, fevers, fatigue, capillary leak syndrome, and significant hepatotoxicity. Thirteen patients generated human anti-mouse antibodies (HAMA) and human anti-ricin antibodies (HARA).

Another phase I trial using B4-blocked ricin involved 43 leukernia/lymphoma patients who received a 7-day continuous infusion rather than bolus injections. Grossbard et al., "Immunotoxin Therapy of Malignancy", in Important Advances in Oncology. DeVita et al., eds., Lippincott, Philadelphia, pp 112-136 (1992). None of the ALL patients responded and only two lymphoma patients had complete responses. Hepatotoxicity was dose-limiting with marked elevation of serum levels of the liver leakage enzymes SGOT and SGPT. Other toxicities included capillary leak syndrome, thrombocytopenia, and fever. Twenty-six patients developed antibodies against HAMA or HARA In fact, B4-blocked ricin trials in

Children's Cancer Groups could not be completed due to marked liver toxicity and development of HARA/HAMA, and no child with ALL had a therapeutic response. Similarly, an anti-T cell immunotoxin containing ricin A chain was found toxic and inneffective in teenage ALL patients of the Pediatric Oncology Group.

In another in vivo study, modified ricin was utilized in an effort to reduce nonspecific toxicity and prevent rapid clearance by the reticuloendothelial system. The modification involved chemically or enzymatically deglycosylating the A chain, while eliminating B chain altogether. In this study, twenty-two patients with advanced B-cell malignancies were treated with the modified (deglycosylated) ricin A chain (dgA) linked to an anti-B cell (CD22) whole antibody or Fab fragment. Vitetta et al., Cancer Res.. 5 ., 4052 (1991). None of the patients had a complete response and significant toxicities were observed when either dgA- whole antibody or Fab fragment conjugates were used, mcluding pulmonary edema, aphasia, rhabdomyolysis with renal failure, and severe capillary leak syndrome.

Capillary leak syndrome has also been observed in human patients treated with immunotoxins comprising intact ricin, ricin A chain, diphtherea toxin or pseudomanas exotoxins. The latter two side effects has not been observed in animals, but can be life-threatening in humans. These results illustrate some recurring problems with immunotoxin therapy for the treatment of leukemias or lymphomas. Overcoming dose-limiting toxicities caused by cross-reactivity with healthy cells and organs, the immunogenicity when antibodies against toxin or carrier are produced by the recipient, and the limited therapeutic responses are among the most obvious and serious challenges.

Therefore there is continuing need for improved therapies to treat ALL patients, particularly those in relapse.

Summary of the Invention The present invention provides a therapeutic method for the treatment of acute lymphoblastic leukemia (ALL). The method comprises parenterally administering to a patient who is afflicted with ALL an effective anti- leukemic amount of a pharmaceutical composition comprising an immunotoxin consisting of monoclonal antibody B43 covalently linked to an effective cytotoxic amount of PAP, in combination with a pharmaceutically acceptable carrier. While the preparation of one B43-PAP immunotoxin is disclosed in U.S. Patent No. 4,831,117, the term "B43-PAP" as used herein is intended to encompass EH 3 functionally linked to cytotoxic amounts of PAP by other means known to the art. As used herein, the term "PAP" refers to any cytotoxic pokeweed antiviral protein, or subunit thereof, mchiding subtypes PAP-II and PAP-S. The term "cytotoxic amount" is defined to mean an amount of PAP that is toxic to the target cell once the immunotoxin has associated with the cell.

The present method is especially suited for the treatment of relapsed ALL patients who have failed other types of therapies. Furthermore, since ALL is the most common childhood malignancy, the method is also of particular value for children, i.e., for patients who are under the age of 18. Monoclonal antibody B43 specifically targets antigen CD19, which is expressed on the surface of leukemic blasts from 100% of B-lineage ALL patients. However, peripheral cancer cells that lack the target antigen may present complications in the treatment of certain patients. In these cases, combined or adjunctive therapies that exploit the diverse cytotoxic mechanisms offered by conventional chemotherapy and radiation assist in the elimination of any cancer cells that lack the target antigen as well as in the suppression of immunotoxin-resistant mutants. Therefore, one aspect of the invention is the use of B43-PAP in conjunction with one or several other known antineoplastic agents, such as cyclophosphamide or etoposide. Thus, one embodiment of the present invention comprises the admmstration of B43-PAP in combination with, e.g., followed by, the parenteral administration of an effective anti-leukemic amount of one or more conventional antineoplastic agents. Preferably, the antineoplastic agent employed is an anti-metabolite or a class I or a class III immunosuppressive agent. Preferably, the antineoplastic agent employed is cytarabine, cyclophosphamide or etoposide. It is also preferred that the antineoplastic agent be combined with a phaπriaraitically acceptable liquid carrier at a concentration of from about 10 mg ml to about 30 mg/ml. In this embodiment of the invention, it is preferred that the cyclophosphamide or cytarabine be adrriinistered intravenously. Preferably, the cyclophosphamide is -ukninistered at the rate of 0.5-3.5 IJM²/24 hours. For some patients, bone marrow transplantation (BMT) may remain the best prospect for survival, in which case immunotoxins might be used in vivo in addition to radiochemotherapy as part of pre-BMT conditioning. Thus, one embodiment of the present invention comprises the systemic administration of B43-PAP followed by radiochemotherapy, prior to BMT.

The present invention is based on my discovery that LPC (i.e, primary clonogenic blasts) from ALL patients are surprisingly sensitive to PAP- containing immunotoxins targeted to appropriate surface antigens, which antigens are capable of antibody-induced intemalization. Uckun et al., Blood. 6, 2449, (1990). This heightened cytotoxicity can be attributed to the fact that the antibody B43 is specific for the surface antigen CD19. CD19 displays exclusive B-lineage specificity, undergoes antibody-induced intemalization, and is expressed on the majority of clonogenic ALL blasts. Extensive studies have provided unambiguous evidence that CD19 antigen is not expressed in non-lymphohematopoietic tissues. Uckun. Blood. 76. 1 08 (1990). The antigen is expressed by virtually 100% of B-lineage ALL cases, is present on each leukemic B-lineage ALL blast at a high density, shows a high affinity for B43 MoAb, and undergoes antibody induced intemalization upon binding of B43 MoAb. Uckun et al., Blood, 21, 13 (1988). Nonetheless, the efficacy of B43-PAP coupled with the absence of dose-limiting side effects was unexpected for a number of reasons. For example, although B43-PAP is specific for B-lineage cells, and was potent in vitro and in mouse models of leukemia, efficacy studies in surrogate mouse models of human cancer or in vitro clonogenic assays have also not been predictive of clinical activity of any immunotoxin.

Apart from the failures of other promising immunotoxins in clinical trials, discussed above, preclinical animal studies using B43-PAP immunotoxin led to observations of marked kidney, liver, muscle and myocardial (heart) toxicity. In mice, B43-PAP had a short half-life similar to other immunotoxins. Its immunogenicity in mice, rabbits and monkeys were similar to those of immunotoxins prepared with toxins other than PAP. However, the Phase I/II study in children and young adults (<30 years) reported hereinbelow, gave rise to a number of unexpectedly positive results. For example, humans, expecially children with leukemia were found to tolerate B43-PAP much better than mice or monkeys. Also, the human patients did not experience the hepatotoxicity or renal toxicity observed in mice and monkey studies. B43-PAP did not give rise to the neurotoxicity fevers or the capillary leak syndrome observed in animals and associated with ricin A chain immunotoxins. A serum concentration of 1 μg ml has consistently led to life threatening capillary leak syndrome when ricin A chain immunotoxins are adrninistered. In contrast, B43-PAP concentrations as high 13 μg ml did not cause capillary leak. B43-PAP is much less immunogenic than any other immunotoxin, with significant human anto-mouse antibody or human anti-toxin antibody responses observed in a very small fraction of patients (3 of 39 patients). Therefore, most patients are able to receive multiple cycles of B43-PAP, unlike patients who have been administered B4- blocked ricin or other immunotoxins containing toxins other than PAP. From a drug delivery standpoint, B43-PAP is very stable in humans, with long half- lives and Area Under Curve (AUC) values that are markedly superior to those reported for any other immunotoxin. Therefore, serum concentrations as high as 13 μg/ml can be attained. Furthermore, the potent anti-leukemic activity of the immunotoxin

B43-PAP is not affected by oxazaphosphorine resistance, classical or atypical multidrug resistance, or radiation resistance of target ALL blast populations.

Another potential problem in using immunotoxins in vivo is related to the presence of carbohydrate residues in the toxin moieties. Reticuloendothelial cells, including Kupffer cells in the liver, express receptors for carbohydrates which may result in rapid clearance and short activity of immunotoxins as well as a significant liver toxicity. In this regard, the method of the present invention is of special value as B43-PAP lacks ∞rbohydrate residues. It is also expected that this immunotoxin will be effective in the treatment of other diseases associated with the proliferation of mammalian cells comprising CD19. Such diseases include other cancers, such as non- Hodgkins lymphomas of B-cell origin, myelomas, or AIDS lymphoma, B43- PAP may also be useful to treat autoimmune diseases including, but not limited to, systemic lupus erythematosus, rheumatoid arthritis, non-glomerular nephrosis, psoriasis, chronic active hepatitis, ulcerative colitis, Crohn's disease, Behcet's disease, chronic glomerulonephritis (membranous), chronic thrombocytopenic purpura, allograft rejection and autoimmune hemo.ytic anemia.

Brief Description of the Figures Figure 1 is a graphical depiction of the effectiveness of proportion of SCID mice with human leukemia surviving event free after treatment with PBS, B43-PAP, ARA-C and B43-PAP+ARA-C.

Detailed Description of me Invention Production and Purification of B43-PAP Immunotoxin

Preferred B43-PAP immunotoxins for use in the method are formed by linking an effective cytotoxic amount of PAP molecules to each molecule of B43. For example, a reagent useful in the practice of the invention is an about 1:1 nτixtιιre of B43-PAP having one and two PAP molecules per B43 molecule, respectively.

The particular B43-PAP employed in the examples hereinbelow is prepared by linking B43 MoAb to PAP as described in U.S. patent no.

4,831,117, to Uckun, which is incorporated herein by reference. A hybridoma secreting B43 is available from the ATCC under designation HB 8903. Further information concerning the production and purification of B43-PAP can be found in Examples 1-4. However, B43 can be linked to effective amounts of PAP by other means disclosed in the art, including those taught in U.S. Patent Nos 4,363,758, Masuho et al.; 5,167,956, Neville, Jr. et al. and 4,340,535, Voisin et al. For example, in addition to N-succinimidyl 3-(2- pyridyldithio)propionate (SPDP), 4-su ;ininτidyloxycarbonyl-methyl-(2- pyridyldithio)-toluene (SMPT) and N-succimidyl 6-[3-(2- pyriαyldthio)propionamido]hexanoate (LC-SPDP) may be used as linking agents. Methods of preparing B43-PAP immunotoxin utilizing these linking agents are given in Example 3, parts 2b and 2c. Adjunct Antineoplastic Agents

A preferred adjunct antineoplastic agent for use with B43-PAP is a Class I or Class III immunosuppressive drug or an antimetabolite. Representative antimetabolites include cytarabine, mercaptopurine, methotrexate, thioguanine, etc. Representative immunosuppressive agents useful in the invention include, but are not limited to asparaginase, cyclophosphamide, daunorubicin, doxorubicin, etoposide, mafosfamide, melphalan and vincristine. Preferably the method of the present invention utilizes either cyclophosphamide or cytarabine in the combination therapy with the immunotoxin.

1. (. clophosphamiάe clophosphamide or "cytoxan" is an alkylating agent and thus mainly affects the short-lived and not the long-lived small lymphocytes. This immunosuppressive agent also suppresses proliferation of macrophages but does not interfere with phagocytosis. Therefore, the primary immune response is mainly affected and cyclophosphamide performs best as a Class I drug.

In vivo efficacy of B43-PAP plus cytoxan against human B-lineage ALL was evaluated in SCID mice. Notably, the combination of B43-PAP with Cytoxan was substantially more effective against human NALM-6 pre-B ALL than B43-PAP alone or Cytoxan alone. Uckun et al., Blood, 22, 3116 (1992). Similar results were obtained using the RS4;11 SCID mouse model of t(4;l 1) infant ALL. Jansen et al., Lej±ej ia, 2, 290 (1992).

Notably, B43-PAP as a single agent is more potent than cyclophosphamide, vincristine, VP-16, methylprednisone, L-asparaginase, adriamycin, BCNU, cytosine arabinoside, topotecan, or taxol against human pre-B ALL in the SCID mouse model system. However, synergjsm was observed with oxazaphosphorines (i.e. cyclophosphamide), cytosine arabinoside, and topotecan. Furthermore, it has been found that etoposide appears to increase immunotoxin toxicity and dexamethasone mitigated immunotoxin toxicity without diminishing immunotoxin effect. 2. Antimetabolites

There are three subcategories of antimetabolites: purine analogs, pyrirnidine analogs and folinic acid analogs. The purine analogs are incorporated into DNA as the deoxyribotides and into RNA as the ribotides, where they interfere with coding and replication. They also act like the natural purine bases in inhibiting synthesis of purine bases by acting through the allosteric feedback systems (pseudo-feedback). The pyrirnidine analogs inhibit enzymes in the biosynthetic pathways for pyrirnidine ribotides and deoxyribotides; thymidylate synthetase, orotic acid decarboxylase, aspartate carbamoyltransferase and dihydroorotase are inhibited. Methotrexate and trimetrexate are the only folinic acid analogues in use; they bind very tightly to dihydrofolate reductase and thereby prevent the conversion of dihydrofolate (folinate to tetrahydrofolate). Antimetabolites useful in the present invention include, but are not limited to, methotrexate, trimetrexate, 5-fluorouracil, Q arabine, mercaptopurine, tWoguanine, 5-azacitidine, floxuridine, 2"- chlorodeoxyadenosine, and the like. a. Cytarabine

Cytarabine is a pyrirnidine nucleoside antimetabolite that is cytotoxic to a number of cell types. It competes with deoxycytidine and also interferes with in orporation of uridine into deoxycytidine nucleotides. This immunosuppressive agent also suppresses primary responses in doses that cause little or no other toxicity. It is component of first choice combinations to treat both acute and chronic myeloblastic leukemias and non-Hodgkin's and Birrkitt's lymphomas. Additionally, by the intraventricular route, it is the drug of choice to treat leukemic metastases in the central nervous system and also other meningeal soft-tissue metastases.

In vivo efficacy of B43-PAP plus cytarabine against human B-cell precursor (BCP) leukemia was evaluated in SCID mice. Notably, the combination of B43-PAP with cytarabine was substantially more effective against human NALM-6 pre-B ALL in a SCID mouse model than B43-PAP alone or cytarabine alone. Furthermore, none of the other combinations tested, including B43-PAP plus vincristine, me ylprednisone, L-asparaginase, carmustine doxorubicin and etoposide proved more effective than B43-PAP alone in this model.

Modes of Admini-rtration of the B43-PAP Immunotoxin

B43-PAP can be formulated as pharmaceutical composition and administered to a mammalian host, such as a human cancer patient, in a variety of forms adapted to the chosen route of administration, i.e., orally or parenterally. In the practice of the present method, it is preferred that the B43-PAP immunotoxin be parenterally administered, i.e., intravenously or mtraperitoneally by infusion or injection. Solutions or suspensions of the immunotoxin can be readily prepared in water, or isotonic saline, such as PBS, optionally mixed with a nontoxic surfactant. Under ordinary conditions of storage and use, these preparations may contain a preservative to prevent the growth of microorganisms.

The pharmaceutical dosage form suitable for injection or infusion use can include sterile aqueous solutions or dispersions or sterile powders comprising the active ingredient which are adapted for the extemporaneous preparation of sterile injectable or infusible solutions or dispersions. In all cases, the ultimate dosage form must be sterile, fluid and stable under the conditions of manufacture and storage. The liquid carrier or vehicle can be a solvent or liquid dispersion medium comprising, for example, water, in combination with minor but effective amounts of ethanol, a polyol (for example, glycerol, propylene glycol, and liquid polyethylene glycols, and the like), vegetable oils, nontoxic glyceryl esters, and suitable nrixtures thereof. The proper fluidity can be rriaintained, for example, by the formation of liposomes, by the maintenance of the required particle size in the case of dispersion or by the use of nontoxic surfactants. The prevention of the action of microorganisms can be accomplished by various antibacterial and antifungal agents, for example, parabens, chlorobutanol, phenol, sorbic acid, thimerosal, and the like. In many cases, it will be desirable to include isotonic agents, for example, sugars, buffers or sodium chloride. Prolonged absorption of the injectable compositions can be brought about by the inclusion in the compositions of agents delaying absorption, for exampie, aluminum monostearate hydrogels and gelatin.

Sterile injectable solutions are prepared by mcorporating the immunotoxins in the required amount in the appropriate solvent with various of the other ingredients enumerated above, and as required, followed by filter sterilization. In the case of sterile powders for the preparation of sterile injectable solutions, the preferred methods of preparation are vacuum drying and the freeze drying techniques, which yield a powder of the active ingredient plus any additional desired ingredient present in the previously sterile-filtered solutions. Dosages

The concentration of B43-PAP in the liquid carrier vehicle can be varied widely, in accord with the size, age and condition of the patient, but is preferably from about 0.05 mg/ml to about 0.25 mg/ml, wherein the amount of B43-PAP administered is preferably from about 0.1 μg/kg/day to about 250 μg/kg/day.

Useful dosages of the B43-PAP immunotoxins were determined by clinical trials or Phase I trials and an effective dose was determined to be 100 μg/kg/day for 5 days, for children or adults as discussed below. This treatment cycle can be repeated for as long as clinical improvement is evident.

1. Combination Therapies

In the cases where B43-PAP is to be administered in conjunction with an adjunct or a second antineoplastic agent, it is preferred that the antineoplastic agent be either an antimetabolite or a class I immunosuppressive drug. More preferably, the adjunct antineoplastic agent is preferably chosen from cytoxan (cyclophosphamide) or cytarabine. When the B43-PAP is administered concurrently with cytoxan it is preferred that the cytoxan be administered in an amount from about 1 mg/kg/day to about 5 mg/kg/day for 5 days. When the B43-PAP is administered concurrently with cytarabine it is preferred that the cytarabine be administered in an amount from about 1 mg/kg/day to about 5 mg/kg day for 5 days. Thus, the invention also provides a product, such as a treatment kit, containing, separately packaged, in association, an effective treatment amount of B43-PAP and an effective treatment amount of the second antimeoplastic agent adapted for simultaneous or sequential use in treating ALL, i.e., by parental administration.

The invention will be further described by reference to the following detailed examples, wherein the following general materials and methods were utilized.

GdLIinss The human NALM-6 pre-B cell ALL cell line was maintained by serial passages in RPMI 1640 medium (GIBCO Laboratories, Grand Island, NY) supplemented with 10% (v/v) heat-inactivated calf-bovine serum (Hyclone Laboratories, Logan, UT), 50 μg/mL streptomycin, 50 IU/mL penicillin, 2 mmpl/L L-glutamine, and 10 mmol/L Hepes buffer. Cells were cultured in tissue culture flasks at 37°C in a humidified 5% C j atmosphere. Before injection into SCID mice, cells were washed twice in phosphate- buffered saline (PBS) and resuspended in PBS at 5 x 10⁶ cells/mL. SCID mice were inoculated intravenously with 0.2 mL of these cell suspensions containing 1 x 10⁶ NALM-6. SJ3D_ιmce.

All SCID mice were produced by specific pathogen-free (SPF) CB-17 scid/scid breeders (originally obtained from Dr. Melvin Bosma, FOx Chase Cancer Center, Philadelphia, PA) and maintained in the AALAC-accredited Research Animal Resources (RAR) SCID mouse Facility of the Childrens Cancer Group ALL Biology Reference Laboratory at the University of Minnesota (Minneapolis, MN). SCID mice were maiirtained in a SPF environment in microisolator cages (Lab Products, Inc., Maywood, NY) containing autoclaved food, water, and bedding. TrimemopririVsulfamemox-G»le (Bactrim) was added to the drinking water of mice and was changed three times a week.

Vinicristine was obtained from Eli Lilly Industries, Inc., Carolina, Puerto Rico, methylprednison was obtained fjτ>m the Upjohn Co., Kalamazoo, MI, L-asparaginase was obtained from Merck Sharp & Dohme, West Point, PA, Doxorubicin was obtained from Adria Laboratories, Columbus, OH, etoposide was obtained from Bristol-Myers Squibb Company, Evansville, IN, Carmustine (BCNU) and Cytarabine (ARA-C) were obtained from Ben Venue Laboratories, Inc., Bedford, OH.

Example 1

Large Scale Production and Piπifif afion of Pokeweed Antiviral Protein rPAP.

PAP from spring leaves was prepared according to the following procedure. 5 kg of frozen pokeweed leaves were juiced in a juicer (Acme model 5001) and clarified by centrifugation. Subsequently, PAP was purified from the supernatant according to the procedure of Irvin, J.D., Arch. Biochem. Biophys., 169. 522 (1975). Purified PAP was concentrated by ultrafiltration over a PM-30 membrane (Amicon), dialyzed against H₂0 for 2 days with one exchange of dialysis fluid, and lyophilized to dryness using a lyophilizer obtained from FTS Systems, Stone Ridge, NY.

Example 2 Laige Scale Production and Purification of B43 (anti-CD19^ Monoclonal Antibody Large scale production of B43 MoAb was performed using a dedicated ACUSYST-Jr (69.8 cm wide X 66 cm deep X 52 cm high) benchtop automated hollow-fiber cell culture system (Endotronics, Coon Rapids, MN). B43 MoAb was purified from the harvested ACUSYST-Jr culture supernatant using the Affi-Prep Protein A MAPS system (obtained from Bio-Rad Laboratories, Richmond, CA) set up in a 49 cαft. chromatography cabinet (Model 450 Puffer Hubbard, New York, NY) equipped with two 15 W germicidal ultraviolet (UV) lamps. Purified antibody was neutralized, concentrated, dialyzed against 40 mM sodium phosphate buffer, pH 7.5, containing 150 mM sodium chloride, and filter sterilized. Antibody concentrations were determined spectrophotometrically using an E^1% ₂₈₀ value of 1.4. All buffers were prepared with endotoxin free water (Travenol Laboratories, Inc., Deerfield, IL) and filter-sterilized just before use. F ramnlβ 3

Pmdiiction and P"ηfif ion of B43-PAP immunotoxin

Highly purified preparations of B43 MoAb and PAP were used as the starting materials for the scaled-up preparation of B43-PAP immunotoxin. All column eluants were tested for sterility and the presence of endotoxin (using the Limulus amebocyte assay) prior to use. All of the following steps in the preparation and purification of B43-PAP immunotoxin were performed under GLP conditions using sterile, endotoxin-free buffers and equipment. L Modification of B43 MoAb and PAP In brief, purified B43 MoAb, at a concentration of 10 mg/ml in 40 mM sodium phosphate, 150 mM sodium chloride, pH 7.5 (PBS) was reacted with a 3:1 molar excess of SPDP (N-succ_rύmidyl 3-(2-pyridyldithio) propionate (Pharmacia LKB, Piscataway, NJ), freshly prepared in DMSO (HybriMax grade, Sigma Chemical Co., St. Louis, MO), at a concentration of 64 mM, and diluted 1/10 in PBS just prior to use. Purified PAP, at a concentration of 10 mg/ml in PBS pH 8, was mixed with a three-fold molar excess of 2-iminothiolane HC1 (Pierce Chemical Co., Rockford, IL), prepared immediately prior to use as a 20 mM solution in 50 mM sodium phosphate, pH 8.6. Both modification reactions were allowed to proceed for 2 hours at room temperature with gentle rocking in sterile, endotoxin-free vials (Miles, West Haven, CT). Excess reagents and low molecular weight reaction products were subsequently removed by gel filtration on Sephadex G-25 PD- 10 prepacked columns (Pharmacia LKB) equilibrated in sterile, endotoxin free PBS, pH 7.5. Individual fractions were monitored at 280 nm and those containing the majority of the protein were combined and the total amounts of antibody and PAP calculated using E^I% ₂₈o_nm values of 1.40 and 0.83 for B43 and PAP respectively.

The extent of arnino group modification of B43 MoAb was determined by measuring the liberation of pyridine-2-thione groups following treatment with dithiothreitol (DTT). The concentration of the liberated groups was calculated by using the extinction coefficient E _343nm = 8.08 X 10³ M^'1. Thiolation of PAP by 2-iminothiolane was dtfermined following treatment with Ellman's reagent (DTNB). The absorbance at 412 nm was measured and the number of sulfhydryl groups mtroduced per mole of PAP was calculated using an extinction coefficient of 13.6 X 10³ M^"1. 2. Conjugation of B43 MoAb and PAP. a_. Utilizing SPDP as a linking agent.

2-irrύnotWolane-derivatized PAP was added to the SPDP-modified B43 MoAb at a final molar ratio of 3.5:1, PAP:MoAB. This rnixture was incubated for 2 hours in sterile, endotoxin free vials at room temperature with gentle rocking and left at 4° C overnight. Gentle rocking was continued for 4-5 hours the following day before the reaction mixture was filtered (0.2 μm Acrodisc, Gelrnan Sciences, Ann Arbor, MI) in preparation for the HPLC step. Utilizing LC-SPDP as a linking agent

Initially, the procedure of part 1 was followed, with the substitution of LC-SPDP for SPDP to mtroduce 2-pyridyl disulfide bonds into B43 MoAb. A 47 mM solution of LC-SPDP in DMSO was freshly prepared and diluted 1:10 in PBS immediately prior to use. Modified PAP was added to the LC- SPDP-modified B43-MoAb at a final molar ration of 3.5:1, PAP:MoAb. This mixture was incubated for 2 hours in sterile, endotoxin free vials at room temperature with gentle rocking and left at 4° C overnight. Gentle rocking was continued for 4-5 hours the following day before the reaction mixture was filtered (0.2 μm Acrodisc, Gelrnan Sciences, Ann Arbor, MI) in preparation for the HPLC step. j Utilizing SMPT as a linking agent For modification with SMPT, the published procedure of Thorpe was used. Cancer Res.. 42, 5924 (1987). Step 1 was replaced with the following procedure. Briefly, 20 mg of B43 were dialyzed overnight against 50 mM sodium borate buffer, pH 9.0, containing 1.7% (w/v) sodium chloride, and subsequently reacted with a 2.4:1 molar ratio of SMPT. Dimemylformamide was added to the MoAb at a final volume of 10% in order to keep the SMPT soluble. Purified PAP (10 mg/ml in PBS, pH 8.0) was modified via its free amino groups with a 3:1 molar excess of 2-iminothiolane HC1 (Pierce Chemical Company) prepared immediately prior to use as a 20 mM solution in 50 mM sodium phosphate buffer, pH 8.6. The modification reaction was carried out in endotoxin-free, glass vials at room temperature for 2 hours with gentle rocking. Excess reagents and low molecular weight reaction products were subsequently removed from the derivaiized PAP and B43 MoAb by gel filtration on Sephadex G-25 PD-10 prepacked columns (Pharmacia LKB) equilibrated in "phosphate-EDTA" buffer containing 10 mM Na₂HP0₄ + 1.8 mM KH₂P0₄ + 170 mM NaCl + 3.4 mM KC1 + 1 mM EDTA, pH 7.5. Individual fractions were monitored at 280 nm and those containing the majority of the protein were combined and the total amounts of PAP and MoAb calculated using E %/_2S0ιm values of 0.83 and 1.4 for PAP and B43, respectively. Thiolation of PAP by 2-iminothiolane was determined following treatment with Ellman's reagent (DTNB). The absorbance at 412 nm was calculated using an extinction coefficient of 13.6 x 10° M^"'. The extent of amino group modification of B43 MoAb was determined by measuring the release of pyridine-2-thione groups following treatment with dithiothrieitol (DTT). The concentration of these liberated groups was calculated using the extinction coefficient E ₃₄₃ „„, = 8.08 x 10³ M^"1.

Modified PAP was added to the SMPT-derivatized B43-MoAb at a final molar ration of 2.5:1, PAP:MoAb. This mixture was incubated for 2 hours in sterile, endotoxin free vials at room temperature with gentle rocking and left at 4° C overnight. Gentle rocking was continued for 72 hours the following day before the reaction mixture was filtered (0.2 μm Acrodisc, Gelrnan Sciences, Ann Arbor, MI) in preparation for the HPLC step. Purification of B43-PAP immunotoxin.

The reaction mixture of Example 3, part 2a was subjected to gel filtration chromatography by HPLC to remove unreacted PAP as well as high molecular weight (> 300 kDa) conjugates/aggregates. A 21.5 X 600 mm Spherogel TSK-3000-SW column (TosoHaas and Beckman Instruments) was used and was equilibrated in 100 mM sodium phosphate buffer, pH 6.8, at a flow rate of 3 ml/min. Ion-exchange chromatography on CM-Sepharose (Pharmacia LKB, Piscataway, NJ) was used to further purify B43-PAP immunotoxin from unconjugated B43 MoAb. 200 mg batches of the semipurified B43-PAP immunotoxin from the HPLC step were concentrated to 10 mg/ml using the Centriprep 30 devices (Amicon, Danvers, MA) and equilibrated by dialysis in 10 mM sodium phosphate buffer, pH 6.2 at 4° C. Spectro/Por 2 tubing was used and the 1500 ml buffer changed twice at 12 hour intervals. The CM- sepharose column (5X12 cm), containing 230 ml of resin was equilibrated in 10 mM sodium phosphate buffer, pH 6.2 and the pH as well as the conductivity of the column effluent were measured. The dialyzed sample (20 ml) was diluted to 100 ml using 10 mM sodium phosphate, pH 6.2, and the pH and conductivity were measured before applying the sample to the column at a flow rate of 1 ml/min. When the sample had completely drained into the resin, the column was washed with the pH 6.2 buffer until the peak of unconjugated antibody came through and the absorbance at 280 nm retumed to baseline.

B43-PAP immunotoxin was subsequently eluted from the CM- Sepharose column using 10 mM sodium phosphate buffer, pH 7.8, containing 20 mM sodium chloride. The ascending portion of the immunotoxin peak was collected in 5 ml fractions as the absorbance at 280 nm began to increase. A small peak or early shoulder occasionally eluted immediately prior to the large immunotoxin peak. This material was contaminated with a small amount of antibody (usually < 5% of the initial amount of B43 MoAb) and was kept separate. The rest of the large peak, containing the 180 kDa and 210 kDa species (i.e., 1 :1 and 2:1 molar ratio of PAP:MoAb) of B43-PAP immunotoxin was collected in two or three fractions and the column washed at pH 7.8, containing 150 mM sodium chloride, was used to elute any remaining immunotoxin. Fractions containing purified 180 kDa and 210 kDa B43-PAP immunotoxin species were combined, brought to 40 mM sodium phosphate, 150 mM sodium chloride, pH 7.5, concentrated to 1.0 mg/ml, filter-sterilized, and frozen at -70° C until use. Protein concentrations were determined for the B43-PAP conjugates using the Bicinchoninic Acid Protein Assay kit obtained from Sigma Chemical Co. (St. Louis, MO). The conjugation of B43 MoAb to PAP was routinely monitored using 5% (non- reduced) SDS-PAGE separating slab gels and the Bio-Rad Mini Protean II apparatus. _. Endotoxin Removal.

The Affi-Prep Polymyxin Support (obtained from Bio-Rad Laboratories, Richmond, CA) was used to remove endotoxin from the purified B43-PAP immunotoxin preparations. Talmadge et al., J. Chromatogr.. 476. 175 (1989). The resin was washed ten times with sterile, endotoxin-free water (Travenol Laboratories, Deerfield, IL), followed by two washes in sterile, endotoxin-free sodium phosphate buffer, pH 7.5, containing 150 mM sodium chloride. 20 ml of B43-PAP, at a concentration of 1.5 mg/ml, were added to 12 ml of washed Affi-Prep Polymyxin resin in a sterile and pyrogen-free 50 ml centrifuge tube. The mixture was gently rotated overnight at 4° C (20-24 hours), then centrifuged to pellet the resin. The ύnmi otoxin-containing supernatant was carefully removed and sterile-filtered into a sterile, endotoxin- free glass vial. 5 ml of additional sterile PBS were added to wash the resin. Following centrifugation, this supernatant was filtered into the same glass vial and a sample removed for the Limulus amebocyte lysate (LAL) assay. Example 4

Quality Control Analysis of PAP toxin. B43-MoAb. and B43-PAP

Immunotoxin Biochemical Analysis of PAP toxin. B43 MoAb. and B43-PAP Immunotoxin. To biochemically confirm the purity of PAP toxin, 1.5-3.0 μg samples of purified PAP protein were analyzed by SDS-PAGE using 15% separating gels and the Bio-Rad Mini Protean II slab gel apparatus (Bio-Rad Laboratories, Richmond, CA) under derrøruring conditions. Similarly, 1.5-3.0 μg samples of purified B43 MoAb were analyzed by SDS-PAGE (Mini Protean II slab gel apparatus of Bio-Rad Laboratories) using a 5% separating gel and 4% stacking gel (non-reduced) or 15% separating gel and 5% stacking gel (reduced). Pre-stained molecular weight standards (Diversified Biotech, Newton Centre, MA) included lactoglobulin (18 kDa), carbonic anhydrase (29 kDa), ovalbumin (43 kDa), glutamate dehydrogenase (55 kDa), bovine serum albumin (66 kDa), phosphorylase B (95.5 kDa subunit) and myosin (205 kDa subunit). Laemmli, U.K., Nature. 222, 680 (1970). Gels were stained with Coomassie Blue G-250, destained in 10% acetic acid/30% methanol, dried, and subsequently scanned using a Beckman DU62 spectrophotometer and Gel Scan Soft-Pac Module software (Beckware Instruments, Fullerton, CA). Gels were stored in H20 and photographed using Ektachrome ASA 100 film.

The purity of PAP was also assessed by ion exchange HPLC using an SP-5PW 7.5 X 7.5 mm polymer based analytical column in the Beckman System Gold HPLC System and System Gold Chromatography Software (Beckman Instruments, San Ramon, CA). A flow rate of 1 rnl/min was used and PAP eluted with a 20 minute, 0-300 mM potassium chloride gradient in 20 mM potassium phosphate buffer, pH 7. 30 μg of highly purified PAP protein, in 100 μl of 10 mM sodium phosphate buffer, pH 7.0, were sequenced according to the automatic degradation procedure originally described by Ed an and Begg (Eur. J. Biochem.. 1, 80 (1967)) and modified by Hunkapiller et al. (Methods Enzymol.. 21, 399 (1983)) using an Applied Biosystems Model 470A gas phase protein sequencer. High performance liquid chromatography using an on-line Model 120A HPLC (Applied Biosystems, Foster City, CA) was used to identify the phenylthiohydantoin amino acids. HPLC chromatograms of the sample, generated for each Edman degradation cycle, were compared to similar HPLC chromatograms obtained for phenylthiohydantoin amino acid standards. Table I summarizes the quality control data on the purified PAP batches (n=<3) used in preparing clinical batches of B43-PAP immunotoxin. TABLE T QUALITY CONTROL ANALYSIS OF PIJRIFTED OraWEEn Ar^TTvTRAI. PROTEIN (PAV

Test Parameters Rsulis

Yield 8 800--8855 m n g purified PAP kg pokeweed leaves

Purity

SDS-PAGE 99.9%

Analytical HPLC 99.9%

Molecular weight 29 kDa

Amino acid composition Confύτriatory

Sterility testing Negative

Ribosome Inhibitory activity IC₅₀ 0.34 ± 0.06 ng/ml

(12.0 ± 2.0 pM)

Acute toxicity in mice LD^i.v.) = 150 ug/mouse

The purity of B43 MoAb was assessed by size exclusion (gel filtration) chromatography using a 7.5 X 300 mm (13 μm) TSK 4000 SW silica based analytical column and Beckman System Gold Chromatography Software Package (Beckman Instrument, San Ramon, CA). The column was equilibrated in 100 mM sodium phosphate buffer, pH 6.8, at a flow rate of 0.5 ml min and the MoAb eluted as a sharp peak with a retention time of 20 min.

For the biochemical evaluation of the purity of B43-PAP immunotoxin,

3.0 μg amounts of CM-sepharose purified B43-PAP protein were boiled in sample buffer containing 40 mM Tris buffer pH 6.8, 2% SDS, 7.5% glycerol and 0.005% bromophenol blue tracking dye and electrophoresed on 5% separating gels or 3-17% gradient gels (Jule, New Haven, CT) according to the method described by Laemmli, cited supra.

Furthermore, the presence of the PAP and B43 moieties in purified B43-PAP immunotoxin was confirmed using a two-step immunoblotting technique and a detection kit obtained from Bio-Rad Laboratories. This kit contained a goat-anti-rabbit IgG-alkaline phosphatase conjugate and was able to detect 0.1 ng of PAP protein which had been electrophoretically transferred to a nitrocellulose membrane, following SDS-PAGE, using a semi-dry Semi- Phor apparatus (Model TE-70 Hoefer Scientific, San Francisco, CA). The anti-PAP antibody was generated in rabbits that had been hyperimmunized with purified PAP. This method was also used to verify the removal of unconjugated PAP from the clinical preparations of B43-PAP immunotoxin. The proteins were transferred from gels to nitrocellulose membranes by electrophoretic transfer. Color transparencies were made from the Western blots using a 4 X 5 Calumet camera and 4 X 5 Daylight Hctachrome color transparency 100 ASA sheet film (Eastman Kodak Company, Rochester, NY). Quantitation of the immunoblots was performed by densitometric analysis using a Beckman DU62 spectrophotometer equipped with a gel scanning attachment and Gel Scan Soft-Pac Module. A plot of the relative density of the PAP band versus the amount of PAP protein present in that lane yielded a curve from which the amount of PAP protein present in samples of B43-PAP immunotoxin could be determined.

Similarly, single step immunoblotting using purified B43 MoAb as a standard was done using alkaline phosphatase conjugated goat anti-mouse IgG (Sigma Chemical Co., St. Louis, MO) to detect unconjugated B43 MoAb remaining in the immunotoxin preparations.

2_ι Immunological Analysis of B43 MoAb and B43-PAP Immunotoxin The immunoreactivity of unlabeled as well as phycoerythrin labeled preparations of purified B43 MoAb was studied by immunofluorescence and multiparameter flow cytometry on a FACStar Plus (Becton Dickinson, Mountain View, CA) as described in Uckun et al., Blood, 21, 13 (1988). The CD19 specificity of purified B43 MoAb and B43-PAP immunotoxin was corifirmed in blocking experiments by examining (1) their ability to block the bmding of radiolabeled or fluorochrome labeled B4 and B43 anti-CD19 MoAbs to CD19+ target cells and (2) to bind to the surface of COS cells transfected with the CD19 cDNA clone (provided by Ivan Stamenkovic, Harvard University, Boston, MA). Equilibrium bmding assays and Scatchard analysis were performed using ¹²⁵I-labeled preparations of purified B43 MoAb and purified B43-PAP immunotoxin to deteπriine and compare their affinity for the target CD19 antigen, as described by Uckun et al., cited supra. Table II summarizes the biochemical, immunologic, and biologic features of purified B43-MoAB as well as the results of the quality control tests.

TABL π

QUALITY CONTROL ANALYSIS OF PURIFIED B43/CD1 MONOCLONAL ANTIBODY

Test Parameters Results

Purity

SDS-PAGE 99.9% Analytical HPLC 99.9%

Molecular weight 150 kDa

Reactivity with CD19 transfectants

(FACS analysis) Positive Immunoreactivity profile (FACS analysis) (+) NALM-6, REH/B lineage ALL (-) MOLT-3 T-ALL, KG-l/AML

Crossblocking (FACS analysis) (+) B4/CD19, Leul2/CD19 (-) 24.1/CD10, IF-5/CD20, G28-7/CD22, J3-109/CD72

Saturating concentration (FACS analysis) 5 ug/ml; (33.3 nM) NALM-6 cells Affinity (kDa) (Scatchard analysis) 50 pM Nalm-6 cells Endotoxin contamination (LAL assay) 3.3 EU/mg

Rabbit pyrogen test Negative

Mouse DNA contamination

(slot blot hybridization) < 2.5 pg/ml

Microbiological tests Sterility testing Negative MAP test Negative

Retrovirus reverse transcriptase assay Negative XC plaque assay Negative S+JL- focus assay Negative Mycoplasma test Negative Acute toxicity in mice LD₅₀(i.v.) = 10,000 ug/mouse Tn Vitro Biological Analysis of PAP Toxin and B43-PAP Immunotoxin.

The ribosome-inhibiting activity of purified PAP was analyzed in a cell-free translation system obtained in kit form from Promega Biotec (Madison, WI) as described by Pelham et al. in Eur. J. Biochem., £2, 247 (1976). In vitro adapted CD19⁺ target (Nalm-6 = pre-B ALL) and CD19- non-target (KG1 = AML and MOLT-3 = T-ALL) leukemic cell lines were used for the biological testing of B43-PAP immunotoxin potency and selectivity. Cell lines were maintained, treated with immunotoxins, and the inhibition of the clonogenic growth of leukemic cells following immunotoxin treatment was evaluated by a quantal serial dilution assay as described by Uckun et al. in J. Immunol.. 124, 2010 (1985). The extent of leukemia cell elimination was expressed as log kill = log (φ control/ φ test ) where φ is the most probably number of clonogenic units (CU) as estimated by the Spearman Karber method.

4_. Sterility and Endotoxin Testing

A manufecturer's working cell bank (MWCB) was established for the B43 MoAb-producing hybridoma cell line, IST-1586. The MWCB, as well as purified B43 MoAB/B43-PAP immunotoxin, were screened for microbial contarninants including mycoplasma, bacteria, and fungi, and sent to

Microbiological Associates, Rockville, MD, for MAP testing for detection of adventitious murine viruses as well as for testing for pyrogenicity, presence of mouse DNA and presence of retroviruses according to the Office of Biologies Research and Review, Center for Drugs and Biologies, FDA "Points to Consider in the Manufacture and Testing of Monoclonal Antibody Products for Human Use". The endotoxin content in clinical batches of B43-PAP immunotoxin was routinely measured by the Limulus amebocyte lysate (LAL assay)using the reagents and protocol from the Associates of Cape Cod, Woods Hole, MA In Vivo Toxicity and Pharmacokinetic Properties of PAP. B43 MoAh and B43-PAP Immunotoxin

All animal studies were performed under GLP conditions and following the U.S. Government Principles for the Utilization and Care of Vertebrate Animals, Used in Testing, Research, and Training and accorxiing to the guidelines of the University of Minnesota Animal Care Committee. Female BALB/c mice (6-8 weeks old, 15 - 17 g) were obtained from NIH and were maintained in the ALAAC accredited facilities of the University of Minnesota Research Animal Resources. In acute toxicity studies, mice were given i.p. or i.v. injections of 0-

250 μg PAP, 0-10,000 μg B43 MoAb; or 0-250 μg purified B43-PAP immunotoxin in 0.2 ml PBS. Deaths were recorded twice daily and LD₅₀ values were determined for 10 day survival. Groups of 5-8 mice were used for each of ten different treatment doses, and the experiments were repeated four times.

In pharrriacokinetic studies, mice were lightly anesthetized with ether and injected i.v. with 50-250 μg B43-PAP immunotoxin in 0.5 ml PBS. Mice were serially bled by retroorbital puncture at 10 minutes, 30 minutes, 1 hour, 2 hours, 4 hours, 8 hours, 12 hours, 24 hours and 48 hours following the administration of immunotoxin. The in vivo stability of B43-PAP immunotoxin was examined by solid-phase ELISA determinations using (1) Falcon Micro Test 111 culture plates coated with affinity purified polyclonal rabbit anti-PAP IgG and (2) goat anti-mouse IgG conjugated to peroxidase, as described above, for the detection of B43 MoAb in ECS supematants. & ¹²⁵I Labeling of B43 MoAb and B43-PAP Immunotoxin and Scatchard Analyses

¹²⁵I-B43 MoAb (3.32 X 10¹⁰ cpm/mg) and ¹²⁵I-B43-PAP (3.17 x 10¹⁰ cpm/mg) were prepared from purified B43 MoAb and B43-PAP immunotoxin by a solid phase iodination technique using Iodo-Beads (Pierce, Rockford, IL), following the specific recommendations of the manufacturer. Ligand binding assays and Scatchard plot analysis of the specific equilibrium binding data were performed to determine the affinity of B43 MoAb and B43-PAP for the target CD19 antigen on NALM-6 pre-B ALL cells, as described in detail by Uckun et al. in Blood, 21, 13 (1988).

Examples 1-4 describe the standardized procedures for producing of the highly purified preparations of B43-PAP .mmunotoxin utilized in the following examples which describe the use of B43-PAP immunotoxin in SCID mice, cynomo-Ogous monkeys and in human clinical trials. The procedures described for production and purification of B43-PAP immunotoxin yield 100 mg of purified immunotoxin/day. Analytical HPLC, SDS-PAGE with gel scanning, and Western blot analyses using anti-PAP or anti-mouse IgG antibodies indicate that the final product is > 95% pure with < 5% free antibody and < 0.5% free PAP. The B43-PAP that is produced is sterile, mycoplasmarfree, free of ecotropic or xenotropic murine type C viruses by the XC plaque and mink S⁺L^" focus assays, free of retroviral reverse transcriptase activity, not pyrogenic when tested in healthy rabbits and its endotoxin contamination was 0.5 EU/mg by LAL assay.

The estimated highest B43-PAP dose to be administered in the projected phase I toxicity study is 0.5 mg/kg. Hence, patients would receive 0.25 EU kg endotoxin at the highest immunotoxin dose, which is 20-fold less than the permissible endotoxin dose of 5 EU/kg determined by the FDA. The biochemical, immunological, or biological properties of B43-PAP immunotoxin, in particular, it's antigen specificity, affinity, chemical composition/purity, did not show significant variations among different batches, confirming the reproducibility of the procedure. Table III summarizes the quality control analysis performed on the purified B43-PAP immunotoxin. TABLE m

OUALTTY CONTROL ANALYSTS OF PURIFIED B43-PAP

TMMUNOTOXIN

Test Parameters Results

Yield 18 ± 2 (% of initial B43 MoAb)

Chemical composition (SDS-PAGE, HPLC, Immunoblotting) 180 kDa immunotoxin species

(1:1 PAP:MoAB ratio) 56%

210 kDa immunotoxin species

(2:1 PAP:MoAB ratio) 41%

Unconjugated B43 MoAb contamination 2.6%

Unconjugated PAP contamination 0.4%

Biochemical stability (SDS-PAGE) > 4 months at 4° C

> 12 months at -70° C

Reactivity with CD19 transfectants

(FACS analysis) Positive

Immunoreactivity profile (FACS analysis) (+) NALM-6, REH, RAJI

(-) MOLT-3, KG-1

Saturating concentration (FACS analysis) 5 ug/ml; (26 nM)

Affinity (kDa) (Scatchard analysis) 65 pM

Ribosome inhibitory activity (Reticulocyte lysate assay)

8 ± 2 ng/ml (41 ± 11 pM) 68 ± 27 ng/ml (341 ± 137 pM)

In vitro anti-leukemic activity (clonogenic assay) I o 100 ng/ml (0.5 nM) TABLE m fcont

QUALITY CONTROL ANALYSIS OF PURIFIED B43-PAP

IMMUNOTOXIN

Test Parameters Results

300 ng/ml (1.5 nM)

IC9 -IC9 9 1000 ng/ml (5.1 nM)

Plasma half life in mice (solid phase ELISA) t^xaa = 3 h; f¹²β = 10 h

Endotoxin contamination (LAL assay) 0.5 EU/mg

Rabbit pyrogen test Negative

Polymyxin contamination (HPLC analysis) < 1 ug/mg

Mouse DNA contamination

(slot blot hybridization) < 2.5 pg/0.5 mg

Microbiological tests

Sterility testing Negative Retrovirus reverse transcriptase assay Negative XC plaque assay Negative

S⁺L^' focus assay Negative

Mycoplasma test Negative

Acute toxicity in mice LD^i.v.) = 20 ug/mouse LD₅₀(i.p.) = 60 ug/mouse

Example 5 Studies in Mouse Models of CP19⁺ Human B-lineagp AT I .

A highly aggressive subclone of the human pre-B acute lymphoblastic leukemia cell line NALM-6 causes disseminated and fatal leukemia in CB.17 mice with severe combined immunodeficiency (SCID) even after intravenous injection of a single cell. An intravenous challenge with lxl 0⁶ NALM-6- UM1 cells caused 15 of 27 (56%) SCID mice to become paraplegic at 31 ± 2 days (median = 33 days) and 27 of 27 (100%) mice to die of disseminated human pre-B ALL at 38 ± 1 days (median = 39 days). Uck n et al., Blood. 22, 2201 (1992). This SCID mouse model of aggressive human pre-B ALL was used to evaluate the in vivo anti-leukemic efficacy of B43-PAP.

A three-day treatment with non-toxic doses of B43-PAP markedly reduced the incidence of paraplegia and improved event-free survival in SCID mice challenged with lxlO⁶ NALM-6-UM1 pre-B ALL cells, as reflected by significantly higher cumulative proportions of mice free of paraplegia or alive for one to seven months, as compared to PBS treated control mice. The Kaplan-Meier estimates and standard errors of the probability of developing paraplegia after inoculation of lxlO⁶ NALM-6-UM1 cells were 64 ± 10 % for PBS treated mice (median time to paraplegia = 37 days) (N=27), 18 ± 8% for mice treated with 15 μg B43-PAP (5μg/mouse/day x 3 days) (N=23) and 5 ± 5 % for mice treated with 30 μg B43-PAP (10 μg/mouse/day x 3 days) (N=21).

While 27 of 27 PBS treated control SCID mice died of leukemia at 38 ± 1 days (range = 24 to 54 days), only 16 of 44 B43-PAP treated mice developed leukemia at 74 ± 12 days (range = 30 to 182 days), consistent with 6 logs kill of clonogenic NALM-6-UM1 cells in 64% of SCID mice. Uckun et al., Blood, 22, 2201 (1992). The Kaplan-Meier estimates and standard errors of the probability of long-term event-free survival after inoculation of lxlO⁶ NALM-6-UM1 cells were 65 ± 10 % for mice treated with 15 μg B43- PAP and 60 ± 11 % for mice treated with 30 μg B43-PAP with a median survival time of >7 months for both groups.

Long-term survivors in the B43-PAP treated groups were electively sacrificed at 7 months to assess their leukemia burden. No histopathologic, flow cytometric, or PCR evidence of pre-B ALL was found. The activity of B43-PAP appears similar or superior to that of other conventional cytotoxic agents at clinically attainable and tolerable AUC's.

Example 6 Studies in Cynomologous Monkeys In non-human primate toxicity studies, a total of 10 cynomologous monkeys were given escalating doses of B43-PAP. No histopathologic lesions wer found in monkeys given 7 intravenous doses of B43-PAP at dose levels of 0.001 mg/kg/day (Total dose = 0.007 mg/kg), 0.01 mg kg/day (Total dose = 0.07 mg/kg), or 0.05 mg/kg (Total dose = 0.35 mg/kg) on alternate days, or in monkeys given 7 intravenous doses of B43-PAP on consecutive days at dose levels of 0.01 mg/kg/day (Total dose = 0.07 mg/kg) or 0.1 mg/kg/day (Total dose = 0.7 mg/kg). In these monkeys, the only clmcal/larxjratory signs of toxicity were a mild (<2-fold) and transient elevation of trarisarriinases and a mild capillary leak syndrome (only at doses 0.35 mg/kg). However, several lesions were found in monkeys given 7 intravenous doses of B43-PAP on 7 consecutive days at dose levels of 0.5 mg/kg/day (Total dose = 3.5 mg/kg) or 1.0 mg/kg/day (Total dose = 7.0 mg/kg). These lesions included acute mild multifocal hepatocellular necrosis and hepatitis at the 1.0 mg/kg/day dose level and severe subacute renal tubular necrosis at both dose levels.

The stability (chemical, biological, and immunological) and immunogenicity (induction of host immune responses to PAP as well as murine IgG moieties) of B43-PAP in cynomologous monkeys was also evaluated. The serum half-life of B43-PAP in cynomologous monkeys ranged from 18.2 hrs to 22.6 hrs. The kinetics as well as the magnitude of the humoral immune response of cynomologous monkeys to the PAP or the murine IgG moieties of B43-PAP were dependent on the immunotoxin dose administered.

Example 7 Phase l/π αinical Studies Twenty-four patients (4 adults and 20 children) with therapy refractory and steroid resistant ALL have received 1-3 cycles of B43-PAP therapy. These patients had relapsed after multiple courses of intensive chemotherapy and/or total body irradiation plus chemotherapy and had failed attempts to control their disease with a combination of multiple standard chemotherapeutic agents. During each 5-day treatment cycle, B43-PAP was administered daily as a 1 hour intravenous infusion. Prior to and following each infusion, the catheter line was cleared with normal saline, or dextrose, 5% in 1/2 normal saline. 50 mg/m² hydrocortisone was added to each B43- PAP bag to minimize the risk of allergic reactions. The infusion was not mterrupted for blood drawing or for adrninistration of other medications. Except for capillary leak and myalgias, no other significant toxicities were observed at dose levels ranging from 0.1 μg/kg/day to 250 μg/kg/day. Significant myelosuppression, nephrotoxicity, hepatotoxicity, or cardiac toxicity was not observed. Five patients achieved complete remission, two patients achieved a partial remission (7% blasts with concomitant recovery of normal hematopoiesis), 5 patients (all with M3 marrow status with large numbers of circulating leukemia cells) had partial responses (marked reduction or eradication of circulating leukemia and shrinkage of lymph nodes/spleen without a significant decrease in the percentage of bone marrow blasts), 9 patients had a stable disease while on B43-PAP therapy and only 3 patients had progressive leukemia despite B43-PAP.

B43-PAP was very stable and therapeutic concentrations of > 0.1 μg/ml of serum could be rnaintained by a single infusion per day for 12-24 hours in all patients receiving B43-PAP at the stage II dose of 100 μg/kg/day. Concentrations as high as 13 μg/ml were well-tolerated by patients.

Only 3 of 24 patients developed antibodies to the murine MoAb moiety as well as the PAP moiety of B43-PAP immunotoxin. This result indicates that at therapeutic doses B43-PAP will effectively kill or inhibit functional B-cells and no significant host immune response will be triggered in the majority of patients.

Example 8 Studies in Mouse Models of CD19⁺ Human B-lineage AIL with B43-PAP + Cytarabine

A highly aggressive subclone of the human pre-B acute lymphoblastic leukemia cell line NALM-6 causes disseminated and fatal leukemia in CB.17 mice with severe combined immunodeficiency (SCID) even after intravenous injection of a single cell. An intravenous challenge with lxlO⁶ NALM-6- UMl cells caused 15 of 27 (56%) SCID mice to become paraplegic at 31 ± 2 days (median = 33 days) and 27 of 27 (100%) mice to die of disseminated human pre-B ALL at 38 ± 1 days (median = 39 days). Uckun et al., Blood, 22, 2201 (1992). This SCID mouse model of aggressive human pre-B ALL was used to evaluate the in vivo anti-leukemic efficacy of B43-PAP + cytarabine.

Female SCID mice (age range 5.4 - 10.1 weeks) were inoculated intravenously with 1 x 106 NALM-6 cells via tail vein injections on day 0 and 24 hours later were subjected to treatment with the regimens depicted in Table IV, below. All doses were given in 0.2 mL PBS solutions unless otherwise specified. The combination of the indicated drug with B43-PAP was given according to the regimen of each drug with B43-PAP given in parallel on 5 consecutive days according to the regimen depicted in Table IV.

TABLE TV TREATMENT PROTOCOLS FOR COMBINATION THERAPTES

Days i l k 21 2£ NALM-6 1 x 10⁶ cells, i.v. Day 0 X

PBS, 0.2 mL i.p. xxxxx

Vincristine 3 mg/m² i.p. X X X

Methylprednisolone XXXXXXXXXX 30 mg/m² i.p. L-Aspariginase 30,000 IU/m² i.p. xxxxx

Doxorubicine 6 mg/m² i.p. XXXXXXXXXX

Etoposide (VP16) X

1500 mg/m² i.p.

Carmustine (BCNU) 150 mg/m2 i.p. x

Cytarabine (ARA-C)

300 mg/m2 i.p. xxxxxxxxx

Mice were observed daily for evidence of leukemia and killed when moribund or unable to obtain food or water. Event times were measured from the day of inoculation of leukemia cells to the day of paraplegia (which results from central nervous system (CNS) leukemia) or death. The probability of event-free survival was determined, and event-free interval curves were generated using the Kaplan-Meier product limit method. The log-rank test was used to assess the effect of various treatment regimens on event-free survival of SCID mice. Mice were necropsied at the time of death or euthanization, histopathology, and polymerase chain reaction (PCR) analyses were performed to assess the burden of human leukemia cells. For each mouse, multiple tissues, including bone marrow, spleen, liver, brain, kidneys, lungs, heart, ovaries, and gut were histologically evaluated. All histopathologic studies were performed by a veterinary pathologist. Human DNA was detected by amplifying a 110-bp fragment from the first exon of the human β-globin gene. The results of this experiment are depicted in Table V, below.

TREATMENT PROTOCOLS FOR COMBINATION THERAPIES

Cumulative Proportion of Mic

Median EFS Surviving Event-Free r% SF

Treatmen # of mice (days) 5ϋd lfjQd 2QQd

PBS 49 40 12*5 O±O O±O

VCR 5 51 40±22 O±O O±O

MP 5 163 lOO±O 60±22 40±22

L-ASP 5 44 O±O O±O O±O

ADR 5 49 40±22 O±O O±O

VP16 10 66 80±13 30±15 O±O

BCNU 5 54 60±22 O±O O±O

ARA-C 5 57 60±22 O±O O±O

B43-PAP 15 164 lOO±O 53±13 46±13

B43-PAP+ 5 NR lOO±O 60±22 60±22

VCR

B43-PAP+ 5 ^" NR lOO±O 60±22 60±22

MP

B43-PAP+L- 5 240 80±13 40±22 40±22

ASP

B43-PAP+ 5 103 lOO±O 60±22 20±18

ADR

B43-PAP+ 5 64 60±22 O±O O±O

VP16

B43-PAP+ 5 103 lOO±O 40±22 40±22

BCNU

B43-PAP+ 5 NR lOO±O lOO±O lOO±O

ARA-C*

* B43-PAP+ARA-C conferred significantly better EFS outcome when compared with B43-PAP alone (P=0.01) or ARA-C alone (P=0.0001). Abbreviations: NR, not reached; EFS, event free survival; VCR, vincristine; MP, methylpredisolone; L-ASP, L-asparaginase; ADR, Doxorubicine; VP16, etoposide; BCNU, Carmustine; ARA-C, Cytarabine.

As is shown in Table V, the 5 SCID mice challenged with lxlO⁶ NALM-6-UM1 pre-B ALL cells that received B43-PAP plus cytarabine according to the schedule shown in Table IV exhibited a markedly reduced incidence of paraplegia and an improved event-free survival, as reflected by significantly higher cumulative proportions of mice free of paraplegia or alive for one to seven months, as compared to PBS treated control mice. See also, Figure 1. While 49 of 49 PBS-treated control SCID mice died of leukemia at 100 days, none of the 5 B43-PAP-rcytarabine treated mice developed leukemia by the end of the study - 200 days.

Example 9 Clinical Data Repaiding Safety and Efficacy of B43-PAP Plus Cytm t C clophosphamide is commercially available as a lyophilized powder.

The drug should be mixed with D₅W to a final concentration of 20 mg/mL but not less than lOOcc/M² for IV use. The drug is administered by IV drip over 60 minutes after the completion of B43-PAP infusion. Adequate diuresis must be maintained before and after cytoxan adπύnistration. The recommended amount is 2 L/M²^ hours. This rate should be started at least 6 hours before cytoxan and continued at least 24 hours after cytoxan.

Eight patients with refractory ALL was treated with 100 μg/kg day x 5 days B43-PAP plus 10 - 20 mg/kg/day x 5 days Cytoxan (cyclophosphamide). On each of the five treatment days, B43-PAP was given i.v. between hours 0 and 1 followed by Cytoxan infusion between hours 1 to 2. Mesna was used for bladder protection at a dose of 2 mg/kg/dose x 5 doses/day i.e., total daily Mesna dose = total daily Cytoxan dose (MESNA is commercially available and should be diluted with D₅W or normal saline). On day 7, patients were started on 5 μg/kg of G-CSF qday. This protocol was tolerated extremely well and all patients but one experienced no capillary leak, no hepatotoxicity, no renal toxicity, and no cardiac toxicity. Firrthermore, the patients had no fever, nausea or any other discomfort. Three patients achieved a complete remission, one patient achieved a partial remission, one patient had a very good peripheral response with disappearance of circulating leukemia cells, 2 patients had stable disease and only one patient has progressive leukemia

Example 10

Use of B43-PAP to Eliminate Residual Leukemia Cells in High Risk

Remission AT I Patients ιιnrieτgoing BMT B43-PAP was used in 3 patients who had residual leukemia and it was able to erradicate this residual leukemia These patients remained free of disease following BMT, demonstrating that B43-PAP can be used safely in the context of BMT, and may iπiprove outcome.

All patents and publications are incorporated by reference herein, as though individually incorporated by reference. While only certain preferred embodiments of this invention have been shown and described by way of illustration, many modifications will occur to those skilled in the art and it is, therefore, desired that it be understood that this is intended herein to cover all such modifications that fall within the spirit and scope of this invention.

Claims

WHAT IS CLAIMED IS:

1. A therapeutic method for the treatment of acute lymphoblastic leukemia (ALL) comprising parenterally administering to a patient who is afflicted with ALL a tlierapeutically effective amount of an immunotoxin consisting of an effective cytotoxic amount of PAP linked to monoclonal antibody B43 (B43-PAP).

2. The method of claim 1 wherein said patient is a relapsed ALL patient.

3. The method of claims 1 or 2 wherein said patient is under the age of 18.

4. The method of claim 1 wherein the B43-PAP is administered in combination with a pharmaceutically acceptable liquid carrier.

5. The method of claim 4 wherein the liquid carrier comprises isotonic saline.

6. The method of claim 4 wherein the composition is administered intravenously.

7. The method of claim 1 wherein 1-2 molecules of PAP are linked to each molecule of B43.

8. The method of claim 4 wherein the concentration of B43-PAP in said liquid carrier is 0.05-0.25 mg/ml.

9. The method of claim 1 wherein the amount of B43-PAP administered is about 0.1 μg/kg/day to 250 μg/kg/day.

10. The method of claim 1 wherein the treatment further comprises the administration of an effective anti-leukemic amount of a second antineoplastic agent.

11. The method of claim 10 wherein the second antineoplastic agent is a class I immunosuppressive drug or an antimetabolite.

12. The method of claim 11 wherein the second antineoplastic agent is a class I immunosuppressive drug.

13. The method of claim 12 wherein the second antineoplastic agent is cyclophosphamide.

14. The method of claim 11 wherein the second antineoplastic agent is an antimetabolite.

15. The method of claim 14 wherein the second antineoplastic agent comprises memotrexate, trimetrexate, 5-fluorouracil, cytarabine, mercaptopurine, t oguanine, 5-azacitidine, floxuridine or 2"- chlorodeoxyadenosine.

16. The method of claim 15 wherein the second antineoplastic agent is cytarabine.

17. The method of claim 10 wherein the second antineoplastic agent is combined with a ptøuτnaceutically acceptable carrier.

18. The method of claim 10 wherein the second antineoplastic agent is administered intravenously.

19. The method of claim 1 wherein the B43-PAP is effective to eliminate residual leukemia in the patient prior to bone marrow transplantation.

20. A treatment kit, comprising, separately packaged in association, an effective anti-leukemic amount of B43-PAP and an effective anti- leukemic amount of a second antineoplastic agent which amounts are adapted for simultaneous or sequential administration.

21. The kit of claim 20 wherein the amounts are adapted for parenteral administration in combination with a pharmaceutically acceptable carrier.

22. The kit of claim 21 wherein the amounts are adapted for intravenous administration.