Biodistribution and antitumour efficacy of long-circulating N-(2-hydroxypropyl)methacrylamide copolymer-doxorubicin conjugates in nude mice

The aim of this study was to evaluate the influence of the molecular weight (mol. wt) of N-(2-hydroxypropyl)methacrylamide (HPMA) copolymer-doxorubicin (DOX) conjugates (P-DOX) on biodistribution and therapeutic efficacy in nu/nu mice bearing human ovarian carcinoma OVCAR-3 xenografts. Copolymerisation of HPMA, a polymerisable derivative of DOX (N-methacryloylglycylphenylalanylleucylglycyl doxorubicin) and a newly designed crosslinking agent, N(2),N(5)-bis(N-methacryloylglycylphenylalanyl-leucylglycyl)ornithine methyl ester monomers resulted in novel, high mol. wt, branched, water-soluble P-DOX containing lysosomally degradable oligopeptide sequences as crosslinks and side-chains terminated in DOX. Four conjugates with mol. wt of 22, 160, 895 and 1230 kDa were prepared. The results indicated that the half-life in blood and the elimination rate from the tumour were up to 28 times longer and 25 times slower, respectively, for P-DOX (mol. wt=1230 kDa) than for free DOX. Treatment with P-DOX (mol. wt > or = 160 kDa) inhibited tumour growth more efficiently than that of 22 kDa P-DOX or free DOX (P<0.02) at a 2.2 mg/kg DOX equivalent dose. In conclusion, the administration of long circulating P-DOX resulted in enhanced tumour accumulation with a concomitant increase in therapeutic efficacy.     

Anticancer agents coupled to N-(2-hydroxypropyl)methacrylamide copolymers. I. Evaluation of daunomycin and puromycin conjugates in vitro

During recent years N-(2-hydroxypropyl)methacrylamide (HPMA) copolymers have been developed as targetable drug carriers. These soluble synthetic polymers are internalized by cells by pinocytosis and they can be tailor-made to include peptidyl side-chains degradable intracellularly by specific lysosomal enzymes. Thus they provide the opportunity fo achieve controlled intracellular delivery of anticancer agents. The anthracycline antibiotic daunomycin, and protein synthesis inhibitor puromycin, were bound to HPMA copolymers via several different peptide side-chains, including Gly-Gly, Gly-Phe-Leu-Gly and Gly-Phe-Phe-Leu. Incubation of polymer-drug conjugates with isolated lysosomal enzymes (either a mixture of rat liver lysosomal enzymes or purified thiol-dependent lysosomal proteinases, cathepsins L and B) showed that significant release of drug occurred over 20 h, more than 20% of daunomycin and more than 80% of puromycin being liberated. To test their pharmacological activity conjugates were incubated with either the mouse leukaemia L1210, or the human lymphoblastoid leukaemia CCRF in vitro. The conjugates tested were all less effective than free daunomycin, but they showed differential toxicity against L1210 depending on the aminoacid sequence of their drug-polymer linkage. Inclusion of fucosylamine-terminating side-chains into the HPMA copolymer structure increased the affinity of conjugates for the L1210 cell membrane and resulted in increased toxicity. In contrast HPMA-daunomycin conjugates with or without fucosylamine affected CCRF cells equally, but this cell line was more sensitive than the mouse leukaemia to both free and polymer-bound daunomycin. Incubation of L1210 cells in polymer-bound daunomycin for 72 h, followed by plating cells out in low density in drug-free medium, showed that a concentration of polymer-bound drug (184 micrograms ml-1) could be selected to achieve a cytotoxic effect.     

Anticancer agents coupled to N-(2-hydroxypropyl)methacrylamide copolymers. II. Evaluation of daunomycin conjugates in vivo against L1210 leukaemia

DBA2 mice were inoculated i.p. with 10(5)L1210 cells. Animals subsequently treated with daunomycin (single i.p. dose, 0.25-5.0 mg kg-1) all died. The maximum increase in mean survival time observed was approximately 135%. Animals treated with N-(2-hydroxypropyl)methacrylamide (HPMA) copolymers conjugated to daunomycin (DNM) showed a significant increase in mean survival time when the polymer-drug linkage was biodegradable (i.e., Gly-Phe-Leu-Gly). Such treatment also produced a number of long term survivors (greater than 50 days). In contrast, HPMA copolymer conjugated to DNM via a non-degradable linkage (Gly-Gly) produced no increase in survival time relative to untreated control animals. The effect observed with biodegradable HPMA copolymer-DNM conjugates was dependent on the concentration of conjugated drug administered (optimum greater than 5 mg kg-1); the frequency of administration (multiple doses were more effective than single); the timing of administration (single doses given on days 1 and 3 were most effective); and the site of tumour inoculation and route of drug administration. Biodegradable HPMA copolymer-DNM conjugates administered i.p. were active against L1210 inoculated s.c. at higher doses than required to curb a peritoneal tumour. Under certain experimental conditions polymer-DNM conjugates containing fucosylamine or galactosamine proved more active than conjugates without the carbohydrate moeity. The mechanism of drug-conjugate action in vivo is at present unclear. Radioiodination of polymer showed approximately 75% of polymer-drug conjugate to be excreted 24 h after i.p. administration. Synthesis of HPMA conjugates containing [3H]DNM showed that polymer containing Gly-Gly-[3H]DNM was excreted (60% of radioactivity in the urine, 24 h) in macromolecular form. In contrast polymer containing Gly-Phe-Leu-Gly-[3H]DNM was largely excreted in the form of low molecular weight species.     

Antitumor activity of IL-2/anti-IL-2 mAb immunocomplexes exerts synergism with that of N-(2-hydroxypropyl)methacrylamide copolymer-bound doxorubicin conjugate due to its low immunosuppressive activity

Interleukin (IL)-2 has been approved for treatment of metastatic renal cancer and malignant melanoma. However, its unfavorable pharmacologic properties, severe side effects and the negative role of IL-2 in maintaining T regulatory cells are severe drawbacks. It has been shown that immunocomplexes of IL-2 and certain anti-IL-2 mAbs possess selective and high stimulatory activity in vivo. Here, we show that IL-2/S4B6 mAb immunocomplexes expand not only CD122(high) subsets and newly activated CD8(+) T cells but also natural killer T cells and γδ T cells. Further, we demonstrate that natural killer (NK) cells expanded by IL-2/S4B6 mAb immunocomplexes in vivo have high cytolytic activity, which can be further increased by coadministration of IL-12. We also demonstrate that IL-2/S4B6 mAb immunocomplexes possess noticeable antitumor activity in two syngeneic mouse tumor models, namely BCL1 leukemia and B16F10 melanoma, but only if administered early in tumor progression. To effectively treat established tumors, we administered the tumor-bearing mice first with N-(2-hydroxypropyl)methacrylamide copolymer-bound doxorubicin conjugate, and subsequently with IL-2/S4B6 mAb immunocomplexes alone or with IL-12 to induce an efficient antitumor immune response. Importantly, we show that the conjugate has significantly lower immunosuppressive activity than free doxorubicin when using dosage with comparable antitumor activity, thus eliminating the majority of tumor cells while leaving the immune system mostly unimpaired for stimulation with IL-2/S4B6 mAb immunocomplexes. Indeed, we demonstrate that the conjugate and IL-2/S4B6 mAb immunocomplexes together have synergistic antitumor activity.     

N-(2-hydroxypropyl)methacrylamide copolymers targeted to the hepatocyte galactose-receptor: pharmacokinetics in DBA2 mice.

N-(2-Hydroxypropyl)methacrylamide (HPMA) copolymers containing doxorubicin (DOX) and galactosamine can be targeted to the hepatocyte galactose receptor for organ-specific chemotherapy of primary and metastatic liver cancer. Here we report the dose-dependent pharmacokinetics of this macromolecular conjugate. Following intravenous administration to mice most efficient liver targeting was seen at low dose (0.05 mg DOX kg-1), with receptor saturation observed using higher bolus doses. Repeated low dose bolus injections did not cause down-regulation of the galactose receptor and targeted drug delivery rates of greater than or equal to 2 micrograms DOX g-1 liver h-1 were achieved. DOX is released from such conjugates intracellularly via action of lysosomal proteinases. It was shown that isolated rat liver lysosomal enzymes (Tritosomes) can release unmodified DOX from the peptidyl side chain Gly-Phe-Leu-Gly at a rate greater than or equal to 3 micrograms DOX g-1 liver h-1 i.e. the hydrolytic capacity is greater than the observed rate of drug delivery to the liver lysosomes in vivo. Although most conjugate would be captured by normal hepatocytes following intravenous administration, it was shown that the human hepatoma cell line HepG2 retains the galactose receptor, accumulating and processing the conjugate efficiently. Potential dose limiting toxicities of such drug conjugates could include cardio- or hepatotoxicity. Administration of conjugate reduced the 15 min heart level of DOX approximately 100-fold compared with that observed for an equivalent dose of free drug. Preliminary experiments showed that plasma levels of alkaline phosphatase, alanine transaminase and asparate transaminase did not change following administration of HPMA copolymer-daunorubicin (DNR) (10 mg DNR kg-1) indicating no significant heptatoxicity.     

Reduced cardiotoxicity of doxorubicin given in the form ofN-(2-hydroxypropyl) methacrylamide conjugates: an experimental study in the rat

Summary A rat model was used to evaluate the general acute toxicity and the late cardiotoxicity of 4 mg/kg doxorubicin (DOX) given either as free drug or in the form of threeN-(2-hydroxypropyl)methacrylamide (HPMA) copolymer conjugates. In these HPMA copolymers, DOX was covalently bound via peptide linkages that were either non-biodegradable (Gly-Gly) or degradable by lysosomal proteinases (Gly-Phe-Leu-Gly). In addition, one biodegradable conjugate containing galactosamine was used; this residue was targeted to the liver. Over the first 3 weeks after the i.v. administration of free and polymer-bound DOX, all animals showed a transient reduction in body weight. However, the maximal reduction in body weight seen in animals that received polymer-bound DOX (4 mg/kg) was significantly lower than that observed in those that received free DOX (4 mg/kg) or a mixture of the unmodified parent HPMA copolymer and free DOX (4 mg/kg;P<0.01). Throughout the study (20 weeks), deaths related to cardiotoxicity were observed only in animals that received either free DOX or the mixture of HPMA copolymer and free DOX; in these cases, histological investigations revealed marked changes in the heart that were consistent with DOX-induced cardiotoxicity. Sequential measurements of cardiac output in surviving animals that received either free DOX or the mixture of HPMA copolymer and free DOX showed a reduction of 30% in function beginning at the 4th week after drug administration. The heart rate in these animals was 12% lower than that measured in age-matched control rats (P<0.05). Animals that were given the HPMA copolymer conjugates containing DOX exhibited no significant change in cardiac output throughout the study (P<0.05). In addition, no significant histological change was observed in the hearts of animals that received DOX in the form of HPMA copolymer conjugates and were killed at the end of the study. However, these animals had shown a significant increase in heart rate beginning at 8 weeks after drug administration (P<0.01). This study demonstrates that covalent binding of DOX to HPMA copolymer conjugates via both stable and biodegradable peptidyl linkages considerably reduces both the general acute toxicity and the late cardiotoxicity of DOX in the rat and could offer the potential for improving the therapeutic index in the clinical application of DOX.     

Free and N-(2-hydroxypropyl)methacrylamide copolymer-bound geldanamycin derivative induce different stress responses in A2780 human ovarian carcinoma cells

The effects of geldanamycin (GA), 17-(3-aminopropylamino)-17-demethoxygeldanamycin (AP-GA), and N-(2-hydroxypropyl)methacrylamide copolymer-AP-GA conjugate [P(AP-GA)] on A2780 human ovarian carcinoma cells at an equitoxic dose (2x IC(50)) were compared by the gene expression array analysis. All treatments resulted in similar gene expression profiles up to 12 h (e.g., down-regulation of CDK4 and up-regulation of APAF-1), although P(AP-GA)-treated cells showed delayed gene expression because of time-dependent internalization of the conjugate and intracellular drug release from P(AP-GA). However, AP-GA-treated cells showed elevated expression of HSP70 and HSP27 after 6 h compared with that observed by GA and P(AP-GA) treatments. Depletion of C-Raf, an HSP90 client protein, was observed in all treatments up to 12 h. Confocal microscopy using mesochlorin e(6) as a model drug revealed that drug release caused by the lysosomal cleavage of glycylphenylalanylleucylglycine oligopeptide spacer, used as GA derivative copolymer attachment/release point, was moderately fast. These results suggested that AP-GA treatment may activate stress-response pathways, whereas P(AP-GA) treatment may suppress them and trigger signaling pathways essential to cell growth arrest and death by inducing an HSP90-active factor. Although GA and P(AP-GA) treatments induced a time-dependent increase in HSP70 and HSP27 protein expression (detected by Western blotting analysis), AP-GA treatment resulted in more rapid and more intense expression of both proteins. Our results suggest that conjugation of AP-GA to N-(2-hydroxypropyl)methacrylamide copolymer may be able to modulate the cell stress responses induced by AP-GA because of differences in its internalization mechanism, subcellular localization, and intracellular concentration gradients.     

A Possibility to overcome P-glycoprotein (PGP)-mediated multidrug resistance by antibody-targeted drugs conjugated to N-(2-hydroxypropyl)methacrylamide (HPMA) copolymer carrier

N-(2-hydroxypropyl)methacrylamide (HPMA) copolymers containing doxorubicin (DOX) and different targeting moieties were developed with the aim of specific chemotherapy. Two of them, HPMA-conjugated DOX and galactosamine-targeted DOX, are in phase II clinical trials in the U.K. We studied the effect of conjugates with different targeting moieties (anti-CD71, antithymocyte globulin, anti-CD4, transferrin) on human or mouse multidrug resistance (MDR) cell lines (CEM/VLB, P388-MDR). It was shown that targeting decreases the level of MDR for DOX and the level of MDR depends on the targeting moiety used. The combination of these conjugates with chemosensitisers (cyclosporin A, D, G) restored almost completely the sensitivity of MDR cell lines to that of parental sublines. These results suggest that different intracellular trafficking of these conjugates (in membrane-limited organelles) in contrast to free diffusion for low molecular weight compounds might partially overcome P-glycoprotein (Pgp)-mediated MDR. We also report here the development of biodegradable HPMA hydrogels suitable for prolonged release of the cytostatic drug and chemosensitiser as a potential approach to overcome MDR mediated by Pgp.     

Antitumor activity of adriamycin (hydrazone-linked) immunoconjugates compared with free adriamycin and specificity of tumor cell killing

Adriamycin (ADM) was chemically coupled to two monoclonal antibodies (MAb) expressed on human B-cell lymphomas. Immunoconjugates were prepared by linking to the MAb an ADM derivative, Adriamycin 13-[3-(2-pyridyldithio)propionyl]hydrazone (ADM-HZN), which releases ADM under mild acidic conditions (see preceding article). The (ADM-HZN) conjugates were tested for antitumor activity on two human B-lymphoma xenografts, Daudi and Ramos, which were growing as solid tumors in athymic mice. The conjugates, injected i.p., significantly inhibited tumor growth when antibody protein doses were greater than or equal to 500 mg/kg (approximately 10 mg/mouse). At these input antibody doses, (ADM-HZN) conjugates were more potent and had greater antitumor activity than free ADM given at an optimized dose and schedule. MAb-conjugated ADM was also tolerated to much higher levels than unconjugated drug. Antitumor activity was not obtained using mixtures of MAb plus free drug or with MAb-drug conjugates that did not bind to the tumor target cell. Thus, the antitumor activity of the immunoconjugate was directed by binding of the MAb portion of the conjugate to target tumor cells.     

Carcinogenic activity of endogenously synthesized N-nitrosobis(2-hydroxypropyl)amine in rats administered bis(2-hydroxypropyl)amine and sodium nitrite

The carcinogenic activity of endogenously synthesized N-nitrosobis(2-hydroxypropyl)amine(BHP) was investigated in male Wistar rats administered bis(2-hydroxypropyl)amine(BHPA) mixed in powder diet at a concentration of 1%, and sodiumnitrite (SN) dissolved in distilled water at concentrationsof 0.15 and 0.3%, for 94 weeks. Urinary excretion of BHP wasdetected in rats given 1% BHPA and 0.3% SN but not in the groupsreceiving either of these precursors alone. Nasal cavity, lung,esophagus, liver and urinary bladder tumors were found in animalstreated with combinations of 1% BHPA and 0.15 or 0.3% SN, suggestingthat the target organs of the endogenously synthesized BHP aresimilar to those affected when the carcinogen is administeredexogenously. The incidences of nasal cavity and lung tumorsreached 74 and 58% in rats given 1% BHPA and 0.3% SN, respectively.Tumors at sites other than target organs were only found atlevels similar to those previously reported for spontaneoustumors in male Wistars. The present results clearly indicatedthe tumor inducibility of a nhrosatable amine, BHA, throughan endogenous nitrosation by feeding to rats in conjunctionwith nitrite, and provide further suggestive evidence that endogenousnitrosations of environmental nitrosatable amines can be a potentialrisk factor in human cancer development.     

Antitumor activity of N-(phosphonacetyl)-L-aspartic acid, a transition-state inhibitor of aspartate transcarbamylase.

N-(Phosphonacetyl)-L-aspartate (PALA) is an analog of the transition state for the aspartate transcarbamylase reaction and has been reported previously to be a potent and specific inhibitor of de novo pyrimidine nucleotide biosynthesis. It is now shown that PALA has considerable antitumor activity against certain transplantable tumors in mice. PALA, unlike other antimetabolites, was less effective against ascitic leukemias than against two solid tumors, B16 melanoma and Lewis lung carcinoma. Another solid tumor, Ridgway osteogenic sarcoma, which is sensitivie to many established chemotherapeutic agents, did not respond to PALA. Daily or intermittent treatment with PALA did not significantly increase the life-span of mice bearing i.p. leukemia L1210. The survival time of mice bearing i.p. P388 leukemia was prolonged by PALA treatment by up to 64%. In a number of experiments mice bearing i.p. B16 melanoma survived 77 to 86% longer than did controls when treated with PALA (490 mg/kg) on Days 1, 5, and 9. Lewis lung carcinoma, a tumor refractory to most established antineoplastic agents, was highly sensitive to PALA. Treatment on Days 1, 5, and 9 following s.c. implantation of Lewis lung carcinoma was curative to 50% of the mice. If treatment was delayed until s.c. Lewis lung tumors had reached about 500 mg, PALA neither cured the mice nor produced significant tumor regression. However, extensive delay of tumor growth and prolongation of survival were still observed.     

Antitumor efficacy of interleukin-2 alone and in combination with adriamycin and dacarbazine in murine solid tumor systems

Recombinant interleukin-2 (IL-2)/chemotherapy combinations have recently entered clinical trial. The rationale for sequencing has primarily been empiric or based on in vitro data. To establish in vivo models for chemoimmunotherapy trials, we investigated IL-2 alone and in combination with dacarbazine (DTIC) and adriamycin. IL-2 (as a single agent given i.v. at 1-3 x 10(5) Cetus units once daily for 5 days, repeated 7-10 days later), was highly active against an immunogenic line of colon adenocarcinoma no. 11/A [tumor growth inhibition (T/C) = 0% with cures]. It was modestly active against colon adenocarcinoma no. 38 (T/C = 39%), mammary adenocarcinoma no. 16/C (T/C = 18%), and B16 melanoma (T/C = 21%). IL-2 was inactive against colon adenocarcinoma no. 7/A (T/C = 83%). Combination trials were done using DTIC and IL-2 against colon no. 7/A and upstaged colon no. 11/A. The combination of adriamycin and IL-2 was tested against mammary adenocarcinoma no. 16/C. In the DTIC/IL-2 combination trials, the combination was superior over either agent used alone. In the IL-2/adriamycin trials, the combination was no better than adriamycin alone at optimum dosages.     

Carcinogenic activity of endogenously synthesized N-nitrosobis(2-hydroxypropyl)amine in rats

The carcinogenic activity of endogenously synthesized N-nitroso-bis(2-hydroxy-propyl)amine (NDHPA) was investigated in male Wistar rats administered bis(2-hydroxypropyl)amine (DHPA), mixed into a powdered diet at a concentration of 1%, and NaNO2 dissolved in distilled water at concentrations of 0.15% and 0.3%, for 94 weeks. Urinary excretion of NDHPA clearly demonstrated its endogenous synthesis in rats given 1% DHPA and 0.3% NaNO2, but not in the groups receiving either of these precursors alone. Tumours of the nasal cavity, lung, oesophagus, liver and urinary bladder were found in rats treated with 1% DHPA and 0.15% or 0.3% NaNO2. The incidences of nasal cavity and lung tumours reached 74% and 58% respectively, in rats given 1% DHPA and 0.3% NaNO2. The tumour distribution was almost the same as that seen in rats given NDHPA. These results indicate that endogenously synthesized NDHPA has similar carcinogenic activity to exogenously administered NDHPA in rats.     

Antitumor activity of alpha fetoprotein and epidermal growth factor conjugates in vitro and in vivo.

Conjugates of carminomycin (Cm) with alpha-fetoprotein (AFP) and epidermal growth factor (EGF) were prepared and their cytotoxic activities were studied in vitro. Both conjugates showed cytotoxic activity which exceeded that of free Cm in tumor cell cultures of MCF-7, SKOV3, QOS, P388 and B16 cells. The antitumor effects of the conjugates were studied in vivo in mice with subcutaneous tumors of B16 and P388 cells. The Cm-AFP and Cm-EGF conjugates inhibited tumor growth and noticeably increased the mean life span in experimental animals. Our results suggest that the therapeutic activity of Cm can be significantly enhanced by conjugation to AFP or EGF.     

Carcinogenicity of N-nitrosobis(2-hydroxypropyl)amine and N-nitrosobis(2-oxopropyl)amine in MRC rats.

Weekly sc injections of equitoxic doses of N-nitrosobis(2-hydroxypropyl)amine (BHP) and N-nitrosobis(2-oxopropyl)amine (BOP) to Wister-derived MRC rats induced tumors. The incidence, latency, multiplicity, morphologic type, and distribution of these tumors varied according to the compound given. The esophagus was the main target organ for BHP (100%), followed by the respiratory tract (87%), pharynx (80%), colon and liver (each 73%), kidneys (20%), thyroid gland (20%), and urinary bladder and urethra (each 7%). BOP was ineffective in the esophagus and pharynx but induced a higher incidence of tumors in the kidneys (27%), thyroid gland (60%), urinary bladder (33%), and urethra (73%) and fewer neoplasms in the respiratory tract (20%), colon (67%), and liver (53%). In addition, BOP caused a few, apparently primary, prostate squamous cell carcinomas. The results are compared with results of BHP treatment in Sprague-Dawley rats and with results of BHP and BOP treatment in Syrian golden hamsters.     

Initiation activity of endogenously synthesized N-nitrosobis(2-hydroxypropyl)amine in the rat liver

The initiation potential of N-nitrosobis(2-hydroxypropyl)amine (NDHPA) endogenously synthesized from bis(2-hydroxypropyl)amine (DHPA) or tris(2-hydroxypropyl)amine (THPA) in the presence of sodium nitrite (NaNO2) was investigated in the rat liver by quantitation of hepatocellular foci showing phenotypic expression of glutathione S-transferase placental form (GST-P). The investigation consisted of two experiments. In the first, male Wistar rats were divided into six groups as follows: group 1 was non-treated; groups 2 and 3 received 0.15% and 0.3% NaNO2, respectively; group 4 received 1% DHPA; groups 5 and 6 received 1% DHPA together with 0.15% and 0.3% NaNO2, respectively. In experiment 2, the same protocol was used except that 2% THPA was substituted for 1% DHPA. The treatments were continuous until sacrifice at week 94 in experiment 1 and week 104 in experiment 2. As a result putative preneoplastic GST-P-positive foci observed in the liver and increased dose-dependently in rats from groups 5 and 6 which had received DHPA and NaNO2, but not in rats administered THPA and NaNO2. The results indicate that endogenously synthesized NDHPA from DHPA and NaNO2 is capable of initiating neoplastic development in the rat liver.     

Etoposide and adriamycin containing combination chemotherapy (HOPE-Bleo) for relapsed Hodgkin's disease.

Forty-four patients with relapsed or resistant Hodgkin's disease were treated with adriamycin 40 mg m-2 i.v. on day 1, vincristine 1.4 mg m-2 i.v. on days 1 and 8, prednisolone 40 mg m-2 orally daily for 8 days, etoposide 200 mg m-2 orally daily for 4 days according to the nadir white cell count, and bleomycin 10 mg m-2 i.v. days 1 and 8 (HOPE-Bleo). Median age was 27 (range 12-71). When stage was considered according to all sites currently or previously involved by Hodgkin's disease (cumulative stage) 26 patients (59%) had stage IV, 13 (29%) stage III and five (11%) stage II disease; 33 (75%) had B symptoms. All patients had received previous chemotherapy and 18 (41%) had received two or more regimens. Twenty-six patients (59%) achieved CR and 10 (23%) PR; the median duration of CR was 22 months and median survival for all patients was 48 months. Eight patients remain in continuous CR; none of these had received extensive previous chemotherapy. Among the 19 patients who had relapsed from CR achieved by a single previous chemotherapy regimen, six (32%) achieved long CR on HOPE-Bleo. The regimen was generally well tolerated but the principal toxicity was myelosuppression. There were two toxic deaths, one due to neutropenic sepsis and the other due to acute peritonitis. The HOPE-Bleo regimen is an effective treatment for relapsed or resistant Hodgkin's disease, with a low probability of carcinogenesis and infertility. These factors suggest that HOPE-Bleo deserves further evaluation as primary treatment for Hodgkin's disease and very careful selection of relapsed patients for high dose salvage chemotherapy with bone marrow transplants must be exercised.