LMB-2. A total of 35 patients were treated. All toxicity was reversible. The maximum tolerated dose was designated as 40 ug/Kg QOD x3 where the most common toxicities were transaminase elevations and fever. Six of the 35 patients made neutralizing antibodies at high enough levels to prevent additional cycles. Eight major responses were observed, including 1 complete remission (CR) and 3 partial responses (PR) in the 4 hairy cell leukemia (HCL) patients treated. PRs were also seen in patients with chronic lymphocytic leukemia (CLL), cutaneous T-cell lymphoma (CTCL), adult T-cell leukemia (ATL) and Hodgkins disease (HD). These constituted the first major responses in cancer to an agent containing an Fv fragment. A third clinical lot of LMB-2 began production by the NCIs Monoclonal Antibody and Recombinant Protein Facility and its completion is expected by the end of this year. Clinical and preclinical studies indicated that the toxicity of LMB-2 was associated with cytokine elevations. Therefore, once the new clinical lot of LMB-2 is ready, we plan to treat additional patients on the LMB-2 protocol combining agents to suppress cytokine production to determine if the maximum tolerated dose can be increased. Once it is determined how best to prevent the toxicity of LMB-2, we plan to initiate phase II testing at NIH and at outside institutions. BL22. The phase I trial of BL22 began 2/99 and so far a total of 19 patients have been treated at doses of 3, 6, 10, 20 and 30 ug/Kg i.v. QOD x3. Toxicity has consisted of a cytokine release syndrome, composed of several signs including fever, hypotension, third spacing of fluid, bone pain, and nausea, vomiting and diarrhea. Unlike toxicity from other toxins, we have been able to prevent toxicity using a variety of agents to prevent cytokine release, including NSAIDs and steroids. We are currently determining the most effective and safe method to prevent cytokine release. Even at low doses, BL22 has shown significant antitumor activity, inducing a CR in one patient with HCL and PRs in two additional HCL patients. One patient with CLL and 3 additional patients with HCL have shown >95 to >99% reductions of malignant circulating cells, with so far < 50% reduction of solid masses or splenomegaly. All of the patients treated had failed standard therapy. Only two patients have been unable to receive additional cycles of BL22 because of neutralizing antibodies, and one of these had significant neutralizing antibodies prior to receiving BL22. Isolation of a new recombinant immunotoxin targeting CD30. Immunization of several different strains of mice with a plasmid encoding full-length human CD30 rapidly resulted in high-titers of CD30-specific antibodies. Analysis of phage libraries through panning on soluble antigen or cells showed that one immunodominant sequence was produced. A recombinant immunotoxin made with this sequence was specifically cytotoxic toward human cells transfected with CD30, with an IC50 of 2 ng/ml. However, both variable domains were determined to be unmutated murine genotype sequences. Since these sequences do not appear to undergo somatic mutation for affinity maturation in immunized mice, our goal is to carry out selective mutagenesis of the variable domains and screen phage libraries of mutants for Fv molecules with high binding affinity. Isolation and characterization of PE35 produced in the plasma after immunotoxin injection. The cleavage of PE between Arg279 and Gly280 is necessary to produce the carboxyl terminal fragment of the toxin which translocates to the cytosol and causes cell death. This cleavage, originally expected to occur within target cells, was found to occur within the plasma after injection of immunotoxin in patients treated with LMB-2 or in non-tumor-bearing mice. Experiments indicate that after cleavage the disulfide bond linking residues 265 and 287 prevents the binding domain from dissociating from the toxin and permits the immunotoxin to bind selectively to target cells. The carboxyl terminal fragment (PE35) was isloated from the plasma of mice receiving high doses of LMB-2, and amino terminal sequencing of the isolated PE35 fragment was consistent with the cleavage event occurring at the Arg279-Gly280 position. Results from this project indicate that cells lacking the ability to cleave PE may appear resistant to recombinant toxins ex vivo, but may not be resistant in vivo since the cleavage event can occur outside the malignant cells.Investigation of the hepatotoxicity of recombinant toxins. Our previous work showed that LMB-2 is 3 to 4-fold more toxic than BL22 to mice and is also more toxic to monkeys because of liver toxicity. This indicates that hepatic toxicity is not mediated through CD25, since murine CD25 does not bind LMB-2. However, it has also been found recently that the hepatic toxicity of LMB-2 is mediated through cytokines, notably TNFa. To determine whether recombinant toxins can kill hepatocytes directly, fresh single-cell suspensions of murine or rat hepatocytes were prepared by Greg Heestand and incubated with recombinant toxins. We found that in isolated rodent hepatocytes, LMB-2 bound non-specifically and had several-fold higher uptake and cytotoxicity compared to BL22. Studies performed using isolated hepatocytes from Cynomolgus monkeys and humans confirmed that LMB-2 is several-fold more cytotoxic than BL22. The IC50s of LMB-2 are 200-1000 ng/ml, which are comparable to peak plasma levels of 298-1040 ng/ml achieved in patients treated with LMB-2 at the MTD. This study suggests that a component of the hepatic toxicity of LMB-2 in humans is due to direct but non-specific interaction with hepatocytes. We have tested a variety of other recombinant toxins in this ex vivo model, including other recombinant immunotoxins and fusion proteins containing growth factor ligands varying from 8 to 20 kDa in length. So far we have found that cytotoxicity toward hepatocytes is generally higher with ligands having a high isoelectric point (pI). Moreover, cytotoxicity appears to be higher with recombinant toxins of lower molecular weight. These studies support efforts to develop recombinant toxins of lower pI, like M1(dsFv)-PE38 in collaboration with Drs. Pastan and Onda, in addition to preventing the cytokine release which facilitates hepatotoxicity.Work on the circularly permuted recombinant IL4-toxin IL4(38-37)-PE38KDEL made in our lab is continuing in collaboration with Raj Puri at the FDA and Steve Marcus at Neurocrine. In phase I/II clinical trials, over 33 patients with recurrent high-grade gliomas in the US and Europe have received IL4(38-37)-PE38KDEL by intratumoral infusion. At this time several patients have responded, including a long-term complete remission in a patient with multiply relapsed glioblastoma multiforme. Major obstacles being addressed currently include determining which patients have tumors with high IL4R expression, and also differentiating toxin-induced necrosis of normal brain from adverse events related to catheter placement or to edema from the necrosis of malignant tissue.Targeting acute myelogenous leukemia with recombinant GM-CSF toxin. The recombinant toxin DT388-GM-CSF made in our lab is currently undergoing phase I testing at Wake Forest University in collaboration with Art Frankel. At this time 16 patients have been treated at dose levels of 1, 2, 3 and 4 ug/Kg i.v. QD x5. Dose-limiting cytokine-release syndrome was observed at the 2 ug/Kg x5 dose level, with confirmed elevations of circulating IL6 and IL-1-receptor antagonist. This syndrome is preventable with high dose solumedrol even at twice this dose level. So far, many of the patients have had considerable reductions of malignant circulating cells, and the maximum tolerated dose with steroids has not yet been determined.
| Kreitman, Robert J (2009) Recombinant immunotoxins for the treatment of chemoresistant hematologic malignancies. Curr Pharm Des 15:2652-64 |
| Kreitman, Robert J; Stetler-Stevenson, Maryalice; Margulies, Inger et al. (2009) Phase II trial of recombinant immunotoxin RFB4(dsFv)-PE38 (BL22) in patients with hairy cell leukemia. J Clin Oncol 27:2983-90 |
| Ahmad, Ejaz; Steinberg, Seth M; Goldin, Lynn et al. (2008) Immunophenotypic features distinguishing familial chronic lymphocytic leukemia from sporadic chronic lymphocytic leukemia. Cytometry B Clin Cytom 74:221-6 |
| Pastan, Ira; Hassan, Raffit; FitzGerald, David J et al. (2007) Immunotoxin treatment of cancer. Annu Rev Med 58:221-37 |
| Arons, Evgeny; Suntum, Tara; Sunshine, Joel et al. (2007) Immunoglobulin light chain repertoire in hairy cell leukemia. Leuk Res 31:1231-6 |
| Arons, Evgeny; Sorbara, Lynn; Raffeld, Mark et al. (2006) Characterization of T-cell repertoire in hairy cell leukemia patients before and after recombinant immunotoxin BL22 therapy. Cancer Immunol Immunother 55:1100-10 |
| Arons, Evgeny; Margulies, Inger; Sorbara, Lynn et al. (2006) Minimal residual disease in hairy cell leukemia patients assessed by clone-specific polymerase chain reaction. Clin Cancer Res 12:2804-11 |
| Arons, Evgeny; Sunshine, Joel; Suntum, Tara et al. (2006) Somatic hypermutation and VH gene usage in hairy cell leukaemia. Br J Haematol 133:504-12 |
| Pastan, Ira; Hassan, Raffit; Fitzgerald, David J et al. (2006) Immunotoxin therapy of cancer. Nat Rev Cancer 6:559-65 |
