Interim Phase I Results of Recombinant Immunotoxin HA22 in Patients with Hairy Cell and Chronic Lymphocytic Leukemias

For the treatment of B cell malignancies, we have developed a high affinity recombinant immunotoxin (HA22), which is composed of the Fv portion of an anti-CD22 antibody fused to a 38 kDa fragment of Pseudomonas exotoxin A. HA22 is an improved form of immunotoxin BL22, which produced a 50–60% complete remission (CR) rate in cladribine-resistant hairy-cell leukemia (HCL). A partial response (PR) rate of 17% was observed with BL22 in chemoresistant chronic lymphocytic leukemia (CLL). The higher response rate of BL22 in HCL compared to CLL is related to its higher cytotoxicity for HCL than CLL cells ex vivo (median IC50 4.3 ng/ml in 33 HCL patients vs 430 ng/ml in 70 CLL patients, p < 0.0001), and the higher CD22 expression on HCL than on CLL cells (median 44,000 vs 1250 sites/cell).To increase cytotoxic activity, the affinity of the Fv of BL22 was increased about 15-fold by phage display by mutating 3 amino acids in HCDR3 from SSY to THW. Patients with HCL and CLL were treated on 2 separate multicenter phase I protocols. To be eligible, patient required at least 2 prior standard therapies and a medical indication for treatment. HA22 was administered by 30 min infusion every other day for 3 doses (QOD ×3) and patients could be retreated at the same dose level every 4 weeks unless progressive disease or immunogenicity occurred. A standard 3–6 dose escalation scheme was used, allowing new patients to be enrolled at a higher dose level if 0 of 3 or 1 of 6 patients at the current level had dose limiting toxicity (DLT). A total of 16 HCL patients were treated with HA22 at 5, 10, 20, 30, and 40 ug/Kg QOD ×3. Three patients were treated at each of the first 4 dose levels, and a 4^th patient received 40 ug/Kg QOD ×3 because one patient at that dose level received only 1 cycle. Of 14 evaluable patients who received a total of 45 cycles, 4 (29%) developed high levels of neutralizing antibodies. No DLT was observed on any cycle, although 1 patient developed a grade II hemolytic uremic syndrome (HUS) during cycle 3 of 30 ug/Kg QOD ×3. Major responses included 6 (38%) CRs and 6 (38%) PRs out of 16 evaluable HCL patients (unmonitored). Responses were observed at all dose levels including 3 of 3 PRs at the lowest dose level, and 1–2 CRs at each of the next dose levels. Like BL22, peak plasma levels of HA22 were inversely related to tumor burden as assessed by circulating malignant cells or spleen size, and increased with repeated doses as patients responded. Phase I testing of HA22 in CLL is at an earlier stage with 11 patients enrolled at dose levels of 5, 10, and 20 ug/Kg QOD ×3, and decreases in circulating CLL cells have been observed. In conclusion, HA22 has not yet shown DLT in phase I testing and is highly active in HCL. Compared to BL22, HA22 shows no evidence of increased toxicity due to enhanced binding affinity to CD22.