Phase I Trial of Recombinant Immunotoxin RFB4(dsFv)-PE38 (BL22) in Patients With B-Cell Malignancies
http://www.100md.com
《临床肿瘤学》
the Laboratory of Molecular Biology, Laboratory of Clinical Pathology, and Medicine Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD
ABSTRACT
PURPOSE: To conduct a phase I trial of recombinant immunotoxin BL22, an anti-CD22 Fv fragment fused to truncated Pseudomonas exotoxin.
PATIENTS AND METHODS: Forty-six pretreated patients with CD22+ non-Hodgkin's lymphoma (NHL; n = 4), chronic lymphocytic leukemia (CLL; n = 11), and hairy cell leukemia (HCL; n = 31) received 265 cycles at 3 to 50 μg/Kg every other day x 3 doses.
RESULTS: BL22 was active in HCL, with 19 complete remissions (CRs; 61%) and six partial responses (PRs; 19%) in 31 patients. Of 19 CRs, 11 were achieved after one cycle and eight after two to 14 cycles. All 25 responders benefited clinically with one cycle. The CR rate was 86% in patients enrolled at 40 μg/Kg every other day x 3, and 41% at lower doses (P = .011). The median duration for CR was 36 months (range, 5 to 66 months), and eight patients remain in CR at 45 months (range, 29 to 66 months). Lower but significant activity occurred in CLL. Neutralizing antibodies occurred in 11 (24%) of 46 patients (all HCL). A reversible hemolytic uremic syndrome requiring plasmapheresis was observed in one patient with NHL during cycle 1 and in four patients with HCL during cycle 2 or 3. The maximum-tolerated dose (MTD) evaluated at cycle 1 was 40 μg/Kg IV. QOD x 3. The most common toxicities at 30 to 50 μg/Kg every other day x 3 included hypoalbuminemia, transaminase elevations, fatigue, and edema.
CONCLUSION: BL22 was well tolerated and highly effective in HCL, even after one cycle. Phase II testing is underway to define the efficacy with one cycle and to study safety when additional cycles are needed for optimal response.
INTRODUCTION
Standard treatment of indolent lymphomas and leukemias is rarely curative, necessitating the development of novel treatment. A number of treatments with novel mechanisms have been developed and include targeted immunotherapy with monoclonal antibodies (MAbs) or growth factors which are either unlabeled1-4 or conjugated to radionuclides,5-7 chemotherapy drugs,8,9 or protein toxins.10-13 Several clinical trials have previously been reported with immunotoxins containing deglycosylated ricin A chain conjugated to either the anti-CD22 antibody RFB414,15 or its Fab fragment,16 resulting in partial response rates of up to 38%.14-16
CD22 is a 135 kDa phosphoglycoprotein adhesion molecule which is expressed on normal B cells and in 70% to 90% of B-cell malignancies.16,17 We have developed a novel targeted therapeutic approach in which a bacterial toxin is fused to the Fv portion of an antibody.10 To target CD22, a recombinant immunotoxin, BL22, was developed containing the variable domains of the anti-CD22 MAb RFB4 fused to a 38 kDa portion of Pseudomonas exotoxin A.18 To stabilize the recombinant immunotoxin, the cloned variable domains of RFB4 were joined with a disulfide bond by engineering cysteine residues into the framework regions of VH and VL.18-20 The VH domain is fused to PE38, which contains the translocating and ADP ribosylating domains of the toxin, but not the binding domain which binds to normal cells.21-27 BL22 is cytotoxic toward fresh CD22+ B-cell leukemic cells from patients and produces complete remissions (CRs) in mice-bearing human CD22+ lymphoma xenografts.28,29 A phase I trial was undertaken in patients with B-cell hematologic malignancies. Of the first 31 patients enrolled onto the trial, 16 patients had hairy cell leukemia (HCL). A high response rate to BL22 was documented in these 16 patients in an interim report,30 including 11 CRs (69%) and two partial responses (13%; PRs). The following is a report of all 46 patients enrolled onto the phase I study, including a total of 31 patients with HCL.
PATIENTS AND METHODS
Patients
The phase I protocol was approved by the National Cancer Institute's (NCI) Investigators' Review Board. Three patients were initially enrolled at each dose level. If no patient developed a dose-limiting toxicity (DLT), subsequent patients were enrolled at the next dose level. However, if one patient developed DLT, dose escalation did not proceed until at least three additional patients without DLT were enrolled at the current dose level. The maximum-tolerated dose (MTD) was defined as the highest dose level on cycle 1 where zero to one patient of six had DLT. To be eligible, patients had to have CD22 documented on tumor cells either by flow cytometry, immunohistochemistry, or radiolabeled binding assay. Patients with HCL required prior treatment with a purine analog if they had a CR of less than a year or lower response to the last course. Patients with non-Hodgkin's lymphoma (NHL) and chronic lymphocytic leukemia (CLL) were eligible after failure of standard chemotherapy, and patients with aggressive NHL required failure of, refusal, or ineligibility for transplantation. All patients had to have disease requiring treatment, adequate pulmonary, renal, and hepatic function, and lack high levels of neutralizing antibodies to BL22 in the serum. Prior chemotherapy within 3 weeks or MAb therapy within 3 months before beginning BL22 was not allowed. Patients received BL22 diluted in 50 mL of saline containing 0.2% human serum albumin over a period of 30 minutes every other day for three doses (QOD x 3). Patients without progressive disease or high neutralizing antibody levels were eligible for re-treatment at 3 to 6 week intervals. The re-treatment dose level could be changed to the highest one that was considered safe. Patients with reversible grade 3 DLT could be re-treated at a lower dose level. Toxicity was graded by the NCI Common Toxicity Criteria version 2.0. DLT was defined as grade 3-4 toxicity except for transient grade 3 transaminase elevations fever, neutropenia, and thrombocytopenia. Response was assessed by computed tomography (CT) and analysis of peripheral blood and bone marrow. CR was defined as an absence of visible HCL by CT, Wright stains of peripheral blood, and hemotoxylin/eosin stains of bone marrow. PR was defined as a reduction in tumor burden by at least 50% by standard criteria with no mass increasing in size. CR and PR were assessed at least 4 weeks after the last dose of BL22. The criteria for major response lacked the need for reversal of cytopenias, which is required by chemotherapy trials31,32 because patients were heavily pretreated. However, hematologic remission (HR) was still tracked and was defined by neutrophil and platelet counts of at least 1.5 and 100 cells/L x 10–9, respectively, and hemoglobin of at least 11 g/dL. Duration of response was measured from the first time point at which patients met all criteria to the first time point of relapse. Marginal response (MR) was defined as a reduction of any lesion by at least 25% without any lesion showing a significant increase. Progressive disease (PD) was defined by at least a 25% increase in any measurable lesion, or more than a 50% increase in leukemic cell count. Stable disease (SD) was defined as the absence of either a response or PD. Of the 46 patients enrolled, 31 patients had HCL, 11 patients had CLL, and four patients had NHL. The first 16 HCL patients (patients 2, 4, 9, 14, 15, 17, 18, 20, 23, 25, and 26 to 31; Table 1) were included in a previous publication30 as patients 1 to 16.
Plasma Pharmacokinetics and Serum Antibody Assays
To determine levels of BL22, dilutions of plasma were incubated with Raji cells.29 Cytotoxicity assessed by the inhibition of [3H]leucine incorporation was used to interpolate on a standard curve as described.28,29 The plasma level was determined by multiplying the plasma dilution calculated to produce 50% inhibition of protein synthesis by the IC50, the concentration of purified BL22 which caused 50% protein synthesis inhibition. To assay for the presence of neutralizing antibodies, mixtures containing 90% serum and 10% BL22 (final BL22 concentration, 1,000 ng/mL) were incubated at 37°C for 15 to 20 minutes, diluted, cultured with Raji cells at a final BL22 concentration of 10 ng/mL, and % inhibition determined by [3H]leucine incorporation. The percent neutralization was defined as % inhibition in the presence of serum, subtracted from the % inhibition in the presence of 0.2% human serum albumin in phosphate-buffered saline (HSA-PBS), divided by the % inhibition in the presence of HSA-PBS, multiplied by 100%.
RESULTS
Patients
As shown in Table 1, patients treated had received a median of four (range, 1 to 11) prior treatments, and had a median Eastern Cooperative Oncology Group performance status of 1 (range, 0 to 2) and a median age of 54 years (range, 30 to 81 years). Among the 42 patients with HCL or CLL, the median circulating leukemic count was 0.53 x 109/L (range, 0 to 600 x 109) cells. Among the 31 HCL patients, three had HCLv, a poor prognosis variant of HCL which responds poorly to even initial chemotherapy.33-36
Dose Levels and DLT
Clinical results are summarized in Tables 1 and 2. Overall, 265 cycles were administered to 46 patients, of which 142 cycles were re-treatment cycles administered on protocol and 77 off protocol by special exemption. The number of cycles per patient depended on patient response, toxicity, drug supply, regulatory concerns, and relapse. Dose levels were 3 to 50 μg/Kg QOD x3. There were no drug-related deaths. The most common DLT was hemolytic uremic syndrome (HUS), a syndrome involving platelet and fibrin deposition in the renal glomerular endothelium, microangiopathic hemolytic anemia due to intravascular fibrin strands, thrombocytopenia, and renal insufficiency.37 HUS was observed in one NHL patient during cycle 1, and four HCL patients during cycles 2 (patients 20 and 28) or 3 (patients 36 and 38). Unlike all other patients, the NHL patient received BL22 by a prolonged infusion due to a grade 1 allergic reaction to the test dose. Complete recovery from HUS in this patient could not be assessed due to progressive NHL and refusal of salvage chemotherapy. This patient died with disseminated intravascular coagulation, including increasing coagulation times and decreasing fibrinogen, abnormalities which were not observed during her HUS event, and based on their timing (> 30 days after BL22) were considered related to rapidly progressive NHL. However, all four HCL patients completely recovered following 6 to 12 days of plasmapheresis, and renal function in these patients remains normal after 39 to 57 months (median, 46 months) of follow-up from study entry. Moreover, these patients achieved complete clearance of HCL cells from the blood and bone marrow with resolution of pre-existing hematologic abnormalities. No other patients had allergic reactions to the test dose. The only other DLT observed was a cytokine release syndrome (high fever, bone pain, hypotension) with systemic (but not pulmonary) vascular leak syndrome (VLS) during cycle 2 (patient 15), completely resolving after 3 days. Because the HUS during cycle 2 at 50 μg/Kg QOD x 3 was the most severe case of toxicity among the HCL patients, and because another patient at this dose level (patient 27) had a grade 1 creatinine elevation during cycle 2, no additional patients were enrolled at this dose level.
Adjunct Measures Tried to Prevent Toxicity
To prevent allergic reactions and fever, patients received prophylactic acetaminophen, hydroxyzine, and ranitidine, and no dose-limiting allergic reactions were observed. Prophylactic anti-inflammatory agents, ultimately judged not to be useful, were administered to most of the first 31 patients enrolled onto the study. As shown in Table 1, these agents consisted of either (1) nonsteroidal anti-inflammatory drugs (NSAIDs) alone, (2) high-dose steroids, or (3) an NSAID plus an anti-TNF agent (etanercept or infliximab). Some of the first 12 patients enrolled at dose levels of 3 to 20 μg/Kg QOD x 3 had prophylactic NSAIDs, but only during re-treatment, and none had DLT. The six patients enrolled at the 30 μg/Kg QOD x 3 dose each received an NSAID during cycle 1 and some re-treatment cycles; two of these six patients had DLT, one patient had DLT during cycle 1 (HUS), and one patient during cycle 2 (VLS). In an attempt to reduce inflammation leading to VLS, we administered prophylactic high-dose steroids to the seventh patient enrolled at the 30 μg/Kg QOD x 3 dose. However, an inflammatory syndrome and Pneumocystis carinii pneumonia occurred after prednisone was decreased. We therefore tried rofecoxib and infliximab in the next 12 patients enrolled at dose levels 30 (n = 3), 40 (n = 6), and 50 (n = 3) μg/Kg QOD x 3. However, one patient each at dose levels 30 and 50 μg/Kg QOD x 3 developed HUS during cycle 2. Once HUS appeared more serious than inflammatory syndromes, we stopped using anti-inflammatory agents, and began using prophylactic intravenous (IV) fluid to minimize renal toxicity. Using prophylactic fluid, we enrolled patients at dose levels of 30 (n = 3) and 40 (n = 6) μg/Kg QOD x 3. Two patients in the latter group had HUS during cycle 3. In both cases, cycle 3 began only 16 days after the last dose of cycle 2. In an attempt to reduce the risk of HUS, the re-treatment interval was increased to 6 weeks and the dose level was decreased to 30 μg/Kg QOD x 3 on cycle 1 and 20 μg/Kg x 3 on subsequent cycles. Six additional patients were treated at this dose and schedule with prophylactic IV fluid, and no DLT was observed. Since no DLT was observed in 12 of 12 patients during cycle 1 of 40 μg/Kg QOD x 3, this dose level was the cycle 1 MTD, but the safe dose and interval for re-treatment were not defined.
Toxicity
At the MTD, BL22 during cycle 1 was well tolerated, with only grade 1-2 reversible toxicities (Table 2). The most common toxicities were fatigue, myalgia, ALT and AST elevations, and hypoalbuminemia each of which affected at least half of the patients. When considering all toxicities on all 172 cycles of BL22 administered at 30-50 μg/Kg QOD x 3 in phase I (Fig 1), hypoalbuminemia was most commonly observed. Transaminase elevations were usually low grade. Grade 3 transaminases were not considered dose limiting due to an absence of functional impairment as assessed by bilirubin, alkaline phosphatase, and prothrombin time and their rapid reversibility, with no grade 3 duration extending beyond 4 days. Likewise, fever, which occurred in 28 of 172 cycles including a patient at the grade 3 level, was not dose limiting. All the other grade 3 and 4 events shown in Figure 1, including the grade 3 VLS, occurred along with serious adverse events already discussed in this section and the two preceding sections. Any patient with at least a 5% weight gain was automatically considered to have grade 2 VLS, but we did not observe pulmonary edema. It was not clear whether systemic third spacing was due to hypoalbuminemia or a consequence of vascular damage. The decrease in albumin was accompanied by a decrease in immunoglobin G from precycle to day 7 to 10 of the cycle (r2 = 0.31; P < .0001; n = 80), indicating that the decrease in intravascular protein was not specific to albumin. Unlike purine analogs, BL22 did not adversely affect the CD4 count. Of 26 patients with available pretreatment and 6-month post-treatment CD4 counts, five, 17, and four patients had pretreatment CD4 counts of less than 50 (grade 4), 50 to 200 (grade 3), and 200 to 500 (grade 2), respectively. None of these 26 patients, treated at dose levels of 3 to 50 μg/Kg QOD x 3, showed any significant decrease in CD4 count with BL22. In contrast, 23 patients (88%) had a 1 to 3 grade improvement in CD4 count.
Immunogenicity
To determine the immunogenicity of BL22, a bioassay was used to detect neutralizing antibodies to BL22. Eleven of 46 patients presented with neutralizing antibodies after one to eight cycles of BL22 at dose levels of 20 to 50 μg/Kg QOD x 3 (Table 2). BL22 was not immunogenic in B-cell chronic lymphocytic leukemia or NHL, and most patients with HCL received multiple cycles of BL22 without presenting with neutralizing antibodies. The development of neutralizing antibodies in HCL was not dose related; high levels of neutralizing antibodies were observed in three of seven patients at the 20 μg/Kg QOD x 3 dose level and in zero of five patients at the 50 μg/Kg QOD x 3 dose level. Immunogenicity was not related to prior splenectomy (2 = 0.02; P > .05), or pre-existing CD4 counts (Wilcoxon, P = .48; n = 13). HCL patients received rituximab therapy less frequently than did non-HCL patients, and prior rituximab therapy in HCL was not associated with increased or decreased immunogenicity (2 = 2.6, P > .05).
Pharmacokinetics of BL22
To quantify BL22 in plasma, dilutions of plasma were incubated with Raji cells to detect biologically active drug levels. Peak levels were dose-related and median half-lives were about 3 hours at the 30 to 50 μg/Kg QOD x 3 dose level. Area under the concentration-time curve (AUC) significantly correlated with dose (Fig 2A), but was highly variable due to the high level of CD22 expression on malignant cells, causing a sink effect. With decreased tumor burden after achievement of CR, the peak level and AUC increased and the volume of distribution fell to the level of the plasma volume (Figs 2B through 2D). This provided the rationale for re-treatment at a lower dose level than the one used in cycle 1 for the last six patients enrolled. In examining the relationship between PK parameters, tumor burden, and response, we determined that HCL patients with CR after one cycle of BL22 had significantly higher AUCs (P = .01; Fig 2E) and lower concentrations of circulating HCL cells (P = .007; Fig 2F), compared with those HCL patients without a CR to cycle 1. Thus patients with lower tumor burden achieve higher AUCs and have a higher chance of a CR in cycle 1, which is particularly important when immunogenicity occurs early.
Response Rate and Treatment Outcome in HCL
As listed in Tables 1 and 2, patients with HCL had high response rates to BL22. A total of 25 of 31 patients responded with 19 CRs (61%) and six PRs (19%). Of the patients with HCLv, (patients 14, 18, and 26), all three achieved CR. Cytopenias improved in all responders and completely resolved in all CRs, in two of six PRs, and in one of patients patients with MR. Responding patients often had rapid reductions in circulating malignant counts (90% in 2 days; 99% in 1 week) and in spleen size. As listed in Table 1, only one HCL patient in CR had response associated with high-dose steroids (patient 18), which were used only in cycle 2; CR was not achieved in this patient until after cycle 4. Lack of a major response was associated with low doses (3 to 6 μg/Kg QOD x 3) in two patients, and massive (> 20 cm) abdominal lymph nodes in two patients. Five patients achieved CR and two patients achieved PR even though high levels of neutralizing antibodies arose after dosing. CRs were detected by bone marrow biopsy after one cycle in 11 patients and after two to 14 cycles in eight patients, so that re-treatment improved the best response. The CR rate was 86% (12 of 14) in patients enrolled at dose levels of at least 40 μg/Kg QOD x 3, compared with 41% (seven of 17) of HCL patients enrolled at lower dose levels (2 = 6.4; P = .011). CR rate after cycle 1 significantly correlated with dose level (P = .03). As listed in Table 1, CR in HCL was achieved in six (55%) of 11 patients with immunogenicity versus 13 (65%) of 20 patients without (2 = 0.33; P > .5).
As listed in Table 1, all HCL patients had prior treatment with purine analogs, with one to three prior courses of cladribine or pentostatin, as well as other agents. The majority of these patients had never achieved CR with prior therapy (patients 14, 18, 20, 23, 26 to 28, 33, 37, and 45), six patients had no major response to the last course of cladribine or pentostatin (patients 25, 30, 31, 35, 39, and 40), and three patients had an inadequate CR to the last course of purine analog (patients 9 and 17, < 6 months; patient 38, < 2 years).
Response Duration
As previously reported,30 three of the first 16 patients relapsed 6 to 12 months after achieving CR, and all three patients were returned to CR with additional BL22. Additional relapsing patients were not re-treated, generally because they maintained HRs. Of the 19 total CRs, the median CR duration was 36 months (range, 5 to 66 months; Fig 3A); eight patients remain in CR at a median of 45 months (range, 29 to 66 months) of follow-up. Of the 11 patients with CR after one cycle of BL22, there were three relapses, compared with six relapses among eight patients achieving CR with more than one cycle (2 = 4.23; P = .04). Thus patients responding more quickly to BL22 may have had a more durable response. The immediate often lifesaving effect of CR is hematologic recovery, and 22 (73%) of 30 patients with pre-existing cytopenias achieved hematologic remission, defined as neutrophils, platelets, and hemoglobin levels of at least 1,500/μL, 100,000/μL, and 11 g/dL, respectively. As shown in Figure 3B, the median HR duration was 37 months and 11 patients (50%) remain in HR at a median 48 months of follow-up.
Response Time Course in HCL
To determine how rapidly patients would respond to immunotoxin therapy, we assessed the numbers of circulating HCL cells at multiple time points, particularly during the first week of therapy. HCL patients varied widely with respect to disease burden at the time of protocol entry. Nevertheless, the decrease in the malignant count was often 90% after 2 days and 99% 1 week after protocol entry. In a typical patient (patient 23; Fig 4B) with classic HCL, in whom the malignant count was low and the spleen was enlarged, CR was achieved with one cycle of BL22 and one additional (consolidation) cycle was given. Also with classic HCL, patient 25 (Fig 4A) presented with a high HCL count, as is common after splenectomy. This patient required three cycles before achieving CR and then had two consolidation cycles. Patients with HCLv had very high malignant counts in the blood, whether the spleen had been removed or not. Patient 14 with HCLv (Fig 4C) required nine cycles to achieve CR (including six cycles after splenectomy performed for HCL-related hemorrhage), and then had two consolidation cycles. In contrast, patient 26 required only one cycle for CR and also had two consolidation cycles (Fig 4D). This exemplifies that disease burden does not prevent achievement of a CR to BL22, but sometimes additional cycles are required.
Minimal Residual Disease in HCL Patients Treated With BL22
HCL patients achieving CR to purine analogs have a 13% to 50% incidence of minimal residual disease (MRD), defined as collections of CD20+ or TRAP+ cells in the bone marrow biopsy by immunohistochemistry.38-41 The presence of MRD after cladribine or pentostatin is considered a risk factor for earlier relapse. Of the 19 patients achieving CR with BL22, only one patient (patient 27) had MRD by this definition. We found that flow cytometry of the peripheral blood is highly sensitive as an assay for MRD, and is able to detect less than 0.01% HCL cells due to strong expression of light chains, CD20, CD11c, and also expression of CD103.42 Flow cytometry was positive in two of 19 CRs (patients 18 and 38). As we previously reported,42 in patients treated with BL22 or other HCL therapies, polymerase chain reaction using consensus primers is less sensitive than flow cytometry for detecting MRD. In the patients in CR after BL22, none were positive by polymerase chain reaction. Thus BL22 was capable of inducing CRs without MRD in HCL.
Response in CLL and NHL
Unlike HCL, CRs were not observed in CLL and NHL, although the number of patients enrolled at the upper dose levels, 30 to 50 μg/Kg QOD x 3, were limited to four and two patients (Table 1). Marginal responses were observed in three CLL patients, one of whom had more than 99.9% eradication of circulating CLL cells and improvement in cytopenias, and received maintenance BL22 for almost 5 years. A less than 50% response in an abdominal mass prevented this patient from being considered a major responder. This patient was one of two MRs in CLL, which occurred in patients not pretreated with steroids.
DISCUSSION
We treated patients with CD22+ malignancies with BL22 and observed high rates of overall response (81%) and CR (61%). This experience extends the previous report of 11 of the first 16 patients achieving CR.30 BL22 is the first agent since purine analogs reported to have a high CR rate in HCL. Unlike most trials of purine analogs, all HCL patients on our protocol had prior treatment with cladribine and most were judged unlikely to respond to an additional course. The effectiveness of BL22 in chemoresistant patients is probably due to its different mechanism of action, in that HCL cells from chemoresistant patients continue to display high levels of surface CD22; this is exemplified by its high efficacy in HCLv, which is poorly responsive to chemotherapy but yet strongly expresses CD22.36
Duration of CR in HCL
Purine analogs are highly effective in HCL but cause long-term (median, 4 years) suppression of CD4+ T cells.43,44 A major advantage of BL22 over purine analogs is that it spares resting T cells, and hence it may be given repeatedly without cumulative myelotoxicity. Eight of 10 patients with a pretreatment CD4 count of less than 500 had a clinical response, and all eight patients (100%) experienced an improvement in CD4 counts of at least 1 to 2 grades after BL22. Although patients were limited to zero to two consolidation cycles due to drug supply and regulatory issues, it is possible that more consolidation cycles would have been effective in eliminating microscopic disease that could eventually cause relapse. Patients are not cured with purine analogs alone,45,46 and it is conceivable that targeting surface CD22 with or without other agents could improve the percentage of tumor cells eradicated.
Efficacy of BL22 in HCL Versus Other Disorders
HCL is particularly responsive to BL22, probably due to extremely high (5,000 to 70,000 sites/cell) expression of CD22. By comparison, CLL cells sensitive to BL22 may express around 1,000 and as few as 350 sites per cell.28 CLL and HCL patients who relapsed had unchanged levels of CD22 on the cells, which probably allowed them to be successfully re-treated. Additional patients with CLL and NHL will need to be enrolled at upper dose levels to define the activity of BL22 in this patient population.
Efficacy of Rituximab for HCL
Rituximab, targeting CD20, has been tested in four small trials of HCL patients,47-50 in which a total of 18 (30%) of 60 patients achieved CR. The two trials with the highest CR rates enrolled some patients without cytopenias,47,50 and in one trial up to 12 doses of rituximab were administered.50 Patients previously treated with rituximab must wait at least 3 months before receiving BL22, due to the long half-life of the MAb. Of 28 HCL patients enrolled at the 30 to 50 μg/Kg QOD x 3 dose level of BL22, eight patients received prior rituximab, and in that group there were three CRs and three PRs (Table 1).
Mechanism of HUS in Patients Treated With BL22
After closing the BL22 protocol to new accrual, HUS was observed in one of two HCL patients who received BL22 by special exemption before the opening of the phase II trial. BL22 was given at the 40 μg/Kg QOD x 3 dose level for cycle 1 and re-treatment was at 30 μg/Kg QOD x 3. In all six patients with HUS (one NHL, five HCL), ADAMTS-13 function37 was unimpaired. Thus, the mechanism of HUS in our patients is different from that of classic Shiga-like toxin-related HUS. Several hundred cycles of other PE38-containing recombinant toxins have been administered to patients without any detectable HUS, indicating that the mechanism is at least in part related to CD22. All five HCL patients who had HUS had been re-treated with BL22 within 4 weeks of beginning the prior cycle, and HUS in all five patients was complete reversible. Moreover, all five HCL patients with HUS experienced hematologic remission and three of the five patients had CR, suggesting that the risk-benefit ratio was clearly favorable even in those patients with HUS. Currently, HCL patients enrolled onto the BL22 phase II trial are re-treated (if re-treatment is deemed necessary) more than 8 weeks after cycle 1. Patients are also treated with continuous normal saline at a dose of 1 L/d for days 1 to 8 to prevent hypovolemia and renal toxicity. Anti-inflammatory agents, which were not considered useful in phase I, are not being used in the phase II trial. So far, HUS requiring plasmapheresis has not occurred in 18 patients.
In conclusion, BL22 is well tolerated and highly effective in HCL after one cycle at 40 μg/Kg QOD x 3. Its safety after re-treatment in HCL is currently under investigation. BL22 is also being tested in CLL and pediatric ALL. A higher affinity mutant of BL22,51 termed HA22, is currently being prepared for phase I testing, particularly for diseases like CLL with lower CD22 expression compared with HCL.
Authors' Disclosures of Potential Conflicts of Interest
Acknowledgment
The authors would like to recognize several research nurses involved in the care of our patients, including Karen Bergeron, Miranda Raggio, Michelle Zancan, Diana O'Hagan, Kelly Cahill, Rita Mincemoyer, and Linda Ellison. We also recognize the valuable contributions of several technicians, including Inger Margulies and Maria Gallo, and valuable data management from Barbara Debrah. We thank Seth Steinberg for statistical assistance. We recognize Elaine Jaffe for evaluating immunohistochemistry in several NHL patients. We also recognize the valuable contributions of Toby Hecht, PhD, and Steve Giardina, PhD, at the Monoclonal Antibody and Recombinant Protein Facility (MARP), Frederick, MD, Jay Greenblatt, PhD, Thomas Davis, MD, Helen Chen, MD, and Julie Rhie, PhD, at CTEP, and David Waters, PhD, at the SAIC, Frederick, MD. We recognize the important preclinical work on BL22 by Elizabeth Mansfield, PhD, the contribution of RFB4 antibody by Ellen Vitetta, MD, and the RFB4 hybridoma by Peter Amlot, MD. Lastly, we thank the many medical oncology fellows and nurses at the NIH Clinical Center who were involved in the care of our patients.
NOTES
Supported by the intramural program of the National Cancer Institute, National Institutes of Health, Bethesda, MD.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Christiansen J, Rajasekaran AK: Biological impediments to monoclonal antibody-based cancer immunotherapy. Mol Cancer Ther 3:1493-1501, 2004
Weiner GJ, Link BK: Antibody therapy of lymphoma. Adv Pharmacol 51:229-253, 2004
Lin TS, Byrd JC: Monoclonal antibody therapy in lymphoid leukemias. Adv Pharmacol 51:127-167, 2004
McDermott DF, Atkins MB: Application of IL-2 and other cytokines in renal cancer. Expert Opin Biol Ther 4:455-468, 2004
Majhail NS, Hussein M, Olencki TE, et al: Phase I trial of continuous infusion recombinant human interleukin-4 in patients with cancer. Invest New Drugs 22:421-426, 2004
Burke JM, Jurcic JG: Radioimmunotherapy of leukemia. Adv Pharmacol 51:185-208, 2004
Cheson BD: Radioimmunotherapy of non-Hodgkin lymphomas. Blood 101:391-398, 2003
Linenberger ML: CD33-directed therapy with gemtuzumab ozogamicin in acute myeloid leukemia: Progress in understanding cytotoxicity and potential mechanisms of drug resistance. Leukemia 19:176-182, 2005
Law CL, Cerveny CG, Gordon KA, et al: Efficient elimination of B-lineage lymphomas by anti-CD20-auristatin conjugates. Clin Cancer Res 10:7842-7851, 2004
FitzGerald DJ, Kreitman R, Wilson W, et al: Recombinant immunotoxins for treating cancer. Int J Med Microbiol 293:577-582, 2004
Frankel AE, Neville DM, Bugge TA, et al: Immunotoxin therapy of hematologic malignancies. Semin Oncol 30:545-557, 2003
Dang NH, Hagemeister FB, Pro B, et al: Phase II study of denileukin diftitox for relapsed/refractory B-Cell non-Hodgkin's lymphoma. J Clin Oncol 22:4095-4102, 2004
Kreitman RJ, Pastan I: Immunobiologic treatments of hairy-cell leukemia. Best Pract Res Clin Haematol 16:117-133, 2003
Amlot PL, Stone MJ, Cunningham D, et al: A phase I study of an anti-CD22-deglycosylated ricin A chain immunotoxin in the treatment of B-cell lymphomas resistant to conventional therapy. Blood 82:2624-2633, 1993
Sausville EA, Headlee D, Stetler-Stevenson M, et al: Continuous infusion of the anti-CD22 immunotoxin IgG-RFB4-SMPT-dgA in patients with B-cell lymphoma: A phase I study. Blood 85:3457-3465, 1995
Vitetta ES, Stone M, Amlot P, et al: Phase I immunotoxin trial in patients with B-cell lymphoma. Cancer Res 51:4052-4058, 1991
Clark EA: CD22, a B cell-specific receptor, mediates adhesion and signal transduction. J Immunol 150:4715-4718, 1993
Mansfield E, Amlot P, Pastan I, et al: Recombinant RFB4 immunotoxins exhibit potent cytotoxic activity for CD22-bearing cells and tumors. Blood 90:2020-2026, 1997
Reiter Y, Brinkmann U, Kreitman RJ, et al: Stabilization of the Fv fragments in recombinant immunotoxins by disulfide bonds engineered into conserved framework regions. Biochemistry 33:5451-5459, 1994
Reiter Y, Kreitman RJ, Brinkmann U, et al: Cytotoxic and antitumor activity of a recombinant immunotoxin composed of disulfide-stablized anti-Tac Fv fragment and truncated Pseudomonas exotoxin. Int J Cancer 58:142-149, 1994
Allured VS, Collier RJ, Carroll SF, et al: Structure of exotoxin A of Pseudomonas aeruginosa at 3.0 Angstrom resolution. Proc Natl Acad Sci U S A 83:1320-1324, 1986
Hessler JL, Kreitman RJ: An early step in Pseudomonas exotoxin action is removal of the terminal lysine residue, which allows binding to the KDEL receptor. Biochemistry 36:14577-14582, 1997
Chiron MF, Fryling CM, FitzGerald DJ: Cleavage of Pseudomonas exotoxin and diphtheria toxin by a furin-like enzyme prepared from beef liver. J Biol Chem 269:18167-18176, 1994
Ogata M, Fryling CM, Pastan I, et al: Cell-mediated cleavage of Pseudomonas exotoxin between Arg279 and Gly280 generates the enzymatically active fragment which translocates to the cytosol. J Biol Chem 267:25396-25401, 1992
Kreitman RJ, Pastan I: Importance of the glutamate residue of KDEL in increasing the cytotoxicity of Pseudomonas exotoxin derivatives and for increased binding to the KDEL receptor. Biochem J 307:29-37, 1995
Theuer C, Kasturi S, Pastan I: Domain II of Pseudomonas exotoxin A arrests the transfer of translocating nascent chains into mammalian microsomes. Biochemistry 33:5894-5900, 1994
Theuer CP, Buchner J, FitzGerald D, et al: The N-terminal region of the 37-kDa translocated fragment of Pseudomonas exotoxin A aborts translocation by promoting its own export after microsomal membrane insertion. Proc Natl Acad Sci U S A 90:7774-7778, 1993
Kreitman RJ, Margulies I, Stetler-Stevenson M, et al: Cytotoxic activity of disulfide-stabilized recombinant immunotoxin RFB4(dsFv)-PE38 (BL22) towards fresh malignant cells from patients with B-cell leukemias. Clin Cancer Res 6:1476-1487, 2000
Kreitman RJ, Wang QC, FitzGerald DJP, et al: Complete regression of human B-cell lymphoma xenografts in mice treated with recombinant anti-CD22 immunotoxin RFB4(dsFv)-PE38 at doses tolerated by Cynomolgus monkeys. Int J Cancer 81:148-155, 1999
Kreitman RJ, Wilson WH, Bergeron K, et al: Efficacy of the Anti-CD22 Recombinant Immunotoxin BL22 in Chemotherapy-Resistant Hairy-Cell Leukemia. N Engl J Med 345:241-247, 2001
Goodman GR, Burian C, Koziol JA, et al: Extended follow-up of patients with hairy cell leukemia after treatment with cladribine. J Clin Oncol 21:891-896, 2003
Cheson BD, Horning SJ, Coiffier B, et al: Report of an international workshop to standardize response criteria for non-Hodgkin's lymphomas. J Clin Oncol 17:1244-1253, 1999
Juliusson G, Liliemark J: Rapid recovery from cytopenia in hairy cell leukemia after treatment with 2-chloro-2'-deoxyadenosine (CdA): relation to opportunistic infections. Blood 79:888-894, 1992
Hoffman MA, Janson D, Rose E, et al: Treatment of hairy-cell leukemia with cladribine: Response, toxicity, and long-term follow-up. J Clin Oncol 15:1138-1142, 1997
Blasinska-Morawiec M, Robak T, Krykowski E, et al: Hairy cell leukemia-variant treated with 2-chlorodeoxyadenosine–a report of three cases. Leuk Lymphoma 25:381-385, 1997
Matutes E, Wotherspoon A, BritoBabapulle V, et al: The natural history and clinico-pathological features of the variant form of hairy cell leukemia. Leukemia 15:184-186, 2001
Moake JL: Thrombotic microangiopathies. N Engl J Med 347:589-600, 2002
Tallman MS, Hakimian D, Kopecky KJ, et al: Minimal residual disease in patients with hairy cell leukemia in complete remission treated with 2-chlorodeoxyadenosine or 2- deoxycoformycin and prediction of early relapse. Clin Cancer Res 5:1665-1670, 1999
Wheaton S, Tallman MS, Hakimian D, et al: Minimal residual disease may predict bone marrow relapse in patients with hairy cell leukemia treated with 2-chlorodeoxyadenosine. Blood 87:1556-1560, 1996
Matutes E, Meeus P, McLennan K, et al: The significance of minimal residual disease in hairy cell leukaemia treated with deoxycoformycin: A long-term follow-up study. Br J Haematol 98:375-383, 1997
Zak P, Chrobak L, Dedic K: Minimal residual disease, its detection and significance in hairy-cell leukemia. Acta Medica (Hradec Kralove) 42:85-88, 1999
Sausville JE, Salloum R, Sorbara L, et al: Minimal residual disease detection in hairy cell leukemia: Comparison of flow cytometric immunophenotyping with clonal analysis using consensus primer polymerase chain reaction for the heavy chain gene. Am J Clin Pathol 119:213-217, 2002
Seymour JF, Kurzrock R, Freireich EJ, et al: 2-chlorodeoxyadenosine induces durable remissions and prolonged suppression of CD4+ lymphocyte counts in patients with hairy cell leukemia. Blood 83:2906-2911, 1994
Seymour JF, Talpaz M, Kurzrock R: Response duration and recovery of CD4+ lymphocytes following deoxycoformycin in interferon-alpha-resistant hairy cell leukemia: 7-year follow-up. Leukemia 11:42-47, 1997
Filleul B, Delannoy A, Ferrant A, et al: A single course of 2-chloro-deoxyadenosine does not eradicate leukemic cells in hairy cell leukemia patients in complete remission. Leukemia 8:1153-1156, 1994
Carbone A, Reato G, Di Celle PF, et al: Disease eradication in hairy cell leukemia patients treated with 2- chlorodeoxyadenosine. Leukemia 8:2019-2020, 1994 (letter)
Hagberg H, Lundholm L: Rituximab, a chimaeric anti-CD20 monoclonal antibody, in the treatment of hairy cell leukaemia. Br J Haematol 115:609-611, 2001
Lauria F, Lenoci M, Annino L, et al: Efficacy of anti-CD20 monoclonal antibodies (Mabthera) in patients with progressed hairy cell leukemia. Haematologica 86:1046-1050, 2001
Nieva J, Bethel K, Saven A: Phase 2 study of rituximab in the treatment of cladribine-failed patients with hairy cell leukemia. Blood 102:810-813, 2003
Thomas DA, O'Brien S, Bueso-Ramos C, et al: Rituximab in relapsed or refractory hairy cell leukemia. Blood 102:3906-3911, 2003
Salvatore G, Beers R, Margulies I, et al: Improved Cytotoxic activity towards cell lines and fresh leukemia cells of a mutant anti-CD22 immunotoxin obtained by antibody phage display. Clin Cancer Res 8:995-1002, 2002(Robert J. Kreitman, David)
ABSTRACT
PURPOSE: To conduct a phase I trial of recombinant immunotoxin BL22, an anti-CD22 Fv fragment fused to truncated Pseudomonas exotoxin.
PATIENTS AND METHODS: Forty-six pretreated patients with CD22+ non-Hodgkin's lymphoma (NHL; n = 4), chronic lymphocytic leukemia (CLL; n = 11), and hairy cell leukemia (HCL; n = 31) received 265 cycles at 3 to 50 μg/Kg every other day x 3 doses.
RESULTS: BL22 was active in HCL, with 19 complete remissions (CRs; 61%) and six partial responses (PRs; 19%) in 31 patients. Of 19 CRs, 11 were achieved after one cycle and eight after two to 14 cycles. All 25 responders benefited clinically with one cycle. The CR rate was 86% in patients enrolled at 40 μg/Kg every other day x 3, and 41% at lower doses (P = .011). The median duration for CR was 36 months (range, 5 to 66 months), and eight patients remain in CR at 45 months (range, 29 to 66 months). Lower but significant activity occurred in CLL. Neutralizing antibodies occurred in 11 (24%) of 46 patients (all HCL). A reversible hemolytic uremic syndrome requiring plasmapheresis was observed in one patient with NHL during cycle 1 and in four patients with HCL during cycle 2 or 3. The maximum-tolerated dose (MTD) evaluated at cycle 1 was 40 μg/Kg IV. QOD x 3. The most common toxicities at 30 to 50 μg/Kg every other day x 3 included hypoalbuminemia, transaminase elevations, fatigue, and edema.
CONCLUSION: BL22 was well tolerated and highly effective in HCL, even after one cycle. Phase II testing is underway to define the efficacy with one cycle and to study safety when additional cycles are needed for optimal response.
INTRODUCTION
Standard treatment of indolent lymphomas and leukemias is rarely curative, necessitating the development of novel treatment. A number of treatments with novel mechanisms have been developed and include targeted immunotherapy with monoclonal antibodies (MAbs) or growth factors which are either unlabeled1-4 or conjugated to radionuclides,5-7 chemotherapy drugs,8,9 or protein toxins.10-13 Several clinical trials have previously been reported with immunotoxins containing deglycosylated ricin A chain conjugated to either the anti-CD22 antibody RFB414,15 or its Fab fragment,16 resulting in partial response rates of up to 38%.14-16
CD22 is a 135 kDa phosphoglycoprotein adhesion molecule which is expressed on normal B cells and in 70% to 90% of B-cell malignancies.16,17 We have developed a novel targeted therapeutic approach in which a bacterial toxin is fused to the Fv portion of an antibody.10 To target CD22, a recombinant immunotoxin, BL22, was developed containing the variable domains of the anti-CD22 MAb RFB4 fused to a 38 kDa portion of Pseudomonas exotoxin A.18 To stabilize the recombinant immunotoxin, the cloned variable domains of RFB4 were joined with a disulfide bond by engineering cysteine residues into the framework regions of VH and VL.18-20 The VH domain is fused to PE38, which contains the translocating and ADP ribosylating domains of the toxin, but not the binding domain which binds to normal cells.21-27 BL22 is cytotoxic toward fresh CD22+ B-cell leukemic cells from patients and produces complete remissions (CRs) in mice-bearing human CD22+ lymphoma xenografts.28,29 A phase I trial was undertaken in patients with B-cell hematologic malignancies. Of the first 31 patients enrolled onto the trial, 16 patients had hairy cell leukemia (HCL). A high response rate to BL22 was documented in these 16 patients in an interim report,30 including 11 CRs (69%) and two partial responses (13%; PRs). The following is a report of all 46 patients enrolled onto the phase I study, including a total of 31 patients with HCL.
PATIENTS AND METHODS
Patients
The phase I protocol was approved by the National Cancer Institute's (NCI) Investigators' Review Board. Three patients were initially enrolled at each dose level. If no patient developed a dose-limiting toxicity (DLT), subsequent patients were enrolled at the next dose level. However, if one patient developed DLT, dose escalation did not proceed until at least three additional patients without DLT were enrolled at the current dose level. The maximum-tolerated dose (MTD) was defined as the highest dose level on cycle 1 where zero to one patient of six had DLT. To be eligible, patients had to have CD22 documented on tumor cells either by flow cytometry, immunohistochemistry, or radiolabeled binding assay. Patients with HCL required prior treatment with a purine analog if they had a CR of less than a year or lower response to the last course. Patients with non-Hodgkin's lymphoma (NHL) and chronic lymphocytic leukemia (CLL) were eligible after failure of standard chemotherapy, and patients with aggressive NHL required failure of, refusal, or ineligibility for transplantation. All patients had to have disease requiring treatment, adequate pulmonary, renal, and hepatic function, and lack high levels of neutralizing antibodies to BL22 in the serum. Prior chemotherapy within 3 weeks or MAb therapy within 3 months before beginning BL22 was not allowed. Patients received BL22 diluted in 50 mL of saline containing 0.2% human serum albumin over a period of 30 minutes every other day for three doses (QOD x 3). Patients without progressive disease or high neutralizing antibody levels were eligible for re-treatment at 3 to 6 week intervals. The re-treatment dose level could be changed to the highest one that was considered safe. Patients with reversible grade 3 DLT could be re-treated at a lower dose level. Toxicity was graded by the NCI Common Toxicity Criteria version 2.0. DLT was defined as grade 3-4 toxicity except for transient grade 3 transaminase elevations fever, neutropenia, and thrombocytopenia. Response was assessed by computed tomography (CT) and analysis of peripheral blood and bone marrow. CR was defined as an absence of visible HCL by CT, Wright stains of peripheral blood, and hemotoxylin/eosin stains of bone marrow. PR was defined as a reduction in tumor burden by at least 50% by standard criteria with no mass increasing in size. CR and PR were assessed at least 4 weeks after the last dose of BL22. The criteria for major response lacked the need for reversal of cytopenias, which is required by chemotherapy trials31,32 because patients were heavily pretreated. However, hematologic remission (HR) was still tracked and was defined by neutrophil and platelet counts of at least 1.5 and 100 cells/L x 10–9, respectively, and hemoglobin of at least 11 g/dL. Duration of response was measured from the first time point at which patients met all criteria to the first time point of relapse. Marginal response (MR) was defined as a reduction of any lesion by at least 25% without any lesion showing a significant increase. Progressive disease (PD) was defined by at least a 25% increase in any measurable lesion, or more than a 50% increase in leukemic cell count. Stable disease (SD) was defined as the absence of either a response or PD. Of the 46 patients enrolled, 31 patients had HCL, 11 patients had CLL, and four patients had NHL. The first 16 HCL patients (patients 2, 4, 9, 14, 15, 17, 18, 20, 23, 25, and 26 to 31; Table 1) were included in a previous publication30 as patients 1 to 16.
Plasma Pharmacokinetics and Serum Antibody Assays
To determine levels of BL22, dilutions of plasma were incubated with Raji cells.29 Cytotoxicity assessed by the inhibition of [3H]leucine incorporation was used to interpolate on a standard curve as described.28,29 The plasma level was determined by multiplying the plasma dilution calculated to produce 50% inhibition of protein synthesis by the IC50, the concentration of purified BL22 which caused 50% protein synthesis inhibition. To assay for the presence of neutralizing antibodies, mixtures containing 90% serum and 10% BL22 (final BL22 concentration, 1,000 ng/mL) were incubated at 37°C for 15 to 20 minutes, diluted, cultured with Raji cells at a final BL22 concentration of 10 ng/mL, and % inhibition determined by [3H]leucine incorporation. The percent neutralization was defined as % inhibition in the presence of serum, subtracted from the % inhibition in the presence of 0.2% human serum albumin in phosphate-buffered saline (HSA-PBS), divided by the % inhibition in the presence of HSA-PBS, multiplied by 100%.
RESULTS
Patients
As shown in Table 1, patients treated had received a median of four (range, 1 to 11) prior treatments, and had a median Eastern Cooperative Oncology Group performance status of 1 (range, 0 to 2) and a median age of 54 years (range, 30 to 81 years). Among the 42 patients with HCL or CLL, the median circulating leukemic count was 0.53 x 109/L (range, 0 to 600 x 109) cells. Among the 31 HCL patients, three had HCLv, a poor prognosis variant of HCL which responds poorly to even initial chemotherapy.33-36
Dose Levels and DLT
Clinical results are summarized in Tables 1 and 2. Overall, 265 cycles were administered to 46 patients, of which 142 cycles were re-treatment cycles administered on protocol and 77 off protocol by special exemption. The number of cycles per patient depended on patient response, toxicity, drug supply, regulatory concerns, and relapse. Dose levels were 3 to 50 μg/Kg QOD x3. There were no drug-related deaths. The most common DLT was hemolytic uremic syndrome (HUS), a syndrome involving platelet and fibrin deposition in the renal glomerular endothelium, microangiopathic hemolytic anemia due to intravascular fibrin strands, thrombocytopenia, and renal insufficiency.37 HUS was observed in one NHL patient during cycle 1, and four HCL patients during cycles 2 (patients 20 and 28) or 3 (patients 36 and 38). Unlike all other patients, the NHL patient received BL22 by a prolonged infusion due to a grade 1 allergic reaction to the test dose. Complete recovery from HUS in this patient could not be assessed due to progressive NHL and refusal of salvage chemotherapy. This patient died with disseminated intravascular coagulation, including increasing coagulation times and decreasing fibrinogen, abnormalities which were not observed during her HUS event, and based on their timing (> 30 days after BL22) were considered related to rapidly progressive NHL. However, all four HCL patients completely recovered following 6 to 12 days of plasmapheresis, and renal function in these patients remains normal after 39 to 57 months (median, 46 months) of follow-up from study entry. Moreover, these patients achieved complete clearance of HCL cells from the blood and bone marrow with resolution of pre-existing hematologic abnormalities. No other patients had allergic reactions to the test dose. The only other DLT observed was a cytokine release syndrome (high fever, bone pain, hypotension) with systemic (but not pulmonary) vascular leak syndrome (VLS) during cycle 2 (patient 15), completely resolving after 3 days. Because the HUS during cycle 2 at 50 μg/Kg QOD x 3 was the most severe case of toxicity among the HCL patients, and because another patient at this dose level (patient 27) had a grade 1 creatinine elevation during cycle 2, no additional patients were enrolled at this dose level.
Adjunct Measures Tried to Prevent Toxicity
To prevent allergic reactions and fever, patients received prophylactic acetaminophen, hydroxyzine, and ranitidine, and no dose-limiting allergic reactions were observed. Prophylactic anti-inflammatory agents, ultimately judged not to be useful, were administered to most of the first 31 patients enrolled onto the study. As shown in Table 1, these agents consisted of either (1) nonsteroidal anti-inflammatory drugs (NSAIDs) alone, (2) high-dose steroids, or (3) an NSAID plus an anti-TNF agent (etanercept or infliximab). Some of the first 12 patients enrolled at dose levels of 3 to 20 μg/Kg QOD x 3 had prophylactic NSAIDs, but only during re-treatment, and none had DLT. The six patients enrolled at the 30 μg/Kg QOD x 3 dose each received an NSAID during cycle 1 and some re-treatment cycles; two of these six patients had DLT, one patient had DLT during cycle 1 (HUS), and one patient during cycle 2 (VLS). In an attempt to reduce inflammation leading to VLS, we administered prophylactic high-dose steroids to the seventh patient enrolled at the 30 μg/Kg QOD x 3 dose. However, an inflammatory syndrome and Pneumocystis carinii pneumonia occurred after prednisone was decreased. We therefore tried rofecoxib and infliximab in the next 12 patients enrolled at dose levels 30 (n = 3), 40 (n = 6), and 50 (n = 3) μg/Kg QOD x 3. However, one patient each at dose levels 30 and 50 μg/Kg QOD x 3 developed HUS during cycle 2. Once HUS appeared more serious than inflammatory syndromes, we stopped using anti-inflammatory agents, and began using prophylactic intravenous (IV) fluid to minimize renal toxicity. Using prophylactic fluid, we enrolled patients at dose levels of 30 (n = 3) and 40 (n = 6) μg/Kg QOD x 3. Two patients in the latter group had HUS during cycle 3. In both cases, cycle 3 began only 16 days after the last dose of cycle 2. In an attempt to reduce the risk of HUS, the re-treatment interval was increased to 6 weeks and the dose level was decreased to 30 μg/Kg QOD x 3 on cycle 1 and 20 μg/Kg x 3 on subsequent cycles. Six additional patients were treated at this dose and schedule with prophylactic IV fluid, and no DLT was observed. Since no DLT was observed in 12 of 12 patients during cycle 1 of 40 μg/Kg QOD x 3, this dose level was the cycle 1 MTD, but the safe dose and interval for re-treatment were not defined.
Toxicity
At the MTD, BL22 during cycle 1 was well tolerated, with only grade 1-2 reversible toxicities (Table 2). The most common toxicities were fatigue, myalgia, ALT and AST elevations, and hypoalbuminemia each of which affected at least half of the patients. When considering all toxicities on all 172 cycles of BL22 administered at 30-50 μg/Kg QOD x 3 in phase I (Fig 1), hypoalbuminemia was most commonly observed. Transaminase elevations were usually low grade. Grade 3 transaminases were not considered dose limiting due to an absence of functional impairment as assessed by bilirubin, alkaline phosphatase, and prothrombin time and their rapid reversibility, with no grade 3 duration extending beyond 4 days. Likewise, fever, which occurred in 28 of 172 cycles including a patient at the grade 3 level, was not dose limiting. All the other grade 3 and 4 events shown in Figure 1, including the grade 3 VLS, occurred along with serious adverse events already discussed in this section and the two preceding sections. Any patient with at least a 5% weight gain was automatically considered to have grade 2 VLS, but we did not observe pulmonary edema. It was not clear whether systemic third spacing was due to hypoalbuminemia or a consequence of vascular damage. The decrease in albumin was accompanied by a decrease in immunoglobin G from precycle to day 7 to 10 of the cycle (r2 = 0.31; P < .0001; n = 80), indicating that the decrease in intravascular protein was not specific to albumin. Unlike purine analogs, BL22 did not adversely affect the CD4 count. Of 26 patients with available pretreatment and 6-month post-treatment CD4 counts, five, 17, and four patients had pretreatment CD4 counts of less than 50 (grade 4), 50 to 200 (grade 3), and 200 to 500 (grade 2), respectively. None of these 26 patients, treated at dose levels of 3 to 50 μg/Kg QOD x 3, showed any significant decrease in CD4 count with BL22. In contrast, 23 patients (88%) had a 1 to 3 grade improvement in CD4 count.
Immunogenicity
To determine the immunogenicity of BL22, a bioassay was used to detect neutralizing antibodies to BL22. Eleven of 46 patients presented with neutralizing antibodies after one to eight cycles of BL22 at dose levels of 20 to 50 μg/Kg QOD x 3 (Table 2). BL22 was not immunogenic in B-cell chronic lymphocytic leukemia or NHL, and most patients with HCL received multiple cycles of BL22 without presenting with neutralizing antibodies. The development of neutralizing antibodies in HCL was not dose related; high levels of neutralizing antibodies were observed in three of seven patients at the 20 μg/Kg QOD x 3 dose level and in zero of five patients at the 50 μg/Kg QOD x 3 dose level. Immunogenicity was not related to prior splenectomy (2 = 0.02; P > .05), or pre-existing CD4 counts (Wilcoxon, P = .48; n = 13). HCL patients received rituximab therapy less frequently than did non-HCL patients, and prior rituximab therapy in HCL was not associated with increased or decreased immunogenicity (2 = 2.6, P > .05).
Pharmacokinetics of BL22
To quantify BL22 in plasma, dilutions of plasma were incubated with Raji cells to detect biologically active drug levels. Peak levels were dose-related and median half-lives were about 3 hours at the 30 to 50 μg/Kg QOD x 3 dose level. Area under the concentration-time curve (AUC) significantly correlated with dose (Fig 2A), but was highly variable due to the high level of CD22 expression on malignant cells, causing a sink effect. With decreased tumor burden after achievement of CR, the peak level and AUC increased and the volume of distribution fell to the level of the plasma volume (Figs 2B through 2D). This provided the rationale for re-treatment at a lower dose level than the one used in cycle 1 for the last six patients enrolled. In examining the relationship between PK parameters, tumor burden, and response, we determined that HCL patients with CR after one cycle of BL22 had significantly higher AUCs (P = .01; Fig 2E) and lower concentrations of circulating HCL cells (P = .007; Fig 2F), compared with those HCL patients without a CR to cycle 1. Thus patients with lower tumor burden achieve higher AUCs and have a higher chance of a CR in cycle 1, which is particularly important when immunogenicity occurs early.
Response Rate and Treatment Outcome in HCL
As listed in Tables 1 and 2, patients with HCL had high response rates to BL22. A total of 25 of 31 patients responded with 19 CRs (61%) and six PRs (19%). Of the patients with HCLv, (patients 14, 18, and 26), all three achieved CR. Cytopenias improved in all responders and completely resolved in all CRs, in two of six PRs, and in one of patients patients with MR. Responding patients often had rapid reductions in circulating malignant counts (90% in 2 days; 99% in 1 week) and in spleen size. As listed in Table 1, only one HCL patient in CR had response associated with high-dose steroids (patient 18), which were used only in cycle 2; CR was not achieved in this patient until after cycle 4. Lack of a major response was associated with low doses (3 to 6 μg/Kg QOD x 3) in two patients, and massive (> 20 cm) abdominal lymph nodes in two patients. Five patients achieved CR and two patients achieved PR even though high levels of neutralizing antibodies arose after dosing. CRs were detected by bone marrow biopsy after one cycle in 11 patients and after two to 14 cycles in eight patients, so that re-treatment improved the best response. The CR rate was 86% (12 of 14) in patients enrolled at dose levels of at least 40 μg/Kg QOD x 3, compared with 41% (seven of 17) of HCL patients enrolled at lower dose levels (2 = 6.4; P = .011). CR rate after cycle 1 significantly correlated with dose level (P = .03). As listed in Table 1, CR in HCL was achieved in six (55%) of 11 patients with immunogenicity versus 13 (65%) of 20 patients without (2 = 0.33; P > .5).
As listed in Table 1, all HCL patients had prior treatment with purine analogs, with one to three prior courses of cladribine or pentostatin, as well as other agents. The majority of these patients had never achieved CR with prior therapy (patients 14, 18, 20, 23, 26 to 28, 33, 37, and 45), six patients had no major response to the last course of cladribine or pentostatin (patients 25, 30, 31, 35, 39, and 40), and three patients had an inadequate CR to the last course of purine analog (patients 9 and 17, < 6 months; patient 38, < 2 years).
Response Duration
As previously reported,30 three of the first 16 patients relapsed 6 to 12 months after achieving CR, and all three patients were returned to CR with additional BL22. Additional relapsing patients were not re-treated, generally because they maintained HRs. Of the 19 total CRs, the median CR duration was 36 months (range, 5 to 66 months; Fig 3A); eight patients remain in CR at a median of 45 months (range, 29 to 66 months) of follow-up. Of the 11 patients with CR after one cycle of BL22, there were three relapses, compared with six relapses among eight patients achieving CR with more than one cycle (2 = 4.23; P = .04). Thus patients responding more quickly to BL22 may have had a more durable response. The immediate often lifesaving effect of CR is hematologic recovery, and 22 (73%) of 30 patients with pre-existing cytopenias achieved hematologic remission, defined as neutrophils, platelets, and hemoglobin levels of at least 1,500/μL, 100,000/μL, and 11 g/dL, respectively. As shown in Figure 3B, the median HR duration was 37 months and 11 patients (50%) remain in HR at a median 48 months of follow-up.
Response Time Course in HCL
To determine how rapidly patients would respond to immunotoxin therapy, we assessed the numbers of circulating HCL cells at multiple time points, particularly during the first week of therapy. HCL patients varied widely with respect to disease burden at the time of protocol entry. Nevertheless, the decrease in the malignant count was often 90% after 2 days and 99% 1 week after protocol entry. In a typical patient (patient 23; Fig 4B) with classic HCL, in whom the malignant count was low and the spleen was enlarged, CR was achieved with one cycle of BL22 and one additional (consolidation) cycle was given. Also with classic HCL, patient 25 (Fig 4A) presented with a high HCL count, as is common after splenectomy. This patient required three cycles before achieving CR and then had two consolidation cycles. Patients with HCLv had very high malignant counts in the blood, whether the spleen had been removed or not. Patient 14 with HCLv (Fig 4C) required nine cycles to achieve CR (including six cycles after splenectomy performed for HCL-related hemorrhage), and then had two consolidation cycles. In contrast, patient 26 required only one cycle for CR and also had two consolidation cycles (Fig 4D). This exemplifies that disease burden does not prevent achievement of a CR to BL22, but sometimes additional cycles are required.
Minimal Residual Disease in HCL Patients Treated With BL22
HCL patients achieving CR to purine analogs have a 13% to 50% incidence of minimal residual disease (MRD), defined as collections of CD20+ or TRAP+ cells in the bone marrow biopsy by immunohistochemistry.38-41 The presence of MRD after cladribine or pentostatin is considered a risk factor for earlier relapse. Of the 19 patients achieving CR with BL22, only one patient (patient 27) had MRD by this definition. We found that flow cytometry of the peripheral blood is highly sensitive as an assay for MRD, and is able to detect less than 0.01% HCL cells due to strong expression of light chains, CD20, CD11c, and also expression of CD103.42 Flow cytometry was positive in two of 19 CRs (patients 18 and 38). As we previously reported,42 in patients treated with BL22 or other HCL therapies, polymerase chain reaction using consensus primers is less sensitive than flow cytometry for detecting MRD. In the patients in CR after BL22, none were positive by polymerase chain reaction. Thus BL22 was capable of inducing CRs without MRD in HCL.
Response in CLL and NHL
Unlike HCL, CRs were not observed in CLL and NHL, although the number of patients enrolled at the upper dose levels, 30 to 50 μg/Kg QOD x 3, were limited to four and two patients (Table 1). Marginal responses were observed in three CLL patients, one of whom had more than 99.9% eradication of circulating CLL cells and improvement in cytopenias, and received maintenance BL22 for almost 5 years. A less than 50% response in an abdominal mass prevented this patient from being considered a major responder. This patient was one of two MRs in CLL, which occurred in patients not pretreated with steroids.
DISCUSSION
We treated patients with CD22+ malignancies with BL22 and observed high rates of overall response (81%) and CR (61%). This experience extends the previous report of 11 of the first 16 patients achieving CR.30 BL22 is the first agent since purine analogs reported to have a high CR rate in HCL. Unlike most trials of purine analogs, all HCL patients on our protocol had prior treatment with cladribine and most were judged unlikely to respond to an additional course. The effectiveness of BL22 in chemoresistant patients is probably due to its different mechanism of action, in that HCL cells from chemoresistant patients continue to display high levels of surface CD22; this is exemplified by its high efficacy in HCLv, which is poorly responsive to chemotherapy but yet strongly expresses CD22.36
Duration of CR in HCL
Purine analogs are highly effective in HCL but cause long-term (median, 4 years) suppression of CD4+ T cells.43,44 A major advantage of BL22 over purine analogs is that it spares resting T cells, and hence it may be given repeatedly without cumulative myelotoxicity. Eight of 10 patients with a pretreatment CD4 count of less than 500 had a clinical response, and all eight patients (100%) experienced an improvement in CD4 counts of at least 1 to 2 grades after BL22. Although patients were limited to zero to two consolidation cycles due to drug supply and regulatory issues, it is possible that more consolidation cycles would have been effective in eliminating microscopic disease that could eventually cause relapse. Patients are not cured with purine analogs alone,45,46 and it is conceivable that targeting surface CD22 with or without other agents could improve the percentage of tumor cells eradicated.
Efficacy of BL22 in HCL Versus Other Disorders
HCL is particularly responsive to BL22, probably due to extremely high (5,000 to 70,000 sites/cell) expression of CD22. By comparison, CLL cells sensitive to BL22 may express around 1,000 and as few as 350 sites per cell.28 CLL and HCL patients who relapsed had unchanged levels of CD22 on the cells, which probably allowed them to be successfully re-treated. Additional patients with CLL and NHL will need to be enrolled at upper dose levels to define the activity of BL22 in this patient population.
Efficacy of Rituximab for HCL
Rituximab, targeting CD20, has been tested in four small trials of HCL patients,47-50 in which a total of 18 (30%) of 60 patients achieved CR. The two trials with the highest CR rates enrolled some patients without cytopenias,47,50 and in one trial up to 12 doses of rituximab were administered.50 Patients previously treated with rituximab must wait at least 3 months before receiving BL22, due to the long half-life of the MAb. Of 28 HCL patients enrolled at the 30 to 50 μg/Kg QOD x 3 dose level of BL22, eight patients received prior rituximab, and in that group there were three CRs and three PRs (Table 1).
Mechanism of HUS in Patients Treated With BL22
After closing the BL22 protocol to new accrual, HUS was observed in one of two HCL patients who received BL22 by special exemption before the opening of the phase II trial. BL22 was given at the 40 μg/Kg QOD x 3 dose level for cycle 1 and re-treatment was at 30 μg/Kg QOD x 3. In all six patients with HUS (one NHL, five HCL), ADAMTS-13 function37 was unimpaired. Thus, the mechanism of HUS in our patients is different from that of classic Shiga-like toxin-related HUS. Several hundred cycles of other PE38-containing recombinant toxins have been administered to patients without any detectable HUS, indicating that the mechanism is at least in part related to CD22. All five HCL patients who had HUS had been re-treated with BL22 within 4 weeks of beginning the prior cycle, and HUS in all five patients was complete reversible. Moreover, all five HCL patients with HUS experienced hematologic remission and three of the five patients had CR, suggesting that the risk-benefit ratio was clearly favorable even in those patients with HUS. Currently, HCL patients enrolled onto the BL22 phase II trial are re-treated (if re-treatment is deemed necessary) more than 8 weeks after cycle 1. Patients are also treated with continuous normal saline at a dose of 1 L/d for days 1 to 8 to prevent hypovolemia and renal toxicity. Anti-inflammatory agents, which were not considered useful in phase I, are not being used in the phase II trial. So far, HUS requiring plasmapheresis has not occurred in 18 patients.
In conclusion, BL22 is well tolerated and highly effective in HCL after one cycle at 40 μg/Kg QOD x 3. Its safety after re-treatment in HCL is currently under investigation. BL22 is also being tested in CLL and pediatric ALL. A higher affinity mutant of BL22,51 termed HA22, is currently being prepared for phase I testing, particularly for diseases like CLL with lower CD22 expression compared with HCL.
Authors' Disclosures of Potential Conflicts of Interest
Acknowledgment
The authors would like to recognize several research nurses involved in the care of our patients, including Karen Bergeron, Miranda Raggio, Michelle Zancan, Diana O'Hagan, Kelly Cahill, Rita Mincemoyer, and Linda Ellison. We also recognize the valuable contributions of several technicians, including Inger Margulies and Maria Gallo, and valuable data management from Barbara Debrah. We thank Seth Steinberg for statistical assistance. We recognize Elaine Jaffe for evaluating immunohistochemistry in several NHL patients. We also recognize the valuable contributions of Toby Hecht, PhD, and Steve Giardina, PhD, at the Monoclonal Antibody and Recombinant Protein Facility (MARP), Frederick, MD, Jay Greenblatt, PhD, Thomas Davis, MD, Helen Chen, MD, and Julie Rhie, PhD, at CTEP, and David Waters, PhD, at the SAIC, Frederick, MD. We recognize the important preclinical work on BL22 by Elizabeth Mansfield, PhD, the contribution of RFB4 antibody by Ellen Vitetta, MD, and the RFB4 hybridoma by Peter Amlot, MD. Lastly, we thank the many medical oncology fellows and nurses at the NIH Clinical Center who were involved in the care of our patients.
NOTES
Supported by the intramural program of the National Cancer Institute, National Institutes of Health, Bethesda, MD.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Christiansen J, Rajasekaran AK: Biological impediments to monoclonal antibody-based cancer immunotherapy. Mol Cancer Ther 3:1493-1501, 2004
Weiner GJ, Link BK: Antibody therapy of lymphoma. Adv Pharmacol 51:229-253, 2004
Lin TS, Byrd JC: Monoclonal antibody therapy in lymphoid leukemias. Adv Pharmacol 51:127-167, 2004
McDermott DF, Atkins MB: Application of IL-2 and other cytokines in renal cancer. Expert Opin Biol Ther 4:455-468, 2004
Majhail NS, Hussein M, Olencki TE, et al: Phase I trial of continuous infusion recombinant human interleukin-4 in patients with cancer. Invest New Drugs 22:421-426, 2004
Burke JM, Jurcic JG: Radioimmunotherapy of leukemia. Adv Pharmacol 51:185-208, 2004
Cheson BD: Radioimmunotherapy of non-Hodgkin lymphomas. Blood 101:391-398, 2003
Linenberger ML: CD33-directed therapy with gemtuzumab ozogamicin in acute myeloid leukemia: Progress in understanding cytotoxicity and potential mechanisms of drug resistance. Leukemia 19:176-182, 2005
Law CL, Cerveny CG, Gordon KA, et al: Efficient elimination of B-lineage lymphomas by anti-CD20-auristatin conjugates. Clin Cancer Res 10:7842-7851, 2004
FitzGerald DJ, Kreitman R, Wilson W, et al: Recombinant immunotoxins for treating cancer. Int J Med Microbiol 293:577-582, 2004
Frankel AE, Neville DM, Bugge TA, et al: Immunotoxin therapy of hematologic malignancies. Semin Oncol 30:545-557, 2003
Dang NH, Hagemeister FB, Pro B, et al: Phase II study of denileukin diftitox for relapsed/refractory B-Cell non-Hodgkin's lymphoma. J Clin Oncol 22:4095-4102, 2004
Kreitman RJ, Pastan I: Immunobiologic treatments of hairy-cell leukemia. Best Pract Res Clin Haematol 16:117-133, 2003
Amlot PL, Stone MJ, Cunningham D, et al: A phase I study of an anti-CD22-deglycosylated ricin A chain immunotoxin in the treatment of B-cell lymphomas resistant to conventional therapy. Blood 82:2624-2633, 1993
Sausville EA, Headlee D, Stetler-Stevenson M, et al: Continuous infusion of the anti-CD22 immunotoxin IgG-RFB4-SMPT-dgA in patients with B-cell lymphoma: A phase I study. Blood 85:3457-3465, 1995
Vitetta ES, Stone M, Amlot P, et al: Phase I immunotoxin trial in patients with B-cell lymphoma. Cancer Res 51:4052-4058, 1991
Clark EA: CD22, a B cell-specific receptor, mediates adhesion and signal transduction. J Immunol 150:4715-4718, 1993
Mansfield E, Amlot P, Pastan I, et al: Recombinant RFB4 immunotoxins exhibit potent cytotoxic activity for CD22-bearing cells and tumors. Blood 90:2020-2026, 1997
Reiter Y, Brinkmann U, Kreitman RJ, et al: Stabilization of the Fv fragments in recombinant immunotoxins by disulfide bonds engineered into conserved framework regions. Biochemistry 33:5451-5459, 1994
Reiter Y, Kreitman RJ, Brinkmann U, et al: Cytotoxic and antitumor activity of a recombinant immunotoxin composed of disulfide-stablized anti-Tac Fv fragment and truncated Pseudomonas exotoxin. Int J Cancer 58:142-149, 1994
Allured VS, Collier RJ, Carroll SF, et al: Structure of exotoxin A of Pseudomonas aeruginosa at 3.0 Angstrom resolution. Proc Natl Acad Sci U S A 83:1320-1324, 1986
Hessler JL, Kreitman RJ: An early step in Pseudomonas exotoxin action is removal of the terminal lysine residue, which allows binding to the KDEL receptor. Biochemistry 36:14577-14582, 1997
Chiron MF, Fryling CM, FitzGerald DJ: Cleavage of Pseudomonas exotoxin and diphtheria toxin by a furin-like enzyme prepared from beef liver. J Biol Chem 269:18167-18176, 1994
Ogata M, Fryling CM, Pastan I, et al: Cell-mediated cleavage of Pseudomonas exotoxin between Arg279 and Gly280 generates the enzymatically active fragment which translocates to the cytosol. J Biol Chem 267:25396-25401, 1992
Kreitman RJ, Pastan I: Importance of the glutamate residue of KDEL in increasing the cytotoxicity of Pseudomonas exotoxin derivatives and for increased binding to the KDEL receptor. Biochem J 307:29-37, 1995
Theuer C, Kasturi S, Pastan I: Domain II of Pseudomonas exotoxin A arrests the transfer of translocating nascent chains into mammalian microsomes. Biochemistry 33:5894-5900, 1994
Theuer CP, Buchner J, FitzGerald D, et al: The N-terminal region of the 37-kDa translocated fragment of Pseudomonas exotoxin A aborts translocation by promoting its own export after microsomal membrane insertion. Proc Natl Acad Sci U S A 90:7774-7778, 1993
Kreitman RJ, Margulies I, Stetler-Stevenson M, et al: Cytotoxic activity of disulfide-stabilized recombinant immunotoxin RFB4(dsFv)-PE38 (BL22) towards fresh malignant cells from patients with B-cell leukemias. Clin Cancer Res 6:1476-1487, 2000
Kreitman RJ, Wang QC, FitzGerald DJP, et al: Complete regression of human B-cell lymphoma xenografts in mice treated with recombinant anti-CD22 immunotoxin RFB4(dsFv)-PE38 at doses tolerated by Cynomolgus monkeys. Int J Cancer 81:148-155, 1999
Kreitman RJ, Wilson WH, Bergeron K, et al: Efficacy of the Anti-CD22 Recombinant Immunotoxin BL22 in Chemotherapy-Resistant Hairy-Cell Leukemia. N Engl J Med 345:241-247, 2001
Goodman GR, Burian C, Koziol JA, et al: Extended follow-up of patients with hairy cell leukemia after treatment with cladribine. J Clin Oncol 21:891-896, 2003
Cheson BD, Horning SJ, Coiffier B, et al: Report of an international workshop to standardize response criteria for non-Hodgkin's lymphomas. J Clin Oncol 17:1244-1253, 1999
Juliusson G, Liliemark J: Rapid recovery from cytopenia in hairy cell leukemia after treatment with 2-chloro-2'-deoxyadenosine (CdA): relation to opportunistic infections. Blood 79:888-894, 1992
Hoffman MA, Janson D, Rose E, et al: Treatment of hairy-cell leukemia with cladribine: Response, toxicity, and long-term follow-up. J Clin Oncol 15:1138-1142, 1997
Blasinska-Morawiec M, Robak T, Krykowski E, et al: Hairy cell leukemia-variant treated with 2-chlorodeoxyadenosine–a report of three cases. Leuk Lymphoma 25:381-385, 1997
Matutes E, Wotherspoon A, BritoBabapulle V, et al: The natural history and clinico-pathological features of the variant form of hairy cell leukemia. Leukemia 15:184-186, 2001
Moake JL: Thrombotic microangiopathies. N Engl J Med 347:589-600, 2002
Tallman MS, Hakimian D, Kopecky KJ, et al: Minimal residual disease in patients with hairy cell leukemia in complete remission treated with 2-chlorodeoxyadenosine or 2- deoxycoformycin and prediction of early relapse. Clin Cancer Res 5:1665-1670, 1999
Wheaton S, Tallman MS, Hakimian D, et al: Minimal residual disease may predict bone marrow relapse in patients with hairy cell leukemia treated with 2-chlorodeoxyadenosine. Blood 87:1556-1560, 1996
Matutes E, Meeus P, McLennan K, et al: The significance of minimal residual disease in hairy cell leukaemia treated with deoxycoformycin: A long-term follow-up study. Br J Haematol 98:375-383, 1997
Zak P, Chrobak L, Dedic K: Minimal residual disease, its detection and significance in hairy-cell leukemia. Acta Medica (Hradec Kralove) 42:85-88, 1999
Sausville JE, Salloum R, Sorbara L, et al: Minimal residual disease detection in hairy cell leukemia: Comparison of flow cytometric immunophenotyping with clonal analysis using consensus primer polymerase chain reaction for the heavy chain gene. Am J Clin Pathol 119:213-217, 2002
Seymour JF, Kurzrock R, Freireich EJ, et al: 2-chlorodeoxyadenosine induces durable remissions and prolonged suppression of CD4+ lymphocyte counts in patients with hairy cell leukemia. Blood 83:2906-2911, 1994
Seymour JF, Talpaz M, Kurzrock R: Response duration and recovery of CD4+ lymphocytes following deoxycoformycin in interferon-alpha-resistant hairy cell leukemia: 7-year follow-up. Leukemia 11:42-47, 1997
Filleul B, Delannoy A, Ferrant A, et al: A single course of 2-chloro-deoxyadenosine does not eradicate leukemic cells in hairy cell leukemia patients in complete remission. Leukemia 8:1153-1156, 1994
Carbone A, Reato G, Di Celle PF, et al: Disease eradication in hairy cell leukemia patients treated with 2- chlorodeoxyadenosine. Leukemia 8:2019-2020, 1994 (letter)
Hagberg H, Lundholm L: Rituximab, a chimaeric anti-CD20 monoclonal antibody, in the treatment of hairy cell leukaemia. Br J Haematol 115:609-611, 2001
Lauria F, Lenoci M, Annino L, et al: Efficacy of anti-CD20 monoclonal antibodies (Mabthera) in patients with progressed hairy cell leukemia. Haematologica 86:1046-1050, 2001
Nieva J, Bethel K, Saven A: Phase 2 study of rituximab in the treatment of cladribine-failed patients with hairy cell leukemia. Blood 102:810-813, 2003
Thomas DA, O'Brien S, Bueso-Ramos C, et al: Rituximab in relapsed or refractory hairy cell leukemia. Blood 102:3906-3911, 2003
Salvatore G, Beers R, Margulies I, et al: Improved Cytotoxic activity towards cell lines and fresh leukemia cells of a mutant anti-CD22 immunotoxin obtained by antibody phage display. Clin Cancer Res 8:995-1002, 2002(Robert J. Kreitman, David)