A Phase I Study of a Combination of anti-CD19 and anti-CD22 Immunotoxins (Combotox) in Adult Patients with Refractory B-Lineage Acute Lymphoblastic Leukaemia

Abstract

Novel agents are needed for patients with refractory and relapsed acute lymphoblastic leukaemia (ALL). Combotox is a 1:1 mixture of two immunotoxins (ITs), prepared by coupling deglycosylated ricin A chain (dgRTA) to monoclonal antibodies directed against CD22 (RFB4-dgRTA) and CD19 (HD37-dgRTA). Pre-clinical data demonstrated that Combotox was effective in killing both pre-B-ALL cell lines and cells from patients with pre-B ALL. A clinical study of paediatric patients in which 3 of 17 patients with ALL experienced complete remission, supported the preclinical work and motivated this study. This study was a Phase I, dose-escalation trial using Combotox in adults with refractory or relapsed B-lineage-ALL. A cycle consisted of 3 doses, with one dose given every other day. Dose levels were 3, 5, 6, 7 and 8 mg/m² per dose. Seventeen patients, aged 19-72 years, were enrolled in this multi-institution study. The maximum tolerated dose was 7 mg/m²/dose (21 mg/m²/cycle) and vascular leak syndrome was the dose-limiting toxicity. Two patients developed reversible grade 3 elevations in liver function tests. One patient achieved partial remission and proceeded to allogeneic stem cell transplantation. All patients with peripheral blasts experienced decreased blast counts following the administration of Combotox. Thus, Combotox can be safely administered to adults with refractory leukaemia.

Introduction

In 2008, the estimated number of new cases of acute lymphoblastic leukaemia ALL in the United States was 5,430 and the number of deaths from this disease was 1,460 (http://www.seer.cancer.gov/statfacts/html/alyl.html). Patients who experience a relapse after remission can be expected to succumb within one year even if a second complete remission (CR) is achieved.

Both anti-CD19 (HD37) (Pezzutto, et al 1987) and anti-CD22 (RFB4) (Campana, et al 1985) are murine IgG₁ monoclonal antibodies (MAbs) that are coupled to deglycosylated ricin A chain (dgRTA, previously called dgA) to create the immunotoxins (ITs) HD37-dgRTA and RFB4-dgRTA, which bind to and kill human pre-B-ALL blasts in vitro (Herrera, et al 2000). In preclinical studies, when used alone in vitro, the cytotoxic activity of each IT was specific and reproducible.

When used together in a 1:1 combination (Combotox), the combination was more effective than either IT alone (Herrera, et al 2000). Individual ITs or Combotox treatment administered to severe combined immunodeficiency (SCID) mice daily for 4 days beginning either the day after inoculation (early disease) with NALM-6-UM1 (a pre-B-ALL cell line) or 2 weeks after inoculation (late disease) was effective. In both models, Combotox was more active than either IT alone, resulting in cures and prolonged survival (Herrera, et al 2000, Herrera, et al 2003). Encouraging clinical data were observed in a clinical trial of paediatric patients with pre-B-ALL, where 3/17 patients experienced CR (Herrera et al 2009). We, therefore, designed and conducted a Phase I study in adults with BALL. The maximum tolerated dose (MTD) was 7 mg/m² × 3 and vascular leak syndrome (VLS) was the dose-limiting toxicity (DLT). Combotox led to decreases in circulating peripheral leukaemic blasts in all evaluable patients and resulted in 2 partial remissions (PRs). Thus, Combotox could be given safely to adults with ALL and had modest clinical activity in this highly refractory disease.

Methods

Study Design and Patient Population

This was a Phase I multi-centre, dose escalation trial in adult patients with relapsed or refractory B-ALL. The aims of this study were: 1) to determine the MTD and DLT of Combotox in adult ALL; (2) to determine whether Combotox was active in ALL; and 3) to determine the pharmacokinetics (PK) and development of human anti-mouse antibodies (HAMA) and human anti-RTA (HARA) antibodies.

Eligibility criteria included: 1) disease that was refractory to conventional agents or recurrent disease based on marrow histology, cytogenetic studies, or polymerase chain reaction (PCR) amplification, 2) absence of HAMA, and 3) CD19 or CD22 had to be expressed on at least 50% of the tumour cells as previously described (Herrera et al 2009). Patients and/or their legal guardians signed informed consent following approval by the Institutional Review Boards of the Albert Einstein College of Medicine, MD Anderson Cancer Center and University of Texas Southwestern Medical Center (UTSWMC).

Administration of Combotox

Both RFB4-dgRTA and HD37-dgRTA were prepared in our good manufacturing practice (GMP) facility at UTSWMC as previously reported (Amlot, et al 1993, Stone, et al 1996). Combotox was prepared at each site by mixing the two ITs in a 1:1 ratio on the day of use. A cycle of treatment with Combotox consisted of 3 doses administered every other day by intravenous infusion over 4 h. Dose levels were 3 mg/m², 5 mg/m², 6 mg/m², 7 mg/m² and 8 mg/m².

Adverse reactions (ADRs) were graded according to the National Cancer Institute Common Toxicity Criteria version 2.0 (http://ctep.cancer.gov/protocolDevelopment/electronic_applications/docs/ctcv20_4-30-992.pdf). Capillary leak was defined as the development of pleural effusions with or without a pulmonary deficit, a 10-20% weight gain, or pulmonary oedema. VLS was defined as the manifestation of the constellation of symptoms of peripheral oedema, weight gain, tachycardia, hypoalbuminemia, with or without hypotension and oliguria (Messmann, et al 2000, Stone, et al 1996). Repeat cycles were permitted in the absence of grade 3 or 4 toxicity and the absence of HAMA/HARA (≤1 μg/ml). Intra-patient dose escalation was permitted in the absence of a clinical response. The MTD was defined as the dose level below which one patient experienced grade 4 toxicity or three or more patients experienced grade 3 toxicity. DLTs were defined as adverse experiences that defined the MTD regardless of whether they required discontinuation of dosing.

Dose escalation plan

3 patients were to be entered at each dose level. If grade 3 toxicity occurred, a maximum of 3 more patients were to be entered at that dose level. If a total of 3 patients experienced grade 3 toxicity, the MTD was defined as that dose level. If a grade 4 toxicity occurred, 6 patients were to be treated at the next lower dose level to determine if it was safe. The time interval between cycles was 1 week.

Disease Assessment and Response Criteria

CR was defined as the attainment of an M1 (definition = < 5% blasts) bone marrow status on day 15 with no evidence of circulating blasts or extramedullary disease. PR was defined as the complete disappearance of circulating blasts and achievement of an M2 (≤ 25% blasts) marrow status on day 15. Progressive Disease (PD) was defined as an increase in the circulating leukaemic blast count of ≥ 25% on day 15 or the development of extramedullary disease. A patient who failed to qualify as a CR, PR or PD was defined as having stable disease (SD). Patients with molecular evidence of disease with PCR testing for Philadelphia chromosome, even with less than 5% blasts, were considered to have disease. For these patients, attainment of PCR negativity was considered a response.

PKs and HAMA/HARA

Serum levels of Combotox from samples obtained pre-dosing, at the end of infusion, and at 4, 8, 20 and 44 h following the end of infusion were determined as previously described (Herrera et al 2009). HAMA and HARA levels were determined on day 15 and, if not retreated, on day 28 (Amlot, et al 1993).

Results

Demographics

The demographics of the patients are shown in Table I.

Table I. Demographics, Disease History and Dose Level by Patient.

Patient Number	Dose level (mg/m2)	No. of Cycles	Sex	Age (year)	Disease Duration at Entry (years)	%RFB4+1	%HD37+1	No. of Prior Therapies/Relapses/Transplants/Radiation therapy	BCR-ABL1
1	3	2	F	26	0.5	100	100	2/0/0/0	+
2	3	2	M	22	2.1	53	76	4/2/0/0
3	3	2	F	24	6	81	60	3/1/0/0
4	5	1	M	34	1.9	80	85	1/1/0/0
5	5	1	F	70	2	18	99	2/2/0/0
6	5	1	M	60	0.5	22	97	5/0/0/0	+
7	6	2	M	55	2	45	64	3/3/0/1
8	6	2**	M	35	1	70	99	1/1/1/0	+
9	6	1	F	32	1	46	99	2/2/0/0
10	7	1	M	47	1	66	69	3/3/0/0
11	7	1*	M	19	1.8	6	94	4/1/0/0
12	7	1	F	52	1	90	80	2/2/0/0
13	7	2**	M	36	2	22	75	2/2/0/0
14	7	2	M	30	1	93	99	2/2/0/0
15	7	1	F	32	1	91	97	2/2/0/0
16	7	1	M	55	3	13	99	3/3/0/0
17	8	1	M	24	6	81	99	3/3/1/0

¹Percentage of tumour cells stained with RFB4 or HD37.

^*Discontinued after 2 doses course 1 for elevated gamma glutamyl transpeptidase

^**Discontinued cycle 2: Patient 8 for pruritus and erythema, and Patient 13 for hypersensitivity reaction.

The 17 patients ranged in age from 19 to 72 years, and included 6 females and 11 males. The number of prior therapies ranged from 1 to 5; the median number of relapses was 2, and 2 patients had undergone stem cell transplantation (SCT). All screened patients fulfilled the criteria of at least 50% positivity for either CD19 or CD22 antigen expression on the lymphoblasts.

Dosing and Toxicity

Table I lists the dose level at which each patient was treated and the number of cycles received. Table II lists the severe (grade 3-4) toxicities that were considered to be possibly or probably related to treatment with Combotox.

Table II. Severe Toxicity Possibly or Probably Associated with Combotox and Response by Dose Level.

Combotox dose levels (mg/m2)	Number of patients	Toxicity ≥ grade 3 and possibly attributed to Combotox (affected patients)	Best response1
3	3	None	1 SD, 2 PD
5	3	None	3 PD
6	3	None	1 PR, 2 PD
7	7	Grade 3 LFT abnormalities (Patients 11, 12)	7 PD
8	1	Grade 4 VLS (Patient 17)	1 PD
TOTAL	17		1 PR, 1 SD, 15 PD

¹PR = partial remission; SD= stable disease; PD = progressive disease LFT, liver function test; VLS, vascular leak syndrome

The DLT was grade 4 VLS experienced by Patient 17, who had refractory disease and had previously received an allogeneic SCT. The patient was heavily pretreated and had received treatment on paediatric oncology protocol 9906 (vincristine, prednisone, daunorubicin, asparaginase, methotrexate, 6-mercaptopurine, cytanbine, and cytoxan), Hyper-CVAD, clofarabine, cytarabine, mitoxantrone and annamycin. He also had a history of VLS during prior chemotherapy. He developed hypotension and fluid overload 3 days after the last dose of Combotox. This was complicated by progressive renal failure. He decompensated over the next five days and expired due to multi-organ failure and possible concomitant sepsis.

Patients 11 and 12 developed grade 3 elevated liver function tests (LFTs). Both patients had a history of preexisting LFT abnormalities, but had normal LFTs at the time of enrollment. Patient 11 developed gamma glutamyl transpeptidase elevations that reversed a few days after stopping the drug. This patient had fatty liver infiltrations at entry as determined by computerized axial tomography scan. Patient 12, who had Hepatitis B, developed elevations in aspartic transaminase and alanine transaminase that peaked at levels of 657 u/l and 541 u/l respectively, 4 days after the last dose of Combotox. These levels had gradually dropped to 306 u/l and 431 u/l when the patient was discharged to hospice for progressive disease.

Two patients discontinued dosing during the second cycle of treatment. Patient 8 received two doses, developed pruritus and rash after the second dose and discontinued treatment. Two other cases of mild erythema occurred. Patient 13, whose dose was elevated from 7 mg/m² for the first cycle to 8 mg/m² for the second cycle, experienced hypersensitivity, hypotension with a reduced O₂ tension of 91% during the second cycle and discontinued treatment after the second dose.

PKs and Immunogenicity

The C_max and T_1/2 for individual patients are listed in Table III.

Table III. Pharmacokinetics Following the First Infusion and Peripheral Blast Count.

Patient number	Dose level (mg/m2)	RFB4-dgRTA: Cmax (ng/ml)	RFB4- dgRTA: T½(h)	HD37-dgRTA: Cmax (ng/ml)	HD37- dgRTA: T½(h)	Absolute blast count (×109/l) (pre-entry)	Change in Blasts	HAMA RFB4 (Post Course 1)	HAMA HD37 (Post Course 1)	HARA (Post Course 1)
1	3	189	<4	131	3.6	0	NE	4706	0	0
2	3	458	<4	256	<4	0.008	↓	0	0	0
3	3	182	<4	194	4-8	10.16	↓	0	0	0
Averages		276	<4	193	4
4	5	354	2.9	274.7	10.1	0.0015	↓	0	0	0
5	5	190	<4	93	∼4	15.48	↓	NE	NE	NE
6	5	NE	NE	NE	NE	47.13	↓	NE	NE	NE
Averages		271	<4	184	7
7	6	367	<4	518	14.3	0.0376	↓	0	0	0
8	6	>43**	NE	>419**	3.4	0.0228	↓	0	0	0
9	6	548	12.8	733	10.0	0.0260	↓	0	0	0
Averages		458	8	626	9
10*	7	>37	<4	>163	18.5	21.664	↓	0	0	0
11	7	335	<4	80	<4	1.003	↓	0	0	0
12	7	1451	9.9	513	11.8	0.0220	NE	0	0	0
13	7	MS				2.709	NE	0	0	0
14	7	66	<4	263	5.8	9.844	↓	0	0	0
15	7	1417	9.2	646	12.6	0.0079	NE	NE	NE	NE
16	7	476	10.2	184	4.3	1.024	↓	NE	NE	NE
Averages		749	7	321	9			NE	NE	NE
17	8	1333	3.5	593	2.7	NE	NE	NE	NE	NE

NE = Not evaluable;

^↓= decreased

C_max levels, especially for RFB4-dgRTA, correlated with dose level and the number of circulating blasts. The mean dose levels of RFB4-dgRTA ranged from 276 ng/ml for the 3 mg/m² dose group to 749 ng/ml for the 7 mg/m² dose group. Within the 7 mg/m² dose group, the Cmax for RFB4-dgRTA ranged from 66 ng/ml [9844 blasts/ml] [Patient 14] to 1451 ng/ml [220 blasts/ml] [Patient 12].

The serum half-life (T½) of Combotox in some patients was extremely short, < 4 h, while in others it was as long as 18.5 h. The T½ was likewise affected by the number of circulating blasts: <4 h [9844 blasts/ml] and 9.9 h [220 blasts/ml] hours for Patients 14 and 12, respectively.

The C_max levels in the two patients with severe ADRs (Patients 12 and 17), were 1,451 ng/ml and 1,333 ng/ml. The C_max in Patient 13, the third patient with a severe ADR, could not be determined.

One out of 14 patients tested for HAMA and HARA was positive for HAMA after course 1. This is lower than in patients without circulating blasts, confirming that circulating blasts reduce the rate of HAMA/HARA.

Response

Table II lists the response and dose level for each patient.

Patient 7 had 80% involvement of the bone marrow with blasts, which decreased to 10% after Combotox treatment. He received a second cycle, experienced a PR and was then eligible for and received an allogeneic SCT. He had successful engraftment but died of septic complication within 2 months post-transplant.

Patient 1 had 3% blasts in the bone marrow with Ph+ leukaemia by fluorescence in situ hybridization before Combotox treatment and had a reduction to <1% blasts post treatment. As she had 3% blasts at study entry, she did not meet criteria for PR and was graded as SD. The patient developed an increase in blast counts at 1 month post-treatment.

In all patients with peripheral blasts, Combotox administration led to reductions in leukaemic blasts immediately after administration. Graphs of peripheral blasts versus time are shown for all patients who had peripheral blasts (Fig 1). The decreases were seen very rapidly after drug administration and were not accompanied by any other cytopenias, demonstrating specific in vivo cytotoxicity of the drug.

Figure 1.

Absolute blast counts (×10⁹/l) vs Day of dosing in selected patients.

Discussion

RFB4-dgRTA and HD37-dgRTA have been used alone to treat over 100 patients with lymphoma and leukaemia. Combotox, a 1:1 mixture of RFB4-dgRTA and HD37-dgRTA, has been used to treat paediatric patients with pre-B-ALL. Both single agents and Combotox have demonstrated clinical activity. Combotox induced CRs in 3/17 paediatric patients with ALL. Therefore, we conducted a trial in adults with ALL.

There were 6 major findings to emerge from this study of 17 heavily pretreated adult ALL patients. 1) The MTD of Combotox in this patient population was 7 mg/m² × 3. This was much higher than in previous studies and was probably a result of a high number of circulating blast cells. The DLT was VLS. Graft-versus-host disease, which was the DLT in paediatric ALL, was not seen in the present cohort. This is perhaps due to the fact that only two patients had undergone a prior transplant, in contrast to the 9 children in the paediatric ALL study (Herrera et al 2009). 2) Elevated LFTs were seen in two patients. Elevated LFTs had not previously been observed in patients treated with Combotox or the individual immunotoxins. These may have been due to the higher than normal individual doses given in this study, or they may have been due to increased sensitization of the liver following prior chemotherapy that had induced elevated LFTs. 3) Combotox can be safely administered to adults with multiply relapsed leukaemia, but reduced toxicity might occur with tailored regimens as discussed below. 4) PK analysis revealed that serum levels of Combotox correlated with both the dose level and the number of circulating blasts. 5) 7% of the patients with B-lineage-ALL developed HAMA and/or HARA after the first cycle. This is about 25% of the rate observed previously in adults with non-Hodgkin lymphoma (NHL) and 50% of the rate observed in paediatric patients with Pre-B ALL (18%) (Herrera et al 2009; Amlot et al 1993; Stone et al 1996; Messmann et al 2000). 6) Combotox was effective in reducing the leukaemic burden in the majority of patients tested and in inducing a PR in 1 patient. It is possible that extended schedules might lead to a higher rate of response and extended duration of response. Further, disease reduction prior to Combotox, might lead to cures as it did in the trial in paediatric ALL (Herrera et al 2009).

Severe adverse events were observed in this study. VLS was the DLT. However, the patient who experienced this had been very heavily pretreated and had a history of VLS in the past secondary to exposure to high dose chemotherapy. He had also received a prior allogeneic transplant, which can lead to prolonged vascular endothelial damage (Schindler et al 2001). Nevertheless, Combotox could not be completely excluded as a contributing factor and should be used with caution in patients with preexisting vascular compromise.

We observed striking reductions in blast counts with Combotox administration. These reductions were rapid and were restricted to only the blast population, demonstrating specific cytotoxicity. The patients with high peripheral blasts counts also had reduced serum levels and reduced the half-life of Combotox, again demonstrating that Combotox is able to bind to the malignant cells. The often rapid rebound in peripheral blasts following the 3^rd dose, suggests that continued dosing with a reduced dose following the loss of peripheral blasts might lead to loss of blasts in the bone marrow and more durable remissions. These data also suggest that Combotox could be combined with chemotherapy in future studies in the setting of relapsed refractory disease or minimal residual disease (MRD). The latter would allow lower doses, as there would not be an antigen sink, and might lead to a meaningful reduction in toxicity and widening of the therapeutic window.

The C_max levels in the two patients with severe ADRs (Patients 12 and 17), were 1,451 ng/ml and 1,333 ng/ml. Levels above 1,000 ng/ml have been associated with severe toxicity in prior studies in patients with non-Hodgkin lymphoma (NHL) (Stone et al 1996). Reduction of serum levels to well below 1,000 ng/ml in some patients with a high number of circulating blasts appears to have reduced the rate of severe toxicity at these dose levels, as reported previously (Schindler et al 2000). Levels above 1,000 ng/ml, as seen in Patient 15, do not always lead to severe toxicity, so there are probably other predisposing factors involved. For example, elevated LFTs, while not seen in prior trials with these immunotoxins, were seen in two patients, both of whom had experienced liver toxicity with prior chemotherapy.

Newer data have shown that leukaemic stem cells in ALL are CD19+ and can persist and result in relapse (Hong, et al 2008). A recent study also demonstrated clinical efficacy of another anti-CD19 antibody in the MRD setting (Topp et al 2009). Thus, using Combotox in this setting might be useful in eliminating the residual non-dividing stem cells.

In summary, Combotox demonstrated single-agent activity in a high percentage of very heavily pre-treated patients with multiply relapsed B-lineage-ALL. The immunogenicity of Combotox was not clinically significant, and occurred at about half the rate observed in adult patients with NHL. The MTD was 7 mg/m² × 3. Grade 3 elevation of the LFTs occurred. These have not been observed previously in patients treated with Combotox.