A Phase I Clinical Trial of CHT-25 a ¹³¹I-Labeled Chimeric Anti-CD25 Antibody Showing Efficacy in Patients with Refractory Lymphoma

In spite of recent advances in therapy, there remains an acute need for improved treatments for Hodgkin lymphoma (HL) and other lymphomas. For instance, there is only a 50% prospect of long-term disease free survival for patients with relapsed HL responding to high-dose chemotherapy and autologous stem cell transplant (1, 2). The interleukin-2 (IL-2) receptor is selectively overexpressed in HL and other lymphomas and therefore is a potential target for new therapies. Radioimmunotherapy (RIT) with CHT-25 [Iodine-131 (¹³¹I)–labeled chimeric antibody to the IL-2 receptor (CD25)] has been developed for this purpose and a phase I study in relapsed HL and T-cell lymphoma is reported.

CD25 is a 55-kDa³ protein with three transmembrane protein chains that is induced on T-cell activation but is not found on B or T lymphocytes or monocytes in the resting state (3, 4). Hodgkin/Reed-Sternberg cells frequently express CD25, and the well vascularized nature of lymphoma tissue makes it accessible to i.v. administered antibodies (5, 6).

Although therapy with unlabeled antibody is effective for many B-cell lymphomas, (7, 8), it has shown limited activity in HL (9) and only short-term benefit in human T-cell lymphotrophic virus–associated lymphoma/leukemia where the IL-2 receptor is part of an important growth pathway (10). The problem of low potency can be addressed by conjugation of antibody to immunotoxins (11, 12) or by radiolabeling with a radionuclide such as ¹³¹I or Yttrium-90 (⁹⁰Y; ref. 13).

Repeated treatment is often compromised with murine antibodies due to their immunogenicity(14, 15). We have addressed this by using a chimeric antibody with mouse variable regions (murine RFT5 cell antibody) and human constant regions with proven low immunogenicity (16–18). The antibody binds with an affinity approximating that of IL-2 itself (17) and in its unlabeled form is licensed for the prevention of acute cellular rejection in renal allografts (18). It is known to have a longer blood half-life than the murine antibodies directed against the same epitope (17) and differs in this respect from the murine antibodies ibritumomab and tositumomab also used in RIT (19–21).

This article is believed to be the first to describe a clinical trial investigating the safety and efficacy of RIT with CHT-25 in refractory CD25-positive HL and other CD25-positive lymphomas.

The study was approved by the local Research Ethics Committee and the U.K. Medicines and Healthcare Products Regulatory Agency. All patients provided written informed consent. Patients with histologically confirmed lymphoma expressing CD25 on >50% of tumor cells (or the surrounding tumor milieu in HL) from a biopsy at diagnosis or relapse were eligible. Other inclusion criteria included the following: lymphoma refractory to prior therapy or relapsed after standard chemotherapy, age 18 y or older, measurable disease by computed tomography (CT), life expectancy of ≥3 mo, no serious comorbidities including HIV infection, WHO performance status of ≤2, bone marrow examination within 4 wk showing ≤25% involvement by tumor with adequate normal hematopoiesis, glomerular filtration rate of >50 mL/min, left ventricular ejection fraction on a multiple gated acquisition scan of >50%, absolute neutrophil count of ≥1.5 × 10⁹/liter hemoglobin of ≥10 g/dl, platelet count of ≥100 × 10⁹/liter, plasma creatinine mol/L, aspartate aminotransferase/alanine aminotransferase of ≤150 mol/L or EDTA clearance ≥50 mL/min, plasma bilirubin of ≤30 2× upper limit of normal, international normalized ratio of prothrombin time (INR) of ≤1.5, and normal thyroid function or stable on treatment. All patients were human anti-mouse antibody negative before therapy and any chemotherapy or radiotherapy was completed at least 4 wk before CHT-25 administration; corticosteroids were permitted as long as the dose was unchanged in the previous 4 wk.

This trial was an open-label, nonrandomized, phase I, multiple dose escalation study. Primary objectives were to determine the maximum tolerated dose (MTD), dose-limiting toxicity (DLT) and pharmacokinetic analysis. Secondary objectives were tumor response assessment and immunogenicity.

An accelerated titration design was used initially with stem cell support offered at doses that may be myeloablative (≥1,480 MBq/m²). A 100% dose escalation was used initially with three patients per cohort until 1,480 MBq/m² with the next increment at 50% and the final at 30% increase. The dose groups were as follows: 370, 740, 1,480, 2,220, and 2,960 MBq/m². The starting dose of 370 Mb/m² was chosen based on medical internal radiation dose (MIRD) dosimetry calculations from a previous study suggesting an absence of significant toxicity with a chimeric antibody. In the initial, possibly myeloablative, dose escalation phase, patients could be retreated if all drug-related toxicities (excluding hematological) had resolved to grade 2 or less. In the MTD finding phase, all toxicity (including hematological) had to recover to grade 2 or less, with a neutrophil count of >1.0 × 10⁹/L and a platelet count of >100 × 10⁹/L.

Escalation was stopped after the first DLT occurred at 2,960 MBq/m², which led to the only treatment-related death. Evidence of response had already been shown above 740 MBq/m² with stem cell support only required above 1,480 MBq/m². A further cohort was therefore included midway between 740 and 1,480 MBq/m² (at 1,200 MBq/m²) to exploit this combination of response together with no requirement for stem cell storage. The number of cycles was then capped at two. The entire dose schedule is represented in Fig. 1A and B.

The MTD was defined as the level in which a DLT occurred in two of three or two of six patients. DLT was defined as any grade 3 or 4 nonhematologic toxicity according to National Cancer Institute criteria version 2.0 excluding nausea, vomiting, or diarrhea in those not receiving adequate premedication. Hematologic toxicity for a DLT included grade IV neutropenia with failure to respond in 10 d to granulocyte colony-stimulating factor, or any duration associated with fever, infection, or thrombocytopenia grade IV, or associated with bleeding or requiring platelet transfusion. Dose escalation to the next level was permitted with a minimum time limit of 4 wk between treatments, with no DLT and evidence of tumor localization (>3% injected radioactivity/kg). The blood count and biochemistry were monitored twice weekly, whereas performance status and toxicity were monitored weekly. Baseline investigations were repeated at day 25 and 53 including tumor response assessment.

CHT-25 was provided by Novartis and supplied as 10-mg sterile vials lyophilized for reconstitution. ¹³¹I was supplied by MDS Nordion or Amersham. Reconstitution of antibody was done during radiolabeling using the N-Bromo-succinamide/L-tyrosine technique (22). CHT-25 was administered at a fixed 10-mg dose with a minimum of 8 wk of follow-up, or until toxicity resolved or another anticancer therapy instigated. CHT25 was administered initially at a rate of 6 mg/h, increasing gradually to 24 mg/h over 40 min. 0.9% NaCl was infused concurrently (150 mL/h) with full resuscitation equipment available. An Administration of Radioactive Substances Advisory Committee license was obtained.

Human antimurine antibody levels were assessed by ELISA prestudy and on days 43 and 57 with RFT5 and the nonspecific isotype matched antibody RDFR2 as positive and negative controls, respectively.

Blood (2.5 mL) was taken in an EDTA blood tube at 1, 3, 6, and 24 h postinfusion then on day 2, 3, 6, and when possible, 9. The samples were stored in an approved radioactive storage area until they had decayed to appropriate levels for analysis. Radioactivity was measured using a Packard Cobra 11th Series Auto γ counter. Monoexponential or biexponential decay curves as appropriate were modeled to the data.

SPECT images were acquired at 4, 24, 48, 72, and 96 h (when possible), for biodistribution, using an ADAC Vertex Plus dual-headed γ camera and reconstructed iteratively (OSEM algorithm; ref. 23) applying scatter and attenuation corrections. Regions of interest were drawn for tumor and organs of interest (heart, lung, liver, spleen, and kidneys) and average radioactivity uptake was then calculated. Olinda/XEM software (24) was used to calculate radiation doses to organs, red marrow (25), and tumor from the radioactivity uptake within the regions of interest, in blood and from the whole body dose-rate data.

Levels of the IL-2 receptor in serum were analyzed by ELISA using the Cellfree Human sIL-2R ELISA kit and a standard methodology.

Response to treatment was assessed using the updated International Working group Response Criteria (26). An ¹⁸F-fluorodeoxyglucose positron emission tomography (PET) scan was done using the hybrid SPECT/PET ADAC Vertex coincidence imaging system (Phillips-ADAC) or GE Discovery LS PET-CT (GE Healthcare) within 2 wk of study entry and then repeated at day 29 and at day 57. Images were reported from the screen of the camera-dedicated computer by a trained nuclear medicine physician (JB) blinded to patient identity and scan order.

Hematologic toxicity measures were assessed using the blood counts as continuous variables as the outcome measures as well as categorization through the National Cancer Institute Common Toxicity Criteria standard toxicity grading (version 2.0). The clinical characteristics were summarized by simple descriptive summary statistics.

Fifteen patients (Table 1) received one or more treatments; they all had >50% of HL cells or T cells (either malignant or surrounding milieu in HL) positive for CD25. The median number of prior chemotherapy regimens was 4 (range, 1-8); 9 had undergone autologous stem cell transplant. One patient was excluded from analysis because a second biopsy during the study showed the presence of second (B cell) lymphoma that was CD25 negative. Two patients were on long-term steroids (5 and 20 mg prednisolone). Both patients had developed progressive lymphoma on that dose, which remained unchanged throughout the study.

Toxicity is summarized in Tables 2 and 3. There was no significant infusion-related toxicity. The principal toxicity was hematologic and developed progressively as administered activity was escalated up to 2,960 MBq/m². The patient treated at this dose had previously received eight lines of therapy and had prolonged myelosuppression after CHT-25. Despite stem cell support, he died at day 36 of neutropenic sepsis and presumed Pneumocystis jiroveci pneumonia. Before treatment, seven patients were lymphopenic (3 grade 2 and 3 grade 3) and counts fell with therapy. On de-escalation to 1,480 MBq/m², grade 4 hematologic toxicity occurred (Table 2) with a median platelet nadir of 15 × 10⁹/L (range, 8-150) at a median of 41 days (range, 31-56) and neutropenia (median nadir, 0.8 × 10⁹/L; range, 0.1-7.8) after a median of 46 days (range, 28-54). A further three patients were treated at 1,200 Mbq/m² with acceptable toxicity (Table 2), and this was determined as the MTD.

Blood transfusion was used to maintain the hemoglobin between 12-14 g/L for 4 weeks posttherapy to maximize radiation effect. Granulocyte colony-stimulating factor was given to all patients with grade 4 neutropenia until count recovery. If the neutrophil count remained ≤0.2 × 10⁹/L for 2 consecutive days, stem cells were reinfused. Two patients had their stem cells infused, both while suffering from neutropenic sepsis (patient 7 and patient 14 treated at 2,960 MBq/m² and 1,480 MBq/m², respectively). Six patients required platelet transfusion at or above the MTD of whom five had platelet counts of <10 × 10⁹/L. Between a platelet count of 10 to 20 × 10⁹/L, administration was at the investigators discretion based on clinical need; below 10 × 10⁹/L platelets were given immediately due to risk of spontaneous hemorrhage. All patients receiving platelet transfusions were classified as having grade 4 toxicity. Septic events are included (Table 3) regardless of presumed causation, due to any possible immunomodulatory effect of CHT-25. There was no evidence of viral reactivation.

There were 14 hospital admissions in nine patients. Of these, six were due to infection (four neutropenic sepsis or Hickman line infection with neutropenia), four were associated with progressive disease (one pain, one renal failure secondary to obstruction, two general deterioration), one was for hemoptysis secondary to lymphoma in the lung with a normal platelet count, one was for 24-h observation after head injury when thrombocytopenic, and two patients for blood transfusion and/or platelet transfusion for practical reasons.

Table 4 shows the pharmacokinetics for 22 treatments. Of these, 13 can be modeled using a biexponential curve. However, nine patients seem to follow a monoexponential decay: the initial, relatively rapid blood clearance might have been masked by the limited number of data points in the first 6 hours. The 50% and 90% median clearance was calculated to be 44 and 193.5 hours, respectively, consistent with the long half-life of the chimeric antibody.

The median pre treatment soluble IL-2R (sIL2R) was 2,803.3 U/mL (range, 580-66,080 U/mL). This compares with median normal values for the assay of 521 U/mL (range, 269-1,116 U/mL). The level of the pretreatment sIL2R was not found to be a prognostic factor for survival on this small patient sample with patients above 2,803 U/mL having a median survival of 407 days compared with 485 days for those below. (Kaplan-Meier method, Mantel-Cox x², 0.17; P = 0.68). The molar ratio between the CHT25 antibody and sIL2R levels in plasma at the moment postinjection (plasma volume from standard man) was calculated to be a median of 49 to 1 (with a range from 2:1-260:1). There was no clear relationship in this small sample between the ratio and response.

Three patients received three treatments, four had two treatments and seven had one treatment. One patient became HACA response positive (patient 1) at 45 days after the first treatment. There was rapid blood clearance of CHT-25 in patient 3 without positive HACA and this was not explained. Patient 9 was nonspecifically positive in the HACA assay to both irrelevant control antibody and PBS at baseline; values fell with therapy suggesting nonspecific binding. One patient was screened for entry to the trial after an allograft but was ineligible due to a positive HACA possibly representing an immune antitumor response.

Data were collected on 11 patients. Figure 2A shows median values and ranges of the percent of injected radioactivity per kilogram in normal organs and tumor with wide interpatient variation. While radioactivity declined from 4 hours in normal tissues, it increased up to 24 hours in the tumor with retention to 96 hours Comparing tumor to blood ratio 4 hours after administration with those after 96 hours, there was an increase in all seven cases with sufficient data for this analysis by a median of 2.2-fold (range, 1.3:1-4.0:1; P = 0.005 by the Wilcoxon Signed Ranks test). Figure 2B shows the absorbed radiation dose in normal tissues (Olinda/XEM software; ref. 24). Figure 2C shows the absorbed radiation dose in tumors of individual patients (Olinda/XEM). Tumor values are not comparable with those for normal tissues (Fig. 2B) because this software does not include the contribution to radiation dose from adjacent tissues as it does for normal tissues. Figure 2D shows the relationship between the absorbed bone marrow dose per unit of injected activity and bone marrow toxicity (Spearman rank value r = −0.66). In addition, a relationship was seen between absorbed tumor dose and response in seven patients. Nonresponders had tumor doses of ≤1,325 mGy and responders had doses of ≥1,990 mGy with no crossover. (Mann-Whitney P = 0.057).

Objective responses were observed in an administered activity-dependent fashion (Table 5). No responses were seen at 370 MBq/m² and one at 740 MBq/m². Six of nine patients treated at a single administration of 1,200 MBq/m² or higher had a response [three complete responses (CR; two HL, and one AITCL) and three partial response (PR; all HL); Fig. 3]. The two patients with adult T-cell leukemia/lymphoma had ≤740 MBq/m² and one had a CR.

Four patients (patient 10, 11, 12, and 14) considered ineligible for allogeneic or autologous transplant because of chemoresistant disease were treated. Three responded; two proceeded with transplant, one entered CR comorbidity prevented transplant.

The median overall survival was 453 days (range, 44-925). Two patients remain alive at the point of data analysis. The progression-free survival of patients who had a response or stable disease was a median of 179 days (range 72-925+). The two patients transplanted post–CHT-25 were excluded from progression-free survival analysis.

There is limited experience with chimeric or human antibodies in the RIT of HL or T-cell lymphomas. A phase I study of a chimeric antibody demonstrating favorable findings at the MTD is reported. It is relevant when considering toxicity and response that the patients had poor prognostic features including multiple lines of prior myelosuppressive therapy. All tumors were positive for CD25 on immunohistochemistry; with a median duration between CD25 staining and treatment of 384 days (range, 39-2,722 days), dedifferentiation and a loss of CD25 status cannot be excluded. Good localization of CHT25 by SPECT imaging suggests that dedifferentiation was not an issue.

Dose-limiting toxicity was myelosuppression, which was readily managed at the MTD. The only treatment-related death occurred in a patient who received more than twice this dose. The rate of non-neutropenic infection was relatively high; this may have been related to the use of long-term central venous access devices. Modulation of CD25-positive immune cells by therapy and other immune defects consequent on previous therapies cannot be excluded. There was no evidence of viral reactivation, but given the associated lymphopenia, appropriate microbial prophylaxis for lymphopenic patients may be warranted. Repeated therapy was feasible after recovery from hematologic toxicity and further chemotherapy was not precluded in patients whose tumor did not respond.

Response assessed using ¹⁸F-fluorodeoxyglucose PET and CT (26–28) was encouraging. Of nine patients treated at or above the MTD, the overall response rate was 67% (three CR and three PR). Two further patients had stable disease, whereas one had progression. The two patients on stable doses of prednisolone responded to CHT25 having previously progressed while taking this steroid dose. The study was not designed to assess progression-free survival or overall survival, and the figures presented in the results section were derived from routine clinical follow-up. It is evident that long-term responses and survival can be achieved.

Pharmacokinetics and dosimetry showed that the chimeric antibody delivered radiation to normal tissues over a more prolonged time than previously reported in a murine CD25 antibody, although with wide variability (⁹⁰Y-antiTac; ref. 29). To exploit this, ¹³¹I with its physical half-life of 8 days was chosen as the therapeutic radionuclide. One advantage of using ¹³¹I is the mean path length (0.4 mm) of the radionuclide, which allows for the eradication of adjacent antigen-negative tumor cells. The longer path length of ⁹⁰Y means a greater proportion of the dose is deposited outside small tumor masses and its shorter half-life means that the prolonged retention of antibody in tumor cannot be exploited. However, there is potential for investigation of other β-emitting radionuclides.

Serum levels of sIL2R were elevated in all patients but antibody was calculated to remain in excess at the time of administration. Because satisfactory tumor targeting was shown and tumor response rate was high, it is concluded that the antibody/antigen ratios were satisfactory for effective therapy. However, it remains possible that further improvements in antibody delivery could be achieved by reducing sIL2R levels before therapy.

Previous studies have shown that an elevated sIL2R level is an adverse prognostic factor in HL (30) and decreases in responding disease. This was not confirmed in the present study but the sample size was small.

There was a correlation between red marrow–absorbed dose and marrow toxicity, although variation in the extent of prior bone marrow damage is a separate parameter and was not considered. Flt-3 levels have been reported as a means of addressing this issue (31) and would be worthy of inclusion in future trials. A clear dose response relationship was not seen in this small study; however, tumor responses were only recorded in patients with absorbed doses of ≥1,990 mGy. One of the objectives of the planned phase II study will be to determine whether this is a consistent finding so that pretreatment dosimetry (including blood pharmacokinetics) could be used to select appropriate patients for therapy.

Although the radiation dose seems important for tumor response, an additional cytotoxic effect may be produced by the antibody. Recent work has considered a possible anticancer effect from the depletion of regulatory T cells (32) or through other immune effects. Currently evidence of a direct anticancer-immune mediated antibody effect is inconclusive. Of note, unlabeled murine antibody without a radioconjugate showed only limited benefit with responses of a short-lived duration (10). For this reason, CHT25 was not trialed in its unlabeled form as an anticancer agent. There is no preclinical, anticancer data for CHT25 due to a lack of a suitable animal model; although rhesus monkeys share the same IL-2R epitope as humans, they are not suitable to be used in therapy studies.

Previously phase I/II studies in HL using RIT have combined ¹³¹I with polyclonal anti-ferritin or anti-CD30–directed antibodies (33, 34). The overall response rates have ranged from 36% to 49% and the major side effect, as observed here, was hematologic toxicity, most notably thrombocytopenia. ¹³¹I-Rituximab has shown good effect in 91 patients with relapsed or refractory NHL achieving an overall response rate of 76% with 53% attaining either a CR or a CR, unconfirmed (Cru). These results are similar to the two licensed murine antibodies; the degree of myelosuppression was similar degree even with the longer circulating time of the chimeric antibody (35).

CHT25 has a potential application in preparation for autologous and allogeneic transplantation, giving the possibility of long-term disease control to a proportion of patients with relapsed or refractory disease. Demonstrable chemosensitivity and maximal reduction in disease burden before transplantation are critical predictors of transplant outcome in HL (2, 36). In this study, we induced PET responses and transplanted patients previously considered inappropriate for this treatment. This justifies further investigation of CHT25 either to render refractory HL patients eligible for transplant or to improve transplant outcome when combining it with standard conditioning regimens. It may also have a place in the early treatment of poor prognosis patients and other CD25-positive lymphomas as it is well known that a range of hematologic malignancies express CD25 (37–41). In our study, we observed a CR in a patient with a T-cell lymphoma, justifying further studies to look at this patient group.

In summary, the CHT25 study shows that RIT with a chimeric antibody can be safely administered at a dose of 1,200 MBq/m², giving objective responses in a high proportion of patients with lymphomas expressing CD25. Pharmacokinetic and pharmacodynamic parameters relevant to producing this effect have been shown. CHT25 seems to have potential as a new treatment in this patient group and is shortly to proceed to further evaluation of efficacy in phase II studies.

CHT25 is described in U.S. Patent 6,383,487; held by P. Amlot, A. Akbar, G. Heinrich, and S. Cammisuli. P. Amlot is coauthor of U.S. Patent 6,383,487. The other authors declare they have no conflict of interest.

We thank Peter Isaacson for assessment of CD25 staining, Andrew Hilson for image reading and administrative support, Ama Rohatiner, Jamie Cavenagh, Ben Mead, Peter Johnson, Kim Orchard, Alan Horwich, Anthony Goldstone, and the hematology department at the Royal Free Hospital for support of the study.