Tumor Regression in Cancer Patients by Very Low Doses of a T Cell–Engaging Antibody

See allHide authors and affiliations

Science  15 Aug 2008:
Vol. 321, Issue 5891, pp. 974-977
DOI: 10.1126/science.1158545


Previous attempts have shown the potential of T cells in immunotherapy of cancer. Here, we report on the clinical activity of a bispecific antibody construct called blinatumomab, which has the potential to engage all cytotoxic T cells in patients for lysis of cancer cells. Doses as low as 0.005 milligrams per square meter per day in non–Hodgkin's lymphoma patients led to an elimination of target cells in blood. Partial and complete tumor regressions were first observed at a dose level of 0.015 milligrams, and all seven patients treated at a dose level of 0.06 milligrams experienced a tumor regression. Blinatumomab also led to clearance of tumor cells from bone marrow and liver. T cell–engaging antibodies appear to have therapeutic potential for the treatment of malignant diseases.

Substantial evidence suggests that cytotoxic T lymphocytes participate in controlling tumor growth and that they can be harnessed in immunotherapeutic settings. For example, their presence in colorectal tumors was shown to strongly predict clinical outcome (1), and adoptive transfer of ex vivo expanded tumor-derived lymphocytes (2) or of T cell receptor gene transfected T cells to melanoma patients (3) has been shown to lead to cancer regression. However, therapies aimed at inducing specific T cell responses against cancer cells, including vaccination (4) and CTLA-4–blocking antibodies (5), have been limited by relatively low response rates and relapses, presumably because of immune escape mechanisms of cancer cells (6, 7).

An alternative approach to harness the high cytotoxic potential of T cells has been the use of T cell–engaging antibodies (8, 9). Conventional antibodies are not able to directly recruit T cells because these cells lack Fcγ receptors. Well-studied antibody constructs for engaging T cells are bispecific T cell engagers (BiTE), which are based on single-chain antibodies (10, 11). These molecules can transiently tether resting T cells to tumor cells, leading to concomitant T cell activation and serial lysis of tumor cells (1214). To date, various mouse models have demonstrated high levels of activity of BiTE antibodies (1520). The independence of this approach from peptide antigen presentation and tumor-specific T cell clones suggests that it can overcome major immune escape mechanisms.

We have investigated clinical activity and safety of increasing doses of BiTE antibody blinatumomab (also called MT103/MEDI-538) with dual specificity for CD19 and CD3 (2124) in an ongoing study with non–Hodgkin's B cell lymphoma (NHL) patients relapsed to standard therapies. Demographics, staging, number of prior therapies, treatment details, and clinical outcome of all 38 patients from the currently completed dose levels are shown in table S1 (25). At dose levels of 0.005 mg/m2 per day and higher, CD19+ target cells rapidly vanished from peripheral blood of patients for the rest of the treatment period (table S1). Figure 1 shows representative pharmacodynamic data for patient 28, who had a complete and lasting tumor regression at a dose level of 0.03 mg/m2 per day. The decline of CD19+ cells (Fig. 1A) was accompanied by expression of apoptosis marker annexin V (Fig. 1B), indicating lysis rather than extravasation of target cells. Peripheral T cells initially disappeared but then reappeared, with counts and kinetics showing a high interpatient variability (Fig. 1C and fig. S1). In this and several other patients treated at higher dose levels, peripheral T cell numbers increased severalfold over baseline, which coincided with expression of activation markers CD69, CD25, and HLA-DR by a significant proportion of CD8+ T cells (Fig. 1D). During or after treatment, total T cell counts either contracted to baseline or remained at a higher level than before treatment (Fig. 1C). The increase in T cell counts under blinatumomab treatment was predominantly caused by an expansion of effector memory CD8+ and CD4+ T cells with CD45RA/CCR7 phenotype, whereas counts of CD8+ T cells with phenotypes of naïve, central memory, and CD45RA+ effector memory T cells remained more or less constant (Fig. 1E). The early disappearance of T cells may be explained by a transient increase in the adhesiveness of T cells to vessels and/or extravasation, whereas their expansion may be due to proliferation of effector memory T cells.

Fig. 1.

Effect of T cell–engaging antibody blinatumomab on B and T cells in peripheral blood of a responding NHL patient. Analyzed blood samples were derived from patient 28 showing a complete response at 30 μg of blinatumomab/m2 per day. Total cell counts per μl and expression of cell surface markers were determined by flow cytometry (25). The first data point shows the baseline value immediately before start of treatment. (A) The effect of blinatumomab treatment on counts of CD19+ B and lymphoma cells in peripheral blood. The percentage of CD5+ lymphoma cells before treatment was 44%. (B) The effect of blinatumomab treatment on surface expression of apoptosis marker annexin V during the decay phase of B cell counts. MFI indicates mean fluorescence intensity. (C) The effect of blinatumomab treatment on total counts of CD4+ and CD8+ T cells. (D) The effect of blinatumomab treatment on activation markers CD69, CD25, and HLA-DR expressed on peripheral CD8+ T cells. (E) The effect of blinatumomab treatment on four distinct CD8+ T cell subpopulations. TCM, central memory T cells; TEM, effector memory T cells; and TEMRA, CD45RA+ effector memory T cells. For experimental details, see (24).

Among 38 patients receiving blinatumomab at doses from 0.0005 to 0.06 mg/m2 per day, a total of 11 major responses have been observed (Table 1 and table S1), which were centrally confirmed according to defined criteria (26). All four complete (CR) and seven partial tumor regressions (PR) occurred at doses of 0.015 mg/m2 per day and higher, indicating a dose response relationship (Table 1). All seven patients treated at the recently completed dose level of 0.06 mg/m2 showed objective responses. Tumor regressions were observed in patients with follicular lymphoma, mantle cell lymphoma, and chronic lymphocytic leukemia. Examples of a complete tumor regression in the pelvic area of patient 16 (Fig. 2A) and of a partial regression in the chest area of patient 15 (Fig. 2B) are shown. In all responding patients, the majority of tumor shrinkage occurred within the first 4 weeks of treatment. Longest duration of a CR currently is 13 months in a patient with mantle cell lymphoma, and three more patients have ongoing regressions lasting for >6 months (table S1). No relapse has thus far been observed in responding patients treated with blinatumomab at dose levels of 0.03 and 0.06 mg/m2 per day. Blinatumomab doses of 0.015 mg/m2 per day and higher led to depletion of tumor cells not only in blood (Fig. 1A), lymph node lesions (Fig. 2, A and B), and spleen (Fig. 2C) but also in bone marrow (Fig. 2D) and, in one case, in infiltrated liver (Fig. 2E). In 9 of 11 cases of bone marrow infiltration, immunohistochemical and flow cytometric analysis of patient biopsies revealed either complete (6/11) or partial elimination of tumor cells from bone marrow (3/11) (table S1).

Fig. 2.

Antitumor activities of T cell–engaging antibody blinatumomab. (A) Computer tomography (CT) images of patient 16 (0.015 mg of blinatumomab/m2 per day) before and after 8 weeks of treatment. The response of this patient was rated as complete. Arrows indicate two lymph node tumors in the pelvic area. (B) CT scans of patient 15 (0.015 mg of blinatumomab/m2 per day) before and 4 weeks after treatment. The response of this patient was rated partial. Arrows indicate lymph node tumors in the mediastinum. (C) CT scans of the abdomen of patient 16 (0.015 mg of blinatumomab/m2 per day) before and 4 weeks after treatment. S denotes splenomegaly. (D) Micrographs of marrow biopsies of patient 13 at baseline and 15 days after start of treatment with 0.015 mg of blinatumomab/m2 per day, Blue indicates tumor cells (hematoxylin staining); brown, T cells (CD3 staining). (E) Micrographs of liver biopsies from patient 33 (0.06 mg of blinatumomab/m2 per day) stained for B cell antigen CD20 at baseline and after 4 weeks of treatment. Representative sections of biopsies are shown.

Table 1.

Relationship between dose level and antitumor activity of BiTE antibody blinatumomab in relapsed NHL patients. Abbreviations used are FL, follicular lymphoma; MCL, mantle cell lymphoma; CLL, chronic lymphocytic leukaemia. For details of trial design, see (25).

View this table:

Adverse events thus far recorded in the trial up to the completed dose level of 0.06 mg/m2 per day are listed by their incidence and grade in table S2 (any event occurring in 3 patients) and table S3 (grade 3 and 4 events occurring in 1 patient). Most frequent side effects included pyrexia, lympho- and leukopenia, chills, and increase of C reactive protein. The majority of adverse events occurred within the first week of treatment and usually normalized to grade 1 or lower during further treatment. Symptoms of the central nervous system—including disorientation, confusion, speech disorders, tremor, and convulsions—have been observed, all of which were fully reversible during or shortly after discontinuation of treatment. All adverse events that led to a discontinuation of treatment with blinatumomab are described in detail in (25). No clinically apparent cytokine-release syndrome was evident in any patient. Unlike with antibodies blocking CTLA-4, no autoimmune phenomena have thus far been observed in patients treated with blinatumomab. Antibodies binding or neutralizing blinatumomab have not yet been detected in patients.

First confirmed responses to blinatumomab as a single agent occurred in NHL patients at a serum level of 0.6 ng/ml (at 0.015 mg/m2 per day). This is about 5 orders of magnitude below serum levels reported for the monoclonal antibody rituximab at a dose of 375 mg/m2 per week, which likewise elicits objective responses as single agent in this disease (27). The enormous potency difference between BiTE antibody blinatumomab and a conventional antibody may relate to the high lytic potential of cytotoxic T cells, which are activated by engagement of only very few CD3 receptor subunits, can rapidly adopt a serial lysis mode, and proliferate at the site of their activation. An ongoing clinical phase II trial is investigating the activity of blinatumomab in patients with acute lymphoblastic leukaemia. It will also be interesting to investigate whether BiTE antibodies are active in solid tumor indications. A clinical trial with a BiTE antibody specific for EpCAM (18), a target antigen widely expressed on human adenocarcinoma (28) and on cancer stem cells (29), has recently been initiated.

Supporting Online Material

Materials and Methods

SOM Text

Fig. S1

Tables S1 to S3

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article