Chronic lymphocytic leukemia (CLL) represents the most prevalent adult leukemia and is characterized by the accumulation of long-lived B lymphocytes. Despite significant advances in its therapy especially during the last two decades, the disease still remains incurable and new treatment options need to be developed. Among targeted approaches, immunotherapy comprises Fc-mediated mechanisms that are independent of the mechanisms exploited by chemotherapy or modulation of signal transduction by small molecule drugs.
Currently the CD20 and CD52 antibodies, rituximab and alemtuzumab are clinically used for treating CLL in combination with chemotherapy. The addition of rituximab to a fludarabine/cyclophosphamide chemotherapy regimen led to a longer overall survival of the chemoimmunotherapy-treated patients defining the current treatment standard.1 Although the efficiency of rituximab is limited by relatively low and variable expression of CD20 on CLL cells and poor response in the patient subgroup with chromosome 17p deletion, alemtuzumab additionally depletes T lymphocytes, which leads to an increased risk of infection. Therefore novel monoclonal antibodies directed to CD20 and additional antigens are being developed for immunotherapy of CLL.2
The target antigen of rituximab is human CD20 (membrane spanning four domains, subfamily A, member 1; MS4A1), a common cell surface antigen of all B cells except stem or plasma cells. Like CD20, the tetraspanin glycoprotein CD37 is strongly expressed on the surface of B lymphocytes and on mature B-cell leukemia and lymphoma cells, but with higher expression on CLL cells than on normal peripheral B lymphocytes. Although the structures of both protein families cross the cytoplasma membrane four times and have a similar size of about 200 amino acids, CD20 and CD37 do not share any sequence homology and known function (Supplementary Material). As the large extracellular loop of CD20, which consists of 44 amino acids, provides distinct epitopes for several therapeutic antibodies, including rituximab, ofatumumab and GA101, an even greater variety of potential epitopes can be expected in the EC2 loop of CD37 comprising the triple number of amino acids.
Immunoreagents targeting CD37 with potential for treating CLL as single agents or in combination are being developed. Among these, the small modular immunopharmaceutical TRU-016 was assessed for use in B-cell malignancies. It induced caspase-independent direct cell death (DCD) in primary CLL cells and mediated antibody-dependent cell-mediated cytotoxicity (ADCC).3 In the meantime, TRU-016 has been examined in a clinical phase I trial.4 As additional immunoreagents targeting CD37 monoclonal antibody (mAb) 37.1 and mAb 37.2 were developed.5 Like rituximab, mAb 37.1 is a mouse–human chimeric antibody of IgG1 isotype, but additionally was Fc-engineered. mAb 37.2 is a humanized version derived from mAb 37.1, which binds CD37 with low nanomolar affinity. To prepare these novel CD37 antibodies for a recently started phase I trial of the German CLL study group, we investigated them in comparison with monoclonal antibodies currently used for CLL therapy, that is, rituximab and alemtuzumab, with regard to antibody-induced apoptosis in isolated CLL cells and to B-cell depletion from whole-blood samples as in our recent characterization of the novel CD20 antibody GA101.6
First, we studied the direct effects of monoclonal antibodies on freshly isolated CLL cells. As immunoreagents targeting CD20 and CD37 were found to trigger programmed cell death in isolated CLL cells in a caspase-independent manner,3, 7 we wish to refer to this type of antibody-induced cell killing as DCD, so as to distinguish it from classical, caspase-dependent, for example, DNA damage-induced apoptosis. Moreover this designation is meant to represent DCD as the counterpart of Fc-mediated killing mechanisms of antibodies. For measuring DCD induction, isolated CLL lymphocytes were incubated in cell culture medium containing 10% calf serum for 24 h, comparing the different antibody preparations at a fixed concentration of 10 μg/ml without addition of crosslinking anti-Fc antibody. Subsequently, the samples were examined flow cytometrically for annexin V-binding and 7-amino-actinomycin D (7AAD) staining as previously described.6 On the average, the percentages of cells showing phosphatidylserine (PS) exposure were significantly increased as compared with untreated samples by all four investigated antibody preparations (Supplementary Figure 1). Annexin V-binding alone was more efficiently induced than PS exposure and membrane disintegration together. Relative viability and DCD induction values were derived from these primary data as described,6 and are detailed for individual CLL samples (Supplementary Table 1). The mean DCD induction according to PS exposure in the complete set of samples was highest by mAb 37.1, with 51% followed by alemtuzumab and mAb 37.2 with 35 and 32%, respectively, and by far exceeding that by rituximab. When three samples, in which the percentage of cells with PS exposure exceeded 50%, were excluded from the analysis, the ranking and order of magnitude of DCD induction by the investigated antibodies was maintained (Figure 1a). Individual CLL samples were arranged according to ascending mAb 37.1-induced DCD, and are listed with associated clinical information and molecular features (Supplementary Table 1). The investigated samples included two with deletion of chromosome 17p (sample identification (ID) E and N), in which the CD37 antibodies efficiently induced DCD. This suggests the potential for high efficacy of these immunoreagents also in the high-risk group of CLL patients with p53 mutation or deletion, although two observations are scarce evidence. Examples using freshly isolated CLL samples did not show a pronounced dose-dependent increase of antibody-induced DCD in the concentration range 1–30 μg/ml.
The DCD induction by the reference antibodies, rituximab and alemtuzumab was largely in line with previous reports.6, 7 Although the marginal DCD induction by rituximab was expected, the average alemtuzumab-induced DCD without crosslinking anti-Fc antibodies and without elevated serum levels may be overestimated in this study owing to three particularly sensitive samples (ID E, J and N). In isolated CLL cells the humanized mAb 37.2 led to similar DCD induction as alemtuzumab, but did not affect T lymphocytes at the same time, thus avoiding a major drawback of alemtuzumab in a clinical setting. The observed exquisite DCD induction in freshly isolated CLL cells by the novel CD37 antibodies is of similar size as that reported for TRU016, a small modular immunopharmaceutical targeting CD37.3 Remarkably, the present results with mAb 37.1 and 37.2 were obtained without crosslinking by anti-human IgG in contrast to the DCD induction reported for TRU-016. In CLL cells DCD induction by the present CD37 antibodies, in particular by the mouse–human chimeric mAb 37.1, clearly exceeded that by rituximab, and apparently even surpassed the improved DCD induction by the glyco-engineered type II CD20 antibody GA101.6
In the case of rituximab, antibody-induced DCD constitutes only a minor fraction of overall antibody-induced cytotoxicity exerted on CLL cells.8 In order to comprise Fc-mediated mechanisms in our assessment in addition to DCD, we next investigated the depletion of CLL cells from whole-blood samples by treatment with CD37 antibodies and controls. The present B-cell depletion assay depends on the individual host-specific capacity of patient samples to mediate complement-dependent cytotoxicity or antibody-dependent cell-mediated cytotoxicity, and thus has potential for predicting clinical responses. For performing this assay we enumerated B and T lymphocytes in antibody-treated whole-blood samples and untreated controls, and calculated the antibody-induced B-cell depletion from the relative changes of B- to T-cell ratios. This was accomplished by three color FACS analysis after staining with differently fluorescence-labeled antibodies specific for the general, T- and B-lymphocyte antigens CD45, CD3 and CD19.6 Previously we calculated B-cell depletion by CD20 antibodies mainly via B- to T-cell ratios, because antibody-mediated depletion in vitro of T lymphocytes can be excluded due to complete absence of target antigen expression in this cell population. In contrast, low levels of CD37 were reported to be expressed and functionally important also on T lymphocytes.9 Therefore, we monitored B- and T- cell counts in six whole-blood samples from healthy donors by calibration with absolute counting beads with and without prior antibody treatment (Supplementary Figure 2). Although mAb 37.1 and mAb 37.2 led to sizable B-cell depletion also in normal blood samples, T-cell counts were not affected. In contrast, the CD52 antibody alemtuzumab considerably reduced B- and T-cell counts, in agreement with strong antigen expression on either cell type. Thus, among the antibodies tested, alemtuzumab was the only antibody to reduce T-cell counts. Lack of significant T-cell reduction by CD37 and CD20 antibodies, but not alemtuzumab was also confirmed in a set of CLL samples (numbers 3, 4, 8 and 11) that was evaluated according to absolute cell counts obtained by calibration with TruCount beads (BD Biosciences, Heidelberg, Germany). As T-cell counts remain unchanged by treatment with CD37 antibodies, they can serve as an internal standard in the assay evaluation. Thus the calculation of B-cell depletion via B- to T-cell ratios is adequately applicable also for the present CD37 antibodies. As CD52, the target antigen for alemtuzumab, is expressed on B and T cells alike, alemtuzumab was not included in the large series of experiments evaluated according to B- to T-cell ratios.
We started our assessment of antibody-induced B-cell depletion from whole blood by comparing mAb 37.1, mAb 37.2 and rituximab at a fixed concentration of 10 μg/ml (Figure 1b). The mean percentages of B-cell depletion after 24 h treatment with these antibodies were 27, 18 and 12%, respectively. The CD37 antibody mAb 37.1 led to significantly superior B-cell depletion than mAb 37.2 and rituximab. B-cell depletion by mAb 37.1 in individual samples ranged 5–85%, and in most samples surpassed the corresponding mAb 37.2 effects (Supplementary Table 1). Cytogenetic aberrations, for example, affecting the p53 and ATM genes, were assessed in some but not all patients. Therefore no conclusions about the efficiency of CD37 antibodies on CLL samples with specific cytogenetic abnormalities can be made. Stratification according to treatment status showed higher antibody-induced B-cell depletion in untreated as compared with previously treated patients (Figure 1c). This difference was marginally significant in the present set of eight samples in each group, but may have been reinforced by absence of samples with high ZAP70 expression among the treated and their over-representation among the untreated samples, as a trend for higher B-cell depletion in ZAP70-negative as compared with ZAP70-positive samples was also noted. In contrast, DCD induction by CD37 antibodies did not notably differ in smaller sets of samples from untreated or previously treated patients.
As a complement to the observed cytotoxic effects of therapeutic antibodies at a fixed concentration on CLL cells, we determined the relative expression levels of CD20 and CD37 in CLL samples (Figure 1d). As different labels were used for detecting CD20 and CD37 and because the degree of conjugation of the two antibody preparations used had not been ascertained to be equal and homogeneous, these measurements served to compare relative expression of CD20 and CD37 on the examined samples rather than providing absolute antigen densities. Despite these limitations, the CD37 expression on CLL samples surpassed that of CD20, as only the top quartile of CD20 signals, but all CD37 mean fluorescence intensities (MFI) values except the bottom quartile were above an arbitrary MFI threshold of 10 000. In addition, the differences in mean and median MFI after CD20 and CD37 staining were sufficiently big to suggest stronger surface expression of CD37 than of CD20 on CLL lymphocytes as previously observed.3 CD37 expression was not significantly correlated with CD20 expression (R2=0.07). Surface expression data of individual CLL samples are labeled with the sample IDs indicating their rank among the samples investigated for DCD induction or B-cell depletion by 10 μg/ml mAb 37.1 (Supplementary Table 1). Five samples with known antibody-induced DCD and CD37 expression (numbers 1, 2, 9, 10 and 11) covered the range of observed MFI values in CLL samples almost completely. Two samples representing the low and high extremes of observed CD37 MFI, both exhibited about 70% DCD induction by mAb 37.1. For all five investigated samples, CD37 expression and mAb 37.1-induced DCD were not correlated with each other (R2=0.24). Analysis of CD37 and C20 surface expression in cell lines was largely in agreement with published data for Mec1, Raji and Ramos cells,3 and showed extremely low CD37 expression in the cell line JVM-3. The number of samples with data on both, antibody effects and surface expression data, is limited and did not indicate any correlation of CD37 or CD20 surface expression and the corresponding antibody effects. In larger data sets, we found a similar lack of correlation of antigen surface expression with antibody-induced DCD for CD20 antibodies despite a significant correlation with antibody-induced B-cell depletion.6 At least for B-cell depletion a correlation with surface expression may therefore be expected, but particularly for small numbers may be obscured in the present assay from whole blood, for example, by variable human IgG levels and target to effector cell ratios.
As a further step in the assessment of antibody effects, the dose-dependency of the B-cell depletion induced by CD37 antibodies was investigated (Figure 2). Although dose-dependency in a concentration range 1–30 μg/ml was not prominent for antibody-induced DCD, CLL cell depletion from whole-blood samples increased with antibody concentrations (Figure 2a). The shape of the observed dose response curves is of saturation type, and suggests an influence on antibody effects of individually different human IgG levels in the assay matrix. Relatively high concentrations of therapeutic antibodies may be required to overcome inhibition of Fc-mediated antibody effects by unspecific IgG present in plasma. Probably for this reason the average CLL depletion by CD37 antibodies rose by more than 50%, when the antibody concentrations were increased from 10 to 30 μg/ml (Figure 2b). This dose-dependent enhancement of B-cell depletion by CD37 antibodies was highly significant in the examined set of whole-blood samples. In addition, the range of B-cell depletion response was narrower at higher than at lower antibody concentrations. For these reasons the potential efficacy of the investigated CD37 antibodies is underestimated, if only B-cell depletion at a concentration of 10 μg/ml is considered, as much higher plasma concentrations of monoclonal antibodies of up to 100 μg/μl can be pharmacologically achieved and lead to higher effects. Due to incomplete representation by a single treatment concentration, the actual dose-dependent antibody efficacy is higher than indicated by the B-cell depletion values observed with 10-μg antibody/ml and also may yield a different ranking.
Although the same treatment period and antibody concentration were applied for investigating DCD induction in isolated CLL cells and B-cell depletion from whole-blood samples, results of these assays cannot be directly compared with each other, for instance due to the different nature of endpoints used, that is, percentages of cells showing signs of early cell death versus decrease of absolute counts of intact cells. Moreover the assay matrix of the whole-blood assay comprises influences, which are not present in the DCD assay, for example, variable levels of human IgG and ratios of target to effector cells. As for antibody mechanisms, the present B-cell depletion assays were meant to capture overall antibody effects on individual blood samples and to comprise the respective status of anti-tumor immune function, rather than characterizing NK cell-mediated ADCC induced by CD37 antibodies using externally added peripheral blood mononuclear cells as previously reported.5 The remarkable DCD induction by CD37 antibodies appears to make a major contribution of DCD to the overall antibody-mediated elimination of tumor cells, as the apparent mean induction of PS exposure numerically even surpassed the observed B-cell depletion.
In conclusion, the present CD37 antibodies, especially mAb 37.1, show clearly superior in vitro effects on CLL cells in terms of DCD induction and CLL cell depletion from whole blood than the monoclonal antibodies in current clinical use for CLL therapy. mAb 37.1 and 37.2 therefore show a great potential for therapeutic applications in CLL patients.
Hallek M, Fischer K, Fingerle-Rowson G, Fink AM, Busch R, Mayer J et al. Addition of rituximab to fludarabine and cyclophosphamide in patients with chronic lymphocytic leukaemia: a randomised, open-label, phase 3 trial. Lancet 2010; 376: 1164–1174.
Jaglowski SM, Alinari L, Lapalombella R, Muthusamy N, Byrd JC . The clinical application of monoclonal antibodies in chronic lymphocytic leukemia. Blood 2010; 116: 3705–3714.
Zhao X, Lapalombella R, Joshi T, Cheney C, Gowda A, Hayden-Ledbetter MS et al. Targeting CD37-positive lymphoid malignancies with a novel engineered small modular immunopharmaceutical. Blood 2007; 110: 2569–2577.
Furman RR, Andritsos L, Flinn IW, Forero-Torres A, Foon KA, Pagel JM et al. Phase 1 Dose escalation study of TRU-016, an anti-CD37 SMIPTM protein in relapsed and refractory CLL. Blood 2010; 116: 56.
Heider K-H, Kiefer K, Ostermann E, Lamche E, Küpcü Z, Jacobi A et al. Novel Fc-engineered antibodies to 37 with pro-apoptotic and enhanced ADCC activity. Blood 2010; 116: 3916.
Patz M, Isaeva P, Forcob N, Muller B, Frenzel LP, Wendtner CM et al. Comparison of the in vitro effects of the anti-CD20 antibodies rituximab and GA101 on chronic lymphocytic leukaemia cells. Br J Haematol 2011; 152: 295–306.
Stanglmaier M, Reis S, Hallek M . Rituximab and alemtuzumab induce a nonclassic, caspase-independent apoptotic pathway in B-lymphoid cell lines and in chronic lymphocytic leukemia cells. Ann Hematol 2004; 83: 634–645.
Voso MT, Pantel G, Rutella S, Weis M, D’Alo F, Urbano R et al. Rituximab reduces the number of peripheral blood B-cells in vitro mainly by effector cell-mediated mechanisms. Haematologica 2002; 87: 918–925.
van Spriel AB, Puls KL, Sofi M, Pouniotis D, Hochrein H, Orinska Z et al. A regulatory role for CD37 in T-cell proliferation. J Immunol 2004; 172: 2953–2961.
This study was supported in part by research funding from BI to G.K. and M.H. This work is part of projects in our laboratory that are supported by grant DKH 109 214 of the German Cancer Aid to G.K. and M.H., by the German Research Council (DFG), SFB832, and by a Research Alliance grant from the CLL Global Research Foundation to M.H.
K.-H.H. is employed by Boehringer Ingelheim (BI), the company that developed the CD37 antibodies under investigation. G.K. and M.H. received research funding from BI. The remaining authors declare no conflict of interest.
Supplementary Information accompanies the paper on the Leukemia website
About this article
Cite this article
Krause, G., Patz, M., Isaeva, P. et al. Action of novel CD37 antibodies on chronic lymphocytic leukemia cells. Leukemia 26, 546–549 (2012) doi:10.1038/leu.2011.233
Biotechnology Letters (2018)
Cytotoxicity of the CD37 antibody BI 836826 against chronic lymphocytic leukaemia cells in combination with chemotherapeutic agents or PI3K inhibitors
British Journal of Haematology (2016)
Idelalisib improves CD37 antibody BI 836826 cytotoxicity against chemo-resistant /relapse-initiating CLL cells: a rationale for combination treatment
Blood Cancer Journal (2016)
Leukemia & Lymphoma (2015)
Frontiers in Physiology (2015)