Antibody to CD47: Preclinical Promise for Chronic Lymphocytic Leukemia

Abstract & commentary

Synopsis: The antibody against CD47 or the natural ligand for CD47, thrombospondin, was able to kill leukemic cells from 42 patients with chronic lymphocytic leukemia. This target may be suitable for clinical application.

Source: Mateo V, et al. Nat Med 1999;5:1277-1284.

Chronic lymphocytic leukemia (cll) is one of the more common lymphoid malignancies. Treatment in the early stages of disease does not improve the prognosis.1 When the disease becomes symptomatic, nucleosides, especially fludarabine, and alkylating agents such as chlorambucil are frequently capable of producing antitumor responses for variable periods. However, no curative therapy is available.

As more is learned about the biology of lymphoid cells and the tumors derived from them, it has become clear that certain signals that are involved in normal lymphocyte regulation may provide an effective means of treating tumors. Thus, there is now preclinical evidence that aggressive histology B-cell lymphomas may be killed by activating the CD40 molecule on their surface.2 The CD30 molecule on the surface of anaplastic large cell lymphoma may be an effective target for therapy.3 In addition, recent data on CLL cells have identified a potential target for biologic therapy in this form of lymphoid malignancy.

Thrombospondin is a major component of alpha granules in platelets and is produced by other cells as well. It is an adhesive glycoprotein that is produced by platelets locally and appears to inhibit fibrinolysis; thus, it helps maintain clot integrity. However, many additional roles for the protein are being uncovered, including roles in angiogenesis and phagocytosis. One receptor for thrombospondin is CD47, also known as integrin-associated protein because it is often found in association with integrin molecules, especially alpha v beta 3, which is also known as the vitronectin receptor. On lymphoid cells, CD47 is expressed independently from integrins.

Mateo and colleagues examined the in vitro effect of CD47 ligation on the survival of CLL cells obtained from 42 different patients. Antibody to CD47 and thrombospondin itself were both used as ligands. Cells exposed to these stimuli adhered strongly to the plastic culture dish and began to undergo programmed cell death or apoptosis. However, unlike most forms of apoptosis, the cells died without activating the caspase pathway (caspases are proteases that destroy many key proteins necessary for damage repair). Among the cells dying from CD47 ligation were at least two cases in which the CLL cells were refractory to glucocorticoid-induced apoptosis, a finding often associated with refractoriness to treatment. Other stimuli, for example, antibody to CD5, can also induce the adherence of CLL cells to plastic, but this is not accompanied by the death of the cells. Thus, signal transduction through CD47 is thought to send a death signal to the cell and adhesion is not sufficient to send that signal. The death signal is strong; after 30 minutes of exposure, death was irreversible.

A number of factors have been identified that help CLL survive in vitro and presumably in vivo. These factors include interleukin-4 and interferon-gamma. Although both of these factors as well as adherence to bone marrow stroma protected CLL cells from spontaneous apoptosis in vitro, none of these factors could prevent CD47-induced apoptosis.


So what will it take to make thrombospondin, or anti-CD47 antibody, an effective treatment for CLL? The steps for clinical development are well-defined. But what are the odds that CD47 represents a useful therapeutic target? Several features of the effects of CD47 stimulation are encouraging. It seems to be working in a fashion that is distinct from drugs and glucocorticoids. The apoptosis induced by these agents is caspase-dependent whereas CD47-induced apoptosis is caspase-independent. It is possible that one mechanism of resistance to standard treatment involves the development of defects in the caspase pathway. Thus, CD47 activation may be effective in drug-resistant cases of CLL. Preliminary evidence suggests this is the case. Of the 42 cases examined by Mateo et al, several were resistant to drug-induced killing but all 42 showed dramatic killing by CD47 ligation. Furthermore, drug-induced killing can be largely reversed by interleukin-4 and interferon-gamma, agents that have no effect on CD47 killing.

Another intriguing feature of the CD47 effect is its speed and irreversibility. CD47 ligation induces signal transduction and the pathways that are activated through the receptor are still being defined. However, cell death signalling is one of the most promising areas of targeted drug development for cancer.

Arguing against CD47 as a therapeutic target are its widespread distribution, its poorly defined function, the widespread distribution of its natural ligand, the poorly defined function of its natural ligand. We have no idea of the range of normal cells that express CD47 or what would happen to them if they suddenly received a signal through it. Will there be a therapeutic index? Will the receptor down-modulate or shed from the cell surface? Will the cells develop resistance to CD47 signalling? Much is yet to be defined. However, it is at least encouraging that we now know enough about the cell biology of CLL cells that an alternative to DNA poisons can at least be tested for its therapeutic potential.


1. Dihiero G, et al. N Engl J Med 1998;338:1506-1514.

2. Funakoshis S, et al. Blood 1994;83:2787-2794.

3. Tian ZG, et al. Cancer Res 1995;55:5335-5341.

Chronic lymphocytic leukemia cells can be killed in vitro by exposure to which of the following agents?

a. Thrombomodulin

b. Thrombospondin

c. Thrombopoietin

d. Thromboglobulin

e. Fibronectin