+86-21-58356573
LANGUAGE

CD47 Targeted Drugs: Avoiding Blood Toxicity while Maintaining Pharmacodynamics

CD47 Targeted Drugs: Avoiding Blood Toxicity while Maintaining Pharmacodynamics

 

Author: Tian Wenzhi

 

       There are many discussions on the hematotoxicity of CD47 targeted drugs in the CD47 target drug field. To solve this problem, Dr. Tian Wenzhi, founder of ImmuneOnco, to discuss in depth for your reference.

      CD47 is a widely expressed membrane protein, which is expressed in almost all normal tissues (including red blood cells), mainly as a self-protective protein. The corresponding ligand of CD47 is SIRP alpha, which is mainly expressed in macrophages. When combined with CD47, it transmits inhibitory signals and inhibits the phagocytic activity of macrophages. The suppression signal induced by CD47-SIRP alpha interaction is the so-called "don't eat me" signal. In order to escape the attack of macrophages, cancer cells also protect themselves by inducing "Don't Eat Me" signal through high expression of CD47.

       Antibody drugs targeting CD47 play an anti-tumor role mainly by activating macrophages. To activate macrophages adequately, two basic conditions are needed: 1) blocking the interaction between CD47-SIRPa (such as CD47 antibody or SIRP alpha-Fc), thus eliminating the inhibitory signal; 2) Fc terminal of IgG1 binds to Fc gamma receptor on the surface of macrophage membrane, stimulating the membrane skeleton structure of macrophages. If these two signals are not satisfied at the same time, macrophages can only be partially activated.

      Because red blood cells also express CD47, if the drug (CD47 antibody or SIRP alpha-Fc) binds to red blood cells, it is possible to destroy red blood cells through two mechanisms: 1) the conjugation of antibodies with different red blood cells leads to red blood cell agglutination and subsequent erythrocyte lysis; 2) after the antibody binds to red blood cells, if the antibody is IgG1, the Fc end of the antibody can pass through macrophages. Fc R binds to activate macrophage phagocytosis of red blood cells. If the antibody is IgG4, it will not activate macrophages, but it can also cause the agglutination of red blood cells through the first mechanism and then the lysis of red blood cells. It should be noted that not all antibodies cause erythrocyte agglutination, so when screening antibodies, it is possible to screen antibodies that do not cause erythrocyte agglutination. T lymphocyte also expresses CD47. Some antibodies can bind to CD47 of T lymphocyte and induce cell apoptosis (J Immunol. 1999 Jun 15; 162 (12): 7031-40.). Therefore, this factor should be considered in screening CD47 antibodies and excluding the antibodies that can cause T lymphocyte apoptosis.

      There are many problems to be considered in the choice of subclass IgG for antibody drugs, which depend not only on the biological characteristics of the target cells and the target itself, but also on the mechanism of action of the antibody drugs. The general principles are as follows: 1) for the target of immune cells, multiple selection of IgG4 (e.g. O drug, K drug); 2) for the target of cancer cells, multiple selection of IgG1 (e.g. for EGFR, Her2, VEGFR2 and PD-L1 targets, etc.); 3) If Fc effect function is needed, IgG1 can only be selected (e.g. Rituximab, Herceptin, etc.).

        For CD47, because of its expression in cancer cells, IgG1 should be chosen reasonably, but because red blood cells also express CD47, if the screened antibody binds to red blood cells, IgG1 can not be selected, otherwise it will cause serious blood toxicity, but even if IgG4 is selected, if it can cause red blood cell agglutination through coupling, it will also destroy red blood cells and produce blood toxicity. Unless this factor is taken into account in antibody screening and antibodies that cause erythrocyte agglutination are excluded (not all antibodies cause erythrocyte agglutination). However, if IgG4 is not supplied by Fc effect, its clinical efficacy will be greatly reduced. It is impossible for a single drug to have a good therapeutic effect. It must be combined with other antibody drugs. In combination, there are two situations (for example, CD47 and CD20 antibodies):

       1)Both antibodies bind to the same target cell (CD20 cell).

       2)The two antibodies bind to different target cells respectively.

     As far as therapeutic effect is concerned, synergistic effect can only occur when the first condition occurs, which can induce macrophages to attack target cells (Fig. 1A). In the second case, only CD20 antibody can induce attacks on target cells through its Fc effect function, and there will be no synergistic effect between them (Fig. 1B). For our molecule (IMM01), because it does not bind to human erythrocyte at all, we chose IgG1, which can remove the "don't eat me" signal by blocking the interaction between CD47-SIRP alpha and exert strong anti-tumor activity through Fc-mediated effect (mainly ADCP) (Fig. 1C). 


                                        Figure 1. Mechanisms of CD47 targeting drugs

      Because CD47 is expressed in erythrocyte, any CD47 target drug (antibody or SIRP alpha-Fc) that can bind to erythrocyte may cause erythrocyte destruction through two mechanisms alone or superimposed. One is erythrocyte agglutination through coupling, which leads to erythrocyte lysis. The other is activation of macrophage phagocytosis through Fc-mediated ADCP and destroy red blood cells.

      The known CD47 antibodies are highly binding to human erythrocyte, while SIRP alpha-Fc only has a weak binding (Fig. 2). So the question is, since both red blood cells and cancer cells express CD47, why does SIRP alpha-Fc not bind to red blood cells but to cancer cells? This is determined by the different conformation of CD47 in two kinds of cells. Of course, it is also related to the SIRP alpha-Fc molecule itself. That is, the binding of SIRP alpha-Fc molecule to CD47 is affected by its conformation in the cell membrane, while the antibody to CD47 is not affected by the conformation of CD47 in the cell membrane. 

 

                                         Figure 2. Erythrocyte binding activity analysis


      The conformation of CD47 in erythrocyte of different species is different, so the binding activity of SIRP alpha-Fc to erythrocyte of different species is also different (Figure 3). SIRP alpha-Fc hardly binds to human erythrocyte, but can bind to cynomolgus monkey and mouse erythrocyte (binding to mouse erythrocyte refers to rat SIRP alpha-Fc, human SIRP alpha-Fc does not bind to mouse erythrocyte).

   Our molecule (IMM01) is a genetically engineered SIRP alpha-Fc, which does not bind to human erythrocyte at all (Fig. 4), but can bind to monkey erythrocyte and activate macrophages to phagocyte monkey erythrocyte through Fc-mediated ADCP, while human erythrocyte has no phagocytic effect at all (Fig. 5). After analysis with blood samples from 100 individuals, it was fully confirmed that IMM01 did not bind to human erythrocyte at all (Fig. 6). That is to say, when monkeys are used to evaluate the safety of drugs, there will be erythrocyte toxicity, which is not expected to occur in the human body.


 

              Figure 3. Conformation of CD47 in red blood cells of different species and genera

 

 

              Figure 4. Conformation of CD47 in red blood cells of different species and genera

 

 

                      Fig. 5. The effect of IMM01 on macrophage phagocytosis of erythrocytes


  

                Figure 6. The effect of IMM01 on the phagocytosis of erythrocytes by macrophages

 

       Hu5F9 is a humanized CD47 antibody, IgG4, which does not have Fc effect. It relieves the "don't eat me" signal simply by blocking the interaction between CD47-SIRP alpha. Therefore, the efficacy of Hu5F9 alone will be greatly reduced. Only when combined with other antibodies (such as Rituximab) can it have a significant effect. At the same time, because Hu5F9 binds highly to human red blood cells and causes erythrocyte agglutination, the clinical incidence of grade 3 anemia is as high as 10% (fig. 7). Because it induces T lymphocyte apoptosis, the probability of grade 3 lymphocyte decrease is as high as 15%. These adverse reactions will restrict the long-term therapeutic effect of the molecule.

 

                                  Figure 7. Clinical toxicity and side effects of Hu5F9

 

     Our molecule (IMM01) has both CD47 blocking activity and Fc effect, so it has strong anti-tumor activity. At the same time, because it does not bind to human erythrocyte and does not induce T cell apoptosis, it is expected that there will be no serious blood toxicity.

     Our experimental studies show that blocking the interaction of CD47-SIRP alpha alone has limited efficacy. When Fc effect was removed, the efficacy of IMM01 was lost by 90% (Figure 8, left). At the same time, we also proved that if macrophages were removed before treatment, the efficacy would also be lost by 80-90% (Fig. 8, right).

 

                             Figure 8. Analysis of the anti-tumor mechanism of IMM01

 

     Our molecule (IMM01) has both CD47 blocking activity and Fc effect, so it has strong anti-tumor activity. At the same time, because it does not bind to human erythrocyte and does not induce T cell apoptosis, it is expected that there will be no serious blood toxicity.

 

       Summary:

     A. To develop CD47 antibody, two exclusive conditions must be set in the process of antibody screening: 1) erythrocyte binding and inducing erythrocyte agglutination; 2) inducing T lymphocyte apoptosis. If we do not exclude the activity of these two antibodies, it will be difficult to achieve clinical success.

       B.  Antibodies targeting CD47 need to retain both target blocking activity and Fc effect function. If only the former is retained, the efficacy will be greatly reduced, only for blood tumors, and need to be combined with other antibodies. In addition, IgG4 antibodies against CD47 are not effective against solid tumors, even when combined with other antibody drugs.