Response to "in vivo secretion of anti-CD3 × anti-tumor bispecific antibodies by gene-modified cells: over a decade of T-cell engagement".

To the editor: We thank Dr Comte and colleagues for their letter1 regarding our article on engager T cells.2 They point out their work on genetically modifying human lymphocytes to secrete an anti-carcinoembryonic antigen × anti-CD3-specific diabody.3 The concept of bispecific antibodies dates back to 1960, when Nisonoff et al. proposed that it should be possible to “prepare antibodies of mixed specificity.”4 Staerz and Bevan demonstrated in 1986 that it was possible to redirect T cells to target antigens with a bispecific antibody.5 Since then, the field of bispecific antibodies has gradually evolved, reflected by some 2,200 publications in PubMed. Among several bispecific antibody formats evaluated in clinical studies, bispecific T-cell engagers (BiTEs),6 which consist of two single-chain variable fragments specific for CD3 expressed on T cells and a tumor antigen, have shown impressive clinical results resulting in approval by the US Food and Drug Administration of the CD19-specific BiTE blinatumomab in 2014. References in papers like ours are not meant to be inclusive. Given the large number of publications in the field, we cited references that are focused on BiTEs, genetically modifying T cells with retroviral vectors, and the tumor antigen EphA2. Our work significantly differs from the publication by Compte et al.,3 in which the authors demonstrate (i) secretion of a carcinoembryonic antigen–specific diabody from lentivirally transduced T cells, (ii) antigen-dependent T-cell proliferation and tumor cell killing, and (iii) in vivo antitumor activity using a Winn assay in which tumor cells and T cells are mixed before subcutaneous injection into immunodeficient mice. We use a different bispecific antibody format (BiTEs), which does not require the synthesis of two polypeptides; also, we target a different tumor antigen (EphA2) and use retroviral vectors to genetically modify T cells. We demonstrate (i) antigen-dependent cytokine production in addition to T-cell proliferation and T-cell killing, (ii) enhanced engager molecule production post–T-cell activation, (iii) the ability of engager T cells to redirect bystander T cells to tumor cells in vitro using transwell assays and bystander T-cell proliferation in vivo, and (iv) in vivo antitumor activity using two xenograft models with day 7 tumors. We wish Dr Comte and colleagues continued success with their approach, especially with the genetic modification of other cell types to deliver bispecific antibodies in vivo.