Cytomegalovirus-Specific T-Cell-Receptor-like Antibodies Target In Vivo-Infected Human Leukocytes Inducing Natural Killer Cell-Mediated Antibody-Dependent Cellular Cytotoxicity

Abstract

Cytomegalovirus (CMV) reactivation after stem cell or solid organ transplantation remains a major cause of morbidity and mortality in this setting. T-cell receptor (TCR)-like antibodies bind to intracellular peptides presented in major histocompatibility complex (MHC) molecules on the cell surface and may have the potential to replace T-cell function in immunocompromised patients. Three previously selected CMV-specific, human leukocyte antigen (HLA)-restricted (HLA-A*0101, HLA-A*0201 and HLA-B*0702) Fab-antibodies (A6, C1 and C7) were produced as IgG antibodies with Fc optimization. All antibodies showed specific binding to CMV peptide-loaded tumor cell lines and primary fibroblasts expressing the corresponding MHC-I molecules, leading to specific target cell lysis after the addition of natural killer (NK) cells. When deployed in combination as an antibody pool against target cells expressing more than one matching HLA allele, cytotoxic effects were amplified accordingly. CMV-specific TCR-like antibodies were also able to mediate their cytotoxic effects through neutrophils, which is important considering the delayed recovery of NK cells after stem cell transplantation. When tested on patient blood obtained during CMV reactivation, CMV-specific antibodies were able to bind to and induce cytotoxic effects in lymphocytes. CMV-specific TCR-like antibodies may find application in patients with CMV reactivation or at risk of CMV reactivation. In contrast to previous HLA/peptide-directed therapeutic approaches, the concept of a TCR-like antibody repertoire covering more than one HLA allele would make this therapeutic format available to a much larger group of patients.

Keywords: T-cell receptor (TCR)-like antibodies; cytomegalovirus; human leukocyte antigen (HLA) alleles; solid organ transplantation; stem cell transplantation.

Figure 1

Cytomegalovirus (CMV)-specific T-cell receptor-like antibodies (TCRLs) bind to in vivo-infected lymphocytes. Peripheral blood samples of CMV-infected patients were stained with CMV-specific TCRLs C1 (A), A6 (B) and C7 (C). CMV viral loads were measured from peripheral blood on the same day and are shown in Table 2. C1 showed the most robust binding to in vivo-infected lymphocytes, but A6 and C7 were also able to bind to CMV peptide-presenting lymphocytes. The binding capacities of CMV-specific antibodies varied between all three antibodies and also between patients. These variations in binding capacities may result from differences in HLA complex presentation for each HLA allele and from differences in CMV load at the time the binding experiments were performed. Grey: negative control, Green: A6 antibody, Red: C1 antibody, Blue: C7 antibody.

Figure 2

Staining intensity of CMV-specific TCRLs correlates with the viral load of patients. For three patients, longitudinal testing of CMV-specific antibodies was performed during the course of their CMV infections. In patient CMV3 (A), C1 showed strong binding to CMV-infected lymphocytes, with an infection load of 2300 IU/mL. In the same patient, staining intensity decreased during treatment and with dropping viral loads. In patient CMV7 (B), C1 showed strong binding to CMV-infected lymphocytes at both timepoints, with high viral loads (1.5 × 10⁶ and 251 × 10³ IU/mL). We believe that above a certain threshold of CMV copies, the loading limit of the existing MHC molecules is reached and a saturation effect occurs. This leads to the intensity of staining with CMV-specific TCRLs not increasing further. After effective treatment and a reduction in CMV copies to 3500 IU/mL, binding capacity was significantly reduced. Similar effects were observed in patient CMV1 (C), in whom the staining intensity of C1 correlated directly with measured CMV copies. The correlation of the staining intensity of CMV antibodies and CMV copy numbers was consistent within each patient, but not between different patients, e.g., 600 CMV IU/mL in patient CMV3 only resulted in very weak binding of C1, whereas strong binding was seen for C1 in patient CMV1 at 700 CMV IU/mL. Grey: negative control, Green: A6 antibody, Red: C1 antibody, Blue: C7 antibody.

Figure 3

Natural killer (NK) cell-mediated antibody dependent cellular cytotoxicity (ADCC) of in vivo CMV-infected PBMCs. ADCC tests against HLA A*0201-expressing PBMC target cells of acute CMV-infected patients mediated by the A*0201/pp65(aa495-503)-specific IgG1 antibody “C1”. PBMCs from patients (A) CMV3, (B) CMV7 and (C) CMV1 (black columns). (D) ADCC assessment against HLA B*0702-expressing PBMC target cells of the acute CMV-infected patient CMV8 mediated by the B*0702/pp65(aa417-426)-specific IgG1 antibody “C7” (black columns). PBMCs of non-infected healthy donors were used as CMV-negative target cells (white columns). Patients and controls in (A–C) express the MHC-I subtype HLA A*0201, whereas the patient and control in (D) express the MHC-I subtype HLA B*0702. NK cells derived from the healthy controls were used as effector cells in the indicated E/T ratio (y-axis labeling) for experiments with patient lymphocytes (allogeneic) as well as for controls (autologous). Since no significantly different lysis (*) of native autologous PBMCs compared to allogeneic patient PBMCs was observed, an allo-specific ADCC by the NK cells of the respective healthy donor against patient PBMCs could be excluded. No significantly stronger ADCC (**) was induced against native PBMCs of healthy volunteers even when treated with the respective IgG1 construct (white columns: native/C1 and native/C7). In contrast, adding C1 or C7 to CMV-infected PBMCs resulted in a significant increase in NK cell-mediated ADCC (black columns: native/C1 and native/C7) compared to healthy PBMCs (native/C1 or C7: white vs. black column; framed p-values). As expected, loading PBMCs with CMV peptide induced strong ADCC after addition of the corresponding IgG construct (CMV/C1 or CMV/C7).