Targeting a Specific Glycosylated Epitope of CD43 with a New Humanized Monoclonal Antibody for the Treatment of Pediatric and Adult T-Cell Acute Lymphoblastic Leukemia (T-ALL)

Introduction: T-cell acute lymphoblastic leukemia (T-ALL) accounts for about 20% of pediatric and adult ALL cases. Despite the use of intensive chemotherapy protocols, 25% of children and 50% of adult patients fail to respond or relapse. The 3-years prognosis for these patients is poor and novel treatment options are needed. The targeting of tumor-associated antigens by monoclonal antibodies (mAb) is among the most investigated immune-therapeutic strategies. Accordingly, we developed a new humanized mAb (hUMG1), directed against a heavy glycosylated epitope of CD43 which presents a high reactivity against T-ALL cells. Here we investigated the pre-clinical therapeutic activity and the mechanisms of action of hUMG1 in experimental models of T-ALL.

Methods: The expression of hUMG1 target was assessed by flow cytometry on tumor cell lines and primary samples from either T-ALL patients (n=48) or healthy donors (n =6). Humanized mAbs were generated by combining the variable domains of the murine antibody to the corresponding human IgG1 constant domains. Through immunohistochemistry (IHC) we screened several tissue microarrays (TMA) including human, cynomolgusmonkey and macacus rhesus (according to FDA/CE guidelines), rat and mouse normal tissues. Complement-mediated cytotoxicity (CDC), antibody-dependent cell-mediated cytotoxicity (ADCC) and cellular phagocytosis (ADCP) on T-ALL cells were evaluated by flow cytometry. Six different in vivo models on NSG mice (3 with administration of NK-92-CD16+ effectors) have been generated to evaluate mAbs activity in different disease settings: orthotopic, subcutaneous advanced disease (treatments started at 100 mmc) and subcutaneous limited disease (treatments started the day after injection).

Results: By screening different cancer cell lines, we observed hUMG1 target to be highly expressed on malignant T-ALL cells. We then tested T-ALL patient-derived blasts from 48 samples (40 pediatric and 8 adults) collected at diagnosis. The antigen was expressed on 23 out of 48 (48%), and most of them belonged to the subset of cortical (EGIL TIII) T-ALL group. By contrast the target antigen was not expressed on normal bone marrow cells from healthy donors. The analysis of the TMA including human normal tissues revealed a specific binding for thymus cortical lymphocytes only, leading us to hypothesize an acceptable safety profile. A humanized mAb, named hUMG1, and an afucosylated version of this mAb (aUMG1) were then developed. By gene expression profiling, western blot and flow cytometry, we observed that target binding by hUMG does not exert any direct activity on neoplastic cells. Subsequently, to investigate CDC, ADCC or ADCP, T-ALL cells were cultured in the presence of complement, peripheral blood mononuclear cells or macrophages, at increasing concentrations of both antibodies. Neither hUMG1 nor aUMG1 were able to induce CDC on target cells. Conversely, both mAbs induced CD16 downregulation, IFN-g production and degranulation on NK cells (more evident with aUMG1) and significant cytotoxicity against both T-ALL cell lines and primary blasts. Additionally, both mAbs induced ADCP. Lastly we observed a mAb-dependent activation of monocytes in the presence of target cells, as demonstrated by the reduction of CD16+ "non-classical" monocytes. Furthermore, we demonstrated potent activity of both mAbs in different T-ALL in vivo models. In an orthotopic model we observed 5 out of 20 treated mice free of disease after 100 days from injection as compared to none of the control group. In both subcutaneous models, we observed a strong ability of our antibody to delay tumor growth and to increase mice survival. Of note, the addition of NK-92-CD16+ strongly improved the activity of aUMG1. In the attempt to bring this antibody from bench to bedside, we assessed, through IHC, the expression of the hUMG1 target in healthy tissues from cynomolgus monkey, macacus rhesus, rat and mouse. We did not observe any reactivity, suggesting that the mAb target is very specific for human cells.

Conclusion: Here, we demonstrated that hUMG1 mAb recognizes an antigen specifically expressed on the majority of T-ALL, and its binding is able to mediate effective cytotoxicity against leukemia in vitro and in vivo, indicating that this antibody, in particular the aUMG1 version, may represent a novel promising immune-therapeutic tool for the treatment of T-ALL patients.