Hydroxyurea Is Most Suitable for Cytoreduction of AML Prior to CD33/CD3 Bispecific BiTE® Antibody (AMG 330) Therapy: Uncompromised T-Cell Proliferation Ex-Vivo and CD33 Upregulation on AML Cells

Bispecific T-cell engager (BiTE®) antibodies represent a promising tool for anti-leukemic immunotherapy. The CD19/CD3-bispecific antibody blinatumomab was shown to be active in refractory and relapse patients with B-precursor acute lymphoblastic leukemia (Topp et al, ASCO 2014). Transient, blinatumomab-mediated cytokine release syndrome has been linked to target cell numbers as this phenomenon is predominantly observed within the first treatment cycle. In our previous work, we demonstrated that the bispecific CD33/CD3 BiTE® antibody AMG 330 is able to induce activation and proliferation of residual autologous T-cells and effectively mediates lysis of primary acute myeloid leukemia (AML) cells (Krupka et al, Blood 2014; 123(3):356-65).

We hypothesize that in AML patients with high initial leukocyte counts (WBC > 30.000/μl) a cytoreductive phase prior to AMG 330 therapy might be beneficial to reduce the incidence and severity of cytokine mediated toxicity. Ideally, the cytoreductive drug does not impair T-cell function or reduce target antigen expression level. In the current study, we evaluated the effect of cytarabine (20 µM), decitabine (5 µM), azacitidine (1 µM and 5 µM) and hydroxyurea (10 µM and 100 µM) on T-cell proliferation and function in close analogy to potential treatment algorithms for AML. Healthy donor (HD) T-cells were pre-incubated with the cytoreductive drugs for 72 hours. T-cells were CFSE-labeled and co-cultured with either HL60 or MV4-11 cells (effector cell:target (E:T) ratio 1:1) in the presence or absence of AMG 330 (5 ng/ml). After 3 days of co-culture, lysis of HL60 cells and T-cell proliferation was assessed by flow cytometry. Pretreatment of T-cells with cytarabine completely abrogated T-cell function (lysis of HL60 cells: untreated (UT): 96.9% vs 20 µM: 4.2%) and significantly impaired T-cell proliferation (UT: 31.2% vs 20 µM: 4.6%). These findings correlated to data using primary AML samples collected 3 and 6 days after discontinuation of cytarabine treatment. After a 3-day chemotherapy-free interval, we observed no relevant T-cell proliferation and lysis of AML cells upon the addition of AMG 330 to the ex-vivo long-term culture system (lysis of AML cells on day 12: 30%; fold change T-cell expansion 0.9). After a 6-day treatment-free interval, high T-cell proliferation and cytotoxicity against primary AML cells were observed (lysis of AML cells on day 12: 61%; fold change T-cell expansion: 3.1). In contrast to cytarabine, decitabine treatment only marginally impaired T-cell function. Similarly, pre-incubation with azacitidine did not convey a negative effect on T-cell function (lysis of HL60 cells: UT: 100% vs 1 µM: 94.9% vs 5µM: 86.8%; proliferation: UT: 90.9% vs 1 µM: 80% vs 5 µM: 66.8%). Pretreatment with hydroxyurea had the least impact on T-cell performance. It did not impair T-cell function (lysis of HL60 cells: UT: 100% vs 10 µM: 100% vs 100 µM: 100%) and proliferation compared to untreated controls (UT: 92.9% vs 100 µM 90.8% vs 10 µM 92.9%).

As we have previously shown that the level of CD33 expression correlates to kinetics of AMG 330-mediated lysis (Krupka et.al, EHA 2014), we analyzed the effect of the cytoreductive agents on CD33 expression level in AML cell lines and primary AML cells. Five AML cell lines (HL60, MV4-11, PL21, OCI-AML3, KG1a) and a primary AML patient sample were cultured in the presence or absence of decitabine (5 µM and 50 µM), azacitidine (1 µM and 5 µM) or hydroxyurea (10 µM and 100 µM) for 72 hours. The change of CD33 expression level was evaluated by flow cytometry (median fluorescence intensity, MFI). No significant changes in CD33 expression level were observed after culture of AML cell lines and primary AML cells with decitabine or azacitidine. In contrast, hydroxyurea upregulated surface expression of CD33 on 2/5 cell lines (HL60 and PL21) in a dose dependent manner (HL 60 MFI Ratio: UT 134.9 vs 10 µM 171.3 vs 100 µM 210; PL21 MFI Ratio: UT 166.9 vs 10 µM 177.9 vs 100 µM 191.8).

In summary, we could show that pretreatment with hydroxyurea did not impair T-cell function and proliferation. In addition, we observed an upregulation of CD33 expression on AML cell lines. As the BiTE® technology relies on T-cell function and target antigen expression level, sequential and combinatorial immuno-chemotherapeutic approaches need to address both issues. Our data support the use of hydroxyurea in AML patients that require cytoreduction prior to AMG 330 treatment.