CD64-directed microtubule associated protein tau kills leukemic blasts ex vivo

Abstract

Fc gamma receptor I (FcγRI, CD64) is a well-known target antigen for passive immunotherapy against acute myeloid leukemia and chronic myelomonocytic leukemia. We recently reported the preclinical immunotherapeutic potential of microtubule associated protein tau (MAP) against a variety of cancer types including breast carcinoma and Hodgkin's lymphoma. Here we demonstrate that the CD64-directed human cytolytic fusion protein H22(scFv)-MAP kills ex vivo 15-50% of CD64+ leukemic blasts derived from seven myeloid leukemia patients. Furthermore, in contrast to the nonspecific cytostatic agent paclitaxel, H22(scFv)-MAP showed no cytotoxicity towards healthy CD64+ PBMC-derived cells and macrophages. The targeted delivery of this microtubule stabilizing agent therefore offers a promising new strategy for specific treatment of CD64+ leukemia.

Keywords: Fc-gamma receptor (CD64); cytolytic fusion proteins; immunotherapy; microtubule associated protein tau (MAP); myeloid leukemia.

Figure 1. Characterization of the CD64⁺ leukemic blasts and H22(scFv)-MAP binding

Blood samples were obtained from seven untreated patients diagnosed with different forms of leukemia (Table 1). A. The proportion of primary CD33⁺ CD64⁺ leukemic blasts from seven patients is compared directly to the isotype control. Percentiles indicate the proportion of double positive cells from all measured events. B. The ability of purified H22(scFv)-MAP to bind the leukemic cells in vitro was tested by flow cytometry. The binding activity of H22(scFv)-MAP (left) and H22(scFv)-ETA’ (right) correlate to the estimated CD64⁺ level. C. Comparative binding activities of both CD64-targeting fusion proteins. Abbreviations: MFI, mean fluorescence intensity.

Figure 2. H22(scFv)-MAP induces apoptosis in CD64⁺ leukemic blasts ex vivo

Leukemic blasts isolated from four AML patients A. and three CMML patients B. were treated ex vivo with 200 nM H22(scFv)-ETA’, H22(scFv)-MAP or Mock-MAP for 12 h. The leukemic cells were stained with Annexin V–eGFP and PI. Dot blots (ungated) show the specific pro-apoptotic effect of the H22(scFv)-MAP compared to Mock-MAP and H22(scFv)-ETA’. The experiment was carried out in duplicate.

Figure 3. H22(scFv)-MAP is cytotoxic towards CD64⁺ leukemic blasts ex vivo

A. Quantitative analysis of the dot blots in Figure 2 representing the proportion of cells undergoing apoptosis (the sum of early (Annexin V⁺/PI⁻) and late (Annexin V⁺/PI⁺) apoptotic cells) in the different experimental groups. Values are presented as box and whiskers plots with the interquartile range and minimum/maximum values indicated by the error bars. Each measurement was taken in duplicate. Statistical analysis was carried out using the non-parametric one-way ANOVA test (Kruskal-Wallis test by ranks): *p ≤ 0.05. B. Cytotoxicity of H22(scFv)-MAP plotted against its binding activity on leukemic blasts. C. Correlation between the cytotoxicity of the CD64-targeting fusion proteins. Abbreviations: MFI, mean fluorescence intensity.

Figure 4. H22(scFv)-MAP shown no cytotoxicity towards healthy PMBC-derived cells

A. Quantitative analysis of the proportion of healthy PMBC-derived cells undergoing apoptosis (the sum of early (Annexin V⁺/PI⁻) and late (Annexin V⁺/PI⁺) apoptotic cells) in the different experimental groups. Dot blots are shown in Supplementary Figure S2. B. TAMs express significantly lower amounts of CD64. The levels of CD64 on differently polarized, healthy PBMC-derived macrophages (n=3 for M0/1/2 and n=5 for TAM) were assessed by flow cytometry. C. Representative dot blots (ungated) confirm that TAMs are resistant to the CD64-directed MAP hCFP but not to the nonspecific cytostatic agent paclitaxel. D. Quantification of TAM apoptosis (n=5 PBMC donors) in the different experimental groups. Values are presented as box and whiskers plots with the interquartile range and minimum/maximum values indicated by the error bars. Each measurement was taken in duplicate. Statistical analysis was carried out using the non-parametric one-way ANOVA test (Kruskal-Wallis test by ranks): *p ≤ 0.05.