Cytolethal distending toxin B as a cell-killing component of tumor-targeted anthrax toxin fusion proteins

Abstract

Cytolethal distending toxin (Cdt) is produced by Gram-negative bacteria of several species. It is composed of three subunits, CdtA, CdtB, and CdtC, with CdtB being the catalytic subunit. We fused CdtB from Haemophilus ducreyi to the N-terminal 255 amino acids of Bacillus anthracis toxin lethal factor (LFn) to design a novel, potentially potent antitumor drug. As a result of this fusion, CdtB was transported into the cytosol of targeted cells via the efficient delivery mechanism of anthrax toxin. The fusion protein efficiently killed various human tumor cell lines by first inducing a complete cell cycle arrest in the G2/M phase, followed by induction of apoptosis. The fusion protein showed very low toxicity in mouse experiments and impressive antitumor effects in a Lewis Lung carcinoma model, with a 90% cure rate. This study demonstrates that efficient drug delivery by a modified anthrax toxin system combined with the enzymatic activity of CdtB has great potential as anticancer treatment and should be considered for the development of novel anticancer drugs.

Figure 1

Cytotoxicity of fusion proteins on RAW264.7 cells (murine leukemic monocyte/macrophages), CHO K1 cells (Chinese Hamster Ovary cells), HeLa cells (human cervical carcinoma cell line), HN6 cells (human head and neck cancer cell line), and LL3 cells (murine Lewis Lung carcinoma cells). Cells (5000/well grown overnight) were exposed to different concentrations of fusion proteins for 72 h. All samples contained a fixed concentration of 250 ng/ml PA or tumor-specific PA-L1 (e) and varying concentrations of LFnCdtB or CdtB (a–d) or LFnCdtB or FP59AGG (e). Viable cells were quantitated in an assay employing 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT). Relative survival was calculated as the percentage of living cells after treatment in relation to untreated cells. Error bars indicate S.E.M. of 2 (CdtB) and 3–6 (LFnCdtB and FP59AGG) independent experiments performed in triplicate

Figure 2

Cell cycle analysis of CHO K1 cells (0.2 × 10⁶ cells) treated with 250 ng/ml PA and 100 pM CdtB or LFnCdtB or 0.1 pM FP59AGG for 4–48 h. As a control, cells were incubated with 1 μM staurosporine for 4–14 h to induce apoptosis. Cells were fixed with ethanol, stained by propidium iodide to quantify cellular DNA, and 25 000 cells were counted by flow cytometry. Cell cycle analysis by FlowJo software (Flow Jo version 7.6; Tree Star, Ashland, OR, USA) indicated the percentage of cells in the G0/G1 phase (green cell population and percentage on the left side in each panel) and in the G2/M phase (turquois cell population and percentage on the right side in each panel)

Figure 3

Apoptosis induction by LFnCdtB analyzed by PARP cleavage (a) and TUNEL staining (b). (a) HeLa cells (1 × 10⁶ cells) were exposed to 250 ng/ml PA and 10 nM LFnCdtB or CdtB for 24–72 h. As a control, cells were incubated with 1 μM staurosporine for 2 h to induce apoptosis. Cells were lyzed and analyzed by simultaneous anti-PARP (red signals) and anti-actin (green signals) immunodetection following western blotting. (b) CHO K1 cells (0.2 × 10⁶ cells) were treated with 250 ng/ml PA and 100 pM CdtB or LFnCdtB or 0.1 pM FP59AGG for 4–48 h. As a control, cells were incubated with 1 μM staurosporine for 4–14 h to induce apoptosis. Cells were fixed with ethanol, stained by TUNEL staining to quantify apoptotic cells, and 10 000 cells were counted by flow cytometry. TUNEL-positive apoptotic cells were gated by high fluorescence and the percentage is indicated in every panel

Figure 4

Intracellular detection (a–c) and nuclear action (d) of LFnCdtB. (a–c) To detect the fusion proteins in the cytosolic or nuclear fraction of cells, CHO K1 cells (1.75 × 10⁶ cells) were incubated with 250 ng/ml PA and 5 nM biotinylated LFnCdtB or FP59AGG for 1 or 4 h. Additional cells were incubated with a mixture of 100 nM each CdtA, CdtB, and CdtC (Cdt). Cytosolic and nuclear fractions were isolated, and 40 μg total protein of each fraction were separated by SDS–PAGE and submitted to western blotting. Immunodetection was performed simultaneously by streptavidin (a, green signal), anti-CdtB (b, red signal), anti-LF (c, red signal), and anti-Pseudomonas exotoxin A (c, green signal), and anti-p84 (nuclear marker protein, red signal in a and green signal in b) and anti-MEK2 (cytosolic marker protein, red signal in a and green signal in b). Merged signals are shown in yellow. The nuclear action of CdtB was detected by immunofluorescence detection of phosphorylated histone H2A.X (d). HeLa cells (0.2 × 10⁶ cells on cover slips) were incubated with 250 ng/ml PA and 10 nM LFnCdtB or 10 nM FP59AGG or with 100 nM CdtA, CdtB, and CdtC each (Cdt) for 1–24 h. Cells were fixed and permeabilized before staining by anti-pH2A.X and DAPI. Results show DAPI fluorescence in blue (nuclear DNA, left panel), phase contrast (middle panel), and pH2A.X fluorescence in red (DNA damage, right panel)

Figure 5

Tumor treatment by LFnCdtB in a mouse model. 10–13 C57BL/6 mice per group were injected subcutaneously with LL3 mouse melanoma tumors on day 0. Injection of PA+LFnCdtB or tumor-specific PA-L1+LFnCdtB fusion into the mouse peritoneum was performed on days 5, 7, 9, 12, 14, and 16 (arrows) using either PBS (black line), 100 μg wild-type PA+100 μg LFnCdtB (blue line) or 100 μg PA-L1+100 μg LFnCdtB (red line). (a) Relative mean mouse body weight until day 19. Error bars indicate the S.D. (b) Mean tumor masses until day 19. Tumor mass was calculated from measurements of the tumor width, depth, and height