DFMO inhibition of neuroblastoma tumorigenesis

Abstract

Background: Most high-risk neuroblastoma patients who relapse succumb to disease despite the existing therapy. We recently reported increased event-free and overall survival in neuroblastoma patients receiving difluoromethylornithine (DFMO) during maintenance therapy. The effect of DFMO on cellular processes associated with neuroblastoma tumorigenesis needs further elucidation. Previous studies have shown cytotoxicity with IC50 values >5-15 mM, these doses are physiologically unattainable in patients, prompting further mechanistic studies at therapeutic doses.

Methods: We characterized the effect of DFMO on cell viability, cell cycle, apoptosis, neurosphere formation, and protein expression in vitro using five established neuroblastoma cell lines (BE2C, CHLA-90, SHSY5Y, SMS-KCNR, and NGP) at clinically relevant doses of 0, 50, 100, 500, 1000, and 2500 μM. Limiting Dilution studies of tumor formation in murine models were performed. Statistical analysis was done using GraphPad and the level of significance set at p = 0.05.

Results: There was not a significant loss of cell viability or gain of apoptotic activity in the in vitro assays (p > 0.05). DFMO treatment initiated G1 to S phase cell cycle arrest. There was a dose-dependent decrease in frequency and size of neurospheres and a dose-dependent increase in beta-galactosidase activity in all cell lines. Tumor formation was decreased in xenografts both with DFMO-pretreated cells and in mice treated with DFMO.

Conclusion: DFMO treatment is cytostatic at physiologically relevant doses and inhibits tumor initiation and progression in mice. This study suggests that DFMO, inhibits neuroblastoma by targeting cellular processes integral to neuroblastoma tumorigenesis at clinically relevant doses.

Keywords: DFMO; ELDA; cell cycle; neuroblastoma; neurosphere; senescence; xenograft.

FIGURE 1

Effect of DFMO on cell viability, apoptosis, and cell cycle. Five established neuroblastoma cell lines (BE2C, CHLA‐90, SMS‐KCNR, SHSY5Y, and NGP) were treated with escalating doses of DFMO (0, 50, 100, 500, 1000, and 2500 μM) for 48 and 72 h. The cells were harvested and stained with a custom flow cytometry panel combining viability (fixable live/dead far red stain‐(Invitrogen), apoptosis markers (caspase‐3 and 7) (Vybrant FAM Casapse‐3 and 7 Assay Kit‐Invitrogen) and cell cycle stain (Vybrant DyeCycle™ Violet Stain‐Invitrogen). At least 100,000 live events were acquired on BD Fortessa cytometer and data analyzed using FlowJo software. Figure 1 (A) top panel shows median percentage of viable cells at each treatment for (i) 48 and (ii) 72 h, respectively, while the bottom panels (iii) and (iv) shows changes in apoptosis (p > 0.05). (B) shows median percentage of cells in G1, S and G2 cell cycle stages for (i)48 and (ii) 72 h post‐DFMO treatment.

FIGURE 2

DFMO treatment increases senescence‐associated βGal while downregulating Cyclin D1 and phospho‐Rb expression in neuroblastoma cell lines. Neuroblastoma cells were seeded in 6‐well plates and treated with escalating doses of DFMO (0, 50, 100, 500, 1000, and 2500 μM) for 48 and 72 h. (A) The cells were fixed and stained for SA‐βGal activity following strict adherence to the manufacturers protocol. Positive cells were detected by the presence of blue color staining. Five different field of view images were taken at 20× magnification. The number of positive cells were counted and expressed as percentage of the total number of cells per field of view. We observed a dose‐dependent increase in SA‐βGal activity in cells treated with DFMO relative to untreated controls at both (i) 48 and (ii) 72 h. (B) Densitometry analysis of immunoblot staining for cell cycle‐associated proteins (i) cyclin D1 (CCND1) and (ii) phosphorylated Retinoblastoma protein (p‐Rb) demonstrated overall decreases in levels of both proteins at 48 h, a further indication of DFMO treatment initiating transition toward a senescent phenotype while inhibiting cell proliferation.

FIGURE 3

Increasing doses of DFMO abrogates neurosphere formation and progression in neuroblastoma cell lines. Two cells per well for each cell line were seeded in 96‐well plates and treated with DFMO doses of 0, 100, 500, and 2500 μM. The frequency and the size of neurospheres were monitored and imaged weekly using Incucyte® ZOOM for 4 weeks. The top panel (A) is graphical representation of neurosphere numbers over time while the lower panel (B) shows the changes in neurosphere sizes over time in the tested cell lines. The bottom panel (C) shows representative images of neurosphere over time. We observed a dose‐dependent inhibition on the frequency and size of neurospheres in the five cell lines tested over time. Significant changes were observed in four of the cell lines (CHLA90, SHSY5Y, BE2C, and NGP) at 500 μM and 2500 μM compared to the vehicle.

FIGURE 4

Effect of in vitro DFMO pretreatment on tumor formation/initiation capacity of neuroblastoma cells. (A) Xenograft tumor model for limiting dilution analysis (LDA) using BE2C cells: mice were injected with 10, 50, or 100 cells pretreated for 10 days or 20 days with DFMO (5 mM) and tumor frequency was evaluated for 60 days. (B) The number of mice in each Xenograft group for LDA using BE2C and SMS‐KCNR cells. Mice were injected with 500, 1000, or 5000 cells pretreated for 10 or 20 days with DFMO (5 mM) and tumor frequency was evaluated for 90 days. (C) LDA of tumor takes at study termination. Tumor‐initiating cell frequencies and p‐values are calculated using online ELDA platform (accessed 20 September 2016), followed by post hoc log‐rank and Fisher's test analysis. Overall p‐value, p < 0.4. (D) Event‐free survival plots of xenograft mice injected with BE2C cells pretreated with DFMO (5 mM) for 0, 10, or 20 days. ANOVA p‐value evaluating the significance of DFMO effect for length of treatment, p < 0.001 compared to vehicle control. (E) Event‐free survival plots of xenograft mice injected with SMS‐KCNR cells pretreated with DFMO (5 mM) for 0, 10, or 20 days. ANOVA p‐value, p < 0.05 compared to vehicle control. (F) Neurosphere assay reporting tumor formation frequencies for DFMO‐pretreated NB cells. BE2C and SMS‐KCNR cells were pretreated with DFMO for 0, 10, or 20 days and allowed to form neurospheres in neurobasal media for 2 weeks.

FIGURE 5

Effect of in vivo DFMO treatment on tumor formation/initiation capacity of NB cells. (A) Xenograft for LDA using BE2C cells. Mice were injected with 10, 50, or 100 cells and treated with DFMO (2%) in drinking water. Tumor frequency was evaluated for 52 days. (B) Xenograft for LDA using SMS‐KCNR cells. Mice were injected with 500, 1000 or 5000 cells and treated with DFMO (2%) in drinking water. Tumor frequency was evaluated for 75 days. (C) Frequencies, and p‐values are calculated using online ELDA platform followed by post hoc log‐rank and Fisher's test analysis. Overall p‐value, p < 0.04 compared with vehicle control. (D) Event‐free survival plots of xenograft mice injected with BE2C cells and treated with DFMO (2%) in drinking water. Overall p‐value, p < 0.001 compared to vehicle control. (E) Event‐free survival plots of xenograft mice injected with SMS‐KCNR cells and treated with DFMO (2%) in drinking water. Overall p‐value, p < 0.005 compared to vehicle control. The number of mice in each group are depicted in 5C.

FIGURE 6

Effect of DFMO treatment on tumor growth and progression in mice. Mice were injected with 2 × 10⁶ BE2C cells. After tumors formed (7 days), the mice were given either normal drinking water (vehicle) or DFMO (2%)‐containing water. Following 7 days of treatment, the tumors were (A) resected and measured for size (B) volumes and (C) weights. The values are expressed as means ± SE (N = 3). 2‐way ANOVA, p < 0.001 compared with vehicle control. (D) Tumor cells dissociated from vehicle‐treated or DFMO‐treated mice were evaluated for neurosphere formation frequency for 2 weeks. The values are expressed as means ± SE (N = 5, 96‐well plates). 1‐way ANOVA followed by Fisher's t‐test, p < 0.005 compared with vehicle control. (E) Dissociated tumor cells were assayed for protein expression ODC1, LIN28B and MYCN by Western blot analysis. Band intensities were quantified and normalized to β‐actin (n = 9). Western blot images are depicted as 3 representative tumors from each group (T1–T3). (F) Bar graph depicting the percentage of xenograft tissue positive for each endpoint. (G) Tabular representation of IHC scoring, including the average score, standard error of the mean, and p‐value for each group; necrosis cellular senescence (p16ink4a), proliferation index (Ki67), apoptosis (cleaved caspase 3), and LIN28B (N = 9 for each group). (H) Representative immunohistochemistry of three tumors from each group (T1–T3) were stained for tumor structure by hematoxylin and eosin (H&E), proliferation index (Ki67), apoptosis (cleaved caspase 3), and LIN28 expression. N = 9 for the number of tissue slices quantified for each end.