Introduction

Triple-negative breast cancer (TNBC) metastasizes, is associated with poor prognosis and lacks effective targeted therapies.
Supported by previous data showing the cytostatic and metabolic effects of Deferiprone (DFP) in multiple cancer cell lines^1,3 we expanded our preliminary TNBC studies investigating the metabolic effects of DFP in the TNBC cell line 4T1¹ to include the human TNBC cell line MDA-MB-231. Here, we monitored and quantified the metabolic effects of DFP in live MDA-MB-231 cells in our MR-compatible bioreactor by multi-nuclear MRS.

Methods

Human MDA-MB-231 and murine 4T1 TNBC cells were grown in Dulbecco’s Modified Essential (DME) medium, containing 25 mM glucose (Glc), 6 mM glutamine (Gln), 10% fetal bovine serum and 1% penicillin/streptomycin, and cultured in humidified 5% CO2 / 95% air at 37 ⁰C.
For the MR-cell bioreactor studies, 6×10⁶ MDA-MB-231 cells or 3×10⁶ 4T1 cells respectively were seeded per 0.5 ml microcarrier beads (total 3.5 ml; 125-212 μm diameter; PP3772, Corning, New York, NY) until reaching ~70% confluence. Cell number and viability were estimated from a representative sample at the start (#cells_init) and end (#cells_end) of each MR experiment by Trypan Blue exclusion assay.
Continually perfused, live cells were studied in an MR-compatible cell bioreactor^1,2,3 on a 500 MHz AVANCE III Bruker scanner equipped with a 10-mm broad-band probe (Bruker BioSpin, Billerica, MA) for ¹H and ³¹P/¹3C MRS. To monitor metabolism changes, cells were continuously perfused for 31 hours with complete DME media. After the setup and the first ³¹P MR spectrum, the un-labelled perfusion medium was replaced with medium containing 25 mM 1-¹³C-labeled glucose for the remaining 30 h of experiment.
During each experiment, ¹H-decoupled, Nuclear Overhauser Effect-enhanced ¹³C MR spectra (1200 s relaxation time (TR), 45 flip angle, 30 kHz sweep width (SW), 8k points, 1000 averages, 20 min acquisition time) were acquired repeatedly interspaced with two consecutive 31P spectra (1200 s TR, 45 degree flip angle, 20 kHz SW, 2k points, 1800 averages, 30 min acquisition time) every six hours.
Three independent experiments each, were performed in untreated medium (CTRL) and three in medium containing 100 µM DFP (DFP-treated).
Quantification of MR spectral peaks was performed using AMARES (jMRUI v5.2). Metabolites were normalized to the β-NTP signal of the first ³¹P MR spectrum (β-NTP_init).
A 2-tailed, unpaired Student’s T test was used to compare DFP-treated and untreated (CTRL) MDA-MB-231 cell data. Metabolic ¹³C-labeling rates for DFP-treated and untreated MDA-MB-231 cells were compared to corresponding rates in previously studied¹ murine 4T1 cells for various glucose-driven metabolites.

Results & Discussion

The comparison between cell count ratios (r_C = #cells_end/#cells_initial) and corresponding cellular energy ratios (r_E=[β-NTP_end/β-NTP_init]) of DFP-treated and untreated MDA-MB-231 cells showed that cell mass and cellular energy ratios were significantly lower in DFP-treated than untreated MDA-MB-231 cells, with a significantly smaller decrease (p<0.05) observed for r_E than for r_C (Fig. 1). These observations agree with our previous findings for murine 4T1 TNBC cells and the cytostatic DFP effect in our IC₅₀studies¹. Figure 2 addresses the effects of DFP on cellular energy and phospholipid metabolism and pH, as measured by ³¹P MRS (representative spectra at 31 h shown in Fig. 2A). Time course data of β-NTP reveal similar β-NTP levels up-to 25 h beyond which β-NTP increases further only in the CTRL MDA-MB-231 cells, consistent with cell growth (Fig. 2B). Beyond the significantly higher PC in CTRL versus DFP-treated MDA-MB-231 cells, PC remains higher in the DFP-treated cells, consistent with reduced phospholipid turnover (Fig. 2C). PE is lower in DFP-treated than untreated MDA-MB-231 cells with a statistically significant difference at 12 h (Fig. 2D). NADH metabolites differ significantly at the latest time points (>15 h) between DFP-treated and untreated MDA-MB-231 cells (Fig. 2E). These results agree with our previous findings in murine 4T1 TNBC1 and the TRAMP-C2 prostate cancer cell lines³.
Figure 3 shows the comparison between representative ¹³C MR spectra of DFP-treated and untreated MDA-MB-231 cells acquired at 31 h. The quantitative evaluation of ¹³C metabolites (Fig. 4) shows various average metabolite levels that differ significantly between DFP-treated and untreated MDA-MB-231 cells. A significant decrease in uptake and metabolization of Glc 1Cα and Glc 1Cβ was observed in the presence of DFP at 12 h and 18 h for Glc 1Cα and at 12 h – 31 h for Glc 1Cβ. The incorporation of Glc into glycogen (Glyc) is significantly higher in DFP-treated than untreated MDA-MB-231 cells while de-novo glutamate (Glu) and lactate (Lac) syntheses significantly decrease in DFP-treated compared to untreated MDA-MB-231 cells (Fig. 4B).
Figure 5 shows the comparison of the metabolite ¹³C-labeling rates at each time point between DFP-treated and untreated human MDA-MB-231 murine 4T1 cells. In both TNBC cell lines, exposure to DFP affects glucose consumption, glucose-driven Krebs cycle activity, cellular energy and fatty acid metabolism in accordance to previous findings in TRAMP-C2 prostate cancer cells³.

Conclusion

The cytostatic and metabolic effects of DFP in the TNBC cell lines MDA-MB-231 and 4T1 is a preliminary step for the preclinical application in combination therapy in ongoing in vivo animal studies followed by clinical translation. Clinical translation is straightforward due to Deferiprone already being clinically approved for non-cancer related diseases.

Acknowledgements

This work was supported by grants W81ZWH-17-1-0525 (JAK) and P30CA008748 (Cancer Center Support Grant). We thank Dr. RV Simões for helpful discussions.