Histone deacetylase inhibition enhances the therapeutic effects of methotrexate on primary central nervous system lymphoma

Abstract

Background: Polyglutamylation is a reversible protein modification that commonly occurs in tumor cells. Methotrexate (MTX) in tumor cells is polyglutamylated and strongly binds to dihydrofolate reductase (DHFR) without competitive inhibition by leucovorin. Therefore, tumor cells with high polyglutamylation levels are supposed to be selectively killed, whereas normal cells with lower polyglutamylation are rescued by leucovorin. This study investigated the combined effects of MTX plus histone deacetylase inhibitors (HDACIs), which upregulate MTX polyglutamylation, in primary central nervous system lymphoma (PCNSL).

Methods: We evaluated cell viability after MTX treatment and leucovorin rescue and compared the expression of folylpolyglutamate synthetase (FPGS), γ-glutamyl hydrolase (GGH), and DHFR in 2 human PCNSL-derived cell lines (HKBML and TK) and a human Burkitt lymphoma cell line (TL-1). Combination treatments were created using 4 HDACIs: panobinostat, vorinostat, sodium butyrate, and valproic acid. The expression of DHFR was examined as well as ratios of FPGS/GGH expression. The combined effects of MTX plus HDACIs were evaluated using a cell viability assay, mass spectroscopy imaging, and subcutaneous and intracranial xenograft models.

Results: HDACIs upregulated the ratio of FPGS/GGH expression resulting in increased polyglutamylation of MTX, but also downregulated expression of the target molecule of MTX: DHFR. The combination of MTX and vorinostat decreased cell viability in vitro (P < .05) and tumor volumes in a subcutaneous model (P < .0001), and prolonged survival in an intracranial model (P < .01), relative to controls.

Conclusion: HDACIs enhanced the therapeutic effect of MTX through increased polyglutamylation of MTX and concomitant downregulation of DHFR expression.

Keywords: histone deacetylase inhibitor; leucovorin; methotrexate; polyglutamylation; primary central nervous system lymphoma.

Figure 1.

Effects of methotrexate (MTX) treatment and leucovorin (LV) rescue were related to the degree of polyglutamylation in lymphoma cell lines. (A) Viability assays using cell lines after incubation with MTX (CellTiter-Glo). Each cell line was analyzed after 72 h of incubation. (b) Cells were treated with MTX for 24 h, followed by the addition of LV. Cell viability was assessed 48 h later. We defined EC₅₀ as the concentration of LV that recovered 50% cell viability. Data are shown as mean value ± SD from 3 independent experiments. *P < .01, compared with MTX. (C) Immunoblotting for folypolyglutamate synthetase (FPGS), γ-glutamyl hydrolase (GGH), and dihydrofolate reductase (DHFR) in the different cell lines. The internal control was α-tubulin. The relative expression of FPGS/GGH represents the ratio of FPGS/α-tubulin and GGH/α-tubulin calculated by densitometry. The FPGS/GGH ratio of HKBML is adjusted to 1. Data are shown as mean value ± SD from 3 independent experiments; *P < .05, **P < .01.

Figure 2.

Changes in expression of folypolyglutamate synthetase (FPGS), γ-glutamyl hydrolase (GGH), and dihydrofolate reductase (DHFR) in TL-1 cells treated by histone deacetylase inhibitors (HDACIs) for 72 h. The relative expression of FPGS/GGH represents the ratio of FPGS/α-tubulin and GGH/α-tubulin calculated by densitometry. The FPGS/GGH ratio of TL-1 control is adjusted to 1. Data are shown as mean value ± SD from 3 independent experiments, *P < .05.

Figure 3.

Molecular imaging of the distribution of methotrexate (MTX) polyglutamates 2–7 in TL-1 cells treated using vorinostat or vehicle followed by MTX. (A) Imaging of phospholipid (PC34:1), the surrogating cell component. (B) Distributions of MTX polyglutamates 2, 3, and 5. In the MALDI-MS imaging software, we could only display up to 3 channels at a time. Therefore, we used MTX-PG 2, 3, and 5 as representatives for figures because there were statistically significant differences in expression of MTX-PG 2, 3, and 5 between MTX and MTX plus vorinostat. (C) Relative amounts of MTX polyglutamates 2–7 are calculated by correcting the numbers of each detected pixel with phospholipid PC (34:1) in TL-1 cells, with or without vorinostat treatment. Error bars indicate standard deviation. *P < .05, compared with MTX.

Figure 4.

Effects of the methotrexate (MTX) plus vorinostat combination on lymphoma cell lines in vitro and in mouse xenograft models. (A) Cells were sequentially treated using MTX (24 h) and vorinostat (24 h), and cell viability was determined after 48 h of incubation in a drug-free medium. Error bars represent standard deviations. The viability of cells treated using MTX + vorinostat was compared to that of cells treated using MTX or vorinostat alone; *P < .05, **P < .01. (B) Effects of MTX and vorinostat in a mouse subcutaneous tumor model. Mice with TL-1 subcutaneous tumors received an intraperitoneal injection of vehicle, MTX (50 mg/kg), vorinostat (50 mg/kg), or MTX (50 mg/kg) + vorinostat (50 mg/kg) on days 15, 18, and 21. Error bars indicate standard deviation; n = 5 for each. *P < .0005, **P < .0001. (C) Xenograft tumor sections derived from the intracranial xenograft mouse model were stained using (Upper) hematoxylin and eosin or (Bottom) anti-CD20 to identify B-cell lymphomas. Scale bar, 2 mm (×20), 50 μm (×400). (D) Kaplan–Meier survival curves for mice harboring the TK intracranial xenografts after treatment using vehicle, MTX (50 mg/kg), vorinostat (50 mg/kg), or MTX (50 mg/kg) + vorinostat (50 mg/kg) The log-rank test was used for the statistical analysis. *P < .05, **P < .01.