Incorporation and Repair of Epigenetic Intermediates as Potential Chemotherapy Agents

Abstract

The incorporation of nucleoside analogs into DNA by polymerases, followed by their removal through base excision repair (BER), represents a promising strategy for cancer chemotherapy. In this study, we investigated the incorporation and cytotoxic effects of several nucleoside analogs-some of which are epigenetic reprogramming intermediates-in the U87 glioblastoma cell line. We found that two analogs, 5-hydroxymethyl-2'-deoxyuridine (5HmdU) and trifluorothymidine (TFT), are both cytotoxic and are efficiently incorporated into genomic DNA. In contrast, the 5-carboxy analogs-5-carboxy-2'-deoxyuridine (5CadU) and 5-carboxycytidine (5CadC)-showed no cytotoxicity and were not incorporated into DNA. Interestingly, 5-hydroxymethyl-2'-deoxycytidine (5HmdC) was cytotoxic but was not directly incorporated into DNA. Instead, it was deaminated into 5HmdU, which was then incorporated and likely responsible for the observed toxicity. 5HmdU is actively removed from DNA through the BER pathways. In contrast, TFT remains stably incorporated and is neither excised by BER nor does it hydrolyze into 5CadU-a known substrate for the DNA glycosylase SMUG1. We also found that N⁶-benzyladenosine (BzAdo), an inhibitor of the enzyme 2'-deoxynucleoside 5'-phosphate N-hydrolase (DNPH1), enhances the cytotoxicity of 5HmdU. However, the thymidine phosphorylase inhibitor tipiracil hydrochloride (TPI) does not increase the cytotoxic effect of TFT in U87 cells. Together, these findings highlight 5HmdU and TFT as promising chemotherapeutic agents for glioblastoma, each with distinct mechanisms of action and cellular processing.

Keywords: 5-hydroxymethyuracil; 5HmdU; BER; DNA; SMUG1; TFT; chemotherapy; epigenetic intermediates; glioblastoma; trifluorothymidine.

Figure 1

Structures of the nucleoside analogs examined in this study.

Figure 2

The nucleoside analogs TFT, 5HmdU, and 5HmdC are toxic to U87 cells in culture whereas 5CadC and 5CadU are not. Cell viability was assessed by the MTT assay at 3 days using 2 × 10³ cells. Decreasing cell viability is seen with increasing nucleoside concentration with active nucleosides. The nucleoside analogs 5CadU and 5CadC are not toxic in this assay. Each experiment was run in triplicate. Error bars represent the standard deviation of three determinations.

Figure 3

TFT and 5HmdU are incorporated into DNA. The incorporation of nucleoside analogs into DNA of cells in culture was measured by LC-MS/MS. Nucleoside analogs TFT and 5HmdU are incorporated into DNA while 5CadU, 5CadC, and 5HmdC are not incorporated. (A) Approximately 8340 ± 1844 TFT per 10⁶ dT molecules were incorporated in 4 h at 5 µM TFT. Approximately 12.9% TFT is hydrolyzed to 5CadU (1236 ± 309 per 10⁶ dT) during DNA digestion. (B) Approximately 698 ± 91 5HmdU per 10⁶ dT are incorporated into DNA. (C) 5HmdC is not incorporated into DNA, but, its enzymatic deamination product, 5HmdU is observed in DNA at 915 ± 118 per 10⁶ dT. (D) LC-MS/MS separation of nucleoside standards. In each experiment, 2 × 10⁶ cells were used. The average numbers of nucleosides and standard deviations were determined from three experimental replicates.

Figure 4

Nucleoside analogs can be modified in the cytoplasm. (A) Control (PBS) showing that nucleoside analogs are not detectable in the cytoplasm of untreated cells. (B) When TFT is added to cultured cells at 50 µM, both TFT and 5CadU can be seen. However, only 2.3% TFT hydrolyzes to 5CadU in that time. (C) When 5HmdC is added to media, both 5HmdC and 5HmdU are observed. At 1 h, approximately 6.3% of the 5HmdC has been deaminated to 5HmdU. (D) 5CadC is not deaminated to 5CadU at 1 h. In each experiment, 1.7 × 10⁶ cells were used. The average number of nucleosides and standard deviations were determined from three separate injections onto the LC-MS/MS.

Figure 5

5HmdU is repaired from DNA but TFT is not. (A) The amount of 5HmdU in the DNA is diminished from 4 to 28 h to 30.8% of its initial value by replication-dilution as well as from BER. (B,C) The TFT and 5BrdU fall to 61.7% and 58.0% of their initial values from 4 to 28 h due to replication-dilution. (B) The amount of 5CadU arising from TFT hydrolysis is 12.9% at 4 h and 11.1% at 28 h and is accounted for by hydrolysis during DNA digestion. In each experiment, 2 × 10⁶ cells were used. The average numbers of nucleosides and standard deviations were determined from three separate injections onto the LC-MS/MS from three experimental replicates.

Figure 6

Repair of duplex DNA containing a TFT:A base pair or a 5CadU:A base pair. A 79 base pair duplex (U:A) was incubated with uracil DNA-glycosylase (UDG) in CutSmart^TM buffer (New England Biolabs, Ipswich, MA, USA), 5 U of apurinic endonuclease (APE1), 6 pmol DNA polβ, 5 U E. coli ligase, 26 mM NAD⁺ and 20 mM of each indicated dNTP in 12.5 µL at 37 °C for 1 h. SMUG1 (10 U, New England Biolabs) was added and reactions were incubated at 37 °C for 1 h. Samples were run on denaturing 15% PAGE. The U:A base pair is processed by UDG and APE1 creating a repair gap. This gap is filled with DNA polβ and a modified dNTP (5CadU-TP or TFT-TP) and ligated with DNA ligase. The 5CadU, but not TFT, is cleaved by SMUG1.

Figure 7

The toxicity of 5HmdU is increased by BzAdo but the toxicity of TFT is not increased by TPI. The 3-day toxicity of TFT (A) and 5HmdU (B) were examined in U87 cells exposed to 1 µM BzAdo, 50 µM TPI or no inhibitor. The IC₅₀ of 5HmdU decreased 4-fold from 48 ± 7 µM to 12 ± 3 µM with BzAdo. In each experiment, 2.2 × 10³ cells were used. GraphPad Prism software (Version 10.5.0) was employed to calculate IC₅₀ values and error bars determined from 3 experimental replicates.

Figure 8

U87 proliferation in the presence of the nucleoside inhibitors as a function of time is in accord the cytotoxicity assays. Approximately 2.2 × 10⁴ cells in 1 mL were seeded per well into 7 12-well plates per experiment. After allowing cells to attach overnight, cells were treated with 48 µM 5HmdU, 10 µM TFT or 48 µM 5HmdU plus 1 µM BzAdo before placing in 10% CO₂ incubator. A single plate was removed and cells counted by Trypan Blue exclusion at 0, 9, 24, 34, 48, 57 and 72 h after adding analogs. Error bars represent triplicate cell counts from four experimental replicates.