Disruption of ATR Signaling by Epstein-Barr Virus Latent Membrane Protein 1 Sensitizes Nasopharyngeal Carcinoma Cells to Cisplatin

Abstract

Nasopharyngeal carcinoma (NPC) occurs with high incidence in Southeast Asia where almost all tumors are associated with Epstein-Barr virus (EBV) infection. Cisplatin is used in combination chemotherapy. In this study, we determined that the EBV oncoprotein, latent membrane protein 1 (LMP1), perturbs DNA damage response (DDR) signaling, activation of cell cycle checkpoints, and sensitivity to cisplatin in NPC cells (HK1). Hypersensitivity was validated by LMP1 knockdown and CRISPR/Cas9 targeting in HK1-EBV cells with latent EBV infection. The conserved PxQxT motif (in CTAR1) and Y384 residue (in CTAR2) were required for the hypersensitivity. Inhibition of ATR (VE821 or AZD6738), but not ATM (KU55933 or AZD0156), phenocopied the G1 arrest and hypersensitivity. Attenuation of DDR signaling and hypersensitivity by LMP1 or ATR inhibition was also observed in the C17 NPC cell line with restored stable LMP1 expression. LMP1 expression in NPC tumors is highly variable. Publicly available RNA-sequencing data from microdissected NPC tumors showed that LMP1 expression in the primary tumors was the lowest in cisplatin-treated patients that experienced recurrence. These findings could have clinical significance in stratifying NPC patients such that tumors with limited or variable LMP1 expression might benefit from ATR inhibitor therapy.

Keywords: DNA damage response signaling; Epstein–Barr virus; latent membrane protein 1; nasopharyngeal carcinoma.

Figure 1

LMP1 disrupts cisplatin‐induced cell cycle arrest. (A) HK1 cells treated with doxycycline (50 ng/mL, 24 h) to induce LMP1 expression were exposed to increasing doses of cisplatin for 72 h and harvested for cell cycle analysis. (B) HK1‐inducible cell lines were treated with increasing doses of doxycycline and analyzed for cell cycle profiles after exposure to cytotoxic levels of cisplatin (IC90, 7 µM) for 72 h. (C) HK1 cells expressing LMP1 (50 ng/mL doxycycline) with or without cisplatin treatment were analyzed for cell division and cell cycle profiles with the CFSE cell tracker dye and propidium iodide costaining. The mean fluorescence intensity (MFI) value for each cell division is indicated for Generation 0 (G0), G3, and G5. *Cells harvested at the start of the assay, before exposure to cisplatin.

Figure 2

Cells arrested in G1 show lower levels of the DNA damage marker γH2AX. HK1 vector control and LMP1‐expressing cells (50 ng/mL doxycycline) were analyzed for γH2AX levels in different phases of the cell cycle, after exposure to 7 µM (IC90) of cisplatin for 72 h. The discontinued lines mark the mean fluorescence intensity for cells in G1 or G2–M. x and y axes are set to the same scale for all plots. N/A (top right panel), HK1 vector control cells treated with cisplatin do not yield a G1 population and therefore a change in MFI could not be determined.

Figure 3

LMP1 attenuates DNA damage response signaling and the removal of cisplatin‐DNA adducts. HK1 vector control and LMP1‐expressing cells (100 ng/mL doxycycline) were analyzed for (A) DNA damage response signaling and G1/S checkpoint proteins by immunoblot. (B) Residual cisplatin‐DNA adducts were analyzed by a genomic DNA dot blot and probed with anti‐cisplatin DNA adduct antibody. Shown is the background‐subtracted signal normalized to the starting signal at recovery time zero (R0).

Figure 4

LMP1 functional mapping by cell cycle profile and phosphorylation of p53 (S15) in response to cisplatin. (A) Schematic of LMP1 deletion and point mutants. LMP1‐inducible HK1 cells were analyzed for LMP1 expression by immunoblot for the HA‐tag. (B) Cell cycle profiles and (C) p53 activation were compared for the full‐length LMP1 (FL) and LMP1 A5, Y₃₈₄G mutant (50 ng/mL doxycycline) post‐cisplatin treatment (IC90, 7 µM for 72 h).

Figure 5

LMP1 sensitizes HK1 cells to cisplatin. (A) HK1 vector control or LMP1‐expressing cells were treated with increasing doses of cisplatin for 72 h. Shown are dose–response (best‐fit) curves from a representative experiment. The IC50 values and the dose–response curves by two‐way ANOVA analysis from three independent experiments are shown in the table. (B and C) Cisplatin sensitivity is phenocopied in HK1 vector control cells treated with the ATR (VE821 and AZD6738) or Chk1 (PF477736) inhibitors, but not with the ATM (KU55933 and AZD0156) or Chk2 (chk2i II) inhibitors. Shown is one representative experiment from three independent experiments quantified in the table.

Figure 6

LMP1 sensitizes HK1 cells after two consecutive rounds of treatment with cisplatin. (A) HK1 inducible cell lines with/without doxycycline induction (100 ng/mL, 24 h) were treated with 7 µM (IC90, 72 h) of cisplatin. Cells recovering after 1 week were treated with increasing doses of cisplatin for 72 h. Dose–response curves were generated by plotting the concentration of inhibitor (cisplatin) on a logarithmic scale vs. response (RFU normalized to untreated cells). The IC50 was defined as the concentration of cisplatin that results in a 50% reduction of the cell viability. (B) Sensitivity of the LMP1 A5, Y₃₈₄G mutant to cisplatin treatment was compared to the vector control or LMP1 FL. (C) Summary of LMP1 functional mapping phenotypes in response to cisplatin.

Figure 7

LMP1 sensitizes HK1‐EBV cells to cisplatin. HK1‐EBV stable cell lines expressing recombinant LMP1 China 1 (Ch1) strain or knockdown of endogenous LMP1 (by shRNA or CRISPR/Cas9) were treated with increasing doses of cisplatin for 72 h, in the absence of IL6 stimulation of the endogenous LMP1 promoter, with or without ATMi (KU55933, 30 µM). Shown are dose–response (best‐fit) curves from a representative experiment. Comparison of the average IC50 values and the dose–response curves by two‐way ANOVA analysis, from three independent experiments, are shown in the table.

Figure 8

Model of cisplatin‐induced DNA damage response (DDR) signaling with or without LMP1. Cisplatin generates interstrand and intrastrand DNA crosslinks. RPA32 binds to the lesion sites and recruits mainly ATR but also ATM. Upon recruitment to the lesion sites, these kinases are activated and trigger DDR signaling that includes phosphorylation of RPA32 and other targets such as Chk1, Chk2, and H2AX. Signaling mediated by ATM and ATR leads to checkpoint activation, cell cycle arrest and recruitment of repair proteins. If DNA damage is repaired, the cell cycle progresses. In NPC tumors where ATM expression is downregulated, LMP1 disrupts the response to DNA damage by attenuating the activation of ATR and Chk1, which results in reduced efficiency of DNA repair, persistent DNA damage, cell cycle arrest, and cytotoxicity.