The DNA damage response network in the treatment of head and neck squamous cell carcinoma

Abstract

Background: We sought to determine whether DNA damage response (DDR)-related aberrations predict therapeutic benefit in cisplatin-treated head and neck squamous cell carcinoma (HNSCC) patients and how DDR pathways are modulated after treatment with olaparib alone or in combination with cisplatin or durvalumab.

Patients and methods: Oxidative stress, abasic sites and DDR-related parameters, including endogenous DNA damage, DNA repair mechanisms and apoptosis rates, were evaluated in HNSCC cell lines and peripheral blood mononuclear cells from 46 healthy controls (HC) and 70 HNSCC patients at baseline and following treatment with cisplatin-containing chemoradiation or nivolumab or enrolled in the OPHELIA phase II trial (NCT02882308; olaparib alone, olaparib plus cisplatin, olaparib plus durvalumab).

Results: HNSCC patients at diagnosis exhibited deregulated DDR-related parameters and higher levels of oxidative stress and abasic sites compared with HC (all P < 0.05). Accordingly, nucleotide excision repair (NER; ERCC1, ERCC2/XPD, XPA, XPC) and base excision repair (APEX1, XRCC1) genes were downregulated in patients versus HC whereas double-strand breaks repair (MRE11A, RAD50, RAD51, XRCC2) and mismatch repair (MLH1, MSH2, MSH3) genes were overexpressed. Corresponding results were obtained in cell lines (all P < 0.001). Excellent correlations were observed between individual ex vivo and in vivo/therapeutic results, with cisplatin non-responders showing higher levels of endogenous DNA damage, augmented oxidative stress and abasic sites, increased NER capacities and reduced apoptosis than responders (all P < 0.05). Also, longer progression-free survival correlated with lower NER capacity (P = 0.037) and increased apoptosis (P = 0.029). Interestingly, treatment with olaparib-containing regimens results in the accumulation of cytotoxic DNA damage and exerts an extra antitumor effect by elevating oxidative stress (all P < 0.05). Nivolumab induced no significant changes in the DDR parameters examined.

Conclusions: Aberrations in DDR signals are implicated in the response to HNSCC chemotherapy and can be exploited as novel therapeutic targets, sensitive/effective non-invasive biomarkers as well as for the design of novel clinical trials.

Keywords: DNA damage response; cisplatin-containing chemoradiation; endogenous DNA damage; head and neck squamous cell carcinoma; olaparib-containing regimens; oxidative stress.

Conflict of interest statement

Disclosure The authors have declared no conflicts of interest. Data sharing Data are available upon reasonable request.

Copyright © 2021 The Author(s). Published by Elsevier Ltd.. All rights reserved.

Figures

Figure 1

DNA damage response in HNSCC cell lines. (A) Alkaline comet assay images of normal and HNSCC cell lines. (B) Bar charts showing distribution of the endogenous DNA damage measured by comet assay. (C) Typical images showing γH2AX staining using confocal microscopy. Upper images, γH2AX staining; middle, cell nuclei labeled with DAPI; bottom, merged. (D) Bar charts showing distribution of the endogenous DSBs measured by γH2AX staining. Bar charts showing distribution of (E) the oxidative stress and (F) AP-sites. (G) Southern blot of the cisplatin-induced N-ras-specific monoadducts in normal and HNSCC cell lines. Bar charts showing (H) the kinetics of monoadducts repair and (I) the accumulation of monoadducts following cisplatin treatment. N CL, normal cell line; HNSCC CL, HNSCC cell line. Bar charts showing (J) the kinetics of γH2AX foci formation/removal and (K) the accumulation of γH2AX foci after cisplatin treatment. (L) Bar charts showing distribution of the lowest concentrations of cisplatin required for the induction of apoptosis 24 h, 48 h and 72 h after cisplatin treatment. Error bars represent SD; ∗∗∗P < 0.001 by Mann–Whitney U test. The experiments shown were based on a minimum of three independent repeats. AP, abasic sites; AUC, area under the curve; CL, cell line; DAPI, 4′,6-diamidino-2-phenylindole; DSBs, double-strand breaks; HNSCC, head and neck squamous cell cancer; IS, internal standard; N CL, normal cell line; SD, standard deviation.

Figure 2

Endogenous DNA damage in PBMCs from HNSCC patients. (A) Alkaline comet assay images of two healthy controls (HC1, HC2) and two HNSCC patients at baseline. (B) Box plots showing statistical distribution of the endogenous DNA damage measured by comet assay in HC and HNSCC patients at baseline. (C) The immunofluorescence γH2AX staining of two HC and two HNSCC patients. (D) Box plots showing statistical distribution of the endogenous DNA damage in HC and patients at baseline. Box plots showing statistical distribution of the oxidative stress (E) and the AP-sites (F) in HC and patients at baseline. Error bars represent SD; ∗∗∗P < 0.001 by Mann–Whitney U test. The experiments shown were based on a minimum of three independent repeats. AP, abasic sites; DAPI, 4′,6-diamidino-2-phenylindole; HC, healthy control; HNSCC, head and neck squamous cell cancer; PBMCs, peripheral blood mononuclear cells; NR, non-responder; R, responder; SD, standard deviation.

Figure 3

DDR signals in PBMCs following ex vivo cisplatin treatment. (A) Southern blot of the ex vivo cisplatin-induced monoadducts in a HC and two HNSCC patients (R, responder; NR, non-responder). (B) The kinetics of monoadducts repair and (C) the statistical distribution of the accumulation of the ex vivo cisplatin-induced monoadducts, in HC and HNSCC patients at baseline. (D) The kinetics of γH2AX foci and (E) the statistical distribution of the ex vivo cisplatin-induced γH2AX foci accumulation in HC and HNSCC patients. (F) Box plots showing statistical distribution of the lowest concentrations of cisplatin required for the induction of apoptosis 24 h, 48 h and 72 h after ex vivo treatment with cisplatin. Error bars represent SD; ∗P < 0.05, ∗∗P < 0.01 and ∗∗∗P < 0.001 by Mann–Whitney U test. The experiments shown were based on a minimum of three independent repeats. AUC, area under the curve; DDR, DNA damage response; HC, healthy control; HNSCC, head and neck squamous cell cancer; IS, internal standard. NR, non-responder; PBMCs, peripheral blood mononuclear cells; R, responder.

Figure 4

Kaplan–Meier PFS curves and expression of DDR-associated genes. (A) Kaplan–Meier PFS curves according to response to cisplatin-containing chemoradiation. Kaplan–Meier curves demonstrating that longer PFS is associated with (B) lower NER resulting in higher DNA damage burden, (C) increased apoptosis, (D) lower endogenous DNA damage, and (E) decreased oxidative stress at baseline. (F) Hierarchical clustergram of 84 DDR-associated genes in eight HNSCC patients at baseline versus eight HC. (G, H) Genes demonstrating at least two-fold difference in the transcription activity between patients and HC. AUC, area under the curve; CI, confidence interval; DDR, DNA damage response; GSH, glutathione; GSSG, glutathione disulfide; HC, healthy control; HNSCC, head and neck squamous cell cancer; HR, hazard ratio; NER, nucleotide excision repair; PFS, progression-free survival.

Figure 5

DDR signals following various in vivo/therapeutic treatments. (A) Box plots showing the kinetics of monoadducts repair in patients following various therapeutic treatments. (B) Box plots showing statistical distribution of the accumulation of monoadducts after cisplatin-only therapy. (C) Correlation between monoadducts burden following therapeutic and ex vivo cisplatin treatment in the same patients. (D) Box plots showing the kinetics of γH2AX foci formation/removal following various therapeutic treatments. (E) Box plots showing the statistical distribution of the accumulation of γH2AX foci after cisplatin-only therapy. (F) Correlation between DSB burden following therapeutic and ex vivo cisplatin treatment in the same patients. Box plots showing statistical distribution of the oxidative stress (G) and the abasic sites (H) after various therapeutic treatments. Error bars represent SD; ∗P < 0.05, ∗∗P < 0.01 and ∗∗∗P < 0.001 by Mann–Whitney U test. The experiments shown were based on a minimum of three independent repeats. AUC, area under the curve; DDR, DNA damage response; DSB, double-strand break; GSH, glutathione; GSSG, glutathione disulfide; NR, non-responder; NS, not significant; PBMCs, peripheral blood mononuclear cells; R, responder; SD, standard deviation.