Purpose:

ERCC1/XPF is a DNA endonuclease with variable expression in primary tumor specimens, and has been investigated as a predictive biomarker for efficacy of platinum-based chemotherapy. The failure of clinical trials utilizing ERCC1 expression to predict response to platinum-based chemotherapy suggests additional mechanisms underlying the basic biology of ERCC1 in the response to interstrand crosslinks (ICLs) remain unknown. We aimed to characterize a panel of ERCC1 knockout (Δ) cell lines, where we identified a synthetic viable phenotype in response to ICLs with ERCC1 deficiency.

Experimental Design:

We utilized the CRISPR-Cas9 system to create a panel of ERCC1Δ lung cancer cell lines which we characterized.

Results:

We observe that loss of ERCC1 hypersensitizes cells to cisplatin when wild-type (WT) p53 is retained, whereas there is only modest sensitivity in cell lines that are p53mutant/null. In addition, when p53 is disrupted by CRISPR-Cas9 (p53*) in ERCC1Δ/p53WT cells, there is reduced apoptosis and increased viability after platinum treatment. These results were recapitulated in 2 patient data sets utilizing p53 mutation analysis and ERCC1 expression to assess overall survival. We also show that kinetics of ICL-repair (ICL-R) differ between ERCC1Δ/p53WT and ERCC1Δ/p53* cells. Finally, we provide evidence that cisplatin tolerance in the context of ERCC1 deficiency relies on DNA-PKcs and BRCA1 function.

Conclusions:

Our findings implicate p53 as a potential confounding variable in clinical assessments of ERCC1 as a platinum biomarker via promoting an environment in which error-prone mechanisms of ICL-R may be able to partially compensate for loss of ERCC1.

See related commentary by Friboulet et al., p. 2369

Translational Relevance

There remains a longstanding interest in understanding how to use tumor DNA-repair deficiencies to predict response to chemotherapy and immunotherapy. Perhaps the best example of such an endeavor was the testing of ERCC1 expression as a platinum biomarker in non–small cell lung cancers in an international, randomized phase III trial, which ultimately did not show clinical benefit for patients with low ERCC1 receiving platinum. This was surprising considering ERCC1/XPF activity is thought to be essential for repair of platinum-DNA damage. We have identified a synthetic viable phenotype of tolerance to cisplatin in a panel of ERCC1-deficient cells that appears to be partially controlled by p53, thus leading to the identification of a new subset of platinum-tolerant, ERCC1-deficient tumors. Furthermore, this differential response to platinums was also observed in 2 patient cohorts, suggesting p53 status as a potential confounding variable in clinical studies attempting to implement ERCC1 as a platinum biomarker.

The structure-specific endonuclease excision repair cross-complementation group 1 (ERCC1)/xeroderma pigmentosum group F (XPF) plays key roles in nucleotide excision repair (NER), interstrand crosslink repair (ICL-R), homologous recombination (HR) repair, and single-strand annealing pathways. Although the role of ERCC1/XPF in NER is well established, the totality of its specific functions in the processing and repair of interstrand crosslinks (ICL) has remained unclear (see refs. 1, 2 for review). ICLs are produced upon exposure to agents that covalently link bases in opposing strands of DNA, and endonucleases are required for cleavage of the phosphodiester backbone adjacent to ICLs in order to initiate repair (3). Much recent evidence indicates ERCC1/XPF is required for ICL-unhooking, whereas other work highlights additional roles for this complex in ICL-R downstream of unhooking (4–7).

Use of interstrand crosslinking agents, including cisplatin, remain a mainstay in the treatment of malignancies. Several mechanisms for resistance to platinum-based chemotherapy have been described, including loss of base excision repair and mismatch repair, decreased drug accumulation, increased sequestering by thiols, decreased apoptosis, and increased translesion synthesis (8–11). Another proposed mechanism of resistance to cisplatin involves increased expression of ERCC1/XPF observed both in vitro with cisplatin-sensitive/resistant cell lines and in relation to survival in patient samples (11–17). Work by our laboratory and others have shown that downregulation of ERCC1/XPF sensitizes cancer cells to cisplatin and that this sensitivity is related to a reduction in ICL and intrastrand adduct (ISA) repair (18). In addition, small molecule inhibitors of ERCC1/XPF can increase cisplatin sensitivity both in vitro and in vivo, indicating the potential of pharmacologically targeting ERCC1/XPF to enhance platinum efficacy (19, 20).

First identified as a potential biomarker for response to platinum-based chemotherapy in the late 1990s, low ERCC1 expression was observed in a relatively high amount of patient tumors including in non–small cell lung, head and neck cancers, and ovarian serous adenocarcinomas. Although preclinical data were promising, many challenges faced the clinical implementation of ERCC1 as the first platinum biomarker including problems with antibody specificity, splice variant expression, and conflicting results from clinical and preclinical studies. However, it is possible that an incomplete understanding of basic biological factors controlling sensitivity to ICL-inducing agents in the absence of ERCC1 may also have contributed to the failure of the ERCC1 clinical trials. What has become clear over the past 10 years is that a DNA-repair deficiency does not necessarily predispose to sensitivity to a particular drug. This is most notably observed with BRCA1/2 deficiencies in the context of PARP inhibition where loss of subsequent secondary factors is capable of making BRCA1/2-mutant tumors resistant to PARP inhibition. In the context of ICL-repair, recent evidence has shown that loss of p53, the deubiquitinase, USP48, or the Blm–RMI1–TOPIIIa signaling axis is capable of increasing resistance of Fanconi anemia (FA) deletion mutants to ICLs both in vitro and in vivo (21–23). In particular, these findings directly implicate increased reliance on DNA repair pathways to deal with interstrand crosslinks that would otherwise be unrepaired as a result of loss of canonical ICL-R.

In this study, we identified p53 status as at least a partial modifier of the sensitivity of ERCC1Δ cells to ICL-inducing agents. Here, we characterize a panel of ERCC1Δ lung cancer cell lines developed with CRISPR-Cas9. We describe a differential phenotype in sensitivity to cisplatin and mitomycin C (MMC) that appears to be correlated with p53 status where ERCC1Δ/p53WT cell lines exhibit hypersensitivity to ICL-inducing agents, but ERCC1Δ/p53mutant/null cells exhibit only mild sensitivity. Finally, we show evidence that tolerance to interstrand crosslinks with ERCC1 deficiency is supported by entry into S-phase and relies on BRCA1 and DNA-PKcs function. Together this evidence suggests that functional loss of p53 may allow for the uncovering of alternate repair mechanisms capable of at least partially overcoming the repair defects associated with loss of ERCC1/XPF activity thus leading to the identification of a new subset of cisplatin-tolerant ERCC1-deficient tumors. These findings have direct clinical ramifications for future studies of ICL-repair in human tumors as well as impacting any attempts to implement biomarkers for sensitivity to ICL-inducing agents in the future.

Cell lines and cell culture

H1299, H460, H522, H1703, H1650, H358 were all obtained from the ATCC, were tested for mycoplasma, and authenticated by the BioBanking and Correlative Sciences Core Facility at Karmanos Cancer Institute. A549 WT and ERCC1Δ cells were obtained from Jean-Charles Soria, Ken Olaussen, and Luc Friboulet (Gustave Roussy Cancer Center). OV2008 and C13* cells were obtained from Stephen B. Howell (University of California San Diego). A549 and OV2008 cells were not further authenticated or tested for mycoplasma. Cell lines were maintained for no greater than ∼15 passages during the course of experiments. H1703, H522, H460, OV2008, C13, H1650, H358, and H1299 cells were cultured in RPMI1640 (Dharmacon) media supplemented with 10% FBS (Atlanta Biologicals) and 1% penicillin/streptomycin (Dharmacon) and grown at 37°C in 5% CO2. A549 cells were cultured in DMEM (Dharmacon) supplemented with 10% FBS, 1% penicillin/streptomycin, 1% MEM, nonessential amino acids (Dharmacon), and 1% HEPES (Dharmacon).

CRISPR-Cas9–mediated gene knockout

Knockout experiments were performed essentially as previously described (Addgene plasmid 52961; ref. 20). Knockout clones were validated by Sanger sequencing unless otherwise stated. crRNA sequences are included in Supplementary Table S1.

Colony survival assays

Colony survival assays were performed as previously described (9). Cells were treated with cisplatin (Sigma-Aldrich) or MMC (Selleckchem) for 2 hours or gemcitabine (Selleckchem), camptothecin (Selleckchem), or etoposide (Selleckchem) for 4 hours in serum-free medium. Cells were treated with Palbociclib (Selleckchem), Ribociclib (Selleckchem), DNA-PK inhibitor (NU-7441; Selleckchem), or XL-413 (DBF4-dependent kinase inhibitor; Tocris) for 24 hours in complete medium. For UV-C treatment, 2,000 cells were seeded in 6-well plates and treated with the corresponding UV-C dose the following day. Plates were fixed and stained with crystal violet 3 days posttreatment and crystal violet was dissolved in 10% acetic acid and absorbance at 595 nm was measured using a SpectraMax M5 plate reader (Molecular Devices). IC50s were estimated using SigmaPlot 10.0 Software.

Viability assays

A total of 12,000 cells were seeded in 24-well plates. Cells were treated with cisplatin for 24 hours and allowed to grow for an additional 24 hours. Live/dead cells were counted using Trypan Blue exclusion and ≥100 cells were counted for each concentration.

Flow cytometry

Apoptosis was measured by flow cytometry using the PE Annexin V Apoptosis Detection Kit (BD Biosciences). Cell-cycle profiles were determined using the propidium iodide (PI) Flow Cytometry Kit (Abcam). For both assays, 5 × 105 cells were seeded in 10 cm plates. The following day cells were treated with 500 nmol/L cisplatin for 2 hours in serum-free medium and cells were allowed to grow for 48 hours (for apoptosis) or were collected at various time points (cell cycle). Flow cytometry was performed on a BD LSR II SORP Flow Cytometer (BD Biosciences). Data were analyzed using ModFit LT (Verity Software House) and FlowJo v10 (FlowJo, LLC).

Modified alkaline comet assay

Modified alkaline comet assays were performed essentially as previously described (9, 20, 24). Cells were seeded in 6-well plates so that they would be ∼70% to 90% confluent at the time of harvesting. H522 and H1299 cells were treated with cisplatin for 2 hours. Control and cisplatin-treated cells were then treated with 100 μmol/L H2O2 (Fisher Scientific) for 15 minutes immediately prior to harvesting by trypsinization at 0, 24, and 48 hours post-cisplatin treatment. Cells were embedded in 0.5% low-melting agarose (Fisher Scientific; Catalog No. BP165-25) and spread on slides coated with 1.5% Standard Low–mr Agarose (Bio-Rad; Catalog No. 162-0100) and allowed to solidify. After 10 minutes, slides were placed in 4°C lysis buffer (2.5 M NaCl, 100 mmol/L EDTA, 10 mmol/L Tris base, 1% Triton X-100, pH 10) for ∼1 hour. Excess buffer was removed and slides were placed in the electrophoresis tank with 4°C alkaline electrophoresis buffer (0.3 mol/L NaOH, 1 mmol/L EDTA) and incubated for 20 minutes. Slides were electrophoresed for 25 minutes at 300 mA (∼22–26 V). Slides were then incubated with 4°C neutralization buffer (0.4 mol/L Tris Base, pH 7.5) for 10 minutes. Slides were fixed in 95% ethanol for 10 minutes and allowed to dry followed by incubation with SYBR-Gold (Invitrogen). Slides were imaged with a Nikon epifluorescence microscope. Approximately 50 cells were analyzed per slide with Komet Assay Software 5.5F (Kinetic Imaging). ICLs were measured as the ratio of the median tail moment of the treated compared with the untreated sample where the ratio at 0 hours post-cisplatin treatment was normalized to 100% for each isogenic cell line.

Patient survival analysis

The Cancer Genome Atlas (TCGA) provisional lung adenocarcinoma cohort was utilized to assess the relationship of ERCC1 tumor expression and TP53 mutational status on patient outcomes (25, 26). Tumor genomic and patient outcomes data were accessed for these TCGA patients on cBioPortal (26, 27). Genomic data were cleaned and normalized prior to release as described previously (25, 26). ERCC1 expression was stratified into 2 groups, high and low, at the upper quartile of expression values. TP53 mutation status was also stratified into 2 categories, mutated or wild-type (WT), based upon the presence or absence of nonsynonymous mutations in the coding region of the gene as detected by whole-exome sequencing. Patient overall survival (OS) was modeled in R (version 3.4.3) using the Kaplan–Meier method and log-rank test. Because treatment data for the TCGA lung adenocarcinoma cohort is not publically available, we also analyzed the 2017 TCGA ovarian cancer data set. Patients with stage 3+4 disease who received a platinum agent were included in the analysis. Patients were selected based upon the presence or absence of a TP53 mutation and patients were stratified based upon ERCC1 expression using the Affymetrix probe ID: 203720_s_at and the “auto select best cutoff” function. Data were analyzed using KMPlotter (kmplotter.org/ovar/; ref. 28).

Immunofluorescence

Cells were seeded on coverslips and treated with 500 nmol/L cisplatin for 2 hours. Forty-eight hours posttreatment cells were fixed with 4% paraformaldehyde and permeabilized in 0.3% Triton-X in PBS. Cells were blocked in 10% FBS in 0.1% Triton-X in PBS for 1 hour and incubated with primary antibody for 1 hour and secondary antibody for 1.5 hours in 1% BSA/0.1% Triton-X. DNA was stained with 300 nmol/L DAPI for 5 minutes and coverslips mounted on slides with DakoCytomation Fluorescent Mounting Medium (Agilent) and sealed with nail polish. Slides were imaged with a Nikon epifluorescence microscope and images were analyzed using ImageJ software and the Find Maxima function. A minimum of 100 cells per group per experiment were analyzed. Antibodies are available in Supplementary Table S2.

shRNA knockdowns and re-expression of p53 and ERCC1

BRCA1 and BRCA2 shRNA was purchased from Sigma-Aldrich. Sequences and identification numbers included in Supplementary Table S1. ERCC1-202 cDNA was purchased from Genscript and was cloned into pCDH-puro lentiviral vector. shRNA lentivirus and ERCC1-202 lentivirus was produced and transductions were performed as previously described (29). TP53 cDNA was purchased from Origene (Catalog No. R200003). Cells were transfected with 2.5 μg/DNA per well with a final concentration of 3 μg/well Lipofectamine for 6 hours. Cells were allowed to rest for 24 hours, after which geneticin sulfate was added. Approximately 2 weeks posttransfection, cells were harvested to assess p53 expression and experiments were performed. Knockdown of BRCA1 and BRCA2 were validated by Western blot or qRT-PCR as previously described using the primers listed in Supplementary Table S1 (24).

Statistical analysis for cell line studies

Flow cytometry and modified alkaline comet assay data were analyzed by 2-sample t test. Data comparing the dose effects of cisplatin or MMC on p53 status stratified by ERCC1 WT and knockout were analyzed by 2-way ANOVA with interaction test. γH2AX foci data were analyzed by Wilcoxon rank-sum test. Drug response, apoptosis, and modified alkaline comet assay experiments were all performed at least 3, unless otherwise stated. Cell cycle and γH2AX foci formation experiments are presented as a representative result from 2 to 3 individual experiments.

Western blot analysis

Protein extraction and Western blot analysis were performed as previously described (24). Antibodies used for Western blot analysis are available in Supplementary Table S2.

ERCC1Δ cells exhibit 2 distinct phenotypes upon cisplatin treatment

We developed a panel of ERCC1Δ cell lines using CRISPR-Cas9 to assess differences in sensitivity to ICL-inducing drugs (Fig. 1A). ERCC1 has 4 known splice variants, which differ by single intron inclusion or single exon exclusion (30), so in order to generate a clean background for our studies, we designed a crRNA targeting ERCC1 exon 2, which is shared by all ERCC1 splice variants. We utilized lung cancer cell lines that differed in p53, EGFR, and K-Ras status (Supplementary Fig. S2A). As expected, loss of ERCC1 led to loss of XPF expression, because both factors generally require each other for stability. However, a self-dimerization–based mechanism for XPF stability in the absence of ERCC1 has been reported in biochemical studies and could explain why 1 H1299 ERCC1Δ clone did not have reduced XPF expression (Supplementary Fig. S2B; ref. 31). In addition, we received A549 ERCC1Δ cells for our investigations (Fig. 1A).

Figure 1.

Cisplatin and MMC sensitivity of a panel of ERCC1Δ lung cancer cell lines. A, Western blot analysis depicting ERCC1 and XPF expression in the WT and ERCC1Δ cell lines generated by CRISPR-Cas9 and the A549 WT and ERCC1Δ cells. B and C, Clustering of cisplatin clonogenicity assays of ERCC1 WT and ERCC1Δ cells by p53 status. D and E, Clustering of cisplatin viability assays of ERCC1 WT and ERCC1Δ cells by p53 status. F and G, Clustering of MMC clonogenicity assays of ERCC1 WT and ERCC1Δ cells by p53 status. Data analyzed by 2-way ANOVA. ***, P < 0.001.

Figure 1.

Cisplatin and MMC sensitivity of a panel of ERCC1Δ lung cancer cell lines. A, Western blot analysis depicting ERCC1 and XPF expression in the WT and ERCC1Δ cell lines generated by CRISPR-Cas9 and the A549 WT and ERCC1Δ cells. B and C, Clustering of cisplatin clonogenicity assays of ERCC1 WT and ERCC1Δ cells by p53 status. D and E, Clustering of cisplatin viability assays of ERCC1 WT and ERCC1Δ cells by p53 status. F and G, Clustering of MMC clonogenicity assays of ERCC1 WT and ERCC1Δ cells by p53 status. Data analyzed by 2-way ANOVA. ***, P < 0.001.

Close modal

Because ERCC1 is necessary for key aspects of NER, HR, and ICL-R, we expected that ERCC1 loss would hypersensitize cells to cisplatin and MMC. Interestingly, upon titration of cisplatin in clonogenic and viability assays, we saw 2 distinct phenotypes, hypersensitivity and modest tolerance. H522, A549, and H460 ERCC1Δ (p53WT) cells were all very sensitive to cisplatin in both clonogenic (IC50s ranging from ∼60 to 240 nmol/L) and viability assays (Supplementary Fig. S1A and S1C). These observations were validated in a second ERCC1Δ clone in H460 cells which showed the same hypersensitive phenotype (Supplementary Fig. S2B and S2C). Interestingly, loss of ERCC1 in H1650, H1703, H358, and H1299 cells (p53null/mutant) only resulted in modest increased sensitivity to cisplatin in clonogenic (IC50s ranging from ∼1.1 to 2.0 μmol/L) and viability assays (Supplementary Fig. S1B and S1D). These effects were validated with multiple ERCC1Δ clones developed with 2 crRNAs, and a XPFΔ clone in H1299 cells, suggesting this modest sensitivity is a true phenotype of loss of functional ERCC1-XPF (Supplementary Fig. S2B, S2D–S2F). We were able to fully restore resistance to cisplatin in H1299 ERCC1Δ cells and partially restore cisplatin resistance in H460 ERCC1Δ cells when ERCC1-202 was re-expressed (Supplementary Fig. S2G and S2H). We also observed this differential phenotype with MMC, a more potent inducer of interstrand crosslinks than cisplatin (Supplementary Fig. S3A and S3B). Conversely, we did not observe increased sensitivity of ERCC1Δ compared with ERCC1 WT cells with etoposide, gemcitabine or camptothecin in clonogenic assays, but both H460 and H1299 ERCC1Δ cells were sensitive to UV-C irradiation (Supplementary Figs. S3C and S3D; S4A and S4B).

We observed that this differential phenotype appeared to be correlated with p53 status and so we performed clustering analysis based upon ERCC1 and p53 status. No differential clustering was observed in ERCC1 WT cells stratified by p53 status in the clonogenic survival or viability assays after cisplatin or MMC treatment (Fig. 1B, D, and F). However, plotting all ERCC1Δ cell lines together displayed 2 distinct phenotypes that appeared to be correlated with p53 status, where p53WT/ERCC1Δ cells were significantly more sensitive to cisplatin and MMC compared with p53mutant/null/ERCC1Δ cells (Fig. 1C, E, and G).

Altering p53 status alters the differential sensitivity of ERCC1Δ cells to cisplatin

The differential phenotype of ERCC1Δ cells to cisplatin and MMC appeared to be correlated with p53 status, and we hypothesized that altering p53 status could alter or reverse the observed phenotypes. To test this hypothesis, we overexpressed WT p53 in the p53-null H1299 cell line and assessed clonogenic potential after cisplatin treatment (Fig. 2A; Supplementary Fig. S5A). We observed that expression of p53 in parental and ERCC1Δ cells increased sensitivity to cisplatin (WT: 4.4 μmol/L vs. 2.3 μmol/L and ERCC1Δ: 1.3 μmol/L vs. 0.4 μmol/L; Fig. 2A). To address whether loss of p53 could increase resistance of p53WT/ERCC1Δ cells to cisplatin, we utilized CRISPR-Cas9 to disrupt the TP53 gene in H460 and H522 parental and ERCC1Δ cells using a crRNA targeted to Exon 7 of TP53. Western blot analyses show loss of p53 in the H460 cell lines at steady-state levels and upon induction with Nutlin-3 (Supplementary Fig. S5A). In addition, we confirmed disruption of TP53 by DNA sequencing (Supplementary Fig. S5E). In H460 ERCC1Δ/p53* cells, we observed modest increased clonogenicity (∼2-fold) after platinum treatment compared with ERCC1Δ alone in shorter-duration colony assays (Fig. 2A). This fold difference could be dramatically enhanced (∼10-fold) by extending the length of the colony assay from 6 to 12 days (Supplementary Fig. S5B). Despite reports that H522 cells harbor a homozygous single-base deletion in codon 191 of the TP53 gene, we were not able to detect this deletion when sequencing exons 5 and 6 of TP53. Therefore, for the purposes of this study, we considered the H522 cell line p53 WT. The p53-edited H522 cell lines were validated by sequencing which showed that the TP53 alleles were disrupted by CRISPR-Cas9 in the H522 p53* cells, including an 8 amino acid in-frame deletion in 1 allele which would account for a slightly reduced molecular weight band near 50 kDa; we also observed the acquisition of a truncated p53 mutant near 25 kDa (Supplementary Fig. S5A and S5E). In H522 ERCC1Δ cells, TP53 was partially disrupted (Supplementary Fig. S5E). This would be consistent with Western blot analysis results showing induction of p53 in the ERCC1Δ/p53* clone (Supplementary Fig. S5A). Partial p53 disruption in H522 cells also resulted in increased colony formation after cisplatin treatment in the ERCC1Δ cells (Fig. 2A). Similar results were also observed in A549 WT and ERCC1Δ cells upon TP53 disruption (Supplementary Fig. S5C–S5E).

Figure 2.

Effects of p53 on sensitivity of ERCC1Δ cells to cisplatin. A, Clonogenic survival after cisplatin treatment in H1299 (p53-null and p53 re-expressed), H460 (p53 WT and knockout), and H522 (p53 WT and knockout) isogenic cell lines differing by p53 and ERCC1 status. n = 3; data plotted as average of 3 independent experiments ± SD. B, Cisplatin viability assays of H460 and H522 isogenic cell lines treated with escalating doses of cisplatin., n = 3; data plotted as average ± SD. C, Compilation of data for H460 and H522 cells from 3 independent flow cytometry experiments representing % Annexin-V positive cells ± single-dose cisplatin treatment. Data represented as average % Annexin-V positive cells ± SD. *, P < 0.05 measured by 2-sided t test. NS, no significance, P > 0.05.

Figure 2.

Effects of p53 on sensitivity of ERCC1Δ cells to cisplatin. A, Clonogenic survival after cisplatin treatment in H1299 (p53-null and p53 re-expressed), H460 (p53 WT and knockout), and H522 (p53 WT and knockout) isogenic cell lines differing by p53 and ERCC1 status. n = 3; data plotted as average of 3 independent experiments ± SD. B, Cisplatin viability assays of H460 and H522 isogenic cell lines treated with escalating doses of cisplatin., n = 3; data plotted as average ± SD. C, Compilation of data for H460 and H522 cells from 3 independent flow cytometry experiments representing % Annexin-V positive cells ± single-dose cisplatin treatment. Data represented as average % Annexin-V positive cells ± SD. *, P < 0.05 measured by 2-sided t test. NS, no significance, P > 0.05.

Close modal

We also assessed changes in viability of H460 and H522 isogenic cell lines with increasing doses of cisplatin by Trypan Blue live/dead assays. Strikingly, we saw that disruption of p53 in ERCC1Δ H460 cells increased the IC50 in viability assays ∼15-fold (100 nmol/L vs. 1.8 μmol/L; Fig. 2B). In H522 cells, partial disruption of TP53 in ERCC1Δ cells increased the IC50 in viability assays ∼2-fold (0.8 μmol/L vs. 3.3 μmol/L; Fig. 2B). We also performed flow cytometry-based analysis of apoptosis with H460 and H522 isogenic cells lines. All paired cell lines were treated for 24 hours with the IC50 dose of the ERCC1Δ determined in the live/dead assays and cells were allowed to grow for an additional 24 hours before proceeding with flow cytometry. In H460 and H522 ERCC1Δ cells, cisplatin treatment resulted in approximately 50% cell death as measured by the percent Annexin-V positive cells (Fig. 2C). In ERCC1Δ/p53* cells, loss of p53 conferred significant protection from cisplatin-induced apoptosis (Fig. 2C). In addition, the level of apoptosis observed in ERCC1Δ/p53* cells was not statistically different from p53* cells. These viability data appear to suggest that p53 loss is critical for limiting apoptosis in the presence of unrepaired ICLs with loss of ERCC1.

Next, we tested whether knockout of ERCC1 had differential effects on cisplatin sensitivity in the OV2008/C13* cell line model of cisplatin resistance. C13* cells exhibit increased levels of ERCC1 compared with OV2008 cells (Supplementary Fig. S5F). Although both cell lines possess WT p53, p53 induction in C13* cells is impaired and p53 is not stabilized upon platinum treatment (Supplementary Fig. S5G; refs. 32, 33). In clonogenic assays, we observe OV2008 ERCC1Δ cells are more sensitive to cisplatin than the C13* ERCC1Δ cells (Supplementary Fig. S5H). These data suggest that the differential phenotype may not be limited to p53 mutations but could be extended to include defects in p53 stability/induction.

Kinetics of the DNA damage response in ERCC1Δ cells

Our hypothesis regarding the role of p53 in this differential phenotype was that p53 predisposed repair-deficient cells to apoptosis and that loss of p53 promoted DNA damage tolerance. To assess general levels of DNA damage signaling in the ERCC1Δ cells, we performed a treatment time-course and measured levels of p53, CDKN1A (p21), PARP1 cleavage, and H2AX phosphorylation at various time points by Western blot analysis. In the cisplatin-hypersensitive H460 and H522 ERCC1Δ cells, we saw induction of PARP cleavage after treatment, consistent with the viability assays (Fig. 3A). In addition, we saw induction of p53 and p21 over time which persisted (Fig. 3A). Consistent with previously reported observations (5), we detected a large induction of γH2AX in the H460 and H522 ERCC1Δ cell lines which continued to the 48-hour time point indicating persistent, unrepaired DNA DSBs in the hypersensitive cells (Fig. 3A). H1299 ERCC1Δ cells exhibited very little induction of cleaved PARP after treatment with cisplatin (Fig. 3A). Unexpectedly, there was very little induction of γH2AX in H1299 ERCC1Δ cells compared with WT cells (Fig. 3A), indicating either 1. DNA damage signaling is defective, or that 2. DSB repair is not defective in these cells. The former possibility is unlikely, considering cells defective in phosphorylation of H2AX are sensitized to DNA damaging agents such as ionizing radiation (34–36). This evidence suggests differential responses to DNA damage may contribute to the bimodal phenotype observed in our panel of ERCC1Δ cells.

Figure 3.

Differential DNA damage signaling and interstrand crosslink repair in ERCC1Δ cells. A, Western blot analysis of a cisplatin treatment time course measuring induction of cleaved PARP, γH2AX, p53, and p21 up to 48 hours in H460, H522, and H1299 WT and ERCC1Δ cells. B, Images and quantification of 1 representative experiment showing yH2AX foci formation in H1299 and H460 cells 48 hours post-cisplatin treatment. ****, P < 0.0001 as measured by Wilcoxon rank-sum test; experiments performed 3 times. C, Modified alkaline comet assay data indirectly measuring ICL-R in H522 and H1299 cells. Data presented as mean of 3 independent experiments. Error bars represent SEM. P < 0.05 as measured by Student t test.

Figure 3.

Differential DNA damage signaling and interstrand crosslink repair in ERCC1Δ cells. A, Western blot analysis of a cisplatin treatment time course measuring induction of cleaved PARP, γH2AX, p53, and p21 up to 48 hours in H460, H522, and H1299 WT and ERCC1Δ cells. B, Images and quantification of 1 representative experiment showing yH2AX foci formation in H1299 and H460 cells 48 hours post-cisplatin treatment. ****, P < 0.0001 as measured by Wilcoxon rank-sum test; experiments performed 3 times. C, Modified alkaline comet assay data indirectly measuring ICL-R in H522 and H1299 cells. Data presented as mean of 3 independent experiments. Error bars represent SEM. P < 0.05 as measured by Student t test.

Close modal

We further confirmed γH2AX results from Western blot analysis by immunofluorescence. In H460 WT cells, cisplatin treatment did not result in increased γH2AX foci formation 48 hours post-cisplatin treatment, whereas in ERCC1Δ cells foci formation was dramatically increased (Fig. 3B). These data would suggest that DSB formation during ICL-R is at least partially independent of ERCC1/XPF activity. This would be consistent with previous reports in ERCC1-deficient cells showing other endonucleases are capable of the initial incision steps of ICL-R including Mus81 and Fan1 (reviewed in ref. 37). In H460 ERCC1Δ/p53* cells, we observed a significant reduction in the number of γH2AX foci present 48 hours after treatment (Fig. 3B). Conversely, in cisplatin-tolerant H1299 ERCC1Δ cells we observed very few γH2AX foci at 48 hours post-cisplatin treatment likely suggesting that ICL-R is largely not defective in these cells (Fig. 3B).

Cisplatin is a bifunctional drug, which induces ISAs between guanine residues on the same strand of DNA and these structures, which are repaired through NER, represent >90% of the damage caused by cisplatin whereas ICLs represent approximately 1% to 5% of total DNA adducts generated by cisplatin. We suspected that ISAs, which should persist in ERCC1Δ cells, may be less toxic than ICLs which can function as complete replication blocks. To tease apart whether the modest sensitivity we observe in H1299 ERCC1Δ cells in clonogenic assays is due to unrepaired ISAs, we generated XPAΔ clones in H1299 cells (Supplementary Fig. S6A and S6D). XPA's only described function is to act as a scaffolding protein during NER, and XPA-deficient cells are less sensitive to MMC than ERCC1-deficient cells (38). So, XPAΔ cells should display the relative contribution of unrepaired ISAs in H1299 cells. In clonogenic assays, XPAΔ cells display the same sensitivity to cisplatin as ERCC1- or XPF-deficient cells, strongly pointing to the modest sensitivity observed as the relative contribution of unrepaired ISAs (Supplementary Fig. S6B and S6E). We also measured sensitivity to MMC, which induces monoadducts and a higher level of ICLs than cisplatin. We observed a classic phenotype, where ERCC1 and XPF knockout cells were ∼2.5-fold more sensitive to MMC than XPAΔ cells (Supplementary Fig. S6C). We hypothesize that the relative amount of ICLs compared with the total amount of DNA damage may impact this differential phenotype with ERCC1 deficiency. To assess differences in ICL-R between our isogenic cell lines, we performed modified alkaline comet assays in the H522 and H1299 isogenic cell lines. Although this assay is an indirect measure of interstrand crosslinked DNA, it is commonly used to measure platinum ICL-R. H1299 cells were treated with 5 μmol/L and H522 cells with 1.5 μmol/L cisplatin for 2 hours followed by measurements of ICL-R at the 0- (immediately after treatment), 24-, and 48-hour time-points. Despite attempts to perform these analyses with H460 isogenic cell lines, the ERCC1Δ cells were too sensitive to cisplatin to observe significant differences between untreated and treated samples at the 0-hour time point when treated with 500 nmol/L cisplatin, making it impossible to accurately monitor ICL DNA repair via this assay. As we expected, H522 ERCC1Δ cells were not capable of ICL-R (Fig. 3C). In H522 ERCC1Δ/p53* cells, ICL-R was at least partially rescued where there was no difference compared with WT or p53* cells at the 24-hour time point, but statistically greater amounts of ICL DNA damage remained at 48 hours posttreatment relative to p53* cells and less amounts of damage remained relative to ERCC1Δ cells (Fig. 3C). This observation would be consistent with partial disruption of TP53 in these cells. Consistent with the DNA damage signaling we observed in H1299 ERCC1Δ cells, no delay in ICL-R compared with H1299 WT cells was detectable (Fig. 3C), however, re-expression of p53 in H1299 ERCC1Δ cells induced a near-complete block of ICL-R compared with parental cells (Fig. 3C).

Cell-cycle arrest profiles differ in ERCC1Δ cells after cisplatin treatment

Many reports with ERCC1 knockout and knockdown cells have shown that after treatment with a crosslinking agent there is a potent G2–M cell-cycle arrest which has been attributed to unrepaired DNA damage. To test whether p53 status affected cell-cycle arrest after treatment with cisplatin, we assessed cell-cycle profiles in a time course after treatment. Although the p53 null H1299 cells exhibited no distinct G2–M arrest after platinum treatment, we consistently observed a slight increase in G2–M arrest in the H1299 ERCC1Δ cells at the 24-hour time point, which resolved by the 48-hour time point (Fig. 4A and C). Conversely, in H460 ERCC1Δ cells, we observed a potent G2–M arrest 24 hours post-cisplatin treatment (Fig. 4B and C). By 48 hours posttreatment, we detected a sharp increase in a sub-G1 population consistent with an increase in cell death. In the H460 ERCC1Δ/p53*cells, we observed the same potent G2–M arrest that we observed in H460 ERCC1Δ cells at 24 hours posttreatment (Fig. 4B and C). Astonishingly, by 72 hours after treatment, we saw a near-complete recovery from G2–M arrest with only a minor increase in the sub-G1 population. We hypothesize that unrepaired DNA damage leads to G2–M arrest but that this G2–M arrest is not permanent. Eventually cells enter into M-phase and subsequently into G1 phase where the presence of DNA DSBs triggers p53-mediated cell death. However, in the absence of p53, cells are either capable of tolerating unrepaired DNA crosslink damage or alternate repair mechanisms may exist that can at least partially compensate for loss of ERCC1.

Figure 4.

Cell-cycle profiles after cisplatin treatment in ERCC1Δ isogenic cell lines. Representative data from flow cytometry experiments measuring cell cycle profiles in A. H1299 and B. H460 isogenic cells at varying time points after cisplatin treatment (n = 2 for each sample). C, Quantification of the percent of cells in G2–M phase following cisplatin treatment. Analysis excludes the sub-G1 population of cells from the quantification.

Figure 4.

Cell-cycle profiles after cisplatin treatment in ERCC1Δ isogenic cell lines. Representative data from flow cytometry experiments measuring cell cycle profiles in A. H1299 and B. H460 isogenic cells at varying time points after cisplatin treatment (n = 2 for each sample). C, Quantification of the percent of cells in G2–M phase following cisplatin treatment. Analysis excludes the sub-G1 population of cells from the quantification.

Close modal

p53 status may act as a confounding variable in clinical assessments of ERCC1 as a platinum biomarker

The potential for utilizing ERCC1 expression to predict clinical response to platinum-based chemotherapy has been extensively tested in multiple cancer types including lung and ovarian cancers with varying results (39–42). We wanted to assess whether p53 status may be a confounding variable. Utilizing the TCGA lung adenocarcinoma data set, we split patients into 2 groups; those whose tumors had WT TP53 and those whose tumors had any amino acid-changing mutation in TP53. Although we did not observe any significant difference in ERCC1 expression between groups (P = 0.156), when we stratified WT p53 tumors based upon ERCC1 high or low expression, we observed a significant ∼50% increase in median OS for patients with low ERCC1 (Fig. 5A, B, and D). However, in patients whose tumors had p53 mutations, no significant increase in median OS for patients with low compared with those with high ERCC1 was observed (Fig. 5C and D). Although TCGA lung adenocarcinoma treatment data are not publically available, nearly 100% of patients with lung adenocarcinoma receive a platinum agent during the course of treatment, and so we hypothesize that these data are contingent upon platinum treatment. We also corroborated these results in the TCGA ovarian cancer data set in terms of both progression-free and OS specifically in the context of platinum treatment (Supplementary Fig. S7). We did not specifically test the effects of including BRCA1/2 mutation status in our analysis; however, BRCA1/2 mutations are generally observed in a p53 mutant context where we did not observe a significant benefit for patients with low ERCC1 in the context of OS.

Figure 5.

ERCC1 expression and p53 status in relation to OS in the TCGA lung adenocarcinoma data set. A, Expression of ERCC1 in lung adenocarcinoma tumors delineated by p53 status. Data compared by Student t test. OS of lung adenocarcinoma patients whose tumor harbored B. WT p53 and stratified by ERCC1 expression, or (C) mutated p53 and stratified by ERCC1 expression. D, Combined model of OS of patients with lung adenocarcinoma accounting for p53 status and ERCC1 expression.

Figure 5.

ERCC1 expression and p53 status in relation to OS in the TCGA lung adenocarcinoma data set. A, Expression of ERCC1 in lung adenocarcinoma tumors delineated by p53 status. Data compared by Student t test. OS of lung adenocarcinoma patients whose tumor harbored B. WT p53 and stratified by ERCC1 expression, or (C) mutated p53 and stratified by ERCC1 expression. D, Combined model of OS of patients with lung adenocarcinoma accounting for p53 status and ERCC1 expression.

Close modal

Mechanistic characterization of ICL tolerance with ERCC1-deficiency

We hypothesized that in a p53mutant/null background where apoptosis and G1 checkpoint activation are diminished, alternate repair mechanisms may exist to deal with the damage from unrepaired ICLs that accumulate as result of loss of ERCC1. It is well documented that ERCC1/XPF activity is critical for ICL-R in G1 phase, where replication-dependent processes are not available for dealing with ICLs, so we tested whether transient inhibition of entry into S phase could sensitize ICL-tolerant ERCC1Δ cells to cisplatin. Palbociclib alone inhibited growth of H1299 WT and ERCC1Δ cells and corresponded with an increase in the percentage of cells in G1 phase compared with untreated cells (Supplementary Fig. S8C and S8D). For combination treatment experiments, we treated cells with cisplatin followed by 24-hour treatment with the CDK4/6 inhibitors, palbociclib and ribociclib. We observed that H1299 WT and ERCC1Δ cells could be sensitized to cisplatin in clonogenic assays even with transient inhibition of CDK4/6 activity (Fig. 6A; Supplementary Fig. S8B), which was very similar to what we observed with reexpression of p53 (Fig. 2A). The addition of palbociclib to hypersensitive H460 or H522 ERCC1Δ cells did not further enhance cisplatin sensitivity, although there were small increases in sensitivity for the parental cell lines (Supplementary Fig. S8A and S8B). Furthermore, blocking replication initiation in H1299 ERCC1Δ cells via inhibiting the DBF4-dependent kinase (DDK) with the inhibitor, XL413, could also sensitize to cisplatin (Fig. 6A). We take this to suggest that in hypersensitive ERCC1Δ cells, ICL-R is completely dependent upon ERCC1/XPF activity whether or not cells are in G1 or S–G2–M phases of the cell cycle. However, the increased sensitivity of H1299, H460, and H522 WT cells with CDK4/6 or DDK inhibition may indicate that timely entry into S-phase is also critical for supporting platinum resistance despite being DNA repair proficient. This requirement for S-phase entry appears to be exacerbated in p53 null H1299 ERCC1Δ cells where platinum sensitization by CDK4/6 or DDK inhibition indicates that ERCC1/XPF activity is critical for ICL-R in G1 phase in ICL-tolerant cells, and that entry into S-phase may lead to ERCC1/XPF-independent mechanisms for tolerating or repairing ICL-DNA damage.

Figure 6.

Molecular pathways contributing to cisplatin tolerance in p53-null cells with ERCC1 deficiency. A, Clonogenic survival assays of H1299 WT and ERCC1Δ cells treated with cisplatin and palbociclib, ribociclib, or DBF4-dependent kinase inhibitor. Left: Plot depicts 1 representative experiment (n = 3). Middle: Plot represents average of 3 independent experiments. Right: Plot represents 1 representative experiment (n = 3). B, Clonogenic survival of H1299 isogenic cells treated with cisplatin ± NU7441 (DNA-PKcs inhibitor; n = 3). C, Western blot analysis showing BRCA1 knockdown and clonogenic assays of H1299 WT and ERCC1Δ with shControl and shBRCA1 knockdown. D, Quantification of γH2AX foci formation 48 hours post-cisplatin treatment in H1299 WT and ERCC1Δ BRCA1 knockdown cells ± cisplatin treatment. Data are representative of 2 independent experiments. ****, P < 0.0001 as measured by Wilcoxon rank-sum test.

Figure 6.

Molecular pathways contributing to cisplatin tolerance in p53-null cells with ERCC1 deficiency. A, Clonogenic survival assays of H1299 WT and ERCC1Δ cells treated with cisplatin and palbociclib, ribociclib, or DBF4-dependent kinase inhibitor. Left: Plot depicts 1 representative experiment (n = 3). Middle: Plot represents average of 3 independent experiments. Right: Plot represents 1 representative experiment (n = 3). B, Clonogenic survival of H1299 isogenic cells treated with cisplatin ± NU7441 (DNA-PKcs inhibitor; n = 3). C, Western blot analysis showing BRCA1 knockdown and clonogenic assays of H1299 WT and ERCC1Δ with shControl and shBRCA1 knockdown. D, Quantification of γH2AX foci formation 48 hours post-cisplatin treatment in H1299 WT and ERCC1Δ BRCA1 knockdown cells ± cisplatin treatment. Data are representative of 2 independent experiments. ****, P < 0.0001 as measured by Wilcoxon rank-sum test.

Close modal

Next, we tested whether factors involved in nonhomologous end joining (NHEJ) and HR supported the tolerance observed in H1299 ERCC1Δ cells. Inhibition of DNA-PKcs activity with the inhibitor, NU7441, selectively increased sensitivity of ERCC1Δ H1299 cells (1.3 μmol/L vs. 0.5 μmol/L), but not parental cells, to cisplatin (Fig. 6B). In addition, we performed shRNA knockdown of BRCA1 in H1299 cells and assessed clonogenicity after cisplatin treatment. BRCA1 knockdown led to increased sensitivity in both H1299 WT and ERCC1Δ cells, consistent with critical roles for BRCA1 in regulating DSB end-resection and contributing to HR, translesion synthesis, and microhomology-mediated end joining (Fig. 6C). Interestingly, the effects of BRCA1 knockdown in sensitizing ICL-tolerant ERCC1Δ cells to cisplatin appear to be independent of BRCA2 as we only observed increased sensitivity in WT cells, but no increased sensitivity of ERCC1Δ cells to cisplatin (Supplementary Fig. S8E). Furthermore, BRCA1 knockdown led to a significant increase in the presence of γH2AX foci persisting 48 hours post-cisplatin treatment in the H1299 ERCC1Δ cells treated with cisplatin compared with ERCC1Δ alone (Fig. 6D). Although the mechanism underlying ICL-tolerance in a subset of ERCC1Δ cells is not entirely parsed out, it appears that this tolerance is dependent upon DNA-PKcs and BRCA1 function, but likely independent of BRCA2.

A recent international randomized phase III clinical trial utilizing ERCC1 expression to predict response to platinum-based chemotherapy did not show clinical benefit for patients with non–small cell lung cancer with low ERCC1 who received a platinum agent (40, 41). In addition, a preclinical study failed to show any correlation between pretreatment ERCC1 expression in ovarian cancers and response to platinum-based chemotherapy (42). Here we showed that our in vitro data may have direct clinical implications where we observed a clinical benefit for patients with lung adenocarcinoma and ovarian carcinoma with low ERCC1 only when WT p53 was retained. These data may provide an additional explanation for conflicting results from clinical studies as to the benefit of using ERCC1 expression to predict responders to platinum-based chemotherapy. Our data may have the greatest impact in cancer types where the p53 mutation rate is markedly high, such as lung adenocarcinoma (∼50% p53 mutant) and ovarian serous carcinoma (∼90%), where ERCC1 has been investigated as a platinum biomarker.

In this study, we characterized a panel of ERCC1Δ lung cancer cell lines. We identified a differentially sensitive phenotype of ERCC1Δ lung cancer cell lines that appears to be at least partially associated with p53 status. If cells harbored WT p53, the ERCC1 deletion clones exhibited hypersensitivity to crosslinking agents, whereas the p53mutant/null cell lines exhibited mild sensitivity. Viability after cisplatin treatment of ERCC1Δ cells was dramatically increased by disrupting p53 by CRISPR-Cas9. However, clonogenicity of ERCC1Δ cells increased by disrupting p53 whereas sensitivity was increased in ERCC1Δ/p53null cells following expression of TP53 cDNA. The modest increases in clonogenicity in isogenic ERCC1Δ cells with subsequent disruption of p53 compared with our panel of ERCC1Δ cells likely suggests additional factors, such as those involved in processing and stability of replication forks (RPA availability for example), may be critical for further enhancing clonogenicity in response to platinums with loss of ERCC1 (43). This would be most significant in terms of factors involved in response to intrastrand DNA damage which also theoretically requires ERCC1/XPF activity for resolution. A similar phenotype was observed by Feng and Jasin, where they observed a similar partial, but significant, rescue of clonogenicity of BRCA2-deficient cells with p53 loss, potentially suggesting that alterations in additional factors or pathways are critical for supporting clonogenic growth in tumors harboring a p53 mutation and loss of BRCA2 (44).

Interestingly, modified alkaline comet assays in H522 and H1299 cell lines showed differential repair of ICLs depending on p53 status. We also demonstrated that ERCC1Δ cells exhibit G2–M arrest following cisplatin treatment. p53 disruption in H460 ERCC1Δ cells did not alter the initial G2–M arrest observed in H460 ERCC1Δ cells, however, at 48 hours posttreatment there was a dramatic increase in a sub-G1 population in H460 ERCC1Δ cells that corresponded with a decrease in the G2–M population. This is opposed to the near complete abrogation of cell death and G2–M arrest in the H460 ERCC1Δ/p53* cells.

Of importance, transient inhibition of entry into S-phase sensitized ICL-tolerant ERCC1Δ/p53* cells to cisplatin, suggesting ERCC1/XPF is indeed critical for ICL-R in G1 phase and that the persistence of these unrepaired ICLs may trigger growth inhibition. This growth inhibition also suggests that entry into S-phase is critical for supporting ICL tolerance in these cells where there may be decreased dependence on ERCC1/XPF for ICL unhooking resulting in the accumulation of DNA DSBs. Based upon our data, it is likely that loss of downstream functions of ERCC1/XPF in ICL-R lead to persistent DSBs that are unresolved, at least initially, leading to G2–M arrest. However, cells eventually escape this arrest and enter into M and subsequently into G1 phase where p53 activity is critical for sensing persistent DNA damage from the previous round of the cell cycle and triggering apoptosis as well as activating the G1 checkpoint. It is most likely that loss of p53 leads to a decrease in this apoptotic potential and loss of G1 checkpoint activation which enables secondary, alternate repair pathways to at least partially contribute to ICL-R either in later stages of the cell cycle or in the subsequent G1 phase where a number of error-prone repair pathways, independent of BRCA2, may be available, including break-induced replication, microhomology-mediated end joining, and NHEJ. This hypothesis is supported by the observation that DNA-PKcs and BRCA1, but not BRCA2, are critical for supporting tolerance to ICLs in the absence of p53 and ERCC1. A similar phenotype with BRCA1 was recently reported in MMC-resistant, FANCC-deficient cells where BRCA1 function was critical for supporting MMC resistance in the absence of USP48, despite loss of canonical ICL-R (22). Together, these data suggest that in both p53 and ICL-R deficient cells there is the potential uncovering of alternate DNA repair or tolerance mechanisms for dealing with unrepaired ICLs that specifically relies upon DNA-PKcs and BRCA1 as well as entry into S phase.

Several other groups have identified similar differential phenotypes in vitro and in vivo with loss of factors involved in ICL-R that appear to be correlated with p53 status including Mus81, BRCA2, and FANCD2 (21, 44, 45), although a mechanism for this differential phenotype has not been described. However, the role of p53 in inducing apoptosis does not appear to fully account for this differential phenotype as a recent report showed in FANCD2-deficient mice that p53 loss completely rescued mice from FA symptoms specifically in the context of aldehyde-DNA ICLs (21). Although the abrogation of FA symptoms was certainly related to a reduction in apoptosis, the authors also observed a dramatic increase in chromosomal aberrations including deletions and translocations suggesting that loss of p53 may uncover an alternate, error-prone ICL-R pathway and that WT p53 serves to suppress this error-prone repair likely through its roles in controlling apoptosis and mediating cell-cycle control.

In conclusion, the work in this study characterizes a novel phenotype of ICL-tolerance in a subset of ERCC1-deficient cells and highlights the potential importance of p53 as a clinically relevant variable in studies evaluating ERCC1, and possibly other ICL-R factors, as a platinum biomarker. The surprising finding that DNA-PKcs and BRCA1 support this phenotype of resistance suggests there are repair mechanisms in place which can at least partially overcome the ICL-R defects associated with loss of ERCC1. Furthermore, it will be important to fully characterize the molecular mechanisms underlying this process and to expand the list of repair factors that are involved in supporting resistance to crosslinking agents despite loss of canonical ICL-R.

Better understanding mechanisms of resistance to DNA crosslinking agents in the context of DNA repair deficiencies may lead to the identification of novel targets for therapeutic intervention that could be developed to improve patient responses to platinum-based chemotherapy.

No potential conflicts of interest were disclosed.

Conception and design: W. Lei, G. Bepler, S.M. Patrick

Development of methodology: J.R. Heyza, W. Lei, D. Watza, G. Bepler, S.M. Patrick

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.R. Heyza, W. Lei, D. Watza, H. Zhang, J.B. Back, G. Bepler

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.R. Heyza, W. Lei, D. Watza, W. Chen, J.B. Back, G. Bepler, S.M. Patrick

Writing, review, and/or revision of the manuscript: J.R. Heyza, W. Lei, D. Watza, J.B. Back, A.G. Schwartz, G. Bepler, S.M. Patrick

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.R. Heyza, W. Lei, G. Bepler

Study supervision: S.M. Patrick

We wish to thank Jean-Charles Soria, Luc Friboulet, and Ken Olaussen for the A549 cells used in this study. We also thank Stephen B. Howell for the OV2008 and C13* cells used in this study. J.R. Heyza and D. Watza are supported by T32CA009531 and A.G. Schwartz is supported by NIH R01CA141769. Generous funding from Karmanos Cancer Institute. The Microscopy, Imaging and Cytometry Resources Core, Biobanking and Correlative Sciences Core, and Biostatistics Core are supported, in part, by NIH Cancer Center Support Grant No. P30CA022453 to the Karmanos Cancer Institute at Wayne State University.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Bergstralh
DT
,
Sekelsky
J
. 
Interstrand crosslink repair: can XPF-ERCC1 be let off the hook?
Trends Genet
2008
;
24
:
70
6
.
2.
Scharer
OD
. 
ERCC1-XPF endonuclease-positioned to cut
.
EMBO J
2017
;
36
:
1993
5
.
3.
Hashimoto
S
,
Anai
H
,
Hanada
K
. 
Mechanisms of interstrand DNA crosslink repair and human disorders
.
Genes Environ
2016
;
38
:
9
.
4.
De Silva
IU
,
McHugh
PJ
,
Clingen
PH
,
Hartley
JA
. 
Defining the roles of nucleotide excision repair and recombination in the repair of DNA interstrand cross-links in mammalian cells
.
Mol Cell Biol
2000
;
20
:
7980
90
.
5.
Niedernhofer
LJ
,
Odijk
H
,
Budzowska
M
,
van Drunen
E
,
Maas
A
,
Theil
AF
, et al
The structure-specific endonuclease Ercc1-Xpf is required to resolve DNA interstrand cross-link-induced double-strand breaks
.
Mol Cell Biol
2004
;
24
:
5776
87
.
6.
Klein Douwel
D
,
Boonen
RA
,
Long
DT
,
Szypowska
AA
,
Raschle
M
,
Walter
JC
, et al
XPF-ERCC1 acts in Unhooking DNA interstrand crosslinks in cooperation with FANCD2 and FANCP/SLX4
.
Mol Cell
2014
;
54
:
460
71
.
7.
Zhang
N
,
Liu
X
,
Li
L
,
Legerski
R
. 
Double-strand breaks induce homologous recombinational repair of interstrand cross-links via cooperation of MSH2, ERCC1-XPF, REV3, and the Fanconi anemia pathway
.
DNA Repair (Amst)
2007
;
6
:
1670
8
.
8.
Kelland
L
. 
The resurgence of platinum-based cancer chemotherapy
.
Nat Rev Cancer
2007
;
7
:
573
84
.
9.
Kothandapani
A
,
Sawant
A
,
Dangeti
VS
,
Sobol
RW
,
Patrick
SM
. 
Epistatic role of base excision repair and mismatch repair pathways in mediating cisplatin cytotoxicity
.
Nucleic Acids Res
2013
;
41
:
7332
43
.
10.
Loh
SY
,
Mistry
P
,
Kelland
LR
,
Abel
G
,
Harrap
KR
. 
Reduced drug accumulation as a major mechanism of acquired resistance to cisplatin in a human ovarian carcinoma cell line: circumvention studies using novel platinum (II) and (IV) ammine/amine complexes
.
Br J Cancer
1992
;
66
:
1109
15
.
11.
Ferry
KV
,
Hamilton
TC
,
Johnson
SW
. 
Increased nucleotide excision repair in cisplatin-resistant ovarian cancer cells: role of ERCC1-XPF
.
Biochem Pharmacol
2000
;
60
:
1305
13
.
12.
Britten
RA
,
Liu
D
,
Tessier
A
,
Hutchison
MJ
,
Murray
D
. 
ERCC1 expression as a molecular marker of cisplatin resistance in human cervical tumor cells
.
Int J Cancer
2000
;
89
:
453
7
.
13.
Bellmunt
J
,
Paz-Ares
L
,
Cuello
M
,
Cecere
FL
,
Albiol
S
,
Guillem
V
, et al
Gene expression of ERCC1 as a novel prognostic marker in advanced bladder cancer patients receiving cisplatin-based chemotherapy
.
Ann Oncol
2007
;
18
:
522
8
.
14.
Ceppi
P
,
Volante
M
,
Novello
S
,
Rapa
I
,
Danenberg
KD
,
Danenberg
PV
, et al
ERCC1 and RRM1 gene expressions but not EGFR are predictive of shorter survival in advanced non-small-cell lung cancer treated with cisplatin and gemcitabine
.
Ann Oncol
2006
;
17
:
1818
25
.
15.
Lord
RV
,
Brabender
J
,
Gandara
D
,
Alberola
V
,
Camps
C
,
Domine
M
, et al
Low ERCC1 expression correlates with prolonged survival after cisplatin plus gemcitabine chemotherapy in non-small cell lung cancer
.
Clin Cancer Res
2002
;
8
:
2286
91
.
16.
Olaussen
KA
,
Dunant
A
,
Fouret
P
,
Brambilla
E
,
Andre
F
,
Haddad
V
, et al
DNA repair by ERCC1 in non-small-cell lung cancer and cisplatin-based adjuvant chemotherapy
.
N Engl J Med
2006
;
355
:
983
91
.
17.
Bepler
G
,
Kusmartseva
I
,
Sharma
S
,
Gautam
A
,
Cantor
A
,
Sharma
A
, et al
RRM1 modulated in vitro and in vivo efficacy of gemcitabine and platinum in non-small-cell lung cancer
.
J Clin Oncol
2006
;
24
:
4731
7
.
18.
Arora
S
,
Kothandapani
A
,
Tillison
K
,
Kalman-Maltese
V
,
Patrick
SM
. 
Downregulation of XPF-ERCC1 enhances cisplatin efficacy in cancer cells
.
DNA Repair (Amst)
2010
;
9
:
745
53
.
19.
McNeil
EM
,
Astell
KR
,
Ritchie
AM
,
Shave
S
,
Houston
DR
,
Bakrania
P
, et al
Inhibition of the ERCC1-XPF structure-specific endonuclease to overcome cancer chemoresistance
.
DNA Repair (Amst)
2015
;
31
:
19
28
.
20.
Arora
S
,
Heyza
J
,
Zhang
H
,
Kalman-Maltese
V
,
Tillison
K
,
Floyd
AM
, et al
Identification of small molecule inhibitors of ERCC1-XPF that inhibit DNA repair and potentiate cisplatin efficacy in cancer cells
.
Oncotarget
2016
;
7
:
75104
17
.
21.
Garaycoechea
JI
,
Crossan
GP
,
Langevin
F
,
Mulderrig
L
,
Louzada
S
,
Yang
F
, et al
Alcohol and endogenous aldehydes damage chromosomes and mutate stem cells
.
Nature
2018
;
553
:
171
7
.
22.
Velimezi
G
,
Robinson-Garcia
L
,
Munoz-Martinez
F
,
Wiegant
WW
,
Ferreira da Silva
J
,
Owusu
M
, et al
Map of synthetic rescue interactions for the Fanconi anemia DNA repair pathway identifies USP48
.
Nat Commun
2018
;
9
:
2280
.
23.
Moder
M
,
Velimezi
G
,
Owusu
M
,
Mazouzi
A
,
Wiedner
M
,
Ferreira da Silva
J
, et al
Parallel genome-wide screens identify synthetic viable interactions between the BLM helicase complex and Fanconi anemia
.
Nat Commun
2017
;
8
:
1238
.
24.
Sawant
A
,
Floyd
AM
,
Dangeti
M
,
Lei
W
,
Sobol
RW
,
Patrick
SM
. 
Differential role of base excision repair proteins in mediating cisplatin cytotoxicity
.
DNA Repair (Amst)
2017
;
51
:
46
59
.
25.
Campbell
JD
,
Alexandrov
A
,
Kim
J
,
Wala
J
,
Berger
AH
,
Pedamallu
CS
, et al
Distinct patterns of somatic genome alterations in lung adenocarcinomas and squamous cell carcinomas
.
Nat Genet
2016
;
48
:
607
16
.
26.
Cancer Genome Atlas Research Network
. 
Comprehensive molecular profiling of lung adenocarcinoma
.
Nature
2014
;
511
:
543
50
.
27.
Gao
J
,
Aksoy
BA
,
Dogrusoz
U
,
Dresdner
G
,
Gross
B
,
Sumer
SO
, et al
Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal
.
Sci Signal
2013
;
6
:
pl1
.
28.
Gyorffy
B
,
Lanczky
A
,
Szallasi
Z
. 
Implementing an online tool for genome-wide validation of survival-associated biomarkers in ovarian-cancer using microarray data from 1287 patients
.
Endocr Relat Cancer
2012
;
19
:
197
208
.
29.
Sawant
A
,
Kothandapani
A
,
Zhitkovich
A
,
Sobol
RW
,
Patrick
SM
. 
Role of mismatch repair proteins in the processing of cisplatin interstrand cross-links
.
DNA Repair (Amst)
2015
;
35
:
126
36
.
30.
Friboulet
L
,
Postel-Vinay
S
,
Sourisseau
T
,
Adam
J
,
Stoclin
A
,
Ponsonnailles
F
, et al
ERCC1 function in nuclear excision and interstrand crosslink repair pathways is mediated exclusively by the ERCC1-202 isoform
.
Cell Cycle
2013
;
12
:
3298
306
.
31.
Das
D
,
Tripsianes
K
,
Jaspers
NG
,
Hoeijmakers
JH
,
Kaptein
R
,
Boelens
R
, et al
The HhH domain of the human DNA repair protein XPF forms stable homodimers
.
Proteins
2008
;
70
:
1551
63
.
32.
Fraser
M
,
Chan
SL
,
Chan
SS
,
Fiscus
RR
,
Tsang
BK
. 
Regulation of p53 and suppression of apoptosis by the soluble guanylyl cyclase/cGMP pathway in human ovarian cancer cells
.
Oncogene
2006
;
25
:
2203
12
.
33.
Fraser
M
,
Leung
BM
,
Yan
X
,
Dan
HC
,
Cheng
JQ
,
Tsang
BK
. 
p53 is a determinant of X-linked inhibitor of apoptosis protein/Akt-mediated chemoresistance in human ovarian cancer cells
.
Cancer Res
2003
;
63
:
7081
8
.
34.
Celeste
A
,
Difilippantonio
S
,
Difilippantonio
MJ
,
Fernandez-Capetillo
O
,
Pilch
DR
,
Sedelnikova
OA
, et al
H2AX haploinsufficiency modifies genomic stability and tumor susceptibility
.
Cell
2003
;
114
:
371
83
.
35.
Bassing
CH
,
Chua
KF
,
Sekiguchi
J
,
Suh
H
,
Whitlow
SR
,
Fleming
JC
, et al
Increased ionizing radiation sensitivity and genomic instability in the absence of histone H2AX
.
Proc Natl Acad Sci U S A
2002
;
99
:
8173
8
.
36.
Kao
J
,
Milano
MT
,
Javaheri
A
,
Garofalo
MC
,
Chmura
SJ
,
Weichselbaum
RR
, et al
gamma-H2AX as a therapeutic target for improving the efficacy of radiation therapy
.
Curr Cancer Drug Targets
2006
;
6
:
197
205
.
37.
Zhang
J
,
Walter
JC
. 
Mechanism and regulation of incisions during DNA interstrand cross-link repair
.
DNA Repair (Amst)
2014
;
19
:
135
42
.
38.
Niedernhofer
LJ
,
Garinis
GA
,
Raams
A
,
Lalai
AS
,
Robinson
AR
,
Appeldoorn
E
, et al
A new progeroid syndrome reveals that genotoxic stress suppresses the somatotroph axis
.
Nature
2006
;
444
:
1038
43
.
39.
Friboulet
L
,
Olaussen
KA
,
Pignon
JP
,
Shepherd
FA
,
Tsao
MS
,
Graziano
S
, et al
ERCC1 isoform expression and DNA repair in non-small-cell lung cancer
.
N Engl J Med
2013
;
368
:
1101
10
.
40.
Bepler
G
,
Williams
C
,
Schell
MJ
,
Chen
W
,
Zheng
Z
,
Simon
G
, et al
Randomized international phase III trial of ERCC1 and RRM1 expression-based chemotherapy versus gemcitabine/carboplatin in advanced non-small-cell lung cancer
.
J Clin Oncol
2013
;
31
:
2404
12
.
41.
Lee
SM
,
Falzon
M
,
Blackhall
F
,
Spicer
J
,
Nicolson
M
,
Chaudhuri
A
, et al
Randomized prospective biomarker trial of ERCC1 for comparing platinum and nonplatinum therapy in advanced non-small-cell lung cancer: ERCC1 Trial (ET)
.
J Clin Oncol
2017
;
35
:
402
11
.
42.
Rubatt
JM
,
Darcy
KM
,
Tian
C
,
Muggia
F
,
Dhir
R
,
Armstrong
DK
, et al
Pre-treatment tumor expression of ERCC1 in women with advanced stage epithelial ovarian cancer is not predictive of clinical outcomes: a Gynecologic Oncology Group study
.
Gynecol Oncol
2012
;
125
:
421
6
.
43.
Belanger
F
,
Fortier
E
,
Dube
M
,
Lemay
JF
,
Buisson
R
,
Masson
JY
, et al
Replication protein A availability during DNA replication stress is a major determinant of cisplatin resistance in ovarian cancer cells
.
Cancer Res
2018
;
78
:
5561
73
.
44.
Feng
W
,
Jasin
M
. 
BRCA2 suppresses replication stress-induced mitotic and G1 abnormalities through homologous recombination
.
Nat Commun
2017
;
8
:
525
.
45.
Pamidi
A
,
Cardoso
R
,
Hakem
A
,
Matysiak-Zablocki
E
,
Poonepalli
A
,
Tamblyn
L
, et al
Functional interplay of p53 and Mus81 in DNA damage responses and cancer
.
Cancer Res
2007
;
67
:
8527
35
.

Supplementary data