Comparison of ERCC1/XPF genetic variation, mRNA and proteinlevels in women with advanced stage ovarian cancer treatedwith intraperitoneal platinum

Julie A. DeLoiaa, Nikhil R. Bhagwatb,c, Kathleen M. Darcyd, Mary Strangee, Chunquio Tiane,Kevin Nuttalle, Thomas C. Krivakf,*, and Laura J. Niedernhoferb,c,**

aSchool of Public Health and Health Services, The George Washington University, 2300 Eye St.,NW, Washington, DC 20037, USAbDepartment of Microbiology and Molecular Genetics, University of Pittsburgh School ofMedicine, 523 Bridgeside Point II, 450 Second Avenue, Technology Center, Pittsburgh, PA15219, USAcUniversity of Pittsburgh Cancer Institute, 5117 Centre Avenue, Hillman Cancer Center 2.6,Pittsburgh, PA 15213-1863, USAdWomens Health Integrated Research Center at Inova Health System, 3289 Woodburn Road,Annandale, VA 22003, USAePrecision Therapeutics Inc, 2516 Jane Street, Pittsburgh, PA 15203, USAfDivision of Gynecologic Oncology, Magee Womens Hospital, University of Pittsburgh MedicalCenter, Pittsburgh PA, 15213

AbstractObjective—Approximately 20% of patients receiving platinum-based chemotherapy forepithelial ovarian cancer (EOC) are refractory or develop early recurrence. Identifying thesepatients early could reduce treatment-associated morbidity and allow quicker transfer to moreeffective therapies. Much attention has focused on ERCC1 as a potential predictor of response totherapy because of its essential role in the repair of platinum-induced DNA damage. The purposeof this study was to accurately measure protein levels of ERCC1 and its essential binding partnerXPF from patients with EOC treated with platinum-based therapy and determine if protein levelscorrelate with mRNA levels, patient genotypes or clinical outcomes.

Methods—ERCC1 and XPF mRNA and protein levels were measured in frozen EOC specimensfrom 41 patients receiving intraperitoneal platinum-based chemotherapy using reversetranscription polymerase chain reaction and western blots. Genotypes of common nucleotidepolymorphisms were also analyzed. Patient outcomes included progression free (PFS) and overallsurvival (OS).

© 2012 Elsevier Inc. All rights reserved*Corresponding author. Fax: +1 412 641 5417. tkrivak@mail.magee.edu (T.C. Krivak). **Corresponding author at: University ofPittsburgh Cancer Institute, 5117 Centre Avenue, Hillman Cancer Center, 2.6, Pittsburgh, PA 15213-1863. Fax: + 1 412 623 7761.niedernhoferl@upmc.edu (L.J. Niedernhofer)..

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.ygyno.2012.05.006.

Conflict of interest statement The authors declare no financial conflicts of interest.

Author status N.B., T.K., L.N. and J.D. designed the experiments. N.B., M.S. and K.N. acquired the data. K.D. and C.T. did thestatistical analysis. N.B., K.D, L.N and J. D. drafted the manuscript. All authors assisted with data interpretation, manuscript revisions,and approved the final version of this manuscript.

NIH Public AccessAuthor ManuscriptGynecol Oncol. Author manuscript; available in PMC 2012 December 11.

Published in final edited form as:Gynecol Oncol. 2012 September ; 126(3): 448–454. doi:10.1016/j.ygyno.2012.05.006.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

Results—Expression of ERCC1 and XPF were tightly correlated with one another at both themRNA and protein level. However, the mRNA and protein levels of ERCC1 were not positivelycorrelated. Likewise, none of the SNPs analyzed correlated with ERCC1 or XPF protein levels.There was an inverse correlation between mRNA levels and patient outcomes.

Conclusion—Neither genotype nor mRNA levels are predictive of protein expression. Despitethis, low ERCC1 mRNA significantly correlated with improved PFS and OS.

KeywordsDNA repair; Platinum therapy; Chemoresistance; Biomarker; Prognosis; Single nucleotidepolymorphism

IntroductionEpithelial ovarian cancer (EOC) remains the leading cause of death from gynecologicmalignancy in this country. The majority of women present with advanced stage disease andare treated with aggressive surgery followed by platinum/taxane-based chemotherapy [1,2].Seventy-five to 80% of women respond to this therapeutic regiment. The other 20–25% ofpatients are refractory, showing no response, or platinum resistance with cancer recurrencewithin 6 months of the end of treatment [3–5]. To date there is no way to predict who willrespond to platinum-based chemotherapy and who is resistant and will suffer an earlyrecurrence. Developing a strategy to identify patients who are resistant to standard first-linechemotherapy is a critical step toward improving treatment of patients with EOC [6,7].

The therapeutic effectiveness of platinum-based therapy is not fully understood, but itscytotoxicity is generally accepted to be mediated by the formation of platinum-DNA adducts[8]. Cisplatin forms primarily 1,2-intrastrand crosslinks between adjacent purines in DNA,and other adducts including 1,3 intrastrand crosslinks and interstrand crosslinks between thetwo strands of DNA [9]. The only mechanism for removing platinum intrastrand cross-linksfrom DNA is nucleotide excision repair (NER) [10]. This mechanism requires over 30proteins to recognize DNA adducts, excise them, and replace the missing nucleotides. TheERCC1-XPF heterodimer is an endonuclease essential for NER [11]. Approximately 5% ofthe DNA adducts formed by DNA cisplatin are interstrand crosslinks [9]. Interstrandcrosslinks are extremely cytotoxic and thought to drive the anti-proliferative effects ofcrosslinking agents used in chemotherapy[12]. Interstrand crosslinks are repaired via amechanism that is distinct from NER, but also requires the ERCC1-XPF nuclease [13]. ThusERCC1 and XPF are the only two proteins that are absolutely required for the repair of allplatinum-induced DNA adducts.

Based on this critical role in DNA repair, much effort has focused on whether ERCC1expression predicts tumor resistance to platinum therapy. In EOC, as with other tumor types,ERCC1 genotype, mRNA levels, and protein levels have been speculated to reflect thefunctional level of ERCC1-XPF nuclease and thereby cellular DNA repair capacity [14–26].In earlier studies, we demonstrated an association between ERCC1 genotype and clinicaloutcomes, which was especially pronounced in women treated through the intraperitoneal(IP) route [16]. Of note, the mechanism of ERCC1-XPF regulation has not yet beenestablished and could be at the transcriptional, translational or post-translational level.Further, it is not known if ERCC1-XPF is rate limiting for NER or interstrand crosslinkrepair. While this does not exclude the utility of either the genotype or expressionlevel ofERCC1 as a clinically useful biomarker, the correlation may be unrelated to DNA repair.

It is important to address these gaps in knowledge by quantitatively measuring ERCC1-XPFmRNA and protein levels in a single set of tumor samples and determining if there is any

DeLoia et al. Page 2

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

correlation between these parameters and clinical outcomes. The aim of this study thereforewas to utilize a well characterized tumor sample set to accurately measure ERCC1-XPFexpression protein levels by immunoblot and determine if there is any correlation withmRNA levels, single nucleotide polymorphisms (SNPs) and clinical outcomes to gainmechanistic insight into the contribution of ERCC1-XPF to tumor resistance to platinumchemotherapy.

MethodsPatients and tissue

Forty-one de-identified frozen ovarian tumor samples were obtained from the MageeWomens Health Tissue Bank of the University of Pittsburgh Medical Center. Inclusioncriteria consisted of women receiving IP platinum-based chemotherapy and a confirmeddiagnosis of EOC. Patients were evaluated every 3 months for the first 2 years after surgeryand then every 6 months for the next 3 years. Progression was defined by RECIST criteria ora doubling of CA125 from the laboratory normal. All samples and clinical data werecollected through an honest broker system from patients who had given informed consent.This study was approved by the University of Pittsburgh Institutional Review Board.

Statistical analysisProgression-free survival (PFS) and overall survival (OS) were measured from the date ofsurgery. PFS was the time until disease recurrence or death, whichever came first. OS wasthe time until death regardless of causes. Patients were grouped into three subgroups basedon the level of ERCC1 or XPF mRNA or protein expression (low, mid and high), with eachsubgroup including approximately the same number of patients. The Kaplan-Meierprocedure was used to estimate the PFS and OS and the log-rank test was used to comparethe group-difference in survival distributions. A Cox proportional hazards model was usedto estimate the hazard ratio (HR) adjusted for age and stage. Associations between ERCC1/XPF genotype and mRNA/protein expression were evaluated using Wilcoxon rank-sum test.Spearman rank correlation was calculated to measure the relationship between ERCC1 andXPF expression.

DNA isolation and genotypingGenomic DNA was isolated from tissue using the Purgene Genomic DNA Purification Kit(Gentra, Minneapolis, MN) as described by the manufacturer's instructions. Seesupplemental data for genotyping details.

Measurement of ERCC1 and XPF mRNATotal RNA was isolated from frozen tumor samples using Trizol reagent (Invitrogen;Carlsbad, CA) according to the manufacturer's instructions. Following isolation, DNA wasremoved by treatment with DNaseI (Invitrogen). RNA quality and quantity were determinedby measuring absorption at 280 and 260 nm. cDNA was synthesized from 2 μg of totalRNA using the High Capacity kit (Applied Biosystems; Foster City, CA). Expression ofERCC1 (Assay ID # Hs00157415) and XPF (Hs00193342) was determined by real-timepolymerase chain reaction (PCR) on an ABI Prism 7700 Sequence Detection System,according to the manufacturer's instructions (Applied Biosystems). PCR was performedusing TaqMan® Gene Expression Assays (Applied Biosystems), which contain TaqManunlabeled primers and a FAM (fluorescein) dye-labeled MGB (minor groove binder) probefor target genes. The GAPDH gene (Assay ID #Hs99999905) was used as an endogenousreference for all samples. All amplification cycles were performed in a single 50 μl reactionwith cDNA equivalent to 100 ng of total RNA. A typical 50 μl reaction sample contained 25

DeLoia et al. Page 3

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

μl of TaqMan Universal PCR Master Mix (containing 1X TaqMan buffer, 200 μM dNTPs,5 mM MgCl2, 1.25U AmpliTaqGold, and 0.5U of Amperase uracil-N-glycosylase (UNG)along with the TaqMan primers and MGB probes. Thermal cycling conditions were 2min at50°C and 10min at 95 °C followed by 45 cycles at 95 °C for 30s and 60°C for 1min.Relative expression was determined using the comparative CT (Cycle-Threshold) method(Applied Biosystems, 2001), which consists of normalization of the number of target genecopies (ERCC1 and XPF) to an endogenous reference gene (GAPDH). All experiments wereconducted in duplicate and an average CT (Cycle-Threshold) value calculated for thereplicates±the standard error of the mean. The CT value refers to the cycle number whereinthe fluorescent intensity crosses the threshold line, which is set in the exponential phase ofthe amplification plot above background levels.

Measurement of ERCC1 and XPF proteinOf the 41 patients, there was adequate frozen tumor specimen for quantitation of ERCC1and XPF protein by immunoblot from 25 patients. Frozen samples were sonicated tohomogenize the tissue in 1 mL of Laemmli's buffer (10% Glycerol, 50 mM Tris-HCl pH 6.8and 2% SDS) containing 7 M urea, 5 mM dithiothreitol, 0.5 mM phenylmethyl sulfonylfluoride, and 1 μg/ml each of protease inhibitors, leupeptin, aprotinin and pepstatin. Thecrude homogenate was cleared of debris by centrifugation at 10,000 rpm for 15 min at 4 °C.The supernatant was stored in aliquots at −80 °C until analyzed. 50 μg of total protein fromeach sample was resolved by 10% SDS-PAGE and transferred to a nitrocellulose membrane.The membrane was cut in half lengthwise. XPF was detected in the upper portion of themembrane (MW 120 kDa) with primary antibody Ab-1 (mouse monoclonal, Neomarkers,1:1000) and alkaline phosphatase (AP)-conjugated goat anti-mouse secondary antibody(1:7500, Promega). ERCC1 was detected in the lower portion of the membrane (MW 37kDa) with primary antibody FL-297 (rabbit polyclonal, Santa Cruz, 1:1000) and AP-conjugated goat anti-rabbit secondary antibody (1:7500, Promega). The loading control, β-actin, was also detected in the lower portion of the membrane (MW 45 kDa) with primaryantibody ab13822 (chicken monoclonal, 1:1000, Abcam,) and AP-conjugated goat anti-chicken secondary antibody (1:1000; Abcam). Recombinant ERCC1-XPF [46] was used asan internal standard for the electrophoretic mobility of ERCC1 and XPF. Protein levels werequantitated by densitometry and corrected for loading using the actin control.

ResultsTumor specimens

Frozen samples of tumors were obtained from 41 patients presenting to Magee WomensHospital with EOC and who consented to having their tissue deposited in the University ofPittsburgh Tissue Bank. Patient characteristics are shown in Table 1. Forty of these patientshad Stage III–IV disease. Specimens were collected during their initial debulking surgery.All patients were treated with platinum-based chemo-therapy immediately followingsurgery. Thirty-three patients completed all 6 cycles of combination IV/IP therapy. Allpatients had at least one cycle of IV/IP chemotherapy. Patients were changed fromcombination IV/IP chemotherapy do to port complications, hemataologic toxicity,abdominal pain, and worsening fatigue. The 8 patients that required discontinuation of IV/IPtherapy all completed a total of 6 cycles of chemotherapy utilizing IV carboplatin andpaclitaxel.

ERCC1 and XPF mRNA levels in ovarian cancerERCC1 and XPF mRNA was measured by qRT-PCR using total RNA isolated from frozentumor specimens using GAPDH mRNA levels as a control (Fig. 1A). There was tremendousvariability in the level of ERCC1 mRNA between tumors (range: 0.91–80; 25th–75th

DeLoia et al. Page 4

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

percentile: 2.6–11.5) and to a lesser extent XPF mRNA (range: 0.27–6.0; 25th–75th pct:0.81–2.6). The median level of ERCC1 mRNA expression was 6.0 and the median level ofXPF mRNA expression was 1.3. There was a highly significant correlation between thelevels of ERCC1 and XPF mRNA (Fig. 1B; r=0.83, p<0.001).

ERCC1 and XPF protein levels in ovarian cancerProtein extracts were prepared from frozen tumor samples and ERCC1 and XPF proteinlevels measured by immunoblot for 25 samples (the subset of tumors for which adequatefrozen material was available) using actin as a loading control (Fig. 2A). Similar to mRNAexpression, the level of ERCC1 and XPF protein expression was variable between tumors(Fig. 2B) and there was a highly significantly correlation between ERCC1 and XPF proteinlevels (Fig. 2C; r=0.85, p<0.001). Surprisingly, there was a significant negative correlationbetween ERCC1 mRNA and protein levels (Fig. 3A and C; r=−0.41, p=0.048). There wasno evidence that XPF mRNA level was positively or negatively associated with XPF proteinlevel (Fig. 3B and D). These data provide definitive evidence that mRNA levels of ERCC1and XPF cannot be used to predict protein levels.

Genetic variant analysisTwo common polymorphisms in both ERCC1 (C/T codon 118 and C8092A) and XPF(G1244A and T2505C) were also analyzed. In order to determine the impact of eachpolymorphism on ERCC1 and XPF expression, we limited our analysis to patients who werehomozygous for the common or variant allele (Table 2). There were approximately an equalnumber of patients homozygous for the C and T alleles of codon 118 (8 vs. 11), but the A/Agenotype at position 8092 was much rarer than C/C (3 vs. 20). Sequencing of XPF revealedonly five patients homozygous for the C allele at position 2505, while 22 had a T/Tgenotype. None of the patients were homozygous for A at position 1244 of XPF. ERCC1and XPF mRNA and protein levels were averaged for all patients homozygous for eachSNP. There was not a significant difference in the ERCC1 mRNA or protein levels betweenpatients with either genotype for the two ERCC1 SNPs. Likewise there was no difference inthe XPF mRNA or protein levels between patients with the C/C or T/T genotype at T2505Cof XPF.

Patient survivalPFS and OS were calculated from the date of their initial surgical debulking. Thirty-three ofthe forty-one patients had disease progression, 17 of whom died by the time of analysis. Themedian duration of follow-up for living patients was 30 months (range: 19–46 months).Overall, the median PFS was 14.3 months (95% CI: 11.2–19.6) and median OS was notreached as of this analysis.

For survival analysis, patients were grouped into three subgroups based on the level ofERCC1 or XPF mRNA expression (low, mid and high; ERCC1: <3.0, 3.0–8.9, ≥9.0; XPF:<1.0, 1.0–1.9, ≥2.0), with each subgroup including approximately the same number ofpatients. High ERCC1 mRNA expression significantly correlated with shorter PFS(p=0.040) and shorter OS (p=0.006) (Fig. 4A and B). After adjusting for age and stage, aone-fold elevation of ERCC1 level was associated with a 26% increase in the risk of diseaseprogression (HR = 1.26, 95% CI = 0.96–1.65, p = 0.093) or 73% increase in the risk of death(HR=1.73, 95% CI=1.16–2.58, p=0.007). There was also a suggestion that higher XPFmRNA expression could be associated with shorter PFS (p=0.204) and OS (p=0.271), butthe associations were not statistically significant (Fig. 4C and D). In contrast, neitherERCC1 nor XPF protein expression level significantly correlated with PFS or OS regardlesswhether patients were grouped or not (data not shown). Collectively, these expression data

DeLoia et al. Page 5

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

support the conclusion that qRT-PCR is the best method for measuring ERCC1 expressionin tumors as a potential predictor of patient response to platinum-based therapy and survival.

DiscussionThere is a need to identify strategies to predict patient response to chemotherapy in order toavoid unnecessary toxicity and improve patient outcomes in cancer. ERCC1-XPF isinvolved in the two repair pathways required to remove cisplatin-induced DNA damage.Therefore this endonuclease is an attractive potential biomarker to predict whether or notplatinum-based therapy will kill tumor cells. There is an extensive and confusing body ofliterature on ERCC1 expression levels in cancer due in part to the fact that is not knownwhether the protein expression is regulated at the transcriptional or post-translational level.Furthermore, the 8 F1 antibody used in numerous clinical trials to measure ERCC1 proteinlevels is non-specific [27,28]. In this study, we took advantage of a unique resource (frozentumor specimens from a cohort of ovarian cancer patients all of whom received platinum-based therapy and outcomes were known) to definitively quantify protein levels of ERCC1and its essential binding partner XPF. The data are novel in that protein levels, mRNA levelsand genotypes of both genes were analyzed in a single set of tumors allowing not only theevaluation of correlation of these variables with patient outcome, but also with one other.

ERCC1 and XPF are unstable in the absence of one other in vivo [29–32]. This implies thatERCC1 and XPF only form heterodimers with one another and do not have separatefunctions. This is reinforced by the fact that Ercc1−/− and Xpf−/− mice have identicalphenotypes [30,33]. The second implication is that in clinical samples ERCC1 and XPFprotein levels should be significantly correlated. This is confirmed in our study (Fig. 2),where protein levels were quantitatively assessed by immunoblot with appropriate loadingcontrols. To our knowledge this is the only study to date measuring ERCC1-XPF proteinlevels in frozen tumor samples by this method using validated antibodies [28].

The C→T polymorphism at ERCC1 codon 118 (Asn) changes the codon from AAC, acommonly used codon, to AAT, which is infrequently used in mammalian cells. This couldretard translation, potentially leading to a reduction in ERCC1 protein level [34]. However,there is no significant difference in ERCC1 protein levels in tumors with a genotype of C/Cor T/T (Table 2). The C8092A ERCC1 SNP is intronic. Thus, there may be differences inERCC1 mRNA stability between genotypes. Despite this prediction, we were unable todetect a significant difference in mRNA or protein levels between the C/C and A/Agenotypes, though our sample size was small. The two XPF SNPs examined in this study arein the coding sequence of the gene. G1244A leads to an amino acid change from arginine toglutamine. It was too rare to analyze in our cohort. No correlation was detected between theT2505 genotype, a silent polymorphism in a serine residue, and mRNA or protein levels.Thus there is no evidence that these frequently studied SNPs in ERCC1 and XPF affectexpression of the DNA repair enzyme.

Based on the highly significant correlation between ERCC1 and XPF protein levels,measuring either could potentially inform about the DNA repair capacity of a tumor ortissue. However, there was no significant correlation between ERCC1 or XPF protein levelsand overall survival or progression-free survival in epithelial ovarian carcinoma treated withplatinum (Fig. 4). This lack of association may be because ERCC1-XPF levels are not rate-limiting for DNA repair or that DNA repair is not driving platinum resistance. Additionally,the evaluation of only one method of DNA repair may be of limited clinical value.Therefore, using a combination of biomarkers evaluating the different DNA repairmechanisms may add more specific tumor information and lead to a predictive modelidentifying patients whose tumors may respond to or be resistant to platinum based therapy.

DeLoia et al. Page 6

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

In contrast to protein expression, the one measure that did correlate with improved patientoutcomes (OS and PFS) was low ERCC1 mRNA expression. This correlation waspreviously reported in multiple tumor types including ovarian and lung carcinoma, whileothers failed to detect a correlation [20–22,24,25,35]. These prior studies invoked amechanism whereby higher mRNA levels were speculated to reflect higher ERCC1 protein,increased DNA repair, and thereby cisplatin resistance [24,36]. However, we establishedthat ERCC1 mRNA levels are negatively correlated with ERCC1 protein levels (Fig. 3A andC). Furthermore, there was no significant correlation between ERCC1 protein levels andpatient outcome. Thus, the association between mRNA levels and patient outcomes could bean artifact of the sample size and warrants further investigation. An alternative explanationfor the correlation between ERCC1 mRNA levels and clinical outcomes could be indirect.For example, mRNA levels of ERCC1 may be regulated at least in part by RAS [37]. RASactivation is linked to aggressive EOC in mice [38]. Interestingly, RAS transformationpromotes resistance to cisplatin [39]. Therefore, it is possible that increased ERCC1 mRNA,correlating with poor patient outcomes, is simply a surrogate marker for RAS activation intumors.

In conclusion, ERCC1 mRNA levels in tumors correlated with outcome in patients withEOC treated with IP cisplatin retrospectively. The next critical step is determining if ERCC1mRNA levels can be used prospectively to identify which EOC patients should receiveplatinum therapy. It must be emphasized that the mechanism behind the predictive value ofERCC1 mRNA expression remains unknown. Higher ERCC1 and XPF mRNA levels do notreflect higher protein levels. Thus, the link between high ERCC1 mRNA and poor PFS andOS is most likely not mediated by ERCC1-XPF dependent DNA repair.

Supplementary MaterialRefer to Web version on PubMed Central for supplementary material.

AcknowledgmentsWe thank Dr. Richard D. Wood, Grady F. Saunders Distinguished Professor in Molecular Biology, The Universityof Texas M. D. Anderson Cancer Center, Smithville, TX for generously providing recombinant ERCC1-XPF forimmunoblotting.

Financial dsclosure N.B. and L.J.N. were supported by the National Institutes of Health ES016114 andES016114-03S1. M.S., T.K. and J.D. were supported by the University of Pittsburgh Medical Center MageeWomens Hospital. K.D. and C.T. were supported by the Gynecologic Oncology Group Statistical and Data Center,Roswell Park Cancer Institute. The funders had no role in study design, data collection and analysis, decision topublish, or preparation of the manuscript.

References[1]. Vergote I, van Gorp T, Amant F, Leunen K, Neven P, et al. Timing of debulking surgery in

advanced ovarian cancer. Int J Gynecol Cancer. 2008; 18(Suppl. 1):11–9. [PubMed: 18336393]

[2]. Escobar PF, Rose PG. Docetaxel in ovarian cancer. Expert Opin Pharmacother. 2005; 6:2719–26.[PubMed: 16316310]

[3]. Piccart MJ, Bertelsen K, James K, Cassidy J, Mangioni C, et al. Randomized intergroup trial ofcisplatin-paclitaxel versus cisplatin-cyclophosphamide in women with advanced epithelialovarian cancer: three-year results. J Natl Cancer Inst. 2000; 92:699–708. [PubMed: 10793106]

[4]. McGuire WP, Hoskins WJ, Brady MF, Kucera PR, Partridge EE, et al. Cyclophosphamide andcisplatin compared with paclitaxel and cisplatin in patients with stage III and stage IV ovariancancer. N Engl J Med. 1996; 334:1–6. [PubMed: 7494563]

[5]. Ozols RF, Bundy BN, Greer BE, Fowler JM, Clarke-Pearson D, et al. Phase III trial of carboplatinand paclitaxel compared with cisplatin and paclitaxel in patients with optimally resected stage III

DeLoia et al. Page 7

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

ovarian cancer: a Gynecologic Oncology Group study. J Clin Oncol. 2003; 21:3194–200.[PubMed: 12860964]

[6]. Chang A. Chemotherapy, chemoresistance and the changing treatment landscape for NSCLC.Lung Cancer. 2011; 71:3–10. [PubMed: 20951465]

[7]. Hofler H, Langer R, Ott K, Keller G. Prediction of response to neoadjuvant chemotherapy incarcinomas of the upper gastrointestinal tract. Adv Exp Med Biol. 2006; 587:115–20. [PubMed:17163161]

[8]. Rabik CA, Dolan ME. Molecular mechanisms of resistance and toxicity associated withplatinating agents. Cancer Treat Rev. 2007; 33:9–23. [PubMed: 17084534]

[9]. Dronkert ML, Kanaar R. Repair of DNA interstrand cross-links. Mutat Res. 2001; 486:217–47.[PubMed: 11516927]

[10]. Martin LP, Hamilton TC, Schilder RJ. Platinum resistance: the role of DNA repair pathways.Clin Cancer Res. 2008; 14:1291–5. [PubMed: 18316546]

[11]. Sijbers AM, de Laat WL, Ariza RR, Biggerstaff M, Wei YF, et al. Xeroderma pigmentosumgroup F caused by a defect in a structure-specific DNA repair endonuclease. Cell. 1996; 86:811–22. [PubMed: 8797827]

[12]. McHugh PJ, Spanswick VJ, Hartley JA. Repair of DNA interstrand crosslinks: molecularmechanisms and clinical relevance. Lancet Oncol. 2001; 2:483–90. [PubMed: 11905724]

[13]. Niedernhofer LJ, Odijk H, Budzowska M, van Drunen E, Maas A, et al. The structure-specificendonuclease Ercc1-Xpf is required to resolve DNA interstrand cross-link-induced double-strandbreaks. Mol Cell Biol. 2004; 24:5776–87. [PubMed: 15199134]

[14]. Darcy KM, Birrer MJ. Translational research in the Gynecologic Oncology Group: evaluation ofovarian cancer markers, profiles, and novel therapies. Gynecol Oncol. 2010; 117:429–39.[PubMed: 20233625]

[15]. Kim HS, Kim MK, Chung HH, Kim JW, Park NH, et al. Genetic polymorphisms affectingclinical outcomes in epithelial ovarian cancer patients treated with taxanes and platinumcompounds: a Korean population-based study. Gynecol Oncol. 2009; 113:264–9. [PubMed:19203783]

[16]. Krivak TC, Darcy KM, Tian C, Armstrong D, Baysal BE, et al. Relationship between ERCC1polymorphisms, disease progression, and survival in the Gynecologic Oncology Group Phase IIITrial of intraperitoneal versus intravenous cisplatin and paclitaxel for stage III epithelial ovariancancer. J Clin Oncol. 2008; 26:3598–606. [PubMed: 18640939]

[17]. Jo H, Kang S, Kim SI, Kim JW, Park NH, et al. The C19007T polymorphism of ERCC1 and itscorrelation with the risk of epithelial ovarian and endometrial cancer in Korean women. A casecontrol study. Gynecol Obstet Invest. 2007; 64:84–8. [PubMed: 17314486]

[18]. Kang S, Ju W, Kim JW, Park NH, Song YS, et al. Association between excision repair cross-complementation group 1 polymorphism and clinical outcome of platinum-based chemotherapyin patients with epithelial ovarian cancer. Exp Mol Med. 2006; 38:320–4. [PubMed: 16819291]

[19]. Li QQ, Yunmbam MK, Zhong X, Yu JJ, Mimnaugh EG. Lactacystin enhances cisplatinsensitivity in resistant human ovarian cancer cell lines via inhibition of DNA repair and ERCC-1expression. Cell Mol Biol (Noisy-le-Grand). 2001; 47 [Online Pub:OL61-72].

[20]. Darcy KM, Tian C, Reed E. A Gynecologic Oncology Group study of platinum-DNA adductsand excision repair cross-complementation group 1 expression in optimal, stage III epithelialovarian cancer treated with platinum-taxane chemotherapy. Cancer Res. 2007; 67:4474–81.[PubMed: 17483363]

[21]. Vergote I, Calvert H, Kania M, Kaiser C, Zimmermann AH, et al. A randomised, double-blind,phase II study of two doses of pemetrexed in the treatment of platinum-resistant, epithelialovarian or primary peritoneal cancer. Eur J Cancer. 2009; 45:1415–23. [PubMed: 19168349]

[22]. Weberpals J, Garbuio K, O'Brien A, Clark-Knowles K, Doucette S, et al. The DNA repairproteins BRCA1 and ERCC1 as predictive markers in sporadic ovarian cancer. Int J Cancer.2009; 124:806–15. [PubMed: 19035454]

[23]. Dabholkar M, Bostick-Bruton F, Weber C, Bohr VA, Egwuagu C, et al. ERCC1 and ERCC2expression in malignant tissues from ovarian cancer patients. J Natl Cancer Inst. 1992; 84:1512–7. [PubMed: 1433335]

DeLoia et al. Page 8

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

[24]. Dabholkar M, Vionnet J, Bostick-Bruton F, Yu JJ, Reed E. Messenger RNA levels of XPAC andERCC1 in ovarian cancer tissue correlate with response to platinum-based chemotherapy. J ClinInvest. 1994; 94:703–8. [PubMed: 8040325]

[25]. Codegoni AM, Broggini M, Pitelli MR, Pantarotto M, Torri V, et al. Expression of genes ofpotential importance in the response to chemotherapy and DNA repair in patients with ovariancancer. Gynecol Oncol. 1997; 65:130–7. [PubMed: 9103402]

[26]. Lin K, Ye D, Xie X. Protein expression levels of excision repair cross-complementation group 1and xeroderma pigmentosum D correlate with response to platinum-based chemotherapy in thepatients with advanced epithelial ovarian cancer. Int J Gynecol Cancer. 2008; 18:1007–12.[PubMed: 18081788]

[27]. Niedernhofer LJ, Bhagwat N, Wood RD. ERCC1 and non-small-cell lung cancer. N Engl J Med.2007; 356:2538–40. [author reply 2540-2531]. [PubMed: 17568038]

[28]. Bhagwat NR, Roginskaya VY, Acquafondata MB, Dhir R, Wood RD, et al. Immunodetection ofDNA repair endonuclease ERCC1-XPF in human tissue. Cancer Res. 2009; 69:6831–8.[PubMed: 19723666]

[29]. Jaspers NG, Raams A, Silengo MC, Wijgers N, Niedernhofer LJ, et al. First reported patient withhuman ERCC1 deficiency has cerebro-oculo-facio-skeletal syndrome with a mild defect innucleotide excision repair and severe developmental failure. Am J Hum Genet. 2007; 80:457–66.[PubMed: 17273966]

[30]. Niedernhofer LJ, Garinis GA, Raams A, Lalai AS, Robinson AR, et al. A new progeroidsyndrome reveals that genotoxic stress suppresses the somatotroph axis. Nature. 2006; 444:1038–43. [PubMed: 17183314]

[31]. Biggerstaff M, Szymkowski DE, Wood RD. Co-correction of the ERCC1, ERCC4 andxeroderma pigmentosum group F DNA repair defects in vitro. EMBO J. 1993; 12:3685–92.[PubMed: 8253090]

[32]. Yagi T, Wood RD, Takebe H. A low content of ERCC1 and a 120 kDa protein is a frequentfeature of group F xeroderma pigmentosum fibroblast cells. Mutagenesis. 1997; 12:41–4.[PubMed: 9025096]

[33]. Tian M, Shinkura R, Shinkura N, Alt FW. Growth retardation, early death, and DNA repairdefects in mice deficient for the nucleotide excision repair enzyme XPF. Mol Cell Biol. 2004;24:1200–5. [PubMed: 14729965]

[34]. Yu JJ, Mu C, Lee KB, Okamoto A, Reed EL, et al. A nucleotide polymorphism in ERCC1 inhuman ovarian cancer cell lines and tumor tissues. Mutat Res. 1997; 382:13–20. [PubMed:9360634]

[35]. Rosell R, Cecere F, Santarpia M, Reguart N, Taron M. Predicting the outcome of chemotherapyfor lung cancer. Curr Opin Pharmacol. 2006; 6:323–31. [PubMed: 16765644]

[36]. Li Q, Gardner K, Zhang L, Tsang B, Bostick-Bruton F, et al. Cisplatin induction of ERCC-1mRNA expression in A2780/CP70 human ovarian cancer cells. J Biol Chem. 1998; 273:23419–25. [PubMed: 9722577]

[37]. Youn CK, Kim MH, Cho HJ, Kim HB, Chang IY, et al. Oncogenic H-Ras up-regulatesexpression of ERCC1 to protect cells from platinum-based anticancer agents. Cancer Res. 2004;64:4849–57. [PubMed: 15256455]

[38]. Fan HY, Liu Z, Paquet M, Wang J, Lydon JP, et al. Cell type-specific targeted mutations of Krasand Pten document proliferation arrest in granulosa cells versus oncogenic insult to ovariansurface epithelial cells. Cancer Res. 2009; 69:6463–72. [PubMed: 19679546]

[39]. Sklar MD. Increased resistance to cis-diamminedichloroplatinum(II) in NIH 3T3 cellstransformed by ras oncogenes. Cancer Res. 1988; 48:793–7. [PubMed: 3276398]

DeLoia et al. Page 9

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

Fig. 1.ERCC1 and XPF mRNA expression in EOC. (A) Graph showing the levels of ERCC1mRNA in individual ovarian tumors on the primary vertical axis and XPF mRNA levels onsecondary vertical axis. Tumors on the horizontal axis are organized by increasing levels ofERCC1 mRNA. (B) Scatter plot showing the ERCC1 mRNA level vs. the XPF mRNA levelin the same tumor. The correlation between mRNA levels was highly significant (r=0.83,p<0.001).

DeLoia et al. Page 10

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

Fig. 2.ERCC1 and XPF protein expression in EOC. (A) Representative immunoblot of proteinextracts from tumors to measure ERCC1 and XPF protein levels. Actin was used as aloading control. The asterisks indicate unrelated cross-reacting bands. (B) Graph showingthe corrected immunoblot-band intensities for both ERCC1 (white bars) and XPF (grey bars)in each of the EOCs analyzed. The solid black line represents the ratio of band intensities ofERCC1 to XPF. Tumors on the horizontal axis were organized by increasing levels of XPFprotein. (C) Scatter plot of corrected XPF protein levels vs. ERCC1 protein levels inindividual tumors. The correlation between ERCC1 and XPF protein levels was highlysignificant (r=0.85, p<0.001).

DeLoia et al. Page 11

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

Fig. 3.Comparison of ERCC1 and XPF mRNA and protein levels. (A) Scatter plot of ERCC1protein vs. mRNA for individual tumors. There was a significant negative correlationbetween mRNA and protein levels (r=−0.41, p=0.048); (B) Scatter plot of XPF protein vs.mRNA for individual tumors. There was no correlation between the two; (C) Graph ofERCC1 protein level (bars; primary vertical axis) for individual tumors and ERCC1 mRNA(grey line; secondary vertical axis). Tumors were organized by increasing levels of ERCC1mRNA; (D) Graph of XPF protein level (bars; primary vertical axis) for individual tumorsand XPF mRNA (grey line; secondary vertical axis). Tumors were organized by increasinglevels of XPF mRNA.

DeLoia et al. Page 12

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

Fig. 4.Comparison of EOC patient outcomes and ERCC1 or XPF mRNA expression. ERCC1mRNA expression was categorized as high, medium or low and its impact on (A) PFS and(B) OS was graphed in Kaplan-Meier plots. Similarly, the impact of XPF mRNA expressionon (C) PFS and (D) OS were charted.

DeLoia et al. Page 13

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

$waterm

ark-text$w

atermark-text

$waterm

ark-text

DeLoia et al. Page 14

Table 1

Patient characteristics (n=41).

Characteristic No. patients (%)

Age (years)

Median (range) 57 (40–76)

Stage

II 1 (2.4)

III 33 (80.5)

IV 7 (17.1)

Histology

Serous 31 (75.6)

Clear cell 4 (9.8)

Carcinomasarcoma 2 (4.9)

Unknown 4 (9.8)

Tumor Grad

2 2 (4.9)

3 39 (95.1)

Debulking Status

aOptimal

37 (90.2)

Suboptimal 4 (9.8)

aOptimal debulking includes microscopic residual and any disease <1 cm gross residual.

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

DeLoia et al. Page 15

Tabl

e 2

ER

CC

1 an

d X

PF m

RN

A a

nd p

rote

in e

xpre

ssio

n le

vels

by

gene

tic p

olym

orph

ism

s.

mR

NA

Exp

ress

ion

Pro

tein

exp

ress

ion

SNP

No.

Med

ian

(25t

h–75

th P

ct)

p V

alue

No.

Med

ian

(25t

h–75

th P

ct)

p V

alue

ER

CC

1

C

odon

0.12

20.

102

11

8

C/C

82.

96 (

1.99

–3.6

8)6

6.75

(4.

54–8

.26)

C/T

226.

42 (

2.77

–17.

51)

144.

14 (

3.21

–6.3

0)

T/T

118.

51 (

4.29

–9.9

2)4

2.95

(2.

06–4

.66)

C

8092

A0.

101

0.37

9

C/C

207.

99 (

3.53

–11.

50)

103.

97 (

2.98

–6.0

2)

C/A

185.

52 (

2.55

–17.

27)

125.

23 (

3.95

–8.2

6)

A/A

32.

23 (

0.91

–3.3

4)2

5.11

(3.

08–7

.14)

XPF

G

1244

A-

-

G/G

391.

23 (

2.57

–2.6

4)24

8.69

(5.

60–1

2.66

)

G/A

21.

23 (

0.74

–2.9

7)0

T

2505

C0.

430

0.53

4

T/T

221.

27 (

1.00

–1.4

5)15

6.93

(4.

38–1

2.24

)

T/C

142.

23 (

1.07

–3.3

2)6

9.08

(6.

30–1

3.08

)

C/C

51.

17 (

0.64

–1.9

5)3

11.0

7 (8

.89–

13.0

9)

Gynecol Oncol. Author manuscript; available in PMC 2012 December 11.