A NEIL1 single nucleotide polymorphism (rs4462560) predicts the risk of radiation-induced toxicities in esophageal cancer patients treated with definitive radiotherapy

Authors

  • Yun Chen MD,

    1. Department of Radiation Oncology, Fudan University Shanghai Cancer Center, Fudan University, Shanghai, China
    Search for more papers by this author
    • The first 2 authors contributed equally to this article.

  • Meiling Zhu MD,

    1. Cancer Institute, Fudan University Shanghai Cancer Center, Fudan University, Shanghai, China
    Search for more papers by this author
    • The first 2 authors contributed equally to this article.

  • Zhen Zhang MD,

    1. Department of Radiation Oncology, Fudan University Shanghai Cancer Center, Fudan University, Shanghai, China
    Search for more papers by this author
  • Guoliang Jiang MD,

    1. Department of Radiation Oncology, Fudan University Shanghai Cancer Center, Fudan University, Shanghai, China
    Search for more papers by this author
  • Xiaolong Fu MD,

    1. Department of Radiation Oncology, Fudan University Shanghai Cancer Center, Fudan University, Shanghai, China
    Search for more papers by this author
  • Min Fan MD,

    1. Department of Radiation Oncology, Fudan University Shanghai Cancer Center, Fudan University, Shanghai, China
    Search for more papers by this author
  • Menghong Sun MD,

    1. Department of Pathology, Fudan University Shanghai Cancer Center, Fudan University, Shanghai, China
    Search for more papers by this author
  • Qingyi Wei MD, PhD,

    Corresponding author
    1. Cancer Institute, Fudan University Shanghai Cancer Center, Fudan University, Shanghai, China
    2. Department of Epidemiology, The University of Texas MD Anderson Cancer Center, Houston, Texas
    • Corresponding authors: Kuaile Zhao, MD, Department of Radiation Oncology, Fudan University Shanghai Cancer Center, 270 Dong An Road, Shanghai 200032, China; Fax: (011) 86-21-64174774; kuaile_z@sina.com; Qing-Yi Wei, MD, PhD, Cancer Institute, Fudan University Shanghai Cancer Center, 270 Dong An Road, Shanghai 200032, China; Fax: (011) 86-21-64172585; weiqingyi@yahoo.com; or Department of Epidemiology, The University of Texas MD Anderson Cancer Center, 1515 Holcomb Boulevard, Houston, TX 77401; Fax: (713) 563-0999; qwei@mdanderson.org

    Search for more papers by this author
  • Kuaile Zhao MD

    Corresponding author
    1. Department of Radiation Oncology, Fudan University Shanghai Cancer Center, Fudan University, Shanghai, China
    • Corresponding authors: Kuaile Zhao, MD, Department of Radiation Oncology, Fudan University Shanghai Cancer Center, 270 Dong An Road, Shanghai 200032, China; Fax: (011) 86-21-64174774; kuaile_z@sina.com; Qing-Yi Wei, MD, PhD, Cancer Institute, Fudan University Shanghai Cancer Center, 270 Dong An Road, Shanghai 200032, China; Fax: (011) 86-21-64172585; weiqingyi@yahoo.com; or Department of Epidemiology, The University of Texas MD Anderson Cancer Center, 1515 Holcomb Boulevard, Houston, TX 77401; Fax: (713) 563-0999; qwei@mdanderson.org

    Search for more papers by this author

Abstract

BACKGROUND

To assess the association between single nucleotide polymorphisms (SNPs) of base-excision repair genes and clinical outcomes, the roles of genetic variants of 3 selected genes—flap structure-specific endonuclease 1 (FEN1), 8-hydroxyguanine DNA glycosylase (hOGG1), and nei endonuclease VIII-like 1 (NEIL1)—were investigated in radiation-induced esophageal toxicity (RIET), radiation pneumonitis (RP), and overall survival (OS) after radio(chemo)therapy in patients with esophageal squamous cell carcinoma (ESCC).

METHODS

NEIL1 reference SNP 4462560 (rs4462560) and rs7402844, hOGG1 rs1052133 and rs293795, and FEN1 rs4246215 and rs174538 were genotyped in 187 patients with ESCC who received definitive radiotherapy with or without chemotherapy. Kaplan-Meier cumulative probabilities and Cox proportional hazards regression models were used to assess the effect of the genotypes on the risk of RIET, RP, and OS.

RESULTS

The authors observed that patients who had the NEIL1 rs4462560 GC/CC genotype had a statistically significantly lower risk of both grade ≥2 acute radiation-induced esophageal toxicity (RIET) (adjusted hazard ratio [HR], 0.421; 95% confidence interval [CI], 0.207-0.856; P = .017) and grade ≥2 acute radiation pneumonitis (RP) (adjusted HR, 0.392; 95% CI, 0.163-0.946; P = .037) compared with patients who had the GG genotype, but the genotype did not affect OS (adjusted HR, 0.778; 95% CI, 0.471-1.284; P = .326). There were no significant findings for other the SNPs under investigation.

CONCLUSIONS

The NEIL1 rs4462560 SNP may serve as a predictor of acute RIET and RP risk but not of OS. Larger prospective studies are needed to validate these findings. Cancer 2013;119:4205–4211. © 2013 American Cancer Society.

INTRODUCTION

Radiation-induced esophageal toxicity (RIET) and radiation pneumonitis (RP) are 2 major side effects of thoracic radiotherapy that have been well studied in patients with lung cancer.[1-3] Nevertheless, there is significant variation in the risk of radiation toxicities among patients, even when their normal tissues have been exposed to the same dose range or volume of irradiation, suggesting that genetic makeup determines an individual's susceptibility to radiation toxicities. Therefore, it is important to identify critical biomarkers that may optimize individual therapy by adjusting the patient's radiation dose.

The base-excision repair (BER) pathway is a major pathway responsible for the removal of DNA damage caused by the action of reactive oxygen species and alkylating agents. BER facilitates the repair of damaged DNA through 2 pathways: a short-patch pathway and a long-patch pathway.[4] For example, nei endonuclease VIII-like 1 (NEIL1) and 8-hydroxyguanine DNA glycosylase (hOGG1) are 2 glycosylases involved in the short-patch pathway that participate in removing oxidatively induced DNA base lesions; whereas flap structure-specific endonuclease 1 (FEN1) is essential for the long-patch pathway, which conducts a repair tract of at least 2 nucleotides.[4-6] It has been demonstrated that suboptimal BER capacity may increase tissue radiosensitivity and produce more severe radiation toxicity.[7-9] In fact, a few studies have suggested that single nucleotide polymorphisms (SNPs) of BER genes may function as biomarkers for susceptibility to radiation-induced normal tissue toxicities and treatment outcomes; however, those studies focused primarily on cancers of the esophagus, lung, breast, and cervix and had relatively small sample sizes.[10-15]

In view of published studies, we hypothesized that functional SNPs of BER genes may be useful biomarkers for predicting clinical outcomes in patients with esophageal squamous cell carcinoma (ESCC) who receive definitive radiotherapy with or without chemotherapy. To test this hypothesis, we performed a case-only study in which we genotyped 6 SNPs in 3 BER pathway genes—FEN1, hOGG1, and NEIL1—and evaluated their roles in the development of acute RIET and RP. In addition, we assessed the association between these SNPs and OS in the same patients.

MATERIALS AND METHODS

Patients

This study included 187 patients with ESCC who received treatment at Fudan University Shanghai Cancer Center between 2008 and 2011. The patient selection criteria included the following: 1) newly diagnosed and histopathologically confirmed ESCC; 2) the intent to administer definitive radiotherapy or radiotherapy combined with chemotherapy (total radiation dose, ≥50 grays [Gy]); 3) the patient received intensity-modulated radiation therapy; and 4) blood samples and signed informed consent forms were obtained before radiotherapy. Acute esophagitis and acute pneumonitis were graded according to the Common Terminology Criteria for Adverse Events version 4.0.[16]

Dosimetric data were calculated using the Philips Healthcare radiation therapy planning system (Pinnacle 8.0; Philips Radiation Oncology Systems, Milpitas, Calif) for each patient, including the mean dose to the entire esophagus (MED), the length of gross tumor volume (GTV) of primary esophageal cancer (LGTV-P), the length of the esophagus that received at least 50 Gy (LE50), the mean lung dose (MLD), and the volume of lung that received at least 20 Gy (V20). The entire esophagus (from the bottom of the cricoid cartilage to the gastroesophageal junction) included the esophageal tumor (GTV-P) and the normal esophagus, which were separately contoured as 2 parts.

Single Nucleotide Polymorphism Selection and Genotyping

The National Center for Biotechnology Information dbSNP database (http://www.ncbi.nlm.nih.gov/, accessed August 25, 2013) and SNPinfo (available at: http://snpinfo.niehs.nih.gov/, accessed August 25, 2013) were used to identify the common, potentially functional SNPs based on the following 3 criteria: 1) located at the regulatory regions of genes (ie, the 5′ near gene, the 5′ untranslated region [UTR], the 3′ UTR, the 3′ near gene, and splice sites) or coding regions; 2) a minor allele frequency (MAF) ≥5% in Han, Beijing (CHB) descendants reported in the dbSNP database; and 3) affected the activity of microRNA binding sites and transcription factor binding sites in the putative promoter region or changed the amino acid in the exons. Consequently, 1 SNP in the 3′ UTR of FEN1 (SNP reference 4246215 [rs4246215]; T→G), 1 SNP in the 5′ near gene of FEN1 (rs174538; G→A), 1 SNP in the 3′ near gene of NEIL1 (rs4462560; C→G), and 1 SNP in the exon of hOGG1 (rs1052133; G→C) were finally selected for genotyping in the study population. In addition, we added 2 more SNPs as references, including 1 in the 5′ near gene of NEIL1 (rs7402844; C→G) and 1 in the 3′ UTR of hOGG1 (rs293795; G→A), because it has been predicted that these are potentially functional; however, the MAF did not completely meet the criteria, because the former had an MAF of 30% in CHB and Japanese in Tokyo combined, and the later had an MAF of 4.8% in CHB. A linkage disequilibrium analysis suggested that these 6 potentially functional SNPs also captured 75 other untyped SNPs within the 3 selected genes or in nearby genes. Genomic DNA was isolated from blood samples, and assays were performed for genotyping (TaqMan; Invitrogen, Inc., Carlsbad, Calif), as described previously,[17] with a successful genotyping rate of 99.8%. The genotyping assays were repeated for 10% of the samples, and the results were 100% concordant.

Statistical Methods

The chi-square test was used to assess the differences in frequency distributions of the selected demographic variables, risk factors, alleles, and genotypes of the selected SNPs between the cases and controls. Univariate Cox proportion hazards regression analysis was performed to calculate the hazards ratio (HR) and confidence interval (CI) for the effects of genotypes on the risk of RP, RIET, and OS. In addition, multivariate Cox hazards regression analysis was performed to adjust for other covariates. Kaplan-Meier curves were used to estimate the cumulative RP and RIET probability and OS.

RESULTS

Patient Characteristics and Clinical Outcomes

The study included 187 patients with ESCC with a median age of 64 years (range, 37-88 years), of whom 73.8% (138 of 187 patients) were men. The median total radiation dose was 62.0 Gy (range, 50-68 Gy) at doses of 1.8 to 2.25 Gy per fraction. Of all 187 patients, 49 (26.2%) received radiation therapy alone, and 138 (73.8%) received chemotherapy plus radiotherapy. The chemotherapeutic agents were either cisplatin-based or fluorouracil-based. Baseline clinical and dosimetric characteristics of the patients are summarized in Table 1.

Table 1. Associations Between Patient-Related, Tumor-Related, and Therapy-Related Characteristics and Grade ≥2 Radiation-Induced Esophageal Toxicity, Grade ≥2 Radiation Pneumonitis, and Overall Survival
ParameterNo.RIETRPOS
Univariate AnalysisMultivariate AnalysisUnivariate AnalysisMultivariate AnalysisUnivariate AnalysisMultivariate Analysis
HR (95%CI)PHR (95%CI)PaHR (95%CI)PHR (95%CI)PbHR (95%CI)PHR (95%CI)Pc
  1. Abbreviations: CI, confidence interval; Gy, grays; HR, hazard ratio; LE50, length of esophagus receiving >50 Gy; LGTV-P, length of gross tumor volume of primary esophageal cancer; MED, mean esophagus dose; MLD, mean lung dose; OS, overall survival; RIET, radiation-induced esophageal toxicity; RP, radiation pneumonitis; V20, volume of normal lung receiving 20 Gy or more radiation.

  2. a

    P values were calculated with adjustment for age, sex, smoking status, chemotherapy history, fraction dose, disease stage, radiation dose, LGTV-P, MED, and LE50.

  3. b

    P values were calculated with adjustment for age, sex, smoking status, chemotherapy history, fraction dose, disease stage, radiation dose, MLD, and V20.

  4. c

    P values were calculated with adjustment for age, sex, smoking status, chemotherapy history, fraction dose, disease stage, radiation dose, and LGTV-P.

  5. d

    Staging was according to the International Union Against Cancer ICC (sixth edition).

Sex             
Men1381.000 1.000 1.000 1.000 1.000 1.000 
Women491.328 (0.690–2.554).3962.102 (0.924–4.778).0761.346 (0.609–2.974).4631.247 (0.521–2.987).6200.855 (0.535–1.366).5110.909 (0.526–1.569).731
Age, y             
≤651121.000 1.000 1.000 1.000 1.000 1.000 
>65750.371 (0.177–0.775).0080.431 (0.199–0.931).0321.291 (0.614–2.712).5011.879 (0.833–4.235).1281.211 (0.809–1.814).3531.204 (0.769–1.884).417
No. of pack-years             
≤181121.000 1.000 1.000 1.000 1.000 1.000 
>18701.334 (0.716–2.487).3641.816 (0.849–3.885).1240.652 (0.286–1.490).3110.688 (0.271–1.742).4301.942 (0.890–2.029).1611.143 (0.718–1.821).573
Chemotherapy             
No491.000 1.000 1.000 1.000 1.000 1.000 
Yes1381.347 (0.644–2.815).4280.772 (0.341–1.749).5352.300 (0.800–6.620).1241.972 (0.610–6.373).2570.711 (0.459–1.102).1270.591 (0.345–1.012).055
Fraction dose, Gy             
≤21411.000 1.000 1.000 1.000 1.000 1.000 
>2460.478 (0.201–1.133).0940.404 (0.143–1.144).0880.814 (0.330–2.008).6551.673 (0.518–5.402).3900.438 (0.254–0.756).0030.552 (0.288–1.056).073
Staged             
I-II671.000 1.000 1.000 1.000 1.000 1.000 
III-IV1201.114 (0.586–2.115).7420.938 (0.467–1.884).8751.467 (0.646–3.331).3601.358 (0.563–3.273).4962.382 (1.477–3.841).0001.935 (1.161–3.225).011
Radiation dose, Gy             
≤61.2871.000 1.000 1.000 1.000 1.000 1.000 
>61.21000.560 (0.302–1.037).0650.683 (0.324–1.441).3170.625 (0.296–1.322)0.2190.607 (0.224–1.642).326.668 (0.443–1.008).0540.702 (0.416–1.184).185
LGTV-P, cm             
≤71331.000 1.000 - - 1.000 1.000 
>7521.114 (0.586–2.115).6761.404 (0.676–2.913).363----1.579 (1.032–2.416).0351.702 (1.066–2.719).026
MED, Gy             
≤45641.000 1.000 - - -   
>451200.960 (0.510–1.800).8900.647 (0.306–1.306).255--------
LE50, cm             
≤10261.000 1.000 - - - - 
>101593.614 (0.873–14.955).0765.035 (1.103–22.991).037--------
MLD, Gy             
≤1392- - 1.000 1.000 - - 
>1392----0.783 (0.367–1.673)0.5280.495 (0.188–1.303)0.154----
V20, %             
≤0.28170- - 1.000 1.000 - - 
>0.2815----2.656 (1.009–6.991).0483.718 (1.075–12.854).038----

At the time of the last follow-up (January 2013), 97 patients (55.1%) had died with a median survival time (MST) of 11 months (range, 1-49 months), and 79 patients (44.9%) were still alive for an MST of 27 months (range, 14-48 months). The OS rates at 1 year, 2 years, and 3 years were 70.5%, 49.9%, and 43.4%, respectively.

Although none of the patients in this study suffered death caused by acute radiation-induced toxicity, 42 patients (22.5%) exhibited grade ≥2 acute RIET (ie, grade 2, 3, and 4 toxicities were observed in 33 patients, 6 patients, and 3 patients, respectively). In addition, 28 patients (15%) exhibited grade ≥2 acute RP (ie, grade 2, 3, and 4 toxicities were observed in 23 patients, 5 patients, and zero patients, respectively). The median time to occurrence was 24 days after the first day of irradiation for patients with RIET and 35 days after the first day of irradiation for patients with RP.

FEN1, hOGG1, and NEIL1 Single Nucleotide Polymorphisms and the Risk of Radiation-Induced Esophageal Toxicity

Table 1 displays results from the univariate and multivariate analyses of patient-related, tumor-related, and therapy-related characteristics and grade ≥2 RIET. Patients aged >65 years had less RIET risk compared with those aged ≤65 years (adjusted P = .032). The LE50 was also associated with RIET risk, and patients who exposed >10 cm of the esophagus to ≥50 Gy of irradiation had a more than 5-fold greater RIET risk compared with those who exposed ≤10 cm of the esophagus (adjusted P = .037). Table 2 displays the Cox proportional hazards regression analyses for RIET risk by FEN1, hOGG1, and NEIL1 genotypes. The NEIL1 rs4462560 GC/CC genotypes were associated with significantly lower risk of grade ≥2 RIET than the GG genotype in a multivariate model with adjustment for age, sex, smoking status, type of treatment, mean esophageal dose, radiation dose, fraction dose, stage, LGTV-P, and LE50 (adjusted HR, 0.421; 95% CI, 0.207-0.856; P = .017) (Fig. 1); whereas the rs7402844 SNP had no association with the risk of RIET. Furthermore, the hOGG1 rs293795 and rs1052133 SNPs and the FEN1 rs4246215 and rs174538 SNPs revealed no significant association with the risk of RIET.

Figure 1.

The cumulative probabilities of (a) radiation-induced esophageal toxicity (RIET) and (b) radiation pneumonitis (RP) are illustrated as a function of time from the start of therapy according to genotypes of the NEIL1 rs4462560 single nucleotide polymorphism. The NEIL1 rs4462560 GG genotype was associated with a statistically significant higher incidence of both RIET and RP compared with GC/CC genotypes.

Table 2. Associations Between Genotypes and Grade ≥2 Radiation-Induced Esophageal Toxicity, Grade ≥2 Radiation Pneumonitis, and Overall Survival
PolymorphismNo.No. (%)aRIETRPOS
Univariate AnalysisMultivariate AnalysisNo. (%)cUnivariate AnalysisMultivariate AnalysisNo. (%)eUnivariate AnalysisMultivariate Analysis
HR [95% CI]PHR [95% CI]PbHR [95% CI]PHR [95% CI]PdHR [95% CI]PHR [95% CI]Pf
  1. Abbreviations: CI, confidence interval; HR, hazard ratio; OS, overall survival; RIET, radiation-induced esophageal toxicity; RP, radiation pneumonitis.

  2. a

    “No. (%)” indicates the number of number of individuals who had grade ≥2 RIET and their proportion in each genotype category.

  3. b

    P values were calculated with adjustment for age, sex, smoking status, chemotherapy history, fraction dose, disease stage, radiation dose, length of gross tumor volume of the primary the esophageal cancer, mean esophageal dose, and length of the esophagus receiving >50 grays.

  4. c

    “No. (%)” indicates the number of number of individuals who had grade ≥2 RP and their proportion in each genotype category.

  5. d

    P values were calculated with adjustment for age, sex, smoking status, chemotherapy history, fraction dose, disease stage, radiation dose, mean lung dose, and the volume of normal lung receiving ≥20 Gy radiation.

  6. e

    “No. (%)” indicates the number of number of individuals who died and their proportion in each genotype category.

  7. f

    P values were calculated with adjustment for age, sex, smoking status, chemotherapy history, fraction dose, disease stage, radiation dose, and length of gross tumor volume of the primary esophageal cancer.

NEIL1 rs4462560                
GG3613 (36.1)1.000 1.000 8 (22.2)1.000 1.000 23 (69.7)1.000 1.000 
GC9118 (19.8)0.486 [0.238–0.991].0470.436 [0.201–0.947].03615 (16.5)0.733 [0.311–1.730].4790.426 [0.164–1.102].07845 (52.3)0.695 [0.420–1.149].1560.923 [0.539–1.581].770
CC6011 (18.3)0.454 [0.203–1.013].0540.399 [0.171–0.932].0345 (8.3)0.357 [0.117–1.090].0710.339 [0.109–1.057].06229 (50.9)0.561 [0.323–0.975].0410.621 [0.345–1.116].111
GC/CC15129 (19.2)0.473 [0.246–0.910].0250.421 [0.207–0.856].01720 (13.2)0.580 [0.256–1.317].1930.392 [0.163–0.946].03774 (51.7)0.637 [0.398–1.019].0600.778 [0.471–1.284].326
NEIL1 rs7402844                
GG7815 (19.2)1.000 1.000 9 (11.5)1.000 1.000 37 (50.7)1.000 1.000 
GC8520 (23.5)1.259 [0.645–2.460].5001.191 [0.597–2.374].62015 (17.6)1.589 [0.695–3.631].2721.439 [0.597–3.467].41847 (57.3)1.305 [0.845–2.016].2301.342 [0.856–2.104].199
CC247 (29.2)1.508 [0.615–3.699].3701.461 [0.538–3.966].4574 (16.7)1.435 [0.442–4.660].5481.679 [0.484–5.828].41413 (61.9)1.467 [0.777–2.770].2371.165 [0.584–2.326].665
GC/CC10927 (24.8)1.315 [0.700–2.473].3941.243 [0.646–2.391].51519 (17.4)1.554 [0.703–3.430].2761.485 [0.639–3.451].35860 (58.3)0.748 [0.494–1.131].1680.767 [0.499–1.178].226
FEN1 rs4246215                
GG7414 (18.9)1.000 1.000 10 (13.5)1.000 1.000 39 (58.2)1.000 1.000 
GT9423 (24.5)1.315 [0.677–2.556].4191.415 [0.704–2.843].33014 (14.9)1.088 [0.483–2.451].8381.032 [0.414–2.571].94651 (56)1.070 [0.703–1.627].7531.272 [0.816–1.983].288
TT195 (26.3)1.470 [0.529–4.082].4601.061 [0.327–3.444].9224 (21.1)1.586 [0.498–5.059].4351.976 [0.563–6.927].2877 (38.9)0.871 [0.388–1.955].7380.906 [0.372–2.203].827
GT/TT11328 (24.8)1.340 [0.710–2.550].3711.355 [0.685–2.679].38318 (15.9)1.170 [0.540–2.535].6901.171 [0.492–2.786].72258 (53.2)1.041 [0.692–1.565].8481.220 [0.790–1.883].369
FEN1 rs174538                
GG7414 (18.9)1.000 1.000 10 (13.5)1.000 1.000 39 (58.2)1.000 1.000 
GA9323 (24.7)1.331 [0.685–2.588].3981.431 [0.712–2.873].31413 (14)1.011 [0.443–2.306].9790.941 [0.372–2.382].89750 (55.6)1.049 [0.688–1.598].8241.247 [0.798–1.947].332
AA205 (25)1.384 [0.499–3.844].5321.002 [0.308–3.256].9975 (25)1.979 [0.676–5.792].2132.545 [0.785–8.258].1208 (42.1)0.992 [0.462–2.132].9841.058 [0.459–2.442].894
GA/AA11328 (24.8)1.340 [0.710–2.550].3711.360 [0.690–2.680].38318 (15.9)1.170 [0.540–2.535].6901.170 [0.490–2.790].72858 (53.2)1.041 [0.692–1.565].8481.220 [0.790–1.883].369
HOGG1 rs1052133                
GG7012 (17.1)1.000 1.000 9 (12.91.000 1.000 28 (42.4)1.000 1.000 
GC8624 (27.9)1.748 [0.874–3.496].1141.707 [0.838–3.478].14115 (17.4)1.373 [0.601–3.138].4521.286 [0.538–3.074].57248 (58.5)1.288 [0.806–2.056].2901.241 [0.762–2.022].386
CC316 (19.4)1.188 [0.446–3.165].7310.904 [0.313–2.612].8524 (12.9)0.975 [0.300–3.167].9670.893 [0.238–3.352].86721 (75)2.008 [1.130–3.568].0171.794 [0.984–3.271].057
GC/CC11730 (25.6)1.600 [0.820–3.120].1711.475 [0.741–2.938].26919 (16.2)1.265 [0.572–2.795].5621.190 [0.520–2.720].69069 (62.7)1.441 [0.927–2.241].1051.376 [0.872–2.171].170
HOGG1 rs293795                
AA17137 (21.5)1.000 1.000 25 (14.6)1.000 1.000 87 (53.4)1.000 1.000 
AG115 (45.5)2.502 [0.982–6.375].0552.968 [1.095–8.042].0322 (18.2)1.247 [0.295–5.265].7641.633 [0.359–7.424].5259 (90)2.188 [1.097–4.362].0261.832 [0.854–3.931].120
GG50 (0)0.000 [0.000].9890.000 [0.000].9781 (20.2)1.478 [0.200–10.909].7020.000 [0.000-].9831 (33.3)0.630 [0.086–4.432].6300.656 [0.087–4.966].683
AG/GG165 (31.3)1.557 [0.610–3.960].3531.933 [0.743–5.031].1773 (18.8)1.316 [0.397–4.358].6541.137 [0.262–4.929].86410 (76.9)1.742 [0.903–3.360].0981.538 [0.748–3.172].244

FEN1, hOGG1, and NEIL1 Single Nucleotide Polymorphisms and the Risk of Radiation Pneumonitis

Table 1 lists the associations between clinical and dosimetric parameters and the risk of grade ≥2 RP. The V20 value was associated significantly with the risk of RP in both univariate and multivariate analyses (adjusted P = .038). None of the other characteristics were associated with RP risk in this study population.

The associations of FEN1, hOGG1, and NEIL1 genotypes with the risk of grade ≥2 RP were analyzed using a Cox proportional hazards regression model. Patients who had the NEIL1 rs4462560 GC/CC genotypes had a decreased RP risk in multivariate analysis compared with those who had the GG genotype (adjusted HR, 0.392; 95% CI, 0.163-0.946; P = .037) (Fig. 1) with adjustment for age, sex, smoking status, type of treatment, MLD, radiation dose, fraction dose, stage, and V20. No significant associations between the risk of RP and the other genotypes were observed (NEIL1 rs7402844, hOGG1 rs1052133 and rs293795, FEN1 rs4246215 and rs174538) (Table 2).

FEN1, hOGG1, and NEIL1 Single Nucleotide Polymorphisms and Overall Survival

Univariate and multivariate analyses demonstrated that the variables LGTV-P and stage were associated significantly with OS (adjusted P = .026 and P = .011, respectively) (Table 1). However, the data failed to demonstrate any association between genotypes of BER genes and OS (Table 2).

DISCUSSION

NEIL1, which is located on chromosome 15q23, was first identified in 2002 as a gene encoding a human glycosylase that participates in the first step of the BER pathway, modifying the bases with oxidative damage.[18] It has been established that NEIL1 is a critical component of BER that may incorporate deoxyuridylate to repair DNA damage caused by thymidylate synthetase pathway inhibitors.[19] Rosenquist et al discovered that NEIL1 acts as compensation for the other BER genes, such as OGG1 and neutral trehalase 1(NTH1). For example, it is reported that NEIL1 repairs lesions like FapyA and 5S-6R thymine glycol, which cannot be excised by OGG1 or NTH1.[8] Rosenquist et al further discovered that embryonic stem cell lines with no NEIL1 expression were approximately twice as sensitive to low levels of irradiation as the control cells.[8] Zhai et al investigated the associations between SNPs of NEIL1/NEIL2 and the risk of squamous cell carcinoma of the oral cavity and oropharynx; and they observed that the NEIL2 rs804270 CC genotype was associated with advanced tumor stages, but they did not observe a risk associated with either NEIL1 rs7182283 or rs4462560 SNPs.[20] However, there is no report for a role of functional NEIL1 polymorphisms in OS or the risk of radiation-induced normal tissue toxicity like RP and RIET.

In this study, the data demonstrated that patients with the NEIL1 rs4462560 GG genotype had a statistically significantly elevated risk of both grade ≥2 RIET and grade ≥2 RP compared with those who had GC/CC genotypes. Nevertheless, none of the 6 SNPs were associated with OS. It is likely that our sample size was not large enough to detect some weak associations, if any.

Our data indicated that the 1-year and 2-year survival rates in this Chinese patient cohort were 71% and 51%, respectively, which were similar to those reported from the Radiation Therapy Oncology Group (RTOG) 0113 or RTOG 9405 trials in the United States.[21] Because of the advances in radiotherapy technology and the improvement of quality assurance, dosimetric parameters like LE50, V20, and MLD have decreased, and the numbers of patients who develop either grade ≥3 RIET or grade ≥3 RP have decreased accordingly. Therefore, we chose grade ≥2 RIET and grade ≥2 RP as the cutoff targets for analysis in the current study. We used the parameter of esophageal length instead of the volume that received a certain dose or more by considering that there were diverse ways to contour the esophagus and measure the length, which are much easier if done consistently. We observed that age was a significant factor affecting RIET, a finding consistent with the study by Ahn et al,[22] and that LE50 was associated with RIET in multivariate analyses, although no statistical significance was detected in univariate analysis. We also demonstrated that V20 was an important dosimetric parameter of predicting the risk of RP, which is highly consistent with other reported studies.[23-26]

We assumed that the HR of DNA SNPs on acute radiation-induced toxicities was >1.5 or <0.67 based on findings from previous studies, and we calculated the power with a significance level of α = .05. Consequently, we observed that our sample size of 187 patients would have 21.1% and 67.9% power to detect an HR of 1.5 for the variant genotype with MAF values of 5% and 30%, respectively. The power increased as the HR and MAF increased. For example, our current study had a power of 99.1% to detect an HR of 0.421 for rs4462560 GC/CC versus GG for the risk of grade ≥2 acute RIET with an MAF of 43.6%. In addition, we performed a multiple comparison analysis (for 6 SNPs) and observed that the true P value was > .05 according to the Bonferroni adjustment. Indeed, because all methods of testing multiple comparisons are very conservative when exploring candidate genes with a prior hypothesis, we call for additional large studies to replicate our findings. Once our results are validated in larger studies, by knowing the NEIL1 rs4462560 genotypes before patients receive radiation therapy, we could help patients to reduce the chance of developing RIET or RP, because radiation oncologists may be able to revise the prescription dose or use other strategies to protect the normal esophagus and lungs. While we look forward to other investigators validating our observations in larger studies, we are planning to substantiate our findings by increasing our sample size and performing molecular mechanism investigations in future studies.

FUNDING SUPPORT

This study was supported by the National Natural Science Foundation of China Research, China (grant 21172043); the China Recruitment Program of Global Experts at Fudan University; and a grant from the Chinese Ministry of Health (grant 201002007).

CONFLICT OF INTEREST DISCLOSURES

Qingyi Wei was supported by the China Recruitment Program of Global Experts at Fudan University and by a grant from the Ministry of Health (grant 201002007). Kuaile Zhao was supported by the National Natural Science Foundation of China (grant 21172043).

Ancillary