Molecular epidemiological studies have reported large inter-individual variations in susceptibility to environmental carcinogens and subsequent cancer risk, which may partly be due to genetically determined variations in nucleotide excision repair (NER) capacity(Reference Andrew, Nelson and Kelsey1–Reference Lockett, Snowhite and Hu4). The NER pathway protects the integrity of the genome by recognising and eliminating a broad spectrum of bulky lesions such as UV-induced pyrimidine dimers, aromatic DNA adducts, and cross-links. Moreover, several NER-related enzymes have been shown to play a role in the cellular protection against certain types of oxidative DNA damage (including thymine glycols, 8-oxoguanine and cyclodeoxyadenosine), most likely by acting as a cofactor in base excision repair (BER)(Reference Friedberg5–Reference Kuraoka, Bender and Romieu7). Thus NER is a versatile DNA repair system, involving the joint action of a variety of enzymes such as XPC–RAD23B, CSB, XPA, XPF–ERCC1, and others(Reference Sancar, Lindsey-Boltz and Unsal-Kaçmaz8, Reference Sugasawa9). It is clear that NER plays a crucial role in cancer prevention, because defects in this pathway lead to several severe human disorders, such as xeroderma pigmentosum(Reference Bootsma, Kraemer, Cleaver, Vogelstein and Kinzler10). Furthermore, several studies suggest that genetic polymorphisms in various NER genes may have a profound impact on the phenotypical activity of this repair pathway(Reference Zhu, Yang and Chen11–Reference Dusinska, Dzupinkova and Wsolova13). In addition, various genetic polymorphisms in DNA-repair genes have been shown to modulate the levels of bulky DNA adducts(Reference Zhu, Yang and Chen11, Reference Matullo, Palli and Peluso14, Reference Shen, Gammon and Terry15) or chromosomal damage(Reference Cheng, Leng and Dai16–Reference Mateuca, Roelants and Iarmarcovai18). So, a number of studies identified associations between polymorphisms in DNA-repair genes with the amount of DNA damage and the capacity to repair these damages.
Next to the effect of single nucleotide polymorphisms (SNP) on DNA-repair activity, also other factors, such as diet and specific dietary compounds, are thought to modulate DNA-repair capacities. Although there is sufficient evidence for chemopreventive effects of certain dietary compounds(19), only a few studies have reported that dietary compounds influence DNA-repair processes (for a review, see Tyson & Mathers(Reference Tyson and Mathers20)). Several of these studies investigated the dietary modulation of BER or the repair of oxidative lesions. For example, a 3-week intervention with one, two or three kiwi fruits resulted in a significant increase of the BER capacity, as measured by a modified comet assay(Reference Collins, Harrington and Drew21). In fact, to the best of our knowledge, there are only two studies that investigated the effect of dietary factors on NER capacity in human subjects. Wei et al. observed an association between low dietary folate intake and reduced NER capacity(Reference Wei, Shen and Wang22), while Tyson et al. reported no detectable effect of micronutrient supplementation on NER capacity(Reference Tyson, Caple and Spiers23). Therefore, there is an increasing need to study the impact of diet on NER capacity.
We previously showed that especially oxidative stress can inhibit NER capacity(Reference Langie, Knaapen and Houben24, Reference Gungor, Godschalk and Pachen25). Thus, enhanced dietary intake of antioxidants may represent an opportunity for improving NER by reducing oxidative stress. Therefore, we studied the dietary modulation of DNA repair by using samples from a 4-week intervention study with healthy volunteers, consuming 1 litre of blueberry and apple juice per d(Reference Wilms, Boots and de Boer26). This intervention was found to be efficient in enhancing antioxidant defence and reducing the levels of ex vivo-induced oxidative DNA damage(Reference Wilms, Boots and de Boer26, Reference Wilms, Hollman and Boots27). In the present study, we hypothesised that NER capacity is determined by polymorphisms in DNA-repair genes and that diet may modulate an individual's NER capacity. Therefore, our aims were to (i) investigate the effect of the blueberry and apple juice intervention on the NER capacity; (ii) determine the effect of genetic polymorphisms in NER genes on the phenotypic NER capacity; (iii) identify possible gene–diet interactions.
Materials and methods
Study population
The study population consisted of 168 healthy volunteers, 114 female and fifty-four male, aged 18–45 years (for more details, see references(Reference Wilms, Boots and de Boer26, Reference Wilms, Hollman and Boots27)). Volunteers were recruited through advertisement in local newspapers and were included if they were non-smokers, did not use any medication (except for oral contraceptives) or any vitamin supplementation. The present study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Medical Ethical Committee of Maastricht University and the Academic Hospital Maastricht. Written informed consent was obtained from all subjects.
Dietary intervention and study design
The design of the present intervention as well as the efficacy of the washout period was based on a pilot study, described previously(Reference Wilms, Boots and de Boer26, Reference Wilms, Hollman and Boots27). Briefly, in a paired design, each subject acted as his or her own control. A 5 d washout period was followed by an intervention period of 4 weeks with a custom-made blueberry and apple juice mixture, produced specifically for the present study by Riedel Drinks (now Friesland Foods, Ede, The Netherlands), of which subjects consumed 1 litre/d. This blueberry and apple juice mixture was about 1·85 times more concentrated than regular fruit juices of Riedel Drinks, consisting of 135 % blueberry juice and 50 % apple juice. As a consequence, it contained high levels of antioxidants, predominantly in the form of flavonoids. Supplementation with this blueberry and apple juice for 4 weeks was reported to be effective, as the intervention significantly increased total plasma antioxidant capacity (trolox equivalent antioxidant capacity (TEAC); P < 0·001) and reduced the levels of ex vivo-induced oxidative DNA damage by 20 % (P = 0·006)(Reference Wilms, Boots and de Boer26). The impact of seasonal variation in dietary habits or increased sensitivity was overcome by year-round random sampling.
Collection of samples
After the 5 d washout period and a second time after the 4-week intervention period, blood samples were obtained between 08.00 and 09.00 hours by venepuncture. Volunteers were allowed to have breakfast before sampling, but no juice. Venous blood samples were obtained into one 10 ml EDTA vacuum tube for plasma analyses and into two 10 ml vacuum lithium heparin tubes (venoject II; Terumo-Europe, Leuren, Belgium) for isolation of lymphocytes. The EDTA tubes were centrifuged for 10 min at 265 g at 4°C to separate plasma for the analysis of the total plasma antioxidant capacity (TEAC) as described previously(Reference Wilms, Boots and de Boer26). All plasma samples were kept at − 80°C until analysis. Lymphocytes were isolated using a standard density gradient centrifugation method(Reference Boyum28), sampled and stored as cell pellets at − 20°C. One sample was used to isolate DNA for genotyping purposes, using standard phenol extraction procedures. Another lymphocyte sample was used to prepare protein extracts to phenotypically assess NER capacity.
Selection of polymorphisms for genotyping
In the present study, twelve SNP in NER genes (Table 1) were selected on the basis of (a) their association with cancer development, or (b) their expected influence on DNA repair based on literature review. DNA sequences and allele frequencies were obtained from the Cancer SNP 500 database (http://snp500cancer.nci.nih.gov). Of the twelve SNP analysed here, two have been described before by Wilms et al. (Reference Wilms, Boots and de Boer26). In order to genotype the remaining ten SNP (Table 2) we further developed the multiplex PCR method. The development and validation of the adapted multiplex PCR for the new set of ten SNP in various NER genes was based on an approach as described before(Reference Knaapen, Ketelslegers and Gottschalk29). The adapted procedure is defined in the following paragraphs.
NER, nucleotide excision repair; XPA, xeroderma pigmentosis, complementation group A; UTR, untranslated region; GGR, global genome repair; TCR, transcription-coupled repair; RPA, replication protein 1; ERCC1, excision repair cross-complementing group 1; XPC, xeroderma pigmentosis, complementation group C; RAD23B, RAD23 homologue B (one of two human homologues of Saccharomyces cerevisiae Rad23); XPF, xeroderma pigmentosis, complementation group F; ERCC4, excision repair cross-complementing group 4; ERCC2, excision repair cross-complementing group 2; XPD, xeroderma pigmentosis, complementation group D; TFIIH, transcription factor IIH; ERCC5, excision repair cross-complementing group 5; XPG, xeroderma pigmentosis, complementation group G; ERCC6, excision repair cross-complementing group 6; CSB, Cockayne syndrome B.
* According to the Cancer SNP 500 database (http://snp500cancer.nci.nih.gov)(Reference Gu, Zhao and Dinney42, Reference Qiao, Spitz and Shen54–Reference Chen, Kamat and Huang62).
† SNP were coded as low-activity alleles, judged on prior knowledge from published literature and their expected modulating effect on the NER capacity: 0 = homozygous for high-activity allele; 1 = heterozygous, carrying one high- and one low-activity allele; 2 = homozygous for low-activity allele.
XPA, xeroderma pigmentosis, complementation group A; XPC, xeroderma pigmentosis, complementation group C; ERCC6, excision repair cross-complementing group 6; ERCC1, excision repair cross-complementing group 1; RAD23B, RAD23 homologue B (one of two human homologues of Saccharomyces cerevisiae Rad23); ERCC2, excision repair cross-complementing group 2; ERCC5, excision repair cross-complementing group 5.
* According to the Cancer SNP 500 database (http://snp500cancer.nci.nih.gov).
† Neutral non-homologous, non-binding tails are underlined.
PCR primer design and multiplex PCR amplification
Primer 3 software (http://www.broad.mit.edu/cgi-bin/primer/primer3_www.cgi) and Netprimer software (http://www.premierbiosoft.com/netprimer/netprlaunch/netprlaunch.html) were used to design PCR primers (for more detailed information, see Knaapen et al. (Reference Knaapen, Ketelslegers and Gottschalk29)). Primers were obtained from Qiagen. First, the isolated DNA containing SNP was amplified in one eightplex and one duplex PCR reaction. For the eightplex PCR, a 10 μl reaction mixture was prepared containing PCR buffer, 10 mm-deoxynucleotide triphosphates, 50 mm-MgCl2, Platinum® Taq Polymerase (5 U/μl; Invitrogen, Carlsbad, CA, USA) and template DNA (40 ng/μl). The final primer concentrations were 2·0 μm (for XPC-03, ERCC6-01, ERCC1-05, ERCC1-06 and ERCC1-30), 0·8 μm (for XPA-02), 1·6 μm (for XPC-01) and 4·0 μm (for RAD23B-04). For the duplex PCR, the final primer concentrations were 2·7 μm (for XPD-02) and 12·3 μm (for ERCC5-01) (for corresponding rs-numbers and PCR primers, see Table 2). PCR was conducted as follows: denaturation at 94°C for 3 min; thirty cycles of 94°C for 30 s, 63°C for 30 s, and 72°C for 30 s; and a final extension at 72°C for 5 min. After PCR amplification, the products were pooled (5 μl of the eightplex and 3 μl of the duplex PCR product) and incubated (37°C for 45 min) with 4 μl of ExoSAP-IT (Amersham, Roosendaal, The Netherlands) to digest residual deoxynucleotide triphosphates and primers. Enzymes were deactivated at 75°C (15 min).
Multiplex genotyping
Genotyping was performed by single base extension (SBE) using SnaPShot™ as described by the manufacturer (Applied Biosystems, Nieuwerkerk a.d. IJssel, The Netherlands), with some modifications. Briefly, SBE primers were designed using Primer 3 and Netprimer software to bind immediately adjacent 5′ to the specific SNP, with a template-specific part of 20 to 33 bp and a Tm of 66°C to 69°C (Table 2). To facilitate detection of ten polymorphisms in one single run, the length of the extension primers was adjusted to a distinct size by the addition of a non-homologous tail to their 5′ side(Reference Knaapen, Ketelslegers and Gottschalk29). To this end, 5·5 μl of the purified PCR product (containing eleven fragments) were mixed with 2·5 μl of SNaPshot reaction mixture, 1 μl of pooled SBE primers and 1 μl of water. The final concentration for all SBE primers was 4 μm, except for XPA-02 and ERCC5-01 (2 μm). SBE was performed using twenty-five cycles of 96°C for 10 s and 60°C for 30 s. Subsequently, the samples were incubated at 37°C for 1 h with 1 U of shrimp alkaline phosphatase (Amersham, Roosendaal, The Netherlands), followed by enzyme deactivation at 75°C for 15 min. The SBE products were finally analysed by capillary electrophoresis, for which 1 μl of the (fivefold-diluted) SBE product were mixed with 13 μl of deionised formamide and 0·4 μl of Genescan-120 LIZ™ size marker (Applied Biosystems, Foster City, CA, USA). Samples were denatured at 95°C for 5 min and run on an ABI-Prism® 3100 genetic analyser using a 36 cm capillary array and POP-6 polymer. Analyses were performed with Genescan™ software (version. 3.7; Applied Biosystems)(Reference Knaapen, Ketelslegers and Gottschalk29).
Measurement of nucleotide excision repair capacity
A subpopulation of thirty-six healthy volunteers (twenty-eight female and eight male, aged 18–45 years) was selected for the phenotypical assessment of NER. Since we previously observed that ERCC1 expression could be a proxy for NER capacity(Reference Langie, Knaapen and Houben24), selection of subjects occurred according to their ERCC1 genotype. More specifically, samples from homozygous wild types and homozygous carriers of the variant allele for the three studied SNP in ERCC1 were selected for DNA-repair analysis (for the number of subjects selected per ERCC1 genotype, see Table 3).
SNP, single nucleotide polymorphism; XPA, xeroderma pigmentosis, complementation group A; UTR, untranslated region; XPC, xeroderma pigmentosis, complementation group C; ERCC1, excision repair cross-complementing group 1; ERCC2, excision repair cross-complementing group 2; ERCC5, excision repair cross-complementing group 5; ERCC6, excision repair cross-complementing group 6; RAD23B, RAD23 homologue B (one of two human homologues of Saccharomyces cerevisiae Rad23).
* A subpopulation was selected for nucleotide excision repair analysis, according to their ERCC1 genotype.
To phenotypically assess the NER capacity in human lymphocytes, we applied a modified comet assay recently developed in our laboratory(Reference Langie, Knaapen and Brauers30). Basically, this assay measures the ability of a cell or tissue extract to incise substrate DNA containing benzo[a]pyrene-diolepoxide (BPDE)–DNA adducts. Thus, this assay reflects an individual's capacity to recognise and incise damaged DNA, which are important first steps in the NER process. The increase in DNA incisions/breaks, leading to increased tail moments (TM) and percentage DNA in the tail (also known as tail intensity; TI), is indicative of the NER capacity of the cell extracts. After subtracting background levels from all data, the final repair capacity was calculated according to Langie et al. (Reference Langie, Knaapen and Brauers30). Analyses were performed in duplicate and samples of the same subject isolated before and after the intervention were paired for analysis. Nucleoids exposed to 3 μm-BPDE were used as positive controls to correct for inter-assay variations (TM of BPDE-exposed cells ranged between experiments (n 19) from 1·73 (se 0·97) to 6·64 (se 0·77)). Percentage DNA in the tail never exceeded 30 %, indicating that the in vitro repair assay in our experiments was not near to saturation.
Statistical analysis
Differences in DNA-repair capacities and TEAC before and after the intervention were analysed by paired-samples t tests. To investigate the effect of the total number of low-activity alleles on the NER capacity, genotypic polymorphisms were coded as number of low-activity alleles, judged on prior knowledge from published literature and their expected modulating effect on the NER capacity (Table 1): 0 (homozygous for high-activity allele); 1 (heterozygous, carrying one high- and one low-activity allele); 2 (homozygous for low-activity allele) (similar approach as previously reported by Ketelslegers et al. (Reference Ketelslegers, Gottschalk and Godschalk31)). Subsequently, the total sum of low-activity alleles was computed and related to NER capacity using linear regression analysis. For obtaining sufficient numbers per group (n ≥ 3) and subsequent optimal statistical analysis, carriers of five, six, seven or eight low-activity alleles were grouped as carriers of less than nine low-activity alleles. For the same reason, carriers of thirteen, fourteen, fifteen or sixteen low-activity alleles were grouped as carriers of more than twelve low-activity alleles. Stepwise multivariate, linear regression was used to assess the impact of sex, age, TEAC and various polymorphisms on the phenotypically assessed NER capacity. The relationship between NER capacity before and after intervention was assessed by linear regression. Statistical analysis was performed using SPSS (version 15.0; SPSS, Inc., Chicago, IL, USA). In each case, mean values with their standard errors are presented and P < 0·05 was considered statistically significant.
Results
Single nucleotide polymorphisms and genotype frequencies
In Table 1, all analysed SNP, amino acid and base changes related to the polymorphism and the expected effect of the variant allele on the NER capacity are listed. Furthermore, the frequencies of the wild-type, heterozygous and variant alleles as observed in the present study population (n 168) are represented in Table 3. Complete genotypes were obtained from all samples and frequencies were in Hardy–Weinberg equilibrium. For validation purposes seventeen of the 168 samples (10 %) were genotyped twice (eight samples of these seventeen were even genotyped in triplicate) and no differences were found.
Effects of dietary intervention on the nucleotide excision repair capacity
The 4-week intervention with blueberry and apple juice was reported to be effective in the total study population (n 168)(Reference Wilms, Boots and de Boer26). The mean TEAC value was significantly (P < 0·001) increased by the intervention from 781 (se 3·95) to 800 (se 4·02) μm. Similar results were found for the selected subpopulation (n 36); mean TEAC values were significantly elevated (P = 0·032) from 791 (se 6·61) to 805 (se 7·90) μm. However, when studying the effect of the dietary intervention on NER capacity, no clear effects of the 4-week blueberry and apple juice intervention on the phenotypically assessed NER capacity and no significant correlations between NER capacity and TEAC values were observed. NER capacity measured as TM before the intervention correlated strongly with the NER capacity detected after the intervention (Fig. 1). Similar results were obtained by using TI as a read-out (R 2 0·79; P < 0·001; slope 0·97).
Effects of genetic factors on the nucleotide excision repair capacity
As a first approach to investigate the influence of the genetic profile on the phenotypic NER capacity, the total sum of putative low-activity alleles was calculated and related to the NER capacity assessed before the dietary intervention. A significant inverse correlation between the amount of low-activity alleles and the NER capacity was observed, when repair capacity was calculated by using TM values (Fig. 2(a)) as well as when TI values were used (Fig. 2(b)).
The impact of all single genetic polymorphisms, as well as age and sex, on NER capacity was assessed by stepwise multivariate linear regression analysis. Sex and age had no effect on NER, while the single genetic polymorphism XPA G23A was revealed to be a significant predictor of the NER capacity (Fig. 3(a) and (c)) before the intervention. Individuals that were homozygous for the variant allele of XPA G23A (n 5) showed a about three times lower NER capacity as compared with those carrying the homozygous wild-type alleles (n 17). This association, between NER capacity and the SNP XPA G23A, was not affected by the blueberry and apple juice intervention (Fig. 3(b) and (d)).
Gene–diet interactions
Although the diet did not affect the repair capacity in general, it can be postulated that individuals with a certain genetic background may show an altered NER capacity due to the blueberry and apple juice intervention. To study possible gene–diet interactions regarding changes in NER capacity, differences between the NER capacities measured after and before the intervention were calculated for each individual (ΔNER capacity = NER capacityafter − NER capacitybefore). Based on calculations using TM values, improved NER capacity was detected upon dietary intervention in individuals carrying multiple low-activity alleles (Fig. 4(a)); a mean ΔNER capacity of − 0·15 (se 0·13) was observed for carriers of eleven or fewer low-activity alleles (n 29), which increased to a mean ΔNER capacity of 0·36 (se 0·51) and 0·96 (se 1·64) for carriers of twelve (n 3) and more than twelve (n 4) low-activity alleles, respectively. Similar results were obtained by using TI as a read-out of the NER capacity (Fig. 4(b)); carriers of eleven or fewer low-activity alleles (n 29) showed a mean ΔNER capacity of − 0·29 (se 0·25), which increased to 0·87 (se 1·04) and 1·63 (se 1·36) for carriers of twelve (n 3) and more than twelve (n 4) low-activity alleles, respectively. It should be noted here, however, that the effects of the dietary intervention on the NER phenotype was considerably smaller than the effect of the genotype alone.
Furthermore, this ΔNER capacity as an indicator of the intervention effect was tested by multiple stepwise linear regression analysis against all individual SNP, ΔTEAC, age and sex. No effects of sex, age and ΔTEAC were observed. However, the intervention differentially affected ΔNER capacity in subjects that carried the RAD23B Ala249Val polymorphism (Fig. 5); homozygous carriers of the low-activity Val-allele (n 4) benefited more from the intervention by a significantly increased NER capacity as compared with subjects homozygous for the wild-type/high-activity Ala-allele (n 20). Interestingly, homozygous carriers of the RAD23B Val-allele showed about 1·3 times lower NER capacity as their wild-type counterparts (1·49 (se 0·21) v. 1·93 (se 0·30) and 2·99 (se 0·51) v. 4·05 (se 0·66), when using TM and TI values, respectively) before the intervention, while after intervention about 1·5 and about 1·3 times higher NER capacity compared with homozygous carriers of the wild-type Ala-allele (2·52 (se 0·92) v. 1·73 (se 0·23) and 4·88 (se 1·71) v. 3·70 (se 0·55), when using TM and TI values, respectively) was observed in these subjects. In other words, improved NER capacity upon dietary intervention was detected in individuals carrying low-activity alleles.
Discussion
Until now only a few studies have investigated the relationship between genetic polymorphisms in DNA-repair genes and fruit and vegetable intake, mostly in relation to cancer risk and not directly linked to actual repair capacities(Reference van Gils, Bostick and Stern32, Reference Misra, Ratnasinghe and Tangrea33). The present study is one of the first to report a joint effect of genetic polymorphisms in NER-related genes and dietary intervention on the phenotypically assessed NER capacity. Twelve genetic polymorphisms in NER-related genes were assessed and related to an individual's phenotypic NER capacity. Furthermore, the effect of a 4-week dietary intervention with an antioxidant-rich blueberry and apple juice on the phenotypic NER capacity was evaluated and possible genotype–diet interactions were studied. Although the NER capacity was not affected by the dietary intervention in general, carriers of multiple low-activity alleles seemed to benefit from the intervention. Therefore, the present results support the hypothesis that genetic polymorphisms significantly affect NER capacity, which can be further modulated by diet.
In a previous in vitro study(Reference Langie, Knaapen and Houben24), we observed that NER capacity was inhibited by oxidative stress. It can thus be postulated that a diet rich in antioxidants may protect NER. However, no overall effects were found in the present intervention study on NER capacity. This is in correspondence with a recent study from Tyson et al. in which they reported no detectable effects of micronutrient supplementation on NER capacity(Reference Tyson, Caple and Spiers23). A reasonable explanation for the absence of dietary effects in the present study is that the study population consisted of healthy non-smoking volunteers. Since healthy volunteers encounter relatively low levels of oxidative stress, it may be difficult to detect small additional effects of dietary intake of antioxidants. Future studies on the beneficial effects of diets rich in antioxidants should thus focus on, for instance, subjects suffering from diseases that involve increased oxidative stress. This is supported by observations from previous antioxidant intervention trials with oxidatively stressed subjects (for a review, see Møller & Loft(Reference Møller and Loft34)). Additionally, in the present intervention study, inter-individual variations in NER capacity were in the range of about 16-fold, which is similar to variations reported previously in human lymphocytes(Reference Tyson, Caple and Spiers23, Reference Langie, Knaapen and Brauers30, Reference Gaivão, Piasek and Brevik35). Moreover, a strong correlation was observed between the NER capacity before and after the supplementation period (R 2 0·69, P < 0·001 and R 2 0·79, P < 0·001 upon using TM and TI as the read-outs, respectively), indicating that inter-individual variations in NER are maintained over a considerable time, which has been reported before for NER(Reference Tyson, Caple and Spiers23) as well as for BER(Reference Collins, Dusinska and Horvathova36). In contrast, BER seems to be modifiable by the intake of antioxidants even in healthy subjects(Reference Collins, Harrington and Drew21), but data are not consistent(Reference Møller and Loft34, Reference Møller, Vogel and Pedersen37).
Our second aim was to further elucidate the genotype–phenotype relationships with respect to the NER process. Although the majority of genes encoding proteins involved in DNA-repair processes are polymorphic(Reference Mohrenweiser, Xi and Vazquez-Matias38), only a limited number of studies have examined the actual phenotypic effects of these genetic polymorphisms. In one of our previous studies, we observed a significant correlation between the ERCC1 expression and the phenotypic NER capacity in vitro (Reference Langie, Knaapen and Houben24). ERCC1 encodes a subunit of the NER complex, which is required for the incision step of NER(Reference Sancar, Lindsey-Boltz and Unsal-Kaçmaz8, Reference Sugasawa9). However, we did not observe any significant correlation of the studied ERCC1 polymorphisms with the NER capacity in the present study. Since the functional relevance of SNP in ERCC1 remains unclear and inconsistent results have been reported(Reference Zhou, Liu and Park39, Reference Zienolddiny, Campa and Lind40), further investigation into the effect of these polymorphisms on DNA repair is needed.
On the other hand, subjects carrying a high number of putatively low-activity alleles showed lower NER capacity as compared with subjects carrying only a few low-activity alleles. This approach, looking at the combined effect of multiple gene variants rather than investigating the effect of a single nucleotide polymorphism, has been applied before: several studies have reported an association between the number of putatively high-risk alleles in DNA-repair genes and levels of bulky DNA adduct(Reference Ketelslegers, Gottschalk and Godschalk31, Reference Matullo, Peluso and Polidoro41), while others have observed increased cancer risk with increasing number of putative high-risk alleles(Reference Gu, Zhao and Dinney42, Reference Crew, Gammon and Terry43). All these observations suggest that, at the individual level, studying the combined effect of multiple gene variants may be important in order to define DNA-repair capacity.
Subsequently, to investigate which polymorphisms may have the highest contribution to the inter-individual variations in NER capacity, multivariate linear regression analysis was performed. The common SNP XPA G23A seemed to be the most relevant polymorphisms for defining NER capacity. Individuals homozygous for the wild-type allele (GG) exhibited a significantly three times higher NER capacity (approximately) compared with carriers of at least one variant allele. The XPA protein is involved in both global genome repair and the transcription-coupled repair pathway of the NER process, playing an essential role in the assembly of the pre-incision complex(Reference Sugasawa9). The common G → A single-nucleotide substitution in the 5′ untranslated region of the XPA gene is located four nucleotides upstream from the ATG start codon(Reference Butkiewicz, Rusin and Harris44). The functional relevance of this SNP is unknown; however, it has been demonstrated that the 5′ untranslated region may regulate gene expression through post-transcriptional control mechanisms(Reference Akiri, Nahari and Finkelstein45, Reference Larsen, Amri and Mandrup46). In addition, several studies have shown that individuals with this G → A substitution in the 5′ untranslated region of XPA have a increased risk of lung cancer(Reference Wu, Zhao and Wei12, Reference Vogel, Overvad and Wallin47, Reference Butkiewicz, Popanda and Risch48). Furthermore, in agreement with the present results, Wu et al. observed a more efficient NER capacity in subjects carrying the XPA GG genotype(Reference Wu, Zhao and Wei12). The association between this XPA polymorphism and NER capacity was not affected by the blueberry and apple juice intervention, indicating that the XPA G23A SNP might be regarded as a predictor for the NER capacity.
Although we initially did not observe an overall effect of diet on NER capacity, it was still possible that a subgroup may benefit from the intervention due to gene–diet interactions. Indeed, enhanced NER capacity was detected in subjects carrying a sum of twelve or more low-activity alleles. Although we did not observe a general effect of the intervention on the NER capacity in our healthy non-smoking study population, we were able to detect a beneficial effect of the intervention in individuals with an initial low NER capacity. Similar observations were reported by Guarnieri et al. detecting beneficial effects of antioxidants only among poorly nourished subjects with low repair activity(Reference Guarnieri, Loft and Riso49). As mentioned above, future studies on the beneficial effects of diets rich in antioxidants should thus focus more on specific susceptible subpopulations.
Furthermore, the single polymorphism RAD23B Ala249Val seemed to be a predictor for the intervention effect (ΔNER capacity). Subjects homozygous for the variant Val-allele of RAD23B had enhanced NER capacity after the intervention, while NER capacity in carriers of the wild-type Ala-allele was unaffected. Moreover, increased ΔNER capacity was observed in subjects that carry two low-activity alleles of RAD23B Ala249Val. The protein encoded by RAD23B binds XPC, forming a heterodimeric complex(Reference Sugasawa9). In the global genome repair pathway XPA binds this protein complex, which is essential for the recruitment of all subsequent NER factors in the pre-incision complex. Although the biological function of the RAD23B Ala249Val polymorphism is not clear, the variant alleles Val/Val were associated with increased lung cancer risk and higher BPDE sensitivity as compared with the homozygous Ala/Ala wild types(Reference Lin, Swan and Shields2, Reference Shen, Berndt and Rothman3), which is in agreement with the low NER capacity that we detected before the intervention in homozygous carriers of the Val-allele. However, these data need to be interpreted with care, because the group of subjects homozygous for the RAD23B Val-allele is small (n 4), and all carried a high number of low-activity alleles. In other words, improved NER capacity upon dietary intervention was especially detected in individuals carrying a high number of low-activity alleles. Nonetheless, our observations suggest that both genetic as well as environmental factors such as diet can modulate an individual's NER capacity, separately or through interaction with each other.
It is not yet clear how this interaction between the genotype and antioxidant intake can be explained. Several studies have suggested that some dietary antioxidants may confer protective properties through mechanisms that are unrelated to their conventional free-radical scavenging abilities, such as up-regulation of antioxidant defence, xenobiotic metabolism or DNA-repair genes(Reference Cooke, Evans and Mistry50, Reference Silva, Gomes and Coutinho51). For example, quercetin and vitamin C were shown to induce different DNA damage-responsive signalling pathways (for example, p53 and activator protein-1 (AP-1)/NF-κB) that can subsequently enhance the expression of, for instance, NER genes(Reference Lunec, Holloway and Cooke52, Reference Ye, Goodarzi and Kurz53). However, results from various in vivo intervention studies have been equivocal(Reference Cooke, Evans and Mistry50). Therefore, further and larger studies are needed to clarify possible correlations between an individual's antioxidant capacity and DNA-repair capacity, both in the whole population as well as in several subgroups.
In the present study, we report a joint effect of genetic polymorphisms in NER-related genes and a dietary intervention on the phenotypically assessed NER capacity. Overall, the present results show that genetic factors have more impact on the NER capacity as compared with the effects of the blueberry and apple juice intervention. Furthermore, the common genetic polymorphism XPA G23A might be a predictor for the NER capacity, as it was not affected by the dietary intervention. Still, the present study suggests that the combined effect of multiple gene variants may be more important than the investigation of single nucleotide polymorphism in order to define an individual's DNA-repair capacity. Improved NER capacity upon dietary intervention with blueberry and apple juice was detected in individuals carrying a high number of putative low-activity alleles. In conclusion, studies of genotype–phenotype interactions seem to be helpful in the identification of susceptible subpopulations that may benefit from specific dietary interventions.
Acknowledgements
We thank Ralph W. H. Gottschalk for his assistance in the set up of the multiplex SNaPShot PCR and Stijn B. J. Lumeij for his technical support in the in vitro DNA-repair assays.
Part of the studies was supported by the European Network of Excellence (NoE) ‘Environmental Cancer risk, Nutrition and Individual Susceptibility’ (ECNIS), sixth Framework programme (FP6), FOOD-CT-2005-513943.
The intervention study was funded by the Netherlands Organisation for Health Research and Development (Nutrition: Health, Safety and Sustainability programme, grant no. 01412012).
The project was conceived of by J. C. S. K., R. W. L. G. and F. J. vS. The intervention was coordinated by J. C. S. K. and L. C. W. TEAC analysis was performed by L. C. W., while DNA-repair measurements were conducted by S. A. S. L. Modification of the multiplex PCR and genotyping was carried out by S. A. S. L., with help from S. H. Data analysis was carried out by S. A. S. L. and R. W. L. G., and all authors contributed in writing the manuscript.
The authors declare no conflicts of interest.