Main

Telomeres, located at the ends of chromosomes, consist of long TTAGGG nucleotide repeats and an associated protein complex. Chromosome ends are protected from end-to-end fusion and degradation by this telomere complex, termed shelterin (de Lange, 2005). The TTAGGG repeats shorten with each cell division, and eventually reach a critical state, at which time cellular senescence and/or apoptosis is normally triggered (Rodier et al, 2005). Tumour cells may survive cellular crisis in the absence of chromosomal stability through the activation or inactivation of alternative pathways. Breast cancer fits the paradigm of dysfunctional telomere-induced genomic instability, because the transition of breast duct hyperplasia to ductal carcinoma in situ likely results from a period of telomere crisis (DePinho, 2000; Chin et al, 2004). As breast cancer progresses further to invasive and metastatic stages, telomere dysfunction and genomic instability become more apparent (Nishizaki et al, 1997; Buerger et al, 1999; Chin et al, 2004). As cells progress through the latter stages of carcinogenesis, telomeres become relatively stable. In addition, low-telomere DNA content was found to be an independent predictor of decreased survival in comparisons of breast cancer specimens to normal tissues (Chin et al, 2004; Fordyce et al, 2006).

Most genes involved in telomere biology are highly conserved between species and have limited nucleotide diversity in humans (de Lange, 2004; Savage et al, 2005). We hypothesized that common genetic variation (minor allele frequency (MAF) greater than 5%) in the form of single nucleotide polymorphisms (SNPs) in these genes could affect cancer risk. This hypothesis was investigated in a population-based case–control study of breast cancer study in Poland, in which we genotyped 24 common SNPs that captured most of the common genetic variation in five genes important in telomere biology. The studied genes included telomerase (TERT (protein name), TERT (HUGO gene name), 5p15.33) (Collins and Mitchell, 2002), telomerase-associated protein (TP1, TEP1, 14q11.2) (Poderycki et al, 2005), telomeric repeat-binding factor 1 (TRF1, TERF1, 8q13) (Smogorzewska et al, 2000), telomeric repeat-binding factor 2 (TRF2, TERF2, 16q22.1) (Chong et al, 1995; Broccoli et al, 1997) and protection of telomeres 1 (POT1, POT1, 7q31.33) (Baumann and Cech, 2001).

Materials and methods

Study population

The design of this population-based breast cancer case–control study has been described (Garcia-Closas et al, 2006a). Eligible cases included women aged 20–74 years who were Polish residents of either Warsaw or Łódź with pathologically or cytologically confirmed in situ or invasive breast cancer, newly diagnosed in 2000–2003. An estimated 90% of eligible cases were identified through a rapid identification system at five participating hospitals. Information from Cancer Registries was used to identify the remaining 10% of eligible breast cancer cases. Eligible control subjects were residents of Warsaw and Łódź who did not have a history of breast cancer at enrollment. Controls were randomly selected from population lists, and frequency-matched to breast cancer cases by city and 5-year age groups. Women provided a personal interview on known and suspected risk factors. Venous blood samples were collected by a trained nurse. The study protocol was reviewed and approved by local and National Cancer Institute (NCI) Institutional Review Boards. All participants provided written informed consent. Of the 3037 eligible cases and 3639 eligible controls identified, 2386 (79%) cases and 2502 (69%) controls agreed to participate in the personal interview. The present study is limited to women with blood DNA samples: 1995 cases (6% in situ) and 2296 controls, which represented 84 and 94%, respectively, of the study population.

Laboratory methods

Genomic DNA for genotype analyses was isolated from buffy coat or whole blood samples using the Autopure LS® DNA Purification System (Gentra Systems Inc., Minneapolis, MN, USA). Twenty-four SNPs in POT1, TEP1, TERF1, TERF2, and TERT were genotyped by investigators blinded to case–control status, using TaqMan or MGB Eclipse platforms at the Core Genotyping Facility of the Division of Cancer Epidemiology and Genetics, NCI (Table 1). Assay conditions are available at http://snp500cancer.nci.nih.gov (Packer et al, 2006). When possible, rs numbers based on the dbSNP database are indicated (http://www.ncbi.nlm.nih.gov/SNP). If an rs number has not yet been assigned, an E number (e.g. E3675_301) is provided, based on nomenclature from the SNP500Cancer project (Packer et al, 2006). Single nucleotide polymorphism locations were determined using the guidelines of the Human Genome Variation Society (den Dunnen and Antonarakis, 2001).

Table 1 Association between 24 single nucleotide polymorphisms in five genes important in telomere biology and breast cancer risk among cases and controls
Full size table

A total of 100 duplicate DNA pairs were ⩾98% concordant for each SNP with the exception of TERF1 IVS9-163T>C (rs3863242, 97%) and TERT Ex2-659G>A (rs2736098, 94%). Genotypes were called for >98% of all SNPs. Genotype frequencies for all loci were in Hardy–Weinberg equilibrium among controls.

Single nucleotide polymorphism selection

Initial SNP selection criteria included MAF greater than 5% in Caucasians from SNP500 Cancer (n=31), even spacing across the gene, SNPs with potential functional implications and/or patterns of nucleotide diversity and linkage disequilibrium (LD) previously determined through extensive re-sequence analysis (Savage et al, 2005; Packer et al, 2006) and assay availability at the time of SNP selection. The SNPs selected using these criteria were evaluated as haplotype-tagging SNPs compared with all common SNPs identified in the prior re-sequence analysis using tagSNPs (Stram, 2004) and TagZilla (http://tagzilla.nci.nih.gov/). R2H was the pairwise correlation coefficient between SNPs determined by these programs. SNPs with R2H ⩾0.8 were considered highly correlated.

TEP1 (54 exons, 40.7 kilobase pairs (kbp)) has minimal LD and eight common SNPs in the 31 SNP500 Caucasians. The five TEP1 SNPs genotyped (Table 1) gave an R2H of 0.84, indicating representative coverage of common genetic variation across TEP1. TERF1 (10 exons, 15.3 kbp) has very limited nucleotide diversity with only four common SNPs in SNP500 Caucasians between introns 7 and 9 (Savage et al, 2005). Three of these SNPs were genotyped and very good correlation for the fourth SNP was noted, R2H=1.0. TERF2 (10 exons, 30.3 kbp) has only four common SNPs between introns 1 and 8 and a very small common haplotype block between introns 6 and 7 (Savage et al, 2005). TERF2 IVS6+27G>A and IVS7-42T>C were highly correlated with the other SNP in this block, TERF2 IVS8+95T>C (E3675_301) (R2H>0.8), but did not cover the SNP in intron 1 (TERF1 IVS1-5C>T, E5055_301), which only had a MAF of 5% in SNP500 Caucasians. Studies of genetic variation in TERT (41.9 kbp, 16 exons) are complex due to low nucleotide diversity and limited LD (Savage et al, 2005). The 10 SNPs genotyped in our study spanned 43 kbp from −1654A>G to Ex16+203C>T and were representative of common genetic variation, R2H=0.63. We were unable to genotype TERT Ex14+7C>T (E3661_301, H1001H) due to lack of assay availability, which would have increased the R2H to 0.83; however, we did genotype Ex16+203C>T (rs2853690), which was only 1776 bp 3′ of TERT Ex14+7C>T. The four SNPs genotyped in POT1 (17 exons, 74.7 kbp) spanned 73.1 kbp (−1386G>A through IVS13-98T>G), a region with strong LD and 11 common SNPs in SNP500 Caucasians (Savage et al, 2005). These SNPs (Table 1) were good representatives of common genetic variation across POT1, R2H=1.0.

Statistical analyses

Odds ratios (OR) and 95% confidence intervals (CI) from logistic regression models with dummy variables for matching factors (age in 5-year categories and study site (Warsaw or Łódź)) were used to estimate relative risks for the genotypes examined. The association between genotypes and breast cancer risk was tested using a 2 degrees of freedom (df) likelihood ratio test and a trend test. Heterogeneity of genotype ORs among groups of women defined by age categories and family history of breast cancer in first-degree relatives were evaluated by introducing interaction terms in logistic regression models. A positive family history was defined for women reporting one or more first-degree relatives diagnosed with breast cancer in the study questionnaire. An additive genetic model was assumed in interaction analyses. Age was considered as a continuous variable in tests for genotype–age interactions. Haplotypes were constructed for cases and controls using PHASE v2.1 (Stephens et al, 2001; Stephens and Donnelly, 2003) and HaploStats (Lake et al, 2003). The global case–control permutation test was performed using PHASE v2.1 (Stephens et al, 2001; Stephens and Donnelly, 2003). HaploStats (Lake et al, 2003) was used also to determine the global score P-value, haplotype frequencies, ORs and 95% CIs.

Results

Most cases (74%) and controls (69%) in the study were postmenopausal, and cases were diagnosed at an average age (standard deviation) of 56 (±10) years. The established risk factors were associated with breast cancer risk in comparable direction with similar estimates of magnitude reported by others (Garcia-Closas et al, 2006b).

Case–control analyses showed no statistically significant associations between the 24 SNPs in TEP1, TERF1, TERF2, TERT and POT1 and risk of breast cancer (Table 1). Specific haplotypes derived from the evaluated SNPs were also not associated with increased risk of breast cancer in this study (data not shown). There were no statistically significant associations among age, SNP and breast cancer risk (Supplementary Table 1).

Case–control analyses suggested inverse associations between homozygous variants of TERT and breast cancer risk at two SNP sites, TERT-1654A>G (OR 0.85, 95% CI 0.72–1.02) and TERT Ex2-659G>A (A305A) (OR 0.76, 95% CI 0.58–1.00) (Table 1). The inverse association of TERT Ex2-659G>A (A305A) and two other linked TERT SNPs appeared to be limited to individuals with a family history of breast cancer in first-degree female relatives, −1381C>T (OR 0.73, 95% CI 0.53–1.00), −244C>T (OR 0.66, 95% CI 0.46–0.95), and Ex2-659G>A (A305A) (OR 0.57, 95% CI 0.39–0.84) (Table 2 and Supplementary Table 2). These SNPs were not significantly related to family history of cancer among the control population, and analyses of breast cancer cases with a family history of breast cancer compared with all controls, regardless of family history, produced similar results (data not shown). These three SNPs appeared to be in LD by D′, but only −244C>T and Ex2-659G>A were strongly correlated with R2H of 0.79. TERT-1381C>T, −244C>T, and Ex2-659G>A had high pairwise D′ values, but the R2H showed that only −244C>T and Ex2-659G>A were highly correlated. This suggests that the associations seen in TERT −1381C>T may not be related to the effects of LD between this SNP, −244C>T and Ex2-659G>A. However, the statistical association seen in −244C>T and Ex2-659G>A could be because they are highly correlated, and in effect, measure the same risk marker.

Table 2 Association between selected single nucleotide polymorphisms in TERF2 and TERT and breast cancer risk among cases and controls, stratified by family history of breast cancer in first-degree female relatives
Full size table

Haplotype analyses were performed for all SNPs studied in TERT and for each of the two major haplotype blocks in TERT (block 1: −1654A>G, −1381C>T, −967T>C, −244C>T and Ex2-659G>A, block 2: IVS10+269C>T and Ex16+203C>T). There were no significant associations for haplotypes in the primary case–control analysis (data not shown). However, a block 1 haplotype (ATCCA) in TERT was associated with protection from breast cancer when only individuals with a family history of breast cancer were studied (OR 0.61, 95% CI 0.38–0.97, P=0.034).

In addition, women with a family history also showed a borderline statistically significant positive association between TERF2 IVS-42T>C variant alleles and breast cancer risk (OR 1.57, 96% CI 0.97–2.55, P interaction 0.06). No other associations were significantly modified by family history of breast cancer (Supplementary Table 2).

Discussion

To our knowledge, this is the first study to investigate genetic variation within genes important in telomere biology (POT1, TEP1, TERF1, TERF2 and TERT) and breast cancer risk. The SNPs genotyped were representative of common genetic variation across the genomic region of interest, and showed no significant overall associations with breast cancer risk. However, data suggested association between variants in TERT among women with a positive family history of breast cancer.

TERT Ex2-659G>A showed a borderline statistically significant association with a reduced risk of breast cancer in analysis of all cases and controls, which appeared to be stronger for individuals with a family history of breast cancer. Similar associations of two other SNPs, −1381C>T and −244C>T, in individuals with a family history of breast cancer were also noted. TERT −244T>C was noted to have increased telomerase activity related to the T allele in a recent study of non-small cell lung cancer (Hsu et al, 2006). TERT −1381C>T also appears to be a functional SNP. Studies of promoter function at this site (noted at −1327 by the authors, but with the same rs number, rs2735940) suggested longer telomere length in with TT homozygotes compared with CC (Matsubara et al, 2006). Our findings suggested that variants in TERT could have an effect in individuals already at increased genetic risk of breast cancer, although the number of individuals with a family history of breast cancer was small.

TERF2 IVS6+27G>A (E3673_301) was also associated with a reduced risk of breast cancer in individuals with a family history of breast cancer, however, the functional significance of the SNP is unknown. It does not appear to affect an intron–exon splice site (Conde et al, 2004).

The SNPs evaluated in this study were chosen based on previous knowledge of common genetic variation resulting from re-sequence analysis, captured most of the common variation in the five studied genes (i.e. POT1, TEP1, TERF1, TERF2 and TERT), and could be related to breast cancer risk based on the role suggested for telomere biology in this disease (Baykal et al, 2004; Wacholder et al, 2004; Savage et al, 2005). Although associations with less common SNPs are possible, our data indicate that common variation in these genes is unlikely to substantially affect overall breast cancer risk. The associations of TERT −1381C>T, −244C>T, Ex2-659G>A and the corresponding haplotype in individuals with a family history of breast cancer are intriguing and warrant follow-up in independent studies.