Mean Telomere Length and Risk of Incident Colorectal Carcinoma: A Prospective, Nested Case-Control Approach

Telomeres are tandem repeats of DNA sequences—special chromatin structures—located at the ends of eukaryotic chromosomes. These structures are believed to protect the telomeric regions from recombination and degradation, thus avoiding a DNA damage cellular response (1, 2). Genome instability is a hallmark of tumorigenesis and is a wildly accepted view as a major contribution to the development of cancer, including colorectal carcinoma (CRC; refs. 1, 3). Furthermore, recent studies have implicated telomere length shortening as an independent marker for the progression and/or prognosis of CRC (4-11) based on the comparison of paired cancerous and adjacent noncancerous tissue specimens from the same individuals. However, to date, no studies have been conducted to examine the relationship of telomere length as a risk predictor with incident CRC.

We thus prospectively examined the possible association of mean peripheral blood leukocyte (PBL) telomere length with risk of incident CRC using a nested, matched case-control sample drawn from the prospective Physicians' Health Study cohort.

We used a nested case-control design within the Physicians' Health Study, a randomized, double-blinded, placebo-controlled trial of aspirin and β-carotene initiated in 1982 among 22,071 male, predominantly white (>94%), U.S. physicians, 40 to 84 y of age at study entry (12). Before randomization, 14,916 participants provided an EDTA-anticoagulated blood sample that was stored for genetic analysis. All participants were free of prior myocardial infarction, stroke, transient ischemic attacks, deep venous thrombosis/pulmonary embolism, and cancer at study entry. Yearly follow-up self-report questionnaires provided reliable updated information on newly developed diseases. For all reported incident CRC events occurring after study enrollment, medical records were requested and reviewed by an end-points committee.

The nested case-control design has been previously described (13). For each case, one or two controls matched by age ±2 y, smoking history (never, past, or current), and length of follow-up were chosen among those subjects who remained free of CRC. The present study consisted of white males only: 76 1-to-1 matched pairs and 115 1-to-2 matched pairs. Median length of follow-up since randomization for the cases was 6.02 y (interquartile range, 3.22-8.95). The study was approved by the Brigham and Women's Hospital Institutional Review Board for Human Subjects Research.

Genomic DNA was extracted from whole blood using the QIAmp Spin Column protocol (Qiagen). Telomere length was determined by a previously described and validated, unified quantitative PCR protocol (14). In brief, two master mixes of PCR reagents were prepared: one for telomere reaction and one for single-copy gene reaction (36B4 on chromosome 12). Telomere repeat copy number to single gene copy number (T/S) ratio was determined on an ABI 7900HT Sequence Detection System (Applied Biosystems) in a 384-well format using the following PCR protocol: 95°C for 15 min to activate Taq polymerase, 40 cycles of denaturation at 95°C for 15 s, and annealing-extension at 54°C for 2 min. Each 5 μL amplification reaction volume contained 1× Qiagen QuantiTect SYBR Green Master Mix (Qiagen) and 10 ng of template DNA. The primer sequences used were described elsewhere (14, 15). All samples for both the telomere and single-copy gene amplifications were done in duplicate on the same 384-well plate. C^t value assignment was carried out by two independent observers, and if necessary, a complete regenotyping was done. The C_t values generated were used to calculate the T/S ratio for each sample using the following equation: T/S = 2^−ΔCt (where ΔC^t = C_{t single-copy gene} − C_t telomere). Results were scored blinded as to case-control status.

As previously noted, the observed T/S ratios had a skewed distribution, and the data were log_e transformed. The T/S ratios between cases and controls were compared using the nonpaired t test. Spearman's correlation analysis was used to assess the effects of age, current smoking, body mass index (BMI), alcohol use, and exercise on relative T/S ratios among all subjects. Risk ratio of CRC associated with log_e-transformed T/S ratios were calculated by conditional logistic regression analysis, adjusting for age, smoking status, and length of follow-up since randomization and further controlling for randomized treatment assignment, BMI, alcohol use (daily/weekly/rarely), and exercise (daily/weekly/rarely). All analyses were carried out using SAS 9.1 package (SAS Institute, Inc.). A two-tailed P value of 0.05 was considered a statistically significant result.

Baseline characteristics of the study participants are shown in Table 1. In concordance with published data, an inverse relationship of the observed T/S ratios with age among all subjects was found (Spearman correlation = −0.093; P = 0.038; Supplementary Table S1) but not BMI, current smoking, daily alcohol use, nor daily exercise (Supplementary Table S1). However, the observed T/S ratios were similar between cases and controls (P = 0.650; Table 1). Furthermore, no association of mean T/S ratios with risk of incident CRC was found in the regression analysis (adjusted odds ratio, 1.249; 95% confidence interval, 0.863-1.808; P = 0.238; Table 2). Stratified analysis by median follow-up time since randomization was also done, and again, similar null findings were observed. The coefficients of variation of the telomere, single-gene, and T/S ratio duplicate assays were all <2%, respectively.

To the best of our knowledge, the present nested, matched case-control investigation is the first to examine the relationship of mean leukocyte telomere length with risk of incident CRC, and we found no evidence for an association. In concordance with previous reports, the present study found an inverse correlation between mean leukocyte telomere length and age.

Recent studies have shown telomere length shortening as an independent marker for the progression and/or prognosis of CRC (4-11). However, these studies were based on telomere length measurements in colonocytes of paired cancerous-noncancerous tissue specimens from the same individuals, as opposed to PBL as used in the current investigation. As previously noted, Craig et al. (16) found no correlation of telomere length with age in normal colonocytes from normal individuals; however, telomere shortening with age was observed in other tissues, including blood. Furthermore, the study by O'Sullivan et al. (10) also observed an association between telomere length in PBL and gastric tissue but not between PBL and colon. Taken altogether, these data suggest that telomere dynamics in colonocytes differ from other tissues, including PBL (10, 16), and may be partly related to the dynamics of telomere-telomerase complex in cell proliferation (8) and exposure/responses to oxidative damage (17). As no prospective, epidemiologic data on mean PBL telomere length and risk of incident CRC are available, a cross-reference comparison cannot be achieved on the present null findings in relation to CRC risk. Nevertheless, the present investigation suggests that PBL telomere length may not play a role in the underlying pathogenesis of CRC.

The nature of the present investigation in which the determination of a case status was based solely on the subsequent development of disease rather than on any arbitrary selection criteria designed by the investigators greatly reduces the possibility of bias and confounding. Nonetheless, our study population consists of white males only so the data cannot be generalized to other ethnic groups, women, and populations with different socioeconomic status. Furthermore, no baseline information on colorectal screening is available in the present sample population (e.g., removal of polyp during colonoscopy). Thus, the potential confounding effect of colorectal screening cannot be assessed in the present context. In our study, we had the ability to detect, based on the present sample size, assuming 80% power, at an α of 0.05, a true difference in the mean telomere length ratio of <−0.153 or >0.153. Because of the stringent matching criteria used, 76 cases could only be matched to one control; thus, power to detect differences may differ by matched case-control subgroups.

As telomere biology represents a rapidly expanding research field, further thoughts on future development of quantitative PCR technique are worthwhile. Of relevant note, a multiplex monochrome quantitative PCR method for telomere length measurement has recently been described (18). Thus, future methodologic comparison with the gold standard Southern blot method is worthwhile.

In conclusion, the present prospective, nested case-control study of middle-aged white U.S. men found no evidence for an association of leukocyte mean telomere length with risk of incident CRC. If corroborated in future prospective studies, our present findings further suggest that mean leukocyte telomere length measurement may not be a useful indicator for risk assessment.

No potential conflicts of interest were disclosed.

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