Our results showed that, after adjustment for known breast cancer risk factors, shorter telomeres on chromosome 9p-short (allelic shorter version) were strongly associated with an increased risk of breast cancer. This finding supports our hypothesis that individuals who possess short telomeres on chromosome 9 are at increased risk of breast cancer. Our results are corroborated by previous reports showing that telomere dysfunction is involved in the development of breast cancer (6
). For example, marked shortening of chromosome telomeres were reported in grade II and III breast cancers (35
). Very short telomeres are also frequent genetic alterations in premalignant lesions (6
), suggesting that telomere dysfunction is an early event in breast carcinogenesis.
A study of genetically engineered mice provided evidence that the average telomere length in a cell may be less important for cancer risk than short telomeres on specific chromosome arms because chromosome fusions associated with telomere dysfunction occur preferentially on chromosome arms with the shortest telomeres (37
). In humans, there are 23 pairs of chromosomes and 92 telomeres, and chromosome-specific telomere lengths are highly polymorphic between chromosomal arms (38
). One potential implication of this chromosome arm-specific telomere length variation is that chromosome arms bearing the shortest telomeres may predispose to the chromosome aberrations and therefore have an impact on the evolution of tumors. This concept is supported by several studies demonstrating that chromosome arms with the shortest telomeres were more often found in the telomere fusions leading to chromosome instability (37
). Our observation that shortest 9p telomere (shorter 9p-short telomeres) was strongly associated with breast cancer risk supports this concept. To the best of our knowledge, our study is the first to report that shorter telomere on chromosome 9p-short is strongly associated with breast cancer risk. If confirmed by other studies, chromosome 9p telomere length, in combination with other risk biomarkers, could potentially improve breast cancer risk assessment for individual women in general population. Based on these promising findings, it is reasonable to anticipate that lengths of other arm-specific telomeres might be associated with breast cancer risk. Thus, future studies should aim to identify such telomeres.
Our results also indicated that telomere lengths on chromosome 9q were not associated with breast cancer risk. We observed no significant correlations between 9p and 9q telomere lengths in controls, except a weak correlation between 9p-short and 9q-short, suggesting telomere lengths on 9p or 9q are independent events. It is well documented that frequent chromosomal abnormalities in early stage breast cancers only involves a handful of chromosome arms, including gains of 1q, 8q, 17q and 20q and losses of 8p, 9p, 16q and 17p (5
). Therefore, it is not surprising to find that chromosome 9q telomeres were not associated with breast cancer risk in our study because chromosome 9q aberration is not a common abnormality in breast tumors. In this study, we also found that there was no significant difference in mean overall (cell total) telomere length between cases and controls () (31
). It would be expected that overall telomere length would not be informative due to the inclusion of large number of ‘irrelevant’ telomeres in the measurement. This may in part explain the null findings by three recent studies that examined the association of overall (cell total) telomere length in blood leukocytes and breast cancer risk (32
The case–control differences reported here are unlikely to be ascribed to assay bias because the samples were processed and measured blinded to case–control status. Measurements of four chromosome 9-arm telomere lengths were made simultaneously from the same cells captured by the fluorescent imaging system. While there were significant case–control differences for chromosome 9p telomere lengths, the mean telomere lengths of chromosome 9q and cell total were very similar between cases and controls. During the study, a systematic quality control plan was implemented to ensure the consistent efficiency of fluorescent in situ
hybridization (the coefficient of variation of the total telomere lengths among 20 repeats of the control cell
12.4%). Co-hybridization of a FITC-labeled chromosome 9-specific probe (green) was used as an internal control for hybridization efficiency, and only slides that showed bright green chromosome 9 signals were accepted. The samples that failed to meet these quality control standards were rejected and repeated (3% of the samples). Additionally, relative telomere lengths of chromosome 9 arms were defined as the ratio of chromosome 9 telomere lengths to the total telomere length, thus minimizing the assay variation between the individual samples. Bias in telomere length measurement is thus unlikely.
Given that this is a case–control study, a theoretical concern is that telomere length in leukocytes is affected by case status (reverse causality). Data by previous studies and by us indicated that the mean overall telomere length of blood leukocytes in breast cancer patients was not significantly shorter than in healthy women controls (31
), suggesting there is no significant shortening of blood leukocyte telomere length associated with having breast cancer. Although previous studies (42
) suggested that chemotherapy and/or radiotherapy can induce telomere shortening in leukocytes, all the blood samples in our study were drawn before any chemotherapy and radiotherapy treatments. Thus, reverse causality is not a plausible explanation for our results.
Recall bias might influence information about self-reported breast cancer risk factors in a case–control study where the data were collected after the diagnosis of cancer. We compared chromosome 9 telomere lengths in control subjects by numerous variables from the questionnaire and did not find any significant differences in the mean chromosome 9 telomere lengths between subgroups defined by race, age, smoking status, alcohol drinking, menopausal status, physical activity in the teens, hormone replacement therapy, history of pregnancy, family history of cancer, education and income (data not shown). Cases and controls were closely matched by age, and age was included as continuous variable in all the logistic models for adjustment. Thus, age would not confound the telomere results.
In conclusion, our data provide the first evidence that short telomere on chromosome 9p is strongly associated with breast cancer risk. If confirmed in future studies, chromosome 9p telomere length could be incorporated into the current prediction models to significantly enhance breast cancer risk prediction. Better risk assessment would improve the efficiency of both population-based preventive programs, such as screening mammography, as well as individual-based preventive strategies such as chemoprevention by targeting women who are at the greatest risk for breast cancer.