Introduction

AURKA (STK15, Aurora kinase) is involved in cell cycle regulation, in particular the passage from G2 to M, through the formation of mitotic spindle formation [1]. While this gene is found to be amplified in many tumor types, including breast cancer [2], the loss of over expression of AURKA has been correlated with the transition of in situ to invasive ductal carcinoma [3].

Ewart-Toland et al. [4] found that the polymorphic Ile allele of F31I was more effective at transforming rat cells to a more malignant phenotype. This and other polymorphisms have been associated with breast cancer risk in four retrospective case-control studies (Table 1). Sun et al. found a significant increase in risk (OR 1.76, 95% CI 1.16-2.66) associated with the homozygous state of the Ile allele at the F31I (rs2273535) polymorphism of AURKA, while Dai et al. did not (OR 1.2, 95% CI 0.9-1.6), in hospital-based [5] and population-based [6] case-control studies of Han Chinese women. Lo et al. also did not observe a statistically significant (p=0.32 of association under an additive model) association between this polymorphism and breast cancer risk in a hospital based case-control study in Taiwan [7], however they did observe association between haplotypes of AURKA and breast cancer. It is interesting to note that the prevalence of the homozygous state of the Ile allele in Asian populations (~45%) is drastically different that that of Caucasians (~4%). Egan et al. described an increase in risk associated with a compound genotype of two non-synonymous polymorphisms, F31I and V57I (rs1047972), with individuals homozygous for the 31I and 57V alleles having a nearly 2-fold increase in risk of postmenopausal invasive breast cancer [8]. We studied these two polymorphisms in a case-control study nested within the prospective Nurses’ Health Study.

Table 1 Prior reports of association between the F31I Polymorphism and breast cancer risk
Full size table
Table 2 Association between AURKA polymorphism and breast cancer risk in the Nurses’ Health Study
Full size table

Materials and methods

Genotyping assays for the AURKA polymorphisms (F31I, rs2273535; and V57I, rs1047972) were performed by the 5′ nuclease assay (TaqMan) on the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA). TaqMan primers, probes, and conditions for genotyping assays are available on request from the authors. Our study included a total of 1259 incident breast cancer cases (1021 invasive, 208 in situ and 30 unknown or undetermined histology) diagnosed after blood draw up to June 1, 2000, and 1742 matched controls, drawn from 32,826 women who gave a blood sample in 1989-90. Controls were randomly selected participants who were free of diagnosed cancer (except non-melanoma skin cancer), and matched to cases based on age, menopausal status, recent post-menopausal hormone use, and time, day, and month of blood collection. Greater than 95% of the samples were successfully genotyped on the first attempt, with samples that failed genotyping being removed from further analyses.

We used SAS v8.2 (SAS Institute, Cary, NC) for most statistical analyses. Odds ratios (OR) and 95% confidence intervals (CI) were calculated using both conditional and unconditional logistic regression, controlling for matching factors, age at menopause, age at menarche, age at first birth and parity, history of benign breast disease, and family history of breast cancer using PROC PHREG (conditional regression) and PROC LOGISTIC (unconditional regression). We tested for departures from Hardy-Weinberg equilibrium using PROC ALLELE. Interactions were tested by likelihood ratio tests comparing the model with main effects for each variable of interest to the model with the two variables cross-tabulated. Meta-analyses were conducted using the rmeta package in R 1.7.1. All p-values reported are two sided.

Results and discussion

Our results were similar to those observed by Egan et al. [8]. Meta-analyses of the prior studies (shown in Table 1) and our study show a significant increase in breast cancer risk for women homozygous for the I allele of F31I (Summary OR 1.28, 95% CI 1.08-1.53). The test of heterogeneity between studies was not significant (p = 0.29).

Both polymorphisms studied were in Hardy-Weinberg equilibrium in controls (p>0.41). There were no significant differences in risk estimates generated by conditional and unconditional models, and Table 2 reports those from the multivariate unconditional regressions. We observed a borderline significant association between homozygosity of the 31I allele and overall breast cancer risk (OR 1.40, 95% CI 0.97-2.02, Table 2). No statistically significant association was observed with the V57I polymorphism. When restricting our analyses to post-menopausal women, those homozygous for the 31I allele were at 57% increased risk of invasive breast cancer (OR 1.57, 95% CI 1.05-2.33, Table 2). Our results in combining the two genotypes are similar to those seen by examining the V31I allele alone. In analyses restricted to post-menopausal women, we found that women homozygous for both the 31I and 57V were at 63% increased risk of invasive breast cancer (OR 1.63, 95% CI 1.08-2.45, Table 2). We observed, a higher risk of invasive breast cancer in lean post-menopausal women homozygous for 31I (BMI < median of 25 kg/m2, OR 1.90; 95% CI 1.11-3.25), as opposed to heavier post-menopausal women (BMI ≥ median of 25 Kg/m2, OR 1.06; 95% CI 0.64-1.76).

Egan et al. had observed that heavier women were at increased risk with the 31I homozygous genotype when contrasted with leaner women. However, since neither the p-value for the interaction in our study (p=0.27) or theirs (p=0.99) was statistically significant between BMI and AURKA genotypes on breast cancer risk, these differences are most likely due to chance. In pre-menopausal women, Egan et al. observed a non-significant inverse association among women homozygous for the I allele at V57I (OR 0.38, 95% CI 0.14-1.07) which is very similar to our study (OR 0.35 95% CI 0.09-1.44). While neither study has large numbers of pre-menopausal women, the similarity between the two findings warrants examination in studies with larger numbers of pre-menopausal women. We observed no difference in the distribution of BMI or duration of menstruation (menarche to first birth or menopause in nulliparous women) between genotypes among cases or controls (data not shown). No differences in risk were observed upon stratification by history of benign breast disease, first degree family history of breast cancer, or estrogen/progesterone receptor status of tumors in the cases.

In conclusion, these polymorphisms in the AURKA gene appear to be associated with an increase in breast cancer risk, which is similar in both Caucasian and Asian populations. This increase in risk is independent of established hormone related risk factors for breast cancer.