The ataxia-telangiectasia mutated (ATM) gene plays a critical role in cell-cycle arrest, apoptosis, and DNA repair. However, to date, no study has directly investigated the association between ATM gene expression and breast cancer survival.
ATM gene expression levels were evaluated in tumor and adjacent normal tissue from patients diagnosed with primary breast cancer or BBD using quantitative real-time reverse transcription polymerase chain reaction (RT-PCR) assays. Cox regression models were used to evaluate the association of ATM gene expression and survival in a cohort of 471 breast cancer patients.
In breast cancer cases, ATM expression in cancer tissues was decreased by approximately 50% compared with adjacent normal tissues from the same patients. In BBD cases, the expression level of the ATM gene was similar in benign tumor tissue and adjacent normal tissues. No apparent difference was found in ATM gene expression levels in adjacent normal tissues obtained from cancer patients or BBD controls. Compared with patients with the lowest tertile of the ATM mRNA, patients in the upper 2 tertiles had more favorable disease-free survival (hazard ratio [HR] = 0.46, 95% confidence interval [CI]: 0.30–0.73 and HR = 0.52, 95% CI: 0.33–0.81, respectively) and overall survival (HR = 0.56, 95% CI: 0.35–0.92 and HR = 0.70, 95% CI: 0.43–1.13, respectively).
The ATM (ataxia-telangiectasia mutated) gene is located on chromosome 11q22∼23, encoding a protein kinase of 350 kDa, which is involved in the maintenance of genome integrity.1–6 Many in vitro and in vivo studies suggest that the ATM gene may play an important role in breast carcinogenesis. Most previous studies in humans have focused on the evaluation of the association between ATM gene variants and breast cancer.1, 3, 5, 7–11 Only a few small studies have directly evaluated ATM gene expression in breast tissues.6, 12 Low ATM expression in breast cancer tissue has been shown to be related to a high rate of DNA mutation in cancer cells and, thus, a progressive cancer phenotype.8, 13–15 In addition, low ATM expression in breast carcinoma has also been correlated with increased neoangiogenesis.16 However, to date no study has directly investigated the association between ATM gene expression and breast cancer survival. In this study we first evaluated the expression patterns of the ATM gene in breast tissues from patients diagnosed with breast cancer or benign breast disease (BBD). To investigate the clinical utility of ATM gene expression patterns, we examined the associations between levels of ATM mRNA expression and breast cancer survival.
MATERIALS AND METHODS
Study Subjects and Tissue Samples
Included in this study was a subset of patients who were recruited as a part of the Shanghai Breast Cancer Study.17, 18 These patients were diagnosed with breast cancer or BBD between 1996 and 1998 and were identified through a network of major hospitals that treat over 80% of breast cancer patients in urban Shanghai. A total of 1602 women who were diagnosed with a primary breast cancer were identified. Among them, 1459 (91.1%) participated in the Shanghai Breast Cancer Study. During surgery, tumor tissue samples were obtained from the tumor and the adjacent normal tissue samples were obtained from the distal edge of the resection. These samples were snap-frozen in liquid nitrogen as soon as possible, typically within 10 minutes. Samples were stored at −70°C until the relevant assays were performed.
All patients were interviewed at the time of recruitment. A structured questionnaire was used in this study to collect information on demographic factors, menstrual and reproductive history, hormone use, previous disease history, family history of cancer, physical activity, tobacco and alcohol use, and usual dietary habits. All participants were measured for current weight, circumferences of the waist and hips, and sitting and standing heights. Medical charts were reviewed using a standard protocol to obtain information on cancer treatment, clinical stages, and cancer characteristics, such as estrogen and progesterone receptor status. Two pathologists reviewed pathology slides to confirm the diagnosis for breast cancer or BBD. BBDs were classified based on the published criteria developed by Page et al.19
All breast cancer patients were followed through July 2005 either in person or via phone contact and through record linkage to the death certificates kept by the Shanghai Vital Statistics Unit. In all, 89.2% of patients successfully completed the follow-up interview either in person or by telephone. For those who could not be contacted in person or by phone, linkage to the death certificate data was conducted to obtain information on the date and cause of death. Subjects who had no match in the death registry were assumed to be alive on December 30, 2004, 6 months before the linkages in order to allow for a possible delay of entry of the death certificates into the registry.
Total RNA was extracted from tissue specimens by homogenization in TRIzol solution (Invitrogen, Carlsbad, Calif), phase separation, precipitation, and washing following the manufacturer's instructions. The quality and quantity of RNA was measured by spectrophotometric analysis. TaqMan Reverse Transcription Reagents (N8080234) or High Capacity cDNA Archive Kits (4322171) were obtained from Applied Biosystems (Foster City, Calif). RNA was reverse-transcribed (RT) in a final volume of 15 μL containing 0.15 μg RNA and 1× RT polymerase chain reaction (RT-PCR) buffer, 5.5mM MgCl2, 500 μM each dNTP, 2.5 μM Random Hexamers, 0.4 U/μL RNase inhibitor, and 3.125 U/μL MultiScribe reverse transcriptase (Applied Biosystems). The mixture was incubated at 25°C for 10 minutes, 37°C for 120 minutes, and 95°C for 5 minutes.
The primers and probes for the ATM gene (Hs00175892) and the β-actin gene (Hs99999903) were obtained from Applied Biosystems. Quantitative real-time PCR was performed using a 384-well optic tray on an ABI PRISM 7900HT Sequence Detection System (Applied Biosystems). A total reaction volume of 5 μL containing 2.2 μL cDNA template at different dilutions, 1× TaqMan Universal PCR Master Mix (without UNG), and 1× Gene Expression Assay Mix, including the primers and marked probes from Applied Biosystems Assay-on-Demand services. The thermal cycling conditions were as follows: 95°C for 10 minutes to activate the AmpliTaq Gold enzyme, followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. Every sample was tested in triplicate. Two control samples were used in each plate to monitor interplate variation, which was found to be smaller than 5% in our study. The threshold cycle (Ct) was determined at 0.1 based on the amplification linear area of the target genes and the β-actin gene (an internal control). The normalized quantity of the target gene was calculated as 2−ΔCt, where ΔCt was obtained directly by subtracting Ct for the target gene from Ct for the β-actin gene. The final result was expressed as 2−ΔCt*1000.20
The initial samples included in the current analysis contained tumor tissues from 128 primary breast cancer patients and 39 BBD patients, along with the paired adjacent normal tissues from 43 cancer patients and 39 BBD patients. For survival analysis, tumor tissue samples for an additional 343 breast cancer patients were included (batch 2) to enhance the statistical power of the study.
The data were skewed to the high value, and thus medians (25th percentile, 75th percentile) were used and compared across study groups using the Mann-Whitney test. Paired data were evaluated using the Wilcoxon signed rank test. The primary outcomes used for this study were overall survival (OS) and disease-free survival (DFS). The endpoint for the analysis of OS was any death and for the analysis of DFS was cancer recurrence/metastasis or death related to breast cancer. Survival time was calculated as the time from cancer diagnosis to the endpoints of the study, censoring at the date of last contact or noncancer death (for DFS only). For subjects who had died of breast cancer without information on the date of recurrence or metastasis the total survival time was substituted for the DFS time.18 The 5-year survival rate was estimated using the Kaplan-Meier method. The log-rank test was applied to test the differences in survival across comparison groups. Cox proportional hazard models were applied for calculating hazard ratios (HRs) using the lowest tertile as the reference after adjusting for age, menopausal status, TNM stages, and estrogen receptor / progesterone receptor (ER/PR) status. All P-values are 2-sided. SAS software was used for statistical analysis (v. 9.1; SAS Institute, Cary, NC).
Table 1 presents the characteristics of study participants, which included 471 breast cancer patients and 39 BBD patients. The mean age was 47.5 years for breast cancer patients and 43.7 years for BBD patients. In BBD patients, 53.8% were nonproliferative lesions, whereas among proliferative benign breast disease fibroadenoma was the most common, accounting for 33% of the total benign diseases diagnosed, and only 1 subject had atypical hyperplasia. There were 23.8% breast cancer patients diagnosed at an early stage (0 or I), 40.3% at stage IIa, 26.3% at stage IIb, and 9.6% at stage III or IV. Most of the breast cancer patients (66.6%) had a stage II cancer, and virtually all patients received surgery (100%) or chemotherapy (96.8%). Radiotherapy was given to 45.2% of patients, whereas 76.9% received tamoxifen therapy. The reason for the high chemotherapy percentage in our study is due to the recommendation in China that neoadjuvant chemotherapy is a major adjunctive treatment for breast cancer.21, 22
Table 1. Characteristics of Study Participants
No. of subjects
SD indicates standard deviation; ER, estrogen receptor; PR, progesterone receptor.
Excluding 95 patients with missing data on ER status and 98 patients with missing data on PR status.
Table 2 compares the expression levels of the ATM gene in tumor tissues and adjacent nonneoplastic tissues in cancer or BBD patients. ATM expression levels were significantly lower in cancer tissue than in paired adjacent normal tissues (P < .001), whereas ATM expression in benign tumor tissues were similar to their corresponding adjacent tissues (P = .433). Further, there was no significant difference between tumor tissues and the adjacent normal tissues in either nonproliferative or proliferative cases among BBD patients (P = .061 and P = .463, respectively). When cancer patients and BBD patients were compared, cancer tissues showed a remarkably reduced ATM expression compared with benign tumor tissues (P < .001). The expression levels of this gene in the adjacent normal tissues from cancer patients and BBD patients, however, were comparable (Table 2). We analyzed the ATM mRNA expression in breast cancer tissue with respect to breast cancer stage. However, no significant association was observed between ATM expression level and TNM stage (data not shown). In addition, ATM expression levels were not related to radiotherapy in either adjacent normal tissues or in tumor tissues of breast cancer patients (data not shown). This is not surprising because tissue samples were collected before any cancer therapies.
Table 2. Comparison of ATM mRNA Expression Levels Between Tumor Tissues and Tumor Adjacent Normal Tissues in Patients Diagnosed With Breast Cancer or Benign Breast Disease (BBD)
The median follow-up time for breast cancer patients was approximately 7 years. Table 3 presents HRs and 95% CIs after adjusting for potential confounding factors, including TNM stage, ER/PR status, radiotherapy, menopausal status, and age. Compared with patients with the lowest tertile of ATM expression in cancer tissue, patients within the upper 2 tertiles of ATM expression had a more favorable DFS and, to a lesser extent, OS. Results from stratified analyses by TNM stage are also presented in Table 3, showing that the inverse association of breast cancer survival with ATM expression exists in both earlier and later stage cancer patients. The curves for OS and DFS are shown in Figure 1. As expected, patients in the upper 2 tertiles of ATM expression in breast cancer tissue had better OS and DFS compared with patients with the lowest ATM expression level throughout the study period.
Table 3. Association of ATM Gene Expression in Tumor Tissues With Overall and Disease-Free Survival Among Breast Cancer Patients
Adjusted for age, menopausal status, TNM stage, radiotherapy, and ER/PR status.
Patients with stage 0, I, or IIa cancer
Patients with stage IIb, III, or IV cancer
We conducted additional analyses treating ATM expression as a continuous variable. Because the ATM expression data were skewed to the right, log-transformed ATM expression values were used. An inverse association between ATM expression and breast cancer risk was observed. The adjusted HRs were 0.80 (95% CI: 0.42–1.51) for OS and 0.66 (95% CI: 0.36–1.24) for DFS. These results were consistent with those based on categorical data analysis (data not shown in table).
In this population-based study we found that the ATM mRNA level was significantly lower in breast cancer tissue than the adjacent normal tissue from the same patients or breast tissues from patients with BBD. Furthermore, high ATM expression levels in breast cancer tissue were associated with increased DFS and OS. This finding is new and consistent with the hypothesis that the ATM gene may play an important role in breast cancer development and progression.
Only 2 small studies have examined ATM mRNA expression levels in breast cancer tissues using semiquantitative competitive RT-PCR methods.6, 12 In the first study conducted in German women, ATM mRNA expression was evaluated in 39 breast cancer patients, 14 benign breast lesion patients, and 4 unmatched control individuals.6ATM transcript levels were the lowest in breast cancer tissue, followed by benign tumor tissue and normal breast tissue specimens. The reduced ATM expression in cancer tissues, however, was not observed in the second study, a hospital-based study of 89 Caucasian women with breast cancer.12 Both studies used competitive RT-PCR. Although the sensitivity of competitive RT-PCR was high, it may have increased the likelihood of contamination and detection of false-positives.23, 24 We used real-time quantitative RT-PCR method, a more accurate and reliable method in assessing gene expression levels.24 Our findings were consistent with the first study described above as well as 2 other studies, in which ATM protein expression levels were found to be lower in breast cancer tissue than in matched control samples.2, 25 These findings indicate that ATM mRNA expression is substantially down-regulated in breast cancer tissues, but not in adjacent normal or benign breast tumor tissue. The cellular mechanisms that lead to the deregulation of ATM expression in breast cancer tissue remain to be determined. Losses on chromosome 11q, where the ATM gene is located, are frequently observed in breast cancer cells, which may explain the reduced mRNA level of the ATM gene in some cancer patients.4, 6 Alternatively, a reduced level of ATM protein expression may be due to methylation of the bidirectional ATM promoter region.26
More favorable survival was found in our study among breast cancer patients whose cancer tissues expressed a high level of ATM mRNA. However, we did not find the interaction between ATM expression and the adjuvant radiotherapy in breast cancer survival. Our findings are novel and biologically plausible. Previous in vitro studies have found that cells defective in the ATM gene appear to be more resistant to irradiation-induced apoptosis.27–29 In addition, ATM protein activates and stabilizes p53 tumor suppressor protein, which results in the induction of G1, S, and G2 cell cycle arrest, and DNA repair and activation of cell death pathways.3, 4, 30, 31 Because radiation and chemotherapy results in extensive DNA damage, a signal to ATM activation, cancer cells with no ATM or reduced ATM expression may respond poorly to radiation and/or chemotherapy. This, in turn, influences the prognosis of breast cancer.4, 28, 31, 32 Our finding is also supported by a recent study in which reduced ATM expression in breast cancer tissue was found to be correlated with increased microvascular parameters, a prognostic factor for breast cancer.16 However, our findings are not consistent with data from some in vitro studies.33, 34 One study found high ATM gene expression was observed in methotrexate-resistant mouse fibroblast cell lines.33 In another study, silencing of ATM protein enhanced radiation-mediated cell killing of human cervical cancer cells.34 These findings suggest that patients who have a higher ATM expression level may be less sensitive to radiotherapy or chemotherapy, and more likely to have a poor survival rate compared with those who possessed a lower ATM expression. The possible explanation for this inconsistency is that ATM may play an important role in neoangiogenesis and activation of cell death pathways instead of affecting the responses to adjuvant therapy,16, 35, 36 or ATM may have a different role in different cancer types. Therefore, further studies are warranted to replicate these findings.
The large sample size of our study provides stableestimates of gene expression patterns and their correlations with cancer survival. In addition, we included both tumor tissue and adjacent normal tissue from patients diagnosed with breast cancer or BBD, which enabled a systematic evaluation of expression patterns of the ATM gene based on the type of tissues and diseases. However, it should be noted that although the median length of the follow-up period (84 months) is considered adequate, reevaluation of the data after extension of the follow-up period might provide additional information. Although we did not adjust for each of the specific prognostic factors, such as invasiveness of cancer, metastatic status, and node involvement, TNM stages were adjusted in our study and used in stratified analyses, showing that these prognostic factors are unlikely to affect the results from this study.
In conclusion, we found that ATM expression was down-regulated in breast cancer tissues and that a high level of ATM expression in breast cancer tissue was linked to improved survival in breast cancer patients. Our findings, if confirmed, may have significant clinical implications in identifying high-risk breast cancer patients for cancer treatment and follow-up.
We thank Drs. Wanqing Wen and Qiong Li for their contributions in data analysis, Brandy Venuti for technical assistance in the preparation of the article, and all of the study participants and research staff of the Shanghai Breast Cancer Study for their support.