CEBP Grants Frontiers in Basic Cancer Research
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
Cancer Epidemiology Biomarkers & Prevention 16, 1812, September 1, 2007. doi: 10.1158/1055-9965.EPI-06-1034
© 2007 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Euhus, D. M.
Right arrow Articles by Lewis, C. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Euhus, D. M.
Right arrow Articles by Lewis, C. M.
Related Collections
Right arrow Risk Assessment
Right arrow Risk Assessment: Biomarkers

Atypia and DNA Methylation in Nipple Duct Lavage in Relation to Predicted Breast Cancer Risk

David M. Euhus1,4, Dawei Bu1, Raheela Ashfaq2, Xian-Jin Xie3,4, Aihua Bian4, A. Marilyn Leitch1 and Cheryl M. Lewis1

Departments of 1 Surgery, 2 Pathology, and 3 Clinical Sciences, and 4 Simmons Comprehensive Cancer Center, University of Texas Southwestern Medical Center, Dallas, Texas

Requests for reprints: David M. Euhus, E6.222, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390-9155. Phone: 214-648-6467; Fax: 214-648-7965. E-mail: david.euhus{at}utsouthwestern.edu


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Background: Tumor suppressor gene (TSG) methylation is identified more frequently in random periareolar fine needle aspiration samples from women at high risk for breast cancer than women at lower risk. It is not known whether TSG methylation or atypia in nipple duct lavage (NDL) samples is related to predicted breast cancer risk.

Methods: 514 NDL samples obtained from 150 women selected to represent a wide range of breast cancer risk were evaluated cytologically and by quantitative multiplex methylation-specific PCR for methylation of cyclin D2, APC, HIN1, RASSF1A, and RAR-ß2.

Results: Based on methylation patterns and cytology, NDL retrieved cancer cells from only 9% of breasts ipsilateral to a breast cancer. Methylation of ≥2 genes correlated with marked atypia by univariate analysis, but not multivariate analysis, that adjusted for sample cellularity and risk group classification. Both marked atypia and TSG methylation independently predicted abundant cellularity in multivariate analyses. Discrimination between Gail lower-risk ducts and Gail high-risk ducts was similar for marked atypia [odds ratio (OR), 3.48; P = 0.06] and measures of TSG methylation (OR, 3.51; P = 0.03). However, marked atypia provided better discrimination between Gail lower-risk ducts and ducts contralateral to a breast cancer (OR, 6.91; P = 0.003, compared with methylation OR, 4.21; P = 0.02).

Conclusions: TSG methylation in NDL samples does not predict marked atypia after correcting for sample cellularity and risk group classification. Rather, both methylation and marked atypia are independently associated with highly cellular samples, Gail model risk classifications, and a personal history of breast cancer. This suggests the existence of related, but independent, pathogenic pathways in breast epithelium. (Cancer Epidemiol Biomarkers Prev 2007;16(9):1812–21)


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
The Gail model (1) has previously been validated as a tool for breast cancer risk assessment. Cohort studies have consistently shown that the model is well calibrated (i.e., for a given population, the ratio of observed to expected breast cancers is near 1.0; refs. 2, 3). However, the model does not discriminate well between women who will develop breast cancer and women who will not. Assessment of biomarkers in breast epithelial cells obtained by minimally invasive approaches has been proposed as an approach for individualized risk stratification with the potential to improve on the discrimination of mathematical models.

Promoter region methylation is a well-established mechanism for silencing tumor suppressor genes (TSG). Studies in cell culture suggest that DNA methylation of some genes is a very early event in transformation that precedes spontaneous immortalization (4). Dense TSG methylation is readily detectable in nearly all breast cancers (5) but not in nonproliferative breast tissue (6). TSG methylation can be observed in benign breast tissue using highly sensitive methods (7). Its occurrence in benign proliferative breast disease (6, 8-10), benign breast tissue adjacent to breast cancer (9, 11), and lobular carcinoma in situ (12) suggests that it may be an early biomarker of carcinogenesis.

Nipple duct lavage (NDL) is a minimally invasive approach for obtaining breast epithelial cells. Cytologic atypia identified in nipple aspirate fluid (NAF; ref. 13) or in random periareolar fine needle aspiration (RP-FNA) samples (14) is associated with increased breast cancer risk. Cytologic atypia diagnosed by NDL is currently being evaluated in a prospective multi-institutional clinical trial, but little is known about the biology of cells obtained by NDL, particularly as it relates to breast cancer risk.

Quantitative multiplex methylation-specific PCR (QM-MSP) measures the proportion of total gene copies in a sample with promoter region methylation (methylation fraction). This study was conducted to measure the prevalence of TSG methylation in duct lavage samples and to assess the relationship between TSG methylation and cellularity, cytologic classification, and predicted breast cancer risk.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Study Subjects
This study was done after approval by the University of Texas Southwestern Medical Center Institutional Review Board and in accordance with an assurance filed with and approved by the Department of Health and Human Services. Informed consent was documented in writing for all subjects. Patients with incident breast cancer and unaffected women over the age of 18 years presenting for breast cancer risk assessment were offered ductal lavage regardless of the calculated risk level. Exclusion criteria included presence of an undefined palpable or mammographic breast lesion suspicious for malignancy; bilateral prophylactic mastectomy; any prior breast irradiation; any systemic chemotherapy in the past; performance status that restricted normal activity for a significant portion of the day; current use of androgens, luteinizing hormone–releasing hormone analogues, prolactin inhibitors, antiandrogens, or glucocorticoids (women were eligible if these drugs were discontinued 3 months before lavage); ever use of tamoxifen, raloxifene, or other selective estrogen receptor modifiers; or pregnancy or lactation within 6 months.

Breast Cancer Risk Assessment
Comprehensive breast cancer risk factor information was collected and breast cancer risk calculated using custom software (BreastC.A.R.E.), which guides a structured, on-screen interview and uses the Gail model (1), Claus model (15), and BRCAPRO (16) to estimate age-specific and cumulative breast cancer probabilities. Portions of this software are available in the cancer genetics risk counseling program CancerGene.5 The Gail model, which is based on age, race, age at menarche, age at first live birth, number of breast biopsies, and family history of breast cancer in first-degree relatives, is well calibrated for estimating breast cancer risk (2, 3). Absolute risk calculated by the Gail model is highly dependent on race and age, so a risk index was calculated by dividing the 5-year Gail risk by age- and race-matched general population risk (17). We have previously reported that Gail model calculations, but not Claus or BRCAPRO model calculations, are associated with TSG methylation in benign breast samples (18). For this reason, women with a 5-year Gail model risk that was greater than or equal to twice the age- and race-matched general population risk were classified as high risk, and all other unaffected women as lower risk regardless of the Claus and BRCAPRO predictions.

Nipple Duct Lavage
Local anesthetic cream (EMLA, Astra-Zeneca) was applied to the nipples 1 to 2 h before the procedure. At the start of the procedure, the patient did a breast self-massage, after which the nipple was dekeratinized with a mild abrasive gel (Nuprep, D.O. Weaver and Co.). The operator then continued the breast massage in an effort to express NAF. If no NAF was elicited manually, a nipple aspirator (FirstCyte, Cytyc Health Corporation) was used. Fluid producing ducts were initially cannulated with a tapered dilator coated with 2% lidocaine jelly, after which a ductal lavage microcatheter (FirstCyte Microcatheter, Cytyc Health Corporation) was inserted. Saline (10 mL) was infused into the duct in 0.5-mL increments and the effluent fluid aspirated. An attempt was made to lavage all fluid-producing and at least one non–fluid-producing duct from each breast.

Duct lavage samples were dispersed into 30 mL of CytoLyt solution as they were obtained. The samples were immediately split, with half of the volume submitted for cytology and half submitted for methylation assays. Samples for methylation were centrifuged at 1,800 x g for 30 min and, after the supernatant had been decanted, the cell pellet was stored frozen at –80°C until the time of DNA extraction.

Cytologic Assessment
Cytology slides were prepared using the thin-prep method and stained using the Papanicolaou technique. All slides were evaluated by the same breast cytopathologist (R.A.). The epithelial cell yield for each sample was estimated as insufficient cellular material for diagnosis, scant cellularity but sufficient for diagnosis (~10 cells), 11-99, 100-999, or ≥1,000 cells. The cytopathologist subjectively classified each sample as nonproliferative, proliferative, mild atypia, or marked atypia, and also assigned a numerical score according to the method of Masood (19). Briefly, each of six cytologic features was assigned a score of 1 to 4. These cytologic features include cell arrangement, pleomorphism, paucity of myoepithelial cells, anisonucleosis, nucleoli, and chromatin clumping. Nonproliferative samples generally score in the 6-10 range; proliferative samples, 11-14; and atypia, ≥15. A description of the cytologic findings from the current study has previously been published (20).

TSG Methylation
After comparing several methods for DNA extraction (21), a sequential protein -> DNA precipitation method was selected (Puregene, Gentra). To estimate the quantity of amplifiable DNA in each sample, glyceraldehyde-3-phosphate dehydrogenase was amplified from 1 µL of the DNA extraction and the PCR products were resolved by electrophoresis on an agarose gel. For samples producing strong bands relative to a 100 ng/µL standard prepared from HCC1954 cells, 5 µL of DNA were sodium bisulfite treated; for samples producing bands similar to the standard, 10 µL were treated; and for samples producing bands that were weaker than the standard, the entire 20 µL of DNA were sodium bisulfite treated. Sodium bisulfite treatment was done using the method of Clark (22). Yeast tRNA was used as a carrier for all sodium bisulfite treatments so that product recovery would not pose a limitation for paucicellular samples.

We selected five genes that are known to be highly methylated in breast cancer as compared with benign breast tissue or that are differentially methylated in benign breast epithelium from high-risk women as compared with lower-risk women (18). Each gene is known to be regulated by promoter region methylation. Primers were specifically chosen to amplify a region of the promoter known to silence gene expression when methylated. Publications supporting our marker and primer selection include cyclin D2 (23), APC (promoter A1; refs. 24, 25), HIN1 (26), RASSF1A (27), and RAR-ß2 (28).

TSG methylation was quantified using the QM-MSP method of Fackler (7). Briefly, multiplex PCR was initially done for all five markers using the Qiagen multiplex PCR kit. An MJ Research DNA Engine Dyad PTC220 thermocycler was used with the following PCR program: 95°C 15 min, and then 40 cycles of 94°C 30 s, 58°C 90 s, 72°C 90 s, followed by 72°C 10 min, with a final hold at 4°C. First-round multiplex primers were designed to bracket the region of interest external to the CpGs subject to methylation (Table 1 ). The second-round uniplex PCR was nested within the region amplified by the first round primers. The uniplex primers and TaqMan probes were designed to bind specifically to methylated or unmethylated CpGs (Table 1). Real-time PCR was done using a Chromo4 machine (MJ Research) running Opticon Monitor 3.00.367 and the following program: 95°C for 10 min, followed by 40 cycles of 95°C 30 s, 60°C 45 s. Copy number standards were prepared for each gene (unmethylated and methylated) using lymphocytes for unmethylated DNA and the following cell lines for methylated DNA: HCC1954 (APC, HIN1, and RASSF1A), HCC1569 (cyclin D2), and MCF7 (RAR-ß2). The mean Ct value for duplicate test samples was converted to DNA copy number based on the 40,000-copy standard run on that plate. Methylation fraction (i.e., proportion of gene copies methylated) for a given sample was calculated as methylated copies / (methylated copies + unmethylated copies).


View this table:
[in this window]
[in a new window]

  		Table 1. Primers and probes for quantitative multiplex methylation-specific real-time PCR

 
Quality assurance standards included (a) standard curve slope within –3.11 and –3.58 without removing any points (this correlates with reaction efficiencies of 110% and 90% respectively); (b) standard curve R2 > 0.985; (c) difference in duplicate Ct values <1.6; and (d) average Ct for the test sample within the average Ct values of the standards. Assays not meeting the quality assurance standards were repeated.

Linearity and accuracy of the assay were assessed by preparing mixtures of DNA that were 100% and 0% methylated for each gene. The assay showed excellent linearity across the entire dynamic range of 0% to 100% methylation as shown by the following R2 values: cyclin D2, 0.993; APC, 0.943; HIN1, 0.947; RASSF1A, 0.915; RAR-ß2, 0.978. The sensitivity of the assay was calculated at one methylated gene copy among 105 unmethylated copies. Intra-assay and interassay variabilities were measured by repeating the analysis four times on the same day or on different days using samples with very low levels of methylation (<1%) as well as samples with high levels of methylation (~80%). The coefficient of variation is calculated as the SD of the four replicates divided by the mean. Intra-assay coefficients of variation ranged from 0.148 to 0.436 for samples with methylation of <1% of gene copies and from 0.003 to 0.305 for samples with methylation of ~80% of gene copies. Interassay coefficients of variation ranged from 0.159 to 0.555. In general, the reproducibility of the assay is lower than that commonly reported for assays like ELISA, which usually show coefficients of variation <0.15. Nevertheless, for RASSF1A, which had the lowest interassay reproducibility, a methylation fraction of 0.06 would fall 2 SDs above a methylation fraction of 0.03, permitting reliable discrimination between these values.

Approach to Data Analysis
Methylation fraction is defined as the proportion of gene copies in a sample that are methylated. A sample was classified as methylation positive for a given gene if the methylation fraction was greater than the 90th percentile for ducts from unaffected women at lower risk for breast cancer according to the Gail model. These values were 0.009 for cyclin D2, 0.028 for APC, 0.020 for HIN1, 0.015 for RASSF1A, and 0.008 for RAR-ß2. Median methylation fraction provides a measure of the intensity of methylation for a group of samples and is the median methylation value for all samples classified as methylation positive according to the preceding criteria.

The prevalence of TSG methylation was compared between groups using {chi}2 if there were ≥5 observations in each group, and Fisher's exact test if there were <5 observations. Median methylation fractions were compared using the Kruskal-Wallis statistic. All statistical tests were two sided.

The prevalence of TSG methylation and of atypical cytology was assessed in relation to age, race, menopausal status, NAF production, cellularity of the sample, and risk level of the breast providing the sample using logistic regression in a series of univariate analyses and then by multivariate analysis that included all covariates generating a univariate P < 0.15. Each duct was considered as an independent unit of analysis and all ducts from women affected or unaffected with breast cancer were included in this analysis. A similar analysis that only included ducts from women unaffected with breast cancer was done to identify Gail risk factors predicting marked atypia or TSG methylation. The primary methylation end point evaluated in these analyses was methylation fraction for ≥2 genes greater than the 90th percentile of samples from women whose 5-year Gail risk was less than twice the age- and race-matched general population risk (Gail lower risk).


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Between 10/16/2001 and 6/21/2005, 150 women were enrolled in the study. Nipple ducts were successfully cannulated in 149 (99.3%) women. The NDL procedure was done in a total of 514 ducts from 290 breasts of 149 women. Samples adequate for cytologic diagnosis were obtained for 134 of the 150 (89.3%) women. The characteristics of the study sample are summarized in Table 2 . On the average, 1.4 dry ducts were lavaged per patient, and 2.1 NAF-producing ducts for a total of 3.5 ducts per patient.


View this table:
[in this window]
[in a new window]

  		Table 2. Characteristics of the study sample

 
Methylation assays were only done for samples with sufficient epithelial cells for cytologic classification. This ensured that only samples from successful lavages were included. Ducts failing to meet all quality assurance standards for ≥4 of the five genes were excluded (15%). Misplaced, ambiguously labeled, or improperly stored samples account for exclusion of an additional 4% of samples. Figure 1 shows the distribution of patients and evaluable samples.


Figure 1
View larger version (19K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Distribution of study subjects and evaluable samples. Values for "Breasts" include only those were a duct could be cannulated. High risk and lower risk for the unaffected women are defined by the Gail model. ICMD, insufficient cellular material for diagnosis.

 
Methylation Profiles in Breast Cancer and Ducts Ipsilateral to Breast Cancer
NDL ipsilateral to a breast cancer rarely retrieved cancer cells. QM-MSP data were available for 34 tumor-tissue FNAs from patients participating in the duct lavage study. Duct lavage convincingly retrieved cancer cells in only 3 of these 34 (9%) cases. The lavage samples from patients 1, 4, and 19 in Fig. 2 contained atypical cells with methylation profiles that were concordant with that of the cancer for ≥4 genes. The frequency of TSG methylation was significantly greater in the tumors than the corresponding NDL samples for every gene (Table 3 ). Methylation of ≥1, ≥2, or ≥3 genes was identified in 74%, 56%, and 44% of the tumor samples as compared with 28%, 13%, and 7% of the NDLs (P = 0.002, P = 0.001, and P = 0.002, respectively). Among the methylation-positive cases, the median methylation fraction was significantly greater for the tumor-tissue FNAs than the corresponding NDL samples for APC, HIN1, and RASSF1A. There was no correlation between the individual methylation fractions of the tumors as compared with the corresponding NDLs for any gene. These data confirm that the QM-MSP assay is reliably measuring TSG methylation, that the selected marker panel is relevant to breast cancer, and that lavage ipsilateral to a breast cancer rarely retrieves cancer cells.


Figure 2
View larger version (53K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Methylation profiles of FNA samples from primary tumors as compared with NDL samples from ducts ipsilateral to these tumors. Cases are sorted by the degree of methylation of the primary tumor from most methylated to least methylated. Hist is histology of the primary tumor: ductal carcinoma in situ (DCIS; Figure 2), infiltrating ductal carcinoma (Figure 2), infiltrating lobular carcinoma (Figure 2), other (medullary or metaplastic; Figure 2); any ductal carcinoma in situ: yes (Figure 2), no (Figure 2); methylation fraction: 0% (Figure 2), >10% (Figure 2), no result (Figure 2); Cyto is NDL cytology: Masood score ≥15 (Figure 2), Masood score 11-14 (Figure 2), Masood score ≤10 (Figure 2). Ducts classified as cytologically atypical with methylation results that were concordant with the primary tumor for ≥4 genes are marked with asterisk.

 

View this table:
[in this window]
[in a new window]

  		Table 3. TSG methylation in 34 primary tumors and corresponding ipsilateral duct lavage samples

 
Intraductal and Interductal Correlations
For breasts unaffected with breast cancer, methylation of cyclin D2 or RASSF1A predicted methylation of other genes in the same duct. For instance, 38% of 21 ducts with cyclin D2 methylation also showed methylation of APC as compared with only 5% of 206 ducts without cyclin D2 methylation (P < 0.0001). Cyclin D2 methylation also predicted methylation of HIN1 (P = 0.017) and RASSF1A (P = 0.007) but not RAR-ß2 (P = 0.450). RASSF1A methylation was less predictive but was associated with methylation of APC (P = 0.025) and HIN1 (P = 0.052) but not RAR-ß2 (P = 0.942).

There were 47 breasts with methylation data for two ducts. Methylation of a specific gene in one duct did not predict methylation of the same gene in another duct. However, if methylation was identified for ≥2 genes in one duct, 55% of the breasts showed methylation for at least one gene in a second duct as compared with only 5% if no duct showed methylation of ≥2 genes (P = 0.002).

There were 80 breasts with cytology data for two ducts. Marked atypia was diagnosed in 8% of these 160 ducts. If one duct showed marked atypia, the second duct also showed marked atypia in 23% (P = 0.209). If one duct showed marked atypia, a second duct showed either mild or marked atypia in 54% as compared with 19% if no duct showed marked atypia (P = 0.011). Mild atypia in one duct did not predict atypia in the second duct.

Factors Predicting Atypical Cytology
Mild atypia was highly related to sample cellularity by both univariate and multivariate analyses [multivariate odds ratio (OR) ≥1,000 cells, 3.58; P = 0.0005] but was unrelated to the risk classification of the breast providing the sample. On univariate analysis, marked atypia was related to race, menopausal status, cellularity, and risk classification of the breast providing the sample (Table 4 ). There was a trend for increasing prevalence of marked atypia with increasing age, but this did not reach statistical significance. The association with race is artifactual as most of the non-Caucasian women were affected with breast cancer and race became insignificant after adjusting for the risk classification of the breast providing the sample in multivariate analysis. Independent predictors of marked atypia identified by multivariate analysis included perimenopausal status, increased cellularity, and increased risk classification (Table 5 ). The prevalence of marked atypia was 5% for Gail lower-risk ducts, 10% for Gail high-risk ducts, 15% for ducts contralateral to a breast cancer, and 17% for ducts ipsilateral to a breast cancer (Fig. 3 ). Results were similar if marked atypia was defined by a Masood score ≥15.


View this table:
[in this window]
[in a new window]

  		Table 4. Univariate analysis to identify factors predicting TSG methylation or marked atypia

 

View this table:
[in this window]
[in a new window]

  		Table 5. Multivariate analysis

 

Figure 3
View larger version (29K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. Prevalence of methylation and marked atypia by gene, sample source, and threshold for classifying a duct as positive. Lower 90% thresholds correspond to the 90th percentile methylation values for ducts from unaffected women at lower risk for breast cancer according to the Gail model; Fackler thresholds were calculated to provide the greatest discrimination between ducts with breast cancer and ducts without breast cancer (29). Threshold values for APC are not reported by Fackler. Prevalence of marked atypia or methylation of ≥2 genes is calculated using the 90th percentile of Gail lower-risk ducts. Figure 3, ducts from unaffected Gail lower-risk women; Figure 3, ducts from unaffected Gail high-risk women; Figure 3, ducts contralateral to a breast cancer; Figure 3, ducts ipsilateral to a breast cancer.

 
To identify Gail risk factors that are associated with marked atypia in NDL samples, a univariate analysis was done for all ducts from women unaffected with breast cancer, which included age, age at menarche, age at first live birth, and number of first-degree relatives with breast cancer as covariates. The prevalence of marked atypia increased with increasing age, decreasing age at menarche, and increasing number of first-degree relatives with breast cancer, but none of these differences was statistically significant by univariate or multivariate analysis.

Factors Predicting TSG Methylation
By univariate and multivariate analyses, methylation of ≥2 genes was predicted by perimenopausal status, increased sample cellularity, and risk classification of the breast providing the sample (Tables 4 and 5). There was a trend for increasing methylation prevalence with increasing age, but this did not reach statistical significance. Marked atypia, but not mild atypia, was correlated with methylation by univariate analysis but not by multivariate analysis that adjusted for cellularity and risk classification. The prevalence of methylation was greater for Gail high-risk ducts than Gail lower-risk ducts for every gene, except APC, but these differences were not statistically significant.

To identify Gail risk factors that are associated with TSG methylation in NDL samples, a univariate analysis was done for all ducts from women unaffected with breast cancer, which included age, age at menarche, age at first live birth, and number of first-degree relatives with breast cancer as covariates. The prevalence of methylation of ≥2 genes increased with increasing age, decreasing age at menarche, increasing number of biopsies, and increasing number of first-degree relatives with breast cancer. Only ≥2 first-degree relatives with breast cancer was significantly correlated with TSG methylation on multivariate analysis (OR, 3.75; 95% CI, 1.08-13.02; P = 0.04).

Alternate Methylation Thresholds
Fackler et al. (29) used receiver operator characteristic curve analysis to identify methylation fractions that provide maximal discrimination between ducts with histologically proven cancer and ducts without cancer. These were 0.026 for cyclin D2, 0.046 for HIN1, 0.043 for RASSF1A, and 0.015 for RAR-ß2. Fackler et al. did not assess APC. Using these thresholds, ≥2 genes were methylated in 2% of ducts from Gail lower-risk women as compared with 5% of ducts from Gail high-risk women (multivariate OR, 5.15; P = 0.101). Of note, these thresholds did not provide significant discrimination between ducts ipsilateral to a breast cancer and ducts from women unaffected with breast cancer. Figure 3 shows the prevalence of TSG methylation by gene, threshold, and risk category of the breast providing the sample.

Although TSG methylation can be detected in benign breast tissue using highly sensitive assays, this methylation occurs at considerably lower levels than those typically associated with breast cancer. Therefore, >0.0001 was evaluated as a threshold value. This is 1 order of magnitude greater than the measured sensitivity of the assay. Thirty-five percent of ducts from Gail lower-risk women showed methylation of ≥2 genes at this threshold as compared with 54% of ducts from Gail high-risk women (multivariate OR, 2.56; P = 0.008).

In addition, the methylation fraction for each gene was summed to generate a composite methylation score for each duct. The threshold for classifying a duct as methylation positive was set to the 90th percentile of the composite score for Gail lower-risk ducts. This approach did not provide any discrimination between risk groups, and there were no statistically significant correlations with sample cellularity or marked atypia.


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Based on cytology and methylation profiles, we have concluded that NDL ipsilateral to a breast cancer rarely retrieves cancer cells. This is consistent with a prior study in patients undergoing NDL immediately before mastectomy that found markedly atypical cells in fluid-yielding ducts communicating with a breast cancer in only 13% (30). Conversely, other investigators have reported that methylation of cyclin D2, RAR-ß, or Twist is identified in 85% of NDL samples from women with malignancy but rarely in women without malignancy (23). That study is not directly comparable with other studies because lavage samples were collected during ductoscopic visualization of tumors. The same group has recently used receiver operator characteristic curve analysis to determine methylation fractions providing optimal discrimination between ducts with histologically proven cancer and ducts without cancer (29). They report that assessment of TSG methylation doubles the discrimination afforded by cytology alone. This is to be expected because breast cancer cells show much higher levels of TSG methylation than benign cells for most genes (Table 3). The difficulty arises in clinical application. There are upwards of 15 independent ductal systems in each breast with six to eight readily identifiable openings on the nipple surface (31). The QM-MSP assay can improve the recognition of cancer cells when they are retrieved, but the anatomic realities of the breast make NDL unsuitable for identification of focal alterations like carcinoma. Using marked atypia or TSG methylation as a classifier did not improve the poor discrimination between breasts with cancer and unaffected breasts, even at Fackler's thresholds, or when only the most abnormal duct from each breast was included in the analysis. High levels of TSG methylation can be observed in NDL samples from breasts unaffected by breast cancer (Table 6 ) but it is not yet clear whether this is a marker of increased breast cancer risk.


View this table:
[in this window]
[in a new window]

  		Table 6. Cases with methylation fractions >0.10 for ≥2 genes

 
Although NDL is unsuitable for the identification of focal process, such as carcinoma, it may be useful for the detection of risk-associated field changes. Our intraductal and interductal correlation data, as well as a recently published tissue mapping study (32), suggest that TSG methylation may be one such field change.

Using a qualitative assay to evaluate random periareolar FNA samples, we have previously reported that the prevalence of methylation of ≥2 genes is greater in benign samples from Gail high-risk women than in samples from Gail lower-risk women (18). The current study confirms this observation in an independent sample set using NDL. Interestingly, the prevalence of TSG methylation in Gail high-risk ducts is similar to that of ducts contralateral or ipsilateral to a breast cancer, suggesting that TSG methylation in benign breast epithelium is indeed an early risk-associated field change. A history of breast cancer in ≥2 first-degree relatives predicted an increased prevalence of TSG methylation on univariate and multivariate analyses. In addition, increasing age, early age at menarche, and benign breast disease trended in the appropriate direction but did not reach statistical significance. Consistent with our prior observations in RP-FNA samples (18), neither Claus nor BRCAPRO probabilities predicted TSG methylation. These family history models are strongly weighted on early age at breast cancer diagnosis, a feature of BRCA1-associated breast cancer but not necessarily of BRCA2-associated breast cancer or familial breast cancer not associated with BRCA gene mutations. The pathogenesis of genomically unstable BRCA1-associated breast cancer is likely distinct from other breast cancers as suggested by the high frequency of TSG silencing by allelic deletion (33) and the low frequency of TSG methylation (34). Our series includes only one BRCA1 and four BRCA2 gene mutation carriers. No TSG methylation was identified in the BRCA1 mutation carrier; one BRCA2 mutation carrier had an acellular lavage; one BRCA2 gene mutation carrier had no TSG methylation; and two BRCA2 mutation carriers showed high levels of methylation for multiple genes (Table 6). TSG methylation may be a feature of BRCA2 mutation–associated breast carcinogenesis.

We have previously reported that in RP-FNA samples cyclin D2 methylation is essentially cancer specific (18). In contrast, cyclin D2 was methylated in 10% of NDL samples from breasts unaffected with breast cancer and was highly correlated with methylation of other genes in the same duct. Patients with the most abnormal methylation results (methylation fraction >0.10 for ≥2 genes) are shown in Table 6. Cyclin D2 methylation is a prominent feature of these cases. These methylation patterns may represent mammographically and magnetic resonance imaging occult breast cancer; alternatively, cyclin D2 gene silencing may predispose benign breast epithelium to methylation of multiple other genes.

We have previously reported that RAR-ß2 methylation is detected in 32% of benign RP-FNA samples from women with breast cancer as compared with 9% for unaffected women (18). Bean et al. (35) identified RAR-ß2 gene methylation in 68% of RP-FNA samples from 38 high-risk women and reported that methylation was correlated with both sample cellularity and atypia. These observations are not confirmed in the current NDL study. Bean et al. used a qualitative assay and two different primer pairs to evaluate two separate, but overlapping, regions of the RAR-ß2 promoter termed M3 and M4. Our primers and probes covered most of the region designated as M4 in Bean's study and identified methylation in 12% of ducts from breasts unaffected by breast cancer as compared with 38% of RP-FNA samples that Bean's group evaluated at M4. We specifically selected primers and probes for each of the genes that would interrogate promoter regions known to be associated with gene silencing when methylated, but it is possible that other, neighboring regions could provide additional biologically relevant information. It is more likely, however, that the exfoliated cells retrieved by duct lavage are biologically distinct from cells collected by RP-FNA.

Determining methylation prevalence from QM-MSP data poses a challenge because the continuous methylation data must first be converted to binary data by setting a threshold for classifying a specific test as positive. We selected the 90th percentile of Gail lower-risk ducts as the threshold for this study to permit comparison with other studies. It must be recognized, however, that many ducts classified as methylation negative by this threshold actually contain small populations of cells with detectable levels of methylation. Methylation of ≥2 genes provided good discrimination between Gail lower-risk ducts and Gail high-risk ducts for methylation thresholds ranging from >0.0001 to ~0.05. Composite scores based on summing methylation fractions for the five genes did not provide any discrimination. Based on the work of Fackler et al. (29) and data from the current study, it seems that high threshold levels are appropriate if the intent is to discriminate between cancer and benign cells, but low-level methylation of multiple genes is the appropriate disease risk discriminator. Biologically, this likely reflects small populations of cells that have gained a survival advantage by methylating multiple TSGs.

Marked atypia, but not mild atypia, was associated with the risk classification of the breast providing the sample and also correlated with TSG methylation by univariate analysis, but not multivariate analysis, that adjusted for menopausal status, sample cellularity, and risk classification. Of the 52 ducts with either marked atypia or methylation of ≥2 genes, only 15% were positive for both markers. Marked atypia provided the same degree of discrimination between Gail lower-risk ducts and Gail high-risk ducts (OR, 3.48) as measures of TSG methylation (OR, 3.51), but provided better discrimination between Gail lower-risk ducts and ducts contralateral to a breast cancer (OR, 6.91 versus 4.21). The combination of marked atypia or TSG methylation improved the discrimination slightly (Fig. 3). Interestingly, perimenopausal status, abundant cellularity, and increasing risk group classification were independently associated with both marked atypia and TSG methylation by multivariate analysis. These data suggest that both marked atypia and TSG methylation are markers of breast cancer risk but the underlying processes giving rise to these end points may be different.

We used the validated Gail model as a surrogate for breast cancer incidence, but this will have resulted in a significant number of misclassifications as the model does not discriminate well between individuals who will develop breast cancer and those who will not (2, 3). The fact that the prevalence of marked atypia and TSG methylation both track with breast cancer risk is encouraging. The ongoing multi-institutional Serial Evaluation of Ductal Epithelium trial sponsored by Cytyc Health Corporation is designed to determine whether lavage atypia predicts increased breast cancer incidence.

In summary, given the anatomic complexity of the mammary ductal system, NDL does not seem to be suitable for the early detection of focal processes, such as carcinoma, and it is unlikely that the addition of molecular biomarker assays would solve this difficulty. Alternatively, NDL may be suitable for the detection of more diffuse risk-associated alterations in the ductal epithelium. Marked atypia and TSG methylation seem to have promise in this regard as each is independently associated with highly cellular samples, Gail model risk classifications, and a personal history of breast cancer. Marked atypia is more predictive of breast cancer risk classification than TSG methylation, but the two markers are largely independent, suggesting that different strategies may be required to target the underlying mechanisms. Cytologic assessment and QM-MSP analysis of NDL samples may provide an approach for identifying high-risk women and for monitoring the effects of interventions designed to reduce breast cancer risk.


   Footnotes
 
Grant support: Department of Defense Breast Cancer Research Program IDEA award DAMD17-01-1-0421.

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.

5 http://www4.utsouthwestern.edu/breasthealth/cagene Back

Received 12/12/06; revised 6/21/07; accepted 6/26/07.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Gail MH, Brinton LA, Byar DP, et al. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. J Natl Cancer Inst 1989;81:1879–86.[Abstract/Free Full Text]
  2. Costantino JP, Gail MH, Pee D, et al. Validation studies for models projecting the risk of invasive and total breast cancer incidence. J Natl Cancer Inst 1999;91:1541–8.[Abstract/Free Full Text]
  3. Rockhill B, Spiegelman D, Byrne C, Hunter DJ, Colditz GA. Validation of the Gail et al. model of breast cancer risk predicition and implications for chemoprevention. J Natl Cancer Inst 2001;93:358–66.[Abstract/Free Full Text]
  4. Tlsty TD, Crawford YG, Holst CR, et al. Genetic and epigenetic changes in mammary epithelial cells may mimic early events in carcinogenesis. J Mammary Gland Biol Neoplasia 2004;9:263–74.[CrossRef][Medline]
  5. Bae YK, Brown EG, Bornman D, et al. Hypermethylation in histologically distinct classes of breast cancer. Clin Cancer Res 2004;10:5998–6005.[Abstract/Free Full Text]
  6. Lehmann U, Langer F, Feist H, Glockner S, Hasemeier B, Kreipe H. Quantitative assessment of promoter hypermethylation during breast cancer development. Am J Pathol 2002;160:605–12.[Abstract/Free Full Text]
  7. Fackler MJ, McVeigh M, Mehrotra J, et al. Quantitative multiplex methylation-specific PCR assay for the detection of promoter hypermethylation in multiple genes in breast cancer. Cancer Res 2004;64:4442–52.[Abstract/Free Full Text]
  8. Parrella P, Poeta ML, Gallo AP, et al. Nonrandom distribution of aberrant promoter methylation of cancer-related genes in sporadic breast tumors. Clin Cancer Res 2004;10:5349–54.[Abstract/Free Full Text]
  9. Umbricht CB, Evron E, Gabrielson E, Ferguson A, Marks J, Sukumar S. Hypermethylation of 14-3-3{sigma} (stratifin) is an early event in breast cancer. Oncogene 2001;20:3348–53.[CrossRef][Medline]
  10. Jeronimo C, Costa I, Martins MC, et al. Detection of gene promoter hypermethylation in fine needle washings from breast lesions. Clin Cancer Res 2003;9:3413–7.[Abstract/Free Full Text]
  11. Li Z, Meng ZH, Chandrasekaran R, et al. Biallelic inactivation of the thyroid hormone receptor ß1 gene in early stage breast cancer. Cancer Res 2002;62:1939–43.[Abstract/Free Full Text]
  12. Fackler MJ, McVeigh M, Evron E, et al. DNA methylation of RASSF1A, HIN1, RAR-ß, Cyclin D2 and Twist in in situ and invasive lobular breast carcinoma. Int J Cancer 2003;107:970–5.[CrossRef][Medline]
  13. Wrensch MR, Petrakis NL, Miike R, et al. Breast cancer risk in women with abnormal cytology in nipple aspirates of breast fluid. J Natl Cancer Inst 2001;93:1791–8.[Abstract/Free Full Text]
  14. Fabian CJ, Kimler BF, Zalles C, Mal E. Short-term breast cancer prediction by random periareolar fine-needle aspiration cytology and the Gail risk model. J Natl Cancer Inst 2000;92:1217–27.[Abstract/Free Full Text]
  15. Claus EB, Risch N, Thompson WD. Autosomal dominant inheritance of early-onset breast cancer. Cancer 1994;73:643–51.[CrossRef][Medline]
  16. Berry DA, Parmigiani G, Sanchez J, Schildkraut J, Winer E. Probability of carrying a mutation of breast-ovarian cancer gene BRCA1 based on family history. J Natl Cancer Inst 1997;89:227–38.[Abstract/Free Full Text]
  17. Ries L, Harkins D, Krapcho M, et al., editors. SEER cancer statistics review, 1975-2003. Bethesda (MD): National Cancer Institute; 2005.
  18. Lewis CM, Cler LR, Bu D-W, et al. Promoter hypermethylation in benign breast epithelium in relation to predicted breast cancer risk. Clin Cancer Res 2005;11:166–72.[Abstract/Free Full Text]
  19. Masood S, Frykberg ER, McLellan GL, Scalapino MC, Mitchum DG, Bullard JB. Prospective evaluation of radiologically directed fine-needle aspiration biopsy of nonpalpable breast lesions. Cancer 1990;66:1480–7.[CrossRef][Medline]
  20. Bushnaq ZI, Ashfaq R, Leitch AM, Euhus D. Patient variables that predict atypical cytology by nipple duct lavage. Cancer 2007;109:1247–54.[Medline]
  21. Cler L, Bu D-w, Lewis CM, Euhus DM. A comparison of five methods for extracting DNA from paucicellular clinical samples. Mol Cell Probes 2006;20:191–6.[Medline]
  22. Clark SJ, Harrison J, Paul CL, Frommer M. High sensitivity mapping of methylated cytosines. Nucleic Acids Res 1994;22:2990–7.[Abstract/Free Full Text]
  23. Evron E, Dooley WC, Umbricht CB, et al. Detection of breast cancer cells in ductal lavage fluid by methylation-specific PCR. Lancet 2001;357:1335–6.[CrossRef][Medline]
  24. Zochbauer-Muller S, Fong KM, Virmani AK, Geradts J, Gazdar AF, Minna JD. Aberrant promoter methylation of multiple genes in non-small cell lung cancers. Cancer Res 2001;61:249–55.[Abstract/Free Full Text]
  25. Virmani AK, Rathi A, Sathyanarayana UG, et al. Aberrant methylation of the adenomatous polyposis coli (APC) gene promoter 1A in breast and lung carcinomas. Clin Cancer Res 2001;7:1998–2004.[Abstract/Free Full Text]
  26. Krop IE, Sgroi D, Porter DA, et al. HIN1, a putative cytokine highly expressed in normal but not cancerous mammary epithelial cells. Proc Natl Acad Sci U S A 2001;98:9796–801.[Abstract/Free Full Text]
  27. Burbee DG, Forgacs E, Zochbauer-Muller S, et al. Epigenetic inactivation of RASSF1A in lung and breast cancers and malignant phenotype suppression. J Natl Cancer Inst 2001;93:691–9.[Abstract/Free Full Text]
  28. Widschwendter M, Berger J, Hermann M, et al. Methylation and silencing of the retinoic acid receptor-ß2 gene in breast cancer. J Natl Cancer Inst 2001;92:826–32.
  29. Fackler MJ, Malone K, Zhang Z, et al. Quantitative multiplex methylation-specific PCR analysis doubles detection of tumor cells in breast ductal fluid. Clin Cancer Res 2006;12:3306–10.[Abstract/Free Full Text]
  30. Khan SA, Wiley EL, Rodriguez N, et al. Ductal lavage findings in women with known breast cancer undergoing mastectomy. J Natl Cancer Inst 2004;96:1510–7.[Abstract/Free Full Text]
  31. Going JJ, Moffat DF. Escaping from Flatland: clinical and biological aspects of human mammary duct anatomy in three dimensions. J Pathol 2004;203:538–44.[CrossRef][Medline]
  32. Yan PS, Venkataramu C, Ibrahim A, et al. Mapping geographic zones of cancer risk with epigenetic biomarkers in normal breast tissue. Clin Cancer Res 2006;12:6626–36.[Abstract/Free Full Text]
  33. Esteller M, Fraga MF, Guo M, et al. DNA methylation patterns in hereditary human cancers mimic sporadic tumorigenesis. Hum Mol Genet 2001;10:3001–7.[Abstract/Free Full Text]
  34. Tisserand P, Fouquet C, Barrois M, et al. Lack of HIN1 methylation defines specific breast tumor subtypes including medullary carcinoma of the breast and BRCA1-linked tumors. Cancer Biol Ther 2003;2:559–63.[Medline]
  35. Bean GR, Scott V, Yee L, et al. Retinoic acid receptor-B2 promoter methylation in random periareolar fine needle aspiration. Cancer Epidemiol Biomarkers Prev 2005;14:790–8.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Clin. Cancer Res.Home page
M. J. Fackler, A. Rivers, W. W. Teo, A. Mangat, E. Taylor, Z. Zhang, S. Goodman, P. Argani, R. Nayar, B. Susnik, et al.
Hypermethylated Genes as Biomarkers of Cancer in Women with Pathologic Nipple Discharge
Clin. Cancer Res., June 1, 2009; 15(11): 3802 - 3811.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
J. T. Loud, A. C.M. Thiebaut, A. D. Abati, A. C. Filie, K. Nichols, D. Danforth, R. Giusti, S. A. Prindiville, and M. H. Greene
Ductal Lavage in Women from BRCA1/2 Families: Is There a Future for Ductal Lavage in Women at Increased Genetic Risk of Breast Cancer?
Cancer Epidemiol. Biomarkers Prev., April 1, 2009; 18(4): 1243 - 1251.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
D. M. Euhus, D. Bu, S. Milchgrub, X.-J. Xie, A. Bian, A. M. Leitch, and C. M. Lewis
DNA Methylation in Benign Breast Epithelium in Relation to Age and Breast Cancer Risk
Cancer Epidemiol. Biomarkers Prev., May 1, 2008; 17(5): 1051 - 1059.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Euhus, D. M.
Right arrow Articles by Lewis, C. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Euhus, D. M.
Right arrow Articles by Lewis, C. M.
Related Collections
Right arrow Risk Assessment
Right arrow Risk Assessment: Biomarkers

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online