Abstract
Background:
Breast cancer risk increases during pregnancy and remains elevated for a number of years thereafter. Cancer-associated proteins that are secreted into breast milk may provide a means to detect cancer in the lactating breast or to assess future breast cancer risk.
Objective:
To determine whether proteins linked to breast cancer would be differentially expressed in matched (both breasts from each participant) human milk samples collected from women with unilateral breast cancer.
Methods:
Five cancer-associated proteins (basic fibroblast growth factor [bFGF], YKL-40, neutrophil gelatinase-associated lipocalin, and transforming growth factor β1 and β2) were analyzed in milk provided by 5 lactating women, 4 of whom were known to have cancer in 1 breast (and the opposite breast clinically disease free) at the time of milk collection and 1 who developed breast cancer 2 years after milk collection.
Results:
Expression was significantly higher for TGFβ2 (P = .03) and bFGF (P =.03) in the breasts with cancer.
Conclusion:
These proteins may play a role in assessing a woman’s risk of pregnancy-associated breast cancer. Because of variable protein concentration among patients and the limited sample size, the results are considered preliminary.
Well Established
Breast cancer diagnosed during pregnancy and lactation is difficult to diagnose and may be more aggressive than breast cancer diagnosed at other times.
Newly Expressed
We identified breast cancer-related proteins that are up-regulated in the milk of lactating women known to have cancer at the time of milk collection or who subsequently developed breast cancer.
Background
Breast cancer is the leading cause of cancer death for women 15-29 years of age. 1 Approximately 11 000 new cases of in situ or invasive breast cancer and 1100 breast cancer deaths occur in women younger than age 40. 2 Young women with breast cancer are more likely than older women to present with advanced disease and have worse outcomes. 3 Breast cancer in young women is more likely to be hormone insensitive and of higher grade, with a higher proliferative rate 4 and a greater likelihood of bone marrow metastases. 5 Cancer is more difficult to detect in a young woman’s breast, which is denser than after menopause, making both physical exam and mammogram evaluation more difficult.
The breast is the second most common site of pregnancy-associated cancer (after cervical), which translated to approximately 5000 annual new cases of pregnancy-associated breast cancer (PABC) in the United States in 2005. 6 Pregnancy-associated breast cancer is most often defined as breast cancer diagnosed during pregnancy or within the first 12 months postpartum, although recent findings suggest that the increased risk of PABC may extend for 5 years or longer postpartum. 7 Currently, the average age of women with PABC is between 32 and 38 years. 8 The incidence of PABC may increase, as a growing number of women are choosing to postpone childbearing until their mid- to late 30s. We demonstrated that the fluid present in the ducts of nonlactaing women contains secreted proteins that are predictive of breast cancer. 9 In the present study, we chose 5 proteins to evaluate in matched milk samples collected from women who had unilateral breast cancer.
The angiogenic protein basic fibroblast growth factor (bFGF) has been detected in human milk. 10 Levels of bFGF in nipple aspirate fluid were positively associated with cancer. 11 YKL-40 is a secreted angiogenic glycoprotein that is involved in normal mammary gland involution.11,12 High serum levels are associated with breast cancer metastases and reduced survival. 13 The homolog of neutrophil gelatinase-associated lipocalin (NGAL), which is expressed in breast carcinomas, has been identified in cow’s milk. 14 Inhibition of NGAL impairs breast tumorigenesis and metastasis, 15 and increased NGAL expression is associated with decreased disease-specific survival. 16 Transforming growth factors β1 and β2 (TGFβ) isoforms regulate normal breast development, apoptosis, and matrix remodeling during breast involution induced by breast weaning, with biphasic effects on tumor progression, acting as tumor suppressors to prevent cancer formation and promoting invasion and metastasis at later stages of cancer progression. 17
Whereas bFGF, TGFβ, and YKL-40 are involved in normal mammary gland development and involution, aberrant expression may promote carcinogenesis.18,19 NGAL has been associated with breast cancer in multiple reports. The objective of this study was to determine the levels of bFGF, YKL-40, NGAL, and 2 isoforms of TGFβ to assess their potential role in PABC by comparing levels in the affected versus the unaffected breast.
Methods
Recruitment
Five lactating women diagnosed with breast cancer within 22 months of pregnancy were recruited to an IRB-approved study being conducted at the University of Massachusetts–Amherst. Recruitment was primarily Web-based and used emails sent by the Love-Avon Army of Women (http://www.armyofwomen.org/). For 38 healthy women, mature milk was collected after providing consent to an IRB-approved study being conducted at the University of North Dakota (UND). Recruitment for the women at UND was through local media and flyers.
Milk Sample Collection
Consented women pumped or hand-expressed milk from each breast on a single day and returned the milk on ice either directly (some UND samples) or in a collection box overnight (all University of Massachusetts–Amherst samples, remainder of UND samples). Samples were processed immediately upon arrival. Each woman completed a health and reproductive history questionnaire. Each woman with cancer provided a copy of her breast biopsy/pathology report.
Assessment of Protein Biomarkers
For the 5 women with cancer diagnosed in 1 breast but not the other, aliquots of frozen whole milk and/or milk samples diluted in sterile phosphate-buffered saline were shipped on dry ice to the laboratory of Dr. Edward Sauter for analysis. For 38 healthy women who were cancer free, matched milk samples from both breasts were collected by Dr. Sauter. For all samples, total protein was determined using a bicinchoninic acid protein assay kit from Pierce (Rockford, IL), and individual protein levels determined as outlined below, in all cases according to the manufacturer’s instructions. Samples were diluted as necessary so that the measurement for each protein was within the range detectable by the kit. All samples were treated in similar fashion and were analyzed by the same technician. Results were then adjusted to reflect the protein concentration in undiluted milk. The intra-assay coefficient of variation (CV) was < 10% for all kits except NGAL, where the CV was 18%.
bFGF
bFGF was analyzed using an enzyme-linked immunosorbent assay (ELISA) kit from R&D Systems (Minneapolis, MN). The kit uses a quantitative sandwich enzyme immunoassay technique. The detection limit of the kit is 10 ng/L.
YKL-40
YKL-40 was analyzed by immunoassay (Quidel Corporation, San Diego, CA). The kit uses a monoclonal anti-YKL-40 antibody conjugated to biotin that binds to streptavidin and captures YKL-40 in the standard or sample. The detection limit is 10 ng/mL.
NGAL
NGAL content in milk was measured by immunoassay (BiPorto Diagnostics, Denmark). The detection limit of the NGAL kit is 4.0 pg/mL.
TGFβ isoforms
The protein expression levels of TGFβ1 and β2 in milk samples were determined by immunoassay (R&D Systems). The detection limits of the TGFβ1 and β2 kits are 4.61 and 7.0 pg/mL, respectively.
Statistical Analysis
To determine whether the difference in expression between matched cancer and healthy breasts was possibly a random event, we compared the expression of bFGF, the protein with the greatest difference in mean expression (Table 1), in mature milk collected from both breasts of 38 healthy women.
Abbreviations: MeanLT, mean of the log-transformed data; NA, not available at time of analysis; ND, not detected; SDLT, standard deviation of the log-transformed data.
Note: Trends for significance for each marker were similar controlling for total milk protein concentration. P values shown are for analyses conducted on log-transformed data; analyses conducted prior to log transformation were not significant.
The data from both women with breast cancer and controls were not normally distributed and were log transformed. Paired t tests were used to compare protein levels between breasts, both for healthy women and for women who developed cancer, on both the untransformed and log transformed data. Statistical significance for the 2-tailed tests was set at P < .05.
Results
The women who developed breast cancer were of similar age at the time of milk donation (mean = 36.2 y; SD = 2.39); age at first birth (mean = 32 y; SD = 4.53); and number of live births (mean = 2.2; SD = 0.84). However, the age of the baby they were nursing varied greatly (median 180 d, range 7-660 d). The time between milk donation and cancer diagnosis also varied among the 5 women. Milk was collected 24 months before cancer was detected in the first woman; immediately before core biopsy in a second; after diagnostic core biopsy but before definitive surgery (either excisional biopsy or mastectomy) in the third; and after definitive surgery in the fourth and fifth women.
The Level of Secreted Proteins in Milk Varies among Women and Breasts
As shown in Table 1, a total of 49 measurements were made on milk samples from the 5 women (each woman provided a sample from each breast, but 1 milk sample was not available at the time of 1 assay). Of the 49 measurements, proteins were detected in 47. For 4 of the 5 women (participants 2-5), the levels of all 5 proteins were higher in the breast with cancer, regardless of the length of time she had been nursing (as indicated by the baby’s age) or the time between milk donation and diagnosis of cancer. In contrast, for participant 1, only TGFβ2 and bFGF were higher in the breast with cancer. Interestingly, participant 1 provided her milk samples roughly 2 years prior to her diagnosis with cancer. Despite the wide range in the relation between collection time and cancer diagnosis, analysis of the log-transformed data indicates that the expression of TGFβ2 and bFGF was significantly higher in milk that came from the breast with cancer as compared to milk from the healthy breast.
Mean expression of bFGF from the 38 controls was 0.86 pg/mL in both the left (SD = 1.39) and right (SD = 1.40) breasts, and it was not significantly different for the untransformed (P = .99) or the log-transformed data. Among the 5 women found to have breast cancer, mean bFGF expression in the diseased breast was 5.74 pg/mL, and in the presumed healthy breast 1.56 pg/mL. The difference in bFGF expression in the breasts of controls versus the breasts of the women with cancer was significant (P = .002), whereas the difference in bFGF expression between breasts of controls versus the presumed healthy breast of the women who had cancer in their contralateral breast was not (P = .41).
Discussion
It is our premise that breast milk contains components that can inform regarding current and/or future breast cancer risk of the mother. Our findings support this premise and provide preliminary evidence that breast milk assessment may prove useful in determining current and/or future breast cancer risk. Breast cancer that develops during or shortly after pregnancy is frequently difficult to diagnose and, at least by some reports, more aggressive than cancer diagnosed at other times in a woman’s life. The proliferative state of the breast during pregnancy and lactation provides a medium of growth factors that allows both normal and malignant breast epithelial growth, whereas breast involution once the baby is weaned presents an environment not unlike that seen during wound healing, in which growth factors are up-regulated.
Consistent with findings in breast tumors16,21 and breast fluid obtained from the breasts of women who are not lactating, 11 we observed higher mean levels of all 5 proteins in the breast milk of women who developed breast cancer than in the opposite healthy breast. On the other hand, we did not find a different level of bFGF expression comparing right to left breast in healthy women. The protein levels in both the healthy breast and breast diagnosed with cancer of participant 1 (Table 1) are higher than for the other participants, and therefore this individual may be an outlier. Interestingly, YKL-40, which plays a role in normal breast involution, was lowest in the breasts of the 2 women who nursed their children for over 1 year.
Conclusion
Our observations are limited by sample size and therefore are more useful for generating hypotheses than drawing conclusions. Nonetheless, we believe our observations are sufficiently compelling to warrant a validation study to assess the expression of TGFβ2 and bFGF in the breast milk of normal and high-risk women, as well as in those with PABC, to assess the usefulness of these proteins in PABC risk stratification.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funded in part by Avon Foundation for Women and the CDMRP-BCRP.