One approach to assess the influence of a specific process such as age at FFTP is global (proteome wide) assessment. Unfortunately, this approach has limited sensitivity [29
] even with sample concentration approaches [30
], and lack reliable quantitation of individual protein concentration. We elected to use immunoassays which are highly sensitive and quantitative, to compare individual protein expression during the course of lactation.
A striking feature of the data is that the expression of most proteins varies during lactation, even after controlling for total protein concentration. A similar observation in a smaller sample was made by another group in the assessment of milk immunoglobulins and lysozyme [31
]. The most common trend that we observed was for the concentration of proteins to decrease from BL to 2 months after starting nursing, and then to increase between 2 months and W. Although the reason(s) for this are uncertain, a possible explanation is that the breast secretes a given amount of each protein daily, with greater milk volume during the 2 month lactation period compared to BL or W periods leading to a lower protein concentration. Indeed, total milk protein concentration has been shown to negatively correlate with 24 hour milk volume [32
], and gradual weaning over a period of three months has been shown to increase total protein as milk volume decreased [33
The most significant changes in expression from the start to the end of lactation in both FFTP age groups were in KLK6, 8 and TGFβ2. There was a significant decrease in KLK3 in women < 26 at the time of FFTP, but not in women with a later age of FFTP. The significance of the change in one but not the other group is uncertain. For KLK6, 8 and TGFβ2, there was a small change early (between BL and 2 months), with a larger change later (between 2 months and W) in lactation. For all three proteins there was an increase in mean expression at the time of weaning compared to BL.
KLK6 and 8 expression is downregulated in breast cancers [34
]. As such, it has been speculated that these genes may function to suppress tumors [13
]. KLK6 [13
] is epigenetically regulated with gene silencing in tumors. KLK6 may play a protective role against tumor progression by inhibiting the epithelial-to-mesenchymal transition [13
]. KLK8 has been shown to function as a serine protease [35
]. The mechanism of KLK8 gene silencing in breast cancer has not been reported, though alternative splice variants of the gene which influence prognosis have been reported in lung cancer [36
]. The increase in KLK8 expression from BL to W was greatest in women who whose FFTP occurred at a young age, which is consistent with known protective effect of early age at FFTP [2
] and a tumor suppressive effect of KLK8.
The significant changes in KLK6 protein expression during lactation and the report of regulation of KLK6 by methylation in breast cancer cells [13
] prompted us to examine KLK6 promoter methylation in cells obtained from breast milk. We detected surprisingly high levels of KLK6 promoter methylation with moderate differences among the three sampling times that did not explain the observed changes in protein expression. Indeed, the lowest mean level of methylation, 34%, occurred at 2 months. While we expected low methylation to be associated with the highest protein expression, KLK6 expression, on average, was highest in wean milk for all age categories (Table ). Whereas methylation scores were relatively high at all sampling times, repeated measures from ten women revealed average methylation changes during lactation of 37%.
To assess the likelihood that the observed variability in KLK6
methylation throughout lactation was an artifact, we examined the promoter methylation of p16
(data not shown), which we would not predict to vary with lactation duration. In contrast to KLK6, there was little methylation of p16 at each point, ranging from 1 to 5%, indicating that the variability in KLK6 methylation is not likely to be an artifact. It is important to note that, unlike the prior report associating KLK6 methylation with KLK6 expression in cultured malignant breast cells [13
], our clinical investigation evaluated breast cells collected from the breasts of healthy women. Regarding the variability in KLK6 methylation which we observed, we believe that this is at least partly due to differing percentages of epithelial cells and leucocytes that were present in the milk at the three sampling times, since DNA was isolated from the total cell pellet. Nonetheless, we confirmed that epithelial cells present in milk from healthy breasts have high levels of methylation by examining several epithelial-enriched cell fractions (data not shown).
Whereas we did not find a association between KLK6 methylation and protein expression, we did observe a highly significant association between KLK6 and TGFβ1 protein expression. TGFβs 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 in early stages of cancer and promoting invasion and metastasis at later stages [37
]. Our observation that the expression of these proteins are directly correlated, whereas they are inversely associated in breast cancer cells in culture [23
], is consistent with the pleiotropic effects of these molecules to suppress cancer initiation and promote later stages of cancer invasion and metastasis.
This report has limitations. Most notable is the limited sample of women whose FFTP was age 35 or greater, as well as the limited BL and 2 month sample sizes for the methylation analyses. We may have missed associations with age at FFTP due to this. A second limitation is that there was some variability between when the wean milk sample was collected and when the woman stopped nursing entirely. A third limitation is that we cannot say with certainty that the proteins identified which are upregulated at the time of weaning would put these women at lower risk of future breast cancer, since we do not have sufficient follow-up to address this. On the other hand, as previously discussed, we know from other reports that higher expression of KLK 6, 8, 10 and TGFβ2 have been associated with a lower risk of breast cancer [34