Early detection is a major factor contributing to the 2.3% annual decline in breast cancer death rates over the past 10 years [17
]. Nonetheless 40,480 women in the U.S. were projected to die from breast cancer in 2008 [18
], in part because currently available breast cancer screening tools such as mammography and breast examination miss 10–40% of early breast cancers and are least effective in detecting cancer in young women, whose tumors are often more aggressive. An invasive needle or surgical biopsy must be performed when an area of suspicion is identified in order to confirm, by cytologic or histologic evaluation, the presence of malignancy, even though 66–85% of abnormalities are benign [19
In principle, there are three different ways in which biomarkers can be used. First, they can be used to differentiate normal from diseased states. Second, they can be used to determine into which prognostic groups subjects should be placed. Finally, biomarkers can be used to monitor response to therapy. The identification of biologic markers such as miRNA species collected noninvasively in the circulation that would distinguish between women with or without breast cancer is therefore of crucial importance in early cancer detection, to improve disease free and overall survival from the disease. There is a recent report which outlines the usefulness of miRNA signatures to distinguish normal from diseased tissues and tissue origin, relying upon miRNA signatures from formalin fixed, paraffin embedded and frozen tissue specimens [20
], demonstrating the stability of miRNAs. Our miRNA findings in archival serum support and extend these findings, with the potential that miRNA analysis may prove useful for noninvasive breast cancer detection.
We first tested if spiked-in cel-miR-39 and has-miR-145 were detectable in a 10 year old human serum and a goat serum specimen. The specimen was divided into five samples of equal volume. Decreasing inputs of single-strand cel-miR-39 and has-miR-145 (105, 103, 10, 10-1 and 0 pg) added to each sample (Figures and ) led to the corresponding mean Ct values (19.2, 25.8, 32.4, 37.1 and 38.7), respectively, with similar values for cel-39 and has-145. The Ct value of 37.1 obtained using the most dilute sample (10-1 pg) was slightly lower than background of 38.7, indicating that the level of endogenous miR-145 in the human serum to which the synthetic single stranded miR-145 mimic was spiked in was very low, and therefore unlikely to have influenced the amplification of the synthesized miR-145. The equidistant curves down to a concentration of 10-1pg/400 μL serum suggest linear recovery (Figure ) and efficient amplification (Figure , slope = -3.0) of the added miRNAs. The Ct values (Figure ) for endogenous GAPDH mRNA in the serial serum samples to which has-miR-145 was spiked in were approximately 35. The results demonstrate that circulating RNA, including endogenous GAPDH mRNA and spiked-in cel-miR-39 and has-miR-145, can be efficiently extracted and amplified from serum.
Different quantities (105, 103, 10, 10-1, and 0 pg) of cel-miR-39 (blue curves) and has-miR-145 (red curves) inputs led to similar corresponding Ct values of 19.2, 25.8, 32.4, 37.10 and undetectable, respectively.
There was linear recovery (slope = -3.0) of the miR-39 and miR-145 at concentrations equal to and greater than 10-1pg/400 μL serum.
The Ct value for endogenous GAPDH mRNA in each of the serial human serum samples to which miR-145 was spiked in was approximately 35.
We then performed a pilot study in serum specimens from 21 women with and without breast cancer (Table ) to assess the expression of miR-16, -145 and -155 using 18s rRNA to normalize the relative expression of the miRNAs in archived serum specimens from women with and without breast cancer. Each miRNA and 18s rRNA was isolated and amplified efficiently in all subjects. The expression of all three miRNAs in serum was similar in samples from healthy women compared to those with breast cancer, unlike a prior report which found differences in the expression of mi-R-145 and -155 miRNAs in breast cancer tissue vs. benign tissue women [6
]. On the other hand, progesterone receptor (PR, p = 0.016) positive tumors had higher miR-155 expression than tumors that were negative for this receptor. miR-155 has been classified as a multifunctional miRNA, having an important role in both normal and pathologic processes in immunity, inflammation, cancer and cardiovascular disease [21
]. The significance of our finding regarding miR-155 and the PR will require a larger study.
The Ct Value of miR-16, miR-145 and miR155 in the Serum Samples with and without Breast Cancer
In determining the proper endogenous control for qPCR, we initially tried three small non-coding nucleoar RNAs (Rnu44, Rnu48 and Rnu66). Unfortunately, none were useful because we could not detect them. GAPDH was also tested but relatively high Ct values were observed, so 18s rRNA was tried levels adequate to serve as a control in all human serum samples. We confirmed that our techniques yielded adequate RNA extraction efficiency by measuring levels of spiked-in cel-miR-39 and has-miR-145. The spiked-in miRNAs had the advantage of lacking homology to and thereby interference from endogenous human miRNAs.