In the current study, we examined SIAH expression in primary DCIS treated with wide excision only and in adjacent normal breast tissue. We found that increased SIAH expression in DCIS was correlated with subsequent disease recurrence and increased risk of progression. Increased SIAH expression was also correlated with more aggressive features including comedo histology and increasing nuclear grade. However, aggressive features alone were not significantly associated with disease recurrence. We feel this potential discrepancy is not due to size of the patient cohort since other studies with larger numbers of patients have also found a lack of correlation between DCIS with aggressive features and increased recurrence in patients treated with wide excision only [16
]. Rather, we feel that SIAH has a relatively strong predictive value that is evident in the current study population.
SIAH expression in DCIS was positively correlated with expression levels in normal breast tissue with younger patients having higher SIAH expression in normal adjacent breast tissue. In a univariate logistic regression analysis, SIAH expression in DCIS and normal adjacent tissue were both able to predict recurrence. A multivariate analysis controlling for age, necrosis, comedo, and histologic subtype was not possible due to the small number of recurrences [17
]. Additionally, a ROC curve using 15% as a cutoff for positive SIAH staining resulted in a sensitivity of 76% and a specificity of 79% for disease recurrence. These data suggest that SIAH may be a potential biomarker for disease recurrence in patients with DCIS. In the future, we hope to expand our study to validate our findings in patients treated with radiation and/or tamoxifen in addition to surgery.
Numerous biomarkers have been investigated for risk stratification of patients with DCIS. Elevated Ki-67 levels, p53 mutations, and HER2 amplification are known to be associated with increased nuclear grade and necrosis which have also been associated with disease recurrence and progression [reviewed in 3
]. p53 mutations and HER2 amplification have been seen in 25 and 30% of DCIS, respectively. Other cell cycle markers have also been studied including p21, p27, and cyclin D1 [reviewed in 3
]. Despite these studies, no single biomarker has been identified to guide proper and effective therapies.
Another group, Tsikitis et al., has suggested a new classification scheme based on the combination of molecular and histologic characteristics of DCIS [19
]. The histological features of aggressive lesions include comedo necrosis, high nuclear grade, and negative hormone receptor status while the molecular features include increased Ki-67, p53 mutation, HER2 amplification, increased COX2 expression, loss of heterozygosity at 11q13, and increased angiogenesis [19
]. Less aggressive DCIS has lower Ki-67, normal p53, no HER2 amplification, and deletion of chromosome 16. They suggested that their classification scheme helps separate patients who will develop invasive breast cancer within a short period of time (2–5 years) necessitating more aggressive treatment from those who will develop invasive breast cancer over a longer period of time (10–20 years) [19
In the current study, SIAH expression was found to be correlated with nuclear grade and comedo histology. However, more importantly, we were able to show an increase in the odds ratio for recurrence with increased SIAH expression in DCIS as well as normal breast tissue. Despite the correlation of SIAH expression with more aggressive tumor biology, the exact mechanism of increased SIAH expression in tumorigenesis is unknown. However, increased SIAH expression has been seen in both neoplastic and non-neoplastic proliferating cells [9
]. Therefore, it is not surprising that SIAH expression is associated with more aggressive lesions given that these lesions also tend to have an increased mitotic index, i.e., increased Ki-67.
Paradoxically, SIAH was initially thought to be a tumor suppressor gene [reviewed in 8
]. However, recent studies have shown that decreased SIAH function results in inhibition of tumor growth [9
, reviewed in 8
]. Indeed, recent studies using anti-SIAH molecules (shRNA knock down and dominant negative SIAH) have shown that SIAH deficiency significantly impedes lung and pancreatic tumor growth in vitro and in vivo [9
]. Also, Bowtell and Möller found that SIAH expression was strongly upregulated in ER-positive breast tumors arguing against a role for SIAH as a tumor suppressor [unpublished observations, reviewed in 8
]. In contrast, we found no correlation between SIAH expression and ER expression in this study.
Interestingly, we also found an inverse correlation between SIAH expression in normal breast tissue and patient age (Spearman correlation = −0.54, P
< 0.001) as well as a trend toward an inverse correlation between SIAH expression in DCIS and patient age (Spearman correlation = −0.23, P
= 0.067). It has long been known that DCIS in younger patients has a higher risk of recurrence. In a recent study of 657 patients, Collins et al. found that DCIS in young women was more likely to be symptomatic and more extensive at presentation and was also more likely to involve cancerization of the lobules. Interestingly, there was no correlation found with histologic features known to be associated with more aggressive disease such as high nuclear grade, comedo necrosis, or architectural pattern. Collins et al. also interrogated currently used molecular markers (ER, PR, and Her-2) and found no correlation with age. They propose that more aggressive disease may be related to less screening in younger patients which causes lesions to become more extensive at presentation [14
Another study addressing molecular markers and histologic features of DCIS in relation to age showed a trend toward higher HER2 expression in younger patients with DCIS. They also showed a correlation of HER2 expression with high-grade lesions, ER and PR negativity, and increased Ki-67. However, due to a small number of patients examined, the study was unable to show a statistically significant difference in the presentation of these aggressive features in DCIS in younger patients [15
]. Again, this study failed to conclusively tie currently used molecular markers to lesions that are more aggressive, i.e., DCIS in younger patients. Our study demonstrated a significant correlation between increased SIAH expression in normal breast tissue adjacent to DCIS lesions and younger patient age suggesting an overall difference in breast tissue biology in younger and older patient populations which may predispose younger patients to the development of more aggressive DCIS.
Another biomarker, Enhancer of Zeste 2 (EZH2), a Polycomb family of transcription repressors, has also been shown to have an increased expression in normal breast tissue adjacent to atypical ductal hyperplasia and DCIS [20
]. The group that discovered this relationship was able to show that this was not simply a “field effect” but rather was a biomarker for breast tissue at risk for the development of invasive carcinoma. This relationship was demonstrated by increased EZH2 expression in benign breast biopsies from patients that later developed carcinoma as well as in morphologically normal mammary tissue from prophylactic mastectomy specimens of patients with BRCA1 mutations. Interestingly, like SIAH, EZH2 expression is increased with increasing nuclear grade and comedo-type histology. The authors of this study suggest that EZH2 could be used as a novel biomarker to predict which patients are at increased risk for the development of invasive breast carcinoma [20
Recently, Chen et al. examined the molecular profile of invasive ductal carcinoma (IDC) and compared it to normal adjacent breast tissue using microarray analysis [22
]. They found that a subset of normal adjacent breast tissue had a molecular signal similar to invasive ductal carcinoma despite the fact that there was no morphological evidence of carcinoma or other premalignant conditions such as DCIS or atypical ductal hyperplasia. The genes comprising the molecular signal for invasive ductal carcinoma were primarily those involved in cell adhesion and proliferation. Interestingly, genes involved in proliferation were over-represented in the molecular signal of the normal adjacent breast tissue with an invasive ductal carcinoma-like signature suggesting that genes involved in proliferation are important early in the development of breast carcinoma [22
]. Another group has found that SIAH expression is associated with proliferating cells [9
]. Therefore, increased SIAH expression may be a biomarker for cells with a more “proliferative” gene expression profile. This is consistent with the function of SIAH as an essential signaling module downstream of the RAS signaling pathway that drives cell proliferation, tumor growth, and metastasis. Also, the active stromal or adjacent pre-malignant tissues may be primed to facilitate tumor progression and invasion or are at an increased risk for the development of primary or recurrent carcinoma. Indeed, in our study, we found that a 1% increase in SIAH expression in DCIS lesions results in a 10% increase in the odds of recurrence while a 1% increase in SIAH expression in normal adjacent tissue results in a 31% increase in the odds of recurrence.
In summary, we evaluated SIAH expression in a cohort of 65 DCIS patients who underwent surgical resection only for local disease and with extensive follow-up data. We found that SIAH expression alone is a good predictor of disease recurrence in DCIS patients with a sensitivity of 76% and a specificity of 79%. It is conceivable that the use of SIAH expression in combination with other biomarkers such as EZH2 may further increase overall specificity and sensitivity of these novel prognostic biomarkers thereby allowing us to more accurately identify those patients at the highest risk for development and/or recurrence of breast carcinoma. This will allow for use of more aggressive treatments in the DCIS population with poor prognosis and may prevent overtreatment in patients with highly favorable prognosis.