Identification of new breast cancer candidate genes associated with stromal invasion.

Abstract

Abstract Abstract #4036 Background: We recently reported our results of a study identifying a number of DNA copy number changes that differentiated low grade ductal carcinoma in situ (DCIS) lesions with and without associated stromal invasion. We now describe a series of validation experiments that support the role of four genes in early breast cancer progression, most of which were not previously implicated in mammary neoplasia.
 Materials and Methods: DNA was extracted from microdissected paraffin sections of formalin fixed low grade DCIS lesions with an invasive component (n=25), as well as morphologically similar DCIS lesions without associated invasion within at least 5 years (n=20). The DNA was used for comparative PCR, with GAPDH as the reference gene. Specific primers were designed for fourteen candiate genes. Amplicon sizes ranged from 80 to 90 base pairs. Using log2 ratio cutoffs of +1.0 and -0.5, respectively, the frequency of copy number gains and losses was determined for each candidate gene. Additional validation studies employed genomic DNA samples from an independent cohort of 35 microdissected DCIS lesions with (n=17) and without (n=18) an associated invasive component.
 Results: Our previous whole genome profiling study using high density BAC arrays revealed that DNA copy number changes on 1q, 3p, 5q, 6p, 9p, 10q, 11q, 12q, 16p, 16q, 18q, 20q, and 22q may occur at different rates in low grade DCIS lesions with and without associated stromal invasion. From the differentially amplified or deleted chromosomal sites, thus far we have selected fourteen candidate genes for validation by comparative PCR, utilizing aliquots of the same DNA samples that had previously been used for array CGH analysis. Four genes were confirmed to have differential copy number changes. NCOR2 (on 12q24.31), DYNLRB2 (on 16q23.2), CELSR1 and UPK3A (both on 22q13.31) were more frequently amplified in DCIS lesions that had no invasive component. Moreover, NCOR2 and DYNLRB2 showed more frequent copy number losses in DCIS lesions that had progressed to invasive carcinoma. For NCOR2, these observations held true in an independent cohort of microdissected low grade DCIS cases. The other ten candidate genes revealed significant frequencies of copy number changes that were not differentially distributed, however. TEME9, CALN1, and CENTA1 were amplified in 30-40% of all cases. IRF8, THAP7, and FIGNL1 were deleted in 20-50% of all cases. FBXO31, CDH13, IKZF1, and MAPK9 revealed both gains and losses in >20% of all cases.
 Discussion: Our validation experiments have provided additional evidence for the role of several novel genes in early breast cancer progression. Four of the candidate genes we assayed so far may function as invasion suppressor genes. Ten additional genes showed significant copy number changes but did not discriminate DCIS lesions with and without the propensity to invade the stroma. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 4036.