Xiaowei Chen, PhD
Office Phone: 215-214-4288
Global DASE analysis to identify novel coding or no-coding genetic factors for breast tumorigenesis
The significant mortality associated with breast cancer suggests a clear need to improve current research strategies to identify new genes that predispose women to breast cancer. Therefore, we hypothesize that DASE is a functional index for cis-acting variants and pathogenic mutations, and global profiling of DASE in breast cancer precursor tissues can be used to identify novel causative alleles for breast cancer susceptibility. Although other approaches have been examined for DASE profiling, SNP array-based technology is the most common approach for global DASE analysis. With the advance of array technology, new generation arrays carry much higher content of SNP probes (> 1 million). In our ongoing projects supported by Susan G. Komen, we have employed the Illumina® Omni1 array for genome-wide allele-specific expression (GWASE) measurements in a set of eight normal HMEC established from breast cancer patients. In order to use meaningful data, the raw results reported from Illumina were subjected to multiple filters, and only heterozygous SNPs that had significant intensity signals were left for DASE analysis. We have developed two statistical methods, SNP- and gene-based approaches, which allow us to identify 90 and 143 genes with significant DASE>2 (P ≤ 0.01, FDR ≤ 0.05), respectively. Importantly, 34 coding and 26 non-coding candidate loci are identified by both approaches. Ingenuity Pathway Analysis revealed two major interaction networks, which involve a variety of biological functions, including cell death and cell-to-cell signaling. One of the networks includes known cancer causative genes, ZNF331, USP6 and DMBT1. Our study demonstrated that global DASE profiling could be a powerful new approach to identify breast cancer risk alleles.
Intra-individual heterogeneity among pre-malignant/pre-invasive lesions
The heterogeneity among multiple DCIS lesions within the same patient, also diagnosed with invasive IDC, has not been well evaluated, leaving clinical and research implications of this intra-individual heterogeneity of DCIS yet to be explored. In our recent study, we have clearly demonstrated that intra-individual heterogeneity in multi-lesional DCIS is very common in patients with concurrent diagnoses of IDC. Our results showed that expression of PR, HER2, Ki-67, and p16 varied significantly in DCIS lesions from a single patient. In addition, seventy-two percent of the individuals had heterogeneous expression in at least 2 out of 6 IHC markers tested. Furthermore, a subpopulation of DCIS lesions (Subgroup IIb) had a higher Ki-67 index and positive p53 expression, but lower p16 staining intensity, than those in DCIS lesions (Subgroup I) with different molecular subtypes from the adjacent IDC. In comparison to other DCIS subgroups (I or IIa), type IIb DCIS lesions had the same molecular subtypes as the adjacent IDC but not the same subtypes as the adjacent normal terminal duct lobular units (TDLU). Our findings suggested the existence of DCIS subgroups with different degrees of “aggressiveness” (i.e. Subgroup I, indolent DCIS; Subgroup IIb, invasion-prone DCIS). Supported by these novel findings, we hypothesize that the “aggressive” DCIS subgroups which give rise to IDC share the same gene signatures as adjacent IDC, and these gene signatures will not only serve as risk markers to differentiate “indolent” or “invasion-prone” DCIS subpopulations but will also play a role in the DCIS initiation and progression.
Identification of breast cancer-associated lincRNA candidates using high-dense SNP arrays
The intergenic portion of the human genome is pervasively transcribed to produce large numbers of long intergenic non-coding RNAs (lincRNAs), which have recently been suggested to regulate a wide array of transcriptional and posttranscriptional processes in both physiological and disease conditions. Results from our previous studies showed that lincRNA genes appear to be less susceptible to copy number variations (CNV) in comparison to coding and other intergenic regions, indicating lincRNAs may be essential for maintaining cell survival. Here, we established a novel genome-wide approach, employing high-density SNP arrays, to globally identifying lincRNA genes whose expression is altered in breast cancers. We hypothesize that the lincRNAs whose expression is frequently altered during malignant transformation likely influence breast tumor progression and thus are new breast cancer candidate genes.