Abstract P3-04-06: Higher gene expression of CSPG4 in the basal-like subtype of invasive breast cancer and its negative association with lymph node metastasis

Abstract

Abstract Background: Chondroitin sulfate proteoglycan 4 (CSPG4) was first identified as a melanoma specific antigen and demonstrated to stabilize cell-substratum interactions during early events of tumor cell spreading on endothelial cells and matrix proteins. In invasive breast cancers (IBCs), CSPG4 has been reported to be highly expressed in the triple-negative (TN) subtype. Western blotting analyses suggest that CSPG4 is expressed on TN IBC lines. Here we further explored the specific gene expression (GE) profiles associated with CSPG4 using the available IBC cases from The Cancer Genome Atlas (TCGA) project towards identification of potential drug targets for patients with TN IBC. Methods: The GE data (Agilent, log2 transformed) of 459 IBC tumors were downloaded from the TCGA data portal. PAM50 classification results of all samples were obtained from the TCGA breast cancer AWG group, including 203 Luminal A (LA), 113 Luminal B (LB), 51 HER2+, 84 basal-like (Basal), and 8 Normal-like (not used). Clinicopathologic data (age, race, T-, N-, and M-stages) were also obtained from the AWG group. Non-parametric Kruskal-Wallis test was used for cross-subtype analysis of CSPG4 expression, followed by Wilcoxon-Mann-Whitney test for pairwise comparison with Bonferroni adjustment. Chi-square tests were performed for categorical variation analyses. Differentially Expressed Gene (DEG) analysis was performed using the Bioconductor “samr” package for two class unpaired comparison, with p < 0.05 and FDR < 0.10 as the cut-off for DEGs. All other statistical analyses were performed using SAS. Results: GE level of CSPG4 was highest in the Basal subtypes (mean±SD, −0.92±1.37) compared to all other subtypes of IBCs, including LA (−1.50±0.66, p = 0.0013), LB (−1.86±0.68, p < 0.0001), and HER2 (−1.69±0.60, p = 0.0014). GE in LA was also significantly higher compared to LB (p < 0.0001). The Basal subtype exhibited a much wider variation of the expression level compared to other subtypes. To better understand this variation, we analyzed the age, T, N, and M across the 4 quartiles of the CSPG4 GE levels in the Basal subtype. Only N was significantly associated with the quartiles (p = 0.04), with N-positive rates of 45.8%, 54.2%, 32.0%, and 16.7% from the lowest to the highest quartiles respectively. Further DEG analysis between the top and bottom quartiles of the samples identified 46 genes, including MIA, SOX10, MSMB etc. Many of these genes showed a strikingly similar expression distribution pattern to CSPG4. Discussion: Our study confirmed the higher GE of CSPG4 in Basal subtype. In this subtype, patients expressing high CSPG4 in the tumor show a lower rate of local lymph node metastasis, implying that these tumors may not metastasize via the lymphatic vascular system. A number of DEGs were identified that are positively associated with CSPG4 expression. Although the roles these molecules play in promoting IBC progression needs to be further evaluated, our study may lead to the identification of a drug target for the Basal subtype of IBC patients. The views expressed in this article are those of the author and do not reflect the official policy of the Department of Defense, or U.S. Government. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P3-04-06.