Study of Gene Expression Profiles of Breast Cancers in Indian Women

Abstract

Breast cancer is the most common cancer among women globally. In India, the incidence of breast cancer has increased significantly during the last two decades with a higher proportion of the disease at a young age compared to the west. To understand the molecular processes underlying breast cancer in Indian women, we analysed gene expression profiles of 29 tumours and 9 controls using microarray. In the present study, we obtained 2413 differentially expressed genes, consisting of overexpressed genes such as COL10A1, COL11A1, MMP1, MMP13, MMP11, GJB2, and CST1 and underexpressed genes such as PLIN1, FABP4, LIPE, AQP7, LEP, ADH1A, ADH1B, and CIDEC. The deregulated pathways include cell cycle, focal adhesion and metastasis, DNA replication, PPAR signaling, and lipid metabolism. Using PAM50 classifier, we demonstrated the existence of molecular subtypes in Indian women. In addition, qPCR validation of expression of metalloproteinase genes, MMP1, MMP3, MMP11, MMP13, MMP14, ADAMTS1, and ADAMTS5 showed concordance with that of the microarray data; wherein we found a significant association of ADAMTS5 down-regulation with older age (≥55 years) of patients. Together, this study reports gene expression profiles of breast tumours from the Indian subcontinent, throwing light on the pathways and genes associated with the breast tumourigenesis in Indian women.

Figure 1

Volcano plots showing the distribution of gene expression by microarray in total breast tumours and early- and late-onset breast tumours as compared to controls. The plot shows gene expression profiles of breast tumours. The plot was obtained between negative log p-value (y-axis) and log fold change (x-axis). Each dot represents one gene, genes shown in green colour had significant fold change (FC ≥ 1.5, and adjusted p ≤ 0.05) while the remaining genes depicted in red, black and orange colour didn’t reach significance. (a) Plot shows gene expression profiles of total tumours vs controls (b) Plot shows the gene expression profiles of early-onset tumours vs controls (c) Plot shows the gene expression profiles of late-onset tumours vs controls.

Figure 2

Unsupervised hierarchical clustering of differentially expressed genes. Heatmap showing the hierarchical clustering of tumours based on their gene expression. 2413 genes were found to be differentially expressed in tumours (FC ≥ 1.5, and adjusted p-value ≤ 0.05) forming distinct up-regulated and dow-nregulated clusters. Red colour represents up-regulation and green colour represents down-regulation. The differentially expressed genes are mentioned on the y-axis, and sample IDs are mentioned on the x-axis.

Figure 3

Showing gene network analysis of differentially expressed genes in breast tumours. Interactive gene networks were identified based on their position in the network by two measures; degree centrality, where the importance of a node is dependent on the number of connections to other nodes and betweenness centrality, which measures the number of shortest paths going through a node. Nodes with a higher degree are hubs of the network, and the size of the nodes is based on their degree values, with a bigger size accounting for larger degree values. The colour of the nodes is related to the expression of genes, where up-regulated nodes are shown in red and down-regulated nodes in green colour while the grey coloured nodes are those genes that are not present in our data set but are part of the PPI network (The network analysis was done with DEGs having FC ≥±5). Among the gene networks, AURKB, CENPA, TOP2A, BUB1, CCNB2, MMP1, and SPP1 were the most interactive nodes.

Figure 4

Heatmap showing hierarchical clustering of predicted molecular subtypes. Molecular subtypes were predicted using PAM50 classifier in breast tumours, consisting of subtypes viz. luminal A, luminal B, HER2/neu, basal and normal-like (FC ≥ 1.5, and adjusted p-value ≤ 0.05). Genes pertaining to each subtype formed distinct clusters. Red colour represents up-regulation and green colour depicts down-regulation. The subtypes are mentioned on x-axis while differentially expressed genes are mentioned on the y-axis.

Figure 5

Venn diagram showing the common and unique genes belonging to each molecular subtypes in breast tumours. Venn diagram showing differentially expressed genes unique in basal subtype (842) followed by HER2/neu (705), luminal B & A (415, 198) and normal-like subtypes (39).

Figure 6

Validation of gene expression of MMPs by quantitative reverse transcription PCR. Scatter plots showing the up-regulation of (a) MMP1 (p = 0.05), (b) MMP11 (p = 0.03), (c) MMP13 (p = 0.018) (d) MMP3 (p = 0.214), (e) MMP14 (p = 0.722) and down-regulation of ADAMTS1 (p = 0.009), (g) ADAMTS5 (p = 0.05) in breast tumours compared to controls. The values are the mean of log fold change normalized to endogenous controls, along with the standard error (shown by vertical bars) as obtained by Mann-Whitney U test.