Candidate luminal B breast cancer genes identified by genome, gene expression and DNA methylation profiling

Abstract

Breast cancers (BCs) of the luminal B subtype are estrogen receptor-positive (ER+), highly proliferative, resistant to standard therapies and have a poor prognosis. To better understand this subtype we compared DNA copy number aberrations (CNAs), DNA promoter methylation, gene expression profiles, and somatic mutations in nine selected genes, in 32 luminal B tumors with those observed in 156 BCs of the other molecular subtypes. Frequent CNAs included 8p11-p12 and 11q13.1-q13.2 amplifications, 7q11.22-q34, 8q21.12-q24.23, 12p12.3-p13.1, 12q13.11-q24.11, 14q21.1-q23.1, 17q11.1-q25.1, 20q11.23-q13.33 gains and 6q14.1-q24.2, 9p21.3-p24,3, 9q21.2, 18p11.31-p11.32 losses. A total of 237 and 101 luminal B-specific candidate oncogenes and tumor suppressor genes (TSGs) presented a deregulated expression in relation with their CNAs, including 11 genes previously reported associated with endocrine resistance. Interestingly, 88% of the potential TSGs are located within chromosome arm 6q, and seven candidate oncogenes are potential therapeutic targets. A total of 100 candidate oncogenes were validated in a public series of 5,765 BCs and the overexpression of 67 of these was associated with poor survival in luminal tumors. Twenty-four genes presented a deregulated expression in relation with a high DNA methylation level. FOXO3, PIK3CA and TP53 were the most frequent mutated genes among the nine tested. In a meta-analysis of next-generation sequencing data in 875 BCs, KCNB2 mutations were associated with luminal B cases while candidate TSGs MDN1 (6q15) and UTRN (6q24), were mutated in this subtype. In conclusion, we have reported luminal B candidate genes that may play a role in the development and/or hormone resistance of this aggressive subtype.

Figure 1. Specific regions targeted by CNAs in luminal B BCs.

Genomic profiles were established for 188 breast tumors (32 luminal B and 156 non luminal B). On the left is shown a hierarchical clustering of genome copy number profiles measured by aCGH on 24,907 probes or groups of probes (without X and Y). Red indicates increased copy number and green indicates decreased copy number. To the left are indicated chromosome locations with chromosome 1pter to the top and 22qter to the bottom. Next on the right, significant copy number amplifications (dark red), gains (red) and losses (green) observed in luminal B compared to non-luminal B tumors (Fisher's exact test), are plotted as a function of chromosome location. Only amplification, gains, and losses associated with luminal B tumors are shown (Fisher's exact test; FDR<0.05) for each chromosome. In addition to the previously reported 8p11-12, 11q13 amplifications, 17q, 20q gains and 18p losses we previously reported (16,20), other luminal B CNAs include 7q11.22, 7q34, 8p11.21-p12, 8q21.12-q24.23, 11q13.3-q14.1, 12p12.3-p13.1, 12q13.11-q24.11, 14q21.1-q23.1, 17q11.1-q25.1, 20q11.23-q13.33 gains and 6q14.1-q24.2, 9p21.3-p24.3, 9q21.2, 18p11.31-p11.32 losses.

Figure 2. Correlation between gene expression and genome alterations on the 6q regions.

Genomic and gene expression profiles were established for 188 breast tumors (32 luminal B and 156 non luminal B identified at the top by blue and grey boxes, respectively) in three 6q regions: 6q14.1-q22.31 (top), 6q22.31-q23.1 (middle) and 6q23.2-q24.2 (bottom). For each region, heatmaps for genome copy number and gene expression profiles are consecutively drawn. Genome copy number was measured by aCGH on probes or groups of probes spanning each of these regions. Red indicates increased copy number and green indicates decreased copy number. In the heatmap tumors are organized from the tumor that presented the most copy number losses to the tumor that exhibited the most copy number gains. The next heatmap was established with the expression of the independent genes located on the corresponding 6q region and profiled in the same 188 tumors similarly organized. For gene copy number and gene expression heatmaps, we used color scale limits from −3 to +3 and −2 to +2, respectively. Next to the right, are plotted genes successively selected by steps I, II and III of the integrated analysis “aCGH & mRNA expression” as defined by the work pipeline (Figure S2). Grey and green lines correspond to rejected and selected genes, respectively. Among genes with an expression level that varied according to CNAs, we retained genes showed significant differences (vertical line) in copy number loss correlated with downregulated expression in luminal B compared to non-luminal B tumors. They were qualified as potential TSGs. For each region, only the first five most significant are listed. PNRC1, NCOA7 and TNFAIP3 genes were the most significant candidate TSGs for the 6q14.1-q22.31 (top), 6q22.31-q23.1 (middle) and 6q23.2-q24.2 (bottom) regions, respectively.

Figure 3. Kaplan-Meier MFS curves and pathological clinical response in luminal BCs according to RECQL4 mRNA expression.

A-Kaplan–Meier survival curves were drawn according to RECQL4 gene expression status of 1,016 luminal BCs established from a large public series of 5,765 BCs (Table S1B). Stratification into high-risk (red curve) and low-risk (black curve) groups were based on relative risk defined by the Cox model using the natural threshold of 1. RECQL4 gene expression is associated with MFS (p = 3.8010⁻⁹). B - From the same large public series, 924 BCs were informed for the pathological response (Table S1B). The RECQL4 overexpression was associated with a better pathological response within the group including both luminal A and B tumors (N = 435) (Fisher, p = 0.013).

Figure 4. DNA Methylation promoter profiles in breast cancers.

A- Hierarchical clustering established with the most variant methylation scores observed in 5,492 gene promoters (SD>0.3) in 109 tumors samples and 8 normal breast tissues samples. Each row of the data matrix represents a gene promoter and each column represents a sample. DNA methylation variations are depicted according to the color scale shown at the bottom. Red indicates increased DNA methylation score and green indicates decreased DNA methylation score. The dendrogram of samples (above matrixes) represents overall similarities in DNA methylation profiles and is zoomed to the right. Three groups of tumor samples (I, IIa and IIb) are associated with various DNA methylation patterns and delimited by orange vertical lines. Below the dendrogram are some histoclinical and molecular features of the samples: from top to bottom, intrinsic molecular subtypes⁴, IHC ER status and SBR grade. Color legends for the various features are illustrated below. B - The median methylation levels of the 4,545 subtype-associated genes were highest in the luminal and ERRB2 subtypes and lowest in the basal subtype (p = 6.24 10⁻¹²). C - Compared to the other molecular subtypes, the DNA methylation levels (three top panels) of ASS1, C6ORF145 and ZFP36L2 gene promoters and their mRNA expression (three bottom panels) were higher (Table S4E) and lower in the luminal B BCs, respectively.

Figure 5. Landscape of specific regions/genes in the 188 breast tumors.

The most significant amplifications/gains, losses and mutations in four major subtypes are shown.