A collection of breast cancer cell lines for the study of functionally distinct cancer subtypes

Abstract

Recent studies suggest that thousands of genes may contribute to breast cancer pathophysiologies when deregulated by genomic or epigenomic events. Here, we describe a model "system" to appraise the functional contributions of these genes to breast cancer subsets. In general, the recurrent genomic and transcriptional characteristics of 51 breast cancer cell lines mirror those of 145 primary breast tumors, although some significant differences are documented. The cell lines that comprise the system also exhibit the substantial genomic, transcriptional, and biological heterogeneity found in primary tumors. We show, using Trastuzumab (Herceptin) monotherapy as an example, that the system can be used to identify molecular features that predict or indicate response to targeted therapies or other physiological perturbations.

Figure 1. Comparison of array CGH analyses of human breast cancer cell lines and primary tumors

A and B: Frequencies of significant increases or decreases in genome copy number are plotted as a function of genome location for 51 cell lines (A) and 145 primary tumors (Fridlyand et al., 2006) (B). Positive values indicate frequencies of samples showing copy number increases [Log2(copy number) > 0.3], and negative values indicate frequencies of samples showing copy number decreases [Log2(copy number) < −0.3]. C and D: Differences (y axis) between frequencies of gains and losses across the genome for the cell lines versus the tumors are represented in C and D, respectively. Genome copy number aberration frequencies are plotted as a function of location position in the genome beginning at 1pter to the left and ending at Xqter to the right (chromosome locations are indicated by numbers above and below graphs). Vertical lines indicate chromosome boundaries. Vertical dotted lines indicate centromere locations.

Figure 2. Unsupervised hierarchical clustering of genome aberrations in 51 breast cancer cell lines and 145 primary breast tumors

Clusters show results at 1952 BAC probes common between the tumor and cell line CGH arrays. Each row represents a BAC probe, and each column represents a cell line or tumor sample. Green indicates increased genome copy number, and red indicates decreased genome copy number. Yellow indicates high-level amplification. The bar to the left shows chromosome locations with chromosome 1pter to the top and 22qter to the bottom. The locations of the odd-numbered chromosomes (shaded black) are indicated. The upper color bar shows columns representing tumors or cell lines. The lower color bar indicates the genomic characteristics of the cell lines and tumors from this study and for the tumors as reported (Fridlyand et al., 2006). Color codes are indicated at the bottom of the figure.

Figure 3. Gene expression profiles of 51 human breast cancer cell lines

A: Hierarchical cluster analysis of breast cancer cell lines with subclusters [(i) through (v)] indicated by colored bars at side. Genes were restricted to those showing significant variance across all samples, resulting in clustering of 1438 probe sets (see Experimental Procedures). B: Cell lines clustered by genes selected by PAM analysis representing the luminal, Basal A, and Basal B clusters. Clustering was performed as described in the Experimental Procedures. Each row represents a gene, and each column represents a cell line sample. As shown in the color bar, black represents no change, red represents upregulation, and green represents downregulation of gene expression.

Figure 4. Comparative analyses of aberration frequencies in basal and luminal primary tumors and cell lines

A–C: A and B show frequencies of genome copy number gains and losses in luminal and basal breast tumors, respectively (Fridlyand et al., 2006). C shows univariate statistical assessments of differences between the two tumor types. D–F: D and E show frequencies of genome copy number gains and losses in luminal and Basal B breast cancer cell lines, respectively. F shows univariate statistical assessments of differences between the two cell line types. p values of 0.05 and 0.01 are indicated as in C. All data are plotted beginning at chromosome 1pter to the left and ending at Xqter to the right. Vertical solid lines indicate chromosome boundaries. Vertical dashed lines indicate centromere locations. p values of 0.05 and 0.01 are indicated by dashed horizontal lines in C and F.

Figure 5. A functional model to investigate lead candidate therapeutic targets

Gene expression and copy number for the 66 candidate therapeutic genes in the 51 breast cancer cell lines. Genes were selected for their overexpression and association with outcome in human breast tumors (Chin et al., 2006). High gene expression (≥2-fold over mean gene expression for all the samples) is shown in red, and high-level amplification (≥0.9 Log2 ratio) is shown in yellow for each cell line. Genes that have high expression and gene copy number are shown in blue. Gene HUGO names are shown to the right, and the corresponding BAC clone ID and chromosome are shown to the left.

Figure 6. Relationship of transcriptional profiles to biological function

A: Morphology of cell lines grown in tissue culture on plastic. B: Invasive potential of 30 breast cancer cell lines as measured by modified Boyden chamber assays (see Experimental Procedures). Each data point represents the mean ± SD of three wells.

Figure 7. Indicators of therapeutic response to Trastuzumab in HER2-positive breast cancer cell lines

A: CGH profiles of nine breast cancer cell lines overexpressing HER2. B: Response of HER2-overexpressing cell lines to 48 hr treatment with 21 μg/ml Trastuzumab (Herceptin) as measured by BrdUrd incorporation (top panel) and relocalization of p27^KIP1 (lower panel).