Molecular profiling of human mammary gland links breast cancer risk to a p27(+) cell population with progenitor characteristics

Abstract

Early full-term pregnancy is one of the most effective natural protections against breast cancer. To investigate this effect, we have characterized the global gene expression and epigenetic profiles of multiple cell types from normal breast tissue of nulliparous and parous women and carriers of BRCA1 or BRCA2 mutations. We found significant differences in CD44(+) progenitor cells, where the levels of many stem cell-related genes and pathways, including the cell-cycle regulator p27, are lower in parous women without BRCA1/BRCA2 mutations. We also noted a significant reduction in the frequency of CD44(+)p27(+) cells in parous women and showed, using explant cultures, that parity-related signaling pathways play a role in regulating the number of p27(+) cells and their proliferation. Our results suggest that pathways controlling p27(+) mammary epithelial cells and the numbers of these cells relate to breast cancer risk and can be explored for cancer risk assessment and prevention.

Figure 1. Cell type-specific differences in gene expression according to parity and BRCA1/2 mutation status

(A) Genome-wide view of genes differentially expressed between nulliparous (N) and parous (P) samples in the four cell types analyzed. Each dot represents a gene. Fold differences between averaged N and P samples and their corresponding p-values are plotted on the y and x-axis, respectively. Green vertical lines and numbers indicate p=0.05 and genes differentially expressed at p<0.05, respectively. (B) Three-dimensional projection of the gene expression data onto the first three principal components. Each ball is a different sample; cell type and parity are indicated. (C) Bar plot of the paired Euclidean distance for each of the four cell types. P-value indicates the significance of difference (Kolmogorov-Smirnov test) between parous and nulliparous groups in CD44⁺ and other cell types. (D) Hierarchical clustering of Norwegian cohort based on Pearson correlation using genes differentially expressed in CD44⁺ cells. (E) Serum estradiol levels for the samples corresponding to Panel D. (F) Hierarchical clustering of CD44⁺ cells from nulliparous and parous control women and parous BRCA1 and BRCA2 mutation carriers with the exception of one BRCA2 sample (N152, highlighted in orange) that was nulliparous. (G) Relative frequency of CD44⁺, CD24⁺, and CD10⁺ cells. 10 samples were analyzed from each of the indicated. Each dot represents an individual sample. Error bars represent mean ± SEM. See also Figures S1–2 and Tables S1–3.

Figure 2. Signaling pathways affected by parity-related differences in gene expression patterns

(A) Dendrogram depicting hierarchical clustering of signaling pathways significantly high in parous or nulliparous samples in any of the four cell types analyzed. (B) Heatmap depicting unsupervised clustering of signaling pathways significantly down or upregulated in parous compared to nulliparous samples in any of the four cell types analyzed. Color scale indicates –log p value of enrichment. Orange rectangles highlight cell type-specific or common altered pathways. (C) Genes differentially expressed between nulliparous and parous samples in each of the four cells types were analyzed for relative enrichment with the indicated protein classes (lower panel) and for relative connectivity (upper panel). y-axes indicate –log p-values for enrichment with the listed protein classes or the number of overconnected objects. (D) Venn diagram depicting the number of unique and common pathways high in CD44⁺ cells from nulliparous women and in mammary glands of virgin rats, respectively. (E) List of top common pathways downregulated in CD44⁺ cells and mammary glands from parous women and rats, respectively. Name of pathways and p-values of enrichment are indicated. See also Tables S4–5.

Figure 3. Epigenetic differences between nulliparous and parous tissues

(A) Genome-wide view of differentially methylated genes in CD24⁺ and CD44⁺ cells between nulliparous and parous samples. All MSDK sites are plotted on the x-axis in the order of p-values of the difference between nulliparous and parous samples in CD44⁺ or CD24⁺ cells. Log ratios of averaged MSDK counts in three N and three P samples are plotted on the y-axis. Green vertical lines indicate p=0.01 and the numbers of significant DMRs (p<0.01) are shown. (B) Pathways enriched with genes in CD44⁺ cells with the indicated difference in DNA methylation between nulliparous and parous women. (C) Genes with promoter and gene body DMRs in CD44⁺ cells from nulliparous and parous samples were analyzed for relative enrichment with the indicated protein classes and for relative connectivity. y-axes indicate –log p-values for enrichment with the listed protein classes or the number of overconnected objects. (D) Pie charts depicting the relative % of genes in different functional categories with the indicated gene expression and DNA methylation pattern in CD44⁺ cells from nulliparous and parous women. (E) Scatter plot for MSDK-seq and SAGE-seq data to depict correlations between differential promoter methylation and differential gene expression for transcription factors. Each point represents a gene with a MSDK-seq site in certain region (promoter: −5kb to +2kb from TSS, gene body: +2kb from TSS to the end of gene), and log10 p-value is plotted for difference of DNA methylation (x-axis) and expression (y-axis) between parous and nulliparous samples. If a MSDK site is hypermethylated or a gene is higher expressed in parous, −1 is multiplied to log10 p-value, providing positive values. MSDK-seq sites that are significantly (p<0.05) hypo- or hypermethylated in parous or nulliparous samples are highlighted in blue. See also Figure S3 and Tables S5–7.

Figure 4. Expression of p27 in normal breast tissue samples

Representative examples of multicolor immunofluorescence analyses of normal mammary epithelium. p-values of differences between nulliparous and parous groups are indicated. Error bars represent median ± SEM. (A) Expression of p27, CD24, and CD44 in breast tissue of premenopausal nulliparous (NP) and parous (P) women. Graphs show the quantification of p27 staining intensity in multiple samples. Immunofluorescence staining for p27 and Ki67 in breast tissue from premenopausal (B) and postmenopausal (C) nulliparous and parous women. Graph shows frequencies of p27⁺ and Ki67⁺ cells in nulliparous (NP) and parous (P) samples.

Figure 5. Hormonal factors and the expression of p27 in normal breast tissues

Asterisks indicate significant (p≤0.05, t-test or Fisher exact test) differences between groups of 4–8 samples. Error bars represent mean ± SD. (A) Representative double immunofluorescence staining for p27 and ER in breast tissue from the indicated groups of women. Graph shows frequencies of p27⁺, ER⁺, and p27⁺ER⁺ cells in each group of samples. (B) Representative double immunofluorescence staining for p27 and AR in breast tissue from premenopausal nulliparous and parous women, and in BRCA1 mutation carriers. Graphs show frequencies of p27⁺ and AR⁺ cells in each set of samples. (C) Representative double immunofluorescence staining for p27 and Ki67 in breast tissue from the indicated groups of women. Graphs show frequencies of p27⁺, Ki67⁺, and p27⁺Ki67⁺ cells in each group of samples. See also Figure S4.

Figure 6. Modulation of p27⁺ breast epithelial cells and proliferation by hormonal and parity-related pathways

Asterisks indicate significant (p≤0.05) differences. Error bars represent mean ± SD. (A) Representative hematoxilin and eosin staining depicting morphology of breast tissue after 8 days in culture. (B) Representative examples of multicolor immunofluorescence analyses of BrdU⁺, p27⁺, and Ki67⁺ cells in control and tissues treated with inhibitors of the indicated pathways. (C) Frequency of Ki67⁺, BrdU⁺, and p27⁺ cells in each of the indicated conditions. (D–E) Representative images of immunofluorescence analysis of p27 and graph depicting the frequency of p27⁺ cells in tissue slices from 3–4 independent cases treated with hormones mimicking the indicated physiologic levels in women. See also Figure S4.

Figure 7. Signaling pathways regulating p27⁺ breast epithelial cells

Asterisks indicate significant (p≤0.05) differences. Error bars represent mean ± SD. (A) Representative examples of multicolor immunofluorescence analyses of pSMAD2, pEGFR, Axin2+, and p27 cells in control and tissues treated with inhibitors of the indicated pathways. (B) Quantitation of differences in the expression of markers reflecting pathway activity between control and inhibitor-treated tissues (inhibitor). (C) RGB spectra demonstrating overlap between the expression of p27 and the indicated marker. (D) Double immunofluorescence staining for p27 and pSmad2 in breast tissues from 3–4 independent cases of the indicated women. Graph shows frequencies of p27⁺, pSmad2⁺, and p27⁺pSmad2⁺ cells in each group of samples.