Limits...
Transcriptome analysis of mammary epithelial subpopulations identifies novel determinants of lineage commitment and cell fate.

Kendrick H, Regan JL, Magnay FA, Grigoriadis A, Mitsopoulos C, Zvelebil M, Smalley MJ - BMC Genomics (2008)

Bottom Line: One of these, Sox6, was specifically expressed in luminal estrogen receptor negative cells and functional assays confirmed that it maintained mammary epithelial cells in a differentiated luminal cell lineage.The mouse mammary epithelium is composed of three main cell types with distinct gene expression patterns.These suggest the existence of a novel functional cell type within the gland, that the basal/myoepithelial cells are key regulators of paracrine signalling and that there is a complex network of differentially expressed transcription factors controlling mammary epithelial cell fate.

View Article: PubMed Central - HTML - PubMed

Affiliation: Breakthrough Breast Cancer Research Centre, The Institute of Cancer Research, London, UK. howard.kendrick@icr.ac.uk

ABSTRACT

Background: Understanding the molecular control of cell lineages and fate determination in complex tissues is key to not only understanding the developmental biology and cellular homeostasis of such tissues but also for our understanding and interpretation of the molecular pathology of diseases such as cancer. The prerequisite for such an understanding is detailed knowledge of the cell types that make up such tissues, including their comprehensive molecular characterisation. In the mammary epithelium, the bulk of the tissue is composed of three cell lineages, namely the basal/myoepithelial, luminal epithelial estrogen receptor positive and luminal epithelial estrogen receptor negative cells. However, a detailed molecular characterisation of the transcriptomic differences between these three populations has not been carried out.

Results: A whole transcriptome analysis of basal/myoepithelial cells, luminal estrogen receptor negative cells and luminal estrogen receptor positive cells isolated from the virgin mouse mammary epithelium identified 861, 326 and 488 genes as highly differentially expressed in the three cell types, respectively. Network analysis of the transcriptomic data identified a subpopulation of luminal estrogen receptor negative cells with a novel potential role as non-professional immune cells. Analysis of the data for potential paracrine interacting factors showed that the basal/myoepithelial cells, remarkably, expressed over twice as many ligands and cell surface receptors as the other two populations combined. A number of transcriptional regulators were also identified that were differentially expressed between the cell lineages. One of these, Sox6, was specifically expressed in luminal estrogen receptor negative cells and functional assays confirmed that it maintained mammary epithelial cells in a differentiated luminal cell lineage.

Conclusion: The mouse mammary epithelium is composed of three main cell types with distinct gene expression patterns. These suggest the existence of a novel functional cell type within the gland, that the basal/myoepithelial cells are key regulators of paracrine signalling and that there is a complex network of differentially expressed transcription factors controlling mammary epithelial cell fate. These data will form the basis for understanding not only cell fate determination and cellular homeostasis in the normal mammary epithelium but also the contribution of different mammary epithelial cell types to the etiology and molecular pathology of breast disease.

Show MeSH

Related in: MedlinePlus

Partial heat map and cluster analysis of differential gene expression across virgin mammary epithelial populations. Clusters representative of the different gene expression patterns are shown. Red indicates high expression, green indicates low expression [see Additional file 3].
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2629782&req=5

Figure 2: Partial heat map and cluster analysis of differential gene expression across virgin mammary epithelial populations. Clusters representative of the different gene expression patterns are shown. Red indicates high expression, green indicates low expression [see Additional file 3].

Mentions: To carry out a comprehensive, whole genome gene expression analysis of the epithelial cell populations in the virgin mammary gland (Figure 1A and 1B), primary mouse mammary cells were isolated and stained with anti-CD24 and anti-Sca-1 antibodies. The CD24+/Low Sca-1-, CD24+/High Sca-1- and CD24+/High Sca-1+ cells were separated by flow cytometry (Figure 1C) and used to prepare RNA. To demonstrate that these populations corresponded to basal/myoepithelial, luminal ER- and luminal ER+ cells respectively, as we have previously described [5], relative gene expression levels of the basal cell marker Keratin 14 (Krt14), the luminal cell marker Keratin 18 (Krt18) and Estrogen Receptor α (Esr1) were measured by quantitative real-time rtPCR (qPCR). The results showed that the CD24+/Low Sca-1- cells expressed Krt14 but not Krt18 and that the CD24+/High Sca-1- and CD24+/High Sca-1+ cells expressed Krt18 but not Krt14. They also showed that the CD24+/High Sca-1+ population was enriched for Esr1 expression compared to the other two populations. This agreed with previous data [5] and with staining of sections through the mouse mammary gland (Figure 1A and 1B) which showed that the basal cells were Keratin 14+ (K14+), the luminal cells were K18+ and that ER+ cells were found exclusively in the K18+ luminal cell layer. Therefore, these data confirmed the identity of the three populations [see Additional file 1] [5]. Next, a whole transcriptome gene expression analysis of the three populations was carried out using the Affymetrix platform (Mouse Expression Arrays MOE430 2.0). To identify genes whose expression in one population was different from its expression in the other two, Significance Analysis of Microarray (SAM) was applied [30,31]. Using this technique in a multiclass setting, the average gene expression within one group was compared with the average gene expression of all samples. Following repeated permutation of the data, the strength of the relationship between gene expression and its expressed group was established and a significance level determined. When a FDR cutoff of 0.0% was applied, 2182 probe sets [see Additional file 2] were identified which showed significant differential expression across the three populations. These 2182 probes corresponded to 1427 well known individual genes. Heat mapping and hierarchical cluster analysis of the genes identified 14 different gene clusters. The cluster analysis showed that some gene sets were characteristic of only one population (e.g. clusters 1, 10 and 11) whereas others were expressed in two of the three (e.g. cluster 14) (Figure 2) [see Additional file 3].


Transcriptome analysis of mammary epithelial subpopulations identifies novel determinants of lineage commitment and cell fate.

Kendrick H, Regan JL, Magnay FA, Grigoriadis A, Mitsopoulos C, Zvelebil M, Smalley MJ - BMC Genomics (2008)

Partial heat map and cluster analysis of differential gene expression across virgin mammary epithelial populations. Clusters representative of the different gene expression patterns are shown. Red indicates high expression, green indicates low expression [see Additional file 3].
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2629782&req=5

Figure 2: Partial heat map and cluster analysis of differential gene expression across virgin mammary epithelial populations. Clusters representative of the different gene expression patterns are shown. Red indicates high expression, green indicates low expression [see Additional file 3].
Mentions: To carry out a comprehensive, whole genome gene expression analysis of the epithelial cell populations in the virgin mammary gland (Figure 1A and 1B), primary mouse mammary cells were isolated and stained with anti-CD24 and anti-Sca-1 antibodies. The CD24+/Low Sca-1-, CD24+/High Sca-1- and CD24+/High Sca-1+ cells were separated by flow cytometry (Figure 1C) and used to prepare RNA. To demonstrate that these populations corresponded to basal/myoepithelial, luminal ER- and luminal ER+ cells respectively, as we have previously described [5], relative gene expression levels of the basal cell marker Keratin 14 (Krt14), the luminal cell marker Keratin 18 (Krt18) and Estrogen Receptor α (Esr1) were measured by quantitative real-time rtPCR (qPCR). The results showed that the CD24+/Low Sca-1- cells expressed Krt14 but not Krt18 and that the CD24+/High Sca-1- and CD24+/High Sca-1+ cells expressed Krt18 but not Krt14. They also showed that the CD24+/High Sca-1+ population was enriched for Esr1 expression compared to the other two populations. This agreed with previous data [5] and with staining of sections through the mouse mammary gland (Figure 1A and 1B) which showed that the basal cells were Keratin 14+ (K14+), the luminal cells were K18+ and that ER+ cells were found exclusively in the K18+ luminal cell layer. Therefore, these data confirmed the identity of the three populations [see Additional file 1] [5]. Next, a whole transcriptome gene expression analysis of the three populations was carried out using the Affymetrix platform (Mouse Expression Arrays MOE430 2.0). To identify genes whose expression in one population was different from its expression in the other two, Significance Analysis of Microarray (SAM) was applied [30,31]. Using this technique in a multiclass setting, the average gene expression within one group was compared with the average gene expression of all samples. Following repeated permutation of the data, the strength of the relationship between gene expression and its expressed group was established and a significance level determined. When a FDR cutoff of 0.0% was applied, 2182 probe sets [see Additional file 2] were identified which showed significant differential expression across the three populations. These 2182 probes corresponded to 1427 well known individual genes. Heat mapping and hierarchical cluster analysis of the genes identified 14 different gene clusters. The cluster analysis showed that some gene sets were characteristic of only one population (e.g. clusters 1, 10 and 11) whereas others were expressed in two of the three (e.g. cluster 14) (Figure 2) [see Additional file 3].

Bottom Line: One of these, Sox6, was specifically expressed in luminal estrogen receptor negative cells and functional assays confirmed that it maintained mammary epithelial cells in a differentiated luminal cell lineage.The mouse mammary epithelium is composed of three main cell types with distinct gene expression patterns.These suggest the existence of a novel functional cell type within the gland, that the basal/myoepithelial cells are key regulators of paracrine signalling and that there is a complex network of differentially expressed transcription factors controlling mammary epithelial cell fate.

View Article: PubMed Central - HTML - PubMed

Affiliation: Breakthrough Breast Cancer Research Centre, The Institute of Cancer Research, London, UK. howard.kendrick@icr.ac.uk

ABSTRACT

Background: Understanding the molecular control of cell lineages and fate determination in complex tissues is key to not only understanding the developmental biology and cellular homeostasis of such tissues but also for our understanding and interpretation of the molecular pathology of diseases such as cancer. The prerequisite for such an understanding is detailed knowledge of the cell types that make up such tissues, including their comprehensive molecular characterisation. In the mammary epithelium, the bulk of the tissue is composed of three cell lineages, namely the basal/myoepithelial, luminal epithelial estrogen receptor positive and luminal epithelial estrogen receptor negative cells. However, a detailed molecular characterisation of the transcriptomic differences between these three populations has not been carried out.

Results: A whole transcriptome analysis of basal/myoepithelial cells, luminal estrogen receptor negative cells and luminal estrogen receptor positive cells isolated from the virgin mouse mammary epithelium identified 861, 326 and 488 genes as highly differentially expressed in the three cell types, respectively. Network analysis of the transcriptomic data identified a subpopulation of luminal estrogen receptor negative cells with a novel potential role as non-professional immune cells. Analysis of the data for potential paracrine interacting factors showed that the basal/myoepithelial cells, remarkably, expressed over twice as many ligands and cell surface receptors as the other two populations combined. A number of transcriptional regulators were also identified that were differentially expressed between the cell lineages. One of these, Sox6, was specifically expressed in luminal estrogen receptor negative cells and functional assays confirmed that it maintained mammary epithelial cells in a differentiated luminal cell lineage.

Conclusion: The mouse mammary epithelium is composed of three main cell types with distinct gene expression patterns. These suggest the existence of a novel functional cell type within the gland, that the basal/myoepithelial cells are key regulators of paracrine signalling and that there is a complex network of differentially expressed transcription factors controlling mammary epithelial cell fate. These data will form the basis for understanding not only cell fate determination and cellular homeostasis in the normal mammary epithelium but also the contribution of different mammary epithelial cell types to the etiology and molecular pathology of breast disease.

Show MeSH
Related in: MedlinePlus