Limits...
Genetic background may contribute to PAM50 gene expression breast cancer subtype assignments.

Hu Y, Bai L, Geiger T, Goldberger N, Walker RC, Green JE, Wakefield LM, Hunter KW - PLoS ONE (2013)

Bottom Line: Recent advances in genome wide transcriptional analysis have provided greater insights into the etiology and heterogeneity of breast cancer.It is thought that the expression patterns of the molecular subtypes primarily reflect cell-of-origin or tumor driver mutations.These results have important implications for interpretation of "snapshot" expression profiles, as well as suggesting that incorporation of genetic background effects may allow investigation into phenotypes not initially anticipated in individual mouse models of cancer.

View Article: PubMed Central - PubMed

Affiliation: Center for Biomedical Informatics and Information Technology, Bethesda, Maryland, United States of America.

ABSTRACT
Recent advances in genome wide transcriptional analysis have provided greater insights into the etiology and heterogeneity of breast cancer. Molecular signatures have been developed that stratify the conventional estrogen receptor positive or negative categories into subtypes that are associated with differing clinical outcomes. It is thought that the expression patterns of the molecular subtypes primarily reflect cell-of-origin or tumor driver mutations. In this study however, using a genetically engineered mouse mammary tumor model we demonstrate that the PAM50 subtype signature of tumors driven by a common oncogenic event can be significantly influenced by the genetic background on which the tumor arises. These results have important implications for interpretation of "snapshot" expression profiles, as well as suggesting that incorporation of genetic background effects may allow investigation into phenotypes not initially anticipated in individual mouse models of cancer.

Show MeSH

Related in: MedlinePlus

Unsupervised clustering of the GSE2034, TCGA, MOLF, DO and NZB PAM50 genes.The position of the mouse samples and the subtypes of the human samples are indicated across the top of the heatmap. The color coding of the clustering across the top of the figure is based on the enrichment of human samples within each major block.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3756056&req=5

pone-0072287-g002: Unsupervised clustering of the GSE2034, TCGA, MOLF, DO and NZB PAM50 genes.The position of the mouse samples and the subtypes of the human samples are indicated across the top of the heatmap. The color coding of the clustering across the top of the figure is based on the enrichment of human samples within each major block.

Mentions: Further investigation was performed by expanding both the mouse and human data sets. In addition to the small NZB backcross two additional mouse expression data sets were generated. 134 tumors from the previously described MOLF mouse backcross [21] and 133 tumors from a cross between Diversity Outbred (DO) [14], [23] and PyMT were arrayed and included in the analysis. These crosses represent increasing genetic diversity (NZB<MOLF<DO) due to the presence of wild mouse-derived polymorphisms in the MOLF and DO crosses, with the DO cross approximating the degree of polymorphism observed in humans. In addition the 286 human tumor samples from the GSE2034 data set were supplemented by the addition of the TCGA (The Cancer Genome Atlas; N = 466) expression data. Expression data from each of the array platforms (MOE430 v2, Mouse Gene 1.0×ST, U133A, Illumina) were normalized, filtered for the genes comprising the PAM50 subtype signature and unsupervised hierarchical clustering performed. As can be observed in figure 2, mouse tumors from all three data sets distributed across all of the human subtypes, consistent with the possibility that tumor subtype classification by PAM50 signature might be significantly influenced by inherited polymorphism.


Genetic background may contribute to PAM50 gene expression breast cancer subtype assignments.

Hu Y, Bai L, Geiger T, Goldberger N, Walker RC, Green JE, Wakefield LM, Hunter KW - PLoS ONE (2013)

Unsupervised clustering of the GSE2034, TCGA, MOLF, DO and NZB PAM50 genes.The position of the mouse samples and the subtypes of the human samples are indicated across the top of the heatmap. The color coding of the clustering across the top of the figure is based on the enrichment of human samples within each major block.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3756056&req=5

pone-0072287-g002: Unsupervised clustering of the GSE2034, TCGA, MOLF, DO and NZB PAM50 genes.The position of the mouse samples and the subtypes of the human samples are indicated across the top of the heatmap. The color coding of the clustering across the top of the figure is based on the enrichment of human samples within each major block.
Mentions: Further investigation was performed by expanding both the mouse and human data sets. In addition to the small NZB backcross two additional mouse expression data sets were generated. 134 tumors from the previously described MOLF mouse backcross [21] and 133 tumors from a cross between Diversity Outbred (DO) [14], [23] and PyMT were arrayed and included in the analysis. These crosses represent increasing genetic diversity (NZB<MOLF<DO) due to the presence of wild mouse-derived polymorphisms in the MOLF and DO crosses, with the DO cross approximating the degree of polymorphism observed in humans. In addition the 286 human tumor samples from the GSE2034 data set were supplemented by the addition of the TCGA (The Cancer Genome Atlas; N = 466) expression data. Expression data from each of the array platforms (MOE430 v2, Mouse Gene 1.0×ST, U133A, Illumina) were normalized, filtered for the genes comprising the PAM50 subtype signature and unsupervised hierarchical clustering performed. As can be observed in figure 2, mouse tumors from all three data sets distributed across all of the human subtypes, consistent with the possibility that tumor subtype classification by PAM50 signature might be significantly influenced by inherited polymorphism.

Bottom Line: Recent advances in genome wide transcriptional analysis have provided greater insights into the etiology and heterogeneity of breast cancer.It is thought that the expression patterns of the molecular subtypes primarily reflect cell-of-origin or tumor driver mutations.These results have important implications for interpretation of "snapshot" expression profiles, as well as suggesting that incorporation of genetic background effects may allow investigation into phenotypes not initially anticipated in individual mouse models of cancer.

View Article: PubMed Central - PubMed

Affiliation: Center for Biomedical Informatics and Information Technology, Bethesda, Maryland, United States of America.

ABSTRACT
Recent advances in genome wide transcriptional analysis have provided greater insights into the etiology and heterogeneity of breast cancer. Molecular signatures have been developed that stratify the conventional estrogen receptor positive or negative categories into subtypes that are associated with differing clinical outcomes. It is thought that the expression patterns of the molecular subtypes primarily reflect cell-of-origin or tumor driver mutations. In this study however, using a genetically engineered mouse mammary tumor model we demonstrate that the PAM50 subtype signature of tumors driven by a common oncogenic event can be significantly influenced by the genetic background on which the tumor arises. These results have important implications for interpretation of "snapshot" expression profiles, as well as suggesting that incorporation of genetic background effects may allow investigation into phenotypes not initially anticipated in individual mouse models of cancer.

Show MeSH
Related in: MedlinePlus