Research Papers:

Reverse phase protein array identification of triple-negative breast cancer subtypes and comparison with mRNA molecular subtypes

Hiroko Masuda, Yuan Qi, Shuying Liu, Naoki Hayashi, Takahiro Kogawa, Gabriel N. Hortobagyi, Debu Tripathy and Naoto T. Ueno _

PDF  |  HTML  |  Supplementary Files  |  How to cite

Oncotarget. 2017; 8:70481-70495. https://doi.org/10.18632/oncotarget.19719

Metrics: PDF 1665 views  |   HTML 3198 views  |   ?  


Hiroko Masuda1,2,4, Yuan Qi3, Shuying Liu1, Naoki Hayashi1, Takahiro Kogawa1,2, Gabriel N. Hortobagyi1, Debu Tripathy1 and Naoto T. Ueno1,2

1Department of Breast Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA

2Morgan Welch Inflammatory Breast Cancer Research Program and Clinic, The University of Texas MD Anderson Cancer Center, Houston, TX, USA

3Department of Bioinformatics and Computational Biology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA

4Department of Breast Surgical Oncology, Showa University Hospital, Tokyo, Japan

Correspondence to:

Naoto T. Ueno, email: [email protected]

Keywords: triple-negative breast cancer, molecular subtype, functional, proteomics, mRNA microarray

Received: October 19, 2016    Accepted: June 27, 2017    Published: July 31, 2017


Background: Reverse phase protein array (RPPA) analysis, allows investigation of potential targets at the functional protein level,. We identified TNBC subtypes at the protein level using RPPA and compared them with mRNA molecular subtypes (TNBCtype, TNBCtype-4, and PAM50) that is unique in its availability of both RPPA and mRNA analyses.

Methods: We classified the samples from 80 TNBC patients using both k-means and hierarchical consensus clustering analysis and performed Ingenuity Pathway Analysis. We also investigated whether we could reproduce the mRNA molecular subtypes using the RPPA dataset.

Results: Both clustering methods divided all samples into 2 clusters that were biologically the same. The top canonical pathways included inflammation, hormonal receptors, and MAPK signaling pathways for the first cluster [“inflammation and hormonal-related (I/H) subtype”] and the GADD45, DNA damage, and p53 signaling pathways for the second cluster [“DNA damage (DD)-related subtype”]. Further k-means cluster analysis identified 5 TNBC clusters. Comparison between sample classification using the 5 RPPA-based k-means cluster subtypes and 6 gene-expression-based TNBCtype molecular subtypes showed significant association between the 2 classifications (p = 0.017).

Conclusions: The I/H and DD subtypes identified by RPPA advance our understanding of TNBC’s heterogeneity from the functional proteomic perspective.

Creative Commons License All site content, except where otherwise noted, is licensed under a Creative Commons Attribution 4.0 License.
PII: 19719