Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Comparative Study
. 2011 Jun 8;11:230.
doi: 10.1186/1471-2407-11-230.

Protein Expression Based Multimarker Analysis of Breast Cancer Samples

Affiliations
Free PMC article
Comparative Study

Protein Expression Based Multimarker Analysis of Breast Cancer Samples

Angela P Presson et al. BMC Cancer. .
Free PMC article

Abstract

Background: Tissue microarray (TMA) data are commonly used to validate the prognostic accuracy of tumor markers. For example, breast cancer TMA data have led to the identification of several promising prognostic markers of survival time. Several studies have shown that TMA data can also be used to cluster patients into clinically distinct groups. Here we use breast cancer TMA data to cluster patients into distinct prognostic groups.

Methods: We apply weighted correlation network analysis (WGCNA) to TMA data consisting of 26 putative tumor biomarkers measured on 82 breast cancer patients. Based on this analysis we identify three groups of patients with low (5.4%), moderate (22%) and high (50%) mortality rates, respectively. We then develop a simple threshold rule using a subset of three markers (p53, Na-KATPase-β1, and TGF β receptor II) that can approximately define these mortality groups. We compare the results of this correlation network analysis with results from a standard Cox regression analysis.

Results: We find that the rule-based grouping variable (referred to as WGCNA*) is an independent predictor of survival time. While WGCNA* is based on protein measurements (TMA data), it validated in two independent Affymetrix microarray gene expression data (which measure mRNA abundance). We find that the WGCNA patient groups differed by 35% from mortality groups defined by a more conventional stepwise Cox regression analysis approach.

Conclusions: We show that correlation network methods, which are primarily used to analyze the relationships between gene products, are also useful for analyzing the relationships between patients and for defining distinct patient groups based on TMA data. We identify a rule based on three tumor markers for predicting breast cancer survival outcomes.

Figures

Figure 1
Figure 1
Overview for conducting a Weighted Correlation Network Analyses (WGCNA) of patient TMA data (Steps 1-4) and follow up analyses (Steps 5-7). Steps 1-4 are numbered to correspond with the WGCNA methods section in the text. After defining WGCNA and WGCNA* patient groups, we compare these results to a more conventional variable selection approach (Steps 5-6). Finally, we validate the WGCNA* and conventional results in independent Affymetrix gene expression data sets (Step 7).
Figure 2
Figure 2
Results of a WGCNA of 82 breast cancer patients and 26 markers. A. Markers were clustered according to their expression levels across patient samples, so that each branch of the tree indicates a patient. The first row of white, grey and black colors below the tree indicates WGCNA patient groups that correspond to clusters of patients that have similar marker expression profiles. The second row consists of WGCNA* groups which is an approximation to WGCNA that relies on only three of the 26 markers. Subsequent rows consist of clinical variable data, where black matches with unfavorable prognostic factors, white is favorable, grey is intermediate, and yellow indicates missing data. Stage was coded as 1-3 with stage 1 colored white (there was one stage 4 patient that we re-coded as stage 3). Grade was coded as 1-3 with grade 1 colored white. Her2+, ER- and PR- were colored black. The presence of lymph node involvement (LNI) and metastasis were colored black. Tumor size was re-coded as quantiles, where tumors smaller than the 25th percentile were colored white, tumors between the 25th-75th percentiles were colored grey, and sizes greater than or equal to the 75th percentile were colored black. B-C. WGCNA patient groups correspond to low, moderate (mod.) and high mortality. D. An approximation to the WGCNA groups "WGCNA*" that uses a subset of three markers (rather than the full marker set) is also highly related to patient survival.
Figure 3
Figure 3
The WGCNA* and COX mortality group definitions. A. Classification trees were used to identify a subset of markers (3 out of 26 total) and their optimal thresholds for approximating the WGCNA groups. Nearly 88% (72 matches out of 82) of the mortality group assignments matched between WGCNA* and WGCNA. The markers and approximate thresholds included: p53 (dichotomized at the 75th percentile), Na-KATPase-β1 (33rd percentile) and TGF β receptor II (66th percentile). High mortality was defined by high p53 and low Na-KATPase-β1. The group with a 17% mortality rate is called "low" because 10 of these 12 patients were assigned to the low mortality group by WGCNA. B. We also conducted a more traditional multimarker analysis by dichotomizing each of the 26 markers at an optimal threshold for survival prediction and then using a step-wise marker selection approach to achieve low, moderate and high mortality "COX" patient groups. This approach defined high mortality as high MED28, and moderate mortality as low MED28 and high Smad4. In both diagrams "cyt" indicates expression in the cytoplasm.
Figure 4
Figure 4
Variable and marker boxplots by WGCNA* mortality group. Kruskal-Wallis p-values are reported for the comparison of each variable and marker to the WGCNA* patient groups, where the WGCNA* patient groups are color coded to indicate low (white), moderate (grey) and high (black) mortality. A. Metastasis, stage, ER+ and death are significantly related to the WGCNA* groups (p < 0.05). B. The top 10 markers related to survival that achieved significance at p < 0.05 in a univariate Cox proportional-hazards model when dichotomized at an optimal cut-point. The boxplots indicate that no variable or marker by itself can define the WGCNA* groups. Abbreviations are as follows, "LNI" stands for Lymph Node Involvement, "cyt" indicates the TMA marker was expressed in the cytoplasm and "nuc" indicates nuclear expression.
Figure 5
Figure 5
Validation of the WGCNA* high mortality group in two independent gene expression data sets (A-B). A. Results for the Miller 2005 data set are shown for the following probe sets ATP1B1: 201242_s_at, TP53: 201746_at, and TGFBR2: 208944_at. The Pawitan 2005 data set validated for all probe set combinations, but results for ATP1B1: 201242_s_at, TP53: 211300_s_at, and TGFBR2: 207334_s_at are shown in B. Data set information can be found in Additional File 2, and additional validation results can be found in Table 6.

Similar articles

See all similar articles

Cited by 17 articles

See all "Cited by" articles

References

    1. Kreike B, van Kouwenhove M, Horlings H, Weigelt B, Peterse H, Bartelink H, van de Vijver MJ. Gene expression profiling and histopathological characterization of triple-negative/basal-like breast carcinomas. Breast Cancer Res. 2007;9(5):R65. doi: 10.1186/bcr1771. - DOI - PMC - PubMed
    1. Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA. et al. Molecular portraits of human breast tumours. Nature. 2000;406(6797):747–752. doi: 10.1038/35021093. - DOI - PubMed
    1. Siziopikou KP, Cobleigh M. The basal subtype of breast carcinomas may represent the group of breast tumors that could benefit from EGFR-targeted therapies. Breast. 2007;16(1):104–107. doi: 10.1016/j.breast.2006.09.003. - DOI - PubMed
    1. Ryden L, Jirstrom K, Haglund M, Stal O, Ferno M. Epidermal growth factor receptor and vascular endothelial growth factor receptor 2 are specific biomarkers in triple-negative breast cancer. Results from a controlled randomized trial with long-term follow-up. Breast Cancer Res Treat. - PubMed
    1. Brennan DJ, Gallagher WM. Prognostic ability of a panel of immunohistochemistry markers - retailoring of an 'old solution'. Breast Cancer Res. 2008;10(1):102. doi: 10.1186/bcr1854. - DOI - PMC - PubMed

Publication types

MeSH terms

Substances

Feedback