Skip to main content

Genome-wide gene expression profiling of formalin-fixed paraffin-embedded breast cancer core biopsies using microarrays

Budczies J, Weichert W, Noske A, Muller BM, Weller C, Wittenberger T, Hofmann HP, Dietel M, Denkert C, Gekeler V

Publication: J Histochem Cytochem (2011) • Vol 59 • Page 146PMID: 21339180View on PubMed

Purpose

The purpose of this paper was to determine whether RNA from archival formalin-fixed paraffin-embedded (FFPE) breast cancer core needle biopsies can be used for genome-wide microarray gene expression profiling.

Conclusion

Classification of estrogen receptor (ER) and human epidermal growth factor receptor 2 (HER2) status based on RNA levels as determined by microarray analysis was feasible. However, unsupervised hierarchical clustering was not able to distinguish ER- and ER+ FFPE specimens, but when a selected subset of the genes was used, an almost perfect separation was achieved. Overlapping ER and HER2 gene expression signatures were found between FFPE core needle biopsies and frozen tissues.

Studies (1)

Purpose

The purpose of this study was to determine whether RNA from archival FFPE breast cancer core needle biopsies can be used for microarray analysis and classification of ER and HER2 status. Expression profiles of FFPE tissue were compared to a publicly available data set from frozen tissues using the same chip platform.

Summary of Findings

The 3-prime/5-prime ratios for ACTB and GAPDH using FFPE tissue were high compared to when frozen tissue specimens were used. RNA fragmentation curves for FFPE tissue specimens were steeper than those for frozen tissue. An average of 21% of transcripts were detected using RNA from FFPE specimens while 41% were detected from frozen specimens on the Affymetrix HG-U133 Plus 2.0 chips. The mean correlation between each pair of expression signatures for FFPE core biopsies was 0.56 while for frozen tissue specimens it was 0.80. Unsupervised hierarchical clustering was not able to distinguish ER- and ER+ FFPE specimens, but when a selected subset of the genes was used, an almost perfect separation was achieved. One of the 9 probe sets for ER was able to detect significant differences in expression between ER- and ER+ FFPE specimens, as determined by IHC and ISH. Using a cut-off of 10 for this probe set led to high specificity (100%) and sensitivity (85.7%) for ER status determination. One of the 3 probe sets for HER2 was able to detect significant differences in expression between HER2- and HER2+ FFPE specimens, as determined by IHC and ISH. Using a cut-off of 12 for this probe set led to high specificity (83.3%) and sensitivity (100%) for HER2 status determination. Significant overlaps of 63% and 5% of differentially expressed genes for ER+ and ER- tumors and HER2+ and HER2- tumors, respectively, were found between FFPE and frozen specimens.

  • Biospecimens

    TissueBreast
  • Preservation Types

    FormalinFrozen
  • Diagnoses

    NeoplasticCarcinoma
  • Platforms

    RNARNA amplification
    RNAIn situ hybridization
    RNADNA microarray
    ProteinImmunohistochemistry
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    Biospecimen PreservationType of fixation/preservation
    Formalin (buffered)
    Frozen
    DNA microarray SpecificTargeted nucleic acid
    ACTB
    ER
    GAPDH
    HER2
    STAT1
    In situ hybridization SpecificTargeted nucleic acid
    ER
    HER2
    Immunohistochemistry SpecificTargeted peptide/protein
    ER
    HER2
    Immunohistochemistry SpecificTechnology platform
    DNA microarray
    In situ hybridization