Accurate Gene Expression Measurement in Formalin-Fixed and Paraffin-Embedded Tumor Tissue

To the Editor-in-Chief:

In the February 2001 issue of The American Journal of Pathology, Specht and colleagues 1 have published technical advances in their article entitled “Quantitative Gene Expression Analysis in Microdissected Archival Formalin-Fixed and Paraffin-Embedded Tumor Tissue.” 1 In contrast to earlier published data, where RNA extraction and subsequent RT-PCR from formalin-fixed and paraffin-embedded tissue (FFPE) has been reported to be cumbersome, 2 Specht et al 1 present an optimized protocol for RNA extraction from FFPE tissue, followed by the powerful TaqMan quantitative RT-PCR methodology. 3 Major improvement of RNA extraction is achieved by prolonged proteinase K digestion at 60°C. In addition, the choice of very short amplicons, tolerating a high degree of RNA degradation, represents a major breakthrough.

In a different system, we have observed a high tolerance of TaqMan measurements towards RNA degradation. Human colon tissue was incubated at room temperature from 10 minutes up to 4 hours prior to RNA extraction. At up to 90 minutes, incubation at room temperature revealed no detectable degradation of RNA as indicated by a ratio of 28S:18S ≥ 2 evaluated by the RNA 6000 LabChip kit with the Agilent 2100 Bioanalyzer. Degradation was observed after 120 minutes by a decreased ratio of 28S:18S of 1.4. Thereafter, the ratio further decreased to zero and accumulation of short RNA fragments was detected. However, the quantitative determination of the amount of GAPDH, β-actin, c-Myc, and Fra-1 expression using the TaqMan technology did not alter over the complete time range. Amplified amplicons were 69 to 83 base pairs in size. 4

Specht et al 1 have chosen an HT29 and A431 xenograft model to test relative gene expression of various genes in adjacent lying frozen and FFPE tumor halves. Whereas several genes showed no difference in expression levels as compared between frozen and FFPE, the levels of FGF-R4 and of EGF-R varied significantly. These differences seem to occur in a non-predictable manner. On the other hand, we observed a clear gene expression difference in tissues with low-quality RNA (eg, FFPE) compared to frozen tissue, especially when genes with low expression levels were compared.

Furthermore, Specht et al 1 have tested the influence of tissue thickness on fixation and RNA degradation. They stated that in tissue thickness of up to 7 cm no differences of RNA expression levels were found, irrespective of whether measurements were carried out at either the surface or at pre-defined levels (1 cm, 2 cm, etc.) inside the tissue. A rule of thumb states that tissue thickness should not exceed 5 mm in at least in one spatial dimension to allow proper fixation. 5 However, the authors do not indicate whether the given tissue thickness addresses all three dimensions (eg, 7 × 7 × 7 cm³) or only one. Therefore, it remains speculative whether the presented data reflect an unfavorable or a rather favorable fixation condition.

Laser microdissection represents a powerful tool to study gene expression in a histomorphological context. The opportunity to investigate archival FFPE tissue would allow one to take advantage of the huge amount of tissue samples stored in pathological institutes. Specht et al 1 investigated the expression of HER-2/neu mRNA in FFPE esophageal adenocarcinomas. Tumors having a HER-2/neu amplification and a 3+ EGFR immunohistochemistry were microdissected and HER-2/neu mRNA was quantitated. The data showed a large variability in HER-2/neu mRNA quantity. The large variation in HER-2/neu mRNA expression may reflect the heterogeneous mRNA expression levels throughout a tumor specimen, where HER-2/neu positive cell clusters have been arbitrarily microdissected and analyzed. However, it cannot be excluded that fixation parameters such as fixation delay, time, and temperature may account for the large expression variability. Frozen sections were not included in the study to test differences due to fixation parameters.

Taken together, the authors present very important improvements for the RNA extraction from FFPE tissue and subsequent quantitation using TaqMan methodology. We still think that a fully controlled standardization of tissue fixation and processing, including testing for RNA quality prior to qualitative analysis, is a prerequisite for accurate gene expression measurement in FFPE tissue and at the same time insures comparative immunohistochemistry analysis.