Towards quantitative mRNA analysis in paraffin-embedded tissues using real-time reverse transcriptase-polymerase chain reaction: a methodological study on lymph nodes from melanoma patients

Abstract

Improved extraction techniques combined with sensitive real-time reverse transcriptase-polymerase chain reaction may allow detection of mRNA in formalin-fixed, paraffin-embedded (FFPE) materials, but the factors affecting mRNA quantification in clinical material using these methods have not been systematically analyzed. We designed analyses using real-time reverse transcriptase-polymerase chain reaction for quantification of MART-1, beta-actin, and beta(2)-microglobulin mRNAs. The analytical intra- and interassay imprecision (coefficient of variation) was in the range 10 to 20% for all three genes studied. Using these protocols, we studied the influence of tissue autolysis and length of formalin-fixation on mRNA detection in metastatic melanoma. Delay in freezing reduced detectable mRNA, although this was less than predicted and mostly occurred early in autolysis. MART-1, beta-actin, and beta(2)-microglobulin mRNAs were consistently detected in FFPE metastatic melanoma even after fixation for up to 3 weeks, although the total mRNA detected was markedly reduced in fixed compared with fresh tissues (up to 99%). Quantification of MART-1 was, however, possible if this was expressed relative to a housekeeping gene. The polymerase chain reaction product from FFPE tissues could be increased up to 100-fold amplifying short (<136 bp) compared with long amplicons. Variations in time before tissue processing and in fixation length seem to be less important sources of imprecision than previously assumed. Our findings suggest that quantitative analysis of mRNA in archive and routine diagnostic tissues may be possible.

Figure 1.

Schematic overview of the sampling protocol for four experiments of prefreezing delay (A) and experiments of fixation time (B). Duplicate RT-measurements were performed on each sample.

Figure 2.

Quantification of β-actin (amplicon size, 99 bp; A) and β₂-M (amplicon size, 85 bp; B) in samples frozen after prolonged storage at room temperature. Each point represents the mean of duplicate measurements of paired samples obtained from melanoma metastases from four different patients (n = 4). Data are expressed as percentage of mRNA normalized to point 0 hour. Vertical lines represent ± SEM.

Figure 3.

Effect of delay in freezing on the expression of mRNA for MART-1. Ten samples from each of four different patients (n = 4) were obtained from melanoma lymph node metastases. Data are expressed as a ratio between arbitrary units of MART-1 mRNA and β-actin mRNA. The measurement of MART-1 and β-actin mRNA levels are described in the Materials and Methods. Amplicon sizes for MART-1 and β-actin were 136 bp and 99 bp, respectively. Each point represents the mean of paired samples measured in duplicate. Vertical lines represent minimum and maximum measurement.

Figure 4.

The effect of different amplicon sizes and different fixation times on the quantification of mRNA for β-actin. Results shown are means of duplicate measurements of samples obtained from two melanoma lymph node metastases (I and II). Data are expressed as percentage of β-actin mRNA present in the corresponding frozen tissue (99 bp).

Figure 5.

The effect of different fixation times and different amplicon sizes on the quantification of MART-1. Mean of duplicate measurements of two (I and II) melanoma lymph node metastases are expressed as a ratio between arbitrary units of mRNA for MART-1 and mRNA for β-actin (amplicon size, 99 bp). Data are expressed in percent normalized to matched frozen sample (99 bp).