Immunoguided laser assisted microdissection techniques for DNA methylation analysis of archival tissue specimens

Abstract

Altered DNA methylation is a fundamental characteristic of carcinogenesis. The analysis of DNA methylation in tumor cells may help to better understand tumor pathogenesis and more importantly may be used as diagnostic tool with therapeutic consequences. To detect targets relevant in tumorigenesis, it is essential to separate neoplastic cells from nonneoplastic cells. An excellent method for isolating specific cells is laser-assisted microdissection (LAM). Target cell identification for immunoguided LAM (ILAM) requires immunohistochemistry (IHC). Yet, it is unclear whether IHC for ILAM influences DNA methylation. The goals of this study were to establish an optimized protocol for antigen retrieval and IHC of formalin-fixed paraffin-embedded (FFPE) specimens suitable for ILAM and to evaluate its effect on the DNA methylome using a high throughput array. Using ten archival FFPE specimens, we showed specific staining suitable for ILAM. Extracted DNA from microdissected cells of immunohistochemically or H&E-stained tissue sections showed identical DNA quality and a strong correlation (r = 0.94 to 0.98) for CpG target methylation of 1505 analyzed sites in a series of five paired samples. No differential methylation between H&E and IHC was detected in 1501 of 1505 CpG targets (99.7%; P < 0.05). These results demonstrate the validity and utility of the herein described protocol, which allows the application of ILAM for large-scale genomic and epigenetic analyses of archival tissue specimens.

Figure 1

Technical steps for immunohistochemistry of formalin-fixed paraffin-embedded tissue sections on membrane slides for subsequent laser-assisted microdissection and DNA methylation profiling.

Figure 2

CD30 staining of Hodgkin-Reed-Sternberg cells in a Hodgkin's lymphoma on a regular glass slide (A) and on a polyethylene naphthalate membrane slide (B). Both sections of the FFPE sample, either on glass slide or membrane slide, are shown without hematoxylin counterstaining and without cover slide; original magnification, ×40.

Figure 3

CD30⁺ Hodgkin-Reed-Sternberg cells in a Hodgkin's lymphoma before microdissection (A), selected for microdissection (B), and after microdissection (C) using the Leica LMD 6000; original magnification, ×63.

Figure 4

PCR products after PCR amplification of the β-globin gene using three different sets of primers. A: Detection of 152-bp amplicon. B: Detection of 268-bp amplicon. C: Detection of 676-bp amplicon. P indicates positive control; H, PCR of microdissected tumor cells stained with H&E; M, molecular weight marker; I, PCR of microdissected tumor cells stained with immunohistochemistry; N, negative control. PCR of controls, H, and I were performed as duplicates.

Figure 5

Illumina GoldenGate Methylation array of five individual cases. Each plot shows the correlation of target CpG methylation of DNA isolated from H&E-stained tumor cells (x axis) compared with DNA isolated from immunohistochemically stained tumor cells (y axis). A minimum correlation of r = 0.94 and a maximum correlation of r = 0.98 was observed.