Maintaining RNA integrity in a homogeneous population of mammary epithelial cells isolated by Laser Capture Microdissection

Abstract

Background: Laser-capture microdissection (LCM) that enables the isolation of specific cell populations from complex tissues under morphological control is increasingly used for subsequent gene expression studies in cell biology by methods such as real-time quantitative PCR (qPCR), microarrays and most recently by RNA-sequencing. Challenges are i) to select precisely and efficiently cells of interest and ii) to maintain RNA integrity. The mammary gland which is a complex and heterogeneous tissue, consists of multiple cell types, changing in relative proportion during its development and thus hampering gene expression profiling comparison on whole tissue between physiological stages. During lactation, mammary epithelial cells (MEC) are predominant. However several other cell types, including myoepithelial (MMC) and immune cells are present, making it difficult to precisely determine the specificity of gene expression to the cell type of origin. In this work, an optimized reliable procedure for producing RNA from alveolar epithelial cells isolated from frozen histological sections of lactating goat, sheep and cow mammary glands using an infrared-laser based Arcturus Veritas LCM (Applied Biosystems®) system has been developed. The following steps of the microdissection workflow: cryosectioning, staining, dehydration and harvesting of microdissected cells have been carefully considered and designed to ensure cell capture efficiency without compromising RNA integrity.

Results: The best results were obtained when staining 8 μm-thick sections with Cresyl violet® (Ambion, Applied Biosystems®) and capturing microdissected cells during less than 2 hours before RNA extraction. In addition, particular attention was paid to animal preparation before biopsies or slaughtering (milking) and freezing of tissue blocks which were embedded in a cryoprotective compound before being immersed in isopentane. The amount of RNA thus obtained from ca.150 to 250 acini (300,000 to 600,000 μm2) ranges between 5 to 10 ng. RNA integrity number (RIN) was ca. 8.0 and selectivity of this LCM protocol was demonstrated through qPCR analyses for several alveolar cell specific genes, including LALBA (α-lactalbumin) and CSN1S2 (αs2-casein), as well as Krt14 (cytokeratin 14), CD3e and CD68 which are specific markers of MMC, lymphocytes and macrophages, respectively.

Conclusions: RNAs isolated from MEC in this manner were of very good quality for subsequent linear amplification, thus making it possible to establish a referential gene expression profile of the healthy MEC, a useful platform for tumor biomarker discovery.

Figure 1

Impact of the freezing system on morphology of fresh mammary tissue sections and on quality of RNA extracted. Alveolar (acini) structures lined by MECs (yellow arrows) can be easily distinguished after staining mammary tissue sections with Cresyl violet AMBION. Immediately after collection, mammary tissue samples were washed in cold PBS solution, cut in cube of 3-4 mm thickness, and frozen in four different conditions: pieces of mammary tissue were either directly introduced into 1.5-ml eppendorf tubes and frozen in liquid nitrogen (A), or in cold isopentane (-80°C) using the SnapFrost™ system (C), or embedded in OCT^® contained in cryomold before to be immediately immerged in liquid nitrogen (B) or in cold isopentane (D). RNA quality (RIN) which was estimated by RNA 6000 Pico LabChip kit and Agilent 2100 Bioanalyzer, was identical (ranging between 8.5 and 9.5) whatever the freezing procedure, as illustrated in the electrophoreris profiles. Some large blisters (red arrows on Figure A and B) appear however in biopsy flash frozen in liquid nitrogen, mainly without cryoprotector, and morphological details are better seen on tissues frozen using the SnapFrost™ system (Magnification: ×60). The green arrow indicates the thermoplastic film stuck on epithelial cells to be captured.

Figure 2

Impact of glass slide temperature on cell transfer efficiency and RNA extraction yield. Frozen tissues cut at 8 μm and transferred on glass slides that have been placed at room temperature (A) or chilled and kept at 4°C, before transfer (D). As evidenced by the number of cells captured (C and F) and the RNA yield, given under each cap (C and F) the efficiency of capture was very low for the slides placed at room temperature. A/D: Cresyl violet^® stained slides before LCM; red arrows indicate the thermoplastic film stuck on epithelial cells to be captured. B/E: mammary tissue section remaining on the glass slide after LCM. C/F: microdissected mammary epithelial cells transferred on the LCM cap. Magnification: ×60.

Figure 3

Impact of histological stain and fixative/protector on morphology of goat mammary tissue and RNA integrity. Frozen mammary tissue sections were stained either with HistoGene^® (A to C) or with Cresyl violet^® (D and E). MEC isolated after treatment with RCL2, a new fixative preserving tissue morphology and Nucleic Acid integrity (C and D), or with RNA later (B) provide RNA of better quality, expressed as ΔRIN which is difference between the RIN value obtained with total RNA extracted from the mammary tissue section (8.5) and the RIN value obtained with RNA extracted from microdissected cells.

Figure 4

Selectivity of MEC capture assessed by real-time quantitative PCR. To estimate contamination of the laser-captured MEC by adjacent MMC as well as by immune cells (macrophages and lymphocytes) mRNA transcript levels of cell-specific markers were assessed by measuring the relative expression of the relevant genes between captured cells and mammary tissue scrapes, i.e. the mammary tissue remaining on slides, after LCM. Relative quantities of specific markers from MMC (Krt14), macrophages (CD68), lymphocytes (CD3e) and MEC (CSN1S2, LALBA) were assessed after normalization (PPAI and/or RPS24). FASN, a gene expressed in a large panel of cell types, including adipocytes and MEC, was also assessed in the same conditions. Mean RQ values are given for each gene. Krt14 decreased (RQ = 0.14; ca. 7-folds reduction) in the captured cells compared with the whole mammary tissue. Likewise the same ratio was observed with CD68 (RQ mean = 0.18) and CD3e (RQ mean = 0.14). Levels of LALBA (RQ mean = 1.47 fold), CSN1S2 (RQ mean = 1.33 fold) were significantly increased. *indicates ratio significantly different between cap and scraped tissue (p < 0.001).